Professional Documents
Culture Documents
of Idioms
A Syntactic Study
of Idioms:
By
Anna Dąbrowska
A Syntactic Study of Idioms:
Psychological States in English and Their Constraints
Reviewed by:
Bożena Cetnarowska
(University of Silesia in Katowice, Poland)
Henryk Kardela
(Maria Curie-Skłodowska University in Lublin, Poland)
By Anna Dąbrowska
All rights for this book reserved. No part of this book may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
electronic, mechanical, photocopying, recording or otherwise, without
the prior permission of the copyright owner.
Acknowledgments ...................................................................................... ix
Introduction ................................................................................................. 1
A - Adjective
Adv - Adverb
AP - Adjective Phrase
AdvP - Adverb Phrase
ApplH - High Applicative
ApplHP - High Applicative Phrase
ApplL - Low Applicative
ApplLP - Low Applicative Phrase
COCA - The Corpus of Contemporary American English
CP - Complementiser Phrase
Det - Determiner
DOC - Double Object Construction
DP - Determiner Phrase
FI - Feature Inheritance
Gen - Genitive
ICE - Idiomatically Combining Expressions
IdP - Idiomatic Phrases
iF[val] - Valued interpretable features
LF - Logical Form
O - Object
OE - Object Experiencer
P - Preposition
PF - Phonetic Form
PP - Prepositional Phrase
Pr - Predication
PrP - Predication Phrase
SC - Small Clause
SE - Subject Experiencer
SEM - Syntax-Semantics Interface
Spec - Specifier
TP - Tense Phrase
uF[ ] - unvalued uninterpretable features
v - Light / Little Verb
vP - Light / Little Verb Phrase
V - Verb
VP - Verb Phrase
LIST OF TABLES
(0.3) The examples of idioms and the OE (class II-III) psych-verbs they
correspond to:
a. X annoys Y (class II):
get the hump
raise Y’s hackles
get a rise out of Y
4 Introduction
1.1 Introduction
The aim of Chapter One is to present the definitions of the notion of an
idiom and its characteristics. The chapter comprises five sections. Section
1.2 provides a wide range of definitions of an idiom, taken from
dictionaries and encyclopaedias, followed by the definitions put forward
by linguists. Section 1.3 discusses different dimensions of idiomaticity,
meant here as the characteristics of idioms. This part of the chapter opens
with an analysis of idiom metaphoricity, and the notion of metaphor and
figurative language. Then, various taxonomies of idioms, idiom processing
mechanisms, their metaphorical interpretation, as well as analysability and
ambiguity of idioms are studied. Besides, this section deals with idiom
non-compositionality, i.e. the fixedness of form and internal structure of
idioms. Finally, among the characteristics of idioms analysed here, there is
idiom literalness, familiarity and predictability, with the focus laid on the
role of context, well-formedness of idioms, and the level of their
formality. Section 1.4 concentrates on the models and hypotheses of idiom
representation and processing commonly referred to in the literature. In
section 1.5, the working definition of an idiom is established, on the basis
of the aforementioned dimensions of idiomaticity. Finally, section 1.6
summarises all the aspects discussed in this chapter.
all entities subsumed under this label. What is more, linguists have not
reached any solution in form of a unified approach and view related to
idioms so far, nor is it possible to offer in this book an explicit description
of what the term idiom refers to. Nevertheless, in this section of the book,
an attempt will be made to provide some clues as to how to define an
idiom. The starting point in the discussion is an encyclopaedic and a
dictionary definition, presented in section 1.2.1, followed by the concepts
of linguists and scientists regarding the definitions of an idiom, outlined in
section 1.2.2.
city,” “to sleep late in the morning” where every word has a meaning that
is contributory to the meaning of the phrase.
(Willey 1939: 221)
not be translated literally from one language into the other has been
considered an idiom” (Wadepuhl 1928: 68). Having noticed that idioms
are hardly ever translated literally, Adkins (1968) adds that often the
dictionary is of little aid to provide the meaning of a particular expression.
Idioms have meanings different from the meanings of the words which
compose them. Moreover, idioms cannot be understood from the way they
have been formed, which has been exemplified by Adkins ((1968: 149) by
means of idioms such as, make a beeline for, meaning “to take the shortest
route,” and be short-handed, meaning “to have insufficient help.” Other
idioms are composed of verbs and prepositions, such as to fill in, meaning
“to substitute for” or “to complete the blanks on a form,” or built of verbs
and adverbs such as to look forward, meaning “to anticipate.”
Besides, O’Dell and McCarthy (2010) define an idiom as a fixed
expression whose meaning is not immediately obvious from looking at the
individual words in the idiom (cf. McCarthy and O’Dell 2002). Likewise,
Fraser (1970) treats idioms as multi-word phraseological units, whose
meaning is not predictable from their constituent parts, “I shall regard an
idiom as a constituent or a series of constituents for which the semantic
interpretation is not a compositional function of the formatives of which it
is composed” (Fraser 1970: 22; cf. Makkai 1972). In short, Fraser (1970)
underlines the fact that the individual elements of an idiom cannot provide
the overall meaning of the idiom. Correspondingly, Palmer’s (1986: 36)
view of a genuine idiom implies a phrase covering more than one word,
whose meaning is unpredictable from the individual idiom constituents.
Additionally, he notices that even though idioms behave like single words
in semantic perception, grammatically they cannot be recognized as such
units since idioms normally do not undergo changes (e.g. cannot form the
past tense).
Furthermore, Kavka (2003) discusses the expressions with idiomatic
nature in general, defining them as “multiword chunks consisting of
elements, or constituents, which are bound together lexically and
syntactically” (Kavka 2003: 12). Yet, he further makes a division of these
idiomatic expressions into collocations (that are semi-compositional) and
idioms proper (genuine idioms that are characterized with non-
compositionality and invariability), and mentions that idioms are “a non-
literal alternative from possible options of a literal interpretation” (Kavka
2003: 14, 25).
Then, the picture of an idiom as “a unique and fixed combination of at
least two elements some of which do not function in the same way in any
other combination (of the kind) or occur in a highly restricted number” is
presented by ýermák (2007: 142). Moreover, he stresses that anomaly is
12 Chapter One
out, most idioms are rather infrequent, i.e. they may be restricted to certain
registers and uses of speech, or to certain accents or dialects of the English
language. Schraw et al. (1988: 424) conclude that both lexicalization and
familiarity contribute to the likelihood of idiomatic preferences, while
only lexicalization contributes significantly to the comprehension of
idiomatic meanings.
Likewise, both Everaert et al. (1995: 3-5; 2010: 81) and Fernando
(1996: 2-3) define idioms, or phraseologisms, so-called in Polish linguistic
literature, as always conventionalised multiword expressions, characterised
by semantic opacity, i.e. the fact that the meaning of the whole is not the
sum of the components (cf. SzymaĔska 2008: 116). “Idioms are
conventionalized linguistic expressions which can be decomposed into
potentially meaningful components and exhibit co-occurrence restrictions
that cannot be explained in terms of rule-governed morphosyntactic or
semantic restrictions” (Everaert 2010: 81). Moreover, for Everaert (2010),
idioms include “all formulaic expressions including sayings, proverbs,
collocations” (Everaert 2010: 77). Similarly, Kjellmer (1994) considers
idioms as a type or subset of collocation, while others consider “restricted
collocations” (e.g. cardinal error / sin / virtue / grace) to be a type of
idiom (Cowie and Mackin 1975; Cowie, Mackin, and McCaig 1983).
Similarly, for Saberian (2011a: 1231), the term “idiom” has been used
to cover a wide variety of different types of multi-word units (MWUs),
which are treated as vocabulary items consisting of a sequence of two or
more words. These words constitute a meaningful and inseparable unit.
Yet, Grant and Bauer (2004) state that the term MWU refers to both
idioms as well as open and restricted collocations, excluding phrasal
verbs. However, for Grant and Bauer (2004), open collocations are the
freest kind of MWU, while core idioms are the most restricted ones.
Similarly, Aisenstadt (1979) argues that collocations differ from idioms as
“R[estricted] C[ollocation]s are not idiomatic in meaning; they do not
form one semantic unit; their meaning is made up as the sum of the
meanings of their constituents. They have a much greater variability and
usually occur in patterns with a number of interchangeable constituents”
(Aisenstadt 1979: 1).
What is more, according to Fillmore, Kay, and O’Connor (1988), some
conditions should be met to name a phrase idiomatic or not, since
“constructions may be idiomatic in the sense that a large construction may
specify a semantics (and/or pragmatics) that is distinct from what might be
calculated from the associated semantics of the set of smaller
constructions that could be used to build the same morphosyntactic object”
(Fillmore et al. 1988: 501). Furthermore, Fillmore et al. (1988: 506-510)
14 Chapter One
features of idioms are more noteworthy than others, several elements are
required for an expression to be categorised as an idiom (cf. various
taxonomies of idioms, outlined in section 1.3.1.2). Still, there are
expressions that are more prototypical idioms than others, and sometimes
it is a mission hardly possible to distinguish idioms from other types of
fixed and / or metaphorical expressions, which can be easily noticed in the
discussion below. The first property of idioms to be embarked on in the
subsequent section is idiom metaphoricity and figurativeness.
Adkins (1968: 149) explains that a language which is not literal, often
employing metaphors, is called figurative language. While no attempt has
been made to classify figurative language, it should be noted that the term
“figure of speech” or “figurative language” covers such examples as
simile, metaphor, personification, and hyperbole.
ýermák (2001: 5) adds that the very notion of metaphor, dating from
Aristotle, appears to have rather unclear boundaries, and doubts whether a
metaphor can be treated as a reliable and general condition to characterise
an idiom. Lakoff and Johnson (1980), on the other hand, underline that the
metaphor is omnipresent in everyday life, in whatever thought, action or
language. They argue that “our ordinary conceptual system, in terms of
which we both think and act, is fundamentally metaphorical in nature”
(Lakoff and Johnson 1980: 4), and add further that “the English
18 Chapter One
expressions are of two sorts: simple literal expressions and idioms that fit
the metaphor and are part of the normal everyday way of talking about the
subject” (Lakoff and Johnson 1980: 46).
In addition, Mäntylä (2004) stresses that figurativeness (metaphoricity)
is one of the most commonly acknowledged features of idioms, and that
the roots of metaphoricity often derive from real situations or acts (e.g.
hang up one's boots), or an image created by the connection between the
idiom and its meaning (e.g. the fat is in the fire). She also makes a remark
that the reason why idioms have been recognized as “dead,” or arbitrary, is
that the relation between the origins of an expression and its meaning has
faded (e.g. kick the bucket, which means “to die”), or the literal context
belongs to a special field unknown to the ordinary language user (e.g. kick
something into touch, which means “to send the ball out of play”), making
the association very difficult to discern (Mäntylä 2004: 28-29).
Moreover, Horn (2003), providing a thorough analysis of idioms,
metaphors and their syntactic mobility, credits Jackendoff (1997), and
Nunberg et al. (1994), for introducing the term “a sort of metaphorical
semantic composition” (Jackendoff 1997: 168; Horn 2003: 246) to
describe mobile expressions. Jackendoff (1997) states that idioms having
this property “can be partitioned into chunks that correspond to the “sub-
idiomatic” readings of the syntactic idiom chunks” (Jackendoff 1997:
168). Fixed idiomatic phrases, in contrast, lack this property, as
exemplified by Horn (2003) by means of the well-known and much-
discussed fixed VP idioms, such as those in (1.1), and the examples of
mobile VP idioms, as in (1.2) below:
Mobile idioms, on the other hand, can occur in the passive, as in (1.4)
below:
1
The sign * [asterisk] is used in the thesis to mark the sentence / phrase as
unacceptable in terms of grammar.
2
This appears to be similar to the property of composition proposed by Nunberg et
al. (1994), who explain that if an idiom is compositional, then elements of its
interpretation can be assigned to its various components “in such a way that each
constituent will be seen to refer metaphorically to an element of its interpretation”
[after the meaning of the idiom is known] (Nunberg et al. 1994: 496, 499).
20 Chapter One
According to Makkai’s study (1972: 117), there are two idiomaticity areas in
English to which an idiom can belong: lexemic and sememic. The lexemic
idiomaticity area, (the class of the so-called “polylexonic lexemes”)
comprises expressions of more than one word, which are “subject to a
possible lack of understanding, despite familiarity with the meanings of the
components, or the erroneous decoding: they can potentially mislead the
uninformed listener, or they can disinform [sic] him” (Makkai 1972: 122).
Disinformation or misunderstanding take place when an idiom is decoded,
or understood in a semantically wrong way. Instead, the semantic
idiomaticity area (the class of the so-called “polysememic sememes3”)
contains expressions of more than one word, which have both a logical
literal meaning and a moral or a deeper meaning, e.g. proverbs. On the basis
of this theory of idiomaticity, Makkai (1972) classifies all idioms either
under the category of lexemic idioms or under the label of sememic idioms.
Lexemic idioms are shorter and function as parts of speech, whereas
sememic idioms function as sentence idioms.
Consequently, Makkai (1972) divides all lexemic idioms longer than one
word and shorter than a sentence into six types, presented in Table 1-1
below.
3
A sememe is a semantic language unit of meaning, analogous to a morpheme,
relevant in structural semiotics (cf. Stanojeviü 2009).
22 Chapter One
As shown in Table 1-1, class (1) of phrasal verb idioms includes both
phrasal and prepositional verbs, with the constituent structure of verb +
particle. A phrasal verb (e.g. put up) can carry a separate literal meaning,
apart from its one or more idiomatic meanings (put up “accommodate” /
“give the idea”) (Makkai 1972: 135-136).
Furthermore, class (2) of Tournure idioms (often verbal idioms) are
made of at least three words and have a phrase-like structure. “Tournure”
means a mode of expression, so tournure idioms are the type of
expressions that people generally identify idioms with, e.g. kick the bucket
“to die” (Makkai 1972: 153-154). Some tournure idioms have a
compulsory it, which differentiates these tournure idioms from phrasal
verb idioms, (e.g. to have it out “to discuss a problem to solve it” (Makkai
1972: 148)). In addition, tournure idioms often contain a compulsory
definite or indefinite article, and they can only show variation in inflection
(past tense, future tense, etc.) (ibid.: 148).
Moreover, class (3) of Irreversible binomial idioms have a fixed
structure since their word order cannot be reversed, e.g. spick-and-span
“very clean and tidy” but not *span-and-spick (Makkai 1972: 164; Travis
1984).
Class (4) of Phrasal compound idioms comprises nominal compounds,
that first have to be institutionalised and widely recognised in their
specific meanings, but which denote a specific, commonly known object,
using common nouns, e.g. “White House.” Makkai even claims that within
this approach all proper nouns could count as idioms, even personal names
(Makkai 1972: 168).
Class (5) of Incorporating verb idioms consists of the first element that
is either a noun or an adjective, which is attached to a verb, e.g. to baby-
sit, to sight-see. For Makkai (1972), due to their ambiguous literal
interpretation, incorporating verb idioms may be idiomatic. For instance,
when to baby-sit is interpreted literally, it may mean “to make baby or
babies sit” or “sitting with regard to, or on account of a baby or babies”
(Makkai 1972: 168).
Finally, class (6) of Pseudo-idioms represents the last type in Makkai’s
(1972) classification, encompassing all idioms in the lexemic idiomaticity
area which have a cranberry morph as a constituent, e.g. hither and yon “in
many different directions or places; here and there,” kit and caboodle “the
whole lot; everything,” tit for tat “an action of revenge” (Makkai 1972:
340).
Additionally, Makkai (1972: 172-179) divides his sememic idioms into
nine types according to their origin or function in a language, as presented
in Table 1-2 below.
Towards Defining an Idiom 23
(6) Proverbial idioms with Curiosity killed One may pay dearly for
a moral the cat. one’s curiosity.
A little more than
(7) Familiar quotations as
kin,and less than
idioms
kind. (Hamlet
I.ii.65)
(8) Idiomaticity in It wasn’t too bad. Approval.
institutionalized It wasn’t exactly Displeasure.
understatement my cup of tea.
As cold as a
(9) Idiomaticity in
witch’s tit Extremely cold.
institutionalized
He won’t lift a He is very idle.
hyperbole
finger..
that idiom figurative meaning. Gibbs and Nayak (1989) and Gibbs et al.
(1989a) exemplify this category with the idiom break the ice, when the
word break links to the idiomatic sense of changing a mood or tense
atmosphere, while the word ice relates figuratively to social tension.
Similarly, in the idiom pop the question, the noun question quite clearly
refers to a “marriage proposal,” while the verb pop to the act of expressing
it (cf. Gibbs et al. 1989b: 59; and CieĞlicka 2004: 95; among others).
The second category of idioms includes the so-called abnormally
decomposable idioms, whose individual components have some
metaphorical relation to their idiomatic referents (e.g. buck in the idiom
pass the buck, with its meaning “to attribute to another person or group
one’s own responsibility”) (Gibbs and Nayak 1989: 109). Likewise, as
argued by Gibbs et al. (1989a), “we can understand the hitting of certain
buttons in hit the panic button as a conventional metaphor for how we
react in extreme circumstances” (Gibbs et al. 1989a: 578). Consequently,
an abnormally decomposable idiom may be viewed as somewhat lexically
flexible (cf. Gibbs et al. 1989b: 65). Additionally, the difference between
normally and abnormally decomposable idioms lies in the fact that,
compared to the former category, in which the words constituting an idiom
denote directly some component of the idiomatic reference, the latter
contains such idioms which only refer to some metaphorical relation
between the individual part and the referent (cf. CieĞlicka 2004: 95).
Thirdly, semantically non-decomposable idioms are the ones whose
individual constituents do not contribute to the overall figurative meaning,
e.g. chew the fat, which means “to gossip or make a small talk” (Titone
and Connine 1999). This category resembles the traditional approach to
idiomatic expressions, which are not compositionally derived from their
constituent words that build the string. Thus, Gibbs et al. (1989a) provide
a definition of semantically non-decomposable idioms in the following
words, “the individual components of phrases such as kick the bucket or
chew the fat are not in the same semantic field as their respective
figurative referents (i.e. “to die” and “to talk without purpose”) and should
not be viewed as semantically decomposable” (Gibbs et al. 1989a: 578;
and cf. CieĞlicka 2004: 95).
Having provided the division of idioms founded on Nunberg’s (1978)
initial taxonomy, Gibbs et al. (1989b: 59) mention the syntactic
variability of idioms. On the basis of the results of their experiments, they
prove that the more decomposable an idiom is, the more syntactically
productive the idiom is expected to be. Therefore, normally decomposable
idioms (e.g. pop the question) have been found much more syntactically
productive than either abnormally decomposable (e.g. carry a torch,
26 Chapter One
sports team can undergo optional elision, e.g. the San Francisco 49ers can
occur as the 49ers, or as a modifier in the compound noun a 49ers player,
etc.
The last subclass of lexicalized phrases, within the taxonomy of Sag et
al. (2002), apart from fixed expressions, and semi-fixed expressions,
comprises syntactically-flexible expressions, which have a wider range of
syntactic variability than semi-fixed expressions. Syntactically-flexible
expressions occur in the form of (i) decomposable idioms; (ii) verb-
particle constructions; and (iii) light verbs. “Decomposable idioms” can be
syntactically flexible to some extent, but it is difficult to predict what kind
of syntactic variation a given idiom can undergo. Moreover, “verb-particle
constructions,” such as write up and look up are made up of a verb and one
or more particles. They may be either semantically idiosyncratic, as brush
up on “to improve,” or compositional as break up in the meteorite broke
up in the earth's atmosphere. In some transitive verb-particle constructions,
as call someone up, an NP argument can occur either between or following
the verb and particle(s): call Tom up or call up Tom, respectively. Besides,
adverbs can often be inserted between the verb and particle as in fight
bravely on. Finally, in the case of “light verb constructions,” as make a
mistake, or give a demo, it is hardly predictable which light verb is
connected with a given noun. Although they are highly idiosyncratic, they
have to be distinguished from idioms: “the noun is used in a normal sense,
and the verb meaning appears to be bleached, rather than idiomatic” (Sag
et al. 2002: 7).
Finally, the taxonomy of Sag et al. (2002: 7), beside lexicalized
phrases, also includes institutionalized phrases which cover
conventionalized phrases / collocations, such as salt and pepper, traffic
light and to kindle excitement. They are semantically and syntactically
compositional. Regarding the phrase traffic light, traffic and light both
retain simple senses but produce a compositional reading by being
combined into constructions. As institutionalized phrases are fully
compositional, they can show full syntactic variability.
4
Lakoff and Johnson’s (1980) research has evoked many studies referring to
conceptual metaphors, such as, e.g. ARGUMENT, TIME and LOVE, which are
said to be used by people continually, and thus stimulating more extensive
linguistic studies.
Towards Defining an Idiom 31
5
As noted by Anudo and Kodak (2017: 168), Lakoff and Johnson’s (1980)
metaphors allow us to comprehend a more clearly delineated concept, e.g. the
concept of love can be comprehended in terms of a “journey.” Moreover,
metaphors should involve two different kinds of activities, as in the “love is
journey” metaphor, love involves emotions, but a journey refers to travelling.
Additionally, metaphors structure our everyday concepts, since “love” is partially
structured in terms of a “journey.” Furthermore, metaphors enable us to understand
one domain of experience in terms of another, e.g. the target domain of love is
referred to by the source domain of a journey. Besides, in the “Love is a journey”
metaphor, different means of travel can be used to explain a love relationship, i.e.
the journey could be undertaken using a train, a car, or a ship, etc. Since these
means of transport are different kinds of vehicles; thus, “vehicles” have become a
superordinate category evoking in human minds rich mental images and rich
knowledge structures. A mapping at the superordinate level enlarges the chances
for mapping rich conceptual structures in the source domain onto the target domain
(cf. Anudo and Kodak 2017: 168; Kövecses 2000, 2005, 2010; and Gavelin 2015;
among others).
32 Chapter One
related to the next feature of idioms, viz. idiom ambiguity, resulting from
idiom metaphoricity, which is to be analysed below.
Chafe (1968) notices that many idiomatic expressions are ambiguous, with
one interpretation (the literal meaning), deriving from the meanings of the
words involved, and the other–the idiomatic meaning.
To begin the discussion concerning idiom ambiguity, it is worth
recalling transformational grammarians’ classification of ambiguity first.
They distinguish (i) lexical; (ii) surface structure; and (iii) underlying
structure ambiguities (cf. Chomsky 1965; MacKay and Bever 1967; Bever
et al. 1969: 225; and Bobrow and Bell 1973; among others). The lexical
ambiguity implicates alternative dictionary meanings of a word with no
differences at the other grammatical levels (e.g. in “The cold was
bothering John,” cold may refer either to John’s illness or the weather
condition which is worrying John). The surface (or derived) structure
ambiguity depends on how words are grouped together into phrases, i.e.
structured (e.g. “Visiting relatives can be nice” may be understood as
follows: the activity of visiting (relatives) is nice, or the relatives who are
visiting us are nice). The underlying structure ambiguity entails a change
in the essential relations between words (e.g. in “The mayor told the police
to stop drinking,” drinking was stopped either by the mayor who
announced it to the police, or the police was ordered to stop drinking) (cf.
Bobrow and Bell 1973: 343).
