Pragmatics
2:1.49-62
InternationalPragmatics
Association
METHAPOR, MEANING, AND COMPREHENSION
BegoflaVcente
Black'sinteraction theory of metaphor has had a formidable impact on the
literature.On the positiveside,his formulation hasproved very enlighteningin many
respects,particularlyin focusing the attention of linguistsand philosophers on the
natureof conceptualstructure and the possibilitiesof growth in human knowledge
throughthe expressivepotential of language.On a less positive side, he could be
heldresponsible,in part, for some of the confusionin the literature concerningwhat
hasbeencalled'metaphoricalmeaning'.
As is well-known, Black argued very strongly against comparative, and in
generalsubstitutivetheories of metaphor on the grounds that they assigned a
peripheraland decorativerole to this type of expression.He was,without using the
pragmaticterminologyavailable today, rebelling againstthe assumptionthat there
are literal equivalentsto metaphoricalexpressions,probably on the correct intuition
that speakersonly rarely engagein verbal decoration for the sake of it. For Black,
paraphrases
of metaphoricalexpressionsare unsatisfactoryto everyone,not because
theyfail to capturethe expressive,emotive or suggestiveelement in metaphors,but
becausethe conventionalresourcesavailablein the languageare not up to the job
of capturing the 'irreducible cognitive content' that metaphors carry, which is
typicallynew and created by the metaphor itself -at least in the case of strong,
interactionmetaphors.This was something that unrefined comparison theories of
metaphorcould not handle, for they could only deal with the transfer of attributes
whichare alreadypart of the meaningsof the terms involved.The widespreadAristoteliannotion of the displacedsign whereby a name is applied to something that
it doesnot conventionallydesignatefor the purposesof shockingthe audience has
beensomewhatout of favour lately, sinceBlack set theoreticiansafter the cognitive
functionof this type of expression.Black's strong creativity thesisassignedmetaphor
a specialkind of meaningresulting from the interaction of the terms in it -or rather
the implicativecomplexesinvolved- and which is:
a) new, or creative,not inferrable from the standard lexicon,
b) is not paraphrasable,at least not completely,
c) brings about conceptualinnovationsand therefore generatesnew
knowledgeand insight. (M. Black 1979:23)
In his 1979 paper he goes to great lengths to show how this can be so, how
"metaphorsenableus to see aspectsof reality that the metaphor's pi.oduction helps
to constitute".(M. Black 1962:39)
Begona Wcente
Now, while we do want to hold on to Black's idea of the creative power of
metaphor I find endlesstrouble with his accountingfor it in terms of a special
cognitivecontent or meaning.As a studentof metaphor my own view is that loose
or carelessand sometimesmisguidedtalk of metaphoricalmeaning,especiallyby
linguists, has caused much confusion in the literature. In this respect I find D.
Davidson's(1980)caution againstthis stateof affairs,highlyvaluable.In a blunt and
provocativestatementhe declaredthat
" [...] metaphors mean what the words in their most literal interpretation
mean, and nothing more." (D. Davidson 1980:238)
This no doubt came as a shock for many who interpretedit in the sensethat there
was nothing worth studying concerningthe topic of metaphor. But this is obviously
a misinterpretation.If he is right, then for a sentencelike (1) below,the sameas for
any other sentencein the language,we have its meaning specifiedby the set of
necessaryand sufficient conditionsfor its truth:
(1) Richard is a gorilla.
As a sentence,(1) means that a certain entity/individual is a member of the set of
all gorillas, and we know its meaning when we know what the world would have to
be like for the sentenceto be truer. That whoeveruttered the sentencemust have
had the intention of conveyingsomethingelse by this obvious flouting of Grice's
maxims of conversationis something that lies outside the scope of a semantic
theory. When we talk about the meaningof a metaphor this is, accordingto Davidson, as far as we can go. Clearly, what he is doing is simply applying the distinction
between meaning and use (or sentencemeaning and speaker'smeaning) to the
study of metaphorical utterances.The effects that the 'imaginative employment of
words' have on us -he went on- need not concern,in principle,a theory of meaning.
When we talk about the specialmeaningof a metaphorwe are simply blurring this
fundamentaldistinction.Thlk of metaphorsas the bearersof cognitivecontentsor
special meaningsarises from a failure to distinguishbetween meaning and use.
Thus, while for Black the problem with paraphrasesis that they "fail to give the
insight that the metaphor did" (Black 1962: 46), for Davidson, if the so-called
content of a metaphor cannot be put into words, then, by definition, it is not a
content, it must be somethingelse.We cannot have it both ways,Davidson seems
to be saying:
' As is well known, this is how meaning is characterized in truth-theoretic approaches.
Blakemore (1987) & Carston (1988) within the Relevance Theory framework argue quite
convincingly for the necessityto draw a distinction between linguistic semanticsand truth-conditional
semantics,something that J. Katz had been insisting on as well, but from a slightly different position.
