Untitled - Caio - Index of

Michael Meier-BruggerIndo-European LinguisticsIn cooperation with Matthias Fritz and Manfred MayrhoferTranslated by Charles GertmenianWaiter de Gruyter . Berlin' New York2003

Michael MeierBriigger holds the Chair in Comparative and IndoEuropean Linguisticsat the Free University of Berlin, Germany.Matthias Frit: is Research Fellow in the Department of Comparative and Indo-European Linguistics at the Free University of Berlin, Germany.Man/red Mayrho/er is Emeritus Professor of Comparative and Indo-European Linguistics at the University of Vienna, Austria.Contents3oJ)Univerzitna kniznicaV BratislaveI I 11 111111111111111111111111111111111111111*1800256290*Contents ............................... ................................... ................................ VPreface to the Seventh German Edition (2000) ...................................... IX,Preface to the Eighth German Edition (2002) ...................................... XIIlPreface to the English Translation of the Eighth German Edition ........... XVTerminology, Symbols, Abbreviations ............................................... XVII1. Terminology ........................................................................... XVII2. Symbols, Writing Conventions, Citing Conventions ............... XVIII3. Abbreviations .............................. ........... ......... ............ . ........... XXII. Introduction ..............................•........... .............. ................................. 1A. The Field and its Study .................................................... ....... ........ 1@ Printed on acidfree paper which falls within the guidelinesor the ANSI to ensure permanence and durability,ISBN 3-1 1-017433-2Bibliographic information published by Die Deutsche BibfiolhekDie Deulsche Bibliothek lists this publication in the Deutsche Nationalbibliografic;detailed bibliographic data is available on the Internet at .© Copyright 2003 by Waiter de Gruyter GmbH & Co. KG, D-I0785 BerlinAll rights reserved, including those of translation into foreign languages. No part of thisbook may be reproduced or transmitted in any form or by any means, electronic ormechanical, including photocopy, recording, or any information storage and retrievalsystem, without permission in writing from the publisher.Cover design: Hansbernd Lindemann, BerlinPrinting and Binding: WBDruck, Rieden am ForggenscePrinted in GermanyUnlver-.,Itna kn,.nLBATIL AVB. Indo-European Linguistics in the Age of the PC and the Internet .... 6C. A Word on the History ofIndo-European Linguistics ..................... 9D. Overview of the Indo-European Languages and their Sources ...... 171. General Information ........ ............................................................ 172. The Individual lE Language Families and their Sources . ........ ...... 18E. The Reconstruction of Pro to-In do-European ......... ....................... 411. Examples of Reconstruction ....................................................... 412. Fundamentals of Reconstruction ................................................. 563. The Time, Place, and Culture of the Proto-Indo-Europeans ........ 63n. Proto-Indo-European Phonology ...................................................... 71A. General Information ..................................................................... 71.. ,., # •• _.

xPrefacePrefaceXIVolume 1. Introduction and Phonology; Volume H. Morphology. " Thetwo volumes of the fifth edition (1966 and 1969) were unchanged. In itssixth edition the work remained unchanged, though issued in a single volume11.Krahe's volume is now more than fifty years old. Although the sixthedition has the external appearance of a new volume of the SammlungGoschen, the content remains a child of the third edition of the I 960s.Typical of the state of research at that time is Krahe's comment onlaryngeals (vol. I p. 101): "A number of Indo-European scholars have recentlyrepresented the so-called laryngeal theory, according to which, inaddition to the phonemes included here, the basic Indo-European languageincluded certain laryngeals (glottals and schwas), which are not taken intoaccount in this short work. The author is of the opinion that in a workthat is primarily conceived for students and as an introduction to a field,only the most proven research should be presented. The laryngeal theoryhas been affirmed neither in its methodology nor in its technique." Thelast sentence is decidedly incorrect: The common expression 'laryngealtheory' is historically conditioned and can be misleading, but this shouldnot lend credence to the idea that the laryngeals are only vague and theoretical.Today laryngeals are part of the attested body of phonemes of theIndo-European language.12 One often finds in Krahe's book informationand reconstructions, about which we either know more, or have acquiredmore adequate insights today.I n the middle of the 1980s, the publisher retained Heiner Eichner andManfred Mayrhofer to write a new Indo-European Linguistics for theGoschen series. By 1985 Dr. Mayrhofer had written the section on phonology;Dr. Eichner was to provide the introduction and the section onmorphology. A 1988 publication date was announced, but the volume wasnever published. A number of difficulties hindered publication, particularlyon the side of Dr. Eichner. Dr. Eichner and his manuscript, which in themeantime had grown, were faced with space and time constraints, includingthat difficulty, known to every author, of putting one's own name onsomething less than completely perfect, which hindered him finally, frombringing the project to completion., 1962 (SG 59) 110 pages.10 1963 (SG 64) 100 pages." 1985 (SG 2227).12see below L 3 14 fT.In December 1993, the publisher offered to me the role of Or. Eichner.After exchanges with all concerned parties, a contract was signed withFebruary 1996 as the agreed upon date of completion. The offer of a positionto me at the Freie University in Berlin rendered the bold timeframeimpossible. The change from Hamburg to Berlin brought everything withit except the necessary leisure to produce a manuscript. l am grateful toBrigitte SchOning, who, while showing kind understanding of my timeconstraints, was able to make sure the actual deadline of the publisher wasmet. Although I would like very much to have expanded and embellishedmy manuscript, I must now give it up to those who will publish it.By fall 1998 I had covered only half the intended material but was informedthat I had considerably exceeded the page allowance for oneGoschen volume. Thus, a currently expanding series, the "de GruyterStudienbUcher," was attractive. They offer greater spatial freedom. Further,they do not demand a particular format of proof. Hence, the manuscript,prepared with the help of Microsoft Word 7.0 for Windows 95, canbe used directly for publication. No more galleys. On the part of the publisher,Ingeborg Klak took care that the typography does not come off tooold-fashioned.While Dr. Mayrhofer and Or. Eichner, following the precedent set byKrahe, intended on onJy covering phonology and morphology, now syntaxand vocabulary are receiving the recognition they deserve.The explanations in the section on morphology are based on the textwritten by M. Mayrhofer in 1985, as Dr. Mayrhofer had given me completefreedom to determine the arrangement and the organization of thedefinitive version. I take sole responsibility for the present version of thesection on phonology. Happily, l was able to have Matthias Fritz write thesection on syntax. Further, he contributed to the part of the introductionOn the history of Indo-European linguistics as well as the overview ofIndo-European languages and their sources.M. Fritz, M. Mayrhofer, Elisabeth Rieken, Bernfried Schlerath andAntje Schwinger read individual excerpts with a critical eye and sharedteir criticisms with me. Verorrika Rittner and Thorsteinn Hjaltason reVIewed individual texts electronically. Mr. Hjaltason and Ms. Schwingerh lped me with certain data processing problems, Mr. Hjaltason especiallyWith the creation of particular symbols, and Ms. Schwinger with the layout.A sincere thanks to you all!

XIIPrefacePrefaceXIllAs is appropriate for a volume in the Sammlung Goschen and a "deGruyter Studienbuch," the book should offer an introduction to IndoEuropean linguistics. The textbook should provide an informative andcomprehensive treatment of the issues and areas of focus from a contemporaryperspective. I allow myself a special style in presenting both informationon the field in general and on the course of study in the introduction.In the main text, on the other hand, I present inconclusive material ina usable way, to encourage possibilities of more in-depth study. Thetreatment proceeds according to good science and good conscience.Completeness is nowhere attempted. To keep this in the reader's mind,the chapter titles begin often with "on." Most of the examples in the sectionsmorphology, syntax and vocabulary come from Latin, Ancient Greekand Indian (Indo-Aryan). In the section on phonology, the examples areintentionally more numerous and are taken from the entire realm of IndoEuropean languages, which include Hittite, Germanic and Slavic, in additionto the above-mentioned three. The index appended at the end shouldgive support to the reader and facilitate his or her access to information.The state of research reflected in the text is that of September 1999. Lastadditions will be inserted after this deadline just before the definitive layoutis made in December.The current textbook is not without competition. Still good is Rix'sHistorische Grammatik des Griechischen (1976). Starting from Greek,Rix presents pertinent information on all aspects of Proto-Indo-Europeanphonology and morphology. But the treatment is done without any referenceto discussion in the field. Recent works worthy of mention are Szemerenyi,Einfuhrung, fourth edition (1990), Beekes, Introduction, (1995),Schmitt-Brandt, Indogermanislik, (1998) and in a certain sense alsoMeiser, Historische Laut- und Formenlehre der lateinischen Sprache,(1998) (compare pp. 27-46, the chapter "Grundziige der urindogermanischenGrammatik"). All four titles have strengths and weaknesses, and soshall it be with mine. The comprehensive bibliographical information inSzemerenyi is a treasure trove, but the skepticism regarding laryngeals isbothersome. Beekes' book is illustrative and very readable, but in phonologyand morphology he leans too much on views shared only by himselfand F. Kortlandt. Schmitt-Brandt's work is in a praiseworthy fashionaimed at beginners and concentrates especially in morphology on a broadargumentation that encourages creativity. However, his book cannot berecommended because he leads the reader into unsignaled idiosyncrasiesthat stray from the communis opinio. Meiser, like Rix, treats only phonol-ogy and morphology. He grounds himself competently, but of necessitybriefly (too briefly), on the realities of the Proto-Indo-European language.I should not speak of typographical errors. He who sits in a glass house iswell advised to throw no stones.I am thankful for any feedback and request that it be sent to the Seminarfur Vergleichende und Indogermanische Sprachwissenschaji at theFreie University in Berlin (FU), Fabeckstralle 7, D-14195 Berlin-Dahlem.Te!': 030-838-55028; Fax: 030-838-54207; e-mail: drmeier@ zedat.fuberlin.de;intemet site: http//www.fu-berlin.de/indogermanistik.I dare try to create a rubric on our web site: "De Gruyter StudienbuchIndogermanische Sprachwissenschaft: Addenda and Corrigenda." I'veresolved, starting with the appearance of the book, on the first weekday ofeach month, to note addenda and corrigenda as I learn about them.Berlin-Dahlem, September 15, 1999Michael Meier-BriiggerPreface to the Eighth German Edition (2002)The occasion to publish the present eighth edition presented itself unexpectedlysoon after the publication of the seventh edition.I know that I am joined by M. Fritz and M. Mayrhofer in my gratitudefor he responses of goodwill and for the positive reception of the previouseditIon, which appeared in March 2000. We are pleased that the text hasfound its place as an introduction to the current status of Indo-Europeanlinguistic research. It is encouraging that this text and E. Tichy's depictionoffirndamentals of Indo-European linguistics (Bremen 2001) complementeach other wonderfully without any prior arrangement having taken place.E. Tlchy sketches a concise treatment of the main characteristics, while the

XIVPrefacePrefaceXVpresent text fills them out with details. T can easily imagine that the interestedreader would first consult Tichy's treatment, and then, given thepage references from that text, would reach for the present text.Happily, the requested feedback was furnished with intensity.On theother hand, my planned internet rubric, "Addenda and Corrigenda" did notcome to be. Among the reasons for this are juridical problems that preventone from presenting special characters that were created at the Seminar onthe Internet.Other reasons include the ever-growing burden of tasks inteaching and administration: Constant reductions and transformations duethe scarcity of public funds claim most of our energies.I am happy about the positive feedback and thankful for having beensent entire lists of comments, inquiries, and tips about typographical errorsand unclear points. The lists complemented each other wonderfully: Eachperson sees something different; no one sees everything.Imperfections,and the corresponding need for correction, were found mostly where generallyaccepted opinion does not apply:Hence, in the assessment of therole of the laryngeal h3 in the Hittite language (see below L 334), or in theevaluation of the stative diathesis and its connection to perfect and mediumdiatheses (see below F 211), or in the case of genders, where onemust likely draw the conclusion that the formation of three-way masculine/feminine/neutertook place only after the splitting off of Anatolian,thus in a phase that followed the phase of the original Indo-European language(see below F303), or in the earlier common assessment of the lE*gWO!!_ 'steer' as an hysterodynamically declined noun. In contrast to thisrepresentation, *gWO!!_ must be categorized as primarily acrostatic with astrong stem *gWO!!_ and a weak stem *gwi!!!_. The adjustment to amphidynamicallydeclined root nouns such as lE *dje!!- 'sky, god of the sky,day' with the strong stem *dje!!- and the weak stem *di!!-' is secondary.The strong nominative singular *gWO!!-S remains, but the weak genitivesingular *gWe!!_s is replaced by the new weak genitive singular *gWo!!-es(see below F 318 § 6). The eighth edition introduces modified content insuch cases.The numbering of paragraphs from the seventh edition wasconserved whenever possible.What I could not yet accomplish in theeighth edition includes the addition of a complete index of vocabulary, andthe expansion of the rather brief treatment of vocabulary.Altogether, as many as possible of the typographical errors and oversightswere corrected.clearly marked as such.Imperfections were either eliminated, or at leastAs a result of such work, an eighth edition appearedwhich I can stand by with a good conscience.The responsibilityfor all remaining mistakes lies with me. M. Fritz kindly agreed to correctthe work, and thus took responsibility for the section on syntax.My particular thanks for criticism goes to: Augustinus Bal (Amsterdam),Irene Balles (Jena), Wolfgang Beck (Wiirzburg), Joachim Becker(Giittingen), Martin Braun (Vienna), Emilio Crespo (Madrid), RobertoGusmani (Udine), Heinrich Hettrich (Wiirzburg), Katharina Kupfer(Freiburg), Christoph Landolt (Ziirich), Gyula Mayer (Budapest), H. CraigMelchert (Chapell Hill), Peter-Arnold Murnm (Munich), Sergio Neri (Salzburg),Oswald Panagl (Salzburg), Bernfried Schlerath (Berlin), DietherSchiirr (Griindau) and Stefan Schurnacher (Halle and Freiburg). KlausJiirgen Grundner help with the corrections on this edition as well as withthe seventh.Ms. Susanne Rade supported me kindly on the part of thepublisher and guided me, such that the eighth edition could appear in thewinter semester of 2002.The addresses are the same as on page XIII.Berlin-Dahlern, September 15, 2001Michael Meier-BriiggerPreface to the English Translation of the EighthGerman EditionIn recent years, the English language has attained a position similar tothat enjoyed by Latin in the Middle Ages. While Indo-European linguisticshas traditionally had a strong basis in the German-speaking world, the fieldis happily becoming ever more internationally and intercontinentally oriented.It is thus not surprising that I have received several inquiries regardingthe possibility of an English translation of the original Germanlanguageedition.

xxTerminology, Symbols, AbbreviationsTerminology. Symbols, AbbreviationsXXIb. Writing conventionsI) For notation ofIndo-European, see below L 100f.2) In the cases of individual Indo-European languages, the accepted modesof writing are respected. I would like to call attention to a couple of details:Latin is always to be taken as /Id, and in contrast, as /kw/,without regard for whether it represents /Iq" or comes from /kw/, see belowE 506 § 3. Please note as well that a word, such as shouldbe read as /majjus/ and a as /kon-jikijo/: ---7 Leurnann LLFL1977 p. 127f.In the examples from Mycenaean Greek, the instance of a word is givenas a rule fIrst, just as it is represented in Mycenaean Greek Linear-B. Theprohable phonological interpretation is then given equally in Latin script,but not in Greek script.(GraphemicaHy indicated phonetic transitionalsounds are taken into account as well, and often a modem morphologicalsegmenting follows.) Compare i-je-ro-wo-ko i.e. hijero-I}orgos 'priest ' .For the sake of simplicity, the symbols < > and / / are not written here. -Where an aspiration is probable, it is marked with an h, such as in the precedmgexample. - In the interpretation of the z-series, either a id, a gj , a11 or a d1 will be inserted according to etymological origin, for example:lo-pe-za I.e. lorped1a 'table' < *(kW)tr-ped-ja '(thing) having four-legs'.The sources of quoted forms are not cited, but they can easily be found:-m Aura Jorro DMic. I 1985 II 1993. - Examples from Greek dialects arerepresented without accents.The model for Ancient Indian (Vedic) is Mayrhofer EW Aia. As iscommon practice in lE linguistics, Vedic and Old Indian nominal forms arecited as a rule as sterns without an ending (thus, for example, the Vedictivi- 'sheep'). Older manuals often cite the nominative singular in thesandhi form with -& (so-called Visarga), cf. a reference such as Latin ovis'sheep' = Vedic tivi&. For more information see below L 309 § 3. - Thecited form for Vedic and Old Indian is the third person singular (seebhtirati 'carries'), also on the subject the full grade verhal root (e.g. Vedicbhar- 'carry'). Accents are used in fInite verb forms only when they aretextually attestable.In the case of Avestan, I follow the example of ---7 Hoffinann / ForssmanAvestische Laut- und Flexionslehre 1996.In Anatolian (Hittite) b should be written in transliteration from cuneiformas well as in the transcription, but s should be written in the transliteration,and not in the case of the transcription, compare, for example,pa-ab-uu-e-na-a.s i.e. pabbl}enas. The z that was introduced in the transliterationis given as r. Further, apparently stressed syllable marks such aslak in ttik-na-a-as do not refer to the place of the accent. In accordancewith the conventions of Ancient Oriental Studies, the mark a is merelysuggests that in addition to the usual form tak (= lak number one) a secondform of symbol (= tak number two) is in use.On Gothic: ---7 Binnig GOlisch 1999.c. Citing ConventionsNote that methods of citing are not handled in the same way in all ancientIndo-European languages, compare for example the verbal forms ofLatin and Greek, in which the fust person singular is cited, whereas in Vedicthe third person singular is the traditionally cited form.In Mayrhofer KEW A, the verb forms are listed according to the thirdperson singular form, whereas in Mayrhofer EW Aia they are indexed bytheir full-grade verbal roots, compare KEWA p. 1Il 562f. the entry svarati'gives off a sound, sounds, sings', with EW Aia II p. 792f., the correspondingentry under SVAR 'giving off a sound, sounding, singing'. Theciting convention of nouns as well varies from the KEWA to the EW Aia'compare in the KEWA svarga&, while the EW Aia lists svarga-.3. AbbreviationsCertain of the most common abbreviations have been omitted from theollowing list. As a rule, the abbreviations that are used are long establishedor clear in their context.N.B.: Abbreviations of periodicals are found in the bibliography.§ paragraph Arm. Armenianabl.AV Atharva Vedaace. accusative Avest. Avestanact.adj.ad I.adv.aor.ablativeactive voiceadjectivead locumadverbaoristcol.coil.dat.encl.etc.columncollectivedativeencliticet cetera

2 IntroductionSwitzerland. Within the German-speaking realm there are also fine differencesthat can have to do with local tradition or teachers.In its informationon curricula in Indo-European linguistics, the fo llowing account concentrateson the situation in Germany. Additionally, the local circumstancesin Berlin may occasionally shine through.The field of Indo-European linguistics is represented at every majoruniversity.The "Indogermanische Gesellschaft" represents the interestsofthe field as a professional association, see below E 20 I § 2. It organizesa conference every four years, see the bibliography under the heading'Fachtagung.' In addition, there are individual colloquia, see the bibliographyunder the heading 'Kolloquium.'E 101. There is a whole series of positions taken regarding the natureand aim of Indo-European linguistics: -7 ArbeitsausschuB der IndogermanischenGesellschaft in Kratylos 13 1968 p. 222f (= Linguistische Berichte9 1970 p. 78-80); Szemenlnyi EinJuhrung 1990 p. 32-36; R. LiihrIndogermanistik am Wendepunkt? "Thesen zur zukunftsorientierten Ausrichtungeiner Disziplin" in the GiefJener Universitiitsbliitter 25 1992 p.77-90; G. Neurnann "Zur Inferdisziplinaritiit der Geisteswissenschaften.Ein Beispiel: Die Vergleichende Spraclrwissenschaft" in the GiefJenerUniversitiitsbliitter 29 1996 p. 61-67; G. E. Dunkel "Zurcher Indogermanistik:t:wischen Vergangenheit and Zukun f' in the InJormafionsblattder Universitiit Zurich 6 1990 p. 10-12.In this context it is further worth taking a look at the definitions of thefield that appear in increasing numbers on the web sites of departments andinstitutes for Indo-European linguistics. Examples from Cologne, Munichand Wiirzburg are printed here as examples. As of March 1999 they wereobtainable under their corresponding addresses and are still worth readingtoday.As is characteristic and typical of today's fast-moving world, thetexts from 1999 are, as a rule, no longer the same as texts of 2001. I encouragethe interested reader to visit these web sites and consult the mostcurrent information.E 102. The Instituf Jur Spraclrwissenschaft at the Universitiit zu Kiilln(-7 Link on our web site [see above E 100] see the heading for IndoEuropean linguistics in Europe) offered the fo llowing definition in March1999:"The fo cus of historical-comparitive linguistics are languages, whichthrough systematic similarities in declension, word fo rmation, syntax,The Field and its Study 3and vocabulary are recognizable as 'related.' The comparison of theselanguages yields information about the history, pre-history, and originsand development of individual traits of each of them that could otherwisenever be obtained. Historical-comparitive linguistics concerns itselfempirically and theoretically both with processes of linguistic history,such as the splitting of originally unified languages into differentdescendent languages, and with language-inherent and extra-lingual circumstancesfor linguistic development. In addition, it investigates whathistorical linguistic description is capable of saying about culturaltransformation. - Historical-comparative linguistics has developedmost fully in the area of the so-called Indo-European languages, whichinclude great European and Asian languages of cultural importance (Indian,Iranian, Greek, Slavic, Latin, Germanic and Celtic languages), andsince its founding at the beginning of the 19t h century it devotes thegreater part of its interest to these languages."E 103. The web site of the InstitutJur Allgemeine and IndogermanischeSprachwissenschaft of the Ludwig-Maximilians Universitiit in Munich (-7Link on our web site [see above E 100] under the heading Indo-Europeanlinguistics in Europe) offered in March 1999 the fo llowing description ofthe field:"Indo-European linguistics is an empirical-historical, theoreticallyoriented discipline.The goal of its research is manifold: On the onehand, through comparisons of individual Indo-European languages(Particularly their earliest available stages, such as Old High German,Vedic, Hittite), it aims to gain knowledge about language and culture ofthe common predecessor of these languages, namely of Proto-IndoEuropean. Meanwhile, the grammatical system of this mother languageand its various changes after the moment of splitting off of theindividual languages is in the foreground.On the other hand, IndoEuropean linguistics contributes to the better understanding of historicalphenomena in the language and culture of all Indo-European peoplesthrough knowledge of established rules that is acquired in the abovementionedprocesses.As a connector of philologies, Indo-Europeanhnguistics includes the cultural realms from northern Europe, theMedtterranean of classical antiquity and the ancient and modem Orient,reaching all the way to India and central Asia.The most importantlanguage groups, or rather individual languages are Old Indian, andGreek, as well as Old Iranian, Latin, Germanic, Celtic, Slavic, Baltic,

4 IntroductionHittite, Armenian, Tocharian, and Albanian. - In the context of itsoutlined research objective, the field is additionally concerned withgeneral linguistic problems, such as those of linguistic change and therelationship between historical and typological linguistic comparison.- By its nature, Indo-European linguistics is interdisciplinary, withnatural connections with neighboring linguistic, philological, andcultural-historical fields (for example, pre- and early history)."E 104. On the web site of the Lehrstuhl fur Vergleichende Sprachwissenschaftof lulius-Maximilians Universitat in Wiirzburg ( Link on ourweb site [see above E 100] under the heading Indo-European linguistics inEurope) the field was defined as of March 1999 as the fo llowing:"Comparative Indo-European linguistics is an empirical-historicaltheoretically based discipline. It has several research goals: It comparesthe individual Indo-European languages (particularly their earliestknown stages of development, fo r example, Latin, Ancient Greek, OldIndian, Gothic, Hittite, among others, but also the more recent stages)and gains, through processes of reconstruction, knowledge aboutgrammar and vocabulary of earlier, non-written stages of language, andof the common predecessor of all these languages, Proto-IndoEuropean. Through the background of comparison furnished hy thelarge number of languages and their diachronic perspectives, compara.tive linguistics contributes to a deeper understanding of grammar andvocabulary of the individual Indo-European languages. - Thoroughresearch on these languages leads both to knowledge about the culturalbackground of a given language community (including history, socialstructure, religion, philosophy, poetry, etc.), to the theoretical study ofgeneral linguistic objects (e.g. the structural organization of languagesystems, phenomena of language development, relationship betweengenetic and typological language comparison.) - Comparative linguis.tics is an interdisciplinarily oriented field and is seen as a connectingmember between the cultural regions of northern Europe over theMediterranean and the ancient Orient through India and central Asia."E 105. There is nothing more exciting and creative than historicalcomparativelinguistics. I am not alone in making this assertion.But first a warning: The way to academic employment in the realm oflinguistics is rocky and full of thorns. Whoever sets out to do this can win ,The Field and its Study 5but can also lose. Those who would like a large amount of money in theiraccount ought to choose another route.The fundamental prerequisite of promising studies in the field is a livelypersonal interest in languages and speech.perspective must also be present.A preference fo r a historicalIn addition to the general maturity expectedat the university level, knowledge of Latin and Greek (preferablylearned in school) are also important. Where today's school curricula havenot provided this option, the necessary knowledge of Latin (to the extentof the so-called kleines Latinum) and Greek can be acquired in the firstyears of university education. Knowledge of English, German, and Frenchis imperative fo r the reading of literature in the field. Spanish, Italian, andRussian are also helpful.Latin, Greek, and Vedic are the pillars of Indo-European linguistics, ifonly because discussion in the field since its beginnings has referred toproblems in terms of these languages.allows independent assessment.Only appropriate language abilityAlong with the study of the three languages mentioned above, it is reccommendedthat one acquire good knowledge of Anatolian Hittite, as wellas an Old Germanic language (such as Gothic, Old High German, or OldSaxon) and Old Church Slavic or Lithuanian.Nothing can replace the reading of primary texts. It is also a personalgain to read, not just partially and according to immediate need, but reallyfrom A to Z, works and essays in Indo-European linguistics that have becomeclassics. I enjoy thinking back to when I read Wackernagel (KleineSchriften I / II 1969 III 1979, and Vorlesungen I 1926 II 1928) andSchulze (Kleine Schriften and Nachtri'ige 2" d ed. 1966) during my ownyears at university. The texts read like detective novels.It is worthwhile setting high personal standards and looking around tosee how other linguists have become what they are: A utobiographischeBerichte 1991 and Portraits I I II 1966.Like every other field, Indo-European linguistics has its unwritten rulesof the guild. Thus, it is the duty of each and every researcher to keep newdevelopements "in dialogue" with past research and to pay homage toprior accomplishments through frequent citation. The constantly growingquantity of data that one must master fo r this purpose is problematic. Butthe dealings with, and the reverence for what has preceded ought not toblock new insights into the future.E 106.At the time of my own studies (1967- 1973), the study of IndoEuropean linguistics was stiJI very simply structured.One was educated

6 Introductionby attending lectures, lower-level seminars, and upper-level seminars; oneengaged in individual reading, wrote papers, and after five or six years,chose a dissertation topic and completed his studies directly after the dissertationwith doctoral exams.Studies of Indo-European linguistics have become ever more reglementeddue to dramatic changes in European universities (in view here areparticularly the German ones).Today, fo llowing as a rule fo ur semesterseach of Grundstudium and Hauptstudium, and a semester of exams, oneattains the Magister (M.A.) degree. But only the completion of a dissertationallows one to think of a career in Indo-European linguistics.Whoeverchooses Indo-European linguistics as an occupation learns fo r a lifetime.And new changes continue to threaten courses of study:The latestfo rm of this threat is the introduction of a Bachelor's degree (B.A.) afteronly six semesters.sufficient thoroughness in three years.Indo-European linguistics cannot be conveyed withWhile the Bachelor candidate canobtain a good Indo-European linguistic education, he still needs at least anM.A. and a dissertation as qualifications in the field.B. Indo-European Linguistics in the Age of the PC andthe IntemetE 200. Indo-European linguistics, like any other science, can no longermake do without computers and the interne!. While PCs are indeed variouslyused, (the palette of uses ranges from those of a simple typewriter,to professional use in the word-processing of texts in a variety of languages),the potential uses of the internet lead to an ever-greater density ofinformation: -7 Studia lranica, Mesopotamica et Anatolica (i.e. SIMA) 2I 1996, published by J. Gippert and P. Vavrousek. Prague 1997. Here,one finds the files of the fourth Internationale Arbeitstagung fur Computereinsatzin der Historischen Sprachwissenschajt 1995 in Vienna, editedby H. Eichner and H. C. Luschiitzky.Today, good, practically oriented introductions for PCs and the internetare available. Examples include those of the Wissentschajtliche Buchgesellschajt:-7 H. Schroder, I. Steinhaus Mit dem PC durchs Studium.Indo-European Linguistics in the Age of the PC and the Internet 7Darmstadt 2000; D. Kaufrnann, P. Tiedemann Internet fur Althistorikerand A ltphilologen. Darmstadt 1999.E 201. I would like to call attention to three institutions specializing inIndo-European linguistics and the information that they offer.Therr websites include further information and a large variety of links. It should beclear to each user that the information on web sites changes constantly(The information included here is current as of March 1999.) and thtmany departments and institutes (and I include our department at the FreleUniversity among these) are in the course of creating better, more informativeweb page.1) T1TUS (Thesaurus Indogermanischer Text- und Sprachmaterialien)Begun by 1. Gippert, the goal of this institution, based for the moment mFrankfurt, is to process all linguistic materials that are relevant m questIons ofIndo-European linguistics into digital form that may easily be anaIyzed: -7http://titus.uni-frankfurt.de. A constantly growing quantity of data is availableunder the headings "Actualia," "Didactica," 'Textus" and 'TITUS." J. Gippertoffers an overview in the above-mentioned (E 200) SIMA-volmne p. 49-76. He also offers further descriptions of his project as of March 1999 underthe heading 'TITUS" as Beschreibung A (called 'TITUS, Das Projekt einesindogermanistischen Thesaurus in LD V-Forum" [i.e. Forum der Gesellschajtfor Linguistische Datenverarbeitung] Band 12 I 1 1995 p. 35-74) andBeschreibung B (which refers to "TITUS: Von der Keilschrift zur TextdatenbanK'in Frankforter Forschung 4 1995 p. 46-56). A more detailed text maybe found under the title Beschreibung C : C.-M. Bunz "Der Thesaurus indogermanischerText- and Sprachmaterialen (T1TUS) ein Pionierprojekt derED V in der Historisch- Vergleichenden Sprachwissenschaft." - 1. GIppertdescribes his comprehensive vision of the future in the Frankfurt research asthe fo llowing:"Beyond the archiving of field-specific data, the project, which since thethird conference on 'The use of computers in historical-comparative linguistics'in Dresden (October 1994) has been led under the succinct name"TITUS" (Thesaurus Indogermanischer Text- und Sprachmaterialien),should be extended increasingly to other areas of linguistic research.comprehensive bibliographical information system should play a centralrole, featuring new material in all areas that touch upon the field WIth .aclaim to the greatest possible timeliness. The internet will perform a deCIsivefunction here as well: The aspired-to timeliness may only be obtamedA

8introductionA Word on the History of lndo-European Linguistics 9if the infonnation is not printed, but rather only processed online; and thecompilation of individual pieces of infonnation, which an individual institutecould scarcely accomplish, should be distributed as quickly as possible tomany partners whose common contact is the interne!. There already existfirm agreements among colleagues at the universities of Prague, Vienna,Copenhagen, Leyden, Maynooth, among others. (A 'test-run' of the bibliographyis was being conducted from Frankfurt at tbe time of publication.)Under the same prerequisites (participation of as many partners as possiblein order to assemble complementary infonnation), a couple of other areasof application of the TITUS project are emerging that should allow it tomature into a comprehensive field-specific infunnation sYstem Thus, currentregular notifications about events in the field (congresses, and conferences,as well as university programs), open positions and offers, projects,research plans, etc. can all be viewed. The assembly of all such infonnationrequires, thanks to the internet, only a very small amount of space and timeon site. In order !o call attention to a conference that will take place at anAmencan uruversIty, the address of the invitation text must merely be enteredon the given internet page -provided, naturally, that event organizersplace their text on the interne!."2) The Indogermanische Gesellschaji, which is momentarily located inHalle (see link ọ our web site [see above E 100] under the heading, lndoEuropean lingUIstICS ill Europe): Up to date news from the IndogermanischeGesellschaft (about, among other things, the nature and goal of IndoEuropean linguistics and work with media) is offered, as well as addresses andgeneral infonnation. The Hallisches Institut fur Indogermanisti/c, Allgemeineand Angewandte Sprachwissenschaji is currently responsible for maintenanceoflhe page.3) The Institute for Linguistics at the University of Cologne. The rubrics"Sprachen and Schrijien der Welt," "Indogermanisch allgemein," "Antikeallgemein" among others, and the links to individual Indo-European languagesare avaiJable through the thankfully provided thematically organized links.There is a link to the University of Cologne on our web site (see above E 100under the heading Indo-European linguistics in Europe).C. A Word on the History of Indo-European LinguisticsE 300. It is not at all uninteresting to look up the entries under the heading'Indo-European linguistics' in a general encyclopedia.I) The groJ3e Knaur (Munich 1 ZUrich 1967) offers an astonishingly competenttreatment, which is given below without changes (the citations inthis paragraph [-4] refer to the encyclopedia; information about particularresearchers is provided in the index of this volume):"Indo-European linguistics, a science that explores the lndoEuropean languages. FoUowing the recognition, already in the 18thcentury, by W. Jones (1786) of the relatedness of Sanskrit to Europeanlanguages, R. Rask (1814), F. Bopp (1816), and 1. Grirnm (1819)founded Indo-European linguistics. Rask and particularly Grirnm(Deutsche Grammatik, 1819 fI.) researched the historical stages of theGermanic languages (see Consonant shift) in an exemplary manner.Whereas Bopp (Vergleichende Grammatik I 833ff.) analyzed and comparedforms, A.F. Pott provided underpinning -4etyrnology (EtymologischeForschungen, 1833-36) through exact comparisons of phoneticequivalences. Working with fixed rules of phonetic development, thefirst to try to attain an original [ndo-European language was A.ScWeicher (Compendium der vergleichenden Grammatik der indogermanischenSprachen, 1861/62); he was also the first to bring Slavic andparticularly Lithuanian into consideration. Researchers next tried todefine methods and phonetic rules more clearly: 1863 H. G. Grassmann'sLaw (dissimilation of aspirates), 1877 K. Vemer's Law(-4Grammatical change), 1876-78 Ausnahmslosigkeit der Lautgesetze(A. Leskien, H. Osthoff and F. K. Brugmann; -4Neogrammarians).Amelung, Brugmann, H. Collitz, F. de Saussure, J. Schrnidt resolvedthe problem of the Indo-European 'a' (European a, e, 0); G. I. Ascolidiscovered the two Indo-European guttural series; Brugmann (,Nasalissonans in der indogermanischen Grundsprache,' 1876), the syllabic mand n; de Saussure (' Memoire sur le systeme primitif des voyelles dansles langues indoeuropeennes,' 1878179) formed the vowel theory ofProto-Indo-European through systematic representation of ablaut degreesof short and long vowels, discovery of and of two-syllable-4Roots. H. Paul (' Prinzipien der Sprachgeschichte,' 1880) contributedthe theory of Analogy, the effect of which Brugmann and Osthoff

10 Introductiontreated in their 'Morphologische Untersuchungen' (1878 if.). H.Hilbschmann recognized the -7Armenian language as a separate languagegroup. B. G. G. Delbrilck contributed his Syntax (1893-1900) toBrugmann's 'GrundrifJ der vergleichenden Grammatik der indogermanischenSprachen' (1886 ff.).Significant investigations into the philology of individual languageswere provided by: Ch. Bartholomae (lndo-Iranian), J. Wackernagel, W.Schulze, and later P. Kretschmer (Greek), Fr. K1uge, H. Paul, E. Sievers,and still later W. Streitberg (Germanic), R. Thumeysen (Celtic). H.Hirt contributed in the areas of -7accent (1895) and -7ablaut (1900) aswell as to those of the original homeland and language of IndoEuropeans ('Die lndogermanen,' 1905-07; 'lndogermanische Grammatile,' 1921-37). Tocharian and Hittite were discovered at the beginningof the 20' h century and worked on by W. Schulze, E. Sieg, W.Siegling, W. Krause (Tocharian), and F. Hrozny, F. Sommer, J.Friedrich (Hittite), H. Pedersen (both). Along with Hittite, Luwian andPalaic were also revealed; Phrygian, Lycian, and Lydian were also researched.Krahe analyzed the remains of the -7Illyrian language. IndoEuropean linguistics became increasingly focused on questions of detailand individual philologies. Since de Saussure's demand for a synchronized,systematic linguistics ('eours de la linguistique generale,'1916), Indo-European linguistics, which is historically ('diachronically')oriented, has been replaced by various movements in modern -7linguistics,particularly abroad (Geneva, Prague, Copenhagen, USA)."2) Meyer's Enzyklopadisches Lexikon (MannheimlWienlZilrich. 9 " ed.1974) offers a comparahly competent overview under the heading. Dissappointing,on the other hand - yet for this day and age, perhaps typical -is a newly conceived reference such as Haremberg, Kompaktlexikon in 3Banden. Dortmund 1996. The field of Indo-European linguistics is nolonger mentioned; one is rather summarily referred to the "Indo-Europeanlanguage group" and the "Indo-Europeans."3) An exhaustive treatment of the history of Indo-European linguisticsfrom its beginning is lacking to this day.Helpful terature: - a) On the subject of the history ofIndo-Europeanlinguistics, with a particular focus on its beginnings and on the subject ofthe history of linguistics in general: -7 Benfey Geschichte derSprachwissenschaft 1869; Delbrilck Einleitung 1904; Windisch SanskritphilologieI 1917 IT 1920; Portraits I I II 1966; Neumann,Indogermanistik 1967; Koerner, Practicing Linguistic HistoriographyA Word on the History of Indo-European Linguistics 111989; Einhauser Junggrammatiker 1989; Szemen!nyi Ein/uhrung 1990 .p.I If. ' Bartschat Methoden der Sprachwissenschaft 1996; Morpurgo DavlesO/lcento 1996. - b) with focus on the 20' h century (until 1960): -7 SzemerenyiRichtungen der modernen Sprachwissenschaft 11 1982.4) A couple of important steps of development from § I are clarified inthe following.E 301. Similarities and relationships in vocabulary between Europeanlanguages such as Latin and Greek and Sanskrit have been increasinglystudied since the eighteenth century. -7 Thumb I HauschIld Handbuch desSanskrit I I I 1958 p. I 68ff. (The work concerns Sanskrit studies inEurope). On Sir W. Jones: -7 p. 173f.; Portraits 1 1966 p. 1-57; LexiconGrammaticorum 1996 p. 489f.; Mayrhofer Sanskrit and die Sprachen Alteuropas1983..In his German-language work, J. K1aproth (Asia polyglo/la ParIS 1823p. 42ff.) refers commonly to the language group that joins Europe andIndia as "indo(-)germanisch." But this term is clearly not K1aproth's invention.He uses the word as an established term, that at the time competedwith the term 'Indo-European' and was employed by Bopp. BeforeK1aproth, the Danish geographer K. Malte-Brun had eVIdently used tbeterm "Iangues indo-germaniques": -7 Thumb I Hauschild Handbuch desSanskrit II I 1958 p. 42f.; F. R. Shapiro "On the Origin of the term 'lndoGermanic"' in HL 8 1981 p. 165-170; K. Koerner "Observations of theSources, Transmission, and Meaning of 'Indo-European ' and RelatedTerms in the Development of Linguistics" in IF 86 1982 p.I-29; by thesame author, Practicing Linguistic Historiography 1989 p. 149-177; Szemen,nyiEinfuhrung 1990 p. 12f. note I; G. Bolognesi "Sul termine 'indogermanisch"' in FS Belardi I 1994 p. 327-338; F. Bader in langues indoeuropeennes1994 p. 23..While the term 'Indo-European' established itself in English and m theRomance languages, the term 'indo-germanisch' has become accepted inthe German-speaking world: -7 Committee of the lndogermanische Gesellschaftin Kratylos 27 1982 [1983] p. 221f. (position with regard to thejuxtaposition of 'lndo-Germanic' and 'Indo-European': "Eine Abkehr vondem eingebiirgerten wissenschaftlichen Terminus 'indogermanisch ' istalso nicht geboten"). A consciously anti-West German development, teterm 'indoeuropiiisch' was the term of choice in tbe German DemocratIcRepublic: -7 E. Seidel in Wissenschaftliche ZeitschriJt der HumboldtUniversitiit zu Berlin, Gesellschafts- and Sprachwissenschaftliche ReiheXVrn 1969 p.297 ("In dealing indirectly with West German servants of

12 Introductionimperialism, I see no reason to avoid using the term 'indogermanischeSprachwissenscha ft ...Nonetheless, I shall respect the wishes of the editorand use ' indoeuropdisch. "')E 302.The actual history ofIndo-European linguistics begins with FranzBopp (1791-1867). In 1816 he proved the relationship of the IndoEuropean languages. The foreword of his fundamental work, "Uber dasConjugationssystem der Sanskritsprache in Vergleichung mit jenem dergriechischen, lateinischen, persischen and germanischen Sprache" isdated 16 May, the date on which Bopp celebrated the birth of IndoEuropean linguistics. Whereas earlier suppositions were only supported bycomparisons of words, Bopp proved the existence of relationships throughgrammatical comparison.Bopp's study of Vedic was prompted byFriedrich Schlegel's "Ueber die Sprache and Weisheit der Indier" (Heidelberg,1808). For more information on Bopp: -? B. Schlerath BerlinischeLebensbilder - Geisteswissenschaftler 1989 p. 55-72; SzemerenyiEin f uhrung 1990 p. 6f.Along with Bopp, Jacob Grimm (1785-1863) is of great significance inthe history of Indo-European linguistics. With his Deutsche Grammatik(I 819ff.), he introduced the historical dimension in linguistic research: linguisticcomparison and history constitute the foundaton of Indo-EuropeanlingUIStiCS: -? Szemerenyi Richtungen I 1971 p. 13ff.E 303.The institutionalization of the field began with the appointment ofBopp in 1821, on the recommendation of Wilhelm von Humboldt to thethen new Berlin University. Bopp received the newly created chair for"Orientalische Literatur und Allgemeine Sprachkunde."In its first decades, the field remained closely associated with Sanskritstudies, given that familiarity with Sanskrit rendered the discovery of theIndo-European language group possible. Because of this tight connectionto Sanskrit studies, Indo-European linguistics in its beginnings was closestto oriental studies, such that professors at the time usually carried thewords 'Sanskrit' and 'oriental' in their titles. Yet the descriptions 'Indianstudies' and 'oriental studies' have just as little to do with their presentmearungs as Bopp's chair of 'allgemeine Sprachkunde' had to do withtoday's understanding of general linguistics. August Wilhelm Schlegelreceived the first chair in Indian Studies in 1818 in Bonn.Chairs in Indo-European linguistics without particular ties to Sanskritwere created starting only in the 1870's.Thus, for example, Karl Brugmann'schair in Leipzig was created by simply renaming the chair for Clas-A Word on the History of lndo-European Linguistics 13sical Philology occupied by his teacher, Georg Curtius, who had been astudent of Bopp in Berlin.E 304. Bopp's circle of students was very large. For example, FriedrichRiickert, who held a chair for oriental studies in Erlangen and taught inBerlin, is significant beyond the confines of Indo-European linguistics.The circle included Wilhelm von Humboldt, August Wilhelm Schlegel andmany representatives ofIndo-European linguistics and Indian studies, suchas August Friedrich Pott, Theodor Aufrecht, Olto von Bohtlingk, AdalbertKuhn, Adolf Friedrich Stenzler, and Albrecht Weber, who became Bopp'ssuccessor.In 1872, Hermann Ebel received the first chair for comparative linguisticsin Berlin. In 1876, Johannes Schmidt, a student of August Schleicher,became his successor.Herrnann Ebel fo unded the "Berliner Schule" ofphilological Indo-European linguistics, which is to be distinguished fromthe "Leipziger Schule" of systematic "neograrnmarians."E 305.The introduction of various new methods is associated with thenames Pot!, Schleicher, and Schmidl: Thus, with A. F. Pot! from Halle isassociated a more rigorous observation of consonant shift in the study ofetymology (-? Etymologische Forschungen auJ dem Gebiete der IndogermanischenSprachen mit besonderem Bezug auJ die Lautumwandlungim Sanskrit, Griechischen, Lateinischen, Litauischen and Gothischen.Lemgo 1833-1836); with A. Schleicher from Jena, reconstruction andtheories of linguistic lineage; (-? Compendium der vergleichenden Grammatikder indogermanischen Sprachen. Weimar 1861); with J. Schmidtfrom Berlin, the wave theory (-? Verwandtschaftsverhaltnisse 1872). Thename Adalbert Kuhn calls to mind both Indo-European mythology as wellas the founding of a review in the field of lndo-European linguistics, a reviewwhich, with only minimal changes in the title, has appeared from1852 to the present day, and is still referred to as "Kuhns Zeitschrijf': HS(previously ZVS or KZ). See the bibliography.E 306.Representatives of different philological disciplines belonged tothe group of so-called neogrammarians, which included for example, theslavist August Leskien and the germanist Hermann Paul. Literature: -?Einhauser Junggrammatiker 1989.The principle that phonetic rules are without exceptions (which includesthe consonant shift as a phonetic rule) can be traced to the neograrnmarians.Numerous durable phonetic rules were discovered by researchers

16 Introductionthe statements that regard issues specific to the situation at the Universityof Base I and Switzerland in general.):"Here I should, in conclusion, accommodate today's utilitarianism bynaming several additional factors that in my opinion clariJY a wellgroundedhistorical-comparative, and particularly an Indo-Europeancurriculum as something useful, even outside of the more narrow areaof classical philology and lndology, factors that can relativize the orchid-statusof Indo-European linguistics: - Firstly, Indo-European linguisticscan like no other science mediate between most languages ofEurope, and particularly between the four languages of our country: Itguides the regard to the COmmon linguistic fund and the differences thathave grown over time. It helps us to be familiarized ourselves, likewisearcheologically, with the confusion of different historical layers, to keepthem distinct from one another, and to be mindful of the historical contextof inherited linguistic material, influences of languages of classicalantiquity and the renaissance, exchanged linguistic material from theRoman-Germanic symbiosis of the early middle ages, scientific terms ofthe high middle ages that originated in Arabic, borrowings from thecourtly culture of France, Anglo-Saxon technological-commercial vocabularyof the last hundred years, and much else. The historicalcomparativeperspective could, and furthermore should be made onceagain increasingly fruitful in language studies in the sense of applied linguistics.This is particularly useful for the Romance languages, ofwhich three are official languages in Switzerland, and a fo urth is amongthe most widespread languages of the world, further in the area ofEuropean common cultural vocabulary, which in many cases is commonto all four official languages of Switzerland, as well as to English andthe other European languages, and finally more or less in the structureof sentences and the typological changes of the last 2000 years, whichlikewise in nearly all of Europe reveal more commonalities than differences.And not least, the historical-analytical perspective, as I see it,encourages individual linguistic competence, trains sensitivity to styleand broadens available linguistic resources. - Secondly, historicalcomparativelinguistics can help general linguistics to regain dimensionsthat the latter has for some time neglected, namely the historical andcomparative dimensions. Lately, a convergence has been discemible,and here in Basel the signs of fruitful collaboration seem to me particularlypromising. - Thirdly, lndo-European linguistics contributesgreatly to the color of a university, fo r it brings its own bases of researchand is at the same time a helpful field for many others, is cen-A Word on the History of lndEuropean Linguistics 17trally important for classical philology ong .other filds, and meingfullycomplementary to most other philologles, and .It bradens sIgnificantlythe selection of available fields through the mcluslon of languagesthat would otherwise never be taught or researche. Through't multilingual integrating effect, Indo-European lingUIStICS creates,I sd" alalong with history and comparative literature, .an ad Itlon , partlCU ar ylinguistically-oriented network of the most dIverse fields. As an .etymologicalscience par excellence, It IS particularly capab!e of findmg abroader audience. - Last but not least, lndo-European ImgUlStIcs doesnot cost very much: This is true in absolute terms, because It requlfes,aside from a minimum to assure continuity, very little personnel; books,reviews and other resources as well are required, commensurately WIththe breadth of the field, in modest quantity. But also considered relativelythe cost-benefit relationship is not at all bad, because here notonly :.re the numbers of students relevant, but also just as uch thequalitative aspect of the contribution of lndo-Europe lingUiStICS tothe functioning of many other fields and of the broadenmg ofthe chOIceof fields ... - It is a particularly desirable and effectIve reinforcementof the successful functioning of this small field that is full of tradItIonthat the many other fields actually take advantage of its capacity tohelp. To achieve this, the field, or rather its representative, must contributehis part in his teaching, in his relations to students and representativesof the other fields, among others, as well as - today more thanever - toward the general public."D. Overview of the Indo-European Languages and theirSources1. General InformationE 400. Attestation and extent of documentation of the Indo-Europeanlanguages varies from language to language. This is dependent on whenthe individual groups of lE language speakers found thir way from thespoken word, originally prevalent in all of them, to the wntten one..As a rule, this development took place at the time of contact WIth anestablished culture that employed writing. Compare for example the-.II

18 IntroductionAnatolian Hittites, who may be integrated in the Mesopotamian cuneifonntradition (cf. E 410 below), or the Mycenaean Greeks, who borrowed LinearB from the Cretan family of scripts (cf. E 418 below), or the Celts,who, depending on the region, wrote their inscriptions with the Greek,Latin, Etruscan, and even Iberian alphabets (cf. E 43 1, §1 below), or theTochari, who, through their participation in the life marked by the Buddhismof the Tarirn Basin of the sixth century AD, created languagemonuments of their own (see below E 408).The earliest attestations of some language branches are translations ofChristian content.These include Gothic, Old Church Slavic, and Armenian.A table of entry dates of individual languages into the world ofwritten language is offered in Benveniste Institutions IT 1969 in the openingtext to the "note bibliographique."In the best ofall cases, the age of the language is consistent with that ofthose who created its records, as is the case with contemporary inscriptions.In other cases, the documents originate at a much later date, such asis the case as a rule with manuscripts. Thus, there is a period of oral tradition,or also of written tradition between the attested linguistic phase andthe point at which the physical document can be dated.Some languages have only come to be known in the last century,whether it be because they had remained undiscovered, or because thewritten documents could only then be deciphered.The capacity to decode the languages concerned here varies both accordingto whether or not the given language has a modem descendent,and according to the length of its philological tradition.E 401. General overviews of the individual representatives of the IndoEuropean language f.amilies are available in: Cowgill Einleitung 1986 p.17ff. ; Lockwood Uberblick 1979; Lingue indoeuropee 1994 = IndoEuropean languages 1998; langues indo-europeennes 1994; Beekes Introduction1995 p. 17ff. ; Convegno Udine (Restsprachen) 1981 [1983].2. The Individual Indo-European Language Families and theirSourcesE 402.A short initial enumeration follows here in the order of the earliestattestation of the individual language.In each case, I shall mention theearliest evidence, and in the case of datable records the actual evidenceshall also be mentioned.Further, an indication is given of an indirectlyOverview of the Indo--European Languages and their Sources 19transmitted text if it is significant for the particular language because ofage or sIZe.The current order is as fo llows: - Anatolian (Old Hittite, originaldocuments from the 16t h century BC, copies of texts from the 17t h centuryBC); - Greek (Mycenaean original documents of the 17t h and 14t h /13t hcenturies BC); - Indian (13th century BC: the handing-down of theRigveda must have taken place purely orally until far into the last millenium,but the creation of individual verses and of some philosophical contentdates probably from the 13t h century BC; further, material transmittedthrough secondary sources, including some names of deities and termsdates back to the Hurrian Mitanni empire of the 16t h to 14t h centuries BC.);- Iranian (the core of the Old Avestan text corpus dates back to Zarathustra,founder of zarathustranism, and thereby to the 10t h century BC,but after a long period of oral transfer, the conserved texts were only preservedin written fo nn starting in the 13t h century AD, in the Middle Persianperiod); - Italic (perhaps the so-called fibula praenestina [if it iseven authentic, and not a forgery: Wachter Altlaleinische Inschriften1987 pp. 55-65] with its inscription can be dated to the first half of the 7thcentury BC; whereas other Latin monuments, such as the so-called Duenosinscription, only date from the 6t h century); - Celtic (continental celticinscriptions date from the 2nd century BC); - Germanic (Wulfila's Bibletranslation in Gothic is dated c. 3 SO AD; Gennanic names on coins and inindirect records are attested from the I " century BC); - Armenian ( 5thcentury AD); - Tocharian (6t h century AD); - Slavic (9t h century AD);- Baltic (14t h century AD); - Albanian (15t h century AD).E 403. The fo llowing somewhat more thorough enumeration proceedsgenerally from geographical east to west, and fo llows, within each region,the order of first attested occurrence.kept quite brief.References to helpful literature are1) Indian subcontinent and Chinese Turkestan: Indo-Inmian with Indianand Inmian (which adjoins from the west) ; Tocharian.2) Asia Minor, Greece, and the Balkan peninsula: - From the 2"" centuryBC Anatolian in the east, Greek in the west. - From the I " millenium BCPhrygian in Asia Minor. - Armenian in the east and Albanian in the Balkansbeginning after Christ.

20 IntroductionOverview of the lndo-European Languages and their Sources 213) Italian peninsula: Italic.4) Europe north of the Alps: Celtic, Germanic, Balto-Slavic.E 404.Indian, or Indo-Aryan:Indian and Iranian are closely linked both linguistically and culturally intheir early stages.Even the name for the people, arja-, shared by bothlanguage branches, is an expression of intertwined Indo-Iranian cultures,c( W 304. For a good overview of Indo-Iranian: ..... M. Mayrhofer inlangues indo-europeennes 1994 p. 101-120; Arbeitstagung Erlangen 1997[2000].I) Indian or - with reference to the non-Indo-European languages of India- Indo-Aryan is first attested (16t h /14t h centuries BC) indirectly in other traditions,and in fact in the fo rm of borrowed words and proper nouns in the Hurrianof kingdom of Mitanni ("Mitanni-Indian"): ..... Mayrhofur lndo-Arier1966; Kammenhuber Arier 1968; M. Mayrhofer Welches Material aus demlndo-Arischen von Mitanni verbleibt for eine selektive Darstellung? in KleineSchriften n 1996 (an essay from 1982) p. 304-322; O. Carruba Zur Oberlieferungeiniger Namen and Appellativa der Arier von Mitanni: "A Luwianlook? " in Arbeitstagung Erlangen 1997 [2000] p. 51-67. On the subject ofHunian itself: ... Neu Hurrilisch 1988; Wegner Hurritisch 2000.2) The oldest layer of the Indian language, which has continually developedthrough the present day, is tangible in the Vedic of the Rigveda, which may bedated to the middle of the 13t h century BC and represents the Indian languageof the Punjab region in the north-west oflndia. The dating concerns howeveronly the language, and not the documents, since written transmission of Vedictexts began only two thousand years later. Because the oldest Vedic texts arecerernonial literature in verse, a genre in which exact phonetic observation andpreservation are of the highest priority, the greatest reliability on the part of theoral tradition is assumed.WIthin Vedic Indian, different phases of languagemay be differentiated, which are connected with the various texts: Rigveda,Sarnaveda, Yajurveda, Atharvaveda, Brahmanas, Upanishads, Aranyakas.Even within the Rigveda, the collected hymns are not all of the same age; theoldest being those of the books two through seven, the so-called 'fumilybooks'. Different dialects ofVedic may also be distinguished.Literature: - a) general: ..... Thumb 1 Hauschild Handbuch des Sanskrit1958 / 1959; Wackernagel l Debrunner Altindische Grammalik 1957 / 1954 11930. - b) Vedic: ... MacDonell Vedic Grammar 1910; Aufrecht Hymnendes Rigveda I 1 11 1877; Geldner RV Obersetzung 1951-1957; GrassrnannWorterbuch 1873; Mayrhofer EWAia; Hoffinann Injunktiv 1967; Narten Sig-malische Aoriste 1964; Goto, I. Priisensklasse 1987 and by the same autbor,Materialien Nr. 1-29 1990-1997; Zehnder AVP 2 1999; M. Witze "Tracingthe Vedic Dialects" in Dialectes indo-aryennes 1986 [1989] p. 97-265; by thesame author, Die sprachliche Situation Nordindiens in vedischer Zeit in ArbeitstagungErlangen 1997 [2000] p. 543-579.3) The earliest directly handed-down Indian pieces of linguistic evidence areinscriptions from the Buddhist Emperor Moka of 250 BC. The developmentof Priikrit begins in around 500 BC. Part of Priikrit is Pali, the canonical languageof southern Buddhism: ..... Geiger Piili 1916; Mayrhofer Piili 1951; vonHinilber Alteres Mittelindisch 1986.4) Classical Sanskrit, which as a literary and intellectual language is used tothe present day, appeared only after Middle Indian in the second half of thefirst millenium BC under the influence of the grammarian PiiI)ini (c. 400 BC)and others: . .. Mayrhofer Sanskrit-Grammatik 1978.5) Among the New Indian languages, Hindi and Urdii are particularlyworth mentioning. For information on the present linguistic situation: ..... P.Gaeflke and H. Bechert in Indologie 1979 p. 32ff.E 405. Already in its oldest attestations, Old Iranian may be divided intoan eastern and a western branch. East Iranian is represented by Avestan;West Iranian by Old Persian. Avestan and Old Persian developed separatescripts: Avestan is written from right to left in an alphabet that is basedupon the cursive Pahlavi alphabet of the 4t h century AD and which, thanksto its weallh of characters, can take into account fine phonetic differences.Old Persian, however, is written in a simple cuneiform that developedseparately around 520 BC.The fragmentary attestation of Old Iranian leads great importance to thesuccessive Middle and New Iranian languages. See an example of Khotanesein L 211 § 5.Literature: ... Compendium Linguarum lranicarum 1989 with articleson all aspects of Iranian (from Old Iranian and Middle Iranian to New iranian);R. Schmitt Die iranischen Sprachen, Eine Einfiihrung in 5 Teilen inSpektrum Iran 8,4 1995 p. 6-27; 9,2 1996 p. 6-32; 9,3-4 1996 p. 6-32;10,1 1997 p. 10-38; 11,1 1998 p. 14-42 = Schmitt lranische Sprachen2000; M. Mayrhofer 'L 'lndo-iranien' in /angues indo-europeennes 1994p. 101-120; Bartholomae Altiranisches Worterbuch 1904 (1979); HoffmannAltiranisch in Azifsdtze I 1975 p. 58-76 (Article from 1958; otherwisecompare the essays I-Ill, which contain Hoffinann's central studies on

22 IntroductionOld Iranian); R. S. P. Beekes "Historical Phonology of Iranian" in JIES25 1997 p. 1-26.E 406. The oldest occurrences of Avestan are the so-called Gathas ofZarathustra ("Gatha-Avestan"), which are hymns to the deity AhuraMazda. These excerpts, together with the Yasna HaptaQhiiiti, a ceremonialtext in prose date from the 10 " century BC. - Later Avestan is a laterdialect of Old Avestan and dates from around the 6' " and 5' " centuries BC.The oldest known manuscript text dates from 1288 AD .Literature: Beekes Gatha-Avestan 1988; 1. Kellens in CompendiumLinguarum Iranicarum 1989 p. 32-55; Kellens / Pirart Textes vieilavestiquesI-m 1988-1991; Hoffinann / Forssman Avestische Laut- andFlexionslehre 1996 (p. 247ff A compilation of writings on Avestan by B.Forssman).E 407. Old Persian is first attested at the time of the Old Persian cuneiformscript, thus c. 520 BC; but the inscriptions from the 3' d century BCalready contain linguistic mistakes that indicate that Old Persian was alreadyno longer a contemporary language. In some cases, Old Persian alsocontains "medisms." The entire fund of Old Persian texts is comprised of asmall corpus of inscriptions: Brandenstein / Mayrhofer Altpersisch1964; Mayrhofer Supplement 1978; R. Schrnitt in Compendium LinguarumIranicum 1989 p. 56-85; Schrnitt Bisitun Inscriptions 1991; by thesame author Altpersische Inschriften 1999; M. Mayrhofer mer die Verschriftungdes Altpersischen in Kleine Schriften II 1996 p. 387-399 (essayfrom 1989).For information on the Parthian Empire under the Arsacids (247 BC -224 AD): Partherreich [1996] 1998 (which contains, among others, R.Schrnitt Parthische Sprach- and Namenuberlieferung aus arsakidischerZeit p. 163-204. The oldest Pahlavi-inscription comes from the founder ofthe Sassanid dynasty, Ardashir I (Papakan = Ardashir) (224-241 n. Chr.): Overviews in Compendium Linguarum Iranicarum 1989 p. 95ffE 408. Two languages may be distinguished in the case of Tocharian:East Tocharian, or Tocharian A, and West Tocharian, or Tocharian B.Both were used in eastern Turkestan starting in the 2" d century BC. Theearliest attestations date from the 6' " century AD; the latest from the 8' "century AD . - Tocharian A was a purely written language. Tocharian B,on the other hand, was a lingua franca in Turfan, Qarasahr, Sorcuq, andKuca. Contents of literary texts include poetry, religious, and scientificOverview of the Indo-European Languages and their Sources 23material. Extant religious, Buddhist texts, of which several are for themost part translations from Sanskrit, including some that are bilingual, andothers that are copies of a known original. In addition, in Tocharian B,there are reports from monasteries, caravan passes, a letter, and text frommural paintings. A modified Northern Indian Brahrni alphabet was used._ Around 1900, expeditions explored the Chinese province Xinjiang. In1904, A Le Coq and A. Grilnwedel discovered a distinctly Tocharian languagein manuscripts from eastern Turkestan. The language was furtherresearched by E. Sieg and W. Siegling: W. Siegling Tocharisch, dieSprache der Indoskythen in Sitzungsberichte der Berliner Akademie 1908p. 915-932 (Sergej Th. Oldenburg first postulated the lndo-Europeancharacter of Tocharian in 1892: E. N. Tyomkin in TIES 7 1997 p.205ff.).Literature: An excellent introduction may be found in Pinault Tokharien1989; Fachtagung Tocharisch Berlin 1990 [1994] (which contains,among others, p. 310ff. G. Klingenschrnitt Das Tocharische in indogermanistischerSicht, see also, by the same author, Tocharisch and Urindogermanischin Fachtagung Regensburg 1973 [1975] p. 148-163); TIES(see the information in the bibliography); Adarns Tocharian 1988; AdarnsDictionary (Tocharian B) 1999; Ringe Sound Changes in Tocharian I1996; Hackstein Sigmatische Prdsensstammbildungen 1995; Carling LokaleKasus im Tocharischen 2000.E 409. With its Old Hittite cuneiform texts from the 16' " century BC, theAnatolian group of languages offers the oldest attested occurrence of anIndo-European language. Eight Anatolian languages are attested: Hittite,Luwian, Palaic, Lycian, Lydian, Carian, Pisidian, and Sidetic. - TheAnatolian languages use three different systems of writing: A type ofbabylonian-Assyrian cuneiform (Hittite, Palaic, Luwian), hieroglyphics(Luwian), alphabet (Lycian, Lydian, Carian, Pisidian, Sidetic).Literature: Melchert "Anatolian" in langues indo-europeennes 1994p. 121ff; N. Oettinger "Die Gliederung des anatolischen Sprachgebietes"in ZVS 92 1978 [1979] p. 74-92; by the same author, in DNP under theheading Kleinasien column 555-559.E 410. An archive of clay tablets was discovered in 1906 inHattusaIBogaik6y, 150 kilometers east of Ankara. The linguistic materialfo und in the Arzawa letters from the Amarna correspondence, found in1887/88 in Middle Egypt moved Knudtzon to express the supposition in

26 Introductionfo und in lasos dates from as early as the 7" century BC; the Carian-Greekbilingual stele from Athens dates from the end of the 6" century BC. Thealphabet is of the Anatolian type, but an independent creation. The appropriateattribution of phonetic values, which was only recently achieved,was confirmed through the discovery of a Carian-Greek bilingual stele inKaunos.Literature: ---7 Adiego Studia Carica 1993; Cario 1993 [1994]; ColloquiumCaricum 1998 (on the bilingual stele from Kaunos); 1. Hajnal inKadmos 36 1997 p. 141-166 and 37 1998 p. 80-108.E 416. In the southern Anatolian area of Pi sidi a and around the city Side,successor languages ofLuwian are attested to have been spoken. We havea whole series ofnarnes as well as a few inscriptions from Side.Literature: ---7 Neumann Kleine Schriften 1994 p. 227ff. (Essays onSidetic may be found under the numbers 33, 39, 43, 48, 49.)E 417. Greek is a language that can be documented over a period of4000 years. - The oldest documents are clay tablets in Linear-B, seebelow, E 418. Greek inscriptions in alphabet script date from as early asthe 8" century BC. With the exception of Cypriot, which employed a syllabicscript similar to that of Mycenaean which dates back to the secondmillenium BC, Greek texts are written with an alphabet. The Greek alphabetwas probably adapted from a northwest Semitic alphabet around 800BC. - In the following, the most important texts are given first. Paragraphshave been added for Mycenaean Greek (see below, E 418), Homer(see below, E 419) and the dialects (see below, E 420).Literature: - a) General: ---7 Schwy2er Griechische Grammatik I 1939;Schwyzer / Debrunner Griechische Grammatik II 1950; E. Risch in LAW1965 under the heading Griechisch; Meier-BrUgger Griechische SprachwissenschaJtI / 11 1992; K. Strunk Vom Mykenischen bis zum klassischenGriechisch in Griechische Philologie 1997 p. 135ff.; B. Forssman in DNPunder the heading Altgriechisch; Meillet Apert;u 1975; HierscheGrundzuge 1970; Risch Kieine Schriften 1981 (compare Indices); RuijghScripta Minora I 1991 IT 1996 (compare Indices). - b) Phonetics andmorphology: ---7 Lejeune Phomitique 1972; Alien Vox Graeca 1987;Zinsmeister Griechische Grammatik I 1954; Rix Historische Grammatikdes Griechischen 1976; Bomemann / Risch Grieschische Grammatik1978. - c) Syntax: ---7 Delaunois Syntaxe 1988; Kolloquium Kuhner Amsterdam1986 [1988]. - d) Vocabulary: ---7 Frisk GEW; ChantraineDELG; DGE (compare Anejo 1II); CEG.Overview of the lndo-European Languages and their Sources 27E 418. The earliest written evidence of Greek that has survived, is comprisedof inscriptions in a syllabic script ("Linear B"). The oldest knowndocument comes from the area around Olympia, contains the proper nounKharokws (compare the Homeric XapoljI), and dates from 1650 BC: ---7 P.Arapogianni, J. Rambach, L. Godard in Floreant Sludia Mycenaea I 1995[1999] p. 39-43. Part of the clay tablets from Knossos originate in the 14"century BC, the rest are dated around 1200 BC. Aside from Knossos onCrete, mainland sites include palaces in Pylos, Mycenae, Tiryns, and (inBoetia) Thebes. The language of the inscriptions is referred to as MycenaeanGreek or simply as Mycenaean. This early phase of Greek came tobe a fo cus of research only in 1952, after Michael Ventris, along with J.Chadwick, deciphered Linear-B script. It is notable that Greek linguisticshas only been able to integrate Mycenaean material in its discussion sincethe 60s. A textbook such as Schwyzer's Griechische Grammalik I 1939can certainly maintain particular assertions that were at the time thinkable,but which today may no longer be maintained, fo r example the fonnerlycommon analysis of EVEKa 'because' < "en-!!eka 'with regard to the will,'which because of Mycenaean e-ne-ka, i.e. eneka must be rejected (In thecase of "en!!eka, "e-we-ka would be expected!). The questionable EvEKamust now be seen as the fo ssilized root word "",lIeK-", in the accusativesingular case, which means 'for the attainment of (with a genitive complement)':---7 Meier-BrUgger Griechische SprachwissenschaJt I 1992 p. 88f.Literature: ---7 SMID; Chadwick Documents 1973; Hiller / PanaglFriihgr. Texte aus myk. Zeil 1976; Aura Jorro Diccionario micenico. I1985 IJ 1993; Meier-BrUgger Griechische SprachwissenschaJt I 1992 p.43ff.; Hooker Linear B 1980; Lejeune Memoires I-IV 1958-1997. - Thelast colloquia: ---7 Colloquium Mycenaeum 1975 [1979]; Res Mycellaeae1981 [1983]; Tractala Mycenaea 1985 [1987]; Mykenaika 1990 [1992];Floreant Studio Mycenaea 1995 [1999]. - Concerning new finds in Thebes:---7 V. Aravantinos in Floreant Studia Mycenaea I p. 45ff. Further,compare the CRAI-Iectures by L. Godart and A. Sacconi: Les dieux Ihebainsdans les archives myceniennes in volume 1996 p. 99-1 13; Les archivesde Thebes et le monde mycenien in volume 1997 p. 889-906; Lageographie des elals myceniens in volume XX 1999 p. 527-546E 419. The first Ancient Greek literary texts are the two epics by Homer;the Iliad and the Odyssey, the writing of which is probably correctly datedto the 8th century BC. The oldest evidence of the written transmittal ofthese texts are papyri from the third century BC. Although there hadprobably been a couple of editions before this date, well into the Hellenis-

32lntroouction- The comedy Poenulus is by Plautus. - Further literature on Punic: M. G. Guzzo Amadasi in LDIA 1978 p. 1013ff.E 427. Latin is by far the best attested of the Indo-European languages ofancient Italy.I) The Latin language was originally the dialect of the city of Rome andwas closely connected to tbe landscape ofLatium: Kolb Rom 1995.2) The earliest occurrences are inscriptions fo und in the city of Rome fromthe 6t h century BC. From the 5t h to the I" centuries BC, the language is knownas Old Latin. The majority of preserved texts come from the period from the, " century BC to the I" century AD. Classical Latin, strictly speaking, includesonly the prose writings of Cicero and Caesar, which date from the I "century BC.Overview of the lndo--European Languages and their Sources 33o6) Of particular importance is also Vulgar Latin, .the spoken language of which in the various Roman provinces the indiVIdual Romance languagg:ta1an, and Portugese) developed. In a unique way, the relatlonshi oomanian, Romansch, Sardic, French, Dalmatian, Italian, Prove .nyaj, Sparush,:: Latin: ViiiWinen Latin vulgaire 1981.ther language to its oflSpring may be observed and documented. nE 428. The oldest occurrences of Faliscan, the language of Falerii and thesurrounding area, are inscriptions from the 6t h cent BC; the latest Falis-. . ean msenp t'o I ns are from the 2,d century BC. ASIde from the sparse m-.script ions there are no attestations of Faliscan. L' Iter t ure: -: Veller .Handbuch der italischen Dialekte I 1953 p. 277ff.; Glacomelli LmguaFalisca 1963; G. Giacomelli in LDIA 1978 p. 505ff.3) For information on Old Latin inscriptions: Emout Recueil 1947;Diehl Alllaleinische lnschriften 1965; Warmington Remains oJ Old Lalin IV1940; Degrassi inscripliones I-IT 1965-1972; by the same author Imagines1965; Romische inschriften, by L. Schurnacher. Stuttgart 1988 ( Reclam,Universal-Bibliothek Nr. 8512) and Die romische Lileralur in Texl and Darslellung.volwne I: Republikanische Zeil r (Poesie) by H. and A. Petersmann.Stuttgart 1991 ( Reclam, Universal-Bibliothek Nr. 8066); Meyer LaleinischeEpigraphik 1973; Bliimel Unlersuchungen 1972; Radke Archaisches Lalein1981; Wachter Alllaleinische Inschriften 1987; Vine Archaic Lalin 1993.4) The first Latin literary texts are only attested from the 3rd century BC: H. Rix "Schrift and Schriftgebrauch im vorlilerarischen Millelilalien" inHoffinann Gedenlif€ier 1996 (1997] p. 27-42.5) Literature on Latin: - a) History of the Latin language: Meillet Esquisse1928; Devoto Lingua di Roma 1940; M. Leurnann Geschichle derlaleinischen Sprache in Leurnann / Hofrnann / Szantyr Allgemeiner Teil 1965p. 10·ff.; Solta Slellung der laleinischen Sprache 1974; Giacomelli Lingualalina 1993; 1. Kramer Geschichle der lateinischen Sprache in LateinischePhilalogie 1996 p. 115-162; M. Meier-Briigger in RGA 18 2001 under theheading 'Latein'. - b) General Latin: Leurnann / Hofrnann / Szantyr AllgemeinerTeil 1965; H. Rix in DNP 6 1999 column 1160- 1163 under theheading 'Latein'; Kolloquium Latein and indogermanisch Salzburg 1986[1992]; Sommer Handbuch 1948; SOmmer / Pfister Lautlehre 1977; LeumannLLFL 1977; Meiser Laut- and Formenlehre 1998; Schrijver Laryngeals inLatin 1991; Benedetti Composti radicali 1988. Further, compare the serieslKLL (I CLL / ClLL), see the bibliography, under lKLL.E 429.Among the languages of the Sahellian branch of languages are socalledSouth Pieene, Oscan, Umbrian, and a couple of other, only sghtlyvarl'ed languages , such as Volscian. Of partlcuiar mterest IS South PlC en e,the oldest attestations of which date from t h e 6t h cen t ury BC . Sout h p1-cene is particularly interesting. Three different alphabets were employe mOscan inscriptions: the Greek alphabet, the Latin Iphabet, and the Oscru:alphabet. The oldest inscriptions date from the 3.century BC.d .The socaUed Iguvine Tables are the main source for Umbnan and come from the3' " _2, d centuries BC.Literature: Meiser Umbrisch 1986; G. Meiser, Pailgmsch, Latemund Sudpikenisch in Glotta 65 1987 p. 104-I 25; H. Rix Umbro e Pto:Osco- Umbro in Convegno Udine (mmor language ) 1991 [1993], arl. . . . ..nett! !scnzlOm sudplcene 1985' , G. M else ' rand H Rix m Tavole dl Agnone'1994 [1996] p. I 87ff. and p. 243ff.; H. Rix, Sudpikemsc . h kd u/u . . III ' HS 1071994 p. 105-122; Schirmer Wortschatz 1998; Untermarm WorterbuchOskisch-Umbrisch 2000.. . .E 430. Venetian represents a separate Italic languạge of what is now t?Veneto region (inscriptions from the 6t h _2, d centurIes BC): PeUegnruProsdocimi Lingua Venelica 1 / II 1967; Lejeune Venele 1974; A. L. Prosdocimi11 venetico in LDIA 1978 p. 257/f by the same author m ConvegnoUdine (minor languages) 1981 [1983] p. 153ff.E 431. Celtic can roughly be divided into Continental Celtic (from theEminent) and Insular Celtic (from the Bnllsh Isl s [and fromth':r::Oh migrations, Brittany connects linguistically WIth Breton D.

36 IntroductionThesen zum Ursprung der Runenschrift in Etrusker niirdlich von Etrurien,Akten des Symposions von Wien - SchlofJ Neuwoldegg 1989, edited by L.Aigner-Foresti. Vienna 1992 ( SbOA W vo!. 589) p. 411 If. ; by the sameauthor, Germanische Runen und venetische Phonetik in Feslschrift O.Werner, Vergleichende germanische Philologie und Skandinavistik, editedby Th. Birkmann et al. Tiibingen 1997 p. 231-248; E. Seebold Fupark,Beith-Luis-Nion, He-Lamedh, Abgad und Alphabel, Ober die Syslemalikder Zeichenaufztihlung bei Buchstaben-Schriften in FS Untermann 1993p. 411-444; Old English Runes 1991; Frisian Runes 1996 (which includes,among others, A. Barnmesberger Frisian and Anglo-Saxon Runes: FromIhe Linguislic Angle p. 14-23; bibliography p. 22f.); Barnmesberger Pforzenund Bergakker 1999; A. Griffiths in IF 104 1999 p. 164-210.I) The main representative of the East Germanic branch is Gothic.Extant attestations include, on the one hand the above-mentioned Bibletranslation, on the other, a few trade documents from the 6'h century AD.Apparently, an emmissary of the Holy Roman Empire, Ogier Ghislain deBusbecq, was able, during a stay in Constantinople between 1554 and1562, to assemble a list of 86 Gothic words (referred to as 'CrimeanGothic'). Unfortunately, only small fragments ofBurgundian and Vandalichave survived.Literature on Gothic: -7 RGA 12 1998 see under the headings, GOlen, GotischeSchrift, and gOlische Sprache (various authors); Krause Handbuch desGotischen 1968; Braune / Ebbinghaus Gotische Grammalik 198 I; Feist GothischesWiirterbuch 1939; Binnig Gotisch 1998. On the phonetic system, seebelow, K. Dietz in L 222 § 5.2) In the North Germanic, or Scandinavian branch (Old Norse), which isfirst attested in runic writings of the 3'" century AD (which one may call 'EarlyNorse'; the name Proto-Norse being misleading in the case of an attested language),are to be included Old Icelandic, Old Norwegian, Old Swedish, andOld Danish, which could all be distinguished from one another at the time ofthe first preserved manuscripts in the 12" century. Old Icelandic and OldNorwegian are classified as Old West Norse; Old Swedish and Old Danish, asOld East Norse. The most attested in terms of literature is Old West Norse(since the 9" century AD), particularly Old Icelandic; that is why Old lce1andicforms are traditionally cited.Literature: -7 Noreen Altisltindisch and Altnorwegisch 1923; NoreenAltschwedisch 1904; Vries Altnordisches etymologisches Worterbuch 1962;Liihr EgiU 2000. - Modern Scandinavian languages: -7 BraunmiiUer SkandinavischeSprachen 1992 (with a review by J. A. Haroarson).Overview of the (ndo-European Languages and their Sources 373) West Germanic, of which only a few runic writings remain, is composedof Old English, Old Frisian, Old Saxon (i.e. Old Low German) and Old LowFrankish (i.e. Old Netherlandish) on the one hand, and Old High German, witha second consonant shift, on the other. Old English, Old Frisian, and OldSaxon are placed under the rubrics, North Sea Germanic or Ingvaeonic. OldEnglish is attested from the early 8" century AD, Old High German from thelate 8'h century, Old Saxon from the 9" century AD, Old Low Frankish fromthe 10'h century AD, and Old Frisian from the 13" century AD.Literature: - a) Old English (Old Saxon): -7 Brunner AllenglischeGrammalik 1965; Krogh Stellung des AllStichsischen 1996. - b) OldHigh German and German: -7 Sonderegger Althochdeutsch 1987; LiihrHildebrandlied I / II 1982; Seebold Etymologie 198 I p. 731f. (Die deutscheSprache); Riecke jan-Verben 1996; Sprachgeschichle I 1998 and 21985; dtv-Allas Deulsche Sprache 1998; Schwerdt 2. LV2000.E 433. The Slavic group of languages may be divided into three subgroups:South Slavic (Bulgarian, Macedonian, Serbian, Croat, Slovenian),East Slavic (Russian, Belarusian, Ukrainian [Ruthenian], and West Slavic(Polish, Upper Sorbian, Lower Sorbian, Czech, Slovakian, [{Dravano}Polabian {Liineburg Wendland}, Pomeranian {Pomeranian Baltic Coast},Slovincian {which, as the last representative of Po me rani an, died out in the20th century}, Kashubian {still spoken today as a dialect}]). The oldestextant Old Church Slavic document is not written in a linguistically unifiedform, but is rather marked by the individual languages Bulgarian, Serbian,and Russian. Additionally, it does not contain a perfectly representativeexcerpt in any one of the concerned Slavic languages.South Slavic: The oldest attested Slavic language (second half of the9" century AD) is the Old Church Slavic that was employed by the Slavicmission in Mahren to translate Christian texts from Greek, and which isfounded on a Salonikian dialekt. Due to predominantly Bulgarian qualitiesof the Salonikian dialect, Old Church Slavic is also called Old Bulgarian.In order to establish a system of writing, the Slavic apostle Cyril (originallynamed Constantine) created, on the basis of the Greek lower-case alphabet,the separate Glagolitic alphabet, which was replaced around 900 ADby the CyriUic alphabet which is based on the capital letters of the Greekalphabet. The earliest inscriptions date from the 10'h and 11 'h centuries AD- Middle Bulgarian begins in the 12'h century. - Serbo-Croatian hasbeen transmitted since the 12'h century in Church Slavic texts with SerboCroatian characteristics. Two different alphabets are used, namely Cyrillicin Serbian texts of the orthodox church, and Glagolitic in the Croatian

38 Introductiontexts of the Roman church. - Slovenian has been continually attestedsince the 15th century; the oldest Slovenian linguistic evidence is fo und inthe Friesinger records of about 1000 AD.East Slavic: Transmittal of Russian begins in the middle of the I1 thcentury AD through documents in Church Slavic which bear characteristicsof East Slavic. Russian uses its own system of writing, the Cyrillicalphabet. - Belorusian and Ukrainian have both been attested since the12th century on Old Russian language monuments, which bear dialectalqualities.West Slavic: Polish has been attested since the Ith century. Polabianwas spoken in the area of the lower reaches of the Elbe; it died out inHannover in the 18th century.Literature: - a) General: Brauer Slavische Sprachwissenschaft1961-1969; Panzer Slavische Sprachen 1991; Pohl Le balte et le slave inlangues indo-europeennes 1994 p. 233-250; Rheder Slav. Sprachwissenschaft1998. - b) Old Church Slavic: Leskien Handbuch der AltbulgarischeSprache 1962; Aitzetmiiller Altbulgarische Grammatik 1978;Koch Old Church Slavic Verbum I / II 1990 (I p. 17f. A good sketch ofthe history of Old Church Slavic).E 434. The Baltic group is composed of three languages: Lithuanian,Latvian (East Baltic) and Old Prussian (West Baltic). Other Baltic languagesmay only be researched through borrowed words in Lithuanian andLatvian: among others, Jatvingian and Curonian.The oldest Baltic linguistic record is the Elbinger lexicon of the beginningof the 14th century AD. It contains 802 Old Prussian equivalents ofOld Middle German words. The oldest Baltic text is Old Prussian as well;it comes from the middle of the 14th century AD and includes only elevenwords. Old Prussian textual material is otherwise also very limited. Thelanguage died out in the 17'h century AD.The first Old Lithuanian and Old Latvian texts come from the 16'h centuryand appear already in book form. The oldest Lithuanian book is acatechism from 1547 translated into Lithuanian. The earliest Latvian textis a translation of the Lord's Prayer from the first half of the 16th centuryAD.In the case of Lithuanian, one can distinguish two dialects, LowerLithuanian (Zemaitisch) in the Northwest of the Lithuanian language-area,and High Lithuanian (Aukstaitisch); both dialects may in turn be dividedinto subdialects.Overview of the Indo-European Languages and their Sources 39Literature: - a) General Baltic: Stang Vergleichende Grammatik1966; Baltische Sprachen 1994; Baltistik 1998. - b) Lithuanian: SennHandbuch der lithauischen Sprache 1966; Fraenkel Lithautisches etymologischesWorterbuch 1962-1965; Bammesberger Abstraktbildungen1973; Petit Lituanien 1999.E 435. Within the group of Indo-European languages, some individuallanguages are more closely associated with one another owing to morphologicalor lexical similarities. The cause for this, as a rule, is a prehistoricgeographic proximity (perhaps even constituting single linguisticcommunity) or a common preliminiary linguistic phase, a middle motherlanguagephase, which would however then be posterior to the period ofthe mother language.In the case of Anatolian, the question is asked inversely. Did it separatefirst as a language branch from Proto-Indo-European, and to what extentwas it thus spared developments common to the remaining Proto-IndoEuropean language group? See below §5.On the general problematics: Porzig Gliederung 1954; Ancient lEDialects 1963 [1966] (which includes, among others, H. M. Hoenigswald,"Criteria for the Subgrouping of Languages" p. I ff); E. Seebold in RGAII 1998 p. 289ff; G. Klingenschrnitt Die Verwandtschaftsverhaltnisse derindogermanischen Sprachen in Kolloquium Pedersen Copenhagen 1993[1994] p. 235ff. (On Anatolian, among others.); W. Hock, Balto-Slavisch,lndo-Iranisch, ltalo-Keltisch: Kriterien Jur die Annahme von Sprachgemeinschaftenin der lndogermania in Aspekte baltistischer Forschung2000 p. 119-145.1) For more information on similarities between Greek, Armenian, andPhrygian (which all likely come from the same geographical area): NeumannPhrygisch und Griechisch 1988; Clackson Armenian and Greek 1994.2) Notions of Halo-Celtic, which was emphasized in earlier research mustbe modified. In all probability, there was no Italo-Celtic preliminary phase.Rather, Celtic contacts with eastern Indo-Europe are ancient. Compare thecase, among others, of relative pronouns, which in Celtic, contrarily to theItalic *kwo_ / *kwi-, is represented by *Hjo- (cf. F 404), a characteristic that itshares with Greek, Phrygian, Indo-Iranian and Slavic. Celtic contacts withItalic are of a later date: C. Watkins "/talo-Celtic Revisited" in AncientIndo-European Dialects 1963 [1966] p. 29-50 (also published as SelectedWritings I 1994 p. 105-126); K. H. Schmidt "Late in und Keltisch " in Kollo-

42 IntroductionE 501. The first example focuses on phonology and is often used in academicinstruction: -t M. Mayrhofer in AOA W 1 17 1980 p. 364.I) The phonetic deviation from the Latin nominative singular nix (with

44IntroductionThe Reconstruction of Pro to-In do-European 45recognizeab1e in the PIE *bher-e-ti 'he carries', cf. F 203 § I with LIV typeIn).5) A clear example, parallel to the Vedic as-, is the Vedic appears in place of i.8) Observations regarding the endings of the as- and . paradigms: The Isg. -m-I, the 2 sg. -s-i, the 3 sg. -I-i and the 3 pi. -anl-i end with -I, as opposedto the I pI. -mtis and the 2"" person plural -IM, where there is no -i. One mustadd, however, that the I pI. -mas-i exists in Vedic as a later variant. Theseendings are indicative of the present tense and are referred to as 'primary endings.'See below § 11.Although not all of the twelve forms of bhar- are attested, they aresafely deducible: Macdonell Vedic Grammar 1910 p. 319 ff. ; concerningthe verb: Gota 1. Priisensklasse 1987 p. 225 ff.; on the thematicvowel -a- / -a- « PIE *-e- / *-0-), see below, F 101 § 4.11) The contrast between present and non-present is established throughtwo different formal means: On tbe one hand through the augment, which isplaced before the verb stem in the past tense (for more information, see below,F 213), on the other hand through the two series of endings: primary endingsfor the present, and the secondary endings fur the non-present. In the I", 2"",and 3'" singular and 3'" plural, the presence or absence of -i differentiates thetwo series, whereas in the I" and 2"" person singular, the endings -mas( -I) vs.-ma and -Iha vs. -Ia(na), which differ from one another only slightly, perfonnthis task. In place of -nl-i : *-nt, in the 3'" person plural, the later -nli: -n isfound (with simplification of -nl to -n in word-final position).12) The Vedic imperfect forms of as- (for information on tbe status of researchon attestations: Gota Malerialien Nr. 3 1990 under appropriatebeading; the Vedic verb forms are only given accents wben they are attested inaccented form; the verb in the main clause, as is generally known, nonna11ycarries no accent; for further information, see below F 211):

46 lntroductionattestedexpected1 sg. asam *a-as-lfl2 sg. as's *a-as-s3 sg. lis, &sit *a-as-t3 pI. asan *a-s-ant13) The investigation of the theoretically expected fonns in § 12 is based onthe analysis from § I in combination with the structure of a-bhar-a-m in § \0.Unproblematically comprehensible, the initial sound in the I' person singular,in which the augment a-, with the full grade a in the root syUable a, throughcontraction directly produces as-. Additional information regarding the ending-am: On the basis of the thematic I' sg. -a-m, in the case of the athematicending following a consonant-final verb stem, one expects to find *-,!" orrather the * -a that phonetically results from it. The -am that is attested in itsplace, may thus be categorized as a form arrived at through analogical differentiation.14) The 2 sg. *a-as-s and the 3 sg. *a-as-I both lead to *as with a contracteda and final Os, because the final consonant gloupS *-ss and *-sl are bothsimplified according to phonetic laws and result in simple -so The rarely attested3 sg. as is thus old. The extent to which it belonged to the contemporarylanguage is questionable. The formal similarity between the 2" " personsingular and the 3' " person singular cannot have been practicable in the longterm, and must have led to a new differentiation. - The new fonns with the2" " person singular as-i-s and the 3 sg. as-i-I appear already in the Rgveda.They evidently have their model in verbs such as brav ; - 'to speak, say, talk' «PIE *-e1jlf-: For further information on the so-called se!-roots, see below L315 § I). There, for example, the form of the subjunctive present active 3 sg.is brav-a-I « PIE * --e-t; the -e- is the subjunctive suffix, see below F 207§ I). In contrast, the form of the indicative present active 3 sg. is bravi-ti «PIE *-eyH-ti; the interconsonantal laryngeal is represented by i in IndoIranian, see below L 325 - L 327) and the form of the indicative imperfectactive 3 sg. is a-bravi-t « PIE *-eyH-t). Parallel to brav-a-I and to a-braY-lot(which is result of new analysis, replacing a-bravi-t), an imperfect active 3 sg.as-i-t could be created on the base of the similar subjunctive present active 3sg. as-a-I. On the unexpected a- of the 3 pI. asan, see below E 504 § 10.The Reconstruction of Proto-Indo-European 47E 503. In a second (this time language-external) step, the correspondingfonns from the classical Latin paradigm are introduced for comparison inorder to establish further clarity.I) I have again chosen the tabular form for the relevant paradigm (For informationon the details: -+ Leumann LLFL 1997 § 400 A):FonnsReconstructionI sg. sum < *es-miinscr. esom2 sg. es < *esi < *es-si3 sg. esl < *es-tiI pI. sumus < *s-mos2 pI. estis < *s-Ie-s3 pI. sunt < *s-ontiinscr. sont2) The reconstruction is based upon analysis of the Vedic paradigm. However,Latin e-vocalism predates Indo-Iranian a-vocalism. See below L 206. Inthe case of the endings (-+ Leumann LLFL 1977 p. 512ff.), -mus results from*-mos, -ti-s from *-te-s, and -unl via -onl from -onti (perhaps preserved in theIremonti, quoted by Festus: -+ Leumann op. cit.p. 92). Next to -enti, -anti issecondary, see below, § 1 1.3) The classical Latin paradigm is most easily understood if one uses the 3sg. est and the 3 pI. sunt as a starting point. The verbal root reveals in boththese fonns the Ablaut change expected from Vedic, from full grade, singulares- to zero gtade, plural so.4) In contrast to the Vedic endings (3 sg. -I-i and 3 pI. -anl-l) the Latinendings (3 sg. -t and 3 pI. -ont) are missing the short final, unstressed i-vowels.While Vedic conserves these, they are eliminated in Latin: -+ Leumann LLFL1977 p. 92 and see below, L 423 § I. Note: The elimination of the -i is mucholder than what is called rhotacism, which, in the fourth century BC, wouldhave to have changed a remaining *esi into 'eri, see below L 309 § I. - Thatwhich applies to the 3' " person singular and plural, also applies to the 2 sg. esas it relates to the postulated Proto-Italic *esi.

