Identity Formation at the Keffer Iroquoian Village: A Relational Network Approach to Communities of Practice by Susan Dermarkar A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Graduate Department of Anthropology University of Toronto © Copyright by Susan Dermarkar 2019 Identity Formation at the Keffer Iroquoian Village: A Relational Network Approach to Communities of Practice Susan Dermarkar Doctor of Philosophy Graduate Department of Anthropology University of Toronto 2019 Abstract This thesis employs settlement pattern studies and social network analysis of ceramics to present a history of the Keffer village and its multiple ceramic practice communities. Keffer is a fifteenth-century Iroquoian site on the north shore of Lake Ontario. Temporally and spatially patterned variability in the diverse local tradition and nonlocal ceramics materialized the fluid, multivalent, and contingent identity of the Keffer potters at the household level. Nonlocal ceramics, particularly locally produced “emergent” ceramics, expressed unique yet shared identities with others within the village. These same analyses produce a nuanced series of development plans that trace the initial settlement, the growth, and the initial steps in the abandonment of the village. Distinct ceramic practices indicate that, contrary to what has been argued in some of the literature, the longhouse clusters that form the initial village occupation may not reflect kin-based social groups. However, longhouses added independently after this first settlement may reflect the arrival of small, related, kin-based groups. Village growth appears to have occurred at the level of the individual household and was likely unplanned. Village abandonment occurred on the same scale, at the individual household level. ii This study also shows that social network analysis of ceramic types is an appropriate mechanism for the exploration of multiscalar group identity. Analysis at the pan-Iroquoian, regional, and village levels demonstrates Keffer’s central position in the highly connected north shore interaction network and its distant relations with St. Lawrence and Eastern Iroquoian groups. Social relations and the identities they sustained, from the most intimate level of the household to the pan-cultural interaction sphere, can be successfully revealed and explored through the application of combined social network, community of practice, and agency approaches. iii Acknowledgments This thesis has been a long-term project I could never have accomplished without the help of a large number of people. First off, I would like to thank Suzanne Needs-Howarth, my copy editor and formatter, for all the support and friendship she has provided. Without Suzanne’s incredible skills I could not have completed this very complicated task. The person most instrumental in getting this project off the ground is my supervisor at the University of Toronto at Mississauga, Dr. David Smith. Dave welcomed me as a Ph.D. student 20 years after I finished my Master’s degree. He has continued to be excited about this project and my research over the years. It was Dave’s desire to see the Keffer site get the attention it deserved that led to this thesis. I would like to thank the members of my committee, Dr. Edward Swenson and Dr. Edward Banning, for providing guidance on theoretical and methodological aspects of my research. Dr. Swenson helped me with my understanding of identity, and Dr. Banning introduced me to the exciting field of social network analysis. Their input has had an immense effect on the direction of the project. My association with Dave Smith and University of Toronto at Mississauga was made possible by Dr. Heather Miller, who welcomed me to audit her courses and introduced me to the department. Before this, Dr. Paul Healy, Dr. Jeffrey Seibert, and Dr. James Conolly of Trent University helped me update my archaeological theory and culture history knowledge by allowing me to audit their graduate- and undergraduate-level classes at Trent University. James Connelly and Dr. Andrew Bevan, of University College London, welcomed me to do observer bias research on their Antikythera crew. Dr. Helen Haines, of Trent University, provided me with the opportunity to renew my love of ceramics on the 2011 Ka’Kabish project. Gaining access to, and temporary possession of, the Keffer material was made possible by the Ontario Museum of Archaeology in London, the collection steward. I would like to thank Dr. Robert Pearce for facilitating this loan and that of the Forget materials from the museum. When I was looking for a suitable contemporary ceramic collection, I had the unfailing support of Dr. Dean Knight of Wilfrid Laurier University. Dean provided me with use of his university lab space for an entire year, as well as access to the Baumann ceramic collection. He also welcomed iv me on the final year of the Baumann site excavations and supplied any data I requested. Over the past 35 years, Dean has given me friendship and advice that I will always be grateful for. I would like to thank Dr. Ronald Williamson, who has been very supportive of my research on the Don valley. Ron has always made time to meet with me and discuss ideas. He and the team at Archaeological Services Inc. have provided me with any data and any archaeological reports, published or unpublished, that I have requested. Robert von Bitter and the staff at the Ontario Ministry of Tourism, Culture and Sport deserve thanks for providing me with access to site reports and archived material. My research has also led me farther afield to look at research collections. I would like to thank James Hunter, then at the Midland Museum, for allowing me access to the collections; April Hawkins of the Royal Ontario Museum, for being interested in my work and giving me support with research into the Risebrough collection, held at the museum; the staff of the Toronto and Region Conservation Authority, for providing access to the Seed-Barker reports; and Kathy David and the Department of Anthropology, University of Toronto St. George campus, for providing me with lab space to view the university’s ceramic collections from the Black Creek, Parsons, and Risebrough sites. Stacey Girling and the staff of the Canadian Museum of History were very helpful with my research into the Draper, Keffer, Jarrett-Lahmer, and Salem sites ceramics housed in Gatineau. I would particularly like to thank Michel Savard at the Musée huron-wendat in Wendake, Quebec, for access to the Keffer and Forget collections there and the warm welcome that the museum afforded me. Dr. Brandi-Lee MacDonald, now at the University of Missouri, did me a great favour and analyzed the Keffer site copper artifacts at short notice. I would like to thank Dr. Jennifer Birch of the University of Georgia, who produced the crucial longhouse post mould density report, for all the support she has given me over the past decade. Dr. Harry Lerner of McGill University, Dr. Bruce Jamieson, and Dr. Michael Spence of Western University have willingly provided me with information on their research on the Keffer site lithics, bone artifacts, and human burials, respectively. Dr. Peter Ramsden has provided ceramic data from the Trent Valley Project. I appreciate his generosity and our discussions. Nicholas Adams provided me with his ceramic data and photographs of the Arbor Ridge site, Dana Poulton, of D.R. Poulton and Associates, was helpful with advice and access to reports, and Gordon Dibb, of North York Archaeological v Services, gave me access to the Snodden and Trent Foster collections and welcomed me in his home. Researchers in New York state have been very supportive of my efforts to become familiar with ancestral Haudenosaunee ceramics and culture. I would like to express my appreciation to the staff at the New York State museum in Albany for providing access to their precontact ceramic collections, and I would like to thank Dr. John Hart of the museum for his academic support over the past few years. Wayne Lenig has helped me on many occasions in my struggles to understand the history of the Otstungo Notched ceramic type. I sincerely appreciate his knowledge and advice. Dr. Timothy Abel, of Jefferson Community College, has welcomed me into his home and freely shared his research on and passion for northern New York state archaeology. Dr. William Engelbrecht, of the State University of New York College at Buffalo, who originally introduced me to Haudenosaunee ceramics, has been a wonderful friend. Bill is always ready to share his vast knowledge of ceramics and Iroquoian archaeology. During my time in the research lab at University of Toronto at Mississauga, I was honoured to have the volunteer help of many students. Rachel Gruber, while working on the Keffer pipe coding, put in more time and effort than could ever have been expected. Erika Johannsson provided the first preliminary study of the Keffer site ceramic castellations, and Andreas Vastikas was a great help with the Forget collection and the photography of the Keffer ceramics. I would like to thank Robert and Elizabeth Dermarkar for the weeks they spent photographing the collection. Elizabeth also spent endless hours producing illustrations and maps. To all these people, I extend my thanks. Many other scholars, too many to mention, have helped me in innumerable ways. I thank them all. Financial support for many aspects of the Keffer site analysis was provided by a University of Toronto Graduate Studies Research Grant, for which I am greatly appreciative. Last, but not least, I would like to thank all the family and friends who have supported me over the years. My co-Scouters in Enniskillen and beyond helped carry my load by taking over many of my scouting duties. My sisters pushed me to continue. My children, Liz, Jim, and Bob, and my friend Quinn Zavagno helped me with computer programs I never learned in school. I want to thank my friends Margot and Steve, who housed and fed me for a year; Ann, Liz, Mara, and many others, who gave me tea; and numerous others, who understood and tried not to interrupt vi my work. I send my love and appreciation to my mother and mother-in-law, who have hardly seen me in the past few years. Above all these wonderful people, I thank my husband for years of understanding and support. vii Table of Contents Acknowledgments.......................................................................................................................... iv List of Tables ................................................................................................................................ xii List of Figures .............................................................................................................................. xiv List of Appendices ..................................................................................................................... xxiv 1 Introduction .................................................................................................................................1 Situating the Keffer Village ........................................................................................................1 Chapter Summary........................................................................................................................4 2 Theoretical Underpinning: The Search for Identity ....................................................................7 Identity: From Ethnicity to Active Community ..........................................................................7 Introduction of the Human Factor through Agency Theory .....................................................11 Symbols and Boundaries ...........................................................................................................12 Messaging of Social Distance ...................................................................................................12 Identity and Communities of Practice .......................................................................................22 Interaction and Social Network Analysis ..................................................................................26 3 Background ...............................................................................................................................31 Iroquoian Development .............................................................................................................31 Late Woodland I: Princess Point, A.D. 500–1000 .............................................................34 Late Woodland II: Early Ontario Iroquoian, A.D. 1000–1300 ..........................................34 Late Woodland III: Middle Iroquoian, A.D. 1300–1420 ...................................................35 Late Woodland III: Pre-Contact Iroquoian, A.D. 1420–1550 ...........................................37 Keffer ........................................................................................................................................41 Keffer Site Environment ....................................................................................................41 Early Research at the Keffer Site .......................................................................................47 Keffer Site Excavation, 1984 and 1988 .............................................................................49 viii 4 Method: Ceramic Type and Attribute Analysis ........................................................................54 Ceramic Method and Theory ....................................................................................................54 History of Research on the Keffer Ceramics ............................................................................60 General Procedures Followed in the Current Analysis .............................................................61 Iroquoian Ceramic Typology at Keffer .....................................................................................63 Local Tradition Types and Attributes .......................................................................................65 “Huron” Types ...................................................................................................................67 Neutral Area Types ............................................................................................................74 High Collar Types ..............................................................................................................76 Erie–Niagara Frontier Undecorated Types ........................................................................80 Early Ontario Iroquoian Types ..........................................................................................81 Nonlocal Tradition Types and Attributes ..................................................................................84 Ceramics of the Central and Eastern Iroquois ...................................................................86 Ceramic Types of the St. Lawrence and Northern New York Iroquoians .........................95 Distinctive Ceramic Attributes and Emergent Vessels at Keffer.......................................99 Exotic Attributes Seen at Keffer ......................................................................................109 Nonlocal Learner Vessels and Practice Vessels at Keffer ...............................................112 Discussion ...............................................................................................................................114 5 Social Network Analysis: Keffer in the Iroquoian World ......................................................117 Social Network Analysis of Pan-Iroquoian Ceramic Practice ................................................117 Social Network Analysis of Local Tradition Ceramics from 140 Ontario Iroquoian Sites ....121 Ontario Iroquoian Ceramic Relations as Seen Through Network Analysis ...........................122 Multiscalar Relations of the Southern Ontario Iroquoian and Keffer Site Ceramics .............125 6 Settlement Pattern Analysis ....................................................................................................130 Spatial Arrangement of Major Settlement Features ................................................................130 Topography .............................................................................................................................133 ix Overlapping and Sharing of Features ......................................................................................137 Definition of Site Clusters.......................................................................................................138 Definition of Site Settlement Phases .......................................................................................141 Site Settlement Phase 1a .........................................................................................................142 Phase 1a, Cluster A ..........................................................................................................144 Phase 1a, Cluster B ..........................................................................................................147 Phase 1a, Cluster C ..........................................................................................................150 Phase 1a, Cluster D ..........................................................................................................153 Site Settlement Phase 1b .........................................................................................................157 Phase 1b, Cluster A ..........................................................................................................157 Phase 1b, Cluster B ..........................................................................................................160 Phase 1b, Cluster C ..........................................................................................................160 Phase 1b, Cluster D ..........................................................................................................162 Site Settlement Phase 2a .........................................................................................................163 Phase 2a, Cluster A-B-D North .......................................................................................165 Phase 2a, Cluster C ..........................................................................................................167 Site Settlement Phase 2b .........................................................................................................169 Phase 2b-2, Cluster A-B-D North ....................................................................................171 Phase 2b-2, Cluster D South ............................................................................................171 Phase 2b-2, Cluster C .......................................................................................................172 Summary of Site Settlement Phases ........................................................................................173 Phase 1a ...........................................................................................................................174 Phase 1b ...........................................................................................................................175 Phase 2a ...........................................................................................................................175 Phase 2b ...........................................................................................................................176 Discussion ...............................................................................................................................176 x 7 Analysis of Keffer Site Local Tradition Ceramics ..................................................................177 Factors Involved in Keffer Site Network Analysis .................................................................177 Social Network Analysis of Local Tradition Ceramic Chronology ........................................181 Application of Local Tradition Ceramics Social Network Analysis Stages to Settlement Phases ...............................................................................................................................182 Summary .................................................................................................................................184 Constructing a Community .....................................................................................................185 Village Settlement Overview ..................................................................................................194 Social Network Analysis of Local Tradition Communities of Practice ..................................195 Discussion ...............................................................................................................................198 8 Communities of Practice at Keffer ..........................................................................................199 The Emergence of Keffer’s Ceramic Practices .......................................................................199 Chronological Development of the Communities of Practice ................................................209 External Relations of Keffer’s Communities of Practice: A Social Network Analysis ..........219 Overview ..........................................................................................................................220 Communities and Practice ...............................................................................................229 External Ties by Time .............................................................................................................235 Social Network Analysis of Nonlocal Tradition Ceramics at Keffer .....................................238 Nonlocal Ceramic Practice at Keffer ......................................................................................249 Shared Local and Nonlocal Practice at Keffer ........................................................................260 Discussion ...............................................................................................................................265 9 Conclusions .............................................................................................................................268 References ....................................................................................................................................279 Appendices ...................................................................................................................................312 xi List of Tables Table 6-1. House and Extension Length(s) and Wall Post Density. ........................................... 130 Table 6-2. Houses and Middens Associated with Each Site Cluster and Elevation. .................. 141 Table 6-3. Possible Concurrent Houses of Phase 1a Clusters. ................................................... 144 Table 6-4. Possible Concurrent Houses of Phase 1b Clusters. ................................................... 159 Table 6-5. Possible Concurrent Houses of Phase 2a Clusters. ................................................... 165 Table 6-6. Possible Concurrent Houses of Phase 2b Clusters. ................................................... 170 Table 6-7. Houses, Middens, and Palisades Associated with Each Cluster of Each Occupation Phase. .......................................................................................................................................... 174 Table 7-1. Houses and Middens Included at BR 170 in Each of the Seven Node Cluster Groups for Each Sample Size. ................................................................................................................. 179 Table 7-2. Average Tie Densities Within and Between Individual Clusters, the Combined Interior Samples, and the Combined Exterior Samples. ............................................................. 180 Table 7-3. Houses and Middens Included in Each Second-Level Temporal Stage Based on Social Network Analysis and Settlement Studies. ................................................................................. 183 Table 7-4. Proportion of Local Tradition Ceramic Types by Second-Level Stage, Based on Social Network Analysis and Settlement Studies. ...................................................................... 187 Table 7-5. Settlement Changes in Maps 1–8. ............................................................................. 190 Table 7-6. Houses and Middens Associated with Each Local Tradition Community of Practice. ..................................................................................................................................................... 197 Table 8-1. Local Tradition Ceramic Percentage Ranges by Community of Practice. ................ 206 Table 8-2. Community of Practice by Local Tradition Ceramic Percentage Ranges. ................ 207 xii Table 8-3. Loci by Map and Local Community of Practice. ...................................................... 209 Table 8-4. Temporal Order of Communities of Practice. ........................................................... 229 Table 8-5. Keffer House and Midden Similarity Ties with Other Ontario Iroquoian Sites........ 236 Table 8-6. Network Ties of Nonlocal Practice. .......................................................................... 251 Table 8-7. Shared Local and Nonlocal Ceramic Practice. .......................................................... 260 Table 8-8. Summary of Ceramic Practices at Keffer. ................................................................. 262 xiii List of Figures Figure 3-1. The geographical extent of Northern Iroquoia. .......................................................... 31 Figure 3-2. The Woodland period in south-central Ontario. ........................................................ 33 Figure 3-3. Location of the Keffer site in the Don River watershed. (modified from Toronto and Region Conservation Authority ca. 2008) .................................................................................... 43 Figure 3-4. The bi-level plateau of the Keffer site. (Ontario Museum of Archaeology) .............. 44 Figure 3-5. Glacial deposits in the vicinity of the Keffer site. (modified from Ontario Geological Survey 1980) ................................................................................................................................. 45 Figure 3-6. Soils in the vicinity of the Keffer site. (modified from Agriculture Canada 1955) ... 46 Figure 3-7. Map of the Keffer village site. (Clark 1925) .............................................................. 48 Figure 3-8. Aerial view of Lot 13, 1985 excavations, and wooded Lot 12 bottom right. ............ 50 Figure 4-1. Keffer site location within pan-Iroquoia A.D. 1350 to 1650. (David Smith) ............ 59 Figure 4-2. Keffer ceramic object chemistry principal components analysis of major elements. (courtesy John Creese) .................................................................................................................. 64 Figure 4-3. Huron Incised vessel, Midden 60, Keffer. ................................................................. 67 Figure 4-4. Sidey Notched vessel, Midden 65, Keffer. ................................................................ 69 Figure 4-5. Copeland Incised motif on Black Necked vessel, Midden 60, Keffer. ...................... 69 Figure 4-6. Sidey Crossed vessel, Midden 65, Keffer. ................................................................. 69 Figure 4-7. Warminster Crossed vessel, Midden 60, Keffer. ....................................................... 70 Figure 4-8. Seed Incised vessel, Midden 77, Keffer. .................................................................... 71 Figure 4-9. Warminster Horizontal vessel, surface find, Keffer. .................................................. 71 xiv Figure 4-10. Black Necked vessel, Midden 65, Keffer................................................................. 73 Figure 4-11. Black Necked vessel with horizontal motif, Midden 65, Keffer.............................. 73 Figure 4-12. Lawson Incised vessel, Midden 57, Keffer. ............................................................. 75 Figure 4-13. Lawson Incised vessel, Midden 57, Keffer. ............................................................. 75 Figure 4-14. Lawson Opposed vessel, Midden 74, Keffer. .......................................................... 76 Figure 4-15. Sopher High Collar vessel, House 19, Keffer. ......................................................... 77 Figure 4-16. Lalonde High Collar vessel, Midden 57, Keffer. ..................................................... 78 Figure 4-17. Lalonde High Collar learner vessel, Midden 65, Keffer. ......................................... 78 Figure 4-18. Lalonde High Collar type variant vessel, Midden 57, Keffer. ................................. 79 Figure 4-19. Lalonde High Collar type variant vessel, Midden 57, Keffer. ................................. 79 Figure 4-20. Niagara Collared vessel, House 12, Keffer. ............................................................. 80 Figure 4-21. Ripley Collared vessel, Midden 65, Keffer. ............................................................. 81 Figure 4-22. Ripley Plain vessel, surface find, Keffer. ................................................................. 81 Figure 4-23. Middleport Oblique vessel, House 6, Keffer. .......................................................... 82 Figure 4-24. Pound Necked vessel, Midden 65, Keffer. ............................................................... 82 Figure 4-25. Pound Blank variant vessel, House 1, Keffer. ......................................................... 83 Figure 4-26. Ontario Horizontal vessel, House 1, Keffer. ............................................................ 83 Figure 4-27. Ontario Oblique vessel, Midden 52, Keffer. ............................................................ 83 Figure 4-28. Cayadutta Incised vessel, Midden 70, Keffer. ......................................................... 87 Figure 4-29. Chance Incised vessel, Midden 57, Keffer............................................................... 88 xv Figure 4-30. Fonda Incised vessel, Midden 65, Keffer................................................................. 88 Figure 4-31. Otstungo Incised vessel, House 12, Keffer. ............................................................. 89 Figure 4-32. Onondaga Triangular vessel, Midden 85, Keffer. .................................................... 89 Figure 4-33. Cayuga Horizontal vessel, Midden 61, Keffer. ........................................................ 90 Figure 4-34. Thurston Horizontal vessel, Midden 65, Keffer....................................................... 90 Figure 4-35. Cayuga Horizontal vessel, House 4, Keffer. ............................................................ 91 Figure 4-36. Richmond Incised vessel, House 11, Keffer. ........................................................... 91 Figure 4-37. Syracuse Incised vessel, Midden 57, Keffer. ........................................................... 92 Figure 4-38. Wagoner Incised vessel, Midden 61, Keffer. ........................................................... 92 Figure 4-39. Otstungo Notched vessel, Midden 65, Keffer. ......................................................... 93 Figure 4-40. Rice Diagonal vessel, Midden 62, Keffer. ............................................................... 93 Figure 4-41. Dutch Hollow Notched vessel, Midden 56, Keffer. ................................................. 94 Figure 4-42. Seneca Notched vessel, Midden 65, Keffer. ............................................................ 94 Figure 4-43. SDEA motif (Smith 1997) outline on Durfee Underlined vessel, Draper. (Canadian Museum of History) ...................................................................................................................... 96 Figure 4-44. Roebuck Low Collar vessel, Midden 60, Keffer...................................................... 97 Figure 4-45. Stamped Low Collar vessel, Midden 65, Keffer. ..................................................... 97 Figure 4-46. Roebuck Low Collar vessel with SDEA motif, Midden 73, Keffer. ....................... 98 Figure 4-47. Durfee Underlined vessel with SDEA motif, Midden 60, Keffer. ........................... 99 Figure 4-48. Otstungo Notched vessel, Salem Cat. #697. (Canadian Museum of History) ....... 100 Figure 4-49. Otstungo Notched vessel, Maynard-McKeown. (Canadian Museum of History) . 101 xvi Figure 4-50. Otstungo Notched vessel, unknown provenience, New York state. (courtesy Wayne Lenig) .......................................................................................................................................... 102 Figure 4-51. Otstungo Notched vessel, Draper Cat. # 41454. (Canadian Museum of History) . 103 Figure 4-52. Otstungo Notched vessel, Parsons Cat. # 3357. (University of Toronto) .............. 103 Figure 4-53. Otstungo Notched vessel, Damiani. (courtesy Ron Williamson) .......................... 104 Figure 4-54. Otstungo Notched vessel, Mantle. (ASI 2006:Plate 37, Cat. #10432) ................... 104 Figure 4-55. Rice Diagonal vessel, Salem Cat. #699. (Canadian Museum of History) ............. 105 Figure 4-56. Rice Diagonal vessel, Cleary Cat. #011. (courtesy David Smith) ......................... 105 Figure 4-57. Rice Diagonal vessel, Lite. (Canadian Museum of History) ................................. 106 Figure 4-58. Rice Diagonal vessel, Draper Cat. #39180. (Canadian Museum of History) ........ 106 Figure 4-59. Durfee Underlined vessel with SDEA motif, Draper Cat. #114617. (Canadian Museum of History) .................................................................................................................... 107 Figure 4-60. Durfee Underlined vessel with SDEA motif, Jarrett-Lahmer Cat. #22897. (Canadian Museum of History) .................................................................................................................... 108 Figure 4-61. Roebuck Low Collar vessel with SDEA motif, Black Creek. (University of Toronto) ..................................................................................................................................................... 108 Figure 4-62. Durfee Underlined vessel with SDEA motif, Parsons Cat. #460 (University of Toronto) ...................................................................................................................................... 109 Figure 4-63. Northern New York Iroquoian vessel, St. Lawrence Site (Abel 2000:Plate 2). .... 109 Figure 4-64. “Exotic” Durfee Underlined vessel, Midden 60, Keffer. ....................................... 110 Figure 4-65. Vessels with poorly executed thumbnail gouges, Midden 65, Keffer. .................. 111 Figure 4-66. Localized Onondaga Triangular vessel, Midden 77, Keffer. ................................. 111 xvii Figure 4-67. Otstungo Notched/Rice Diagonal Everted Rim collarless learner vessel, Midden 60, Keffer. ......................................................................................................................................... 113 Figure 4-68. Durfee Underlined learner vessel, Midden 61, Keffer. .......................................... 113 Figure 4-69. Wagoner Incised learner vessel, Midden 74, Keffer. (courtesy Steven Dorland) .. 114 Figure 5-1. Pan-Iroquoian network BR 100 spring layout. ........................................................ 118 Figure 5-2. Pan-Iroquoian network BR 140 spring layout. ........................................................ 118 Figure 5-3. Pan-Iroquoian network t2–t3 BR 93 spring layout. ................................................. 120 Figure 5-4. Local tradition ceramics of Ontario Iroquoian sites BR 99 spring layout. .............. 123 Figure 5-5. Local tradition ceramics of Ontario Iroquoian sites BR 140 main component MDS layout........................................................................................................................................... 124 Figure 6-1. Map showing all post moulds. ................................................................................. 132 Figure 6-2. Map showing all post moulds and the 1 m contour intervals. .................................. 135 Figure 6-3. Digitized sketch plan showing houses, middens, and palisades and the 1 m contour intervals. ...................................................................................................................................... 136 Figure 6-4. View of the high town and low town longhouses at Keffer’s north end during excavation. .................................................................................................................................. 137 Figure 6-5. Plan showing the houses (solid fill), middens (dotted fill), and palisades (dotted lines) associated with each of the four site clusters, and the 1 m contour interval. .............................. 139 Figure 6-6. Plan showing Phase 1a. ............................................................................................ 143 Figure 6-7. Plan (left) and map (right) showing Phase 1a, Cluster A. ........................................ 145 Figure 6-8. Map detail showing House 2 sweatbath postmoulds over House 3 north wall. ....... 145 Figure 6-9. View of House 6 wall post mounds under Feature 2 of House 7. ............................ 146 xviii Figure 6-10. Plan (left) and map (right) showing Phase 1a, Cluster B. ...................................... 148 Figure 6-11. Plan (left) and map (right) showing Phase 1a, Cluster C. ...................................... 150 Figure 6-12. Map detail showing House 13b wall posts crossing Palisade 1. ............................ 152 Figure 6-13. Plan (left) and map (right) showing Phase 1a, Cluster D. ...................................... 154 Figure 6-14. Map detail showing south end of House 11. .......................................................... 155 Figure 6-15. Plan showing Phase 1b. .......................................................................................... 158 Figure 6-16. Plan (left) and map (right) showing Phase 1b, Cluster A. ..................................... 159 Figure 6-17. Plan (left) and map (right) showing Phase 1b, Cluster B. ...................................... 160 Figure 6-18. Plan (left) and map (right) showing Phase 1b, Cluster C. ...................................... 161 Figure 6-19. Plan (left) and map (right) showing Phase 1b, Cluster D. ..................................... 163 Figure 6-20. Plan showing Phase 2a. .......................................................................................... 164 Figure 6-21. Plan (left) and map (right) showing Phase 2a, Cluster A-B-D north. .................... 165 Figure 6-22. Plan (left) and map (right) showing Phase 2a, Cluster D south. ............................ 166 Figure 6-23. Plan (left) and map (right) showing Phase 2a, Cluster C. ...................................... 168 Figure 6-24. Map detail showing House 16b extension. ............................................................ 169 Figure 6-25. Plan showing Phase 2b-1. ...................................................................................... 169 Figure 6-26. Plan showing Phase 2b-2. ...................................................................................... 169 Figure 6-27. Plan (left) and map (right) showing Phase 2b-1, Cluster D south. ........................ 170 Figure 6-28. Plan (left) and map (right) showing Phase 2b-2, Cluster A-B-D north. ................ 171 Figure 6-29. Plan (left) and map (right) showing Phase 2b-2, Cluster D south. ........................ 172 xix Figure 6-30. Plan (left) and map (right) showing Phase 2b-2, Cluster C. .................................. 173 Figure 6-31. Plans of (left to right) site settlement Phases 1a, 1b, 2a, 2b-1, and 2b-2. .............. 173 Figure 7-1. Local tradition ceramics Over 1 Vessel BR 170 MDS layout with BR 190 Blue Backbone structure...................................................................................................................... 181 Figure 7-2. Original Stages 1–5 as established through social network analysis. ...................... 182 Figure 7-3. Plan showing possible features of Phases 1a to 2b-2. .............................................. 182 Figure 7-4. Second-level temporal Stages 1–5 as established through social network analysis and settlement studies. ....................................................................................................................... 183 Figure 7-5. Undecorated neck Huron, decorated neck Huron, and Neutral ceramics percentages sorted by decorated neck percentages, highest to lowest. ........................................................... 189 Figure 7-6. Keffer development sequence Maps 1–8. ................................................................ 190 Figure 7-7. Local tradition ceramics Over 1 Vessel BR 185 Gower layout. Temporal patterning, from earlier to later, is represented by the physical placement in the figure, left to right. ......... 196 Figure 7-8. Local tradition ceramics Over 1 Vessel BR 171 Gower layout ............................... 197 Figure 8-1. Local tradition ceramics Over 29 Vessels BR 190 Gower layout. .......................... 200 Figure 8-2. Local tradition ceramics Over 1 Vessel BR 185 Gower layout. Temporal patterning, from earlier to later, is represented by the physical placement in the figure, left to right. Colours indicate community of practice group. ....................................................................................... 202 Figure 8-3. Local tradition ceramics Over 1 Vessel BR 171 Gower layout. .............................. 203 Figure 8-4. Local tradition communities of practice. ................................................................. 208 Figure 8-5. Keffer development sequence Maps 1–8. ................................................................ 209 Figure 8-6. Map 1, with community of practice 5. ..................................................................... 210 Figure 8-7. Map 2, with communities of practice 2, 3, 4, and 5. ................................................ 211 xx Figure 8-8. Map 3, with communities of practice 1–5. ............................................................... 213 Figure 8-9. Map 4, with communities of practice 1–5. ............................................................... 214 Figure 8-10. Map 5, with communities of practice 1–5. ............................................................. 215 Figure 8-11. Map 6, with communities of practice 1–5. ............................................................. 216 Figure 8-12. Map 7, with communities of practice 1–5. ............................................................. 217 Figure 8-13. Map 8, with communities of practice 1–5. ............................................................. 218 Figure 8-14. Pan-Iroquoian network BR 93 spring layout. ........................................................ 219 Figure 8-15. 140 Ontario Iroquoian sites, including Coulter and Kirche, and Keffer loci local tradition ceramics BR 120 spring loaded. ................................................................................... 220 Figure 8-16. Locations of selected ancestral Wendat sites along the north shore of Lake Ontario. (Williamson 2014:Figure 2) ........................................................................................................ 221 Figure 8-17. Selected Ontario Iroquoian sites, including Coulter and Kirche, and Keffer loci local tradition ceramics BR 150 spring loaded. ................................................................................... 222 Figure 8-18. Selected Ontario Iroquoian sites, not including Coulter and Kirche, and Keffer loci local tradition ceramics BR 170 spring loaded. .......................................................................... 223 Figure 8-19. Local tradition ceramics ego Logan BR 170.......................................................... 224 Figure 8-20. Local tradition ceramics ego Damiani BR 170. ..................................................... 224 Figure 8-21. Connected Ontario sites, including Coulter and Kirche, BR 170 Gower layout main component. .................................................................................................................................. 225 Figure 8-22. Selected Ontario Iroquoian sites, not including Coulter and Kirche, with Keffer loci local tradition ceramics BR 160 spring loaded. .......................................................................... 226 Figure 8-23. Local tradition ceramics ego Damiani BR 160. ..................................................... 227 Figure 8-24. Local tradition ceramics ego Logan BR 160.......................................................... 227 xxi Figure 8-25. Connected Ontario Iroquoian sites and Keffer loci local tradition ceramics BR 150 spring loaded. .............................................................................................................................. 228 Figure 8-26. Local tradition ceramic communities of practice plan Maps 1–8. ......................... 235 Figure 8-27. Local tradition ceramic practice by settlement plan Maps 1–8.............................. 238 Figure 8-28. Nonlocal ceramics by sample size BR 100 spring layout. ..................................... 239 Figure 8-29. Nonlocal ceramics by sample size BR 130 spring layout. ..................................... 240 Figure 8-30. Keffer village local tradition practice. ................................................................... 241 Figure 8-31. Nonlocal ceramics by local tradition communities of practice BR 150 Gower layout. ..................................................................................................................................................... 242 Figure 8-32. Nonlocal ceramics by local tradition communities of practice BR 140 spring layout. ..................................................................................................................................................... 243 Figure 8-33. Nonlocal ceramics by local tradition communities of practice BR 130 spring layout. ..................................................................................................................................................... 244 Figure 8-34. Nonlocal ceramics by local tradition communities of practice BR 120 spring layout. ..................................................................................................................................................... 245 Figure 8-35. Nonlocal ceramics by local tradition communities of practice BR 110 spring layout. ..................................................................................................................................................... 246 Figure 8-36. Nonlocal ceramics by local tradition communities of practice BR 100 spring layout. ..................................................................................................................................................... 247 Figure 8-37. Nonlocal ceramics by local tradition communities of practice BR 70 spring layout. ..................................................................................................................................................... 247 Figure 8-38. Nonlocal ceramic group practices. ......................................................................... 250 Figure 8-39. Otstungo Notched emergent ceramic from Keffer. ................................................ 256 xxii Figure 8-40. Reconstructed Durfee Underlined emergent ceramic from Parsons. (Williamson and Powis 1998:Figure 20) ................................................................................................................ 257 Figure 8-41. Shared local and nonlocal ceramic practice at Keffer. ........................................... 261 xxiii List of Appendices Appendix A: BCal Radiocarbon Dates for the Keffer Site ....................................................... 312 Appendix B: X-ray Fluorescence Analysis of Copper Artifacts from the Keffer Site .............. 314 Appendix C: Previous Research on the Keffer Site .................................................................. 316 Appendix D: All Ceramic Types Represented in the Keffer Site Assemblage ....................... 318 Appendix E: Appendix to Chapter 4: Methodology Brainerd-Robinson Adjacency Matrix ... 340 Appendix F: Appendix to Chapter 4: Pan-Iroquoian Sites Employed in the Social Network Analysis..................................................................................................................................... 355 Appendix G: Appendix to Chapter 5: Social Network Analysis of 140 Ontario Iroquoian Local Tradition Ceramic Collections .................................................................................................. 370 Appendix H: Appendix to Chapter 7: Factors Affecting Social Network Analysis Results for Keffer Site Local Tradition Ceramics ....................................................................................... 381 Appendix I: Appendix to Chapter 7: Chronological Patterning in Keffer Local Tradition Ceramic Social Network Analysis and Its Effect on Interpretations of Village Settlement Patterns...... 393 Appendix J: Appendix to Chapter 7: Social Network Analysis of Keffer Site Social Relations and Communities of Practice ........................................................................................................... 431 xxiv 1 Introduction Situating the Keffer Village In this thesis, I employ a multiscalar social network analysis approach to explore the social relationships of the inhabitants of the fifteenth-century Ontario Iroquoian Keffer village. I further employ a communities of practice approach to illuminate the ongoing emergence of group identity among village households. The formation of identity is mediated in part by quotidian engagements with material culture that anchor identity in the habitual practices and dispositions of social agents (habitus in the spirit of Bourdieu 1977). I show that some Keffer households were members of separate, but concurrent, traditional and non-traditional communities of ceramic practice. Each household assemblage (after DeLanda 2006; Swenson 2018) shared ceramic ties at multiple scales of increasing social distance, or territoriality (Swenson 2018), with other Keffer households and local, regional, and, in some cases, pan-Iroquoian communities. The shared, matrilineally based, kin-oriented identities portrayed by these ties do not reflect ethnicity but are a product of close social relations constructed and maintained through quotidian domestic interaction. The Iroquoian inhabitants of the Keffer village and those of the surrounding area are commonly referred to as ancestral Wendat (Birch and Williamson 2013), one of several geographically circumscribed Northern Iroquoian groups living in close proximity to the lower Great Lakes of eastern North America. To the west of Lake Ontario are the peoples referred to as the ancestral Neutral or Attiwandaron. Differentiation of these two groups is primarily based on their historic relations, and here they are united as southern Ontario Iroquoians. Other Northern Iroquoians include the ancestral Haudenosaunee (Seneca, Cayuga, Onondaga, Oneida, and Mohawk) south of Lake Ontario; the Northern New York Iroquoians of Jefferson County, at the eastern end of Lake Ontario; and the diverse St. Lawrence Iroquoian communities spread eastward along the shores of the St. Lawrence River. Located amidst a geographically vast territory of Algonquianspeaking peoples, these groups are united in belonging to the Northern Iroquoian language family, distantly related to that of the more southerly Cherokee. 1 2 Northern Iroquoian groups shared several similar, but not identical, cultural practices that distinguished them from their less sedentary Algonquian neighbours. By the turn of the twelfth century, maize-based horticulture dominated the subsistence pattern. This transformation was accompanied by dramatic shifts in residential patterns. Iroquoian groups settled into semipermanent communities of two to six residential, bark-covered longhouses by the turn of the fourteenth century (Birch and Williamson 2018). Residences were inhabited by extended matrilocal families, and matrilineal clan ties linked residents to families and longhouses within the village and to other communities, both local and farther afield. Exhausted agricultural fields and deteriorating domestic structures prompted village relocation, usually within the local drainage system, at intervals of 10 to 50 years (Warrick 1988). For reasons unknown to us, former village locations were rarely re-occupied. Therefore, most archaeological Iroquoian village sites represent the occupation of a single generation, in the timeline of one community. Much of the late-fourteenth- to late-sixteenth-century north shore Iroquoian occupation occurred on the southern slopes of the Oak Ridges Moraine, to the north of the western and central portions of the northern Lake Ontario shoreline, in today’s densely occupied Greater Toronto Area. This area surrounding the City of Toronto has seen rapid residential expansion in the past half century. With the introduction of mandatory archaeological assessment in the later part of the twentieth century, a large number of Iroquoian village and special-purpose sites have recently been subject to intensive investigation and excavation. Southern Ontario is distinguished by an archaeological record of temporally and spatially distinct small-scale communities almost unparalleled. As a small-scale society, Northern Iroquoians are often characterized as tribal or Neolithic (Creese 2010; Kvetina and Hrncir 2013). While seemingly innocuous, the term tribal is laden with problematic assumptions (sensu Sahlins and Service 1960; Service 1962). Tribe is the second of a four-tier neoevolutionary classification scheme presented by Marshall Sahlins and Elman Service (1960). Band, tribe, chiefdom, and state represent the four stages of sociopolitical development. The first of these, bands, are small, nomadic, egalitarian groups. The second stage, tribes, are characterized by horticultural or pastoralist economies. Social relations are based on kinship and on reciprocal exchange within larger social groups numbering from a hundred to a few thousand. Leadership at this level remains egalitarian. In chiefdoms, hierarchical leadership of groups commonly numbering in the tens of thousands is supported by 3 the introduction of intensive agriculture. The final evolutionary stage is the state, with centralized authority and social stratification and specialized craft production supported by intensive agriculture. The term tribe has been used in North American archaeology to describe the general characteristics of small-scale agricultural groups without the intention of implying neoevolutionary connotations, while the term Neolithic is more common in archaeology across the globe (Creese 2010:1) but is now gaining more acceptance in Iroquoian archaeology (Birch 2018). Swidden agricultural societies, characteristic of both the Neolithic and tribal categories, appear in the ethnographic record worldwide and in the archaeological record at least as far back as 10,000 years B.P. Ethnographic studies of such groups normally deal with the micro-scale dynamics of social relationships and their articulation with the larger community and society within a relatively short temporal scale, that of months, years, or, in some cases, decades, at the level of “evenéments” as defined in the Annales school (Braudel 1980). The incorporation of the ethnography of small-scale societies provides an unparalleled resource for interpretation of the archaeological record. Within this record, the processes of these relations can be discerned at the medium, and sometimes longue, durée scale (Harris and Cipolla 2017). The resolution of the southern Ontario Iroquoian archaeological record provides the opportunities for the intensive study of the social dynamics and processes at the micro-, macro-, and multiscalar levels of “nonstate” society. This thesis aims to contribute to the investigation of the complexity of the social relations of heterarchical societies as seen in the Iroquoian occupation of southern Ontario by means of the application of relevant theoretical perspectives including social network analysis; modern perspectives; the integration of historical contingency, practice theory, and social agency; and a rich and detailed database. I pursue this goal through the multiscalar examination of Iroquoian ceramic assemblages from the household to the pan-Iroquoian level. Keffer ceramic assemblages, analysed employing the commonly used typological taxa developed by Richard MacNeish (1952), are adopted as proxies representing communities of practice of ceramic production and/or consumption at the household level. The Keffer assemblage includes a significant number of “exotic” or “foreign” pottery vessels relative to most Ontario Iroquoian collections. These vessels, labelled “nonlocal tradition” ceramics here, are considered to embody distinct communities of practice not seen in the south-central Ontario Iroquoian tradition. This 4 label is not meant to indicate nonlocal production. In contrast, ceramic types described by MacNeish as characteristic of ancestral Neutral and Wendat production share a common ceramic heritage and constitute a large majority of the ceramics of Ontario Iroquoian sites. These ceramic assemblages represent the “local tradition” practice community. Multiscalar examination of the relationships between these two domains of ceramic practice, at both the intra- and inter-village levels, aims to decipher the origins and possible destinations of the peoples of Keffer. Chapter Summary In chapter 2, I present an in-depth discussion of the various theoretical and methodological approaches employed in this research. I explore the intersecting concepts of ethnicity and identity and their relations with communities of practice. In the case of egalitarian pre-contact Iroquoian society, a relational construction of identity proves more accurate than bounded categorical identities. The place of communication in these complex forms of identity is also explored. I examine nuanced elements of individual and group identity through the clinal plane of social distance. I explore in turn the role of interaction theory and the social network approach in the analysis of social distance, group identity, and the resulting communities of practice. Chapter 3 situates the Keffer site within the social and historical context of Iroquoian society in Ontario and New York state. The Keffer site itself is then discussed in the context of the recent history of the extensive excavations, research projects, and publications pertaining to the siterelated research which has occurred during the past 35 years. I explain the methodology used in this thesis to categorize and analyze both the Keffer site ceramics and pan-Iroquoian ceramics in Chapter 4. Here, the types included in the study and the guidelines established to assign ceramic types are clearly put forth and explained. These ceramic types, based on the MacNeish typology, are then assigned to one of two macro-scale conventions. These include the local tradition ceramics, as described above, and nonlocal ceramics. This group is comprised of pottery whose style originates to the south and east, among the New York state or the St. Lawrence Iroquoians. This distinction is made on the basis of perceived origin and main geographical location of the type’s dominance. Following this, I describe similarity matrices created using the Brainerd-Robertson coefficient from both the local and nonlocal ceramic databases and I explain their incorporation into network analysis. 5 In Chapter 5, I present social relations of Northern Iroquoians as mediated by ceramic practice. I introduce a depiction of the relations of 252 ceramic assemblages at the pan-Iroquoian level as seen through social network analysis. This is followed by network analysis of the local tradition ceramics of 142 Iroquoian sites from across southern Ontario. Together they form the diachronic and synchronic contextual backdrop for the positioning of the Keffer village ceramics. Chapter 6 then presents an in-depth analysis of the settlement patterns of the village, resulting the construction of four major phases of development of the village. Each of these phases likely comprise contemporaneous settlement features, houses, middens, and palisades. In chapter 7, I present the results of the network analysis performed on local tradition ceramics from each of the village houses and middens. I interpret network graphs, of increasing levels of similarity and sample size, into two separate domains, temporal and social. A rough timeline of the Keffer village occupation, displayed as a series of stages, is produced through the interpretation of the locations of the village houses and middens along one axis of the graph topology, the arrangement of ties and nodes in a network graph. In the second domain of interpretation, the presence of ties among these features is interpreted, displaying social ties among longhouses. The strongest of these node connections form the core of several partially distinct spheres of ceramic practice across the site. In the second section of Chapter 7, the communities of practice are applied to the settlement phases developed in Chapter 6. The incorporation of both ceramic and settlement analysis provides a more refined and detailed depiction of village development and results in the division of the four proposed occupation phases into a series of eight sequences. In Chapter 8, I discuss the separate intra-village local ceramic spheres of practice, inferred from network analysis, and I investigate the relationships which exist among them. Further network analysis of the Keffer local tradition ceramics, in comparison with other southern Ontario sites, reveals close similarities of these Keffer household communities with those of surrounding settlements. These results are then interpreted to propose possible population origins for the households of the growing Keffer village. Network analysis of nonlocal tradition ceramic similarities identifies three additional interrelated spheres of ceramic practice. The first of these depicts Haudenosaunee influence through north interaction networks. The second reflects possible direct ties with Iroquoians of the St. Lawrence valley. The third sphere is composed of two distinct, new and “emergent” (Lee 2006), vessel types, local products reflecting emergent identities resulting from diverse influences. Overlap in these practice 6 domains is examined in the context of household origin, identity, and internal and external social relations. Chapter 9 brings together the results of the settlement pattern and ceramic analysis. A summary of the growth of the village traces the proposed origins of each new household and the developing social relationships of ceramic practices within the village over its lifetime. The decline of the village population is then described, and I offer some hypotheses on the dispersal and social reconstitution of the final Keffer village inhabitants. In the last section of this chapter, I evaluate the appropriateness of the methodological and theoretical approaches employed to unravel the synchronic and diachronic relations of the Keffer village and propose future lines of research. Over the course of these nine chapters, I employ settlement pattern studies, in addition to network and typological analysis of the Keffer village ceramics, to produce a nuanced portrayal of intra-village social relations within the context of the village, regional, and pan-Iroquoian developmental history. 2 Theoretical Underpinning: The Search for Identity Identity: From Ethnicity to Active Community The terms “ethnicity” and “ethnic group” are the subject of much controversy in the social sciences. In archaeology, they have often been used as synonyms for “group identity” and “social group.” In many cases, the authors cited in this chapter refer to these terms interchangeably, either without reference to a concrete definition or, in some cases, using a definition that does not accord with contemporary social theory dealing with ethnicity and identity. Archaeological material culture items displaying common attributes have traditionally been labelled as individual “types,” while prevalent combinations of various distinctive spatially associated artifacts found throughout a geographical region area were seen as “diagnostic” of “cultures” (Childe 1929). Geographically bounded assemblages of material remains were assumed to derive from a common culture and common language and, ergo, to hold a common identity (Peeples 2011). Specific diagnostic material culture combinations were then assigned ethnic identity labels; the “people”—and their associated group identity—were merely a byproduct of the artifacts. These “holistic” units were viewed as disconnected, discrete, and localized in nature, existing and persisting with little to no external contact (Cunningham 2001). This culture-historical approach in early twentieth-century Western archaeology compartmentalized material culture territories into discrete packages of peoples, slotting them into standardized, geographically and chronologically demarcated containers of culture conforming to taxonomic system patterns of Phase, Horizon, and Tradition, as exemplified in McKern’s Midwestern Taxonomic System (Childe 1929; Kossina 1911; McKern 1939; Ritchie 1951, Willey and Phillips 1958). Sassaman and Rudolphi argue that the normative perspective that characterized the culture-historical approach assumed that learning processes were crossculturally and temporally invariant (Sassaman and Rudolphi 2001:407). It was assumed that each generation produced a replica of earlier material culture attributes and ethnic identity through time (Trigger 1989), thereby reducing emic relationships of identity and ethnicity to etic classifications of artifacts (Beisaw 2010:244). Identity, then, from the late nineteenth to the mid- 7 8 twentieth century, registered only at the regional level, and people were viewed only in terms of the larger “culture groups” to which they were assigned (Stone 2003). In contrast, by the late 1920s, the term “culture” had been replaced in anthropological circles by the notion of the “tribe,” then seen as a distinct, self-reliant social group (Diaz-Andreu 2015:4818), which itself was replaced during the mid-twentieth century by “ethnic group,” a term first theorized at the end of the 1920s (Diaz-Andreu 2015; Jenkins 2008). Those sharing one ethnicity were seen as having genetic, linguistic, cultural, and, often, spiritual traditions in common (Peeples 2011). Ethnic identity was seen as directly resulting from the geographic and social isolation of “culture” groups, separating them from communities outside the materially conscribed borders. Although the label “ethnic group” would suggest the incorporation of the concept of “persons,” culture was viewed as homogeneous and homeostatic, and change in structure only occurred with the introduction of unfamiliar “culture” from external sources, whether caused by warfare, immigration, or other factors (Kroeber 1939; Linton 1944; Wissler 1914). The categorization of diagnostic material culture attributes formed the basis for the recognition of group identity in the application of this model. The first large-scale culture histories of Northern Iroquoian tribal groups, influenced by the Midwestern Taxonomic System (McKern 1939), appeared in the mid-twentieth century. William Ritchie’s influential The Pre-Iroquoian Occupations of New York State (1944) and Ritchie and Richard MacNeish’s The Pre-Iroquoian Pottery of New York State (1949) were followed by MacNeish’s Iroquois Pottery Types: A Technique for the Study of Iroquoian Prehistory (1952). The Chance horizon: An early stage of Mohawk Iroquois Cultural Development, by Ritchie, appeared in 1952 and this work was expanded on by Donald Lenig in 1965. In 1974, Engelbrecht examined archaeological patterns reflected at the tribal level among the New York state Iroquois. It was not until the publication of J. V. Wright’s The Ontario Iroquois Tradition (1966) that such a synthesis was available for the pre-contact period Ancestral Wendat (Huron) materials of southern Ontario. These culture histories were products of the unquestioning colonial thinking of the early twentieth century. As elsewhere, these pre-contact societies were viewed as homeostatic and unchanging. Aboriginal peoples were seen as uncreative and unimaginative (Trigger 1980). These “groups,” were not really acknowledged as being constituted of individual beings but were 9 analyzed in isolation within the bounded geographical areas arbitrarily assigned to their population. Any apparent change in material culture could only be explained as the result of some external cultural influence. The normative paradigm viewed prehistoric human behaviour occurring within unquestioned and unchanging societal rules. Only through contact with more advanced peoples could socio-political, ideological, or economic change occur. This contact took place in the form of a uni-directional influx caused by either diffusion—the spread of “successful” or “attractive” attributes (Childe 1950) from one culture to another—or full-scale population migration into the area. Decades later, in the 1990s, Dean Snow (1995, 1996) proposed the appearance of Iroquoian cultural traits within a sea of Algonkian peoples to be the result of migration into the Great Lakes area and not independent in-situ development. Several Iroquoianists, particularly Crawford and Smith, refuted this hypothesis (Crawford and Smith 1996; Smith and Crawford 1995, 1997, 2002, n.d.). On a smaller scale, major Ontario Iroquoian scholars, such as Norman Emerson and Frank Ridley, spent many years in a battle over the direction, not the existence, of fifteenth-century migration between two areas of marginally distinctive ceramic typologies (Emerson 1998; Emerson and Popham 1952; Pendergast 1965; Ridley 1952, 1958, 1963). During the following decades, northeastern archaeologists continued to produce culture histories within a framework of diffusionist–migrationist paradigms to explain culture change. Wright’s The Ontario Iroquois Tradition (1966) ignored any possibility of internal cultural transformation as an explanation for apparently sudden shifts in material culture characteristics. As an example, the conquest of the Glen Meyer peoples to the west by the Pickering branch of the Early Ontario Iroquois during the thirteenth century was presented as the only likely cause of widespread ceramic transformation in central Ontario during the thirteenth century (Wright 1966:53). As a reflection of the persistence of the culture-historical mindset, which based the definition of “culture” solely on material culture, The Ontario Iroquois Tradition remained the “bible” of Ontario archaeology for many decades. Its delimitation of what were considered two geographically separated ancestral Wendat culture regions was based on the greater presence of one ceramic vessel type, Lalonde High Collar, to the north of the Oak Ridges Moraine and a lesser frequency of this type south of the moraine (Wright 1966:68–74). After this period of intense interest in culture history paradigms, the groundbreaking introduction by Fredrik Barth of the role of perception in the concept of identity (1969) sparked a renewed 10 interest in the study of identity in anthropology and sociology. This precipitated a radical shift in social theory. In fact, the concept of the subjective creation of identity had been introduced more than 40 years earlier, in Max Weber’s Economy and Society, first published in 1922. Weber proposed that “race” (a word which Jenkins (2008:10) suggests Weber equates with “ethnic group) “creates a “group” only when it is subjectively perceived as a common trait: this happens only when a neighbourhood or the mere proximity of racially different persons is the basis of joint (mostly political) action, or conversely, when some common experiences of members of the same race are linked to some antagonism against members of an obviously different group” (Weber 1978:385). In this view, the ethnic group is defined by the belief in shared descent to support a common interest (Jenkins 2008:10). Barth’s reintroduction of this idea transposed “identity” from a static entity into a flexible representation of the interplay of subjective, socially meaningful factors chosen and ascribed by the players themselves (Diaz-Andreu 2015:4818; Peeples 2011:9). Significantly expanding on this foundation, Ronald Cohen introduced the idea that identity (termed ethnicity, in his definition), in all its forms, is both fluid and situational (Cohen 1978), and that it transforms, in accordance with changing relations and social, economic, and political factors. Social identity was no longer an attribute of a people, but a relationship created and maintained through all forms of interaction (Peeples 2011). Currently the definition of “identity” is commonly accepted as “the way individuals relate themselves and others to larger groups based on perceived similarities and differences socially defined as important” (Peeples 2011:2; see also Barth 1969; Calhoun 1995:193–197; DiazAndreu et al. 2005; Emberling 1997; Jenkins 2004:1–8; Jones 1997). Perhaps the most revolutionary element of this definition is the concept of self-perception. From this perspective, identity is an acknowledgement of feelings of belonging or not belonging originating from within. Cunningham suggests that identity consists of one’s stories, the cumulative narratives of one’s life (Cunningham 2001; see also Ferris 1999, 2006, 2017). It is through the identification of the differences between individuals and groups that self-ascribed identity is formed (Kolb and Snead 1997). This self-ascribed identity is purposely and strategically bounded (Barth 1969; Bowser 2000: Duff 2002; Hodder 1985; Mantha 2009), and according to Cameron (2013:220), following Barth (1969), agents construct their social identity to optimize their social standing. The fluid and situational nature of identity (Cohen 1978) is essential to this definition. Since 11 some believe that identity is often manipulated for social advantage, it follows that even identities as predictable and non-threatening as those associated with marriage and aging may be altered and signalled to maximize advantages in changing circumstances. The extent of this maximization is contingent on distinct cultural logic. Such perspectives stress the strong relationship between the agent and identity (see also Hodder 1982). Divisions in the ceramic typology employed in Iroquoian archaeology are based on MacNeish’s perception of Iroquoian ethnicity. This study will show that while ethnicity may be reflected in some aspects of ceramic production and use, it is not actually reflected in typology. It is, however, more clearly reflected in the accepted norms of production, or the chaîne opératoire; form; and, perhaps, function of specific vessel “types” in the case of the Northern Iroquoians. These vessels are consciously employed as symbols of unique, nonlocal community, or household relations and identity in the Keffer village. Introduction of the Human Factor through Agency Theory With the introduction of agency theory, British researchers, such as Hodder (1982) and Shanks and Tilley (1987), successfully challenged the non-involvement of humans, either as individuals or as groups, in the creation and maintenance of their culture. Agency theory, a theoretical tenet of postprocessualism along with creation and negotiation of meaning, proposes that people were actively, often intentionally, involved in directing and historically affecting their social, political, economic, and ideational environment (Dobres and Robb 2000). This theory was immediately integrated into European archaeology but remained on the periphery of human interaction and identity research in North America until somewhat later (Hegmon 2003). The merger of Frederick Barth’s “self-ascribed identity,” Cohen’s concepts of identity as fluid and situational, and the actors of agency theory gave researchers such as Cohen (1978), Hegmon (1992), and Emberling (1997) the theoretical tools that they coalesced into the interactionist approach to group identity, or “ethnicity,” as described by Stone (2003; see also Svensson 1985). This interactionist approach rejects the characterization of identity as a concrete entity, a consequence of spatial and social group isolation or local environmental adaptation defined by material traits that remain unchanging prior to the introduction of new cultural attributes imported with the influx of outside populations or ideas. The approach defines identity as a process whereby individuals consciously create, use, and manipulate material and non-material 12 symbols of group inclusion and exclusion. In the formation and re-formation of identity through interaction within and between groups, this social categorization of self and others is in a constant state of negotiation (Emberling 1997). Ascription to any and all groups required recognition and acceptance of such status by both self and all others in the applicable social milieu. Symbols and Boundaries Symbols of inclusion are meaningful only when they are recognized by both those referenced and those excluded. The tensions involved in this dichotomy bore the potential for conflict in the reification of us–them categorization and the creation of social boundaries. It is in the complex agential manoeuvring of social relations involved in interaction that the process of identity creation occurs (Weber 1978). This identity creation takes place across all dimensions of society, from the immediate family, to the most comprehensive scale of interaction and can encompass all aspects of culture—social, political, economic, and ideational. These social relations link participants through both space (i.e., within a household) and time (i.e., across generations). Messaging of Social Distance This prioritization of social relations as the driving force in identity creation, in concert with the exclusion of the material mediation of the interactions (Jones 1996), is seen as a major fault of the interactionist approach. Researchers approaching identity studies from the enculturationist school point out that a relational view emphasizes differences to the detriment of similarities, thereby stressing boundary maintenance (Bentley 1987), while overlooking the internal unity of social groups. In contrast, in the enculturationist model, culture content and the categorization of diagnostic material culture attributes formed the basis for the recognition of group identity. Consumers and producers of less visible symbols may unconsciously communicate inclusion within social groups. Change in production methods of less visible material objects is considered to reflect divergence in learning environments or inculcation of outsider behaviours into the daily habitus at the small scale. In contrast, highly visible symbols were sometimes employed consciously to demarcate similarities and differences (Peeples 2011). 13 The enculturationist approach employs Bourdieu’s concept of habitus as the main source of identity formation. Habitus encompasses the deep-seated dispositions and the embodiment of the socialized norms that guide human behaviour and thinking within a structure of socially conditioned and acceptable limits to that behaviour. Habitus thus forms the basis for the creation, maintenance, and re-creation of “culture” (Bourdieu 1977). Identity orientation stems from the cultural matrix in which people experience, use, learn, and act within their daily lives, and within which they construct an ongoing sense of themselves and an understanding of their fellows. Culture and identity are not “‘something” that people “have,” or, indeed, to which they “belong’” (Jenkins 2008:19); instead, they are best thought of as ongoing processes of identification. Additionally, critics such as Jones (1996, 1997) reject the assumptions of the interactionist approach that all behaviour and interaction is consciously performed in direct pursuit of perceived personal benefit. They insist that unity is created through the everyday habitus of practice shaped by unspoken social norms and that action is often performed without conscious intention. Interactionists reject this enculturationist approach as a return to primordialist thinking, in which identity formation is dependent on isolation and in which humans lack any input, either conscious or unconscious, directing their lives. As will be more fully discussed below, one of the fundamental characteristics ignored by the two approaches, but essential to their differences, is the variable of scale. Enculturationists fail to recognize that the foundation of their approach lies in interaction. This interaction, however, is at the most intimate level, the micro-scale of family relationships. Habitus is formed through repeated and daily practice within the cultural context of the micro-scale of the immediate family and/or household. Habitus is interaction. Without such small-scale, intra-household interaction, there could be no learning, no enculturation. Interactionists, on the other hand, most commonly investigate patterns at levels of interaction above this intimate familial level, looking at evidence for interaction among bounded groups in the search for larger-scale, collective group identities (e.g., Braun and Plog 1982; Duff 2002; Emberling 1997:319–325; Jones 1997; Peeples 2011:14; Plog 1980, 1983; Shennan 1989; Upham 1982). But identity is not just based on cultural affiliation. Gender, age, ritual societies, and clan sodalities operate at much smaller scales. In the end, the selection of an appropriate theoretical and methodological approach for research of past identity depends on the social scale of social interaction, whether at the individual or at any of the multitude of social group levels. 14 The social distance approach was proposed by Lyons and Clark (2008) to overcome the shortcoming of the two perspectives discussed above and to bring together in one analysis the positive aspects of the interactionist approach to ethnicity, with its emphasis on agency and the purposeful use of style to communicate identity (as seen by Wiessner 1984; Wobst 1977), and the enculturationist analysis of style as a passive element embedded in the habitus of everyday life that reproduces the cultural structure within which it exists (Sackett 1973, 1977). Structure and agency are given equal weight in Lyons and Clark’s approach, in contrast to previous approaches to identity, which have tended to emphasize either an enculturationist or an interactionist view, each to the exclusion of the other. The enculturationist approach measures the magnitude of social and physical learning frameworks of artifact production in social distance or, more accurately, social proximity. In the study of these communities of practice, social closeness, or proximity, of members is assumed and generally taken for granted. Lyons and Clark (2008) examine the ethnographic study of the historic relationship of the Tewa and Hopi peoples of the American southwest and the material and nonmaterial consequences of their interaction over time (Kroskrity 1993) to uncover patterns relating to increased social interaction and its effect on cultural convergence. The results of this research produced two general observations. The first states that increased social relations between two groups (in this case specifically what they consider “ethnic groups”) will lead to a decrease in social distance. The second states that these increased cultural similarities, seen in both material and ideational forms, may not occur in all areas of culture. Distinctive core traits may be retained or emphasized, or both, by one or both groups as “badges of ethnicity.” Lyons and Clark (2008) admit that we do not yet have the capability to address ancient ethnicity directly. They therefore suggest that social distance can be used as a proxy to measure relative group identity. Consequently, it seems plausible that these could be equally applied to non-ethnic collective identities. Following the work of Kroskrity, Lyons and Clark conclude: 1. Ethnicity (Collective Identity) as a force of social agency is quite durable and may remain after inter-group contact. 2. The diacritica (those traits which emphasize difference) of ethnicity (Collective Identity) can change as social distance decreases and cultural similarity increases. 3. Some diacritica may disappear after contact but others may remain and be emphasised. 15 4. Time depth is necessary for elucidating the dynamics which reflect social distance. 5. With increasing cultural similarity, as seen in material or ideational culture, there will be a decrease in social distance and subsequently a decrease in diacritica. 6. Knowledge of the distinct cultural traditions of both groups prior to increased interaction is necessary in order to ascertain changes [Lyons and Clark 2008:195–196]. Lyons and Clark (2008) also incorporate the bridging perspective of Abbott (2000). This approach, closely based on that of Moerman (1965), explicitly brings together the effects of enculturation, as seen through the analysis of ceramic technology, and of interaction, represented by ceramic intergroup exchange, to examine changes in social affiliation over time. Specifically, Abbott sees an increased similarity in ceramic technologies and the intensification of intergroup ceramic exchange as evidence of a decrease in social distance. Interpersonal closeness is characterized by Lyons and Clark as an ever-expanding circle of relationships. In contrast to the implications of the permanence of boundary maintenance characteristic of the interactionist camp (Cohen 1978; Emberling 1997; Hegmon1992; Stark 1998) social distance “evokes process, change, and scale” in the examination of interpersonal distance between groups or individuals. Lyons and Clark (2008) attempt to determine circumstances in which different “classes” of diacritica mediate social interaction. Following Barth (1969) and Stone (2003), they suggest that, under stable conditions, there will be many forms of ethnic diacritica. Highly visible, public “badges” of group membership, manifest in body adornment, public architecture, and so forth, appear. In periods of crisis or a demographic or power imbalance between populations, diacritica of alternate self-ascription will be salient and more commonly employed. This element of identity may be materialized in the possession of nonlocal ceramics as a public sign of alternate identity. These traits are flexible and are easily emphasized or de-emphasized according to specific and fluctuating contexts. This manipulation of diacritical symbols is based on the advantage or disadvantage of stressing particular identities in differing circumstances. Crisis and/or power imbalance and the ensuing social disruption can cause a breakdown in the continuity of learning frameworks and enculturative traditions, thereby causing a discontinuity in cultural traditions as communities of practice are disrupted or destroyed (for an Iroquoian example see Martelle 2004). 16 In the culture-historical approach, and to a lesser extent in the interactionist approach, “difference” was employed to create the designation of concrete divisions separating groups of people, termed boundaries. Boundaries, which are often evident in the presence of socially acknowledged symbols, existed simultaneously at various scales. Consequently, inclusion/exclusion at one level did not determine inclusion/exclusion at another. These boundaries and their respective symbols and social relations are always in a state of flux because interaction and the factors that determined it changed. The existence of spatial, physical boundaries between groups at any level is not essential in this approach. Boundaries, such as the one-row palisades seen in fourteenth- and fifteenth-century Iroquoia, most often take the form of social distinction (Stone 2003) marking social inclusion and exclusion within the village community. Although physical boundaries, in the form of village palisades and internal house partitioning, do exist, these divisions are interpreted as relational, not categorical, in nature following Creese (2010, 2012) and, following Pauketat and Ault (2005), should be viewed as the result of shared practice binding a community. In a similar vein, Mills (2018) has recently employed material culture boundary markers to explore relational identity in the American Southwest. In social distance and network studies, difference and similarity exist on a gradual plane of distance among social relationships, marking a sliding scale of association. In the study of boundary objects, the boundary is composed of an intermediary space linking two practices, not a line of demarcation. However, I view this approach as based on the concrete boundaries of discrete practices. It is in the union of two welldefined and dissimilar objects or attributes that the boundary is bridged. This bridging object is reducible to its two parts as unrelated entities and therefore does not recognize clinal social distance. Identity in this case may be based on ethnicity. I suggest emergence theory (Lee 2006), and the related concept of relational typologies within assemble theory (DeLanda 2006, 2016; Fowler 2017) are more useful concepts for the study of relational identity through ceramic analysis. In emergence theory, the appearance of a new and distinctive entity is due to the interaction of its constituent parts through collective action. The characteristics of this new entity are not explicable in terms of the base parts, and the object produced is irreducible back to these components (Lee 2006). This process of transformation and evolution was, and remains, an integral element of Iroquoian ideology as demonstrated in the recurring theme of metamorphosis in Iroquoian creation stories (Martelle 2002:448). Emergence is fluid, and if the newly emergent 17 entity has no productive effect and is epiphenomenal, it will vanish with time. With the integration of the emergence perspective, assemblage theory interprets archaeological types as dynamic, contextual, emergent, and relational (Fowler 2017). Incorporation of the relational perspective of assemblage theory in Iroquoian ceramic studies provides a direct link to the multiscalar exploration of relational identities. Within Iroquoian society, social relations, and therefore identities, were configured through kinship (Birch 2008; Sioui 1999). In non-state societies, such as the Northern Iroquoians, kinship is thought to be the universal or near-universal model structuring social behaviour (Chernela 1992; Shennan 1989). Shennan suggests that kinship may in fact be the only basis for group identity formation in these societies (Shennan 1989). He ignores the obvious social connections of such characteristics as gender and age, among many others. Although important, kinship is but one element of identity in non-state societies. Kinship escalates from the level of the individual, through nuclear and progressively less closely related family ties, to extensive clan networks. Each individual is related, whether this is acknowledged or not, to all kin at some level of social distance. Therefore, all individuals acquire multiple, multiscalar identities (Peeples 2011:2). As these identities are formed in relation to one’s perceived connection with others, along the clinal relationship scale from intimacy to enmity, identity is always situational and fluid (Ortman 2010) because contact with others is continually in a state of flux, even though some social identities may claim the ethos of permanence (Peeples 2011:2). Material remains, as uncovered through excavation, particularly in the case of Northern Iroquoian sites, encapsulate a brief period in the life of a community. In the past this led to a view of these communities as static and fixed. However, non-state populations, including Northern Iroquoians (Finlayson 1985; Ramsden 2009) were characterized by a pattern of perpetual circulation of individuals, families, and larger groups (Preucel 1988; Schachner 2010), and what appear to have been static, settled communities were unceasingly in a state of flux. Ethnographic and ethnohistoric studies confirm fluidity of social group membership in non-state societies (Cameron 2013:221). All relocations, including those on the smallest scale, resulted in a change in group composition and therefore a transformation in all its interconnected and interacting social units (Cameron 2013). This flux in membership, therefore, replicates itself in changes in identity, likely at a variety of scales simultaneously. Creese’s application of “personhood,” as it pertained to Northern Iroquoians, provides a more nuanced perspective that 18 moves beyond possible biases of present-day identity politics in simply acknowledging the multiplicity and contingency of individual identities (Creese 2012, 2016). In Creese’s definition of personhood, individual or nuclear family identity was “fractal” or “part-in-whole” and could not be separated from one’s collective identities, particularly those of the house and the lineage (Creese 2012:366). This multifaceted identity was formed in large part through habitus within the household and the longhouse. Continuous change in the membership of these groups was accompanied by constant recreation of this identity. Changes in identity, on the other hand, may have caused transformation in household membership. Approaching Iroquoian identity as based on “part-in-whole” personhood, it may be possible, through the analysis of ceramics associated with specific longhouses and the larger village, to uncover relationships and attendant identities of groups in terms of local, nonlocal, and emergent, household-focused spheres of practice at Keffer as the village grew through time. In the social distance approach, relative distance, measured by ceramic similarity, is evaluated in terms of those traits that are passively produced and those traits that are actively created. I see passive traits in the ceramics of Keffer as those conforming to the general practice of the local tradition chaîne opératoire. Agency is materialized at graduated, increasing levels of identity communication through decorative variation at the lowest level and the production of vessels through the auspices of nonlocal tradition chaîne opératoire. Lyons and Clark (2008; see also Stone 2003), following the work of Jenkins (2008), invoke the “social constructionism of ethnicity,” suggesting that the diacritica (those traits which emphasize difference) of passive and intentional messaging are evident in different aspects of material culture. It is only through the evaluation of the visibility and permanency of these signals within environments of social interaction that the nature of this messaging can be ascertained. The transmission of passive or intentional identity may be dependent on the nature of the identity that is being communicated. According to Peeples (2011), following the “New York School” of relational sociology, and the work of Christopher Tilly (1978, 2001), identity takes two separate but interconnecting forms, namely, relational and categorical. Relational identity is formed in the course of both casual and formal relationship networking and interaction in the form of kinship or exchange. It is based on personal positions (Hall 1990:225) within networks of interaction among individuals and, as such, has no perceivable boundaries. Categorical identity, on the other hand, acknowledges inclusion in a more formally recognized group based on perceived similarity 19 and differences with others in a well-defined social unit, such as those of state, religion, age, and gender. To be formally defined, a categorical identity must have tight boundaries of inclusion and exclusion. Categorical identity is the actively expressed affiliation with a specific social group that persists despite changing membership and participation (Barth 1969:10). Similarly, categorical identity is named and needs to be symbolized to facilitate recognition. Some researchers of prehistory believe that categorical identities were not significant in the past (Creese 2013; Shennan 1989), while others (Stone 2003) suggest that this stance ignores the existence of inequality in earlier societies. In the case of Northern Iroquoians of the pre-contact and pre-Confederacy era, it is widely held that strong social strictures were deeply integrated into the Iroquoian social system to counteract the existence of social inequality (Trigger 1990), therefore reducing the influence of most categorical identity. Categorical identity based on the sexual division of labour is, however, assumed, specifically in the realm of ceramic production, to the point where this study is presumably based entirely on the social influence of one gender and age group, that of the mature female producers and consumers of pottery. Here, identity will be explored as relational and emergent in nature and will be viewed from a relational approach, incorporating the enculturation perspective through the micro-scale relations of intra-village spheres of ceramic practice. For Rowlands (1998), relational identity only appears as a product of contact with “others.” In Middle Bronze Age Germany, relational identity appeared and expanded into more complex signalling through dress or costume as trade networks increased in size. These traits are not apparent earlier, in what Rowlands views as a period of isolation (Rowlands 1998:223). Indeed, Comaroff and Comaroff insist that identity is only constructed to emphasize contrast, not similarity or resemblance. Identity is only performed, or created, when one is faced with others “who are different” (Comaroff and Comaroff 1992:245). Since identity is formed through the constant negotiation and renegotiation of relationships, it can be said that any single identity does not exist as an essentialized construct (Peeples 2011; Rowlands 1998), although “some social identities may maintain strong perceptions of permanence, which can structure the actions of individuals through time” (see Bentley 1987:24–27; see also Jenkins 2000; Jones 1997:84–105; Peeples 2011:2). 