The issue is important because pragmatic aspectsare shown to play a decisive role in the latter but
not in the former, since linguistic meaning is decoded and is truly context-free, while truthconditional content is partly inferred in context. For the purposesof the point I want to make here,
however, this simpler characterizationwill do.
Methaphor, meaning and comprehension
51
"[...] the usual view wants to hold that a metaphor does somethingno plain
prosecan do, and on the other hand it wants to explainwhat a metaphor does
by appealing to a cognitive content -just the sort of thing plain prose is
designedto express."(D. Davidson1980:251)
From here Davidsonconcludesthat it is not the meaningof the metaphor that we
'effects
find difficult to p'araphraseor grasp, for that lies on the surface, but the
metaphorshave on us':
"The common error is to fasten on the contentsof the thoughts a metaphor provokes and to read these contents into the metaphor itself."
(Davidson1980:251)
There is no denyingthat metaphor makes us notice aspectsof reality we had not
discoveredbefore, but not by 'standingfor or expressingthe fact'. The paraphrase
failswhenwe want it to capturewhat the metaphorbringsto our attention.Because
that is not finite in scope-which is just what Black was suggestingwhen he talked
aboutthe ambiguityand open-endedness
of metaphoricalinterpretations-nor need
it be propositionalin nature. For if it were, it should be expressiblein literal
language.Metaphors may trigger all kinds of thoughts, impressions,suggestions,
imagesand feelings.The number and strength of the implications one may derive,
the correspondences
it may invite us to draw between the terms or conceptual
domainsinvolved cannot be called the metaphor's cognitive content or special
meaning.The richnessof metaphorical uses of languagedoes not lie in their
semanticmeaningbut in their pragmaticor communicativevalue, and as Levinson
(1983)puts it, there's nothing denigratingin approachingmetaphor as a use of
languageand therefore as a pragmatic phenomenon.What we have to do is simply
'division
apply the
of labour' between semanticsand pragmatics to the study of
metaphor:
"A pragmaticapproachwill be based on the assumptionthat the metaphoricalcontentof utteranceswill not be derivedby principlesof semantic
interpretation;rather the semanticswill just provide a characterizationof
the literal meaning or conventionalcontent of the expressionsinvolved,
and from this,togetherwith detailsof the context,the pragmaticswill have
to providethe metaphoricalinterpretation."(S. lrvinson 1983: 156)
Metaphorsare no doubt cognitivelyhelpful. But we can only speakof the cognitive
contentof an utterance-as Warner (1973:368)argues-"in so far as it is a possible
object of knowledge,and hence capable of being true or false". He thinks, very
muchin the sameline as Davidson,that Black goeswrong because
"He appearsto be running together the thesisthat good metaphors play
a distinctiveand significantrole in enablingus to understandourselvesand
the world, with the thesisthat they conveyinformation unparaphrasable
by any nonmetaphoricalutterance."(M. Warner 1973:369)
52
Begona Vicente
Manns (7975:360)argues againstWarner that Black should be read as equating
'cognitive content'
with 'insight' since in Black's own words, that is what the
paraphrasecannot handle, and so,
"[...] Black's own suggestionthus points forcibly towardsan interpretation
'cognitive
of
content'far broader than 'that which makes a truth claim'."
I agree with Manns that
'cognitive
"[...]
is a term applicable to a range of phenomena far more
extensivethan the classof verifiable or falsifiable statements,[...]"(J.W.
Manns 1975:361)
but I am sure that it is not with the word 'cognitive'that Davidsonfound fault, but
with the word 'content', that is why the rest of the quote from Manns brings us back
into the same confusion:
"[...] un utterancecan have cognitivecontentwhen it is employedin any of
the following forms: (1) directing our attention towards a feature or
featuresof our perceivedenvironment;(2) aiding us to cultivate a skill or
master a practical obstacle;(3) altering our way of categorizingthe world
t...h (+) making a truth claim."(J.W.Manns L975:361)
On pain of being accusedof splittinghairs I think it's saferto talk about 'cognitive'
'communicative
or
import'when referring to information implicitly conveyedby an
utterance and differing from the propositional cohtent of the sentence used. The
word 'content'is linked to conventionalmeaningand semanticsunderstoodas that
branch of linguisticswhich dealswith the invariant context-freeaspectsof meaning
the languagesystem assignsto words, phrasesand sentences.2
We gain nothing, in
theoreticalterms, by usingthe sameword to refer to both the insights,thoughtsetc,
a certain use of an expressionmay trigger in a certain interpreter and to the
information it invariablycarrieswith it, even when it may not be what the speaker
is primarily interestedin communicating.[t is not merely a terminologicalquestion
that is at stake here, but a clear failure to distinguishpurely semantic from
pragmatic aspectsof verbal communication.In his later papers (1979 and 197819)
Black really seemsto waver betweenan awarenessof the existenceof thesenotions
and a tacit dismissalof their value: when he usesthe word 'say'it encompasses
the
explicit as well as the implicit3; his use of the word 'meaning' is likewise totally
2 Perhaps it could be argued that philosophers use
'content',
the term
in a pre-theoretical sense,
'form'
as opposed to
without further specification. What I am trying to show is that this kind of use
becomes grossly inadequate for the purpose of explaining non-literal usesof language. See also note
l.