48 lntroduction5) In the singular paradigm, sum stands out in relation to the expected ablaut.Starting from the Vedic, first Proto-ltalic 'esmi, or at least 'esum is tobe expected. The newly discovered South Picene esom ( Marinetti lscrizionesudpicene 1985 p. 214) confirms this hypothesis. An inscribed Latin esomwas also found recently: M. Cristofuni in Quademi di ArcheologiaEtrusco-Ilalica 25 1996 p. 21. One may thus quite rightly accept that an existingProto-ltalic 'lis-mi became esom after it eliminated the finaI -i from 'esmand a weak vowel was inserted, via the form 'es"m: H. Rix in KolloquiumLalein und lndogermanisch Salzburg 1986 [I992) p. 230. The laterreplacement of esom by sum came about through the analogy of the I" pe=nplural sumus. There, the cause is the regular, zero grade, 's-mos (in which,the initial ' sm- > 'som- > sum-).6) In the plural paradigm, the initial full grade es- of the 2"" pe=n pluraleslis stands out. It should be regarded as analogous to the 3 sg. esl.7) The new analogous relationships between the first J"' =n singular andplural (§ 5) and between the 3'" person singular and the 2 person plural (§ 6)show that the ablaut change between the singular es- and the plural SO, at thetime of the replacement of the Latin esom by sum and of' sle- by eslis was nolonger in use as a living tool for the formation of verbs.8) Obscure areas of older ablaut structures may also be observed in theclassical Latin paradigm i- 'to go' (On the material: Leumann LLFL 1977 §399):FormsReconstructionI sg. eo < ·ej-a2sg. is < 'ei-si3 sg. il < *ej-liI pI. imus < *i-mos2 pI. ,lis < *i-/e-s3 pI. eunl < *j-enti9) The forms on the right may be postulated starting from Vedic (seeabove, E 502 § 5). Taking into account the typical Latin monophthongizationof ei > , (see below, L 220 § I) as well as the changes that we have alreadyobserved in the area of endings in the case of sum (see above, § 4), the 2 sg. is

50 Introductionfonusfirst prelirninaryreconstructionI sg. etllL < *es-miT2 sg. et < *esi < • es-si3 sg. E

52 IntroductionGreek/Doric -ti. 1be ending of the 3' " person plural in -am may be comparedwith eam from § 5.9) The questionable Greek peculiarity of the constant es- fo r full grade andzero grade, may most easily be explained by accepting that the Proto-IndoEuropean verbal root *es- originally contained a word-initial laryngeal, andthus read *h/es-, but otherwise was conjugated in a normal proterodynarnicfushion. Full grade *h/es- and zero grade * h,s- both produce in fuct, accordingto phonological rules, es- in Greek, see below, L 322 § I and L 328 § I.Full (e-) grade Greek *h/es- lives on after the disappearance oflaryngeals before a vowel as es-, whereas the word-initial consonance of the zero gradeGreek *h/s- developed with the help of a weak vowel via *h/os_ > *h/es- >also to es-. - This is not the case in Vedic and Italic, in which the full gradees- stands in clearly visible contrast to the zero grade s-: While the full grade"h/es- does resuh, parallel to Greek, in Vedic as- (on the intermediary grade*es-, see below, L 206) and the Latin es-, the zero grade *h/s- leads, in contrast,through simplification of the double consonance, from "h/s- to So. Thesingular of the Vedic participial negation &sat- (with a long initial vowelRgveda IV 5, 14 and VII 104,8; also attested several times with a short initialvowel asal-) 'not being' proves indirectly that at the time of this constructionthe initial double-consonance must still have been present, for this alone explainsthe stretched initial secondary form: PIE *IJ-h/s-nt- > Proto-Indo-•.. .LIranian * a-Hsal- > Vedic asat-. The short vowel is to be understood in contrastas a later analogous form of the participial sal-, with the usual prefix ofnegation 0- < * Q-.10) The postulation of PIE * h/es- leads to an adequate understanding ofte Vedic unperfect and, in SO doing, confirms its correctness as in the case ofiisant- in § 9. It was namely always noticeable that the Vedic singular andplural augmented forms exhibited equally lengthened as-, although the language-internalcomparison does not at first lead one to expect this, and, in theplural, should lead to a short initial vowel, see above, the table in E 502 § 12with the analysis of the I sg. &sam < *a-aS-Tf! and the 3 pI. &san < *a-s-ant, inwhich the reconstruction cannot explain the long initial vowel. - While in thecase of a postulated Proto-Indo-European root form "es- / *s- one could expecta first grade Proto-lndo-European singular *e-es- / plural *e-s- or ratherProto-lndo-Iranian singular iis- / plural *as- and an analogous transfer oflength into the plural, the PIE *h/es- / *h/s- produces directly the attested,The Reconstruction of Proto-lndo-European 53unified paradigm: PIE *e-h/es- > Proto-Indo-Iranian *il-Has-' > *a.as- > iisandthe plural *e-h/s- > Proto-lndo-Iranian *a-Hs- > iis-.11) Contrary to today's research and its laryngeal *h,es-, earlier research,with its postulation of the simple non-larygeal *es- /*s-, had considerable argumentativedifficulties. As representative of the communis opinio of thetime, one may cite Thumb / Hauschild Handbuch des Sanskrit I / 2 1959 § 488(as well as Leurnann LLFL 1977 p. 522 § 400.A.lb). Thumb and Hauschilddraw a direct parallel between the Vedic sanli and the Doric Greek Ev"ti. andtrace it via "henti back to "senti. Through analogy to the singular, the expectedaspiration is eliminated. Correspondingly, 10niclAttic Greekelai would be traced back to *hensi or rather *sensi (with Southern Greek -siinstead of -11). But in the case of the I" person plural "smen and the 2" " personplural "sle the initial e- would be analogously introduced. But the extension ofthe analogous e- to the stems of the optative and the participle cannot be explained.Further, the Mycenaean -e-o-Ie, i.e. -ehontes, known since 1952,fo rces the pushing forward of the analogy in question into the 2nd milleniumBC. 1be communis opinio leaves one utterly helpless before the Mycenaeane- of the 3 pI. e-e-si, i.e. ehensi, particulatly if one, in the cases of Ionicl Atticelai and Doric Ev"ti., believed he could succeed with the use of *senti.But despite all evidence: Even today there are researchers who happilycontinue to use *es- / *s- as a starting point and build hypotheses fromthere: -+ K. Shields "On the Origin oJ Dialeclal Ablaut Patterns oJ IhePresenl Aclive Indicative oJPIE "es To Be"' in HS 110 1997 p. 176-180.E 505. As a result of the first three steps of E 502-504, both of the followingparadigms may with certainty be attributed to Proto-IndoEuropean:Proto-lndo-European*h1es- 'to exist'*hJes-mi*h/esi < " hIes-si*h/es-ti*h1s-me*h/s-Ie"h/s-entiProto-lndo-European*hJej- 'to go'*h1ej-mi*h/ej-si*h/ej-ti*h/i-me*h/i-Ie*h/j-enti3 Nota bene: Because the laryngeal quality of Indo-Iranian laryngeals is no longerrecognizeable, J note it nonnally with H.

54 IntroductionI) To the given meaning of the PIE ·hJes- 'to exist, to be there' may beadded that this verb must have had this strong meaning in the Proto-IndoEuropean period. Nominal sentences of the type, 'the floor is dry' were thusnot, as we would expect from the standpoint of German, formed with 'is,'which developed as a helping verb, but rather the simple series of PIE·dhei!'i5m with ·trsteh, was enough. See below, S 206 on the nominal sentence.Proto-Indo-European ·hJes- was often clarified with local particles ofthe type, 'ab-wesend,' cf. Latin ab-sent-, Mycenaean Greek a-pe-o-te i.e. apehontes.2) One further remark on ·e1-: It may not be omitted here that the rootoriginally contained a word initai laryngeaJ, and was thus ·hJej-: LIV 1999p. 207. This renders plausible the surprising notion that, as in the case of as- inthe imperfect singular as well as in the plural, despite the ablaut ·e1- : ·i-, bothbegin in Vedic with a long vowel (cf. I sg. 4Yam and 3 pI. 4Yan): I sg. ilyam

56 Introduction5) A word on the construction of the feminine 'mare.' Along with thechoice of words of difierent origins to express masculinity and femininity (e.g.'mother,' 'mther'), another option is inflection, i.e. the changing or extentionof the masculine ending by means of suffixes. For example: the Vedic asviifeminine= Old Vedic and Late Vedic aspii- feminine. So-called communiapresent another possibility. 'Thus are called ... such substantives which denoteanimated, sexually characterized beings, in the case of whih both themasculine and the feminine specimen can be denoted such that the substantiveform does not change, but rather, according to the gender of the specimen, thecorresponding pronoun (or adjective) is in masculine or feminine form." (Forexample, Ohi t and Old Latin lupus flmina 'she-wolf): WackernagelVorlesungen IT 1926 p. 23f., further p. 10 and p. 315. As a rule, the communiaprove to be older than the corresponding feminine inflected forms. Forthat reason it is advisable to ascnbe the commune for 'mare' (*ekyo- m/f.) toProto-Indo-European, but the formation of the feminine *t!kueh,- from theVedic word asvii- to post-PIE languages. - See below L 421 -§ 2'on *snusof.'daughter-in-law.'6) The further analysis of *(hJ)elCljo- is uncertain. It is often grouped togethr the word fumily of PIE *hleK- 'quick' (CompareGreek c.OKU;,Vedic asu-, Latm acu-, 6cior 'quick(er),' and *OK- from the PIE *hlohIK-?)and then analyzed as PIE *h/'ICIj-o- 'provided with *hlelClj- ('quickness')': Mayrhofer EW Aia I p. 140 and p. 179f.; Rix Termini der Unfreiheit 1994 p.10 and .by the same author, in Kratylos 41 1996 p. 156 (on Latin acu-); I.BaIles m HS 110 1997 p. 220 note 8. However, this cannot be rigorouslyproven.7! Concerning concrete fuets, e.g. whether the wild horse living in the forestsIS meant, see below E 512 § 3 and 4b.2. Fundamentals of ReconstructionE 507. The individual Indo-European languages provide the point ofdeparture for .Indo-European linguistics. All of these languages showeqUlvalences m great number in all possible sub-areas. The equivalencesare often very extensive: As a rule, individual lexemes and morphemes notonly exhibit sar or even identical forms, their content is just as similaror even Idenllcal, for example Vedic as-t-i, Greek ea-1-t and Latin es-t(above E 502ff.), in which not only the external form and construction areunmistakably similar, but also the meaning ('to be') and position of theThe Reconstruction ofProto-lndo-European 57forms in the paradigm (present, strong, third person, singular) are thesame. Vedic asti, Greek ecrti and Latin est offer only one example amongcountless others. With the help of phonetic rules, the equivalences may beproven even where the individual words are outwardly very different. Forexample, compare the Vedic sromata 'honor, good reputation' with OldHigh German hliumunt 'reputation, rumor,' where both evidently, accordingto rules of phonetics, come from the PIE */{;/e!!-mlp-: Mayrhofer EW Aia II p. 667I) The presence of numerous firmly established commonalities within theindividual Indo-European languages can in no way be explained as if thecommonalities were universal aspects of human language.Indeed, there aresurely universal givens, of which the future-oriented linearity of speaking is anexample: On the axis oftime, one may only speak in a forward/future orientedmanner, and not in a backward/past oriented manner. The linguistic signs are,as a rule, not universal, but rather are arbitrary, created through conventionwithin each linguistic community. In the case of 'to be,' the verbum existentiaeis fully different in Semitic languages, for example, Hebrew hiijiih 'to happen,to become, to be,' etc.2) Coiilcidence is equally unsuitable as an explanation. Of course there areastounding coiilcidences, such as the apparent similarity of classical Latin deus'God' to Greek Se&; 'God.' A short examination of the older and oldest occurrencesof both words (Old Latin deivos and Mycenaean thehOs) clearlyshows that both nomina are different from their inception: deivos is derivedfrom PIE *djey- '(God of the daytime) heavens' in the sense of 'heavenly(being),' thehOs « *dhhls-O-) is derived from PIE *dhehls- 'God, divine': Meiser, Laut- und Formenlehre 1998 p. 107. Contrarily, the tonal similarityof the Vedic as-t-i, Greek ea-1-t and Latin es-t excludes the possibility ofcoiilcidence. Accordingly, the older the examples, the more their content andform are similar.3) Borrowing is also not valid as an explanation. Borrowing takes place ofcourse, in the lexical area, for example Latin miichina and Doric Greekaxava or German Kaiser and Latin Caesar. In both cases the routes of theborrowings are known; in the former case, miichina was taken by the Romansfrom Dorie Greeks of Lower Italy. The original *miicana was integrated intothe Latin lexicon and later, through the typical Latin vowel weakening,changed to *miicina, and finally established as miichina, with the h to indicatethe aspiration of the original Greek. In the latter case, Kaiser was given by theRoman dictator Caligula (officially C. Caesar) around 37 to 41 AD to theGermanic peoples, from which came his bodyguards at that time. ( H. Rix

58 Introduction"Latein - wie wurde es ausgesprochen?" in Beitrage zur miindlichen Kulturder Romer, edited by G. Vogt-Spira. Tiibingen 1993. p. 14).4) Alone the hypothesis of the derivation from a common preliminary phaseof the language groups in question is a suitable explanation of the numerousequivalences in the areas of phonetics, construction, and semantics. In otherwords: Vedic as-i-i, Greek eO"-'-1 and Latin es-t are related to one anotherand are based upon the reconstructible common preliminary PIE *h,esti (seeabove, E 505). This language was spoken by the Proto-Indo-European linguisticcommunity. No system of writing was known to this community.From the Proto-Indo-European form * h,esti follow individual routes througha continuum of tradition and generations of speakers to Latin, to Greek, and toVedic forms:Proto-Indo-European *h,estiVedic asti Greek eO"n Latin est5) For the sake of simplicity in representation no tree tables are used, butrather a correspondence set is used in its place as an abbreviation: PIE * h ,esti= Vedic as-t-i = Greek eO"-'-1 = Latin es-t. One must not misunderstand anddraw the conclusion that the Vedic asti comes from the Greek ecrn, or thatthe Latin est comes from Greek.E 508. Since the time of Bopp (see above, E 302), Indo-European linguisticshas been a methodically established and functional science. Itsprimary method is the internal and external comparison within, and amongthe Indo-European languages, as well as the constant review of the relationof the Indo-European languages to the original language and viceversa.I) The specialized knowledge in all questions of Indo-European 1inguisticsand in the reconstruction of assertions about the original language gainedthrough comparison has in the meantime become very comprehensive andcomes from the work of generations of Indo-European linguists. Owing to thedialogue in the international community of researchers, this specialized knowledgeis subjected to an ongoing process of review. New finds and new ideascontinually show that we are on the correct path with our specialized knowledge.Time and time again, new pieces, some larger, some smaller, can beThe Reconstruction of Pro to-in do-European 59added to the puzzle of Indo-European 1inguistics. It attests to the accuracy ofour progress when a form that had been unclear on the hasis of earlier researchsuddenly is understood without problem.A small example from my research: The Epic Homeric verb formea81j (in aO"1tt ea81j Iliad N 543 and ::: 419) has been unclear sinceantiquity and has given occasion to untenable speculations (-7 ChantraineDELG sub voce: "rien de clair"). In another context entirely, I asked myselfone day, what the Greek equivalent of German singen is. Known tome from textbooks was the associated nominative active-voiceOJ.Li] (Epic-Poetic Greek since Homer) '(divine) voice' « *songwhti).The underlying verb is believed to be lost. (-7 Frisk GEW s. v.: "das zugrundeliegende primare Verb ist nur im Germanischen erhalten"). Anidea struck me: as an aorist, ea811 belongs to this lost verb « *e-s(lgwh_).The meaning (the word occurs in two battle scenes) fits remarkably well:"(upon him) sounded the shield." For the verbal form in question, thismeans that it was translated from the Homeric singers in the correct context,but was very early no longer understood. On the details of the explanation,see below, L 345 § l.2) Thanks to the numerous comparisons of equivalent forms we are thebest and most accurately informed in the problems of phonetics. Thus, theform of the Indo-European root * snejgwh_ is for certain correctly determined,see above E 50 I. However, while the determination of its meaning as 'tosnow' is rather certain, the question of how this relates to the Vedic meaning'to be sticky' cannot be answered with certainty. The degree of certainty ofassertions varies in semantics just as it does in morphology and still more insyntax, see below S 101. Each case must be checked individually. For example,PIE *h,es- (see above E 505), of which the reconstructed 3 sg. *h,esti isquite probably the form of the original language, whereas in the I pI. *h,s-metheexact form of the present ending cannot be ascertained.E 509. The reality of the linguistic continuity of traditions from the postulatedProto-Indo-European original language to the known individualIndo-European languages of historical times cannot be doubted. To whatextent linguistic relation implies familial clan and a common genetic originof the dispersing groups of speakers remains to be discovered. Concerningnecessary information as to the location, time and culture of this originallanguage, as well as ideas on the 'how' of the genesis of the individuallanguages, see below E 511 ff.

60 IntroductionI) However, our reconstruction has its very clear limits. We are decisivelydependent on extant attestations of the individual languages.This varieswidely, see above E 400. Just imagine the situation if we had only contemporaryGreek and the first Middle GreekIByzantine attestations after the full ofConstantinople in 1453. Greek, which is extremely well documented, wouldhold a position similar to that of Albanian within the Indo-European realm, andwould not have the weight that it has long heen accorded, owing to is! uniquedocumentation in the Mycenaean and Ancient Greek periods.2) Our reconstructions are not fixed for all time. Depending on argumentation,judgements may differ from researcher to researcher.Further, newinformation and finds can certainly require one to make corrections. However,corrections of both kinds do not mean that the postulated language haschanged, rather only that we have adapted our assertions to the latest level ofresearch.3) Our reconstructions lead us indeed directly from the lexical ofiSpring inquestion to the underlying uniform phase of language that last unified them.But the route from the last uniform language phase to the individual languagewas not as straight as is necessarily suggested by the existing forms, whichhave prevailed in the process of linguistic development."(One can) say thatcomparative reconstruction determines the average quantity and not the commonquantity ... of the linguistic phenomena concerned, and thus only grasps apart of reality.": -4 H. Eichner in Akten 13. Qsterreich. Linguistentagung1988 p. 15. "We must not delude ourselves into believing that our retrogressivemethod of reconstruction matches, step by step, the real progression oflinguistic history" : -4 E. Pulgram "Proto-Indo-European Reality and Reconstruction"in Language 35 1959 p. 423. Time and time again, there must haveheen forms, for example, that were once a part of the lexicon or grammar, butthen became uncommon and disappeared.Only in the best case do we haveknowledge of them, such as when, for example, they are handed down in lexicalizedform. We can reconstruct, for example, that there was a Proto-IndoEuropean nominal root *"]f!nt- '(side of) fo rehead.'While it is still alive inHittite, other Indo-European languages such as Greek only show it in lexicalizedforms, e.g. Qv'ti. 'in fuce of7towards' (which is used not only as a preverband preposition, but also as a prefix in compound fo rms and is formally a fossilizedlocative singular form of the fo rmer nominal root: -4 Friedrich I KarnmenhuberHW III p.1581I See also L 322 § 2.Further literature on reconstruction: -4 M. Mayrhofer Ober sprachlicheRekonstruktionsmethoden in AOA W 117 1980 [1981] p. 357-366; D. M.Job Zur Bewertung von Rekonstrukten in GS Kronasser 1982 p. 46-71;The Reconstruction of Proto-lndo-European 61Akten 13. Qsterreich. Linguistentagung 1988 (p. 7ff. Contributions to thepodium discussion: "Sprachwandel und Rekonstruktion," among whichthat of H. Eichner (p. 10-40] is particularly fruitful); Language Typology1988 [1991] (among others, H. M. Hoenigswald 'MorphemiC Change,Typology, and Uniformitarianism: A study in reconstruction ' p. 17-26; E.P. Hamp "On Reconstructing Morphology and Syntax " p. 105-110). Formore information see below, end of § 4.4) It is important to recognize that linguistic change is not a plann prcess, according to which a language is constructed m an orderly fushion., Alanguage such as German, Swahili, or Italian is a spontaneous systemIt ISthe non-intended result of human interaction": -4 R. Keller "Sprachwandel,ein Zerrspiegel des Kulturwandels?" in Kulturwandel im Spiegel des .Sprachwandels, Achtes Partnerschaftskolloquium 1991 in Diisseldorf, editedby K.-E. Lanne. Tiibingen IBasel l995 (= Culture and 0:-vledge 11) p. 13;by the same author, "Zur Erklarungskraft der Naturilchke.'tstheone mSprachwandel and Sprachgeschichte. CommemoratIVe publicatIon for H.Llidtke, edited by J. Schrnidt-Radefeldt and A. Harder, Tiibingen 1993, p.109-1 16; see also by the same author, Sprachwandel l994 (subtitle: Von derunsichtbaren Hand in der Sprache).Further literature on linguistic change and reconstruction: -4 H. Eichner(For title, see above at end of § 3); Linguistic Change and ReconstructionMethodology 1990 (which includes, among others, H. M. Hoenigswald"Is the 'comparative ' method general or family-specific? " p.375-383; Diachrony within Synchrony 1990 [1992] (which mcludes,among others, A. Barnmesberger "Phonology, analogy, and how languageschange: Notes on the development of some grammatical categoriesin English " p. 359-375; H. M. Hoenigswald "Semantic change and'regularity ': A legacy of the past" p. 85-105); "Explanation in ;listoricalLinguistics " 1992 (which includes, among others, R. Anttila , HistOrIcalexplanation and historical linguistics" p. 17-39; B. D. Joseph "Diachronicexplanation: PUlling speakers back into the picture " p. 123-144); HIStoricalLinguistics 1993 (which includes J. Anderson "Parameters ofsyntactic change: a notional view" p. I ff. ; B. Comrie "Typology and reconstruction" p. 74ff.); R. Gusmani, "Ursprache, Rekonstrukt, hermeneutischeModelle" in FS Meid *60 1989 p. 69-77 (further literature may befound in note I.)E 510. The language that is analyzed with the help of reconstruction, theIndo-European mother language, called Proto-lndo-European, should notbe considered a 'primitive' original language.Rather, it was a fully 'nor-

62 Introductionmal' language, with its own long prior history, and was spoken, to ourknowledge, by a linguistic community that had no system of writing.I) 1be resulting reconstructions of Pro to-Indo-European lay on a uniformline and may necessarily provide only a one-sided portrait, without spatial andhistorical perspectives (see above, E 509 § 3). "It is the nature of every reconstruction,that in it all diachronic, dialectal and other qualified differences beneglected": B. Schlerath in ZVS 95 1981 p. 180. Of course, whereasProto-Indo-European was concretely a living language with dialects and acomplex structure , the only Proto-lndo-European that we can attain, i.e. thereconstructed language, can furnish us with no clues about this. It is thusmethodologically erroneous to equate the reconstructed Proto-Indo-Europeanto the concrete Proto-Indo-European, and to place it in a time/space model. Idirect this negative conclusion toward the discussion surrounding the'time/space model' proposed by W. Meid and shall not conceal that the counter-argumentby B. Schlerath has convinced me: W. Meid "Probleme derrtiumlichen und zeitlichen Gliederung des 1ndogermanischen" in FachtagungRegensburg 1973 [1975] p. 204-2 19; B. Schlerath "1st ein Raum!Zeit-Modellfor eine relwnstruierte Sprache moglich?" in ZVS 95 1981 p. 175-202; by thesame author, "Sprachvergleich und Relwnstruktion: Methoden undMoglichkeiten" in 1ncontri Linguistici 8 1982-1983 p. 53-69; 1ncontri Linguistici9 1984, which includes on p. 63ff. a debate on the subject, sketched byB. Schlerath in Volume 8 (written by E. Carnpanile, F. Crevatin, M. Doria, R.Gusmani, R. Lazzeroni, E. Neu, P. Ramat, K. H. Schmidt and K. Strunk.);Incontri Linguistici 10 1985 yields on pages p. 11-18 the conclusion of B.Schlerath, "Probleme der Relwnstruktion: SchlujJwort und AusbliclC'; J.Tischler, "Bemerlcungen mm 'Raum-Zeit-Modell ·' in FS Meid *60 1989 p.407-429; W. P. Lehmann "Earlier stages of Proto-Indo-European " in FSMeid *60 1989 p. 109-131; F. R. Adrados "The new Image of 1ndoEuropean" in IF 97 1992 p. 1-28; E. Seebold in RGA 15 2000 under theheading, Indogermanische Sprache und Sprachfamilien.2) 1be concrete Proto-Indo-European must itself be the product of a longhistorical development. Several assertions about early phases of Proto-IndoEuropean may be gained from language-internal comparison ( H. RixModussystem 1986 p. 6f.):"Whereby comparative reconstruction is based upon a group of sirnilarforms in a number of languages, internal reconstruction takes its point ofdeparture from irregularities or inhomogeneities of the system of a singlelanguage . ... The fundamental supposition of language-internal reconstructionis that such an irregularity or inhomogeneity in the grammar of a Ian-The Reconstruction of Proto-lndo-European 63guage is the result of a diachronic process, in which an older pattern, orhomogeneity is eclipsed, but not fully supressed . ... 1bere is no reason torenounce the use of language-internal reconstruction, which was developedparticularly in corpus linguistics, on knowledge of Proto-lndo-Europeangained through comparative reconstruction. 1be use of the epithet, 'glottogonicspeculation,' a term occasionally used to defame language-internalreconstruction, is, even in the case of Proto-Indo-European, not merited,because the process neither can nor purports to investigate the creation ofhuman language."Rix works back from Late Proto-Indo-European Phase B (reconstructibleProto-Indo-European) using deducible information about an Early Proto-IndoEuropean Phase A, and gathers in his work related evidence on the ProtoIndo-European verbal systemThe organizers of the Fachtagung Ziirich 1992 [1994] planned to researchthe route from Proto-Indo-European to Early Proto-IndoEuropean, but the subject was not treated sufficiently: G. E. Dunkel, p.VIIIf.; by the same author, "Early, Middle, Late Indo-European: Doing itMy Way " in lncontri Linguistici 20 1997 p. 29-44.3. The Time, Place, and Culture of the Proto-Indo-EuropeansE SIl. No precise statement concerning the exact time period of theProto-Indo-European linguistic community is possible. One may onlystate that the ancient Indo-European languages that we know, which datefrom the 2'd milIenium B.C., already exhibit characteristics of their respectivelinguistic groups in their earliest occurrences, thus allowing one topresume the existence of a separate and long pre-history. For example, te-0, i.e. theMs, 'God' was already common in Mycenaean Greek. Basedupon finds in Indo-Iranian and Italic, one must conclude that *dejyo- wasone of the Proto-Indo-European terms for 'divinity' (see below, W 202§2), which was replaced in Ancient Greek by ae6. The period of 5,000BC to 3,000 BC is suggested as a possible timeframe of a Proto-IndoEuropean language.E S12. Efforts to determine the region and culture of the Proto-IndoEuropean community offer special difficulties. In the case of both questions,assertions in pre- and early history play a decisive role. Further information may come from the reconstructible lexicon, as well as from po-

64 IntroductionThe Reconstruction of Proto-lndo-EuTopean 65tential conclusions drawn from the later geographical region of the linguisticcommunities of the individual Indo-European languages: -? J. Unterrnann,"Ursprache und hislorische Realittil" in Elhnogenese 1985 p. 133-164 (which includes the chapters: "Die Iheoretischen Grundlagen derhistorisch-vergleichenden Sprachwissenschaji und das Stammbaummodelf';"Indogermanistik und Vorgeschichtsforschung"; "InnersprachlicheEvidenz fur vorgeschichtliche Ereignisse" with the subtitles "Die Krilikam Stammbaummodell," "Wortinhalte als Gegenstand der Sprachvergleichung"and "Ethnische Strukturen im Wortschatz der indogermanischenGrundsprache ?"); by the same author, in Kratylos 34 1989 p. 48ff.("Forderungen und Bedenken, die meines Erachtens bei der ErschliefJungder 'indogermanischen ' Kultur, Geisteswelt oder Gesellschaji zu beachtensintf' in the context of the review of Studien zum IndogermanischenWortschatz, 1987). Further basic information: -? A. Scherer"Hauptprobleme der indogermanischen Altertumskunde (seit 1940)" inKratylos 1 1956 p. 3-21; by the same author, Indogermanische Altertumskunde(seit 1956), loc. cit. 10 1965 p. 1-24; W. Dressier "MethodischeVorfragen bei der Bestimmung der 'Urheimat"' in Sprache II 1965 p. 25-60; Zimmer Ursprache 1990.I) As a rule, the early historic and Prehistoric finds do not coihcide with thelinguistic facts: -? B. HfuJsel in FS &hlerath 1992 (1994) p. 26f.:"Linguistic development may be described in steps that, although logicallycomprehensible, are not precisely analyzable without a timescale. Thearchaeologist pursues certain areas of cultural development, the logic ofwhich (if one exists) remains a mystery to him, or is only accessible in a fewaspects of its complex causality." On the other hand, he is provided withconcrete ideas with regard to time, as vague as these may be, and workswith a concept of culture that tbe Indo-European linguist cannot attain.For the archaeologist, culture is understood in the sense of a sociologicaldefinition, as W. E. Miihlmann puts it (-? W6rterbuch der Soziologie, editedby W. Bemsdorf. Stuttgart 1969. p. 598f.): "the totality of typical waysof life in a population, including the mentality upon which they rest, and inparticular, the value system, whereby the typical ways of life include ...even the technical bases of existence, including a material substrate ofclothing, shelter, tools, equipment, etc." The archaeologist has direct accessto the latter areas, and indirect access to the former. The definition ofour concept of culture in archaeology is fur too open and imprecise in itsborders in order for us to equate archaeological culture with 'a people' or'a closed linguistic community,' in the sense of a politically unified group.Their concurrence is at best imaginable, but a priori unlikely. - We archaeologistsknow about the instability and ephemeral quality of early societies,we know of the most diverse fuctors of social coherence, amongwhich language is but one - an important one - yet one from the realm ofcommunications.We archeologists pursue only sub-areas of cultural developmentwithin open, changing communities witbout clear borders. -Cultural contexts, as they are understood for example in marriage communitiesof particular upper classes, or with regard to burial customs, havevery different borders from, for example, contemporaneous types of settlementsand their relation to their landscapes. But which of these two aspectsof culture and which others are relevant to language? Limited to ourcurrent methods, we shall never be able to answer this question. - Thearchaeological concept of culture is composed of so many components,that by its very nature its contours must remain blurred. But languages arequite different. Of course there are connections; no one can imagine culturalconnections witbout any possibility of verbal communication. But it istoo much to ask that archaeologists equate their concept of culture, whichis open and incorporates references on various levels, to the single dimensionof linguistic community. Archaeology and linguistics are so fundamentallydifferent that, while points of agreement may be expected, parallelsand congruency may not. The advantage of linguistic research is itsability to precisely distinguish between individual languages and the regularityof developments. The strength of archaeology is its precision in developingtimelines. What one can do, the other cannot. They could complementeach other beautifully, if only there were enough commonality."2) Nothing more precise may be said regarding the localization of theProto-Indo-European linguistic community.Based upon the localization oflater languages such as Greek, Anatolian, and lndo-lranian, a swathe of land insouthern Russia north of the Black Sea is often proposed as the native area ofthe speakers of Proto-lndo-European: -? Urheimat 1968 (Views of variousauthors from the years 1892-1963). On the region south of the Caucasus: -?Garnkrelidze / lvanov lE and IEs I 1995 p. 85Of.; Th. V. Gamkrelidze, "Neuereszum Problem der indogermanischen Ursprache und der indogermanischenUrheimaf' in ZVS 100 1987 p. 366-377.3) The reconstructible lexicon permits several assertions about the cultureof the speakers of Proto-lndo-European. For example, they kept and bredlivestock (For example, the Proto-lndo-European words *gWO!!_ 'cow,'*h,o!!i- 'sheep,' *pehr 'to lead to pasture, to shepherd' and *poh2i-men-

66 Introduction[Details on the research: -4 Hackstein, Sigmatische Prtisensstammbildungen1995 p. 176f.] 'shepherd,' *Ii(u)yon- 'dog,' among others).B. Forssman offers an accurate summary: --7 Sprache - Funf Vortrage1991 p. 63f. (The bracketed text in the fo llowing quotation is my additionand refers to corresponding Proto-Indo-European forms.):"The keeping of cows yielded 'milk' [Compare Greek,y(a)A.a(K)(t)- and Latin lact-] and 'to milk' [Compare PIE *dhe1}gh-: --7LW 1998 p. 129]; the keeping of sheep yielded 'wool' [Proto-IndoEuropean *",1}/h,-neh,-: --7 M. Peters in Sprache 33 1987 p. 114f.] and'to process wool' [Compare PIE *pe/(- 'to pluck (wool or hair)' and*kes- 'to comb wool': --7 N. Oettinger in MSS 53 1992 p. 149f.]. Thehorse [proto-Indo-European *(h,)ej(yo-, see above, E 506] pulled thechariot [proto-lndo-European *rot-II,-o-: --7 Mayrhofer EWAia Il p.429f.] ... Like other societies that bred livestock, the speakers of Pro toIndo-European lived with extended family [Compare PIE *dem- 'houseclan/community' and *1}i/(- 'settling clan/community': --7 MayrhoferEWAia I p. 697 and Il p. 561; cp. also below L 217 § I]. At the headof the extended family was a master of the house [proto-Indo-European*dems poti-: --7 Mayrhofer EW Aia I p. 699; cp. also below W 211]; themarried sons [proto-Indo-European * suH-ju-, or *suH-nu-: --7 MayrhoferEW Aia II p. 74 1] also belonged to the extended family and were,with their respective families, subordinated to the master of the house.... But a purely male-dominated hierarchy surely did not apply. Alonethe fact that the wife of the master was called 'mistress' [Proto-IndoEuropean *potnih,-: -4 Mayrhofer EW Aia II p. 74f.; cp. also L 211 §4], confers respect to her position . ... In their polytheistic religion, theyworshiped many gods, including fo rces of nature such as the 'father'god ofthe heavens [proto-Indo-European vocative, *d(iJjey ph,ter: --7Mayrhofer EW Aia I p. 751; cp. also F 318 § 6a] ... , the 'mother' earth[Proto-lndo-European *dheghom-, see below F 321 § 1], and the dawn[Proto-Indo-European *h,eys-os-: --7 Mayrhofer EWAia I p. 236; cp.also L 310, F 321 § 2 and W 303] . ... Man considered himself ... acounterpart to the gods; he named things such as himself 'earthly'[compare the PIE *dhghom-ijo-: --7 Meid Gaulish inscriptions 1992 p.22] and "mortal" [proto-lndo-European *mr-to-: --7 Mayrhofer EWAiaII p. 327]. But death [proto-Indo-European *mr-ti-, see below F 317 §7] could be ... overcome by indestructible fame. Our distant linguisticancestors clearly believed such things. Thus, one can reconstruct theProto-lndo-European expression */(Jeyos 6dhgwhitom 'indestructibleThe Reconstruction ofProtcrlndo-European 67fame,' in which *Kleyos actually means the 'heard news' ... In a societycompletely without writing ... fame spreads principally through oralretelling on the part of a creative singer and through auditive receptionon the part of other people, for example, at a festival of the Gods. Ofwhat was the fame of men composed, of which the singers then ...spoke and sang? Surely, to a large extent of men's [Compare PIE*uiH-ro-: --7 Mayrhofer EWAia II 569f.] deeds of battle. The communityof speakers of Proto-lndo-European declared their support forheroism and thus fundamentally for the subjugation of the weak. PeoplesofIndo-European languages have conquered, in the course oftirne,large parts of the world ..."4) Following is a small selection of literature on individual themesamong those mentioned. - a) On several areas: --7 Buck A Dictionary ofSelected Synonyms in the Principal Indo-European Languages. 1949;Hehn Cultivated Plants and Domesticated Animals (1885) 1976; Studienzum indogermanischen Wortschatz 1987; Scardigli Weg zur DeutschenSprache 1994 p. 43ff; Gamkrelidze / Ivanov lE and IEs I 1995 p. 377ff.(Part Two: Semantic Dictionary of the Proto-Indo-European Languageand Reconstruction of the Indo-European Proto-Culture); Mallory / AdarnsEncyclopedia 1997; A. Hausler in RGA 15 2000 under the headingIndogermanische Altertumskunde. - b) Focused on problems concerningthe horse: --7 FS Sehlerath 1992 [1994]; P. Raulwing, "Pferd, Wagen undIndogermanen: Grundlagen, Probleme und Methoden der Streitwagenforschung"in Fachtagung Innsbruck 1996 [1998] p. 523ff.; by the sameauthor, Horses 2000; also by the same author, in DNP 9 2000 under theheading Pfe rd. - c) On the social structure: --7 Benveniste Institutions I +Il 1969; B. Schlerath, "K6nnen wir die urindogermanische Sozialstrukturrekonstruieren?" Methodologische Erwiigungen in Studien zum indogermanischenWortschatz 1987 p. 249-264; S. Zirnmer, "Linguistische Rekonstruktionund Geschichte" in Bopp-Symposium 1992 [1994] p. 302-313. - d) On religion: --7 G. E. Dunke1, "Vater Himmels Gat/in" in Sprache34 1988-1990 [1992] p. 1-26 and Sprache 35 1991-1993 p. I; B.Schlerath, Religion der Indogermanen in Fachtagung Innsbruck 1996[1998] p. 87ff. - e) On poetic language: --7 Schmitt Dichtersprache1967; Indogermanische Dichtersprache 1968; Watkins How to Kill aDragon 1995.E 513. That the various lodo-European languages have developed from aprior unified language is certain (see above, E 435). Questionable is, how-

68 IntroductionThe Reconstruction of Proto-lndo-European 69ever, the concrete 'how' of this process of differentiation.Various hypotheseshave been formulated on this subject.I) The most suitable model fo r our tmderstanding of our reconstructionsremains the tree diagram first proposed by A. Schleicher: -4 Schleicher Compendium1866 p. 9. While all other models (including those of 1. Schmidt andO. Ho!ler) give a good account of observable linguistic development, they arelacking in that they are not usable and verifiable in the context of uniform reconstructionsas we receive them in our work. See above E 507 § 4, furtherH. Rix cited in E 426 § I. - J. Schmidt opposed the tree diagram with a waveimage "which spreads itself in concentric rings and becomes weaker as thedistance from the center increases": Schmidt (1.) Verwandtschajtsverhiiltnisse1872 p. 27f. - O. Holler "Slammbaumtheorie, Welientheorie. Entfaltungstheorie"in PBB 77 1955 p. 30ffgoes beyond J. Schmidt. On the wholecomplex in the light of nwneric-taxonomic processes of classification: -4 H.Goebl '" Stammbaum ' und 'Welie " in Zeitschrijt fur Sprachwissenschajt 21983 p. 3-44.2) As little certainty as may be gained concerning the localization of the nativeland and determining of the time period of Proto-lndo-European (seeabove, E 511 and E 512 § 2), it may in fact be possible to reveal how theProto-lndo-European community dispersed and developed.Several modelsare conceivable: Thus, for example the gain of new land through campaigns ofconquest, or expansion of the region of the language in the wake of a perfectlypeaceable spread of agriculture. Much speaks for the former, i.e. conquest, inmy opinion. (See above, E 512 § 3, at the end, on heroism.)B. Schlerath (-4 B. Schlerath inZVS 95 1981 p. 199.):Tn the words of''The belligerent operations for the spread of the Indo-Europeans seemto me fur more important than the relocation of furming populations thatoccurred in that context, or rather in their wake. I thus achieve an immediateconnection with that which we must accept from the lndo-Europeanconquest of Asia Minor, Iran, and Greece, and I may consider the Celticand Germanic migrations as immediate continuations of this process undervery similar conditions":The vital question of why the campaigns of conquest are not archaeologicallydocwnented remains unanswered.3) We have quite clearly voiced our support of the 'diverging treemodel,' i.e. of a uniform Proto-Indo-European language, with later divisionin daughter languages. However, one may not ignore that our modelis opposed by a 'converging association of languages model,' in whichlanguages that are in spatial and temporal contact [and are not necessarilyrelated to each other] exchange linguistic elements and rules, thus developingand acquiring from each other.Among the prominent advocates isN. S. Trubetzkoy (-4 A text from 1939 is printed in Urheimat 1968 p.214f.):"The supposition of an original Indo-European language is not totallyimpossible. However, it is not at all important; and one can verywell do without it.The term, 'language family' does not presupposethe common descent of a quantity of languages from a single originallanguage.We consider a 'language family' a group of languages, inwhich a considerable quantity of lexical and morphological elementsexhibit regular equivalences.In order to clarifY the regularity of thephonetic equivalences it is not necessary for one to suppose commondescent, since such regularity may also originate througb borrowingsbetween neighboring unrelated languages (so-called Fremdlautgesetze).Neither is agreement in rudimentary lexical and morphological elementsa proof of common descent, since all elements of human language maybe borrowed . ... Thus there is in fact no compelling reason for the acceptanceof a uniform original Indo-European language, from which individualIndo-European languages might have descended.It is just asconceivable that the ancestors of the Indo-European language brancheswere originally different from each other, but through constant contact,mutual influence, and borrowings, approached each other, withouthowever ever becoming identical to one another."Compare further: A. Hiiusler "Archiiologie, das lndogermanenproblemund der Ursprung der Hellenen" in Ohlstadt 1996 [1998] p. 79- 123.However, the thesis of a 'converging association of languages' mayimmediately be dismissed, given that all Indo-European languages arebased upon the same Proto-Indo-European lIexion morphology. As H.Rix makes clear, it is precisely this morphological congruence that speaksagainst the language association model, and for the diverging tree model(-4 1nL 17 1994 1994 18f.; G. Neurnann argues similarly in Ohlstadt 1996[1998] 262f.):"Linguistic elements and rules are not exchanged ad libitum, at leastnot in the langue. There is always a motive fo r borrowings from anotherlanguage: in phonetics, the imitation of an idiom with higher prestige orexotic attraction; in vocabulary, the narning of a borrowed thing, or againthe use of a more prestigeous language; in syntax, a rule that seems fittingor exemplary, etc.; this motive must be investigated . ... Understandably,

70 Introductionborrowings take place more easily in open sub-systems of language than inclosed sub-systems, in which it is more difficult to insert new elements.Open sub-systems include phonems and vocabulary, and to a lesser extentsyntax. Flexion morphology, on the other hand, is the closed sub-systempar excellence. Thus, if one wishes to investigate 'Proto-Italic,' he or shemust first investigate the f1exion morphology in languages of typologicallyIndo-European structure. Even the connection between the lndoEuropean languages was not discovered through phonetics, vocabulary,and syntax, but rather through the system of conjugation."Since H. Krahe there have been efforts to formulate assertions, throughlinguistic analysis of ancient European hydronomy, about the language(presumably Proto-Indo-European) of its creators: W. P. Schmid "AIleuropiiischeGewiissernamen" in Onomaslik 1995 p. 756-762; by thesame author, Schriften 1994 passim. However, the material is very problematic.The controversial thesis of a 'converging language association'plays a central role here. Criticism: Th. Andersson Zur Geschichle derTheorie einer alleuropiiischen Hydronymie in Probleme der Namenbildung1986 [1988] p. 59-90. For more information, see W 305 § 3.11. Proto-Indo-European PhonologyA. General InformationL 100. In anticipation of the individual sections below, the fo llowingtable of relevant Proto-Indo-European phonemes may be reconstructed:vowels (see below, L 200ff.): ·i ·esemivowels (see below, L 212ff): *j ·11·a·0 ·u*i ·ii ·ii ·0 ·udiphthongs (see below, L 219ff.): ·ej *oj ·oj*etJ *ay *oyliquids, nasals (see below, L 300ff): ·1 [·n ·r [·r]·m [.f/l] ·n [.(1]continuants (see below, L 308ff.): ·slaryngeals (see below, L 314ff.):. ", . ", ·h,dentals (see below, L 336): ·1·d ·dhlabials (see below, L 337):.p·b ·bhpalatals (see below, L 339ff.): ·K .g .ghvelars (see below, L 339ff): ·k .g .ghlabiovelars (see below, L 343£): ·kW .gW .gwhL lOt. The system of denotation used in L 100, including the h as superscriptto indicate an aspirated occlusive that is a single phoneme, follows

72Proto-Indo-European Phonologythe recommendations ofB. Forssman in Kratyios 33 1988 p. 61 (with note30) on the occasion of his review of Mayrhofer Lautlehre 1986. Thissystem is used consistently throughout the present work.L 102. The present description of the Proto-lndo-European inventory ofphonemes is consciously phonological and not phonetic. Giving priority tophonemes elevates indeed the transparency of reconstructions, but onl; atthe cost of their presumably concrete phonetic realization, e.g., PIE msd-o- 'nest,' and not [*nizdo-] (surely with a voiced z before the Olced d);PIE *s/eh,- 'to walk, to position oneself,' and not [*slah,-] (With vowelquality assimilation in the fo llowing laryngeals); PIE *dhh,IO-, and not[*dhh/olo-] (with weak vowel, see below, L 103); Proto-lndo-Europeanaccusative plural *-ms, and not [*-ns] (with assimilation, see below, F 104;but in this case one is not consistent and *-ns can also easily be fo und);PIE *dems-poli-, and not [*dens-poli-] (see below W 211).L 103. The Proto-lndo-European continuant 'thorn,' which fe aturesvoiceless *p, and voiced *0 , was often reconstructed in older research,and is considered here, as in 1. Schindler, to be a secondary product thatappeared in tautosyllabic position in initial consonant groupS such as *Ikand *dhg., see below, L 313. I thus write PIE *dhg.em- 'earth' and not*g.o em-, and further, PIE */liej- 'inhabit, settle' and not PI *kjJel-· .Laryngeals are here considered as consonants. Accordmg to this approach,their postulated Proto-Indo-European allophones *af, *a" *a, arethe result of the development of an initially phonologically irrelevant weakvowel between a laryngeal and a consonant: Since the laryngeal is positionedimmediately before a weak vowel, the existence of which it broughtabout, it disappears in accordance with phonetic rules. For example, comparethe PIE *dhhrIO- 'placed' with the spoken [*dhh/olo-] or rather[*dha/lo-] and the descendents from individual lndo-Europan .languages,such as Latin *datus (in conditus < * con-datus, and also m ItS sImplex,fac-tus), Greek ge16 and Vedic hi/o-. That which remains after the disappearanceof the laryngeal is the weak vowel, which in ilieek gets thequality of the laryngeal by assimilation (thUS e, a, or 0), or receIVes anothersole bias independent of its quality (Latin a, Vedic i, etc., see below L325).The series of occlusives are expressly not written in glottal fo rm. Seebelow, L 335 § 3.General Information 73L 104. One must be aware that we are entirely dependent on written recordsfor our knowledge of ancient Indo-European languages.The stepfrom the grapheme (of the letter or character) to the phoneme, which is as! rule what the grapheme denotes, must be respected. Statements such as1- > .Greek 1;- are m error, because the suggested phonetic change is firstdescnbed on the level of the phoneme, and then is completed as I; on thelevel of the grapheme. ObJecltvely correct is in this case alone the statement,that PIE *j- > Greek dj-, or rather dz-. On problems of phonetics(which IS mterested m phonemes in relation to their articulation) and phonology(which depicts language on the level of its relevant sounds or phonems):-7 M. Leumann "Phonologie der toten Sprachen" in Kleinechriften 959 . 398-40? On the history of the origins of the termsgrapheme and phoneme (and also on the confusion of sounds and lettersas well as the difference between written and spoken language): -7Kohrt Problemgeschichte 1985 p. 4/fL 105.I shall here discuss only briefly universal observations and causes?f dlachroruc phonological evolution (which concerns all areas of languagem varymg degrees).Language is sujecṭ to constant change.As an example, compare theOld Latm de/vos dlvuuty' > Classical Latin deus, see below L 217 § 3.Smce th neogrnanans (see above, E 306) it is established thatphonological evollton takes place according to certain laws.Thanks tomore precise empmcal mvestigation, among which research performed inthe region of Frankfurt and Cologne (i.e. in the Rheinische Facher) iswell-known, today one can better describe and understand the processesand spread of phoneltc developments:Initially, changed and unchangedfo rms coeXlSt; the variant applies at the beginning to few words; during theprocess of spred of the new fe ature, there is variance and irregularity; ifthere IS no SOCl pressure that resists the spread, the new feature canspread to t e entrre vocabulary in the entire community of speakers. Forfurther detas: -7 Bynon Historical Linguistics 1977 p. 173ff.Furter Hterature: -7 Boretzky Hislorische Linguistik 1977 p. 79ff.;SzemerenYI Emfuhrung 1990 p. 14ff. (Language in transformation)' HockLanguafe History 1996 p. I 26ff. (,Some types of sound chane'), p.143/f ( :-vny sounds change?') and p. 54 1f. ('Chapter notes and suggestedreadmgs ). Further, see above, E 509 § 4.L 106.Development according to rules of phonetics may be disturbedthrough the use of analogies.Thus, the development proceeds, not ac-

74 Proto-lndo-European Phonologycording to the given phonetic pattern, but rather takes as its model anotherpattern, one which already existed within the specific language. That iswhy the key expression ' Lautgesetz und Analogie' was chosen at the timeof the neogrammarians. - Compare as an example the Latin equos /equus 'horse': According to phonetic rules, one expects in the nominativesingular ecus < "ecos < equos. However, the normal paradigm of ClassicalLatin does not yield a nominative singular ecus, but rather equus: Theanalogical pressure of the equ-forrns such as genitive singular equi (phoneticrules here yielding the form equ-) was clearly so strong that even thenominative singular equos (> Classical Latin equus) was preserved, eventhough !!, when it is before o/u and not at the beginning of a word, normallydisappeared in the 3" century B.C. See above, E 506 § 3.Literature on analogy: -7 Szemerenyi Einfuhrung 1990 p. 29f.L 107. The absolute dating of phonetic, and other linguistic changes isonly possible in rare cases. Approxirnative assertions as to dates are therule. Compare as an example the formulation of M. Leumann LLFL 1977p. 62 on the dating of the phonetic change in Latin from ei > i:"The i of Classical Latin appears sometimes as i and sometimes as eiin the Old Latin of inscriptions , and indeed until 150 B.C., the phoneme,which is based upon one of the i-diphthongs, is written as i orperhaps ei, but the ancient monophthong i only as i; then, there appearsvariance in the orthography between ei and i, until finally only i remains.Thus, around 150 B.C., a phonetic fusion of the older ei and itook place, producing i."The information necessary to precisely determine a date is most oftenmissing. Nonetheless, we are sometimes in a position to chronologicallyplace a dubious change A in relation to a verified change B, and thus toclassify A chronologically before B, or vice versa.For examples of relative chronology, see the notes on the palatal law inL 206 § 2 and the comments on Germanic accents in L 421 § 4.Literature: -7 R. Gusrnani, "Marginalien zum Problem der relativenChronologie" in FS Szemerenyi *75 I 1992 p. 143-152.L 108. In the fo llowing description, it is important to distinguish betweencontext-independent and context-dependent development. The formerrepresents the normal case, which is unaffected by its phonetic context. Inthe case of the latter, the phonemes or accents in its context lead to greateror lesser deviations from the norm.General Information 75An example from Latin: PIE *e is context-independent and remains inLatin est 'is.' PIE *e fo llowed *!!, appearing as 0, as in the Latin novo'new,' on the other hand, is context-dependent in contrast to the Greekcontext-independent ne!!o- (Mycenaean ne-wo in the sense of 'of the currentyear'). However, when fo llowed by *ng, PIE *e becomes Latin i (example:Latin tingo 'to moisten') whereas the context-independent GreekT£yyro 'that' the *e remains unchanged. The case of Latin in 'in, into' inrelation to the Greek EV 'that' is more complex: According to the Greek,the Proto-Indo-European form of the local particle was *en. Therefore,we must accept that an early phase of Latin featured variation of the particle(according to the beginning of the following word) from the unchangeden to the changed (phonetically context-dependent) in and that in the fo llowing period this double representation en/in was decided in favor of in.Incidentally, the same explanation applies in the case of the e in the negation*en- < Proto-lndo-European*(I-: The original double representation*en-/in- is simplified to in-. In the case of Latin sedeo 'to sit,' comparedwith obsideo 'to besiege,' sed- is context-independent, whereas -sid-, inan initial secondary position is a weakened form of -sed-. (For informationon vowel weakening in short middle syllables, see below, L 204.) Formore information on the Latin examples above: -7 Leurnann LLFL 1977 p.45ff.; Sommer / Ptister Lautlehre 1977 p. 53ffL 109. In order to more fully document the descendents of Proto-IndoEuropean phonemes, comparisons not only of the habitual Latin, Greek,and Vedic are treated in the section on phonology, but also of Hittite,Lithuanian, Old Church Slavic, and the Old Germanic languages.B. Proto-Indo-European Vowels1. VowelsL 200. A methodical comparison of Indo-European languages yields aProto-Indo-European system oftive short vowels and five long vowels:*i *u*e '0"a* I"e"ii"0*i'l

76 Proto-lndo-European PhonologyL 201. In this chapter, only those occurrences of the vowels, 'a, ' 0, 'iiand '0, are treated that are not the products of laryngeal influence.Developmentssuch as 'a < 'h,a < PIE 'h",; ' 0 < 'h,o < PIE 'h,e; 'ii < 'ah,

78 Prot-Ind-European Phonologyever the difference between the native u and the Ionic-Attic ii . Forexample, cp. Boetian [tuid ' a] in comparison with Attic

80 Proto-lndo-European PhonologyProto-lndo-European Vowels 81undisturbed. Individual Old Germanic dialects undergo context-sensitivedevelopments with results that sometimes vary: Thus, from i > e (which inthe case of Gothic, when directly preceding h, Iv and r [called breaking],was preserved as the grapheme «11>; and which in North and West Germanic,among others, preceding a in the following syllable [so-called aumlaut)) and in analogous conditions from u to 0 (Gothic «1u» , otherwisefrom e > i (as in Gothic with the exception of the 'break-position,'North and West Germanic preceding nK or i, j and to some extent alsopreceding u in the fo llowing syllable). - Among the long vowels, PIE *0and *a are simplified to Old Germanic *0 (which, in stem syllables, in turnbecame Old High German uo, or New High German u). In North Germanic,and in a large part of West Germanic, PIE *e became a relativelylate: -7 Krahe I Meid Germanische Sprachwissenschajt I 1969 pp. 42, 51,57ff.L 209. Lithuanian, to which other Baltic languages correspond in termsof short vowels, exhibits, like Hittite, Germanic, and Slavic, a simplificationof PIE *a and *0 in a, whereas PIE *e, *i and *u are retained. _The long-vowels PIE *z (Lithuanian 0, but which is retained as a in Old Prussian): -7 Stang VergleichendeGrammatik der ballischen Sprachen 1966 p. 22ff.L 210. In Old Church Slavic, PIE *i and *u are retained as b and b respectively;PIE *e remains e; and *0 and *a are simplified as o. - The'replacement of PIE *zis Old Church Slavic i; that of PIE *u is Old ChurchSlavic y; PIE *e became Old Church Slavic e, PIE *0 and PIE *a bothproduced Old Church Slavic a. The discussion of the extent to whicholder Proto-Slavic preliminary phases precede the established Old ChurchSlavic phonological data has not as yet been resolved. Much speaks infavor ofthe theory that before the consolidation of PIE *0 and *a to 0 andof PIE *0 and *a to a, there existed a Proto-Slavic intermediate phasewith *a and *a:PIE *0 *a *0 *a\ I \ IProto-Slavic *a *a.l. ,/.OCS 0 aEarly Slavic loan-words in Finnish potentially offer evidence of this ProtoSlavic intermediate phase. Compare, for example, Finnish akkuna withOld Church Slavic ohno 'window' < PIE 'containing 0' *",kw_: -7Aitzetmilller Altbulgarische Grammatik 1991 p. 8ff., p. 19ff.; J. Udolph inIF 87 1982 [1983] p. 366f.; M. Trumrner in Die slawischen Sprachen 71984 p. 117ff.; by the same author, Aus der alteren slavischen und.balkonischenSprachgeschichte 1. Fragen des urslavischen Vokolismus. Graz1985 (non vidi: short reference in Indogermanische Chronik 32aNr.1084).L 211. Following is a list of correspondence sets for the individual vowelphonemes.I) Proto-Indo-European *i: - PIE *kwi-s 'who?,' PIE *kwi-d 'what?' =Latin quis, quid, Greek n.;, n, Hittite ku-is. ku-it i.e. lajis, lajit; further, cf.Vedic cit 'sogar, se1bst'; Gothic Ivi-Ieiks, Old English hwi-Ic, 'which'; OldChurch Slavic Cb-to 'what?.' - PIE *ni-sd-o- masculine or neuter 'nest' (onthe problem of phonetic realization, cf. L 102) = Latin nidus and Vedic nit/ti(with -z- as compensatory lengthening from Early Latin *-iz- or Early Vedic*-ii-), Old English, Old High German nest (Proto-Germanic *nista- n.). -PIE *dili- (the zero grade of PIE *dejli-) 'to indicate, announce' = Latin dicare'formally announce,' Cretan Greek 7tpO-oilc-vmt (2"" century B.c. in anepigram: -7 Bile C"itois 1988 note 297 b) 'helshelit shows,' Vedic dis-Oti'he/shelit shows, explains,' Gothic ga-taih-un 'they announced' (with breaking,cf. L 208), Old High German bi-zih-t f. 'accusation.' - PIE *h,olji- ID f.'sheep' = Latin ovis, Greek Ote;, in dialect oFt.;, Vedic tivi-, Luwian iJa-a-u-i-isi.e. Mljzs < *!JGlji-, Lithuanian avis; further, cf. Old Church Slavic OVb-ca f.'sbeep'; Old High German awi-st ID 'sheep stall.'2) Proto-Indo-European *u: - PIE 'jugti- n. 'yoke' (For infonnationabout the word beginning in Proto-Indo-European and Greek, cf. L 213 § I) =Latin iugum, Greek l;uy6v, Vedic yugam, Hittite i-u-mn in the sense of jug an,Proto-Germanic 'juko- (compare Gothicjuk, Old High Germanjoh). - PIE'd"ugh,ter- f. 'daughter' = Greek OUYa1£p-a accusative singular, Vedic duhitar-,Gothic dauhtar (for more infonnation on 0 «1U> cf. L 208), Lithuaniandukte, Old Church Slavic doSli. For more information on this word: -7 MayrhoferEWAia I p. 737f. - PIE *h,rudhrti- 'red, blood-colored' = Latin ruber< Proto-ltalic *rujJro-; Greek epuSp6c;; Vedic rudh-ira- (-7 MayrhoferEWAia II p. 453f.); cf. further, Lithuanian rUdas 'redish, red-brown'; OldEnglish rudian 'to be red.'

82 Proto-indo-European Phonology3) Proto-Indo-European *e: - PIE *mibhelehr f and *nebh-es- n. 'cloud'= Latin nebula f, Greek VEqlEAT] f, vE

84 Proto-indo-European PhonologyLatin iecinoris and F 314 § 6; for information on the word beginning, c[ L213). - PIE "hzSter nominative singular 'Stem' = Greek oo"t1\p. - Anancient PIE "e is presumed to underly the connection of Latin rex, reg-is'king' to Vedic raj-.An excursus: It is appropriate here to refer to the typically Celtic phoneticrule, e> i. It reveals the Gallic names ending in -rix (Vercingetorixetc.; cf. Old Irish rig 'of the king') as genuine and proves that commonwords in Germanic languages are borrowings from Celtic, including, NewHigh German reich, Reich and (Fried-)rich, Gothic (with in place ofi) reiks m. 'ruler,' reiki n. 'Reich,' Old High German rihhi 'powerful,'rihhi n. 'Reich,': W. Meid in Althochdeutsch I 1987 p. 10[ - For informationon Hittite-Luwian be and eb, derived from PIE "h,e and "eh" inwhich the e was conserved intact, c[ L 331.9) PIE "0, like PIE "e (c[ § 8), appears predominantly in lengthenedgrade constructions. In individual IE language constructions it is mergedwith "eh, (c[ L 323 § 3). - Examples: - PIE "!,,!dor n. 'body of water'= Hittite u-i-da-a-ar i.e. lIidar, Greek oorop (On its !lexion, c[ F 314 § 6).- The Proto-Indo-European root word "dom- / dem- 'house' had an accusativesingular of PlE "dom (c[ L 303 and F 320 § la), cf. Armeniantown 'house' and Greek o6i-a n. 'house, apartment, temple': MayrhoferLautlehre 1986 p. 172 and note 312.10) Primary PIE "a is rare, as is PIE "a. - The usual IE a developedfrom PIE *ah, (older PIE *eh,). - Although an original *eh, may not beruled out, the Proto-Indo-European word for 'mother,' which was preservedin some form in most IE languages, possibly contained the primaryPIE *a: Latin mater-; Mycenaean Greek ma-te i.e. mater, Doric a"tT]p,Ionic-Attic ""tT]P, Vedic matlir-; Old Norse mOder, Old English mOdor;Old Church Slavic mati (genitive singular mater-e); Lithuanian mote 'wife'(genitive singular, moter-s). Indeed, along with the problem of determiningthe vowel quality of the first syllable (-a- = Proto-Indo-European "-aor< *-ehr), the indications regarding !lexion and accentuation of theword 'mother' are equally ambiguous. Two possibilities are thinkable: -a) PIE "meh,ter- (or "mater-) originally (as in Ancient Greek ""tT]p), secondarilyadapted to the paradigm of "ph,ter- (thus Vedic matlir-; also representedin Proto-Gerrnanic *moiler-, as shown in Old Norse mOder andOld English modor, c[ L 421 § I). - b) PIE "meh,ter- (or "mater-) =Vedic and Germanic (the Greek stress on the beginning of the word in thenominative comes from the vocative; the suffixal stress in cases other thanthe nominative is older, for instance in the accusative singular l1't£pa: Mayrhofer EWAia II p. 345).Proto-Indo-European Vowels 852. SemivowelsL 212. For the two high vowels, PIE "i and "u, reconstruction yields thenon-syllabic equivalents PIE * j and *11. At first glance, PIE * i and • j, and"u and *!I appear to be two pairs of allophones, each of a single phoneme(with complementary distribution of PIE *i before consonants and PIE *jbefore vowels, cf. Vedic i-mas 'we go,' compared with y-anti 'they go'),cf. L 218. Whether this impression is misleading and each of PIE • i and"j, and *u and *!I, do in fact possess the property of a phoneme is open todiscussion: This is affirmed by Mayrhofer in Lautlehre 1986 p. 160(,while B. Forssman voices criticism in Kratylos 33 1988 p. 63.However, independently of these phonological considerations, the assessmentof PIE *j and *11 is important fo r the comparative method. PIE*1 and *11 are also included as secondary components of diphthongs, c[ L221 and L 223.L 213. [n the languages presently considered, PIE • j is, as a rule, preservedin context-free development. Graphemically, it appears in Latin as, in Vedic as , and in Germanic, Lithuanian, and OCS as . Precedinga, Hittite cuneiform employs ja, otherwise, i.\) Since tbe 8'" century B.C., Greek exhibits as a rule h-

86 Prot-lnd-European Phonologythesis I. According to lndo-Iranian evidence, the verbal root *jeljdh_, whichWlderlies the noun, likely featured a word-initial laryngeal and would thus be* Hjel}dh- 'to enter in motion.' Compare Vedic yUdh- f. 'fight' with the compositeamitriiyUdh- 'fighting enemies' < *amitra-Hjudh-, characterized by along linking vowel -

88 Prot(}-Ind(}-European PhonologyMycenaean dialects with the digamma . However, !! disappeared in lonicAttic dialects before the introduction of the alphabet and before the Homericepics took their final fonn. Although the Attic-Ionic alphabet, like all Greekalphabets, includes the digamma in sixth position, it is only used to indicate thenumerical value six.Many Homeric words and phrases presume the existence of !!, and musthave belonged to the Epic language at a period when !! was still an integralpart of the Epic phonetic system, cf. among many examples (another of whichis found in L 217 § I) the regular hiatus before av 'lord.' This is only understandableif it is caused by the disappearance of a word-initial !!. In fiIct, theword is preserved in Mycenaean as wa-na-ka, i.e. !!anaks 'king' (For informationon Post-Mycenaean semantic development: -7 LfgrE IT s. v. IlamA.e{x;; P.Carlier La royaute en Grece avant Alexandre. Strasbourg 1984). Concerningthe digamma in Homer: -7 B. Forssman in Colloquium Rauricum IT 1992 p.283f § 73.L 217. Correspondence sets for PIE !!:I) Word-initial: - PIE *!!ojlto- m. 'settlement, dwelling' = Lat. vicus 'village,settlement,' Myc. Gr. wo-(i-)ka-de i.e. !!ojkan-de, Arcadian-ThessalianFOtK, Ionic-attic OtK (Compare Homer, Rias A 606 and its formulaicejlav ohov&:, with a metrically long (owing to its position) -ay before OtKov,which is understandable when we accept that while the singers of epics nolonger knew the !! from their mother language, they learned from their scholarlyoral tradition to pronounce oh with a word-initial vowel insertion [*dos and >*d?viB.C.c)3"12"" *d. os > de. us but clivicent. B.c.Commentary: In individual cases, the phonetic situations descrihed in § 2have led to the disintegration of paradigms. Thus, deus 'god/divinity' vs.clivus 'as a divinity, worshiped.' (Several of the forms featured on the tableare not attested; although they are marked with *, their assertion is noncontroversial.)The 'a' level shows a uniform Old Latin paradigm of theroot dej7Jo- 'divinity' < 'heavenly.' (For more information on this word, cfW 202 § 2.) On level 'b' is shown the monophthongization of ej > e, cf. L220 § 1. A divergent development is shown in level 'c': disappearance of !!when followed by olu (on *d.os > de. us cp. Leumann LLFL 1977 p. 10Sf.),retention when followed by elila. From the last stage (c) of development beganthe formation of two new paradigms: On the one hand, deus with thenew genitive sg. dei 'god'; on the other hand, clivi with the new nom. sg.clivus 'as a divinity, worshiped.' - Compare equally the ClassICal Latmwords oleum 'olive ai!' vs. oliva 'olive, olive tree': The initial forms are*elaj!!om (from which, in a first step, the phenomenon of vowel weakeningof middle syllables yields *olei7Jom, and in a second -e!!Om > -e. om >-eum) and *elaj!!ii (from which comes first *olej!!ii, then *-!!a > -ilia).4) Postconsonantal !! is also fur the most part preserved. However, severalphonetic rules of Latin have obscured iroportant correspondence sets. Com-

90 Proto-Indo-European Phonologypare the fo llowing examples with • S!le- > suo- > so- ( Leurnann LLFL 1977p. 47): - PIE 'S!lesor- f. 'sister' (compare Ved. sVllsar-, Got. swistar) = Lat.soror (concerning -r- < -S-, cf L 309 §I). - PIE 'slfep-n- 'sleep' ( MayrhoferEWAia IT p. 791f, compare Ved. svopna-, Old Norse svefo) = Lat.somnus. - PIE 'S!leliru-hr (itself probably derived from 'S!lelb;r-hr [cf. L304 § 3] 'syeliur-o- m. 'stepfuther') f. 'mother-in-law' (compare Ved. ivas- *djm, cf. L 303.3. DiphthongsL 219. The three vowels PIE 'e, '0 and *a combine with the semivowelsPIE 'j and 'y and form what are called 'falling' or 'true' diphthongs suchas PIE 'aj and ' ay, compare as an example NHG Baum [bOom]. 'Climbing'or 'false' diphthongs, fo r example the French roi [ryo] 'king,' are notattested in Proto-Indo-EuropeanThe expected, and, in correspondence sets widely attested diphthongs are:*ej*aj*oj *ey *0'1}In addition it may be asserted that under certain conditions long diphthongssuch as PIE 'oj or *elf may have existed, cf. L 224.L 220.languages:·ayInformation regarding the fate of -j-diphthongs in Indo-EuropeanI) The three short -j- diphthongs are preserved as , and inpre-Classical Latin. - In the course of development of Classical Latin in the2"" century, ej became the monophthong I, after a transitory phase. - Unaffectedby the phonetic circumstances, oj became u; however, in certain positions,the diphthong oj was preserved in the slight)y modified form oe .For example: poena 'punishment' vs. punlre 'to punish' and Poenus 'inhabitantofCarthage' vs. Punicus 'Punic.' Important correspondence sets such asLat. VICUS (= Myc. Or.! dialect yojko-) show the phonetic transformation l}Oj> Zlej- > ui-, influenced by the phonetic environment, in this case, by If-· -SU:;ce the - 2" " century B.C., aj is replaced by , which denotes a diphthong

92 Proto-Indo-European Phonologyjust as oe above. An attestation of this is Caesar. As the New High Germanloanword Kaiser shows, the pronunciaton of the word at the time of the borrowingwas [Ka.esar], cf. E 507 § 3. "The pronWlCiation of ae as a monophthongis a post -Classical development.For further information on all aspects of monophthongization in Latin:-7 Leumann LLFL 1977 p. 60ff.; Wachter, Altlateinische lnschriften 1987p. 477ff.; H. Rix, "Late in - wie wurde es ausgesprochen?" in Beitrage zurmiindlichen Kultur der Ramer, edited by G. Vogt-Spira. Tiibingen 1993 p.I1 ff.; Meiser Laut- u. Formenlehre 1998 p. 57ff.2) The PIE -j- short diphthongs are preserved unchanged in Greek asEl, al, ot.3) In the case of lndo-Iranian, in which PIE *e, *0 and *a merged (cf.L 206), of the three PIE short -j- diphthongs, one would expect aj alone tohave been preserved, and, that the PIE short -l!- diphthongs would havedeveloped in an analogous way (cf. L 222 § 3). This aspect is preserved inthe Old Iranian languages and must also be postulated for the Early Vedicperiod. In extant lndo-Aryan records, aj becomes the monophthong ifwhen it precedes a consonant. In traditional Indological transcriptionpractices, this is denoted as «t>, since there is no «t> in Old lndo-Aryan(-7 Wackemagel / Debrunner Altindische Grammatik I 1957 p. 35ff.;Thumb / Hauschild Handbuch des Sanskrit III 1958 p. 228f.), compareprevocalic ay-ani 'I want to go,' but preconsonantal e-mi 'I go.' On theother hand, Ved. if «t> may be traced to Indo-Iranian az, compare Ved.nM-i$lha- 'the next' with Old AvestanlLater Avestan nazd-ista-; Ved.e-dhi 'be!' < *as-dhi [azdhi] with as-ti '[you] are': -7 Wackemagel / DebrunnerAi. Grammatik l J I 1957 p. 37ff.4) A tendency toward monophthongization is discernible in Hittite _whereas initial PIE *oj and *aj became an open , PIE *ej became a closed. The latter could become i following a velar. Where open and closedversions are preserved, they both appear (at least in terms of graphical notation)to be denoted as e: -7 H. Eichner in MSS 31 1972 p. 76ff.; MelchertAnatolian Historical Phonology 1994 p. 148f.; S. E. Kimball in Sprache36 1994 p. I ff.5) In Germanic, PIE *oj and *aj merged to form *aj (as in the case ofPIE *0 and *a, which became *a, cf. L 206). PIE *ej appears in all ancientGermanic languages as i (Got. «tl>, OE, OHG, etc. Classical Lat. deus / divus, cf. L 217 § 3), Ved. devam,Lith. dievas m 'god,' ON tiv-ar nom plural 'gods.' For more informationon this subject, cf. W 202 § 2. - PIE *dejli- 'to point, indicate' (For examplesof zero grade PIE *dili-, see ahove L 211 § I.) = Lat. dicere 'to announce,determine, speak' (ex-deic-endum is preserved in Old Latin), Gr. 8ei1cV1l!ll 'Ishow,' Ved. deM- m 'area' « *'direction'), Got. ga-teihan 'to report, proclaim,'OHG zihan = NHG zeihen. - PIE *i!'ej-m- 'winter' = Gr. XEllla n.'winter (storm),' Ved. he-man loc. sg. 'in the winter,' Hitt. gimmant- (-7Melchert Anatolian Historical Phonology 1994 p. 102, p. 145), Lith. iie-maf., OCS zi-ma f. 'winter.'2) Correspondence sets for PIE *oj: - PIE *(H)6j-no- 'one' = OLoino(m) m 'the one (acc.)' (Classical unus), Gr. 01VT] f. 'one (as on a die)' (-7Chantraine DELG S. v.), Got. ains 'one,' Old Prussian ains 'alone, only,' cf. F502 § I. - PIE *lojkW-o- 'remaining, left over' (with -0- ablaut, changes toPIE *lejkW-, Gr. Ael= etc.) = Gr. A.ol1l6 OHG lifhan 'fief.' - Also compare PIE *!!6jlio- and"l!ojde in L 217 § 1.3) Correspondence sets for "aj: - PIE *Iajl!o- 'left' = Gr. A.al Lat. lae\IUS, OCS lev. 'left.' - PIE "kajka- 'blind (in one eye)' = Lat. caecus 'blind,'Got. haihs 'one-eyed.' - For information on problems concerning Protoludo-Europeana, cf. L 211 § 5.L 222. The fate of short -l!- diphthongs in the lE languages:I) "The Proto-Indo-European short -l!- diphthongs *el!, *oy and *ay, whichare analog to the short -j- diphthongs are preserved in Greek as £u ou, and QU.2) While is preserved in Latin until the classical period, PIE "el! and" Ol! are already unified as *Ol! in Proto-ltalian; it has been suggested that isolatedattestations of OL correspond to e!! (compare Leucesie, attested in aCarmen Saliare): -7 Leumann LLFL 1977 p. 71; Meiser Laut- u. Formenle-

94 Proto-Indo-European Phonologyhre 1998 p. 59 § 47,5. Old Latin ou became the monophthong it in the formationof Classic Latin. But, between word-initial I and a labial, a LatinoFaliscan o!! appears as oj (from which ej and ; develop according to phoneticrules. Cf. L 223 § I and the example of * h,le!!dhero-).3) In a manner parallel to the assimilation of all Proto-lndo-European short-j- diphthongs in Vedic to e (, cC L 220 § 3), the preconsonantal short PIE-!!- diphthongs were assimilated in Vedic to 0 (

96 Proto-lndo-European PhonologyC. Proto-Indo-European ConsonantsI. Liquids and NasalsL 300. While the PIE liquid consonants ·r and ·1, and the PIE nasals .mand ·n are related to the PIE semivowels .j and . !!, their syllabic allophones,PIE .[r], ·W, .['!'] and .[(1], are related to full vowels. In theirnormal, non-syllabic quality they remain unchangedEuropean languages.L 301. Individual Languages:in most IndoI) In the case oflndo-lranian languages, one observes the phenomenon ofdialects that assirnilate PIE ·r and ·1, forming r (e.g.Ved. rih-, Avest. riz- 'tolick' as opposed to OCS lizati; Ved. roca- as opposed to Gr. AeUK, cf L223 § I) as well the phenomenon of those that retain PIE ·1 (e.g. Late Ved.lih-, Mod. Pers. listan 'to lick'; Ved. lokti- [cf L 223 § 2]), and finally that ofthose dialects in which PIE ·r and PIE ·I are assimilated to become I (e.g.Ved. lup- 'to break apart' along with the original Ved. rup- 'to tear apart,'Lat. rumpere 'to break'; Mid. lndo-Aryan [ASoka] laja 'king' as opposed toVed. raja, Lat. rex 'king'). See also L 306 § 3 below, with PIE ., representingVed. r2) The Mycenaean Linear B alphabet does not differentiate I and r, a featurealtnbutable to the fuet that the creators of Linear B spoke a oon-IndoEuropean language. Although the difference between I and r was phonologicaIIyanchored in their language, the Mycenaean Greeks adopted this peculiarquality in their writing system: Heubeck in Res Mycenaeae 1983 p. 163 f3) A finaJ position PIE ·-m becomes -n in Greek, Hittite, Germanic, and inBaltic as well. (In Lithuanian, postvocalic -n is only discemible in the purelygraphical nasalization). - Final PIE ·-er and ·-or and PIE ·-en and *-onappear in several languages less the finaJ -r (or -n). Compare Gr. 1tOtTtP 'father'and 8uyOTl]P 'daughter' with Ved. pita 'father' and duhita (= Lith. dukte,OCS d7>.5ti < *dukt€). Compare Gr. MV with Ved. S(u)va 'dog' and Gr.iiKIlOlV 'anvil' with Lith. akmuo 'stone.'L 302. Correspondence sets:Proto-lndo-European Consonants 97I) PIE or: - PIE ·doru- n. 'wood' (For declension, cf F 317 § 8) = Gr.oopu, Ved. daru, Hitt. ta-ru-u-i, i.e. ta,!!i dat. _loc. sg.; compare Got. (rlU,OCS drevo 'tree.' - PIE *mijes 'three' (Lat. Ires etc., cf. L 215 § I).mel, . IffiAl-O ah, and the assimilation of -ms > -n ) leads to _Pre-PIE*-ans, and from there, > PIE ·-as. Compare Ved. -as, Got. -os. ( J.

98 Prot Pre-PIE "gWans, and from there, just as thepreceding > PIE "gWas, compare Ved. gas, Doric Gr. . For more informationon PIE "gWoy_, cf. F 318 § 6b. - On the PIE acc. sg. "dam'house' < pre-PIE "dom-m, cf. F 320 § I a; on the PIE acc. sg. ·dhlig"am< pre-PIE ·dhlig"om-m c( F 321 § 1.L 304. In many positions, but particularly between consonants (KRK)and in final position fo llowing a consonant (-KR#), the so-called sonorants,PIE or, "I, "m or "n (which includes PIE *j and !!, cf L 212) becomethe syllabic allophones .[r), ·ll), .[",), and "[(I) (with the vowels "iand "u). Compare PIE "bher- 'to carry': The -to- verbal adjective requiresthe zero grade form of the root, which thus appears phonetically as PIE"bhr-to-.I) The fundamental rule for the production of syllabic allophones functionsiteratively (nonsyllabic-syllabic-nonsyllabic, etc.). Further, in the series KRRK ,the second sonant is affected first. On the subject of the iterative quality, comparePIE "f(yn-es gen. sg. 'dog' => "ICun-lis (the ending -lis remains, the preceding-n- is non-syllabic and includes syllabic u with itself) = Ved. slinas.Concerning the use of the second sonant before the first, compare PIE"f(yn-bh- with _bh_ case => *f(y(l_bh_ (and not *lCun-bh-) = Ved. swlbhis.2) Several exceptions to this fundamental rule are known, fo r example theword-initial PIE ·mn-eh,- 'to think of (as opposed to the expected "",n-eh,-),or the -n-present form PIE "jung- 'to harness' (as opposed to the expectedform *iy(lg-), or the paradigm of PIE "trej-es 'three' the genitive plural"trijam with syllabic i + transitional sound j (c( L 218; according to the rule,one would expect ·1[j6m with syllabic r and non-syllabic ]). For further information: J. Schindler in Sprache 23 1977 p. 56t: Further, for information onPIE "mOg-, ct: L 203.3) The metathesis of PIE "yr hefore KI# > "ru presents an interesting specialcase. Compare the example of Gr. iopixo 'to seat, to install a place ofworship < ·to create the seat of a divinity' < Proto-Gr. "sOdyr-je- (On thesubject of the weak vowel 0, the so-called schwa secundum, ct: L 203; forinformation on s- > h- c( L 309 § 2.): The verb represents a fuctitive -iepresentform of "slid-yr n. 'seat'; compare the parallel Attic constructnA.i't'tOl 'to harvest honey, to cut honeycombs' < "mlit-je- from "melit'honey' (ct: L 302 § 2 above). From whom it was that I first learned of theexample iopixo, I can no longer say, although perhaps it came up in a coursegiven by J. Schindler at Harvard in the Fall Term of 1980. The excellence ofthe questionable analysis oftopixo is shown by dubious efforts such as "obviOUS derivation from "iop\)(;" ( Peters Laryngale 1980 p. 98). A treatmentthat approaches a solution, however without bringing metathesis into account,is: C. 1. Ruijgh in OS Kurylowicz I 1995 p. 353 note 34 with the remark"derive de ·slid-ru-?" - Further examples: Mayrhofer Lautlehre 1986 p.161(; concerning PIE ·sylif(yr-h,- also cf. L 217 § 4 above; fo r information onPIE "kwtru-K- < ·kWtyr-K- 'four' also ct: F 501 below. For a derivation of'four' with metathesis, bringing Hittite kutruen- 'witness' into account: Rieken Nom. Stammbildung 1999 p. 289 note 1385.Perhaps the juxtapositioning of PIE "y!kWo_ (= Ved. vfka- etc., c( L307 § 2) and "lukWo- may he explained by the same process. Concerning*Iukwo-, compare Lat. lupus and Gr. A.'J1CO. The Latin lupus is likely aloanword from Oscan. The development of -p- < _kw_ is not Latin.·Iukwo- is dissimilated to luko- in Greek.L 305. In constrast to the clarity of the state of affairs concerning liquids,only through the astute combination of comparative and language-internalmethods could the syllabic nasals "[ "') and "[(I) be designated as allophonesof PIE "m and on. Since the relationship Ved. man- 'to think':malli- 'thought' corresponded perfectly to that of Ved. bhar- : bhṛta-, itwas possible, given the relation bhar- : bhr- = man- : x for x (= Ved.ma-), to postulate "mlJ-. The language comparison shows that the -a- ofthe Vedic ma-(ta-) and ma-ti- ( 'thinking' corresponds only in Greek to an-a- (cf. the compound, am6-l1a0 'it/himself striving, of itslhis own volition,'attested since Homer), but in Latin corresponds rather to an -en- (cf.menti- 'thought power, thought technique, thought' with the nom. sg.mens, gen. sg. mentis: Reichler-Beguelin Typ mens 1986 p. 19ff.), inGermanic to -un- (Got. ga-munds 'remembrance'), in Lithuanian, -in(mintis 'thoughts'), and in Old Church Slavic -{'- (pa-m{'tb 'memory').These disparities led precisely to the common denominator "-(1-, whichwas originally taken from Old Indian material.The existing knowledge of nasalis sonans may be traced to Karl Brugmann"Nasalis sonans in der indogermanischen Orundsprache" in Studienzur griechischen und lateinischen Grammatik (edited by G. Curtius and,starting with volume 9, together with K Brugman[n)) 9 1876 p. 285ff.Brugmann had predecessors, including the fifteen year old pupil Ferdinand deSaussure (1857-1913). According to his own account, he revealed Gr. -a-

100 Prote>-Inde>-European PhonologyNous lumes ... un texte d'Herodote ... [qui) contenait la formeEaxaat. La forme 'tEaXatat etait pour moi completement nouvelle... A I'instant ou je vis la forme 'tEaxatat, mon attention, extremementdistraite en general ... fut subitement attiree d'une maniere extraordinaire,car je venais de faire ce raisonnement, qui est encore present amon esprit it I'heure qu'il est: AEy0!J.EOa : Atyonat, par consequent'tEaY!J.EOa : 'tEaxNat, et par consequent N = a-7 C. Watkins Remarques sur la methode de Ferdinand de Saussure comparatistein CFS 32 1978 [1979) p. 61 (= Watkins Selected Writings 1 1994 p.266); Kohrt Problemgeschichte 1985 p. 112ffL 306. The remnants of the PIE allophones *lr), *[j), *["') and *[(1) varyin nearly all of the IE languages. Only in the case of Old Indo-Iranian isthe relationship between bhar- and bhrta- still quite clear. That is why theOld Indian grammarians were able to develop an ingenious ablaut system,cf. L 413 § I below. For example, the Ved. If) is represented in Latin by-or- (cf. Jo rs, for-li- 'chance' < *bhr-li-), in Germanic, by -ur- (cf. NHGGe-bur-I) etc.I) Latin retains the PIE syllabic liquids *lr) and *W as or and 01, and thePIE syUabic nasals *["') and *[(1) as em and en (and secondarily as in, cf. L108).2) In Greek, PIE *lr) and *W became pa/ap and Wa'!..; in dialects, p% p,'NJ/o'!.. as well. For details: -7 Lejeune Phonetique 1972 § 201; Risch Kl. Schr.1981 (in an essay from 1966) p. 266ff PIE *["') and *[(1) yielded a; in dialects,also 0: -7 Risch Le. Preceding vowels, semivowels, and laryngeal/vowelcombinations, af! and QV represent the normal case. On the subject of *IJH inGreek, cf. L 332 below.3) IE *[r) and *W are preserved as syUabic liquids in Old Indo-Aryan (notto mention lndo-lranian), and as r (which corresponds to Avestan ara) in Vedic.The only Vedic root containing -/- in its zero grade, kalp- / kip- 'to add,equip' has no sure explanation. (perhaps PIE *kwerp-, if it corresponds toVedic krp- 'figure' and Latin corpus: -7 Mayrhofer EWAia 1 p. 324). Thesyllabic PIE nasals *["') and *[(1) appear as a, preceding vowels and semivowels,as am or an.4) In Hittite, PIE *[r) and *W developed into ar and al. The PIE syUabicnasals are presumably represented by an: -7 Melchert Anatolian HistoricalPhonology 1994 p. 125f., which contains further details.Prot Gt. '!..UKO *h,hp-to- cp.F 502 § 13 below.) 'hundred' = Lat. centum, Gr. EKmov (Arcadian EKmov),

102 Protcrlndo-European PhonologyVed. satam (Avest. satam), Got. hund- 'hundred,' Lith. simta-s 'hundred';PIE *de/iJrz-to- 'the tenth' = Gr. OClCmo *asme > *amme = Aeolic/Homeric& (with Lesbian retraction of the accent), in Doric, in place of*amme, with disposal of the double-consonant through compensatory lengtheningo, in Ionic-Attic again compensatory lengthening *ame with secondaryaspiration and the additional nom pI. ending -

104 Proto-Indo-European PhonologyA similar s > S change in Lithuanian affects only a part of the cases that comeinto question: --7 Slang Vergleichende Grammatik 1966 p. 94ff.L 310. Correspondence sets for PIE os: - PIE *segh- (zero grade*sgh-) 'to overpower' = Ved. sahate, OHG sigu m. 'victory,' Gr. zerograde thematic aor. E-ax-OV of the full grade thematic pres. EXW 'to hold,to posess' (EXW by aspirate dissimilation [cf. L 348 § 2] < *hekh- < *segh[with the typically Greek transformation of mediae aspiratae into lenuesaspiratae, cr. L 336 § 2]) and the Fut. Ew (here heks- without aspiratedissimulation < *hekhs-: The consonant group -ks- < -khs- was formed justearly, and thus, during the period of activity of Grassrnann's Law, couldno longer play a dissimulative role: --7 Meier-Briigger Gr. Sprachw. I 1992p. 59 § E 211.4; also cf. Ved. bandh- vs. Fut. bhantsyati, cf. L 348 § 3).- PIE *h,e'1s- 'ear' = Lat. auris ( - aus-culto 'eavesdrop'), Got. auso,OHG ora n. 'ear' = Lith. ausis f., OCS laO n. 'ear.' - PIE *hles-ti 'he is'and PIE *hlesi 'you are' (For information on the simplification, during thePIE period, of original ° "Ies-si, cf L 312) = Lat. est, es, Gr. eern, et,Ved. asti, asi, Hilt. es-r'i, esi (for information on the lengthening ofstressed vowels in Hilt., cf. L 207), Got. ist, is, Lith. esti, esi. - PIEo",o-sd-o- m. 'branch' (concerning phonetic pronunciation *[ozdo-], cf. L102; on word-initial sounds: --7 H. C. Melchert in HS IQ I 1988 p. 223note 16 with the comparison, already made by E. H. Sturtevant, with Hilt.,asdyer 'timber, brushwood') = Gr. o1; (with = [zd]), Got. asts. -PIE *genhl-os nom. sg. 'gender' with gen. sg. *genhres-os 'gender' =Lat. genus, generis (with -eri- < *-ese-), Ionic-Epic Gr. y£vo, Y£VE(with -e'1- < -e.o- < -eho- < *-eso-). - PIE ''1es- 'to wear, to be clothed'= Ved. vas-te, Hilt. u-e-eS-ta i.e. yes-to 'wears,' compare Gr. EVVllllt (presentstem *yes-nu- along with Aor. Ea-am) 'I dress (myself),' Lat. ves-ti-s'robe,' Got. wasjan 'to clothe' and cf. L 21 1 § 4. - PIE 'pis-to- 'hecrushes' = Lat. pistus, Ved. pi$!a-, compare OCS pbXati 'to push.' - PIE'stehr (> *stahr, cf. L 323 § 2) 'to stand' = Lat. stare (which is not directlytraced to a root aorist as is the communis opinio, but is rather contractedfrom *steh,-je-: Meiser Laut- und Formenlehre 1998 p. 187),Dor. Gr. '(aallt « *si-stahr), cf. Ved. stha-tra- n. 'location,' OHG stan'to stand,' Lith. stoti, OCS slaU 'to place oneself' - PIE *h,eys- (>*",OIjS-, cf the examples in L 323 § 2) 'to light up (sunrise),' compareLat. aurora « *aysos-ii), Proto-Gr. *a'1sos > *ay'1os > Lesbian Aeolicoilw

