Professional Documents
Culture Documents
By Word of Mouth
By Word of Mouth
Editorial Address:
Linguistics (GER)
University of Antwerp (UIA)
Universiteitsplein l
B-2610 Wilrijk
Belgium
Editorial Board:
Norbert Dittmar {Free University of Berlin)
Bruce Fraser (Boston University)
John Heritage {University of California at Los Angeles)
David Holdcroft {University of Leeds)
Catherine Kerbrat-Orecchioni {University of Lyon 2)
Beatriz Lavandera {University of Buenos Aires)
Marina Sbisà {University of Trieste)
Emanuel A. Schegloff (University of California at Los Angeles)
Paul O. Takahara (Kobe City University of Foreign Studies)
Sandra Thompson {University of California at Santa Barbara)
Daniel Vanderveken {University of Quebec at Trois-Rivières)
Teun A. van Dijk (University of Amsterdam)
33
By Word of Mouth
BY WORD OF
MOUTH
METAPHOR, METONYMY AND LINGUISTIC
ACTION IN A COGNITIVE PERSPECTIVE
LOUIS GOOSSENS
PAUL PAUWELS
BRYGIDA RUDZKA-OSTYN
ANNE-MARIE SIMON-VANDENBERGEN
JOHAN VANPARYS
Introduction vii
References 245
It may have become clear by now that, although there is considerable common
ground underlying them, the following studies have their own individual vantage
points, and are only partially intended to form one coherent whole. To bring the
reader's expectations in line with what follows, let us survey the conliibutions
one by one, in the order in which they succeed each other in the volume.
In À survey of metalinguistic metaphors, Johan Vanparys first explains
in some detail how our main database (from LDOCE?) was established, and
INTRODUCTION iX
evaluates its merits and its biases. Taking for granted that by and large it reflects
general usage, he then proceeds to a systematic survey of the source domains
showing up in it. He is able to give an idea of how the diversity of source do-
mains ties up with the diverse complexity of the linguistic action domain. One of
his conclusions is that Reddy's (1979) claim that our language about linguistic
communication is almost exclusively dominated by the conduit metaphor is a
gross overstatement.
In Body parts in linguistic action: underlying schemata and value
judgements, Paul Pauwels and Anne-Marie Simon-Vandenbergen focus on
metaphors of linguistic action whose source domain is the (human) body and its
functioning. As the title of their paper makes clear, they investigate a specific as-
pect of metaphorical extension, namely the expression of value judgements.
They examine various concrete subdomains which are frequently exploited, and
for which, because of this concreteness, the experiential grounding is fairly ob-
vious. Such concrete experiential grounding is not always available, however.
For a considerable number of instances they have to fall back on more abstract
image schemata (in the sense of Johnson (1987)). The way in which the con-
crete subdomains interact with the more abstract image schemata is also ex-
plored. This yields interesting conclusions about scales that are relevant to value
judgements, as well as new insights into image schemata: they emphasize the
importance of a control schema (which is not provided for by Johnson), the pos-
sibility of coming up with a hierarchization among image schemata, and the
overall importance of image schemata in the metaphorical expression of value
judgements.
In more than one respect Anne-Marie Simon-Vandenbergen's paper As-
sessing linguistic behaviour: a study of value judgements is complementary
to the preceding one. Making use of British native speakers' intuitions (which
were obtained via questionnaires), she examines to what extent conventional-
ized metaphors reflect positive or negative evaluations of certain aspects of lin-
guistic behaviour. She finds out that the expression of value judgements is the
rule, and that most of the time they are negative. Her study also sheds light on
the contribution of specific donor domains to the expression of value judge-
ments, where it is revealed, amongst other things, that the value judgement may
be transferred from the source domain directly, or may come in only after the
extension to a specific aspect of the target domain. Also the role of the image
schema of scale in speakers' assessments of linguistic (inter)action is clarified.
The whole is put into further perspective by emphasizing that value judgements
X INTRODUCTION
The research underlying this book has been carried out under the auspices of the
National Science Fund of the Flemish Community, and for two of the authors
(Louis Goossens and Brygida Rudzka-Ostyn) in the context of a research pro-
gram supported by a Belgian government grant (IUAP-II, contract number 27).
It constitutes the continuation of work by the research group which issued The
scene of linguistic action and its perspectivization by SPEAK, TALK, SAY and
TELL (by René Dirven, Louis Goossens, Yvan Putseys and Emma Vorlat; 1982,
also in the series Pragmatics and Beyond, vol. 111:6).
Other work that is the direct or indirect outcome of our collaboration is
reported on in Goossens (1986, 1987, 1989, 1990a, 1990b, 1993a, 1993b and
1994), Rudzka-Ostyn (1988) and (1989), and Simon-Vandenbergen (1991).
Three contributions included here (with slight modifications) have been pub-
lished elsewhere, as Goossens (1990), Pauwels and Simon-Vandenbergen
(1993) and Rudzka-Ostyn (1994). We are grateful to Mouton De Gruyter
(Berlin) and to Peeters (Leuven) for permission to reprint this material.
Each of the authors takes responsibility for her or his contribution(s),
though it must be emphasized that we have greatly profited from each other's
comments. This introduction is the result of combined efforts as well. Finally,
we wish to thank Hans Paulussen (Notre-Dame de la Paix University, Namur)
for sharing his computer expertise with us.
A Survey of Metalinguistic Metaphors
Johan Vanparys
Notre-Dame de la Paix University, Namur
0. Introduction
The goal of cognitive linguistics has been defined as an attempt "to characterize
those psychological structures that constitute a speaker's linguistic ability, i.e.
his grasp of established linguistic convention" (Langacker, 1988: 130). Our
knowledge of language is largely implicit, so that it needs to be explicated by
expert theories. However, there are also intuitive conceptions of language, as
reflected in the metalinguistic repertoire of the lexicon. For instance, the
existence of verbs such as promise, request, assert, apologize, etc., points to
pre-theoretical conceptions about speech acts (cf. Searle 1979, Searle &
Vanderveken 1985). Similarly, hedges like par excellence, loosely/strictly
speaking, technically, etc. indicate our intuitive understanding of linguistic
categorization (cf. Lakoff 1973). Such conceptions are of special relevance to
linguistics, since what we do with language is at least in part determined by
what we think we do (Verschueren 1985a).
It is therefore not surprising that metalinguistic terms have received due
attention in linguistics (eg Ballmer & Brennenstuhl (1981), Dirven et al. (1982),
Leech (1983: ch. 9), Rudzka-Ostyn (1989), Searle & Vanderveken (1985: ch.
9), Vanparys (1993), Verschueren (1985a, (ed.) 1987), Wierzbicka (1985a,
1985b, 1987)). Most studies deal with items that literally designate aspects of
verbal communication. However, "a great many concepts (perhaps most
abstract concepts) are structured metaphorically" (Lakoff 1989: 103). As a
result, metaphorical conceptualizations need to be investigated, as well.
2 JOHAN VANPARYS
1. Previous research
reveal the pervasiveness of THE ARGUMENT IS WAR metaphor, for example (id.:
4):
(7) Your claims are indefensible.
(8) He attacked every weak point in my argument.
(9) His criticisms were right on the target.
This is not the only metaphor in terms of which we frame arguments; alterna-
tives include AN ARGUMENT IS A JOURNEY (id.: 89 ff.) and AN ARGUMENT IS A
BUILDING (id.: 92 ff.). These metaphors bear little if any affinity at all to the
conduit metaphor. They show that, in general, we employ various, sometimes
mutually inconsistent, metaphors to conceptualize the same phenomenon and,
regarding linguistic communication, there exist alternatives to the conduit
metaphor, which focuses exclusively on the encoding-transfer-decoding process.
To conclude, partial or skewed sampling of data may distort the relative
importance attributed to certain metaphors.
More recently, metaphors of verbal communication have been studied by
Rudzka-Ostyn (1988). She has compiled a corpus of 700 examples from various
sources and narrowed it down to the 367 items that involve movement in space.
Within this subset, the primary focus is put on phrasal verbs. A detailed analysis
of these data reveals a large variety of metaphorical extensions from the domain
of spatial motion into that of verbal communication. This study differs in several
important respects from Reddy's. First, the data - although restricted from a
semantic point of view, too - have not been selected on an a priori basis.
Second, the analysis is much more fine-grained. The conduit metaphor operates
with notions such as 'object', 'container' and 'transfer', all of them situated at
the level of abstraction of Johnson's (1987) image schemata. Rudzka-Ostyn
goes into greater detail, adding richer images that have more content. Both
abstract image schemata and rich images play a role in figurative language (cf.
Lakoff 1987: 444 ff.). Finally, Rudzka-Ostyn is aware of her study's limits
("Many [...] areas remain unexplored." (id.: 551)) and counters Reddy's claim
that English is its own worst enemy as a metalanguage by stressing that the
metaphors she has uncovered offer "an extremely powerful means of expres-
sion" (ibid.), which leads to the conclusion that English is "its own best friend"
(id.: 552). (See also Goossens (1994) on the conduit metaphor.)
The study undertaken by Verschueren (1985b), based on a series of news-
paper articles covering the 1960 U2-incident, is restricted not only in scope but
also regarding its objectives (id.: 38). The limitations inherent in such a case
study allow a very detailed analysis of the role played by metalinguistic
4 JOHAN VANPARYS
metaphors in one genre. The conclusions drawn from it, however, constitute but
one piece of a huge jigsaw puzzle that still remains to be assembled.
Other clues can be found in research that is not specifically geared to-
wards verbal communication. Talmy's (1985) work on force dynamics,
elaborated by Johnson (1987: 57-61), shows how the force schema pervades
our thinking about interpersonal relations and communication. Folk beliefs
about emotions and their expression have been studied in detail by Kövecses
(1986) and Lakoff (1987: 380-416).
To conclude, existing research results on metalinguistic metaphors offer
only a partial, somewhat distorted, survey of the topic.
2. Objectives
The aim of this paper is to remedy this situation. An attempt will be made to
construct a representative list of conventional metalinguistic metaphors, which
will then be narrowed down to the set of verbials, that is verbs and longer units
built around a verb (eg phrasal verbs, idiomatic clauses, etc.; cf. Verschueren
1985a).
The general background is the seminal study of Lakoff & Johnson (1980),
which entails a number of commitments. The "essence of metaphor is under-
standing and experiencing one kind of thing in terms of another" (Lakoff &
Johnson 1980: 5). Thus, a speaker who says (1) understands, and consequently
codes, words in terms of containers. In a metaphorical expression, two
conceptual domains are involved: a donor domain and a recipient domain (or
source and target domain). In example (1), the referent of words - an entity in
the domain of verbal communication - is understood in terms of physical
objects, more specifically containers. Very often, the donor domain is more
concrete than the recipient domains, since we tend to understand abstract
phenomena in terms of more concrete things (Johnson 1987). The main
objective of this paper is to identify the range of donor domains serving as the
source for metalinguistic metaphors.
Not anything can be conceptualized in terms of anything else. In order for
a metaphor to be successful, there must be some perceived structural resem-
blance between the two domains. For example, the ARGUMENT IS WAR metaphor
relies on an identification of the speaker (S) with the aggressor, the hearer (H)
with the defendant, the words with weapons, the different stages in an argument
A SURVEY OF METALINGUISTIC METAPHORS 5
with the stages in a fight, etc. (Lakoff & Johnson 1980). For an adequate
understanding of a metaphorical utterance or a set of related metaphors, one
needs therefore a structural description of boh domains involved. The domain of
verbal communication has been described in great detail by Dirven et al. (1982),
Rudzka-Ostyn (1988) and Verschueren (1987b), among others. In this paper,
the ways in which elements from this recipient domain are receive mappings
from various donor domains will be explored.
Typically, a metaphorical mapping between two domains is instantiated by
a whole cluster of examples; the number of instantiations is to some extent
indicative of a metaphor's cognitive salience. Thus, the evidence adduced by
Lakoff & Johnson for the ARGUMENT IS WAR metaphor and the numerous
illustrations provided by Reddy for the conduit metaphor point to a high degree
of salience. Contrary to Reddy and Lakoff & Johnson, the material on which
this paper relies is corpus-based. Since the examples are taken from a corpus-
based general-usage dictionary, the analyst's creativity in making up metaphori-
cal expressions plays no role, which lends more plausibility to the pervasiveness
of the metaphors under investigation. '
3. The data
In principle, there are several ways of constructing a corpus. One might, for
instance, make up one's own examples. This is a useful procedure if the aim is
to verify/falsify a given theory, but the potential pitfalls are obvious: one risks
coming up with artificial examples, self-concocted items reveal nothing about
their relative salience, one cannot be sure that important metaphors have not
been overlooked, etc. Another possibility is to scan texts for metalinguistic
metaphors. This will yield a database with rich contextual information. In order
to arrive at a representative sample, however, one would have to go through a
huge amount of text representing different genres, since metalinguistic
metaphors do not have a high frequency. A text-based corpus, therefore, is only
efficient for case studies with a restricted scope (eg Verschueren (1985b)).
Authenticity and efficiency can be combined in a third option: a diction-
ary-based corpus. Indeed, every dictionary contains a wealth of material that is
relevant to the issue at hand. Moreover, the items found in a general-usage
dictionary constitute a representative sample. They cover all types of language
use and they are not restricted to one particular topic. In addition, conventional-
ized metaphors, when compared with novel constructions, are of special rele-
6 JOHAN VANPARYS
vance (cf. Traugott 1985). When looked at from a diachronic viewpoint, they
may be said to have been more successful. In order for a new lexical item or a
novel sense of an existing item to penetrate the lexicon, it must be used to a
considerable extent by a sufficient number of speakers. In order to stay there, it
must moreover stand the test of time. From a synchronic point of view,
conventionalized metaphors are part of a language's lexicon to the extent that
they have acquired unit status. They are used automatically and spontaneously,
without any constructive effort on the part of S. In sum, it is the dead meta-
phors that we really live by (cf. Mackenzie (1985: 69)), and many of these can
be found in a dictionary.
The corpus serving as the basis of this paper and several other contribu-
tions to this volume is taken from the Longman Dictionary of Contemporary
English (LDOCE). It was compiled in two stages. First, the contents of the
dictionary were narrowed down to entries that involve aspects of verbal
communication. Part of this work was done by computer. This database was
then split up into literal and figurative (metaphorical and metonymic) expres-
sions.
LDOCE was chosen for several reasons. It is a corpus-based general-
usage dictionary representing ordinary, present-day English. Though British
English is the main locus of attention, other varieties are taken into account as
well. The feature that renders LDOCE the ideal dictionary for our purposes is
its restricted defining vocabulary: lexical items are defined by means of a set of
about 2,000 defining words. We hypothesized that definitions of metalinguistic
items (literal and figurative) would include one or more metalinguistic defining
words themselves. Consider the following examples (my italics, J.V.):
accuse to charge (someone) with doing wrong or breaking the law
beg 1 to ask humbly for (food, mone, or other necessary things) 2 to ask
humbly (something not material)
exclaim (mostly used with the actual words of the speaker) to say sud-
denly, because of strong feeling
prophesy 1 to give (a warning, statement about some future event, etc.) as
a result of a religious experience directed by God or a god 2 to say in ad-
vance
shout to give a loud cry (of); speak or say very loudly
LDOCE defines accuse in terms of charge, beg is explained by means of ask,
the definition of exclaim includes terms such as words, speaker and say, the
paraphrase of prophesy includes warning and statement and shout, finally, is
defined on the basis of cry, speak and say. The hypothesis was tested on the
A SURVEY OF METALINGUISTIC METAPHORS 7
basis of the 240 speech act verbs listed in Wierzbicka (1987: 395-397) and
largely confirmed. Most of these verbs are defined in LDOCE by means of one
or more metalinguistic terms. Those that are not turn out to be only marginal
speech act verbs.
(i) Some do not necessarily involve verbal communication:
• calculate, conjecture, compare, conclude, credit, deduce, estimate,
guess, presume, prove, reckon, speculate, suppose, and suspect belong
more to the epistemic than the metalinguistic domain;
• agree, approve, blame, deplore, disapprove, object and resolve refer to
mental states and actions that can but need not be expressed overtly;
• blackmail, dissuade, ridicule and reassure denote actions that can be
brought about by non-linguistic means;
• counter profiles retaliative acts in general, not only those that are lin-
guistic.
(ii) Institutional declarations (acquit, appoint, book, convict, excommunicate,
reserve, resign, sentence) represent "a very special category of speech acts"
(Searle 1979: 18-19). They require a set of extra-linguistic, institutional
rules for their successful performance. Thus, they are situated on the area
of overlap between linguistic and institutional action. Moreover, they are
not essentially communicative, which sets them apart from central speech
acts (cf. Bach & Harnish 1979: 117). Farewell features only as an interjec-
tion in LDOCE, not as a verb.
(iii) The rest of the verbs listed by Wierzbicka but not defined in LDOCE by
means of metalinguistic items are:
bemoan to be very sorry because of
bet to risk (money) on the result of a future event
criticize 1 to find fault with (someone or something); judge severely
2 to make judgments about the good and bad points of (someone or
something)
enthuse to show enthusiasm
exult to rejoice; to show delight (in victory, success, etc.)
justify to give a good reason for
note to call attention to; make known; show
point out to draw attention to (something or someone)
The definitions given for bemoan and criticize are somewhat inaccurate.
They make reference only to one's having a certain feeling/opinion, whereas
other dictionaries indicate that these psychological states are also ex-
8 JOHAN VANPARYS
pressed. The paraphrases for bet, note and point out make no explicit refer-
ence to verbal communication. The element show in the definitions of en-
thuse and exult refers to the exteriorization of emotions, without however
putting the use of language into focus. Justify, finally, is defined in a meta-
phorical way by means of give.
These verbs reveal a disadvantage of LDOCE: the commitment to a
restricted defining vocabulary sometimes results in poor definitions. The
Collins COBUILD English Language Dictionary, for instance, makes ex-
plicit reference to verbal communication for seven out of the eight re-
maining verbs. The exception is bet, a notorious troublemaker in speech act
theory (cf. Fotion (1981)). All in all, however, it seems justified to rely on
the LDOCE definitions for drawing up a list of metalinguistic lexicaliza-
tions. In quantitative terms, the loss - 8 items out of 240 - is negligible.
Moreover, the loss is random, the absent items do not go together semanti-
cally, so that the representativeness remains unaffected. In our analyses, the
poverty of the LDOCE definitions has been compensated for on the basis of
other dictionaries.
An additional source of information is the example sentences given in LDOCE.
Consider the following items (abridged from LDOCE):
airy (derog) having little substance; empty Nothing results from his airy
plans and promises
hover (of people) to wait around one place (fig) A question hovered on his
lips
mingle to mix (different things) together a speech that contained praise
mingled with blame
plaster to put wet plaster on (fig) They plastered over the difficulties with
fine words
spice interest or excitement, esp. as added to something else a few good
stories to add spice to the speech
Though the definitions make no explicit reference to verbal communication, the
examples clearly show the item's relevance. The import of such evidence should
not be underestimated. The point of example sentences in dictionaries,
especially those aiming at a non-native public, is to illustrate typical usages. The
presence of metalinguistic metaphors in the example sentences, especially those
labelled 'fig' indicates therefore that they are to some extent conventionalized;
they are on their way to acquiring unit status. Moreover, LDOCE is to a great
extent corpus-based.
A SURVEY OF METALINGUISTIC METAPHORS 9
barrage (of speech or writing) a large number of things put forward at al-
most the same time or very quickly aftert he other a barrage of questions
blurt out to say (something which should not be said) suddenly and with-
out thinking Peter blurted out the news before he considered its effect
breath: waste one's - to talk uselessly, without effect
The corpus takes the shape of an alphabetical list of dictionary entries. In order
to improve semantic accessibility, it was indexed on key-words (metalinguistic
and other) occurring in the definitions. In this way, lists were generated by
computer of items defined in terms of the same defining word. Thus the defining
word secret yielded semantically related lists such as the following, which reveal
the metaphor TO MAKE KNOWN IS TO MAKE VISIBLE:
leak to make known (news, facts, etc. that ought to be secret)
leak out (of news, facts etc. that ought to be secret) to become known
let the cat out of the bag to tell a secret (often unintentionally)
spill to let out or tell (secret information)
spill the beans to tell a secret too soon or to the wrong person
Describe yields a cluster illustrating the same kind of extension from the visual
to the communicative realm, containing chalk out, delineate, depict, picture,
portray, and sketch. Under the defining word interrupt, several verb+in(to)
combinations are listed (barge into, break in (on), break into, burst in on, butt
in, chime in, cut in, get in, horn in, push in, put in) revealing the metaphor A
CONVERSATION IS A CONTAINER. These lists proved to be useful instruments for
bringing together semantically related items that were scattered across the
alphabetically arranged corpus.
Very often, linguistic expressions are reified, not as containers for ideas, but in
some different fashion. As objects, linguistic expressions can undergo all sorts
of manipulations:
Impaired visibility, as instantiated by fog and throw dust in s.o. 's eyes, re-
sults in a lack of understanding or misunderstanding. When visibility is restored,
things becomes 'clear': clear up (eg a mystery), clear the air.
In sum, a whole range of verbials present linguistic expressions as objects
that can be produced (4.2.1), adorned (4.2.2), expanded or compressed (4.2.3),
altered in some other way (4.2.4) or perceived (4.2.5). Unlike the conduit
metaphor, these reifications do not specifically involve the container schema
posited by Reddy. These cases indicate that the conduit model is not the only
framework in terms of which we conceptualize (or even reify) linguistic
expressions.
not involve transferral of ownership. Moreover, the object need not pass from
hands to hands. However, the elements of the prototype and their interrelations
correspond quite closely with those involved in a typical speech-act scenario:
the two participants (speaker/donor, hearer/recipient, S entertains a conceptu-
alization (owns an object) and assumes a desire on the part of H to become
informed of it.
An example like (29) is fairly neutral with respect to these parameters of
variation. Here, give simply profiles a donor (subject), an object (direct object)
and a recipient (indirect object). Ownership does not seem to be involved, the
exact location of the object (in the hands or elsewhere) is irrelevant. However,
the principal participants are there: speaker/donor, the hearer/recipient, and the
idea/object. In addition there is willingness on the part of S to give the
information (compare he gave her the number with he let slip the number) and
a desire on the part of H to receive it.
(32) No one in the class could furnish the right answer to the question.
(LDOCE)
(33) ... the unreliable gossip that had been peddled for so long ...
(COBUILD)
Other verbials highlight specific elements not predicted by the conduit meta-
phor. Deliver (a lecture/speech) and furnish (eg an answer) place the act of
giving in a commercial environment, where the donor satisfies a demand. In
(32), this element is transferred into the domain of verbal communication in a
classroom setting, where the teacher indicates a desire to be told something.
Peddle (33), on the other hand, derives its pejorative connotation from the
negative value judgments our culture stereotypically places on peddlers.
(34) I didn't feel like giving away more information than I had to.
(COBUILD)
(35) The examiners have given away the answer. (LDOCE)
(36) to give the game/show away (= 'to tell a secret5)
Verbs of giving combined with away focus on a different aspect. Away profiles
movement of an entity from its initial location into another, unspecified region.
In literal giving away, this is carried over into a loss of control over the object in
question. When used as a metaphor of verbal communication (34-36), lack of
intentionality is highlighted.
Grant (37) and offer (38), on the other hand, are motivated by a desire on
the part of the recipient/hearer to receive/be told something:
(37) I had to grant him the reasonableness of his argument. (LDOCE)
A SURVEY OF METALINGUISTIC METAPHORS 19
position {lay down conditions, lay down the law). These examples are subject to
two different, though not mutually exclusive, explanations. For one, they may
involve the placement of documents on a table, for instance by a person filing an
official complaint. This is the preferred interpretation for table (eg a report). On
this account, the metaphors have a metonymic origin. In a more abstract
interpretation, the object is of unspecified nature, undergoing downward move-
ment instigated by S, ending within the immediate proximity of H and resting
there. The downward movement is then explicable in terms of the superordinate
position of S vis-à-vis H (equally in hand down). A difference between laying
down something and giving it to a person resides in the terminal stage of the
process. In a prototypical act of giving, the recipient accepts the object. That is,
the object ends up in the recipient's hands. When I lay it down (in front of him),
however, it may take some time before he takes it up - or he may refuse to
accept it at all. Indeed, official claims, charges etc. are not dealt with on the spot
and may be subject to an acceptance procedure.
(46) put a question
(47) Let me put to you a practical problem. {LDOCE)
(48) The organization put forward eight candidates for the NUS execu-
tive. {COBUILD)
(49) They rejected every proposal put forward. {COBUILD)
(50) I put in a request for an interview {COBUILD)
(51) not to put too fine a point
(52) put it briefly/simply/mildly ...
Put is frequently used in connection with verbal communication (cf. Pau-
wels, this volume). It shares with lay the fact that in the terminal stage the
object rests at a certain location. Examples (46)-(50) all involve speech acts that
require further action on the part of H, in the form of providing an answer (46),
working out a solution (47), considering a proposal (49) or a request (50). The
verb put profiles these speech acts as 'waiting' to be acted on. Regarding put
forward, a similar interpretation as for bring forward can be given: in bringing
the object/message to the fore, S makes it readily accessible.
ideas come or get across, or if one manages to get or put them across or get
them through, or if one's speech goes over well, the object/message equally
reaches its destination, but this time after overcoming some obstacle (across,
over) or following some predefined trajectory (through). In other words, across,
over and through indicate successful communication involving some effort on
the part of S, owing to the presence of resistance.
these cases, the destination of the word-objects does not correspond with H, but
with the person positively or negatively affected by the speech act. Several items
evoke more than mere patient-directed transfer. For instance, when one lays or
puts the blame on someone, the speech act places a burden on him, something
that is fixed may be hard to detach, while heap and pour imply abundance.
In sum (4.3-4.5), verbials indicating information transfer suggest various
qualifications of the conduit metaphor. First, the notion 'conduit' itself is
virtually absent from the database: despite the large number of expressions
denoting transfer of information, this process is rarely presented as taking place
through the kind of pipeline suggested by Reddy. (See also Goossens 1994).)
Second, though the verbials in question do introduce ideas as objects
(reification), most of them can be understood without making reference to
words-as-containers. Third, in most cases the image evoked by an expression is
much richer in content than suggested by Reddy's account, so that other, more
specific conceptual elements need to be drawn into the analysis.
Not only linguistic expressions, but also their source (S) and target (H) are
often presented as containers. Thus, the corpus includes numerous combinations
with out: bleat out, blubber out, blurt out, bluster out, boom out, dish out, give
out, grind out, hand out, ladle out, let out, pour out, put out, roll out, spill out,
spout out, thrash out, throw out. This particle designates movement of an entity
away from its initial location, where it was enclosed in a container. In the
examples, S is therefore conceptualized as a container where information is
stored in reified form. Speaking, then, consists in releasing the idea-objects.
(75) The date of the election will be given out soon. {LDOCE)
(76) He put out a statement denouncing the commission's conclusions.
(COBUILD)
(77) throw out (= 'to say s.th. carelessly and without considering the re-
sult')
(78) He likes dishing out good advice. {LDOCE)
(79) the knowledge that is ladled out daily in high schools {COBUILD)
(80) I don't need you handing out that sort of advice. {LDOCE)
(81) You were always spouting some theory to us at the table.
{LDOCE)
Give out (75) and put out (76) refer to this process in a neutral fashion, while
the other verbials add extra dimensions. The majority focuses on the acoustic
A SURVEY OF METALINGUISTIC METAPHORS 25
aspect of speech: bleat out, blubber out, blurt out, bluster out, boom out, roll
out. Let out (eg a cry of pain) and spill out (eg a secret) highlight the uninten-
tional nature of the speech act. Throw out (77) evokes the negative result of
speaking carelessly, motivated by the destructive action of projectiles. Dish out
(78), ladle out (79) and hand out (80), on the other hand, focus on the authori-
tative position assumed by the speaker-distributor and the abundance of his
output. As already explained above, grind out and thrash out evoke monoto-
nous abundance, by evoking S as a machine. The same effect is produced by the
fountain-metaphor spout (81).
The speaker-as-container metaphor is instantiated not only by verb + out
combinations, but also by other verbs requiring a container-like subject (82) or
object (83):
(82) John emitted a few curses. (LDOCE)
(83) A whistle of surprise escaped him. (LDOCE)
A different realization of the same metaphor can be found in verbials profiling a
person trying to extract an object from a speaker-container. Most of them
indicate unwillingness on the part of S to release information and a concomitant
effort on the part of his interlocutor:
(84) The truth had to be dragged out of him. (COBUILD)
(85) It was impossible to draw the truth out of him. (COBUILD)
(86) Sir James had extracted from Francis a fairly detailed account.
(COBUILD)
(87) She pretended that she couldn't hear, so that they would not force
an answer from her. (LDOCE)
(88) I managed to pry out of the adults the reason for her disappear-
ance. (COBUILD)
(89) I pumped him discreetly about his past. (COBUILD)
(90) The truth had been wormed out of him by his lawyers.
(COBUILD)
A speaker-container is also evoked by verbials of opening and closing,
designating unimpeded speech and termination respectively:
(91) She found it difficult to open out to people. (COBUILD)
(92) He was impressive on paper, but he completely clammed up in the
interview. (COBUILD)
(93) Shut up.