20 In the greater field of identity studies, there has been a tendency to conflate identity to singular categories, such as race, sex, and religion. The deep integration of ethnographic studies has been instrumental in producing more nuanced views of identity within the field of anthropology (Somers 1994). This new view of identity generally aligns with Creese’s use of the fractal nature of personhood in Iroquoian society (Creese 2012), which is formed through everyday interaction with all aspects, both human and material, of the surrounding world, but also meaning that an object belonging to a person constitutes a partible and living extension of the person in question. The replication of these interactions at progressively smaller scales correlates well with the commonly accepted definition of identity as “the way individuals relate themselves and others to larger groups based on perceived similarities and differences socially defined as important” (Peeples 2011:2; see also Barth 1969; Diaz-Andreu et al. 2005; Emberling 1997; Jenkins 2004:1–8; Jones 1997). In the archaeological literature, the term “ethnicity” continues to reference a great variety of social group identities. Often the terms “ethnicity” and “identity” are used interchangeably within a single context (Peeples 2011), and yet these words are not synonymous. The conditions necessary for a group identity to be labelled ethnicity are currently the subject of much debate among scholars, though most acknowledge that ethnicity is only one of multiple forms of collective identity. Current understandings of ethnicity largely adopt Jones’s perspective (1997), as demonstrated by Diaz-Andreu’s recently formulated definition: “that aspect of a person’s selfconceptualization and his or her conceptualization by other individuals that result from identification with one or more broader groups in opposition to others, on the basis of perceived cultural differentiation or common descent” (Diaz-Andreu 2015:4818). In many ways, this varies greatly from the term’s earliest usage as bounded, static, genetically related groups that shared material and ideological cultural traditions. As is evident in the above definition, many researchers in anthropology and archaeology view ethnic identity or ethnicity (sensu Comaroff and Comaroff 1992) as an actively created identity consolidating a group of disenfranchised or politically overpowered and “oppressed” persons with a self-perceived common heritage, who actively engage “ethnic” symbols or signals as unifying elements. The display and embodiment of symbols in specific public contexts powerfully integrates peoples of profoundly varying genetic and cultural backgrounds who might, in other contexts, not view themselves as members of interrelated units. This 21 confederation may embody a method of resistance to the economic and political supremacy of others (e.g., the Assembly of First Nations, created to resist systemic injustice of the Indian Act of the government of Canada). Some researchers insist that there can be no ethnicity and, in some cases, no identity without power asymmetries (Jenkins 1997; Weber 1922). Ethnicity is a problematic concept (Huntley 2004). A number of historically known ethnic and tribal groups were created as a response to contact with state-level society (Bentley 1987; Emberling 1997; Jones 1997; Lockwood 1981; MacEachern 1998; Shennan 1989; A.D. Smith 1986). Researchers in the field of sociology and anthropology argue that ethnicity is either uniquely a product of Indigenous/European colonization (Comaroff and Comaroff 1992), or, as presented by Gellner (1983) and Nash (1987), it constitutes a product of capitalism (Emberling 1997:307–308). Emberling (1997) also argues that few studies have been able to identify ethnic divisions in stateless societies. He thus suggests that the concept of ethnicity is not relevant in non-state studies. In contrast, Stone proposes that ethnicity, as a socially constructed category, “is evoked during interactions between individuals of different social identity” and therefore involved “the channeling of power relative to other individuals and groups” (Stone 2003:40). She argues that ethnicity is not the exclusive domain of groups interacting at the state level, but exists in any situation of power difference at all levels (Stone 2003:41). Supporting this view, Cameron (2013) suggests that power differences occur in any situation where a group, whether large or small, is moving into an area occupied by, or claimed by, others, as seen in the incorporation of nonlocal ceramic practices common to fifteenth- to seventeenth-century Iroquoian communities. As proposed by Emberling (1997), ethnogenesis is accepted here to be the creation of an ethnic identity as primarily occurring in contexts of inequality, where creation of a unified identity is employed as an act of social dominance or resistance. In adopting this position, I accept that the terms “ethnic identity” and “ethnicity” may be acceptable in the context of exclusive ceramic chaîne opératoire, as they reflect differing cultural understandings within the pan-Iroquoian scale. This categorical identity may co-exist with multiple relational collective identities. Therefore, I endorse group identity in terms of the unbounded and clinal perception of relationships, as being formed, following the social distance (Lyons and Clark 2008) model, of ever-increasing waves of interconnectedness (Abbott 2000) in concert with more categorical identities. The analysis of multiscalar collective identities will include the concept of Iroquoian 22 personhood (Creese 2012) as polyvalent and fractal (after Fowler 2004). Participation in multiple ceramic community of practice groups is viewed as a reflection of the self-ascription to multifaceted identity. Each aspect of which is dependent upon the fluid and situational nature of these communities multiscalar group identities in the fifteenth-century Iroquoian context will be examined as negotiated and relational. Identity and Communities of Practice As discussed, identity is multifaceted, “fractal,” multiscalar, fluid, and emergent. Only a minute fragment of any group identity can ever be discerned through archaeological study of material culture. However, communities of practice, at whatever level, constitute a significant aspect of group identity that can be inferred through a study of their material culture. A community of practice is a group of people who are tied together through “mutual engagement” in prolonged execution of some process of creation according to a set of mutually accepted guidelines or standards (Wenger 1998). The social identity and relations of the group members are reinforced and reproduced in the product of this process (Hodder 1982). Changes in the identity and membership are constantly in flux through time and across geographical and social distance. Ethnographic analogy has clearly shown that identity materializes as people socially construct themselves (Mauss 1935). Material expressions of the artisans reflect these changes (Sassaman and Rudolphi 2001:408). Ongoing engagement within the community and the maintenance of co-production ties constantly reproduce the community despite accrued changes. Membership within one community of practice does not preclude involvement in others, and each person identifies, whether consciously or not, with multiple communities. Identity is the reflection of the relations created and employed in the commission of tasks as they are reproduced through tracing similarities among social groups. These groups are defined by association with an archaeological location, such as a household or village community. Practice communities, in the case of Iroquoian ceramics, have no clear boundaries but may, instead, be characterised by ever-expanding circles of weaker social ties as expressed in decreasing similarity of artifacts. In this study, the communities of practice to be studied are those relating to ceramic production, use, and disposal (Gifford-Gonzalez 2014; Gonzalez 2014; Mills 2016:256). With the exception of a minor group of ceramics, which exhibit unique form and motif 23 execution, the majority of ceramic producers within the Keffer village belong to the larger north shore Iroquoian community of practice. In Iroquoian studies, pottery production is accepted as gender specific and is generally regarded as being an element of women’s domain of practice as documented for the “Huron” of the historic period by the French missionary Father Gabriel Sagard (Wrong 1939) and in the ethnographic writings compiled by Elisabeth Tooker (1991:59).Trigger (1978:60) emphasized the strict division of labour between sexes. This strict separation of tasks and Sagard’s allocation of pottery production as a female endeavour together provide a strong argument for the interpretation of ceramic messaging as a reflection of female identity. Engelbrecht (2003:87), following Fenton (1978:303), suggests that domestic ceramics “stood as symbols of family and hospitality” in Iroquoian society. In Iroquoian society women were central to the family units. As kin identity was traced through the female lineage in Iroquoian society, female identity was at the core of the domestic social structure. Women were the permanent residents of the household and longhouse. The fragile nature of marriage resulted in frequent divorce, with men often entering and leaving the family unit (Tooker 1984). Brown (1970, 1975) proposed that, as producers and distributors of food, women yielded great economic, social, and political influence, suggesting it outweighed that of the men. Tooker (1984) and Trigger (1978) refute this interpretation on several grounds. They believe Brown has misinterpreted the economic contribution of men in the precontact period, and therefore the relative social positions of the genders. As Tooker (1984:118) points out, social relations in Iroquoian society were not based on economic or political control, as manifest in ownership. It is reciprocal obligations embedded in these, mostly kin-based, relations that structured society. For Trigger, “Matrilocality and the influential role played by women in matters concerning their extended families were closely related to their role in the economic pattern of their communities and the activity patterns associated with this role. The result was not a matriarchy, but equality based on the separation and complementarity of sexual roles” (1978:61). Like men, women had their own sphere of influence and power (Mann 2000). Thus, the household was the nexus of female influence and power and the central portal of identity. It was in the domain of these matrilineally based family groups that ceramic production was centred. Therefore, it is the members of these families, as practice groups, whose identity was communicated in this production. 24 Ontario Iroquoian ceramics of the fifteenth century followed a single vessel form. These vessels had globular to round bodies with wide mouths and collared rims. These thick-walled vessels had a high centre of gravity and were designed for cooking (Latta 1991). Collars were an integral element of this form. Thickening of the rim through the application of collars added to the strength of the vessel neck, whereas shaping of the collar improves pouring and filling while providing surfaces for easier handling (Martelle 2002:239). This pottery may also have served for storage and transport purposes (Holterman 2007), although Martelle suggests very small pots seen in collections from the turn of the seventeenth century may have been used for individual servings (Martelle 2002). Residue analysis has produced no clear distinction in function by size (Martelle 2002:199). Vessels of similar shape are found in sizes ranging from the 1.5 litre capacity size to “cauldrons” or “kettles” of 10 litres or more capacity (Martelle 2002:214). The majority of vessels conform to a medium size of 4 to 5 litres and were most likely used most often. All of these ceramics would have been on display in a central location in the household within domestic or ceremonial contexts at some time. In the domestic arena, specifically when placed on the central fire, the collars of these vessels would have been the central visual focus of the pot and contained identity messaging (Chilton 1996). Decoration on vessel necks, directly below the collar, would also have been visible. Martelle, citing ethnographic evidence, suggests different size vessels were used in different contexts (2002:233). Medium-sized vessels may have been used on a quotidian basis, ever present on the central fire with the meals for the day. Family members helped themselves at will from collective family pots (Waugh 1916:46). Decorative elements, in this case, would be intended to communicate messages of inclusion and unity. Hospitality was a fundamental tenet of Iroquoian culture (Latta 1991; Tooker 1984; Trigger 1976:51; Wright 1999). Cooking and pottery were central to social and political events such as feasting and burial ceremonies. The largest kettles were prominent in feasting, where the vessel remained at the central fire, while individuals, seated around the fire, were served from it (Martelle 2002:233). Messages of identity inculcated in these ceramics may have served a similar purpose or a quite different purpose, but again it is the identity of the matrilineal, matrilocal family that is being projected. Ceramic projection of male identity has been hypothesized to exist in the creation and use of smoking pipes (Woolfrey et al. 1976). 25 Since the Iroquoian groups are matrilineal and practice matrilocal residence, it has long been assumed that pottery production occurred within the context of household communities of practice. The smallest units of communities of ceramic practice are composed of the female members of matrilineally based extended family households. Each residence is therefore considered to represent a single practice community. Increasing scale—from residence to longhouse cluster, village community, river drainage basin and the village clusters within—is accompanied by increasing social distance seen in decreasing similarity in the profiles of ceramic type frequency. With the exception of a minor group of ceramics determined to be of nonlocal origin, which exhibit unique form and motif execution, the majority of ceramic practitioners within the Keffer village share a larger pan-regional community of practice, that of the southcentral Ontario Iroquoians. At the most intimate level, a community of practice might include only those who produced and fired their ceramics in a single location and within a single fire. Due to the small size of the ceramic samples available for several residences at Keffer, I employ the household, defined as a single longhouse and its associated middens, as the lowest discernible level of community. As household membership is continually in a state of flux in Iroquoian communities, so, too, is membership within the community of practice. The accepted ceramic repertoire of these groups is also in flux (Birch and Williamson 2018), although some groups may be more conservative, experimental, or innovative. Learning within the community is situational and hands-on (Crown 2014, Dorland 2018). Acquiring expertise is a gradual process, involving interaction and participation (Dorland 2018; Striker et al. 2017). Those with less skill first engage on the periphery of production; they are drawn towards the group core as skill levels increase, as skill develops through continuous learning (Lave and Wenger 1991). As practitioners move into, through, and out of the group core, the repertoire of practice and resulting assemblages gradually experiences change (Birch and Williamson 2018). With this change, the identities of both the individual, and the group, have potential for transformation. The emergence of new elements of practice may occur at any tempo. New attributes may develop over the long course of the village occupation or within a more abrupt moment of change. Production technique is likely to change more gradually, at the temporal scale of generations (Arnold 1998:357–358), while decorative transformation occurs more rapidly, over the short term. Changes in both production technique and decorative repertoire, some intentional and directed, others more passive and obscure, can 26 mark changes in group identity. All change, and sometimes continuity, is the result of decision making and marks the implementation of agency within the community (Dobres 2000) as existing forms are repositioned socially, or “resocialized,” and newly introduced ones become standard (Wendrich 2012:16). Interaction and Social Network Analysis Until recently, few quantification procedures were available for the analysis and presentation of social distance. The introduction of social network analysis into material culture studies in archaeology (Brughmans 2010; Knappett 2013; Mills et al. 2015; Peeples 2011) has provided a useful approach for portraying these relations in both absolute and relative terms, taking into account the caveat that archaeological data sets are always partial samples of unknown wholes and whose absolute measures remain speculative. Most social network analysis in the past has concentrated on large-scale, regional or pan-regional, interaction (Tom Brughmans, personal communication, March 2018; Knappett 2018). Studies at the regional inter- and intra-site level have recently been introduced but remain in a small minority (see Blair 2015; Pailes 2014). The research presented in this thesis will employ social network analysis of ceramic vessel types to determine the dynamics of social relationship and social distance of groups across the Keffer site and also among its neighbours in the Don River valley and farther afield. This work will attempt to address temporal, spatial, and social dimensions of identity through ceramic proxies in order to historicize relational constructions of identity among the Keffer village occupants and their pan-Iroquoian relations. I view human culture as inextricably intertwined with interpersonal communication and relations, whether taking the form of artistic expressions of the experienced world, trade, the sharing of everyday utilitarian items, or storytelling throughout time. Communication, in whatever form, has connected peoples of the world throughout history and prehistory (Sindbaek 2013) because societies have never existed or developed in isolation, devoid of outside human contact (Wolf 1982). The essential role of social interaction in the explanation of spatial distributions of archaeological material remains must be acknowledged (LaBianca and Scham 2006). The influence and incorporation of “habitus” (Bourdieu 1977) and structuration (Giddens 1979), and the works of such British archaeologists as Ian Hodder, Christopher Tilley, and Michael Shanks during the 1980s, injected human agency into the explanation of past cultures. In 27 this new perspective, humans were seen, not as passive victims of external pressures, but as active agents consciously or unconsciously producing and affecting change. This shift in anthropological theory coincided with the initial integration or first participation of Ontario and North American indigenous peoples within the field of archaeology. The involvement of descendant individuals forced archaeologists to acknowledge that the ancestors, the subjects of our studies, were actual persons. While these innovations were taking place at the central core of archaeological theory the definition and characteristics of culture itself were also undergoing change as archaeologists slowly began to “abandon the view that societies or cultures were closed or tightly bounded units” (Trigger 2006:437). There was a realization that, as diffusionist scholars of the past had suggested, cultures do not develop in total isolation from other human contact (Wolf 1982), and researchers interested in cultural similarities began to investigate the multidirectional movement of information and contact (Hodder 1982:1). But interaction studies were not new. Archaeologists had long been tracing a variety of network forms, such as the Hopewell Interaction Sphere (Caldwell 1964; Streuver 1963), Mesoamerican obsidian trade routes (Drennan 1984; Zeitlan 1982), the spread of the “Olmec Horizon” (Coe 1962), and Inca road networks (Hyslop 1984; Regal 1936; Topic and Topic 1983). In the influential work of Renfrew and Cherry on peer polity, the effects of social entities of similar size were seen as co-dependents, affecting change upon each other through non-oppressive, egalitarian mechanisms (Renfrew 1986; Renfrew and Cherry 1986). Williamson and Robertson’s (1984) revised peer polity model for northeastern Iroquoians was one the first examples of the application of this new way of looking at cultural exchange in Ontario. It marked an increased interest in theoretical research in exploring the many and diverse factors that influence culture change (Williamson and Robertson 1994). This work was in direct opposition to studies which focused on centralized sources of power, where an Iroquoian village’s largest longhouses, labelled “chiefs’ houses,” were viewed as being the source of historical acts of power (Ramsden 1990) and where the underlying social dynamics which facilitated power were unrecognized. Interaction in pre-contact indigenous culture was not directly explored. David Anthony, with his study of migration that analyzed culture change in the context of larger social dynamics, exerted a great influence on interaction studies (Anthony 1990). Anthony’s approach provided a theoretical stepping stone for studies of group movement (Peeples 2011; Stone 2003). Trigger suggested that archaeologists of this period studying the Northeast no longer viewed societies as 28 bounded, closed systems, but saw them as developing within the constraints and influences of the greater social network (Trigger 2006:439). However, it seems apparent that, to the contrary, a number of Iroquoianists continued studying “ethnic” “tribal” groups in isolation for many years (Hart and Brumbach 2003). During this same period, in the field of sociology, researchers who had always been deeply concerned with human action began to investigate direct relationships among individuals. This effort to understand individual practices and interpersonal relations and the structure of societies was based on the philosophy of Georg Simmel, who proposed that the networks of these relations are the basis of society (Simmel 1922). Each social connection formed a fundamental element of a greater web of relationships (kinship, friendship, citizenship, etc.). During the 1970s, these insights were coalesced into the field of social network analysis (Knappett 2013:3). Social network analysis (SNA), is not meant to be either a theory or a methodology for problem solving (for a contrary opinion, see Borgatti and Halgin 2011). Rather, it is an approach providing an alternative paradigm aimed at examining human culture through the investigation of interaction and relationships (Brughmans 2010:277; Marin and Wellman 2009:23). As Wasserman and Faust state, “The social environment can be expressed as patterns or regularities in relationships among interacting units” (1994:3). These relationships are drawn using graph theory and portrayed as network graphs. In graph theory, networks have two main components: nodes, which are points on the graph representing entities; and edges, ties, or links, which are lines connecting the nodes and signify direct relationships between them. Interpretation of the network is essentially visual, although mathematical measurements are often used to assess the strength and positioning of the nodes and related ties. In social networks, the nodes represent individuals or groups of people, while the ties denote some form of relationship, perhaps kinship or possession of similar characteristics. Although sharing the same methodical foundation with other social sciences, the study of social networks in archaeology is differentiated by one main, and quite substantial, distinction. Nodes in other social sciences generally represent living beings, whose relationships can be ascertained first-hand through direct observation. In archaeology, nodes must be constructed from excavated remains. The creation of a node entails close scrutiny in reference to its suitability for the proposed research. Although any entity can be used as a node, I propose that the selection of inappropriate archaeological units can lead to meaningless or misleading results. 29 Archaeological “network analysis” of both hierarchical and non-stratified societies, like those mentioned above, is deficient in formal methodology and generally follows an amorphous set of constructs dealing with power, wealth, and population distributions to describe and classify (prevailing) societal frameworks (Van der Leeuw 2013:2). The resulting connections demonstrated between archeological entities lack clarity and precision of definition. In the respect that these analyses merely trace the flow of material culture across space, they do not reject any of the long-held normative paradigms characterizing culture as static and people as lacking choice. Social network analysis, in contrast, endeavours to investigate the network causes of a phenomenon (Hanne man and Riddle 2005) and is based on the underlying principle that a regular pattern in relationships, or interaction, “equals” the “structure” of a social network, “a set of nodes (or network numbers) that are tied by one or more sets of relations” (Wasserman and Faust 1994:3). This structure is uncovered and recreated through analysis of the quantifiable variables of closeness, degree, and centrality of the edges, or ties (Sindbaek 2013). While the overall objectives of any research project may be much broader, the three primary goals of the analysis of social networks are the elucidation of the centrality measures of individual nodes; the detection of communities of nodes; and the characterization of weak and strong links, or relationships (Knappett 2013). These results are found through an examination of the qualities, such as strength or weakness, of ties between nodes, not the characteristics of the nodes themselves. In the social sciences, relationships are categorized as economic, political, interactional, and affective (Wasserman and Faust 1994). They are commonly investigated as individual networks, entirely separate from, and unrelated to, the others. In most cases, the network, or groups of nodes, investigated is given some form of arbitrary limits but in fact, social networks have no boundaries (Borgatti and Halgin 2011). Social networks are also dynamic and constantly in a state of flux as the structure of society is changing and therefore “always becoming” or emergent (Van der Leeuw 2013:2). The characteristics of interaction, and therefore the structure of the network, can change gradually or abruptly due to shifting internal or external conditions (Sindbaek 2013). Although SNA can reveal factors affecting a transformation, the results cannot be applied to a situation with the prospect of predicting the form a network will take as an outcome of change. SNA provides direction for in-depth research (Marin and Wellman 2009:23). It is not meant to provide simplistic resolutions to complex inquiries. 30 Despite the profound ideological shift in Northeast archaeology, our understanding of late precontact history remains grounded in the culture history of the 1960s and 1970s, and much of current research focuses on dating cultural traits and mapping their occurrence (Beisaw 2010:244; Martelle 2002:9–10). Hart and Brumbach acknowledge this and criticize current use of culture-historical taxa, which, they insist, should not be assumed to bear “any social or cultural reality in the past” (Hart and Brumbach 2003:750). In an effort to overcome this, I have rejected the use of the controversial ethnic labels employed by MacNeish for pre-contact Ontario Iroquoian peoples, replacing these with references to geographical areas. In terms of New York state pre-contact Iroquoians, I continue to use these ethnic labels, but only in the sense that they represent geographically oriented territories of separate communities. At its core, human society is essentially a vast network of multiscalar interconnecting social relations. These relations are the basis of identity at all levels because there can be no identity in isolation. Relations, and therefore identity, occur along a scale of intimacy and distance that is in constant flux. The female potters of the Keffer village, as members of matrilineal and matrilocal households, communicate diverse and expanding spheres of collective female identity and social relations through the production and use of three domains of pottery. These domains are (1) pottery characteristic of the local north shore tradition, (2) exotic pottery acquired through regional and pan-regional interaction, and (3) locally produced vessels created as symbols of nonlocal relations and identity. These ceramics lend themselves to exploration, through network analysis of the fluid temporal and social multiscalar relations of intra-community, inter-regional, and pan-regional identity as reflected in ceramic communities of practice. These communities emerge, flourish, and recede with changing social relations realized in ceramic vessels. 3 Background Iroquoian Development Figure 3-1. The geographical extent of Northern Iroquoia. The Keffer village is a fifteenth-century pre-contact community, part of the Iroquoian occupation of the north shore of Lake Ontario, hereafter called the north shore Iroquoians. As such, it is one manifestation in the long, continuous sequence of Iroquoian and ancestral Iroquoian peoples in the North American Northeast (Figure 3-1). Iroquoian peoples, characterized in the pre-contact period by a common language; a lifestyle centred around egalitarian social structure; sedentary, maize-based agriculture; distinctive bone, stone, and ceramic material culture; and villages of bark-covered matrilineal, matrilocal, multifamily residences, have occupied lands surrounding the eastern section of the lower Great Lakes and the St. Lawrence valley from the turn of the twelfth century to the present day. By the mid-seventeenth century, territories in central Ontario and along the shores of the St. Lawrence River were generally abandoned, while the New York 31 32 state Iroquois and surrounding Iroquoian groups receded from the south shores of these large waterways as a result of various factors, including the Wars of the Iroquois (1649–1651). The Iroquoian occupation has been traced back in time almost two thousand years, to the Early Woodland occupation of Ontario, Quebec, New York, and surrounding areas in the northern United States. Early efforts to make sense of the archaeological record followed approaches of the taxonomic systems common in the early to mid-twentieth century, namely, the Midwestern Taxonomic System of McKern (1939) and James Wright’s ground-breaking Ontario Iroquois Tradition (1966), which established a framework for the chronological and spatial placement of Iroquoian sites in the province, giving structure to the limited, but growing, data available. The rapid expansion of this database in the past half century, along with changing theoretical perspectives towards more relational thinking, have moved views of Iroquoian culture beyond this paradigm. Based on the still relevant central divisions of the Woodland occupations of Ontario, but rejecting the concepts of phases and horizons, Creese (2010:26) has presented a broad framework that he terms the Ontario Iroquoian Tradition (after Ellis and Ferris 1990). Divisions within this framework are flexible and are based on the gradual changes in settlement and subsistence patterns seen in the archaeological record. Patterns seen in the south-central Ontario tradition, as described below, are generally replicated in neighbouring New York State and along the St. Lawrence River. As interaction with Iroquoian groups of these areas is central to this work, I will compare datasets from this region with the ceramic assemblages at Keffer. In this system, gradually changing characteristics of the Woodland occupation of the Northeast form the basis for the creation of three main periods: the Early, Middle, and Late Woodland. Archaeological remains from the Early Woodland, dating from the ninth to second centuries B.C., document small, mobile, extended family groups subsisting on a foraging, fishing, and hunting–based lifestyle. These groups are found throughout southern Ontario and into Quebec and western New York (Spence et al. 1990:125). The first appearance of pottery in the area occurred in the Woodland period (Spence et al. 1990:125). Maize appeared around 1,000 B.C. in New York state (Thompson et al. 2004). 33 Figure 3-2. The Woodland period in south-central Ontario. The Early Woodland occupations of Ontario are closely related to contemporary occupations of southern Quebec and New York state. The late Early Woodland Middlesex complex (Figure 3-2) is represented almost exclusively by burial contexts in the northern United States and along the St. Lawrence River. The appearance of Adena-like artifacts suggests widespread macroregional shared decorative traits, indicating regional-scale interaction (Ferris 2003; Hart 2000; Hart and Brumbach 2003; Spence et al. 1990). A gradual and spatially inconsistent transformation in settlement and subsistence patterns heralds the Middle Woodland stage, circa the beginning of the fourth century A.D. During this time, foraging, fishing, and hunting remained the main subsistence sources, but food acquisition strategies differed from the earlier period. Evidence would suggest a growing dependence on a seasonal movement subsistence strategy, following an annual round within a band-based territory (Creese 2010:27; Smith 2000). During the warmer months, microband groups converged in large, seasonal campsites in areas of large-scale seasonal food resources available in centralized locations, such as river mouths (Spence et al. 1990:168). These locations show evidence of longterm use and may have been a focal point of annual rounds. In the winter months, the larger 34 groups dissolved into small, scattered family groups. Seasonal movements resulted in longdistance interaction as a routine characteristic of this period (Ferris 2003; Hart 2000; Hart et al. 2003). Late Woodland I: Princess Point, A.D. 500–1000 The gradual change to a semi-sedentary lifestyle began in the closing of the Middle Woodland, as the earliest Late Woodland lifeways, termed Princess Point, emerge in Ontario’s southwest, while hunter-gatherer-foragers continue to inhabit much of the province (Smith 1997). In this period of innovation, the first introduction of maize “gardening” occurs within floodplain areas of the lower Grand River; at Cootes Paradise, at the west end of Lake Ontario; in the Waterloo area; and along the north shore of Lake Erie (Crawford and Smith 1996, 2002; Smith and Crawford 1997; Crawford et al. 1997, 1998; Warrick 2000). Similar settlements are found to the west, in the Rivière au Vase tradition, and to the east, in the Sandbanks tradition of Prince Edward County (Warrick 2000). Evidence of large-scale warm-season gatherings is seen in the poorly preserved, scattered temporary wood framed shelters (Smith and Crawford 2002). This marks a change in the annual round, as these riverine settlements became the permanent, central hub of small family groups (Smith 2000). Towards the end of Princess Point, intensification of maize cultivation laid the foundations for the following Glen Meyer subperiod. Glen Meyer peoples were fully dependent on cultivated crop species (Ounjian 1998). Settlement and social systems experienced major transitions in adopting crop-based subsistence patterns (Ferris 2003). Late Woodland II: Early Ontario Iroquoian, A.D. 1000–1300 The ensuing Late Woodland II, Early Ontario Iroquoian period in central and southwestern Ontario sees the proliferation of nucleated, sedentary, semi-permanent settlements between A.D. 900 and 1100. These small villages of four to five 10–20 m longhouses (Dodd 1984) held 75 to 150 people. Most settlements are now located in upland areas, away from navigable waterways (MacDonald 2002; Pearce and Williamson 1999) and are sometimes palisaded (Warrick 2000; Williamson 1990). Many episodes of rebuilding in the same place meant certain locations could experience long-term use (Kapches 1993:11). Sites such as Calvert may have experienced up to one hundred years of occupation (Timmins 1997). Settlements were located in distinct, autonomous clusters by A.D. 1150 (Timmins 1997; Williamson 2014). Kapches (1990) argues that limited house floor space and heterogeneous pottery within villages suggest the absence of 35 matrilocal residence at this time. Each settlement is economically and politically independent, as reflected in the heterogeneity in ceramics between site clusters (Timmins 1997:239). Maize accounts for up to 20 percent of the diet by the end of the thirteenth century (Katzenberg et al. 1995; Warrick 2000:437; Williamson 2014:9), but traditional foraging, hunting, and fishing provide important food sources (Williamson 1985). Seasonal movements in search of these resources bring people out of the villages for large portions of the year (Kapches 1987; Timmins 1997). Changes in village organization over the course of the Calvert occupation indicate increasing social integration, the existence of matrilineal and matrilocal residence within the larger longhouses of Early Iroquoian villages, and the “socio-political institutions” necessary for village planning (Timmins 1997:239-240). Similarly, after intensive analysis of house and village settlement plans, Creese proposes that the new “physicality of the longhouse” reflects a new sense of belonging and place with membership in co-residential longhouse groups, and that the forming of community occurs through “an increase in the anticipated stability of the constituent social groups” (Creese 2010:34). This subsistence settlement pattern is adopted at different times by different groups throughout southern Ontario and New York, with nucleated settlements not appearing until after A.D. 1300 in some areas (Hart and Brumbach 2003; Snow 1996; Warrick 2000). Increasing and unprecedented interaction between village clusters occurs towards the Early Iroquoian period, resulting in province-wide similarity in material culture, specifically in the pervasive linearimpressed oblique ceramic motif (Timmins 1997:240). By the end of the period, increasing population numbers and village fission leads to the migration from isolated clusters of two or more villages, to colonize new areas across southern Ontario and New York (MacDonald and Williamson 1995; Niemczycki 1984; Warrick 2000). Late Woodland III: Middle Iroquoian, A.D. 1300–1420 The rapid population growth leads to dramatic transformations in subsistence, settlement, and socio-political organization throughout Iroquoia during the Middle Iroquoian period of the fourteenth and early fifteenth centuries. By the fourteenth century, most traits deemed characteristic of Iroquoian culture are widespread (Trigger 1976:91–104; Williamson 2014:9). An overall trend in the homogenization of subsistence, settlement, and socio-political strategies, along with the inter-related processes of population expansion and increasing maize dependency, 36 are the major hallmarks of this time. These mechanisms, the result of increasing interaction among newly established, independent community clusters, led to greater similarity overall. A major marker of increased cultural uniformity is seen in the widespread adoption of the horizontal ceramic motif across Iroquoia, which first appeared in the previous century (Dodd et al. 1990; Wright 1966). At the same time, the geographic isolation of the clusters accounts for preservation of distinctiveness within communities. The first three decades of the fourteenth century experienced a steady but rapid growth in population. This resulted in the expansion of both longhouses and villages throughout Ontario and New York (Engelbrecht 1985; Niemczycki 1984; Warrick 1990). Although population growth is rapid, the large increase in settlement size was due to the merging of some smaller communities into villages (Creese 2010; Dodd et al. 1990; Warrick 1990). Small settlements, characterized by single clusters of four to five longhouses, held groups of 250 to 350, while large villages held 400 to 500 occupants (Warrick 2000:440). Social stress caused by the congregation of these large numbers results in the first examples of village fission (Warrick 2000:441). The fluidity of membership within household groups and villages became an integral component of Iroquoian social relationships hereafter (Hart and Brumbach 2009; Warrick 2000:444). The emergence of kin–based, sometimes fictive kin, longhouse groups, often referred to as clans, and the formalization of matrilineages and matrilocal residence, provided the basis of the social framework (Kapches 1995; Warrick 1984) that is essential for the integration of diverse groups (Warrick 2000). Although in later Wendat communities, marriage is documented to have taken place between local community members within a family’s circle (Tooker 1991:125–128; Trigger 1990a:78–79), existing ties within the marital residence were based on female descent. New mechanisms for the integration of the large numbers of unrelated men brought together through the practice of matrilocal residence include small-scale rituals of group sweats. These are seen in the presence of semi-subterranean sweat lodges in fourteenth-century Iroquoian longhouses (MacDonald 1988), larger ossuary burials, and the accompanying feasts. The co-mingling of bones of the deceased from each house in the village, and perhaps multiple villages, creates both the performance and the concretization of social unity (Ferris and Spence 1995). 37 After 1330, the population exploded (Warrick 2000:439). The heavy reliance on maize, to the point where it now made up as much as 50 percent of the diet (Williamson 2014), was not the sole cause of this population growth; improved childhood health appears to be a major factor (Warrick 2000). As a result of this large-scale growth, Middle Iroquoian villages often experienced fission, as small groups left the community and migrated into new areas of southern Ontario. Small village clusters appeared to the southwest, almost to the territory of the ancestral Western Basin Algonquians (Lennox and Fitzgerald 1990); along the north shore of Lake Ontario; eastwards into the St. Lawrence valley (Jamieson 1990); and, for the first time, in the Historic Wendat homeland of Simcoe County (Sutton 1999). The population explosion also led to an increase in longhouse length (Dodd 1984) and a concomitant increase in house numbers. Village size was highly variable, and a single village at this time could hold more than 660 people. Integration of these large numbers was facilitated by the organization of the village into kin-based clusters with an axial focus on outside open space (Creese 2010). Semi-subterranean sweat lodges become ubiquitous in the last half of the fourteenth century. Social mediation occurred through the performance of large-scale feasts, evidenced by the appearance of large feasting vessels at this time (Finlayson 1988; Tooker 1991:73), and smaller-scale rituals, evidenced by the appearance of elaborately decorated pipes (Smith 1997). The development of kinship-based clans (whether this kinship is fictive or otherwise) may have been one major mechanism of social integration (Engelbrecht 1985; Trigger 1985; Warrick 2000). Late Woodland III: Pre-Contact Iroquoian, A.D. 1420–1550 The early fifteenth century saw stabilization in the rate of population growth throughout Iroquoian Ontario (Warrick 2000:452). However, the increase in population due to the earlier “baby boom” (Warrick 2000:449) resulted in continued expansion of communities in the area southwest of Lake Ontario (Lennox and Fitzgerald 1990); into Waterloo County (Lennox and Fitzgerald 1990); into Simcoe County (Sutton 1999); eastward along the north shore of Lake Ontario (Birch and Williamson 2013); into the lower Trent valley (Ramsden 1990b; Sutton 1990); and still farther east, into Prince Edward County (Adams 2003). Small settlement clusters within restricted physiographic boundaries, generally river watersheds, characterized this period. Geographic separation and clustering of community groups were instrumental in new regional diversity. 38 Small, unpalisaded villages display a range of settlement variability in both size and configuration, but generally are composed of one or two clusters of three to six aligned longhouses (Birch 2010:51; Lennox and Fitzgerald 1990; Warrick 2000). Spacing of longhouses within settlements is inconsistent. In some cases, no clustering is apparent, while in others the houses are highly clustered (Birch and Williamson 2013). Birch suggests that settlements of two clusters, such as the mid-fifteenth-century Alexandra, Hope, and Orion-Murphy-Goulding sites, formed as the first instances of community amalgamation within drainages (Birch 2010:84). Formation of larger social entities within these growing communities likely resulted in the creation of larger production groups beneficial for agricultural production (Dodd 1984). The absence of palisades reinforces the argument that communities did not aggregate for defensive purposes (Birch 2010). The presence of numerous unprotected associated hamlets, agricultural cabins, and other special-purpose buildings in the vicinity of the villages indicates that subsistence activities kept large portions of the population away from the villages for most of the year, spring to fall (Warrick 2000:449), further supporting the supposition that there was a lack of concern for defence in the early part of the fifteenth century. Community congregation coincided with an unprecedented increase in house size. Growth within communities in the form of multiple, large house expansions, not in the creation of new houses; 30 percent of longhouses in the fifteenth century show evidence of expansion. In the first half of the fifteenth century, longhouses across Ontario increased in length to an average of 62 m, from 35 m in the preceding period (Dodd et al. 1990). Structures in the Neutral area expanded to incredible sizes, some being more than 90 m in length (Warrick 2000), and one house at the Coleman site reached the remarkable length of 124 m (MacDonald 1986). The increasing size of membership within households may be the cause of, or may reflect, the growth of kin-based corporate groups focused on subsistence activities. Meanwhile, the sustained need for social integration mechanisms is documented by the continuing use of semi-subterranean sweat lodges into the middle of the fifteenth century. The establishment of local, spatially separated community clusters oriented around geophysical land features led to increased regionalization across southern Ontario (Ramsden 1990a; Warrick 2000). Ramsden notes that a strong regional difference in ceramic traits characterized this period within ancestral Wendat populations (Ramsden 1990a:381). The scarcity of exotic materials and artifacts in these unpalisaded communities indicates regional isolation and limited long-distance 39 interaction in the early decades of the century (Dermarkar 2013; Lennox and Fitzgerald 1990:427). This pattern continued in southwestern Ontario throughout most of the fifteenth century, possibly due to social isolation. Neutral area groups were more widely spaced than any other Iroquoian groups in this period. However, they likely did not constitute a homogenous entity, despite similarity in ceramics (Anderson 2009; Lennox and Fitzgerald 1990). Lack of evidence for long-distance exchange in Simcoe County during the first half of the fifteenth century suggests that a similar pattern of isolation may have existed in this area (Stopp 1982). The introduction of nonlocal ceramics, steatite, and exotic shell artifacts to the north shore of Lake Ontario began towards the middle of the century, with the first significant collection documented at the Joseph Picard site, in the recently defined Oshawa cluster (Williamson 2016b). As the first signs of long-distance interaction and perhaps the first immigration of small, nonlocal social groups (Williamson 2016b) began to occur, large-scale change took place across most of Iroquoia. This transformation did not occur in ancestral Wendake. Long-distance interaction increased across most of Iroquoia throughout the mid-century (Kuhn 2007:336; Snow 1994:37; Warrick 2000). The presence of Onondaga chert and quantities of eastern Durfee Underlined and Roebuck Low Collar pottery on mid-century north shore sites points to harmonious relations with both the western Neutral and the eastern Iroquoians (Williamson and Robertson 1998:147–149; Warrick 2008:157). Construction of new villages along the southern reaches of the Canadian Shield indicates increasing trade between central Ontario Iroquoians and Shield Algonquians (Warrick 2000:451). The increased interaction seen in the Late Woodland III period (Kuhn 2004; Snow 1994:37; Warrick 2008) contrasts with the growing differentiation among regional clusters across Iroquoia as large-scale settlement nucleation begins at both the community and the regional level. In Ontario, communities grew in size in the mid-fifteenth century as existing villages experienced the influx of social groups of various sizes. As in the previous periods, village and household membership remain fluid (Engelbrecht 1995; Niemczycki 1991:32; Snow 1994:84). This inherent social flexibility was noted in the seventeenth century by the Jesuit Jejune (Thwaites 1896-1901). While these immigrant groups may have had nonlocal origins (Robertson and Williamson1998:147), the increase in village size generally reflected the integration of social groups from various communities within the local watershed. At the Draper village site in the 40 Rouge–Duffins watershed large-scale growth resulted from the coalescence of several complete village settlements from within the watershed (Birch 2010, 2012; Birch and Williamson 2013). This same form of coalescence may also have occurred at the Lawson site far to the west and the Coulter site in the Trent valley (Damkjar1982; Lennox and Fitzgerald 1990; Ramsden 1990). This contraction of population distribution is commonly attributed to the rise of incessant intertribal warfare (Warrick 2000:446). Larger, 2–4 ha villages are surrounded by multiple palisades (Warrick 2008:155). Defensive earthworks in combination with palisades distinguish the Neutral and St. Lawrence regions at such sites as Lawson and Southwold (Anderson 2009; Wintemberg 1939), as well as in upper New York—the Onondaga Burke site being one example (Tuck 1971:126). Scattered human skeletal elements and human bone artifacts are found on many of these sites, including the Neutral Lawson site; the Toronto-area Damiani, JarrettLahmer, Keffer, and Parsons sites; the Rouge–Duffins-area Draper site; the Prince Edward County Lite site; and sites in the St. Lawrence valley and New York (ASI 2015; Finlayson et al. 1985; Jenkins 2016; Pendergast 1972; Pradzynski 2013; Rainey 2002; Warrick 2008; Williamson 2000; Williamson and Pfeiffer 2003). However, archaeologists have interpreted the presence of substantial quantities of nonlocal ceramics within villages throughout the latter half of the fifteenth century (Pearce 1978; Pendergast 1980; Pihl 1984; Robertson and Williamson 1998; Smith 1991) as an indication of either peaceful, long-distance interaction or absorption of refugees (Finlayson, Smith et al. 1987; Warrick 2000:451). Conflict appears to have been local or regional in nature in Ontario (Anderson 2009:16; Dupras and Pratte 1998). South of Lake Ontario, hostilities among contemporary ancestral Haudenosaunee seemingly occurred between regions, as local groups appear to have already established formalized alliances within historic territories (Anderson 2009:16). The middle of the fifteenth century sees increased regionalism and warfare throughout Iroquoia (Kuhn 2004), but this occurred earlier in Ontario than in New York and may have been on a different scale (Birch and Williamson 2013:22). Endemic warfare lessened in intensity as communities abandoned the mid- to late fifteenth-century villages. Heavily fortified villages were built, with multiple rows of palisades, but other evidence of conflict was slight. At the turn of the sixteenth century, changes in settlement location and socio-political organization continued. The most striking of these changes is the fusion of previously dispersed regional communities into restricted territories (Engelbrecht 1999). By 1550, all Neutral area 41 villages were located to the east of the Grand River (Lennox and Fitzgerald 1990:411; Warrick 2000) and north shore communities moved north along the valleys of the south slopes of the Oak Ridges Moraine, away from the north shore of Lake Ontario (Williamson 2014). Jefferson County in northern New York was abandoned (Abel 2002, 2017), and St. Lawrence Iroquoians also relocated. Villages at this time were much fewer in number, larger in scale, and spaced at a greater distance (Birch 2010:52; Lennox and Fitzgerald 1990; Warrick 2008). Large villages on the upper reach of the north shore were characterized by smaller, tightly spaced houses in parallel alignment. The decline in house length, to less than 30 m on average, may represent a gradual change in social structure from matrilineal lineages to clan segments (Warrick 2000:449) and more fluid group membership (Engelbrecht 1985:15–17). By the end of the sixteenth century, most of these large villages fissioned, with the component communities moving across the Oak Ridges Moraine into Historic Wendake (Birch 2010:53) and possibly joining groups in the upper Trent River valley (Ramsden 1990a:375). With the final abandonment of the upper Humber and Rouge River drainages, permanent habitation on the central north shore of Lake Ontario ceased as villages flourished in the Historic Wendake homeland. A consolidated presence of the ancestral Neutral remained at the western end of Lake Ontario into the contact period. Keffer Keffer Site Environment The Keffer site represents the remains of a 2.1 hectare (Stewart 1991:2), ancestral Wendat village composed of 18 longhouses and 26 middens and an encircling palisade. It is the last known of a presumed series of settlements of the Don River Iroquoian community, located on the upper reaches of the Don River drainage system, north of present-day Toronto in the province of Ontario, Canada. Preliminary settlement and ceramic studies have suggested a tentative date for the Keffer site occupation to the mid- to late fifteenth century of the Late Woodland period (Finlayson et al. 1985). BCal standard calibration (http://bcal.shef.ac.uk/; Buck et al. 1999) of seventeen AMS radiocarbon dates from Keffer interior pot residue lend support to this chronology. This analysis was run without a priori assumptions about chronology because none exist for the Keffer site. At the 95 percent probability level, the results have a tendency to fall between A.D. 1380–1550 (Appendix A) with a reasonable amount of overlap. The mode, or 42 most common, result lies between A.D. 1450–1510. A newly proposed chronology for Iroquoian sites of the Rouge–Duffins drainage sequence (Manning et al. 2018) raises some questions concerning the conventionally accepted chronology of Ontario Iroquoian sites. These results suggest that much more research will be necessary in this field to clarify these issues. With the present information, a mid- to late fifteenth-century date is put forth for the occupation of the Keffer site. However, in this present study, ordinal rather than absolute chronometric positioning of Keffer and other Iroquoian sites anchor the analysis. A second source for Keffer’s chronological position is X-ray fluorescence analysis of two Keffer copper pieces, one scrap and one rolled bead, performed by Brandi-Lee MacDonald. Both pieces are of indigenous manufacture, supporting the assignment of the Keffer site to the precontact era (Appendix B). Keffer’s appearance followed the large-scale growth in settlements throughout the upper Lake Ontario river watersheds that occurred in the earlier decades of the century. The village is situated on land located on the east bank of a small, unnamed tributary of the upper West Don River, less than 2 km south of the Oak Ridges Moraine. The moraine, a major geological feature of the region, divides the north- and south-flowing drainage systems of south-central Ontario to the north of Toronto. The Don River is the major drainage system of the central and eastern portions of Toronto (Figure 3-3). 43 Figure 3-3. Location of the Keffer site in the Don River watershed. (modified from Toronto and Region Conservation Authority ca. 2008) 44 The Keffer village is located at a linear distance of 24 km from the shore of Lake Ontario. This distance increases greatly if one follows the course of the water flow along the tributaries of the drainage system, a route that was perhaps the preferred path of least resistance for resource gathering groups. With distance from the shoreline comes increasing elevation, as the southern slopes of the Oak Ridges Moraine rise to the north. The village site sits at an elevation of 205– 208 m ASL, 125 m above the level of Lake Ontario. The site is located on a steep-sided, bi-level terrace (Figure 3-4), 5–8 m above the Don River tributary below. Surface runoff and agricultural ploughing have caused the erosion of several longhouse extremities, including end walls, particularly near the western terrace edge. The village occupies both levels of the terrace; the upper village is located on a 3 m high plateau overlooking a tributary of the Don River to its west and a low-lying plain, or depression which encircles it to the north, east, and south (Spence 1988). The plateau comprises roughly a third of the village area (Figure 3-4). Figure 3-4. The bi-level plateau of the Keffer site. (Ontario Museum of Archaeology) The Keffer site is situated within the Peel Plain physiographic region (Chapman and Putman 1984:174–175, Figure 3-5). This region is crossed from north to south by several watersheds draining into Lake Ontario and is characterized by pockets of a variety of soil types resulting from ancient glacial activity in the area. 45 Figure 3-5. Glacial deposits in the vicinity of the Keffer site. (modified from Ontario Geological Survey 1980) The Peel Till Plain, a widespread glacial deposit of boulders, clay, and sandy silt till deposited in the last ice age, underlies large swaths of this area north of Lake Ontario. Two sizeable pockets of glacial pond–deposited Malton Clay, a stone-free lacustrine clay, define the immediate vicinity of the site to both the north and south (Figure 3-6, Figure 3-7). Southern lobes of the Oak Ridges Moraine extend southward to within a couple of kilometres of the site. 46 Figure 3-6. Soils in the vicinity of the Keffer site. (modified from Agriculture Canada 1955) 47 The site itself is located on a bed of Chiguacousey clay loam with imperfect drainage. This deposit stretches to the east and south between the east and west branches of the Don River, while soils to the north and east consists of clay deposits of the Peel Plain less advantageous for agriculture (Agriculture Canada 1955). Tracts of well-drained Oneida clay loam are present within 1–2 km of the site. The steep slope of the Peel Plain, rising more than 120 m from the shoreline northward, has facilitated the formation of deep valleys cut by both the Humber and Don Rivers. Two small outcrops of the underlying shale bedrock occur along the Don River to the south of the Keffer site, along with small areas of sandstone and chert-bearing limestone (Kerr and Spelt 1965:14, in Stewart 1991:9). This till plain geology resulted in an area of welldrained, agriculturally productive soils supporting a climax deciduous forest of mixed hardwood species, dominated by sugar maple, beech, and oak, along with hickory, basswood, and white pine. White ash, elm, and white cedar predominated in depression areas (Chapman and Putnam 1984:289). No swamps or bogs are present on this till plain (Stewart 1991:9), although large wetlands existed at the mouth of the Don River. With consistent levels of precipitation occurring throughout the year and a long frost-free spell, from late May to October, this upper south-slopes area provided an excellent venue for maize horticulture (Stewart 1991a:10), with access to a wide variety of faunal and floral species and rock and clay sources. Early Research at the Keffer Site David Boyle, of the Canadian Institute, initially reported the existence and partial excavation of a large ossuary located on Lot 12, Concession 3, in the Township of Vaughan in 1889 (1907:16). The site itself was first noted and mapped in detail by A.J. Clarke in 1925 (Figure 3-7) in the course of Clarke’s archaeological surveys of the Toronto area, during which he explored, recorded, and mapped the location of numerous sites. A small collection of artifacts acquired during this exploration now resides at the Canadian Museum of History in Gatineau, Quebec. Clark’s map portrays the village site as occupying parts of both Lot 12 and Lot 13 of Concession 3, with the associated ossuary located below a local farm barn erected in 1884 on a knoll in the southwestern section of Lot 12 (Spence 1988). Victor Konrad reported visiting the village site in the early 1970s (Konrad 1973), during his survey of the archaeological resources of the Metropolitan Toronto Planning Area. Konrad registered the village site as Borden number AkGv-14 and the ossuary as AkGv-15 with the Ontario Ministry of Citizenship and Culture. Although Konrad considered that the protected location of the Keffer ossuary lessened the threat 48 of modern disturbance, he suggested that the 6–10 acre village site, already subject to partial disturbance, was threatened by development in the near future. Figure 3-7. Map of the Keffer village site. (Clark 1925) 49 Keffer Site Excavation, 1984 and 1988 The proposed development of Lot 13, Concession 3, by Magna International in the early 1980s led to renewed interest in this archaeological site. In 1984, the Museum of Indian Archaeology (MIA) was hired to assess and, subsequently, salvage-excavate the northern portion of the site, on lands slated for development by the Magna Corporation. Peter Timmins undertook the initial reconnaissance and testing, in the fall of 1984 (Timmins 1984). During this investigation, seven extensive exploratory trenches were laid out across the presumed occupation area. These trenches were mechanically stripped of the plough zone soil, after which areas of interest were shovel shined. The excavators uncovered the remains of a two-row palisade in six of the seven trenches (Finlayson et al. 1985a). These preliminary investigations suggested that 1.56 ha of the site located within Lot 13 were threatened by development. At this time, there were no immediate plans for development of Lot 12, the southern portion of the site. On May 6, 1985, the MIA began salvage operations of the northern area of the site, having been awarded the contract in March (Finlayson et al. 1985a). The large scale of the Keffer project was made possible by funding from all three levels of government and private corporate support. Independent grids of slightly different orientation were laid on the upper and lower levels of the site prior to excavation. Disturbed middens were identified in the ploughed portion of the site through site survey. Middens were then excavated by shovel in one metre squares and the matrix was screened through 6.4 mm mesh (Stewart 1991:3). Basal middens appear under surface midden deposits at depths of 10–40 cm. Therefore, natural stratigraphy was followed where it was present. Where stratigraphy was not apparent, middens were excavated in arbitrary, 10 cm levels. Following excavation of the middens, the remainder of the soil in the ploughed portion of the site was removed mechanically, without screening for cultural material (Finlayson et al. 1985a:3). Post moulds and features in the subsoil were exposed with the mechanical stripping. Features, specifically subsurface pits and some posts, were excavated by hand, sectioned, and drawn, and the fill was screened on 6.4 mm mesh for cultural remains. Flotation samples were taken from all features. Post moulds and features were measured by triangulation from the southwest grid stake of the square. All information was recorded on standardized field note forms. Grid squares were then plotted on personal computers on-site. After being printed, the field forms were verified for accuracy. Human burials found within the village were investigated and reported on by Michael Spence (1988). The appearance of a burial ground of undefined limits exterior to the village 50 palisade and the requirement for further testing for a sewer right-of-way led to further exploratory trenching from beyond the outer palisade in the late fall of 1985. Figure 3-8. Aerial view of Lot 13, 1985 excavations, and wooded Lot 12 bottom right. Further plans for development on Lot 12, Concession 3, prompted MIA to excavate the remainder of the Keffer village site. A large portion of Lot 12, unlike Lot 13, had not been subject to agricultural ploughing and hosted a hardwood forest (Figure 3-8). Excavation of this area followed a different methodology. This portion of the site was excavated by hand in one metre squares. The grid layout continued that of the 1985 excavation of the upper part of the site. The short time allotted for this second excavation season meant that while the houses were hand excavated in one metre squares, sizable areas between the houses and palisades were only investigated through the use of one metre test trenches in the village’s southern portion. In addition to providing descriptions of the 1985 excavations, the Keffer field report (Finlayson et al. 1985) offered interpretations of the temporal and cultural placement of the site within the greater Ontario Iroquoian occupation. The report proposed an early fifteenth-century date for the 51 site based on preliminary analysis of the settlement data and ceramics and noted the unusual bilevel occupation of the site for a “Huron” (now termed ancestral Wendat) village. The superpositioning of structures and the presence of two distinct palisade lines led to the presentation of two differing construction scenarios. The first of these entailed contraction of the village over time, with the removal of numerous longhouses and a reduction of the two-row palisade into a smaller, single-row fence line. An alternate site history proposed village expansion from a small group of six longhouses surrounded by a one-row palisade on the lower level of the site to a larger, bi-level settlement surrounded by a two-row palisade. The replacement, expansion, and renovation of some lower-town houses would have occurred in conjunction with the erection of this new palisade. The settlement history of the upper plateau area is not discussed in this interpretation, perhaps due to incomplete nature of the excavation in this area at the time of publication. The Rouge–Duffins drainage Draper site, at that time the only other extensively excavated contemporary north shore Iroquoian village, provided the most appropriate comparison for settlement pattern and material culture. Although house structures were found to be very similar to those of Draper, the overall shorter length of the Keffer longhouses led the authors to suggest an earlier date for the Keffer site. The team also noted a much lower feature-to-postmould ratio for Keffer. Possible socioeconomic implications of this difference remain unresolved. Similar to Draper, however, the Keffer site contained a large number of broken and burnt human skeletal elements (n=1,237) within its middens and houses (Rainey 2002; Spence and Rainey 2017:25), pointing to warfare of at least regional nature during the fifteenth century (Finlayson et al. 1985:8). The presence of two deer scapula pipes, thought to be characteristic of the St. Lawrence Iroquoians, and a small percentage of St. Lawrence Iroquoian ceramics provide possible evidence of long-distance interaction, perhaps warfare, with St. Lawrence Iroquoians to the east (Finlayson et al. 1985:9). The authors also argued that the ceramics indicate similar contact with Seneca to the south and west (Finlayson et al. 1985:9). The large-scale excavation operation at Keffer provided essential information on the complexity and diversity of Iroquoian villages, enabling more in-depth study of transformations in settlement pattern into the early Late Iroquoian period (Finlayson et al. 1985:9–10). Preliminary reports of the first excavations presented an initial analysis of botanical samples undertaken by Charles Turton (Finlayson et al. 1985:8), which documented the presence of the 52 agricultural cultigens maize, beans, squash, sunflower, and tobacco and also provided evidence of long-term occupation for the Keffer village. According to Turton, the presence of maple and beech charcoal at the base of a village midden overlain by strata containing evidence of more varied mixed trees marks the initial settlement in an established forest. The non-climax trees seen in later deposits suggest that the area was entirely cleared of the original forest and that forest regrowth occurred during the occupation period (Finlayson et al. 1985:9). The use of the secondary growth forest by Keffer residents argues for a long duration of occupation of the village. Finlayson and his team published a summary of the 1985 excavation season and a description of the artifacts, in 1987 (Finlayson et al. 1987), prior to the 1988 excavations of the southern portion of the village. Large-scale research continued as part of the project into the 1990s. Published reports include David G. Smith’s (1991) ceramic study, Frances Stewart’s (1991) analysis of Keffer site fauna and Dawn Wright’s (1991) examination of botanical remains. Numerous unpublished studies (Appendix C) add to large body of research undertaken on the Keffer site. In addition, Keffer has been the subject of several academic presentations and publications. These include the work of John Creese on the Keffer pipes and identity (2013, 2014), my own (Dermarkar 2012, 2013a, 2013b, 2014b, 2016) presentations on the Keffer site and its ceramic relations, Tara Jenkins’ work on the human bone artifacts (Jenkins 2014, 2016), Michael Spence and Dori Rainey’s (2017) study of the cultural significance of the Keffer site human burials and most recently Steve Dorland’s (2018) thesis on learner vessels. In summary, the Keffer site is the remains of a Late Iroquoian village located on a small tributary of the upper Don River, north of Toronto and Lake Ontario. Keffer sits on a bed of clay loam and is surrounded by many pockets of arable land on a glacial till plain just south of the Oak Ridges Moraine. The local vicinity is characterized by a variety of floral, faunal, and geological resources, and the climate is amenable to seasonal agriculture. The presence of an Iroquoian village site in this location has long been known. Previous research on the Keffer village has established it as a fifteenth-century ancestral Wendat village which experienced multiple changes in its settlement pattern throughout its history. External relationships may have been characterized by warfare, as seen in scattered human bone and perhaps less violent interaction, as evidenced in Keffer’s nonlocal ceramics. 53 Although the ceramics of the Keffer site have previously been examined and reported on by David Smith (1991), the main purpose of his report was to provide a preliminary introduction to the ceramic and pipe data and an exploratory review of the decorative variation found among longhouse and middens locations. Smith also examined the utility of ceramic attribute combinations and typology as analytical units (Smith 1991:1). A significant amount of research on the habitation of the village has occurred over the past 40 years. The earliest studies, focused on material culture description of the large collection available for the Keffer site, have supported some ground-breaking work in archaeological method and theory (Creese 2013, 2014; Dorland 2016a, 2016b, 2017, 2018). The body of research briefly described above has not yet been compiled into a cohesive record of the Keffer occupation and its place within greater Iroquoian culture. Although this is not the purpose of this thesis, information from these works will be incorporated here where appropriate. 54 4 Method: Ceramic Type and Attribute Analysis Ceramic Method and Theory Ceramics have played a central role in the study of Iroquoian society since the earliest days of research (Beauchamp 1894, 1900; Boyle 1889, 1891, 1892, 1896, 1897a, 1897b, 1898, 1902; Hunter 1900, 1902, 1907; Laidlaw 1891, 1917; Wintemberg 1900, 1901, 1903). Early research concentrated on the collection of artifacts from ash pits and middens visible in the agricultural fields of Ontario and New York. Ceramic study focused on description and illustration, with an emphasis on traits of form and decorative style (Emerson 1954; Houghton 1916; Jury and Jury 1955; Ridley 1952b). The appearance of the unique Iroquoian ceramics and other material culture within an area surrounded by Algonkian groups presented a puzzle to researchers of the day. Archaeologists proposed a model of intrusive migration from the south into the lower Great Lakes region to explain this anomaly (Beauchamp 1894, 1900; Parker 1916, 1922; Skinner 1921). As early as 1939, Phileo Nash at the University of Toronto hypothesized a local origin for the Northern Iroquoians based on his work at the Pound site, where he noted continuity in ceramic practice from earlier occupations (Pearce 1996:23–24). In the 1940s, the in situ hypothesis became the subject of great interest (Griffin 1944; Kraus 1944). In order to test this hypothesis, Richard MacNeish developed a pottery typology based on his examination of assemblages from all regions of Northern Iroquoia. The resulting typological series, published in Iroquois Pottery Types (MacNeish 1952), dominated Ontario Iroquoian pottery analysis until the mid-1960s, and it continues to be a mainstay of Ontario Iroquoian ceramic study. As recently summarized by Matthew Peeples, types “are descriptive units that identify sets of objects that share some, though not generally all, analytically important characteristics. The selection of these “important” characteristics often renders the typology a subjective one (Peeples 2011:239). While typologies can be the product of a systematic combination of attributes (see Gifford and Smith 1978), in MacNeish’s taxonomy, ceramics are grouped and arranged according to MacNeish’s intuitive separation of vessels, derived in large part from examination of the rim and collar attributes as entire vessels are rarely available in Iroquoian assemblages. He cites an earlier joint statement with William Ritchie to explain the approach: We conceive of the typing of archaeological materials as an attempt to recapture the stylistic concepts in the minds of the original makers…. We believe that, at 55 least to a certain degree, our types reflect aesthetic and utilitarian standards of value which operated as cultural compulsives on the minds of artisans and therefore, they possess some genuine measure of intrinsic validity [Ritchie and MacNeish 1949:98, cited in MacNeish 1952:2]. MacNeish used the analysis of ceramic rims, both their shape and their decoration, combined into types, to define ethnic identity and trace peoples through time. Each of his ceramic types has been directly designated to one of the major “ethnic” groups. It is almost universally accepted today that some communities did, in fact, remain within a single river drainage area, moving within its boundaries on a fairly regular basis (see Abel 2002; Bamann et al. 1992; Birch 2010; Birch and Williamson 2013; Bradley 1987; Pearce 1996; Snow 1997; Tuck 1971; Warrick and Molnar 1986; Wray and Schoff 1953). Other Iroquoian individuals, families, and communities, on the other hand, often relocated into either uninhabited or occupied, distant or adjacent, lands from their home communities, sometimes into existing villages. Widespread interaction throughout the region is reflected in shared ceramic types and attributes. Although MacNeish (1952:2) assigned each of his ceramic types to a historically documented, ethnic tribal nation, this identity is not intrinsic to the nature of typology. Types, as well as attributes, can be employed as tools distinct from any implied ethnic connotations. As they are used today, types are handy tools for preliminary ceramic analysis, providing a large-scale overview of general ceramic affinities and reflecting some level of social relations (Birch et al. 2017). Newer approaches to typology in use in Britain today, such as emergence (Lee 2006) and assemblage theory (DeLanda 2006, 2016), have not yet been introduced into precontact Iroquoian archaeology. Yet these linked perspectives offer theoretical support to the recent introduction of relational network studies within the discipline (Dermarkar et al. 2016, Hart and Engelbrecht 2012; Hart et al. 2016). Of particular importance to this study is Chris Fowler’s (2017) presentation of a relational interpretation of typology which grounded in assemblage theory following Lucas (2012). In contrast to the MacNeish-Ritchie view of types as static concepts in the mind of the ceramicist Lucas argues that types in fact have no essence. There is no unchanging idea, or mental template. Types are a process of recurrent citation and are composed of “constituent populations of things” that are produced from “a similar iterative process” (Fowler 2017:99). The object, here ceramic vessels, and the type, perhaps Huron Incised, are co-emergent. As the vessel is formed, and the action repeated, the type arises. In this 56 view types are not static and inflexible. New elements may be introduced in successive iterations but they will be removed in other iterations if they prove ineffective. In this way types are an important tool in detecting relations and making sense of different past ways of becoming (Fowler 2017:104). This new perspective on typology provides a path forward for the incorporation of ceramic types in Iroquoian archaeology after decades of deliberation regarding its utility. Debate on the meaning and validity of types (Ford 1954; Ford and Steward 1954; Pratt 1960; Spaulding 1953; Whallon 1972, 1980) of the typological series has defined lower Great Lakes archeology since its inception in the early 1950s (MacNeish 1952; Ritchie and MacNeish 1949). However, in Ontario, change did not occur until 1966, when J.V. Wright published a critique, in his seminal work The Ontario Iroquois Tradition, regarding the limitations of typologies, including their subjective and therefore irreplicable nature. As an alternative he proposed their replacement with individual variable analysis. Attribute analysis became the focus of major studies during the 1970s and 1980s following the work of Peter Ramsden (1977). Attribute analysis is available for a large number of collections. However, in Ontario the introduction of attribute analysis has not been entirely successful due to the lack of a common methodology (ASI 2010c; Pendergast 1973; Ramsden 1977; Smith 1990). A common ceramic attribute database therefore does not exist. Since roughly 1990, however, many analyses of Iroquoian pottery have included both typological and attribute analysis. While attribute analysis has been successfully employed for the production of ceramic-based chronologies and the reconstruction of social relations across the province in individual and incomparable studies (Ramsden 1977; Smith 1990), MacNeish’s typology has most often been employed to create chronological site and regional sequences in Ontario Iroquoian studies (Bursey 1993; Wright 1966). In fact, while Wright espoused the use of attribute analysis, he himself relied on the MacNeish typology in his production of the framework of Ontario Iroquoian culture predominant during the latter half of the twentieth century. The ubiquitous use of typology (following Lenig 1965; MacNeish 1952; Noble 1968, 1974; Pratt 1976; Wright 1966) in Ontario, and the production of a large quantity of ceramic type data from New York state over the past half century, much of it available in the published literature, makes comparable large-scale ceramic data easily obtainable and provides an extensive databases for analysis. 57 The assumption that the occurrence of ceramic types and attributes varied in a patterned way over time and that specific types pre-date others across the southern portion of the province formed the basis of the production and interpretation of regional site sequences. The relative chronological position of sites was explored through seriation, using the battleship-shaped curve or the production of Brainerd Robinson co-efficient of similarity (BR) matrix of the relative frequencies of assemblage types (Bursey 1993; Chapdelaine 1989; Emerson 1954; Engelbrecht 1971; Niemczycki 1984; Pendergast 1984; Ramsden 1977; Smith 1990; Wright 1966). It is now clear, however, that while there may be general temporal patterns of popularity among certain “time-sensitive” ceramic types and traits, percentages of these may vary greatly among drainage basin–delineated community clusters (Birch 2010; Birch and Williamson 2013; Williamson and Powis 1998; also see Ramsden 1977). It is imperative, therefore, in assessing the chronological position of ceramic collections, to be cognisant of the geographical, and often resulting cultural, associations of the sites employed in the BR analysis. Network analysis of ceramic attributes presents some problems. The formation of the data matrix employs a sole variable for comparison of the ceramic samples investigated. This results in the selection of only one attribute for comparison, as seen in the collar motif network analysis of Iroquoian ceramics (Dermarkar et al. 2016; Hart and Birch 2017; Hart and Engelbrecht 2012; Hart et al. 2015a, 2015b; Hart et al. 2017), or the use of attribute combinations consolidated into a single variable. In either case, network analysis of attributes entails the ranking of vessel attributes in relation to the questions posed. The specific attributes deemed important by the producer and the consumer—and according to which the vessel was formed or chosen—remain beyond our understanding. While it may be possible to propose a rank order for ceramic attributes according to specific hypotheses, the multiscalar tacking (sensu Wylie 1989) among several scales of social relations employed in this research, from the household to the panIroquoian level, requires an analytical unit of ceramic measurement applicable at multiples scales. Attributes or traits portray only specific elements of practice. These must be combined with other specific vessel attributes, deemed meaningful to the research question, to illustrate social practice above the level of one component of action in the production of pottery. Communities of practice occur at multiple scales simultaneously. At these differing scales they can primarily reflect aesthetic orientations of production and use, as seen in Ann Stahl’s (2013) examination of figurines in the Volta Basin, or explorations of production technique (Dorland 58 2018). Comparison of types in some instances, as is evident in the MacNeish typology, can reflect diversity in many aspects of the chaîne opératoire of two vessels, as is evident when comparing Otstungo Notched everted-lip and Durfee Underlined pots. In contrast with the comparison of Huron Incised and Sidey Notched vessels, only small aesthetic modifications to the rim are defined. Therefore, the ability of a type to reflect a single component of production, a complete chaîne opératoire of ceramic practice at the level of the individual completed vessel, or a larger household or village community of practice through the examination of entire assemblages makes typology well suited to multiscalar analysis. In this study, I employ types in various combinations at descending geographic and social scales to unveil decreasing social distance and increasingly intimate social relations as represented by ceramic use and production. However, the exploration of nuanced relationships between assemblages often entails investigation at the smaller scale, i.e., that of attribute variables. Single attributes may be employed to explore agency at the level of the macro-community or at the level of micro-scale influences and, perhaps, direct contact. For instance, the relative quantities of neck decorated vessels present in an assemblage may aid in the chronological placement of an occupation within a regional grouping or cluster. In contrast, the relationship between occupants of two longhouses within a site might require a micro-scale, nuanced analysis of the number of horizontal lines decorating the neck of vessels in those locations. Although some Iroquoian ceramic types, such as Black Necked and Pound Necked, are prized as chronological markers, others are conjectured to have geographical and socio-cultural affiliations. This is explicit in reference to the St. Lawrence Iroquoian ceramic types Roebuck Low Collar and Durfee Underlined, which occur at Keffer, as well as Corn Ear, Lanorie Corded, Lanorie Crossed, Lanorie Mixed, and Swarthout Dentate, which do not appear at Keffer. MacNeish originally determined Roebuck Low Collar and Durfee Underlined to be of Onondaga ancestry (MacNeish 1952:60–61). These types have now all been designated as St. Lawrence Iroquoian in origin (Dermarkar et al. 2016; Funk 2004:150; Hart et al. 2016). Investigation into the origin and distribution, both pan-regionally and across the village, of Rice Diagonal and Otstungo Notched, two “Mohawk” (MacNeish 1952:74–75) types that occur in significant numbers at Keffer, also provides multiscalar ceramic relations. In contrast, I view types believed to have originated with the pre-contact Neutral in southwestern Ontario, widespread across southern Ontario, as reflecting long-term interregional interaction. 59 Figure 4-1. Keffer site location within pan-Iroquoia A.D. 1350 to 1650. (David Smith) A different level of interaction is seen in the examination of specific ceramic attributes. The presence of nonlocal, St. Lawrence Iroquoian production elements, or attributes: reed punctates (indentations assumed to have been created with the use of cut reeds), basal thumbnail gouges, and finely detailed decorative implementation not generally seen on Ontario Iroquoian ceramics mark the appearance of exotic vessels at the site (Chapdelaine 1991; Jamieson 1990). The minor presence of these attributes in limited loci across the site traces common external ties within the village. The distribution of other attributes, including particular motifs, though lacking in specific geographical or temporal importance, traces distinct practices across the site and in some cases across Iroquoia. In summary, I have chosen ceramic typology following MacNeish and others as the analytical unit anchoring this research. Typed pottery assemblages best represent the agency of the 60 nontechnical aspects of ceramic practice and are useful at multiple scales of social distance. Iroquoian collections in particular are well suited for multiscalar network research because of their overall uniformity and widespread availability. Individual ceramic attributes provide smallscale supplementary insights. History of Research on the Keffer Ceramics Cataloguing of the Keffer site artifact assemblage was performed under the direction of David Smith. The better part of a year was devoted to matching and reconstructing vessels and rim segments, facilitating further analysis of the ceramics. The large size of the collection made it impractical to lay out all of the pottery for extensive visual comparison. Thus, only ceramics from adjoining units within specific site features (that is, houses or middens) were laid out in the lab at any one time, according to their original grid system provenience relationships. These ceramics were then roughly sorted by type or highly visible attribute similarities and compared among the excavation units for vessels match, whether “empirical” (i.e., actual physical mends) or “inferred” (i.e., close similarity of shape and decoration) (Smith 1991:3). Rim sherds were selected for coding and analysis on the basis of some simple criteria. Foremost of these was the presence of the entire exterior design sequence, encompassing at least part of the vessel lip, the complete upper rim exterior, and enough of the neck to ascertain its primary design. Presence of the interior surface was not deemed crucial for inclusion in the study. The sherds matching these criteria came from almost 5,000 individual vessels. All sherds not assigned to a vessel rim section and not meeting these criteria, which numbered in the thousands, were considered fragmentary and were removed from the analyzable collection. The rims included in the original analysis were initially sorted within their larger house or midden loci according to the basic typological system proposed by MacNeish. Each rim was then labelled with a unique “Unit” identification number directly related to its provenance, or locus of origin. In general, the vessels were ordered sequentially, generally according to ceramic type, ranging from the simplest exterior collar decoration to the most complex exterior collar decoration. For example, Huron Incised rims originating in Midden 65 might be sequentially labelled as 650001, 650002, 650003, etc., while those of the Lalonde High Collared type, with complex collar motifs, could be much farther along in the sequence, labelled 650222, 650223, 650224, etc. With the large number of vessels in this collection, this cataloguing method makes 61 identification of vessel origins immediately obvious at the macro, village level. Following field procedure, ceramics originating in Houses 1 to 20 are labelled beginning with the digits 01 to 20, while ceramics originating in middens are labelled beginning with the digits 60 to 80. Information on specific intra-house feature location of artifacts is available in the catalogue database. Due to the small number of features found within houses (Finlayson et al. 1985), intrahouse-level analysis proved uninformative. While preliminary analysis of the collection was a focus of the initial research on the ceramics, Smith (1991:xi) states that one of the main objectives of the work was to make the data accessible and useful for other researchers. To facilitate future analysis an extensive list of attribute variables was included. The code developed by Smith for his Ph.D. research on Middleport ceramics (Smith 1997) was modified for the Keffer site ceramic analysis. The cataloguing, coding, and typing of the ceramic vessels was partially completed by temporary employees supported by the Ontario government’s Summer Employment Program. These individuals had little or no archaeological background and little experience or training in ceramic coding and identification. The resulting coding and typological assignments of the rim sherds therefore tended to be inconsistent. General Procedures Followed in the Current Analysis Early in this research, the need for a re-evaluation of the nearly 5,000 vessels to assure coding uniformity throughout the collection became evident. This re-evaluation ensured that the results could be comparable with those of other assemblages coded using Smith’s Version 4. In the present analysis, material from each locus was laid out according to the original cataloguing scheme, generally by type. This resulted in vessels of similar characteristics being displayed roughly adjacent to one another, which enabled the matching of many rim segments previously attributed to separate vessels. In addition, sherds deemed to be too fragmentary for coding in the previous study were re-evaluated for inclusion in this research, since the criteria previously used were restrictive to certain forms and prioritized some variables to the detriment of the greater study (i.e., evidence of complete neck motif over its presence/absence). With the integration of these additional vessels and the ability to visually match rim segments with others from the same house or midden, the analyzable assemblage has grown to a total of a minimum number of 5,674 vessels. This substantially increased the size of what was already one of the largest collections of 62 vessels in all of Iroquoia. Although MacNeish’s types are the unit of analysis for this study, I also reassessed the attribute coding of the entire collection. In the course of re-analysis of the collection I added several new attributes, dealing with sherd size, reed punctates, and secondary rim notching to the database. I employ these attributes, most notably the presence of reed punctates, in the non-network-based analysis to gain further insights into specific temporal and geographic origins of the vessel. Since type assignment poses several challenges, the specific application of the assigned types in this study requires some explanation. To assure consistency in the intra- and inter-village correlation of ceramic typology, it is essential that I define and describe the Iroquoian ceramic types to be used. MacNeish (1952) provided the basic underlying definitions. I have modified his typological catalogue with some additions and clarifications from later work by Latta (1983), Lenig (1965), Noble (1968, 1974), Pratt (1980), Ridley (1952b), and Wright (1966). I describe issues pertaining to type definitions that have affected the Keffer analysis below. Complete definitions of all types as they are employed in this work are included in Appendix D. It has not been possible to perform comparison through actual sherd-to-sherd assessment for all of the material because the collection has been physically sorted and stored according to the original individual material provenience. The immensity of the Keffer collection and the large size of some of the vessels, ranging from a couple of centimetres to one metre in height and/or diameter, make physical comparisons among the vessels of the various loci untenable and even impossible. Therefore, I created a photographic library of the entire analyzable ceramic rim/vessel collection during the recoding process. The ease of accessing photographs for comparison and analysis purposes also makes the collection now more available and accessible to other researchers, as was intended by the original team. The main purpose of the library is not to produce publication-quality images, but, rather, to create a useful visual catalogue which will be easily accessible for basic visual character information on the collection, without having to pull sherds from storage once again, piece by piece. The photography of all the Keffer vessels, which were sorted and stored primarily according to provenance—most importantly the photography of those with more complex motifs—also allows the exploration of the relationships among large numbers of similar vessels through visual analysis/comparison of the photographs on large viewing screens. 