106 Proto-indo-European Phonology2) First researcbed in the twentieth century, the IE languages Hittite (cf. E410) and Tocharian (cf. E 408) bave contributed significantly to examples ofthis type: The sequence ithJ_/hJ, attested in Greek, corresponds to an inversesequence (/hJ-ithl) in Hittite and Tocharian. In Hittite, 'earth' appears innom/acc. sg. as le-e-kOn i.e. tegan (further attestations in Tischler, HEG III p.292ff.), and in Tocharian A as lka",. The Hittite nom sg. iJar-ttig-ga-as i.e.!Jarlkas, appears to mean 'bear' (further attestations, including suggestions forfurther reading: Puhvel HED III p. 20 I f.).3) Thus, in place of PIE *gha m- ('earth'), a reconstruction ofa paradigmwith conventional consonants appears correct: Hitt. strong stem with nom sg.lekan < PIE *dhegr6m and weak stem with gen. sg. takn- < Proto-Anat.*dh°ghm_ < PIE *dhgrm_. While Anatolian, thanks to the schwa secundumwas able to keep the sound group paradigm-internal, elsewhere PIE *dhgrm_was simplified to *grm-. Compare: Gr. XOl1oi 'on the earth,' Lat. humi 'onthe ground' (from which, secondarily, humus f. 'ground' developed: Wackemagel Vorlesungen IT 1928 p. 32), Lith. ieme, OCS zemlja 'earth.'Finally, in such a case as that of the loc. sg. PIE *dhgrem, the tautosyllabic*dhgr_ was transformed to *grdh_, upon the basis of which both Gr. f

was already discernible in Vedic grammar. Saussure's great achievement wastbe abstraction of comparing nasaI present forms with the more transparent -ninfixverb type of the Old Indian seventh class. Comparing Ved. yu-iz-k-Ia'harness!' with the participle yuk-Ia- 'harnessed,' the insertion of -n- before thefinal phoneme of the root yuk- was clearly visible. Saussure recognized thisnasaI infix present pattern not only in verbs of the Old Indian fifth class (Ved.Sr-I;l-u-Ia 'hear!' [imperative, pl.] and srn-tO- 'heard'; on the ani,-root srn-, c£§I), but also in the relation ofVed. pu-n-I-Ia 'purilY!' to pii-Ia- 'purified.' Infuct, he anived at the correspondence set: Ved. yok-Ira- n. 'harness' (withyak- < *yayk-, cf. L 222 § 3) : yu-iz-k-ta : yuk-Ia- = Sra-Ira- (above §I) :Sr-I;l-u-Ia : srn-tO- = pavi-Ira- : pu-nl-Ia : p/'-Ia-. If this was true, then the se!root must have contained a root-final feature, which we shall call X (not correp108 Proto-lndo-European Phonologyeuropiennes by Ferdinand de Saussure, written as a student at age 21 (Leipzig1879, first appeared 1878; published again in Saussure Recueil 1922 p. I fl).Among his credits, he adopted the organization, by Indian grammarians, ofSanskrit roots into ani,- and se,- roots, for modern comparative linguistics. -The sages of ancient India had observed that some roots, when followed bycertain grammatical elements, did not exhibit the -i- (thus called an-i-" i.e.'without -i-') that other roots did exhibit (which were thus called se, < 'sa-i-,Le. 'with -i-'). Thus, Vedic man- 'to think' is an ani,- root: Compare Ved.man-Iar- m. 'thinker,' mim-Ira- m. 'speech, suggestion' without -i- when preceding-Iar-, -Ira-. In contrast, Ved.jan- 'to create' and san- 'to gain,' (whichare in modern lE linguistics more accurately noted as jan'- and san'-, respectively)are se,-roots. They exhibit -i- preceding the suffixes -Iar-, -Ira-. Compare:Ved. jani-Iar- m. 'creator,' jani-Ira- n. 'place of birth,' sani-Iar- m.'winner, one that benefits,' sani-Ira- n. 'gain, wage.' The Indians also distinguishamong tbeir zero grade roots those that exhibit -i- from those that exhibit-@- preceding certain suffixes. They represent the verbal root pii- (fullgrade pav'-) as heavy (= se!): Compare pavi-Iar- m. 'purifier,' pavi-Ira- n.'sieve.' The verb Srn- 'to hear' (full grade srov-), on the other hand, is consideredlight (= ani!). Compare srotar- m. 'listener' and sro-Ira- n. 'hearing(sense),' both with preconsonantal Sro- < *sray-, c£ L 222 § 3. Further, comparethe light Vedic verbal root bhr- (full grade bhar-) 'to carry,' Ved.bhOr-iu-m 'to be carried,' and bhar-Iar- m. 'porter' with the heavy Ved. fullgrade la';- 'to overcome,' Ved. d-!arl-tu- 'insuperable.' (Here, as elsewhere,the -i- is sometimes lengthened: Wackernagel / Debrunner Ai. Or. /1957 §18; Hirt Idg. Or. IT 1921 § 135.)2) It is possible that the connection of se,-roots such as Ved. pii- (palI-) 'topurilY' (c£ the participle pii-Ia- 'purified') with present stems of the Old Indianninth class (cf. Ved. pu-ni'i-li 'be purifies,' pu-nl-Ia imperative plural 'purilY!')Proto-lndo-European Consonants 109sponding to Saussure's term) for the time being. It appeared following consonants(-v-, -n-) as -i- (and is sometimes also lengthened, c£ §I: d-,6rI-IU-),following short vowels, as -u-; but in both cases as a lengthening. If one fo rmulates accordingly the last series, the analogy to the first series is conspicuous:pavi(lra-}< 'payX(lra-}pu-n-I(ta}< *pu-n-X(ta}pii(la-}< *puX(ta-}3) That Indian grammarians, in proposing the presence of an additional -iinse,-roots, recognized an immanent linguistic trait, is evidenced by roots withliquid and nasaI sounds. If the ani,-root man- (cf. man-Iar-, above §I) had*m!!- as its zero grade (cf. Ved. ma-Ia-), then the zero grade fo rms of the se!rootjan'- (cf. Ved. jani-tOr-) may be traced back to an -IJ- + -X- (c£ Ved.jatO- 'create' with -a- < *-aX- < *-1JX-). Likewise, only ani,-roots such ashar-Ibhr- 'to carry' have a zero grade fo rm with -r- (cf. bhr-Ia- 'carri ed').However, in the case of se,-root forms such as full grade ta';- 'to overcome'or full grade par1- 'to fiII,' -r- + -X-, and -/- + -X- must be added with the result-Ir-, and -iir-, cf. Ved. tir-lhO- 'ford' and piir-I;la- 'filled').L 316. To summarize, while this X is only preserved fo llowing consonants(and non-syllabic sonants) as Ved. -i- (corresponding to the i-' fromOld Indian grammar), after vowels and syllabic nasals and liquids it hascaused a lengthening, cf PIE • peyX- > Ved. pav'- compared with PIE*puX- > Ved. pii-; PIE *genX- > Ved. jani_ compared with PIE *g1JX(corresponding to *g(i- in older transcriptions) > Ved. ja-; PIE *lerX- >Ved. la';- compared with PIE *trX- (corresponding to It- in older transcriptions)> Ved. Ilr-. However, where this element was preserved as thevowel Ved. -i-, it largely corresponded to -a- in non-Indo-Iranian languages.Thus the verb type of the Old Indian ninth class, with zero grade-nl- in pu-ni-Ie 'is purified' and full grade -na- in pu-ni'i-ti 'he/she/it purifies'(cf. L 315 § 2), may be compared with the zero grade Gr. -vo- inoa-vO-tat 'to control' and the full grade lon.-Alt. -VTj- « -na-) inoa-VTj-m 'controlled.' - The equating of Ved. slhila- to Gr. (natOC;'arranged' further indicates that even the zero grade of long vowel rootssuch as Ved. slha- and Gr. (non-Ion.-At!.) sla- (in lon.-At!. '(-cr't1]-l

110 Proto-Indo-European PhonologyProto-Indo-European ConsonantsIIIL 317. Not only did Ferdinand de Saussure identuy the element herereferred to as X, he also postulated two such values, to which a thirdwould later be added by researchers. Thus, today's view of three differentlaryngeaJs underlying the Gr. the-, stii- and do-: Gr. thii- 'to place, lay, sit '(cf. n-I}11-!11) < PIE *dheh,-; non-Ion.-Att. stii- (cf. lon.-Att. t-0'tI]-!11 withe < ii) < PIE *stehr; do- 'to give' (cf. Ot-OCJ)-!11) < PIE *dehr. De Saussurehad seen vocalic elements (coefficients sonantiques) in these values.L 318. Classical Indo-European linguistics of the period followed deSaussure in many ways, but not in all. For example, in place of the X, forwhich he had postulated two values, and for which his laryngealist successorseven proposed a third, a single schwa, a, the schwa indogermanicumwas postulated. The field further adopted his assertion that the sel-rootscould be traced to Proto-Indo-European.An example such as the Ved.jan'- was traced (with the denotation of the period) back to PIE *gena-.In the case of the corresponding zero grade in jii-tti- ,jii- was traced to aheavy PIE *g;;-, which was thus denoted at the time.On the history of the reception of de Saussure's early work in classicalIndo-European linguistics: M. Mayrhofer, "A hundred years after thework of de Saussure," in SbHA W 1981 Bericht 8 p. 261f. (review: O.Szemerenyi in Kratylos 28 1983 [1984] p. 54-59 = Scripta Minora J 1987p. 557-562).L 319. In 1912, Albert Cuny contributed significantly by refuting theinsight, universally accepted around the turn of the century, that the zerograde of sef- roots containing -n- and -r-, were -ii- and -f-, respectively: A. Cuny "lndo-europeen el semilique" in Revue de phonhique 2 1912 p.101 If.I) His argumentation was the fo llowing: If the full grade of sel- roots infuct were to end in PIE *-a, as the understanding at that time of full grade PIE'gena- > Ved. jan'- (cf. L 318) suggested, then one should expect the zerograde PIE *gna, with vocalic -a- fo llowing consonantal -n-. This would havebeen preserved as Ved. *jiii-. However, this form does not exist. Instead, wefind Ved. jii(-tti-), which may be traced via the pre-historic transitory form*jaX-tti- to PIE *-16-.2) Cuny's treatment contains a suggestion of the consonantal nature of X,which, in the case of de Saussure's 'coefficients,' had already been acceptedby Herrnann Mtiller in 1879 / 1880. Mtiller thought of laryngea1s as they arefumiliar in Semitic languages: H. Moller Semitisch und lndogermanisch,Teil l (Konsonanten), Copenhagen 1906, p.VT:"When Ferdinand de Saussure made his brilliant discovery of; as he calledthem, the 'phonemes' A and 0 (Memoire . .. ), I immediately voiced thesupposition (1879) that these fundamental elements, to which I added athird, were consonantal, and indeed glottaIs ... and claimed in 1880 that'they were probably gutturals as exist in Semitic languages'"L 320.Current insights concerning laryngeals take into account that themeaning and use of the term 'laryngeal' have developed throughout thehistory of Indo-European linguistics without a precise phonetic connotation.It is fundamental, firstly, that the existence of three consonantal valuesis assumed, and secondly, that the majority of their effects, as they aretaught in the present introduction, already appeared in Cuny's essay of1912. At that time, Hittite and its !J phoneme had not yet been discovered.The basis of what is called '1aryngeal theory,' i.e. the continuation of deSaussure's insights, was already complete in 1912 and relied upon thecomparative analysis of root structure and ablauts, particularly in AncientGreek and Vedic. For information on the early history of laryngeal theorywith pertinent information on Cuny, Moller, and other researchers, see theexciting portrayal by O. Szemerenyi: La theorie des laryngales de SaussureCl Kurylowicz et Cl Benveniste in BSL 68 / I 1973 p. IIf. (= by thesame author, Scripta Minora 1 1987 p. 1911f.).The assertion by Jerzy Kurylowicz in 1927 that the Hittite !J correspondsperfectly to Cuny's second laryngeal had as its chief consequencethat the consonantal understanding of de Saussure's coefficients, initiallyconsidered heretical, finally gained acceptance.It is certainly of importancethat the earliest attested Indo-European branch confirms the consonantalcharacter of laryngeaIs.Nonetheless, it is worth emphasizing thatthe conclusions reached by de Saussure and Cuny based upon the nonAnatolian languages, owing to their eminent logic, would have found acceptanceeven if Anatolian texts had never come out of the earth: MayrhoferKleine Schriften II 1996 (in an essay from 1987) p. 416.L 321. The fo llowing effects of PIE oh" *h, and *h, are important forthe central lE languages and their early phases.The usual PIE root structure, as it is found in Latin and Greek formssuch as leg-o,fer-o, cap-io, cJep-o, A£t7t-OO, lCel-!1at, is composed of oneor more non-syllabic word-initial phonemes, for example 1-, kl-, k-, fo llowed by a middle -e- (or, seldom, -a-) and' finally one or more non-

112 PrO!-lnd-European Phonologysyllabic phonemes, fo r example -g, -j, _jkw. PIE roots, with an initial (cf. L322 on 'He-) or final (cf. L 323 on '-eH-) sound that must be traced to*"}, *'" or *"" easily fit into this root type thanks to their non-syllabicnature. For more information on PIE root structure: Szemen!nyi Einfuhrung1990 p. 102-104; LlV 1998 p. 5ff. (on verbal roots).L 322. Roots of the type 'He- with initial laryngeal:I) An example fo r PIE 'h,e-: - PIE *",es- 'to exist' = La!. es-I etc. Theinitial position 'h,- is evidenced in forms containing the zero grade *h,s-, cf. E504 § 9. The * h, in the full grade PIE * h,es- seems to have disappeared in allknown IE languages.2) Examples for PIE *",e-: - PIE *",e1l1- 'fuce.' Even into the ProtoIndo-European period, *'" contributed to the evolution of PIE *",e to ''']0;the lexeme thus became *h]Onl-. The fo rmer presence of PIE 'h,e is still reflectedin the lengthened grade, PIE * h,e, in which the e is preserved despitethe preceding oh" cf. L 33 1. The same conclusion may be drawn from thepresence of forms containing the ablaut '-ohr (cf. ay1(X; below): These formsare best understood if they are traced back to a time when '-ehr, with its e,provided the starting point for the ablauts -e- : -0- and -e- : -e-. The initialposition H of'h]O is preserved in AnatoIian, cf. Hitt. bant-s 'fuce, front.' Theother languages conserve *allt- directly, cf. PIE loc. sg. *h,ent-i = La!. ant-e(cf. L 423 § I) 'before,' Or. Ctvt-i (the original Or. *iivn is not attested; theactual accent is abstracted from the usual unstressed procliticaI avn before thecorresponding genetive) 'in the fuce of,' Ved. anl-i 'before.' - PIE *h,eg- (>*h]Og-; along with the fo rm PIE *h]Og- with --o-ablaut) 'to drive, lead' = La!.ag-ere, Or. ay-ElV (also compare the -0- full grade noun ay-liD

114 Proto-lndo-European Phonologyhas as an ablaut "bhoh,- with its preserved -0- preceding -hr, cf. Gr.

116 Proto-lndo-European Phonologycerning the root PIE 'deh,- 'to give' with the development, already in thePIE era > 'doh,-, cf. L 323 § 3) = Gr. E-lio-o, Ved. a-di-Ia and Gr.lio6, Lat. dalus, but (with 0) Ved. (deva-}lIa- 'given (by the gods)'(concerning the early replacement by -dalla- and dala- of -lIa-: MayrhoferEWAia I p. 715). - PIE 'ph,-Io- 'drunk' (PIE root 'peh,- becoming'poh,-; sometimes 'peh,- appears, with the addition of -j-, whichcomes from the present stem, cf. Gr. E1ll0V < 'e pih,-onl with 'pih,-

118 Prot-lnd-European Phonology2) The present tense of the PIE root *pehr (with change > *pohr) 'todrink' (cf. L 323 § 3) should be postulated as *pi-phre-ti. Ved. pibati, OldIranian ibid < 'pibeti 'drink!' (imperative, pI.) and Lat. bibil may be traced, ina first step, to 'pibeti. Both fonns, 'pibeti and 'piph,eti are accounted forwhen we asswne that PIE *h, (in this case prevocalic) lent the characteristic[+voiced] to the preceding unvoiced plosive and thus, that already in the PIEperiod, 'ph, became 'b. - In the case of PIE 'h]€p- 'water' (= Ved. cipRV+ 'water' with apas apa cidbhis and others, cp. Gr. eUpt71O

120Proto-Indo-European PhonologyProto-lndo-European Consonants 121Level II after transformation:a) *h,a- : *h,o- and• hjo- : *h30-b) *h,a- : *h,e- and* h,o- : *h,e-*-ah,- : *-oh,*-oh,- : *-oh,*-ahr : *-ehr*-oh,- : *-eh,-3) Examples: - For Hitt. mefJ-ur 'appropriate, correct time' < PIE*meh'-!!T Eicbner in §1. - For PIE *h,eIC- (becoming *h,aK-) 'pointed'and *h,aIC- with ablaut, one may cite Gr. aK-poc; 'located on the point' andOK-ptC; 'point, sharp edge,' further Hitt. !J€-kur i.e. fJek-ur 'clifltop' (However,this interpretation by H. Eichner [in MSS 31 1973 p. 71] is not uncontroversial: Melchert Anatolian Historical Phonology 1994 p. 144; Rieken NominaleStammbildung 1999 p. 287-289). - Along with PIE *seh2!!el-, becoming*sah,!!el- 'sun' (> Proto-Gr. *hii!!el-ijo- = Cre!. [Hesychius] c'tIl£A1Oc;,Epic-Ionic iJEhO Gr. PT] Ipa I pw etc.Correspondingly, preconsonantal PIE2) Additionally, there are a number of two syllable examples in Greek ofthetype ec'tvaoc; 'death' alongside the regular 9VTJ6c; 'mortal.' The explanationof this appearance of two syllables is disputed. One possible reason is the ac-cent: Thus, the development of PIE *-6h,- > *_onh,- > *-ana- etc. would beconceivable (with assimilation of the weak vowel from the series "n [itselfhaving originated in f)] into the quality of the laryngeal that immediately fo llows): Rix Historisehe Grammatik des Grieehisehen 1976 p. 73; Vine"Deverbative *-et6- " 1998 p. 12ff. (including the suggestion of another solution);C. Rico in IF !O5 2000 p. 161-200; H. Rix in Kratylos 41 1996 p. 158(concerning Latin).3) Further languages that are not treated here, such as Armenian and Tocharian,exhibit developments that vary according to the nature of the laryngeal.Compare the material for lE *rhl vs. lE *rh, and *rh,. Further reading: Mayrhofer Loutlehre 1986 p. 128.4) Correspondence sets for preconsonantal PIE *,..H , PIE *1JllI , PIE *!Hand PIE *rH:4a) PIE *f}hl = Gr. ne. In the other languages that are more thoroughlytreated here, PIE * f}H has in each case only one result: Lat. nii < *1}H,Ved. ii < *aH < *f}H (concerning *,!,H: M. Mayrhofer in Quademidell '/stituto di Glottologia 6 1994 [1995] p. 197ff.), Germ. unK < *unHK(with interconsonantal laryngeal disappearance) < *f}HK. For informationon Lithuanian and Slavic, cf. 4d). - Cf. PIE *t1Jhrto- part. 'created,born' = Gr. Kacri-yvr)o 'brother,' Lat. (g)niitus, Ved. jiita-, Got.(airpa-)kunds 'of earthly origin' (cf. L 334 § 4). - Also compare Ved.asat- < PIE *(Z-hls-1}t- above E 504 § 9.4b) PIE *lJh, = Gr. nil (ion.-Att. > ne), otherwise as in 4a (Ved. ii, Lat.nii, Germ. un etc.). - Cf PIE *dhIJl)h,-to- 'filled with smoke' = Ion. Gr.eVTJ6C; (with e < il) 'mortal' (older *'breathed out,' *'dead'; the meaning'mortal' is a secondary development from aOavmoc; 'immortal' [

122 Proto-lndo-European Phonology'use'): PIE *1J-h3Ph- > Gr. noph- yet in Myc. no-pe-re-ha i.e. nophelehaacc. pI. n. 'useless' (the same word appears later with a clearer negation asaVro$EA'; *fulna- = Got.fulls), Lith. pi/nas (in comparison with Lith. vilkas'wolf < *!!Ikwo_ =Ved. vfka-).4e) PIE "rh, / 'Ih, = Gr. ra / la, otherwise as *h, in 4d: - PIE "/irhrto- part. 'mixed' (lE root "/(erhr with Gr. aorist ICEpa-cr[cr]at) = Gr.a-lCp *-ii. V-) are missing.The source of thepreceding material is F. B. J. Kuiper: Sprache 7 1961 p. 14ff. (=Selected Writings 1997 p. 359tI).Kuiper

124 Protcrlndo--European Phonology3) The sandhi variant that is no longer attested in Vedic appears to be furnishedby Greek and Old Church Slavic. The PIE *-ahr « *-ehr) stems alsohave, along with a unifonn nom sg. -ah, > -a (cf. Gr. $11 'young woman'and OCS zena 'woman'), voc. sg. -a forms (cL Gr. cpa and OCS feno).Clear traces are missing that would confirm a PIE ablaut withfull grade *-ehrand zero grade *-hr. (For potential clues in Mycenaean: --7 I. Hajnal inFloreanl Studia Mycenaea / 1995 [1999] p. 265-276). That is why it appearsas if the differentiation between the nominative and vocative singular in thiscase could be traced to sandhi-influenced double forms that were common at atime when the stems were still composed of *-ah" and the contraction *-ahr> *-a- had not yet occurred. The fonn *-ah" at the time situated in the position-VH#K-, with the later development *-ah,.K- > -a.K-, became the nonnalnominative fonn; while -il (with actual *-a.h, V- > *-il. V-), which occurred inthe position -VH#V-, was assigned to the vocative.4) Another example of the individuality of the various IE languages in theirreception of the then extant laryngeal is furnished by a past participle participlesuch as PIE *fRlhl-IO- 'created, born' (cf. L 332 § 4a), in which various IElanguage forms of PIE *(1 first appear, and only afterward is PIE *hl eliminated:The laryngeal is retained in Vedic until the conclusion of the *n > *ilvocalization process, after which it lengthens the short vowel which it produces,forrningjala-. In the case of Greek, the same basic fonn develops via*gnOhl-Io- > *gnelo- (et: lCacri-YII'l (g)nalus. In a similar fashion as in Vedic, a *kunHda- with -un- < -niscreated in Proto-Germanic via *fRlhIIO-, with the simplification of the cosonantgroup *-nHd- to *-nd-. Cf. Got. airpa-kunds 'of earthly origin.' - Onthe possibly two-syllable Old Avestan va.ala- < *HIjliHata- cf. L 323 § Ib. On Vedic I sg. *ca-kar-Ha (the H still preserved in the period of the activityofBrugrnann's law), cf. L 412. - On Ved. na < *naHus, cf. F 318 § 6c.4. OcclusivesL 335. The faithful phonetic handing down of the Vedic language revealsa group offour plosives, also called occlusives. For example, in the realmof dentals Vedic contains the phonemes I, Ih, d and dh. The transcriptionof simple symbols in Indian script into the digraphs Ih, dh, etc. is establishedin lndology. However, these are in fact single phoneme values thatdistinguish themselves from I, and d, for example, only by the characteris-Proto-lndo-European Consonants 125tic [ + aspirated]. Thus, equivalents such as Ih, dh or I,' d' would be moreappropriate.I) This Vedic group of four provided the model for Proto-lndo-Europeanreconstructions. The Proto-Indo-European preliminary phase of the Vedicphonemes was reconstructed (here, for the sake of simplicity, dentals are used)from unchanged *1, *Ih, od, *dh2) Several reconstruction approaches for the tenuis aspirata *Ih must beabandoned in light of current knowledge of laryngeals, according to whichVedic Ih might have originated from PIE *1 + *h" cC L 329 § I. Thus, reconstructionswhich include the true PIE tenuis aspirata are seldom For example,compare PIE * skehllh- with Gr. am

126 Proto-lndo-EuTopean Phonologyindogermanische Lautverschiebung?" in Althochdeutsch I 1987 p. 3-11(p. 11: 'The explanation of a loanword such as the Germanic *rik- withinthe context of a 'new' theory bumps against so many problems that theappropriateness of the theory must be seriously questioned"); New Soundof Indo-European 1989 (The complete spectrum of glottal theory is discussedbroadly by various researchers, p. 83ff.); J. Gippert "Die Glottaltheorieund die Frage urindogermanisch-kaukasischer Sprachkantakte" inKol/oquium Pedersen Copenhagen 1993 [1994] p. 107-123; by the sameauthor, in BNF 33 1998 p. 41-45 (p. 43 "serious reservations," p. 45 "suchargumentative weaknesses"); M. Job, "Did Proto-Indo-European haveGlottalized Stops?" in Diachronica 12 1995 p. 237ff.; Lamberterie, Armenienc/assique 1992 p. 251-255 (p. 255: "Le traitement du groupe *dw(> Arm. erk-) ... oblige en outre a partir d'une sonore od, non d'unesourde glottalisee *t "'); Ch. de Lamberterie "Latin pignus et la theoriegloltalique" in Aspecls of Latin 1993 [1996] p. 135-151; F. Kammerzell inIF 104 1999 p. 234ffIn the opinion of the authors of the present work, doubts as to the justificationof the glottal approach to Proto-Indo-European have not beendismissed. Thus, glotta1s shall not be taken into account in present reconstructions.4) Along with the dentals presented above, there are labials (cf L 337f.)and tectals (cf. L 339ff.). The latter are composed of palatals, velars andlabiovelars. Following the example of H. Eichner in Kol/oquium Lat. u.Idg. Salzburg 1992 p. 65, the inventory of PIE occlusives may thus bedisplayed:5. Dentals1p b I I Iddh1I KI 1kL 336. In the Indo-European languages:I gI 1 I 1g1 1 ghI ghProt-lnd-European Consonants 127I) The PIE tenuis * t, media * d and media aspirata * dh reconstructions are,in their articulation type, preserved in Vedic. The point of articulation alsoremains the same in the case of dentals, in contrast to PIE * K > Ved. s and PIE*kw > Ved. k, cf L 339 § 2: - Compare PIE *trejes 'three' = Ved. trtiyas;PIE *deKm 'ten' = Ved. di!Sa-; PIE *dhuh2/nb- 'smoke' = Ved. dhwmi-, PIE*medhjo- ' adj. 'middle' = Ved. mtidhya-, PIE *h/rudhr6- 'red' = Ved. rudhira-.2) Only the articulation type of the PIE mediae aspiratae changes in Greek.Already in the pre-Mycenaean period, they lose their [+ voiced] characteristic.Thus, Gr. lh takes the place of PIE *dh. Concerning the examples in § I,compare tpE1

128 Proto-Indo-European Phonologydhuma-; further cf. L 313 § 3 concerning PIE *dhegn_ = Hitt. le-e-lain, loc. sg.da-ga-an 'earth'). - Yet there exists what is called the Sturtevant Rule inorder to distinguish tenuis particularly between vowels through doubling of themedia, cf. u-il-li = Lat. vel-us, cf. L 217 § I. - Rules of lenition must also betaken into account: H. Eichner in MSS 31 1973 p. 79-83 and p. 100 note56; Melchert Analolian Hislorical Phonology 1994 p. 6Of. (A tenuis followingan accented long vowel changed to a lenis articulation already in the ProtoAnatolian era, cf. PIE *dheh{ti 'stellt' > Proto-Anatolian *dMi = Lycian ladi).- Finally, what is called assibilation of I > t (common notation: z ) precedingi cf. 3'" pers. sg. e-es-zi i.e. esti = Ved. asti 'he is'. In the case of *Ih, (= Ved.Ih), the assibilation is stopped preceding i. On a similar development in Gr.dialects see § 2 above. -Concerning Hitt. Isl < PIE */1, cf. L 347 § I.6) The tenuis is preserved in Baltic and Slavic (thus, PIE *1 as I); while mediaand media aspirata merge into media (PIE *d and PIE *dh > d). CompareLith. Irys, OCS tlbje 'three'; Lith. desimtis, OCS desto 'ten'; Lith. dzmai,OCS dym 'smoke' (cf. L 211 § 7); Lith. rUdas 'reddish,' Russ. rudyj 'redhaired.'The existence of a Baltic or a Slavic version of the medialmedia aspiratadichotomy has not been confirmed: Mayrhofer Laullehre 1986 p. %note 21.6. LabialsL 337. The labial tenuis and media, PIE *p and PIE ob, are retained inmost of the lE languages. In Germanic, however, PIE *p led to I (and, incases where lex Verner applies, to 0, er. L 42 1) and PIE *b led to p. InCeltic, PIE *p first becomes *f, then, when preceding s and I, it appears asx, and otherwise as h (preceding a disappearance), cf. Irish alhir 'father'

130 Proto-Indo-European PhonologyProto-Indo-European Consonantsl313) The satem languages (Indo-Iranian, Baltic, Slavic, Armenian and Albanian)feature two separate developments of the PIE phonemes of( and ok. Thepalatal of( becomes an unvoiced sibilant (Ved. s, Lith. s, OCS s); while thevelar k remains unchanged (Ved. le, Lith. k, OCS k). - In the satem languages,the velars lead to the same resuhs as the Iabiovelars, as a rule. - Justas occurred most often in later Roman Latin, where a k preceding e or i waspaIatized (indeed centum [i.e. kentum) > Logudorian [of Sardinia) kentu, butItalian cento, French cent, etc.), k-phonemes of satem languages are subjectedto secondary paIatizations preceding front vowels. It should be noted that alsoin Indo-Iranian, *J caused the palatization of a preceding k phoneme. Onlyafter this process did *e become a, cf. L 206 § 2. Parallel to the headingcentum, the Later Avestan satam 'hundred,' usually written in the simplifiedsatem form, provided the name for this group oflanguages.4) The categorization ofIE languages into centum and satem languagesbore too much emphasis in earlier research. The affiliation ofIE languagesas centum and satem languages cannot have played a decisive role in theevolution of IE languages. Consider, for example the centum languageGreek and the satem language Armenian: Despite this difference, theyboth must be considered quite near to one another, cf. E 435 § I.An overview of research: J. TiscWer "Hunder/ Jahre ken/um-safemTheorie" in IF 95 1990 p. 63-98. Cf. particularly p. 94:"The centum-satem isogloss is not to be equated with a division ofIndo-European, but rather represents simply one isogloss among many... examples of 'centum-like aspects' in satem languages and of 'satemlikeaspects' in centum languages that may be evaluated as relics of theoriginal three-part plosive system, which otherwise was reduced everywhereto a two-part system."Cf. further G. van Driem and S. R. Sharma, "In Search of Ken/um IndoEuropeans in the Himalayas" in IF 101 1996 p. 107-146; Lipp Pala/ale1994.L 340. Correspondence sets:I) For PIE *It. - PIE *Kerd- / */ird- 'heart,' PIE */ir-n- 'hom,' PIE*(d)Krpto- 'hundred,' PIE *dehp/o- 'the tenth,' cf. L 307 § I and 3. - PIE*Jaey- 'to hear' with the participle PIE *Jau-to- 'famed' = Lat. in-cJu/us, Gr.KAm

132 Proto-Indo-European Phonologysg. aoga 'the power' « PIE * h2OlJgos) alongside instr. sg. aojOl)h-ii 'by meansof power' « PIE *h2OlJges-). In Vedic, the palatized fonn ojas- n. 'power'(which corresponds to Old Avestan aojOl)h-) was formed in an analogousmanner. Taking evidence from Iranian into account, the PIE velar *g withsecondary palatization (= Ved. and Avest.j) may clearly be distinguished fromthe PIE palatal *g (= Ved.j, but Avest. z). - PIE *jug6- n. 'yoke' (cf. L 213§ I) with the verbal * je!Jg- 'to yoke, to harness,' cf. Lat. iungere 'to connect,'Or. l;e\JyvUj.ll 'I yoke' along with Ved. y6ga- m 'the harnessing' (with thepalatized -j-, however y6j-ana- n. 'tearn (of draft animals)' and Old Avestanyaoj-ii 'I want to yoke'), Lith. jungiu 'I harness to the yoke,' OCS igo n.'yoke.'L 342.Only one retained fonn is expected in the centum languages forpalatal and velar media aspirata PIE oft and *gh In Greek, this is thetenuis aspirata kh

134 Proto-Indo-European PhonologyProto-Lndo--European Consonants135times', cp. L 203], OCS eetyre 'four'), correspond to word-initial PIE 'kW_ in*kwelyer- 'four,' and *kwo- 'who? '3) In a few cases it appears that the merging of velars and Iabiovelars insatem languages did not take place. This is important insofar as the simultaneousexistence of the three tectal groups is thus proven, cf. L 339 § I concemingLuwian.Other examples are disputed, however. The various types of vocalizationfor PIE '-rH(V)- might thus indicate that, while the g- in gir-tis 'ofthe ode (gen.)' may be traced hack to the velar PIE 'grH-es (cf. Gr. yiipu.;'voice'), the g- in Ved. gur-u- 'heavy' is traced hack to the labiovelar PIE"gWrH-u- (cf. Gr. ap). However, the different types of vocalizationcannot be interpreted as a proof of phonetic difference: The root gir- alsoprobably fe atured a labiovelar and would thus be reconstructed as PIE'gWerH- (cf. Lat. gra-I- in nom. pI. grales, Osc. bra-I- in gen. sg. braleis): Mayrhofer EW Aia 1 p. 469 (with the corrections contained in MayrhoferLaullehre 1986 p. 104f.); H. Rix in FS Narlen 2000 p. 216ff.4) In most of the Greek dialects, PIE "kW preceding 0 and a, and alsopreceding consonants, is represented as a labial by p; preceding it and i, itis represented as a dental by I; and when located near il or y, it is representedby the loss of the labial characteristic as the tectal k. Aeolic alsofeatures p preceding front vowels.Examples: - Cf. PIE "penk'e 'five' with Ion., Att., etc. nEV-t-E, butnEll-ll-tO Gr. AUKO 'naked.' - PIE '1Jgwen-'bulge' with Lat. inguen n. 'bulge in the genital region, genitals,' Gr. ooilv f.m. 'gland.' - PIE "gWihrYD- 'living' = Lat. vivus, Got. qiwans acc. pI. 'theliving,' Lith. gfNas 'living'; further Ved.fivti-; Old Persianfi:a- and OCS Zivo'living' (all three with palatalization of 'g- « 'gW_] before -,-); the case ofGr.iovo

136 Proto-Indo-European Phonology'to burn,' OCS goreli 'to burn,' Russ. fur m 'heat, emhers.' - Further, cfverbal PIE *sengwh_e_ with the nom act. *songWh-ehr: The former is continuedin Got. siggwan 'to sing, to recite'; the latter in Gr. 0111\ f. 'divine voice,oracl :' « *homp"li). The verbal zero grade form *-sTJgwh- is apparently evidentm the Homeric form aam MIh1 in the sense of 'the shield soundedupon him),' the lineage of which is to this day unclear: -7 M. Meier-Briiggerm MSS 50 1989 p. 91 ff and cf. E 508 § I. - Concerning Latin nix niv-is'snow' and ninguit 'it snows' < PIE *snejgwh_ cf. E 501.2) .The Germanic form *_gW_ is only preserved in Gothic in a post-nasalposItIon as gW, compare Got. siggwan in § I. Concerning OHG sniwan 'tosnow' < PIE * snejgwh_ cf. E 50 I § 2. Correspondence sets for the wordinitialGermanic *g- < PIE *gwh_ are only given when they show g-, 111-, orev n b- as contm lIons. Thus, as an example, PIE *gwher_ 'to bum' [above §I] IS connected WIth the Germanic fumilies ofNHG gar, warm, and brennen.However, only one of these three groups can he the correct one: -7 E. Seebold"Die Vertrelung von anlaulend idg. g'h und ghl} im Germanischen" inFachlagung Wien 1978 [1980] p. 450-484.8. Assimilations and DissimilationsL 346. Direct contact between occlusives leads to assimilation. Thus atenuis preceding a media results in a media; a media preceding a tenisresults m a tenuis. Assimilations are also otherwise produced c[ L 215 §2 and L 303.*Examples: - The zero grade of PIE *ped- 'foot,' namely *pd-, becomes-bd-, cf. LAv. acc. sg· fra-bd-am 'front foot,' cf. Or. €7n-flO-o f. in the senseof 'the day following (* at the foot of) the celebration.' However, the Greekterm IS nly superficially formed from -bd-; it is more precisely based upon the-la- ferrurune *epi-bd-jii. An exceptional phonetic development *-bdj- > -bdwouldthus be suggested: -7 Schwyzer Gr. Gr. I 1939 p. 475 and Frisk GEW. v .. -.The zero de oot form PIE *nigW_ 'to wash' (Yed. nir-nih-iintiwIU

which excludes the possibility of a causal relationship with the analogous developmentin Vedic. Thus, according to the model of Si-O{J}-!11 'I give' in thecase of PIE *dheh/- one should expect ·dhidheh/llli. The Greek -riEh1!11 ref138 Protcrlndo-European PhonologyIranian are rare, they can be seen as certain. Thus, the juxtaposition ofsuffixes such as ·-110- (in OHG s/a-dal 'barn' < Proto-Germ. ·sta-jJIa-)and ·-dhlo- (cf. La!. sta-bulum 'location, stable' with -bul- < -bl- < *-dhlorGr. X\n:A.oV 'container for pouring,' dissimilated from the older ·/d'uthlo-,and Pre-Proto-Gr. *ghu-dhlo-) may be explained in that suffixes ofthe latter type were formed from 'Bartholomae contexts,' in which phonetically*-dhlo- < ·-ddhlo- (concerning the consonant simplification ofddh l > dh.l, cf. L 407) < *-dh-tlo-. In other words, *-dhlo- can be explainedas a suffix that first appears as a -tlo- derivation of roots with finalpositiondh but then also found use elsewhere. cr also W 205 § 3.Literature and discussion: --t Mayrhofer Lautlehre 1986 p. 115ff.; H.Scharfe "Bartholomae 's Law Revisited or how the lJ.gveda is dialectallydivided" in Studien zur lndologie und lranistik 20 1996 p. 351-377.L 348. A rule of dissimilation that concerns Vedic and Greek is connectedwith the name of the mathematician, Veda-researcher, and linguistHerrnann Grassrnann. The rule applies independently to each of these twolanguages as a natural process of dissimilation that also is evidenced outsideọf the l do European language family. Efforts to assert a hereditaryrelatIOnship linking the effects of Grassrnann's Law in both Greek and Vedichave not been convincing: --t Mayrhofer Lautlehre 1986 p. 113ff.I) Concerning Vedic material: Grassmann's Law says that in the case ofan indirect sequence of two aspirated plosives (or -h-), the fir;t of them losesits aspirated quality. According to tbe example ofVed. da-da-mi 'I give' (rootdii- 'to give') one should expect from Ved. dhii- 'to lay, to place,' a presentstem ·dhti-dha-mi. The resuhing form, however, is Ved. dtidhami 'I lay.'The 2 sg. mid. dhtitse shows that *dha-dh- was the initial form: In this case-dh- preceding -s- Iost its voiced and aspirated qua1ities (·-dhs- > ·-ds- > -IS-);the dh- in the initial position retained its aspirated quality since no more anaspirated phoneme fo llowed. - cr also the root word formed from the root·dhrugh- 'to deceive, to injure,' with the nom. sg. dhrok 'damaging' (withfina1-position ·-ghs > -k[$]), hut the gen. sg. druh-as « ·drugh-as < ·dhrughas).- Further, ct: the regular Ved. budh- 'to awaken' « *bhudh-), but I sg.aor. a-bhut-s-i.2) A rule ofthe same type affects Greek. It applies to the tenues aspiratae(cf. L 336 § 2), which developed within Greek from media aspiratae -a fuctProto-Indo-European Consonants 139veaIs that the PIE .dhidheh/mi became in a first step, as yet unaffected,.thi/hemi and only afterward was subjected to aspirate dissimilation. A PIE*didhemi should have resulted in the Greek ·dithemi. - PIE ·bhlJff'U- 'ahundant'> Ved. *bhatu- and Gr. ·phakhu-, and afterward with dissimilation ruleof individual IE languages Ved. ·baghu-, becoming bahU- and Gr. pakhU(7tQxi>firs.h Gr ·dh. h > *th ikh*pinxe- > finge-, Proto-Gr. t WIt out assmann eig - e_ -,and afterward with Grassrnann > tejkh- ; Ved. With Grassmann*dh aji - > *dajjh- > deh-.Grassrnann's proof of 1862 helped at the time to resolve an apparentdiscrepancy in the Germanic consonant shift: --t Szemert!nyi Einfiihrung1990 p. 20. The PIE *bhendh- 'to bind' serves as a good example: ThatGothic bindan and not *pindan corresponds to the Vedic bandh- 'to hind'is explained by the fact that the basic PIE is the not yet dissimulated*bhendh- and not (as researchers earlier believed) the dissimulated*bendh-. Working from basis of PIE .bhendh-, Proto-Germanic *bend- isunderstandable. _ However, * bhandh- leads to the dissimulated formbandh- in the case of Vedic. In contrast, the initial position of the futureform bhantsyati remained untouched, because the phonetic sequence.-dhsya- of the form *bhandh-sya- resulted immediately in non-aspirated-tsya-, through exclusion of the aspirated quality in the group *-dhsya- andthrough assimilation of * -dsya- > * -tsya-, cf. also dhatse and abhutsi in §1. -In Greek, the development of *bhendh- may be traced via *phenth- topenth-, cf. 1lEv1}-Ep6 'father-in-law' (this term for a fami1ial relationshipbelongs, via the image of relational 'fustening,' to 'to bind,' cf. the similarVed. btindhu- 'family member').

140 Proto-Indo-European PhonologyLarger Phonetic Unities141D. Larger Phonetic Unitiesdu discours, elle a eu pendant tres longtemps un nom glorieux: la RMtorique"1. Word, Sentence, Text2. Accent; Word-initial, Word-fmal and Sandhi PhenomenaL 400. The word: Although the word is the most important element of alanguage (cf W 100), it exists in the larger context of sentences and textand receives from these only then its legitimation and definition. "The unit'word' must exist, in order that on this basis the signals of sentence hierarchyor sentence perspective may be ... realized." : 4 Hj . Seiler in SeilerSprache und Sprachen 1977 (in a contribution of 1962) p. 39; cf. by thesame author, "On defining the word" loc. cit. (in a contribution of 1964)p. 70-73.However, the definitions of what words really are, are even more multilayered:4 Bussrnann Lexikon d. Sprachw. 1990 p. 849f under theheading 'Wort.'In this context, the following contributions are interesting:4 A. Morpurgo Davies "Folk-linguistics and the Greek word" in FSHoenigswald 1987 p. 263-280 (for example, for information on wordseparatorsin Mycenaean ["mainly determined by accentual criteria"], andon literary and grammatical opinions); LALIES 10 1992 with differentcontributions regarding 'Le mol' (Cf, among others, I. Lallot "Le moldans la Iradition pregrammaticale el grammaticale en Grece"; M.-I.Reichler-Beguelin "Perceplion du mol graphique dans quelques systemessyllabiques el alphabetiques"; G.-J. Pinault "Le mol el ['analyse morphologiqueselon la grammaire indienne").L 401. The sentence:Following are a couple of cross-references regardingthe sentence level and the possibility of syntactical reconstruction,cf S 102; regarding sentence accent and sentence-initial position, cf. S209; further, cf. F 214.L 402. Text: For information on the text level and its syntax, cf. S 200.R. Barthes goes beyond the text level. He emphasizes correctly that inaddition to sentence syntax, the discourse has its own syntax (4 MeierBrilgger Gr. Sprachw. I 1992 p. 97f.):"le discours lui-meme (comme ensemble de phrases) est organise ...le discours a ses unites, ses regles, sa 'grarnmaire' ... Cette linguistiqueL 403. Word accent is a central characteristic of the word. Assertionsregarding Proto-Indo European word accent are possible, cf L 419.Further, individual Indo-European languages exhibit more or less highdegrees of restriction and simplification in word endings: - Latin: Consonantgroups are simplified for the most part, cf. nom. sg. n. cor 'heart'

p142 Proto-lndo-European Phonologyers' Law. According to Sievers' rule, the variant PIE *diiJus of PIE*djJ1js is determined by the final position of the preceding o;d: On theone hand, we have ## ... V#dijJ1js ## and ## .. .vR#dijJ1jS##, on the other## .. .K#*djJ!!S##, cf. L 218. - Another possible example is the mobile -softhe type PIE *(s)teg- 'to cover, to spread over' (-7 LIV 1998 p. 535),cf. Lat. tego vs. Gr. cr-reyoo: -7 Mayrhofer Lautlehre 1986 p. 120 ("inother cases, sandhi -S# #K- may have led to #sK-") and cf. L 311.Occasionally the supposition of prehistoric sandhi can be helpful in theunderstanding of phenomena in individual IE languages. - Cf. the diverginginitial position ofthe Greek pronoun -IV vs. V-IV (cf. F 406), onthe subject of which J. Wackernagel in Wackernagel Kleine Schriften I1969 (in a contribution of 1892) p. 10) comments: "It seems to me ...easiest to attribute - and v- to sandhi." - In the case of the problematicGreek initial position p(t)- (cf. 1I0AtI; 1 1ITOAIC; 'city'; which, according tothe evidence of the proper noun po-to-ri-jo i.e. Ptolijon including the element'city,' dates from Mycenaean) O. Szemen,nyi also appeals to prehistoricsandhi phenomena, considering exemplary phrases such as *Jluthetp6lin 'he came into the city' [with the preserved -t in the Greek final position1 and its regrouping to Jluthe *tpolin, and then, with consonant positionreversal, pt6lin): -7 Szemen:nyi Scripta Minora III 1987 (in a contribtionof 1975 [1979]) p. 1491f. Whether his supposition is correct, isdIsputable.On futther sandhi phenomena, cf. L 334 § 2-3.3. SyllableL 406. Between the levels of entire words and individual phonemes,there are syllables. General information: -7 Bussrnann Lexikon d.Sprachw. 1990 p. 684f.In Ternes' Phonologie 1987 p. 170-172, it is claimed:"A syllable is composed of an obligatory syllable core and a (fucultative)syllable margin . ... The syllable margin is composed, as a rule,of one or more consonants that precede and/or follow the syllable core.The syllable margin is facultative insofar as a word can also be composedof only a vocalic syllable core . ... The description of the syllabicstructure of a language is hardly less important than the determining ofits phoneme system."Larger Phonetic Unities 143The limits and division of syllables is handled variously depending onperspectives and approaches, including those of orthography, pronunciation,metrics, and theory.A syllable is considered open when the syllable limit is placed immediatelyafter the vocalic syllable core. A syllable is considered (naturally)metrically long or short according to the long or short quality of the core.Word internal syllables are 'closed' when the syllable limit immediatelyfo llows one or more consonants, which would then form the syllable margin.The vocalic syllable core may then be long or short, but for the sakeof metrics, every closed syllable is long (by position).The consonant group muta (i.e. occlusive) cum liquida plays a specialrole. It can be shown that in prehistoric times and in Proto-IndoEuropean in particular, the syllable division was made between mula andliquida, and that in Latin and Greek it could be before the muta and liquida,and thus no longer represented a position that could affect the lengthof a vowel (i.e. correptio altica). Cf. the Latin nom. pI. integri 'undamaged,'of which the syllabification is attested in poetry as in. te.gri as wellas in. teg.ri. The word integri may be traced back to *in-tag-ro-. Thevowel weakening of -tag- to -teg- shows that at the time it took place.tag. was a closed syllable: Had the syllable been open at the time, a .ta.would have led to .ti., yielding *intigri. However, no traces of this exist.Further sources: - Latin: -7 Leumann LLFL 1977 p. 21 f. - Greek:Changes in syllabification are possible, cf. again the muta cum liquidagroup, in which both lIOT.pOC; and 110. TpOC; are attested. Compare futthergenitive singulars ending in -ojjo < *-osjo, in which likely both -oj.jo (=Myc. -o-jo and Aeol. -Oto) and -o.jjo I -o.jo are evidenced (the latter fo llowing the disappearance of j in the [onicl Attic contraction to -ou, and theDoric contraction to -(0). However, the facts are not indisputable. literature:-7 Meier-Brtigger Gr. Sprachw. IT 1992 p. 105f. and p. 79f.L 407. Clues as to the syllabic structure of Proto-Indo-European may bediscovered indirectly. Thus, Gr. TPOV (attested from Homer on) 'measure'can be understood as a -tro- derivation from PIE *med- 'to measure'(-7 LIV 1998 p. 380f.), if PIE *med-tro- or, assimilated, *met-tra- i.e.*melt.ra- (with simplification from *melt. to *met. ) was valid. B. Forssmanremains skeptical: -7 Kratylos 33 1988 p. 63). Similarly, (-7 Got6 l.PrasenskIasse [987 p. 218 note 454) the Vedic imperative bodhi (RV) isoften traced to PIE *bhe!fd-dhi, i.e. *bhe!!.dhdhi, with .dhdh simplified to.dhi. But the facts in this case are deceptive. The questionable form bodhimay be better understood as originating from Vedic bhliva (> bho) +dhi:

144Proto-lndo-European PhonologyLarger Phonetic Unities145 St. W. Jamison "Syntactic Constraints on Morphological Change: TheVedic Imperatives bodhi, dehi, and dheht' in Syntaxe des langues indoiraniennesanciennes 1993 [1997] p. 63-80.If the explanation of the type *metl.ro- is correct, then the instances ofsyllabification of the type VKK. RV and VR. KKV may be traced to theProto-Indo-European period: Saussure Recueil 1922 (in an essay from1889) p. 420ff. (with further examples); Mayrhofer Lautlehre 1986 p.III fFurther information on PIE syllabification may be derived from thevariant forms of the type -j-V- and -ij- V-. J. Schindler makes this clear (Sprache 23 1977 p. 56-65 in his review of Seebold Halbvokale 1972; cfL 218). Namely, Sievers' Law was valid in the fo llowing context: R > I# .... V.K _ VK.# and # .. . vR.K _ VK.#, compare with the conspicuously constantconsonantal sonant, Ved. mat.sya- (RV +) 'fish' and iir.dhva- (RV+) 'upright' « *ljrH.d'ljo-). Further reading: Szemerenyi Ein/iihrung1990 p. 11 Off.The detailed discussion by J. Schindler of problems surrounding theapparent PIE fricative *p (cf L 313) secondarily include further assertionsconcerning PIE syllabic structures: "A Thorny Problem" in Sprache23 1977 p.33f. (with discussion of the example of PIE *hzer.tlW- vs.*h,rl. lW- 'bear' among others).4. AblautL 408.All ancient lE languages exhlbit regularly a quantitative or qualitativevowel change in certain groups of fo rms, such as in different casesof a substantive, or derivations from a common root.An example: The suffixal element --rep- in the Greek word fo r 'father'appears with -e- in the accusative singular 7ta--rep-a, but with -0- in thecompound accusative singular a-7ta-op-a 'someone who has no father.'The vowel is absent in the dative singular 7ta-p-i. The dative plural7ta-pa-m '(to) our fo refathers' contains the continuation of -Ir- (cf. Ved.pi-If-$u '[with] the ancestors'), the expected syllabic alIophone of *-trthatone expects to find before the consonant-initial -m. The nominativesingular 7ta-n\p reveals a quantitative 'increase' of -e- to the correspondinglong vowel; the nominative singular a-7t6-OlP reveals the same, butwith the corresponding -0- fo rm.This change clearly did not initiate in Greek, but rather reveals itself asa PIE inheritance through parallels in other lE languages.Compare7ta-rep-a, 7tatpi, 7tan\p with Ved. acc. sg. pi-Iar-am 'father,' dat. sg.pi-tr-e, nom. sg. pi-la (with -la < '-Iar, cf. L 301 § 3).L 409. Two facts are essential concerning this change: The change isdatable to the PIE period and is limited to a few fo rmal possibilities whichare used in morphology. The above case of 7tan\p I a7t6Olp includes onlythe qualitative changes of e and 0, and the three quantities e I 0, e I 0 andzero (0).In a language that, in terms of vowel quality and vowel quantity, haschanged as little as Greek, this consistent regularity may be demonstratedby citing just a fe w phonological-historical examples. This regular changeof vowel quantity and quality, called ablaut, may be subdivided on twofurther levels: qualitatively as e/o- alternation; and qualitatively as full, or'normal' grade, lengthened grade, and zero or 0 grade.Compare the fo llowing schematic depiction:eoa7t6-op-ae7ta-p-ina-'t a-mL 410. The table in L 409 shows five grades, all of which are representedin a single word family in a very fe w cases, such as that of 7ta-rep- I67t6op-: F(e) = full grade with -e-; F(o) = full grade with -0- nuancing;L(e) = lengthened grade with -e-; L(o) = lengthened grade with -0-; Z (0)= zero grade. - Examples: - In Greek, compare the evidence of fo urgrades in n£-oal 'I fly' = F(e), 7to-eOal '1 fly, flap' = F(o), 7tOl-"EI'flapping' = L(o) and 7t-Eaeat 'flying up' = Z(ol; the fifth formal possibilityL(e) = *1t11- was not used in any known fo rm.The changes elo, elo and fJ may be demonstrated in many further rootswith intemal -e-, also in those in which -j- or -n- fo llows -e-, or in which-r- precedes the -e-. - Examples: - Cf Gr. }£i7t-Ol 'I leave behind' =F(e), AOI7t-6 'staying behind' = F(o), (e)-At7t-E 'he left' = Z(o) . - Cf-rev-m 'I will harness' = F(e); 6v-o m. 'tension, string, note' = F(o) ; 6-m f 'tension, stretching' with la- < *11}- = Z(o) (concerning I}, cf. L 305)._ Cf. pe7t-Ol 'I turn' = F(e) ; "tp67t-o m. 'turn' = F(o) ; "tPOl7t-60l 'I twist,turn, change' = L(o); EU-pa1t-EAo '(lightly) twisting, moveable' with-Iro- < *-Ir- = Z(o).

146 Proto-lndo-European PhonologyLarger Phonetic Unities147L 411. The changes between e / 0 / e / 0 /0 are also visible in other lElanguages.than in Greek.However, one must expect much greater phonetic changesFollowing are examples from Latin, in which phonetic changes amongvowels and diphthongs have obscured the original ablaut relationships: -Ideal cases with unchanged PIE values include, fo r example, Lat. leg-o 'Icover, hide' = F(e) , leg-ula f. 'rooftile' = L(e) and log-a f. 'article ofclothing, toga' = F(o). On the other hand, along withfoedus n. 'alliance'(cf. Old Latinfoideralei 'allied') = F(o) andjides f. 'faith' = Z(o) ,feid- =F(e) is only by chance attested in early Latin (cf. diWeidens from inscriptions):In 150 B.C., the normal development led from ej via e to , andthus to jido 'I trust.' A simultaneous jido is no longer identifiable as F(e) .- Cf. as a further example Lat. fer-o 'I carry' = F(e) , fors. for-ti- f.'chance, providence' = Z(o) with -or- < *-r- (a simultaneous ablautchange fer- / for- appears; however, -li- derivations traditionally have zerograde roots, CL Ved. bhr-li- f. 'maintenance' and NHG Ge-bur-I; thus, allseems to indicate the presence of a former *-r-). Also, in the case of therelated root word for m. 'thief,' the L(o) is no longer recognizable, becausein the phonetically expected *for = Gr. q,o\p 'thief, one who carriesaway' 0 has been replaced by U.The case of Latin me-min-' 'I bear something in mind' is also not immediatelyclear:Although mon-eo 'I warn' attests F(o) in the first syllable,the middle-syllable F(o) of *me-mon- (cf. Gr. Ilov-a 'I commemorate')is changed through vowel weakening to-min- (cf. L 204). _Further, the ablaut change *-en- : *-(1- is no longer discernible becauseevery *-(1- became -en-: Cf. the -ti- abstractum men-/i- (mens) f.'thought,' in which it is no longer clear whether the stem may be traced toan Z(o) of *m(l-ti- as expected, or whether the F(e) of *men-li providesthe basis, cf. F 317 § 7.L 412.In Vedic and its Indo-Iranian preliminary phase, since the differencebetween PIE *e, *0 and *a as well as that between PIE *e *0 and *iiwas eliminated, only the quantitative ablaut is clearly disceble, cf. L206.Through linguistic comparison, however, one may detect in some developmentsthe pre-Proto-Indo-Iranian passage of the qualitative ablautfrom PIE *J to PIE *0. An example of this may be fo und in the effects ofthe palatal law, cf. L 206 § 2. Thus, the ablaut PIE *-ge- : *-go- is reflectedin Old Avestan as -Ja- : -ga-, cf. L 341 § 2. Other example: TheVedic 3 sg. perfect form ca-kiir-a 'has done,' with its ca-, suggests *-e- inthe reduplication syllable, but with -kiir-, suggests *-0- in the root syllable,as do the Greek perfect fonns A.£-AOl1t-E and oe-OOplC-E.Further information on Vedic -kiir-: In this case, PIE *-0- clearly beearne1ndo-Iranian -ii-, and thus reveals its contrast to PIE *-e- > IndoIranian -a-. Called Brugmann's Law, this assertion explains the differencebetween the Vedic acc. sg. dii-Iar-am 'donor' and acc. sg. pi-Iar-am 'father'as a reflection of the PIE qualitative vowel difference *-Ior- : *-Ier-,compare Greek oo\-top-a with 1ta-'r£p-a. However, the details are morecomplex. As a rule, this affected open syllables. Proof of this is providedby the I" person singular of the same perfect fo rm, caktira, in which theoriginal vowel in the middle syllable can only have been -0-. However, the-a ending of the I" person singular, unlike the 3'd person singular (withPIE *-e), may be traced hack to *-Ha < PIE *-h2e. In the period of act ivityof Brugmann's Law, the I " person singular with its *-Ha fo rmed aclosed middle syllable, thus *-/car. Ha. While in the 3 sg. *-/ca.ra became-kii-ra, *-/car.Ha remained intact and did not even change when the laryngealH disappeared.Further literature: Mayrhofer Laullehre 1986 p. 146 ff; CollingeLaws 1985 p. 13ff. ; Vol.kart Brugmanns Geselz 1994; I. Hajnal "DasBrugmannsche Geselz in diachroner Sichl und seine Giilligkeit innerhalbder arischen a-Sliimme" in HS 107 1994 p. 194-221 .L 413. The insightful Indian grammarians brought the quantitative ablaut,the only one known to them, into a system, of which the zero grade(such as bhr- 'to carry,' for example in the participle bhr-Ia- 'carried')was the starting point.I) To the zero grade basic fonns, -a- is added a first grade, called GUI,Ja(presumably in the sense of 'high grade'), cf. bhtir-aQO- n. 'burden.'Thehighest grade, called Vrddhi (in the sense of 'increase, growth'), adds -a- tothe root, cf. bhiir-ia- 'to be carried, to be cared for.' A further prime example: - Concerningji- 'to be victorious over' (cf. Ved.ji-Ia- 'conquers') comparewith the GUI)a *jaj- (= prevocalic Ved.jay-, but preconsonantalje-, cf L220 § 3; cf jay-U$- 'victorious,' je-Iar- m. 'victor') and Vrddhijiij- (= Ved.prevocalic j4)l-, preconsonantal jai-; cf. j4)l-u- 'victorious,' jai-tra- 'victorybringing').- Corresponding to sru- 'to hear' (Ved. STUla-) compare withGUI)a *sray- (= Ved. prevocalic Srov-, preconsonantal Sro-; cf srtiv-as- n.'fume,' Sro-Iar- 'Horer') and 'Vrddhi' *Srii!!- (= Ved. prevocalic Sriiv-, preconsonantalsrau-; cf Sriiv-tiyati 'renders audible,' a-srau-, aor.heard').'you

148 Proto-lndo-European PhonologyLarger Phonetic Unities1492) But because of its adoption of the zero grade as the fundamental rootform, this astute observation of the Indian ablaut grade system was not alwayssuccessful. In instances such as man- 'to think' or pat- 'to fly,' because thezero grade fonns in cases like *lnl)- > ma- (cC L 306 § 3) or pt- (in the aoristfo rm a-pa-pt-at) were not discermble to the Indian grammarians, tbey wereforced to assert full grades as root fonns.The case svap- 'to sleep' (withsvap-na- m 'sleep') is similar, in which a gradation based upon the zero gradesup- (sup-ta- 'slept,' Gr. ilrr-vo PIE *Jilu- and > PIE *sup-, produces no problems.L 414. In Hittite, the set of mles regarding the quantitive ablaut is quiteclear, for example the change between e vs. 0 in verbal paradigms, cf. 3sg. ku-en-zi i.e. lajen-t'i 'he kills' vs. 3 pI. ku-na-an-zi i.e. kun-ant'i 'theykill' (=Ved. han-ti vs. ghn-anti, cC L 345 § I). - Laryngeal effects alsoplay a role (cC L 314ff.): Cf. concerning PIE *"Ies-ti 'he is' (=Lat. estetc., Hitt. e-es-zi i.e. es-t'i < *es-t'i; the so-called plene-spelling is explainedas a lengthening due to the accent) the zero grade plural ""Is-enti= a-sa-an-zi i.e. as-ant'i 'they are' = Myc. e-e-si i.e. e"ensi, but Ved.santi, NHG sind, cf. E 504 § II and cf. L 419 § 3 below. - Concerningthe important statements in Hittite (Anatolian) on the lengthened gradesPIE *",e and *eh" cf. L 331.L 415. In Germanic, the ablaut was retained and expanded as a centralmeans of expression in the realm of morphology.An old Germanic languagewithout ablaut (i.e. strong) verbs is unimaginable. In new Germaniclanguages such as English, there are form series (e.g. sing : sang : sung :song) in which the ablaut is the only remaining difference. Along withrevelations in the Indian grade system (cf. L 413), it was through work inGermanic that helped early Indo-European linguistics in its understandingof the ablaut systemThe uniform PIE ablaut schema was transformed by the numerous phoneticchanges that took place concerning vowels, diphthongs, and also r. /.l'[l. and I). For instance, NHG binden : band : ge-bunden may be tracedback to PIE * bhendh- : * bhondh_ : * bh1Jdh- and may be associated with thegrades F(e), F(o) and Z(o), cf. L 410.Further examples: - Got. niman 'to take, to take up' < PIE *nem- =F(e), Got. nam 'took' < *nom- = F(o), Got. nemun 'they took' with fo r e = L(e) , Got. in-numan 'taken in' < *nl'[l- = ZM. - Got. wairjJan 'tobecome' < PIE "yert- = F(e), warjJ 'became' < *'Jort- = F(o), waurjJun'they became' < *lJrl- = Z(o) (concerning Gothic -or- < Proto-Germ.*-ur-, cf. L 208). - Got. tiuhan 'to pull, to lead' < PIE *del}k- = F(e),ga-tauhans part. 'pulled' < *duk- = Z(o) (concerning Gothic -oh- < Germ.*-uh-, cf. L 208).The various Germanic continuations of the uniform PIE ablaut typessuch as *er *or *r or *el} *oy *u are the major reason fo r the differencesbetween the first five series of strong verbs in the Old Germanic languages.L 416. Many instances of inherited ablauts are also attested in the Balticand Slavic languages: - Cf. Lith. tew, OCS teko 'I run, flow' = F(e);Lith. takas m 'canal, pipe, barrel,' OCS toh 'run, flow, threshing floor'F(o) ; Lith. leU 'deep place in a river,' OCS lex" 'ran' < *Iek[-s-] = L(e) ;Russ.-CS lakati 'to drive an animal' = L(o) . - Cf. Lith. merifti 'to starve,'OCS mreli 'to die' < PIE *mer- = F(e) ; Lith. maras 'epidemic,' OCSmoro 'plague' = F(o); Lith. mirtis, OCS s,,-mrolb f. 'death' < *mr- = Z(o) .- Slavic even exhibits the ideal case: all five possible ablaut grades,cC L410: OCS greb-o 'I dig, scratch' = F(e), grob" m. 'grave' = F(o), gres"'dug' < *greb-s- = L(e) , grabit; 'to steal' = L(o) , Cz. po-hfbili 'buried'

150 Proto-lndo-EuTopean PhonologyLarger Phonetic Unities151Further, Ved. yate 'drives' comes from PIE *h2f'g- (Gr. ayw etc.), ifin fuet itmay be traced to the reduplicated present stem *h1i-h1g-e-: -7 K. Strunk inMayrhofer EWAia I p. 51.2) The understanding presented above (L 323) of roots such as Gr.(ti-)!h](-t), ((-)O'ta(-t) and (oi-)&o(-t) as originating from PIE *dhehr,* stehr and *dehr leads one to abandon the hypothesis of a separate ablauttype with primary long vowel endings, the zero grade of which would be -a-,cf. '(-O'ta-t : O'ta- (= Ved. sthi-ta-). In fuct, the ablaut e / 0 / e / 0/0 ispresent even bere. - Examples: - PIE *dhehr > Gr. -!h]- = F(e), PIE*dhoh,- > Gr. SorT] f. 'punishment, sentencing' = F(o) and PIE *dhh,_ > Gr.E-e.:-O 'has sat,' Ved. a-dhi-ta = 2(0) (concerning the vowels, cf. L 325). -PIE *bhehr 'to speak' (becorning PIE *bhahr) = Myc. Gr. 3 sg. pa-si i.e.phiisi 'he c1aims'= F(e) ; PIE *bhohr = Gr. $

152 Prot-lnd-European PhonologyLarger Phonetic Unities1534) The nuancing of *-e- to *-0- (cf. Gr. 1ta--ocp-, but -1t6-op-, cf. L 408)may no longer be, as was for a long time suggested, traced back to an effect ofaccent that is separate from the origin of the zero grade. Nevertheless, it mayhave phonetic causes. Thus, the game of qualitative ablauts in thematic verbs(cf. Gr. q£p-o-!I£v, but q£p-E-1E) creates the impression that post-tonic -eprecedingr (and I m n) might have become -0-: Cf. in addition to Gr. 6naop-a also PIE *de-dor/{- 'have seen' = Ved. dadIu:i- (Gr. with secondarystress OCOopK-): It is explained by Pre-PIE I" person sg. *dederk-h,e with*-e- > *-0- preceding -h,e: --t W. Manczak in Lingua 9 p. 1960 277ff. Forinformation on the history of research on this subject, cf. Szemen!nyi Einfuhrung1990 p. 124-127.5. AccentL 419. Only the independent word accent is discussed here. For informationon nominal accent classes, cf. F 314ff.accents cf. F 214. On sentence accent, CL S 209.For information on verbalI) Among the lE languages, some of those whose surviving documents doreveal information ahout accent placement show free accent stress, i.e. accentstress that is independent of word structure.A Greek word such as opoo: 'useful, profitable'). In Vedic, the gen. sg. brahmaf)GSoccurs with stress on the first syllable (brahma(las: 'the prayer formula') aswell as on the second syllable (brahma(lOs: in the sense of 'of one who prays,of the priest'). These examples show that the supra-segmental phenomenonof accents placed ahove phonemic segments is the only means of distinguishingsuch pairs of cognates.2) It is important for the historical comparative observation of word accentsthat these free accents correspond in several languages, a fuet which canonly be explained by the existence of a common prelirninary phase. However,the agreements with word structure independent accents have been obscuredby language-internal rules.Research by Karl Verner on the effects of PIEaccent placement on the Germanic fricative system has revealed Germanic as afurther essential source of information on the original accents, cf. L 42 1. Thefree accent systems of Baltic and several Slavic languages may be traced tosubsequent developments of the inherited status, cf. L 422. Among the languagesthat have exchanged the PIE accent in favor of a secondary accentsystem, Latin is the most important, cf. L 423.3) Concerning the efforts to ascertain PIE accents fro the basis of photicresults in IE languages that have no certain graphic representatIon of:ccent positions, a few hints shall have to suffice. - For informatin.on Htrite,et: L 207. Further, cf. L 414 above, where in the example e-eS-Zl ;e. ti the singular plene-spelling may be understood as the resuh of h,es-,above which is placed the accent; cf. L 336 § 5 above, in which the lenitionrule requires the presence of a preceding accented long vowel. Literature: .--tMelchert Anatolian Historical Phonology 1994 p. 106f. - For informatIonon implications from Avestan and PaSt6: --t M. Mayrhofer in CompendiumLing. Iran. 1989 p. 12f.L 420.placement.Many Greek - Vedic word pairs show full agreement in accentI) Ct: the familial terms Gr. 1tO-ocp-a = Ved. pitar-am 'father,' cppcr1Ep-a =Ved. bhrtitar-am 'brother,' 9uya-ocp-a = Ved. duhitar-am 'daughter,'11'1-ocp-a = Ved. matar-am 'mother' (in each case acc. sg.); Gr. ve no ,Ved. nabhas- n. 'cloud'; Gr. 9'0116

154 Proto-indo-European PhonologyLarger Phonetic Unities155on the next to last syllable, cf. ltatpoKt6v 'killing the father') as in anexternal comparison (the form, *poikilo- 'colorful,' similar in morphologyto Ved. pesa/Q- 'beautiful, decorated' becomes, according to the dactylosrule, ltotKiA.o). Literature: Meier-Briigger Gr. Sprachw. n 1992 p. 39.4) For further information on the accent in Greek and its peculiarities: Risch Kl. Schr. 1981 (in an essay from 1975) p. 187ff.; M. MeierBriigger in Fachlagung Leiden 1987 [1992) p. 283ff.L 421. According to Verner's Law, published in 1876, the retention ofan unvoiced word-internal fricative in Germanic proves that in an accentuation,similar to that of Vedic and Greek, and which was still present inthe oldest Germanic, the stress was placed intmediately before the fricatives: K. Verner "Eine Ausnahme der erslen Lautverschiebung" in KZ23 1876 p. 97-130. According to an account by O. Jespersen ( PorIrails I 1966 p. 539), Verner made his discovery shortly before taking anap, as he briefly leafed through Comparalive Grammar by F. Bopp. Hisglance fe ll upon pilar- vs. bhralar-: "It struck me that it was strange thatthe one word had a I in the Germanic languages and the other a Ih ... andthen I noticed the accent-marks on the Sanscrit words": Further reading: A. Calabrese and M. Halle, "Grimm 's and Verner's Laws: A New Perspective" in FS Walkins 1998 p. 47-62; K.-H. Mottausch in IF 104 1999p. 46ff.1) Pro to-Germanic f, p, z ;(" s < PIE p, I, k, kW, s were only retained whenthey intmediately fo llowed a word-internal accent in a voiced context, cf. L336 § 4. In all other cases they became voiced fricatives (0, d, g, g ' , z). Theretained Germanic -J;- in the Gothic brojJar 'brother' thus suggests ProtoGermanic *or6jJer-, the accent position of which agrees with that of Ved.bhralar-, Gr.

156 Proto-Indo-European PhonologyLarger Phonetic Unities157Akzent.Ablaut 1968 p. Ill, Kurylowicz writes "It is this mobility and relativefreedom that inspire the comparative linguist to search for a historicalcontext [which relates the Balto-Slavic and Vedic-Greek (Germanic) accentpositions)."However the inherited situation in Balto-Slavic wassubjected to several changes, concerning the dimensions of which as yetno unanimity has been achieved.Thus, it remains disputed whether the accent position variations in Lith.dukte 'daughter,' gen. sg. (old) dukteres, and acc. sg. dilkteri are an echoof the PIE (Ved.) accent positions: compare -te with Ved. duhita 'daughter,'compare -teres (for "duktr-es) with the weak case of the type Ved.dat. sg. duhitr-e and Gr. gen. pI. avyap-6iv, compare dilkteri (with regularaccent "dukteri) with Ved. duhitar-am and Gr. avya-rep-a. Nevertheless,in most cases one must expect further developments that in principlewill be traceable to Proto-Indo-European forms.Accent positions thatremain the same, such as Russ. nebo, Serbo-Croatian nebo 'sky' PIE"mibhos (Gr. VE, Ved. nabhas 'cloud,' cf. L 338), resemble curiouscomcidences. One is tempted to formulate rules, for instance that a syllablecontaining a laryngeal, followed by a stressed syllable in Proto-IndoEuropean, pulled the accent upon itself in Baltic and Slavic: for example,PIE "dhuhrmo- (Ved. dhuma-, Gr. au6, cf. L 211 § 7) Baltic andSlavic dtim-, Lith. dtimai, Russ. dym (gen. sg. dym-a), Serbo-Croatian dim'smoke'; PIE "gWriH-!!ehr (Ved. grlva- f. 'neck') Latvian griva 'rivermouth,' Russ. griva, Serbo-Croatian griva 'mane.' If such a result werecombined with the continuation of the PIE accent position in syllables thatcontain laryngeals (cf. PIE "!!/H-nahr 'wool' [Ved. tir(Zo-) > Latvianvilna, Russ. volna, Serbo-Croatian vlina 'wool'), the value of assertionsbased on Baltic and Slavic regarding PIE accent positions would be relativized.Although the movable and free accents in Baltic and Slavic havethe PIE accent positions as a starting point, at the most they can serve toconfirm in a case by case manner the assertions already made based on theevidence in Vedic, Greek, and Germanic.Further reading: -4 Kurylowicz Akzent.Ablaut 1968 p. III f., p. 123;Stang Vg!. Gramm. 1966 p. 134f.; Kortlandt Slavic Accentuation 1975;Illic-Svityc Nominal Accentuation 1979; Collinge Laws 1985; 1. Hajnal inHS 109 1996 p. 314-3 16 (a review of R. Derksen, Metatony in BalticAmsterdam / Atlanta 1996).L 423. In the case of Latin, no obvious traces of PIE (i.e. Vedic andGreek) accent positions remain. At most is fo und indirect evidence, cf § I.The original PIE accent positions were replaced in a first step by the initialaccent (cf. § 3), from which Classical Latin then broke away (cf. § 2).I) The tate ofword-final -i may provide a clue as to the original PIE accentposition: In non-stressed cases such as est < PIE "h,esti the -i is dropped,whereas in cases such as the final stressed loc. sg. pede < PIE "ped-i, it maybe preserved: -4 H. Rix in Kratylos 41 1996 p. 158 note 7 (based upon casessuch as "ped-i , the locative -e is generalized, even when [as m the case ofante < "hzent-i, cf. L 322 § 2) in the paradigm it was not accented).2) Classical Latin is subjected to a rule that is dependent on the form of theword, according to which two-syllable words that are stressed carry the accenton the second to last syllable (paenultima), while words with more than twosyllables carry the accent on the paenultima when it is long (either by positionor by nature), otherwise the accent is on the third to last syllable, or antepaenultima:regit, regunt; regebat, regUntur; regitur.3) However this system of accent placement must have been preceded byan older one, which was equally restricted, but which fixed the accent on thefirst syllable of the word. Word stress in the very oldest Latin was thus:"faCio, "con-facio, "con-factom. Since word internal short "-a- in an opensyllable becomes -i-, and in a closed syllable becomes -e- (see L 204), the twolatter forms resulted in Lat. conficio and con/ectum. As further examples, cf.ago : ex-igo « "ex-agO), sedeo : ob-sideo ("ob-sed-, cf. L \08), talentum «"talantom, a loanword from the Greek aA.avov), and finally igitur 'thus,because' (separated from the question formula qUid igitur < quid agitur 'whatis it about, what then?)' - Information about the prehistoric initial accent(which reached into the historical era) as well as its effects is important fo rlinguistic comparison since the results that one would expect are often affectedby the initial accent: PIE "genh,tor- (cf. F 101 § 2) becomes, in the earliestLatin, "genator, which then becomes genitor (cp. "tixago > exigO). That Lat.short "-a-, in a closed syllable becomes -e-, is illustrated by the feminine formgenetr"ix < "genatrik- ( Ved. janitri- f. 'one that gives birth'). The Lat."(ab-), "(con)-datos « PIE "dhh,tO-, cf. L 325), which corresponds to theGreek 6E, accordingly appears as ab-ditus, con-ditus. - Aliterative fo rmulae from religious and juridical texts that have been handed down are finalallusions from the classical period to this older epoch of word-initial accents,cf. Cato's Pastores pecuaque salva servassis dUisque duonam salutem "(FatherMars, I bid you, that you) keep herdsmen and herds unharmed and giveguard well" in De Agricultura 141,2).

158 Prot-lnd-European Phonology4) The first step away from the Proto-lndo-European accent system towardthe word-initial accent was necessarily influenced by the Middle-Italian coine,to which belonged, in addition to Latin, also Sabellic and Etruscan. Furtherreading: H. Rix "Die laleinische Synkope als hislorisches und phonologischesProblem" in Kratylos 11 1966 p. 156-165; Leurnann LLFL 1977 p. 235-254; Meiser Laul- u. Formenlehre 1998 p.53.Ill. Proto-Indo-European MorphologyA. General InformationF 100. In traditional granunar, morphology, in the broadest sense, issituated between phonology (see part IT, above) and syntax (see part IV,below).I) A verb form such as Lat. ind. pres. 3 sg. act. gignil 'creates, bringsabout' is composed of a word stem (or, more precisely, of a verbal stem, ormore precisely yet, of the indicative present stem), in this case, gignV- and the3 '" person sg. active ending -I. A substantive form such as the Lat. nom sg.genelrix (" creator [f.]' » 'mother' is based upon the word stem (or, moreprecisely, the nominal stem) genelric- and the nom sg. ending -s (the characterx indicates the double consonance c+s, the character c indicates a voiceless k).The initial, unchanged part of both forms, the word stem, carries the signification,and includes the lexeme. The latter part, the ending, is part of a largerinterchangeable set, or paradigm (cf. in the case of gign V- among others, thepres. ind. act. endings -0 -is -it -imus -ilis -unl, cf. in the case of genetrlcamongothers, the sg. endings -s -is -, -em -e). The endings indicate thegrammatical category of the word form in question, or morpheme.2) The summary in § I outlines the general tasks of morphology. Thebranches etymology, word formation, and semantics are concerned with wordstems, which shall in the present work be treated under the heading 'Vocabulary'in part V.The presently discussed, more narrow sense of morphology (Frenchmorphologie; German, Formenlehre) is concerned with paradigms andtheir endings. Since the structure of a verb combines various tense stems{compare the above present stem gi-gn-V-, and perfect stem 'gena-/}- withthe classical Latin I sg. genul etc.), they shall also be discussed here.

160 Prot-Ind-European Morphology3) Concerning the definition of inflection and morphology: -4 BussmannLexikon der Spraehwissensehaf! 1990 p. 244f. and p. 504f.; LewandowskiLinguisl. Worlerbueh 2 1994 p. 306 and p. 729f.F 101. As a rule, every verbal or nominal word stem may be subdivided.I) A first example: The Greek verb form Ep1]Mamce (Hom, poetic) 'tohinder' contains, along with the ending -0 «' -1; which here indicates 3 sg.Aor.) the elements EPTl-ru+aa+mce-. The nominal stem EPTlTU- serves as thebasis to which both of the suffixes -sa- (which indicates the aorist stem), and-ske- (which indicates an iterative quality) are added. The nominal stem maybe further subdivided into the verbal root EP- « PIE 'yer- 'to hinder,' cfVed. var-Iave 'to be hindered,' var-Ira- 'protective embankment') and thecomplex nominal suffix -1]TU- (which belongs to the group of Greek verbalabstracta with -Iu-; the secondary fonn, with the addition of -e-, and -ii-, iscomparable to the familiar substantives of the type CryoPTlW- [Hom Od.]'speech'; my suggested analysis is supported by the Homeric substantiveEOT]W- 'food,' which is constructed in a parallel manner, and is clearly basedupon the verbal stem EO- 'to eat' < PIE 'h,ed-).2) A second example: The two Latin nominal stems gene-tr-i-e- f.'mother' (older (from before the short vowel weakening in middle syllables: -4Leumann LLFL 1977 p. 82f.; see L 204] 'gena-tr-i-c-) and genilor- m 'futher'(older 'gena-Ior-) are both composed of the full grade verbal root'gena- « PIE 'genh,-) and the suffix -t(JI)r- (which is indicative of the nominaagentis, cf. W 205 § 1). In the case of the masculine, the suffix form isthe lengthened grade -Ior-, in that of the feminine, it is the zero grade -tr- withthe addition of the complex suffiix -i-e- (i.e. -i-k-) in order to clearly mark thefemuune (on -/- < '-ihr, cf. W 204 § I).3) Along with the countless word stems that may be divided into the root(the core that is left after removing all suffixes) and its suffixes, there are alsothose that are solely composed of the root, ct: from the verbal reahn, the rootpresents of the type Ved. as-mi « PIE 'h,es-mi, cf. E 502 § 4) and the rootaorists of the type Gr. EV « PIE '" bhuh,-m), or, from the realm of substantives,the root nouns of the type Lat. v6c- 'voice' « PIE 'uokw- - / *uokwwhichcorresponds to the verbal form PIE '!iekw- 'to- ,say').4) Among word stems, one may distinguish athematic from thematic wordstems. The criterion for this distinction is the presence or absence of the thematicvowel -e- at the end of the stem, just before the ending. The ablaut ofthe thematic vowel is -0-. In the case of substantives, the -0- takes prece-General Information 161dence; however, the -e- ablaut is found in the vocative singular (in masculineand neuter with -e) and in a part of the locative singular and instrumental singularforms (along with -oj and -oh, there are indications of -ej- and -eh,forms), cf. F 311 § I. In the case of verbs, -e- and -0- alternate. The phoneticcontext appears to play a role here. Namely, the -0- vowel is only found beforeendings that begin with -m-, -nl- , -hr, or -ih,-, cf. F 209 and ct: L 418 §4. For more information on the subject in general: -4 Rasmussen Morphophonemik1989 p. 136ff.As a rule, in order to simplifY reconstructions, thematic nominal formsare given an -0- and thematic verbal forms are given -e-. A nominal stemsuch as PIE '!ierg-o- 'work' is thus determined to be every bit as thematicas the verbal stem PIE ·bh"r-e-. The nominal stem PIE '/i(u)!ion- 'dog,'for example or the verbal stem PIE 'h,es- 'to exist' are, in contrast, athematic.In the course of the development of the individual lE languages,the quantity of thematic substantives and verbs has increased, while thequantity of athematic nouns and verbs has decreased. The endings of thematicverbal and nominal stems are somewhat different from those of theathematic forms, cf. F 209 fo r information on the verb, and F 311 for informationon the noun.F 102. Verbal inflection (conjugation) and nominal inflection (declension)shaU be treated separately; the verb, below in F 200ff., the noun, belowin F 300ff. A couple of characteristics, however, are common to bothinflections.I) Paradigms may be defined for conjugation and declension, for examplethe paradigm for the Latin present stems of the type gignV- (namely, in theindicative, present, active, 1 sg. -0, 2 sg. -is, 3 sg. -il, etc.) or the masculinesingular paradigm for Greek -0- stems (nom sg. -0

162 Proto-Indo-European MorphologyFor example, place names require a Iocative; proper noWlS, a vocative; thingsrequire an instnunental, etc.: --+ Risch Kleine Schriften 1981 p. 736; concerningAnatolian, E. Neu in HS 102 1989 p. 13f. ("Much evidence indicates thatthe construction and extension of case inflection for nouns of the personal andobject type did not develop Wlifonnly and simultaneously").For further information on changes in the nominal paradigms, cf. F 302.F 103. As a rule, a single root forms the basis of the entire diversity offorms of a word: Compare, for example, the Latin present stem laud-a-jewiththe slightly varied perfect stem lauda-!!-; as well as the Latin dominusm. 'head of the household, owner' (from the nominal stem dom-ino-),domina f. 'mistress of the houshold, wife' (from the nominal stemdom-ina-), dominare 'to be the head of the houshold' (from the verbalstem 'domin-a-je-), dominatio f. 'rule' (from the norninal stem domina-lion-),and others.In a few old examples, genetically different stems complete each otherthrough suppletion.In contrast to the above-mentioned examples, comparethe Latin present stem esse (sum) 'to be,' which has the perfect stemfUI (a perfect stem 'erul or the like is not attested); or compare Gr. aV1jpm. 'man,' to which the corresponding feminine is yuV1j f. 'woman'('avopia does not exist). For information on the subject of suppletion: --+K. Strunk, "Oberlegungen zu Defektivitiit und Suppletion im Griechischenund lndogermanischen" in Glolta 55 1977 p. 2-34.F 104.Nominal endings give information regarding the paradigmaticcategories of casus, numerus, and genus.Their verbal counterparts giveinformation regarding numerus, person, voice, aspect, mode, and tempus.The description in terms of content of these categories is part of syntax, cf.part IV, with C on the morphosyntax of the verb, and D on the morphosyntaxof the noun.However, mention of content shall be included incases in which it is important for the understanding of the formal side.It is characteristic of the Indo-European languages that there is not aformans that is analytically proper to each grammatical category, butrather that the individual endings combine, synthetically and cumulatively,two, three, or more statements concerning content in the same bundel.For instance, compare, from the verbal realm, the Latin ending -tur (in, fo rexample, laudatur 'he/one is praised'), in which the qualities 3' " person,singular, present, and passive are all expressed. Another example from theverbal realm is the PIE ending '-ent-i (in, for example PIE 'hls-enti 'theyexist,' cf. E 505), which includes the information:3" person, plural, ac-General Information 163tive, and present.From the norninal realm, consider .th example of Lat.-us (in the example dominus) which expresses: nommalIve, smgular, andmasculine.An indication that the analytical process too was known in Pre-ProtoIndo-European is found in the accusative plural *-ns: Assuming .the comparisonof the PIE acc. sg. *-m is not misleading, the accusalIve plural*-ns resulted from the combination oft-m (acc.) and -(e)s (pl.).B. VerbsI. General Info nnationF 200. The verb is the center of the verbal phrase. It offers a muchgreater wealth offorms and contents than the substantive. For informationon syntax, cf. S 300ff.Occasionally the reproach is voiced that reconstruclIons of the PIE verbalsystem are disproportionately based upon knowledge of Greek andIndo-Iranian. Critics of the 'graeco-aryan' reconstruclIon model oftenpoint out the differentness of Anatolian, Tocharian, and Celtic. H ? wever,alternative models could not be offered. In fact, where exact data IS available,it is again and again shown to be the case, that this differentness maybe seen as a secondary variation from the model as it is offered in Greekand Indo-Iranian. Thus, the privileged place of Greek and Indo-Iranianseems not to be attributable to any bias in the research, but rather to thehistory of Greek and Indo-Iranian: --+ H. Rix "Das kellische Verbalsystemauf dem Hintergrund des indo-iranisch-griechischen Rekonstruktionsmodells"in Kolloquium Keltisch Bonn 1976 (1977) p. 132-158.F 201. Literature: - a) Concerning the verb in general: --+ BussmannLexikon d. Sprachw. 1990 p. 828f. s. v. 'Verb'; Metzler Lexikon Sprache1993 p. 675 s. v. 'Verb'; Lewandowski Linguist. Worterbuch 3 1994 p.1221f. s.v. 'Verb'; R. Stempel "Aspekt und Aktionsart, Tempus und Modus:Zur Strukturierung von Verbalsystemen" in IF 1 04 1999 p. 23ff. -b) Concerning the verb in Proto-Indo-European: --+ LIV 1998; BrugmannGrundrifJ 11-3 1916; Krahe ldg. Sprachw. II 1969 p. 50-89; Watkins Vrbaljlexion1969; Jasanoff Stative and Middle 1978; Szemen5nYI Em-.

164Proto-lndo-European MorphologyVerbs 165Jiihrung 1990 p. 244-370 (see Einleitendes iiber das idg. Verbalsystem;Personalendungen, Themavokal, Diathesen; Modusbildungen; Tempusstamme;Synthese: Paradigmen mit Anmerkungen; Verbum infinitum;Vorgeschichte); J. H. Jasanoff "Aspects oJ the Internal History oJ the PIEVerbal System" in Fachtagung ZUrich 1992 [1994] p. 149-168. - c)Concerning the Latin verb: Leumann LLFL 1977 p. 505-624 (see DasVerbalsystem und seine Formen; Proesenssystem mit Praesensstammenund FlexionsJormen; PerJektsystem mit PerJektstammen und FlexionsJormen);Meiser PerJekt 1991; Meiser Laut- und Formenlehre 1998 p. 178-228 (see the sections: Flexion des Verbums; Aufbau des Paradigmas; DieBi/dung des Prasensstamms; Tempora und Modi des Prasensstamms; DerPerJektstamm; Die Endungen; Unregelmajiige Paradigmen; Infinite Verbaiformen).- d) Concerning the Greek verb: Meier-Briigger Gr.Sprachw. IT 1992 p. 46-63; Hauri Futur 1975; Tucker Early Greek Verbs1990; Rijksbaron Verb in Class. Greek 1994. - e) Concerning Vedic andIndo-Iranian verbs: Goto Materialien Nr. 1-29 1990-1997; WerbaVerba IndoArica I 1997; Kellens Verbe avestique 1984 + 1994; Further: Hotlinann AuJsatze I-ill 1975-1992 and Narten Kleine Schriften I 1994with numerous and methodically exemplary treatments of Indo-Iranian andPIE verbal systems; Narten Sigmatische Aoriste 1964; Schaefer Intensivum1994; Kilrnrnel Stativ und Passivaorist 1996. - f) Concerning H;lIiteand Anatolian verbs: Oettinger Verbum 1979. - g) Concerning theGermanic verb: Seebold Germ. starke Verben 1970; BarnmesbergerGerm. Verbalsystem 1986; R. Liihr "Reste der athematischen Konjugationin den germanischen Sprachen" in Kolloquium Germanisch Freiburg 1981[1984] p. 25-90. - h) Concerning Balto-Slavic verbs: Koch Aks!. VerbumI I IT 1990; Petit Lituanien 1999 p. 75ff. - i) Concerning the Celticverb: H. Rix "Das keltische Verbalsystem auJ dem Hintergrund desindo-iranisch-griechischen Rekanstruktionsmodells" in KolloquiumKeltisch Bonn 1976 [1977] p. 132-158. - k) Concerning the Annenianverb: Klingenschrnitt Altarm. Verbum 1982; Lamberterie Armenienclassique 1992 p. 269-276 (see morphologie verbale). - I) Concerningthe Tocharian verb: Pinault Tokharien 1989 p. 123-162 (see morphologieverbale); Hackstein Sigmat. Prasensstammbildungen 1995.2. Verbal Stem FormationF 202. Every PIE verbal form may be split into parts. The first centraldivision separates the verbal ending from the verbal stem that fo llows it, cfF 10 I. - Verbal stems are, according to their form, either athematic (e.g.PIE *h1es-) or thematic (e.g. PIE *bhere-), with final -e- (with ablaut -0-),cf F 101 § 4 and F 203 below, toward the end of the introductory text.The increase in the number of thematic verbs, and the corresponding decreasein the number of athematic verbs may be observed in all IE languages: LIV 1998 p. 12f.I) Each verbal stem without an ending is initially considered a tempusmodusstem (called a secondary stem in H. Rix's terminology). After theelimination of the tempus-modus suffixes (which Rix correspondinglynames secondary suffixes: suffix -(J- for indicative, and in present andaorist, also for imperative and injunctive; suffix -e- for the subjunctive;suffix -jeh1-I-ih1- for the optative) remains only the actual verbal stem(called the primary stem by Rix). Rix's terminology is explained, for examplein Rix Hist. Gramm. d. Gr. 1976 p. 190ff. The common terminologyis non-uniform. It allows itself to be led astray from the realities of thelE languages. Depending on the presence or absence of an aspect systemor a tempus system (the latter being derived from the former), one speakseither of the aspect or the tempus stem.2) The actual verbal stem is either composed of the verbal root (as isthe case of root present and root aorist forms), or is further divisible intothe verbal root and one or more suffixes.The actual verbal stem is in use either as the present stem, the aoriststem, or the perfect stem. The terms present, aorist, and perfect all indicateaspect, which is a grammatical dimension. The aorist stem indicatesthe perfective aspect. The present stem indicates the imperfective aspect.The perfect stem indicates a sort of resultative aspect. For further information,cf. S 304 and S 306-308. The formation types of present, aorist,and perfect stems, which are demonstrable in Proto-Indo-European, arediscussed in F 203, with reference to LIV 1998 p. 14ff.Along with aspect, the lex;cal aspect (i.e. the manner of action, or Aklionsart)also plays a role. The manner of action (or Aklionsart) is a qualityof the verbal meaning which relates both to the process of taking placeof the action that the verb describes, and to the verb's agens or paliens.For more information, cf S 305. The action types, which are demonstrablein Proto-Indo-European are discussed in F 204 below, with referenceto LIV 1998 p. 22ff.3) The present, aorist, or perfect stem forms the basis of the tempusmodusstem, which serves in the expression of the categories of tempusand modus, and is created through the addition of tempus-modus suffixes:

166Prot-Ind-European MorphologyVerbs 167indicativesuffixsubjunctivesufftxathematic-IJ--einalternance with-0-Qptative suffix -jeh,-in ablaut with-ih,-thematic-e-+-IJ- = -einalternance with-O-+-IJ- = -0--e-+-e- = -einalternance with-0-+-0- = -0--o-+-ih,- = -oj-The stem with the suffix -IJ- is autQmatically the indicative stem. In thepresent and aorist systems, the injunctive and the imperative are bothformed from, and attributed to, the indicative stem. With his use of theindicative stem, the speaker indicates that he attributes validity to the contentsof his statement, CL S 310. Stems that are marked with the additionof -e- (in alternance with -0-) indicate the subjunctive; while those featuringthe suffix -jehr (ablaut -ih,-) indicate the optative, cf. F 207. For furtherinformation on modi and considerations of content, cf. S 313.4) To the modus-tempus stem are affixed endings. These endingsserve, With the help Qf accent and ablaut differences in the verbal stem, toexpress the categories of person, numerus, and diathesis.HQwever, the endings contribute more than just this. On the one hand,thanks to. characteristic ending sets, they help to distinguish the present/aorist,imperative, and perfect systems from one another. - Furtherowing to the existence of two sets of endings (primary and secondary), thtwo additIonal tempus types present and preterit/past (nQn-present) arefo rmed. The prunary endings indicate the present tense, the here and now.In contrast, the secondary endings are unmarked in relation to tense. Theymdlcate freedQm frQm temporality, and further, in the case of conciousrefusal of a present form, clearly the past. The choice Qf the two sets ofendings is partially dependent on the desired aspect, and partially dependenton the desired mode. For reasons of aspect, only secondary endingsmay be comḅine with aoris! stems. When combined with present stems,prunary endmgs mdlcate the present tense, and secondary endings the pasttense (or unperfect). While the subjunctive accepts both primary and secondaryendings, the optative, owing to its content, may only accept secondaryendings.5) In order to clearly indicate the past tense, using secondary endings,the fucultative use of a temporal adverb PIE ·h,e 'at that time' couldachieve this emphasis. FQr more information on the augment, CL F 2 I3.In contrast, in order to mention an action without temporal precision (injunctive),the secondary endings may be used, but nQt the temporal adverb,cf. F 213 and S 311.6) Ancient IE languages employ various methods in order to indicatethe future tense. For example, -se- formations in Greek, -s(i)je- formationsin Baltic and Indo-Iranian: -7 Szemerenyi Einfuhrung 1990 p. 307-312. The subjunctive mode of the aorist stem was also a possibility, cf. S306.F 203. Through the presentation of a large amount of data, this sectioncontains an overview of present, aorist, and perfect stem formatiQn. Thereference on this subject is LIV 1998 p. 14ff. with the numbers 1 (a-v) fo rpresent, 2(a-c) for aorist, and 3(a) for perfect. Also worthy of mentiQn isthe list of recQnstructed sterns on pp. 649-661. Also helpful are twoworks by H. Rix: -7 "Einige laleinische Priisensslammbildungen zuSe,-Wurzeln" in GS Kurylowicz I 1995 p. 399-408; "Schwach charakterisierlelaleinische Priisenssliimme zu Se,- Wurzeln mil VolIslufe " in GSSchindler 1999 p. 515-535. Szemerenyi's Einfiihrung 1990 p. 244ff.contains an abundance of suggestions fo r further reading.As I depend upon LIV 1998 for my treatment, and thus follow theprecedent set by Rix, it should be clear that I consider this approach adequateand capable of creating a consensus. Most other treatments andworks on the PIE verb system lack a clear image. My fundamental agreementwith the approach ofLIV 1998, however, does not imply that I recognizeevery analysis featured in LIV. For example: PIE .gWjeh,- 'tolive' with the Latin present VIVO = Ved. ;1vati is traced back to the PIEpresent stem .gWjeh3/·gWih,-u- (= LIV-Typ le) on page 192. In this caseI rather fo llow Rix, in his Termini der Unfreiheil 1994 p. 79, in which PIE.gWih,-!!e-li 'he lives' is directly connected with the (verbal) adjective.gWih,-!!o- 'lively.' The verb in question is presumably nothing other thana very archaic denominative formation with a IJ suffix. The nominal -uostemis used directly as a basis of verbal inflexion, yielding the nomlnaJinflections nom. sg .•gWih3!!o-s, acc. sg .•gWih3!!o-m etc., and the verbalinflections Pras. 3 sg .•gWih3ye-li, 3 pI. .gWih3!!0-nti, etc.Meanwhile a second improved edition of LIV has appeared 200 I,which, unfortunately could not be assimilated here.