H is metaphorized as a container by verb + in(to) combinations and other
verbials requiring a container-like object:
26 JOHAN VANPARYS
4.9.1 Respiration
To speak is to modify one's breath in such a way that one produces conven-
tionalized sounds. In other words, respiration is part of the speaking process.
This gives rise to a number of metonymic expressions involving respiration or
bodyparts used in speech production (the mouth, lips and tongue). Given the
metonymic relationship between breathing and producing speech sounds,
verbials indicating manner of respiration can be used to indicate manner of
speaking. The most neutral verb here is breathe, in the sense of 'to whisper'.
A SURVEY OF METALINGUISTIC METAPHORS 29
Other verbs add extra dimensions, following the pattern 'to V = to say by/as if
V-ing' (to gasp (out), to pant, to puff, to snivel, to snort, to wheeze). The notion
of breath as a limited resource is highlighted by save/waste one's breath. The
expression take one's breath awayrefers to one's inability to speak.
Of course, the primary function of the mouth is to take in food. Its func-
tion as the interface between the digestive system and the outer world gives rise
to a cluster of metaphorical expressions where speaking is conceptualized as the
exteriorization of words, either in the form of food or of saliva.
In the case of regurgitate (as in the case of a student regurgitating what
the teacher said), S is represented as exteriorizing undigested food, which
stands for words he does not understand. The underlying metaphor equates
understanding with digestion. In the case of drool and slobber, on the other
hand, S is portrayed as exteriorizing saliva, which we usually associate with
foolish behaviour.
Pour ('Curses poured from his lips' LDOCE) represents the words as an
excessive amount of some liquid substance. Whether it is food, saliva or some
other specific substance remains indeterminate. Similarly, spit (it) out and cough
(up) characterize the words as an indeterminate substance that clogs the
digestive or respiratory channel. (Bodyparts as a donor domain are discussed in
greater detail by Pauwels & Simon-Vandenbergen, this volume.)
s.o., make mincemeat of s.o., needle s.o., rub s.o.'s nose in the dirt, pull/tear
s.o. apart/to pieces, pulverize s.o., cut s.o. to the quick, round on s.o., run s.o.
down, slam s.o., slash s.o., tackle s.o., take/tear s.o. apart, tear into s.o., tear
s.o. off a strip, tear s.o. to shreds, cast/throw in s.o. 's teeth, tilt at s.o. and wade
into s.o.
Other verbials do situate the aggressive speech event in a context of war-
fare, eg campaign/crusade (eg for a cause), cross swords with s.o., give/lose
ground, hold/shift one's ground, bring up one's big guns and stick to one's
guns.
For a related category of aggression metaphors, the donor domain is tor-
ture or punishment: crucify, drag/haul s.o. over the coals, excoriate, flay, grill,
keelhaul, pillory, rack, give s.o. a (good/real) roasting.
4.12 Sounds
same expressions are used to describe speech events where the designated
gesture is not made. (See also Goossens, this volume.)
In quite a number of cases, the non-verbal act of communication qualifies
as a more or less ritualized event in an institutionalized setting, such as duels
(throw down/take/pick up the gauntlet) performances (applaud s.th., blow one's
own trumpet/horn), hierarchies (bow to s.o., bow down the neck to s.o.,
bow/bend the knee (to) s.o., fall/go on one's knees), religion and supersition
(beat one's breast, cross one's heart and hope to die), the navy and merchant
navy (lower/haul down one's colours/flag, nail one's colours to the mast, sail
under false colours).
Even within the domain of verbal communication, extensions are possible
between various subdomains. For instance, the following verbials require a
religious setting in their literal senses but they can also be applied to describe
other speech events: say amen to, cast aspersions on, damn, pronounce s.o.'s
doom, speak in parables, pontificate, pray, preach, prophesy, swear.
make it drink, parrot, pick over, pick people's brains, rabbit, trot out, talk
turkey, ...
5. Conclusion
Acknowledgement
The LDOCE computer-tape was kindly put at our disposal by Prof. Em. Dr.
L.K. Engels.
Body Parts in Linguistic Action
Underlying Schemata and Value Judgements
Paul Pauwels
Catholic Flemish Institute of Higher Education
Anne-Marie Simon-Vandenbergen
University of Ghent
0. Introduction
In this paper we will take a closer look at metaphors which originate in the
domain of 'body parts' 1 . These are metaphors which either contain explicit
names of parts of the body or refer to processes in which body parts are implicit
but play a central role. Our material consists of 175 items, selected from the
corpus of linguistic action metaphors described by Vanparys in this volume.
The view of metaphor we subscribe to has been outlined in Lakoff and
Johnson (1980) and Johnson (1987). In this view, metaphor is not merely an-
other, non-congruent way of referring to phenomena (processes and partici-
pants), but rather an important mode of understanding and a way of structuring
experience. The concepts used in metaphorisation are "concepts for natural
kinds of experience and objects [which] are structured clearly enough and with
enough of the right kind of internal structure to do the job of defining other
concepts. [These other concepts] are less concrete or less clearly delineated in
their own terms" (Lakoff and Johnson 1980: 118). One important function of
metaphor is then to structure abstract domains by means of projections from
more concrete domains. The domain of body parts and bodily functions is one
such structured concrete domain. The study of metaphors of linguistic action
which refer to parts of the body and their functioning may thus contribute to a
clearer understanding of how physical experience is projected onto linguistic
action. More specifically, the aims of this study are the following.
36 PAUL PAUWELS & ANNE-MARIE SIMON-VANDENBERGEN
Through linguistic action (henceforth LA) human beings interact both with one
another and with their environment. However, LA is only one type of action and
interaction, and other types are more concrete; they involve the body more
'tangibly'. It is therefore logical according to our view of metaphor that the
latter types of interaction should be used as donor domains for the metaphori-
sation of LA.
In what follows we will present a survey of the major types of donor do-
mains involving body parts and outline how they give structure to the LA do-
main.
scribed as useless because it does not have an effect; it does not pay off. In
cough up, the focus is on the removal from the body of something (blood,
phlegm) which causes discomfort. While breathing involves both in- and exha-
lation, the directionality of coughing is obviously an outward one. The domain
interacting with that of breathing is clearly that of illness, and relief can be at-
tained by pushing air out from the throat. The parallel with speech is to be
sought then not only in the use of air, but in the removal of some disturbing
substance from the body (donor domain) and from the mind (LA domain). A
relevant example is: "Even after the police threatened him, Smith refused to
cough" {LDOCE). In choke back, the speaker prevents his feelings from escap-
ing his body by holding them in the throat. The metaphor describes controlled
silence.
Other metaphors focus on silence that is beyond the control of the
speaker. For instance, take someone's breath away ('make unable to speak'
{LDOCE)) attributes the inability to speak to an external agent.
Use of articulators
The metaphor drool ('talk foolishly' (LDOCE)), in which the mouth is neces-
sarily involved, describes a natural reflex which human beings should have learnt
to control. The metaphor brings in three elements: the speaker's behaviour is
described as uncontrolled, the substance brought out is characterized as val-
ueless, and the action as a whole as distasteful and socially unacceptable. In the
metaphor spit out 'say or express with effort, force or anger' {LDOCE), the
value of the substance (whether it be food or phlegm) and the unacceptability of
the action are likewise meaning elements.
A number of expressions focus on the movement of the visible speech or-
gans, and clearly have a metonymic basis:5 keep one's mouth shut, open one's
lips, and close lipped describe presence or absence of speech; lip, chinwag and
jaw describe a type of LA, and are more clearly metaphorical. Other examples
are tongue in cheek and lie in one's teeth/throat. The listener's role is focused
on in turn a deaf ear to, go in one ear and out the other and have a word in
someone's ear. (Although ears are not articulators, they are salient body parts
involved in oral communication). All of these metaphors seem to rely exclu-
sively on the body part domain, in contrast to the majority of instances in our
corpus, where the body part domain plays only a secondary role in the meta-
phorisation, by being the chief instrument in other kinds of (inter)action.
BODY PARTS IN LINGUISTIC ACTION 39
off and tear apart) focus on the violent interaction between animals. The vio-
lence is often extreme, and in the metaphorisation the focus is, in most cases, on
the unmotivated nature of the action. The LA is described as disproportionately
hostile under the circumstances. The agent is overreacting, not controlling him-
self. Backbite describes a LA in which the victim is defenceless because of his
absence. Bite merely describes the effect of the LA on the hearer in terms of a
painful sensation. (The example from LDOCE is: "His words had a cruel bite to
them which we all felt keenly.")
Restricted movement
The domain of restricted movement gives rise to metaphors such as: tonguetied,
muzzle, bridle one's tongue, guard one's tongue and (less unambiguously)
watch one's words6.. The first metaphor describes the inability to speak, the
second one the imposing of silence, while the other three describe self-control.
In muzzle and bridle one's tongue the donor domain is clearly that of animal
control, while in the others the donor is less specific. All metaphors describe
silence or carefully controlled speech. The movement of the tongue as an articu-
lator is a prominent element in three of the metaphors.
Manipulating objects
Another important donor domain is that of manipulating objects. The aspects in
focus are the (lack of) skill of the manipulator (fumble, shoot one's mouth off),
his physical capacities (heavyhanded, a lefthanded compliment), or linguistic
action as exchange of information (hand it to someone).
The metaphor fumble has three slightly different uses, which are illus-
trated by the following examples:
(1) He often has to fumble for the right word. (LDOCE)
(2) [We] are made to feel inferior if we fumble an unusual word.
(Webster's)
(3) What I'm fumbling to say is that I felt different about you.
(COBUILD)
(4) He fumbled in answering and made them suspicious. (Webster's)
In (1) the act is clearly that of looking for an object. The metaphor says that the
speaker is unable to produce a word, and this inability is put down to word-
finding difficulty. Words are conceived as objects, which are stored away and
have to be produced quickly if needed. In (2) the situation is different. The
focus is in this case on the maltreatment of the object. The word is produced,
but as a result of clumsiness, it comes out wrong. The descriptions in (3) and
BODY PARTS IN LINGUISTIC ACTION 43
(4) pay attention to the clumsiness of the speaker, rather than to the manipula-
tion of the objects. This clumsiness is of course the common ground in all these
examples, but the first two stand out in that they allow a more specific transfer
from the donor domain to the LA domain.
The metaphor shoot one's mouth off is still more specific. The mouth is
described as a gun, which in this case causes harm because it is handled care-
lessly or unskilfully. In shooting off a gun, one wastes a bullet and one attracts
attention; one might also hurt someone. Similarly, when one shoots one's mouth
off, one talks 'foolishly about what one does not know about or should not talk
about' (LDOCE).
Pick a quarrel and split hairs are instances of skill used wrongly, in the
first case because of the domain of application, in the second because of the
excessive and useless nature of the action.
Walking
A final domain is that of walking, used to describe the LA as a movement of the
speaker. Again, different metaphors focus on different aspects. Still, the ele-
ments of skill and efficiency seem to play a role in all of them. In backtrack the
aspects in focus are the direction of the movement, and the fact that this is the
consequence of a mistake. The LA is shown as a correction of an error, which
involves a movement opposite to the normal direction of speech. In put one's
foot in it, and never put a foot wrong, the focus is mainly on the (un)skilful
handling of the body. In the first metaphor there is also the aspect of stepping
into some substance which is obviously to be avoided.
1.2.0 Introduction
Following Johnson's (1987) approach, we will now describe the metaphors
from another angle. According to Johnson, metaphors are based on our abstract
bodily experience of the world, which we translate into basic schemata. He
defines a schema as "a recurrent pattern, shape, and regularity in, or of, [our]
ongoing ordering activities. These patterns emerge as meaningful structures for
us chiefly at the level of our bodily movements through space, our manipulation
of objects and our perceptual interactions." (Johnson 1987: 29). Johnson dis-
tinguishes several such basic schemata, and he provides a list (1987: 126) which
is said to be highly selective, but contains those schemata which Johnson thinks
44 PAUL PAUWELS & ANNE-MARIE SIMON-VANDENBERGEN
most pervasive. We have included the list here, and will refer to it in the discus-
sion:
him as using force to keep the container closed. The emphasis differs from
metaphor to metaphor. In cast/throw something in someone's face/teeth, the
intensity of the force is salient; in have a word in someone's ear, the actual entry
is in focus; get an earful pays attention to the lack of control on the part of the
experiencer, while rub in, force/thrust/ram something down someone's throat,
shut someone's mouth combine all three aspects.
It may be noted that the metaphors describing entry are not just restricted
to the normal way of entry in LA (the ears), but also use the mouth as an en-
trance, which links up with the donor domain of eating, or in this case
force-feeding. Tear into and rub in stand apart in this respect. Tear into
('criticize very strongly' (COBUILD)) does not describe the locus of entry; it
simply focuses on the destructive violence involved in effecting entrance. In rub
in, the nature of the force - not extreme, but repeated - is at stake. The defini-
tions in different dictionaries stress different aspects of this metaphor: 'say re-
peatedly' (LDOCE), 'force a lesson into someone's mind' (OALDCE), 'remind
someone of something he does not want to be reminded of, usu. because it is
embarrassing' (COBUILD). The LDOCE definition focuses on the repetitive-
ness of the rubbing, the OALDCE definition combines the force used in the
process with the description of the intended effect (lesson), while the COBUILD
definition refers to the side effect on the hearer (embarrassment).
The extent of entry into the container may vary, as can be seen from put
words into someone's mouth and force/ram/thrust something down someone's
throat. The extent of entry here correlates inversely with the subtlety of the
agent's action. Furthermore there is another difference: whereas the second
metaphor is prototypical of forced entry in that it describes a LA in which the
speaker controls the effect on the hearer, the first metaphor draws attention to
the hearer as speaker in another LA which either follows ('to tell someone what
to say' (LDOCE)) or precedes ('to claim, falsely, that someone has said a par-
ticular thing' (LDOCE)). In either case, the focus is on the control of someone
else's speech.
Yet another use of the CONTAINER image describes the LA as the removal
of something from a container: pick someone's brains, brainwash, take the
words out of someone's mouth, take one's breath away. The actions violate the
self control of the affected person. His words, ideas or capacity to speak/act are
influenced by an outsider. In pick someone's brains the interaction described is
that of eliciting information. The LA is conceived as an act of stealing, but be-
cause of the skilfulness the experiencer does not notice this (cf. picking some-
BODY PARTS IN LINGUISTIC ACTION 47
one's pocket). The information, ideas, etc. stored in the brain are looked upon
as valuable possessions, which should not be interfered with8. The same applies
to brainwash.
A final use of the CONTAINER schema is found in poke one's nose into
something and horn in; here an ongoing process is described as a container, and
the LA as an unacceptable attempt at entering, because it is construed as a
violation of private territory. In horn in, this violation is combined with the use
of force.
To conclude this section, let us survey the role of the body parts in these
metaphors. They function as parts of the container (especially exit and entry:
lips, mouth, teeth, skin), as contents (heart, breath), as containers (brain, mouth,
bosom, cheek, throat, teeth) and as instruments used in the exertion of force
(hands, implicitly in most cases, and horns).
7.2.2 Force
According to Johnson (1987: 42), the FORCE schema is necessarily involved in
our interaction with the environment. Our body exerts and/or undergoes force in
any kind of (inter)action. In normal circumstances, however, our awareness of
these forces is very low, because we take them for granted, and we only notice
them "when they are extraordinarily strong, or when they are not balanced off
by other forces." (Johnson 1987: 42).
Such experiences of force should have the required salience to be used in
metaphorisation. It is therefore not accidental that a large group of instances in
the corpus actually have violent actions or fighting as a donor domain. In
these cases force is clearly in focus, either because of its extreme nature, or
because of the opposition of forces from different sources. The elements of the
FORCE schema which are most salient in that context are the degree of force
and/or the interactional character. In many cases, however, the FORCE schema is
further specified in different ways through the combination with other schemata
such as PATH, CONTAINER, and BALANCE.
Compare the following:
(5) He had studied very little and has no reason to kick about low
grades. (Webster's)
(6) He is always kicking against the system. (COBUILD)
(7) He kicks his children around a good deal. (Webster's)
48 PAUL PAUWELS & ANNE-MARIE SIMON-VANDENBERGEN
In the first two examples, the metaphor kick is shown to rely exclusively on the
FORCE schema, though the emphasis differs: in (5) the intensity of the force is at
stake, in (6) there is specification of a target as well. Example (7) combines the
FORCE schema with the PATH schema. The movement of the experiencer ("his
children") becomes a relevant element.
An example of the combination of FORCE with the CONTAINER and
BALANCE schemata can be found in:
(8) The widow sat fuming and blowing off steam. (OED)
Here, the FORCE schema plays a secondary role, as the focus is clearly on the act
of restoring balance by relieving pressure in the container of the body.
The different types of FORCE schemata distinguished by Johnson were
found in our corpus. Example (7) above illustrates the COMPULSION schema, as
is clear from the combination with the PATH schema. The experiencer becomes
subject to the force and loses control of his movement. The same explanation
would apply to a metaphor like boot out. Examples of BLOCKAGE can be found
most clearly in combination with the CONTAINER schema. Here, the force is used
to effect either exit from the container (spit out) or entry into it
(force/ram/thrust something down someone's throat). Clear examples of
COUNTERFORCE were not found in this corpus, although muzzle could be ex-
plained in those terms, and perhaps also bite one's lips. An example of
REMOVAL OF RESTRAINT is found in bang out ('to write in haste' (LDOCE)).
The image of FORCE is again combined with the CONTAINER schema.
1.2.3 Path
A number of metaphors from various concrete donor domains can be explained
in terms of the PATH schema. From the domain of walking we have the metaphor
backtrack, which focuses on the directionality of the movement along the path.
In some metaphors from the domain of eating, the path becomes relevant be-
cause of the opposition between the directionality inherent in eating and that
inherent in speaking. Examples are found in metaphors like eat one's words and
eat crow, which are instantiations of a more general image of 'taking back one's
words'.
Earlier on we discussed some other combinations of the CONTAINER and
PATH schemata, where the path referred to the movement in the container (bite
back, force/ram/thrust something down someone's throat). In these, as in digest
and regurgitate, the image schema of CENTRE-PERIPHERY also plays a role. In
BODY PARTS IN LINGUISTIC ACTION 49
bite back, the path is cut off at the periphery, in force/ram/thrust something
down someone's throat an attempt is made to reach beyond the periphery (oral
cavity); regurgitate describes how the substance did not succeed in reaching the
centre or in staying there long enough; and digest says that it did.
Finally, the concept of removal, which operated in some of the cases dis-
cussed under the CONTAINER schema, also involves a PATH schema (pick some-
one's brains, take the words out of someone's mouth). In a number of other
metaphors, the schema plays a similar role, but the CONTAINER schema is absent.
Tear a strip off someone combines PATH and FORCE, as does get something off
one's chest. These exemplify several types of removal, however. In the former
case it violates the control of someone over his property or domain. In the latter
instance, removal involves a reassertion by the agent of his self-control. Johnson
(1987) would describe the second metaphor as an example of BALANCE, more
specifically of freedom from external pressure.
ronment or with other human beings. A metaphor like slip of the tongue may be
explained as loss of control, drool as lack of control, while muzzle involves the
control of one actor over another. A human being tries to control himself and
his environment in several ways, and this is a precondition for being able to
make use of that environment. In fumble, for instance, there is a clear absence
of control over the object which the agent tries to manipulate.
The CONTROL schema of course very often interacts with the FORCE
schema where the control of outside agents is involved. Going through the
corpus again, we found that a lot of metaphors can be explained from this angle:
pick someone's brain, brainwash, take the words out of someone's mouth, put
words into someone's mouth all violate the self-control of the experiencer, talk
over someone's head can be construed as language passing the hearer by beyond
his controlled environment. In one of its interpretations, the metaphor mouth-
piece relies on the lack of control of the speaker over the content of his utter-
ances: he is only the channel through which the information passes. (See also
2.2.2. below).
The prototypical instantiations of the BALANCE schema, too, can be rein-
terpreted in terms of SELF-CONTROL. Metaphors like get something off one 's
chest and blow off steam describe the reassertion of self-control after an experi-
ence of outside or inside pressure. A metaphor like cough (up) describes an
uncontrollable action, a temporary loss of self-control. Since coughing can also
be interpreted as having a liberating effect, as taking away internal pressure, it
can also be a restoration of self-control. The context will decide which interpre-
tation is called for.
1.2.5 Contact
The final schema we would like to discuss is the CONTACT schema, which is also
at work in a wide range of metaphorical expressions. In a number of metaphors
it is, however, of secondary importance; it may be implicit, as in muzzle or kick,
or explicit, as in kick against the pricks, which primarily focuses on the useless
exertion of force.
In those metaphors where the schema is basic, it is used to describe two
different aspects of the LA. In in/out of touch, catch someone's ear, have some-
one's ear, the schema is used to describe the relationship between interactants,
while in touch ('to deal with or treat in speech or writing' (LDOCE), touch on
and fumble it describes the relationship between speaker and utterance.
BODY PARTS IN LINGUISTIC ACTION 51
The type of contact, its intensity, duration and the effort involved also play
a role in the metaphors. Touch and in/out of touch (with) are fairly neutral in this
respect; they simply use the presence/absence of contact to describe the pres-
ence/absence of a relationship between speakers or between the speaker and his
subject. With touch on and fumble the contact with the object is neither exten-
sive, nor firm - the metaphors rely on the contrast between mere contact and
control. Absence of firm, extensive contact leads to absence of control. In have
someone's ear and catch someone's ear the contact is with a body part obvi-
ously involved in listening. Have someone's ear involves both contact and con-
trol, or control in terms of contact, and stresses the permanent nature of the
communicative channel. Catch someone's ear focuses on the effort involved in
establishing such a channel, which results in achieving contact and control.
spar can indeed be explained on the basis of a donor domain situation, whereas
for bare one's heart/soul this causes problems, since a situation of this kind is
just not available. It is there that the image-schematic approach is most reward-
ing, or, to put it differently, has a greater explanatory value.
2. Value judgements
When describing linguistic action, a speaker may at the same time express a
judgement or give an evaluation of that LA or of any component of that LA.
When we look at the LA scene, it is clear that the value judgement should be
located at the level of the secondary speech act, the describing act, and that it
refers to aspects of the primary speech act. The primary speaker's intentions and
behaviour, the linguistic form, manner of presentation, hearer's attitude etc. and
the relations between these components all come in for evaluation (see Ver-
schueren 1984 and Rudzka-Ostyn 1988).
In the metaphors examined, two types of criteria are used for evaluating LA:
more concrete ones and more abstract ones. The more concrete criteria are
based on donor concepts; within the donor domains, certain types of behaviour
are approved or disapproved of on various grounds, and these evaluations are
then transferred to the recipient domain.
Apart from such criteria provided by concepts in the donor domains, there
are also more abstract criteria based on a number of VALUE SCALES.
cipient domains. For instance, there are certain social rules which dictate what is
acceptable or unacceptable in the action of eating. Spitting out food (as spitting
in general) is socially disapproved of in our society, so that a negative value
judgement accompanies the LA verb spit out. In the domain of fighting, unfair
behaviour and tactics are objectionable, as in throw dust in someone's eyes.
Lack of skilfulness is valued negatively in the domain of manipulating objects,
so that fumble as a LA verb expresses disapproval, too.
The adjective silver-tongued expresses a positive value judgement of the
manner of speaking. The donor domain of manipulating objects here structures
the linguistic action metonymically as the manipulation of an instrument, more
specifically of the articulator 'tongue'. The positive evaluation is carried over
from the donor domain, in the sense that silver, as a precious metal, has positive
connotations in our society. (Cutlery is such an example where the silver in-
strument gets a positive evaluation as a marker of 'class'.)
The verb backbite, on the other hand, carries a negative evaluation of the
linguistic action. Coming from the domain of violent actions performed by ani-
mals, it expresses a strong disapproval by S2. The negative features which are
carried over from the donor domain are 'aggression', 'pain inflicted on the
victim', and the cowardly aspect of 'attacking someone from the back', i.e.
when he is defenceless. The link between donor domain and recipient domain is
that the mental harm and pain are viewed as similar to physical harm and injury.
2.7.2 Scales
The more concrete criteria illustrated above, however, do not always suffice as
an explanation for the evaluative meaning of metaphors. In addition, more ab-
stract criteria are relied on, based on VALUE SCALES. Johnson includes the image
schema of SCALE in his list (cf. 1.2.0. above), and points out that this structure
is "experientially basic" and "one of the most pervasive image-schematic struc-
tures in our understanding". (1987:123). Further, what is typical of scales is
their normative character: "Having more or less of something may be either
good or bad, desirable or undesirable" (1987:123). In what follows we use the
SCALE schema to account for value judgements expressed by the metaphors in
our corpus. A SCALE schema is based on our experience of 'more' or 'less', and
this 'more' of 'less' takes various more specific forms, depending on what is
measured. The SCALES which we found to be relevant are those of what we will
call 'INTENSITY', 'QUANTITY', 'FREQUENCY', 'SPEED', and 'DURATION'. These
54 PAUL PAUWELS & ANNE-MARIE SIMON-VANDENBERGEN
scales interact whith the image schemata of FORCE, CONTROL, CONTAINER, PATH
and CONTACT We shall now take a look at how they operate in metaphorisation.
Intensity
As pointed out in 1.2.2. above, the image schema of FORCE is particularly salient
in the metaphors with violent actions or fighting as donor domains. Force can be
measured, so that a LA can be evaluated in terms of having stronger or weaker
INTENSITY. One may compare kick someone when he's down and rap someone
over the knuckles in this respect. The former metaphor involves more force and
refers to more violent action than the latter. The value judgement is also more
negative in the former, because of the use of the feet to kick someone who is
already defenceless. The feature 'aggression' combines with the feature 'unfair',
and with the high degree of force being exerted. The INTENSITY SCALE is also
salient in a metaphor such as pull one's punches, where the speaker is restrain-
ing his force. Here a weak degree of force expresses less violent linguistic action
than might be possible in the circumstances.
The INTENSITY SCALE is however not only relevant in the violent action
metaphors. It is also present in metaphors such as shut one's mouth. Here the
mouth as a container is shut, which implies a sudden forceful closing action. The
FORCE structure operating here is that of installing a BLOCKAGE: the closing of
the mouth raises a barrier to LA. The degree of force exerted is clearly stronger
than in the non-metaphorical expression speak no longer. The same applies to
keep one's mouth shut, which involves more deliberate force than e.g. not
speak, or remain silent. However, whether force in such examples is evaluated
positively or negatively is not so clear, and will depend on the context in which
they are used.
Similarly there may be different degrees of intensity in the making of a
speech act. As Johnson says (1987: 59): "The difference lies in the force with
which the sentence-container is thrust upon the hearer". For instance, a warning
may be mild or strong. A strong warning is made more forcefully. This, as John-
son says, is mainly a question of emphasis. In a number of metaphors in the
corpus, such degrees of force of the speech act are particularly salient. Good
examples are applaud and raise one's eyebrows, which both have nonverbal
communication as a donor domain. The meaning of applaud can be paraphrased
as 'express strong agreement with' (LDOCE). Whether the nonverbal act ac-
companies the linguistic action or not, it is clear that the latter is judged by S2
to be a forceful form of approval, agreement, support etc. In contrast, the meta-
phor raise one's eyebrows means 'express surprise, doubt, displeasure by or as if
BODY PARTS IN LINGUISTIC ACTION 55
moving the eyebrows upwards' (LDOCE). The speech act must here be a 'mild'
one, as can be judged from the fact that it would be inappropriate in a context
where S1 starts shouting, crying etc. to express feelings of displeasure.
In conclusion, strong and weak FORCE, measured on the INTENSITY
SCALE, are not positive or negative on their own. Degrees of force may however
combine with other concepts and thus acquire a positive or a negative value.
When force combines with violent behaviour, for instance, the negative value
judgement will be strengthened.
The bodily experience of CONTROL was suggested in 1.2.4. above as a
way of correlating the schemata of FORCE and BALANCE. In general, there are
three types of control:
• control over oneself, which is typically evaluated positively,
• control over others, which is typically evaluated negatively, and
• control over the environment, which is also typically valued positively.
Hence, what may be focused on in metaphorisation is the simple presence or
absence of control. A metaphor in which control is salient is flounder. The
donor domain of this metaphor is walking, and its literal meaning is 'move about
with great difficulty'. Non-fluent speech is referred to in terms of lack of control
of one's movements. The value judgement is hence clearly negative. In muzzle,
the presence of control is valued negatively, because it is exerted over others.
The donor domain of controlling animals adds a strong negative value judge-
ment to the expression, since a strong form of control of humans over other
humans in the way one normally deals with animals is disapproved of. Finally,
instances such as shoot one's mouth off illustrate salience of the third type of
control. The donor domain is the manipulation of objects, here a gun. One
should be in control of objects, and not shoot off a gun when it is not ap-
propriate to do so. In the metaphor, which refers to 'talking foolishly about
something one should not talk about' (LDOCE), lack of control is valued nega-
tively.
