Sie sind auf Seite 1von 234

CSERÉP ATTILA

Motivation Behind Idioms of Criticizing


DEBRECENI EGYETEM

2
CSERÉP ATTILA
Motivation Behind Idioms of Criticizing

Debrecen, 2001
DOKTORI ÉRTEKEZÉSEK

3
Készült
a Debreceni Egyetem BTK Doktori Tanácsa,
az Angol-Amerikai Intézet és
az Angol Nyelvészet doktori alprogram
támogatásával

Lektorálták:
Dr. Martsa Sándor, a nyelvtudomány kandidátusa
és Dr.Csapó József, Ph.D.

Technikai szerkesztő:

Fedélterv:

CONTENTS

ACKNOWLEDGEMENTS..................................................................................................................................

1. INTRODUCTION.............................................................................................................................................
1.1 Scope and aims................................................................................................................................................
1.2 Data.................................................................................................................................................................
1.3 Formal and cognitive semantics: differences..................................................................................................
2. A SURVEY OF VIEWS.....................................................................................................................................

4
2.1 Introduction.....................................................................................................................................................
2.2 Hockett.............................................................................................................................................................
2.3 Institutionalization..........................................................................................................................................
2.4 The scope of phraseology................................................................................................................................
2.5 Weinreich.........................................................................................................................................................
2.6 Analyzability....................................................................................................................................................
2.7 Chafe...............................................................................................................................................................
2.8 Idiosyncrasy.....................................................................................................................................................
2.9 Fraser...............................................................................................................................................................
2.10 Grammatical variation..................................................................................................................................
2.11 Jackendoff......................................................................................................................................................
2.12 Schematicity..................................................................................................................................................
2.13 Makkai...........................................................................................................................................................
2.14 Figurativity....................................................................................................................................................
2.15 Structural classification.................................................................................................................................
2.16 Arnold and other Russian phraseologists.....................................................................................................
2.17 Motivation.....................................................................................................................................................
2.18 Repetition of sounds and syllables................................................................................................................
2.19 Földes............................................................................................................................................................
2.20 Hadrovics.......................................................................................................................................................
2.21 Fernando........................................................................................................................................................
2.22 Lexical variation...........................................................................................................................................
2.23 Howarth.........................................................................................................................................................
2.24 Relation between the criteria........................................................................................................................
2.24.1 Introduction...........................................................................................................................................
2.24.2 Compositionality and analyzability.......................................................................................................
2.24.3 Analyzability and lexicogrammatical variation....................................................................................
2.24.4 Analyzability and figurativity................................................................................................................
2.24.5 Analyzability and structure...................................................................................................................
2.24.6 Analyzability and motivation................................................................................................................
2.24.7 Idiosyncrasy and other properties.........................................................................................................
2.24.8 Lexicogrammatical variation and other properties...............................................................................
2.24.9 Schematicity and other properties.........................................................................................................
2.24.10 Motivation and other properties..........................................................................................................
2.25 Combination of the criteria used in overall classifications..........................................................................
2.26 Moon..............................................................................................................................................................
2.27 Other properties.............................................................................................................................................
2.28 Thematic grouping........................................................................................................................................
2.29 Summary of the criteria................................................................................................................................
3. IDIOMS, METONYMIES AND METAPHORS..............................................................................................
3.1 Psycholinguistic evidence for analyzability and motivation..........................................................................
3.2 Metonymy........................................................................................................................................................
3.3 Metaphor.........................................................................................................................................................
4. IDIOMS OF CRITICIZING: ANALYSIS.........................................................................................................
4.1 Introduction.....................................................................................................................................................
4.1.1 Selection of relevant data........................................................................................................................
4.1.2 Treatment of dictionary and corpus data................................................................................................
4.2 Analysis...........................................................................................................................................................
4.2.1 Compositionality.....................................................................................................................................
4.2.2 Analyzability............................................................................................................................................
4.2.3 Extragrammaticality................................................................................................................................
4.2.4 Lexical uniqueness..................................................................................................................................
4.2.5 Lexicogrammatical variation..................................................................................................................
4.2.5.1 Measuring variation.............................................................................................................................
4.2.5.2 Grammatical variation.........................................................................................................................
4.2.5.3 Lexical variation...................................................................................................................................

5
4.2.6 Figurativity..............................................................................................................................................
4.2.7 Motivation...............................................................................................................................................
5. IDIOMS OF CRITICIZING: MOTIVATION...................................................................................................
5.1 Introduction.....................................................................................................................................................
5.2 Motivation by conceptual metaphors and metonymies..................................................................................
5.2.1 ABSTRACT IS CONCRETE...........................................................................................................................
5.2.2 CRITICIZING IS WAR..................................................................................................................................
5.2.3 CRITICIZING IS PHYSICALLY HURTING.......................................................................................................
5.2.4 ACCEPTING IS TAKING...............................................................................................................................
5.2.5 BAD IS DOWN............................................................................................................................................
5.2.6 ACTIONS ARE TRANSFERS.........................................................................................................................
5.2.7 ANGER IS HEAT, ANGER IS A DANGEROUS ANIMAL...................................................................................
5.2.8 The CONDUIT METAPHOR.........................................................................................................................
5.2.9 GENERIC IS SPECIFIC.................................................................................................................................
5.2.10 UNIMPORTANT IS SMALL.........................................................................................................................
5.3 Other metaphors and metonymies..................................................................................................................
5.4 Opaque idioms.................................................................................................................................................
5.5 Conclusion.......................................................................................................................................................
6. SUMMARY.......................................................................................................................................................

7. REFERENCES..................................................................................................................................................

8. ABBREVIATIONS............................................................................................................................................

9. APPENDIX 1.....................................................................................................................................................
9.1 fettle.................................................................................................................................................................
9.2 kilter*..............................................................................................................................................................
9.3 umbrage...........................................................................................................................................................
9.4 WORD1 with a capital X, WORD2 with a small Y........................................................................................
9.5 catch + ship/boat/taxi......................................................................................................................................
9.6 swallow + N*...................................................................................................................................................
9.7 broad/wide + choice/range*............................................................................................................................
9.8 dudgeon...........................................................................................................................................................
10. APPENDIX 2..................................................................................................................................................

11. APPENDIX 3...................................................................................................................................................


11.1 Native speaker cloze test answers.................................................................................................................
11.2 Native speaker judgement of analyzability...................................................................................................
11.3 Native speaker judgement of motivation......................................................................................................
12. APPENDIX 4..................................................................................................................................................

13. APPENDIX 5..................................................................................................................................................


13.1 Idioms of criticizing......................................................................................................................................
13.2 Collocations of criticizing.............................................................................................................................
13.3 Figurative single-word (near-)synonyms of criticize and criticism.............................................................
14. APPENDIX 6..................................................................................................................................................

15. APPENDIX 7..................................................................................................................................................


armchair*..............................................................................................................................................................
V + flak*................................................................................................................................................................
Aunt Sally*............................................................................................................................................................
backbiting*............................................................................................................................................................

6
back-seat driving/driver........................................................................................................................................
V + aspersion(s).....................................................................................................................................................
come down on sb like a ton of bricks....................................................................................................................
V + fire*................................................................................................................................................................
faint praise*...........................................................................................................................................................
salvo*.....................................................................................................................................................................
lashing...................................................................................................................................................................
stick*......................................................................................................................................................................
rap*........................................................................................................................................................................
roasting..................................................................................................................................................................
get on sb’s back......................................................................................................................................................
get on sb’s case......................................................................................................................................................
V + lecture.............................................................................................................................................................
give sb the rough side of one’s tongue..................................................................................................................
haul sb over the coals............................................................................................................................................
have a thick skin*..................................................................................................................................................
have the hide of a rhinoceros................................................................................................................................
honeymoon*..........................................................................................................................................................
in the line of fire*..................................................................................................................................................
leave oneself open..................................................................................................................................................
like water off a duck's back...................................................................................................................................
look who is talking................................................................................................................................................
nitpick*..................................................................................................................................................................
not pull one's punches*.........................................................................................................................................
people in glass houses...........................................................................................................................................
pot shot..................................................................................................................................................................
put the boot into sb................................................................................................................................................
rap sb on the knuckles...........................................................................................................................................
read sb the riot act.................................................................................................................................................
run the gauntlet.....................................................................................................................................................
sitting duck/target..................................................................................................................................................
V + hammering.....................................................................................................................................................
take it on the chin..................................................................................................................................................
take to task*...........................................................................................................................................................
talk to sb like a Dutch uncle..................................................................................................................................
the pot calling the kettle black..............................................................................................................................
tongue-lashing.......................................................................................................................................................

7
ACKNOWLEDGEMENTS

Many people have helped me with their advice and encouragement. I am especially indebted
to my supervisor, Béla Korponay, for his suggestions and relentless support throughout my
work. I would like to thank Péter Pelyvás and József Andor, who read the first version of
the dissertation and made numerous invaluable comments. Many thanks are also to all my
colleagues.

8
"The whole is more than the sum of its parts."
Aristotle, Metaphysics

1. INTRODUCTION

1.1 Scope and aims

The present study investigates English word combinations and makes an attempt to
investigate their properties and relationships. The focus of the dissertation is on the
semantics of idioms.
The main theoretical framework used for analysis is cognitive grammar, as outlined
in Lakoff (1987) and Langacker (1987). In section 1.3 we summarize the differences
between cognitive and formalist-logical approaches to meaning and idioms. One of our main
aims is to describe the place of idioms in relation to other phraseological units. Chapter 2
will survey the previous literature and examine the criteria that are used to identify
phraseological units. In chapter 4 we will attempt to apply these criteria to a number of
English expressions of criticizing. Another of our goals is to test cognitive grammar's claims
concerning the motivation of idioms by conceptual metaphors and metonymies. These
conceptual mechanisms are discussed in chapter 3, and chapter 5 attempts to find the
(synchronic) motivation underlying idioms of criticizing.

1.2 Data

The focus of this study is on English; therefore, the majority of our examples are
English. However, Hungarian examples will also be given, since the author's mother tongue
is Hungarian and Hungarian data can add further support to the claims made or can throw
light on the differences between the two languages. Furthermore, this work surveys not only
British and American but also Hungarian phraseologists. Hungarian examples are included
not for the purpose of systematic comparison. The dissertation is not meant to be a
contrastive study.

9
The structure of this study is such that chapters 1 and 2 (the introduction and the
survey) contain all types of phraseological units, whereas in chapters 3, 4 and 5 the
examples are mostly idioms. Single-word examples are also found, especially in chapters 2,
3 and 5.
In chapters 4 and 5 we analyze a number of idioms (or idiom-like expressions).
Idioms with meanings related to the words criticize, criticism and critic have been
considered relevant, while idioms used to criticize people (couch potato) have been
excluded. "Related" in the previous sentence is to be interpreted in the following way:
included in our analysis are a) idioms that are synonymous or near-synonymous with the
above-mentioned words (come under fire), b) idioms whose meanings are more specific
than the above-mentioned words (Monday morning quarterback), c) idioms related in
meaning to the above words (cover one's back 'make sure that you cannot be blamed or
criticized later for something you have done' (CIDI: 82)) as well as d) idioms that collocate
with criticism (take...on the chin).
English examples are taken from three sources: 1) the set of examples used by the
authors that are referred to in this dissertation, 2) various thesauruses and dictionaries, the
latter are general as well as specialized, and most of them are corpus-based (14 out of 20
completely corpus-based, LDEI partly corpus-based) and printed 1 and 3) a corpus of 45
million words, the 1997 issues of The Times. To search this corpus we used MonoConc for
Windows. Examples taken from our corpus are shown in Appendix 1 and Appendix 7.
Hungarian examples are taken from 1) the author's own lexicon and 2) various dictionaries.

1.3 Formal and cognitive semantics: differences

Idioms constitute a central category of conventional expressions, and they are


usually defined with reference to their noncompositionality, as in Moon (1995: iv): "[An
idiom] is a group of words which have a different meaning when used together from the one
it would have if the meaning of each word were taken individually". Similar definitions can
be found in many places (cf. Cowie et al. 1993: x, Gulland and Hinds-Howell 1986: 5, Seidl
and McMordie 1988: 13, Carter 1987: 58, as well as Hadrovics 1995: 30, O. Nagy 1999:

1
A list of the dictionaries referred to can be found under "Abbreviations".

10
19, Földes 1987: 13 and Bencédy et al. 1988: 494, 496). Compositionality, and meaning in
general, is handled in different ways by various linguistic theories, but formal theories share
certain basic assumptions, which are in sharp contrast with views held by cognitive
semantics, as described by Lakoff (1987), Langacker (1987) and Lakoff and Johnson
(1999). Formal semantics is a version of what Lakoff (1987: 167) calls "objectivist
semantics". Among the basic tenets of the latter we find the following: a) the world consists
of entities which have fixed properties, and which are in certain relations with each other, b)
the properties are independent of the human body and mind, in other words they are
objective, c) entities that have one or more properties in common form objective categories
with sharp boundaries, d) meaning is independent of the human body and mind, e) linguistic
expressions get their meaning via their correspondence to the world.
Formal semantics claims that both the world and human language can be described
via mathematical, set-theoretical models. Objective entities can be represented by abstract
mathematical entities and properties by sets of entities, which means that we can use set
theory to model the world. For example, red objects in the world fall into a category which
can be represented as a set, the set of red objects. Red as a property exists independently of
the human body and mind, so the category red is an objective category. Set theory can also
model natural language, since formal semantics conceptualizes natural language in terms of
a formal language. Lakoff and Johnson (1999: 445) refer to this as THE FORMAL
LANGUAGE METAPHOR: sentences of the natural language correspond to symbol sequences,
linguistic elements correspond to the individual symbols, and syntax consists of rules for
combining symbols or transforming symbol sequences into other symbol sequences. Thus,
syntax is separate from semantics, since the symbols themselves are meaningless. The task
of semantics is to provide interpretation for the meaningless symbols. Meaning is a
relationship between symbols and entities (Lakoff and Johnson 1999: 447).
Meaning is based on correspondence and truth, as it is correspondence and truth
that are capable of linking abstract, meaningless symbols with the world. Sentences either
correspond to the world or fail to do so. In the former case, they are true. In formal
semantics, the meaning of a sentence is a function that assigns a truth value to the given
sentence (Lakoff 1987: 229). Parts of sentences (noun phrases, verb phrases, etc.)
correspond to the world by picking out correct referents (Lakoff 1987: 230).

11
In formal theories, only expressions with literal meaning can fit the world, only
literal meaning can be true or false. There is therefore a close relationship between truth and
literal meaning. Nonliteral aspects of meaning are analyzed by pragmatics, poetry or
rhetoric, but not semantics, which has a narrowly defined scope. Metaphorical language is
regarded as deviant, because the words are not used in their proper literal senses. Since only
literal concepts can fit the world, metaphor cannot be conceptual. Idioms are usually treated
as dead metaphors, i.e. expressions that have lost their metaphoricity and are now
completely literal. For example, have come a long way now has the literal meaning 'have
made a lot of progress'. Similarly, kick the bucket may have been metaphorical in the past,
but now it simply means 'die' (Gibbs 1994: 273).
Meaning is not only literal, but also compositional, i.e. the meaning of a composite
expression is fully predictable from the meanings of the constituents and their arrangement. 2
Compositionality was introduced to explain the fact that speakers are able to produce and
understand an infinite number of novel utterances (Gibbs 1994: 60). It is compatible with
the formalized, "mathematical-logical" view of semantics, since the values of the component
parts of a mathematical expression fully determine the value of the whole expression.
Compositionality applies not only to sentences, but also to phrases. Noncompositional
expressions simply have to be listed in the lexicon, where the meaning is linked to the lexical
item as a whole. Syntax is sharply distinguished from the lexicon, the former being
considered to be fully regular and predictive, the latter containing the irregularities and
idiosyncrasies. Consequently, idioms have to be listed in the lexicon. Formal theories
assume what Langacker (1987: 29) calls the rule/list fallacy: there are only two mutually
exclusive options, either the speaker can produce an expression using the regular rules of
syntax, or the expression is irregular and it has to be part of the lexicon.
In Lakoffian-Langackerian cognitive grammar meaning is defined broadly as
conceptualization (Langacker 1987: 5). Semantics is encyclopedic, there being no sharp
boundary between semantics and pragmatics (Langacker 1987: 154). Concepts, i.e.
meanings of linguistic units, are understood relative to background knowledge structures,
called domains or idealized cognitive models (ICMs) (Clausner and Croft 1999: 2). 3

2
Note that in kick the bucket, one of the most commonly used examples in discussions of idioms, neither
the meanings of the constituents nor their arrangement (V + NP) helps us predict the idiomatic meaning.
3
Terms such as "domain", "ICM", "frame" are roughly synonymous (Clausner and Croft 1999: 4, Brederode
1995: 34-35).

12
Domains can vary from the relatively simple "leg" domain to the complex "hospital"
domain. The meaning of knee cannot be grasped without conceptualizing the leg, so that
knee is understood via the background concept "leg". The complex "hospital" domain is
necessary, if we are to understand the word nurse. In cognitive grammar there are no a
priori restrictions on which aspect of our knowledge can be included in a given domain, i.e.
which aspect of our knowledge is part of the meaning of a word. Part of the meaning of
nurse is the fact that they look after the sick, they work in hospitals, they wear white clothes
at work, etc. The encyclopedic view of semantics, however, does not mean, as Langacker
(1987: 159-60) points out, that all aspects of our knowledge are equally central to the
meaning of an item. If we know that two of our colleagues are allergic to cats, this
knowledge is peripheral to the meaning of cat. But the dividing line between central and
non-central can only be drawn arbitrarily.
Meaning in cognitive grammar is not limited to literal meaning. Given the
pervasiveness of figurative language, cognitive grammar regards figurativity as natural and
seeks to account for it in terms of various cognitive processes. Metaphor is not seen as
deviant at all. It is based on what Langacker (1987: 94) labels "semantic extension", i.e. the
literal meaning is extended, but the phonological/written form remains the same.
The separation of syntax from semantics and the lexicon is felt to be arbitrary, and it
is claimed that instead of an autonomous syntax it is more appropriate to imagine a
continuum (Langacker 1987: 36). Idioms and other non-idiom expressions (broad range)
are placed in the lexicon, but the regularities that they display are formulated in terms of
rules (i.e. schemas) (Langacker 1987: 42). Idioms and other conventional expressions have
unit status, but that does not mean that the internal structure of an expression disappears
after being established as a unit (Langacker 1987: 59). The idiom lazybones is a unit, but
speakers are aware of the component parts (lazy and bones). At the same time, idioms are
viewed as gestalts, i.e. units that are cognitively simpler than their parts put together.
Speakers need not pay attention to the constituents and discover how to combine them in
order to produce the idiom. It is argued that collocations (cf. 2.22) also form gestalt
structures (Brederode 1995: 26). Gestalt perception is characteristic not only of linguistic
units, but also of certain entities and actions. For example, the action of walking is complex,
since it involves moving the muscles in an appropriate way, lifting one leg and putting it

13
forward, then lifting the other leg, etc. We are aware of these "parts" of walking, but
normally they are not salient.
In cognitive grammar partial compositionality is claimed to be more common than
full compositionality (Langacker 1987: 449). Blackbird has a meaning that is more specific
than the meaning of black and bird together. Likewise, knowledge of the meaning of lazy
and that of bones does not enable the speaker to predict the meaning of lazybones. Instead
of full versus zero compositionality, cognitive grammar operates with the notion of
motivation: the constituent parts of a composite expression motivate - in various degrees -
the expression. Compositionality in formal semantics implies a building-block view of
meaning, the meanings of the component parts are added to one another. There are only
two types of expressions: compositional (i.e. predictable) and noncompositional (i.e.
unpredictable). Cognitive grammar discards the building-block metaphor, and instead of the
formalist-logical notion of compositionality emphasizes discrepancy, predictability and
motivation.4 In the dissertation we will use "(non)compositionality" in the sense in which it
is used by the linguist that we are discussing, otherwise we will use it in the same sense as
"discrepancy/predictability".
The discrepancy between the meaning of the whole and the combined meanings of
the constituents can vary, as in blackbird, gin and tonic, lazybones, promise the
moon/earth, the bee's knees. In some cases the meaning of the whole is just the opposite of
the combined meanings of the parts (You'll be lucky 'used to tell sb that sth that they are
expecting probably will not happen' (OALDb: 767), cf. Hungarian még csak ez hiányzott
'That's all I need'). While predictability is related to semantic discrepancy, this relationship is
not very simple. A small degree of discrepancy will make an expression unpredictable.
Motivation can be viewed as the naturalness of the relation between the combined meanings
of the constituents (in idioms the literal interpretation of the expression) and the meaning of
the composite structure (in idioms the nonliteral sense of the expression). This notion is
equivalent to opacity or transparency, the former emphasizing the lack, the latter
emphasizing a high degree of motivation. The use of lazy and bones in lazybones is more
motivated than the use of bee and knees in the bee's knees.
Compositionality is a matter of degree (in cognitive grammar). A relatively high
degree of compositionality can be found in phraseological units whose meaning includes the
4
Cognitive linguists are not the first to distinguish these notions. For example, discrepancy and motivation
are viewed as distinct by Cruse (1986: 39).

14
(literal) meanings of the component parts but also contains an unpredictable addition. The
examples blackbird and gin and tonic are of this type. The meaning of blackbird is more
specific than the expected meaning of black and bird combined in the given way. Mel'čuk
(1998: 30) refers to similar examples as quasi-idioms, which is a category separate from his
collocations and idioms. Barkema (1996: 76) describes such examples as "pseudo-
compositional". Gläser (1986: 66) also considers them nonidiomatic, while Makkai (1972:
315-17, 321) treats some of them as idioms (blackbird, gin and tonic, whisky and soda)
others as non-idioms (fish and chips, spaghetti and sauce, facts and figures). We regard
them as non-idioms.
There are idiomatic word combinations that differ from the previous examples in
that their idiomatic meaning contains some unpredictable addition apart from the literal
meaning of only one of their constituents, such as lazybones, drop names, run a
temperature or promise the moon/earth. The senses 'lazy', 'names', 'temperature' and
'promise', respectively, are parts of the meaning of the whole. Similar examples are labelled
"partly-compositional" by Barkema (1996: 76). The meaning of similes (shake like a leaf,
free as a bird) also contains the meanings of one of their constituents, and the meaning of
the whole is relatively predictable.
In drop names and run a temperature, 'names' and 'temperature' are part of the
respective idiomatic meanings 'over-use the names of celebrated or influential persons in
order to impress others with one's acquaintance with, or knowledge of, them' (ODEI: 159)
and 'have a body temperature higher than normal' (ODEI: 275). It is difficult to decide
which combination has a higher degree of noncompositionality. Drop names contains more
additional information, and the verb drop in the same sense ('mention') is used in let it drop,
and it is used in drop a hint, drop a bombshell/bomb and drop a brick in very similar senses:
'say', 'tell'. However, the number of these expressions is so small that such a combination
could hardly be called regular, and predictability implies regularity (cf. Langacker 1987:
448, 450). At the same time, words are often seen metaphorically as objects (cf. on the tip
of one's tongue, get one's tongue round sth, take the words out of sb's mouth), and drop
names evokes the image of names treated as objects which can be dropped; therefore, it
makes use of the same metaphor. In run a temperature there seems to be less addition, and
run can be used in the same sense in combination with fever as well, but apart from these

15
two expressions we can hardly find examples of the metaphorical relation between running
and having (a particular condition of the body).
More noncompositional are idioms in which the literal meanings of the constituents
are not part of the figurative meaning, but within this category there are various degrees of
predictability. The more natural the relationship between the figurative and the literal
meaning (i.e. the more motivated the idiom) is for the speaker, the more predictable the
figurative meaning is. The meaning of roll up one's sleeves 'get ready to work hard, often as
part of a group of people' (CCDI: 357) seems to be more predictable than that of the bee's
knees 'if you say that something or someone is the bee's knees', you are saying in a light-
hearted way that you like them a great deal' (CCDI: 25), although both idioms are to a large
extent noncompositional. In our view, predictability depends on both discrepancy and
motivation.

16
2. A SURVEY OF VIEWS

2.1 Introduction

Natural languages contain a large number of ready-made word combinations such as


explode a myth, lose one's head, how do you do, say when, never look a gift horse in the
mouth. All these examples can be taken as illustrating what is called formulaic or
prefabricated language, the characteristics of which include fixed structure (say/*tell
me/*speak when/*what time, *say it when) and noncompositionality. These properties
manifest themselves in various degrees, and they should therefore not be treated as all-or-
nothing phenomena. Moon (1998a: 6) reports that noncompositional or semi-compositional
units have been found in pidgin languages and animal communication as well. The sub-field
of linguistics that studies formulaic expressions is called phraseology.
Phraseological discussion is beset with difficulties arising from differing
terminologies. There are a number of superordinate terms covering the whole spectrum of
expressions. Western, especially British and American, researchers use the term "word
combination"5 or "idiom", while Russian scholars and those influenced by their work tend to
use "phraseme" or "phraseological unit" (Moon 1998a: 5, Cowie 1998b: 5). Other terms are
also found: "formulaic expression" (Coulmas 1994: 1292), "fixed expression" (Moon
1998a: 2), "conventionalized multi-word expression" (Fernando 1996: 37), "conventional
expression" (Langacker 1987: 35), "set expression" (Arnold 1986: 165), "set phrase"
(Mel'čuk 1998: 28), "phrasal lexeme" (Moon 1998b: 79, Lipka 1992: 80), "multi-word unit"
(Cowie 1988: 131), "set combination" (Zgusta 1971: 142), etc. Working within eastern
linguistic tradition, Hungarian phraseologists use "frazéma", "frazeológiai egység",
"frazeológiai kapcsolat", "állandó(sult) szókapcsolat" or "frazeologizmus". All of these can

5
This term sometimes includes free combinations as well (cf. Howarth's (1996: 11-12, 37) use of "composite
unit", which is a subtype of word combination, but it includes institutionalized as well as non-
institutionalized combinations). The meanings of free combinations are predictable to a great extent from
the meanings and arrangement of the components, and they are lexicogrammatically variable, e.g. drink the
tea - sip/find/pour etc. the tea, drink the tea/coffee/wine etc., drinks/drank/drinking etc. the/some/a lot of
etc. tea, the drinking of the tea, etc. Terms including both free combinations and conventional expressions
are "combination of word" and "combination" (cf. Zgusta 1971: 138, Benson et al. 1986)

17
also be found in a narrower sense in the literature, "állandó(sult) szókapcsolat" referring to
phrase-like units, and the other terms used only for noncompositional items (Juhász 1980:
83-84, Bencédy et al. 1988: 494-95, Földes 1987: 12-13).
"Phrasal lexeme", "multi-word unit", "phraseological unit", "frazeológiai egység"
and "phraseme", coined by analogy with "lexeme", all focus on the unit-like nature of these
expressions. The term "unit" is too broad, since not only words but phrases and sentences
are also (linguistic) units. The term "unit" is therefore not very informative, though adding
"phraseological" improves the situation, since there is an allusion to the formulaic nature of
these units. "Phrasal lexeme" emphasizes both the word-like and phrase-like nature of our
examples, but it fails to cover sentences. "Multi-word" implies that these units consist of
more than one word, and this is true of all the examples, but phrases such as the man in the
corner could also be described as multi-word units. Besides, Zgusta (1971: 143, n. 16)
points out that phraseological units may consist of only two words, and what is implied in
Zgusta's argument is that "multi" suggests more than two. "Multi-word" is more informative
than "unit", since it excludes single-word items, but it is still too broad and two narrow at
the same time.
Other terms ("word combination", "frazeológiai kapcsolat", "állandó(sult)
szókapcsolat") imply that we are faced with combinations of words, and this is not
appropriate either, since our examples are different from free combinations such as the man
in the corner (cf. the man/girl/student etc. in the corner, the man in the corner/behind the
bus/on the sofa etc.), and several of the expressions are not genuine combinations of words,
because the items that combine do not have independent figurative meanings (the bee's
knees). However, if we use the term "combination of words" for any type of combination
and reserve "word combination" for other than free combinations, then "word combination"
seems applicable. "Expression" appears to be a term closer to "combination" than to "unit",
but it is more revealing than the others, suggesting something special (i.e. idiosyncratic).
"Fixed" and "set" are misleading in that most of these units are not completely fixed. Moon's
(1998a: 2) general term for phraseological unit is FEI, which stands for "fixed expression
(including idiom)", although she admits that the term is unsatisfactory, since many units are
not actually fixed. All of them are conventional. "Idiom" is confusing, because it is also used
as the label for a subtype of conventional expressions, those that are noncompositional (cf.
1.3). We will be using "idiom" in this sense. On the whole "conventional/phraseological

18
expression" would be the best choice for the general term, but since it is not wide-spread we
will be using "phraseological unit" and "word combination" as well. Although
"phraseological unit" and "word combination" seem to emphasize different properties, they
are treated as synonyms in the literature. "Combination of words" will be used as the
broadest term (cf. Zgusta 1971: 138), including all the examples that we have given so far in
this section, i.e. free combinations and phraseological units.
Identifying conventional expressions is not an easy task, and linguists have
established several criteria in order to capture their essential features and classify them.
Below is a survey of how linguists have treated them and on which properties they have
focused. Discussion of a particular researcher is followed by examination of one or more
properties with which the given researcher is primarily concerned.

2.2 Hockett

Hockett (1958) is concerned with a subtype of word combinations, idioms. As the


following quotation shows, Hockett distinguishes between motivation and compositionality
(more precisely, predictability), though he does not use these terms: "it is one thing to
consider a meaning reasonable after we know it, and quite different to deduce the meaning
of a form from its structure" (Hockett 1958: 171-72). The first part of the quotation refers
to motivation, the second part describes (non)compositionality approached from the point
of view of comprehension. This distinction is also emphasized by Lakoff (1987: 450) and
Nunberg et al. (1994: 495, 496-97, 498).
Hockett (1958: 303-309) also examines idioms from a diachronic point of view, and
his interest lies in how a nonce form comes to be an idiom. He notes that the idiosyncrasy of
an idiom may reside in the original circumstances that attach something special to the
meaning. His example is That's a nice shade of blue, first uttered by a husband upon the
arrival of the wife with a new blue-green blouse and meant as a compliment, then used by
the wife teasingly while pointing to obviously green objects (Hockett 1958: 305). Hockett is
right in recognizing idiom status, though the example is idiomatic only to those who are
familiar with how the expression was born, the husband and the wife, and perhaps a few

19
friends or relatives, for whom the expression is no longer a nonce form. It remains a nonce
form for other speakers.
Hockett (1958: 173, 308-18) subsumes under "idiom" single morphemes (-ed),
nonce forms (pigwards), compounds (blackboard), whole sentences (Now is the time for all
good men to come to the aid of the party), pro-forms (he), numbers, proper names, clipped
forms (cello), acronyms (UNESCO), elliptical forms (You take the red cloth, and I'll take
the yellow), figures of speech (He married a lemon) and slang words. Hockett was a
structuralist, and this partly explains why he extends the notion of idiom to single words and
morphemes. Any unit whose meaning is not deducible from its structure is an idiom for him.
Since the smallest meaningful units are morphemes, single morphemes are by definition
idioms. Focusing on compositionality to the exclusion of other properties is also conducive
to treating single words as idioms.
Note that Hockett's definition of idiom leads to undesirable consequences. The
definition runs as follows: "Let us momentarily use the term 'Y' for any grammatical form
the meaning of which is not deducible from its structure. Any Y, in any occurrence in which
it is not a constituent of a larger Y, is an idiom" [my emphasis] (Hockett 1958: 172). From
this we have to conclude that the words red and herring are idioms when they occur in
constructions such as a red bus, the herring she caught, or red herring 'a type of fish of a
particular colour', but not when they occur in red herring 'an unimportant fact, idea, event,
etc. that takes people's attention away from the important ones' (OALDb: 1064).
Furthermore, as Howarth (1996: 28) points out, idioms can collocate with other idioms. His
example is go at sb hammer and tongs, which he claims is a collocation of go at sb and
hammer and tongs.6 In Fernando's (1996: 144) example - ALP plans to tighten govt fat
cats' belts - we also find two co-occurring idioms. In fall for sth hook, line and sinker, an
idiom, more specifically a phrasal verb (fall for sth), collocates with a trinomial (hook, line
and sinker). Following Hockett we would have to say that go at sb, hammer and tongs, fall
for sth, hook, line and sinker and fat cats are not idioms in the above examples, since they
are parts of larger idiomatic units. This is counter-intuitive.
We claim that pro-forms, numbers, most proper names, and ellipsis are not regarded
as idioms. A number of forms are idiomatic for Hockett, simply because they constantly
change their reference. Proper names and anaphoric pro-forms refer to different entities in
6
Viewed from a cognitive perspective go at evokes an image of a person attacking another one. Adding
hammer and tongs to the verb further specifies the nature of the attack.

20
different speech situations. Hockett (1958: 310) mentions the term "anaphoric substitutes",
but he refers to not only the textual environment, "the preceding speech", but also "factors
of the non-speech environment", i.e. he considers pro-forms idioms, whether their reference
is anaphoric or extralinguistic. However, meaning should not be identified with reference
alone.
Similarly, a unit will not be an idiom simply because it is a clipped form or an
abbreviation. Fernando (1996: 43) excludes from phraseology acronyms such as UNESCO
or UN, but includes examples like VIP and RSVP. VIP is fixed and encodes a culturally
salient phenomenon, while RSVP is a formula with special pragmatic function.
Hockett devotes most of his discussion to the unpredictability of the meaning of
idioms, but he only briefly mentions figurativity (Hockett 1958: 317-18). He does not
discuss the distinction between compositionality and motivation in detail and does not deal
with degrees of motivation and lexicogrammatical restrictedness. He seems to believe in the
lexicon versus syntax division, claiming that a dictionary contains idioms (Hockett 1958:
173). At the same time, he draws our attention to the similarity of single words and idioms
and notices that an idiom may have a literal counterpart (Hockett 158: 172).

2.3 Institutionalization

The process by means of which a nonce form starts spreading in a speech


community and is finally accepted as a familiar item is known as institutionalization (Bauer
1983: 48). This feature does not distinguish single words from phraseological units, but it
does exclude free combinations such as affect world trade. Fernando (1996: 5) points out
that for Makkai institutionalization is one of the criteria of idiomaticity, and this is certainly
true, since it helps us distinguish idiomatic and non-idiomatic combinations (pepper and salt
versus coffee and cream) (Makkai 1972: 159). Yet it is not a sufficient criterion, because
some institutionalized binomials (deaf and dumb) and phrasal compounds (quick as a flash)
are considered non-idiomatic by Makkai (1972: 124, 316-17, 338-39).
Hockett's example (That's a nice shade of blue) has a very low degree of
institutionalization, since it is familiar as a unit only to a couple of people. Most
conventional expressions have a high degree of institutionalization (Bencédy et al. 1988:

21
495, O. Nagy 1999: 11, Moon 1998a: 7). But some are limited to relatively small
communities. Fernando (1996: 67) gives a few examples of idioms used by Australian
street-kids only, such as snowdrop 'steal clothes off the line'. There are word combinations
that are limited to a dialect or a register, among the latter we find mocsári gólyahír 'marsh
marigold' and éjjeli nagy pávaszem 'peacock-butterfly', which are technical terms (Juhász
1980: 85). Makkai (1972: 321-35) lists similar examples in English (bitterhead 'a fish,
Notemigonus crysoleucas', cattail 'a tall reedlike marsh plant, Typha latifolia'), and some
idioms associated with particular registers can also be found in Gläser (1986: 37). A further
Hungarian example is lehúzza a rolót, which in the sense '[of a goalkeeper] save all the
shots' is used only in the context of sports. Some idioms are used only in British (and
Australian) English (have green fingers), others only in American English (have a green
thumb). Apart from technical terms of the type listed above, institutionalization is not used
to distinguish (sub)categories of phraseological units. Although catchphrases may be
familiar only to some speakers, as Crystal (1995: 178) and O. Nagy (1985: 9) point out, the
distinction made between them and other categories is usually based on origin as well as
syntactic and pragmatic function, rather than institutionalization alone.
In cognitive grammar the term "entrenchment" corresponds to what is described
above as institutionalization. The two terms differ only in that the former puts the emphasis
on the mental and neural aspects of language, while "institutionalization" focuses on the
social side. A nonce form becomes entrenched through continuous repetition and achieves
unit status when it is deeply entrenched and easily evoked as a whole (Langacker 1987: 59,
100).
Langacker (1987: 59) claims that the degree of entrenchment depends on frequency
of occurrence, and in corpus linguistic approaches institutionalization is also usually
measured in terms of frequency of occurrence. But corpus studies of idioms reveal that unit
status does not necessarily presuppose a high frequency. Moon (1998a: 60) shows that most
idioms that she has studied have low frequencies. Her statistics reveal that 40 per cent of
the phraseological units drawn from an 18 million-word corpus occur less than five times (in
corpus linguistic parlance this means random chance frequency), and over 70 percent of the
units occur with a frequency of less than 1 per million tokens. Moon's data may be skewed,
since her corpus has certain limitations. It is unbalanced, containing a great proportion of

22
journalism and only a small amount of spoken English (Moon 1998a: 48). 7 By the standards
of today, it is also very small. Furthermore, it must be borne in mind that Moon (1998a: 2)
excludes phrasal verbs and many restricted collocations such as explode a myth, make a
decision, etc. from the set of units she examines. Despite these limitations, she claims that
the general tendencies she has found would be observed in other corpora as well (Moon
1998a: 49). CCDI (xvii), listing about 4400 idioms based on a 211 million-word corpus (cf.
Moon 1998a: 66), ranks 30 per cent of its units as very rare, occurring less often than once
in every 10 million words. The next group, items occurring between 1 and 3 times in every
10 million words, contains over 1500 units. It must be noted, however, that CCDI does not
include phrasal verbs and many conversational formulae such as you know, so to speak, by
the way, all of which are listed in ODEI (603, 508, 87).
The frequency of phraseological units depends on the genre to some extent. The
least genre-bound are those that organize discourse (let alone, on the one hand and in a
nutshell are among the examples listed by Moon (1998a: 236-39)), the most genre-bound
are tied to speech (Moon 1998a: 69). Phraseological units are fairly common in journalism
and fictional dialogue, but the latter does not necessarily reflect authentic spoken language
(Moon 1998a: 69). Moon (1998a: 64-68) reports research conducted by others and shows
that the tendencies observed in her corpus are mirrored in other corpora, although
comparison of different corpora is difficult. She also claims that idioms on the whole are not
significantly common in ordinary spoken English (Moon 1998a: 72-73).
A property which is related to institutionalization and which is often believed to be
typical of idioms is their uniqueness to a particular language. Uniqueness implies that word
for word translation is impossible, and Smith (1943: 176-77) claims that translation is a
good test of idiomaticity. For example, far and away loses its idiomaticity, if translated
literally. Trying to dispel the belief in the uniqueness of idioms, O. Nagy (1999: 17) explains
that several Hungarian idioms have been borrowed from other languages. He also points out
that similarity in meaning and lexicogrammatical structure between two idioms from two
different languages is not necessarily the result of borrowing, since human experience is
similar in different places in the world (O. Nagy 1985: 10).
The notion that an idiom has no counterpart in other languages is reflected in some
of the definitions. NSOED (1305) gives the following paraphrase for one of the senses of

7
Note that this should skew the frequency data in favour of a relatively high frequency.

23
"idiom": "a form of expression, grammatical construction, phrase, etc., peculiar to a person
or language". This sense is given as separate from but related to the sense mentioned in
(1.3), but not all definitions separate the two senses, as can be seen in the following: "an
expression unique to a language, especially one whose sense is not predictable from the
meanings and arrangement of its elements" (McArthur 1992: 497). Avoidance of "idiom" by
some scholars is due to the polysemy and the conflicting use of the term. For example,
Arnold (1986: 166) and Moon (1998a: 3-5) mention various senses of "idiom" and decide
not to use the term.

2.4 The scope of phraseology

In Hockett's (1958) discussion the question arises what types of units are idioms.
The scope of phraseology varies depending on the linguist's emphasis on certain significant
properties of formulaic word-combinations. If we focus on some properties to the exclusion
of others, we can decide to ignore various types of phraseological units (Juhász 1980: 83-
86). Hockett's focus on noncompositionality leads to the omission of many non-idioms that
are phraseological units (explode a myth, I don't know about you, but...).
As has been mentioned, the lower limit is usually taken to be two words. However,
Hockett (1958: 172), as we have seen, analyzes single morphemes as idioms, since their
meanings are unpredictable. We agree with those who do not accept Hockett's view (Moon
1998a: 10, Howarth 1996: 17, Makkai 1972: 38). The notion of (non)compositionality
implies the presence of more than one meaningful element, i.e. more than one morpheme,
otherwise it would be irrelevant to discuss composition. Although we have pointed out
above that noncompositionality implies unpredictability, we have to add that it implies
unpredictability in situations when predictability is expected. Essentially the same line of
argumentation can be found in Makkai (1972: 28), who holds that single morphemes have
no internal morphological structure, and nothing leads to false expectations in terms of
meaning.
Hockett (1958: 173, 308-309) treats most affixed words as non-idiomatic but some
unpredictable examples (pigwards) are called idioms. If the two-or-more-word criterion is
waived and the focus is on the unpredictability of meaning, then the lower limit is the

24
polymorphemic word. Morphemes are meaningful and can be combined to form larger
units; consequently, speakers will have certain expectations of what sense a polymorphemic
unit could have. In this approach a word like mouthful 'a word or a phrase that is long and
complicated or difficult to pronounce' (OALDb: 832) would be classified as an idiom.
However, words have a type of unit status that bound morphemes do not seem to have.
Aitchison (1994: 130) gives psycholinguistic evidence that it is words, rather than separate
morphemes, that are stored in the mind. It must also be borne in mind that phraseologists'
interest lies in idiosyncratic restrictions on combinations of words. Phraseological units
display formal and semantic idiosyncrasies that are unexpected in the light of the evidence
that can be collected from a scrutiny of free combinations. The ill-formedness of *give away
the game does not follow from well-formedness of They felt like they were giving away
company secrets (CCED: 712). Most researchers do not consider affixed words as a
relevant target to study in phraseology, and we also exclude them from this field.
The temptation to treat words as idioms increases if we realize the similarities
between single words and idioms. The idea of similarity can be traced back to Charles Bally,
whose approach was semantic (Juhász 1980: 79). He stressed the possibility of finding
single-word synonyms for phraseological expressions and this property of idioms is often
referred to in comparisons of words with idioms (cf. Arnold 1986: 175, Gläser 1986: 16-17,
Zgusta 1971: 148).8 Single-word synonymy can be traced back to the view that figurative
language is parasitic on literal language, and idioms stand for words used in their literal
senses. Compositionality has similar roots. It is a consequence of truth-conditional
semantics, which also emphasizes literal meaning. Single-word synonymy is supposed to
show the unitary (i.e. unanalyzable) meaning of idioms. For example, Quirk et al. (1985:
1162) distinguish idiomatic and non-idiomatic 'verb + particle' constructions partly on the
basis of the semantic criterion of single-word synonymy. Thus, put up with is judged to be
idiomatic because it can be replaced by tolerate. Cowie and Mackin (1993: 424-25) also list
the same criterion: step up (in one of its meanings) is synonymous with 'improve' or
'enhance', and take off (in one of its meanings) is equivalent to 'mimic' or 'imitate'. However,
as Quirk et al. (1985: 1162) point out, this test is not absolutely reliable, and it should be
supplemented, since get across 'move across' is not classified as idiomatic despite the

8
Zgusta (1971: 148) mentions as one of the properties of phraseological units that they can have one-word
equivalents in a foreign language. This is basically the same property, but it is approached cross-
linguistically.

25
existence of the synonym cross. One could also argue that over the moon is not exactly
synonymous with the single word happy, because it expresses an intensive degree of
happiness. Its meaning is more properly paraphrased as 'very happy'.
There are also cases where the idiom and its non-idiom equivalent may differ in their
referential range. Fernando (1996: 102) claims that (all) the world and his wife tends to
refer to the social elite, and every Tom, Dick, and Harry has the additional meaning of 'no
special importance', while their equivalent, everybody, does not have these restrictions. The
different ICMs evoked by these idioms greatly contribute to the meaning differences.
In cases where an idiom is synonymous with a word there may still be stylistic or
register differences. A baker's dozen 'thirteen' is labelled as old-fashioned in CCDI (15),
CIDI (20), CIDE (94) and OALDb (80). Go west 'die' is also marked as old-fashioned in
CCDI (417) LID (379) and CIDI (420). Pop off is an informal equivalent of die according
to ODPV (260), OALDb (979), and CCED (1276), while it is humorous according to CIDE
(1095), and round the bend 'crazy' is also considered informal in CCDI (27), CIDI (29),
CCED (144) and OALDb (102), though not in CIDE (116). The idiom breathe one's last is
listed as a literary synonym of die in CIDE (162), CCED (197) and OALDb (144).
From a semantic viewpoint other similarities between words and idioms can also be
found. For example, many idioms have figurative senses, just as many words do. Idioms and
words display various meaning relationships, such as polysemy, antonymy and synonymy.
However, polysemy, the existence of two or more non-compositional meanings, is less
common in idioms. Moon (1998a: 188), who does not consider phrasal verbs, estimates that
about 5 per cent of her database word combinations are polysemous and cites Klappenbach,
who reports investigations suggesting that 8-9 per cent of Russian word combinations are
polysemous. Földes (1987: 22) reports 15 and 17 per cent for Russian and 9 per cent for
German. The following can be considered polysemous idioms: call it a day 'stop doing sth',
'retire' (CCDI: 94), bite the dust 'fail or cease to exist', 'die' (CCDI: 114), come a cropper
'suffer a sudden and embarrassing failure', 'accidentally fall and hurt oneself' (CCDI: 88),
and go up in smoke 'catch fire, burn down', 'be destroyed' (Moon 1998a: 189). Polysemy
will rarely lead to ambiguity in context, since polysemous expressions are often associated
with different collocates (Moon 1998a: 188-93).
It is sometimes difficult to decide whether two nonliteral meanings are related or
not, i.e. distinguish polysemy and homonymy. Moon (1998a: 189) lists on the rocks 'served

26
with ice', 'in trouble, shaky' as a polysemous item, but the figurative meanings seem to be
unrelated.
No statistics can be found in the literature for how common antonymy is, but Moon
(1998a: 129) notes that of her 533 phraseological units with the syntactic head in, 37 have
antonymous parallels with out of (in touch  out of touch). Földes (1987: 25) also
considers antonymy among phraseological units less common than among single words,
claiming that it is more commonly found in units functioning as verb phrases than in other
types. We have found no statistics to support or refute this claim, but Moon's (1998a: 129,
156-58) antonymous examples are mostly verb phrases and prepositional phrases, the latter
functioning as adverbials. Antonymous idioms often contain antonymous words (quick off
the mark  slow off the mark), but examples can also be found in which the different
constituents are not antonymous outside the idiom (take heart  lose heart). This is noted
by both Földes (1987: 25-26) and Moon (1989a: 156, 158).
Viewing idioms as idiosyncratic word equivalents is unlikely to help account for the
existence of antonymous idiom pairs. If we assume, as cognitive grammar does, that idiom
constituents can have independent meanings, we will be in a better position to understand
why we find antonymous idioms. However, we do not claim that analyzability is the only
explanation for antonymy. In order to investigate this issue large-scale corpus research
would be necessary.
As far as synonymous idioms are concerned, Földes (1987: 23) claims that, as
opposed to single words, many phraseological units are fully synonymous. This contradicts
claims made by Cowie et al. (1993: xliii) and Gläser (1986: 26-27, 111). Gläser (1986: 27)
lists the following synonyms of quickly: by leaps and bounds 'not formal, very quickly', in
less than no time 'rapidly', in a flash 'quite or very suddenly, quickly', at the drop of a hat
'immediately and willingly' and before you could say Jack Robinson 'coll., very quickly'. She
also gives examples of synonymous proverbs (Once bitten, twice shy, A burnt child dreads
the fire) but notes that full synonymy is rare (Gläser 1986: 111-13). Cowie et al. (1993:
xliii) claim that this is true of idioms in general, and Gibbs (1994: 301-303) shows that the
absence of full synonymy in a particular context is partly due to the differences in underlying
metaphors (cf. 3.1). As has been mentioned, different ICMs give rise to differences in the
conceptual content of expressions.

27
Words and idioms also display similarity from a syntactic viewpoint. Cruse (1986:
38) explains that impossibility of the re-ordering of parts and impossibility of interruption
are characteristic of idioms and words alike. Among the examples he uses to illustrate
interruption and re-ordering we find ?Arthur has a chip, apparently, on his shoulders and ?
John has a bee about it in his bonnet respectively. The latter, however, is a case of
interruption, rather than re-ordering, since about it is a phrase that typically co-occurs with
the idiom have a bee in his bonnet. In other words the idiom have a bee in one's bonnet is
often followed by a prepositional phrase headed by the preposition about, but this
prepositional phrase is not part of the actual idiom. Although it is not always easy to
determine the boundaries of idiomatic expressions, support for our claim comes from
dictionaries that do not show about + noun/pronoun in some of their examples and/or in the
citation forms (CCDI: 26, CIDI: 27, ODEI: 255, OALDb: 98, CCED: 137, CIDE: 112). 9
LID (22) puts about sth in brackets in the citation form, which also suggests optionality.
Notice also that words or expressions with a meaning similar to that of the idiom often
require prepositional phrases: keep thinking about sth, keep talking about sth, be obsessed
with sth. A more suitable example to show the blocking of re-ordering would be give the
game away versus *give away the game. Structurally invariable expressions can be seen as
the prototypical examples of idioms, but many idioms exist where a certain degree of
variation is permitted (turn the clock back/turn back the clock).
Notwithstanding the difficulty of determining which words fall within the boundaries
of an idiom, végre kapcsolt 'it has finally clicked' is not a two-word idiom as Bencédy et al.
(1988: 496) claim, since végre has the independent meaning 'finally', it is optional in
combination with the metaphorical verb kapcsol 'it clicks' and can be replaced by other
adverbials such as idejében 'in time' or késõn 'late' without changing the meaning of the
verb. Similarly, the verb kapcsol has an independent meaning (cf. also MSZKT (256), which
gives kapcsol alone as one of the synonyms of felfog 'grasp'), and it can also be replaced by
a large number of other verbs used in literal or figurative senses without changing the
meaning of végre.
Idioms also resemble single words in that they can co-occur with a more or less
limited range of open-class or closed-class words (look daggers + at sb,
statement/description/account/theory, etc. + hold water) (Cowie 1998c: 221, Fernando
9
The optionality of a word does not necessarily mean that it is not part of the idiom. The position of this
word (initial or final versus medial) in a given string is crucial in configurational languages.

28
1996: 56-59, Moon 1998a: 116-19).10 Cruse (1986: 38-39) draws our attention to the
similarity between idioms and (free) phrases as well, showing the occurrence of idiom-
internal inflections and transformational variants (leg-pull), and he is more inclined to view
idioms phrase-like than word-like. Arnold (1986: 174-77) also emphasizes the similarity
between idioms and phrases: the more lexicogrammatical variation is possible, the more
phrase-like the expression is.
(Simple or complex) words could be placed on an "idiomaticity scale", which is
viewed here merely as a measure of the resemblance that words bear to idioms. The word
kapcsol, used in the sense given above, is certainly closer to idioms than gyurmázik 'model
with Plasticine', since it is metaphorical. Like idioms, words can be used figuratively, and
they can also be motivated by the same metaphor as idioms. For example, the ANGER IS AN

OPPONENT metaphor is illustrated by examples such as I'm struggling with my anger and
I'm finally coming to grips with my anger, where the first example contains a single word
and the second example an idiom (Lakoff 1987: 391). As we can see, words can share with
idioms the property of figurativity. Even higher up on the "idiomaticity scale" we find words
which derive from or are related to idiomatic expressions. The English verb carpet 'criticize'
is related to the idiom be on the carpet, and finger 'accuse' is related to point the finger at
sb. Hadrovics (1995: 159) claims that the word kosár can be used in the same sense as the
idioms kosarat ad 'turn sb down' and kosarat kap 'be turned down'. A slightly different
example is csavaros, given by MSZKT (5) as a synonym of agyafúrt 'cunning', which seems
to be related to the restricted collocation csavaros eszû (literally 'screw-minded/witted').
In an attempt to delimit the scope of phraseology the boundaries we draw will be to
some extent artificial. The above examples show the inappropriateness of sharp boundaries,
though a comprehensive study of the properties of conventional expressions may help us
distinguish phraseological expressions from other types of units. A study of the metaphors
underlying idioms will be more illuminating, if simple or complex words are also considered.

10
The list of nouns that can function as the subject in combination with hold water is taken from Cowie
(1998: 221). Other subject nouns that occur in dictionary examples and/or explanations are argument
(CCDI: 412, CIDI: 416, LID: 372, ODEI: 287, PDEI: 25, CCED: 1886, OALDb: 1460, LDCE: 681, CIDE:
676), reason (CCDI: 412, ODEI: 287, CIDE: 676), application (CCDI: 412), opinion (CIDI: 416), notion
(LID: 372), plan (LID: 372), idea (LID:372), explanation (ODEI: 287, PDEI: 25, LDCE: 681, CIDE: 676),
excuse (ODEI: 287, OALDb: 1460), belief (ODEI: 287), need (ODEI: 287), alibi (ODEI: 287, EI: 169,
CIDE: 676). In our corpus the idiom occurs only with values above the significance threshold (five times),
arguments is found twice, while interpretation, comparisons, exclusions, reservations and model occur only
once.

29
Note that PDEI and Smith (1943) contain some words used figuratively.11 Handbooks
intended for foreign learners frequently include single words in their inventory of idioms.
Ball (1958) is a good example. Hadrovics (1995: 28) also includes words, though he uses
the label "frazéma/szólás értékû magányos szó" 'single word functioning like a
phraseme/idiom' for metaphorical and metonymic examples such as zaboláz 'bridle',
bábáskodik 'be in at the birth of sth', szakáll 'man', konty 'woman', hárpia 'wife' (Hadrovics
1995: 72, 159, 222, 224-25). In other words, he distinguishes them terminologically from
genuine idioms and other word combinations, which are called "frazéma" 'phraseme' or
"frazeológiai kapcsolat" 'phraseological combination', but this distinction is not maintained
when the discussion focuses on the semantic aspects rather than on the structure of word
combinations (Hadrovics 1995: 29, 177, 222). Cross-linguistic studies can also benefit from
a scrutiny of words. The Hungarian idiom özönvíz elõtti and the English word antediluvian
have similar (though not exactly the same) meanings and they have the same underlying
metaphor. But one is a polymorphemic word, and the other is a two-word unit.
We do not accept Hadrovics's (1995: 28, 68) argument that the verbs kimúlik 'pass
away' and letesz vmirõl 'give up sth' should be considered idioms, simply because they are
etymologically derived from the idioms kimúlik e világból (literally 'pass out of this world')
and leteszi elméjét vmirõl (literally 'take one's mind off sth'). There is no synchronic
relationship between the verbs and the expressions, since the expressions are no longer
current. Obviously, this does not mean that diachronic considerations are always irrelevant
in phraseological studies, but synchronic relations are given priority. We will consider the
above verbs idioms, because combinations of a co-verb with a verb will be treated as two-
word units.
Compounds lie on the boundary between single-word and phraseological units. It is
not surprising that researchers are not uniform in their treatment of compounds. Moon
(1998a: 2) is not concerned with compounds. Gläser (1986: 15, 84-85) states that idiomatic
compounds are dealt with in word-formation rather than phraseology but discusses phrase
words such as hand-me-down, forget-me-not and stick-in-the-mud, whereas Fernando
(1996: 41), Makkai (1972: 164-68) and some others include compounds in the scope of
phraseology. Lipka (1992: 79-80, 96) claims that phraseology is concerned with phrasal
lexemes and distinguishes these from compounds but considers compounds such as callgirl,

11
For example, PDEI (60) contains the word rat, and Smith (1943: 199) lists fleece.

30
and holiday idioms. Most Hungarian phraseologists are reluctant to view compounds as a
relevant phraseological category (Juhász 1980: 90, O. Nagy 1985: 8, Földes 1987: 14).
Compounds and phrases are not always easy to distinguish. Several criteria have
been offered (spelling, stress, syntactic isolation of the first element, i.e. the first member of
a compound cannot be modified independently or pluralized, the possibility of inserting
words, etc.), but none of them has been found reliable (Arnold 1986: 112-16, Lipka 1992:
83-84, Bauer 1998, Benson et al. 1986b: 256). Phrases can be transformed into
compounds, a category called phrase words, e.g. their come-and-fight-me attitude (the
example is taken from Quirk (1985: 1566)). Compounds and idioms also share many
properties. Compounds can be viewed as the habitual co-occurrences of their constituents,
and they can be classified into word classes, as can many idioms (Fernando 1996: 41).
Furthermore, compounds can be figurative (pick-me-up) and can derive from phrase-length
idioms (break the ice  ice-breaker, rock the boat  boat-rocking, carry on  carry-on)
(Fernando 1996: 41). Other examples are head-scratching, mind-boggling, cold-shoulder,
tongue-in-cheek, behind-the-scenes, etc. (CCDI: 197, 261, 347, 495, 335, CIDI: 254, 76,
397, 339, LID: 162, 230, 356, 296). Jackendoff (1997: 164-66) also emphasizes the
parallels between idioms and compounds: compounds and idioms can contain unique words
(cranberry, run amok), they can have a word "in the wrong syntactic category" (aloha
shirt - a word used as a greeting in a position typical of nouns and adjectives, in the know -
verb in a position typical of nouns), and they can have words of the right syntactic category
but with the "wrong meaning" (strawberry, eat humble pie).12 Földes (1987: 32) notes that
a compound in one language can evoke the same image as an idiom in another (the German
idiom unter vier Augen corresponds to négyszemközt 'in private' in Hungarian). Therefore,
we will include compounds in the scope of phraseology. Although O. Nagy (1985: 8)
discards compounds, he is aware of the similarity between compounds and idioms: their
ready-made, prefabricated nature, as opposed to free combinations, which are constructed
on the spot. Furthermore, his well-known etymological dictionary of idioms contains some
compounds (ágrólszakadt 'down-and-out', cserbenhagy 'leave sb in the lurch') (O. Nagy
1999: 37, 84).

12
Jackendoff (1997: 165) claims that straw has the wrong meaning, because strawberry has nothing to do
with straw. However, an alternative solution would be to assume homonymy: there are two words, straw1
and straw2, and one of them occurs only in strawberry.