168 Prot()-lnd()-European MorphologyIn the case of the present stem formations (nr. I) I shall not commenton all of the types a-v, from LIV, equally. I shall rather limit myself to themost common and most interesting, giving examJ'les (always the indicative3"' person singular, active, and sometimes the 3' person plural) and possiblysuggestions for additional reading. In order to simplifY comparisonwith the Vedic and Old Indian verbal system, I shall also mention the correspondingOld Indian indication of verb class.Among the formations mentioned in F 204 and F 205, the majority aredeverbative (of verbal origin), some however are denominative (of nomina1origin), cf. comments on the types I r and 4a. - Included among the denominativeformations are the archaic *-eh,- denominatives, cf the nominalPIE *ne!!-o- 'new,' from which the verbal formation *ne!!-eh,-ti 'torenew' is derived, cf. Hilt. 3 pI. ne-wa-ab-ba-an-zi i.e. ne!!aM-anr'i. Thebasis for derivation is the collective formation * ne!!-e-h,-, which is simplyconjugated from the aforementioned schema *gWih3!!o-s : *gWih3!!e-ti.Further reading: Rix Modussystem 1986 p. 13; Steinbauer Denominativa1989 p. 85-90; H. C. Melchert "Denominative Verbs in Anatolian " inFS Puhvel 1997 p. 131-138.Concerning the differentiation of thematic and athematic verbs, cf F101 § 3 and 4. The athematic formation of active verbs shows ablaut betweensingular and plural. More particularly, this is most often an e gradesingular with an accent on the root syllable vs. a zero grade plural with anaccent on the ending [cf E 505 with the example *hles-ti vs. *hls-enti],and relatively seldom a case of a singular lengthened grade vs. e gradeplural. The athematic middle voice uses the ablaut of the correspondingplural. In contrast, thematic formations show no difference between singularand plural. The alternation of -e- with -0- is of phonetic origin, cf. F101 § 4.Terminology: concerning the technical terms 'amphidynamic' and 'acrodynamic,'cf. F 315 § 4. Concerning the terms 'root present' and 'rootaorist,' cf. cf F 202 § 2 and F 101 § 3. - The accents ofOreek and Vedicverb forms pose their own problems, cf. F 214.I) The most important present stem formations ( LIV p. 14-20; N.B.:All references to LIV refer to LIV 1998, even when 1998 is not alwaysnamed.).LIV type la - is the amphidynamic, athematic root present (whichcorresponds in Old Indian to the 2"" class). LIV offers 139 sample rootreconstructions (of which, 102 cases are certain). - A prime example isPIE *gwhen_ 'to strike (down)' ( LIV p. 194-196) with the active 3 sg.*gwhen_ti vs. 3 pI. *gwhn-enti, cf. L 345 § I. The middle voice forms thatVerbs 169correspond in terms of formation are treated separately as zero grade rootstatives (type Ic).LIV type I b - is the acrodynamic, athematic root present, also calledthe Narten present. LIV offers 46 sample roots, of which 3 I are certain._ Consider the prime example, PIE * stelj- 'to be/to make manifest, toprize' ( LIV p. 546; concerning its meaning in Greek and Indo-Iranian: Puhvel HED 2 1984 p. 483 -485 s. v. istuwa-) with the active 3 sg.*srJ!!-ti vs. 3 pI. *stelj-TJti. - The middle voice forms that correspond interms of formation are treated separately as e grade root statives (type Id).I shall include them here. A prime example is PIE middle voice 3 sg.*ste!!-o(-D (the -0 ending is archaic and is most often replaced by -to-, cf.*kj-, in which case the Vedic alludes to -0, but the Greek only alludes to-to, or to -taj, which came yet later) 'was/is manifest' = Or. ""

170 Proto-Indo-European Morphologyment from the LIV type I h: - A prime example is PIE *genhr 'to produce'(-7 LIV p. 144-146) with the 3 sg. active form *gi-fJ1Jh,-e-ti = Lat.gigno = Gr. yiYVOllat (for information on the phonetics of which, er. L330). - Further examples of Se\- roots in Latin are offered in: -7 H. Rixin GS Kurylowicz I 1995 p. 406f. - Examples of Se\- roots in Greek maybe found in: -7 Giannakis Reduplicated Presents 1997 p. I 22ff.Ll V type I k - is the athematic nasa1 infix present, which includes theOld Indian classes 5, 6, 8, and 9. LIV gives 237 roots, of which 170 arecertain: - A prime example is PIE *lejkW- 'to leave (behind)' (-7 LlV p.365f.) with the 3 sg. active form *1i-ne-kW-ti vs. 3 pI. *1i-n-kW-enti = Ved.3 sg. rit/ok-ti (-7 Mayrhofer EWAia rr p. 457f.) = Lat. 3 pI. linqu(-unt).- Cr. PIE *pe!!H- 'to clean , to reform' (-7 LlV p. 432) with the 3 sg.active form *pu-ne-H-ti = Ved. pu-n!i-ti (cf. L 315 § 2). - Cr. PIE*demhr 'to tame, to make compliant' (-7 LIV p. 99f.) with the 3 sg. activeform *d".-nti-hrli (cf. L 316). - The type -ne!!-I-nu-, whicb has as abasis roots with final position -U-, is included in the LTV under II (with 51roots, 35 certain), cf. PIE */{je!!- 'to bear' with the 3 sg. active *K!-ne-!!-Ii= Ved. srn6ti (older * Sr-na-!!-ti [with Proto-Vedic s < K, r < ! and a < e].cf L 315 § 2). - For examples of Se\- roots in Latin: -7 H. Rix in GSKury/owicz I 1995 p. 401 -406. - Nasal infix present fo rmations (or partsof It) has been treated several times, sometimes controversially: -7 KuiperNasalprtisentia 1937; Strunk Nasalprtisenlien und Aoriste 1967; K.Strunk "A nhaltspunkte fur urspriingliche Wurzelabstufung bei den indogermanischenNasalprtisentien" in 1nl 5 1979 [1980] p. 85-102; by thesame author, RefJexions sur l'infJxe nasal in CoJloque E. Benveniste IT1983 p. 151-160; McCone Old Irish Nasal Presents 1991; G. Meiser "ZurFunktion des Nasalprtisens im Urindogermanischen" in FS Rix 1993 p.280-313; (concerning Hittite:) S. Luraghi I verbi derivati in -nu e iJ lorovalore causativo in Grammatica ittita 1992 p. 153-180.LTV type In - is the e grade present with the thematic suffix -e-, and isrepresented in the LIV in 425 roots, of which 23 1 are certain. - A primeexample IS PIE *bher- 'to carry, to bring' (-7 LlV p. 61f.) with the 3 sg.active form * bher-e-Ii = Lat. fert « *iereti with syncope: -7 Meiser LautundFormenlehre 1998 p. 224 § 3) = Gr. EPEl (concerning the ending -ej:-7 M. Kiimrnel in PFU 2-3 1996-1997 p. 121f.; see also F 209) = Ved.bhtirali. - Further reading: -7 Goto I. Prtisensklasse 1987; J. H. Jasanoff"The Thematic Conjugation Revisited" in FS Watkins 1998 p. 301-316.LIV type 10 - is a present with a zero grade root and thematic suffix-e-, and corresponds to the Old Indian 6'h class. The LTV cites 47 roots ofwhich 20 are certain. - A prime example PIE *gWerh,- 'to devour' '(Verbs 171L1V p. 189) with the 3 sg. active form *gWrh,-e-1i = Ved. girati (-7 MayrhoferEWAia I p. 469f.).LIV type I p - is a present form with a zero grade root and thestressed suffix -SKe-, and is cited in LIV with 70 roots (50 certain). - Aprime example is PIE *gWem- 'to go (somewhere), to come' (-7 LIV p.187f.) with the 3 sg. active form *gw".-sKe-ti = Horn. Gr. (in its simplex[non-compound] form, found only in the imperative) aoKE 'go!,' in thecompound form 1tOPEO

172 Proto-[ndo-European Morphology2) Aorist stem fo rmation (---7 LIV p. 20f.).LIV type 2a - is an athematic root aoris! which sometimes becamethematic in IE languages. LIV gives 392 roots, of which 265 are certain.- A prime example is PIE 'gWem- 'to come' (---7 LIV p. I 87f.) with the 3sg. active fo rm '(e") gWem-1 vs. 3 pI. '(e") gWm-ent = Ved. ti-gan (with -an< *-an-I < '-am-I) vs. ti-gm-an (For more information on Vedic and IndoIranian: ---7 Mayrhofer EWAia I p. 465f.). - Further reading: ---7Hardarson Wurzelaorisl 1993.LIV type 2b - is an athematic aorist with the suffix -S-, which, for thisreason is also called the sigmatic aorist. LIV gives 174 roots, of which 79are certain. - A prime example is PIE 'prel

174 Proto-lndo-European MorphologyVerbs175and KVN-Hs-, it appears that complex suffix -Hs- must have also found itsway into the non-laryngeal stems with final -1-, -r-, -m-, and -n-.LIV type Sa - is a desiderative form with the suffix -(1I)s-, of whichLIV cites 72 examples (28 certain): - A prime example is PIE *!!ejd- 'tosee, catch sight of ( LIV p. 606-608) with *!!ejd-s- 'to wish to see' =Lat. vlsere 'to visite.'LIV type 5b - is a reduplicated desiderative with the thematic suffix-(H)se-. LIV cites 36 roots, of which 9 are certain: - A prime example isPIE *!!en- 'to overpower, to win' ( LIV p. 622) with *!!i-!!,}-Hse- =Ved. v;viisali 'would like to win.'3) The stem formation of the action type 'intensive' carries the connotationof "repeated bringing about of a state of affairs" ( LIV p. 24f.).LIV type 6a - is a reduplicated athematic inlensivum, of which LIVcites 14 roots (5 certain). - A prime example is PIE *kwer- 'to cut (off),to carve' ( LlV p. 350f.) with *kwerkwr_ = Ved. participle lairi-kr-al'doing again and again.' - Further reading: Schaefer Inlensivum 1994;M. Fritz, "Keine Spuren von Laryngalen im Vedischen: Die Laryngalkiirzungbeim Inlensivum" in FS Narlen 2000 p. 55-61.4) The stem formation of the action type 'fientive indicates "the entry ofthe subject into a new state of being" ( LIV p. 25).LIV type 7a - is a fientive stem with a suffix -eh,-I-hr, of which LIVcites 58 roots (19 certain): - A prime example is PIE *men- 'to hold athought' ( LIV p. ) with *mn-eh,- = Gr. E!lOVllV 'became furious.' -Further reading: LIV 1998 p. 25; J. A. Hardarson in Fachlagung Innsbruck1996 [1998] p. 323ff.5) The stem formation of the action type 'essive' conveys "a subject'sstate of being ... without stressing the entry of the subject into the state ofbeing" ( LIV p. 25).LIV type 8a - is an essive with the thematic suffix -hde-. LlV offers94 roots, of which 44 are certain. As a derivative of -je-: the compoundsuffix -hde- is based upon the fientives with -eh,-I-h,-. - A prime exampleis PIE *Ien- 'to tighten, to twist' ( LIV p. 569) with *tQ-hde- = Lat.lenere 'halten.' - For further reading, see § 4 type 7a.F 205.The majority of the stem fo rms mentioned in F 203 and F 204 are,according to their aspect, present stems. However, if the speaker wants toswitch from the present to the aorist, or the perfect, or from the aorist tothe present, or to the perfect, or from the perfect to the aorist or the pres-ent, he must add to the given present, aorist, or perfect stem, a new 'aspect'stem.Either the speaker uses what is called a regular stem form series of thetype as in the Greek 7tatOeUro 7tatOeOCt v (present stem) 'to raise' vs.£7toioeooo 7tatoeiicrat (aorist stem) vs. 7tE7toioeulCo 7tE7tatOeUICEvat(perfect stem) or, as in rarer cases, he uses a suppletive stem form series ofthe type as in the Greek £PX0!lat £Pxea6at (present stem) 'about to go,to just now be going' vs. A80v EA8e1v (aorist stem) 'to come to someone,to come to the goal' vs. EA1\AU80 EAllAu8evat (perfect stem) 'tohave come,' in which case two or three different stems with similar meaningare combined. (The meanings cited here are Homeric: LfgrE II col.535f. and col. 726f.). The verbal stem form series are not predictable withthe same certainty, with which the verbal and nominal ending paradigmscould be predicted.That, for example, Or. epX0!lat is paired in aspectwith A90v, the speaker must simply know, or deduce from the context.It is possible to describe rather well the stem form series of the individualIE languages. In fact, one may suspect that Proto-Indo-European alreadyfeatured both types of stem form series.For example, on the onehand, PIE *gWem- 'to come' ( LIV p. 187f.) with the present stems*gw,!,-slie- and *gw,!,_je_ vs. the aorist stem gWem_ vs. the perfect stem*gWe_gW6m_, and on the other hand PIE *h,es- 'to be, to exist' with thepresent stem *h,es- vs. the aorist stem *bhuH- (Gr. £

176 Proto-lndo-European Morphology1) In the Pre-PIE period (phase A) there existed only stem fonns for themanner of action (or Aktionsart), including subjWlCtive and optative stems, themeanings of which were the same as later: SubjWlCtive indicating a voluntative/prospectivequality; optative indicating a cupitive/potential quality.2) Moving from the Pre-PIE phase A to the PIE phase B, two decisive innovationsdeveloped: a) The introduction and application of the presentstemlaorist stem aspectual dichotomy (s. § 3); and b) the polar reorientation ofsubjWlCtives and optatives to create fonns that could then modally determineall present and aorist stems (cf F 207).3) The initial impetus to furm this aspectual dichotomy originated with theroot present fonns which, being action-type neutral, according to their meaningcould be durative or pWlCtuaI, for example the durative PIE * h/es- 'toexist' vs. the pWlCtual PIE *gWem_ 'to go (somewhere), to come.' The durativequality permitted both primary and secondary endings, while the pWlCtualquality only allowed the secondary endings. The decisive break in the developmentof the aspect system was the spread of the possibility of forming apWlCtual s-stem from every poSSIble present stem. - Further reading on theorigin of the aspectual dichotomy: ---7 K. Strunk "Relative Chronology andIndo-European Verb-System: The Case of Present- and Aorist-Stems " in JIES22 1994 p. 417-434. Note the synopsis on p. 417:"Some evidence will be discussed in favour of the preliminary conclusionthat 'Aktionsarten' expressed by different types of present-sterns [occurringboth in Hittite and other lE languages) already existed in early PIE,whereas the category of aspects expressed by contrastive present- andaorist-sterns [lacking in Hittite) did not develop before a later period of thePIE verb-system."4) The aspect system of most early IE languages was largely replaced by atense system which is hased upon time relationships: ---7 R. Stempe "Zur Vorgeschichteund EntwickJung des lateinischen Tempus- und Modussystems" inHS I11 1998 p. 270-285; E. Tichy "Vom indogermanischen TempuslAspektSystem zum vedischen Zeitstufensystem" in Kolloquium Delbriick Madrid1994 [1997) p. 589-609; H. C. Melcbert "Traces of a PIE Aspectual Contrastin Anatolian?" in 1nL 20 1997 p. 83-92.F 207. Concerning the reorientation of the subjunctive and optative actiontype sterns to tempus/modus sterns (cf F 206):I) The subjuctive suffix is PIE *-e-. In the case of athematic verbalstems, the rule is -K+o- (indicative stem), -K+e- (subjunctive stem); corre-Verbs 177spondingly, that of thematic verbs is -e+o- (indicative stem), -e+e- (subjunctivestem) cf F 202 § 3.The formal identity of the athematic subjunctive stem (e.g. PIE"h/es-e-) to the thematic indicative stem (e.g. the type PIE *bher-e- fromF 203 § 1 n) is no coincidence. This identity may be understood if we supposethat the subjunctive with -e- was first an action type. The voluntative/prospectivemeaning was neutralized when the primary endings, whichemphasized the present tense, and thus the immediacy of the action type,were used and could give the impetus for the formation of indicative -esterns.At the same time, the -e- stem voluntative/prospectives provedvery lasting and established themselves, together with the optatives, as amode which could be attached to every stem, lastly even the indicative -estems.For details on the present description: ---7 Rix Modussystem 1986 p. 14f.with note 20. - Further reading on related problems: ---7 E. Risch "ZumProblem der thematischen Konjugation" in Risch Kleine Schriften 1981(in a contribution from 1965) p. 702-709; B. Barschel "Zu oepKO}Lat undeinigen anderen thematischen Wurzelprasentien des Griechischen" inBeitrage zur historischen und vergleichenden Sprachwissenschajt Jena1990 p. 4-8 (which discusses the Greek form as a subjunctive aorist); K.Strunk "Zur diachronischen Morphosynlax des Konjunktivs" in Kol/oquiumKahner Amsterdam 1986 [1988) p. 291-312.2) One may presume an origin for the optative that is similar to that ofthe subjunctive. - In terms of form, the suffix in athematic formations isPIE *-jell/" in the full grade, and *-ih/- in the zero grade. - A PIEo-o-ih/- or o_O_PI/_ must be the starting point for thematic stems. Accordingto K. Hoffrnann (---7 Aufsatze II 1976 p. 615 note 12) a 3 sg.*-o-ih/"t leads, in the case of laryngeal disappearance, via *-o-,t > *-oiland *-ojt, a I sg. *-o-ih/-'!' via o-oio > *-oija > *-ojja. In contrast, H. Rix(---7 Hist. Gramm. d. Gr. 1976 p. 233) uses *-o-jh/- as a basis. J shareHoffrnann's position in accepting O_o_ill/- and postulate that, as in the caseof Gr. 8eil'£v < *the-,-men < °dhell/-ih/-me- (---7 Risch Kleine Schriften1981 [in an essay of 1975) p. 193) there was originally a morpheme divisionbetween -0- and -ih/-. - Further reading: ---7 1. H. Jasanoff"The Ablautof the Root Aorist Optative in Proto-Indo-European" in MSS 52 1991p. 101-122; H. Eichner in Bopp-Symposium 1992 [1994) p. 80ff. (wherehe considers, in light of F. Bopp, whether the optative should be understoodas a periphrastic formation, cf. Lat. edim < Pre-PIE *h/ed+jeh/-m 'Iask for food' in the sense of 'J would like to eat').

178 Proto-Indo-European MorphologyVerbs1793) Anatolian and Hittite feature neither optative nor subjunctive modes.It appears that Anatolian left the community of rE languages before thereorientation of the action types voluntative/prospective and cupitive/potentialto the modes subjunctive and optative, respectively. Whileno traces of the action type cupitive/potential may be found, perhaps theredo exist traces of the voluntative/prospective action type ( RixModussyslem 1986 p. 20f.).Further reading on the subject: K. Strunk "Probleme der Sprachrekonslruktionund das Fehlen zweier Modi im Helhilischen" in 1nl 9 1984[1985] p. 135-153; J. A. Hardarson "Der Verlusl der ModuskalegorieOplaliv" in HS 107 1994 p. 31 f. (It is argued that the loss of the optativedid not take place at the stage of action type, but rather after it had alreadybecome a mode.)3. The Verb EndingF 208. The following descriptions are intentionally kept quite short. Theestablished facts concerning Proto-Indo-European are named whereas fordetails concerning the individual lE languages the menti;n of furthersources shall have to suffice.Sources concerning endings in general: E. Neu "Zum Verhdllnis dergrammalischen Kalegorien Person und Modus im Indogermanischen" inFS Poloml! 1988 p. 461-473; Szemerenyi Einfiihrung 1990 p. 247ff.Concerning the differentiation of the primary and secondary endings, cfE 502 § I1 and F 202 § 4. - 'Dual' fonns are not discussed here ' cf F304 § I.F209. The active endings of the presentlaorist system:secondary endingsprimary endin g sa) at hem. b) them. a) athem. b) them.I sg. -m -o-m -m-; -0 or -o-Jzz2 sg. -s -e-s -s-; -e-s-i3 sg. -I -e-I -I-i -e-I-iI pi. -me -o-me -mes / -mos -o-me-2 pI. -le -e-Ie -le -e-le3 pI. -fJl / -enl -o-nl -fJl-i / -enl-i -o-nt-iComment: - The endings shown above are divided into athematic andthematic endings. The endings a) and b) are the same, with the exceptionof the I sg. thematic primary ending. Otherwise, the only difference is thepresence or absence of the thematic vowel. - The thematic vowel: - Inthe case of indicative stems and the subjunctive stems (formed from athematicstems) that are formally and genetically (cf F 207 § I) identical tothem, the thematic vowel -e- is exchanged with -0- (preceding -m-, -nl-,and possibly -hz; -0- is also used before -ih,- in the optative), see F 101 §4. - Endings: - In the case of the I sg., the secondary ending is "-o-m(as determined by the athematic -m), however, the primary ending is surprisinglynot "-o-m-i, but rather *-0 (or possibly "-o-hz: Rix Hisl.Oramm. d. Or. 1976 p. 250; K. Strunk in Kolloquium Kiihner Amsterdam1986 [1988] p. 304f.). - In the 1 and 2 pI., the lE language attestationsare not clear enough to clarilY the differentiation of the primary and secondaryendings in tenns of form (I pI. probably -mes, secondary -me, or-men?).Further reading: - a) General LE: Rix His/. Oramm. d. Gr. 1976 p.239ff. (endings) and p. 206 (thematic vowel); Szemerenyi Einfiihrung1990 p. 247-252 (endings) and p. 266-268 (thematic vowel); Meiser LaulundFormenlehre 1998 p. 40f.; M. Kiimmel in PFU 2-3 1996-1997 p.120-122. - b) Particularly Latin: Leurnann LLFL 1977 p. 512ff.;Meiser Laul- und Formenlehre 1998 p. 216f - c) Particularly Greek:Meier-Briigger Or. Sprachw. II 1992 p. 53f; E. Risch "Ein Problem desgriechischen Verbalparadigmas: Die verschiedenen Formen der 3. PersonPluraf' in FS Neumann 1982 p. 321-334; F. Kortlandt "The GreekIhird person plural endings " in MSS 49 1988 p. 63-69. Concerning thethematic present endings 2 sg. -£t, 3 sg. -Et: H. M. Hoenigswald in FSHamp I 1997 p. 93ff.; see also M. Kiimmel in F 203 § 1 (LJV type In). -d) Particularly Vedic and lndo-Aryan: Hoffinann / Forssman Avesl. LaulundFlexionslehre 1996 p. 179f. (offering a good tabular overview of theactive endings) and p. 190ff. (active paradigm). - e) Particularly Celtic:F. Kortlandt "A bsolule and Conjuncl Again " in MSS 1994 p. 61-68.F 210.The middle voice endings of the presentlaorist system:I sg.2 sg.3 sg.I pI.seconda) b) a)-Jzze-so and -lhze--10 and -0-medhhz-maj and -Jzze-j-so-1-10-1 -o-j-mesdhhz

1802 pI.3 pi.I-(S)dhyeProto-Indo-European Morphology-nto-jComment: - The PIE middle voice endings (or secondarily passive)are those m column a). They are formed presumably in the Pre-PlE erafrom the basis of the active voice endings, cf, in the case of the secondaryendmgs 2 sg. -s-o, 3 sg. -1-0 and 3 pi. -nl-o. In several ancient lE languages.(e.g. Italic, Celtic, Hittite, Tocharian, and Phrygian), the middle/paSSIveendmgs are marked with a special -rei). - The endings in columnb) (as well as the -r- forms) show forms that are identical to perfectforms, cf. F 211. They are usually said to be the output of the 'stativevoice'. On the whole problem see now 1. H. Jasanoff Hiltite and Ihe lEverb Oxford 2003.Further reading: - a) General: Rix Hisl. Gramm. d. Gr. 1976 p. 246-249; Jasanoff Slallve and Middle 1978; H. Rix "The Proto-Indo-EuropeanMiddle: Content, Forms and Origin" in MSS 49 1988 p. 101-119 (proposesthat one see, m the -0 ending, an anaphoric pronoun); SzemerenyiEmfuhrung 1990 p. 257-259; J. H. Jasanoff in Fachlagung Zurich 1992[1994) p. 152ff. - b) Particularly Latin and Italic: Leumarm LLFL 1977p.515ff. (concerning the -r- forms); H. Rix "Zur Enlslehung des laleinischenPerJektparadigma " in Kol/oquium Lal. u. Idg. Salzburg 1986[1992] t 221-240; MelSer Laul- und Formenlehre 1998 p. 218f.; G.Melser Die sabelbschen Medialendungen der 3. Person" in FachlagungLeiden 1987 [1992) p. 291-305; J. H. Jasanoff "An Ilalic-Celtic Isogloss:The .3 pi. MedlOpassive .in *-nlro " in FS Hamp I 1997 p. 146-161. _ c)PartIcularly Greek: Meler-Briigger Gr. Sprachw. II 1992 p. 54. In thec of the verb lCEll1Ut 'to lie,' the replacement of the older middle voiceendmgs -mai -Ioi (as is still the case in Cypriot: Egetmeyer W6rlerbuch1992 p. 66) y the newer (Homeric) -mai -Iai directly attested. - d) ParttcularlyVedlc and Indo-Iranian: Hoffinann / Forssman Avesl. Laul- undFlexlOnslehre 1996 p. 180f. (with a tabular overview of the middle voiceendmgs) and p. 194ff. (middle voice paradigm). - e) Hittite: Neu MedlOpaSS/v 1968 and Neu Interprel . Mediopassiv 1968; Yoshida Endings in. .-n 1990, G.-J. Pmault (m his reVIew of Yoshida) in BSL 86 / 2 1991134-/41.F 211. Perfect endings:I sg. -ille2 sg. -lh1e3 sg. -ep.Verbs1 pI. -me2 pI.3 pI. -rComment: - The perfect endings represent a system of their own. Thedifferentiations between athematic and thematic forms, primary ending andsecondary endings, and active and middle voices are missing. However,the middle voice of the presentlaorist system shows individual sub-formsthat are identical to perfect forms, cf. F 210 column b. - The Hittite -biconjugationposes problems of its own, but is part of the whole business.An adequat insight is now given by J. H. Jasanoff Hit/ile and Ihe lE verbOxford 2003.Bearing in mind that a communis opinio is not on the horizon, followingare some further sources of information on the perfect and the stative:- a) General LE: Rix Hisl. Gramm. d. Gr. 1976 p. 255-257 (concerningthe perfect); Szemerenyi EinJuhrung 1990 p. 259f. (concerning the perfect);LlV 1998 p. 22 (concerning the stative). - b) Particularly on theLatin perfect: Leumann LLFL 1977 p. 606ff.; Meiser Laul- und Formenlehre1998 p. 217f. - c) Particularly on the Greek perfect: MeierBrilgger Gr. Sprachw. II 1992 p. 54f.. - d) Particularly on Vedic andIndo-Iranian: Hoffinann / Forssman Avesl. Laul- und Flexionslehre1996 p. 179f. and p. 236ff.; T. Got6 "Oberlegungen zum urindogermanischen'Stativ"' in Kol/oquium DelbriJck Madrid 1994 [1997) p. 165-192.F 212. The imperative: The basis upon which the PIE imperative isformed is the bare verbal stern, used in the 2 sg. For example, PIE thematic*bher-e 'carry,' but athematic PIE *h/s-dhi 'sei,' with a particle thatcharacterizes the imperative. The thematic form 3 sg. *bher-e-Iod 'heshould carry' is also Proto-Indo-European. The expansion of furtherforms was Post-Proto-Indo-European. For information on syntax, cf. S312.Further reading: B. Forssman "Der Imperaliv im urindogermanischenVerbalsyslem" in Fachlagung Berlin 1983 [1985] p. 181-197; E.Neu "Betrachlungen zum indogermanischen Imperaliv" in FS Schmeja1998 p. 119-127. Concerning details: K. Strunk in FS Dihle 1993 p.486-472 (Gr. 9£, i!. and 06); B. Forssman p. 185 note 12 (A referenceto Th. Benfey, who was the first to suggest that the Vedic imperative typene$i 'lead' could be traced back to the 2 sg. subjunctive aorist ne$asi).181

182 Prolo-Indo-European Morphology4. The AugmentF 213. The augment *(hl)" is common to Greek, Phrygian, Armenian,and Indo-Aryan as an indication of the past tense, e.g. PIE *(hl)" bherel =Gr. e-.pepe = Ved. a-bharal. The PIE augment *(hl)" was quite probablyan adverb with the meaning 'at that time' and could be employed facultativelywhere indicative fonus of present and aorist stems were combinedwith secondary endings to produce a clear past tense, cf. F 202 § 5. Theestablishment of the augment as a norm in the indicative aorist, indicativeimperfect, and indicative pluperfect took place in a post-Proto-IndoEuropean phase. Other IE languages such as Latin or Germanic developedtheir own suffixal means of indicating past tense fonus, e.g. Lat.pres. 3 sg. eS-1 'is,' but imperfect 3 sg. er-a-I « *es-ii-I) 'was': MeiserLaul- und Formenlehre 1998 p. 197.K. Brugmann gave the impulse to call the Vedic indicative fonus thatdo not feature an augment 'injunctive.' Vedic features the injunctive as amode of its own, with a 'memorative' semantic connotation.Suggestions for further reading: K. Strunk "Der Ursprung des verbalenAugmenles - Ein Problem Franz Bopps aus heuliger Sichf ' inBopp-Symposium 1992 [1994] p. 270-284. - Particularly concerningVedic and its memorative: Hoffinann lnjunktiv 1967. - ConcerningGreek, in which the augment was facultative in poetic language and Mycenaean,but in the Classical language was, as a rule, non-facultative: I.Hajnal in MSS 51 1990 p. 50-55; Meier-Briigger Gr. Sprachw. II 1992 p.50-52; E. J. Bakker "Pointing to the Past: Verbal Augment and TemporalDeixis in Homer" in Euphrosyne, FS Dimilris N Maronitis, Stuttgart1999, p. 50-65. On the New Greek situation see at the end of the paragraph.- In Armenian, the augment is only found as a functionless elementin single-syllable verb fonus that permits phonetic reinforcement, e.g.aor. ind. act. 1 sg. beri '1 carried,' but 3 sg. e-ber: J. Wackemagel inWackernagel Kleine Schriften I 1969 (in a contribution from 1906) p.148-155 (Wackemagel refers to the phenomenon of the Homeric eoxov'(they) had,' in which *oxov is missing). - Further, compare the characteristicsof New Greek verbs of the type of3 sg. Eoeo" 'he has bound' vs.1 pI. MOOIl£ 'we have bound,' in which the augment is preserved when itwas stressed.Verbs 1835. The Verbal AccentF 214. The finite verb of an PIE main clause was normally placed followingthe subject and the object, at the end of the sentence, where thesentence accent usually decreases. However, when the verb was stressedat the beginning of the sentence, or in a subordinate clause, it carried itsnormal accent.1) Researchers agree that Vedic generally reflects the fundamental characteristicsof Pro to-Indo-European, and thus, that the finite verb in a main clausewas unstressed: Wackemagell Debrunner Ai. Grammalik I 1957 p. 290;Klein Verbal A ccenluation 1992 p. 90; Hettrich Hypotaxe 1988 p. 779.2) It may also be sbown that this assertion about Proto-lndo-European alsoat one time held for Greek. The characteristics of Greek may best be understoodif we start from the premise of an unstressed finite verb as an establishednorm. Finite verbal forms also featured normal word accents that were used incase the word was in a position of emphasis. - The introduction of the typicallyGreek restriction of all accents to the last three syllables brought a significantchange, forcing all unstressed finite verb forms, which formed accentunity with the preceding word and had generally three or more syllables, to beaccented according to the schema ["- x u], and [x,,- -], cf L 420 § 2. The newaccent was often used as a word accent, and the old (which often differed),often dropped. This explains, among other things, why a verb form such asIlalvoll£v 'we go' carries its accent on the first syllable, although its origin ofPIE *gwTfI-j6-mes would suggest rather *llatvOIl£V. Only the two-syllableroot presents ellll and $rIlll escaped the effects of the radical new accentuation.Not least owing to their shortness, they remained enclitical. For furtherdetails: Meier-Briigger Gr. Sprachw. II 1992 p. 48-50.3) It remains disputed whether the second position of the finite verb, common to the modem Germanic languages such as German, originated from theinherited phenomenon of enclitics, or whether it appeared secondarily: J.Wackernagel in Wackernagel Kleine Schriflen 1 1969 (in an essay of 1892) p.427 ("the german rule of word order was already valid in the mother language");Th. Eyth6rsson "Zur HislOrisch-vergleichenden Synlax des Verbumsim Germanischen" in Fachlagung Innsbnlck 1996 [1998) p. 407 note 16:('The hypothesis [is] ... not supported by findings'').

184 Proto-lndo-European MorphologyVerbs1856. Infinitive Verb FormsF 215. Alongside the finite, or conjugated verb forms, which, thanks totheir endings are marked (grammarians of antiquity speak of 'limited'forms), and specifY qualities among the verbal categories of aspect-actiontype, tempus-modus, person-number, and voice, there are also infinite verbforms that are not conjugated and, in comparison with the finite forms, arenot as strongly marked (grammarians of antiquity speak of 'unlimited'forms). In the case of infinite verb forms, for example, no information isgiven with regard to person. Among the infinite verb forms are the infinitivesand the participles.F 216. Infinitives are verbal nominal forms, which are based on generalizedcase forms of verbal abstract nouns. It is doubtful that Proto-IndoEuropean featured a specific infinitive suffix. The development of meansof differentiation of voice, aspect, and tempus in the infinitive formations ispost-Proto-Indo-European. For information on syntax, cf. S 202.Further sources of information on the infinitive: - a) Proto-IndoEuropean: H. Rix "Die umbrischen Infinitive au! -fi und die UrindogermanischeInfinitivendung -dhjiir in FS Palmer 1976 p. 319-331; DisterheftInfinitive 1977; Gippert Infinitive 1978; H. Rix in FS Szemminyi*65 II 1979 p. 736ff. (containing a list of PIE verbal abstract nouns); J.-L.Garcia-Ramon "Infinitive im Indogermanischen? Zur Typologie der Infinitivbildungenund zu ihrer Entwicklung in den iilteren indogermanischenSprachen" in 1nL 20 1997 p. 45-69; . - b) Particularly concerningLatin: Leumann LLFL 1977 p. 580-582; Risch Gerund. 1984 p. 26f.(for a general characterization of the Latin infinitive); Meiser LauI- undFormenlehre 1998 p. 225; for information on Umbrian, cf. Rix. - c) Particularlyconcerning Greek: Meier-Briigger Gr. Sprachw. Il 1992 p.60f.; K. Stiiber in MSS 60 2000 p. 138f. (containing information on -Etv

186 Proto-Indo-European MorphologyVerbs187laudiiv- back to periphrastic expressions of the active perfect participleand substantivized verb esse).3) The PIE suffix of the middle voice participle is *-m.Xno- (athematic)and *-o-m.Xno- (thematic), where X = Vor H, or even RV. The exactform is uncertain. G. Klingenschmitt makes a case for PIE *-l1Ih,no(Fachtagung Regensburg 1973 [1975] p. 159-163 f.). The field is notunited regarding this assertion: H. C. Melchert supports -l1In-o- (inSwache 29 1983 p. 24f.); Mayrhofer finds the suggestion significant, butpomts out that even proponents of the triple-laryngeal theory share doubtsabout it (in Lautlehre 1986 p. 130f.); Szemerenyi remains with -l1In-o(Einfiihrung 1990 p. 349f.); further, see B. Forssman in Kratylos 45 2000p. 69f., where be comments on the treatment of suffixes from LindemanLaryngeal Theory 1997. As M. Fritz indicated to me, the two competingzero grade suffix variants *-l1Ih,no- and *-mno- may be combined via thefull grade *-mh,eno-, provided the laryngeal in the suffix disappears wbenthe suffix IS added to a root or stem with a non-syllabic final position precedingthe full vowel e. The non-laryngeal full grade form *-meno- wouldthen have the newly constructed zero grade form *-l1Ino-. - Examplesmclude, Myc. Gr. nom. sg. f. ki-ti-me-na i.e. kti-men-ii 'cultivated, builtupon(and thus inhabited)' and Ved. /q1Jv-iin-ti- to kar- 'to do, to make tobring about' with the present stem !crnctlj- / !crnu- (one must admit the Ieg ce of Klingenschmitt's explanation -iin- < *-aHn- < *-1[Ih,n-; the alternative.Wlth -mn- has its difficulties, cf. Szemerenyi: "the origin of thisformation IS not certain"). - The differentiation of the perfect *-mh,n-ovs.the present * '-o-mh/llo- in the various IE languages may t>e' tracedback to the athematic/thematic dichotomy: Rix Hist. Graml1l. d. Gr.1976 p. 236.4) The verbal adjectives *-to- and *-no- function as past participles inindividual IE languages, cf. W 203.7. Periphrastic ConstructionsF 218. Relative to the ancient IE languages, periphrastic constructions ofthe type Lat. quid futurum est 'what should that become' or quod habeotol/ere 'what I intend to take' are considered new. However such formsare attested in the Hittite of the 2"d millenium RC., e.g. the iJark- construCtionsfor the perfect and pluperfect. If in fact the Latin perfect of thetype portiivi may be traced to the periphrase *portii'1osis esom (i.e. anactive perfect participle with -'1OS- + verbum substantivum [cp. F 217 §2]), then also it must date from prehistoric period. Thus, it may not beruled out that Proto-Indo-European already featured several periphrasticconstructions.I also consider cases such as the following to be similar to paraphrases:Lat. vendere < venum *dide- 'to put up for sale,' in the sense of 'to sell'vs. venire < venum ire 'to go for sale' in tbe sense of 'to be sold' (didemusthere be traced to PIE *dheh,- and not to *deh,-!: Meiser Laut- u.Formenlehre 1998 p. 192). Or, similarly, interficere 'to separate (fromlife), to make disappear' in the sense of ' to kill' vs. interire 'to go and disappear'in the sense of 'to decline' (For details relative to inter: H.Hettrich in MSS 54 1993 p. 169-172; -facere makes clear that this is acase of PIE *dheh,-). This combination of substantivized verb or preverband *dheh,- (in tbe active sense), or *h,ei- 'to go' (in tbe passive sense)certainly dates from a pre-individual language period. For a further example,cf. F 207 § 2 with H. Eichner's reflections on the optative.Further sources: Rosen Periphrase 1992; Boley Hittite harkconstruction1984; Cotticelli-Kurras Hill. 'seill' 1991; by the same authors"The Hittite periphrastic constructions" in Grammatica illita 1992 p. 33-59; S. Luraghi "1 verbi ausiliari in ittita" in FS Ramat 1998 p. 299-322.C. Nouns and Adjectives1. General InformationF 300. Nouns as nominal members expand, complete, and clarify thecontent of the central verb form of the sentence. In contrast to verbs,nouns have a relatively limited number of forms. Through the basicmeaning of the word, which is contained in its nominal stem, the endingscommunicate information about the number and genre of the content, aswell as information about the role of the noun, which is assigned to it bythe speaker in the syntactical context, cf. S 400ff.F 301. Substantives and adjectives are closely related. This closeness isdemonstrated by the fact that an adjective can take the place of a substantive,e.g. Classical Gr. O KQKO" 'the bad' and Lat. IUna 'moon' « Proto-

188 Proto-Indo-European MorphologyItal. *lolJk-snii 'the shining'). Substantives and adjectives are exchangeablein norninal sentences.Adjectives often qualifY substantives and, in order to externally markassociation, often use congruence in genre, number, and case. The dimensionof genre is particularly variable. Unlike the noun, the adjective has nofixed association with a particular genre, cf. S 400. Reading: -7 J. Untermann"L 'aggellivo, Forma e Junzione" in Quademi Patavini di Linguistica7 1988 p. 3-21.Formal peculiarities of the adjective are discussed in F 323ff.F 302. The peculiarities of nominal paradigms in lE languages may ingeneral be traced to Proto-Indo-European. However, every family ofIndo-European languages and its individual languages has undergone inthe course of time a greater or lesser quantity of changes that are characteristicof itself. In the area of gender, a history may be written of coexistenceand replacements among the maseulina, ferninina, and neutra, cf. F303. In the area of number, the rise and fall of the dual must be noted, cf.F 304. In that of case, individual forms are changed over time or evenfully disposed of, cf. F 305.Reading: E. Risch "Betrachtungen zur indogermanischen Nominaljlexion"in Kleine Schrifien 1981 p. 730-738; by the same author, "Diemykenische Nominaljlexion als Problem der indogermanischen und griechischenSprachwissenschajt" in Sprache 32 1986 p. 63-77.F 303.Gender developments:I) Ancient lE languages generally feature three genders: masculine, feminine,and neuter, cf. S 416. On the subject of the term 'neuter,' cf. p. XVII.Although the three-part system is well attested, one should not be misled intoregarding it as a fixed triad. For example, in the Romance languages and inLithuanian, the neuter gender was abandoned, being replaced by a new masculine/fumininedichotomy. Today's English makes do without any genders. -For information on gender in the transformation from Latin to the Romancelanguages: -7 Scoon Neutrum und Kollektivum 1971. - For information onEnglish: Leisi Slreijlichter 1995 p. 107-111 (How English rejected sexism).2) Anatolian features a two-part system, separating 'animate' (commongender) from 'inanimate' (neuter gender). Whether a feminine gender neverreally developed, or developed and vanished without leaving clear traces, iscurrently discussed by researchers. If the former theory is correct (which evidenceseems to suggest), it would indicate that the speakers of Proto-Nouns and Adjectives 189Analolian left the PIE language community before the three-part mascu!inelfumininelneuter system had been established. A consequence of thiswould be that a two-part system had existed in Proto-lndo-European muchlike that preserved in Anatolian. This two-part system was composed, on theone hand, of a class A, in which a nominative and accusative were differentiated',on the other hand, a class B, in which this differentiation was ruled out ' ,cf. S 416. - See also L 211 § 6 (on Lat. virus n.) and F 403 (on PIE *kwi_with no distinction between masculine and feminine)3) The impulse which led to the change from a two class system to a threegendersystem occurred with the inclusion of natural gender in class A. Thefirst step consisted, in all likelihood, of the naming of individual natural feminina(in fumilies and among the animals) clearly as such. The first such developmentsmay well have taken place in Proto-Indo-European. However, onlyafter the migratory departure of the Proto-Anatolians from the community ofPIE language speakers did the change pick up momentum. By the end of thisdevelopment, all nouns of the class A had been firmly assigned a masculine orfeminine gender that was often unrelated to natural sexus.4) How did the speakers of the ancient IE languages proceed concretely inthe formal marking of the feminine gender? - A first possibility (called heteronymy;concerning this and the following information: -7 WackemagelVorlesungen IT 1928 p. 9-11) is shown by examples such as PIE * ph,ter- 'father'vs. PIE *meh,ter- (or *mtiter-, cf. L 211 § 10) 'mother,' in which thegender differentiation is performed by inserting two different lexemes. - Concerningthe second possibility, that of the communia of the type Lat. lupusjemina, see E 506 § 5. - The third and most often used possibility is the alterationor extension of the word ending. The suffixes used for this were*-hr, as well as the complex suffixes *-e-hr and *-i-hr, the latter of whichwas as a rule expanded to -i-k- in Latin, e.g. Lat. *gena-tor- 'creator, i.e. futher'vs. *gena-tr-i-k- 'female creator, i.e. mother,' cf. F 101 § 2. This *-hrmay not be separated from the -hr forms of class B (indicative of collectiva),cf F 313. In Proto-lndo-European, the -hr suffix supposedly had the functionof forming abstract/collective derived forms: From there, one lineage leads tothe PIE collective nouns of B, which function in the case of neuter furms as This topic is discussed by researchers under various headings which indicate thatthe objects in question are conceivable to the speaker as operators in a verbal discourse:commune, genus animo/urn, genus distinctum. personal class..5 Further headings for this group of objects which are conceivable to the speaker asnot operating in a verbal discourse: neuter, genus inanimo/urn, object class.

190 Proto-Indo-European Morphologyplurals of the type Lat. iuga vs. sg. iugum 'yoke'; the other lineage leads to thePIE feminine fonns of A (which include, along with the true inflected femininefonns, abstract nouns of the type Lat. fuga 'escape,' iiislitia 'justice,' etc.).Pronominal fonns such as masc./fum sg. PIE 'se-hr 'this' (class A) vs. n. PIE'Ie-flr 'this' (class B) must have played a central role in these developmentprocesses.5) Roughly stated, the masculine and feminine nouns of ancient IE languagesare thus the continuation of lE class A, while the neuter forms are thecontinuation of class B. While neuter forms reveal more or less the contentsof the ancient PIE class B, the question of sexus, including the formation offeminine forms on the basis of the suffix -hr (which comes from class B) obscures the content of the ancient class A considerably.6) Bearing in mind that a communis opinio will not be reached in the immediatefuture, the following are sources of further information on variousrelated subjects: - a) General information: -7 K. Strunk "Grammalischesund naliirliches Geschlechl in sprachwissenschafilicher Sichl" in Frau undMann, GeschlechlerdijJerenzierung in Nalur und Menschenwell, edited by V.Schubert, St. 0ttiIien 1994, p. 141-164; Leisi Slreijlichler 1995 p. 112-116("The descriplion and naming of women as a linguistic problem ") . - b)Proto-lndo-European: -7 E. Tichy KoUektiva, "Genusfemininum und relativeChronologie" in HS 106 1993 p. 1-19; M. Fritz "Die urindogermanischen sSltimme und die Genese des drillen Genus" in Fachlagung Innsbruck 1996[1998] p. 255-264. - c) Anatolian: -7 E. Neu "Zum Aller der personijizierenden-anl-Bi/dung des Hethilischen, Ein Beilrag zur Geschichle der indogermanischenGenuskalegorie" in HS 102 1989 p. 1-15; J. A. Hardarson"Der Ver/us{ des Genusfemininum" in HS 107 1994 p. 32-35.F 304.Number developments:I) While singular and plural are relatively fixed values, the dual has provento be unstable. Generally speaking, the rise and decline of the dual may bedirectly investigated in individual IE languages, for example in Greek, in whichthe dual is a fixed component of the language, while it is missing altogether inIonic and Lesbian. Attic inscriptions attest its presence in Attic through the 4 thcentury B.c.: -7 Meier-Brtigger Gr. Spraclrw. I 1992 p. 144f.The dual shall not be further discussed in the present work. As yet, thefield lacks a monography on the subject. Matthias Fritz is in the process offilling this gap with his "Unlersuchungen zum indogermanischen Dual,Vom Werden und Schwinden einer grammalischen Kalegorie," on thesubject of which he added:Nouns and Adjectives 191"The externally reconstructible Proto-Indo-European numerus category'dual,' which includes all grarnatically inflectable forms and is thusfirmly anchored in the linguistic system, developed in the course ofproto-Indo-European history first to a systematic numerus category.The origins of the dual are contained in two word types: On the onehand, the personal pronoun is a starting point of the numerus dual; onthe other, among nouns, terms for paired body parts are of great importance.While pronouns in the first and second person feature thedual as grammatical category as far back as they can be traced, the dualcategory initially does not exist among substantives. In the case of theterms for paired body parts the duality is lexically founded. To theseterms for body parts was added a particular suffix, that probably did notinitially have the meaning 'pair,' which of course was already providedin the terms for paired body parts. Instead, this meaning would performa deictic function, emphasizing the inalienability of the body parts.While inalienability is a property of all body parts, since most bodyparts are present in pairs, a reinterpretation of the suffix from an indicationof inalienability to one of a paired quality could take place. Thesymbol PIE '-i is a stem suffix for terms for body parts. Since the dualmeaning is also contained in the dual forms of personal pronouns,which are indicated by PIE '-hl, this clear indicator is transferred tobody parts, where, combined with the existing suffix, it continues toserve as a stem suffix. This is primarily caused by the fact that furthercases are formed, based on the nominative/accusative type, and thatseveral of the concerned body part terms retain the stem element assuch yet in the individual lE language. The formation of verbal dualforms based upon the first person personal pronoun takes place wherethe formation was no longer completed in the Proto-Indo-European period,which then does not take place in the language branches. Thus,the secondary endings may be reconstructed. In the case of syntagmata,using the substantive as a basis, a dual form and the number wordfor 'two' transferred the dual inflection over to the numera, thus echoingthe relation of syntagmata to pronouns and adjectives."Suggestions for further reading (a small selection): - Proto-IndoEuropean: -7 M. Fritz "Der urindogermanische Dual - eine Kiasse fursich?" in Graz 2000 p. 133-137; M. Malzahn "Die nominalen Flexionsendungendes idg. Duals" in HS 112 2000 p. 204-226; by the sameauthor "Die Genese des idg. Numerus Duaf ' in Graz 2000 p. 291-315. -Greek -7 Meier-Briigger Gr. Spr. IT 1992 p. 68f - Germanic: -7 K.

192 Prot()-Ind()-European MorphologyNouns and Adjectives193Strunk "War auch das andere Horn gemeint? Horn B von Gallehus undFragen des Duals" in PBB 114 1992 p. 179-211. - Tocharian: -+ J. Hilmarsson"The Dual Forms of Nouns and Pronouns in Tocharian' " Reykjavik1989 (= TIES, Suppl. Series I); O. Hackstein "On the Prehistory ofDual Inflection in the Tocharian Verb" in Sprache 35 1993 p. 47-70; M.Malzahn in TIES 9 2000 p. 45-52.2) A word on the singular and plural: Singular and plural are grammaticalcategories that are common to the verb and the noun. They permitone to indicate by means of congruence the association of the noun withthe subject of the action, indicated by the verb form employed. The relationshipof singular to plural is a question of syntax. Differentiations interms of content may effect formal changes: In the case of plurality theattention may be focused on the distributive-additive aspect (ct: Horn. Gr.1..aoi 'people, men, warriors': -+ M. Schrnidt in LfgrE IT Sp. 1634, 60ff.;ct: Horn. Gr. llpoi 'the individual thigh parts'), but one may also emphasizethe comprehensive-collective aspect of plura1ity, seeing in the collectivuma singular unity, e.g. (corresponding to the examples above) Horn. Gr.woe; 'people as a collective unity' and iipa 'all thigh pieces as a unifiedmass.' The distributive-additive forms belong to class A; the comprehensive-collectiveforms belong to class B, ct: F 303 § 2. For further information,cf. F 3l3. See also E 504 § 4.Suggestions for further reading: -+ Schmidt (J.) Neutra 1889; H. Eichner"Das Problem des Ansatzes eines urindogermanischen Numerus'Kollektiv ' ('Komprehensiv')" in Fachtagung Berlin 1983 [1985] p. 134-l69; J. A. Harilarson "Zum urindogermanischen Kollekliv" in MSS 481987 p. 71-113; E. Neu "Zum Kolleklivum im Hethitischen" in GrammaticaIlIita 1992 p. 197ff.; Prins Hittite neuter 1997; H. Craig Melchert"Tocharian Plurals in -nt- and Related Phenomena" in TIES 9 2000 p.53-75.F 305.Developments in case:I) External changes may be caused by phonemic changes, e.g. the attestedchange from Old Latin to Classical Latin in the nominative plural ending from-oj> -ej > ->-i.2) Changes may also take place through mutual influence of forms within asingle paradigm, or between two paradigms. For example, the typical Greekreplacement of the loc. pI. PIE *-su by -si, which took place in a preMycenaean period. The immediate cause of this paradigm-internal straighten-ing is the loc. sg. eoding with -i in combination with the instr. pI. eoding -JJ"i,which also contains -i-.3) Formal changes may also be understood against tbe backdrop of content-shifts.Thus, it is possible that paradigmatic categories that were onceclearly distinct became unified to form one single category. As a rule, this'casual syncretism' leads to an excess of inflected forms, because at the time ofeach differentiation a separate form was in use. In a first phase, the forms inquestion become alIornorphs. In a second phase, one of the alIornorphs generallyasserts itself as a norm, while the other(s) become disused. It ages and isgenerally forgotten by the speakers. - Concerning casual syncretism, cf. S404 and F 324 (with a Latin exemple). - Further: -+ H. Rix "Morph%gischeKonsequenzen des Synkretismus" in Proceedings of the Fourteenth InternationalCongress of Linguists n, edited by W. Bahner, J. Schildt and D.Viehweger, Berlin 1990, p. 1437-1441; by the same author, Hist. Gramm. dGr. 1976 § 121 £: Essential information is also offered in: Wackemagel Vorlesungen1 1926 p. 302f. Concerning the term 'syncretism,' which comes fromthe Greek for 'mix': -+ Kl. Pauly 5 Sp. 1648ff.2. Nominal Stem FormationF 306. The formation of nominal stems, and further the history of thesuffixes and suffix groups involved, is part of morphology, cf. W 200ff.The stem ending is of primordial importance for the understanding ofnoun inflection in ancient IE languages. It is thus important whether anoun stem ends in a vowel or a consonant. In the case of the acc. sg., dependingon the ending one finds the variants *-m or *-Tf/. Compare forexample the Greek acc. sg. -a that one finds following consonantal stemending (the type Kl]pUlc-a with -a < *-Tf/), but -n that is found following avocalic stem ending (the type 1to1..t-V with -n < *-m). A further ancientPIE distinction is made between athematic and thematic stems, cf. F 101 §4.In contrast, for an understanding of PIE inflection quite another thing isof importance, namely, the assignment of a nominal stem to an accent, orablaut class.F 307. In the organization and presentation of nominal stems, currentmanuals on ancient IE languages used the stem class principle. Stems are

194 Prote-Inde-European Morphologyclassed according to stem ending and additionally differentiated accordingto individual differences in the ending or the gender.I) Classical Latin declension is traditionally presented in grammars in termsof five declension types: I" declension with the type capra f. 'goat' (-iistems),2"" declension with the type lupus m. 'wolf and iugum n. 'yoke'(-o-stems), 3"' declension with the type 3A rex m 'king' and nomen n. 'name'(consonant stems) and type 3B ignis m. 'fire' and mare n. 'sea' (-i- stems), 4"declension with the type manus f. 'hand, flock' and genu n. 'knee' (-u- stems),5th declension with the type dies m. '(light of) day' (-e-stems). - The 5"declension was modeled according to dies diem. In the case of type 3B of the) '" declension, various inherited consonant stems were inserted, e.g. Lat.cani-s m. f. 'dog' < PIE *f{yon-, cf. Lat. iuveni-s m./f. 'young; young man,young woman' < PIE *hziey-h,on- ( Mayrhofer EWAia IT p. 413£; H. Rixin Etrnsci e Roma 1981 p. 108; for information on suffixes, cf. W 204 § 4), cf.La!. niivi-s f. 'ship' < PIE *neh,u- (cf. F 318 § 6c). - For more information: Leumann LLFL 1977 § 347; E. Risch "Dos System der lateinischen Deklinationen"(published in 1977) in Kleine Schriften 1981 p. 599ff.; Meiser LautundFormenlehre 1998 p. 129ff.2) Classical Greek is presented with the following declensions: The firstdeclension with the type lA, 'tt1!1\ 'ttl!fi f. 'honOT' and type IB tpa!leatpan:£" f. 'table,' 2"" declension with the type 'i m. 'horse' and uy6v n.'yoke,' and the 3 '" declension which includes various subgroups.Amongthem, there are those with, following the stem ending, an occlusive such asaly- f. 'goat,' those with -r- such as 1tat1\p 1tatp- m. 'futher,' those with anasal of the type 1tOtl!1\v m. 'herdsman,' those with -i- and -u- such as 1t6At f.'city' and 1l/iu.; . adj. 'sweet,' those with -ey- such as flaO'tA.£U; flaO'tA£OJc;(-eOJc; lon.-At! with metathesis of the older -fi [as in Homer], Myc. andCypr. -ey-os) m. 'king,' those with -s- such as y£v n. 'gender,' etc. - Formore information: Rix Hist. Gramm. d Gr. 1976 p. 127ff.; Meier-BruggerGr. Sprachw. IT 1992 p. 72ff.3) For the Old Indian corpus, Thumb I Hauschild Handbuch des SanskritI/2 1959 p. 3Off. presents declensions for the -a- and -iistems, -i-, -u- anddiphthong stems, -r- and -n- stems, and further plosive stems, sibilants, and, asa separate group, the heteroclites.4) Concerning Anatolian: Rieken Nom. Stammbildung 1999. - ConcerningGermanic, including its stems with -a-, -0-, -i-. -U-. -n-, root nouns andfurther consonant stems: Bammesberger Urgerm. Nomen 1990 p. 13ff. -Nouns and Adjectives 195For information on Old Church Slavic stem classes: Aitzetmiiller Abulg.Gramm. 1991 p. 68ff.F 308. Efforts to describe Proto-Indo-European declension are usuallybased upon the principles of stem class, which are commonly applied in theindividual lE languages: Szemerenyi Einfuhrung 1990 p. 173ff. (whichclassifies plosive stems, nasal and liquid stems, s-stems, i-, u- and diphthongstems, and thematic stems).However, a view that is becorning more and more prevalent proposesaccent and ablaut as the two most relevant criteria determining declensionclass in Proto-Indo-European: H. Eichner in Sprache 20 1974 p. 27f.with notes 1-2.In the present treatment, the sets of endings that are common to all substantivesshall first be listed (cf. F 309ff.), fo llowed by information on PIEaccent and ablaut classes (cf. F 314ff.).3. Nominal EndingsF 309. Substantives are generally composed of a nominal stem and anending, cf. F 1 0 I.The differentiation of thematic and athematic stems is a problem in itself,cf. F 101 § 4. It may be demonstrated that nominal endings wereinitially uniform, including for the thematic -0- stems, e.g. PIE athem. dat.sg. * ph,tr-ej 'father' vs. PIE them. da!. sg. *!Iergoj 'work.' The lattermay best be understood as the product of contraction of Pre-PIE*yergo-ej. In contrast to the vocalic -i- and -u- stems, in which the allophones-i- and -u- may alternate from being syllabic to non-syllabic (cf. L212), thematic stems followed by vowel-initial endings were contracted.Further, some of the -0- stems formed quali.fYing adjectives and thus cameinto relation with pronouns. Under the influence of these -0- stems, the -0- stem paradigm developed into an independent inflection, cf. F 311.Each substantive of an ancient lE language must belong to a gendercategory, whether it be masculine, feminine, or neuter. In terms of theirdevelopment, masculine and feminine may be grouped together, cf. F 303§ 2. Neuter forms must be distinguished from the others, cf. F 313.For information on nominal morphosyntax and casual categories, cf. S401 ff.

196 Proto-Indo-European MorphologyF 310. Proto-Indo-European athematic nominal endings (excluding neuterendings) may thus be presented in tabular form:s . !. duovoc. -9 as nom.nom. -s /-9 -es -h,ace. -m / -m -ns / -(IS < -m+sgen. l -es / -os / -s -omab!. J l -mdat.Jinstr. -eh, / -hi J _bh_loco -9 / -i -su-ejI) The set of endings above is the norm for substantives with root-finalconsontants. The substantives may be of masculine or feminine gender. Aspecial formal mark for this distinction is not necessary, e.g. PIE m. "ph,ter'mther' vs. PIE f. "meh,ter- 'mother,' cf. F 303 § 4. The best means of distinguishingthe gender of a noun is offered by adjectives and pronouns with threeendings that reveal congruence with the concerned noun, cf. F 301.2) The ending -9, which appears equally in the vocative, nominative, andlocative, reveals that this ending alone cannot have served to differentiate thecases in question. Differences in accent and ablaut must have provided decisivecriteria.3) While vowel and plosive stems are characterized by the ending -s in thenominative singular, the ending -n, combined with a long vowe is the normwith nasa liquid, or -s- stems, e.g. PIE "ph,tlr 'mther' or PIE "K(u)1j6n'dog.' This conspicuous distnbution may best be understood if we supposethat the -s- ending was initially used in the nominative singular of all sterns, andthat this ending was dropped under certain circumstances. Compare, in thecase of the -r- stems, the presumably Pre-PIE starting point "-Vrs, which wassimplified to PIE "-Vr via the transitory phase "-Vrr: Szemerenyi Einfiihrung1990 p. 121 f.; E. P. Harnp in Baltistica 31 /2 1996 (1998) p. 139f.4) The various genitive singular and instrumental singular endings may beexplained by the effects of accent and ablaut, e.g. in the stressed full grade gen.sg. -es in PIE "ph,tr-es 'mtber' in contrast to the unstressed zero grade -s ingen. sg. PIE "m(lt-ej-s 'thought.'5) In superficially comparing dative singuJars featuring PIE -ej with locativesingulars featuring PIE -i one is tempted, as in the case of § 4, to think of anold ablaut difference. However, full grade dative -ej and zero grade locative -iNouns and Adjectives 197were, aocording to all evidence established differentiated values already inProto-lndo-European. For this reason, full grade -ej could certainly be placedbefore a stressed full grade suffix, e.g. PIE dat. sg. *m(ll-ej-ej 'thought.' -Concerning the Mycenaean coexistence of -ej and -i: Hajnal Sprachsehiehten1997 p. 21ff. (p. 6Off. includes information on the Vedic typedivedive 'day for day'). - The locative had a special status with regard to theaccent and ablaut types, cf. F 318 § 6a and F 321 § I.6) The accusative singular with PIE "-m / -lfI and the accusative plural withPIE *-ns / -(IS are examples of ending variants that are determined by nominalstems with a consonant or a vowel in final position: *- V-m and -V-ns vs. -K-lfIand -K-(lS, cf F 306.Concerning the origin of the accusative plural, cf. F 104 (end). Concerningvarious accusative plural forms in lE languages (cf. among others,the -u- stem PIE "-u-ns and its three descendents in early IE languages"-uns; *-uns; "-us): H. Rix in FS Riseh 1986 p. 586-590.7) While the singular forms of the cases genitive and ablative are formallynot differentiated, both showing the endings (-es/-os/-s), they are inmct differentiated in the plural. The genitive features the separate ending"-om, while the ablative is linked to the dative:ab!.dat.instr.La!.-bus-busGr._phiIndo-lr.*-bhjas*-bhjasProto-Germ.*-m-*-m-*-m-Balto-Slav.*-mos*-mos*-mi(s)8) According to evidence in the individual lE languages, the instrumentalplural is linked with the dative and ablative plurals. Evidenceseems to indicate that while the dative and ablative plural were markedwith "-mos, the instrumental plural was marked with *-bhi: J. Katz PersonalPronouns 1998 p. 248f. Thus, _bh_ would have established itself inItalic and !ndo-Iranian as the sole initial consonant, replacing -m-. Conversely,-m- would have established itself in Balto-Slavic and Germanic.lndo-Iranian *-bhjas can thus be regarded as a cross between the instrumental"-bhi and the dative/ablative "-mos. See also J. Matzinger "Die"m-Kasus" des Balto-Slawisehen und Cermanischen" in CS HartmutKatz 2001 p. 183-208. - Concerning the problem of whether and to whatextent the Greek _phi « *bhi) bears singular traits (c[ Horn. dat.tq.l '(with) force' [already in Myc. as wi-pi- in the prefixes of two propernouns]): Meier-Briigger Cr. Spraehw. n 1992 F 302 § 2.

198 Prot(}-lnd(}-European MorpbologyF 311. The thematic, or -0- stem substantives form a group that is separatefrom all other substantives (cf. F 309) and features several formal pe_culiarities. Its range of forms reveals close connections to the pronouns.1) The following Proto-Indo-European schema was established based uponendings in the various ancient IE languages:svoc.as nom.nom. *-o-s *-os < *-o-es / *-ojace. *-o-m *-o-ns < *-o-m-sgen. '-o-sjo / adj. *-om < *-a-omab!. *-ot < *-a-et *-ojs < *-o-ojsdat. *-oj < *-o-ej as instr.instr. '-o-h, / '-e-h, '-o-mos / '-o-bh(j)osloc. *-0-; / *-e-i *-oj-su < *-ojs-su?'-e2) I t is clear that thematic substantives generally use the athematicendings, which is contracted with the thematic vowel -0-. The contractionspresumably took place at a Pre-PIE stage, c( F 309.3) Influences of pronouns that may be dated to the Proto-TndoEuropean period are revealed in the genitive singular, in which one findspronominal -o-sjo in place of the expected '-o-s or '-o-es (N.B.: the genitiveis sometimes replaced by a possessive adjective); the ablative singular,in which the genitive and ablative are differentiated through the addition ofthe pronominal ablative -et; the nominative plural, where -os is often replacedby -oj; the ablative plural, where pronominal -ojs is inserted; andthe locative plural, where one finds pronominal -oj[ s]- in place of simple-0-. For more information, c( F 405.4) The genitive singular: - Concerning '-osjo: Literature regardingthe genitive singular is difficult to assess. Newer sources are: H. Rix inMSS 49 1988 p. 107 (H. Rix makes the case that '-o-s is the expectedgenitive singular and that '-os-jo, and '-os-o was formed in the Pre-PIEperiod on the basis of nominal syntagmata such as 'pods h,e/tyos-jo andh,e/tyos-o 'the foot, that of the horse': -(H)jo or -0 should then be considereda relative pronoun, or an anaphoric demonstrative pronoun, whichcould be added to the substantive in question in the Pre-PIE period; concerningnominal relative clauses, c( Chr. Koch in S 205 at the end of § 2);L. A. Prosdocimi in Studi Etruschi 57 1991 p. 152ff. (which concerns theLepontic xosiosio); A. Nikolaev "PIE Ergativity and the Genitive in ._O.!JIO " in UCLA lE Conference 1999 [2000] p. 293-309. - In place of!.Nouns and Adjectives 199'-osio, sometimes the ending '-I is found in ancient IE languages. Theendg '-I is the only genitive singular ending in Celtic. Although a coupleof -osjo- forms are attested in Latino-Faliscan (e.g. the incription fromSatricum, 500 B.C.: Popliosio Valesiosio suodales 'the comrades of PubliusValerius'), starting in the 51h century B.C. the -1- genitives are thenorm. In Osco-Umbrian one finds -eis in place of -1-. Concerning '-1 inProto-Indo-European: G. Klingenschrnitt in Kolloquium Lat. u. ldg.Salzburg 1986 [1992] p. 98-104. The author proposes an 'appertinentive'featuring PIE '-iH-, the chief purpose of which would have been to indicatefamilial possession. This would explain the perceptible competition inancient IE languages between *-1 and the genitive ·-os(j)o. - An adjectivemay be used in place of a genitive to indicate possession. A significanttreatment: 1. Wackemagel "Genetiv und Adjektiv" in Kleine SchriftenIT 1969 (an essay of 1908) p. 1346-1373. Suggestions for further reading: F. Bader "Les genitifs-adjectifs determimis et le probleme de l'article:comparaison typologique entre l'etrusque et les langues indoeuropeennes"in FS Rix 1993 p. 12-45; 1. Hajnal "Der adjektivische Genetivausdruckder luwischen Sprachen" in Graz 2000 p. 159-184.5) For more information on -0- stem paradigms in ancient IE languages(including plural forms): Rix Hist. Gramm. d. Gr. 1976 p. 135ff. (2'ddeclension); G. Klingenschrnitt in Kolloquium Lat. u. ldg. Salzburg 1986[1992] p. 93ff.; Sihler New Comparative Grammar 1995 p. 256ff. - Concerningthe Vedic -ena: Hauri -ena 1963.F 312. The PIE feminine '-e-hrstems (becoming ·-ah,. and then -astems)are, from their origins with final h" athematic consonant stems.The complex suffix '-e-h,. must be seen as an '-h,. derivation from thematicstems, cf. W 204 § 1. For more information on the understanding of·-h,. and '-e-hr, c( F 303 § 4. The ·-e-h,. suffix originally belonged tothe PIE class B described in F 303 § 2. This explains why the ending Os,which marks a subject, is strangely missing in the feminine nominative singular,c( F 313.Under pressure from the ' -0- stem adjectives, which included feminine•-e-h,. in order to mark gender congruence, the athematic paradigm of·-e-h,. was adapted to the thematic paradigm of the masculine '-o-stems,Particularly since ·-eh,. was inflected usually without an ablaut (cp. F 322)and the ·-eh,., followed by vowel-initial endings, contracted after disappearanceof ·-h,., making the new forms externally similar to the '-0-stems, which were themselves already contracted. Compare the Latin -astemsof the I" declension, which, in traditional grammars, are grouped

200 Prot-lnd-European Morphologywith the -0- stems of the 2" " declension, and at the same time separatedfrom the nouns of the 3' d declension. Further, compare (in the nominativeplural) PIE '-e-hres (> '-as), retained in Vedic (as senas) and Gothic (asgibOs) vs. renewed fern. -aj in Latin (e.g. terrae) and in Greek (e.g. 't1.!J,ai)fo llowing the example of thematic pronominal '-oj.Further references: For information on the vocative singular, c( L 334§ 3; concerning the accusative singUlar and plural, cr. L 303.Suggestions fo r further reading: For information concerning problemsassociated with -ehr stems: -? Rix Hist. Gramm. d. Gr. 1976 p. I 29ff; G.K1ingenschrnitt in Kolloquium Lat. u. Idg. Salzburg 1986 [1992] p. 89ff. ;Sihler New Comparative Grammar 1995 p. 266ff; I. Hajnal "Die lykischena-Sttimme: Zum Werdegang einer Nominalklasse" in KolloquiumPedersen Copenhagen 1993 [1994] p. 135-171; by the same author, see L334 § 3 above.F 313. Neuter fo rms generally feature the same endings as masculineforms, but reveal different in the nominative and accusative cases: Thefo rmal identity of nominative and accusative may be explained by the factthat a main characteristic of PIE neuter forms (and oftheir underlying PIEclass B) was to identifY things that were not conceivable as agents in averbal phrase, cf. F 303 § 2. Thus, the formation of a nominative case (inthe sg. with the ending *-s to mark the subject) did not initially come intoquestion.Nominative/accusative neuter singular:While athematic substantiveshave no endings (zero) in the nominative and accusative, thematic substantivesfeature the single '-m ending, e.g. PIE athem. nom./acc.'spennTJ-8 'the planted' vs. PIE them. nom./acc. *!!ergo-m 'the work.'Nominative/accusative neuter plural: In terms of content, the idea of acollective mass is certainly dominant. Therefore, the collective suffix (=athematic *-hr and thematic *-e-hr) is used, no ending (zero) added. If,however, the distributive/additive aspect were to be stressed, a plural ofclass A could have been used. Cr. F 304 § 2.The understanding of the neuter plural as collective explains the ancientlE characteristic, observable in isolated cases, of combination of the neuterplural and the singular of a verb: -? 1. A. Harllarson in MSS 48 1987 p.81ff. with mention of the practice in Attic Greek (A good example is thesentence: 7lovta pet 'everything flows': -? Meier-Briigger Gr. Sprachw. I1992 p. 157), in Vedic (the RV only contains three attestations, which arenot necessarily based on old formations), Old Avestan, in which it is regular,and in Hittite, in which it is exclusively used. - Further sources: -? H.Nouns and Adjectives 201Craig Melchert in TlES 9 2000 p. 53-75 (p. 61ff. "Collective vs. CountPlural in PIE and in Anatolian").4. Inflection Paradigms and their Ablaut ClassesF 314. As a quick glance in the research history shows, it was onlygradually that the insight became established that the fundamental organizationof PIE nominal inflection is defined by the accent classes, or moreprecisely stated, by the ablaut classes that guide them. - The fo llowingdepiction is oriented according to research history, culminating in the currentunderstanding of the subject in F 319. Research in this area is verymuch in progress.1) As is generally known, the athematic nominal paradigm shows a changebetween the strong and the weak stem The weak stem may be distinguishedfrom the strong by a different accent behavior, e.g. in the case of the paradigmof'futher' PIE strong 'phr/{!r- (the suffix is full grade) vs. PIE weak 'phrtr- '(the suffix is zero grade). Accordingly, in Greek there is nom sg. strongM-niP, gen. sg. weak 7lO-tp-6

202 Proto-lndo-European MorphologyIn tbe case of the fonner, the suffix in the weak stem is zero grade and theending full grade; in that of the fonner, the suffix of the weak stem is fullgrade, and the ending is zero grade. Pedersen uses the terms (p. 24 note I):'' 'F/exion forle ' el 'jIexion faible ' de F de Saussure" (--? Saussure Memoire1879 = Recueif 1922 p. 187, I 94ff., 205ff.)."Mais ces lermes prelenl a la confusion avec les expressions 'slark ' el'schwach ' de la grammaire des langues germaniques, el en soi ifsn 'exprimenl pas la vraie nalure du contrasle entre les deux types. Enadoplanl le mol OVYallll; au sens de 'degre vocalique fori ' on pourrailpeul-elre forger les lermes 'jIexion hyslerodyname ' [of the type 1lO-nlP, orpilcl] et 'jIexion prOlerodyname ' [ofthe type sUnu-s]."In other words, Pedersen distinguishes on the one hand what he caJls the hysterodynamicinflection type with its strong stem (according to the schema:zero grade unstressed root, full grade stressed suffix, and zero grade unstressedending) and weak stem (according to the schema: zero grade unstressedroot, zero grade unstressed suffix, and full grade stressed ending), andon the other hand, what he calls the proterodynamic inflection type with astrong stem (according to the scbema: full grade stressed root, zero gradeunstressed suffix and zero grade unstressed ending) and weak stem (accordingto the scbema: zero grade unstressed root, full grade stressed suffix, and zerograde unstressed ending).4) Pedersen makes clear that the original combinations of stressed fullgrade and unstressed zero grade' are only rarely preserved in the lE languages.Thus, a large number of analogies and innovations have taken place in extantlE paradigms. In the case of nom sg. sUnUs, for example, neither the zerograde stem, nor the stressed zero grade suffix correspond with the postulatedproterodynamic ending form *se1}il-nu-s. In that of the Vedic gen. sg. pi-tur(with -Iur

204Proto-lndo-European MorphologyNouns and Adjectives 205strong stem nom. acc.weak stemloc. sg.singular*ued-n-collective*ud-n- '*ud-en(i)8) Further, in contrast to the singular form *!!odr-, PIE * jJkwr- perhapshad the weak stem * jokWr- (--7 Nussbaum Two Studies 1998 p. 150 note179), from which the Latin ioc- might have received its -0-. But this is notthe last word on this subject: --7 Meiser Laut- und Formenlehre 1998 p.142.F 315. Various authors have dealt with accent and ablaut paradigms inrecent years.I) The first to be named is 1. Schindler. Concerning his life's work: --7 R.Schmitt in A1mbA W [Almanach der Osterreichischen Akademie der Wissenschaften]145 1994 / 1995 p. 584ff.). Following are his major essays: --7 "Dasindogermanische Wortfor 'Erde ' und die dentalen Spiranten" in Sprache 131967 p. 191-205; "Zu hethitisch nekuz" in ZVS 81 1967 p. 290-303; "L'apophoniedes noms racines indo-europliens" in BSL 67/1 1972 p. 31-38; "ZumAblaut der neutralen s-Stiimme des lndogermanischen" in Fachtagung Regensburg1973 [1975] p. 259-267 (p. 262f. on the four PIE ablaut classes);"L 'apophonie des themes indo-europliens en -rln" in BSL 70/1 1975 p. 1-10;''Alte und neue Fragen zum indogermanischen Nomen" in Kolloquium PedersenCopenhagen 1993 [1994] p. 397-400.2) From the abundance of other sources, I shall name: - a) H. Eichner"Die Elym% gie von Hilt. mehur" in MSS 31 1973 p. 91 (in note 33 regardingterminology, cf. § 3 and 4); by the same author, "Zu Elymologie undFlexion von vedisch str, und pumiin" in Sprache 20 1974 p. 26-42. - b) RixHist. Gramm. d Gr. 1976 p. 121ff. -c) Beekes IENominal Injlection 1985.(Among other things, Beekes postulates boldly the existence of a hysterodynarnictype with nom sg. KeK-K, ace. sg. KK-eK-", and gen. sg. KK-os, aswell as various sub-paradigms. Neuter forms, on the other hand, use a proterodynarnicparadigm); Lubotsky Nominal Accentuation 1988. - d) Onvarious related subjects: W. Hock "Der urindogermanische Flexionsakzentund die morphologische Akzentologiekanzeption" in MSS 53 1992 p. 177-205; N. Oettinger "Der Akzent des indogermanischen Kollektivums im Lichtedes Hethitischen" in MSS 53 1992 p. 207-214; by the same author, "Der Ablaulvon 'Ahorn' im lndogermanischen" in HS 107 1994 p. 77-86. - e) X.Tremblay "Un nouveau type apophonique des noms alhlimatiques sujfaauxde /'indo-europeen" in BSL 91 /1 1996 p. 97-145. The author tries to establishthe additional 'anakinetic' inflection type according to the mode nom sg.Kek-6s and *Kek-Ior- and gen. sg. K

206Proto-Indo-European MorphologyNouns and Adjectives 2072. hysterokinetic (transfer of accent from the suffix to the ending)3. amphikinetic (transfer of accent from the root to the ending)4. holokinetic (transfer of accent between root, suffix, and ending)strongweaknom. sg. Ved. su-nu-sinstead ofPIE * seuH-nu-sgen. sg. Ved. su-no-s < * su-miy-s < PIE * suH-ney-sF 316.In PIE nominal inflection, the ablaut classes proterodynamic andhysterodynamic were on the one hand well established, cf. F 314 § 3ff. Asshall be shown in F 319, the additional categories acrodynamic and amphilholodynamicmust also be included.In tbe fo llowing depiction, I shall use the abbreviations W (= root), S (=suffix) and E (= ending), and z (zero grade; borrowing from J. Schindler, Ishall use z [= zero] and not 8: -7 Sprache 1 5 1969 p. 144; by the sameauthor, Fachlagung Regensburg 1973 [1975] p. 262ff.), e (stressed, egrade), 0 (stressed, o-grade) and 0 (unstressed, o-grade). 'Strong' standsfo r 'strong stem'; 'weak' stands for 'weak stem.'The appropriate accent pattern is primary.The ablaut pattern associatedwith it is originally only the product of the accent circumstances, accordingto the principle, 'stressed syllable = e-grade,' 'unstressed syllable= zero grade or o-grade.'In the evolution of Proto-Indo-European into the individual lE languages,the ablaut was more often retained than the accent.Among theablauts, the circumstances in the suffix and at the ending remained themost stable. Ablauts in the root were most often dropped in favor of formaluniformity.F 317.The proterokinetic/proterodynamic nominal class:I) A schematic depiction ofthe Proto-lndo-European proterokinetic accentpattern (the underlined elements carry an accent):::g I li I;2) A schematic depiction of the Proto-Indo-European proterodynamic airlaut pattern:I! I3) Consider the model example Ved. 'son' (cf. F 314 § 3; for furtherstrongweakdetails: -7 Mayrhofer EWAia II p. 74 1):4) Commentary on § 3: While the suffixal ablaut change is attested inV ed·lC, such evidence is missing fo r the ablaut change m the root syllable_Thus the reconstruction must remain uncertam · · m thi s regar d.Asansu-. ,alternative to full grade * seyH-, stressed, zero gra d e * su 'H- canno t he. . .ruled out. - The paradigm of Ved. sunu- consIStently exhibits a stabesuffixal accent. However, this state may not be very old. The accent ISexpected on the full grade suffix of the weak stern, but not on the zero .rade suffix of the strong stem. In other words: The ablaut behaVlor ofhe suffix may only be explained if an older proterokinetic accent precededthe newer static accent.5) The various PIE -i- and -u- stems are distinctively proterodynamic:*KeK-i-s*KeK-u-s* KzK-ej-s *KzK-ey-s6) As a rule, IE languages have retained the -i- and -u :suffixal ablautmarker.However, in root ablaut fo rmation and accentualton, stmplificationis the rule: Starting from the weak stems, zero grade roots were genera1izedand accents became static on the root or the suffix.7) As examples of § 5, I shall e a few abstract Ii-/-Iej- nouns: -Latin: mors morlis f. 'dying, death < PIE weak *mr-leJ- (which IS an .abstract-li- noun from PIE *mer- 'to disappear, to die') with the generalizedzero grade mor- < *mr-. For information on other Latin descendents fabstract -ti- nouns: -7 Leurnarm LLF 1977 p. 344f.; Retchler-BeguelinType mens 1986 p. 23. - Greek: Hom. 1toen

208 Proto-Indo-European MorphologyNouns and Adjectives209'city' shows the same pattern as the -si-abstracts. - Vedic: mati- f.'thinking, thought, sense' < PIE weak *mr;-tej- with generalization of theweak stem's zero grade in the root, and of the weak stem's accent on thesuffix. A genitive singular form such as PU$les 'flourishing, growth'

2\0 Proto-Indo-European Morphology J. Schindler in BSL 7011 1975 p. 10. Concerning Hittite: CHD P-I1994 p. 12-16; Rieken Nom. Stammbildung 1999 p. 331-333.F 318. The hysterokineticlhysterodynamic nominal class:I} A schematic depiction of the Proto-lndo-European hysterokinetic accentpattern (the underlined elements carry an accent):stronglE I Iweak2} A schematic depiction of the corresponding Proto-lndo-European hysterodynamicablaut pattern:strongweakI3} Cf the prime example, Ved. 'father' (cf F 314 § Iff.; on phonetics, cf. L324 § 3):strong Ved. nom. sg. pi-ta: < PIE *phrter

212 Prolo-Indo-European Morphologynom. sg. *djeys a well established Proto-Italic nom. sg. dies, attested particularlyin Diespiter: This may best be understood when we presume thatthe form was very early restructured, in analogy to acc. sg. *djem . Concerningthe pre-Lucian form dipoteres: H. Rix in FS Hamp 11 1997 p.146-149 « *Di, pater < *Die pater). - Concerning the Greek and Mycenaeanparadigms (note: Mycenaean Oat. sg. Diyej is parallel to HeroD: Aura Jorro DMie. I 1985 p. 180 s. v. di-we; E. Risch "Die mykenisehenPersonennamen au! _e" in Traetata Myeenaea 1987 p. 281- 298 (on thetype ku-ne / ku-ne-u, among others); HI. Hagen "Die Diskussion um dieSchreibweise von Zijv () im homerisehen Epos" in Glolla 72 1995 p. 98-104; J. Martfnez Garcfa "Quod lieet 10ui, non lieet boui, Zum grieehisehenNamen ZeV;' in HS 110 1997 p. 211-214 (He asserts that the accentuationof !Xlii

214 Proto-Indo-European MorphologyI shall first give an overview of all four Proto-Indo-European classes,using, for the sake of simplification only the 'dynamic' ablaut patterns.The exceptional status of the locative has to be noted.I) The acrodynamic nominal class (cf. F 320):strong 0w S E"or e z zweak e z zloco z eIl) The proterodynamic nominal class (cf. F 317):w S Estrong e z zweak z e z"locoz eIll) The hysterodynamic nominal class (cf. F 318):W S Estrong z e zweak z z eloco z eIV) The amphi- or holodynamic nominal class (cf. F 321):W S Estrong e 0 zweak z z eloco z eF 320. The acrostatic, or acrodynamic nominal class I.This class reveals the fo llowing accent schema:strongweakI!IThe ablaut in the root syllable produces the differentiation betweenstrong and weak sterns. - The normal schema is strong -0- vs. weak -e-.J. Schindler showed ( BSL 6711 1972 p. 32-36) that a part of the rootnouns belongs to this class, namely the type with the strong stem *dom-Nouns and Adjectives 215vs. weak stem *dem-, cf. § I below. Depending upon their content, theseroot nouns may be considered, according to Schindler p. 36, partially as"subs/an/ifs jeminins a valeur resultative ou passive," and partially as"noms d'agent (substantifs et adjectifs), souvent avec une nuance iterative."The amphidynamic root nouns may be distinguished from these acrostaticroot nouns, cf. § 3 below. It is worth noting that individual acrostaticroot nouns secondarily became amphidynamic, cf. PIE * lajon- vs.weak *liun-' instead of exspected *lajen- 'dog' (details on the subject inSchindler, ibid. p. 33-36). - In addition to the class with root vowel -0-vs. -e-, a separate sub-class may be distinguished with the schema rootvowel strong -e- vs. weak -e- (or -0-), cf. § 2 below.I a) A prime example for the type -0- vs. -e- in the root syllable: - PIE*dom- 'house' (details: Mayrhofer EWAia I p. 697 S.V. dam- and 699S. v. dampati-; R. Lipp in LIV 1998 p. 98 note I suggests tracing *dombackto *domh,-, because of the verbal form *demh,- 'to unite, to build',cf. L 330; however the addition of", is not mandatory; on the acc. sg. seealso L 303 ):PIEancient IE languagesstrong nom. sg. *dom-sacc. sg. *dam Arm . tun; Gr. OW< *dom-mweak gen. sg. *dem-s used in Gr. oecr( no'l11C;) and inVed. dampati-=; *dm-es Arm . tanloco sg. *dem-i Ved. dame (with athematic -e?)Ib) A further example for the type with -0- vs. -e- in the root syllable:- PIE *n6kW-t- 'night' (for details: Rieken Nom. Stammbildung 1999p. 128f.; the noun represents the verbal abstract noun corresponding toPIE *nekw- 'to dawn': Mayrhofer EWAia Il p. 2f.):PIEIEstrong nom. sg. *nokW-t- s La!. noxacc. sg. *n6kW-t-mweak I gen. sg. *nekW-t-s Hitt. nekur'loco loco sg. *m!kW-t2) PIE strong *jekW-r vs. weak *jokw-f}- 'liver' suggest a separate subclass,cf. F 314 § 6 and 8.

