Available via license: CC BY 4.0
Content may be subject to copyright.
STUDIA SEMIOTYCZNE, t. XXXIII, nr 1 (2019), s. 105–134
ISSN 0137-6608, e-ISSN 2544-073X
DOI: 10.26333/sts.xxxiii1.07
A rt i cl e
RICHMOND KWESI *
SEMANTIC MEANING AND CONTENT: THE INTRACTA-
BILITY OF METAPHOR
SU M M A R Y : Davidson argues that metaphorical sentences express no propositional
contents other than the explicit literal contents they express. He offers a causal account,
on the one hand, as an explanation of the supposed additional content of a metaphor in
terms of the effects metaphors have on hearers, and on the other hand, as a reason for the
non-propositional nature of the “something more” that a metaphor is alleged to mean.
Davidson’s account is meant to restrict the semantic notions of meaning, content, and
truth, to literal sentences. I argue that the Davidsonian causal account does not satisfacto-
rily account for metaphor’s figurativeness, speakers’ assertion and hearers’ uptake of
metaphor, and our discursive practices of using metaphors in disagreements and argumen-
tation. I offer a non-compositional analysis of a semantic account of metaphor within
which one can make sense of the applicability of the notions of meaning and content to
metaphor. This analysis shows that metaphorical sentences have meanings other than, and
in addition to, their literal meanings and what speakers can use them to mean.
KE Y W O R D S : metaphor, compositionality, Davidson, meaning, content, causal account.
1. INTRODUCTION
Semanticists who have worked on the semantic notions of meaning, content,
truth and assertion have had to grapple with the phenomenon of figurative lan-
guage in general and metaphor in particular. Some of them have engaged with
the following questions: Do metaphorical sentences have contents other than, or
in addition to, the literal contents they express? If metaphorical sentences have
* University of Ghana, Department of Philosophy and Classics. E-mail:
rkwesi@ug.edu.gh. ORCID: 0000-0001-5286-1935.
106
RICHMOND KWESI
any such non-literal contents, are these contents propositional in nature? How do
metaphorical sentences come to have such additional (propositional) contents?
And, how do users of metaphors—speakers and hearers—associate, derive, or
capture, these additional contents? An adequate semantic theory or a semantic
account of our linguistic practices has to provide an answer to the above ques-
tions by either explaining away and dispensing with the issue of metaphor, that it
is not amenable to the semantic notions properly understood, or that the semantic
notions are not appropriately applicable to the phenomenon of metaphor, or that
the senses in which the semantic notions are applicable to metaphor are different
from those in which they are applicable to ordinary literal uses of language.
Let us consider an intuitive story of how the notions of meaning, content, and
truth are indispensable to thinking about metaphor. In the sentence, “Gabriele is
a crocodile—he is impulsive and angry, he changes like the wind”, the first part
of the sentence—“Gabriele is a crocodile”—is a metaphor;
1
and the second
part—“he is impulsive and angry, he changes like the wind”—is considered
variously as the content, meaning, interpretation, or paraphrase of the content of
the metaphor. The metaphor is about a subject and a predication of a property to
the subject, the predicate presents us with a description of the subject—it is an
attribution of a particular property to the referent of the subject term of the meta-
phor. The metaphorical statement has a semantic value—it is true or false if the
subject “fits” the description, or if it is the way in which it is being described.
Taken literally then, the description is false (on the assumption that Gabriele is
human and not a crocodile) but since the statement is a metaphor (or is being
used as a metaphor), it is the second part of the remark—the paraphrase—which
gives the interpretation of the property being attributed to Gabriele in literal
terms that confers truth (or falsity) on the metaphor. That is, the metaphorical
statement has truth value, and the truth value is derived from, and dependent
upon, the truth or falsity of the corresponding interpretation or paraphrase of the
metaphor. In this example then, the description of Gabriele as a crocodile is true
if it is true that Gabriele is impulsive and angry and changes like the wind. This
intuitive story is unpersuasive to Donald Davidson and other theorists who are
sympathetic to his account of metaphorical meaning and content.
I discuss in this paper Davidson’s treatment of the meaning and content of
metaphorical sentences. Davidson maintains that metaphorical sentences express
no propositional contents other than the explicit literal contents they express. He
offers a causal account, on the one hand, as an explanation of the supposed addi-
tional content of a metaphor in terms of the effects metaphors have on hearers,
and on the other hand, as a reason for the non-propositional nature of the “some-
thing more” that a metaphor is alleged to mean. In the analysis of the Davidsoni-
an position, I will argue that what metaphors cause us to do, and the effects they
1
Another way of talking about the sentence is to say that it is being used as a meta-
phor. The sentence itself could also be another figure of speech like irony or overstate-
ment, or simply an insult, although both the linguistic and the non-linguistic contexts will
disambiguate the particular use.
SEMANTIC MEANING AND CONTENT…
107
have on us, does not preclude their having contents that can be propositional in
nature. I will point out two primary defects of the causal account proposed by
Davidson: one, it combines the theses of literalism and compositionality to the
analysis of metaphor and in so doing mischaracterizes metaphors as having only
literal meanings;
2
and two, it presents a one-sided perspective on the use of met-
aphors in terms of the effects they have on hearers, and, thereby, fails to appreci-
ate the value in the making of metaphors when hearers become speakers. The use
of metaphors in disagreements, deductive and inductive arguments, and the pos-
sibility of retracting metaphorical utterances, making of inferences from meta-
phors—all these practices establish one crucial thing: contra Davidson, meta-
phors could have contents that are propositional in nature.
2. AGAINST PROPOSITIONAL CONTENTS OF METAPHORS
The causal theorist (Davidson, 1979; Rorty, 1979; 1987; 1989; Reimer, 2001;
Lepore & Stone, 2010) is motivated to restrict the semantic notions of meaning
and truth to the more familiar literal uses of language. She is averse to both re-
vising her ontological commitments, and broadening the use of truth and mean-
ing, to include, or apply to, metaphorical and other figurative uses of language.