In addition, the underlying structural level of sentences represents the
essential logical relations which the words bear to each other. In other
words, at a deeper level the logical relational concepts, subject, predicate,
and object, express the internal relations among the words and phrases of a
sentence (cf. Bever et al. 1969: 225). Some explanation and illustration of
ambiguity between alternative syntactic structures underlying a sentence is
provided by Wundt (1900: 268) and Bever et al. (1969: 225-226). The
actual order of the words in a sentence does not always correspond to the
underlying relations. For example, in (1.5) active and (1.6) passive
sentences, presented below, the underlying relations are the same although
the word orders differ:
In the two sentences, inn (1.5) and (1 1.6), the actinng person (su
ubject) is
Caesar in both cases. Buut this phrase is the topic off the statemen nt only in
the first, butt not in the seecond sentencee (cf. Wundt 1900: 268). These
T two
examples arre analysed as a having thee same underrlying phrase structure
represented in the tree diagrams in n Figure 1-11 and in Fig gure 1-2,
respectivelyy.
and the figurative notion they designate determine the comprehension and
use of the idiom. Thus, paying attention to semantic productivity of idioms
would be helpful to investigate the role that idiom constituents and their
meanings play in the idiom overall figurative interpretation. Glucksberg
(1993) describes semantic productivity as “the ability of people to create
new idiomatic meanings by changing relevant aspects of an idiom's
constituent elements” (Glucksberg 1993: 15). According to Glucksberg
(1993), semantically productive processes are motivated by communicative
intentions, and thus they enhance communicative functions, in
contradistinction to unmotivated synonym substitutions. Hence, semantic
productivity entails an interpretable relationship between original idiom
components and their substitutes, so that the speaker’s communicative
intention may be deduced (cf. CieĞlicka 2004: 100). To provide
Glucksberg’s (1991) example, replacing shatter the ice for break the ice
generates a new idiomatic sense that is based on the meaning of the
original idiom and the relation between the original constituents and their
substitutes. The difference between the meaning of break and its
substituted element shatter produces a new interpretation, defined by
Glucksberg (1991) as “to break down an uncomfortable and stiff social
situation flamboyantly in one fell swoop” (Glucksberg 1991: 149). Such
examples of semantically productive idiom variants, as observed by
Glucksberg (ibid.), can be encountered regularly in the media, literature
and everyday conversation; henceforth, being comprehended by language
users without difficulty.
Finally, Gibbs (1993) strongly argues that analysability, strictly related
to metaphoricity, is crucial in comprehending and learning idioms.
Besides, Gibbs (1994) sees semantic productivity as linked to the degree
of idiom analysability; which is quite interesting, due to the assumption
that the more analysable a given idiom is, the more probable it is for it to
undertake various semantic modifications. Nevertheless, Glucksberg
(1991) does not hold this belief, emphasising that analysability is “neither
a necessary nor a sufficient condition for an idiom to be varied
productively” (Glucksberg 1991: 151). Instead, what matters for
Glucksberg, (1991) is whether the semantics of the idiomatic phrase has a
direct functional relation to the idiom meaning. Idiom semantic
productivity is closely connected with idiom lexical flexibility, which is
discussed in the following section.
40 Chapter One
6
Fraser (1970) points out that in cases where the entire sentence is idiomatic (e.g.
proverbs) and when they permit passivization, the appropriate level for them in
this hierarchy would be L5-Reconstitution, not L4-Extraction.
42 Chapter One
level, but also the transformations included in the lower levels, levels L2-
Insertion and L1-Adjunction. For example, pass the buck to “to attribute to
another person or group one’s own responsibility” is analysed as
belonging to level L5. This means that any reconstitution operation is
permissible, including the action nominalization, but also any other
operations lower in the hierarchy are also tolerable for this idiom. But, the
idiom blow off some steam “to give vent to one’s repressed emotions”
belongs to level L1, and the only modification it allows is an adjunction
operation (the gerund nominalisation), with no other alterations possible.
Finally, a phrase keep watch over “to observe with continuous attention” is
marked as belonging to level L4, which predicts that extraction (the
passive and prepositional phrase pre-posing), insertion (adverbial
placement) and adjunction (gerundive nominalization) can be applied to it,
but that reconstitution (the action nominalization) cannot (Fraser 1970: 39-
40).
In short, Fraser’s (1970) hierarchy ranges from completely frozen
idioms to free collocations. Even though Fraser’s (1970) hierarchy has
never received empirical support, undoubtedly, it is a light in the tunnel to
understand the lexical flexibility of idioms. It is still valid today, being
often cited and referred to.
Then, a year before Fraser (1970), Weinreich (1969) attempted to set
out mathematical formulae to express the structures of idioms. Indeed,
Weinreich’s (1969) and Fraser’s (1970) work is respected and well-known
in the area of idiom studies. There were also other noteworthy studies on
idiom structure during the 1970’s, such as for instance, those of Makkai
(1972), and Newmeyer (1974). Makkai (1972) examines the question of
frozenness and restricted grammatical transformations by his
morphological bans. “The compulsory plural and compulsory singular
ban” means, in practice, that in idiomatic expressions the plural and
singular form cannot be changed freely (Makkai 1972: 122-123). Makkai
(1972) exemplifies the rule as follows: hammer and tongs “to argue, fight”
ĺ *hammer and tong; the skin of my teeth “miraculously” ĺ *the skin of
my tooth. Here, Makkai (1972) states that the second form of the two
idioms is not grammatical.
Interestingly, in Longman Dictionary of English Idioms (1979: xiv-
xix), it is noted that even though many idioms are so frozen that any
additional words cannot be put within the phrase, there are still some
exceptions, i.e. adjectives and adverbs are often allowed to be inserted
within many idioms, e.g. to go (all) to pieces “to break down.” Secondly,
but perhaps not preferably, impolite and swear words may be inserted in
frozen idioms, playing the role of intensifiers, e.g. he went the whole
Towards Defining an Idiom 43
(e.g. play with fire “to do something dangerous or risky”). On the other
hand, opaque idioms are those whose meaning cannot be derived from the
analysis of their constituents (e.g. kick the bucket “to die”). Idioms also
vary as regards their well-formedness, that is their syntactic structure can
be well-formed, as in paddle your own canoe “to handle one’s own
affairs,” or ill-formed, as in go bananas “to become wildly irrational.”
Lastly, idioms can be literal (ambiguous), when they offer both the literal
and figurative interpretations, e.g. break the ice, which if literally taken,
means “to cut the frozen surface,” and if figuratively treated, it means “to
do or say something to relieve tension or get conversation going.”
However, idioms can also be non-literal (unambiguous), when only the
figurative interpretation is probable, e.g. drink somebody under the table
“to be able to drink more alcohol than someone else” (cf. CieĞlicka et al.
2008).
Yet, the impact of literalness on the comprehension of idiomatic
phrases is also shown in the studies of Titone and Connine (1994a),
supporting the compositional approach to idioms. They confirm the
significant contribution of literalness to the activation of idioms literal and
figurative interpretation in the course of idiom processing.
So far, the study of the influence of literalness on idiom processing has
yielded inconsistent results. Brannon (1975), and Popiel and McRae
(1988) find longer classification times for literal idioms, compared to
idioms with only a figurative meaning. Reverse outcomes are obtained by
other researchers in their multiple tasks (Swinney and Cutler 1979; Estill
and Kemper 1982; Schweigert 1986; Mueller and Gibbs 1987; Schweigert
and Moates 1988; and Cronk and Schweigert 1992; among others).
Tabossi et al. (2011: 113) add that in the incidental-memory experiment
conducted by Botelho da Silva and Cutler (1993), the participants
remembered idioms much better than control strings, but with no
difference in their ability to remember literal and non-literal idioms.
Titone and Connine (1994a) investigate the impact of literalness on the
processing of idiomatic expressions, and they confirm that literalness
determines the activation of the literal and figurative meaning of idioms
during their processing. In brief, literal idioms, unlike non-literal ones,
reveal less activation of the figurative meaning and more activation of
their literal interpretation of the constituent words of the idiom string.
In tum, familiarity, as the other key factor in idiom comprehension, is
identified by Titone and Connine (1994b) as the “frequency with which a
listener or reader encounters a word in its written or spoken form and the
degree to which the meaning of a word is well known or easily
understood” (Titone and Connine 1994b: 250). In addition, Cronk and
48 Chapter One
lot of attention has been paid to these characteristics, which concern the
role of context, well-formedness of idioms, and the level of their
formality. So far, in the study of idiom syntax, knowing the syntactic
behaviour of individual idiomatic expressions has been a precious tool that
can be applied to manipulate the syntactic variable experimentally
(Tabossi et al. 2011: 113). Undoubtedly, both the syntactic productivity
and the lexical creativity of idioms are matters of degree, depending on the
idiom compositional properties. This conclusion indicates that idioms do
not form a unique class of linguistic items, but share many of the
properties with more literal language. Indeed, understanding the syntactic
behaviour of idioms is a fundamental mission for any theory of idiom
representation and processing, which is to be analysed in section 1.4, and
several alternatives are still under debate (Gibbs and Nayak 1989; and
Sprenger et al. 2006).
The third option, the Direct Access Hypothesis (Gibbs 1980, 2002),
illustrated in Figure 1-5 below, assumes that an idiom’s figurative
meaning can be activated without the literal meaning being processed first.
Gibbs’s studies show that, given a suitable context, the idiomatic meaning
is processed sooner than the literal meaning (cf. Gibbs 1985; 1986;
Schweigert 1986; Schweigert and Moates 1988; and Needham 1992;
among others). It is possible that people completely bypass such mode
when faced with a highly familiar idiom, or when they have a sufficient
context to infer an idiomatic interpretation. In these cases, the idiomatic
meaning is directly accessed.
Towards Defining an Idiom 57
Figure 1-5. The Direct Access Hypothesis according to Gibbs (1980, 2002)
(own source)
7
Cf. Footnote 5.
Towards Defining an Idiom 59
Gibbs and his colleagues (Gibbs and Nayak 1989; and Gibbs et al. 1989a)
introduce the Idiom Decomposition Hypothesis, within the compositional
framework. Although often cited as such, the Idiom Decomposition
Hypothesis is not an assumption about the processing of idioms;
decomposability is an influencing variable with regard to comprehension
or representation of idioms (Gibbs et al. 1989a).
To begin with, the Idiom Decomposition Hypothesis is about the
analysability of idioms, viz. the “speaker’s assumptions about how the
meaning of the parts contribute to the figurative meanings of the whole”
(Gibbs and Nayak 1989: 104). A decomposable idiom is an idiom whose
individual components contribute to its figurative meaning (e.g. play with
fire “to do something dangerous or risky”), while idioms whose individual
elements do not make such a contribution are non-decomposable (e.g. kick
the bucket “to die”). But true decomposability is a feature of idioms that is
relevant from a psycholinguistic point of view and it is based on speakers’
judgements; whereas compositionality is a theoretical assumption about
the combination of syntactic constituents and their phrasal or sentential
meanings, which is important within linguistic theories such as generative
grammar.8 From the generative, syntactic point of view, only the literal
8
The compositionality principle, as a basic assumption of generative grammar,
goes back to Frege (1884 / 1980), and holds that the “meaning of an expression is a
function of the meanings of its parts and of the way they are syntactically
60 Chapter One
combined” (Partee 1984: 153; and cf. also Dowty 2007: 25). It is responsible for
the great problems that idioms pose within this framework and is the reason why
generative grammarians have proposed various, complex assumptions for the
description of idiomatic expressions. These assumptions were intended to make
idioms fit into the overall compositional pattern (e.g. Weinreich 1969; Fraser 1970;
Katz 1973; and Everaert 1993; among others).
Towards Defining an Idiom 61
Even though over the last five decades, linguists and psycholinguists have
developed a number of hypotheses to describe the distinctive grammatical
characteristics of idioms, and to provide some explanation for their
processing and representation; there are three aspects that have been
neglected. According to Abel (2003), these are: first, the studies deal
exclusively with the native mental lexicon, and do not try to integrate the
second language (L2) lexicon. Second, the studies concentrate either on
lexical representations or on conceptual aspects, but do not try to combine the
two into one theoretical model. Lastly, most of the studies do not allow for
frequency effects to play a role in the representation or processing of idioms.
The Model of Dual Idiom Representation (from now on the DIR
Model), offered by Titone and Connine (1999), is a psycholinguistic
model which includes the above-mentioned aspects. Not only does it
combine the lexical and the conceptual level but it also integrates the
representation of idioms in the first language (L1) and the L2 lexicon.
Abel (2003) presents some supporting evidence for the DIR Model based
on empirical studies on the decomposability of idioms with native and
non-native speakers of English. The insights and results of experimental
morphological studies are also used to confirm the assumptions of this
model. Thus, the Model of Dual Idiom Representation (DIR Model)
assumes that non-decomposable idioms require an idiom entry (at the
conceptual level), whereas decomposable idioms can be represented via
constituent entries, and can additionally develop an idiom entry (both at
the lexical level), as illustrated in Figure 1-8 below.
Figure 1-8. Titone and Connine’s (1999) Model of Dual Idiom Representation
(own source)
Towards Defining an Idiom 63
1.4.3.3 The Hybrid Model of Idiom Production (Cutting and Bock 1997)
Cutting and Bock (1997) represent the first attempt to address questions
concerning the storage and retrieval of idiomatic phrases during language
production. They ran a series of error-elicitation studies in which
participants were briefly presented with two idioms and, after a short
pause, were asked to produce one of them as quickly as possible. The
dependent measures are production latencies and blending errors, that is,
unconventional combinations of two idioms. In the first experiment,
Cutting and Bock (1997) employed pairs of idioms with similar (kick the
bucket, or meet the Maker “to die”) or different (shoot the breeze “to
converse aimlessly; to chat,” or raise the roof “to show great enthusiasm”)
idiomatic meanings, and with the same syntactic form (chew the fat “to
gossip or make a small talk,” or raise the roof “to show great enthusiasm”)
or different (chew the fat “to gossip or make a small talk,” or nip and tuck
“inconclusive as to outcome; close or just even in a race or comparison”)
syntactic forms. What follows is an assumption that if idioms are stored as
unitary forms, then their syntactic structure should have no effect on the
resulting idiom blends. In contrast, if idiomatic expressions do submit to
syntactic analysis in the course of their production, then idioms with a
similar structure should produce more blends than those with differing
structures. Cutting and Bock’s (1997) experiment reveals that same-syntax
idioms, with similar figurative meanings, are more likely to blend than
different-syntax idioms, with different figurative meanings. This evidence
64 Chapter One
9
The very term lemma refers to a representation of a lexical item grammatical
class information plus a pointer to the word forms (Roelofs 1992; and Levelt and
Meyer 2000; among others).
Towards Defining an Idiom 65
structure idioms, such as meet your Maker and kick the bucket “to die,” are
activated simultaneously and will compete in the course of language
production, which might lead to blend errors such as ‘meet the bucket’.
Moreover, with the purpose of testing the hybrid account of idiom
production, Sprenger, Levelt and Kempen (2006) examined the production
of idioms in a series of studies employing reaction time paradigms. In their
first experiment, Sprenger et al. (2006) tested the claim that idiom
constituents are the same lemmas which get activated in the course of
producing literal phrases and that idiom phrases have their unitary
representations in the mental lexicon. They, thus, employed a cued-recall
task, in which native speakers of Dutch produced either idiomatic or literal
phrases they had learned earlier, as an answer to a visually displayed
prompt word. While looking at the prompt word, participants heard a
prime, which was either a word identical to the noun of the phrase to be
produced, or a word unrelated to any of the phrase’s elements. The results
showed that both idiomatic and literal phrases are produced faster when
clued-up by one of their content words. This result, according to Sprenger
et al. (2006), supports the view that idioms are compositional phrases,
made up of the same simple units which are activated in the course of
literal language production. In addition, the comparison of the priming
effects found for idiomatic and literal phrases reveals that idioms are
primed significantly stronger than literal utterances. Sprenger et al. (2006)
explain this difference in the strength of priming by postulating a separate
meaning representation for idioms in the mental lexicon. Lemmas which
belong to an idiomatic phrase are, in this way, incorporated into a unitary
lexical entry. Therefore, priming one of the lemmas, belonging to this
common idiomatic representation, results in the spreading of activation to
the remaining lemmas of the idiom entry, and makes them more available
for retrieval.
These conclusions, confirmed in the remaining experiments (cf.
CieĞlicka 2010), explicitly support the Hybrid Account of Idiom
Representation, in which idioms are both unitary and compositional
phrases. Besides, in a post hoc analysis of the results obtained in all the
three experiments, Sprenger et al. (2006) find that idiom decomposability
does not influence the extent to which literal meanings of idiom elements
become activated in the course of idiom production. Taking everything
into account, the study conducted by Sprenger et al. (2006) confirms the
validity of the Hybrid Account of Idiom Representation, proposed by
Cutting and Bock (1997). In an effort to make the hybrid model applicable
to production and comprehension, Sprenger et al. (2006) propose an
extension and modification of the original hybrid model, which they call
66 Chapter One
10
A superlemma is a separate all-inclusive representation of an idiom, introduced
into the hybrid model at the lexical-syntactic processing level in order to account
for the evidence that the syntactic properties of an idiom are in some way
independently represented. This superlemma is linked with the individual lemmas
which make up the idiomatic phrase.
Towards Defining an Idiom 67
specific categories each. Proverbs with their purpose to draw a moral and
teach a lesson are closely related to cultural aspects; sayings are generally
not figurative; and conversational phrases, at last, with their special
interactional function are more literal than figurative. In a nutshell, Figure
1-9 illustrates the overlap between the boundaries of various multi-word
expressions, making them nearly inseparable from one another.
Fourthly, the terms that are expected to occur in the book, i.e.,
“idioms,” “phraseological units,” “idiomatic expressions,” “multi-word
expressions” or “fixed expressions” seem to be more widely known in
English. They have been chosen to make the subject matter more
straightforwardly accessible to scholars not directly working in the field of
phraseology. Gläser’s (1998) definition of “phraseological units,”
reformulated in (1.8) below, appears to be the most suitable one, as it
comprises virtually all the possible notions, characteristics and entities that
should be encompassed by the term.
meanings of its components; (ii) some subset of idioms has a fixed word
order, which implies they have a restricted set of variants and should be
treated as entities since the link between their form and meaning has not
been recognized; (iii) recently, with a great input of psycholinguistic
studies (cf. Fernando 1996; and Moon 1998; among others), a great
number of idioms far from being dead or frozen has been found, but
instead, they may be marked with possible alterations, metaphoricity and
perceptible origins of their meanings.
Besides, some crucial characteristics of idioms have been established.
Firstly, it is idiomatic diversity, which results in various taxonomies of
idioms, classified in multiple ways, based on idiom semantics, syntax, and
function, e.g. the taxonomy of Makkei (1972), the one of Nunberg (1978)
and of his followers (Gibbs and Nayak 1989: 104; Titone and Connine
1999), Cacciari and Glucksberg (1991), Nunberg et al. (1994), the
taxonomy of Sag et al. (2002), and the classification of Yoshikawa (2008),
to name just a few. Yet, the kind of typology chosen for the purpose of this
book is the bipolar classification offered by Nunberg et al. (1994), who
divide idioms into (i) idiomatically combining expressions (e.g. pull
strings “to use connections”), and (ii) idiomatic phrases (e.g. kick the
bucket “to die”).
Additionally, the feature that most idiomatic phrases share, even
though the boundaries are sometimes overextended, is the metaphorical or
figurative nature (Mäntylä 2004: 28-29). What is more, the ambiguity of
many idiomatic expressions has been pointed out, as some idioms have
one interpretation (the literal meaning) derived from the meanings of the
words involved and/or the other–the idiomatic meaning. Subsequently,
some attention has also been paid to the further characteristics of idioms,
i.e., idiom analysability, the issue of non-compositionality, and idiom
semantic decomposition; all of which refer to the extent to which idiom
constituent parts contribute to the idiom overall interpretation (cf. Nunberg
1978; Cacciari and Tabossi 1988; Glucksberg 1991; and Gibbs 1994;
among others). As the third characteristics of idioms, the fixedness of form
and internal structure of idioms have been examined. Jackendoff’s (1997)
approach to idioms as linguistic units with an internal linguistic structure
constrained by syntax, semantics, morphology, and phonology, has shed
light on idiom variability. Fraser’s (1970) six-point Frozenness Hierarchy
places idioms on the scale, grouping them from totally frozen forms that
permit no grammatical or lexical changes to idioms that tolerate
unrestricted variation. Undoubtedly, both the syntactic productivity and
the lexical creativity of idioms are matters of degree, depending on the
idiom compositional properties. Moreover, the literalness of idioms, their
72 Chapter One
familiarity and predictability, investigated with the focus put on the role of
context, well-formedness of idioms, and the level of their formality, have
been discussed. Furthermore, some space has been devoted to the models
of idiom representation and processing most commonly cited in the
literature; namely, the non-compositional models (e.g. Swinney and Cutler
1979; Gibbs 1980), the compositional hypotheses (e.g. Cacciari and
Tabossi 1988; Gibbs, Nayak, and Cutting 1989), and the hybrid
approaches with the Model of Dual Idiom Representation (e.g. Titone and
Connine 1999).
In brief, in Chapter One some main properties and models of idioms
representation and processing have been provided as an essential
background to understand the syntactic and semantic variability of idioms,
which is to be studied in Chapter Four of the book. Moreover, the working
definition of an idiom as a multi-word phrase and phraseological unit has
been established, to be adopted for the purposes of this book. Gläser’s
(1998) definition of “phraseological units” has been chosen as the one
comprising all the possible notions and characteristics that should be
encompassed by the term under consideration.
Since the book is to focus on idioms referring to psychological states in
English, it is important to learn first the basic syntactic, semantic, and
aspectual characteristics of psychological verbs. These characteristics and
the structure of psychological verbs, as well as an overview of syntactic
approaches to psych-verbs, are the main focus of the subsequent Chapter
Two.
CHAPTER TWO
2.1 Introduction
1
Arad (1998) makes a suggestion that psych-verbs are neither lexically nor
syntactically distinctive, since they share the same structure as their main predicate
(locative, dative etc.). However, their unique syntactic properties rely more on
their stativity rather than on their being assigned a specific structure (cf. Grafmiller
2013).
74 Chapter Two
verb, on account of the fact that the sentence involves no state of mind
either of Charles or of Nina.2
Psych verbs can be divided into two classes, depending on the syntactic
position of the Experiencer (e.g. Croft 1993; Pesetsky 1995; and Iwata
1995; among others), such as Subject Experiencer (SE) verbs and Object
Experiencer (OE) verbs, as illustrated in (2.2a-b):
(2.2)
2
According to Klein and Kutscher (2005: 2), from the semantic point of view,
psych-verbs can be classified into verbs denoting emotions (love, frighten, etc.),
perception verbs (see, taste, etc.), cognitive verbs (think, assume, muse, etc.), and
evaluating verbs (respect, appreciate, etc.). However, some of the verbs listed here
do not satisfy Landau’s (2010) definition of psych-verbs.
76 Chapter Two
3
The concept of semantic roles relates to the notion of thematic relations, i.e. the
relations which are semantic in nature. The thematic role is, thus, a semantic
Syntactic and Semantic Characteristics of Psychological Verbs 77
(cf. Belletti and Rizzi 1988; Grimshaw 1990; Pesetsky 1995; Arad 1998;
2000; Anagnostopoulou 1999; Primus 1999; Pylkkänen 2000; Verhoeven
2010; and Landau 2010; among others). This challenge that psych-verbs
pose is to be discussed in sections 2.3 and 2.5 in more detail.
In a nutshell, psych-verbs are expressed in different ways across
languages. In languages like German, Italian, or English, psych-verbs
permit word orders or binding relations that are impossible for non-psych-
verbs. A comprehensive investigation of psych-verbs in and across a
variety of world languages is provided by Landau (2010). In addition,
Bennis (2000), and Drijkoningen (2000) offer some detailed analysis of
psych-verbs in Dutch. Pesetsky (1987) analyses psych-predicates in
English; Anagnostopoulou (1999, 2008) in Greek; while Ruwet (1972,
1993), Legendre (1989, 1993), and Bouchard (1992, 1995) in French.
Moreover, Arad (1998), Reinhart (2002), and Anagnostopolou (2008)
discuss psych-verbs in Hebrew; Klein and Kutscher (2005), and Temme
(2014) in German; whereas Belletti and Rizzi (1988), Cresti (1990), and
Arad (1998) in Italian. Besides, Biaáy (2005) and Rozwadowska (1992,
2012, 2014) work on Polish psych-verbs; Franco (1990), and Jiménez-
Fernández (2014) analyse Spanish psych-verbs; while Kim and Larson
(1989) study Korean psych-predicates. Indeed, psych-verbs have become a
much debated issue in the literature.