216 Proto-lndo-European Morphology3) In addition to the aerodynamic root nouns of the type strong -6- vs.weak -e- in the root syllable (cf. § I), a second group of root nouns of thetype strong -e- vs. weak -z- may be demonstrated: J. Schindler in BSL6711 1972 p. 36-38. This type falls under the category of 'amphidynamic'nouns, cf. F 321 § 3. A prime example is *djey- 'sky, sky god, day' withthe strong stem *djey- vs. weak stem *diy- ', cf. F 318 § 6. In terms ofcontent, according to Schindler (p. 38) one finds "noms d'actions" and"noms d'agent tires de verbes d'etat."F 321. The amphidynamic and the holodynamic nominal class IV.The amphidynamic paradigm is characterized by its direct accent transferfrom the root to the ending: The full grade root with zero grade endingis correspondingly exchanged with the zero grade root and its fullgrade ending. This exchange is typical for a part of the root nouns. However,when, along with the root, a suffix is also added, and thus a locativesingular with a stressed full grade suffix, then the whole is to be considereda holodynamic ablaut pattern: J. Schindler in Fachtagung Regensburg1973 [1975] p. 262f.A schematic depiction of the amphi- and holokinetic accent pattern (theformer with W+E; the latter with W+S+E):W S Estrong W S Eweak W Sloc. sg. W S EI) The Proto-Indo-European word for 'earth' furnishes the classic exampleof a holodynamic ablaut paradigm ( Mayrhofer EW Aia I p. 424f. with furtherinformation on details; a central work concerning all issues associatedwith this word is given by 1. Schindler in Sprache 23 1977 p. 31; on the ace.sg. see L 303):PIElEstrong nom. sg. *dh e 'goom-Hitt. te-e-luin; Gr. X9-u\vacc. sg. *dheghom Ved. 4tim< *dheK'om-mweak gen. sg. *dhgom-esHitt. tak-na-a-as::;* gh",_es cf Gr. xal1-ai*f!tm-es Ved.jmas *dhog"m_esENOUDS and AdjectivesIloe. sg.1* dhgoem::;*goo em1 Ved. 4ami; Gr. initial X9-2) Further examples of the holodynamic class are furnished by nonneuter(animate) simplicia of the type PIE fern. *h,eys-os- 'dawn redness'with PIE strong nom. sg. * h,eys-os (cf. Horn. Gr. 1\

218 ProtD-lndD-European MorphologySchindler's research ( Wurzelnomen 1972 and cf. F 315 § I above), theroot nouns featuring both PIE accent types (acrostatic and amphikinetic)represent a particularly interesting group, cf F 320. Lastly: K.-H.Mottausch in HS 113 2000 p. 29-52. - The classes are further importantfo r the understanding of the suffix system named after W. Caland, cf. W206.5. Fonnal Characteristics of the AdjectiveF 323. The dimension of gender is not connected with the lexeme in thecase of adjectives.On the contrary, adjectives are capable of fo rmingmasculine, feminine, and neuter fo nns in order to show congruence, cf. F301.As a rule, adjectives may accept all three sets of endings. However, inlE languages, there are also adjectives with only two endings, cf § 3 below.1) The majority of adjectives have the -o-stem, among which I includeverbal adjectives and compound adjectives, e.g. PIE *dej!!-o- 'belongingto the sky(god), heavenly,' cf. PIE *Klu-to- 'hear, heard named, famous,'cf. Ved. an-udr-o- 'waterless.' The corresponding feminine fo rms featurePIE *-ehr (>-ii- in the lE languages), cf. F 312.2) But there are also adjectives (again, including verbal adjectives andcompound adjectives) with -i- and -u- stems, and also with consonantstems, e.g. PIE *s!!eh2du- 'sweet' (= La!. sviivis < *s!!iid,l-i-; Gr. 1\ou

220 Prot-lnd-European Morphologyand -mh20- 'recall the ordinal nwnbers (cf. F 503) as they developed fromthem. The ordinal numbers were often used to mark the concluding, andmost important elements of a series, thus receiving superlative traits: ,Benveniste Noms d'agent 1948 p. 144ff.; Risch Kleine Schriften 1981 p.684ff. - Each of the individual IE languages reveals its own typical characteristics.Suggestions for further reading: , Seiler Sleigerungsformen1950; Szemerenyi Einfiihrung 1990 p. 203-214.la) In the primary comparative form, the comparative suffix PIE '-ios(with ablaut '-is-; the ablaut pattern is holodynarnic, cp. F 321 § 2) replacesthe positive suffix directly fo llowing the (generally) full grade rootof the adjective stem, e.g. La!. positive magnus (i.e. mag-no-) 'great' withcomparative maior « 'mag-ios-; while the a in the root is short, the syllableis counted as long because of its double consonant; clever schoolbooks thus often cite maior 'greater,' although it is not completely correct)and mag-is 'more'; Gr. positive KaK-o- 'bad' with the comparativeKaK-irov 'worse'; Ved. positive nav-a- 'new, fresh, young' with the comparativenav-yas- / nav-/yas- 'new, newer, newest (in case forms such asthe accusative singular neuter, instrwnental, and dative) anew.' - Theform PIE '-ios- varies alIophonically with PIE '-iios (which replaces it innominative singular masculine and fe minine forms with the structureKV.K- and KVR.K-; whereas, according to rules of phonetics, '-ios- isexpected in forms with three or more syllables), cf. L 218 § I on Siever'sLaw. - Evidence seems to suggest that the suffix -ios- belongs to thegroup of fo rms derived from -es-stems, which feature masculine/feminineand neuter forms, e.g. Lal. m dolor (stem 'dolhl-os-: from the PIE verbalroot 'delhl- 'to hew, to split') 'pain' vs. n. genus (stem 'genhl-es-) 'gender.'The form in question, -ios- should then be considered a complexsuffix, which was initially in pre-PIE time formed as -es-derivative from-ie- verbal stems, but then, when newly re-anaJyzed, could be linked dirctlyto the verbal root: root-je- + -es-> root- + -ies-. If Gr. KaUo

222 Proto-indo-European Morphology(Xenophanes) with -sson < -k-jon 'sweeter,' but yl..uK-iOlV (in Homer'sOdyssey). Material for further discussion: -t J.-L. Perpillou in BSL 69 / I1974 p. 99ff - A further problem, specific to Attic Greek, is the lengtheningof the syllable preceding -jos-, e.g. Att. -ijon- vs. Ion. -ijon-, cf. Atl.].I£iOlV vs. Ion. OlV 'better', cf Atl. KpeiUOlv vs. Ion. KP£ClClOlV'stronger' (where Ion. -ss- and Atl. -/t- < -Ij-: -t Risch Kleine Schriften1981 p. 506). According to information provided by forms like(lClClOV 'nearer' and iiUov 'more, rather' with Attic --a- and not ---e-, thedevelopment cannot be ancient (an old Ionic-Attic --a- would have beenaffected by the ii > if vowel shift, cf L 205). Risch (Worlbi/dung 1974 p.89) sees a replacement for the initially common full grade of the root inthis lengthening. Other full grade forms are KP£ClClOlV «* krel-jOn: butsuperlative Kpa1-lCl-10- < *lq-t-iS-IO-) and 6A£iOlV (concerning Atticinscriptions: -t Threatte Attic IT 1996 p. 309) 'smaller in number' «*-lejg-jon: but positive 6Uy-o-). - In the superlative, the Greek is -is-Io-.The form is attested as early as the Mycenaean period, cf nom. pI. n.me-Id-ta i.e. meg-isla.Id) A prime example in Vedic: - positive taw is- 'strong, powerful'with comparative Iliv-iyas-, and tav-yas- 'stronger, more powerful.' Forthe superlative, cf yilv-io$!ha- 'recently, recently born' fromyuv-an- « PIE*h,je!!-h,on-. For details: -t Mayrhofer EWAia II p. 413; concerning thesuffix, cf W 204 § 4) 'young.' - Comparative: While the coexistence of-iyas- and -yas- is of post-PIE origin, it can (as can Greek) in its essentialsbe traced back to PIE alIophonic arrangement mentioned at the beginning(la). The cause of the lengthening of i is not clear,. although one reasoncould be an origin in sel-roots, cf. just above Iliviyas- < *tti!fiHjas- withlaryngeal metathesis < *ta!!Hijas- < PIE *te!!h,-ijos-. - The superlative inVedic is -i$-!ha- < *-is-th,o-.2a) The secondary comparative form uses the PIE suffix *-(t)ero-. -Good examples include Gr. OiKOLO

224 Proto-indo-European MorphologyD. Pronouns1. General InformationF 400. Pronouns are among the most important deictic and anaphoricelements of the sentence. Only pronouns share the important grammaticalcategory 'person' with the verb, ct: S 301. Contrarily to the 3" person"celui qui est absent, la non-person ne, " the I" person marks "le 'ie' quienonce" and the 2nd person, "le 'tu I auquel 'je ' s 'adresse." The correspondingplural forms are more complex. Whereas the 3'" person permitsa true plural, the I" person plural "nous" is either to be interpreted as "moi+ vous" or as "moi + eux"; "le 'nous ' annexe au 'ie ' une globalite indistincted'autres personnes." And for the 2"" person plural: '' 'vous, ' qu 'i/s 'agisse du 'vous ' collecti/ ou du 'vous ' de politesse, on reconnaft unegeneralisation du 'tu "': E. Benveniste "Structure des relations de personnedans le verbe" in Benveniste Problemes I 1966 (essay from 1946) p.225ff.I) It has been shown that Proto-Indo-European had, on the one hand, whatare called single-gender personal pronouns, including the possessive pronounsderived from them, and, on the other hand, what are called the double-genderinterrogative and indefinite pronouns as well as, in ancient IE languages, thetriple-gender demonstrative and relative pronouns. The conservation of theoriginal forms in the IE languages varies. The most changes affected demonstrativepronouns, which could lose their referential effect, and then must benewly determined.2) In comparison with the normal order and enclisis in a nominal or verbalsyntagma, the position of emphasis at the beginning of a sentence increases thepronominal information. The difference between enclisis and emphasis (orthotony)was not manifested exclusively by the presence (or absence) of theaccent, but rather also morphologically, i.e. through the shortened form in thecase of enclitica as opposed to the longer stressed form.3) Pronouns played an important role in the development of nominal inflection.Their influence on the o-stems is unmistakable, cC F 3 11.4) Suggestions for fiuther reading: - a) general: Benveniste (see above);Wackernagel Vor/esungen IT 1928 p. 75ff; Metzler Lexikon Sprache 1993 Sp.487 s. v. 'Pronomen.' - b) Proto-lndo-European: Krahe Idg. Sprachw. ITPrOrtouns 2251969 p. 38-46; Szemerenyi Einfuhrung 1990 p. 215-24; R. S. P. kes"The origin 0/ the Indo-European pronominal mj/ectlOn m FS Poome 1 ,?8873-87; F. Bader "Les pronoms dans les langues indo-europeennes mtemoireS de la Societe de Linguistique de Paris N.S. I 1990 p. 23-35.- c)Latin: Leumann LLFL 1977 p. 460-484; Meiser Laut- und Formenlehre1998 p.156-169. - d) Greek: Rix Hist. Gramm. d G . 1976 p. 174-189:Meier-Briigger Gr. Spraclnv. IT 1992 p. 85-93. - e) VedIC: J. Wackernagelin Wackernagel / Debrunner Ai. Grammatik ill 1930 p. 431-59 . .- f)Anatolian: E. Neu "Zu einigen Pronommalformen des Hethlflschen m FSPuhvel I 1997 p. 139-169. - g) Celtic: Schrijver Celtic Pronouns andParticles 1997. - For more information, see the sections on the individualgroups below, particularly in F 401 § 4.2. Personal and Possessive PronounsF 401. Personal pronouns are part of the hereditary lexicon of the lElanguages. For information on gyntax, cf. S 400.I) However incomplete, the fo llowing table shows the most importantcharacteristics of PIE personal pronouns, which do not distinguish betweengenders. - Thus, in the case of '1' (speaker) and 'you' (interlocutor)there is no distinction between masculine and feminine gender. -The nominative is distinguished from other cases in having its own stern,cC PIE nom. sg. ·eg-oh, vs. ace. sg. PIE ·m-. - Since every finite verbform automatically indicates the 'person' of the verb, the nominal pronounforms are already adequately marked. - In comparison with the orthotonicforms, which are often strengthened by particles, the enclitic formsfeature the minimal word stem and may be used in multiple cases, cf. orthotonicPIE nom. sg. ·eg- with suffix -oh, or -hrom vs. enclitic ·m-oj forgen. and dat. sg. - For the 3" person, several demonstrative pronounsare possible. These are presented in F 405f. Only the 3 sg. reflexive pronounis treated in § 3 below. Even externally, one may discern that PIEacc. sg. ·se belongs to I sg. ·me and 2 sg. ·te.I se. 2 se. I nl. 2 pI.nom. sg. ·eg-oh,·yejs ·juHsonly stressed ·eiIh,-omacc. sg. encl. ·me ·te / ·'ye *nos *yosace. sg. *me *te *1Js-me • us-mestressed·/"h,

226 Proto-Indo-European Morphologygen. dat. sg. *moj *tojencl.gen. sg. *mene ·(eyestresseddat. sg. *me-g'rej *te-bhei *7}s-me-i ·us-me-istressed *me-fthi-om *te-bhi : o2) Details concerning Proto-Indo-European: - J. T. Katz ( "Archaischekeltische Personalpronomina aus indogermanischer Sicht" inFachtagung Innsbruck 1996 [1998] p. 265-291;) asserts, based on obliquecases in Celtic material, the existence of an ending *-me for the 1" personand an ending *-!!e for the 2'd person: In place of the commonly reconstructed1 pI. *1Js-me and 2 pI. *us-me, the author thus suggests a new 1pI. *1Js-me vs. 2 pI. *us-Ife and consequently holds the balance in favor of*-me in the 2" person plural fo r secondary; thus, from this perspective the2"" person singular form *t!!e is old (the !! from *t!!e is generally seen ascoming from the nominative 'tu-H-), and further *me < om-me. - Latin: Leumann LLFL 1977 p. 461ff; Meiser Laul- und Formenlehre 1998 p.156ff. (concerning dissimilation in Pre-PIE dat. sg. 'me-bhej to PIE*me-gi'ej etc.). - Details on Greek: Rix Hist. Gramm. d. Gr. 1976 p.176ff. The fo rms of the I" and 2" person plural are in Proto-Greek newlyconstructed on the basis of the acc. * Qs-me and *us-me. For the 1 pI. cf.Aeol. nom. and acc. iillIlE; in At!. cf. nom. "l!Ei

228 Proto-Indo-European Morphologynine. This absence is ancient and particularly plausible in Proto-IndoEuropean, in which presumably the feminine was only firmly establishedafter the splitting off of Anatolian, cf. F 303 § 2.While the sentence-initial orthotone forms mark questions such as'who?,' 'what?,' 'from whom?,' 'whose?,' etc., the enclitic forms could be'moderated' to form indefinite pronouns meaning, for example, 'whoever'or 'anyone,' etc.The situation in lE languages varies. The range of application of * kWi_vs. *kwo- is difficult to determine, particularly since *kWi- and *kwo- werealso used in restrictive relative clauses just as the appositive * Hjo-. Forinformation on the apparent connection of *kwo_ with the copula *kwe, cf.G. Dunkel in FS Narten 2000 p. 9-29. For further information, cf. F 404.The fo llowing table serves to clarify a few points concerning the inflectionof the substantive interrogative pronoun. The reconstructed forms aretaken from Rix Hist. Cramm. d. Cr. 1976 p. 187. The present work isinappropriate for a discussion of the extent to which Rix's postulated differentiationof m.lf. *kwo- (cf. Lat. cuius < *kwo;lo-) from n. *kwe_ (cf.Avest. cahiiii mit ca- < *ce- < *ke-

230 Prot-lnd-European MorphologyPronouns23 1The fo llowing table illustrates the most important masculine and neuterforms:sg. m. sg. n. 101. m. 01. n.nom. *s6 I "to-d *16-j *te-h,acc. *to-m I "to-d *to-ns *te-h,gen. *to-s(j)o *tojs-omdat.*to-smojloco*t6jsu < *'6js-su?instr. "to-hi *IoisThe inflection of pronouns reveals characteristics that distinguish itfrom nominal inflection: cf. -d (and not -m) in nom. acc. sg. n., cf. -j innom. pI. m. (and not -es), cf. -s(j)o in gen. sg., cf. -sm- in dat. sg. and cf.stem toj(s)- in gen. and loco pI. The -0- stem nominal forms show this influence,cf. F 311 § 3. Concerning the much discussed problem of gen. sg.-s(j)o, cf. F 311 § 4.Further sources: - Concerning the dubious invariable * so (called the"sa-fige" in Vedic, Greek, and Anatolian): G. E. Dunkel in KolloquiumWackernagel Basel 1988 [1990) p. 100lf; S. W. Jarnison Vedic 'sa-jige ':"An inherited sentence connective?" in HS 105 1992 p. 213-239: J. S.K1ein 'sa-jige ' and Indo-European Deixis " in HS 109 1996 p. 21-39; G.E. Dunkel in IF 102 1997 p. 176f.; C. Watkins in FS Narten 2000 p. 263-281. - Concerning the coexistence of "so / "to-: K. Strunk in Glotta55 1977 p. 7-9.F 406.A further demonstrative pronoun with a Proto-Indo-Europeanbasis is PIE *i- (or, with ablaut, *ej-): cf. in Latin, anaphoric is, ea, id, andi-dem, ea-dem, id-em ( Meiser Laut- und Formenlehre 1998 p. 159-161); cf. in Greek the formal remnants ll-lV (already min as early as Mycenaean)and V-l v ( Wackernagel Kleine Schriften I 1969 p. 10); cf. Ved.m. ay-am, f. iy-am, n. i-d-am ( Mayrhofer EW Aia I p. 103). Concerningthe entire body of issues: Szemen:nyi Einfuhrung 1990 p. 218-220.The form *(I'I}e (unstressed "-o) is also datable to the Proto-IndoEuropean period. It found use as an augment (cf. F 213); and further presumablyplayed a decisive role in the formation of middle voice endings(cf. F 210), as in the genitive *-os-o (cf. F 311 § 4).The individual lE languages developed (mostly on the basis of ProtoIndo-European material) an entire series of new pronouns, cf. Lat. hichaec hoc, iste ista istud, and ipse ipsa ipsum, cf. Gr. OOE, OU10

232 Proto-Indo-European MorphologyNumerals233coine of the Oscians, Latins, and Etruscians: --7 H. Rix "Buchs/abe, Zahlwortund Ziff er im AI/en Milleli/alien" in FS Pisani IT 1969 p. 845-856; by thesame author, "Die Al/ersangabe in der oskischen Inschrifi Ve. 70 und OscanUmbrian akno- 'Jahr"' in MSS 37 1978 p. 149-163. As Rix is able to show,the presumption of older research that Middle Italian nwneraJs were in fact reinterpretedwriting symbols must be corrected.3) Further sources: - a) General: --7 Melzler Lexikon Sprache 1993 p.430 s. v. 'Numerale.' - b) Proto-Indo-European: --7 Krahe Idg. Sprachw. IT1969 p. 46-49; Szerrerenyi Numerals 1960; Eichner Numeralia 1982 (on 2-5);Szemerenyi Einfuhrung 1990 p. 234-243; Numerals 1992 (an anthology witha complex history; the fo reword gives details; the individual books refer to thevolwne that was left for a considerable time unpublished, cf LewandowskiLinguist. W6rlerbuch 2 1994 p. 754 s. v. 'Numerale' with the entry by Ross A.c., ed., "Indo-European numerals." 1981 and Metzler Lexikon 1.c. in section AJ. Grozdanovic [sic] (Gvozdanovic) (ed.), "Indo-European NwneraJs." Berlin,1988; contains descriptions of the situation in various rE languages: cf. RColeman 'Italic,' F. M. J. Waanders 'Greek,' R Emmerick 'Old Indian,' H.Eichner 'Anatolian,' A. S. C. Ross and J. Bems 'Germanic' and B. Comrie'Balto-Slavonic'); B1aZek Numerals 1999. - c) Particularly concerning Latin:--7 Leumann LLFL 1977 p. 484-495; Meiser Laul- und Formenlehre 1998 p.170-177. - d) Particularly concerning Greek: --7 Rix Hisl. Gramm. d. Gr.1976 p. 171-173; Meier-Brilgger Gr. Sprachw. IT 1992 p. 93-98; F. Kortlandt"Greek nwneraJs and PIE glotta1ic consonants" in MSS 42 1983 p. 97-104. -e) Particularly on Vedic: --7 J. Wackemagel in Wackemagel / Debrunner Ai.Grammalik 111 1930 p. 329-430; Thumb / Hauschild Handbuch des Sanskrit I/ 2 1959 p.1 54-166. - f) Particularly on Anatolian: --7 H. Eichner in Numerals1992 p. 29-96; O. Carruba "Betrachlungen zu den ana/olischen und indogermanischenZahlwdrlern" in Fachlagung Innsbruck 1996 [1998] p. 505-519. - g) Particularly on Germanic languages: --7 Krahe / Meid Germ.Sprachw. IT 1969 p. 87-94. - h) Balto-Slavic: --7 B. Comrie in Numerals1992 p. 717-833. - i) Armenian, Tocharian: --7 F. Kortlandt "ProtoArmenian numerals" in Kolloquium Pedersen Copenhagen 1993 [1994] p.253-257; W. Winter "Tocharian" in Numerals 1992 p. 97-161.2. Cardinal NwnbersF SOL The cardinal numbers are part of the inherited lexicon quite likethe terms fo r farni1ial relationships of the type PIE 'ph,ter- 'futher.' -bly the cardinal numbers were not declined initially in pre-PIEPresuro a ,.' . . . _ The genitive was used to comrnurucate quantItIes, as IS th e casedividual lE languages from the number 'five' on (e.g. Old Church.)_SW .In order to mark congruence, in Proto-Indo-European themina! endings are included fo r the lower numbers. The para d' Igms finoorthe num bers 'one' to 'four' presumably fe atured a single set of endings fo r. . .masculine and fe minine fo rms WIth addItIonal fo rms fo r neuter, an d sometunesalso fo r feminine fo rms. - Numerals were often inserted as prefixes.oli ' d" Wf possessive compound forms, cf. the type Gr. tPl-lt - tnpo m08. In this case they are generally zero grade, cf. PIE *s1[I- 'one' = La!..sern- > sim- in sim-plex 'simple,' Gr. a- in ii-1t 'one tune'; V .ed. sa- msa-/cft 'one time' or cf. PIE *kWtljr- 'four' in Homenc Greektp';tj>

234 Proto-Indo-European MorphologyNumerals 2354) 'four': - PIE * kWetljor- = Lat. qualluor; Gr. Att. tEnop (

236 Prolo-Indo-European Morphology3. Ordinal NwnbersF 503. In Proto-Indo-European, and naturally also later, ordinal numbersbad tbe function of signaling the end of a series, e.g. 'We traveled for ninedays. But on the tenth ...: 1. Wackemagel in Wackemagel / DebrunnerAi. Gramm. 1I1 1930 p. 400ff.; Risch Kleine Schriften 1981 (in an essay of1962) p. 684ff. (including many examples from ancient lE languages). Asis otherwise common in ancient IE languages, ordinal numbers are -o-stemadjectivs derived from the cardinal numbers. In the case of 'first' and'second,' non-cardinal derived fo rmations were possible too, cf La!.primus 'foremost' in the sense of 'first' and secundus 'following (adj.)' inthe sense of 'second.' The ordinal numbers are the starting point for superlativeformations, cf F 325.The common PIE ordinal suffixes are -10- and -111]0-. One also finds themore ancient -0-, sometimes alternating with the complex form -h,o-, cf.PIE *sept'l'-(h,)o-'seventh.' The old -o-stem formation *dtiK'l'-lh,-o- 'belongingto the decad', derived from the PIE abstract noun *dtili1[l-lehr is apotential starting point for the ordinal suffix variant PIE *-h,o- and alsofor the PIE variant *-Ih,o- of *-10-': Concerning the understanding of theabstract noun *-I(e)h,-: Kurylowicz Eludes I 1935 p. 49; the VedicsuffIX -Iha- can indeed best be traced hack to PIE *-Ih,-o-, cf. L 329 § Iwith further examples for Ved. -Ih - < PIE *-Ih,-. - Concerning detailsin Latin: Leumann LLFL 1977 p. 49 1ff. and Meiser Laul- und Formenlehre1998 p. 174f. ; - Greek: Rix Hisl. Gramm. d. Gr. 1976 p.170f.; - Vedic: J. Wackemagel l.c. - On the subject as a whole: -4G. Schrnidt "Indogermanische Ordinalzahlen" in IF 97 1992 p. 197-235.Numerals. . --t Leurnann LLFL 1977 p. 494f.; Wackernagel I Debrunner Ai.000.Gramm. ill 1930 p. 422ff.2374. Nwnerical AdverbsF 504. In the case of the numerical adverbs of the type of multiplicatives'once,' 'twice,' the word formations of the lowest numbers have their originsin Proto-Indo-European, cf. PIE *dtli-s 'twice' (= La!. bis with theconsonant shift dl}- > b-) and PIE *Iri-s 'thrice' (= Gr. tpi). To express'once' there existed a compound form with the prefix *s'l'-, cf. in F SOlthe three PIE representatives with varying suffixes. For further informa-7 The "2 in *delbflth:ro- is primarily part of the abstract noun formatiom, but is thenadded by the speakers to the suffix through a restructuration from *deKTflthro- to*deKrp-th,o- and to *de/i1[lt-h,o-.

General Information 239IV. Proto-Indo-European SyntaxA. General Information- by Matthias Fritz -S 100. Berthold Delbrtick's Vergleichende Syntax der indogermanischenSprachen was the first comprehensive and fundamental treatment of thefield and has not been replaced in the hundred years since its appearance: Delbriick Vgl. Syntax I-TTI 1893-1900. The importance of this workwas underscored by a conference dedicated to its author on the hundredthanniversary of its publication: Kolloquium Delbruck Madrid 1994[1997). For information on the role of this work in the history of the field 'cf. K. Strunk p. 571 IfI) Whereas since the time of A. Schleicher efforts have been made to reconstructthe phonology and morphology of Proto-Indo-European (see E305), B. Delbriick does not reconstruct the syntax of Proto-Indo-European,but rather compares and analyzes phenomena of the individual lE languages.Only the future may tell whether or not this 'pre-Schleicherian' approach toresearch on syntax, particularly on that concerning larger parts of sentences, isInSurmountable. Only those sentence parts may be reconstructed which revealcomparable information in both content and expression. However it is oftendifficult to decide whether these comparable phenomena are preset in ProtoIndo-European, or Proto-Indo-European merely provided the conditions forrelated developments in various individual languages.2) Alongside B. Delbriick's work, the comprehensive treatments of syntaxby Karl Brugrnann ( Brugrnann Einfacher Satz 1925, by the same author,Kurze Vgl. Gramm. 1902-1904; Grundrifi 11-3 1916 on particles in the simplesentence) Hermann Hirt ( Hirt Idg. Gramm. VI 1934 and VII 1937) areworth narrung. Although a contemporary equivalent of these works does notexist, one is planned within the scope of the Indogermanischen Grammatik byJ Kurylowicz and M. Mayrhofer and is presently being prepared by a numer'us staff led by Jose Luis Garcia Ram6n, Heinrich Hettrich, and Oswaldanagl ( For more information, see the web sites of the universities of Wiirzburg and Salzburg. See links on our web site [see E 100), under the heading'lrxlo-European linguistics in Europe'). The Indogermanische Gesellschaftheld a conference on this subject Wiirzburg from 29 September to 3 October1999: Indogermanische Syntax, - Fragen und Perspektiven -, H. Hettriched. , WteSbaden 2002.3) Winfred P. Lehmann ( Lehmann PIE Syntax 1974) and Paul Friedrich( P. Friedrich Syntax 1975) offer manageable treatments which, however,are more concerned with the more narrow syntactical question of word orderin view of typology and which, owing to their limited scale, cannot offer acomprehensive treatment. For an in-depth discussion of the subordinated importanceof word order in Proto-Indo-European with its wealth of inflectedforms, see C. Watkins "Towards Proto-lndo-European Syntax: problems andpseudo-problems" in Watkins Selected Writings I 1994 (in a contribution from1976) p. 242-263.4) A reconstruction of Pro to-Indo-European syntax is dependent on an idealizednotion ofa standard Proto-Indo-European. Questions of style, to whichparticularly O. Panagl has dedicated himself, demand particular sensitivity.The language of poetry is of particular importance due to its traditional formulas: R. Scbrnitt Dichtersprache 1967; Watkins How to kill a dragon1995. Owing to the very nature of the coexistence of oral innovations andwritten tradition, coUoquial language is only with difficuhy accessible. Potentialsources include above all texts in which dialogues occur, of either dramatic,philosophical, or lyrical nature: Hofrnann Lat. Umgangssprache1978.5) Basic literature on Proto-Indo-European and Indo-European syntax includes: Brugrnann Grundrifi IJ-2 1911 and 11-3 1916; by the same author,Einfacher Satz 1925; Biibler Sprachtheorie 1934; Chantraine Gramm. hom. 111953; Delbriick Vgl. Syntax I-ill 1893-1900; by the same author, Gr. Syntax1879; by the same author, Ai. Syntax 1888; Havers Erkltirende Syntax 1931;Hirt Idg. Gramm. VI - VII 1934- 1937; Hofrnann / Szantyr Syntax 1965;Kieckers Gr. Gr. ill-IV 1926; Krabe Vergl. Syntax 1972; Kiibner / Gertb 111898-1904; Kiibner / Stegrnann IT 1955; Kurylowicz Inflectional Categories1964; Matthews Syntax 1981; Meier-Briigger Gr. Sprachw. I 1992; MiklosichVergl. Syntax 1868-1874; Monro Homer. Dialect 1891; Pinkster LateinischeSyntax 1988; Scherer Lat. Syntax 1975; Sch,wyzer / Debrunner Gr. Gr. 11

240Prol-Ind-European Syntax1950; Sommer Verg/. Syntax 1931; Speyer Syntax 18%; Wackemagel Var/e_sungen I-ll 1926-1928.S 101. Given t?eir dependence on questions of that which signifies andthat which tS SIgnified, efforts at reconstructing Proto-Indo-Europeansyntax are necessarily more reliant on hypotheses and probabilities than arethe reaintS of morphology and phonology. Internal syntactical reconstruc_tIon ?oes even one step further, since the observation of divergences ascẹrtamedthrough external reconstruction is based upon preconceived notionsof a syntactịcal system and also because the target idea that resultsfrom the comparIson of historical syntactical systems ('cleartSing' them)can only be made plausible in terms of typology.The pr ?blematic nature of syntax reconstruction is comparable to thatof semantics fa r as syntactic meanings, like semantic meanings, belongto the lingUIstic realm of content: Thus, morphemes, like lexemes, consistof slgnijica s and signijicatum and are used by the speaker to indicate hisconceptual Idea (conceptual level: designatum) of the language-externalworl. As in the case of lexemes, morphemes may be assigned basicmeanmgs that are not dependent on context: Thus, the functions of amorpheme may vary according to the lexemes that appear in a sentence.The ?ngUlst'c code employed by the speaker directly refers to his conceptal ,?ea which, ill turn, tS dependent on his interpretation of the actuals, uatlon such that various speakers may discuss the same real situationWIth varying linguistic codes.The fo llowing schematic diagram describes the relationship of linguisticcode to conceptuality and reality in different terms:signifiant significanssemantic level signifie silffiificatumconceptual level designatumlanguage-external realityS 102. Syntax is the scientific description of sentences. From the outsetquestions must be addressed as to the nature of a sentence and the criterithat a ntence must fulfill in order to be considered as such. In the casef prehistoc linguistic research it becomes important to address the additionalquestion of how sentences may be reconstructed.Even from the standpoint of contemporary languages, the question ofwhat constttutes a sentence is not simple to artSwer. It may in any caseonly be clearly and definillvely answered, if sentences in themselves revealGeneral Information 241quisites to such an artSwer; i.e., when sentences contain certainthe pre ttre • Stics which make them sentences, and without which they wouldcharac ernotbe sentences. If sentences do not contain these charactenstlcs, t e.questiondefinition.of the fundamental criteria of the sentence becomes a matter of. .. .f b'Taking, fo r example, the sentence criterion of the pamng 0 a su ect;' d' we may also accept as sentences phrases that are quite different, fo r·th a predicate, then, along with such sentences as 'The cat caught ae:ple the use of the German noun, 'Hi/fet' .The latter t may ofc e in such cases as well, as it does in another formulation that could bese be explained in such a way that the predlcahon of a subject takesh:sen in the same situation: 'Hi/f mirt' Thus, the differences are limitedto the signijiant.One might suggest that in order fo r a sentence to be taken as such, Itmust be complete and meaningful. Of course, incomplete sentences arenot necessarily meaningless, and complete sentences are not necessarilymeaningful. Examples give the impression that complete sentences aremore or less meaningful, or rather, in themselves meaningful. However,sentence examples are not linguistic reality. In the natural linguistic reality,each expression exists within a larger context and serves the goal of communication.Errors of execution, i.e. miscommunication, are natural in thisas in any other human activity. However, miscommunication may occur incomplete sentences as well as in incomplete ones .On the otm:r hand, theincomplete sentence may fulfill its purpose in a larger context Just as wellas the complete sentence, and is, of course, omnipresent in daily speech.The decisive factor is understandability in context. Thus, understandabilityand completeness of an individual sentence apart from its context carmotbe considered fundamental and essential qualities of a sentence.With regard to syntactical reconstruction, the assertion that real sentences,i.e. sentences in their full context, carmot be reconstructed, is selfevident.This is not a failing of syntactical reconstruction, fo r, as we havejust seen individual sentences without a context, like individual words andsounds, e not linguistic reality. That which carmot be reconstructed isthe linguistic reality; that which can be reconstructed is the system of alanguage. Individual examples can only serve to illustrate the reconstructedsystem. Whether or not the system of an entire language can bereconstructed is quite another question, the artSwer to which is more dependenton the state of research than on research methods. Also dependenton the state of research is the ability to fo rmulate reconstructed sentences,fo r the fo rmulation of a sentence, which has no claim to linguistich.

242 Proto-Indo-European Syntaxreality, is hindered not by the fact that sentences cannot be reconstructedbut rather by the fact that one may not be able to verifY that the recon:structed elements all belong to the same historical phase of language. Ifthe historical contemporaneousness of all of the elements in a formulationwere assured, nothing would be left to hinder its assertion. While this inno way guarantees that the formulated sentence existed in its given fonn,this limitation affects all reconstructions.Suggestions fo r further reading: C. Watkins "Preliminaries to the reconstructionof Indo-European sentence structure" in Proceedings of the9'h International Congress of Linguists Cambridge Massachusetts, editedby H. G. Lunt. The Hague 1964 p. 1035-1042; by the same author, "TowardsProto-Indo-European Syntax: Problems and Pseudo-Problems " inWatkins Selected Writings I 1994 (in a contribution from 1976) p. 242-263; W. Dressier "Ober die Rekonstruktion der indogermanischen Syntax"in ZVS 85 1971 p. 5-22; W. Winter "Reconstructional ComparativeLinguistics and the Reconstruction of the Syntax of Undocumented Stagesin the Development of Languages and Language Families " in HistoricalSyntax 1984 p. 613-625; S. Jamison "Determining the Synchronic Syntaxof a Dead Language " in Historical Linguistics 1989 p. 21 1-220.B. Sentence Syntax1. General InformationS 200. Whether compound or simple, the individual sentence is not thelargest unit of syntax. Also included in syntax are the syntactical and semanticrelations that go beyond the limits of the single sentence. Hence,one speaks of 'text syntax' in current research. In the appearance of hypotaxisfrom parataxis in main clauses and subordinate clauses, the directconnection between text syntax and sentence syntax becomes clear, owingto the supra-sentential effect of pronouns. This is illustrated in Latin bythe use of sentence-initial relative pronouns.S 201. Phrases may be divided into hypotactic and paratactic phrasesaccording to whether or not they are included within other phrases. Accordingto whether they are independent in terms of content, they can besentence Syntax 243divided into main and subordinate clauses. According to the intention ofthe speaker they may be divided into statements, questions, and imperativesentences. According to whether or not they contain a finite verb, theymay be divided into verbal and nominal sentences.S 202. There are infinitive and participial constructions that may betraced to the use of verbal nouns in certain Proto-Indo-European cases.Among the verbal nouns in Proto-Indo-European are included those verbalsubstantives and verbal adjectives that are included in verbal paradigms oflE languages as infinitives and participles. While it is as yet unclearwhether one may postulate the use of particular infinitive or participialconstructions in Proto-Indo-European, the existence of participles maysafely be asserted. Additionally, infinitive constructions with final dative,accusative of direction, and the locative of destination are presumed tohave existed. While infinitives are defined by syntax, the very presence ofparticiples reveals that there were participial constructions in Proto-IndoEuropean. According to J. L. Garcia Ramon, "[the assertion] of a trueProto-Indo-European infinitive ending in ·-sen(i) is justified,'" whichcould be traced to a locative form, which, according to K. Stiiber, appearsin the case of s-stem abstract nouns with the locative forms ending in ·en"Even in what are known as absolute constructions, there was the possibilityin Proto-Indo-European of employing various cases to insert a phrasein a sentence. According to Keydana the locative was used in such cases.For more information of morphology, cf. F 216 10Suggestions for further reading: Gippert Infinitive 1978; by the sameauthor, "Zum 'prtidikotiven ' Injinitiv" in ZVS 97 1984 p. 205-220; by thesame author, "Ein keltischer Beitrag zur indogermanischen Morphosyntax:Das altirische Verbalnomen" in Kolloquium Delbriick Madrid 1994[1997] p. 143-164; H. Hettrich "Zur historischen Syntax der nomina actionisim f!.gveda: Der 'doppelte Dativ "' in MSS 43 1984 p. 55-106; bythe same author, "Nochmals zu Gerundium und Gerundivum" in FS Rix1993 p. 190-208; Risch Gerund. 1984; J. L. Garcia Ramon "Zur Konkurrenzvon Dativ und Akkusativ von Nomina actionis und Abstrakta im Indogermanischen"in GS Kurylowicz 1995 [1996] p. 101-1 13; by the same8 Infinitive im Indogermanischen? Zur Typologie der lnfinilivbildungen und zu ihrerEntwicklung in den tilteren indogermanischen Sprachen in 1nl 20 1997 p. 45-699 K. StUber Zur Herkunft der altindischen Infinitive auf -slmi in MSS 60 2000 p.135-16710Absolute Konstruktionen 1997 p. 33

244Prol-Ind-European SyntaxSenlence Syntax245author, "Infinitive im Indogermanischen? Zur Typologie der Infinitlvbtldungenund zu ihrer Entwicklung in den alteren indogermanischenSprachen" in InL 20 1997 p. 45-69; Keydana Absol. Konstr. 1997; K.Stiiber "Zur Herkunft der altindischen Infinitive auf -sant' in MSS 602000 p. 135-167.2. Parataxis and HypotaxisS 203.Parataxis is the coordination of sentence clauses, which maythemselves be statements, questions, or imperative sentences.Statementsand questions may contain the same elements, differing only slightly inword order and intonation, such that the one can easily be transformedinto the other. Imperative sentences, on the other hand, like the imperativemode, constitute a special case.I) In statements, the speaker takes a position regarding the validity he attachesto the content of his expression; fo r example, whether the validity ispresupposed (injunctive) or claimed (indicative), or whether he regards thestatement as a possibility (optative), or as taking place in the future (subjunctive).Only with the perfective aspect could the indicative mood be used todescribe future events, since the perfective aspect does not permit statementsabout the immediate present. However, in Proto-Indo-European the perfectIve.aspect of the indicative aorist, paired with the augment and secondaryendmg, IS only used to describe past events. When paired with the primaryendmg ,:"thout the augment, the perfective aspect only appears in the subjunctiveaoftSt, revealing the probable origin of the association of the future tenseand the subjunctive mood.2) Unlike statements, questions express the incomplete or uncertain idea ofthe speaker . .Accordingly, one may distinguish completive questions from yesor.-no questIons. In completive questions, interrogative pronourtS and pronominaladverbs take the places of nominal (including pronominal) and adverbialsyntagmata, respectively. In yes-r-no questions, no replacement of word typetakes place; rather the entire verbal action of the sentence is put in question.This may be accomplished through variation of the word order and intonationof the statement fo rm. However, the thematization ofa finite verb may not beconfirmed on the basis of evidence from individual languages.For information on the Proto-Indo-European questions *kwis hlesi?'Who are you?' and *kwosjo hlesi? 'Whose (son) are you?': R. SchmittDichtersprache 1967 p. 136£S 204.Along with parataxis (coordination), there is also evidence ofhypotaxis (subordination) in Proto-Indo-European.HypotaxISmay be.understood as the insertion of one phrase into another. The essentIal charteristicof hypotaxis is the possibility of including one phrase with a finite:rb fo rm in another such phrase, i.e. the integration of one verbal clauseinto another.. . . . . .The insertion of one clause mto another IS accomparued by var13tlons mthe form of the subordinate clause from the form that it had as a mainclause.Subordinate clauses are thus differentiated from ll\3in clauses inthat their signijicans is of a fo rm that is not fo und in coordinated clauses,whereby the fo rmal difference may in fact be limited to suprasegmentalphonology. In terms of content, given that a subordinate clause only representsa completion of, or additional information about, a main clause, aparatactic relationship may exist between independent sentences that issimilar to the hypotaxis between main and subordinate clauses. However,one can speak of hypotaxis only when this relationship is expressedthrough the fo rmal integration of the subordinate clause in the main clauseas in the case of adverbs, attributes, and appositions. The fo rmal characteristicsof subordinate clauses vary among the individual lE languages. InProto-Indo-European, the accentuation of the finite verb is accepted as afo rmal characteristic of the subordinate clause as opposed to the mainclause, in which the finite verb is not accentuated, except when it establishesthe theme at the beginning of the sentence. A summary of the criteriathat distinguish main clauses from subordinate clauses may be fo und in E.Hermann "Gab es im Indogermanischen Nebensalze?" in KZ 33 1895 p.481-535.In Syntax III 1900, Delbriick asserts, "that ... originally all sentenceswere coordinated alongside one another." (p. 411) He continues:"Thehistorical view, as it is generally accepted today, must have as its point ofdeparture the hypothesis that there was a time at which there were onlymain clauses." (p. 412) Further, speaking of his time, Delbruck states:"Tbe assertion that hypotaxis developed from parataxis has become thecommon heritage of the field." (p. 413) In order to systematize subordinateclauses, Delbriick proposes, (1900, 413f.) their division into a prioriand a posteriori subordinate clauses according to "thought relationship."S 205.In Proto-Indo-European, the existence of subordinate clauseswith relative pronouns may be asserted, which, in addition to filling therole of a relative clause, may also serve as adverbial phrases. The comparisonofIE languages does not support the existence of special conjunc-

246 Proto-lndo-European Syntaxlions for the introduction of adverbial phrases. Adverbial phrases tbat areintroduced by counctions bave their origin in relative clauses. This isevid 7nced by the fact that conjunctions may be traced to certain cases ofrelatIve pronouns, e.g. Lat. cum < quom = acc. sg. ill of the relative pronoun'kwo_. Following the origin of adverbial phrases, whereby conjunc_lions found their origins in relative pronouns, other parts of speech werealso transformed into conjunctions. Thus, subordinated clauses that areintroduced ?y relative pronouns can perform the function of subject, obJect, adverbIal phrase, appositional phrase, and attribute.I) Proto-Indo-European features two types of relative clauses: attributiveand appositive. These are distinguished from one another by sentence-initialpronouns: The attnbutive relative clause is introduced by the pronoun *k"'i- /'k'" 0-; \be appositive, by the pronoun ' Hjo-. There is one semantic differencebetween the two types of relative clause: While the attributive relative clausedescribes an additional quality, \be appositive relative clause states a knownquality of tbat to which the relative pronoun refers, i.e. the referent. The attributiveelative clause may thus appear without a referent, while \be apposit,verelallve clause may only be used in connection with a referent.In thepost-Proto-Indo-European period, this fonnal characteristic of the two typesof relallve clause was abandoned, such tbat the individual lE languages eachuse just one of\bese relative pronouns.2) According to Ch. Lehmann, Proto-Indo-European features only postnominaland sentence-initial relative clauses, and none that are prenominaJ andsentence-final." 'PostnominaJ' and 'prenominaJ' indicate the placement of therelative clause relative to \be referent; 'sentence-initial' and 'sentence-finaJ'indicate the placement of the relative clause in relation to the main clause. InLehrnann's .view, \be .'Hjo- relative clause occurs only in a postnominaJ posilionand ongmates eIther from an attribute that was added to 'Hio- or froman independent sentence introduced by 'Hio-. In fuet, • Hjo- an anaphoricfuntion. The sentence-initial k"'i-/k'" 0- phrase, according to Lehmann,serves to indIcate a \berne, to which \be fo llowing main clause refers: ''Thenucleus of a restri c:tive relative clause [sic] is necessarily semantically indefirute;thus, \be relallve pronoun that detennines it must be indefinite. The useof a relative pronoun tbat is closely related both to the interrogative pronounand t ?the indefinite pronoun is thus explained . .. " (p. 163).According toHettrich (Hypotaxe, 1988 p. 776-778), fo llOwing a linguistic phase which fea-sentence Syntax 247relative participles, relative clauses develop fro paratactically coorditU:clauses with anaphoric. pronouns, process which leads m turn to thena . n of appositive-explicallve relallve sentences (from sentences WIth:aJ anaphoric 'kwi_/'k'" 0-). For ịnformation n morplogy, c .f F 404.furrn;:::c 'Hjo-) and attributive-restrictive relative clauses (from sentences.. . . .FM m ore information regardmg the termmology apposlllve vs. attnbu-t ·d· Ch,. Seiler Relativsatz 1960; concerrung apposItIve vs. restnctlve: LehmanO Relativsatz 1984 and Hettrich Hypotaxe 1988. Accor mg toK It, there was no relative attribute in Proto-Indo-European, but rather Itoc· f · 1 1 ·developed in individual lE languages upon the baSIS 0 nomma re aliveclauses.123. Verbal and Nominal PhrasesS 206. In a verbal phrase, the finite verb determines the actants, whichare governed by rules of dependence. In the nominal phrase, there IS nofinite verb to determine the actants. Nominal phrases are not simply verbalphrases without a finite verb (with what is called elipsis of the copula), butrather constitute an independent type of clause. Thus, the predIcate nounin nominal phrases is always stressed, unlike the verbal predicate in verbalphrases. In fact the term 'elipsis' is not exact, since the copula is not essential.Contrarily, the use of the copula should rather be seen as an adaptationto the common pattern of verbal phrases, which always feature afinite verb form. This use of the copula is in fact a sort of explicative signification,in which the content of the copula is expressed through theconnection of the various sentence elements and is given particular emphasisalone through an independent linguistic symbol of comparable meaning.S 207. For more information on parataxis: -7 Kieckers Steflung desVerbs 191 1; Ammann Untersuchungen I 1922; by the same author, "Untersuchungenzur homerischen Wortfolge und Satzstruktur. 2. Te il: DieSteflung des Verbums, im Einzelnen untersucht" in IF 42 1924 p. 149-171and 300-322; W. Dressier "Eine textsyntaktische Regel der idg. Wortstellung"in ZVS 83 1969 p. 1-25; R. Harweg "Zum Verhiiltnis von Satz,Hauptsatz und Nebensatz" in ZDL 38 1971 p. 16-46; C. Watkins "SomeIndo-European verb-phrases and their transformations " in Selected... .I JCh. Lehmann "Der indogermanische ·kwi-Ikwo-Relativ.vatz im typ% gischenVergleich", Kol/oquium Syntax Pavia 1979 [1980J p. 155-16912"Zur Vorgeschichle des relaliven Atlribu/ivkonnexes im Baltischen und Slavischen"in Kolloquium ldg., Slow. u. Bolt. Jen. 1989 [1992J p. 45-88

248 Proto-Indo-European Syntaxsentence Syntax249Writings I 1994 (in a contribution from 1975) p. 189-209' K HParataxe und H t ' K.'. egerfo axe m wartalnik Neojilologiczny 24 1977 p. 279_86, A.. Scherer Rekonstruktion grundsprachlicher Satzbauplane" in PSzemerenYl 1979 p. 755-762; Andersen Word Order Typology 1983' FBader "Structure de I 'enonce indo-eur0'Peen" in Historl'cal L '1987 13 34' L". . .. ' . .mgU/stlcsp. .- , uraghi Old Hlltlte Sentence Structure 1990' A HintParataxIs and Hypotaxis in the Avesta" in Syntaxe des la s ilramennes anciennes 1993 [1997] p. 51-62.guzendo-! 20. For mor :; information on hypotaxis: --+ Seiler Relativsatz 1960't2i;:;:,emann Tempora/e und konditionale Nebensatze des Hethi.chem MllIellungen des InslItuts fur Orientjorschung 11 1965 P 23 174 and 377-415; R. Schmitt-Brandt "Vergleich der indogerman;scheebensatzkostruktionen" in Fachtagung Bern 1969 [1973] p. 125-141'. Bossong Typologle der Hypotaxe" in FoL 13 1979 p. 33-54; H. Jfu"Abf::sa: Ko;;Plemente Im Urindogermanischen" in FS Szemerenyi 19797p. '. ' . Hettrch "Zur Entwicklung der Fina/satze altindoger_Samsc er 'Prachen:' m ZVS 100 1987 p. 219-237; by the same authorLatelmsche KondlllOnalsatze in sprachverg/eichender Sichf' in Kollo ''f:,um Lat. u. Jdg. Salzburg 1986 [1992] p. 263-284; by the same author I t:: taxeS 209. With regard t.1988; Lehmarm Re/ativsatz 1984; Krisch Konditiona/satz;. 0 sentence accent, one may note that the word thatbegms the se?te?ce IS stressed. Sentence-initial position implies the function0 establishing the topic: In nominative If. .tence initial position is considered the normal=k\ SUbJec m senrogattvesentences the nominal I. ype. n mteraskedestabli h '.s es the theme. It IS thus the interrogative pronoun that begmst e sentence.is ah 'e:n!.e ement, about which the question isntic ::; ded as a seo?d word in the sentence; a further encliticLaw: --+ J ;'acke/" . .I::: :: ;;;word. ,.rhis IS known as the (Delbriick-)Wackernagelrnagel Ober em Gesetz der indogermanischen Wort-acr;gel Kleine Schriften I 1969 (in a contribution fromI. ' , 0 ge Laws 1985 p. 217-219; T. Krisch "B Delbrucks[;;t/theuliger Sichf' in Kolloquiu;" Delbruck.In relations of dependence there is a dominant component and a de dent component.The dominant component creates a syntacticalpenning fo r the dependent component, which the latter fills. Phrasal ele::t do not fill openings left by the finite verb are circumstants.O ts that fill the openings left by finite verbs are actants; phrasal elementsSuggestions fo r further reading: --+ Tesniere Syntaxe structurale 1959;Jlapp Grundfragen 1976; Pinkster Lat. Syntax 1990; Haudry Cas envedique 1977.I) Agreement and government have in conunon that they appear in dependencyrelations.This does not mean that components that participate inagreement and govermnent are thus equivalent to each other; rather, the one isindependent and the other dependent.In addition, the components are notequivalent to each other within the rea1ms of agreement and government: --+Matthews Syntax 1981 p. 249: "( ... ) that government and agreement are directionalrelations . " However, agreement and government are not equal intheir accordance with the dependency relation: --+ Matthews Syntax 1981 p.249: ''But the direction does not always match that of dependency."2) By 'agreement' is only meant the accordance of a declined fo rm of onegrammatical category with another: --+ Matthews Syntax 1981 p. 246:"Agreement (or concord) is usually described as a relation between words thatshare a morphosyntactic feature." However, the agreement may also takeplace between a lexical and a granunatical category, which is then evidenced ina syntactical incongruence, such as when a singular quantity is given as a subjectof a verb that is in a plural form. For example: eaesar Bellum Gallicum 2,6, 3 cum tanta multitudo lapides ac tela conicerent "when such a large group(of people) hurled stones and projectiles"Syntactical disagreement is thuspossible when, in place of syntactical agreement, agreement between a grammatical and a lexical category takes place. A similar relationship also existsbetween personal pronouns and the verbal category of person.3) Agreement is not dependent upon the fuet that an identical (morphologically)linguistic symbol is used to express the category of agreement. AsScherer puts it in his Lat. Syntax 1975 p. 97, "... it is characteristic ofancient rElanguages that morphological elements, which are equivalent to each other, arenot generally phonetically equivalent...... In a main clause, the category ofagreement may also remain fully unexpressed.S 210. Syntactical relations4) Various gradations among the various categories of agreement may beobserved: --+ Ch. Lehmann Kongruenz in Syntax 1993 I p. 725a: ''the essentialis that it is more probable that a category in a language functions in agree-

250 Proto-Indo-European Syntaxment, the more the category is grammaticized C···) Gender and nominal classonly appear in the context of agreement. But precisely they are the moregrammaticized variants of other processes of nominal classification, such as thepossessive or number classification, which themselves are not bound byagreement." In the case of internal agreement Ci.e. within the nominal syn_tagrna), both parts of the agreement have the same referent. In the case ofexternal agreement, the predicate is in agreement: Matthews Syntax 1981p. 250: "It appears that agreement fo llows the direction of dependency whenthe dependant is a modifier or a determiner C ... ) but is the opposite when it is acomplement, or at least the complement ofa predicator."5) Ch. Lehmann, to the origin 0 f affixes of agreement, asserts C"Kongruenz"in Syntax 1993 1 p. 729a): "Affixes of agreement uhirnately originatefrom pronouns C ... ) In internal agreement, demonstratives, and in externalagreement, pronouns are the basis C ... ) In a diachronic perspective, reference isthus the primary function of agreement."6) Government is a syntactical relation in which the subordinated compontentof a dependence relation fiIJs a semantic blank created by the dominantcomponent: Matthews Syntax 1981 p. 250: ''Government, which is traditionallyrecognized only in complement constructions, follows the direction ofdependency throughout." On the other hand, when the dominant componentfiIJs the semantic role of the subordinate component, there exists what is calleda relation of modification. Government and modification may thus be definedwith precision in relation to one anotber: If; in a sentence, an element r, issemantically relational and syntactically dominant, and, another element r, issemantically absolute and syntactically dependent, while filling the syntacticalblank left by r" then government occurs, with r, governing r,. If, in a sentence,an element m, is semantically relational and syntactically dependent, andanother element m, is semantically absolute and syntactically dominant, andeach of these fiIJs a blank left by the other, then modification occurs, with m,modifYing m,.verbal MorphosyntaxC. Verbal Morphosyntax\. General lnforrnationIn addition to its lexical meaning, the finite verb coists ofS 300 .tical categories, which are in turn composed of the followmg fivegr . number mode tense-aspect, and diathesIS. For m-dimenSIOns. person, , ,fo rmation on morphology, cf. F 200ff.Th tegories in themselves are three categories of number Csmular, an:; ;u ral) the five modes (indicative, injunctive, imperative, conJuncdand optativ), the three tense-aspects (present, aoris!, perfect), and thelive, .' )three diatheses C active, rruddle, passive .Suggestions for further reading: K. Hoffmann "Das Kategonen ?, stem des indogermanischen Verbums" in Hoffmann Aujsatze IT 1976 Cm acontribution from 1970) p. 523-540.2) Transitivity is not marked morphologically, but rather is expressedthrough the presence of an accusative morpheme m the comleent. Inhis "Zur Funktion des Nasalprasens Im Urmdogermams hen Cm FS R,x1993 p. 280-313), G. Meiser tries to sho . that the nasal infix of nasal presentswas the original expression oftranslltvlty.2. Person and NwnberS 301 . Within the dimension of person, PIE features three catgorieswhich are normally numbered following the example of grammarIans ofantiquity, a practice which should not be understood as a statemeṇt boutthe meanings of these categories. In the singular, the first person mdlcatesthe speaker; the second, the person to whom he speaks; and the third, thatabout which one speaks. Thus, the first person refers m ev,:ry case to ahuman being, or rather to an object that is thought of as ted :Thes econd person essentially refers to a being that is thought of as hstenmg, oran accordingly conceived object. The t pern, on the other hand, hasno natural tendency to indicate either hvmg bemgs, or objects, and canindicate the one just as well as the other. For more informallon, .cf. .F 400.The plural of the first or second person does not necessartly mdlcatethat there is more than one speaker, or people, to whom one speaks, butmay simply indicate that the speaker and listener represent groups. The..25 1

252Proto-lndo-European SyntaxVerbal Morphosyntax 253distinction between the inclusive first person plural (,we,' i.e. including thespeaker, his group, and the listener) and exclusive first person plural (,we 'I.e. the speaker and his group, without the inclusion of the listener) canntbe reconstructed in Proto-Indo-European. That which is true of the plural,also applies to the dual: separate subcategories for the inclusive dual ('wetwo,' i.e. the speaker including the listener) and the exclusive dual ('wetwo,' I.e. the speaker and his group without the listener) have not yet beenobserved.S 302. In the verbal as in the nominal number categories, Proto-Indo_European features a singular, a plural, and a dual.. In the case f t?e verb, number refers to the number of living beings ortgs that are mdlcated by the subject noun. The number plural does nottndlcte tha the verbal activity takes place repeatedly or over a longerduratIon; this IS expressed by the activity type of the verb: -7 DressierVerb. Pluratittit 1968.Number is the dimension in which, in general, agreement exists betweenverbal and nominal inflection, namely the agreement between the finiteverbal form of the predicate and nominal form of the nominative subject.The.verbal aspect of 'person' must, with regard to agreement, be consldredID another Iigh: Aside from its reflection in the verb form, 'personIS not a grammattcal category, but rather a lexical one one that isfirmly linked to the personal pronoun.3. Tense-Aspect and MoodS 303. The dimensions 'tense-aspect' and 'mood' are linked in theirfuncions .and appear together. While tense and aspect are representedWIthin a smgle morpheme and are thus connected with each other in termsof content, the catego of 'mood' is sometimes expressed using a propermOdal .morpheme, as ID the cases of the subjunctive and optative, and issometunes expressed through the use of different endings (indicative, inJuncttve, unperattve).Suggestions for further reading: -7 Mutzbauer Gr. Tempuslehre 1893-1909; Mutzbauer KOlifunktiv und Optativ 1908; P. Kiparsky "Tense andMood ID Indo-European Syntax" in FL 4, 30-57; Rix Modussystem 1986;K. Strunk "A propos de quelques categories marquees et non-marqueesdans la grammOlre du grec et de l'indo-europeen" in Colloque P. Chantraine1989 (1992) p. 29-42.'304. Using aspect, the speaker places the verbal action in a chronologIC: . . .S . at relati onship whereby he specifies whether the the verbal action ispleted (perfecttve aspect), or m course (unperfecttve aspect). When:grammar of a language includes this distinction, the language in questtOsion.H. Rix theorizes that an earlier phase of Proto-Indo-European featuredlarger number of categories within the aspect dimension, and otherwise. n 'IS considered an aspectual language. Aspect IS a grammattcal duneno longer distinguished between grammatical aspect and lexical aspect,rather uniting the two under the rubric 'aspect-action type': -7 H. Rix inL1V 1998 p. 10: "Aorist, present, perfect, as well as, departing more orless from the communis opinio, in causative-iterative, desiderative, intensive,fientive and essive." For further information, cf. F 206. For furtherinformation: -7 Rix Modussystem 1986.S 305. Unlike grammatical aspect, lexical aspect (Le. manner of action,or Aktionsart) is a property of the verbal meaning, and thus belongs to thelexical realm Lexical aspects are not uniform in terms of content: Theysometimes refer to the process of verbal activity, sometimes to the subject.Lexical aspects that refer to the process of verbal activity may be dividedinto telic and atelic lexical aspects, the fo rmer being those that only last amoment, the latter, those of more significant duration. Telic lexical aspectmay be the result of the verbal activity as a whole (momentative), of itsbeginning (initial-terminative), or of its end (final-terminative). Atelic lexicalaspect is a feature of verbal activities that last longer (durative), or arerepeated (iterative). Lexical aspects that refer to the subject may concerna desire of tbat subject (desiderative), or the bringing about of a state ofaffairs (factitive), or the cause of an event (causative). In terms of contents,lexical aspects are sometimes similar to grammatical categories, e.g.the desiderative lexical aspect and the optative mood, which overlap in thefirst person singular when the speaker and subject are one and the same.The factitive and the causative lexical aspects correspond to the activevoice. Because lexical aspect and aspect are sometimes similar, the lexicalaspect system can be carried over into the grammatical aspect system andvice versa, as may be observed in the evolution of Proto-Indo-Europeaninto the individual lE languages. Evidence of the change from lexical aspectto a grammatical aspect system may be even be observed in suppletiveverbal paradigms with stem forms of different verbal roots, e.g. 'carry,''bring': Lat. pres.Jero vs. perf. tuli; Gr. pres. $€pw vs. aor. ijveyKOV.

254 Proto.-indo.-EuTopean SyntaxS 306. Proto-Indo-European features three tense-aspect stems for expressingtense and aspect: aorist stem, present stem, and imperfect stem.The imperfect stem is formed from the present stem. The indicative formsof the tense stem only indicate the present (indicative present, perfect) andpast (indicative aorist, imperfect); future actions are expressed through thesubjunctive mood.In the post-Proto-Indo-European period, there are, aside from the languagesthat continue the use of the subjunctive, various other means ofexpressing future actions, for example the Latin -b- future ( LeumannLLFL 1977 p. 577-580; Meiser Laul- und Formenlehre 1998 p. 199£), c£Hitt. uyami ("to come") / paimi ("to go") + present ( E. Neu "Fulllr imHelhitischen?" in FS Strunk 1993 p. 195-202), ct: Ved. -Iar- (nomenagentis) + copula ( E. Tichy "Wozu brauchl das Allindische einperiphrastisches FUlur?" in ZDMG 142 1992 p. 334-342).S 307. In its task as an indicator of tense, the present stem reveals asimilarity to both the perfect stem and the aorist stem. The perfect, likethe present indicative, refers to the present tense, indicating a state of affairsto which the verbal action led, e.g. 'the goat has eaten,' i.e. 'the goatis sated.' On the other hand, the imperfect, which derives from the presentstem, has in common with the aorist a reference to the past tense: Theimperfect and the indicative aorist differ only in their stem forms and areotherwise formally identical. The indicative aorist cannot be used to indicatethe present tense, since the indicative aorist paradigms do not featureprimary endings that indicate the 'here and now' of the communicationprocess. This is due to the perfective aspect, which in the indicative excludesthe possibility of referring to the present and has more of a futuremeaning. The Proto-Indo-European aorist and perfect categories fusedinto the perfect in Latin.S 308. Examples: - a) present: Plautus Trinummus 400, aperiunluraedes ''the house is opened"; A 100 vocrtov oi..,at EAt"'OEa "You seekhoney-sweet homecoming"; RV 10, 107, 7 da/qit;lisva", da/qit;ii gli",dadiili ''the Dakit).a gives a steed, the Dakit).a gives a cow." - b) imperfect:Plautus Casina 178, nam ego ibam ad le ''for I came to you"; M 152aAa yap KpatEp

256 Prol()-lnd()-European Syntaxperfect pr se ? t stem is used to express 'inhibitive' interdictions. Preven_tlve mterdlctlons avoid an event from the outset; inhibitive interdictio?alt an .eve t that is in co se. The common term for these phrases . prohibltlve ; a useful termmological differentiation of the two types ofmterdlctlon 15 that between prohibitive and inhibitive interdictions.EXa mPles: a) prohibitive: V 3, 53, 2 ma parii gii/.l "don't go away" -b) inhibItIve: AV 10, 1,26 ma ti!ha/.l "don't stay (standing)."Suggestlons for further reading: Hoffinann !,yunktiv 1967; H. Ammann"Die ditesten Formen des Prohibilivsatzes im Griechischen undLateinischen" in IF 45 1927 p. 328-344.S 312. The perative, particularly the true, 2 "" person perative, holdsa special . plce m the verbal paradigm, similar to that in the nominal paradIgmoccupIed by the vocative, which is equally directed to a listener and.WIth which th imperative shares the formal characteristic of having singularform which 15 composed of the stem without an ending with no signof its connection to the sentence.Examples: - Plautus, Mostellaria, 387 habe bonum animum "havegood c ? age"; B,331 aU' aye !It!lvete !l(XV'tEC; "come now, stay"; RV I,16, 6 tam mdr sahase p iba "Drink this, oh Indra, fo r strength."For further informatIOn: B. Forssman "Der !mperaliv im urindogerman schen erbaisystem" in Fachtagung Berlin 1983 [1985] p. 181-197.For mformatIon on morphology, ct: F 212.In addition to the true .perative, which expresses a request or an orderthat de ds the lnunedlate execution of the verbal activity, another formof expressIon for mstructions and requests with a temporal function developedfrom the association of this perative form with the ablative of thedemonstratIve pronoun PIE 'tad. While these instructions and requestshave the medlate validIty of the true perative, they do not bring aboutthe munedlate executIon of the verbal activity.Interdictions, or negative orders or requests, are, unlike positive orders?D d equests, not expressed with the perative, but rather with the inJunctIve m connection with the negation PIE 'mehl• The use of the infinitIveto name the prohibited verbal activity in IE languages is comparable.S 313. According to Delbriick's investigations of fundamental notions(AI. Syntax 1888 p. 302), the SUbjunctive mood expresses a will, while theoptatIve mood expresses a wish. It is portant to note that the will or theWIsh (as the case may be) that is meant is that of the speaker, and not thatofthe subject, or, more precisely stated, that of the actor that is designated'•• \ Morphosyntax'l/erlJGby ttbrouge nb its own derivational verbal form, namely, the desiderative. Forh orninative form. The wish of the subject was originally expressed. . .da (Character of the Moods 1956), the characteTlStlc propertIes are? OO uaIization' for the subjunctive (p. 69ff.), and 'eventuality' for the optaVIS (p 52). According to A. Scherer ("Die ursprungliche Funktion des.t . b'Konjunktivs" in Fachtagung Be n 1969 [197] p. 99-106, the su unctIvedraWS the conclusion from a gIven Ituatlo (p. 101) :'The subjunctIvewould then indicate a state of affiurs, which accordmg to the reievantjacts, may be accepted factual (i . .e. concluded from the CITcumstances tobe necessarily true), while the mdlcatlve reports what the speaker kMWS( believes to know), or asserts as a fact. The optatIve characterIZesmT., IPmerely that the state of affairs was thought." (p. 101). he ongma rotolodo-Europeancategories 'optative' and 'subjunctive' fused in Latin.Suggestions for further reading: Delbriick Conjuncliv und Optativ1871' Hahn Subjunctive and Optative 1953; Gonda Indo-EuropeanMoods 1956; K. Strunk "Zur diachronischen Morphosyntax des Konjunktivs"in Kolloquium Kuhner Amsterdam 1986 [1988] p. 291-312. For information on morphology, cf. F 206f.1) 'The subjunctive, which originally indicates the future, has two functions:J 0 its prospective function, it serves to express things that happen in the future,while in its voluntative function, it indicates the will of the speaker. 'The subjunctive is used to express his will when he considers that it is within his powerto bring about the verbal action. A declaration of will in a strict sense is onlypossible when the speaker has direct influence on events, such that that whichis desired may also be executed. This means that a true expression of will mayonly be in the first person singular, while all other cases are equally requests . .Ifthe first person subjunctive is taken as a request made of oneself, a connectIonto the second and third person subjunctive is possible in which the speaker hasno direct influence on the reali7ation of the verbal action, so that the statementmay only be understood as a request. A further connection may be made withthe I" person plural, in which the speaker communicates his own will, and atthe same time directs a request to others.2) Examples: _ a) I sg.: Plautus Bacchides 1049 quod perdundumst properemperdere ''what may be lost, J wiIVwant to hurry up and lose"; u 296ID' iJ:ye Ot KOt EyriJ &00 eivtOv "thus I wi\Vwant to give a gift of welcomealso to him"; RV 10, 39, 5 purii vii". virya pra bravii jane "your earlierheroic deeds I wi\Vwant to announce to all people"; RV 6, 59, 1 pra nu vociisutesu vam ''On the occasion of the pressing, I thus wi\Vwant to announce theheric deeds of both of you." - b) 1 pI.: Q 601 vuv liE !lVllcrW!1ESO OOpltOU257

258 Prot-lndc>-European Syntaxvorbal Morphosynlax259"now we wiWwant to think about the meal"; RV 5, 51, 12 svastaye vQyUmupa braviimahai "We wiWwant to call to Viiyu for the sake of welfare." - c)2 "" person: Plautus, Mostellaria, 388 taeeas "you should remain silent"; RV 431, 3 abh! u 1;U1/.1 sakhiniim avita jaritftuim satam bhaviisi iitibhi/.l "you, ohhelper of the singer's fiiends, wiWshould protect us well with a hundredhelps." - d) 3 "' person: Plautus, Captivi 1 15 sed uti adserventur magna diligentia"but they should he guarded with great care"; H 197ou rap n fl£ f31u re EKcl>V aeKovra oll1at." For none wiWshould force meto leave against my own will"; H 87 Kal 1tO "tt

260 Proto-Indo-European Syntaxmiddle voice in the following way: "It is quite obvious how to distribute thetwo functions of the more recent middle voice among these two older voices:the content of the middle was the reflexive along with the passive, and thecontent of the stative was the deponent. ,,14 Rix emphasizes that the middlevoice is more related to the stative voice than to tbe perfect mood: ''It is thisvoice 'stative' and not the mode of action 'perfect,' that is the partner of thevoice middle." (p. 104).Examples: - RV 4, 21, 4 rayo brhato ya tse ''who has great wealth athis disposal"; RV 7, 101, 2 yo visvasya jagato deva tse ''the god, whichhas the entire world at his disposal."Suggestions for further reading: Neu Mediopassiv 1968; C. Watkinsin Selected Writings I 1994 (in a contribution from 1971 [1973)) p. 146-188; N. Oettinger "Der indogermanische Stativ" in MSS 34 1976 p. 109-149; by the same author, "Zur Funktion des indogermanischen Stativs" inFS Rix 347-361; Jasanoff Stative and Middle 1976; K. Strunk "Zum idg.Medium und konkurrierenden Kategorien" in FS Seiler 1980 p. 321-337;H. Rix "The Proto-Indo-European Middle: Content, Forms and Origin" inMSS 49 1988 p. 101-1 19; Kiimmel Stativ und Passivaorist 1996' T. Goto"Uberlegungen zum urindogermanischen 'Stativ " ' in Kolloquium DelbruckMadrid 1994 [1997] p. 165-192; R. Stempel "Staliv, Perfekt undMedium: Eine vergleichende Analyse fur das Indogermanische und Semilische"in GS Kurylowicz 1995 p. 517-528.4) The function of the category 'passive,' which appears in many lElanguages, but did not exist as a grammatical category in Proto-IndoEuropean, was performed by the middle voice. The various lE languagesthat feature a passive voice each formed it independently from each other.Suggestions for further reading: E. Schwyzer "Zum personlichen Agensbeim Passiv, besonders im Griechischen" in Schwyzer Kleine Schriften1983 (in a contribution from 1943) p. 3-79; H. Jankuhn Passive Bedeulung1969; Hettrich Agens 1990.5) According to I. Mel'cuk, there is a difference between 'diathesis' and'voice.''' According to his view, one speaks of 'diathesis' in cases inwhich forms of the same verb that are commonly said to differ in voicecannot be used to describe the same real situation. Such is the case ofp. 105.14 "The Proto-Indo-European Middle: Content, Forms and Origin" in MSS 49 1988J5 I. Mel'tuk "The inflectional category o/ voice: towards a more rigorous definition" in Causatives and Transitivity 1993 p. 1-46verbal Morphosyntax 261:'d, the possibility of referring to the same r al situation exists in the case'ddle voice forms and their corresponding active forms. On the otherof the passive forms that correspond to actIve forms. In this case, onespeaks of 'voice.'D. Nominal MorphosyntaxI. Nominal PropertiesS 400. The verb, with its system of categories, presents a contrast withall other inflectable parts of speech, which share a common system ofcategories. For this reason, one speaks of nominal categories whenspeaking not only of the noun, which includes substantives and adjectives,but also when speaking of pronouns. The commonalities that combinethese word types are the case and number categories.In the case of adjectives and gendered pronouns, the dimension of genderis not directly linked to the lexeme, cf F 323.The personal pronoun plays a special role among the pronouns andnouns, not just because it does not distinguish between gender, but alsobecause personal pronouns, unlike other pronouns, do not in fact take theplace of nouns, which is why it would be better to use the term 'personals.' Unlike the case of verbs, the dimension 'person' in personal pronounsis lexical. For information on morphology, cf. F 401.a) Case (S 401 - S 414)S 401. To each case may be attributed a certain meaning. To be certain,the meaning may vary from the central meaning in certain cases. Meaningsof cases vary as do lexical meanings, according to context. However, twoopposing meanings may not be unified in a single linguistic symbol. Themeaning of a case is generally independent of context, while the variousfunctions are determined by the context. H. Hettrich proposes a processfor the description of meaning and function of lE casual categories: "z Ur funklionalen Varialionsbreite altindogermanischer Kasus: Der Ablativim f!.gveda" in FS S/runk 1993 p. 53-55. In the wider context of a

262Prot-Ind-European S Ynlaxsentence, there are certain roles that may be assigned to the various n0rninal forms that appe !'t the sentence. These roles, however, are inde:pendent of the lingUlSlIc s ?W bol and concern the actual situation, whichmay be descnbed qUite vanously by the speaker. The same actual situat"ay thus be described in an active construction, or in a passive consttlon: The cat ate the mouse. - The mouse was eaten by the cat .Tn the 0case the nominat:ve fo 'cat' corresponds to the agens, in the other, tnommalIve form mouse c rresponds to the patiens. Agens and patiens:rre two opposmg roles, which may neither be assigned as different meanmgs? f a smgle linguistic symbol, nor classified as functions of a singlemeanmg. As roles, agens and patiens are separated from the linguisticsymbol of the nominative and may not be indicated by the nominative.Rather, the nominative indicates that which is in the foreground, thus, thetheme; whether the agens or the patiens provides the theme is unimportant.For further information: E. Tichy "Transponierte Rollen und Ergtinzungenbeim vedischen Kausativ" in FS Rix 1993 p. 436-459.S 402. The claim is often made that case meaning is least distinct in thecase .of complements, and most distinct in the case of extensions. Accordmgto w.. U. Dressl r: .' .. case forms are obligatory completions ofverbs ... subjects and objects are automatic results of the use of verbswhich, in their dependence schemes, present corresponding fiIIablspaces"," and further: " ... there remains the function of case in the facultativeextension of the sentence. Here, the case has syntactical value of itsown 17". H dry "A. , au : s a general rule, one may assert that government.tends to depnve the case of Its own semantic contents; a governed use isdefin? by a ction. :ositive semantic contents may only appear in free8.uses. , Pinkster: , ... the semanlIc relations within a sentence are revealedby the cases only to a very limited extent because: _ within thenuclear predication the predicate determines the ssibility of lexemes tooccur as arguments with the predicate; the number and nature of the semantic.functions .are .fixed for each verb; - outside the nuclear predicationthe lexIcal mearung Itself determmes to a high degree whether a lexememay be used with a given semantic function."" However, the claim mayI. " .17 .,uber de Rekonstrulttion der indogermanischen Syntax " in ZVS 85 1971 p. J Of.18.. Ober die Rekonstrulction der indogermanischen Syntax " in ZVS 85 1971 p. 12.Cas en w§dique 1977 p. 14" Lat. Syntax 1990 p. 47(o01inade with such comprehensive validity: H. Hettrich "Rektionbe ': utonomer Kasusgebrauch" in Kolloquium WacJrernagel Baselnot1 MorphosyntaxiJlr " 9:0 ] p. 82-99; by the same author, "Semantische und syntaktischeBe/1988lu111- 134. Because the nominative does not occur as an extenht ngen zum doppelten Akkusativ" in Fachtagung Zurich 1992[ 19 'tP 'meaning as a complement cannot be compared with that of an" on I S . d'' .The locative on the other hand, may always m Icate a spatiaextensIon.'1 t ' nship regardless of whether it is a complement or a gIven. aID ,Sug. ..gestions for further reading: Fraenkel Syntax der Llth. Kasus. . .M I, n"1928; E. Risch "Betrachtungen zur indogermanischen263. I, mmav,ex/O mR' h Kleine Schriften 1981 p. 730- 738; S. Luraghi Der semanllsche11:'; funktionelle Bau des althethitischen Kasussystems" in ZVS 99 1986p. 23-42.S 403.Although they have meanings that sometimes vary greatly, differentcases fit into a single paradigm: Thus, in terms of content, the nominativecase, when used to indicate a grammatical subject, is completelydifferent from the locative case when it is used to indicate the spatIal aspectsof the verbal action.I) The order in which cases are listed originates in Sanskrit grammar, inwhich the cases in the paradigm that were formally identical were groupedtogether in each of the three numbers. However, this formal criterion is not apurely external characteristic. This formal identity is also generally defensiblein relation to meaning, just as the partial formal fusion of various case formsmay be seen as a preliminary phase of case syncretism2) Proto-lndo-European cases may be classified into groups according toaspects of content: There are cases with rather abstract meaning, that crossreferencewithin the language system, and others that have rather concretemeaning, referring primarily to language-external reality. This differentiation isnot new, but must not be seen as an absolute classification, since individualcases are situated between the two poles, able to be used concretely or syntactically.Cases assume particular meanings in the establishment of spatial relationsof the verbal action: The spatial cases are the locative (where?), theaccusative (where ... to?), and the ablative (where ... from?). The noun thatindicates the place to which the verbal action refers is declined in one ofthese cases, allowing that which is signified by the subject (in the case ofintransitive verbs), or that which is signified by the object (in the case ftransitive verbs), to be spatially situated. That which is spatially situated IS

264Prolo-Indo-European Synlaxreferred to as the locatum; that which refers to the place of reference, islthe re atum.Suggestions for further reading: Fraenkel Syntax der Lith. Postposi_tionen und Priipositionen 1929; Starke Kasus und Adv. im A Hitt. 1977 .Luraghi Casi e preposizioni 1996.S 404. A common phenomenon of the linguistic development fromProto-Indo-European to the lE languages is case syncretism, which meansthat cases which were originally separate from each other and distinguish_able by their endings, were subsumed into a single ending. The spectrumof meaning 0: the rsulting case bec?mes correspondingly broad, renderingthe task of dlscerrung a basIC mearung of the case more difficult. For informationon morphology, cf. F 305 § 3.An intermediate stage in the fusion of one or more cases may be observedin which, according to number, or the context of gender or stemformation, the number of endings is smaller than the number of cases andthus, within one and the same paradigm, not all cases are formally durerentfrom ach ?ther in all numbers. The differentiation of cases is no longerparadigmatic, but rather syntactic. Thus, already in Proto-Indo-European,all eight of the cases may be differentiated only in the singular of nonneuterthematic nominal stems. The plural and dual of the same paradigmno longer perrmt the reconstruction of eight different case endings.In Latin, the ablative represents the merger of three cases: instrumental,ablative, and locative. In Greek, the PIE instrumental and locative casesmerged to form the dative, and the ablative was subsumed in the genitive.SuggestIons for further reading: Delbriick Synkretismlls 1907' H.Hettrich "Zum Kasussynkretismus im Mykenischen" in MSS 46 1985 p.111-122; M. Meier-Briigger "Zum VerhOllnis von Form und Funktiongrammatischer Kategorien" in Fachtagung Berlin 1983 [1985] p. 271-274.S 405. Nominative.The nominative occupies a special position within the nominal paradigmsof lE languages. This position is revealed by, among other things,the fact that in Old Indian -apart from neuter forms- all three numbers areformed on the basis of the strong stem and that tbe columnal nominal accentin Greek follows tbe accent position in the nominative case. Withinthe realm of syntax as well, tbe nominative traditionally plays a special roleas the casus rectus, which contrasts with all the other casus obliqui of theparadigm.'..,(JIl1lf\. aI Morphosynlaxminative indicates the theme of the sentence which, in a non-The no . . 'aI. .Oth- - , " ' 1 ..u_.265thi hd tence is placed in sentence-lfUtI poSItion. er sen encelernents are also thematized in taking the sentence-lfUtla poSition, w c , h non-marked sentence, is reserved for the subject.U1t.,';;'e Proto-Indo-European nominative .does not indicate the subject oft· 'm the logical sense but rather m the sense that appears to thean ac Ion ' . 'r to be bearer and middle-point of the actIon that IS expresse ythe ver .observehi dd bb " ( Delbriick Gr. Syntax 1879 p. 78). However, t S oes not. 'fi.h 'ddlI to the interrogative pronoun, which places ItS re erent m t e nn e-app even when it does not take the subject position and is not the bearer verbal action. The concept of the subject is itself difficult to grasp;or H.-J. Sasse it is, " ... a syntactical relation with semantic and pragmatIcfunctions ... [the] sentence element that is indicated as the subject has adoubled function as it is both pragmatic (as an indicator of the tOpIC of thesentence) and semantic (as an identifier of the agent). This double-. . . . . . ,,20function finds expressIon m ItS syntactIcal characterIstIcs.For further information: G. Serbat "Der Nominativ und seine Funklionals Subjektkosus im Lichte moderner Sprachtheorien" in G/otta 591981 p. 119-136.S 406.VocativeThe vocative is the nominal form that is used for addressing a listener.There is only a distinct vocative in the singular, and even then, not allnominal paradigms feature a separate vocative form. Where there IS novocative, its function is taken by the nominative. The same occurs whentwo actions of addressing are linked: While the first is in the vocative, thesecond is in the nominative. - Examples: r 276f. Zeii ltatep ... 'HEAlO,6' "Oh father Zeus and Helios"; RV 3, 25, 4 agna indras ca "Oh Agni andlndra"Suggestions for further reading: Svennung Anredeformen 1958;Zwolanek Anrufungsformen 1970.I) The vocative element in the sentence receives no accent. -Example:Rv I, 184, 2 asme u $U Vr$af}ii miidayethiim "Enjoy yourselves nicely, youtwo heroes, in our company."2) In Old Indian, when the vocative forms a sentence of its own, and is thusin sentence-initial position, it receives stress, regardless of its normal nominalaccent, on its first syllable, i.e. on the first syllable of the sentence. In this case,20 Subjektprominenz in FS Slimm 1982 p. 270

266Proto-Indo-European Syntaxntence stress is meant and not word stress. - Example: AV 19, 70, I devajivata "Gods! Live!"S 407.AccusativeThe accusative has two apparently very different functions: On the onehand, it indicates the direct object in the case of transitive verbs (i.e. accusativeobject), on the other hand, it expresses that the verbal action bearsan orientation in terms of space (i.e. directional accusative).The accusativeis further used to express spatial or chronological expanse (i.e. accusativeof expanse). In addition, it expresses the relation of the verbal actionto a referent in a non-spatial sense (relational accusative). Finally, theaccusative is also used when the contents of a verb are additionally expressedthrough a noun which appears in the accusative (i.e. accusative ofcontents): The technical term for this use of a substantive and a verb withthe same lexical contents is figura etym% gica. The original meaning ofthe accusative is probably that of direction, in the sense of spatial relation.The additional meanings that developed upon this basis include extent,relation, object and contents.I) According to Hiibschmann (Casus/ehre 1875), the accusative indicatesthe "completion or narrower definition of the verbal concept" (p. 133), distinguishingan obligatory accusative, i.e. the object accusative, from a fucultativeaccusative. Delbriick (Gr. Syntax 1879) thus describes the use of the accusative:"Originally, it served neither to indicate the object, nor the destination,nor the relation, etc., but rather simply to complement the verb. The choice ofsenses in which this complement was to be understood was left to the listener."(p. 29). He further asserts that there were "different types of uses ...already in the Proto-Indo-European period" (p. 29).2) As an indicator of place, the accusative is similar to the locative, which isalso used to indicate the arrival at a destination toward which a movement wasoriented. In contrast, ahhough the accusative does not exclude the arrival at adestination, it is semantically indifferent to the question of arrival at a destination: J. L. Garcia Ram6n "Zum Akkusativ der Richtung im Vedischen undim lndogermanischen" in FS Strunk 1995 p. 33-52.It remains disputed whether the local or grammatical meaning of theaccusative is original. According to G. De Boel ("The Homeric accusativeof limit of motion revisited" in Kolloquium Kuhner Amsterdam 1986(1988] p. 53-65), the directional accusative is not inherited, but rathernewly created:. al Mor phosyntaxNQI11iOsatlV"With causative motion verbs, first of all, the accumulation of acu. eS is clearly caused by the addition of a secondary goal accusallveb' ect accusative that was already present in the construction.to "."ar/ the accusative with intransitive motion verbs carmot be shownSimiltl t 0 'ginal use It is restricted to a lexical subclass, m which suchto re ec n ' . I h -of the accusative seems bkely to develop spontaneous y, as ap:d again in Modem Greek." (p. 64f.)._.. ,-c. ormation· De Boel Goal accusative 1988.For lID".er UIl'•3) Equally unclear is the relationship in Proto-lndo- uropean of a specializeddirectional case, the 'directive,' .which was contmu d m Anatoli::0 t " in Fachtagung Zurich 1992 (1994] p. 17-36), the directive only mdlhe directional accusative. Accordmg to G. Dunkel ( The lE DỊrec:::es the direction: "It expressed only the .aim or direction of a mov :i meaningS: the accusative indicates "attainment of the goal and en.t" (p. 34). In comparison, the accusallve and the locallve have addl't" (p 34)' and the locative "attamment of the goal ... and ... stattermg 1 . •of rest" (p. 34).S.'. "s h '267c. further readm ' g' W P SchmJd prac WlssenuggestlOnslor .:, .' . .." .schaftliche Bemerkungen zum hethitischen DlrekfIV m FS Ollen 1973p. 291-30 I; Neu Lokativ 1980.4) Only miscellaneous remnants of the accusative of direction without. .d "(to) home " rusthe usc of a preposition are extant m Latm, e.g. omum ,''to the countryside.". " ' "Examples: _ a) Accusative of direcllon: A 322 EpXEaeO lmllV goboth of you to your tent"; K 195 OOOl KEKA1\a10 .IloullV who weresummoned for consultation"; TS 6, 2, 11, 4 yadii mukha", gachaty "athodaram gachati "if it goes to the mouth, then it goes to the stomach. _ b) accsative of extent: Plautus, Trucu/entus 278 nocte "!i str ?m,ntlspemoctare "to pass one night in tbe straw"; 'l' 529 ooupo EPOYllV at aspear throw's distance"; A. 190 XE1a "in the winter"; TB 1, 3, 6,3 saptcldaSapravyiidhtin aj i", dhavanti "they run a race for a distance of seven,.teen times the range of one shot"; TB I, 1, 3, 9 so asvattM sa",vatsaramati$that "be remained in the tree for one year." - c) accusal1ve ofrelatlon.Platus, Menaechmi 511 f. indutum ... pal/am "clothed in a dres .s"; E 35 ,Aat VE10 oe Xpoa KaMv "and she was redd .ened o her beautiful skm ,SB 14, 7, 2, 27 naina", krtakrte tapata/.! "neither things done, or.tbin?undone hurt this one." - d) object accusative: SB 14, 7, I, 24 Jlghran v betcid ghratavya", na jighra/i ''truly smelling, he smells not what IS to

268 Proto-lndo-European SYntaxsmelled." - e) accusative of content: Plautus, Caplivi 358 quod bo .bene fil beneficium "which charitable act is well directed to the goo::0414 iiUoI 0' acp' ii),,),,1Jm aXT]v Ea-xov'o vEecrmv "here and tberthey fo ught the fight for the ships"; RV 8, 7, 4 ytid yamalJl yanli vayubhih ."when they go the way with the winds."5) Suggestions for further reading: La Roche A ccusaliv 1861; GaedICkeAccusallV lm Veda 1880; Miiller Nominaliv und Akkusaliv 1908;.Jacqumod Double accusalif en grec 1989; H. Hettrich "Semanlische undsynlakllsche Belrachlungen zum doppellen Akkusaliv" in Fachlagun gZiirich 1992 [1994] p. 111-134;S 408. InstrumentalThe instrumental case indicates that which accompanies the verbal activity.This meaning forms the basis from which other meanings have developed:In the case of inanimate objects, the instrumental indicates them by which the verbal action is executed; in the case of a person, itmdlcates that the person executes, or helps to execute the action' in thecase of places, it indicates where movement takes place. The instrental:urther indicates constitution, accompanying circumstances, a reason, andm comparISons, the distinguishing characteristic. The function of the instrumentalthat relates to people, or 'sociative' function may be reconstructedm L te Proto-Indo-European. However, this function presumablyfinds Its ongms m a use that is purely related to inanimate objects: K.Strunk " Synlaktlsche Bemerkungen zum helhitischen und indogermanischn Inslrumenlal In ISloriceskaja lingvislika i tipologija" edited by G. A.Klirnov et al. Moscow 1993 p. 81-91. In the indication of temporal circurnst ces, te instrumental bears a resemblance to the temporal locative.In Latm, the trumental, like the locative, has merged into the ablative.In Greek, the mstrumental has merged with the dative, c( S 404.Examples: - a) Instrumental of accompaniment: Plautus, Amphitruo219 poslquam Ulrimque exitum eSI maxuma copia "after they marched up great umbers on both sides"; )" 160( Ev9ao' lKavel vT]i e Kat, .eW Olm 'you !'rnve here WIth the ship and the companions"; RV I, I, 5,devo de ,ebhlr a ga L "'.al 'the god should come here with the gods" RV 5,L51, I V1SValr u "'"eblr a gahi "come here with all helpers"; RV I, 92, 7dlVli stav duhlla gotamebhiiJ "the daughter of the heavens is prized by the,Gotamas. - b) Instrumental of me8flS: Plautus, Truculenlus 526f. nequeel/Om queo / pedibus mea sponle ambulare "and I carmot even walkarotmd mdependently on my own feet"; Lucretius 4 387 vehimur navi"we sail with the ship"; A 527 Kecpa),,1j Kaavew(j) ':1 will nod with myN()!I1inaJ Morph osynlax 269head"; M 207 nEO 1tVOl1j aVEoLO "he flew with a breath of the wind";RV I, 128, 3 sala". caanO aabhi/.! "the god that sees with a htmdredeyes"; RV 3, 32, 14 naveva yanlam "as to those who go with the ship." cl Instrumental of route: Plautus, Curculio, 35 nemo ire quemquam publicaprohibet via "no one hinders another from walking on a public street";Plautus, Poenulus, 1105 lerra marique "on earth and sea"; RV I, 25, 7anlliriena palalam "which fly in the air"; RV 3, 58, 5 eha yalam palhibhirdevayanai/.! "comes this way on divine paths"; RV 5, 64, 3 mitrasyayayam palM "I would walk on Mitra's path." - d) Instrumental of constitution:Cato, De agricullura 88, I amphoram deJracto colla " an amphorawith a broken neck"; PY Ta 64 1.1 li-ri-po e-me po-de i.e. Iriposheme pode "a tripod with one leg"; RV 4, 7, 3 dyam iva slibhi/.! "like theheavens with the stars." - e) Instrumental of accompanying circumstances:A 555 el T]6n 9w

270Prot-Ind-European Syntax229 quoniam vox mihi prope hic sonal? "what voice thus sounds for meso near?"; Plautus, Rudens 274 nunc libi ampleclimur genua "now Wshall seize your knees"; Plautus, Truculenlus 378 mihi quidem alque ocu lis meis "indeed fo r me and my eyes"; H 423 Ot o· T]"vt£ov aMT]' Aolan t ey met oe another"; H 101 tq;o£ 0' Eyrbv a{ltC'> 8!llpi]0Ilal "and" d hfor this one I will arm myself'; A 4 airtoix; liE EMl>pW tEUx£ KUV£cr",v"and he gave them to the dogs as prey"; E 249f. IlT]OE 1l0l OUt'8uv£ "do not rage so to me"; B 142 tol", liE 8uIlOV EVt crtiJ8£crcrtv OPlV£:'and he stirred the soul in their chests"; '¥ 595 Oatll0crtv dVal aAl'rp'to be a smoer to the gods"; RV 4, 12, 3 dOdhiili rtilna". vidhale ...mtirlyiiya "he distributed wealth to the devoted mortal'" RV I 15 12devlin devayalli yaja "sacrifice to the gods for the worshlpper ;f g;ds'"RV 2, 2, 8 tililhis carur iiytive "a dear guest for the sohn of Ayu." _ b)dallvus finlIs: Plautus, Poenulus 626 ul quaeslui habeanl male loquimellonbus 'that they have It as a gam, that they speak badly of their betters";H 285 XaPIlU llpOKaAtcrcrato "he called out to battle"; RV I, 30 6urdhvtis li$!hii na ultiye "be there upright to support us."For further irtformation: Havers Kasussyntax 191 1; Oertel Dativi finales1941.S 410. AblativeThe ablative expresses the place of origin of the verbal action. Accordingly,the ablative is principally featured when a locatum moves, or ismov ,: d, away .from a relalum. To this central meaning may be traced theablalIve functIons elating to origin, which refers to a spatial idea, relatingto separalIon, .whl h is accompanied by a movement away, relating tocomparisons, ID which the ablative is used to indicate the object in relationto. which a compared object differs. r n Greek, the ablative was subsumedWIthin the genitive, cr. S 404.Examples: - a) Ablative of place of origin: Cato, De agricultura 5pnmus cubllu surgal "he gets up out of bed first"; Plautus, TrinummusO cunclo7 exturba aedibus "drive all from the house"; 0 655 v£ciiv IlEV£!llpT]a 'the ;treated from the ships"; E 456 OUK av OT] tovo' avopaIlax:'l ,£paLO,c?uld you not push this man from the fight?; RV 7, 18,10 lyur gavo na yavasad tigopii/:z ''they went like cows from the fieldWIthout a herdsman"; RV 7, 5, 6 /Vti". dtisyiimr 6kasa agna iija/:z "you, ohAgm, drIve the Dasyus from their homeland." - b) Ablalivus originis:Plaut, CapllvI 277,quo < de ge

272Proto-indo-European S}Tltax'I' 485 fl pl1t? ?