Literal uses of language can be evaluated for truth partly because there are gen-
erally accepted ways for fixing the contents and propositions expressed by literal
sentences (or utterances), and partly because, unlike in the case of metaphorical
sentences, literal truth conditions, usually, can be assigned to sentences irrespec-
tive of the particular contexts in which they are used. Every metaphorical claim
or sentence, when construed literally, has a literal content or expresses a literal
proposition. The causal theorist is of the view that the literal content or the prop-
osition the metaphor literally expresses is the only content possessed or proposi-
tion expressed by a metaphor; the non-literal aspect of a metaphor is nothing
propositional. This view implies that metaphors do not have propositional con-
tents in addition to their literal contents, and hence, metaphorical sentences qua
metaphors cannot be truth-evaluable. This presents a bit of a puzzle: on the one
hand, in virtue of being a metaphor, a metaphorical sentence is meaningful and
has a non-literal content, and yet the metaphor itself is non-truth-evaluable; and
2
I have pointed out in a previous paper (Kwesi, 2018a) that the theses of literalism
and compositionality are often linked with another thesis, representationalism, to support
the view that the relevant criterion of truth is the capacity to represent states of affairs as
they really are. For instance, Cooper, a defender of Davidson’s view, has made these
remarks: “The notion of truth, as we normally understand it, is used to appraise utterances
in terms of what they achieve. A true statement is one which successfully achieves what
statements generally aim to achieve—telling how things really are. To employ the notion
of truth in the appraisal of metaphor, therefore, wrongly suggests that metaphors, too,
have the dominant aim of getting us to see how things actually are” (1986, p. 250).
I argued there that the combined theses of literalism and representationalism do not sup-
port the view that metaphors are not truth-evaluable.
108
RICHMOND KWESI
on the other hand, a metaphor has or expresses only a literal proposition which
makes the metaphor either literally true or false, and thereby, making the meta-
phor truth-evaluable. The causal theorist argues for the second part of the puzzle
by showing that
i. the words of a metaphor have only literal meanings, and, therefore, the
metaphorical sentences they compose only have literal meanings; and
ii. in light of (i), metaphorical sentences have literal truth conditions which
makes most, if not all, metaphors patently or literally false.
For the first part of the puzzle, she supports her position by arguing that
iii. the supposed additional non-literal meaning or content of a metaphor is not
propositional in nature;
iv. this non-propositional meaning of the metaphor is merely the effects meta-
phors have on their recipients; and
v. a metaphorical sentence does not have a single definite content or meaning;
rather, it has many, and perhaps, an infinite number of contents.
Davidson’s anti-truth account of metaphor gives expression to the tenets (i)—
(v) above. Davidson’s (1979) main claim (as he himself calls the “thesis” of his
paper) is that “metaphor means what the words, in their most literal interpreta-
tion mean, and nothing more” (p. 30). This thesis is borne out of a commitment
to two views about language: literalism and compositionality. Davidson’s literal-
ism acknowledges a distinction between the literal and metaphorical uses of
language, but claims that sentences can only have ordinary literal meaning and
truth, and that a distinction between the literal and the metaphorical does not
entail that metaphorical sentences have “special” meaning and truth i n add i-
ti on to their literal senses and truth. What metaphors mean, and what their truth
values are, are no different from their assessment from a literal point of view. In
his commitment to compositionality, Davidson is of the view that the meaning of
a sentence is determined from the meanings of the individual words that com-
pose it. If a metaphor can only be explained by appealing to the literal meanings
of the words that compose it, then for Davidson “sentences in which metaphors
occur are true or false in a normal, literal way, for if the words in them don’t
have special meanings, sentences don’t have special truth” (p. 39). Combining
his literalist and compositionalist views, Davidson’s claim is that the words of
a metaphorical sentence have no special meanings other than their ordinary lit-
eral meanings, and hence the sentences they compose only have literal meanings.
In view of the fact that metaphorical sentences only have literal meanings and
literal truth conditions, metaphorical sentences have no contents except the con-
tents that they literally express. This is why most metaphors are literally false, if
not absurd. That metaphors have no contents (except what they literally express)
SEMANTIC MEANING AND CONTENT…
109
implies that there is n othi ng else that is communicated or conveyed by the use
of metaphor, nothing else propositional that can be grasped and evaluated as true
or false. Davidson entreats us to give up “the idea that a metaphor carries a mes-
sage, that it has a content or meaning (except, of course, its literal meaning)” and
see the supposed content of metaphor as “something about the effects metaphors
have on us” (p. 43). A metaphor can provoke thoughts and ideas in us, it can
make us attend to some likeness and similarities between two things, it can cause
us to notice something in a different way, but all these are effects metaphors have
on us: metaphors “make us appreciate some fact—but not by standing for, or
expressing, the fact” (p. 44). Davidson’s denial of the cognitive claims of meta-
phor presents us with an account of metaphor that is causal in nature: it is a caus-
al account in the sense that it explains metaphor both in terms of what it cau se s
us to do and the eff ect s it has on us. In this cause-effect view, metaphor has no
content other than what it literally means and expresses, which is usually false or
absurd; if we mistakenly think that there is an additional f ig urat iv e or m et-
ap ho ric al content to a metaphor, it is merely because we are confusing
effect with content. What metaphor directs our attention to, what it makes us see,
cannot be propositional in character; for as Davidson exclaims: “seeing as is not
seeing that” (p. 45). In this regard, Davidson likens metaphor to a joke or
a dream or “a bump on the head”—these acts have effects on us by making us
come to notice or observe some fact without their expressing those facts. Meta-
phors can lead one to s ee something as, but not tha t; they can inti ma te,
nudge, or poke one to view something in a different way, but intimation is not
the same as meaning; they can ca us e one to have certain beliefs, but they do
not express those beliefs;
3
like jokes and bumps on the head, they can have ef-
fects on others, but such effects are not propositional elements that can be evalu-
ated on the basis of semantical notions like meaning, truth and reference.
Davidson argues also that our inability to paraphrase or decide exactly what
the content of a metaphor is, is not primarily because metaphors are non-
paraphraseable, but because there is no content to be paraphrased or expressed.
He thinks that we imagine there is a content to be captured when all the while we
are in fact focusing on what the metaphor makes us notice; we are merely focus-
ing on the effects metaphor has on us. He writes:
If what the metaphor makes us notice were finite in scope and propositional in na-
ture, this would not in itself make trouble; we would simply project the content
the metaphor brought to mind onto the metaphor. But in fact, there is no limit to
what a metaphor calls to our attention, and much of what we are caused to notice
3
By causing us to form certain beliefs that such-and-such is the case, there is some-
thing “propositional” about metaphor, that is, the acquisition of propositional attitudes.