31
Since many compounds are not compositional, unpredictability of meaning is not
sufficient to treat a compound as an idiom. Compounds occupy various positions on the
cline of idiomaticity. Those deriving from phrase-like idioms (ice-breaker, boat-rocking)
will be classified as idiomatic. Moon (1998a: 88) holds that the key factor in distinguishing
noun compounds from nominal idioms is the defectiveness (i.e. idiosyncrasy) of the latter. 13
Defectiveness can be syntagmatic, inflectional or collocational. Syntagmatically noun phrase
idioms tend to be restricted to certain clause positions. 14 For example, Moon (1998a: 88)
has found that ivory tower is used as an object or prepositional complement. Hungarian
examples can also be found. The idiom fû-fa 'every Tom, Dick and Harry' is not normally
used as the subject of a sentence. Note, however, that Moon is concerned with grammatical
idiosyncrasies of the unit as a whole, not with structural defectiveness within the unit. But
grammatical restrictions on the unit as a whole are also found in the case of non-idiomatic
compounds such as high-heeled, which can be used only attributively (cf. CCED: 795).
Apart from morphological and syntactic criteria, semantic considerations will also
help identify a phraseological unit as an idiom. Moon (1998a: 88) explains that word
combinations classified as fixed expressions in her study tend to be metaphorical and
evaluative rather than descriptive or denotative. Some examples of idiomatic compounds
are given here for illustration: sea change 'a strong and noticeable change in a situation'
(OALDb: 1150), grass roots 'ordinary people in society or in an organization, rather than
the leaders or people who make decisions' (OALDb: 561), path-breaker 'someone who
achieves something path-breaking' (CCDI: 291), rat race 'the way of life of people living
and working in a large city where everyone competes in an aggressive way with each other
in order to be more successful, earn more money' (OALDb: 1050), guinea pig 'a person
used in medical or other experiments' (OALDb: 573), crystal clear 'very easy to understand;
completely obvious' (OALDb: 304). Hungarian examples include cserbenhagy 'leave sb in
the lurch', égimeszelõ 'spindle-shanks', pálfordulás 'sb's road to Damascus', álomszuszék 'lie-
a-bed', boldog-boldogtalan 'every Tom, Dick and Harry' and many compounds implying

13
In descriptions of defectiveness researchers usually focus on idiosyncrasies internal to the phraseological
unit, presumably because restrictions on preceding determiners or on co-occurring grammatical schemas
can also be found with single words and in free combinations. Moon's (1998a) treatment is more
comprehensive and justifiable inasmuch as idiomatic and non-idiomatic compounds are difficult to
distinguish on the basis of internal defectiveness.
14
Given that Hungarian word order is relatively free, it would be more appropriate to use the term
"syntactic" instead of "syntagmatic" and to refer to restrictions on syntactic functions rather than restrictions
on clause positions in discussions of Hungarian idioms.

32
comparison (jéghideg 'ice-cold', pofonegyszerû 'dead easy', sziklaszilárd 'rock-solid',
mézédes 'honey-sweet', etc.).
Idiomatic compounds are not always labelled "idiom" in dictionaries. Some
dictionaries will list a particular idiomatic compound in the entry for one of its members,
others will have separate entries for the same compound. For example, in OALDa (218)
cold comfort is labelled "idiom" in the entry for cold, whereas in OALDb (231) it is a
separate entry.
Having discussed the lower limit of phraseology, we will address the problem of the
upper limit. Fernando (1996: 41) claims that the upper limit for formulaic expressions is the
complex sentence. Hadrovics (1995: 28) and Juhász (1980: 85) hold the same view.
Coulmas (1994: 1292), however, classifies ritualistic formulas as types of formulaic
language. Ritualistic formulas can span several sentences. O. Nagy (1985: 7) notes the
similarity between certain idioms and children's rhymes. Makkai (1972: 43) claims that
children's counting rhymes are idiomatic inasmuch as they have special institutionalized
functions and they are noncompositional. Due to its special intonation it is unclear whether
a counting rhyme consists of one sentence or more than one sentence, but it can be of
considerable length (e.g. Ecpec kimehetsz... [beginning of a counting rhyme]) and it has a
frozen structure. Therefore we will include in phraseology constructions longer than a
complex sentence. Very long constructions are not common because of the pressure that
they exert on memory.
Decisions concerning the scope of phraseology will also depend on theoretical
assumptions about the place of this field in relation to other fields, especially lexicology
(Földes 1987: 15-16). Phraseology can be subsumed under lexicology, but this may bias the
researcher in favour of excluding sentence-like items, since lexicology is primarily a study of
words (Juhász 1980: 91, Gläser 1986: 103). The same view is referred to in the following:
"Idioms of the types considered throughout this Introduction [...] occupy syntactic units
longer than the word but smaller than a complete simple sentence. Indeed, many linguists
would take the view that nothing more extensive should concern the student of
idiomaticity" (Cowie et al. 1993: xiv). The traditional "division of labour" between a fully
regular syntax and an irregular lexicon has also contributed to the separation of words and
sentences. Alexander, as reported by Howarth (1996: 6), and Gläser (1986: 22) consider
phraseology as a sub-discipline of lexicology, while Lipka (1992: 79) argues for not treating

33
phraseology as part of lexicology. Given that phraseological units display both word-like
and phrase-like/sentence-like properties and can have the same length as an average phrase
or sentence, it is probably more appropriate not to consider phraseology as a sub-discipline
of lexicology, morphology or syntax, but to see it as both independent of and overlapping
with them. We will return to the problem of the scope in 2.8 below.

2.5 Weinreich

While Hockett explores language within a structuralist framework, Weinreich's


(1969) approach is generative. Another difference between Hockett and Weinreich is that
Weinreich considers a broader range of expressions, idioms and non-idioms alike. Idioms
are recalcitrant to analysis in the generative tradition due to their partly word-like (i.e.
irregular) partly phrase-like (i.e. regular, productive) nature. Since they are not fully regular,
they are usually placed in the lexicon, the storehouse of idiosyncrasies.
Weinreich (1969: 76) emphasizes the arbitrary nature of idioms, separating
motivation from compositionality. He describes the relation between the idiomatic and the
literal meanings as unsystematic. As we have seen, cognitive grammar also separates
compositionality and motivation, and in this respect it is not different from Hockett's and
Weinreich's approaches. However, cognitive grammar's motivation is an important facet of
meaning, while for Hockett and Weinreich it is of secondary importance. This is partly
because they are unable to free themselves from the building-block metaphor of linguistic
structure, partly because they do not investigate conceptual metaphors.
Though Weinreich subscribes to the central ideas of (early) generative theory, his
approach shows many insights that reappear later in cognitive theories. He finds it
inadequate to put idioms in the lexicon (his term is "dictionary") either in the form of
phraseological entries or in the form of listing the constituent words separately and
providing the idiomatic senses of these words (Weinreich 1969: 54-57). Instead he
postulates an idiom list against which terminal strings can be matched, and the procedure
results in phrases with nonliteral interpretations and arbitrary lexicogrammatical restrictions,
i.e. idioms (Weinreich 1969: 57-59). However, Weinreich (1969: 57, n. 21) notes that the
difference between separating the lexicon from an idiom list and dividing the lexicon into an

34
idiom and a non-idiom part is not crucial. The matching procedure is optional, since idioms
have literal counterparts (Weinreich 1969: 59). Apart from words, the units whose inclusion
in the lexicon is justified are those without literal counterparts: extragrammatical
expressions (by and large)15 and binomials with two unique words (spick and span)16
(Weinreich 1969: 68-69). But they are not regarded as idioms because they are claimed to
have no literal interpretation (Weinreich 1969: 44, 68). Chafe (1968: 116) points out rightly
that the listing of extragrammatical idioms in the lexicon cannot account for the fact that in
some of them particular constituents can take inflections. We have found that in go for
broke the verb go can appear in the inflectional forms in which it appears outside idioms: In
London's West End there is a reluctance to take risks with new plays while going for broke
on musicals (CCDI: 50), The team was aware that we needed to make changes, so we went
for broke (LID: 138).
In an attempt at a unified treatment of collocations and idioms, the emphasis is laid
on meaning and context, which shows the influence of Russian phraseologists. A distinction
is made between idioms and non-idiomatic phraseological units. The former involve a
"reciprocal contextual selection of subsenses" (red herring 'phoney issue'), while subsense
selection is unidirectional in other phraseological units (blind alley) (Weinreich 1969: 42,
45). The subsense 'phoney' of red is selected in the context of herring, and the subsense
'issue' of herring is used only in the context of red. One of Weinreich's merits is that he
realizes the analyzability of idioms, but the unmotivated nature (opacity) of his examples
(by heart, red herring, shoot the breeze), the non-occurrence of the words in their idiomatic
meanings outside the given idiom and the architecture of his generative theory ultimately
prevent him from assigning independent senses to idiom constituents. The lexicon is not an
appropriate place for idioms, because a) if idioms are lexical units, we cannot account for
idiom-internal inflections (shot the breeze versus *shoot the breeze-d) and transformational
variants; b) if idiom constituents are lexical items, nothing can prevent our fully regular
syntax from generating *red issue, where red has the idiomatic sense 'phoney'. Therefore,
15
We will use the label "extragrammatical" for idioms violating the grammar of free combinations, while
retaining "ill-formed" for those violating the lexicogrammatical structure of existing idiomatic word
combinations (*a huge fish, *the bucket was kicked).
16
Weinreich (1969: 69) gives three examples to illustrate the latter type (spick and span, hem and haw, tit
for tat), but none of these is a good example according to NSOED, which claims that spick, hem, haw and
tit (all having the same word class as in the idioms) are used outside phraseological units as well. Spick is
an abbreviation of spick and span; hem is used in the sense 'clear away with a cough', though this sense is
rare; haw means 'utter "haw", especially as an expression of hesitation', and a note says "frequently in hum
and haw"; finally, tit is used in the sense 'a light stroke or tap, a slap' (NSOED: 2982, 1217, 1199, 3320).

35
Weinreich (1969: 58) chooses to place idioms in an idiom list, but in this list the idiomatic
meaning is linked with the idiom as a whole, not with the constituents (shoot the breeze
'chat idly'). This is odd, given his definition of idiom and his separation of the idiom list from
the lexicon.
The definition itself is not perfect, since it is not applicable to unanalyzable idioms,
in which the constituent words do not seem to possess independent idiomatic senses (the
bee's knees). Furthermore, Weinreich (1969: 41) classifies blind date as an idiom, since both
constituents are polysemous, and one subsense is selected in construction with the other
constituent. Though blind date is given in several idiom dictionaries (CIDI: 38, LID: 79,
PDEI: 125, EI: 64), from the viewpoint of the directionality of subsense selection it is not
different from collocations such as make a claim, in which both constituents are polysemous
and subsense selection is bidirectional. Instead of the directionality of subsense selection, it
would be more helpful to emphasize the figurativity as well as the degree of motivation of
the subsenses.
Makkai (1972: 49) criticizes Weinreich, saying that "subsense-assigning can be done
only ex post facto, after the meaning of the idiom is already known to the analyst". This is
true, but analyzability is a relevant notion and can have an effect on the lexicogrammatical
variability of idioms (cf. 2.24.3).
Weinreich (1969: 44) holds that not only idioms but other phraseological units are
also ambiguous between a literal and a nonliteral reading, i.e. all phraseological units have
literal interpretations. Therefore, contrary to Fernando's (1996: 6-7) claim, spick and span
is not a phraseological unit in Weinreich's terminology. The discrepancy between the literal
and nonliteral senses varies, with some idiomatic word combinations (bacon and eggs)
being closer to their literal reading than others (red herring) (Weinreich 1969: 30-31, 43).
In other words, Weinreich assumes degrees of compositionality, which is at odds with the
formalist view. Although the focus is on idioms and their grammatical deficiencies,
Weinreich is atypical in the generative tradition in that he examines phraseological units
other than idioms and also draws our attention to lexical restrictions in his discussion of the
role of context, the extreme case of restriction being the impossibility of any substitution
(blind date/*appointment/*rendezvous, rain/*pour/*snow cats and dogs) (Weinreich 1969:
38, 41). Apart from the possibility of literal interpretation, special semantics and
grammatical idiosyncrasies are required of phraseological units. This is a rather narrow view

36
of phraseology. Weinreich (1969: 71) comments that there is nothing phraseological about
"expressions that are distinguished by nothing but their familiarity, and have no grammatical
defects or semantic properties related to their specialized subsenses [...] they are merely
stable and familiar". Among the examples to illustrate non-phraseological word
combinations he lists the proverb Two wrongs don't make a right. Moon (1998a: 19, 22)
distinguishes semantic and pragmatic noncompositionality, and considers proverbs
pragmatically specialized. However, metaphorical proverbs are also semantically
noncompositional. It seems that Weinreich would exclude many formulae such as Good
morning.

2.6 Analyzability

Analyzability means that the meaning of the whole expression can be distributed
over the constituents. This is the case in spill the beans, because it is possible to see the
contribution that the individual words make (spill 'reveal' + the beans 'secret information').
Similarly, certain parts of the idiomatic meaning can be attached to the constituent words in
grasp the nettle ('tackle...with determination and without delay' + 'something difficult'),17
back the wrong horse (back 'support' + the 'the' + wrong 'wrong' + horse 'person') and mend
one's fences (mend 'improve' + one's fences 'one's relationship').
It is important to note that full analyzability does not require us to attach each
"chunk" of the idiomatic meaning to one of the constituents. It is enough for each
constituent to carry some idiomatic meaning. In other words, examples such as gin and
tonic will be treated as fully analyzable, since each constituent has an independent meaning.
Furthermore, in an ideal case the constituent in question is a word, as in back the wrong
horse, but full analyzability is often taken to apply to idioms in which the (short) noun
phrase, rather than the determiner and the noun separately, carry idiomatic meanings, as in
grasp the nettle. These points are not stated explicitly in the literature, but they are implied.
Langacker (1987: 24, 93) argues that umbrage in the idiom take umbrage at carries
the sense 'offence', and the cat is out of the bag can also be broken down into the cat 'the
information', out of 'out of' and the bag 'concealment'. Similarly, Gibbs (1990: 423, 1994:

17
This is a slightly modified form of the paraphrase found in Moon (1998a: 23).

37
279) claims that let off steam can be analyzed into let off 'release' and steam 'anger'. Note
that analyzability is not intended to entail the claim that the given word in the given
idiomatic meaning can always be used outside the idiom. As opposed to the above
examples, idioms such as shoot the breeze, over the moon, a dog's breakfast and the bee's
knees remain mysterious and unanalyzable.
We have asked a number of native speaker informants in order to see whether they
can attach independent meanings to various idiom chunks. The answers are shown in 11.2.
The idioms have been selected partly from those that are discussed in the literature, partly
from the author’s own data. As can be seen, rarely do we find full agreement among the
speakers. The most obviously analyzable parts are those that can be replaced by their
meaning paraphrases while keeping the context. For example, cool in lose one’s cool was
judged to be analyzable, and this is partly due to the fact that the verb lose can be used with
the meaning paraphrases ‘temper’, ‘calm’, ‘patience’. It is interesting to note that strings
was thought to have no independent meaning, when pull strings is one of the analyzable
idioms found in the literature. The answers did not meet our expectations in the case of cut
little ice and crying for the moon, since we would analyze ice as ‘influence’ and the moon
as ‘the impossible’. It seems that native speakers strongly feel that idiomatic expressions are
units, and they are rather unwilling to attach an independent meaning to the constituents. In
fact, some answers are best regarded as giving the meaning for the whole phrase, rather
than the underlined parts. Two speakers answered ‘see’ for set eyes on him, with only eyes
as the underlined part. In other cases our expectations were borne out, since most speakers
felt that the screws in put the screws on sb, fences in mend fences, or cahoots in in cahoots
with carry independent senses. On the whole our answers would be similar, but we would be
more willing to analyze some expressions (bury the hatchet with bury ‘end’, set the ball
rolling with the ball ‘the situation/process’, crying for the moon with the moon ‘the
impossible’.
Compositionality and analyzability are viewed by many linguists as the same notion,
and noncompositionality is often understood to imply that speakers are unable to analyze
the meaning of the composite structure into smaller chunks that correspond to the
component parts. It is not clear, for instance, how it is possible to attach parts of the
idiomatic meaning 'very happy' to the constituents over, the and moon. Thus, over the moon
is noncompositional. We follow Nunberg et al. (1994: 498) and Langacker (1987: 448),

38
who draw a distinction between the two notions, though Nunberg et al.'s "compositionality"
is synonymous with our "analyzability" and their "conventionality" corresponds to our
"compositionality".
Compositionality can be interpreted in terms of production and comprehension as
well. Viewed from the production side, a (fully) compositional expression is one that
speakers familiar only with the component parts would be expected to be able to generate
and produce with the regularity with which it is used in the language (Nunberg et al. 1994:
495). Many combinations of words that are not regarded as idioms would fail this type of
test. It is not at all certain that the compound trade secret would be easily generated by
speakers to describe information about manufacturing a product that is kept secret. It could
just as well be referred to as manufacturing secret or production secret (cf. Hungarian
gyártási titok 'production secret'). In terms of language comprehension, the meaning of a
fully compositional unit is predictable in an uninformative context, since the meaning of the
whole unit is a regular compositional function of the meanings of the parts. Thus, a speaker
is unlikely to predict that over the moon means 'extremely happy and excited', spill the
beans means 'tell sb sth that should be kept secret or private', a dog's breakfast means 'a
mess' and the bee’s knees means 'an excellent person or thing' (OALDb: 825, 1244, 371,
98). Nunberg et al. (1994: 495) point out that the typical context for idioms is rarely
uninformative, and the literal meaning combined with the context can help the listener to
guess the intended meaning.
Langacker (1987: 448) notes that the analyzability of an expression is determined at
the level of the individual expression, while compositionality concerns the regularity of the
compositional relationship. This implies that compositionality presupposes the existence of
other contexts where constituents with the same meaning are combined in the same way as
in the word combination we are examining. Mend one's fences is analyzable ('improve' +
'one's relationship'), but noncompositional, since fences does not occur elsewhere in the
same sense.
Before the emergence of cognitive grammar idioms were treated as unanalyzable,
but there were some linguists who considered certain idiom constituents to have
independent meanings. Nunberg et al. (1994: 499, n. 13, 505, n. 17) mention a few
researchers who recognized that parts of idioms can have idiomatic meanings. Among those
listed we find Weinreich (1969), who shows that certain idiom chunks can have idiomatic

39
meanings. Although not included in the list, Smith (1943: 184-85) also gives some examples
of words whose obsolete meanings are preserved in idioms (mind in the sense 'memory' is
used in keep in mind). Some linguists currently working outside the framework of cognitive
grammar also acknowledge the analyzability of at least some idioms (cf. Radford 1988: 319,
442, Jackendoff 1997: 168). We have syntactic evidence that analyzable idioms exist.
Nunberg et al. (1994: 500-502) show that modification by relative clause (Pat got the job
by pulling strings that weren't available to anyone else) and other transformations can
apply to parts of idioms, and therefore these parts must have independent interpretation (cf.
also 2.10).
It is not known what per cent of idioms are analyzable. Nunberg et al. (1994: 499)
claim that most phrasal idioms are analyzable, and Gibbs (1994: 278) also talks about many
idioms being analyzable, but we have not found any statistics in the literature.
Unanalyzable idioms do occur (shoot the breeze, kick the bucket), and analyzable
ones display various degrees of analyzability, as Gibbs (1990: 423, 1994: 278-79) points
out. He reports research done by Nayak and himself on how people rate the contribution of
the individual words of certain idioms to the overall figurative interpretation (Gibbs 1990:
423-24, Gibbs 1994: 279). The findings suggest that people's intuitions are reasonably
consistent and that there is a scale of analyzability, with 1) highly analyzable idioms (pop the
question, miss the boat), 2) unanalyzable idioms (kick the bucket, shoot the breeze) and 3)
abnormally analyzable idioms (carry a torch, hit the panic button). In carry a torch it is
possible to link the meaning 'love' to the word torch, only because of the conventional
association between love and fire (cf. Kövecses and Szabó's (1996: 333) examples - The
fire between them finally went out, I am burning with love, She carries a torch for him, The
flames are gone from our relationship - as well as expressions such as the flames of
passion, an old flame). It is less likely that we would identify part of the idiomatic meaning
with carry, i.e. it seems arbitrary to link the 'be in' part of 'be in love' with this verb. It must
also be noted that the idiom is used especially when love is unrequited, but this is not
expressed by any of the parts. Nunberg et al. (1994: 496-97) focus on analyzability alone to
subclassify idioms into two groups: analyzable (pull strings 'exploit + personal connections')
and unanalyzable (shoot the breeze). The classification offered by Gibbs is more
appropriate, since in partially analyzable idioms only some of the constituents carry
independent meanings. In laugh one's head off the verb laugh has an independent meaning,

40
but the rest of the idiomatic meaning ('a lot', 'for a long time') cannot be distributed over the
other constituents. We therefore assume three levels of analyzability: 1) fully analyzable
(pull strings), 2) partially analyzable (know sth like the back of one's hands 'know + sth +
very well') and 3) unanalyzable (round the bend 'crazy'). Since prototypical idioms are
unanalyzable, shoot the breeze has a higher degree of idiomaticity than pull strings, at least
from the point of view of analyzability.
In phraseological studies analyzability is generally taken to be the decisive criterion
for distinguishing idioms from restricted collocations (cf. Howarth (1996: 23) and Cruse
(1986: 40), who states that "each lexical constituent [of a collocation] is also a semantic
constituent"). It can indeed be used to distinguish idioms from collocations and many
formulae, but, as we have seen, analyzable idioms do occur. Analyzability is therefore not a
reliable criterion. Analyzability is not reliable in separating free combinations from
collocations either, since both are fully analyzable.
From a lexicographic viewpoint analyzability means that a given idiom need not be
listed at the end of a dictionary entry. If an idiom constituent has a meaning of its own, and
that constituent is a polysemous word in the language, then the lexicographer may wish to
consider the option of including the idiom in the relevant section of the entry. If, for
example, one of the meanings of take is 'accept', then idioms containing this verb in the
same sense could be given together with other non-idiom examples under the meaning
'accept'. Although in both take sth as it comes and take it or leave it the verb take means
'accept', in OALDb (1325) they are found at the end of the entry among other idioms.
Furthermore, take sth lying down, take sth on the chin, take sth in good part, take the bad
with the good, which are given in OALDb (741, 204, 922, 78) in various entries, could just
as well be listed under the meaning 'accept' in the entry for take. This solution would have a
more beneficial effect on language learners, since they could see what is common in a given
group of idioms.

2.7 Chafe

Before criticizing Katz and Postal's and Weinreich's views, Chafe (1968: 111) gives a
list of those properties of idioms that a generative grammar is expected to account for.

41
These are as follows: a) idioms are noncompositional, and the idiomatic meaning is similar
to that of a single word; b) idioms are subject to transformational restrictions (*The bucket
was kicked by Sam); c) some idioms are extragrammatical (by and large, trip the light
fantastic); d) a particular idiom that has a literal counterpart is more frequently interpreted
idiomatically than literally. Chafe (1968: 110-11) believes that failure to provide plausible
explanations for the phenomenon of idiomaticity has sounded the death knell for generative
syntax, especially the Chomskyan paradigm. In Chafe's opinion, Katz and Postal's theory
can handle only the property mentioned in "a" above, but it collapses when faced with the
other peculiarities of idioms (Chafe 1968: 112, 114-15). Weinreich's solution is likewise
dismissed.
Chafe (1968: 118) gives priority to semantics, which is his starting point. He
proceeds from semantics to phonetics, since semantic units have to be converted into
speech sounds, and sets up two intermediary levels, one of which (the post-semantic level)
is roughly equivalent to Chomskyan surface structure (Chafe 1968: 120). He claims that
kick the bucket cannot be passivized, because it is a single semantic unit with the meaning
'die', and die is an intransitive verb, which cannot undergo the passive transformation (Chafe
1968: 122). In contrast, pull sb's leg permits the passive, because semantically it is a
transitive verb. Moon (1998a: 109) cites evidence that supports a similar claim to some
extent. However, Gibbs (1994: 272-73) reports the finding by Nunberg that give up the
ghost, throw in the towel and pop the question can be passivized even though their literal
paraphrases are intransitive verbs (die, resign and propose).
Chafe (1968: 122) makes an attempt to explain why kick can inflect but bucket
cannot in kick the bucket. Essentially, he posits a rule that reverses the historical process of
idiom formation, and claims that idioms are semantic units interpreted figuratively at the
semantic level but they are arrangements of semantic units interpreted literally at the post-
semantic level (Chafe 1968: 121). The word bucket cannot inflect, since it is not present at
the semantic level, being introduced post-semantically. However, the same argument can
apply to kick as well, which is also introduced post-semantically in Chafe's terminology;
therefore, Chafe (1968: 122) is unable to explain why Sam has kicked the bucket is correct,
but *Sam has kick the bucket-ed is not. Thus, Chafe's explanation of the transformational
deficiency of idioms is not convincing. He claims that the difference between the semantic
and the post-semantic levels is not sharp, so that the grammatical variation of some idioms

42
is the result of attaching semantic units such as "passive" to post-semantic units, hence the
correctness of The hatchet was buried (Chafe 1968: 124-25). However, only a tentative
explanation is offered as to what determines whether an idiom behaves like kick the bucket
or bury the hatchet. Chafe (1968: 125) attributes their syntactic difference to their
difference in terms of motivation: bury the hatchet is more motivated than kick the bucket.
Chafe (1968: 123) claims that the greater text frequency of the idiomatic meaning is
due to the usefulness of the figurative sense and to the simplicity of idiomatic units as
opposed to the complexity of their literal counterparts. Howarth (1996: 19) rightly argues
that only a fraction of idioms have single-word synonyms, but in generative theories
noncompositional expressions are units, and as a result of the building-block metaphor,
units are always simple. Cognitive grammar considers idioms as gestalts, which are simpler
than the component parts, but their simplicity does not mean that speakers are not aware of
any internal complexity. One mark of the simplicity of idioms is the availability of single-
word synonyms, but, as we have seen above (cf. 2.3), single-word synonyms cannot always
be found.
Finally, we will note that Chafe (1968: 126) follows Hockett (1958: 173) in treating
grammatical categories (progressive, perfect, passive) as idioms.

2.8 Idiosyncrasy

Chafe lists four properties of idioms, of which we will examine here only
extragrammaticality. Noncompositionality and single-word synonymy have already been
discussed, while transformational restrictions and figurativity will be dealt with later.
Conventional expressions, especially idioms, have often been regarded as
idiosyncratic in terms of their lexicogrammar and semantics. Smith's (1943: 168) definition
of the term reflects this view: "I shall use 'idiom' [...] in its narrow sense, meaning the
idiosyncrasies of our language, and, above all, those phrases which are verbal anomalies,
which transgress, that is to say, either the laws of grammar or the laws of logic". The ill-
formedness of certain lexical and grammatical variants (i.e. lexicogrammatical restriction) is
one sign of the idiosyncratic nature of phraseological units. The ill-formedness of the
citation form is another. There is no generally agreed terminology for expressions which

43
violate the syntactic rules of the language, since Cruse (1986: 38) uses "asyntactic idioms",
Moon (1998a: 21) refers to them as "ill-formed collocations", Fernando (1996: 30-31)
labels them as "noncanonical idioms", in Jackendoff's (1997: 211) terminology they are
"nonsyntactic idioms", and Fillmore et al. (1988: 505) describe them as "extragrammatical
idioms". Examples include phraseological units such as by and large, all of a sudden, go for
broke, long time no see, I kid you not, believe you me, the biter gets bit in English and
kötélnek áll 'accept', oda se neki 'never mind', egy szó mint száz 'in brief', jobbnál jobb 'one
better than the other', mellre szív 'take sth to heart', se lát, se hall 'sb neither sees nor hears'
in Hungarian.
The idiosyncrasy of idioms has been noted by many scholars, but in generative
grammar there is a sharp division between a regular syntax and an irregular lexicon, as well
as a requirement that the grammar should be maximally economic, i.e. regularities should be
captured by rules alone and the lexicon should contain only the irregularities. Thus, the
idiom go for broke is problematic, because it cannot be generated in a regular way, yet it
displays a certain amount of regularity; namely, go can take inflections. Cognitive grammar
sees no sharp division between regular and irregular (i.e. idiosyncratic) features and lays no
special emphasis on economy (Langacker 1987: 40-42, 45-46). Thus, a schema can be
extracted from examples such as in the know, on the make, on the go, since they have the
structure 'preposition + the + verb'. The schema is far from being productive, but it is still a
schema.
While not disputing the existence of extragrammatical expressions, Fernando (1996:
34) claims that the majority are grammatical, but Nunberg et al. (1994: 515) claim their
number is "not so small". Moon's (1998a: 61-62) statistics show that extragrammatical units
comprise less than five per cent of her database word combinations, which seems to
corroborate Fernando's claim. However, most of Moon's (1998a: 61) data are based on
primary classifications, i.e. with each phraseological unit assigned to only one subcategory.
If dual classification is permitted, which can indeed capture the overlapping nature of the
categories, then Moon (personal communication) finds 8.5 per cent of units are
extragrammatical. This is not a high figure, but it certainly discourages us from viewing
extragrammaticality as an occasional, insignificant phenomenon.
Various subtypes can be distinguished within the subcategory of extragrammatical
units depending on the nature of ill-formedness: odd phrase structure (give sb what for),

44
ellipsis (come to think of it), inflections (dog eat dog), archaic mood (come what may) or
strange word class (in brief) (Moon 1998a: 81-82). Some might wish to consider
extragrammatical those combinations that violate selectional restrictions and truth
conditions such as put one's best foot forward, live a lie, white lie (cf. also borotvaélen
táncol 'be on a razor-edge', alszik, mint a bunda 'sleep like a log', fából vaskarika
'nonsense', kétbalkezes 'be all thumbs), a decision that rests ultimately on whether we wish
to maintain the autonomy of syntax and what phenomena we prefer to treat as syntactic. We
will not regard the above examples as extragrammatical. Moon (1998a: 83) also discusses
idioms such as bite the bullet in the section on extragrammaticality, since the definite article
has no anaphoric or cataphoric reference (and clearly no situational reference either),
although syntagmatically the idiom is grammatical. In this case the idiosyncrasy of the item
comes from a consideration of the co-text, not from the structure of the idiom. But Moon
(personal communication) did not classify these as extragrammatical in her typology, and we
will also ignore them.
No subclassification of extragrammatical idioms exists according to the degree of
idiosyncrasy, the only exception is Smith (1943). But is idiosyncrasy a graded property? We
claim it is. It has been argued, and not only in cognitive grammar, that grammaticality
judgements cannot always be formulated with a yes or no. The use of symbols such as ?, ??,
*, ** suggests a cline with various degrees of ill-formedness. For example, the idiom
dressed up to the nines seems to be more deviant than in short, since it is extragrammatical
due to the use of the definite article before the numeral and the plural of the numeral. The
expression if it ain't broke, don't fix it also contains two peculiar features: the use of ain't
and broke. Similarly, believe you me is odd, because of the word order and the overt subject
in the imperative. However, it is not clear, for instance, how many degrees should be set up.
The difficulty of measuring idiosyncrasy is partly due to its subjectivity. Smith (1943: 182-
83) establishes two idiosyncratic groups, one in which "the laws of grammar seem to be
openly flouted", and another containing "slighter anomalies". Expressions of the former type
include double negation, the use of prepositions at the end of sentences, double genitives (a
picture of the king's), none of which are idioms in our interpretation. Slighter anomalies are
the use of words in strange word classes (whys and wherefores) or the use of intransitive
verbs as transitive (come it over). We follow Smith in assuming three levels: 1) grammatical,

45
2) slightly extragrammatical (on the make) and 3) heavily extragrammatical (give sb what
for).
Smith (1943), Moon (1998a: 21), Carter (1987: 64), Fillmore et al. (1988: 505) and
Cruse (1986: 37-38) are among those who use the criterion of extragrammaticality in their
classification of phraseological expressions. For Fillmore et al. it is a major criterion. Moon
uses it in order to make a distinction between subcategories. Howarth (1996: 45, 46, 89)
also identifies the extragrammatical group, but syntactic criteria are not regarded as
important in his classification. The property of extragrammaticality is usually taken to divide
word combinations into two groups: grammatical and extragrammatical.
At this point we return briefly to the problem of the scope of phraseology. If a
researcher attaches great importance to idiosyncrasy, then they may consider phraseological
those phrases that are distinguished by nothing else but their odd grammar or illogicality.
This is often the case in books and dictionaries written for non-native speakers, where the
learner's needs may well influence the choice of what to include. Idiosyncrasy can be
approached cross-linguistically, in which case grammatical features that are present in
English but absent in other languages are also "idiomatic". For example, Ball (1958: 2, 4-5)
considers idioms the use of the past tense in It's time we went home, inversion in Had you
come earlier, you would have seen her and the meaning change that results from slight
differences in word order as in It may well be ahead of time versus It may be well ahead of
time. We do not regard these as idioms.
Ball's (1958: 1) definition of "idiom" - "the use of familiar words in an unfamiliar
sense" - is too broad and too narrow at the same time. It seems that the word unfamiliar in
his definition means 'noncompositional' or 'different from the literal/different from the most
common'. This is too broad, because it covers various grammatical peculiarities as well as
single words used figuratively. But it is also too narrow, because it focuses on fully
noncompositional items, disregarding any variation in degree.
Smith's (1943: 170-71, 178-84) treatment of idioms is also loose, since many
grammatical and logical peculiarities are dealt with: the use of prepositions (find a fault in a
person versus find fault with him), the split infinitive, the use of the plural determiner with a
singular noun in these sort of things, the double genitive (a picture of the king's), the
omission of the definite article in at least, etc. These differ from Ball's examples in that
Smith's examples are idiosyncratic from an intralinguistic point of view alone, even if we

46
ignore cross-linguistic comparison. Most of Smith's examples given above are
compositional to a great extent (with the exception of at least); therefore, we do not
consider them to be idioms.
Smith (1943: 187) claims that "there is a certain irrelevance in the human mind, a
certain love for the illogical and absurd, a reluctance to submit itself to reason, which breaks
loose now and then, and finds expression for itself in idiomatic speech". We certainly agree
that several idioms are peculiar, but we do not hold that the human mind is particularly
inclined towards the illogical. As far as semantic idiosyncrasy (opacity) is concerned, we
will see that many idioms are motivated by conceptual metaphors that are deeply rooted in
our everyday experience. Furthermore, examples of folk etymology can be found in idioms,
which runs counter to any "love for the illogical and absurd". A well-known example is
humble pie, whose original form probably contained the word umble, which referred to the
entrails of a deer (DIO: 79-80). Umble pie was a dish usually given to lowly people, while
the lord with his family and guests enjoyed the venison. In the course of time umble lost its
motivation and has been replaced by the word humble. This replacement leads to a higher
degree of motivation, given that the idiom denotes the act of apology and humiliation.
The idiom put a damper on sth means 'to stop [a situation] being as successful or as
enjoyable as it might be' (CCDI: 92), and its variant is put a dampener on sth. Since the
latter variant is found only in recently published dictionaries, we can safely assume that the
original form was damper. The word damper refers to a device on piano strings, which is
used to make the sound less loud (CIDI: 88, DIO: 67), and just as dampers can reduce the
degree of loudness, so the idiom denotes the stifling effect that somebody or something
exerts on the enjoyment of others. The emergence of dampener is probably linked to the
metaphor ENTHUSIASM IS FIRE. This metaphor underlies for example the idiom fire in the
belly. We know that water can extinguish fire, and dampness can slow down or prevent
burning, it is therefore not surprising that dampener has emerged in addition to damper.
The variant dampener may also be motivated by the fact that the verb dampen collocates
with enthusiasm.
As far as grammatical idiosyncrasies are concerned, they often result from ellipsis. A
diachronic study of idioms can show that in several cases the original forms were not
illogical at all. For example, the Hungarian idiom résen van (literally ‘be on slit’) 'be on
guard' is an elliptical form of résen van a füle (literally ‘sb's ears are on slit’) (Hadrovics

47
1995: 167). The English idiom through thick and thin was originally through thickets and
thin woods (Funk 1948: 58).
Although Chafe does not discuss lexical uniqueness, it can be viewed as a lexical
counterpart of the type of grammatical idiosyncrasy we have described, since both consist in
characteristics that are not found outside phraseological units (cf. also Moon (1998a: 80)
for a parallel between the two notions). Lexical uniqueness means that a given word is not
used outside idioms. The English examples to and fro, kith and kin, in fine/good fettle and
the Hungarian examples fittyet hány 'cock a snook', dugába dõl 'fall through', vérszemet kap
'get carried away' all contain unique words. Although Földes (1987: 46) identifies ludas a
dologban 'it's his fault' as a phraseological unit containing a unique word, we consider it a
free combination containing the grammatical collocation of ludas with -ban/ben. (cf.
MSZKT (539, 102, 863, 904), which lists ludas without a dologban as a headword and as a
synonym of various words).
The parallel between lexical and grammatical uniqueness is not perfect. A unique
word in a given combination will usually not be found not only outside phraseological units
in general but also in other phraseological units. The words fro, kith and fettle are
encountered only in the expressions listed above. Concordance lines for fettle are given in
9.1. Other adjectives than good or fine occur only once or twice. According to NSOED
(3450, 25) umbrage and kilter are exceptions, since the former occurs in both take
umbrage at sth and give umbrage to sb, while kilter occurs in out of kilter and in in (good)
kilter. However, in our corpus kilter occurs in the form out of kilter (cf. 9.2), and umbrage
occurs only with take (cf. 9.3). In contrast to lexical uniqueness, a particular grammatical
feature that is "unique" may well occur in more than one word combination. The expression
long time no see does not contain an overt subject, nor does come to think of it. Another
unusual grammatical schema is a preposition followed by an adjective, and this is found not
only in of old but also in in brief and some other word combinations. As we have seen, the
'preposition + the + verb' schema also occurs in more than one idiom.
For Fillmore et al. (1988: 506) expressions containing unique words are necessarily
also unique from a structural point of view. They argue that if a lexical item is unique, the
productive syntactic rules cannot combine it with other words. We need knowledge of the
word class in order to apply the syntactic rules, and with unique words the word class
category is not clear. In cognitive approaches, however, lexical and structural uniqueness do

48
not necessarily co-occur. For example, kith and kin will be analyzed as exemplifying the
regular schema 'noun + and + noun' and in fine/good fettle will be an example of
'preposition + adjective + noun'. These schemas can be extracted from other combinations
and applied to these examples.
Moon's (1998a: 61-62) statistics show that less than 2 per cent of her database
phraseological units contain a unique lexical item, which means that this type of lexical
idiosyncrasy is not common in word combinations. There are various types of lexical
uniqueness. The following classification ignores diachronic aspects: a) some of the
monomorphemic lexical items are unique, because they never occur outside certain word
combinations (kith is found only in kith and kin); b) others have homographs outside
phraseological units (lurch is used as a noun or a verb in free combinations and lurch, a
different word, is found in leave sb in the lurch). There are polymorphemic words whose
morphemic composition is unique: c) because one or more morphemes occur only in the
expression (make amends)18 or d) because the given combination of the constituent
morphemes occurs only in the expression (at gunpoint). The latter type can be viewed as
including the former, since a unique morpheme could imply a unique combination.
Alternatively, it can be viewed as a different category, in which case it includes only those
items whose constituent morphemes can be found outside phraseological units only when
they are not combined. No importance will be attached to viewing it one way or another.
Alternative groupings are possible and a slightly different classification is found in Moon
(1998a: 78-80).19
Lexical uniqueness is not regarded as a graded property in the literature, although an
expression containing two unique words would be more idiosyncratic than one with only a
single unique word. However, we have not found examples with two or more unique
words. But it must be borne in mind that for some native speakers Weinreich's (1969: 69)
examples (spick and span, hem and haw, tit for tat) may have two idiosyncratic
constituents. The example make amends could be considered more idiosyncratic than the
expression at gunpoint, since the former contains a constituent morpheme that does not

18
Amend as a noun is obsolete or dialectal (NSOED: 65).
19
Since the words are uncommon, it may not always be clear whether the native speaker analyzes the given
word as consisting of one morpheme or not. Moon (1998a: 79) does distinguish a polymorphemic group,
where "items have compositional or familiar morphemic structures", but excludes from this group the
expressions go/be haywire, at loggerheads and on tenterhooks, although they seem to be morphemically
composite. NSOED (1619, 3250) lists both logger and tenter.

49
occur elsewhere (amend), while the latter contains morphemes that occur elsewhere (gun,
point). We assume three levels: 1) lexically ordinary, 2) slightly idiosyncratic (at gunpoint)
and 3) heavily idiosyncratic (kith and kin).
There are several reasons why a particular lexical item is unique. Strictly speaking,
unique words are those that are obsolete, borrowed from foreign languages, or simply not
found elsewhere. Often a word will become obsolete because its referent is no longer used
(Moon 1998a: 40). We will also consider unique those items that are absent from the
standard dialect but may occur in various regional varieties. Moon (1998a: 78-79) lists
several examples in what corresponds to our group "a", of which kith is obsolete and
cahoot does not occur except in the idioms kith and kin and in cahoots respectively
(NSOED: 1459, 316). But she also enumerates expressions whose "unique" lexical items do
occur in the standard variety outside the expressions, though most of these words are
restricted to various registers or they are simply rare according to NSOED (Moon 1998a:
79). For example, grist occurs in grist to one's mill and also independently (NSOED: 1146).
The question arises whether such words should be considered unique or not. We hold the
view that native speakers' vocabularies obviously differ, so that what is unique for some is
not necessarily unique for others. Therefore, we will not insist on treating NSOED as the
only arbiter. Some of the other expressions mentioned by Moon (1998a: 79) contain words
which are homographic with other independently occurring words according to NSOED
and could therefore be placed in the homographic group (group "b" in our classification).
Such an example is kilter. It is used in out of kilter, in (good) kilter, and its homograph is
restricted to the register of card games (NSOED: 1488).
Most Hungarian unique words are polymorphemic. Examples where one or more
morphemes are unique include dugába dõl fall through', fabatkát sem ér 'not worth a row of
beans', fittyet hány 'cock a snook', kordában tart 'keep a tight rein', kapóra jön 'come in
very handy', pellengérre állít 'pillory sb'. The word kapó is obsolete, fitty is simply not used
outside the idiom, while duga, batka, korda and pellengér were borrowed from other
languages, became part of dialectal Hungarian (e.g. duga) or the standard language but
have become obsolete (cf. O. Nagy 1999: 200, 133-34, 97-98, 119, 228-29 and ÉKSZ: 647,
412, 246, 94, 761, 1096). The following contain morphemes (marked in bold) that have
homonymic counterparts outside the phraseological unit: írmagja se marad ‘be completely
gone’, hoppon marad 'be left on the shelf', se szeri se száma 'innumerable'. Finally, the

50
combination of the morphemes is unique in egérutat nyer 'manage to escape', hadilábon áll
'be very poor at sth', szemügyre vesz 'examine' and se szeri se száma 'innumerable'.
The criterion of lexical idiosyncrasy is used by Moon (1998a: 21) and Makkai
(1972: 123) to establish the subtypes called "cranberry collocations" and "pseudo idioms"
respectively. Both researchers use it as a secondary criterion in order to subdivide certain
categories. For Makkai (1972: 185, 340, 164) lexically unique idioms comprise a category
within lexemic idioms (one of the two major groups), but they are also regarded as
constituting a subcategory within various categories of lexemic idioms, since they are
doubly classified. Thus, kith and kin is found among both irreversible binomials and pseudo-
idioms (Makkai 1972: 320, 340).

2.9 Fraser

Fraser's (1970) major concern is transformational restrictions on idioms. It is


important to note that he examines the transformational potential of only a subclass of
idiomatic expressions: verb phrases (lay down the law, hit the nail on the head, lend a hand
to, shoot the bull, stew in one's own juice) (Fraser 1970: 40-41). Only one of the examples
is sentence-like: the cat has someone's tongue. Furthermore, he is not interested in affix
attachment or number attachment, which are considered to be transformations in early
generative grammar (Fraser 1970: 37, n.12).
He claims that parts of an idiom make no contribution to semantic interpretation,
arguing that gapping and clefting (*It was the riot act that John read to me) cannot occur, it
is impossible to conjoin parts of an idiom (*Mary took heed of John's warning and later
steps to rectify the situation) and it is ungrammatical to replace a noun phrase constituent of
an idiom with a pronoun or attach a restrictive relative clause to a noun phrase constituent
(Fraser 1970: 33).
Fraser (1970: 39) establishes six levels of frozenness, with a seventh level (L6
below) characteristic of free combinations:

51
L6 Unrestricted no idiom belongs here
L5 Reconstitution He laid down the law to his daughter  His laying down of the
law to his daughter
L4 Extraction look up the information  look the information up
L3 Permutation lay down the law  lay the law down
L2 Insertion read the riot act to the class  read the class the riot act
L1 Adjunction John hit the ball  John's hitting the ball
L0 Completely Frozen kick over the traces

What are regarded as different types of transformations in the generative tradition


may belong to the same level. For example, the passive transformation in The law was laid
down by her father and the particle movement resulting in look the information up both
move part of the idiom to a position outside the idiom (L4 - Extraction). Conversely, the
same type of transformations may belong to different levels. The particle movement relating
lay down the law and lay the law down and that relating look up the information and look
the information up are different, since in the former case it exchanges two successive
constituents of the idiom (L3 - Permutation), while in the latter case it extracts an idiom
constituent from the idiom (L4 - Extraction). As we move up the levels, we find more and
more variable idioms.
Fraser (1970: 39) claims that any idiom belonging to one level automatically belongs
to any lower level as well. This is a somewhat imprecise formulation, but it is made more
accurate by Fraser (1970: 39-40) himself: any idiom belonging to a particular level will also
permit transformations belonging to any lower level. Thus keep watch over belongs to L4;
consequently, extraction, insertion, and adjunction are possible, but reconstitution cannot
apply. Fraser (1970: 41) does not claim that all his idioms have the same degree of
frozenness for all native speakers, but he suggests that the hierarchy does hold.

2.10 Grammatical variation

Word combinations have certain grammatical restrictions, by which we mean that


they are subject to restrictions on the inflectional forms (call the shots/*shot), and
transformations20 (rearrangement, passivization, etc.) frequently result in ill-formed units
20
We will use this term in a theory-neutral sense for various types of structural changes. We do not wish to
imply that passivization and similar phenomena should be handled in transformational rather than non-

52
(kith and kin, *kin and kith) or well-formed non-idiomatic, non-figurative units (he kicked
the bucket, the bucket was kicked). The possibility of transformations and idiom-internal
inflections is supposed to show the variability of conventional expressions, while the
impossibility of the same features is taken to be evidence for the unit-like, word-like nature
(i.e. the gestalt nature) of idioms. Though this is true, the possibility of a certain
transformation can also show gestalt structure, depending on how that transformation
applies. Moon (1998a: 177) has found that when 'adjective + noun' idioms are modified by
an adjective, the modifying adjective is attached to the beginning of the idiom, irrespective
of the usual order of adjectives (political hot potato, legal red herring), which shows that
the idiom is a gestalt.
Fraser (1970) claims that certain transformations are impossible. Moon (1998a: 104-
16) does not comment on gapping, clefting, and co-ordination, but she observes that
embedding and pronominalization are rare in authentic texts, although isolated examples can
be found, e.g. Dean Ryan's early return stirs up waters they themselves have muddied,
Anyway, if there is ice, Mr Clinton is breaking it with a visit to the Canadian capital on
February 23rd and 24th.21 Moon (1998a: 108) finds corpus evidence in support of Nunberg
et al.'s (1994: 524) claim that idioms with double passives are rare. Nunberg et al. (1994:
501-502) and Pulman (1993: 252) give counterexamples to Fraser's claim about the
meaninglessness of idiom constituents: Those strings, he wouldn't pull for you
(topicalization), Pat tried to break the ice, but it was Chris who succeeded in breaking it
(pronominalization), It was thin ice that they were skating on (clefting) and Chickens that
are counted before they are hatched aren't a good basis for a bank loan (relative clause).
Nunberg et al. (1994: 499-501, 503) use their examples - together with modification by an
adjective (leave no legal stone unturned), quantification (touch a couple of nerves) and
ellipsis (My goose is cooked, but yours isn't) - to claim that constituents of idioms
contribute to the interpretation of the whole idiom, but they argue that it is not the literal
meaning of the constituent that makes the contribution, and this does not contradict Fraser

transformational frameworks. Omission and insertion, which result in lexical changes as well as structural
changes, are considered in this dissertation transformations, i.e. types of grammatical variation. Inflecting
parts of phraseological units will not be considered a transformation, but it will also be treated as
grammatical variation.
21
The example of embedding contains additional modification of the noun phrase constituent (waters
instead of the waters), but only one of Moon's (1998a: 110-11) six examples has no additional modification
of the noun.

53
(1970: 33).22 However, in Fraser's theoretical framework idioms are lexical units, so that the
meaning is carried by the idiom as a whole. Idioms are unanalyzable.
Moon (1998a: 120) has found that around 40 per cent of her phraseological units
permit lexical or transformational variation. No data is given separately for lexical,
inflectional and transformational variation, and Moon (1998a: 105) emphasizes that her
corpus is too small to yield conclusions concerning transformational variation. Negative
evidence does not necessarily mean that the given transformation is impossible. Certain
transformations may be strongly linked to certain types of phraseological units. Proverbs,
for example, are often shortened.
Among the transformational variations exemplified by Moon (1998a: 106-16, 139-
45) we find negation, passivization (sth is borne in mind), nonfinite forms (to add insult to
injury), embedding (another straw at which we can clutch), pronominalization (Anyway, if
there is ice, Mr Clinton is breaking it with a visit...), nominalization (coming and going),
transformation to adjectives (round-the-clock), transformation to predicates (make hay),
variations expressing the notion of possession (have (no, an) axe to grind - with(out) an
axe to grind), causative and resultative structures (let the cat out of the bag - the cat is out
of the bag), aspectual variation (cross one's fingers - keep one's fingers crossed), reciprocal
structures ((X is) at loggerheads with Y - (X and Y are) at loggerheads), ditransitive
structures (drop sb a line - drop a line to sb) and delexical structures (rap sb on the
knuckles - give sb a rap on the knuckles). She also discusses insertion (in extremely hot
water) (Moon 1998a: 174-77).
Before continuing the discussion, we must note an important point. The possibility
of a transformational variant need not automatically lead to the frequent occurrence of that
variant in a corpus. Even if the corpus is well-balanced, it may contain no examples or only
a few occurrences of the given variant. Therefore, a distinction must be made between the
possibility of a transformational variant and its frequency of occurrence.
It is difficult to judge how correct Fraser's hierarchy is, but on the whole it tends to
be rejected. Gibbs and Gonzales (1985: 246) have tested native speakers' intuition and
concluded that there was no regular correspondence between subjects' ratings and Fraser's,
while Moon (1998a: 105) reports Reagan, who has found an 86 per cent agreement
between informants and the hierarchy. Barkema (1996: 74-75) has examined noun phrase
22
However, as we will see, several idioms contain literal constituents in addition to nonliteral ones (know
the ropes).

54
idioms and found that there is a kind of hierarchy, but he also reports counterexamples.
However, it must be borne in mind that Barkema (1996) studied inflectional variation and
other grammatical changes, but only noun phrases, while Fraser, as we have seen, was not
interested in inflectional changes and considered only verb phrases. Howarth (1996: 19)
claims that " 'transformational deficiency' applies to individual idioms (and not only to
idioms) haphazardly, and one cannot draw general conclusions from it". Similarly, Cowie
(1994: 3170) holds that "specific restrictions do not apply evenly to idioms of a given
structural type [...] and may affect some collocations as well [...] while spill the beans (true
idiom) can be passivized, mark time (figurative idiom) cannot, and neither can foot the bill
(restricted collocation)".
Is grammatical variation a graded property? Fraser's hierarchy suggests gradation,
and Fraser (1970: 39) himself notes that it "reflects, from bottom to top, an increasing
degree of distortion permitted to the basic idiom shape of the untransformed idiom". Within
the framework of early generative grammar the degree of distortion can be measured via the
number of transformations as well as the degree of structural change. Extraction (L4) and
Insertion (L2) seem to be similar, since in both cases the transformed variant differs from
the citation form of the idiom in that a word/phrase that is not part of the idiom ends up
within the idiom. It is therefore not clear why they manifest different levels. Nevertheless,
grammatical variation is graded, but the number of subtypes varies. Fraser sets up six
subtypes within idioms, while Carter (1987: 63-64) arranges word combinations into three
subgroups along the cline of grammatical variation (and grammaticality): flexible (break sb's
heart), regular with certain constraints (drop a brick), and irregular (go it alone). Barkema
(1994: 41-42) also divides combinations of words into three types: fully flexible (a bird in
the garden), semi-flexible (a foot in the door) and inflexible (out of date). We assume four
levels: 1) highly variable (for example, free combinations), 2) variable 3) restricted 4) frozen
(for example, verb phrase idioms in which the verb inflects but no other variation is
permitted belong here).
Grammatical variation of the type discussed above must be distinguished from
exploitation, which is most strongly associated with journalism and metaphors (Moon
1998a: 170).23 Exploitation is the creative manipulation of conventional expressions, and it
leads to various non-institutionalized but contextually appropriate forms. In one type of

23
The term "exploitation" is taken from Moon (1998a: 170).

55
exploitation one or more words that are not part of the expression are added or inserted (on
the horns of a property dilemma) (Moon 1998a: 173, 174-77). Further examples of
exploitation are given in 2.22 below.

2.11 Jackendoff

Weinreich (1969), Chafe (1968) and Fraser (1970) can be described as early
generativists, whereas Jackendoff (1997) analyzes conventional expressions using the
apparatus of current generative theory. The remarks that we have made about generative
grammar apply only to early versions of the theory, so that it is desirable to summarize
Jackendoff's views. Furthermore, given the range of various generative theories (some using
others discarding transformations), we must bear in mind that generalizations imply a
certain amount of simplification.
Jackendoff's position is much closer to the cognitive approach than the position of
early generativists. The division between a fully regular syntax and irregular lexicon is less
sharp than in early generative approaches, as a certain amount of redundancy is felt to be
necessary (Jackendoff 1997: 15, 151). Syntax has a less central role in grammar, and its
separation is less clear-cut, because it shares a single interface with both phonological and
conceptual structures (Jackendoff 1997: 19, 99-100). Jackendoff (1997: 61-62) also prefers
an encyclopedic view of semantics, since world knowledge can be part of the lexical-
conceptual structure of words. Thus, lexical-conceptual structure comes closer to what is
referred to as "domain" or "idealized cognitive model" in cognitive grammar. The theory
rests on three pillars: phonological structures, syntactic structures and conceptual structures
as well as correspondences that link these structures (Jackendoff 1997: 38-39). Though
sharing some assumptions with cognitive grammar, this theory regards syntax as a mediator
between phonology and semantics, and the conceptual structure contains no imagery, only
propositional information. We find that the use of both propositional information and
imagery in domains (or ICMs) more closely reflects how language works.
One of the problems that generative theories must tackle is the insertion of idioms at
some point in the derivation of sentences. In early versions of the theory it was assumed
that idioms are lexical items that are inserted in deep structure in the same way as words.

56
Jackendoff (1997: 158-60) draws our attention to the disadvantages of this view. If let the
cat out of the bag is simply a verb, it has no internal structure, so that idiom-internal
inflections cannot be used. Furthermore, there are sentence-like idioms such as the cat's got
sb's tongue, which can be inserted under an S, not a V, or N, etc. Two other options can be
considered: a) the idiom constituents are listed separately in the lexicon, or b) the idiom has
phrase-like structure. Jackendoff (1997: 160) argues rightly that option a) requires ensuring
that, for example, bucket should have a special interpretation in the context of kick and vice
versa. The mechanism for this contextual specification performs the same task as option b),
but it does so in a more complicated way; therefore, option b) is preferred.
Option b) could mean that idioms are inserted in the deep structure at nonterminal
(phrasal) nodes, but Jackendoff (1997: 158) explains that this possibility has never been
exploited, so that the only solution is to analyze kick the bucket as a verb with a VP
structure: [V[VP[V kick] [NP[Det the] [[N bucket]]]]. We agree with Jackendoff (1997: 159) that
this analysis encounters difficulties. It implies phrasal syntax below the word level, and it
suggests that ordinary syntactic structure in the majority of idioms is accidental.
Jackendoff (1997: 159) claims that the analysis of 'verb + particle' constructions
(look up) as [V[V look] [Prt up]] is wrong on purely syntactic grounds, since the underlying
form must contain a noun phrase, because this is what generality demands (cf. put in the
books, put the books in, put the books in the box). We believe that it is misleading to ignore
semantic considerations and rely on syntax alone. Examples such as look up differ from put
in, exactly because up is not used in its spatial sense, so that up cannot be followed by noun
phrases in the way in can: *look the answer up a dictionary/the shelf vs put the books in
the box. Cognitive grammar analyzes look NP up and look up NP as related units that are
assembled through different compositional paths (Langacker 1987: 476). It is not necessary
to posit a very general underlying structure that is capable of handling variants through
transformations. It is likely that speakers extract the schemas [[V P] NP] and [[V NP] P]
from examples such as look up NP, turn on NP, look NP up, turn NP on, etc. (Langacker
1987: 477). But neither of these schemas has to be basic or underlying. Furthermore, it is
not obligatory to unite these schemas in a more schematic form.
Jackendoff (1997: 161-63) introduces lexical licensing, which can handle many of
the difficulties that insertion cannot. Licensing means that a lexical entry is unified
simultaneously with independently generated phonological, syntactic and semantic

57
structures. This is possible because a lexical entry contains phonological, syntactic and
semantic (i.e. conceptual) information, where certain chunks of the information are
coindexed. The figure below taken from Jackendoff (1997: 162) illustrates the point:

1) phonological structure:
take bto cthe dcleaners
a

2) syntactic structure:
VPx

aV PP

P
b NP

Det
c dN plur
3) conceptual structure:
[Event GET ([ ] A[ALL OF [MONEY OF[ ] A]])]x

Analyzing sentences as IPs, this theory also accommodates sentence-like idioms. However,
as Jackendoff (1997: 236, n. 6) admits it cannot apply to extragrammatical idioms. He
claims that they are truly marginal, which may not be the case (cf. 2.8). He also claims that
idiosyncratic grammar is found outside idioms as well, and the licensing of these examples
could be extended to idioms. For example, galore is an adjective that follows rather than
precedes the noun. However, galore does so in all cases, whereas the grammatical
idiosyncrasy of idioms is not found in non-idiom examples.
Another area studied by generative grammarians is the transformational recalcitrance
of idioms. In early generative theories it is assumed that idioms are inserted in the derivation
at the level of deep structure, and in Government and Binding theory certain phrases receive
idiomatic interpretation at D-structure as well. The problem with this approach is that there
are several idioms that occur only at surface structure (S-structure), i.e. only in transformed
variants (be hoist by one's own petard, play hard to get, Is the Pope Catholic, etc.)
(Jackendoff 1997: 167, Nunberg et al. 1994: 516, Moon 1998a: 109). These idioms must

58
undergo only certain transformations, but it is not clear what mechanism can ensure this.
Jackendoff (1997: 167) suggests insertion or licensing (depending on one's theory) at S-
structure, which has the advantage of stipulating correct surface forms. Since all idioms
have surface forms, these forms can be checked against the stipulated forms.
Jackendoff (1997: 168, 170) accepts the idea that many idioms are analyzable and
that this can contribute to their transformational variability, noting that not all analyzable
idioms can undergo transformations that would be possible (raise hell 'cause a serious
disturbance', *Hell was raised by Herodotus). There is no difference between cognitive
grammar and lexical licensing at this point.
Jackendoff's (1997: 171-76) theory is also capable of accounting for what he calls
"constructional idioms". These are partially compositional or noncompositional examples
with highly variable lexis: 'V + NP + PP/AP' (water the tulips flat, cook the meat black), 'N1
+ P + N1' (day by day, inch by inch). They have syntactic and conceptual structures but no
phonological structure.