Where CONTROL involves FORCE it may also be graded in terms of FORCE
and hence entail an evaluation on the INTENSITY SCALE. For instance, rub
something in involves less force , and hence a weaker form of control than force
something down someone's throat
Quantity
The relevance of QUANTITY may be exemplified by full-blooded (argument,
style, etc.). In this metaphor, the donor domain is 'blood', as the source of life
56 PAUL PAUWELS & ANNE-MARIE SIMON-VANDENBERGEN
and energy, and the positive value judgement of a forceful manner of speak-
ing/writing is related to the QUANTITATIVE SCALE
The SCALE OF QUANTITY is salient in a number of metaphors (all based on
the CONTAINER schema), though there is once again not a one-to-one rela-
tionship between much/little on the one hand and positive/negative on the other.
Clearly, when 'too much' is the meaning, the value judgement is negative, since
too means lack of balance. An instance of this is loudmouth. This metonymy
expresses a negative evaluation of someone who talks too much. High degree of
loudness (phonetic aspect) here expresses a form of 'obtrusive', hence disturb-
ing linguistic action.
In other cases, such as pour from/out, the value judgement is not clear. In
such utterances as (9), (10) and (11) below, the focus is on amount and speed,
combined with the physical schema of a CONTAINER. Whether S2 evaluates this
as positive or negative depends on other factors:
(9) Curses poured from his lips.
(10) Words poured out of her.
(11) She poured out her thoughts/feelings.
Frequency
As with the other scales, when extremes are focused on, this is valued nega-
tively. Examples are rattle-brained, rub something in and brainwash. In all
three, repeated talking is disapproved of. In rattle-brained, the brain is com-
pared to an instrument making a lot of quick little noises which however do not
mean anything. The adjective is used metonymically of a speaker who talks all
the time. The verbal rub something in refers, in its donor domain, to a repetitive
movement of the hands applying an ointment to the skin, until it is absorbed. It
is based on the CONTAINER, FORCE and CONTACT schemata, and the FREQUENCY
of the movement is required for effect. In the same way, repeated LA leads to
'absorption' of the message. It is valued negatively because the experi-
encer/listener loses control. Brainwash is very similar: again repeated LA is
valued negatively because it forces the listener to absorb the speaker's ideas.
Speed
LA may be perceived by S2 as being fast or slow. In such cases as pour out,
quantity and speed go together, but are not particularly approved or disap-
proved of. SPEED may however combine with a value judgement. Snap provides
a clear example. In the sense of 'say quickly, in an angry way' (LDOCE) it ex-
presses a negative evaluation of the LA, both through the feature 'violence'
BODY PARTS IN LINGUISTIC ACTION 57
(from the animal donor domain) and through the feature 'quickness': the unex-
pectedness and quickness make the victim defenceless. SPEED adds to the ag-
gression in this case.
In other cases, speed may imply fluency and hence be valued positively,
while slow speech is often associated with non-fluency and awkwardness. The
positive value judgement is apparent in have a ready tongue (ready here mean-
ing 'quick'), while the negative evaluation is clear in instances such as falter,
flounder, fumble.
Duration
DURATION interacts with FORCE, to the extent that the exertion of force may be
of long or short duration. An instance is chew the fat, which has a high value on
the duration scale. It may also relate to CONTACT, as in touch on, where the
contact is brief.
Further, the schema of CONTROL may interact with the schema of
CONTACT. In such metaphors, control may be graded in terms of DURATION. For
instance, have someone's ear involves permanent control and contact (see
1.2.5.). In contrast, go in at one ear and out the other implies a very brief con-
tact and hence almost total absence of control.
In c o n c l u s i o n , the SCALES of INTENSITY, QUANTITY, FREQUENCY, SPEED,
and DURATION may interact both with one another, and with a positive-negative
value judgement. A value judgement in terms of positive/negative is, however,
not a necessary corollary, and there is certainly no one-to-one relationship be-
tween say a high or low value on one of the scales and a positive or negative
value judgement.
Metaphors expressing value judgements can be grouped into two main classes,
viz. those cases where the value judgement is context-independent, i.e. inherent
in the metaphor, and those where it is context-dependent.
hands, hence inefficiency, and this evaluation is carried over into the LA domain
to refer to clumsiness in speech.
Another instance of context-free value judgement is found in drool. The
meaning of the verb in the donor domain, viz. 'letting liquid flow from the
mouth' (COBUILD) gives the metaphor its negative value judgement of 'talking
nonsense, talking foolishly' (LDOCE). The action in the donor domain is indeed
associated with imbecility, i.e. lack of CONTROL over one's own body, and
therefore meets with social disapproval. In the LA domain this negative evalua-
tion is always present.
neutral, simply denoting a person who publicly states the opinions, policies etc.
of another person or organization. In sentences like (16), there is no positive or
negative evaluation:
(16) He became the official mouthpiece of the moderate leadership.
(COBUILD)
However, the metaphor is frequently derogatory. The negative value judgement
comes from the fact that the speaker is compared to an instrument used by
someone else. Thus, whereas the metaphor may be used in a neutral way, the
negative evaluation is clear in other contexts.
A third type of metaphor combines a positive with a negative value
judgement. The type may be illustrated by means of blow off steam. As ex-
plained in 1.2.4. above, the image schemata of FORCE and BALANCE (or
CONTROL) are particularly salient in this metaphor. The restoration of balance is
valued positively: by expressing one's feelings one restores mental balance. On
the other hand the comparison of a human being with a steam engine may ex-
press a certain amount of scorn, contempt, etc. Compare example (17):
(17) They can let off steam in pubs where nobody knows them.
(COBUILD)
This type is different from the type illustrated by jaw: whereas jaw is a case of
ambiguity, blow off steam is a case of merger. The two value judgements are
present at the same time, and the context may strengthen the positive or the
negative side of the coin.
Another example of the merger type is find one's tongue. Finding an ob-
ject which one needs for a certain purpose is a positive experience. Hence, the
ability to speak again (after the removal of psychological barriers) is valued
positively. However, finding implies previous loss, which is valued negatively.
Example (18) implies annoyance of the speaker, which may also be present in
(19):
(18) Lost your tongue ?
(19) Oh, you've found your tongue, have you. (COBUILD)
In (19), contextual factors reinforce the negative value judgement; the expres-
sion of scorn is salient.
It needs to be emphasized that the role of context - both situational and
linguistic - cannot be overestimated. First, even though a particular item proto-
typically expresses a positive or negative value judgement, pragmatic factors
may always reverse it. In utterances such as (20), the metaphor which negatively
60 PAUL PAUWELS & ANNE-MARIE SIMON-VANDENBERGEN
refers to excessive control over someone, has a positive meaning for the
speaker, because of his ideology:
(20) Fortunately, we managed to brainwash him.
However, such factors do not detract from the inherently negative value judge-
ment present in the metaphor.
Secondly, the semantic-syntactic contexts in which the metaphors fre-
quently or typically occur may contribute to the value judgement. For instance,
pull punches and put a foot wrong nearly always occur in the negative, as in
(21) and (22):
(21) I didn't pull any punches.
(22) He never puts a foot wrong.
Although the former is inherently positive, and the latter inherently negative,
their most usual contexts, i.e. negatives, reverse the value judgement of the LA.
Another example is the metaphor spit out. The imperative (23) gives a positive
value judgement to the LA referred to (though the expected message still quali-
fies as 'difficult', 'possibly unwelcome'), as contrasted with an utterance like
(24), in which the value judgement is negative. Mood thus also contributes to
value judgement:
(23) Spit it out. (LDOCE)
(24) She spat out the hateful word. (Webster's)
There are two types of interaction between value judgements in donor domains
and the recipient domain of LA: the value judgement may be directly trans-
ferred, or it may be new.
2.4 Links between the most important donor domains, image schemata, and
value judgements
boundaries (hence ambiguity, etc.) and the negative experience of tasting bad
food.
Thirdly, eating involves the destruction of food. In the same way, LA may
lead to the destruction of someone. In such metaphors, strong force exerted on
a person is valued negatively. An example is chew someone out.
Finally, eating may be focused on as an activity with a certain duration.
This can be valued positively (focus on the social aspect) or negatively (focus
on DURATION without much result). Both evaluations may be expressed in chew
the rag/fat, meaning 'have a conversation together' (positive) or 'complain
together' (negative).
2.4.2 Breathing
is 'inner'. (cf. Johnson 1987: 124ff.). It may be noted in passing that the
CENTRE-PERIPHERY schema is also particularly salient in metaphors referring to
the expression of feelings, in which the heart is seen as the centre of our inner
feelings: bare one's heart, mbosom, hearty, heartwarming.
2.4.6 Walking
LA as compared with movement along a path, is valued negatively in the fol-
lowing two cases.
BODY PARTS IN LINGUISTIC ACTION 65
In this final section we shall summarize the main findings of our research and
point out which aspects need further exploration.
Although the main principle underlying the collection of our corpus was
the explicit or implicit reliance of the metaphors on the functioning of the body,
it is clear that there are also important differences cutting across the expressions
composing the material.
• One distinction, mentioned in 1.1. above, is that between metaphors based on
(i) those body parts used in both speaking and some other activity, and
(ii) those body parts not present in LA, such as feet, hands etc.
Within each of these classes, further distinctions can be made. In class (i), a
distinction can be drawn between metaphors where another domain is clearly
relevant (e.g. eating, breathing) and those where no such other domain seems
to be involved. An instance of the latter type is open one's lips. However,
this type is, in our view, not merely metonymicai, because at a higher level
transfer does indeed take place. In the case of open one's lips, for instance,
the body is seen as a container for language. The focus on language as com-
ing out of a container (the speaker's body) and going into one (the listener's
body) links a large number of at first sight unrelated expressions. At this
higher level, such expressions as open one's lips, shut up, pour out, rub it in,
have a word in someone's ear, etc. can all be shown to be related conceptu-
ally. In class (ii) above, a further distinction can be made between those ex-
pressions where body parts are used in nonverbal communication and those
where body parts are used in some type of non-linguistic (inter)action (e.g.
fighting, walking, etc.).
66 PAUL PAUWELS & ANNE-MARIE SIMON-VANDENBERGEN
• Finally, expressions such as put words into someone's mouth, take the words
out of someone's mouth have no donor domains except the most general
ones of image schemata: 'the body as a CONTAINER' and 'CONTROL' (see
Pauwels, this volume, for a detailed discussion of such metaphors).
The conclusion we draw from these types of expressions is that the image
schematic approach to metaphorisation is essential, both for establishing links
between metaphorical expressions originating within a particular domain and for
generalizing over domains. However, some metaphors seem to work only on
this most general level, while others, though ultimately also based on image
schemata, depend more directly on very specific domains of experience. It fol-
lows that the relevance of these different levels of abstraction may vary from
expression to expression, depending on its position on a scale. The exact rela-
tionship between domains and image schemata needs further investigation.
Value judgements are among the most important factors that facilitate
metaphorical transfer. As pointed out, they may either simply be explained from
more specific donor domains or they may, in addition or exclusively, be based
on a number of SCALES. In other words, donor-based judgements also involve
scales but there are judgements which seem to be based on scales only. Since
'too' is negative, talking too much, too slowly, too long, too frequently etc. is
valued negatively: e.g. rub something in derives its negative value judgement
from 'too frequently'. It was further emphasized that scales may be relevant
without involving a value judgement: in touch on {a subject) the brevity of the
contact is relevant in the metaphorisation, but the expression is not inherently
positive or negative.
Only when we show how all the factors mentioned (i.e. the relevant image
schemata, domains, scales, value judgements) interact, can we capture the intri-
cacies of metaphorisation.
It is suggested in this paper that the schematic framework proposed by
Johnson (1987) may have to be modified and/or refined. One such refinement
involves the introduction of the CONTROL schema. Presence or absence of con-
trol and the type of control indeed appear to be important variables determining
the value judgement. Its exact status in terms of links with the other schemata
still needs to be examined, but so does, in our view, the status of Johnson's
schemata. It seems possible, for instance, that some schemata are more basic
than others, or, in other words, that there may be a hierarchy of schemata. Fur-
ther, we have found that the image schema of SCALE is indeed pervasive in that
68 PAUL PAUWELS & ANNE-MARIE SIMON-VANDENBERGEN
Acknowledgements
Notes
6. Guard one's tongue, tonguetied and bridle one's tongue have been classified here, rather
than under 1.1, since the tongue only provides the link with the LA domain, while the
main metaphorical element is clearly the absence of movement/speech.
7. The term "experiencer" is used here as a cover term for listener, addressee and patient.
8. Another possible interpretation for this metaphor was suggested by L.Goossens. Birds of
prey also pick the brains of their prey. The container schema is still salient in this inter-
pretation.
Assessing Linguistic Behaviour
A Study of Value Judgements
Anne-Marie Simon-Vandenbergen
University of Ghent
0. Introduction
1. Approach
The figures indicate that between 100% and 70% agreement was reached in
48% of the cases; and that at least 50% agreement was reached in 85.6% of the
cases. The two items on which all informants agreed were fluent and expand on
a subject. Both were judged to express a positive evaluation, the former expect-
edly so, the latter slightly more surprisingly so. We shall come back to fluent
below. At this stage, the unanimous agreement on expand on a subject needs to
be examined because it points to a second 'limitation' of the questionnaire.
It is possible that in some cases the order in which items were presented
may have influenced the answers. The fact that all informants judged expand on
a subject as positive, may have been caused by the fact that it occurred in a list
of definitely (more) negative items such as keep on about a subject, be on about
a subject, carry on about a subject, etc. (See Table B3). This may explain why
it is considered a positive rather than a neutral item, which it appears to be,
judging from dictionary definitions. (See e.g. COBUILD 'give more information
or details about it when you write or talk about it'). This explanation finds some
support in the results for go into a subject, which, whereas I also expected it to
be judged as neutral, only received 3 ticks in the 'neutral' box, while 85% of
informants saw this expression also as a positive one.
76 ANNE-MARIE SIMON-VANDENBERGEN
On the other hand, the explanation does not account for the fact that go
into a subject does not reach a 100% agreement. Perhaps such small
'deviations' must be allowed for in this kind of questionnaire and interpreted as
a matter of coincidence, perhaps there is an explanation in terms of the kinds of
contexts with which informants associate items. In the case of expand on a
subject, it seems to me possible that its unanimously positive evaluation may
have something to do with the academic context in which students meet the
word, especially in comments on essays, in interviews, etc. in the exhortative
form: 'Expand on this!' or 'Could you expand on that?'. In other words, ex-
panding on a subject is positive because it suggests that expectations are met.
Further, the person who can 'expand' is knowledgeable and able to present
information in a structured way.
That the order in which items are presented may influence answers, how-
ever, is further apparent in the reactions to an item which was presented twice,
i.e. once in set A and once in set B. The expression stick to a subject was pre-
sented as no. 10 in Table A5 and as no. 10 in Table B5. As the figures show, an-
swers were very much divided over positive and neutral in A5 (9 positive - 8
neutral), while 90% of informants judged it to be a positive item in B5 (i.e. 18
positive v. only 2 neutral). The explanation may be that the preceding item in B5
was meander, which is almost like an antonym of stick to a subject and which
was judged rather negatively. The preceding item in A5, viz. tackle a subject, is
semantically less close to stick to a subject and would therefore not cause this
kind of reaction.
On the other hand, it seems to me important that neither in A5 nor in B5
was the item judged to be negative (only 2 ticks in A5, and none in B5). I there-
fore believe that the general trends are significant - on condition the results are
interpreted with caution.
Another item which occurred twice was chatty, this time in the same set
(A4), once as no. 2, once as the very last item, no. 20. This means that in this
case the same informants gave their opinion twice on the same item. The figures
show that the answers are largely, though not completely, the same: 85% posi-
tive v. 75% positive; 3 neutral answers in both cases; 1 negative and 1 blank in
no. 20. In other words, my informants on the whole primarily thought of chatty
as a positive word, but it is obvious that one may also think of contexts in which
it acquires a negative meaning. (See the example in LDOCE, quoted in section 1
above, and see further below).
ASSESSING LINGUISTIC BEHAVIOUR 77
3. The data
In this section I shall look at expressions referring to the social function of talk,
to various aspects of turn-taking, and to various ways of dealing with a topic.
type of discourse (i.e. talking for the sake of contact) as positive, negative, or
neutral? Secondly, what types of metaphors refer to conversation?
A list of twenty nouns all referring to verbal events was presented to the
twenty informants, who were asked to evaluate the verbal events as positive,
negative or neutral. The answers show a number of interesting tendencies (cf.
Table Al):
Table 2
positive neutral
conversation: 50% 50%
talk: 50% 50%
dialogue: 35% 65%
(1) The nouns which one would intuitively classify as 'neutral', viz. conversa-
tion, talk , and dialogue also get a fairly high score for 'positive' (cf. Table 2). It
appears that the more formal word dialogue, which also refers to a more formal
occasion, gets a higher percentage of 'neutral' evaluations than the more every-
day words. It is further noteworthy that the word colloquy (60% neutral) had 2
ticks under 'negative' as well. This may be due to the formality of the event
referred to. On the other hand it is possible that a number of the students had no
sharp intuitions about the word. In any case, there is nothing in the definition
given in dictionaries to predict a negative value judgement (cf. LDOCE: 'a
formal conversation'; COBUILD: 'conversation or meeting; a formal word'),
unless formality has negative connotations to some subjects. It is important,
though, that the other three words had empty slots for 'negative'. (Dialogue:
35% positive; talk/conversation: 50% positive). This shows that, if a value
judgement is attached to these words - out of context - it is a positive one.
Hence, the mere act of 'talking, having a conversation' is felt to be a positive
one, possibly because it is a signal of non-inimical relations. (Compare: not to be
on speaking terms).
(2) Words which scored higher for 'positive' are listed in Table 3. What they
share, as opposed to the four discussed above, is reference to 'informality', and
'lightness of topic' (talking about unimportant things), which are felt to be
pleasant.
82 ANNE-MARIE SIMON-VANDENBERGEN
Table 3 Table 4
In order to test this hypothesis, the same twenty words were presented to an-
other group of twenty informants, who were asked to evaluate the verbal events
referred to as: a) light, heavy, neither; and b) as: deep, shallow, neither.
(Appendix, Table Bl). The results were the following for the above words:
(i) Colloquy and dialogue are the only two words which have more than 50%
answers for 'heavy' and 'deep' (Colloquy: 70% heavy; 65% deep; dialogue:
90% heavy; 90% deep). This seems to indicate that it is the 'seriousness' which
gives these words a less positive connotation than talk and conversation (cf.
talk: 95% neither heavy nor light; 75% neither deep nor shallow; empty boxes
for light and shallow; conversation: 55% neither heavy nor light; 55% neither
deep nor shallow; empty boxes for light and shallow.) These figures show that
'light' and 'shallow' are not associated with these four words, and that talk is
the most neutral one in these respects.
(ii) In contrast, the more positive words chat, chit-chat, chinwag and natter are
associated with 'light' and 'shallow' (cf. Table 4). These figures show two
things: first, that lightness of subject-matter may be valued positively; secondly,
that chat, which scores lower for 'shallow' than the other three positive words,
is the most positive word of all (80%). If one compares chat and chinwag, there
is nothing in the dictionary definitions which would explain a less positive value
judgement in chinwag (55% positive). It is defined in LDOCE as 'informal
conversation; chat'; and in COBUILD as 'long enjoyable conversation between
friends'. Hence, the slightly more negative connotation must come from the
feature 'shallowness', which is less pervasive in chat (cf. further on the meta-
phor chinwag). This hypothesis was confirmed by QII and QUI. Students who
evaluated chinwag, chit-chat, and natter as negative words all referred to the
trivial, inconsequential nature of the activity. It further appears from QII and
QUI that the absence of a high degree of agreement shown up by QI must be
explained from both the context and collocations that come to informants'
ASSESSING LINGUISTIC BEHAVIOUR 83
minds first, as well as from their sub-culture and personalities. Examples given
by those evaluating the above words as positive were such as:
(9) Let's have a good chinwag.
(10) We must meet and have a chinwag.
(11) They sat down for a chit-chat.
(12) We had a good old natter this morning.
Typical collocations of chinwag seem to be, judging from the examples: a good
-, a nice -, a jolly old -.
Examples given to illustrate the negative connotation of these same words
were such as:
(13) I want positive ideas, not chit-chat.
(14) Don't waste your time on chit-chat.
(15) This parish magazine is full of nothing but chit-chat.
Strikingly, an illustration of the negative connotation of chinwag was the fol-
lowing:
(16) Ethel came over for coffee and we had a bit of a chinwag.
Although the example does not make clear the negative connotation, the stu-
dent's comment does: 'It's a bit old-ladyish'. Such evaluations of words and of
the contexts they are associated with point to the relevance of the sub-culture.
In this case, age seems to be an important factor. Other instances will confirm
the role played by the sub-culture.
(3) Words which combined the features 'negative' - 'light' - 'shallow' were, in
descending order of 'negativeness': prattle (80%), babble (75%), yap (75%),
idle talk (70%), trivial talk (70%), gossip (65%), tattle (65%), small talk
(60%), jaw (50%). It is clear, however, that the relevance of the features 'light'
and 'shallow' varies: in prattle, for instance, they are very salient and directly
relevant to the negative value judgement (See Appendix, Table B1: 90% light;
85% shallow); in contrast, jaw scores only 55% on both 'light' and 'shallow',
which means that the negative judgement may be due to other factors. (On the
metaphor jaw, see further below).
In any case, it appears that 'light' and 'shallow' can go together with
positive or negative evaluations. 'Shallowness' in a more extreme form may
indeed become 'foolishness', 'emptiness'.
Let us now have a closer look at the metaphors in this set, viz. chinwag,
jaw, small talk, cackle, yap (Tables Al and Bl in the Appendix).
84 ANNE-MARIE SIMON-VANDENBERGEN
Chinwag and jaw are comparable in that they share the same donor do-
main, viz. body parts. They both refer to parts of the face which are seen mov-
ing when one talks. Hence, in both cases LA is referred to in terms of something
which is visible, an outward sign of talk, articulation rather than meaning or
content. This focus on the superficial level of language conveys the meaning
aspect 'talk for talk's sake', i.e. the primary function of casual conversation. In
both cases, also, no value judgement is attached to the words in the donor do-
main. What about the recipient domain?
When compared with chat, a non-metaphorical expression which has
more or less the same denotative meaning, chinwag is seen to contribute a
negative element to the concept of 'conversation', while the shift towards the
negative pole is even more obvious with jaw. Compare: chat: 80% positive (no
negatives); chinwag: 55% positive (other answers divided over negative and
neutral); jaw: 50% negative (other answers divided over positive and neutral).
Hence, the metaphors seem to have the function of expressing a negative value
judgement. Chinwag, as pointed out above, emphasizes the lightness or shal-
lowness of the subject-matter, expressed metaphorically by reference to the
surface level of articulation. In contrast, jaw seems to go in the other direction:
it scores considerably lower on 'light' (compare: chat: 95%; chinwag: 90%;
jaw: 55%). The definitions do point to a difference in value judgement through
the absence of the element 'enjoyable', 'friendly' in jaw. Compare:
chat COBUILD informal, friendly conversation, usu. about things which
are not serious or important LDOCE friendly informal conversation
chinwag COBUILD long enjoyable conversation between friends LDOCE
informal conversation; chat
jaw (vb) COBUILD talk to each other, often for a long time, without saying
anything important LDOCE (informal, sometimes derog.) talk/to talk
Furthermore, jaw can also mean 'sermonizing talk, lecture' (CODCE), which
may have caused the negative responses.
The less positive meaning of jaw as compared with chinwag can in fact be
explained from the donor domain: on the one hand the movement of the jaws is
a slower, heavier one than that of the chin, so that speed and frequency of
movement seem to be associated with 'lightness'; on the other hand, the word
wag may have associations with 'tail wagging', an outward sign of wellbeing in
animals.
The metaphors cackle and yap are based on the animal domain. Both ex-
press a negative value judgement, according to dictionary definitions as well as
according to questionnaire responses (Compare: yap: 75% negative; cackle:
ASSESSING LINGUISTIC BEHAVIOUR 85
85% negative). These expressions are discussed further in section 3.2.2, because
they focus on manner of speaking (loudness) rather than on the function of
'maintaining contact'. Their relevance here lies in the feature 'empty, unim-
portant subject-matter'. Compare:
yap LDOCE noisy empty talk; vb: talk noisily about unimportant things
cackle LDOCE foolish useless talk
It appears, then, that 'unimportant subject-matter' may have positive as well as
negative connotations, which lie on a scale (cf. Figure 1). Different metaphors
focus on different 'degrees' on that scale. 'Talk for talk's sake' becomes nega-
tive when seen as 'useless noise'. The animal domain seems to emphasize this
particular meaning aspect: the barking of dogs in yap; the noise made by hens in
cackle.
Figure 1
The metaphor small talk was valued by 60% of informants as having a negative
connotation. Yet, there is nothing in dictionary definitions to point to a negative
meaning element. Dictionaries refer to the lightness of subject-matter. Compare:
small talk COBUILD light conversation that people make at social occa-
sions about unimportant things LDOCE light conversation on unimportant
or nonserious subjects CODCE unimportant social conversations
QII yielded the following negative comments on small talk: 'very trivial',
'banal', 'a waste of time', 'semi-ritualistic practice', 'inconsequential'.
One explanation for its more negative connotation than chat or chinwag
may be the social context associated with it, i.e. talk for talk's sake between
strangers, not friends. Talking about unimportant subjects is on such occasions a
signal of the absence of real contact or of the superficiality of the relationships.
The donor domain is space: the word small denotes 'unimportant'. Although it
is neutral in the donor domain, it frequently expresses a negative value judge-
ment when it acquires a metaphorical sense in various collocations. Compare:
feel small ('feel ashamed or humble'); small fry ('young or unimportant per-
son'); small beer (s1. 'unimportant': He thinks he's wonderful, but he's really
rather small beer). (Definitions from LDOCE). Such examples support and are
coherent with Lakoff & Johnson's UP-DOWN spatialisation metaphors: '"Bigger
86 ANNE-MARIE SIMON-VANDENBERGEN
3.1.2 Turn-taking
As pointed out by Good (1979), quoted above, equal access to speaker and
hearer roles is a convention to be observed by participants who want to pre-
serve a balance of power. It appears from value judgements that speakers who
break the unwritten rules of conversational behaviour in this respect are evalu-
ated negatively.
self but moves objects (words) into it. These expressions are neutral with regard
to an evaluation of the LA, because get and put are themselves neutral items.
In conclusion, the degree of force which is used in entering conversation
directly correlates with the extent to which the LA is viewed negatively. De-
grees of force are expressed by means of types of physical movement and ways
of damaging or destroying objects.
Figure 2
When a subject has been brought up, one starts talking about it, which is typi-
cally expressed as physical movement. An example is come to (the point).
Again, the way of moving is responsible for differences in value judge-
ment: plunge into (a subject) adds the meaning of 'suddenly or hastily': 'She
plunged at once into a description of her latest illness'. (Example from
LDOCE). Here, the scales of speed and force are relevant. As Table A5 in the
Appendix shows, the expression was valued positively by 70% of informants,
while 2 informants ticked the negative slot and 4 the neutral one. QII clarifies
these evaluations. On the positive side the comments were: 'suggests enthusi-
asm', 'suggests dedication'; on the negative side the comments referred to lack
of thought. One informant, who ticked the neutral slot, explained his answer as
'it shows initiative, but may reflect stupidity'.
90 ANNE-MARIE SIMON-VANDENBERGEN
track of what you are saying'. Hence, lack of control is involved, which is dis-
approved of.
3.2.1 Speed
When a person is said to speak fast or slowly, this is an evaluation on the scale
of speed, which does not necessarily imply a positive or negative judgement.
The question arises, however, to what extent a positive or negative evaluation
correlates with either fast or slow speech, and how relevant the scale of speed is
in the LA lexicon.
burble LDOCE talk quickly but foolishly or in a way that is hard to hear
clearly
jabber CODCE 1 speak volubly and with little sense 2 utter (words) rap-
idly and indistinctly
prattle LDOCE talk meaninglessly or lightly and continually in a simple
way, or like a child
In this respect, then, the sound-imitative words are metaphorical in a broad
sense of the term, in that reference to the sound-level implies absence of the
deeper level of meaning. The explanation for the frequent association of fast
speech with lack of deeper meaning may be that thought requires time, so that if
one talks too quickly, one just produces sounds without meaning. Further, the
features 'fast, quickly' frequently co-occur with 'continually'. Examples are
rattle on, twitter, burble, prattle, but also talk nineteen to the dozen, pour
('flow steadily and rapidly'), etc. The same explanation applies to continuous
speech, viz. that if one speaks continually, LA becomes merely the 'mechanical'
production of sound. That such an association of 'fast' with 'continual' is a real
one is also apparent from the fact that, although to gas and to run on have no
feature 'fast' in the dictionary definitions, which only refer to continuous speech
and empty talk, 70% of informants attributed the feature 'fast' to gas, and 65%
did so for run on. (In the latter case, the donor domain is clearly salient). (See
further: 3.2.3)
The same link between high speed and lack of deeper meaning or value is
present in the metaphors knock off ('write quickly') and throw off ('say/write
something easily'). The focus here is on the quick separation of words from the
speaker (the particle off): the donor domain is that of manipulating objects in a
brusque manner, or more generally, it is spatial.