But the Davidsonian contention is that the metaphorical sentence itself does not express
the proposition that such-and-such is the case; the metaphor itself does not make a state-
ment or communicate something that is propositional.
110
RICHMOND KWESI
is not propositional in character. When we try to say what a metaphor “means”,
we soon realize there is no end to what we want to mention. (1979, p. 44)
Davidson’s line of thought has been expanded by Lepore & Stone (2010) in
their thesis statement that “though metaphors can issue in distinctive cognitive
and discourse effects, they do so without issuing in metaphorical meaning and
truth, and so, without metaphorical communication” (p. 166). Like Davidson,
they take a pragmatic view of metaphor as involving some sort of speaker inten-
tions and not co mm un ic at ed mea ni ng . They argue that metaphor should
be catalogued among practices such as “hinting, joking, trash-talking, flirting,
and flattering” (p. 166). By joking, one aims to cause certain effects in one’s
audience rather than to assert something that can be appraised for truth. And
metaphor is no different from jokes. An interlocutor may use a metaphor with the
intention that his hearers see a particular point but this point “is not a property of
the metaphor itself” (p. 173). Lepore and Stone contend that “interlocutors use
their metaphorical discourse not to assert and deny propositions, but to develop
imagery and to pursue a shared understanding” and that “such practices can
account for our interactions in using metaphor, without appealing to metaphori-
cal meaning or metaphorical truth” (p. 177). In effect, they argue for distinguish-
ing “metaphorical thinking—developing imagery, seeing one thing as another,
noticing similarities—from merely grasping a proposition, namely the one that is
speaker meant, brought about through an intention to present information
through coordination or intention recognition” (p. 178).
For Richard Rorty, Davidson’s causal account enables us to see the distinc-
tion between the literal and the metaphorical not as two sorts of meaning or truth
but a distinction between “familiar and unfamiliar uses of noises and marks”
(1989, p. 17). The literal is the regular and familiar uses of language that are
marked by predictability and a generally accepted procedure for determining
meaning and truth. The metaphorical, Rorty thinks, is an unfamiliar noise—a use
of familiar words in unfamiliar ways. As an unfamiliar noise, metaphor has no
fixed place in the language game. Uttering a metaphorical sentence is not to say
something true or false; it is not to say something that has a meaning. Rather,
uttering a metaphor only produces an effect in one’s audience and causes them to
have certain beliefs or act in certain ways. In one characterization of metaphor,
Rorty has this to say:
Tossing a metaphor into a conversation is like suddenly breaking off the conversa-
tion long enough to make a face, or pulling a photograph out of your pocket and
displaying it, or pointing at a feature of the surroundings, or slapping your inter-
locutor’s face, or kissing him. Tossing a metaphor into a text is like using italics,
or illustrations, or odd punctuation or formats. All these are ways of producing ef-
fects on your interlocutor or your reader, but not ways of conveying a message. To
none of these is it appropriate to respond with “What exactly are you trying to
say?” If one had wanted to say something—if one had wanted to utter a sentence
with a meaning—one would presumably have done so. (1989, p. 18, italics mine)
SEMANTIC MEANING AND CONTENT…
111
In another characterization, Rorty likens metaphor to thunderclaps and bird-
song to make the same point. A novel metaphor is like the noises of a bird we are
not acquainted with. The noise c au ses us to believe that there is, for instance,
a quetzal in the forest. But the noise itself does not convey the information nor
express the fact that there is a quetzal in the forest. In the same way, a metaphor
causes us to change our beliefs and desires without representing or expressing
any facts of the world. According to Rorty, we should see metaphor in its func-
tions as
causes of our ability to do lots of other things—e.g., be more sophisticated and in-
teresting people, emancipate ourselves from tradition, transvaluate our values,
gain or lose religious faith—without having to interpret these latter abilities as
functions of increased cognitive ability. (1987, p. 284–285)
Rorty, therefore, allows metaphors to have functions, that is, to be causes of
beliefs, just as Davidson endows metaphor with the ability to direct our attention
to notice similarities between things. Yet, these functions of metaphor are not to
be interpreted as conveying any message that will add to our knowledge.
Both Rorty and Davidson rely on a distinction between “cause of belief” and
“justification of belief” (or “reason for belief”) and argue that it is a conflation of
this distinction that seems to give some credence to the cognitive claims of met-
aphor. As it pertains in sensory observations of birdsong and other unfamiliar
noises, we can draw a distinction between the unfamiliar noise as a stimulus to
knowledge and the claim that it conveyed that knowledge. The noise is merely
a stimulus to knowledge or a cause of the belief that there is a bird in the forest,
but it is not a reason for, nor a justification of, the belief or information that there
is a bird in the forest. What causes belief and knowledge is not necessarily that
which expresses or conveys belief and knowledge. Metaphor as an unfamiliar
noise belongs not to cognition but to stimulus. It has a place in a causal scheme
of things, but it does not have in addition a place in a pattern of justification of
beliefs. By confining the interpretation and meaning of metaphor to the literal
and explaining away the supposed additional content of a metaphor in terms of
the effects metaphors have on us, Davidson, Rorty, Lepore and Stone, and others,
limit the semantic notions of truth, meaning and content to regular and literal
uses of language.
3. ANALYSIS AND CRITIQUE OF THE CAUSAL ACCOUNT OF METAPHORICAL
CONTENT
Davidson’s account of metaphor has been discussed extensively in the litera-
ture, mostly in two main directions: there are those who criticize his literalist
account and argue for the cognitive claims of metaphor, especially Black (1979;
1993), Goodman (1979), Leddy (1983), Hesse (1987; 1988), Farrell (1987),
Moran (1989), Camp (2006a; 2006c; 2008), Johnson (2008), and most works in
cognitive linguistics; and there are others who have defended his account, partic-
112
RICHMOND KWESI
ularly, Davies (1982), Davies (1984), Rorty (1987; 1989), Cooper (1986),
Reimer (2001), Lepore & Stone (2010; 2015). The critique of the Davidsonian
account in the literature has primarily focused on showing that there is a cogni-
tive dimension (Black, 1979) or a propositional dimension (Moran, 1989; Camp,
2006a; 2006b; 2008) to metaphor. This propositional aspect of a metaphor is
usually derived from, or associated with, the intentions of the speaker—what the
speaker means by uttering a metaphor (Searle, 1979).