2.12 Schematicity

Our discussion of compositionality in 1.3 mentioned two significant features: the


meanings of the constituents and the arrangement of the constituents. In phraseological
studies less attention is paid to the latter than the former. To redress the balance to some
extent, we will here focus on the noncompositionality displayed by the structure of idioms,
which leads us to a discussion of examples that Jackendoff labels "constructional idioms".
The noncompositionality of the structure entails that the meaning of a particular
expression does not follow directly from the arrangement of the constituents. If we say He
watered the tulips flat, we mean something like 'he caused the tulips to become flat by
watering them', or 'he flattened the tulips by watering them'. As Jackendoff (1997: 172-72)
explains, the syntactic head of the VP is the verb, but "it is not the semantic head; rather, its
meaning is embedded in a manner constituent". The meaning of this construction is
causative. Similarly, Him be a doctor? conveys more than would be predicted on the basis
of the structure alone. It is clear that idiomaticity (noncompositionality) is attached to the
structure in these expressions, since the words keep their literal meanings and can be

59
replaced by a large number of other words (We cooked the meat black, He talked himself
hoarse, Your brother help me?, Her write a novel about Spanish Inquisition?) (Jackendoff
1997: 171, Fillmore et al. 1988: 511). We have not found any statistics on what percentage
of word combinations are schematic.
Although the above examples are noncompositional, their noncompositionality is not
as high as that of grasp the nettle. First of all, the discrepancy between the strictly
compositional interpretation and the actual meaning is greater in grasp the nettle than in We
cooked the meat black. Secondly, owing to the productivity of constructional idioms,
speakers encounter many examples with the same structure, similar discrepancy in meaning,
but with different words. Langacker (1987: 451-52) claims that when the nature of the
discrepancy is constant from one expression to the next, speakers may capture this
regularity in a schema, and they may be able to predict the meaning of a novel expression.
Therefore, despite the discrepancy, predictability may be high.
Constructional idioms are called "formal idioms" by Fillmore et al. (1988: 505) and
"free realizations" by Moon (1998a: 158-61). In Langacker's (1987: 314) terminology they
are schematic, we will therefore call them schematic idioms. Fillmore et al. (1988: 505)
contrast these schematic idioms with what they call "substantive idioms", in which the
words are fully specified (grasp the nettle).
Schematicity is a graded property. Apart from expressions which exemplify two
extremes with highly variable lexis but fixed structure at one end (Him be a doctor?) and
with highly restricted lexis but variable structure at the other (put the cart before the horse,
with the cart before the horse, cart-before-the-horse, cf. CCDI: 61), a number of in-
between cases can be found, where part of the idiom is schematic, the rest is specified. High
schematicity is characteristic of examples such as The more confidence you build up in
yourself, the greater are your chances of success, in which the schematic part is the
comparative construction (shown by underlining in the structures), the specified part is the
definite articles: 'the + COMPARATIVE + the + COMPARATIVE'. Moon's (1998a: 159-
60) examples also show high schematicity: a few clowns short of a circus, one sentence
short of a paragraph and two slices short of a toast rack are all based on the schema
'QUANTIFIER + NOUN + shy/short + of + NP'. They are used to indicate mild insanity.
The following Hungarian idioms are also schematic: jobbnál jobb 'very good' ('AA + -bb +
-nál/nél + AA + -bb') and enni ettem 'approx. I did eat, but...' ('VERB STEMB + infinitive

60
suffix + VERB STEMB + past tense inflection + personal suffix'). The superscript letters
show that the same adjectives and the same verbs are used in these constructions. Less
schematicity is characteristic of idioms such as life with a capital L, based on the structure
'W with a capital X', where W can stand for a large number of words, and X marks the
initial sound/letter of word W, and what is specified - apart from the string with a capital -
is that the initial sound/letter of the first word must be the same as the last item in the idiom.
The variability of this idiom can be seen in the examples taken from our corpus: each
occurrence of the idiom contains a different first constituent (cf. 9.4). The idiom W with a
small X is also schematic, having an opposite meaning (cf. 9.4). A corresponding Hungarian
example is a víz az úr 'water has control', based on 'a + N + az úr'. Finally, most of the
idiom is specified in bury one's head in the sand or catch sb red-handed, where the only
schematic bit is the possessive determiner in the first example, and the noun phrase in the
second (catch NP red-handed). We can see a cline of schematicity/specificity within which
we establish three levels: 1) highly schematic (Him be a doctor?), 2) partially schematic
('QUANTIFIER + NOUN + shy/short + of + NP'), 3) specified (catch NP red-handed).
The entries for catch sb red-handed in idiom dictionaries are not the same. CIDI
(322) uses bold type for catch and red-handed only, implying that sb and the words that can
replace it are not actually part of the idiom, while CCDI (320) and LID (55) seem to include
sb in the idiom. The question arises whether the variable constituents that have literal
meanings are part of the idiom or not. The answer is not simple, but it can have an effect on
lexicographical practice. Jackendoff (1997: 161-62) argues that object noun phrases in verb
phrase idioms are semantically necessary, in order to satisfy the argument structure of the
verb, but he excludes them from his lexical entries. Langacker (1987: 313-14) treats catch
NP red-handed and lose PRONOUN's cool as units, in which a constituent is characterized
only schematically. In other words catch...red-handed and lose...'s cool seem to be
discontinuous units, but they are units only by virtue of being embedded in catch NP red-
handed and lose PRONOUN's cool, respectively.

2.13 Makkai

61
Makkai's (1972) study of word combinations was written in the framework of
Stratification Grammar, and it is one of the most detailed treatments of the field. He uses
"idiom" as a general label for various types of phraseological units (including compounds)
and identifies two groups of idiom: idioms of encoding and those of decoding (Makkai:
1972: 25). Idioms of encoding, also called phraseological idioms, are peculiar and fixed
ways of saying something, but they are not characterized by disinformation. Disinformation,
or "the ability to mislead", and ambiguity are key properties of idioms of decoding, also
called semantic idioms, such as hot potato (Makkai 1972: 25, 57, 122). The use of at in
Kim was driving at sixty miles an hour illustrates idioms of encoding, and restricted
collocations (take a bath, take a hair cut, take a walk) are also classified as such (Makkai
1972: 56-57). In fact, the term "idiom of encoding" is used in two slightly different ways,
one of which has been illustrated above. The other use of the term is found in Makkai's
(1972: 25, 57) claim that all idioms of decoding are simultaneously idioms of encoding as
well, but not vice versa, i.e. the category of idioms of encoding includes idioms of decoding.
In other words, all the examples above are idioms of encoding, since the speaker has to
encode in a certain way (Makkai 1972: 57). The distinction made by Makkai between the
two uses of "idiom of encoding" draws our attention to two different but related aspects of
language: production and comprehension.
The potential ambiguity of idioms (of decoding) arises out of the possibility of the
constituent words to occur outside the idiom, usually in a literal sense (Makkai 1972: 122-
23, 148). (We will ignore for the moment ambiguity due to a particular idiom having two or
more idiomatic meanings, an ambiguity that is also potential rather than real (cf. Moon
1998a: 187-93 and Földes 1987: 22-23)). Unique words cannot occur in other
environments, as a result of which idioms containing them cannot mislead the listener and
they are called "pseudo-idioms" by Makkai (1972: 123). Moon (1998a: 178-79) and
Fernando (1996: 6) argue against attaching too much importance to homonymy and
ambiguity. They explain that the context disambiguates in most cases, we will therefore not
consider the criterion of ambiguity a decisive one in assigning an expression to the class of
idioms. Phrases describing literally impossible situations, such as walk on air, make a
mountain out of a molehill, get one's mind around sth, are not genuinely misleading, but
they are idioms. Makkai (1972: 311-13) includes in his tournure group be caught between

62
the devil and the deep blue sea, rain cats and dogs, dance on air, be born with a silver
spoon in one's mouth. Whether they are literally possible, not impossible but highly unlikely,
or impossible depends on one's experience and worldview.
Each morpheme of an idiom normally occurs outside the idiom in a different sense.
This is clear from Makkai's (1972: 122-23) treatment of the singular, the plural and the
articles as exceptional morphemes (i.e. they do occur outside the idiom in the same sense),
as well as from the following statements: "[a]ny polylexonic lexeme which is made up of
more than one minimal free form or word [...] each lexon of which can occur in other
environments as the realization of a monolexonic lexeme is a lexemic idiom" [my emphasis],
and "lexemic idioms 'disinform' or 'mislead' the uninformed listener because their constituent
lexons occur in other environments as the realization of (other) monolexonic lexemes" [my
emphasis] (Makkai 1972: 122, 148). 24 In spite of this, Makkai's (1972: 311-13) detailed
classification includes examples one of whose words (apart from the articles) can occur in
the same sense outside the idiom. In the following list Makkai's own paraphrases are given:
to be caught between the devil and the deep blue sea 'to be caught between two equally
unpleasant alternatives', to call somebody on the carpet 'to summon somebody to one's
office for reprimand', to rain cats and dogs 'to rain profusely', to call a spade a spade 'to
call a thing by its right name'. Apparently, an idiom can mislead the listener even if it
contains a word whose meaning is not always different in other environments.
Within idioms of decoding two groups are distinguished: lexemic and sememic
idioms (Makkai 1972: 122, 128). This division roughly corresponds to that of word-like and
sentence-like units, though some word-like units (never get to first base) belong to
Makkai's sememic group. The two groups of idioms comprise the two idiomaticity areas
established by Makkai (1972: 117). Lexemic idioms are subclassified on the basis of the
syntactic structure. The following categories are listed together with a number of examples:
a) phrasal verb idioms (take off), b) tournure idioms (kick the bucket, pull a fast one, rain
cats and dogs), c) irreversible binomial idioms (sink or swim), d) phrasal compound idioms
(hot dog, dog in the manger), e) incorporating verb idioms (eavesdrop) and f) pseudo-
idioms (spick and span) (Makkai 1972: 135-69, 191-340).
Some idioms are doubly classified. Makkai (1972: 153, 159) feels that the
combination of a verb or preposition with an irreversible binomial (rain cats and dogs,

24
The term "lexon" corresponds to the traditional concept of morpheme.

63
through thick and thin) belongs to both tournures and binomials. Despite this only some of
the examples (by fits and starts, through thick and thin) can be found in both groups in the
more detailed classification, given in the second part of the book, while others are listed
only in one of the groups (rain cats and dogs, without let or hindrance) (Makkai 1972:
312, 314-20). Pseudo-idioms are also claimed to be doubly listed, which is true except for
cranberry, hara-kiri and kit and caboodle (Makkai 1972: 164, 169, 321-40). Incorporating
verb idioms are also placed in two different groups, though this is not stated explicitly. They
are found in a group of their own, at the same time all incorporating verbs could be treated
as a subtype of tournure idioms: verbs without direct objects (Makkai 1972: 313, 339-40).
There are some inaccuracies in the assignment of certain schemas. For example,
Makkai (1972: 163) holds that the binomials ups and downs and ins and outs have the
schema 'preposition/adverb + Plural + and + preposition/adverb + Plural', but the conjoined
words are nouns, since they take the plural suffix. Both idioms are analyzed as conjoined
nouns in Seidl and McMordie (1988: 81, 82) and ODEI (318, 573), and the syntactic
structures in which the idioms are used also tip the balance in favour of regarding them as
noun phrases: ins and outs is usually preceded by the definite article and followed by of, ups
and downs is often used as the object of have. The grammatical idiosyncrasy of these idioms
arises from assigning to two of their constituents a strange word class, not from attaching
the plural to a preposition/adverb. However, it must be admitted that Makkai's (1972: 163)
approach is justified within his own framework, since the same words that are conjoined in
these idioms cannot be used as nouns elsewhere (except for up), but the constituents of
lexemic idioms can occur elsewhere. Within tournures a subcategory is set up whose
structure is described as follows: "the tournure headed by a verb does not contain either the
compulsory IT or either one of the compulsory articles A or THE, and is followed by a
direct object after the verb, followed by further optional modifiers" (Makkai 1972: 154).
However, some of the examples used to illustrate this do contain articles (build castles in
the air, have a beef against sb, make a federal case out of sth, take a raincheck on sth, call
a spade a spade) (Makkai 1972: 312-13). It is not clear either why fish and chips is put into
the non-idiomatic binomial group in the detailed classification, when it refers to an
institution and it is said to be of the same type as the idiom gin and tonic (Makkai 1972:
161, 317).

64
Sememic idioms are subclassified according to their pragmatic function and their
type of source as well ("First Base" idioms and familiar quotations). Among sememic idioms
we find examples of politeness formulae (May I...?, more precisely May I VP?), proposals
(How about a drink?, a partially schematic idiom: How about NP?), greetings (So long),
proverbs (Birds of a feather flock together), etc.
Similes that are considered idiomatic by most researchers are classified as non-
idiomatic by Makkai (1972: 338). Conversely, some binomials that we would treat as non-
idiomatic are analyzed as idiomatic (Makkai 1972: 314-16): Adam and Eve 'biblical couple',
brace and bit 'carpenter's tool', flotsam and jetsam 'remnants of a shipwreck', nuts and
bolts 'common fastening devices for wood and metal, etc.'. The phrase hammer and tongs is
considered by Makkai (1972: 124) as non-idiomatic in the sense 'traditional blacksmith's
tools', but brace and bit in the sense given above is treated as an idiom despite the fact that
in terms of compositionality the examples are very similar.

2.14 Figurativity

Phraseological units often involve metaphors (take the bull by the horns),
metonymies (lend a hand) and other figurative devices, for example simile (like a bear with
a sore head), hyperbole (a storm in a teacup), litotes (not hold water), irony (a fine/pretty
kettle of fish), etc. (Nunberg et al. 1994: 492, Moon 1998a: 193-200). Figurativity is typical
of idioms, and it is also found in other types of phraseological units, though to a lesser
extent. The figurativity of conventional expressions is also noted by Hungarian researchers
(cf. Hadrovics 1995: 30, O. Nagy 1999: 11). As we have seen, for Weinreich (1969: 43-44)
there is nothing phraseological about expressions that cannot have a literal interpretation,
while for others many expressions are literal. For Makkai (1972) idioms (of decoding)
without literal interpretation are not genuine idioms. The literal and figurative
interpretations of a sequence of words are often associated with different collocations or
structures. Moon (1998a: 183) reports that the adjectives scalding, steaming and boiling
occur in the Bank of English25 in combination with in hot water only with the literal
interpretation of this phrase.
25
A corpus of authentic English, which contained over 300 million words at the time of Moon's (1998a)
report.

65
Defining what is literal and distinguishing between literal and figurative meanings is
notoriously difficult (cf. Gibbs 1994: 75-79, McArthur 1992: 403). Nevertheless, there is a
difference between them. In the dissertation "figurative" is used as a synonym of
"nonliteral", except where it could lead to misunderstanding.
We have not found any data about the ratio of figurative units. Idioms, a number of
collocations and many proverbs are figurative, while formulae (Good morning) tend to be
literal.
The criterion of figurativity divides phraseological units into four groups: a)
combinations having only literal interpretation (Good morning); b) expressions with both
literal and figurative interpretations (grasp the nettle); c) expressions whose literal
interpretation is illogical or impossible (kith and kin, shoot the breeze) and d) expressions
with partly literal partly figurative meanings (explode a myth). But, as Moon (1998a: 178)
points out, in context many units can be interpreted only nonliterally, so it is misleading to
divide idioms with and without literal counterparts so sharply. She gives a comparison of
the literal and idiomatic frequency of eight phrases, including break the ice, (skate) on thin
ice and spill the beans, and her statistics support the greater likelihood of idiomatic readings
(Moon 1998a: 182).
Literal interpretation is potential rather than actual in groups b) and c). When
interpreted literally, combinations belonging to group b) are free combinations (grasp the
nettle, rock the boat) or collocations (take a walk, mind one's step, strike gold), not idioms
(for further examples see Howarth (1996: 117)).
One of the categories of word combinations, called "restricted collocations" are
partially literal. Carter (1987: 63) and Fernando (1996: 254) use the term "restricted
collocation" in a different sense, but in this dissertation we follow Cowie et al. (1993: xiii).
Howarth (1996: 65, 92-101) shows that the nonliteral constituent (i.e. the verb) in a 'verb +
noun' restricted collocation can have a technical meaning (bring an action), a delexical
sense (take a decision), or a figurative sense (follow instructions). The technical sense is
usually a specialization of one of the figurative senses of the verb, restricted to a specific
register (Howarth 1996: 93). Some of Howarth's (1996: 65) examples of restricted
collocations satisfy the criteria mentioned above (crying shame, richly deserve, cast a vote),
others do not. He lists binomials (trial and error, bloodied but unbowed) and trinomials
(lock, stock and barrel, left, right and centre), which are clearly not restricted collocations

66
even by his own criteria. In trial and error all the words are used in their primary, literal
senses, in the other examples the words do not seem to have independent senses (except for
bloodied but unbowed, where but is used in its usual sense, but the other words are
figurative), and lexical substitution is possible only in bloodied but unbowed.
Though partial figurativity/literalness is often used to identify restricted collocations
(explode a myth, make a decision), it is not a reliable criterion; therefore, it is supplemented
by other criteria. Some fully figurative units are sometimes classified as restricted
collocations. For example, Howarth (1996: 97) discusses among restricted collocations
expressions such as take a stance and make strides, in which none of the constituents is
literal, though he notes that they could equally be classified as idioms. In other cases
partially figurative/literal combinations are treated as free combinations or idioms. Howarth
(1996: 101) explains that reflect in its figurative sense can collocate with a large number of
nouns, and such combinations as reflect the disintegration are free combinations. Moon
(1998a: 83, 89), who studies idioms, includes examples such as look daggers at sb and
dressed to kill, in which one of the constituents has a literal meaning.
Gläser (1986: 43) illustrates the three categories of restricted collocations, free
combinations and idioms in a table which is given below:

67
free combination restricted collocation idiom
to run a race to run a farm to run the gauntlet 'to risk danger, anger,
to run a mile to run a hotel criticism, etc.'
to run a horse to run a club
to run a car to run the country to run it fine/close '(not formal) to leave the
(at a rally) 'control it' least amount of time, money, etc., possible
to run a train to run a candidate before the end of a limit for finishing or doing
(on oil) to run a risk something'
to run the water to run a temperature
(till it gets hot) to run one's head against a brick wall '(coll.)
to run a hot bath to try to do or obtain something difficult with
very little hope or success'

to run with the hare and hunt with the hounds


'(rather lit.) to try to remain friendly with or
offer support to two very different types of
people or groups'

The table suggests that the boundaries of these categories are clear-cut and that the defining
criterion is the literal/figurative distinction. Contrast also drop a pen (free combination),
drop a hint (restricted collocation) and drop a bombshell (idiom). However, the boundaries
are fuzzy, and emphasis on the criterion of lexical variation can lead to a combination being
regarded as a free combination or an idiom rather than a restricted collocation. Thus, the
first three examples of restricted collocations in the table (run a farm, run a hotel, run a
club, where run means 'manage') seem to be free combinations. Howarth (1996: 31)
suggests that restrictions on the range of nouns that can co-occur with run in the given
sense can be stated in general semantic terms, and the set of nouns is not limited, which
supports analysis of run a farm/hotel/club as free combinations. Fernando (1996: 32, 36)
also considers run a business/company, etc. as unrestricted, although she shows that other
combinations of the same type can have arbitrary limitations. For example, she claims that in
catch a bus/plane/ferry the nouns refer to public transport, but ship cannot be used,
although boat can, while private transport nouns are unacceptable, though catch a taxi is

68
possible (Fernando 1996: 36). This suggests that catch a bus is slightly different from run a
business. Fernando's claims are borne out by our corpus (cf. 9.5), which contains one
example of ship, and this is below our significance threshold (cf. 4.1.2). Both taxi and boat
are found more than twice in combination with catch. Restrictions on combinability can be
discovered with the help of corpus investigations, so that it is especially in this area that
phraseology could benefit from corpus linguistics.
Deciding whether a combination is an idiom or not is also problematic for
lexicographers. Run a risk and run a temperature are classified as restricted collocations in
the table above, but in OALDb (1105, 1338) only the former is labelled as an idiom, the
latter is not. CCED (1458) does not regard run a temperature as an idiom and shows that
fever can be substituted for temperature.26 However, while run a temperature is only
partially compositional, since the fact that the temperature is high cannot be predicted from
the combination, run a risk is more compositional. Some inconsistencies can also be
observed. In OALDb (1043) the expression fly into a rage is given in bold within an
example sentence in the entry for rage, which suggests that it is a collocation. However, the
same expression is labelled as an idiom in the entry for fly (OALDb: 494)! Many partially
literal combinations are listed in dictionaries of idioms. The table below shows some
examples:

figurative literal figurative


promise the moon/earth
swallow one's pride
teach/show sb/learn/know the ropes
know the score
lie through one's teeth
teething problems/troubles
white lie
rain cats and dogs
hate sb's guts
pay through the nose
cold comfort
dressed to kill
laugh/shout etc. one's head off
look daggers

26
Both OALDb and CCED are corpus-based dictionaries.

69
talk shop
square meal

These examples are included in most dictionaries of idioms, and they are classified as
idioms in OALDb (1014, 1112, 741, 1044, 591, 930, 384, 595, 314, 1327, 1255), except
for swallow one's pride, know the score, teething problems/troubles, white lie and cold
comfort (OALDb: 1003, 1143, 1335, 1477, 231). 27 In all these examples one part of the
expression is used in its literal sense, while the other part is figurative, though in lie through
one's teeth, rain cats and dogs, and pay through the nose a substantial part is figurative.
In terms of lexical variation, the examples above allow no or only very limited
substitution, except for swallow one's pride. Since swallow can be used with a number of
other words in the sense 'to hide your feelings' (OALDb: 1312), this expression is a
restricted collocation rather than an idiom. It may well be that there are strong lexical
preferences, so that pride, and anger are much more likely to co-occur with swallow than
other words denoting feelings, but other nouns are not unacceptable. CCED (1686) shows
that anger is also possible, OALDb (1312) gives doubts and anger, LDCE (1454) lists
doubts, CIDE (1473) shows disappointment and anger.
It must be noted, however, that in restricted collocations of the 'noun + verb' or
'adjective + noun' types, it is typically the noun that is used in a literal sense, and the verb or
adjective is nonliteral, and this could be used as an additional criterion to distinguish
examples such as promise the moon from restricted collocations. In other cases, as in cold
comfort or square meal our additional criterion will not help, because it is the adjective that
is figurative. Since a ptototypical idiom is unmotivated, the opacity of the figurative element
tips the scales in favour classification as an idiom.
Despite what is implied in Gläser's table above, for her idioms can be partially literal
(white lie) or fully figurative (white elephant, burn the candle at both ends) (Gläser 1986:
55-56). The former are called "unilateral idioms", the latter are "bilateral idioms" when they
consist of two open-class words (white elephant, smell a rat) and "multilateral idioms"
when they contain more than two open-class words (burn the candle at both ends). Some
of the examples are in the wrong groups (green light 'the sign, or permission, to begin an
action', the red carpet 'not formal, especially good treatment' are identified as unilateral,

27
Teething problems/troubles, white lie, and cold comfort are given separate entries, presumably because
they are treated as compounds.

70
whereas they are fully figurative (cf. Gläser 1986: 72)). In deciding whether a unit is an
idiom or a restricted collocation, Gläser (1986: 38-41, 87) emphasizes figurativity rather
than nonliteralness and has recourse to lexicogrammatical criteria. Thus, 'delexical verb +
noun' combinations (pay a visit) are not idioms (Gläser 1986: 68).
We consider figurativity to be a graded property, so that phraseological units can be
1) literal (Better late than never), 2) partially literal (explode a myth), 3) figurative (i.e.
nonliteral) with a potential literal interpretation (grasp the nettle) and 4) figurative with no
literal interpretation (shoot the breeze).

2.15 Structural classification

One of the lexicogrammatical aspects of phraseological units is their syntactic


structure. Structural classification of word combinations is often mingled with functional
classification. The primary division is between word-like units (lose one's head, zöld ágra
vergõdik 'get on') and sentence-like units (the buck stops here, kicsi a világ 'it's a small
world') (Cowie 1998b: 4, O. Nagy 1985: 13, Hadrovics 1995: 34, 114). The former
function as sentence constituents, i.e. noun phrases, verb phrases, etc., the latter as
sentences. Cowie et al.'s (1993: xi) division into phrase idioms and clause idioms does not
correspond to the above-mentioned grouping, since a unit containing a verb and its
complement, object or adverbial (lose one's head, paint the town red, zöld ágra vergõdik
'get on') is considered a clause idiom (cf. Quirk et al.'s (1985: 69) verb phrase, which does
not contain objects, complements or adverbials and Fernando's (1996: 39, 42) use of "semi-
clause" for examples such as lose one's head).
Hungarian being a pro-drop language, the division between word-like and sentence-
like is less sharp than in English. For example, Hadrovics (1995: 30, 114) classifies vki
ingatag talajon áll 'stand on unsteady ground' and similes (ravasz, mint a róka 'cunning as a
fox', reszket, mint a nyárfalevél 'shake like a leaf') as sentence-like idioms. Since overt
subjects are not obligatory in Hungarian, similes are analyzed as sentence-like units,
consisting of a main clause without subject and an elliptical subordinate clause of
comparison. O. Nagy (1985: 13) treats similes as word-like, though he is also aware of their
sentence-like structure.

71
Word-like units can sometimes be traced back to sentence-like combinations. The
source of the idiom birds of a feather is the proverb birds of a feather flock together.
Similarly, ajándék ló 'gift horse' comes from ajándék lónak ne nézd a fogát 'don't look a gift
horse in the mouth'. These cases must be distinguished from allusions, when a fragment of a
proverb stands for the whole proverb (You know what they say, no smoke..., eh?).
Within word-like units the commonest type is the predicate, especially 'verb + object'
(bend the rules), 'verb + adjunct' (fall on deaf ears) and 'verb + object + adjunct' (lay one's
cards on the table) (Moon 1998a: 83, 85-87). The asymmetry between 'verb + object' and
'subject + verb' idioms (the penny drops) has been noted by several linguists (Langacker
1987: 236, Nunberg et al. 1994: 525). Adjuncts, most of which are prepositional phrases
(by heart), are common, noun phrases (a blessing in disguise) are less common, and
adjective phrases (wet behind the ears) are infrequent (Moon 1998a: 84, 88, 89). Some
phraseologists use a specific label for word-like units, such as "composite", "nomination" or
"állandósult szókapcsolat" (Cowie 1998b: 5, Juhász 1980: 85).
Word-like units are subclassified according to the syntactic function of the whole
unit (a flash in the pan - noun phrase, complement, teljes gõzzel 'at full throttle' - noun
phrase, adverbial) or - with various degrees of precision - according to the structure (word
class or syntactic function and arrangement) of the constituents (a flash in the pan - 'noun +
preposition + noun'; teljes gõzzel 'at full throttle' - 'adjective + noun'; lay one's cards on the
table - 'verb + object + adverbial') (cf. Moon 1998a: 85-94, Arnold 1986: 172-73, Cowie et
al. 1993: xxix-xxxvii, Földes 1987: 34-36). Some (sub)categories of word combinations
distinguished structurally have specific labels. A noncompositional 'verb + particle'
combination is called a phrasal verb.28 Although phrasal verbs are idioms, they are not
always discussed together with idioms and not all dictionaries of idioms cover phrasal verbs
(cf. LDEI: x, CCDI: v). Being numerous in English and easily separable from the rest of
phraseological units by a structural criterion alone, phrasal verbs are often listed in separate
dictionaries and excluded from studies of idioms. Idioms with the schema '(as) + adjective +
as + noun phrase' (as good as gold) or 'verb + like + noun phrase' (work like magic) are
called similes, corresponding Hungarian units (ravasz, mint a róka 'cunning as a fox',
reszket, mint a nyárfalevél 'shake like a leaf') are referred to as "szóláshasonlat". Binomials
and trinomials are also distinguished on the basis of their syntactic structure alone.

28
Many linguists also consider compositional 'verb + particle' combinations to be phrasal verbs.

72
Binomials are parallel units, usually two conjoined words linked by a hyphen or by the
words and, or, nor, or but, and the term is also used for two juxtaposed prepositional
phrases (in English) (part and parcel, wine and dine, slowly but surely, give or take, back
to front, from cradle to grave). Hungarian examples of the corresponding category are
éjjel-nappal 'day and night', boldog-boldogtalan 'every Tom, Dick and Harry', se füle, se
farka 'nonsense', se szó, se beszéd 'out of the blue', lépésrõl lépésre 'step by step', etc.
Trinomials contain three words, the last of which is linked to the previous word usually
through the word and (lock, stock and barrel). Within collocations a distinction is made
between lexical and grammatical collocations, and this distinction is also based on the
structure alone (Benson 1985: 61). Grammatical collocations are combinations of an open-
class word with one or more closed-class words (account for), lexical collocations consist
of at least two open-class words (adopt a policy) (Benson 1985: 61-62). Cowie (1998c:
225) warns against using the term "grammatical collocation" for combinations of a word
and its finite or non-finite clause complementation (an agreement that she would represent
us), but this is sometimes found. Grammatical collocations are often excluded from
phraseology despite the frequently arbitrary nature of the restrictions. The division between
these subcategories is sharp, but lexical collocations can involve constraints on certain
grammatical aspects as well, such as the use of determiners (Howarth 1996: 35). Thus, a
given combination can combine lexical and grammatical collocations (in fly into a fury, the
words fly and into form a grammatical collocation, and this combination forms a lexical
collocation with fury).
Within word-like units the subcategories have sharp boundaries, since the criterion
of structure is not graded. Similarly to analyzability, grammatical structure cannot
distinguish free combinations from phraseological units.
We have found some inaccuracies in some of the structural classifications. Makkai's
(1972) typology has already been discussed. Földes (1987) offers a structural classification
of Hungarian, German and Russian word-like units. The Hungarian expression fejvesztve
rohan is classified as an adverbial, but in fact the whole word combination functions as a
verb phrase (Földes 1987: 34). It is a collocation of a verb with an idiomatic compound, the
compound functioning as an adverb. It is not clear why véget vet vminek and tudomásul
vesz vmit are placed in different groups ("frazeologikus kapcsolatok" 'phraseologial
combinations' and "funkcionális igei kapcsolatok" 'functional verbal combinations'), since

73
both seem to be combinations of a figurative verb with a literal noun. Similarly, lépésrõl
lépésre and keresztül-kasul are of different types for Földes (1987: 19-20), although both
seem to be binomials. Földes (1987: 35-36) shows only the word-class label of the
constituent words in the schemas; consequently, the 'V + (d) + N' schema is listed twice for
Hungarian, illustrated by örül a szíve 'be happy' and tartja a száját 'keep one's mouth shut'.
Obviously, the nouns have different syntactic functions in these examples (subject and
object). It is only after a detailed scrutiny that we realize that the abbreviation "suff" does
not simply mean 'suffix', but it stands for case inflections. However, not all case inflections
are marked. For example, the word képet, a noun in the accusative, is simply an 'N' in
savanyú képet vág 'pull a long face', whereas karokkal is an 'N + suff' in tárt karokkal fogad
'welcome with open arms', otherwise these idioms conform to the same structure: 'adjective
+ noun + verb' (Földes 1987: 36). We can only speculate that the reason for this is probably
Földes's effort to match Hungarian structures with corresponding German and Russian ones
as closely as possible, and Hungarian inflected nouns as well as 'noun + postposition'
combinations often correspond to 'preposition + noun' combinations in these two foreign
languages. Exceptions include the Hungarian accusative and genitive.
Sentence-like units can have various pragmatic functions and a number of them are
to a large extent (semantically) compositional (Good morning, How do you do, You can say
that again, Yours sincerely, You can't have your cake and eat it, Meghiszem azt! 'I should
think so', Jobb késõn, mint soha 'Better late than never'). The term "idiom" in British
phraseology is not normally used for sentence-like items with various pragmatic functions
(Cowie 1988: 132-33, Carter 1987: 58-59, Benson et al. 1986b: 253), although it is clear
that noncompositional expressions can be found among proverbs, catchphrases and
formulae. Cowie et al. (1993: xv) illustrate metaphorical proverbs and catchphrases (The
early bird catches the worm, The buck stops here, If you can't stand the heat, get out of the
kitchen), and noncompositional formulae can also be encountered (You can say that again).
Zgusta (1971: 151) also notes that at least some idioms belong to "set groups of words",
which is his cover term for sentence-like items. In contrast, Benson et al. (1986b: 253)
consider idioms and proverbs different categories.
Phraseological units with pragmatic function are distinguished by many researchers
(Makkai 1972: 172-79, Fillmore et al. 1988: 506, Moon 1998a: 19, Carter 1987: 60).
Allerton (1984: 36-39) discusses co-occurrence restrictions on words and claims that there

74
is a pragmatic level (e.g. Fares, please!). The label "formula" is typically used to denote
expressions with various discourse-structuring or speech-act functions (Good morning,
How do you do, You can say that again, Yours sincerely, Fares, please!) (Cowie 1994:
3169, 3170). One of Moon's (1998a: 19) major categories is formulae, which are
pragmatically noncompositional. She subdivides this category using a mixture of criteria, as
is usual in subclassifications of sentence-like units. There is some disagreement as to how
many categories should be set up and what type of defining properties they have. The
various types of sentence-like units are distinguished on the basis of criteria including the
type of pragmatic function: greetings (Good morning), units expressing agreement (You can
say that again), proverbs (You can't have your cake and eat it), etc.; and various aspects of
the source: catchphrases (The buck stops here, May the Force be with you), slogans (Safety
first), proverbs (You can't have your cake and eat it), quotations (A thing of beauty is a joy
for ever) (Moon 1998a: 22, Crystal 1995: 178, 180, 184, Cowie et al. 1993: xvii, xl, Nagy
1996: 26-29).
Quotations, catchphrases and slogans all have identifiable sources, but they differ in
other properties of the source and in their functions. Quotations tend to have a written
origin, and they are semantically more profound than catchphrases, which tend to originate
with a prominent figure (If you can't stand the heat, get out of the kitchen, first used by
President Truman) or in the world of entertainment (Crystal 1995: 178, 184, Cowie et al.
1993: xvii, xli). Catchphrases tend to be of spoken origin, relatively trivial in subject matter
and popular for only a short period. Although Crystal (1995: 184) puts the emphasis on
their differences, he admits that "catch phrases are, indeed, a species of quotation".
McArthur (1992: 944) sees slogans as a subtype of catchphrases, while Crystal (1995: 178-
80) discusses them as separate categories. Slogans are essential parts of various campaigns
in the fields of advertising, politics, health, environment, etc., and they encourage people to
behave in a certain way or buy something (Crystal 1995: 180).
Cowie et al. (1993: xv, xl) and Benson et al. (1986b: 253) do not distinguish sayings
from proverbs, while Arnold (1986: 179) gives the following heading to one of the sections
in her book "Proverbs, sayings, familiar quotations and clichés", suggesting that sayings and
proverbs are separate categories, but the section contains no discussion of the former.
Similarly, Zgusta (1971: 151) and Nunberg et al. (1994: 492) mention both proverbs and
sayings but do not explain the difference. Moon (1998a: 22) subsumes under "saying"

75
catchphrases and quotations and claims that proverbs have "deontic functions". This
function is also mentioned by McArthur (1992: 818), who defines "proverb" as "a short
traditional saying of a didactic or advisory nature, in which a generalization is given specific,
often metaphorical, expression". At the same time, the definition of the term "saying" is not
very different: "an informal, general term for anything said, especially if it is brief and to the
point; a pithy or concise observation that expresses folk wisdom that has been handed down
orally, or represents a basic principle, fundamental teaching, or the like" (McArthur 1992:
888-89). Nagy (1996: 27) also notes the similarity but claims that "sayings are less likely to
give advice and more likely to be philosophical". He adds that most sayings originate in
works of literature, and their authors can generally be identified (Nagy 1996: 27). This
means that his term "saying" roughly corresponds to "quotation" or "aphorism".
Similar categories are distinguished in Hungarian: közmondás 'proverb' (Aki másnak
vermet ás, maga esik bele ‘the biter gets bit’, literally ‘He who digs a pit for others will fall
into it himself’), szállóige 'catchphrase' (Nyelvében él a nemzet ‘The nation lives in its
language’), társalgási fordulat or formula 'formula' (Hogy vagy? ‘How are you?’), etc. (O.
Nagy 1985: 9, 15, Bencédy et al. 1988: 495-98, Hadrovics 1995: 28-29, 128).

2.16 Arnold and other Russian phraseologists

Arnold (1986) discusses a wide range of phraseological units and recognizes the
graded nature of lexical variation. She divides word strings into free phrases, semi-fixed
combinations and set expressions, the latter including restricted collocations and idioms
(Arnold 1986: 167-68). Examples of various degrees of lexical restrictedness are given (cut
bread/cheese, cut/eat bread versus cut no ice), and pronominalization is also illustrated,
though the example is a restricted collocation rather than an idiom (black frost) (Arnold
1986: 168-69).
Arnold reports various Russian linguists' classifications, which influenced British
models (Cowie 1998b: 4). Vinogradov's scheme is based on semantic and lexicogrammatical
factors, especially the motivation of the unit (Arnold 1986: 170). He draws a three-way
distinction between phraseological fusions (tit for tat), phraseological unities (stick to one's
guns) and phraseological combinations (meet the demand) (Arnold 1986: 170). The first
category is characterized by a high degree of frozenness and lack of motivation, while

76
phraseological unities are motivated, since their sense can be seen as a metaphorical
extension of some original sense (Arnold 1986; 170, Cowie 1998b: 4-5). Lexical
substitution is possible, but limited in the latter (stick/stand to one's guns) (Arnold 1986:
170). The most controversial category is that of phraseological combinations. They are units
one of whose constituents (an open-class word) is used in a literal meaning, and another
constituent (an open-class word) is used figuratively (Arnold 1986: 170, Cowie 1998b: 5).
Vinogradov's classification is compatible with and corresponds to the distinctions made by
Cowie et al. (1993: xii-xiii). Vinogradov's phraseological fusions correspond to Cowie's
pure idioms, phraseological unities are figurative idioms in Cowie's scheme, and
phraseological combinations are the same as Cowie's restricted collocations (Cowie 1998c:
214-15).
Amosova challenges Vinogradov's divisions, subdividing phraseological
combinations into two further types (Cowie 1998c: 215). The word that is used in a
figurative sense is considered to have a phraseologically bound meaning, the other word is
the binding or determining context. If this context is completely fixed, i.e. no substitution of
the literal word is possible without changing the meaning of the figurative word, as in small
change, the combination is classified as a phraseme and regarded as part of phraseology. If
the context is not fixed, as in meet the demand/the necessity/the requirements, the
combination is called phraseoloid and viewed as falling outside phraseology (Arnold 1986:
171, Cowie 1998c: 215). Thus, Amosova provides another example of how focusing on a
given criterion can affect the scope of phraseology. Other analysts, especially those with an
applied linguistic approach, are more liberal than Amosova and include units in the
phraseological spectrum which permit a more or less limited substitution of both (or all)
their constituents (Cowie 1998c: 216, Howarth 1996: 43).
Amosova also recognizes the division between motivated and unmotivated idioms,
but uses the label "idiom" for both (Arnold 1986: 171, Cowie 1998c: 215). She is not the
only one to establish one category for both (partially) motivated and unmotivated word
combinations. Stroyeva, reported in Arnold (1986: 171), and Mel'čuk, reported in Cowie
(1998c: 215), are reluctant to set up two groups, arguing that motivation depends on the
speaker's cultural, educational and linguistic experience (Arnold 1986: 171, Cowie 1998c:
215).

77
After a discussion of the similarities and differences between idioms and single
words, Arnold (1986: 177-79) describes various properties that increase the frozennes and
unitary nature of phraseological units as well as reducing the burden they impose on
memory. These are rhythm (far and wide), rhyme (high and dry), alliteration (with might
and main) and imagery (metaphor, metonymy, simile and contrast). Finally, she addresses
proverbs, familiar quotations and clichés (Arnold 1986: 179-81). Contrary to Howarth's
(1996: 9) claim, she disagrees with Amosova and prefers to treat proverbs and familiar
quotations as phraseological, pointing out that a proverb can serve as the source of
particular idioms (A drowning man will clutch at a straw  clutch at a straw) (Arnold
1986: 179-80). She gives examples of familiar quotations, which come from literary
sources, such as Shakespeare's The rest is silence (Arnold 1986: 180). She mentions the
oft-cited properties of clichés (hackneyed, stale, poor compensation for a lack of thought or
precision), and she also recognizes the subjectiveness of the category: "[o]pinions may vary
on what is tolerable and what sounds an offence to most of the listeners or readers, as
everyone may have his own likes and dislikes" (Arnold 1986: 181).

2.17 Motivation

We briefly described motivation in 1.3. It is a key concept in cognitive grammar.


Moon (1998a: 61, 64) shows that 37 per cent of metaphorical expressions that she has
studied are transparent (i.e. motivated), 51 per cent semi-transparent (i.e. partially
motivated) and 12 per cent opaque (i.e. unmotivated), though the division of metaphors into
transparency groups is highly subjective, as Moon (1998a: 63-64) herself points out.
Motivation is a graded property. The relations between the literal and figurative
senses of scratch one's head and throw in the towel are more natural than that between the
two interpretations of shoot the breeze. Moon (1998a: 22-23) subclassifies the category of
metaphors into transparent (behind sb's back), semi-transparent (grasp the nettle) and
opaque (red herring). Howarth (1996: 24) mentions two categories, motivated and
unmotivated idioms, noting that there is a "diachronic scale" from figurative idioms via
expressions such as stop the rot or bury the hatchet, which are motivated for some and
unmotivated for others, to unmotivated idioms. He follows Cowie et al.'s (1993: xii-xiii)

78
typology, in which motivation is used to distinguish the two categories of pure idioms (kick
the bucket) and figurative idioms (close ranks). Cowie et al. (1993: xiii) claim that these
two types have fuzzy boundaries and they can merge. Carter (1987: 64) establishes four
major groups: 1) transparent (long time no see), 2) 'semi-'idioms, idiomatic similes (we are
all in the same boat, as sober as a judge), 3) semi-transparent (a watched pot never boils),
4a) opaque and overt (uninterpretable without contextual/cultural knowledge) (bottoms up),
and 4b) opaque and covert (to kick the bucket). Carter (1987: 62) claims that semi-idioms
require a process of analogizing that is direct rather than oblique and "the figurative
specialization occurs in one part of the expression not in the whole". In other words, they
are partially literal.
A cursory glance at examples of "képes beszéd" 'imagistic speech' and "körülírás"
'paraphrase' suggests that these categories are identified on the basis of figurativity.
Hadrovics (1995: 157) classifies fejet hajt 'resign oneself' (literally 'bow head') with both
literal and figurative interpretations as "képes beszéd" 'imagistic speech' (cf. Cowie et al.'s
(1993: xiii) figurative idioms), while nem köti az orrára 'not let sb into the secret' (literally
'not tie sth onto sb's nose'), which cannot be interpreted literally, is an example of
"körülírás". A deeper examination however reveals that both types can have possible or odd
literal interpretations: cserbenhagy 'leave sb in the lurch' and nagy fába vágja a fejszéjét
'bite off more than one can chew' (literally 'cut one's axe into a big tree') are both examples
of "képes beszéd" 'imagistic speech', while lába kel 'get lost' (literally 'grow legs') and talpa
alatt elfúj a szél 'get hanged' (literally 'the wind blows under sb's soles') are labelled as
"körülírás" (Hadrovics 1995: 159, 222). What is at issue here is whether there is a
(diachronic or synchronic) connection between the literal and the figurative interpretations
(Hadrovics 1995: 158-59). Diachronically or synchronically motivated expressions are
identified as "képes beszéd" 'imagistic speech'.
As we saw in 2.14, motivation can also be used as a supporting criterion, in order to
classify expressions as idioms or restricted collocations. In white lie and know the ropes, the
relation between the literal and figurative senses of white and ropes is probably not clear to
most native speakers. It is therefore likely that the given examples are closer to idioms than
to collocations. Since in most restricted collocations of the 'adjective + noun' schema it is
the adjective that is figurative, white lie, cold comfort and square meal can be distinguished

79
from restricted collocations only on the basis of motivation. They are idioms, since the
figurative item is unmotivated.
We assume three levels of motivation: 1) highly motivated (scratch one's head), 2)
partially motivated (as sober as a judge, grasp the nettle) and 3) unmotivated (the bee's
knees).
It is standard practice to use the term "homonymy" and their derivatives in
discussions of the relationship between the literal and figurative interpretations of idioms.
Thus, many idioms are said to have homonymous literal counterparts. Lipka (1992: 135-39)
discusses the criteria that are usually employed to distinguish homonymy from polysemy
and concludes that they form the endpoints of a continuum. Weinreich (1969: 42-43) holds
the same opinion and illustrates the various degrees of difference between the literal (i.e.
compositional) and the nonliteral senses with binomials such as Latin and Greek, bacon and
eggs, and milk and honey. Though only the last example is an idiom, Weinreich proves
himself to be a keen observer. Both Lipka (1992: 138-39) and Langacker (1987: 398, n. 20)
draw our attention to the subjectiveness of deciding whether the relation between two
words is polysemy or homonymy. From a cognitive point of view we can say that the
meanings 'an animal like a mouse with wings, that flies and feeds at night' (OALDb: 90) and
'a piece of wood with a handle, made in various shapes and sizes, and used for hitting the
ball in games such as baseball, cricket and table tennis' (OALDb: 90) are related simply
because they share the same phonological sequence (bat). However, this relatedness is
distant, in other words the link between the two meanings is not motivated. The view that
the relation between the literal and idiomatic interpretation of idioms is a case of homonymy
rather than polysemy can probably be traced back to the treatment of idioms as unmotivated
expressions. If we accept the view that many idioms are more or less motivated, we will be
able to draw a more realistic picture. The relation between the literal and idiomatic senses of
the bee's knees is homonymous, but it appears to be less appropriate to use the term
"homonymous" or "homonymy" for motivated idioms (scratch one's head). Note that
metaphorical and metonymic extension of the literal sense of a single word results in
polysemy, not homonymy; therefore, the term "polysemy" could be used for idioms by
analogy. The most obviously polysemous examples are twiddle one's thumbs, one's jaw
drops, tear one's hair out, scratch one's head, etc., which denote gestures and the emotions
associated with them. In addition to regarding idioms as unmotivated units, another factor

80
may have contributed to the avoidance of the terms "polysemy" and "polysemous". As we
noted in 2.14, an idiom is no longer an idiom when interpreted literally, and having different
names for the same string of sounds/letters (idiom and collocation for take a walk, or idiom
and free combination for grasp the nettle) has probably prevented analysts from developing
an awareness of dealing with the same expression.

2.18 Repetition of sounds and syllables

Apart from Arnold (1986: 177), Smith (1943: 174-75), Gläser (1986: 51, 114, 127)
and Hadrovics (1995: 108-13) devote some discussion to various types of repetition.
Alliteration (part and parcel, You can be sure of Shell), assonance (A stitch in time saves
nine) and rhyme (wine and dine, Early to bed and early to rise makes a man healthy,
wealthy and wise) lend a rhythmic quality to conventional expressions. Repetition can serve
stylistic or mnemonic purposes.
We have not found statistics about what percentage of phraseological units contain
repetition. It is common in proverbs, slogans, binomials and similes, but it is not a defining
property of these categories, and it can be found outside them (everything but the kitchen
sink).

2.19 Földes

Interested in second language acquisition, Földes (1987) explores phraseological


units from a cross-linguistic perspective. He expresses the view that most of the transfer
errors that intermediate learners make are phraseological, especially where the
corresponding expressions have similar structure and lexis (Földes 1987: 8-9). Howarth
(1996: 135-36, 156) shares this opinion in relation to advanced learners and puts the
emphasis on the difficulty inherent in the collocational end of the spectrum.
Földes (1987: 12-13) uses the terms "frazeologizmus" 'phraseologism' and
"frazeológiai egység" 'phraseological unit' as general labels for conventional expressions and
points out that "idiom" is not a synonym of these, since idioms comprise a subtype of
phraseological units characterized by noncompositionality. Single words used figuratively

81
and taking obligatory prepositions or case inflections as well as compounds are excluded
from phraseology (Földes: 1987: 14).
Földes (1987: 27-40, 44-47) also addresses the problem of cross-linguistic
equivalence, both from semantic and structural aspects. Equivalence can be found in various
degrees, since two units from different languages can have the same idiomatic meaning, the
same literal meaning or can be based on the same image, or they can share two or even all
three of these properties. The following types of equivalence are exemplified: 1) the
constituents (interpreted literally) and the structures of the idioms are equivalent, and 1a)
the idiomatic meanings are the same (a jobb keze valakinek - js. rechte Hand sein), 1b) the
idiomatic meanings are slightly different, 1c) the idiomatic meanings are rather different (jd.
hat Grütze im Kopf means 'be very clever', but каша в голове means 'be stupid'), 1d)
the idiom in one language has additional meanings that the idiom in the other language does
not have (seine Haut zu Markte tragen - vásárra viszi a bõrét, where both idioms have the
sense 'risk one's life', but the German phrase can also mean 'show one's naked body for
money'), 1e) the idiomatic meanings are opposite; (2) the idiomatic meanings differ only in
style (die Köpfe zusammenstecken is ironic, while összedugják a fejüket is neutral or
colloquial); 3) the idiomatic meanings and the structures in the two expressions are
equivalent, but the constituents are not exact equivalents (a tenyerén hordoz - jn. auf
Händen tragen). The use of the terms "poliszémia" 'polysemy' and "homonímia"
'homonymy' (cf. 1c above), and "hipero-hiponímia" 'hypero-hyponymy' (cf. 1d above) for
the meaning relation between idioms taken from two different languages is unorthodox
(Földes 1987: 31). The label "stilisztikai szinonímia" 'stylistic synonymy' (cf. 2 above) is
misleading, since it describes cases where the equivalent idioms differ in style (Földes 1987:
31).
Structural differences are also discussed, though Földes (1987: 37-40) uses a
mixture of structural and semantic criteria in the discussion. Thus, one group consists of
idioms where the German or Russian prepositions are not equivalent to the Hungarian
postposition (ül a négy fal között literally 'sit between the four walls' and in seinen vier
Wänden sitzen literally 'sit in his four walls').
From a formal point of view, a phraseological unit in one language can have an
equivalent in the form of a single word, a phraseological unit, or a paraphrase in another
language. Földes (1987: 29) claims that Russian лёгок на помине has no idiomatic

82
equivalent in Hungarian, yet the paraphrase that he gives ('vki hirtelen, váratlanul
megjelenik, amikor róla beszélnek vagy rá gondolnak' 'somebody turns up suddenly,
unexpectedly, when he/she is being talked about or thought about') can also be expressed
idiomatically as farkast emlegetnek, a kert alatt kullog (cf. O. Nagy 1985: 197). It is also
claimed that the word order in the Hungarian idiom csupa csont és bõr is the same as that in
its English equivalent, but the opposite is true (cf. skin and bones) (Földes 1987: 44).
Földes (1987: 41-47) briefly discusses some universal tendencies in idiom formation
and in formulaic language as well, noting the predominance of idioms describing negative,
unpleasant situations (Földes 1987: 43).

2.20 Hadrovics

Hadrovics (1995) surveys Hungarian phraseological units from a diachronic


perspective. He discusses the historical development of word-like and sentence-like
combinations, also showing obsolete variants and senses. Rich exemplification and the wide
range of word combinations analyzed are among the notable features of Hadrovics's study.
In the first part of the book phraseological units are discussed from a structural perspective,
which is followed by a survey based on semantic criteria (thematic groups and types of
idiomatic meanings). Finally, proverbs, familiar quotations, catchphrases and units borrowed
from foreign languages are exemplified. Hadrovics (1995: 133, 237) notes that some
phraseological units have highly variable lexical content (úgy pofon vágja, hogy
megemlegeti/hogy a nevét is elfelejti/hogy leröpül a feje, etc. 'give sb a big slap’), and he
also draws attention to the frequent shortening of proverbs. Priority is given to diachronic
aspects, which explains why some idioms describing situations that no longer occur
(kiteszik a szûrét 'turn sb out') are still identified as "képes beszéd" 'imagistic speech'
(Hadrovics 1995: 159). The historical focus can also be seen in the inclusion of many
obsolete phrases such as hogy egy sommába megmondjam 'in brief' (Hadrovics 1995: 121).
Hadrovics (1995: 108) is aware of the fuzziness of the boundary between what is
phraseological and what is not, it is therefore not surprising that we can find examples
whose phraseological status has been questioned or even rejected by other researchers.
Analysis of proverbs, catchphrases, idiomatic compounds and familiar quotations as part of
phraseology is justified, but there are a number of single words as well, simple or complex

83
words used figuratively (kockáztat 'risk', préda 'prey', pásztor 'shepherd', zaboláz 'bridle')
(Hadrovics 1995: 76, 212, 207, 196, 190). They are not usually classified as phraseological
in the literature. Hadrovics (1995: 44, 46, 50) also includes single words if they form
grammatical collocations with obligatory case inflections, e.g. dicsekedik vmivel 'boast of
sth', üt vkire 'take after sb', függ vmitõl 'depend on sth'.
For Hadrovics a genuine word-like phraseological unit has a figurative
interpretation. Combinations of a literal and a figurative word (cégéres lator 'arrant rogue',
égbekiáltó bûn 'cardinal sin', hétpróbás gazember 'utter scoundrel') are not genuine
phraseological units (Hadrovics 1995: 27-28, 76). Despite this some of the examples given
in the book have straightforward literal interpretations, which Hadrovics (1995: 190)
himself points out, e.g. "hol megállapodik, [lova] fékit rándítja, erõs kézzel hirtelen
megállítja", "az több fogságban mellettünk lévõ nemességet, patkószegeket vervén fejekben,
iszonyú halállal rekkentették el", "mert abban az idõben szintén úgy hagyigáltanak azok
parittyából a célhoz az miképpen mostan lövöldöznek nyilakkal és golyóbisokkal"
(Hadrovics 1995: 191, 211, 215).

2.21 Fernando

In Fernando's (1996) study of word combinations the emphasis is laid on the


discoursal function of idioms, with the three functions identified as ideational, interpersonal
and relational. Fernando (1996: 1) established these functions after Halliday's model. The
use of authentic data enables her to illustrate how the novel and the conventional interact in
language and to observe exploitations, which she calls "inventive imitation" (Fernando
1996: 145). Exploitation can lend context-specific meaning to idioms, especially ideational
ones (Fernando 1996: 119). A distinction is made between idioms and idiomaticity. The
latter refers to the habitual co-occurrence of words, whereas idioms are lexicogrammatically
highly restricted word combinations (Fernando 1996: 30). All idioms display idiomaticity,
but there are many non-idioms that also possess this feature (Fernando 1996: 30). Lexical
variability is a key feature in her classification. There are invariant units (devil-may-care),
units with restricted variance (explode a myth) and combinations with unrestricted variance
(catch a bus/plane/ferry, etc.) (Fernando 1996: 32). Some psycholinguistic aspects are also

84
mentioned, such as the minimum effort required for the use of conventional forms
(Fernando 1996: 75).
The bulk of the book comprises an analysis and exemplification of ideational,
interpersonal and relational idioms. Ideational idioms roughly correspond to what others
label word-like idioms functioning as noun-, verb-, adjective- or prepositional phrases
(Fernando 1996: 98). They denote various aspects of situations, such as participants,
actions, processes, etc. (Fernando 1996: 97). Compared with single words, they often have
more specific meanings (Fernando 1996: 99). Even idioms with a seemingly fairly generic
sense (every Tom, Dick, and Harry; every man and his dog, (all) the world and his wife)
have narrower reference than their non-idiom equivalents, 'everybody', 'average people' (cf.
Hungarian boldog-boldogtalan, fû-fa) (Fernando 1996: 101-102). Ideational idioms tend to
be imagistic and evaluative (Fernando 1996: 100). Although Moon's functional grouping is
different, she also finds that informational idioms are frequently evaluative (Moon 1998a:
239).
Fernando's (1996: 183) interpersonal idioms (Believe (you) me, See you later, by the
way) are basically formulae, having various pragmatic functions. They ensure smooth
interaction between the participants in a conversation, and they can structure discourse
(Fernando 1996: 154, 183). They are typically lower in information value, they have less
imagery and are less susceptible of exploitation than ideational idioms (Fernando 1996: 155,
160, 183). Fernando (1996: 157-83) analyzes the language of service encounters, small talk,
institutionalized wishes and greetings, as well as discussions and conversations to show how
interpersonal idioms mark politeness or conflict. Relational idioms (e.g. in order that, round
the clock) structure discourse by linking phrases or sentences, or they locate it in time
(Fernando 1996: 187-88). Although the three types of idioms are distinguished functionally,
they are claimed to have distinct lexicogrammatical or semantic characteristics as well. For
example, (personal) pronouns often mark interpersonal idioms (e.g. You're kidding), while
the schema 'A + N' is typical of ideational idioms (Fernando 1996: 187).

2.22 Lexical variation29

29
Other terms with the same meaning are "substitution" in Zgusta (1971: 144), "commutability" in Howarth
(1996: 41) and "collocability" in Barkema (1996:77).

85
Lexical fixedness implies that lexical substitution of the constituent words is limited
(a weight/load/*burden off sb's mind/*brain/*head) or impossible (a big/*large/*huge
fish). Though grammatical variation has been discussed separately, they often co-occur.
Insertion and omission modify both the lexical content and the structure of the expression.
Variants such as learn the ropes and teach sb the ropes also show both structural and lexical
differences.
Lexical variation must be distinguished from exploitation, which is used for stylistic
purposes or to introduce a pun. Non-institutionalized forms of idioms can sometimes be
found in graffiti. For example, Abstinence is the thin end of the pledge contains an
exploitation of the idiom the thin end of the wedge, and the grass is always greener on the
other fellow's grave is a non-institutionalized version of the grass is always greener on the
other side of the fence.30 Advertising slogans often use non-institutionalized forms of idioms
as well. For example, after a restructuring of one of the biggest Hungarian banks the
following advertisement appeared in a daily: Tiszta bankkal indulunk ( 'We start with a clean
bank'), which relies on readers' familiarity with the idiom tiszta lappal indul 'start with a
clean sheet', a case of quasi-homophonous substitution. 31 In our corpus we have found Now
it appears that the Germans are ready to swallow the Maastricht pill, a non-
institutionalized variant of swallow a bitter pill.
Moon (1998a: 170-74) surveys several types of exploitation and notes that some
nonce-uses have become institutionalized, such as call a spade a shovel. One form of
creative manipulation is lexical substitution, whereby constituent words are replaced by
more formal, literary or euphemistic synonyms, or by non-synonymous but context-related
words (when the brown stuff hits the fan, burn the candle at five ends) (Moon 1998a: 170-
73). In other types of exploitation a quasi-homophonous word is substituted (skip the light
fantastic). The purpose of exploitation is to provide humour or make the expression
contextually more appropriate (Moon 1998a: 170). For example, CCDI (335) shows that
Jack Robinson can often be replaced with other contextually relevant words in before you
could say Jack Robinson, e.g. She was on the phone to New York before you could say
long-distance.
As was mentioned in 2.10, we have found no statistics about the percentage of
lexically variable expressions. For Zgusta (1971: 144-45) lexical restrictedness is the most
30
The examples are taken from Rees (1979: 14, 31).
31
The example is from a daily newspaper: Hajdú-Bihari Napló, 27th January, 1999.