3 M. Black: 'The author of a metaphor says something'; "[...]
he [the author of the metaphor]
was saying various things, many of them implicitly'. (1978i9:185)
Methaphor, meaning and comprehension
53
unrestricted4'
If the so-calledchangeof meaning that
takes place in metaphorical
interpretationis not permanent,as-Black
.".o!iir"s, but ,"ii", on the ,,metaphor
maker(is) attaclringan altered senseto the
*o.i, he is usingin ,ont""t,, (his italics)
(M' Black 1979:
are clearlytalkingof a widespreadphenomenon
?9),Y"
in linguistic
communication
that far exceedstire terriiory of the rn.tuphorical
and brings us well
into the domain of utterance interpretation,
where we have to fincl the general
principleswhereby sentencescan be
used to convey something additional or
differentfrom their literalmeaning.In view
of this,Jissatisfaction
*iit metaphorical
paraphrasing
cannot be phrasedin terms of a lossof content,
but has to be linked
to the natureof the distinctionbefween
explicit content and implicit import. Thus,
no 'translation'from the latter into the former
will ever fuily satisfyproducersand
interpreters,
sincethe implicit typicallyleavessomemargin -which
,un u" very wide
or verynarrow-for personalvariation,a margin
which ii not p."*niin the caseof
explicitcommunication.
But notice that dissati"staction
with paraphraseso expressed
makesit difficult to maintain that what is
lost is a ,cognitivecontent,.s
Davidson'stheory of meaning -if correctneed not demolish an interactive
theoryof metaphor.All we would fru" to do
is reformulate what the cognitive value
of metaphorconsistsin. Besides,our enthusiasm
for the cognitivepower of metaphorsshouldnot let us forget that not every
interestingmetaphoricalutterancehas
thesamepowerto restructureconceptual
domains,amongother thingsbecausethat
need not be the point of every conceivable
meiaphorical utterarice. Analogies,
models and examples also help in our growth
of knowledge and are of great
cognitivevalue; besides,they alsb work by
Iorcing ,r," inr"rpreter to draw structural
correspondences'
and yet, we are not as prone to talk about
their ,irreducible
cognitivecontent' or 'speciarmeaning', und
.ori-p"opre are happy over their
paraphrasibility.Davies,sexample
(2) Malcom is a thief
4M' Black:'l
usethe word 'meaning'herefor whatevera competent
speakermaybe saidto have
grasped
when he succeedsin respondingadequately
to the actual or hypotheticalverbal action
mnsistingin the seriousutteranc€of thi sentence(s)
in question
". (lg7g:2\ He also insiststhat
whenhe talksabout 'shiftsof meaning'he
means'shifis in the speaker,smeaning,
ltslsrzsl.
5I think
the issueoverthe paraphrasability
of m,etaphorcan be settledquite satisfactorily
using
thenotionsof explicitano impiicit communication.
If ui pr.rpi'*se we understandfinding a literal
alternative
wayof wordingthi literal contentor an
e*piesrioi, itr"n sentences
usedmetaphoricary
areparaphrasable
to the sameextentthat any other sentence
is. If, on the other hand,we want the
paraphrase
to capturewhat the metaphoricalu.tterance
is-impficitry taken to conveyin speech
situations'then again' this is, in principle,equally
possibleor impossibleto the sameextent
for
metaphorical
and nonmetaphoricalutt".un."r. A key notion
here is that of the indeterminacyof
implicatures,
of which.Gr.iceltslsl spokeand which spe.u"i*
wilson givea decisiverote in their
pragmatics'
and which has to do with the degree
tr ,p"J", backing we can atrach to the
implicatures
we derivewhen interpretingan utterance.
JuJt thint of a possibleparaphrasefor an
utterance
like "Mary is quite dissatisfied
*ittr you, workn,or *nlcn we wourddemand
irrat it give us
thefull communicative
import.
54
Begofia Vicente
'a
in which one is informed that a friend whom one took to be pillar of probity and
integrity' (Davies 198213:75), has stolen a large sum of money is a good one to
illustrate this point. Notice that (2) would have a great impact on the hearer'sview
of his friend and it would certainlyforce a reinterpretationof a great many assumptions about this Malcom, and yet, we would not want to talk about metaphor in this
case.
There's the further point I mentioned before, that the openendednessand
suggestivepotential also found characteristicin metaphoricalutterances,can be
found in other forms of figurative as well as non-figurative language,and can be
shown to be a matter of degree for utterances in general rather than a property
exclusiveto metaphoricalspeech.Think of a sincerereply to the question"What's
your girlfriend's name?" of the form
(3) B: I don't know.