274 Proto-Indo-European SyntaxNominal Morphosyntax275adpositionadverbgovernment + -modification + +The adverbs in lE languages that correspond to adpositions are positionedfollowing their referents: Benfey Vedica et Linguistica, Strasbourg/ London 1880 p. 101-114.Suggestions fo r further reading: Starke Kasus und Adv. im A Hill.1977; O. Dunkel "Preverb repetition" in MSS 38 1979 p. 41-82; by thesame author, "Die Grammatik der Partikeln" in Fachtagung Leiden 1987[1992] p. 153-177; Horrocks, Space and Time 1981; Ch. Lehmann "LatinPreverbs and Cases" in Latin Linguistics and Linguistic Theory. Proceedingsof the 1" International Colloquium on Latin linguistics 1981 inAmsterdam, edited by H. Pinkster. Amsterdam / Philadelphia 1983 p. 145-161; Krisch Konstruktionsmuster 1984; J. Boley "Hillile and IndoEuropean Place Word Syntax " in Sprache 31 1985 p. 229-241 .; by thesame author, Sentence Particles 1989; H. Hettrich "Syntax und Wortartender Lokalparlikeln des lJgveda. J: Mh!"' in MSS 52 1991 p. 27-76; O.-J.Pinault "Le probleme du pr

276Proto-lndo-European Syntax---t K. H. Schmidt "Probleme der Ergativkonstruktion" in MSS 36 1977 p.97-116; F. Villar "Ergativity and animate/inanimate gender in lndoEuropean" ZVS 97 1984 p. 167-196.V. The Proto-Indo-European LexiconA. General InformationW 100. The most important element of a language is the word, in whichare represented phonetic, phonological, morphological, syntactic, lexical,and stylistic aspects.All words in a language compose the lexicon of that language.Eachparticipant in a community of speakers acquires a more or less large partof the lexicon, and is capable of using these words in a manner consistentwith the common practice of his time.more consequential than a mistaken inflexion.renders comprehension impossible.A mistaken use of a word stem isThe former immediatelyAs a rule, most words are acquired by the child from his immediatefamily, whose members, in turn, have acquired words from their parents.Because ofthis tradition, most words may be traced hack for generations.According to conventional thinking, the lexicon is presumably organizedin the brain according to themes, model phrases, and chains of association.While organization of entries according to letters of the alphabetdoes not correspond to linguistic reality, this concept is practical and effective.A lexicon is not a static quantity. It is, like human life itself, subjectedto ongoing fluctuations. Tt is dependent upon consensus in the community?f speakers. The lexicon is always open to the expression of new meanmgs.On the one hand, new words may be formed analogically, based uponexisting models and elements, cp. for example the Gr. model a()'tAeu

278 The Proto-Indo-European LexiCOnand Lat. testa 'earthenware, pot, jug, shard,' becoming 'head' in the Romancelanguages. In place of the elegant caput, testa was later used toindicate the skull, eventually developing into the standard term for head.The Latin examples are taken from H. Rix: -) Gentilnamensystem 1972 p.714.On the other hand, foreign words can at any time be incorporated intothe lexicon. As soon as the orthography, pronunciation and inflection areadapted to the receiving language, one considers the word a loan-word: -)Bussmann Lexikon der Sprachw. 1990 p. 253. In Latin, fo r example, theloan word 'ampora 'two-handled clay container with a narrow neck witha poted bottom' < Gc.. acc. sg. aJ.lpEa (the Lat. -a is resumably bestexplamed If we assume mstead of -ea a spoken form amp orea with consonantale; the accusative form serves here for the speakers as a basis becauseit was in daily use for inventories and recipes, etc.). The phonemeph, uncommon in Latin, was replaced by p and the word integrated into thetypical Latin -a- declension). Proof of the complete integration of the loanword in the language is the purely Latin-internal diminutive ampulla'ointment flask, bottle.' Finally, the basis word 'ampora was recognizedagain as a fo reign term in the Classical Latin form amphora, with ph replacingp, this time following the Greek. On an other example see above E507 § 3 (machina).Each word was re-created at a certain time with a certain meaning.Once created, each word has its own history: It may find daily use; it maybe limited to a certain style or level of language; it may one day fall intodisuse and disappear from the lexicon, etc. Etymology and word historymay not be separated from each other: -) Seebold Etymologie 1981 p. 58.The explicit goal of etymology is to determine the 'true' first meaning(motivation) of a word, and to thus gain information about the 'true' natureof that which it indicates. But this is naIvely optimistic. Language isalways arbitrary and, as such, information about 'true' nature cannot thusbe attained, but rather, in the best case, one may find information about themotivations of the speakers in naming the object, etc., as they did and nototherwise: -) Rix Termini der Unfreiheit 1994 p. 9f with suggestions forfurther reading in note 21.In most cases, one carmot trace a word all the way back to its creation,cf. e.g. PIE 'K(u)ljan 'dog,' which, owing to correspondence sets in the lElanguages (cf. E 507 § 5), may with certainty be traced to Proto-IndoEuropean. What we can no longer determine, is how the 'dog' receivedits name in pre-Proto-Indo-European. E. P. Hamp proposes a relationshipto PIE *pelfu- 'livestock,' deriving 'liIjon- from a basic form *pelilj-on-General Information.responsible for herds' (evidence to the contrary: -) Mayrhofer Laut/ehre< , h ' hi h1986 p 118). On the other hand, the word lor toot , w c , agam ow-.tOg Ot IE' correspondence sets, may be traced to PIE *h,d-ont-. However,' hi dth verbal root PIE '",ed- 'to bite, to eat' allows us m t s case to un ered pIE 'h,d-ont- as a participial derivation, in the sense of 'biter.'stan" f IEven though the age and/or motIvatIon 0 a term are no onger .ascertainable,the possibility remains, within the limits of documentatIon, oftracing back the Iifespan of a concerned word m IE languages.Suggestions for further reading: - a) General: -) Wo rterbucher 1-31989-1991. - b) Specific to Proto-Indo-European: -) Pokorny IEW1959; C. Watkins, et aL, Indo-European roots m The American HeritageDictionary of the English Language, edited by W. Morris. Boston, et aL1969 [1980] p. 1505-1550; R. S. P. Beekes "Een nieuw lndo-Europeesetymologisch woordenboeK' in MKNA (= Mededelingen van de KoninkIijkeNederlandse Akndemie van Wetenschappen, afdelmg Letter!cunde) 61 / 9 1998 (see also the presentation of the project TITUS /Frankfurt in Actualia / Projekte / Leiden: There is a link to the Frankfurtpage on our web site [cr. E 100] s.v. 'Indo-European .linguistics inEurope'). - c) Individual perspectives on the reconstructIon of the PIElexicon: -) Stud. z. idg. Wortschatz 1987; R. Wachter "Wortschatzrekonstruktionauf der Basis van Ersatzbildungen" in Fachtagung Innsbruck1996 [1998] p. 199-207. - d) Latin: -) Leurnann / Hofrnann / SzantyrAUg. Teil 1965 p. 74'ff.; Szemerenyi Lat. Wortschatz 1989. - e) Greek:-) Meier-BrOgger Gr. Sprachw. II 1992 p. 7ff.W 101. While synchronic semantic research varies among the individualIE languages according to the availability of documentation, the reconstructionof Pro to-In do-European word semantics is particularly limited.Assertions may however be made concerning groups of words, for example,names of familial relations, body parts, natural elements (fire, water,light, etc.), home, and family, etc., cf. E 512 § 4, which includes suggestionsfor further reading.W 102, Etymology and word history have always been the fo cus of greatinterest. All single-language etymological dictionaries offer information onProto-lndo-European, including statements to the effect that a term is 'inherited'or ex. gr. 'pre-Greek' which are termed 'etymology,' althoughstrictly speaking only statements concerning the concret crealton of a wordhave etymological value.".279

280The Proto-lndo-European Lexiconword Formation281Suggestions for further reading: A. Bammesberger "Geschichte deretymologischen Forschung seit dem Beginn des 19. Jahrhunderts" inSprachgeschichte I 1998 p. 755-786. - Concerning the individual IElanguages, see (among many others): Etymologisches Worterbuch 1983(including contributions on various IE languages); Walde / Hofinann LEW1965; Emout / Meillet DELL 1959; Frisk GEW 1960-1972; ChantraineDELG 1968-1980; Mayrhofer EW Aia; Tischler HEG; CHD; Kluge / Seebold1995; Seebold Etymologie 1981; Vries AnordEW 1962; EWAhd;Vasmer REW 1953-1958; Fraenkel Lit. etym. Worterbuch 1962-1965 .Demiraj A1ban. Etymologien 1997; etc. - The 'checklist' by K. Hoffinaru:and E :Tichy regarding the assertion, or appraisal of etymological interpretations: Hoffinann Aujsdtze III 1992 (first published, 1980) p. 761ff.W 103. Wbether or not the Proto-Indo-European lexicon contained foreignor lo words is unknown, but, as in the case of any living language,qUlt pOSSIble. If yes, it remains to be known from what language the borroWIngsmIght have taken place: The Finno-Hungarians (cf E 436) andHanuto-Senutes (cf E 437) have been suggested as possibilities.B. Word FormationI. General InfonnationW 200. As a rule, the creation of new words only takes place accordingto odeJs. Normally, the speaker takes existing material from his ownlexI on as a model from which to abstract the basis for rules of word fo rmation, cf. the example from Greek, apxeoo in W 100..An unlimited number of relationships can be created and re-establishedm word formatIon by analogy, which is not confined to the real historicaluffixes and entire conglomerates of suffixes are created, cf. ex. gr. La!..s _ um _Ion . Starting from a regular series of examples, e.g. audire 'to listen'. ,dilor 'listener' d . .>au> auditorium 'room for listeners a !fect connectIOn. ,was later drawn from audire to audilorium 'a place Lor " li sterung ( ll d_ c e asuffix shift). Later, based upon this example, the word dormlfonum,'room for sleeping,' was formed from dormire 'to sleep,' regardless of thefact that the requisite step dormitor was in this case uncommon: Leumann/ Hofinann / Szantyr AUg. Teil 1965 p. 72*. - Suffix conglomeratessuch as *-ih, and *-ehr have been shown to date from the Proto-IndoEuropean period, cf W 204 § 1. For an other example see above F 507with note 7.W 201. Word formation of the ancient lE languages has been quite thoroughlyresearched. Following are fundamental recommendations for allareas of word formation: - a) Latin: M. Leumann "Gruppierung undFunktionen der Wortbildungssuffixe des Lateins" in Leumann KleineSchriften 1959 (in an essay dated 1944) p. 84-107; Leumann LLFL 1977p. 273ff. (nominal stem formation), p. 383ff. (norninal composition). - b)Greek: Debrurmer Gr. Wortbi/dung 1917; Chantraine Forma/ion desnoms 1933; Schwyzer Gr. Gr. I 1939 p. 415-544 (including introductorymaterial and information on root nouns, composite nouns, and nominalsuffixes); Risch Wor/bildung 1974. - c) Vedic: J. Wackemagel Einlei/ungzur Wortlehre, Nominalkomposilion, A. Debrunner "Die Nominalsuffixe"in Wackernagel / Debrunner Ai. Gramm. II / 1 1957 and II /21954. - d) Hittite: Rieken Nom. S/ammbildung 1999. Further bibliographicalsuggestions may be found at appropriate places in the text.In the following section, individual suffixes and compounds shall bepresented that can be traced to Proto-Indo-European. With this purposein mind, I shall permit myself to follow the overview offered in MeierBriigger Gr. Sprachw. II 1992 p. 20ff. and shall limit myself to referring tothe Proto-Indo-European core of suffixes and types of composite nouns.developme ts . .Lat substance adjectives of the type -no- present an exampleof this, mcludmg lIignus 'made of oak' and aenus 'made of bronze'« *aies-no-, cf L 215 § 1): A substance adjective that would correspondto terra 'earth,' should thus be *terrii-no- 'made of earth.' The attestedform terr-enus 'made of earth' shows that the historically inaccurate reanalysISof aenus as a-enus was so common as to make possible the transnusslonof the new suffix -enus. - Through such transformations, new2. Word Fonnation Using Suffixes; Suffix SystemsW 202. Adjectives that are derived from substantives:I) Concerning possessive adjectives, in the sense of 'belonging to,' or'in relation to': - The suffix that is most widespread among the IE languagesis *-io-.

282The Proto-Indo-European LexiCOnWord Formation283I a) However, this reconstruction, common in older treatments, is not sutlicientlyprecise. The fo llowing three endings are theoretically conceivable: I)*-jo-; IT) *-i(j)-o-, Ill) -iH-o-. To all appearances, the forms vary in their individualuses: Peters Laryngale 1980 p. 131 in note 79; Mayrhofer Laut/ehre1986 p. 161 with note 267; A. Hardarson in F.S'Rix 1993 p. 164 with note 25.These three authors consider at least fo ur different suffixes, namely, the verbal*-jo-, a nominal *-jo- with a vague meaning (according to its origins, presumablyan -i-stem with a thematic vowel); a nominal *-ih,-o- with a specialindication of possession (by its origin, possibly an -o-derivation from an abstractnoun featuring *-ih,-: G. Klingenschrnitt in Fachlagung Regensburg1973 [1975] p. 154 in note 10; or, what 1 consider more likely, an-o-derivation from -ih,- constructions of the type ofVed. vrkt: Rubio OrecillaSufljo de derivacion nominal 1995 p. 316f.); a nominal locative *-i(O-o(by its origin an -o-derivation from the nominal locative singular forms in -i, cf.the type Ved. damiya- 'domestic'). - The precise determination of the detailsis made more difficult by Proto-lndo-European variant fo rms resulting fromthe prior syllabic structure in the case of the fo rms of type 1 (for information onthe change from *-jo- to *-ijo-, cf. L 218 § I on Sievers' Law), by PIE vanishingof H, which, when occuring in type Ill, leads to merger with type IT, andfinally by generalization in the IE languages of *-ijo- in type I, such that, afterthe varushing of the laryngeal under *-ijo-, both true *-ijo- forms and thosethat were ori y *-jo- and *-ihzO- forms can he united. - The complexityf the sItuation IS revealed by the example PIE * ph,-Ir-ijo- 'helonging to theph2tr-, located at the *ph2Ier-, corning from the *ph2Ier-' = Lat. palrius =Gr. 1tmpto = Ved. pilriya-, etc. If one assumes the presence of an originalsuffix -jo- (evidence in Italic suggests this basic form for the group of patronyrns: Rix Genlilnamensyslem 1972 p 718f. with note 60), one should thenet to find *ph21{-jo-; the -Ir- (and not -tr-) double consonance is only conceIvable,if It was re-formed according to the model of the type gen. sg.*pIl2Ir-es already in the Proto-Indo-European era; the variant *-iio- (in placeof *-jo-) would then simply he the phonetic consequence of -the doubleconsonant-Ir-. The form *-ijo-, suggested in IE 1anguages, may just as wellhe traced back to *-ill,-o-; further, it should not be excluded that both locative*-i(O-o-, as well as *-jo- and *-ih,-o- derivatives of*ph21r- first coexisted, andthen merged. - See also the issue of Lat. dius = Myc. Gr. di-wi·/o I di-u-joi.e. di!!ijon (secondarily also inner-Myc. di!!jon) 'belonging to (in the realm of)*djey- " 'shrine to Zeus' and di-wi-ja I di-u-ja i.e. di!!ijii (secondarily innerMyc. di!!jii) female divinity 'belonging to *djey-,' 'daughter of *djey-' = Ved.div(i)ya- (RV+) 'heavenly, divine.' The substantive Horn. &a = Ved. devI isseparate from the adjective. Suggestions for further reading: Risch Kleine1981 p. 580f.; Mayrhorer EWAia I p. 727 (which justly emphasSchriflen . sthat the Vedic form devi cannot share a heritage WIth Gr. &a); Hardarson IDFS J/jx 1993 p. 164-166. Particularly concerning Mycenaean: Aura JOTTODMic I 1985 p. 178fI; A. Leukart in Mykenaika 1992 p. 394 note 44. Forinformation on the special group ofpatronyrns, cf. W 302 § 3.Ib) l. Balles proposes a new view of the entire issue: Sprache 39 1997[2000] p. 141-167. The contribuion is based upon thesis paper fro Vienna.which the author submrtted ID 1996 under the directIon of J. Schindler.Following are the results in tabular form:A) -o-ab!. of AoOOCIOV < *16!!ki-o--i- stemwp6 (Myc. lu-rjo)< *tuHrj-6-III-io- -i(j) o-cf. Aves!. tuiri- n.'whey'B) -o-ab!. of a) relational * medhj-o- > IlEcr( cr) mediusloc. -i adjective *a/i-o- > iill.o aliusb) composl- E-aA.-tOstional suffixc) denomi- 9aA.6crcr-lnative ad- *sokWh,-jo- > *sokjo- > soc-iusectivelerl-iusd) verbal ii1tElPO < *1J-per-jo- iiy-tO, infer-ius,adjectiveexim-iusVed. ajur-ya- 12x (-ia- Ved. mGd-ia-Ix) 12xJ:!ta- 1&Commentary: - The author postulates (ibid., p.161f.) that the commonstarting point for the entire group are the -0- derivatives from -i locatives(B), which must he distinguished from the less numerous -0- derivativesof -i- stems (A), cf. AooocrOV (Theophrastus) 'white core of pine'

284 The Proto-indo-European LexiCOnsons, as well as for reasons of clarity (the hiatus filled by the glide j to the-ijo-form). The group Bd however represents a special case in which, inthe case of Vedic verbal adjectives, the form -ijo- in longer word Constructions,is shortened to -jo-.2) A further possibility for marking possession through construction isattested by PIE 'd-e-jy-a- 'belonging to the heavens' = OL (Duenos inscription)deivo- and Class.Lat. deus I divus (cf. L 217 § 3) = Ved. dew i« 'dajya-) 'divine; god' = OAv. daeuua- 'demon' = Lith. dievas, etc.: Mayrhofer EW Aia I p. 742f. The PIE *d-e-jy-a- is considered a Vrddhiderivativefrom PIE 'djey- I *diy- 'god of the day and/or sky' and hasbeen described as an -0- derivative with additional insertion of -eo: Darrns Vrddhi 1978 p. 376/f - In just the same way, • n-e-y-o- 'now;new, young' may be interpreted as a derivative of the temporal adverb PIE'nu 'now.' - Further sources of information on this phenomenon whichis common and has been systematized in lndo-Iranian include: WackernagelI Debrunner Ai. Grammalik IT I 2 1954 § 34/f W. Schulze wrotethe classic work on this form of derivative: Schulze Kleine Schriften1966 p. 60ff. (in an essay from 1907). Schulze considers the example PIE"syefura- (in which • syeli- < "s!!-e-eli-) meaning 'who belongs to the father-in-Iaw,'i.e. 'son of the father-in-law' with the Proto-Germanic descendant'S!!egura- (cf. NHG Schwager) as a derivative from PIE• s!!efuro- 'father-in-law' (for information on PIE • suefur-h,- 'mother-inlaw,'cf. L 217 § 4) with the Proto-Germanic desc;nd;t "s!,exuro- (cf.NHG Schwaher). For information on phonetics, cf. L 42 1 § I, and further,cf. Kluge I Seebold 1995 p. 657 s. v. 'Schwager' and 'Schwaher.' _Cf. also L 33 1 §3 (which concerns Luw. si!Jyal).3) Adjectives indicating material composition: - PIE "-ej-o-, cf. Lat.aur-eus 'golden,' Class. Gr. apyup-oUc; 'made of silver' (the contractedform may be traced to -eo- < -ejo- [Myc.]; the Homeric accentuation ofthe t nom. sg. apyupEoC; vs. dat. sg. apyupeq> was secondarily formedaccordmg to the schema iivElpCll1toC; vs.avElpomq>; the older pattern"apyupeo is indirectly attested by the Classici Attic Greekcontracted fo rms), Ved. hirany-aya-. - Suggestions for further reading: l. Hajnal "Die jriihgriechische Flexion der SIoffadjeklive und derenere ble Grundlagen" in Fachlagung Zurich 1992 [1994] p. 77-109 (Follow,?g the lead of A. Heubeck, the author postulates, using MycenaeanrelatIons to the suffix forms -e-(j)o and -(i-}jo, the existence of a PIEfeminine fo rm *-ih,- in addition to the masculine and neuter '-ej-o-). _word Fonnation 285Concerning Latin: Leumann LLFL 1977 p. 286f. (-eus) and p. 321(-inus).4) Adjectives indicating fullness, in the sense of 'rich in ... ': - For information on PIE '-yenl-, cf. W 305.5) Smaller groups: - Temporal adjectives: O. Szemerenyi "Latinhibemus and Greek XEl!1EPl voc;, the formation of time-adjectives in theClassical languages" in Szemerenyi Scripta Minora III 1987 (in an essay of1959) p. 1141-1 159.W 203. Adjectives that are derived from verbs:PIE '-Ia-, '-na- and *-10- are components of verbal adjectives whichhave been partially integrated into the various lE verbal systems as passive(or intransitive) perfect participles. The roles of these suffixes vary fromone IE language to another, e.g. '-10-, which, in Slavic forms the perfectactive participle, but which in Greek is very seldom: Risch Wortbildung1994 p. 107. In the case of the passive perfect participle, the languagesalternate with *-to- and "-no-. Examples of the former are Latin, Greek,and Vedic; an example of the latter is Slavic: Leumann LLFL 1977 p.611. For further sources concerning '-to- and '-no-: Szemen!nyi Ein/iihrung 1990 § 351f. Despite the limited use of'-no-, there are also constructionsthat are common to lE languages that are attributable to ProtoIndo-European, cf. PIE *pJhrna- = Lat. ptenus (in place of "pianos) =Ved. purnti- 'full, filled' = Got. fulls and NHG voll « Proto-Germ.'fulna- < "fuIHno-) = Lith. pi/nas: Mayrhofer EWAia II p. 156; cf.also L 332 § 4d.W 204. Substantives that are derived from substantives or adjectives:I) Terms that indicate feminine beings and that are derived from masculinefonns: - The suffixes "-i-h,- and "-e-h1 are examples of -hr constructions on-i- and -0- stems (cf. F 323 § 2 and F 312) and may with certainty be traced toProto-Indo-EuropeartOne example (among many) of '-e-h,- is the Proto-Indo-European construction"ne!,-eh,- '(the) new one (feminine, sc. woman)' = Lat. nova =Gr. vea = Ved. nava- etc. Some constructions that appear to be of Pro toIndo-European origin, are in fact of later origin, cf. E 506 § 5.The situation of "-i-h,- is more complex. Vedic fe atures two differentt YPes of inflection: Type I = devt-, Type 11 = vrkt-. - The inflectionalmodel Type I (according to MacDonell Vedic Grammar 1910 p. 274)

286 The Proto-Indo-EuTopean Lexiconfeatures: nom. sg. devt, acc. sg. devtm; nom. pI. devts, acc. pI. devtn, instr.pI. devtbhis, etc. vs. gen. sg. devyas, dat. sg. devyai, and instr. sg. devya: Mayrhofer EWAia I p. 744 s. v. devt. - An example of the inflectionmodel Type IT (according to MacDonell Vedic Grammar 1910 p. 270ff,marked by B) is rathi- (here masculine) 'wagon driver, charioteer': nom.sg. rathis (with -si); dat. pI. rathtbhis, acc. pI. ralhin etc. vs. acc. sg. rathiyam(written rathyam), gen. sg. rathiyas (-yas), dat. sg. rathiye (-ye),instr. sg. rathi; nom. acc. pI. ralhiyas (-yas). - Concerning the inflectionof Type I, which is found almost purely apart from the -a- stems, and Typen, which indicates feminine gender in -a- stems and is common as a feminineform of possessive composites: Wackemagel / Debrunner Ai.Gramm. III 1930 p. 163ff. and II / 2 1954 p. 368ff.; Mayrhofer EWAia IIp. 570f s. v. vfka-; by the same author, "Zu iranischen Reflexen desvrkt-Typus" in Mayrhofer Kleine Schriften n 1996 (in an essay of 1980) p.353ff. - Type I is widely represented in the Indo-European branch oflanguages. Among the many examples, compare the feminine active presentparticiples with -nt-, including PIE *h,s-lJt-ih,- 'existing' = Proto-Gr.*-ehat-ja > Myc. Gr. -ehassa = Ved. sat-t- (for information on the verbPIE *h,es-, cf. E 502ff.; for more information on attestations of feminineparticiples in Greek: Meier-Briigger Gr. Sprachw. II 1992 p. 63). -Type II is primarily present in Vedic; non-Indian examples are seldom:compare, on the one hand, Ved. puru$t- (RV+) 'woman,' from puru$a(RV+) 'person' and Ved. naptt- 'granddaughter,' from napiit- 'grandson';and on the other hand, compare Ved. ahi- 'mother cow' with correspondingforms in Old Avestan and Later Avestan azi- and Ved. vrkt- with thecorresponding Old West Norse form ylgr 'she-wolf.' - The origins ofType I are unproblernatic: nom. sg. dev-t- / gen. sg. dev-ya-s < PIE*dejlj-ih,- / *dilj-jid,,-s (the accent of the nominative singular form is secondary,in place of *devi-). While the nominative singular accent in theHomeric Greek example Oio is retained, its stem form originates fromweak case forms. Suggestions for further reading: H. Eichner in Sprache20 1974 p. 28. - Additional information on *-ih,- and *-itih,-:While zero grade *-ih,- yielded *-i- phonetically in IE languages (e.g. theVedic and Latin suffix -tr-i-k-, cf. F 101 § 2), a newly formed *-jh,- (withconsonantal D often resulted from -jeh,-, as is the case in the Greek normalform -ja. Remnant -i- forms are however attested, cf nom. sg. *yA.o\X-1-'tongue' (cf Homeric YACOJ(tC;, 'yoke strap end'; a secondary -in- stem wasformed by sandhi via the acc. sg. -in V-, in the sense of -in' V- ; the accentedsuffix is secondary as well) vs. Ion. gen. sg. YAoaaiic; (with the expectedzero grade stem and -ss- [Att. -11-] < *_kh_j_; on the other hand,Word Formation 287Class. Att. gen. sg. YMYr't11

288 The Proto-lndo-European lexiCOnW 205. Substantives that are derived from verbs:I) Indicator of the agent (nomina agentis): - For infonnation on PIE*dh,-Ier- 'occasional giver' vs. *deh,-Ior- 'habitual giver': Tichy "Die NominaAgenlis au! -Iar- im Vedischen," 1995. Cf. F 101 § 2 for an examplefrom Latin.2) Verbal abstract nouns (nomina aClionis): - H. Rix offers a list of Pro toIndo-European constructions in PS Szemertinyi *65 IT 1979 p. 737f.: Amongthe types featured on the list are those with suffixes such as *-li- (cf. F 317 §7), *-Iu-; those with an -0- stem such as PIE *r6ydh-o- 'to whine'; feminineforms such as *bhug-ehr 'escape'; root nouns of the type *ne"- 'destruction'and -I- constructions of the type *slu-I- 'high praise'; and also -es- neuterfo rms of the type *tep-es- 'heat, warmth.'3) Indicators of tools, means, and places: - For infonnation on PIE*-Ir-o-, *-11-0-, *-dhr-o- and *-dhl-o- (which find their point of departure in the-0- derivations from nomina agentis with -Ier- I -Ior-, which, due to variationsin root forms, are often misleadingly transformed to *-Il-o- or *-dhr-o- or*_dhl_o_): Risch Worlbi/dung 1974 p. 41; Olsen Instrumenl Noun SuffIX1988; M. V. Southern in MSS 60 2000 p. 89-133 (which concerns Lat. labulaand the suffix -dhlo-). Further, cf. L 347 § 2.W 206. Suffix systems, suffix associations: - Parts of the group describedby W. Caland are demonstrably of Pro to-In do-European origin.I) In his treatment of the Later Avestan word xruu-i-drau- 'who carries aterrible, bloody, wooden weapon,' W. CaIand called attention to the fact thatAvestan adjectives with -ra- and -ma- as the initial element in the case of compoundadjectives replace the -ra- descriptively with -i-. See, for example theadjective xro-ra- (corresponding to the compound example) 'bloody, temble'or cf Later Avestan draz-ra- 'tough, strong, capable' vs. Later Avestandraz-i-ratJa- 'who has a strong wagon': KZ 31 1892 p. 267 and 32 1893p. 592 ("It now [seems] probable to me that this characteristic lasts into theIndo-Eranic period"). In his piece "APrIKEPA rNOE und Genossen" publishedin Wackernagel Kleine Schrifien 1 1969 (but which dates from 1897) p.770, J. Wackernagel showed that the phenomena is of PIE origin: "The CalandRule is thus corrunon Indo-European." Wackernagel refers to Greek examplessuch as apyo -rg- or directly from PIE*hrg-6-, cf. L 333 § I) vs. apY-l-KEpauvo

290 The Proto-Indo-European LexiconA. root noun B. adiective C. abstract noun O. adiectivelE str. *kroljh,-, a. *krnh,-ro- str. *krOljh,i-, *krulzzi-no- -wk. *krnh,- ' b*krouh,-o- wk. *krliuh,ilaterA vestan xru- a. Avest. A vestan prefix Lith. kruvinasxriira- xruuib.German rohComment:Situated at the beginning of the chain of derivations is theroot noun, *krOyh,- (A) = LAv. xrn- 'bloody, raw meat,' from which thepossessive adjectives *krnh,-ro- (Ba) = OAv. and LAv. xrnra- 'bloody,terrible' and *kroyh,o- (Bb) = NHG roh were derived externally. Startingfrom the adj. * kroljh,o- (Bb) the external derivation leads, via replacementof -0- by -i-, to *kroyh,i- (C).It is incorporated as a prefix in the zerograde of compound forms, e.g. *krnh,i- = LAv. xruui-drau- 'who carries abloody weapon.' The -i- abstract noun may be fo und indirectly in the possessiveVed. -0- derivation kravya- 'bloody' « *kroyh,j-o- 'possessing*kroljh,i-'). The derivation then continues to the possessive *krnh,i-no(0) = Lith. kruvinas 'bloody,' etc. For more detailed infonnation: -7 A. 1.Nussbaurn in GS Schindler 1999 p. 402.6) A central insight for further understanding of compound fo rmationsis that there are no adjectives, only abstract nouns. The compound fo rmxruui-drau- may first be understood in the sense of 'whose (wooden)weapon is characterized by blood.' However, the -i- abstract noun wasquickly out of use, whereby the prefix was accorded adjectival value, andxruui-drau- was then interpreted as 'who has a bloody (wooden) weapon.'From the appearance of the commonly used adj. xrnra-, the impressionarose that the adjective was also present in the prefix, with replacement of-ra- by -i-.7) It may also be explained why, in contrast to the -ro- adjectives (e.g.xriira-), the -u- adjectives (e.g. Gr. 1tOI..) appear as prefixes of compoundfo rms. Even here appearances are misleading. The point of departurein this case is the internal derivation from the aerostatic abstractnoun *p6lh/u- 'fullness' and proterokinetic possessive adjective str."pelh/u-, wk. "p!h/elj- 'having fullness, much,' cf § 4. The abstract nounis present in the prefixes of compound fo rms, but since both the abstractnoun and the adjective, owing to their internal deduction remain -u-stems,the impression thus arose much later, at a time when the abstract noun haddisappeared, that the still-productive -u- adjective was reflected in the prefix.The -u- adjectives have generally established themselves with zeroword Formation 291stem roots, cf. here Ved. puru-. The -0- of Gr. 1tOI.. was long misunderstood,but is now recognized as coming from an abstract noun.3. Word Formation by CompositionW 207. The possibility of fo rming compound words dates from ProtoIndo-European and consists of the joining of two nominal stems, a firstelement, and a second element. -The first element is norrnally composedof a simple stem while the second element may be inflected: The stem ISeither used directly, or in combination with a suffix such as " -0-, *-i-,*-ijo-, etc. The new noun receives its own accent aod a new meaning. Two types of PIE models are the possessive compounds (called bahuvnhlin Old Indian terminology, meaning 'having much rice') and the verbalgovernment compounds. Proto-Indo-European itself features several establishedtypes of both models.1) In the case of the possessive compound, nominal phrases such as Gr.t = !!EA.a 'the horse [is] black' becomes the noun I..OV-t1t1tO-. Intltis case, the compound fo rm serves to characterize a person, namely one,'who owns a black horse/rides on a black horse.' As a rule, prefixes findtheir origin in abstract nouns, cf. W 206 § 6. See further W 208 below.2) In the case of the verbal government compound, a verbal sentence ofthe type Gr. ICOUpOV PEq,et '(the wet nurse) feeds a male child' is nominalizedto fo rm ICoupo-p6$O-. In order to make the compound form possiblethe finite verb is transformed into an abstract noun, here in the fo rmof .;p6$O"The newly formed word indicates a person 'who goes aboutthe feeding of children.' See further W 209 below.3) The genesis of compound forms that is described here, with the exampleof Greek, most likely took place in the pre-Proto-Indo-Europeanperiod: -7 J. Wackernagel in Wackernagel / Oebrunner Ai. Gramm. II / 11957 (first edition, 1905) p. 289 (dealing with possessive compoundfo rms) and p. 186 (on verbal government compounds); Risch KleineSchriften 1981 (in a contribution from 1945) p. 124, note 21.4) Terminology concerning compound fo rms is not completely uniform.On the one hand, there is a relatively comprehensive Old Indian terminologywhich is fully adapted to the classical circumstances peculiar to OldIndian and built up purely in terms of content: -7 J. Wackernagel in Wackernagel/ Oebrunner Ai. Gramm. II / 1 1957 p. 140-142. On the other

292 The Proto-indo-European LexiconWord Fonnation293hand, a separate tenninology has evolved within lE linguistics, includingterms already used by E. Riseh in his early work such as: possessive compound,verbal government compound, determinative compound (cf. W21 1), and prepositional government compound (cf. W 210): RisehKleine Schriften 1981 (in a contribution from 1944) p. 5 and p. 36 and (ina contribution of 1945) p. 112f.; Riseh Wortbi/dung 1974 p. 182.5) To date, there exists no comprehensive treatment of Proto-IndoEuropean compound forms. Sadly, his passing prevented 1. Schindlerfrom completing the planned volume on Indo-European grammar. However,the plan of his Madrid lecture is accessible: "Zur intemen Syntaxder indogermanischen Nominalkomposita" in Kolloquium Delbruck Madrid1994 [1997] p. 537-540.6) More recent suggestions for further reading: Mikkola KompositumI 1971 ; Th. Linder Lateinische Komposita Innsbruck 1996; V. Sadovskiin Arbeitstagung Erlangen 1997 [2000] p. 455-473 (concerning whatare caUed entheos compounds and prepositional government compoundsin the Rigveda); G. E. Dunkel "On the origins oJ nominal composition inlE " in CS Schindler 1999 p. 47-68.W 208. FoUowing is further infonnation on lE possessive compoundfo rms (see W 207 § I): - The contents of ancient lE possessive compoundforms often relate to appearance, property, fame, beauty, andpower. The first element may be an adjective, substantive, number, oreven a preposition, e.g. La!. con-cord- 'having a heart that is together'meaning 'unanimous,' Gr. pl-1t08- 'having three feet' in the sense of 'acontainer with three feet' or Ved. su-sravas- 'having powerful, positivefame' in the sense of 'glorious,' etc.Suggestions for further reading: Riseh Wortbildung 1974 p. I 82ff.;J. Schindler "Zu den homerischen PoOooaKTVA-oq-Komposita' in FS Risch1986 p.393-401; J. UhIich 'Der Kompositionslyp 'Armstrong' in den indogermanischenSprachen' in HS 110 1997 p. 21-46.W 209. The fo Uowing section contains further infonnation on verbal governmentcompound forms (see W 207 § 2): At least two different typesof constructions are traceable to Proto-Indo-European:I) Verbal government compound fonns with verbal first element, e.g., Gr.exe-1tCJlAo-C; 'he who keeps young horses' and Vedic /rasa-dasyu- 'he whocauses his enemies to tremble.' Suggestions for further reading: RischWortbi/dung 1974 p. 190ff.2) Verbal government compound fonns with a verbal second element: -What are caUed root compound forms represent an archaic group, e.g., La!.,in-ceps < "primo-cap- 'he who takes the first part' and con-iug- 'boundfogether,' Gr. xep-vt!l- (which was represented y in Mycenaean,aske-ni-qa i.e. id'err-nigW-) '(water) that washes hands and VedlC havzr-ad'eating the food of sacrifice.' The second element is not, as is generaUy supposed,a nominal root, but rather a verbal root: Scarlata Wurzelko"!positaim !J.g- Veda 1999 p. 765f. Suggestions for further reading: In additIon toScarlata, see particularly Benedetti Composti radicali 1988. - The morecommon type in ancient rE languages is that exemplified by Gr. lCoupmp04>

294 The Protlndo-European LexiCOnThe Lexicon of Names295with genitive) dates to the Proto-Indo-European period, the fusion is postProto-Indo-European. For information on the archaic gen. sg. PIE "dems,cf. F 320 § la.Suggestions for further reading: E. Risch "Grieehisehe De/ermina/iviromposi/a" in Riseh Kleine Schrifien 1981 (in the post-doctoral workpublished in IF 59 Hefi I 1944 and IF 59 Hefi 3 1949) p. I If. Owing tothe index in the Kleinen Schrifien, a large part of the examples were explainedfo r the first time.W 212. For information on what are called Amret;iila, e.g. Ved. diwidive'day for day,' se F 318 § 6a (at the end of the section).C. The Lexicon of Names1. General InformationW 300. In addition to the value of the word as an appellative, it also hasits place as a name. "Each word of a language is a 'name' with regard tothe act of designation": B. SchJerath "Name and Word in IndoEuropean" in FS Hamp IT 1997 p. 164-169.The abstract noun " hJmih,-ml}- may be traced with certainty to ProtoIndo-European and means 'name, appellation, qualification' ( La!.nomen, Gr. Qvol1a, Ved. naman-, Hitt. liiman-, Got. namo, oes imf etc.: Mayrhofer EWAia 11 p. 35-37 and cf. F 317 § 10).According to the evidence in individual IE languages, *"Jmih,-men- +*dhehJ- 'to give a name to someone' was a feature of Proto-IndoEuropean. The naming of a newborn child, generally performed by thefather, presumably took place on or after the ninth day after the baby'sbirth: R. Schrnitt in Namenjorsehung I 1995 p. 616.W 301. Proper nouns have a special role in the lexicon of a language. Incontrast to appellatives, they indicate an individual and not a class, orgenre.The marking of the individual is not only important in designating individualpeople, but also in designating groups of people, geographical features,etc.Due to space limitations, the following treatment is quite brief.more information on the subject: NamenJorsehung 1-2 1995-1996.2. Names of Persons and DeitiesW 302. The common manner of naming a person in our western worldconsists of a first name and a surname. The first name (German: Vorname;French: prenom) serves to identifY a person. The surname, or family name(German: Naehname, French: nom de Jamille) generally reveals the geneticmembership of a person in a fa mily.First names are determined by the parents (or rather, the father) andgenerally remain valid until the death of the named individual.ForThe parents'choice of first names is dependent on a number of factors, which melude (among many others): family tradition (that the son inherits the nameof the grandfather), vision of the future, the parents' plans for the life oftheir child, e.g. that he should be successful, etc. While Iirst names arefreely chosen, family names are predetermined and may not be varied.The current European naming practices including a Iirst name and afamily name took shape in a long evolutionary process starting in the 12' hcentury A.D. The family name was originally an addition to the Iirst name,and only later became inherited.One suggestion for further reading,among many others: Bach Deu/sehe Personennamen 1943.1) All evidence suggests that, in the ancient IE period, only the individualname was common, e.g. the Homeric Greek practice with examples such as"AXlAAe{x;, "EK'tWp, nmpoKA (with -es < "-eyes; however, in place of thefull name, the sbortened form na"tpO is common, which is formed fromthe vocative form napoKI..e, which itself is abbreviated from the full vocativeform "n6poKA.u:

296 The Proto-Lndo-European LexiCOnnamen und Kurznamen bzw. Kosenamen im Indogermanischen" ibid. p.419ff.3) Further, the syntactical indication of the father's name also dates fromProto-lndo-European, whether by adding the name of the futher in the genitive,in the sense of 'son of X,' or by adding a possessive adjective that is derivedfrom the name of the futher. An example of the former is HadubrandHeribrandes suno; an example of the latter is Myc. Gr. a-re-ku-tu-ru-woe-te-wo-ke-re-we-i-jo i.e. Alekiruljon Eteljokleyehijos 'Alektruyon, son ofEteyokleyes' and Russ. Nikolaj Sergejevic. - Patronymics ending in -ios(later -ius) led to what is called the nomine gentile in Rome, compare Gaiuslulius Caesar with Gaius = praenomen < individual name, IlIlills = nomengentile < patronymic and Caesar = cognomen. The first step towards thesystem that is widespread in all of Middle Italy is the transferral to each personof the patronymic from the futher. For more comprehensive information: Rix Gentilnamensystem 1972 passim (For information on suffix functions inProto-Indo-European, p. 71.). - For information on Greek patronymics: Meier-Briigger Gr. Sprachw. Il p. 21 (in section 3 below).4) When considering the giving of names to individuals, one departs generallyfrom the basis of the free male. Individual names of women have theirown special issues and are directly related to the role of women in IndoEuropean society, as is shown in forms of address. Whereas the man is addressedusing the individual name, a simple 'oh woman' suffices in the case ofthe woman. "The woman is treated more as a typus, the man as an individual": J. Wackemagel "Uber einige antike Anredeformen" in WackernagelKIeine Schrifien Il 1969 (in an essay of 1912) p. 970ff. and particularly p. 993(Wackemagel makes clear that the same forms of address were adopted forinteractions with the gods. To say that the [ndo-Europeans were not verydifferent from the Romans and Greeks would not likely be too far from themark). - In Rome, women generally carried only the nomen gentile, cf.Cornelia, IlIlia etc.: Rix Gentilnamensystem 1972 p. 704. - In the caseof the Greeks, most names of women are simply feminine forms of masculinenames of individuals, e.g. (already) Myc. a-re-ko-sa-da-ra, i.e. Aleksandrii(which corresponds to *Aleks-anor, meaning 'who fights off men'), Hom'AvOpo!'

298 The Proto-Indo-European LexiconThe Lexicon of Names299W 304. There are names of tribes, names of peoples, names of warriorgroups, etc, that correspond to the diverse characters of groups of peoplethat are determined socially and economically.It is important to note whether the name of a tribe or group originatesfrom within the group of from a neighboring group, e.g., the case of theClassical Greek word for 'Greek,' "EAA'lVec; vs. Lat. Graeci (which originatesfrom the Italian penninsula, from which perspective the name of theGraeci of Epirus [modem northwestern Greece and Albania] came to bethe Latin term for all inhabitants of the region): Biville Emprunts II1995 p. 178 note 39 with suggestions for further reading.The extent to which names of groups, tribes, and peoples may be tracedto Proto-Indo-European is not clear. The lndo-Iranian self-descriptiveterm *ar(i)ja- is old, for example. For more information on the associatedissues: Mayrhofer EW Aia I p. 174f. ; F. Bader "Les noms des aryens:Elhniques el expansion" in Langues Indo-europeennes 1994 p. 65ff. -The Mycenaean Greeks apparently called themselves AkhajlJo-: CbantraineDELG I 1968 p. 149; Latacz Troia und Homer 200 I p. 150ff.(which includes information on the Homeric terms Akhaioi, Danaoi andArgliioi).3. Fonnation of Place NamesW 305. In addition to individuals and groups of individuals, the environmentsthat surround the individuals and the places where they live are importantpoints of orientation: mountains, rivers, bodies of water, settlements,cities, etc. These place names are stable and are transferred fromgeneration to generation.While individual lE languages are clearly linked to particular geographicregions, no such clarity exists for the Proto-Indo-European period,cf. E 512. It is thus practically impossible to determine whether particularplace names are of Pro to-In do-European origin.The comparison ofIE habits in the formation of place names leads oneto suppose, in the case of commonality among several languages, thatthese hahits may be traced to speakers of Proto-Indo-European. Amongthese habits is surely the use of the lE suffix *-Ijenl- (sometimes with theablaut *-yonl-, fern. *-y,!l-ih,) to characterize typical merits of a particularplace, e.g. the Mycenaean place name dat. sg. sa-ri-nu-wo-/e i.c. Sa!in-yont-ej '(place) where celery flourishes' (in which case alphabeticGreek features LeA- in place of Sal-. For more information: LeukartFriihgr. Nomina 1994 p. 116f. note 233); the Indo-Iranian n,am for a ri er_d region 'saras-yal-iH- (= Ved. sarasvatz-, LAv. harax alll-) mearungiver or region) where there e many .swamps';.the L !c:an place nameXada-wiili- (hellenized as Kaouavoa), I.e. *bada-lJanll- (place) that ISrich in grain' ( Hajnal Lyk. Voka!ismus 1995 p. 88). Suggestlons forfurther reading: R. Schmitt in NamenJorschung 1995 p. 633f.I) As a rule, place names originate from general appellatives e.g., the Vedieriver name sindhu- 'Indus,' which was originally an appellatlVe that t'river': The Indus, like the Nile and the Rhine, is the unique prototypical nverin its region.2) Tnbe names may be used to indicate places, e.g. Gr. &A4>oi as an accusative(of direction) plural tribe name 'to the Delphians' .and .dat. loc. pI.6EA4>6tc; 'with/among the Delphians.' Since the Delphians inhabIt a particularregion and town, the accusative and dative!locative can be direc!\Y used toindicate them: &A4>oi, 'city in Phokis at the foot of mount Parnass. Similar 15the case of Old Indian loc. pI. madre$U which means 'with/among the mat1ra-people,'which may be interpreted, 'in the area where the madra-peopledwell.' This characteristic of equating the tribe name with the name of theplace where it lives is in all likelihood of Pro to-Indo-European origirI.3) Place names and prehistory: Because place names are stable and n:mainlinked to their objects, they are particularly instructive regardmg the prehisto ,?,of a place. In addition to the place names that may be mterpreted on the basl5ofthe locally spoken language, there are as a rule names as well that .may notbe interpreted by any single language. When these names are very anCient, onemay conclude that their appearance occurred at a time at which the locallyspoken language was not yet known and another language was in use. However,when further documentation of such a language is missing, it is practicallyimpossible to extrapolate information about the language based on placenames. One difficulty among many are adaptations of the name to subsequentlanguages. _ Greece presents an example: Place names that may not be understoodon the basis of the Greek language, such as K6ptv6oc; and 'AEnjvat(!) date, unlike comprehensible names such as LeAtVOUc; (meaning 'rich incele ry ,' see above), from before the settlement of the land by the Greeks tookplace, and may thus be considered pre-Greek: E. Risch "Ein Gang durchdie Geschichle der griechischen Orlsnamen" in Kleine Schrifien 1981 (m anessay from 1965) p. 145ff.. - Examples of European hydron y: Names ofrivers such as Elbe and Rhine are incomprehensible on the basl5 of Gennarucand must therefore be of pre-Gerrnanic origin. However, whether these names

300The Proto-Indo-European LexiCOnare based upon Proto-Indo-European linguistic material and were thus Creatby speakers of Proto-[ndo-European remains disputed, cf. E 513 § 3.ed4) Helpful resources concerning place names: -? l-L. Garcia-Ram6n inDNP 4 .1998 Sp. 93 0-934 s. v. "Geographische Namen"; Namenjorschung 21996 WIth chapter X ( ' Namengeographie ') and the chapters :xv -xvn (' Or _n men I: Siedlungnamen;' 'Orlsnamen 11: Flurnamen;' 'Gewassername:');TlSChler KlemaslOlIsche Hydronymie 1977' Zgusta Kieinasiatische Ortmen 1984; KiemaslOtlsche Onomaslik in Namenjorschung 1995 p. 63611". . .RGA s. v. 'Liinder- und LandschaJisnamen.''sna ·.,VI. Bibliography and Key to ReferenceCitationsMore than simply a list of works used in the preparation of this book andtheir shortened citations, the following bibliography aims at furnishing thereader with a representative overview of current lndo-European specialtyliterature as of mid-2001. Included are titles which will not he referred toelsewhere in the present work. Exhaustiveness is not pretended in anyarea All publications are not equally good, but none of those listed is sopoor that one cannot learn something from it, even if what is learned issimply how things are not.Reviews and discussions of books constitute a separate genre of literature.Particular works are mentioned here and there in order to call attentionto this type of information. It is heartily recommended to those whoare penetrating a new field of study, or searching the newest positions on acertain problem, that they take into account the opinions, corrections, andcompletions of appropriate reviewers. However, it was never intended inthis work that all the concerned reviews he named. I especially refer thereader to the publication for reviews and reports, Kratylos, published bythe Indogermanische GesellschaJi, and of which the most recent volumeavailable to me at the time of publication is 46 2001. Occasionally a noteOr comment from 'ldg. Chr.' will be referred to (see entry below).The basis for the abbreviations of journals used in this text is the BibliographieLinguistique / Linguistic Bibliography, the last volume of whichthat was consulted is that from 1994. In addition, a couple of particularlyrelevant titles appear in the list. For habitual lecture and consultation, thefollOwing (in alphabetical order) are recommended, among others: DiachrOnica,Glotta, HS (older abbreviations ZVS, and KZ), IF, JIES, MSS andSprache.

302 Bibliography and Key to Reference CitationsBibliography and Key to Reference Citations303The fo llowing bibliographical information is not always unifonnly Organized.In the case of newer works, indications are included at the end ofthe entry, when the work appeared in an particular series, such as a Supplementaryvolume to the ZVS / HS, an IBS volume, or the review of anacademy.In the case of publications which have long been well-known,only the place of origin and year, and no additional bibliographical information,are given.AA WL = A bhandlungen der Akademie der Wissenschajien und der Literaturin Mainz.Adarns Dictionary 1999 = D. Q. Adams, D.Q. 1999 A Dictionary oJ TocharianB. Amsterdam and Atlanta, Ga (= Leiden Studies in IndoEuropean 10).Adarns Tocharian 1988 = Adarns, D. Q. Tocharian Historical Phonologyand Morphology. New Haven 1988 (= American Oriental Series 71).Adiego Studia Carica 1993 = Adiego, I.-J. Lajara Studia Carica, Investigacionessobre la escritura y lengua carias. Barcelona 1993.Adrados Manual, see below, Manual de lingiiistica indoeuropeaAitzetmUller Abulg. Gramm. 1991 =R. AitzetmUller 1991 AltbulgarischeGrammatik als Einfiihrung in die slavische Sprachwissenschaft. 2d ed,revised and expanded, Freiburg (= Monumenta Linguae Slavicae Nr.30).Akten 13. Osterreich. Linguistentagung 1988 = Zinko, c., ed. 1988 Aktender 13. Osterreichischen Linguistentagung 1985 in Graz mU den Beitragender Tagung 1983 in Salzburg. Graz.AlIen Vox Graeca 1987 = AlIen, W. S. 1987 Vox Graeca, A Guide to thePronunciation oJ Classical Greek. 3d ed. Cambridge.Althochdeutsch I / II 1987 = Bergrnann, R., H. Tiefenbach and L. Voetzeds. 1987 Althochdeutsch. Heidelberg: Volume I (Grammatik. Glossenund Texte); Volume IT (Warter und Namen. Forschungsgeschichte).Ambrosini Linguistica Indo-Europea 1 111 1996 = R. Ambrosini 1996 Introduzionealla linguistica indo-europea. 1 (La ricostruzione dell' indoeuropeo),IT (Le lingue indo-europee orientali e centrali). Lucca.Ammann Untersuchungen I 1922 = H. Arnmann Untersuchungen zurhomerischen Wortolge f und Satzstruktur. I. allgemeiner Teil. Leipzig1922; by the same author, vol 2: Die Stellung des Verbums, im einzelnenuntersucht in IF 42 1924 p. 149-171 and 300-322.Ancient lE Dialects 1963 [1966] = Ancient Indo-European Dialects, ProceedingsoJ the ConJerence on Indo-European Linguistics 1963 in LosAngeles, H. Birnbaum and 1. Puhvel (eds). Berkeley I Los Angeles1966.Andersen (H.) Prehistorical Dialects 1996 = H. Andersen ReconstructingPrehistorical Dialects, Initial Vowels in Slavic and Baltic. Berlin 1996.Review: Chr. Koch in Kratylos 45 2000 p. 146-154.Andersen (P. K.) Word Order Typology 1983 = P. K. Andersen Word OrderTypology and Comparative Constructions. Amsterdam / Philadelphia1983. Review: Ch. Lehrnann in Kratylos 29 1984 [1985] p. 25-30.Anttila PIE Schwebeablaut 1969 = R. Anttila Proto-Indo-EuropeanSchwebeablaut Berkeley I Los Angeles 1969. Review: J. Schindler inKratylos 15 1970 [1972] p. 146-152.A6A W = Anzeiger der Osterreichischen Akademie der Wissenschajien,philosophisch-historische Klasse.Arbeitstagung Erlangen 1997 [2000] = Indoarisch, Iranisch und die Indogermanistik,Arbeitstagung der Idg. Gesellschaji 1997 in Erlangen,edited by B. Forssman and R. Plath (eds), Wiesbaden 2000. Review: B.Schlerath in OLZ 962000 p. 306-3 16.Arbeitstagung Osk.-Umbr. Freiburg 1991 [1993] = Oskisch-Umbrisch,Texte und Grammati/c, Arbeitstagung der Idg. Gesellschaji und der SocietaItaliana di Glottologia 1991 in Freiburg, edited by H. Rix, ed.Wiesbaden 1993.Arbeitstagung (100 Jahre) Tocharologie SaarbrUcken 1995 [1997] = Arbeitstagung100 Jahre Tocharologie, Kolloquium der Idg. Gesellschaji1995 in SaarbrUcken, published as TIES 7 1997.Arens Sprachwissenschaji I + II 1969 = H. Arens Sprachwissenschaji,Der Gang ihrer Entwicklung von der Antike bis zur Gegenwart.Frankfurt am Main 1969, 1 (Von der Antike bis zum Ausgang des 19.Jahrhunderts), IT (Das 20. Jahrhundert).Arlotto Introduction 1972 = A. Arlotto Introduction to Historical linguistics.Boston 1972.Arumaa Urslav. Grammatik I 1964 II 1976 III 1985 = P. ArurnaaUrslavische Grammatik. Heidelberg: vo!. I (Einleitung, Lautlehre mitVokalismus und Betonung) 1964; vo!. II (Konsonantismus) 1976; vo!.III (Formenlehre) 1985.ASNP = Annali della Scuola Normale Superiore di Pisa: lel/ere, storia efilosofia.

304 Bibliography and Key to Reference CitationsAspects of Latin 1993 [1996] = Aspects of Latin, Papers from the SeventhInternational Colloquium on Latin Linguistics Jerusalem 1993, editedby H. Rosen, ed. Innsbruck 1996 (= IBS 86); see below, s. v. IKLL(JCLL / CILL) VII 1993 [1996].Aspelde baltist. Forschung 2000 = Aspelde Baltistischer Forschung, editedby J- Range, ed. Essen 2000 (= Schrijien des Instituts fur Baltistikder Ernst-Moritz-Arndt-Universittit Greifswald 1).Assmann Kulturelles Gedtichtnis 1997 = J. Assmann Das kulturelleGedtichtnis. Schriji, Erinnerung und politische Identittit in fruhenHochkulturen. Munich 1997 (and re-editions).Aufrecht Hymnen des RV 1877 = Th. Aufrecht Die Hymnen des Rigveda.2 vols. 2d ed. 1877 (and reprints).Aura Jorro DMic. I 1985 II 1993 = F. Aura Jorro Diccionario micenico.Madrid, vo!. 1 (a-n) 1985, vo!. " (o-z) 1993.A utobiographische Berichte 1991 = Wege in der Sprachwissenschaji,Vierundvierzig autobiographische Berichte, Festschriji fur M Wandruszka,edited by H.-M. Gauger and W. Piickl, eds. TUbingen 1991 (=Tubinger Beitrtige zur Linguistik, vo!. 362).Bach Deutsche Personennamen 1943 = A. Bach Die deutschen Personennamen.Berlin 1943.Baltische Sprachen 1994 = Die baltischen Sprachen, Eine Einfuhrung, R.Eckert, Elvira-Jutia Bukeviciiite, F. Hinze. Leipzig / Berlin 1994. Review:F. Scholz in Kratylos 42 1997 p. 126-/30.Baltistik 1998 = Baltistik, Aufgaben und Methoden, edited by A.Bammesberger. Heidelberg 1998 (= Indogermanische Bibliothek, Reihe3, vo!. 19). Review: R. Matasovic and W. P. Schmid in IF 105 2000 p.342-351; F. Heiderrnanns in Kratylos 45 2000 p. 154-162.Bammesberger Abstraldbildungen 1973 = A. Bamrnesberger Abstraldbildungenin den baltischen Sprachen. Giittingen 1973.Bammesberger Germ. Verbalsystem 1986 = A. Bammesberger Der Aufbaudes germanischen Verbalsystems. Heidelberg 1986 (= Untersuchungenzur vergleichenden Grammatik der germanischen Sprachen,vo!. I).Bammesberger Laryngaltheorie 1984 = A. Bamrnesberger Studien zurLaryngaltheorie. Giittingen 1984 (= supplement to ZVS 33). Review:R. S. P. Beekes in Kratylos 31 1986 p. 70-75; F. O. Lindeman in IF 911986 p. 349-351. See also below s.v. Laryngaltheorie 1988.Bibliography and Key to Reference Citations 305Barnmesberger Pforzen und Bergakker 1999 = Pforzen und Bergakker,Neue Untersuchungen zu Runeninschrijien, edited by A. Bamrnesberger.Giittingen 1999 (= HS, Ergtinzungsheji 41).Barnmesberger Urgerm. Nomen 1990 = A. Bammesberger Die Morphologiedes urgermanischen Nomens. Heidelberg 1990 (= Untersuchungenzur vergleichenden Grammatik der germ. Sprachen Nr. 2).Bartholomae Air. Worterbuch 1904 = C. Bartholomae AltiranischesWorterbuch. Strasbourg 1904. The second photo mechanical reprint of1979 contains supplements and improvements (vo!. 1881-1900), as wellas supplementary material and groundwork from 1906. Berlin 1979.Bartschat Methoden der Sprachwissenschaji 1996 = B. Bartschat Methodender Sprachwissenschaji. Von Hermann Paul bis Noam Chomsky.Berlin 1996.BCH = Bulletin de Correspondence Hellenique.Becbtel Gr. Dialekte 1921-1924 = F. Bechtel Die griechischen Dialelde, 3vols. Berlin 1921-1924.Becbtel Hauptprobleme 1892 = Fr. Bechtel Die Hauptprobleme der indogermanischenLautlehre seit Schleicher. Giittingen 1892.Beekes Gatha-Avestan 1988 = R. S. P. Beekes A Grammar of GathaAvestan. Leiden 1988. Review: J. E. Rasmussen in Kratylos 36 1991 p.109-116. C[ also, by the same author, "Historical Phonology of lraruan"in JIES 25 1997 p. 1-26.Beekes lE Nominal Inflection 1985 = R.S.P. Beekes, The Origins of theIndo-European Nominal Inflection. Innsbruck 1985 ( = IBS Nr. 46).Review: J. Schindler in Idg. Chr. 31a 1985 Nr.85.Beekes Introduction 1995 = R. S. P. Beekes Comparative Indo-EuropeanLinguistics, An Introduction. Amsterdam 1995. Review: A. Bamrnesbergerin HS 109 1996 p. 310-3 16; M. KUmme1 in PFU 2-3 1996 /1997 p. 113-125; Ch. de Lamberterie in BSL 92 / 2 1997 p. 143-149;SI. Zimmer in PBB 119/2 1997 p. 276-282.Beekes Laryngeals 1969 = R. S. P. Beekes The Development of theProto-Indo-European Laryngeals in Greek. The Hague / Paris 1969.Review: C. J. Ruijgh in Lingua 26 1970 / 1971 p. 181-198 (= ScriptaMinora 1 1991 p. 330-347); H. Rix in Kratylos 14 1969 [1972] p. 176-187. C[ also R. S. P. Beekes Laryngeal Developments: A survey inLaryngeal Theory 1988 p. 59-105.Benedetti Composti radicali 1988 = M. Benedetti I composti radicalilatini: Esame storico e comparativo. Pisa 1988.Benfey Geschichte der Sprachwissenschaji 1869 = Th. Benfey Geschichteder Sprachwissenschaji und orientalischen Philologie in Deutschland

306 Bibliography and Key to Reference CitationsBibliography and Key to Reference Citations 307seit dem Anfange des i9. Jahrhunderts mit einem Rilckblick auf diefrilheren Zeiten. Munich 1869 (reprinted 1965).Benveniste Hittite et indo-europeen 1962 = E. Benveniste Hit/ite et indoeuropeen,Etudes comparatives. Paris 1962.Benveniste institutions I + IT 1969 = E. Benveniste Le vocabulaire desinstitutions indo-europeennes. 2 vols. Paris 1969. Translated into Englishwith the title "Indo-European Language and Society," London1973. Translated into Gennan as "Indoeuroptiische Institutionen."Translated by W. Bayer, D. Hornig, K. Menke, edited, and with an afterwordby SI. Zimmer. Frankfurt / New York 1993. Review: R.Schrnitt in Kratylos 39 1994 p. 183f.Benveniste Noms d'agent 1948 = E. Benveniste Noms d'agent et nomsd'action en indo-europeen. Paris 1948.Benveniste Origines 1935 = E. Benveniste Origines de la formation desnoms en indo-europeen. Paris 1935.Benveniste Problemes I 1966 II 1974 = E. Benveniste Problemes de linguistiquegenerale. Paris vol. 1 1966, vol. II 1974.Benveniste, see also s. v. Colloque E. Benveniste.Berlinische Lebensbilder - Geisteswissenschajiler 1989 = BerlinischeLebensbilder, edited by W. Ribbe: vol. 4, Geisteswissenschajiler, editedby M. Erbe. Berlin 1989.Bibliographie d. Hethitologie 1-3 1996 (1998) = Systematische Bibliographieder Hethitolagie 1915-1995, 3 vols., assembled by V. Soucekand J. Siegelova. Prague 1996.Bile Cretois 1988 = M. Bile Le dialecte cretois ancien, Etude de la languedes inscriptions, Recueil des inscriptions posterieures aux IC (= InscriptionesCreticae). Paris 1988.Binnig Gotisches Elementarbuch 1999 = W. Binnig Gotisches Elementarbuch.Berlin 1999 (Tbis de Gruyter Studienbuch replaces H. HempelGotisches Elementarbuch, Grammatik, Texte mit Obersetzungen undErltiuterungen, Berlin 1966, Sammlung Goschen vol. 79/79a).BiOr = Bibliotheca Orientalis.Birkhan Kelten 1997 = H. Birkhan Kelten. Vienna 1997. Review: J.Ublich in CMCS 39 2000 p. 65-73.Bitte1 Hattusha 1970 = K. Bittel Hattusha, The Capital of the Hittites.Oxford 1970. See also, s. v. Neve Hat/usa 1996.Bitte1 Hethiter 1976 = K. Bittel Die Hethiter: Die Kunst Anatoliens vomEnde des 3. bis zum Anfang des i. Jahrtausends vor Christus. Munich1976.Biville Emprunts I 1990 II 1995 = F. Biville Les emprunts du latin augrec. Approche phonetique. Louvain, Paris: vol. I (Introduction et consonantisme)1990; vol. II (Vocalisme et conclusion) 1995.BL = Bibliographie linguistique / Linguistic Bibliography. Boston / London.Blaiek Numerals 1999 = V. Blaiek Numerals: Comparative-EtymologicalAnalysis of Numerals Systems and their Implications. Brno 1999. Review:M. de Vaan in Sprache 39 / 2 1997 [2000] p. 239-245; V. Bubenikin JIES 28 2000 p. 450-454.B10ch Suppletive Verba 1940 = A. Bloch Zur Geschichte einiger suppletiverVerba im Griechischen. Base1 1940.BlUmel Aiol. Dialelete 1982 = W. B1Ume1 Die aiolischen Dialelete, Phonologieund Morphologie der inschriftlichen Texte aus generativerSicht. Giittingen 1982 (= supplement ZVS Nr. 30).B1Urne1 Untersuchungen 1972 = W. B1Umel Untersuchungen zu Lautsysternund Morphologie des vorklassischen Lateins. Munich 1972 (=MSS, Beiheji, Neue Folge 8).BNF = Beitrtige zur Namenforschung.Boh! Besitzverhtiltnis 1980 = S. Boh! Ausdrucksmiltel fur ein Besitzverhtiltnisim Vedischen und Griechischen. Louvain-la-Neuve 1980.Boisacq DELG 1950 = E. Boisacq Dictionnaire etymologique de lalangue grecque etudiee dans ses rap ports avec les autres langues indoeuropeennes.4t h edition, augmented with an index by Helmut Rix. Heide1berg1950.Boley Hitlile hark-construction 1984 = J. Bo1ey The Hit/ite harkconstruction.Innsbruck 1984 (= IBS 44).Boley Sentence Particles 1989 = J. Bo1ey The Sentence Particles and thePlace words in Old and Middle Hiltite. Innsbruck 1989 (= IBS 60).Cf, by the same author, The Hiltite Particle -z / -za. Innsbruck 1993 (=IBS 79).Bopp Albanesisch 1855 = F. Bopp Ober das Albanesische in seinen verwandtschajilichenBeziehungen. Berlin 1855 (= Abhandlung derPreuBischen Akadamie der Wissenschaften, Philosopbisch-bistorischeKlasse).Bopp Conjugationssystem 1816 = F. Bopp Ober das Conjugationssystemder Sanskritsprache in Vergleichung mit jenem der griechischen,lateinischen, persischen und germanischen Sprache. Neben Episodendes Ramajan und Mahabharat in genauen metrischen Obersetzungenaus dem Originaltext und einigen Abschnilten aus den Veda 's. Frankfurt1816.

308 Bibliography and Key to Reference CitationsBopp-Symposion 1992 (1994] = Bopp-Symposium 1992, Humbo1dtUniversitat zu Berlin, documents from the 1992 conference given tocommemorate the 200'h birthday ofFranz Boppon 14 September 1991,edited by R. Sternemann. Heide1berg 1994.Boretzky Historische Linguistik 1977 = N. Boretzky Einfiihrung in diehistorische Linguistik. Reinbek, near Hamburg (Rowohlt) 1977.Bornemann / Risch Gr. Gr. 1978 = E. Bornemann and E. Risch GriechischeGrammatik. 2d ed. Frankfurt 1978. Additionally, compare W.Kastner Sprachgeschicht/iche Erliiulerungen zur Griechischen Grammatik.Frankfurt 1988.Brauer Slav. Sprachw. I 1961 11 1969 III 1969 = H. Brauer SlavischeSprachwissenschaft. 3 vols. Berlin 1961-1969: I (Einleilung, Laut/ehre)1961, TI (Formenlehre, 1. Teil) 1969, III (Formenlehre, 2nd Teil)1969 (= Sammlung Goschen Nr. 1191, 1192 and 1236).Brandenstein / Mayrhofer Allpersisch 1964 = W. Brandenstein and M.Mayrhofer Handbuch des Altpersischen. Wiesbaden 1964. See also,s. v. Mayrhofer Supplement 1978.Braunmiiller Skandinav. Sprachen 1991 = K. Braunmiiller Die skandinavischenSprachen im Oberblick. Tiibingen 1991 (= UTB 1635). Review:J. A. Hardarson in PFU 4 1998 p. 85-96.Braune / Ebbinghaus Got. Gr. 1981= W. Braune GOlische Grammatik.19th ed, revised by E. A. Ebhinghaus. Tilbingen 1981.Brixhe Grec anatolien 1987 = C. Brixhe Essai sur le grec analolien audebut de noire ere. new revised edition, Nancy 1987.Brixhe Koine I 1993 = La Koine grecque antique 1: Une langue introuvable,edited by C. Brixhe. Nancy 1993.Brixhe Pamphylie 1976 = C. Brixhe Le dialecle grec de Pamphylie,Documents el grammaire. Paris 1976.Brixhe Phonetique et phonologie 1996 = C. Brixhe Phonetique et phonologiedu grec ancien I, Quelques grandes questions. Paris 1996 (=Bibliolheque des Cahiers de I' lnslitul de Linguistique de Louvain Nr.82).Brixhe / Lejeune Pa/eo-phrygien 1984 = C. Brixhe Corpus des inscriptionspaleo-phrygiennes. 2 vols. Paris 1984.Brugmann Einfacher Satz 1925 = K. Brugmann Die Syntax des einfachenSalzes im Indogermanischen. Berlin / Leipzig 1925.Brugmann Grundriji 1 1897 !I-I 1906 11-2 1911 U-3 1916 = K. Brugmann[and B. Delbriick] Grundriji der vergleichenden Grammalik der indogermanischenSprachen. 2d revised edition. Strasbourg: vo!. I 1897(Einleilung, Laut/ehre: die ersle Hiilfte p. 1-622; die zweile Hiilfte p.Bibliography and Key to Reference Citations 309623- 1098; unchanged reprint of both parts 1930); vo!. II containingthree parts, namely 11-1 1906 (Allgemeines, Zusammensetzung [compoundforms], Nominalstiimme: p. 1-688), U-2 1911 (Zahlw6rter, Genera,Kasus- und Numerusbildung, Pronomina, Adjektiv, Adverbia,Priipositionen: p. 1-997) and !I-3 1916 (Verbumftnilum and infinitum,Partikeln im einfachen Satz: die erste Hiilfte p. 1-496, die zweileHiilfte p. 497-1052). reprint Berlin / New York 1967. Regarding vols.III - V, see s. v. Delbriick Vgl. Syntax.Brugmann Kurze vg!. Gramm. 1902-1 904 = K. Brugmann Kurze vergleichendeGrammatik der indogermanischen Sprachen. Strasbourg: I.Lieferung (Einleitung und Lautlehre: p. 1-280) 1902, 2nd Lieferung(Lehre von den Wortformen und ihrem Gebrauch: p. 281-622) 1903, 3.Lieferung (Lehre von den Salzgebilden und Sach- und W6rlerverzeichnis:p. 623-677) 1904. unchanged reprint Leipzig 1933, reprint 1970.Brugmann / Thumb Gr. Gr. 1913 = K. Brugmann Griechische Grammalik.4th ed. by A. Thumb. Munich 19\3.Brunner Aengl. Gr. 1965 = K. Brurmer Allenglische Grammalik nach derangelsiichsischen Grammatik von Eduard Sievers. 3rd ed. Tiibingen1965.Bryce Lycians 1986 = T. R. Bryce The Lycians in Literary and EpigraphicSources. Copenhagen 1986 (= volume I of the work by T. R.Bryce and J. Zahle: The Lycians, A Study of Lycian History and Civilisalion10 Ihe Conquest of Alexander the Great).Bryce Kingdom 1998 = T. Bryce The Kingdom of the Hilliles. Oxford1998.BSL = Bullelin de la societe de linguistique. Paris.Buchholz I Fiedler Alban. Gramm. 1987 = O. Buchholz and W. FiedlerAlbanische Grammatik. Leipzig 1987.Buck Comparative Grammar 1963 = C. D. Buck Comparative Grammarof Greek and Latin. 9'h ed. Chicago I London 1963.Buck Dict. of selecl. Syn. 1949 = C. D. Buck A Dictionary of SelecledSynonyms in the principal Indo-European Languages. Chicago 1949.Buck Greek Dialects 1955 = C. D. Buck The Greek Dialects, Grammar,Selected Inscriptions, Glossary. 2nd ed. Chicago / London 1955 (andreprint).BOhler Sprachtheorie 1934 = K. Biihler Sprachtheorie. Die Darslellungsfunktionder Sprache. Jena 1934.Burker! Griechische Religion 1977 = W. Burker! Griechische Religionder archaischen und klassischen Epoche. Stuttgart 1977. Translatedinto English as: Greek Religion. Cambridge I Mass. 1985.

310 Bibliography and Key to Reference CitationsBurkert Orientalisierende Epoche 1984 = W. Burkert "Die orientalisierendeEpoche in der griechischen Religion und Literatur," lecture givenon 8 May 1982, Heidelberg 1984 (= SbHA W 1984 I 1). English translation:"The Orientalizing Revolution, Near Eastern Influence on GreekCulture in the Early Archaic Age." Cambridge I Mass. 1992.Bussrnann Lexikon d. Sprachw. 2nd ed. 1990 = H. Bussrnann Lexikon derSprachwissenschaJt. 2nd, fully revised ed. Stuttgart 1990.Bynon Hist. Linguistics 1977 (also: Hist. Linguistik 1981 = Th. BynonHistorical Linguistics. Cambridge 1977 [and reprints]) = HistorischeLinguistik, Eine Eirifiihrung. Munich 1981 (revised and expanded editionof the English original). Review of the German edition: 1. Udolphin IF 96 1991 p. 258-262.Campanile Ricostruzione 1990 = E. Campanile La ricostruzione dellacultura indoeuropea. Pisa 1990.Cas et prepositions en grec ancien 1994 = Cas et prepositions en grecancien, Actes du colloque international de Saint-Etienne 1993, editedby B. Jacquinod. Saint-Etienne 1994.Cardona Them. Aorists 1960 = G. Cardona The Indo-European ThematicAorists. Ann Arbor 1960.Cario 1993 [1994] = La deciJrazione del cario, Alii del 10 Simposio Internazionale1993 in Rome, edited by M. E. Giannotta, R. Gusmani, L.Innocente, D. Marcozzi, M. Salvini, M. Sinatra, P. Vannicelli. Rome1994 (= Consiglio Nazionale delle Ricerche, Monograjie Scientifiche).Carting Lokole Kasus im Toch. 2000 = G. Carting Die Funktionen derlokolen Kasus im Tocharischen. Bertin I New York 2000.Carruba Palaisch 1970 = O. Carruba Das Palaische: Texte, Grammatik.Lexikon. Wiesbaden 1970 (= StBoT 10).Causatives and Transitivity 1993 = Causatives and Transitivity, edited byB. Comrie and M. Polinsky. Amsterdam I Philadelphia 1993 (= Studiesin Language Companion Series 23).CEG = Chronique d'etymologie grecque, see below, Chantraine DELG.Celtic Languages 1992 = The Celtic Languages, edited by D. Macaulay.Cambridge 1992 (from the series Cambridge Language Surveys).Celtic Languages 1993 = The Celtic Languages, edited by M. J. Ball togetherwith 1. Fife. London / New York 1993 (Routledge).CFS = Cahiers Ferdinand de SaussureChadwick Documents 1973 = 1. Chad wick Documents in MycenaeanGreek. 2nd ed. Cambridge 1973.Bibliography and Key to Reference Citations 311Chantraine DELG 1968-1980 = P. Chantraine Dictionnaire etymologiquede la langue grecque, Histoire des mots. Paris I 1968, IT 1970, III1974, IV I I 1977, IV I 2 1980. Since 1999, Chantraine DELG isavailable in a single-volume reprint of slightly reduced size, which includesa supplement (see below, CEG 1-3). Helpful addenda and corrigendamay be found in Ruijgh Scripta Minora I 1991 p. 571-632 andSzemerenyi Scripta Minora III 1987 p. 1559-1607, among others. Cfalso the ongoing Chronique d'etymologie grecque (CEG), edited by A.Blanc, Ch. de Lamberterie and J. L. Perpillou: CEG I in RPh 70 1996[1997] p. 103-138; CEG 2 in RPh 71 1997 [1998] p. 147-179; CEG 31998 [1999] in RPh 72 1998 [1999] p. 119-148; CEG 4 1999 [2000] inRPh 73 1999 [2000] p. 79-108.Chantraine Formation des noms 1933 = P. Chantraine La formation desnoms en grec ancien. Paris 1933 (and reprints).Chantraine Gramm. homo I 1958 11 1953 = P. Chantraine Grammairehomerique. 2 vols. Paris: I (Phonetique et Morphologie) 3rd ed. 1958,11 (Syntaxe) 1 st ed. 1953.Chantraine Morphologie 2nd ed. 1961 = P. Chantraine Morphologie historiquedu grec. 2nd ed. Paris 1961 (and reprints).CHO = The Hittite Dictionary of the Oriental Institute of the Universityof Chicago, edited by H. G. Giiterbock and H. A. Hoffuer. Currentlyavailable are the volumes: L-N 1989; P-I (pa - para) 1994; P-2 (para -(UZU)pattar A) 1995; P-3 «UZU)pattar A - putkiya-) 1997.CILL see below, IKLL.Clackson Armenian and Greek 1994 = J. Clackson The Linguistic Relationshipbetween Armenian and Greek. Oxford / Cambridge 1994. Review:Ch. de Lamberterie in Kratylos 42 1997 p. 71-78.CMCS = Cambrian Medieval Celtic Studies.Collinge Laws 1985 = N. E. Collinge The laws of Indo-European. Amsterdam/ Philadelphia 1985. An update through the year 1998 by thesame author may be found in nES 27 1999 p. 355-377.Coil. Myc. 1975 [1979] = Colloquium Mycenaeum, Actes du sixiemeColloque International 1975 in Chaumont, edited by E. Risch and H.Miihlestein. Neuchiitel l Geneva 1979.Colloque E. Benveniste I / IT 1983 [1984] = E. Benveniste aujourd'hui,Actes du colloque Tours 1983, edited by G. Serbat. 2 vols. Paris 1984.Colloque P. Chantraine 1989 [1992] = La langue et les textes en grecancien, Actes du colloque P. Chantraine Grenoble 1989, edited by F.Letoublon. Amsterdam 1992.

312 Bibliography and Key to Reference CitationsColloquium Caricum 1997 [1998] = Colloquium Caricum, Akten der InternationalenTagung iiber die karisch-griechische Bilingue von Kaunos1997 in Feusisberg near ZUrich, published as Kadmos 37 1998,Coil. Raur. 2 1991 = Colloquium Rauricum, vo!. 2, Zweihundert JahreHomer-Forschung, Riickblick und Ausblick, edited by J. Latacz. Stuttgart/ Leipzig 1991. Review: R. Schmitt in Kratylos 38 1993 p. 73-79.Compendium Ling. Iran. 1989 = Compendium Linguarum lranicarum,edited by R. Schmitt. Wiesbaden 1989.Comptetiv,es 1998 [1999] = "Les comptetives en grec ancien," ColloqueSaint-Etienne 1998, edited by B. Jacquinod. Saint-Etienne 1999.Convegno Udine (Restsprachen) 1981 [1983] = "Le lingue indoeuropee diframmentaria allestazione, Die indogermanischen Restsprachen", Aliidel Convegno della Societa Italiana di GIOllologia e della Idg. Gesell.1981 in Udine, edited by E. Vineis. Pisa 1983.Coseriu Synchronie, Diachronie 1974 = E. Coseriu Synchronie, Diachronieund Geschichte, Das Problem des Sprachwandels. Munich 1974.Cotticelli-Kurras Heth. 'sein ' 1991 = P. Cotticelli-Kurras Das hethitischeVerbum 'sein '. Heidelberg 1991 (= Texte der Hethiter 18).Cowgill Einleitung 1986 = W. CowgiU Einleitung, translated into Germanand with a bibliography by A. Bammesberger and M. Peters = 1.Halbband von vo!. [ of ldg. Gr. (see below). Heidelberg 1986.CR = Classical Review.CRAI = Comptes rendus de I 'Academie des Inscriptions et Belles-Lellres.Crystal Enzyklopadie 1995 = D. Crystal Die Cambridge-Enzyklopadie derSprache. Frankfurt / New York 1995 (= German translation of theoriginal in English, The Cambridge Encyclopedia of Language. Cambridge/ New York / Melbourne 1987).Darms Vrddhi 1978 = G. Darms Schwaher und Schwager, Hahn undHuhn. Die V rddhi-Ableitung im Germanischen. Munich 1978 (= MSSBeiheft 9, Neue Folge).De Boel Goal accusative 1988 = G. De Boel Goal accusative and objectaccusative in Homer. A contribution to the theory of transitivity.Briissel 1988;Debrunner Gr. Wortbildung 1917 = A. Debrunner Griechische Wortbildungslehre.Heidelberg 1917.Debrunner, see below under Schwyzer / Debrunner and Wackernagel /Debrunner.Degrassi Inscriptiones 1-11 1965-1972 = A. Degrassi Inscriptiones LatinaeLiberae Rei Publicae. 2nd ed. Florence: I (Nr. 1-503) 1965 (and re-Bibliography and Key to Reference Citations 313prints); 11 (Nr. 504-1277 and indexes) 1972. By the same author,Imagines. Berlin 1965.Delaunois Syntaxe 1988 = M. Delaunois Essai de syntaxe grecque classique,Reflexions et recherches. Leuven / Brussels 1988.Delbrilck Ablativ Localis Instrumentalis 1867 = B. Delbriick Ablativ LocalisInstrumentalis. Berlin 1867.Delbriick Ai. Syntax 1888 = B. Delbriick Altindische Syntax. HaUe 1888(reprint Darmstadt 1968).Delbriick Einleitung 1904 = B. Delbrilck Einleitung in das Studium derindogermanischen Sprachen. 4th ed. Leipzig 1904.Delbriick Gr. Syntax 1879 = B. Delbriick Die Grundlagen der griechischenSyntax. Halle 1879.Delbriick Synkretismus 1907 = B. Delbriick Synkretismus. Ein Beitrag zurgermanischen Kasuslehre. Strasbourg 1907.Delbriick Vgl. Syntax I 1893 II 1897 III 1900 = B. Delbriick VergleichendeSyntax der indogermanischen Sprachen. 3 parts, Strasbourg:I 1893, II 1897, III 1900. The three parts are included in Brugmann'sGrundrifJ as vols. Ill -V. Along with the GrundrifJ, they werereprinted in 1967.DELG see above, Chantraine DELG.DELL see below, Emout / Meillet DELL.Derniraj (B.) Alban. Etymologien 1997 = B. Derniraj Albanische Etymologien,Untersuchungen zum albanischen Erbwortschatz. Amsterdam /Atlanta 1997. Review: G. Bonnet in BSL 93 /2 1998 p. 256-262.Derniraj (S.) Albanisch 1993 = S. Derniraj Historische Grammatik deralbanischen Sprache. Vienna 1993.Denniston Greek Particles 1954 = 1. D. Denniston The Greek Particles.2nd ed., revised by K. J. Dover. Oxford 1954.Der Neue Pauly = DNP = Enzykloptidie der Antike, edited by H. Cancikand H. Schneider. Stuttgart: vo!. I (A - Ari) 1996; vo!. 2 (Ark - Ci)1997; vol. 3 (Cl - Epi) 1997; vo!. 4 (Epo - Gro) 1998; vo!. 5 (Gru -lug) 1998; vo!. 6 (lul - Lee) 1999; vol. 7 (Lef- Men) 1999; vo!. 8 (Mer- Op) 2000; vo!. 9 (Or - Poi) 2000; 10 (Pol - Sal) 2001. Further, volumesof the Rezeptions- und Wissenschafts-geschichte that have alreadyappeared include: vo!. 13 (A-Fo) 1999 and vo!. 14 (Fr - Ky)2000.Deutschsprachige Keltologen 1992 [1993] = Akten des Ersten Symposiumsdeutschsprachiger Keltologen 1992 Gosen near Berlin, edited byM. Rockel and SI. Zinuner. Tiibingen 1993. Further, see below, Keltologen-SymposiumII 1997 [I 999].

314 Bibliography and Key to Reference CitationsBibliography and Key to Reference Citations315Devoto Lingua di Roma 1940 = G. Devoto Storia della lingua di Roma.Bologna 1940 (reprint 1944, further 1983 foreword by L. Prosdocimi;translated into German by 1. Opelt as "Geschichte der Sprache Roms".Heidelberg 1968).DGE = Diccionario griego-espanol. Madrid, under the direction of F. R.Adrados. vo!. I (0 -aUO) 1980. The work is currently vo!. IV 1997and the word lhci>vuxo. Anejo I / 11 see above, Aura Jorro DMic;Anejo 1II = Repertorio bibliogrtijico de la lexicograjia griega 1998.Diachrony within Synchrony 1990 [1992] = Diachrony within Synchrony:Language History and Cognition, Papers from the International Symposium1990 in Duisburg, edited by G. KeIJerrnann and M. D. Morrissey.Frankfurt 1992 (= Duisburger Arbeiten zur Sprach- und Kulturwissenschaji14).Dialectes indo-aryennes 1986 [1989] = Dialectes dans les lilleraturesindo-aryennes, edited by C. Caillat. Paris 1989.Dialectologica Graeca 1991 [1993] = Dialectologica Graeca, Actas del11 Coloquio Internacional de Dialectologia Griega Madrid 199 I, editedby E. Crespo, 1. L. Garcia Rarn6n, A. Striano. Madrid 1993. Concerningthe third colloquium, ct: Kata ditilekton 1996 [1999] below.The texts of the first colloquium (Rencontre internationale) in Nancy /Pont-a-Mousson were published in Verbum ID 1987.Dialektologie 1 1982 2 1983 = Ein Handbuch zur deutschen und allgemeinenDialektforschung, edited by W. Besch, U. Knoop, W.Putschke, H. E. Wiegand. Berlin / New York. 2 half-vols.: I 1982; 21983 (= HSK I. I and 1.2).Diehl All/at. Inschriften 1965 = E. Diehl Alllateinische Inschriften, milIndizes. 5th ed. Berlin 1964 (= Kleine Texle, commented by H. Lietzrnann,Nr.38 / 40).Di Giovine see below, Giovine.Dionysios Thrax (2" ' century B.C.) Grammalik = La grammaire de Denysle Thrace, Iraduile el annolee par Jean Lallol. 2nd ed. Paris 1998.Disterheft Injinitive 1977 = D. Disterheft The Syntax of the Irifinilive inIndo-European: Evidence from Indo-1ranian, Celtic, and Hit/ite. Ph.D. dissertation. University of California, Los Angeles 1977 (microfilmAnn Arbor 1979).DNP = see above, Der Neue Pauly.Dobias-Lalou Cyrime 2000 = C. Dobias-Lalou Le dialecte des inscriptionsgrecques de Cyrene. Paris 2000.Dressier Verb. Pluralitat 1968 = W. DressIer Sludien zur verbalen Pluralilat,Ilerativum, Dislributivum, Inlensivum in der allgemeinenGrammalik, im Lateinischen und Helhilischen. Vienna 1968.dtv-Allas Dt. Sprache 1998 = dtv-Atlas Deutsche Sprache 1998. Munich1 st ed. 1978, 12. ed. 1998.Dubois Arcadien 1986 = L. Dubois Recherches sur le dialecte arcadien.Louvain-La-Neuve 1986: I (Grammaire), IT (Corpus dialectal), ill(Noles, Index, Bibliographie).Duden Grammatik 1995 = Der Duden in 12 Banden, vo!. 4. Grammatikder deulschen Gegenwartssprache, 5th ed., fully revised and expanded,edited and revised by G. Drosdowski, together with P. Eisenberg, H.Gelhaus, H. Henne, H. Sitta and H. WeUrnann. Mannheirn / Leipzig /Vienna / ZUrich 1995.Egetmeyer W6rterbuch 1992 = M. Egetmeyer W6rterbuch zu den Inschriftenim kyprischen Syllabar. Berlin 1992.Egli Gelenkheteroklisie 1954 = J. Egli Heteroklisie im Griechischen milbesonderer Berucksichtigung der Falle von Gelenkheleroklisie. Zurich1954.Eichner Numeralia 1982 = H. Eichner Studien zu den indogermanischenNumeralia (2-5). (unpublished) postdoctoral thesis. Regensburg 1982.Einhauser Junggrammatiker 1989 = E. Einhauser Die Junggrammatiker,Ein Problem fur die Sprachwissenschajisgeschichtsforschung. Trier1989.Ernout Recueil 1947 = A. Ernout Recueil de textes latins archarques,Textes epigraphiques et lilleraires. 2nd ed. Paris 1947 (and reprints).Ernout / Meillet DELL 1959 = A. Emout and A. Meillet Diclionnaireetymologique de la langue latine. Histoire des mots. 4 t h ed., revised,corrected, and augmented with an index. Paris 1959.Ethnogenese 1985 = Studien zur Ethnogenese, A bhandlungen derRheinisch-Westfalischen Akademie der Wissenschajien vo!. 72.Opladen 1985.Etruschi e Roma 1979 [1981] = GIi Etruschi e Roma, Atti dell' incontrodi studio in onore di Massimo Pallottino Rome 1979, edited by G.Colonna. Rome 1981. Review: D. Steinbauer in GGA 235 1983 p.210-232.Etrusker 1985 = Die Etrusker, edited by M. Christofani. Stuttgatt / ZUrich1985.Etter Fragesatze 1985 = A. Etter Die Fragesatze im Sgveda. Berlin / NewYork 1985. Review: J. S. Klein in Kratylos 33 1988 p. 79-83.