63 Iroquoian Ceramic Typology at Keffer The ceramic collection of the Keffer village comprises 37 individual ceramic types. (A description of all of these types is available in Appendix D.) This assemblage has the largest diversity of types of any Ontario Iroquoian collection known to date. The great majority, 87.2 percent, of the rimsherds included in the current analysis conform to MacNeish’s Huron and Neutral wares, here called local tradition ceramics. Earlier Ontario Iroquois types make up 1.0 percent of the collection, and undecorated collar Niagara and Ripley types represent 1.1 percent. The remaining rimsherds comprise a large variety of types, present in much lower quantities, that share characteristics of shape and decoration distinctive to geographically more distant Iroquoian groups. Generally labelled “exotics” by others (Birch and Williamson 2013; Kuhn 2004:446), these ceramics are here collectively termed nonlocal ceramics. Most of these ceramics appear to have been produced at Keffer or nearby, roughly following decorative formulas originating farther afield, beyond the ancestral south-central Ontario Iroquoian homelands. Although the decorative motif and rim shape, or both, loosely conform to type definitions, the simplification of the motif and lack of precision in the application of the motif suggest a local versus nonlocal production for the majority of these vessels. To confirm this belief, Laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS) was performed by John Creese on vessels of local and nonlocal ceramic types from Keffer. The local types examined in Creese’s study are Huron Incised rims from three separate middens, one Black Necked rim, and one Lalonde High Collar rim. The nonlocal types examined are one Otstungo Notched vessel, one Onondaga Triangular sherd, and several Durfee Underlined vessels. In all, 15 vessels were tested, only one of which was deemed ahead of time to be exotic, of probable St. Lawrence origin due to the precision and complexity of the collar decoration. In Creese’s principal components plot (Figure 4-2); all vessels are represented by blue triangles. The generally homogenous nature of these samples is apparent when compared with the more scattered distribution on the plot of the pipe bowls (represented by green diamonds), pipe stem beads (represented by black circles), and effigy pipes (represented by red squares). The uniformity of the ablation results for these vessels, as seen in their clustering to the middle left of the graph, indicates local, on-site manufacture of the majority of both the local tradition and the nonlocal tradition ceramics sampled. In addition, Creese found that ceramic production waste from across the Keffer site has essentially the same profile of major chemical elements as the 64 majority of the tested pottery vessels, supporting the suggestion of on-site production of both local and nonlocal ceramic types (John Creese, personal communication 2014). Figure 4-2. Keffer ceramic object chemistry principal components analysis of major elements. (courtesy John Creese) A small portion of these nonlocal ceramics at Keffer consists of several vessels deemed to be of foreign manufacture. These are found across the site in association with several diverse features. The hypothesis of off-site manufacture for these vessels is based on subjective judgement of vessel attributes. Primary among these are the elaborate, complex motifs, with fine detail executed with visible precision. The reed punctates motif commonly decorates these vessels. The designation of these vessels as “exotic,” or foreign-made, is supported by Creese’s analysis. The circled blue triangle on the bottom left of the graph represents a finely made Durfee Underlined vessel with reed punctates on the collar. The presence of these vessels of foreign manufacture within the Keffer village indicates long-distance external social ties, materialized in the form of trade items or gifts. Pots located within the main body of the cluster belong to both the local and Haudenosaunee traditions and emergent vessels indicating likely local manufacture 65 Local Tradition Types and Attributes I deem concave interior rim forms on local tradition ceramics in this study to represent the influence of a unique community of practice whose origin archaeologists have often assigned to the southwestern region of Iroquoian Ontario (MacNeish 1952; Wright 1966). Ceramics typical of sites culturally designated as “Neutral,” namely, Lawson Incised and Lawson Opposed, are characteristic of these western ceramic collections of the mid-fifteenth century. They are also present, although in much smaller proportions, on Iroquoian sites throughout most of southcentral Ontario during this period. These mid-fifteenth-century Neutral ceramics are characterized by concave interior rim form and vertical and opposed incised lines (where a single oblique or a set of obliques meet obliques of the opposite direction), respectively. Pound Necked and Pound Blank share this rim form but are most popular in the preceding century. Convex interior rims may reflect both time and area of influence. Convex rims became increasingly common into the seventeenth century, when they form the majority of interior vessel rim shape in Wendat collections. The height of a portion of Iroquoian vessel rims increased from medium (3.5 cm) to high (well over 5 cm) in the fifteenth century. These high collars became less common in the sixteenth century. General height characteristics are subsumed in ceramic types and are reflected to some extent in the network topology. Neck decoration, a general temporal indicator within local community groups, is represented here by the Black Necked and Pound Necked types. Interpretation of temporal position in the network graphs partially relies on the presence or absence of these types in the detailed exploration of resultant ties. The ceramic collar attributes of interior rim form, collar height, and neck decoration are reflected in the categories of local tradition pottery created for this analysis. Interior rim form distinguishes between “Huron” and “Neutral” ceramics. Neck decoration is marked in the categories of decorated neck and undecorated neck. Collar height is subsumed under the Huron, Neutral, and High Collar labels. These specific attributes are integral to this study and integral to the local tradition categories generated. The local tradition ceramics found at the Keffer village are grouped into three major categories, undecorated neck Huron, decorated neck Huron, and Neutral, following MacNeish’s divisions. No tribal or ethnic designation is implied by these labels in this work. Each of the groups comprised one or more individual MacNeish, Noble, Ridley or Wright types. Brief descriptions and reference information for each of the types are given in Appendix D. A list of the types 66 comprising each larger group of ceramics is given at the end of each section, in alphabetical sequence. The distinction between the “Huron” type Huron Incised and the “Neutral” type Lawson Incised is of primary concern in the typing of local southern Ontario Iroquoian vessels. (Note that hereafter, in the text and captions, these terms are mostly used without quotation marks, as a shorthand.) Ceramics designated to be of Huron tradition see their highest frequencies in the geographical area of the Lake Ontario north shore area and Simcoe County. The Neutral types are ceramics originating to the west of Lake Ontario that are accepted as characteristic of the ancestral peoples of the historic-period Neutral area. The name Neutral was applied by the first Europeans in the area to the people whom the historic Wendat called the Attiwandaron. It should be noted that, although MacNeish (1952:14, 34) assigned Huron and Neutral ceramics to different Iroquoian traditions, their true relationship remains unresolved because both types appear to stem from a common Middleport base. Types of both of these traditions have a strong presence at Keffer, though the number of Huron vessels far surpasses the number of Neutral ones. In the fifteenth century, collar decoration, in the form of closely spaced parallel vertical or oblique lines encircling the rim, characterized ceramics of both traditions. Martelle (2202) in her study of late sixteenth—early seventeenth century Wendat ceramics presents a similar argument noting that that the combination of decorative attributes on Huron rims built on the baseline of the Huron Incised motif of a set of either vertical or oblique lines (2002:447). “The majority of decorative types…merely add to modifications to these baselines” (2002:447). The final size, shape and decoration of the vessels varied according to the intended social and functional context of use but Huron Incised was always the baseline from which the potter worked. Martelle interprets the resulting ceramics as transformations which always remain connected to their origin. This “sense of connection is something integral to Iroquoian metaphysics” (2002:448). From a related perspective, I view the presence of the oblique background on central Ontario Iroquoian pottery rims as indicative of the assertion of collective identity at the macro Ontario Iroquoian level. Engelbrecht supports this hypothesis of passive action in the production of ceramics in his assertion that “Women shaped this clay into culturally prescribed [italics mine] forms” (Engelbrecht 2003:87). I believe it is in the performance of alterations to these patterns that the expression of identity occurs. Likewise, plain collars should be viewed as the result of a conscious decision to remove the oblique line 67 background, thus altering the “norm.” These vessels therefore should not be considered undecorated, but, rather, the product of active choice. Production and use of pottery conforming to accepted norms may reflect a sense of inclusion and acceptance. I argue that it is in the modification of these markers of belonging that agency expresses new ways of self and group identification and dynamic associative membership. “Huron” Types One of the primary markers of inclusion in the global pre-contact ceramic practice of the Lake Ontario north shore area Iroquoians is the ubiquitous use of the Huron Incised type ceramic. The Huron Incised type (Figure 4-3) is defined as having “oblique or vertical lines on an outflaring short collared pot with the inner surface opposite the collar straight to convex” (MacNeish 1952:34). For Iroquoians to the west, this is mirrored in the dominance of the Lawson Incised type vessel. Diagnostic features of Lawson Incised are “oblique or vertical lines on short poorlydefined, channeled (convex interior) collars” (MacNeish 1952:14). In the current recoding of the Keffer ceramics, the MacNeish definitions were closely followed, with the poorly defined and channelled collar of Lawson Incised employed as the main traits separating the two types. Figure 4-3. Huron Incised vessel, Midden 60, Keffer. North shore Lake Ontario Iroquoian ceramics are divided here into two main groups, those with neck decoration and those without. For these local tradition types, the presence of geometric 68 decorative motifs below the collar base is deemed temporally sensitive (Ramsden 1977:184; Wright 1966). Undecorated neck vessels of the mid-fifteenth century consist of two distinct ceramic forms, those commonly considered to have Neutral traits—including low collars, convex interior profiles, and simple incised exterior collar motifs (Ramsden 1977)—and those with a straighter high collar form, such as the Lalonde High Collar type, defined by Ridley (1952b), and Noble’s Sopher Incised type, a minor type with a small presence at Keffer. These high collar types represent local manifestations of a widespread practice and are therefore assigned to the independent high collar category described below. In this thesis, the word Neutral is used only in the context of ceramic categorization. It refers to vessel types which I include in a general class based on MacNeish’s (1952) Huron Incised type. Along with interior collar profile—straight or convex in earlier times and predominantly convex towards the period of European contact—the oblique or vertical parallel line pattern is an essential element of the Neutral community of practice. Therefore, Huron Incised is the basic Neutral type around which most other Neutral types are defined. The closely spaced parallel oblique lines, most often incised or stamped, are predominantly sloped from bottom left to upper right of the rim, but they may also slope to the upper left or be entirely vertical. As noted, this pattern forms the basic background common to most Neutral vessels, with other types generally displaying some form of manipulation or superimposition over the basic design. Variations primarily consist of the addition of short, regularly spaced, repetitive incised marks overtop of the pre-existing definitive background of the characteristic parallel oblique or vertical linear incisions on the collar. The simplest variant is embodied in the addition of vertical lip incising, which defines the type Sidey Notched (Figure 4-4). The Copeland Incised, Sidey Crossed, Sidey Notched, and Warminster Crossed types are characterized by slight linear incised variations placed over this basic Huron Incised type motif. Copeland Incised (Figure 4-5) (Ridley 1952b:205; Wright 1966:73) sees the basic oblique line background being completely interrupted by a series of horizontal lines forming panels along the collar. Sidey Crossed (Figure 4-6) involves the superimposition of horizontal lines, either interrupted or continuous, near the base of the collar. 69 Figure 4-4. Sidey Notched vessel, Midden 65, Keffer. Figure 4-5. Copeland Incised motif on Black Necked vessel, Midden 60, Keffer. Figure 4-6. Sidey Crossed vessel, Midden 65, Keffer. 70 In Warminster Crossed (Figure 4-7), superimposed diagonal lines cross the background design at a wider line spacing interval. Seed Incised (Figure 4-8) shares vessel shape with other Neutral ceramics but has a distinctive motif, with a band of short, vertical slashes, or gashes, at the base of the lip and the base of the collar. MacNeish considers this design a later derivative modifying the vertical impressions of the basic Huron Incised motif (MacNeish 1952:35). It appears in small numbers at Keffer. Unlike on the undecorated neck types mentioned above, the Huron Incised oblique background may not be clearly visible on Warminster Horizontal (Figure 4-9) collars. This ceramic shares collar form and shape with other Neutral types but is characterized by superimposed horizontal trailing. This trailing often overwhelms the oblique pattern, which may be visible at the top and the base of the collar. Undecorated neck Neutral vessels increase in frequency with time within their geographically or socially bounded extended communities of practice (Ramsden 1977) (for example, the watershed community of Rouge–Duffins). Figure 4-7. Warminster Crossed vessel, Midden 60, Keffer. 71 Figure 4-8. Seed Incised vessel, Midden 77, Keffer. Figure 4-9. Warminster Horizontal vessel, surface find, Keffer. The Black Necked type (Figure 4-10 and Figure 4-11) is the only local tradition ceramic with prominent and large-scale neck decoration. Single lines of punctates are sometimes found just below the collar on other vessel types. The presence of these punctates does not inform on ceramic type, however. Therefore, the Black Necked vessels represent the attribute of neck decoration in the north shore Iroquoian ceramics of this study and form the entire neck decorated Huron category. High percentages of Black Necked pottery are seen as a marker of an early manifestation of north shore ceramic assemblages (Ramsden1977:184). Black Necked vessels are distinguished from the above types by large-scale decoration on the vessel neck, generally 72 commencing directly below the break of the collar base and often continuing towards the shoulder. While the presence of upper, or whole, neck decoration is the main characteristic of this type, a well-defined collar is also fundamental. Black Necked vessels can be adorned by either encircling horizontal lines at the top of the neck or any combination of vertical or oblique patterns, sometimes quite elaborate. Uncollared vessels or vessels with poorly defined collared rims with encircling horizontal lines are assigned to the earlier, Middleport-era Pound Necked type. They are not considered Neutral vessels. Other ceramic analysts have in the past followed different sets of unstated ceramic typing priorities. A clear statement of procedures used in the application of any typology can and will produce concise, interpretable, and replicable results. The definition of the Black Necked ceramics has raised problems in previous work on the Keffer ceramics because no clear distinction was being made regarding the dominant attribute used for assigning Neutral vessels. Often, without explanation, other attributes, such as lip decoration or collar motif, were given precedence over the presence of neck decoration in typing vessels. In this study, decoration on the neck overrides any decoration of the collar in the definition of Huron types because this is the definitive definition of the Black Necked type. All Huron vessels with significant neck decoration have been retyped as Black Necked. All vessels with substantial upper neck decoration, excluding the presence of a sole linear display of encircling punctates, are typed as Black Necked unless they fall under the auspices of the Middleport Pound Necked or Middleport Oblique (Figure 4-23) types. Where no neck decoration is present on local tradition sherds, rim motif and collar shape are the primary elements used in the definition of types. 73 Figure 4-10. Black Necked vessel, Midden 65, Keffer. Figure 4-11. Black Necked vessel with horizontal motif, Midden 65, Keffer. To summarize, for the purpose of this ceramic analysis, the local Neutral tradition ceramics found at the Keffer village are grouped into three major categories, each of which comprises one or more individual MacNeish, Noble, Ridley, or Wright types. Undecorated neck Huron 74  Copeland Incised  Huron Incised  Seed Incised  Sidey Crossed  Sidey Notched  Warminster Crossed  Warminster Horizontal Decorated neck Huron  Black Necked Neutral Area Types While the motif of the Lawson Incised type (Figure 4-12 and Figure 4-13) is identical to that of the Huron Incised type, other vessel characteristics, predominantly exterior collar shape and interior form, distinguish these two main ceramic traditions. Neutral ceramic identification is based on the works of MacNeish (1952) and Ridley (1961). Neutral types are described as having constricted necks on flaring collared or collarless vessels, and generally short collars (Lennox and Fitzgerald 1990:415). The common modification of the main background oblique pattern of the Huron Incised motif so often seen in Wendat ceramics—like the late-fourteenth- to seventeenth-century increases in the Sidey Crossed, Warminster Crossed, and Copeland incised types—does not occur to the same extent in the Neutral equivalent type (i.e., Lawson Incised). Alteration of this oblique motif is generally only seen in the appearance of the Lawson Opposed motif/type. Lawson Opposed (Figure 4-14) is decorated with bands of opposing obliques, a motif commonly seen adorning the necks of Black Necked Huron vessels. This motif is not common on Huron collars. This type is likely derived from Lawson Incised (MacNeish 1952:14) and is much less common than Lawson Incised, particularly in north shore Wendat sites. Both types 75 appear at Keffer. Since they are closely related and share the time-sensitive characteristic of undecorated necks, they are placed here within the general category of Undecorated Necked Neutral. Figure 4-12. Lawson Incised vessel, Midden 57, Keffer. Figure 4-13. Lawson Incised vessel, Midden 57, Keffer. 76 Figure 4-14. Lawson Opposed vessel, Midden 74, Keffer. Neck decoration is not common on Neutral vessels after the early fifteenth century. Undecorated collar vessels of the Ripley Plain type are common in Neutral assemblages, but less so in areas where the Huron ceramics predominate. Earlier in the Neutral area sequence though, neck decoration is common. It is a distinctive trait of the Pound Necked and Ontario Horizontal vessels found here and across southern Ontario, primarily in the late fourteenth century. A minor occurrence of these necked types, most notably the Pound Necked vessels, is found on sites across southern Ontario throughout the fifteenth century. In summary, the Neutral types used in this analysis are as follows: Undecorated Neck Neutral (also called Neutral)  Lawson Incised  Lawson Opposed High Collar Types High collar vessels (Figure 4-16 to Figure 4-19), that is, local tradition ceramics with straight collars over 3 cm in height, appear regularly, in small numbers, throughout Iroquoian Ontario, from the Neutral area in the west to the St. Lawrence valley in the east and Simcoe County in the north, during the mid- to late fifteenth century (Lennox and Fitzgerald 1990; William Fitzgerald personal communication 2014; Dana Poulton personal communication 2018) and as seen in this work. They are one of three minor groups included here. The Huron Lalonde High Collar type is highly characteristic of early fifteenth-century Simcoe County, although the type appears in low 77 numbers throughout the north shore and sporadically in outlying sites for the next two centuries (Nougerat 2012). Sopher High Collar vessels (Figure 4-15), a type first described by Noble (1968), are uncommon on the north shore. They generally appear in low proportions in late precontact Simcoe county collections (see Latta 1983:39). Lalonde High Collar vessels (Figure 4-16 and Figure 4-17) have rims measuring above 35 mm, often above 60 mm. They are characterized by straight to slightly convex interior and exterior profiles and complex triangular motif inside or above upper or lower collar horizontals, or both. The type name Lalonde High Collar is often applied to fifteenth-century high collar ceramics which do not conform to the relatively rigid Ridley definition (Dermarkar 2014; Ridley 1952b) but which are characterized by rims greater than 35 mm in height. Unlike some New York Iroquoian types, local Ontario high collar vessels have well-defined collar bases. High collar vessels are a characteristic of this period—in fact this rim form predominates in Iroquoian ceramic collections of fifteenth-century New York state (Engelbrecht 2003:85)—and, as in Ontario, rim heights decline in the sixteenth and seventeenth centuries. Figure 4-15. Sopher High Collar vessel, House 19, Keffer. 78 Figure 4-16. Lalonde High Collar vessel, Midden 57, Keffer. Figure 4-17. Lalonde High Collar learner vessel, Midden 65, Keffer. The Lalonde High Collar basal lines and vertical interruption of base lines on this vessel and the nubbin castellation show Lalonde High Collar production. 79 Figure 4-18. Lalonde High Collar type variant vessel, Midden 57, Keffer. Figure 4-19. Lalonde High Collar type variant vessel, Midden 57, Keffer. The high collared ceramics are as follows: High Collared  High Collared, no type name  Lalonde High Collar  Sopher High Collar 80 Erie–Niagara Frontier Undecorated Types Undecorated collared vessels appear in small proportions at Keffer and on other central north shore Iroquoian sites, but in greater proportions, up to 25 percent, at western sites, such as Lawson. They are the second minor group included here. The appearance of these vessels at Keffer is therefore significant and of particular interest. As mentioned above, it is the presence of short, parallel incised lines on vessel collars that characterizes both ancestral Neutral area and north shore Iroquoian pottery. The absence of this element represents a specific and intentional choice by the maker. The purpose behind this choice is unknown, but it should be included in any analysis as an indication of ceramic diversity and ceramic practice. Examples of the Erie-Niagara Frontier undecorated rim types are shown in Figure 4-20 to Figure 4-21. Figure 4-20. Niagara Collared vessel, House 12, Keffer. 81 Figure 4-21. Ripley Collared vessel, Midden 65, Keffer. Figure 4-22. Ripley Plain vessel, surface find, Keffer. These types can be summarized as follows: Erie-Niagara Frontier  Niagara Collared  Ripley Collared  Ripley Plain Early Ontario Iroquoian Types Though they are generally found in small numbers, the presence of vessels from the Early Ontario Iroquoian group aids in the chronological placement of Keffer features and of the site. 82 This pottery makes up the third minor group included in this analysis. The Middleport Oblique type is represented by only two fragmentary sherds (Figure 4-23). The early Neutral decorated neck Pound Necked type (Figure 4-24) is characterized by horizontal lines on the upper portion of the neck. Due to its infrequent occurrence in this collection, its early temporal assignment, and its limited presence at Keffer (49 vessels), the Pound Necked (Figure 4-24) sample has been incorporated into a category here labelled Early Ontario Iroquoian ceramic types. With only one vessel in the Keffer collection, Pound Blank (Figure 4-25) has also been subsumed into the Early category. Other Early Ontario Iroquoian types that are present in the Keffer assemblage are Ontario Horizontal (Figure 4-26) and Ontario Oblique (Figure 4-27). Single examples of these types are commonly found on fifteenth-century sites and appear in very small proportions at Keffer. Figure 4-23. Middleport Oblique vessel, House 6, Keffer. Figure 4-24. Pound Necked vessel, Midden 65, Keffer. 83 Figure 4-25. Pound Blank variant vessel, House 1, Keffer. Figure 4-26. Ontario Horizontal vessel, House 1, Keffer. Figure 4-27. Ontario Oblique vessel, Midden 52, Keffer. 84 In summary, the Early Ontario Iroquoian types are as follows: Early Ontario Iroquoian  Middleport Oblique  Ontario Horizontal  Ontario Oblique  Pound Blank  Pound Neck Nonlocal Tradition Types and Attributes As noted above, ceramic types originating in areas outside of the territories occupied by the southern Ontario Iroquoians of the north shore, Simcoe County, and Neutral areas are designated as nonlocal. These ceramics are subdivided into three main groups: Central Iroquois, Eastern Iroquois, and St. Lawrence –Northern New York Iroquoian. Ceramics of the Northern New York Iroquoians of Jefferson County have no specific types or attributes known to originate within their community of practice and generally share the majority of their types and attributes with ceramics of the St. Lawrence Iroquoians, particularly in reference to the Durfee Underlined type. New York state Iroquoian ceramics were divided into types by MacNeish according to geographical location within historic Haudenosaunee Iroquois confederacy territories. Based on limited data of the time MacNeish assigned the origin of each specific type in accordance with its greatest frequency at sites within each of these territories, regardless of the temporal placement of the site in the local sequence. Each type was thus assigned to an ethnic identity linked with a historic Iroquois Confederacy territory through the concept of the direct historical approach and named according to the geographic locale of discovery. Affiliation of many of these types with historic tribes or Nations has created confusion since MacNeish’s publication of Iroquois Pottery Types in the early 1950s. Vessels showing similar characteristics but appearing in separate regional developmental sequences bear different names (Engelbrecht 2004) as a result of MacNeish’s application of 85 individual labels to similar ceramics found in these separate geographic areas, which he associated with historic Iroquois confederacies. In addition, types assigned as characteristic of one tribal group may actually have originated with another group or have a shared origin. The separate village sequences of the Seneca, Cayuga, Onondaga, Oneida, and Mohawk developed from the Middle Iroquoian, formerly called the Owasco, tradition and the later Oak Hill horizon in upstate New York. This makes designation of a single origin for these types difficult and ambiguous. The earliest researchers considered Onondaga and Mohawk ceramics a unified body (Dyck 2004:13). Later researchers, namely, Dean Snow (1989), Donald Lenig (1965), and William Starna (1980), established a distinctive history for the Mohawk, demonstrating in situ development from the Oak Hill horizon into historic times (see also Chapdelaine 1990). More recently still, social network analysis (SNA) (Dermarkar et al. 2017; Hart and Engelbrecht 2012; Hart et al. 2016) shows the distinctiveness of Mohawk area vessels from those from other regions. Similarly, MacNeish designated ceramics from the Ivey, Lanorie, Putman, and Roebuck sites, along with collections from northern New York and along the shores of the St. Lawrence ethnically Onondaga (MacNeish 1952:56–57). The Durfee Underlined, Lanorie Corded, Lanorie Mixed, Lanorie Crossed, and Roebuck Low Collar types, placed in this “Onondaga” ethnic category by MacNeish, are currently considered to be of St. Lawrence Iroquoian origin (Clermont et al. 1983; Funk 2004:150). With these changes, only Onondaga Triangular and Syracuse Incised remain as types specifically affiliated with the historic Onondaga nation area. The recent application of novel theories has exposed the inherent limitations of assigning ethnic labels to these ceramics. Pratt (1980) merged several MacNeish types together, suggesting that minor attribute variations were not significant. These include the Wagoner-Syracuse Incised type, reminiscent of fifteenth-century Ontario high collars. This composite type was formerly divided into Syracuse Incised and Wagoner Incised on the basis of basal notching in Wagoner Incised vessels. Cayadutta-Otstungo Incised was also divided by MacNeish into Cayadutta Incised and Otstungo Incised due to basal notching on Cayadutta Incised pottery. In this study, these ceramics are considered to originate within the greater Eastern Iroquois tradition because the ceramic sequences share a common origin. In the Ontario and pan-Iroquoian analysis, these combined Pratt types are used in the database, since a significant proportion of the data available in the literature follows this typology. I have retained type names assigned by MacNeish in the 86 intra–Keffer village analysis because the attribute of basal notching is deemed significant at the micro scale. In the first decades of the fifteenth century, ceramics with design and/or form considered typical in terms of type and/or attributes of St. Lawrence and Eastern Iroquois, that is, the Onondaga, Oneida, and Mohawk of the ancestral Haudenosaunee (Engelbrecht 2004) area, began to trickle into the north shore area. The mechanism responsible for their introduction, whether trade or population movement, is now the subject of much dialogue (Abel 2016; Birch 2016; Dermarkar et al. 2016; Engelbrecht and Jamieson 2016; Gates-St. Pierre 2016; Gaudreau and Lesage 2016; Lesage et al. 2016; Ramsden 2016; Richard 2016; Steckley 2016; Warrick and Lesage 2016; Williamson 2016). The nonlocal ceramic types found within the Keffer village are divided into three groups: Central Iroquois, Eastern Iroquois, and St. Lawrence Iroquoian. The first two are further subdivided following the MacNeish subgroups, as noted below. The significant presence of many of these nonlocal ceramics in the Keffer collection necessitates an overview of these pottery traditions. Following convention, they are discussed here in geographical sequence. The Otstungo Notched and Rice Diagonal types are considered in some depth due to the surprisingly large presence of these types in the Keffer collection. Ceramics of the Central and Eastern Iroquois These ceramics share a tradition of uncollared, thickened, everted-lip vessels and collared vessels. The latter are derived from the horizontal motifs of earlier Point Peninsula/Early Owasco times (Lenig 1965:79). On later, Iroquoian collared vessels from the region, the horizontals are reduced in number. One to three horizontals ascend to the upper portion of the medium to high, upright rim. Horizontals do not appear on the base of Eastern Iroquois collar motifs. The presence of one or more upper horizontals characterizes several other types, all of which MacNeish assigned to one of the pre-Confederacy tribes, Onondaga, Oneida, or Mohawk. These include the Cayadutta Incised (Figure 4-28), Chance Incised (Figure 4-29), Fonda Incised (Figure 4-30 and Figure 4-31), and Otstungo Incised (Figure 4-31) types, designated as Mohawk, and Onondaga Triangular (Figure 4-32), long considered Onondaga in origin. Each of these ceramics has the distinctive upper framing over vertical or obtuse triangular motifs, with or without various forms of basal notching or upper rim verticals, or both. Related types with primarily horizontal motifs include the Mohawk Fonda Incised and Cayuga Horizontal (Figure 87 4-33) types, each with a horizontal motif interrupted by a series of parallel oblique–lined panels. Cayuga Horizontal, as well as the Oneida Thurston Horizontal (Figure 4-34), lack upper rim horizontals and are instead characterized by a series of mid-collar horizontals. Unlike the types with discontinuous horizontal designs, Thurston Horizontal is typified by a continuous and unbroken band of horizontals framed on the top and bottom with bands of short, vertical incisions. Thurston Horizontal and the Huron Warminster Horizontal (Figure 4-9) share similar motifs but are dissimilar in collar shape. Once again, the MacNeish definition was closely followed here. Warminster Horizontal, as a motif modification of Huron Incised, is characterized by a constricted neck with a short, outflaring collar and a convex interior rim profile (MacNeish 1952:35). The slightly constricted neck of Thurston Horizontal, on the other hand, is surmounted by a well-defined, incurving, interior-channelled collar of medium height, 25–55 mm (MacNeish 1952:68). The main similarity between these two types then lies in the mid-collar encircling horizontals. Figure 4-28. Cayadutta Incised vessel, Midden 70, Keffer. 88 Figure 4-29. Chance Incised vessel, Midden 57, Keffer. Figure 4-30. Fonda Incised vessel, Midden 65, Keffer. 89 Figure 4-31. Otstungo Incised vessel, House 12, Keffer. Figure 4-32. Onondaga Triangular vessel, Midden 85, Keffer. 90 Figure 4-33. Cayuga Horizontal vessel, Midden 61, Keffer. Figure 4-34. Thurston Horizontal vessel, Midden 65, Keffer. According to Lenig (1965), horizontal motifs are generally earlier than opposed motifs and are derived from the Point Peninsula and Owasco linear ceramics. The poorly delineated rim–neck junction of the Cayuga Horizontal (Figure 4-33 and Figure 4-35) type is mirrored in the somewhat later Cayuga ceramic type Richmond Incised (Figure 4-36). This type is defined by the presence of vertical or oblique incisions on a relatively high collar. Of similar design are the Onondaga Syracuse Incised (Figure 4-37) and basally notched Mohawk Wagoner Incised (Figure 4-38) types. Well-defined collar bases separate the latter two ceramics from Richmond Incised vessels. The high collars of Eastern Iroquois vessels are reminiscent of the Simcoe County Lalonde High Collar (Figure 4-16 and Figure 4-17) vessels, though their motif and collar shape 91 are not identical. It is recognized that there is a relationship between the height and motifs of the high collar of the Eastern Iroquois vessels and that of the Lalonde High Collar ceramic (Ridley 1952b). The nature of the relationship, however, remains unclear. Figure 4-35. Cayuga Horizontal vessel, House 4, Keffer. Figure 4-36. Richmond Incised vessel, House 11, Keffer. 92 Figure 4-37. Syracuse Incised vessel, Midden 57, Keffer. Figure 4-38. Wagoner Incised vessel, Midden 61, Keffer. The Mohawk types Otstungo Notched (Figure 4-39) and Rice Diagonal (Figure 4-40) and the Seneca types Dutch Hollow Notched (Figure 4-41) and Seneca Notched (Figure 4-42) are discussed in more detail in the section titled Distinctive Ceramic Attributes and Emergent Vessels at Keffer. 93 Figure 4-39. Otstungo Notched vessel, Midden 65, Keffer. Figure 4-40. Rice Diagonal vessel, Midden 62, Keffer. 94 Figure 4-41. Dutch Hollow Notched vessel, Midden 56, Keffer. Figure 4-42. Seneca Notched vessel, Midden 65, Keffer. To summarize, the Central and Eastern Iroquois types are as follows (sequenced geographically, as noted above): 95 Seneca   Dutch Hollow Notched Seneca Notched Cayuga   Cayuga Horizontal Richmond Incised Onondaga  Onondaga Triangular  Syracuse Incised Oneida  Thurston Horizontal Mohawk        Cayadutta Incised Chance Incised Fonda Incised Otstungo Incised Otstungo Notched Rice Diagonal Wagoner Incised Ceramic Types of the St. Lawrence and Northern New York Iroquoians Until the mid-1970s, Iroquoian sites on the south shore of the St. Lawrence valley in Quebec were considered remnants of early Onondaga tribal movements throughout the area (Dyck 2004; MacNeish 1952:57). It was not until Bernard Hoffman (1961), the first researcher to differentiate these sites, proposed a unique identity for the peoples of the river valley that St. Lawrence Iroquoians were acknowledged as a distinct regional or social group in the archaeological literature (Pendergast cited in Dyck 2004:15). As noted by Abel (2017), sites situated in the St. Lawrence valley were ignored as part of Wright’s (1966) greater Ontario Iroquois Tradition. Acceptance of St. Lawrence Iroquoians as a distinct archaeological construct finally occurred with the 1972 publication of Pendergast and Trigger. Then, in 1996, general acceptance of this redesignation occurred with the publication of a National Geographic article on the St. Lawrence 96 Iroquoians (Dyck 2004:16). Today, while it is acknowledged that groups of the St. Lawrence valley were independent of other Iroquoian groups, it is also recognized that great diversity existed within peoples of the extended St. Lawrence geographic area and that they cannot be seen as one cohesive social group or “nation” (Chapdelaine 1991). As a result of these developments, the ceramic types Durfee Underlined (Figure 4-43) and Roebuck Low Collar (Figure 4-44), which were defined according to vessels from the St. Lawrence Roebuck site, formerly considered Onondaga in origin, now are considered part of the St. Lawrence and Northern New York ceramic groups, as separate from the Eastern Iroquois traditions. Pendergast’s Stamped Low Collar (Figure 4-45) is now also considered part of the St. Lawrence group. Figure 4-43. SDEA motif (Smith 1997) outline on Durfee Underlined vessel, Draper. (Canadian Museum of History) 97 Figure 4-44. Roebuck Low Collar vessel, Midden 60, Keffer. Figure 4-45. Stamped Low Collar vessel, Midden 65, Keffer. The type Durfee Underlined, while partially based on form and collar height, is distinguishable from incised Eastern Iroquois types by its primary defining trait, the presence of one or two horizontal lines near both the base and the upper rim of the medium-height collar. All Eastern Iroquois types lack basal collar horizontals. MacNeish (1952:61) suggests that this ceramic may be related to Lanorie Corded, often found in eastern Quebec. The higher-collared Durfee Underlined pottery of the fifteenth century reflects the general trend for high collars in this period. General form and motif similarities shared by Durfee Underlined and the Lalonde High Collar type most popular in fifteenth-century Simcoe County were at the centre of debate during the mid-twentieth century (Emerson 1959a; Emerson and Popham 1952; Ridley 1952b). This issue eventually fell to the background of Iroquoian ceramic and interaction studies unresolved. 98 Further research in this area is outside the realm of this thesis but remains central to the understanding of pan-Iroquoian relationships of the fifteenth century. In some cases, Roebuck Low Collar (Figure 4-46) may share the distinctive lower collar line motif of Durfee Underlined (Figure 4-47), but, unlike in Durfee Underlined, this collar line is not “the” defining element of the type. The short, channeled collar can be decorated with one of any number of incised oblique designs (MacNeish 1952:61). St. Lawrence Low Collar is a label commonly used to incorporate pottery of this rim shape decorated with a myriad of motifs. I have incorporated a greater group of Low Collar vessels into the analysis but retained the Roebuck Low Collar type. Figure 4-46. Roebuck Low Collar vessel with SDEA motif, Midden 73, Keffer. 99 Figure 4-47. Durfee Underlined vessel with SDEA motif, Midden 60, Keffer. The St. Lawrence Iroquoian types are thus as follows: St. Lawrence Iroquoian  Durfee Underlined  Roebuck Low Collar  Stamped Low Collar Distinctive Ceramic Attributes and Emergent Vessels at Keffer In the New York Iroquois ceramic traditions, uncollared vessels with thickened lips comprise up to 20 percent of vessel assemblages (Bradley 1987:121). This vessel form does not appear in the southern Ontario traditions. Engelbrecht (2003:85) suggests that variation in pottery forms may relate to alternate food preparation functions, since preliminary analysis shows a low level of carbonized food remains on these vessels compared with other types. He postulates that these uncollared vessels may serve for tea brewing or storage. At Keffer, uncollared vessels are divided into two groups. The first group, composed of the Dutch Hollow Notched type (Figure 4-41), is present in low numbers across the site and conforms to the more crudely formed vessels typical of New York uncollared pottery. The second group, composed of the Otstungo Notched 100 (Figure 4-39) and Rice Diagonal (Figure 4-40) types, as seen at Keffer and other north shore fifteenth-century sites, appears to be quite different in several aspects from Keffer’s Dutch Hollow Notched pottery. These vessels are consistently well formed, with intricate rim and, sometimes, neck decoration. MacNeish identified uncollared and decorated thickened lip Rice Diagonal and Otstungo Notched vessels as Mohawk in origin and ethnicity, but their presence in significant proportions on earlier Onondaga and Oneida valley sites (Engelbrecht 2003:14) suggests they are part of a larger, more widespread tradition. There is no concrete evidence to suggest that the Otstungo Notched and Rice Diagonal vessels found at Keffer, Draper, and other Ontario sites are of foreign manufacture, or “exotics.” Vessels with form and decorative attributes similar to the Otstungo Notched vessels at Keffer do not appear to be present in other Northern Iroquoian collections (Timothy Abel, personal communication 2015; Bruce Jamieson, personal communication 2018; Wayne Lenig, personal communication May 2018) with the exception of identical vessels at Parsons and Draper and similar but smaller sherds at Jarrett-Lahmer and Damiani. Otstungo Notched vessels at Salem (Figure 4-48) and Maynard-McKeown (Figure 4-49) conform to the MacNeish definition of the type) but show variations in form from those found in the north shore area. Figure 4-48. Otstungo Notched vessel, Salem Cat. #697. (Canadian Museum of History) 101 Figure 4-49. Otstungo Notched vessel, Maynard-McKeown. (Canadian Museum of History) The horizontal line motif, lacking parallel obliques, common on New York Otstungo Notched vessels, is mirrored in one vessel found in Midden 61 at Keffer. Vessel form, however, suggests local production. Wayne Lenig (personal communication, 2018) acknowledges that Otstungo Notched vessels appearing at Keffer follow the type definition of both MacNeish (1952) and Lenig (1965) (see Figure 4-50 for an example). However, the Keffer vessels are more robust and show increased precision in decoration when compared with New York vessels. Otstungo Notched vessels found at the Rouge–Duffins Draper (Figure 4-51) and Parsons (Figure 4-52) sites, both dating to the fifteenth century, are identical to those seen at Keffer. Vessels at JarrettLahmer (ASI 1999:59) and Damiani (Figure 4-53) are quite similar. One of the later sixteenthcentury Mantle site vessels (Figure 4-54) matches those of western New York, suggesting an alternate origin pathway. It is apparent, given the high quality of production of the north shore renditions of the Otstungo Notched vessels that these ceramics are not intended to mimic or replicate examples found outside the north shore. These emergent vessels are a new creation, communicating a new identity. They are not hybrids that merely combine those elements, but material statements of a previously unseen identity. The north shore Lite site Rice Diagonal vessel (Figure 4-56) appears to be an intermediary between the simplistic motif of the St. Lawrence Otstungo Notched vessels and the more complex motif of those on the central north shore. The trend to larger, more rounded notches found on north shore Rice Diagonal and Otstungo Notched vessels, away from the tiny, grooved notches of the eastern vessels from the St. Lawrence valley and New York, also characterizes this Lite site vessel. One might even call this a north shore type, as it appears exclusively in this area. A number of Rice Diagonal vessels found on the St. Lawrence Salem (Figure 4-55) and Cleary (Figure 4-56) sites and the Prince 102 Edward county Lite (Figure 4-57) site share similar form and motif with the Keffer (Figure 4-40) and Draper (Figure 4-58) site Rice Diagonal vessels. The Rice Diagonal vessels found at Keffer and Draper are close copies of many found on St. Lawrence sites. I suggest that north shore ceramicists combined the precision of form and decorative techniques found on St. Lawrence Rice Diagonal vessels with the Otstungo Notched type motif, thereby creating a ceramic unique to their practice communities while retaining the original Rice Diagonal template. In addition, complex neck decoration was commonly, but not always, applied to these new vessels. Neck decorated ceramics are very common in fifteenth-century Rouge–Duffins drainage ceramic assemblages and seen in lesser, but varying, amounts across the contemporary north shore. In contrast, neck decoration is not common on Rice Diagonal or Otstungo Notched vessels of St. Lawrence or New York assemblages. This might suggest a Rouge–Duffins origin for these ceramics but, as network analysis shows, interaction among the fifteenth-century north shore communities was very active. Figure 4-50. Otstungo Notched vessel, unknown provenience, New York state. (courtesy Wayne Lenig) 103 Figure 4-51. Otstungo Notched vessel, Draper Cat. # 41454. (Canadian Museum of History) Figure 4-52. Otstungo Notched vessel, Parsons Cat. # 3357. (University of Toronto) 104 Figure 4-53. Otstungo Notched vessel, Damiani. (courtesy Ron Williamson) Figure 4-54. Otstungo Notched vessel, Mantle. (ASI 2006:Plate 37, Cat. #10432) 105 Figure 4-55. Rice Diagonal vessel, Salem Cat. #699. (Canadian Museum of History) Figure 4-56. Rice Diagonal vessel, Cleary Cat. #011. (courtesy David Smith) 106 Figure 4-57. Rice Diagonal vessel, Lite. (Canadian Museum of History) Figure 4-58. Rice Diagonal vessel, Draper Cat. #39180. (Canadian Museum of History) Another example of this modification of nonlocal ceramic traits is seen in the almost exclusive use of a distinctive complex motif designated SDEA (horizontal lines over vertical lines beside oblique right lines over horizontal lines), following Smith Version 4 ceramic attribute code (Smith 1997:Appendix C:122, 138) (Figure 4-43), on the majority of Durfee Underlined (Figure 4-47) and some Roebuck Low Collar (Figure 4-46) vessels at Keffer. The opposed motif without 107 horizontal lines also appears in small numbers on Ontario Iroquoian high collar and low collar (Figure 4-45) vessels at Keffer. The SDEA motif itself is not unique to north shore site nonlocal vessels; it is the ubiquity of this motif on locally manufactured ceramics of the fifteenth-century north shore that stands out. Although the identical motif is present at sites such as Draper (Figure 4-43 and Figure 4-59), Jarrett-Lahmer (Figure 4-60), Black Creek (Figure 4-61), and Parsons (Figure 4-62), preliminary visual analysis of these collections suggest it is more common at Keffer. On St. Lawrence and Northern New York Durfee Underlined vessels, this specific motif is less common by ratio than other motifs. When present, it usually displays a more complex and finely executed version of the motif (Figure 4-63) than is visible on vessels from Keffer and its neighbours. Durfee Underlined vessels with the SDEA motif are present at all sites on the central north shore where north shore Otstungo Notched ceramics are present. They are more numerous at Draper, Parsons, and, perhaps, Jarrett-Lahmer than Otstungo Notched and appear at sites such as Black Creek, where the later vessels have not been found. I suggest that this motif, like the Ontario Otstungo Notched vessels, is a local redesign of a nonlocal form that doesn’t appear elsewhere. Figure 4-59. Durfee Underlined vessel with SDEA motif, Draper Cat. #114617. (Canadian Museum of History) 108 Figure 4-60. Durfee Underlined vessel with SDEA motif, Jarrett-Lahmer Cat. #22897. (Canadian Museum of History) Figure 4-61. Roebuck Low Collar vessel with SDEA motif, Black Creek. (University of Toronto) 109 Figure 4-62. Durfee Underlined vessel with SDEA motif, Parsons Cat. #460 (University of Toronto) Figure 4-63. Northern New York Iroquoian vessel, St. Lawrence Site (Abel 2000:Plate 2). Exotic Attributes Seen at Keffer The nonlocal ceramic category also includes the few sherds of “foreign”, that is, vessels of nonlocal manufacture. These sherds are labelled “exotic.” Such designation occurs when specific attributes, that are not characteristic of the local tradition practice, are present in combination on vessels lacking local tradition motifs or decorative attributes. Basal notching is a nonlocal 110 attribute, commonly seen on Eastern Iroquois and St. Lawrence vessels. However, thumbnail notching, in which the imprint of the thumb nail is visible (Figure 4-64 and Figure 4-65), is considered a St. Lawrence attribute. While basal notching is apparent on a large number of nonlocal vessels at Keffer, basal thumbnail notching is much less common and appears only in concert with “exotic” reed punctates. The St. Lawrence Iroquoian ceramic tradition is characterized by specific decorative attributes rather than by specific MacNeish types. These include annular “reed” punctates, generally below sloping castellations; the corn ear motif; ladder plait incising (Chapdelaine 1990; LeMoine 2015) (believed by Wonderly [2004] to represent a stylized corn ear motif); and precisely formed and intricate motif patterns (Jamieson 1990:92; Ramsden 1977). At Keffer, these “exotic” attributes include “reed” punctates, thumbnail-gouged notches, and elaborate, finely executed complex design motifs (see Figure 4-64 for an example). These are found almost exclusively on foreignproduced pottery. Only one vessel of probable local production with reed punctates was found at Keffer (Figure 4-66). Figure 4-64. “Exotic” Durfee Underlined vessel, Midden 60, Keffer. 111 Figure 4-65. Vessels with poorly executed thumbnail gouges, Midden 65, Keffer. Figure 4-66. Localized Onondaga Triangular vessel, Midden 77, Keffer. The motif seen on this vessel, in Figure 4-66, oblique right diagonals over verticals with repeating pattern, commonly occurs on Durfee Underlined ceramics at Keffer and other north shore sites. This vessel, however, is Onondaga Triangular in type as it lacks basal horizontal lines and has multiple upper lines. Multiple upper horizontal lines are not common at Keffer or other north shore sites. This vessel, with its combination of reed punctates, a St. Lawrence characteristic, and Onondaga motif, fits into Pendergast’s hybrid category. The lack of intricate detail, the misalignment of the vertical punctates, and the simplified SDEA motif support the hypothesis that this is locally produced pottery. This vessel is one of very few examples of hybrid ceramics at Keffer. 112 Thumb-nail-gouged basal rim notches are also most apparent on exotic vessels (Figure 4-64). Like reed punctates, thumbnail-gouged notches appear very rarely on Keffer vessels. At Keffer, such gouges appear on one locally produced nonlocal vessel. Again, the precision of the foreignproduced ceramics is missing in the execution of the thumb nail gouges (Figure 4-65). Exotic attributes are very rarely applied to locally produced pottery in the Keffer assemblage. Instead, the local north shore potters of the late fifteenth century have redesigned nonlocal type vessels into north shore ceramics not seen elsewhere. These include the Rice Diagonal with neck design, the Otstungo Notched, and the Durfee Underlined vessels, all of which appear at Keffer and several other north shore sites. Nonlocal Learner Vessels and Practice Vessels at Keffer Learner vessels can provide a wealth of information regarding the dynamics of local communities of practice (Dorland 2018:11). The presence of learner vessels produced following the general concepts or attributes of specific ceramic types is a positive indication that vessels of this type were produced on site. It confirms the existence of associated communities of practice. The motor skills required for the creation of different vessel and rim forms are acquired during initial learning experiences, in the Iroquoian case most likely in childhood (Dorland 2018). While certain aspects of this production, generally dealing with decorative elements (Arnold 1991), can be more easily altered in later life, the motor skills employed in vessel formation are more difficult to relearn. Learner vessels with nonlocal traits, such as those illustrated below, display characteristics of the three most numerous nonlocal vessels found at Keffer, namely, Ostungo Notched/Rice Diagonal (Figure 4-67), Durfee Underlined/Roebuck Low Collar (Figure 4-68), and Wagoner Incised (Figure 4-69), confirming the local, on-site production of these types. The everted-lip vessel shown in Figure 4-67 displays elements of vessel form which distinguish the Otstungo Notched and Rice Diagonal types. This sherd, originating in Midden 60, provides important evidence in support of the hypothesis of on-site production of these nonlocal vessels. These vessels are characteristic of the nonlocal community of practice as exemplified by ceramics of Houses 12 and 18, on the village plateau. 113 Figure 4-67. Otstungo Notched/Rice Diagonal Everted Rim collarless learner vessel, Midden 60, Keffer. Figure 4-68. Durfee Underlined learner vessel, Midden 61, Keffer. 114 Figure 4-69. Wagoner Incised learner vessel, Midden 74, Keffer. (courtesy Steven Dorland) The Durfee Underlined learner vessel shown in Figure 4-68, from Midden 61, associated with House 14 provides evidence of the production of this St. Lawrence–linked type in Keffer ceramic practice. Interestingly, the opposed motif may be an attempt to conform to that of the SDEA code motif. The production of this vessel and that of the learner everted-lip collarless vessel shown in Figure 4-67 support the hypothesis that a new form of previously nonlocal ceramics was produced on the north shore and specifically at Keffer. More traditional nonlocal Iroquois vessels were also included in the learning process, as illustrated by the Wagoner Incised learner vessel in Figure 4-69. Chemical ablation of 15 Keffer local, nonlocal, and exotic vessels appears to confirm that local tradition, nonlocal tradition and emergent pottery was produced by practice communities at Keffer. Discussion Typology, following MacNeish (1952), forms the core of ceramic analysis in this study at all levels, from intra-village to pan-Iroquoian. This is supplemented with the examination of specific local and nonlocal tradition attributes. The pottery examined here falls into two basic categories 115 of local tradition and nonlocal tradition ceramics. The nonlocal group is further divided into vessels conforming to the Haudenosaunee tradition, vessels of foreign or “exotic” production and newly emergent pottery. In the next chapters analysis of the Keffer site ceramic material aims to seek out relationships in a multiscalar fashion and at many levels, from intra-site to inter-regional. To this end, the analysis operates at many levels, from the most specific to the most general, from micro scale to macro scale, from the household-level community to the pan-Iroquoian community. Types and attributes are both employed. Alone and in combination, they complement each other to form a more complete and richer picture of the inter-relationships of ceramics at Keffer and within greater Iroquoia. Combined types of all local and nonlocal ceramics, often used in regional ceramic analysis, are adopted for large-scale analysis. Analysis at the regional, pan-Ontario Iroquoian level local tradition types, as defined here, forms the database for network exploration of the chronological and social connections, in the context of ceramic practice, prevailing among the north shore of Lake Ontario communities. These analyses result in chronologies and social relations of the Keffer village settlement, the Ontario Iroquoian community and the panIroquoian community. On the face of it, use of the MacNeish/Ridley typology may appear contrary to these aims, because MacNeish’s typological framework is essentially one of ethnic classification based on deeply entrenched ideas of the perpetual, unchanging, and conservative nature of indigenous pre-contact cultures. A typological approach can unveil the nature of these relationships while avoiding the preconceived notions of ethnic divisions they original reflected. Instead, ceramic types are viewed as representations of individual and practice community choices, whether conscious or unintentional. Choices occur within the structure of perceived acceptability at the lowest level, that of the household. Such actions result from quotidian relationships and habitus. As such, types are not situated in this work as ethnic identifiers but as representations of contingent choices. The Brainerd-Robinson similarity measure (See Appendix E) in the R program (R, Core Team 2013) is used to produce data matrices which are then manipulated in the Ucinet social network analysis program. Visual analysis of the resulting graphs (see Appendix E), with statistical backup, are the basis of resulting interpretations regarding inter-group relations. At the regional (that is, Ontario) and Keffer village scales, the factors producing these patterns are explored in more detail through the analysis of the specific typological configurations of the samples. 116 Individual communities of ceramic practice are constructed as a result of this micro-scale exploration. Integration of resulting social and chronological relationships at the village and Ontario and pan-Iroquoian scales is employed to examine the position of the Keffer community in the fifteenth-century Iroquoian world. 117 5 Social Network Analysis: Keffer in the Iroquoian World Social Network Analysis of Pan-Iroquoian Ceramic Practice Situating the Ontario Iroquoian and Keffer ceramic community within the wider pan-Iroquoian cultural context is crucial to understanding the social relations and the divergent identities of the ceramic practice communities of the Keffer village inhabitants. Social network analysis has been demonstrated to be an effective mechanism for elucidating patterns with macro-scale ceramic datasets (Brughmans 2010; Knapp 2013; Mills 2013, 2017; Peeples 2011). Large-scale network analysis of pan-Iroquoian ceramic relations has been the focus of numerous studies during the past decade (Birch and Hart 2018; Dermarkar et al. 2016; Hart and Engelbrecht 2012; Hart et al. 2016, 2017). These studies have produced an overview of the relationship and development across Northern Iroquoia of vessel collar motifs, which, the authors suggest, are employed as social signals. The correlation between the possible social signalling function of vessel rim motif decoration and the communication of social relations involved in the production of ceramic types, cannot be equated with any certainty. In fact, collar motif and vessel type may be entirely unrelated. For example, specific collar motifs are not an essential element of the definition of several Iroquoian ceramic types, including Black Necked, Sidey Notched, Durfee Underlined, Roebuck Low Collar, and several of the ancestral Haudenosaunee ceramics with complex collar motifs. Consequently, the results of these previous analyses are not directly applicable to the present work. In this study, vessel types are employed in the exploration of communities of practice. Although social communication may be involved in some aspect of ceramic production, social signalling is not a focus of this research. Instead, I analyse a dataset of ceramic types from 252 Iroquoian sites (Appendix F) from southern Ontario and New York to investigate the strength and direction of the ceramic relations among Keffer’s Iroquoian neighbours, the origins of these interaction patterns, and Keffer’s place within this social network. (I present a full explanation of the analysis undertaken and the methods of network interpretation used in Appendix E.) Network analysis of pan-Iroquoian ceramic types (Figure 5-1, Figure 5-2) clearly shows a strong division between the ceramic practices of the New York state Iroquois and those of southern Ontario. The role of geography in the ties of the Ontario sites is less evident in this graph. 118 Figure 5-1. Pan-Iroquoian network BR 100 spring layout. Figure 5-2. Pan-Iroquoian network BR 140 spring layout. The network analysis portrays several distinct ceramic groups. Those of New York State adhere closely to their geographic or territorial boundaries. At first, it may appear that these groupings are a product of the historic tribal affiliation of types used by MacNeish. However, the original ascription by MacNeish of the Northern New York and some St. Lawrence sites, as Onondaga argues that this is not the case. The Northern New York sites form distinct territorial and topological clusters yet are characterized by ceramics of other territories. In general, New York 119 state Iroquoians shared ceramic practice is restricted to geographical locations within the historic tribal homelands to various extents. The division between Eastern and Central Iroquois groups is highly evident, with the first ties appearing between these groups at low similarity levels only in the later, t5 and t6, periods, suggesting closer ties occurred with the confederation of the Five Nations Iroquois. Much higher levels of ceramic interaction are present throughout southern Ontario in all periods, with the highest uniformity in the t1 and t6 periods. The t6 consolidation reflects the creation of the Wendat, Petun, and Neutral confederacies in the contact period. T2 to t5 ceramic practice across the north shore is generally well integrated across both time and space and remains so up to BR 140, or 70 percent type similarity (Figure 5-2). Network analysis shows the ceramic practice communities of Ontario and New York to have been quite distinct throughout the period A.D. 1350–1650. Ontario groups form a compact network cluster. In New York, more dispersed spatial clustering occurs within the Five Nations traditional territories. Nodes from within each of the traditional territories of the Eastern Iroquois and the Northern New York Iroquois cluster together, with each cluster lightly tied to one of the others. Nodes representing the Cayuga and Seneca ceramic collections of the Central Iroquois are weekly tied together and also weakly connect sixteenth-century Onondaga and Oneida sites through contemporary sites to the Ontario component. These Cayuga and Seneca sites connect with sixteenth-century site nodes of the group often designated as Erie (Dermarkar et al. 2016; Engelbrecht 1991, 2004). This group, in turns, bonds tightly with the Neutral sites of the middle temporal range of the Ontario cluster and most likely represents movement of Neutral populations south and east from their traditional homeland, as Lennox and Fitzgerald (1990) and Engelbrecht (1991) note. The late timing of this connection precludes this path being the conduit for the arrival of eastern-influenced ceramics into Ontario practice in the mid-fifteenth century. The shared Owasco heritage of the ancestral Central and Eastern Iroquois and the lack of ties with Ontario in the earlier periods imply an ancestral tradition quite different from that of the fourteenth-century Ontario Iroquois. In contrast, the unified early Ontario ceramic tradition of the fourteenth century may have exerted a strong influence on, or been the origin of, the Ontario St. Lawrence and Northern New York ceramic practices. This is indicated by the presence of a weakly tied topological bridge comprised of the earliest St. Lawrence sites connecting the fourteenth-century Ontario and early fifteenth-century Northern New York Iroquois sites. This is the last evidence of shared ceramic 120 practice between St. Lawrence Iroquoians and north shore Ontario peoples. The close ties shared by St. Lawrence Iroquoians with Northern New York Iroquois potters during the fifteenth century are replaced by strong bonds between the St. Lawrence and later sixteenth-century Oneida with the abandonment of Jefferson County by the Northern New York Iroquois in the early sixteenth century. At this time, the St. Lawrence Iroquoians abandoned the St. Lawrence valley, a movement completed by the end of the century. Like the Iroquoians of south-central Ontario, the peoples of New York retreated from the open shorelines in the sixteenth century. Connections between the New York and Ontario ceramic communities occur, but only at very low levels, and only in the earliest and latest periods of the Iroquoian occupation of Ontario. Aside from one weak mid-fifteenth-century tie between Northern New York and Prince Edward County sites, there is no evidence of any strong ceramic relationships between the two fifteenthto early sixteenth-century populations. While groups in New York restricted ceramic practice within defined territories, fifteenth-century ceramicists in southern Ontario pursued more integrative activities. Here, there is a general trend in the change of ceramic practice, both over time and across geographical distance. The turn of the fifteenth century witnesses the greatest diversity of practice, with nodes of this period in Ontario found throughout most of the component body. This is the era of the introduction of ancestral Wendat (“Huron”) and Attiwandaron (“Neutral”) ceramics across the province, a time of ceramic transformation. Figure 5-3. Pan-Iroquoian network t2–t3 BR 93 spring layout. 121 In summary, network topology indicates a high level of interaction throughout Iroquoian Ontario, which generally stopped at the barrier created by Lake Ontario and the Niagara and St. Lawrence rivers. Interaction of a slightly different character occurred south of this boundary. These territorially oriented groups shared somewhat less similar ceramic practices within local communities and more varied practices between regions. The ancestral Wendat and Neutral appear to have been more socially integrated peoples than the populations of New York. The low level at which the New York and Ontario ceramic samples connect portrays somewhat sporadic contact between the two groups, with no indication in the network topology of large-scale movement or contact between contemporary ceramic practice groups of these two Iroquoian peoples. The Keffer site sits within the core of the consolidated t2–t3 and t3–t4 Ontario components, fully integrated into the large-scale ceramic community of practice of southern Ontario. This cluster is well separated from those of New York state and the St. Lawrence Iroquoians and gives no indication of the ties evidenced by the presence of nonlocal and exotic ceramics at Keffer and other fifteenth-century north shore sites. Social Network Analysis of Local Tradition Ceramics from 140 Ontario Iroquoian Sites Determining ceramic affiliations of the Keffer communities of practice on a clinal scale of social distance within the highly connected network of the greater Ontario Iroquois population is a major focus of this study. The history and relationships of the Keffer ceramic communities are explored and uncovered through examination of these connections. Prior to this micro-scale analysis, the Keffer site itself must be situated in the greater Ontario Iroquoian ceramic network. These relationships are explored here through a meso-scale use of social network analysis. The Keffer local tradition ceramic collection is compared with other Iroquoian collections from across the province. I examine local tradition ceramics from 140 sites (Appendix G) spanning the period A.D. 1350–1650, previously known as the Late Ontario Iroquoian period. Local tradition ceramics are exclusively employed in this analysis to deduce underlying long-term shared ceramic traditions. Nonlocal tradition ceramics, as seen in relatively high proportions at such sites as Damiani, Draper, Mantle (now renamed Jean-Baptiste Lainé), Parsons, and Seed-Barker, are not included in this analysis. It is argued in this study that these nonlocal ceramics are 122 reflective of external ties, not historical practice. I suspect that inclusion of this pottery may obscure strong long-term relationships of local practice tradition. Ontario Iroquoian Ceramic Relations as Seen Through Network Analysis The Keffer village is the last-known manifestation of Iroquoian occupation in the Don River valley and is centrally located geographically within the Ontario Iroquoian homeland. Although archaeologists have argued that communities moved further north with time, towards the headwaters of the rivers draining into Lake Ontario, many earlier t2 and t3 Don River sites are located to the north, upslope from the Keffer location. Evidence for occupation of the Don River valley is heavily skewed to the t2 (A.D. 1400–1450) and t3 (A.D. 1450–1500) periods (Williamson 2014). Only one site, Moatfield, provides documentation of earlier Iroquoian occupation, and sites later than Keffer have yet to be discovered (Williamson 2014:21). Discussion of ceramic relations within Ontario therefore centres most heavily on this 100-year period of the fifteenth century. Ontario ceramics in the network analysis (Figure 5-4) divide into two main components. This division is primarily temporal, as early Late Ontario Iroquoian ceramics from across the study area form a distinctive assemblage almost entirely disengaged from later sites at higher similarity levels, above 70 percent, in the late fourteenth century. 123 Figure 5-4. Local tradition ceramics of Ontario Iroquoian sites BR 99 spring layout. There is, however, evidence of a strong continuation in ceramic practices throughout the early fifteenth century exhibited across all studied geographic groups. The lack of evidence for early sites in the archaeological record of the Don and Humber drainages, while unfortunate, does not invalidate the hypothesized continuation of early ceramic practices in this area. The maintenance of local ceramic practice is extrapolated from the prevalence of this pattern as seen across Ontario. Early fifteenth-century sites of all areas display ties with fourteenth-century sites at BR 130 (65 percent similarity). Across the north shore of Lake Ontario, shared ceramic traditions are continuous in most areas. Although the Humber t2 Black Creek site ceramics are integrated with those of the larger ceramic culture, evidence suggests that sites of the Humber valley do not form a single, unified ceramic community. The early Downsview ceramics are most closely aligned with London-area Neutral ceramics. Parsons and Bosomworth also share close ceramic similarities with Neutral sites near the western end of Lake Ontario and with late fifteenthcentury PEC ceramics. The late fifteenth-century Humber River drainage Damiani and Logan sites share close ceramic ties with the fifteenth-century Don River drainage Keffer and JarrettLahmer sites. To the west, Neutral sites began to disengage from the earlier ceramic conformity by the early fifteenth century, similar to a distinctive group of sites in Simcoe County. This topological separation of Simcoe County sites may reflect Sutton’s suggestion of two separate movements 124 into the area from the south (Sutton 1995, 1996, 1999). Another relatively unified ceramic community of separate, but highly interacting, groups existed throughout this century from the Don valley eastward to the Trent valley and into Prince Edward County. Slightly later, A.D. 1450–1550, the Trent valley sites, Kirche and Coulter, were also present in this formation, clearly separated from the sizeable later sixteenth- and seventeenth-century aggregate of sites representing the final Iroquoian occupation of Ontario, centred in northern Simcoe County. Terminal sites of drainage-centred communities eastward from the Humber valley shared a higher level of ceramic similarity with the late Simcoe cluster than with slightly earlier sites in their area of origin. The Humber Damiani and Logan sites link the contemporary late fifteenthcentury eastern sites, Keffer and Jarrett-Lahmer, to the Simcoe core. The Rouge–Duffins Mantle and Aurora sites sit within this core. The later Humber valley McKenzie and Seed-Barker sites bind the seventeenth-century Neutral Bradt and earlier Neutral and Humber sites to those of the highly integrated Simcoe cluster. Figure 5-5. Local tradition ceramics of Ontario Iroquoian sites BR 140 main component MDS layout. Network analysis indicates that the interrelated fifteenth-century sites of the Trent, Rouge– Duffins, and Don drainages formed the central core of the Ontario Iroquoian local ceramic 125 tradition, binding fourteenth-century ceramic practices with sixteenth- and early seventeenthcentury assemblages. The central location of the Keffer and other Don valley sites in this bridge structure highlights Keffer’s integral role in the transition of ceramic practices from the fifteenthcentury north shore to the later home of the Huron-Wendat in Huronia, in Simcoe County. Fifteenth-century Humber and Neutral, and some Simcoe County ceramic practices are to some extent peripheral to these developments. It is clear that the north shore of Lake Ontario was an area of widespread ceramic interaction among local tradition ceramic communities during the century leading up to the dispersal of the region’s occupants. This interaction pattern continued from an earlier period of province-wide ceramic unity so well displayed in the t1 cluster of sites. Thus, network analysis places the Keffer practice community as a central player, both chronologically and geographically, in the ceramic relations of north shore Iroquoian movement, from its origins on Lake Ontario’s north shore to the final settlement in Huronia. Multiscalar Relations of the Southern Ontario Iroquoian and Keffer Site Ceramics Pan-Iroquoian network analysis confirms distinct ancestral ceramic traditions for the Ontario Iroquoian peoples and most New York Iroquoian peoples, although Seneca and Cayuga sites in the earliest period shared closer ties with early Ontario sites than with those of the Eastern Iroquoians (Figure 5.1). The division between these two main groups persisted throughout time despite sporadic ceramic connections at the eastern and western ends of Lake Ontario. The close ties among Ontario, St. Lawrence, and Northern New York sites in the fifteenth century reflect interaction among these groups at that time. In general, New York–based Iroquoians were much more restricted in their ceramic interaction by geography than were Ontario-based groups. Close ties among later (t4–t6, A.D. 1500–1650) Neutral and Erie site nodes support the conjectured movement of the Ontario Neutral southward and eastward. This explains the near disappearance and disengagement of the Neutral nodes from the Ontario cluster at this time. Other than this movement, there is no strong evidence in the networks of large-scale movement of ceramic practices among the New York state groups or between these groups and the Iroquoians of the Ontario cluster who consolidated in the Simcoe area, called Huronia, towards the end of the sixteenth century. The tightly bound cluster of Ontario Iroquoians seen in the networks portrays a much higher level of ceramic interaction across the southern part of the province than is seen to the in New York state to the south. 126 Within the Ontario Iroquoian ceramic history, both network graphs and typology illustrate a highly interactive network of ceramic tradition, which is summarized here following a temporal framework. The main periods are described from west to east and then north into the Simcoe County area. Early communities are widespread across southern Ontario in all areas surrounding the Don and Humber drainages, which lack sites from this time. These communities shared in the widespread Early Ontario Iroquoian ceramic heritage of the fourteenth century. Two somewhat distinct practices developed from this heritage. These include ceramics of the Neutral tradition, with only undecorated neck forms, and of the “Huron” tradition, with both decorated and undecorated neck forms. Neutral ceramics are generally more common in the area west of Lake Ontario, while “Huron” ceramics dominate the area east of this along the north shore of Lake Ontario and into the Simcoe County area. In the early fifteenth century, “Huron” ceramics, in the form of undecorated neck vessels, began to appear in sites to the north of the Milton Neutral cluster. This pattern continued in the Credit valley, where ties to the east now became stronger than those to the west. These patterns link this area to sites on the other side of Toronto, as the Don and Humber valleys lack sites from this period. Contemporary early sites in the Highland Creek and Rouge–Duffins drainages seem to share a single, or affiliated, ceramic community of practice. These collections are composed of 50 to 85 percent of Early Ontario Iroquoian types, with almost no decorated neck “Huron” ceramics in the earliest sites and a very limited presence, of up to 6 percent, of undecorated neck “Huron” ceramics. Towards the turn of the fifteenth century, decorated neck vessels began to have a greater presence, with 6 to 30 percent at Rouge–Duffins sites. Neutral ceramics appeared at moderate levels, 3 to 9 percent, in these early sites and retained this low presence throughout the community history. After A.D. 1400, decorated neck “Huron” vessels saw a great surge in popularity, rising to 80 percent at Joseph Picard, while undecorated neck vessels remained below 35 percent. In the post–A.D. 1450 period, undecorated neck vessels rise in favour and match decorated neck vessels in popularity. By the end of the occupation of this area in the mid-1600s, decorated neck pottery had almost disappeared, diminishing to just 2 percent at Mantle. Undecorated neck “Huron” ceramics continued to represent less than 70 percent of the assemblages, as locally manufactured nonlocal tradition ceramics gained prominence in the collections. 127 During the late fourteenth century, in the more easterly Trent River watershed, the Early Ontario Iroquoian ceramics that were so prevalent elsewhere at this time occurred in smaller frequencies, between 40 and 65 percent. Decorated neck pottery occurred in larger proportions than before, now composing up to 20 percent of the Trent valley ceramic sample. Likewise, undecorated neck ceramics, representing less than 15 percent, were also more numerous at this time in the Trent valley than they are in the Rouge–Duffins/Highland Creek area. After A.D. 1400, the Rouge– Duffins and Trent drainages shared equal proportions of undecorated neck ceramics, up to 36 percent, while decorated neck proportions skyrocketed to dominance, making up 40 to 85 percent of the total ceramics. As the fifteenth century ended, decorated neck pottery remained slightly more popular in the Rouge valley, at 40 percent, than in the Trent valley, where a lower level, of 30 percent, characterizes contemporary collections. After A.D. 1500, however, decorated neck ceramics disappeared almost completely in the Rouge–Duffins region. They retained a substantial minority position, upwards of 30 percent, in the Trent valley, alongside a small but increasing majority of undecorated neck ceramics. The small presence of Neutral pottery found earlier in the Trent valley (less than 10 percent) slowly shrank over time. In the post–A.D. 1500 period, no evidence of these ceramics remained. The absence of these ceramics in the Trent valley collections raises questions concerning their significance and the causes of their absence in this area. The rise of popularity in the Neutral ceramics in the Rouge–Duffins drainage at the Mantle site in this later period, in conjunction with the almost complete disappearance of neck decorated ceramic here, portrays a division in the ceramic pathways of the two watersheds that had earlier shared close similarity. Similarly, some Simcoe County sites contain up to 10 percent Neutral ceramics, while many show no evidence of this pottery. With the exception of those sites in the Prince Edward County cluster of Arbor Ridge, Hillier, and Lite, all sites of the early fifteenth century geographically surrounding the Don–Humber region consistently contain appreciable percentages of Neutral ceramics. Decorated neck vessels (in this case, Black Necked) never achieved the level of popularity in Simcoe County that they experienced in the Trent and Rouge–Duffins areas. Proportions fluctuate across Simcoe County sites of the period. On the other hand, undecorated neck “Huron” ceramics display more consistency and overall popularity in Simcoe County. After A.D. 1450, undecorated neck “Huron” ceramic proportions increased steadily throughout Simcoe County, until they constituted almost 100 percent of the collections at several early sixteenth-century 128 village sites. Decorated neck ceramics are almost non-existent in most villages by this time. While a general decrease in decorated neck ceramics and a respective increase in undecorated neck ceramics is true for the period A.D. 1400–1600 in Simcoe County, the pattern is, in fact, more highly variable and sporadic than patterns seen in other drainage communities, due to several waves of peoples from more southerly areas, along the Lake Ontario north shore. The chronological separation of the network as portrayed in Figure 5-5, with earlier sites to the left and later sites to the right of the graph, reflects the shift from the neck decorated pottery of earlier, Middleport-based ceramics to the rising popularity of DNH ceramics present in the t1–t2 bridge between the two sides of the component. The introduction in t2 and subsequent increased in popularity in t3–t6 of undecorated neck “Huron” pottery produces the tighter formation of the later contact period. The spread of sites from the bottom of the graphs shows the range of the influence of Neutral ceramics, with large proportions in the Neutral-area sites and almost none in the t3 Rouge–Duffins, Trent, and some Simcoe sites. The increasing dominance of undecorated neck “Huron” pottery is manifest in the late Simcoe-area sites at the right of the networks. The continuing popularity of decorated neck ceramics into the sixteenth century the Trent valley is demonstrated by the position of the Kirche and Coulter site nodes, which are integrated with earlier t2 and t3 Rouge–Duffins and other north shore sites. The absence of ties between Trent valley sites and the t5 Rouge–Duffins Mantle site, which possesses almost no neck decorated vessels, exemplifies a disconnection in shared traditions. The Don valley group of sites, with the centrally located Keffer site, sits solidly in the centre of the Ontario local tradition post-t1 core. This centrality illustrates the central role of the Don valley occupations as a conduit of ceramic tradition though both time and space. Collections of the numerous t2 and t3 sites in this centrally located watershed are characterized by their diversity. Neutral, decorated neck, and undecorated neck “Huron” vessels appear in varying proportions at most sites. This diversity illustrates the complex ceramic interaction occurring in this central area throughout the fifteenth century. Network analysis indicates that the interrelated fifteenth-century sites of the Trent, Rouge– Duffins, and Don valleys form the central core of the Ontario Iroquoian local ceramic tradition, binding fourteenth-century ceramic practices with those of the sixteenth and early seventeenth centuries. The central location of the Keffer and other Don valley sites in this bridge structure 129 highlights Keffer’s integral role in the transition of ceramic practices from the fifteenth-century north shore to the later home of the Huron-Wendat in Huronia (Simcoe County). Fifteenthcentury Humber, Neutral, and some Simcoe County ceramic practices are somewhat peripheral to these developments (Figure 5-3). It is clear that the north shore of Lake Ontario was an area of widespread ceramic interaction among local tradition ceramic communities during the century up to the dispersal of groups from the region. This followed an earlier period of province-wide ceramic unity so well displayed in the t1 cluster of sites (Figure 5-3). Network analysis thus places the Keffer community of practice as a central player, both chronologically and geographically, in the ceramic relations of Huron-Wendat, from their origins on Lake Ontario’s north shore to their final settlement in Huronia. Although some interaction between St. Lawrence and New York sites and those of the north shore is apparent in the various network graphs, the disconnection between the ceramic heritages of the two areas is well visualized, and the independent and highly intra-connected nature of the Ontario ceramic tradition is plainly evident. The Don–Humber ceramic communities, and the Keffer site itself, are located at the centre of this Ontario Iroquoian ceramic tradition development and show no close ties to ceramic communities to the south. An in-depth, multiscalar investigation of the internal and external relations of the Keffer ceramics will help to locate Keffer more precisely in the larger network and provide insights into ceramically mediated social relations and interactions of the village inhabitants. 130 6 Settlement Pattern Analysis In this chapter I follow the research of Dodd (1984) and Warrick (1984) and analyze the settlement features of the Keffer site, houses, middens, and palisades into four spatial groups, or clusters. Settlement clusters likely reflect some aspect of social relations, generally kinship-based formations. Consideration of the inter-relationships of the archaeological features within each cluster and between clusters is aimed at providing potential chronological links within clusters and across the village as a whole. This chronological framework, grounded in settlement data, will provide a basis for a detailed temporal and social interpretation of the network analysis of the Keffer village ceramics in chapter 7. Spatial Arrangement of Major Settlement Features This settlement pattern analysis examines three types of features, namely, longhouses, palisades, and middens. The Keffer village site is composed of 18 longhouses, ranging in length from 15 to 65 m, and one small, circular structure of unknown function. The latter building was originally labelled House 5 in the field and has retained that name is subsequent work. All structures thought to be houses were numbered in the field during the course of excavation. One of these, “House 17,” was deemed not to be a house in post-excavation analysis; therefore, house numbering is not contiguous. Several of these houses were extended after their initial construction. These house extensions were given letter suffixes (Table 6-1). Post mould densities are highest in the walls of the longest houses (with one exception), indicating long-term occupation (Table 6-1, Figure 6-1). The exception is House 15, which, despite being the longest house, at 65 m, has the lowest post mould count, indicating that it was likely short-lived. Table 6-1. House and Extension Length(s) and Wall Post Density. House/Extension Length (m) Addition Length (m) # Wall Posts per m 1a 37.5 1b 45 2 35 4.1 3a 22 4.7 3b 30.5 10.7 7.5 8.5 131 4 42.5 4.8 6 35 4.3 7a 28 13.4 7b 45 8 40 7 9a 27.5 7.9 9b 40 10a 22 10b 29 11a 40 11b 55 15 11c 62 7 12 45 6.4 13a 32 5.4 13b 40 7.5 14a 28 7 (north) 14b 40.5 7.5 (south) 15a 37.5 15b 65 16a 32.5 16b 40 18a 24 18b 36.5 5 (north), 7.5 (south) 18c 44 7.5 19 25 7.7 20 40 6.3 17 4.9 6 (north), 7.5 (south) 4.3 7 10.5 4.5 3.8 27.5 7.7 7.5 6 132 Figure 6-1. Map showing all post moulds. 133 Twenty-seven middens, or centralized refuse deposits, were identified during excavation and have been included in this site synthesis. Some midden features, such as Midden 76, lack clear physical contexts. Some, such as Midden 58, possess samples of unverified origin. Others, including Midden 64, lack analysable ceramics, while Middens 69 and 70, were re-classified as other types of features during post-excavation analysis. Therefore, not all 27 “middens” are discussed here, and the numbering of the middens is not contiguous. At first glance, the settlement plan of the Keffer village suggests a slightly dispersed pattern of three longhouse clusters, each spatially disconnected from the four longhouses scattered around the eastern and southern perimeters of the site. These houses are situated within one or both of two well-defined palisades, one double-row palisade exterior to the village (labelled Palisade 2), and a second, much smaller, single-row palisade entirely within the limits of the first. A 50 m fence line (labelled Palisade 3) partially surrounds House 10 at the base of the plateau slope. Several surface middens are scattered throughout the site both inside and outside Palisades 1 and 2, but the largest middens appear on the riverside terrace slope and within a slight depression at the southeastern edge of the site. Again at first glance, this suggests a transparent pattern of growth, from a small, clustered, 11 longhouse occupation encircled by a single-row palisade, to larger village of 15 houses, encircled by a double-row palisade. Topography The superposition of the topographic features of the greater site area over the excavated settlement layout (Figure 6-3, Figure 6-2) provides a supplementary and essential explanatory context for the comprehension of the village layout. The geophysical aspects of hillside slopes, most relevantly the 3 m rise in elevation from the basal elevation of the “low town” to the crest of the upper plateau of the “high town,” at the western limit of the site, can account for the specific structural characteristics of the village features. The topographic analysis reveals that the double-row-palisaded village of 15 houses consisted of two coexisting groups of houses separated by the differing elevations. The 25 m separation of the high town longhouses from the rest of the residential structures is explained by the steep slope (Figure 6-4) encircling the plateau. What appear to be interior surface middens throughout this 25 m band are now revealed as hillside middens originating from households throughout the high town. The materials from 134 these middens most likely derive from the high town and are not by-products of deposition from nearby houses at the lower elevation. The middens thus take two general forms: surface middens, spread in shallow deposits across the interior village living space, and hillside middens, consisting of debris deposited away from the living space, down the slopes between the two elevations or on the village outskirts. Most large middens at Keffer, like those at other Iroquoian villages of this period (such as Baumann and Jarrett-Lahmer, see DRPA 1996a; Stopp 1985), are hillside middens, as is evident from their location in relation to the contour intervals (Figure 6-3). 135 Figure 6-2. Map showing all post moulds and the 1 m contour intervals. 136 Figure 6-3. Digitized sketch plan showing houses, middens, and palisades and the 1 m contour intervals. 137 Figure 6-4. View of the high town and low town longhouses at Keffer’s north end during excavation. Overlapping and Sharing of Features An analysis of overlapping and shared features can help sort out the settlement sequence. As the Keffer settlement plan attests, throughout all stages of the village occupation, wall post moulds of some structures are either shared with, or are directly in contact with, those of other structures. Many structures share this characteristic of physical contact. For example, the northwest corner and the east wall of House 18a, the earliest manifestation of this residence, are in direct physical contact with Palisade 1 and House 15a, respectively. In a similar manner, House 1a is located in direct contact with Palisade 1, and House 13a may share posts with the southern limit of the palisade. The earliest edition of House 11a is also in physical contact with this palisade. There are also indications that Houses 7a and 15b may also connect with the palisade. Later structures, Houses 4 and 19, show evidence of direct contact with the second palisade construction. The positioning of these structures in such a manner that each of these houses aligns exactly with the nearby palisade—and such that there is no evidence of structural contact between House 4 and 138 Palisade 2 or House 19 and Palisade 2b—strongly suggest the contemporaneity of the structures. Instances of non-contemporary overlapping of features are clearly apparent in other places on the site plan. Despite the common occurrence of midden–palisade interfaces, the presence and location of these middens adjacent to the palisades does not greatly inform settlement sequences. Interior middens, in contrast, often affect village mobility patterns, impeding access throughout the settlement, and therefore can be informative regarding settlement sequences. The closely spaced and overlapping nature of the different house clusters prevents refuse accumulation in these areas. Apparently, the initial structural proximity within these clusters already restricted mobility between houses. This is illustrated in the placement of several houses at Keffer (namely, Houses 1, 4, 10, and 11) in direct contact with palisades at one of the structural corners. It is only around the peripheries of these clusters, skirting the interior village limits or hilltops, that ground surface middens appear, in locations where they do not seriously impede traffic flow. Definition of Site Clusters Visual analysis of the spatial distribution of the longhouses in conjunction with the site topography indicates that the main occupation features of the village are spatially separated into four distinct sectors, or clusters (Figure 6-5). The placement of houses in clusters assumes that they were constructed together as a planned unit (Warrick 1984). Each of the middens in the village is correlated with one or more houses, according to its proximity to the structure, the topography of the landscape surrounding the midden, the ease of intra-site mobility between the house and the midden, and the overall size of the midden. With this in mind, the houses and their correlated middens are ascribed to one of the four “clusters,” Clusters A, B, C, and D (Table 6-2). The clusters are colour coded for ease of viewing (dark orange = Cluster A; light orange = Cluster B; light green = Cluster C; dark green = Cluster D). 139 Figure 6-5. Plan showing the houses (solid fill), middens (dotted fill), and palisades (dotted lines) associated with each of the four site clusters, and the 1 m contour interval. 