4
Interestingly, Tenny (1987), Rappaport Hovav and Levin (1988), and Grimshaw
(1990), on the other hand, hold the view that there seems to be no direct relation
between syntax and the lexical semantics of predicates (LCS) but only between
syntax and the AS.
Syntactic and Semantic Characteristics of Psychological Verbs 79
verbs, i.e. pairs such as fear/frighten, in which the argument bearing the
Experiencer role is mapped, firstly, onto the subject position and secondly,
onto the object position; (iv) variable behaviour verbs, i.e. intransitive
verbs which may appear as both unaccusatives and unergatives, with the
existence of a locative PP (in Dutch and Italian) or without it (in Hebrew).
In short, psych-verbs are among these items which defy the UTAH.
Accordingly, to save Baker’s (1988) hypothesis, alternative views on
psychological predicates have been proposed in the literature, as discussed
in section 2.5.
Furthermore, for Grimshaw (1990), the lexical-syntactic representa-
tion, viz. the argument structure (AS) of a verb is related to the event
structure of a verb. For instance, an accomplishment verb like x constructs
y is analysed as an activity in which x engages in construction plus a
resulting state in which existence is predicated of y (Grimshaw 1990: 26),
as represented in (2.6):
(2.6)
and van Valin and LaPolla (1997: 91-102), among others, decompose
Vendler’s classes in terms of independent features, as reproduced in (2.8).
5
For example, for Pustejovsky (1991) accomplishments and achievements fall into
the so-called “transitions,” for Tenny (1987: 20) achievements and
accomplishments are dissimilar in terms of the duration of an event, while Verkuyl
(1993: 48) concludes that the distinction between achievements and
accomplishments, viz. the duration of events, is a matter of real-world knowledge.
Syntactic and Semantic Characteristics of Psychological Verbs 83
(Bach 1986: 6)
6
Typical examples are: (a) sit, stand, lie + LOC (b) be drunk, be in New York, own
x, love x, resemble x (c) walk, push a cart, be mean (Agentive) (d) build x, walk to
Boston (e) recognize, notice, flash once (f) die, reach the top.
7
Rozwadowska (2012: 535) uses the term “psych-eventualities,” to identify mental
and emotional states and their beginnings. Developing further the ideas of Arad
(1998) and Pylkkänen (1997), she argues that psych-eventualities are what
Rappaport Hovav and Levin (1988) treat as temporally dependent co-existing sub-
events with one event variable, which is identified through one structure
participant only.
8
Cf. Piñón (1997), who distinguishes between two fundamentally different types
of eventualities: (i) those with some duration, which he calls happenings; and (ii)
boundary happenings, which are the initial or final boundaries of some happening.
Boundary happenings are absolutely instantaneous. Piñón (1997) argues that
predicates such as begin or arrive denote boundary happenings, whereas a change
of state predicate like cool does not.
84 Chapter Two
9
Besides, for Rappaport Hovav and Levin (1998), each constant is associated with
a name (i.e. a phonological string). A set of “canonical realization rules” governs
the compatibility of different constant types with different event types (i-vi):
i. manner ĺ [ x ACT<MANNER> ]
(e.g., jog, run, creak, whistle, etc.)
ii. instrument ĺ [ x ACT<INSTRUMENT> ]
(e.g., brush, hammer, saw, shovel, etc.)
iii. placeable object ĺ [ x CAUSE [ BECOME [ x WITH <THING> ]
]]
(e.g., butter, oil, paper, tile, wax, etc.)
iv. place ĺ [ x CAUSE [ BECOME [ x <PLACE> ] ] ]
(e.g., bag, box, cage, crate, garage, pocket, etc.)
v. internally caused state ĺ [ x <STATE> ] (state)
(e.g., bloom, blossom, decay, °ower, rot, rust, sprout, etc.)
vi. externally caused state ĺ [ [ x ACT ] CAUSE [ BECOME [ y
<STATE> ] ] ]
(e.g., break, dry, harden, melt, open, etc.)
(Rappaport Hovav and Levin 1998: 109)
10
The Davidsonian (1967) account of adverbial modification, initially used with
action verbs, has been used to separate the ontological properties of situation types.
Davidson argues that adverbials are attached to an event argument that must
therefore be present in the structure of (eventive) verbs. Since statives do not allow
for such adverbials, they are taken to lack this argument, referred to as the
Davidsonian argument.
11
Maienborn (2003, 2005, 2007) offers a theory of nondynamic expressions,
which distinguishes between state verbs (sleep, sit, stand, lie and wait), consistent
with the criteria for the so-called Davidsonian eventualities, and stative verbs
(know, weigh, own and resemble), which refer to the Kimian (1969) state criteria.
86 Chapter Two
(2009) observes that the nature of stative verbs has not yet been
completely explained, but at least four suppositions about statives can be
made. First, stative predicates are recognised as the smallest and simplest
building blocks of event structure. Second, stative verbs have a less
complex structure compared to eventive verbs. Third, stative verbs are
very similar to copular constructions. Finally, stative verbs contain only a
single (structural) argument, which is the holder of the state (Rothmayr
2009: 39). Taking the distinction between Kimian (1969) and Davidsonian
(1967) statives (cf. Maienborn 2003) as a preliminary assumption, the
results of Rothmayr’s (2009) study of different types of stative verbs
reveal that these predicates cannot be seen as the smallest building-blocks
of event structure. Indeed, it is possible for a verb to express both a state
and an event (either by referring to an intentional activity by an agent, or
by expressing a change of state). Therefore, the event structure must be a
property of the lexical-semantic structure of stative predicates in a verb.
Finally, Rothmayr (2009) offers a two-group division of stative verbs into:
i) verbs that allow stative reading only (SE verbs, such as love, and verbs
of position); and ii) verbs that display systematic stative-eventive
ambiguity (e.g. verbs with both a causative and stative reading, such as
obstruct, dispositional verbs, such as help, verbs with a modal operator
and a feature of action, such as threaten, and verbs of body posture).
According to Rothmayr (2009), the stative/eventive ambiguity can be
explained by a cause-operator in a verb’s semantic structure that relates
two sub-eventualities to one another; the stative interpretation arises when
both sub-eventualities are stative (Rothmayr 2009: 80-82).
Unquestionably, Vendler’s (1967) event classification may be used as
a guide for lexical semantic representations, but it may not be the final
determining factor of event-based theories of argument structure. Apart
from the work of Dowty (1979), Rappaport Hovav and Levin (1998), and
Rothmayr (2009), many theories of lexical semantic representations of a
verb have been developed so far, e.g. Carter (1976), Jackendoff (1983),
Pinker (1989), Parsons (1990), Pustejovsky (1991), Croft (1998), and
Rothstein (2004, 2008), among many others. In general, these alternative
frameworks concern lexical semantic representations of a verb that
decompose events into more primitive predicates.
The latter group comprises copular constructions, regardless of whether they are
stage-level predicates (denoting a temporary property) or individual-level
predicates (carrying a more or less constant property). Kimian states, to be precise,
reflect Kim’s (1969) notion of temporally bound property exemplifications–they
are not discernible, and they cannot be modified by event-related adverbials
(Maienborn 2007), while Davidsonian states can.
Syntactic and Semantic Characteristics of Psychological Verbs 87
12
An interested reader is referred to Alexiadou and Iordăchioaia (2014), who note
some ambiguity among SE psych-verbs in Romanian (between eventive and stative
readings), and in Greek (either only eventive or only stative).
88 Chapter Two
13
Arad (1998) notes that both the stative reading and the non-stative reading are
causatives (as is evident by the causative morphology on OE verbs in Finnish,
Hebrew and Japanese). The type of causation is different in each case: for non-
stative readings it is an active causation, causing a change of state, the other is
stative causation, or triggering a concomitant state (Arad 1998: 6).
14
Cf. Pylkkänen (2000: 431-432) and the correlation between “affectedness” and
“eventiveness” in Parodi and Luján (2000).
15
Likewise, Pylkkänen (2000) identifies a class of causative psych-verbs in
Finnish, which represent temporary ‘stage-level’ states rather than permanent
“individual-level” states (e.g. inho-tta “disgust,” saali-tta “cause to pity,” sure-tta
“cause to be sad;” see also Marín and McNally 2005). Pylkkänen (2000)
recognizes as well another set of causative psych-verbs which are non-stative and
involve the inchoative morpheme (e.g. raivo-stu-tta “cause to become furious,”
kauhi-stu-tta “cause to become terrified,” viha-stu-tta “cause to become angry”).
Syntactic and Semantic Characteristics of Psychological Verbs 91
16
Guidi (2011: 42) argues that this problem of ambiguous aspectual interpretation
can occur for verbs with dative Experiencers too, but he does not discuss this issue
further.
92 Chapter Two
Perlmutter and Postal 1984). These facts make Pesetsky (1995) suggest
that all these verbs are unaccusative as they do not form passives. 17
17
Further evidence for the unaccusativity of escape and elude comes from the fact
that they do not form middles or –er nominals (similarly to concern and interest),
as in (i) (cf. Pesetsky 1995; Levin 1986).
(i) a. * Great ideas elude/escape/concern/interest easily.
b. * an eluder, *an escaper, *a concerner, *an interested.
(Landau 2010:50)
94 Chapter Two
(2.22) The criteria for identifying verbal passive with eventive (non-
stative) psych-verbs:
Grafmiller (2013: 91) explains that the sentences modified by the adverb
suddenly describe a punctual change in the Experiencer, hence the
predicate is required to be interpreted as an event, not as a state.
Consequently, only those verbs which are acceptable with such
interpretations allow verbal passives.
Finally, the needs V-ed construction, as in (2.22c)(i)-(ii), provides
another piece of evidence for the existence of verbal passives with some
OE verbs. The sentence in (22c)(ii) is fine to Grafmiller (2013), even
though it involves two verbs that are normally listed as stative OE psych-
verbs, viz. concern and depress. Indeed, Grafmiller’s (2013) analysis of
the corpus data shows that eventive and stative uses are available for all
OE verbs in both the active and passive, which runs counter to many
authors’ claims (Arad 1998; Bouchard 1995; and Landau 2010; among
others).
Additionally, the needs V-ed construction can be mainly found in
dialects of western Pennsylvania, central Ohio, and other parts of the Great
Lakes area of the U.S. Landau (2010: 51) points out that the Pittsburghese
dialect of English provides further evidence for eventiveness, rather than
agentivity of verbal psych passives (cf. Tenny 1998; Grafmiller 2013: 91-
95).18 Indeed, the more eventive the verb, the more felicitous verbal
passives are. However, Tenny (1998) notes that “a complex of factors
influences the degree of eventiveness, including not only agentivity but
also volitionality, punctuality, and the affectedness of change of state in
the experiencer. ... Individual speakers vary in how strict they are with this
scale in making verbal passives” (Tenny 1998: 595). Therefore, relying on
Tenny’s (1998) analysis, Landau (2010: 51) draws the conclusion that
English provides evidence from independent sources for the possibility of
verbal passive to justify the non-stativity of class II verbs.
What is more, Verhoeven (2010: 18-19, 42-44) carries out some
diagnostic tests for agentivity and stativity, in order to identify semantic
properties of particular verbs of different psych-verb classes in five
different languages. Especially in those languages which display a
18
In her analysis of this dialect, Tenny (1998) explains that the verbal passive
participle construction is well-matched to eventive adverbials, progressive aspect
and idiom chunk passives, and mismatched with the adjectival un-passive, as in
(i)a-(i)d, respectively.
(i) a. The dog needs scratched hard.
b. The car has been needing washed for a long time now.
c. Tabs need kept on the suspect.
d. * The house needs unpainted.
(Landau 2010: 51 (102))
96 Chapter Two
a. class I: SE psych-verbs
Mary loves / fears dogs. (stative reading)
Test 1: not possible in the progressive tenses
*Mary is loving / fearing dogs.
Test 2: not used in imperatives
Love / *fear dogs!19
19
The SE psych-verb “to love” is stative, but shares some properties with eventive
verbs, e.g. the imperative Love children! sounds good.
Syntactic and Semantic Characteristics of Psychological Verbs 97
20
On the unaccusative status of stative OE verbs, cf. section 2.5.
98 Chapter Two
For Belletti and Rizzi (1988), the only aspect in which the three verb
classes differ from one another is the specification of case-grid, which
governs diverse derivations from deep to surface structure. They claim that
the Experiencer is a deep structure subject with verbs belonging to class I
(Nominative Experiencer, accusative Theme). Thus, the D-structure
configuration of sentence (2.36a), which comprises an SE psych-verb
(class I), is illustrated in (2.37):
Hypothesis:
syntactic configurations projected from a given ș-grid should
reflect the hierarchy, so that for every pair of ș-roles in the ș-
104 Chapter Two
In addition, on the basis of the hierarchy and the hypothesis in (2.39), the
following assumption can be made:
Accordingly, Belletti and Rizzi (1988: 344) propose for the three
psych-verb classes the Lexical Representations generating the required
syntactic configurations, as in (2.41):
21
Cf. Pesetsky’s (1987) comments on Belletti and Rizzi (1988)’s assumption about
the identical ș-grids of all the three classes of psych-verbs.
22
Cf. Chomsky (1982) for the Extended Projection Principle (EPP), which
concerns the obligatoriness of subjects.
Syntactic and Semantic Characteristics of Psychological Verbs 105
b. Binding:
The Experiencer in the object position can bind an anaphor
in the subject position, e.g.
Questi pettegolezzi su di sé preoccupano Gianni piú di ogni
altra cosa
These gossips about himself worry Gianni more than
anything else
However, Belletti and Rizzi (1988: 312-313) argue that these examples
involve D-structure binding. Instead, they propose that Principle A is an
‘anywhere principle’, thus it can be satisfied at D-structure, or at S-
Structure (or LF), e.g.
23
Cf. Belletti and Rizzi (1988: 306-308) for a complication with fare + Infinitival
PPs.
106 Chapter Two
24
An interested reader is referred to Belletti and Rizzi (1988: 295-324) for more
details and other arguments in favour of the unaccusativity of class II psych-verbs,
such as: anaphoric cliticization, focus and agentivity effects, arbitrary pro, and
causatives.
Syntactic and Semantic Characteristics of Psychological Verbs 107
25
For an explanation why in the order EXP-V-THEME the Experiencer takes a
Subject rather than a Topic position, cf. Belletti and Rizzi (1988: 339 fn. 32).
108 Chapter Two
Landau (2010) adopts Arad’s (1998) argument that even stative SE verbs
can denote locative relations. Thus, the Experiencer (Monica, he) is either
conceived of as the “substance” contained in the mental state or the
container in which the mental state resides, as illustrated in (2.45).
Landau (2010: 11) argues that his contention that the Experiencer
denotes a mental location holds true, even when the Experiencer occurs as
a bare nominal, as in the case of SE verbs in Hebrew, French and Navajo.
He emphasises that in those languages and in many others, there occur
frequently periphrastic constructions, comprising the verbs be / have, a
psych noun and an Experiencer location, as illustrated in (2.45a)-(2.46a).
Besides, in Irish and Scottish Gaelic, Experiencers are solely introduced
by locative prepositions, as exemplified in (2.46b).
In addition, Landau (2010) emphasises the fact that SEs, which are
stative transitives, behave in a unique way when they co-occur with a
locative preposition (cf. Doron 2003). Accordingly, Landau (2010: 12)
proposes that SE psych-verbs can also be treated as having an oblique
Experiencer, even in languages like English, where they always take the
nominative (nonoblique) form. To justify this assumption, he evokes
Speas’ (1990) arguments concerning SE verbs, which are said to introduce
a path, either as a goal or a source, unlike non-Experiencer Subjects, as
illustrated in (2.47) and (2.48).
Landau (2010: 14) argues that only in the case of SE verbs the
preposition al-yedey “by” is supplanted by a different preposition. In the
beynoni pa’ul of these verbs, the original external argument (the
Experiencer) surfaces with the locative preposition al “on,” as illustrated
in (2.49b). What is more, the same preposition al “on” also occurs with SE
transitive verbs in lexical causativization. There the original object
remains accusative, and the original subject (the Experiencer) becomes
oblique, with the preposition al “on,” as exemplified in (2.50).
(2.51)
(Landau 2010: 8)
26
As stated by Landau (2010), the specific syntactic behaviour of OE is best
visible in peculiarities concerning binding, extraction/islandhood, reflexivization
and argument linearization, etc. (cf. Mohanan and Mohanan 1990)
112 Chapter Two
which assigns the dative case. The VP structure of class III psych-verbs is
then as in (2.52).
27
To recall briefly, Belletti and Rizzi (1988) identify for Italian Class I of
psychological verbs as the one with the uncontroversial transitive structure, since
the Experiencer has the external ș-role, and no inherent Case is assigned. Class II
and class III of psych-verbs, with no external ș-role, and the Experiencer
associated with an inherent Case (accusative or dative) have, according to Belletti
and Rizzi (1988), an unaccusative structure.
114 Chapter Two
28
Landau (2010: 19-20) deals with Belletti and Rizzi’s (1988) claim concerning
the unaccusativity of II and III class psych-verbs. He says that it is unproblematic
for class III verbs, which assign dative case to the Experiencer and select the
auxiliary essere. Nonetheless, class II verbs with these two properties are never
found: they select the auxiliary avere, and assign the accusative case, violating
Burzio’s (1986) generalization.
29
The trees by Landau (2010) and Harwood et al. (2017) (e.g. (4.54) in Chapter
Four) are syntactic trees containing reference to some semantic roles.
116 Chapter Two
Landau (2010: 75), into core and peripheral ones, with the information
which languages they are typical of, listed in (2.56):
30
This approach is compatible with Marín and McNally’s (2011) account and an
earlier analysis of psych-verbs offered by van Voorst (1992). For a further
discussion concerning the aspectual properties of psych-verbs cf. Grimshaw
(1990), Pesetsky (1995) and Pylkkänen (2000), among others.
118 Chapter Two
Landau (2010: 29-30) argues that the kind of violation exhibited in PP-
extraction in (2.60d) is as unacceptable as extraction of genuine
Experiencer direct objects, which he takes to support his analysis of
Experiencers as arguments of null prepositions. Landau offers a syntactic
analysis of English OE verbs in which they do not take complement NPs
(or DPs) as do canonical transitive verbs, but instead select for PP
complements headed by a null preposition (Øȥ). Since objects of OE
verbs are arguments of null prepositions, extraction from within these null-
headed PPs should reveal the same degree of unacceptability as extraction
from overt-headed PP complements found with other verbs.31
Furthermore, Landau (2010) faces the controversial question of
whether class II verbs have a verbal passive. This is where the subdivision
of class II into three groups becomes vital. Hence, stative class II verbs
(and all class III verbs) generally fail to passivize since they are
unaccusative. Agentive eventive class II verbs are usual transitive verbs,
and therefore universally allow passivization. Non-agentive eventive class
II verbs, which are not unaccusative, allow pseudopassives (i.e.
prepositional passive, a form of English passive voice in which the object
of a preposition becomes the subject of a clause). However, the verbs with
the oblique nature of Experiencers are expected not to passivize unless the
language can resort to the special strategies, presented in (2.61).
31
Moreover, Landau (2010: 30-31), similarly to Grimshaw (1990: 15), analyses
some other peculiarities of OE verbs, namely their inability to form synthetic
compounds, involving a deverbal head and its object (a god-fearing man, a fun-
loving teenager, *a man-frightening god, *a parent-appalling exploit). Another
characteristic of OE verbs is their resistance to Heavy NP Shift (HNPS), analogous
to the inner object in the double object construction, as in (i)-(ii). While overtly
prepositional Experiencers, as in (iii) below, are perfectly moveable.
(i) * These things bothered yesterday the man who visited Sally.
(ii) * We told these things (yesterday) the man who visited Sally.
(iii) These things appealed yesterday to the man who visited Sally.
120 Chapter Two
Pseudopassive:
[TP [DP Exp]1 [T’ Aux [VP [V VPASS + Øȥ ][DP t1 ] ]]]
Quirky passive:
[TP [PP Øȥ [DP Exp]]1 [T’ Aux [VP VPASS [PP t1 ] ]]]
(Landau 2010: 48)
32
Cf. Landau’s (2010) arguments for psych-verbs, which can passivize in Finnish.
There the aspectual distinctions are morphologically marked, so “the relevant
judgments need not appeal to subtle semantic intuitions” (Landau 2010: 54).
Syntactic and Semantic Characteristics of Psychological Verbs 121
33
In fact, it was Pesetsky’s (1995) original claim that OE psych-verbs subsume SE
psych-verbs.
122 Chapter Two
(2.64)
(2.65)
34
I would like to thank Prof. Henryk Kardela for suggesting the term “cognitive-
semantic-syntactic trees” for the syntactic structures of SE and OE psych-verbs
offered by Fábregas and Marín (2015). In their structure the notion of state
belongs to event structure. This type of event structure is not mentioned either in
Landau’s (2010) or Harwood et al.’s (2017) syntactic structures. In addition, the
trees by Fábregas and Marín (2015) refer both to Dowty (1979) and his semantic
concepts of State, or Causer, and to Jackendoff (1990) and his cognitive-semantic
decomposition analysis. In turn, the trees offered by Koopman (1991, 2010) for
verb particle constructions (cf. (4.69)) are clear syntactic structures.
35
Class I (SE) verbs comprise individual-level and stage-level predicates (love vs.
worry), and the latter are often related to inchoative or reflexive morphology,
which is likely to have an agentive interpretation (cf. Pesetsky 1995; Reinhart
2002; and Pylkkänen 2000).
Syntactic and Semantic Characteristics of Psychological Verbs 123
36
Grafmiller (2013) ascertains that providing further insight into the relationship
between language and emotion concepts lies at the heart of his research. To
understand this relationship, he explores how the conceptual properties of
emotions are encoded in the words and constructions used to describe them.
Assuming that human beings build mental concepts of various kinds which reflect
their experience, Grafmiller (2013: 1-2) notes that many of these concepts are
encoded in the meanings of individual words, often conceived of as “entries” in a
mental lexicon (e.g. Jackendoff 1989; and Pustejovsky 1995). Lexical entries
comprise varying degrees of semantic information, conceived of as sets of
privative features, thematic role lists, and/or event structures, and that words are
124 Chapter Two
individuated conceptually in terms of the information that they denote. With regard
to verbs, Grafmiller (2013: 2) argues that in a wide variety of theoretical
approaches a verb’s semantic representations determine the range of syntactic
realizations of its arguments (e.g. Dowty 1991; Jackendoff 1990; Langacker 1987;
Pinker 1989; Rappaport Hovav and Levin 1998; van Valin 1990; and van Valin
and LaPolla 1997; among others).
Syntactic and Semantic Characteristics of Psychological Verbs 125
3.1 Introduction
The aims of Chapter Three are twofold. On the one hand, we strive to
select psych-verbs, which are a representative set of this type of
predicates. On the other, we intend to determine idioms which are to be
analysed in Chapter Four, and which correspond to the psych predicates
listed earlier. Besides, the methodology adopted in data selection is also
explained in Chapter Three. Two stages of the corpus study need to be
conducted to meet the objectives just stipulated. The first stage involves
eliciting the most frequent class I and III psych-verbs (cf. Belletti and
Rizzi 1988), while class II top psych-verbs are adopted after Grafmiller
(2013). The second stage is based on searching for idiomatic phrases
corresponding to the psychological verbs selected.
The guidelines adopted in data selection in the first stage of the study
are as follows: (i) the psych-verbs to be elicited correspond to the
psychological predicates in Belletti and Rizzi’s (1988) tripartite typology
of psych-verbs, described in detail in Chapter Two of the book; (ii) only
the most frequent psych-verbs, with the top occurrence in The Corpus of
Contemporary American English (COCA), are taken into account; and (iii)
the psych-verbs to be selected belong to various emotion domains.