We may find this answervery suggestive,but again, it's the interaction between B's
words and the assumptionsA brings to bear on their interpretationthat makes B's
words suggestive,not just their meaning.
My proposal would be then that we first look at what the availablepragmatic
theories have to say on the generalquestionof the mechanisms-linguisticas well
as non-linguistic-underlyingutteranceinterpretationin general,before we rush to
'special'
label
every aspectof the behaviourof metaphoricalutterances.
The interestof the points Davidsonmakesis that they bring into the open one
of the main sourcesof confusion in the literature on metaphor, namely, the lack of
terminological and theoretical clarity and precision.There is far too much talk on
the meaning of metaphor that relies on an unfortuhately nonexistingagreement on
'meaning'.
the scope of the term
That's why Davidson'sessayis a useful starting
point.
What is badly needed in the literature is a rigorous application of the
terminology available since Grice brought it to our attention that linguistic
communication relies fundamentallyon much more than the coded meaningsof
terms in the language.He talked about what is said -the truth-conditionalaspects
of the sentence uttered- and what is implicated -whatever other information is
derived inferentially from the safng of x in the circumstances6.
A very good therapy would be to stop talking of metaphors or metaphorical
sentencesand concentrateon metaphoricalutterances.No sentenceis a metaphor.
In spite of the deeply rooted idea that there is a classof sentencesthat can be
instantlyrecognizedas metaphorsnobodyto my knowledgehasbeen able to list the
characteristicsthat mark off metaphoricalsentencesfrom the categorymistakesof
the child or the second-languagelearner, for example. As Black himself acknowledges, "...every criterion for a metaphor's presence,however plausible is
defeasiblein specialcircumstances"
(M. Black 1979:36).The reasonfor this being
6 See Wilson & Sperber (1981),D. Blakemore(1987)and R. Carston (1988),for a revised
characterizationof the distinction.
Methaphor, nteaning and comprehension
55
that it is speaker'sintentionsthat make an utterancemetaphorical,as M. Reddy
notedalreadytwo decadesago.Somepeople have rejectedthis idea on the grounds
that
"Suchintentionsare neither necessarynor sufficientfor determining that
an utteranceis metaphorical.On the one hand, we may intend but fail to
makemetaphor-we may not be sufficientlycompetentin the language.On
the other hand,we may interpret as metaphoricalstatementswhich were
neverso intended."(E.F, Kittay 1987:46)'
This is certainlyso only if we considerspeaker'sintentionsin independencefrom
'bound',
by the very
discourse
situations,that is, situationsin which the speakeris
natureof the form of behaviour she has engagedin, to be relevant and cannot
normally afford to be misunderstood: an utterance is always the product of an
ostensive
intention to inform an audienceof something,and the audiencecannot
help but take it that way. Since communication is all about the providing of
evidenceof jntentionson the part of speakersand the recognitionof those intentionson the part of hearers,(as Grice (1975)argued,and more recentlySperber &
Wilsonhave convincinglyexploited to its most interestingconsequences)
talk of a
sentence
as being meant metaphoricallycan only mean that the speaker assumes
that her intention will be recognizedby the hearer on the grounds that she has
chosenthe most relevantverbal stimulusat her disposalto conveythe information
sheintendedto communicate.Of course,she may go wrong and the hearer may
interpretthe utterancein a different way from that envisagedbut that applies as
muchto metaphoricalutterancesas to any other utterance.Communicationtakes
placeat a risk, relevancetheory reminds us, and that's only too true8.
The feeling that one needs neither a context nor any assumptionsabout the
intentionsof the speakerto know that a sentencelike (a) below is to be interpreted
metaphorically,
(for -it is argued-nobody in their senseswould want to produce or
interpretit literally) has mistakenlyreassuredsome theoreticiansthat metaphor is
a semanticmatter.
(4) New York City is a pressurecooker.
To the counterargumentusually levelled againstthis line of reasoningthat some
sentencesare undoubtedly perceived as metaphorical only against a particular
contextof utterance,one can alwaysretort by arguing that literal languagecan also
be shownto be context dependentto a large extente.
7
Seealso Ortony (1980:71),for similar views.
8 SeeSperber & Wilson's Relevance:Communication & Cognition (1985). I'm relying throughout
their
view of communication and on some of the terminology they have put fonvard.
on
' L.J. Cohen (1979)'The semantics of metaphor' (in Ortony 1979) has used this argument in
favour of a semantic theory of metaphor.