86
important criterion in identifying phraseological units, but in contrast to Fernando (1996) he
does not emphasize that it is a graded property. Carter (1987: 63) divides phraseological
expressions into four groups based on lexical variation: unrestricted collocations (take a
rest/a holiday/a walk, etc.), semi-restricted collocations (harbour
doubt/grudges/uncertainty/suspicion), familiar collocations (unmitigated disaster), and
restricted collocations32 (stark naked). Moon (1998a: 145-50, 158-70) discusses frames,
free realizations and idiom schemas. In frames "clusters of FEIs share single or common
structures, but the realizations of one constituent vary relatively widely, though usually still
within the bounds of a single lexical set" (in a fix/hole/mess/paddy/spot) (Moon 1998a:
145-46). "Free realization" is Moon's term for schematic idioms, in which "the lexis is
routinely varied without any apparent limits" (Moon 1998a: 158). Idiom schemas have a
metaphor in common and cognate lexis, but they do not necessarily have a fixed structure or
lexis (fan the fire/fires/flames, add fuel to the fire/flame/flames, fuel the
fire/fires/flame/flames) (Moon 1998a: 163). It seems that Moon (1998a) establishes five
types of idioms on the basis of the degree of lexicogrammatical variation: 1) invariable
idioms, 2) idioms with restricted variation (play/keep one's cards close to one's chest, not
have the foggiest idea/without the foggiest idea), 3) variable idioms with a fixed structure
called "frames" (in a fix/hole/mess/paddy/spot), 4) schematic idioms, i.e. highly variable
idioms with a fixed structure (cf. 2.12), and 5) highly variable idioms, called "idiom
schemas", in which both the structure and the words are variable (fan the fire/fires/flames,
add fuel to the fire/flame/flames, fuel the fire/fires/flame/flames).
We will establish four levels of lexical restrictedness: 1) highly variable units,
including highly schematic idioms (take a rest/a holiday/a walk, etc.), 2) variable (in a
fix/hole/mess/paddy/spot), 3) restricted (play/keep one's cards close to one's chest) and 4)
completely fixed (lock, stock and barrel).
In 2.14 we briefly mentioned lexical variation as a property that is frequently used in
distinguishing restricted collocations from idioms. Collocations are usually described as
habitual co-occurrences of words (cf. Cruse 1986: 40, Moon 1998a: 26, Benson 1985: 61),
but the term "collocation" sometimes covers ad hoc combinations as well (cf. Howarth
1996: 37, Fernando 1996: 250). Cowie (1994: 3169) claims that collocations have
"arbitrary limitations of choice at one or more points", as in light/?heavy exercise or slash

32
This term is not to be confused with Cowie et al.'s (1993: xiii) "restricted collocations".

87
one's wrists/*throat, and contrasts them with free combinations such as drink one's tea, in
which selectional restrictions constrain the choice of words, and the rules of co-occurrence
can be stated in terms of general features based on the meanings of the words. We will
follow this practice, but one can also find that the term "(open/free) collocation" is used for
free combinations (cf. Cowie et al. 1993: xiii, Howarth 1996: 34). Carter (1987: 54) also
contrasts selectional and collocational restrictions when he writes "[t]he examination of
collocational ranges of items begins where semantic analysis of selection restrictions leaves
off".
Free combinations often permit the replacement of their constituents with a number
of words, without changing the meaning of the other constituent (fill/empty/drain, etc. the
sink, fill the sink/basin/bucket, etc.) (Cowie et al. 1993: xiii). However, a high degree of
lexical restriction imposed on co-occurring words is not necessarily arbitrary. Allerton
(1984: 27) discusses examples such as purse one's lips, gruff voice, shrug one's shoulders.
If we try to replace the nouns in these combinations while retaining the sense of the other
element, we will see that hardly any other nouns can be substituted. Yet we can easily
account for co-occurrence restrictions if we know the meaning of purse, gruff and shrug.
Arms akimbo is a similar example, since the impossibility of replacing arms follows from the
meaning of akimbo. The lexical restrictions on these examples are therefore semantically
motivated, not unpredictable. Similarly to Zgusta (1971: 141), we will classify them as free
combinations. The combination broad range is different from the above examples, not only
because some degree of substitution of the noun is possible (broad spectrum, coalition,
etc.), but also because the unacceptability of ?broad choice is not predictable on the basis of
selectional restrictions or the meaning of broad. For example, both wide range and wide
choice are common as can be seen in 9.7, and wide seems to have the same meaning in
these combinations as broad.

2.23 Howarth

Howarth (1996: 2, 141) is concerned with how non-native advanced level speakers'
use of restricted collocations compares with native speakers' norms. He ventures into
uncharted territory by analyzing academic writing from a phraseological perspective. Before

88
a detailed analysis of native and non-native data, he makes a survey of phraseological
investigations, including research into the processing of idioms (Howarth 1996: 3). The
focus of researchers has been on idioms, as a result of which there is no fully developed
theoretical framework for and very little psycholinguistic data on collocations (Howarth
1996: 3). He urges a shift of attention away from idioms, since he finds that idioms are not
the dominant type of word combinations in authentic texts, whereas restricted collocations
are much more numerous, and most errors in non-native writing occur not in idioms but in
restricted collocations (Howarth 1996: 121, 133, 162).
Howarth (1996: 11, 32, 34) points out that it is essential to arrange the criteria for
classifying combinations of words in the following hierarchical order: syntactic structure,
institutionalization, semantic opacity/transparency, commutability (i.e. lexical substitution),
semantic unity and motivation. It seems that his hierarchy is determined by his method of
attempting to narrow down the focus of investigation gradually by moving from the free end
of the spectrum towards the opaque end and his desire to reproduce the continuum model
(Howarth 1996: 33, 46).
Howarth (1996: 11-12) first separates sentence-like from word-like units, since he is
concerned with restricted collocations. Within word-like units, he argues for using the
criteria of syntactic structure (a complex feature including well-formedness, grammatical
structure and transformational restrictions) and institutionalization before others, which
helps the researcher discard examples that are not syntactic units and many syntactic units
that are free rather than restricted (Howarth 1996: 34-38). Although syntactic structure and
institutionalization are discussed together with the other criteria, they are not offered as
relevant for the classification of phraseological units (Howarth 1996: 34, 45). It is claimed
that the criterion of lexical restriction should apply before semantic considerations of
opacity, since the set of opaque units can be seen as a sub-set of lexically restricted
combinations (Howarth 1996: 32). Despite this claim, their order is reversed in the detailed
discussion and in the summary table (see below), which is supposed to show the ideal order
of the criteria (Howarth 1996: 38-45, 47).
Howarth's (1996: 38) criterion of semantic opacity/transparency seems to be a
complex criterion subsuming several features, of which one or the other is in focus
depending on whether he wishes to distinguish idioms or restricted collocations from the
rest of word strings. Opacity/transparency refers to the analyzability of a combination and

89
semantic specialization, i.e. whether the constituents have literal or nonliteral meanings
(Howarth 1996: 38-41). These "subcriteria" are supposed to identify restricted collocations
as different from idioms on the one hand and free combinations on the other hand.
However, this identification cannot always be successful, since, as we saw in 2.14, the
literal/nonliteral distinction cuts across the division between free and restricted, though
nonliteralness tends to imply a certain degree of restrictedness (Cowie 1994: 3169, Howarth
1996: 101). At the same time, several idioms are analyzable. Howarth (1996: 23-24)
introduces additional complication into his criteria by using "semantic unity" as an
equivalent to the subcriterion of analyzability and "opacity/transparency" as an equivalent to
the criterion of motivation in his discussion of idioms. Kick the bucket is treated as
semantically unanalyzable, i.e. semantically unitary, and "transparency" is paraphrased as
"the extent to which [a phraseological unit] is motivated" (Howarth 1996: 23-24).
Howarth's (1996: 47) tabular summary of the criteria provided at the end of his discussion is
reproduced below:

well- institution- specialized collocation- semantically unmotivated


formed alized element ally restricted unitary
free collocations  /x x x x x
restricted coll.     x x
figurative idioms      x
idioms      

The ordering of the criteria of syntactic structure ("well-formed" in the table above)
and institutionalization with respect to each other is also problematic if we apply these
criteria to phraseological units only. But Howarth (1996: 37) applies them to free
combinations as well and takes the view that all combinations of words (except for
extragrammatical idioms) are grammatically well-formed. This is a simplification, since
syntactic well-formedness and institutionalization yield overlapping groups.
Howarth (1996: 47) uses six criteria to classify combinations of words, two of
which serve to separate free combinations from phraseological units. In fact, only four
criteria are used to classify phraseological units: 1) figurativity ("specialized element" in the
table above and subsumed under "semantic opacity/transparency" in his discussion), 2)
lexical variation ("collocationally restricted" in the table and "commutability" in his

90
discussion), 3) analyzability ("semantically unitary" in the table and subsumed under
"semantic opacity/transparency" in his discussion) and 4) motivation.
Using the criteria mentioned above, Howarth (1996: 84-132) analyzes 'verb + noun'
restricted collocations found in authentic texts, noting native speaker deviations from the
norm. He refines the criterion of lexical variation in an attempt to provide a classification
that will take account of degrees of restrictedness, which are claimed to be captured well
with the help of lexical restrictedness (Howarth 1996: 102). Five levels are established, with
Level 1 (a small number of synonymous verbs + an open set of nouns, eg.
adopt/accept/agree to a proposal/suggestion/recommendation/convention/plan etc.) on the
boundary of free and restricted combinations. Level 2 is characterized by a small range of
nouns in combination with a small number of synonymous verbs (introduce/table/bring
forward a bill/an amendment). If only one noun can co-occur with a given set of
synonymous verbs (play/take heed), the phrase belongs to Level 3, while complete
restriction on the verb in combination with a small number of nouns (give the
appearance/impression) is typical of Level 4. Levels 3 and 4 could be seen as showing the
same degree, as Howarth (1996: 103) himself notes. The rationale behind separating these
levels is that the starting point for the speaker is claimed to be the noun, and the most
immediate mark of restrictedness is the appropriateness of a single verb in the given sense.
Level 5 includes examples such as curry favour, which could be treated as restricted
collocations or idioms (Howarth 1996: 115-16).
Howarth (1996: 126-28) identifies two types of collocational errors within lexical
substitution: blends (*make an assurance) and overlaps (*propose caution). Blends are the
result of combining a verb from one collocation with a noun from another (e.g. make a
promise, give an assurance  *make an assurance), while collocational overlaps are more
complex, involving a cluster of verbs and nouns subject to arbitrary restrictions with
unpredictable gaps: (e.g. propose an action, advise an action, advise caution  *propose
caution). In native-speaker texts deviance is very low and mostly of the blend type
(Howarth 1996: 131). Non-natives use more free combinations and fewer restricted
collocations than natives, and they have been found to have more problems with
overlapping collocations than natives (Howarth 1996: 158, 193). One of the most surprising
findings is that there is no correlation between the number of collocational errors and the
language proficiency level of the non-native student (Howarth 1996: 159).

91
Due to the ill-understood nature of collocations and the limited phraseological
research, teachers cannot give learners the expected assistance in this area (Howarth 1996:
189, 193). It is proposed that rather than focusing on the teaching of a large number of
actual collocations, we need to draw learners' attention to the general mechanism of
collocation, which inevitably involves learning collocations (Howarth 1996: 168).

2.24 Relation between the criteria

2.24.1 Introduction

We agree with Howarth's (1996) claim that a combination of criteria can lead to a
more successful identification of different types of expressions. Before discussing how the
criteria have been combined by various analysts, it is necessary to survey their relations. Our
discussion may have given the impression that the properties are independent, but many of
them are in fact related.

2.24.2 Compositionality and analyzability

Cognitive grammar's separation of analyzability from both compositionality and


motivation does not entail that the three notions have nothing to do with one another. If the
discrepancy between the compositional/literal interpretation and the
phraseological/idiomatic meaning is slight, the degree of analyzability is high. For example,
gin and tonic and I'm sorry to say are to a large extent compositional, and so they are
analyzable. However, partial compositionality can still co-occur with full analyzability:
explode a myth 'prove wrong + a myth', meet the requirements 'satisfy + the requirements'.
Noncompositional expressions are in a number of cases also fully analyzable: back the
wrong horse 'support + the + wrong + person' and mend one's fences 'improve + one's
relationship'. In other cases noncompositionality goes with unanalyzability, as in over the
moon 'very happy'.

92
2.24.3 Analyzability and lexicogrammatical variation

Nunberg et al. (1994: 509) express the view that passivization and other structural
variation (cf. 2.10) can be explained with the help of the distinction between analyzable and
unanalyzable idioms as well as other factors. What these other factors might be is
summarized as follows: "interacting factors, mostly having to do with the nature of the
discourse function of particular constructions and the particular figures underlying various
idiomatic combinations [i.e. analyzable idioms], have a critical role to play as well"
(Nunberg et al. 1994: 509). This dissertation will not attempt to subject idioms to a detailed
analysis of discourse functions, because it is beyond the scope of the present work. One of
the aims of phraseological research in the future could be to discover what factors
determine the grammatical restrictedness of conventional expressions and whether
analyzability has a special place among these factors. Nunberg et al. (1994: 512) also claim
that "syntactic variability [...] does not always require semantic analyzability".
The results of psycholinguistic research support the view that analyzable idioms are
judged to be syntactically more variable than less analyzable idioms, with abnormally
analyzable units being closer to unanalyzable ones (Gibbs 1990: 425, Gibbs 1994: 281).
Gibbs (1994: 281) concludes that syntactic variation "can at least partially be explained in
terms of an idiom's semantic analyzability". The relation between analyzability and variation
is not simple, because not all analyzable idioms are variable and not all variable idioms are
analyzable. Jackendoff (1997: 170) notes that Postal gives some analyzable examples that
cannot be passivized: raise hell ('cause' + 'a serious disturbance') and give the lie to X
('show' + 'X' + 'to be a falsehood').
Lexical variation seems to be related to analyzability as well. Nunberg et al. (1994:
504) claim that analyzability can explain why "the same NP (with a single idiomatic
interpretation) may occur with more than one idiomatic verb", as in keep/lose/blow one's
cool. Some of their examples, however, are wrong. In search/hunt/look for NP high and
low the noun phrase does have an independent sense, but it does not need to be idiomatic.
In talk/argue/complain until you are blue in the face the verb is not idiomatic, and the noun
phrase does not seem to have an independent idiomatic meaning. Nor do the noun phrases
have independent meanings in stop/turn on a dime and should/ought to/need to/better
have/get one's head examined/seen to/looked at/tested.

93
Gibbs (1990: 426, 1994: 282-83) claims that replacing lexical items in analyzable
idioms seems to be far less disruptive, so that punt the pail instead of kick the bucket is less
acceptable than burst the request instead of pop the question. This was the result of several
psycholinguistic studies in which speakers were presented with expressions that differed
from the institutionalized forms of idioms in that either the verb or the noun constituent had
been replaced with a synonymous word.

2.24.4 Analyzability and figurativity

We have examined the relation between analyzability and compositionality and


concluded that a high degree of compositionality will lead to full analyzability.
Compositionality and figurativity are related, since a compositional expression is usually
interpreted literally. It is therefore not surprising that there is a similar relationship between
analyzability and figurativity. Fully literal word combinations (gin and tonic, See you later)
are usually analyzable to a great extent, while partially literal idioms are partially or fully
analyzable, since the literal sense saliently contributes to the idiomatic meaning of the whole
(promise the moon/earth 'promise + something impossible', dressed to kill 'dressed +
smartly/glamorously', look daggers 'look + (very) angrily', light as a feather 'very + light'
etc.). Not all similes are partially literal. In read sb like a book the verb is used figuratively.
Literalness implies a certain degree analyzability, but not vice versa.

2.24.5 Analyzability and structure

This relationship has not been examined in the phraseological literature, but we
claim that it is important. There may be a tension between the structures of a combination of
words or the functions of the constituents, when they are interpreted literally and
idiomatically. This tension may make it less likely that the native speaker attaches the given
meaning to the constituent. Weinreich (1969: 34) notes that in fight tooth and nail and go
at sb hammer and tongs the binomial parts have different functions. Interpreted figuratively
they are adverbials, while interpreted literally they are noun phrases functioning perhaps as

94
complements. Noun phrases can function as adverbials in standard English, but mostly
nouns with a temporal meaning can function as adverbials of time. Further examples with
different structures/functions in their literal and figurative senses are given in the table
below. The structures and the functions in the idiomatic interpretation have been established
on the basis of the paraphrases given below, which may not be the only possible ones, but
we have attempted to formulate them in a way that their structure is as close to the literal
structure as possible. Differences are shown in bold. ADV stands for 'adverb' or 'adverbial',
C for "complement" and P for "preposition" or "particle".

expression structure/function interpreted structure/function interpreted


literally figuratively
step/tread on sb's toes 'V + P + NP/V + ADV' 'offend/annoy sb': 'V + NP/V + O',
the figurative NP/O = the
determiner of the literal NP/the
ADV
hear/see the last of 'V + NP + P/V + O (?)' 'hear/see for the last time': 'V + P +
NP/V + ADV'
the figurative adjectival sense
'last' = the literal noun last
laugh/shout etc. one's head off 'V + NP + P/V + O' 'laugh/shout etc. very noisily/very
much': 'V + ADV/V + ADV'
rain cats and dogs 'V + NP/V + O (?)' 'rain heavily': 'V + ADV/V +
ADV'
strike it rich 'V + PRONOUN (it) + A/V + O + 'become rich (unexpectedly)': 'V +
A' A + (ADV)/V + C + (ADV)'
over the moon 'P + NP/ADV' 'very happy': 'ADV + A/C'

2.24.6 Analyzability and motivation

Motivation and analyzability are also related, since metaphorical or metonymic


motivation can lead to the recognition that the constituents of the idiom have independent
meanings. The metonymic motivation of from the cradle to the grave contributes to the
analyzability of this idiom: the cradle stands for birth and the grave stands for death.
Grady (1997: 83) argues that there is a metaphorical relation between swallowing
and accepting (not only in an intellectual sense), and this helps us realize that swallow means
something like 'accept (often unwillingly)' in swallow a bitter pill. There are a number of

95
non-idiom examples in which swallow can be used in the same sense, as in I was surprised
that he just sat there and swallowed all their remarks (OALDb: 1312). The meaning
'accept' is also found in the idioms swallow the bait and strain at a gnat and swallow a
camel, and this is additional support for the link between swallowing and accepting, which
also provides partial motivation for ram/shove/force sth down sb's throat. Grady (1997: 83-
84) explains that the motivation behind this link is not surprising, since swallowing food
means we have made a decision to accept it in our bodies, and swallowing entails an
absence of resistance. Grady is right that in prototypical cases we want to swallow the food,
once we have made the decision, but it is not always a case of making a conscious decision
(sometimes we swallow accidentally), and it is not always a case of wanting to swallow
(sometimes we do not want to eat more).
It is probable that swallow a bitter pill and swallow the bait are easily analyzed by
speakers in the following way: swallow 'accept' + a bitter pill 'a difficult/unpleasant
fact/situation' (or even bitter 'difficult/unpleasant' + a pill 'a fact/situation') and swallow
'accept' + the bait 'something that has been said/offered to tempt somebody'. The relation
between the literal meanings of bitter and bait and their figurative senses seems to be
motivated, while the link between 'pill' and 'fact/situation' is less so. Langacker (1987: 94)
also shows that the link between 'bag' and 'concealment' is natural and salient, while that
between 'cat' and 'secret' is not clear in let the cat out of the bag. The motivatedness of the
idiom constituents varies.
The motivatedness of a relation does not necessarily depend on whether the
constituent can be used outside the given word combination in the same (or a very similar)
figurative meaning. For example, neither cat nor bag seems to have the given figurative
senses outside let the cat out of the bag and its variant the cat is out of the bag, yet bag is
more naturally associated with concealment. However, examples for a motivated link can
often be found outside a given unit as well. Moon (1998a: 201) points out that the
figurative meaning that rock has in rock the boat 'upset a calm situation' (CCED: 173) is
institutionalized outside the idiom, and this suggests that 'upset' is naturally assigned to
rock, while the meaning 'a calm situation' is less easily attached to the boat. (Note that the
boat is used in a related, though not exactly the same, sense in be in the same boat.) Moon
(1998a: 201) also claims that it is easier to see the similarity between revealing a secret and
spilling something than between beans and secrets, since spill is used in the same sense in

96
spill one's guts and spill it. There is however a difference between the examples with spill
and those with swallow. In the first case the same (or a very similar) sense can be found in
other idioms but not outside those idioms, while swallow can be used in the same figurative
sense outside idioms as well. Another analyzable example is go through the mill 'experience
a very difficult period or situation' (CCDI: 260), in which go through is easily linked with
the sense 'experience', since it can be used outside idioms in the same sense.
Despite their relatedness, motivation and analyzability are independent properties.
As we have seen, constituents with a low degree of motivation can have independent senses
(a pill 'a fact/situation', cat 'secret'). At the same time, motivation does not necessarily imply
analyzability, as the following idioms of anger show. Hit the roof/ceiling, flip one's lid and
blow one's stack are (partially) motivated, but the constituents do not seem to carry
independent meanings. Idioms showing the same degree of analyzability can have different
degrees of motivation, as in happy as a pig in muck and happy as a sandboy. Both are
partially analyzable (happy 'happy' + as a pig in muck 'very', happy 'happy' + as a sandboy
'very'), but happy as a pig in muck is more motivated than happy as a sandboy. At first
sight clear as a bell is less motivated than clear as crystal, clear as day, or clear as mud.
Clear as crystal and clear as day are relatively motivated, and clear as mud can be
considered to have a similar degree of motivation, since it means 'not clear at all'. This idiom
is oxymoronic and ironic. Moon (1998a: 151) claims that clear as a bell has a different
meaning, since it is used of sounds. OALDb (101) and ODEI (18) support her claim, but
CCDI (73) and CCED (142) contain examples in which the idiom is used in the same sense
as clear as crystal/day. When used of sounds, clear as a bell is motivated and metonymic,
a bell stands for the sound of a bell.

2.24.7 Idiosyncrasy and other properties

There seems to be no relation between extragrammaticality and analyzability, since


some expressions are fully analyzable (believe you me), some are partially analyzable (in the
know), others are unanalyzable (sleight of hand).

97
Fraser (1970: 31) claims that he has found no transformational variation in
extragrammatical idioms.33 Whether this is true or not depends on what is considered to be a
transformation. Most of Moon's (1998a: 81-83) examples are invariant, no
lexicogrammatical variants are given, except for go (the) whole hog, in which the article can
be deleted, though only in American English (cf. CCDI: 209, CIDI: 423). In our
terminology, and in early generative grammar, deletion is a transformation. Similarly, you is
deletable in believe you me according to OALDb (101), but not according to CCED (142).
If we accept OALDb's claim, the extragrammatical idiom can undergo the deletion
transformation.
The previous example reveals that there are cases where one of the variants of an
idiom is extragrammatical, while the other variant(s) is (are) not idiosyncratic. Believe you
me and believe me can be considered variants, but the latter is not extragrammatical. Tear a
strip off sb and tear sb off a strip are also variants. Considered in isolation, both conform to
grammatical structures, but tear sb off a strip would be an ill-formed variant of tear a strip
off sb outside the idioms. The idiom for the sake of sth also has an extragrammatical
variant: for sth's sake.
Though Fraser (1970: 30) claims that extragrammatical idioms have no literal
interpretation, we hold that many of them can be interpreted literally (dog eat dog, rain cats
and dogs, I kid you not, if it ain't broke, don't fix it, cf. also Hungarian se lát se hall etc.).
There seems to be no relationship between extragrammaticality and motivation.
Idiosyncratic idioms can be motivated (long time no see), partially motivated (the biter gets
bit) or unmotivated (hard done by).
We suggest that several unique words carry independent meanings, as shown in the
table below, which contains most of Moon's (1998a: 78-79) examples :

unique word meaning of unique word word combination

amok run amok


beck be at sb's beck and call
boot to boot
cahoots 'co-operation', 'collusion' in cahoots with sb
cropper come a cropper
curry 'try to gain' curry favour
33
Fraser (1971: 30-31) uses the term idioms "without literal counterpart", but all the examples are
extragrammatical idioms.

98
dint 'means' by dint of
dudgeon 'anger', 'resentment' in high dudgeon
fettle 'health', 'condition' in fine/good fettle
fro 'back again', 'backwards' to and fro
grist grist to sb's mill
haywire 'wrong', 'confused', go/be haywire
'disordered'
hue hue and cry
ken 'knowledge' beyond one's ken
kibosh 'end' put the kibosh on sth
kilter 'good condition/health' in/out of kilter
kith 'friends' kith and kin
loggerheads at loggerheads
lurch 'difficult situation' leave sb in the lurch
purview 'range' within the purview of sth
queer 'spoil' queer sb's pitch
scruff 'nape/back (of the neck)' by the scruff of sb's neck
shebang 'affair', 'matter' the whole shebang
sleight sleight of hand
slouch be no slouch (at sth)
snook cock a snook
spick spick and span
tenterhooks on tenterhooks
trice 'instant', 'moment' in a trice
truck 'dealings', 'connection' have no truck with sb
umbrage 'offence' take umbrage
wend wend one's way
yore 'old (times)', 'the past' of yore

To test our intuition, we have asked our native informants in the form of a cloze test
(cf. Appendix 2) and then gave them various idioms with certain parts underlined and asked
them to supply the meaning. The cloze test was used to test whether our informants
recognize unique words with the help of the context. Most of them did. The test was also
useful to check what kind of meaning our native speakers attach to a unique word. For
example, collusion or dealings given instead of cahoots suggests that this unique word
carries an independent sense. Although we have suggested no independent senses for amok,
cropper and slouch, we may have been too cautious, as our informants seem to attach
independent meanings to these unique words, as can be seen from their answers.
Moon (1998a: 78-80) shows the following lexicogrammatical variants of her idioms
containing unique words: in fine/good fettle, go/be haywire, out of kilter/off kilter, at/from

99
the outset and for sb's/sth's sake / for the sake of sb/sth. NSOED (761, 939, 1146, 1488,
3450) gives these variants of Moon's (1998a) examples: in high/great etc dudgeon, in
fine/good etc fettle, for/to the mill, in (good) kilter, give umbrage (to sb)/take umbrage (at
sth). In our corpus dudgeon occurs with the adjective high, but not with great (cf. 9.8).
Fettle collocates with fine and good (cf. 9.1), though a larger corpus might yield other
collocates as well. No examples of grist for the mill, in (good) kilter and give umbrage can
be found in our corpus. Most lexically idiosyncratic idioms seem to be invariable.
Lexically idiosyncratic expressions cannot have (fully) literal interpretation, since
unique words have no literal meanings. It is difficult to judge the degree of motivation in
these expressions, as the prerequisite to any judgement is the separation of literal and
figurative meanings. It seems that lexically idiosyncratic units can be partially motivated
(leave sb in the lurch) or unmotivated (come a cropper).34
Grammatical and lexical idiosyncrasy sometimes combines in one and the same
idiom: stand sb in good stead, come a cropper, sleight of hand.

2.24.8 Lexicogrammatical variation and other properties

No relation has been found between variation (lexical and transformational) and the
frequency of the phraseological unit (Moon 1998a: 120-21). This is in accordance with
Barkema's (1996: 81) finding, which is one of the results of a study of the relationship
between the grammatical (transformational and inflectional) variation of two thousand noun
phrase units and various factors such as compositionality, lexical variation, structure,
syntactic function, genre, medium and frequency. Of these factors only the syntactic
structure of the unit ('premodifier + head', 'head + PP postmodifier', etc.), syntactic function
(subject, direct object, etc.) and medium (spoken vs written) were found to have a
significant influence on grammatical variation (Barkema 1996: 81). In particular, 'head + PP
postmodifier', 'premodifier + head' and 'head + clause postmodifier' units are claimed to be
more variable than 'determiner + head' or co-ordinated units. The grammatical variability of
the noun phrase was found to be higher when the phrase functioned as a subject, subject
complement or direct object, while variability was less when the phrase functioned as

34
Come a cropper in the sense 'fall over/off' is not motivated.

100
prepositional complement, still less when it functioned as a single utterance. Variability was
low when the function was adverbial, indirect object or premodifier.
Interestingly, no significant relation was found between the degree of
compositionality and the degree of grammatical variation, though this may be due to the
narrow scope of the investigation. At the same time, fully compositional expressions were
found to be more sharply different from less than fully compositional items, than any of the
partially compositional or noncompositional types from each other (Barkema 1996: 76).
Another surprising finding was that lexically fixed or semi-fixed expressions showed a
tendency to be more grammatically variable than lexically highly variable units (Barkema
1996: 77).
From the discussions in 2.14 and 2.22 we can conclude the following: figurativity
tends to go hand in hand with lexicogrammatical restrictions, but (arbitrary)
lexicogrammatical restrictions can be found in all types of word combinations, literal
(Good/*Beautiful/*Nice morning, slash one's wrists/*throat), partially figurative (break
one's journey/*trip/*voyage) and fully figurative (a big/*large/*huge fish).
Moon (1998a: 166-70) links metaphorical motivation with lexicogrammatical
variation to some extent, claiming that a metaphorically motivated idiom schema (in her
terminology) has several variants, but the metaphor is common and the open-class
constituents must belong to a particular lexical set.

2.24.9 Schematicity and other properties

Schematicity always implies lexical variation. The more schematic an expression is,
the more it is variable. A high degree of lexical substitution of one constituent in the idiom is
probably the result of that constituent being schematic, but more restricted substitution does
not presuppose schematicity.
The schematic part of a phraseological unit is usually literally interpreted. In life
with a capital L the word life has its literal meaning. However, occasionally the schematic
parts have no independent meanings, as in a few clowns short of a circus (cf. 2.12).

101
2.24.10 Motivation and other properties

The relationship between motivation and analyzability as well as that between


motivation and lexicogrammatical variation has been discussed. Literalness always
presupposes motivation, but figurative items can have various degrees of motivation. The
claim that many idioms are motivated is not to be interpreted as a claim for
compositionality.
We cannot predict the meaning of over the moon, despite its (partial) motivation by
the metaphor HAPPY IS UP. Happiness is not the only notion related to the concept 'up',
success (on the up) and control (gain the upper hand) are also 'up'. Furthermore, we cannot
be sure which idiom is motivated by a given conceptual metaphor, if all we know is the
literal meanings of the constituents and their combination. For example, grasping something
metaphorically denotes understanding something (UNDERSTANDING IS GRASPING), as can be
seen in get hold of the wrong end of the stick. The word grip is a synonym of grasp, and the
same metaphor motivates (to a lesser extent) the phrase come/get to grips with sth, but
keep a firm/tight grip/hold on sth denotes control, not understanding (ODPV: 197).
Although motivation and compositionality are not unrelated, they are distinct notions.
Despite partial motivation the above examples are unpredictable.
To some extent discrepancy and motivation are related. As discrepancy increases,
motivation decreases (Langacker 1987: 464). The motivation of gin and tonic is high, it is
less in scratch one's head and even less in let the cat out of the bag.

2.25 Combination of the criteria used in overall classifications

In overall classifications of combinations of words semantic criteria are used in


combination with lexicogrammatical variation in most cases, but the emphasis varies from
researcher to researcher.
As we have seen, Carter (1987) establishes three separate continua: 1) degrees of
lexical variation: unrestricted collocations (take a rest/a holiday/a walk, etc.), semi-

102
restricted collocations (harbour doubt/grudges/uncertainty/suspicion), familiar collocations
(unmitigated disaster), restricted collocations (stark naked); 2) degrees of grammatical
variation: flexible (break sb's heart), regular with certain constraints (drop a brick),
irregular (go it alone); 3) degrees of motivation: transparent (long time no see),
'semi-'idioms, idiomatic similes (we are all in the same boat, as sober as a judge), semi-
transparent (a watched pot never boils), opaque and overt (uninterpretable without
contextual/cultural knowledge) (bottoms up), opaque and covert (to kick the bucket).
However, he sees the separation of the various criteria as rather artificial. This is clear from
his comments. As he writes "it will be argued that the notion of the cline can continue to
help us to range [phraseological] units in terms of sets of continua with fixed points but
several intermediate categories" (Carter 1987: 63). And he adds that "it is necessary to
separate the clines but it is also clear that there are points of intersection and overlap
between the clines which allow us to define the most fixed expression as those which are
'closed' in more than one category" (Carter 1987: 64). It seems that he would calculate the
degree of the overall fixedness of a unit on the basis of the three criteria and classify
expressions ultimately on the basis of their overall frozenness, but no such classification is
offered. The only alternative grouping shown is similar to Alexander's typology, which is
developed using a mixture of syntactic and pragmatic criteria (Carter 1987: 60, Howarth
1996: 8).
Cowie et al. (1993: xii-xiii) contrast restricted collocations with open collocations
(i.e. free combinations), figurative idioms (close ranks) and pure idioms (kick the bucket) on
the basis of three criteria: figurativity, lexical variation and motivation. In free
combinations each constituent is typically used in a literal sense, and the phrase allows a
high degree of lexical variation, while in restricted collocations one word (in the case of
two-word units) is used in a figurative sense, and the other word has a literal meaning.
Figurative idioms are to some extent motivated, while pure idioms are unmotivated. The
idioms and collocations are also classified according to their structure and syntactic
function (do sb credit - 'verb + indirect object + direct object) (Cowie et al. 1993: xxix-
xxxvii).
Howarth (1996: 38-47) combines four criteria in his classification of phraseological
units: figurativity, lexical variation, analyzability and motivation, with the emphasis on
figurativity and lexical variation. In Fernando's (1996: 31-33, 35-36, 71-72) typology the

103
emphasis is on the combination of figurativity and lexical variation, which identifies the
following types: nonliteral invariant (smell a rat), nonliteral restricted (seize/grasp the
nettle), semi-literal invariant (drop names), semi-literal restricted (explode a
myth/theory/notion/idea/belief), semi-literal unrestricted (catch a bus/plane/ferry, etc.),
literal invariant (arm in arm), literal restricted (happy/merry Christmas) and literal
unrestricted (weak/strong/black, etc. coffee). This scale shows the tendency, mentioned
earlier, that figurativity implies restrictedness. For Fernando (1996: 31-32, 43, 56) the
primary criterion of idiomaticity is lexical variation, but she resorts to grammatical variation
in cases where the expressions have the same degree of lexical restrictedness. She also
offers a grouping based on discourse function alone (ideational idiom denoting action
(spill the beans), ideational idiom denoting attributes (cut-and-dried), interpersonal idiom
expressing directive (believe you me), interpersonal idiom expressing rejection (you're
kidding), concessive relational idiom (at the same time), etc.) (Fernando 1996: 72-74).
Moon (1998a: 19-25) uses lexicogrammatical variation and idiosyncrasy,
motivation and pragmatic function combined with various criteria, but she does not
combine the three major criteria. Although dual classification is discussed briefly, the three
criteria yield three separate groupings.
Gläser (1986: 63-152) uses the combination of figurativity with structure in
classifying word-like expressions, and pragmatic function together with source type (cf.
2.15) in grouping sentence-like expressions. Makkai (1972: 135-79) uses the same criteria.
For him figurativity is used only to distinguish idioms without literal interpretation from the
rest, and the emphasis is on structure and pragmatics, while for Gläser (1986: 55-56, 71-
101) figurativity is a significant property showing degrees of idiomaticity: partially literal
(white lie) or short fully figurative (white elephant), and long fully figurative (burn the
candle at both ends). Figurativity combined with structure establish groups such as partially
figurative 'A + N' (cold comfort), 'V + PP' (rise to the occasion) etc.; short fully figurative
'N's + N' (a baker's dozen), 'V + one's+ N' (lose one's head) etc.; long fully figurative (every
Tom, Dick, and Harry), 'V + NP + PP' (get a word in edgeways), etc.
It is instructive at this point to show Gläser's (1998: 128) classification of word
combinations, reproducing her chart, which is the English version of the German chart given
in Gläser (1986: 48):

104
As is usual in the literature, word combinations are divided into the categories called
nominations, reductions of propositions and propositions on the basis of syntactic function.
Nominations function as word-like units, and in this sense they are referred to as word
equivalents (Gläser 1986: 45). Propositions are sentence-like units, while reductions of
propositions occupy a transitional area, since units belonging here are related to or can be
traced back to sentence-length expressions, but they function as word-like units (Gläser
1986: 45). Therefore, they are subtypes within nominations. Gläser (1998: 126, 1986: 14)
approaches the word-like/sentence-like distinction from a semantic point of view.
Nominations are described as units that "designate a phenomenon, an object, an action, a
process or state, a property or a relationship in the outside world", while propositions
"designate a whole state of affairs" (Gläser 1998: 126).
In 2.15 we mentioned the derivational relation between word-like and sentence-like
units. As the chart shows, the transitional area contains fragments of proverbs (a new

105
broom), proverbial sayings (put the cart before the horse), allusions and fragments of
quotations (the green-eyed monster), irreversible binomials (touch and go) and similes (as
busy as a bee). Gläser (1986: 46, 66, 73) notes that irreversible binomials vary in
compositionality (fish and chips, bits and pieces) and that similes are to a large extent
compositional. Fragments of proverbs are those that derive from proverbs through
shortening, e.g. too many cooks, the last straw, a new broom (Gläser 1986: 45, 106-107).
Moon (1998a: 115, 131) also gives similar examples (make hay, birds of a feather, stick to
one's last, etc.). Proverbial sayings are expressions that could easily be transformed into a
proverb, or a didactic maxim or aphorism by the addition of a verb or some other words
(additions are shown in brackets in the examples), e.g. the thin end of the wedge (is
dangerous), (Don't) make a mountain out of a molehill (Gläser 1986: 45). Nagy (1996: 26)
probably has in mind proverbial sayings when he claims that "proverbs are often derived
from proverbial phrases or idioms". Among allusions and fragments of quotations we find
biblical allusions (the/a doubting Thomas, hide one's light under a bushel), allusions to
Greek and Roman history and mythology (the/an apple of discord), allusions to Greek and
Roman literature (the golden mean), and allusions to English literature (a Jekyll and Hyde)
(Gläser 1986: 123-25). As in other classifications, the boundaries of certain categories are
not clear-cut. Like proverbial sayings, many allusions can be easily changed into sentence-
length expressions (Gläser 1986: 123, 124). The terms "allusion" and "fragment of
quotation" are used as synonyms, though only "allusion" tends to be used for word
combinations whose source is a mythological or historical event (the/an apple of discord,
cross the Rubicon) (Gläser 1986: 122, 123-24).
Propositions include proverbs (Let sleeping dogs lie), commonplaces (We live and
learn), quotations and winged words (A thing of beauty is a joy for ever, Brevity is the soul
of wit), commandments and maxims (Thou shalt not kill), slogans (Safety first) and routine
formulae (How do you do?) (Gläser 1986: 46-48, 126, Gläser 1998: 127). Commonplaces
are trite formulae and truisms without advisory and didactic function, often serving as
conversation fillers, and the category of winged words contains proverbs that can be traced
back to literary sources. Gläser (1986: 47) regards catchphrases as a subtype of quotations,
since they are sentence-length expressions with identifiable sources. The difference between
our classification given in 2.15 and hers is that we have drawn a sharper distinction between
catchphrases and quotations, and we have included nonmetaphorical proverbs in the

106
proverb category, rather than separating them as an independent group. The latter decision
affects our judgement of the idiomaticity of proverbs (see the paragraph below).
In the chart above the unshaded area shows idiomatic units. Although irreversible
binomials and similes are unshaded (i.e. idiomatic), the former include non-idioms (gin and
tonic, loud and clear) as well, and the latter are only partially idiomatic (Gläser 1986: 46,
66-67, 70). We do not regard all proverbs idiomatic (i.e. noncompositional). The proportion
of idioms to non-idioms is a matter of debate among phraseologists. Statistics suggest that
within word-like units non-idioms are predominant, so that the innermost circle in the chart
above may not reflect the proportion accurately (cf. Howarth 1996: 120-21). It is to this
circle that we will turn now.
Nominations cover idioms and non-idioms, and among the latter we find fixed
phrases of various types; for example, technical terms such as indecent exposure, phrases
functioning as proper names such as the Black Sea, the Red Cross and the Iron Lady, and
the category of collocations, which is explained and exemplified below. Since
compositionality is a matter of degree, it is sometimes difficult to decide whether a partially
compositional expression is an idiom or a non-idiom. The decision depends on where we
place the cut-off point in the degree of discrepancy between the expected meaning, based on
a fully compositional interpretation of the constituents, and the actual meaning of the
expression. It could be argued that the Black Sea, the Red Cross and the Iron Lady can
misinform the listener. Furthermore, the Red Cross is based on metonymy, since the phrase
denoting the symbol of the organization is used as the name of the organization, it is
therefore idiomatic. At the same time, the capital letters in the written form of the
expression suggest that the reader should interpret it as a proper name. In the Iron Lady and
the Black Sea one word is used figuratively, while the other word (the noun) is literal, and
this is a typical feature of restricted collocations.
We have pointed out the difficulty of distinguishing restricted collocations from
idioms. Gläser (1986) tends to link figurativity with idiomaticity more strongly than we do,
others emphasize lexical variation, but most linguists do not set up an in-between category.
In contrast, Benson et al. (1986b: 254) establish a transitional category between idioms and
collocations, claiming that examples such as foot the bill, all dressed up and nowhere to go,
curry favour, be in a tight spot, on the spur of the moment are more lexically restricted than
ordinary collocations, but they "seem to have a meaning close to that suggested by their

107
component parts". In other words, they are claimed to have a higher degree of lexical
restrictedness and a higher degree of compositionality than (prototypical) restricted
collocations. But Benson et al.'s examples seem to be rather different. We hold the view that
be in a tight spot is noncompositional, whereas on the spur of the moment is more
compositional. We claim that there is no need for a transitional category if we recognize the
graded nature of idiomaticity. Establishing a third category creates the problem of
distinguishing idioms and restricted collocations from this transitional category.

2.26 Moon

Moon (1998a) gives a detailed description of the forms, functions and frequencies of
English phraseological units based on a study of an 18 million-word authentic corpus of
contemporary English, which yielded over 6000 units to be investigated. Some restricted
collocations (wide awake, fast asleep) are included, but most restricted collocations are not
investigated, and compounds and phrasal verbs are also excluded (Moon 1998a: 2, 89,
221). Moon (1998a) aims at comprehensiveness and focuses on providing an overview
including others' views, rather than theorizing.
Moon (1998a: 19) distinguishes three major categories of word combinations,
depending on whether they are problematic on the grounds of pragmatics (formulae),
semantics (metaphors) or lexicogrammar (anomalous collocations). In contrast to other
groupings, "formulae" is thus a general term used for items called simple formulae, sayings,
proverbs and similes. Metaphors are further distinguished on the basis of their degree of
opacity, while anomalous collocations are divided into subcategories according to
lexicogrammatical and semantic criteria (Moon 1998a: 21, 22-23).
Moon (1998a) emphasizes the relative variability of idioms, illustrating variation of
lexis and structure, as summarized in 2.10 and 2.22.
In her discussion of the semantics of word combinations, Moon (1998a: 201) notes
that fixed expressions are sometimes asymmetrical in the sense that some parts are more
decodable than others. For example, it is easier to draw an analogy between spilling and
revealing than between beans and secret in spill the beans (cf. spill one's guts, spill it, and
spill out). Some metaphors are exemplified, but not all of them seem to be appropriate
(Moon 1998a: 202-205). For example, DESTRUCTION IS FIRE, DANGER IS FIRE and DANGER

108
IS HEAT are introduced to account for some idioms, but the metaphors illustrated should
perhaps be re-formulated as DIFFICULTY/UNPLEASANTNESS IS HEAT and FAILURE IS FIRE.

The former underlies idioms such as a hot potato, if you can't stand the heat, get out of the
kitchen, in the hot seat, the latter seems to motivate crash and burn, get one's fingers
burned and one sense of go up in smoke. Play with fire and too hot to handle seem to be
based on DANGER IS FIRE/HEAT, but the figurative meanings of like a moth to the flame and
there's no smoke without fire are not related to danger or destruction (although destruction
in the figurative sense is close to failure).
Chapter 8 surveys the discoursal function of phraseological units, while Chapter 9
looks at the interpersonal role of word combinations. There is a strong tendency for verb
phrases to be informational or evaluative, while noun and adjective phrases tend to be
evaluative, and adjuncts are informational (Moon 1998a: 220). Metaphors and formulae (in
Moon's terminology) are typically evaluative (Moon 1998a: 221, 225). Negative evaluation
has been found more common than positive (Moon 1998a: 247). Apart from informational
and evaluative expressions the following types are exemplified: situational (those that are
responses to or occasioned by the extralinguistic situation, e.g. it's a small world),
modalizing (those conveying truth values, probability, advice, warnings, preference, etc.,
e.g. on the cards, I kid you not, at one's peril) and organizational (those organizing text and
signalling discourse structure, e.g. let alone) (Moon 1998a: 225-39). Moon (1998a: 267)
claims that phraseological units enable speakers to express evaluations politely, but they also
increase solidarity between the speaker and the hearer.
Word combinations also provide cohesion in a text. Several units can function as
conjunctions (so that), but cohesion is also ensured lexically through the insertion of
context-related words in the idiom (Moon 1998a: 279-81, 284). Although idioms typically
interrupt the cohesive flow of the text, if interpreted literally, there are cases when the lexis
of the context, including the lexis of other idioms in the preceding or following chunk of
text, influences the choice of a particular idiom or the form of a non-institutionalized
variant, often resulting in humour. Examples given by Moon (1998a: 286-93) include It is
better, he believes, to rock the boat than keep it on an even keel, where the same nautical
image is used, and Abstinence is the thin end of the pledge.

109
2.27 Other properties

Idioms typically describe psychologically and socially salient situations that tend to
recur (Nunberg et al. 1994: 493, Fernando 1996: 35). Nunberg et al. label this feature
"proverbiality" and explain that situations such as becoming restless, talking informally or
divulging a secret are described in terms of scenarios "involving homey, concrete things and
relations - climbing walls, chewing fat, spilling beans". They also point out that idioms imply
evaluation and are not normally used to describe situations that are seen as neutral; for
example, reading a book.
Although idioms have often been associated with informal language, several
lexicographers emphasize that many idioms can also be used in other styles (Seidl and
McMordie 1988: 12, Cowie et al. 1993: xxxix, Moon 1995: vi-vii). Cowie and Mackin
(1993) in their dictionary of phrasal verbs use the labels "formal", "informal", "slang" and
"taboo", while Cowie et al. (1993) in their dictionary of idioms, which excludes phrasal
verbs, use the same labels and add "neutral", but "slang" and "taboo" are not regarded here
as terms of style. Gläser (1986: 33) distinguishes three levels and subdivides them into a
total of eight. Her examples include I beg to differ (formal), between Scylla and Charybdis
(literary, poetic/rhetorical), Thou shalt not kill (archaic), cum grano salis (foreign), build
castles in the air (neutral), as right as rain (colloquial), and off one's rocker (slang) (Gläser
1986: 33-37). The eighth stylistic level is termed vulgar, and it is said to be hardly
distinguishable from taboo, such as shit bricks (Gläser 1986: 31, 37). Nagy (1996: 25)
identifies four stylistic levels of phraseological units: formal, neutral, informal and slang. He
also establishes subgroups: poetic/literary, formal, business and legal at the formal level,
informal, folksy and conversational at the informal level, slang, rude/rough and vulgar/taboo
at the slang level.
Conventional expressions can be used to convey the speaker's attitude towards what
he or she is describing; consequently, they are often evaluative (Moon 1995: vi). Moon
(1998a: 246-47) informs us that about a third of her database expressions are evaluative and
negative attitude is roughly twice as common as positive attitude. Some dictionaries provide
information about what type of attitude is associated with the use of certain idioms (cf. the
list of labels in LDEI: xv, ODPV: 474 and ODEI: xl).

110
2.28 Thematic grouping

Idioms can also be arranged into thematic groups. Classification of this type is based
on various semantic criteria. One criterion is the semantic field of one of the lexical words
(the keyword) of the idiom. For example, PDEI (16) lists catch sb red-handed under
COLOURS. Another criterion is the field associated with the source of the expression,
which explains why Smith (1943: 198, 284) puts neck and neck under RACE TRACK and
not BODY PARTS. The idiom has its origin in horse-racing (PDEI: 90). A third criterion is
the semantic field related to the idiomatic meaning of the expression, such as
DEPENDENCE (tied to sb's apron strings), IMPORTANT ((the) top dog), or WEALTH
(have/with money to burn) (Heaton 1987: 1-2). Thematic groupings of the first two types
tend to focus on open-class words; therefore, they omit word combinations consisting of
closed-class items only and those whose open-class items result from conversion, such as
all in all and ins and outs (cf. the nuts and bolts and at all hours, both of which are
included in PDEI (212, 37)). The table below shows a contrastive list of the thematic
categories established by Smith (1943: 189-227, 279-92), and Hadrovics (1995: 159-221,
260-72). Idioms coming from foreign sources and from literature are not shown.
Comparison is difficult, because it is not always clear whether a particular category is major
or minor for Smith. Major categories (unlabelled in the table below) are marked with a
bold-line square around them, minor categories are separated with dashed lines or simply
printed under one another, if several minor categories of one of the authors correspond to
one category of the other author (eg. Smith's DOGS, OTHER TAME ANIMALS and
DOMESTIC POULTRY correspond to Hadrovics's DOMESTIC ANIMALS).

111
Smith Hadrovics Smith Hadrovics
Dogs Domestic animals Hunting Hunting
Other tame animals
Domestic poultry
Horses Horse Fish, fishing Fishing
Harnessed horses, carts Farming Farming, tools
Wild animals Wild animals Gardening, logging
Wild birds
Snails, worms and Insects, reptiles, birds Blacksmith, Handicrafts Occupations
insects
Body parts Body parts Sea Man and forces of nature
Inland, freshwater
Wind, rain and weather
Open-air scenes/objects Minerals Houses, buildings House, furniture
Woods, trees Plants Furniture, household Household
Fruit/vegetable garden Kitchen, cooking
Flower garden Fire
Sewing, mending
Games Card games Speech of soldiers Camping
Card games Personal contest Fight
Other indoor games Firearms Open fight
Dicing Siege
Victory, loot
Discipline, punishment
Popular sports Contest, Sports
Horseracing
Food, eating Food, eating Family, friendship
Drinking Drinking Contact, visit
Birth, death Transport
Motion Business Trading, market
Commerce
Clothing Clothing Music, dancing Music, dancing
Law Law, legal proceedings Painting
Coins and metals Finances, taxation Theatre
Authorities Books, reading, schools
Old religious allusions Beliefs, superstitions Folk customs
God, faith
Devil Devil
Bible Bible

Thematic classification varies from analyst to analyst due to the subjectivity of


establishing groups, selecting the keyword and choosing the criterion. Smith (1943) seems
to consider the source of the idiom more important, though many of his expressions are

112
doubly listed. For example, shoulder to shoulder is given under SPEECH OF SOLDIERS
and BODY PARTS, a red herring is considered both as a FISH AND FISHING idiom and a
HUNTING idiom, and win one's spurs can be found under both PERSONAL CONTEST
and HORSES (Smith 1943: 193, 194, 195, 196, 197, 284). Hadrovics's (1995) grouping
seems to be based on the keyword and the source, but it is not clear which one has priority.
The idiom kiteszik a szûrét 'turn sb out' (literally 'put out sb's szûr [felt cloak of Hungarian
shepherds]') is found under RUHÁZAT, LÁBBELI 'CLOTHING, FOOTWEAR' not under
NÉPSZOKÁSOK 'FOLK CUSTOMS' (Hadrovics 1995: 178, 217-18). But elõre iszik a
medve bõrére 'count one's chickens before they are hatched' (literally 'drink to the bear's
hide in advance') is given under the label IVÁS 'DRINKING' but not under VADÁSZAT
'HUNTING' (Hadrovics 1995: 176, 194). Double classification is sometimes found, as with
nem árul zsákbamacskát 'speak sincerely, reveal one's plans' (literally 'not sell a cat in a
sack'), which is found under HÁZIÁLLATOK 'DOMESTIC ANIMALS' and
ADÁSVÉTEL, VÁSÁR 'TRADING, MARKET', and ujjat húz valakivel 'pick a quarrel
with sb' (literally ‘draw finger with sb'), which is listed both under TESTRÉSZEK 'BODY
PARTS' and VERSENYZÉS 'CONTEST' (Hadrovics 1995: 170, 187, 204, 214). In
thematic classifications the keyword tends to be the noun, but other word classes may also
be selected. PDEI (73, 19) shows draw a red herring across the path under FISH, but
white elephant is put in the category COLOURS.

2.29 Summary of the criteria

Below is a summary of the criteria that we have discussed together with the levels
we have set up:
Compositionality: 1) fully compositional (gin and tonic), 2) partially compositional
(promise the moon/earth), 3) noncompositional (roll up one's sleeves);
Institutionalization: all phraseological units are institutionalized, no levels can be set up to
distinguish types of expressions;

113
Analyzability: 1) fully analyzable (pull strings), 2) partially analyzable (laugh one's head
off), 3) unanalyzable (shoot the breeze);
Extragrammaticality: 1) grammatical, 2) slightly extragrammatical (on the make), 3)
heavily extragrammatical (give sb what for);
Lexical uniqueness: 1) lexically ordinary, 2) slightly idiosyncratic (at gunpoint), 3) heavily
idiosyncratic (kith and kin);
Grammatical variation: 1) highly variable (free combinations), 2) variable ((put the cart
before the horse, with the cart before the horse, cart-before-the-horse), 3) restricted, 4)
frozen;
Schematicity: this property cannot distinguish types of phraseological units but the
following degrees can be set up: 1) highly schematic (Him be a doctor?), 2) partially
schematic ('QUANTIFIER + NOUN + shy/short + of + NP'), 3) specified (catch NP red-
handed);
Figurativity: 1) literal (Better late than never), 2) partially literal (explode a myth), 3)
figurative (i.e. nonliteral) with a potential literal interpretation (grasp the nettle), 4)
figurative with no literal interpretation (shoot the breeze);
Structure: no levels can be set up;
Type of source: no levels can be set up;
Pragmatic function: no levels can be set up;
Motivation: 1) highly motivated (scratch one's head), 2) partially motivated (as sober as a
judge, grasp the nettle), 3) unmotivated (the bee's knees);
Repetition of sounds and syllables: degrees of repetition can be set up, but they do not
mark degrees of idiomaticity and they cannot distinguish conventional expressions from free
combinations;
Lexical variation: 1) highly variable units, including highly schematic idioms (take a rest/a
holiday/a walk, etc.), 2) variable (in a fix/hole/mess/paddy/spot), 3) restricted variation
(play/keep one's cards close to one's chest), 4) completely fixed (lock, stock and barrel).
Of these properties the following can be used to determine degrees of idiomaticity:
compositionality, analyzability, extragrammaticality, lexical uniqueness, grammatical
variation, figurativity, motivation, and lexical variation.