Once the set of psych-verbs is determined, the second stage of the
study is embarked upon. It aims at investigating any possible
phraseological units, collected from English dictionaries and thesauri,
which correspond to simple psych-verbs. Not only can these idiomatic
phrases be used as substitutes for simple psych-verbs, enriching our
language inventory, but also these idioms will be compared with their
basic psych-verb counterparts in Chapter Four of the book, as regards their
syntactic and aspectual characteristics. As a result, the research is expected
to provide some insight into the complex nature of psychological
predicates.
130 Chapter Three
The chapter is organised into four sections. In section 3.2 the three
above-mentioned criteria of selecting the basic psych-verbs are discussed
in detail. Here, the focus is laid on determining the most frequent class I
(SE) and class III (OE) psych-verbs, based on the corpus analysis. In
section 3.3, the purpose of the research into psych-idioms, the data
collection method, and the methodology adopted in the corpus study are
outlined. Section 3.4 reviews the data analysed and offers some discussion
related to the results of the research. The chapter ends with a brief
summary and provides a scope of further investigation, carried out in
Chapter Four of the book.
Even though psych-verbs have already been defined and discussed in the
preceding chapter, some crucial points concerning those predicates are
repeated here, for the sake of convenience. On the basis of the well-known
studies represented by Belletti and Rizzi (1988), Pesetsky (1995), and
Landau (2010), among others, it can be specified that psych-verbs express
(a change in) mental or/and emotional state and a relation between the two
arguments: an Experiencer and the Cause / Theme of such a psychological
condition. An Experiencer can be realized as either a subject or as an
object, following Belletti and Rizzi’s (1988) tripartite classification, as
introduced at the very beginning of this book, and illustrated in (2.3) in
Chapter Two, but repeated in (3.1) below for the sake of convenience.
Due to the fact that certain psychological states are potentially more
common than others, it is expected that the frequency of usage of these
psych-verbs is far higher than the frequency of other emotion / mental
verbs in given corpora. For instance, everyone can recognise the priority
of the state of “love “ over “charm “ or “awe “ with the help of some
basic intuition. Moreover, corpus linguists suggest that “the frequency
distribution of tokens and types of linguistic phenomena in corpora have
(..) some kind of significance. Essentially, more frequently occurring
structures are believed to hold a more prominent place, not only in actual
discourse but also in the linguistic system, than those occurring less often”
(Schmid 2010: 101). Finally, assuming that the essential requirement for
132 Chapter Three
1
Interestingly, some specialists in modern English lexicology (Ginzburg et al.
2004; among others) argue that “frequency value of the word is as a rule the most
reliable and objective factor indicating the relative value of the word in the
language in general and conditioning the grammatical and lexical valency of the
word” (Ginzburg et al. 2004: 180). Besides, the frequency value of the word can
be in many cases “sufficient to judge of its structural, stylistic, semantic and
etymological peculiarities, i e. if the word has a high frequency of occurrence one
may suppose that it is monomorphic, simple, polysemantic and stylistically neutral.
Etymologically it is likely to be native or to belong to early borrowings” (ibid.:
180).
In Search of Psychological Idioms: A Corpus-based Account 133
Grafmiller (2013) lists seven more top psych-verbs, i.e. captivate (313)
and amaze (268), which I have left aside since they are synonymous with
the verb fascinate, listed in (3.3). The verb frighten (202) is synonymous
with scare; thus, it is omitted as well. Verbs, astonish (169), startle (133),
please (130), and upset (121) are not analysed either, because of their
134 Chapter Three
3.2.2.2 The most frequent class I (SE) and class III (OE) psych-verbs
in the COCA Corpus
In order to elicit the representatives from all Belletti and Rizzi’s (1988)
classes of psych-verbs, Grafmiller’s (2013) results obtained for the most
common class II OE psych-verbs must be extended to cover both class I
(SE) and class III (OE) psych-verbs. Therefore, class I and III psych
predicates have become the material of my study.
With the aim to select the most frequent members of class I and class
III psych-verbs which appear in the COCA Corpus, I have searched this
online corpus with respect to the occurrence of the predicates under
scrutiny. For the purpose of this research, the list of psych-verbs offered by
Levin (1993) has been adopted as the dataset. Class II psych-verbs,
represented in Levin’s typology by Amuse Verbs (220 members, e.g. amuse,
bother, concern, frighten, scare), have been excluded in order not to repeat
the study, carried out by Grafmiller (2013) (cf. section 3.2.2.1 above).
In comparison with the variety of verb taxonomies offered in the
literature, it is Levin’s (1993) classification which is considered to be far
more sophisticated and thoroughly researched.2 Indeed, in her seminal
work, Levin (1993) categorised over 3,000 English verbs, along with their
shared meaning and behaviour, which makes her classification one of the
most extensive ones. All of the over 3,000 English verbs are put into 57
semantically coherent classes and around 200 subclasses.3 In her
2
There have been numerous attempts to classify English verbs with various
differentiating factors, e.g. argument structure and semantic characteristics, event
structure or thematic role structure (cf. Vendler 1967; Croft 1991; Fillmore and
Baker 2001; and Kipper-Schuler 2005).
3
Those include, e.g. Verbs of Putting; Verbs of Removing; Verbs of Sending and
Carrying; Verbs of Exerting Force: Push / Pull Verbs; Verbs of Change of
Possession; Learn Verbs; Hold and Keep Verbs; Verbs of Concealment; Verbs of
Throwing; Verbs of Contact by Impact; Verbs of Contact: Touch Verbs; Verbs of
In Search of Psychological Idioms: A Corpus-based Account 135
those most frequent psych-verbs which represent class I (SE) and class III
(OE) psych-verbs in Belletti and Rizzi’s (1988) syntactically-oriented
typology.
When Levin’s (1993) group of psych-verbs is organized in accordance
with Belletti and Rizzi’s (1988) typology, the classification of psych-verbs
looks as in (3.4).
(3.4) Levin’s (1993) set of 374 psych-verbs within Belletti and Rizzi’s
(1988) tripartite syntactically-oriented typology:
7
http://corpus.byu.edu/coca/ and http://www.wordandphrase.info/frequency
List.asp
In Search of Psychological Idioms: A Corpus-based Account 137
8
Online International Dictionary available at http://idict.org, retrieved 5/12/2016.
138 Chapter Three
contact with (something aimed at): a laser-guided bomb had missed its
target; he shot twice at the cashier, but missed both times.” The verb
engage is defined as “to arrange to employ or hire (someone): he was
engaged as a trainee copywriter;” whereas the verb care means “look after
and provide for the needs of someone.” Undoubtedly, these verbs can refer
to a psychological state, i.e. miss “fail to notice someone,” engage
“attract someone’s interest,” and care “feel concern or interest” but these
are not their sole and main readings, in contradistinction to the other verbs
from the top group under scrutiny. As a result, it would hardly be possible
to discern manually a psychological reading from the non-psychological
one, and present accurate frequency details afterwards. In short, any psych
predicate with an ambiguous reading from the list of 154 items tested has
been excluded from my further research.
Indeed, the search has yielded quite a big number of psych-verbs with
a high frequency of occurrence in the COCA Corpus. It has been found
out that in the dataset of 154 psych-verbs, there are 13 verbs with over
10,000-token-occurrence, excluding those with ambiguous meanings. Two
of them represent class III psych predicates, i.e. matter and appeal;
whereas eleven belong to class I verbs in Belletti and Rizzi’s (1988)
typology, as illustrated in (3.5) below.
(3.5) 13 psych-verbs of class I and III with the highest frequency in the
COCA:
The top 13 psych predicates, listed in (3.5), occur in the COCA Corpus
with the frequency over 10,000. They are also, as expected, included in the
COCA “5,000 frequency word list,” viz. the first 5,000 most frequent
words / lemmas in the whole COCA Corpus.9 Interestingly, Nation (2006),
9
Schmitt and Schmitt (2014) state that for English, “high-frequency vocabulary”
has traditionally been understood as around the first 2,000 most frequent word
families. On the other hand, low-frequency vocabulary has been characterized in
various ways, ranging from anything beyond 2,000 word families all the way up to
all of the word families beyond the 10,000 frequency level. Any frequency list is
accurate provided the corpus (collection of texts) that it is based on is a reliable
source as well. The 450 million word COCA Corpus, which has been chosen as the
source corpus for this study, also contains its own frequency data of English
In Search of Psychological Idioms: A Corpus-based Account 139
and Schmitt and Schmitt (2014), among others, take the number 2,000 as
the borderline for high-frequency vocabulary. Thus, psych-verbs ranked
from 1 up to 2,000 can be treated as representatives of “high-frequency
vocabulary” in the language, while those below the 2,000th position are
supposed to be representatives of mid-frequency or low-frequency
vocabulary (Schmitt and Schmitt 2014: 501). Table 3-1 below presents the
ranking position and the exact frequency rates on the COCA frequency
word list for the 13 psych-verbs from class I and III.
Table 3-1. The ranking position of the top 13 psych-verbs on the COCA 5,000
frequency word list10 (above the number of 10,000, ordered from highest to lowest)
ranked third in the set of 13 top psych-verbs. Then come the verbs love,
enjoy and worry, which occupy the ranking positions up to the first top
1,000 words. Next, the verbs such as hate, fear, appreciate, matter, and
trust belong to the period between 1,000 and 2,000 of the most frequent
words. Finally, the bottom of the list is covered by the verbs respect and
appeal, which are in the 2836th and 2927th position, respectively. The
frequency of all of the 13 psych-verbs ranges from 514,972 highest to
10,745 lowest in the whole COCA Corpus. Indeed, being so frequently
used in the language inventory, these are the predicates which deserve to
be selected as the starting point for the further stage of the research,
devoted to idiomatic phrases and carried out in section 3.3 of this chapter.
Obviously, since the first top 13 psych verbs have been extracted from the
initial set of 154 class I and III psych-verbs, the remaining 141 verbs have
been classified as less frequent, thus they are outside of our interest and
are left aside here.
In addition, 9 verbs from class II, adopted from Grafmiller’s (2013)
study, have been checked against the current COCA 5,000 frequency word
list. The results obtained are presented in Table 3-2. However, it should be
borne in mind that the number of occurrence of the tokens given by
Grafmiller (2013) was obtained by means of his original Python
programming and thicker filters he had applied to choose the best
examples of psych-verbs and their argument structures.
On the basis of the results presented in Table 3-2, it can be observed
that some of the psych-verbs examined by Grafmiller (2013) are not
included in the list of the first 5,000 most frequent words, updated in 2016.
Only the verbs surprise, scare, and concern can be found between 3,000
and 4,000 in the ranking position. Nonetheless, without shadow of a doubt,
exactly these predicates are commonly referred to in the linguistic
literature, therefore they are highly eligible for further analysis.
In Search of Psychological Idioms: A Corpus-based Account 141
Table 3-2. Top class II OE psych-verbs from Grafmiller’s (2013) study and their
ranking position checked against the COCA 5,000 frequency word list11 (with the
number of tokens after Grafmiller’s filtering)
Class II psych-verbs
No Rank with the number of tokens Frequency
(Grafmiller 2013)
1. 3086 surprise 389 10023
2. <5000 annoy 366 >10000
3. <5000 fascinate 285 >10000
4. <5000 amuse 283 >10000
5. 3844 scare 272 7367
6. <5000 depress 210 >10000
7. <5000 anger 207 >10000
8. <5000 horrify 159 >10000
9. 3439 concern 137 8702
Finally, the list of verbs most popular in the literature and in the COCA
Corpus relating to emotional and psychological conditions, is expected to
comprise the complete section of all emotional states, both negative and
positive ones. Thus, the set of 9 class II psych-verbs from Grafmiller’s
(2013) work, and 11 class I and 2 class III psych-verbs from my study (cf.
Table 3-1), elicited so far, should be filtered for the final selection, to
comply with the 5 basic emotion domains referred to in the literature
cross-linguistically.
11
Available at http://www.wordfrequency.info/top5000.asp, last retrieved
5/01/2017.
142 Chapter Three
12
Another vital model of viewing emotions is offered by Scherer (2005), who calls
it the Component Processing Model of Emotion, comprising five crucial elements,
i.e. the cognitive component, the peripheral efference component, the motivational
component, the motor expression component, and the subjective feeling
component. From Scherer’s (2005) perspective, emotion experience makes all of
these items coordinated and synchronized for a short period of time, driven by
appraisal processes. Besides, emotions are described as discrete and consistent
responses to internal or external events which have a particular importance for the
organism (cf. Scherer 2005; Frijda 1986, 2007). Scherer (2005) classifies emotions
under the group of short-lived affective phenomena, in contradistinction to moods,
attitudes and personality traits, which are long-lasting. Rozwadowska and Willim
(2016) draw a conclusion that if Johnson-Laird and Oatley’s (1989) approach were
combined with Scherer’s (2005) model and Hartshorne et al.’s (2010) proposal of
distinguishing brief psychological states (fright, anger, delight) from stable
tendencies, viz. dispositions (love, liking, hatred), then “from the psychological
perspective dispositions / attitudes / emotional relations tend to be long-lasting,
whereas causatives and responses to external or internal stimuli are short-lived”
(Rozwadowska and Willim 2016: 19; cf. Myers 2004).
In Search of Psychological Idioms: A Corpus-based Account 143
The wheel comprises eight primary emotions grouped into positive and
negative ones, i.e. joy versus sadness; anger versus fear; trust versus
disgust; and surprise versus anticipation. In addition, some basic emotions
can blend to form complex emotions. The latter could arise from blending
cultural conditioning or connotation with the basic emotions, e.g. a
mixture of interpersonal anger and disgust may lead to contempt (Plutchik
2001: 350). What is more, Plutchik (2001) argues that advanced emotions,
such as love, are a combination of two or more generic emotions; thus,
love, experienced as a strong feeling of attachment and attraction, is a
combination of joy and trust. While remorse, also regarded as an advanced
emotion, is characterised by feelings of regret and sometimes shame,
whereas regret is a combination of surprise and disgust. In short, Plutchik
(2001) suggests that besides the eight basic emotions, the two, i.e. love and
144 Chapter Three
remorse, form a set of the top ten emotions which each human being
mostly faces in his/her life.13
Additionally, certain emotion words characterised as culture-specific
phenomena can be, nevertheless, comparable to their translational
counterparts. As maintained by Johnson-Laird and Oatley (1989, 1992),
Ekman (1999), Turner (2007), Rott and Verhoeven (2016), some basic
emotional modes are assumed to be stable across cultures. These are
happiness, sadness, fear, anger, and disgust. Plutchnik’s (1980) much
more complex wheel of emotions (see Figure 3-1) comprises these five
basic modes, as well. And these precisely are the most common subject
fields, called emotion domains,14 which are mostly alluded to in the
literature while discussing emotions.
13
What is more, having examined a wide range of emotion theories across
different fields of research, Turner (2007, 2009) identifies four emotions that all
researchers relate to human neurology, viz. assertive-anger, aversion-fear,
satisfaction-happiness, and disappointment-sadness. These four categories are
termed primary emotions, which, when combined, may produce more elaborate
emotional experiences, called first-order elaborations, including sentiments such as
pride, triumph, and awe. In Turner’s theory (2007), emotions can be experienced at
different levels of intensity; thus, feelings of concern are a low-intensity variation
of the primary emotion aversion-fear, whereas depression is a higher intensity
variant. Two of the main eliciting factors for the arousal of emotions within this
theory are expectation states and sanctions.
14
In linguistics the term “domain” denotes “a subject field which has a particular
set of vocabulary associated with it” (Macmillan Dictionary retrieved 30/11/2016
from http://www.macmillan dictionary.com/dictionary /british/domain). This
general definition of “domain” is meant in this context, without making reference
to a far broader concept of “domain” brought by cognitive linguistics.
Cognitive linguistics, in turn, introduces the terms “conceptual domain,”
”conceptual metaphor,” or ”cognitive metaphor,” referring to the understanding of
one idea, or conceptual domain, in terms of another, “mapping the structure of one
domain onto the structure of another” (Lakoff 1987). A conceptual domain deals
with any coherent organization of human experience, and it can be exemplified by
the understanding of quantity in terms of directionality (e.g. “the price of peace is
rising”) (Lakoff 1987). Besides, within the framework of Conceptual Metaphor
Theory (Lakoff and Johnson 1980), metaphors link two conceptual domains, the
“source” domain and the “target” domain. While the “source” domain comprises a
set of literal entities, attributes, processes and relationships, linked semantically
and apparently stored together in the mind, the “target” domain tends to be
abstract, and takes its structure from the source domain, through the metaphorical
link, or “conceptual metaphor.” Target domains are supposed to have relationships
between entities, attributes and processes which mirror those found in the source
domain. At the level of language, “source” domains can be expressed through
In Search of Psychological Idioms: A Corpus-based Account 145
As can be seen in (3.6), 15 of the initial list of the 22 top psych-verbs seem
to match the basic emotion domains intuitively well. The positive emotion
of HAPPINESS comprises 7 psych-verbs which are generally associated
with the state of being delighted or positively influenced. The remaining
four domains, i.e. SADNESS, ANGER, FEAR, and DISGUST, are
negative ones, covering altogether 8 psych-verbs. Thus, the balance is
maintained between positive and negative emotions. However, using the
common intuition, some verbs may be cross-domain specific, e.g. worry15
refers both to SADNESS and FEAR; and hate is linked to DISGUST,
ANGER and SADNESS. In these cases, the definitions of the verbs
provided by Johnson-Laird and Oatley (1989) are decisive, i.e. worry
denotes ANXIETY / SADNESS, while hate refers to DISGUST (Johnson-
Laird and Oatley 1989: 114, 121). Likewise, the verbs surprise, appeal to,
and matter to have been marked with an asterisk to indicate some
difficulty to classify them. In fact, they can be defined as “causing a
sudden unexpected onset of an emotion (generally positive ones),” and the
key to group them under the “happiness” label lies in the positive feeling
they refer to. Besides, the verbs appeal to and matter to are the only
representatives of class III (OE) psych-verbs; that is why, their presence in
the set of psych-verbs, selected for further research, is more than required.
Moreover, seven of the initial list of the 22 top psych-verbs (cf. Table
3-1 and Table 3-2) have not been categorised into the basic emotion
domains at all, and as a result have been left aside. They are systematically
ambiguous due to their multiple meanings, even polar ones. For instance,
the verb concern may refer to anxiety or sympathy for someone else, and
thus, it may belong either to the HAPPINESS (positive) or to SADNESS
(negative) domain. Similarly, want / need are defined as “to have a goal
which if attained causes happiness (or makes good deficiency);” hence,
they do not refer to one specific emotion. Another kind of ambiguity and
difficulty to assign certain psych-verbs to the basic emotional domains, as
Clore et al. (1987) point out, arises in the use of words that do not, strictly
speaking, imply emotions but convey an emotional / mental state, e.g.
“feeling inspired” or “feeling in need of something.” This refers to the
verbs trust “to believe and rely on,” and respect “to judge that someone
deserves to be admired” from Table 3-1. Therefore, these items are also
excluded from further research. Finally, the verb like (cf. Table 3-1) has
been crossed out from the final list of psych-verbs since it has a substitute,
i.e. the verb enjoy. Finally, the verb appreciate from Table 3-1 has not
been chosen either, on account of the fact that, representing the positive
emotional domain, it may range in scale of emotion, from extreme
exhilaration to a neutral state of accepting something.
As can be seen in (3.8), the figurative interpretation of the idiom show the
white feather is based upon the elements show, the, white, and feather.
In summary, the working definition provided here complies with the
definition of idioms, called also phraseological units or phrases,
introduced in Chapter One of the book. It should be remembered that in
most cases idiom constituents do not contribute to the overall meaning of
the idiomatic phrase, then the idiomatic unit should be recognised as a
metaphor and a cohesive entity treated as a whole.
The research reported here is based on English data collected from English
dictionaries and thesauri listed in (3.9), in which the idiomatic synonyms
of the psych-verbs under scrutiny have been searched for.
16
In this book, # is used to indicate loss of the idiomatic meaning.
In Search of Psychological Idioms: A Corpus-based Account 149
Every single verb has been first checked and associated with a specific
definition, for which an idiomatic equivalent has been searched.
The study focuses on psychological verbal idioms, i.e. instances in
which a psych meaning occurs within idiomatic expressions that begin
with a verb. A number of criteria defining a verbal idiom have been
identified in the literature (Marantz 1984; Kiparsky 1987; Nunberg et al.
1994; Croft and Cruse 2004; Svenonius 2005; Evans and Green 2006; and
Harwood et al. 2016; among others). These are summarised in (3.10)
below.
Accordingly, any expressions which form a literal meaning and lack any
figurative / idiomatic meaning, have been omitted. The precise instances
excluded from the study by the criteria mentioned in (3.10) are listed in
(3.11).
(3.11) Phraseological verbal units which have been excluded from the
study by the criteria listed in (3.10):
In total, the database thus created contains 161 English idiomatic units,
which are grouped according to the surface syntactic patterns they occur in
for each of Belletti and Rizzi’s (1988) classes.
Taking into consideration the pattern the core of an idiom forms and
the position an argument is placed in, the canonical idioms I have
scrutinised fall into 12 distinct types, schematized in (3.12).17
17
Interestingly, Bruening (2010), cited in Kim (2014), argues that in English there
are three existent classes of idioms with Double Object CConstructions (DOC) and
prepositional dative constructions / prepositional datives (PD). He suggests the
following overall distribution of idiomatic patterns with DOC and PD
constructions (idiomatic parts are underlined, X in brackets is a variant):
a. Class I: Verb NP NP (give X the creeps)
b. Class II: Verb NP to NP (give rise to X)
c. Class III: Verb NP to NP (send X to the showers)
d. Class IV: V NP NP (nonexistent)
(Kim 2014: 216)
In Search of Psychological Idioms: A Corpus-based Account 153
The data in (3.13)-(3.17) below illustrate all the types of idioms possible
for five SE psych-verb which have been on the research list, i.e. love,
enjoy, hate, fear, worry. All the elicited idiomatic phrases are arranged
according to a syntactic pattern they exhibit, while the exemplary
sentences for those idioms, taken from the COCA and/or obtained via the
Google Search, are listed in APPENDIX 1.
18
The genitive possessor in the case of this idiom has to agree with the subject,
represented by an Experiencer (Y).
In Search of Psychological Idioms: A Corpus-based Account 155
(3.15) hate - feel strong dislike for or hostility toward (10 idioms)
Type (B): V + NP + preposition + NP (complement of a P)
pour scorn on X
bear ill will toward X
bear / owe a grudge against X
bear / feel aversion / malice / hostility / repugnance
toward (to) X
have no stomach / taste / use for X
show contempt for X
19
“Red” in this idiom is an optional element, though if used, it is a part of a
resultative phrase.
20
“The particle “off” is the constituent of the verb “let/blow” itself, not a
preposition requiring a complement.
21
The idioms hold X in contempt and look down Y’s nose at X refer to one’s
CONTEMPT and mean “to spurn, to regard oneself as superior to others and thus
act in a haughty or snobbish manner”. Nonetheless, these idioms have been
grouped into the basic emotion domain HATE, which consists of disgust, anger,
and/or sadness (cf. Johnson-Laird and Oatley 1989: 121). Truly, as suggested by
Prof BoĪena Cetnarowska, the aforementioned idioms might have been grouped as
156 Chapter Three
Type (A): V + NP
have the blues
24
Cf. (3.11) in Chapter Three, where the examples of phraseological verbal units,
excluded from the study by the criteria listed in (3.10), are listed. The phrase have
butterflies in the stomach comprises the verb have and a NP, and it obtains a
figurative meaning “to worry.” Therefore, it is not excluded from the list of idioms
to be examined.