56
Begofia Vicente
But this will not do. In the first place,as a sentence(4) has no literal range of
application -it is false- becauseof its sortal anomaly.As an utterance it would
certainly be easy to attach to it communicativevalue, but this is so becauseno
linguistic information is ever interpreted in a null context -leaving aside the
metalinguisticusesof linguistsl0.We must distinguishour generalcapacityto make
senseof the things around us, especiallywhen they come -as utterancesdo- in the
shape of ostensivestimuli, from the formal characterizationof the sentencesin the
languagesystem.Unlike utterances,sentencesare abstractobjectswith only formal
propertiesand thereforewhat we are doingwhen we claim that we can interpret (4)
as a sentence,metaphorically,is simply insert it in some form of default context, and
attach to it the most stereotypic assumptionsaccessedvia the linguistic and
encyclopedicinformation that the terms in the sentencemake available-and thus
we treat it as a form of purposeful behaviour. This is never difficult since as
speakersof the same language-or very similar ones- we share not just linguistic
information but much information that is non-linguistic.In principle there is no
absolute guarantee that nothing but a metaphorical reading will do for a sentence
like (a). Berg (1988:698-699)has used this example to argue, againstDavidson's
claim that "metaphors cannot be paraphrased,[...] becausethere is nothing to
paraphrase,"the following:
"How is Davidson's position to be reconciled with the seemingly overwhelming evidence against it? For instance, in a discussion about the
quality of life in different cities someoneonce said to me "New York is a
pressure cooker". I see no difficulty in paraphrasinghis remark by saying,
He told me that life in New York City is very stressful.And we can easily
imagine the speaker confirming my report. How can Davidson maintain
that no propositional content can thus be ascribedto the metaphorical
remark?"
Firstly, the paraphrase he offers could hardly be accepted as the propositional
content of (a). Rather, it qualitiesas a strong implicatureof (4) in the context he
provides for its interpretation, that of "a discussionabout the quality of life in
different cities". We just need to change the context in which (4) was uttered and
a different implicature comes to the front of what is being communicated.I can
easilyimagine a conversationalsituationin which what the hearer would construct
as the main implicitly conveyedinformation would be somethinglike (5)
(5) In New York City artistic activity is more
intense than in other places.
But of course,we do not want to claim that every implicaturean utterancemay be
taken to convey in a context -howeverstrongly-is its propositional content, for what
would be the use of such a notion?
1oThis is a point that J. Searle (19'79b) has made.
Methaphor, nteaning and comprehension
57
Another problem that so-called semantic theories of metaphor (and also
Searle's)have to face is precisely their failure to reconcile two apparently
'proposition
viewsof metaphor,which we can call -followingDavies-the
contrasting
theories'-those who seek to capture in propositionalform what the metaphor
'image theories'-thosewho see this way of going about
and the
communicatesmetaphoras the renderingof what is the most preciousingredientof metaphorical
reorganizingpower,etc.The questionof whether
speech,
its richness,suggestiveness,
we are dealingwith two different theories for two different types of metaphor can
be easilyoverridden if we show that one and the same utterance -whatever its
syntactictype- can be used to convey a certain proposition -somethingspecificother than that literally expressedby the sentenceused, while at the same time
communicatea whole range of weaker assumptions,indeterminatein number and
form, i.e., while staying open-ended.This time we can have it both ways. It's
importantto emphasizethat in both caseswe are talking about implicit information,
hencethe indeterminacy.What differs is only the degreeof speaker'scommitment
with the assumptionsconstructedor the degree of responsibilitythat the hearer
'common ground'll
takesin their derivation,and here a crucial notion is that of
betweenspeakerand hearer.Both Gibbs (1987)and Bergmann(1982) -outsidethe
'deliberate
relevancetheory framework- have seen it this way. Gibbs talks of
ambiguity',and the example he givesis "This room is a toilet", something his father
usedto sayto him when he was a teenager.His impressionisticline of argument is
the following: "Each time he said this over the years I always felt as though I
understoodhis intended meaning,which to me implied that he thought the room
wasmessyand in dire need of cleaning.However, his metaphoricalutterance (...)
struckme differentlyevery time he said it. (...) What I did senseeach time he said
"Thisroom is a toilet" was some kind of negativeassociationwith toilets, as though
toiletssometimessmelled unpleasant,or were unpleasantto look at on occasion.
Each of these inferencesseemed to be a part of the meaning of my father's
metaphoricalstatements,and I always sensedthat these were indeed authoized".
(p.44).This seemsto me to be a lessdistortingsort of approach to pursue in the
studyof metaphoricalutterances.
The other two notions.which one very commonly encounters in the literature
on metaphoricalmeaning and cognitive or propositional content are those of
metaphoricaltruth and assertivepower.
Searle(7979a),for example,startedoff his analysisof metaphoricalutterances
by placingit neatlywithin the domain of pragmatics.For him, metaphoricalutterancesare just one special case of a more general phenomenon in language
consisting
in the sayingof somethingbut meaninganother.He insiststhat there are
not two kinds of meaning one literal and one metaphorical,that sentencesand
wordshaveonly the meaningthat they have and nothing else.The work to be done
consistsin formulating the principles relating the literal meaning of the expression
usedto its metaphoricalinterpretation.This relationshiphad to be systematicand
restricted,not random. It was probably this requirement of systematicity,which
1rSperber & Wilson speak of 'mutual manifestness',a more elaborated nolion.