The critique of the account I offer here is partly diagnostic, intended to reveal
the ways in which the Davidsonian tenets (i)—(v) above are flawed and untena-
ble; and it is partly prescriptive, meant to provide evidence that metaphorical
sentences have propositional contents. The analysis pursued here is to show how
Davidson’s account does not adequately and satisfactorily deal with the issue of
the meaning and content of metaphor. It is often regarded as implied by Da-
vidson’s account that once one accepts his central thesis that a metaphor has no
additional meaning and truth-value other than its literal meaning and truth-value
then one is committed to seeing metaphor only in terms of its functions—in
terms of its causes and effects. However, the inference from literalism—“only
literal meaning”—to a causal explanation—“only causal role”—is not a logically
necessary one. It is possible to accept Davidson’s central thesis without adducing
a causal explanation for how metaphor works, and more importantly, without
singling out a causal explanation as the only explanation one could give to meta-
phor. Similarly, the conclusion that metaphors have no propositional contents
cannot be premised on the fact that metaphors have causes and effects on their
users. It is possible to accept a causal explanation of how metaphors work—that
is, that they cause us to acquire certain beliefs, that they direct our attention to
see similarities between two things, etc.—and posit that they have propositional
contents in addition to their causal role. In other words, that metaphors have
causes and effects does not preclude their having propositional contents. Meta-
phors do have functions, they do cause us to do certain things, they have effects
on us; but their having functions and effects is not a reason for, nor a limitation
of, their capacity to be something else, or have something more—something
propositional.
3.1. Metaphor and Compositionality
Davidson’s ultimate position on metaphor is that metaphorical sentences have
only literal meanings and hence only literal truth conditions. This position is as
a result of combining his thesis of literalism—that the words of a metaphor have
only literal meanings—with compositionality—that the meaning of a sentence is
derived from the meanings of the individual words that compose it. That is, if the
words of a metaphor have only literal meanings then metaphorical sentences
have only literal meanings. However, this analysis is flawed: the mistake lies in
the conjunction of the two theses—literalism plus compositionality—to generate
the solution that metaphorical sentences have only literal meanings:
SEMANTIC MEANING AND CONTENT…
113
Literalism (L) + Compositionality (C) = Literal Meaning (LM)
To see the flaw, we have to take a critical look at the two principles of literal-
ism and compositionality. Compositionality is regarded as one of the essential
properties of language which is used to explain, among other things, our linguis-
tic and cognitive abilities to learn a language by learning the meaning of a finite
number of expressions and yet be able to produce and understand an infinite
number of meaningful sentences (Davidson, 1967, 1984; Fodor & Lepore, 2002;
Pagin & Westerstahl, 2010a; 2010b). On one definition, the principle of compo-
sitionality is the claim that “the meaning of a complex expression is determined
by its structure and the meanings of its constituents” (Szabo, 2010, p. 255). This
determination of the meaning of the complex expression is usually construed in
functional terms; that is, “the meaning of the complex expression is a function of
the meanings of its parts and the mode of composition by which it has been ob-
tained from these parts” (Kracht, 2011, p. 57). Compositionality is a semantic
phenomenon, for it determines the semantic value of a complex from the values
of its constituents, thereby constraining the relevant factors involved in the de-
termination of meaning. In a strict sense of compositionality, what is necessary
and sufficient for determining the semantic value (meaning, content, denotation)
of a complex expression is the semantic information and contribution derived
from the parts of the complex expression and its mode of composition. This is
akin to what Dever (2008) has called the “semantic closure” constraint of the
principle of compositionality.
However, it has been questioned in various ways in the literature as to wheth-
er the meaning or content of a complex expression is determined purely from the
semantic values of, or the semantic information provided by, its constituents and
their mode of composition. This questioning arises out of the observation that the
meaning or content of an expression is underdetermined by the semantic infor-
mation provided by the parts of the expression, and that, there are certain con-
stituents of the meaning of an expression that are provided purely on pragmatic
grounds, usually by a process of “free pragmatic enrichment” (Carston, 1988;
2002; Recanati, 2004; Sperber & Wilson, 1995; Hall, 2009). The utterances of
“I have had breakfast” and “It is raining” have their truth-evaluable contents
<I have had breakfast this morning> and <It is raining in Cape Town> respec-
tively, where the time (of breakfast) and the location (of rain) are freely pragmat-
ically supplied by the context of the utterance. These additional constituents of
the meaning or content of an expression are not traceable overtly or covertly to
the encoded meanings of the parts of the expression; they are provided and con-
strained by purely pragmatic factors.
4
Generalizing from this observation, con-
4
Some semanticists like Stanley (2000; 2002), and King & Stanley (2005) have ar-
gued that indexicality and other contextual factors can be traced to the logical form of the
expressions which suggests that the so-called free pragmatic enrichments are constrained
semantically. Lasersohn (2012) has argued that the context-sensitivity nature of most
expressions and the fact that speakers rely on pragmatics to arrive at the contents of cer-
114
RICHMOND KWESI
textualists and pragmatists argue that the intuitive meaning and content of an
expression cannot be given solely by a compositional semantics.
Compositionality itself as a principle for the determination of the semantic
value of a complex expression does not discriminate between literal and non-
literal meaning even though it seems to presuppose literal meaning. All that is
required for compositionality is that the meaning of the complex be a function of
the meanings of its parts and their mode of composition. This does not imply that
the kind of meaning
5
of the complex be determined from a function of its parts
and the ways in which they are composed. That is, compositionality does not
specify or stipulate that the meaning-type of the complex be derived from those
of its units. The requirement that the meanings of complex expressions be literal
because their units are literal is an additional constraint on the meaning of com-
plex expressions. Up to this point we had noted that compositionality requires
that the meaning of the complex expression is a function of both
(a) the meanings of its constituents, and
(b) their mode of composition.