Another connotation of fast speech may be that it is aggressive. A suc-
cession of rapid questions is felt to be threatening because the addressee is
pushed into a position where he has to produce answers quickly, which is face-
threatening. It can be seen that a series of questions put rapidly one after the
other is typically expressed by means of metaphors from the donor domain of
war and fighting: a barrage of questions, besiege/bombard/pelt s.o. with ques-
tions. The metaphor a volley (as in 'He directed a volley of curses at the driver'
LDOCE) fits into the same pattern.
The explanation for these metaphors is that an object moving at high
speed will hit a target harder. Hence, if that object is directed at someone with
the intention to attack, speed becomes threatening. As there is in these meta-
94 ANNE-MARIE SIMON-VANDENBERGEN
phors a link between the scales of speed and frequency, they are discussed fur-
ther in section 3.2.4.
Summarising, we can say that fast speech frequently has the negative con-
notations of lack of clarity and lack of content. A typical donor domain is that
of sound, in that a high number of items referring to fast speech are sound-
imitative. Some of those have links with the animal domain, such as twitter
(birds); others are also used for sounds made by objects, such as rattle, clack
(cf. CODCE: 'sharp sound as of clogs on stone or of boards struck together'),
rap (LDOCE: 'sound of a quick light blow: a blow on the door'); still other
sound-imitative words are also used for the sound made by water flowing, such
as burble (cf LDOCE: 1 'to make a sound like a stream (flowing over stones)').
The same link between liquid and fast speech is present in pour (out), which has
no positive or negative value judgement but merely refers to a smooth move-
ment. Other donor domains are the brusque manipulation of objects and fight-
ing. The latter gives rise to the connotation of aggression, sometimes a con-
comitant of rapid talk directed at someone.
Table 5
3.2.2 Loudness
When one examines the metaphors in the LDOCE corpus which in their defini-
tions have the terms 'loudly, noisily, etc', the following connotations emerge.
Loud speech is uninhibited, free, forceful. The assumption is that, if
someone speaks loudly, he wants to be heard by everyone, and hence is not
afraid of reactions. These features are present in the non-metaphorical items
vocal and vociferous, on which informants were divided with regard to value
judgement in terms of positive, negative or neutral (Table A4). The metaphor to
sound off on the other hand, was found by 50% of informants to be negative
(Table A6). The explanation is that the positive characteristic 'uninhibited'
becomes negative when speech is 'imposed' on an audience: loud speech cannot
be ignored. The definition in COBUILD makes this clear: 'express your opin-
ions strongly or loudly to everyone without being asked'. This also explains the
divided answers for vociferous (8 positive, 8 negative, 2 neutral). Comments in
QII indeed show that the positive evaluation applies to speakers who 'state their
opinion clearly', 'are well-spoken and coherent'. Comments on the negative
evaluation were: 'implies over-loud arguing', 'suggests someone who is inclined
to be overbearing'.
A similar metaphor is to air, which literally means 'expose to air', and re-
fers to 'make known to others (one's opinions, ideas, complaints, etc.), esp. in a
96 ANNE-MARIE SIMON-VANDENBERGEN
noisy manner' (LDOCE). The example given in LDOCE is: 'They were tired of
the doctor airing his knowledge'. However, to air one's ideas was given in QUI
and valued by 7 out of 10 students as positive, and by the 3 others as neutral.
The 'positive' comments referred to: 'sharing one's views with others',
'circulating ideas, so that others can think about them', 'suggests free discus-
sion, a forum for sharing thoughts', 'implies healthy discussion and exchange of
thought'. It seems, then, that the feature which was valued positively by my
informants is 'openness'. The example given by LDOCE to illustrate the nega-
tive connotations collocates to air with knowledge, which conveys a different
attitude as compared with ideas, since the former suggests authority and control
over others. This shows that the linguistic context in which such metaphors
occur may determine which meaning aspect is more salient. Further, I would
hypothesise that the sub-culture plays a role, as well: young people seem to
value openness.
One step further, loudness becomes offensive. The metonym loudmouth
indeed refers to someone who talks 'in an unpleasant, offensive or stupid way'
(COBUILD). The feature of aggression is present in animal-sound based meta-
phors for loud speech, such as bark ('say something in a sharp loud voice'
LDOCE), and bellow ('shout something in a deep voice' LDOCE).
The donor domain of nature gives rise to the metaphors bluster and thun-
der. In both cases, loud speech is compared to threatening weather conditions,
i.e. wind blowing roughly and thunder respectively. Finally, the aggressive con-
notation of loud speech is expressed by reference to the manipulation of objects
in hurl, whose literal meaning is 'throw with force, throw violently from some
position' and whose metaphorical meaning is 'shout out violently' (LDOCE).
The general pattern in these metaphors is that loud speech is forceful
speech, which becomes negative when force is directed against an addressee, in
which case it is threatening.
A different value judgement attaching to loudness is stupidity. This feature
is present in loudmouth (cf. above), as well as in a number of sound-imitative
words which combine 'fast and loud', such as jabber, rattle on, gabble, at least
according to questionnaire responses. All of these were valued negatively. Jab-
ber also refers to the noise made by monkeys, and a number of other metaphors
show that animal sounds are a typical donor domain for loud, stupid talk: cackle
(hens), yap (dogs), gaggle (geese). The latter, according to LDOCE, refers to:
'1. number of geese together; 2. group of noisy people, esp. women, who talk a
lot.'
ASSESSING LINGUISTIC BEHAVIOUR 97
The opposite of loud speech, viz. soft, quiet speech, appears to be less sa-
lient, since I did not find metaphors for it in the corpus. The word softspoken
seems to indicate, though, that it is valued positively. This may help to explain
why loud speech should be associated with stupidity. On the one hand, loud
speech reminds one of sounds made by animals (cf. the animal-based meta-
phors), on the other hand it is a signal of anger, aggression, and hence lack of
self-control. The positive appreciation of soft speech may reflect the high value
put on 'education', 'fine breeding', etc. (Cf. 'Keep your voice down'). It is
possible that such a negative appreciation of loud speech should be culturally
determined, and possibly absent in other cultures.
3.2.3 Quantity
gasbag COBUILD person who talks a lot, esp. about things that are not
important (informal and offensive word) CODCE bag for holding gas; bal-
loon's or airship's gascontainer; (derog.) empty talker
In these metaphors the link between donor and recipient domains is emptiness.
It is significant, though, that both words have the meaning element 'large
quantity of speech'.
The explanation for the frequent association of continual talk with empti-
ness or foolishness must be the same as that suggested above in connection with
quick speech: serious thinking requires time; if one talks continually, there can-
not be much thinking behind it. Emptiness of content is expressed by focusing
on the sound level in the sound-imitative words, or by reference to something as
rare or light as 'gas'.
Talking (too) much is further associated with lack of control, because
there is the risk of saying something one did not really want to say or something
one should not have said. The metaphor let something slip is based on the do-
main of movement, and has a negative meaning due to the loss of balance and
control, i.e. to the element 'involuntary'. In the questionnaire, 80% of in-
formants valued the term as a negative one (Table A7). The importance of self-
control is further corroborated if one looks at evaluations of the opposites, viz.
of terms denoting carefulness in speech, such as guard one's tongue (75% posi-
tive).
However, the issue is slightly more complicated. The expression watch
one's words, which seems to be near-synonymous with guard one's tongue, was
rated as positive by only 65% of informants (Table A7). This may be an indica-
tion of a gradient: self-control is positive, but over-prudence is negative. The re-
sults of QUI for watch one's words further reveal the importance of collocation
and situation in value judgement. The expression was indeed valued negatively
by 8 out of 10 informants. (One ticked 'neutral', one left it blank). The explana-
tions all refer to over-carefulness imposed on one. Examples given by subjects
to illustrate the expression were:
(17) You have to watch your workaround him; he's easily offended.
(18) You'd better watch your words. She's his girl-friend, you know.
(19) If you don't watch your words, you'll find yourself in a difficult
situation.
(20) Make sure you watch your words when she arrives.
(21 ) The teacher told the boy to watch his words.
ASSESSING LINGUISTIC BEHAVIOUR 99
3.2.5 Directness
I shall now address the question to what extent being direct or indirect is
evaluated positively or negatively by speakers of English.
ASSESSING LINGUISTIC B E H A V I O U R 103
Results from the questionnaire show that being direct tends to be evalu
ated positively, though not unreservedly so. Compare: give a straight-out an
swer (95%), tell s.o. in plain terms (90%), tell s.o. straight (90%), speak one's
mind (80%), tell s.o. flat out (70%), use plain words (70%), speak plainly
(65%), call a spade a spade (60%), not to mince one's words (60%), tell s.o.
point-blank (55%) (Tables A6, B6).
There is clearly a gradient here, ranging from near-unanimous agreement
on the positive connotations to a nearly equal distribution of positive and nega
tive evaluations. The explanation is that 'directness' implies honesty and clarity,
which are evaluated positively, but may lead to rudeness, and thus hurt people.
The comments given in QII and QUI support this. Again, as with other items
discussed above, however, it is not the case that each individual expression has
either the positive or the negative connotation. Rather, informants agree on the
connotations to a greater or lesser extent, so that here again individual appre
ciations of types of behaviour come into the picture. A comment on not to
mince one's words given in QIII may clarify this:
It implies someone who is not afraid to say what they think and although I
think people should consider other people's feelings, we spend too much
time skirting around a subject for the sake of unnecessary taboos and con
ventions in our society.
There is apparently a clear link between the negative evaluations and flouting
politeness principles which require indirect speech acts.
The following items received a negative evaluation: plaster over an issue
(95%), hedge one's answer (90%), wrap up one's meaning (85%), beat about
the bush (85%), give an evasive answer (85%), talk round the issue (70%)
(Tables A6, B6). Two factors in particular are involved here, viz. deceit and lack
of clarity. The reason why plaster over an issue gets near-unanimous agree-
104 A N N E - M A R I E SIMON-VANDENBERGEN
ment, as compared with say talk round the issue is that in the former deceit is
salient, in the latter mere absence of clarity.
Opinion was very much divided on speak baldly (Table 6). The expla
nation must be that judgements depended on which features of the expression
informants had in mind. (cf. the LDOCE definition: 'spoken plainly, even cru
elly'). Further, as suggested before, the informant's personality and his whole
set of values play a role.
When we look at the donor domains which provide metaphors for direct
ness and indirectness, it appears that there are two major ones.
The first important domain is that of movement in space. Directness in
speech is expressed in terms of movement along a straight path, without turn
ings. Indirectness is compared with the opposite. Contrast: a straight-out an
swer, flat out v. talk round an issue, an evasive answer.
The second domain is that of covering and concealing. Underlying these
expressions is the metaphor 'Seeing is understanding'. The more one covers up
the meaning, the greater the indirectness. Compare: plain words, speak plainly
v. wrap up one 's meaning, plaster over an issue, hedge one 's answer. In (not) to
mince one's words, the object in question is food, which is (not) made more
palatable by mincing it.
This second donor domain leads us to another group of metaphors which
express some form of 'indirectness', viz. by referring to various ways of
'dressing up' one's language (to use another metaphor). Examples are: adorn,
embellish, polish՛, florid, flowery, frilly, fustian, refined. These words also
focus on speech as an object which is treated, decorated and/or covered up in
various ways. Value judgements here seem to depend on whether the decora
tions are exaggerated or not. The expression polish one's speech was included
in Q I , and opinion was divided: 9 positive, 5 negative, 5 neutral (Table 6).
This was surprising in view of the rather positive definitions in LDOCE and
COBUILD. The latter even adds 'used showing approval'. The expression was
therefore included in QII and QUI, and the following 'negative' comments were
given: 'if one has to polish one's speech it suggests one feels inferior',
'snobbery', 'false, rather pretentious speech'. These comments clearly testify to
an ideology which favours 'the natural' rather than adaptation to certain norms.
The evaluations of and comments on a refined way of speaking go in the same
direction. In QUI, 4 out of 10 informants thought it was a negative term, and
gave the comments 'put on', 'false', 'snobbish' 'posh'.
ASSESSING LINGUISTIC B E H A V I O U R 105
Finally, there are the definitely negative terms to refer to particular styles,
such as inflated, swollen, turgid, padding. The motivating factor in these is the
contrast between a big shape and inner emptiness.
I shall now examine more closely the donor domains most frequently drawn
upon for the expression of the various aspects of conversational behaviour
discussed above. These domains vary in degrees of abstraction and may inter
sect. For instance, the domains of 'movement in space' and 'liquids' intersect,
and so do the domains of 'liquids' and 'sounds'. Hence, when a metaphorical
expression is given as an example of a particular donor domain, this does not
imply exclusive membership of that domain. Rather, the implication is that the
feature which is most relevant to the meaning of the metaphor comes from that
domain.
Chinwag and jaw refer in the donor domain to parts of the face which are seen
moving when one talks. When these body parts are used to refer to informal
conversation the focus is on the superficial level of language, viz. articulation of
sounds. This emphasis conveys the meaning of 'absence of a deeper content'.
The noun loudmouth used metonymically to refer to a person who talks
loudly and in an offensive or stupid way focuses on the most important body
part used in speech. Again, the articulatory or phonetic level is used to convey
something about the content of the message. The same negative value judge
ment is present in blabbermouth to refer to a person who talks too much and
hence cannot keep a secret.
Expressions with tongue such as bridle one's tongue, guard one 's tongue,
tongue-tied, have a loose tongue, lost your tongue? and with lips such as close-
lipped, tight-lipped, keep a stiff upper lip, my lips are sealed, focus on the most
movable articulators and convey a positive value judgement or a negative one
depending on
• the extent to which the speaker is able to control his articulators;
• the extent to which the speaker is seen as behaving in a 'sociable' way. (Self-
control is positively valued only insofar as it does not hinder social contact
and co-operation.);
106 A N N E - M A R I E SIMON-VANDENBERGEN
4.3 Liquid
Apart from rivers and streams, other elements in nature are used as donor do
mains in my corpus: bluster and thunder are both used for the element 'force',
combined with loud noise. Hence, they are used to link loudness with aggres
sion.
ASSESSING LINGUISTIC B E H A V I O U R 107
4.5 Animals
Speakers are seen as dealing with words, messages, topics in the way one ma
nipulates objects. Value judgements depend on such factors as:
• the speaker's skill or absence of it;
• the extent to which the speaker-manipulator wants to 'disguise' the real
object by treating it in various ways;
• the relevance and usefulness of his actions (e.g. belabour is negative);
• the extent to which the speaker's action is destructive (e.g. chip in, cut in are
negative).
Expressions based on the domain of war and fighting all express a negative
value judgement of the type of L A referred to. They are used to convey the
meaning of aggressive behaviour. In this corpus they were found to be typically
used for the expression of repeated questions (cf. 3.2.4. above), which are
'threatening' to the experiencer.
4.9 Sound
Apart from those expressions where sounds specifically come from water (4.3),
natural elements (4.4) or animals (4.5), there are a great number of sound-imi-
108 A N N E - M A R I E SIMON-VANDENBERGEN
tative expressions which refer to L A only. They typically end in the suffixes -le
and -er denoting the frequentative repetitive nature of the sound. They are
metaphorical in the broad sense of the term only. Though primarily referring to
fast, continuous and/or loud speech, they frequently express a negative value
judgement, in that focus on the sound level tends to imply absence of content,
hence foolishness, stupidity, etc.
It is clear from the present corpus that in the evaluation of L A the scales of
quantity, frequency, speed, duration and intensity, which are suggested as rele
vant in Pauwels & Simon-Vandenbergen, play an important role.
Firstly, there is a direct relationship between the scale of speed and the
evaluation of fast/slow talk. It appears that metaphorical expressions for fast
talk nearly always express a negative value judgement. This judgement is typi
cally expressed in words focusing either on the sound level (the sound-imitative
ones) or on the mechanical production of speech (by means of donor domains in
which the 'sources' are machines or guns). This means that when fast speech is
expressed metaphorically it is because it is judged to be 'too fast', i.e. an ex
treme position on the scale is focused on.
The other extreme, i.e. speaking too slowly, is also valued negatively, ei
ther because of the lack of variation (leading to boredom) or because of the lack
of smoothness. The latter aspect is typically expressed by drawing on the donor
domain of bodily movement and skill, especially through focusing on lack of
control.
It appears, however, that 'fast speech' is more frequently metaphorised
than 'slow speech', and that in those items which were judged to denote slow
speech, other features are more salient.
Secondly, quantity and duration are focused on in metaphors for talking
too much or too little. Again, as with speed, 'too much', 'too long' are more
frequently metaphorised than 'too little'.
It is striking that in a great number of cases the sound-imitative words de
note 'fast and continuous' speech, which means that the same value judgements
attach to both types of LA. Mere quantity is focused on in a number of expres
sions based on the donor domain of 'liquid', and value judgements depend on
the context. Again, quantity and speed may be combined (as in gush, torrent).
ASSESSING LINGUISTIC BEHAVIOUR 109
A final question asked at the outset was how realistic the dichotomy context-
dependent v. context-independent value judgements would be. The findings of
this study suggest the following answers.
• Firstly, apart from just a few exceptions, the words presented in the ques
tionnaires did not receive unanimous judgements. On the other hand it is
clear that some expressions invite a greater amount of agreement than others.
Therefore a gradient-view seems realistic. Words which are judged to be
positive by say 90% of informants must be different from those which are
judged as positive by say 45%, as negative by 25% and as neutral by 30%.
The difference between the types must lie in the fact that some words are
prototypically positive or negative, whereas others are not.
• Secondly, we must account for the lack of agreement. It appears from the
results of the questionnaires that informants in some cases had different col
locations and contexts in mind. The former are particularly interesting and
ASSESSING LINGUISTIC B E H A V I O U R 111
worth further examination. In any case it is obvious that the same expression
may, depending on contextual factors, be positive or negative. Compare:
(22) We had a good old natter this morning and it cleared the air.
(Positive).
(23) These women nattering all the time got on his nerves. (Negative).
Of course, pragmatic factors may always override value judgements (even
prototypical ones), as in:
(24) I hate nice and cosy chats. (Negative).
The word chat, whose positive meaning is reinforced by nice and cosy, here
occurs in an utterance which, through the superimposition of the speaker's
individual judgement, expresses negative feelings about what are typically
positive events. However, it is useful to distinguish between such clear cases
of 'reversal' of judgements and others where it is not so obvious whether the
positive or the negative meaning aspect is predominant.
• Thirdly, it appears that sub-cultural as well as individual values and attitudes
play a role in judgements. The same expression may evoke contexts which by
one group/individual are evaluated in a positive way and by another one in a
negative way. For instance, self-control and refinement are obviously not al
ways valued positively by everybody. The association of types of behaviour
with particular evaluations depends on one's broader value system. (See e.g.
the discussion of keep a stijf upper lip in section 3.2.3.2.).
7. General conclusions
Acknowledgements
her students and for devoting some of her lecturing time to their completion. I
am very grateful to the Newcastle students themselves for their invaluable help.
114 A N N E - M A R I E SIMON-VANDENBERGEN
Questionnaire I: set A
I Evaluate the type of verbal event denoted by the following nouns as basically
positive/pleasant, negative/unpleasant or neutral.
Put a tick in the appropriate column.
Table Al
Noun Positive Negative Neutral Blank
1 chat 16 = 80% 4
2 babble 3 15 = 75% 2
3 dialogue 7 13 = 65%
4 chinwag 11 =55% 6 2 1
5 conversation 10 = 5 0 % 10 = 50%
6 jaw 4 10 = 50% 5 1
7 natter 11=55% 4 4 1
8 Smalltalk 2 12 = 60% 6
9 gossip 2 13 = 65% 4 1
10 cackle 1 17 = 85% 1 1
11 pow-wow 7 1 9 3
12 prattle 2 16 = 80% 1 1
13 colloquy 4 2 12 = 60% 2
14 chit-chat 12 = 60% 4 2 2
15 tattle 1 13 = 65% 4 2
16 yap 2 15 = 75% 2 1
17 talk 10 = 50% 10 = 50%
18 idle talk 2 14 = 70% 4
19 trivial talk 2 14 = 70% 4
20 light talk 9 11=55%
ASSESSING LINGUISTIC B E H A V I O U R 115
II Evaluate the type of verbal event denoted by the following verbs as positive,
negative or neutral.
Put a tick in the appropriate column.
Table A2
Table A3
Table A4
Table A5
Table A7
Expression Positive Negative Neutral Blank
1 guard one's tongue 15 = 75% 1 3
2 let something slip 16 = 80% 3
3 spit out one's feelings 9 7 3
4 watch one's words 13 = 65% 6
5 bite one's lip 7 6 6
Questionnaire I: set
The nouns below denote a verbal event. Which adjectives from the following sets
do you primarily associate with them: (A) light or heavy or neither; (B) deep or
shallow or neither. Tick twice: choose one item from set A and one item from set
B. Put ticks in the appropriate columns.
118 A N N E - M A R I E SIMON-VANDENBERGEN
Table BI
A
Noun Light Heavy Neither Deep Shallow Neither Blank
1 chat 19 1 1 12 7
= 95% = 60%
2 babble 17 3 16 4
= 85% = 80%
3 dialogue 18 2 18 2
= 90% = 90%
4 chinwag 18 1 1 3 16 1
= 90% = 80%
5 conversation 9 11 9 11
= 55% = 55%
6 jaw 11 5 4 4 11 5
= 55% = 55%
7 natter 19 1 2 17 1
= 95% = 85%
8 Smalltalk 17 3 1 15 4
= 85% = 75%
9 gossip 17 3 3 15 2
= 85% = 75%
10 cackle 13 1 6 2 8 10
= 65% = 50%
11 pow-wow 3 10 7 9 4 7
= 50%
12 prattle 18 1 1 2 17 1
= 90% = 85%
13 colloquy 1 14 5 13 1 6
= 70% = 65%
14 chit-chat 20 1 17 2
= 100% = 85%
15 tattle 18 1 1 1 17 2
= 90% = 85%
16 yap 17 3 2 15 3
= 85% = 75%
17 talk 1 19 5 15
= 95% = 75%
18 idle talk 17 1 2 2 16 2
= 85% = 80%
19 trivial talk 13 1 6 2 14 4
= 65% = 70%
20 light talk 16 4 2 11 7
= 80% = 55%
II The verbs below denote a verbal event. Which adjectives from the following sets
do you primarily associate with them: (A) loud or soft or neither; (B) fast or slow
ASSESSING LINGUISTIC BEHAVIOUR 119
or neither. Tick twice: choose one item from set A and one item from set B. Put
ticks in the appropriate columns.
Table B2
A
Verb Loud Soft Neither Fast Slow Neither Blank
1 chatter 10 4 6 20
= 50% = 100%
2 natter 5 8 7 15 5
= 75%
3 converse 1 14 5 17 3
= 70% = 85%
4 palaver 16 1 2 10 5 4 1
= 80% = 50%
5 drone on 7 6 7 19 1
= 95%
6 gas 7 2 11 14 1 5
= 55% = 70%
7 rattle 11 2 7 20
= 55% = 100%
8 babble 9 6 5 20
= 100%
9 clack 14 2 4 8 2 11
= 70% = 55%
10 gabble 12 1 7 19 1
= 60% = 95%
11 run on 4 4 12 13 3 4
= 60% = 65%
12 prattle 4 4 12 18 2
= 60% = 90%
13 jabber 14 6 19 1
= 70% = 95%
14 patter 1 12 7 15 2 3
= 60% = 75%
15 splutter 7 4 9 9 1 10
= 50%
16 twitter 1 15 4 19 1
= 75% = 95%
17 talk nineteen 11 1 8 17 2 1
to the dozen = 55% = 85%
18 falter 12 8 1 15 4
= 60% = 75%
19 burble 3 11 6 9 6 5
= 55%
20 flounder 1 8 11 14 6
= 55% = 70%
120 A N N E - M A R I E SIMON-VANDENBERGEN
Table B3
IV Indicate whether the following expressions are typically associated with men,
women or both.
Put a tick in the appropriate column.
Table B4
Table B5
Table B6
Table B7
Expression Positive Negative Neutral Blank
1 blow off steam 12 = 6 0 % 3 5
2 have a loose tongue 1 18 = 90% 1
3 throw something out 5 2 13 = 65%
4 bridle one's tongue 7 3 10 = 50%
5 keep a stiff upper lip 6 9 5
122 A N N E - M A R I E SIMON-VANDENBERGEN
Questionnaire III
The expressions below all denote verbal events, types of verbal behaviour, ways of
talking, etc.
a) According to you, do they express a positive, a negative or a neutral value judge
ment? Put a tick in the appropriate slot.
b) Briefly explain why you judge the expression to be positive, negative or neutral.
c) Use the expression in a sentence to show the kind of context which you associate
with it.
1. a chinwag
a) Positive □ Negative □ Neutral □
b) Why?
c) Example sentence
c) Example sentence
c) Example sentence
ASSESSING LINGUISTIC B E H A V I O U R 123
4. to r u n on
a) Positive □ Negative □ Neutral □
b) Why?
c) Example sentence
c) Example sentence
c) Example sentence
7. to twitter
a) Positive □ Negative □ Neutral □
b) Why?
c) Example sentence
124 ANNE-MARIE SIMON-VANDENBERGEN
c) Example sentence
9. a chatterbox
a) Positive □ Negative □ Neutral □
b) Why?
c) Example sentence
c) Example sentence
Paul Pauwels
Catholic Flemish Institute of Higher Education
further insight into the interaction of different donor domains, the existence of
different levels of metaphorization, and the role of image-schematic structures.
Because so many of the concepts that are important to us are either ab
stract or not clearly delineated in our experience [...] , we need to get a
grasp on them by means of other concepts we understand in clearer terms.
(Lakoff and Johnson, 1980: 115)
From the above, it is clear that an essential part of metaphor is the presence of a
salient concept, which is used to structure a non-salient experience. Salience can
be a consequence of concreteness, visual impact, specificity, but also of perva
siveness and familiarity. It is important to note that salience is not an absolute
property, inherent in a concept, but that it is relative, in so far as it depends on
the experience of the language users who constitute the linguistic community. A
linguistic community, moreover, is never homogeneous, and the experiences of
its members may differ substantially, which is why salience, too, may differ for
individual language users. Furthermore, linguistic communities and their lan
guage evolve, and concepts may gain or lose salience. Because salience is rela
tive, metaphor, which relies on salient concepts, is relative too, and it becomes
important that one should think in terms of metaphoricity, i.e. metaphor as a
matter of degree.
In my account, I propose to use the notion of 'recoverability of donor do
main/situation' to describe the extent to which a specific donor concept, and the
way in which the metaphorical expression relates to it, are salient for a linguistic
community.
One problem which any investigation of a concept relying on the knowl
edge of the 'average language user' is confronted with, is verifiability. For the
purposes of this paper, I will take it that the expressions and definitions con
tained in a collection of explanatory dictionaries such as COBUILD, OALDCE,
LDOCE and Webster's provide a picture which certainly does not fall short of
the average language user's lexical competence. More specifically, I take it to be
the case that concepts and meanings not represented there, constitute specialist
knowledge, available to some language users, but certainly not to the majority
of the present-day linguistic community of native speakers of English. This leads
me to posit two basic kinds of metaphorical expression: those in which the
LEVELS OF METAPHORIZATION 127
metaphor relies on specialist knowledge, versus those where the concepts are
fully salient. When faced with metaphors relying on non-salient concepts, the
average language user will either (i) perceive/learn them as nonfigurative, or (ii)
perceive them as figurative and use them, or try to interpret them, on the basis
of knowledge available to him. In the second case he may well recover different
concepts and donor domains to give meaning to the metaphor than were at
work in the original - from a diachronic point of view - metaphorizationprocess.
But even when the concept used in a metaphor can safely be con
sidered as salient and transparent, the problem of verifiability remains: the
knowledge contained in general explanatory dictionaries can only provide an
insight in the possibilities which are open to most of the language community,
but does not show which of that knowledge will, or indeed needs to, be acti
vated in actual situations. It is likely that the use of a metaphorical expression in
a familiar pattern often precludes the need to recover its non-figurative basis,
whereas for example an ironic or punning use reactivates the metaphor and its
non-figurative basis, and heightens its metaphoricity.
In the remainder of this section, I will try to demonstrate how different
degrees of metaphoricity correspond to the degree to which, and the way in
which, a donor domain/situation can be recovered. Some of the examples used
here were taken from Pauwels & Simon-Vandenbergen (this volume).
A metaphor like bandy words with 'argue, quarrel with' has, for present-
day users of English, no recoverable donor domain. Bandy used to be a ball-
game like tennis, and the meaning the metaphor relies on, i.e. 'throw or strike a
ball to and fro, as in the games of tennis and bandy' (OED) is, like other literal
meanings of bandy, indicated as archaic; in our set of explanatory dictionaries
bandy is only mentioned in its LA-meaning. This metaphor would, in conven
tional terminology, be called 'dead' - the primary meaning of the expression
now lies in the L A domain, and the expression does not point the language user
to another domain. 2
The picture is different for expressions like tear off a strip and kick
against the pricks, which may not be completely recoverable as military degra
dation resp. the futile protest of oxen against the goad, but which are at least
partly recoverable. Both clearly fit into the more general donor domain of vio
lent actions, which is a frequent donor for LA.