58
Begofia Wcente
indeed has to be met, that led him into proposingthe following analysiswhere he
clearly proves unable to maintain the basic distinctionshe had started off by
establishing.
First, he proposesthat in analyzingmetaphor we should have two sentences,
one for the 'metaphoricalassertion'asutteredby the speaker,the other its paraphrase,expressing'literally'whatthe speakermeant by the metaphor (J. Searle 7979a:
82). This preliminary formulation which Black had criticisedalmost twenty years
before, might not have been so unfortunatehad he not rushed on to claim that
"In each case the speaker'smetaphoricalassertion(MET) will be true if
and only if the correspondingassertionusingthe (PAR) sentenceis true."
(J. Searle 1979a:82).
An example of this would be:
(6) (MET) Richard is a gorilla.
(PAR) Richard is fierce, nastyand prone to violence.
Thus, unlike ordinary sentences,(6) (MET) does not have its truth-conditionsand
value specifiedby the meaningsof the elementsin it: "...thetruth-conditionsof the
assertionare not determinedby the truth conditionsof the sentenceand the general
term" (J. Searle 1979a:84-5).So, before we can decide on the truth value of (6)
(MET) we need to know first what the truth value of the corresponding (PAR)
sentenceis.
lraving asidethe question-ofthe particularexamplechosen,which considerably
facilitatesSearle'sformulatiort2, I discerntrvo confusionshere, one having to do
with the notion of truth and the other with that of assertion.Concerningthe first we
might as well insist once more that if notions like sentencemeaning and truth
conditionsare to be of any use,then there'sonly one meaningand one set of truth
conditions-barring ambiguity-that can be assignedto a metaphoricalutterance.By
assigningtwo different sets of truth conditions -one for the sentenceas taken
literally, and the other for its interpretation-Searleis clearlyreducingmetaphor to
polysemy,for if a sentencecan have two different sets of truth conditions,that
sentencehas two literal meaningsassignedto it by the rules of the languagel3.If
we assign truth values to speaker'smeaning, how are we to distinguishit from
sentencemeaning?
12Searle does mention at the end of his paper that there are metaphors which
differ from the
'simple'
(5) in that their interpretation is open-endeci.However, no explanation is given for how the
interpretation proceeds in such cases.
13In fact some theoreticians with an interactive frame of mind such as Mac Cormack
(1985) A
Cognitive Theory of Metaphor, M.I.T Press, have made a point concerning the production of
polysemy through metaphor: "Metaphors both depend on and produce polysemy" (...) "for unless a
word can be taken in more than one sense,metaphor is impossible'. But while this may well be true,
I don't think this is what Black had in mind when he talked about the cognitive content of
metaphorical expressions.
Methaphor; meaning and comprehension
59
As for the questionof what sort of thing a 'metaphoricalassertion'might be,
notice firstly, that it is speakersthat assert, not sentences.The intuition that
metaphoricalutterancesare often -though not always-used to convey beliefs is
correct.When we assert something,we represent ourselvesas believing that a
certainstate of affairs is the case. But there is no necessarylink between the
linguisticpropertiesof a sentencelike its grammaticaltype or its meaningand what
it is usedto convey.Therefore there is nothing odd with a speakerusing a certain
declarativesentencelike (a) or (5) above to communicatea belief or a very strong
commitmenton a propositionother than that literally expressedby her sentence.la
The hearer'staskwill be, one of constructing-rather than identifyrng- a proposition
whichhe might representthe speakeras believing,as usual,on the evidencehe'll
gatherfrom the contextas well as from the literal meaningof the words used.Still,
whentalkingof truth, we cannot make it apply to what has thus been 'metaphorically asserted',
for that is not public, it never leavesthe realm of representationsin
the speakeror hearer'smind, but it is the truth of the utterancesof sentencesthat
we are interestedin in semantics,not of thoughtsor other internal representations.
To repeat,when a speaker utters (4) or (5) above she does not communicate a
beliefin the state of affairs describedby the sentenceused.This is typically the case
with most tropes and with non-assertivespeechacts. What she is after is not that
thehearerrepresenther as believingthat Richard is a gorilla or that New York City
is a pressurecooker -for if she were, we would not be talking of metaphor but of
madness.Whatever the proposition(s) the hearer arrives at as the one(s) the
speakerintended -the assumptioncommunicatedin this way- cannot reach the
status,so to speak,of the literally expressedproposition,for in so far as they are
implicitlycommunicated,they are partly the responsibilityof the hearer to a greater
or lesserextent.15
I will concludeby looking at one last exampleof what I describedat the beginningasa very muddledstateof affairsconcerningtheoriesof metaphoricalmeaning.