158 Chapter Three
A A’ A’’ B B’ C D D’’
Verb/
Tokens V+ V+ V+ V+ NP V+ NP V+NP V + V+
NP possessor possessor +P +P+ +P +NP possessor possessor
+N + N+ +NP possessor +N +N
particle +N +P+NP +P+
possessor
+N
love 13 --- --- --- 11 --- --- 2 ---
enjoy 11 4 1 --- 6 --- --- --- ---
hate 10 --- --- --- 6 --- 1 3 ---
fear 9 6 --- --- 1 1 --- --- 1
worry 6 1 --- 1 1 3 --- --- ---
TOTAL /
11 1 1 25 4 1 5 1
49
In Search of Psycchological Idiom
ms: A Corpus-bbased Account 159
Syn
ntactic patteerns for SE psychologicall idioms A
(in percenta
age)
A'
D''
D
C 2% A A''
%
10%
2% 23%
B
B' A'
8% 2% B'
A''
2% C
B D
51%
D''
Figure 3-2. S
Syntactic patterrns for SE psych
hological idiom
ms (in percentag
ge)
25
In Type A, all the idiom components are fixed, and an Experiencer (Y) is
realised covertly, but it can be found in the context of a particular sentence, e.g. in
(i) below, American fighters are Experiencers, while in (ii) Tony is an Experiencer
(placed in the position of a subject), who got annoyed by us and the fact that we
hadn”t invited him to the party.
(i) play (a game of) cat and mouse [against /towards / with Y] “to annoy Y”
Enemy warplanes have been playing a deadly game of cat and mouse [against
American fighters], trying to bring American fighter planes into rang of their
missiles. (Google)
(ii) get the hump “to get annoyed” Tony got the hump because he thought we
hadn”t invited him to the party. (Google)
In Search of Psychological Idioms: A Corpus-based Account 161
26
If the component “to death /to pieces” in the idiom “tickle Y to death / to pieces
/ pink” is changed into “pink,” the idioms forms Type (E”): V + a complement
small clause.
In Search of Psychological Idioms: A Corpus-based Account 163
27
This idiom belongs to Type A, in which all the idiom components are fixed,
whereas an Experiencer (Y) is realised covertly, but recognisable from the context
of the sentence, e.g. in (i) below, you is the Experiencer, who experiences the state
of being depressed.
(i) upset the applecart [in Y] “to depress Y”
I don”t want to upset the applecart now by asking you to change the date
for the meeting. (Google)
164 Chapter Three
28
This idiom represents Type A, with all components fixed. Therefore an
Experiencer (Y) is realised covertly, but still it can be noticed from the context of
the sentence, e.g. in (i) below, she is the Experiencer, who experiences the state of
being angered.
(i) fan the fire / flame (of something) [in Y] “to anger Y”
She already found him attractive, but his letters really fanned the flames
[in her]. (Google)
29
The particle “up” is a constituent of the verb “get.”
In Search of Psychological Idioms: A Corpus-based Account 165
30
The particles “away,” and “off” work as particles of the phrasal verbs.
31
The particles “down / over” work as particles of the phrasal verb “knock down”
or “knock over.”
32
In the idiom “catch Y up short,” “short” is an adjective, which works as a
complement of the preposition “up,” and may be treated as a resultative phrase
(Type E”).
33
If the component “off balance” in the idiom “catch Y off balance / up short /
napping” is changed into “napping,” the idiom forms Type (E”): V + a
complement small clause.
34
All these idioms include a resultative phrase, i.e. “open-mouthed / at a loss for
words / dumb / with awe / with wonder,” which are treated as complements of the
small clauses.
166 Chapter Three
Table 3-4. Types of syntactic patterns for OE (class II) psychological idioms
V+ V+ V+ V+ V+ V+N V+ V + V V V+a
NP posse- posse- NP + NP + P posse- posse- +NP +NP+ compl.
ssor+ ssor+ P P+ +P ssor+ ssor+ +P+ NP small
clause
N N+par +NP possess +NP N +P N posse-
-ticle
or+ N +NP +P+N ssor+
P N
annoy
2 5 2 4 1 7 1 1 --- 2 1
/ 26
fasci-
--- 5 --- 3 --- 2 --- --- 1 --- ---
nate/ 11
amuse
--- 1 --- 2 --- 1 --- --- --- --- ---
/ 4
scare
--- 1 --- 4 --- 3 --- 1 --- 2 ---
/ 11
In Search of Psycchological Idiom
ms: A Corpus-bbased Account 167
depress
/ 20
1 4 --- 4 2 7 --- ---- --- 1 1
anger
/ 10
1 1 2 2 --- 3 --- 1 --- --- ---
surprise
/ 15
---- --- 2 --- --- 7 --- 1 --- 1 4
horrify
/ 8
---- 1 --- 4 --- 2 --- ---- --- 1 ---
TOTAL
/105
4 18 6 23 3 32 1 4 1 7 8
D'
14% A'
A'
43%
B
B
43%
D'
Figure 3-4. S
Syntactic patterrns for OE (classs III) psychologgical idioms
(iin percentage)
170 Chapter Three
PSYCHOLOGICAL IDIOMS:
SYNTACTIC CONSTRAINTS
AND ASPECTUAL PROPERTIES
4.1 Introduction
Idioms represent “a multidimensional lexical space, characterized by a
number of distinct properties: semantic, syntactic, poetical, discursive, and
rhetorical” (Nunberg et al. 1994: 492); therefore, any attempt to categorise
idioms along single-criterion definitions or rules is futile. Taking into
account various dimensions of idiomaticity, two main opposing
approaches to idioms are usually distinguished in the literature, which
have been outlined in Chapter One of the book, and are now briefly
returned to, for the sake of convenience. These are: (i) the more standard
non-compositional approach, in which idioms are recognised as arbitrary
configurations of words with nonliteral meanings, syntactically and
semantically behaving as lexical entries; and (ii) the non-traditional
compositional approach, which refers to the non-arbitrary internal
semantic and syntactic structure of idioms.
In the traditional approach, typically adopted in the generative
grammar, it is believed that the meanings of idioms cannot be derived
compositionally by the morpho-syntactic rules of a language (e.g. the
meaning of the canonical idiom kick the bucket (“to die”) cannot be
derived from the meanings of its constituents, i.e. “kick,” “the,” and
“bucket”); thus, idioms are thought to be arbitrarily stipulated in memory
(cf. Chomsky 1980; Cruse 1986; Fraser 1970; and Katz 1973; among
others). In the generative literature, idiom chunks are not associated with
meanings, and it is said that “there is no relation between the meanings of
the parts and the meaning of the whole from the viewpoint of
“synchronic” structure” (Kiparsky 1976:79). Chomsky (1980: 149) names
idioms “non-compositional” structures; Machonis (1985: 306) defines
them as “frozen” expressions, “not predictable from the composition;”
while van der Linden (1992) follows Katz and Postal (1963: 275) in
172 Chapter Four
claiming that the meaning of an idiom “is not a function of the meanings
of its parts and the way these are syntactically combined; that is, an idiom
is a noncompositional expression” (van der Linden 1992: 223). Consequently,
since there is no relation between the linguistic forms of such idioms and
their meanings, and due to their lack of internal syntactic or semantic
structure, hardly any modification is possible in the case of this type of
“frozen” idioms (called idiomatic phrases, after Nunberg et al. 1994).
However, this standard view of idioms as non-compositional strings or
long words has been objected to in the current research. It has been
demonstrated that not all idioms are “frozen,” and as opaque or
unanalysable as kick the bucket. For most idioms (called idiomatically
combining expressions, after Nunberg et al. 1994), some relations between
their meaning and form can be established. In fact, the meanings of
particular components of idioms do play a role in the way idioms are used
and understood (cf. Cacciari 1993; Cacciari and Glucksberg 1991; Gibbs
1990; Keysar and Bly 1995, 1999). What is more, this type of idioms can
be syntactically transformed in various ways and their parts can be
modified, while some elements within an idiomatic expression can be
semantically productive (cf. Makkai 1972; Nunberg et al. 1994; O’Grady
1998; and Ifill 2002; among others). Thus, a question arises of what sort of
syntactic modifications can compositional idioms undergo.
Following this second research trend, attention is paid here to the
linguistic side of the distinction between so-called “decomposable vs. non-
decomposable idioms” (Nunberg 1978). Taking this division into
consideration, the purpose of Chapter Four is to discuss and analyse
idioms, with a special focus laid on the psychological idioms elicited in
Chapter Three, with the aim of gaining better theoretical and empirical
insights into syntactic rules they are governed by. More precisely, issues
that are to be addressed here include the following: (i) semantic properties
of idioms, with special reference made to psych-idioms, and a distinction
between idiomatically combining expressions and idiomatic phrases
(section 4.2); (ii) syntactic and semantic flexibility of psychological
idiomatically combining expressions (section 4.3); (iii) syntactic
constraints on psychological idiomatic phrases and the structure of idioms
analysed in previous accounts (section 4.4) and within the Phase Theory
(section 4.5); and (iv) the position of an Experiencer in psychological
idioms (section 4.6). The chapter ends with an overview of the aspectual
properties of psychological idioms (section 4.7), followed by the
concluding remarks, outlined in section 4.8.
Psychological Idioms: Syntactic Constraints and Aspectual Properties 173
4.2.1.1 Conventionality
1
The meanings provided for the idioms refer to psych-verbs corresponding to the
idiomatic phrases. In fact, the exact meanings may differ in the degree from these
basic psychological domains they belong to.
2
The etymological background and the information about the cultural /
conventional origins of idioms have been retrieved from http://www.phrases.
org.uk/meanings/118400.html, in March, 2017.
Psychological Idioms: Syntactic Constraints and Aspectual Properties 175
4.2.1.2 Inflexibility
3
The sign * [asterisk], as explained earlier, will be used to mark the sentence /
phrase as ungrammatical, while # [hash] indicates the loss of idiomatical
interpretation.
Psychological Idioms: Syntactic Constraints and Aspectual Properties 177
of its parts. Instead, idioms are viewed to be listed as phrasal entries in the
lexicon, associating each entry directly with a single semantic representation.
Furthermore, once an idiom is recognised, we are able to establish
correspondences between the parts of the structured denotation of the
phrase, e.g. the relation of making someone dispirited and depressed in the
idiom give Y the blues in (4.1f), and the components of this idiom (give
and the blues), in such a way that each constituent is to be predictable in
terms of metaphorical reference to an element of the interpretation. Thus,
the idiom is given a compositional, i.e. idiosyncratic, analysis (cf. Nunberg
et al. 1994: 496).
In addition, a lot of experimental research has been done concerning
the variability in the degree of idiomatic compositionality (analysability).
As a result, it has been shown that people appear to have strong intuitions
enabling them to judge an idiom as being decomposable or non-
decomposable (Gibbs and Nayak 1989). It can be easily identified how
each part of the idiom combines in the derivation of the figurative
interpretation for decomposable / compositional idioms (e.g. pull strings,
cast a spell on Y, get cold feet), but not for non-decomposable / non-
compositional idioms (e.g. kick the bucket, raise the roof, paint the town
(red)).
While analysability defines the range to which idiom components
contribute to the overall idiomatic interpretation; transparency refers to
the relative ease with which any assumptions about the meaning of an
idiom are accessed and implications derived. Idiom transparency is
sometimes called “literalness” of an idiom, and is also identified by some
scholars as the degree to which the meaning of the idiom can be derived
from the phrase. Thus, analysability and transparency are so closely
related that they are mostly used interchangeably.
According to CieĞlicka (2004: 98), Mäntylä (2004: 28-29), and
Peacock (2009: 2), among others, there are directly / fully transparent
idioms, for which the literal meaning of their constituents is clearly linked
to the figurative overall meaning, e.g. give the green light “to accept
something.” These idioms display a high degree of analysability and
transparency. However, in the case of semi-/ partially / relatively
transparent idioms, the literal meaning of their components gives only
some hint of the figurative meaning (e.g. quake in your shoes “to fear”)
but the link is not as noticeable as with fully transparent idioms. Finally,
there are (transparently) opaque (semi-opaque) idioms, in which the
motivation behind the figurative meaning is hard to perceive unless the
etymology is already known, e.g. be home and dry “to succeed.”
Psychological Idioms: Syntactic Constraints and Aspectual Properties 179
meaning of the phrase, the more transparent the idiom seems to be, and the
easier the idiomatic interpretation will be obtained. For example, in the
idiom spill the beans “to divulge a secret,” spill directly implies “divulge,”
and the word the beans refers to the noun “a secret,” as a result of the
long-lasting convention concerning this expression. McGlone, Glucksberg
and Cacciari (1994) propose that the idiom spill the beans is more
transparent than an alternative expression spill the mud, which may stem
from the fact that the beans make a great contribution to the idiom
meaning in such a way that beans, like “secrets,” are many and countable.
Finally, it is decomposable / compositional idioms which are more
flexible (syntactically, semantically, and lexically) than non-decomposable
ones. As argued by Gibbs and Nayak (1989), Gibbs et al. (1989a,b),
Glucksberg (1993, 2001), among others, the more analysable and
transparent an idiom is, the more syntactic flexibility it is expected to
reveal, which is to be discussed in section 4.3 below.
4
The online version of The Merriam-Webster Collegiate Dictionary, available at
https://www.mer riam-webster.com/ was retrieved on March 11, 2017.
Psychological Idioms: Syntactic Constraints and Aspectual Properties 181
5
Online International Dictionary available at http://idict.org, retrieved on March
11, 2017.
182 Chapter Four
Indeed, as seen in (4.6), to eat Y’s heart out is an exaggerated way to mean
that someone is worried. In the same vein, trying to express somebody’s
annoyance the idiom cut Y to the quick / bone reveals someone’s psychical
and emotional condition clearly enough. To mean that someone or
something scares or horrifies a person, the idioms curdle / chill Y’s blood
or put the screws on Y may be used, and the emotional state is almost
visibly exposed.
6
Retrieved from American Heritage Dictionary of the English Language,
available at https://www. ahdictionary.com/word/search.html?q=online on March
11, 2017.
Psychological Idioms: Syntactic Constraints and Aspectual Properties 183
f. the social problem of anger, viz. filling someone with anger; expressed
in the following idioms, e.g.:
kindle Y’s wrath
get Y’s back / dander up
put / set Y’s back (up)
blow a gasket on Y
breathe fire over Y
drive Y out of mind
drive Y to distraction
drive Y round the bend/ twist
vent X’s spleen at Y
7
The online version of The Merriam-Webster Collegiate Dictionary, available at
https://www.merriam-webster.com/ retrieved on April 12, 2017.
8
Importantly, as commented by Prof BoĪena Cetnarowska, some idioms are
currently used less frequently, e.g. in the Oxford English Dictionary (online
version https://en.oxforddictionaries.com), the idiomatic phrase give someone a pip
is marked as “dated,” while give someone a turn and pay court to someone are
noted by the Longman Dictionary (http://www.ldoceon line.com/dictionary) as
“old fashioned.”
186 Chapter Four
9
Literary Devices Editors. 2013. Colloquialism. Retrieved April 12, 2017, from
https://literarydevices. net/ colloquialism/
Psychological Idioms: Syntactic Constraints and Aspectual Properties 187
As can be seen in (4.8), the most common registers that idioms are likely
to occur in are newspapers, magazines, fiction, and spoken discourse.
Academic sources, in turn, record the lowest scores of the idioms under
scrutiny. In fact, newspapers, magazines, and fiction are not taken into
consideration, since there is no clear-cut distinction between the sources
written in colloquial language and the ones which contain more scientific
and formal language. In what follows, having compared the spoken
discourse to the academic text the psychological idioms occur in, the
superiority of the former over the latter type of register may be evidently
noticed. That is why, the assumption that informality is one of the features
of idiomaticity, including psychological idioms, has been confirmed.
10
The online version of The Merriam-Webster Collegiate Dictionary, available at
https://www.mer riam-webster.com/ retrieved on April 12, 2017.
11
Online Etymology Dictionary, available at http://www.etymonline.com retrieved
on April 12, 2017.
188 Chapter Four
12
Interestingly, it should be repeated here, for the sake of convenience that, taking
into account the relation between the form of an idiom and its meaning, Gibbs and
Nayak (1989), and Nunberg (1978), among others, distinguish three types of
idioms. First, when a one-to-one relation occurs, i.e. a relation in which each word
contributes independently to the figurative interpretation (e.g. the semantic relation
between “pop” and “utter” and “question” and “marriage proposal” in pop the
question), the idioms are typically known as “normally decomposable” idioms.
Second, they mention “abnormally decomposable” idioms, in which there is an all-
to-one relation with the (literal) meaning of the whole phrase being semantically
connected with the figurative interpretation (e.g. bury the hatchet, and push the
panic button). Finally, in case of “non- decomposable” idioms, the relation may be
none-to-one in that the component words neither individually nor as a whole are in
a semantic relation to the idiomatic meaning (e.g. chew the fat, and break a leg).
Moreover, the relation may differ in terms of transparency between the constituent
192 Chapter Four
explicit. For Wasow et al. (1984), Fillmore et al. (1988), Nunberg et al.
(1994), Everaert et al. (1995), Harwood et al. (2016), and Corver et al.
(2017), the extent to which idioms can be syntactically and lexically
modified accounts for the difference between IdPs and ICEs. Especially
the possibility for an idiom to undergo passivization, topicalization, and
adjectival modification, while leaving the idiomatic interpretation intact,
are the most commonly cited diagnostics to distinguish IdPs from ICEs, as
these tests produce the most categorical results. The tests on the level of
conventionality, compositionality or opacity of an idiom are, in turn, far
more ambiguous and a matter of one’s individual interpretation (cf. Corver
et al. 2017: 12). Consequently, IdPs are expected to remain completely
resistant to any syntactic or lexical modifications, in contradistinction to
ICEs, which are to show a considerable degree of syntactic and lexical
flexibility (cf. Nunberg et al. 1994; and Gibbs et al. 1995; among others).
The issue of syntactic and lexical variability of idioms is to be discussed
more thoroughly in section 4.3 of the chapter.
13
There have been 20 native speakers who have been asked for their judgements.
Two of them (10%) were aged 20-25, the other two (10%) were in their 30s-40s,
whereas the remaining sixteen (80%) were aged 60-70. All of them either teach
English presently or used to do so before they retired.
194 Chapter Four
14
For further discussion concerning anaphora see Nunberg et al. (1994: 501-503).
Psychological Idioms: Syntactic Constraints and Aspectual Properties 195
15
The expression may be also interpreted literally “to get cold,” referring to a
bodily sensation, but in our analysis of the phrase only its figurative interpretation,
viz. “to worry,” is taken into consideration.
196 Chapter Four
but, instead, it denotes a figurative reading of the idiom, which means here
“annoying the conservative part of the society.” In (4.16b), the adjective
total does not describe the shit, but, in turn, it builds the overall idiomatic
interpretation of “being scared totally.” Similarly in (4.16c-d), a sudden
liking and a horrifying shadow refer to the context in which the
expressions should be interpreted; viz. “Suddenly, she took a liking,” and
“Horrifyingly, the TB-ridden slums cast a shadow.” In brief, external
modification exists outside of the domain of idioms. Thus, examples like
(4.16a-d), allow an interpretation where the NP-internal modifier does not
work as restricting the reference of the nominal constituent, but, instead,
as an operator taking the nominal within its scope (cf. Nunberg et al. 1994:
500).
In (4.17a-d), there are some instances of the internal adjectival
modification of the object NP, attested for psychological idioms.
16
As given by American Heritage Idioms Dictionary (2002). Retrieved from
http://www.dictionary. com/browse/blue-funk--in-a
200 Chapter Four
(1978), Nunberg et al. (1994), Folli and Harley (2007), and Stone (2013),
among many others. For decades, many theoretical proposals have been
offered to account for the distinction between idioms which can be
passivized, while still retaining their idiomatic meaning in the passive (e.g.
The beans were spilled “The secret was divulged”), and idioms which are
non-passivizable since their idiomatic interpretation is then lost (e.g. #The
bucket was kicked receives only a literal interpretation).
One of the most widely-recognised accounts for (lack of) passivization
of certain idioms is the one discussed in Nunberg et al. (1994). They argue
that compositionality of idioms is the key factor for an idiom to be
passivized. Due to semantic non-compositionality of kick the bucket “to
die,” which has its idiomatic meaning distributed over the entire phrase,
the idiom cannot passivize and still retain its idiomatic meaning. Whereas
a semantically compositional idiom spill the beans “to divulge a secret” is
passivizable, because the meaning of the idiom is shared among the
constituent elements of the idiom; as spill denotes “to divulge” and the
beans refer to “a secret.”17 In addition, Folli and Harley (2007), and Stone
(2008, 2013) offer a syntax-based account for distinguishing passivizable
and non-passivizable idioms.18
There is a fair number of psychological idiomatically combining
expressions which retain their idiomatic interpretation in the passive, as in
(4.19a-d) below, while some of them, when passivized, sound
grammatically unacceptable, as in (4.19e).
17
Cf. McGinnis’s (2002) counterarguments against Nunberg et al.’s (1994)
account of passivization. For McGinnis (2002), aspect plays a significant role
since it is compositional even in non-passivizable idioms.
18
Cf. Stone (2013: 4-5) for a more detailed analysis.
202 Chapter Four
19
Cf. Petersen’s (2016) comment on raising out of some idioms, which, if
modified this way, lose their idiomatic meaning, as in (i).
(i) O kombos fenete oti exi ftasi sto xteni (Greek)
the knot seem-3SG that have-3SG reached to-the comb
“The knot seems to have reached the comb.”
Idiomatic reading: #”Things seem to have come to an end.”
(Petersen 2016: 248)
Petersen (2016: 247-248) claims that (parts of) an idiom may not carry discourse
effects, and their raising, as an instance of A-movement, renders an idiomatic
interpretation infelicitous.
Psychological Idioms: Syntactic Constraints and Aspectual Properties 203
Van der Linden (1991: 27) claims that the control-construction is hard to
test for idioms.20
20
Interestingly, control has been thoroughly studied in the Government and
Binding framework in the 1980s, while in the days of Transformational Grammar,
it has been referred to in terms of Equi-NP deletion. Since control is said to be only
applicable to meaningful expressions; thus, an expression (the controller) is related
to an abstract pronominal (the controllee), and both the controller and the
controllee have to be meaningful expressions (cf. Schenk 1995: 260-261). Idiom
chunks cannot become controllers, as illustrated in the following examples:
(i) spill the beans ĺ “to reveal a secret”
a. Pete instructs John to spill the beans.
b. John tries to spill the beans.
c. *John instructs the beans to be spilled.
d. *The beans try to be spilled.
(van der Linden 1991: 27)
(ii) paint the town ĺ “to enjoy”
a. Mary instructs James to paint the town.
b. James tries to paint the town.
204 Chapter Four
changed into different ones, viz. cast ĺ hang over, and dampen ĺ ravish,
the sentences get improved, but the psychological idiomatic meaning is
lost, even though some figurativeness of the sentences is still retained.
Moreover, the instance illustrated in (4.25d) is an example of Inverted-
clefts, which is completely licit as regards grammar, but its meaning is
ambiguous between the idiomatic one, viz. give Y the blues “to depress
someone,” and the literal interpretation “to provide somebody with this
specific kind of music, i.e. blues.” Likewise, the idiomatic meaning is lost
for most of the Pseudo-cleft structures, illustrated in (4.25e-h). Only the
cleft examples with the idioms show contempt for X “to hate” (4.25e), and
have / get goose bumps “to fear” (4.25f), seem to be acceptable and retain
their idiomatic interpretation, but native speakers assess these structures as
rarely used in a real life language.