An expression of the type get it straight from the horse 's mouth is in all
probability recoverable to a greater degree: it benefits from the presence of an
other expression requiring the same encyclopaedic knowledge (looking a gift
128 P A U L PAUWELS
horse in the mouth) and maybe also from the fact that the donor domain has a
greater degree of salience than in the cases described above.
Some expressions stand to be misinterpreted because the lack of specialist
knowledge may lead to wrong inferences. One such example is burden (of an
argument, story, message...), meaning the 'main subject or point' (LDOCE).
(1) The burden of his message did not strike me as being very original.
(COBUILD)
Here the meaning which should be recovered is not 'something that is carried, a
load' (OALDCE). This appears on closer scrutiny, since this meaning would
involve a reversal in value judgement from donor domain to recipient domain -
something which is unlikely, as is shown by the data examined in Pauwels &
Simon-Vandenbergen (this volume) - as well as a reversal of roles in the situa
tion: the burden in the metaphor has a neutral to positive connotation, and it is
described as being at the basis of the argument. Rather, the meaning which
should be recovered in that of bourdon 'bass, undersong, or accompaniment
'(OED), from which it was derived via 'the refrain or chorus of a song'(OED)
via figurative uses like (2) to refer to 'the chief theme, leading idea, prevailing
sentiment' (OED) of an LA.
(2) This was the burthen of all his song - 'Everything which we could
reasonably hope from war, would be obtained from treaty.' (OED)
It is not unlikely that there are other metaphors for which this kind of mistaken
identity is possible. The question which needs to be investigated is in how far
this influences metaphoricity.3
The viability of these metaphors probably depends to a high degree on the
way in which they fit into more general ways of conceptualizing, and in how far
this is recognizable. Another example may illustrate the point.
In an expression like lace into someone 'attack someone with words' the
use of lace to mean 'whip' may not be immediately recoverable. The presence
of other, similar, collocations in expressions like tear into someone and even
lam into someone - where lam, too, is intransparent - may help the language
user recover the relevant donor domain of violent action. The actual donor
situation might remain beyond his grasp, but the collocation with into someone
will probably serve to situate the expression with other, similar ones.
With blow the gaff 'let some secret be known' (LDOCE) the situation is
different. For the average language user, the original donor domain of sailing is
far from recoverable,4 and although a link to the domain of breathing may pro-
LEVELS OF METAPHORIZATION 129
vide some ground for understanding (blowing as contrasted with keeping one's
mouth shut), it does not go a long way. Individuals with a specialist knowledge
of sailing will find the element gaff less of a problem. For them, the metaphoricity
of the expression is higher. A similar case, also from the domain of
ships, is bilge 'foolish talk' (LDOCE). In this case, though, the item does not
form part of a larger pattern, nor is there a fixed collocation, so the lay person
has no basis for a possible (re)interpretation.
Of course, quite a few metaphors have easily recoverable donor domains.
Metaphors like eat one 's words and kick someone around are clearly and com
pletely recoverable from their donor domains of resp. eating and fighting.
Moreover, it is unlikely that these metaphors will ever lose their recoverability,
since they refer to basic bodily experiences, and as such have a near-ontological
basis.
The picture which emerges from the above can be outlined as follows.
Metaphors originate in situations/domains which are recoverable for the lan
guage users. Historical change may cause a change in salience of those concep
tual situations/domains, so that the metaphor becomes irrecoverable (dead, in
traditional terminology). In cases where the original literal meaning of a given
item falls out of use, the process of 'dying' can only be slowed down through
the presence of a (large) number of current metaphors drawn from the same
domain, which provide clues for each other's interpretation.
However, historical change can also have the opposite effect. Domains
which used to be 'specialist' can acquire the necessary overall salience to war
rant recoverability. The growing importance of computer and computer science
has given rise to a number of metaphors with this domain for a donor, and most
of these should be classified as recoverable. Feedback, for example, was origi
nally a metaphorization of the use of energy in an (electric) circuit: 'the return
of a fraction of the output signal from one stage of a circuit, amplifier etc. to the
input of the same or preceding stage ... to increase, decrease the amplification
etc.; also a signal so returned'(OED). From there, it was extended to the com
puter domain, where it later came to metaphorize the flow of (electronically
coded) information. In a final stage, it has become a metaphor for L A , describ
ing 'a response' (OED) as in (3).
(3) We began to get a fairly good feedback from most people who
know about it, and it looked as though the concerts would be
good scenes. (OED)
130 PAUL PAUWELS
Notice how the purpose described in the first of the above OED definitions is
paralleled in the aim of the L A of giving feedback, viz. the modification of the
behaviour of the system/person addressed. Further, it is interesting to see how
the metaphorization in this case also relies on an existing and at the time already
well-established pattern of meaning extension for the verb feed, which can mean
'gratify, minister to the demands of (a person's vanity, desire of vengeance, or
other passion); to sustain or comfort a person with (usu. fallacious) hopes'
{OED). In this usage, exemplified by (4), feeding also describes an action which
affects the mind of the addressee.
(4) He feeds himself and his friends with hopes of a speedy peace.
{OED)
Other expressions from the domain of computer technology, like input, output,
interface, etc. have been extended in similar directions. Together with the in
creasing popularity of the domain (or because of it), there occurs an increase in
the number of metaphors which relate to it as a donor domain.
The metaphors we have been discussing up to this moment all rely on a
specific donor situation, which may or may not be recoverable. In Kövecses'
(1986) terms, these would all be "basic level metaphors" which are "linked more
directly to experience". Kövecses' other class of metaphors, "constitutive meta
phors", are based on "superordinate concepts". Lie in one's throat/teeth 'lie
shamelessly' {LDOCE), as in (5), would be a case in point.
(5) Whoever asserts that lies in his throat. {OED)
The relevant domain for this metaphor is that of movement in space. More spe
cifically, any explanation would probably refer to image-schematic notions like
container, path, centre-periphery, as distinguished by Johnson (1989: 126). The
key factor in the metaphor is that the origin of the L A is situated along the path
of speech, but nearer the periphery of the body-container than one would expect
of something which, according to our cultural conventions, should preferably be
kept hidden. At this stage, again, the question of metaphoricity arises. Indeed,
the metaphor can be explained, but in how far the image it relies on is recover
able for the average language user seems a moot point.
From the foregoing, it should be clear that Kövecses' (1986) binary clas
sification is insufficient to fully grasp the complexity of different types of meta
phor. Although it is probably an accurate picture of the two basic ways in which
original metaphors can arise, it fails as a model to explain how conventionalized
metaphors are used and experienced, and how, and to what extent, the average
language user recovers which basis for a metaphorical interpretation. In what
LEVELS OF METAPHORIZATION 131
3. һ put-metaphors
For this investigation use was made of the LDOCE-bastd corpus of L A meta
phors as described in Vanparys (this volume), which was then further extended
on the basis of other explanatory dictionaries: Oxford Advanced Learner's Dic
tionary of Current English (1989 4 ) (OALDCE), Collins COBUILD English
Language Dictionary (1987) (COBUILD) and the Longman Dictionary of
Contemporary English (1987 2 ) (LDOCE).
The set of expressions under investigation is restricted in two ways. First,
there is the syntactic restriction to verbal expressions. Metaphorization patterns
for verbials, nominals and adjectives will probably exhibit differences, but these
are not our concern here. A second restriction imposed for this investigation of
a semantic nature. As in the main corpus, we chose to only investigate LA-
metaphors. Rather than focus on the variety of donor-domains used in meta-
phorizing LA, 5 we chose to focus on the extensions from one central scene
which is lexicalized by the verb put.
In all, the corpus consists of 87 sentences exemplifying 42 expressions.
The corpus is limited, and as such we cannot hope to provide a comprehensive
picture of metaphorization. Still, it is well suited to demonstrate the complexity
of metaphorical mappings, even in this limited set, and further highlight the
problems of recoverability and metaphoricity.
This section will provide a survey of the main senses of put in the spatial do
main, which is shown by a frequency count6 to be the primary domain for put.
132 PAUL PAUWELS
We will first describe the relations between those meanings, before attempting
to describe the extensions into the L A domain.
Basically, put can be described as a three place predicate which profiles
two trajector-landmark relationships (Rudzka 1988b: 508, henceforth TR and
L M ) . The basic patterns are exemplified by (6-13).
(6) She put the book on the table. (OALDCE)
(7) Did you put sugar in my tea? (OALDCE)
(8) She put the baby in its cot.
(9) It's time to put the baby to bed. (OALDCE)
(10) She put her arm around his shoulders. (OALDCE)
(11) Put down that knife before you hurt someone. (OALDCE)
(12) She put her hat and coat on. (OALDCE)
(13) She put some makeup on.
The prototype7 for put is exemplified by (6). We can try and describe this pat
tern in terms of basic image schemata. First of all, it singles out an entity as a
(manipulated) object. This object is presupposed to be something the agent is
holding in his hands - presupposed, because this is not something which is fo
cused by the verb. Put describes a movement away from the agent, which entails
a path, a direction, and two types of location i.e. an origin (which is not pro
filed) and a destination. The agent is in control of the object until it has reached
its destination. This destination is focused separately in the syntactic structure,
and is expressed by means of a prepositional phrase describing a location in the
spatial domain. The first TR-LM relationship holds between the agent and the
object, the second between the object and the destination. Two elements from
this prototype, i.e. the presence of a human agent as primary TR and the move
ment in space of a secondary TR to a secondary L M , constitute the common
core of the basic meanings.
The pattern exemplified by (7) is characterized by a difference in the na
ture of the secondary TR and L M . The sugar referred to can be (but need not
be) a substance rather than a single solid object, and the manipulation would
then be indirect, by means of an instrument such as a spoon. Also L M 2 is non-
prototypical, and the final relationship between TR2 and L M 2 is not one of
support, but rather one of 'merging'.
In (8) the secondary TR/primary L M is also different in nature, but this
does not affect the relationship with the primary TR. A baby can be conceptual
ized as being sufficiently small and manageable to be lifted, carried and left
somewhere in the same way as an object. This pattern with a human being as
LEVELS OF ΜETAPHORIZATION 133
primary L M is further extended via human beings not in physical control of their
bodies (asleep, drunk, ill or unconscious) and hence manipulatable, to human
beings not in mental control of themselves (e.g. prisoners). Example (9) shows
another extension of the pattern, which is metonymic in nature. The ritual of
putting a baby to sleep is referred to by means of a description of the central
element in the process.
Sentence (10) exemplifies another highly frequent pattern in the use of put
in the spatial domain. In this case, the primary L M is not an object, but a body-
part. The extension is quite straightforward, since hand and arm are both central
elements of the prototypical act of manipulating as conceptualized by put. Still,
the movement of other bodyparts such as head and foot is also conceptualized
by means of put.
Although the meanings illustrated by (7,8,10) are central, I would argue
they are also extensions from the prototype in that they all conceive the primary
L M as an object, which involves crossing a (sub)domain boundary. The question
is now in how far these domain boundaries are salient to the average language
user in this context. In view of the frequency and the well established nature of
these patterns, it is likely that metaphoricity is low. Still, the use of put strength
ens the conceptualization of a variety of primary LMs as manipulatable objects.
In cases like (10), the inherent duality of the human being between the mind as
the origin of agency and the body as the primary instrument of that agency is
exploited and emphasized.
In examples (11-13), LM2 is not expressed. According to Rudzka-Ostyn
(1988b) we are dealing with a 'covert' or 'hidden' landmark, which "does not
mean that the landmark is absent from our conceptualization of the scene; it
simply means that it is either too obvious or not worth bothering about" (p520).
Still, it should be noted that in all of these cases there is a particle/adverb further
specifying the verb. There are two main subtypes in this set. In (11), the hidden
L M can be inferred as being the agent (the primary TR) who functions as source
of the movement. In (12) and (13), it is the agent functioning as destination of
the movement. Example (13) can be seen as a further extension of the pattern in
(12), in that the putting involves a substance rather than an object. Still, (12) too
is already further away from the prototype, since it involves a more complex act
of manipulation, a nonprototypical secondary L M and a more complex final
relationship between the secondary TR and L M .
In sum, it should be clear that put has several (non)figurative extensions in
the spatial domain, and even in the more specific subdomain of manipulating
134 P A U L PAUWELS
For each of the expressions under investigation, I will try to determine the ex
tent to which a donor situation can be recovered from the spatial domain, and
the relevance of the prototypical meaning of put. I hope to demonstrate that
transfers occur at different levels, and that metaphor is a relative notion.
Figure la
Put one's cards on the table 'clearly state one's intentions' (LDOCE). A unifying scene
in the donor domain at the basis of the metaphorical transfer is unequivocally recov
erable. (Full lines: in focus; dotted lines: out of focus; T: transfer; R: recovery.)
The main distinction to be made is that between expressions which rely on one
clearly recoverable donor scene (Fig. la), and those that can only be recovered
from different (superordinate) donor domains (Fig. lb). In the second group,
present-day average language users have no access to a nonfigurative scene
where the elements present in the metaphor are unified. In the first group, we
find expressions which are recoverable from the domain of the manipulation of
objects (3.3.1), specific subdomains of this more general domain (3.3.2.), other
domains which have been metaphorized through the domain of manipulating
objects (3.3.3), especially the domain of bodily movement (3.3.4).
LEVELS OF METAPHORIZATION 135
Figure lb
Put in for something 'apply for ' (LDOCE) There is no single unifying scene on which
the metaphor relies, for present day users. Recovery in various (superordinate) do
mains.
Figure 2
Put on the agony 'to say that one's sufferings, feelings etc. are stronger, greater, or
worse than they really are, esp. in order to influence others ' (LDOCE)
There are four possible paths of recovery, viz.
(1) putting on a mask, hence hiding;
(2) putting on a play, hence acting;
(3) putting on a machine; hence activating;
(4) putting on an extra measure; hence exaggerating.
These rely on two different strategies of recovery, viz.
(a) that which is being 'put on': a mask (1), or a machine (3);
(b) that which it is being 'put on ': a stage (2) or a pile (4).
Put the record straight 'give a correct account of facts or events' (COBUILD)
and put someone right/straight 'give correct information to someone who has a
wrong opinion, often rather sharply'(LDOCE) both rely on the donor domain of
arranging objects. As in arranging, what is correct or right depends on the
judgement of the agent. This is especially relevant in the second metaphor,
where control over the hearer-object is in focus. On top of that, this expression
has another possible donor situation where persons are actually manipulated;
soldiers on parade (or schoolchildren for that matter) are put straight in the
literal sense.9 This may be done by hand, or by means of language, which again
LEVELS OF METAPHORIZATION 141
provides a metonymic basis for the metaphor. In the latter interpretation, the
metaphor is once again based on an extension from the prototype in which the
secondary TR is a person not in control of his own body.
Put someone up (for something) 'suggest (someone) for a job, position
etc.' (LDOCE) relies on the situation of putting up notices or pictures. In cer
tain contexts it has an element of metonymy as well, as in (22)
(22) They put up several candidates in Cornwall... (COBUILD)
The names and pictures of candidates in elections are literally put up on walls,
hoardings etc. This metonymic link was probably the basis for the metaphor,
which later got a more general meaning. Put something up 'present (an idea
etc.) for discussion or consideration'(OALDCE) seems to have a similar basis.
Both expressions ultimately rely on a more general mapping of spatial concepts
where up means visible, as contrasted with down.
Another metaphor which is probably recoverable from a more or less
clearly recognizable donor situation in the field of the manipulation of objects is
put across/over (OALDCE) oneself/one's meaning 'cause to be understood'
(LDOCE), which is clearly regarded as a skill, judging from examples (23-28).
(23) That speaker doesn't know how to put himself across. (LDOCE)
(24) I'm not putting my meaning across very well. (LDOCE)
(25) It's very hard to put across the facts. (COBUILD)
(26) He doesn't know how to put himself across at interviews.
(OALDCE)
(27) You need the skill to put your ideas across. (COBUILD)
(28) She's very good at putting her ideas across. (OALDCE)
The donor situation is that of moving something across a barrier or a gap, to
another person. Again, one can trace a metonymie element, as people involved
in L A are prototypically seated across each other, and in quite a few situations
with an object/obstacle in between (table). In the case of speaking before an
audience, there is even a real gap, which consists of the physical distance and
very often also the difference in physical level between speaker and audience. In
putting oneself across, as in (23) and (26) there is a metonymy embedded within
the metaphor, of the type 'speaker for utterance' (see Goossens (this volume).
The collocation with the negative not seems to focus the lack of (self)-control
or skill exhibited by the speakers.
142 PAUL PAUWELS
(31) The letter from her daughter set her mind at rest. (L2)
We are dealing with descriptions of an intended perlocutionary act, and in (31),
for example, it is not even clear whether it is the contents of the letter or the fact
of getting the letter which does the trick. It could be claimed in view of other
evidence (expressions like peace of mind, turn something over in one 's mind)
that we are actually dealing with a metaphorization of a state of mind, on which
another metaphor has been superimposed. The mind is conceived as moving,
worry as excessive movement, and rest as lack of movement. Expressions like
turn/set one's mind to something, and a mind working at full speed demonstrate
that the mind is conceived as a piece of machinery. On this piece of machinery,
various operations can be performed, among other things via LA: in (30-31)
affecting the mind by L A is described as a manipulation of the machinery - some
kind of adjustment, or tuning. The metaphorization of the L A is in this case
shown to be closely linked with, or even dependent upon the metaphorization of
the mind. Given the generally accepted relation between language and thought,
this should not come as a surprise. The core of the LA-metaphor is the control
which is achieved by means of language.
L A of the stated type. Both expressions have shifted away from metonymy to
cover less specific contexts as well, and this is made possible by the meta
phorical pattern they fit into. As in the other metaphors, control over one's body
is stressed in the paraphrase. Put one 's foot down may also fit into the more
general pattern of making a stand, and could hence be looked at from a com
pletely different angle: the situation of horsemen dismounting to stand and fight.
We would then be dealing with an example of the 'argument is war' system.10
Here, however, the recoverability is questionable.
As was already clear from the foregoing, metaphors very often rely on more
than one cognitive donor domain. The previous set of expressions could still be
situated in one rather specific domain, with in some cases a secondary aspect
contributed by another domain. In the set which now follows, various domains
make equally important contributions.
At the outset (3.2.) it was stated that the put-prototype could be specified
through the type of secondary TR and L M , and the relation between them as
expressed in the preposition/particle. In the previous paragraphs we have seen
how such specifications allowed us to recover donor domains in different ways:
through the secondary TR (put one's foot in it), through both secondary TR and
L M (put the blame on someone), through the secondary L M (put the mouth on
someone), and, finally, via relations which are typical of certain situations (put
someone up for something). In the latter case, the metonymic basis for the
metaphor also played an important role.
In what follows, the specifications do not support the recovery of one do
nor domain: the secondary TR and L M , if specified, are usually specified within
the L A domain, and the prepositions which specify relations between them are
either atypical in combination with put (put over on, put off) or too vague (put
in). There is one exception to this pattern. In put words into someone's mouth,
the secondary L M is specified outside the L A domain - but mouth is of course
still related to LA. Still, the specification does point to a clearly identifiable do
main, so we should consider whether this expression is maybe a borderline case.
Put words into someone's mouth is also treated in Pauwels & SimonVandenbergen
(this volume). It is classified there as a case of 'abstract' meta-
LEVELS OF METAPHORIZATION 145
ting go of something. In this metaphor, the second element, about, brings this is
focus: when things are about they are per definition scattered at vague loca
tions. This metaphor uses the contradiction between the control which is present
when one gives information to an individual, and the lack of control once that
information reaches more than one person and starts to spread.
3.4.3 Vagueness
The final set of expressions do rely on prototypical uses of the scene of putting,
but it is in no way clear on exactly which use each relies. Put in(to) and put out
occur in so many different situations, that it would be unacceptable to pin them
down on just one donor domain. The notion which is central here is the con
tainer schema. It is used to describe different aspects of the L A situation: the
addressee (put in a request/claim, put in for), the ongoing L A or conversation
(put in, put in a good word for), language (put feelings into words), the speaker
(put out a statement).
Put feelings into words, as in (36) and (37), conceives of words as con
tainers of meaning (a frequent metaphor), and feelings as manipulated objects.
This metaphor fits the conduit metaphor, as described by M.Reddy (1979: 286
ff).
(36) They cannot put their feelings into words. (COBUILD)
(37) ... trying to put her feelings into words. (LDOCE)
The examples seem to emphasize the difficulty and the skill required - feelings
are difficult to 'grasp' and to control. A variant on this metaphor allows for
quantification, which is implicit in (36-37), but comes out more clearly in (38).
(38) He put a lot of feeling into it. (my example)
In put in and put in a good word for, we are once again dealing with a hidden
landmark, which can be construed as ongoing LA, conceived as a container (see
also Rudzka-Ostyn 1988b: 529). The speaker is perceived as putting a contri
bution into this container. One of the effects is to reify that contribution, which
is then open to quantification. The container itself is in both metaphors con
ceived as closed, and rather difficult to get into, which results into connotations
of interrupting as in (39) and (40) or conceives entry as a privilege as in (40)
and (41). The ongoing interaction is conceived as the privileged territory of the
participants, which remains closed to 'outsiders'.
(39) "But what about us," he put in. (LDOCE)
(40) Could I put in a word at this point ? (OALDCE)
LEVELS OF METAPHORIZATION 149
(41) I'll put in a good word at the meeting if I get a chance ...
(COBUILD)
Figure 3
Recovery on superordinate domain; donor situation not fullly salient, or insufficient
for a full grasp.
1. put words into someone's mouth 'tell so what to say/suggest or claim falsely that so
has said a particular thing' (LDOCE);
2. put forward 'offer (a suggestion) for consideration' (LDOCE);
3. put in for 'apply for' (LDOCE);
Similarly, put in a request/claim, put in for and put out a statement, where ad
dressee, resp. speaker are conceived as containers, seem to rely on the more
general idea of territories closed to outsiders. The addressee/resp. speaker in
these metaphors is an organisation, which is often metonymically identified with
its office, which is of course literally inside a building. The relevant aspect in the
metaphors is again lack of access, which leads to lack of information about what
goes on inside. It is possible to find a parallel here with the metaphors based on
put up (3.3.2) and put forward (3.4.1). In all cases there is a probable me
tonymic basis in the manipulation of written documents. On the basis of the
parallel, we should construe the hidden L M 1 of put in for as oneself. Self-
determination/control would again be a relevant feature in the metaphorization,
which is logical in context like those provided by (42) and (43).
150 P A U L PAUWELS
3.5 Metonymy
Finally, the corpus contains five expressions which are basically metonymic. In
put, put (the) pen to paper, put something down, put someone down for some
thing, and put a sock in it the action described is usually contiguous with L A ,
i.e. "the mapping occurs within a single domain." (Goossens 1990b: 325)
Put in isolation can be used to refer to the process of writing, as exempli
fied in (44-46).
(44) I don't know what to put. {LDOCE)
(44b) Put a question mark. (LDOCE)
(45) Put your name here. (OALDCE)
(46) Put all the details on the card. (COBUILD)
(46b) He couldn't read what Ken had put for his address. (COBUILD)
From these examples, and the definitions (44'-46') which accompany them, it is
clear that the process of writing here is only considered in its 'physical' aspect,
and not as a full LA.
(44') write down, make a written mark (LDOCE)
(45') write or mark something on something (OALDCE)
(46') write or type (a particular word, sentence etc. somewhere)
(COBUILD)
In other words, we are not really dealing with a metonymy for L A , but with a
metonymy based on an act of manipulation which accompanies the L A of writ
ing. The mapping occurs within the L A domain, so we are clearly dealing with
metonymy.
Put (the) pen to paper 'start to write' (LDOCE) is also a clearcut case.
The only divergence from a purely metonymic interpretation here lies in the
possible use of another instrument of writing; in that case the pen would not be
part of the scene which it is used to describe. The requirement of contiguity
would be broken. Still, one would hesitate to call this metaphor, since we are
still moving within the same cognitive domain of writing. In cases where there is
LEVELS OF METAPHORIZATION 151
an obvious conflict - e.g. someone sitting down at a PC and saying (47) - the
divergence would in all probability be understood as ironic and metaphorical.
(47) Let's put pen to paper. (my example)
Also put something down relies on the prototypical scene of writing, where the
movement of the pen (or other writing utensil) towards the paper follows a
downward trajectory. Again contiguity is not assured in all cases - writing on a
wall would be an obvious case, but then again it would not normally be de
scribed by means of put down. Here, though, there is a metaphorical potential in
the combination which can be interpreted as an act of controlling - for instance
in order to remember as in (48).
(48) Let me put down your telephone number before I forget.
(LDOCE)
Writing moreover is a less volatile means of communication than speech, which
strengthens this possibility of interpretation.
Put someone down for something is basically the same. Here, it is the
presence of a human object for put brings out the dimension of control. Again,
this corresponds to the generally perceived value of a name on a list: if your
name is put down, this is usually binding. Moreover, the expression contains a
metonymy of the type person-for-name. This second metonymy further en
hances the importance of the control element, since it is not 'just' the name
which is on the list.
Put a sock in it, finally, stands apart in that it is only used performatively,
as in (49).
(49) "Why don't you put a sock in it. I'm trying to work." (OALDCE)
Here, the act of gagging, which entails silence, is called up indirectly, since we
are dealing with an action the addressee is supposed to perform on himself. The
expression again illustrates the way in which metaphor and metonymy interact.
There is an element of contiguity, since gagging entails silence, but the relation
ship is not symmetrical, since silence does not entail gagging. It is not clear
whether in this case one could say that the expression is used to 'refer to' the
act of shutting up. This would lead us to posit a metaphorical interpretation, in
which shutting up (in a certain context) is explained as an act of gagging one
self. The expression draws attention to an apparent unwillingness of the ad
dressee to shut up, the fact that this is considered to be a nuisance, and the pos
sible solution to this problem.
152 P A U L PAUWELS
Figure 4
Metonymy and the extension into metaphorization (dotted lines)
put a pen to paper 'start to write', put sth down 'write sth down'
4. Conclusions
This investigation has shown that different types of metaphors can be distin
guished along different dimensions. First of all, there is the dimension of meta
phoricity - i.e. the degree to which an expression is perceived as metaphorical
by the present-day average language users - which depends on the recoverability
of salient donor situations or donor concepts. Second, there is the dimension of
specificity. Some metaphors rely on clearly recoverable, highly specific, situa
tions, which are salient because of their visual, sensory etc. imminence. Other
metaphors rely on the recoverability of image schematic structures, which are
salient because of their pervasiveness.
LEVELS OF METAPHORIZATION 153
The contribution of put varies according to the specificity of the metaphor. With
the most specific cases, put at most only adds a secondary aspect. In ex
pressions like put one 's cards on the table the contribution is minor: it is the
scene as a whole which is recovered, and the focus is not on the mapping of put
LEVELS OF METAPHORIZATION 155
Next, I would like to come back to a number of metaphors based on the con
tainer schema. Put in, put in a good word for, put in for, put in a request/claim
and put out a statement do not seem to rely on a specific donor domain. Hence,
one would assume that they follow the pattern of metaphors based on different
donor domains, but this does not seem to be the case. Conversely, the meta
phors seem to rely on the conception of the act of putting in/out, without fur
ther specification.
One reason for this could be the fact that the container schema, in contrast
with other image schemata, is still a relatively rich concept, with a large number
of entailments, such as contents (fluid or object), central and peripheral ele
ments, an inside-outside orientation, a containing structure, and a potential for
movement in and out and for quantification (full-empty). Container is a superordinate
concept, as is clearly demonstrated by basic level realizations like
cup, box, glass etc. Still, one could hypothesize that this is only true in one per
spective, while from another angle (the domain of objects), it is more specific.
156 P A U L PAUWELS
Another factor which plays a role in this set of expressions, is the strong
interaction between put in and the inferred destination (i.e. addressee, or ongo
ing L A ) . To begin with, it is striking that we are dealing with particle verbs,
where the destination of the putting is not specified, in cases where there is a
hidden landmark. The interaction, then, concerns the conflict between an act of
putting in and the perceived nature of the destination, on the image schematic
level. In putting in, one clearly enters a domain, and if that domain is another
person (addressee) or something which is perceived as belonging to someone
(ongoing L A ) , this entails a clear violation of territory, and of control. The
conflict can in some cases be resolved by relinquishing control over that which
is put in. In put in a good word for, for example, the limited nature of the con
tribution seems to indicate a minimal violation of territory. The good word is put
in, and then left to undergo the forces of ongoing interaction. With put in, on
the other hand, the direct object allows for more variation. Examples (39) and
(40) show that the situation is more confrontational; in being a request, (40)
clearly shows who is perceived to be in control of the situation.
(39) "But what about us," he put in. (LDOCE)
(40) Could I put in a word at this point ? (OALDCE)
Put out a statement is different, since there is no conflict situation. The speaker
is in full control, but chooses to move something out of his domain so that oth
ers can act upon it. That which is put out gives the outsiders an idea of what is
inside, or going on inside. In this way they do acquire some kind of control.