140 Cluster A is comprised of those houses (Houses 1, 2, 3, 5, 6, and 7) and their affiliated middens (most notably the two large hillside middens, Middens 57 and 60), located in the northernmost area of the lower level of the site. These features are closely positioned and share a generally parallel east–west orientation. While the western edges of the structures appear to be aligned facing Palisade 1, their eastern ends open onto a 3 m2 area that is largely devoid of refuse and of exterior features and posts. “House 5,” a round structure 8 m in diameter with no evidence of a hearth or other occupation features, is not a true longhouse but a construction of unknown function. It is included in Cluster A, along with three surface middens (Middens 51, 63, and 64), and the two large, adjoined hillside middens (Middens 57 and 60). Midden 64 is a small surface deposit located directly south of Midden 52. It has no analyzable ceramics and no clear relationship with any other archaeological features. Therefore, it is not included in this analysis. Separated from Cluster A by orientation, rather than distance, are the three houses of Cluster B: Houses 8, 9, and 10. While House 8 shares an east–west alignment with House 7 in Cluster A, its radial orientation with Houses 9 and 10 around an open plaza area suggests its inclusion in the spatially distinct Cluster B. Cluster B is therefore deemed to consist of three houses, Houses 8, 9, and 10, oriented at a 40-degree angle, southeast–northwest. The houses of this semi-circular cluster focus on a small open area to the west. Middens associated with these houses are Midden 55 (located in the western plaza area, where the three house ends meet), Midden 56 (to the south of House 9), Midden 58 (located overtop of House 10), and Midden 67 (located just to the south of the fence line surrounding House 10). Cluster C is composed of a concentrated grouping of six houses oriented north–south, perpendicular to Houses 9 and 10 of Cluster B. This group is composed of all houses positioned on the hilltop plateau, in the so-called high town area (Houses 12, 13, 14, 15, 16, and 18). These structures generally articulate radially towards the centre of the plateau. Several middens along the surrounding hill slope are associated with the heavy occupation of this hilltop area. These include the surface deposit Midden 59 (just outside Houses 14 and 16), Midden 61 (to the north of House 14), Midden 62 (a hillside midden associated by proximity with House 18), Midden 66 (located on the western slope of the plateau), Midden 72 (on a hill slope to the south of Houses 12 and 13), Midden 73 (on the plateau next to the south end of House 13), Midden 74 (a very large midden surrounding the southern and western portions of House 12), Midden 75 (on the 141 hill slope south of House 15b), Midden 78 (to the east of Midden 75), and Midden 80 (adjoining and to the north of Midden 75). The last group of longhouses, Cluster D, is a collection of four structures, Houses 4, 11, 19, and 20, located around the site periphery. These houses do not form a physically close-knit group or cluster per se, but they share uncommon positioning. The occupants of Cluster D are spatially isolated from those of the village core by the location of the longhouses exterior to the original palisade. At the northeastern periphery of the village, two houses, House 4 and House 11, are better integrated and share the lower-level terrace with Clusters A and B. The two longhouses at the southern end of the village periphery, Houses 19 and 20, are separated from the village centre by the extensive slope surrounding the east and south sides of the village plateau housing Cluster C. Several interior middens within the cluster block easy navigation among these houses. These are Middens 53, 65, 71, and 77. Also interrupting movement are surface Middens 52 and 56, located between Clusters B and D. Table 6-2. Houses and Middens Associated with Each Site Cluster and Elevation. Cluster Houses Middens Elevation A 1, 2, 3, 5, 6, 7 51, 57, 60, 63 low town B 8, 9, 10 52, 53, 54, 55, 56, 58, 67 low town C 12, 13, 14, 15, 16, 18 59, 61, 62, 66, 68, 72,73, 74, 75, 78, 80 high town D 4, 11, 19, 20 65, 71, 77 low town Definition of Site Settlement Phases The Keffer site occupation history can be divided into two main phases of construction, based on the location of the houses and middens in relation to the palisades. Phase 1 consists of those occupation features which, according to settlement pattern analysis, could have existed prior to the construction of, or during the lifetime of, Palisade 1. This phase is further subdivided into Phase 1a, defined as all those features which could have co-occurred during the initial stage of the period, and Phase 1b, defined by features which might have co-occurred as some of the initial structures were replaced while Palisade 1 remained intact. The demolition of Palisade 1 and the expansion and rebuilding of existing houses, together with the construction of new houses and a second, more extensive village palisade, mark the 142 beginning of Phase 2a. The major expansion of the palisaded area during the village occupation provides a clear break in the settlement history. A small extension of the second palisade, Palisade 2, in the southwest corner of the site, marks the final stage of palisade construction, in Phase 2b. Phase 2b is further separated into 2b-1 and 2b-2, the final structural change seen in the village. The separation of the occupation into these phases of development is not meant to suggest that this scenario represents temporal snapshots of any period of the village occupation. Nor is it meant to suggest that changes within sub-phases a or b happened simultaneously in all clusters. This allocation simply represents the potential of co-occurrence of features within the complex village construction history. Overlapping longhouses, palisades, and middens throughout the site make it clear that all houses and areas of the village could not have been occupied simultaneously and help to reveal a relative chronology of settlement pattern. This chronology is examined in detail and illustrated through the use of post mould maps and site plans. Plans indicate newly appearing or newly altered settlement features of each phase by progressively lighter colours in the colour coding (burgundy = Phase 1a; bright red = Phase 1b; orange = Phase 2a; ochre = Phase 2b-1; yellow = Phase 2b-2). Site Settlement Phase 1a Phase 1 is defined by the presence of the first (and smaller), oval palisade. It is thought to correspond to the establishment of the hamlet or village. This phase is broken down into two main events, 1a and 1b (Table 6-3, Figure 6-6). This division is marked by the restructuring of the Cluster C high town settlement before the major village expansion that occurs with the erection of Palisade 2. 143 Figure 6-6. Plan showing Phase 1a. 144 Table 6-3. Possible Concurrent Houses of Phase 1a Clusters. Cluster Houses A 1a, 3a/b, 5, 7a B 8, 9a/b, 10a/b C 15a/b, 12 D 4, 11a, 20 Phase 1a, Cluster A The five aligned houses of Cluster A (Figure 6-7) are efficiently fit into the northernmost level area, inside Palisade 1. The cluster is comprised of two sets of overlapping buildings (Houses 2 and 3, and Houses 6 and 7); a small, circular structure of unknown function (House 5); and one spatially independent residence (House 1). House 1 lies east–west along the southern edge of the palisade. The initial construction of House 1a, in Phase 1a, measured 37.5 m in length. A wall post density of 10 posts per metre (ppm) suggests a long period of occupation for this house. This is supported by the extensive midden, Midden 1, located just metres from its western end wall. The northeast corner of House 1a makes direct contact with the northeastern portion of Palisade 1. The rest of the building sits 10 m from the northern part of the palisade. Less than a metre to the south of House 1 is the north wall of House 2, a 35 m long, lightly occupied dwelling with evidence of a single construction phase. The short-term occupation of House 2 is suggested by the paucity of interior posts and features, as well as the lack of expansion of the residence. It is also supported by the low wall post density of 4.1 ppm, when compared to the site average of 6.5 ppm. 145 Figure 6-7. Plan (left) and map (right) showing Phase 1a, Cluster A. House 3 parallels and overlaps the entire southwest half of House 2. The first construction of House 3a measures 30.5 m and has a wall post density of only 4.1 ppm. House 3b is later extended by 7.5 m. Even with the extension, House 3 appears to have a relatively short lifespan for a Keffer site house. House 2 features a large interior sweatbath (a circular pattern of post moulds approximately 1 m in diameter marking a temporary, tent-like sauna structure used for spiritual cleansing). It is characterized by an extremely high density of posts. This sweatbath overlies part of the north wall of House 3, thus indicating that House 2 definitely postdates House 3 and placing House 2 in a later construction sequence (Figure 6-8). Figure 6-8. Map detail showing House 2 sweatbath postmoulds over House 3 north wall. 146 Also overlapping House 2 on the east is House 5, an 8.6 m circular structure of unknown function. In its first guise, House 3a is located more than 7 m from House 5, but the 7.5 m extension brings the House 3b east wall in contact with that of the House 5 structure. The circular wall of House 5 is also in direct contact with House 6 to its south. House 6, a 35 m building of one construction episode, lies at an angle to Houses 3 and 7 in this cluster, but parallel to Houses 1 and 2. The east end wall of House 6 crosses Palisade 1, while the west end wall is covered by Feature 2, an occupation feature of House 7b (Figure 6-9). House 7a, which in its original form is 28 m long, is situated 3 m south and parallel to Houses 3 and 5. Along with House 3b, Houses 1 and 7, the two longest-occupied structures in the cluster (as indicated by the high wall post density and interior post density [Figure 6-7]) show evidence of structural extensions. House 1a, initially 37.5 m long in Phase 1a, was expanded to 42 m in a later phase, as seen by its growth across the palisade line. The initial construction, fitting very snugly within the palisade, is likely an early construction and is therefore assigned to Phase 1a. Figure 6-9. View of House 6 wall post mounds under Feature 2 of House 7. 147 Post mould density studies do not support the previous interpretation of multiple eastward extensions to House 7 as has previously been argued (Finlayson et al. 1987). At Keffer, as clearly indicated by the House 11 expansions, walls of the final stage of longhouse construction (expansions) are characterized by the lowest post density for the structure. The rebuilding or fortifying of previous expansions, as in House 11b, results in wall segments with higher post densities in the final construction, but lower post densities than in the original section of the building, such as the walls of House 11a. In House 7, Birch (2011) has identified only two episodes of construction. While the initial House 7a has the highest post density of any longhouse yet documented in Northern Iroquoia (Jennifer Birch, personal communication 2014), at 13.4 ppm, the extension of House 7b has a much lower density, below the site average, at 4.9 ppm. This density is constant across a distance of 15.9 m of the 17 m extension side wall. The consistency in wall post density across the length of the addition supports a single house expansion, to a final length of 42 m. The superpositioning of House 7b Feature 2 over the end wall of House 6, which dates to a later phase, indicates that the expansion of House 7b occurred later in the site history, post–Phase 1, after the removal of House 6 and Palisade 1. As noted above, a definitive temporal relationship for all features cannot be ascertained from settlement evidence alone. However, it is clear that House 2 must postdate House 5 and, therefore, House 3. House 6 must postdate the removal of Palisade 1, and according to its parallel alignment with Houses 2 and 7, it probably coexists with these structures, although it pre-dates the extension of House 7b. Therefore, the initial settlement of Cluster A in Phase 1a potentially includes Houses 1a, 3a, 3b, 5, and 7a. Houses 2, 6, and 7b belong to a later construction phase. With the exception of Midden 54, which overlies Houses 6 and 7b, all middens in this area may have been formed during this early phase. These include the small surface middens 51, 63, and 64 and the large hillside middens 57 and 60. Phase 1a, Cluster B Cluster B (Figure 6-10) is composed of three longhouses (Houses 8, 9, and 10), four middens (Middens 54, 55, 56, and 67), and a semi-circular fence line 50 m in length, here labelled Palisade 3. House 8 lies parallel to the Cluster A House 7 and 5 m to its north. Although it sits in parallel alignment with Houses 1 to 7, and thus appears to belong with Cluster A, House 8’s eastern end is not aligned with those of Houses 1 to 7. Together with the north ends of Houses 9 148 (38 m long) and 10 (28 m long), the western porch/storage area of House 8 opens onto a central plaza and refuse area probably shared by these three houses, indicating they form a socially linked group, Cluster B. Post lines and interior feature data advocate for a single construction event for House 8. A wall post density of 7 ppm, slightly above the site average, supports a relatively long lifespan for the house. The post layout (Figure 6-10) and the presence of wall trenches suggest a rebuilding of this longhouse at some point, but no extensions. Instead, at the west end of House 8 is an interior end wall, commonly used to separate the living and storage/exit areas of longhouses, labelled “vestibules” by Creese (2010:257). Figure 6-10. Plan (left) and map (right) showing Phase 1a, Cluster B. The closely aligned and parallel Houses 9 and 10 are angled at 40 degrees to the southeast of House 8, forming a rough semi-circular cluster at their western ends. The parallel, side-by-side alignment of Houses 9 and 10 indicates coexistence at some point. House 9a, with a wall post density of 7.9, was originally a small building, 27.5 m in length. It experienced expansions at both ends of the building, 6 m to the north and 7.5 m to the south. The final structure is 40 m in length. House 9 was probably erected in this phase (Phase 1a), since the first village palisade, Palisade 1, skirts around it and House 8 at a distance of 10 m, a common buffer zone size at Keffer. The heavy concentration of interior sweatbaths present in House 9 suggests a long or 149 intensive occupation. House 10 was also in the first instance built as a small residence, of only 22 m. It was later expanded by 7 m, and the exact fit of House 10b (at 28 m) within the eastern and western limits of Palisade 3 suggest that the house was extended from its original length after the erection of this interior palisade or fence line. The paucity of interior features, low density of wall posts, at 4.3 ppm, and the superpositioning of Midden 58 over the interior of House 10, portray a short habitation period for this house and confirm its abandonment prior to the end of the village occupation. The purpose and timing of the erection of a 50 m barrier or fence line (Palisade 3), bordering House 10 on three sides is not clear. It likely provided a barrier for protection of some kind, perhaps related to climatic, geographic, or social factors. The placement of Palisade 3 in direct contact with the end of the House 10 sidewall replicates the not uncommon practice at Keffer of abutting longhouse corners to palisade structures. Logic would suggest, however, that the palisade was constructed prior to the extension of the smaller House 10a (21 m), when a sizable, 10 m gap would have existed between the two constructions, allowing easy passage. While the second iteration of the house fits neatly within the palisade, it would have restricted access to and from House 10. The co-presence of the very small refuse accumulation of Midden 67, at the base of the slope from the high town, directly adjacent to—and possibly contemporary with—Palisade 3 would have had little effect on intra-site mobility. Since the interior fence line, Palisade 3, surrounding Houses 9 and 10 on the west, already limits much of the pedestrian access for the inhabitants of these structures; the addition of Midden 55 at their northern egress would have almost entirely cut off access among these three structures. This situation suggests that Houses 8, 9, and 10; Palisade 3; and Midden 55 were not entirely coeval. However, the lack of superpositioning leaves the temporal positioning of these features unknown. Also associated with this grouping (specifically with Houses 9 and 10), is Midden 56, a surface midden at the southeastern end of House 10, which abuts the interior palisade. Its presence likely posed only a minor obstacle while Palisade 1 remained in place, as it lay nested alongside the interior fence wall. in an alternate scenario this fence line as either predated or post-dated the village. 150 Phase 1a, Cluster C Figure 6-11. Plan (left) and map (right) showing Phase 1a, Cluster C. The high town plateau at Keffer gradually rises to a height of 3 m above the level ground of the low town. Tightly fit within this tableland are six overlapping longhouses, 12, 13, 14, 15, 16, and 18, which together form Cluster C (Figure 6-11). The six houses involved in this development include five with some evidence of growth over time (Houses 13, 14, 15, 16, and 18). Three structures, Houses 12, 13, and 14, display evidence of long-term occupation, as seen in the high density of interior posts. Houses 15, 16, and 18 contained fewer interior posts and features, suggesting they experienced a shorter occupation. The longest house in the high town, House 15, at its greatest extent measured 65 m long, but it was originally constructed to a length of 37.5 m, roughly two-thirds the final dimension. The west wall of House 15a merges with the east wall of House 18a (25 m long). It is unlikely these two structures co-existed, as large areas of the plateau around the buildings would thereby remain open, while the area between the buildings would be impassible. All other houses in Cluster C (the high town) except House 14 could have existed simultaneously with either House 151 15a or House 18a, but not both. However, because the earliest House 18 construction crosses the surrounding palisade, Palisade 1 clearly belongs to a later village construction phase. The overlap of up to four constructions, Houses 14a (32 m), 14b (38 m), 15a, and 15b, within a small area makes post mould patterns difficult to decipher. Nonetheless, a distinct end wall for House 15a can be deciphered (Figure 6-11). More than 50 percent of the area covered by both stages of House 14’s construction overlaps almost the entirety of House 15a and portions of House 15b. The parallel alignment of Houses 14 and 18 suggest they are concurrent structures, indicating that House 14 also belongs to this later stage. House 15 therefore appears to be the earliest of these three longhouses. The early construction of House 15 is supported by its central location across the middle of the plateau. As Houses 13, 14, and 16 cross House 15b, which is the latest extension of this structure, they, too, must post-date House 15. Post mould studies show House 15 to have the lowest wall post density of all the houses at Keffer, at 3.8 ppm, with no evidence of repairs or rebuilding of the original structure. Internal sweatbath features are present, but not in the high density seen in the neighbouring House 12. This indicates that House 15 witnessed a relatively short occupation. Yet House 15b was expanded to the south an additional 27.5 m, making it the village’s largest longhouse, at 65 m in length. Midden 78, located on the slope just beyond the house end, may have originated as a result of this expansion. House 12, built at 45.1 m, shows no evidence of rebuilding, but the wall post density, of 6.4 ppm is much higher than that of House 15, and a dense pattern of sweatbath features covers the entire extent of the interior. The high concentration of sweatbath posts located throughout the longhouse interior indicates that this house experienced the longest occupation of all six houses of the high town. These characteristics, along with House 12’s lack of building superimpositions, indicate a long occupation for this house. The existence of House 12 prior to the final, lightly occupied (and therefore probably short-lived) extension of House 18c is evidenced in the overlap of this expansion across the north end porch of House 12. Since House 12 did not survive to the end of the village occupation, as revealed by the overlapping ends walls of House 18c, it must have been constructed early in the village history. Consequently, House 12 may have co-existed with House 15, which pre-dates Houses 13, 14, 16, or 18a/b. Although an early presence of the smallest version of House 18 is possible in the high town, in the earliest stage its abutment with House 15a in a sparsely built area, and its parallel alignment 152 with the later House 14, suggest it dates to a later construction phase. There are clearly two sets of post moulds in the west end of House 18b. This would seem to confirm that this end of the house was rebuilt once, following the same outline. Both these sets of posts cross the Palisade 1 line, demonstrating that this portion of the house was erected after the palisade was moved. While Houses 13a (28 m), 13b (39.9 m), and 16 (approximately 33.5 m, although the true length is impossible to obtain due to erosion) may have existed in conjunction with the earliest House 15 (15a), their overlap of House 15b suggests they were constructed later. The overlap of the southernmost limit of House 13 with Palisade 1 (Figure 6-12) shows that House 13b post-dates the palisade and that it coincides with the construction of House 18b. Therefore, Houses 15a and 15b must both pre-date the appearance of Houses 13, 14, 16, and 18. Archaeological settlement data, therefore, suggest that only Houses 12 and 15 could have co-existed in this earliest stage. Figure 6-12. Map detail showing House 13b wall posts crossing Palisade 1. Long-term occupation of the six houses on the plateau resulted in the deposition of several extensive hillside middens and two surface middens. Midden 59, lying at the apex of the cluster of Houses 14, 15, and 16, is situated at the top of the plateau slope but does not form a communication barrier. Surface Midden 73 slightly overlies the House 15b end wall, and thus most likely post-dates the removal of this structure. As neither midden intersects any of the structures, they provide no clues to house sequence. Middens 61 and 68, on the hillside of the northeastern plateau, inside Palisade 1, and Middens 72, 75, 78, and 80, exterior to Palisade 1, cross no palisade or house post lines, and thus cannot be attributed to any specific era of the 153 hilltop occupation. These exterior middens may therefore date to a later phase. Middens 62 and 74 cross both the Palisade 1 and the Palisade 2 lines, suggesting long-term deposition. The possible Phase 1a Cluster C features thus include Houses 12 and 15a/b and Middens 59, 61, 68, 73, 74, 75, and 80. Phase 1a, Cluster D To the east of the original village palisade, Palisade 1, on the remaining level ground south and west of a rise to the north and east, four additional longhouse structures were constructed— House 4 to the northeast, House 11a directly to the east, House 20 to the southeast, and House 19 farther south. These structures and the six associated middens, Middens 52, 53, 65, 67, 71, and 77, are designated Cluster D. Whether this collection of occupation features combines to form a true cluster is uncertain, but their existence within the expanded village area between the two village palisades suggests some form of connection. Unlike Houses 4, 11, and 20, the southernmost of the houses, House 19 (a shorter, 25 m, single-construction longhouse), undoubtedly belongs to a post–Phase 1 stage, as it overlies the construction of the second village palisade, Palisade 2. Similarly, Midden 52 can also be assigned to a later phase, as it partially covers Palisade 1, next to House 4. Conversely, Midden 53, situated just south of House 4, is composed of isolated surface refuse and may derive from this early time in Phase 1a. Excavation notes suggest that Midden 76, originally assigned to an area overlapping the west end of House 19, is part of the post-occupation disturbance and should not be considered an archaeological or cultural feature. 154 Figure 6-13. Plan (left) and map (right) showing Phase 1a, Cluster D. Settlement pattern evidence alone suggests that House 11 may have been the first residence constructed in a central location in this open area outside the original village boundary. At some point prior to the removal of the first palisade, perhaps even early in the village’s lifespan, the first, 40 m construction of House 11 (11a) was erected, partially abutting the exterior east wall of the palisade. A further sign of an early (perhaps Phase 1a) date for this structure is indicated in the features of the house itself. While the interiors of most of the intensively occupied longhouses at Keffer, including House 11, show evidence of multiple aboveground sweat lodges, common in the late fifteenth century (Williamson 2014:11), House 11 is the only structure at the site with evidence supporting the presence of a possible semi-subterranean sweat lodge within its interior (MacDonald and Williamson 2001:66–67; Williamson 2014:9). A large (203 × 183 cm) feature (Figure 6-14) abuts the construction of the House 11 east wall. The rebuilding of this wall 155 may have obliterated any evidence of a passage into the house. Although its depth is unknown, the large size and location of this feature (Feature 4) suggest it may have functioned as a sweat lodge. The proposed function of this feature remains speculative, however. The occurrence of a sweatbath in this location would argue for a relatively early occupation of House 11a in the village history, as no evidence of additional semi-subterranean sweat lodges has been found at Keffer. House 11a has a high wall post density, of 10.5 ppm. This density, second only to that of House 7a, together with the numerous interior occupation features in the south end of the house, indicate it is one of the longest occupied households of the site. It also supports an early date for the construction of House 11a’s southern section. Figure 6-14. Map detail showing south end of House 11. Whether Houses 4 and 20 were constructed concurrently with House 11 remains unknown, since neither of these structures possess wall post lines which intersect with those of other features, although House 4 sits in almost direct contact with the outer palisade, Palisade 2. This house was originally built with a length of 42.5 m and was not extended. With the exception of Houses 15 and 11, at 65 m and 62 m, respectively, most extended houses at Keffer have a final length of this same approximate dimension, of 40 m. House 4 is aligned in parallel with, but 5 m from 156 Palisade 1. This house was most likely constructed prior to the removal of this palisade, before the elongation of House 11b/c, and prior to the erection of House 6, thereby affording it a wide exterior pathway running from end to end. House 4 has a low interior post density, with the exception of a large sweatbath feature, with a high density of post moulds suggesting multiple constructions, in its southern end. It is also characterized by a low wall post density, of 4.8 ppm. This density measure is very similar to those of the short-term structures, Houses 2, 3, and 6, as well as the extension of House 7b, at 4.1, 4.7, 4.3, and 4.9 ppm, respectively. The low ppm, together with the absence of a large midden deposit in the immediate area, suggests a relatively short occupation for this residence. House 20 has a heavier feature density (and therefore longer lifespan) than House 4. It is located, along with House 19, at a distance from the other site residences. Its north end lies 25 m from House 11 and approximately 40 m southeast of the next closest house. House 20 is post-dated by House 19, which lies only 10 m to its southwest. Surface Midden 71, located directly north of House 20, probably appeared with the first occupation of House 20. As it is physically separated from other village features, Midden 71’s existence in Phase 1a is uncertain and cannot be ruled out. In contrast, Midden 77’s overlap of Palisade 2 indicates its deposition in a later phase. Lying between Houses 11 and 20 (two long-term residences) is Midden 65, the largest refuse deposit in the village. The position of Midden 65 within a wide depression on the village’s southeastern periphery, distant from the main village occupation, suggests it could have been an early location for refuse deposition. There is a lack of evidence for the intersection of Midden 65 with other village features; therefore, there is a possibility that it, too, may originate in Phase 1a. The archaeological settlement data for Phase 1a of the village residential development allow for the erection of 11 longhouses concurrently with the construction of Palisade 1 in this early period. The palisade was evidently designed to skirt House 1a along the northern portion of the village, as it runs at a distance of some 20 m beyond the north wall of House 3a and 15 m east of House 5. Following this same reasoning, Houses 8 and 9 must have also been present at the time of the construction of Palisade 1. Palisade 3, a small, interior, semi-circular structure surrounding and abutting the northwest wall of House 10, seems to isolate the area of Houses 9 and 10 from the upslope occupation of the high town. The wall trench of House 9 (Feature 1, Excavation Square Form 470-515) interrupts the Palisade 3 (Palisade 301 on excavation records) wall, with no evidence of the palisade posts intruding into the construction trench. Palisade 3 is therefore 157 earlier than House 9 in this spot. Whether Palisade 3 is earlier or later than House 10 is unknown. At its greatest expanse, after the mostly northerly portion of the barrier was disassembled, this fence line enclosed both Houses 9 and 10. Site Settlement Phase 1b Phase 1b, Cluster A Phase 1b, is illustrated in Figure 6 15 and summarized in Table 6 4. The maps of the individual clusters within it are shown in Figure 6 16 to Figure 6 19. 158 Figure 6-15. Plan showing Phase 1b. 159 Table 6-4. Possible Concurrent Houses of Phase 1b Clusters. Cluster Houses A 1a, 3a/b, 5, 7a B 8, 9a/b, 10a/b C 12, 13a, 14, 16, 18a D 4, 11a, 20 Figure 6-16. Plan (left) and map (right) showing Phase 1b, Cluster A. The shift into Phase 1b is marked by only minor alterations in the configuration of longhouse and midden features throughout the lower level of the village. In contrast, occupation on the plateau, in Cluster C, experiences fundamental realignments. The changes marking the end of Phase 1a in each cluster are specific to those clusters and have no relation to the timing of changes delineating the commencement of Phase 1b in other clusters. The relative temporal relationship of the events among clusters in Phase 1b is unknown. This restriction disappears with the advent of Phase 2 and so, consequently, the description of modifications via clusters is used for Phase 2. Possible modifications in Cluster A during Phase 1b include the expansion of House 3a to its final length of 26.3 m (House 3b), perhaps in 160 conjunction with the removal of the House 5 structure, which would have opened up transit paths in this section of the cluster. Figure 6-17. Plan (left) and map (right) showing Phase 1b, Cluster B. Phase 1b, Cluster B Nearby, in Cluster B, it may be that, while House 8 remained essentially unchanged, Houses 9 and 10 experienced small expansions, of 6 m to the north and 6 m to the south, respectively. It is also possible that these areas, with little evidence of use, served as porch/storage areas. On the other hand, the southern extension of House 9 shows evidence of intense use. Palisade 3, the small fence/palisade surrounding Houses 9 and 10 and protecting them from the west, may have been erected before the extension of House 10a. Changes to House 9 may have occurred at any point in the village occupation. Phase 1b, Cluster C Cluster C experienced the greatest transformation in Phase 1b. While changes in this latter part of Phase 1 are minimal in most areas of the village, Cluster C was subject to almost total renovation. This transformation is initiated with the removal of the village’s longest residence, House 15b. House 15a/b may possibly constitute the most briefly occupied dwelling of the village, as demonstrated by its low interior feature density and the village’s lowest wall post density. (Table 6-1). Unlike those residences with more lengthy occupation, it possesses no wall trenches and shows no evidence of rebuilding or renovation. In contrast, the adjacent and 161 concurrent House 12 has a density of 6.4 ppm. Although House 15 intersects with four other houses, their radial alignment makes most of their associated interior features relatively easy to separate, indicating their residential affinity. Figure 6-18. Plan (left) and map (right) showing Phase 1b, Cluster C. The plateau settlement layout appears to have been planned ahead of construction, with House 18 well integrated into the layout, in a parallel alignment with Houses 12, 13, and 14 in the post– House 15 settlement. Four shorter longhouses were erected in the significant stretch of hilltop that opened to the north and east of House 12 following the demolition of House 15b. These new structures include Houses 14a/b and 16a/b, as well as the initial versions of Houses 13a and 18a, both of which fit snugly within the original village palisade. The close proximity of the five structures left a small central plaza area but little extra exterior transit space. House 14 was erected almost directly over what had been the north end of the House 15. House 18a sits parallel and 2 m to the east of House 14. House 13a, 10 m south of House 14, is similarly aligned, 5 m east of House 12. Examination of the four structures in isolation from the rest of the high town features would suggest that Houses 13 and 16 either pre-date the southern extension of House 15 or post-date 162 the structure entirely, as they overlay the House 15b extension, but not the initial House 15a construction. Houses 13 and 18, however, both feature house expansions overlapping Palisade 1, suggesting they post-date both House 15 and this early palisade. The low interior post feature density within House 18 indicates a short occupation history. The highest feature and post mould density are seen in the northern end of the building, affirming this as the area of longest occupation. Yet the wall post density of House 18a, at 6 ppm, is quite low for a structure with evidence of two expansions and renovations. The intersection of the House 18b wall posts with those of the Palisade 1 post line indicate a later date for this house extension. It is conceivable that House 18 was not constructed during Phase 1b. House 18 and Midden 62 may belong to the later Phase 2. There is no evidence for the removal of the first village palisade during this stage, and all surrounding refuse deposits may have continued to amass along the hillside slopes, in Middens 59, 61, 62, 68, 72, 74, 75, 78, and 80, throughout Phase 1b. Although it is clear that the residences of the high town, Cluster C, cannot all be contemporaneous, a definitive relative occupation sequence cannot be ascertained solely from the settlement evidence. Phase 1b, Cluster D The position of Cluster D (Figure 6-19), made up of Houses 4, 11a, and 20, between the first and second palisades, and the lack of any physical relationship with structures of Clusters A, B, and C, within the first palisade, obscure the timing of any changes which might have occurred during Phase 1. All dynamics attributable to Phase 1a are equally possible in Phase 1b. House 19 was not present during this phase. 163 Figure 6-19. Plan (left) and map (right) showing Phase 1b, Cluster D. Site Settlement Phase 2a The shift to Phase 2 is marked by the removal and replacement of Palisade 1. Post mould evidence suggests that the entire palisade was removed and rebuilt, expanding the interior of the village 25 m to the east, to the point where the land begins to rise uphill again; 10–15 m to the west, down the edge of the ravine; 10 m to the north; and 35 m to the south, to the edge of the level terrain (Figure 6-20). In these areas, the second palisade, Palisade 2, is double-rowed, with 1.5 m between the rows. The expansion of the village limits 35 m southward, off the hillside to flatter ground, allowed for the extension of several of the existing houses at the site and the addition of new ones. The houses, middens, and palisades belonging to Phase 2a are summarized below. 164 Figure 6-20. Plan showing Phase 2a. 165 Table 6-5. Possible Concurrent Houses of Phase 2a Clusters. Cluster Houses A, B, and D merged 1b, 2, 4, 6, 7a, 8, 9b, 10b, 11b/c C 12, 13a/b, 14a/b, 16, 18a/b D 19, 20 Phase 2a, Cluster A-B-D North In Phase 2a, the removal of Palisade 1 and construction of Houses 4 and 11 change the configuration of the inter-longhouse relations. Because inter-household patterns are no longer clear the houses of Clusters A and B and the houses of the north end of Cluster D are grouped together (Figure 6-21). The southern part of Cluster D forms its own cluster (Figure 6-22), as does the entirety of Cluster C (Figure 6-23). Figure 6-21. Plan (left) and map (right) showing Phase 2a, Cluster A-B-D north. In Cluster A, House 1b was extended to its final length of 44.5 m, crossing the former palisade line. Although the relative timing of the replacement of House 3 with House 2 and the coincident removal of House 5 cannot be determined in relation to the expansion of the village palisade in 166 Phase 2, it is probable that the erection of House 6, which crosses the earlier palisade line and is in parallel alignment with House 2, occurred at this time. House 5 posts overlap with those of House 2, indicating they cannot be coeval. The coexistence of the converging walls of Houses 5 and 6, which block all passage in the vicinity, is improbable, suggesting that the removal of House 5 also occurred with this restructuring of Cluster A. Houses 2 and 6 were therefore likely erected simultaneously with the destruction of Houses 3 and 5. House 6, which underlies the House 7b extension, was placed to the end of the original House 7a building. Although there is little to no change evident in the structure of Houses 8, 9, and 10 of Cluster B in conjunction with the elimination of Palisade 1, the removal of the palisade profoundly increases accessibility among houses of the three clusters and contributes to the integration of all houses on the lower town elevation. At some point subsequent to the removal of the palisade, deposition of Midden 52 began over the earlier barrier between Houses 4 and 8, suggesting shared usage of space across the previous barrier. Palisade 2 was erected directly east of House 4, at the very edge of the flat terrain. The limitations on available space led to Palisade 2 being erected in very close proximity, side-by-side with House 4. Just south of House 4, House 11 was extended by 10 m over the former village palisade, to a length of 35 m. Figure 6-22. Plan (left) and map (right) showing Phase 2a, Cluster D south. 167 There is no evidence of any changes to House 20 at this stage, but the village area now encompassed by the new palisade is greatly expanded, leaving more than 10 m of free space around it and Midden 65. To the south of House 20, post mould evidence suggests that House 19 was erected outside the village area, with its northeast corner in direct contact with the southern edge of Palisade 2. Alternatively, House 19 may have appeared in conjunction with Palisade 2, being purposely excluded from the larger village. There is little evidence in the way of interior posts or features to support an intense occupation history for this structure. However, a relatively high density of wall posts, 7.7 ppm, may indicate that the longhouse was rebuilt after a brief period of occupation. Midden 77 may have initially appeared between Houses 19 and 20 around this time. Phase 2a, Cluster C On the plateau, in Cluster C, the expansion of the interior village area allowed for the extension of House 18b to the north and, possibly simultaneously, to the south, bringing its length to 36 m. The final and very short-lived expansion to a total length of 43 m overlies the very north end of House 12. The destruction of House 12 and expansion of House 18 may have occurred in this phase or later, in the next stage. 168 Figure 6-23. Plan (left) and map (right) showing Phase 2a, Cluster C. The growth of House 13b to 40 m also happened after the removal of Palisade 1. House 11c, the longest structure in this latter part of the occupation, at 62 m, may also have been extended in Phase 2a. Directly south of House 13, at the top edge of the hillside slope, Midden 72 also appears at this time, as seen by the sub-midden post moulds of Palisade 1. Midden 66, down the west slope of the embankment and accessible to all houses of the high town, also accumulates with the eradication of the palisade. Located across the plateau and not associated with the palisades, House 14b may have expanded from 35 m to 41 m at any time during its existence. Similarly, House 16b appears to have experienced enlargement, from approximately 32 m to 40 m, at an unspecified time (Figure 6-24). 169 Figure 6-24. Map detail showing House 16b extension. Site Settlement Phase 2b Phase 2b marks the final episode of construction in the Keffer village. Phase 2b is further divided into Phase 2b-1 (Figure 6-25) and Phase 2b-2 (Figure 6-26). Figure 6-25. Plan showing Phase 2b-1. Figure 6-26. Plan showing Phase 2b-2. 170 Two main events mark this final stage. The first, seen in Phase 2b-1, is the extension of the enclosure (becoming Palisade 2b) around the southern flank of House 19 at some point after its construction on the south side of the pre-existing Palisade 2 (Figure 6-27), in this way bringing the residence within the physical and social limits of the village proper. The second major change is seen in Phase 2b-2. Table 6-6. Possible Concurrent Houses of Phase 2b Clusters. Cluster Houses Palisade(s) A-B-D 1b, 2, 4, 7a, 8, 9b, 11c Palisade 2b, 3 C 13b, 14b, 16b, 18c Palisade 2b D 20 Palisade 2b Figure 6-27. Plan (left) and map (right) showing Phase 2b-1, Cluster D south. 171 Phase 2b-2, Cluster A-B-D North Following the relocation of the southern portion of the palisade to incorporate House 19 into the larger community, the number of occupied longhouses in the village began to decline (Figure 6-28). This is seen in the removal of three of the shortest and least densely occupied longhouses, Houses 6, 10, and 19. The deposition of interior Middens 54, 58, and 77, respectively, over the occupation layers of these three houses confirms the end of their occupation. The presence of Midden 54 over the eastern extension of House 7b indicates that this structure was either shortened or abandoned at this time. Deposits of Midden 57 overlie the western end of House 7, although it is not clear whether this is due to post-occupation disturbance of the large midden or post–House 7 use. Although Midden 58 is quite small and may therefore represent a short period of deposition, both Middens 54 and 77 are substantial, and their presence is indicative of a longer period of use. Figure 6-28. Plan (left) and map (right) showing Phase 2b-2, Cluster A-B-D north. Phase 2b-2, Cluster D South The expansion of Midden 77 over the east end of House 19 indicates the removal of this structure at a time when House 20 was still occupied (Figure 6-29). The sparse nature of the house’s interior features and low post mould density support a short lifespan for this structure. 172 Figure 6-29. Plan (left) and map (right) showing Phase 2b-2, Cluster D south. Phase 2b-2, Cluster C The expansion of Midden 74 over the southern end of House 12, along with the extension of House 18c over its northern end, signals the removal of this longhouse while other houses on the plateau survive (Figure 6-30). 173 Figure 6-30. Plan (left) and map (right) showing Phase 2b-2, Cluster C. Summary of Site Settlement Phases Each of the phases—Phases 1a, 1b, 2a, and 2b—illustrate a range of potential co-existing features within the village. Thus, the preceding discussion can be visually summarized in a short series of five village layout maps displaying the development of the village layout in a general order determined by settlement feature relationships and superpositioning (Figure 6-31, Table 6-7). Figure 6-31. Plans of (left to right) site settlement Phases 1a, 1b, 2a, 2b-1, and 2b-2. 174 Table 6-7. Houses, Middens, and Palisades Associated with Each Cluster of Each Occupation Phase. Phase 1b Cluster A B C D A B C 2a D 4, 11a, 20 A, B, and D north merged 1b, 2, 4, 6, 7a, 8, 9b, 10b, 11b/c C 12, 13a/b, 14a/b, 16, 18a/b 1a Houses 1a, 3a/b, 5, 7a 8, 9a/b, 10a/b 15a/b, 12 4, 11a, 20 1a, 3a/b, 5, 7a 8, 9a/b, 10a/b 12, 13a, 14, 16, 18a D south 19, 20 A, B, and D north merged 1b, 2, 4, 7a/b, 8, 9b, 11c 2b-1 C 12, 13a/b, 14a/b, 16, 18a/b D south 19, 20 A, B, and D north merged 1b, 2, 4, 7a/b 8, 9b, 11c 2b-2 C 13b, 14b, 16, 18c D south 20 Middens 51, 57, 60, 63 55, 56, 67 59, 61, 68, 75, 78, 80 53, 65, 71,77 51, 57, 60, 63 55, 56, 67 58, 59, 61, 62, 68, 72, 73, 74, 75, 78, 80 53, 65, 71, 77 51, 52, 53, 55, 56, 57, 60, 63, 65, 67 59, 61, 62, 66, 68, 72, 73, 74, 75, 78, 80 Palisades 1, 3 1, 3 2, 3 65, 71, 77 51, 52, 53, 55, 56, 57, 60, 63, 67 59, 61, 62, 66, 68, 72, 73, 74, 2b, 3 75, 78, 80 65, 71, 77 51, 52, 53, 54, 55, 56, 57, 58, 60, 63, 67 59, 61, 62, 66, 68, 72, 73, 74, 2b, 3 75, 78, 80 65, 71, 77 Phase 1a The initial configuration of the Keffer village was likely shaped by the concurrent construction of a single-row palisade encompassing longhouses belonging to three of the four residential clusters at the site (Clusters A, B, and C). However, there is no concrete evidence to confirm contemporary construction. Settlement pattern evidence suggests the co-presence of three longhouses, Houses 1a, 3a, and 7a, and one poorly understood circular structure, House 5, in Cluster A, within the northern periphery of Palisade 1. Each of these houses was later extended to its final configuration. Houses 3 and 7 may have been expanded during the existence of this first palisade. House 1 was expanded in a later phase. These three houses form a tightly knit, parallel cluster and are associated with large hillside middens, Middens 57 and 60, to the west. The small surface midden, Midden 63, may have been deposited simultaneously, although its position in the centre 175 of what may have been an outdoor activity area makes this timing uncertain. To the south, Cluster B, composed of three radially situated longhouses, Houses 8, 9a, and 10a, opened onto Midden 55, a sizeable surface deposit. A large open area existed between Houses 8 and 9. House 10, surrounded by a small palisade or fence line was originally a shorter structure but may have been expanded during this phase to fit neatly within the fence enclosure. Deposition of Midden 56, to the south of Houses 9 and 10, may also have begun in the period. Midden 67, just outside the House 10 fence line, may be the product of this cluster or of the high town Cluster C House 15. This house, along with House 12, is one of the first structures to appear in the high town. Phase 1b House 15 was expanded sometime after its initial construction, but the evidence of limited internal features suggests that this expansion had a short duration of occupation. Middens 59, 73, and 78 may have started to accumulate after House 15a expanded into the southern portion of the hilltop, to form House 15b. Numerous middens surround the plateau, and while most could have been used in Phase 1a, it is probable that only a few would have been. These are Middens 61, 68, 74, 75, and 80. Outside the palisade, Houses 4, 11a, and 20, along with their associated middens—Middens 53, 65, and 71—may have emerged soon after the construction of the first village palisade, and certainly would have emerged prior to its removal and the building of a double-row palisade that included a much larger portion of the site. On the plateau, a major reorganization resulted in the replacement of House 15 with four new houses, Houses 13, 14, 16, and 18. Phase 2a The next major step in the residential development occurred with the erection of an expanded village palisade, Palisade 2, which now incorporated the three houses of Cluster D, Houses 4, 11, and 20, inside the village area. House 19, a small and short-occupation residence, was constructed outside and touching Palisade 2 soon after it was erected. The increase in internal village area allowed for the expansion of Cluster A Houses 1b and 3b, Cluster D House 11b/c, and Cluster C Houses 13b and 18b, across the previous palisade line. Houses 14b and 16b in Cluster C also experienced expansion. In Cluster A, the new House 6 is placed in the area opened up with the removal of the interior palisade. Houses 3 and 5 are removed and replaced with the new house, House 2. This structure shares better alignment with Houses 1 and 6 than the 176 now-removed House 3. The small House 19 residence was constructed abutting the exterior wall of Palisade 2 soon after this barricade was erected. Phase 2b Shortly afterward, the southern limits of this portion of Palisade 2 were expanded 15 m to the south, to incorporate House 19 into the community. Around this time, or perhaps slightly earlier, House 12 in the high town cluster is removed to make way for the southern expansion of House 18c. Low wall post and interior feature densities indicate that this expansion was extremely short-lived. Deposition of Midden 77 on top of the east wall trenches of House 19 confirms the short lifespan of this residence. In the last stages of the village occupation, Houses 6, 10, 12, and 19 are abandoned, while House 18, and possibly House 11, are slightly expanded, to form Houses 18c and 11c. Discussion The complex history of the Keffer settlement, with its constantly shifting village residential patterns and the consequent indications of both social inclusion and social exclusion, are evident in the archaeological settlement pattern data. The allocation of each of the village features, houses, middens, and palisades to phases depicting their first potential presence in the village indicates an early appearance of most components. With this settlement sequence as a framework, network analysis of the village ceramics will be employed in the next chapter to provide a more nuanced portrayal of the temporal and social relationships among the Keffer village inhabitants. 177 7 Analysis of Keffer Site Local Tradition Ceramics Factors Involved in Keffer Site Network Analysis I employ network analysis of the local tradition ceramics to produce two sets of linked relations, concerning both chronological and social ties. In social terms, I aim to reveal ceramic communities of practice as they appear in the village both synchronically and diachronically. Communities of practice are seen in highly connected topological subgroups within the network structure. To understand the development of these ceramic communities throughout the lifespan of the village, demonstration of chronological control proves essential. Production of this chronology is therefore pursued prior to the examination of individual communities of practice. Both network and conventional ceramic seriation approaches are employed in conjunction with the settlement pattern analysis performed in Chapter 6. This chronology is then integrated with the newly revealed communities of practice to produce a temporal view of ceramic practice development. As network analysis of ceramic collections at the intra-site level is a novel approach a highly detailed investigation of the Keffer ceramics is described in Appendix H to fully explore this process. In Appendix E the effect of graph layout on the interpretation of the Keffer dataset was intensively examined. Graph layout was seen to have a substantial effect on visual interpretation, and it was concluded that network graphs portrayed in Gower metric and non-metric multidimensional scaling (MDS) best display the true strength of relationships both between individual houses and middens and throughout the greater village. Consequently, these two graph algorithms are adopted here in the examination of local tradition ceramic relationships and communities of practice across the Keffer village. The data matrices used in this analysis are composed of 12 values, in this case Iroquoian ceramic types, within 1 variable, that of ceramic type. This list of values is exclusive to all those types deemed by the author to be typical of ceramics produced by post–A.D. 1350 Iroquoian peoples of south-central Ontario west of the lands occupied by St. Lawrence Iroquoians in the precontact period. These types are assigned to the local tradition ceramic category. They have not been assigned to the more exclusive categories of ethnic Iroquoian affiliation or time periods within the data structure, so as not to bias the interpretation. 178 The total ceramic collection from each archaeological locus, whether residential or refuse deposit, represents the entire period of use of that feature. In the absence of extensive research into the location of each sherd, none of these ceramic collections can be considered indicative of one specific event or time period in the locus’s lifespan. Episodes of deposition, whether initial or final, cannot be separated from the long-term praxis of use. The sample, as a whole, instead, represents the entire temporal life span of the feature. This results in highly skewed samples sizes within the collection. To mitigate against the problems associated with the great range in size of the loci samples, from 6 to 966, and the extremely small size of some of these samples, matrices of three differing sample size parameters are used to examine the effect of sample size on network results. The first of these includes all samples containing between 2 and 19 vessels (labelled Over 1 Vessel); the second includes all samples containing between 20 and 29 vessels (labelled Over 19 Vessels); and the third includes all samples containing 29 or more vessels (labelled Over 29 Vessels). Statistically, sample sizes larger than 30 are considered more robust, and results produced from the Over 29 matrix are therefore considered the most reliable. However, several important loci on this site are characterized by samples of fewer than 29 vessels. While those features with 20 to 29 vessels (Over 19 Vessels) may provide statistically weaker results, their inclusion contributes essential evidence regarding the nature of their relationships. There can be little confidence in results arising from network analysis of samples under 20 vessels (Over 1 Vessel), but they offer the only hints regarding the relationships of the archaeological features containing these vessels with other features within the greater ceramic community of the village. Three loci, Houses 14, 16, and 18, are characterized by only one vessel each, and these ceramics have not been included in the analysis. Three network graph runs were completed with each of the three matrices, for a total of nine runs. In the first step, the most exclusive (Over 29 Vessels) database was analyzed at increasing similarity levels: BR 170, 180, and 190. The Over 19 Vessels and Over 1 Vessel matrices were then run following the same procedure. When the similarity level is increased with each run, the number of nodes included in the graph decreases and the pull exerted on the remaining nodes is also reduced. Through visual comparison of these ties, the effect of smaller samples on the overall structure of the graphs can be evaluated. Network statistics, graph density, and node degree are employed to confirm these observations. 179 These analyses confirmed that both smaller sized samples and decreases in similarity levels (Table 7-1) affected the relationships among larger samples with stronger ties. In both cases the inclusion of additional nodes on the peripheries of the graphs caused an alteration in original topology. Accordingly graph interpretation is founded on the highest-level similarity graphs of samples with a minimum of 30 vessels. These observations form the core of the results with lower-level ties, and those of smaller samples providing additional inferential information. Table 7-1. Houses and Middens Included at BR 170 in Each of the Seven Node Cluster Groups for Each Sample Size. Node Cluster Group House/Midden Over 29 Vessels House/Midden Over 19 Vessels House/Midden Over 1 Vessel 1 a 52, 54, 60 52, 54, 60, 65, 72, 73, 74 52, 54, 60, 65, 72, 73, 74 b 65, 72, 73, 74 2 59, 61, 66, 71 1, 4, 7, 59, 61, 66, 71 1, 61, 63, 66, 71, 10, 77, 80 3 13, 20 13, 20 13, 20 4 56, 62 56, 62 56, 62 5 77, 80 77, 80 6 51, 55 7 2, 4, 6, 7, 11, 59 Network researchers suggest that it is generally accepted that the closer together in geographic space two groups are, the more similar they are likely to be (Borgatti et al. 2013; de Nooy et al. 2011:346; McPherson et al. 2001; Watts and Strogatz 1998). In archaeological terms, this is interpreted as the axiom proposing that the closer two populations are, the more they will interact and ultimately the more similar their material culture will be (Hart 2012:128; Willey and Phillips 1958). Thus, geographical proximity should be a predictor of material culture similarity. Accordingly, and taking into account that most Iroquoian villages represent a single, short-term occupation lasting 10 to 20 years in general, a ceramic type similarity network graph should illustrate that settlement loci within a village will share the highest similarity values with those closest to them in space. The distance employed, or measured, should ideally reflect topography 180 and intervening obstacles, since these factors can greatly affect path length. Network analyses of the Keffer ceramic data on the effect of geographic distance (see Appendix H) concluded that physical location of the samples within the village were not the primary determinants of highlevel similarity ties. Social factors affecting ceramic change are elusive, however. Longhouse clusters, thought by Iroquoianists to represent matrilineally related social groups of formed pottery manufacturing bodies, or communities of practice, are not highly visible in the weak pattern of intra-cluster ties seen in the statistical evidence (Table 7-2). These results indicate that temporal and social factors not related to cluster origin were therefore likely factors in the network graph topology of the Keffer ceramics. Table 7-2. Average Tie Densities Within and Between Individual Clusters, the Combined Interior Samples, and the Combined Exterior Samples. Cluster A Cluster B Cluster C Cluster D Cluster A 167 161 165 175 Cluster B 161 167 164 163 Cluster C 165 164 166 170 Cluster D 175 163 170 181 Exterior 167 162 166 169 Interior 167 167 166 181 Interior/exterior difference 0 +5 0 +12 Percent difference Total difference 2.5% 14 Note: Boldface denotes intra-cluster density. 6 6% 6 18 181 Social Network Analysis of Local Tradition Ceramic Chronology In this section, I examine the chronological patterning of the network graphs to interpret the occupation sequence of the Keffer village. This results in a series of five village plans, labelled Stages 1 to 5. These results are then correlated with the settlement phases, Phases 1a to 2b-2, discussed in Chapter 6. This analysis permits the development of a related but alternate plan of village development through time, labelled second-level Stages 1–5. Interpretation of the network topologies dealing with intra-site relations is dependent on the establishment and use of a frame of reference (see Betts 2008), provided by the settlement phases illustrated in section 6.5. From this, two houses, Houses 15 and 19, have been assigned chronological positions as, respectively, the first and last longhouses constructed in the village. Their appearance as topologically opposed isolated pendants circumscribing most house and midden collections in the local tradition network indicates (Figure 7-1) chronology may be a fundamental component of the Gower and MDS layout graph structure. Therefore, the axis which lies between these two nodes, approximately horizontal in orientation, is assumed to loosely represent temporal change in the ceramic practice across this village. The resulting chronology is illustrated in a series of five stages (Figure 7-2). Figure 7-1. Local tradition ceramics Over 1 Vessel BR 170 MDS layout with BR 190 Blue Backbone structure. 182 Figure 7-2. Original Stages 1–5 as established through social network analysis. Application of Local Tradition Ceramics Social Network Analysis Stages to Settlement Phases The village occupation sequence is further refined through the integration of the SNA Stages 1–5 (Figure 7-2) and settlement pattern Phases 1a to 2b-2 (Figure 7-3). This is an essential step in the development of an occupation history of the village. Detailed analysis is provided in Appendix I. Phases indicate the viable physical coexistence of village features but lack social elements. SNA stages on the other hand lack the physical control of settlement studies. The incorporation of this physical control results in an amended, or second level, series of stages (Figure 7-4, Table 7-3) in which several houses and middens have been reallocated to alternate stages. Figure 7-3. Plan showing possible features of Phases 1a to 2b-2. 183 Figure 7-4. Second-level temporal Stages 1–5 as established through social network analysis and settlement studies. Table 7-3. Houses and Middens Included in Each Second-Level Temporal Stage Based on Social Network Analysis and Settlement Studies. Second-Level Stage Cluster A Cluster B Cluster C Cluster D Cluster A-B-D 1 M51, M57 M55, M67 H15a, M68, M78 M53 2 H3, H5, M51, M57 H8, M56, M55, M67 H12, H15b, M62, M75, M68, M78 M53 3 M60, H3, H5, M51, M57 H10, H8, M56, M55, M67 H13, M61–H14, M72, M73, M74, M80, H12, M62– H18a, M68, M75, M78 H20, M65, M77, M53 4 H1b, H2, M63 M60, H3, M51, M57 H9, H10, H8, M56, M55, M67 M59–H16, M66, H13, M61–H14, M72, M73, M74, M80, H12, H15b, M62–H18a, M68, M75, M78 H4, H11, M71, H20, M65, M77, M53 H6 5 H7, H1b, M60, H2, H3, M51, M57, M63 H9, H10, H8, M56, M55, M67 M59–H16, M66, H13, M61–H14, M72, M73, M74, M80, H12, H15b, M62–H18a, M68, M75, M78 H19, H4, M71, H20, M65, M77, M53 M52, M54, H6 Notes: Boldface indicates new additions in the stage, added at the beginning of the cell. M = Midden; H = House. 184 Summary The combined ceramic social network and settlement pattern analyses convey a picture of gradual growth across the Keffer village with the development, or introduction, of several small, but not dissimilar, ceramically linked residential groups. The initial occupation may have consisted of one residence, located on the plateau, and several small refuse deposits across the site. Whether the erection of Houses 3, 8, and 12 occurred concurrently or somewhat later is unclear because nearby middens, Middens 51, 53, 55, and 57, of Group 5, display slightly earlier positions on the temporal axis of each graph. Midden 62, as a proxy for House 18, appears in this early period, most likely indicating the coinciding erection of House 18a and the removal of House 15. House 18’s alignment with House 12 would also argue for their concurrent construction. The third stage is characterized by at least one large-scale, tightly knit ceramic group appearing within the village. On the plateau, this is seen in the ceramics of House 13, its associated Group 2 hilltop Midden 73, and hillside Middens 72 and 80, along with the northern neighbour, House 14, and its related large hillside midden, Midden 61. The high-level bond shared between House 13 and House 20 in Group 3 indicates that this community extended outside the village proper even before the expansion of the village palisade. The high-level integration of these groups across the site is seen in the membership of the large Midden 60 to the north and House 9 at the site’s core. The placement of Middens 52 and 54 in Stage 3, as shown in the network topology, may be due to social rather than chronological factors, as settlement studies show both middens to be of a later stage, as they overlie longhouses which appear in this stage. The expansion of the palisade in Stage 4 is mirrored by population growth in the northern part of the village, with four houses of shared ceramic characteristics constructed between pre-existing houses. Two houses, Houses 1 and 9, show evidence of long occupation, and these structures likely date to an earlier part of the sequence. House 4 is a new addition within the growing boundaries of the village. The last part of the sequence, Stage 5, is marked by the distinctive Group 4, here composed of Houses 2, 7, and 11. Archaeological features of House 2 show light occupation indicating the structure may have only experienced short-term use. Houses 7 and 11 are each characterized by expansions which also experience short-term occupation. The associated ceramics may derive from these later expansions. At the periphery of the network is House 19. In accordance with evidence from 185 settlement Phases 2b-1 and 2b-2, ceramics of this last residence may be characteristic of the last stages of communal ceramic production and use near the end of the village occupation, a profile not differentiated in earlier and/or longer occupations of other loci at the site. Interpretation of the network based on the temporal positioning of peripheral nodes, as indicated by settlement pattern analysis, appears to have been generally successful. Reliance on small samples is problematic but not insurmountable, provided the robustness of samples is not overlooked or ignored but, rather, taken into account during interpretation. Constructing a Community Through the use of the network approach, the gradual changes in ceramic practice can be traced. In the specific case of the Keffer village, this process may be correlated with time and has produced a timeline of ceramic practice transformation from the appearance of the first residence to that of the last residence constructed in the village. However, specific changes, resulting from in situ development or outside influences in the form of trade or immigration, are not accessible through network analysis, but might be inferred from the integration of local tradition ceramic typology into the settlement sequence. In-depth typological analysis should also aid in the evaluation of the validity, or soundness, of the suggested village development and the utility of a visual analysis of networks approach for small-scale intra-site exploration. In this analysis the placement of each house and midden within one of these five stages is correlated with the relative quantities of the local tradition vessels in their samples. Because the two main categories, decorated neck Huron and undecorated neck Huron, index chronological as well as social distinctions, the ratio of these two groups can be used to assess the placement of each house and midden within the stages. In this way, a more nuanced timeline is formed. As with the majority of Iroquoian ceramic analyses in this study, typology, following the work of MacNeish (1952), forms the basis of the classification. However, in this in-depth investigation into the factors influencing the topology of the network graphs, the local tradition ceramics have been allocated to more inclusive type categories, as explained in Chapter 4. This amalgamation of types aims to reduce the isolating effect produced by the division of generally similar ceramic vessels on the basis of minor differences. When each type is given equal weight in the analysis, types that separate according to slight variations, for example in collar motif, can complicate 186 analysis and have a significant influence on inter-sample comparison results. Therefore, local tradition ceramics have been allocated to one to the following categories, as explained in Chapter 4:  undecorated neck “Huron” (UNH)  decorated neck “Huron” (DNH)  “Neutral”  high collar  early Ontario Iroquoian types (Early) Readily discernible differences in assemblage characteristics are apparent when ceramic frequencies are investigated. Variations as they apply to the ceramics of each house and midden are examined in terms of the three of the five main local tradition ceramic type groups, as they appear within the suggested stages. These are:  decorated neck “Huron” (here restricted to Black Necked)  undecorated neck “Huron” (here consisting of all low collar “Huron” incised types)  “Neutral” (here restricted to Lawson Incised and Lawson Opposed) Primary sequential relationships based on local tradition ceramic type are advanced according to the temporally suggestive type categories, the possible influence on the proportions of undecorated neck Huron and decorated neck “Huron” (i.e., Black Necked). Ceramics in the village’s collections of ceramics from the west will be assessed by the relative presence of Neutral pottery (represented by two types, Lawson Incised and Lawson Opposed). These numbers vary significantly at locations across the site, from zero to 20 percent (Table 7-4), with an average of 6.5 percent. 187 Table 7-4. Proportion of Local Tradition Ceramic Types by Second-Level Stage, Based on Social Network Analysis and Settlement Studies. Cluster Sample Size SecondLevel Stage Locus UNH DNH Neutral High Collar Niagara & Ripley Early A 3 1 M51 60.3 24.1 10.3 1.7 3.4 0.0 A 7 1 M57 62.3 19.0 12.3 3.5 0.4 2.5 B 4 1 M55 60.9 24.8 9.3 0.6 0.6 3.7 B 1 1 M67 61.1 11.1 16.7 0.0 0.0 11.1 C 2 1 M78 45.5 18.2 27.3 9.1 0.0 0.0 C 2 1 M68 50.0 22.2 22.2 0.0 5.6 0.0 C 2 1 H15 55.0 40.0 0.0 0.0 5.0 0.0 D 3 1 M53 57.1 28.6 3.6 3.6 7.1 0.0 A 1 2 H3 66.7 0.0 16.7 16.7 0.0 0.0 B 1 2 H8 66.7 20.0 6.7 6.7 0.0 0.0 B 3 2 M56 66.7 25.6 2.6 5.1 0.0 0.0 C 3 2 M75 62.1 10.3 20.7 6.9 0.0 0.0 C 5 2 M62 67.5 23.8 5.3 1.3 2.0 0.0 C 4 2 H12 70.4 21.4 2.0 1.0 5.1 0.0 A 7 3 M60 73.6 15.7 7.2 2.0 0.6 0.8 B 1 3 H10 100.0 0.0 0.0 0.0 0.0 0.0 C 3 3 M73 68.9 15.6 6.7 4.4 2.2 2.2 C 7 3 M74 69.6 20.2 6.9 1.2 0.7 1.4 C 3 3 M72 69.9 19.3 9.6 1.2 0.0 0.0 C 4 3 H13 73.1 25.0 0.0 0.9 0.0 0.9 C 3 3 M80 75.3 15.1 5.5 2.7 1.4 0.0 C 6 3 M61 75.7 15.5 5.6 1.3 1.6 0.3 D 7 3 M65 67.0 19.9 8.7 1.4 1.5 1.5 188 D 5 3 H20 73.8 24.1 0.0 1.3 0.8 0.0 D 4 3 M77 75.2 17.7 4.4 1.8 0.9 0.0 A 1 4 H6 71.4 14.3 0.0 0.0 0.0 14.3 A 1 4 M63 72.7 18.2 0.0 9.1 0.0 0.0 A 3 4 H1 76.5 17.6 0.0 2.9 0.0 2.9 A 1 4 H2 87.5 12.5 0.0 0.0 0.0 0.0 B 3 4 H9 74.3 8.6 14.3 0.0 2.9 0.0 C 3 4 M59 76.3 15.0 2.5 0.0 6.3 0.0 C 3 4 M66 79.3 12.1 6.9 0.0 1.7 0.0 D 4 4 M71 79.3 16.3 4.3 0.0 0.0 0.0 D 3 4 H11 83.9 12.9 0.0 0.0 0.0 3.2 D 2 4 H4 84.2 15.8 0.0 0.0 0.0 0.0 A 2 5 H7 84.0 12.0 0.0 0.0 0.0 4.0 B 4 5 M54 74.8 18.5 5.9 0.0 0.8 0.0 D 4 5 M52 72.2 20.1 5.6 0.7 0.0 1.4 D 3 5 H19 91.7 6.3 2.1 0.0 0.0 0.0 Note: UNH = undecorated neck Huron, DNH = decorated neck Huron Legends for Table 7-4 Sample Size 1 2 3 4 5 6 7 Number of Vessels 6–19 22–29 30–49 50–99 100–299 300–499 500–999 Colour Local Tradition Ceramic Percentage Ranges 40–62% UNH/10–20% DNH/12–28% Neutral 65–70% UNH/15–25% DNH 50–60% UNH/25–40% DNH 70–80% UNH/12–25% DNH 80–90% UNH/6–12% DNH The presence of Neutral ceramics (represented by two types, Lawson Incised and Lawson Opposed) in occupation features across the site also ranges significantly, from zero to 20 percent, although the average is much lower, namely, 6.5 percent. This disparity in the presence of these distinctive ceramics is significant. It represents alternative influences on the communities of ceramic practice across the village, whether in terms of production, consumption, or both. There 189 appears to be no correlation between the proportions of decorated neck Huron and undecorated neck Huron vessels and Neutral vessels (Figure 7-5), suggesting that their presence is not a temporal marker but reflects some aspect of social relations. Decorated neck Huron vessels, represented by Black Necked–type vessels, on the other hands are generally present in larger proportions than the Neutral ceramics and are seen in widely divergent frequencies across the site, from 0 to 40 percent. Like all the ceramic classifications in this study, this range may be partially attributable to the great range in ceramic sample sizes across the site, as previously discussed. Taking this into account, these frequencies are employed with care so that inferences relating to the relationships of all household groups across the living village can be evaluated. A detailed explanation of this analysis is presented in Appendix J. 100.0 90.0 80.0 70.0 60.0 50.0 40.0 30.0 20.0 10.0 0.0 H15 M56 M55 H20 M68 M74 H8 M72M54 M63 H1 H4 M73 M80 H6 H2 H7 M75 H19 Undecorated Neck Huron Decorated Neck Huron Neutral Figure 7-5. Undecorated neck Huron, decorated neck Huron, and Neutral ceramics percentages sorted by decorated neck percentages, highest to lowest. The combination of settlement, network, and ceramic type category analyses has produced a comprehensive picture of the development of the Keffer village occupation history (Figure 7-6, Table 7-5). 190 Figure 7-6. Keffer development sequence Maps 1–8. Table 7-5. Settlement Changes in Maps 1–8. Time Palisades in Use Map 1 Houses Added or Expanded Houses Removed Middens in Use 15a 67, 68 51, 53, 55, 56, 57, 74, 75, 78 Map 2 1, 3 1a, 3a, 5, 7a, 8, 9a, 10a, 12, 15b Map 3 1, 3 3b, 9b, 10b, 13a, 14a/b, 18a, 20 Map 4 1, 3 11a Map 5 2a 1b, 2, 4, 6, 7a, 11b, 13b Map 6 2a 19 Map 7 2a 7b, 11c, 18c Map 8 2b 15 60, 61 (H14), 62 (H18), 65, 72, 73, 77, 80 3b, 5 59 (H16), 63, 66, 71, 52 6, 10, 12 7a/b, 19 54, 58; expansion of 74, 77 The site was first settled by the occupants of House 15a (Figure 7-6, Map 1). This house was placed in a prime location in the central portion of the riverside hilltop, or plateau, of the site. Its 191 diagonal placement across the plateau suggests that further occupation of the location may not have been planned at this time. The settlement may have originally served as a special purpose hunting or agricultural cabin within newly expanding fields in the area, similar in purpose to the occupation of the nearby Hidden Spring site. Alternately, House 15a may have been the home of an advance construction party established as a foundation for planned future movement into the location. This hypothesis is supported by the following expansion of the longhouse to 65 m, becoming the largest longhouse at the site. Large-scale development of the village occurs in concert with this expansion (Table 7-6, Map 2). On the plateau, creating the newly formed Cluster C, House 12 is constructed to the west of the newly expanded House 15b. This marks a population influx on this upper elevation. Occupation of the lower level, at the base of the plateau, also begins around this time, with the appearance of two small residential clusters. At the northern edge of the lower ground, three longhouses, Houses 1a, 3a, and 7a, are built, along with the structure labelled House 5, to the east of House 3a. These structures comprise Cluster A. The second low town cluster, Cluster B, is also composed of three longhouses, Houses 8, 9a, and 10a. House 10 is protected to the west by a fence line, Palisade 3, erected sometime during its life history. Construction of this initial village settlement coincides with the erection of the first village enclosure, Palisade 1 (Figure 76, Map 2). This first community at Keffer therefore appears to be composed of three social units as displayed through inter-longhouse proximity and clustering. The next major development is defined by a series of distinct, yet seemingly related, events occurring across the site (Figure 7-6, Map 3). In Cluster A, to the north, the small House 3a is extended to meet the edge of the round structure, House 5. Houses 9a and 10a in Cluster B, to the south, also experience expansion, possibly at the same time. Interestingly, House 9b shows clear evidence of total reconstruction with this expansion. This is not a common practice in the extension of Iroquoian longhouses on the north shore. Also occurring after this initial construction of Palisade 1 is a restructuring of the plateau settlement layout, as seen in Cluster C. Perhaps in order to make room for denser occupation, House 15b is removed at this point and three new longhouses are constructed in a new orientation, parallel with the alignment of the remaining house, House 12. House 13a is located alongside House 12 at the south end of the plateau. Houses 14a and 18a are similarly aligned directly to the north. The construction of House 20 to the south of the village palisade marks the arrival of the first household physically, 192 and therefore socially, excluded from the residents of the village proper, in the area labelled Cluster D (Figure 7-6, Map 3). The high level of similarity between the local tradition ceramics of Houses 20 and 13, however, suggests that a strong social relationship existed between these two households. The lower level of the Keffer village was situated in a small, valley-like depression, with hill slopes to both the east and west and somewhat lower ground present to the south. House 20 is built in the most accessible area of this lowland. After its construction, land for further building becomes limited. When a second longhouse, House 11a, was built outside the palisade, the narrow space remaining between the eastern palisade and the eastern hill slope provides the best location for this construction (Figure 7-6, Map 4). After the appearance of House 11a, Palisade 1 was removed and most of the surrounding lowlying land was encompassed within the growing village (Figure 7-6, Map 5). The erection of Palisade 2 was part of the large-scale growth and change across the site. The expansion of several longhouses and the addition of new houses mark a significant event of population growth. House 11a, which formerly intersected with the exterior of Palisade 1, was now expanded to the north and entirely rebuilt. A second longhouse, House 4, was added in the newly opened-up area of the village interior to the north of House 11b. The ratio of local tradition decorated versus undecorated neck Huron vessels shows Houses 4 and 11 to be very similar ceramically and to share a common heritage. Spatial limitations are apparent in the tight positioning of House 4 directly adjacent to the east wall of Palisade 2. A second new longhouse, House 6, was built between the end of House 7a in Cluster A and House 4. At this point, Houses 3a and 5 are removed and House 2 was built to replace them, between Houses 1 and 7. House 1 was expanded beyond the former palisade line and was entirely rebuilt. House 7 also appears to have been rebuilt at this time. The expansion of Houses 9b and 10b likely occurred simultaneously with the increase in growth across the community, although their expansion may have occurred earlier. The new settlement configuration in the lower level of the site reflects changing social dynamics. Clusters A and B were no longer spatially segregated, and Houses 4 and 11 are integrated into a larger collective area. Houses 1b, 2, 4, 7a, 8, 9b, 10b, and 11a are joined in Cluster A-B-D. House 20, in the southern end of Cluster D, may now have been more socially aligned with its nearest neighbour on the plateau, House 13b, which also experiences expansion and rebuilding in the same location with the relocation of the palisade. Houses 14a 193 and 18a, to the north, were expanded and rebuilt as well. To the east, a new structure, House 16a, was added on the only remaining hilltop land, a narrow projection of the plateau to the east, as House 12 was rebuilt in its original footprint, without structural expansion. Large-scale village renewal is most clearly seen in the relocation of the palisade. With this comes the erection of three new longhouses, Houses 4, 6, and 16. Four longhouses, Houses 1b, 13b, 14b, and 18b, are extended in the course of their complete reconstruction. House 20 was rebuilt but not extended. Houses 9b and 10b may also have been rebuilt and extended at this time, while Houses 3b and 5 were removed and replaced with House 2. This reconfiguration of the village layout was a significant event in the social reconstruction of the Keffer community. It reflects the immigration, acceptance, and inclusive integration of others, previously considered outsiders, into the social world of the Keffer community. The construction of House 19, outside and to the south of Palisade 2 marks the final period of population growth in the village. Sometime after the House 19 construction, a portion of Palisade 2, to the north, was removed and expanded to encompass House 19 within the village. Evidence of the short occupation of the recently built Houses 2, 6, and 19 suggests that the village was not occupied long term after the erection of Palisade 2. The removal of House 6 in Cluster A-B-D and of House 12 on the plateau appears to signal the first episodes of household groups relocating outside the village. The rebuilding and expansion of Houses 7b, 11c, 16b, and 18c, on the other hand, reflect growth within their respective households. The spread of refuse from Midden 74 over the south wall post line of House 12, in the meantime, shows that large-scale occupation on the plateau continued after the removal of House 12. The expansion of Midden 77 over the east wall of House 19 also suggests the continued occupation of House 20 after the removal of House 19 in the southern end of the site. To the north, the appearance of the surface deposit of Midden 58 over the central portion of House 10b marks the removal of House 10 around this time, while the spread of Midden 57 over the west wall of House 7b and the deposition of Midden 54 over its eastern end confirm the removal of House 7b prior to the end of the Keffer village occupation (Figure 7-6, Map 8). During the final habitation of Keffer, as seen through archaeological evidence, the village population decreases significantly from its height, as seen in Map 5. The removal of Houses 6 and 7 now isolates Houses 1 and 2 in Cluster A. Cluster B was somewhat re-formed to include 194 Houses 4, 8, 9b, and 11c. Occupation on the plateau continues in Houses 13b, 14b, 16b, and 18c. House 20 remains in the southern part of the low land. Whether Keffer was finally abandoned en masse or gradually, household by household, is unknown. The preceding section has presented a lifecycle history of the Keffer village. This was achieved through the integration of network analysis of ceramic types, local tradition ceramic proportions, and the physical layout of site structures as indicators of the temporal position of habitation, refuse, barrier, and palisade features. The ceramic evidence was integrated into the five stages derived from network analysis settlement data phases as a framework. It is clear that these proposed changes occurred over the course of the occupation of the village, but while the data allow for the sequence to be reconstructed, the absolute temporal distance between each alteration is inaccessible. Some changes may have been contemporaneous or may have occurred at a rapid pace, one after another, while other innovations may have required lengthy planning between episodes of change. Village Settlement Overview The life history of the Keffer community is unique, like that of other north shore Iroquoian villages whose settlement history has been studied in depth. At Keffer, the original settlement of one longhouse quickly expanded with the addition of three residential clusters. This first village was surrounded by a palisade. Sometime after expansion of lower-elevation houses, longhouse additions and cluster reorganization on the plateau mark a period of increased population. At this point, longhouses began to be added outside the village, leading to the eventual expansion of the palisade to include the new houses, while expansion and replacement of older houses lead to infilling of the village interior. A final house was added outside the palisade, which is then expanded once again, bringing the last household into the village. In the final stages of the village’s life, individual houses were removed as the village population gradually decreased and several households left the community, perhaps as an advance construction party for the next home of the village community. While neither settlement nor ceramic data can reveal the final habitation and abandonment of the village, ceramic links with other, nearby Iroquoian settlements may provide clues to its synchronic and diachronic relations with other communities in the area. 195 To access the agency of the people who built and lived in the village, it is necessary to delve more deeply into the subtleties of communities of practice at the household level. Households are the basic social units (see Creese 2010; Timmins 1997), and these combine into larger household clusters, thereby forming the next level of the social network. Social Network Analysis of Local Tradition Communities of Practice The dynamic structure of the graphs and the changing patterns of clusters and outliers apparent at different similarity levels and sample sizes inform on the nature of the ceramic relationships at Keffer. The Over 29 Vessels BR 190 (Figure 7-1) network provides the backbone structure of the closest relationships among local tradition ceramic samples. This framework is composed of a one small world formation, Group 2, composed of several highly similar ceramic midden collections from across all settlement clusters of the village. Members of this small world appear to date to the middle of the village occupation, although they have close ties to both earlier and later deposits. A second minor small world, Group 3, includes primarily deposits of the western side of high town. As tie levels are lowered and new constituents are added, shifting allegiances and new bonds emerge. The network fills in and relationships become more numerous and more complex across the village as a whole. Outliers encircling the core relationships materialize, reflecting diverse influences on local traditions. At the more inclusive end of the spectrum, the Over 1 Vessel BR 170 network, new ties hint at weaker relationships, providing supplementary information to be integrated into analysis after higher-level connections are taken into account. Detailed analysis of high similarity level, BR 190 and 180, graph topology, discussed in Appendix J, has produced a revised series of groups, believed to represent the local tradition ceramic communities of practice present throughout the village occupation (Figure 7-7 and Figure 7-8, Table 7-6). 196 Figure 7-7. Local tradition ceramics Over 1 Vessel BR 185 Gower layout. Temporal patterning, from earlier to later, is represented by the physical placement in the figure, left to right. 197 Figure 7-8. Local tradition ceramics Over 1 Vessel BR 171 Gower layout Table 7-6. Houses and Middens Associated with Each Local Tradition Community of Practice. Local Tradition Community of Practice Houses 1 13, 20 2 1, 9, 10 54, 61, 77, 80, 52, 60, 71, 66, 59,63 3 8, 12 72, 74, 73, 65, 56, 62 4 2, 4, 6, 7, 11, 19 5 3, 15 Middens 51, 53, 55, 57, 67, 68, 75, 78 Note: Boldface indicates initial high-level core nodes of communities. Order indicates approximate strength of ties. 198 These groups represent social units that produced or consumed ceramics as a collective. The divisions are to some extent indicative of chronological patterning in the graph and reflect social ties across time within the village. The five communities of practice (CoP) are numbered according to the relative strength of their highest ties. That is, the members of community of practice 1 (CoP 1) are more similar than those of CoP 2, and so on. The strongest ties within each group are indicated in boldface type and are located within the list next to those with which the group shares this high-level tie. In the last group, CoP 5, lower similarity level relations produce a less cohesive and more loosely tied community. Discussion Social network analysis of local tradition ceramics has resulted in a chronology of ceramic practice at the Keffer village. This pattern has been interpreted along two separate but linked pathways. A nuanced sequence of the village occupation was produced through the incorporation of this chronology with settlement pattern information provided in chapter 6 and an examination of the ceramic typological characteristics of the settlement features. The interpretation of high levels of similarity among the ceramics of house and midden collections, as demonstrated in graph topology, has further defined a series of social related ceramic communities of practice. Closer examination of these communities of practice and their significance in the development of the village community occurs in Chapter 8. 8 Communities of Practice at Keffer The Emergence of Keffer’s Ceramic Practices This study looks at the household as the minimal group size in which identity is reflected in ceramic production. Investigation at the level of intra-longhouse divisions is not possible with the samples available from Keffer even though it has one of the largest ceramic collections in Iroquoia. Most longhouses on the site are characterised by relatively small sample sizes. Three houses contain only one vessel each. Most have samples in the low end of the range from 6 to 131 vessels. Only House 20, which was excavated by hand, with 261 vessels, contains an adequate sample. In addition to the paucity of interior features at Keffer most longhouses experienced taphonomic processes and/or inconsistent excavation resulting in minimal or uneven samples across the length of the structure. Network analysis was performed on three matrices of escalating ceramic sample sizes. Distinctive local tradition practice groups are evident, or “hinted at,” in the network graphs at every level of similarity examined, from BR 190 down to BR 170, where all nodes join the network. Conclusions are compiled from the foundation of the highest level and most exclusive relationships, as identified in the network backbone (Figure 8-1), progressively changing community configurations as lower-level results are applied. 199 200 Figure 8-1. Local tradition ceramics Over 29 Vessels BR 190 Gower layout. Change in ceramics assemblages are now accepted as resulting from gradual transformation in the accepted practices within communities of practice, the producers and consumers of pottery, often as a result of changing social relations (Crown 2014; Dorland 2018; Striker et al. 2016). Network graphs of the Keffer ceramic assemblages clearly demonstrate that ceramic practice changes through time. The network graphs of the Keffer ceramics reveal that the temporal axis constitutes one of the main components affecting node layout. Incorporation of this hypothesis is used above in the development of an occupation sequence based on the network analysis results at Keffer. Integrating the temporal axis, which in this case (Figure 8-2) appears to run approximately horizontally across the graph, traces development from the earliest ceramics on the left to the latest on the right. However, the vertical span of the network confirms that time is not the only variable affecting the graph topology. But the strength of network analysis lies in its portrayal of patterns in data structure. The approach does not provide causes of the resulting structural patterning. Therefore, other factors influencing the topological location of nodes must be determined elsewhere. In this case, in-depth ceramic typological analysis will be employed 201 (see Appendix I). This section aims to identify groups of nodes representing similar local ceramic systems. These node clusters are interpreted as proxies illustrating multiscalar differences in ceramic communities of practice. As nodes are constantly in motion across the graph structure with changing similarity levels, the ultimate goal of the analysis is to uncover the closest ties of each node as illustrated through colour coding. At the BR 195 level several separate components are present in the network. As nodes appear and connect with existing components with decreasing similarity, they are coloured to match the first node with which they connect. This process resulted in five discrete ceramic communities. At the very high BR190 level (Figure 8-1), close ties emerge between some loci. By BR 185 (Figure 8-2), I identified five separate and widely isolated components. 202 Figure 8-2. Local tradition ceramics Over 1 Vessel BR 185 Gower layout. Temporal patterning, from earlier to later, is represented by the physical placement in the figure, left to right. Colours indicate community of practice group. By BR 171 (Figure 8-3) each node was assigned to one of the five main groups. The BR 185 (Figure 8-2) graph structure shows a clear bifurcation of the practice community which appears to be temporal in nature. The main component is divided into two large groups, Group 2, and 3, the nodes of which fit into a narrow column on the horizontal temporal axis. The intervening gap is bridged by the Group 1 pair, Houses 13 and 20, and a more distant pair comprised of Midden 52 and the larger Midden 60. Group 3, located to the left, is interpreted as the first of the two large groups. This group is composed of features located on the western half of the high town plateau, Midden 62, and its affiliated House 18, and Midden 74, which is a product of House 12 and likely House 13. Interestingly Midden 73, first connecting with Group 2, shares many strong 203 ties with Group 3, and by the BR 185 similarity level joins it. Midden 73 likely results from the occupation of House 13. House 8 and Midden 56, both located in the central portion of the low town, are also members of Group 2. In the early period of the Keffer village, Midden 56 lay towards the limits of the village interior near the wall of Palisade 1. This refuse was likely a result of deposition from the nearby House 9, and possibly neighbouring House 8. Figure 8-3. Local tradition ceramics Over 1 Vessel BR 171 Gower layout. Connecting the earlier Group 5 community to Group 2 is the village’s largest midden, Midden 65. This deposit is located outside the first palisade downhill from the early occupation of the plateau, Cluster C. Its massive assemblage of 966 vessels argues for a long period of use and therefore it may have been one of the first large-scale dumps for the residents of southern section 204 of Cluster C. House 12, a fairly long structure near the edge of the western slope, may have blocked access to this hillside for occupants of the extended House 15b. Midden 52, which overlies Palisade 1, originates in the newly combined Cluster A-B-D area. Though linking closely with Midden 74 Midden 52 is also heavily tied with Group 2. Like Midden 52, Midden 60 shares close relationships with both Groups 2 and 3. Although its topological position suggests Midden 60 dates slightly earlier than Group 2, at low similarity levels. Midden 60 is pulled by the new nodes of Houses 6 and 9 into Group 2 itself. Midden 60 lies at the western hillside end of Houses 1, 2, and 3, and somewhat north of House 7 in Cluster A. There is little doubt the majority of its contents are the product of these residences. Although Midden 60 ties closely with Group 2 there are no ceramic samples in Cluster A closely connected with this group. This indicates that there was a clear separation, at least in the earliest days of occupation, between the northern low-town inhabitants and those of the west side of the plateau, while there was some joint practice between the Cluster B Houses 8 and 9 and these Group 3 practitioners. The very strong connection between House 13, located on the eastern side of the high town, and House 20, outside the first palisade, reflects the strong similarity in ceramic practices of Groups 2 and 3. Originally isolated between the two large components, House 20 drifts towards the western Group 2. House 13 becomes entirely enfolded into Group 3 as similarity levels decrease. At BR 171 the two houses and their practice communities are clearly separated. Group 2, towards the east, is dominated by the strongly tied small world ceramics of House 10, the nearby late Midden 54, and Midden 61, which represents the occupants of house 14, and Midden 80, one of the main repositories of House 13. Midden 59, most closely associated with the eastern hilltop House 16, is slightly outside of this core but links with House 13 and the slightly isolated Cluster A House 1, which itself links with the adjacent Midden 63. Midden 77, a late deposit of House 20, and possibly the later House 19, may bring the House 20 community into Group 3. The origins of the refuse in Midden 66 are uncertain. Despite its location adjacent to the Group 3 western residences Houses 12 and 18 the ceramic material relates more closely with that of the eastern Houses. Midden 66 may have been accessed by residents of Houses 14 and 16 sometime before the expansion of House 18c and/or after the removal of House 12. This would explain its 205 membership in the eastern Group 2 community. Midden 71 is a later deposit from House 20. Adjacent Midden 77, located between Houses 19 and 20, may have been abandoned during the erection and occupation of House 19. Deposits from this midden would therefore by-products of the last residents of Houses 19 and 20. For this reason, Midden 71 links directly with the latest practice community Group 4, a pendant stretching to the right, later, side of the network. Some Group 4 samples are quite small, but several contain enough material to provide a reasonable picture of the local traditions employed. This pendant connecting with the Group 2 core is characterized by sample of descending similarity which may be almost entirely temporally based. It is comprised of the last houses to be constructed at Keffer, as seen in settlement studies. Each of these structures was either newly built, House 4, 6 and 19, outside or across the original palisade, constructed after the removal of an earlier building; House 2 was erected after the dismantling of House 3, contracted and possibly rebuilt as seen in House 7, or reconstructed and expanded as seen in the parallel wall post lines of House 11. The earliest community of ceramic practice seen at Keffer is distinct from those of the later and more widespread occupation. This group, Group 5 is internally diverse and characterized topologically by three pendants emerging off the large 522 vessel Midden 57. This midden’s strongest connection is with Groups 2 and 3 via Midden 65, an early midden with a long period of use. The tie between these two middens may originate in their similar early date at opposite locations of the site while intra-village travel was wide open. Midden 57’s long period of use may result in a later date for this collection than any of the other smaller Group 5 deposits. Middens 53 and 55, surface deposits, are both located in the central low town area in Clusters B and D. Midden 51, also a surface midden, is found north of House 1 and may have been deposited prior to the House 16 expansion. The early House 3 structure is central to all of these middens as well as hillside Midden 57. House 15, in the prime location on the plateau, most likely created the four small middens surrounding it on the plateau slopes, Middens 68 and 75 to the south and 78 to the north as well as the small deposit, Midden 67 beside Palisade 3. Although Houses 3 and 15 are both depicted as primary occupations of the low town and high town, respectively, ceramics of these Houses 3 and 15 do not share much in common. This may be due to the very small samples of these houses. Despite this, House 3 shares some similarity with the House 15 middens being more similar to those of Middens 67 and 75. All of these samples, Houses 3 and 15, and Middens 67, 68, 75 and 78, include less than 206 32 vessels, some much less. Group 5 samples with larger collections, Middens 51, 53, 55 and 57, are marginally closer to the main component. The Keffer village local tradition ceramic collections portray a large amount of intra-village unity with all collections connecting to the main network component by the 85.5 percent (BR 171) similarity level. Network investigations at higher levels suggest greater unity among specific nodes forming slightly differing, separate practice communities. As illustrated in graphs at the BR 185 and 171 levels (Figure 8-2 and Figure 8-3), time is a factor in the separation of these practices. It is also apparent that other factors are involved in the formation of the graph topology. Ceramic profiles, in terms of the local tradition categories, of the five communities of practice are incorporated in the examination of the divisions and connections seen in the graph structures. Although network analysis was performed at the level of individual types, summarizing samples through the individual analysis of 20 types would result in a lack of clarity. Local tradition categories provide a more concise illustration of results (Table 8-1). Table 8-1. Local Tradition Ceramic Percentage Ranges by Community of Practice. Local Tradition Community of Practice Undecorated Neck Huron (%) Decorated Neck Huron (%) Neutral (%) 1 74 25 0 2 72–79 12–20 3–14 3 67–70 16–25 2–10 4 71–92 6–16 0–2 5 (all vessels) 46–67 10–46 4–27 5 (>29 57–62 10–29 4–21 vessels) Note: In this and similar tables, the colour coding in the headers indicates the three main local ceramic categories. Practice communities 1 to 5 were originally labelled according to the strength of their internal ties. Divergence among the core of each of the groups is clearly visible at the BR 185 level (Figure 8-2) but small sample size exaggerates this difference in the pendants of communities 2 (green) and 4 (yellow). At the lower BR 171 measure, these pendants are absorbed into the communities in the main component. Those nodes in the community 5 pendant are also now more integrated, but the diversity in the smaller samples produce three new pendants at BR 171. 