To sum up, the cleft structures, provided in (4.25b-h), seem to capture
the fact that the more transparent and compositional an idiom is, the easier
it forms the cleft structure. Furthermore, Gramley and Pátzold (1992)
propose that some of the reasons why certain idioms do or do not allow
transformations seem to be idiosyncratic; while for others the semantic
reasons can be given. Therefore, most idioms usually resist the isolation of
one formative for emphasis, as in the case of cleft structures (*It was her
throat that he jumped down, jump down someone’s throat “to respond to
what someone has said in a sudden and angrily critical way”), because in
this operation word forms are treated as semantic constituents, which they
are not. Throat in he jumped down her throat has no isolable meaning in
the idiom, and that is why it cannot be modified (cf. Gramley and Pátzold
1992: 57).21
21
Reeve (2010) provides some remarkable account for possible or disallowed cleft
structures. He argues that VO idioms, such as keep track and make headway, must
“be base-generated as a constituent.” This contributes to the fact that the idiomatic
object is dependent on the verb. However, in some cases, it is possible for the
object to undergo movement and still retain its idiomatic meaning, e.g. What kind
of track was she keeping t of her expenses?, or The careful track that she’s keeping
t of her expenses pleases me (cf. Carlson 1977, cited in Reeve 2010: 65). Whereas,
other cleft variations are found unacceptable, e.g. *What she is keeping of her
expenses is CAREFUL TRACK (den Dikken et al. 2000, cited in Reeve 2010: 65),
or * CAREFUL TRACK is what she is keeping of her expenses (cf. Reeve 2010:
65-66).
210 Chapter Four
only to Jason are the free slots of the idioms, which, when shifted into the
topic position at the beginning of the sentence, do not result in breaking
the core meaningful structure of these idioms. Therefore, such
constructions are acceptable, although judged by native speakers as quite
rare. A more detailed analysis of topicalization is provided in Wasow et al.
(1984), Gazdar et al. (1985), and Osborne et al. (2012), among others.
The example in (4.27a) shows object incorporation with give Y a turn “to
horrify,” which appears as the adjectival expression turn-giving. Whereas,
in (4.27b) the adjectival compound blood chilling derives from the idiom
chill Y’s blood “to scare,” and it comprises the idiomatic NP object blood
placed pre-verbally, before the idiomatic verb chill. The determiner of the
NP, represented in (4.27b) by a genitive possessor, which reveals the
identity of the Experiencer Y, must be omitted in this type of modification.
212 Chapter Four
Things that drive me crazy might not bother you, while something
that has you climbing the walls might hardly ruffle my feathers.
(Google)
22
While discussing extendibility, Egan (2008) provides the following example:
(i) Speaker A: I hear Mr. Jones kicked the bucket.
Speaker B: Yeah. He almost connected yesterday; today he really put the
boot on it.
(kick the bucket; Egan 2008: 393).
In Egan’s (2008) example above, extendibility means that once Speaker A, in (i).
has invoked the idiom kick the bucket to mean that Mr. Jones has died, Speaker B
relates to the same semantic space of death, using terms connected, and boot.
23
Egan (2008) argues that the line between extendibility (and perhaps even
figurative modification) and wordplay is tenuous, and no dividing line of this kind
should be drawn.
214 Chapter Four
24
It is worth recalling here, for the sake of convenience, what has been introduced
in Chapter Three of the book. Specifically, all psychological idioms are not
provided with detailed meanings they have, but classified into the general
psychological domains, directly related to psych-verbs. In fact, the interpretation of
every single idiom differs, e.g. both the idiom carry a torch for X “to secretly love
someone who does not love you,” and set store by X “to regard as valuable or
worthwhile, worthy to be loved” are classified into the same semantic domain of
LOVE, and the psych-verb related to these idioms is “to love.”
216 Chapter Four
25
These numbers are based on the corpus study, presented in Chapter Three.
Psychological Idioms: Syntactic Constraints and Aspectual Properties 217
dash / wither Y’s hopes; rattle Y’s cage / chain; or get Y’s dander /
hacklers / Irish up.
To sum up, in section 4.3, the issue of syntactic and lexical variability
of ICEs (on the example of psychological idioms) has been discussed.
Quantification, topicalization, ellipsis, and anaphora, among others, are the
numerous examples of variation forms idiomatically combining
expressions can occur in. As exemplified by psychological idioms, ICEs
tend to exhibit also a degree of lexical substitution, due to their more
compositional nature. IdPs, instead, are usually entirely resistant to any
such alterations, as has been discussed in section 4.2.1.2, and exemplified
by inflexibility of psychological IdPs in (4.10). Table 4-1 summarises the
differences between IdPs and ICEs. Compositionality
Conventionality
Flexibility
Variation
Syntactic
Opacity
Lexical
Open
Slots
Idiom
Non-
IdP High Opaque No No No
compositional
Trans- More
ICE Low Yes Yes Yes
parent compositional
To begin with, it has been Nunberg, Sag and Wasow (1994) who, in their
seminal work, distinguish idiomatic phrases (IdPs), e.g. kick the bucket “to
die,” or raise the roof “to enjoy,” from idiomatically combining
expressions (ICEs), e.g. spill the beans “to reveal a secret,” or give Y a
fright “to scare.” This distinction, adopted for the sake of the book, and
analysed thoroughly in section 4.2.2, seems to be useful to understand the
behaviour of the so-called non-compositional and fixed idioms, on the one
hand, and compositional and flexible ones, on the other.
What is more, while analysing idioms, generative linguists have never
dealt satisfactorily with idiomatic compositionality, productivity and
syntactic variance, which has led them to treating idioms as “extra-
grammatical,” placed at the periphery of grammar. Nunberg et al. (1994),
instead, argue that certain components of an idiom can be assigned an
interpretation, and that “modification, quantification, topicalization,
220 Chapter Four
ellipsis, and anaphora provide powerful evidence that the pieces of many
idioms have identifiable meanings which interact semantically with other”
(Nunberg et al. 1994: 503). Contradicting this way a well-established
assumptions made in generative grammar, Nunberg et al. (1994: 503)
conclude that the meanings of idiom chunks are not their literal meanings,
but, instead, idiomatic meanings are largely derived from literal meanings
in the conventionalized, but not entirely arbitrary, manner. As a result,
conventionality should not be identified with non-compositionality, and
differences in syntactic flexibility among idioms can be explained by
means of the compatibility of semantics with the semantics and pragmatics
of various constructions (cf. Nunberg et al.’s (1994: 504-505) postulation
of the existence of homomorphic families of idioms, discussed in section
4.3.3).
Furthermore, Nunberg et al. (1994) comment on Marantz’s (1984) and
Kiparsky’s (1987) generalization concerning the rarity of idiomatic Agents
and Goals in idioms, providing some additional argumentation about why
idioms contain far fewer animate NPs than inanimate ones in general.
Nunberg et al. (1994) argue that, in normal discourse, verbs mostly take
animate objects, while hardly any animate objects are used in idiomatic
expressions (e.g. hit the ceiling “to get very angry,” hit rock bottom “to
reach the lowest point,” kiss the dust “to fall down due to being shot / hit,”
or kiss the cup “to drink”). Indeed, this assumption seems to be true when
considering the case of psychological idioms under scrutiny. Truly, out of
161 English idioms related to a psychological condition, which have been
elicited in Chapter Three, there is only one instance including a literally
animate NP, i.e. get Y’s Irish up “to annoy,” providing the nationality
adjective Irish, used in this idiom, is an NP. Moreover, Nunberg et al.
(1994) assume that the predisposition of metaphorical mappings to shift
from concrete to abstract in idioms does not offer a full explanation of the
extreme rarity of concrete / animate idiomatic meanings in idioms. Such
metaphorical reference to concrete things, actions, or situations does not
commonly occur.
Likewise, according to Nunberg et al. (1994), it is relatively rare for an
idiom to feature a fixed Goal or Possessor argument. This also follows
from their generalization concerning the scarcity of fixed Agent
arguments, since Goals and Possessors tend to be animate. Nunberg et
al.’s (1994) assumption is confirmed by psychological idioms. In the set of
161 psychological idioms, both idioms with a possessor (get Y’s goat;
raise Y’s hackles “to annoy”) and a fixed Goal (tickle Y to death “to
amuse”) comprise fewer than one fourth of all the dataset.
Psychological Idioms: Syntactic Constraints and Aspectual Properties 221
26
It is significant to add that 3 years later after Nunberg et al.’s (1994) seminal
works, Jackendoff (1997) in his framework of Representational Modularity (RM),
postulates that both syntactic and lexical conceptual structures (LCS, Jackendoff
1997: 49) are involved in the production of meaning. Jackendoff (1997, 2002)
argues that idioms have phonological structure, syntactic structure, and conceptual
structure, but due to the fact that not all of the syntactic constituents of an idiom
correspond to conceptual constituents, the idiom, in fact, may give rise to idiomatic
interpretation. Most idiomatic expressions do not have compositional meaning;
thus, they comprise complex lexical items whose meaning is not syntactically
analysed but rather determined by the syntactic-conceptual structure interface
component. In short, in his representational modularity approach, Jackendoff
(1997) postulates the lexical licensing of units larger than Xº.
27
A chain is identified by O’Grady (1998: 284) in the sense that “iff x [in the
string x…y…z…] licenses y and z, or if x licenses y and y licenses z.”
222 Chapter Four
to be illustrated via a tree structure, the continuity chain will go down the
tree, from heads down to their dependents.
Let us take all the syntactically different patterns of the psychological
idioms, elicited in Chapter Three, to examine how they behave with
respect to the Continuity Constraint.The twelve types, which 161
psychological idioms under scrutiny represent, show the patterns of chains
specified in (4.35) below.
(4.35) The patterns of chains for the psychological idioms (cf. O’Grady
1998):
Psychological Idioms: Syntactic Constraints and Aspectual Properties 223
the role of a complement of the P. In four types of the idioms analysed, the
chain is broken due to the fact that Y (the Experiencer) or X (the Theme)
occurs in between the verb and other dependents and interrupt the chain
relation between them. For the four types of idioms, listed in (4.36) below,
the Continuity Constraint does not hold. Thus, they cannot be captured in
O’Grady’s (1998) model.
Table 3-4 in Chapter Three shows that the number of OE (class II)
psychological idioms, occurring in the aforementioned illicit idioms is
extremely big for Type (C), quite big as for Types (E) and (E’), while
Type (D’) is marginal. Therefore, based on the chain-like patterns of
psychological idioms, analysed in (4.35), it can be concluded that the
chain cannot replace all syntactic structures ever possible, but, instead, it
works as a bare minimum constraint on the very structure of an idiom.
Nonetheless, the Continuity Constraint seems to be applicable to most
cases.
All in all, in O’Grady’s (1998) Continuity Constraint, the lexical
choices sound explicitly fixed within idiomatic constructions, i.e. the
selection restrictions are based on specific Head-to-Head relationships.
The Continuity Constraint accurately specifies the organization of existing
idioms, including non-constituent idioms, simultaneously predicting that
certain types of patterns are impossible. Additionally, it sheds new light on
the relevance of argument structure for idiom formation, pointing to some
asymmetries in the composition of idioms, making space, this way, for
thematic hierarchy effects. Finally, non-idiomatic open slots are placed by
Psychological Idioms: Syntactic Constraints and Aspectual Properties 225
O’Grady (1998) at the very end of chains, just to preserve the chain from
breaking. Instead, there should be more possibilities for idiomatic patterns
with open positions. Therefore, the existence of psychological idioms with
open slots realised as NP objects, makes the Continuity Constraint flawed
and, thus, worth revising.
What is more, developing Kiparsky’s (1987) and Baker’s (1989)
Thematic Hierarchies, reproduced in (4.37a) and in (4.37b) respectively,
O’Grady (1998) formulates the Hierarchy Constraint, as in (4.38).
28
As summarised by Landau (2003: 2-3), T, v [little v], and C are core functional
categories. While v expresses transitivity, selects V, has ij-features (object
agreement), selects external argument, has optional EPP feature (second Merge)
for object shift; T expresses tense/event structure, has ij-features (subject
agreement), obligatory EPP feature. If T is selected by C, it has a full set of ij-
features, whether expressed (finite) or not (control). If selected by V
(raising/ECM), it has only a subset of ij- features. C, in turn, expresses force/mood,
has ij-features, and an optional EPP feature (for wh-phrases).
Psychological Idioms: Syntactic Constraints and Aspectual Properties 229
Besides, “v* is the functional head associated with full argument structure,
transitive and experiencer constructions, and is one of several choices for v, which
may furthermore be the element determining that the selected root is verbal”
(Chomsky 2008: 142).
What is more, a sentence is usually said to comprise two phases, namely the
complementizer C (CP) and the light verb v (vP). Chomsky (2000) suggests that if
the light verb v represents a transitive verb, then it is ij [phi]-complete. Likewise,
finite and control C are recognised as ij-complete. Then, transitive vPs (ij-
complete v), and finite CPs and control CPs (with PRO subjects), represent strong
phases.
29
Chomsky (2000, 2001) argues that phases display some degree of phonological
and syntactic or LF independence. While semantic or LF independence is marked
by the fact that the edges of phases are reconstruction sites for quantifier and
operator movement; phonological independence is recognised by the phonological
isolation of phases.
30
However, Citko (2014) admits that it is difficult to explain the reason why the
argument structure of unaccusative or passive verbs is less complete than the
argument structure of transitive verbs. Likewise, unergative verbs may be treated
as incomplete in terms of their argument structure, when they lack an internal
argument (John ran vs. John ran a race), or many transitive verbs may also be
found incomplete in terms of their argument structure, because they form double
object constructions (John baked a cake vs. John baked Mary a cake) (cf. Citko
2014: 29-30; and Epstein 2007).
230 Chapter Four
As can be seen in (4.41b), there are only two types of features visible to
the computational system (valued interpretable features iF[val], and
unvalued uninterpretable features uF[ ]). Val stands for any feature value
complex, while empty brackets signify the lack of value, and filled
brackets signify valued features. A feature is interpretable if it can be
interpreted at the interface level LF, making direct contributions to
meaning, while a feature is uninterpretable if it cannot be interpreted at the
interface level LF. In short, feature interpretability is determined in the
lexicon, being maintained throughout the derivation, including the LF-
branch of the derivation (Chomsky 2001: 6). The examples of interpretable
features include the ij-features of nouns (DPs) and the tense feature of T;
whereas uninterpretable features comprise, e.g. the ij-features of T and v,
the tense feature of V and the Case feature of DPs (cf. Willim 2012: 764).
Importantly, uninterpretable features have to be valued in the course of the
derivation. Valuing unvalued features occurs via the mechanism called
Agree. Unvalued features need to be deleted by the time of Transfer to the
interfaces, because uninterpretable features cannot be interpreted by the
interfaces.
Furthermore, within the Minimalist Program, it is uninterpretable ij-
features which are supposed to trigger computational operations (agreement
and movement). Moreover, only phase heads are said to have
31
Cf. Pesetsky and Torrego (2007), and Boškoviü (2011), who suggest
distinguishing other combinations, i.e. uF [val] an uninterpretable and valued
feature, and iF [ ] an interpretable and unvalued feature.
Psychological Idioms: Syntactic Constraints and Aspectual Properties 231
(4.44)
Movement out of the phase proceeds through the phase edge, and a
constituent is only permitted to move out of a phase (the Spell-Out
domain), providing the constituent has first moved to the phase “edge” (cf.
Citko 2014: 32). Hence, Chomsky (2000: 109) claims that phase heads
have the requisite features to trigger movement.
Furthermore, even though the generally cited form of the PIC is the
one reproduced in (4.43), various versions of the PIC have been proposed
in the literature. Indeed, all of them refer to the same general
configuration, presented in (4.45a) below, in which Z and H are phase
heads, while X is a non-phase head placed between them. A clausal
structure of this configuration is shown in (4.45b), in which C and v are
phase heads, while T is not.
operations within the smallest strong ZP phase, but not beyond. In other
words, elements inside the strong phase ZP can target the edge Į of HP.
Once the derivation reaches ZP, the complement of Z is sent to Spell-Out,
and it remains impenetrable (inaccessible) for further computation. The
domain of H (here, YP) is not accessible to operations at ZP, but only H
and its edge.
Besides, the successive-cyclic movement invariably targets the edge of
cyclic domains. The fact which stems from the cyclic determination of
strong phases and the Phase Impenetrability Condition is that accessibility
of the edge of a strong phase is only up to the next strong phase. Making a
supposition that the Spell-Out operation sends a syntactic object to both
PF and LF, the appropriate generalization can be formulated as in (4.46)
below, following Chomsky (2001).
In (4.46), Phase 1 is strong and Phase 2 is the next higher strong phase,
where a strong phase is a CP or a vP.
In a nutshell, as noted by Radford (2000), Chomsky’s (2000) Phase
Impenetrability Condition32 is a natural consequence of the locality
constraint, referring to the phase boundary of syntactic and phonological
operations. Hence, the domain of a subordinate (strong) phase (i.e. vP, CP
or possibly DP) is not penetrable to the head of an immediately
superordinate phase. Syntactic movement operations, according to the
assumptions made within the Phase Theory, involve two sub-operations of
copying and deletion. Thus, the constituent which is moved is first copied
into the position to which it moves, and then the original one may be
32
As argued by Citko (2014: 33), Müller (2004), Richards (2011), among others,
there are at least two versions of the Phase Impenetrability Condition, i.e. (i) one
from Chomsky’s (2000) “Minimalist Inquiries,” generally referred to as “a Strong
PIC / PIC1;” and (ii) the other from Chomsky’s (2001) “Derivation by Phase”
(referred to as “a weak PIC / PIC 2”). They are reproduced below:
(i) PIC1: The domain of H is not accessible to operations outside HP; only H and
its edge are accessible to such operations.
(ii) PIC2:The domain of H is not accessible to operations at ZP; only H and its
edge are accessible to such operations.
(Chomsky 2001: 13-14)
234 Chapter Four
deleted, viz. given a null phonetic Spell-out. It is also typical of the Phase
Theory that “Spell-Out is cyclic, at the phase level” (Chomsky 2001: 9).33
33
In the Minimalist Program the thematic roles are assigned as a result of the
operation Merge. These roles are mapped according to the hierarchy of thematic
roles (cf. Kiparsky 1987; Baker 1989; and O’Grady 1998; among others) and
according to the UTAH (Baker 1988). Psych verbs do not satisfy the latter
requirement.
Psychological Idioms: Syntactic Constraints and Aspectual Properties 235
(4.48)
34
Cf. Jackendoff (1997), Marantz (2001), Svenonius (2005), and Harwood et al.
(2016), among others, who provide various proposals concerning how idiomatic
meanings are primed at the syntax-semantics interface.
236 Chapter Four
(4.50)
Moreover, the nouns in IdPs (e.g. the bucket, or the roof) do not denote
generic and unique referents, but, instead, are “non-denoting nouns” for
Fellbaum (1993), or “no identifiable idiomatic referents” for Grégoire
(2009). Hence, the nominal expression (NP) in an idiomatic phrase (IdP) is
non-referential, and it cannot be spelled out self-sufficiently, but always as
a part of the vP-phase, to evoke the idiomatic non-literal interpretation.
Following the account provided by Harwood et al. (2016), a syntactic
representation for Type (A): V+ NP psychological idiomatic phrases,
listed in (4.49) can be as in (4.51) below.
(4.51)
As can be seen in the tree structure in (4.51), the nominal expression the
roof cannot become an argument, i.e. receive the theta-role, as in the case
of all IdPs, because only a referential category, i.e. only DPs, can occur in
argument positions (Stowell 1991; Longobardi 1994; and Harwood et al.
2016; among others). Likewise, the determiner of the direct object in IdPs
lacks a direct referent in the discourse, and thus it does not project. Since
there is no DP-phase in idiomatic phrases (IdPs), there is no DP-phase
boundary to be straddled by IdPs. Instead, the direct object of an idiomatic
phrase directly forms part of the vP-phase, rather than projecting its own
phasal domain (cf. Harwood et al. 2016: 2, 5).
Nonetheless, among the instances of psychological idiomatic phrases
(IdPs) listed in (4.49) above, there are some idioms which lack the definite
determiner, in contradistinction to the canonical example kick the bucket
“to die,” or raise the roof “to enjoy.” Instead, an NP is preceded by an
indefinite article (e.g. have a ball “to enjoy”), a null article [ø] (e.g. turn
Psychological Idioms: Syntactic Constraints and Aspectual Properties 239
tail “to fear”), a fixed adjective (e.g. show the white feather “to fear”),
takes a plural form (e.g. get pins and needles “to fear”), or is coordinated
with another NP (e.g. have / get pins and needles “to fear”). This fact does
not bring any consequences for the syntactic structure of this type of
psychological idioms, since all the idiom constituents are, anyway,
expected to be confined to a single transitive vP-phase, which can be
represented by means of the structure, as in (4.52a-b) below.
35
The existence of the so-called “little-n” (by analogy with “little-v”) is justifiable
in the structure of complex head-initial nominal phrases. Even though an NP does
not need a theta-assigner for its external argument since there is no external
argument; the shell structure of complex head-initial NPs is the same type of shell
structure as that of complex head-initial VPs (cf. Haider 2012: 55). Introducing n,
Chomsky (2007) draws a parallel between VPs and NPs; thus, n* is like v* and D
is like N.
240 Chapter Four
referential properties with indefinites (cf. the discussion above), and thus
lack n* in their structure the same way indefinites do. Therefore, Chomsky
(2007) makes a proposal that only definite nominal phrases (which are in
fact n*Ps in this view) constitute a phase, while indefinite nominals are not
phases. Similarly, Adger (2003), Radford (2004), Hinzen (2012), Corver
et al. (2017), and Harwood et al. (2017) state that indefinite DPs are not
phases, but only complete referential expressions are phasal, and the
instances of idioms in (4.49) and (4.52) can be treated as “complete
referential expressions,” with an assigned idiomatic interpretation only if
their constituent elements (V+NP) form a single unit. In other words, the
object of these psychological IdPs will not be spelled out independently,
but only as part of the vP-phase.36
All in all, with the overall idiomatic interpretation not derived from
their constituent parts, IdPs do form single semantic units, necessarily
interpreted as a whole. Therefore, it is highly justifiable to treat fixed IdPs
as restricted to a single vP-phase boundary.37 This fact sheds light on
further consequences stemming from it, viz. vP-phase boundary accounts
for an upper limit of the size of verbal IdPs. To be more precise, the
syntactic material found within the vP-phase specifies the extent of a
verbal idiom (Svenonius 2005; Harwood 2013, 2015, 2016, 2017; Kim
2014; Harwood and Temmerman 2015; and Corver et al. 2017; among
others). What is more, having based our line of reasoning on psychological
verbal IdPs (e.g. paint the town / raise the roof “to enjoy”), the nominal
(NP) objects in these IdPs have been proved to be non-referential, and
always spell-out not as a separate DP-phase but as an NP of the vP-phase,
to evoke the idiomatic interpretation. Hence, with no DP-phase present in
IdPs, no problem of straddling the DP-phase boundary arises.
36
Following Harwood et al. (2016), it can be concluded that the determiner in a
NP of IdP idioms (e.g. raise the roof) is non-referential; whereas, the determiner
(definite article) in ICE-type idioms (e.g. spill the beans / fan the fire) is
referential, forming a DP.
37
Cf. Harwood and Temmerman’s (2015) findings concerning several idioms
which depend on material external to the vP-phase, i.e. modality and aspect. They
prove that if an idiom is deprived of the relevant aspectual form or modal verb, its
figurative interpretation is altogether lost. However, Harwood and Temmerman’s
(2015) claim is contradicted by Kitagawa (1986), Ifill (2002), and Svenonius
(2005), among others.
Psychological Idioms: Syntactic Constraints and Aspectual Properties 241
38
Since phasehood is often characterized in terms of interface considerations, the
diagnostics are usually classified into two major groups: LF diagnostics and PF
diagnostics, with reference given to the two interfaces. However, some researchers,
e.g. Matushansky (2005), Citko (2014), among others, divide phasehood
diagnostics into more types, adding a separate class of syntactic (or
morphosyntactic) diagnostics to the two based on the interfaces.
39
According to Citko’s (2014) syntactic diagnostic evidence, DPs are phases since,
e.g. DP is a domain for feature valuation; D is the locus of uninterpretable features;
movement out of DP proceeds through the edge; and D determines Spell-Out (cf.
Citko 2014: 123).
242 Chapter Four
Dem > Art >Num >unit >Pl / sort41 >Adj > n42 > N
(Svenonius 2007: 26)
40
Cf. Citko (2014: 109), who proposes a DP structure with more common nominal
functional projections, suggesting to accommodate the DP-elements. Nonetheless,
according to Citko (2014), nP is placed below DP, which is in contradistinction to
Chomsky (2007), Corver et al. (2017), and Harwood et al. (2016, 2017), for whom
nP marks referentiality and is placed higher than DP (cf. (4.54)-(4.55)).