Again it is the perceived nature of the object which plays a central role: in this
case the statement is perceived as still linked to its originator.
ing). The largest set of expressions refer to the agent as landmark, or to his en
vironment with respect to his position in that environment. Here we include put
about lies, put someone off, put something/someone/oneselfforward, put out a
statement. In two expressions the movement of a bodypart is profiled with re
spect to the agent (put one's foot down, put out feelers), and one expression
uses a bodypart (face) as landmark (put on the agony). In this respect, the AmE
expression put one's foot in one 's mouth offers an interesting insight in the in
terpretation of metaphors. It makes explicit the construal of the landmarks in
terms of the recipient domain, whereas the BrE expression put one's foot in it
relies on the donor domain for the recovery of its vague landmark.
In this paper I have introduced the notion of recoverability to describe the ex
tent to which, and the way in which, the average language user is able to link a
metaphorical expression to a (usually non-figurative) donor scene. I have pre
ferred recoverability to salience, because, to my mind, it is not the overall sali
ence of the donor scene or concept which is at stake, but the salience of that
scene together with the availability of the scene in using the metaphor. In other
words, a scene or situation may be salient, but still not be recoverable to the
language user in the context of a certain metaphor. What makes the donor scene
recoverable seems to be
(i) the overt signalling in the metaphorical expression,
(ii) the awareness of other, similar, metaphorical patterns or mappings,
(iii) the overall salience of the donor scene or concepts, and
158 P A U L PAUWELS
(iv) possibly also the wider context in which the metaphor is embedded.
The contributions of (i-iii) have been discussed and analyzed in this paper. The
role of (iv) remains beyond the scope of this investigation, since we worked
from a lexicographic database, and we have no access ot the context beyond the
example sentences.
Finally, the judgements on recoverability in this paper are subject to two
limitations. To begin with, they rely on a notion of average language user as
defined in relation to a number of explanatory dictionaries; furthermore, the
interpretation of the data is my own, and although it has been verified with a
number of competent language users, there has been no empirical testing of
either the specific interpretations or indeed the model of interpretation.
Notes
L o u i s Goossens
University of Antwerp
1. Introduction
The purpose of this paper is to deepen our insight into the ways in which
metonymy interacts with metaphor in figurative language. Although in principle
metaphor and metonymy are distinct cognitive processes, it appears to be the
case that the two are not mutually exclusive. They may be found in combination
in actual natural language expressions. In that sense there might be room for the
neologism in the title of this paper, for which I suggest the phonological
realization [metæf'tɔnimi] to help the reader along if (s)he wishes to know
whether the word is also pronounceable. It will be shown, however, that the
interaction can take several forms, for which a single term may be misleading
rather than helpful. In other words, I would like to assign metaphtonymy the
status of a mere cover term which should help to increase our awareness of the
fact that metaphor and metonymy can be intertwined.
To explore the interaction I have used a data base made up of figurative
expressions where the target domain is that of linguistic action. This database
is restricted and can therefore not be expected to provide an exhaustive account
of the possible interaction patterns. On the other hand linguistic action is
sufficiently complex and the data base exhibits enough diversity to allow us to
come up with the main patterns, which in the final section of this paper will be
put into a somewhat broader perspective.
In what follows I first remind the reader of a couple of basic insights into
metaphor and metonymy (section 2) as well as into the target domain (section
3). Next, I provide a brief account of the data base and of the donor domains
160 Louis GOOSSENS
figuring in it (section 4). Sections 5, 6 and especially 7 constitute the bulk of the
paper: in them I explore the different ways in which metaphor and metonymy go
together for the three donor domains in the data base in succession. Section 8
surveys these findings and tries to come up with a few generalizations about the
interplay between metaphor and metonymy.
In other words, the crucial difference between metonymy (as well as synecdo
che) and metaphor is that in a metaphoric mapping two discrete domains are
involved, whereas in a metonymy the mapping occurs within a single domain.
Obviously the hierarchy among cognitive domains, as well as their de
limitation, which are important areas for exploration within cognitive linguistics
anyway (see e.g. Langacker (1987: chapter 4)), are important issues in this
context. For the purposes of what follows we simply posit the existence of
complex domains built up by the combination of other domains which them
selves may either be complex or basic in the sense of Langacker (1987). It
should also be expected in this view that the boundary lines between domains
are often fuzzy, which is one of the reasons why metaphor and metonymy may
interpenetrate.
4.1.2 Sound
More specifically, our third donor domain is that of physical violent action,
which itself is a subdomain of the vast domain of human action. Again this is an
important donor domain for linguistic (inter)action, not unexpectedly, given the
connection with the Argument-is-war-metaphor identified by Lakoff and
Johnson (1980).
METAPHTONYMY 163
As a rule the donor domain is clearly distinct from the target domain, there is a
mapping from one domain into another, hence we get pure metaphors. Typical
examples are:
• bark 'say something in a sharp loud voice': the loud, penetrating sound of
barking dogs is mapped onto linguistic action where the sound is perceived
as loud, harsh or sharp;
• blow one's own trumpet 'say good things about oneself, perhaps immodestly,
so that others will know them' : the public and festive character of trumpet
blowing, in combination with the added reflexive dimension, is mapped onto
self-praise. To the extent that we may conceive of a scene in which the
trumpet blowing is followed by a public statement in which the announcer
'says good things about himself', we might accept a metonymic basis for the
expression. Since such an interpretation is far removed from the prototypical
scene of trumpet blowing, however, such a metonymic basis is very weak, to
say the least.
The items in the first group, on the other hand, but only those, usually have a
metonymie ingredient. Let us have a closer look at giggle 'express by or utter
with a giggle' as a paradigm case. A typical example would be (1).
(1) 'Oh dear', she giggled, ' I ' d quite forgotten'
One interpretation is that she said this while giggling: in that case there is a
synecdochic relationship; we express part for the whole, we have a pure
metonymy. Another way to interpret it is that she said this as if giggling; hence
there is a crossing of domain boundaries, we have a metaphor. Very often,
however, it is not clear whether the domain of giggling and that of
(light-hearted) linguistic action are conceived separately. We are still aware of
the metonymie basis in the metaphorical interpretation. This is what I would like
to call metaphor from metonymy. Figure 1 tries to visualize this. On the left
hand side of the figure two potentially discrete domains, A and B, intersect; they
are fused in a single scene (the surrounding circle). On the right hand side, A
and are separated, but, as the broken arrow indicates, there remains a
conceptual link with the scene in which the two are together.
The double possibility (metaphor from metonymy or metonymy only)
holds for most items in group (i) (19 out of 27). Obviously for some of them the
metaphorical reading is the natural one, for example, for applaud 'express
strong agreement with (a person, idea, etc.)', as in (2).
(2) These changes will be applauded
METAPHTONYMY 165
Another example of this sort is snort 'express anger in a snorting way', where it
is difficult to conceive of a scene where the non-linguistic and the linguistic
activity go together and where we get an as-if-reading, hence a (pure) meta
phor.
In the whole subcorpus of 100 verbials there were only 6 or 7 items for
which a metonymic ingredient can be suggested. All of these are of a type where
the violent action could be accompanied by verbal action, for example, throw
mud at 'speak badly of, especially so as to spoil someone's good name unnec
essarily'. It is conceivable that people may combine the violent action with
shouting names, which is linguistic action: this would be an instance of
metonymy. The metaphorical interpretation can easily be established, however,
without this metonymic backing. What this adds to our insight, is that metaphor
from metonymy occurs with varying degrees of cognitive saliency; instances like
the one discussed here provide us with the limiting case. Another instance of
this sort is give a rap on/over the knuckles 'attack with sharp words', an
item which also occurs in the body part corpus.
• 10 items are connected with the hand, including hand itself (6 times),
palm (1), finger (1), knuckles (2).
• The leg or part of it are represented 7 times: legs (1), foot/feet (4), knee(s)
(2).
• Finally, there are two items with blood, again a 'body part' only in the loose
sense of the word.
Secondly, it must be pointed out that the role played by the body part plays
varies according to whether we have a verbial, an adjectival, or a nominal. In
the case of verbials and adjectivals the body part is necessarily integrated into
some broader scene. Nominals, on the other hand, may be directly related to an
aspect of linguistic action, though also here there may be a combination with
another item, so that the body part is instrumental in a broader scene as well. As
we shall find, this considerably increases the complexity with which metaphor
and metonymy may interact.
pure metonyms, and mixed cases in the corpus. I have added the distribution
over the verbials, adjectivals and nominals, because it is not insignificant.
In the context of this paper it is, of course, the mixed cases that are of interest;
they will be explored under the following subheadings. Before proceeding with
that discussion, let me draw the reader's attention to the high proportion of pure
metonyms for the nominals as opposed to their complete absence for adjectivals.
This hangs together with the fact that it is easier to select entities which are part
of, or otherwise associated with, other entities as representatives for those other
entities than it is to represent properties by partial or associated properties
(where I take for granted that the categorial meaning for nominals is the
denotation of entities and for adjectivals the denotation of properties). As will
appear from the instantiations for the mixed cases, it is usually (but not always)
the integration of a nominal element into the verbial or the adjectival that is
responsible for the metonymic ingredient in an otherwise metaphorical context.
This pattern, which frequently occurs when the donor domain is (non-linguistic)
human sound (see section 5), is also well represented in the body part data. In
my analysis there are 27 items (24 verbials, 2 adjectivals, 1 nominal), i.e.
practically one fourth of the data, that belong here. Again the boundary lines
with pure metaphors and pure metonyms are sometimes a little hazy, but there is
no doubt that the great majority of those 27 can safely be assigned to this type.
For all of them it is possible to use them metonymically, that is with reference to
a scene where both the non-linguistic and the linguistic action reading are
relevant, and it is that metonymic reading which is the basis for the metaphorical
use. As a rule, however, there is an idea of transfer from a distinct scene; in
other words, we get metaphors for which there is a link with their metonymie
origin. In the following exemplification it will also appear that the relevant
scenes have to be characterized in their own right; the body part is just an
ingredient in a broader scene. This accounts for the fact that there is no
significant correlelation with any of the subgroups distinguished in 7.1. Let me
provide a few instances now with a word of explanation.
• Say something /speak / talk with one's tongue in one's cheek 'say
something and mean the opposite, esp. in an insincere or ironic way'. The
metonymie basis is a scene in which someone literally (and visibly) pushes his
tongue into his cheek while saying something that he does not really mean; in
this metonymie reading - unlike in the (admittedly improbable) literal inter-
METAPHTONYMY 169
In this pattern, which appears in the body part data only, we get metaphors
(involving therefore a mapping from a donor domain A onto the discrete target
domain B, which in our data evidently is linguistic (inter)action), but with a
built-in metonymy. This metonymy involves the body-part which is a shared
element in both domains (A and B). This situation can be pictured as in Figure
2, where the shared element, the body-part, is represented as x. Because of its
different function in the two domains, it is differentiated as and x' in the donor
and recipient domain.
170 Louis GOOSSENS
Note that this representation does not yet give us the whole story about the
shared item x/x'. A couple of examples will show that as a rule it functions
metonymically in the target domain only, whereas it is interpreted literally or
(more often) (re)interpreted metaphorically in the donor domain.
• Bite one's tongue off <informal> 'be sorry for what one has just said',
typically in contexts like (3).
(3) I should/could bite my tongue off
Here tongue can be processed literally in the donor scene. Because of the
counterfactual contextualization this donor scene can only be one that does
not directly tie up with everyday experience. Perhaps the best way to charac
terize it is in terms of self-punishment, where the punishment hyperbolically
involves a rather unlikely kind of self-mutilation. Mapping this onto linguistic
action we get something like 'depriving oneself of one's ability to speak',
where the metonymy is from tongue to the speech faculty as a whole. The
hyperbolic nature then generates an implicature in the Gricean sense along
the lines of 'I'm terribly sorry for having said something so foolish, rude, or
the like'.
• Shoot one's mouth off 'talk foolishly about what one does not know about
or should not talk about'.
The donor domain is the foolish or uncontrolled use of firearms: the foolish
(and therefore potentially, though not intentionally, dangerous) use of a gun
is mapped onto unthoughtful linguistic action. By integrating mouth into a
scene relating to the use of firearms it is reinterpreted as having properties of
a gun in the donor domain; this is the metaphorization in the donor domain.
In the recipient domain, however, there is a first level of interpretation which
amounts to something like 'using one's mouth foolishly', in which mouth is
a metonymy for speech faculty. Hence an utterance like (4) comes to mean
'Don't say anything rash'.
METAPHTONYMY 171
Also this type is represented by one instance only: be/get up on one's hind legs
'stand up in order to say or argue something, esp. in public'.
172 Louis GOOSSENS
The peculiarity about this item is perhaps best revealed if we leave out
hind: being I getting up on one's legs with reference to 'standing up in
order to say something in public' is metonymic, there is an overall scene of
somebody standing up and saying something publicly. The addition of hind
forces us to reinterpret in terms of an animal standing up. This suggests a
greater effort, an event which attracts more attention. At the same time there is
a bathetic effect, because a human being is interpreted as being involved in the
pseudo-achievement of standing on two legs. One may, of course, also argue
that the addition of hind makes the expression as a whole metaphorical; it is
only to the extent that we process it with an awareness of the metonymy, that it
is more adequate to view this as a metaphor embedded into a metonymy.
L o u i s Goossens
University of Antwerp
1. Introduction
In this paper I will try to add to our insight into metonymy, its relationship to
'literal' usage and to metaphor, and its conventionalization. Making use of data
from Ælfric, Chaucer and Shakespeare (the three 'respectable horses' of my
title), as well as of the categorization of metonymic expressions with mouth
proposed in Goossens (1993b), I will clarify some of the ways in which
metonymy works as a tool for meaning extension in a diachronically diversified
data base. As regards my understanding of metonymy within the framework of
Cognitive Grammar I have profited from Croft (1993).
It will be shown, amongst other things, that metonymy shades off into
'literal' meaning extensions, but also that it interacts with metaphor along the
lines indicated in my Metaphtonymy paper (this volume), and that there is
something of a metonymy-metaphor continuum. The diachronic spread of our
data (which range from Old English over Middle English to Early Modern
English, and occasionally include Present-day items) will allow us to have an
eye for at least some aspects of conventionalization with respect to metonymy.
As such this paper brings together a number of insights about metonymy
formulated in three earlier papers of mine (Goossens 1990b, 1993a and 1993b),
and supplements them in various ways.
176 Louis GOOSSENS
1.2 Structure
2. Metonymy
but "[s]ince Proust's claim to fame is that he is a writer, and the work produced
is a salient element in the domain of creative activity, the metonymic shift is
quite natural" (p. 348).
Still following Croft, we accept that the difference between intrinsic and
extrinsic is a continuum, where (2) exemplifies the highlighting of very intrinsic
facets of a concept, and (3) and (4) the highlighting of clearly extrinsic ones. In
between are instances like (5) and (6).
(5) She came in through the bathroom window (Croft's (24))
(6) We need a couple of strong bodies for our team (Croft's (25),
taken over, as the reader may have found out, from Lakoff and
Johnson)
In (5) "[t]he interpretation of [WINDOW] as an opening in the shape domain is
somewhat extrinsic because it makes crucial reference to what is around it -
contrast the use of window to describe a physical object in a hardware store
showroom - though it appears to be less extrinsic to the concept [WINDOW] than
the publishing company and writings in examples (18) and (20) above [our (3)
and (4)]" (p. 349-350). Whereas (5) is still intrinsic enough not to be taken as
an instance of metonymy, (6), which exemplifies synecdoche, is generally
treated as a subtype of metonymy: the selection of bodies "is sanctioned by the
need to highlight the physical strength/ability domain underlying the domain of
human beings" (p. 350).
Summarizing this position, we accept with Croft that metonymy is part of
a more general process, domain highlighting, where metonymy involves the
highlighting of an extrinsic entity/facet in a possibly nonprimary domain of the
concept. I take instances like (5) to be out, but synecdoches like (6) to be in; but
it follows from this conception of metonymy that the boundary with 'literal
uses' will not always be easy to draw.
3. Metaphtonymy
turned up in our diachronic data: see sections 6.2 and 6.3 respectively. In
addition, also demetonymization, one of the less current types, showed up.
Again we do no more here than refer the reader to the metaphtonymy paper and
to the discussion in 6.4 in the present one.
The only point that we would like to emphasize here, is that in the same
way as we can discern a continuum from 'literal' to 'metonymic', there is one
from metonymy to metaphor. The difference between metonymy and metaphor
is that for the former the mapping occurs within the domain matrix of a given
concept, whereas for the latter the mapping is between two discrete domains.
To the extent that two domains can on some occasions be conceptualized as
joined together in one domain matrix, and on others as being dissociated, which
is exactly what we claim to be characteristic of metaphor from metonymy, an
area can be postulated where the distinction between metonymy and metaphor
gets blurred. The rule is, however, that we have a clear notion in specific
contexts which of the two is at stake. In our corpus a few instances of metaphor
from metonymy turned up in context, though it also appeared here that its main
importance is as a diachronic process.
4. The data
Old English (as I found out in the preparation of Goossens (1993a; for some
evidence, see there). Given its frequency in Chaucer, and especially in Shake
speare, moreover, there are good grounds to believe that it is an extremely
suitable item for the kind of diachronic investigation that I wish to undertake
here.
Granting then that this is a somewhat restricted database in more than one
respect (only 108 items in all; domain of linguistic action as the only common
denominator), it will appear that a considerable number of complexities
concerning metonymy can be illustrated from it, and that at least some
diachronic perspective on metonymy can be gleaned from it.
Certain features are shared by some of these domains: (3) and (7) are
'communicative', for example, the others are not; (2), (6) and - at least partially
- (7) are 'ingressive', whereas (3), (4) and (5) are 'egressive'.
The diagram above (Fig.1) is an informal way to represent this. The bold
circle indicates the 'thing' [MOUTH], the relevant domains are indicated as
ellipses, with a full line for the primary domain, a broken line for the secondary
domains and dotted lines for the tertiary domains.
The prototype for mouth is no doubt the human mouth and can be charac
terized, following the paraphrase in (7), as "the opening on the face through
which a person can take food/drink into the body, and speak or make sounds",
but this characterization is, of course, not absolute, since there may be variation
in the amount of detail that is included.
Note, furthermore, that what we are going to be concerned with, is the
mapping of mouth onto an element or structure in its secondary domain (iii)
only, viz. linguistic action. This implies that in our corpus data speakers choose
to access elements or structures in the domain of linguistic action through a
symbolic unit that denotes a 'thing' belonging to a secondary domain from its
viewpoint.
6. Metonymy in perspective
6.1 Preliminaries
The foregoing discussion is not complete, though. For one thing, because the
salience with which the elements/aspects of linguistic action are made accessible
by the item mouth is variable, and should in each instance be evaluated
186 Louis GOOSSENS
individually; for another, because the distinction between the different types may
be blurred.
Let's tackle the blurring among types first. An obvious case is type (ii), in
which the phrase X's mouth is mapped onto both [x's WORDS/WHAT X SAYS OR
SATO] and on [X AS SPEAKER], according to whether we process mouth indi
vidually or carry out the mapping for the complex noun phrase with the genitive
(or of-phrase). Clearly, we can execute the conceptual mapping in one or both
of these ways. Note that in (12), for example, the selection restrictions for
dampne(n) 'damn' invite a personal interpretation of mouth (as [SPEAKER]), but
the third person of the finite verb militates against a simple paraphrase of your
owene mouth as 'you'.
Another instance quoted above which is open to more than one me
tonymic interpretation is (15). can be mapped onto something like 'the
capacity to say what has to be said' (which for me would come under [SPEECH
FACULTY]), as well as onto 'the right words to be spoken' (which would rather
belong to type (i)). Note that in this instance the absence of the article increases
the necessity of a metonymic reading.
This lack of determinacy is not exceptional in my data. Ultimately, it is a
principle like the Gricean maxim of Relevance that will help us to decide on the
appropriate interpretation. In some instances, for that matter, a contextual
implicature will have to be inferred on top of the metonymie interpretation, as in
the next two examples which can be adduced to illustrate this point.
(19) (Chaucer, Boece V, prosa 3, 124)
But not only to trowe [believe] that God is disseyved [mistaken],
but for to speke it with mouthe, it is a felonous [wicked] synne
This instance belongs to type (i), in that mouthe is mapped onto 'words',
[SPEECH] in the typology adopted; however, the contextual interpretation has to
be pushed beyond this: with mouthe implicates [ALOUD], or even [PUBLICLY].
The fact that we end up in contextually, or, if you wish, pragmatically deter
mined meaning is, of course, no difficulty for CG, which rejects the difference
semantics/pragmatics anyway.
Something similar holds for the following instance from Shakespeare.
(20) (As like it, iii 1 11)
Thy lands, and all things that thou dost call thine
Worth seizure, do we seize into our hands,
Till thou canst quit thee by thy brother's mouth
Of what we think against thee
F R O M THREE RESPECTABLE HORSES' M O U T H S 187
Thy brother's mouth was assigned to category (i), mouth can be said to stand
for [WORDS]. At the same time, thy brother's mouth can be interpreted as [THY
BROTHER (AS SPEAKER)], type (ii). But ultimately the most adequate meaning
specification for this combination with quit (= 'acquit') would be [THY
BROTHER'S TESTIMONY], which builds on the interpretation as type (i).
We are still left with what we can refer to as the saliency question. What I
mean by this, is that in some instances in our data the metonymic mapping is
indispensable in order to make sense contextually, whereas in others it would
seem that a literal interpretation might do as well.
None of the instances in my corpus has the salience of such show exam
ples of metonymy as the ham sandwich in Lakoff and Johnson's The ham
sandwich is waiting for his check. Still, instances like (9), (10), (14) and (15)
will not work in a literal interpretation; we might say that the metonymy has
medium saliency there. Cases like (11), (13), (16) and (17), on the other hand,
could also be interpreted literally, though ultimately a more satisfactory
interpretation is obtained if we read them metonymically. Again this is no real
difficulty with a CG view of meaning. Indeed, it only proves that a particular
item which gives access to a meaning network can be made compatible with a
given context in more than one way. Our point about additional implicatures
that can be built into contextual interpretations is in the same line of thinking.
In the same way as 'literal' and 'metonymic' form a continuum, or, more
generally, in the same way as it is difficult to keep literal meaning and metonymy
apart in certain instances, there is also an interpenetration of metonymy and
metaphor.
Following up on my contribution on metaphtonymy and section 3 in this
paper, I will now look at this interpenetration, as it shows up in my data, from
four angles:
• the blurring of the distinction between metonymy and metaphor (6.4.1.)
• instantiations of metaphor from metonymy (6.4.2.)
• metonymy within metaphor (6.4.3.)
• demetonymization in a metaphorical context (6.4.4.).
Note beforehand that these rubrics are not necessarily a hundred percent
discrete.
188 Louis GOOSSENS
This kind of combination, where metonymy and metaphor 'hold their own', is
well represented in our data, especially in the Shakespeare sample, though it is
present from Ælfric onwards. I adduce a couple of examples with a minimum of
discussion.
(28) (ÆLS (Auguries) 34)
[...] wyrgendras Þæra. bið symle
[...] evil-doers whose mouth is always
F R O M THREE RESPECTABLE HORSES' M O U T H S 191
6.4.4 Demetonymization
7. Conventionalization
7.1 Preliminaries
It is obvious that our three samples can do no more than give us a glimpse of
the developments connected with the symbolic unit mouth and its metonymie
mapping onto various aspects of linguistic action. What we find in this material
is the reflexes of the metonymie uses of mouth in three bodies of writing which
F R O M THREE RESPECTABLE HORSES' M O U T H S 193
differ not only in time, but also in register and in dialect. It should be remem
bered, for instance, that there is no direct continuation between the late West-
Saxon standard used by the homilist Ælfric and the early modern English
standard of Shakespeare, whereas Chaucer's language is no more than an
important ingredient in the standard that is emerging at the end of the Middle
English period.
On the other hand, we have found that there is a certain stability in the
metonymic subtypes with which we have worked. A question that can therefore
be asked is to what extent the different metonymic uses are conventionalized,
where we understand conventionalization as degree of conceptual en
trenchment, which is directly connected with ease of activation.
Whether and to what extent a particular usage is conventionalized is diffi
cult to make out for older language stages; sometimes it is not easy for Present-
day English either. A prototypical usage can, of course, safely be taken to be
fully conventionalized. Metonymie uses, which we defined as involving
relatively extrinsic meaning elements (section 2) should be expected to have a
low degree of conventionalization, unless they occur with considerable
frequency.
It is with this in mind that we will confront the different metonymie types
in the three samples with Present-day English. As a criterion for conventionali
zation in today's English we consider whether a particular pattern has filtered
through into general usage dictionaries like LDOCE, COBUILD, OALDCE,
Heritage, and Webster's Third.
72 Observations
In order to detect some of the factors that are operative in the conventionaliza
tion of metonymies, let us review the subtypes we distinguished earlier from that
point of view. The (restricted) generalizations that follow from these observa
tions come in 7.3.
(i) mouth - - > [WHAT IS SAID]
Judging from the constant frequency with which this pattern occurs (Ælfric 11
times; Chaucer 14; Shakespeare 14), it would seem that it is fairly well
conventionalized.
Note, however, that the Ælfric instances are still largely an echo of their
biblical origin (with the Psalms as the chief donor source). Though Ælfric shows
some degree of independent conceptualization of linguistic action in terms of
194 Louis GOOSSENS
the symbolic unit (as I hope to have demonstrated in Goossens (1993a)),
his works are still characterizable as belonging to the religious register for which
the Hebraistic tradition of the Bible remains an obvious ingredient. Instance (28)
quoted above is representative of this, as are (34)-(36:.
(34) (ÆCHom I I , 39 1 294.231)
and cwæð to ðam leasan mid
and said to the wicked-one with
gelærnedum ð
learned mouth
(35) (ÆCHom I, 39 604.21)
hæbbe eower gehwylc halwende lare on
have of-you each holy lore in mouth
and sealmboc on handa
and psalm book in hand
(36) (ÆLS (Basil) 116)
he cwæð to Basilie, beo ðin muð afylled mid
he said to Basil be your mouth filled with
haligre herunge æfter Þinre bena
holy praise after your prayer/service
All of these retain their biblical ring when transposed into Present-day English;
conventionalization must be taken to be restricted to the religious register. Note,
moreover, that in (35), the addition of a determiner would be required ('in your
mouth').
Chaucer is less dependent on religious contexts, witness (37).
(37) (Boece I, Prosa 5, 57-60)
al [though] myghten tho [those] same thynges betere and plentevously
[abundantly] ben couth [be known] in the mouth of the
peple that knoweth al this
Still, the vast majority of instances occur in the Parson's tale, which is
somewhere in between a homily and a treatise of moral theology. A clearly
conventionalized pattern for Chaucer is in mouth, of mouth, with mouth(e) in
the meaning 'oral(ly)', occasionally even 'public(ly)'. It is illustrated in (9)
above and in (38).
(38) (.., Parson, 167)
Contriccioun [remorse] of herte, and shrift [confession] of mouth
As the comparison with Present-day English reveals, such (temporary)
conventionalization is no guarantee that it will be retained in later language
F R O M THREE RESPECTABLE HORSES ' M O U T H S 195
Perhaps the most obvious, but not necessarily unimportant, conclusion is that
the metonymies we have investigated here do not follow rectilinear paths to
conventionalization.
Type (ii) comes closest to straightforward conventionalization (in the
sense that it can be transposed into Present-day English without difficulty), as
do some (but not all) of the verbal expressions with mouth as (part of) the direct
object considered under type (v). For type (i) the conventionalization appears to
be more restricted than its sheer frequency would make us believe at first sight.
Types (iii) and (iv) have the lowest frequency in our database; but, whereas type
(iii), in which mouth is mapped onto [SPEAKER], appears to be on the path of
increasing conventionalization, type (iv), where mouth is used for [SPEECH
FACULTY], would seem to go the opposite way.
As to the factors which turned out to be relevant in the conven
tionalization process (or its opposite), the following are worth pointing out.
• The (relative) continuity of the metonymic patterning hangs together with the
stability of the conceptual network that mouth gives access to. This includes
the stability of its prototype as well as of the conceptual ingredients involved
in linguistic action with which this prototype is connected. In other words,
there do not appear to be noteworthy differences between the basic concep
tualization of our mouths and its functionality in speaking by the speakers of
Old English and by those of Modern English.
• Conventionalization may be restricted to a specific register. As we have
noted, most uses in Ælfric still betray their biblical origin. It is in this bibli
cal/religious register that they often continue to date, whereas they have not
percolated into general usage.
• Among the factors that appear to work against conventionalization, the
following may be singled out.
- A first barrier, though not an absolute one, is the shift in subcategorization
in the instances where mouth came to be used as an uncountable (for ex
ample Chaucer's in mouth, of mouth 'oral(ly)'; in a way this shift points to
increased conventionalization in Chaucer's idiolect (and beyond Chaucer,
given the currency of these phrases in other Middle English texts), but
apparently the conventionalization was not strong enough in most of these
instances for the increased conceptual distance that this status as uncount-
200 Louis GOOSSENS
able entails not to militate against a smooth transition of the pattern to the
following generations.
- Secondly, the embedding of the metonym in an idiosyncratic metaphorical
context (which we often find in Shakespeare) usually appears to be a rea
son for the conventionalization not to go through.