Now, there is a further constraint on the meaning of complex expressions or
sentences in general in relation to the parts that compose them:
(c) the kind of meaning (or the meaning-type) of the complex expression is
a function of the meaning-types of its constituents.
This additional constraint (c) is what informs the literalism of Davidson’s ac-
count of metaphor.
We should distinguish between two strands of literalism: word-literalism (Lw)
and sentence-literalism (Ls). Davidson actually argued for word-literalism, indi-
cating that words themselves do not have “extra” or non-literal meanings, and by
extension, he argued for sentence-literalism through a compositional analysis.
The fact that Davidson argued for word-literalism has been observed also by
Farrell (1987) who shows that in his essay on metaphor, Davidson hardly treats
metaphor at the sentential level; rather, he treats it at the level of word meaning.
At the beginning of his paper where he states his main thesis, Davidson writes:
“This paper is concerned with what metaphors mean, and its thesis is that meta-
phors mean what the words, in their most literal interpretation, mean, and noth-
ing more” (1979, pp. 29–30, emphasis mine). He again writes that “my disa-
greement is with the explanation of how metaphor works its wonders. To antici-
pate: I depend on the distinction between what words mean and what they are
tain expressions do not undermine, but are compatible with, the principle of composition-
ality.
5
By “kinds of meaning” I mean meaning-types such as literal meaning, metaphorical
meaning, symbolic meaning, etc.
SEMANTIC MEANING AND CONTENT…
115
used to do” (p. 31). And when he discusses and rejects other views on metaphor,
Davidson has these things to say: “The idea, then, is that in metaphor certain
words take on new, or what are often called “extended” meanings” (p. 32); “Per-
haps, then, we can explain metaphor as a kind of ambiguity: in the context of
a metaphor, certain words have either a new or an original meaning” (p. 32);
“I have been making the point by contrasting learning a new use for an old word
with using a word already understood” (p. 37). And after discussing a number of
views, Davidson concludes:
The argument so far has led to the conclusion that as much of metaphor as can be
explained in terms of meaning may, and indeed must, be explained by appeal to
the literal meanings of words. A consequence is that the sentences in which meta-
phors occur are true or false in a normal, literal way, for if the words in them don’t
have special meanings, sentences don’t have special truth. (1979, p. 39)
As Farrell points out, Davidson’s focus on the word instead of the sentence
“serves his strategy in the article” for “he interprets his opponents to be making
a claim that metaphorical meanings constitute an extra layer of word meanings,
and consequently, that metaphor is analogous to ambiguity, in that a word may
have two different meanings” (1987, p. 637).
6
So, Davidson inveighs against
positing additional m et aph or ica l or f ig urat iv e meanings to the words that
compose a metaphor and then argues that since the words do not have extra
meanings other than what they mean literally, metaphorical sentences only have
literal meanings. In doing so, Davidson rescues semantics from accounts based
on extended word-meanings and also from the multiplicity of meaning and truth
with respect to the words in a metaphor. Now, so long as we are dealing with the
words of a metaphorical sentence, Davidson’s account seems plausible.
A point of departure with Davidson here is that both his attack on the theories,
and the theories he was attacking, miss one crucial point about metaphor: a met-
aphor is not necessarily a metaphor because a word has been used metaphorically
or in an unfamiliar way. It is only when we take the word, be it the focal word of
the metaphorical sentence, as the unit of analysis that we worry as to whether the
word has an “extended” meaning or reference. Indeed, words in every sentence
have no “extra” meanings other than what they mean literally, but their composi-
tion into sentences marks an important difference between figurative and literal
sentences. That is, word-literalism does not imply sentence-literalism when the
expressions in question have been construed metaphorically or figuratively. It is
one thing to say that the words in a metaphor only have literal meanings and
6
In Davidson’s general theory of meaning, however, word and sentence meaning go
hand in hand. For him, other than a semantic theory being compositional, it must also be
interpretative, in the sense that it should be possible for the theory to be used to under-
stand speakers and their linguistic behaviour. Since a theory of meaning is a theory of
truth for Davidson, one constructs a systematic truth theory from both the meanings of the
words and sentences of a language.
116
RICHMOND KWESI
another thing to say that the metaphorical sentence has only a literal meaning or
interpretation. One can endorse the claim that the words of a metaphor have no
special, extra, non-literal meanings without further endorsing the claim that the
metaphorical sentences composed out of the individual words have literal mean-
ings. Idiomatic expressions are paradigmatic cases of counter-examples not only
to compositionality in general but more particularly to the constraint (c) on the
meanings of complex expressions or sentences in general which requires that the
meaning-type of the constituents transfer to the meaning-type of the complex
expression.
Idioms are generally considered to be expressions whose meanings are con-
ventionalized in the sense that “their meaning or use can’t be predicted, or at
least entirely predicted, on the basis of a knowledge of the independent conven-
tions that determine the use of their constituents when they appear in isolation
from one another” (Nunberg, Sag, & Wasow, 1994, p. 492). An idiomatic expres-
sion defies the principle of compositionality in that the meaning of the idiomatic
expression is not determined by a compositional function of the meanings of its
constituents (Chomsky, 1965; 1980; Katz, 1973; Kracht, 2011).
7
The meanings
of idiomatic expressions like “kick the bucket” and “take the bull by the horns”
are not determined by the meanings of their component parts despite their having
syntactic structures. Interestingly also, the words of these idiomatic expressions
do not acquire extra meanings other than their literal meanings, but the idiomatic
meanings of the expressions are not dependent on the literal meanings of the
words even where their composition into a whole fails. That is, in spite of the
fact that the parts do not compose into a whole in determining their idiomatic
meanings, the idiomaticity of the expressions is not a function of the idiomaticity
of the words that make them up.
8
In other words, if the constraint of the meaning
of complex expressions (c) holds, then when the expressions are given idiomatic
meanings this should result from the constituent words having idiomatic mean-
ings (just as when they are interpreted literally, the words should have their lit-
eral meanings at play). But although the expressions have idiomatic meanings
their constituent words do not acquire any extra meanings other than their literal
7
Nunberg, Sag, & Wasow (1994) have shown that not all idiomatic expressions are
non-compositional. They distinguish idiomatically combining expressions like “pull
strings” whose meanings could be distributed among its parts, from idiomatic phrases like
“kick the bucket” which do not distribute their meanings to their parts. My concern in the
main is with idiomatic phrases.