In her book Metaphor lts Cognitive Force and Linguistic Stntcture,I(ittay takes up
Black'sclaim that metaphorsare the bearersof an irreducible cognitive content.
Relyingon a relationaltheory of meaning-which sometimesshe calls a "contextual
conceptionof meaning"(E,E Kittay 1987:17) -followingthe Saussureannotion of
languageas a system of interconnectedelements mutually dependent for their
to See Sperber & Wilson (1986, chapter 4),
'explicatures'
especially the section on
and
'implicatures'.
1t M. Bergmann (1982, Metaphorical assertions. Philosophical
Review 9I,229-245) and I.
lnewenberg (1975,Identiffing metaphors. Foundations of Langaage 12, 315-38) both deal with the
issueof metaphorical assertions and metaphorical truth. While Bergmann is right in that na person
who uses a sentence metaphorically does not use it to assert the proposition which is literally
expressedby the sentence. In the case of assertive metaphor we must distinguish sentence meaning
and speaker'smeaning" (p.234), but wrong when she says that truth and falsehood'Should be tied
to the assertion made, rather than to the sentence used" (p.238), l.oewenberg (1975) argues in a
completelyopposite direction: metaphorical utterances are not used to make truth claims, and if they
were, they would simply be false in most cases,and from this she is concludes, wrongly, that they
must not be taken as assertions and that 'Mets are statements without truth value' (p.338).
60
Begofia Wcente
definition- the double semanticcontentof the metaphoris this time capturedby her
distinctionbetweenfirst and second-ordermeaning-a functionof the former- which,
she claims,"cuts acrossthe semantic/pragmatic
divide" (E.F, Kittay 1987:35).Firstorder meaning is defined very much along the lines of Katz & Fodor's semantic
theory with their semanticfeatures,semanticcombinationrules and, in her case,
semantic field indicators. As for second-ordermeaning, there are several types,
depending on the function which operateson first-order meaning:indirect speech
acts,irony, etc. In her approach,speaker'sintentionsdo not have much of a role to
play, and the context becomes-via the expressibilityprinciple- "another utterance
of greater complexity of which the first (the metaphoricalone) is a constituent"
'locus of
(E.F, Kittay 1987 78) and thus, she catersfor thosecasesin which the
the
conceptual incongruity' necessaryfor an utterance to qualify as a metaphor is
located outside the sentence.When the semanticincongruity(conceptualanomaly
or semantic deviation) she finds necessaryfor metaphor is not located within the
sentence,there is typically "an implicit invocation" of a semantic field outside the
sentence which is integrated and the procedure of tension resolution that
understandinga metaphor requiresin her theory proceedsin the same manner as
for metaphorical utterances whose incongruity is found within the sent€nce.
Situationalcontextsare thus assimilatedto linguisticonesand any elementswe may
draw from the context in the interpretation of the metaphorical utterance are thus
rendered linguisticin nature and the explanationremainssemantic.This trick will
not do, for the key issue in the interpretation of an utterance is precisely the
inferring of relevant information from what is said in a certain context and what a
theory has to account for is preciselyhow this is done, what criteria guide both
speakerand hearer in this task. Once the selectionof the relevantinformation has
been made, the interpretationfollows automatically:but a lot more than semantic
mechanismshas been involved in the process.
Kittay's formulation raises some obvious questionsconcerning the motivation
as well as for the way it operates.What is
for her notion of second-order-meaning
the psychologicalstatus of these second-orderrules? They have been devised to
capture the intuition that it is through the restructuringof one conceptualdomain
by another, by the projectingof one implicativesystemonto another -as Black put
it- that utterers of this type of expressioncommunicate.But while it may well be
true that that is one of the many ways in which certain utterancescan achieve
communicativevalue -contextualeffects,or simplyrelevance-throughthe perceiving
and forcing of structural similaritiestypicallyunexploredand unperceivedprior to
the formulation of the utterancel6,one must not forget that more often than not
there is no widespreadagreementin the caseof highlypoetic metaphorsas to what
the correct interpretationsare, and one wonderswhy these second-orderrules do
not operate acrossspeakersto yield the samesortsof interpretations.Notice as well
that literary criticismsharessome characteristics
with ordinary linguisticexchanges,
but not all. Kittay & lrhrer (1981) offer a highly elaborate interpretation of
16See Verbrugge& MacCarrell(1977)(Metaphoriccomprehension:
studiesin remindingand
resembling".CognitivePsychologt9) on this.
Methapho4 meaning and comprehension
6l
Shelley'sBeesof England and other poems,over which, I'm sure, they spent a good
many hours, trying to establishall conceivableconnectionsand using all their
knowledgeof the authors, the literary conventionsof their time and a whole lot of
information that could bring out their point more clearly. Ordinary linguistic
communicationtakes place at a much more rushed pace,even if the utterancesthat
are being used are not being meant literally.