207 Table 8-2. Community of Practice by Local Tradition Ceramic Percentage Ranges. Local Tradition Community of Practice Undecorated Neck Huron (%) Decorated Neck Huron (%) Neutral (%) 5 (all vessels) 5 (>29 vessels) 46–67 10–46 4–27 57–62 10–29 4–21 3 67–70 16–25 2–10 1 74 25 0 2 72–79 12–20 3–14 4 71–92 6–16 0–2 Here the communities of practice (Table 8-2) have been reordered following their temporal positions. As previously mentioned, the percentage of decorated neck vessels generally decrease with time, but absolute quantities must be compared to those within the watershed communities, as these may vary significantly. Undecorated neck Huron vessels are believed to have increased with time within central Ontario Iroquoian collections. These two hypotheses form the premise of the ceramic chronology. Temporal characteristics of Neutral ceramics are not so clearly understood, although observations made in the course of examining hundreds of site ceramic profiles in this study suggest that they appear to decrease in time east of the Credit River. Neutral ceramic quantities are not considered highly significant temporal factors in the Keffer samples, however. Neutral ceramics may be a stronger indication of inter-regional interaction. The absence of Neutral ceramics in community 1 separates these nodes from other communities at the BR 190 level, this but has less effect as more nodes exert pull on the House 13 node at BR 171. Allocation of the house and midden ceramics to individual communities of practice was originally accomplished through visual topological analysis. Application of ceramic profile percentages to the communities indicates that the undecorated neck Huron ceramics are a major factor in the separation of these groups, as they comprise between 57 and 92 percent of the Over 29 Vessels samples. Neutral and neck decorated vessels are characterized by much larger percentage ranges within most communities of practice, signalling less influence on node allocation. The second-largest group, decorated neck Huron ceramics, varies from 6 to 29 percent. Several Keffer samples contain no Neutral vessels, but in some instances these ceramics constitute up to 27 percent. 208 The spatial distribution of the local tradition practice communities (Figure 8-4) indicates quite diverse patterns in the history of ceramic usage across the village, particularly between the high town and low town residences. Figure 8-4. Local tradition communities of practice. 209 Chronological Development of the Communities of Practice Changes in the social relationships among the village households throughout the course of the village occupation emerge through the integration of communities of practice (Table 8-2 and Table 8-3) and the Keffer development sequence, as illustrated in Maps 1 to 8 (Figure 8-5). Figure 8-5. Keffer development sequence Maps 1–8. Table 8-3. Loci by Map and Local Community of Practice. Map 1 2 2 2 2 3 3 3 4 5 5 8 Local Tradition Community of Practice 5 2 3 4 5 1 2 3 4 2 4 2 Loci 15, 67, 68 1, 9, 10 8, 12, 56, 74 7 3, 51, 53, 55, 57, 75, 78 13, 20 60, 61, 77, 80 62, 65, 72, 74 11 52, 59, 63, 66, 71 2, 4, 6, 19 54 210 The first occupation of the Keffer village is composed of the small House 15a, placed in a central location on the plateau, and its two small middens, Middens 67 and 68 (Figure 8-6) located at either end of the longhouse. Ceramics of this first settlement belong to community of practice (CoP) 5 and are generally characterized by the lowest level of UNH (undecorated neck Huron) vessels, below 65 percent of the sample, and almost equal proportions of DNH (decorated neck Huron) and Neutral ceramics (Table 8-2). Figure 8-6. Map 1, with community of practice 5. The expansion of House 15b is indicated by the deposition of CoP 5 Middens 75 and 78 to the south of the extension (Figure 8-7). Following, or perhaps contemporary to, the House 15b expansion, ceramic practices change as the population of the village increases. House 12, with 211 the adjoining Midden 74, both members of CoP 3, with higher levels of UNH and lower levels of Neutral ceramics, appear on the plateau to the south of House 15b. At the base of the hill slope, Houses 1a, 3a, and 7a form cluster A to the north. House 3a and nearby Midden 57 and Midden 51 to the north of House 1 belong to Community 5, supporting their early appearance. House 1 shares CoP 3 with House 12. The small CoP 4 sample of House 7 has high levels of UNH characteristic of later ceramic samples. A second residential group, Cluster B, consisting of Houses 8, 9a, and 10 and Middens 53, 55, and 56, also occurs at this time. Middens 53 and 55, located at either end of House 8, belong to CoP 5 and may be early deposits from these Cluster B houses. House 8 and Midden 56 belong to CoP 3, like House 12 on the plateau. Houses 9 and 10, along with House 1, are the first members of CoP 2. Figure 8-7. Map 2, with communities of practice 2, 3, 4, and 5. 212 Continued growth at Keffer is seen in the construction of three additional longhouses on the plateau. House 14a’s construction is signalled by the appearance of its associated CoP 2 Midden 61 on the northern slope (Figure 8-8). House 18a beside it is represented by the new CoP 3 Midden 62. The appearance of CoP 3 Midden 72 on the south slope and Midden 73 directly east of House 13 may indicate that the first construction of House 13a, built following the removal of House 15b, also belonged to the increasingly common CoP 3. However, the inclusion of the House 13 ceramics in CoP 1, together with those of House 20, newly constructed outside the village palisade, may indicate changing directions in regional interaction, mirroring those of the new arrivals. No Neutral ceramics are present in this new practice community. The presence of Midden 77, adjacent to House 20, and Midden 80, to the south of House 13, both of which belong to CoP 2 (the strong small world cluster seen in the graphs), may represent deposits from specific, and perhaps unique, families within these houses. The large size of surface Midden 65, located in a slight depression at the western edge of the lower level, argues for long-term use as a refuse area by occupants of the village’s south end. The inclusion of this midden in CoP 3 may reflect the increasing population in this area over time. Population growth occurs in the northern section of the village, in Clusters A and B, during this same period. Expansion of Houses 3b, and continued heavy occupation of House 1a, may account for the spread of refuse deposition north of Midden 57 into the new CoP 2 Midden 60. 213 Figure 8-8. Map 3, with communities of practice 1–5. The ceramics of House 11, which settlement analysis indicates was the next structure built, belong to CoP 4. Although the first iteration of the longhouse was erected before the demolition of Palisade 1, the ceramics reflect the long-term occupation of House 11a, b, and c (Figure 8-9). 214 Figure 8-9. Map 4, with communities of practice 1–5. The expansion of the interior village area with the construction of Palisade 2 is paralleled by further population expansion (Figure 8-10). The new House 16, as seen in the appearance of Midden 59 and perhaps Midden 66, shares CoP 3 with House 14, beside it on the plateau. The removal of Palisade1 and House 3 opened up the north end of the village, allowing the expansion of House 1b, whose new inhabitants created the nearby surface Midden 63, also part of CoP 2. Midden 52, to the south, also belongs to this community of practice. It was begun after the removal of the first palisade. The expansion of the palisade to the bottom of the eastern hill slope also allowed for the construction of new longhouses, House 2, 4, and 6, and the expansion of House 11. All of these houses belong to the latest ceramic group, CoP 4. The late date of these ceramics is marked by a sharp increase in the presence of UNH, a decrease in DNH, and the 215 almost total absence of Neutral vessels. These later ceramic collections do not appear anywhere on the plateau. Figure 8-10. Map 5, with communities of practice 1–5. A new residence, House 19, belonging to the late CoP 4, is the last structure erected at Keffer and marks the end of its major growth period. This house, although sharing ceramic practice with those houses most recently constructed on the low ground, was built outside of the village proper (Figure 8-11). The acceptance of this household into the village community, which may be partially related to its membership in CoP 4, resulted in the final configuration of the settlement’s largest occupation. At this point, members of this later practice community occupied many of the 216 low-town structures, and Houses 7b and 11c expanded to their greatest extent. Due to their great length, they, along with House 4, held a large percentage of the village population and their community of practice grew. Figure 8-11. Map 6, with communities of practice 1–5. The dismantling of House 12, on the western plateau, may be a sign of, or a consequence of, this change. However, as House 12, an extremely long-lived structure with evidence of repairs but no conclusive proof of rebuilding, was removed, perhaps due to it poor condition, House 18c expanded over a large area of the plateau (Figure 8-12). Shared practice between Houses 12 and 18 suggests that the members of the House 12 household may have moved into the greatly expanded House 18c. Continuation of this community of practice on the west side of the plateau is seen in the expansion of Midden74 over the remains of House 12. In Cluster A-B-D, on the northern lowland, the extension of House 7b may similarly be the result of the acceptance of House 6, which was removed and covered by House 7b. They also share a community of practice, CoP 4. 217 Figure 8-12. Map 7, with communities of practice 1–5. The removal of House 12, on the plateau, was followed by the elimination of several houses on the lowland (Figure 8-13). To the west of House 9, a small deposit, Midden 58, marks the removal of House 10. This ceramic sample was not suitable for examination due to adverse taphonomic processes. Midden 54, a large surface midden, covers the area where both House 6 and 7b had been removed. This midden shares CoP 2 with the extended Houses 1b and 9b. The latest structure added to the village, House 19, was also removed at this time, as the large-scale expansion over its east wall indicates. The deposits of these three surface middens over the previously existing houses signify the last evidence for occupation of the Keffer village. 218 Figure 8-13. Map 8, with communities of practice 1–5. The first residents at the Keffer village brought with them a ceramic practice characterized by relatively low amounts of UNH ceramics and high proportions of DNH pottery. These collections also contain Neutral ceramics in varying proportions, from 4 to 21 percent. As the village expanded, two new practice groups emerged. Ceramics of those households on the west side of the plateau, in Houses 12 and 18, and those of House 8, on the lowland, appear slightly earlier than those of the east, Houses 13, 14, and 16. These ceramics mirror the ceramic practice of lowland Houses 1, 9, and 10, which were built early in the village’s construction. Houses 13, on the plateau, and House 20, at its base, may have been constructed during this episode of expansion, but the absence of Neutral ceramics suggests that their ceramic origins may have differed slightly from those of the other residences. With the removal of the first palisade and 219 expansion of the village, a later ceramic practice, with very high levels of UNH pottery, appeared with the influx of new residents in the last, short-lived houses. Houses, 4, 6, and 19 and the newly extended House 11c are squeezed into the only available land. The growth of House 7b only occurs after the removal of House 6, both characterized by these late ceramics. Gradual and large-scale transformation in the communities of local tradition ceramic practice during the growth, expansion, and early stages of abandonment at Keffer indicate a long period of occupation at the village, characterized by constant movement of people into, within, and out of the village. High dwelling concentration and ubiquitous longhouse expansion and refurbishment support these conclusions. External Relations of Keffer’s Communities of Practice: A Social Network Analysis The Keffer ceramic assemblage is typical of north shore Iroquoian ceramics of the fifteenth century. As part of the Don River drainage occupation (Don River nodes are yellow), it shares a high level of similarity with other central north shore communities of the period and is part of the highly connected Ontario Iroquoian tradition (Figure 8-14). The local tradition ceramics of Keffer, like those of other non–St. Lawrence Ontario sites, are distinct from those of the New York Iroquois. The search for strong influences on the Keffer ceramic practice communities is therefore confined to local Ontario sites. Figure 8-14. Pan-Iroquoian network BR 93 spring layout. 220 Overview Earlier network analysis of local tradition ceramics in this study was employed to postulate differing regional-level practices and intra-village practice communities along both the temporal and social scales. Here, connections among the Keffer loci collections and other Ontario Iroquoian communities are investigated using a single similarity matrix comprised of all Keffer house and midden samples and those of 140 additional sites. Due to the large number and close similarity of the intra-site Keffer nodes, the resulting network graphs of over 200 nodes present tight hairball clusters of the Toronto area sites. Formerly, Ontario ceramic site collections have commonly been recorded as single sample entities. It is now acknowledged that many settlements are not characterized by a single community of practice (Ramsden 2009, 2016; Williamson and Powis 1998). In this work, the Keffer house and midden local tradition ceramic database is combined with that of other central Ontario Iroquoian settlements to demonstrate ties among the sub-village-loci–level practices at Keffer and those of associated communities. I will propose possible communities of origin for the original House 15a settlers and those of subsequent newcomers to the village. The resulting network graphs portray high-level similarities and display possible relations among these ceramic communities (Figure 8-15). The location of north shore sites is illustrated in Figure 8-16. Figure 8-15. 140 Ontario Iroquoian sites, including Coulter and Kirche, and Keffer loci local tradition ceramics BR 120 spring loaded. 221 Figure 8-16. Locations of selected ancestral Wendat sites along the north shore of Lake Ontario. (Williamson 2014:Figure 2) Both social and temporal factors influence differing praxis within village ceramic communities. At Keffer, the existence of these dynamics is manifest in the village palisade expansion and placement of houses exterior to both the first and second palisades. As such, it is evident, as seen in settlement development studies, that Keffer was home to more than one incoming social community. At high similarity measures, specific links between Keffer loci and other sites become evident. Again, sample size must be taken into account in interpreting these results. This is clearly illustrated by in the intra-village Keffer results as nodes of small samples sizes, generally below 20 vessels, in Houses 2, 3 6, and 10 and Middens 67 and78 are distributed along the peripheries and are loosely connected to other sites (Figure 8-17). 222 Figure 8-17. Selected Ontario Iroquoian sites, including Coulter and Kirche, and Keffer loci local tradition ceramics BR 150 spring loaded. Though network graphs of the Keffer ceramics portray a village with a high level of internal ceramic overall similarity (Figure 8-17), characterized by an average network density of BR 173.8 (in the Over 29 Vessels samples), the variance in the values of highest edge tie per node (a range of BR 171–198) illustrates a level of heterogeneous praxis, perhaps attributable in some part to distinct origins of the local ceramic practices. This topological distribution is evident even at the low BR 120 similarity level as internal village nodes intersect with those of diverse geographic regions and chronology, connecting earlier t2; contemporary t3; and later t4, t5, and (Figure 8-15) from as far east as the Trent valley to as far north as Wendake, in Simcoe County. The interlinked component of the Ontario sites and the 38 Keffer nodes provides a very good overview of Keffer’s position in the province-wide Iroquoian network. For illustration and analysis purposes, however, extraneous nodes, which are topologically distant from the core area of interest surrounding the Keffer nodes, and their connections have been removed. This includes all t1 and many t2 ceramic collections that are peripheral to the fifteenth-century Keffer ceramic interaction sphere. Several sites in the Don watershed—the t2 Baker, Boyle-Atkinson, McGaw, Orion-Murphy-Goulding, and Over sites—as well as several in the neighbouring Humber valley—including the t2 Downsview, t3 Black Creek and Parsons, t4 Seed-Barker, and t5 223 McKenzie-Woodbridge sites—do not connect closely with this core and are therefore absent from the resulting higher-level graphs (Figure 8-18). Figure 8-18. Selected Ontario Iroquoian sites, not including Coulter and Kirche, and Keffer loci local tradition ceramics BR 170 spring loaded. The network structure begins to take form at the BR 170 similarity level as four regional sites connect to the mesh of Keffer house and midden nodes, which are already highly connected. Nearby settlements on the Don and Humber rivers share the largest number of high-level similarity ties with Keffer houses and middens, particularly those features with larger sample sizes. These sites include the t2 Don valley Risebrough site, located to the south; the nearby t3 Don valley Jarrett-Lahmer village; the t3 Humber valley Damiani site; and the t3 Logan site, situated at the headwaters of the Humber River on the Oak Ridges Moraine. In the spring layout graph at this level, the position of the Logan site between early and mid-stage development nodes within the Keffer core communicates the close association of Logan’s ceramics with those of Keffer’s first occupants. The later Damiani site is located on the fringes of the component, while the Don valley t2 Risebrough and t3 Jarrett-Lahmer sites connect to two separate nodes via pendants. The Damiani site shows more similarity with the overall ceramics of Keffer than do either of the Don valley sites. The tightly clustered nature, or hairball topology, of the Keffer nodes that occurs when the similarity levels are lowered to a point where external ties occur, makes the decipherment of the 224 individual ties difficult or next to impossible. Ego network plots of the two Humber valley sites, Logan and Damiani, demonstrate these relationships more clearly (Figure 8-19 and Figure 8-20). Figure 8-19. Local tradition ceramics ego Logan BR 170. Figure 8-20. Local tradition ceramics ego Damiani BR 170. At BR 170, the ego plot of the Logan node (Figure 8-19) uncovers several strong ties with House 12 and Middens 52, 60, 71, 74, and 77. In the larger component graph, Logan is invisible, though it is present in a central position, behind Midden 74 (Figure 8-21). This would suggest that the Logan site has deep-rooted ties to many components of the Keffer network. The BR 170 Damiani ego network (Figure 8-20) is smaller and does not connect with any of the nodes of the Logan network, indicating differing communities of ceramic practice. The nature of this difference, whether social, spatial, or temporal, is not provided by network analysis. Included in this small component are House 9 from the low town and Middens 61 and 73, thought to be products of high town end-to-end Houses 14 and 13, respectively. 225 Figure 8-21. Connected Ontario sites, including Coulter and Kirche, BR 170 Gower layout main component. The t2 Risebrough site, located approximately 5 km south of Keffer, connects at this level, BR 170, with the early low town Midden 57, while the later and geographically closer t3 JarrettLahmer site ties in a similar pendant to the large Midden 65. House 15, the earliest structure at Keffer, connects at this high level with the nearby t3 McNair site. It is only through a bridge created by connections between the Trent t4 Coulter site and Midden 56 that this component connects to the main structure of the graph. In the absence of these later intermediary sites (Figure 8-18), House 15 and the other members of this isolated group, the Don t2 Walkington 2, t2 Watford, t3 Hidden Spring, and t3 McNair sites, remain unattached. Interestingly, the Rouge–Duffins t3 Spang site is highly connected with House 15. This is undoubtedly due to the high proportions of DNH vessels present in the Rouge–Duffins tradition. For this reason, the t2 Best site is also part of this structure but does not connect directly with House 15. Neither Spang nor any other Rouge–Duffins site share other high-level, at or above BR 150, ties with any of the Keffer loci (Figure 6-22). 226 The Kirche and Coulter nodes tie elements of the main component to the earlier House 15 and the early Don and Rouge–Duffins community sites. Many features of the early settlement at Keffer—including House 15 and Middens 51, 53, and 55 and those of slightly later occupation of the plateau, Houses 12 and 13 and Middens 62—also tie with the t4 Trent Kirche community at the BR160 measure. The Coulter and Kirche nodes do not, however, connect these Keffer loci nodes with nodes of other sites, even as similarity decreases to BR 160. For this reason, graphs both with and without the Coulter and Kirche pair, will be used only to illustrate inter-site Keffer ceramic relations. Figure 8-22. Selected Ontario Iroquoian sites, not including Coulter and Kirche, with Keffer loci local tradition ceramics BR 160 spring loaded. 227 The large increase in network ties that results from lowering the similarity level to BR 160 results in a dense maze of nodes within the central component of the network (Figure 8-22 and Figure 8-23). Figure 8-23. Local tradition ceramics ego Damiani BR 160. Figure 8-24. Local tradition ceramics ego Logan BR 160. Because Damiani carries several additional ties with Keffer loci at this level, it becomes partially integrated into the dense mesh of the graph (Figure 8-25). An expanded Damiani ego network at the BR 160 level (Figure 8-24) now includes Houses 2, 7, 9, and 19 and Middens 54, 59, 60, 61, 66, and 73. Damiani also connects with the Robitaille site, in Wendake. The more deeply embedded Logan ego (Figure 8-25) network displays ties with mid-life Houses 12 and 13 and 228 Middens 51, 52, 54, 57, 60, 62, 66, 71, 72, 74, and 80. At this level, the upper Humber Logan and Damiani nodes share ties with Middens 54, 60, and 66. The Don sites, meanwhile, are generally linked to the earliest houses and middens. Risebrough and Watford both tie with Middens 51, 55, and 68. Risebrough is also connected with Midden 75, affiliated with House 15b. Watford and Hidden Spring, a site suggested to be a satellite of Keffer (ASI 2010b), and shown by Spence and Rainey (2017) to be socially related, are linked to House 15. In addition, the t2 Don Hope site joins the House 15 component at this level. The Watford collection now connects with other Keffer midden samples, specifically, Middens 57, 72, and 74. Figure 8-25. Connected Ontario Iroquoian sites and Keffer loci local tradition ceramics BR 150 spring loaded. Most of the Keffer nodes have numerous ties as the similarity level declines to BR 150. Only a few small-sample nodes remain on the periphery, with limited ties. House 3, with 6 vessels, lingers as the only Keffer node tied to the main component through an external site, the t3 Credit valley Emmerson Springs site. Midden 56, a deposit for House 9 and possibly House 11, sees its first exterior connection with Damiani at this point. It is also at this level that a second tie to the Rouge–Duffins drainage appears when the ceramics of Midden 59, associated with House 16, 229 link with those of the later t4 Mantle collection. House 11 also gains its first external ties, with t4 Kirche. Small-sample loci Houses 6 and 8 and Midden 63 connect with the t2 and early t3 Don valley Jarrett-Lahmer, Risebrough, and Watford collections, while Midden 78 attaches to Logan and House 3 joins with the later t5 Credit valley Emmerson Springs site. The late Middens 54 and 66 also share ties with sites to the north, in ancestral Wendake. These include the t4 Gwynne and the t5 Hunter’s Oro 17 and Starr sites. Gwynne also ties with Middens 60 and 71, and Starr connects with Midden 59, while the t6 Alonso site ties with the very late House 19. Communities and Practice Network analysis of Keffer ceramics (Chapter 7) has been employed to determine separate communities of local tradition ceramic practice within the Keffer village and to create a timeline of settlement development. Of course, these results are inextricably intertwined. The local tradition communities of practice, divided into five subgroups, and the community development, divided into eight maps, or plans, form the background data structure, as seen in Table 8-4, used to detect the development of extra-village ceramic relations at Keffer. Table 8-4. Temporal Order of Communities of Practice. Local Tradition Community of Practice Map Loci 1 5 15, 67, 68 2 5 3, 51, 53, 55, 57, 75, 78 2 3 8, 12, 56, 74 3 3 62, 65, 72, 74 3 1 13, 20 2 2 1, 9, 10 3 2 60, 61, 77, 80 5 2 52, 59, 63, 66, 71 8 2 54 2 4 7 4 4 11 230 5 4 2, 4, 6, 19 Community of Practice 5: Houses 3 and 15, Middens 51, 53, 55, 57, 67 68, 75, and 78 Nodes of the first, and earliest, of the communities of practice, CoP 5, which consists of Houses 15 and 3 and Middens 51, 53, 55, 57, 67, 68, 75, and 78, connect most highly with earlier Don valley sites. The early House 15, ties very highly with the nearby McNair site and, at lower levels, with Watford and Hidden Spring (Figure 8-22). Spence and Rainey (2017) proposed that the people of these sites are biologically affiliated with the people of Keffer. The presence of a rare genetic dental trait at both Keffer and Hidden Spring led them to suggest that close kin ties may have existed between Keffer and Hidden Spring. Williamson (2014) proposed that Hidden Spring is a special purpose site affiliated with Keffer. The t3 Spang site, in the upper Rouge– Duffins drainage, is strongly linked, above the BR 170 level, to House 15. This anomaly is explained by the unique nature of the local tradition community of practice of the Rouge–Duffins watershed. Ceramic collections from this area are characterized by very high proportions of decorated neck Huron vessels, specifically, Black Necked. Neck-decorated pottery is generally accepted to decrease within each community over time, but the overall larger proportions seen in both the Rouge–Duffins and the Trent sites skew the appearance of their temporal relationships, placing them chronologically with earlier ceramic collections of other drainage basin–oriented communities. The smaller middens of this community of practice, Middens 67 and 68, which appear with House 15a in the first settlement, connect most closely with the nearby JarrettLahmer site and with the Risebrough site to the south. At a lower level, just above BR 150, Midden 67 also ties with the Don valley McNair site. With the expansion of House 15b in the next development of the village, Midden 75, located south of House 15b, ties to the Risebrough site on the Don River. Midden 78 is also created to the south of House 15b. Ceramics from this deposit connect with the Logan site, officially in the Humber watershed but closely affiliated with Keffer. Midden 57, situated to the west of the first houses in Cluster A, in the low town, shares close ties with the earlier Don valley Risebrough site (Figure 8-22). The large size of Midden 57 suggests a significant timespan for the use of this midden, by a large population. This suggests that the first community of practice was not short lived at Keffer. Midden 57 also shares ties with Logan and Jarrett-Lahmer. This may indicate an 231 early to mid-century date for the Logan site. Also appearing in the low town with the first village expansion are Middens 51, 53, and 55. Midden 53’s position near later Houses 4 and 11 would suggest a later date for this midden, but its close ties with the earlier Don Watford and McNair sites indicate it is an early midden that ties to the first longhouses of Cluster B, its neighbours, Houses 8 and 9. At the west end of these houses, Midden 55 links closely with the slightly later Don Jarrett-Lahmer and Risebrough sites. The first occupation of the Keffer site is characterized by a community of practice based in the larger Don drainage community of the mid-fifteenth century. Community of Practice 3: Houses 8 and 12, Middens 56, 62, 65, 72, 73, and 74 The introduction of a new and slightly different community of practice is seen in this first expansion. In the low town, Cluster B House 8 shows some affinity with the earlier Don Macartney site but shares ties to the Logan site. At the south end of the cluster, Midden 56, appearing at this time, also shows low-level ties with Logan and with one Don drainage site, the McNair site. A similar level tie with the later Trent valley Coulter and Kirche sites reflects the significant amount of decorated neck Huron vessels in the sample. Its presence at this time helps to confirm that House 9, directly north, existed in this early period. This practice community also appears on the plateau. The first house added to the high-town cluster is House 12. Midden 74 beside it begins with the erection of this house. Both ceramic collections share their highest ties with Logan and lower ties with the Don Jarrett-Lahmer and Risebrough sites. House 12’s ties with Coulter and Kirche reflect the neck decorated vessels present here, but not seen to such an extent in Midden 74. This community of practice expands with next growth of the high town occupation. With the removal of House 15b, new longhouses are constructed on the plateau. Refuse from the first occupation of House 13a is found in Middens 72 and 73 to its south end, and deposits from House 18a are found nearby on the northwest slope in Midden 62. Middens 62 and 72 both tie most closely with the Logan sites and, at a slightly lower level, with the Don Jarrett-Lahmer and Risebrough sites, supporting the early placement of Houses 13a and 18a in the village. Midden 73 links at its highest level with the later Humber valley Damiani site and may date to a slightly later occupation of House 13. At some distance from the base of the plateau slope, outside the 232 village palisade, refuse began to accumulate in Midden 65, the site’s largest midden, around this time. Midden 65, located between House 20 and the later House 19, shares a high-level tie, above BR 170, with the neighbouring Jarrett-Lahmer site, indicating that it may have experienced considerable use in this early period by members of this practice community located in Houses 8 and 9 of Cluster B. Community of Practice 1: Houses 13 and 20 In this period, a new longhouse, House 20, was built outside the first palisade. It has a unique ceramic practice, characterized by the absence of Neutral ceramics, an unusual occurrence at Keffer. This is matched by an alteration in the ceramic practice of House 13 from that seen in its original refuse deposits, Middens 72 and 73. The House 13 and 20 pair, which tie at BR 190, form the small group of CoP 1. The strongest level external link shared by both Houses 13 and 20 is with the Trent Kirche site at BR 160, although House 13 also ties with the Humber Logan site at this level. House 20 shares a slightly weaker bond with Logan and Coulter. It is the high level of decorated neck vessels, that is, Black Necked, and the total absence of Neutral ceramics that distinguish this community of practice. Community of Practice 2: Houses 1, 9, and 10, Middens 52, 54, 59, 60, 61, 63, 66, 71, 77, and 80 The large membership of CoP 2 grows from the appearance of Houses 1a, 9a, and 10a in the settlement of the low town. The weak external ties of these houses, just above BR 150, suggest that their long-term occupation and the changing practices occurring over that period within the houses lack signs of distinctive relations with any specific village or village community. House 1 shares a low-level connection with the Logan site, but Houses 9 and 10 are linked to more distant sites, the Simcoe County Bidmead site and the Prince Edward County Ames site, respectively. Midden 60, started sometime after the construction of House 1 and directly to the west, shares a much higher tie, above BR 170, with the Logan site as well as ties above BR 160 with the later Damiani site, in the upper Humber valley, and with the Simcoe County Gwynne site. Midden 61, associated with House 14, on the north slope of the plateau, directly south of Midden 60, also shares ties above BR 170 with Damiani and with Starr, another Simcoe County site. Midden 80, probably associated with the occupation of the House 13b, is more closely related to the Logan 233 site. In cluster D, Midden 77, associated with House 20, shares ties not only with Logan, but also with the earlier Don Risebrough site. The slightly later House 20 midden, Midden 71, is tied at a high level with Logan as well, but its slightly lower ties are with the Simcoe County Gwynne site. To the north, in the newly openedup lowland area inside Palisade 2, Midden 52 forms between House 8 and the newer House 4. Ceramics of this midden link with Logan above BR 170 and with earlier Don sites, Risebrough and Jarrett-Lahmer, at lower levels. New middens of CoP 2 forming on the plateau at this time no longer tie with earlier Don sites. Midden 59, related to the new residence, House 16, shares its highest ties with Starr in Simcoe County and slightly lower-level ties with Damiani. Midden 66, earlier suggested as relating to House 16 and other late residences on the plateau, ties most closely with Damiani and then Logan. Lower-level ties link it to the Gwynne. Near the end of the village occupation Midden 54 appears over the remains of Houses 6 and 7b. This late midden is tied most closely with the later Damiani site. Ceramic practices reflected in this community of practice are dominated by their affinity with the two differing Humber valley sites. Earlier samples link most highly with the Logan site and some early Don sites while ceramic practices of the later Damiani site are associated most strongly with the last longhouses. Many houses in this community are also characterized by developing ceramic similarities with later sites in the Wendake homeland. Community of Practice 4: Houses 2, 4, 6, 7, 11, and 19 Small-sample Houses 2, 4, 6, 7, and 11 and the larger-sample House 19 make up CoP 4. Construction of these houses occurred at several different stages of the village’s growth. House 7 likely dates to the first settlement of Cluster A in the low town. House 11 was originally constructed outside of Palisade 1, before this palisade’s removal. Houses 2, 4, and 6 were likely constructed in conjunction with the erection of the second palisade, while the extension of House 7b post-dates the removal of House 6. House 19, like House 11, was originally built exterior to the village palisade but was later incorporated into the village. Traces of the early occupation of these longhouses are present in their accompanying middens. The practice reflected in the house ceramics dates to later occupation of the houses. Longhouses at Keffer had very few pit features. Ceramics from these houses therefore probably date to their late occupation. We see practice changing in this late period, as undecorated ceramics make up a much larger percentage of the 234 samples. These changes are reflected in the links appearing at low levels with Simcoe sites. The community has no external ties above BR 170 and few below, as these nodes sit peripheral to the main component of the network. Houses 2 and 19 and, to a lesser extent, House 4 tie with the later Damiani site. The earlier House 7 links with Logan above BR 160. House 6, with a very small sample, of 8 vessels, has no external ties at or above BR 150. House 11 shares its highest external similarity with the later Simcoe Orr Lake and Bidmead sites at just above BR 140. Strong ties existing between Logan and nodes of the each of the later strongly tied communities, CoP 1 to 4, are particularly interesting as the absence of nonlocal ceramics at Logan suggests it is an early fifteenth-century settlement. Latta (1980) dates Logan ceramics to the same period as the mid-fifteenth-century Oshawa-area Macleod site. The influence of external communities on the dynamic process of change and the development of Keffer’s ceramic practices (Figure 8-26) may lie in the multiple connections linking Keffer ceramic practice groups to other southern Ontario Iroquoian settlements through time. 235 Figure 8-26. Local tradition ceramic communities of practice plan Maps 1–8. External Ties by Time As far as the ceramic tradition of Keffer, the people of this village, in all its stages of development, appear to have come from the local Don valley community. Nonlocal ceramics are not the majority in any of the collections, and those that follow other traditions are almost certainly made on site. There is no evidence of a “foreign” group establishing of any of the households or communities of practice. 236 Table 8-5. Keffer House and Midden Similarity Ties with Other Ontario Iroquoian Sites. Sites with Highest Ties Map 1 Locus H15 Local Tradition Commun ity of Practice 5 1 M67 5 Don 1 M68 5 Don 2 H3 5 Credit, Don 2 M51 5 Don 2 2 M53 M55 5 5 Don Don 2 M57 5 Don, Logan 2 2 M75 M78 5 5 Don Don 2 H12 3 Logan, Trent, Don 2 H8 3 Don 2 M56 3 Logan, Don 2 M74 3 Logan, Don 2 2 H1 H10 2 2 2 H9 2 Logan Prince Edward County, Humber Simcoe, Humber Geographical Areas with Highest Ties Don Ties above BR 170 McNair, Spang Ties above BR 160 Watford, Kirche, Coulter Ties above BR 150 Ties above BR 140 JarrettLahmer, Risebrou gh McNair Emmers on Springs Risebro ugh Ties above BR 130 Jarrett-Lahmer, Risebrough Risebrough, Logan Risebrough, Logan, JarrettLahmer, Coulter Watford Jarrett-Lahmer, Risebrough Jarrett-Lahmer Risebrough Logan Kirche Coulter McNair Logan Logan Gwynne Risebrou gh JarrettLahmer, Risebrou gh, Fournier Logan, Macart ney Logan, Kirche, Coulter, McNair Logan Jarrett-Lahmer, Risebrough, Gwynne Logan Ames Bidmead Damiani Damiani , 237 Emmers on Springs 2 3 H7 H13 4 1 3 H20 1 Humber Logan, Trent, Don Logan, Trent Damiani Logan Kirche 3 M60 2 Logan, Humber Logan 3 M61 2 Damiani 3 M77 2 Humber, Simcoe Logan, Don 3 3 M80 M62 2 3 Logan, Humber Logan, Don 3 M65 3 Don, Logan 3 M72 3 Logan, Don 3 4 M73 H11 3 4 Humber, Don Simcoe Damiani 5 M52 2 Logan, Don Logan 5 M59 2 Starr, Damiani 5 M63 2 Simcoe, Humber Simcoe, Humber 5 M66 2 5 M71 2 Humber, Simcoe Logan, Simcoe Damiani, Logan, Gwynne 5 H2 4 5 5 6 H4 H6 H19 4 4 4 8 M54 2 Kirche Logan JarrettLahmer Logan Logan Ames Coulter, McNair Logan, Coulter, Fournier Damiani, Gwynne, Risebrough Starr Gwynne, Risebrough Logan Logan, JarrettLahmer, Risebrough, Gwynne Logan Damiani Damiani Jarrett-Lahmer, Risebrough, Fournier Jarrett-Lahmer Humber, Simcoe Humber Simcoe Humber, Prince Edward County Damiani Humber, Logan Damiani, Logan, Starr Damiani, Ames Orr Lake, Bidmead Risebrou gh, JarrettLahmer Logan Fournier , Damiani Gwynne Orr Lake, Bidmead Bidmead , Ames Damiani Logan Fournier Robitaill e, Bidmead Damiani 238 Social Network Analysis of Nonlocal Tradition Ceramics at Keffer The Keffer local tradition ceramics have been separated into a series of five spheres of practice. The hard boundaries are archaeological constructs and are not recognizable among the practices within the village. Instead, ceramic practice occurred within a temporal and social continuum, as demonstrated in these plans of ceramic practice development, labelled Plan Maps 1–8 (Figure 8-27). Figure 8-27. Local tradition ceramic practice by settlement plan Maps 1–8. In this section, in the social network analysis of the above-mentioned emergent, nonlocal, and exotic ceramics are examined within one database. Further analysis of the graph results, using detailed typological information for the collection, is then used to separate the ceramics of these three categories, discussed in section 9.5. Network analysis of the nonlocal collection includes all vessels not belonging to the local tradition as defined in section 4.5. This includes locally made vessels and those made outside the village. All pottery fitting a type definition is hypothesized to reflect similar external influences. The first set of network graphs employing nonlocal ceramic data are characterized by the node attribute of sample size to determine the effect of this attribute on graph topology at both high 239 and low similarity levels (Figure 8-28 and Figure 8-29). At BR 100, all of the small-sized samples (fewer than 50 vessels) are excluded from the core of the graph; they are connecting as pendants or, in some cases, totally detached from the component. Larger samples are generally located within the graph core. Figure 8-28. Nonlocal ceramics by sample size BR 100 spring layout. At the higher BR level of 130 (Figure 8-29), several large collections, specifically Midden 74, the second largest sample at the site, with 644 vessels, and Midden 62, with more than 200 vessels, do not connect with the main component. Midden 57, with more than 500 vessels, does not connect with any other sample nodes. Thus, the size of small samples, those under 50 vessels, is at least partially a determinant of network topology. These small samples often include very few, sometimes only one, nonlocal vessels. The percentage of nonlocal ceramics in those samples that do include nonlocal vessels ranges from 3.2 percent to 25 percent. This proportion is not a direct result of total sample size, since Midden 57, with 522 vessels, has 6 percent nonlocal pottery, while Midden 62, with 208 vessels, has 25 percent nonlocal ceramics. Instead, it is a result of the fact that smaller samples, even those with close ties to other nodes, do not contain the wide diversity of nonlocal ceramic types, 15 in all, that is possible in larger collections and therefore cannot connect highly with the largest samples at the core of the component. Large samples may not include this high level of diversity either and likewise remain on the graph periphery. 240 Figure 8-29. Nonlocal ceramics by sample size BR 130 spring layout. The primary network analysis of nonlocal ceramics at Keffer explores possible distinct practice communities for ceramics of this category and examines their relations with the established local tradition communities of practice. To facilitate this last aim, graph nodes are colour-coded here according to their local tradition communities of practice, CoP 1–5, as seen below (Figure 8-30). 241 Figure 8-30. Keffer village local tradition practice. Ties formed at the highest level display the fundamental differences in these relationships, and lower-level connections bridge the separate practices (Figure 8-31). Delineation of separate practices is visually assessed and thus the product of subjective choice. 242 Figure 8-31. Nonlocal ceramics by local tradition communities of practice BR 150 Gower layout. Nonlocal ceramic samples separate into two distinct groups at BR 150, or 75 percent similarity. The largest collection, from CoP 3 Midden 65 anchors CoP 2 Middens 61 and 77 into the closeknit component of nonlocal ceramics (NLC) Group 1. Midden 65 sits on the periphery of the core in the local tradition graphs at BR 185 and is not closely linked with either Midden 61 or Midden 77. Two other nodes of CoP 2, Middens 80 and 66, are linked by House 13 of the small, tightly bound CoP 1. House 20, with which it shares very high similarity in local tradition ceramics, does not appear with House 13 at this level, but its closest ties, Middens 80 and 66, are located close to House 13 on the plateau. CoP 3 Middens 60 and 71 also link to House 13 at BR 140 (Figure 8-32). These middens are more distant from House 13, with Midden 71 located next to House 20 and Midden 60 farther away, in the north end of the site. No new nodes join the Midden 65–centred NLC Group 1 at this level, but two new small component pairs appear. Middens 62 and 74, which form the core of local tradition CoP 3, are linked in Group 3. CoP 5 Midden 53 and CoP 3 Midden 54 do not share local tradition practices but are closely positioned beside House 8 in the low town. This pair is labelled NLC Group 4. 243 Figure 8-32. Nonlocal ceramics by local tradition communities of practice BR 140 spring layout. House 20, which ties closely with House 13 in CoP 1, links to House 13 and the adjoining CoP 2 Midden 60 and 80 nodes at BR 130 and ties this growing component with the NLC Group 1 Middens 65 and other CoP 2 nodes Middens 61 and 77 (Figure 8-33). House 20 and Midden 65, which are of differing local tradition practices but are located side by side, bridge these two components. CoP 1 House 20 links through House 13 with NLC Group 2. Midden 72, part of CoP 3, and located south of House 13 and beside Midden 80, ties through Midden 80 to House 13. The early Midden 55, a member of CoP 5, links to House 13 and Midden 71 at BR 130. Midden 55 is not found in close proximity to any other village features of this component. No changes are seen in the two pair groups, NLC Groups 3 and 4. One new pair appears. CoP 5 Midden 51 and House 2, a late house in the northern part of the village’s lowland, are located in close proximity on the site. Few nonlocal vessels are found in these samples, and this may be the reason for their high similarity level, as discussed above. 244 Figure 8-33. Nonlocal ceramics by local tradition communities of practice BR 130 spring layout. The components of the graph continue to consolidate as the similarity level decreases to BR 120 (Figure 8-34). CoP 2 Midden 59 links to this core through Midden 66, both of which are related to House 16 according to the settlement analysis. To the right of the component, Midden 65 acts as a bridge connecting the three main clusters of NLC Groups 1, 2, and 3 with the CoP 3 nodes of NLC Group 3, Middens 62 and 74, which are connected in peripheral pendant formation. It is only at this level that the large, early CoP 5 Midden 57 joins the main component of NLC Group 2, through Midden 61. Midden 57’s topological position as a pendant confirms its distinctive character when compared with the growing component centred on the later House 13. CoP 4 House 19 similarly links in a pendant to Midden 77, which overlies it in the later occupation of the site. 245 Figure 8-34. Nonlocal ceramics by local tradition communities of practice BR 120 spring layout. At the similarity level of BR 110 (Figure 8-35), the nodes of CoP 1 and CoP 2 compose the main core of the large component. At this level, CoP 3 House 12 connects with the peripheral CoP 3 NLC Group 3 through Midden 62. The CoP 3 Midden 73 node, on the other hand, connects with the main core through House 13, which is directly beside it on the plateau. CoP 5 nodes are scattered as pendants. Middens 68 and 75, related to the last occupation of House 15b link with House 12, which was erected at the same time as House 15b. The pair NLC Group 4 ties through early Midden 57 to the component. The low level and peripheral connections of these CoP 5 nodes illustrate their temporal difference from the main body of the CoP 1 and NLC Group 3 core. The second pair group from the north end of the village, House 2 and Midden 51, are joined by the nearby Midden 52, located at the north end of House 9, and Midden 56, situated to its south. Connections made at these low levels are not indicative of strong similarities. 246 Figure 8-35. Nonlocal ceramics by local tradition communities of practice BR 110 spring layout. Below BR 120, additional nodes, with weak ties, have only a small effect on the main structure of the main component, but they pull stronger ties away from each other and hide the original nonlocal groups apparent at higher levels. These peripheral ties suggest possible relationships of those samples with very few, often only one, nonlocal vessels (Figure 8-36 and Figure 8-37). Ties at these levels are not considered significant. 247 Figure 8-36. Nonlocal ceramics by local tradition communities of practice BR 100 spring layout. Figure 8-37. Nonlocal ceramics by local tradition communities of practice BR 70 spring layout. 248 The lower similarity levels characteristic of nonlocal networks has resulted in generally less distinct separations within village practice of these ceramics. Smaller sample sizes appear to be partially responsible for these results. However, network topology indicates that nonlocal ceramic practice within the Keffer village is highly varied across the site. This variation spreads along a continuum from three distinct centres of unique, dissimilar practice. The original components seen at the 75 percent (BR 150) (Figure 8-31) level of similarity, include NLC Group 1, composed of Middens 61, 65, and 77; NLC Group 2, a larger star component with five nodes, House 13, Middens 60, 61, 71 and 80; and the NLC Group 3 component of Middens 62 and 74. The NLC Group 4 ceramic samples from Middens 53 and 54, with three and seven vessels, do not provide sufficiently robust data to form the core of a community of practice. The position of these two nodes within a pendant unconnected to nodes other than the Midden 57 bridge node at the low BR 100 level (Figure 8-36) support this conclusion. At BR 130, the NLC Group 2 component grows with the addition of Houses 1 and 20 and Midden 55. This group then links directly to the core node of House 13 (Figure 8-33). Midden 72 joins the component as a peripheral node. With the appearance of the House 20 node in a connection to the NLC Group 1 House 13 node at the BR 130 level, a bridge is created linking the two small world NLC Groups 1 and 2 through the large Midden 65 node. NLC Group 3 remains as a separate component, suggesting that it has unique properties. It eventually links to the main component through the Midden 65 node at BR 120. At this level, the distinct nature of these three components remains visible as NLC Groups 1 and 3 project outward from the main body of the NLC Group 2 component. As similarity levels decrease below this level, additional pendants attach to this main body and it becomes apparent that NLC Group 2 is the centre of most nonlocal practice in the village. NLC Group 1 was joined by House 19 at BR 120, and House 12 attached to NLC Group 3 at BR 110, with Middens 68 and 75 also connecting as pendants. By this level, no separation is visible between NLC Groups 1 and 2, and these practices join to form a larger and more diverse community of practice. NLC Group 3 also links with this larger practice community at this level, but it remains somewhat divergent on the periphery of the graph (Figure 8-35). 249 Nonlocal Ceramic Practice at Keffer Three connected and related communities of practices (Figure 8-38, Table 8-6) can be discerned from visual analysis of the nonlocal ceramic network graphs. The relations among these three components vary in strength and direction, with the members of NLC Group 1, specifically Midden 65, taking a central bridging location. Tied with Midden 65, the other core nodes of this group, Middens 61 and 77, became integral to holding the components together. Many nodes also tie to the main component as weakly connected components, not really members of any of the three groups. 250 Figure 8-38. Nonlocal ceramic group practices. 251 Table 8-6. Network Ties of Nonlocal Practice. Loci/Nonlocal Ceramic Practice Groups with Highest Ties Nonlocal Ceramic Practice Group 1 BR 150 61, 65, 77 BR 140 BR 130 NLC Group 2 2 66, 13, 80 60, 71 NLC Group 1, 20, 55, P1, P72 3 62, 74 BR 120 NLC Group 3, P19, P57 P59 BR 110 NLC Group 1 P12, P75 BR 100 P73 NLC Group 2 Note: P means the node ties as a pendant. As a mediator between the NLC Groups 2 and 3 practices, NLC Group 1 is dominated by the large and very diverse collection of Midden 65. This midden was accessible to most occupants of the southern half of the low town at different times during the habitation of the village. Midden 77, at the opposite end of House 20 from Midden 65, is also a core member of this NLC Group. The position of House 20, on the other side of the bridge linking NLC Groups 1 and 2, is of interest as Midden 77 is a refuse deposit located between Houses 19 and 20 and therefore associated with both. House 19, the last house to be erected at the village, joins this community at the BR 120 level. To the north, on the plateau slope, Midden 61 and its associated plateau-top house, House 14, is the third member of this core NLC Group. Located to the north of Midden 61, on the riverside slope of the low town Cluster A, the large and long-term Midden 57 connects this NLC Group via a pendant to Midden 61. The proximity of Midden 57 to the west end of House 3 and to House 7, which it overlaps, and perhaps to the later House 2, suggests that these houses may also be part of the NLC Group 1 community. The second practice community centres on the neighbouring eastern plateau-top House 13 and nearby Middens 66 and 80. At BR 130, Midden 72 joins this NLC Group as a pendant, suggesting that it is tied to the practice here but not integral to it. The connection of Midden 59, at BR 120, to the core indicates that the associated longhouse, House 16, which is situated next to House 13, is also part of the community. This may be due to its spatial location between NLC Groups 2 and 3. Midden 71, a late deposit located at the base of the plateau slope in this area is also a core member of this practice NLC Group. Midden 71 relates to the latest occupation of 252 House 20, which joins this NLC Group at the slightly lower BR level of 130. Situated in the extreme north of the village’s low town, Midden 60 also belongs to the core of this second practice NLC Group. The addition of House 1, situated to the east of Midden 60, into this NLC Group at BR 130 illustrates that the use of similar emergent ceramics was also occurring in this area of the low town. Midden 55, located just south of the House 1 cluster, is also integrated into this NLC Group at BR130. This is significant as the low volume of pottery from excavations in the northern part of the low town makes associations for households in this zone conjectural. The probable creation of this midden by the occupants of Houses 8 and 9 suggests that these two households may have also been part of this larger nonlocal ceramic practice community. The third NLC Group is composed of Middens 62 and 74. These repositories for the waste of Houses 12 and 18 are located on the west side of the plateau, along with Houses 12 and 18. House 12 joins this component at BR 110, as does Midden 75, southeast of House 12. Core members of this NLC Group are confined to Houses 12 and 18 on the west side of the plateau. Membership within both NLC Groups 2 and 3 appears to be spatially oriented, most noticeably by the position of the communities on opposite sides of the village plateau. While the smaller NLC Group 3 community of two houses is limited to the west side of the plateau surface, members of the NLC Group 2 community located on the eastern edge of the hilltop, in Houses 13 and 16, spread both north and south into Cluster A and D of the low town. With the later construction of the NLC Group 1 House 14, all three practice NLC Groups reside, side by side, on the plateau. Ceramic evidence indicates that the highly linked NLC Groups 1 and 2 are both spread across the lower level of the site, to the north and the south. Most houses and middens not included in these highly tied components connect loosely with Group 2, on the opposite side of the graph from Group 3. The NLC Group 3 community, more distinct in its nonlocal practice, perhaps due to its small size, is confined, and perhaps limited to, one extended family, although local tradition ties suggest that it is affiliated with House 8 on the lower level. These differing practices, as determined by the graphic separation of NLC Groups 1 and 2 above the BR 130, or 75 percent, similarity level, and at BR 120 for NLC Group 3, can only be understood through detailed analysis of their ceramic samples. This is particularly valid in the case of nonlocal collections. 253 The large proportion of vessels in the Keffer site collection which no do fit within the definition of local tradition ceramics have been categorized as “nonlocal” here and by others (Birch et al. 2017; Spence and Rainey 2017). It is now apparent that this term conflates multiple unique nonlocal ceramic practices. In this section, nonlocal ceramics are divided into three categories. The first category is composed of those ceramics which fit well within the MacNeish type definitions of ceramics believed to have originated within one of the historic Haudenosaunee territories. Those vessels characteristic of Seneca, Cayuga, Onondaga, Oneida, and Mohawk types appearing on north shore sites in the mid-fifteenth century appear to follow the traditions of communities within their respective Haudenosaunee territories. However, visual examination of the Keffer vessels in this category cannot determine their location of manufacture. No distinguishing local tradition traits have been observed in examples of the Keffer collection. The presence of a well executed Wagoner Incised learner vessel suggests the presence of potters with knowledge of nonlocal vessel manufacture although the wide diversity of types present in low numbers across the site does not support an on-site centre of production. These vessels may be of local or foreign manufacture, but only further investigation and chemical testing can resolve this issue. These ceramics are referenced under the label Haudenosaunee. Here, the term “exotic,” often used for any nonlocal ceramics on Ontario sites, is reserved for vessels of almost certain off-site, or foreign, production. The second category of nonlocal ceramics is thus that of vessels of foreign manufacture, or “exotics.” These vessels make up a small minority of the nonlocal collection and consist almost exclusively of the Durfee Underlined type, belonging to the St. Lawrence Iroquoian and Northern New York tradition. Roebuck Low Collars are also found. Both of these vessel types may be embellished with reed punctates and thumb-nail gouged basal notches characteristic of St. Lawrence Iroquoian sites of southern Ontario. Other hallmarks of the St. Lawrence Iroquoian tradition are complex glide reflection motifs (David Smith, personal communication 2018) and precision of decoration. The presence of a small proportion of these ceramics at various locations throughout the Keffer village suggests that direct long-distance relations occurred between members of these households, or communities of practice, and peoples producing St. Lawrence Iroquoian ceramics, hundreds of kilometres to the east. St. Lawrence and Northern New York–type vessels found in north shore collections, specifically Durfee Underlined and Roebuck Low Collar, on the other hand, often do not display the 254 characteristics of St. Lawrence valley manufacture. These vessels conform to the type definition as set out by MacNeish and are therefore typed as such. When these vessels are compared with vessels of known St. Lawrence provenience, obvious differences become apparent. Potters of these north shore vessels do not follow the accepted chaîne opératoire of the original practice community. It now appears that many, if not all, of these north shore vessels were manufactured locally within north shore villages, while a minority, namely, those characterized by St. Lawrence Iroquoian production traits, are believed to be of foreign manufacture. The presence of locally made “St. Lawrence Iroquoian-style” (Trigger et al. 1980) ceramics on north shore Iroquoian sites has been acknowledged for several decades. A study involving chemical analysis of these anomalous vessels concluded that “there is no reason to believe that the St. Lawrence Iroquoian-style sherds were not being manufactured locally” and that “no significant difference” in the chemical distribution of sherds from either the “local Ontario Iroquoian-style” or St. Lawrence Iroquoian–style ceramics examined from the Parsons site was apparent (Trigger et al. 1980:128). I have examined vessels from the Parsons site included in this 1980 study and can confirm that they coincide in appearance with Durfee Underlined vessels found at Keffer and on other north shore sites. The uniformity of production within the collection of “St. Lawrence Iroquoian–style” ceramics follows the concept of emergent ceramics (Lee 2006; Fowler 2017). These new ceramics have often been regarded as “hybrid.” However, in Pendergast’s (1980) study of St. Lawrence Iroquoian–like ceramics from the Draper site, he regarded hybrid pottery as a mixture of random ceramic attributes combined haphazardly on various vessels. It is believed to be the result of contact among distinct ceramic traditions that involved the adoption of individual traits of interest. Hybrid vessels, following this definition, are found in very small numbers at Keffer. In a recent study of mixed tradition ceramics in the American Southwest, Barbara Mills pronounced the term “hybrid” as unhelpful because it describes an object without providing an explanation for the processes responsible for its appearance (Mills 2018). Following Wenger’s (1999) work on communities of practice, Mills has categorized these vessels as “boundary objects” (Mills 2018). These ceramics, however, reflect a distinctive type of hybrid object. Here, defining traits of ceramics from two different traditions are blended in a standardized manner, producing vessels of a new “type.” These ceramics clearly reflect the influence of both traditions and are meant to negotiate a new social reality where the boundary between the two traditions is erased. The new St. Lawrence—Northern New York “type” vessels 255 found in significant numbers at Keffer are not hybrid ceramics, characterized by random nonlocal traits applied to Huron vessels, or vice-versa. Neither do they fit the definition of “boundary objects.” Although they display obvious characteristics or the original iterations of production of these types, there is no visual evidence of purposeful blending of two traditions to signify a new joint identity. Along with the locally made Durfee Underlined vessels, three additional ceramic types found at Keffer fit this description. These are Roebuck Low Collar, whose prototype also originated in the St. Lawrence valley, and Otstungo Notched and Rice Diagonal, which are thought to have Mohawk origins. As discussed in Chapter 4, the fact that these ceramics are found in the collections of only a few sites on the north shore, and nowhere else in Iroquoia, indicates that they are local redesigns, or recreations, of nonlocal ceramic types. These types have been divided into two main categories. The first of these is composed of locally made vessels conforming to the MacNeish Otstungo Notched type, with 118 vessels, and second is the related Rice Diagonal type, with 46 vessels. The Otstungo Notched and Rice Diagonal types are collarless, everted-lip forms characterized by a long neck and slightly oblong body shape whose widest diameter matches that of the rim. Otstungo Notched and Rice Diagonal ceramics of Keffer and other sites on the north shore closely follow the type description of MacNeish but are clearly different from the New York ceramics (Figure 8-39). These vessels are of more “bold” (Wayne Lenig, personal communication 2018) construction and decoration. Vessels and rims are of heavier and more regular form, with precise and well-executed motifs. Unlike New York State vessels, those of the north shore are commonly decorated with large-scale horizontal line or complex oblique neck decoration, a trait particularly common east of the Don River at this period, in the Rouge– Duffins and Trent drainages. The two types are commonly found together on north shore sites and separated from each other only by a variation in the lip decoration. These everted-lip vessels belong to a production tradition, or chaîne opératoire, quite distinct from that of north shore Iroquoian ceramics. They have been amalgamated into a category of ceramics labelled “evertedlip.” Examples found at Keffer and other north shore sites are very well executed and would have required a great deal of skill to produce. Although these ceramics are found in large proportions at Keffer, they appear in much lower proportions on other fifteenth-century north shore sites. 256 Figure 8-39. Otstungo Notched emergent ceramic from Keffer. Ceramics fitting the Durfee Underlined type, numbering 76 vessels, and those of the Roebuck Low Collar type, numbering 50 vessels, are characterized by a very different vessel form from that of the everted-lip vessels, one much more along the lines of north shore Iroquoian vessels, with a globular body, constricted neck, and prominent collar. These traditions share a similar chaîne opératoire. The complex motif and vessel formation of this pottery is generally not well executed, and less effort appears to have been put into vessel production in general. These vessels do not show the precision of either construction or decoration seen in St. Lawrence Iroquoian vessels. They are, however, uniform in construction and decorative characteristics across the north shore, forming a “subtype” per se. Durfee Underlined vessels in this category are most often decorated with the simplified SDEA motif (Figure 8-40), as described in Chapter 4. It is seen on some of the north shore Roebuck Low Collar vessels as well, but not with the same frequency. This motif is not common on St. Lawrence valley–constructed vessels. Distinction between the two types is heavily based on collar height, with collars of the low collar 257 vessels being under 25 mm and those of Durfee Underline types measuring over 35 mm. High collar vessels with complex or opposed collar decorations are a hallmark of mid-fifteenthcentury Ontario Iroquoian pottery. Due to the prevalence of the higher-collared Durfee Underlined vessels over Roebuck Low Collar vessels at this site, ceramics of this second category are known here simply as Durfee Underlined ceramics. Both the Durfee Underlined and Roebuck Low Collar vessels, identical to those seen in several other sites across the north shore, fit the definition of emergent ceramics. Figure 8-40. Reconstructed Durfee Underlined emergent ceramic from Parsons. (Williamson and Powis 1998:Figure 20) These locally manufactured Durfee Underlined and Everted-lip ceramics adhere closely to characteristics of emergent ceramics that forms the third category of nonlocal ceramics at Keffer. In the Keffer collection, these emergent vessels make up the majority, 59 percent, of the 495 nonlocal ceramics of known house or midden provenience. Haudenosaunee ceramics comprise 258 38 percent of the nonlocal ceramics, and “exotics” account for only 4.4 percent of the total nonlocal pottery. The nonlocal ceramic database used for network analysis subsumes these three nonlocal ceramic categories into a single variable, that of MacNeish type. Resulting network graphs, as seen above (Figure 8-34 and Figure 8-35), portray a large component representing one overall ceramic practice community, segmented but tied together by bridging practices. The distinct nonlocal practices of NLC Groups 2 and 3 are connected by the practice of NLC Group 1 to form a unified continuum of ceramic practice. Participants of NLC Group 1 incorporate practices of NLC Groups 2 and 3 in almost equal proportions. This is seen most clearly in the almost equal proportions of the two emergent ceramic categories, with a range of 20 to 36 percent Durfee Underlined vessels and 24 to 26 percent everted-lip ceramics. The total proportion of the emergent vessels in this NLC Group is, however, much lower than in either of NLC Groups 2 or 3. NLC Group 1, instead, is characterized by higher usage of vessels assigned by MacNeish to the Five Nations, ranging from 33 to 48 percent in various ceramic samples. Pottery accredited to each of the Seneca, Cayuga, Onondaga, Oneida, and Mohawk territories exist at Keffer, and each of these types is present in this collection. In NLC Group 1, and in most Keffer collections, the largest portion of the Haudenosaunee ceramics is of Onondaga affiliation, the vast majority of which are of the Syracuse Incised type. The second most common type is Mohawk Wagoner Incised. Pratt (1980) has conflated these two types into the joint Syracuse-Wagoner Incised type, which composes up to 50 percent of the Haudenosaunee type ceramics in NLC Group 1 and 41 percent at Keffer overall. In addition to a substantial presence of Haudenosaunee ceramics, the core NLC Group 1 ceramic samples of Middens 65, 61, and 77 contain 11 of the 21 exotics found at Keffer, or 52 percent. Of these, 10 are found in the southern Cluster D Middens 65 and 77. Compared with NLC Group 1, NLC Group 2 has a ceramic profile dominated by slightly higher proportions of Durfee Underlined ceramics, ranging between 30 and 50 percent, and lower proportions of everted-lip ceramics, ranging between 17 and 36 percent, in its robust sample. Haudenosaunee ceramics comprise a smaller proportion of the ceramics sample, ranging from 14 to 28 percent of NLC Group 2. With the exception of Midden 80, a collection with only six nonlocal vessels, exotic ceramics are found in the collections of all core members of NLC 259 Groups 1 and 2. Four exotic vessels are found in the Cluster A Midden 60. The remaining three belong to House 20 and Midden 71, both of which are located alongside Middens 65 and 77, in Cluster D. The nonlocal ceramics of NLC Group 3 are characterized by distinctive differences from those of NLC Groups 1 and 2, as seen in the network graphs. The ceramic profile of Middens 62 and 74, which form the practice core, is distinguished by the high proportion of Everted-lip ceramics, 65 and 52 percent, respectively. Durfee Underlined vessels, on the other hand, make up only 6 and 13 percent of the nonlocal pottery, whereas 25 to 35 percent of the samples are composed of Haudenosaunee type ceramics. This is similar to the proportions in NLC Group 2, but much lower than those in NLC Group 1. Midden 74, the second largest ceramic collection at the site, with 60 nonlocal vessels, also has 1 exotic vessel. The large size and long duration of this midden may have allowed for refuse deposition from houses of other nonlocal ceramic NLC Groups, particularly after the removal of House 12. Allocation of this exotic to another NLC Group is supported by the absence of exotics in Midden 62. This collection has the largest proportion of nonlocal ceramics, 25 percent, of any ceramic sample in the village. The divergent ceramic profiles of NLC Groups 2 and 3 suggest a continuum of nonlocal practice within both the settlement cluster and the village. This scale of ceramic similarity and difference is marked at either extremity by the practice of these NLC Groups and is reflective of their node positions at alternate topological ends of the high-level network components. The topological bridging position of NLC Group 1, between the NLC Group 2 and 3 components, is indicative of the central role its members play in bonding these two communities of practice. Lower, but equal, proportions of both emergent ceramic sets and the larger proportion of Haudenosaunee tradition ceramics in the NLC Group 1 samples, when compared with those of NLC Groups 2 and 3, suggests that the emergent ceramics are less important as markers of distinctive identity for this intermediary NLC Group. Identities associated with the two distinct emergent ceramic sets are of more concern to the members of NLC Groups 2 and 3 communities. What aspect of identity these ceramics relay, perhaps kin relations or clan membership, is the subject of conjecture. 260 Shared Local and Nonlocal Practice at Keffer The ceramics characteristic of each of these three individual practices are situated along the continuum of overall nonlocal ceramic practice at Keffer. As most of the earliest and latest deposits are absent from the core NLC Groups, social factors, rather than chronology, appear to exert the most influence on this continuum. Local tradition ceramics, separated into delineated spheres of practice, also occur along a cline. Both social and temporal elements are factors in the distribution of these practices, however. The assignment of the Keffer nonlocal ceramics into separate, but linked, practice NLC Groups, which are archaeological constructs, permits comparison of elements of nonlocal practice with those of the local tradition practices for the exploration of separate, multiple, coinciding, or multiscalar identities, in order to understand the processes that created them and why they disappeared. The separation of the local and nonlocal ceramic databases facilitates exploratory analysis into the possible overlap in these two practices at Keffer. Correlation of the NLC Group members results in a series of four arrays of shared local and nonlocal practice (Table 8-7, Figure 8-41). As a reflection of the tight small world clusters with few nodes displayed in the nonlocal practice graphs, early and late settlement features and those with small nonlocal ceramic samples are absent from these arrays. Table 8-7. Shared Local and Nonlocal Ceramic Practice. Shared Practice Array NLC Group/Community of Practice Core Nodes at BR 140 1 NLC Group 1/CoP 2 61 (14), 77 2 NLC Group 2/CoP 2 60, 66, 71, 80 2 NLC Group 2/CoP 1 13, 20 3 NLC Group 3/CoP 3 74 (12), 62 (18) Nodes at < BR 130 1, 59 (16) 12 261 Figure 8-41. Shared local and nonlocal ceramic practice at Keffer. Network topology of the nonlocal ceramics demonstrates the close relationship between NLC Groups 1 and 2, linking them at the BR 130 similarity level. This consonance is reflected in the presence of these NLC Groups within the same local tradition community of practice, CoP 2. The assignment of Midden 65 to the earlier practice community, local tradition CoP 3, is a reflection of the partially temporal nature of the local practice communities and explains Midden 65’s absence in Array 1 despite its high proportion of nonlocal ceramics. The early chronological 262 placement of Midden 65 is partially due to its extensive period of use throughout the lifetime of the village. During early periods, of less dense occupation, perhaps prior to the introduction of large proportions of nonlocal ceramics, Midden 65, located in a slight depression in open area easily accessible to residents of the village’s east side, was possibly a repository with large-scale use, thus leading to the large deposits of early local tradition ceramics. Later village infilling, and crowding with additional longhouses, would lead to more localized use of Midden 65 by nearby residences, Houses 4, 11, 20, and, perhaps, also House 19, a member of the same nonlocal ceramic practice. This would account for differences in use patterns of this midden, as seen in the two separate spheres of local and nonlocal ceramic practice. The co-existence of Middens 77 and 61 within Array 1 (Figure 8-41) suggest the existence of shared overall practices for Midden 77, just outside House 19, and Midden 61, a proxy for House 14 (Table 8-8). Both middens appear on the site in the Map 3 stage of development (Figure 8-27) and are characterized by large proportions, over 70 percent, of undecorated neck ceramics. Midden 61’s (and therefore House 14’s) shared local and nonlocal practices, which separate it from the practices of Houses 12, 13, and 18, illustrate the existence of very different ceramic practices side by side in this plateau-top settlement cluster. Table 8-8. Summary of Ceramic Practices at Keffer. Loci H15 M67 M68 H1 M74 M57 M55 H12 M75 H10 H9 H8 M56 H7 H3 M51 Map 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 Local Tradition Practice Sphere 5 5 5 2 3 5 5 3 5 2 2 3 3 4 5 5 Nonlocal Practice Group Shared Local– Nonlocal Practice Array 2 3 1 2 3 3 2 3 Geographical Area with Highest Ties Don Don Don Logan Logan, Don Don Don Don Don Prince Edward County, Humber Simcoe, Humber Don Logan, Don Humber Credit Don 263 M53 M78 M61–H14 M77 H13 H20 M60 M80 M62–H18 M65 M72 M73 H11 M59–H16 M66 M71 M52 M63 H2 H4 H6 H19 M54 2 2 3 3 3 3 3 3 3 3 3 3 4 5 5 5 5 5 5 5 5 6 8 5 5 2 2 1 1 2 2 3 3 3 3 4 2 2 2 2 2 4 4 4 4 2 1 1 2 2 2 2 3 1 2 2 1 1 2 2 2 2 3 2 2 2 2 2 2 1 Don Don Humber Logan, Don Logan, Don Logan, Trent Humber Logan Logan, Don Don Logan, Don Humber, Don Simcoe Simcoe, Humber Humber, Simcoe Logan, Simcoe Logan, Don Simcoe, Humber Humber, Simcoe Humber Simcoe Humber, Prince Edward County Humber, Logan In Array 2, members of NLC Group 2 and CoP 2 include Midden 60 and House 1 in Cluster A to the north; Middens 59, 66, and 80 on the plateau; and Midden 71, beside House 20, in Cluster D to the south. Like the pottery of Array 1, the local tradition ceramics of this NLC Group date its members to the time of the expansion and infilling of the original village before the extension of Palisade 1. Local tradition ceramics of this period includes high levels of undecorated neck vessels and a small proportion of Neutral ceramics. Local tradition ceramics of CoP 1, composed of the highly similar local tradition ceramics of Houses13 and 20, differ only in the absence of Neutral ceramics. The House 13 nodes are located at the heart of the nonlocal NLC Group 3 component. Nonlocal ceramics of this NLC Group are dominated by the Durfee Underlined emergent ceramics and possess several exotic vessels, implying direct links with ceramic practices of and people to the east. This eastern orientation in the ceramic practice might explain the absence of the western influence, seen in absence of Neutral ceramics. The very close relationship in the local tradition ceramic practices of Houses 13 and 20 has led to the merging of the two local tradition communities of practice, CoP 1 and CoP 2. Houses 13 and 20 therefore join Array 2. The shared local and nonlocal practices of this NLC Group spread throughout the 264 village. Spatially, House 13 and Midden 80 anchor this practice NLC Group on the east side of the plateau, alongside House 14 of Array 1 and House 12 of Array 3. To the east of House 13, the later House 16 is represented by Midden 59. Both local and nonlocal ceramics suggest that the members of House 16 share similar practices with House 13. The use of Midden 66 by the members of House 13b after the removal of House 12 has been suggested earlier. The creation of a pathway across the remains of House 12, from Houses 13 and 16 to Midden 66, at this time would explain the matching ceramic patterns of the two residences. The late erection of House 16 and its short period of occupation may suggest, along with the similarity in ceramic practices that the residents of House 16 were in some way matrilineally related to those of House 13. The long occupation of the village makes it possible that this household is a spinoff from that of House 13. Down slope from House 13 are House 20 and the neighbouring, late Midden 71. The earlier profile of the ceramics from Midden 77, to the south of Midden 71, indicates its use by early occupants of House 20. Midden 71 was added later, when House 19 was built, towards the end of the village occupation. Midden 71, then, represents the later stage of the House 20 communities of practice. The ceramic practices of House 20, as noted, are highly connected with those of House 13. Midden 60 and House 1, beside it, in the northern low town area, also belong to Array 2. The high proportion of emergent type Durfee Underlined vessels in Midden 60 indicates that other houses besides House 1—possibly Houses 2, 6, and 7—may have used this midden for refuse disposal and thus belonged to this NLC Group. There is a direct correlation between the members of the nonlocal ceramic NLC Group3 and the local tradition CoP 3, as seen in Array 3 (Table 8-7). The ceramics of these two households on the west side of the plaza are distinct in both the local and nonlocal ceramic spheres. This close and distinct ceramic relationship is complicated by the temporal differences between the longhouses. House 12 is one of the earliest houses to be constructed at the site. House 18a may have been constructed soon after, but it was not extended until the removal of House 12. The first, small construction of House 18a was likely home to kin of the House 12 residents and may have been created when some members of this very crowded house decided to begin a new household. The distinct nature of the ceramic ties between the houses suggests that the residents 265 of House 12 relocated to the northern portion of the plateau and joined the House 18 household upon the large-scale expansion of House 18c. It is unlikely that this house was constructed by members of another community because the distinctive emergent ceramics of these two houses are not known to be present in large proportions elsewhere on the north shore, or in Iroquoia. Local tradition ceramics of these deposits, with low proportions of undecorated neck ceramics, 67 to 70 percent, and high proportions of decorated neck ceramics, 16 to 25 percent, attest to an early date for the construction of House 12 and a slightly later date for House 18. This chronology, with both the local and nonlocal ceramics of Array 3 predating most collections of Arrays 1 and 2, may signify a temporal disparity between the employment of the Durfee Underlined and Otstungo Notched ceramic sets. The production of the Otstungo Notched and Rice Diagonal vessels may have first appeared at Keffer with the construction of House 12, in the first stage of village growth, while House 15b occupied the central portion of the plateau. With the construction, occupation, and expansion of House 18, the popularity of these vessels, particularly Otstungo Notched, increased. Although these are the only two residences where this pottery dominated the nonlocal practice to this extent, the practice spread throughout the village. Members of NLC Group 1 show evidence of consumption and deposition of these vessels in quite large proportions. Participants of NLC Group 2 practice also used these vessels, but in much smaller proportions than seen in other locations. Production of the vessels may have required special skills not possessed by all residents of the village, and it is probable that these well-made ceramics originated, or were produced, in the practice communities of Houses 12 and, later, 18, and the subsequently shared with other village occupants. Discussion Ceramic practices at the Keffer village are characterized by continuity, diversity, and change through time. Ceramics of the first settlement at Keffer indicate close relations, and possibly the origin of the village, with slightly earlier and contemporary local Don valley communities. The early stages of growth at the village, seen in the construction of three residential clusters surrounded by a one-row palisade, occur as the first ceramics of St. Lawrence and New York State Iroquoians appear on sites along the north shore. Local tradition ceramics already in use at Keffer display some diversity among the practices of the various households. This diversity rests on a continuum of small-scale temporal differences and is the product of the origin and kin ties of these families. Even at this early stage in the village’s history, local tradition ceramic practice 266 separate into two dominant communities, CoP 2 and 3. Members of these separate communities of practice are dispersed in various settlement clusters across the village in a transition from the early Don valley community influence. An influx of ceramics that are not part of the traditional practice within the Keffer community occurs soon after the establishment of the village, with small proportions of vessels typical of the New York Finger Lakes region. These occur in varying amounts in refuse deposits across the village and suggest that the Keffer residents were part of the larger north shore interaction network that was moving these nonlocal ceramics into the area. Concomitantly, an entirely new form of ceramic, totally foreign to the practices of north shore Iroquoians, appears in large proportions in two highly related households on the village plateau that share a unique local tradition sphere of practice with one early longhouse situated in the lower level of the village. This beautifully produced, everted-lip vessel form is a re-creation of vessels originally established in the Onondaga region. This new ceramic is found on only a few sites along the north shore of Lake Ontario, and its production may have centred on this community of practice at the Keffer village because manufacture of this pottery entails knowledge and skills not used in the production of Ontario Iroquoian ceramics. This new and unique pottery was meant to be noticed by others, communicating a message of a divergent identity despite local tradition ceramics indicating a local Don valley origin for the inhabitants of these houses. The debut of a previously unused vessel form might indicate an accompanying new culinary, perhaps feasting, practice. As use of this pottery spreads to varying extents across the village, a second new ceramic is introduced into community practice. This new emergent Durfee Underlined ceramic is more common than the everted-lip pottery on other contemporary north shore sites. At Keffer, emergent Durfee Underlined is associated with the practice of a second, but slightly newer, residential group also centred on the village plateau. Shared local and nonlocal ceramic practices with the newly established longhouse erected outside the original palisade indicate that strong, perhaps kin-related, regional connections influenced village growth. External ties of village households began to shift away from the Don drainage, with new ties developing with a possible northern special-purpose site and to later Humber and Trent valley sites. These new relations may be reflected in the newly appearing pottery. The Durfee Underlined ceramics, though quite different from other ceramics of the north shore, were produced with similar manufacturing techniques and could possibly have been produced by most potters at Keffer. This new ceramic does not display the investment of time and skill seen in the everted-lip ceramics, and this is true of emergent Durfee Underlined vessels found across the 267 north shore. It appears that the message communicated by these ceramics is quite different than that of the everted-lip pottery. This new practice spread more widely throughout the village, but it continued to be most popular in the practice community in which it first appears. The two distinct practice groups adopt each other’s unique emergent vessels, though in much lower proportions than those of their own. Ceramic messaging of these groups was mediated by a third group, which shared in both of these new practices equally, but which also seems more connected to external ties that brought in pottery from New York state communities. The largest local community of practice, which includes most houses in the lower village and three of the five houses in the upper village, was split between these newer nonlocal practice groups. By this time, household ties were strongly oriented to the later Humber Damiani village and various settlements in Simcoe County. The late introduction of exotic Durfee Underlined vessels of foreign production indicates that these latter two groups maintained stronger external ties with St. Lawrence and Northern New York groups, while the first group remains more isolated. With the introduction of two new longhouses and the expansion of the palisade to include them, local tradition practice of the village shifts as neck decoration becomes less popular. The amalgamation of the earlier, first practice group into one expanded longhouse on the plateau towards the end of the village occupation signals their diminishing influence. Although evertedlip ceramics are seen in small proportions on contemporary north shore sites, the practice disappeared with the demise of the Keffer village. Emergent Durfee Underlined vessels also diminished in proportions as communities from the Don valley eastwards moved out of the north shore. The importance of communicating the diverse identities expressed in newly emergent ceramics weakened with the abandonment of Keffer and surrounding villages. 