8
One could point out that this is so because idioms are lexicalized expressions that
should be treated as single words. However, this view is unattractive. As Titone and Con-
nine point out, “there is evidence showing that idioms possess a great deal of internal
sematic structure. Idioms are modifiable with adjectives or relative clauses (e.g. “She did
not spill any of those precious beans”), and parts of idioms may be quantified (e.g., “She
didn’t spill a single bean”), emphasized through topicalization (e.g., “She didn’t spill the
beans yesterday, but spilled them today”) without disrupting comprehension or awareness
of their idiomaticity” (Titone & Connine, 1999, p. 1659).
SEMANTIC MEANING AND CONTENT…
117
meanings, and hence, the constraint on the meaning of complex expressions (c)
cannot be accurate.
The point here is that we can allow that the words that compose any figura-
tive expression maintain their literal meanings, but this concession does not
imply that figurative expressions only have literal meanings. This is because the
analysis and interpretation of a figure of speech like an idiom starts rather at the
phrasal or sentential level. An idiom obviously is different from a metaphor—
a metaphorical expression can be live and novel in characterizing one thing in
terms of another thing, while an idiom is a set phrase whose meaning cannot be
inferred from the meanings of the words that make it up, and whose usage is
characteristic of a group of people. A significant difference between a metaphor
and an idiom is that unlike an idiom, an understanding of the literal meanings of
the words in a metaphor aids in the interpretation of the metaphor. However, the
analysis and interpretation of metaphor takes a cue from idioms: a metaphorical
sentence can have a meaning, a meaning other than what it literally means even
though the words that compose the metaphor as Davidson has strongly argued
only have literal meanings.
9
How does this cash out?
In any context of use, both metaphorical sentences and idiomatic expressions
can be understood and interpreted literally. Compare the idiom “she kicked the
bucket” to the metaphor “Gabriele is a crocodile.” When we combine the literal-
ist thesis with compositionality “she kicked the bucket” just means that she
kicked the bucket, and similarly, “Gabriele is a crocodile” means that Gabriele is
a crocodile. Construed figuratively, it seems okay to say that “she kicked the
bucket” means that she is dead. Or perhaps, we should say that in an appropriate
context, one utters “she kicked the bucket” to mean that she is dead. “That she is
dead” becomes the content or the proposition asserted by the idiom-user. (This is
quite different from the effect the idiom might have on an audience, if any.) If the
sentence “she kicked the bucket” could mean both she kicked the bucket and she
is dead then we can say that the sentence has two meanings depending on the use
to which it is put: used literally, it has the meaning (LM) that she kicked the
bucket, and used figuratively (as an idiomatic expression), it has the meaning
(MM) that she is dead. The difference between LM and MM lies in the role
9
It is possible for one to argue that idiomatic expressions are not necessarily breaches
of compositionality, and that the cases of idioms neither affect nor make compositionality
false. The rule of compositionality is meant to apply to non-idiomatic uses of language.
This argument seems right. But the point here is not that the rule of compositionality is
breached or made false by idiomatic expressions. The point rather is that idiomatic ex-
pressions, in being figurative expressions, do not require that their meanings imply that
the words that make up the expressions also acquire figurative meanings. This suggests
that the meaning of a figure of a speech does not imply that the words of that figure of
speech have figurative meanings or applications. So, in the case of idioms, their meanings
do not depend on the meanings of the words that compose them, and there is no further
requirement that the words should have figurative meanings. This is the principle that
I am claiming holds in the case of metaphors.
118
RICHMOND KWESI
compositionality plays in the determination of the meaning of the sentence:
whereas LM results from the meanings of the individual words of the sentence,
MM does not; MM is not worked out from the meanings of the individual words
of the figurative expression.
A similar situation is what obtains in metaphor. In terms of LM, “Gabriele is
a crocodile” means that Gabriele is a crocodile, which might seem false or ab-
surd. But understood figuratively, it can have the MM meaning that Gabriele is
impulsive and angry. What is interesting about the metaphorical case is that the
MM meaning, while it does not result out of the composite of the literal mean-
ings of the words (for then we will have LM), is linked in a peculiar way to the
words of the sentence, not in terms of literal meaning, but usually, in terms of
certain cultural and idiosyncratic features or connotations associated with the
words of the metaphor. These cultural, religious, moral, aesthetic and idiosyn-
cratic features we associate with certain words and phrases are, in an important
sense, not part of what we will ordinarily call the literal meaning. It is not part of
the literal meaning of Gabriele being a crocodile that he is impulsive and angry.
This suggests that a determination of the meaning of the metaphor from a com-
posite of the literal meanings of the words will mischaracterize the expression as
a metaphor. In both the idiomatic and metaphorical cases, the words of the ex-
pressions retain their usual literal meanings, but a compositional determination
of their meanings misses the point of their figurativeness; that is, apprehending
their LM meanings is just to take the expressions literally.
The idiom, “she kicked the bucket” means she is dead, period. However, if
we understand both metaphors and idioms as figurative devices which defy the
laws of compositionality and constraint (c) resulting in metaphorical and idio-
matic expressions having MM as characterized above, then it seems that we
cannot think of the MM of a metaphor as merely a paraphrase (or effect or in-
sight) and that of an idiomatic expression as a meaning or proposition. It is true
that there could be more than one interpretation or meaning we could come up
with for a metaphorical sentence; but this will not yield different kinds of mean-
ings. There are only the two kinds of meanings here—LM and MM—depending
on whether the meanings are calculated based on compositionality or not. Just as
a literal sentence could be ambiguous or have multiple meanings under LM, so
a metaphorical sentence could have a variety of meanings under MM; the various
meanings under MM are all possible meanings that are partly determined and
constrained by the contexts and circumstances in which the metaphorical sen-
tence is used. The distinction between LM and MM in terms of whether they are
faithful to the principle of compositionality can be used to mark a difference
between metaphors and ambiguous sentences. The different meanings of an
ambiguous sentence are all determined by a compositional analysis of the literal
meanings of the words where either the different lexical meanings of the words
are used in the analysis (as for instance in the case of “he went to the bank”) or
that the compositional structure is permuted (as in the case of “he killed the man
with an umbrella”). Metaphors are not ambiguous either lexically or structurally.