- Thirdly, the fact that a given metonym does not fill a lexical gap (in other
words, the fact that there is another conventionalized way of saying what
the metonym expresses) may be another barrier against smooth conven
tionalization of the metonym.
• Among the factors that promote conventionalization, there is first the
salience of the mapping that is involved. This may account, for example, for
the fact that of the two least frequent patterns, type three {mouth for
[SPEAKER]) and type four {mouth for [SPEECH FACULTY]), it is the mapping
onto [SPEAKER] that appears to be most conventionalized in Present-day
English. Experientially the connection between mouth and speaker is more
obvious ('salient') than that between the mouth and the associated speaker's
speech capacity.
With respect to type (ii), which is the one that is most easily transposable to
later language stages, there is the fact that this metonymy can be accessed
doubly. First, through a mapping of mouth onto [WHAT is SAID/SPEECH],
secondly, by way of mapping the phrase as a whole onto [X AS SPEAKER]:
double access promotes ease of activation, and apparently conventionaliza
tion as well.
Then we found that in some instances phrases tend to be conventionalized as
phrases. This applies to by word of mouth and from mouth to mouth. Why
exactly these are picked out and not others is difficult to explain. Perhaps the
availability of the frames by {way) of and from {time) to (time)/from (case) to
(case) may have helped. In the second example it may also have to do with
the fact that it provides speakers with a compact way to express a relatively
complex conceptualization.
Finally, the reason why some of the verbal expressions in group (v) may be
so accessible and open to conventionalization would seem to be that they are
close to being synecdochic, in the sense that they express a more complex
(linguistic action) event by way of a simpler, partial event (e.g. open one's
mouth). Note that also type (iii), and to some extent type (ii), are synecdo
chic.
FROM THREE RESPECTABLE HORSES' MOUTHS 201
Not only the difference between literal and metonymie uses may be blurred in
context, we also observed that it is sometimes difficult to distinguish among the
different metonymie subtypes. Some instances appear to be compatible with
more than one interpretation, according to the stretch of context considered. A
notable example of this sort was our second type, where mouth can be mapped
onto [WORDS, WHAT IS SAID SPEECH], whereas the nominal phrase as a whole
(X's mouth/mouth of X) stands for [X AS SPEAKER]. Ultimately, relevance with
respect to overall context is decisive as to which interpretation fits best (where
the possibility of a double fit must be one of the options).
202 Louis GOOSSENS
In the same way as there is a continuum between the literal and the metonymic
senses of the item mouth, we find indications of a continuum metonymy-
metaphor, be it only for the verbal expressions where mouth is part of the object
nominal. As we have pointed out, these metonymic expressions may give rise to
metaphorical ones (metaphors from metonymy). In our diachronic data base it
was revealed that some contexts are compatible with both the metonymie and
the metaphorical interpretation, but that more often context will disambiguate
the expression as one or the other. Metaphor from metonymy therefore as a rule
denotes the quite current type of metaphor for which a metonymie origin can be
postulated; the metonymy, then, gives the metaphor its experiential underpin
ning.
Otherwise metaphtonymy manifests itself as the insertion of a metonym
into a metaphorical context. Again this requires considerable conceptual
flexibility from the decoder, and tends to give the combination an idiosyncratic
effect.
8.4 Conventionalization
Metonymies, at least of the current types studied here, form a continuum with
the so-called literal senses with which they are connected on the one hand, and,
at least in the instance of metaphor from metonymy, with metaphor on the other.
This makes it difficult sometimes to notice them, especially because there may
be some blurring among different metonymie extensions, as in the data studied
here. In addition, they often require close scrutiny of the contexts in which they
come in order to be interpreted properly: it is important to give them their me
tonymie reading also if they occur in metaphorical contexts, or to be prepared
for multiple interpretation as the context unfolds. All this make them a
204 Louis GOOSSENS
Brygida Rudzka-Ostyn
University of Leuven
1. Introduction
2. Verbs of Answering
The verbs to be discussed are answer, rejoin, reply, respond, and retort. The
discussion will be based on examples drawn from several sources, including the
Leuven University Theatre Corpus.2 The focus here being on mechanisms of
meaning extension, we will present the verbs' semantics without an analysis of
the lexico-syntactic frames in which they occur. Needless to say, these frames
will play a major role in our attempts to identify the different meanings of each
verb.
208 BRYGIDA RUDZKA-OSTYN
2.1 Answer
The most common of the five verbs is undoubtedly answer. It has already re
ceived a great deal of attention, though primarily in its role as a verb of speak
ing. Of special interest are Wierzbicka's (1987: 37If.) observations on this vari
ant of answer. In her paraphrases intended to capture its meaning, she mentions
the speaker's awareness of being expected to react verbally to another's utter
ance as well as his/her awareness of having to react appropriately. The speaker
knows that what (s)he is going to say must somehow fit that utterance, but not
necessarily in terms of providing missing information. Wierzbicka (op. cit.: 372)
correctly observes that the meaning of answer cannot be restricted to question-
answer contexts, as is often the case in the literature. The verb is quite common
in responses to non-interrogative utterances, as in:
(1) "My God, it's true!" Yossarian shrieked and collapsed against
Nately in terror.
"There is no God," answered Dunbar calmly, coming up with a
slight stagger. (CAT: 129)
(2) "Goodbye, Tony. God bless you." Tony doesn't answer. They go
out. (TC)
However, the fact that analysts tend to link answer with questions may indicate
that this usage is typical for the verb. But then why is it typical? The explanation
is straightforward: since an answer is normally given in a context where a verbal
response is expected, and since questions, at least prototypical ones, epitomize
such expectations, the verb answer becomes most natural, and indeed proto
typical, in designating responses to questions rather than other types of utter
ance (Wierzbicka idem; Rudzka-Ostyn 1989). Moreover, as has often been
pointed out, 3 a true question is incomplete; it contains an empty slot, which is
precisely what a response needs to count for an answer. Thus, although other
speech acts may be followed by answers, questions seem to provide an optimal
context for this type of response. Hence the frequent occurrence of examples
like the following:
(3) "Aren"t you Chaplain Tappman?" demanded the obese colonel.
"He's the one," sergeant Whitcomb answered. (CAT: 387)
(4) I'm willing to answer any questions you like to put. (TC)
While genuine questions are inherently incomplete, other types of utterance of
ten acquire this property via their interaction with various pragmatic factors. To
return to the situations coded by (1) and (2) above, the utterance which pre-
METAPHOR, SCHEMA, INVARIANCE 209
cedes the response is not a question, but because it (a) allows "the assumption
of a gap (...) to be filled" (Wierzbicka op. cit.: 372) and (b) the response given
fits that gap, one can use answer to report the event. And as the gap created is
not prompted by a desire for information, the response does not have to be in
formative. Informativeness is required, however, in the case of responses typi
cally associated with answer.4 Wierzbicka (op. cit.: 373) identifies them as
"knowledge-providing devices" and rightly points out that the knowledge sup
plied is supposed to be sufficient and satisfactory. To these concepts we could
add that of control: under normal circumstances the speaker is in control of
what he says and may therefore be held responsible for his utterance, be it a true
answer or not.
To sum up, a speech act reported by answer is typically subsequent vis-à-
vis another speech act and is expected to be informative, appropriate and suffi
cient. In less typical codings, some of these notions may be suspended or elabo
rated further, and thus restricted to specific situations. In (1) and (2) above, the
requirement of informativeness is dropped, whereas in (5) and (6) answering is
confined to a very restricted context of accusations or criticism. It amounts to
acts of verbal self-defence. We will return to this particular sense of answer at a
later stage because this was the original meaning of the verb.
(5) ... was held in bail to answer a charge of petty larceny. (Webster's)
(6) How would you answer the criticism that your government has in
creased the level of taxation? (LDOCE)
Given the fact that the concepts invoked in the description of answer-related
speech events also characterize non-linguistic events, it should be possible to
use answer in situations that have nothing to do with verbal exchange. This is
indeed what happens, but the extensions are not unconstrained. Answer cannot
be used with reference to just any event that follows another; the use must con
form to some schema, and to discover this schema it is necessary to go into the
place of answers and speech acts in general within a broader categorial context.
As a category comparable with other categories, speech events constitute
a subclass of what we can call "acts" or "events". The latter subdivide along
various, often intersecting, parameters. One of them pertains to the acciden
tal/non-accidental opposition. Non-accidental acts include acts performed in
reaction to or caused by other acts, and it is, among others, to this broad sub
category that answers belong. As a verbal response to another's utterance, an
answer is a subcategory of acts triggered by other acts. It is an elaboration, a
special case of something done in reaction to something else. How such acts
210 BRYGIDA RUDZKA-OSTYN
link up with speech acts in general and answers in particular can be represented
as in Figure 1, where full-line arrows indicate a relationship of instantia
tion/elaboration and its opposite, schematicity, and interrupted-line arrows in
dicate extension. Sequences of three dots indicate elements that could be added
to complete the legend or the network of (sub)categories.
METAPHOR, SCHEMA, INVARIANCE 211
Simplistic as it is, Figure 1 depicts several paths by which answer could gain
access to non-linguistic acts. To see how exactly answer arrived at these ex
tended senses, it may be instructive to go back to its prototypical version. As
already mentioned, this version is appropriate when a speaker reacts verbally to
another's question in a satisfactory manner, i.e. provides the information sought.
Since questions are a subcategory of utterances, the meaning of answer can
easily be generalized to include verbal responses to any triggering speech act
(cf. (1) and (2) above). And since speech acts are a subtype of acts, the process
of generalization can be carried beyond the domain of linguistic action. At this
point, answer acquires a very general, metaphoric meaning 'to react to an
other's action'.
(7) Jesus told us how to answer a blow on the cheek but he never told
us how to answer a kiss. (TC)
(8) He answered his country's call when he wasn't much older than
you are ... (TC)
As in the case of questions and answers, one event comes as a reaction to an
other, and it is this prior-subsequent schema5 that licenses the extension. An
other motivating factor is the nature of the prior event; like a question it is as
sumed to require, or at least be able to trigger, a response. It is not surprising
then that, in addition to this general meaning of answer as illustrated above,
speakers of English have come up with more specific variants focusing on
counteracts and responses to acts meant as signals. Examples (9)-(14) may
serve as illustrations:
(9) ... a major incursion which would be answered by massive retalia
tion. (COBUILD)
(10) ... answered the enemy's fire shell for shell (Webster's)
(11) Sam answered her look with a grin. (COBUILD)
(12) A spotlight comes up between the towers. Two answering owl
hoots are heard. (TC)
(13) ... the second voice answers in the dominant... (Webster's)
(14) ... underground treasures ... seldom answer clearly when they are
queried by the geologist's instruments. (Webster's)
In the situations coded by (9) and (10) the reaction is meant as an act of retalia
tion. What comes to the fore is the concept of opposition, a concept that is not
unexpected if we bear in mind the etymology of answer and its present-day use
in the context of accusations and criticism. This concept is, however, absent
from the variants illustrated in (11)-(14), where making a responsive sound or a
212 BRYGIDA RUDZKA-OSTYN
signal matters and the responding entity does not have to be human or even
animate. English has further extended this use of answer to such events as
picking up the telephone when it rings or going to the door when one hears a
knock or the bell (see examples (15)-(16)). Again the extension is in keeping
with the prior-subsequent schema and the expectancy pattern associated with
prototypical answer. We have here a combination of metaphor and metonymy --
a metaphtonymy, as Goossens (1990) aptly identifies such combinations -- be
cause the front door and the telephone actually refer to sounds signalling a de
sire to be let in or to communicate.
(15) The telephone rings. Ben answers it. (TC)
(16) Andrew answers the front door. (TC)
The prior-subsequent schema itself incorporates information captured by a
higher-order correspondence schema. At a certain level of abstraction, trigger
ing acts and responses to these acts can be seen as having some common
ground by virtue of either belonging to the same category of socio-perceptual
signals or complementing one another in terms of informativeness or reflecting
the same kind of intention, good or bad, on the part of the agent. Unlike the
prior-subsequent schema, which dominates sequences of events, the correspon
dence schema generalizes over dynamic and static relations. This being the case,
it is easy to understand why answer has come to mean simply 'correspond,
match, be similar (in number, size, shape, position, character, etc.)', as in:
(17) A man answering his description has been seen in the Bedford
area. (COBUILD)
(18) parts that answer each other on a blueprint (Webster's)
Whereas in the examples discussed so far the notion of semi-similarity between
prior and subsequent events determines the path for metaphoric extension, in
other cases the concepts of fulfilment and appropriateness become a critical
factor. As already pointed out, a prototypical answer must satisfy a desire for
information and it must be appropriate. It must fit the gap created by the given
question. Since the notions of appropriateness, fulfilment or conformity are also
present in the semantics of other domains, a new possibility of metaphoric ex
tension arises. This possibility has been exploited in English, and so answer can
be used in contexts in which verbs such as fit, suit, satisfy, fulfil, serve, or meet
normally occur. Here are a few examples, one of which (cf.(20)) lends itself to a
'metaphtonymic' interpretation:
(19) These discussion groups are obviously answering a need.
(COBUILD)
METAPHOR, SCHEMA, INVARIANCE 213
Figure 2
a) = exs. 5, 6; b) = exs. , 4; ) = exs. 1, 2; d) = exs. 7, 8; e), f) = exs.
17, 18; g) = exs. 11, 12; h) = exs. 9, 10; i) = exs. 13, 14; j) = exs. 15, 16;
k) = exs. 24, 25; l) = exs. 26, 27; m) = exs. 19, 20; n) = exs. 21, 22, 23.
Figure 2 brings together all the variants of answer discussed in this section and
shows the main interconnections between them. The descriptions enclosed in
the boxes are not meant as exhaustive definitions. Rather, they function as
shorthand intended to label the most important concepts associated with a par-
METAPHOR, SCHEMA, INVARIANCE 215
our day. The similarity between their organizational pattern and what we have
identified above as a correspondence schema gave rise to numerous extensions
of andswarian into the domain of imitative acts as well as that of different cor
respondence relations. In turn, the similarity between a speaker's readiness to
meet the expectations of his interlocutor and acts of recompense, assumption of
responsibility or atonement led to extensions into the semantic territory nor
mally controlled by verbs such as pay, repay, be responsible, and atone. By
around 1200, andswarian was already common as a non-communicative verb
meaning 'correspond, be similar', 'assume responsibility for' and 'atone for'.
Then came variants whose meanings approached those of recompense, return,
and retaliate. And finally a number of extensions developed around the concepts
of satisfaction and fulfilment, allowing answer to become a synonym of satisfy,
fulfil and accomplish. Several of the metaphoric and non-metaphoric variants of
andswarian/ nswer acquired more specific senses like 'reply favourably to a
petition', 'sing antiphonally', 'react to a sign/signal', etc.
Figure 3 presents the main stages in the evolution of andswarian/answer
together with a few illustrations of the more interesting developments. All the
'historical' examples cited below as well as in the following sections come from
the Oxford English Dictionary or its Supplement. The dates provided inside the
network indicate the first and last occurrence of the given variant as recorded in
this dictionary. If only one date is given, the variant is still in use. In any event,
the dates are mere approximations as the different variants of answer must have
been well entrenched in the language before they were recorded. Note, moreo
ver, that the meanings isolated here do not always conform to the groupings
distinguished in the Oxford Dictionary due to the considerable semantic overlap
among these groupings.
As is evident, not all variants of andswarian have remained productive. In
some cases, moreover, their range of applications has changed. At the beginning
of the 16th century, answer acquired the meaning 'react verbally to another's
utterance in an impertinent way'. This meaning is still alive, but it now requires
the presence of back and an intransitive frame. Restrictions of a different kind
affected sense (e), which now occurs primarily in the context coded by Your
prayers have been answered. To draw attention to one more example, the vari
ant of answer which lexicalized the correspondence schema could in the 18th
century be used in causative constructions (see (o)). This is no longer possible.
METAPHOR, S C H E M A , INVARIANCE 217
Figure 3
d) How they will answer it ... at the last day I know not (1680); e) With
reluctancy to answer my Request (1689); f) Hush, Frank, never answer
your father (1853); g) The Glebe will answer to the Sylvan Reign, Great
Heats will follow, and large Crops of Grain (1697); The girl instantly an
swered to the action in her sculling (1865); i) Able to answere feast with
feast (1601);Answering love for love (1827); j) Knock but at the gate,
and he himself will answer (1597); k) The holwe rokkis answerden hire a-
gayn (1385); m) Or hit ountrid h to aunsware Ector agayne (1400); n)
218 BRYGIDA RUDZKA-OSTYN
Our short diachronic survey also demonstrates that the metaphoric senses of
answer are extensions from the domain of communication. This domain is their
source domain, at least as far as the Old English data go. The factors that de
termined the growth of the andswarian category are the same as the ones in
voked to account for the uses of Modern English answer. The reason why this
should be so is clear: a word that has taken root in a language cannot acquire
just any new sense. There must be a link between what the word already means
and the sense it is to adopt. And since the present-day network of senses is the
result of earlier processes, whatever has motivated them is necessarily reflected
in this network.
2.2 Reply
Another very common verb of answering is reply. Like answer, it reports speech
acts triggered by other speech acts. But while typical answers fill some informa
tion gap, typical replies seem to focus more on "covering the same ground,
giving an appropriate return" (Webster's). This return may, but does not have to
be, triggered by a question, and, as Wierzbicka (1987: 373-75) rightly notes, its
relation to the triggering act can be quite loose. My data base provides numer
ous examples of reply (both as a verb and a noun) referring to the second
speaker's views on or reactions to earlier utterances. Here is just a small sample:
(1) "... Your job is to bomb the ammunition dumps at Bologna."
METAPHOR, SCHEMA, INVARIANCE 219
Reply is certainly more formal than answer, which may have something to do
with the Latin origin of reply.
The origin is worth examining in connection with the other meanings that
reply developed over the centuries. Dictionaries link reply with Old French
replier and Latin replicare, both of which meant 'fold back, fold again, turn
back' {Webster's, OED). This meaning was still recorded between 1450 and
1574, but already in a 1385 work by Chaucer reply occurred as a verb of com
munication. In addition to the general meanings it exhibits now, it also had more
specific senses such as 'say that one did not mean what one said, retract' and
'say in response to the defendant's plea', as in these illustrations:
(5) Whiche thing is wonder, that they knowing me saiying but soth
arne nowe tempted to reply her olde praysinges. (1388)
(6) To that that he hath aunsuerd y have replyed yn such wyse ...
(1453)
(7) The plaintiff may plead again, and reply to the defendant's plea.
(1768)
From the vantage point of their Latin and Old French ancestors, the communi
cative versions of reply functioning in Middle English appear to be metaphoric
extensions -- extensions facilitated by the structural similarity exibited by acts of
folding/turning an object back and utterances made in response to other ut
terances. In both cases we witness a repetition of an earlier act and a reversal of
directionality. An object can be folded back only once it has been unfolded;
likewise, a speaker can reply only after he has been spoken to. Moreover, like
the object that returns, by being folded back, to its earlier position or form, a
verbal reply goes back to the speaker who initiated the exchange. We will have
more to say about the similarity between responsive speech acts and this type of
physical manipulation at a later stage. What needs stressing now is that since the
communicative and non-communicative variants were operating in English at
the same time, they can all be regarded as instantiations of one and the same
schema. Once the schema became available as a sufficiently salient frame, it
could be applied to any new domain organized around a similar schema, which
explains the use of replien in reference to acts of repudiation, as in:
(8) The quene Gwendolyne ... Whome Kyng Locryne forsoke and re
ply ed, And Estrylde weddid againe. (1470)
In (8), the extended sense 'send away, reject' was no doubt motivated by the
prior-subsequent schema and the concept of reversing the effect of the prior act.
METAPHOR, SCHEMA, INVARIANCE 221
Figure 4
a) = exs. 1, 2, 3, 4; b) = exs. 6, 7; c) = ex. 5; d) = exs. 9, 10, 11; e) = 12,
13; f) = exs. 14, 15; g) = ex. 8.
Within the domain of speech acts, new concepts came to be associated with this
schema, concepts that the schema did not have in the domain of physical move
ment. One of them was the notion of producing an utterance in reaction to the
triggering act. Via generalization this notion gave rise to an extended meaning
of reply: 'do something in direct/immediate reaction to'. In the situation coded
by (8), the subsequent act reverses the effect of some prior act which could have
been performed many years earlier. The subsequent act cannot be said to consti-
222 BRYGIDA RUDZKA-OSTYN
tute a direct and immediate reaction to whatever preceded it. The concept of
such a reaction is associated with the variant exemplified in (9)-(11):
(9) .... see ... thine eyes replying To the hues of yon fair heaven.
(1818)
(10) He sang his song, and I replied with mine. (1848)
(11) He replied with a nod. (OALDCE)
Related to the above variant is a more specific sense which can be paraphrased
as 'return an attack'. This retaliatory sense may have been reinforced by the
legal use of communicative reply as well as the concept of counterforce incor
porated in the original meaning of the verb. Folding or turning back an object is
not necessarily done in reaction to some prior act but it may involve overcoming
the object's resistance. Like the general sense, the more specific one (see (12)-
(13) below) is still alive.
( 12) The besieged replied sharply. (1829)
(13) ... poured broadside after broadside into the forts, which replied
continuously. (Webster's)
The general sense can be linked to yet another very specific meaning, para-
phrasable roughly as 'return a sound'. As the following examples indicate, this
meaning has also stood the test of time.
(14) What man that in the wodes crieth, Withoute faile Eccho replieth.
(1390)
(15) We were calling for help, but only the rocks replied.
The most important senses of reply are presented in Figure 4, which combines
synchronic and diachronic information. From the vantage point of Modern Eng
lish, sense (a) appears to be prototypical for the reply category. Note again that
if only one date is provided, the variant is still in use. Note also that a two-
headed arrow (like the one linking the meaning of Latin replicare with Middle
English replien) symbolizes a relationship of identity.
While in the case of answer all the extended uses could be related to the
domain of speech acts as their source domain, the metaphoric extensions of
reply pose a few problems. At first glance directed physical motion appears to
be the source domain for the extensions, though not for all. Variant (d), for in
stance, dominates much more directly the communicative and acoustic variants
of reply than the variant denoting physical motion. It should therefore be treated
as an extension from these variants. The fact that they were in use much earlier
would support such a conclusion (but see section 3). On the other hand, variant
(d) is not unrelated to the original meaning of replye. What is then its source
METAPHOR, SCHEMA, INVARIANCE 223
domain? If its direct linkage with communicative and acoustic reply is taken to
be the decisive factor, we are still left with two source domains: communication
and acoustics. Which one should we choose? Or should we argue that all three
domains are in fact source domains for the most general variant of reply? We
will return to these questions in the final section.
2.3 Respond
(2) ... I've heard you say you were a gnu and Ann respond with the
unlikely information that she's a hippopotamus ... (TC)
(3) "Peckem?" repeated General Dreedle, still squinting with bewil
derment. "Just what the hell does Peckem have to do with it?" ...
"Absolutely nothing, sir!" Colonel Cathcart responded sprucely ...
(CAT: 224)
(4) ... responded negatively to the question. (Webster's)
(5) The priest says "The Lord be with you," and the people respond,
"And also with you." (LDOCE)
Whereas answers and replies are prototypically verbal reactions to utterances,
responses are often reactions to the attending circumstances, including the first
speaker's behaviour or intentions. In other words, a verbal response may be as
much a reaction to verbal as to non-verbal stimuli. Moreover, the response to a
verbal stimulus may be non-verbal. Consider in this respect example (6):
(6) He responded to my suggestion with a laugh/by laughing.
(LDOCE)
One may, of course, argue that the laughter is a symbolic representation of a
scornful utterance, but the fact that answer and reply would not fit as well in the
context of (6) shows that respond has a yet broader range of applications. The
differences become even more conspicuous in the frame suggested by L.
Goossens and A.-M. Vandenbergen (p.c.): "Do you want it?" He didn't an֊
swer/reply/*respond; he just nodded. What makes respond odd here is precisely
its stronger association with non-verbal acts. Its linkage with verbal stimuli may
be suspended altogether, in which case the meaning of respond comes very
close to that of react and can be paraphrased as 'do in reaction to', 'react to a
stimulus'. Such extended meanings are given below:
(7) ... responding to the threat of death with behavior that is a degra
dation of the human spirit. (Webster's)
(8) An electric circuit will respond most readily to impulses which
come timed to its own natural rate of vibration. (OED-S)
(9) The pupil of the eye responds to change of light intensity.
(Webster's)
As in the case of verbal responses, we have here a succession of events, one
being triggered by another. Like a speaker that reacts to another's utterance, an
inanimate entity reacts in accordance with the stimulus. This reaction is concep
tualized as originating with this entity, its directionality is thus opposite to the
stimulating force or influence. The motivating force behind such extensions is
METAPHOR, SCHEMA, INVARIANCE 225
again the similarity perceived in the structure of verbal responses and other
types of responsive acts.
Frequently respond means not only 'react' but 'react favorably', as the
examples below testify.
(10) This case proves the importance of adults responding to children's
needs as and when they arise. {COBUILD)
(11) The disease failed to respond to drugs. (LDOCE)
We could consider such uses as a case of extension via specialization and say
that speakers of English decided at one point to associate the concept of benefit
with the verb respond, and with time this association became conventionalized.
The question remains, however, what prompted them to do so in the case of re
spond and not, for instance, reply, which takes us back to Latin respondere.
This ancestor of respond already carried positive overtones and meant, among
other things, 'to pledge, promise (in return), warrant'. The concept of a positive
reaction has been carried into English and, as etymological evidence indicates,
has always been linked with respond. For instance, in the 17th century it
emerged in the variant of respond which came very close to the verb recipro
cate. Instructive here is the following example:
(12) The King should not be denied the means, by which he may re
spond the great confidence placed on him. (1642)
Against this variant, the uses illustrated in (10)-(11) can be seen as cases of gen
eralization rather than specialization since the specific constraints operating on
acts of reciprocation have been suspended. What this example makes clear is
that the notions of specialization and generalization are relative, not absolute,
and their application cannot be divorced from one's vantage point and standard
of comparison. With respect to the very general meaning 'react to some stimu
lus' as expressed by certain variants of respond, the meaning 'react favorably'
(cf. exs. 10-11) is a specialization; but with respect to the 'reciprocate' variant
of respond (ex.12), it is a case of generalization. There is, however, a usage (see
ex. 13) which appears to be a specialization vis-à-vis the 'reciprocate' variant:
(13) If one diamond were opened and partner responded two clubs,
there would not be a sound rebid. (OED-S)
Here the favorable reaction is confined to a very restrictive context of bridge
playing, specifically to acts of making a bid in reply to a partner's bid. These
acts may combine verbal and non-verbal elements.
226 BRYGIDA RUDZKA-OSTYN
Figure 5
a) = exs. 1, 2; b) = exs. 3, 4; c) = ex. 5; d) = ex. 13; e) = ex. 6; f) = ex.
7; g) = ex.14; h) = exs. 7, 8, 9; i) = ex. 12; j) = exs. 10, 11.
METAPHOR, SCHEMA, INVARIANCE 227
This variant of respond gradually fell out of use, but as long as it was employed
it was fully motivated by the structural similarities holding between sequences of
triggering acts plus responses and entities brought into a relationship of corre
spondence. An additional motivating factor was no doubt the morpho-phonetic
affinity between respond and correspond.
With respect to the original meaning of respond, the meanings illustrated
in examples (7)-(12) and (14)-(15) are metaphoric extensions. The source do
main is that of verbal communication and the recipient domains include a host of
acts performed in reaction to other acts or stimuli. But from the present-day
perspective, that is, from the perspective of the very general meaning 'do
something in reaction to', the communicative uses of respond may be seen as
cases of specialization. If indeed this general meaning has become for speakers
of English the most representative meaning of the whole respond category, what
is the exact status of the more restrictive senses vis-à-vis the general sense? Can
some of them be treated as metaphoric extensions, or are they all de facto cases
of synecdoche? This is another problem which we will take up in the final sec
tion.
Figure 5 is an attempt to bring together the different variants of respond
discussed and to make explicit the links between them. As elsewhere, the de
scriptions in the boxes are intended to signal the most important concepts ex
pressed by the given variants. A question mark accompanying a box means that
it is unclear when the particular sense was recorded. Although some speakers of
English see sense (f) as prototypical for the whole respond category, others are
inclined to accord similar status to sense (a). In view of this discrepancy, we
have not used any prototypicality markings in Figure 5.
2.4 Rejoin
Unlike the verbs discussed so far, rejoin refers to speech acts which express
disagreement and/or annoyance on the part of the second speaker. The utterance
of the first speaker peeves him, which explains why his response is often quick
and sharp. The triggering utterance may, but need not, be a question; what is
important is that the response be related to it and that, as in the case of answers
and replies, some symmetry obtain between the two speech acts. A rejoinder,
however, not only takes up a point made by the first speaker but it also counters
228 BRYGIDA RUDZKA-OSTYN
that point, thus introducing an element of opposition. The example below illus
trates a typical context for the use of rejoin:
(1) "He can't sleep comfortably on that ship", she said. "In his present
state of mind", rejoined Andrew, "he might not sleep comfortably
anywhere". (Webster's)
The notion of opposition has been carried over from the original Middle English
use of rejoin. The verb was first used, and in fact still is, to denote a defendant's
answer to a plaintiff's replication (Webster's, OED). Webster also defines Mid
dle English rejoinen as 'join one's own plea to that of the plaintiff'. Given the
fact that the verb's French ancestor meant 'join two things together', the com
municative use of rejoin could have resulted from a metonymic extension: what
at one stage designated a legal document joined to another legal document, both
bearing on the same case, came to be linked with the content of one of those
documents and, by further extension, with a reply to a legal charge or to just
any utterance likely to arouse opposition. Indeed, in example (1) above any ref
erence to a legal context is absent; the only concept that has been preserved is
that of symmetry and comparatively mild opposition.