114
3. IDIOMS, METONYMIES AND METAPHORS

3.1 Psycholinguistic evidence for analyzability and motivation

In 1.3 we summarized the differences between the formalist and cognitive


approaches to idioms. From a cognitive perspective idioms are gestalt structures with
various degrees of discrepancy, analyzability and motivation. They cannot be sharply
distinguished from non-idioms. Idioms can range from fully analyzable to unanalyzable
expressions and from highly motivated to unmotivated units. They are noncompositional or
partially compositional.
In the previous chapter we discussed the relationship between various properties. In
this chapter we focus on two of these properties, analyzability and motivation, which play a
significant role in the cognitive view. Some psycholinguistic evidence was mentioned in
favour of analyzability (cf. 2.6 and 2.24.3). In this section we give more evidence
suggesting that several idioms are analyzable and motivated.
Analyzability affects not only the lexicogrammatical variability but also the
comprehension of idioms (Gibbs 1990: 426-29, Gibbs 1994: 284-88). It seems that the high
degree of analyzability of some idioms more directly facilitates recognition that the idioms
are to be interpreted nonliterally. In a series of psycholinguistic experiments it took people
less time to process normally and abnormally analyzable idioms than their literal control
phrases, and unanalyzable idioms took longer time to process than their control phrases
(Gibbs 1990: 427, Gibbs 1994: 285). Gibbs (1990: 428, 1994: 285-86) rejects the
hypothesis that people combine the context-free literal meanings of the constituent words in
an idiom during comprehension, pointing out that many idioms are extragrammatical or
violate selectional restrictions in their literal interpretation (in the know, promise the moon).
The claim is true, but it loses some of its strength, if we realize that several examples he
gives can be considered restricted collocations (swallow one's pride) or idioms that are very
similar to restricted collocations, and in these examples one of the constituents does have a
literal interpretation. In crack a joke, swallow one's pride and promise the moon either the
verb or the noun is used in its literal sense (joke, pride, promise). Although the literal

115
interpretation of all the constituents is not combined, some words keep their literal sense. It
is unclear what role the literal meanings of the words play in idiom comprehension, or at
what point people recognize the figurativeness of the expression, but there is evidence that
some analysis of the individual parts of the idiom takes place during understanding (Gibbs
1994: 286-88). Furthermore, Gibbs (1990: 429, 1994: 289) reports experiments which
suggest that children also attempt to use compositional analysis while comprehending
idioms.
We have already seen examples of idioms that were motivated by conceptual
metaphors. However, conceptual metaphor is only one of the conceptual mechanisms
motivating idioms. Metaphors and idioms can be regarded as two overlapping sets, with
many idioms being metaphorical and others that are not. Among the latter we find examples
such as bricks and mortar 'houses and other buildings, especially when they are considered
as an investment or safeguard for the future' (CCED: 199), which is based on metonymy.
The materials used in building a house stand for the whole house. The idiom new blood
'new people in an organization who will provide new ideas and energy' (CIDI: 269) is also
motivated by the PART FOR WHOLE metonymy. Idioms denoting various emotions and
evoking the image of bodily effects of these emotions or conventionalized human gestures
associated with the emotions are also metonymic. Surprise can be expressed as sb's jaw
drops (cf. Hungarian leesik az álla 'be surprised', literally 'sb's chin drops'), where the
emotion is described with one of the gestures commonly associated with it.
Apart from metaphor and metonymy conventional knowledge can also contribute to
the motivation of idioms. Such knowledge is shared by speakers about various aspects of
life. For Kövecses and Szabó (1996: 338) conventional knowledge is equivalent to the
background knowledge structure that is used in understanding concepts. In fact,
conventional knowledge can be thought of as encyclopedic knowledge. Kövecses and
Szabó (1996: 338-39) give examples of idioms which contain the word hand and are
motivated by encyclopedic knowledge alone. We have shared information about the shape,
size, use and function of the human hand, and we can use it to make sense of idioms. It is
clear why have one's hands full means 'be busy'. We know that if our hands are full, we
cannot use them to do anything but hold whatever we have in our hands.
Motivation typically comes from a combination of two or more conceptual
mechanisms (Kövecses and Szabó 1996: 340). Indeed, one could argue that the metonymic

116
examples given above are also motivated by our knowledge of what bricks are used for or
what physiological effects or gestures accompany a particular emotion. Metonymy and
conventional knowledge jointly motivate the examples. In other cases conventional
knowledge and metaphor seem to provide motivation. Underlying the idiom run before one
can walk 'to do things that are difficult, without learning the basic skills first' (OALDb:
1453) is our knowledge that running is a more difficult activity than walking, and you
normally have to learn to walk before you learn to run. Thus, motivation comes partly from
conventional knowledge, partly from the conceptual metaphor ACTION IS MOTION. This
metaphor underlies many other idioms such as where does sb go from here 'used to ask what
action sb should take, especially to improve the difficult situation that they are in' (OALDb:
549), follow in sb's footsteps 'do the same job, have the same style of life, etc. as sb else,
especially sb in your family' (OALDb: 497), a mountain to climb 'something that is very
difficult to do’ (CIDI: 261), go the extra mile 'make a special effort to achieve sth, help sb,
etc.' (OALDb: 807) (cf. also the Hungarian akadályt gördít vkinek az útjába literally 'roll an
obstacle in sb's way').
Metaphor and metonymy can also combine in one and the same expression. The
idiom rack one's brains contains a metonymy, since there is a metonymic relation between
the body part (brain) and thinking, but the expression as a whole is metaphorical. The
metaphor is based on an underlying metonymy. Catch sb's eye is also motivated by a
combination of metaphor and metonymy, but it is different from rack one's brains in that the
word eye can be replaced by the word attention in the given idiom, whereas the word
brains cannot be replaced by the word thinking or thought(s). Catch sb's eye is a metonymy
based on an underlying metaphor.
On the traditional view, idioms are dead metaphors, i.e. they once had metaphorical
origins, but they have lost their metaphoricity over time (Gibbs 1994: 273, Lakoff and
Johnson 1999: 119). For example, spill the beans and kick the bucket have arbitrary
idiomatic meanings, though they may have been figurative in the past (cf. Hungarian
kiteszik a szûrét 'turn sb out' (literally 'put out sb's szûr [felt cloak of Hungarian
shepherds]')). In such cases conventions of usage determine the meaning of an idiom, as
well as the appropriateness of idioms in different situations (Gibbs 1994: 274). Cognitive
grammar does not deny that idioms that can be regarded as dead metaphors exist, but it
holds that many seemingly "dead" metaphors found in idiomatic expressions are part and

117
parcel of everyday ordinary language. Many idioms are believed to be motivated by
conceptual metaphors. These metaphors are called conceptual, because they are deeply
entrenched in our thinking, they help us reason about and understand one concept in terms
of another and they can be found outside spoken/written language.
We have already mentioned some examples of conceptual metaphors. Here is a list
of those metaphors with examples, most of which are idioms:

ACCEPTING IS SWALLOWING swallow a bitter pill


ACTION IS MOTION run before one can walk, where does sb go from
here, follow in sb's footsteps, a mountain to climb,
go the extra mile, akadályt gördít vkinek az útjába
ANGER IS AN OPPONENT come to grips with one's anger
CONTROL IS UP gain the upper hand, hold sb down, under sb's heel
DIFFICULTY/UNPLEASANTNESS IS a hot potato, if you can't stand the heat, get out of
HEAT the kitchen, in the hot seat
ENTHUSIASM IS FIRE fire in the belly
FAILURE IS FIRE crash and burn, get one's fingers burned, go up in
smoke
HAPPY IS UP over the moon
SUCCESS IS UP on the up
UNDERSTANDING IS GRASPING get hold of the wrong end of the stick, come/get to
grips with sth, have a grasp of sth, megfoghatatlan
'inconceivable'

Many other examples of metaphorically motivated idioms can be found in the


literature. Spill the beans is partially motivated by what is called the CONDUIT METAPHOR,

which entails that THE MIND IS A CONTAINER and IDEAS ARE ENTITIES (Lakoff 1987: 449-
50). The beans correspond to the information and the container is the head. Spilling is an act
of letting the information out. Evidence that these are not isolated metaphors motivating
only a single idiom comes from the large number of idiom and non-idiom examples that are
based on these metaphors, such as I gave you that idea (the example is taken from Lakoff
and Johnson (1980: 11)). The CONDUIT METAPHOR is taken to imply that ideas, thoughts
and meanings are objects that can be put into linguistic signals, which function as containers
and are sent along a conduit to the listener, whose task is to find the ideas and take them
into their heads (Marsha got those concepts from Rudolf, It is very difficult to put this

118
concept into words, Everybody must get the concepts in this article into his head by
tomorrow or else) (Reddy 1993).
It is important to realize that the examples mentioned above are felt to be perfectly
natural ways of talking about the given aspects of life, they are used automatically and
unconsciously. Lakoff and Turner (1989: 129) point out that it is this automatic,
unconscious nature that leads people to believe that they are dead metaphors. Live
metaphors are mistakenly associated with conscious use that requires extra effort to
understand.
Psycholinguistic research has found some evidence that many idioms are not dead
metaphors. The images they evoke can be traced back to various conceptual metaphors,
which constrain these images. As a result, we can find remarkable similarities between
individuals' mental pictures of a given idiomatic expression or of several idioms with similar
figurative meanings. It is not claimed that all aspects of the image that people have are the
same. For example, spill the beans evokes images that vary as to where the beans are before
they are spilled, or whether they are cooked or uncooked (Lakoff 1987: 449). However, it
is interesting that for most people the container for beans is about the size of the human
head, the spilling is accidental, the beans are not in a neat pile after they are spilled, and they
are difficult to retrieve (Lakoff 1987: 449).
In one experiment Gibbs and O'Brien (1990: 39, 41) asked subjects to define idioms
of anger (hit the ceiling), control (lay down the law), secretiveness (button your lips),
insanity (lose your marbles) and revelation (let the cat out of the bag) and verbally describe
their image in as many details as possible. There was a high degree of consistency in the
images for idioms with similar figurative meanings (Gibbs and O'Brien 1990: 43).
Similarities were also detected at fairly specific levels. For example, in the case of flip your
lid and hit the ceiling subjects imagined "some force causing a container to release pressure
in a violent manner", although "lids can be flipped and ceilings can be hit in a wide variety of
ways, caused by many different circumstances" (Gibbs and O'Brien 1990: 43). Traditional
views cannot account for why people are so uniform. Participants in the same experiment
were also asked specific questions about causation, intentionality, manner, consequences
and reversibility after the first stage of image description (Gibbs and O'Brien 1990: 39-40).
Their responses were very similar (Gibbs and O'Brien 1990: 46). Gibbs and O'Brien (1990:
46-47) argue that people's consistency in their intuitions about the causes, manner, etc. of

119
events is due to the constraining influence of underlying conceptual metaphors, which help
us understand experiences, such as getting angry, revealing a secret, etc. For example, we
conceptualize anger in terms of heat, more specifically heat of a fluid in a container, and we
know that after a certain point the fluid or heat escapes violently from the container. This is
why subjects' images for blow your stack and flip your lid consisted of a person's head
blowing up with steam coming out. Two other experiments were conducted to test the
hypotheses that the uniformity in people's responses was simply due to the figurative
meaning of the idiom and that people formed images solely on the basis of their knowledge
of basic-level prototypes for the objects and actions depicted in idioms (Gibbs and O'Brien
1990: 51, 57). Both hypotheses were rejected (Gibbs and O'Brien 1990: 62).
Emotion concepts, which are often conceptualized in terms of more concrete human
experience, are dynamic, rather than static, i.e. they have a temporal structure. Lakoff
(1987: 397-98) gives a brief description of the prototypical model of anger, which includes
several stages of getting angry. Gibbs (1990: 437-41, 1994: 295-300) reports experiments
which were conducted to test whether people are sensitive to the temporal structure of
concepts and whether they realize that different idioms can denote different temporal stages.
The results show that speakers' sensitivity to the similarity in the meanings of idioms is
based partly on which temporal stage the idioms refer to and that people use their
understanding of the temporal sequencing in their appropriateness judgements. For
example, it was shown that the idiomatic expression got hot under the collar fits better the
sentence When Billy told his father that he had to stay late for detention, his father... than
the sentence When Billy told his father he had totaled his new Porshe, his father... because
of people's expectation that the story character should experience a particular degree of
anger. Traditional views of idioms can account for these facts by stipulating arbitrarily that
different anger idioms should have meanings such as 'get angry', 'get very angry', 'get
extremely angry', etc. What traditional views cannot explain is why two idioms which have
the same meaning, i.e. which refer to the same degree of anger, receive different
appropriateness judgements in one and the same context! The cognitive view argues that
this difference is due to the different conceptual metaphors that underlie the idioms. Lakoff
(1987: 383, 393), based on research done by Kövecses and himself, shows that expressions
of anger are motivated by several metaphors, two of which are ANGER IS THE HEAT OF A

FLUID IN A CONTAINER and ANGRY BEHAVIOUR IS AGGRESSIVE ANIMAL BEHAVIOUR (Lakoff

120
1987: 383, 393). The first metaphor can be seen in blow one's top, the second in bite sb's
head off. Gibbs (1990: 441-43, 1994: 300-302) has found that people prefer to use idioms
whose underlying conceptual metaphor is coherent with the metaphorical information given
in a particular context. For example, readers judge blow one's top more appropriate than
bite sb's head off in contexts which prime the fluid metaphor, rather than the animal
metaphor. Such a context can be seen below (Gibbs 1990: 442, 1994: 301):

Mary was very tense about this evening's dinner party. The fact that Bob had not
come home to help was making her fume. She was getting hotter and hotter with
every passing minute. Dinner would not be ready before the guests arrived. As it got
closer to five o'clock the pressure was really building up. Mary's tolerance was
reaching its limits. When Bob strolled in at ten minutes to five whistling and smiling,
Mary...

The use of very tense, making her fume, getting hotter, the pressure was really building up
and reaching its limits reinforces the image of a hot fluid in a container.
Since conceptual metaphors seem to motivate idioms, but not their literal
paraphrases, despite traditional assumptions idioms are often not identical in meaning to
their paraphrases. Conceptual metaphors yield certain entailments of the cause,
intentionality, manner, etc. of events, which are absent from their literal equivalents. This
nonequivalence hypothesis has also been tested (Gibbs 1994: 303-306). Participants were
given written stories about different events, such as getting angry, revealing a secret, etc.
Some stories contained violations of certain entailments. For example, spill the beans
entails an unintentional act, but the story below suggests intentionality (Gibbs 1994: 304):

John heard some interesting gossip about Paul and Mary. Even though Paul and
Mary were married to other people, they had recently started having a passionate
affair. John was very surprised when he found out about the affair. So John called up
another friend who knew Paul and Mary and quickly blurted out what he knew. The
friend commented to John that he had really

spilled the beans

121
(or)
revealed the secret35

In no-violation story contexts people gave the same appropriateness rating to idioms and
their literal paraphrases, but in violation contexts idioms were judged to be less appropriate
(Gibbs 1994: 305). Though there is evidence in favour of conceptual metaphors, Gibbs
(1994: 306) warns us that it has not yet been established that people actually access these
conceptual metaphors during on-line processing.
Discovering the conceptual metaphors in a language will enable us to see the
motivation behind several idioms. Consider the following expressions: have sth on the
brain, on sb's mind, a weight/load off sb's mind, sb's mind is on sth. All these idioms have
meanings related to thinking, which is not surprising, since the brain and the mind are the
locations for mental processes. However, while sb's mind is on sth can be paraphrased as
'think about sth' (cf. CCED: 1054, CIDI: 255), the others seem to be close in meaning to
'worry about sth, think about a problem' (cf. ODPV: 176, CCDI: 262, CIDI: 48, 255, 254
LID: 40, 230,379, OALDb: 139, 809). The explanation can probably be found in the
conceptual metaphor PROBLEMS ARE BURDENS, which explains why if something is on our
mind, we keep worrying about it, but if our mind is on something, we simply think about it.
Problems and difficulties are conceptualized as concrete burdens that weigh heavily on our
mind. The mind/brain is located under the burden. The explanation sounds simple, but we
must warn the reader against jumping to conclusions. Counterexamples such as take
one's/sb's mind off sth do occur. The idiom does not simply mean 'stop thinking about sth',
but it refers to not worrying about a problem (cf. ODPV: 374, CIDI: 256, OALDb: 809).
Given the metaphor PROBLEMS ARE BURDENS, it would be more "logical" to say take sth off
one's/sb's mind.
Since many idioms are claimed to be motivated by metaphors and metonymies, we
consider these conceptual mechanisms in detail below.

35
Gibbs (1994: 304) presents this text as a no-violation context, and another one in which "John
accidentally said something about what he knew" as a violation context, which is clearly not the case.
However, we assume that the two texts were simply exchanged accidentally in printing. In Gibbs (1994:
304) they are separated from the main body of the text both in terms of lines and indentation, just as we
presented our text.

122
3.2 Metonymy

Traditional definitions of metonymy make reference to words and a relationship


described as contiguity, which implies spatial, temporal or causal association (Geeraerts
1994: 2477, McArthur 1992: 656, Croft 1993: 347). One word is said to stand for another,
as in The buses are on strike, where the object stands for the user, or in The White House
isn't saying anything, where the place is used for the institution associated with that place,
or in He filled up the car, where the whole (i.e. the car) is used instead of the part (i.e. the
tank). The latter type of metonymy, as well as the one in which the part stands for the whole
(He now works as a farm hand), is called synecdoche (McArthur 1992: 1014). The entity
to which the actual word or expression refers is called the vehicle (hand in the last
example), and the entity that we have in mind is the target (labourer in the last example).
Similarly to metaphor, in traditional approaches metonymy is viewed as a matter of
language, and this is reflected in definitions, such as that found in McArthur (1992: 656): "a
figure of speech which designates something by the name of something associated with it".
Cognitive grammar treats metonymy as a conceptual phenomenon, rather than linguistic.
Evidence for this is cited by Lakoff and Johnson (1980: 37), Lakoff (1987: 79-90) and
Gibbs (1994: 324-33). Support in favour of the conceptual nature of metonymy comes
partly from the systematicity of metonymic expressions, as shown by examples such as He
bought a Ford, He's got a Picasso in his den, I hate to read Heidegger. As Lakoff and
Johnson (1980: 39) point out, they are all based on the PRODUCER FOR PRODUCT

metonymy. Gibbs (1994: 325) explains that ?Mary was tasty cannot be used, if we want to
refer to the product (e.g. a cheesecake) made by Mary, but this is due to the general
principle that the metonymic use of an expression must be sanctioned by a body of beliefs
encapsulated in an appropriate frame. Thus, there is a belief in our culture that value
attached to a work of art comes from the genius of the artist who created it. But no such
relationship holds between a cake and the person who makes the cake.
Metonymic reasoning plays an important role in how we view categories. Lakoff
(1987: 79-90) shows that a particular member of a category can be considered to be more
representative of the whole category than other members; consequently, this representative
member stands for the whole category in our thinking. For example, housewife mothers are
taken as better examples of mothers than nonhousewife mothers. Another member of the

123
same category, the working mother, is defined in contrast to the stereotypical housewife
mother, since not all mothers who work are regarded as working mothers, but only those
who stay at home with their children and nurture them (Lakoff 1987: 80-81). We often
reason and make inferences on the basis of stereotypical, prototypical, ideal or salient
examples of categories. In one experiment, the probability of a California earthquake (a
salient example of natural disasters) causing a flood was rated higher than the probability of
a massive flood in North America (Lakoff 1987: 90).
The metonymic nature of thinking can also be detected in areas such as language
comprehension, the use of gestures and the visual arts (Gibbs 1994: 327-33). We easily
understand who the pronoun they refers to in I need to call the garage (where my car was
being serviced). They said they'd have it ready by five o'clock. The reason is that the
singular garage metonymically stands for the people who work there. Speakers and
listeners use metonymic reasoning when they produce or understand indirect speech acts
(Gibbs 1994: 351-57). In making an indirect request, people prefer to highlight the potential
obstacles that may prevent the listener from complying with the request, so that Do you
know what time you close? is inappropriate as a request to a shop owner, because the likely
obstacle in this situation is the owner's unwillingness to provide the information. By asking
about salient obstacles, speakers assume that listeners can infer the whole sequence of
actions that must be done. Although Lakoff and Turner (1989: 103) claim that metonymy is
used primarily for reference, Lakoff and Johnson (1980: 36) note that it is not merely a
referential device, and the above discussion supports this. Having exemplified the
conceptual nature of metonymy, below we will discuss its other properties and its types.
Radden and Kövecses (1999: 18-19) argue that metonymy should not be conceived
of as a simple substitution, since metonymy interrelates entities, rather than replacing one
entity with another. They borrow an example from Warren, saying that we do not refer to
music in I like Mozart, but to music composed by Mozart. Two remarks will be made in
connection with this. First, to be absolutely precise, it is not entities but words or
expressions denoting various entities that are substituted for other words or expressions.
Entities per se are not used in a sentence, they are referred to. (This is another metonymy
also found in Lakoff and Johnson's (1980: 35) definition: "we are using one entity to refer
to another that is related to it".) Second, we could say that substitution is an appropriate
description of the relationship between the words/expressions involved, since Mozart

124
substitutes for music composed by Mozart, and the fact that it does not substitute for the
word music alone does not change the essential features of substitution. Similarly, the
examples given above suggest that we replace words/expressions with other
words/expressions; for example, we say buses instead of bus drivers, a Picasso instead of a
painting by Picasso. Despite all this we claim that Radden and Kövecses are right, since
examples can be found in which the substituting word/expression sounds much more normal
than the word/expression that could also be used in its place. Langacker (1993: 38, n. 8)
points out the awkwardness of *The dog's teeth bit the cat, as opposed to The dog bit the
cat. 36 It must be added that sentences such as The dog bit the cat are not treated as genuine
examples of metonymy (Radden and Kövecses 1999: 31). They illustrate what Langacker
(1987: 273) refers to as the discrepancy between active zone and profile. An example that is
more metonymic is How much wine do we have left?, in which the upper end of a scale is
used for the whole scale, but it would be odd to say ?What quantity of wine do we have
left? or ?What is the quantity of wine left?, i.e. we do not normally use the whole scale.
Langacker (1993: 30) analyzes metonymy as a reference-point phenomenon,
whereby "the entity that is normally designated by a metonymic expression serves as a
reference point affording mental access to the desired target". In She heard the trombone,
the musical instrument is the vehicle, and this entity serves as a reference point that provides
access to the target, the sound of the trombone. Treating metonymy as a type of reference
point construction places it in the context of a pervasive phenomenon.
The contiguity of the vehicle and the target is viewed by cognitive approaches as
conceptual instead of existing in the extralinguistic reality. Following Lakoff (1987: 78),
Radden and Kövecses (1999: 20) account for the contiguity of the vehicle and the target
within the framework of an idealized cognitive model (ICM), while Croft (1993: 348)
operates with domains and domain matrices, claiming that the vehicle and the target are
included in the same domain matrix. Below we summarize Radden and Kövecses's (1999)
cognitive theory of metonymy.

36
Langacker (1993: 37) accounts for the strangeness of *The dog's teeth bit the cat by claiming that the
animal is perceived as an integral whole and biting is attributed to the animal as a whole. While we agree
that this is the normal case, we believe that the cognitive salience of the animal as a whole can be
overridden in special situations where the dog and its teeth can be separated. Imagine a comic strip or an
animated cartoon in which an unlucky dog happens to have teeth that tend to jump out of its mouth and
snap at other animals and people. In such a case the asterisked sentence becomes fully acceptable.

125
Radden and Kövecses (1999: 23) distinguish three ontological realms, the world of
concepts, forms and things/events, all of which involve ICMs. Metonymy occurs within or
across these realms. The table below provides an overview of various types of metonymy as
they relate the realms together with Radden and Kövecses's examples. The Sign ICM is the
cognitive model of the linguistic sign relating form and concept, while Reference ICMs
relate real world things/events to other ontological realms. In concept metonymies, there is
a shift from one concept to another, either accompanied or not by a shift in form (Radden
and Kövecses 1999: 26).
Sign metonymies:
1.FORM FOR CONCEPT the word form dollar for 'money'
2.CONCEPT FOR FORM "tip of the tongue" experience

Reference metonymies:
3.FORM-CONCEPT FOR THING/EVENT the word cow for a real cow
4.THING/EVENT FOR FORM-CONCEPT ---
5.CONCEPT FOR THING/EVENT the concept 'cow' for a real cow
6.THING/EVENT FOR CONCEPT Bobby Charlton for 'sportsmanship'
7.FORM FOR THING/EVENT the word form cow for a real cow
8.THING/EVENT FOR FORM ---

Concept metonymies:
9.FORMA-CONCEPTA FOR FORMB-CONCEPTB buses-'buses' for bus drivers-'bus drivers' in
The buses are on strike
10.FORMB-CONCEPTB FOR FORMA-CONCEPTA I-'I' for my car-'my car' in I am parked over
there
11.FORM-CONCEPTA FOR CONCEPTB mother-'mother' for 'housewife mother'
12.CONCEPTB FOR FORM-CONCEPTA ---
13.FORMA-CONCEPTA FOR FORMA-CONCEPTB White House-'place' for White
House-'institution' in The White House did not
intervene
14.FORMA-CONCEPTB FOR FORMA-CONCEPTA Central Station-'institution' for Central
Station-'place' in Let's have an oyster dish at
Central Station
15.FORMA-CONCEPTA FOR FORMB-CONCEPTA UN-'United Nations' for United
Nations-'United Nations'
16.FORMB-CONCEPTA FOR FORMA-CONCEPTA [possible in translation equivalence]

126
As can be seen, the possibility of the reversal of each metonymic relationship is also
examined. Reversibility has two slightly different interpretations, depending on whether the
vehicle and the target belong to different ontological realms or not. If they do, reversal
simply means that the same realm that can occur as the vehicle can also occur as the target
and vice versa. (We are simplifying at this point, since it is not the realm that functions as a
vehicle or target, but entities belonging to a particular realm or realm combination.) This
type of reversibility is exemplified in sign and reference metonymies above (1-8), where we
find four pairs of (potential) metonymic relations, and the relations in each pair (e.g. 7 and
8) differ only in terms of their directionality. If the vehicle and the target come from the
same realm (see 9, 10, 13-16 above), the metonymy carries some conceptual content in
addition to interrelating the two realms. Since the same type of realm, or - to be precise -
the same realm combination, can be found at both ends of the metonymic relation, the
metonymy is reversible. For example, the FORM-CONCEPT pairing can occur both as the
vehicle and the target. But this type of reversibility is inherent in the metonymy.
The second type of reversibility pertains to the conceptual content. What we mean
by this term is whether the metonymy is an example of CONTROLLED FOR CONTROLLER,

PART FOR WHOLE, PLACE FOR INSTITUTION, etc. In other words, The buses are on strike,
which is of the CONTROLLED FOR CONTROLLER type, has its reverse counterpart in a
CONTROLLER FOR CONTROLLED metonymy, and one example of this is I am parked over
there. If the vehicle and the target belong to overlapping realms (see 11 and 12 above), both
types of reversibility can be examined. The two types of reversibility we have described are
also noted by Radden and Kövecses (1999: 27), since in their discussion of the
mother-'mother' example they distinguish the metonymic relationship from the specific
metonymic relationship. CONTROLLED FOR CONTROLLER, PART FOR WHOLE and PLACE FOR

INSTITUTION would be examples of specific metonymic relationships, while FORM FOR

CONCEPT is simply a metonymic relationship. But in their discussion of the other examples
they use only "the metonymic relationship".
Radden and Kövecses (1999: 23) use the semiotic triangle as their starting point to
set up the relevant ontological realms. Most of the examples to illustrate the ontological
realm of forms are word forms, but there is also an example of a graphic symbol, the dollar
sign ($) (Radden and Kövecses 1999: 24). Concepts can be accessed via strings of sounds,
strings of letters or, in some cases, various pictograms or symbols. If we have more than

127
one form with more or less the same meaning, the question arises whether it is possible for
one of these forms to stand metonymically for another form? In other words, in examples
such as I  words, does the heart-shaped symbol stand for the word form like? Before
considering the possibility of treating this example as a sign metonymy, we will assume that
it is a concept metonymy. In concept metonymies it is the whole form-concept unit (-'like')
that stands for another form-concept unit (like-'like'), forms do not normally occur without
concepts. Therefore, this will be an example of the type of metonymy given in 15 above:
FORMA-CONCEPTA FOR FORMB- CONCEPTA.

All the examples of variant forms (UN for United Nations, exam for examination,
What the heck are you doing? for What the hell are you doing?) are assigned to the above
type of concept metonymy by Radden and Kövecses (1999: 28). They also make a
distinction between metonymies in which the whole ICM stands for its parts or vice versa
(Whole ICM and its part(s)) and metonymies in which a part of an ICM stands for another
part (Parts of an ICM) (Radden and Kövecses 1999: 30, 36). Examples such as UN or
exam, where the variant form of the sign is a reduction of the original form are classified as
belonging to the "Whole ICM and its part(s)" type, since part of the form (exam) stands for
the whole form (examination). However, if we treat UN or exam as a concept metonymy,
then it is clear that the ICM will include not only the form, but the whole form-concept unit,
and the same criterion should be used in determining its type as in the case of the other
concept metonymies. Although the form America is a reduction of the form the United
States of America, this is not an example of the PART FOR WHOLE metonymy. It is the
meaning that matters, not the form. Extensionally, America is the whole, and the United
States is one of its parts.
Reduced and modified forms seem to be examples of the "Parts of an ICM"
metonymy. For example, crude for crude oil could be seen as a case of using a
property/attribute for an entity having that property/attribute. Since attributes are often
treated as possessions, the specific metonymic relation is POSSESSED FOR POSSESSOR. It is
more difficult to classify the other examples of reduction (UN, exam). If we assume that the
concept remains the same, it is difficult to see any metonymic relation, since metonymy
always implies conceptual differences. Concept metonymies are described as involving "a
shift from ConceptA to ConceptB" (Radden and Kövecses 1999: 26). Where reduction or
modification is accompanied by a slight change in meaning, the type of metonymic relation

128
should be determined on the basis of this change. Radden and Kövecses (1999: 28, 36, 43)
decide to treat UN for United Nations, exam for examination, crude for crude oil, or tgif
for Thank God, it's Friday as a PART FOR WHOLE metonymy of the "Whole ICM and its
part(s)" type, and euphemisms (heck for hell, effing for fucking) and substitutions by pro-
forms as belonging to the MODIFIED FORM FOR ORIGINAL FORM and SUBSTITUTE FORM FOR

ORIGINAL FORM metonymies respectively, both of the "Parts of an ICM" type. This suggests
that the above examples could be regarded as involving forms only, and as such they are not
concept metonymies but sign metonymies of a special type. Within the Sign ICM the
reduced or modified form is taken to stand for the original form. However, Radden and
Kövecses (1999: 23) do not show that forms can stand for forms. Nevertheless, it is more
appropriate to regard variant forms, including forms such as , as FORM FOR FORM

metonymies, in addition to considering them FORM FOR CONCEPT metonymies.


Examples of stand-for relations can also be found in rhyming slang. Within the Sign
ICM the following word forms stand for the concepts indicated: apples and pears for
'stairs', bowl of water for 'daughter', butcher's (hook) for 'look', trouble and strife for 'wife'
(McArthur 1992: 868). Within the Concept ICM the rhyming slang form-concept units
replace standard language form-concept units, as in Take a butcher's at him 'Take a look at
him'.
Rebuses often mix letters, symbols and pictures. They can be taken to stand for
linguistic expressions, i.e. form-concept units (rebuses) stand for form-concept units (words
or sentences). For example, H& stands for hand. But in order to find the linguistic
expression, we must follow a mental course starting with the forms given in the rebus (&),
going to the relevant concept ('and') and proceeding to reach a form (and), which is the
usual form expressing the same concept. Putting together the final forms (h and and) will
yield the final concept, which the whole rebus is supposed to denote. The forms in the rebus
stand for concepts, which in turn stand for forms. Similarly, in rebuses that include pictures
(see Appendix 4), we need to start with the picture first (a form), then establish mental
contact with the concept it stands for, then we need to reach a written form that normally
stands for the same concept. In our first rebus in Appendix 4, the form we start with is a
picture, the concept it stands for is 'pan', and the written form associated with this concept
is pan. It is only when we have reached this written form that we can perform the required
operation of attaching to it a t.

129
The analysis of some examples as forms standing for other forms raises the question
whether stand-for relations exist within the other two ontological realms of the semiotic
triangle. Prototypical examples of metonymy involve a CONCEPT FOR CONCEPT relation.
Since concepts rarely occur without linguistic forms in communication, forms are typically
paired with concepts.
Examples of substitution can be found in extralinguistic reality. Shop-window
dummies stand for people, scarecrows also replace people, a brick can be used instead of
the missing leg of an object, and a nail can be hammered into the wall not only with a
hammer. When children are playing, they often re-name various objects that replace the
genuine entity; for example, a plastic chip can be regarded as a pill that will help the doll get
better, so that it is referred to as pill or medicine. We do not wish to treat these examples
metonymic.
Radden and Kövecses (1999: 31-43) describe various metonymic relationships,
arranging them into two major groups; "Whole ICM and its part(s)" and "Parts of an ICM".
Apart from PART OF A FORM FOR THE WHOLE FORM (crude 'crude oil') and MODIFIED FORM
FOR ORIGINAL FORM (effing 'fucking'), which were discussed above, several examples are
given. Metonymies of the "Whole ICM and its part(s)" type include PART FOR WHOLE and
WHOLE FOR PART, which we have already illustrated, as well as the following: UPPER END OF
A SCALE FOR WHOLE SCALE (How old are you? 'what is your age?'), OBJECT FOR MATERIAL

CONSTITUTING THE OBJECT (I smell skunk), SUBEVENT FOR WHOLE EVENT (Mary speaks
Spanish 'Mary speaks, understands, reads and writes Spanish'), PRESENT FOR FUTURE (I am
off 'I will be off'), ACTUAL FOR POTENTIAL (He is an angry person 'he can be angry'),
CATEGORY FOR A MEMBER OF THE CATEGORY (the pill 'birth control pill'), SPECIFIC FOR

GENERIC (A spider has eight legs 'spiders have eight legs'), DEFINING PROPERTY FOR

CATEGORY (blacks 'black people'), CATEGORY FOR SALIENT PROPERTY (Boys will be boys
'boys will be unruly'). Most of these metonymies are reversible. "Parts of an ICM" include
MANNER FOR ACTION (to tiptoe into the room), PERCEPTION FOR THING PERCEIVED (sight
'thing seen'), CAUSE FOR EFFECT (healthy complexion 'the good state of health bringing
about the effect of healthy complexion'), PRODUCER FOR PRODUCT (I've got a Ford 'I've got
a car'), CONTROLLED FOR CONTROLLER (The Mercedes has arrived), POSSESSOR FOR

POSSESSED (That's me 'that's my bus'), CONTAINER FOR CONTENTS (The bottle is sour 'the

130
milk is sour'), PLACE FOR INHABITANTS (The whole town showed up 'the people showed up'),
WORDS FOR THE CONCEPTS THEY EXPRESS (a self-contradictory utterance).
The reversibility of metonymy does not mean that both directions are equally
common. For example, metonymies of the type given in 11, 13 and 15 in the table above
exemplify the preferred (or the only possible) direction. The CONCEPT FOR FORM metonymy
is not common. The examples given tentatively by Radden and Kövecses (1999: 24)
illustrate the phenomenon that we have a concept in mind but cannot think of the word
form, and when learners must find a word in a foreign language for a concept with which
they are familiar. However, in these cases the concept cannot evoke the word form, but it
does not actually stand for the word form.
The preferred direction seems to be determined by cognitive and communicative
principles (Radden and Kövecses 1999: 44). In our human experience, concrete objects
have more salience than abstract objects, it is therefore natural that targets tend to be
concrete and vehicles abstract, as in a book written in a careful hand 'a carefully written
book' (Radden and Kövecses 1999: 45). The principle of CONCRETE OVER ABSTRACT also
motivates many examples of conversion, such as blanket the bed, porch the newspaper.
One disadvantage of Radden and Kövecses's classification is that it treats different
subtypes of metonymy alike. Geeraerts (1994: 2478) distinguishes metonymy as a means of
creating a new lexical item (e.g. braille after its inventor Braille, or volt after the scientist
Volta) from other types of metonymy. Several of Radden and Kövecses's (1999: 37)
examples belong to the former type, e.g. to blanket the bed, writer, to porch the newspaper.
These should be distinguished from examples such as I am parked over there, since blanket
as a verb does not refer to the object involved in the action, as Radden and Kövecses (1999:
37) suggest. Metonymy is present only in the process of word-formation, in converting the
noun blanket into the verb blanket, or in deriving writer from write.
Most of Radden and Kövecses's examples are not idioms. It is beyond the scope of
the present study to examine which (sub)type of metonymy can be found in the idioms of
English and what percentage of idioms are metonymic. This would be the task of future
studies. The metonymic examples that we gave in 3.1 exemplify the PART FOR WHOLE

metonymy (bricks and mortar, new blood) and the SUBEVENT FOR WHOLE EVENT metonymy
(sb's jaw drops, leesik az álla 'be surprised'). The latter underlies roll up one's sleeves, in
which a subevent (rolling up one's sleeves) is used for the whole event (preparation). The

131
idiom drop names illustrates the CATEGORY FOR A MEMBER OF THE CATEGORY metonymy,
since the word name (category) stands for 'the names of famous people' (subcategory). The
previous two examples are metaphorically motivated as well.
More complex is the idiom a hard/tough act to follow, which is metaphorical not
metonymic, but its use in a sentence results in metonymy as well. The subject slot is often
filled with a personal name or pronoun, so that examples such as He'll be a tough act to
follow are not uncommon. In this example he stands for his acts/actions.

3.3 Metaphor

A number of scholars have studied metaphors since the time of Aristotle. Ricoeur
(1986: 16-24) draws attention to some characteristics of Aristotle's treatment of metaphor.
For Aristotle, metaphor is something that happens to the noun, it is defined in terms of
movement, and it is the transposition of a name. Accordingly, the first part of the definition
is "metaphor 'is the application to something of a name belonging to something else' "
(Steinhart and Kittay 1994: 2452-53). Another feature of Aristotle's definition is that the
rest of it contains a typology of metaphor (Ricoeur 1986: 20). The transposition can be
"either from the genus to the species, or from the species to the genus, or from a species to
another species, or according to analogy" (Steinhart and Kittay 1994: 2453). An example of
the species-to-genus transfer is Indeed ten thousand noble things Odysseus did, where ten
thousand is used instead of many, and an example of metaphor by analogy is the evening of
life used in the sense 'old age', the analogy being Old age is to life as evening is to day
(Gibbs 1994: 210).
Metaphor is a matter of words, since it is words that are transposed (Gibbs 1994:
210). Ricoeur (1986: 20) claims that Aristotle's notion of metaphor includes the idea of
deviation from ordinary usage, and the idea that a word is borrowed from its original
domain and is used as a substitute for another word, which is absent but available. The
ornamental function of metaphor can be traced back to these ideas, since the substitute
conveys no new information, because the absent word could be brought back in. Aristotle is
credited with introducing the Elliptical Simile theory of metaphor, which later gave rise to
the comparison theory (Steinhart and Kittay 1994: 2453). On his view, a metaphor is a
condensed simile, from which the word like is dropped.

132
Lakoff and Johnson (1999: 376) find that Aristotle's reasoning is based partly on the
conceptual metaphors IDEAS ARE ESSENCES and ESSENCES ARE FORMS, and his theory of
metaphor is a consequence of these conceptual metaphors, as well as certain folk theories
that he held. For Aristotle, ideas are aspects of the physical world, and it is impossible for an
idea to be conceptualized in terms of another. It follows that metaphor is linguistic, not
conceptual, and that it is a deviant use of a word. It also follows that the only meaning a
metaphorical word can have is some other literal meaning.
Since Aristotle, scholars have developed various theories of metaphor, such as the
substitution view, the comparison view and the interaction view. As its name implies, the
substitution theory holds that metaphors are decorative replacements of literal words; for
example, Man is a wolf is an indirect way of saying Man is fierce (Gibbs 1994: 212,
Steinhart and Kittay 1994: 2453). This view relegates metaphors to a secondary category,
they are stylistic devices that are not essential. Lakoff and Turner (1989: 120-22) point out
in their criticism that using a literal paraphrase cannot convey all the inferences that a
metaphor can. Making a choice can be viewed metaphorically as taking a path, but the
metaphor entails, for example, that just as people on different paths are not together, so
people who make different choices are metaphorically separate.
Both the substitution and the comparison views assume the primacy of literal
meaning. The comparison view claims that a metaphor "A is B" means "A is like B in certain
respects" (Gibbs 1994: 212). A is referred to as the target or topic, B is the vehicle. This
view is taken to imply that the metaphor is based on shared features and that there are
preexisting similarities between the target and the vehicle (Gibbs 1994: 218, Lakoff and
Turner 1989: 198). Lakoff and Turner (1989: 198) claim that metaphorical interpretation is
more complex. In Achilles is a lion, the courage of Achilles belongs to his character, while
it is an instinctual property of the lion, so that character is understood metaphorically in
terms of instinct. Furthermore, the quintessential nature of the lion's courage is mapped
onto Achilles' courage, which becomes steadfast and quintessential. Gibbs (1994: 214-18)
also provides some criticism. Certain metaphors do not appear to be based on similarity, as
in Sally is a block of ice, where there is no similarity even between unemotional behaviour
and coldness interpreted literally (Gibbs 1994: 218). But the comparison view cannot
account for how we understand metaphors even when there is some underlying similarity.
Whether the target and the vehicle are considered to be similar and to what extent depends

133
on the context, and a given metaphor does not necessarily presuppose any association
between the two terms in people's minds before they encounter these terms in the metaphor
itself (Gibbs 1994: 216-17). Lakoff and Johnson (1980: 147-155, 214-15) also argue that
conceptual metaphors are based on similarities as experienced by people, not on objective
features. For example, there is no inherent similarity between time and moving objects
despite the TIME IS A MOVING OBJECT metaphor. This metaphor underlies, for example, the
idioms race against time, keep up/move with the times.
Many linguists attempt to capture metaphors by emphasizing their anomalous
nature, which can be traced back to Aristotle's notion of deviance. This deviance may be
grammatical or semantic, and in the latter case the sentence interpreted literally is assumed
to be false (Gibbs 1994: 222). For example, The stone died violates selectional restrictions,
and the recognition of deviation is the first step in metaphor comprehension (Gibbs 1994:
223). However, several examples can be found that are grammatically deviant but are not
judged to be metaphorical (The book who you read was a best seller) (Gibbs 1994: 223).
Other examples do not seem to be false, yet they are metaphors (No man is an island)
(Gibbs 1994: 224). Views based on deviance suggest that metaphor comprehension takes
place only after speakers recognize that the sentence is ungrammatical or literally false, but
psycholinguistic evidence does not support this (cf. Gibbs 1994: 225).
The Interaction view, as Gibbs (1994: 233) explains, fares better than the
comparison view in that it does not presuppose that the metaphor is based on preexisting
similarities. The target and the vehicle interact in a parallel implicational complex, which is
the result of the listener selecting certain features of the vehicle in the light of the target and
vice versa. Lakoff and Turner (1989: 132) criticize this view by stressing the unidirectional
nature of metaphor, in which the mapping goes from the source domain (journey) to the
target domain (life). Gibbs (1994: 239) mentions psycholinguistic evidence in favour of this
argument. For example, the properties of the vehicle are more effective in the recall of
metaphors than the properties of the target. Gibbs (1994: 237) also points out that
according to the interaction view the predicates that listeners project from the vehicle to the
target are understood literally rather than metaphorically. But it is not likely that marriage
can be described literally as a game (cf. Marriage is a zero-sum game).
Cognitive grammar challenges the traditional view of metaphor, which claims that
everyday conventional language is literal, that all concepts can be understood literally, and

134
that only literal language can be contingently true or false (Lakoff 1993: 247). On the
cognitive view, metaphor is understanding one conceptual domain in terms of another
(Lakoff and Turner 1989: 103). A domain, called target domain, is understood in terms of
another domain, which is called source domain. The mechanism through which this happens
is mapping, i.e. the source domain is mapped onto the target domain. For example, love is
understood and experienced in terms of a journey, and this can be expressed in the form of
the conceptual metaphor LOVE IS A JOURNEY (Lakoff and Johnson 1980: 44-45, Lakoff
1993: 206-207). Examples illustrating this metaphor as well as other metaphors were given
above in 3.1. Anger is understood in terms of a heated fluid, and this is expressed by the
conceptual metaphor ANGER IS THE HEAT OF A FLUID IN A CONTAINER, which underlies
examples such as I had reached the boiling point, He was bursting with anger, She flipped
her lid, etc. (Lakoff 1987: 383-85). Another metaphor mentioned above was THE MIND IS A

CONTAINER, which is exemplified by idioms such as turn sth over in one's mind, keep sth in
mind, not enter sb's head/mind. Numerous examples of metaphors have been discussed in
the literature. Lakoff (1987: 274) notes that social and interpersonal relationships are often
understood in terms of links, as in make connections and break social ties. Since marriage is
a type of human relationship, it is also conceptualized in terms of a physical link, which can
be described as the MARRIAGE IS A TIE metaphor. This metaphor motivates examples such
as get spliced, tie/untie the knot, and get hitched.
The correspondences between source and target domains are conceptual, rather than
linguistic. Lakoff (1993: 208-209) as well as Lakoff and Turner (1989: 50) emphasize the
distinction between the linguistic and the conceptual levels of analysis. Lakoff (1993: 209)
claims that if metaphors were linguistic, we would have different metaphors in sentences
such as We've hit a dead-end street, We can't turn back now, Their marriage is on the
rocks. However, they are linguistic manifestations of one and the same metaphor, and it
makes sense why particular expressions have the given meaning. Metaphorical mappings
influence people's reasoning as well, since various inferences in the target domain can be
drawn on the basis of the source domain. Furthermore, if metaphors were merely linguistic,
we would not be able to understand novel uses and poetic metaphors as easily as we seem
to do. All this evidence in favour of the conceptual nature of metaphors will be exemplified
below.

135
The question arises whether any type of experience can serve as a source domain or
target domain. Since metaphor implies discrepancy as well as similarity, source and target
domains cannot be very similar. We have already seen several metaphors which suggest that
abstract concepts tend to be understood in terms of more concrete experience. As Lakoff
and Johnson (1980: 59) say, "we typically conceptualize the nonphysical in terms of the
physical". Emotion (love, anger, etc.) is especially susceptible to metaphorical
understanding. Many relatively basic concepts seem to be metaphorical, as Lakoff's (1993:
212) list shows: time, quantity, state, change, action, cause, purpose, means, modality and
category. For example, time is understood in terms of motion (Lakoff and Johnson 1980:
41-44, Lakoff 1993: 216-18). Sometimes times are conceptualized as objects which move
towards the observers (future), pass them (present) and then move away from them (past),
as in The time will come when..., The time for action has arrived, The time has long since
gone when.... The idioms race against time and beat the clock also evoke the image of time
as a moving entity, but in this case the present time moves towards the future, and we try to
race against the present (Lakoff and Turner 1989: 45-46). The same image of people and
the present time moving in the same direction (towards the future) can be seen in keep
up/move with the times and be behind the times. Time can also be thought of as a flowing
substance (cf. the flow of time), in which case the amount of the substance corresponds to
the duration of time and the motion of the substance past the observer to the passage of
time (Lakoff and Johnson 1999: 144-45). The idiom a lot of water has gone under the
bridge with its flowing water imagery provides us with an apt metaphor that underlies the
idiomatic meaning 'a lot of time has passed'. Sometimes time does not move. Times are
locations on a path, and the observers are moving along that path: As we go through the
years..., We're getting close to Christmas. The idioms down the road and way back (in) are
also motivated by the same conceptualization.
As we have seen, conceptual metaphors usually motivate not only idiomatic
expressions, but also single words. Further examples are grasp 'understand' and see 'know',
which are motivated by UNDERSTANDING IS GRASPING and KNOWING IS SEEING respectively
(Lakoff and Turner 1989: 129, 130, Gibbs 1994: 276). The former metaphor seems to
provide partial motivation for get hold of the wrong end of the stick, the latter metaphor
underlies idioms such as the scales fall from sb's eyes, keep one's cards close to one's chest,
light dawns on sb, the blind leading the blind or in the dark.

136
We used the terms "concept" and "domain" interchangeably, but they are not
(always) synonymous. Clausner and Croft (1999: 2) describe domains as background
knowledge structures in the context of which concepts are understood. Langacker's (1987:
147) use of the term is similar. Concepts can serve as domains for other concepts, but
domains are typically more complex knowledge structures (Langacker 1987: 147). An
example of a concept functioning as a domain is 'finger', when it is used as the necessary
context for 'knuckle' (Langacker 1987: 147-48). The background context used in
understanding the concept 'bird' includes knowledge about the shape of birds, about their
activities (e.g. flying, eating), the avian lifecycle, etc. (Clausner and Croft 1999: 7). These
aspects are specified relative to a variety of domains, such as space, life, time, etc. In other
words, the concept 'bird', just as most concepts, presupposes a whole collection of domains,
called its domain matrix (Clausner and Croft 1999: 7, Langacker 1987: 147). It seems that
the concept-domain relationship can be regarded as a part-whole relationship, i.e. the
concept is part of its domain or domain matrix (Clausner and Croft 1999: 6).
We agree with Langacker (1987: 63) that linguistic semantics is encyclopedic, and
this raises the following question: if concepts are understood relative to a domain matrix,
and if a domain matrix contains more than one domain, can metaphorical mapping occur
between domains within the same domain matrix? Croft (1993: 348) rejects this idea, since
in the case of She's feeling down, there is no spatial orientation domain within the domain
matrix of the concept 'sad'. Thus, SADNESS IS DOWN involves a mapping between domains
which are parts of different domain matrices. However, Barcelona (1997: 4) claims that the
behavioural effects of sadness are part of the encyclopedic knowledge we have about 'sad',
and one such behavioural characteristic is drooping posture (John drooped his head, She
walked with drooping shoulders/downcast eyes after the news of her child's death), which
presupposes the domain of spatial orientation. Consequently, spatial orientation is a domain
within the domain matrix of the emotion, and therefore SADNESS IS DOWN appears to be a
mapping between domains within the same domain matrix. Barcelona (1997: 4) prefers to
view metaphor as a mapping between domains from different matrices, claiming that in our
unconscious, encyclopedic knowledge the spatial domain may be included in the domain of
sadness, but our (relatively) conscious folk taxonomy does not include it.
The correlation between source and target domains is often grounded in direct
human experience, but this is not a necessary condition for metaphor. Love and journeys are

137
not directly linked in our experience, nor are understanding and grasping. Similarly, Lakoff
and Turner (1989: 84) note that there is no direct experiential motivation that connects the
source (people) and the target (plants) in the PEOPLE ARE PLANTS metaphor (a young
sprout), as opposed to MORE IS UP (Prices went up) or PURPOSES ARE DESTINATIONS (reach
one's goal). Idioms exemplifying the above metaphors include put down roots, rooted to the
spot, a shrinking violet, a broken reed (PEOPLE ARE PLANTS), go through/hit the roof, turn
sth up (MORE IS UP), and go for sth, lose one's way as in I feel that the project has lost its
way (OALDb: 1464) (PURPOSES ARE DESTINATIONS). As Lakoff (1987: 276) explains, it is a
common daily experience that the level of a substance rises, whenever we add more to it.
The association between quantity and verticality seems natural. Similarly, there is a close
relation between our purposes and destinations. As Lakoff (1987: 277) and Johnson (1987:
115) explain, one of our everyday experiences is our purpose of getting to a particular
location. Sometimes our only purpose is to move to another place, but it is more common
that this is only a prerequisite for doing something else; for example, taking a desired object.
All this shows that MORE IS UP and PURPOSES ARE DESTINATIONS are rooted in everyday
human experience.
The PURPOSES ARE DESTINATIONS metaphor is related to other conceptualizations
which are described by Lakoff (1993: 220-22) as parts of the event structure metaphor. This
metaphor conceptualizes actions, changes, states, etc. in terms of space, motion and force.
Among the metaphors that comprise the event structure metaphor Lakoff (1993: 220) lists
STATES ARE LOCATIONS (fly into a rage), CHANGES ARE MOVEMENTS, ACTIONS ARE SELF-

PROPELLED MOVEMENTS, PURPOSES ARE DESTINATIONS, DIFFICULTIES ARE IMPEDIMENTS TO

MOTION. However, as Grady (1997: 101-108) points out, some of these are independently
motivated and not necessarily related to motion along a path. He suggests that examples
such as He's really straightjacketed in that job - his boss gives him no latitude whatsoever
are motivated by the metaphor ACTION IS BODILY MOTION, which is based on recurring
experiences such as lifting our arm to drink or shifting our position in order to feel more
comfortable, i.e. motion of a general sort that does not imply paths and destinations (Grady
1997: 103-104). Anything that hinders our motion, not only motion along a path, is
experienced as a constraint on our action. The idiom elbow room supports Grady's claim
that cramped surroundings are seen as constraining our motion. Rules and regulations
imposed on us restrict our actions, and it is not difficult to see why stretch the rules means

138
'to do something or to allow someone to do something which is not usually allowed' (CIDI:
332). By stretching the rules we can make them less tight, so that there is more freedom to
act.
Grady (1997: 104) also shows that difficulties are not necessarily associated with
whatever blocks our motion towards a particular destination, but burdens in general
(DIFFICULTIES ARE BURDENS). The same metaphor is also posited by Lakoff and Turner
(1989: 25, 149), whose examples include My job is weighing me down, Don't burden me
with your problems and Get off my back. This association is based on the everyday
experience of the discomfort we feel while lifting heavy physical burdens (Grady 1997:
104). There need not be any reference to motion, as in the above sentences or in Grady's
(1997: 104) example The tax burden on people in their bracket has grown considerably,
but the metaphor is compatible with ACTION IS MOTION and underlies sentences such as
Lakoff's (1993: 220) He's carrying quite a load. Mental and emotional difficulties or
problems are seen as burdens in the above examples, as well as in the idioms a weight/load
off sb's mind, carry the weight of the world on one's shoulders, on sb's mind, get sth off
one's chest.
The association between states and locations is well-entrenched, and it may be based
on the link between our subjective state and the place where we are (Grady 1997: 106).
Additionally, this association is reinforced by the metaphor CHANGE IS MOTION, which
motivates examples such as Things have gone from bad to worse lately (Grady 1997: 106).
Change and motion are related, since the motion of objects around us is a type of change in
our environment. All this can explain why permanent states are not understood in terms of
location, as shown by the awkwardness of ?My dog is in a state of mammalhood or ?These
cliffs are in/at a state of verticality (Grady 1997: 107).
Grady (1997: 41) also notes that there is no direct experience that could possibly
link theories with buildings despite the existence of the THEORIES ARE BUILDINGS metaphor
(Your facts are solid, but your argumentation is shaky). He convincingly argues that this is
a compound metaphor arising out of the unification of two more basic metaphors:
ORGANIZATION IS PHYSICAL STRUCTURE (the fabric of society) and VIABILITY IS ERECTNESS
(The speed record for the mile still stands/fell/was toppled) (Grady 1997: 40-52). This
explains why certain salient elements in the domain of buildings (tenants, windows,
functions of buildings, etc.) have no counterparts in the target domain and why not only

139
buildings can serve as the source domain (Grady 1997: 43, 51). In They tore the theory to
shreds the source domain of textiles is mapped onto the target domain of theories. Finally,
the more basic metaphors are motivated by the common experience of interacting with
complex objects and forming mental images of logical and causal relations existing in those
objects (Grady 1997: 52).
Grady (1997: 24-25) uses the term "primary metaphor" for experientially grounded
metaphors. Their source and target domains arise out of direct perceptions of the world and
responses to these perceptions. The source concepts are related to physical perception (e.g.
the detection of movement, the perception of weight, etc.). Grady's (1997: 25) examples of
primary metaphors include the two basic metaphors mentioned above as well as DESIRE IS

HUNGER, grounded in the correlation between the sensation of hunger and the focus on
finding food, DIFFICULTY IS HEAVINESS, which is based on the difficulty we experience when
lifting heavy objects, KNOWING/UNDERSTANDING IS SEEING, based on the fact that most of
the information about our surroundings comes to us through the visual channel. The DESIRE

IS HUNGER metaphor motivates the idioms whet sb's appetite, forbidden fruit, lick one's lips
and examples such as But Jules was not eager for classroom learning, he hungered for
adventure (CCED: 825). One type of desire is lust, and it is not surprising that Lakoff
(1987: 409) lists several examples showing the LUST IS HUNGER metaphor: He is sex-
starved, She's quite a dish, I hunger for your touch, etc.
Desire is also associated with thirst, as the following sentences show: Children show
a real thirst for learning (CCED: 1738), People should understand how thirsty for revenge
they are (CCED: 1738), Her thirst for knowledge will never be quenched (CIDE: 1161), I
don't think Dick will ever manage to slake his lust for power (CIDE: 1348). The use of
quench and unquenchable with the noun desire may be doubly motivated. On the one hand,
quench can be used literally to talk about satisfying one's thirst, and given the metaphorical
relation between thirst and desire, it collocates not only with thirst, as in the last example
above, but also with desire. On the other hand, OALDb (1037) suggests that one meaning
of quench is 'to stop a fire from burning', and it is this sense that is used figuratively in
quench sb's desire. The domain of fire is often mapped onto the domain of intense feelings,
as in the idioms sparks fly, fan the flames, breathe fire, fire in the belly, and strong desire
can be described as burning desire. The experiential basis of DESIRE IS THIRST seems to be

140
just as natural as that of DESIRE IS HUNGER, but we have not found idioms that are
motivated by the former.
Although Grady (1997: 25) pairs the source domain of itch with the target domain
of compulsion to act, the target domain of desire is also linked to the same source domain.
In fact, compulsion can be regarded as a strong desire to do something. The idioms itchy
feet and itchy fingers denote desire 'to travel or move to a different place' (OALDb: 690)
and desire 'to get involved in a particular activity' (CCDI: 143) respectively. Most dictionary
examples of itch that we have examined support the mapping between itch and desire to
act: He's itching to get back to work (OALDb: 690), The general was itching for a fight
(CCED: 896), He had an itch to change things (CIDE: 756) (cf. also the Hungarian
examples viszket a tenyere 'itch to slap sb' (literally 'sb's palm is itching') and feltûnési
viszketegség 'morbid desire to attract attention' (literally 'sensation itchiness')). Other
examples, however, denote a desire for something to happen, for mental or emotional
experience, as in students itching for the lesson to end (OALDa: 634), He was itching to
hear the results (CIDE: 756) and in the idiom the seven-year itch.
Bodily experience can serve as the source domain in many cases, because it is
structured by image schemas, which are described in detail by Johnson (1987). They are
schematic patterns which constantly recur in our bodily movement, in our interaction with
objects and in perception, and they help us make sense of and organize our experiences
(Johnson 1987: 29-30). Since source domains are mapped onto target domains in
metaphors, image schemas also structure target domains. Johnson's (1987: 126) list of
image schemas includes CONTAINER, PATH, CENTRE-PERIPHERY, and several others. One
pervasive organizing principle in our experience is the CONTAINER schema, which defines
boundaries, an interior and an exterior. It is used in structuring experiences such as getting
into a car, going out of a room, pouring water into a glass, etc. The PATH schema imposes a
source (starting point), a goal (end point) and a path on various types of motion. It also
typically implies directionality.
Metaphorical mappings are not unlimited. One major constraint is the Invariance
Hypothesis, which holds that metaphorical mappings preserve the image schematic structure
of the source domain, provided it is consistent with the inherent structure of the target
domain (Lakoff 1993: 215). For example, in the LINEAR SCALES ARE PATHS metaphor,
exemplified by sentences such as John is far more intelligent than Bill, the starting point of

141
the path is mapped onto the bottom of the scale, and the distance travelled is mapped onto
quantity (Lakoff 1993: 214). This means that inherent target domain structure limits the
mapping possibilities. Sources are mapped onto sources, paths onto paths, and so on. We
agree with Barcelona (1997: 11), who claims that not only the inherent structure of the
target domain but conventional knowledge about the target domain has to be preserved as
well. This is implied in Lakoff's (1993: 216) explanation that, although we know that a
recipient possesses the object given after the giving, when actions are conceptualized as
transfers (She gave him a kick), the action does not exist after it occurs.
The source-to-target mapping can be thought of as a set of correspondences.
Ontological correspondences link entities in the two domains, while epistemic
correspondences map knowledge about the source domain onto knowledge about the target
domain. The metaphor ANGER IS THE HEAT OF A FLUID IN A CONTAINER involves the
following ontological correspondences: the container is the body, the heat of fluid is the
anger, pressure in the container is internal pressure in the body, explosion is loss of control,
etc. (Lakoff 1987: 387). Some of the epistemic correspondences are the following: we
know that intense fluid heat leads to container heat, internal pressure and agitation, so the
effect of intense anger is body heat, internal pressure and agitation. An explosion is
damaging to the container and dangerous to bystanders; therefore, a loss of control is
damaging to an angry person and dangerous to other people (Lakoff 1987: 387). Similarly,
we know that darkness deteriorates visibility, and this knowledge as well as the metaphor
KNOWING IS SEEING can be used to interpret idioms such as in the dark, keep sth dark, a
leap in the dark, which all convey lack of knowledge.
Metaphorical mappings are unidirectional. Entities from the source domain are
mapped onto entities in the target domain, but not vice versa. Target domain expressions are
not used to talk about source domains. For example, several types of human experience are
understood in terms of verticality, as shown by the orientational metaphors HAPPY IS UP,

SAD IS DOWN (My spirits rose/sank), MORE IS UP, LESS IS DOWN (My income rose last year,
His income fell last year), GOOD IS UP, BAD IS DOWN (Things are looking up, Things are at
an all-time low), etc. (Lakoff and Johnson 1980: 15-17). But we cannot reverse the
directionality of these mappings and refer to a balloon rising as "getting happier/better".
Metaphors such as PEOPLE ARE MACHINES and MACHINES ARE PEOPLE, which underlie the
respective examples have a screw loose, fire on all cylinders and The computer is

142
punishing me are not genuine counterexamples to our claim. As Lakoff and Turner (1989:
132) explain, these are different metaphors, since in the former the various parts and
characteristics of functioning of the machine are mapped onto people, while in the latter the
will and desire of people are attributed to machines.
Evidence for the conceptual nature of metaphor also comes from the ease with
which speakers interpret novel extensions of conventional metaphors, which are often
encountered in poetry. Lakoff and Turner (1989: 67, 158) claim that poetic language uses
the same conceptual metaphors as ordinary language but extends and combines them, as a
result of which poetic expressions differ from conventional phrases. When Horace refers to
death as the "eternal exile of the raft", he uses the DEATH IS DEPARTURE metaphor, adding
conceptual content that is absent from everyday thought. He describes death as an exile and
specifies the vehicle, which is a raft (Lakoff and Turner 1989: 67-68).
Since metaphor pervades our thinking, we should be able to find evidence for
metaphor outside language. One piece of evidence is described by Johnson (1987: 110-11),
Gibbs (1994: 162) and Grady (1997: 14), who report experiments designed to test
analogical reasoning. Subjects with little understanding of electricity were taught to think of
electric circuits in terms of either flowing water or a moving crowd. Those who used the
flowing water analogy had a better understanding of certain aspects of the system, while
those with the moving crowd model understood other aspects better. This finding shows
how reasoning can be influenced by metaphors.
It has also been found that gestures can also be based on metaphors (Gibbs 1994:
164-65, Grady 1997: 15). The gesture for direct limits consists of moving the hand along a
straight line in front of the body, with the movement ending in a tensed stop. This gesture
was used by a mathematician in a conversation, even though he made a speech error by
saying "inverse limits", when he meant "direct limits". The gesture was obviously motivated
by what he had in mind, not by the word he was uttering.
Pictorial representations of metaphors are often found in cartoons (Lakoff 1993:
241). Anger is commonly depicted by steam coming out of the cartoon character's ears, the
realization of the ANGER IS THE HEAT OF A FLUID IN A CONTAINER metaphor. What these
examples show is that metaphor is not simply a matter of language.