SEMANTIC MEANING AND CONTENT…
119
For the different meanings of a metaphor belong to MM which does not entertain
the use of compositional analysis.
So, in agreement with Davidson, the words in a metaphor, like those of most
figurative expressions, retain their ordinary meanings and significations, but
unlike Davidson, the meaning of a metaphorical sentence is not computed from
the literal meanings of the words that make it up. Metaphors and other figurative
expressions defy the principle of compositionality. Since the principle of compo-
sitionality does not apply in the case of metaphors, it implies that metaphorical
sentences, contra Davidson, do not have only literal meanings and should not be
evaluated with literal truth conditions. Does this imply that metaphorical sen-
tences have additional meanings other than their literal meanings? This will
amount to similarly asking whether an idiomatic expression has an additional
meaning other than its literal meaning. Is the meaning that she is dead an addi-
tional meaning of the idiomatic expression “to kick the bucket”? It is obvious
that the idiomatic meaning of “she kicked the bucket” just is that she is dead,
because the sentence has been construed figuratively or idiomatically. And as we
have seen, this meaning is the MM that is not a resultant of the compositional
analysis of the words of the sentence. This MM is not a meaning in addition to
the LM of the sentence, since the sentence has been construed figuratively. In the
same vein, construing the sentence “Juliet is the sun” literally and realizing that
the sentence is false or absurd is not an indictment on the sentence when it is
construed metaphorically. The MM of the metaphor is not a meaning extra or
additional to its LM, as if they are derived from the same analysis. What exists
here is a meaning difference in kind, which reflects a difference in construal of
the sentence: a sentence construed literally employs a compositional analysis in
determining what it means literally; that same sentence construed metaphorically
or figuratively, adopts a non-compositional analysis in determining what it
means non-literally.
In summary, we have three models for associating meaning and content with
metaphors in relation to the literal:
120
RICHMOND KWESI
Table 1
Model
Word
Progress
Sentences
Model 1
Literal meanings
Metaphorical
meanings
Compositionality
Literal meanings
Metaphorical
meanings
Model 2
Literal meanings
Compositionality
Literal meanings
Model 3
Literal meanings
Literal meanings
Compositionality
Non-compositionality
Literal meanings
Metaphorical
meanings
The first model is the view that Davidson attacks which posits that metaphors
trade on the ambiguity of words, that words have literal and metaphorical senses,
and hence metaphorical sentences have two meanings (Beardsley, 1962; 1978;
Goodman, 1968; 1979). This model ought to explain how words come to acquire
metaphorical meanings and how they are composed to form metaphorical mean-
ings of sentences. A major problem for this model is how it can satisfactorily
explain the phenomenon of dead metaphors: as Davidson puts it, when the dead
metaphor “he was burned up” was active, “we would have pictured fire in the
eyes or smoke coming out of the ears” (1978, p. 38). Davidson’s own view is the
second model which posits that words have only literal meanings and the sen-
tences they compose also have only literal meanings. But this view is not able to
satisfactorily explain how one comes to fail to grasp the metaphor even though
one understands its literal meaning; and as I will show below, the view is also not
able to account for how two people can disagree over the proposition expressed
by a metaphor even in situations where the literal meaning of the metaphor
seems irrelevant to the disagreement or where the two parties can engage in
disagreements even though they agree on the literal meaning of the metaphor.
The view I have tried to formulate above is the third model which grants that
words in a sentence have only literal meanings, but the meanings of the sentenc-
es they constitute are either literal or metaphorical depending on whether the
meanings are derived from a compositional analysis or not. The affinity of meta-
phors with idioms that I drew above suggests that the non-compositional transi-
tion from literal-word-meaning to metaphorical meaning is a matter of sentence,
as opposed to speaker, meaning. Speaker-meaning is a viable alternative route
that is non-compositional in nature, but this route need not make any pro-
nouncement about the literal-metaphorical status of the words of a metaphor. The
third model pursued here is an affirmation of the fact that the words of a meta-
phor do not acquire mythical or mysterious non-literal meanings. But more im-
SEMANTIC MEANING AND CONTENT…
121
portantly, one can use an idiom to speaker-mean (SM) something other than its
LM or MM; SM is pragmatic meaning and it arises in the use of both literal and
figurative expressions and in our linguistic practices in general. This means that
for literal sentences, we can have LM and SM as in the case of implicatures; and
for figurative expressions we can also have LM and SM, as for instance in the
cases of understatements and ironies. But SM alone does not establish that fig-
urative expressions like idioms have, in addition to LM, MM which is semantic
meaning. The model pursued here is that in addition to SM and LM, metaphori-
cal expressions have MM, and this semantic meaning can be different from, or
similar to, although not necessarily derived from, the pragmatic speaker-meaning.
3.2. The problem of Many Contents
A possible objection to the analysis above is that there is a kind of definite-
ness associated with literal meaning and content such that even if we allow both
metaphors and idioms to have MM, that of the idiom is definite and given. We
cannot appropriately talk of the meaning or the content of a metaphor as we do
with an idiomatic expression. If there is no definite content to a metaphor, this
will suggest that it is not a genuine linguistic item that we should be concerned
with associating it with meaning and content. However, this objection is not well
motivated. The point of the “inexhaustibility” (Cohen, 1975) of the interpretation
of metaphor cannot, and should not, be construed as a defect of metaphor. It
should also not be construed as the yardstick for attributing content to metaphor.
Inability to paraphrase a particular metaphor and/or the indeterminacy of the
right kind of paraphrase for a metaphor, are not in themselves indicators of the
absence of any content that the metaphor might have. Rather, the ability to para-
phrase (most) metaphors into propositional form is an indication that metaphors
have contents.
If a metaphor expresses two or more propositions or if it has more than one
interpretation, or if it can be paraphrased into more than one sentence, then it is
not a matter of its having no content but that it has “many contents”. A denial of
the content of metaphor rests on the flawed principle that many contents mean no
content at all; it is like when you say too much, you end up not saying anything
at all. Although a metaphor says too much, it at least says something. And it is
because it says something that we are able to give at least one paraphrase of it.