The real chiillengelies in finding the cognitivemechanismsthat tell metaphor
productionand interpretationapart from non-metaphoricalspeechonce we have
established
that the generalmechanismsof significationare the same for both -the
searchfor relevanceguidingthe makingmanifestof intentionsand their recognitionand that literal and nonliteral communication constitute the two ends of a
continuumof cases.Once we get rid of the assumptionthat literal languageis the
norm in linguisticcommunication and apply rigorously the distinction between
explicit and implicit communication, highlighting the importance of speaker's
intentionsover formalistic characterizations,we realize that the full import of a
metaphor or any other utterance, cannot be given in independence from a
communicativesituation and therefore from a specific communicative intention.
Interestingmetaphors typically -and here I'm following Sperber & Wilson (1986)make manifest an indefinite range of implicatures some very strongly conveyed,
others only very weakly so, and of course, not necessarilythe same in every
situation,or for every person. But here metaphorical utterances share with other
typesof less fascinatingUpes of expressionslike vague and loose utterances the
sameproperty of communicatingassumptionswith different degreesof strength or
speakerbacking.When we look for the rule for the perfect paraphrasethat'll give
us the key to the workings of metaphorical language,we are missing the point
altogether,for linguistic communication is only partly coded behaviour, and what
makes it such a powerful instrument for the conveying of information is the
possibility
of exploitinggeneralprinciplesof human cognitionlike the attribution of
purposeto the behaviourof others and relfng on others similarly trying to make
senseof our behaviouron the same grounds.
There are certainly many studiesof metaphoricalspeechwithin a pragmatic
perspective,
mainly along Gricean lines.The dangerhere lies in relying uncritically
on the assumption-a prejudice really- that literal languageand the search for truth
take precedenceover all other requirementsin verbal communication.But that
shouldbe the topic of another study.
62
Begofia Vicente
References
Berg,J. (1988) Metaphor,meaning,and interpretation.Joumalof Pragmatics12:695-78.
Black,M. (1962)Metaphor.ln Models& Metaphors.Ithaca,NewYork: CornellUniversityPress,2547.
Black, M. (1978/1979)How metaphorswork: A reply to D. Davidson.In Sacks,S. (Ed.) On
Metaphor Chicago:Universityof ChicagoPress,181-192
Black, M. (1979)More about metaphor.In A Ortony (ed.)Metaphor& Thought.C.U.P.,19-43.
Blakemore, D. (1987) SemanticConstraintson Relevance.Blaclavell,Oxford.
Carston,R. (1988)Implicature,explicature,andtruth-theoreticsemantics.
In Kempson(ed.)Mental
Representations:
The InterfacebetweenLanguageand Reality.C.U.P.
Cohen,L.J. (1979)The semanticsof metaphor.In A Ortony (ed.) Metaphorand Thought.C.U.P.,
186-201.
Davidson,D. (1980)What metaphorsmean.In M. Platts(eA.)Reference,
Tiuth& Reality.l-ondon:
Routledge& Kegan Paul,238-254.
Davies, M. (1982/3) Idiom and metaphor.hoceedingsof theAristotelianSociety83: 67-85.
Gibbs,R. (1987)What doesit meanto saythat a metaphorhasbeenunderstood?.In R.E. Haskell
(ed.) Cognitionand SymbolicStructures.Nonrood, NJ: Ablex PublishingCorporation.
Grice, P.H. (1975) l,ogic & conversation.In P. Cole & J. Morgan (eds.).Synrar& Semanrtcs3.
Speech,4cls.
New York: AcademicPress.
Kittay, E.E (1987) Metaphor: Its CognitiveForceand Stntcrure.Oxford: ClarendonPress.
Kittay, E.E & A lrhrer (1981)Semanticfieldsand the structureof metaphor.Sndiesin Language
5: 31-63.
lrvinson, S. (1983)hagmatics.C.U.P.
Manns,J.W. (1975) Metaphorand paraphrase.
Bitish JournalofAesthetics15(a):358-66.
Ortony, A. (1980) Some psycholinguistic
aspectsof metaphor.In R.P.Honeck & R.R. Hoffman
(eds.)Cognitionand FigurativeLanguage.l,awrenceErlbaumAssociates.
Searle,J. (1979a)Metaphor.In J. Searle,Expression
& Meaning.C.U.P.,76-116.
Searle,J. (1979b)Literal Meaning.In J. Searle,Expression
& Meaning.C.U.P.,
Ll7-136.
Sperber,d. & Wilson, D. (1986)Relevance:
Communication
and Cognition.Oxford: Blackwell.
Warner,M. (1973) Black'smetaphors.Bitish Joumalof Aesthetics13(4):367-372.
Wilson, D. & Sperber, D. (1981) On Grice's theory of conversation.In PN. Werth (ed.)
Conversation& Discourse.l,ondon: Croom Helm.