143
4. IDIOMS OF CRITICIZING: ANALYSIS

4.1 Introduction

4.1.1 Selection of relevant data

Our aim in this chapter is to apply the criteria listed at the end of chapter 2 to idioms
and idiom-like constructions of criticizing in order to rank the examples on a scale of
idiomaticity.
The word criticism seems to have three different meanings, which CCED (389)
paraphrases as 1) 'the action of expressing disapproval of something or someone', 2) 'a
statement that expresses disapproval' and 3) 'a serious examination and judgement of
something such as a book or play'. The first and second senses, as the paraphrases suggest,
are more closely related to each other than the third to either of them, and it is criticism in
the first two senses that we are primarily concerned with. Criticism in the first two senses is
directed at a person or an object, whereas criticism in the third sense is typically levelled at
an object. The picture, however, is more complicated in the latter case, because we can use
metonymy and refer to the person instead of the object. We are primarily interested in
idioms which denote criticism of a person. In order to avoid confusion and
misunderstanding, the action of expressing disapproval will be referred to as criticizing (see
the first paraphrase above), and criticism will be used for the actual words or sentences (see
the second paraphrase above). In the quotations and the dictionary/corpus examples we will
keep the original words.
As a starting point three dictionaries were used to collect idioms of criticizing. Two
of the dictionaries were thematic (Using Idioms and Picturesque Expressions), and the third
had a thematic section (Metaphorically Speaking). After a scrutiny of the meanings and
cross-checking in various dictionaries some of the expressions were discarded. They were
judged to be not sufficiently close in meaning to our target domain: be in the hot seat, be on
the warpath, box sb's ears, bulldoze some proposal through, face the music, nail one's
colours to the mast, blue-pencil, I knew him when, slings and arrows, a jaundiced eye,

144
mote in the eye, point a finger at sb, raise an eyebrow, strain at a gnat and swallow a
camel, trigger-happy, Don't hurry Hopkins, skin sb alive. Though phrasal verbs are also
idiomatic we are concerned with other types of idioms. In discussing the motivation of our
examples, however, we consider phrasal verbs and single words as well, since motivation by
conceptual metaphor is found behind phrasal verbs and single words used figuratively.
Twenty-two phrasal verbs and a number of single words have been collected. They are listed
in Appendix 5.
Selecting the relevant idioms was not very easy, since dictionary paraphrases of the
meanings are different and in several cases it is difficult, if not impossible, to decide whether
the given idiom is used primarily in the sense ‘criticize’ or in some other sense, such as
‘punish’ or ‘blame’. A number of expressions were added to the initial list taken from
various dictionaries (cf. the heading of the table in Appendix 6 and the list of abbreviations),
including an electronic dictionary, WordNet 1.6. This resulted in 72 examples to be
subjected to analysis. They are shown in Appendix 5. Appendix 6 shows the
lexicogrammatical variants given by several dictionaries, and Appendix 7 gives corpus
examples in the form of concordance lines. Some idioms or idiom variants are old-fashioned
(give sb the rough side of one's tongue) or rare, but most examples are current.
We are aware of the fact that a comprehensive phraseological analysis of the field of
criticizing must consider not only idioms, but all types of conventional expressions. This is,
unfortunately, beyond the scope of the present work. However, a far-from-exhaustive list of
collocations is given in Appendix 5 to present a fuller phraseological picture. The
collocations are taken from BBI and monolingual medium-sized dictionaries.
As we have seen, there is a fuzzy boundary line between idioms and collocations.
Consequently, our division of the examples into idioms and collocations (cf. 13.1 and 13.2)
involves a certain amount of arbitrariness. We have kept in the idiom list those expressions
that are included in most dictionaries of idioms. However, note that, for example, take a lot
of flak and come in for a lot of flak are treated as collocations not idioms in OALDb (483).
It is assumed that a more comprehensive and fine-tuned analysis than our present study will
eventually show whether they are more idiom-like or more collocation-like.

4.1.2 Treatment of dictionary and corpus data

145
We will assume that if a particular lexicogrammatical variant is given by at least half
of the dictionaries that contain the idiom, then the given variant exists. We do not treat all
the dictionaries in the same way, more weight will be attached to information coming from
recently published corpus-based dictionaries. Some problems arise due to the fact that
dictionaries do not always give clear indication of what is possible. For example, LPVD
(xiii) marks phrasal verbs that are usually passive, those that are never passive and those
that are normally only passive. In the rest of cases there is no special indication. We have
decided to use the examples in the dictionaries as well as the dictionary entry forms to
extract the relevant information. However, it is not certain at all whether the absence of
passive examples (or other grammatical variants) is due to the infrequency/impossibility of
the passive (or the given variant) or simply due to lexicographic omission. The reader
should bear this in mind throughout the discussion.
Since our corpus has limitations (unbalanced, of relatively small size), we will show
some caution. In particular, as is usual in corpus studies, we set a significance threshold,
which means that if a lexicogrammatical variant occurs three times or more, we will assume
that it is a normal (institutionalized) variant of the idiom. If the variant occurs twice or less,
we will ignore it. It may be exploitation or an error. As a general guideline, we will follow
the dictionary information, where there is disagreement between the corpus and dictionary
data.
The corpus also contains repetitions of some idioms together with the context
(paragraph(s) or separate sentence). We will follow the standard practice of not tampering
with the corpus, but a question mark will be used in front of a particular variant, if its
occurrence above the significance threshold is due to repetition.
The reader should assume that the degree of a given property indicated in the tables
below applies to all variant forms of the given idiom. If particular variants differ with
respect to the given property we will mention the difference.

4.2 Analysis

4.2.1 Compositionality

146
In armchair critic, etc. the word critic, etc. is used in its literal sense, and the ‘A +
N’ pattern matches the idiomatic meaning, so that the expression is partially compositional,
as opposed to armchair general.
Look who is talking seems to be a formula and it seems to be partially
compositional, since the look who is part often carries the sense 'sb shouldn't'. The variants
you can talk and you are a fine one to talk are noncompositional.
Although we have suggested that tell sb where to get off is noncompositional, tell
may carry its literal sense for some speakers.

armchair general 3 attract flak 3 Aunt Sally 3


backbiting 3 back-seat driver 3 be a sitting shot 3
call sb on the carpet 3 cast aspersions 2 cast sth in sb's teeth 3
come down on sb like a 3 come under fire 3 cover one's back 3
ton of bricks
curtain lecture 3 damn sb with faint praise 2 dip one's pen in gall 3
Don't come the uncle over 3 fire a salvo 3 get a lashing 3
me
get a lot of stick 3 get a rap on the knuckles 3 get a roasting 3
get a rocket 3 get on sb's back 3 get on sb's case 3
give sb a lecture 3 give sb down the banks 3 give sb grief 3
give sb Jesse 3 give sb the rough side of 3 give sb what for 3
one's tongue
haul sb over the coals 3 have a go 3 have a shot 3
have a thick skin 3 have one's ears slapped 3 have the hide of a 3
back rhinoceros
honeymoon 3 in the firing line 3 jump down sb's throat 3
kale through the reek 3 lash of scorpions 3 lay out in lavender 3
leave oneself open 3 like water off a duck's 3 look who's talking 2
back
Monday-morning 3 nitpick 3 not pull one's punches 3
quarterback
peanut gallery 3 people in glass houses 3 pin sb's ears back 3
shouldn't throw stones
pot shot 3 put the boot into sb 3 rap sb on the knuckles 3
read sb the riot act 3 roll with the punches 3 run the gauntlet 3
shoot sb down in flames 3 sitting duck 3 stand in the breach 3
stop-watch (critic) 3 take a hammering 3 take it on the chin 3
take sb to task 3 talk to sb like a Dutch 3 tear a strip off sb 3

147
uncle
tear sb to pieces 3 tell sb where to get off 3 the pot calling the kettle 3
black
throw sb to the wolves 3 tongue-lashing 3 turn one's guns on sb 3

4.2.2 Analyzability

We assume backbiting to be unanalyzable synchronically. However, NSOED (234)


says the verb bite used to have the meaning ‘speak sharply or deprecatingly (against); carp
(at)’.
Suggestions for analysis (the meaning chunks in brackets are not attached to any of
the constituents):
armchair general/critic ‘who has no direct experience + critic’
attract flak ‘attract + criticism’
be a sitting shot ‘be + whose actions make him very vulnerable to/easy to +
criticism/criticize’
cast aspersions ‘make + critical remarks’
come down on sb like a ton of bricks 'criticize + sb + severely'
come under fire 'come under + criticism'
cover one's back 'protect + oneself'; back carries no independent sense
curtain lecture '? + criticism'
damn sb with faint praise 'criticize + sb + with + weak + praise'
fire a salvo 'make + a + critical remark'
get a lashing 'get + criticism'
get a lot of stick 'get + a lot of + criticism', a stick to beat sb with is somewhat less
analyzable: 'sth + (you can use+) + to criticize + sb'
get a rap on the knuckles 'get + criticism'
get a roasting 'get + criticism'
get a rocket 'get + criticism'
give sb a lecture 'express + criticism'
give sb grief 'express + criticism'
give sb Jesse 'express + criticism'

148
give sb what for 'express + criticism'; what and for have no independent meanings
have a go 'make + an attack'
have a shot 'make + an attack'
have a thick skin, have the hide of a rhinoceros 'be + insensitive to criticism'
have one's ears slapped back 'get + criticized'
in the line of fire/in the firing line ‘in + a position/situation + (where you are likely
to be +) criticized/attacked’
leave oneself open 'make + oneself + vulnerable to criticism'
look who's talking 'look + who + is + talking/criticizing'
nitpick 'find + unimportant faults'
not pull one's punches 'not + weaken + one's + criticism'
people in glass houses shouldn't throw stones 'people + who have faults + shouldn't
+ criticize'
pot shot 'sudden/unexpected/unfair + criticism'
roll with the punches 'be able to deal + with + criticism'
shoot sb down in flames 'criticize + sb + very much/strongly'
sitting duck 'easy + target for criticism'
take a hammering 'receive + criticism'
take it on the chin 'accept + criticism + bravely'; on, the and chin seem to have no
independent senses
tell sb where to get off; if we take it as partially compositional and literal then it is
partially analyzable: 'tell + sb + angrily + what you think of them'
the pot calling the kettle black 'sb with a fault + criticizing + sb with the same fault';
calling...black corresponds to 'criticizing'
tongue-lashing '? + criticism'
turn one's guns on sb 'direct + one's + criticism + at sb'
We have asked our native speaker informants to judge the analyzability of some of
our idioms and provide plausible meanings (cf. 11.2). Except for pot shots, their intuitions
seem to be similar to the author's, as far as analyzability and the particular meanings are
concerned.

armchair general 1 attract flak 1 Aunt Sally 3

149
backbiting 3 back-seat driver 3 be a sitting shot 1
call sb on the carpet 3 cast aspersions 1 cast sth in sb's teeth 3
come down on sb like a 2 come under fire 1 cover one's back 2
ton of bricks
curtain lecture 2 damn sb with faint praise 1 dip one's pen in gall 3
Don't come the uncle over 3 fire a salvo 1 get a lashing 1
me
get a lot of stick 1 get a rap on the knuckles 2 get a roasting 1
get a rocket 1 get on sb's back 3 get on sb's case 3
give sb a lecture 1 give sb down the banks 3 give sb grief 1
give sb Jesse 1 give sb the rough side of 3 give sb what for 2
one's tongue
haul sb over the coals 3 have a go 2 have a shot 2
have a thick skin 2 have one's ears slapped 2 have the hide of a 2
back rhinoceros
honeymoon 3 in the firing line 1 jump down sb's throat 3
kale through the reek 3 lash of scorpions 3 lay out in lavender 3
leave oneself open 1 like water off a duck's 3 look who's talking 1
back
Monday-morning 3 nitpick 1 not pull one's punches 1
quarterback
peanut gallery 3 people in glass houses 2 pin sb's ears back 3
shouldn't throw stones
pot shot 1 put the boot into sb 3 rap sb on the knuckles 3
read sb the riot act 3 roll with the punches 1 run the gauntlet 3
shoot sb down in flames 2 sitting duck 1 stand in the breach 3
stop-watch (critic) 3 take a hammering 1 take it on the chin 2
take sb to task 3 talk to sb like a Dutch 3 tear a strip off sb 3
uncle
tear sb to pieces 3 tell sb where to get off 3 the pot calling the kettle 2
black
throw sb to the wolves 3 tongue-lashing 2 turn one's guns on sb 2

4.2.3 Extragrammaticality

The following idioms are extragrammatical: Don't come the uncle over me, give sb
down the banks, give sb what for, take sb to task, where there is no article before the

150
countable singular noun. The variant tear sb off a strip would be an ill-formed
transformational variant of tear a strip off sb if these were free combinations.

armchair general 1 attract flak 1 Aunt Sally 1


backbiting 1 back-seat driver 1 be a sitting shot 1
call sb on the carpet 1 cast aspersions 1 cast sth in sb's teeth 1
come down on sb like a 1 come under fire 1 cover one's back 1
ton of bricks
curtain lecture 1 damn sb with faint praise 1 dip one's pen in gall 1
Don't come the uncle over 3 fire a salvo 1 get a lashing 1
me
get a lot of stick 1 get a rap on the knuckles 1 get a roasting 1
get a rocket 1 get on sb's back 1 get on sb's case 1
give sb a lecture 1 give sb down the banks 3 give sb grief 1
give sb Jesse 1 give sb the rough side of 1 give sb what for 3
one's tongue
haul sb over the coals 1 have a go 1 have a shot 1
have a thick skin 1 have one's ears slapped 1 have the hide of a 1
back rhinoceros
honeymoon 1 in the firing line 1 jump down sb's throat 1
kale through the reek 1 lash of scorpions 1 lay out in lavender 1
leave oneself open 1 like water off a duck's 1 look who's talking 1
back
Monday-morning 1 nitpick 1 not pull one's punches 1
quarterback
peanut gallery 1 people in glass houses 1 pin sb's ears back 1
shouldn't throw stones
pot shot 1 put the boot into sb 1 rap sb on the knuckles 1
read sb the riot act 1 roll with the punches 1 run the gauntlet 1
shoot sb down in flames 1 sitting duck 1 stand in the breach 1
stop-watch (critic) 1 take a hammering 1 take it on the chin 1
take sb to task 2 talk to sb like a Dutch 1 tear a strip off sb 1
uncle
tear sb to pieces 1 tell sb where to get off 1 the pot calling the kettle 1
black
throw sb to the wolves 1 tongue-lashing 1 turn one's guns on sb 1

151
4.2.4 Lexical uniqueness

Although most dictionaries suggest that aspersions does not occur outside the
expression cast aspersions, in our corpus it occurs in other contexts.

armchair general 1 attract flak 1 Aunt Sally 1


backbiting 1 back-seat driver 1 be a sitting shot 1
call sb on the carpet 1 cast aspersions 1 cast sth in sb's teeth 1
come down on sb like a 1 come under fire 1 cover one's back 1
ton of bricks
curtain lecture 1 damn sb with faint praise 1 dip one's pen in gall 1
Don't come the uncle over 1 fire a salvo 1 get a lashing 1
me
get a lot of stick 1 get a rap on the knuckles 1 get a roasting 1
get a rocket 1 get on sb's back 1 get on sb's case 1
give sb a lecture 1 give sb down the banks 1 give sb grief 1
give sb Jesse 3 give sb the rough side of 1 give sb what for 1
one's tongue
haul sb over the coals 1 have a go 1 have a shot 1
have a thick skin 1 have one's ears slapped 1 have the hide of a 1
back rhinoceros
honeymoon 1 in the firing line 1 jump down sb's throat 1
kale through the reek 1 lash of scorpions 1 lay out in lavender 1
leave oneself open 1 like water off a duck's 1 look who's talking 1
back
Monday-morning 1 nitpick 1 not pull one's punches 1
quarterback
peanut gallery 1 people in glass houses 1 pin sb's ears back 1
shouldn't throw stones
pot shot 1 put the boot into sb 1 rap sb on the knuckles 1
read sb the riot act 1 roll with the punches 1 run the gauntlet 1
shoot sb down in flames 1 sitting duck 1 stand in the breach 1
stop-watch (critic) 1 take a hammering 1 take it on the chin 1
take sb to task 1 talk to sb like a Dutch 1 tear a strip off sb 1
uncle
tear sb to pieces 1 tell sb where to get off 1 the pot calling the kettle 1
black
throw sb to the wolves 1 tongue-lashing 1 turn one's guns on sb 1

152
4.2.5 Lexicogrammatical variation

4.2.5.1 Measuring variation

Comparing the lexicogrammatical variability of expressions is difficult, since idioms


can have two, three or more constituent words arranged in different patterns. Variability
must be measured not in absolute terms, but in relation to what is permitted by the rules that
apply to free combinations and in relation to the structure of the given idiom. We will
discuss grammatical variation in sections, dividing our examples into groups on the basis of
their structure. This will give us the opportunity to compare idioms conforming to the same
pattern, so that a more objective picture of variability can be drawn. Like Fraser (1970), we
will consider types of transformation, rather than individual transformations: inflection,
insertion, deletion and rearrangement.
In our treatment of lexical variants we follow the general practice of recording
lexical variation independently of meaning change. In certain cases the reader will see
idioms with opposite (figurative) meanings treated as variants (get on sb’s case, get off sb’s
case). They may be considered separate idioms. Our decision to treat these as variants is
justified inasmuch as we believe that one mark of the lexical variability of free combinations
is the substitution of words with opposite meanings.

4.2.5.2 Grammatical variation

The table below shows which transformations are possible. If a transformation is


inapplicable, the mark “-“ is placed in the slot. Where the information was insufficient to
determine variability, a question mark is used.

Inflection Addition Deletion Rearrang.


armchair general  - -
attract flak    
Aunt Sally  - -
backbiting - -
back-seat driver  - -
be a sitting shot ? ? ? ?
call sb on the carpet  
cast aspersions   -

153
cast sth in sb's teeth ? ? ? ?
come down on sb like a ton 
of bricks
come under fire   -
cover one's back 
curtain lecture -
damn sb with faint praise  
dip one's pen in gall ? ? ? ?
Don't come the uncle over me ? ? ? ?
fire a salvo  
get a lashing  
get a lot of stick  
get a rap on the knuckles 
get a roasting 
get a rocket 
get on sb's back 
get on sb's case 
give sb a lecture  
give sb down the banks ? ? ? ?
give sb grief 
give sb Jesse ? ? ? ?
give sb the rough side of  -
one's tongue
give sb what for  -
haul sb over the coals  
have a go 
have a shot 
have a thick skin  
have one's ears slapped back ? ? ? ?
have the hide of a rhinoceros  
honeymoon  - -
in the firing line - 
jump down sb's throat 
kale through the reek - ? ? ?
lash of scorpions - ? - ?
lay out in lavender ? ? ? ?
leave oneself open  
like water off a duck's back - 
look who's talking 
Monday-morning quarterback
nitpick  - - -
not pull one's punches  
peanut gallery - ? - ?
people in glass houses 
shouldn't throw stones
pin sb's ears back ? ? ? ?

154
pot shot  -
put the boot into sb 
rap sb on the knuckles  
read sb the riot act  
roll with the punches 
run the gauntlet  
shoot sb down in flames
sitting duck -
stand in the breach ? ? ? ?
stop-watch (critic) ? ? ? ?
take a hammering  
take it on the chin 
take sb to task  
talk to sb like a Dutch uncle  
tear a strip off sb  
tear sb to pieces  -
tell sb where to get off 
the pot calling the kettle
black
throw sb to the wolves  
tongue-lashing -
turn one's guns on sb ? ? ? ?

N+N
Our corpus does not have examples of armchair critic, and has only one occurrence
of armchair general. The word armchair is attributive in all our examples. Singular as well
as plural nouns can be used as the second element, and insertion of words between
armchair and the following noun seems to be possible when the inserted element forms a
compound with the head noun (football fan), so that we assume that modifying words
cannot be inserted. This is the reason that we have indicated non-occurrence of addition in
the table above.
Aunt Sally occurs in the corpus with adjectives only twice, with other determiners
than the indefinite article in the singular once and in the plural form once. Dictionary
examples give plural forms as well. We assume the following variants: an Aunt Sally and
Aunt Sallies.
Dictionaries show that the plural back-seat drivers does occur, although there is
only one plural example in our corpus. However, only one dictionary mentions the variant
back-seat driving, and it is found just above the significance threshold in our corpus, we
will assume that both the plural and the -ing variant exist.

155
Nitpick occurs in our corpus as a verb in the infinitive form and in an inflected form.
We assume it can be inflected. The forms nitpicking and nitpicker seem to be more
common.

armchair general 2 Aunt Sally 2 backbiting 3


back-seat driver 1 curtain lecture 3 honeymoon 2
Monday-morning 4 nitpick 1 peanut gallery ?
quarterback
pot shot 3 stop-watch (critic) ? tongue-lashing 3

A+N
The variant sitting target seems to be more variable.
sitting duck 4

N + P + NP
kale through the reek ? lash of scorpions ?

V + NP
Dictionary examples show that in get/take the flak the article can be omitted, but this
is not shown in our corpus examples. The zero article occurs below the significance
threshold with get, and with take 5 out of 6 examples are repetitions of the same sentence.
On the whole flak is more common with the definite than with the zero article. As far as
other determiners are concerned, in the corpus we find some, a lot of. Others occur twice or
once. Relative clause (cf. 6) and postmodifying past participle (cf. 11, 15) exemplify
rearrangement. Inflections cannot be attached to the noun, but this is not considered to be a
restriction, since the noun is uncountable.
We have not found information about be a sitting shot. It is probably a less common
variant of sitting duck/target.

attract flak 1 be a sitting shot ? cast aspersions 2


cover one's back 1 fire a salvo 2 get a lashing 3
get a lot of stick 3 get a rap on the knuckles 4 get a roasting 4
get a rocket 4 have a go 4 have a shot 4

156
have a thick skin 2 not pull one's punches 3 run the gauntlet 3
take a hammering 3

V + NP + NP
give sb a lecture 3 give sb grief 4 give sb Jesse ?
give sb the rough side of 4 read sb the riot act 3
one's tongue

V + NP + PP
Our corpus shows that addition of determiners/adjectives is possible in damn sb with
faint praise, and it also shows that the passive is permitted, though only two dictionaries
out of 8 give passive examples.
It is not clear whether modifiers can be added to boot in the idiom put the boot in.
Our corpus shows repetitions, so one must be cautious.
Our corpus shows several examples of the passive variant of take sb to task, and
despite the dictionary data, we consider the passive possible. There are only two examples
of take to task + NP, below the significance threshold.
call sb on the carpet 3 cast sth in sb's teeth ? damn sb with faint praise 3
dip one's pen in gall ? haul sb over the coals 3 put the boot into sb 4
rap sb on the knuckles 2 take it on the chin 4 take sb to task 3
tear a strip off sb 3 tear sb to pieces 4 throw sb to the wolves 3
turn one's guns on sb ?

V + P + NP
Though come under fire, hold fire and draw fire are given separate entries in idiom
dictionaries, we will assume that these are variants. They have the same metaphor. The
variants hold fire and draw fire seem to permit deletion of the determiner, they are therefore
more variable.
Get on sb's back is variable to some extent, but grammatical and lexical variation go
together.

come under fire 3 get on sb's back 3 get on sb's case 4

157
jump down sb's throat 4 roll with the punches 4 stand in the breach ?

P + NP
in the line of fire 3

Miscellaneous

Our data is not sufficient to determine the variability of talk to sb like a Dutch uncle,
but we rely here on our dictionaries. Similarly, although CCDI (307) says that the idiom is
variable, most other dictionaries do not show this. We assume invariability (on the basis of
our data), but a larger corpus may well show that we are wrong.

come down on sb like a 4 give sb down the banks ? Don't come the uncle over ?
ton of bricks me
give sb what for 4 have one's ears slapped ? have the hide of a 3
back rhinoceros
lay out in lavender ? leave oneself open 3 like water off a duck's 3
back
look who is talking 3 people in glass houses 2 pin sb's ears back ?
shouldn't throw stones
shoot sb down in flames 3 talk to sb like a Dutch 3 tell sb where to get off 4
uncle
the pot calling the kettle 4
black

4.2.5.3 Lexical variation

Although armchair occurs with many different nouns in our corpus, only six nouns
(audience, ballonist, investor, fan, traveller, tycoon) occur three or more times. Of these
ballonist is repeated in the same context. The meaning of armchair seems to be a bit
different in collocation with the noun audience, because it does not necessarily have a

158
negative connotation. It is simply used to contrast audience at home with audience at the
show, game, etc. (Near-)synonyms of fan also occur: enthusiast, devotee, aficionado.
Interestingly, not a single example of armchair critic can be found, but there is one example
of criticize…from a far-off armchair. Similarly, only one example of armchair general
occurs in the corpus, and only one dictionary (PE: 294) mentions the phrase. A larger and
more balanced corpus may well show that several other nouns can be used in the place of
general, so that the phrase may be marked as highly variable. Dictionary examples and
corpus examples yield the following lexical variants: armchair critic/traveller/audience/?
ballonist/investor/tycoon. Some variants (not all the idioms are listed below, for variants see
Appendix 6 and the corpus examples):
attract flak: take/get/prepare for/?aim + flak;
come under fire: attract/concentrate one's/draw (one's)/hold (one's)/be under/turn
(one's)/direct one's;
cover one's back: cover one's ass;
get a lashing: give sb a lashing, lexical and grammatical variation;
get a lot of stick: come in for/take/give sb a lot of stick; a stick with which to beat is
best regarded as a separate idiom, it occurs with no other verb;
get a rap on the knuckles: give sb a rap over/on the knuckles; dictionaries show a
range of verbs, which suggests that the verb is variable;
give sb a lecture: our corpus shows a number of verbs, although each occurring only
once or twice, but the verb may be variable;
not pull one's punches: although the any-variant occurs in our corpus only twice,
dictionaries show that it is common;
put the boot into sb: put the boot in, which changes the transitivity of the
expression;
take a hammering: our corpus shows hammering in three sentences with three
different verbs, which suggests relatively high variability, as opposed to the dictionary data;
tear sb to pieces: pick/pull sb to shreds;

armchair general 3 attract flak 3 Aunt Sally 4


backbiting 4 back-seat driver 4 be a sitting shot ?
call sb on the carpet 3 cast aspersions 4 cast sth in sb's teeth ?
come down on sb like a 4 come under fire 2 cover one's back 3

159
ton of bricks
curtain lecture 4 damn sb with faint praise 4 dip one's pen in gall ?
Don't come the uncle over ? fire a salvo 4 get a lashing 3
me
get a lot of stick 3 get a rap on the knuckles 2 get a roasting 3
get a rocket 3 get on sb's back 3 get on sb's case 3
give sb a lecture 2 give sb down the banks ? give sb grief 4
give sb Jesse ? give sb the rough side of 3 give sb what for 4
one's tongue
haul sb over the coals 3 have a go 4 have a shot 4
have a thick skin 3 have one's ears slapped ? have the hide of a 2
back rhinoceros
honeymoon 4 in the firing line 3 jump down sb's throat 4
kale through the reek ? lash of scorpions ? lay out in lavender ?
leave oneself open 3 like water off a duck's 4 look who's talking 3
back
Monday-morning 4 nitpick 4 not pull one's punches 3
quarterback
peanut gallery ? people in glass houses 3 pin sb's ears back ?
shouldn't throw stones
pot shot 4 put the boot into sb 3 rap sb on the knuckles 3
read sb the riot act 4 roll with the punches 4 run the gauntlet 4
shoot sb down in flames 4 sitting duck 3 stand in the breach ?
stop-watch (critic) ? take a hammering 3 take it on the chin 3
take sb to task 4 talk to sb like a Dutch 4 tear a strip off sb 4
uncle
tear sb to pieces 3 tell sb where to get off 4 the pot calling the kettle 4
black
throw sb to the wolves 3 tongue-lashing 4 turn one's guns on sb ?

4.2.6 Figurativity

Armchair general is fully figurative, but armchair critic, etc. are partially figurative.

armchair general 3 attract flak 4 Aunt Sally 3


backbiting 4 back-seat driver 3 be a sitting shot 4
call sb on the carpet 3 cast aspersions 4 cast sth in sb's teeth 3
come down on sb like a 3 come under fire 3 cover one's back 3

160
ton of bricks
curtain lecture 3 damn sb with faint praise 2 dip one's pen in gall 3
Don't come the uncle over 4 fire a salvo 3 get a lashing 3
me
get a lot of stick 3 get a rap on the knuckles 3 get a roasting 3
get a rocket 3 get on sb's back 3 get on sb's case 3
give sb a lecture 3 give sb down the banks 4 give sb grief 3
give sb Jesse 4 give sb the rough side of 4 give sb what for 4
one's tongue
haul sb over the coals 3 have a go 4 have a shot 3
have a thick skin 3 have one's ears slapped 3 have the hide of a 4
back rhinoceros
honeymoon 3 in the firing line 3 jump down sb's throat 4
kale through the reek 4 lash of scorpions 4 lay out in lavender 3
leave oneself open 4 like water off a duck's 3 look who's talking 2
back
Monday-morning 3 nitpick 3 not pull one's punches 3
quarterback
peanut gallery 3 people in glass houses 3 pin sb's ears back 3
shouldn't throw stones
pot shot 3 put the boot into sb 3 rap sb on the knuckles 3
read sb the riot act 3 roll with the punches 3 run the gauntlet 4
shoot sb down in flames 3 sitting duck 3 stand in the breach 3
stop-watch (critic) 3 take a hammering 4 take it on the chin 3
take sb to task 4 talk to sb like a Dutch 3 tear a strip off sb 3
uncle
tear sb to pieces 3 tell sb where to get off 3 the pot calling the kettle 4
black
throw sb to the wolves 3 tongue-lashing 4 turn one's guns on sb 3

4.2.7 Motivation

The motivation of our examples will be examined in chapter 5.

161
5. IDIOMS OF CRITICIZING: MOTIVATION

5.1 Introduction

Lakoff and Johnson (1980: 4) introduce the metaphor ARGUMENT IS WAR, and
among the examples illustrating it we find His criticisms were right on target. They mention
several types of argument and distinguish rational argument as well as one-party rational
argument from everyday argument (Lakoff and Johnson 1980: 87-88). In this study we are
interested in the latter type, which can be paraphrased as 'disagreement' or 'quarrel'.
Criticizing37 and arguing can be viewed as similar activities, but there are also differences
between them. Arguing is perhaps more typically bidirectional, with the participants taking
turns. Criticizing is more unidirectional. Lakoff and Johnson's example quoted above
suggests that criticism is conceptualized in the same way as argument, i.e. one of the
conceptual metaphors people use is CRITICIZING IS WAR. At the same time it is possible that
a closer scrutiny will reveal other metaphors as well.
We will explore various examples with similar meanings in order to answer the
following questions. To what extent are idioms of criticizing motivated by conceptual
metaphors, metonymies and encyclopedic knowledge? What source domains are mapped
onto the target domain? Are there any idioms that do not seem to be motivated by the above
mechanisms? Can we find examples other than idioms that show the same mapping? To
answer this question we will consider collocations and single words as well. A list of
figurative single words and collocations can be found in Appendix 5. We will attempt to
provide plausible synchronic motivation for our examples, which may or may not be in
accordance with the actual origin and development of the word or expression. Where the
origin of an idiom is known, or believed to be known, it is also given.

37
For the use of criticizing and criticism see 4.1.1.

162
5.2 Motivation by conceptual metaphors and metonymies

5.2.1 ABSTRACT IS CONCRETE

This is a widespread metaphor, and most of the examples given under 5.2.2 are
motivated by it. For example, finding sb’s faults (something abstract) is conceptualized as
picking nits (something concrete) in nitpick. Here we are interested primarily in more
specific versions of this metaphor, in which the domain of physical is mapped onto the
domain of emotional/psychological/social. Criticizing is a not only a verbal activity, but it is
also a psychological, emotional and social event. If criticizing is public, it can lead to a loss
of the social standing or rank of the people criticized. It is also unpleasant for them
psychologically/emotionally. Criticizing in private can equally be an emotional or
psychological ordeal.
More specific but still fairly general versions of the metaphor in the heading are
Lakoff and Johnson's (1980: 50) EMOTIONAL EFFECT IS PHYSICAL CONTACT and Lakoff's
(1987: 448) SOCIAL/PSYCHOLOGICAL HARM IS PHYSICAL HARM metaphor. Both metaphors
map something abstract (emotional/social/psychological) onto something concrete
(physical). Unpleasant emotional events are often conceptualized as physical forces that
affect the person, as in get a lashing, get a lot of stick, get a roasting, get a rocket, etc. In
some of our examples, the nature of this physical force is clear. In roll with the punches and
not pull one’s punches it is a punch, in put the boot into sb it is a kick. Below we will
attempt to find some metaphors that are compatible with ABSTRACT IS CONCRETE and
EMOTIONAL/SOCIAL/PSYCHOLOGICAL IS PHYSICAL but more specific.

5.2.2 CRITICIZING IS WAR

Given what we said in 5.1, it is perhaps best to begin our investigation with
idiomatic examples (including phrasal verb) motivated by this metaphor: attract flak38, be a
sitting shot, come under fire, cover one's back, fire a salvo, get a rocket, have a shot, in the
firing line, jump on sb, leave oneself open, pot shot, shoot sb down in flames, stand in the

38
In this discussion we will not list all the variants of a given idiom, and the reader is expected to assume
that the variants are motivated by the same metaphor.

163
breach, turn one's guns on sb; shoot down sb. We also find a number of conventional
expressions other than idioms and words that instantiate the same metaphor. For example,
adverse/devastating/hostile/sharp criticism, barrage of criticisms, level criticism at sb,
attack, attack sb (cf. Hungarian támad 'attack sb' in the same sense), blast, blast sb, barb,
cover oneself, fire, flak, onslaught, snipe at sb, target, target sb. All the idiom and non-
idiom examples are metaphorical except for be a sitting shot, which combines metaphor
with metonymy. Shot stands metonymically for 'person' still within the source domain and
the whole phrase is metaphorical.
Criticizing is conceptualized in terms of firing or shooting (shoot down sb, turn
one's guns on sb), or - more generally - attacking a person (attack, attack sb). A typical
manner of attacking is firing guns (attract flak, level criticism at sb, snipe at sb), but other
weapons could also be used (barb). The attacker is mapped onto the critic, and the weapon
corresponds to the criticism (level criticism at sb, barb, turn one's guns on sb). The person
criticized corresponds to the person who is attacked.
Note that the above correspondences do not seem to be completely fixed. It could
be argued that criticism (i.e. a critical remark/statement) does not always correspond to the
weapon, rather it is the bullets or the fire from the guns that maps onto criticism (attract
flak, fire a salvo at sb). If we are to decide which entity of the source domain corresponds
to which entity of the target domain, we must be able to determine whether a given noun is
used in the sense 'criticizing (i.e. the act of criticizing)' or 'criticism (i.e. the critical
remark/statement)' Unfortunately, the act of criticizing is not easy to separate from the
statements that we make when we are criticizing. In several dictionaries the two meanings
are not given separate subentries, and consequently they are not illustrated by separate
examples, as in CCED (389) or OALDb (298). Although CIDE (326) follows the same
practice, it marks the use of criticism as countable or uncountable and occasionally provides
paraphrases, so that it gives more guidance. In OALDa (277) and LDCE (327) the two
senses are illustrated separately. Collocations are helpful in showing the intended meaning,
since come under, attract and take tend to be followed by the word criticism in the sense
'(action of) criticizing', while level or make tend to be followed by criticism in the sense
'critical remark(s)/statement(s). It seems that criticizing (the action) rather than criticism
(the remarks) corresponds best to flak or fire in our idioms. It must be noted that there is a

164
strong metonymic connection between criticizing and criticism, just as there is a metonymic
relationship between guns and bullets or fire.
We can use our knowledge of the source domain to draw certain inferences about
the target domain. For example, we know that if people leave themselves open and do not
cover themselves, they can easily be attacked or shot at. In target domain terms, they can
easily be criticized (leave oneself wide open). We also know that if people are not protected
by the walls of a fortress, because there is an opening in the wall, they have to take most of
the attack. Therefore, it makes sense that stand in the breach should have the meaning 'bear
the brunt of criticisms' (MS: 437). Similarly, it is not completely arbitrary that hold fire
means 'defer criticizing' (MS: 436).
Although we have suggested that the source domain is war, it is not certain that all
the examples necessarily evoke a war scenario. Note first that with some examples listed
above it is not clear whether there are any weapons involved (attack sb). Second, the use of
weapons does not automatically lead to a war, since participants in various types of fighting
may use knives, guns and other objects, and hunting is an activity that also presupposes the
use of weapons. In fact, an etymological enquiry will reveal that pot shot 'a criticism of
someone which may be unexpected and unfair' (CCED: 1285) originally denoted random
shots taken at game without careful preparation. The word pot in the idiom refers to the
fact that the purpose of shooting was to provide a meal, i.e. fill the pot. (PE: 161). After
some time, the term was used outside hunting with the sense 'a shot taken at a defenseless
person or thing at close range from an advantageous position' (PE: 161). We haven't been
able to trace the origin of be a sitting shot and take a shot at sb, which might have taken the
same path in their semantic development as pot shot.

5.2.3 CRITICIZING IS PHYSICALLY HURTING

The discussion in the last paragraph of 5.2.2 prompts us to search for examples
which, interpreted literally, denote some type of a fight or clash between people, without

165
necessarily evoking a war scenario: cast sth in sb's teeth, get a lot of stick, get a rap on the
knuckles, get on sb's back, give sb the rough side of one's tongue, have a go, have a thick
skin, have one's ears slapped back, have the hide of a rhinoceros, jump down sb's throat,
not pull one's punches, pin sb's ears back, put the boot into sb, rap sb on the knuckles, roll
with the punches, take a hammering, take it on the chin, tear a strip off sb, tear sb to
pieces (cf. Hungarian szétcincál), tongue-lashing; go for sb, hit back, hit out, lash into sb,
lash out, lay into sb, rip into sb, slap down sb, tear into sb. Since criticizing is a verbal
activity, it is not surprising that the tongue is also referred to (give sb the rough side of
one's tongue, tongue-lashing). These idioms are metonymic, since the domain of speaking
includes the subdomain of the tongue, and criticizing can be seen as speaking angrily.
Metonymy can also be found in get a lot of stick, since the stick stands for beating
(INSTRUMENT FOR ACTION). Additionally, this variant is also motivated by the ACTIONS ARE

TRANSFERS metaphor, which is used to conceptualize actions as objects that are transferred
from the agent to the patient (Lakoff 1993: 216). Non-idiom examples whose source
domain (literal interpretation) denotes a type of fight or clash are assail sb, assault, assault
sb, bash sb, bashing, basher, belabour sb, engage in backbiting, flay sb, hammer sb,
hammering, knock sb, knocker, lash sb (cf. Hungarian ostoroz ‘whip’), lashing, rap sb,
slam sb and blistering/damaging/scathing criticism. All of them are metaphorical. Hammer
sb, hammering, lash into sb, lash out, and lash sb are also metaphorical ways of saying
'criticize' and 'criticism', though the formation of the verb hammer from the noun hammer
and the verbs lash (into/out) from the noun lash is based on the INSTRUMENT FOR ACTION

metonymy (cf. the discussion of Radden and Kövecses's examples in 3.2).


In the above examples physically hurting or hitting maps onto criticizing, the hurter
maps onto the critic, and the person who suffers injuries or pain corresponds to the person
who receives criticism. Since hurting and hitting often imply an attack, it is difficult to
separate these events. Some examples denote attack in their literal senses (have a go, go for
sb), others denote hitting (rap sb on the knuckles, hit out, bash sb) or hurting (tear a strip
off sb, flay sb). If taken literally, some of the examples denote unlikely situations (engage in
backbiting, jump down sb's throat), but they would also probably involve pain or injuries. It
is interesting to note that, as opposed to the examples in 5.2.2, the weapon is not a gun, but
a whip (lash) or a stick (give sb a lot of stick). In several cases no weapon is specified, or it

166
is clear that only the hands (or legs) are used (not pull one's punches, put the boot into sb,
tear sb to pieces).
One of the most common human experiences in which a person can hurt another
one, causing pain, is fight. Although the examples above need not evoke an actual fight, this
is likely to be the typical situation. The critic is conceptualized as causing the criticized
person some pain, and the one who undergoes criticism probably tries to defend themselves.
Genuine fighting is imitated in certain sports, where hurting the opponent to some extent
may be one of the goals, as in boxing. The expression pull one's punches actually has as its
origin a boxing term meaning 'to strike with less than one's full weight, to strike with a light
blow' (PDEI: 250). War is another common experience, but we take it to be a subtype of
fight. We believe that fight is more basic, more widespread. Lakoff and Johnson (1980: 61-
62, 84) also point out the ubiquity of fighting, and AN ARGUMENT IS A FIGHT is one of the
metaphors they use, though ARGUMENT IS WAR is more commonly found in their book.
Another common experience is (corporal) punishment, which also causes some pain without
any fight between the participants. The idioms get a lot of stick, have one's ears slapped
back and rap on the knuckles originally denoted forms of punishment (PE: 571, 572, 573).
Criticizing and punishment are so closely related that several idioms are polysemous, having
both the idiomatic sense ‘criticize’ and ‘punish’ (cf. for example OALDb: 237 and CIDPV:
52 for come down on sb). It is not uncommon for the punisher to be higher in rank or
authority than the person taking the punishment. This is reflected in the meaning of rap on
the knuckles, which refers to criticism made by an authority according to CCDI (232),
CCED (1360) and PDEI (99).
Having explored the metaphorical motivation behind many idioms of criticizing, we
can now understand why there are so many references to physical pain. Typically criticizing
is triggered by some action that has not been permitted, or that is foolish or dangerous, and
it may well be accompanied by punishment, including corporal punishment. Viewed in this
light, many expressions and words of criticizing may be metaphorically and metonymically
motivated at the same time, since the contiguous relation between criticizing and
punishment can be seen as a type of metonymy. Criticizing is mapped onto attacking, hitting
or hurting, and the conceptual metaphors involved are CRITICIZING IS A FIGHT, CRITICIZING

IS WAR and CRITICIZING IS PUNISHMENT. The following correspondences can be established:

167
Correspondences: Source domains/ICMs:
critic  attacker war, fight
criticizing  attacking
person criticized  person attacked
critic person hurting/hitting fight, punishment
criticizing  hurting, hitting
person criticized  person hurt/hit

Knowledge about the source domains can be used to make various inferences. The
idioms above suggest that the hurter (the critic) hurts the skin, the surface of the body,
rather than internal organs. Given that the typical weapon is the stick or the whip, this
seems to be a safe assumption. Encyclopedic knowledge tells us that thickness of the skin
can alleviate the pain, it is therefore not surprising that have a thick skin and have the hide
of a rhinoceros denote that the person is not affected by criticism. We also know that if
somebody is being taken to pieces, they experience excruciating pain, so that take sb to
pieces implies severe criticism.

5.2.4 ACCEPTING IS TAKING

This metaphor seems to motivate the idiom take it on the chin and the collocation
take criticism. This metaphor has not been mentioned before but it is rooted in our everyday
experience. We take various objects that are offered to us, and taking implies that we
“accept” those objects in the sense that we agree to use them. Taking is one of the most
common motor activities that the child does. We also take food, i.e. we accept what is
offered and eat it. Swallowing as one of the source domains for accepting has been
discussed and exemplified above.

5.2.5 BAD IS DOWN

As has been mentioned above, criticizing is associated with some unpleasant or


negative situation. Typically, the person who receives criticism is seen as negative or bad, at

168
least temporarily. One of the conceptual metaphors prevalent in our thinking is BAD IS

DOWN, and the following idioms are partially motivated by this same metaphor: come down
on sb, do down sb, dress down sb, put down sb. Note that several Hungarian words
denoting criticizing contain the co-verb le- 'down': lecikizés, ledorongolás, levágás, those
of criticize include leszól, ledorongol, and lehord (MSZKT: 498). Note also that
ledorongol (down + log + SUFFIX), ledorongolás (down + log + SUFFIX + SUFFIX) and
levágás (down + cut + SUFFIX) are motivated partly by the metaphor BAD IS DOWN, partly
by CRITICIZING IS A FIGHT or CRITICIZING IS PUNISHMENT. Furthermore, gáncsol 'trip, make
sb fall' in the sense 'criticize' and gáncs 'trip, an act of tripping' in the sense 'criticizing' also
seem to be motivated to some degree by BAD IS DOWN.

5.2.6 ACTIONS ARE TRANSFERS

We have already seen one example (get a lot of stick) for this metaphor, further
examples are get a lashing, the variants get a lot of/the flak and take a lot of/the flak, get a
roasting, the variant get/give sb a lot of stick, give sb a lecture. The act of criticizing is seen
as an object that can be transferred from one person to another. It is less likely that give sb
down the banks, give sb Jesse and give sb what for also evoke a similar image. The first and
the third are extragrammatical, while Jesse may imply transfer of an object, but the nature
of this object is not clear. Give sb Jesse is thought to have its origin in falconry, in which a
jess or jesse, the strap which secured the bird's leg to the falconer's wrist, was used as a
punishment for poor performance (PE: 571-72). As far as give sb what for is concerned, we
have been unable to trace its origin.
The idiom give sb grief also implies transfer, though not of action, but of a state.
Lakoff and Johnson (1999: 196) refer to this as the CAUSATION IS TRANSFER OF

POSSESSIONS metaphor.

5.2.7 ANGER IS HEAT, ANGER IS A DANGEROUS ANIMAL

The critic in a situation of criticizing is often angry. Therefore, we have examined


our idioms in order to check whether any of the anger metaphors posited by Lakoff (1987:

169
380-415) is also instantiated. We suggest that get a roasting is partially motivated by the
ANGER IS HEAT metaphor. Lakoff (1987: 383) gives two versions of this metaphor: ANGER

IS THE HEAT OF A FLUID IN A CONTAINER and ANGER IS FIRE. We hold the view that our
idiom is compatible with both versions of the metaphor, but especially with ANGER IS FIRE.

The critic is conceptualized as the heat source, and the criticized person suffers physical
pain. The ANGER IS HEAT metaphor is based on the physiological effects of anger, especially
increased body heat. Among the single-word synonyms flame sb, roast sb and roasting also
seem to be motivated by the ANGER IS FIRE metaphor.
Another anger metaphor posited by Lakoff (1987: 393) is ANGER IS A DANGEROUS

ANIMAL. At first sight there are two idioms that show this metaphor: lash of scorpions and
throw sb to the wolves. However, only the latter is a genuine example of the metaphor, since
in lash of scorpions 'an extremely severe punishment; an unusually harsh, vituperative, or
vitriolic chastisement or criticism' (PE: 572), the word scorpion refers to an ancient
instrument of punishment. It was a whip or lash with steel spikes. It may well be the case
that this idiom is partially motivated both by the CRITICIZING IS PHYSICALLY HURTING

(CRITICIZING IS PUNISHMENT) metaphor and the ANGER IS A DANGEROUS ANIMAL metaphor


for those who are unaware of the fact that scorpion does not denote the animal in this
idiom.
Though in the ANGER IS A DANGEROUS ANIMAL metaphor, the animal corresponds
to anger (cf. Lakoff 1987: 393), metonymically the angry person (the critic) can stand for
anger, so that the wolves to which the poor victim is thrown correspond to the critics. The
same metaphor also underlies the collocation fierce critic. The source domain of a
dangerous animal is especially fitting, because it reinforces the image of attacking/hurting. A
dangerous animal is likely to attack you and hurt you. Indeed certain idioms (get on sb’s
back, jump on sb) listed above under these metaphors could also derive some of their
motivation from the ANGER IS A DANGEROUS ANIMAL metaphor. Similarly, the collocations
biting criticism, savage attack/criticism, criticize sb savagely and the words claw sb,
savage sb all derive their motivation from mapping the source domain of dangerous animals
to the target domain of criticizing, or mapping the source domain of fight/punishment onto
the target domain of criticizing.

170
5.2.8 The CONDUIT METAPHOR

This metaphor is a combination of several metaphors, implying that ideas are


objects, these objects can be placed into containers, which are sent to the listener/reader.
The containers correspond to words or sentences, which are regarded as containing various
meanings, thoughts or ideas. The listener/reader can take them out of their containers and
take them into their heads, if the communication is successful. Another version of the
CONDUIT METAPHOR implies that ideas, thoughts and feelings are objects, but they are
simply ejected into external space by speakers/writers, and they are reified in this space, i.e.
they have independent existence, and they may or may not find their way into people's heads
(Reddy 1993: 170-71). It is possible that give sb the length of one's tongue is (partially)
motivated by the CONDUIT METAPHOR. It evokes the image of something transferred, though
it is not a typical example, because in this idiom it is not a linguistic expression that is
transferred. We prefer to view it as metonymical, the length of one's tongue standing for
criticism (i.e. critical remarks/statements).
Although the CONDUIT METAPHOR as a whole does not appear to underlie the other
examples, many expressions are partly motivated by the IDEAS ARE OBJECTS metaphor
(which is a more specific version of the ABSTRACT IS CONCRETE metaphor), especially those
that map criticism onto weapons or fire (attract flak, fire a salvo at sb, cast sth in sb's
teeth, etc.). The idiom be like water off a duck’s back can also be considered as partly
motivated by the same metaphor, since criticism is conceptualized as an object that leaves
the criticized person unaffected, just as water comes off a duck. Indeed, the idiom makes
sense only because we can conceptualize criticism as an object which can make physical
contact with the person criticized.

5.2.9 GENERIC IS SPECIFIC

Lakoff and Turner (1989: 162) introduce this metaphor to account for how we
interpret proverbs. Proverbs evoke fairly concrete images, specific events and actions, and
these specific schemas are mapped onto generic schemas. For example, when we encounter
the proverb the pot calling the kettle black, we extract some generic-level information. We

171
have two entities (the pot and the kettle in the specific proverb), and one entity points out
one of the properties of the other entity (the pot points out the blackness of the kettle). We
know that pots and kettles are both black, because they have been exposed to fire, but this
is a piece of conventional knowledge that is not expressed overtly in the proverb. Using this
knowledge, we conclude that one entity points out a property in the other that is present in
both entities. We also know that (the relevant parts of) pots and kettles often have the same
degree of blackness, so that our generic schema will contain the information that the
property can be present in both entities to the same extent. We also know that the property
we refer to is probably a negative one. Several expressions show that one of the conceptual
metaphors people make use of is BAD IS BLACK (blacken sb's name/character, The future
looks black, black magic, the black sheep of the family, She's not as black as you paint her,
etc.). This metaphor helps us interpret the word black as referring to something bad.
However, all this does not mean that the meaning is predictable!
In order to arrive at an interpretation of people in glass houses shouldn't throw
stones, we use conventional knowledge again. We know that a glass house is not ideal,
because it has several disadvantages over houses made of other materials. Similarly, people
who have faults are not ideal either. We know that throwing stones in a glass house is
foolish. But how do we know that the proverb is about criticizing? The expression throw
stones may imply this, because we have seen that there is a strong link between criticizing
and attacking/hurting somebody, but we do not want to claim that this link is evident in the
given proverb. Nevertheless, some people may rely on the mapping between criticizing and
hurting unconsciously, together with the pieces of information we have mentioned. Reliance
on the conceptual metaphor could provide additional motivation, since taken together with
the knowledge that throwing stones can break the glass and stone-throwers are exposed to
injuries caused by sharp pieces of glass (i.e. they can easily be hurt), the metaphor suggests
that those who criticize can just as easily be criticized by others. Those who do not rely on
this metaphor may remember the passage from the Bible ‘He that is without sin among you,
let him first cast a stone at her' (John VIII: 7).

5.2.10 UNIMPORTANT IS SMALL

172
The only idiom that is partially motivated by this metaphor is nitpick. Nitpicking, i.e.
pointing out small, unimportant faults, highlights an aspect of criticism that is different from
those that we have seen so far. The nit is the egg of a louse; therefore, it is very small, and
the image suggests that minor faults and errors are conceptualized as very small objects
which are found by the critic. Motivation comes via the metaphor only if the speaker has the
necessary encyclopedic knowledge about the size of nits.
One aspect of criticizing is expressing the low value or unimportance of somebody
or something that is criticized. This loss of importance and loss of value is also reflected in
rubbish sb and trash sb, which are not related to size, but the single-word synonyms belittle
and deflate, like the idiom mentioned above, are. Importance is conceptualized in terms of
size through the metaphors IMPORTANT IS BIG and UNIMPORTANT IS SMALL. They underlie
several idioms in the language (big fish/cheese/gun/wheel, bigwig, get too big for one’s
boots, make a big/great play of sth, etc.), and the basis of these metaphors is easy to
understand if you think of people in high/important positions. They tend to have big cars,
big offices, big desks, etc.

5.3 Other metaphors and metonymies

Some of our examples seem to be metaphorical, though not motivated by the


conceptual metaphors mentioned above. The examples in this group are different from those
in 5.4 below, since the idioms analyzed here evoke situations and experience that are
familiar to native speakers. Note that knowledge about a given situation may be the result
of personal experience or may come from cultural knowledge, education, etc. The relation
between the literal and the figurative meanings is relatively easily understood, i.e. the
examples are relatively motivated. In contrast, opaque idioms are unmotivated, with literal
interpretations that do not help the speaker see the connection with the idiomatic meaning.
It must be borne in mind that motivation varies from speaker to speaker, and it is a scalar
notion; consequently, the division between the examples listed here and those given under
5.4 below is not sharp. Some of the expressions that we have put in this group could easily
be considered as opaque by others. Conversely, some of our opaque examples may well be
partially motivated for other speakers.

173
armchair general

The meaning of the phrase armchair general is paraphrased in PE (294) as 'a person
removed from a given situation who thinks he could do a better job of directing it than
those actually in charge'. The word armchair can collocate in the same sense with critic and
other nouns as well, and it implies that the person referred to by the collocating noun has no
direct experience of a particular subject. They may have read or heard about it. Typically,
gaining experience requires leaving one's armchair, and in most cases one's home as well,
and moving about in the world. Those who are reluctant to abandon the comfort of their
armchair cannot expect to know much about a particular subject. Our example is therefore
(partially) motivated, and armchair critic is more motivated than armchair general if our
intended meaning is 'someone who criticizes without having direct experience'. The word
general derives its motivation from CRITICIZING IS WAR. Note that one of our conceptual
metaphors perhaps contributes to a certain degree to the motivatedness of our examples. If
LIFE IS MOTION, then living one's life and gaining experience involves motion, rather than
rest.

back-seat driver

A back-seat driver is someone who gives unsolicited advice or criticism, and the
relation between the literal and the idiomatic meanings seems to be clear. Passengers in the
back seat will often criticize car drivers, telling them how they should drive, what they
should do or should have done. ODEI (42) gives the following paraphrase for the idiomatic
sense: 'sb who criticizes, or tries to influence, the decisions, actions etc. of others which do
not concern him or which (perhaps because of his inferior position) he is unable, or
unwilling, to take himself'. It is the responsibility of the driver to drive the car, just as it is
the responsibility of those involved to make the decisions. The idiom is metaphorically
motivated, but it is not the whole source domain (driving) that is mapped onto the target
domain (criticizing), since we do not have other words and expressions that would show the
same mapping. We do not use driver in the sense 'person criticized', and sitting in the back
(of the car), or telling the driver what to do, etc. do not correspond to criticizing. Instead,

174
one concrete, rich image is used, and one of the participants is selected for target domain
exploitation.

bawl out

One of the meanings of bawl is 'to shout loudly, especially in an unpleasant or angry
way' (OALDb: 92), which means that there is a metonymic connection between the literal
and the figurative senses. When we criticize, we often speak loudly and angrily, but the
content of what we say is more important. The literal meaning refers to one of the common
or typical manners of reprimanding; therefore, the underlying metonymy could be called
MANNER FOR ACTION. This would be a plausible explanation of the origin of the idiom, but
the figurative sense is believed to have emerged on the ranches of the United States, where
it was used by cowboys and denoted the bawling of angry cattle (PE: 570).

curtain lectures

In this example as well as in give sb a lecture, lecture and lecture sb the emphasis is
not so much on anger and punishment as on teaching the proper behaviour. This is the basis
of the motivation. Lecturing in the literal sense involves teaching somebody various points,
and criticizing can be seen as a kind of teaching. It is the rational and educational rather
than the emotional and confrontational aspects of criticizing that are emphasized in the
above-mentioned examples.
As we can see, metaphorical usage allows us to highlight different aspects of one
and the same target concept. Sometimes it is the critic's anger or wish to punish the other
person that is salient, sometimes teaching. Curtain lectures is less motivated than lecture or
lecture sb, since the use of curtain is not well-motivated synchronically. It goes back to the
times when curtains hung around the bed, and the phrase originally referred to the wife's
nagging talk addressed to the husband in bed with the curtains drawn (ODEI: 126, LDEI:
71, PE: 571). CDWPO (64) says that the idiom variant Caudle lecture is derived from
papers called Mrs Caudle's Curtain Lectures, written by Douglas Jerrold and published in
Punch. In these papers Mr Caudle suffered the naggings of his wife after they had gone to
bed. Although we suggested above that punishment is not in the centre of our attention,

175
teach, a verb close in meaning to lecture, can be used in the sense 'punish', as shown by one
of the examples in OALDb (1333): I'll teach you to call me a liar (i.e. punish you for
calling me a liar).

damn sb with faint praise

This expression means that you praise someone so weakly that it is obvious that you
do not have a high opinion of them (CCDI: 308). The individual constituents seem to have
independent senses, which are used outside the expression as well. Damn means 'criticize'
outside this phrase as well, though it usually takes inanimate objects, and faint also retains
one of its senses in this word combination. Faint is used figuratively, but praise is used in its
literal sense. There is a relatively high degree of motivation, which is however not the result
of underlying conceptual metaphors.

dip one's pen in gall

PE (160) paraphrases the meaning of this idiom as 'to write with bitterness and spite;
to malign another'. The metonymy involved in the phrase is complex, since the instrument of
the action (the pen) is referred to in the expression, and the literal meaning of the whole
idiom denotes the start of the activity of writing, or more accurately, an activity which
preceded actual writing in the past, dipping one's pen in liquid ( SUBEVENT FOR WHOLE

EVENT). Metaphorical motivation comes from the use of gall, which is a bitter substance
secreted by the liver. The negative emotion associated with the idiomatic meaning, anger
and spitefulness, is linked with negative perception, bitter taste, another specific version of
the ABSTRACT IS CONCRETE metaphor.

honeymoon

The word denotes 'a short period enjoyed by a newcomer to some activity before his
performance is criticized' (MS: 436). Just as a newly married husband and wife are nice to
each other at the beginning of their marriage, especially during honeymoon, so newcomers
are treated with patience for a short period of time. In a stereotypical marriage, criticism is

176
likely to appear as the couple spend more and more time together. Although we would not
like to propose a conceptual metaphor whose source domain is married life (and the target
domain is criticism), note that this example, as well as curtain lectures, shows that certain
aspects of the husband-wife relationship can be used to talk about criticizing.

Monday morning quarterback

A Monday morning quarterback is someone who 'criticizes or judges something


unfairly, because although they now have full knowledge of the way things happened, the
people involved could not possibly have had that knowledge and so could not have behaved
any differently' (CCDI: 314). Since most football games are played on Sunday in the United
States, the phrase refers to those who discuss the particulars of the game and tell people
what the coach should have done to win. Motivation here depends on whether the speaker
is familiar with the given aspect of American culture.

5.4 Opaque idioms

haul sb over the coals, run the gauntlet

There are several idioms that seem to be unmotivated for the average native speaker
from a synchronic point of view. However, after tracing their origin, we see some of them in
new light. The examples listed in the heading all go back to a situation of causing someone
pain, hurting somebody, which is an important source domain in one of our conceptual
metaphors. Haul over the coals refers to the punishment of heretics (LDEI: 58, Funk 1948:
48). Run the gauntlet was formerly a military or naval punishment in which the offender was
forced to run between two rows of people who would hit him with a whip, leather thong or
rod. The word gauntlet is a corrupted form of a Swedish word given variously as gatloppe,
gatlop or gatulopp meaning 'a running of the lane' (Funk 1948: 178-79, CDWPO: 216,
DIO: 91). This type of punishment was also used in public schools (PDEI: 172).