268 9 Conclusions In this thesis, I reveal the independent and active nature of local and nonlocal ceramic practices of household social units, thus uncovering the multilevel nature of relational identity as reflected in joint communities of practice. I differentiate individual spheres of local tradition ceramic practice and discern co-existent, on-site, nonlocal tradition practices within the social and temporal village practice continuum through the use of social network and typological seriation analysis. The main focus of this research is the detection of multiple domains of emergent group identity as identified through ceramic relations. The underlying foundation of this research is the relational, fluid, contingent, and emergent nature of identity and society. This identity itself is intricately linked with social distance among people (Lyons and Clark 2008), which eschews boundaries in relationships, instead acknowledging their clinal nature. Social distance is in a constant state of flux between two entities as a result of unconscious and intentional agency within these relationships. Communities of practice entail a basic form of social relationship that occurs at multiple levels and is materialized in the archaeological record. Social network analysis has explored changing social relations within and beyond the Keffer village. I identified social relationships forming communities of practice through the identification of patterned spheres of practice as reflected in the networked ceramic data. Thus, network analysis is an appropriate tool for the detection of communities of practice within ceramic production and consumption across the Keffer village. I reconstructed the changing occupation and spatial organization of the Keffer village through the integration of multiple methodologies. Settlement pattern analysis combined with stratigraphic studies and the network analysis of ceramic connectivity revealed different periods of aggregation. In-depth research into specific house and midden ceramic type profiles verified network interpretations and added nuance to the sequence. This multilayered process resulted in an array of eight synthetically delineated stages illustrating the progression of the village’s construction and occupation. The combination of methodologies results in a nuanced developmental sequence documenting the internal and external ceramically affiliated relations and consequent population movement within the village. Further network analysis, combining village loci and pan-Iroquoian ceramics, implemented for both the local and nonlocal spheres, adds insight into external relations of the village site as a 269 historical and social entity, and also into the internal and external relations of each of its households and ceramic communities of practice. Local tradition network analyses led to the identification of reasonable and specific local origins for many of the original village households. Ceramic similarities among later houses and other southern Ontario Iroquoian villages document the general path of shared ceramic change at Keffer through its lifetime, ultimately suggesting the hypothetical destination of the final inhabitants as they moved on from the Keffer village. Network analysis in archaeology has been used almost exclusively during the past decade to examine regional and long-distance trade and interaction. In contrast, the exploration of the dynamics of close, matrilineally tied, kin-based social relationships within the village and local intra-drainage community performed in this thesis demonstrates the potential of social network analysis for multiscalar investigation—from the microcosm of household refuse deposition patterning and matrilineally based intra-village relations to macrolevel, pan-Iroquoian ceramic emulation and interaction. Analysis of two ceramic realms, local tradition and nonlocal tradition ceramic types, proved instrumental in the detection of two interrelated spheres of ceramic practice. Long-held, matrilineal, intra-lineage-based praxis of pottery production and consumption is displayed in the patterning of traditional ceramic types. At Keffer, ceramic distribution patterns reflect two primary communities of traditional Ontario Iroquoian ceramic practice within the village. These two groups are only partly spatially defined, but their influence is spread throughout households across the village, suggesting that social proximity in the form of matrilineal relations is not the prime factor in longhouse location. Change within this praxis over time is revealed through the application of the temporal patterning, as seen in village development, to the local tradition communities of practice. As discussed in Chapter 8, the analysis also identified unique nonlocal tradition ceramic practice communities within the village, as well as extra-village interaction spheres. Results of this work confirm the value of this approach in multiscalar applications when used in conjunction with the validation of visualized patterning through a secondary methodology. Large-scale, pan-Iroquoian networks affirm that Iroquoian peoples of southern Ontario, and beyond, cannot be lumped together into one larger group, or polity. Instead, each drainage-based community, and in fact each village and household, were characterized by distinct historical trajectories and fluid, constantly emerging social identity. Kinship-based social relations are at the core of this multifaceted identity (Birch 2008; Creese 2012; Sioui 1999). These complex, kin- 270 based relational identities are expressed by the separate, but associated, local and nonlocal pottery traditions. Individual, matrilineally based household social groups comprised the fifteenth-century Iroquoian centre of power and decision making (Creese 2010). These social groups within the Keffer village were highly involved in a complex network of locally oriented, kin-based ceramic communities of practice. Simultaneous involvement in independent transactions with longdistance partners, as seen in the incorporation of nonlocal ceramic practices into the local household practice, reflects the multiscalar and complex nature of these kin-based identities. The separation of the ceramic collection at Keffer into local tradition and nonlocal tradition types enabled the exposure of alternate yet related spheres of social interaction in terms of ceramic practice. Local tradition ceramics produced within long-standing matrilineal kinship networks reflect the habitus and the chaîne opératoire of the producers and the enculturation of the learners within domestic contexts. Shared practices therefore may reflect kin-related households within the village. The village itself is a fluid alliance of extended matrilineal families, each of which is constantly manipulating its relations and identity through agency. Membership within communities of ceramic practice, both local and nonlocal, does not reflect only one aspect of identity. The independent agency of the households within kin groups is evident in the diversity of external ties they hold, as seen in the distinct nonlocal ceramic collections. External ties and interaction patterns may thus be shared according to alternate guidelines prescribed by inclusion in these, or other, peer collectives. Communities of practice, in this case matrilineally based families, are fluid. Alliances and external relationships are changing. At some level of similarity, all Iroquoian ceramicists belong to the same community of practice and share in one larger identity, as “kinship is the core of Iroquoian society...through which all social relations were configured” (Birch 2008:207). This work, however, illustrates the multiplicity of relational identities that can be detected when identity is broken into its constituent parts, and it underscores the importance of rejecting the normative view of bounded culture, where the archaeological record of individual sites is constituted as a bounded and unified whole. Local tradition network ties suggest that the first settlers of Keffer arrived from a nearby community in the Don River drainage. This may have been prior to the establishment of widespread interaction among the Don valley communities and those of eastern Ontario, since no 271 nonlocal ceramics appear in this early settlement. They were followed soon after by additional Don valley inhabitants, and by the mid-life occupation of the village, newcomers bring indications of strengthening long-distance interaction with the communities of the St. Lawrence valley and the Eastern Iroquois. These long-distance relationships are materialized in multiple ceramic domains and reflect varying communities of ceramic practice. Ceramics of all five Haudenosaunee regions are found in most nonlocal ceramic collections at Keffer. They begin to appear after the first quarter of the fifteenth century and decline after the turn of the sixteenth century. Their presence likely does not indicate direct personal contact of north shore communities with those of the Haudenosaunee, it does but reflect some level of participation in this extensive north shore interaction sphere, as many north shore sites contain very similar ceramics. In contrast, the presence of exotics from the St. Lawrence valley in limited deposits likely suggests that the households who created these deposits may have held direct ties with communities to the east. The recognition of a new, previously unidentified ceramic practice at the Keffer village reflects the appearance within the community of expressions of newly emerging identities. Emergent vessels are characterized by two distinct and unrelated ceramic forms, the well-executed, everted-lip jar form, likely produced at the Keffer site, and the locally produced north shore pottery conforming to the Durfee Underlined type. Unlike the everted-lip vessels, the Durfee Underlined vessels belong to a pottery “subtype” seen in other north shore collections. Therefore, their presence does not indicate direct contact with the producers of the original Durfee Underlined type vessels. The everted-lip ceramics are almost unique to the Keffer village and communicate a very different identity and/or related consumption activity than any previously seen. The unique vessel form may also reflect new feasting activity. These ceramics are purposely employed to indicate one distinct aspect of this group identity. They are not indicative of external, ongoing trade networks because they are unique throughout Iroquoia to this practice group. The emergence of the everted-lip pottery reflects the creation of a new identity within the female potting community. It may also reflect change in the identity of related males, as they are the most prominent participants in feasting activities. The emergence, proliferation, and demise of this ceramic by the end of the Keffer occupation indicates that dynamic fluctuation in social relations and identity formation occurred with changing membership in the village community. The appearance of vessels of this type at the Parsons and 272 Draper sites may indicate direct interaction among the three villages, conventionally believed to be contemporary occupations. In contrast, the emergent Durfee Underlined vessels are found on many fifteenth-century sites across the north shore. The conformity in motif on these vessels may indicate specialized production within one site, but more research is needed to confirm this. The more widespread presence of this pottery indicates a more extensive interaction sphere for these vessels. Both of these emergent ceramics are the basis for differing ceramic practice communities at Keffer, which correspond with somewhat unique local tradition ceramic practices. Numerous layers of practice are performed simultaneously by members of the same household. There may be correlation of these layers in some cases, but in other instances potters use local and nonlocal ceramic practice to communicate flexible and disparate identities. This study has shown that the separation of ceramic collections into separate and unique practices reflecting different spheres of praxis aids in our understanding of the processes involved in the formation of multilevel identities. Each sphere is reflective of the existence of different levels in the clinal relationship of social distance. Ceramics produced within longstanding matrilineal kinship networks, termed local tradition ceramics, reflect the habitus of the producers in the chaînes opératoires and enculturation of learners within domestic contexts. The distinction between locally made and foreign-made ceramics helps us to distinguish between relationships. This study has also reinforced recent pronouncements by other researchers (Birch et al. 2017; Bursey 1996; Habiba et al. 2018; Mills 2017) on the relevance of ceramic typology in the study of regional relationships. In this instance, it has proven essential to the understanding of dynamics involved in the patterns detected, but not explained, through network analysis. Through a combination of network analysis, ceramic seriation, and settlement pattern study, a complex history of the Keffer settlement is unveiled. At Keffer, the first construction of the village occurs with the erection of a small longhouse or cabin, possibly a construction or specialpurpose cabin, on the central plateau by local peoples from nearby Don River drainage communities, perhaps the contemporary McNair village. Ceramics suggest a lack of external nonlocal interaction at this first cabin, unsurprising considering the labour-intensive focus of these first inhabitants. An influx of local Don drainage people follows after a short period of occupation. At this time, the first house is doubled in size and a second longhouse is added to the 273 plateau to accommodate these new residents. This growth is accompanied by the construction of two clusters of three longhouses each on the lower terrace. This newly expanded settlement is surrounded by a one-row village palisade, marking the social unification of the community. Inhabitants of the Don valley Risebrough and Watford communities or other nearby communities are likely settlers of these lower-terrace longhouses, as seen in associated early surface refuse deposits. Two of these new households in the low town share ceramic practices with the residents of the first cabin and may be kin from nearby settlements. With growth of the village population, the need for additional housing results in the demolition of the first longhouse and its replacement with three tightly spaced houses crowded on the plateau. Residents of the original longhouse, now occupying one of the newly built structures, are soon joined by a kin group in a newly constructed longhouses located close by, just outside the village palisade wall. Pottery of these two households displays shared practices and relations in the realms of both local and nonlocal ceramic practices. By this time, the small, household-associated surface middens are generally abandoned because new, shared hillside middens are required for the large amount of refuse produced by the growing number of households. Over the course of the occupation, three distinct but related spheres of local ceramic practice exist among the high town residents. The two households aligned along the west side of the plateau share ceramic profiles with at least one household in the low town. A second practice sphere with similar but slightly later ceramics appears on the east side of the plateau, and a third practice of marginally later local ceramics, reflecting the long-term occupation of the plateau, is seen towards the north of the plateau. These three ceramic practice spheres occur in three neighbouring households within one settlement cluster. At this point, the Keffer village is densely packed within the first palisade. There is little room for structural expansion and no room for additional longhouses. Therefore, the newest incoming household, though related by kinship to the village’s first occupants, as reflected in the form of their ceramic practice, must build outside the palisade. Genetic and ceramic similarities suggest people of this household are close biological relatives of the occupants of a small special-purpose cabin at Hidden Spring, located nearby, perhaps an agricultural cabin. This new longhouse is built in the closest proximity possible to their kin on the plateau surface. Many households from this middle period of the occupation at Keffer display their highest ceramic similarities with the small Logan site in the resource-rich higher lands of the Oak Ridges Moraine. Hidden Spring had 274 only two examples of nonlocal ceramics. With no nonlocal ceramics, Logan may also be a special-purpose site connected with Keffer. After the construction of a second longhouse exterior to the village wall, the entire original palisade was removed, and the palisaded area was expanded to incorporate both households. The large-scale remodelling and expansion of the village included the extension of several longhouses on both elevations. Additional longhouse realignments and new longhouse construction in the low town mark a period of great population growth in the village. These new villagers, like those before them, arrived from local Don River valley settlements. With little room left in the village area, one last longhouse is squeezed inside the palisade. Finally, another, the last to be built at the site, is erected in the remaining space outside the palisade. With these last households, ceramic practices shift and become more reminiscent of the later Wendat ceramics of the sixteenth and seventeenth centuries. This last household is accepted into the village proper with the expansion of the outer palisade around the house exterior. The occupation of this last house is brief. Soon after its inclusion within the palisade, the house is torn down, along with a second recently built house in the north end of the village and a third located in the village centre. Occupation of the village continued for some time after this event; however, the central planning which coincided with the expansion of the village now disappears. The relatively small size of the surface middens which appear at this time suggests that occupation of the village did not continue for long after the first houses were removed. Ceramic similarities suggest that the marginally later Damiani village, in the neighbouring Humber River watershed, was the destination of at least a portion of the Keffer population. Ties with northern sites indicate that a strengthening of relations in that direction was also occurring at the end of the fifteenth century. Planning objectives of both the household and the community level changed over time at Keffer. This began with the first houses being planned at the individual household level, with no view to longer-term community growth. This is apparent in the erection of this first house diagonally across the plateau in an orientation that precluded large-scale development in the area. Refuse deposits in this early period also reflect a lack of long-term planning, as small surface middens are located across the low town area. In the low town, the alignment of two clusters that appeared in the first stage suggests local family decisions, not a unified community–level process. As the village grew, increased planning occurred. The erection of Palisade 2 and the 275 realignment of new houses at the community level reflect community planning. At least one of these houses was soon dismantled, and the occupants relocated to the neighbouring house, which expanded overtop its footprint. With the fluid nature of households in Iroquoian communities, this change could simply reflect a change in household configuration. It may not be indicative of poor planning. As Birch mentions for Mantle (2010), use of large, community-based middens, which begins with the first infilling of the village, also suggests village-level organization. Village-level planning, however, was sporadic, as continual change and fluid, perhaps unplanned, village membership was the norm at Keffer. The unplanned and flexible nature of the Keffer village occupation may have been dictated to some extent by appeals or requests of close kin living in local villages. Keffer was a small hamlet that continued to sporadically attract small numbers of new residents. Local tradition ceramics indicate that most newcomers originated in the local Don River drainage area. Nonlocal tradition ceramics suggest that the groups who arrived after the first households already had wellestablished long-distance ties to communities in the east. This is a period when interaction across the north shore was well established and expanding and when interaction was intensifying with Iroquoian communities in Quebec and New York. During the period of Damiani’s first occupation, ceramics characteristic of long-distance interaction continued to appear, but in lower proportions, indicating weakening ties with people of the St. Lawrence valley and Eastern Iroquois area. This study has shown that newcomers placed outside the village palisade may not, in fact, be in any way socially distant. Pragmatism may have been the deciding factor in the location of new structures, rather than social exclusion, as was previously believed (for example, by Finlayson [1985]). Space may be a determinant not just in village-level organization, but also in decisions regarding possible destinations of households as they relocated outside of the village into new communities. It is possible that these communities, like Keffer, experienced processes of longterm, unplanned, and casual growth fuelled by the immigration of various kin-related households over the course of decades. This pattern of growth contrasts sharply with the unique coalescence seen at the contemporary Rouge–Duffins Draper site. While in the Keffer village growth was at least partially based on movement of households of nearby villages into a spatial relationship with existing members after the initial settlement construction, no episode of expansion is sizable enough to represent the incursion of entire village populations. Village growth as seen at Keffer 276 may be more typical of Don–Humber drainage villages, reflecting the complex web of interdrainage ceramic relations negotiated at the level of the family and household, as illustrated in the network graphs, and controlled by the agency of each household and each family. As Birch and Williamson (2013) suggest, each community followed its own path in the course of community aggregation and separation. The creation, growth, and final abandonment of the Keffer village do not represent the lifespan and movement of a single, united community. This village is instead an alliance of extended matrilineal families originating in related but separate north shore Iroquoian communities from within the local drainage system. Village growth and decline depended on the fluid agency of the individual household and those of its kin as they moved into, within, and out of the village, changing alliances and constantly increasing and decreasing the social distance of each relationship. Settlement clusters may have originated as kin-based units, but long-term occupation over the course of several decades eradicated evidence of this early habitation. The last period of occupation in the archaeological record contains traces of only the last longhouse residents. At Keffer, the last houses within cluster formations do not share either local or nonlocal ceramics practices, indicating that kin-based ties may not be the basis of this cohabitation. This residential pattern may instead be subject to local social and practical contingencies. Without further research in this area, we can no longer assume that kin ties are the core of cluster formation. Though it is not the largest north shore site, the Keffer village may be the longest-occupied settlement in the region. Keffer alone shows evidence of repeated reconstruction of longhouses within the same footprint, most likely due to structural deterioration. The rebuilding, in lieu of relocation, of the village may be due to its beneficial location. Situated at the edge of the moraine, and south of the passageway between the Logan and Beeton sites, Keffer had access to the vast resources of the till plain agricultural fields, a large river drainage basin, and the forests and fauna of the uplands. Ceramic similarity indicates a close connection between Keffer and Logan. The small Logan settlement may have served as a hunting cabin for Keffer or as a trading post for the exchange of food resources and deer between Keffer and recently settled communities in Simcoe County, north of the Oak Ridges Moraine. 277 This study contributes directly to three domains of scholarship. Most noticeably, it augments the growing empirical record pertaining to the culture history of the greater pre-contact Iroquoian world and the occupation history and dynamics of Iroquoian lifeways on the north shore of Lake Ontario, while providing insights into the small-scale relations of one village. This work also introduced the integration of alternative theoretical treatments of this empirical data through some novel, and increasingly popular, approaches and methodologies. In the paradigm employed here, identity is seen as relational, social distance is presented as a measure of social relations, and identity emerges in the constructs of communities of practice. The integration of a communities of practice approach at the level of the household has shown that identity and agency are accessible at the most intimate scales. I have shown that communities of practice underwrote the construction, change, and continuity of identity. The relational and multiscalar nature of identity as mediated by ceramic practices is amenable to social network analysis. While network analysis of macro-regional and regional-scale relations is an area of widening archaeological interest, network analysis at the microscale of intra-village relations as employed here is novel and has been proved effective in detecting micro-scale social relations and likely kin-based identities. Examination of relations at this scale is essential to interpreting the culture of fifteenth-century Iroquoians and that of other small-scale societies, based on the household (Creese 2010) and organized through the “vast system of kinship” (Sioui 1999:xi). With some recent exceptions, the Iroquoian village is commonly depicted as a complete social unit in and of itself. Analysis of the Keffer communities of practice demonstrates that this conflation masks the true nature of the social interactions and relationships occurring both within the village, within settlement clusters in the village, and in the external relationships of households, communities, and the entire village. My analysis has demonstrated that even co-eval families within the same cluster maintained differing external interaction spheres. Each cluster is not inevitably representative of an individual village moving into the site. The reconfigurations of internal village layout at Keffer, which occurred with the expansion of the palisade and the integration of previously external households, is reflective of social fluidity, as previously excluded kin-based households were brought into the social relations of the village. 278 The possibilities for future research are endless. Future work could place more emphasis on social relations as the foundation and mechanism bonding Iroquoian society at all levels, from the household, to historic-period Nations, to the Wendat of today. Such work will necessitate division and analysis of the archaeological record into pieces that can reflect these relations at all levels. Social network analysis and the concept of emergent “assemblages” hold great promise for the examination of social relations at multiple scales. Network analysis has provided an effective tool to explore social relationships within the microcosm of the village. Further investigations of additional forms of material culture and reanalysis of more detailed, intra-house provenance information available in recent site reports will further improve the study of smallscale social relations at the level of the nuclear family. One of the core goals of this work was to offer a contribution to the growing body of research that promotes a greater understanding of the southern Ontario Iroquoians as active agents of culture. The in-depth ceramic seriation produced here has produced a solid relational chronology which may provide a model for further refinement of carbon dating through Bayesian analysis. The Iroquoian archaeological record is almost unsurpassed, due to the short-term, singleoccupation character of its sites, in its ability to afford research opportunities at a microlevel. Thus, Iroquoian archaeology provides a great resource for the study of social relations in heterarchical agricultural societies. References Abbott, David R. 2000 Ceramics and Community Organization among the Hohokam. University of Arizona Press, Tucson. Abel, Timothy J. 2002 Recent Research on the Saint Lawrence Iroquoians of Northern New York. Archaeology of Eastern North America 30:137-154. 2016 The Iroquoian Occupations of Northern New York: A Summary of Current Research. Ontario Archaeology 96:65-75. 2017 The Iroquoian Occupations of Northern New York. Electronic document, https://www.academia.edu/30484816/The_Iroquoian_Occupations_of_Northern_New_York. Adams, Nicholas R. 2003 The Arbor Ridge Site: A Study in Settlement Dynamics and Population Movement during the Fifteenth Century at the Eastern End of Lake Ontario. Master’s thesis, University of Leicester. Agriculture Canada 1955 Soil Survey of York County. Report 19. Electronic document, http://sis.agr.gc.ca/cansis/publications/surveys/on/on19/index.html. Allen, Kathleen M. S. 2009 Temporal and Spatial Scales of Activity among the Iroquois: Implications for Understanding Cultural Change. In Iroquoian Archaeology and Analytic Scale, edited by L. E. Miroff and T. D. Knapp, pp. 153-177. University of Tennessee Press, Knoxville. Anderson, Jacob M. 2009 The Lawson Site: An Early Sixteenth Century Neutral Iroquoian Fortress. Special Publication No. 2, Museum of Ontario Archaeology, London, Ont. Anthony, David W. 1990 Migration in Archaeology: The Baby and the Bathwater. American Anthropologist 92:895-914. ARA (Archaeological Research Associates) 2003 Stage 3&4 Archaeological Assessment, Lougheed Site (BdGw-13) City of Barrie, Simcoe Township, Former Township of Innisfil. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. Arnold, Dean E. 1991 Ethnoarchaeology and Investigations of Ceramic Production and Exchange: Can We Go Beyond Cautionary Tales? In The Ceramic Legacy of Anna O. Shepard, edited by R. L. Bishop and F. W. Lange, pp. 321-345. University Press of Colorado, Niwot, Colo. 1998 Andean Ceramic Technology: An Ethnoarchaeological Perspective. In Ceramic Production in the Andes: Technology, Organization, and Approaches, edited by I. Shimada, pp. 353-367. MASCA Research Papers in Science and Archeology 15. Philadelphia. ASI (Archaeological Services Inc.) 1998 Final Report on Archaeological Salvage Excavation of the Murphy-Goulding Site (AlGu-3), Town of Richmond Hill, Regional Municipality of York. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. 279 280 2001 The Stage 2 to Stage 4 Salvage Excavations at the Jarrett-Lahmer Site (AlGv-18). Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. 2004 Report on Stage 4 Salvage Excavation of the Serena Site (AhGx-274) Allison Estates, Subdivision (25T-91014), City of Hamilton, Regional Municipality of Hamilton, Ontario. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. 2005a The Stage 3–4 Salvage Excavations of the Burkholder 2 Site (AlGt-35) Lot 8, Concession 9, Geographic Township of Markham Box Grove Secondary Planning Area, Box Grove Hill Development Lands, Town of Markham, Regional Municipality of York, Ontario. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. 2005b The Archaeology of the Wellington Site (BcGw-55): A Report on the Stage 4 Salvage Excavations of the Wellington Site, Holly Secondary Planning Area (43T-92023), Part of the East Half of Lot 3, Concession 12, City of Barrie, Simcoe County, Ontario. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. 2006a The Stage 4 Salvage excavation of the Baker site (AkGu-15) Lot 11 Concession 2 (WYS) Block 10 O.P.A. 400 Former Township of Vaughan, City of Vaughan, Regional Municipality of York, Ontario. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. 2006b The Archaeology of the Dykstra Site (BbGw-5): A Report on the Stage 4 Salvage Excavations of the Holly Secondary Planning Area (43T-92026), Part of the Northwest Half of Lot 2, Concession 12, City of Barrie, Simcoe County, Ontario. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. 2006b Stages 1–3 Investigation of the Mantle Site (AlGt-334) Part of Lot 33, Concession 9, Town of Whitchurch-Stouffville, Regional Municipality of York, Ontario. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. 2007 McGaw Site (AlGu-88) Archaeological Interpretive Program Results of the 2003 Field Season. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. 2008a Report on the Stage 3–4 Salvage Excavation of the Alexandra Site (AkGt-53), Draft Plan of Subdivision SC-T20000001 (55T-00601), Geographic Township of Scarborough, now in the City of Toronto, Ontario. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. 2008b The Stage 4 Salvage Excavation of the Orion Site (AlGu-45), Lot 56, Concession 1W.Y.S. Town of Richmond Hill, Regional Municipality of York, Ontario. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. 2009 The Archaeology of the Holly Site (BcGw-58): Stage 4 Salvage Excavation of the Holly Site, Dykstra Subdivision, Holly Secondary Planning Area (43T-92026), Part of the Northeast Half of Lot 2, Concession 12, City of Barrie, Simcoe County, Ontario. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. 2010a Report on the Salvage Excavation of the Antrex Site (AjGv-38), City of Mississauga, Regional Municipality of Peel, Ontario. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. 2010b The Archaeology of the Hidden Spring Site (AlGu-368): Stage 4 Salvage Excavation of the Hidden Spring Site, Oxford West Subdivision Development, Part of Lots 13-16 and 37-40, 281 Registered Plan 1931, Town of Richmond Hill, Regional Municipality of York, Ontario. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. 2010c The Archaeology of the Robb Site (AlGt-4): A Report on the Stage 4 Mitigative Excavation of the Angus Meadows Subdivision 19T-95030 (Revised), Part of Lot 1, Concession 8, Town of Markham, Regional Municipality of York, Ontario. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. 2010d The Archaeology of the Walkington 2 Site (AlGu-341): A Report on the Stage 3 and Stage 4 Mitigative Excavations of the of the Nine-Ten Property, Draft Plan of Subdivision 19T-95066 (Revised) Part of Lots 16 and 17, Concession 2, City of Vaughan, Regional Municipality of York, Ontario. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. 2011 The Stage 3–4 Archaeological Excavation of the Hope Site (AlGv-199), Draft Plan of Subdivision 19T-02V07 and 19T-02V08, City of Vaughan, Regional Municipality of York, Ontario. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. 2012a Stage 3&4 Archaeological Assessment Joseph Picard Site (AlGs-376), Highway 407 East, Lot 32, Concession IV Whitby Township, Former Ontario County, Regional Municipality of Durham, Preliminary Excavation Report. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. 2012b Report on the Stage 3–4 Mitigative Excavation of the McNair Site (AlGu-8), City of Vaughan, Regional Municipality of York, Ontario. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. 2014a Report on Stage 3—4 Mitigative Excavation of the Damiani Site (AlGv-231), City of Vaughn, Regional Municipality of York, Ontario (draft). Report on file, Archaeological Services Inc., Toronto. 2014b The Archaeology of the Mantle Site (AlGt-334): Report on the Stage 3–4 Mitigative Excavation of Part of Lot 22, Concession 9, Town of Whitchurch-Stouffville, Regional Municipality of York, Ontario. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. 2014c The Archaeology of the Yatsihsta’ Site (AlGs-452): A Report on the Stage 4 Mitigative Excavation of the Yatsihsta’ Site (AlGs-452) Highway 407 East, Lot 1, Concession VI, Pickering Township and Lot 35, Concession VI, Whitby Township, Former Ontario County. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. 2015 The Archaeology of the Damiani Site (AlGv-231): A Report on the Stage 4 Mitigative Excavation of the Damiani Site (AlGv-231), Part of Lot 25, Concession 7, Pine Heights Estates, Blocks 40 and 47, City of Vaughan, Regional Municipality of York, Ontario. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. Bamann, Susan, Robert Kuhn, James Molnar and Dean Snow 1992 Iroquoian Archaeology. Annual Review of Anthropology 21:435-460. Barth, Fredrik 1969 Introduction. In Ethnic Groups and Boundaries: The Social Organization of Culture Difference, edited by F. Barth, pp. 9-38. Little Brown, Boston. Beauchamp, William M. 1984 Curious Iroquois Pottery. Science 23 282 1900 Aboriginal Occupation of New York. Bulletin of the New York State Museum, University of the State of New York, Albany. Beisaw, April M. 2010 Memory, Identity, and NAGPRA in the Northeastern United States. American Anthropologist 112(2):244-256. Bekerman, André and Gary A. Warrick (editors) 1995 Origins of the People of the Longhouse: Proceedings of the 21st Annual Symposium of the Ontario Archaeological Society. Ontario Archaeological Society, Toronto. Bentley, G. Carter 1987 Ethnicity and Practice. Comparative Studies in Society and History 29(1):24-55. Betts, Matthew W. 2008 Subsistence and Culture in the Western Canadian Arctic: A Multicontextual Approach. Canadian Museum of Civilization, Gatineau, Quebec. Birch, Jennifer 2008 Rethinking the Archaeological Application of Iroquoian Kinship. Canadian Journal of Archaeology 32(2):194-213. 2010 Coalescent Communities in Iroquoian Ontario. Unpublished Ph.D. dissertation, Department of Anthropology, McMaster University, Hamilton, Ontario. 2011 Unpublished Report on Keffer Site Wall Post Density. Report on file, Department of Anthropology, University of Toronto at Mississauga. 2012 Coalescent Communities: Settlement Aggregation and Social Integration in Iroquoian Ontario. American Antiquity 77(4):646-670. 2014 Post Mould Densities of the Keffer Longhouses. Report on file, Museum of Ontario Archaeology, London, Ont. 2016a Geopolitics and Dimensions of Social Complexity in Ancestral Wendake c. A.D. 1450-1600. Ontario Archaeology 96:35-46 2016b Interpreting Iroquoian Site Structure through Geophysical Prospection and Soil Chemistry: Insights from a Coalescent Community in Ontario, Canada. Journal of Archaeological Science: Reports 8:102-111. 2018 Iroquoian Archaeology. In Encyclopedia of Global Archaeology, edited by C. Smith, pp. 1-7. Springer International Publishing, Cham. Birch, Jennifer and John P. Hart 2018 Social Networks and Northern Iroquoian Confederacy Dynamics. American Antiquity 83(1):1333. Birch, Jennifer and Ronald F. Williamson 2012 The Mantle Site: An Archaeological History of a Huron-Wendat Community. AltaMira Press, Lanham, MD. 2013 Organizational Complexity in Ancestral Wendat Communities. In From Prehistoric Villages to Cities: Settlement Aggregation and Community, edited by J. Birch, pp. 153-178. Routledge, New York. 283 2018 Initial Northern Iroquoian Coalescence. In The Archaeology of Villages in Eastern North America, edited by Jennifer Birch and Victor D. Thompson, pp. 89-105. University of Florida Press, Gainesville. Birch, Jennifer, Robert B. Wojtowicz, Aleksandra Pradzynski and Robert H. Pihl 2017 Multi-Scalar Perspectives on Iroquoian Ceramics: Aggregation and Interaction in Pre-Contact Ontario. In Process and Meaning in Spatial Archaeology: Investigations into Pre-Columbian Iroquoian Space and Place, edited by E. E. Jones and J. L. Creese, pp. 111-144. University Press of Colorado, Boulder. Blair, E. 2015 Making Mission Communities: Population Aggregation, Social Networks, and Communities of Practice at 17th Century Mission Santa Catalina de Guale. Ph.D. dissertation, University of California, Berkeley. Borgatti, Stephen P. 2002 NetDraw Software for Network Visualization. Analytic Technologies, Lexington, KY. Borgatti, Stephen P. and Daniel S. Halgin 2011 On Network Theory. Organization Science 22(5):1168-1181. Borgatti, Stephen P., Jeffrey C. Johnson and Martin G. Everett 2013 Analyzing Social Networks. Sage, Los Angeles, CA. Bourdieu, Pierre 1977 Outline of a Theory of Practice. Cambridge University Press, Cambridge. Bowser, Brenda J. 2000 From Pottery to Politics: An Ethnoarchaeological Study of Political Factionalism, Ethnicity, and Domestic Pottery Style in the Ecuadorian Amazon. Journal of Archaeological Method and Theory 7(3):219-248. Bowser, Brenda J. and John Q. Patton 2008 Learning and Transmission of Pottery Style: Women's Life Histories and Communities of Practice in the Ecuadorian Amazon. In Cultural Transmission and Material Culture: Breaking Down Boundaries, edited by M. T. Stark, B. J. Bowser and L. Horne, pp. 105-129. University of Arizona Press, Tucson. Boyle, David 1889 Township of Nottawasaga. Annual Archaeological Report for Ontario 1888-1889:4-11. 1891 The Southwold Earthwork. Annual Archaeological Report for Ontario 1890-1891:8-10. 1892 Malahide. Annual Archaeological Report for Ontario, 1891:11-12. 1896 Ontario Earthworks. Annual Archaeological Report for Ontario 1895-1896:37-40. 1897 Ontario Mounds. Annual Archaeological Report for Ontario, 1896-7:14-37. 1897 Ossuary in Beverly. Annual Archaeological Report for Ontario, 1896-7:46. 1898 Earthwork in Malahide. Annual Archaeological Report for Ontario, 1897-98:44. 1902 The Yellow Point Mound. Annual Archaeological Report for Ontario, 1901:25-29. Bradley, James W. 1987 Evolution of the Onondaga Iroquois: Accommodating Change, 1500-1655. Syracuse University Press, Syracuse, N.Y. 284 Brainerd, George W. 1951 The Place of Chronological Ordering in Archaeological Analysis. American Antiquity 16(4):301314. Braudel, Fernand 1980 On History. University of Chicago Press, Chicago. Braun, David P. 1983 Pots as Tools. In Archaeological Hammers and Theories, edited by J. A. Moore and A. S. Keene, pp. 107-134. Academic Press, New York. 1991 Why Decorate a Pot? Midwestern Household Pottery, 200 B.C.-A.D. 600. Journal of Anthropological Archaeology 10(4):360-397. Braun David P., and Stephen Plog 1982 Evolution of “Tribal” Social Networks: Theory and Prehistoric North American. American Antiquity 47:504-525. Brown, Judith K. 1970 Economic Organization and the Position of Women among the Iroquois. Ethnohistory 17:151167. 1975 Iroquois Women: An Ethnohistoric Note. In Toward an Anthropology of Women, edited by Rayna R. Reiter, pp. 235-251. Monthly Review Press, New York. Brughmans, Tom 2010 Connecting the Dots: Towards Archaeological Network Analysis. Oxford Journal of Archaeology 29(3):277-303. Buck, C. E., J. A. Christen and G. N. James 1999 BCal: an on line Bayesian radiocarbon calibration too. Internet Archaeology 7. (http//intarch.ac.uk/journal/issue7/buck/). Bull, Steven 1987 Notes on an Investigation of the Relation Between Whole Vessels and Their Rim Sherds. Arch Notes Sep/Oct:29-30 1989 Reconstructed Ceramic Pots: Relationships Between Whole Vessels and Their Rim Sherds. Canadian Journal of Archaeology 13:219-222. Bursey, Jeffrey A. ca. 1990 Ceramic data forms on file, Department of Anthropology, University of Toronto at Mississauga. 1993 Prehistoric Huronia: Relative Chronology Through Ceramic Seriation. Ontario Archaeology 55:3-34. Bush, D. R. 1976 The Chronological Position of the CRS Site, Simcoe County, Ontario. Ontario Archaeology 28:17-32. Caldwell, Joseph R. 1964 Interaction Spheres in Prehistory. In Hopewellian Studies, edited by J. R. Caldwell and R. L. Hall, pp. 133-143. Illinois State Museum, Scientific Papers. Illinois State Museum, Springfield, Ill. 285 Calhoun, Craig J. 1995 Critical Social Theory: Culture, History, and the Challenge of Difference. Blackwell, Cambridge, Mass. Cameron, Catherine M. 2013 How People Moved among Ancient Societies: Broadening the View115(2):218-231. Carr, Christopher 1985 Introductory Remarks on Artifact Analysis. In For Concordance in Archaeological Analysis: Bridging Data Structure, Quantitative Technique, and Theory, edited by C. Carr, pp. 502-508. Westport Publishers, Kansas City. 1995 A Unified Middle-Range Theory of Artifact Design. In Style, Society and Person: Archaeological and Ethnological Perspectives, edited by C. Carr and J. E. Neitzel, pp. 171-258. Plenum, New York. Centola, Damon 2015 The Social Origins of Networks and Diffusion120(5):1295-1338. Chapdelaine, Claude 1989 Le Site Mandeville à Tracy, Variabilité Culturelle des Iroquoiens du Saint Laurent. Signes des Ameriques 7. Recherches amérindiennes au Québec, Montréal. 1991 Poterie, Ethnicité et Laurentie Iroquienne. Recherches amérindiennes au Québec 21:44-52. Chapman, L. J. and D. F. Putman 1984 The Physiography of Southern Ontario. Ontario Research Foundation, Toronto. Chernela, Janet M. 1992 Social Meaning and Material Transaction: The Wanano-Tukano of Brazil and Colombia. Journal of Anthropological Archaeology 11(2):111-124. Childe, V. Gordon 1929 The Danube in Prehistory. Oxford University Press, Oxford 1950 The Urban Revolution. The Town Planning Review 21(1):3-17. Clermont, Norman, Claude Chapdelaine and Georges Barré 1983 Le Site Iroquoien de Lanoraie: Témoignage d'une Maison-Longue. Signes des Amériques 3. Recherches amérindiennes au Québec, Montréal. Coe, Michael D. 1962 Mexico. Frederick A. Praeger, New York. Cohen, Ronald 1978 Ethnicity: Problem and Focus in Anthropology. Annual Review of Anthropology 7:379-403. Collar, Anna, Fiona S. Coward, Tom Brughmans and Barbara Mills 2015 Networks in Archaeology: Phenomena, Abstraction, Representation. Journal of Archaeological Method and Theory 22(1):1-32. Comaroff, John L. and Jean Comaroff 1992 Ethnography and the Historical Imagination. Studies in the ethnographic imagination. Westview Press, Boulder. Cooper, Martin 286 2016 "In Order to Bring Them to Trade": Neutral Exchange during the Sixteenth Century. In Contact in the 16th Century: Networks among Fishers, Foragers, and Farmers, edited by B. Loewen and C. Chapdelaine, pp. 257-267. Canadian Museum of Civilization, Gatineau, Quebec. Crawford, Gary W. 2003 The Wallace Site (AkGx-1), 1984 and 1985. Manuscript on file, Department of Anthropology, University of Toronto at Mississauga. Crawford, Gary W. and David G. Smith 1996 Migration in Prehistory: Princess Point and the Northern Iroquoian Case. American Antiquity, Vol. 61:782-790. 2002 Early Late Woodland in Southern Ontario: An Update (1996-2000). In Northeast SubsistenceSettlement Change: A.D. 700-1300 by John P. Hart and Christina B. Rieth. New York State Museum, University of the State of New York, The State Education Department, Albany. Crawford, Gary W., David G. Smith and Vandy E. Bowyer 1997 Dating the Entry of Corn (Zea mays) into the Lower Great Lakes Region. American Antiquity 62:112-119. Crawford, Gary W., David G. Smith, Joseph R. Desloges and Anthony M. Davis 1998 Floodplains and Agricultural Origins: A Case Study in South-Central Ontario. Journal of Field Archaeology 28(2):123-137. Creese, John L. 2010 Deyughnyonkwarakda—“At the Wood’s Edge”: The Development of the Iroquoian Village in Southern Ontario A.D. 900-1500. Unpublished Ph.D. dissertation, Department of Anthropology, University of Toronto. 2012 The Domestication of Personhood: A View from the Northern Iroquoian Longhouse. Cambridge Archaeological Journal 22(3):365-386. 2013 Making Pipes and social persons at the Keffer site: A Life History Approach. Paper presented at the 78th Annual Meeting of Society for American Archaeology, Honolulu 2014 Wendat Smoking Pipes and Social Networks on the North Shore. Paper presented at the 41st Symposium of the Ontario Archaeological Society, Peterborough, Ontario. 2016 Making Pipes and Social Persons at the Keffer Site: A Life History Approach. In Perspectives on the Archaeology of Pipes, Tobacco and other Smoke Plants in the Ancient Americas, edited by E. A. Bollwerk and S. Tushingham, pp. 27-49. Springer, New York. Crown, P. L. 2014 The Archaeology of Crafts Learning: Becoming a Potter in the Puebloan Southwest. Annual Review of Anthropology 43:71-88. Cunningham, Jerimy J. 2001 Ceramic Variation and Ethnic Holism: A Case Study from the "Younge-Early Ontario Iroquoian Border" in Southwestern Ontario. Canadian Journal of Archaeology 25:1-27. Damkjar, Eric R. 1982 The Coulter Site and Late Iroquoian Migrations to the Upper Trent Valley. Master’s thesis, Simon Fraser University, Burnaby, B.C. DeLanda, M. 287 2006 A New Philosophy of Society: Assemblage Theory and Social Complexity. Continuum, London. 2016 Assemblage Theory. Edinburgh University Press, Edinburgh. de Nooy, Wouter, Andrej Mrvar and Vladimir Batagelj 2011 Exploratory Social Network Analysis with Pajek. Rev. and expanded 2nd ed. Structural analysis in the social sciences. Cambridge University Press, Cambridge. Dermarkar, Susan 2012 The Keffer Site (AkGv-14): An Overview of Our Current Archaeological Knowledge. Paper presented at the Annual Symposium of the Ontario Archaeological Society, Windsor, Ontario. 2013 North Shore Interaction in 15th Century Iroquoia. Paper presented at the Annual Symposium of the Ontario Archaeological Society. Niagara Falls, Ont. 2013 International Interaction in the 15th Century: The Ancestral Wendat Keffer Site and Its Ties to the Northeastern New York State Iroquois. Paper presented at the Iroquois Conference, Cortland, New York. 2014 A Preliminary Look at the Definition, Spatial Distribution and Cultural Significance of the Lalonde High Collar Ceramic Type. Paper presented at the Annual Conference of the Canadian Archaeological Association, Calgary. 2014 The Materialization of Interaction, the Land Between and What I Learned about Social Network Analysis at the Keffer Site. Paper presented at the Annual Symposium of the Ontario Archaeological Society, Peterborough, Ont. 2016 Explorations of Ethnicity, Identity and Ethnogenesis at the 15th Century Iroquoian Keffer Site. Paper presented at the Annual Conference of the Canadian Archaeological Association, Whitehorse. Dermarkar, Susan, Jennifer Birch, Termeh Shafie, John P. Hart and Ronald F. Williamson 2016 St. Lawrence Iroquoians and Pan-Iroquoian Social Network Analysis. Ontario Archaeology 96:87-103. Diaz-Andreu, Margarita 2015 Ethnic Identity/Ethnicity and Archaeology. In International Encyclopedia of the Social & Behavioral Sciences, edited by J. D. Wright, pp. 4817-4821. Second edition. Elsevier, Amsterdam. Dobres, Marcia-Anne 2000 Technology and Social Agency: Outlining a Practice Framework for Archaeology. Social archaeology. Blackwell Publishers, Oxford. Dobres, Marcia-Anne and John E. Robb 2000 Agency in Archaeology: Paradigm or Platitude? In Agency in Archaeology, edited by M.-A. Dobres and J Robb, pp. 3-18. Routledge, New York Dodd, Christine F. 1984 Ontario Iroquois Tradition Longhouses. National Museum of Man, Ottawa. Dodd, Christine F., Dana R. Poulton, Paul A. Lennox, David G. Smith and Gary A. Warrick 1990 The Middle Ontario Iroquoian Stage. In The Archaeology of Southern Ontario to A.D. 1650, edited by C. J. Ellis and N. Ferris, pp. 321-360. Occasional Publication No. 5. London Chapter, Ontario Archaeological Society, London, Ont. 288 Dorland, Steven G. H. 2016 Practicing Informal Apprenticeship: A Study of Learning Landscapes in 15th Century Great Lakes Potting Communities. Paper presented at the Annual Meeting of the Society for American Archaeology. 2016 Learning Landscapes in the Great Lakes Region: An Analysis of 15th Century Ceramic Assemblages. Paper presented at the Monthly Meeting of the Ontario Archaeological Society, Toronto Chapter. 2017 Were Pre-contact Children Really Innovators? Paper presented at the Annual Meeting of the Canadian Archaeological Association 2018 Childhood Learning on Turtle Island: A Ceramic Analysis of Late Woodland Villages in the Lower Great Lakes Region. Unpublished Ph.D. dissertation, Department of Anthropology, University of Toronto. Drennan, Robert D. 1984 Long-Distance Movement of Goods in the Mesoamerican Formative and Classic. American Antiquity 49(1):27-43. DRPA (D. R. Poulton and Associates Inc.) 1996a The Stage 3 Archaeological Assessment of the Jarrett-Lahmer Site (AlGv-18), Draft Plan 19T95079, City of Vaughan, Ontario. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. 1996b The 1992–1993 Stage 3–4 Archaeological Excavations of the Over Site (AlGu-120) (W.P. 23389-00), vol. 1. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. Duff, Andrew I. 2002 Western Pueblo Identities: Regional Interaction, Migration, and Transformation. University of Arizona Press, Tucson. Dupras, Tosha L. and David G. Pratte 1998 Craniometric Study of the Parsons Crania from Midden 4/Feature 245. Ontario Archaeology 65/66:140-145. Dyck, Ian 2004 Private and Public Passions: James Pendergast's Archaeology and the National Museum of Canada. In A Passion for the Past: Papers in Honour of James F. Pendergast, edited by J. V. Wright and J.-L. Pilon, pp. 5-48. Canadian Museum of Civilization, Gatineau, Quebec. Ellis, Christopher J. and Neal Ferris (editors) 1990 The Archaeology of Southern Ontario to A.D. 1650. Occasional Publication No. 5. London Chapter, Ontario Archaeological Society, London, Ont. Emberling, Geoff 1997 Ethnicity in Complex Societies: Archaeological Perspectives. Journal of Archaeological Research 5(4):295-344. Emerson, J. Norman 1954 The Archaeology of the Ontario Iroquois. Ph.D. dissertation, University of Chicago. 1959 A Rejoinder upon the MacNeish-Emerson Theory. Pennsylvania Archaeologist 29:98-107. 1959 The Bosomworth Site. Ontario History 51(1):61-63. 289 1998 The Research Potential of Huronia (1965). In Five Manuscripts by Dr. John Norman Emerson, compiled by C. Garrad. Special Publication, The Ontario Archaeological Society 13-14. Toronto. Emerson, J. Norman and R. E. Popham 1952 Comments on "The Huron and Lalonde Occupations of Ontario". American Antiquity 18:162164. Engelbrecht, William E. 1969a Barker Site Regrouped Ceramic Data. Electronic database, The Digital Archaeological Record (tDAR), https://core.tdar.org/document/373025/barker-site-regrouped-ceramic-data, accessed July 27, 2016. 1969b Simmons Site Regrouped Ceramic Data. Electronic database, The Digital Archaeological Record (tDAR), https://core.tdar.org/document/373067/simmons-site-regrouped-ceramic-data, accessed July 27, 2016. 1970a Cornish Site Regrouped Ceramic Data. Electronic database, The Digital Archaeological Record (tDAR), https://core.tdar.org/document/373026/cromwell-site-regrouped-ceramic-data, accessed July 27, 2016. 1970b Powerhouse Site Regrouped Ceramic Data. Electronic database, The Digital Archaeological Record (tDAR), https://core.tdar.org/document/373006/, accessed July 27, 2016. 1971 A Stylistic Analysis of New York Iroquois Pottery. Ph.D. dissertation, University of Michigan. 1974a Englebert Site Regrouped Ceramic Data. Electronic database, The Digital Archaeological Record (tDAR), https://core.tdar.org/document/373076/englebert-site-regrouped-ceramic-data, accessed July 27, 2016. 1974 The Iroquois: Archaeological Patterning on the Tribal Level. World Archaeology 6(1):52-65. 1974b Silverhouse Site Regrouped Ceramic Data. Electronic database, The Digital Archaeological Record (tDAR), https://core.tdar.org/document/373006/, accessed July 27, 2016. 1977 Ellis Site Regrouped Ceramic Data. Electronic database, The Digital Archaeological Record (tDAR), https://core.tdar.org/document/373051/ellis-site-regrouped-ceramic-data, accessed July 27, 2016. 1978 Kleis Site Ceramic Data. Electronic database, The Digital Archaeological Record (tDAR), https://core.tdar.org/document/375156/kleis, accessed July 27, 2016. 1985 New York Iroquois Political Development. In Cultures in Contact: The European Impact on Native Cultural Institutions in Eastern North America A.D. 100-1800, edited by William Fitzhugh. Anthropological Society of Washington Series. Smithsonian Press, Washington and London. 1991 Erie. The Bulletin, Journal of the New York State Archaeological Association 102:2-12. 1993 Smokes Creek (R. Haas). Electronic database, The Digital Archaeological Record (tDAR), https://core.tdar.org/document/375148/smokes-creek-r-haas, accessed July 27, 2016. 1995 The Case of the Disappearing Iroquoians: Early Contact Period Superpower Politics. Northeast Anthropology 50:35-59. 1999 Iroquoian Ethnicity and Archaeological Taxa. In Taming the Taxonomy: Toward a New Understanding of Great Lakes Archaeology, edited by R. F. Williamson and C. M. Watts, pp. 51-59. Eastendbooks, Toronto. 290 2003 Iroquoia: The Development of a Native World. Iroquois and their neighbors. Syracuse University Press, Syracuse, N.Y. 2004 Northern New York Revisited. In A Passion for the Past: Papers in Honour of James F. Pendergast, edited by J. V. Wright and J.-L. Pilon, pp. 125-144. Canadian Museum of Civilization, Gatineau, Quebec. Engelbrecht, William and Bruce Jamieson 2016 Stone Tipped versus Bone- and Antler-tipped Arrows and the Movement of the St. Lawrence Iroquoians from Their Homeland. Ontario Archaeology 96:76-86. Fenton, William N. 1978 Northern Iroquoian Culture Patterns. In Northeast, edited by Bruce G. Trigger, pp. 296-321. Handbook of North American Indians, Vol. 15, William C. Sturtevant, general editor, Smithsonian Institution, Washington, D.C. Ferris, Neal 1999 Telling Tales: Interpretive Trends in Southern Ontario Late Woodland Archaeology. Ontario Archaeology 68:1-62. 2006 In Their Time: Archaeological Histories of Native-Lived Contacts and Colonialisms, Southwestern Ontario, A.D. 1400-1900. Ph.D. dissertation, Department of Anthropology, McMaster University, Hamilton, Ontario. 2017 Being Iroquoian, Being Iroquois: A Thousand-year Heritage of Becoming. In Rethinking Colonial Pasts Through Archaeology. Edited by Neal Ferris, Rodney Harrison and Michael V. Wilcox, pp. 371-396. Oxford University Press, Oxford. Ferris, Neal and Michael W. Spence 1995 The Woodland Traditions in Southern Ontario. Revista de Arqueología Americana 9:83-138. Finlayson, William D. 1985 The 1975 and 1978 Rescue Excavations at the Draper Site: Introduction and Settlement Patterns. National Museum of Man, Ottawa. Finlayson, William D., David G. Smith and Bern Wheeler 1987 What Columbus Missed! St. George Press for the Museum of Indian Archaeology, London, Toronto, Ontario Finlayson, William D., Michael W. Spence, David G. Smith and Peter Timmins 1985 The 1985 Salvage Excavations at the Keffer Site: A Field Report. Kewa (Newsletter of the London Chapter, Ontario Archaeological Society) 85(8):2-10. Ford, James A. 1954 The Type Concept Revisited. American Anthropologist 56:42-53. Ford, James A. and Julian H. Steward 1954 On the Concept of Types. American Anthropologist 56(1):42-57. Fowler, Christopher 2004 The Archaeology of Personhood: An Anthropological Approach. Themes in Archaeology Series, Routledge, London. 2017 Relational Typologies, Assemblage Theory and Early Bronze Age Burials. Cambridge Archaeological Journal 27:95-109. 291 Garrad, Charles, Jean-Luc Pilon and William A. Fox (editors) 2014 Petun to Wyandot: The Ontario Petun from the Sixteenth Century. Canadian Museum of History, University of Ottawa Press, Ottawa. Gates-St. Pierre, Christian 2016 Iroquoians in the St. Lawrence River Valley before European Contact. Ontario Archaeology 96:47-64. Gaudreau, Marianne and Louis Lesage 2016 Understanding Ethnicity and Cultural Affiliation: Huron-Wendat and Anthropological Perspectives. Ontario Archaeology 96:6-16. Gellner, Ernest 1983 Nations and Nationalism. New perspectives on the past. Cornell University Press, Ithaca. Giddens, Anthony 1979 Central Problems in Social Theory: Action, Structure, and Contradiction in Social Analysis. University of California Press, Berkeley. Gifford, James C. and Watson Smith 1978 Gray Corrugated Pottery from Awatovi and Other Jeddito Sites in Northeastern Arizona. Papers of the Peabody Museum of Archaeology and Ethnology, Harvard University. Peabody Museum of Archaeology and Ethnology, Harvard University, Cambridge, Mass. Gifford-Gonzalez, Diane P., and Jun Ueno Sunseri 2007 Foodways on the Frontier: An Early Colonial Pueblo. In We Were What We Ate: The Archaeology of Food and Identity. Edited by Kathryn Twiss, pp. 260-287. Annual Visiting Scholar Conference, Occasional Papers 34. Center for Archaeological Investigations, Southern Illinois University, Carbondale. Gonzalez, Diane P. 2014. Constructing Community Through Refuse Disposal. African Archaeological Review 31:339-382. Granovetter, Mark S. 1973 The Strength of Weak Ties. American Journal of Sociology 78(6):1360-1380. Griffin, James B. 1944 The Iroquois in American Prehistory. Papers of the Michigan Academy of Science, Arts and Letters 29:357-374. Gruber, Rachel 2014 Examining the Relationship between Pipe Style Frequency and Social Interaction Among Prehistoric Huron-Wendat Groups of Ontario. Manuscript in possession of the author. Habiba, H., Jan C. Athenstädt, Barbara J. Mills and Ulrik Brandes 2018 Social Networks and Similarity of Site Assemblages. Journal of Archaeological Science 92:6372. Hall, Stuart 1990 Cultural Identity and Diaspora. In Identity: Community, Culture, Difference, edited by Jonathan Rutherford, pp. 222-237. Lawrence & Wishart, London. Hanneman, Robert and Mark Riddle 2005 Introduction to Social Network Methods. Electronic document, http://faculty.ucr.edu/~hanneman/nettext/. 292 Harris, Oliver J. T. and Craig N. Cipolla (editors) 2017 Archaeological Theory in the New Millennium: Introducing Current Perspectives. Routledge, Abingdon, UK. Hart, John P. 2000 New Dates from Classic New York Sites: Just How Old are Those Longhouses? Northeast Anthropology 60:1-22. 2012 The Effects of Geographical Distances on Pottery Assemblage Similarities: A Case Study from Northern Iroquoia. Journal of Archaeological Science 39(1):128-134. Hart, John P. and Jennifer Birch 2017 Social Networks and Northern Iroquoian Confederacy Dynamics. American Antiquity 83:13-33. Hart, John P, Jennifer Birch, Susan Dermarkar, Termeh Shafie and Ronald F. Williamson 2015 St. Lawrence Iroquoians and Pan-Iroquoian Social Network Analysis. Paper presented at the Annual Symposium of the Ontario Archaeological Society, Midland, Ontario. Hart, John P, Jennifer Birch, Susan Dermarkar, Ronald F. Williamson and Termeh Shafie 2015 Iroquoian Social Network Change in Southern Ontario A.D. 1350-1650. Paper presented at the SUNBELT International Social Network Conference, Brighton. Hart, John P., Jennifer Birch and Christian Gates St-Pierre 2017 Effects of Population Dispersal on Regional Signaling Networks: An Example from Northern Iroquoia. Science Advances 3(8):e1700497. Hart, John P. and Hetty Jo Brumbach 2003 The Death of Owasco. American Antiquity 68(4):737-752. 2009 On Pottery Change and Northern Iroquoian Origins: An Assessment from the Finger Lakes Region of Central New York. Journal of Anthropological Archaeology 28:367-381. Hart, John P. and William Engelbrecht 2012 Northern Iroquoian Ethnic Evolution: A Social Network Analysis. Journal of Archaeological Method and Theory 19(2):322-349. Hart, John P., Termeh Shafie, Jennifer Birch, Susan Dermarkar and Ronald F. Williamson 2016 Nation Building and Social Signaling in Southern Ontario: A.D. 1350-1650. PLoS ONE 11(5). Hart, John P., Robert G. Thompson and Hetty Jo Brumbach 2003 Phytolith Evidence for Early Maize (Zea mays) in the Northern Finger Lakes Region of New York. American Antiquity 68(4):619-640. Hawkins, Alicia L. 1992 Just Who Made Those Frilled Ceramics? Compositional and Stylistic Analyses of Ceramics from Ossossane and Freelton. Master’s Research Paper, Department of Anthropology, University of Toronto. 2001 Genoa Frilled Pottery and the Problem of the Identification of the Wenro in Huronia. Ontario Archaeology 72:15-37. 2004 Recreating Home? A Consideration of Refugees, Microstyles and Frilled Pottery in Huronia. Ontario Archaeology 77/78:62-80. Hayden, Brian (editor) 293 1979 Settlement Patterns of the Draper and White Sites, 1973 Excavations. 6. Dept. of Archaeology, Simon Fraser University, Burnaby, B.C. 1979 The Draper and White Sites: Preliminary and Theoretical Considerations. In Settlement Patterns of the Draper and White Sites, 1973 Excavations, edited by B. Hayden, pp. 1-28. Publication No. 6. Department of Archaeology, Simon Fraser University, Burnaby, B.C. Hayden, Brian and Aubrey Cannon 1983 Where the Garbage Goes: Refuse Disposal in the Maya Highlands. Journal of Anthropological Archaeology 2:117-163. 1984 Interaction Inferences in Archaeology and Learning Frameworks of the Maya. Journal of Anthropological Archaeology 3(4):325-367. Hayes, Charles F. III 1980 An Overview of the Current Status of Seneca Ceramics. In Proceedings of the 1979 Iroquois Pottery Conference, edited by C. F. Hayes III, pp. 87-93. vol. 13. Rochester Museum and Science Center, Research Records. Hegmon, Michelle 1992 Archaeological Research on Style. Annual Review of Anthropology 21:517-536. 2003 Setting Theoretical Egos Aside: Issues and Theory in North American Archaeology. American Antiquity 68(2):213-243. Heidenreich, Conrad E. 1972 The Huron: A Brief Ethnography. Discussion Paper No. 6. York University, Department of Geography, Toronto. Hodder, Ian 1982 Symbols in Action: Ethnoarchaeological Studies of Material Culture. Cambridge University Press, Cambridge. 1985 Postprocessual Archaeology. Advances in Archaeological Method and Theory 8:1 Hoffman, Bernard G. 1961 Cabot to Cartier: Sources for a Historical Ethnography of Northeastern North America, 14971550. University of Toronto Press, Toronto. Holterman, Carrie 2007 So Many Decisions! The Fonger Site: A Case Study of Neutral Iroquoian Ceramic Technology. Master’s thesis, Department of Anthropology, McMaster University, Hamilton, Ontario. Houghton, Frederick 1909 Indian Village, Camp and Burial Sites on the Niagara Frontier. Bulletin of the Buffalo Society of Natural Sciences 9(3):263-374. 1927 The Migrations of the Seneca Nation. American Anthropologist 29(2):241-250. Hunter, Andrew F. 1900 Notes on Sites of Huron Villages in the Township of Tay, Simcoe County. Annual Archaeological Report for Ontario, 1899:51-82. 1902 Notes on Sites of Huron Villages in the Township of Medonte (Simcoe Co.). Annual Archaeological Report for Ontario, 1901:56-100. 1907 Huron Village Sites. Annual Archaeological Report for Ontario, 1906:1-56. 294 Huntley, Deborah L. 2004 Interaction, Boundaries, and Identities: A Multiscalar Approach to the Organizational Scale of Pueblo IV Zuni Society. Ph.D. dissertation, Arizona State University. Hyslop, John 1984 The Inca Road System. Academic Press, St. Louis Jamieson, J. Bruce 1982 The Steward Site: A Study in St. Lawrence Iroquoian Chronology. Master’s thesis, Department of Anthropology, McGill University, Montreal. 1990 The Archaeology of the St. Lawrence Iroquoians. In The Archaeology of Southern Ontario to A.D. 1650, edited by C. J. Ellis and N. Ferris, pp. 385-404. Occasional Publication No. 5. London Chapter, Ontario Archaeological Society, London, Ont. Jenkins, Richard 2000 Categorization: Identity, Social Process and Epistemology. Current Sociology 48(3):7-25. 2004 Social Identity. Routledge, London. 2008 Rethinking Ethnicity: Arguments and Explorations. Second edition. Sage, Thousand Oaks, California. Jenkins, Tara 2014 Contexts, Needs and Social Messaging: Situating Iroquoian Human Bone Artifacts in Southern Ontario, Canada. Paper presented at the Annual Symposium of the Ontario Archaeological Society. 2016 Contexts, Needs, and Social Messaging: Situating Iroquoian Human Bone Artifacts in Southern Ontario, Canada. In Theoretical Approaches to Analysis and Interpretation of Commingled Human Remains, edited by A. J. Osterholtz, pp. 139-183. Springer, New York. Johannsen, Erika 2014 Investigating Castellation Diversity among Iroquoian Potters at the Keffer Site. Paper presented at the Annual Symposium of the Ontario Archaeological Society. Jones, Sîan 1996 Discourses of identity in the interpretation of the past. In Cultural Identity and Archaeology: the construction of European communities, edited by P. Graves Brown, S. Jones and C. Gamble, pp. 62-80. London: Routledge. 1997 The Archaeology of Ethnicity: Constructing Identities in the Past and Present. Routledge, New York. Jury, Wilfrid and Elsie McLeod Murray Jury 1955 St. Louis: Huron Indian Village and Jesuit Mission Site. Bulletin 10. Museum of Indian Archaeology, University of Western Ontario, London, Ont. Kapches, Mima 1981 The Middleport Pattern in Ontario Iroquoian Prehistory. Unpublished Ph.D. dissertation, Department of Anthropology, University of Toronto. 1987 The Auda Site: An Early Pickering Iroquois Component in Southeastern Ontario. Archaeology of Eastern North America 15:155-175. 1990 The Spatial Dynamics of Ontario Iroquoian Longhouses. American Antiquity 55(1):49-57. 295 1993 The Ontario Iroquoian Longhouse ca 1350 to ca. 1640: A Study in Vernacular Archaeology. In Vernacular Architecture in Ontario, edited by J. Beck and A. Keefer, pp. 3-15. Architectural Conservancy of Ontario, Toronto. 1995 Chaos Theory and Social Movements: A Theoretical View of the Formation of the Northern Iroquoian Longhouse Cultural Pattern. In Origins of the People of the Longhouse: Proceedings of the 21st Annual Symposium of the Ontario Archaeological Society, edited by A. Bekerman and G. Warrick, pp. 86-96. Ontario Archaeological Society, Toronto. Katzenberg, M. Anne, Henry P. Schwarcz, M. Knyfe and F. Jerome Melbye 1995 Stable Isotope Evidence for Maize Horticulture and Paleodiet in Southern Ontario, Canada. American Antiquity 60:335-350. Kerr, Donald P. and Spelt, Jacob 1965 The Changing Face of Toronto: A Study in Urban Geography. Memoir 11, Geographical Branch, Department of Mines and Technical Surveys. Queen's Printer, Ottawa. Knappett, Carl (editor) 2013 Network Analysis in Archaeology: New Approaches to Regional Interaction. Oxford University Press, Oxford. 2018 From Network Connectivity to Human Mobility: Models for Minoanization. Journal of Archaeological Method and Theory 25:974-995. Knight, Dean and Margot Snyder 1981 The 1981 Excavations at the Ball Site. Report on file, Department of Sociology and Anthropology, Wilfrid Laurier University, Waterloo, Ontario. Kolb, Michael J. and James E. Snead 1997 It's a Small World after All: Comparative Analyses of Community Organization in Archaeology. American Antiquity 62:609-628. Konrad, J. V. 1973 The Archaeological Resources of the Metropolitan Toronto Planning Area: Inventory and Prospect. Discussion Paper No. 10. York University, Department of Geography, Toronto. Kossinna, G. 1911 Die Herkunft der Germanen: Zur Methode der Siedlungsarchäologie. Kabitzsch, Leipzig. Kraus, Bertram S. 1944 Acculturation: A New Approach to the Iroquoian Problem. American Antiquity 9(3):302-317. Kroeber, Alfred L. 1939 Cultural and Natural Areas of Native North America. University of California Publication in American Archaeology and Ethnology 38. University of California Press, Berkeley. Kroskrity, Paul V. 1993 Language, History, and Identity: Ethnolinguistic Studies of the Arizona Tewa. University of Arizona Press, Tucson. Kuhn, Robert D. 2004 Reconstructing Patterns of Interaction and Warfare Between the Mohawk and Northern Iroquoians during the A.D. 1400-1700 Period. In A Passion for the Past: Papers in Honour of 296 James F. Pendergast, edited by J. V. Wright and J.-L. Pilon, pp. 145-166. Mercury Series Paper No. 164. Canadian Museum of Civilization, Gatineau, Quebec. Kvetina, Petr and Vaclav Hrncir 2013 Between Archaeology and Anthropology: Imagining Neolithic Settlements. Anthropologie 12:323-347. LaBianca, Øystein Sakala and Sandra A. Scham (editors) 2006 Connectivity in Antiquity: Globalization as a Long-Term Historical Process. Equinox Pub., London, Oakville, Conn. Lafitau, Joseph-François 1724 Moeurs des sauvages amériquains, comparées aux moeurs des premiers temps. Saugrain l'aîné et Ch. E. Hochereau, Paris. Laidlaw, George E. 1891 Balsam Lake. Annual Archaeological Report for Ontario, 1891:13-77. 1917 Indian Sites in Victoria County and Vicinity. Annual Archaeological Report for Ontario, 1917. Latta, Martha A. 1976 The Iroquoian Cultures of Huronia: A Study of Acculturation through Archaeology. Unpublished Ph.D. dissertation, Department of Anthropology, University of Toronto. 1980 Controlling the Heights: The Iroquoian Occupations of the Albion Pass Region. Archaeology of Eastern North America 8:71-77. 1983 Ceramic Types in Ontario. Report on file, Department of Anthropology, University of Toronto at Mississauga. Lave, J. and E. Wenger 1991 Situated Learning: Legitimate Peripheral Participation. Cambridge University Press, Cambridge. 1993 The Ontario Iroquoian Longhouse ca 1350 to 1640: A Study in Vernacular Archaeology. In Vernacular Architecture of Ontario, edited by J. Beck and A. Keffer, pp. 3-15. Architectural Conservancy of Ontario, Toronto. Lee, Jaegwon 2006 Emergence: Core Ideas and Issues. Synthese151:547-559. Lemonnier, Pierre 1986 The Study of Material Culture Today: Toward an Anthropology of Technical Systems. Journal of Anthropological Archaeology 5:147-186. Lenig, Donald 1965 The Oak Hill Horizon and its Relation to the Development of Five Nations Iroquois Culture. Researches and transactions of the New York State Archeological Association 15(1). New York State Archaeological Association, Buffalo, N.Y. Lennox, Paul A 1981 The Hamilton Site: A Late Historic Neutral Town. Mercury Series 103. National Museum of Man, Ottawa. 2000 The Molson Site: An Early Seventeenth Century, First Nations Settlement, Simcoe County, Ontario. Bulletin 18. London Museum of Archaeology. 297 Lennox, Paul A. and William R. Fitzgerald 1990 The Culture History and Archaeology of the Neutral Iroquoians. In The Archaeology of Southern Ontario to A.D. 1650, edited by C. J. Ellis and N. Ferris, pp. 405-456. Occasional Publication No. 5. London Chapter, Ontario Archaeological Society, London, Ont. Lennox, Paul A. and Ian T. Kenyon 1984 Was That Middleport Oblique or Pound Oblique? A Study in Iroquoian Ceramic Typology. Ontario Archaeology 42:13-26. Lerner, Harry J. 1997 The Keffer Site (AkGv-14) Projectile Point Assemblage: A Preliminary Morphological and Distributional Analysis. Report on file, Museum of Ontario Archaeology, London, Ont. 2000 Static Types to Dynamic Variables: Re-assessing the Methods of Prehistoric Huron Chipped Stone Tool Documentation and Analysis in Ontario. Master’s thesis, Department of Anthropology, McGill University, Montreal. 2011 Dynamic Variables and the Use-Related Reduction of Huron Projectile Points. Paper presented at the 78th Annual Meeting of Society for American Archaeology, Sacramento, California 2015 Dynamic Variables and the Use-Related Reduction of Huron Projectile Points. In Works in Stone: Contemporary Perspectives on Lithic Analysis, edited by M. J. Shott, pp. 143-161. University of Utah Press, Salt Lake City. Lesage, Louis, Neha Gupta and Georges Sioui 2016 Introduction. Ontario Archaeology 96:3-6. Lockwood, W. G. 1981 Religion and Language as Criteria of Ethnic Identity: An Exploratory Comparison. In Ethnicity and Nationalism in Southeastern Europe, edited by S. Beck and J. W. Cole, pp. 71-81. Antropologisch-Sociologisch Centrum, Universiteit van Amsterdam, Amsterdam. López Varela, Sandra L. (editor) 2017 Innovative Approaches and Explorations in Ceramic Studies. Archaeopress, Oxford. Lucas, G. 2012 Understanding the Archaeological Record. Cambridge University Press, Cambridge. Lyons, Patrick D. and Jeffery J. Clark 2008 Interaction, Enculturation, Social Distance, and Ancient Ethnic Identities. In Archaeology without Borders: Contact, Commerce, and Change in the U.S. Southwest and Northwestern Mexico, edited by L. D. Webster and M. E. McBrinn, pp. 185-207. University Press of Colorado, Boulder, Colo. MacDonald, Robert I. 1986 The Coleman Site (AiHd-7): A Late Prehistoric Iroquoian Village in the Waterloo Region. Master’s thesis, Department of Anthropology, Trent University, Peterborough, Ontario. 1988 Ontario Iroquoian Sweat Lodges. Ontario Archaeology 48:17-26. 2002 Late Woodland Settlement Trends in South-Central Ontario: A Study of Ecological Relationships and Culture Change. Unpublished Ph.D. dissertation, Department of Anthropology, McGill University, Montreal. MacDonald, Robert I. and Ronald F. Williamson 298 1995 The Hibou Site (AlGo-50): Investigating Ontario Iroquoian Origins in the Central North Shore Area of Lake Ontario. In Origins of the People of the Longhouse: Proceedings of the 21st Annual Symposium of the Ontario Archaeological Society, edited by A. Bekerman and G. A. Warrick, pp. 9-42. Ontario Archaeological Society, Toronto. 2001 Sweat Lodges and Solidarity: The Archaeology of the Hubbert Site. Ontario Archaeology 71:2978. MacEachern, Scott 1998 Scale, Style, and Cultural Variation: Technological Traditions in the Northern Mandara Mountains. In The Archaeology of Social Boundaries, edited by M. T. Stark, pp. 107-131. Smithsonian Institution Press, Washington. MacNeish, Richard S. 1952 Iroquois Pottery Types: A technique for the study of Iroquois prehistory. National Museum of Canada, Bulletin 124. Canada Dept. of Resources and Development, National Parks Branch, National Museum of Canada, Ottawa. Manning, Sturt W., Jennifer Birch, Megan A. Conger, Michael W. Dee, Carol Griggs, Carla S. Hadden, Alan G. Hogg, Christopher Bronk Ramsey, Samantha Sanft, Peter Steier and Eva M. Wild 2018 Radiocarbon re-dating of contact-era Iroquoian history in northeastern North America. Science Advances 4(12). DOI:10.1126/sciadv.aav0280. Mantha, Alexis 2009 Territoriality, social boundaries and ancestor veneration in the central Andes of Peru. Journal of Anthropological Archaeology 28:158-176. Marin, Alexandra and Barry Wellman 2010 Social Network Analysis: An Introduction. In The Sage Handbook of Social Network Analysis, edited by J. Scott and P. J. Carrington, pp. 11-25. Sage, London. Martelle, Holly 2002 Huron Potters and Archaeological Constructs: Researching Ceramic Microstylistics. Unpublished Ph.D. dissertation, Department of Anthropology, University of Toronto. 2004 Some Thoughts on the Impact of Epidemic Disease and European Contact on Ceramic Production in Seventeenth Century Huronia. Ontario Archaeology 77/78:22-44. Mauss, Marcel 1935 Techniques of the Body. Journal de Psychologie Normal et Pathologique XXXII:271-293. McKern, William C. 1939 The Midwestern Taxonomic Method as an Aid to Archaeological Study. American Antiquity 4(4):301-313. McNutt, Charles H. 2005 Seriation: Classic Problems and Multivariate Applications. Southeastern Archaeology 24(2):209222 McPherson, Miller, Lynn Smith-Lovin and James M. Cook 2001 Birds of a Feather: Homophily in Social Networks. Annual Review of Sociology 27(1):415-444. 299 Mills, Barbara J. 2003 Ceramics and Zuni Identity. In Zuni Origins: Anthropological Approaches on Multiple Scales, edited by David Gregory and David Wilcox. University of Arizona Press, Tucson. 2007 Performing the Feast: Visual Display and Suprahousehold Commensalism in the Puebloan Southwest. American Antiquity 72:210-239. 2016 Communities of Consumption: Cuisines as Constellated Networks of Situated Practice. In Knowledge in Motion; Constellations of Learning across Time and Place, edited by Andrew P. Roddick and Ann B. Stahl, pp. 247-270. University of Arizona Press, Tucson. 2017 Social Network Analysis in Archaeology. Annual Review of Anthropology 46(1):379-397. 2018 Intermarriage, Technological Diffusion, and Boundary Objects in the U.S. Southwest. Journal of Archaeological Method and Theory 25(4):1051-1086. Mills, Barbara J., Jeffery J. Clark, Matthew A. Peeples, W. R. Jr. Haas, John M. Jr. Roberts, J. Brett Hill, Deborah L. Huntley, Lewis Borck, Ronald L. Breiger, Aaron Clauset and M. Steven Shackley 2013 Transformation of Social Networks in the Late Pre-Hispanic US Southwest. PNAS 110(15):5785.5790. Mills, Barbara J., Matthew A. Peeples, W. Randall Jr. Haas, Lewis Borck, Jeffery J. Clark and John M. Jr. Roberts 2015 Multiscalar Perspectives on Social Networks in the Late Prehispanic Southwest. American Antiquity 80(1):3-24. Moerman, Michael 1965 Ethnic Identification in a Complex Civilization: Who Are the Lue? American Anthropologist 67(5, Part 1):1215-1230. MPP (Mayer, Pihl, Poulton and Associates) 1987 The 1984 Salvage Excavation at the Boyle-Atkinson Site (AlGu-1), Town of Richmond Hill, Ontario. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. Nash, M. 1987 Ethnicity in Peninsular Malaysia: The idiom of communalism, confrontation and co-operation. In Dimensions of Social Life: Essays in Honor of David G. Mandelbaum, edited by R. Hockings, pp. 559-571. Mouton de Gruyter, Berlin. Nash, Philleo 1939 The Pound Village Site, Elgin County, Ontario. Society for American Archaeology Notebook 7475. Niemczycki, Mary Ann P. 1984 The Origin and Development of the Seneca and Cayuga Tribes of New York State. Research records / Research Division, Rochester Museum and Science Center. Rochester Museum and Science Center, Rochester, N.Y. 1995 Ceramics and Ethnicity In West-Central New York: Exploring Owasco-Iroquois Connections. Northeast Anthropology 49:43-54. Noble, William C. 300 1968 Iroquois Archaeology and the Development of Iroquois Social Organization (1000-1650 A.D.). Ph.D. dissertation, University of Calgary. 1974 The Jackes (Eglinton) Site: Another Facet of Southern Huron Development in the Toronto Region. Ontario Archaeology 22:3-31. Nougaret, Sarah 2012 La Céramique Lalonde High Collar: Comme outil de recherche pour l’étude des échanges dans le Nord-Est de l’Amérique du Nord (1450-1550 A.D.). Ph.D. dissertation, Laval University, Quebec. Ontario Geological Survey 1980 Quaternary Geology Toronto and Surrounding Area. Geological Series, p. 2204. Electronic document, https://mdl.library.utoronto.ca/collections/scanned-maps/quaternary-geology-torontoand-surrounding-area-p-2204. Ortman, Scott G. 2010 Evidence of a Mesa Verde Homeland for the Tewa Pueblos. In Leaving Mesa Verde: Peril and Change in the Thirteenth-Century Southwest, edited by T. A. Kohler, M. Varien and A. M. Wright, pp. 222-261. University of Arizona Press, Tucson. Ounjian, G. L 1998 Glen Meyer and Prehistoric Neutral Paleoethnobotany. Unpublished Ph.D. dissertation, Department of Anthropology, University of Toronto. Pailes, M. 2014 Social network analysis of early classic Hohokam corporate group inequality. American Antiquity 79(3): 465-486. Parker, A. C. 1916 The Origin of the Iroquois as Suggested by their Archaeology. American Anthropologist 18:479507. Pauketat, Timothy R. and Susan M. Alt 2005 Agency in a Postmold? Physicality and the Archaeology of Culture-Making. Journal of Archaeological Method and Theory 12(3):213-237. Pearce, Robert J. 1978 A Preliminary Report on the Draper Site Rim Sherds. Research Report 1. Museum of Indian Archaeology, University of Western Ontario, London, Ont. 1996 Mapping Middleport: A Case Study in Societal Archaeology. Research Report 25. London Museum of Archaeology, London, Ont. 1997 The Watford Site (AlGu-5), Richmond Hill: License Report. Report on file, Ontario Ministry of Tourism, Culture and Sport, Toronto. Peeples, Matthew A. 2011 Identity and Social Transformation in the Prehispanic Cibola World: A.D. 1150-1325. Unpublished Ph.D. dissertation, Department of Anthropology, Arizona State University. 2011 R Script for Calculating the Brainerd-Robinson Coefficient of Similarity and Assessing Sampling Error. Electronic document, http://www.mattpeeples.net/br.html, accessed September 25, 2014. Pendergast, James F. 301 1962 The Crystal Rock site: An early Onondaga-Oneida site in eastern Ontario. Pennsylvania Archaeologist 32(1):21-34. 1963The Payne Site: An Iroquoian Manifestation in Prince Edward County, Ontario. In Contributions to Anthropology V: Archaeology and Physical Anthropology, pp. 126-257. Bulletin 206. National Museum of Canada, Ottawa. 1964 The Waupoos Site—An Iroquois Component in Prince Edward County, Ontario. Pennsylvania Archaeologist 34:69-89. 1965 Other Ideas on “The Ontario Iroquois Controversy”. Ontario Archaeology 8:39-44. 1966 Three Prehistoric Components in Eastern Ontario: The Salem, Gray’s Creek and Beckstead Sites. Bulletin 208. National Museum of Canada, Ottawa. 1972 The Lite Site: An Early Southern Division Huron Site near Belleville, Ontario. Ontario Archaeology 17:24-61. 1973 The Roebuck Prehistoric Village Rimsherds: An Attribute Analysis. Archaeological Survey of Canada Mercury Series Paper no. 8. National Museum of Man. Ottawa. 1980 Origins of the St Lawrence Iroquoian Pottery on the Draper Site. Report on file, Museum of Ontario Archaeology, London, Ont. 1984 The Beckstead Site—1977. National Museum of Man, Ottawa. Pendergast, James F. and Bruce G. Trigger 1972 Cartier's Hochelaga and the Dawson Site. McGill-Queen's University Press, Montreal and Kingston. Pihl, Robert H. 1984 Final Report on the Draper Rim Sherd Collection: 1975/1978 Excavated Samples. Report on file, Archaeological Services Inc., Toronto. Pihl, Robert H. and Stephen Cox Thomas 1997 The Finch Site: A Late Iroquoian Special-Purpose Site on West Catfish Creek. Ontario Archaeology 63:37-84. Plog, Stephen 1980 Stylistic Variation in Prehistoric Ceramics. Cambridge University Press, Cambridge. 1983 Analysis of Style in Artifacts. Annual Review of Anthropology 12:125-142. Pradzynski, Aleksandra 2013 Damiani (AlGv-231) Ceramic Analysis. Report on file, Archaeological Services Inc., Toronto. Pratt, Peter P. 1960 A Criticism of MacNeish's Pottery Types. Pennsylvania Archaeologist 30:106-110. 1976 Archaeology of the Oneida Iroquois. Occasional Publications in Northeastern Anthropology 1. Man in the Northeast, George's Mill, N.H. 1980 Oneida Iroquois Pottery Typology. In Proceedings of the 1979 Iroquois Pottery Conference, edited by C. F. Hayes III, pp. 35-49. vol. 13. Rochester Museum and Science Center, Research Records. Preucel, Robert W. 302 1998 Learning from the Elders. In Excavating Voices: Listening to Photographs of Native Americans, edited by M. Katakis, pp. 17-25. University of Pennsylvania Museum of Archaeology and Anthropology, Philadelphia. R, Core Team 2013 R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna. Rainey, Dori L. 2002 Challenging Assumptions: An Analysis of the Scattered Human Remains from the Keffer Site (AkGv-14). Master’s thesis, University of Western Ontario, London, Ont. Ramsden, Peter G. 1977 A Refinement of Some Aspects of Huron Ceramic Analysis. National Museum of Man, Ottawa. 1978 An Hypothesis Concerning the Effects of Early European Trade Among Some Ontario Iroquois. Canadian Journal of Archaeology 2:101-110. 1990 The Hurons: Archaeology and Culture History. In The Archaeology of Southern Ontario to A.D. 1650, edited by C. J. Ellis and N. Ferris, pp. 361-384. Occasional Publication No. 5. London Chapter, Ontario Archaeological Society, London, Ont. 1990 Saint Lawrence Iroquoians in the Upper Trent River Valley. Man in the Northeast 39:87-95. 2009 Politics in a Huron Village. In Painting the Past with a Broad Brush: Papers in Honour of James Valliere Wright, edited by D. L. Keenlyside and J.-L. Pilon, pp. 299-318. Canadian Museum of Civilization, Gatineau. 2016 Becoming Wendat: Negotiating a New Identity around Balsam Lake in the Late Sixteenth Century. Ontario Archaeology 96:121-132. Regal, Alberto 1936 Los Caminos del Inca en el Antigua Peru. Sanmarti, Lima. Renfrew, Colin 1986 Introduction: Peer Polity Interaction and Socio-Political Change. In Peer Polity Interaction and Socio-Political Change, edited by C. Renfrew and J. Cherry, pp. 1-18. Cambridge University Press, Cambridge. Renfrew, Colin and J. Cherry 1986 Peer Polity Interaction and Socio-Political Change, edited by C. Renfrew and J. Cherry. Cambridge University Press, Cambridge. Richard, Jean-François 2016 Territorial Precedence in Eighteenth- and Nineteenth-Century Huron-Wendat. Ontario Archaeology 96:26-34. Ridley, Frank 1952 The Huron and Lalonde Occupations of Ontario. American Antiquity 17(3):197-210. 1952 The Fallis Site, Ontario. American Antiquity 18:7-14. 1958 Did the Huron Really Migrate North from the Toronto Area? Pennsylvania Archaeologist 28(3 & 4):143-144, 150. 1961 Archaeology of the Neutral Indian. Etobicoke Historical Society, Ontario. 1963 The Ontario Iroquoian Controversy. Ontario History 55(1):49-59. 303 Ritchie, William A. 1944 The Pre-Iroquoian Occupations of New York State. Rochester Museum of Arts and Sciences, Memoir No. 1. Rochester, New York. 1951 A Current Synthesis of New York Prehistory. American Antiquity 17:130-136. 1952 The Chance Horizon: An Early Stage of Mohawk Iroquois Cultural Development. New York State Museum Circular. University of the State of New York, Albany. Ritchie, William A. and Richard S. MacNeish. 1949 The Pre-Iroquoian Pottery of New York State. American Antiquity 15(2):97-124. Robertson, David A., Stephen G. Monckton and Ronald F. Williamson 1995 The Wiacek Site Revisited: The Results of the 1990 Excavations. Ontario Archaeology 60:40-91. Robertson, David A. and Ronald F. Williamson 1998 The Archaeology of the Parsons Site: Summary and Conclusions. Ontario Archaeology 65/66:146-150. Robinson, W. S. 1951 A Method for Chronologically Ordering Archaeological Deposits. American Antiquity 16(4):293-301. Rowlands, Michael 1998 Centre and periphery: a review of a concept. In Social Transformations in Archaeology: Global and Local Perspectives, edited by Kristian Kristiansen and Michael Rowlands, pp. 214-235. Routledge, London. Sackett, James R. 1973 Style, function and artifact variability in Palaeolithic assemblages. In The Explanation of Culture Change, edited by Colin Renfrew pp. 317-325. University of Pittsburgh Press, Pittsburgh. 1977 The Meaning of Style in Archaeology: A General Model. American Antiquity 42:369-380. Sahlins, Marshall and Elman R. Service (editors) 1960 Evolution and Culture. University of Michigan Press, Ann Arbor. Sassaman, Kenneth E. and Wictoria Rudolphi 2001 Communities of Practice in the Early Pottery Traditions of the American Southeast. Journal of Anthropological Research 57(4):407-425. Schachner, Gregson 2017 Corporate Group Formation and Differentiation in Early Puebloan Villages of the American Southwest. American Antiquity 75(3):473-496. Service, Elman 1962 Primitive Social Organization: an evolutionary perspective. Random House, New York Shanks, Michael and Christopher Y. Tilley 1987 Re-Constructing Archaeology: Theory and Practice. New studies in archaeology. Cambridge University Press, Cambridge. Shennan, Stephen J. (editor) 1989 Archaeological Approaches to Cultural Identity. Unwin Hyman, London and Boston. Silliman, Stephen W. 304 2015 A Requiem for Hybridity? The Problem with Frankensteins, Purées, and Mules. Journal of Social Archaeology 15(3):277-298. Simmel, Georg 1955 [1922] Conflict and The Web of Group-Affiliations. Free Press, [S.l.]. Sindbæk, Søren M. 2010 Re-Assembling Regions: The Social Occasions of Technological Exchange in Viking Age Scandinavia. In The Archaeology of Regional Technologies: Case Studies from the Palaeolithic to the Age of the Vikings, edited by R. Barndon, A. Engevik and I. Øye, pp. 263-289. Edwin Mellen Press, Lewiston, N.Y. Sioui, Georges E. 1999 Huron-Wendat: The Heritage of the Circle. University of British Columbia Press, Vancouver. Skinner, Alanson 1921 Notes on Iroquois Archaeology. Indian Notes and Monographs. Museum of the American Indian, Heye Foundation, New York. Smith, Angele 1988 Keffer Field Notes Volume 1. Report on file, Museum of Ontario Archaeology, London, Ont. Smith, Anthony D. 1986 The Ethnic Origins of Nations. B. Blackwell, Oxford, UK New York, NY, USA. Smith, David G. 1990 Iroquoian Societies in Southern Ontario: Introduction and Historical Overview. In The Archaeology of Southern Ontario to A.D. 1650, edited by C. J. Ellis and N. Ferris, pp. 279-290. Occasional Publication No. 5. London Chapter, Ontario Archaeological Society, London, Ont. 1991 Keffer Site (AkGv-14) Rim Sections and Ceramic Smoking Pipes. Museum of Indian Archaeology, Research Report 23. 1997 Archaeological Systematics and the Analysis of Iroquoian Ceramics: A Case Study from the Crawford Lake Area, Ontario. Bulletin 15. London Museum of Archaeology, London, Ont. 2014 Pottery Decoration as an Index of Heterarchical Social Relations in Ancestral Wendat Society. Paper presented at the Annual Symposium of the Ontario Archaeological Society. 2014 The Keffer Site: 26 Years Later. Paper presented at the Annual Meeting of the Canadian Archaeological Association. Smith, David G. and Gary W. Crawford 1995 The Princess Point Complex and the Origins of Iroquoian Societies in Ontario. In Origins of the People of the Longhouse: Proceedings of the 21st Annual Symposium of the Ontario Archaeological Society, edited by A. Bekerman and G. A. Warrick, pp. 55-70. Ontario Archaeological Society, Toronto. 1997 Recent Developments in the Archaeology of the Princess Point Complex in Southern Ontario. Canadian Journal of Archaeology 21:9-32. 2002 Recent Developments in the Archaeology of the Princess Point Complex in Southern Ontario. In Northeast Subsistence and Settlement Change, A.D. 700-1300, edited by J. P. Hart and C. B. Rieth, pp. 97-116. New York State Museum Bulletin. Vol 496. New York State Museum, Albany. 305 ca. 1997 The Roots of Iroquoian Sedentism: Middle to Late Woodland Settlement Shifts in South-Central Ontario. Report on file, Department of Anthropology, University of Toronto at Mississauga. Snow, Dean R. 1989 The Evolution of Mohawk Households A.D. 1400-1800. In Households and Communities, edited by S. MacEachern, D. Archer and R. Gavin, pp. 293-300. University of Calgary Archaeological Association, Calgary. 1994 The Iroquois. Blackwell, Oxford. 1994 Paleoecology and the Prehistoric Incursion of Northern Iroquoians into the Lower Great Lakes Region. In Great Lakes Archaeology and Paleoecology: Exploring Interdisciplinary Initiatives for the Nineties, edited by R. I. MacDonald, pp. 283-293. Quaternary Sciences Institute, Waterloo, Ontario. 1995 Migration in Prehistory: The Northern Iroquoian Case. American Antiquity 60(1):59-79. 1996 More on Migration in Prehistory: Accommodating New Evidence in the Northern Iroquoian Case. American Antiquity 61(4):791-796. 1996 Mohawk Demography and the Effects of Exogenous Epidemics on American Indian Populations. Journal of Anthropological Archaeology 15(2):160-182. 1997 The Architecture of Iroquois Longhouses. Northeast Anthropology 53:61-84. Somers, Margaret R. 1994 The Narrative Constitution of Identity: A Relational and Network Approach. Theory and Society 23(5):605-649. Spaulding, Albert C. 1953 Statistical Techniques for the Discovery of Artifact Types. American Antiquity 18:305-314. Spence, Michael 1988 The Excavation of the Keffer Site Burials. Report on file, Museum of Ontario Archaeology, London, Ont. Spence, Michael W., Robert H. Pihl and Carl R. Murphy 1990 Cultural Complexes of the Early and Middle Woodland Periods. In The Archaeology of Southern Ontario to A.D. 1650, edited by C. J. Ellis and N. Ferris, pp. 125-169. Occasional Publication No. 5. London Chapter, Ontario Archaeological Society, London, Ont. Spence, Michael W. and Dori Rainey 2017 Mortuary Practices and Their Social Implications at the Keffer Site, Ontario. Ontario Archaeology 97:1-43. Stark, Miriam T. 1998 Technical Choices and Social Boundaries in Material Culture Patterning: An Introduction. In The Archaeology of Social Boundaries, edited by M. T. Stark, pp. 1-11. Smithsonian Institution Press, Washington. 1998 The Archaeology of Social Boundaries. Smithsonian Institution Press, Washington. Stark, Miriam T., Ronald L. Bishop and Elizabeth Miksa 2000 Ceramic Technology and Social Boundaries: Cultural Practices in Kalinga Clay Selection and Use. Journal of Archaeological Method and Theory 7(4):295-331. Starna, William A. 306 1980 Mohawk Iroquois Populations: A Revision. Ethnohistory 27(4):371-820. Starna, William A. and Robert E. Funk 1994 The Place of the In Situ Hypothesis in Iroquoian Archaeology. Northeast Anthropology 47:4554. Steckley, John 2016 St. Lawrence Iroquoians among the Wendat: Linguistic Evidence. Ontario Archaeology 96:1725. Sterner, Judy 1989 Who is Signalling Whom? Ceramic Style, Ethnicity and Taphonomy among the Sirak Bulahay. Antiquity 63(240):451-459. Stewart, Frances L. 1991 The Faunal Remains from the Keffer Site (AkGv-14), a Southern Ontario Iroquois Village. Museum of Indian Archaeology, Research Report. 1997 Proto-Huron/Petun and Proto-St. Lawrence Iroquoian Subsistence as Culturally Defining. Unpublished Ph.D. dissertation, Department of Anthropology, McGill University, Montreal. Stone, Tammy 2003 Social Identity and Ethnic Interaction in the Western Pueblos of the American Southwest. Journal of Archaeological Method and Theory 10(1):31-67. Stopp, Marianne P. 1982 An Archaeological Examination of the Baumann Site, a Precontact Lalonde Settlement in Simcoe County, Ontario. Master’s thesis, Memorial University of Newfoundland, St. John's, Newfoundland. 1985 An Archaeological Examination of the Baumann Site: A 15th Century Settlement in Simcoe County, Ontario. Ontario Archaeology 43:3-29. Striker, Sarah, Linda Howie and Ronald Williamson 2017 Forming Pots and Community: Pottery Production and Potter Interaction in an Ancestral Wendat Village. In Innovative Approaches and Explorations in Ceramic Studies, edited by Sandra L. Lopez Varela, pp. 53-70. Archaeopress Archaeology, Oxford. Struever, Stuart 1964 The Hopewell Interaction Sphere in Riverine-Western Great Lakes Culture History. In Hopewellian Studies, edited by J. R. Caldwell and R. L. Hall, pp. 85-106. vol. 12. Illinois State Museum, Scientific Papers. Sutton, Richard E. 1990 Hidden Amidst the Hills: Middle and Late Iroquoian Occupations in the Middle Trent Valley. Occasional Papers in Northeastern Archaeology 3. Copetown Press, Dundas, Ontario. 1995 New Approaches for Identifying Prehistoric Iroquoian Migrations. In Origins of the People of the Longhouse: Proceedings of the 21st Annual Symposium of the Ontario Archaeological Society, edited by A. Bekerman and G. A. Warrick, pp. 71-85. Ontario Archaeological Society, Toronto. 1996 The Middle Iroquoian Colonization of Huronia. Unpublished Ph.D. dissertation, Department of Anthropology, McMaster University, Hamilton, Ontario. 307 1999 The Barrie Site: A Pioneering Iroquoian Village Located in Simcoe County, Ontario. Ontario Archaeology 67:40-87. Svensson, T. G. 1985 Basic anthropological perspectives in the study of ethnicity. In Ethnicity in Canada, edited by A. Pletsch, pp. 30-50. Marburger Geographische Schriftern, Marburg, Germany. Swenson, Edward R. 2018 Assembling the Moche: The Power of Contemporary Gatherings on the North Coast of Peru. World Archaeology. DOI:1080/00438243.2018.1474132. Thompson, Robert G., John P. Hart, Hetty Jo Brumbach and Robert Lusteck 2004 Phytolith Evidence for Twentieth-Century B.P. Maize in Northern Iroquoia. Northeast Anthropology 68:25-40. Thwaites, Reuben Gold, editor 1896-1901 The Jesuit Relations and Allied Documents. 73 Vols. Burrows Brothers, Cleveland. Tilly, Charles 1978 From Mobilization to Revolution. [Don Mills, Ontario.] Addison-Wesley Pub. Co., Reading, Mass. 2001 Mechanisms in Political Processes. Annual Review of Political Science 4(1):21-41. Timmins, Peter A. 1997 The Calvert Site: An Interpretive Framework for the Early Iroquoian Village. Canadian Museum of Civilization, Gatineau, Quebec. Tooker, Elisabeth 1984 Women in Iroquois Society. In Extending the Rafters: Interdisciplinary Approaches to Iroquoian Studies, edited by Michael K. Foster, Jack Campisi and Marianne Methun, pp. 109-124. State University of New York Press, Albany. 1991 An Ethnography of the Huron Indians, 1615-1649. Syracuse University Press, Syracuse, N.Y. Topic, John R. and Theresa Lange Topic 1983 Coast Highland Relations in Northern Peru: Some Observations on Routes, Networks and Scales of Interaction. In Civilization in the Andes: Essays in Honor of Gordon R. Willey edited by Richard M. Leventhal and Alan L. Kolata, pp. 237-259. University of New Mexico Press, Albuquerque. Trigger, Bruce G. 1976 The Children of Aataentsic: A History of the Huron People to 1660. McGill-Queen’s University Press, Montreal and Kingston. 1978 Iroquoian Matriliny. Pennsylvania Archaeologist 48(1-2):55-65. 1980b Archaeology and the Image of the American Indian. American Antiquity 45:662-76. 1981 Archaeology and the Ethnographic Present. Anthropologica 23(1):3-17. 1985 Draper: Past and Prologue. In The 1975 and 1978 Rescue Excavations at the Draper Site: Introduction and Settlement Patterns, edited by W. D. Finlayson, pp. 5-18. National Museum of Man, Ottawa. 1985 Natives and Newcomers: Canada's “Heroic Age” Reconsidered. McGill-Queen's University Press, Montreal and Kingston. 308 1989 1990 A History of Archaeological Thought. Cambridge University Press, Cambridge. The Huron: Farmers of the North (Second Edition). Harcourt Brace Jovanovich College Publishers, New York. 1990 Maintaining Economic Equality in Opposition to Complexity: An Iroquoian Case Study. In The Evolution of Political Systems: Sociopolitics in Small-Scale Sedentary Societies, edited by S. Upham, pp. 119-145. Cambridge University Press, Cambridge. 1995 Expanding Middle-range Theory. Antiquity 69(264):449-458. 2006 A History of Archaeological Thought, Second Edition. 2nd ed. Cambridge University Press, Cambridge. Trigger, Bruce G., L. Yaffe, M. Diksic, J.-L. Galinier, H. Marshall, and J. F. Pendergast 1980 Trace-element Analysis of Iroquoian Pottery. Canadian Journal of Archaeology 4:119-145. Tuck, James A. 1971 Onondaga Iroquois Prehistory: A Study in Settlement Archaeology. Syracuse University Press, Syracuse, N.Y. Tyyska, Allen E. 2017 Huron-Wendat Sweat Baths. Ontario Archaeology 97:64-78. Upham, Steadman 1982 Polities and Power: An Economic and Political History of the Western Pueblo. Academic Press, New York. van der Leeuw, Sander E. 2013 Archaeology, Networks, Information Processing, and Beyond. In Network Analysis in Archaeology: New Approaches to Regional Interaction, edited by C. Knappett, pp. 335-348. Oxford University Press, Oxford. Vatistas, Andreas 2011 A Typology and Seriation of Rim Sherds from the Forget Site. Report on file, Department of Anthropology, University of Toronto at Mississauga. Vygotskiĭ, Lev S. 1978 Mind in Society: The Development of Higher Psychological Processes. Harvard University Press, Cambridge, Mass. Warrick, Gary A. 1984 Reconstructing Ontario Iroquoian Village Organization. National Museum of Man, Ottawa. 1988 Estimating Ontario Iroquoian Village Duration. Man in the Northeast 36:21-60. 1990 A Population History of the Huron-Petun, A.D. 900-1650. Unpublished Ph.D. dissertation, Department of Anthropology, McGill University, Montreal. 2000 The Precontact Iroquoian Occupation of Southern Ontario. Journal of World Prehistory 14(4):415-466. 2008 A Population History of the Huron-Petun, A.D. 500-1650. Cambridge University Press, Cambridge. Warrick, Gary and Louis Lesage 2016 The Huron-Wendat and the St. Lawrence Iroquoians: New Findings of a Close Relationship. Ontario Archaeology 96:133-134. 309 Warrick, Gary and James Molnar 1986 An Iroquoian Sequence from Innisfil Township, Simcoe County. Arch Notes May/June:21-34. Wasserman, Stanley and Katherine Faust 1994 Social Network Analysis: Methods and Applications. Structural Analysis in the Social Sciences 8. Cambridge University Press, Cambridge. Watts, Duncan J. and Steven H. Strogatz 1998 Collective Dynamics of “Small-World” Networks. Nature 393:440-442. Waugh, Frederick 1916 Iroquois Foods and Food Preparation. Government Printing Bureau, Ottawa. Weber, Max 1978 [1922] Economy and Society: An Outline of Interpretive Sociology. 