316 Bibliography and Key to Reference CitationsBibliography and Key to Reference Citations 317Etymologisches Worterbuch 1983 = Das etymologische Worterbuch,Fragen der Konzeption und Gestaltung, edited by A. Bammesberger.Regensburg 1983 (= Eichstdtter Beilrdge 8).Euler Gemeinsamkeiten 1979 = W. Euler Indoiranisch-griechische Gemeinsamkeitender Nominalbildung und deren indogermanischeGrundlagen. Innsbruck 1979 (= ms Nr. 30).Evidence Jor Laryngeals 1965 = Evidence Jo r Laryngeals, edited by W.Winter. London / The Hague / Paris 1965.EW Ahd = Etymologisches Worterbuch des Althochdeutschen. Gottingen /ZUrich. I (-a - bezzisto) 1988 by A. L. Lloyd and O. Springer; IT (bi ezzo) 1998 by A. L. Lloyd, R. Liihr and O. Springer. A separate bookletof word indexes belongs to vol. 1.EW Aia see below, Mayrhofer EW Aia.Explanation in Historical Linguistics 1992 = Explanation in HistoricalLinguistics, edited by G. W. Davis and G. K. Iverson. Amsterdam /Philadelphia 1992 (= Amsterdam Studies in the Theory and History oJLinguistic Science, IV Current Issues in Linguistic Theory 84).Fachtagung Berlin 1983 [1985] = Grammatische Kategorien, Funktionund Geschichte, Akten der VII. Fachtagung der Idg. Gesellschaft 1983in Berlin, edited by B. Schlerath. Wiesbaden 1985.Fachtagung Bern 1969 [1973] = Indogermanische und allgemeineSprachwissenschaft, Akten der IV Fachtagung der Idg. Gesellschaft1969 in Bern, edited by G. Redard. Wiesbaden 1973.Fachtagung Innsbruck 1961 [1962] = Akten der n. Fachtagung der Idg.GeseUschaft 1961 in Innsbruck. Innsbruck 1962.Fachtagung Innsbruck 1996 [1998] = Sprache und Kultur der Indogermanen,Akten der X Fachtagung der Idg. Gesellschaft 1996 in Innsbruck.Innsbruck 1998 (= ms Nr. 93).Fachtagung Leiden 1987 [1992] = Relative Chronologie, Akten der VlIIFachtagung der Idg. Gesellschaft 1987 in Leiden, edited by R. S. P.Beekes. Innsbruck 1992. Review: B. Forssman in Kratylos 39 1994 p.48-55 (WIth a useful glossary [po 53-55] in place of an index).Fachtagung Regensburg 1973 [1 975] = Flexion und Wortbi/dung, Aktender V Fachtagung der Jdg. Gesellschaft 1973 in Regensburg, edited byH. Rix. Wiesbaden 1975.Fachtagung Tocharisch Berlin 1990 [1994] = Tocharisch, Akten derFachtagung der Idg. Gesellschaft 1990 in Berlin, edited by B. Schlerath.Reikjavik 1994 (= TIES, Suppl. Ser. 4).Fachtagung Wien 1978 [1980] = Lautgeschichte und Etymologie, Aktender VI Fachtagung der Idg. Gesellschaft 1978 in Vienna, edited by M.Mayrhofer, M. Peters, O. E. Pfeiffer. Wiesbaden 1980.Fachtagung Zurich 1992 [1994] = Fruh-, Mittel-, Spdtindogermanisch,Akten der IX. Fachtagung der Jdg. Gesellschaft 1992 in Ziirich, editedby G. E. Dunkel, G. Meyer, S. Scarlata, Chr. SeidJ. Wiesbaden 1994.Review: J. S. K1ein in Kratylos 42 1997 p. 24-32.Feist Got. Worterbuch 1939 = S. Feist Vergleichendes Worterbuch dergotischen Sprache. 3" revised ed. Leiden 1939. See also, LehrnannGothic Etymological Dictionaryl986.Feist Indogermanen 1913 = S. Feist Kultur, Ausbreitung und Herkurift derIndogermanen. Berlin 1913.Floreant Studia Mycenaea 1995 [1999] = Floreant Studia Mycenaea, Aktendes X Internationalen Mykenologischen Colloquiums Salzburg1995, edited by S. Deger-Jalkotzy, S. Hiller and O. PanagJ. 2 vols., Vienna1999 (= Denkschriften 274, Osterr. Ak. der Wissenschaften).FoL = Folia Linguistica.Formazione dell ' Europa Iinguistica 1993 = La Jormazione dell ' EuropaIinguistica, Le Iingue d'Europa Ira la fine del I e del 11 millennio, editedby E. Banfi. Florence 1993. Including, among others, "Le Iinguegermaniche" by M. Meti; "Le lingue slave" by A. Cantarini; "Le Iinguebaltiche" by P. U. Dini; "Le Iingue celtiche" by P. Cuzzolin; La linguagreca by E. Banfi; "La lingua albanese" by S. Demiraj.Forssman Pindar 1966 = B. Forssman Untersuchungen zur Sprache Pindars.Wiesbaden 1966.Forssman, see also, Hoffinann / Forssman.Fraenkel Lit. etym. Worterbuch 1962-1965 = E. Fraenkel Litauisches etymologischesWorterbuch. Heidelberg 1962-1965.Fraenkel Syntax der lit. Kasus 1928 = E. Fraenkel Syntax der IitauischenKasus. Kaunas 1928.Fraenkel Syntax der lit. Postposilionen und Prdpositionen 1929 = E.Fraenkel Syntax der litauischen Postpositionen und Prdpositionen.Heidelberg 1929.Friedrich (J.) Elementarbuch I 1960 = J. Friedrich Hethitisches Elementarbuch.I (KurzgeJajJte Grammatik). 2nd ed. 1960.Friedrich (1.) Kleinas. Sprachdenktndler 1932 = J. Friedrich KleinasiatischeSprachdenktndler. Berlin 1932.Friedrich (1.) / Kammenhuber HW = 1. Friedrich Hethitisches Worterbuch.2nd ed. by A. Kammenhuber fully revised on the basis of edited Hittite

318 Bibliography and Key to Reference CitationsBibliography and Key to Reference Citations319texts. Heidelberg. installment 1 1974. Status at the end of 1998: instalhnent13 (I::f) 1998.Friedrich (p.) Syntax 1975 = P. Friedrich Proto-Indo-European Syntax:The Order of Meaningful Elements. Washington 1975 (= JIES, MonographI).Frigi e Frigio 1995 [1997] = Frigi e Frigio, Alii del 1° Simposio InternazionaleRome 1995, edited by R. Gusmani, M. Salvini, P. Vannicelli.Rome 1997 (= Consiglio Nazionale delle Ricerche, Monografie Scientifiche).Frisian Runes 1994 [1996] = Frisian Runes and Neighbouring Traditions,Proceedings of the First International Symposium on Frisian Runes atthe Fries Museum 1994 in Leeuwarden, edited by T. Looijenga and A.Quale Amsterdam 1996 (= Amsterdamer Beitrage zur alteren Germanistik,Band 45).Frisk GEW 1960-1972 = H. Frisk Griechisches etymologisches W6rterbuch.Heidelberg 3 vols. 1960-1972: I (A - Ko) 1960, II (Kp - Q) 1972,111 (Nachtrage, Wortregister, Corrigenda, Nachwort) 1972. also reprints.Fritz Lokalpartikel 1997 = M. Fritz Die syntaktischen und semantischenRelationen der Lokalpartikeln mit drei Kasus bei Homer. Berlin (dissertationat the Freie Universitat) 1997.FS = Festschrifi (commemorative publications): For the sake of simplicity,only the name of the honored person shall be given, along with thedate of publication. A longer list of Festschrifien may be fo und inMayrhofer EW Aia I p. XXV - XXX and 11 p. XIII - XV. A few newerFestschrifien are additionally listed here.FS Beekes 1997 = Sound Law and Analogy, Papers in Honor of R. S. P.Beekes, edited by A. Lubotsky. Amsterdam / Atlanta 1997.FS Belardi I 1994 = Miscellanea di studi linguistici in onore di WaiterBelardi, edited by P. Cipriano, P. di Giovine, M. Mancini. Rome vol. I(Linguistica indoeuropea e non indoeuropea) 1994.FS Dihle 1993 = Philanthropia kai Eusebeia, Festschrifi fur AlbrechtDihle zum 70. Geburtstag, edited by G. W. Most, H. Petersmann andA. M. Ritter. Gottingen 1993.FS Forssman 1999 = gering und doch von Herzen, 25 indogermanistischeBeitrage B. Forssman zum 65. Geburtstag, edited by J. Habisreitinger,R. Plath, S. Ziegler. Wiesbaden 1999.FS Hamp 1990 = Celtic Language, Celtic Culture: A Festschrifi for EricP. Hamp, edited by A. T. E. Matonis and D. F. Melia. Van Nuys, California 1990.FS Hamp 1/11 1997 = Festschrifi for Eric P. Hamp. Washington 1997 (=JIES Monographs 23 and 25).FS Hoenigswald 1987 = Festschrifi fo r Henry M. Hoenigswald, edited byG. Cardona and N. H. Zide. TUbingen 1987.FS Knobloch 1985 = Sprachwissenschaftliche Forschungen, Festschrififiir J Knobloch, edited by H. M. Qlberg, G. Schrnidt. Innsbruck 1985(= IBS 23).FS Kuiper 1968 = Pratidiinam, Indian, Iranian and Indo-European StudiesPresented to Franciscus Bernardus Jacobus KUiper on his sixtiethbirthday, edited by J. C. Heesterrnan, G. H. Schokker, V. I. Subramoniam.The Hague / Paris 1968.FS Lejeune 1978 = Etrennes de septantaine, Travaux de linguistique et degrammaire comparee offerts a Michel Lejeune, edited by a group of hisstudents. Paris 1978 (= Etudes et Commentaires 91).FS Meid "60 1989 = Indogermanica Europea, Festschrift fiir W Meidzum 60. Geburtstag, edited by K. Heller, O. Panagl, J. Tischler. Graz1989.FS Meid "70 1999 = Studia Celtica et Indogermanica, Festschrift fiir WMeid zum 70. Geburtstag, edited by P. Anreiter, E. Jerem. Budapest1999.FS Narten 2000 = Anusantatyai, Festschrift J Narten, edited by A.Hintze, E. Tichy. Dettelbach 2000 (= MSS, Beiheft 19. NF).FS Neumann 1982 = Serta Indogermanica, Festschrift fiir Giinter Neumannzum 60. Geburtstag, edited by J. Tischler. Innsbruck 1982.FS Ollen 1973 = Festschrift Heinrich Ollen, edited by E. Neu and Chr.Roster. Wiesbaden 1973.FS Ollen 1988 = Documentum Asiae Minoris Antiquae. Festschrift furHeinrich Ot/en zum 75. Geburtstag, edited by E. Neu and Chr. Riister.Wiesbaden 1988.FS Palmer 1976 = Studies in Greek, Italic, and Indo-European Linguisticsoffered to Leonard R. Palmer, edited by A. Morpurgo Davies andW. Meid. Innsbruck 1976.FS Puhvel I 1997 = Sludies in Honor of J Puhvel. 1: Ancient Languagesand Philology, edited by D. Disterheft, M. Huld and J. Greppin.Washington 1997 (= JIES Monograph 20).FS Ramat 1998 = Ars Linguistica, Studi offerti a Paolo Ramat, edited byG. Bernini, P. Cuzzolin, P. Molinelli. Rome 1998.FS Risch 1986 = o-o-pe-ro-si, Festschrift fur Ernst Risch zum 75. Geburtstag,edited by A. Etter. Berlin / New York 1986.

320 Bibliography and Key to Reference CitationsBibliography and Key to Reference Citations 321FS Rix 1993 Indogermanica et Italica, Festschrift fur Helmut Rix zum65. Geburtstag, edited by G. Meiser. Innsbruck 1993.FS Schlerath 1992 [1994] Die Indogermanen und das Pferd, Akten desInternationalen interdisziplinaren Kolloquiums an der Freien UniversittitBerlin, 2nd-3. Juli 1992 [ Festschrift fur Bernfried Schlerath],edited by B. Hiinsel and SI. Zimmer. Budapest 1994. Cf. report by SI.Zimmer in Ethnographisch-archtiologische Zeitschrift 33 1992 [1993]p. 297-301.FS Schmeja 1998 Wort - Text - Sprache und Kultur, Festschrift furHans Schmeja zum 65. Geburtstag, edited by P. Anreiter and H. M.O lberg. Innsbruck 1998 ( lBK 103).FS Schmid 1999 Florilegium Linguisticum, Festschrift fur Wolfgang P.Schmid, edited by E. Eggers, 1. Becker, J. Udolph, D. Weber. Bern,Frankfurt am Mainet al. 1999.FS (K. H.) Schmidt 1994 Indogermanica et Caucasica, Festschrift furKarl Horst Schmidt zum 65. Geburtstag, edited by R. Bielmeier and R.Stempel. Berlin / New York 1994.FS Seebold 1999 Grippe, Kamm und Eu/enspiegel, Festschrift fur EImarSeebold zum 65. Geburtstag, edited by W. Schindler and J. Untermann.Berlin 1999.FS Stimm 1982 Fakten und Theorien, Beitrage zur romanischen undallgemeinen Sprachwissenschajt, Festschrift fur Helmut Stimm zum 65.Geburtstag, edited by S. Heinz and U. Wandruszka. Tiibingen 1982 (TBL 191).FS Strunk 1995 Verba et structurae, Festschrift fur Klaus Strunk zum65. Geburtstag, edited by H. Hettrich, W. Hock, P.-A. Mumm and N.Oettinger. Innsbruck 1995. Review: R. S. P. Beekes in Kratylos 421997 p. 36-39.FS Szemerenyi *65 1 / II 1979 Studies in Diachronic, Synchronic, andTypological Linguistics, Festschrift for Oswald Szemerenyi, edited byB. Brogyanyi. 2 vols. Amsterdam 1979.FS Szemerenyi *75 I / IT 1992, IIJ 1993 Prehistory, History and Historiographyof Language, Speech and Linguistic Theory, Papers inHonor of Oswald Szemerenyi I, edited by B. Brogyanyi; HistoricalPhilology: Greek, Latin, and Romance, Papers in Honor of OswaldSzemerenyi 11, edited by B. Brogyanyi and R. Lipp; ComparativeHistorical Linguistics: Indo-European and Finno-Ugric, Papers inHonor of Oswald Szemerenyi Ill, edited by B. Brogyanyi and R. Lipp.3 vols. Amsterdam / Philadelphia 1992-1993 ( Current Issues in Lin-guistic Theory, Nr. 64, 87 and 97). Review: M. Egetmeyer in PFU (PFU) I 1994 [1995] p. 47-53 and p. 55-61; 4 1998 p. 69-76.FS Thomas 1988 Studia Indogermanica el Slavica. Feslgabe fur WernerThomas zum 65. Geburtstag, edited by P. Kosta. Munich 1988.FS Untermann 1993 Sprachen und Schriften des antiken Mittelmeerraumes,Festschrift fur Jurgen Untermann, edited by F. Heidermanns,H. Rix and E. Seebold. Innsbruck 1993 ( lBS 78).FS Watkins 1998 Mir Curad, Studies in Honor of Calvert Watkins, editedby J. Jasanoff, H. Craig Melchert, L. Oliver. Innsbruck 1998 (lBS 92).FS Zoic 1989 Phonophilia, Untersuchungen zu Phonetik und Phonologie,F. Zaic zum 60. Geburtstag, edited by W. Grosser et al. Salzburg1989.Fulk Quantitative Ablaut 1986 R. D. Fulk The Origins of IndoEuropean Quantitative Ablaut. Innsbruck 1986 ( lBS 49). Review:G. Schmidt in Kratylos 32 1987 p. 37-46.Gaedicke Accusativ im Veda 1880 C. Gaedicke Der Accusativ im Veda.Breslau 1880.Gamkrelidze / Ivanov lE and IEs 1995 T. V. Gamkrelidze and V. V.Ivanov Indo-European and the Indo-Europeans. A Reconstruction andHistorical Analysis of a Proto-Language and a Proto-Culture. 2 vols.Berlin / New York 1995. The work is the English translation of theoriginal Russian Indoevropejskij jazyk i indoevropejcy that was publishedin 1984. The translation, dated 1995, reflects the status of researchin 1984. Review: M. Mayrhofer in Kralylos 42 1997 p. 21-24;J. Gippert in BNF 33 1998 p. 39-54.Gaters Lettische Syntax 1993 A. Gaters Lellische Syntax, Die Dainas,posthum edited by H. Radtke. Frankfurt am Main et al. 1993.Geiger Piili 1916 W. Geiger Piili, Literatur und Sprache. Strasbourg1916.Geldner RV Obersetzung 1951-1957 Der Rig- Veda, Aus dem Sanskritins Deutsche ubersetzt und mit einem laufenden Kommentar verse henvon K. F. Geldner. 4 vols. Leipzig 1951-1957. Although the translationwas already complete in 1923, circumstances were such that thethe work was first published only after the Second World War, cf. thefo reword by 1. Nobel (vol. IV p. V-VII).Germanenprobleme in heutiger Sicht 1986 Germanenprobleme in heutigerSicht, edited by H. Beck. Berlin / New York 1986.GGA G6ttingische Gelehrte Anzeigen.

322 Bibliography and Key to Reference CitationsGiacomelli Lingua falisca 1963 = G. Giacomelli La lingua falisca. Florence1963.Giacomelli Lingua latina 1993 = R. Giacomelli Storia della lingua latina.Rome 1993 (in the series Guide allo studio de/la Civilita romana IV .1)Giannakis Reduplicated Presents 1997 = G. K. Giannakis Studies in theSyntax and Semantics of the Reduplicated Presents of Homeric Greekand Indo-European. Innsbruck 1997 (= IBS 90). Review: E. F. Tuckerin Kratylos 45 2000 p. 111-116.Giovine Perfetto I 1990 II I III 1996 = P. Di Giovine Studio sui Perfettoindoeuropeo. Rome: Parte I 1990 (La funzione originaria del perjettostudiala nella documentazione delle lingue storiche); Parte 11 1996 (Laposizione del perfetto all' interno del sistema verbale indoeuropeo);Parte III 1996 (Indici); also by the same author, Le lingue anatoliche eil perjetto indoeuropeo: Una "petitio principii " ? in FS W. Belardi l.Rome 1994 p. 113-130. Review: J. A. Hardarson in Kratylos 46 2001p. 36-44.Gippert Infinitive 1978 = J. Gippert Zur Syntax der infinitivischenBildungen in den indogermanischen Sprachen. Frankfurt I Bern I LasVegas 1978.Gippert Iranica 1993 = J. Gippert iranica Armeno-Iberica, Studien zu deniranischen Lehnortern im Armenischen und Georgischen. 2 vols. Vienna1993 (= SbOA W Nr. 606).Giotta = Giotta. Zeitschriji fur griechische und lateinische Sprache. Gottmgen.GmGr Memoire 1986 = R. GmGr Das Schicksal von F. de Saussures"Memoire". Eine Rezeptionsgeschichte. Bern 1986. Review: M. Mayrhoferm Kratylos 33 1988 p.I-15 = Kleine Schrijien II 1996 p. 271-285.Godel Classical Armenian 1975 = R. Godel An introduction to the Studyof Classical Armenian. Wiesbaden 1975.Gonda Indo-European Moods 1956 = J. Gonda The Character of theIndo-European Moods. Wiesbaden 1956. Review: E. Risch in Gnomon33 1961 p. 174-178; J. Kurylowicz and Hj. Seiler in Kratylos 1 1956 p.123-135.GORILA I - V 1976-1985 = L. Godart, J.-P. Olivier Recueil des inscriptionsen lineaire A. Paris: I (Tablettes Mitees avant 1970) 1976; 1I(Nodules, ScelLes et rondelles Mites avant 1970) 1979; III (Tablettes,nodules et rondelles edites en 1975 et 1976) 1976; IV (Autres documents)1982; V (Addenda, Corrigenda, Concordances, 1ndex etBibliogr aphy and Key to Reference Citations 323planches des signes) 1985 (=Ecole Fram;aise d'Athimes, Etudes CretoisesNr. 21, 1 -5).GotO l. Prtisensklasse 1987 = T. Goto Die "I. Prtisensklasse" im Vedischen.Untersuchung der voUslufigen themalischen Wurzelprtisentia.Vienna 1987. The 2nd ed. (1996) includes supplements and improvements.Review der 2nd ed.: SI. W. Jamison in Kratylos 34 1989 p. 59-65.Goto Materialien Nr. 1-3 1990 Nr. 4-7 1991 Nr. 8-15 1993 Nr. 16-291997 = T. Goto, Materialien zu einer Liste altindischer Verbalformen.Bulletin of the National Museum of Ethnology, Osaka (Japan) : Nr. 1-3(1. am', 2. ayli, 3. asls) in Vol. 15 14 1990 p. 987-1012; Nr. 4-7 (4.doghldughldohlduh, 5. savlsu, 6. l sav i lsu, 7. ' (sav'l)su in Vol. 16 / 31991 p. 681-707; Nr. 8-15 (8. ardlrd, 9. i$, 10. Uks, 11. e$li$, 12. elM,13. 0k/oc/uc, 14. ka(l, 15. vaks1uks) in VoLl8 I I 1993 p. 119-141; Nr.16-29 (16. chad, 17. chandlchad, 18. chardlchrd, 19. dagh/dhag, 20.dve$ldvi$, 21. bandhlbadh, 22. 1 man, 23. ' man, 24. mna, 25. l yavlyu,26. ' yavlyu, 27. san', 28. star/str, 29. st/stf) in Vol. 22 I 4 1997[1998] p. 1001-I059.Grammatica ittita 1992 = Per una grammatica iIIita, Towards a HitliteGrammar, edited by O. Carruba. Pavia 1992 (= Studia Mediterranea7).Grassmann Worterbuch 1873 = H. Grassmann Worterbuch zum Rig- Veda.Leipzig 1873 (and reprints). 6'" revised, updated ed. by M. Kozianka1996. Review of the 6th ed.: Th. Zehnder in PFU 4 1998 p. 77-84.Graz (125 Jahre Idg.) 2000 = 125 Jahre 1ndogermanistik in Graz, FestbandanltifJlich des 125jtihrigen Bestehens der Forschungs-einrichtung"Indogermanislik" an der Karl-Franzens-Universittil Graz, edited byM. Ofitsch and Chr. Zinko, Graz 2000.Greek Language in Cyprus 1988 = The History of the Greek Language inCyprus, Proceedings of an International Symposium Larnaca 1986,edited by J. Karageorghis I O. Masson. Nicosia 1988.Greek Particles 1996 [1997] = New Approaches 10 Greek Particles, Proceedingsof Ihe Colloquium Amsterdam 1996, edited by A. Rijksbaron.Amsterdam 1997.Greek Personal Names I 1987 1I 1994 III A 1997 = A Lexicon of GreekPersonal Names, edited by P. M. Fraser, E. Matthews. Oxford. vol. I(The Aegean 1slands, Cyprus, Cyrenaica) 1987; vol. 11 (Attica, editedby M. J. Osborne, S. G. Byme) 1994; vol. III A (The Peloponnese,Western Greece and Magna Graecia) 1997.

324Bibliography and Key to Reference CitationsBibliography and Key to Reference Citations 325GrojJere altkelt. Sprachden/analer 1993 (1996) = Die grojJeren altkeltischenSprachden/analer, Akten des Kolloquiums 1993 in Innsbruck,edited by W. Meid and P. Anreiter. Innsbruck 1996 (= IBK,Sonderheft 95). Review: J. Uhlich in Kratylos 44 1999 p. 144-154.Gr. Phi/ologie 1997 = Einleitung in die griechische Philologie, edited byH.-G. Nesselrath. Stuttgart I Leipzig 1997.GS = Gedenkschrift (memorial publication); a listing may be fo und in MayrhoferEW Aia I p. XXXII f, II p. XVII.GS Brandenstein 1968 = Studien zur Sprachwissenschajt und Kulturkunde,Gedenkschrift fur W Brandenstein, edited by M. Mayrhofer.Innsbruck 1968 (= IBK Nr. 14).GS Cowgill 1987 = Studies in Memory of Warren Cowgil/ (1929-1985),Papers from the Fourth East Coast Indo-European Conference 1985,edIted by C. Watkins. Berlin I New York 1987 (= Untersuchungen zurIndogermanischen Sprach- und Kulturwissenschajt, N.F. 3).GS Katz 2001 = Fremd und Eigen, Untersuchungen zu Grammatik undWortschatz des Uralischen und Indogermanischen, in memoriamHartmUl Katz, edited by H. Eichner, P.-A. Mumrn, O. Panagl, E. Winkler.Vienna 200 I.GS Kronasser 1982 = Investigationes Philologicae et Comparativae.Gedenkschrift fur Heinz Kronasser. Wiesbaden 1982.GS Kurylowicz I I II 1995 = Kurylowicz Memorial Volume, vo!. I, editedby W. Smoczynski. Cracow 1995. vo!. IJ was published as vo!. 4 1995of Linguistica Ballica. Review by I: A. Christol in BSL 92 I 2 1997 p.131-135.GS Schindler 1999 = Compositiones indogermanicae in memoriam JochemSchindler, edited by H. Eichner and H. Chr. Luschiltzky. Prague1999.GS van Windekens 1991 = Studia etymologica indoeuropea, Memoriae A.J van Windekens (1915-1989) dicata, edited by L. Isebaert. Leuven1991.Guiraud Phrase nominale 1962 = Ch. Guiraud La phrase nominale engrec. Paris 1962.Gusmani Lyd. Worterbuch 1964 = R. Gusmani Lydisches Worterbuch.Heidelberg 1964.Gusmani Lyd. Wh. Erg. 1986 = R. Gusmani Lydisches Worterbuch. Erganzungsband.Heidelberg 1986.Haas Heth. Religion 1994 = V. Haas Geschichte der hethitischen Religion.Leiden I New York I Koln 1994.Hackstein Sigmat. Prasensstammbildungen 1995 = O. Hackstein Untersuchungenzu den sigmatischen Prasensstammbi/dungen des Tocharischen.Gottingen 1995 (= HS, Erganzungshejt Nr. 38). Review: I. H.W. Penney in Kratylos 43 1998 p. 92-96; H. Craig Melchert in TIES 92000 p. 145f.; G.-I. Pinault in BSL 95 I 2 2000 p. 157-163.Hahn Naming-Constructions 1969 = E. A. Hahn Naming-Constructions insome Indo-European Languages. Ann Arbor 1969.Hahn Subjunctive and Optative 1953 = E. A. Hahn Subjunctive and Optative,Their origin asfutures. New York 1953.Hajnal Lyk. Vokalismus 1995 = l. Hajnal Der Iykische Vokalismus:Methode und Erkenntnisse der vergleichenden anatolischen Sprachwissenschajt,angewandt auf das Vokalsystem einer Kleincorpussprache.Graz 1995. Review: G. Neumann in HS II1 1998 p.372-376.Hajnal Myk. Kasussystem 1995 = 1. Hajnal Studien zum mykenischen K sussystem. Berlin / New York 1995. ReVIew: N. Guilleux (-Maunce) mBSL 92 / 2 1997 p. 200-216.Hajnal Myk. u. homo Lexikon 1998 = l. Hajnal Mykenisches und homerischesLexikon, Ubereinstimmungen, Divergenzen und der Versuch einerTypologie. Innsbruck 1998 (= IBS, Vortrage und kl. Schriften 69)..Hajnal Sprachschichten 1997 = I. Hajnal Sprachschichten des MykemschenGriechisch. Salamanca 1997 (= Minos Sup/. 14).Hamm Sappho und Alkaios 1957 = E.-M. Hamm Grammatik zu Sapphound Alkaios. Berlin 1957.Handbuch der Onomastik 1 1995 = Namenforschung, Name Studies, Lesnoms propres, Ein internationales Handbuch zur Onomastik, edited byE. Eichler, G. Hilly, H. Lamer, H. Steger, L. Zgusta. part I. Berlin /New York 1995.Happ Grundfragen 1976 = H. Happ Grundfragen einer Dependenzgrammatikdes Lateinischen. Gattingen 1976.Hardarson Wurzelaorist 1993 = J. A. Hardarson Studien zum indogermanischenWurzelaorist und dessen Vertretung im Indoiranischen undGriechischen. Innsbruck 1993. Review: K. Strunk in Kratylos 39 1994p. 55-68; M. Peters in Idg. Chr. 35 Nr. A 789.Haudry Cas en w!dique 1977 = J. Haudry L'emploi des cas en w!dique.Introduction a I'etude des cas en indo-europeen. Lyon 1977. ReVIew:G. Cardona in Kratylos 23 1978 [1979) p. 71-82..Haudry Les indo-europeens 1985 = J. Haudry Les indo-europeens. Pans1 st ed. 1981, 2nd ed. 1985 (= Collection Que sais-je ? Nr. 1965).Haudry L'indo-europeen 1979 = I. Haudry L'indo-europeen. Paris 1979 (=Collection Que sais-je ? Nr. 1798).

326Bibliography and Key to Reference CitationsBibliography and Key to Reference Citations327Hauri -ena 1963 = Ch. Hauri Zur Vorgeschichte des Ausgangs -ena desInstr. Sing. der a-Sttimme des Altindischen. Gottingen 1963 (= ZVSErganzungshejt Nr. 17).Hauri FUlur 1975 = H. W. Hauri Kontrahiertes und sigmatisches FUlur.Einjlusse von Lautstruktur und Aktionsart auf die Rildung des griechlschenFuturs. Gottingen 1975 (= ZVS, Erganzungshejt Nr. 24).RevIew: C. J. Ruijgh in Kratylos 20 1975 [1977] p. 82-91 (= ScriptaMmora T 1991 p. 368-377); F. M. 1. Waanders in Mnemosyne 33 1980p. 369-374.Hauschild, see below, Thumb / Hauschild.Haverling sco-Verbs 2000 = G. Haverling On sco- Verbs, Prefixes andSemantic Functions. Goteborg 2000.Havers Erklarende Syntax 1931 = W. Havers Handbuch der erklarendenSyntax, Ein Versuch zur Erforschung der Bedingungen und Triebkrtiftein Syntax und Stilistik. Heidelberg 1931.Havers Kasussyntax 191 1 = W. Havers Untersuchungen zur Kasussyntaxder indogermanischen Sprachen. Strasbourg 1911.Havers Sprachtabu 1946 = W. Havers Neuere Literatur zum Sprachtabu.Vienna 1946.Hawkins Corpus I 2000 / 2 1999 = D. Hawkins Corpus of HieroglyphicLuwian inscriptions. Berlin / New York. vol. I (in 3 parts: Inscriptionsfrom the Iron Age) by D. Hawkins 2000; vol. 2 (Karatepe-Aslantas:the inscriptions; facsimile edition) by H.

328 Bibliography and Key to Reference CitationsHirt Idg. Gramm. 1- Vll 1927-1937 = H. Hirt Indogermanische Grammatik.7 parts Heidelberg: I (Einleitung, Etymologie, Konsonantismus)1927; II (Vokalismus) 1921; III (Nomen) 1927; IV (Doppelung,Zusammensetzung, Verbum) 1928; V (Akzent) 1929; VI (Syntax I,Syntaktische Verwendung der Kasus und der Verbalformen) 1934; Vll(Syntax I!, Die Lehre vom einJachen und zusammengesetzten Satz)1937, posthumous.Hirt Indogermanen I 1905 II 1907 = H. Hirt Die Indogermanen, [hreVerbreitung, ihre Urheimat und ihre Kultur. 2 vols. Strasbourg 1905-1907.Hirt Urgerm. I-Ill 1931-1934 = H. Hirt Handbuch des Urgermanischen.Heidelberg 193/-1934.Historical Linguisti fs Papers 1985 [1987] = Historical Linguistics, PapersJrom the 1 InternatIOnal Conference on Historical LinguisticsPavia 1985, edited by A. G. Ramat, O. Carruba and G. Bernini. Amsterdam/ Philadelphia 1987 (= Current Issues in Linguistic Theory 48).Historical LinguiSti fs Papers 1989 [1993] = Historical Linguistics, Papersfrom the 9' InternatIOnal ConJerence on Historical LinguisticsRutgers University i 989, edited by H. Aertsen and R. J. Jeffers. Amsterdam/ Philadelphia 1993 (= Current issues in Linguistic Theory106).Historical Linguistics Problems 1993 = Historical Linguistics: Problemsand Perspectives, edited by Ch. 10nes. London / New York 1993 (=Longman Linguistics Library).Historical Morphology 1980 = Historical Morphology, edited by 1. Fisiak.The Hague / Paris / New York 1980 (= Trends in LinguistiCS, Studiesand Monographs 17).Historical Syntax 1984 = Historical Syntax, edited by 1. Fisiak. Berlin etal. 1984 (= Trends in Linguistics, Studies and Monographs 23).HL = Historiographia linguistica.Hock Historical Linguistics 1986 = H. H. Hock PrinCiples oJ HistoricalLinguistics. Berlin 1986.Hock Language History 1996 = H. H. Hock Language History, LanguageChange and Language Relationship: An Introduction to Historical andComparative Linguistics. Berlin / New York 1996. Review: Th. Krischin Kratylos 44 1999 p. 174-176.Hodot Eolien 1990 = R. Hodot Le dialecte .!olien d'Asie. La langue desinscriptions, Vl/e s. a. C. - IVe s. p. C. Paris 1990.Hoenigswald Historical linguistics 1973 = H. M. Hoenigswald Studies inFormal Historical Linguistics. Dordrecht / Boston 1973.Bibliography and Key to Reference Citations 329Hoenigswald Language Change 1960 = H. M. Hoenigswald LanguageChange and Linguistic Reconstruction. Chicago 1960.Hoffinann AuJsiitze 1 1975 II 1976 III 1992 = K. Hoffinann AuJsiitze zurIndoiranistik. Wiesbaden 1975- 1992.Hoffinann Gedenlifeier 1996 [1997] = Akademische Gedenlifeier Jur ProJessor Dr. Karl Hoffmann am 1 i. Juli 1996. Erlangen 1997 (= AkademischeReden und Kol/oquien, Friedrich-Alexander-Universitiit Erlangen-Niirnberg,Bd. 12).Hoffinann Injunktiv 1967 = K. Hoffinann Der Injunktiv im Veda. Einesynchronische Funktionsuntersuchung. Heidelberg 1967.Hoffinann / Forssman Av. Laut- und Flexionslehre 1996 = K. Hoffinannand B. Forssman Avestische Laut- und Flexionslehre. Innsbruck 1992(= IBS Nr. 84). Review: X. Tremblay in BSL 92 / 2 1997 p. 180-184;R. S. P. Beekes in Kratylos 44 1999 p. 62-71.Hoffinann / Narten Sasanid. Archetypus 1989 = K. Hoffinann and J.Narten Der Sasanidische Archetypus, Untersuchungen zu Schreibungund Lautgestalt des Avestischen. Wiesbaden 1989.Hofrnann Lat. Umgangssprache 1936 = 1. B. Hofrnann Lateinische Umgangssprache.Heidelberg 2nd cd. 1936, 4. ed. with index 1978.Hofrnann see s. v. Hofrnann / Szantyr Syntax, and Leumann / Hofrnann /Szantyr Lat. Gr. , and Walde / Hofrnann LEW.Hofinann / Szantyr Syntax 1965 = J. B. Hofrnann Lateinische Syntax undStilistik, revised by A. Szantyr. Munich 1965 (= 2nd vol. by Leumann /Hofrnann / Szantyr Lat. Gr. in the Handbuch der Altertumswissenschafl)·Homeric Questions 1995 = Homeric Questions, Essays in Philology, AncientHistory and Archaeology, edited by J. P. Crielaard. Amsterdam1995.Homers Ilias Gesamtkammentar 2000ff., see Homers !lias Prolegomena.Homers flias Prolegomena 2000 = Homers Ilias, Gesamtkammentar, editedby J. Latacz. Prolegomena as an independent volume by F. Graf, l.de Jong, J. Latacz, R. Niinlist, M. Stoevesandt, R. Wachter, M. L.West. Munich / Leipzig 2000. Gesamtkammmentar: vol. I, I. Gesang(Faszikel I Text und Obersetzung; Faszikel 2 Kommentar) 2000.Hooker Linear B 1980 = J. T. Hooker Linear B, An Introduction. Bristol1980.Hooker Scripta Minora 1996 = 1. T. Hooker Scripta Minora, Selectedessays on Minoan, Mycenaean, Homeric and Classical Greek subjects,edited by F. Amory, P. Considine, S. Hooker. Amsterdam 1996.

330 Bibliography and Key to Reference CitationsBibliography and Key to Reference Citations 331Horrocks Greek 1997 = G. Horrocks Greek, A History of the Languageand its Speakers. London 1997.Horrocks Space and Time 1981 = O. C. Horrocks Space and Time inHomer. Prepositional and Adverbial Particles in the Greek Epic. NewYork 1981.Houwink ten Cate Luwian Population Groups 1965 = Ph. H. 1. Houwinkten Cate The Luwian Population Groups of Lycia and Cilicia Asperaduring the Hellenistic Period. Leiden 1965.HS = Historische Sprachforschung, Oiittingen. Prior to vol. 100 the abbreviationwas ZVS (Zeitschrijt fur Vergleichende Sprachforschung),although It was commonly referred to as KZ (Kuhns Zeitschrift). Forthe vols. I-lOO, there now exists an index (Sachindex), edited by A.Bammesberger, revised by I. Hajnal, Chr. Schaefer, O. Schaufelbergerand S. Zlegler. Oiittingen 1997. On the current internet site of the departmentfor Englische und Vergleichende Sprachwissenschaji of theKatholische Universitat Eichstall, A. Bammesberger offers an additionalindex to the first hundred volumes.HSK = Handbucher zur Sprach- und Kommunikationswissenschaji: Berlin'New York. vol. 1.1 and 1.2 see above, Dialelctologie 1-2 1982-1983;vols. 2. I and 2.2 see below, Sprachgeschichte 1-2 1998-1985; vols. 5. Iand 5.2 see below, W6rterbucher 1-3 1989-1991; vols. 9. 1 and 9.2 seebelow, Syntax 1-2 1993-I 995; vol. 10. I and 10.2 see below, Schrijtund Schrijtlichkeit I -2 1994-I 996; vol. 11.1 and I 1.2 see below ' NamenjorschungI -2 1 995-I 996.Htibschmann Casuslehre 1875 = H. Htibschmann Zur Casuslehre. Munich1875.Hiibschrnann Kleine Schrijten 1976 = H. Hiibschmann Kleine Schrijtenzum Armenischen, edited by R. Schmitt. Hildesheim ' New York 1976.Further, cC R. Schmitt "Schrijtenverzeichnis Heinrich Hubschmann" inHS I II 1998 p. 185-190.IBK = Innsbrucker Beitrage zur Kulturwissenschaji.IBS = Innsbrucker Beitrage zur Sprachwissenschaji.ICLL, see below, IKLL.Idg. Chr. = Indogermanische Chronik in Sprache, since vol. 13 1967 including13a. Last available: Idg. Chr. 35 ' IT in Sprache 37 ' 3 1995.Idg. Gr. I , I 1986 [ , 2 1986 II 1968 III , I 1969 = IndogermanischeGrammatik, founded by J. Kurylowicz, edited by M. Mayrhofer. Heidelberg:1 , I (Einlei/ung) 1986 see above, Cowgill Einleitung 1986; I ,2 (Laut/ehre) 1986 see below, Mayrhofer Lautlehre 1986; 11 (A kzent .Ablaut) 1968 see below, Kurylowicz Akzent . Ablaut 1968; III (Formenlehre), I (Geschichte der indogermanischen Verbaljlexion) 1969see below, Watkins Verbaljlexion 1969.lE and IEs 1970 = Indo-European and Indo-Europeans, edited by O.Cardona, H. M. Hoenigswald, A. Senn. Philadelphia 1970.lE Subgrouping 1999 = Special Session on Indo-European Subgroupingand Internal Relations, edited by B. K. Bergen, M. C. Plauche, A. C.Bailey. Berkely 1998 (= Berkely Linguistic Society, Proceedings of theTwenty-Fourth Annual Meeting).IEW, see below, Pokorny IEW.IF = Indogermanische Forschungen.ITJ = Indo-Iranian Journal.IKLL (YCLL ' CILL) Y 1981 (1983); IV 1987 (1991]; V 1989; VII 1993[1996]; VIII 1995 [1996]= Internationales Kolloquium zur lateinischenLinguistik [IKLL] (International Colloquium on Latin Linguistics[ICLL]' Colloque international de linguistique latine [CILL]): - I1981 [1983] Latin Linguistics and Linguistic Theory, Proceedings ofthe 1" ICLL Amsterdam 1981, edited by H. Pinkster. Amsterdam ,Philadelphia 1983. - IV 1987 11991]: New Studies in Latin linguistics,Selected papers from the 4 ICLL Cambridge 1987, edIted by R.Coleman. Amsterdam ' Philadelphia 1991. - V 1989: Actes du V'CILL Louvain-la-Neuve / Borzee 1989, edited by M. Lavency and D.Longre. Louvain-Ia-Neuve 1989. - VII 1993 [1996] see above, Aspectsof Latin 1996. - VIII 1995 (1996]: Alcten des VllI. IKLL Eichstall1995, edited by A. Bammesberger and F. Heberiein. Heidelberg1996.Iliad I-VI 1985-1993 = The Iliad, A Commentary, edited by G. S. Kirk.Cambridge: I (Books 1-4) and II (Books 5-8) by G. S. Kirk; III (Books9-12) by B. Hainsworth; IV (Books 13-16) by R. Janko; V (Books 17-20) by M. W. Edwards; VI (Books 21-24) by N. Richardson.Illic-Svityc Nominal Accentuation 1979 = V. M. Illic-Svityc Nominal Accentuationin Baltic and Slavic, translated from Russian by R. L. Leedand R. F. Feldstein. Cambridge ' Mass. 1979.Indo-European Languages 1998 = The Indo-European Languages, editedby A. G. Ramat and P. Ramal. London 1998. See also the Italian original,s. v. Lingue indoeuropee 1994.Indo-Europeanization of Northern Europe 1996 = The Indo-Europeanizationof Northern Europe, Papers Presented at the International Conference in Vilnius 1994, edited by K. Jones-Bley and M. E. Huld.

332Bibliography and Key to Reference CitationsBibliography and Key to Reference Citations 333Washington 1996 (= llES, Monograph 17). Review: B. Hiinsel and B.Schlerath in Kratylos 46 2001 p. 55-62.Indogermanische Dichtersprache 1968 = Indogermanische Dichte"pra_che, edited by R. Schmitt. Dannstadt 1968 (= Wege der ForschungCLXV).Indologie 1979 = EinJuhrung in die Indologie, Stand, Methoden, AuJgaben,edited by H. Bechert and G. von Simson. 1st ed. 1979 2nd ed.1993.1nL = Incontri Linguistici.Insler Giithiis 1975 = S. Insler The Giithiis oJZarathustra. Teheran 1 Liege1 Lelden 1975 (= Acta Iranica 8). Review: H.-P. Schrnidt in IIJ 211979 p. 83-1 15.Iranisches Personennamenbuch = Iranisches Personennamenbuch, editedby M. Mayrhofer and R. Schmitt. Vienna. vot. I 1979: M. MayrhoferDie altlramschen Namen (which includes Avestan and Old Persiannames: Cf. .by the same author, Zum Namengut des Avesta. Vienna1977 [= SbOA W vot. 308, Abhandlung 5]). The Iranische Personennamenbuchan ambitious project which comprises altogether 10 vols.Along With vot. I, other fitscicles and volumes have already been published,cf. vot. [[ Faszikel 2 1986 Ph. Gignoux Noms propres sassamdesen moyen-perse epigraphique, vot. V Faszikel 4 1982 RSchmitt Iranische Namen in den indogermanischen Sprachen Kleina slens (Lykisch, Lydisch, Phrygisch), vot. V Faszikel 6a 1990 Ph. HuyseIramsche Namen In den griechischen Dokumenten A·gyptens.Itaha alumna 1990 = Italia omnium terrarum alumna, La civiltiI dei Veneti,Reti, Liguri, Celti, Piceni, Umbri, Latini, Campani e lapigi. Milan:1st ed. 1988, 2nd ed. 1990 (= Credito Italiano, Antica Madre,Col/ana di studi sul/' ltalia antica vo!. 11, under the direction of G. P.Carratelli).Italia parens 1991 = Italia omnium terrarum parens, La civiltiI degli Enotn,Chom, Ausoni, Sanniti, Lucani, Breltii, Sicani, Siculi, Elimi. Milan:1st ed. 1989, 2nd ed. 1991 (= Credito Italiano, Antica Madre Col/anadi studi sul/' Italia antica vot. 12, under the direction of G. P. Carratelli).Jackson Early Britain 1953 = K. Jackson Language and History in EarlyBntaln, A Chronological Survey oJ the Briltonic Languages First toTwelfth Century A.D. Edinburgh 1953.'Jacquinod Double accusatif en grec 1989 = B. Jacquinod Le double accusatif en grec d'Homere iI la fin du V' siec/e avant J-c. Louvain-LaNeuve 1989Jarnison -aya- 1983 = S. W. Jamison Function and Form in the -tiyaFormations oJ the Rig Veda and Atharva Veda. Gottingen 1983 (=ZVS, Ergiinzungshefl Nr. 31). Review: G. Pinault in Kratylos 29 1984[1985] p. 47-5 1; M. Peters in ldg. Chr. 30a Nr. 242.Janda Stock und Stein 1997 = M. Janda Ober "Stock und Stein ", Die indogermanischenVariationen eines universalen Phraseologismus. (dissertation)Vienna 1995 = MSS, Beihe fl l8, N.F. 1997.Janda Eleusis 1999 = M. Janda Eleusis, Das indogermanische Erbe derMysterien. Innsbruck 1999 (= IBS 96).Jankuhn Passive Bedeutung 1969 = H. Jankuhn Die passive Bedeutungmedialer Formen untersucht an der Sprache Homers. Gottingen 1969(= supplement to ZVS 21).Jasanoff Stative and Middle 1978 = J. H. Jasanoff Stative and Middle inIndo-European. Innsbruck 1978.JAOS = Journal oJ the American Oriental Society.Jensen Altarm. Gr. 1959 = H. Jensen Altarmenische Grammatik. Heidelberg1959.JIES = Journal oJ Indo-European Studies. Washington D. C.Joachirn MehrJachpriisentien 1978 = U. Joachirn MehrJachpriisentien im8gveda. Frankfurt am Main 1 Bern 1 Las Vegas 1978 (= EuropiiischeHochschulschriften Reihe XXI vo!. 4).Joki Uralier und Indogermanen 1973 = A. 1. Joki Uralier und Indogermanen.Die iilteren Beruhrungen zwischen den uralischen und indogermanischenSprachen. Helsinki 1973.Iok! Albanisch = N. Jokl Linguistisch-kulturhistorische Untersuchungenaus dem Bereiche des Albanischen. Berlin 1 Leipzig 1923.Kadmos = Kadmos. Zeitschrift for vor- und Jriihgriechische Epigraphik.Berlin.Kammenhuber Arier 1968 = A. Kammenhuber Die Arier im VorderenOrient. Heidelberg 1968.Kammenhuber Kleine Schriften 1993 = A. Kammenhuber Kleine Schriftenzum Altanatolischen und Indogermanischen (1. Teilband 1955-1968,2. Teilband 1969-1990) Heide1berg 1993. Review including muchcomplementary information and many references: 1. Catsanicos in BSL92 /2 1997 p. 156-179.

334 Bibliography and Key to Reference CitationsKastner Adjektive 1967 = W. Kastner Die griechischen Adjektive zweierEndungen auJ -OJ:. Heidelberg 1967. Review: E. Neu in IF 74 1969 p.235-242.Kata ditilekton 1996 [1999] = Kata ditilekton, Alii del III Colloquio Internationaledi Dialectologia Greca Napoli / Fiaiano d'ischia 1996edited by A. C. Cassio, Naples 1999 (= Annali dell ' istituto Universitario Orientale di Napoli, Sezione Filologico-Lelleraria, 19 [1997]).See above, Dialectologica Graeca 1991 [1993].Katicic Languages oJ the Balkans 1976 = R. Katicic Ancient LanguagesoJ the Balkans I. The Hague / Paris 1976. Review: C. de Simone inKratylos 22 1977 [1978] p. 113-1 19.Katz (H.) Lehnworter 1985 = H. Katz Studien zu den alteren indoiranisc?enLehnwortern in den uralischen Sprachen. postdoctoral paper MunIch1985. A posthumous publication is presently being prepared by RP. Ritter for Verlag Winter.Katz (J.) Personal Pronouns 1998 = 1. T. Katz Topics in Indo-EuropeanPersonal Pronouns. dissertation. Harvard 1998. Cf. also, by the sameauthor, the review of Schrijver Celtic Pronouns and Particles 1997 inKratylos 46 2001 p. 1-23.KeUens Noms racines 1974 = 1. KeUens Les noms-racines de l'Avesta.Wiesbaden 1974 (= Beitrage zur iranistik vo!. 7). Review: R Sclunittin Kratylos 19 1974 [1975] p. 56-60; 1. Schindler in Sprache 25 1979p. 57-60.KeIJe?s Verbe avestique 1984 + 1995 = 1. KeIJens Le verbe avestique.Wlesbaden 1984; by the same author, Liste du verbe avestique. Wiesbaden1995.KeUens / Pirart Textes vieil-avestiques I 1988 IJ 1990 III 1991 = 1. Kellens,E. Pirart Les textes vieil-avestiques. Wiesbaden: vo!. I (introduction,texte et traduction) 1988; vo!. IT (Repertoires grammaticaux etlexique) 1990; vo!. 1II (Commentaire) 1991. Reviews of vol. I 1988: P.O. Skjrerv0 in JAOS 111 1991 p. 659-662; S. W. Jamison in IIJ 361993 p. 244-25 1. Review of vols. 1-lI/: N. Oettinger in Kratylos 381993 p. 43-49.KeUer (M.) Verbes latins a irifectum en -sc- 1992 = M. KeUer Les verbeslatins a irifectum en -sc-, Etude morphologique a partir des inJormationsat/estees des I'epoque preclassique. Brussels 1992 (= CollectionLatomus 216).KeUer (R) Sprachwandel 1994 = R. KeUer Sprachwandel, Von der unsichtbarenHand in der Sprache. 2nd ed. Tilbingen and Basel 1994 (=UTB Nr. 1567).Bibliography and Key to Reference Citations 335Keltologen-Symposium IT 1997 [1999] = Akten des Zweilen DeutschenKeltologen-Symposiums in Bonn J 997, edited by St. Zimmer, R. KMderitzsch,A. Wigger. Tilbingen 1999. See above, DeutschsprachigeKeltologen 1992 [1993].KEWA, see below, Mayrhofer KEW A.Keydana Absol. Konstr. 1997 = G. Keydana Absolute Konstruktionen inaltindogermanischen Sprachen. Gottingen 1997 (= Erganzungsheft HSNr. 40).Kieckers Gr. Gr. lI/-IV 1926 = E. Kieckers Historische Grammatik desGriechischen. 4 parts. Berlin / Leipzig (Sammlung Goschen: vols. 117,118, 924, 925) 1925-1926: I (Lautlehre) 1925; II (Formenlehre) 1926;III (Synlax, ersler Teil) 1926; IV (Syntax, zweiter Teil) 1926.Kieckers Stellung des Verbs 1911 = E. Kieckers Die Stellung des Verbs imGriechischen und in den verwandten Sprachen, Erster Teil: Die Stellungdes Verbs im eirifachen Hauptsalze und im Nachsatze nach dengriechischen Inschriften und der alteren griechischen Prosa, verglichenmil den verwandten Sprachen. Strasbourg 1911.Kienle Histor. LFL d. Dt. 1965 = R. von Kienle Historische Laut- undFormenlehre des Deutschen. 2nd ed. Tilbingen 1965.Kilian lndogermanen 1983 = L. Kilian Zum Ursprung der Indogermanen,Forschungen aus Linguisti/c, Prahistorie und Anthropologie. Bonn1983. Review: F. Lochner von Hilttenbach in Kratylos 29 1984 [1985]p. 160-163.Kimball Hittite Historical Phonology 1999 = S. E. Kimball Hittile HistoricalPhonology. Innsbruck 1999 (= IBS 95).Klaproth Asia polyglot/a 1823 = J. Klaproth Asia polyglot/a. Paris 1823.The importance of this pioneering work should not be underestimated.Cf. G. Bolognesi in FS Belardi I 1994 p. 334ff.Klein Discourse Grammar 1985 = J. S. Klein Toward a Discourse GrammaroJ the Rigveda. 2 parts. Heidelberg 1985. Review: H. Hettrich inKratylos 33 1988 p. 72-79.Klein Particle u 1978 = J. S. Klein The Particle u in the Rigveda, A Synchronicand Diachronic Study. GlIttingen 1978 (= ZVS, Erganzungsheft27).Klein Personal Deixis 1996 = J. S. Klein On Personal Deixis in ClassicalArmenian, A Study oJ Ihe Syntax and Semantics oJ the n-, s- and dDemonstratives in Manuscripts E and M oJ the Old Armenian Gospels.MSS, Beiheft 17, N.F. 1996.

336 Bibliography and Key to Reference CitationsKlein Verbal Accentuation 1992 = J. S. K1ein On Verbal Accentuation inthe Rigveda. New Haven 1992. Review: S. Migron in Kratylos 401995 p. 190-192.Klingenschmitt Altarm. Verbum 1982 = G. Klingenschmitt Das AltarmenischeVerbum. Wiesbaden 1982. Review: G. R. Solta in Kratylos29 1984 [1985) p. 59-74.KI. Pauly = Der Kleine Pauly, Lexi!wn der Antike. 5 vols. Stuttgart / Munich1964-1975 (= dtv 1979 Nr. 5963). Sec above, s. v. DNP.K1uge Stammbildungslehre 1926 = F. Kluge Nominale Stammbildungslehreder altgermanischen Dialekte. 3rd ed. Halle 1926.KJuge Urgermanisch 1913 = F. K1uge Urgermanisch, Vorgeschichte deraltgermanischen Dialekte. Strasbourg 1913.K1uge / Seebold 1995 = F. K1uge Etymologisches Worterbuch derDeutschen Sprache. revised by E. Seebold. 23rd expanded ed., Berlin /New York 1995.Koch Aksl. Verbum I / II 1990 = Ch. Koch Das morphologische Systemdes altkirchenslavischen Verbums. Munich 1990: vol. I (Text); vo!. IT(Anmerkungen).Kolver Sekundtire Kasus 1965 = B. Kolver Der Gebrauch der sekundtirenKasus im Tocharischen. dissertation. Frankfurt 1965.Koerner Practicing Linguistic Historiography 1989 = K. Koerner PracticingLinguistic Historiography, Selected Essays. Amsterdam 1 Philadelphia1989 (= Amsterdam Studies in the Theory and History of LinguisticScience 50).Kohrt Problemgeschichte 1985 = M. Kohrt Problemgeschichte des Graphembegriffs und des fruhen Phonembegrifft. TUbingen 1985 (= Reihegermanistische Linguislik Nr. 61).Koim! 1 1993 1I 1996 III 1998 = La !wine grecque antique, sous la directionde Cl. Brixhe. Nancy (= Etudes anciennes 10, 14, 17). vol. I 1993(une langue introuvable ?); II 1996 (La concurrence); ill 1998 (Lescontacts).Koivulehto Uralische Evidenzfur die Laryngaltheorie 1991 = J. Koivulehto.Uralische Evidenz fur die Laryngaltheorie. Vienna 1991 (=SbOA W, 566. vol.). Review: T. Hofstra and O. Nikkilii in Kratylos 381993 p. 36-39; R. P. Ritter in PFU 1 1994 / 1995 p. 3-8.Kolb Rom 1995 = F. Kolb Rom, Die Geschichte der Stadt in der Antike.Munich 1995.Kolloquium Delbruck Madrid 1994 [1997) = Berthold Delbruck y la sintaxisindoeuropea hoy, Aetas del Coloquio de la ldg. GesellschajiBibliography and Key to Reference Citations 3371994 in Madrid, edited by E. Crespo and J.-L. Garcia-Ram6n. Wiesbaden1997.Kolloquium Germanisch Freiburg 1981 [1984) = Das Germanische unddie Rekonstruktion der indogermanischen Grundsprache, Aklen desFreiburger Kolloquiums der Idg. Gesellschaji 1981 in Freiburg, editedby J. Untermann and B. Brogyanyi. Amsterdam / Philadelphia 1984.Kolloquium Idg., Slaw. u. Bait. Jena 1989 [1992) = Indogermanisch,Slawisch und Baltisch, Materialien des Kolloquiums Jena 1989, editedby B. Barschel, M. Kozianka, K. Weber. Munich 1992.Kolloquium Keltisch Bonn 1976 [1977) = Indogermanisch und Keltisch,Kolloquium der Idg. Gesellschaji 1976 in Bonn, edited by K. H.Schmidl. Wiesbaden 1977.Kolloquium Kuhner Amsterdam 1986 [1988) = In the Footsteps of R.Kuhner, Proceedings of the International Coil. in Commemoration ofthe 150th Anniversary of the Publication of R. Kuhner's AusfohrlicheGrammatik 11: Syntaxe 1986 in Amsterdam, edited by A. Rijksbaron,H. A. Mulder, G. C. Wakker. Amsterdam 1988.Kolloquium Lat. u. Idg. Salzburg 1986 [1992) = Latein und Indogermanisch,Aklen des Kolloquiums der Idg. Gesellschaji 1986 in Salzburg,edited by O. Panagl and Th. Krisch. Innsbruck 1992.Kolloquium Pedersen Kopenhagen 1993 [1994) = In honorem HolgerPedersen, Kolloquium der Idg. Gesellschaji 1993 in Copenhagen, editedby J. E. Rasmussen. Wiesbaden 1994.Kolloquium Syntax Pavia 1979 [1980) = linguistiC Reconstruction andIndo-European Syntax, Proceedings of the Colloquium of the " IndogermanischeGesellschaji" 1979 in Pavia, edited by P. Ramal. Amsterdam1980.Kolloquium Wackernagel Basel 1988 [1990) = Sprachwissenschaji undPhilologie, Jacob Wackernagel und die Indogermanistik heute, Kolloquiumder Idg. Gesellschaji 1988 in Basel, edited by H. Eichner and H.Rix. Wiesbaden 1990.Kortlandt Slavic Accentuation 1975 = F. H. H. Kort1andt Slavic Accentuation,A Study in Relative Chronology. Lisse 1 Holland 1975.Krahe Idg. Sprachw. I 1966 IT 1969 = H. Krahe IndogermanischeSprachwissenschaji. Berlin: vol. I (Einleitung und Lautlehre) 5th ed.1969 (= Sammlung Goschen Nr. 59); vol. 1I (Formenlehre) 5th ed.1969 (= Sammlung Goschen Nr. 64). 6th ed. (unchanged) in one vo!.1985 (= Sammlung Goschen Nr. 2227). Concerning older editions, aswell as previous work by R. Meringer, see p. IX above.

338 Bibliography and Key to Reference CitationsKrahe IlIyrier I 1955 II 1964 = H. Krahe Die Sprache der IlIyrier. Wiesbaden:vo!. [ (Die Quellen) 1955; vo!. II 1964 (contains Die messapischenInschrifien by C. de Simone and Die messapischen Personennamenby J. Untermann).Krahe Vergl. Syntax 1972 = H. Krahe Grundziige der vergleichendenSynlax der indogermanischen Sprachen, edited by W. Meid and H.Schmeja. Innsbruck 1972 (= ms 8).Krahe / Meid Germ. Sprachw. I 1969 11 1969 I1I 1967 = H. Krahe and W.Meid Germanische Sprachwissenschaft. 3 vols. Berlin: I (Einleilungund Laut/ehre) 4th ed. 1969; n (Formenlehre) 4th ed. 1969; III (Worlbildungslehre)1st ed. 1967 (= Sammlung Goschen Nr. 238, 780,1218).Kralylos = Kralylos. Krilisches Berichls- und Rezensionsorgan fiir indogermanischeund allgemeine Sprachwissenschaft. Wiesbaden.Krause Handb. d. GOI. 1968 = W. Krause Handbuch des GOlischen. 3. ed.Munich 1968.Krause / Thomas Toch. Elemenlarbuch I 1960 II 1964 = W. Krause andW. Thomas Tocharisches Elemenlarbuch. Heidelberg: vo!. I 1960; vo!.II (W. Thomas) 1964.Kretschmer Einleilung 1896 = P. Kretschmer Einleilung in die Geschichleder griechischen Sprache. Oiittingen 1896.Krisch Kondilionalsiilze 1986 = Th. Krisch Oberiegungen zur Herkunftund Enlwicklung der irrealen Kondilionalsiilze des Allgriechischen.Innsbruck 1986 (= ms, VOrlriige und Kleinere Schrifien 38).Krisch Konslruktionsmusler 1984 = Th. Krisch Konslrulaionsmusler undBedeulungswandel indogermanischer Verben. Anwendungsversuchevon Valenzlheorie und Kasusgrammalik auf Diachronie und Rekonstrulaion.Frankfurt am Main, et al. 1984. Review: Chr. Koch in Anzeigerfiir Slavische Philologie 17 1986 p. 187-198.Krisch Perfelaa mil langem Reduplikolionsvokal 1996 = Th. Krisch ZurGenese und Funktion der allindischen Perfelaa mil langem Reduplikolionsvokal,Mil kommenlierler Malerialsammlung. Innsbruck 1996 (=IBS 87). Review: O.-J. Pinault in BSL 93 / 2 1998 p. 139-143; St. W.Jamison in Kralylos 44 1999 p. 59-62.Krogh Slellung des Altsiichsischen 1996 = S. Krogh Die Slellung desAllsiichsischen im Rahmen der germanischen Sprachen. Oiittingen1996 (= Sludien zum Althochdeulschen 29).KS = Kleine Schrifien.Kiihner / Blass Ausfiihrliche Gramm. d. gr. Sprache I / I 1890 I / 2 1892= R. Kiihner Ausfuhrliche Grammalik der griechischen Sprache. Han-Bibliography and Key to Reference Citations 339nover: Ersler Teil (Elemenlar- und Formenlehre) revised by F. Blass,3rd ed. in 2 vols. 1890 and 1892.Kiihner / Oerth Ausfiihrliche Gramm. d. gr. Sprache Il / I 1898 II / 21904 = R. Kiihner Ausfiihrliche Grammalik der griechischen Sp rache.Hannover / Leipzig: Zweiler Teil (Salzlehre) revised by B. Oerth, 3rded. in 2 vols. 1898 and 1904.Kiihner / Holzweissig Ausfiihrliche Gramm. d. lal. Sprache I 1912 = R.Kiihner Ausfiihrliche Grammalik der laleinischen Sprache. ErslerBand. (Elemenlar-, Formen- und Laullehre) revised by F. Holzweissig.Kiihner / Stegmann A usfiihrliche Gramm. d. lal. Sprache n / I and 11 / 21955 = R. Kiihner Ausfiihrliche Grammalik der laleinischen Sprache,Salzlehre, revised by C. Stegmann, 2 parts. 3rd ed. revised A. Thierfelder.Leverkusen 1955.Kiimmel Perfela 2000 = M. Kiimmel Das Perfela im Indoiranischen, EineUnlersuchung der Form und Funlaion einer ereblen Kalegorie desVerbums und ihrer Weiterbildung in den allindoeuropiiischen Sprachen.Wiesbaden 2000.KUmmel Slaliv und Passivaorist 1996 = M. KUmmel Slaliv und Passivaorislim Indoiranischen. Oiillingen 1996 (= HS Ergiinzungsheft 39).Review: I. Hajnal in Kralylos 44 1999 p. 50-54.Kuhn Lelzles Indogermanisch 1978 = H. Kuhn Das lelzle Indogermanisch.Mainz 1978 (= AA WL 1978 Nr. 4). Critical review: B.Schlerath in Kralylos 23 1978 [1979] p. 44-57.Kuiper Nasalprasentia 1937 = F. B. J. Kuiper Die indogermanischen Nasalpriisenlia.Ein Versuch zu einer morphologischen Analyse. Amsterdam1937 (first appeared as a dissertation in a shorter version in 1934).Kuiper Selected Writings 1997 = F. B. J. Kuiper Selected Writings on IndianLinguislics and Philology, edited by A. Lubotsky, M. S. Oort, M.Witze!. Amsterdam 1997.Kuiper Ved. Noun-Inflexion 1942 = F. B. J. Kuiper NOles on Vedic NounIriflexion. Amsterdam 1942 (reprinted in Selecled Writings 1997 p.437-530); see below, F 314 § 5.Kurylowicz Accenlualion 1952 = J. Kurylowicz L'accenlualion deslangues indo-europeennes. Cracow 1952, 2" " ed. 1956.Kurylowicz Akzenl . Ablaul 1968 = J. Kurylowicz Akzenl . Ablaul = vo!.n der Idg. Gr. (see above,). Heidelberg 1968.Kurylowicz Apophonie 1956 = J. Kurylowicz L'apophonie en indoeuropeen.Wroclaw 1956.Kurylowicz Esquisses I 1973 II 1975 = J. Kurylowicz Esquisses linguisliques.2 vols. Munich: I 1st ed. 1960, 2nd ed. 1973; II 1975.

340 Bibliography and Key to Reference CitationsKurylowicz Etudes [ 1935 = J. Kurylowicz Etudes indo-europeennes I.Cracow 1935.Kurylowicz Inflectional Categories 1964 = J. Kurylowicz The InflectionalCategories of Indo-European. Heidelberg 1964.Kurylowicz, see also GS Kurylowicz 1995 above.KZ = Zeitschrift fur Vergleichende Sprachforschung. founded by A.Kuhn. Presently published in Gottingen, see above, HS.LALIES = LALIES. Actes des sessions de linguistique et de lillerature.Paris.Lambert Langue gauloise 1994 = P.-Y. Lambert La langue gauloise. Paris1994.Lamberterie Ad}. en -u; 1990 = Ch. de Lamberterie Les adjectifs grecs en- U; 2 vols. Louvain-Ia-Neuve 1990. Review: J. L. PerpiUou in RPh 641990 [1992] p. 197-200; F. Bader and L. Dubois in BSL 86 I 2 1991 p.145-149; M. Peters in Idg. Chr. 34 Nr. G 60 1; F. Mawet in REArm 241993 p. 301-305.Lamberterie A rmenien c/assique 1992 = Ch. de Lamberterie Introductiona I'armenien classique in LALIES 10 1992 p. 234-289.langues indo-europeennes 1994 = langues indo-europeennes, edited by F.Bader. Paris 1994. reprint, with addenda 1997. Review: A. Blanc inBSL 92 /21997 p. 141-143.Language Typology 1988 [1991] = Language Typology 1988, TypologicalModels in the Service of Reconstruction, edited by W. P. Lehmannand H. J. Hewitt. Amsterdam I Philadelphia 1991 (= Amsterdam Studiesin the Theory and History of Linguistic Science, Series TV: Current lssuesin Linguistic Theory 81).Lanszweert Bait. Grundwortschatz 1984 = R. Lanszweert Die Rekonstruktiondes baltischen Grundwortschatzes. Frankfurt am Main 1984.La Roche Accusativ 1861 = J. La Roche Homerische Studien, Der Accusativim Homer. Vienna 1861Laroche Hieroglyphes hit/ites I 1960 = E. Laroche Les hieroglyphes hittitesI (L'ecriture). Paris 1960.Laroche Hourrite 1980 = E. Laroche Glossaire de la langue hourrite.Paris 1980.Laroche Louvite 1959 = E. Laroche Dictionnaire de la langue louvite.Paris 1959.Laryngales 1990 = La reconstruction des laryngales, edited by J. KeUens.Liege 1990. Review: G.-J. Pinault in BSL 86 I 2 1991 [1992] p. 116-118; F. O. Lindeman in Kratylos 37 1992 p. 58-62.Bibliography and Key to Reference Citations 341Laryngaltheorie 1988 = Die Laryngaitheorie und die Rekonstruktion desindogermanischen Laut- und Formensystems, edited by A. Bammesberger.Heidelberg 1988. Index by S. Ziegler 1990. Review: M. Petersin ldg. Chr. 33 Nr. G 231; F. O. Lindeman in HS 102 1989 p. 268-297;R. S. P. Beekes in Amsterdamer Beitrage zur A'lteren Germanistik 331991 p. 237-245.Latacz Homer 1989 = J. Latacz Homer, Der erste Dichter des Abendlandes.2nd ed. Munich I Ziirich 1989.Latacz Troia und Homer 2001 = J. Latacz Troia und Homer, Der Weg zurLosung eines alten Ratsels. Munich I Berlin 200 I.Lat. Philologie 1996 = Einleitung in die lateinische Philologie, edited byF. Graf. Stuttgart I Leipzig 1996. Review: W. Pfaffel in Kratylos 441999 p. 94-98.LAW 1965 = Lexikon der Alten Welt. Ziirich I Stuttgart 1965.LDlA 1978 = Lingue e dialelli dell' Italia antica, edited by A. L. Prosdocimi.Rome 1978. revised and with indexes by A. Marinetti 1982.Lehmann (Chr.) Relativsatz 1984 = Chr. Lehmann ner Relativsatz, Typologieseiner Strukturen, Theorie seiner Funktionen, Kompendiumseiner Grammatik. TUbingen 1984 (= Language Universals Series 3).Lehmann (W. P.) Gothic Etymological Dictionary 1986 = W. P.Lehmann, A Gothic Etymological Dictionary, based on the third editionof Feist Gothisches Worterbuch. Leiden 1986.Lehmann (W. P.) Idg. Forschung 1992 = W. P. Lehmann Die gegnwartigeRichtung der indogermanischen Forschung. Budapest 1992. Review:M. Kiirnmel in PFU 4 1998 p. 51-59.Lehmann (W. P.) PIE Syntax 1974 = W. P. Lehmann Proto-IndoEuropean Syntax. Austin 1974.Leisi Sprach-Knigge 1992 = I. and E. Leisi Sprach-Knigge oder Wie undwas soli ich reden ? 1st ed. TUbingen 1992 (3rd ed. 1993).Leisi Streiflichter 1995 = E. Leisi Streiflichter: unzeitgemafte Essays zuKuitur, Sprache und Literatur. TUbingen 1995.Leisi Wortinhalt 1974 = E. Leisi Der Wortinhalt. 4th ed. Heidelberg 1974.Lejeune Lepontica 1971 = M. Lejeune Lepontica. Paris 1971 (=Monographies Linguistiques I).Lejeune Memoires I 1958 IT 1971 lIT 1972 IV 1997 = M. LejeuneMemoires de philologie mycenienne. 4 series. Paris: I (1955-1957)1958; Rome: 11 (1958-1963) 1971; III (1964-1968) 1972; IV (1969-(996) 1997.

342Bibliography and Key to Reference CitationsBibliography and Key to Reference Citations 343Lejeune Notice 1993 = M. Lejeune Notice biographique et bibliographique,suivie de I'expose "D'Alcoy a Espanca: Rejlexions sur les ecriturespaleo-hispaniques". Leuven 1993.Lejeune Phomltique 1972 = M. Lejeune Phonetique historique du mycenienet du grec ancien. 2nd ed. Paris 1972.Lejeune Venete 1974 = M. Lejeune Manuel de la langue venete. Heidelberg1974.Leskien Handb. d. abulg. (aksl.) Sprache 1962 = A. Leskien Handbuchder altbulgarischen Sprache. Heidelberg 1962.Lesky Gr. Lit. 1957-1958 = A. Lesky Geschichte der griechischen Literatur.1st ed. Bern 1957-1958, 3rd ed. Bern and Munich 1971.Leukart Fruhgr. Nomina 1994 = A. Leukart Die fruhgriechischen Nominaauf -tas und -as, Untersuchungen zu ihrer Herkunji und Ausbreitung(unter Vergleich mit den Nomina auf -ells). Vienna 1994 (= SbOAW558). Review: 1.-L. Perpillou in Kratylos 42 1997 p. 81-86.Leurnann Homerische Worter 1950 = M. Leurnann Homerische Worter.Basel 1950 (and reprint).Leurnann Kleine Schrifien 1959 = M. Leurnann Kleine Schrifien, edited byH. Haffier, E. Risch and W. Riiegg. ZUrich 1959.Leumann LLFL 1977 = M. Leumann Lateinische Laut- und Formenlehre.new ed. Munich 1977 (= vo!. I of Leurnann / Hofmann / Szantyr Lat.Gr.). The new edition replaces an earlier version (1926-1928) from thehand of M. Leurnann. The 1926-1 928 version was included as the 5thedition of Lateinischen Laut- und Formenlehre, part of the greaterwork, Handbuch der Altertumswissenschaji and was written as a thoroughrevision of the 4 th edition by F. Stolz. It is unsurpassed in brevityand conciseness.Leurnann Neuerungen 1952 = M. Leumann Morphologische Neuerungenim altindischen Verbalsystem. Amsterdam 1952.Leumann / Hofrnann / Szantyr AlIg. Teil 1965 = M. Leumann, J. B. Hofmannand A. Szantyr AlIgemeiner Teil. This work presents general reflectionson Leumann / Hofmann / Szantyr Lat. Gr. and constitutes anappendix in Hofmann / Szantyr Lat. Syntax 1965.Leurnann / Hofmann / Szantyr Lat. Gr. = M. Leumann, 1. B. Hofrnann andA. Szantyr Lateinische Grammatik. Munich 1977 and 1965 (as part ofthe Handbuch der Altertumswissenschaji). The first volume is includedin the present bibliography as Leumann LLFL 1977; the second volume,as Hofmann / Szantyr Lat. Syntax 1965. The general section (AlIgemeinerTeif) is included as Leumann / Hofrnann / Szantyr AlIg. Teil1965.LEW = A. Walde Lateinisches etymologisches Worterbuch. 3rd reviseded. by J. B. Hofmann. 3 vols. Heidelberg 1938-1956.Lewandowski Linguist. Worterbuch 1-3 1994 = Th. Lewandowski LinguistischesWorterbuch. 6th ed. Heidelberg / Wiesbaden 1994 (= UTB1518): I (A-H), 2 (I-R), 3 (S-Z).Lex(icon) Gramm(aticorum) 1996 = Lexicon Grammaticorum, Who'sWho in the History of World Linguistics, General Editor H. Stammerjohann.TUbingen 1996.LfgrE = Lexikon des frilhgriechischen Epos, edited by Thesaurus LinguaeGraecae in Hamburg. Founded by B. Snell. Gottingen: Fascicle I(0 - OetKi]s) 1955. The fascicles 1-9 were published as vo!. I (A) 1979;10-14 as vo\. IT (B-A) 1991. The newest fascicles are 17 (oli6t; - 0pooo)1999 and 18 (oPY'i -Itat;) 2000.Lindeman Laryngeal Theory 1997 = F. O. Lindeman Introduction to the"Laryngeal Theory". revised and augmented edition of the Oslo 1987edition. Innsbruck 1997 (= IBS Nr. 91). Review: B. Forssman in Kratylos45 2000 p. 68-75. Review of the 1987 edition: H. Rix in IF 961991 p. 269-274; M. Mayrhofer in Kratylos 36 1991 p. 92-95. Cf. alsoF. O. Lindeman in HS 102 1989 p. 268-297 (review of Laryngaltheorie,1988).Lingue e dialelli 1978 = Lingue e dialelli dell' Italia antica, edited by A.L. Prosdocimi. Aggiornamento e indici a cura di A. Marinetti. Padua1982.Lingue indoeuropee 1994 = Le lingue indoeuropee, edited by A. GiacaloneRamat and P. Ramal. Bologna 1994. Review: B. Schirmer inKratylos 42 1997 p. 39-43. Cf. the English translation, s. v. IndoEuropean Languages 1998.Linguistic Change and Reconstruction Methodology 1990 = LinguisticChange and Reconstruction Methodology, edited by Ph. Baldi. Berlin /New York 1990 (= Trends in Linguistics, Studies and Monographs45).Lipp Palatale 1994 = R. Lipp Die indogermanischen Palatale im Indoiranischen.(dissertation). Freiburg 1994. Publication is imminent.UV 1998 = Lexikon der indogermanischen Verben, Die Wurzeln und ihrePrimtirstammbildungen, under the direction of H. Rix, revised by M.Kiimmel, Th. Zehnder, R. Lipp, B. Schirmer. Wiesbaden 1998; see below,F 203. New edition, see below, UV 2001. At the time of publication,it was no longer possible to integrate information in LIV 2001.,Thus all references to UV refer to LIV 1998. Reviews of LIV 1998:

344Bibliography and Key to Reference CitationsBibliography and Key 10 Reference Citations 345E. Seebold in IF 104 1999 p. 287-299; Ch. de Lamberterie in BSL 95 122000 p. 139-145. On the new revised edition see LIV 2001.LIV 2001 = Lexikon der indogermanischen Verben. New revised editionby H. Rix and M. Kiimrnel. Wiesbaden 2001.Lockwood Idg. Sprachw. 1982 = W. B. Lockwood IndogermanischeSprachwissenschaft. Tiibingen 1982. English original: Indo-EuropeanPhilology. London 1969.Lockwood Uberblick 1979 = W. B. Lockwood Uberblick iiber die indogermanischenSprachen. Tiibingen 1979. English original: A Panoramaof Indo-European languages. London 1972.Lohmann Genus und Sexus 1932 = J. Lohmann Genus lInd Sexus, Einemorphologische Studie zlIm Ursprung der indogermanischen nominalenGenus-Unterscheidung. Gottingen 1932.Lomrne1 Kleine Schriften 1978 = H. Lomrne1 Kleine Schriften. Wiesbaden1978.Lubotsky Nominal Accentuation 1988 = A. M. Lubotsky The System ofNommal Accentuation in Sanskrit and Proto-Indo-European. Leiden1988. ReView: S. W. Jamison in JAOS III 1991 p. 419-422.Lubotsky RV Word Concordance 1997 = A. Lubotsky A IJgvedic WordConcordance. 2 Parts. New Haven 1997 (= American Oriental Series82-83).Liihr Egil/ 2000 = R. Liihr Die Gedichte des Skalden Egill. Dettelbach2000 (= Jenaer Indogermanistische Textbearbeitllng 1).Liihr Expressivittit 1988 = R. Liihr Expressivittit und Lautgesetz im Germanischen.Heidelberg 1988.Lilhr Hildebrandlied I 1 II 1982 = R. Liihr Studien zur Sprache des Hildebrandliedes.Frankfurt am Main 1982: Part I (Herkunft und Sprache)'Part II (Kommentar).Liihr Neuhochdeutsch 1986 = R. Liihr Neuhochdeutsch, Eine Einfiihrungm d,e Sprachwissenschaft. 4th ed. Munich 1986.Luraghi Casi e preposizioni 1996 = S. Luraghi Studi su casi e preposizioninel greco antico. Pavia 1996.Luraghi Hit/ite 1997 = S. Luraghi Hit/ite. Munich 1 New Castle 1997 (=Languages of the World, Materials 114).Luraghi Old Hittite Sentence Structure 1990 = S. Luraghi Old Hit/iteSentence Structure. London 1 New York 1990.MacDonell Vedic Grammar 1910 = A. A. MacDonell A Vedic Grammar.Strasbourg 1910 (Indian reprints).'Macqueen Hit/ites 1986 = J. G. Macqueen The Hittites and their Contemporariesin Asia Minor. 2nd ed. London 1986. paperback, 1996.Mallory In Search of the Indo-Europeans 1989 = J. P. Mallory In Searchof the Indo-Europeans, Language, Archaeology and Myth. London1989. Critical review: B. Schlerath in Praehistorische Zeitschrift 671992 p. 132-137.Mallory 1 Adams Encyclopedia 1997 = J. P. Mallory and D Q. Adams:Encyclopedia of Indo-European Culture. London and Chicago 1997.Review: SI. Zimrner in TIES 27 1999 p. 105-163 and m Kratylos 452000 p. 46-52; A. Hliusler in IF 105 2000 p. 314-318.Mann lE Comparative Dictionary 1984 1 1987 = S. E. Mann n IndoEuropean Comparative Dictionary. Hamburg 1984 1 1987. CntIcal review:M. Mayrhofer in Kratylos 34 1989 p. 41-45 (p. 45: "ein soschlechtes Buch").Manual de lingiiistica indoeuropea I-III 1995-1998 = F. R. Adrados, A.Bemare, J. Mendoza Manual de lingiiistica indoeuropea. Madnd: I(Prologo, introduccion fo mitica) 1995; Il (Mor/ologia nominal y verbal)1996; III (Morfologia: pronombres, adverbios, particulas y numerales;Syntaxis; Dijferenciacion dialectal) 1998.Marazzi Gerogliftco 1990 = M. Marazzi 11 geroglifico anatolico: Problemidi analisi e prospettive di ricerca. Rome 1990.Marinetti Iscriz. sudpicene 1985 = A. Marinetti Le iscrizioni sudpicene, I:Testi. Florence 1985 (= Lingue e Iscrizioni dell ' Italia Antica 5). Review:G. Meiser in Kratylos 32 1987 p. 110-118.Marouzeau Latin litteraire 1949 = J. Marouzeau Quelques aspects de laformation du latin lit/eraire. Paris 1949.Martfnez Garcia Nombres en -v = F. J. Martfnez Garcia Los nombres en-vdel griego. Frankfurt am Main 1994 (= Europtiische Hochschulschriften,series XXI, vo!. 166).Masson (E.) Immortalite 1991 = E. Masson Le combat pouri'immortalite, Heritage indo-europeen dans la mythologie anatolienne.Paris 1991.Masson (0.) Anatolian Languages 1994 = O. Masson Anatolian Languagesin Cambridge Ancient History, vo!. III 1 2nd ed. Cambndge1991. Phrygian, p. 666-669; Lydian, p. 669-671; Lyclan p. 671-674;Carian p. 674-676.Masson (0.) ICS 1961 + Add. 1983 = O. Masson Les inscriptionschypriotes syllabiques, Recueil critique et com:ente. Pru:is 1961. reprintwith addenda nova 1983. ReView of the l ed. 1961. E. Risch InKratylos 10 1965 p. 88-94.

346 Bibliography and Key to Reference CitationsMatthews Syntax 1981 = P. H. Matthews Syntax. Cambridge et al. 1981.Maurach Lat. Dichtersprache 1983 = G. Maurach Enchiridion Poeticum,Hilfsbuch zur lateinischen Dichtersprache. Darmstadt 1983. 2 " ed.1989.Mayrhofer Altiranische Namen 1979 = M. Mayrhofer Die altiranischenNamen. Vienna 1979. The work is part of the Iranischen Personenna_menbuches, see above.Mayrhofer EW Ala = M. Mayrhofer Etymologisches Worterbuch des Altindoarischen.Heidelberg 1986ff. Vols. I (A-DlI) 1991 and " (N-lI)1996 are of great value.Mayrhofer Indo-Arier 1966 = M. Mayrhofer Die Indo-Arier im AltenVorderasien. Mit einer analytisrhen Bibliographie. Wiesbaden 1966.Cf. also, by the same author, Die Arier im vorderen Orient _ ein Mythos? Vienna 1974 (= SbOA W, 294. vol., 3. Abhandlung).Mayrhofer KEWA = M. Mayrhofer KurzgeJafJtes etymologisches Worterbuchdes Altindischen. 4 vols. Heidelberg 1956-1980.Mayrhofer Kleine Schriften I 1979 " 1996 = M. Mayrhofer AusgewahlteKleine Schriften. Wiesbaden: I 1979 edited by S. Deger-Jalkotzy and R.Schmitt; II 1996 edited by R. Schmitt.Mayrhofer Lautlehre 1986 = M. Mayrhofer Indogermanische Grammatik,vol. I, 2nd half-volume, Lautlehre. Heidelberg 1986. Review: B.Forssman in Kratylos 33 1988 p. 56-63.Mayrhofer Piili 1951 = M. Mayrhofer Handbuch des Piili. Part I: Grammatile;Part 1I: Texte und Glossar. Heidelberg 1951.Mayrhofer Sanskrit-Gramm. 1978 = M. Mayrhofer Sanskrit Grammatikmit sprachvergleichenden Erlauterungen. 3rd ed. Berlin / New York1978 (= Sammlung Goschen Nr. 2207).Mayrhofer Sanskrit und die Sprachen Alteuropas 1983 = M. MayrhoferSanskrit und die Sprachen Alteuropas, Zwei Jahrhunderte des Widerspielsvon Entdeckungen und Irrtumern. Gottingen 1983 (= NA WG,1983, Nr. 5).Mayrhofer Supplement 1978 = M. Mayrhofer Supplement zur Sammlungder altpersischen Inschriften. Vienna 1978 (= SbOA W, 338. vol.).Mayrhofer see also, Brandenstein / Mayrhofer above.McCone Old Irish Nasal Presents 1991 = K. McCone The Indo-EuropeanOrlgms oJ the Old Irish Nasal Presents, Subjunctives and Futures.Innsbruck 1992.McCone Relative Chronology 1996 = K. McCone Towards a RelativeChronology oJ Ancient and Medieval Celtic Sound Change. Maynooth1996 (= Maynooth Studies in Celtic Linguistics I).Bibliography and Key to Reference Citations 347McCone / Sirnms Progress in Medieval Irish Studies 1996 = K. McConeand K. Simms Progress in Medieval Irish Studies. Maynooth 1996.Meid Archtiologie und Sprachwissenschaji 1989 = W. Meid Archtiologieund Sprachwissenschaji, Kritisches zu neueren Hypothesen der Ausbreitungder Indogermanen. Innsbruck 1989 (= IBS, Vortrage undKleinere Schriften 43). Critical review: B. Schlerath in PraehistorischeZeitschrift 67 1992 p. 137-139.Meid Ib/ 1989 = W. Meid Das Problem von indogermanisch /b/. Innsbruck1989 (= Vortrage und Kleinere Schriften 44).Meid Botorrita 1993 = W. Meid Die erste Botorrita-Inschrift: Interpretationeines keltiberischen Sprachdenkmals. Innsbruck 1993 (= IBS Nr.76). Cf. by the same author, Celtihlrian Inscriptions. Budapest 1994and Kleinere keltiberische Sprachdenkmaler. Innsbruck 1996 (= IBS,Vortrage und kieinere Schriften 64).Meid Gaulish inscriptions 1992 = W. Meid Gaulish Inscriptions: TheirInterpretation in the Light oJ Archaeological Evidence and their Valueas a Source oJ Linguistic and Sociological ln/ormation. Budapest 1992(2nd Rev. Ed. 1994).Meid, see also,Krahe / Meid.Meier-Brtigger Gr. Sprachw. I / II 1992 = M. Meier-Briigger GriechischeSprachwissenschaji. 2 vols. Berlin / New York 1992 (= SammlungGoschen Nr. 2241 and Nr. 2242): I (Bibliographie, Einleitung, Syntax);1I (Wortschatz, Formenlehre, Lautlehre, Indizes).Meillet Aper,u 1975 = A. Meillet Aper,u d'une histoire de la languegrecque. Paris 1913. Reprints. 8th ed. 1975 with a bibliography by o.Masson. Regarding the reception of this epoch-making work: A.Morpurgo Davies in Meillet e/ la linguistique de son temps 1988 p.235-252.Meillet Armenien classique 1936 = A. Meillet Esquisse d'une grammairecompart!e de l'armenien classique. 2nd edition, completely revised. Vienna1936.Meillet Esquisse 1928 = A. Meillet Esquisse d'une histoire de la languelatine. Paris 1928.Me illet et la linguistique de son temps 1988 = Antoine Meillet et la linguistiquede son temps, edited by S. Auroux. Lille 1988 (= HEL [Histoire,Epistemologie, Langage 1 10 / Il).Meillet Introduction 1949 = A. Meillet lntroduction a l'etude comparativedes langues indo-europeennes. Paris 1949 (reprint of the 8th ed. Alabama1964).

348Bibliography and Key to Reference Ci' • .. I.4liol)sMeillet Linguislique hislori qze el r "A.mgulsllqueMeillet Lingenerale Iisli ue1921 II:;;IU;19361921,el II 1936ImgulSlique(!:'d v!iousgenerale. Paris IMellIet Mhhode comparmive 192;:enA M 'II elinguisliqueL melhode 'hislorique. Oslo 1 Pari 1;comp araliveMeillet 15Vendryes Grammaire comparee 1924 repnm s).dryes TraiM de- A. Meilletgrammaireandcom 1.a 'Ven_1924. r;".e dAcomparablelangueswork isc/assiques.Smkr ewParisomparallve Grammar 1995(q. v.).Meiser (Hisl.) Laul- u. Formenlehr e (d. lal.Hislorische Sprache)LaUl- und1998 =FormO. Meiserstadt 1998. Review ' B V ' nt;;;re der Ia/einischen:"Sprache. mDarm_MeiserarylosPeifekl 1991 G46 200 I p. I 18-126.elserlaleinischen undPeifeklsYSI;ms .AusbildungPOs tdVorgeschichledes1991 .octoral work Freiburg un '.B relsgauMeiser Umbrisch 1986 = 0 M : 'che.LaInnsbruckulgeschichle1986der(= IBSumbrischenI)Spra-Melster Homer. Kunslsprache 1921 ' = K.. .Kunslsprache.MelsterLeipzigDie1921.homerischeMelSterbans 1 Schwyzer All. .derInschr. 1900 =allischen Inschrifien. 3rd.K. Melsterhansed .,Grammalikzer.revIse dBerlinand1900.expanded by E. Schwy-Melchert Ab/. and Inslr. 1977 = HinC ' M e I cHillile.h ert Ablalive and (dissertation). Harvard 1Inslrumenlal977Melchert Ana/olian HisloricalbanPhonolo gy . 1994 -Hislorical Phonology Am t d- H. C. Melchert AnaloRieken in BiOr 55 1998 Sp: 473:45 . a;:; 6 Atlanta . 1994. Review: E.p. 96-1'08.. ettmger m Kra/ylos 43 1998Melchert Cuneiform Luvian 1993 = H .Lexicon.C.ChapelMelchertHill 1993.Cuneiform LuvianMelchert Hillile HisloricalHl IlIePhonology 1984 =Hislorical Phonology . O"tf 0 mgen 1984 (=ZVS, Ergtinzungsheji32).H. C. Melchert Sludies inMelchert Lycian 1993 = H C. Med. Chapel.Hill 1993'elchert LycIQn Lexicon. 2nd fully revisedMendez Dosuna Noroe;le 1985 = JdoriosM' ddel 0norOesle). Salamanca i 98 n ez osuna Noroesle (DialeclosMelzler Lexikon Sprache 1993 and .2000 =lted by H. GUick. I st cd. Slut! g art. Melzler Lexlkon Sprache, 1 Wedpandeded.eunar2000.1993. 2nd revised and ex-Bibliography and Key to Reference Citations 349Meyer Lat. Epigraphik 1973 = E. Meyer Einfuhrung in die lateinischeEpigraphik. Darmstadt 1973.MH = Museum Helveticum.Mikkola Abstraklion 1964 = E. Mikkola Die Abstraklion, Begriff undStruktur, Eine logisch-semantische Untersuchung auf nominalistischerGrundlage unter besonderer Berucksichligung des Lateinischen. Helsinki1964.Mikkola Kompositum I 1971 = E. Mikkola Das Kompositum, vol. I (seeparticularly the first chapter, 'Die Zusammensetzung, Begriff undSlruktur' p. 5ff. and the bibliography p. 56ff.). 2nd ed. 1971.Miklosich Vergl. Syntax 1868-1874 = Fr. Miklosich Vergleichende Syntaxder slavischen Spraehen. Vienna 1868-1874.Minos = Minos. Revista de jilologia egea. Salamanca.Monro Homer. Dia/eel 1891 = D. B. Moneo A Grammar of the HomericDialect. 2nd ed. Oxford 1892.Morpurgo Davies Ollocento 1996 = A. Morpurgo Davies La linguislicadell ' Olloeento. Bologna 1996.MSS = Munchener Studien zur Sprachwissenschaji.Muller Norninativ und Akkusativ 1908 = C. F. W. MiiUer Syntax des Norninativsund Akkusativs im Lateinischen. Leipzig 1 Berlin 1908.Mutzbauer Or. Tempuslehre I 1893 2 1909 = C. Mutzbauer Die Grundlagender grieehischen Tempuslehre und der homerische Tempusgebrauch.Ein Beitrag zur historisehen Syntax der griechischen Sprache.2 vols. Strasbourg 1893-1909.Mutzbauer Konj. und Opl. 1908 = C. Mutzbauer Die Grundbedeutung desKonjunklivs und Optalivs und ihre EnIWieklung im Griechischen. Leipzig1 Berlin 1908.Mykenai"ka 1990 [1992] = Mykenai"ka. Acles du IXe Colloque internationalsur les textes myceniens et egeens 1990 in Athens. Paris 1992 (=BCH, Suppl. 25).Nagy = Greek Dialects 1970 = O. Nagy Greek Dialects and the Transformationof an Indo-European Process. Cambridge 1 Mass. 1970.Review: A. Morpurgo Davies in CR 22 1972 p. 371-374; C. J. Ruijghin Scripta Minora I 1991 p. 635-650.Namenforschung 1 1995 1 2 1996 = Namenforschung / Name Studies / Lesnoms propres, Ein internalionales Handbuch zur Onomaslik, edited byE. EicWer, G. Hilty, H. Llifl:ler, H. Steger, L. Zgusta. Berlin 1 NewYork. 2 vols.: I 1995; 2 1996 (= HSK 11.1 and 11.2). Review: R.Bergrnann in BNF N.F. 32 /4 1997 p. 457-471.