41
In Svenonius’ (2007) hierarchy, the categories “Pl/sort” and “unit” apply to such
cases in which plurals cooccur with classifiers, involving either “unit” classifiers or
noun classifiers, not “sort” classifiers. Classifiers individuate masses for
quantification and counting (cf. Svenonius 2007: 23).
42
The “n” here is a gender or noun class marker (cf. Svenonius 2007: 26).
Psychological Idioms: Syntactic Constraints and Aspectual Properties 243
As can be seen in (4.55), idioms which are ICEs, and which follow the
pattern V + the + NP, as in Type (A) of psychological idioms, comprise a
244 Chapter Four
43
Canonical V-Object IdPs do not form a DP-phase, but give rise to the idiomatic
V+NP structure, which is confined to a single vP-phase, and spelled out at one go,
to evoke an idiomatic interpretation, as illustrated in (4.51), and repeated in (4.54).
44
It is worth mentioning here the structural constraint on idiomatic interpretations,
introduced by Koopman and Sportiche (1991: 224), who deal, among other issues,
with idiomatic constituents with open positions realised in possessors, such as the
ones found in our dataset of psychological idioms, e.g. get Y’s goat “to annoy,” try
Y’s patience “to annoy,” ruffle Y’s feathers “to annoy,” catch Y’s eye “to
fascinate,” win Y’s heart “to fascinate,” or turn Y’s head “to fascinate.” Koopman
and Sportiche (1991) argue that an open position never bears a direct
correspondence to the structural head of the idiomatic constituent, since no
idiomatic interpretation should be based on, for example, a pair of co-occurring
words viz. a modifier of the subject or a modifier of the VP. Therefore, they
propose the following structural constraint on idiomatic interpretations, formulated
in (i) below.
Psychological idioms of Type (A’), e.g. raise Y’s hackles, kindle Y’s
wrath, and ruffle Y’s feathers, also include the instances of alienable
possession, such as hackles, wrath, or feathers.46 They have been checked
against syntactic diagnostics, applied by Harwood et al. (2017), i.e.
passivization, topicalization and modification, to check if these idioms are
ICEs or IdPs. As a result, the psychological idioms which have passed the
diagnostics are ICEs and are presented in (4.56a-c); whereas an example
of a psychological idiom (with inalienable possession eye), which has not
met the requirements of the syntactic diagnostics, is shown in (4.56d).
Nonetheless, the idiom in (4.56d) is not an IdP.47
46
I would like to thank Prof BoĪena Cetnarowska for her suggestion that if hackles
and feathers in the psychological idioms raise Y’s hackles, and ruffle Y’s feathers
“to annoy” are personified, then these nouns may be treated as the parts of one’s
human body, which results in making them the instances of inalienable possession
not the alienable one.
47
Cf. a counterexample of what Corver et al. (2017) postulate about no ICEs with
inalienable possession, viz. a psychological idiom in (i), which includes inalienably
possessed heart, and still passes the diagnostics to be an ICE; and a psychological
idiom in (ii), which comprises alienable possession boats, and fails the diagnostics
to be an ICE:
(i) gladden Y’s hearts “to amuse”
His heart has been gladdened because of the presence of God’s Holy Spirit.
(Google) (passivization)
Her heart, the painting gladdened. (native speakers’ judgement)
(topicalization)
A lovelier June morning never gladdened innocent hearts. (Google)
(modification)
(ii) float Y’s boats “to appeal to”
# His boats were floated onto cradles.(native speakers’ judgement)
(passivization)
*Definitely, her boats, comedies float most.(native speakers’ judgement)
(topicalization)
# A calm sea can float some of her boat. (native speakers’ judgement)
(modification)
246 Chapter Four
My feathers were ruffled and the more I thought about it the more
ruffled they got. (Google) (passivization)
Her feathers, the misbehaving child ruffled. (topicalization)
(native speakers’ judgement)
This love story ruffles my emotional feathers every time I see it.
(Google) (modification)
d. catch Y’s eye “to fascinate”48
48
The fact that the idioms catch Y’s eye “to fascinate” fails the canonical syntactic
diagnostics does not make simultaneously the idiom an IdP, because of the open
slot realised in the position of the possessor, which is not expected in IdP-fixed
idioms. Instead, the idiom is expected to be a “mismatch” of ICE-idioms. An
example of an IdP-idiom with a possessor in its object can be open Pandora’s box
“to start something that causes many new and unexpected problems,” which does
not belong to our dataset of psychological idioms.
Psychological Idioms: Syntactic Constraints and Aspectual Properties 247
(4.57) Alienable possession in ICE idioms, e.g. kindle Y’s wrath “to
anger:”
45
An interested reader may also find some phase-bound account, proposed by
Svenonius (2005), which is left here aside for the lack of space. Svenonius (2005),
referring to multidominant trees, suggests structures called Banyan trees, which
include more than one undominated node. Banyan trees are created when first
Merge as well as internal Merge can target subconstituents of structures already
built by Merge. Thus, a psychological idiom with an open possessor position, e.g.
raise Y’s hackles “to annoy,” would be illustrated by means of a Banyan tree, as in
(i).
248 Chapter Four
(4.58) Inalienable possession in some ICE-idioms, e.g. catch Y’s eye “to
fascinate”
(i)
All in all, as shown in (4.57) and (4.58), the higher functional head is
absent in inalienable possession structures (cf. Alexiadou 2003; and Lin
2007; among others). Interestingly, Alexiadou (2003) notes that a
subject/agent relation between the possessor and the possessed noun is
reflected in the case of inalienable possession in ICE idioms, which
accounts for the presence of n* in alienable possession structures. Only
this type of idioms contains a DP-phase. Inalienably possessed nominals,
in turn, do not project n*P, and thus do not exhibit a DP-phase. This kind
of phase-bound analysis correlates well with the explanation provided in
section 4.5.2 above for nominal phrases in IdP fixed idioms, which lack
n*, and thus do not constitute a DP-phase, but belong to the vP-phase the
whole IdP idiom is confined to. However, idiomatic nouns in ICE-idioms
which are alienably possessed build independent n*Ps and form phases.
50
I refer an interested reader to some previous approaches to DOC constructions,
e.g. to Larson’s (1988) DOC framework, who posits a hierarchical structure for the
VP, involving two VP-shells, i.e. higher VP (V1P) and lower VP (V2P). The
structure is known as a “VP-shell” structure since the highest VP contains the
lower VP as a complement. In his proposal, Larson (1988) claims that the subject
is generated in SpecV1P; the Theme, i.e. the indirect object (IO) is generated as the
specifier of the lower VP (V2P); while the direct object (DO) as the complement of
V2. The verb starts out in the lower head position V2’ and moves to the higher
position V1’. The psychological idiom with a DOC, e.g. give Y the blues “to
depress,” is supposed to be illustrated in Larsonian shell structure, as in (i) below:
(i) Larsonian shell structure for the psychological idiom with a DOC
e.g. give Y the blues “to depress”
250 Chapter Four
refers to some direct referent in the discourse, and takes the role of an
indirect object. The role of a direct object is associated with a fixed idiom
component, as shown in (4.59a-f).
As shown in (4.59a-f), the fixed idiom element, in the role of the direct
object, can be represented either by a definite noun (e.g. the blues, the
wrong way), by an indefinite nominal unit, both singular (e.g. a curve, a
ball, a turn, a fright) and plural (e.g. pimples), by a compound noun (e.g.
goose pimples), or by a noun modified with an adjective (e.g. the wrong
way), or a quantifier (e.g. quite, a bit of). The verbal predicate of this type
of psychological idioms is exemplified by an agentive transitive verb give,
throw, or rub (up), the last of which can be accompanied by a particle up.
Nunberg et al. (1994), and Espinal and Mateu (2010) assume that open
object positions are one of the characteristics ICEs do allow for, which
makes them more compositional, while fixed slots are typical of canonical
IdPs. Following this claim, the psychological idioms in (4.59a-f), can be
treated as ICEs, composed of a lexical verb and a theme, with an open
position for the thematic goal.
The starting point in our phase-based analysis of psychological ICEs
with DOCs is McGinnis’ (2001) claim that certain types of double object
constructions may be phases, while the relationship between the direct and
the indirect object in a DOC is mediated by a functional head, called an
applicative head51 (cf. Pylkkänen 2008). Pylkkänen (2002, 2008),
developing a suggestion made by Marantz (1993), distinguishes high and
low applicative structures, as presented in (4.60) and (4.61), respectively.
51
The applicative head conveys a large number of meanings: benefactive,
malefactive, instrumental, locative and comitative meanings, among others (cf.
Citko 2014: 152).
252 Chapter Four
(4.62)
52
The Extended Projection Principle (EPP), proposed by Chomsky (1982), is a
linguistic hypothesis about subjects. The EPP refers to clauses which are required
to contain a NP or DP in the subject position viz. in the specifier of TP or IP (cf.
Chomsky 1982: 10). Importantly, the EPP-feature forces an item equipped with it
to project a specifier.
254 Chapter Four
53
In his idiom-as-selection proposal, built on O’Grady’s (1998) work, Bruening
(2010) argues that in order for two syntactic constituents, X and Y, to form an
idiom, one must select the other; and it is via selection that two elements are
combined together and interpreted contextually (cf. Bruening’s (2010) Constraint
on Idiomatic Interpretation, formulated in (4.39), and repeated in (i) for
convenience.
(i) If X selects a lexical category Y, and X and Y are interpreted idiomatically, all
of the selected arguments of Y must be interpreted as part of the idiom that
includes X and Y.
Lexical categories are V, N, A, and Adv.
(Bruening 2010: 532 (25-26))
To be more specific, in V-NP idioms (e.g. pull strings, or paint the town), the
verbs select direct objects, and then, taken together, they can be interpreted
idiomatically. In the case of verb-theme English ditransitive idioms (e.g. give NP
the boot, give NP a turn), they are proposed by Bruening (2010) to have ApplHP
and the following structure [ApplHP NP [ApplH' [ApplH [VP [V give] theme]]]].
ApplH selects V, which selects the theme. ApplH and V are interpreted
idiomatically, and thus all selected arguments of V have idiomatic interpretation.
Nonetheless, Bruening’s (2010) analysis lacks an explanation why the specifiers of
ApplH, i.e. the NPs in e.g. give NP the boot, give NP a turn, are excluded from the
idiomatic interpretations. Kim (2014: 216; 2015: 648) argues that it is not clear
why the specifier of ApplH is excluded from the idioms, provided they are also
selected by ApplH. Therefore, Bruening’s (2010) idiom-as-selection approach
seems not to rule out the specifiers of functional heads in a sufficiently explanatory
way.
Psychological Idioms: Syntactic Constraints and Aspectual Properties 255
54
The psychological idiom send Y into a funk “to depress” seems to be of the same
PD (postpositional Dative) pattern as Bruening’s (2010) original example for class
III ditransitive idioms (Class III: Verb NP to NP (send X to the showers “to remove
from a job”). However, the status of other idioms with the preposition “to” is
ambiguous. For instance, some psychological idioms of this type appear to be
regular VP-NP+PP idioms (e.g. drive Y up the wall “to annoy”), while others seem
to be resultative structures (e.g. frighten / scare Y to death “to horrify”).
256 Chapter Four
(4.64)
55
Svenonius (2005), in his idioms-listed in the lexicon approach, also predicts a set
of permissible idioms. Illustrating his patterns of permissible idioms with
psychological examples, they would be as follows: (i) idioms with free Goal
objects in a double object structure (give Y a turn “to scare”); (ii) idioms with open
possessor positions (raise Y’s hackles “to annoy”); (iii) but no idioms with free
Theme objects in a double object structure (*give a fright Y), which is also
confirmed by the data of our study on psychological idioms.
Psychological Idioms: Syntactic Constraints and Aspectual Properties 257
(4.65)
(4.66)
their idiomatic parts comprise verbs and objects that appear in the domain
of the VPs. Indeed, Kim’s (2015) proposal accounts for DOC-idioms,
including psychological idioms, but the puzzle why the phase boundary in
phase-bound Class I idioms is restricted to VP (which is commonly not
treated as a phase), and not to ApplHP (which is assumed to be a phase)
remains unsolved.
(4.67) b. A syntactic representation for Types (B), (B’), (C), and (C’)
As shown in (4.67b), in Type (B), Type (B’), Type (C), and Type (D’) ICE
idioms, a phase boundary for idiomatic interpretation is marked by a PP,
which is the complement of a VP. Since the prepositions (for, down, in,
up, off, etc.) in all the idiom types are functional, the PPs are capable of
being phases (cf. Marantz 1984, 1997; Chomsky 2000; and Citko 2014;
among others). The complements of the Ps are either fixed nominals (e.g.
in contempt, or up the wall) or nominals that comprise possessors (down
Y’s back, off Y’s feet). The object DPs in the aforementioned types are
represented either by fixed nominals with indefinite articles “a” (a torch, a
belly) or realising an open position (X or Y), and they belong to the VP.
Regardless of whether or not the DPs in these ICE-type idioms are
referential or not; and hence whether there are DP-phases formed or not,
Psychological Idioms: Syntactic Constraints and Aspectual Properties 261
these idioms do not pose a problem, since ICEs can straddle a number of
different phases.
Furthermore, the idiomatic phase-based structure in (4.68b) represents
Type (C’), Type (D), and Type (D’’) psychological ICE idioms, with the
patterns as in (4.68a).
As shown in (4.68b), the syntactic representation for Types (C), (D), and
(D’’) Type (B), differs from the one in (4.67b) only in the type of a DP-
object, which is a specifier of the VP. In (4.68b), the DP would be given a
chance to become a DP-phase if only the nominals were alienably
possessed as in (4.57). Instead, the DPs in these idioms refer to parts of
one’s body, which makes them inalienable; hence, no DP-phase is attested
in a VP in (4.68b). Nonetheless, these ICE-idioms build PP-phases,
formed by functional Ps (e.g. out of, to, in) and the complements of the Ps,
i.e. fixed NPs (joint), or NPs which realise an open position (e.g. X), or
NPs which comprise a possessor realising an open position (e.g. X’s
mouth).
In short, assuming that PPs can work as phases as long as they, among
other requirements, include functional Ps (cf. Kim 2014, 2015; and
Harwood et al. 2016), the psychological ICE-idioms which represent the
Psychological Idioms: Syntactic Constraints and Aspectual Properties 263
Is it strange that her judgment was perverted, and her heart was
eaten out. (Google) (passivization)
Her heart, the disappointed love for him was eating out.
(native speakers’ judgement) (topicalization)
Napoleon at St. Helena is eating his proud heart out with rage.
(Google) (modification)
If his socks had been knocked off, she couldn’t tell a word.
(Google) (passivization)
My socks, the news knocked off.
(native speakers’ judgement) (topicalization)
Psychological Idioms: Syntactic Constraints and Aspectual Properties 265
56
Cf. the structure in (4.57) for alienably possessed DPs in ICE-type idioms.
57
Cf. the structure in (4.58) for inalienably possessed DPs in ICE-type idioms.
266 Chapter Four
a. V + [NP AP]:
drive Y batty “to annoy”
leave Y open-mouthed “to surprise”
strike Y dumb “to surprise”
b. V + [NP AdvP]:
knock Y sideways “to scare”
bring Y low “to depress”
catch Y unawares “to surprise”
c. V + [NP PP]:
strike Y with wonder “to surprise”
strike Y with awe “to surprise”
leave Y at a loss for words “to surprise”
d. V + [NP NP]:
drive Y nuts/ bananas / bonkers “to annoy”
58
Cf. Bruening’s (2016) discussion concerning resultative constructions analysed
as small clauses, and other Small Clause approaches to the argument structure.
59
Jiménez-Fernández and Spyropoulos (2013), similarly to Cinque (1990) and
Haegeman (2010), assume that SC structure includes a functional projection F
above the projection of the category that functions as the predicate. Jiménez-
Fernández (2000) recognises some more possibilities of including other functional
categories in the structure of SCs, e.g. Aspectual Phrase. Bowers (1993, 2001)
posits a Predication Phrase above the SC; while Starke (1995) holds that SCs are
full clauses, and hence project a CP. For reasons of space, it is not possible to
discuss these different proposals related to the functional spine of SCs in detail.
Thus SCs are analysed here, after Bower (1993) and Citko (2014), as Predication
Phrase. Importantly, SCs are treated as reduced clauses, which lack a CP and a TP.
Psychological Idioms: Syntactic Constraints and Aspectual Properties 269
In brief, a PrP structure is not the only possible structure for small
clauses (cf. Citko (2011) for an overview of possibilities), but it seems to
work efficiently for the instances of psychological idioms of Type (E’).
them. Moreover, nearly all idioms from our dataset can be characterized as
having an open position located, e.g. in the direct object, small clause
subject, or possessor.
60
If an Experiencer is the trigger of emotion, it behaves like an agent, then “the
experiencer (…) resembles the agent to the extent that his involvement in the
action is volitional and he has control over the execution of the mental process”
(Dąbrowska 1997: 94). Therefore, the performer of the process appears rather
“active than passive, conscious rather than lacking consciousness, rational rather
Psychological Idioms: Syntactic Constraints and Aspectual Properties 275
than irrational, [and] capable of referring to the objective world and seeing things
that are really ‘out there’” (Dąbrowska 1997: 97).
276 Chapter Four
As can be seen in (4.77) and (4.78), the position of the Experiencer in the
aforementioned psychological idioms corresponds to the position an
Experiencer takes in psych-verbs. In all the cases the parallelism goes in a
straight line, viz. if an Experiencer (Mark for all these instances) occupies
the preverbal position, having the role of a subject, it plays the same role
and takes a preverbal position in the case of psychological idioms. On the
other hand, if the Experiencer has the role of an object, its position both in
psych-predicates and their idiomatic counterparts is postverbal.
What is more, the Experiencer, both for SE and OE psychological
idioms, is realised mostly overtly as a DP. However, in the case of the
idioms corresponding to OE psych-verbs, the Experiencer is realised either
as a DP-direct object (e.g. gave Mark the pip), or as a possessor of the DP-
object (e.g. tickles Mark’s fancy), or as a possessor of the DP-complement
of P (e.g. brought a hornet’s nest round Mark’s ears), or as a PP
complement (e.g. thumbed their noses at Mark). The Experiencer may be
also covert, so that only the context of the discourse reveals who the
Psychological Idioms: Syntactic Constraints and Aspectual Properties 277
Experiencer is (e.g. The screaming children raised the hump, so that Mark
couldn’t sleep).
As can be seen in (4.79), four types out of twelve elicited in Chapter Three
consist of a PP, with an Experiencer overtly placed, which comprises
altogether 35 psych-idioms out of 112 OE class II and III (ĺ 32%). Types
(B)-(B’) are the most numerous, covering 26%. These types include a verb
and a fixed object, followed by a Prepositional Phrase realising an open
position in a complement of the P, i.e. either in an NP (V+NP+P +NP), or
in a possessor modifying the fixed noun (V+NP+P +possessor +N). Types
(D)-(D’), in turn, realise two open positions, i.e. in the object position and
within a PP, hardly ever occur in this idiomatic dataset (6%).
Interestingly, in 5 SE idioms out of 49 SE idioms (which comprises
10% of Type (B’) and (D’’)), the Experiencer is located in a PP, viz. the
Experiencer is realised as a possessor modifying the fixed noun placed in a
PP. The instances of these SE idioms are shown in (4.80) below.
(4.81) a. X frightens Y
b. [CS+ ([X]Į, [INCH [BE ([FEAR ([Į])], [AT [Y]])]])]
c. X causes fear of X to come to be IN Y.
(Jackendoff 1990: 300 n.4, cited in Landau 2005: 7)
61
Following Jackendoff (1990), the meaning of each sentence relies on the
universal semantic categories that the relative construction is made up of. The
inventory of such categories include EVENT, STATE, THING, PATH, PLACE,
PROPERTY, and TIME. All these universal semantic categories can combine with
each other by means of functions such as IN, AT, BE, INCH, and so on.
62
Cf. Baker (1997) for a similar suggestion, Iwata’s (1995) “reversed” option for
the Experiencer to be placed within the mental state.
280 Chapter Four
Jackendoff’s (1990) analysis the target of fear equals its cause, Pesetsky
(1995) does not find this equation necessary.
Moreover, Bouchard (1995) treats one’s mental state as an independent
semantic argument, called psy-chose, which he names also a syntactic
argument since “in mental space, the psy-chose is somehow put in contact
with the argument it affects” (Bouchard 1995: 272). This “argument” can
absorb the emotion or feeling that the psy-chose denotes, as shown in the
periphrastic psych construction63 provided in (4.82a), or illustrated by
means of the OE in (4.82b).
Landau (2010: 10) further assumes that despite the fact that psych-
verbs are decomposed conceptually into an “action” light verb plus a
mental state (psy-chose), this does not imply that this decomposition
happens on the syntactic level as well.64 Instead, the locative preposition is
syntactically active no matter if the Experiencer is a bare nominal or not,
while syntactic activity happens in the mental state only when it is visible,
i.e. in periphrastic constructions.
In short, according to Landau’s (2010) localist approach, the
Experiencer designates a mental location, thus it is placed within the
structure of a PP. This PP may be headed by a lexical P (as in English
obliques) or a null P (as in Latin obliques); nevertheless, both cases are
structurally distinct from bare DPs (Landau 2010: 21-22). On the ground
of these fundamental assumptions of Landau’s (2010) localist and
decomposition theory of psych-verbs, the verbal idiomatic units that
63
Cf. section 2.5.2.1. in Chapter Two for more details.
64
Landau (2010: 137 fn. 2) mentions that, actually, some evidence has been
provided to justify the claim that periphrastic and synthetic psych constructions
have different forms in some semantic aspects that cannot be ascribed as the single
factor of incorporation. While in non-agentive contexts, periphrastic forms are telic
(as in (ii)), synthetic forms are not (as in (i)). Thus, simple N-to-V incorporation
does not imply such aspectual shifts.
i. The movie horrified / enraged Mary for/*in 15 minutes.
ii. The movie filled Mary with horror / awoke rage in Mary in/*for 15 minutes.
(Landau 2010: 137 fn. 2)
Psychological Idioms: Syntactic Constraints and Aspectual Properties 281
Test 2: imperatives
Carry a torch for children!
65
Even though the native speakers’ judgements vary concerning the
grammaticality of this sentence with the adverbs on purpose / deliberately /
intentionally, the most natural in this sentence is the adverb intentionally.
284 Chapter Four
Test 2: imperatives
*Have eyes for children!
The SE psych-verb “to love” is stative, but shares some properties with
eventive verbs, e.g. the imperative Love children! sounds good (cf. tests in
(2.23a) in Chapter Two). As shown in (4.85), the predicate “to love” can
be replaced by idioms corresponding to it, which do not always show the
same aspectual property. To be precise, e.g. the idiom carry a torch for X
has a non-stative viz. eventive agentive or eventive non-agentive reading,
since it satisfies the progressive tense test and the imperative test for
eventivity, and it can be used non-agentively (with the adverb
unintentionally) or agentively (with the adverb intentionally), as shown in
(4.85b). In turn, the idiom have eyes for X, fails the tests for eventive
(agentive vs. non-agentive) reading, and satisfies the tests for stativity (it
does not appear in the progressive or the imperative), which confirms its
stative reading. Thus, psychological idioms with an Experiencer Y in the
subject position can retain the same aspectual property (i.e. stative
reading) as the usual psych-verb which corresponds to them. However,
some SE idioms give rise to an eventive (agentive or non-agentive)
reading, in contradistinction to the corresponding psych-verb, as illustrated
in (4.85b).
Furthermore, the sentences in (4.86)-(4.88) below are to check the
aspectual property of psychological idioms, in which the Experiencer
Mary is realised in an object position of class II and III psych-verbs. In
(4.86), the meaning of the idioms is “to depress;” in (4.87) the idioms
mean “to annoy;” while in (4.88) the idioms mean “to matter to.”