2 vols. University of California Press, Berkeley. Wendrich, Willeke 2012 Archaeology and Apprenticeship: Body Knowledge, Identity, and Communities of Practice. In Archaeology and Apprenticeship: Body Knowledge, Identity, and Communities of Practice, edited by W. Wendrich, pp. 1-19. University of Arizona Press, Tucson. Wenger, Etienne 1998 Communities of Practice: Learning, Meaning, and Identity. Learning in doing. Cambridge University Press, Cambridge. Whallon, Robert 1972 A New Approach to Pottery Typology. American Antiquity 37(1):13-33. 1980 On the Monothetic Nature of “Traditional” Types: A Contribution from Analysis of Owasco and Iroquois Ceramics. In Proceedings of the 1979 Iroquois Pottery Conference, edited by C. F. Hayes III, pp. 9-20. vol. 13. Rochester Museum and Science Center, Research Record. White, Marian E. 1971 Ethnic Identification and Iroquois Groups in Western New York and Ontario. Ethnohistory 18:13-18. Wiessner, Polly 1983 Style and Social Information in Kalahari San Projectile Points. American Antiquity 48:253-276. 1984 Reconsidering the Behavioral Basis of Style: A Case Study among the Kalahari San. Journal of Anthropological Archaeology 3(3):190-234. 1985 Style or Isochrestic Variation? A Reply to Sackett. American Antiquity 50:160-165. 2002 The Vines of Complexity: Egalitarian Structures and the Institutionalization of Inequality among the Enga. Current Anthropology 43(2):233-269. Willey, Gordon R. and Philip Phillips 1958 Method and Theory in American Archaeology. University of Chicago Press, Chicago. Wissler, C. 1914 Material Culture of North American Indians. American Anthropologist:16:447-505. Williamson, Ronald F. 1985 Glen Meyer: People in Transition. Unpublished Ph.D. dissertation, Department of Anthropology, McGill University, Montreal. 310 2007 "Otinontsiskiaj ondaon" ("The House of Cut-Off Heads"): The History and Archaeology of Northern Iroquoian Trophy Taking. In The Taking and Displaying of Human Body Parts as Trophies by Amerindians, edited by R. J. Chacon and D. H. Dye, pp. 190-221. Springer, New York. 2014 The Archaeological History of the Wendat to A.D. 1651: An Overview. Ontario Archaeology 94:3-64. 2016 East-West Interaction among Fifteenth-Century St. Lawrence Iroquoian and North Shore of Lake Ontario Ancestral Wendat Communities. Ontario Archaeology 96:104-120. Williamson, Ronald F., Meghan Burchell, William A. Fox and Sarah Grant 2016 Looking Eastward: Fifteenth-and Early Sixteenth-Century Exchange Systems of the North Shore Ancestral Wendat. In Contact in the 16th Century: Networks among Fishers, Foragers, and Farmers, edited by B. Loewen and C. Chapdelaine, pp. 235-255. Canadian Museum of Civilization, Gatineau, Quebec. Williamson, Ronald F. and Susan Pfeiffer (editors) 2003 Bones of the Ancestors: The Archaeology and Osteobiography of the Moatfield Ossuary. Mercury Series Paper No. 163. Canadian Museum of Civilization, Gatineau, Quebec. Williamson, Ronald F. and Terry G. Powis 1998 Parsons Site Ceramic Vessels. Ontario Archaeology 65/66:53-71. Williamson, Ronald F. and David A. Robertson 1994 Peer Polities Beyond the Periphery: Early and Middle Iroquoian Regional Interaction. Ontario Archaeology 58:27-48. 1998 The Archaeology of the Parsons Site: A Fifty Year Perspective. Ontario Archaeology 65/66. Williamson, Ronald F., Stephen Cox Thomas and Robert I. MacDonald 2003 The Archaeology of the Moatfield Village Site. In Bones of the Ancestors: The Archaeology and Osteobiography of the Moatfield Ossuary, edited by R. F. Williamson and S. Pfeiffer, pp. 19-88. Mercury Series Paper No.163. Canadian Museum of Civilization, Gatineau, Quebec. Wintemberg, William J. 1900 Indian Village Sites in the Counties of Oxford and Waterloo. Archaeological Report for Ontario 1899:83-92. 1901 Indian Village Sites in the Counties of Oxford and Waterloo. Annual Archaeological Report for Ontario 1900:37-40. 1903 Archaeology of Blenheim Township. Archaeological Report for Ontario 1902:58-70. 1913 The Archaeology of Blandford Township, Oxford County, Ontario. National Museum of Canada, Bulletin 1:187-200. 1939 Lawson Prehistoric Village Site, Middlesex County, Ontario. National Museum of Canada, Bulletin 94. Canada Dept. of Mines and Resources, Mines and Geology Branch, Ottawa. 1942 The Geographical Distribution of Aboriginal Pottery In Canada. American Antiquity 8(2):129141. Wobst, H. Martin 311 1977 Stylistic Behaviour and Information Exchange. In For the Director: Research Essays in Honor of James B. Griffin, edited by C. E. Cleland, pp. 317-342. , Anthropology Paper No. 61. University of Michigan, Museum of Anthropology, Ann Arbor. Wolf, Eric 1982 Europe and the People Without History. University of California, Berkeley and Los Angeles. Woolfrey, Sandra, Prince Chitwood and Norman E. Wagner 1976 Who Made the Pipes? A Study of Decorative Motifs on Middleport Pipes and Pottery Collections. Ontario Archaeology 27:3-12. Wonderley, Anthony W. 2004 Oneida Iroquois Folklore, Myth, and History: New York Oral Narrative from the Notes of H. E. Allen and Others. 1st ed. Iroquois and their neighbors. Syracuse University Press, Syracuse, N.Y. 2005 Iroquois Ceramic Iconography: New Evidence from the Oneida Vaillancourt Site. Ontario Archaeology 79/80:73-87. 2005 Effigy Pipes, Diplomacy, and Myth: Exploring Interaction between St. Lawrence Iroquoians and Eastern Iroquois in New York State. American Antiquity 70(2):211-240. Wray, Charles F. and Harry Schoff 1953 A Preliminary Report on the Seneca Sequence in Western New York—1550-1687. Pennsylvania Archaeologist 8(2):53-63. Wright, D. M. 1991 An Archaeobotanical Report Based upon Flotation Analysis of Plant Materials from the Keffer Site, Vaughn, Ontario. Museum of Indian Archaeology Research Report no. 22, London. Wright, James V. 1966 The Ontario Iroquois Tradition. National Museum of Canada, Bulletin 210. National Museum of Canada, Ottawa. 1974 The Nodwell Site. National Museum of Man, Ottawa. 1980 The Role of Attribute Analysis in the Study of Iroquoian Prehistory. In Proceedings of the 1979 Iroquois Pottery Conference, edited by C. F. Hayes III, pp. 21-26. vol. 13. Rochester Museum and Science Center, Research Record. Wrong, G. M. (editor) 1939 Sagard's Long Journey to the Huron. The Champlain Society, Toronto. Wylie, M. Alison 1989 Archaeological Cables and Tacking: The Implications of Practice for Bernstein's "Options Beyond Objectivism and Relativism". Philosophy of the Social Sciences 19(1):1-18. Zeitlin, Robert N. 1982 Toward a More Comprehensive Model of Interregional Commodity Distribution: Political Variables and Prehistoric Obsidian Procurement in Mesoamerica. American Antiquity 47(2):260275. Appendices Appendix A: Bcal Radiocarbon Dates for the Keffer Site The radiocarbon dates for the Keffer site are summarized in Table A-1 and Figure A-1. Table A-1. Calibrations of all the radiocarbon dates for the Keffer site. 312 313 Figure A-1. The 17 radiocarbon dates from the Keffer site run through Bcal Standard Calibration at 2 Sigma, 95% probability (calibrated by D. G. Smith 2019). 314 Appendix B: X-ray Fluorescence Analysis of Copper Artifacts from the Keffer Site 315 316 Appendix C: Previous Research on the Keffer Site Preliminary reports of the first excavations presented an initial analysis of botanical samples undertaken by Charles Turton (Finlayson et al. 1985:8), which documented the presence of the agricultural cultigens maize, beans, squash, sunflower, and tobacco and also provided evidence of the relatively long-term occupation of the Keffer village. According to Turton, the presence of maple and beech charcoal at the base of a village midden overlain by strata containing evidence of more varied mixed trees marks the initial settlement in an established forest. The non-climax trees seen in later deposits suggest that the area was entirely cleared of the original forest and that forest regrowth occurred during the occupation period (Finlayson et al. 1985:9). The use of the secondary growth forest by Keffer residents argues for a long duration of occupation of the village. Dawn Wright (1991) undertook a larger, more thorough analysis of selected midden and house botanical remains. Frances Stewart’s (1991) analyzed a portion of the Keffer site fauna; David G. Smith’s (1991) studied the site ceramics; Harry Lerner’s (1997) examined the projectile points and chipped lithics; Bruce Jamieson’s (1991) analyzed the modified bone, antler, tooth, and shell artifacts. Michael Spence’s (1988) also reported on the excavation and analysis of the burials, and Stephen Cox Thomas (n.d.) analyzed the dog burial. Finlayson and his team published a summary of the 1985 excavation season and a description of the artifacts, in 1987 (Finlayson et al. 1987), prior to the 1988 excavations of the southern portion of the village. With financial support from a University of Toronto Graduate Research Grant to Susan Dermarkar, Brandi-Lee MacDonald performed x-ray fluorescence analysis of two Keffer copper pieces, one scrap and one rolled bead. Both pieces are of indigenous manufacture, supporting the assignment of the Keffer site to the precontact era. This research grant also supported six radiocarbon AMS dates, performed by Direct AMS and Beta Analytic. These analyses examined pot residue originating from vessels found in diverse areas of the village (see Appendix A). The Keffer site collections from the 1985 and 1988 excavations, housed at the Museum of Ontario Archaeology (formerly Museum of Indian Archaeology) in London, have also been the subject of a large number of academic research projects. Analyses of previously unstudied house and midden faunal samples from the site have been undertaken by undergraduate students in the University of Toronto St. George campus faunal analysis lab class, taught by Max Friesen, in 317 2016. The results of these analyses are on file at the Department of Anthropology at University of Toronto St. George campus. University of Toronto student Rachel Gruber (2014) performed a preliminary analysis of Keffer pipes and their relation with other north shore Iroquoian collections. Harry Lerner completed his Master’s thesis on chipped lithic artifacts, using the Keffer collection (Lerner 2000), and he continues to publish on the material (Lerner 2015). Dori Rainey’s Master’s thesis investigating the implications of scattered and butchered human bone found throughout the site was completed at the University of Western Ontario (Rainey 2002). Frances Stewart’s (1996) PhD thesis, comparing faunal resource use at the Keffer site and the St. Lawrence Iroquoian McKeown site, was the first thesis to come out of the Keffer site excavations. This was followed by Bruce Jamieson’s PhD thesis (2016), based on the Keffer bone artifacts, and, most recently, by Steven Dorland’s PhD thesis (2018), which examines the Keffer learner ceramics in the interpretation of the dynamics of learning within Wendat ceramic communities of practice. Working with Keffer whole vessels, Steve Bull (1987, 1989), at the University of Western Ontario, published two papers on the relations between complete pots and their rim sherds. In addition, Keffer has been the subject of several academic presentations and publications These include the work of John Creese on the Keffer pipes and identity (2013, 2014), my own (Dermarkar 2012, 2013a, 2013b, 2014b, 2016) presentations on the Keffer site and its ceramic relations, Tara Jenkins’ work on the human bone artifacts (Jenkins 2014, 2016), Erika Johannsen’s (2014) research into Keffer castellation typology, Harry Lerner’s (2011) study of the chipped lithic artifact typologies, David G. Smith’s presentation on Keffer’s settlement pattern (2015), Michael Spence and Dori Rainey’s (2017) study of the cultural significance of the Keffer site human burials, and Frances Stewart’s (1991b) examination of the implications of screen mesh size on faunal recovery rates. Although a large amount of research has been done on the Keffer site, this large and relatively complete collection lends itself to further study. 318 Appendix D: All Ceramic Types Represented in the Keffer Site Assemblage In most cases, the information for these types is taken from text or, where no text is available, from plates, in MacNeish (1952). The page numbers have been noted below. In some cases, other sources have been used, as noted below. EARLY ONTARIO Middleport Oblique (MacNeish 1952:16–17) Vessel shape: probably globular Neck shape: slightly constricted Collar shape: slightly outflaring rims with poorly defined, incipient collars Collar height: short, 1 to 2.5 cm Exterior profile: unspecified Interior profile: unspecified Decoration: short, parallel oblique lines on upper rim; horizontal line or lines on lower rim or at base of incipient collar; sometimes notches or linear punctates on base Castellations: unknown Ontario Horizontal (MacNeish 1952:16) Vessel shape: globular Neck shape: slightly constricted Collar shape: earlier, outflaring and poorly defined; later, sharply defined, straight, vertical Collar height: short, 1 to 2.5 cm Exterior profile: convex to straight Interior profile: concave Decoration: two to five horizontal incised lines on collar; earlier, sometimes vertical or oblique gashes above and/or below horizontal lines; later, sometimes ovoid notches at base of collar Castellations: unknown Ontario Oblique (MacNeish 1952:18) 319 Vessel shape: globular Neck shape: slightly constricted Collar shape: poorly defined Collar height: short, 1 to 2.5 cm Exterior profile: slightly convex to straight Interior profile: concave Decoration: bands of horizontal obliques of opposing direction on top of each other, often with more bands of obliques or horizontal lines below Castellations: rarely Note: MacNeish (1952:18) states that this type has a “criss cross” design, but that implies lines crossing each other, which they do not. Pound Blank (MacNeish 1952:15–16) Vessel shape: globular Neck shape: slightly constricted Collar shape: well defined Collar height: short, 1 to 2.5 cm Exterior profile: convex Interior profile: concave Decoration: alternating wide-based, undecorated isosceles triangles, separated by four oblique lines; commonly punctates or gashes form the base of the triangles Castellations: four, with vertical band of four parallel lines with two undecorated, right-angle triangles on either side, under castellations Pound Neck (MacNeish 1952:14–15) Vessel shape: globular Neck shape: slightly constricted Collar shape: poorly defined, with increasing thickness towards the rim Collar height: short, 1 to 2.5 cm Exterior profile: straight to convex Interior profile: concave 320 Decoration: horizontal incisions encircling the neck; oblique or vertical, parallel incisions on collar Castellations: unknown SOUTH-CENTRAL ONTARIO “Huron/Neutral” Copeland Incised (Wright 1966:73) Vessel shape: unknown Neck shape: unknown Collar shape: sharply defined collar base Collar height: short, 1 to 2.5 cm Exterior profile: unknown Interior profile: concave or convex Decoration: incised horizontal line broken by vertical or oblique lines; may have ovate basal and/or rim punctates; no lip or interior decoration Castellations: unknown Huron Incised (MacNeish 1952:34) Vessel shape: globular body with outflaring rim Neck shape: constricted Collar shape: well defined Collar height: short, 1 to 2.5 cm Exterior profile: commonly concave; sometimes straight Interior profile: straight or convex Decoration: parallel oblique or vertical lines around the collar, rarely, inside opposed triangles; sometimes notches above or below decoration Castellations: one to four; earlier, pointed; later, sometimes pointed, commonly squared Seed Incised (MacNeish 1952:35) Vessel shape: globular body with outflaring collar Neck shape: constricted Collar shape: well defined Collar height: short, 1 to 2.5 cm 321 Exterior profile: concave Interior profile: usually convex; more rarely, straight Decoration: short, vertical gashes at top and base of collar; sometimes gashes in middle of collar Castellations: pointed, squared, or bifurcated Sidey Crossed (MacNeish 1952:36) Vessel shape: globular body with outflaring collar Neck shape: constricted Collar shape: well defined Collar height: short, 1 to 2.5 cm Exterior profile: concave Interior profile: straight or convex Decoration: closely spaced oblique or vertical parallel lines that are crossed in the centre by a solid horizontal line or a broken horizontal line Castellations: present Sidey Notched (MacNeish 1952:33) Vessel shape: globular body with outflaring collar Neck shape: constricted Collar shape: well defined Collar height: short, 1 to 2.5 cm Exterior profile: concave Interior profile: convex Decoration: closely spaced oblique or vertical parallel lines; rarely, forming opposed triangles; lips always notched; neck not decorated Castellations: one, two, or four Note: MacNeish (1952) does not include notched Warminster Crossed in this type. In this analysis, what MacNeish (1952) refers to as notched Warminster Crossed is assigned to the Sidey Notched type. Warminster Crossed (MacNeish 1952:32) Vessel shape: globular body with outflaring collar 322 Neck shape: very constricted Collar shape: well defined Collar height: short, 1 to 2.5 cm Exterior profile: concave Interior profile: convex Decoration: closely spaced oblique lines crossed by widely spaced oblique lines in the opposite direction, on short, outflaring collar that is either convex or straight Castellations: squared and notched Note: This type may have opposed incised-line triangles on the neck. In this analysis, these vessels are typed as Black Necked. Warminster Horizontal (MacNeish 1952:34–35) Vessel shape: globular body with outflaring collar Neck shape: constricted Collar shape: well defined Collar height: short, 1 to 2.5 cm Exterior profile: concave Interior profile: usually convex; more rarely, straight Decoration: three to seven horizontal lines on the collar, usually with vertical or oblique gashes at base of collar; no neck deck Castellations: usually pointed; sometimes squared or bifurcated Black Necked (MacNeish 1952:36) Vessel shape: globular body with outflaring collar Neck shape: constricted Collar shape: well defined Collar height: short, 1 to 2.5 cm Exterior profile: straight to concave Interior profile: straight or convex Decoration: a variety of oblique lines, etc., vertical lines, opposed triangles, horizontal lines, with or without broken horizontals (Sidey Crossed) on the rim, with or without basal collar notches (i.e., Warminster Horizontal); there may be notches on the top of the collar (Sidey Notched); 323 opposed triangles filled with oblique lines on the neck; rarely, only horizontal incisions on the neck Castellations: pointed Lawson Incised (MacNeish 1952:14) Vessel shape: globular body Neck shape: constricted Collar shape: collar blends gradually into the neck; poorly defined collar Collar height: short, 1 to 2.5 cm Exterior profile: straight or concave Interior profile: convex Decoration: incised parallel line, usually oblique to the right or vertical; less commonly obliques to the left; rarely, bases notched Castellations: one, pointed Lawson Opposed (MacNeish 1952:13–14) Vessel shape: globular body Neck shape: constricted Collar shape: collar blends gradually into the neck; poorly defined collar Collar height: short, 1 to 3.5 cm Exterior profile: straight or concave Interior profile: concave Decoration: opposed triangles or bands of oblique, parallel lines Castellations: one, pointed Lalonde High Collar (Ridley 1952:205) Vessel shape: elongated body more than 30 cm in height; shoulder diameter greater than collar diameter Neck shape: constricted and long, at least two times the collar height Collar shape: well defined Collar height: 8 cm or greater Exterior profile: concave 324 Interior profile: straight Decoration: two to three closely spaced incised horizontal lines on collar base, sometimes on upper rim, sometimes on both the collar base and the upper rim; lines connect or stop at castellation hachuring; central area of collar has open triangles circumscribed with punctates; sometimes other spaces hachured with oblique or horizontal lines Castellations: two, weak, opposed to each other, with numerous vertical hachure lines under them Sopher Incised (Latta 1983:39; Noble 1968:172–173) Vessel shape: unknown Neck shape: slightly constricted Collar shape: wider at the base and tapered to the lip Collar height: high, no specifics Exterior profile: straight Interior profile: straight to concave Decoration: vertical punctates at base of collar, with vertical and oblique plaits divided by a diagonal ladder plait; sometimes short, vertical strokes below the rim High Collared (Fitzgerald 2014; Sutton 1990) Vessel shape: unknown Neck shape: slightly to well constricted Collar shape: generally well defined; upright Collar height: greater than 35 mm; generally greater than 50 mm Exterior profile: slightly concave or straight Interior profile: straight, convex, or concave Decoration: large variety of oblique and triangular motifs Castellations: often; present in various shapes Note: This is a mode, not a type. NIAGARA AREA/ERIE Niagara Collared (MacNeish 1952:26) Vessel shape: globular body 325 Neck shape: slightly constricted Collar shape: poorly defined Collar height: short, 2.5 cm Exterior profile: convex Interior profile: concave Decoration: usually absent; rarely, lips incised Castellations: one Ripley Plain (MacNeish 1952:25–26) Vessel shape: globular body Neck shape: in-sloping Collar shape: in-sloping Collar height: short, 2.5 cm Exterior profile: convex Interior profile: concave Decoration: usually absent; rarely, lips incised Castellations: unknown Ripley Collared (MacNeish 1952:26–27) Vessel shape: globular body Neck shape: slightly constricted Collar shape: poorly defined Collar height: short, 1 to 2.5 cm Decoration: gashes or punctates at base of collar Castellations: unknown Exterior profile: unspecified Interior profile: unspecified WESTERN NEW YORK STATE Dutch Hollow Notched (MacNeish 1952:43) Vessel shape: small, globular body with outflaring rim Neck shape: long; slightly constricted 326 Collar shape: thickened lip Collar height: n/a Exterior profile: commonly thickened lips of flat or convex shape Interior profile: convex Decoration: notches on the rim adjacent to the lip or notches or gashes cutting the thickened lip and/or outer rim Castellations: none Seneca Notched (MacNeish 1952:43) Vessel shape: globular body Neck shape: short; slightly constricted Collar shape: appliquéd; poorly attached Collar height: n/a Exterior profile: straight or slightly concave Interior profile: various; some concave Decoration: long notches at base of rim; sometimes notches at top of collar Castellations: rarely; if present, one or two Cayuga Horizontal (MacNeish 1952:52) Vessel shape: globular body with incipient collar Neck shape: short; constricted Collar shape: poorly defined Collar height: high, 6 to 9 cm Exterior profile: from convex to concave Interior profile: concave Decoration: earlier, horizontal line motifs; later, horizontal lines cut at wide intervals by oblique bands of three to four parallel incised lines; sometimes horizontal lines have band of verticals crossing them (usually under castellations); sometimes round notches at base Castellations: two to four Richmond Incised (MacNeish 1952:51) Vessel shape: globular body with incipient collar 327 Neck shape: short; constricted Collar shape: poorly defined Collar height: high, 6 to 9 cm Exterior profile: from convex to concave Interior profile: concave Decoration: incised vertical or oblique incised lines, alone or in opposed triangles, usually with notches on collar base; rarely, one or two horizontal lines at top of collar Castellations: two to four, often with bands of four to six vertical lines EASTERN NEW YORK STATE Onondaga Triangular (MacNeish 1952:60) Vessel shape: unknown Neck shape: slightly constricted Collar shape: well defined Collar height: medium, 2.5 to 3.5 cm Exterior profile: concave to convex Interior profile: concave to straight Decoration: right-angle triangles filled with parallel incised lines; a single or double horizontal line encircles the top of the rim, with a band of short, vertical impressions at the top of the collar; variant has a diagonal line of punctates or ladder plaits across the hypotenuse of the triangles; later, oval-shaped notches at base Castellations: two to four Syracuse Incised (MacNeish 1952:59) Vessel shape: unknown Neck shape: unstated Collar shape: well defined; thin Collar height: medium, 2.5 to 3.5 cm Exterior profile: convex to concave Interior profile: usually concave; sometimes straight Decoration: oblique line decoration, generally of opposed triangles filled with oblique lines; no notching on collar base or rim; some variants may have punctates above and below rim 328 Castellations: present Thurston Horizontal (MacNeish 1952:66, 69) Vessel shape: globular body Neck shape: slightly constricted Collar shape: well defined Collar height: medium to high, usually 2.5 to 4 cm, may be up to 6 cm Exterior profile: usually convex; sometimes concave Interior profile: well defined; concave Decoration: two to seven horizontal lines in middle of collar; above and below are bands of short, closely spaced vertical lines; sometimes notched or unnotched base; in historic period, often effigies under castellations Castellations: two Cayadutta Incised (MacNeish 1952:77–78) Vessel shape: globular body Neck shape: slightly constricted Collar shape: well defined Collar height: medium, 2.5 to 3.5 cm Exterior profile: convex Interior profile: concave Decoration: one to two horizontal lines, usually below bands of short, closely spaced vertical lines just below lip; verticals may be absent; below this are parallel oblique or vertical incised lines, opposed triangles filled with oblique lines, or oblique lines separated by a band of vertical, incised oblique line with crescent punctates or oblique notches; earlier, oblique bands of short linear punctates or opposed triangles outlined by four parallel lines on the neck Castellations: low, pointed Chance Incised (MacNeish 1952:78) Vessel shape: globular body Neck shape: constricted Collar shape: sharply defined base 329 Collar height: low to medium Exterior profile: straight to convex Interior profile: concave Decoration: single or double horizontal line below band of short, vertical impressions on top of collar; below this are a variation of design motifs, including oblique lines, opposed oblique lines, vertical lines, vertical and oblique lines, horizontal lines, horizontal lines broken by triangles filled with oblique parallel lines; no collar base notches Neck decoration: oblique plaits of linear punctates; commonly four oblique lines forming triangles or herringbone design Castellations: present Fonda Incised (MacNeish 1952:76) Vessel shape: globular body Neck shape: constricted Collar shape: well developed Collar height: medium, 2.5 to 3.5 cm; increasing with time Exterior profile: highly convex to straight Interior profile: commonly concave Decoration: multiple horizontal lines below band of short, vertical impressions on top of collar; horizontal lines may be broken by oblique line–filled triangles Neck decoration: earlier, opposed-line triangle; crescent-shaped or elliptical notches on collar base Castellations: two to four Otstungo Incised (MacNeish 1952:76–77) Vessel shape: round body Neck shape: slightly constricted Collar shape: well developed Collar height: medium to high, greater than 3.5 cm Exterior profile: straight Interior profile: concave but with a lot of variation 330 Decoration: one to two horizontal lines, usually below bands of short, closely spaced vertical line just below lip; verticals may be absent; below this are parallel oblique or vertical incised lines, opposed triangles filled with oblique lines, or oblique lines separated by band of vertical incised oblique line; collar base has round or elliptical notches or punctates Castellations: incipient (deduced from Plate xxxi) Otstungo Notched (MacNeish 1952:60) Vessel shape: globular body with outflaring rim Neck shape: straight or very slightly constricted Collar shape: outflaring; thickened rim Collar height: n/a Exterior profile: straight lip on everted rim Interior profile: convex Decoration: on top of thickened lip; earlier, dominant motif of one or two encircling incisions parallel to edge of lip, with interior and/or exterior bands of notches or short lines at right angles to encircling incisions; other motifs have opposed oblique lines Castellations: none Note: Originally seen as Mohawk but may have regional origin. A minority ware throughout Mohawk history; reaches its peak in historic times. Rice Diagonal (MacNeish 1952:60) Vessel shape: globular body with outflaring rim Neck shape: constricted Collar shape: outflaring; thickened rim Collar height: n/a Exterior profile: straight lip Interior profile: convex Decoration: on top of thickened lip; oblique parallel lines on thickened lips or straight parallel lines at right angles to lip; rarely, interior and exterior rim notching Castellations: none Note: Originally seen as Mohawk but may have regional origin. 331 Wagoner/Wagner Incised (MacNeish 1952:75) Vessel shape: globular body Neck shape: constricted Collar shape: sharply defined base Collar height: medium, 5 cm Exterior profile: concave Interior profile: concave opposite collar base Decoration: oblique or vertical parallel incised lines; commonly opposed triangles filled with parallel lines; circular notches at base Castellations: incipient Durfee Underlined (MacNeish 1952:60) Vessel shape: unknown Neck shape: slightly constricted Collar shape: well defined Collar height: medium Exterior profile: convex to concave Interior profile: commonly concave; rarely, straight Decoration: one or two horizontal lines at base and rim of collar; sometimes between these horizontals, oblique-line, opposed triangles or opposed, right-angled triangles filled with parallel lines; sometimes undecorated triangles outlined by incised lines with or without gashes along the edge; later, basal notches Castellations: two to six; earlier, sometimes hollow reed punctates below castellations; later, sometimes effigy faces or figures Note: Originally seen as Onondaga by MacNeish (1952). Now seen as St. Lawrence valley and Northern New York (William Engelbrecht 2019 Personal Communication). Roebuck Low Collar (MacNeish 1952:61, Figures 24.1–24.20) Vessel shape: globular body Neck shape: constricted Collar shape: well defined Collar height: short 332 Exterior profile: convex to straight Interior profile: concave Decoration: commonly oblique lines, often forming opposed, isosceles or right-angled triangles on the bottom half of the collar, with one or two horizontal lines at the top; shows designs common to Chance Incised, Otstungo Incised, or Durfee Underlined; sometimes basal notches Castellations: one or two; sometimes three hollow reed punctates under castellation(s); rarely, handles under castellation(s) Note: Originally seen as Onondaga by MacNeish (1952). Now seen as St. Lawrence valley and general area. Lanorie Mixed (MacNeish 1952:63) Vessel shape: globular or elongated–globular body with rounded base Neck shape: slightly constricted Collar shape: well defined Collar height: medium, 2.5 to 3.5 cm Exterior profile: usually straight; sometimes convex Interior profile: concave Decoration: designs are complex and can vary; made by two or more decorative techniques of incising, dentate stamping, or cordwrapped-paddle-edge impression Castellations: unknown Note: This name represents a spelling error in the name Lanoraie by MacNeish (1952). Summary statistics for the ceramic types discussed in the text are provided in Tables D-1 and D2. 333 Table D-1. Ceramic Types Present at the Keffer Site. Type Huron Incised Black Necked Lawson Incised Sidey Crossed Sidey Notched Warminster Crossed Otstungo Notched Durfee Underlined High Collar 15th cent. Niagara Collared Pound Necked Rice Diagonal Warminster Horizontal Dutch Hollow Notched Syracuse Incised Roebuck Low Collar Wagoner incised Lalonde High collar Seed Incised Copeland Incised Onondaga Triangular Thurston Horizontal Cayuga Horizontal Fonda Incised Lawson Opposed Richmond Incised Ripley Plain Ontario Horizontal Cayadutta Incised Chance Incised Ripley Collared Middleport Crisscross Otstungo Incised Middleport Oblique Ontario Oblique Pound Blank Unknown Total Number of Vessels 2728 960 357 303 246 157 126 99 56 50 49 49 49 47 47 45 37 33 25 19 19 19 17 13 13 9 8 5 4 4 4 2 2 1 1 1 69 5674 Number of Vessels as % of Site Total 48.1 16.9 6.3 5.3 4.3 2.8 2.2 1.7 1.0 0.9 0.9 0.9 0.9 0.8 0.8 0.8 0.7 0.6 0.4 0.3 0.3 0.3 0.3 0.2 0.2 0.2 0.1 0.1 0.1 0.1 0.1 0.0 0.0 0.0 0.0 0.0 1.2 100.0 Total Number of Vessels of Unknown Provenience 104 36 33 6 12 3 4 4 8 2 3 1 3 4 0 0 1 0 1 2 1 1 0 0 1 2 2 0 0 0 0 0 0 1 1 0 5 241 Number of Vessels of Unknown Provenience as % of Site Total 3.8 3.8 9.2 2.0 4.9 1.9 3.2 4.0 14.3 4.0 6.1 2.0 6.1 8.5 0.0 0.0 2.7 0.0 4.0 10.5 5.3 5.3 0.0 0.0 7.7 22.2 25.0 0.0 0.0 0.0 0.0 0.0 0.0 100.0 100.0 0.0 7.2 4.2 334 Table D-2. Keffer Site Ceramic Types Percentage by Community of Practice. Local Tradition Total Vessels 39 9 6 22 8 25 17 40 13 37 112 131 1 20 1 55 261 1 62 153 60 127 169 49 522 87 530 413 208 12 966 73 18 21 103 87 51 644 31 144 25 80 241 5674 Local CoP 2 4 5 4 4 4 3 2 2 4 3 1 2 5 3 4 1 2 5 2 5 2 5 3 5 2 2 2 3 4 5 2 5 5 2 3 3 3 5 2 5 2 Non-House /Midden Total Vessels House/ Midden H1 H2 H3 H4 H6 H7 H8 H9 H10 H11 H12 H13 H14 H15 H18 H19 H20 H16 M51 M52 M53 M54 M55 M56 M57 M59 M60 M61 M62 M63 M65 M66 M67 M68 M71 M72 M73 M74 M75 M77 M78 M80 Black Necked 15.4 11.1 0 13.6 12.5 12.0 17.6 7.5 15.4 10.8 18.8 20.6 0 40 0 5.5 21.8 0 22.6 19.0 26.7 17.3 23.7 20.4 17.4 13.8 14.5 14.0 17.3 16.7 17.5 9.6 11.1 19.0 14.6 18.4 13.7 18.2 9.7 13.9 16.0 13.8 Huron Incised 43.6 66.7 50 59.1 37.5 84 23.5 60.0 61.5 48.6 44.6 42.7 0 50 0 67.3 54.4 100 41.9 51.6 38.3 56.7 52.1 44.9 44.4 56.3 53.0 62.5 39.4 58.3 39.0 47.9 50 38.1 52.4 44.8 56.9 48.3 51.6 43.1 32.0 55.0 Copeland Incised 0 0 0 0 0 0 0 0 0 0 0 0.8 0 0 0 0 0 0 1.6 0 0 0.8 0.6 0 1.3 0 0.4 0.2 0.5 0 0.2 0 0 0 0 0 0 0 0 0 0 0 Seed Incised 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1.8 1.1 0 0 0 0 0.8 0.6 0 0.4 0 0.6 0.7 0.5 0 0.7 2.7 0 0 0 0 0 0 0 0 0 0 Sidey Crossed 15.4 0 16.7 13.6 12.5 0 17.6 5.0 7.7 8.1 4.5 9.2 0 0 0 5.5 3.4 0 6.5 11.8 5.0 4.7 2.4 6.1 4.8 5.7 7.4 2.7 2.4 0 4.1 4.1 0 0 7.8 13.8 0 5.4 3.2 9.7 8.0 12.5 Sidey Notched 0 0 0 0 12.5 0 11.8 0 0 13.5 2.7 0.8 0 0 0 0 3.8 0 3.2 2.6 6.7 6.3 1.8 0 2.9 5.7 3.2 1.5 1.9 8.3 9.5 4.1 11.1 4.8 2.9 4.6 3.9 4.7 3.2 2.8 0 1.3 10.4 960 53.3 2728 0.2 19 0.1 25 4.7 303 5.6 246 335 Table D-2. Keffer Site Ceramic Types by Community of Practice. (cont.) Local House/ CoP Midden 2 H1 4 H2 5 H3 4 H4 4 H6 4 H7 3 H8 2 H9 2 H10 4 H11 3 H12 1 H13 2 H14 5 H15 3 H18 4 H19 1 H20 2 H16 5 M51 2 M52 5 M53 2 M54 5 M55 3 M56 5 M57 2 M59 2 M60 2 M61 3 M62 4 M63 5 M65 2 M66 5 M67 5 M68 2 M71 3 M72 3 M73 3 M74 5 M75 2 M77 5 M78 2 M80 Non-H/M Total Vessels War minster Crossed 2.6 11.1 0 0 0 0 5.9 0 0 0 7.1 4.6 0 0 0 3.6 3.4 0 3.2 1.3 3.3 0 0 0 3.1 1.1 2.6 1.2 3.8 0 4.1 2.7 0 0 5.8 3.4 0 3.3 0 2.8 0 0 0.9 157 Local Tradition (cont.) Warm Lalonde inster High High Lawson Horizontal collar Collar Incised 5.1 2.6 0 0 0 0 0 0 0 0 16.7 16.7 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5.9 5.9 0 0 0 12.5 0 0 0 7.7 0 0 0 0 2.7 0 0.9 1.8 2.3 0.8 0.0 0 0 100 0 0 5 0 0 0 100 0 0 0 1.8 0 0 0 0.8 0.8 0.4 0 0 0 0 0 0 0 1.6 9.7 0.7 0 0.7 5.2 0 1.7 1.7 3.3 0.8 0 0.0 5.5 0.6 0 0.6 8.9 2.0 0 4.1 2.0 0.4 1.5 1.7 11.1 1.1 0 0 2.3 0.8 0.8 1.1 5.8 0 0.2 1.0 4.6 0.5 0.5 0.5 3.4 0 0 8.3 0 1.2 0.7 0.5 7.5 1.4 0 0 5.5 0 0 0 16.7 0 0 0 19.0 1.9 0 0 3.9 0 1.1 0 9.2 0 2.0 2.0 5.9 0.8 0.2 0.9 6.1 0 6.5 0 19.4 0.7 0 1.4 3.5 0 0 8.0 24 0 1.3 1.3 5.0 0.8 0.0 4.8 10.0 49 33 56 357 Lawson Opposed 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1.8 0 0 0 0 0 0 0 0 0.2 0 0.8 0.5 0.5 0 0.2 0 0 0 0 0 0 0.2 0 0 0 0 0.1 13 336 Table D-2. Keffer Site Ceramic Types by Community of Practice. (cont.) Early Ontario House/ Midde n H1 H2 H3 H4 H6 H7 H8 H9 H10 H11 H12 H13 H14 H15 H18 H19 H20 H16 M51 M52 M53 M54 M55 M56 M57 M59 M60 M61 M62 M63 M65 M66 M67 M68 M71 M72 M73 M74 M75 M77 M78 M80 Local CoP 2 4 5 4 4 4 3 2 2 4 3 1 2 5 3 4 1 2 5 2 5 2 5 3 5 2 2 2 3 4 5 2 5 5 2 3 3 3 5 2 5 2 Non-H/M Total Vessels Middle port Crisscros s 0 0 0 0 12.5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.2 0 0 0 0 0.0 2 Middle port Obliqu e 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.5 1 Ontario Horizonta l 0 0 0 0 0 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.6 0 0.2 0 0 0 0 0 0 0 0 0 0 0 0 0.3 0 0 0 0 0.0 5 Ontario Obliqu e 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.5 1 Erie area Poun d Blank 2.6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.0 1 Pound Necke d 0 0 0 0 0 0 0 0 0 2.7 0 0.8 0 0 0 0 0 0 0 1.3 0 0 3.0 0 2.1 0 0.8 0.2 0 0 1.3 0 11.1 0 0 0 2.0 0.8 0 0 0 0 0.9 49 Niagara Collare d 0 0 0 0 0 0 0 2.5 0 0 3.6 0 0 5 0 0 0.4 0 3.2 0 5 0 0.6 0 0.4 5.7 0.6 1.5 1.4 0 0.9 1.4 0 4.8 0 0 0 0.5 0 0.7 0 1.3 0.8 50 Ripley Collare d 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.8 0 0 0 0 0 0 0 0 0.2 0 0 0 0 0 0 0.2 0 0 0 0 0.0 4 Riple y Plain 0 0 0 0 0 0 0 0 0 0 0.9 0 0 0 0 0 0.4 0 0 0 1.7 0 0 0 0 0 0 0 0 0 0.2 0 0 0 0 0 2.0 0 0 0 0 0 0.2 8 337 Table D-2. Keffer Site Ceramic Types by Community of Practice. (cont.) Northern New York–St. Lawrence Local House/ CoP Midden 2 H1 4 H2 5 H3 4 H4 4 H6 4 H7 3 H8 2 H9 2 H10 4 H11 3 H12 1 H13 2 H14 5 H15 3 H18 4 H19 1 H20 2 H16 5 M51 2 M52 5 M53 2 M54 5 M55 3 M56 5 M57 2 M59 2 M60 2 M61 3 M62 4 M63 5 M65 2 M66 5 M67 5 M68 2 M71 3 M72 3 M73 3 M74 5 M75 2 M77 5 M78 2 M80 Non-H/M Total Vessels Durfee Underlined 2.6 0 0 0 0 0 0 0 0 5.4 0 3.8 0 0 0 3.6 2.3 0 0 0 0 0 0.6 0 0.6 1.1 2.6 1.7 1.4 8.3 2.6 4.1 0 0 1.9 2.3 2.0 0.8 0 4.9 4.0 3.8 1.4 99 Haudenosaunee Territory Roebuck Low Collar Otstungo Notched 0 0 0 0 0 0 0 2.5 0 0 0 2.3 0 0 0 0 1.5 0 0 0 0 0.8 1.2 2.0 0.4 2.3 0.6 0.7 0 0 1.4 2.7 0 0 1.0 0 2.0 0.5 0 0 4 1.3 0.0 45 0 0 0 0 0 0 0 0 0 0 3.6 1.5 0 0 0 3.6 1.1 0 0 0.7 0 0 0.6 2.0 0.8 1.1 0.9 1.5 14.9 0 2.4 1.4 0 4.8 1.9 2.3 0 3.9 3.2 3.5 0 1.3 1.0 126 Rice Diagonal 2.6 0 0 0 0 0 0 0 7.7 0 0 2.3 0 0 0 1.8 0 0 0 0 1.7 2.4 0.6 2.0 0.2 0 1.1 0.7 2.4 0 0.4 4.1 0 0 2.9 0 0 1.4 0 1.4 0 0 0.2 49 338 Table D-2. Keffer Site Ceramic Types by Community of Practice. (cont.) Haudenosaunee Territory (cont.) Local House/ CoP Midden 2 H1 4 H2 5 H3 4 H4 4 H6 4 H7 3 H8 2 H9 2 H10 4 H11 3 H12 1 H13 2 H14 5 H15 3 H18 4 H19 1 H20 2 H16 5 M51 2 M52 5 M53 2 M54 5 M55 3 M56 5 M57 2 M59 2 M60 2 M61 3 M62 4 M63 5 M65 2 M66 5 M67 5 M68 2 M71 3 M72 3 M73 3 M74 5 M75 2 M77 5 M78 2 M80 Non-H/M Total Vessels Dutch Hollow Notched 2.6 11.1 0 0 0 0 0 2.5 0 2.7 0.9 2.3 0 0 0 0 0.8 0 3.2 1.3 1.7 0 0 10.2 0.6 1.1 0.6 0.2 2.4 0 0.2 4.1 0 0 1.0 0 2.0 0.2 0 0.7 0 1.3 0.6 Seneca Notched 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.1 0 0 0 0 0 0 0 0 0 0 0 0.0 Cayuga Horizontal 0 0 0 0 0 0 0 0 0 0 1.8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.2 0.2 1.0 0 0.4 0 0 4.8 1.0 0 2.0 0.2 0 2.1 0 0 0.0 Thurston Horizontal 0 0 0 4.5 0 0 0 0 0 2.7 0.9 0 0 0 0 3.6 0 0 0 0.7 0 0 0 0 0 0 0 0.2 0 0 0.4 1.4 0 0 0 0 0 0.3 0 2.8 0 0 0.1 Onondaga Triangular 0 0 0 4.5 0 0 5.9 0 0 0 0.9 0 0 0 0 0 0.8 0 0 0 0 0 0 0 0 0 0.8 0 0.5 0 0.2 0 0 0 0 0 0 0.6 0 0.7 4.0 0 0.7 Richmond Incised 0 0 0 0 0 0 0 0 0 2.7 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1.1 0.2 0.2 0 0 0.1 1.4 0 0 0 0 0 0.2 0 0 0 0 0.2 Syracuse Incised 0 0 0 0 0 0 5.9 5.0 0 2.7 0 0 0 0 0 0 0.8 0 0 1.3 1.7 1.6 0 2.0 1.1 0 0.4 1.2 1.9 0 1.0 0 0 0 0 0 0 0.8 0 2.1 0 0 0.0 47 1 17 19 19 9 47 339 Table D-2. Keffer Site Ceramic Types by Community of Practice. (cont.) Haudenosaunee Territory (cont.) Local CoP 2 4 5 4 4 4 3 2 2 4 3 1 2 5 3 4 1 2 5 2 5 2 5 3 5 2 2 2 3 4 5 2 5 5 2 3 3 3 5 2 5 2 Non-H/M Total Vessels House/ Midden H1 H2 H3 H4 H6 H7 H8 H9 H10 H11 H12 H13 H14 H15 H18 H19 H20 H16 M51 M52 M53 M54 M55 M56 M57 M59 M60 M61 M62 M63 M65 M66 M67 M68 M71 M72 M73 M74 M75 M77 M78 M80 Wagoner incised 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.4 0 1.6 0 1.7 1.6 0 0 1.9 0 0 1.2 1.0 0 0.8 0 0 0 0 0 0 0.6 0 1.4 0 0 0.1 37 Cayadutta Incised 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.8 0 0 0 0 0 0 0 0.2 0 0 0 0 0 0.1 0 0 0 0 0 0 0 0 0 0 0 0.0 4 Chance Incised 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.1 0 0 0 0 0 0.2 0.3 0 0 0 0 0.0 4 Fonda Incised 0 0 0 4.5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1.1 0 0.7 1.4 0 0.2 1.4 0 4.8 0 0 0 0 0 0.7 0 0 0.0 13 Otstungo Incised 0 0 0 0 0 0 0 2.5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.1 0 0 0 0 0 0 0 0 0 0 0 0.0 2 Untypable (%) 5.1 0 0 0 12.5 0 0 0 0 0 4.5 5.3 0 0 0 0 0.8 0 1.6 2.0 0 0 1.8 2.0 2.3 0 0.4 0.5 0.5 0 1.2 0 0 0 1.0 0 2.0 0.6 3.2 1.4 0 1.3 1.9 69 340 Appendix E: Appendix to Chapter 4: Methodology Brainerd-Robinson Adjacency Matrix The approach employed in order to explore the presence of ceramic practice communities within the larger Keffer village community requires the application of some form of measurement to the ceramic types used as proxies for consumer choice. In a network approach, this measurement is based on relationships between entities, in this case, house and midden ceramic samples. The similarity seen among the collections represents the strength of these relationships. A commonly employed method for assessing archaeological ceramic samples is the BrainerdRobinson co-efficient of similarity or dissimilarity matrix, signed specifically for this purpose (Brainerd 1951; Robinson 1951). This approach measures the similarity in the proportions of the values of the selected variable. In this case, the values selected are relative frequencies of ceramic types within the variable of the MacNeish typology. As suggested by McNutt (2005), the results of the similarity measures may be misleading, and it is therefore important to understand the data being used. The same level of similarity between two entities can be derived in quite different ways. The presence of one dominant value in both samples can result in a high similarity co-efficient. On the other hand, equal, or very close, proportions of all values, whether they are high or low, within the two cases may result in the same high BR index. Sample Size and Geographical Proximity In this study, similarity in the realm of shared ceramic types is used as a proxy for social ties. It must be realized, however, that such similarity reflects interaction in only in regard to ceramics. It is not indicative of similarity in terms of any other index. It is important to be cognisant of dominant values which may affect similarity results. In this case, the Huron Incised type, or undecorated neck Huron category, can, in rare cases, comprise up to 100 percent of the sample. Dominant values can cause skewing in small samples, where diversity of types is less likely to occur. In order to control for the possibility of this skewing in the Keffer ceramics, the data have been divided into three matrices. Two of these, with sample sizes ranging below 30 vessels, are considered less robust. The third, with a minimum sample size of 30 vessels, is used to form the base graphs on which the analysis is grounded. 341 This value matches the minimum sample size chosen by the Southwest Social Networks Project team of Barbara Mills and associates, leaders in archeological ceramic network analysis (Habiba et al. 2018). There are 42 categories of ceramic wares among the more than 4 million vessels in the Southwest Social Networks database. The actual similarity measured, as noted by Habiba et al. (2018:64) and Mills (2017), is the consumption and disposal of ceramics. In the absence of technical information pertaining to manufacturing processes, the production of vessels cannot form part of the study of practice communities. BR matrices for large archaeological datasets are easily created with the use of Matt Peeples’ BR subprogram in the open source statistical program R (Peeples 2011). The program is capable of transforming either frequency or percentage data input in csv (comma-separated variables) format. The relative weighting of each loci collection is an inherent property of frequency data, in contrast to percentage data, where each sample is given equal weight and where sample size may skew results in the output matrix files. To counteract skewing effect, Peeples (2011) has included a Monte Carlo simulation of 1,000 iterations with frequency input. This results in the production of a probability matrix output file predicting the likelihood (between 0 and 1) that each cell in the output BR matrix represents randomized data. According to Peeples (http://www.mattpeeples.net/br.html; Matthew Peeples, personal communication 2015), the numbers generated are unlikely to be within the acceptable range (P<.05) if the samples compared originate in the same population. Peeples suggests that collections from within the same region, and particularly with the same site, may be too similar to fit the randomized sampling level generally accepted. This does not mean that these samples are unsuitable for comparative analysis; it merely demands that extra vigilance in their interpretation. SNA has advanced greatly in the past few years, and some researchers (Dermarkar et al. 2016; Hart and Engelbrecht 2011; Hart et al. 2016) now employ the robust interpretive powers of SNA statistics, such as External-Internal Index, which compares the strength of internal ties to those of external ties, indicating strong homophily (−1) or strong heterophily (www.analytictech.com/essex/Exercises/Interpreting_UCINET.doc) and centrality measures, to verify visual interpretation of SNA plots in the exploration of social relationships. The extreme proximity, in social, spatial, and temporal terms, of the ceramic samples from within the Keffer settlement suggests that results from many network statistics are unlikely to be robust at this level. Therefore only frequency-based matrices are used for Keffer intra-site 342 network analysis. At the inter-site, regional, and pan-regional levels, in contrast, frequency data is often not available and percentage data is relied upon for the BR matrix production. Social Network Analysis Interpretation The intended analysis was proposed to take two fundamental forms, visual and statistical investigation of ceramic type through social network analysis (SNA) software programs and basic statistical analysis. Although ceramic types form the core of this analysis, I also examined vessel attribute variables deemed to be of chronological or social significance, both singularly and in combination, to unmask and interpret patterns. The raw data were restructured into several assemblages, each of which was constructed in order to uncover different kinds of relationships within the database. Initially, the finalized data was sorted in Excel, according to loci, into house and midden contexts. Specific palisade designations were not available, though in most cases it was possible to identify middens which post-dated palisade removal or collapse. I selected Ucinet 6 for social network analysis software for this project for a variety of reasons. As Ucinet 6 is a common platform, detailed software directions and information are available online. This program is accessible without in-depth computing or statistical background knowledge and is therefore relatively easy to use. Ucinet Netdraw (Borgatti 2002) contains many features of use which produce relatively simple, clear, yet well-nuanced portrayals of similarity data. It is well suited for network illustration, such as node attribute and tie-strength characteristics, as discussed below. SNA programs require the storage of input data in a square matrix form. Ucinet data files are specifically designed for Netdraw. Two forms of data files are necessary for fluid SNA graph interpretation: Brainerd-Robinson coefficients of similarity and node attribute data. The original matrix portrays the inter-node similarity levels. A second file encodes attributes of the nodes, or physical origins of the samples. The node attribute files created for this study describe individual characteristics of the nodes. Graph Node Attributes Sample Size at Keffer I produced network graphs of the Keffer ceramic data for each of the three increasingly exclusive data sets, with sample sizes of 6–19 vessels, greater than 20 vessels, and greater than 30 vessels. 343 Although statistical analysis is generally believed to be more robust with sample sizes above 30 units, smaller data samples are used here to produce the most inclusive network graphs possible. The weak ties seen in these plots provide hints for directions of inquiry for internal village relationships involving these poorly represented loci. Several longhouses yielded very limited material culture remains due to a combination of the low number of longhouse pit features and the extensive mechanical topsoil stripping of ploughed topsoil in the northern portion on the site. With 37 distinct type variables present in the Keffer ceramic collection, and therefore in the database, entities with small samples will not reflect the diversity present in the overall collection. This sampling bias must be accounted for in analysis (Habiba et al. 2018:66). In order to visualize this discrepancy for immediate recognition during graph pattern exploration, the relative size of each ceramic sample is included in the node attributes. This first node attribute, which portrays the number of vessels, or sample size, is an ordinal variable. These sample size categories are not separated by uniform divisions; rather, the quantitative differences between the categories are uneven. At the lower end of the spectrum, the values are divided into several smaller-range categories, while at the higher end of the spectrum, values represent greater size ranges. The relative size of the small samples gives a more precise picture of the variability possible within each sample. Larger samples, defined as between 200 and 499 vessels or more than 500 vessels, hold greater likelihood of variability. I have divided sample sizes into intuitively colour-coded groups based on the number of vessels within them, with the colours reflecting the amount of caution needed. Darker colours portray more robust samples. These assignments are illustrated in the figure legends. Node Type: House or Midden The second node attribute variable, as seen in node shape, is nominal and describes the archaeological context of the sample origin, either longhouse or midden. House nodes are portrayed by circles and middens by squares. The association of the loci in relation to the village palisades is integral to the temporal positioning of site features in the site development sequence. This affiliation is portrayed in the third variable, also nominal, of settlement cluster location. Cluster assignment is also displayed by colour. Where cluster affiliation is used in network analysis, it replaces the variable of sample size as seen in the figure legends. 344 Node Variables for Ontario and Pan-Iroquoian Network Ontario and pan-Iroquoian network nodes have two integrated attributes. The first, portrayed by node colour, is the geographic region of the site. The second, portrayed by node shape, is the time period assigned to the site. Methods of Visualization In order to establish a background network for the ceramic social relations across the village, the BR matrix includes all ceramics with identifiable occupation contexts, using the Ucinet Netdraw program. Network graphs are produced at several BR levels of similarity for each data matrix. BR cut-off levels are arbitrarily defined by the researcher, with the objective of the analysis in mind (Habiba et al. 2018). The lowest similarity level displayed for each of these graph series is set at the BR measure where all nodes of the matrix are connected in one main component. From this foundation, network graphs are produced at BR value increments of 10 points, or 5 percent of similarity, until in the highest BR levels no ties remain between nodes. At any point where there appears to be a significant increase in the number of broken ties, indicating numerous changes in similarity among nodes, plots are generated at a midpoint BR level of 5, as in BR 145. Netdraw plots are automatically produced in a “spring embedded” layout. This layout (also termed an arrangement or an organization) produces a graph where strongly linked (i.e., high BR value) nodes are attracted closely together, but where the program lightly repulses nodes from each other to produce an easily legible plot. This form of layout, however, does not visually reproduce a linear relationship in the distance between two nodes. Instead, it aims to cover as much of the graph surface area as possible while reflecting relative similarities. As mentioned, these plots render nodes and their relative positions readily identifiable. More realistic representations of the relationships between nodes are produced with the complimentary Gower layout. With its overlapping nodes and closely spaced ties, this layout can be difficult to read but more instructive, because more realistic, proportional distances are illustrated. 345 A combination of the two layouts is advantageous. Ease of the graph reading, that is, decipherment, is also aided by software options that allow for the portrayal of differing levels of tie strengths and several options for altering the appearance of nodes based on their characteristics. In addition, the visual depiction of various centrality vectors with Netdraw clearly illustrates the influence of each of the nodes in the overall network through node size enhancement, thus facilitating comprehension of the intensity of interaction. With binary data, used to indicate the presence or absence of a tie, Degree Centrality node size is indicative of the relative number of ties a node shares with other nodes, not the strength of those ties. The Degree Centrality option is useful in visually connecting locations with specific ceramic types or attributes. It will be employed to illustrate the co-existence of exotic ceramics in the pursuit of direct ties among occupants within the village. A “valued” graph goes beyond noting the presence or absence of a tie, by also measuring the strength of a tie between two nodes, such as BR value. In such a “valued” graph, the Eigenvector and Degree Centrality options also portray node dominance through node size and colour, with larger nodes indicating greater influence, or pull, on the total network. Network Graph Layouts Visualisation of networks in NetDraw can take an almost infinite number of forms. The software is capable of producing graphs in numerous diverse layouts, combined with a multitude of relationship analysis programs displayed through a variety of tie transformation applications. While many layouts and analysis procedures can be useful for applications in archaeology, it is necessary to select, generally through trial and error, those most suited to the needs of the specific project pursuant to the nature, size, and scale of the research data. In this thesis, the high similarity level of the relationship ties and the nature of the community of practice bonds being investigated may be best explored through simple graph layouts and basic centrality measures, because this high level of similarity does not produce well-defined or useful outcomes in most of the analysis procedures. Some centrality measures, such as Eigenvector and Network Density, are useful and will be incorporated in support of the various graph layouts employed. Basic graph layouts in NetDraw include the default graph theoretic (spring embedded), Gower, non-metric multidimensional scaling (MDS), and ego network configurations, each of which is 346 employed in the investigation of the Keffer site ceramics. Spring embedded loading is characterized by node repulsion, resulting in links of equal length between each node. This layout portrays the existence of a tie between two nodes at the given similarity value, but both the absolute and relative strengths of the ties are obscured in weighted networks. While this default layout is the result of the direct illustration of the matrix data, other layouts are produced from the application of secondary algorithms on this initial default graph. Metric layouts, such as Gower scaling, formally called Torgersen-Gower scaling, on the other hand, produce configurations which are highly representative of Euclidean space. In other words, nodes are located in any space by co-ordinates, one for each dimension of space being represented. The high similarity levels seen in the Keffer ceramics produce Gower layout graphs characterized by dense, and often indecipherable, clusters of nodes. MDS layouts result in more visibly interpretable structures, which incorporate a slight repulsion of nodes, making the graph easier to read and nodes more visible. In this layout, although the precise measure of tie strength is lost, the relative spacing of nodes is preserved. This is not true of other layouts, such as spring embedded loading. MDS algorithms may be applied to various existing graph layouts. The most consistently accurate non-metric MDS graph diagrams are achieved through the application of this layout on existing Gower network diagrams, in contrast to the default layout, in which Euclidean placement is not a priority and which may produce final network graphs that differ from one another (Borgatti et al. 2013). Therefore, in this work, non-metric MDS layout is applied systematically to Gower metric graph layouts to produce comparable results in each of the iterations. Ego layouts, both spring embedded loading and Gower, display only those nodes connected to ego at this level in the network. Extraneous nodes are removed, simplifying relations to the ego node. In the spring embedded layout, distance of these nodes from the ego node is not indicative of the strength of the ties (Figure E-1). However, networks are rendered centring on the ego node, thereby emphasizing outliers in the relationship. In Gower ego networks, those nodes most similar to ego are placed in alignment with the Euclidean distance of their relations (Figure E-2). The less cluttered ego graphs aid in the visual interpretation of individual node relationships at all levels. 347 Figure E-1. Local tradition ceramics Over 29 Vessels sample BR 170 default spring embedded layout. Figure E-2. Local tradition ceramics Over 29 Vessels sample BR 170 ego Midden 54 default spring embedded layout. 348 Figure E-3. Local tradition ceramics Over 29 Vessels sample BR 170 ego Midden 54 Gower layout. The Brainerd-Robinson co-efficient of similarity matrix is well suited to both the metric spatial representation of Gower scaling and the more readable node placement of non-metric MDS layout. The true nature of the relationship between two nodes is most accurately displayed at the highest cut-off levels of similarity in the Gower layout, followed closely in precision by non-metric MDS graphs. In both cases, as similarity levels decline, the more numerous, but lower strength, ties to the additional nodes exert pulls on these original nodes, thereby drawing them away from their most similar counterparts and into more complex, lower-level relationships. These higherstrength relationships must be preserved in the analysis results when lower-level similarity graphs are interpreted, introducing relationships with the newly introduced nodes at these lower levels. In the Keffer case, the relationships visible in the Gower and MDS layout graphs at BR 190 (Figures E-4 and E-5) are seen as most accurately reproducing the strength of the connections between the original nodes. The force of these ties is visually softened with the addition of the new nodes at the lower BR 180 (Figures E-6 and E-7) and 170 levels (Figures E-8 and E-9). 349 Figure E-4. Local tradition ceramics Over 29 Vessels sample BR 190 Gower layout. Figure E-5. Local Tradition ceramics Over 29 Vessels sample BR 190 MDS layout. 350 Figure E-6. Local tradition ceramics Over 29 Vessels sample BR 180 Gower layout. Figure E-7. Local tradition ceramics Over 29 Vessels sample BR 180 MDS layout. 351 Figure E-8. Local tradition ceramics Over 29 Vessels sample BR 170 Gower layout. Figure E-9. Local tradition ceramics Over 29 Vessels sample BR 170 MDS layout. Network graphs of the exclusive Over 29 Vessels matrix at BR 193 (Figures E-10 and E-11) and each of the chosen similarity levels in both Gower and MDS layout illustrate this point. The utility of the two layout algorithms in terms of node visibility at different similarity cut-off levels 352 is also revealed. The enhanced Midden 54 node is observed changing position relative to other nodes as the similarity level decreases. Figure E-10. Local tradition ceramics Over 29 Vessels sample BR 193 Gower layout. Figure E-11. Local Tradition ceramics Over 29 Vessels sample BR 193 MDS layout. At BR 193, Midden 54 (the large, bright yellow node) ties to Midden 60 (BR 193.1) and Midden 77 (BR 195.1). In the Gower layout (Figure E-8), these three nodes rest on top of one another 353 due to their similarity at BR 193. Middens 61 and 80 are linked in the network but do not tie directly with Midden 54 at this level and are therefore not incorporated in this mass. Only Middens 77 and 80 are this closely associated with Midden 54 at the BR 190 benchmark. When the BR is lowered to BR 180, Midden 80 moves away, leaving Midden 77 alone overtop of Midden 54. Midden 60 reconnects with Midden 54 at BR 170, bringing Midden 52 along with it, while Midden 77 separates to rejoin Midden 80. These movements are due to attraction of other, newly introduced nodes at this level, particularly the over–BR 170 ties that the original nodes have with their new partners. The Gower layout portrays the ties in accordance with their Euclidean distance. Two nodes with the same relations to all other nodes will therefore overlay one another. This does not imply that these nodes are 100 percent similar, merely that their relations to other nodes at this BR level are the same. Usually these nodes are closely tied themselves. This layout can be confusing and misleading because some nodes, and their labels, may be totally obscured by others. Therefore, MDS, which places nodes in the graph in relational distance, with a small amount of repulsion between them, allowing the visibility of all nodes, is often preferred. As the level of similarity in these samples is extremely high, Gower layout provides more precise demonstration of proximal relationships. The use of non-metric MDS as a backup facilitates the decipherment of the nodes involved in these compact associations. Statistical Analysis Visual interpretation of network graphs comprises the majority of the analysis. Network statistics are used in some instances to confirm or support properties of the Keffer site ceramics relations. These statistics are produced with the Ucinet program and are almost entirely confined to the evaluation of network densities produced from the various-sized sample databases. The measures employed focus on comparison of valued edges, the similarity measures between each pair of nodes. Statistics resulting from comparison of the presence or absence of edges at certain similarity levels, such as the E-I Index, are not useful in this case. The high level of similarity of the values generally shared among the Keffer site nodes results in the simultaneous appearance or disappearance of edges, or ties, as similarity threshold levels decrease or increase, respectively. Resulting statistics show little differentiation among the village nodes or within highly connected site clusters. The use of social network analysis as an exploratory tool used to 354 generate hypotheses in this study supports the prioritization of visual examination of network components over statistical analysis, which would be more relevant for hypothesis testing. 355 Appendix F: Appendix to Chapter 4: Network Analysis and Keffer’s Place in the Iroquoian World Table F-1 lists the pan-Iroquoian sites employed in the social network analysis. Table F-1. Name, Date, Location or Historic Territory, and Sources for the Sites Used in the Social Network Analysis. Site Name A. Brown Time Period (A.D.) 1400–1450 Location/Historic Territory Simcoe County Source Bursey ca. 1990 Adams 1550–1600 Seneca Niemczyski 1984 Alexandra 1350–1400 Scarborough–Highland Creek ASI 2008a Alonzo 1600–1650 Simcoe County Bursey ca. 1990 Ames 1400–1450 Trent valley Bursey ca. 1990 Angoutenc 1500–1550 Simcoe County Bursey ca. 1990 Antrex 1350–1400 Credit River ASI 2010a Arbor Ridge 1400–1450 Prince Edward County Adams 2003 Atsista (Yatsihsta) 1450–1500 Lynde Creek–Oshawa ASI 2014c Atwell 1500–1550 Onondaga Engelbrecht 2004:141 Auger 1600–1650 Simcoe County Bursey ca. 1990 Aurora 1500–1550 Rouge River–Duffins Creek Birch et al. 2017 Bach 1550–1600 Oneida Engelbrecht 2004:142 Baker 1400–1450 Don valley ASI 2006a Ball 1550–1600 Simcoe County Knight and Snyder 1981 Bark 1400–1450 Trent valley Sutton 1990 Barker 1550–1600 Mohawk Engelbrecht 1969a Barnes 1500–1550 Onondaga Engelbrecht 2004:142 Barrie 1350–1400 Simcoe County Wright 1966 Bathurst 1350–1400 Humber valley ASI 2014b Baumann 1400–1450 Simcoe County * Beckstead 1400–1450 St. Lawrence valley Pendergast 1966 Belcher 1500–1550 Seneca Niemczyski 1984 Bernaults 1600–1650 Simcoe County Bursey ca. 1990 Best 1400–1450 Rouge River–Duffins Creek Birch et al. 2017 Beswetherick 1350–1400 Simcoe County Bursey ca. 1990 Bidmead 1600–1650 Simcoe County Bursey ca. 1990 Black Creek 1400–1450 Humber valley Wright 1966 Bosomworth 1400–1450 Simcoe County Wright 1966 Boyle Atkinson 1450–1500 Don valley MPP 1987 Bradt 1550–1600 Milton area Bursey ca. 1990 Buffum 1500–1550 Erie Wright 1966 356 Burkeholder 2 1400–1450 Rouge River–Duffins Creek ASI 2005a Buyea 1450–1500 Oneida Engelbrecht 2004:142 Cameron 1600–1650 Oneida Niemczyski 1984 Campbell 1350–1400 Milton area Smith 1997 Carlos 1400–1450 Northern New York Engelbrecht 2004: 137 Carmichael 1450–1500 Collingwood area Garrad et al. 2014:276-277 Carson 1450–1500 Simcoe County Bursey ca. 1990 Cayadutta 1500–1550 Mohawk Lenig 1965:100 Cedar Point 1600–1650 Simcoe County Latta 1976:392 Cemetery 1450–1500 Onondaga Engelbrecht 2004:141 Chance 1400–1450 Mohawk Lenig 1965:100 Charlebois 1550–1600 Simcoe County Latta 1976 Chase 1550–1600 Onondaga Engelbrecht 2004:141-142 Chaumont 1450–1500 Northern New York Engelbrecht 2004:137 Chew 1600–1650 Simcoe County Bursey ca. 1990 Chypchar 1350–1400 Milton area Smith 1997 Cleary 1400–1450 Simcoe County Warrick and Molnar 1986 Clifton Springs 1400–1450 Cayuga Niemczyski 1984 Connor-Rolling 1600–1650 Collingwood area Garrad et al. 2014:276-277 Cooper 1550–1600 Simcoe County Bursey ca. 1990 Copeland 1450–1500 Simcoe County Bursey ca. 1990 Cornish 1600–1650 Seneca Engelbrecht 1970a Coulter 1500–1550 Trent valley † Cranston 1400–1450 Simcoe County Bursey ca. 1990 Crawford Lake 1400–1450 Milton area Smith 1997 Cromwell 1500-1550 Mohawk Lenig 1965:100 Crystal Rock 1350–1400 St. Lawrence valley Pendergast 1962 CSR 1450–1500 Simcoe County Bush 1976 Damiani 1450–1500 Humber valley ASI 2014a: 2015 Dansville 1400–1450 Seneca Lenig 1965:88 Davey 1350–1400 Simcoe County Sutton 1999 Deowingo 1350–1400 Mohawk Lenig 1965:100 Deshambault 1400–1450 Simcoe County Latta 1976 Diable 1550–1600 Oneida Engelbrecht 2004:143 Doncaster 1350–1400 Don valley Wright 1966 Downsview 1400–1450 Humber valley Wright 1966 Draper 1450–1500 Rouge River–Duffins Creek Pihl 1984 Drumholm 1350–1400 Milton area Smith 1997 Dunn 1400–1450 Simcoe County Bursey ca. 1990 Dunsmore 1450–1500 Simcoe County Robertson and Williamson 2003 Durfee 1450–1500 Northern New York Engelbrecht 2004:137 Durham 1450–1500 Northern New York Engelbrecht 2004:137-138 357 Dutch Hollow 1600–1650 Seneca Lenig 1965:88 Dykstra 1350–1400 Simcoe County ASI 2006b Eaton 1550–1600 Erie Wright 1966 Edwards 1350–1400 Neutral Smith 1997 Ellery 1600–1650 Simcoe County Bursey ca. 1990 EllesmereMorrison Ellis 1450–1500 Simcoe County Bursey ca. 1990 1600–1650 Erie Engelbrecht 1977a Elwood 1550–1600 Mohawk Snow 1995 Emmerson Springs Englebert 1500–1550 Credit valley Hawkins 2004 1550–1600 Susquehannah Engelbrecht 1974a Factory Hollow 1600–1650 Seneca Lenig 1965:88 Farlain Lake 1450–1500 Simcoe County Latta 1976:337 Farrell 1350–1400 Seneca Niemczyski 1984 Finch 1400–1450 Southwestern Ontario Pihl and Thomas 1997 Footer 1350–1400 Seneca Niemczyski 1984 Forget 1600–1650 Simcoe County Vastistas 2011 Fort Drum 1400–1450 Northern New York Engelbrecht 2004:138 Fournier 1450–1500 Simcoe County Bursey ca. 1990 Frank 1450–1500 Northern New York Engelbrecht 2004:138 Freeman 1450–1500 Northern New York Engelbrecht 2004:138 Ganada 1500–1550 Mohawk Snow 1995 Garoga 1450–1500 Mohawk Lenig 1965:100 Genoa Fort 1550–1600 Cayuga Lenig 1965:87 Gervais 1400–1450 Simcoe County Sutton 1999 Getman 1450–1500 Mohawk Lenig 1965:100 Glebe 1600–1650 Collingwood area Garrad et al. 2014:276-277 Goff 1450–1500 Oneida Pratt 1976 Goodeve 1500–1550 Simcoe County Bursey ca. 1990 Goodyear 1550–1600 Erie Wright 1966 Gostick 1400–1450 Rouge River–Duffins Creek ASI 2014d GrahamFerguson Graham-Rogers 1600–1650 Collingwood area Garrad et al. 2014:276-277 1550–1600 Simcoe County Wright 1966 Grandview 1400–1450 Lynde Creek–Oshawa McDonald 2002 Green Lake 1550–1600 Erie Wright 1966 Grey's Creek 1400–1450 St. Lawrence valley Pendergast 1966a Gwynne 1500–1550 Simcoe County Bursey ca. 1990 Hamilton 1600–1650 Milton area Lennox 1981 Hanes 1500–1550 Milton area Bursey ca. 1990 Haney-Cook Lower Haney-Cook Upper 1600–1650 Collingwood area Garrad et al. 2014:276-277 1600–1650 Collingwood area Garrad et al. 2014:276-277 358 Heath 1450–1500 Northern New York Engelbrecht 2004:138 Heron 1400–1450 Simcoe County Bursey ca. 1990 Hidden Springs 1450–1500 Don valley ASI 2010b Hillier 1400–1450 Prince Edward County ‡ Holly 1350–1400 Simcoe County ASI 2009 Hope 1400–1450 Don valley ASI 2011 HOro17 1550–1600 Simcoe County Bursey ca. 1990 Howie 1550–1600 Collingwood area Garrad et al. 2014:276-277 Hubbert 1400–1450 Simcoe County McDonald and Williamson 2001 Hummel 1350–1400 Cayuga Lenig 1965:87 I. Elliot 1400–1450 Bruce County area Bursey ca. 1990 Inverhuron 1450–1500 Bruce County area Wright 1966 Ivey 1350–1400 St. Lawrence valley Lenig 1965:86 Jackes 1400–1450 Don valley Noble 1974 Jarrett-Lahmer 1450–1500 Don valley ASI 2001 Joseph Picard 1450–1500 Lynde Creek–Oshawa ASI 2012a Keffer 1450–1500 Don valley § Kelly-Campbell 1600–1650 Collingwood area Garrad et al. 2014:276-277 Kienuka 1550–1600 Erie Wright 1966 Kirche 1500–1550 Trent valley † Kleis 1600–1650 Erie Engelbrecht 1978 Klinko 1450–1500 Cayuga Niemczyski 1984 Klock 1550–1600 Mohawk Lenig 1965:100 Laidlaw 1400–1450 Simcoe County ARA 2003 Lalonde 1400–1450 Simcoe County Bursey ca. 1990 Lanoraie 1350–1400 Lower St. Lawrence valley Lenig 1965:86 Latimer 1550–1600 Collingwood area Garrad et al. 2014:276-277 Lawson 1450–1500 Southwestern Ontario Wright 1966 Lite 1450–1500 Prince Edward County Pendergast 1972 Logan 1400–1450 Humber valley Bursey ca. 1990 Long Point 1450–1500 Seneca Lenig 1965:88 Macartney 1400–1450 Don valley Pearce 1998 MacMurchy 1550–1600 Collingwood area Garrad et al. 2014:276-277 Mantle 1500–1550 Rouge River–Duffins Creek ASI 2014c Martin 1550–1600 Mohawk Engelbrecht 2004:143-44 Matteson 1450–1500 Northern New York Engelbrecht 2004:139 McAllister 1550–1600 Collingwood area Garrad et al. 2014:276-277 McCrae 1400–1450 Simcoe County Bursey ca. 1990 McEwen 1600–1650 Collingwood area Garrad et al. 2014:276-277 McGaw 1400–1450 Don valley ASI 2007 McKenzieWoodbridge McNair 1500–1550 Humber valley Birch et al. 2017 1400–1450 Don valley ASI 2012 359 McQueen 1550–1600 Collingwood area Garrad et al. 2014:276-277 Melville 1600–1650 Collingwood area Garrad et al. 2014:276-277 Messenger 1350–1400 Milton area Smith 1997 Middleport 1350–1400 Southwestern Ontario Wright 1966 Millroy 1350–1400 Rouge River–Duffins Creek Wright 1966 Milton 1400–1450 Milton area Smith 1997 Molson 1550–1600 Simcoe County Lennox 2000 Morse 1450–1500 Northern New York Engelbrecht 2004:139 Moyer 1350–1400 Waterloo County Smith 1997 MudCreek 1450–1500 Northern New York Engelbrecht 2004:139 Munsville 1600–1650 Oneida Lenig 1965:93 Myers Station 1600–1650 Cayuga Lenig 1965:87 New 1350–1400 Rouge River–Duffins Creek ASI 2008a Newton Hopper 1550–1600 Erie Engelbrecht 1977b Nichols Pond 1450–1500 Oneida Engelbrecht 2004:143 Nodwell 1400–1450 Bruce County area Wright 1966 Nohle 1450–1500 Northern New York Engelbrecht 2004:139 Orion MurphyGoulding Orr Lake 1450–1500 Don valley ASI 1998, 2008a 1600–1650 Simcoe County Wright 1966 Otsungo 1450–1500 Mohawk Lenig 1965:100 Over 1400–1450 Don valley DRPA 1996b PaddisonBellwood Parsons 1500–1550 Collingwood area Garrad et al. 2014:276-277 1450–1500 Humber valley Williamson and Powis 1998 Payne 1450–1500 Prince Edward County Pendergast 1963 Peacock 1600–1650 Collingwood area Garrad et al. 2014:276-277 Pengilly 1400–1450 Credit valley Bursey ca. 1990 Pipeline 1350–1400 Milton area Smith 1997 Plater-Fleming 1600–1650 Collingwood area Garrad et al. 2014:276-277 Plater-Martin 1600–1650 Collingwood area Garrad et al. 2014:276-277 Pompey Center 1600–1650 Onondaga Lenig 1965:88 Potacki 1400–1450 Northern New York Engelbrecht 2004:139 Pound 1400–1450 Southwestern Ontario Wright 1966 Power House 1600–1650 Seneca Engelbrecht 1970b Pretty River 1600–1650 Collingwood area Garrad et al. 2014:276-277 Pugh 1400–1450 Rouge River–Duffins Creek Birch et al. 2017 Putnam 1400–1450 Northern New York Engelbrecht 2004:140 R[aymond] Reid 1400–1450 Milton area Bursey ca. 1990 Rices Woods 1550–1600 Mohawk Snow 1995 Richmond Mills 1500–1550 Seneca Niemczyski 1984 Rife 1350–1400 Milton area Smith 1997 Ripley 1550–1600 Erie Wright 1966 360 Risebrough 1400–1450 Don valley ‖ River 1400–1450 Credit valley Bursey ca. 1990 Robb 1350–1400 Rouge River–Duffins Creek ASI 2008c Robitaille 1600–1650 Simcoe County Latta 1976:427 Rock Bottom 1600–1650 Collingwood area Garrad et al. 2014:276-277 Roebuck 1500–1550 St. Lawrence valley Jamieson 1982; MacNeish 1952 Roof 1550–1600 Simcoe County Warrick and Molnar 1986 Rumney Bay 1450–1500 Trent valley † Salem 1400–1450 St. Lawrence valley Pendergast 1966a Second Lake 1400–1450 Simcoe County Bursey ca. 1990 Second Woods 1400–1450 Mohawk Snow 1995 Seed Barker 1500–1550 Humber valley Birch et al. 2017 Serena 1350–1400 Milton area ASI 2004 Sewell 1350–1400 Scarborough–Highland Creek ASI 2008a Shelby 1550–1600 Erie Wright 1966 Sidey-Mackay 1550–1600 Simcoe County Bursey ca. 1990 Silverheels 1600–1650 Erie Engelbrecht 1974b Simmons 1550–1600 Erie Engelbrecht 1969b Slack Caswell 1350–1400 Milton area Smith 1997 Smith 1450–1500 Mohawk Engelbrecht 2004:144 Smokes Creek 1600–1650 Erie Engelbrecht 1993 Snodden 1350–1400 Trent valley ¶ Sopher 1550–1600 Simcoe County Noble 1968 Southwold 1400–1450 Southwestern Ontario Wright 1966 Spang 1450–1500 Rouge River–Duffins Creek Birch et al. 2017 St Lawrence 1400–1450 Simcoe County Engelbrecht 2004:140 Starr 1450–1500 St. Lawrence valley Bursey ca. 1990 Steward 1450–1500 Northern New York Jamieson 1982 Swarthout 1450–1500 Northern New York Engelbrecht 2004:140 Talcott Falls 1450–1500 Northern New York Engelbrecht 2004:140-141 Temperance 1500–1550 Onondaga Engelbrecht 2004:142 Thomson 1350–1400 Scarborough–Highland Creek ASI 2008a Thurston 1600–1650 Oneida Engelbrecht 2004:143 Train 1400–1450 Simcoe County Bursey ca. 1990 Unick 1350–1400 Milton area Smith 1997 Uren 1350–1400 Bruce County area Wright 1966 VanEden 1350–1400 Milton area Smith 1997 Vints 1600–1650 Simcoe County Bursey ca. 1990 Wagners Hollow Walkington 2 1550–1600 Mohawk Engelbrecht 2004:144 1400–1450 Don valley ASI 2010d Wallace 1450–1500 Credit valley Crawford 2003 Warminster 1600–1650 Simcoe County Wright 1966 361 Warren 1600–1650 Seneca Lenig 1965:88 Watford 1400–1450 Don valley Pearce 1997 Waupoos 1450–1500 Prince Edward County Pendergast 1964 Webb 1400–1450 Simcoe County Bursey ca. 1990 Wellington 1350–1400 Simcoe County ASI 2005b White 1500–1550 Collingwood area Garrad et al. 2014:276-277 Whitford 1450–1500 Northern New York Engelbrecht 2004:141 Wiacek 1350–1400 Simcoe County Robertson et al. 1995 Wilson 1350–1400 Trent valley Sutton 1990 Woodley 1450–1500 Cayuga Lenig 1965:87 Wylie-Bowman 1400–1450 Collingwood area Garrad et al. 2014:276-277 Wylie-Coyle 1550–1600 Collingwood area Garrad et al. 2014:276-277 Yong-McQueen 1550–1600 Collingwood area Garrad et al. 2014:276-277 † Based on the author's examination of collections held at the Department of Anthropology, Wilfrid Laurier University, Waterloo. Based on the author's examination of drawings of rimsherds by Peter Ramsden. ‡ Based on the author's examination of collections held at the Canadian Museum of History. § Based on the author's examination of collections held at the Museum of Ontario Archaeology, on loan to the Department of Anthropology, University of Toronto at Mississauga. Based on the author's examination of collections held at the Department of Anthropology, University of Toronto, and the Royal Ontario Museum. Based on the author's examination of collections held at York North Archaeological Services Inc., Peterborough, Ontario. * ‖ ¶ 362 Social Network Analysis of Pan-Iroquoian Ceramic Practice The ceramic types included in this analysis were chosen on the basis of common usage. Types present on only a small number of sites in small proportions were removed to prevent overweighting of minor types. In addition, several pairs of types have been combined under compound names, such as Pound-Black Necked, since it has become common in Iroquoian research to combine types with similar characteristics (see Lennox and Kenyon 1984; Pratt 1980) Data available in many Iroquoian ceramic reports include conflated type names following the work of Pratt 1980. These combined types are used here. At the low similarity measure of BR 100, or 50 percent similarity, (Figure F-1), the panIroquoian network, including sites of all periods, t1–t6, separates into three basic components. The largest of these includes all of the Ontario-based sites. A degree of temporal and spatial clustering is apparent across this component, portraying patterned interaction across time and space. New York sites form two well-separated groups, with t2–t6 Cayuga and Seneca sites roughly laid out by time and location. The larger New York state component displays clear separation of site collections by location. In this network, north shore St. Lawrence Iroquoian sites are most tightly connected with the Northern New York cluster. A pendant off this group ties the later t4 St. Lawrence Roebuck and t4 Northern New York Swarthout and Durfee sites to t5 and t6 Oneida sites, perhaps suggesting community movement in this direction over time. It is only through the t3–t6 Oneida and Onondaga sites that the large Mohawk-area cluster connects with the St. Lawrence, Northern New York, and later t5 and t6 Oneida nodes at the BR 100 level. 363 Figure F-1. Pan-Iroquoian network BR 100 spring layout. As the similarity level rises, New York groups begin to disentangle and disperse, with several smaller components being formed. By BR 140, or 70 percent similarity (Figure F-2), these smaller segments are visibly divided by both time and geography. The Ontario component remains a connected group, although temporal and geographical patterning becomes more apparent here as well. Figure F-2. Pan-Iroquoian network BR 140 spring layout. 364 It is only at the low similarity value of BR 95, or 47.5 percent (Figure F-3), that ties appear between the two geographic regions. As this level, connections are only present through a chain of nodes, across multiple time periods. Sites of t4–t6 Eastern Iroquois Oneida, Onondaga, Mohawk; Northern New York; and the St. Lawrence are tied to predominantly t2 and t3 Neutral Ontario sites, through t5 and t6 Erie sites and a t3–t6 Cayuga, Seneca, and Erie bridging segment. This extremely weak connection, tied through nodes spanning two centuries, merely indicates that certain traditional ceramic practices of the western Iroquois, represented through the lens of types, were shared with nearby neighbours at one point in time. Figure F-3. Pan-Iroquoian network BR 95 spring layout. Contemporary New York and Ontario sites first connect at the BR 93 level (Figure F-4). Northern New York sites from south of the St. Lawrence River tie with nearby north shore of Lake Ontario Prince Edward County sites. At this level, early t1 and t2 western Iroquois Seneca and Cayuga sites are connected with t1 Ontario sites and do not link to later western Iroquois sites. In fact, t1 Seneca sites are more closely tied with some t1 Neutral-area sites than are other Ontario t1 sites. 365 Figure F-4. Pan-Iroquoian network BR 93 spring layout. Figure F-5. Pan-Iroquoian network BR 81 spring layout. In this time period, the pan-Iroquoian network is quite dispersed at BR 94, or 47.5 percent similarity, with the Ontario component well separated from those of the Mohawk and Seneca clusters to the bottom (Figure F-6). St. Lawrence sites are tied to the larger Northern New York component. 366 Figure F-6. Pan-Iroquoian network t2–t3 BR 94 spring layout. Direct ties between New York and Ontario sites in the latter half of the fifteenth century, contemporary with the occupation of the Keffer village, do not appear until the similarity level is lowered further, to BR 93 (Figure F-7). Here ties appear between the Northern New York Matteson and Mud Creek sites and the nearby ancestral Wendat Prince Edward County Waupoos site. Waupoos is not itself a highly integrated node in the Ontario component, falling out at BR 127. Its own weak position adds to the fragile nature of this bridging tie with the Ontario cluster. 367 Figure F-7. Pan-Iroquoian network t2–t3 BR 93 spring layout. The t2–t3 network graph at BR 80 (Figure F-8) displays the similarity level, 40 percent, at which all the main components tie together. The two linked main components, the first comprised of St. Lawrence t2 and t3 sites and the second of the tight cluster of the Northern New York sites, are lightly tied to the t2 and t3 ancestral Eastern Iroquois Oneida, Onondaga, and Mohawk nodes, from with a pendant of two western Cayuga-area sites hangs. Two Seneca sites form a small component isolated from both larger ones. In this time period, the Northern New York–St. Lawrence cluster is more highly tied with Ontario sites than with New York sites. The second, denser cluster of ancestral Ontario Iroquoian Neutral and Wendat sites is characterized by a slight separation of t2 and t3 nodes, with the t2 nodes displaying a measure of spatial segregation as well. This suggests a pattern of widespread interaction within the southern Ontario region, with a small bias towards intra-area concentration and a smooth transition to later t3 ceramic practices. Figure F-8. Pan-Iroquoian network t2–t3 BR 80 spring layout. 368 Figure F-9. Pan-Iroquoian network t3–t4 BR 80 spring layout. A similar pattern is seen in the t3–t4 network graphs at BR 90 (Figure F-10). Again Eastern Iroquois and Northern New York Iroquois form two clusters linked with slightly looser ties. In all components there is less separation of nodes of the two time periods. Seneca sites remain isolated once again, and the southern Ontario component is composed of closely tied nodes at the BR 80 level. The role of geography in the ties of the Ontario sites is less evident in this graph. Figure F-10. Pan-Iroquoian network t3–t4 BR 90 spring layout. 369 With both datasets it is the t3 Northern New York Mud Creek and Matteson sites and the Prince Edward County Waupoos sites which bridge the divide between the groups at BR 93 (Figure F11). No t2 or t4 sites connect the main groups at this level. These weak connections suggest that some level of interaction was occurring in t3 between eastern Ontario ancestral Wendat and some Northern New York t3 peoples. Figure F-11. Pan-Iroquoian network t3–t4 BR 94 spring layout. Appendix G: Appendix to Chapter 5: Social Network Analysis of 140 Ontario Iroquoian Local Tradition Ceramic Collections This appendix describes the steps taken to produce a network analysis of the similarities found among local tradition ceramic collections of the Iroquoian sites in southern Ontario. Local tradition ceramics from 140 sites (Appendix C) spanning the period A.D. 1350–1650, previously known as the Late Ontario Iroquoian period, are examined. Collections with very low proportions of Ontario local tradition pottery, including Salem, Grey’s Creek, and the eastern New York sites, are therefore also not included in this analysis. In these cases, the removal of high percentages of nonlocal tradition ceramics from the data matrix produces a skewed representation of the small proportions of local tradition ceramics remaining. In addition, the interaction responsible for the low proportions of nonlocal ceramics present in most assemblages cannot be reflected in value-based network analysis. The ceramics samples employed in this analysis are derived from excavation, surface survey, and casual/amateur collections across southern Ontario, from the Neutral/London area northward to Simcoe County, near Georgian Bay, and eastward across the north shore of Lake Ontario to Prince Edward County (PEC). Although the number of collections is considerable, ceramic data for many additional sites were either not suitable or not available. Time restrictions limited the number of collections which could be analyzed to those deemed most chronologically and geographically pertinent to this study. As such, results of this research are not considered definitive, merely suggestive. For this analysis, sites are categorized temporally into six 50 year periods following the work of Engelbrecht (2003) and Hart and Engelbrecht (2012).  t1 1350–1400  t2 1400–1450  t3 1450–1500  t4 1500–1550  t5 1550–1600 370 371  t6 1600–1650 The framework is composed of arbitrary divisions, which do not reflect actual historical boundaries. The dates assigned to these sites by researchers often do not neatly fit into these categories, as some straddle two periods. Pigeonholed assignments within these categories merely give a frame of reference for relative temporal positioning and do not reflect steps in ceramic or culture change. Network ties may be more indicative of temporal similarity than these categories suggest. Sites of all six time periods are included in the analysis to show historical and regional relationships. Thus, the resulting graphs are not strongly predetermined based on, or delineated by, time periods. Collections were assigned to groups based on the physical presence of the site within, or associated with, specific drainage catchments or geographical regions. These include, roughly from west to east, the Neutral area, with sites to the south and west of Lake Ontario; the Credit River drainage; the Humber River drainage; the Don River drainage; the Scarborough Creek and Highland Creek drainage; the Rouge River and Duffins Creek drainage; the Lynde Creek drainage and the Oshawa area; the Trent River drainage; and, at the eastern limit, Prince Edward County. The final area is that of Simcoe County, to the north of Lake Ontario. 372 Figure G-5. Local tradition ceramics of Ontario Iroquoian sites BR 99 Gower layout. At BR 99 (Figures G-5 to G-7), all 142 site nodes are connected into the Ontario Iroquoian ceramic practice network. The Simcoe Angoutenc collection drops out of the networks at BR 100, or 50 percent similarity among nodes. However, aside from this and a few other peripheral nodes, a picture of several highly clustered, but heavily linked, largely temporally defined groups, appears. Even at this level, the chronological separation of the early sites (t1) from the latest Simcoe-area sites (t6) is immediately obvious. The division of these t1 sites into two large and distinct clusters suggests the presence of two unique but linked communities of practice in this early period. These divisions do not appear to be along geographically defined parameters, and nodes of various geographically defined groups, including those of the Neutral area, are found within both clusters. Interestingly, when looking at Neutral-area smoking pipes of this period, Smith (1990) also noted two separate practice communities, not pre-determined by geographic location. Several of the early t1 sites form a bridge to later t2 sites located in the more easterly watersheds of the greater Toronto area, from the Don valley to Prince Edward County, while more westerly t2 and t3 sites of the Neutral, Humber, and Simcoe areas, along with the later, eastern Payne site, form a separate, more direct, path to later t5 and t6 sites, the majority of which are located in Simcoe County. The t2 and t3 sites of the Don, Rouge–Duffins, Trent, and Prince Edward County areas are highly topologically dispersed, with some forming 373 small clusters spreading along both axes across the central portion of the graph. The t2 and t3 Simcoe sites are scattered within these clusters, suggesting loosely shared ceramic practices among most eastern groups during this 100 year period. Humber valley and Neutral sites, with the exception of the Black Creek site, remain removed from this interaction but appear interrelated in a separate graph area. Figure G-6. Local tradition ceramics of Ontario Iroquoian sites BR 99 MDS layout. 374 Figure G-7. Local tradition ceramics of Ontario Iroquoian sites BR 99 spring layout. MDS and spring layout graphs at this level, BR 99, show Trent valley sites to be moderately peripheral, while Don valley sites are scattered, but central to the t2–t3 network. Figure G-8. Local tradition ceramics of Ontario Iroquoian sites BR 120 spring layout. As the similarity level rises to the 60 percent mark, BR 120 (Figure G-8), a number of Londonarea Neutral and central Simcoe County Huron sites become decisively more peripheral from the central core, with nodes of the two areas separating in opposing directions. This may be an 375 indication that ceramic practices in these two areas are following unique paths during this vibrant 100 year period around the turn of the fifteenth century, while dynamic changes were occurring in Iroquoian culture. Figure G-9. Local tradition ceramics of Ontario Iroquoian sites BR 130 Gower layout. 376 Figure G-10. Local tradition ceramics of Ontario Iroquoian sites BR 130 MDS layout. By the BR 130 level (Figures G-9 and G-10), several Neutral sites are connected only by a pendant, and at BR 140 (Figure G-11), they form a small component separated from the main network. The t2 and t3 Simcoe sites form a pendant off the earlier Trent valley Bark site node at BR 140 (Figure G-11), but they no longer tie to the main component at BR 150 (Figure G-12). Separation at this level indicates that a minimum of 25 percent of the ceramic practices of these Neutral and Simcoe sites are unique from those of the main component sites. Meanwhile, at the lower, BR 130, measure, sites of the Don, Rouge–Duffins, and Trent areas, along with the t2 Humber Black Creek site, remain interwoven and form the networks’ central core as Simcoe County sites are removed to the periphery. The presence of t2 sites in the core up to this point argues for a 60 percent base level of similarity in ceramic practice among most north shore and Simcoe County groups. 377 Figure G-11. Local tradition ceramics of Ontario Iroquoian sites BR 140 main component MDS layout. Non-Neutral-area t1 sites disentangle from, but remain connected to, the relatively homogenous t1 Neutral communities at BR 140, and the schism between the earlier and later ceramic assemblages becomes evident as earlier types rapidly decline in popularity around the turn of the fifteenth century. 378 Figure G-12. Local tradition ceramics of Ontario Iroquoian sites BR 150 main component spring layout. At BR 150 (Figure G-12), the Don valley sites, along with the later Trent valley Coulter and Kirche sites, bridge the gap between the t2 and t3 Rouge, Humber, and St. Lawrence cluster and the Simcoe cluster, in which the later t4 and t5 Humber, Rouge–Duffins, and St. Lawrence nodes reside. Figure G-13. Local tradition ceramics of Ontario Iroquoian sites without the Kirche and Coulter sites BR 150 spring layout. 379 Removal of the later Trent nodes, Coulter and Kirche, results in relations where t3 Don and Rouge–Duffins sites link the eastern Toronto communities with those of the Humber t2–t5 Simcoe t2–t6 component, to the right, and its related but peripheral late Toronto sites (G-13). The addition of ceramic data from other Trent valley sites, such as Jamieson, Benson, and TrentFoster, may provide an alternative conduit for connections among Trent valley sites and those of Simcoe County, and in so doing may explain the intermediate position of the t4–t5 Coulter and Kirche sites within the BR 150 network. However, typed data for these sites was not available for this analysis. Figure G-14. Local tradition ceramics of Ontario Iroquoian sites BR 160 spring layout. The Humber Damiani and the Rouge–Duffins Aurora and Mantle sites have more in common with the late t4–t6 Simcoe County sites at 80 percent, BR 160 (Figure G-14), than they do with earlier or contemporary Toronto-area sites. At this level, these more southerly sites merge with this large, homogenous cluster and separate entirely from clusters of earlier sites. Sites of the greater Toronto area and those to the east sustain a common ceramic praxis as levels rise to BR 160 (Figure G-14), but here division among ceramics of these sites begins. This breach is not, however, mainly along drainage basin, geographic, or temporal lines, unlike that 380 seen in the earlier Neutral and later Simcoe components. Ties linking Jarrett-Lahmer, Logan, Risebrough, and Keffer through the Damiani (Humber) and Fournier and Gwynne (Simcoe) sites to the main Simcoe component are broken, and the eastern Toronto-area cluster fragments into four smaller components. Only that component composed of Risebrough, Jarrett-Lahmer, Logan, and Keffer appears to have separated along geographic lines. The two early t2 groups may be chronologically constituted, but this is uncertain, as the other four components also include several t2 nodes. In the larger cluster, Don and Rouge–Duffins sites link earlier Trent and PEC sites to later Trent sites. No Neutral-area nodes interface with other nodes outside their region at this level. 381 Appendix H: Appendix to Chapter 7: Factors Affecting Social Network Analysis Results for Keffer Site Local Tradition Ceramics Effect of Sample Size on Network Similarity Ties In Appendix E, the effect of graph layout on the interpretation of the Keffer dataset was intensively examined. Graph layout was seen to have a substantial effect on visual interpretation, and it was concluded that network graphs portrayed in Gower metric and non-metric multidimensional scaling (MDS) best display the true strength of relationships both between individual houses and middens and throughout the greater village. Consequently, these two graph algorithms are adopted here in the examination of local tradition ceramic relationships and communities of practice across the Keffer village. To mitigate against the problems associated with the great range in size of the loci samples, from 6 to 966, and the extremely small size of some of these samples, matrices of three differing sample size parameters are used to examine the effect of sample size on network results. The first of these includes all samples containing between 2 and 19 vessels (labelled Over 1 Vessel); the second includes all samples containing between 20 and 29 vessels (labelled Over 19 Vessels); and the third includes all samples containing 29 or more vessels (labelled Over 29 Vessels). Statistically, sample sizes larger than 30 are considered more robust, and results produced from the Over 29 matrix are therefore considered the most reliable. However, several important loci on this site are characterized by samples of fewer than 29 vessels. While those features with 20 to 29 vessels (Over 19 Vessels) may provide statistically weaker results, their inclusion contributes essential evidence regarding the nature of their relationships. There can be little confidence in results arising from network analysis of samples under 20 vessels (Over 1 Vessel), but they offer the only hints regarding the relationships of the archaeological features containing these vessels with other features within the greater ceramic community of the village. Three loci, Houses 14, 16, and 18, are characterized by only one vessel each, and these ceramics have not been included in the analysis. Three network graph runs were completed with each of the three matrices, for a total of nine runs. In the first step, the most exclusive (Over 29 Vessels) database was analyzed at increasing similarity levels: BR 170, 180, and 190. The Over 19 Vessels and Over 1 Vessel matrices were 382 then run following the same procedure. When the similarity level is increased with each run, the number of nodes included in the graph decreases and the pull exerted on the remaining nodes is also reduced. Through visual comparison of these ties, the effect of smaller samples on the overall structure of the graphs can be evaluated. Network statistics, graph density, and node degree are employed to confirm these observations. Network analysis is applied to each of these matrices in order to test the effect of small samples on the greater network. As each node, regardless of sample size, is given equal weight within the database, it has the potential to exert the same strength of node attraction as each of the others. Nodes have been colour coded to visually display sample size differences among the relationships portrayed, as described in Chapter 6. Houses and Middens are shown as circles and squares, respectively. A comparison of relative node position within the network diagrams resulting from the application of the three matrices will show the effects of small sample sizes on the graph structure. These networks will be depicted in Gower layout, where the strength of the ties is most accurately reflected, and also in the somewhat less accurate but more visually accessible MDS layout. Both the Gower and the MDS graphs of local tradition ceramics of Over 29 Vessels sample sizes at the BR 170 level visually illustrate several tight node clusters in an otherwise relatively dispersed network (Figure H-1). The greater core area is split in four directions and has no central focus. Several outliers surround the more central groups, that is, Groups 1 to 5. Of these five groups, four tight, internally linked clusters are apparent at first glance: Group 1a (Middens 52, 54, and 60), Group 1b (Middens 65, 72, 73, and 74), Group 5 (Middens 77 and 80), and Group 6 (Middens 51 and 55). Nodes of these clusters are positioned one on top of one another at this BR level, and these groups sit in close proximity within the graph. Three more loosely tied groups, Group 2 (Middens 59, 61, 66, and 71), Group 3 (Houses 13 and 20), and Group 4 (Middens 56 and 62), are placed somewhat distant from the “core” area of the graph. The most isolated nodes are typically of small sample size, between 31 and 55 sherds. Interestingly, Midden 57, which, with 522 vessels, is one of Keffer’s largest collections, is quite distant from, yet has several ties to, the core area. It is these outliers, in their connections with the tighter core groups, which exert an attractive force on the central groups of nodes, pulling them away from each other. 