177
Aunt Sally, lay out in lavender

These examples evoke a situation that is very similar to fight or punishment, which
means that the literal meaning involves some type of attacking or beating, but the target of
the attack is not a human being. ODEI (41) gives the following paraphrase for Aunt Sally:
'sb singled out as a target for abuse, criticism, or ridicule [...]; an object, or idea, deliberately
invented in order to attract destructive criticism, with the object of leading to constructive
thought'. The relationship between the literal and the figurative meaning is relatively clear, if
we know that Aunt Sally was originally an effigy, and people threw sticks or balls at it in
order to win a prize (LDEI: 283). PE (572) paraphrases lay out in lavender as 'to chastise
harshly and in no uncertain terms; to give someone a dressing down; to knock someone
down or unconscious; to kill someone'. The origin of lavender is uncertain, but the phrase is
thought to have referred to a physical act of beating. Branches of the lavender were once
used to beat freshly washed clothes (PE: 572).

call sb on the carpet

Both this expression and the verb carpet 'reprimand' date from the time when
servants were called into the sitting room or other carpeted room to be scolded by the
master (LDEI: 48, PE: 573). A slightly different explanation is offered by PDEI (152),
which places the scolding scenario into the boss's office, with the boss and the employee as
the participants. In the old days a carpet could be found only in the boss's office. Whichever
may be the true origin, the verb and the idiom are not motivated by any of our conceptual
metaphors.

Don't come the uncle over me

Probably totally opaque now for most native speakers, the phrase dates from ancient
Rome. The uncle's function was to point out the nephew's shortcomings, and uncles were
regarded as severe critics (PE: 161). BDPF (1134) gives the Latin form of the idiom: Ne sis
patruus mihi, i.e. do not overdo your privilege of reproving or castigating me.

178
give sb Jesse

This idiom was discussed in 5.2.6 above.

kale through the reek

The meaning is 'bitter language; unpleasant treatment; severe punishment' (PE: 572).
Synchronically the only motivation is the relationship between the unpleasantness denoted
by reek and the unpleasantness of criticism. The phrase can be traced back to cooking, more
specifically, to the preparation of the kale broth. The kale was cooked over an open fire, and
it had a bad smell (PE: 572).

peanut gallery

The idiom denotes a source of unimportant or insignificant criticism (PE: 393). The
phrase comes from the theatre, where peanuts and popcorn were sold only to the people in
the least expensive seats, those in the rear of the balcony, called peanut gallery. Those
occupying these seats were thought to have little appreciation of the arts; therefore, their
comments and criticism were not taken seriously.

read the riot act

Under the Riot Act (1715) if the people in a riotous gathering did not disperse after
they were given a warning in the form of the reading of the Riot Act, they risked being
arrested as felons (PE: 573). The penalty for disobedience was penal servitude or
imprisonment (Funk 1948: 141). Once the historical origin is known, it is relatively easy to
see why it can be used to refer to reprimanding.

179
stop-watch (critic)

The phrase comes from Laurence Sterne's Tristram Shandy, in which a person
listens to a soliloquy, paying close attention to when and how many times the speaker stops,
how long he pauses, but when the same person is asked about whether any gesture has filled
the speaker's silence or whether the line of thought has been broken as well, he replies 'I
looked only at the stopwatch, my lord' (PE: 162). The figurative meaning is paraphrased in
PE (162) as 'a hidebound formalist, whose focus is so riveted on traditional criteria or
irrelevant minutiae that he fails to attend to or even see the true and total object of his
concern'.

talk to sb like a Dutch uncle

Funk (1948: 183-84) claims that the idiom probably originated in the United States,
but the reference is not clear. In areas where Dutch colonists settled people were notorious
for their harsh discipline and "woe betide the unfortunate child who, having lost his own
parents, was obliged to depend upon an uncle as a foster parent" (Funk 1948: 183-84). The
idiom is slightly motivated, since the word Dutch is used in several expressions with
negative meanings.

be/get on sb's case, give sb down the banks, give sb what for, take sb to task

The expressions listed in the heading are likely to be the most opaque examples,
which does not necessarily mean that they have exactly the same degree of opacity for every
speaker.

5.5 Conclusion

Not counting the ABSTRACT IS CONCRETE and EMOTIONAL/SOCIAL/PSYCHOLOGICAL

IS PHYSICAL metaphors, which are rather general, out of 94 idioms 58 examples seem to be

180
motivated by the conceptual metaphors discussed above. In fact, the number is higher if we
count all the variants separately. The table below shows the degree of motivation, as seen by
the author of this work.

armchair general 2 attract flak 2 Aunt Sally 3


backbiting 2 back-seat driver 2 be a sitting shot 2
call sb on the carpet 3 cast aspersions 1 cast sth in sb's teeth 2
come down on sb like a 2 come under fire 2 cover one's back 2
ton of bricks
curtain lecture 3 damn sb with faint praise 1 dip one's pen in gall 2
Don't come the uncle over 3 fire a salvo 2 get a lashing 2
me
get a lot of stick 2 get a rap on the knuckles 2 get a roasting 2
get a rocket 2 get on sb's back 2 get on sb's case 3
give sb a lecture 2 give sb down the banks 3 give sb grief 2
give sb Jesse 3 give sb the rough side of 2 give sb what for 3
one's tongue
haul sb over the coals 3 have a go 2 have a shot 2
have a thick skin 2 have one's ears slapped 2 have the hide of a 2
back rhinoceros
honeymoon 2 in the firing line 2 jump down sb's throat 2
kale through the reek 3 lash of scorpions 3 lay out in lavender 3
leave oneself open 2 like water off a duck's 2 look who's talking 1
back
Monday-morning 2 nitpick 2 not pull one's punches 2
quarterback
peanut gallery 3 people in glass houses 2 pin sb's ears back 2
shouldn't throw stones
pot shot 2 put the boot into sb 2 rap sb on the knuckles 2
read sb the riot act 3 roll with the punches 2 run the gauntlet 3
shoot sb down in flames 2 sitting duck 2 stand in the breach 2
stop-watch (critic) 3 take a hammering 2 take it on the chin 2
take sb to task 3 talk to sb like a Dutch 3 tear a strip off sb 2
uncle
tear sb to pieces 2 tell sb where to get off 3 the pot calling the kettle 2
black
throw sb to the wolves 2 tongue-lashing 2 turn one's guns on sb 2

181
We assume that idioms strongly motivated by conceptual metaphors are partially
motivated, though within this type there are differences. At present it is not clear how to
distinguish further degrees of motivation. Since this type of judgement is subjective, we
have asked our native speaker informants to grade some of our idioms from the point of
view of motivation. The results are shown in 11.3. It is clear that speakers often disagree.
On the whole, backbiting, get on sb's back, read sb the riot act and take sb to task seem to
be more motivated for our informants than for the author.
We do not want to claim that all the idioms marked "2" have the same degree of
motivation. It is probable that a particular idiom that is motivated by two or more specific
conceptual metaphors that reinforce each other (jump on sb) is more motivated than an
idiom behind which there is only one conceptual metaphor. If a given metaphor is
widespread, so that there are numerous examples, it may also be felt more motivated than
an idiom with a "less productive" metaphor.
Finally, we grade our examples on the basis of their scores as shown in this chapter
and the previous one. The scores are simply added up to yield a number, which shows a
range between 8 and 27. The higher the score, the more idiomatic the expression is assumed
to be. The grading is inevitably subjective. Our method implies the equal treatment of the
criteria. Future research should show whether some properties have to be given more
weight or not. Where we cannot determine lexicogrammatical motivation, a question mark
is used.

8 = least idiomatic, 27 = most idiomatic


armchair general 16 attract flak 16
Aunt Sally 20 backbiting 21
back-seat driver 15 be a sitting shot 12 + ?
call sb on the carpet 20 cast aspersions 16
cast sth in sb's teeth 13 + ? come down on sb like a ton of bricks 20
come under fire 16 cover one's back 16
curtain lecture 20 damn sb with faint praise 15
dip one's pen in gall 13 + ? Don't come the uncle over me 17 + ?
fire a salvo 17 get a lashing 17
get a lot of stick 17 get a rap on the knuckles 18
get a roasting 18 get a rocket 18
get on sb's back 19 get on sb's case 21
give sb a lecture 16 give sb down the banks 17 + ?
give sb grief 19 give sb Jesse 15 + ?
give sb the rough side of one's tongue 21 give sb what for 24

182
haul sb over the coals 20 have a go 21
have a shot 20 have a thick skin 17
have one's ears slapped back 12 + ? have the hide of a rhinoceros 18
honeymoon 19 in the firing line 17
jump down sb's throat 22 kale through the reek 15 + ?
lash of scorpions 15 + ? lay out in lavender 14 + ?
leave oneself open 18 like water off a duck's back 20
look who's talking 13 Monday-morning quarterback 20
nitpick16 not pull one's punches 17
peanut gallery 14 + ? people in glass houses shouldn't throw
stones 17
pin sb's ears back 13 + ? pot shot 18
put the boot into sb 20 rap sb on the knuckles 18
read sb the riot act 21 roll with the punches 19
run the gauntlet 22 shoot sb down in flames 19
sitting duck 18 stand in the breach 13 + ?
stop-watch (critic) 14 + ? take a hammering 18
take it on the chin 19 take sb to task 23
talk to sb like a Dutch uncle 21 tear a strip off sb 20
tear sb to pieces 20 tell sb where to get off 22
the pot calling the kettle black 21 throw sb to the wolves 19
tongue-lashing 20 turn one's guns on sb 12 + ?

It is expected that the above grading of idiomaticity should not be too different from
the intuition of native speakers, especially from the person who does the analysis (in this
case the author of this work). Some of the scores do not meet our intuition. For example,
we feel that Aunt Sally is highly idiomatic, and we expected a higher score, around 25. As a
tendency the scores and our intuition do not differ substantially, but we must bear in mind
that idiomaticity is a subjective notion and speakers differ in their intuitions. On the whole,
idioms motivated by conceptual metaphors (attract flak, get a lashing, take a hammering)
are less idiomatic than those not motivated by conceptual metaphors (call sb on the carpet,
take sb to task), though the appropriacy of metaphor need not depend on whether it is
conceptual or not. For example, back-seat driver has a fairly low degree of idiomaticity on
our scale (score 15), though it is not motivated by conceptual metaphors.
If the grading above does not meet the intuition of the reader, it is possible that
treating the criteria with equal weight is not the best method. Future research will have to
decide whether there are any criteria that are more important. Further research will also tell
us whether it is enough to set up 3 or 4 degrees of a given property.

183
6. SUMMARY

Our aim was to survey phraseological research in order to determine what criteria
can be used to measure the idiomaticity of expressions and what types of phraseological
units are distinguished in the literature. We have also made an attempt to determine the
idiomaticity of idioms of criticizing and discover conceptual metaphors underlying the same
idioms. The most common property of idioms is their noncompositionality. We have
distinguished discrepancy, predictability and motivation, and followed cognitive grammar’s
theoretical framework. These properties, as most other properties of conventional
expressions, are scalar. We can see a cline of discrepancy between the compositional and the
intended (figurative, idiomatic) interpretation in such phrases as gin and tonic, drop names,
and roll up one’s sleeves.
We have set the lower boundary of the field of phraseology between single words
and compounds. We believe that phraseologists’ approach to word combinations as well as
the notion of compositionality presuppose more than a single word for analysis. Single
words and many idioms are, however, similar in certain respects (structural restrictedness,
semantic unity) and they can be motivated by the same metaphor (struggle, come to grips,
where anger is metaphorically seen as an opponent), or can go back to the same fairly
specific image (point the finger at, finger sb).
Weinreich (1969) rightly pointed out the analyzability of many idioms, though he
wanted to treat all types of idioms (unanalyzable as well as analyzable) in the same fashion
as collocations, which are analyzable. Chafe (1968) focused on the idiosyncratic nature of
idioms. Among these we find extragrammaticality and lexical uniqueness. We have
suggested that extragrammaticality and lexical idiosyncrasy can be considered graded
categories, as can grammatical variation. The transformational variability of idioms can be
checked ideally only on a large corpus (in the hundred millions).
Figurativity is an important property of idioms and many collocations. We have
suggested that if the figurative meaning is motivated, the idiom does not have a
homonymous literal counterpart, since homonymy implies lack of relatedness. We have also
pointed out that many idioms that can have literal interpretation without context are not

184
interpreted literally in the given context. Figurativity, as well as lexical restrictedness, is
generally used to distinguish restricted collocations from idioms, but other criteria must also
be used, since figurativity and lexical variation cannot always help us (promise the moon,
white lie, talk shop, square meal). Motivation and the relation between figurativity and the
structure must also be considered. In typical ‘verb + noun’ collocations, the verb is
figurative and the noun tends to be literal. In terms of structure, there is a basic division
between word-like and sentence-like units, but we have seen that pure structural
classification is mixed with functional and other criteria (such as whether the source is
written or spoken, known or unknown).
Although prototypical idioms are unmotivated, many idioms seem to be motivated
(partially), so that motivation is also a graded criterion. Phraseologists differ in their
preferences. Cognitive linguists emphasize the relative motivatedness of idioms, others (e.g.
Fernando (1996)) focus on lexical variation, rather than motivation and figurativity.
There is a complex relationship between the various criteria, and this has also been
discussed. Analyzability seems to depend on partly motivation, partly the structure of the
expression. Due to the motivatedness of swallow in the sense ‘accept’, it carries this
independent meaning in swallow a bitter pill. At the same time, analyzability need not
presuppose a high degree of motivation, since cat seems to have the independent sense
‘secret’ in let the cat out of the bag. Motivation and analyzability, just as motivation and
predictability, are independent notions.
We have established he following criteria of idiomaticity: compositionality,
analyzability, extragrammaticality, lexical uniqueness, grammatical variation, figurativity,
motivation and lexical variation. Before we applied these to idioms of criticizing, we
surveyed the conceptual mechanisms of metaphor and metonymy. Metonymy does not imply
domain mapping, while metaphor does. There are idioms motivated by metonymy alone
(bricks and mortar), metaphor alone (a load off sb’s mind). We have surveyed the
conceptual metaphors posited by linguists and suggested new metaphors ( DESIRE IS THIRST

(thirst for knowledge)).


We have applied the criteria of idiomaticity to a number of idioms of criticizing to
see whether their degree of idiomaticity can be established in this way. The answer seems to
be affirmative, though the analysis is offered as tentative. We have also investigated the
motivation of the same idioms and posited new conceptual metaphors ( CRITICIZING IS

185
PHYSICALLY HURTING, CRITICIZING IS A FIGHT). We believe that our analysis can be applied
to idioms with other meanings and/or in other languages.

186
7. REFERENCES

Aitchison, Jean. 1994. Words in the Mind. Oxford: Blackwell.

Allerton, D. J. 1984. Three (or four) levels of word co-occurrence restriction. Lingua 63:
17-40.

Arnold, Irina Vladimirovna. 1986. The English Word. Moscow: Vysshaja Shkola.

Ball, W. J. 1958. A Practical Guide to Colloquial Idiom. London: Longman.

Barcelona, Antonio. 1997. Guidelines for the application of the theories of metaphor and
metonymy to textual examples. ESSE/4 conference paper. Unpublished.

Barkema, Henk. 1994. Determining the syntactic flexibility of idioms. In: Fries, Udo,
Gunnel Tottie and Peter Schneider. (eds.) Creating and Using English Language Corpora.
Amsterdam-Atlanta: Rodopi.

Barkema, Henk. 1996. The effect of inherent and contextual factors on the grammatical
flexibility of idioms. In: Percy, Carol E., Charles F. Meyer, and Ian Lancashire. (eds.)
Synchronic Corpus Linguistics. Amsterdam-Atlanta: Rodopi.

Bauer, Laurie. 1983. English Word-formation. Cambridge: Cambridge University Press.

Bauer, Laurie. 1998. When is a sequence of two nouns a compound in English? English
Language and Linguistics 2/1: 65-86.

Bencédy, József, Pál Fábián, Endre Rácz and Mártonné Velcsov. 1988 (1st ed 1968
reprinted). A mai magyar nyelv. Budapest: Tankönyvkiadó.

187
Benson, Morton. 1985. Collocations and idioms. In: Ilson, Robert. (ed.) Dictionaries,
Lexicography and Language Learning. Oxford: Pergamon Press and The British Council.

Benson, Morton, Evelyn Benson and Robert Ilson. 1986a. The BBI Combinatory
Dictionary of English: A Guide to Word Combinations. Amsterdam and Philadelphia: John
Benjamins.

Benson, Morton, Evelyn Benson and Robert Ilson. 1986b. Lexicographic Description of
English. Amsterdam and Philadelphia: John Benjamins.

Brown, Lesley. (ed.) 1993. The New Shorter Oxford English Dictionary I-II. Oxford:
Clarendon Press.

Brederode, Thomas van. 1995. Collocation Restrictions, Frames and Metaphor. PhD
dissertation. Amsterdam.

Carter, Ronald. 1987. Vocabulary. London: Unwin Hymen.

Chafe, Wallace. L. 1968. Idiomaticity as an anomaly in the Chomskyan paradigm.


Foundations of Language 4: 109-127.

Clausner, C. Timothy and William Croft. 1999. Domains and image schemas. Cognitive
Linguistics 10-1: 1-31.

Coulmas, Florian. 1994. Formulaic language. In: Asher, R. E. and J. M. Y. Simpson. (eds.)
The Encyclopedia of Language and Linguistics. Oxford: Pergamon Press.

Cowie, Anthony P. 1988. Stable and creative aspects of vocabulary use. In: Carter, Ronald
and Michael McCarthy. (eds.) Vocabulary and Language Teaching. London: Longman.

Cowie, Anthony P. 1994. Phraseology. In: Asher, R. E. and J. M. Y. Simpson. (eds.) The
Encyclopedia of Language and Linguistics. Oxford: Pergamon Press.

188
Cowie, Athony P. 1998b. Introduction. In: Cowie, Anthony P. (ed.) Phraseology. Oxford:
Clarendon Press.

Cowie, Anthony P. 1998c. Phraseological dictionaries: some east-west comparisons. In:


Cowie, Anthony P. (ed.) Phraseology. Oxford: Clarendon Press.

Cowie, Anthony P. and Ronald Mackin. 1993. Oxford Dictionary of Phrasal Verbs. Oxford:
Oxford University Press.

Cowie, Anthony P., Ronald Mackin and Isabel R. McCaig. 1993. Oxford Dictionary of
English Idioms. Oxford: Oxford University Press.

Croft, William. 1993. The role of domains in the interpretation of metaphors and
metonymies. Cognitive Linguistics 4-4: 335-370.

Crowther, Jonathan. (ed.) 1995. Oxford Advanced Learner's Dictionary of Current English.
Oxford: Oxford University Press.

Cruse, D. A. 1986. Lexical Semantics. Cambridge: Cambridge University Press.

Crystal, David. 1995. The Cambridge Encyclopedia of the English Language. Cambridge:
Cambridge University Press.

Cserép, Attila. 1998. Comparative Analysis of Some English and Hungarian Body Part
Idioms Connected With Communication and Thinking. In: Heid, Ulrich (ed.) 3rd
International Symposium on Phraseology: ISP-3 Proceedings. Stuttgart: Universität
Stuttgart.

Cserép, Attila. (forthcoming) The Compositionality of Idioms. In: Korponay, Béla, and
Péter Pelyvás. (eds.) Studies in Linguistics IV. Debrecen: Debreceni Egyetem.

189
Evans, Ivor Henry. 1989. Brewer’s Dictionary of Phrase and Fable. London: Cassell.

Fernando, Chitra. 1996. Idioms and Idiomaticity. Oxford: Oxford University Press.

Fillmore, Charles J., Paul Kay and Mary Catherine O'Connor. 1988. Regularity and
idiomaticity in grammatical constructions: the case of let alone. Language 64/3: 501-538.

Flavell, Linda and Roger Flavell. 1992. Dictionary of Idioms and Their Origins. London:
Kyle Cathie.

Földes, Csaba. 1987. Magyar-német-orosz beszédfordulatok. Budapest: Tankönyvkiadó.

Fox, Chris. (ed.) 2000. Longman Phrasal Verbs Dictionary. Harlow: Pearson Education
Limited.

Fraser, Bruce. 1970. Idioms within a transformational grammar. Foundations of Language


6: 22-42.

Funk, Charles Earle. 1948. A Hog on Ice. New York: Harper and Row.

Gadsby, Adam. (ed.) 1995. Longman Dictionary of Contemporary English. Harlow:


Longman Group Ltd.

Geeraerts, D. 1994. Metonymy. In: Asher, R. E. and J. M. Y. Simpson. (eds.) The


Encyclopedia of Language and Linguistics. Oxford: Pergamon Press.

Gibbs, Raymond W., Jr. 1994. The Poetics of Mind. Cambridge: Cambridge University
Press.

Gibbs, Raymond W., Jr. and Gayle P. Gonzales. 1985. Syntactic frozenness in processing
and remembering idioms. Cognition 20: 243-259.

190
Gibbs, Raymond W. and Jennifer E. O'Brien. 1990. Idioms and mental imagery: the
metaphorical motivation for idiomatic meaning. Cognition 36: 35-68.

Gläser, Rosemarie. 1986. Phraseologie der englischen Sprache. Leipzig: VEB Verlag
Enzyklopädie.

Gläser, Rosemarie. 1998. The stylistic potential of phraseological units in the light of genre
analysis. In: Cowie, Anthony P. (ed.) Phraseology. Oxford: Clarendon Press.

Grady, Joseph. 1997. Foundations of Meaning: Primary Metaphors and Primary Scenes.
Doctoral dissertation, U.C. Berkeley.

Gulland, Daphne M. and David Hinds-Howell. 1986. The Penguin Dictionary of English
Idioms. Harmondsworth: Penguin Books.

Hadrovics, László. 1995. Magyar frazeológia: történeti áttekintés. Budapest: Akadémiai


Kiadó.

Heaton, John Brian and T. W. Noble. 1987. Using Idioms. New York: Prentice Hall.

Hockett, Charles F. 1958. A Course in Modern Linguistics. New York: The Macmillan
Company.

Howarth, Peter Andrew. 1996. Phraseology in English Academic Writing. Tübingen: Max
Niemeyer Verlag.

Jackendoff, Ray. 1997. The Architecture of the Language Faculty. Cambridge,


Massachusetts: The MIT Press.

Johnson, Mark. 1987. The Body in the Mind. Chicago: The University of Chicago Press.

191
Juhász, József. 1980. A frazeológia mint nyelvészeti diszciplína. In: Rácz, Endre and István
Szathmári. (eds.) Tanulmányok a mai magyar nyelv szókészlettana és jelentéstana körébõl.
Budapest: Tankönyvkiadó.

Juhász, József, István Szõke, Gábor O. Nagy and Miklós Kovalovszky. (eds.) 1972.
Magyar értelmezõ kéziszótár. Budapest: Akadémiai Kiadó.

Kiss, Gábor. (ed.) 1998. Magyar szókincstár. Budapest: Tinta Könyvkiadó.

Kövecses, Zoltán and Péter Szabó. 1996. Idioms: a view from cognitive semantics. Applied
Linguistics 17/3: 326-355.

Lakoff, George. 1987. Women, Fire, and Dangerous Things. Chicago: The University of
Chicago Press.

Lakoff, George. 1993. The contemporary theory of metaphor. In: Ortony, Andrew. (ed.)
Metaphor and Thought. Cambridge: Cambridge University Press.

Lakoff, George and Mark Johnson. 1980. Metaphors We Live By. Chicago: The University
of Chicago Press.

Lakoff, George and Mark Johnson. 1999. Philosophy in the Flesh. New York: Basic Books.

Lakoff, George and Mark Turner. 1989. More than Cool Reason. Chicago: The University
of Chicago Press.

Langacker, Ronald W. 1987. Foundations of Cognitive Grammar I. Stanford: Stanford


University Press.

Langacker, Ronald W. 1993. Reference-point constructions. Cognitive Linguistics 4-1: 1-


38.

192
Lipka, Leonhard. 1992. An Outline of English Lexicology. Tübingen: Max Niemeyer
Verlag.

Long, Thomas Hill. (ed.) 1979. Longman Dictionary of English Idioms. Harlow and
London: Longman Group Limited.

Makkai, Adam. 1972. Idiom Structure in English. The Hague: Mouton.

McArthur, Tom. (ed.) 1992. The Oxford Companion to the English Language. Oxford:
Oxford University Press.

Mel'čuk, Igor A. 1998. Collocations and lexical functions. In: Cowie, Anthony P. (ed.)
Phraseology. Oxford: Clarendon Press.

Moon, Rosamund. (ed.) 1995 Collins Cobuild Dictionary of Idioms. London:


HarperCollins Publishers.

Moon, Rosamund. 1998a. Fixed Expressions and Idioms in English. Oxford: Clarendon
Press.

Moon, Rosamund. 1998b. Frequencies and forms of phrasal lexemes in English. In: Cowie,
Anthony P. (ed.) Phraseology. Oxford: Clarendon Press.

Nagy, György. 1996. Angol-magyar idiómaszótár. Budapest: Akadémiai Kiadó.

Nunberg, Geoffrey, Ivan A. Sag and Thomas Wasow. 1994. Idioms. Language 70/3: 491-
538.

O. Nagy, Gábor. 1985 (1st ed 1966 reprinted). Magyar szólások és közmondások.


Budapest: Gondolat.

O. Nagy, Gábor. 1999 (3rd ed 1979 reprinted). Mi fán terem? Budapest: Talentum.

193
Procter, Paul. (ed.) 1995. Cambridge International Dictionary of English. Cambridge:
Cambridge University Press.

Pulman, Stephen G. 1993. The recognition and interpretation of idioms. In: Cacciari,
Cristina and Patrizia Tabossi. (eds.) Idioms: Processing, Structure, and Interpretation.
Hillsdale, New Jersey: Lawrence Erlbaum Associates.

Pye, Glennis. (ed.) 1997. Cambridge International Dictionary of Phrasal Verbs.


Cambridge: Cambridge University Press.

Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech and Jan Svartvik. 1985. A
Comprehensive Grammar of the English Language. London: Longman.

Radden, Günter and Zoltán Kövecses. 1999. Towards a Theory of Metonymy. Offprint from
Panther, Klaus-Uwe and Günter Radden. (eds.) Metonymy in Language and Thought.
Amsterdam/Philadelphia: John Benjamins Publishing Company.

Radford, Andrew. 1988. Transformational Grammar. Cambridge: Cambridge University


Press.

Reddy, Michael J. 1993. The conduit metaphor: a case frame conflict in our language about
language. In: Ortony, Andrew. (ed.) Metaphor and Thought. Cambridge: Cambridge
University Press.

Rees, Nigel. 1979. Graffiti Lives, OK. London: Unwin Paperbacks.

Rees, Nigel. 1996. The Cassell Dictionary of Word and Phrase Origins. London: Cassell.

Renton, Nick E. 1992. Metaphorically Speaking. New York: Warner Books.

194
Ricoeur, Paul. 1986. Between rhetoric and poetics: Aristotle. In: The Rule of Metaphor.
Translated by Robert Czerny with Kathleen McLaughlin and John Costello. London:
Routledge.

Seidl, Jennifer and W. McMordie. 1988. English Idioms. Oxford: Oxford University Press.

Sinclair, John. (ed.) 1995. Collins Cobuild English Dictionary. London: HarperCollins
Publishers.

Smith, Logan Pearsall. 1943. Words and Idioms. London: Constable and Company Ltd.

Steinhart, E. and E. F. Kittay. 1994. Metaphor. In: Asher, R. E. and J. M. Y. Simpson. (eds.)
The Encyclopedia of Language and Linguistics. Oxford: Pergamon Press.

Stern, Karen. (ed.) 1998. Longman Idioms Dictionary. Harlow: Longman.

Urdang, Laurence. (ed.) 1985. Picturesque Expressions: A Thematic Dictionary. Detroit:


Gale Research Company.

Walter, Elizabeth. (ed.) 1998. Cambridge International Dictionary of Idioms. Cambridge:


Cambridge University Press.

Wehmeier, Sally. (ed.) 2000. Oxford Advanced Learner's Dictionary of Current English.
Oxford: Oxford University Press.

Weinreich, Uriel. 1969. Problems in the analysis of idioms. In: Puhvel, Jaan. (ed.) Substance
and Structure of Language. Berkeley: University of California Press.

Zgusta, Ladislav. 1971. Manual of Lexicography. The Hague: Mouton.

195
8. ABBREVIATIONS

BBI Benson, Morton, Evelyn Benson and Robert Ilson. 1986a. The BBI
Combinatory Dictionary of English: A Guide to Word Combinations.
Amsterdam and Philadelphia: John Benjamins.
BDPF Evans, Ivor Henry. 1989. Brewer’s Dictionary of Phrase and Fable. London:
Cassell.
CDWPO Rees, Nigel. 1996. The Cassell Dictionary of Word and Phrase Origins.
London: Cassell.
CCDI Moon, Rosamund. (ed.) 1995 Collins Cobuild Dictionary of Idioms. London:
HarperCollins Publishers.
CCED Sinclair, John. (ed.) 1995. Collins Cobuild English Dictionary. London:
HarperCollins Publishers.
CIDE Procter, Paul. (ed.) 1995. Cambridge International Dictionary of English.
Cambridge: Cambridge University Press.
CIDI Walter, Elizabeth. (ed.) 1998. Cambridge International Dictionary of Idioms.
Cambridge: Cambridge University Press.
CIDPV Pye, Glennis. (ed.) 1997. Cambridge International Dictionary of Phrasal Verbs.
Cambridge: Cambridge University Press.
DIO Flavell, Linda and Roger Flavell. 1992. Dictionary of Idioms and Their Origins.
London: Kyle Cathie.
EI Seidl, Jennifer and W. McMordie. 1988. English Idioms. Oxford: Oxford
University Press.
ÉKSZ Juhász, József, István Szõke, Gábor O. Nagy and Miklós Kovalovszky. (eds.)
1972. Magyar értelmezõ kéziszótár. Budapest: Akadémiai Kiadó.
LDCE Gadsby, Adam. (ed.) 1995. Longman Dictionary of Contemporary English.
Harlow: Longman Group Ltd.
LDEI Long, Thomas Hill. (ed.) 1979. Longman Dictionary of English Idioms. Harlow
and London: Longman Group Limited.
LID Stern, Karen. (ed.) 1998. Longman Idioms Dictionary. Harlow: Longman.
LPVD Fox, Chris. (ed.) 2000. Longman Phrasal Verbs Dictionary. Harlow: Pearson
Education Limited.
MS Renton, Nick E. 1992. Metaphorically Speaking. New York: Warner Books.
MSZKT Kiss, Gábor. (ed.) 1998. Magyar szókincstár. Budapest: Tinta Könyvkiadó.
NSOED Brown, Lesley. (ed.) 1993. The New Shorter Oxford English Dictionary I-II.
Oxford: Clarendon Press.
OALDa Crowther, Jonathan. (ed.) 1995. Oxford Advanced Learner's Dictionary of

196
Current English. Oxford: Oxford University Press.
OALDb Wehmeier, Sally. (ed.) 2000. Oxford Advanced Learner's Dictionary of Current
English. Oxford: Oxford University Press.
ODEI Cowie, Anthony P., Ronald Mackin and Isabel R. McCaig. 1993. Oxford
Dictionary of English Idioms. Oxford: Oxford University Press.
ODPV Cowie, Anthony P. and Ronald Mackin. 1993. Oxford Dictionary of Phrasal
Verbs. Oxford: Oxford University Press.
PDEI Gulland, Daphne M. and David Hinds-Howell. 1986. The Penguin Dictionary of
English Idioms. Harmondsworth: Penguin Books.
PE Urdang, Laurence. (ed.) 1985. Picturesque Expressions: A Thematic Dictionary.
Detroit: Gale Research Company.
UI Heaton, John Brian and T. W. Noble. 1987. Using Idioms. New York: Prentice
Hall.

197
9. APPENDIX 1

Concordances taken from our corpus can be seen below For reasons of space for
some words/expressions only a selection of the total set of corpus examples is given here.
This is marked by an asterisk (*) at the end of the heading

9.1 fettle
1.ers out here." That Stewart looks in better fettle to open the batting in the Tests is not
2.couse fowl (Jane Horrocks, in fine clucking fettle) who, contemptuously surveys her
3.d and pawed the ground.They were in decent fettle.Santa hitched them up for dry runs
4.edon.She approaches the match in excellent fettle, having beaten Arantxa Sánchez
5.Portuguese and English remains in excellent fettle.Immediately within, and just as
6.ith all their equipment and in good fighting fettle, Kalamata presented a scene of
7.ng claims.John Dunlop's horses are in fine fettle and _Will You Dance_ has sound
8.World No 1 in fine fettle for US Open challenge _Norman happy to sta
9.sex miss Blewett_* + Golf: Faldo in fine fettle for Augusta_* + Hockey: Oxford set
10._Faldo in fine fettle for Augusta_
11.st." Indeed, the new year finds him in fine fettle, upbeat once again about his own,
12.--===The Times: Sport:Faldo in fine fettle for Augusta ===-- [Navigation Controls]
13.Golf: Faldo in fine fettle for Augusta Copyright 1997.
14.ssing more chemicals than ICI seems in fine fettle today.Greenfield was also dining
15.strangely joyless.But Midori was in fine fettle for Kreisler's *Praeludium and Allegro
16.e Crown, Mr Robinson declared he was in fine fettle."If you start worrying about
17.s to keep their existing businesses in good fettle.The former Hanson group and BTR
18.ack in London, both man and bike are in good fettle."The bike looks remarkably
19.m heaped upon your team, you appear in good fettle._MA:_ I am.I have not enjoyed
20.me skipping back into the world in pristine fettle.If, on the other hand, I am too

9.2 kilter*
1.ring his overall game back into competitive kilter was less clear.As he walked towards
2.ear's colour cycle sometimes appears out of kilter with the season.In autumn and
3.estival jigsaw, is thrown gloriously out of kilter today when three of the country's most
4.rosity.Wasps' game-plan was thrown out of kilter by the retirement of Mark Weedon,
5.isting domestic rules would still be out of kilter with the new and harmonised rules in
11.l round.German public finances went out of kilter, partly through reparations in the
15.st-wave entry.The business cycle was out of kilter with the Continent, and sterling
17.oetry in the pens _*+ Opinion - Out of kilter with the Commons_* The position of
21.makes talk of a booming economy seem out of kilter.But Rob Buckland, equity
27.achieving too little.Something is out of kilter at the Oval.Half of the side is
28.--===The Times: Opinion: Out of kilter with the Commons ===-- August 25 1997
31.egy in this race is already seriously out of kilter, bearing in mind that he planned to

198
9.3 umbrage
1.rly *seven years* ago.(I took considerable umbrage at this and advised him, with all
2.Asfordby coal mine, has taken good-humoured umbrage at my suggestion that just
3.nconscious level, however, we take a little umbrage at the claim.This is because none
4.to clarify why he or she has taken such umbrage.The Fault Finder picks at
5.outstanding miler in Europe, taking suitable umbrage at having his tail tweaked by the
6.as a workaholic control freak, quick to take umbrage when her decisions were
7.And while many IT professionals might take umbrage at his comments, he has been
8.g or irresponsible.It is a mistake to take umbrage without reading them.And the
9.understood perfectly officials who had taken umbrage at the British team's reluctance
10.of Death'." Windows, it seems, having taken umbrage at the attempted intrusion, has
11.discovers pleasure in life.The court takes umbrage.Rumours of scandalous intimacy
12.GREG RUSEDSKI may not be alone in taking umbrage at Henmania.As the
13.ranquillisers, imagining insults and taking umbrage when sports officials failed to
14.Many women who had read the male volume took umbrage at his remarks.Some
15.andoned when Diana, Princess of Wales, took umbrage at the contents of the
16.player by 21.On hearing this, Afzaal took umbrage.He told Lewis that he could
17.ave, thrifty Middle England immediately took umbrage at the £50,000 ceiling put on
18.Back in 1960, Manchester City Council took umbrage at the film title *Hell is a

9.4 WORD1 with a capital X, WORD2 with a small Y


1.ld tell Jack Straw a thing or two about life with a capital L, because I served as an
2.with a small p, manner, but with commitment with a capital C.At this rate it could go
3.gian architectural style: "It will be modern with a capital M." _Cycling high_ THE
4.unt Indigo of the Indigo club."It's a show with a capital S.Indigo is made up of short
5.great reservation: that he would be trouble with a capital T.In 1979 he married Lolicia
6.r mad scene was crassly ineffectual acting with a capital A.There are still moments
7.now there has been no local resource for Art with a capital A.No gallery, study centre,
8.Bull used to like saying, sounds like a bet with a capital B._*
9.top in tennis by unrelenting professionalism with a small "p".He was helped by the
10.ceeds.The club is run in the professional, with a small p, manner, but with
11.onal."My sense was that it was conservative with a small 'c'.It looked like a very high-
12.is obviously a liberalish body, isn't it? "With a small l, yes, with a small l," he
13..I'm not afraid of being called a liberal with a small l on these matters, I've always
14.e friend describes her as "Welsh Nationalist with a small n", which might make for
15.alish body, isn't it? "With a small l, yes, with a small l," he says."But it is a broad
16.Kerr dismissed Tony Blair as a "conservative with a small c" who was determined to

9.5 catch + ship/boat/taxi


1.classes, pausing over a brandy before they catch a taxi, agree that the great era of the
2.ch offers splendid views of the city, is to catch the airport shuttle, a seven-minute boat
3.ent.Frolic for the evening, but be sure to catch the boat back to paradise.*_SOPHIE
4.should not be taken abroad, but the father caught a boat to Holland and then a flight to
5.o, looking to kill a few hours in Sydney, I caught a taxi across town to Coogee, to
6.s last seen on August 15 last year, when she caught a taxi to Walsall.There she was
7.million pounds.I reached Knightsbridge and caught a taxi to Smith Square, then sought
8.ended up at Brighton College.Their parents caught the ship home.Abraham managed

199
9.6 swallow + N*
1.dership was so swift and sweeping that Brown swallowed his ambition and never even
2.PROVIDED the voters of Tunbridge Wells swallow their *amour propre* and elect a
3.t always right: it was cruel to make people swallow their deepest feelings if their
4.person who has died before we breathe in to swallow any fear the person may be
6.man who knows how to deal with emotions.He swallows them, just as every red-
7.n county cricket match.As Knight and Hemp swallowed their pride to help themselves
9., sooner or later, you are going to have to swallow your pride and talk business with
10.nd to say, 'That's what I'm going to do' and swallow your pride and get on with it."
12.nths of procrastination, Roy Hattersley has swallowed his pride and decided to take
14.an businessmen also hope Iran's regime will swallow its pride and do nothing more to
17.rtisements." Later in the day, the company swallowed its reservations and contacted

9.7 broad/wide + choice/range*


1.s just fantastic." Given the competition's broad age range, how will Söderström and
2.s for sale for £1.75 million.HOWEVER, the broad architectural range is very
3.and grow.Events during the week offer a broad choice of venues including London,
4.tial of providing the viewing public with a broad choice of TV channels; at the time,
5.nefit from their strong market positions and broad product range throughout the
6.nt.He commands the respect and support of a broad range of Conservative opinion
7.PEAKS_ _The new Bank advisers must bring a broad range of views_ _* Gordon
10.K and Asia Pacific, focused on a relatively broad range of services to industry and
30.years, possibily this year.We have a very broad range of products - over 3,000 in one
40.ector, said: "The level of awareness of the broad range of finance options available
50.will enable all family doctors to provide a broader range of patient services within
66.d in Britain for the next century, from the broadest range of sectors as possible," she
67.[Line] A wide choice exists for those wanting a rural lifestyle w
70.said.The results give President Zeroual a wide choice to select a prime minister from
80.ocking should ensure that consumers enjoy a wide choice in the January sales.The
90.ns from £50.Income seekers have an equally wide choice.At one end of the range are
96.able polychromatic self.The gallery owns a wide range of Victorian painting, from
97.new drugs are also better at controlling a wide range of psychotic symptoms." _DR
98.d 200 paintings, pastels and sculpture on a wide range of subjects._Christie's_, 8
99.at, for every modern city woman, there is a wide range of Miss Moneypenny pencil
100._Q: I use Word Perfect 7 which comes with a wide range of fonts.Although I tend
110.m viability of some trusts."Boards have a wide range of options available to them,
120.Dimbleby to answers viewers' questions on a wide range of political issues (T)
124.logy businesses, serving customers across a wide range of industries including
automotive and aeros

9.8 dudgeon
1.gley Test? Caddick left the ground in high dudgeon, as well he might.Against the
2.forward to Martin.Smith also left in high dudgeon when he was given out leg-before
3.Agassi, one dinner guest said, left in high dudgeon.There could be no excuse from
4.d, recently left Sheffield Wednesday in high dudgeon claiming David Pleat, the
5.taken their children away.One left in high dudgeon, and I am afraid to say there are

200
6.go home in disgrace, or flounce off in high dudgeon, or limp off in a beany-hat?
7.shurst Technology is now withdrawing in high dudgeon, irked by the small-minded,
8.where the second Test starts today, in high dudgeon.Clive Lloyd, their manager, who
9.departure of Cantona, in ill-concealed high dudgeon, at the end of last season, United
10.for a further three years but left in some dudgeon after Esser was replaced and a new
11.ho once telephoned a younger don in tactical dudgeon to demand that he hand over

201
10. APPENDIX 2

Could you answer the questions below? The data will be used in my PhD
dissertation. I am interested in native speaker intuition! Please do not look at section B
until you have completed A!

Attila Cserép

The data supplier is a native speaker of British/American/Australian English. (Please delete


as appropriate!)

Section A
Please fill in the sentences with an appropriate word!
1. Political sex scandals are all ......................... to the mill of the tabloid newspapers.
2. If John is going to the States, presumably that'll put the ......................... on his
relationship with Eva.
3. He was one of three children injured when Horrett Campbell, a paranoid schizophrenic,
ran ......................... with a 2ft machete in July.
4. If you don't do your research, you're likely to come a .........................
5. Don't expect to have me at your ......................... and call.
6. Since the departure of Cantona, in ill-concealed high ........................ at the end of last
season, United have been linked with more than half a dozen players.
7. It is all very well for a right-wing party, in ........................ with local reactionaries, to
find seats for their placemen.
8. As an American, she had no ......................... with the painful formality of English life.
9. Owen wondered what Roz was planning; he didn't want her .........................ing his pitch
in front of the area manager.
10. Horowitz, who'd been raised in Hungary, was no ......................... at complex
communication and surveillance systems.

Section B

202
If you think that the underlined parts in the following sentences have meanings of their own,
supply those meanings! You can answer by supplying a single word (eg. good), by giving
(near-)synonyms (eg. good, fine), or by giving a short phrase (eg. the good things). If you
think that the underlined parts do not have meanings in these sentences, cross out the space!

1. By dropping a sprinkling of the right names and enough technical jargon I was able to
pick up secret files left lying on desks. MEANING:
2. It had been over a year since the incident and I thought it was time we buried the hatchet.
MEANING:
3. Helen thought management would tell her to pack her bags but instead they offered her a
new contract. MEANING:
4. They put the screws on him until eventually he was forced to resign. MEANING:
5. Yesterday he was publicly criticized for not doing enough to mend fences with his big
political rival. MEANING:
6. He let it drop that the Prime Minister was a close friend of his. MEANING:
7. He has always carried a torch for Barbara. MEANING:
8. She was back in a trice. MEANING:
9. Advertisers can turn on a dime and produce new commercials in a couple of hours.
MEANING:
10. The boxer beat the tar out of his opponent. MEANING:
11. People who work in television should keep their feet firmly on the ground. MEANING:
12. Airbus will soon be passing the hat around again for an enormous 700-seat aeroplane.
MEANING:
13. I've loved him ever since I first set eyes on him. MEANING:
14. Steve beat the daylights out of him with a hefty length of bike chain. MEANING:
15. Plastics recycling was not done much in this country until some of the big supermarkets
set the ball rolling. MEANING:
16. He won the Nobel Peace Prize, but this cut little ice at home. MEANING:
17. If something gets lost, damaged or stolen, you bite the bullet and cover the cost
yourself. MEANING:
18. He wanted to get into medical school but he failed to make the grade. MEANING:

203
19. Since the departure of Cantona, in ill-concealed high dudgeon at the end of last season,
United have been linked with more than half a dozen players. MEANING:
20. We would have liked to invite all our relatives, but you have to draw the line
somewhere. MEANING:
21. I want to bury the hatchet. I still love her. MEANING:
22. At this I lost my cool and shouted 'for goodness sake, stop!' MEANING:
23. Tony is sure he can pull a few strings and get you in. MEANING:
24. I'm a bit reluctant to show my hand at this stage in the proceedings. MEANING:
25. She decided to go for broke and pursue her acting career full-time. MEANING:
26. Political sex scandals are all grist to the mill of the tabloid newspapers. MEANING:
27. If John is going to the States, presumably that'll put the kibosh on his relationship with
Eva. MEANING:
28. Horowitz, who'd been raised in Hungary, was no slouch at complex communication and
surveillance systems. MEANING:
29. He accused Mr MacGregor of 'sleight of hand'. MEANING:
30. If you don't do your research, you're likely to come a cropper. MEANING:
31. The business cycle was out of kilter with the Continent and sterling was overvalued.
MEANING:
32. It is all very well for a right-wing party, in cahoots with local reactionaries, to find seats
for their placemen. MEANING:
33. He was one of three children injured when Horrett Campbell, a paranoid schizophrenic,
ran amok with a 2ft machete in July. MEANING:
34. They wanted to keep alive the memory of their kith and kin who had died in the war.
MEANING:
35. Don't expect to have me at your beck and call. MEANING:
36. Owen wondered what Roz was planning; he didn't want her queering his pitch in front
of the area manager. MEANING:
37. As an American, she had no truck with the painful formality of English life. MEANING:
38. Joe decided he had spent enough time crying for the moon. MEANING:
39. There are always some risks in breaking the mould and introducing a new style to a
long-established publication. MEANING:

204
40. If I ever allowed a nurse or a technician to work alongside me without wearing gloves,
I'd be called on the carpet immediately for not protecting our staff. MEANING:
41. The race organizers cover their backs by saying they can't take responsibility for any
injuries. MEANING:
42. I'll tell your father and he'll give you what for. MEANING:
43. It's clear the film director was having a shot at the government. MEANING:
44. Their campaign was taking pot shots at Clinton's personal life. MEANING:
45. We'll just have to roll with those punches, but they're not too serious. MEANING:
46. I thought I'd made a sensible suggestion, but they just shot me down in flames.
MEANING:
47. He clearly believes the Americans invented Christmas; you can’t trust people who eat
their turkey in November. But he may have been right about who invented the whole
shebang. MEANING:

Section C
Mark the underlined idioms using the following scale! If the idiom is highly motivated for
you, mark it with the number 1, if it is partially motivated, mark it with 2, and if it is not
motivated write 3 next to the idiom. "Highly motivated" means that you can predict the
meaning using real-world knowledge, as in behind sb's back, scratch one's head, or pack
one's bags, in which the literal meanings are closely related to the idiomatic interpretation.
"Partially motivated" means that the literal meaning is related to the idiomatic interpretation,
and you can more or less guess the meaning but some specialist knowledge is required to do
that, as in throw in the towel or grasp the nettle. "Unmotivated" means that the literal and
idiomatic interpretations are unrelated, as in the bee's knees. All the idioms below have
meanings related to criticizing.
1. backbiting
2. be/get on sb's back
3. come down on sb
4. curtain lectures
5. jump down sb's throat
6. lay into sb
7. look who is talking

205
8. pin sb's ears back
9. pot shot
10. put sb down
11. rap on the knuckles
12. read sb the riot act
13. a sitting duck
14. take sb to task

206
11. APPENDIX 3

Native speakers’ answers to the tasks given in Appendix 2 are shown below. A total
of 9 speakers were asked, of whom 6 were British, 2 were American and 1 was Canadian.

11.1 Native speaker cloze test answers

Unique word Number of Other words given in the same context


right guesses (if a word is given by two speakers, this is shown as "2x")
(out of 9)
grist 7 fodder
kibosh 4 lid; mockers; quash; brake; icing
amok 9
cropper 9
beck 9
dudgeon 5 drama; spirits; jinx
cahoots 6 tandem; dealings; collusion; league
truck 4 affinity; sympathy; experience; empathy
queer 8 steal-2x;
slouch 0 amateur; mug; wizard; stranger; novice; expert; shakes;
great shakes

11.2 Native speaker judgement of analyzability

The table shows whether the underlined parts of the following expressions are
judged by native speaker informants to have an independent meaning () or not (). The
meaning paraphrases given by the native speakers are also shown. If a particular meaning is
given by more than one speaker, this is shown as "2x", "3x", etc. Where it is clear that the
speaker's paraphrase is that of the verb, I have given the paraphrase in the bare infinitive
form, otherwise I have kept the original paraphrase.

Expression   Meaning paraphrase


given by native speakers
dropping ... the right names 9 let fall; leave; give; mention casually; mention-

207
2x; remove; let go; use casually
buried the hatchet 5 4 get rid of; make up (the quarrel); throw away;
dispose of;
to pack her bags 7 2 fire her; get ready;
put the screws on him 2 7 intense pressure and victimization; force;
pressure-5x
mend fences 2 7 differences-2x; links; quarrel; make peace; his
differences; broken agreements and boundaries
let it drop 2 7 be known unexpectedly; be known-4x; be
rumoured; become known;
carried a torch 6 3 been infatuated with; hold; have;
in a trice 1 8 brief moment, without delay; quickly-3x; a very
short time; short time; instant; a moment
turn on a dime 6 3 a small area; quickly; suddenly
beat the tar 5 4 brutally; the fight-2x; life
keep their feet firmly on the 8 1 focus
ground
passing the hat around 5 4 begging bowl; the begging bowl; collecting;
begging
set eyes on him 6 3 my vision; see-2x;
beat the daylights 3 6 normal consciousness and awareness;
consciousness; very badly; life-2x, the life
set the ball rolling 6 3 the movement; start; the process
cut little ice 6 3 attention; mean little; not impress
bite the bullet 7 2 take the consequences; take
make the grade 2 7 the acceptable standard; passing point; necessay
mark; the required level; the standard; the
standard needed; the cut
in ill-concealed high dudgeon 2 7 drama; bad temper; fury; temper, tantrum;
anger; spirit; a fit of pique or anger
draw the line 3 6 the boundary; a halt; decide; limit; the limit; the
stopping point
bury the hatchet 4 5 bad feeling; ill-feeling;make up-2x; the anger
lost my cool 9 sanguinity, calm; patience-2x; reason; temper;
equilibrium; composure-3x; self-control
pull a few strings 7 2 manipulate things; levers of influence
show my hand 3 6 position-2x, advantage, ploy; strengths; cards;
strength; intention
go for broke 8 1 devote herself

208
all grist to the mill 6 3 raw material; corn; to be ground up
put the kibosh on his 3 6 the mockers; end-2x; a finish; the end; jinx, final
relationship blow
no slouch 9 slowcoach, idiot, ?innocent; novice-2x;
incompetent; dunce; fool; slow; inept person;
lazy person
sleight of hand 4 5 deceit; cheating; deceitful; trickery; tricks
come a cropper 3 6 to disaster; loser; crash; fall; victim; unstuck
out of kilter 1 8 sync; synchrony; sychronicity; not in line with;
sequence; balance-2x; synchronization
in cahoots with local 1 8 well in with, thick as thieves; agreement,
reactionaries partnership; league-2x; partnerships; conspiring
with; close collaboration; shady negotiations
ran amok 9 wild-4x; riot; (really) wild; mad; around wildly;
crazy; crazy in a violent way
their kith and kin 6 3 relatives and family; relations; relatives
at your beck and call 5 4 summons; immediate; gesture; command
queering his pitch 2 7 muck up; ruin-2x; mess; mess up-2x; question
had no truck 4 5 time for; patience-3x; sympathy, taste for;
willingness to accept
crying for the moon 6 3 complete success; everything; the impossible
breaking the mould 1 8 the mode; the traditions; the norm; existing
pattern; the established style; rules; the tradition;
the pattern, prior practices
called on the carpet 4 5 for reprimand; to task; to defend myself against
superiors; disciplined; to account
cover their backs 3 6 protect-5x; insure themselves
give you what for 3 6 gyp, punishment; some punishment; a scolding;
a telling off-2x; punish
having a shot 2 8 critical-2x; a go-2x; a dig; an attack; blaming,
attacking; taking an opportunity to be critical
taking pot shots 6 3 well-aimed; aim; easy
roll with those punches 2 7 hits; criticisms-2x; attacks-2x; blows; problems
shot me down in flames 3 6 destroy my position; pull it apart; put me down;
criticized-2x; ridicule
the whole shebang 1 8 business; thing-4x; combination; event;
situation; caboodle, shooting match, affair or
concern, occasion or activity

209
11.3 Native speaker judgement of motivation

If the speaker did not indicate the motivation, we have crossed out the slot in the
table below. Highly motivated = 1, partially motivated = 2, unmotivated = 3; NS 1 = native
speaker 1, NS 2 = native speaker 2, etc.

NS 1 NS 2 NS 3 NS 4 NS 5 NS 6 NS 7 NS 8 NS 9 Aver
age
bacbiting 2 1 2 1 1 2 1 2 1 1.4
be/get on sb's back 2 1 1 1 2 2 1 2 1 1.4
come down on sb 2 2 1 2 2 3 2 2 2 2
curtain lectures - 3 3 2 3 - - - 3 2.8
jump down sb's throat 2 2 1 2 2 2 3 2 3 2.1
lay into sb 3 2 2 1 2 2 1 1 3 1.8
look who is talking 1 1 2 1 1 1 1 1 3 1.3
pin sb's ears back 3 1 2 2 2 2 3 2 2 2.1
pot shot 3 3 3 3 3 2 1 2 1 2.3
put sb down 2 3 1 2 2 3 2 3 2 2.2
rap on the knuckles 2 1 2 1 2 2 1 2 1 1.5
read sb the riot act 2 2 2 1 2 3 2 2 1 1.8
a sitting duck 2 2 3 1 3 2 3 2 1 2.1
take sb to task 2 2 2 2 2 3 2 - 3 2.2

210
12. APPENDIX 4

211
13. APPENDIX 5

13.1 Idioms of criticizing


For the lexical variants of variable idioms see 4.2.5.3.

armchair general attract flak


Aunt Sally backbiting
back-seat driver be a sitting shot
call sb on the carpet cast aspersions
cast sth in sb's teeth come down on sb like a ton of bricks
come under fire cover one's back
curtain lecture damn sb with faint praise
dip one's pen in gall Don't come the uncle over me
fire a salvo get a lashing
get a lot of stick get a rap on the knuckles
get a roasting get a rocket
get on sb's back get on sb's case
give sb a lecture give sb down the banks
give sb grief give sb Jesse
give sb the rough side of one's tongue give sb what for
haul sb over the coals have a go
have a shot have a thick skin
have one's ears slapped back have the hide of a rhinoceros
honeymoon in the firing line
jump down sb's throat kale through the reek
lash of scorpions lay out in lavender
leave oneself open like water off a duck's back
look who's talking Monday-morning quarterback
nitpick not pull one's punches
peanut gallery people in glass houses shouldn't throw stones
pin sb's ears back pot shot
put the boot into sb rap sb on the knuckles
read sb the riot act roll with the punches
run the gauntlet shoot sb down in flames
sitting duck stand in the breach
stop-watch (critic) take a hammering
take it on the chin take sb to task
talk to sb like a Dutch uncle tear a strip off sb
tear sb to pieces tell sb where to get off
the pot calling the kettle black throw sb to the wolves
tongue-lashing turn one's guns on sb

212
bawl out sb chew out sb come down on sb
do down sb dress down sb dump on sb
go for sb hit back hit out
jump on sb lash into sb lash out
lay into sb put down sb rip into sb
run down sb shoot down sb slag off sb
slap down sb tear into sb tell off sb
tick off sb

13.2 Collocations of criticizing

accept criticism adverse criticism


armchair critic arouse criticism
attract criticism barrage of criticisms
biting criticism blistering attack
blistering criticism criticize sb
harshly/heavily/savagely/severely/sharply,
etc.
counter criticism damaging criticism
devastating criticism draw criticism
face criticism fierce critic
harsh criticism heap criticism
heavy criticism hostile criticism
level criticism at sb open to criticism
provoke criticism reject criticism
savage attack savage criticism
scathing criticism sharp criticism
stir up criticism strong criticism
subject sb to criticism sweeping criticism
take criticism temper one's criticism
trenchant criticism unsparing criticism
widespread criticism withering criticism

13.3 Figurative single-word (near-)synonyms of criticize and criticism

assail sb assault assault sb


attack attack sb bash sb
basher bashing barb
belabour sb belittle sb blast

213
blast sb carpet sb claw sb
cover oneself damn sb deflate sb
fire flak flame sb
flay sb hammer sb hammering
knock sb knocker lash sb
lashing lecture lecture sb
onslaught rap rap sb
roast sb roasting rubbish sb
savage sb slam sb snipe at sb
target target sb trash sb

214
14. APPENDIX 6

In the table below "x" marks the occurrence of the example (or a related form of the
example) in the given dictionary. If an example is found in ODPV, it is marked "x1", if it is
found in ODEI, it is marked "x2". Nonoccurrence is marked by leaving the slot empty.