350 Bibliography and Key to Reference CitationsNarten Kleine Schriften I 1995 = J. Narten Kleine Schriften, edited by M.Albino and M. Fritz. vo!. I Wiesbaden 1995.Narten Sigmalische Aorisle 1964 = J. Narten Die sigmalischen Aorisle imVeda. Wiesbaden 1964.Narten Yasna Hapla1Jhaili 1986 = J. Narten Der Yasna Haplat/haili.Wiesbaden 1986.Narten, see also Hoffinann / Narten.NA WG = Nachrichlen der Akademie der Wissenschafien in Gollingen.Neu Ahelh. Glossar 1983 = E. Neu Glossar zu den allhelhilischen Rituailexlen.Wiesbaden 1983 (= StBoT 26).Neu Ahelh. Riluallexle 1980 = E. Neu Allhelhilische Riluailexle in Umschrift.Wiesbaden 1980 (= StBoT 25).Neu Anilla 1974 = E. Neu Der Anilla-TexI. Wiesbaden 1974 (= StBoT18). See also: -) G. Steiner SIruktur und Bedeulung des sog. AnillaTextes in OA 23 1984 p. 53-73.Neu Hurrilisch 1988 = E. Neu Das HlIrritische: Eine altorientalischeSprache in neuem Licht. Mainz 1988 (= AA WL 1988 Nr. 3).Neu Interpret. Mediopassiv 1968 = E. Neu Interpretation der hethilischenmediopassiven Verbalformen. Wiesbaden 1968 (= StBoT 5).Neu Lokativ 1980 = E. Neu Studien zum endungslosen "Lokativ" des Hethitischen.Irmsbruck 1980 (= ms, Vortrage und kIeinere Schriften 23).Neu Mediopassiv 1968 = E. Neu Das hethitische Mediopassiv und seineindogermanischen Grundlagen. Wiesbaden 1968 (= StBoT 6).Neu, see also ROster / Neu.Neumann Indogermanistik 1967 = G. Neumann IndogermanischeSprachwissenschafi 1816 and 1966. Innsbruck 1967.Neumann Kleine Schriften 1994 = G. Neumann Ausgewiihlte kleineSchriften. edited by E. Badali, H. Nowicki and S. Zeilfelder. Innsbruck1994 (= lBS 77).Neumann Phrygisch und Griechisch 1988 = G. Neumann Phrygisch undGriechisch. Vienna 1988 (= SbOAW 499).Neumann Weilerleben 1961 = G. Neumann Untersuchungen zum Weiterlebenhelhilischen und luwischen Sprachgutes in hellenistischer undromischer Zeil. Wiesbaden 1962.Neve ljat/usa 1996 = P. Neve ljallusa - Siadl der Goller und Tempel,Neue A usgrabungen in der Hauplstadl der Helhiler. 2nd ed. Mainz1996 (= Sonderhefi der Anliken Weil).New Sound of Indo-European 1989 = The New Sound of Indo-European,Essays in Phonological Reconslruclion, edited by Th. Vennemann.Bibliography and Key to Reference Citations 351Berlin 1989 (= Trends in Linguistics 41). Review: O. Szemerenyi inDiachronica 6 1989 p. 237-269.Noreen Altisliindisch und Ailnorwegisch 1923 = A. Noreen Allisliindischeund altnorwegische Grammalik unler Beriicksichtigung des Urnordischen.Halle 1923.Noreen AIIschwedisch 1904 = A. Noreen Altschwedische Grammalik, MilEinschlufJ des Allgutnischen. Halle 1904.Nowicki s-Sltimme 1976 = H. Nowicki Die neutralen s-SIiimme im indoiranischenZweig des Indogermanischen. (dissertation in philology).WUrzburg 1976.NTS = Norsk Tidskriftfor Sp rogvidenskop.Numerals 1992 = Indo-European Numerals, edited by J. Gvozdanovic.Berlin I New York 1992 (= Trends in Linguistics, SIudies and Monographs57). Concerning the work, see below F 500 § 3b.Nussbaum Head and Horn 1986 = A. J. Nussbaum Head and Horn inIndo-European. Berlin I New York 1986. Review: R. S. P. Beekes inKratylos 34 1989 p. 55-59.Nussbaum Two SIudies 1998 = A. J. Nussbaum Two Studies in Greek andHomeric Linguislics. G6ttingen 1998 (= Hypomnemata 120). Review:B. Forssman in Kratylos 46 2001 p. 1l3-1 17.OA = Oriens Antiquus.Odyssey J (Books i-viii) 1988 IT (Books ix-xvi) 1989 III (Books xvii-xxiv)1992 = A Commenlary on Homer's Odyssey. Oxford: I by A. Heubeck(with General Introduclion), S. West (with The Transmission of IheTexI and Books i-iv) and J. B. Hainsworth (with The Epic Dialeel andBooks v-viii); II by A. Heubeck (with Books ix-xii) and A. Hoekstra(with Books xiii-xvi) 1989; III by J. Russo (with Books xvii-xx), M.Ferruindez-Galiano (with Books xxi-xxii) and A. Heubeck (with Booksxxiii-xxiv) 1992. The Oxford edition is a "Revised version, without textand translation, of the six-volume edition commissioned by the FondazioneLorenzo Valla and published by Mondadori Milano 1981-1986".Oertel Dalivi finales 1941 = H. Oertel Die Dativi finales abslrakter Nominaund andere Beispiele nominaler Satzfiigung in der vedischenProsa. Munich 1942.Oettinger Verbum 1979 = N. Oettinger Die Stammbildung des hethitischenVerbums. NUmberg 1979. The same author offers the supplementaryDie hethitischen Verbalstiimme in Grammatica It/ita 1992 p.214-252.

352 Bibliography and Key to Reference CitationsOhlstadt 1994 [1996] = Tagungsband "Hellenische Mythologie / Vorgeschichte" (title also in Modem Greek) 1994 in Ohlstadt / Oberbayem,edIted by N. Dimoudis and A. Kyriatsoulis. Altenburg 1996 (organizers:Verein zur Forderung der Aufarbeitung der hellenischenGeschichte Weilheim (Upper Bavaria) and the Club GriechischerAkademiker Miinchen).Ohlstadt 1996 [1998] = Tagung " Die Geschichte der hellenischen Spracheund Schrift" (title also in Modem Greek and English) 1996 in Ohlstadt/ Oberbayem, edited by N. Dimoudis and A. Kyriatsoulis. Altenburg1998 (organizers: Verein zur Forderung der Auforbeitung derhellenischen Geschichte Weilheim [Upper Bavaria]).OLD = Oxford Latin Dictionary. Oxford 1968-1982.01denberg Kleine Schriften 1967 = H. Oldenberg Kleine Schriften, editedby K. L. Janert. 2 parts. Wiesbaden 1967.01denberg Noten 1909-1912 = H. Oldenberg 8gveda, Textkritische undexegetische Noten. I - VI Berlin 1909, VII - X Berlin 1912.Old English Runes 1991 = Old English Runes and their ContinentalBackground, edited by A. Bammesberger. Heidelberg 1991.Olsen Instrument Noun Suffix 1988 = B. A. 01sen The Proto-IndoEuropean Instrument Noun Suffix "-tlom and its variants. Copenhagen1988.01sen Noun 1999 = B. A. 01sen The Noun in Biblical Armenian, Originand Word-Formation - with special emphasis on the Indo-Europeanhen/age. Berlin 1999 (= Trends in Linguistics, Studies and Monographs119). Review: S. Zeilfelder in Sprache 40 1998 p. 105-109; R.Schmitt in Kratylos 46 200 I p. 80-88.OLZ = Orientalistische LileraturzeilungPaImer Descriptive and Comparative Linguistics 1972 = L. R. PaImerDescriplive and Comparative Linguistics, A Critical Introduction.London 1972.Palmer Greek Language 1980 = L. R. Palmer The Greek Language London/ Boston 1980. German translation by W. Meid. Innsbruck 1986 (=IBS Nr. 50).Panzer Slav. Sprachen 1991 = B. Panzer Die slavischen Sprachen in Gegenwartund Geschichte. Sprachstruk/uren und Verwandtschaft.Frankfurt am Main 1991 (= Heidelberger Publikalionen zur Slavistik,A. Linguistische Reihe 3).Partherreich [1996] 1998 = Das Partherreich und seine Zeugnisse / TheArsacid Empire: Sources and Documentation, Beitrage des interna-Bibliography and Key to Reference Citations 353tionalen Colloquiums 1996 in Eutin, edited by J. WiesenhOfer. Stuttgart1998.PBB = Beitrage zur Geschichte der deutschen Sprache und Literatur.(Halle and) Tilbingen (The abbreviation comes from H. Paul, W.Braune, Beilrage ... ).Pedersen Cinquieme declinaison 1926 = H. Pedersen La cinquieme declinaisonlaline. Copenhagen 1926.Pedersen Hillitisch 1938 = H. Pedersen Hillitisch und die anderen indoeuropaischenSprachen. Copenhagen 1938.Pedersen KI. Schr. zum Arm. 1982 = H. Pedersen Kleine Schriften zumArmenischen, edited by R. Schmitt. Hildesheim / New York 1982.Pedersen Tocharisch 1941 = H. Pedersen To charisch vom Gesichtspunktder indoeuropaischen Sprachvergleichung. Copenhagen 1942.Pedersen Vgl. Gramm. d. kelt. Spr. I 1909 1I 1913 = H. Pedersen VergleichendeGrammatik der keltischen Sprachen. 2 vols. Giittingen: 1(Einleitung und Lautlehre) 1909; I1 (Bedeutungslehre) 1913.. .Pellegrini / Prosdocimi Lingua Venetica I / Il 1967 = G. B. PellegnTU andA. L. Prosdocimi La lingua Venelica. Padua 1967. I: Le iscrizioni byG. B. Pellegrini and A. L. Prosdocimi; II: Studi by A. L. Prosdocimi.Peters Laryngale 1980 = M. Peters Untersuchungen zur Vertretung derindogermanischen Laryngale im Griechischen. Vienna 1980. Revịew:R. S. P. Beekes in Kratylos 26 1981 [1982] p. 106-115; J. Catsarucosin BSL 77 / 2 1982 p. 89-95; B. Forssman in ZVS 96 1982 [1983] p.290-292; C. J. Ruijgh in Mnemosyne 36 1983 p. 373-380.Petit 'sue- 1999 = D. Petit "Sue- en grec ancien: La famille du pronomrejIe hi, Linguistique grecq e et comparaison indo-europEienne. Louvan1999 (= Collection Linguistique, publiee par la Societe de Linguistiquede Paris, 79).Petit Lituanien 1999 = D. Petit Lituanien. Paris 1999 (in LALIES 19,Aussois 1998).Pfeiffer Klass. Philologie I 1970 = R. Pfeiffer Geschichte der klassischenPhi/ologie, Von den Anfangen bis zum Ende des Hellenismus. Reinbeknear Hamburg 1970. Klass. Philologie is a translation (by M. Amold)of the English original: "History of Classical Scholarship". Oxford1968.PFU = Phi/ologia Fenno-Ugrica..Pinault Tokharien 1989 = G.-J. Pinault Introduction au tokharien. ParIs1989 (in LALIES, A ctes des sessions de Iinguistique et de !illeratureNr.7, Aussois 1985).

354 Bibliography and Key to Reference CitationsPinkster Lateinische Syntax 1988, i.e. Latin Syntax 1990 = H. PinksterLateinische Syntax und Semantik. Tiibingen 1988 (= UTB Nr. 1462published by the Francke Verlag). The text is a translation of the original,published in Dutch in 1984. An English version appeared in 1990in London I New York with the title: Latin Syntax and Semantics.Place de I' Armenien 1986 = La place de I'armenien dans les languesindo-europeennes, edited by M. Leroy and F. Mawet. Louvain 1986.P1ath Streitwagen 1994 = R. Plath Der Streitwagen und seine Teile imfrl1hen Griechischen. Sprachliche Untersuchungen zu den mykenischenTexten und zum homerischen Epos. Niimberg 1994 (= Erlanger Beitragezur Sprache, Literatur und Kunst 76). Review: I. Hajnal in Kratylos42 1997 p. 78-8 1.Poccetti Nuovi Documenti Italici see below, s. v. Vetter.Pokorny lEW 1959-1969 = J. Pokorny Indogermanisches etymologischesWorterbuch. 1/1l Bern I Munich 1959-1969.Portraits T III 1966 = Portraits of Linguists, A Biographical Source Bookfor the History of Western Linguistics, 1764-1963, edited by Th. A.Sebeok. Bloornington I London 1966: J (From Sir William Jones 10Karl Brugmann); Il (From Eduard Sievers 10 Benjamin Lee Whorf).Porzig Gliederung 1954 = W. Porzig Die Gliederung des indogermanischenSprachgebiets. Heidelberg 1954.Porzig Satzinhalte 1942 = W. Porzig Die Namen fur Satzinhalte im Griechischenund im Indogermanischen. Berlin 1942 (= Untersuchungenzur indogermanischen Sprach- und Kullurwissenschajt 10).POrzIg Wunder der Sprache 1971 = W. Porzig Das Wunder der Sprache,Probleme, Methoden und Ergebnisse der Sprachwissenschajt. 5th ed.revised by A. Jecklin and H. Rupp. Munich 1971 (= UTB 32). 8. ed.1986. The ftrst edition of this outstanding work was published in 1950.Prins Hillite neuter 1997 = A. Prins Hillite neuter singular - neuterplural, Some Evidence for a Connection. Leiden 1997.Probleme der lat. Gramm. 1973 = Probleme der lat. Grammatik, editedby K. Strunk. Darmstadt 1973 (= Wege der Forschung 93).Probleme der Namenbildung 1986 [1988] = Probleme der Namenbildung,RekonslrukJion von Eigennamen und der ihnen zugrundeliegenden Appellative,AkJen eines internationalen Symposiums 1986 in Uppsala'edited by Th. Andersson. Uppsala 1988.Puhvel Analecta Indoeuropea 1981 = 1. Puhvel Analecla Indoeuropea,Delectus operum minorum ... annos 1952-1977 complectens. Innsbruck1981 (= IBS 35).Bibliography and Key to Reference Citations 355Puhvel HED 1 /2 1984, 3 1991, 4 1997, 5 2001 = J. Puhvel Hillite EtymologicalDictionary. Berlin I New York I Amsterdam: vo!. 1 (Wordsbeginning wilh A) and 2 (Words beginning with E and I) 1984; vo!. 3(Words beginning with H) 1991; vo!. 4 (Words beginning wilh K)1997; vo!. 5 (Words beginning with L) 2001. Concerning vo!. 4, cf. I.Hajnal in HS 112 1999 p. 305-315Puhve1 Laryngeals and Ihe lE Verb 1960 = J. Puhvel Laryngeals and theIndo-European Verb. Berkeley 1960. Review: W. Cowgill in Language39 1963 p. 248-270.Radke Archaisches Latein 1981 = O. Radke Archaisches Latein. Darrnstadt1981 (= Ertrage der Forschung 150).Ramat see above, Lingue indoeuropee 1994.Rasmussen Morphophonemalik 1989 = J. E. Rasmussen Studien zur Morphophonematikder indogermanischen Grundsprache. Innsbruck 1989(= IBS Nr. 55).Rasmussen Selected Papers 1999 = J. E. Rasmussen Selected Papers onIndo-European Linguistics, 2 vols. Copenhagen 1999.Raulwing Horses 2000 = P. Raulwing Horses, Chariots and IndoEuropeans, Foundations and Melhods of Chariotry Research from IheViewpoint of Indo-European Linguistics. Budapest 2000. Review: K.Jones-Bley in JIES 28 2000 p. 440-449.REArm = Revue des Etudes Armeniennes.Redei Idg. -ural. Sp rachkontakte 1986 = K. Redei Zu den indogermanischuralischenSp rachkontakJen. Vienna 1986 (= SbClAW 468).Rehder Slav. Sprachw. 1998 = P. Rehder Einfuhrung in die slavischeSprachwissenschajt. Darmstadt 3rd ed. 1998.Reichler-Beguelin Type mens 1986 = M.-J. Reichler-Beguelin Les nomslatins du type mens. Brussels 1986 (= Collection Lalomus vo!. 195).Review: F. Mawet in Kratylos 34 1989 p. 96-102.REL = Revue des etudes latines.Renfrew Archaeology and Language 1987 = C. Renfrew Archaeology andLanguage: The Puzzle of Indo-European Origins. London 1987. Review:E. Campanile in Kratylos 33 1988 p. 53-56.Renou Bibliographie 1997 = Bibliographie des travaux de Louis Renou,by O. Pinault. Paris 1997 (= a supplement to the Bulletin d'etudes indiennes13-14 1995-1996 [1997]).Res Mycenaeae 1981 [1983] = Res Mycenaeae, AkJen des Vll. InternalionalenMy kenologischen Colloquiums 1981 in Nurnberg, edited byA. Heubeck and O. Neumann. 05ttingen 1983. Review: M. Peters in

356 Bibliography and Key to Reference CitationsIdg. Chr. 30a Nr. 581; W. B1Ume1 in GGA 236 1984 p. 121-136; F.Bader in Kratylos 30 1985 p. 105-1 12.RGA = Reallexikon der Germanischen Altertumskunde, founded by J.Hoops. 2nd ed., fully revised and considerably expanded by numerousexperts under the editorship of R. Miiller edited by H. Beck, H. Steuerand D. Tiempe. Status of publication: vo!. 1 1973 - vo!. 18 2001.Rieckejan-Verben 1996 = J. Riecke Die schwachen jan- Verben des Althochdeutschen,Ein Gliederungsversuch. Gottingen 1996 (= Studienzum Althochdeutschen 32). Review: O. W. Robinson in Kratylos 441999 p. 127-130.Rieken Nom. Stammbildung 1999 = E. Rieken Untersuchungen zur nominalenStammbildung des Hethitischen. Wiesbaden 1999 (StBoT 44).RlG = Recueil des Inscriptions Gauloises. Paris. vo!. I 1985: M. LejeuneTextes gallo-grecs; vo!. IT 1 1988: M. Lejeune Textes Gallo-Etrusques,Textes Gallo-Latins sur pierre; vo!. III 1986: P.-M. DuvaJ Les Calendriers.Rijksbaron Verb in Class. Greek 1994 = A. Rijksbaron The Syntax andSemantics of the Verb in Classical Greek. 2nd ed. Amsterdam 1994.Ringe Sound Changes in Tocharian I 1996 = D. A. Ringe On the Chronologyof Sound Changes in Tocharian. I (From Proto-IE to ProtoTocharian). New Haven 1996 (= American Oriental Series 80).Risch Gerund. 1984 = E. Risch Gerundivum und Gerundium. Gebrauchim klassischen und iilteren Latein. Entstehung und Vorgeschichte. Berlin/New York 1984.Risch Kleine Schrijien 1981 = E. Risch Kleine Schrijien. edited by A. Etterand M. Looser. Berlin I New York 1982.Risch WOr/bildung 1974 = E. Risch Wortbildung der homerischen Sprache.2nd ed. Berlin 1974.Ritter Armeno antiguo 1996 = R.-P. Ritter Introduccion al armeno antiguo.Madrid 1996.Rix Etr. Texte I 111 1991 = H. Rix Etrllskische Texte. Editio minor, editedby H. Rix together with G. Meiser. Tiibingen 1991: vo!. T (Einleitung,Konkordanz, Indizes); vo!. TT (Texte).Rix Gentilnamensystem 1972 = H. Rix Zum Ursprung des romischmittelitalischenGentilnamensystems. Publishe as part of Aufstieg undNiedergang der romischen Welt, Geschichte und Kultur Roms imSpiegel der neueren Forschung, edited by H. Temporini. vo!. I I 2.Berlin I New York 1972 p. 700-758.Rix Hist. Gramm. d. Gr. 1976 = H. Rix Historische Grammatik des Griechischen.Darmstadt 1976. 2" revised ed. 1992. Review of the 2ndBibliography and Key to Reference Citations 357ed.: G. DunkeJ in AJPh 97 1976 p. 416-420; M. Peters in Sprache 231977 p. 65-67; F. Bader in BSL 72 I 2 1977 p. 134-140; C. J. Ruijgh inMnemosyne 31 1978 p. 298-307.Rix Kleine Schrijien 2001 = H. Rix Kleine Schrijien, Festgabe fur HelmutRix zum 75. Geburtstag. Bremen 2001.Rix Modussystem 1986 = H. Rix Zur Entstehung des urindogermanischenModussystems. Innsbruck 1986 (= IBS, Vortriige und KleinereSchrijien Nr. 36). Review: E. Risch in Kratylos 32 1987 p. 46-50.Rix Riitisch und Etruskisch 1998 = H. Rix Riitisch und Etruskisch. Innsbruck1998 (= IBS, Vortriige und Kleinere Schrijien 68).Rix Termini der Unfreiheit 1994 = H. Rix Die Termini der Unfreiheit inden Sprachen Alt-Italiens. Stuttgart 1994 (= Forschungen zur AntikenSkJaverei 25).Rosen Periphrase 1992 = H. B. Rosen Die Periphrase, Wesen und Entstehung.Innsbruck 1992 (= IBS, Vortriige und kleinere Schrijien 57).RPh = Revue de Philologie.Rubio Orecilla Sufljo de derivacion nominal 1995 = F. J. Rubio OrecillaEl sufijo de derivacion nominal *-ijo-I *-jo- en los gerundios y gerundivosdel Rg- Veda y el Avesta, Un estudio historico-comparativo.Zaragoza 1995. Review: M. KUmmeI in Kratylos 43 1998 p. 81-83; J.Haudry in BSL 93 I 2 1995 p. 134-139; see below, W 202 § 1.Riister I Neu Heth. Zeichenlexikon 1989 = Chr. Riister and E. Neu HethitischesZeichenlexikon, Inventar und Interpretation der Keilschrijizeichenaus den Bogazlcoy-Texten. Wiesbaden 1989.Ruijgh Etudes 1967 = C. J. Ruijgh Etudes sur la grammaire et le vocabulairedu grec mycenien. Amsterdam 1967.Ruijgh Scripta Minora J 1991 JI 1996 = C. J. Ruijgh Scripta Minora adlinguam Graecam pertinentia I, edited by J. M. Bremer, A. Rijksbaron,F. M. J. Waanders. Amsterdam 1991; Scripta Minora ad linguamGraecam pertinentia TT, edited by A. Rijksbaron, F. M. J. Waanders.Amsterdam 1996.Ruijgh " te epique " 1971 = C. J. Ruijgh Autour de "te epique ", Etudes surla syntaxe grecque. Amsterdam 1972.Ruiperez lIias und Odyssee 1999 = M. S. Ruiperez Anthologie lIias undOdyssee. Wiesbaden 1999. Translation of the Spanish original that waspublished in 1963.Ruiperez Opuscula 1989 = M. S. Ruiperez Opuscula selecta, AusgewiihlteArbeiten zur griechischen und indogermanischen Sprachwissenschajt,edited by J. L. Garcia-Rarn6n. Innsbruck 1989 (= IBS Nr. 58).

358 Bibliography and Key to Reference CitationsBibliography and Key to Reference Citations359Sandhi Phenomena 1986 = Sandhi Phenomena in the Languages 0/Europe, edited by H. Andersen. Berlin I New York I Amsterdam 1986.Saussure Cours 1916 = F. Saussure Cours de linguistique gemirale, editedby Ch. Bally I A. Sechehaye I A. Riedlinger. Paris 1916 (and reprints).Concerning the diachronic aspect of this work: -4 P. Wunderli Principesde diachronie, Contribution a I'exegese du " Cours ". Frankfurt1990 (an announcement of the publication by M. Mayrhofer appearedin Idg. Chr. 34 Nr. A 89).Saussure Memoire 1879 = F. Saussure Memoire sur le systeme primilijdes voyelies dans les langues indo-europeennes. Leipzig 1879(Wiederabdruck in Saussure Recueil 1922 p. 1 ff.; separate reprint Hildesheim1987). Concerning the work and its reverberations: -4 M. Mayrhofer Nach hundert Jahren, Ferdinand de Saussures Friihwerk undseine Rezeption durch die heutige Indogermanistik, Heidelberg 1981 (=SbHA W 1981 I 8). Further, see above Gmiir Memoire 1986, as wellasSaussure Saggio below.Saussure Recueil 1922 = F. Saussure Recueil des publications scientifiquesde F. de Saussure. Geneva 1922 (and reprints).Sausse Sa gg io 1987 = F. Saussure Saggio sui vocalismo indoeuropeo,Itahan edlllon of the Memoire with an introduction and comments by G.C. Vincenzi. Bologna 1987.SbBA W = Sitzungsberichte der Bayerischen Akademie der Wissenschaften,phil.-hist. Klasse, Munich.SbHA W = Sitzungsberichte der Heidelberger A kademie der Wissenchafien, phil. -hist. Klasse, Heidelberg.SbOA W = Silzungsberichte der Osterreichischen Akademie der Wissenschajien,phil.-hist. Klasse, Vienna.Scardigli Goten 1973 = P. Scardigli Die Goten, Sprache und Kultur. Munich1973.Scardigli Weg zur deutschen Sprache 1994 = P. Scardigli Der Weg zurdeutschen Sprache, Von der indogermanischen bis zur Merowingerzeit.Bern et al. 1994 (= German. Lehrbuchsammlung 2).Scarlata Wurzelkomposita im lJg- Veda 1999 = S. Scar lata Wurzelkomposilaim lJg- Veda. Wiesbaden 1999.Schaefer fntensivum 1994 = Ch. Schaefer Das Intensivum im Vedischen (=ZVS Ergtinzungshefl 37). Review: E. Seebold in IF 101 1996 p. 299-302; SI. W. Jamison in Kratylos 42 1997 p. 50-55; A. Lubotsky inJAOS 117 1997 p. 558-564.Scherer Lat. Syntax 1975 = A. Scherer Handbuch der lateinischen Syntax.Heidelberg 1975.Schindler Wurzelnomen 1972 = 1 Schindler Das Wurzelnomen imArischen und Griechischen. (dissertation) Wiirzburg 1972.Schirmer Wortschatz 1998 = B. Schirmer Studien zum Wortschatz derfguvinischen Ta/eln. Frankfurt am Main et al. 1998 (= EuroptiischeHochschulschriften, Reihe XXI Linguistik, vo!. 205). Review: M.Weiss in Kratylos 46 2001 p. 131-134.Schleicher Compendium 1866 = A. Schleicher Compendium der vergleichendenGrammatik der indogermanischen Sprachen, KurzerAbrifl einer Laut- und Formenlehre der indogermanischen Ursprache,des Altindischen, Alteranischen, Altgriechischen, Altitalischen, Altkeltischen,Altslawischen, Litauischen und Altdeutschen. 2nd ed. WeimarI London I Paris 1866 (1 st ed. 1861-1862).Schlerath fndogermanen 1973 = B. SchJerath Die Indogermanen. DasProblem der Expansion eines Volkes im Lichte seiner Sozialstruktur.Innsbruck 1973 (= ms, Vortrtige 8). Review: K. Strunk in BNF N.F. 91974 p. 388-390.Schlerath Kleine Schriften 2000 = B. Schlerath Kleine Schriften. 2 vols.Dettelbach 2000.Schrnid Schriften 1994 = W. P. Schrnid Linguislicae scientiae co/lectanea,Ausgewtihlte Schriften, edited by 1 Becker. Berlin I New York 1994.Schrnid Studien 1963 = W. P. Schrnid Studien zum baltischen und indogermanischenVerbum. Wiesbaden 1963.Schmidt (G.) Personalpronomina 1978 = G. Schrnidt Stammbildung undFlexion der indogermanischen Personalpronomina. Wiesbaden 1978.Schrnidt (l) Neutra 1889 = 1 Schrnidt Die Pluralbildungen der indogermanischenNeutra. Weimar 1889.Schrnidt (l) Verwantschqftsverhtiltnisse 1872 = J. Schrnidt Die Verwantschaflsverhtiltnisse(sic !) der indogermanischen Sp rachen. Weimar1872.Schrnidt (J.) Vocalismus I 1871 Il 1875 = 1 Schrnidt Zur Geschichte desindogermanischen Vocalismus. Weimar: T 1871, II 1875.Schrnidt (K. H.) Celtic 1996 = K. H. Schrnidt Celtic: A Western IndoEuropean Language? Innsbruck 1996 (= IBS, Vortrtige und KleinereSchriften 66).Schrnidt (K. T.) Medium im Toch. = K. T. Schmidt Die Gebrauchsweisendes Mediums im Tocharischen. Gottingen 1974 (dissertation, philology)Gottingen 1969.Schrnitl Ap. Inschriften 1999 = R. Schrnitl Beilrtige zu altpersischen Inschriften.Wiesbaden 1999.

360Bibliography and Key to Reference CitationsBibliography and Key to Reference Citations 361Schmitt Bisitun Inscriptions 1991 = The Bisitun Inscriptions of Darius theGreat, Old Persian Text, edited by R. Schmitt. London 1991 (= Schoolof Oriental and African Studies, Corpus Inscriptionum Iranicarum, Pt.I (Inscriptions of Ancient Iran), Vo!. I (The Old Persian Inscriptions,Texts 1). Review: Ch. H. Werba in Sprache 35 (1991-1993) p. 140-145; S. W. Jamison in IlJ 37 (1994) p. 168-171. Also by R. Schmitt,Epigraphisch-ex getische Noten zu Dareios' Blsitin-Inschrijien. Vierma1990 (= SbOA W, 561. vo!.); and also Ap. Inschrijien 1999.Schmitt Dichtersprache 1967 = R. Schmitt Dichtung und Dichtersprachein indogermanischer Zeit. Wiesbaden 1967.Schmitt Gr. Dialekte 1977 = R. Schmitt Einfiihrung in die griechischenDialekte. Darmstadt 1977.Schmitt lr. Sprachen 2000 = R. Schmitt Die iranischen Sprachen inGeschichte und Gegenwart. Wiesbaden 2000.Schmitt Klass. Arm. 1981 = R. Schmitt Grammatik des KlassischArmenischen mit sprachvergleichenden Erltiuterungen. Innsbruck 1982(= IBS Nr. 32).Schmitt-Brandt Einfiihrung 1998 = R. Schmitt-Brandt Einfiihrung in dieIndogermanistik. TUbingen / Basel 1998 (= UTB Nr. 1506). Review:G. Keydana in IF 104 1999 p. 281-286; K. Stelter in Kratylos 46 2001p. 200-202; see above, p. VIn.Schneider Lautgesetz 1973 = G. Schneider Zum BegrijJ des LautgesetzesIn der Sprachwissenschaji seit den Junggrammatikern. TUbingen 1993(= TBL 46).Schon Neulrum und Kollektivum 1971 = I. Schon Neutrum und Kollektivum,Das Morphem -a im Lateinischen und Romanischen. Innsbruck1971 (= IBS Nr. 6). Review: A. Morpurgo Davies in CR 25 1975 p.248f.Schriji und Schrijilichkeit I 1994 2 1996 = Schriji und Schrijilichkeit /Writing and its Use, Ein interdisziplintires Handbuch internationalerForschung, edited by H. Giinther, O. Ludwig. Berlin / New York. 2parts: I 1994; 2 1996 (= HSK 10.1 and 10.2).Schrijver British Celtic Phonology 1995 = P. Schrijver Studies in BritishCeltic Historical Phonology. Amsterdam / Atlanta 1995.Schrijver Celtic Pronouns and Particles 1997 = P. Schrijver Studies in theHistory of Celtic Pronouns and Particles. Maynooth 1997. Review: J.T. Katz in Kratylos 46 2001 p. 1-23.Schrijver Laryngeals in Latin 1991 = P. Schrijver The Reflexes of theProto-Indo-European Laryngeals in Latin. Amsterdam / Atlanta 1991.Review: H. Rix in Kratylos 41 1996 p. 153-163.Schulze Kleine Schrijien + Nachtr. 2' d ed. 1966 = W. Schulze KleineSchrijien, published by the ldg. Seminar der Universittit Berlin. 2nd ed.with supplement (Nachtrtige), edited by W. Wissmann. Gottingen 1966.Schulze-Thulin o-stufige Kausativa IIterativa und Nasalprtisentien 2001= B. Schulze-Thulin Studien zu den urindogermanischen o-stufigenKausativa I lterativa und Nasalprtisentien im Kymrischen. l.nnsbruck2001 (= IBS 99).Schwerdt 2. LV 2000 = J. Schwerdt Die 2. Lautverschiebung. Heidelberg2000.Schwyzer Kleine Schrijien 1983 = E. Schwyzer Kleine Schrijien, editedby R. Schmitt. Innsbruck 1983 (= IBS 45).Schwyzer Gr. Gr. J 1939 = E. Schwyzer Griechische Grammatik, vo!. I(AlIgemeiner Tei/, Lautlehre, Wortbi/dung, Flexion). Munich 1939.Schwyzer / Debrunner Gr. Gr. 11 1950 = E. Schwyzer and A. DebrunnerGriechische Grammatik, vo!. 11 (Syntax und syntaktische Stilistik),completed and edited by A. Debrunner. Munich 1950.Seebold Etymologie 1981= E. Seebold Etymologie. Eine Einfiihrung amBeispiel der deutschen Sprache. Munich 1981.Seebold Germ. starke Verben 1970 = E. Seebold Vergleichendes und etymologischesWorterbuch der germanischen starken Verben. The Hague/ Paris 1970.Seebold Halbvokale 1972 = E. Seebold Das System der indogermanischenHalbvokale. Untersuchungen zum sog. "Sieversschen Gesetz" und zuden halbvokalhaltigen SujJixen in den indogermanischen Sprachen,bes. im Vedischen. Heidelberg 1972. Review: J. Schindler in Sprache23 1977 p. 56-65.Seebold Personalpronomina 1984 = E. Seebold Das System der Personalpronominain den friihgermanischen Sprachen. Gottingen 1984 (=ZVS, Ergtinzungsheji 34).Seebold, see also Kluge / Seebold above.SEG = Supplementum Epigraphicum Graecum.Seiler Relativsatz 1960 = Hj. Seiler Relativsatz, Allribut und AppOSition.Wiesbaden 1960.Seiler Sprache und Sprachen 1977 = Hj. Seiler Sprache und Sprachen,Gesammelte AuJstitze. Munich 1977 (= Structura I I).Seiler Steigerungsformen 1950 = Hj. Seiler Die primtiren griechischenSteigerungsformen. Hamburg 1950.Senn Handb. d. lit. Sp rache 1966 = A. Senn Handbuch der litauischenSprache. 2 vols. Heidelberg: I (Grammatik) 1966; II (Lesebuch undGlossar) 1957.

362 Bibliography and Key to Reference CitationsBibliography and Key to Reference Citations363Serbat Cas el fonclions 1981 G. Serbat Cas el fo nclions, Elude desprincipales doctrines casuelles du Moyen Age a nosjours. Paris 1981.Review: R. Amacker in Kratylos 27 1982 p. 5-8.Sergent Indo-Europeens 1995 B. Sergent Les indo-europeens, Hisloire,langues, mYlhes. Paris 1995.SihJer New Comparalive Grammar 1995 A. L. SihJer New ComparaliveGrammar of Greek and Latin. New York / London 1995.SMEA Sludi Micenei ed Egeo-Analolici.SMID I 1968 II 1986 1979 1980-1981 Sludies in Mycenaean Inscriptionsand Dialecl. The work offers a comprehensive bibliography withindexes. - SMID I (1953-1964) assembled by L. Baumbach from theSMID-vols. I-X (originally published by: The Institute of ClassicalStudies of the University of London). Rome 1968. - SMID II (1965-1978), assembled by L. Baumbach from the SMID-vols. XI-XXIIf(originally published by: as part I, but the final fascicles were publishedby the British Association of Mycenaean Studies, Cambridge). Rome1986. - SMID 1979, edited by E. Sikkenga. University of Texas atAustin, Department of Classics, Program in Aegean Scripts and Prehistory.1995. - SMID 1980-1981, dito, 1997.Smyth Greek Grammar 1956 H. W. Smyth A Greek Grammar (forColleges), revised by G. M. Messing. Cambridge MA 1956 (and reprints).Solmsen Eigennamen 1922 F. Solmsen Indogermanische Eigennamenals Spiegel der Kullurgeschichte, edited by E. Fraenkel. Heidelberg1922.Solrnsen Unlersuchungen 1901 F. Solmsen Unlersuchungen zur griechischenLaul- und Verslehre. Strasbourg 190 I.Solta Balkanlinguistik 1980 G. R. Solta Einfuhrung in die Balkanlinguislikmil besonderer Berficksichligung des Subslrals und des Balkanlaleinischen.Darrnstadt 1980.Solta Slellung der lal. Sprache 1974 G. R. Solta Zur Slellllng derlaleinischen Sprache. Vienna 1974 ( SbOAW 291. vol., 4. Abhandlung).Solta Slellllng des Arm. 1960 G. R. Solta Die Stellung des Armenischenim Kreise der indogermanischen Sprachen. Eine Untersuchung der indogermanischenBestandteile des armenischen Wortschatzes. Vienna1960.Sommer Handbuch 1948 F. Sommer Handbuch der laleinischen LautundFormenlehre. Eine Eirifuhrung in das sprachwissenschajilicheStudium des Lateins. Heide1berg 1948. See also Sommer / Pfister below.Somrner Heth. 1947 F. Somrner Hethiter und Hethitisch. Stuttgart1947.Somrner NachlafJ 1977 F. Sommer Schrijien aus dem NachlafJ. Munich1977 ( MSS Beiheft 1, Neue Folge).Somrner Nominalkomposita 1948 F. Somrner Zur Geschichte der griechischenNominalkomposita. Munich 1948.Sommer Vergl. Syntax 1931 F. Sommer Vergleichende Syntax derSchulsprachen. 3rd ed. Stuttgart 1931 ( 4th, unchanged reprint Darmstadt1959).Sommer / Pfister Lautlehre 1977 F. Somrner Handbuch der lateinischenLaut- und Formenlehre. Heide1berg 1948. 4th revised ed.: vol. I (Einleitungund Laut/ehre) by R. Pfister. Heidelberg 1977. The plannedsecond volume was not published. Sonderegger Althochdeutsch 1987 S. Sonderegger Althochdeutsche Sprache und Literatur. Eine Einfuhrungin das alteste Deutsch. Darstellung und Grammalik. 2'" revisedand expanded ed., Berlin / New York 1987.Sonderegger Deutsche Sprachgeschichte I 1979 S. SondereggerGrundzuge deutscher Sprachgeschichte. Diachronie des Sprachsystems.vol. I (Einfuhrung, Genealogie, Konstanten). Berlin / New York1979.Speyer Syntax 1896 J. S. Speyer Vedische u. Sanskrit-Syntax. Strasbourg1896.Sprache Die Sprache. Zeitschriji fur Sprachwissenschaji. Wiesbaden /Vienna. Contains the important Idg. Chr. , see above, s. v.Sprache Funf Vortrage 1991 Sprache, Funf Vortrage by H. Kol3ler, J.Richter, B. Forssman, M. v. Engelhardt and R. Slenczka. Erlangen1991 ( Erlanger Forschungen, Reihe A Geisteswissenschajien 54).Sprachen im rom. Reich 1980 Die Sprachen im romischen Reich derKaiserzeil, Kolloquium 1974, edited by G. Neumann and J. Untermann.Koln / Bonn 1980.Sprachgeschichte I 1998 2 1985 Sprachgeschichte, Ein Handbuch zurGeschichte der deutschen Sprache und ihrer Erforschung, edited by W.Besch, A. Betten, O. Reichmann, S. Sonderegger. Berlin / New York. 2parts: I (2'" edition, fully revised) 1998; 2 1985 ( HSK 2.1 and ?.2).Stair na Gaeilge 1994 Stair na Gaeilge in omos do Padraig 0 Fiannachta,edited by K. McCone et al. Maigh Nuad 1994. Which includesa review of current research in Old Irish phonology and morphology byK. McCone p. 61-219.

364Bibliography and Key to Reference CitationsBibliography and Key to Reference Citations 365Stang Opuscula 1970 = Ch. S. Stang Opuscula linguislica, AusgewdhlteAufsdtze und Abhandlungen. Oslo / Bergen / Troms", 1970.Slang Vgl. Gramm. 1966 Ergdnzungsband 1975 = Ch. S. Stang VergleichendeGrammatik der ballischen Sprachen. Oslo / Bergen /TroffiS0: Vgl. Gramm. 1966; Ergdnzungsband, Register, Addenda undCorrigenda zur Vgl. Gramm. 1975.Slarke Kasus und Adv. im Aheth. 1977 = F. Starke Die Funktionen derdimensionalen Kasus und Adverbien im Althethitischen. Wiesbaden1977 (= StBoT 23).Starke Keilschr. -luw. Nomen 1990 = F. Slarke Untersuchungen zurStammbildung des keilschrift-luwischen Nomens. Wiesbaden 1990 (=StBoT 31).Slarke Keilschrift-luw. Texte 1985 = F. Starke Die keilschrift-luwischenTexte in Umschrift. Wiesbaden 1985 (= StBoT 30).Starke Kikkuli 1995 = F. Starke Ausbildung und Training von Streitwagenpferden:Eine hippologisch orientierte Interpretation des Kikkuli-Textes.Wiesbaden 1995 (= StBoT 41).StBoT = Studien zu den Bogazkoy-Texten.Steinbauer Denominativa 1989 = D. H. Steinbauer Etymologische Untersuchungenzu den bei Plautus belegten Verben der lateinischen erstenKonjugation. Unter besonderer Berucksichtigung der Denominative.Altendorfnear Bamberg 1989.Stempel Diathese 1996 = R. Stempel Die Diathese im Indogermanischen,Formen und Funklionen des Mediums und ihre sprachhistorischenGrundlagen. Innsbruck 1996 (= IBS, Vorlrdge und Kleinere Schriften67).Storia d'Europa Il 1994 = Storia d'Europa, vol. II (Preistoria e antichita),edited by 1. Guilaine and S. Sett is. Turin 1994.Streitberg Got. Bibel 1971 = Die gotische Bibel, edited by W. Streitberg.2nd ed. Part I (Der gotische Text und seine grieehisehe Vorlage) Heidelberg1919; 2nd ed. Part 11 (Gotiseh-grieehiseh-deutsehes Worterbuch)Heidelberg 1928. 6th ed. Heidelberg 1971. 7th ed. (Part I with anafterword by P. Scardigli).Streitberg Urgerm. Gr. 1896 = W. Streitberg Urgermanisehe Grammatik,Einfuhrung in das vergleichende Studium der altgermanisehenDialekte. Heidelberg 1896 (reprint 1943).Strunk Lachmanns Regel 1976 = K. Strunk Lachmanns Regel fur dasLateinisehe. Giittingen 1976 (= ZVS Ergdnzungshejt Nr.26). Review:A. Morpurgo Davies in CR 29 1979 p. 259(Strunk Nasalprasentien und Aoriste 1967 = K. Strunk Nasalprdsentienund Aoriste. Heidelberg 1967.Strunk 'Vorhersagbarer ' Spraehwandel 1991 = K. Strunk Zum Postulat'vorhersagbaren' Spraehwandels bei unregelmdjJigen oder komplexenFlexionsparadigmen. Munich 1991 (= SbBAW 1991, Hejt 6).Studi di linguistica greea 1993 [1995] = Studi di linguistiea greea, Materialilinguistici, dossier from a conference in Pavia 1993, edited by P.Cuzzolin. Milan 1995. Review: R. Hodot in BSL 93 / 2 1998 p. 165-168.Stud z. idg. Wortsehatz 1987 = Studien zum indogermanischenWortsehatz, edited by W. Meid. Innsbruck 1987. Review: 1. Unlermannin Kratylos 34 1989 p. 45-54.Sturnpf Wesllochariseh 1990 = P. Stumpf Die Erscheinungsformen desWesttoeharischen. Reykjavik 1990 ( = TIES, Suppl. Ser. 2).Svennung Anredeformen 1958 = 1. Svennung Anredeformen, VergleichendeForschungen zur indirekten Anrede in der drillen Personund zum Nominativ fur den Vokativ. Uppsala / Wiesbaden 1958.Syntax I 1993 2 1995 = Syntax, Ein internationales Handbuch zeitgenossischerForschung, edited by J. Jacobs, A. v. Stechow, W. Stemefeld,Th. Vennemann. Berlin / New York. 2 parts: I 1993; 2 1995 (= HSK9.1 and 9.2).Syntaxe des langues indo-iraniennes anciennes 1993 [1997] = Syntaxedes langues indo-iraniennes anciennes, Colloque international Sitges1993, edited by E. Pirart. Barcelona 1997.Szantyr see above, Hofrnann / Szantyr and Leumann / Hofrnann / Szantyr.Szemen:nyi Einfuhrung 4. ed. 1990 = O. Szemen:nyi Einfuhrung in dievergleichende Sprachwissenschajt. 4 th revised ed. Darrnstadt 1990.The Einfiihnmg first appeared in 1970, the 2 "" ed. in 1980, 3' d ed. in1989. - Italian translation: Introduzione alia linguistica indoeuropea.edited by G. BoccaIi, V. Brugnatelli and M. Negri. Milan 1985. -English translation: Introduction to Indo-European Linguistics. Oxford1996. - Many reviews have been written. Reviews of the I " ed. 1970include: W. Meid in Kratylos 16 1971 [1973] p. 41-49; B. Forssman mAnglia 94 1976 p. 441-450. Reviews of the 3rd ed. 1989: R. Schmittin Gnomon 62 1990 p. 365-367; W. Meid in Kratylos 36 1991 p. 87-91 ; E. Eggers in IF 96 1991 p. 261 -266 (including in note I the reviewsfrom BL (q.v.) of the first and second editions up to 1986). ReVIews ofthe 4t h ed. 1990 include: 1. Hajnal in PFU I 1994-1995 p. 39-46.

366 Bibliography and Key to Reference CitationsBibliography and Key to Reference Citations367Szemerenyi Lal. Worlschalz 1989 O. Szemerenyi An den Quellen deslateinischen Worlschatzes. Innsbruck 1989. Review: M. Peters in Idg.Chr. 34 Nr. H 659.Szemerenyi Numerals 1960 O. Szemerenyi Studies in Ihe IndoEuropean System oJ Numerals. Heide1berg 1960.Szemerenyi Richtungen d. mod. Sprachw. T 1971 II 1982 O. SzemerenyiRichtungen der modernen Sprachwissenschaji. Heide1berg: I (VonSaussure bis Bloomjield, 1916-1950) 1971; II (Die JunJziger Jahre,1950-1960) 1982. Review: M. Mayrhofer in Sprache 29 1983 p. 182-186.Szemerenyi Scripta Minora I 1987 II 1987 III 1987 IV 1991 Supp!. 1992 O. Szemerenyi Scripla Minora. Selected Essays in Indo-European,Greek, and Latin, edited by P. Considine and J. T. Hooker. Parts I-mInnsbruck 1987 ( IBS 53 with 3 parts): I (Indo-European) p. 1-588; II(Latin) p. 589-1076; III (Greek) p. 1077-1643; Part IV (Indo-EuropeanLanguages other than Latin and Greek) Innsbruck 1991 ( IBS 63). Asa supplement, Word Index 1992. By the same author Summing up aLife, Autobiographie und Schrijienverzeichnis. edited by B. Brogyanyi.Freiburg 1991. Concerning Szemerenyi, see also FS 0. Szemerenyi *651979 and *75 1992 above. Also cf. HS 110 1997 p. 1-3 which includescomments on O. Szemerenyi's publications. Lastly, cf. B. BrogyanyiSchrijienverzeichnis: O. Szemerenyi (1913-1996) in PFU 2-3 1996-1997 p. 53-80.Szemerenyi Syncope 1964 O. Szemerenyi Syncope in Greek and IndoEuropean and the Nalure oJ Indo-European Accent. Naples 1964. Review:G. Cardona in Language 43 1967 p. 757-773.Tavola di Agnone 1994 [1996] La Tavola di Agnone net contestoitalico, Convegno di Studio 1994 in Agnone, edited by L. del TuttoPalma. Florence 1996 ( Lingue e Iscrizioni dell ' Italia Antica 7). Review:E. Nieto BaUester in Kratylos 44 1999 p. 98-106.TBL Tubinger Beitrtige zur Linguistik.Tense and Aspect in lE 1997 Tense and Aspect in Indo-European: Theory,Typology, Diachrony, edited by J. Hewson and V. Bubenik. Amsterdam/ Philadelphia 1997. Review: E. C. Po10me in JIES 25 1997 p.482.Temes Phonologie 1987 and 1999 E. Ternes EinJuhrung in die Phonologie.1st ed. Darmstadt 1987; 2nd revised and expanded ed. 1999.Tesniere Syntaxe structurale 1959 L. Tesniere Elements de synlaxestructurale. Paris 1959.Textdatierung 1979 S. Heinho1d-Krahmer, I. Hoffinann, A. Kammenhuber,G. Mauer Probleme der Textdatierung in der Hethitologie. Heide1berg1979 ( Texte der Hethiter 9).Thieme Heimat 1954 P. Thieme Die Heimal der indogermanischen Gemeinsprache.Wiesbaden 1954.Thieme Kleine Schrijien I 1971 II 1995 P. Thieme Kleine Schrijien.Wiesbaden: I. vo!. 1972; II. vo!. 1995.Thieme Studien 1952 P. Thieme Studien zur indogermanischen Wortkundeund Religionsgeschichle. Berlin 1952.Thomas Der tocharische Obliquus 1983 W. Thomas Der locharischeObliquus im Sinne eines Akkusativs der Richlung. Wiesbaden 1983.Thomas ErJorschung des Toch. 1985 W. Thomas Die Erforschung desTocharischen (1960-1984). Stuttgart 1985.Thomas Vergangenheilslempora 1957 W. Thomas Der Gebrauch derVergangenheitstempora im Tocharischen. Wiesbaden 1957.Thomas, see also Krause / Thomas above.Threatte Atlic I 1980 II 1996 L. Threatte The Grammar oJ Atlic Inscriptions.Berlin: r (Phonology) 1980; II (Morphology) 1996.Thumb see above, s. v. Brugrnann / Thumb.Thumb / Hauschild Handb. d. Skr. 1/1 1958 1 /2(+3) 1959 A. ThumbHandbuch des Sanskrit. Eine EinJuhrung in das sprachwissenschajilicheSludium des Altindischen. 3rd ed. by R. Hauschild. Heidelberg: I /1 (Einleilung und Lautlehre) 1958; 1 /2(+3) (Formenlehre, Compositumund Satzbau) 1959.Thumb / Kieckers Gr. Dial. I 1932 A. Thumb Handbuch der griechischenDialekte. Heide1berg. 2nd ed.: vo!. 1 1932 by E. Kieckers.Thumb / Scherer Gr. Dial. II 1959 A. Thumb Handbuch der griechischenDialekte. Heide1berg. 2nd ed.: vo!. I1 1959 by A. Scherer.Thurneysen Old Irish 1946 R. Thurneysen A Grammar oJ Old Irish.Dublin 1946.Tichy Grundwissen 2001 E. Tichy Indogermanislisches GrundwissenJur Studierende sprachwissenschajilicher Disziplinen. Bremen 2001.Tichy Nom. ag. auJ -tar- 1995 E. Tichy Die Nomina agenlis auJ -tar- imVedischen. Heide1berg 1995. Review: H. Hettrich in Kratylos 43 1998p. 84-91.Tichy Onomatop. Verbalbildungen 1983 E. Tichy OnomalopoelischeVerbalbildungen des Griechischen. Vienna 1983 ( SbOA W, vo!. 409).TIES Tocharian and Indo-European Studies. vo!. 1 1987 - vo!. 6 1993Reykjavik; vo!. 7 1997 Copenhagen. (vo!. 7 1997 included the dossierArbeitstagung 100 Jahre Tocharologie Saarbrucken 1995 [1997]).

368 Bibliography and Key to Reference CitationsThe Supplementary Series includes Stumpf Wesllocharisch 1990 as vol.2, Hilmarsson Nasal Prefixes in Tocharian 1991 as vol. 3, FachtagungTocharisch Berlin 1990 [1994] as vol. 4, Hilmarsson Tocharian Dictionary1996 as vol. 5.Tischler HEG = J. Tischler Hethitisches etymologisches Glossar. Innsbruck(= IBS Nr. 20): Part I (A-K) 1983; Part I1 with fascicles 5-6 (LM) 1990 and fascicle 7 (N) 1991; Part III with fascicle 8 (T, D / I)1991, fascicle 9 (T, D / 2) 1993 fascicle 10 (T, D / 3) 1994.Tischler Kleinasiatische Hydronymie 1977 = J. Tischler KleinasiatischeHydronymie, Semantische und morphologische Analyse der griechischenGewtissernamen. Wiesbaden 1977.TPhS = Transactions of the Philological Society. Oxford.Tract. Myc. 1985 [1987] = Tractata Mycenaea, Proceedings of the EighthInternational Colloquium on Mycenaean Studies Ohrid 1985, edited byP. H. llievski and L. Crepajac. Skopje 1987.Tucker Early Greek Verbs 1990 = E. F. Tucker The Creation of MorphologicalRegularity: Early Greek Verbs in -eo, -00, -60, -uo and -10.Gottingen 1990 (= HS, Ergiinzungshefi 35). Review: R. Schrnitt in HS103 1990 p. 301-304; M. Peters in Idg. Chr. 34 Nr. G 615.UCLA lE Conference 1999 [2000] = Proceedings of the Eleventh AnnualUCLA lE Conference 1999, edited by K. Jones-Bley, M. E. Huld, A.Della Volpe. Washington 2000.UCLA lE Studies = University of California Los A ngeles Program in lEStudies, edited by V. V. Ivanov and B. Vine. Vol. I 1999.Uhlich Komponierte Personennamen des Air. 1993 = J. Uhlich Die Morphologieder komponierten Personennamen des Altirischen. Witterschlick/ Bonn 1993.Untermann Monumenta (Linguarum Hispanicarum) = J. UntermannMonumenta Linguarum Hispanicarum. Wiesbaden. vol. 1 (Die Miinzlegenden:1 Text, 2 Tafeln) 1975; vol. 1I (Die Inschriften in iberischerSchrift aus Sudfrankreich) 1980; vol. III (Die iberischen Inschriftenaus Spanien: 1 Literaturverzeichnis, Einleitung, indices, 2 Die Inschriften)1990; vol. IV (Die tartessischen, keltiberischen und lusitanischenInschriften) 1997. Review of Monumenta IV: F. Villar andC. Jordan in Kratylos 46 200 I p. 166-181.Untermann Wb. Osk.-Umbr. 2000 = J. Untermann Worterbuch desOskisch-Umbrischen. Heidelberg 2000.Urheimat 1968 = Die Urheimat der indogermanen, edited by A. Scherer.Darmstadt 1968 (= Wege der Forschung 116).Bibliography and Key to Reference Citations 369UTB = UTBfur Wissenschaft, Uni-Taschenbiicher.Villiniinen Latin vulgaire 1981 = V. Vlliinanen Introduction au latin vulgaire.2nd ed. Paris 1967; 3rd ed. Paris 1981.Vaillant Vieux slave 1948 = A. Vaillant Manuel du vieux slave. Paris1948.Vasmer REW 1953-1958 = M. Vasmer Russisches etymologischesWorterbuch. 3 vols. Heidelberg 1953-1958.Vecerka AklL Syntax I 1989 IT 1993 III 1996 = R. Vecerka Altkirchenslavische(altbulgarische) Syntax, prepared by F. Keller and E. Weiher.Freiburg (Monumenta Linguae Slavicae Dialecti Veteris): I (Dielineare Satzorganisation) 1989 (= MLS 17); Il (Die innere Satzslruktur)1993 (= MLS 34); III (Die Satztypen) 1996 (= MLS 36).Veller Handb. d. ital. Dialekte I 1953 = E. Veller Handbuch der italischenDialekte. I (Texte mit Erkliirung, Glossen, Worterverzeichnis). Heidelberg1953. A useful supplement is P. Poccetti Nuovi Documenti ltalica complemento del Manuale di E. Vetter. Pisa 1979. H. Rix has beenpreparing a new edition of Vetter's Handbuch for some time; furthersee above, Untermann Wb. Osk.-Umbr. 2000.Villar Celtiberian Grammar 1995 = F. Villar A New Interpretation ofCeltiberian Grammar. Innsbruck 1995 (= IBS, Vortriige und KleinereSchriften 62).Vine Archaic Latin 1993 = B. Vine Studies in Archaic Inscriptions. Innsbruck1993 (= IBS Nr. 75). Review: R. Gerschner in Sprache 38 / 21996 [1999] p. 23 1-237.Vine Deverbative *-et6- 1998 = B. Vine Aeolic 0p1TETovand Deverbative*-eto- in Greek and Indo-European. Innsbruck 1998 (= IBS, Vortriigeund Kleinere Schriften 71).Volkart Brugmanns Gesetz 1994 = M. Volkart Zu Brugmanns Gesetz imAltindischen. Bern 1994 (= Universitiit Bern, institut fur Sprachwissenschafi,Arbeitspapiere 33). Review: Th. Zehnder in CFS 48 1994[1995] p. 177-184; R. Lubotsky in Kratylos 42 1997 p. 55-59.von HinUber see above, HinUber.Vottero Beotien I 1998 = G. Vottero Le dialecte beotien (7' s. - 2' s. av.J.-C). Nancy. vol. I (L 'ecologie du dialecte) 1998. At least three furthervolumes have been announced.Vries AnordEW 1962 = J. de Vries Altnordisches etymologisches Worterbuch.2nd, revised ed. Leiden 1962 (= 3rd ed. 1977).

370 Bibliography and Key to Reference CitationsWaanders Local Case Relations 1997 = F. M. J. Waanders Studies in LocalCase Relations in Mycenaean Greek. Amsterdam 1997.Wachter Altlat. Inschriften 1987 = R. Wachter Altlateinische Inschriften.Sprachliche und epigraphische Untersuchungen zu den Dokumentenbis etwa J 50 v. Chr. Bern 1987. Review: M. Peters in Idg. Chr. 32bNr. 1045; M. Lejeune in REL 65 1987 [1989) p. 285-287.Wackernagel Kleine Schriften I / II 1969 III 1979 = J. WackernagelKleine Schriften, 3 vols., edited by the Akademie der Wissenschaften zuG6t1ingen. I - II (p. 1-1426, edited by K. Latte) 1st ed. Gottingen 1953,2nd ed. 1969; III (p. 1427-1905, edited by B. Forssrnan) Gottingen1979.Wackernagel Untersuchungen 1916 = J. Wackernagel Sprachliche Untersuchungenzu Homer. Gottingen 1916.Wackernagel Vorlesungen I 1926 II 1928 = J. Wackernagel Vorlesungenuber Syntax, 2 Reihen. 2nd ed. Basel l926-1928 (1st ed. 1920-1924).Wackernagel, see also Kolloquium Wackernagel above.Wackernagel, see also Wackernagel / Debrunner below.Wackernagel / Debrunner Ai. Grammatik I + II / I 1957 II / 2 1954 III1930 = J. Wackernagel Altindische Grammatik. Gottingen: vol. I(Lautlehre), reprint of the 2nd ed. 1896 with an introduction by L. Renouand additional contributions by A. Debrunner 1957; vol. II / I(Einleitung zur Wortlehre. Nominalkamposition), reprint of the 1905text with addition contributions by A. Debrunner 1957; vol. II / 2 (DieNominalsuffixe) von A. Debrunner 1954; vol. III (Nominaljlexion,Zahlwort, Pronomen) by A. Debrunner and 1. Wackernagel 1930 (Thesections on number words and pronouns are by J. Wackernagel, as wellas §§ 83-101 and §159). R. Hauschild published the index to vols. I -1II in 1964; vol. IV (Verbum und Adverbium) remains missing.Walde VgI. Wh. 1927-1932 = A. Walde Vergleichendes W6rterbuch derindogermanischen Sprachen, edited by J. Pokomy. Berlin 1927-1932 =reprint 1973.Walde / Hofmann LEW = A. Walde Lateinisches etymologisches W6rterbuch,revised, starting with the 3 '" ed., by 1. B. Hofmann. 4th ed. Heidelberg1965 (and reprints).Warrnington Remains of Old Latin I-IV 1935-1940 = E. H. WarrningtonRemains of Old Latin, newly edited and translated. 4 vols. Cambridge /Mass.: I (Ennius and Caecilius) 1935 (reprint 1956); II (Livius Andronicus,Naevius, Pacuvius and Accius) 1936 (reprint 1957); TU(Lucilius, The Twelve Tables) 1938 (reprint 1957); IV (Archaic Inscriptions)1940 (reprints since 1953)Bibliography and Key to Reference Citations 371Wathelet Traits eoliens 1970 = P. Wathelet Les traits eoliens dans lalangue de /'epopee grecque. Rome 1970.Watkins How to kill a dragon 1995 = C. Watkins How to kill a dragon,Aspects of lE Poetics. New York / Oxford 1995. Review: G. E.Dunkel in The Classical Journal 92 1997 p. 417-422; F. Bader in BSL93 /21998 p. 116-130; B. Schlerath in Kratylos 45 2000 p. 36-46.Watkins Selected Writings 1994 = C. Watkins Selected Writings, edited byL. Oliver. 2 vols. Innsbruck 1994 (= lBS 80).Watkins Verbaljlexion 1969 = C. Watkins Geschichte der indogermanischenVerbaljlexion = vol. III I of the Idg. Gr. (see above). Heidelberg1969. Concerning Watkins, see also FS Watkins above.Wegner Hurritisch 2000 = L Wegner Einfuhrung in die hurritische Sp rache.Wiesbaden 2000.Weinrich Linguistik der Luge 1974 = H. Weinrich Linguistik der Luge,Kann Sprache die Gedanken verbergen ? 5. ed. Heidelberg 1974.Weinrich Textgrammatik 1993 = H. Weinrich Textgrammatik derdeutschen Sprache. Mannheim / Leipzig / Vienna / Ziirich 1993.Weiss Italic Nominal Morphology 1993 = M. Weiss Studies in ItalicNominal Morphology. (dissertation). Cornell University 1993.Weitenberg Heth u-Stiimme 1984 = J. J. S. Weitenberg Die hethitischenu-Stiimme. Amsterdam 1984. Review: H. C. Melchert in Kratylos 291984 [1985) p. 79-82.Werba Verba IndoArica (VIA) I 1997 = C. Werba Verba IndoArica: Dieprimiiren und sekundiiren Wurzeln der Sanskrit-Sprache. Part I (Radicesprimariae). Vienna 1997. Review: B. Schlerath in HS 111 1998 p.369-371.West Ilias I 1998 = M. L. West Homerus, J/ias, Part I: Rhapsodiae I-XII.Stuttgart / Leipzig 1998.West Theogony 1966 = M. L. West Hesiod, Theogony, edited with Prolegomenaand Commentary. Oxford 1966.West Works & Days 1978 = M. L. West Hesiod, Works & Days, editedwith Prolegomena and Commentary. Oxford 1978.Wheeler Nominalaccent 1885 = B. L Wheeler Der griechische Nominalaccenl.Strasbourg 1885.Windisch Sanskritphilologie I 1917 II 1920 = E. Windisch Geschichte derSanskritphilologie. 2 parts: I Strasbourg 1917; IT Berlin / Leipzig 1920.Winkler Germanische Casussyntax I 1896 = H. Winkler GermanischeCasussyntax, I. Der Dativ, Instrumental, 6rtliche und halbiirtlicheVerhiiltnisse. Berlin 1896.

372Bibliography and Key to Reference CitationsWissenschaft vom Altertum am Ende des 2nd Jt. n. Chr. 1995 = Die Wissenschaftvom Altertum am Ende des 2nd Jahrtausends n. Chr. , editedby E.-R. Schwinge. Stuttgart and Leipzig 1995.Worterbiicher 1 1989 2 1990 3 1991 = Worterbiicher / Dictionaries /Dictionnaires, Ein internationales Handbuch zur Lexikographie, editedby F. J. Hausrnann, O. Reichmann, H. E. Wiegand, L. Zgusta. Berlin 1New York. 3 parts: 1 1989; 2 1990; 3 1991 (= HSK 5.1, 5.2, 5.3).Bibliography and Key to Reference Citations 373Zweihundert Jahre Homer-Forschung 1991 .= Zweihundert Jahre HomerForschung. Riickblick und Ausblick, edited by J. Latacz. Stuttgart 1Leipzig 1991.Zwolanek Anrufungsfor:;enStudien zu AnrufungsJormen Imchen. Munich 1970 (= MSS, Beiheft 5, N. F.).1970 = R Zwolanek " vayav indrasca ",.Vedi;chen Avestischen und Griechis-'Yoshida (D.) Aheth. Gen. 1987 = D. Yoshida Die Syntax des althethitischensubstantivischen Genitivs. Heide1berg 1987 (= Texte derHethiter 13).Yoshida (K,) Endings in -ri 1990 = K. Y oshida The Hittite MediopassiveEndings in -ri. Berlin 1 New York 1990 (= Untersuchungen zur idg.Sprach- und Kulturwissenschaft, N.F. 5). Review: G. Pinault in BSL86 /2 1991 p. 134-141.ZCP = Zeitschrift fiir celtische Philologie.ZDL = Zeitschrift fiir Dialektologie und Linguistik.ZDMG = Zeitschrift deT Deutschen Morgenlandischen GeseUschaft.Zehnder A VP 2 1999 = Th. Zehnder, A tharvaveda-Paippaliida, Buch 2,Text, Obersetzung, Kommentar, Eine Sammlung altindischer Zauberspriichevom Beginn des 1. Jahrtausends v. Chr. ldstein 1999.Zgusta Kleinasiatische Ortsnamen 1984 = L. Zgusta KleinasiatischeOrtsnamen. Heidelberg 1984 (= BNF N.F. Beiheft 21).Zgusta Kleinasiatische Personennamen 1964 = L. Zgusta KleinasiatischePersonennamen. Prague 1964.Ziegler Ogam-lnschriften 1994 = S. Ziegler Die Sprache der altirischenOgam-Inschriften. Gottingen 1994 (= HS Ergtinzungsheft Nr. 36).Zimrner Satzstellung 1976 = SI. Zimrner Die Satzstellung des finiten Verbsim Tocharischen. The Hague 1 Paris 1976.Zimrner Ursprache 1990 = SI. Zimmer Ursprache, Urvolk und Indogermanisierung.Zur Methode der Indogermanischen Altertumskunde.Innsbruck 1990 (= 18S, Vortrtige u. kleinere Schriften Nr. 46). Review:J. Untermann in Kratylos 39 1994 p. 68-70.Zinsmeister Gr. Gr. I 1954 = H. Zinsmeister Griechische Grammatik I,Laut- und Formenlehre. Munich 1954. revised edition. Heidelberg1990.ZVS see above, HS.

VII. IndexThe fo llowing index has two goals: In addition to serving as a traditionalmdex, It drrects the reade not only from an entry to its correspondingparagrapḥs, but. also occasIonally informs the reader on subjects that arenot expliCItly dIscussed in the text and offers additional bibliographicalinfotlon. T?ere .are also entnes for individual Indo-European linguistsand .histoncal ImgUlsts, particularly those from the first generations fo llowmg 1816.In aḍdition to Melzler Lexikon Sprache 2000, general assistance may befound m the Sachindex of HS 1-100 (1997), see bibliography, s. v. HS. Seealso MlIory/Adams Encyclopeia 1997. What I could not yet accomplishm the eIghth ed,tIon IS the addItIon of a complete index of vocabulary.abbreviated name W 302 (2)abbreviations p. XXI f.ablative S 410ablaut L 408ff.absolute chronology L 107accent L 419accusative S 407acrodynamic F 315 (4), F 320acrostatic F 315 (4), F 320active S 314(1)adjectives F 30 I, F 323, S 400,W202adjectives of material W 202 (3)adposition S 413f.adverb F 324, S 413agens S 401agglutinate F 104agreement S 210 (1-5)Aktanten S 210Albanian E 425Alexandrian grammarians: ---7 Lex.Grammalicorum, 1996 p. 18f.allomorphs F 305 (3)allophones L 212, L 305, L 306,L 308amphidynamic F 315 (4), F 321amphikinetic F 315 (4), F 321Amredita-compounds W 212analogy L 106, W 200analytic F 104anaphoric F 405anaptyptic vowels L 202anaptyxis, see weak vowelsAnatolian E 409, L 334, F 207 (3)Ancient Greek E 417ff.anil-root L315 (1)aorist F 203 (2), S 309aorist stem F 203 (2)Apollonios Dyskolos, Greek grammarian(2 '" c. A.D.): --> Lex.Gramm. 1996 p. 34ff.appellatives W 30 IArmenian E 424articles F 405articulation (place) L 335 (4)articulation (type) L 335 (4)Indexaspect F 202 (2), S 304aspirate dissimilation cf.Grassmann's Lawaspirates L 335assibilation L 336 (2),(5), F 317 (7)assimilation L 215, L 218 (2),L 303, L 346athematic E 502 (4), F 101 (4)Aufrecht, Theodor (1822-1907):--> Lex. Gramm. 1996 p. 52augment E 502 (11), F 202 (5),F 213, F 406a-umlaut L 208, L 307(1)Avestan E 406bahuvrihi W 207Balkan E435 (1)Baltic E 434Balto-Slavic E 435 (3)Bartholomae, Christian (1855-1925): Bartholomae's LawL 347 (2)basic format (format of citalion)p. XIXBechtel, Friedrich (1855-1924): -->Lex. Gramm. 1996 p. 81Behagbel, Ono (1854-1936): -->Lex. Gramm. 1996 p. 84f.Benfey, Theodor (1809-1881):--> Lex. Gramm. 1996 p. 91 f.Benveniste, Emile (1902-1976):--> Szemerenyi Richtungen 111982 p. 121ff.Bloomfield, Maurice (1855-1928):--> Lex. Gramm. 1996 p. I1I f.Bohtlingk, Ono von (1815-1904): -->Portraits I 1966 p. 261-268;Lex. Gramm. 1996 p. 117f.Bopp, Franz (1791-1867): E 302;--> Portraits 1 1966 p. 200-250;375Lex. Gramm. 1996 p. 121f.;Bopp-Symposium 1992 [1994]borrowing E 507 (3)Breal, Michel (1832-1915): --> PortraitsI 1966 p. 440-453; Ch. deLamberterie in BSL 92 12 1997p. 10-13Breton E 43 1 (2)Brugman(n), Karl (/849-1919): E306; --> Portraits I 1966 p. 575-580; Lex. Gramm. 1996 p. 134f.;Brugmann's Law L 412; discoveryof the nasalis sonans L 305Buck, Carl Darling (1866- 1955): -->Portrails 11 1966 p. 266-277;Lex. Gramm. 1966 p. 139f.Caland-system W 206cardinal numbers F 50 ICarian E415case S 401 - S 414casual syncretism F 305 (3), F 324,S 404casus obliquus p. XVIII, F 314 (I),S 405casus reclus p. XVIll, F 314(1),S 405categories (system of) S 300causative F 204 (4)Celtic E 430, L 211 (8)Celto-lberian E 430 (I)Celts E 431 ; concerning politicaVideologicalinstrumentalizalion:--> P. De Bernardo Stempel,R. KOdderitsch, P. 6. Riain, H.Pilch, K. H. Schmidt Phil% gieund ihre Inslrumentalisierung inZCP 49-50 1997 p. 1055-1067;concerning "celtomania" in theJ 91h century M. Decimo in BSL93 /1 1998 p. 1-40

376IndexIndex377centum languages L 339 (2), L 341chance E 507 (2)chronologie (absolute, relative)L 107cognomina W 302 (3)Coine (Gr.) E 420collective S 415collectives F 304 (2)colloquial language S 100 (4)eommunia E 506 (5)comparative linguistics, see I ndo-European linguisticscomparatives F 325compensatory lengthening L 211(I), L 303, L 418 (3)complexive compounds W 210compound W 207ff.comprehensive S 415computer E 200f.coniugatio periphraslica F 218conjugation F 102consonant shift (Germ.) L 336 (4)consonant system L 300ff.context dependent L I 08context independent L 108Continental Celtic E 43 1continuum E 507 (4)Contraction E 502 (13)copula S 206correlatives F 408eorreplio alliea L 406correspondence sets E 506 (I) E507 (5)cover symbols p. XVlllf., L 3 14Cowgill, Warren C. (1929-1985):-> GS Cowgi1l 1987; Lex.Gramm. 1996 p. 208Cri mean Gothic E 432 (I)cuneiform E 409, L 207,Curtius, Georg (I 820-1885):-> Porlrails I 1966 p. 311-373Cyrillic alphabet E 433dative S 409declension F 102, F 300ff.declension classes F 307definition of field E 102ff.delabialization L 206 (2)Delbruck, Berthold (1842-1922):-> POr/rails I 1966 p. 489-496;Lex. Gramm. 1996 p. 229f.;Kolloquium Delbruck Madrid1994 [1997]deIocutive: verba de/ocUliva: ----+ G.Darms in MH 37 1980 p. 20 I ff.demonstrative pronouns F 405derivation W 202ff.derivative compounds W 210desiderative F 204 (5), S 313determinative compounds W 21 Idevl-inflection W 204 (I)Devoto, Giacomo (1897-1 974):-> Lex. Gramm. 1996 p. 236f.dialects (Gr.) E 420digamma L216diminutive forms W 204 (2)Dionysios Thrax, Greek grammarian(2. lh. v.): -> Lex. Gramm. 1996p. 245-248diphthongs L 219ff.directive S 407 (3)dissimilation L 346dual F 304 (I), S 301, S 415Dumezil, Georges (1898-1986):-> B. Schlerath in Kralylos 401995 p. 1-48 ; 41 1996 p. 1-67early and prehistory E 512e-grade L 409Eichner's Law L 33 1 (I)ellipsis of copulae S 206enclitics F 214, F 400 (2), F 40 I(I), F 403, S 209ergative S 413, S 416 (2)essive F 204 (8)Etruscan E 426 (3A)etymology W 100, W 102external reconstruction E 50 I (2)Faliscan E 428family names W 302female names W 302 (4)feminine F 303Fick, August (1833-1916):-> Porlrails I 1966 p. 435-468fientive F 204 (7)flgura elym% giea S 407final position L 403Finno-Hungarian E 436, W 103first ('Christian') name W 302 (3)first name W 302foreign word W 100Fortunatov, Filipp (1848-1914):Lex. Gramm. 1996 p. 303fFraenkel, Ernst (1881-1957):-> Lex. Gramm. 1996 p. 307fricative, cf. thornFriedrich, lohannes (1893-1972): ->Lex. Gramm. 1996 p. 31 Of.fusion W 21 1future F 202 (6), S 306Gallic E431 (lb)giithiis E 406geminate reduction L 312general linguistics E 307genitive S 411genus F 303, F 323, S 416genus verbi cf. voicegeographical names W 305Georgiev, Vladirnir (1908-1 986):-> Lex. Gramm. 1996 p. 335[Germanic E 432glottal theory L 335 (3)glotto-chronology: -> S.-G. Andersenin IF 89 1984 p. 39-52Gothic E 432government S 21 0 (6)graeco-aryan model F 200grammatical categories S 300grammatical change L 42 1 (2)grammatical terms p. XVllf.Grassmann, Hermann (1809-1877):Grassmann's Law L 348Greek E 417ff.Grimm, lacob (1785-1863):-> Porlrails 1 1966 p. 120-179;Lex. Gramm. 1996 p. 370-372; E302; Grimm's Law L 336 (4)gUQa L 413 (1)guttural, cf. tectalHahn, E. Adelaide (1893-1967):-> Lex. Gramm. 1996 p. 383Havers, Wilhelm (1879- 1961):-> Lex. Gramm. 1996 p. 398f.Hermann, Eduard (1869-1950):-> Lex. Gramm. 1996 p. 410heteroclitics F 314 (6)Hirt, Hermann (1865-1936):-> Lex. Gramm. 1996 p. 419f.historical-comparative linguistics E100Hittite E 410Hoffmann, Karl (1915-1996):-> Hoffmann Gedenkfeier 1996[1997]; M. Witzel in llJ 40 1997p.245-253Hofrnann, lohann Baptist (1886-1954): -> Lex. Gramm. 1996 p.423holodynamic F 315 (4), F 321holokinetic F 315 (4), F 321Homer E 419

'Y378lndexIndex379homorganic glide L 218Hilbschmann, Heinrich (1848-1908): E 424; -7 HilbschmannKI. Schr. 1976Humboldt, Wilhelm von (1767-1835): -7 Portraits I 1966 p. 71-120Hurrian E 404 (I)hydronymie E 513 (3), W 305 (4)hypocorisms W 204 (2), W 302 (2)hypotaxis S 204, S 208hysterodynamic F 315 (4), F 318hysterokinetic F 315 (4), F 318Iberian inscriptions: The language isnot [ndo-European: -7 DNP 5col. 623f.l11yrian E 422imperative F 212, S 312imperfect F 202 (4), S 308imperfektive F 202 (2)indefinite pronouns F 403Indian E 404indicative S 310individual names W 302[ndo-Aryan E 404Indo-European E 30 IIndo-European linguistics E 100Indo-Germanic E 301Indogerm. Gese/lschaft E 20[ (2)[ndo-Hinite E 435 (5)Indo-lranian E 404infinitive F 216, S 202infmitive verb form F 215infixes F 202 (1 k)inflection F 100inhibitive S 311initial position S 209injunctive F 202 (4), F 203 (I),F 213, S 311instrumental S 408intensive F 204 (6)intermediary source language E 435internal reconstruction E 50 I (I)internet E 100, E 20 Iinterrogative pronouns F 403interrogative sentence S 203 (2)Iranian E 405Irish E 43 1 (2)Italic E 426 (I)Italo-Celtic E 435 (2)Jacobsohn, Hermann (1879-1933):-7 Lex. Gramm. 1996 p. 468fJoki, Aulis Johannes (1913-[989):-7 Lex. Gramm. 1996 p. 484Jones, Sir William (1746-1 794):-7 Portraits I 1966 p. I-57; Lex.Gramm. 1996 p. 489fKent, Roland G. (1877-1952):-7 Lex. Gramm. [996 p. 508Klaproth, Julius E 301Krahe, Hans (1898-[965): -7 Lex.Gramm., 1996 p. 528f; RGAs. v.Krause, Wolfgang (1895-1970):-7 Lex. Gramm., 1996 p.532fKretschmer, Paul (1866-1956):-7 Lex. Gramm., 1996 p. 533;GS Kretschmer 1956Kuhn, Adalbert (1812-1881):-7 Lex. Gramm., 1996 p. 536f.;E 305Kurylowicz, Jerzy (1895-1978):-7 Lex. Gramm., [996 p. 540-542; Szemerenyi Richtungen 111982 p. 127-137k ... etuores rule: K. Stiiber inMSS 60 2000 p. 153labia Is L 337fflabiovelars L 343Lachmann's Law: -7 Strunk LachmannsRegel 1976; Meiser La.tundFormenlehre 1998 p. 79f.language external E 503language family E 402ff.language internal E 502laryngeal theory p. X, L 320,L417laryngeals L 314ffLatin E 427Latvian E 434Law, phonetic; law(s) and rule(s):-7 Collinge Laws 1985; cf thisindex s. v. Bartholomae, Brugmann,Ca land, Eichner, Grassmann,Grimm, kWelyores, LexRix, liltera, Lindeman, Meillet,ruld. Saussure, Sievers, Stang,Verner and Wackernagellengthened grade (V rddhi) L 409lenition (Hin.) L 336 (5)Lepontian E 431 (Ic)Leskien, August (1840-1916): -7Portraits I 1966 p. 469-473; Lex.Gramm., 1996 p. 565-567Leumann, Manu (1889-1977): -7Lex. Gramm., 1996 p. 567f; B.Forssman in Gnomon 49 1977 p.830-832; E. Risch in Kratylos 231978 [1979] p.213-222; cf. alsoMH 47 [990 p. 3-27 three contributionsin memory of Leumannlexemes E 502 (2), F 100 (I)lexicon W 100Lindeman's Law L 218 (2), L 405Linear A: Cretan writing system(2000 B.C.) for a non lE language:-7 GORlLA 1-V 1976-1985Linear B E 418, L 30 I (2)linguistic change E 509 (4)liquids L 300ff.Lithuanian E 434littera rule L 332 (4c)livestock E 512 (3)loan words W 100, W 103local particle S 413locative S 412locatum S 403 (2), S 410, S 413Luvian E 412Lycian E 413Lydian E 414Macedonian E 421Macedonian E 433Martinet, Andre: -7 SzemerenyiRichtungen 11 1982 p. 153ff.masculine F 303media (aspirata) L 310, L 336mediopassive F 210Meillet, Antoine (1866-1936):-7 Portraits 11 1966 p. 201-249(Writings in his honor contributedby J. Vendryes and A.Sommerfelt); Lex. Gramm., 1996p.622-624; Meillet's Law F 502(2)memorative F 213Meringer, Rudolf (1859-1931):-7 Lex. Gramm., 1996 p. 628fmesostatic F 322Messapians E 422metathesis L 304 (3), L 327, F 307(2), F 317 (7), F 318 (6c), F 325(Id)middle voice F 212, S 314 (2)migrations E 513 (2)Miklosich, Franz Xaver Riner von(1813-1891): -7 Lex. Gramm.,1996 p. 639f

380IndexIndex381miscellaneous languages (lE): .....Convegno Vdine 1981 [1983]M itanni-Ind ic E 404 (I)mobilia F 303 (2), W 204 (I)mode S 303ff.monophthongs (Lat.) E 503 (9), L220 (I)morpheme E 502 (2), F 100 (I)morphology F 100 (2)morphonology L 418murmel vowel L 202, L 318muta cum liquida L 406mutation (Germ.) L 208, L 307 (I)Mycenaean E 418names of gods W 303names W 300fT.Narten present F 203 (I b)nasalis sonans L 305nasals L 300fT.National Socialism: Abuse of linguisticsin support of thencurrentviews. As a tTighteningexample, see W. Wust " lndogermanischesBekennlnis" in DieWeltliteratur, Folge 7, 17 1942p. 134-142.negation: ..... Meier-BrUgger Gr.Sprachw. I 1992 p. 108ft:neo-grammarians E 306neuter p. XVIII, F 303, F 313noem: The smallest unity of thatwhich one wishes to express: .....Hoffmann Aujslitze 11 1976 p.524fT.nomen action is W 205 (2)nomen agentis W 205 (I)nomen gentile W 302 (3)nomen loci W 205 (3)nominal compounds W 207nominal root F 101 (3)nominal sentences S 206, W 207nominal stems F 100 (I)nominal suffixes W 202ff.nominal types F 304nominative S 405Northwest Ind-Germ. E 435 (4)Nostratic E 437notations p. XVIlft:number F 304, S 302, S 415numbers F 500numerals F 500numerical adverbs F 504Nuristani: ..... Buddruss in MSS 361977 p. 19ff.occlusive L 335ft:Ogam inscriptions E 431 (2)Old Bulgarian, cf Old ChurchSlavicOld Church Slavic E 433Old Indian E 404 (2)Old Iranian E 405Old Irish E 431 (2)Old Persian E 407Old Phrygian E 423Old Prussian E 434Old Saxon E 432 (3)onomatopoeia: ... Tichy Onomatop.Verbalbildungen 1983optative F 207 (2), S 313orality E 400ordinal numbers F 503orthoepic diaskeuasis: (RV authoritativetext): ..... Hoffmann Aufs. 111976 p. 546orthotone F 400 (2), F 40 I (I), F403Oscian E 429Ostholf, Hermann (1847-1909):..... Portraits I 1966 p. 555-562;Lex. Gramm., 1996 p. 683; Os-thoWs Law: . .. Collinge Laws1985 s.v.PaQ,ini, ancient Indian grammarian(4. Jh. v. Chr.): ..... G. Pinault inLex. Gramm., 1996 p. 692-697;E 404 (4)Palaic E 411palatal Law (11) L 206 (2f.)palatalization (11) L 206 (2f.), L339 (3), L 341 (2), L 412palatals L 339pali E 404 (3)paradigm F 102 (I)parataxis S 203f., S 207participial constructions S 202participle F 217passive S 314 (2)past tense F 202 (4)patiens S 40 Ipatronymics W 302 (3)Paul, Hermann (1846-1921):..... Lex. Gr. 1996 p. 706-708pausal form L 404Pedersen, Holger (I 867-1953):... Portraits 11 1966 p. 283-287;Lex. Gramm., 1996 p.710; KolloquiumPedersen Copenhagen1993 [1994]; F 314 (3) (4)perfect F 203 (3), F 211, S 309perfect stem F 203 (3)periphrastic F 218Persian E 407person S 301personal pronouns F 40 IPersson, Per (1857-1929): ... Lex.Gramm., 1996 p. 715pertinentive S 411phoneme L 104phonetic L 102phonetics E 306phonetics L 104phonologic L 102phonology L 104Phrygian E 423Pisani, Villore (1899-I 99 I):..... Lex. Gramm., 1996 p. 731Pisidian E416place names W 204 (4)plural S 415plurale tantum F 102 (2)poetic language E 512 (4e),S lOO (4)Pokorny, Julius (1887-1970):..... Lex. Gramm., 1996 p. 739fPorzig, Waiter (1895-1961):..... Lex. Gramm., 1996 p. 746f.positive F 325possessive adjectives W 204 (I)possessive compounds W 208possessive pronouns F 402POll, August Friedrich (1802-1887):E 305; ... Portraits I 1966 p.251 ff. ; Lex. Gr. 1996 p. 749fPrakrit E 404 (3)prepositional government compoundW210pre-Prot-[nd-European (pre-PIE)L 303, L 3 I 2, L 323 (2), L 4 I 7(3), F 104, F 206 (I)(2), F 309,F 310 (3), F 311 (4), F 501present S 309present stem F 203 (I), S 308present tense F 202 (4)primary ending E 502 (8) (I I),F 202 (4), F 209primary stems F 202 (2)pronouns F 3 I I (3), F 400proper noun W 30 Iproper nouns W 302prospective S 3 I 3 (5)

382IndexIndex383proterodynamic F 3 15 (4), F 3 I 7proterokinetic F 315 (4), F 317Proto-Germanic culture E 512 (3)Proto-Indo-European E 410psilosis (Gr.) L 309 (2)Punic E 426 (3B)qualitative ablaut L 409quantitative ablaut L 409quantitative metathesis F 317 (7), F318 (6c)-r-/-n- heteroclitics F 314 (6)Rask, Rasmus Kristian (1787-1832): -. Portraits I 1966 p.I 79ff.; Lex. Gr. 1996 p. 774ff.reconstruction methods E 507reduction grade L 203referent S 403 (2), S 410, S 413reflexive pronouns F 40 I (2)relative chronology L 107relative clause S 205relative pronouns F 404Renou, Louis (1896-1966):-4 G. Pinault in Lex. Gramm.,1996 p. 785f.rhotacism E 503 (4), L 309 (I), F325 (lb)Rigveda (R V) E 404 (2)Rix's Law L 333Romance languages E 427 (6)root F 101 (3)root aorist F 203 (Typ 2a)root compound W 209root noun -. Schindler Wurzelnomen1972; L 211 (6), W 206(5)root present E 502 (4), F 203 (la),F 206root structure L 321ruki rule E 502 (6), L 309 (3)runes E 432Sabellian E 426 (I), E 429sa-fig" F 405sandhi L 404f.Sanskrit E 404satem languages L 339 (3), L 341Saussure, Ferdinand de (1857-1913): E 307, L 315; -. PortraitsII 1966 p. 87-1 10; Saussure'sLaw L 330Scandinavian languages E 432 (2)Schindler, lochem L 103, L 304 (3),L 313 (4), F 315 (I), W 207Schleicher, August (1821-1868): E305; -. Portraits I 1966 p. 374-395; K. Koerner in Lex. Gramm.,1996 p. 835f.Schmidt, lohannes (1843-190 I): E304f.; -. Lex. Gr. 1996 p. 837f.school grammar F 102Schulze, Wilhelm (1863-1935):-. B. Schlerath in Lex. Gramm.,1996 p. 843f.schwa primum (Indogerm.) L 202,L 318, L 324 (2)schwa secundum L 203, L 304 (3),L 313 (3), see also s.v. weakvowelsSchwebeablaut L 417 (3)Schwyzer, Eduard (1874-1943):-. Lex. Gramm., 1996 p. 845f.secondary ending E 502 (9) (I I),F 202 (4), F 209secondary stem F 202 (I), F 209secondary suffix F 202 (I)semivowels L 212ff.sentence S 102sentence accent S 209Serbian E 433Serbo-Croatian E 433sepoots E 502 (14), L 315 (I)sexus F 303 (2)sibilant (fricative) L 308Sidetic E 416Siebs, Theodor (1862-1941):-. Lex. Gramm., 1996 p. 860.Sievers, Eduard (1850-1932):-. Portraits II 1966 p. I-52;Lex. Gramm., 1996 p. 860f.;Sievers' law L218 (1)sigrnatic aorist F 203 (2)significans S 101significatum S 101Silbenstruktur L 407singular S 415singulare tantum F 102 (2)Slavic languages E 433s-mobile L 311, L 405socio-linguistic L 213 (4)Sommer, Ferdinand (1875-1962):-. Lex. Gramm., 1996 p. 873sonorant L 304, L 330sound change L 105South Picene E 429, E 503 (5)Specht, Franz (1888-1949): -. Lex.Gramm., 1996 p. 874f.spiritus asper L 309 (2)Stang, Christian Schweigaard:-. Lex. Gramm., 1996 p. 881;Stang's Law L 303Stang' s Law L 303statement S 203 (I)stative F210,F211,S314(3)Steinthal, Heymann (1823-1899):-. Lex. Gramm., 1996 p. 885f.stem forms F 205Streitberg, Wilhelm August (1864-1925): -. Lex. Gramm., 1996 p.891f.strong cases F 3 14 (I)strong verbs (Ger.) L 415Sturtevant, Edgar Howard (1875-1952): -. Portraits 1I 1966 p.365-384; Lex. Gr. 1996 p. 892f.;Sturtevant's Rule L 336 (5)stylistics S lOO (4)subjunctive (Lat.) < lE Opt. S 313subjunctive F 207 (I), S 313subordinate clauses S 204suffix conglomerates W 200suffix shift W 200suffixes W 202ff.superlatives F 325suppletion F 103, F 205, F 325 (3)syllable L 406syllable core L 406syllable limit L 406symbols (linguistic) S 101syncope (Lat.) L 204syncretism ef. casual syncretismSzemerenyi, Oswald (1913-1996):cf bibliography s. v.tasks oflE linguistics E 101tecta Is L 339ff.tenses S 306tenues L 336ff.tenuis aspirata L 310, L 329 (I)terminology p. XVllf.text syntax S 200textual linguistics L 402, S 200thematic F 101 (4)thematic vowels E 502 (10), F 101(4)thorn L 103, L 313, L 407Thracian E 422three syllable rule (Gr.) L 420Thurneysen, Eduard Rudolf(l857-1940): Lex Gramm., 1996 p. 918T1TUS E 201 (I)Tocharian E 408

384Indextransformation (through laryngeals)L 211 (8), L 323 (2)transitivity S 300 (2)tree models E 513types of actions F 202 (2), F 206,S 305typology S I 00 (3)Umbrian E 429Umlaut (Germ.) L 208, L 418Universalien E 50 I (I)Uralic languages E 436Vrddhi derivation L 331 (3), W 202(2)Vrddhi L 413 (I)v[kt-inflection W 204 (I)Vas mer, Max (1886-1962) --> Lex.Gramm., 1996 p. 959f.Vedic E 404 (2)Vendryes, Joseph (1875-1960) ->Portrails 11 1966 p. 385-393,Lex. Gramm., 1996 p. 962-964Venetian E 430VeMemaM, Th. E 437verbal accents F 214verbal adjectives W 203verbal government compounds L215 (2), L 420 (3), W 207, W209verbal inflection F 102, F 200ff.verbal root E 502 (2)verbal stem F 100 (I), F 202 (I)Verner, Karl (1846-1896) -> PortraitsI 1966 p. 538-548; Lex.Gramm., 1996 p. 965; Verner'sLaw L 421visarga p. XX, L 309 (3)vocative S 406voice S 314vowel weakening (La!.) L 108,L 204, L 217 (3), L 325, L 411vowels L 200ff.Vulgar Latin E 427 (6)Wackernagel, Jacob (1853-1938):-> Portraits 11 1966 p. 52-55; R.Schrnitt in Lex. Gramm. 1996 p.986f.; Wackernagel's Law S209wave theory: -> J. Schrnidt Verwantschajlsverhiiltnisse1872; J.Goebl in Zeilschr. f Sprachwiss.2 1983 p. 3-44; E 305weakcases F314 (1)weak vowels E 503 (5),E 504 (9), L 103, L 202f.,L 304 (3), L 324 (I)Whitney, William Dwight (1827-1894): -> Portraits I 1966 p.399-439; Lex. Gramm., 1996 p.1007-1009word L 400word accent L 403, L 419word formation W 200ff.word groups W 101word limits L 403writing E 400zero grade L 409Zeuss, Johann Kaspar (I 806-1856):-> S. Ziegler in Lex. Gramm.1996 p. 1041f.Zipf, George Kingsley (1902-I 950):->Lex. Gr. 1996 p. 1043f.Zirkumstanten S 210

Untitled - Caio - Index of

Create successful ePaper yourself

Delete template?

Save as template?