A similar symmetry underlies the non-communicative uses of rejoin. As
examples (2)-(5) below illustrate, in each case a subsequent event is a repetition
of a prior event, but with the opposite effect. Specifically, two or more entities
come or are brought together after they had (been) separated. And it does not
matter whether the entities are parts of a concrete object, a spatial configu
ration, or members of a group.
(2) Rejoin the two wires. (LDOCE)
(3) ... the road rejoins the highway two miles east (Webster's)
(4) He was determined to rejoin the RAF. (COBUILD)
(5) You are entitled to rejoin your escort if you wish. (TC)
To a certain extent, Modern English keeps the two versions of rejoin apart by
assigning to them different stress patterns and phonetic values. According to
dictionaries, communicative rejoin is pronounced as [ri'dჳƆin], and the non-
communicative variant as[ 1 ri'dჳƆin]. 10However, this difference is not observed
in all contexts. Non-communicative rejoin often assumes the phonetic value of
its communicative counterpart (cf. examples (3), (4), and (5)), which does not
mean that speakers of English are necessarily aware of the semantic links be
tween them.
METAPHOR, SCHEMA, INVARIANCE 229
In view of all this ambivalence, what is then the categorial status of the
uses illustrated in (2)-(5) above? Are they to be regarded as metaphoric exten
sions of communicative rejoin? Although the earliest (1456, 1530) records lo
cate rejoin in a communicative, specifically legal context, the uses encountered
since the verb became well entrenched in English indicate that the com
municative and non-communicative meanings developed simultaneously. As
already mentioned, communicative rejoin spread into the domain of more gen
eral replies, and non-communicative rejoin extended the concepts of bringing
and coming together to people and abstractions. The following examples illus
trate this evolution:
communicative rejoin
(6) This is the reioynyng of Nicholl Marshall vnto the replicacion of
Robert Bale. (1456)
(7) To this aunswere the Duke of Orliaunce replyed, and King Henry
reioyned. (1568)
(8) Knox rejoyndeth, it is not enough. (1637)
(9) Several passages both of the Preface and Body of the Discourse I
am rejoyning to. (1665)
non-communicative rejoin
(10) Whan they be drye sewe them subtylly and the lyppes wyl reioyne
togydre. (1541)
(11) As tin-soder doth knit and rejoyne a crackt peece of brasse. (1603)
(12) Her great spirite, rejoyned to the spirite Of this great masse, is in
the same enwombed. (1591)
(13) Receive the one, and soon the other Will follow to rejoin his
brother. (1769)
Taking into account the simultaneity noted and the very plausible metonymic
origin of communicative rejoin, the non-communicative uses cannot be viewed
as its extensions. If my interpretation of the historical data is correct, non-com
municative rejoin simply continues to express the meanings it has always had,
and it is communicative rejoin that is de facto an extension. It was possible to
use a verb which denotes acts of physical reunion in the context of this particu
lar type of speech act because of the underlying similarity. Whether one rebuts a
legal charge or rejoins broken pieces, one tries to undo the effects of the first
event, i.e. of the accusation and the breaking respectively. This concept of undo
ing, of reversing, some prior effect has allowed rejoin to extend into the domain
of communication. Interestingly, rejoin has not encroached upon the territory of
230 BRYGIDA RUDZKA-OSTYN
Figure 6
a) = ex. 6; b) = ex. 1; c) = exs. 8, 9; d) = ex. 7; e) = exs. 2, 11; f) = exs.
3, 4, 5, 10, 12, 13.
METAPHOR, SCHEMA, INVARIANCE 231
2.5 Retort
The last verb to be discussed is retort. While Webster and Longman include it
among verbs of answering, Wierzbicka (1987: 136f.) groups it together with
disagree, refute, dispute, deny, contradict and the like under the rubric of
'argue verbs'. This fact alone points to the fluctuation of field boundaries. The
network approach accommodates such fluctuation in a straightforward manner
and, what is more, it provides a rationale for the divergent groupings. Once
meaning representations are viewed as conceptual clusters, and these clusters
are found to include one or more identical concepts, it is easy to show what
licenses the proposed division.
In the case of retort, there are several concepts that link it with both verbs
of answering and arguing. In addition to being subsequent vis-à-vis another act,
a retort is triggered off by that act. It comes as a reaction to blame or criticism
which, as Wierzbicka (1987: 136) rightly observes, may be implied rather than
explicit. There is again a certain symmetry, but not necessarily at the level of
expected information. Rather a retort is meant to counter an assumed attack,
and this countering may involve repeating the words used by the attacking party
(see exs. (1) and (2) below). It is often an act of retaliation, quick11 and rather
aggressive, one "that turns the first speaker's statement or argument against
him" (Webster's). A typical retort exhibits considerable skill and wit on the part
of the speaker, though it is primarily the concepts of countering, implied criti
cism, and disagreement that connect the verb retort with the field of arguing.
Another common element is the second speaker's annoyance at the hostile im
plications. The double affinity — with verbs of answering and verbs of arguing ~
is reflected in the syntactic frames in which retort can occur. It can take the
same prepositional adjuncts and nominal complements as certain verbs of ar
guing. In addition, it can combine with sentential complements and take the
noun question as its direct object, which brings it close to verbs of answering.
The following may serve as illustrations:
(1) "You've got peanut brittle crumbs on your face," Appleby re
marked to him.
"I'd rather have peanut brittle crumbs on my face than flies in my
eyes," Havermeyer retorted. (CAT: 48)
(2) "Are you ready?" Why should I be ready when you're not?" she
retorted... (LDOCE)
(3) ... will retort the question ... by another question ... (Webster 's)
(4) "We'd probably shoot you," ex-P.F.C. Wintergreen replied.
232 BRYGIDA RUDZKA-OSTYN
(8) We shall retort these kind favours with all alacrity of spirit. (1598-
9)
Once retort becomes entrenched in the context of returning weapons, blows, or
injuries, an analogy with negatively charged utterances is easily established.
Verbal retaliation becomes a special case of returning ill-treatment in general.
Hence by the end of the 16th century retort had already occurred as a verb of
communication:
(9) I shall straight retort all the blame ... upon yourself. (1596)
(10) I scorne to retort the obtuse jeast of a foole. (1602)
(11) Hipponax retorted their pleasantry with such keen strokes of satire
that they hanged themselves. (1734)
As retaliation involves acting against somebody, it should be possible to use
retort where intellectual opposition rather than sheer verbal abuse is highlighted.
This possibility was indeed pursued, and so retort came to designate acts of
responding to propositions by means of similar propositions to the contrary.
Here are two illustrations:
(12) Our Experiments may ... enable us to retort their Arguments
against themselves. (1660)
(13) So apt is the Comparison in most Respects ... that my Adversary
in vain Labours to retort it. (1713)
From acts of casting back weapons, returning blows, wrongs and incivilities, we
have come to acts of countering arguments. The objects acted upon have be
come increasingly abstract, but the organizational pattern has remained the
same. In each case, the second act is an imitation, or, better, a mirror image, of
the first one, which leads to a reversal of the attacker-attacked relationship.
When used as a verb of speaking, retort does not necessarily mean that
the first speaker's utterance is an explict attack on the one who responds. As
already mentioned, bad intentions may be only implied, which brings us back to
the central variants of retort as it functions in Modern English. From the dia-
chronic perspective sketched above, these variants may be viewed as meta-
phoric extensions of a schema originally grounded in physical manipulation. And
when the present-day perspective is adopted? What status should we accord to
communicative retort and its non-communicative counterparts? Note that even
now retort is still used in reference to acts which have nothing to do with verbal
communication. It can mean 'return', 'reflect' or 'retaliate', as in:
(14) retort heat (Webster's)
(15) retort a wrong (Webster's)
234 BRYGIDA RUDZKA-OSTYN
(16) There exists in the animals the impulse to retort upon offenders.
(Webster's)
We have already identified a few factors motivating the use of retort in the do
main of retaliatory acts. Very similar factors account for its reference to reflec
tion of heat or light. In order to be reflected, heat and light must first reach the
given surface, at least according to naïve physics. When retorted, they go back
to their initial locus, in which they resemble questions and incivilities 'returned'
to the first speaker. But, to go back to our question, should these uses be
treated as metaphoric extensions of communicative retort? We usually think of
metaphor when a structure from a more concrete domain is mapped onto a
more abstract one. The domain of speech acts is experientially more salient than
that of retaliatory acts in general, and probably more concrete than the domain
of light and heat transmission. From a synchronic perspective, the domain of
communication may then be treated as the source domain for the variants illus
trated in (14)-(16). Yet etymological data like (17) and (18) below point to a
more direct connection with retorquere. Such non-communicative uses were
recorded at the same time as, or even before, certain communicative variants of
retort developed, so diachronically, they are all extensions from the domain of
physical motion. Once again, the concept of source domain turns out to be inti
mately related to the perspective adopted.
(17) As when his vertues shining vpon others, heate them, and they re
tort that heate againe To the first giuer. (1606)
(18) ... a precious stone which, as a burning glasse, receiueth, and re
tortes the Sunne-beames (1611)
An alternative interpretation of the data is possible, however. When we look at
16th-18th century English, we notice that retort had a much broader range of
meanings than today. And since several of them arose almost simultaneously, it
could well be that the metaphoric uses resulted not so much from transfers from
more concrete to less concrete domains but from the application of the schema
underlying retorquere and its early English translations to an increasingly wider
range of domains. What I am suggesting here is that once a schema has been
abstracted and recognized as an organizational pattern for a given linguistic
category, be it a lexical item, a morpheme or a syntactic construction, it can
serve as a link with categories exhibiting similar patterns, and thus mediate ex
tensions across category boundaries. Already in the 17th century retort en
croached upon the territories of retaliate, reflect, return, respond, reciprocate,
METAPHOR, SCHEMA, INVARIANCE 235
and the like simply because all these verbs exhibited similar schematic structure.
Figure 7 below is an attempt to represent these similarities.
Figure 7
a) = exs. 1, 2, 3, 4, 5; b) = exs. 12, 13; ) =exs. 9, 10, 11; d) = ex. 6, 15,
16; e) = ex. 7; f) = exs. 14, 17, 18; g) = ex. 8.
236 BRYGIDA RUDZKA-OSTYN
Although several of the old uses of retort have been lost, to account for the pre
sent-day situation we have to postulate a network that is almost identical to the
one reflecting the historical development. Why? Because it charts conceptually
plausible links. In fact, on the basis of this network one could guess what paths
the item had followed in the past. Again, it does not mean that the evolution
was totally determined from the start since the actual exploitation of the given
possibilities depends on a number of factors, including pressure from neighbour
ing categories. In the case of retort, the choice went in the direction of negative
connotations, so the sense of 'reciprocate' gradually fell out of use.
3. Discussion
Having looked into the semantics of individual verbs of answering, we are now
in a position to identify the concepts that hold the whole field together. One of
them is undoubtedly the prior-subsequent schema. On several occasions we
have seen that all the central variants of the verbs in question and most of their
extended uses denote acts that follow other acts. The prior acts are the cause ~
direct or indirect — of the subsequent acts, and here we come to the next organ
izing schema which indicates that the subsequent acts are performed in reaction
to what precedes them. As a result, the effect of the prior act is, depending on
the type of activity, reversed or undone, and the primary agent becomes the goal
of the subsequent act. It is at him that the activity is now directed. This notion
of effect/direction reversal is another element that binds the verbs to one an
other. And since not only linguistic responses but also acts of retaliation in gen
eral and certain acoustic and thermodynamic phenomena are organized around
the same element, it is no wonder that these verbs expanded their semantic terri
tory the way they did.
In the case of retort, rejoin and, to a certain extent, reply, another concept
emerged as a unifying element, namely that of verbal opposition. The opposition
schema is an elaboration of the effect-reversal schema, and so an additional link
with the other verbs of answering is established.
While the semantics of retort and rejoin revolves around verbal opposition
and physical movement, answer, respond and reply allow more abstract senses.
The highest degree of abstraction has been reached by answer (and, in the past,
by respond՛, see Figure 5), which brings us to the correspondence schema. This
schema evokes situations in which two entities — things or relations — are
brought together, usually for the sake of comparison. The standard of compari-
METAPHOR, SCHEMA, INVARIANCE 237
son may, but does not have to, involve intrinsic properties. Totally different en
tities may be brought into a correspondence relation simply because they are
functionally equivalent or because one is the cause of the other. Since such
causal links exist between responses, verbal or non-verbal, and the
acts/situations that trigger them, we can regard the sequences as elaborations of
the correspondence schema. Another type of elaboration is effected by a schema
which says that the given entities are similar because e.g. they are of the same
kind. This schema appears to operate more on intrinsic elements, although the
distinction between correspondence and similarity may well involve other fac
tors. The two concepts certainly deserve further investigation, also because
categorization in general involves similarity judgments.
Of more immediate concern here is the role of the similarity schema in the
semantics of verbs of answering. As Figures 2, 4, 5, 6, and 7 demonstrate, it
provides a bridge between these verbs, specifically the variants which denote
acts of retaliation, reciprocal bids (as in bridge), echoes, or antiphonal singing.
Another unifying element is its subschema which relates to value judgments. In
extensive studies on the role of values in semantic extensions, Simon-Vanden-
bergen (this volume), Pauwels and Simon-Vandenbergen (this volume), and
Krzeszowski (1993) all point to the importance of the axiological parameter to
categorial growth (cf. also Rudzka-Ostyn 1988). This finding is confirmed by
the present investigation as responses to prior acts often subdivide along ax
iological lines.
Figure 8 presents the upper part of the schematic hierarchy that structures
the semantic space covered by verbs of answering and shows how this space is
parcelled out among them. Parentheses indicate that the verb in question plays a
minor role in instantiating the given schema,12 whereas square brackets mean
that the verb no longer instantiates it. Rejoin and retort belong to the latter
category for they have lost their general communicative sense of 'react verbally
to another's utterance'.13 They are mentioned simply to highlight the common
semantic core that the field had at one time. In order not to overload the figure,
we have introduced schemata that contain disjunctive information. The left-most
schema, for instance, comprises at least six different schemata, and to see how
exactly it relates to the individual verbs, the preceding figures must be con
sulted.
238 BRYGIDA RUDZKA-OSTYN
Figure 8
METAPHOR, SCHEMA, INVARIANCE 239
When we follow the evolution of the semantics of these verbs, we notice that
sense extensions do not necessarily proceed step by step. Often intermediate
stages that might be postulated on the basis of all the meanings expressed by a
given verb are skipped. Highly abstract variants come into use before less ab
stract ones. To take an example, answer had lexicalized the correspondence
schema 300-400 years before certain of its more concrete submeanings devel
oped (see Figure 3: (d), (e), (1), or (m)). Similarly, respond had expressed cor
respondence relations before it came to denote acts performed in reaction to
other acts (see Figure 5). One could, of course, say that this development was
strongly influenced by translations from Latin. But such influence cannot be
invoked in the case of answer, and so a few words of explanation are due. Be
fore, however, attempts are made in this direction, it may be useful to return to
the Invariance Hypothesis.
Its latest version, which both Lakoff and Turner accept, reads as follows:
"In metaphoric mapping, for those components of the source and target do
mains determined to be involved in the mapping, preserve the image-schematic
structure of the target, and import as much image schematic structure of the
source as is consistent with that preservation" (Turner 1990: 254). 14 We want
to know, of course, what would constitute the relevant image-schematic struc
ture in the case of our source and target domains. More concretely, when an
swer extends into e.g. the domain of satisfaction and fulfilment, what precisely is
preserved? On the side of the source we have verbal responses which, proto-
typically, are meant to meet the primary speaker's desire for information; and on
the target side we have a multitude of acts intended to satisfy someone's mate
rial needs, feelings, hopes, or expectations. We would argue that answer is used
in this particular target domain because the schema that can be abstracted from
the multitude of acts is the same as the one that underlies typical verbal answers.
In each case, the subsequent act or situation corresponds to what is wanted.
The concept of meeting a desire constitutes the common schema that sanctions
the use of answer in place of satisfy or fulfil. It is this schema, it seems, that is
preserved by the metaphoric mapping. Most, if not all, of the other concepts
associated with answer are suspended, and what is mapped is consistent with
the structure of the target. So Lakoff and Turner are right? Yes and no. As al
ways, there is a snag.
We have already mentioned problems posed by targets which do not have
any preexisting image-schematic structure (Brugman 1990). The target domains
encountered in this study are not unproblematic either. To return to the domain
240 BRYGIDA RUDZKA-OSTYN
though not exclusively, on the perspective adopted. If you view the communi
cative meanings of retort from the perspective of its earlier uses as a verb of
physical retaliation, the communicative uses are an extension via mapping. If,
however, you take the present-day situation as your vantage point and, follow
ing some dictionaries, treat the communicative uses as primary, then all the
other uses may be regarded as extensions from the communicative meanings. In
the case of respond matters become further complicated by the shift of its centre
of semantic gravity towards non-communicative domains. For some speakers
the very general meaning 'do something in reaction to prior act' has become the
most representative for the whole category. What is then the status of the other
meanings? The communicative uses of respond are instantiations of the general
meaning, but instantiations of what kind?
This is where the concept of metonymy may be brought in. In his work on
metaphtonymy, Goossens (1990) makes us aware of the close symbiosis be
tween metonymy and metaphor and explains its mechanism. Once we adopt the
schematic-network approach, a yet broader dimension of this symbiosis reveals
itself. To appreciate its value, let us pause for a moment and consider the notion
of metonymy. Traditionally, the term has been used to cover a variety of conti
guity relations. It has applied primarily to referents of concepts. Alongside this
prototypical conception of metonymy, an extended one has arisen. Lakoff's
(1987) research into categorial structure has shown that a very similar type of
relation is attributed to concepts themselves. Underlying this attribution is again
a metaphor, this time a metaphor according to which concepts that are associ
ated with one another are viewed in terms of physically contiguous objects.
Since both language users and language analysts seem to rely on such an ex
tended conception of metonymy (cf. Rudzka-Ostyn 1989), we may want to
probe it further. One result of this probing is that any extension effected via ab
straction, metaphoric or not, can be seen as involving a metonymic disso-
ciation. Answer, which typically denotes the whole cluster of concepts isolated
in section 2.1, comes to be metonymically used to denote only part of the clus
ter. Where meaning extension involves generalization, i.e. concept suspension,
metonymic mechanisms are at work. Questions immediately arise with respect
to the reverse process. What happens when a meaning is extended via speciali
zation, via acquisition of new concepts? And is this process still metaphoric?
These questions bring us back to the semantic structure of respond. When
queried about the relationship between the most general meaning of the verb
and its communicative uses, native speakers of English refused to identify it as
242 BRYGIDA RUDZKA-OSTYN
Notes
* I am grateful to Peter Kelly for his valuable suggestions and to Hans Paulussen for com-
puterdrawing my figures.
1. Cf. Lehrer (1978, 1990), Ortony (1979), Honeck & Hoffman (1980), Lakoff & Johnson
(1980), Kittay & Lehrer (1981), Johnson (1981, 1987), Miall (1982), Paprotté & Dirven
(1985), and Kittay (1987).
2. The following abbreviations will be used to refer to the sources:
CAT = Catch-22 by Joseph Heller, New York: Dell, 1968.
METAPHOR, SCHEMA, INVARIANCE 243
Goossens, Louis
1986 "Framing the Linguistic Action Scene in Old and Present-day English: Old Eng
lish cwepan, secgan, sprecan and Present-day English speak, talk, say and tell
compared." In J. Fisiak (ed.) Papers from the Sixth International Conference on
Historical Linguistics. Amsterdam: John Benjamins, 149-170.
1987 "Dealing with Linguistic Action Verbs in Depth." In J. Verschueren (ed.) Lin
guistic Action: Some Empirical-Conceptual Studies. Norwood, N.J.: Ablex Pub
lishing Corporation, 83-106.
1989 "Prototypical and Non-prototypical Meaning: Say Revisited." In F.J. Heyvaert &
F Steurs (eds.) Worlds Behind Words. Essays in Honour of F.G. Droste on the
Occasion of His Sixtieth Birthday Leuven: Leuven University Press, 155-161.
1990a "Framing the Linguistic Communication Scene: ask vs. acsian and biddan." In H.
Andersen & K. Koerner (eds.) Historical Linguistics 1987, Amster
dam/Philadelphia: John Benjamins, 191-209.
1990b "Metaphtonymy: The Interaction of Metaphor and Metonomy in Expressions for
Linguistic Action." Cognitive Linguistics 1.323-340.
1993a "The Rise of a New Conceptualization Pattern: Old English lippe, weler, tunge
and with Reference to Linguistic Action." In H. Aertsen & R. Jeffers (eds.)
Historical Linguistics 1989. Amsterdam/Philadelphia: John Benjamins, 141-151.
1993b ", mouth(e), mouth, Denoting Linguistic Action: Aspects of the Development
of a Radial Category." Acta Linguistica Hungarica 38.61-81.
1994 "Metonymy in the Pipeline: Another Way of Looking at the Container Metaphor".
In K. Carlon, . Davidse & . Rudzka-Ostyn (eds.) Perspectives on English.
Studies in Honour of Professor Emma Vorlat. Leuven: Peeters, 386-394.
Hiź, Henryk, ed.
1978 Questions. Dordrecht: Reidel.
Honeck, Richard P. & Robert R. Hoffman, (eds.)
1980 Cognition and Figurative Language. Hillsdale, N.J.: Erlbaum.
Jaeger, Jeri J.
1986 "Concept Formation as a Tool for Linguistic Research." In Jeri J. Jaeger & John
Ohala (eds.) Experimental Phonology. Orlando: Academic Press, 211-237.
Johnson, Mark
1987 The Body in the Mind. The Bodily Basis of Meaning, Imagination and Reason.
Chicago/London: University of Chicago Press.
----- (ed.)
1981 Philosophical Perspectives on Metaphor. Minneapolis: University of Minnesota
Press.
Kay, Paul
1987 "Linguistic Competence and Folk Theories of Language. Two English Hedges."
In D. Holland & N. Quinn (eds.) Cultural Models in Language and Thought.
Cambridge: Cambridge University Press, 66-77.
Kittay, Eva F.
1987 Metaphor: Its Cognitive Force and Linguistic Structure. Oxford: Oxford Uni
versity Press.
REFERENCES 247
Mackenzie, J. Lachlan
1985 "Metaphor in Contemporary Semantics." Dutch Quarterly Review of Anglo-
American Lettters 15.52-72.
Miall, David S. (ed.)
1982 Metaphor: Problems and Perspectives. Hassocks, Sussex: Harvester Press.
Miller, George A. & Philip N. Johnson-Laird
1976 Language and Perception. Cambridge, Mass.: Harvard University Press.
Newman, John
1993 "The Semantics of Giving in Mandarin." In R. Geiger & . Rudzka-Ostyn (eds.)
Conceptualization and Mental Processing in Language. Berlin/New York: Mou
ton de Gruyter, 433-485.
Oxford Advanced Learner's Dictionary of Current English. 1980, 1989 Oxford: Oxford Uni
versity Press. [OALDCE]
The Oxford English Dictionary. 1933. Oxford. Oxford University Press. [OED]
Papproté, Wolf & René Dirven (eds.)
1985 The Ubiquity of Metaphor. Amsterdam/Philadelphia: John Benjamins.
Pauwels, Paul & Anne-Marie Simon-Vandenbergen
1993 "Value Judgement in the Metaphorization of Linguistic Action." In R. Geiger &
. Rudzka-Ostyn (eds.) Conceptualization and Mental Processing in Language.
Berlin/New York: Mouton de Gruyter, 331-367.
The Penguin Roget's Thesaurus. 1966. Harmondsworth: Penguin.
Reddy, Michael
1979 "The Conduit Metaphor. A Case of Frame Conflict in Our Language about Lan
guage." In A. Ortony (ed.) Metaphor and Thought, Cambridge: Cambridge Uni
versity Press, 284-324.
Rudzka-Ostyn, Brygida
1985 "Metaphoric Processes in Word Formation." In W Papproté & R. Dirven (eds.)
The Ubiquity of Metaphor. Amsterdam/Philadelphia: John Benjamins, 209-241.
1988 "Semantic Extension into the Domain of Verbal Communication." In B. Rudzka-
Ostyn (ed.) Topics in Cognitive Linguistics. Amsterdam/Philadelphia: John Ben
jamins, 507-553.
1989 "Prototypes, Schemas and Cross-Category Correspondences: The Case of Ask."
Linguistics 27.613-661.
1994 "Metaphor, Schema, Invariance: The Case of Verbs of Answering." In K. Carlon,
. Davidse and . Rudzka-Ostyn (eds.) Perspectives on English. Studies in Hon
our of Professor Emma Vorlat Leuven/Paris: Peeters.
----- (ed.)
1988 Topics in Cognitive Linguistics. Amsterdam/Philadelphia: John Benjamins.
Searle, John R.
1979 Expression and Meaning. Studies in the Theory of Speech Acts. Cambridge:
Cambridge University Press.
& Daniel Vanderveken
1985 Foundations of Illocutionary Logic. Cambridge: Cambridge University Press.
REFERENCES 249
Simon-Vandenbergen, Anne-Marie
1991 "The Body in Talking about Language. A Study of Metaphors." In S. Granger
(ed.) Perspectives on the English Lexicon. A Tribute to Jacques Van Roey. Ca
hiers de l'Institution de Linguistique de Louvain 17 (1-3), Louvain-la-Neuve, 51-
63.
1993 "Speech, Music and Dehumanization in George Orwell's Nineteen Eighty-Four:
A Linguistic Study of Metaphors." Language and Literature 2.157-182.
1994 "Metaphorical Verbal Clauses in George Orwell's Nineteen Eighty-Four." In K.
Carlon, K. Davidse & . Rudzka-Ostyn (eds.) Perspectives on English. Studies in
Honour of Professor Emma Vorlat. Leuven: Peeters, 395-407.
Stainton, Caroline
1987 "Interruptions: A Marker of Social Distance?" Occasional Papers in Systemic
Linguistics 2.75-132.
A Supplement to the Oxford English Dictionary. Vols 1-4. 1972-1986. Oxford: Oxford Uni
versity Press. [OED-S]
Talmy, Leonard
1985 "Force Dynamics in Language and Thought." In W.H. Eilfort, P.D. Kroeber &
K.L. Peterson (eds.) Papers from the Parasession on Causatives and Agentivity at
the Twenty-First Regional Meeting of the Chicago Linguistics Society. Chicago:
Chicago Linguistics Society, 293-337.
1988 "The Relation of Grammar to Cognition". In B. Rudzka-Ostyn (ed.) Topics in
Cognitive Linguistics. Amsterdam/Philadelphia: John Benjamins, 165-205.
Traugott, Elizabeth Closs
1985 "'Conventional' and 'Dead' Metaphors Revisited." In W. Papproté & R. Dirven
(eds.) The Ubiquity of Metaphor. Amsterdam/Philadelphia: John Benjamins, 17-
56.
Turner, Mark
1990 "Aspects of the Invariance Hypothesis." Cognitive Linguistics 1.247-255.
1993 "An Image-Schematic Constraint on Metaphor". In R. Geiger & . Rudzka-Ostyn
(eds.) Conceptualization and Mental Processing in Language. Berlin/New York:
Mouton de Gruyter, 291-306.
Vanparys, Johan
1993 The Categorization of Illocutionary Verbs. Doctoral Dissertation, K.U.Leuven.
Verschueren, Jef
1981 "Problems of Lexical Semantics." Lingua 53.317-351.
1984 "Basic Linguistic Action Verbs. A Questionnaire." Antwerp Papers in Linguistics
37.
1985a What People Say They Do With Words: Prolegomena to an Empirical-Conceptual
Approach to Linguistic Action. Norwood, N.J.: Ablex Publishing Corporation.
1985b International News Reporting: Metapragmatic Metaphors and the U2 Incident.
Amsterdam/Philadelphia: John Benjamins.
1987 "Metapragmatics and Universals of Linguistic Action." In J. Verschueren (ed.)
Linguistic Action: Some Empirical-Conceptual Studies. Norwood, N.J.: Ablex
Publishing Corporation, 125-140.
250 REFERENCES
Webster's Third New International Dictionary of the English Language. 1976. Springfield,
Mass.: Merriam. [Webster's]
Wierzbicka, Anna
1985a "A Semantic Metalanguage for a Crosscultural Comparison of Speech Acts and
Speech Genres." Language in Society 14.491-514.
1985b "Different Cultures, Different Languages, Different Speech Acts. Polish vs Eng
lish." Journal of Pragmatics 9.145-178.
1987 English Speech Act Verbs. Sydney: Academic Press.
Subject Index