383 Figure H-1. Local tradition ceramic groups Over 29 Vessels BR 170 Gower layout. 384 Figure H-2. Local tradition ceramic groups Over 19 Vessels BR 170 Gower layout. The BR 170 network graph created with the inclusion the Over 19 Vessels nodes portrays a slightly different picture (Figure H-2). The addition of five new nodes with samples sizes ranging from 20 to 25 vessels changes the dynamics of the visible structure. Three of these nodes become network outliers, and they have only a small effect on the structure as they join with House 11 and Midden 66 to form a small peripheral cluster. Houses 4 and 7 pull Midden 66 towards them in the periphery. Newly appearing Middens 69 and 78 pull Midden 51 outwards, separating it from the Midden 55 node. House 15 has little effect on the overall network, although it places a slight attraction on Middens 51, 55, and 68–78. Group 3, Houses 13 and 20, moves from the periphery, away from House 19 and Midden 53, to the centre of the network. 385 Figure H-3. Local tradition ceramic groups Over 1 Vessel BR 170 Gower layout. When additional loci of Over 1 Vessel samples are introduced, seven new nodes emerge, each with between 6 and 18 vessels (Figure H-3). Outliers House 3 and Midden 67 have no visible effect on the network. Houses 2 and 6 appear to join a small group exterior to the core body, along with Houses 4, 7, and 11, but otherwise also have little effect. Midden 63 joins the Houses 13 and 20 cluster. House 10, with only 6 sherds, becomes tightly integrated into Group 2, and along with Midden 63, pulls House 1 out of the cluster to the left and attracts the large Group 1 (Middens 52, 54, 72, and 74) node towards it. At the same time, the appearance of the House 8 node repulses the group outward while attracting Midden 65. This analysis of the three sample size matrices at the inclusive BR 170 similarity level has resulted in the production of three slightly different depictions of node cluster relationships. The effect of smaller sample size nodes on the larger sample size is visible in the movement of nodes such as Middens 77 and 80 from Group 5 into Group 2 when the smaller samples of House 10 and Midden 63 are added. On the other hand, the strength of the ties binding nodes within 386 Groups 1, 2, 3, and 4 is apparent (Table H-1), suggesting these groups form the basis for the communities of practice at the site. This hypothesis is further investigated below through more in-depth analysis of the relationships at higher similarity levels. Table H-1. Houses and Middens Included at BR 170 in Each of the Seven Node Cluster Groups for Each Sample Size. Node Cluster House/Midden Over 29 House/Midden Over 19 House/Midden Over 1 Group Vessels Vessels Vessel 1 a 52, 54, 60 52, 54, 60, 65, 72, 73, 74 52, 54, 60, 65, 72, 73, 74 1, 4, 7, 59, 61, 66, 71 1, 61, 63, 66, 71, 10, 77, b 65, 72, 73, 74 2 59, 61, 66, 71 80 3 13, 20 13, 20 13, 20 4 56, 62 56, 62 56, 62 5 77, 80 77, 80 6 51, 55 7 2, 4, 6, 7, 11, 59 Smaller-sample nodes often appear as outliers and pendants. This should, logically, pull the network apart, but the opposite occurs. In the Over 29 BR 170 network, the central area is empty, with several node clusters encircling it, forming a roughly rectangular shape. As more, smallersample nodes are introduced, these clusters become larger, in some cases due to the merging of previously separate node clusters. The larger node clusters are attracted towards one another, lowering the level of the difference between clusters while changing the general shape of the structure from rectangular to a more compact, square configuration. Nodes underlying others in 387 the Over 29 Vessels sample begin to drift apart, finally becoming visible in the Over 1 Vessel BR 170 network. Increasing attraction among the most similar nodes draws them into more complex relationships, thus rendering the strongest relationships more visually unclear. Network density decreases from an average of BR 173.8 for the Over 29 Vessels network to BR 168.9 for the Over 19 Vessels network, with five additional nodes, a decrease in the BR value of 4.9 (Table H-2). The addition of seven new nodes in the Over 1 Vessel network results in a much smaller decrease in network density. The BR reduction of just 2.1, to BR 166.8, confirms that these small samples do not have a large effect on the overall network. The standard deviation averages change in the opposite direction, rising from 12.7 to 17.1 and then to 17.03, respectively. The Over 29 Vessels network is visually and statistically a tighter network. Yet, the variation of only 7 points from the most exclusive to the most inclusive matrices in a scale measured out of 200 is only a 3.5 percent decrease in the strength of the average ties when all nodes are included. This suggests that, although these small samples do affect network structures, the effect is minimal and does not necessitate their exclusion. Consequently, while the larger-sample networks form the basis for the main corpus of observations, the smaller-sample nodes provide weaker, yet suggestive relationships and additional directions for investigation. The greatest dissimilarities across the village will be disclosed through the examination of the larger samples. Accordingly, exploration of the local tradition ceramic network, and the resultant communities of practice, will begin with visual examination of these relations, using the Over 29 Vessels matrix to build a foundation of ceramic interaction, with additional insight being gleaned from the smaller-sample matrices subsequently introduced. Table H-2. Network Density, BR Co-efficient Reduction, and Standard Deviation by Sample Size. Network Density BR Co-efficient Sample Size (BR Co-efficient) Reduction Over 1 Vessel 166.8 Over 19 Vessels 168.9 2.1 17.1 Over 29 Vessels 173.8 4.9 12.7 Standard Deviation 17.03 388 A second factor which greatly influences the graph structure is the measure selected for the lowest level of similarity at which ties will be depicted. Low levels produce infinitely more ties than do higher ones. This can result in almost unreadable graphs, as the multitude of lines can create a basically undecipherable “hairball” effect. Experimentation with different levels of similarity, in this case based on the Brainerd-Robinson Co-efficient of Similarity (Brainerd 1951; Robinson 1951), determine the level at which all nodes are linked into the main component of the graph with no excess production of lines. In this case, the most inclusive—that is, the Over 1 Vessel—data are chosen as the base measure. Although the cut-off level at which all nodes are incorporated increases slightly as the sample size rises, the BR 170 level is established as the standard for all matrix applications to ensure comparability. The graphs examined up to this point include only those of the lowest BR level, 170, at which all nodes are connected within the main component. As the Over 1 Vessel matrix has the lowest level at which this occurs, namely, BR 171, a BR value of 170 was chosen for the introductory verification of the decision to include all loci samples of Over 1 Vessel in this research. Although many general relationship patterns are elucidated at this measure, details of these relations become more explicit at high levels, where higher-value ties, or connections, are more clearly illustrated with the denser mass, also referred to as a hairball (Dermarkar et al. 2016; Hart et al. 2016), of lower-level connections. Values of BR 180 and BR 190 similarity have also been selected for examination. At these levels, large-scale changes in the network layout are evident. The few ties which exist above BR 190 are clearly evident and discernible in the BR 190 graph. Effect of Geographic Distance on Similarity Ties Network graphs produced by NetDraw, as described in Chapter 4, suggest directions of inquiry regarding possible relationships which might exist within the databases. Visual analysis and description of these relationships are proposed to reflect the dynamics of the communities of practice—that is, production and usage—of the local tradition ceramics. One of these dynamic forces is physical distance. The effect of geographic distance as it pertains to similarities in local tradition ceramic types within the Keffer village is explored through network graphs illustrating the approximate intravillage location of all house and midden ceramic samples (Figure H-4). In this figure, red lines represent ties at and above the BR 190 measure, or 95 percent similarity. Thick black lines 389 represent ties between BR 180 and BR 189, or 90–94.5 percent similarity, and thin black lines depict lower ties, of BR 170 to 179, or 85–89.5 percent similarity. Visual inspection indicates no apparent correlation between geographic proximity and ceramic similarity values. Edges of all three measures appear to span every sector and settlement cluster of the village equally. Figure H-4. Local tradition ceramics Over 1 Vessel BR 170 by cluster and approximate respective location. Line thickness denotes relative strength of tie. Cluster A is Yellow, Cluster B is Blue, Cluster C is Orange, Cluster D is Green. A good example of this lack of spatial correlation is manifest in the Over 29 Vessels BR 180 and 190 graphs (Figures H-5 and H-6). Adjoined cluster A Middens 57 and 60, both with large, 500+ vessel collections, share no ties at BR 180, although Midden 60 experiences five connections at BR 190 with other loci across the site, in Clusters B, C, and D. 390 Figure H-5. Local tradition ceramics Over 29 Vessels BR 190 Gower layout. Figure H-6. Local tradition ceramics Over 29 Vessels BR 180 Gower layout. 391 Closely related to the influence of physical location on local tradition ceramic relationships is the effect of settlement cluster origin. The lack of any spatially correlated colour patterning indicates that, at the lowest level, cluster origin may not be a primary factor in the ceramic relationships. Statistics on the average density of ties between and inside clusters, both interior and exterior (Table H-3), support this conclusion in a general sense but point to a more complex situation in regards to intra- and inter-cluster relationships. Table H-3. Average Tie Densities within and between Individual Clusters, the Combined Interior Samples, and the Combined Exterior Samples. Cluster Cluster A B Cluster C Cluster D Cluster A 167 161 165 175 Cluster B 161 167 164 163 Cluster C 165 164 166 170 Cluster D 175 163 170 181 Exterior 167 162 166 169 Interior 167 167 166 181 0 +5 0 +12 Interior/exterior difference Percent difference Total difference 2.5% 14 6 6% 6 18 Note: Boldface denotes intra-cluster density. These figures represent the average density of only those node ties within each cluster in relation to ties with those nodes outside the cluster. The average density of ties across all four clusters is 392 BR 167.3, with a range of only 7 points (or 3.5 percent), from BR 162 to BR 169. The general picture here is one of overall similarity. Two clusters, A and B, display identical interior average tie weights. Slightly stronger ties are seen within the remaining two clusters. In no instance does the average of the exterior ties exceed that of the average interior ties, indicating that in general shared cluster origin appears to have a weak positive effect on tie similarity. While the cluster density measures display a slightly different picture from the visual impression of dispersed cluster relations in the network graphs, there is an overall lack of consistency in the results. It is, therefore, apparent, despite omitting physical barriers or topographic differences, that geographic position within the village is not a major determinant of ceramic interaction. High-level connections may instead be temporal in nature, perhaps tracing change in ceramic community of practice through time, or within social groups, when influence of the founding communities of practice spread throughout the village. Social factors affecting ceramic change are elusive, however, as longhouse clusters, thought to represent matrilineally related social groups which formed pottery manufacturing groups, or communities of practice, are not highly visible in the weak pattern of intra-cluster ties seen in the statistical evidence of the non-existent or negligible intra-cluster tie densities and those of the interior-exterior values seen in Table H-3. This is also demonstrated in the graphic evidence portrayed by the dispersed nature of some cluster nodes seen in more inclusive graphs, below the BR 180 similarity level. In a related macro-scale study of collar motifs composed of a large number of pan-Iroquoian ceramic collections, John Hart determined that “geographic distance has little effect on pottery assemblage similarity” (Hart 2012:128). Taking this conclusion in concert with that seen above, distances between archaeological features are not considered highly influential factors in the following cross-village ceramic network analysis. 393 Appendix I: Appendix to Chapter 7: Chronological Patterning in Keffer Local Tradition Ceramic Social Network Analysis and Its Effect on Interpretations of Village Settlement Patterns Chronological Interpretation of Keffer Ceramic Network Analysis In this section, the network graphs are examined from the viewpoint of temporal patterning in order to interpret the occupation sequence of the Keffer village. This results in a series of five village plans, labelled Stages 1 to 5, suggested to be chronological in nature. These results are then combined with those of the settlement phases, Phases 1a to 2b-2, produced in Chapter 6 to verify that the patterns seen in the network topology reflect archaeological patterns in the ground. This culminates in a second level of possible temporal village development plans, labelled second-level Stages 1–5. Interpretation of the network topologies dealing with intra-site relations is dependent on the establishment and use of a frame of reference (see Betts 2008). Results of settlement studies in the form of settlement phases produced in section 6.5 provide this context in terms of a temporal backdrop or setting. Following the proposed stages, two houses, Houses 15 and 19, have been assigned chronological positions as, respectively, the first and last longhouses to be built at Keffer. Their appearance as topologically opposed isolated pendants circumscribing most house and midden collections in Over 19 Vessels BR 170 (Figure I-1) local tradition network intimates that chronology may be a fundamental component of the Gower and MDS layout graph structure. Therefore, analysis will continue on the supposition that an axis of the graph approximately horizontal in orientation roughly represents temporal change in the ceramic practice across this village. 394 Figure I-1. Local tradition ceramics Over 1 Vessel BR 170 MDS layout with BR 190 Blue Backbone structure. Following this logic, Midden 62, at the left end of the BR 190 backbone of highly connected nodes, is deemed to be the earliest component of the structure. Midden 66 located at the opposite extremity, or terminus, is regarded as the latest. Houses 13 and 20, forming a second and independent component at BR 190, rest together, in line with the Midden 62 node, as they do in the Over 29 BR 190 graph, indicating their similarly early presence. The absence of Houses 15 and 19 from the exclusive backbone graph (Figure 7-1) supports the interpretation of their positions as temporally peripheral to the main component; since the first and last residences in the village are expected to differ somewhat from those characterizing the midlife stage of Keffer’s long occupation. Production of a temporal and social sequence for residence occupation of the village relationships employs the backbone graph as its base. The sequence develops through time in this structure, from the relatively early Midden 62 (House 18), followed by the largest and presumably long-term Midden 65 (Houses 11 and 20). The tight pair of Midden 72 (House 13) and Midden 74 (House 12), Group 3, is the next to appear, tied by Midden 52 (Houses 4 and 8), a deposition over Palisade 1, to the somewhat later small world group (Group 2), which dominates this component. The closely related Midden 54 (Houses 7 and 8), Midden 60 (Houses 1, 2, and 3), Midden 61 (House 14), and Midden 80 (House 13 and possibly House 20) appear to be simultaneous depositions. The presence of Midden 54 over House 6, which itself lies over 395 Palisade 1, suggests a relatively late date for these depositions because Midden 52 must post-date the removal of Palisade 1, in Phase 2. Midden 66 (Houses 12 and 18) is the last deposition manifest within this backbone structure. It’s very late position over Palisade 2 may be the result of post-occupation slumping of the deposits. Houses 4, 11, and 7 form a small but tight group, Group 4, which appear around the same late time as Midden 66, in the Over 19 Vessels BR 190 network (Figures J-2 and J-3). The very small, 6 vessel, sample of House 10 is integrated into Group 2 with the introduction of the Over 1 Vessel data (Figures J-5 and J-6). At the same time, House 2, with 9 vessels, is added to the late House 7–House 11 pair. Although suggestive, the small sample results cannot be considered robust. As BR levels are decreased, new nodes are introduced into relationships with the existing nodes of the structural foundation. The pull exerted on the structure by these nodes misshapes it, resulting in altered relationships, both temporal and social, which are applied to the pre-existing backbone structure, not to the newly transformed structures. Still, these additional nodes must be viewed in the context of their closest ties at the level at which they first appear. For instance, nodes which first appear at BR 180 closer to Midden 72 than to Midden 74 must be characterized by this relationship in the comprehensive analysis. Final results will therefore not closely resemble any single graph produced from the data matrices but will reflect the combined results of all earlier network graph dissection. With the incorporation of 11 newly appearing nodes of the Over 29 Vessels matrix graph at BR 180, the temporal span of the previous network is expanded. House 19 forms the right-hand terminus of the graph. It is connected by the small world of Houses 2, 4, and 7 through House 11, Group 4, linking the network with the final occupation of Keffer with House 19. Midden 59 and Houses 1 and 9 appear somewhat earlier than this, as contemporaries of Midden 71 and Houses 13 and 20, Group 1. Middens 57, 55, 51, and 53, Group 5, extend the graph linearly towards the earlier limits of the graph, while Middens 56 and 73 and Houses 8 and 12 belong somewhat later, near Midden 65. Nodes appearing with the incorporation of BR 170 matrices, with the exception of Midden 75, are all characterized by samples of less than 30 vessels. These include the earliest house, House 15, which is temporally close Middens 67, 68, and 78. Houses 3 and 8 appear somewhat closer to the temporal middle of the graph, as does Midden 75. Midden 63 seems to share a similar age with the larger small world group, Group 2, being almost contemporary with the nearby House 1. Likewise, House 6 may be concurrent with the neighbouring House 11, towards the end of the village occupation. 396 This suggested chronology is an interpretation of the combined results from the various network graphs. As each graph displays relationships in a different light, each interpretation simply provides a direction for further investigation. Steps in the sequence, as seen below (Figure I-2), were divided by apparent gaps between nodes and node clusters as seen in the graphs themselves. The maps and plans are oriented with north up unless otherwise noted. Figure I-2. Original Stages 1–5 as established through social network analysis. The data employed in the original matrices were not selected for their chronological characteristics. Although some ceramic types have been postulated to possess temporal significance, this is not true of all types. The appearance of Neutral ceramics on central north shore sites may be more indicative of social factors. Nonetheless, the topology of the resulting graphs does appear to have temporal meaning in general. Stages 1–5, produced from visual analysis, propose a chronology of the site as a basis for further exploration of the data. Application of Local Tradition Ceramics Social Network Analysis Results to Phases The illustrations below show the phases as determined in Chapter 6, which portray the sum of all village features which potentially co-existed within that specific phase. As is apparent in Stage 1 (Figure I-3), a much greater number of features are present in the phase maps than are proposed in the sequence images. It is essential to include distinct natures of the house and midden deposition processes in the interpretation of the network analysis results. Midden deposits are not the direct product of primary occupation; they are the result of, and are reflective of, post-production and postconsumption deposition by individuals associated with principal residences. The assumption is made here that middens are the result of deposits of material culture originating from the 397 residence of closest proximity with clear, direct, and accessible pathways. That this is not true in all cases can be seen in the physical matching of a number of vessel sherds (House 9 and Midden 60) and pipe fragments (House 12 and Middens 72 and 74) from across various areas of the village. In most cases, however, vessel matches occur on sherds of loci in close physical proximity. Ceramics originating in these midden deposits are therefore interpreted as proxies for those houses to which they are assigned (Table I-1). Table I-1. Possible House–Midden Associations. House Middens 1 51, 57, 60, 63 2 51, 54, 57, 60, 63 3 57, 60, 63 4 52, 53, 54, 63 6 52, 55, 63 7 54, 55, 57, 60 8 52, 54, 55 9 55, 56, 58, 67 10 55, 56, 67 11 53, 54, 65 12 66, 72, 73, 74 13 66, 72, 73, 74, 75, 80 14 61, 66, 67, 68, 74 398 15 59, 67, 68, 75, 78 16 59, 66, 73, 78 18 61, 62, 66 19 71, 77 20 65, 71, 77 Note: Boldface indicates most likely associations based on proximity. Stage 1 (Figure I-3) sees the appearance of the first residence, House 15, and some mostly distant middens, all removed from the plateau surface. Midden 57 is located on the side of a large, open hill slope easily accessible to all areas of the village. Its deposition and growth, eventually spreading northward into the area of Midden 60, cannot be considered indicative of the concurrent habitation of nearby houses. As evidenced by the physical mend between sherds of House 9 and Midden 60, this area was most likely used as a refuse area by occupants across the site throughout the village history. Figure I-3. Stage 1: House 15a, Middens 51, 53, 55, 57, 67, 68, 78 (left); Phase 1a (centre); second-level Stage 1 (right). 399 In Stage 2 (Figure I-4), the village begins to expand, as Houses 3, 8, and 12 are erected in the area of future Clusters A, B, and C, respectively. On the hill slope south of the House 15b, expansion the new Midden 75 appears. Also in Cluster C, Midden 62 is present on the northern hill slope, presumably accompanied by the erection of House 18a. In the low town area downhill from House 15b, Midden 56 is accumulating inside the first palisade. Figure I-4. Stage 2: Houses 3, 8, 12, Middens 56, 62, 75 (left); Phase 1a (centre); secondlevel Stage 2 (right). In Stage 3 (Figure I-5), network graphs suggest that the small world group of House 10 and Middens 54, 60, 61, 77, and 80, Group 2, may materialize next. House 14, the likely origin of the Midden 61 material, may appear in parallel with House 13 to its south and House 20, a residence of similar age that is closely linked to House 13. Middens 72, 73, and 74 (Group 3), linked with Houses 13 and 20 (Group 1), are also from this period. The village’s largest midden, Midden 65, also dates to this period. The size of the ceramic assemblage from the midden, with 966 vessels almost 20 percent of the village total, points to long-term use of this refuse deposit, with local ceramic profiles averaging to this period. The presence of a large number of nodes within the central portion of the graph core, thought to be the temporal midpoint of village occupation, suggests that this large-scale growth occurred in its midlife. In general, there is little conflict between the sequences and settlement phases which reflect the possible order of the construction of the village features, houses, and middens. However, the presence of Middens 52 and 54, both of which overlie other village features, in the third sequence does not accord well with the 400 phases. It may be that the first palisade had been removed by this time, but Midden 54 must postdate House 6, over which it was deposited. Network position suggests that Midden 77, located at the southern end of House 20, is earlier than House 19. Settlement analysis shows that this midden overlies the eastern end of House 19. However, the longer duration of the House 20 occupation may account for the post–House 19 accumulation in Midden 77. Visual survey of all network graphs shows House 19 as consistently very peripheral in all cases, supporting its designation as the final residential construction. Figure I-5. Stage 3: Houses 10, 13, 14, 20, Middens 52, 54, 60, 61, 65, 72, 73, 74, 77, 80 (left); Phase 1b (centre); second-level Stage 3 (right). In Stage 4 (Figure I-6), several houses (Houses 1, 4, 6, and 9) and the small Midden 63 (near House 1, in the northern section of the village) share a highly linked position within the network and seemingly shared a community of practice. Although the small sample of House 6 makes its relationships tenuous, its construction across Palisade 1 would require the rebuilding of the village palisade, assigning this fourth sequence to the Phase 2a configuration. The ceramic relationships of Midden 59, which is also located within this sequence, argue for the existence of the associated House 16 simultaneously. Similarly, the co-existence of Houses 12 and 18 on the plateau is implied by the appearance Midden 66, although the origins for material in this deposit are not as assured. The occurrence of Midden 71 suggests a change in depositional habits of the south end residents of House 20 around this time. 401 Figure I-6. Stage 4: Houses 1, 4, 6, 9, Middens 59, 63, 66, 71 (left); Phase 2a (centre); second-level Stage 4 (right). Houses 2, 7, and 11 (Figure I-7) share similar temporal topological axis graph positions—distinct from other ceramic samples—and may represent the penultimate ceramic development of Group 4. The physical relationship of House 7 with Houses 6 and 7 and the overlying Midden 54 is complex, and it seems probable that House 6 was removed around this time, after a short period of occupation. Figure I-7. Stage 5: Houses 2, 7, 11 (left); Phase 2b-1 and Phase 2b-2 (centre); second-level Stage 5 (right). The last house to be erected in the village, as indicated by both the settlement pattern and the graph position on the periphery of the networks, is House 19. Its removal prior to the end of the 402 village occupation can be seen in the expansion of Midden 77, as suggested by the excavation records, across eastern end of House 19. The postulated short-term nature of the occupation is supported by its ephemeral nature. Final Keffer Village Settlement Plans, Maps 1–8: The Incorporation of Local Tradition Ceramic Profiles In the following section, the local tradition ceramic characteristics of each house and midden sample are employed to aid in the temporal placement of the features. This analysis results in a final set of eight settlement plans as described below. Map 1 Visual analysis of the local tradition network graphs and settlement evidence firmly anchor House 15 as the first occupation of Cluster C and organize the nodes of Middens 51, 53, 55, 57, 67, 68, and 78, the middens farthest removed from the plateau surface, within this first settlement stage. The nearby small deposits of Middens 68, 67, and 78 are also associated with Cluster C. The more distant Midden 51 is located at the far end of Cluster A to the northeast. Two larger surface middens, Midden 55 of Cluster B and Midden 53 of Cluster D, are located at the far eastern edge of the lower level. The large Midden 57 is located down slope from House 15, in Cluster A (Figure I-8, Table I-2). 403 Figure I-8. Stage 1 and Map 1. Table I-2. Settlement Changes in Map 1. Palisades Houses Added or in Use Expanded 15a Houses Removed Middens in Use 67, 68 The introduction of local ceramic profiles (Table I-3) for these loci provides a basis for the determination of a more nuanced sequence for the first occupation at Keffer. House 15 has by far the largest proportion of UNH vessels, at 40 percent, and the lowest of DNH vessels, at 55 percent. These numbers support the early appearance of this structure. Although the ceramic sample from House 15 is quite small, numbering 20 vessels, no other deposit on the site is characterized by such an even ratio of DNH and UNH ceramics. The ceramics of Midden 67, down slope from the south end of House 15a, and Midden 68, directly north of the house, contain slightly differing, but still low, proportions of undecorated neck vessels, 61 and 50 percent, respectively. The presence of Neutral ceramics in these samples, 22 percent and 17 percent, results in smaller proportions of decorated neck Huron vessels, 11 percent and 22 percent. Midden 67, with only 18 vessels, dates slightly earlier than its counterpart, Midden 67, due to the 404 presence of early, fourteenth-century horizontal motif, vessels. Located to the south of the fence line at the base of the plateau, labelled Palisade 3, Midden 67 may pre-date the erection of this fence line. The absence of nonlocal or exotic ceramics further supports the early temporal assignment of House 15 and Midden 67 and their affiliation early in the occupation of the high town. Given the small size and light occupation of House 15a, and the large area encompassed by the first palisade, it is unlikely that Palisade 1 was erected in this early period. Table I-3. Proportion of Local Tradition Ceramics in Second-Level Stage 1. DNH High Neutral Collar Niagara & Ripley 55 40 0 0 5 20–29 50 22.2 22.2 0 5.6 B <20 61.1 11.1 16.7 0 0 D 50–99 57.1 28.6 3.6 3.6 7.1 60.9 24.8 9.3 0.6 0.6 Times Locus Built Sample Map Cluster Size UNH H15 1 C 20–29 M68 1 C M67 1 M53 2 2 M55 2 B 100– 199 M51 2 A 50–99 60.3 24.1 10.3 1.7 3.4 M57 2 A >500 62.3 19 12.3 3.5 0.4 M78 2 C 20–29 45.5 18.2 27.3 9.1 0 Although Midden 78, which network analysis has suggested belongs in Stage 1, has an early ceramic profile, its location, near the south end of the House 15b extension, argues for a close relationship between this midden and the later extension of House 15b. For this reason, Midden 78 is not incorporated within the first occupation of House 15a in Map 1 (Figure I-8). The large size, slightly later ceramic profiles, and distance from the lightly occupied House 15a, suggests that neither Midden 51 nor Middens 53, 55, and 57 belong to this first construction. The prime location, diagonally oriented across the central portion of the plateau, suggests that the 32 m House 15a was initially erected without forethought to future development. This structure may, like those of the Hidden Spring site, have originated as a special purpose site and may not have originally been intended to serve as the foundation for a future village. If that is the case, 405 the external post moulds, originally labelled House 17 in the field, may relate to this early construction. Alternatively, House 15’s unique and early ceramic profile, not replicated elsewhere on the site, may signify a simplified ceramic assemblage reflecting the practical orientation of an advance construction party, the initial settlers of the community. A similar topographic position, also crossed by two later longhouses, is suggested for the first house at the later Rouge–Duffins Mantle village (Birch and Williamson 2013:69). Like that of House 15 at Keffer, the orientation of this first house at Mantle does not correlate with that of the later structures. Thus, settlement and ceramic data suggest that the first presence at Keffer is most likely composed of House 15a and the small, nearby Midden 67 and Midden 68 refuse deposits, as illustrated in Map 1 (Figure I-8). Map 2 The Keffer village sees a large-scale expansion in its second occupation stage, with the addition of eight new structures (Figure I-9, Table I-4). The configuration of the first palisade, neatly enclosing the ends of these buildings but leaving room for intra-village movement, suggests a planned construction of this first village within the Palisade 1 walls. Figure I-9. Stage 2 and Map 2. 406 Table I-4. Settlement Changes in Map 2. Houses Remove d Palisade s in Use Houses Added or Expanded 1, 3 1a, 3a, 5, 7a, 8, 9a, 10a, 12, 15b Middens in Use 51, 53, 55, 56, 57, 74, 75, 78 Network analysis deals only with complete household ceramic samples and cannot portray settlement feature changes such as structural expansions of houses already considered present. In the case of House 15 the ceramic profiles of nearby associated middens, thought to originate in the extension, can be used as a proxy for the ceramics of House 15b. Ceramic profiles from Middens 75 and 78 (Table I-5), located down slope from the end of the House 15b southern extension, display an early date, suggesting that they are related to this expansion. Midden 78 has the lowest proportion, 45.5 percent, of UDH ceramics on the site, but the small, 25 vessel, size of the sample and the appearance of 9 percent high collar vessels, characteristic of the mid-fifteenth century, in addition to an 18.2 presence of DNH decorated neck Huron ceramics and a 27.3 percent presence of Neutral ceramics, suggest that this is not the earliest midden associated with House 15. Table I-5. Proportion of Local Tradition Ceramics in Map 2. Times Locus Stage Built Map Sample Cluster Size UNH DNH Niagara High & Neutral Collar Ripley M53 2 D 50–99 57.1 28.6 3.6 3.6 7.1 60.9 24.8 9.3 0.6 0.6 1 M55 1 2 B 100– 199 M51 1 2 A 50–99 60.3 24.1 10.3 1.7 3.4 M57 1 2 A >500 62.3 19 12.3 3.5 0.4 M78 1 2 C 20–29 45.5 18.2 27.3 9.1 0 M56 2 2 B 30–49 66.7 25.6 2.6 5.1 0 407 H12 2 1 2 C 100– 199 H8 2 1 2 B <20 66.7 20 6.7 6.7 0 M75 2 2 C 30–49 62.1 10.3 20.7 6.9 0 H3 2 2 A <20 66.7 0 16.7 16.7 0 M74 3 2 C >500 69.6 20.2 6.9 1.2 0.7 2 70.4 21.4 2 1 5.1 Cluster C In Midden 75, the low level (66.7 percent) of undecorated neck Huron vessels and small presence of decorated neck vessels (11.1 percent) is offset by the sizable incidence of Neutral ceramics, at 22.2 percent. Surprisingly, no Neutral ceramics were found in House 15. The expansion of House 15b marks the beginning of large-scale growth at Keffer, the beginning of a village community. The extension of House 15b to 65 m makes it the longest house at the settlement. However, the low density of internal post moulds and features and the lowest wall post density of any longhouse at the site, at 3.8 ppm, argue for a short occupation span for this residence. Ceramics suggest that the next construction in the growth of the Cluster C group is that of House 12. The large increase in undecorated neck Huron vessels, to 70.4 percent, up from 55 percent, at House 15, and the significant drop in the proportion of decorated neck Huron vessels, from 45 percent at House 15 to 21.4 percent at House 12, indicate that the main occupation of this residence occurred sometime after the construction of House 15b. House 12, a relatively large structure at 44 m, is located directly west of House 15b in the largest open space left on the plateau. This longhouse has abundant evidence of long-term occupation in the form of a high concentration of interior posts, pits, hearths, and sweatbaths and a higher density of exterior wall posts (6.4 ppm). House 12 is surrounded on the down slope surface to the south and west by the second largest midden at the site, Midden 74. The extreme proximity and very similar ceramic profile of this midden to that of House 12, with 69.6 percent undecorated neck Huron pottery and 20 percent decorated neck Huron ceramics, testify to a strong relationship between this residence 408 and the refuse deposit. Although it was in use throughout most of the village’s lifetime, as seen in the proportionally high level of nonlocal tradition ceramics (9.6 percent), this midden probably originated with the erection of House 12. Cluster A With most of the available space on the plateau occupied by Houses 12 and 15b, further growth in the settlement was restricted to the lower area, called the low town by the field crew, nestled to the east, between the plateau and nearby hillside. On this lower level of the site, Middens 51 and 57, which were assigned by topographic position in the network graphs to the first stage, display higher proportions of UNH, at 60.3 and 62.3 percent, respectively, than the very small middens associated with House 15a. These midden samples are matched more closely with those of House 3, another small sample, of only 6 vessels, with 68 percent UNH, in Cluster A. This house is placed through network analysis into Stage 2. While House 3a may have been the first residence, the small ceramic sample renders this early temporal assignment insecure. A somewhat later assignment may be possible since no DNH ceramics are present in the sample, which holds both Neutral and High Collar vessels. The first construction of House 3a measures a mere 20 m, making it the shortest house at the site. This structure shows evidence of short-term occupation, with a relatively low wall post density, of 4.1 ppm, and a low density of interior features. In addition, its side and end walls are entwined with those of the later House 2, indicating it does not remain throughout the occupation of the village. Wall post alignment strongly supports the co-presence of Houses 3 and House 5 (Figure I-10), a round structure of unknown function with a diameter matching the width of the village longhouses, at just under 8 m. The western edge of this structure directly abuts the eastern limits of the extended House 3b and likely pre-dates this expansion, therefore placing the House 5 construction during this first village growth period. No analyzable ceramics were retrieved from “House” 5, although a central hearth is evident in the structure. 409 Figure I-10. Map showing postmoulds of House 5. The orientation of Houses 3 and 5 and the contemporary House 8 in Cluster B suggests that these houses were erected in isolation. The parallel alignment of Houses 1 and 7 (Figure I-11) in Cluster A suggests, following the work of Dodd (1984) and Warrick (1984), that these houses were, if not erected in one preplanned event, at least contemporary, and that they may have constituted a social unit. Directly south and parallel to Houses 3a and 5 is House 7. The placement of House 7a within this first stage of village development relies almost exclusively on strong settlement data. Its contemporaneity with Houses 1 and 3 is supported by its parallel alignment with these two structures. The exterior wall post density of 13.4 ppm represents the highest wall density of any Iroquoian longhouse investigated to date (see Chapter 6). Analysis of the wall post configuration suggests that this density is the result of at least two reconstructions of House 7 in the same location, with the first two structures measuring 30 m. Iroquoian longhouses were often subject to repair during their lifespan. Evidence of such repair is generally accepted to indicate a long occupation for the structure. Rebuilding is not as common and may have occurred most often after destruction of buildings by fire. There is no evidence of fire noted for this or any structure at Keffer, and reconstruction of House 7 is presumed to be an indication of structural deterioration due to long-term occupation. The existence of a House 7b hearth over top of the walls of the later House 6 also confirms a prolonged occupation for House 7. A high density of interior hearth and sweatbath features also supports this conclusion. The small ceramic sample (25 vessels) is composed of 84 percent UNH and 12 percent DNH ceramics and one early vessel. The extreme longevity of House 7, in conjunction with its small ceramic sample size, easily explains the network topology trend to a later occupation date, as the large majority of the 410 ceramics associated with the house may have been relocated during successive cleanings, repair, and reconstruction. Figure I-11. Map showing wall post and hearth features of Houses 1 (top); 2, 3, and 5 (centre); and 7 (bottom). The parallel alignment of House1 with Houses 3, 5, and 7 in Cluster A indicates that House 1a was likely constructed in concert with Houses 3a, 5, and 7a. Measuring 37 m, House 1a is longer than House 7a and much longer than House 3a. Like House 7, House 1 displays characteristics of long occupation. Wall post configuration and the exterior wall post density, of 10.7 ppm, show clear evidence of total rebuilding and expansion of the structure (Figure I-12). The long occupation is also seen in the high interior feature density, primarily in the form of numerous sweat lodges. These small post features form an almost continual column down the central core of the structure. The relatively high proportion of DNH vessels, making up 17.6 percent of the local tradition ceramics, hint at comparatively relatively early occupation, while the UNH ceramics, which make up 76.5 percent of the sample, illustrate a trend towards a later occupation. The longevity of the household occupation could easily account for these proportions. 411 Figure I-12. Map showing posts and features of House 1. Midden 51, just outside Palisade 1, near the east end of House 1a, may originate with the construction of this first Cluster A house. Decorated neck vessels make up 24 percent of the ceramic sample. Undecorated neck vessels compose just 60 percent, with Neutral ceramics adding another 10 percent. These numbers support an early appearance of this midden. A pipe fragment match with Midden 64, a small surface midden located just inside Palisade 1, near the end of House 1a, suggests that refuse in Midden 51 originated in this area of the cluster. Midden 64 has no identifiable vessels and is therefore not included in this analysis. Midden 57, one of the largest middens on the site, is located on the riverside slope to the west of Houses 3a and 7a and close to House 1a. A vessel match between Midden 57 and House 7 confirms their association, although both the midden and the house show evidence of long-term use suggesting Midden 57 may have not overlapped House 7 for a long period. With a very similar ceramic profile to that of Midden 51, of 19 percent DNH, 62 percent UNH, and 12 percent Neutral, Midden 57 may also be first created with the construction of these first houses. Ceramic and settlement data suggest that Houses 1a, 3a, 5, and 7a were built in Cluster A during this first occupation of the low town. Middens 51 and 57, at opposite ends of this cluster, originated at the same time. Parallel to Houses 1a, 3a, and 7a is House 8. It is located 5 m south of House 7a, in the area of the second low town residential group, Cluster B. This cluster is composed of three longhouses: Houses 8, 9a, and 10a. Village layout suggests that Houses 9a and 10a, which form a cluster with House 8 around the open area containing Midden 55, are erected at the same time as House 8. The ceramics of House 8 support its early temporal assignment and synchronic presence in this 412 cluster. Decorated neck Huron vessels make up 17 percent of the local tradition ceramics in the sample. Undecorated neck Huron vessels portray a relatively early profile for the house, making up 67 percent of the local tradition ceramics. House 8 is unique in that it is the most isolated structure of both the densely populated area of the northern portion of the low town and the plateau. In these areas, houses are generally constructed very close together and often overlay one another. Yet the low density of interior features and lack of structural expansion suggest short-term use for this longhouse. The only other houses on the site with these characteristics, Houses 2, 6, and 19, are also thought to be short-term residences. Yet the presence of wall trenches in House 8 and the relatively high wall post density of 7 ppm suggest that this house may have been renovated or rebuilt. In addition, the deposition of Midden 54, late in the village history, occurs directly north of House 8, and while it covers both Houses 6 and 7, it is not noted as covering House 8. Midden 55 is located just outside the western end of House 8. Its early occurrence in this cluster is suggested by the high, 27 percent, proportion of DNH ceramics and low, 60 percent, proportion of UNH pottery. Network topology places this midden, House 8, and Midden 56 in this building stage. The higher proportion of UNH pottery, at 67 percent, but similar proportion of DNH ceramics, at 26 percent, may indicate a slightly later date for this midden than the first appearance of Midden 55. The location of Middens 55 and 56, at opposite ends of Houses 9 and 10, supports the placement of these two houses in this stage. However, the coexistence of Houses 9 and 10, though supported by close parallel alignment, cannot be confirmed ceramically. Due to the superimposition of the later Midden 58, only 6 vessels are definitively assigned to House 10. All of these House 10 ceramics fit in the UNH category, with no evidence of neck decorated pottery in the sample. The larger ceramic sample of House 9, composed of 40 vessels, is similar to that of House 1, with a proportion of 74 percent undecorated neck Huron and 7 percent decorated neck Huron. Unlike most samples with a high level of UNH, indicating a possibly later date for this house, House 9 has a relatively high proportion, 14.3 percent, of Neutral ceramics, which are more common in earlier samples. A substantial line of sweatbath features along the central hearth line, extensions at both ends of the house and a wall post density of 7.9 ppm give credence to long-term occupation of House 9. The ceramics, which date it slightly later, may be reflective of this long occupancy. Settlement evidence suggests that the parallel buildings Houses 9a, at 36 m, and 10a, first erected at 22 m, were constructed together. House 10a is surrounded 413 on three sides with a fence line. There is a 5 m gap between the house and the surrounding fence line, labelled Palisade 3. The purpose of this barrier is speculative, but its position at the base of the plateau’s eastern slope implies some form of protection from events commencing at the hilltop, perhaps runoff from precipitation. The co-existence of this fence line with the Houses 9 and 10 forms a barrier to movement across the base of the slope. However, a mend between sherds from House 9 and Midden 60 shows that occupants of Cluster B used the lower-level hillside Middens 57 and 60 for refuse deposition. Access to this area may have been restricted for the occupants of House 10, and possibly House 9, with these obstructions. This situation may have resulted in the deposition of the surface Midden 55 at the apex of the three Cluster B houses. Although Midden 55 appears to hinder interaction with the more northerly Cluster A, its presence in the early sequence is substantiated by the relatively high proportion of 26.5 percent decorated neck Huron and the relatively low proportion of 60 percent undecorated neck Huron pottery. With only 57 percent UNH and 28.6 percent DNH ceramics, Midden 53, located 20 to 50 m from the Cluster B houses, at the base of the eastern hillside, may be an alternate refuse deposit for the occupants of this area in this early period of occupation. It is likely that the first palisade, Palisade 1, was constructed at this time. In most cases, there is a distance of 10 m between the barrier and the long house ends. The exception to this appears to be House 1a, which may have been built right up to the palisade wall in the east. Map 2 (Figure I-9) portrays the characteristics of the first large-scale occupation of Keffer, as new inhabitants create socially distinct residential groups embodied in the clustering of these first houses. Construction of the 44 m long House 12 and the expansion of the already lengthy House 15a from a 40 m structure to a 65 m structure dominating the plateau left little room for further expansion in Cluster C. Although House 15 was originally responsible for small middens located off the hilltop, heavier occupation led to the creation of Midden 74, to the west of House 12. It may be at this time that settlement spread to the base of the plateau, with the establishment of two residential clusters. To the north, cluster A was composed of four structures aligned basically east–west, in a parallel fashion. The first construction of these three longhouses ranged in length from the House 3a, at 20 m, to House 7a, at 30 m, and House 1a, at 37 m. Refuse from these new houses formed Midden 57, on the open hillside to the west of House 7a, and the small Midden 51, outside the palisade, near the eastern end of House 1a. In Cluster B, three more 414 houses were erected. House 8 alone was built at its final length, of 40 m. Although it is equidistant from both House 7a in Cluster A and House 9a in cluster B, the fact that its orientation forms a circular plaza with Houses 9a and 10a suggest that it may be part of this social grouping. This plaza area was the site of Midden 55. A 50 m fence line was erected around the northern end of the 22 m House 10a, perhaps protecting it from upslope erosion. House 9a, at 36 m, is situated parallel to House 10a. The presence of Palisade 1, erected at this time, may have provided a sense of unity and belonging for settlers of this new village. Map 3 Differences in ceramics suggest that the large Group 2 community appears to have developed in two stages. New additions to the village display two principal ceramic practice communities in Stage 3 (Figure I-13, Table I-6). The location of the construction of House 20, outside the village palisade, represents one of the defining characteristics of this changing village. The establishment of House 20 is mirrored by the coinciding construction of House 13a nearby, on the plateau, just inside the palisade. This highly connected pair of houses displays ties most similar to the Trent valley Kirche site, followed by Logan. The ties which remain with Logan when the Trent ties are removed are deemed more relevant, although further investigation into the nature of the Trent ties may prove informative in terms of their significance. 415 Figure I-13. Stage 3 and Map 3. Table I-6. Settlement Changes in Map 3. Palisades Houses Added or in Use Expanded 1, 3 3b, 9b, 10b, 13a, 14a/b, 18a, 20 Houses Removed Middens in Use 15 60, 61 (H14), 62 (H18), 65, 72, 73, 77, 80 Ongoing changes in the village population dynamics continue to be centred on the occupation of House 15, on the central plateau. Network and settlement analyses indicate the introduction of several new households on the hilltop. Ceramic profiles suggest the introduction of Houses 13a, 14a, and 18a. It is possible that House 16a appears as part of this new settlement organization, although slightly differing ceramic ratios suggest that it may be somewhat later. The low level of interior posts and features seen in House 16 supports a later date for this longhouse (Figure I-14). The three new longhouse structures sit in parallel alignment with the pre-existing House 12, and their construction necessitates the removal of House 15b, which up until this point dominated the village. 416 Figure I-14. Map showing posts and features of House 16. House 13 overlies the southern expansion of House 15 and must have been established following the latter’s removal. The early ceramic character of House 13, with 25.2 percent DNH and a relatively modest presence of 72.9 percent UNH vessels (Table I-7), suggests that the occupation of House 15 through both its stages was short-lived and early in the village’s life. This brief lifespan is supported by the fact that House 15 has the site’s lowest post mould density, at 3.8 ppm, and that, unlike those residences with more lengthy occupation, it possesses no wall trenches and shows no evidence of rebuilding or renovation except for its initial expansion. Table I-7. Proportion of Local Tradition Ceramics in Map 3. Locus SecondLevel Times Stage Built Sample Cluster Size UNH DNH Niagara High & Neutral Collar Ripley M65 3 - D >500 67.0 19.9 8.7 1.4 1.5 M62 2 - C 200– 499 67.5 23.8 5.3 1.3 2 M73 3 - C 50–99 68.9 15.6 6.7 4.4 2.2 M72 3 - C 50–99 69.9 19.3 9.6 1.2 73.1 25 H13 3 2 C 100– 199 M60 3 - A >500 73.6 15.7 H20 3 1 D 200– 499 73.8 24.1 0.9 7.2 2 0.6 1.3 0.8 417 M77 3 - D 100– 199 M80 3 - C 50–99 75.3 15.1 5.5 2.7 1.4 M61 3 - C 200– 499 75.7 15.5 5.6 1.3 1.6 75.2 17.7 4.4 1.8 0.9 The first version of House 13, 13a, measuring 32 m, sits just within the southern limit of Palisade 1, parallel to and 8 m east of House 12, a slightly earlier structure. House 13 has a somewhat earlier ceramic profile than other loci associated with this period, and the absence of Neutral ceramics marks it as unique in Cluster C. To the southeast of House 13a, a small midden appears around this time. Midden 73’s placement over a portion of House 15b’s southern end attests to this placement. That Midden 73 post-dates House 15 and is contemporary with House 13 is also evident from its ceramics, with 68.9 percent UNH and only 15.6 percent DNH. Midden 73 is directly adjacent to the east wall of House 13. Like most other deposits on the plateau, Midden 73 is characterized by a small proportion, 6.7 percent, of Neutral ceramics. Network analysis and local tradition ceramic seriation suggest that Midden 72, located somewhat farther southwest of House 13, is contemporary with Midden 73. The ceramic profile of this midden almost mirrors that of Midden 73, with 69.9 percent UNH and 19.3 DNH. Midden 72 possesses a slightly larger proportion of Neutral ceramics, 9.6 percent. Another new deposit linked with the appearance of House 13a is Midden 80, on the hill slope to its south. This midden, with slightly less, 14 percent, DNH and a more, 74 percent, UNH, illustrates the continuing occupancy of House 13 as these houses grow with future extension. Midden 80 is similar in ceramic profile to Middens 61 and 62. These three refuse deposits likely appear simultaneously and are fundamental to the illustration of village growth at this time. Midden 61, one of the larger deposits at the site, with 413 vessels, is located on the northern slope of the plateau, directly adjacent to House 14a. The position of this deposit to the east of the house, distant from other hilltop residences, leads to the assumption that Midden 61 is directly related to the occupation of House 14. The ceramic profile of this midden, with 16 percent decorated and 76 percent undecorated neck vessels, and therefore of its associated house, House 14, is well matched with that of House 13, suggesting their co-existence. In addition, a vessel 418 mend (Vessel 740023) among three sherds from House 14, House 12, and Midden 74, confirms the coexistence of House 14 with other residences on the plateau. The absence of a ceramic rim sample from House 18, located 2 m to the west of House 14, is unfortunate. However, information on the temporal and social relations of the occupants can be surmised by employing the characteristics of the adjoining midden sample as a proxy (Table I-7, Figure I-13). This is supported by the close proximity of Midden 62, a large midden directly north of and down slope from the plateau at the north end of House 18. Local tradition ceramic type proportions of 24 percent decorated neck and 68 percent undecorated neck Huron support the chronological placement of Midden 62 and House 18 at this time. The presence of a small proportion of Neutral pottery, 5 percent, ties it with samples of Cluster C. Network topology originally placed Midden 62, and therefore House 18, earlier in the village history. However, the ratio of decorated to undecorated neck Huron vessels suggests that this midden, and its associated house, House 18, are not characteristic of an earlier period. The site plan, which shows the walls of Houses 15b and 18b touching each other, also supports the suggestion that these two houses are not contemporary. The appearance of cultural material above both the first and second palisades demonstrates that the bulk of the deposits are related to this later residence, House 18a. The close proximity of Houses 14 and 18, at 2 m, suggests that space for construction at the top of the plateau may have once again been an issue. While large-scale transformation in palisade expansion affecting all areas of the site may be reflected in simultaneous settlement pattern modifications, changes in village dynamics on the plateau and in the lower level cannot be temporally correlated on a small scale. The renovation of the plateau Cluster C therefore does not directly correlate with that elsewhere. However, ceramic profiles related to these changes support relative temporal contemporaneity for these events. In low town Cluster A, to the north the concentrated refuse deposit of Midden 57, which was generated by the initial housing, clusters begin to spread northward, and a new midden, Midden 60, with ceramics dating slightly later that those of Midden 57, develops. Only 15 percent of the low town local tradition rims are characterized by neck decoration, while 73 percent are undecorated. The expansion of House 3b from 20 m to 34 m at this time may be partially responsible for this growth, though a mend between Midden 60 and House 9 suggests that an 419 influx of refuse from Cluster B may be occurring as well. The increases in sizes of House 9b, extended by 6 m, and House 10b, extended by 7 m, are the only significant changes in Cluster B. Increasing population in this group may also be responsible for the increased deposits on the steep riverside slope. The most significant developments at Keffer during this period are the appearance of House 20 to the south, isolated at a distance of 30 m from the village palisade, and the creation of the adjacent refuse deposit, Midden 65. This surface midden with 944 identified vessels is the largest of all middens at Keffer. The local tradition ceramics portray an early temporal placement for Midden 65, with 19.9 percent neck decorated pottery and relatively few, 67 percent, undecorated neck vessels, perhaps hinting at an origin predating that of Middens 60, 61, and 80 on the plateau. Neutral ceramics account for 8.7 percent of the collection, and Early ceramics account for 1.5 percent. In fact, only the earliest ceramics—that is, late Middle Woodland and early Late Woodland, such as Ontario Oblique, which is very rarely seen in the greater Keffer collection— are absent from Midden 65. The extensive nature of this ground midden and its well-preserved state, together with the large physical size of a significant proportion of the sherds, may possibly be attributed to the midden location within a wide but shallow depression on the village’s southeastern periphery (Figure I-13). Contributing to the creation of this new midden were the occupants of a newly constructed longhouse, House 20, outside the palisade wall. Directly adjacent to Midden 65, on somewhat higher ground to the south, House 20 shows a similar, but slightly earlier, proportion of Huron vessels, with 24.3 percent decorated neck to Midden 65’s 20 percent and a slightly higher proportion, 74.5 percent, of undecorated neck rim sherds to Midden 65’s low 67.3 percent. No Neutral vessels are found in House 20. Unlike other houses with evidence of long-term, heavy occupation, House 20 does not appear to have been the subject of expansions or rebuilding. Construction of House 20 prior to Houses 4 and 11 is supported by the comparatively favourable location of this longhouse. House 20 enjoys a wideopen location on the southern portion of the site, with few physical constraints and abundant room for movement. Houses 4 and 11 do not share this benefit, as they are both located close to a palisade wall. The wall post density of 6.6 ppm may instead reflect structural repair over the course of its lifespan. Midden 77, a small surface midden with 57 vessels, is located at the southwest corner of House 20 and near the base of the surrounding hillside slope. The ceramics of this midden suggest that it originated with the creation of House 20. The 17.7 percent 420 decorated and 75.5 percent undecorated neck ceramics correspond well with a slightly later ceramic profile than that of House 20 and the existing Midden 65 repository. The erection of longhouses exterior to the village palisade can sometimes be attributed to longterm, or overwinter, visits from Algonkian groups to the north. In this case, however, the ceramic profile of House 20 is almost identical to that of House 13 on the plateau. These two ceramic collections in fact form the tight-knit Group 3 of the network analysis, suggesting a close common ceramic tradition for the occupants of both houses. All signs suggest that House 20 was not the home of temporary visitors. Map 4 The next development within the Keffer village also takes place outside the village palisade, in Cluster D. The ceramic profile and wall post alignment of House 11a (Figure I-15) suggest its construction at this time just north of Midden 65. The northwest corner of the 41 m long House 11 lightly touches the palisade wall, projecting from it at an angle of 10 degrees. Its existence prior to the removal of the early village palisade, Palisade 1, can be seen in its well-aligned and snug fit abutting the palisade wall. In a similar fashion, the outside corners of Houses 1 and 13 briefly meet the palisade. In this location, House 11 is not crowded into the available level space and has good access to all areas across the village. This first rendition of House 11 (House 11a), was the last longhouse to be erected during the Palisade 1 lifespan (Figure I-16). Despite the virtually identical ceramic profiles of Houses 4 and 11 (Tables I-8 and I-9) (with 84 and 16 percent and 84 and 13 percent undecorated neck vessels and decorated neck vessels, respectively), settlement pattern analysis suggests that House 11a was erected prior to House 4, which was likely constructed concurrently with the second palisade. The possible appearance of a keyhole-shaped semi-subterranean sweat lodge adjacent to the east wall of House 11 (Figure I15) would support an early date, as these features are characteristic of the late fourteenth and early fifteenth century and begin to disappear by the mid-fifteenth century (MacDonald and Williamson 2001; Williamson 2014:11). 421 Figure I-15. Detail map showing posts and features of House 11a. East is up. Figure I-16. Stage 4 and Map 4. Table I-8. Settlement Changes in Map 4. Palisades Houses Added or in Use Expanded 1, 3 11a Houses Removed Middens in Use 422 Table I-9. Proportion of Local Tradition Ceramics in Map 4. SecondLevel Times Built Locus Stage Sample Cluster Size UNH DNH Niagara High & Neutral Collar Ripley H11 D 12.9 0 4 3 30–49 83.9 0 0 Map 5 The changes seen in the last three maps took place in a gradual manner, with houses being built, removed, repaired, and rebuilt in a constant pattern of change (Figures I-17 to I-19, Table I-10 and Table I-11). After the construction of the first settlements in the village, it is unlikely that wholesale renovation and rebuilding of the village happened in steps. A project on this scale would have entailed a huge workforce, interrupting the ongoing production and procurement of food. Figure I-17. Stage 4 and Map 5. 423 Table I-10. Settlement Changes in Map 5. Palisade Houses Added or in Use Expanded Houses Removed Middens in Use 2a 3b, 5 1b, 2, 4, 6, 7a, 11b, 13b 59 (H16), 63, 66, 71, 52 Table I-11. Proportion of Local Tradition Ceramics in Map 5. SecondTimes Level Locus Stage Built Sample Cluster Size UNH DNH Niagara High & Neutral Collar Ripley D 100– 199 79.3 16.3 4.3 0 0 A <20 71.4 14.3 0 0 0 M71 4 H6 4 M52 5 D 100– 199 72.2 20.1 5.6 0.7 0 M63 4 A <20 72.7 18.2 0 9.1 0 M59 4 C 50–99 76.3 15 2.5 0 6.3 M66 4 D 50–99 79.3 12.1 6.9 0 1.7 H4 4 1 D 20–29 84.2 15.8 0 0 0 H2 4 1 A <20 87.5 12.5 0 0 0 1 However, the introduction of these two new longhouses, House 11a and House 20, outside the first village palisade marks the beginning of large-scale change at Keffer. This transformation is exemplified in the next step of the village development. Sometime after House 11a was squeezed into the largest remaining flat surface in the village area, the need for further expansion may have prompted the removal of the first palisade, which was greatly constricting growth of the existing longhouses. The removal of Palisade 1 and its replacement with Palisade 2, which encircled the greatly expanded community, was the foundation of settlement and social modification. Evidence of rapid population growth, as represented by increased indoor living 424 space, is seen throughout the site. Social changes are reflected in the changing dynamics of residence location. With the expansion of House 11b to 55 m and the construction of the 43 m long House 4, just inside the eastern palisade wall, existing social relations may have expanded or changed. The area formerly separated into Clusters A and B and the northern section of Cluster D is more open with the palisade removal and may have re-formed into a more cohesive social unit. Major changes take place in the layout of this area. The removal of the first palisade allows for the expansion of the heavily occupied House 1b by an additional 8 m, to 45 m. At the same time, Houses 3 and 5 are taken down and two new longhouses, Houses 2 and 6, replace them. House 2 replaces House 3. House 2 has a slightly more parallel alignment with Houses 1 and 6, while the previous House 3 was more in line with House 7a. It may be that House 7a was also rebuilt at this time, but its position does not change. Directly to the east of House 7a, House 6 is built very close to House 4 but parallel to the other houses surrounding it, Houses 1b, 7, and 8. The perpendicular alignment of House 4 to this group is most likely due to spatial limitations imposed by the rising slope of the hillside to the east. Houses 1, 4, 6, and 9 and the small Midden 63 share a highly linked position within the network. Network analysis groups these houses on the later periphery of the network, suggesting contemporary construction and occupation. Although each of these houses, Houses 2, 4, 6, and 7, have small ceramic samples, fewer than 30 vessels, their ceramic profiles consistently portray a late date in the village history. The ceramic sample of House 7, which was arguably built in the first major construction period, may represent only one period of the house’s occupation, as ceramics found in the overlap with House 6 were not included in the analysis. Although the same holds true for House 6, the sparse interior features suggest it did not experience a long-term occupation. The eight ceramics of House 6 cannot be relied upon for temporal assignment, and its relative date within the village lifecycle is not therefore refuted ceramically (the proportions are 14.3 percent undecorated and 71.4 percent decorated neck Huron). Likewise, the small ceramic sample of House 2, at nine vessels, does not produce robust statistics, but the 12.5 to 87.5 percent ratio of decorated to undecorated neck Huron hints at a late date for this household, particularly when compared with the 17 and 67 percent proportions of the earlier, now-replaced, House 3. The ceramic profile of House 4, with 84 percent UNH and 16 percent DNH, closely matches that of House 2. Although the majority of refuse from these houses most likely ended up in the large hillside middens to the west of the structures, a small 425 surface midden, Midden 63, with 12 vessels, appears near the end of House 2, between Houses 1b and 6. A vessel match between this midden and House 2 supports its association with House 2 and its late date. The somewhat low, 73 percent, UNH and high, 18 percent, DNH ceramics for this time period are overpowered by the match with House 2, whose late date is supported by its replacement of House 3. A second surface midden, Midden 52, is located between House 4 and the end of House 8, over top of the Palisade 1 post line. The ceramic profile of 72 percent UNH and 20 percent DNH parallels that of Midden 63. A vessel match with House 4 supports the late date of House 4 in that it post-dates Palisade 1. It also provides a more solid dating for the house due to its robust sample of 153 vessels. The growth seen in the low town during this period was paralleled by longhouse expansion in the high town. House 13, whose southern end previously abutted the southern reaches of Palisade 1, is extended 8 m after the palisade’s removal. On the northern end of the plateau, House 18 sees and expansion to both the north and south with the removal of the palisade, growing from 22.5 to 33 m. At this point, the only real space left on the plateau is utilized for the building of a new structure, House 16. With only one ceramic vessel, the dating of the event is based on the small midden located at its northwestern end. No other midden exists in proximity to this house, as it rests on a promontory with little flat ground surrounding it. The late construction of this building is suggested by its limited access and low density of interior features, particularly in its extension. The ceramic sample from Midden 59, which rests beside it on the top of the slope, is composed of 76 percent UNH and 15 percent DNH ceramics. Like most deposits on the plateau, it also has a small proportion, 2.5 percent, of Neutral ceramics, making House 16 the latest longhouse to be built in Cluster C. The effects of this population growth on the plateau may have resulted in the production of a new midden over the remnants of Palisade 1 on the western hillside. The ceramic characteristics of Midden 66—situated down slope, to the west of Houses 12 and 18b (Table I-11)—with 12 percent decorated neck and 79 percent undecorated neck vessels, indicate a late date for the creation and use of this deposit. Its central position on the slope makes it accessible to most hilltop residences before the final extension of House 18c. To the south, a new midden, Midden 71, forms just north of Midden 77, outside the southern end of the densely occupied House 20. Its appearance at this time is seen in the high ratio of UNH to DNH vessels, 79 percent to 16 percent. With a small proportion of Neutral ceramics, 4.3 percent, 426 the profile of this sample closely matches that of Midden 66, a relationship portrayed in the network analysis. This is a period of high-density occupation throughout the Keffer village. Map 6 The last longhouse added to the Keffer village is House 19 (Figure I-18, Tables I-12 and I-13), as suggested by network graphing. The late date of this occupation is indicated by the very high proportion of undecorated neck Huron vessels, 92 percent, and the low proportion of decorated neck vessels, 2 percent, along with its position, removed from the main community and at the extreme edge of relatively flat ground. House 19 was built outside of the original walls of Palisade 2, where it appears to have been squeezed in beside Middens 71 and 77. According to excavation records, a deposit located at the western edge of House 19 was originally assigned the label Midden 76. Subsequently it was determined this area had been subject to a large-scale disturbance, perhaps a tree fall. This resulted in the cancellation of the midden designation. Ceramics from this area may have originated with House 19, but they have not been included in the House 19 sample due to the uncertainty of their origin. Figure I-18. Stage 5 and Map 6. 427 Table I-12. Settlement Changes in Map 6. Palisade in Use Houses Added or Expanded Houses Removed Middens in Use 2a 19 Table I-13. Proportion of Local Tradition Ceramics in Map 6. SecondLevel Times Built Locus Stage Sample Cluster Size UNH DNH Niagara High & Neutral Collar Ripley H19 D 6.3 2.1 5 1 50–99 91.7 0 0 To the north of House 19, the second extension and rebuilding of House 11c may have occurred during this period of continued population expansion. Neither network nor settlement analysis can confirm this event. The increase in the longhouse length to a final total of 62 m may have resulted in a situation where, once again, pathways across this section of the village may have been difficult to negotiate, with the northern extension of House 11c located very close to the side wall of House 4. At some point in the period, the lightly occupied House 6, also located in close proximity to the House 4 western side wall, was removed. Its short occupation period is supported by a low wall post density, of 4.3 ppm, and the presence of the House 7b extension and hearth overtop of House 6’s western end wall, which may have occurred at this time. Map 7 The next significant event in the village settlement is marked by the demolition of a small section of Palisade 2 to the north of the newly constructed House 19 and its rebuilding several metres to its south in order to bring the House 19 inhabitants into the interior of the village (Figure I-19, Table I-14). It is possible that the extension of Houses 11c may have occurred at this point, but the relatively high interior feature density of House 7b and evidence of its rebuilding or renovation suggest that it is longer lived than House 11c. Also happening near the end of the village life is the expansion of Midden 74 over the southwest wall of House 12. The final growth and rebuilding of House 18c over its northern end suggest that House 12 may have been 428 removed. The late timing of this longhouse removal and the rebuilding of House 18c is implied by the very short-term occupation of this extension to House 18. House 18c displays very few internal features and light construction in terms of visible wall post density. In fact, the only part of House 18 with any appreciable interior feature presence is the first occupation, of the north end of structure, labelled House 18a. Figure I-19. Stage 5 and Map 7. Table I-14. Settlement Changes in Map 7. Palisade Houses Added or in Use Expanded Houses Removed Middens in Use 2a 6, 10, 12 7b, 11c,18c 429 Map 8 The final occupation of the Keffer village visible archaeologically sees the continuing removal of longhouses, following the earlier removal of Houses 6 and 12 (Figure I-20, Tables I-15 and I16). This process is seen in the late deposition and expansion of village middens. The appearance of Midden 58, which hosts an unreliable and tiny vessel sample, in the central portion of House 10’s occupation area confirms the removal of this longhouse during the village occupation. The House 7b extension, and perhaps the entire longhouse, was also removed while substantial habitation remained in the area. Midden 54, which overlies it and House 6, has a quite sizable sample size, with 127 vessels. Network analysis placed Midden 54 somewhat earlier. The sample profile, with 18.5 percent DNH and 74.8 percent UNH, supports this placement, but its position over the extended House 7b indicates a later date for this event. The ceramic profile of Midden 54 is closely matched by that of House 1, a long-lived residence, suggesting that this midden likely contains refuse from the removal or cleaning of other houses in the area. In the southern portion of the village, the expansion of Midden 77 over top of the eastern end of House 19, as reported in excavation notes, signals the removal of this short-lived residence as well. Figure I-20. Stage 5 and Map 8. 430 Table I-15. Settlement Changes in Map 8. Palisade Houses Added or in Use Expanded Houses Removed Middens in Use 2b 7a/b, 19 54, 58; expansion of 74, 77 Table I-16. Proportion of Local Tradition Ceramics in Map 8. Locus M54 SecondTimes Level Built Stage 5 Sample Cluster Size UNH DNH Niagara High & Neutral Collar Ripley 100– 199 18.5 5.9 B 74.8 0 0.8 431 Appendix J: Appendix to Chapter 7: Social Network Analysis of Keffer Site Social Relations and Communities of Practice Social Network Analysis of Keffer House and Midden Local Tradition Ceramic Samples Confirmation that neither sample size nor geographic distance are major factors in the site-level network topology supports further exploration of Keffer houses and midden local tradition ceramics via a network approach. Here, the strength and direction of social ties as displayed in network graphs are examined, with the ultimate goal of revealing communities of ceramic practice throughout the village. In this work, each node represents ceramic praxis of a single household or, in the case of middens, neighbouring households. The strongest ties among these practices are seen in the BR 190, Over 29 Vessels, graph (Figure J-1). This component represents the basic backbone structure, that is, the strongest inter-node connections, of all local tradition ceramic relations at Keffer. Each subsequent graph builds on the foundation of the highest-level relationships demonstrated here. Onto this robust graphic framework, nodes characterized by slightly weaker ties are sequentially added. Lower-count nodes are then introduced at uniform, consecutive dissimilarity levels. In this way, the strongest to weakest intra-village relationships emerge in succession. Stronger relationships are acknowledged prior to changes caused by the introduction of lower-level attractive forces. Figure J-1. Local tradition ceramics Over 29 Vessels BR 190 Gower layout. 432 At this level, many of the matrix loci do not appear in the network. Those absent are dominated by the smaller sample sizes, accentuating the significant omission of the large Midden 57, with 522 vessels, from the structure at this level. The network at this cut-off point is divided into two components, with the House 13–House 20 pair, Group 3, clearly dissociated from the larger, main component. Central to this main component is a “small world” network of five spatially distant loci nodes. A small world structure is composed of a cluster of internally linked nodes held together by short paths, or high edge values—in this context, an edge being a line linking two nodes. Nodes of this array type are highly connected to other nodes within the cluster but have few, or no, external ties (Borgatti et al. 2013; Collar et al. 2015:11; de Nooy 2011; Watts and Strogatz 1998). “Enhanced signal propagation speed” characteristic of small worlds spreads the results of contact, such as diseases, more easily than alternate network topologies (topology in this context is defined as “the manner in which constituent parts of a graph are interrelated or arranged” [Oxford Dictionary.com]) (Watts and Strogatz 1998:440). According to Watts and Strogatz, small worlds are typified by homophily (McPherson et al. 2001), in which people with more similar traits tend to cohere closely together therefore interacting more regularly within the small world. The few external interactions usually occur with other small worlds. These outside contacts are made directly and not through intermediaries (Watts and Strogatz 1998). This ceramically defined small world, Group 2, is composed of the late middens, Middens 54 and77 in the Palisade 2–era unified low town; Middens 61 and 80, occupying opposite ends of the high town plateau; and Midden 60, from the north end of the village. To the left of the small world cluster (Group 2), Midden 52, which overlies Palisade 1, forms a bridge to neighbouring Cluster C high town Middens 72 and 74 (Group 1b). These are then linked towards the graph’s periphery with the context that yielded the site’s largest ceramic collection, Midden 65. Appended still farther out from Midden 65 is Midden 62. Forming a bridge in the opposite direction between the small world group to the right and Midden 66 is the smaller Midden 71. The Over 29 Vessels local tradition BR 190 small world is informative in terms of pottery communities of practice. The tightly knit small world body—as seen in the strong ties of Middens 54, 60, 61, 77, and 80—has intense relationships with others within the community, but the cluster links to other parts of the network through one or very few ties, as described by Watts and Strogatz (1998). This implies close temporal and/or social relationships among the producers and/or consumers of the vessels in these loci. As settlement data, in the form of overlapping 433 features, show Middens 54 and 77 to be very late in the village occupation, it is proposed that the strong tie strengths shared with Middens 60, 61, and 80 indicate these deposits to be of comparable age, that is, also dating towards the end of the village lifespan. The extremely strong similarity, 95 percent of local tradition ceramics, among Middens 54, 60, 61, 77, and 80 suggests deposition from inhabitants with extremely similar ceramic tradition backgrounds and presumably the same, or closely related, community of practice, as these are considered temporally late deposits. The split of the network in two divergent directions from the core body illustrates that this unity is not characteristic of all loci (Figure J-2). Midden 71 and its “pendant node” (referring to a node linked to the graph by only one tie), Midden 66, show a clear separation from those of the more highly connected left arm of the network, where Midden 52 forms a bridge between the small world formations of Group 3 and Group 2. The Midden 72– Midden 74 pair, in turn connects Midden 65, by far the largest midden at the site, and the more distant Midden 62 to the core component. Only a small change is seen in the network configuration when the nodes of the Over 19 Vessels samples are integrated (Figures J-2 to J-4). The House 4 node appears in an independent pendant linked by the existing Midden 77 bridge node. Connected by House 4, Houses 7 and 11 emerge as a tight pair of outliers (Group 4). Although most similar to Midden 71, this tripartite group is dissimilar to Midden 66 and all other members of the network at this level. Figure J-2. Local tradition ceramics Over 19 Vessels BR 190 spring layout. 434 Figure J-3. Local tradition ceramics Over 19 Vessels BR 190 Gower layout. Figure J-4. Local tradition ceramics Over 19 Vessels BR 190 MDS layout main component. 435 The meager ceramic collection (6 vessels) of House 10 is highly integrated into the small world group centred on Midden 54 when the Over 1 Vessel samples are merged. The size of this sample renders its inclusion here unsubstantiated, however. This is also true of the articulation of the House 2 node, with 9 vessels, into small world Group 4, composed of Houses 4 and 11 and the newly incorporated House 7 node (25 vessels). Despite these small sample sizes, the projection of the Group 4 pendant away from the network central component reveals the increasingly divergent nature of these local tradition ceramics. Additionally, the disconnection of the Group 3, House 13–House 20 pair, accentuates the uniqueness of these ceramics (Figures J-5 and J-6). Figure J-5. Local tradition ceramics Over 1 Vessel BR 190 Gower layout. 436 Figure J-6. Local tradition ceramics Over 1 Vessel BR 190 MDS layout. In summary, network graphs at the BR 190 similarity level illustrate a tight small world core of Group 2 with solidly linked pendants on its opposing extremities. The addition of low-vesselnumber nodes has little to no effect on the relationships of the two small world components, Groups 2 and 3, of the network. The unique House 13–House 20 pair remains remote and isolated (Figure J-5). The purpose of multidimensional scaling (MDS) is to provide a visual representation of the pattern of proximities among a set of objects. By proximities we mean any symmetric, one-mode matrix of similarities, tie strengths, dissimilarities, distances, etc., among a set of objects. MDS places points (corresponding to our objects) in space such that the distances between the points correspond in a predetermined way to the proximities among objects in the data. The presence of nodes situated in contrasting directions off the main component suggest they possess ceramic samples with characteristics of opposing nature (this is implicit in the topological distance between two nodes and in statistical centrality closeness measures; see Centola 2015:1313 on homophily; see also Collar et al. 2015:14). As similarity measures 437 decrease, the number of ties (or edges, meaning a line linking two nodes) in the network graph greatly increases. Typically, each node acquires numerous additional links affecting the nodes’ degree centrality measures. Although these new connections are at lower BR levels, the increase in tie numbers creates strong attractions in new directions, often overpowering the higher-level similarity connections. Ties appearing with the newly emerging nodes pull existing nodes from their previous positions in the graph. But these visually observable changes within the existing relationships are not relevant. It is only at the highest similarity level at which new nodes and edges appear that the strength of ties is relevant. When the interval is dropped by only 5 points, or 2.5 percent, to BR 185, numerous changes are seen in the Over 29 Vessels graphs (Figures J-7 and J-8). Eight new loci nodes—Houses 1 and 12 and Middens 56, 59, and 73, as well as the outlying House 13–House 20 pair (Group 3)— become attached in numerous directions onto the periphery of the network. The closest association is that of Midden 73, a southern hilltop midden, with the northern Midden 60 node, one removed from each of the existing Groups 1, 2, or 3. House 1 and Midden 59, on the far side of the graph, share very similar ties with the Group 2 cluster but are not tied to each other. At this level, the House 13–House 20 pair, Group 3, joins the main network component, linking with the newly centralized Midden 62. House 12 connects tightly with House 20, more remotely with House 13. All three tie to Group 1 and Midden 62. More distantly, Midden 56 ties to Group 1 and Midden 62 and Midden 65. The small world core of Group 2 remains solidly bound and central, although some of its components, Middens 52, 72, and 74, now cluster together more tightly while being slightly distanced from the others. Although several new nodes materialize at this level on the periphery of the graph, it is the appearance of the pendant nodes of Middens 57, 55, and 51, Group 5, projecting from the solitary node of Midden 65 that is most striking. Node placement on this divergent path (where nodes are only traversed once) (Borgatti et al. 2013) denotes sequentially decreasing similarity, signifying more distinctive ceramic practices with distance from the core, most acutely seen in Midden 51 located in the northeast corner of Cluster A. The emergence of this distinctive feature is the reason for the inclusion of the network graph at BR 185 (Figures J-7 and J-8). This formation is obscured at both the higher BR 190 and the lower BR 180 levels. With the connection of the numerous disparate luminal nodes, the overall patterning is more dispersed, and the graph no longer possesses a consolidated central core. At 438 the same time two main areas of densely connected nodes, Groups 2 and 3, become prominent. In the Over 19 Vessels there are no new nodes occurring at BR 185. Figure J-7. Local tradition ceramics Over 29 Vessels BR 185 Gower layout showing newly integrated nodes in dark blue. 439 Figure J-8. Local tradition ceramics Over 1 Vessel BR 185 Gower layout showing newly integrated nodes in light blue. Although the underlying dispersed pattern remains in the Over 1 vessel BR 185 network (Figure J-8, the nodes of Houses 4, 7, and 11, previously connected in the higher-level Over 1 vessel BR 190 network, provide a foundation for the attachment of House 2 and, ultimately, House 19, to the graph at BR 185. The formation of this pendant, Group 4, mirrors the progressive dissimilarity seen in Group 5. In this network, the newly integrated Midden 73 and House 8 are drawn more closely into the core of the structure than seen only in the Over 29 Vessels graph at this level. Of the Over 29 Vessels nodes, only Midden 75 remains isolated, or unattached, at BR 180 (90 percent similarity), emphasizing the variance between it and the adjoined Midden 80, a node centrally located in the Group 2 small world (Figures J-9 and J-10). Nodes previously connected 440 to the network at higher levels from the Over 1 vessel matrix, Cluster D Houses 11 and 19; now connect in a pendant from House 1 and Middens 66 and 71. House 9 is joined at a distance from the main Group 2 small world through Midden 61 and ties in its periphery with Midden 66. Connecting at the furthest point of the network from House 19, Midden 53 extends the Group 5 pendant a greater distance from the core. Middens 65 and 62, which were formerly peripheral, become more integrated as new connections are formed with House 12 and Midden 73. Figure J-9. Local tradition ceramics Over 29 Vessels BR 180 Gower layout. 441 Figure J-10. Local tradition ceramics Over 29 Vessels BR 180 MDS layout. When the similarity matrix is expanded to include those nodes in the Over 19 Vessels category (Figures J-11 and J-12), no new loci appear. The House 7 node remains hidden, as it was in the BR 190 diagram, behind that of House 11, to which it is linked above this level of similarity. While there is some minor alteration in the structure of the overall graph, it retains the basic configuration it acquired with the Over 29 Vessels samples. 442 Figure J-11. Local tradition ceramics Over 19 Vessels BR 180 Gower layout. Figure J-12. Local tradition ceramics Over 19 Vessels BR 180 MDS layout. As with the Over 19 Vessels matrix, no new nodes are introduced at the BR 180 level in the Over 1 Vessel network visualization (Figures J-13 and J-14). Additional edges in the network overpower higher-value ties resulting in a rearrangement of the central nodes and the interior of the network becomes less highly clustered. 443 At BR 180, several nodes remain outliers to the networks. These include the Over 29 Vessels nodes; Midden 75, a relatively small midden with 31 vessels; the Over 19 Vessels nodes of House 15 and Middens 68 and 78; and the Over 1 Vessel nodes of Houses 3 and 6 and Midden 67. Figure J-13. Local tradition ceramics Over 1 Vessel BR 180 Gower layout. 444 Figure J-14. Local tradition ceramics Over 1 Vessel BR 180 MDS layout. Ultimately, Midden 75 connects as a pendant to Midden 57 at BR 170 (Figures J-15 and J-16), having little effect on the network in general. 445 Figure J-15. Local tradition ceramics Over 29 Vessels BR 170 Gower layout. Figure J-16. Local tradition ceramics Over 29 Vessels BR 170 MDS layout. As with other terminal pendant nodes, House 15’s appearance at BR 170, Over 19 Vessels (Figures J-17 and J-18), off the outlying Midden 53 node, marks its unique ceramic characteristics even as it joins Group 5, a very loosely tied assemblage. A third pendant is 446 formed to the bottom right of Midden 51 as Midden 68 bridges the gap between it and the new Midden 78, diametrically opposed to House 19. Figure J-17. Local tradition ceramics Over 19 Vessels BR 170 Gower layout. 447 Figure J-18. Local tradition ceramics Over 19 Vessels BR 170 MDS layout. At BR 170, three very small sample nodes join the Over 1 Vessel matrix graph (Figures J-19 and J-20). House 3 appends itself to the external Midden 75. Slightly closer to the core, Midden 67 links with isolated Midden 57, while the House 6 node ties closely with House 11 of Group 4. 448 Figure J-19. Local tradition ceramics Over 1 Vessel BR 170 Gower layout. Figure J-20. Local tradition ceramics Over 1 Vessel BR 170 MDS layout. 449 Note that lower levels of similarity, in this case BR 170, and larger sample sizes, as seen here in the Over 1 Vessel matrix, naturally yield infinitely more ties than the higher (BR 180 and 190) levels and smaller (Over 19 Vessels and Over 29 Vessels) matrices, yet the resulting networks share similar overall form. Modifications to the network structures which result from the introduction of these lower thresholds, particularly lower similarity measures, are not highly significant but are useful as guides directing further investigation of the noted patterning. Summary of Social Network Analysis The dynamic structure of the graphs and the changing patterns of clusters and outliers apparent at different similarity levels and samples sizes inform on the nature of the ceramic relationships at Keffer. The Over 29 Vessels BR 190 (Figure J-1) network provides the backbone structure of the closest relationships among local tradition ceramic samples. This framework is composed of a one small world formation, Group 2, composed of several highly similar ceramic midden collections from across all settlement clusters of the village. Members of this small world appear to date to the middle of the village occupation, although they have close ties to both earlier and later deposits. A second minor small world, Group 3, includes primarily deposits of the high town. As tie levels are lowered and new constituents are added, shifting allegiances and new bonds emerge. The network fills in and relationships become more numerous and more complex across the village as a whole. Outliers encircling the core relationships materialize reflecting diverse influences on local traditions. At the more inclusive end of the spectrum, the Over 1 Vessel BR 170 network, new ties hint at weaker relationships, providing supplementary information to be integrated into analysis after higher-level connections are taken into account. This in-depth analysis of graph topology at similarity levels between BR 180 and 190 has produced a revised series of groups, believed to represent the local tradition ceramic communities of practice present throughout the village occupation. These groups, as seen below (Figures J-21 and J-22, Table J-1) represent social ties. These divisions are not fundamentally chronological. The presence of possible temporal patterns may reflect social ties across time within the village. 450 Figure J-21. Local tradition ceramics Over 1 Vessel BR 185 Gower layout. Temporal patterning, from earlier to later, is represented by the physical placement in the figure, left to right. Colours indicate community of practice group. 451 Figure J-22. Local tradition ceramics Over 1 Vessel BR 171 Gower layout. 452 Table J-1. Houses and Middens Associated with Each Local Tradition Community of Practice. Local Tradition Community of Practice Houses Middens 1 13, 20 2 1, 9, 10 54, 61, 77, 80, 52, 60, 71, 66, 59,63 3 8, 12 72, 74, 73, 65, 56, 62 4 2, 4, 6, 7, 11, 19 5 3, 15 51, 53, 55, 57, 67, 68, 75, 78 Note: Boldface indicates initial high-level core nodes of communities. Order indicates approximate strength of ties. The five communities of practice (CoP) are numbered according to the relative strength of their highest ties. That is, the members of community of practice 1 (CoP 1) are more similar than those of CoP 2, and so on. The strongest ties within each group are indicated in boldface type and are located within the list next to those with which the group shares this high-level tie. In the last group, CoP 5, lower similarity level relations produce a less cohesive and more loosely tied community. Closer examination of these communities of practice occurs in Chapter 8.