Psychological Idioms: Syntactic Constraints and Aspectual Properties 285
Test 2: imperatives
Give Mary a hard time!
Test 2: imperatives
Break Mary’s heart!
Test 2: imperatives
Drive Mary bananas!
Test 2: imperatives
Bring a hornet’s nest round Mary’s ears!
Psychological Idioms: Syntactic Constraints and Aspectual Properties 287
Test 2: imperatives
# Carry some weight with Mary!
Test 2: imperatives
# Make a difference to Mary!
288 Chapter Four
widely used test for distinguishing telic and atelic predicates (the “in-for
test”), since Verkuyl (1972), Dowty (1979), Hinrichs (1985), Levin and
Rappaport (1995), Hay, Kennedy and Levin (1999), among others. It is
said that in-adverbials constitute a test for telicity (Mary painted the wall
in two hours/*for two hours), while for-adverbials for atelicity (Mary
walked for two hours/*in two hours).
b. Mary was raising the roof *in two hours / for two hours.
ĺ “to enjoy”
Mary raised the roof in two hours / *for two hours. ĺ “to
enjoy”
(the roof ĺ a count noun vs. “to enjoy” ĺ a telic / an
atelic idiomatic predicate)
66
An interested reader is referred to McGinnis’ (2002) detailed analysis in favour
of her claim that “the meaning of idioms is not entirely arbitrary: the structural
component of meaning (specifically, aspect) is both systematic and compositional.
This observation supports the claim of Distributed Morphology that structural
meaning, but not idiosyncratic meaning, is built in the syntax” (McGinnis 2002:
671).
67
Glasbey (2003) suggests regarding these mismatching idioms as aspectually
compositional, provided the aspectual composition is to include Krifka’s (1992)
“thematic relations” as part of its input.
Krifka (1989, 1992) argues that all thematic relations are cumulative in the
sense that if a predicate applies to x and y, it applies to the sum x + y. What is
more, Krifka (1989, 1992) states that telicity is a function of the structure of the
“incremental theme” argument of the verb. Telicity refers to the relation between
the structure of the argument and the described event (cf. Dowty 1991; Filip 1999;
Jackendoff 1996; Tenny 1987, 1994; Verkuyl 1993; Ramchand 1997; and
Rothstein 2004, 2008; among others). For Krifka (1989, 1992), a defining
characteristic of the incremental theme role is that it can satisfy “Mapping to
Objects” and “Mapping to Events” (cf. Punske and Stone 2015).
Psychological Idioms: Syntactic Constraints and Aspectual Properties 291
For the purpose of the book, the compositional model of idioms, with a
bipolar division of idioms into IdP and ICEs (cf. Nunberg et al. 1994;
Harwood et al. 2016), has been adopted.
Furthermore, the type of idioms that has been chosen as the research
dataset refers to emotional and mental condition of a human participant
(called an Experiencer). Thus, some semantic correspondence between
A Syntactic Study of Idioms 295
Cacciari and Tabossi 1988; Glucksberg 1991; and Gibbs 1994; among
others).
2. Both the syntactic productivity and the lexical creativity of idioms are
matters of degree, depending on the idiom compositional properties.
The results of the syntactic examination of idioms (on the example of
psychological idioms) have revealed their challenging properties.
These are: the syntactic and lexical variability of ICEs, viz. their
possibility to undergo quantification, topicalization, ellipsis, and
anaphora, among others. As exemplified by psychological idioms,
ICEs, thanks to their more compositional nature, tend to also exhibit a
degree of lexical substitution; thus, parts of idioms may be replaced
with other homomorphic lexical counterparts. IdPs, in turn, are usually
entirely resistant to any such alterations, as summarised in Table 4-1.
from section 4.3 in Chapter Four, repeated here for convenience.
Compositionality
Conventionality
Open Slots
Flexibility
Variation
Syntactic
Opacity
Lexical
Idiom
Non-
IdP High Opaque No No No
compositional
More
ICE Low Trans-parent Yes Yes Yes
compositional
3. The most current research performed within the scope of the Phase
Theory (cf. Svenonius 2005; Stone 2009; Harwood 2013, 2016, 2017;
Harley and Stone 2013; Kim 2014, 2015; and Corver et al. 2017;
among others) has shed light both on idiomaticity, and idiom
constrained variability. As a result of the analysis of psychological
idioms in the light of the Phase Theory, it has been noted that, in
addition to CPs and vPs, also DPs, ApplHP, or PrPs may act as phases.
4. The vP-phase imposes an upper bound on verbal IdP idioms. In other
words, IdPs are said to be confined to a single phase, namely the
clause-internal phase;
A Syntactic Study of Idioms 297
5. ICEs are able to span multiple phase boundaries and can depend on the
material beyond them. Moreover, nearly all idioms from our dataset
can be characterized as having an open position located, e.g. in the
direct object, small clause subject, or possessor.
The set of all idioms elicited for the 5 top SE psych-verbs under scrutiny
(3.13) love - feel deep affection (13 idioms)
a) carry a torch for X
You carry a torch for a girl you dated in high school.
(COCA)
b) fall head over heels in love with X
He fell head over heels in love with your sister Alice.
(Google)
c) set store by X
Her father had taught her not to set store by icons.
(COCA)
d) think the world of X
We both thought the world of my dad.
(COCA)
e) take (great) delight/ interest / joy / satisfaction in X
God likes us, takes delight in us, in fact.
(COCA)
f) take a fancy / a liking / a shine to X
John began to take a fancy to Sally late last August at the picnic.
(Google)
g) have a soft spot for X
Harry, I've got a soft spot for you, in my heart.
(Google)
h) have a thing about X
I have a thing about Maggie. I guess I'm in love.
(Google)
i) have a weakness for X
John has a weakness for Mary. I think he's in love.
(Google)
j) have eyes for X
“That may be true but I've only got eyes for you, Pretty Girl,” he says.
(Google)
k) go a bundle on X
I don't go a bundle on Anne's new haircut.
(Google)
300 Appendix 1
(3.15) hate - feel strong dislike for or hostility toward (10 idioms)
a) pour scorn on X
He poured scorn on the defendants: And then came this cynical part of the
alleged plot.
(COCA)
b) bear ill will toward X
Do you bear ill will toward Lady Di?
(COCA)
c) bear / owe a grudge against X
Dugliss has borne a grudge against me since that day he lost his first son,
the Fourth of July.
(COCA)
d) bear / feel aversion / malice / hostility / repugnance toward (to) X
But those with the large rifts persist in their slanders and bear malice
towards one another out of anger.
(Google)
e) have no stomach / taste / use for X
Pat has no stomach for violent movies.
(Google)
f) show contempt for X
Art critics for years ignored or showed contempt for De, especially for
political reasons.
(COCA)
g) hold X in abomination / contempt
His created beings hold him in contempt and despise him.
(Google)
h) turn Y’s back on X
But I don't think now is the time to turn your back on each other. So,
therefore, I support him.
(COCA)
i) turn Y’s nose up at X
John turned his nose up at Ann, and that hurt her feelings.
(Google)
302 Appendix 1
The set of all idioms elicited for the 8 top OE (class II) psych-verbs under
scrutiny, illustrated with the help of sentences extracted from the COCA /
via the Google Search.
t) knock Y sideways
The news of her brother’s death knocked her sideways.
(Google)
The set of all idioms, exemplified with the sentences taken from the
COCA / via the Google Search elicited for the 2 top OE (class III) psych-
verbs under scrutiny.
Aarts, Bas. 1992. Small Clauses in English: The Nonverbal Types. Berlin,
New York: Mouton de Gruyter.
Abel, Beate. 2003. English idioms in the first language and second
language lexicon: A dual representation approach. Second Language
Research 19, 329-358.
Abeillé, Anne. 1995. The flexibility of French Idioms: A representation
with lexicalized tree adjoining grammar. In Everaert, M.; and van der
Linden, E.-J.; Schenk, A.; and Schreuder, R. (eds), Idioms: Structural
and Psychological Perspectives, 15-42. Hillsdale, NJ: Lawrence
Erlbaum Associates.
Abney, Steven P. 1987. The English Noun Phrase in its Sentential Aspect.
PhD Dissertation. Cambridge, MA: MIT Press.
Abraham, Werner. 2010. Misleading Homonymies, Economical PPs in
Microvariation, and P as a Probe. In Cinque, G.; and Rizzi, L. (eds.),
Mapping Spatial PPs, 261-294. New York: Oxford University Press.
Adger, David. 2003. Core Syntax: A Minimalist Approach. Oxford:
Oxford University. Press.
Adkins, Patricia G. 1968. Teaching Idioms and Figures of Speech to
Non-Native Speakers of English. The Modern Language Journal 52
(3), 148-152.
Aelbrecht, Lobke; and den Dikken, Marcel. 2013. Preposition doubling
in Flemish and its implications for the syntax of Dutch PPs. The
Journal of Comparative Germanic Linguistics 16, 33-68.
Aisenstadt, Esther. 1979. Collocability restrictions in dictionaries. ITL:
Review of Applied Linguistics 45 (6), 71-74.
Alexander, Robin J. 1987. Problems in understanding and teaching
idiomaticity in English. Anglistik und Englischunterricht 32, 105-122.
Alexiadou, Artemis. 2003. Some notes on the structure of alienable and
inalienable possessors. In Coene, M.; and D’hulst, Y. (eds.), From NP
to DP. Volume 2: The Expression of Possession in Noun Phrases, 167-
188. Amsterdam: John Benjamins Publishing Company.
Alexiadou, Artemis; Anagnostopoulou, Elena; and Everaert, Martin
(eds.). 2004. The Unaccusativity Puzzle: Explorations of the Syntax-
Lexicon Interface. Oxford: Oxford University Press.
320 References
Becker, Joseph D. 1975. The phrasal lexicon. In Webber, B.; and Schank,
R. (eds.), Theoretical Issues in Natural Language Processing, 70-73.
Cambridge, MA: Association for Computational Linguistics.
Beeman, Mark. 1998. Coarse semantic coding and discourse comprehension.
In Beeman, M.; and Chiarello, C. (eds.), Right Hemisphere Language
Comprehension: Perspectives From Cognitive Neuroscience, 255-284.
Mahwah, NJ: Erlbaum.
Benjafield, John G. 1992. Cognition. Englewood Cliffs, New Jersey:
Prentice-Hall.
Belletti, Adriana; and Rizzi, Luigi. 1988. Psych-verbs and ș theory.
Natural Language and Linguistic Theory 6, 291-352.
Bennis, Hans. 2000. Adjectives and argument structure. In Coopmans, P.;
Everaert, M; and Grimshaw, J. (eds.), Lexical Specification and
Insertion, 27-67. Amsterdam: John Benjamins Publishing Company.
Berman, Arlene. 1973. A constraint on tough-movement. In Corum, C.;
Smith-Stark, T. C.; and Weiser, A. (eds.), Papers from the Ninth
Regional Meeting of the Chicago Linguistic Society, 34-43. Chicago,
IL: University of Chicago.
Bever, Thomas G.; Lackner, James R.; and Kirk, R. 1969. The
underlying structures of sentences are the primary units of immediate
speech processing. Perception and Psychophysics 5 (4), 225-234.
Biaáy, Adam. 2005. Polish Psychological Verbs at the Lexicon-Syntax
Interface in Cross-linguistic Perspective. Frankfurt am Main: Peter
Lang.
Bobrow, Samuel A.; and Bell, Susan M. 1973. On catching on to
idiomatic expressions. Memory and Cognition 1 (3), 342-346.
Bos, Johan; and Spenader, Jeniffer. 2011. An annotated corpus for the
analysis of VP ellipsis. Language Resource and Evaluation 45 (4),
463-494.
Boškoviü, Željko. 2005. On the locality of left branch extraction and the
structure of NP. Studia Linguistica 59, 1-45.
Boškoviü, Željko. 2011. On Unvalued Interpretable Features. In Lima, S.;
Mullin, K.; and Smith, B. (eds.), NELS 39: Proceedings of the 39th
Annual Meeting of the North East Linguistic Society, 109-120.
Amherst, MA: GLSA.
Boškoviü, Željko; and Lasnik, Howard. 2006. Minimalist Syntax: The
Essential Readings. Oxford: Blackwell.
Botelho da Silva, Teresa; and Cutler, Anne. 1993. Ill-formedness and
transformability in Portuguese idioms. In Cacciari, C.; and Tabossi, P.
(eds.), Idioms: Processing, Structure, and Interpretation, 129-143.
Hillsdale, NJ: Erlbaum.
322 References
Egan, Andy. 2008. Pretense for the complete idiom. Noûs 42 (3), 381-
409.
Ekman, Paul. 1992. An argument for basic emotions. Cognition and
Emotion 6 (3), 169-200.
Ekman, Paul. 1999. Basic emotions. In Dalgleish, T.; and Power, M.
(eds.), Handbook of Cognition and Emotion, 45-60. Chichester: Wiley.
Emonds, Joseph. 1976. A Transformational Approach to English Syntax.
New York: Academic Press.
Emonds, Joseph. 1985. A Unified Theory of Syntactic Categories.
Dordrecht: Foris Publications.
Epstein, Samuel. 2007. On I(nternalist)-Functional Explanation in
Minimalism. Linguistic Analysis 33, 20-53.
Ernst, Thomas. 1981. Grist for the linguistic mill: Idioms and “extra”
adjectives. Journal of Linguistic Research 1 (3), 51-68.
Espinal, M. Teresa; and Jaume, Mateu. 2010. On classes of idioms and
their interpretation. Journal of Pragmatics 42 (5), 1397-1411.
Estill, Robert B.; and Kemper, Susan. 1982. Interpreting idioms.
Journal of Psycholinguistic Research 11, 559-568.
Evans, Vyvyan. 2009. How Words Mean. Oxford: Oxford University
Press.
Evans, Vyvyan; and Green, Melanie. 2006. Cognitive Linguistics: An
Introduction. Mahwah, NJ: Erlbaum.
Everaert, Martin. 1993. Verbal idioms, participant idioms and theta
theory. In Smith, L. (ed.), Papers from the third annual meeting of the
formal Linguistics Society of Midamerica, 43-59. Bloomington, IN:
Indiana Linguistics Club.
Everaert, Martin.2010. In Rappaport Hovav, M; Doron, E.; and Sichel, I
(eds.), Lexical Semantics, Syntax, and Event Structure, 76-98. Oxford:
Oxford University Press.
Everaert, Martin; van der Linden, Erik-Jan; Schenk, Andre; and
Schreuder, Rob (eds.). 1995. Idioms: Structural and Psychological
Perspectives. Mahwah, NJ: Lawrence.
Fábregas, Antonio; Marín, Rafael; and McNally, Louise. 2012. From
psych verbs to nouns. In Demonte, V.; and McNally, L. (eds.), Telicity,
Change, and State: A cross-categorial View of Event Structure, 162-
184. Oxford: Oxford University Press.
Fábregas, Antonio; and Marín, Rafael. 2015. Deriving individual-level
and stage-level psych-verbs in Spanish. The Linguistic Review 32 (2),
227-275.
A Syntactic Study of Idioms 329
Literature and the Formula, 73-106. Ann Arbor: Center for the
Coordination of Ancient and Modern Studies.
Kiparsky, Paul. 1987. Morphology and grammatical relations. Ms.,
Stanford, CA: Stanford University.
Kipper-Schuler, Karin. 2005. VerbNet: A Broad-coverage, Comprehensive
Verb Lexicon. PhD Dissertation. Philadelphia: Computer and
Information Science Dept., University of Pennsylvania.
Kitagawa, Yoshihisa. 1985. Small but clausal. Chicago Linguistic Society
21, 210-220.
Kitagawa, Yoshihisa. 1986. Subjects in Japanese and English. PhD
Dissertation. Amherst, MA: University of Massachusetts.
Klein, Katarina; and Silvia Kutscher. 2005. Lexical Economy and Case
Selection of Psych-Verbs in German. Ms., Köln/Bochum.
Knappe, Gabriele. 2004. Idioms and Fixed Expressions in English
Language Study before 1800. A Contribution to English Historical
Phraseology. Frankfurt am Main: Peter Lang.
Koopman, Hilda. 1991. The Verb Particle Construction and the Syntax of
PPs. Ms., UCLA.
Koopman, Hilda. 2010. Prepositions, postpositions, circumpositions, and
particles. In Cinque, G.; and Rizzi, L. (eds.), Mapping Spatial PPs:
The cartography of Syntactic Structures, vol. 6, 26-73. Oxford: Oxford
University Press
Koopman, Hilda; and Sportiche, Dominique. 1991. The position of
subjects. Lingua 85, 211-258.
Koster, Jan. 1975. Dutch as an SOV language. Linguistic Analysis 1.2,
111-136.
Kövecses, Zoltán. 2000. Metaphor and Emotion: Language, Culture and
the Body In Human Feeling. Cambridge: Cambridge University Press.
Kövecses, Zoltán. 2005. Metaphor in Culture: Universality and
Variation. New York: Cambridge University Press.
Kövecses, Zoltán. 2010. Metaphor: A Practical Introduction. Oxford:
Oxford University Press.
Krifka, Manfred. 1989. Nominal reference, temporal constitution and
quantification in event semantics. In Bartsch, R.; van Benthem, J.; and
van Emde Boas, P., Semantics and Contextual Expression, 75-115.
Stanford, CA: CSLI Publications.
Krifka, Manfred. 1992. Thematic relations as links between nominal
reference and temporal constitution. In Sag, I.; and Szabolsci, A.
(eds.), Lexical Matters, CSLI Lecture Notes, 29-53. Chicago, IL:
University of Chicago Press.
338 References
Liontas, John. I. 2003. Killing two birds with one stone: understanding
Spanish VP idioms in and out of context. Hispania 86 (2), 289-301.
Lipka, Leonhard. 2002. English Lexicology. Lexical Structure, Word
Semantics and Word Formation. 3rd ed. of An Outline of English
Lexicology, 1990. Tübingen: Gunter Narr Verlag.
Lin, Chien-Jer Charles. 2007. Processing (In)alienable Possessions at the
Syntax- Semantics Interface. Paper presented at the Conference on
Linguistic Interfaces, University of Ulster, Northern Ireland, June 1-3,
2007.
Lin, Jimmy. 2004. Event Structure and the Encoding of Arguments: The
Syntax of the Mandarin and English Verb Phrase. PhD Dissertation.
Cambridge, MA: Massachusetts Institute of Technology.
Literary Devices Editors. 2013. Colloquialism. Retrieved from
https://literarydevices. net/colloquialism/
Liu, Dilin. 2008. Idioms: Description, Comprehension, Acquisition, and
Pedagogy. New York: Routledge.
Longobardi, Giuseppe. 1994. Reference and proper names: a theory of
N-raising insyntax and logical form. Linguistic Inquiry 25, 609-665.
Machonis, Peter. 1985. Transformations of verb phrase idioms:
Passivization, particle movement, dative shift. American Speech 60,
291-308.
MacKay, Donald G.; and Bever, Thomas G. 1967. In search of
ambiguity. Perception and Psychophysics 2, 193-200.
MacWhinney, Brian. 2008. A unified model. In Ellis, N. C.; and
Robinson, P. (eds.), Handbook of Cognitive Linguistics and Second
Language Acquisition, 341-371. New York: Routledge.
Maienborn, Claudia. 2003. Against a Davidsonian Analysis of Copula
Sentences. In Kadowaki, M.; and Kawahara, S. (eds.), NELS 33
Proceedings, 167-186. Amherst: GLSA.
Maienborn, Claudia. 2005. On the limits of the Davidsonian approach:
the case of copula sentences. Theoretical Linguistics 31 (3), 275-316.
Maienborn, Claudia. 2007. On Davidsonian and Kimian states. In
Comorovski, I.; and Heusinger, Klaus von (eds.), Existence: Semantics
and Syntax, 107-130. Dordrecht: Kluwer.
Makkai, Adam. 1972. Idiom Structure in English. Mouton: The Hague.
Marantz, Alec. 1984. On the Nature of Grammatical Relations.
Cambridge, MA: MIT Press.
Marantz, Alec. 1993. Implications of Asymmetries in Double Object
Constructions. In Mchombo, S. (ed.), Theoretical Aspects of Bantu
Grammar, 113-150. Stanford, CA: CSLI Publications.
A Syntactic Study of Idioms 341
Popiel, Stephen J.; and McRae, Ken. 1988. The figurative and literal
senses of idioms, or all idioms are not used equally. Journal of
Psycholinguistic Research 17 (6), 475-487.
Postal, Paul. 1971. Crossover Phenomena. New York: Holt, Reinhart and
Winston.
Postal, Paul. 1974. On Raising: One Rule of English Grammar and Its
Theoretical Implications. Cambridge, MA: MIT Press.
Primus, Beatrice. 1999. Cases and Thematic Roles: Ergative, Accusative
and Active. Tübingen: Niemeyer.
Pulman, Stephen. G. 1993. The recognition and interpretation of idioms.
In Cacciari, C.; and Tabossi, P. (eds.) Idioms: Processing, Structure
and Interpretation, 249-270. Hillsdale, NJ: Lawrence Erlbaum
Associates.
Pustejovsky, James. 1991. The syntax of event structure. Cognition 41,
47-81.
Pylkkänen, Liina. 1997. Stage and Individual-Level Psych Verbs in
Finnish. Paper presented in the Workshop on Events as Grammatical
Objects. Ithaca, NY: Cornell University.
Pylkkänen, Liina. 2000. On Stativity and Causation. In Tenny, C. L.; and
Pustejovsky, J. (eds.), Events as Grammatical Objects: The
Converging Perspective of Lexical Semantics and Syntax, 417-444.
Stanford, CA: CSLI Publications.
Pylkkänen, Liina. 2008. Introducing Arguments. Cambridge, MA: MIT
Press.
Punske, Jeffrey; and Stone, Schildmier Megan. 2015. Inner Aspect and
the Verbal Typology of Idioms. Presented at The 8th Brussels
Conference on Generative Linguistics.
Radford, Andrew. 1988a. Small Children’s Small Clauses. Transactions
of the Philological Society 86, 1-43.
Radford, Andrew. 1988b. Transformational Grammar: A First Course.
Cambridge: Cambridge University Press.
Radford, Andrew. 2000. T-contraction in a phase-based theory of
grammar. Essex Research Reports in Linguistics 33, 21-32.
Radford, Andrew. 2004. Minimalist Syntax: Exploring the Structure of
English. Cambridge: Cambridge University Press.
Ramchand, Gillian C. 1997. Aspect and predication. Oxford: Clarendon
Press.
Ramchand, Gillian. 2008. Verb Meaning and the Lexicon: A First Phase
Syntax. Cambridge: Cambridge University Press.
346 References
https://www.merriam-webster.com/
O’Dell, Felicity; and McCarthy, Michael. 2010. English Idioms in Use
Advanced. Cambridge: Cambridge University Press.
Online Etymology Dictionary. Retrieved from
http://www.etymonline.com
Online International Dictionary. Retrieved from http://idict.org
Online Oxford English Dictionary. Retrieved from
https://en.oxforddictionaries.com
Phrases and their origins. Retrieved from
http://www.phrases. org.uk/meanings/1184 00.html
Power Thesaurus. 1969. Retrieved from http://www.power thesaurus.org
Seidl, Jennifer; and McMordie, W. 1978. English Idioms and How to
Use Them (4th ed.). Oxford: Oxford University Press.
Simpson, John A.; and Weiner, E. S. C. (eds.) 1989. The Oxford English
Dictionary on CD-ROM (2nd ed.). Oxford: Oxford University Press.
Wielki multimedialny sáownik angielsko-polski i polsko-angielski [Great
Multimedia English-Polish and Polish-English Dictionary]. 2005.
Oxford: PWN.
WordNet 3.0, Farlex Clipart Collection. 2012. Princeton: Princeton
University, Farlex Inc. Retrieved from http://www.free thesaurus.com