C C OD L P E U C O L C
C I PV, I D I I C A D I
D D OD D E E L C D
I I EI I D D E E
b
armchair general
armchair critic-PDEI 150, CCED 78, OALDb 54, LDCE 59, EI 63, armchair traveller-PDEI 150, CCED
78, OALDb 54, LDCE 59, CIDE 63, armchair fan-CCED 78, armchair gardener-CIDE 63, armchair
cricketer-CIDE 63, armchair expert-EI 63, armchair travelling-EI 63
attract flak x x x x x x
get/take (the) flak-CIDI 140, get a lot of flak-CIDI 140, CCED 639, catch/take/get (a lot of) flak/flack-
LID 124, take the flak-CCED 639, LID 124, take a lot of flak-OALDb 483, LDCE 530, come in for a lot
of flak-OALDb 483, get/take flak-LDCE 530, take/catch some flak/run into/come into a lot of flak-
CIDE 529, they have attracted more than their fair share of flak-CCED 639
Aunt Sally x2 x x x
Aunt Sallies-ODEI 41, LID 10, the Aunt Sally of-LID 10, an Aunt Sally-ODEI 41,PDEI 133, EI 58
backbiting x x x x
backbiter-LDCE 79
back-seat driver x x2 x x x x x x
back-seat drivers-ODEI 42, PDEI 141, EI 64, CCED 108, back-seat driving-ODEI 42
be a sitting shot
sitting duck x x x2 x x x x x x x
a sitting duck-CCDI 113, CIDI 353, ODEI 503, PDEI 68, EI 200, UI 12, CCED 1559, sitting duck-
OALDb 1204, LDCE 1339, sitting-duck + N-CCDI 113, sitting ducks-CIDI 353, LID 95, CCED 1559,
LDCE 1339, a sitting target-ODEI 503, sitting targets-ODEI 503, that sitting target-ODEI 503, be a
sitting duck-LID 95
call sb on the x x x x x x
carpet
on the carpet-CCDI 60, EI 94, OALDb 178, CIDE 198, be called on the carpet-CCDI 60, OALDb 178,
get called on the carpet- LID 52, OALDb 178, call sb on the carpet-LID 52, put sb on the carpet-UI 37,
be carpeted-PDEI 151
cast aspersions x x x x x
cast sth in sb's x
teeth
cast in the/sb's teeth-PDEI 88
come down on sb x x x2 x x x x x x x
like a ton of bricks
come down on sb like a ton of bricks -CCDI 394, CIDI 397, LID 355, PDEI 144, EI 239, CCED 1763,
LDCE 1522, CIDE 266, ODPV 64, come down on sb + like a ton of bricks ODEI 357, OALDb 1368,
CIDE 1534, CIDPV 52, LPVD 88, be down on sb like a ton of bricks-CCDI 394, CIDI 397, ODEI 357,
have sb down on you like a ton of bricks-CIDI 397

215
come under fire x x x1 x x x x x x
be under fire-CCDI 144, LID 121, CCED 631, OALDb 477, come under fire-CCDI 144, CIDI 137, ODPV
72, LID 121, UI 37, CCED 631, OALDb 477, under fire-LDCE 522, CIDE 523
cover one's x x x x
back/rear
cover one's back-CCDI 85, Br & Am & Aus CIDI 82, LID 12, LDCE 317, cover one's rear-CCDI 85,
cover one's ass-CCDI 85, Am & Aus CIDI 82, LID 8, cover one's butt LID 8, plural subject + cover their
backs-CIDI 82, LID 12
curtain lecture x2 x
suffer a curtain lecture -PDEI 151
damn sb with faint x x x2 x x x x x
praise
damn sb with faint praise -CCDI 308, CIDI 125, ODEI 130, LID 271, OALDb 315, LDCE 343, CIDE
344, faint praise-CCDI 308, passive: sb is damned by faint praise-CCDI 308, damned with faint praise-
ODEI 130, damning with faint praise the mediation efforts-CCED 1289
dip one's pen in x
gall
Don't come the
uncle over me
draw fire x x x2 x x x
draw sb's fire-CCDI 144, ODEI 157, OALDb 382, draw (sb's) fire-CIDI 137, draw fire-CCDI 144,
CCED 630, draw the fire of sb-CCDI 144, draw fire from sb-CCDI 144, CIDI 137, CCED 630, LDCE
414
fire a salvo
get a lashing x x x
the lashings he got-CCED 934, he was given a severe tongue lashing-OALDb 723, gave him a verbal
lashing-CIDE 797
get a lot of stick x x x2 x x x x x
get a lot of/a bit of/etc stick-CCDI 368, CIDI 370, ODEI 534, LID 326, CIDE 1422, get stick-LID 326,
CIDE 1422, take a lot o/some/etc stick-CIDI 370, ODEI 534, take stick-CIDE 1422, come in for a lot
of/some/etc stick-CIDI 370, CIDE 1422 give sb stick-CCDI 368, ODEI 534, LID 326, LDCE 1411, CIDE
1422, give sb a lot of stick-CCDI 368, give sb a lot of/some/etc stick-CIDI 370, ODEI 534, LID 326,
CCED 1636, a stick to beat sb with-CIDI 370, LID 327, ODEI 482, OALDb 95, a stick to beat sb-CCDI
369, a stick with which to beat sb-CCDI 369, ODEI 482, a rod to beat sb with-ODEI 482, OALDb 95,
sticks to...-ODEI 482
get a rap on the x x x x x x x
knuckles
rap sb on the knuckles-CCDI 232, CCED 1360, OALDb 1048, rap sb over the knuckles- CCDI 232,
OALDb 1048, LDCE 1171, rap sb's knuckles-CCDI 232, CCED 1360, OALDb 1048, passive: be rapped
on the knuckles-CCDI 232, was rapped over the knuckles-CCED 1360, OALDb 1048, LDCE 1171, get
rapped over the knuckles-CIDE 1173, have one's knuckles rapped-CCDI 232, LID 281, get one's
knuckles rapped-LID 281, CIDE 1173, (give sb/get) a rap on/over/across the knuckles-OALDb 1048,
get/receive a rap on the knuckles-CCDI 232, got/was given a rap on/over the knuckles-CIDE 1173,
received an official rap over/on the knuckles-LDCE 1171, need a rap across the knuckles-LID 281, give
sb/be given a rap on the knuckles-PDEI 99, earned him a rap on the knuckles-CCED 1360, gave them a
diplomatic rap over the kuckles-CCED 1360, be given a rap on/over/across the knuckles-LID 281, take
the rap-PDEI 99
get a roasting x x x x
give sb a roasting-CCED 1439, OALDb 1107, LDCE 1232, get a roasting-OALDb 1107, I got a real
roasting from..., be given a roasting-CIDE 1229, a roasting from...-CCED 1439
get a rocket x x x x x x
give sb a rocket-CIDI 328, LID 287, OALDb 1108, LDCE 1233, CIDE 1230, get a rocket-CIDI 328, LID
287, PDEI 233, OALDb 1108,CIDE 1230
get on sb's back x x x1 x x x x x x x x
be on sb's back-CCDI 12, CIDI 16, LID 12, get off sb's back-CCDI 12, CIDI 16, PDEI 100, EI 213, UI
71, CCED 106, CIDE 89, get sb off one's back-CCDI 12, CIDI 16, LID 12 get on sb's back-ODPV 139,

216
keep sb off one's back-LID 12, get off my back-LID 12, LDCE 78, if you'd just get off my back-LID 12
get on sb's case x x x x x
be on sb's case-CCDI 61, CIDI 60, LID 53, OALDb 180, LDCE 197, get on sb's case-CCDI 61, CIDI 60,
LID 53, get off sb's case-CCDI 61, CIDI 60, get off my case-LID 53, OALDb 180, LDCE 197
give sb a lecture x x x
gave us a stern lecture-CCED 948
give sb down the
banks
give sb grief x x x x x
give sb grief-CIDI 163, LID 143, OALDb 565, LDCE 625, CIDE 623, not give me any grief-CIDI 163,
gave me a lot of grief-CIDE 623, get grief-CIDI 163, get a load of grief-CIDI 163
give sb Jesse
give sb the rough x x x2 x x x x
side of one’s
tongue
give sb the rough side of one's tongue-CCDI 395, CIDI 331, LDCE 1239, CIDE 1236, give sb the rough
edge of one's tongue-CCDI 395, give sb the rough edge of one's tongue-PDEI 89, give sb/get the (rough)
edge of one's/sb's tongue-ODEI 225, get the rough side/edge of sb's tongue-LID 308
give sb the length
of one's tongue
give sb what for x x2 x x x x
I'll give you what for-CIDI 154, give sb/get what for-ODEI 232, OALDb 1473, give sb what for-UI 37,
LDCE 598, CIDE 596
haul sb over the x x x1 x x x x x x
coals
haul sb over the coals-BrE CCDI 76, CIDI 74, ODPV 173, BrE LID 65, PDEI 224, CCED 300, BrE
OALDb 592, LDCE 250, CIDE 251, rake sb over the coals-CCDI 76, AmE LID 65, AmE OALDb 592,
LDCE 250, drag sb over the coals-CIDI 74, CCED 300, LDCE 250, CIDE 251, passive: sb is hauled over
the coals-CCDI 76, CIDI 74, ODPV 173, LID 65 only passive examples, PDEI 224 passive example, CCED
300 only passive examples, CIDE 251 only passive examples
have a go x x x x x x
have a go at sb-CCDI 164, CIDI 156, LID 138, CCED 719, OALDb 551, CIDE 606, had a real go at me-
LID 138
have a shot x
have a thick skin x x x2 x x x x x x x
have/develop/grow etc a thick skin-ODEI 276, have a thick skin-BrE LID 312, PDEI 108, EI 25, LDCE
1497, have thick skin-AmE LID 312, grow a thick skin-EI 25, need a thick skin-CCDI 355, CCED 1562,
develop a thick skin-CIDI 387, thick-skinned-CCDI 355, CIDI 387, ODEI 276, LID 312, PDEI 108, EI
25, OALDb 1348, LDCE 1497, CIDE 1511, thickest-skinned-ODEI 276, My skin is thick/I've got thick
skin-CIDE 1346
have a thin skin x x x2 x x x x
have/develop/grow etc a thin skin-ODEI 276, a thin skin-CCDI 355, have got a remarkably thin skin
CIDI 389, have such a thin skin-OALDb 1349, thin-skinned-CCDI 355, CIDI 389, ODEI 276, PDEI 108,
OALDb 1351, LDCE 1501, CIDE 1511
have one's ears
slapped back
have the hide of a x2 x
rhinoceros
have/need/with etc a hide/skin like a rhinoceros, have hides like a rhinoceros, needed the skin of a
rhinoceros her hide is as thick as a rhinoceros-ODEI 264, have (got) a hide like a rhinoceros-EI 240
hold fire x x2
hold fire-CCDI 144, hold one's fire-CCDI 144, ODEI 285
honeymoon x x x x x
honeymoon-LDCE 686, CIDE 681 a honeymoon period-CIDI 196,a/the honeymoon period LID 172,
(LDCE 686),the CIDE 681, the honeymoon is over-LID 172, PDEI 159, honeymoon period [collocation]-
OALDb 623, CIDE 681, Labour's brief honeymoon period-CIDI 196

217
in the firing line x x x x x x x x x
in the firing line-CCDI 244, UI 56, CCED 632, CIDE 523, on the firing line-Am CIDE 523, in the line of
fire-CCDI 244, EI 28, be in the firing line-Br, Am & Aus CIDI 138, Br LID 121, Br OALDb 478, LDCE
828, be on the firing line Am & Aus-CIDI 138, Am LID 121, Am OALDb 478, be in the line of fire-
LID121, LDCE 828, out of the firing line-CCDI 244, CIDI 138, CCED 632, be out of the firing line-LIDE
121, LDCE 828, be out of the line of fire/be out of the line of the critics' fire-LID 121
jump down sb's x x x1 x x x x x x
throat
kale through the
reek
lash of scorpions
lay out in lavender
leave oneself open x x x2 x x x x x
leave oneself wide open-CCDI 420, LID 205, leave oneself open-LID 205, lay oneself wide open-CCDI
420, LID 205, UI 104, LDCE 991, lay oneself open-LID 205, UI 104, CCED 941, LDCE 991, lay
oneself/sb (wide) open-ODEI 343, lay sb open-CCED 941, be (wide) open-CIDI 284 leave them wide
open, lay yourself open to attack-CIDI 284, laid himself wide open-OALDb 887 not idiom
like water off a x x x2 x x x x x x x x
duck's back
like water off a duck's back-CCDI 412, ODEI 357, be (like) water off a duck's back-CIDI 415, LDCE
1614, sth is (like) water off a duck's back-LID 372, (like) water off a duck's back-PDEI 69, UI 128,
CCED 514, OALDb 1460, CIDE 1642, (be) like water off a duck's back-EI 239
look who is talking x x x x x
look who is talking-CIDI 383, LID 217, OALDB 1327, LDCE 1470, CIDE 1488, you're a (fine) one to
talk-LID 217, you're a fine one to talk-CIDI 135, OALDb 1327, LDCE 1470, CIDE 1488, you can talk-
LID 217, CIDI 383, OALDb 1327, LDCE 1470, CIDE 1488, you can't talk-LID 217, OALDb 1327, CIDE
1488, you should talk-CIDI 383
a Monday- x x x x x
morning
quarterback
a Monday-morning quarterback-CCDI 314, CIDI 258, Monday morning quarterback-LID 277, LDCE
1158, CIDE 913, Monday-morning quarterbacking-CCDI 314,...quarterbacks-CCDI 314, LID 277,
the...quarterback-LID 277
nitpick x x x x x
nit-picking (n)-CCED 1116, OALDb 859, LDCE 957, CIDE 956, nitpicking (adj)- nitpicking people PDEI
75, nitpicking detail CCED 1116, OALDb 859, LDCE 957, nitpicking attitude CIDE 956, nit-picker-
OALDb 859, LDCE 957, CIDE 956, nitpick-CIDE 956
not pull one's x x x2 x x x x x x x x
punches
not pull one's punches-CCDI 312, UI 37, CCED 1334, CIDE 1144, pull no punches-CCDI 312, ODEI
469, LID 275, not pull any punches-CCDI 312, CIDI 311, ODEI 469, LID 275, UI 37, LDCE 1145, pull
few punches-ODEI 469, passive: no punches had been pulled-ODEI 469, pull one's punches-ODEI 469,
PDEI 250, EI 175, usually used in neg sentences OALDb 1205, never pulling her punches-CCED 1334,
never pulls any punches-CIDE 1144
peanut gallery
people who live in x x x2 x x x
glass houses
shouldn't throw
stones
people who live in glass houses shouldn't throw stones-CCDI 163, ODEI 453, people who live in glass
houses (shouldn't throw stones)-CIDI 295, LID 177, people (who live) in glass houses shouldn't throw
stones-OALDb 937, people in glass houses shouldn't throw stones-LDCE 601, people in glass houses
really shouldn't throw stones-CCDI 163, people who live in glass houses should be careful about their
stones-CCDI 163, people who live in glass houses should avoid throwing stones-LID 177, aim a brick
straight through their own precious glass house-ODEI 453
pin sb's ears back x
pin sb's ears back-AmE CCDI 117

218
pot shot x x x x
take a pot shot-CIDI 307, a pot shot-CCED 1285, taking pot shots-CCED 1285, took constant potshots-
OALDb 987, potshots-CIDE 1103
put the boot into x x x
sb
put the boot into sb-CCDI 43, put the boot in-CCDI 43, LID 37, OALDb 133, stick the boot into sb-
CCDI 43, stick the boot in-OALDb 133
read sb the riot act x x x2 x x x x x x
read the riot act to sb-CCDI 323, ODEI 476, CCED 1435, CIDE 1178, read (sb) the riot act-CIDI 326,
LID 286, PDEI 231, OALDb 1053, LDCE 1228, read sb the riot act-CCED 1435, passive: the riot act is
read-ODEI 476
roll with the x x x x
punches
roll with the punches-CCDI 312, CIDI 311, LID 275, roll with those punches-LID 275
run the gauntlet x x2 x x x x x
run the gauntlet-CIDI 153, ODEI 485, LID 134, EI 178, CCED 699, OALDb 532, LDCE 585, run a far
severe gauntlet than-ODEI 486
shoot sb down in x x x x
flames
shoot sb down in flames-CCDI 148, CIDI 140, LDCE1316, passive: be shot down in flames-CCDI 148,
CIDI 140, LID 124, LDCE 1316
stand in the
breach
stop-watch (critic)
take a hammering x2 x x
take a beating/pounding/hammering/pummelling/thrashing-ODEI 530, taken quite a beating-ODEI 530
take a beating/hammering-LID 21, take a hammering-CCED 758, come in for a bit of a hammering-
CIDE 639, took quite a pounding-CIDE 1104
take it on the chin x x x
take it on the chin-CCDI 69, CIDI 66, LID 60, take criticism etc on the chin-CCDI 69, took it all on the
chin-CIDI 66, take sth on the chin-LID 60
take sb to task x x1 x x x x x x x
take sb to task-CIDI 384, ODPV 378, LID 340, EI 180, UI 37, CCED 1709, OALDb 1330, LDCE 1474,
CIDE 1492, passive: the Minister was taken severely to task-ODPV 378, I didn't expect to be taken to
task for it-EI 180, The country's intellectuals are also being taken to task for their failure-CCED 1709
talk to sb like a x x
Dutch uncle
talk to sb like a Dutch uncle-PDEI 189, talk (to sb) like a Dutch uncle-LDCE 428
tear a strip off sb x x x1 x x x x x x
tear a strip off sb-CCDI 375, CIDI 376, ODPV 382, LID 331, UI 37, CCED 1655, OALDb 1333, LDCE
1478, CIDE 1495, tear sb off a strip-CCDI 375, CIDI 376, ODPV 382, LID 331, UI 37, CCED 1655,
OALDb 1333, LDCE 1478, CIDE 1495, he tore strips off both of us-CCDI 375, He had a strip/strips
torn off him ODPV 382, passive: got torn off a strip-CCDI 375, He was torn off a strip, A strip
was/Strips were torn off him-ODPV 382
tear sb to pieces x x1 x x x x x x x
tear sb to pieces-ODPV 382, LID 307, UI 37, CCED 1243, OALDb 953, LDCE 1063, tear sth to pieces-
ODPV 382, LID 307, UI 37, OALDb 953, tear sb/sth to ribbons-ODPV 382, tear sb to shreds-ODPV 382,
LID 307, CCED 1543, OALDb 953, CIDE 1494, tear sth to shreds-ODPV 382, LID 307, OALDb 953,
CIDE 1494, rip sb to shreds-LID 307, CCED 1543, LDCE 1228, rip sth to shreds-LID 307, LDCE 1228,
rip sb to pieces-LDCE 1063, rip sth to pieces/shreds/tatters-ODPV 300, pick/pull sb/sth to shreds-
OALDb 953, pull sb to pieces-CIDI 298, UI 37, OALDb 953, LDCE 1063, CIDE 1064, pull sth to pieces-
CIDI 298, UI 37, CCED 1243, OALDb 953, CIDE 1064, pick sth to pieces-CIDI 298, CCED 1243,
OALDb 953, LDCE 1060,CIDE 1064, pick sb to pieces-CIDI 298, OALDb 953, LDCE 1060, CIDE 1064,
take sb/sth to bits/pieces-EI 183, passive: ODPV 382, 300, was torn to pieces LID 307, CCED 1243, was
ripped to shreds-LDCE 1228
tell sb where to get x x x x x x x x

219
off
tell sb where to get off-CCDI 385, CIDI 385, LID 342, PDEI 141, EI 183, UI 37, OALDb 1337, LDCE
1482, tell sb where to go-LID 342, tell sb where they get off-OALDb 1337
the pot calling the x x x2 x x x x x
kettle black
the pot calling the kettle black-CCDI 306, CIDI 307, LID 270, PDEI 154, UI 37, OALDb 986, LDCE
1100, Pots should not call kettles black-CCDI 307, a bit of pot-and-kettle about-CCDI 307, the pot calls
the kettle black-ODEI 463, The pot is often entitled to call the kettle black-ODEI 463
throw sb to the x x x x
wolves
throw sb to the lions-CCDI 247, CIDI 232, LID 215, feed sb to the lions-CIDI 232, throw sb to the
wolves-CCDI 247, BrE, AmE, AusE CIDI 429, LID 388, OALDb 1488, leave sb to the wolves-AusE CIDI
429 throw sb to the dogs-CCDI 105, CIDI 101, LID 90, passive: she's been thrown to the lions CCDI 247,
he was thrown to the wolves CCDI 426, he had been thrown to the dogs-CCDI 105, I had been thrown
to the dogs CIDI 101, I had been fed to the lions CIDI 232, I 'd been thrown to the wolves CIDI 429
tounge-lashing x x x x x
tounge-lashing-CCDI 395, CIDI 397, CCED 1764, CIDE 1534, tongue-lashings-CCED 1764, give sb a
tongue-lashing-CCDI 395, CIDI 397, LID 356, CCED 1764, CIDE 1534, tongue-lash sb-CCDI 395,
variant: a cruel tongue-lashing-CCDI 395, CCED 1764, passive: he was given a severe tongue lashing-
OALDb 723
turn one's guns on
sb

220
15. APPENDIX 7

Below you can find concordance lines taken from our corpus. For reasons of space
for some idioms only a selection of the total set of corpus examples is given here. This is
marked by an asterisk (*) at the end of the heading. The first and the last sentences are
never omitted, so that the number of the last sentence also shows the total number of
examples. If no concordance lines are given for a particular idiom, no examples have been
found.

armchair*
1.rld.It is the sort of event that makes the armchair adventurer gulp with the
3.arts provision can only be good news for the armchair aficionado. Various TV
4.t for would-be oil analysts: the oil price.Armchair analysts can have some fun here
5.etroleum should at least remind investors armchair and professional of the
6.lengths like 26 miles.This is why we at the Armchair Athletics Association (AAA)
7.Times: PoliticsDeals put brand names before armchair audience of
8.y have invented it to delight the late-night armchair audience. Until television took
9.hip history_ _Deals put brand names before armchair audience of millions_ BY
10.sorship and advertising at events watched by armchair audiences of millions became
12.mes Next page: High honour for armchair ballonist [Todays election stories] [Pr
13.High honour for armchair ballonist BY SIMON DE BRUXELLES
14.--===The Times: Britain: High honour for armchair ballonist ===-- April 24 1997
16.hat lies in store from Benaud and Co for the armchair cricket fan this summer © A
19.r reports, I remind myself of the danger of armchair cynicism.It is all too easy to blame
21.on.Trussed up in his dressing gown, Klass's armchair democrat sings old SA anthems
22.ay know of, or live with, one of the army of armchair detectives who ... pointed to
23.*From Mr B*.*J*.*Goodchild *Sir, As an armchair devotee of bowls for some years,
24.it needs a steady hand on the tiller if the armchair enthusiast is to be nursed through
25.and reliable and they won two years ago. Armchair enthusiasts can follow their
26.ents, hippies, yuppies, film location folk, armchair explorers and writers including
27.villed by youthful eagerness". Millions of armchair fans were also disappointed.At
28.I-TECH advances are revolutionising life for armchair football fans.While engrossed
29.to criticise a touring team from a far-off armchair, for even the foreshortening virtues
31.eas, Archie is a stormtrooper rather than an armchair general. He leads loudly from the
32.Government was getting "dubious advice from armchair gurus" who were suggesting
33.tertaining reference work for any student or armchair historian.0171-306 1100 + _Play
34.ting has supplanted food as Britain's newest armchair hobby are the TV programmes
35.too. For most of us, interior design is an armchair hobby: we read an article about
36.kes him more difficult to judge harshly than armchair intellectuals who plumped for
37.chael Joseph, £16.99) 0 0 305 18 THE ARMCHAIR INVESTOR Bernice Cohen
40.y, is to offer some of its services free to armchair investors through a deal with

221
41.ovember 1) and the international audience of armchair jurors flick on to the next
42.and could fetch as much as £3,500. For the armchair matelot, there is a lovely Parisian
44.erm? Inserting gloomy clips of DrPersaud's armchair psychiatry seemed a misguided
45.w up, get married and mortgaged, and become armchair reactionaries themselves.But
46.ialist realism.The face that once inspired armchair revolutionaries across the world
47.s another full-time committed anarchist, an armchair revolutionary working this
48.iny, not only by experts but by thousands of armchair sailors, many of whom have no
49.ame to the UK as a schoolboy, is moving his armchair share-picking skills into a new
50.e best travel writing has been done from the armchair.The best travel writers,
52.ors _ Airtours 01760 260 000 Armchair Tours 0181-560 8008 Aston Coaches 01905
53.off *Malaria Capers* by Robert Desowitz.The Armchair Traveller website, **carries
54.an browse abroad and shop at home - and the armchair traveller only has to wander
55.n's *Stones of Aran* transformed me from an armchair traveller into a real
57.el Next page: Reading test for armchair travellers with a passion for facts and fig
61.Sunday's Travel Reading test for armchair travellers with a passion for facts and figu
62.r the next series of *Wish You Were Here?*, armchair travellers are offered shows
64.ts and figures THE ONE place where armchair travellers, package holidaymakers and
66.ue - remember all gambling is a risk." + Armchair Tycoon *is published by Robson
67.hat only comes with time." Stacey's book, *Armchair Tycoon*, makes the business of
69.s] Next page: How to be an armchair tycoon [Line] Copy
70.The Times: Features: How to be an armchair tycoon November 24 199
71.ng, says Tina Gaudoin_* + How to be an armchair tycoon_* Malcolm Stacey has
72.clothes and boyfriends?_* + How to be an armchair tycoon_* + Mind and Matter -
73.o protect her.Was he wrong? What would the armchair warriors who comment on our

V + flak*
1.ent." But Labour comes in for just as much flak."They will do as the Tories do, which is
2.hit *Walk This Way,* then attracted tabloid flak by touring with their wayward
3.e losses and have been attracting increasing flak from their largest institutional clients.
4.bad times, but I never let my head drop.The flak I got was unbelievable.I was so low at
5.s just melted away, leaving ENO to face the flak alone. As so often, the Arts Council has
7.more senior colleagues, perhaps fearing the flak, lit up. Bad move.The sensitive
8.managing Morecambe and Wise to fielding the flak over *The Word*, Grade has
19.if successful, enable both to flourish. The flak directed at the BAF for failing to
10.in the same position to be prepared for the flak if your jokes go flat, but it is better to
11.personality. Dr Ratey is prepared for some flak when he starts promoting the book in
12.esman said: "She was well prepared for this flak and knew the salary would be
14.y translator, Dzibrila, is that she gets the flak from me, and Dr A just gets Dzibrila's
17.AS IF the Government was not getting enough flak over benefits reform, the Council
19._Eurostar_ may have experienced flak on the track but its French language-traini
20.The crowd was always going to give him some flak after everything that's been going
21.for £46 million in September and into some flak for allegedly paying over the odds for
22.ain at a time when the Government is taking flak over its Welfare to Work programme
25.e was up in Scotland and he's taken a lot of flak. Why the hell wasn't he there? "I
26.not fashionable at school."I get a lot of flak.They think I'm old before my time.But ther
27.e weekend.But Matthew was getting a lot of flak from the governors for what we did
28.for Steel contract_* + LucasVarity takes flak over dividend change_* + Societies to
29.Next page: LucasVarity takes flak over dividend change

222
30.--===The Times: Business:LucasVarity takes flak over dividend change ===-- April
33.Next page: Flak aimed at air force candidate Arts (Mon - F
34.Next page: Flak aimed at air force candidate [Contact Us]
35.ones, who have drawn a huge amount of ageist flak at every stage of their career over
36.nd truculent passengers.Taking most of the flak in tonight's episode is Jeremy Spake, a
37., the Health Minister, who took much of the flak for the Government in the Formula
39.nd Goold had to take a great deal of public flak.Before he stepped down in 1993, he
40.re Secretary, Chris Smith, has received more flak than most of his Cabinet colleagues
41.on.I get rid of them.I'm here to take the flak." _*_*
47.would move early: let the councils take the flak" (report, February 28).This will come
48.s, and salaries, of those who have taken the flak at Camelot. _*
49.eam is not performing, the captain takes the flak," he said, utterly without self-pity."I
50.ng nasty to each other; the agent takes the flak.You can't put buyers and sellers together
51.that her brother has also been taking some flak from his mates. "I decided to write the
52.Working mums have been taking some flak recently.In narrowly-based research by
54.el Henderson_ meets a motivator taking heavy flak © Lloyd, with back to camera,
56.ce, and could absorb much of the inevitable flak from the teaching profession. The
57.mme not, I trust, to absorb the resulting flak. *_JASPER GERARD _ _*
58.ebbed away.People became so used to getting flak for any decisive act that, with their
59.to be independent.I know I am going to get flak for this decision, but I have to consider
60.id.She won at Evian last year but took some flak for having a television set in her bag
61.s looking for someone to blame, I took more flak.But the F-factor concept isn't
62.s than critics."The programme we took most flak on originally was *Hollywood
65.+ Doctors deny that baby was shaken_* + Flak aimed at air force candidate_* + Carter
66.put up by Dunemann but they dealt with the flak and responded in style.Hassan,

Aunt Sally*
1.ary Mabbutt, has become more and more of an Aunt Sally.But the crisis, surely, is one of
2.h Maldini, Ignatiev, the team's coach, is an Aunt Sally, ridiculed as a tactician, criticised
4.ss and to be middle class today is to be an Aunt Sally.We may pay our taxes, get by with
6.wit and commonsense.Social workers are easy Aunt Sallies for Conservative politicians,
7.ies are forewarned. They have been Labour's Aunt Sally since they were privatised.They
8.rest of the world loves to bash a political Aunt Sally.Historical cost figures give you a

backbiting*
1.8.00 EastEnders_ After much speculation and backbiting, Grant and Tony finally find
2.money yet because of all the infighting and backbiting on the federation, they were
3., with some of the insinuations, rumours and backbiting at last to be ended, the prospect
5.efurbishment. Political intrigue, artistic backbiting, architectural incompetence and
6.the local press, which is notorious for its backbiting. As for the rest of former
7.ne association insider, a "hell of a lot of backbiting", with furious letters from the
8.ader insisted that he would not tolerate the backbiting and feuding that marred John
12.lawyers entered the fight After all the back-biting, the dirty tricks, the spin doctors,
13.e said: "I have no time for name-calling or back-biting, the self-interested or the self-

back-seat driving/driver
1.at." Lady Thatcher, who was once seen as a back seat driver for John Major, always

223
2.public life" but he was rarely, if ever, a back-seat driver. Foster's biography also
3.She has no wish to be the Prime Minister's back-seat driver. Nonetheless, political
4.es.For one thing, he will have millions of back-seat drivers watching over his
5.uld not take advice, lest she be accused of "back-seat driving". Mr Blair, too, would be
6.errors of the war and a classic case of "backseat driving" leading to a disaster. Ewing
7.aints or of prodding "to get on with it" "backseat driving" made technically possible by

V + aspersion(s)
1.Senator out to cast aspersions FROM IAN BRODIEIN
2.eryone knew that he had not intended to cast aspersions on the Jewishness of his left-
3.kets yesterday may routinely, nowadays, cast aspersions in the direction of the
4.==THE TIMES: FOREIGN NEWSSenator out to cast aspersions ===-- [Navigation
5.who recently took an advertisement to cast aspersions on a critic's qualifications. And
6.Next page: American senator out to cast aspersions Copyright 1997.
7.s has taken this a stage further by casting aspersions on the legal systems of several
8.of siding with Miss Woodward and of casting aspersions on the quality of American
9.e £250 compensation. This casts considerable aspersions upon the insurance companies
10.to execute their duties unfettered by false aspersions." A solicitor from Allen Gledhill,
11.at press conferences. And it will make foul aspersions of sex scandals or midnight
12.Residents resent aspersions on American justice, reports _Tunku_ Varadar
13.ts about au pair's guilt_* Residents resent aspersions on American justice, reports
14.ar implication was that he expected serious aspersions to be cast on the plans laid out
15.hrough seemed no reason to cast toffee-nosed aspersions on his self-evident abilities in
16.ream Believer*, which cast cruel, unrhyming aspersions on both Strachan and his team.
17.ed and furious with rage, cast the ultimate aspersion at his Leicester players midway
18.ed careers, bankrupt businesses? I cast no aspersion on Dr Stuttaford's agenda or the

come down on sb like a ton of bricks


1.ussed the possibility of sterner discipline. Coming down on him "like a ton of bricks"

V + fire*
1.n Brown had been led to the dispatch box and concentrate his fire on what was said in
2.ready." In his speech last night, Mr Hague concentrated his fire on the Government's
3.whole approach of the tests. Head teachers concentrated their fire on the way the tables
4.bove all, for an end to introspection and a concentration of fire on Labour. To that end,
5.it when Andre Agassi is playing. Agassi can draw the media's fire; Sampras can sneak in
6.d-about widow, Dimitra Liani-Papandreou, has drawn a fresh round of fire from the
7.TIMES: FOREIGN NEWS Cigar-smoking supermodel draws fire July 10 1997
11.can resort with British skiers, Breckenridge draws heavy fire: it's boring, it's flat. But I
12.g 69 1/2 p to 880p. But the deal immediately drew fire from unions representing the
15..5 per cent. The Government has come under fire from employers who say that its plan
17.elieve that Mr Major has a resolution under fire that has been underestimated in the
21.lender in this market, which has come under fire from borrowers for its punitive level
23.=THE TIMES: FOREIGN NEWSBelgian police under fire ===-- April 15 1997
26.Eisenhower's Vice-President. Nixon was under fire for accepting gifts, including
31.es tend to be upset that their hero is under fire, while some Americans worry that
35.building society turned bank, has come under fire from its staff union over plans to

224
36.riding bans yesterday. O'Connor came under fire for his riding of First Village, the
37.fairs. The Paymaster General has been under fire over the purchase of £12.5 million
38.cht*, the writer's reputation has been under fire. Professor Fuegi argued that Brecht's
39..2 billion merger last year, has come under fire for the clumsy debate it has held over
42.spend_* Gordon Brown will today come under fire over his refusal to boost economic
46.secretive, wary body coming under increasing fire from Holocaust survivors in
47.straight on the chin. He remained cool under fire, quietly trying to put his version of
49.ear as backing his party's values turned his fire on the Labour leader, accusing him of
50.not for the likes of me. I therefore hold my fire. _*_*
51.News - 'Energy' drinks under fire for caffeine content Next page: Tears flow
52.new war zone _ THERE was no mortar fire on the Cheshire roads yesterday, just
53.en with eggs_* + 'Energy' drinks under fire for caffeine content_* Robin Young
54.was riddled by a burst of political friendly fire yesterday. *The New Yorker* magazine,
55.ommons. The former Governor's critics, under fire in the Dimbleby book on Mr
56.Roy Hattersley, who has kept up a withering fire on what he sees as the Government's
58.ociety *Sir, Cardinel Hume is training his fire on those who campaign for the right of
59.rs' Union of Scotland, had been under heavy fire recently from grassroots members for
63.HE TIMES: FOREIGN NEWS_Netanyahu stays under fire as he escapes charges_
64.he fault of the coach which puts me under fire," Dwyer said. "If you can play one,
69.The sale of the homes came under political fire. Critics claimed that the former
72.dwide ban on landmines have drawn political fire. Last month she cancelled a meeting
76.and rape victims. Judges still come under fire over seemingly crass remarks to rape
77.re happy to seize any opportunity to return fire. It is good-humoured almost. The rest
78.opularity, now reinforced by his decision to hold fire on EMU, which gives him a claim
79.. The Bar Council and the Law Society have directed their fire at the proposal that
80.ide, naturally. If I had known, I would have directed my fire at less dedicated
82.crime rates under a Labour government as he turned his fire on Mr Blair's failure to
83.+ Rome File by Richard Owen_* Loose cannon turns fire on art world; Olympic
84.--===THE TIMES: FOREIGN NEWS Loose cannon turns fire on art world ===--
86.no proper debate," Sir Leon said. He also turned his fire on the Tories. Sir Leon said:
87.t it was their turn to be favoured and have turned the fire of their television companies
88.--===The Times: World News:Cartoon attracts Asian-American fire ===--
89._WORLD NEWS_ Cartoon attracts Asian-American fire FROM QUENTIN LETTS
90.Next page: Cartoon attracts Asian-American fire Copyright 1997.
91.transmitter tower. Earlier this year, Orange attracted fire from design critics for
92.The Bar Council and the Law Society have directed their fire at the proposal that
93.ide, naturally. If I had known, I would have directed my fire at less dedicated

faint praise*
1.ted yesterday to a torrent of criticism and faint praise, with the European Union leading
2.the reaction that it aroused. Damnation by faint praise has become an occupational
3.previous batch of amateurs were damned with faint praise by their instructors. Others,
4._* DAMNED with faint praise and gently castigated for choking the centre
5.good blocks." It was a case of damning with faint praise. Bassett was nearer the mark
6.faint praise was the order of the day when ministers yest
7.of two of Derby's goals, he damned him with faint praise. "Wanchope? He has the
8.particularly impressed, his manager uttered faint praise, and as for the locals well, it
10.of M Jospin and M Strauss-Kahn offered only faint praise for the Chancellor's plans

225
11.awn that many use to damn German sides with faint praise. When the Hungarian
13.ey after this match in terms that suggested faint praise, such as "hard-working".
14.atting and decent bowling, although that was faint praise for Shine, who could never
15.original director of this century". This is faint praise when set against the enthusiasm of
16.em and the critics had damned them with the faint praise of being second best to a
17.had players worthy of being damned with such faint praise. Without Thorpe in the

salvo*
1. deputy-super-monster, John Prescott, with a salvo of killer-quotes from the monster's
2. o defend the Assisted Places Scheme after a salvo against it from one of Labour's
4. of geriatric sex. Letting loose salvo after salvo of crude humour against the backdrop
5. relevant to the trial.Saving his biggest salvo until last, Mr Jones sought to discredit
7.Next page: Stena Line fires salvo at MMCArts | British News | Business | Co
8. Stena Line fires salvo at MMCBY GEORGE SIVELL STENA LI
12. edge of collapse_* + Stena Line fires salvo at MMC_* Swedish cross-channel ferry
13. tton voters did it for him.Between firing salvos at Labour, William Hague should
14. eliberately restrained from firing the first salvoes. The word in the ranks was: when
15. k, Alec Broers of Cambridge fired the first salvo in what could become a
16. to Scotland even before he fired his first salvo at Scottish Amicable. Since Abbey
17. ind, so great is their anger over the latest salvo of complaints from Paolo Di Canio,
18. ate."Mr Mandela's outburst was the latest salvo in a deepening diplomatic row that
20. n Had Lady Macbeth?* was one of the opening salvoes in the battle against the
21. S on Capitol Hill yesterday fired an opening salvo in the annual debate over
22. as cats to creamy milk.After an opening salvo on America's "cultural imperialism"
23. ban Dublin bigot. This particular satirical salvo from one of the godfathers of
25. ll-scale attack launched in three successive salvoes: *Shakespeare's Sonnets*
27. t of view, John Major felled Tony Blair with salvo after salvo of withering

lashing
1. appeared remarkably impervious to the tongue lashings of Economic Secretary
2. trictures_ BY SIMON WILDE _* THE lashing that English cricket received from
3. n free two-minute calls home. _* _Tongue lashing_ The Government has launched

stick*
1. ship group stages, British clubs have had a stick taken to them by the Australian media
3. be recognised. Each bad review feels like a stick on your back. A prize like this lifts
4. today and the airport delays were used as a stick to beat the British. Ever since, people
5. be they Croatian, Bosnian or Albanian, as a stick with which to beat the Serbs. On the
6. nd others were resentful or feared it was a "stick" with which to beat them. Although
7. he Arsenal manager and set about him with a stick. Before they were homogenised,
8. . "Patten's political agenda has provided a stick for foreign lobby groups with which to
9. nsciously in others, nobody is looking for a stick with which to beat the Labour Party,
11. result, the temptation to use it as another stick with which to beat the English game
13. er's* Editor, _Will Hutton_, has come in for stick from *Campaign*. In a recent
14. o abuse at Barnsley last Saturday. "Getting stick from supporters isn't very pleasant,
15. istic of a £1 million pay-out."I was getting stick from the boys and I thought: 'I'll show
16. ed. "I did well on Saturday and I still got stick for it," he said. "Too fat, too thin, the sa

226
17. an should be given the job. "I'd never had stick from a crowd before I came here and it
18. irl should have come in for some ideological stick, but I have never quite understood
19. ents from beating him with a civil liberties stick. However, a pungent irony will not
20. nority who were giving the previous manager stick," Hateley said. "They were very
22. have never understood why I've taken so much stick and I'm not going to lie and say
24. Sir Cliff Richard would be used to a bit of stick from the critics. Not a bit of it. Sir Clif
25. harp as a tack now. "He has taken a lot of stick because of the money we paid for him.
26. r today. You have got to expect the sort of stick he got. It was a bit over the top, but
not
27. because I have been getting an awful lot of stick from the rest of the lads for my lack
28. ing, Danny has come in for a fair amount of stick for his voice. "I think people forget
29. ularly the players, who took a little bit of stick when we had one or two injury
30. es that easily. Jason had to take a lot of stick, but he just laughed with the rest of us. Th
31. targets such as the media. We got plenty of stick after our FA Cup defeat at Chelsea,
36. with the lads getting together to give some stick to one unfortunate player. Sadly, that
37. en a tough week. Some of the boys took some stick in the press, but they responded to
40. release of a year-long tension. "I got some stick for not putting the ball down, but I new
41. 1993. But the Australians have been taking stick for years and they have not
42. men are always getting the rough end of the stick. Butter doesn't just melt, it dribbles
43. al" profit in 1998, adding: "I've taken the stick. I want to be around to take the praise."
44. might be resigning, saying: "I've taken the stick. I want to be around to take the
45. ying like that, he will soon forget all the stick he took last season. While I was in
46. ALTHOUGH he is routinely beaten with the stick of his own past achievements,

rap*
1. ilar transgression). "They got me on a bum rap," Lenahan seethes. "I argued that I was
3. delivering a stern admonition; a well-timed rap on the knuckles from the headmaster.
4. e: Pru's attention turns to savers after FSA rap Arts (Mon - Fri) | Books (Sat) (Thu) | Bri
7. e: Pru's attention turns to savers after FSA rap [Line] Copyright 1
8. once, mangle my metaphors: he attempted to rap them on the knuckles but succeeded
9. communicated via contaminated water, gets a rap over the knuckles, since he neglected
10. inguished piano teacher who believed in the "rap the knuckles" approach. I became
11. t really happen; or if it did it was just a rap on the knuckles for M Chirac, and they
12. operations, was one of the nine to take the rap. A closer look may now be taken at
14. , they cannot plead ignorance; they take the rap, along with the underlings. Extend it
15. tag in a train tunnel, unjustly taking the rap when a gang of youths derail a goods train
18. r at Taunton Cider and the man who took the rap, unfairly I always thought, for last
19. he Post Office, but having investigated and rapped Channel 4 over the knuckles, it
20. king risks because they are afraid of being rapped over the knuckles. They are
21. Skipton Building Society has been rapped over the knuckles by the Advertising

roasting
1. with her rescue plan.Ms Allen is given a roasting in Mr Kaufman's report, mainly for
2. er meeting said that Ms Harman was given a "roasting" when at least ten MPs made
3. owe, the PIA's chief executive, were given a roasting by the Treasury Parliamentary
4. ence but, even after an inevitable half-time roasting from Roy Evans, their manager,
5. Robinson stands proxy for Labour hypocrisy. Roasting time is overdue. The press

227
6. &GovernmentWily political bird survives Paxo roasting===--
7. pshott__Wily political bird survives Paxo roasting _ THE torrent of election
8. She has even come to terms with the public roasting she got over her "behind closed

get on sb’s back


1.ong the way and there will be people getting on my back for getting booked.I mean,
2. to a worker you might find the Low Pay Unit on your back. Also, men and women must
3. being pumped high and far and the crowd is on your back, there is no worse place for a
4.ssle.People need to encourage him, not get on his back." Therein lies the secret.Tell him
5.se his form has been down, people have been on his back, which is understandable, but
6.crum and pop the ball in and the referee got on my back for a while.Once we had sorted
7.sues without ever getting the Moral Majority on his back and, rarely for a pop star, he
8.supporters to barrack him [Reed], really get on his back.Use words like 'fraud' and 'you
9. parliament." Ronnie wants parliament to get off everyone's back, so that everyone can
10. years and getting all those former players off the team's back.Yorkshire were tipped
11. and next summer is to keep the little sods off my back and to try to stop them taking
12. Next page: How to get the world off your back Arts (Mon - Fri) | Books (Sat) | British
13. Next page: How to get the world off your back [Contact Us] [Subscribe to the paper]
14.==The Times: Interface: How to get the world off your back ===-- October 29 1997
15. How to get the world off your back A FIVE-MINUTE sandwich break at th
16. like John Edmonds, leader of the GMB union, off the Government's back.But while

get on sb’s case


1.to do, regardless of whether my parents were on its case or not. And anyway, why

V + lecture
1.Chinese leaders are bracing themselves for a lecture on human rights. However, the
2.and financial leaders politely listened to a lecture from none other than Kenneth Clarke
3.eve when he was arrested Danielle gave him a lecture that if he broke the law then he
4.ing," he said. By November they were getting lectures from the stage on the importance
5.e, but in both decided to settle for a sharp lecture, and the hope that ill-managed
6.minutes to level: Johnson received a stern lecture after a confrontation with Fidler, and
7.improvement in Japan. But it gives a stern lecture on unemployment, calling on
8.urkin, the referee, handed out only a stern lecture. "I don't know why I did it,"Sinclair
9.rough the show he launched into a terrifying lecture on the evils of drink which put a

give sb the rough side of one’s tongue


1. and is as choleric as Herr Kohl. But his rough tongue is typical of both the political

haul sb over the coals


1. be known that he is not alone in casting hot coals onto the leaders of the profession. His
2. onservative Cabinet minister raked over old coals by attacking the CBI for criticising
3. k and, for his pains, he was hauled over the coals by a committee that preferred to "see
4. urites, please let me know. _"Rake over the coals (3,4,2,4)"_ is one of mine. Hazel
5. d vote for him. We once got hauled over the coals for having too many Labour MPs, so

228
have a thick skin*
1. ival material later. The other defence is a thick skin. With exceptions for the bizarre,
2. imes told © Simon: "I have thick skin and will not be deflected"
5. is obstinate nature, has helped to develop a thick skin and a tendency for defiance. If he
6. lass of champagne, a packet of Nurofen and a thick skin, the most crucial accoutrement
8. I am blessed with a fairly loud voice and a thick skin." Having enjoyed the benefits of
9. ree World hates your business, you develop a thick skin. BAT has survived the
10. y perturbed. They are right up there in the " thick skin" stakes, perched somewhere on
11. the media or politics you have to develop a thick skin," says Mr Swash. After feeling
13. him," Maloney said. "But you have to have a thick skin and deep pockets and be
14. t of what he has got. He has certainly got a thick skin. Maybe it's to make the most of
15. nies' highly skilled tax accountants and a thick skin. Inspectors need to know how
16. der if his pen nib would ever penetrate her thick skin. Ann Widdecombe, the former

have the hide of a rhinoceros


1. ith parliamentary nous, PR experience and a rhinoceros hide. Whatever turns up,
2. lomacy, negotiating skill and the hide of a rhinoceros. Plus enthusiasm and
3. ly thin-skinned in a job where the hide of a rhinoceros is required. Pressure, and the

honeymoon*
1. , it will have to face the first settling of honeymoon accounts. Next week's Budget will
14. abour. But, as Downing Street will confirm, honeymoons do not last forever. "I don't
20. ris Political Sketch] _Icy smiles as one honeymoon ends and another begins _ Festive
21. The Times: Politics Icy smiles as one honeymoon ends and another begins
43. ll claim God led them to him. All had brief honeymoons in southern Utah. All say their
49. elancholy exchanges do take place, once the honeymoon is over. _*
50. bour Government. It is quite clear that the honeymoon is almost over. Let's be clear, if
55. nce_* + Matthew Parris sketch_* The honeymoon is over despite effort to make up_*
72. ingless the size of his Commons majority. Honeymoons never last. Already darkening
108. e. Past experience suggests that political honeymoons seldom last for long. Back in
113. en Dorrell meanwhile claimed that Mr Blair's honeymoon was over as he stepped up
114. nly the second Labour Government. Labour's honeymoon was brief. On September

in the line of fire*


1. The Times: Business: Merrydown board in firing line July 12 1997 _BUSINESS
7.Times: Cricket: Proud Gooch retires from the line of fire July 21 1997 _CRICKET_
8.manager, has withdrawn Michael Owen from the firing line today, with the experienced
19. are," he muses.Yet he puts himself in the line of fire to document the dinosaurs left to
29. Next page: Social chapter in the firing line as Britain becomes more competitive Arts
30.olidays seems to think so _* + In the line of fire _* The EC is set to toughen up on
39.===The Times: Sport:Kallis stays calm in the line of fire ===-- May 16 1997
52.olidays seems to think so _* + In the line of fire _* The EC is set to toughen up on
60.ath of a diamond smuggler puts Quincy in the firing line between the FBI and the
61.olidays seems to think so _* + In the line of fire _* The EC is set to toughen up on
62.." The Attorney-General is regularly in the firing line over decisions he takes in his
63.one tariff war_* + Social chapter in the firing line as Britain becomes more
70.ing wear than anyone else, are right in the firing line.Designers favoured in the mid-

229
87.bour's campaign manager, was kept out of the firing line even when an important
90.ong with other utilities in the windfall tax firing line, the Grid is lobbying to keep its
91.e] _Gore in firing line over cash collected at temple_ FROM IAN BR

leave oneself open


1. uckrik's first chapter is not one which will leave her open to accusations of literary
2. by not fully declaring their interests, MPs leave themselves open to accusations of
3. ries lacking much competition or regulation leaves it open to future criticism. Nor, if the
4. s did, on a change of heart a phrase that leaves new Labour open to mockery from the
5. has shared a room or tent with a homosexual leaves them open to accusation and
6. . Their own parents are sending them out and leaving them open to abuse of all kinds."
7. . Their own parents are sending them out and leaving them open to abuse of all kinds."
8. en did tend to surround themselves with men, leaving them open to the charge that they
9. nusual in local government circles. "We have left ourselves open to ridicule, but the

like water off a duck's back


1.as happened to Lady Diana, it is water off a duck’s back." He concluded: "I hate them, I
2., Cambridgeshire. May 24. _Water off a duck’s back _ *From Mr Mike Peacey *Sir,
3.g your colouring." It was like water off a duck’s back."I wear blue most of the time,"
4. as good as his father?', it was water off a duck’s back.If I was good enough it would
5.out it, very abusive.But it is water off a duck’s back.The public will make their own
6. the s*** they write about me is water off a duck’s back," he says."I feel absolutely
7. Next page: WaterAid off a duck’s back Arts | British News | Business | Court pag
8.imes Next page: WaterAid off a duck’s back [Contact Us] [Line]
9. --===The Times: Britain: WaterAid off a duck’s back ===-- May 27 1997 _BRITAIN_
10. _WaterAid off a duck’s back _
11.ed his resolve; no doubt it was water off a duck’s back off the backs of 24 ducks, to
12.ew Mexico, 87321. May 27. _Water off a duck’s back _ *From Mr Tony Fuller *Sir,

look who is talking


1.relations with colleagues, he can snap back "Look who's talking, Mister 67 per Cent!"

nitpick*
1. one. Dr Carey will doubtless say that I am nitpicking. His words were really
2. pany is brimming with opportunities but the nitpickers complain that it has paid too
3. complicated statement about Nato, a hail of nitpicking questions and more work. Tony
5. s who lead such sad lives that they have to nitpick over details? Your picture of Helen
7. Me_* + Home Life - Bring back the school nit-picker_* + Treating Lice_* + Life and
8. The Times: Weekend: Bring back the school nit-picker August 2 1997 _HOME LIFE_
13. ion of academic philosophy into the kind of nit-picking that Magee deplores, behind
14. the Net in real time. But such technical nit-picking aside, this is an entertaining romp,
15. , adjudicator of fair trade rather than mere nit-picking regulator. Oftel's concern is BT's
16. s received. Palace officials talked of the "nit-picking" that has overwhelmed the trip's
17. es not yet take note form. These range from nit-picking details on the redenomination
18. lous, enticing abundance on the shelves, the nit-picking visitor can spot gaps. On

230
not pull one's punches*
1. deficit with China. Mr Gore said he "did not pull any punches" on the subject since
5. comes to the Turner, the heavyweight critics pull no punches. But, says _Richard
6. for literature made it impossible for her to pull punches. She was not afraid to pass the
7. rely cruel, he had a black belt in judo and pulled few punches in his work. At his sharp-
9. sential for racing's coffers, and it has not pulled its punches. Setting a target of
11. de a name for himself as an interviewer who pulled no punches. "The public's
13. rmed by other artists. Aware perhaps of the pulled punches, Cash signed off with one
14. and grappling with serious themes. "I'm not pulling any punches," he says. "Besides,
15. h, trained as a lawyer and is not known for pulling her punches. She is a keen critic of
16. of being seen either as a damp squib or as pulling its punches. Worse, many key
17. settings of William Blake. *Bright as Fire* pulls no emotional punches, Blake's
19. ther organisations," she says. Ms Fullick pulls no punches about the borough's poor
23. to talking about past mistakes, Frank Warren pulls no punches. Talking about his most

people in glass houses...


1.from a long line of aristocrats. People in glass houses shouldn't throw stones." Mr
2.d Davies for not penalising them. People in glass houses should not throw stones. The

pot shot
1. re are potential singles a-go-go, and enough pot shots taken at sacred cows (*Keep On
2. rance profonde* and taking a few well-aimed pot shots at the left-wing Government. M
3. at Albert Fisher is that no one has taken a pot shot at this company before now. Fisher's
4. block a deal. Of course, Siebe once took a pot shot at APV, but it would be odd for a
5. stello Aragonese, at which Nelson took some pot shots. Below, in the crypt, are ranks of
6. ". Over the past week this has largely meant potshots at the Tories' flailing ducks.
7. the rising stars of Britpop gleefully took potshots at will. Suede's new guitarist, Richard

put the boot into sb


1. newspaper *El País*, putting his Hispanist's boot into the critics for sticking their
2. iness: Youngsters feel Spurs supremo put the boot in ===-- October 23 1997
3. + Youngsters feel Spurs supremo put the boot in_* Chris Ayres watches as the Brown-
4. page: Youngsters feel Spurs supremo put the boot in Arts (Mon - Fri) | Books (Sat) |
5. + Youngsters feel Spurs supremo put the boot in_* + Stock Market - Michael Clark_*
6. page: Youngsters feel Spurs supremo put the boot in [Contact Us] [Line]
7. l* leader. Other notables who have put the boot into lawyers include (in no particular
8. e case, and I know they are going to put the boot in. But if they are so intolerant and
9. ll paid to put his boots on, not to put the boot in," Kendall said. "We are at the bottom
10. esdner Kleinwort Benson, the broker, put the boot into the housebuilders by suggesting
11. atch Youngsters feel Spurs supremo put the boot in Alan Sugar should perhaps have
12. e Times: Britain: England sponsor puts legal boot in ===-- June 7 1997 _BRITAIN_
13. evidence_* + England sponsor puts legal boot in_* + Aitken denies misleading Butler
14. Next page: England sponsor puts legal boot in Arts | British News | Business | Court
15. Next page: England sponsor puts legal boot in [Contact Us] [Line]
16. England sponsor puts legal boot in BY JOAN

231
rap sb on the knuckles
1. inguished piano teacher who believed in the "rap the knuckles" approach. I became so
2. he Post Office, but having investigated and rapped Channel 4 over the knuckles, it
3. king risks because they are afraid of being rapped over the knuckles. They are worried
4. Skipton Building Society has been rapped over the knuckles by the Advertising

read sb the riot act


1. [Valerie Grove Interview] Bland reads riot act" said one headline yesterday. Bland
2. ck did the research for his first novel *The Riot Act*. Stock, a stranger to the world of
3. stand for it any longer. We really read the riot act, telling both parties that we wanted
4. , the enormity of their offence. He read the Riot Act: it had no effect. They then
5. to the sentence of death. He then read the Riot Act, when four of the principal of these
6. Previous to Sir John Hippisley's reading the Riot Act, he informed these infuriated
7. ffer her critical assistance by reading the riot act to Slobodan Milosevic, the Yugos

run the gauntlet


1. Bologna University, was 12 when his father ran the gauntlet of Italians carrying
2. BY POLLY NEWTON _* WOMEN MPs run a gauntlet of sexist comments and
3. ming an accountant. Many examinees have to run an extra gauntlet of parent-created
4. Next page: Troops run media gauntlet to take Albanian beaches
5. le: Work for St Helena Next page: Troops run media gauntlet to take Albanian beaches
6. perial relics happy to be pink_* + Troops run media gauntlet to take Albanian
7. Photograph: LUCA BRUNO/AP _Troops run media gauntlet to take Albanian
8. --===THE TIMES: FOREIGN NEWS_Troops run media gauntlet to take Albanian
9. ) he vanishes with his flock, leaving her to run the gauntlet of folk in a high chapel
10. idley almost ten years ago, but still had to run the gauntlet of a lengthy High Court
11. seemed a crazy thing for Richards, who had run the gauntlet with all the great fast
12. her husband's behaviour that she decides to run the gauntlet of social ostracism. As
13. d, the chairman, and fellow directors are to run the gauntlet at a special meeting
14. hat state-of-the-art 60s should then have to run the gauntlet of a maelstrom at the
15. was not Eton's fault. I believe I would have run the same gauntlet or worse at any
16. Photograph: HAL GARB/REUTERS Star runs gauntlet of the lover he scorned
17. --===THE TIMES: FOREIGN NEWS Star runs gauntlet of the lover he scorned ===-
18. s call off challenge to Yeltsin_* + Star runs gauntlet of the lover he scorned_* + El
19. Next page: Star runs gauntlet of the lover he scorned Arts (Mon - Fri) |
20. Next page: Star runs gauntlet of the lover he scorned [Contact Us]
21. initiated," party leader said_* + Star runs gauntlet of the lover he scorned_* Clint
22. usader learning politics the hard way as he runs the gauntlet of divided local Tories ---

sitting duck/target
1.nothing on under their buckskins. They are sitting targets for the sneers of next-door
2.SUMMIT SKETCH © Clinton: sitting target for cartoonists and satirists

V + hammering
1. British Leyland's Princess comes in for a hammering from Norman Milne of
2.and the Baebes are bracing themselves for a hammering from the classical press. Even
3.so much. If it's justified, then I deserve a hammering. I'm just trying to cut down the

232
take it on the chin
1. people who founded the country will have to take it on the chin. The term Pom is not
2. d family with patience and sensitivity, and taking on the chin drunken abuse meted out
3. ed family with patience and sensitivity and taking on the chin drunken abuse meted out
4. sations along the way. All of which Mr Major took straight on the chin. He remained

take to task*
1. Next page: Charltons take Banks to task Arts | British News | Business | Court pag
7. ic clash with Mugabe_* + Jewish students take Yale to task over dormitory rule_* +
10. ing George, but since then the critics have taken him to task. They questioned his
14. Next page: Water companies taken to task over dividends Arts (Mon - Fri) | British
17. ct Us] Next page: Friedman taken to task on euro [Line] Cop
21. ore recent years white Australians have been taken to task for their appalling crimes
22. en Watt *Sir, Once again, Mother Teresa is taken to task for her views on abortion and
24. rm. _Small print_ BARCLAYS BANK is being taken to task by the Advertising
25. r tours ended and now the players are being taken to task by the most revered and
30. --===The Times: Football: Referees taken to task by Wenger ===-- October 25 1997
33. office has taken such action. Ofwat has been taken to task in its handling of contracts
34. went down a treat. However, Mrs Beckett was taken to task for not mentioning the
35. this week, Mr Brown was quietly and firmly taken to task by Ruth Kelly, Labour MP
36. arcelona, one national newspaper scribe was taken to task for having described the
37. llion, then its senior executives should be taken to task for sloppy management. If it
38. tly through Mr Greer." Mr Hamilton is also taken to task for failing to declare the two
39. TOR_ McAlpine memoir taken to task __ *From Baroness Falkender *S
40. tain my excitement. __ + *I AM gently taken to task by Jonathan Haslam, director of
42. ONEY _ _Sara McConnell_ takes Abbey National to task over advertising campaign
49. en Alliance *Sir, Your editorial of April 7 takes environmentalists to task for failing to
50. charges. (3652) _7.30 EastEnders_ Lorraine takes Grant to task over his treatment of
52. __ *From Lord Quinton *Sir, Lord Rees-Mogg takes me to task (article, January 23)
55. Sir, Mr Q. S. Anisuddin (letter, November 4) takes to task a former head of the UK
56. 76) _7.30 Weekend Watchdog_ AnneRobinson takes tour operators to task for
59. vo Mr Simon Goulden (letter, January 23) for taking Joe Joseph to task for suggesting
60. rnter reserves particular ire for the press, taking newspapers to task for describing
62. 5 Neighbours_ Karl decides enough is enough, taking Tim to task over his obsession
63. e and shy fund managers became media stars, taking to task company directors for
64. g the past week, and we know that we deserve taking to task. There are no excuses.
65. inally rung up by an indignant investor who took him to task for not being the Lloyds
66. an article in a foreign magazine, allegedly took Collymore to task for his selfish play

talk to sb like a Dutch uncle


1.ea, talked to Haaland like, you could say, a Dutch uncle; no immediate booking. Nor

the pot calling the kettle black


1.Court. Talk centred on the Lord Chancellor's pot*-*and*-*kettle criticism of fat cat
2.he accused, in a memorable exchange between pot and kettle, of turning track and
3.e says he might stay here." Talk about the pot calling the kettle black. Ballesteros

233
tongue-lashing
1. will let language police give shopkeepers a tongue-lashing FROM QUENTIN LETTS
2. + Language police can give shopkeepers a tongue-lashing_* + Russians 'building
3. ange from Clive James giving Liza Minnelli a tongue-lashing to a segment from
4. page: Language police can give shopkeepers a tongue-lashing [Contact Us] [Line]
5. with Julian Clary unless they wanted a good tongue-lashing (so to speak). The secret
6. their player/ manager delivered any sort of tongue-lashing to his men afterwards it
7. appeared remarkably impervious to the tongue lashings of Economic Secretary
8. n free two-minute calls home. _* _Tongue lashing_ The Government has launched

234

Das könnte Ihnen auch gefallen