The objection that many contents imply no genuine linguistic item loses its
sway when we consider treatments of vagueness and borderline predicates in the
literature where vague sentences are made truth-evaluable by such methods as
supervaluationism (Fine, 1975; Keefe, 2000; 2008; Cobreros, 2008). Vagueness
is considered a semantic phenomenon
10
(Keefe, 2008; Cobreros, 2008) resulting
10
Williamson (1994) for instance, regards vagueness as an epistemic phenomenon by
treating the proposition a vague sentence expresses as a borderline case which is either
true or false, but we are ignorant of which value it is.
122
RICHMOND KWESI
from semantic indecision in the sense that “nothing in the world, either in the use
or in any other factor relevant to the determination of the meaning of a vague
predicate, decides which of the ways in which we could make precise the predi-
cate is correct” (Cobreros, 2008, p. 292). Vague sentences are therefore consid-
ered to be indeterminate; they are neither true nor false. However, a supervalua-
tional model can be applied to a vague sentence to make it either (determinately)
true or false by means of an a dmi ss ibl e p re cis if ica tion whereby the
sentence is made more precise. In this way, the vague sentence is true if and only
if it is true on all ways of making it precise, and false if and only if it is false on
all ways of making it precise, and neither true nor false otherwise (Fine, 1975;
Keefe, 2000; 2008). The point of supervaluationism in relation to vagueness is to
show how a vague sentence or a multiple-referring expression can be made truth-
evaluable and be accorded a definite truth value. If we can provide a semantics
for vague sentences, then, despite the differences between metaphors and vague
sentences which there might be, the indefiniteness objection to metaphors cannot
be used to deny its capacity to be appraised for truth. For, on a supervaluational
operation on a metaphor, one can take a metaphor to be true or false on all ad-
missible ways of precisifying it, where the precisification could be in the form of
literalizing or paraphrasing the metaphor. The claim here is that metaphors, like
vague sentences and borderline cases, have contents which admit of many possi-
ble precisifications/paraphrases; and just as the many contents a vague sentence
may have do not preclude it from being appraised for truth, it cannot be correct
that metaphors’ having many contents implies that they are not genuine linguistic
items that can be truth-evaluable.
3.3. Metaphor from the Perspective of the Metaphor-Maker
Besides the rather disparaging remark about metaphor as a noise, both Rorty
and Davidson explain metaphor with respect to the effects it has on the hearer.
While this may be true—that is, metaphors have certain effects on hearers—the
explanation is one-sided and inadequate: on the one hand, it gives no explanation
of metaphor from the speaker’s perspective; and on the other hand, the explana-
tion it gives cannot effectively be extended to the maker of a metaphor. How is
the speaker to understand her metaphorical utterance if she is merely making an
unfamiliar noise? What effect is metaphor to have on the maker of the metaphor?
How is the metaphor to cause a change in beliefs and desires if it is to be con-
strued as merely a noise from the speaker’s perspective? Talk of the effects of
metaphor seems accurate when we are considering the role of metaphor from the
point of view of the audience or hearer, but it seems inappropriate to suggest that
the metaphor also causes certain effects in the one making it. The causal account
of metaphor fails to note that there can be both internal and external noises. Ex-
ternal noises may have effects on us and cause us to do certain things or behave
in certain ways; but internal noises are internally generated, and hence the effects
the noises may have on others may not necessarily apply to the generator of the
SEMANTIC MEANING AND CONTENT…
123
noise. In the case of metaphor, the effects metaphor is seen to produce do not
apply to the maker of the metaphor. Hence, an account of metaphor that only
explains metaphor in terms of noises and effects on the part of the audience is an
inadequate one.
It is one thing to say that a metaphor can cause one to entertain certain beliefs
and propositions, and another thing to say that a metaphor is an outward expres-
sion of the beliefs and propositions one has (or is) entertained (or entertaining).
We can agree with Davidson and Rorty (for the sake of argument), that the sorts
of things that a metaphor may cause one to entertain are not propositional in
nature, but this agreement does not imply that the metaphor itself cannot express
a proposition that has been entertained by the maker of the metaphor. One cannot
use the non-propositional character of the sort of things a hearer is caused to
entertain to deny the propositional character of the metaphorical statement that
the speaker of the metaphor may assert. What a metaphor may be “used to do”,
what a metaphor may “cause” one to do or entertain, and the “effect” of what
a metaphor may have on anyone, do not offer an analysis of, and cannot be used
to explain, what a metaphor is.
11
What a metaphor is—a statement or utterance
borne out of the beliefs and propositions conceived and entertained by a speak-
er—and what a metaphor may “suggest” or “point out”, are also separate issues.
One has to be cautious not to conflate, first, the essence and work of metaphor,
and second, the analysis of metaphor from the perspectives of the hearer and
speaker. We can delineate the activities of the speaker and hearer of a metaphor
from the “work” of the metaphor itself. “Nudging”, “poking” and “directing of
attention”, a metaphor can do, but this work of the metaphor does not say any-
thing about whether metaphors can be associated with the expression of proposi-
tional contents. If we are interested in what a metaphor can be used to do, and the
causes and effects associated with a metaphor, the analysis can begin from the
metaphor itself and the force it has on hearers. If we are interested in the mean-
ing or interpretation of a metaphor, if we are poised to give paraphrases of
a metaphor, we can attempt this from the perspective of the hearer by developing
strategies and mechanisms the hearer could use, although in most cases, this
cannot be done adequately independently of the intentions of the speaker. But we
cannot use our conclusions about the causes and effects of a metaphor on the
hearer to posit certain assumptions about the making of the metaphor or about
the essence of the metaphor itself.
Let us suppose with Davidson and Rorty that we should understand metaphor
as seeing one thing as another thing. From a causal account then, we can explain
how a metaphor (or perhaps the metaphor-maker) causes the hearer to see one
thing as another thing. But we cannot appropriately explain by the account that
11
Here, the distinction is between the functions and effects of metaphor on the one
hand, and the constitution of metaphor on the other hand. The functions and effects may
be used to elucidate what metaphor is, but they cannot stand for what constitutes a meta-
phor. Similarly, the things a metaphor may cause a hearer to entertain can be distinct from
the thing—a proposition, perhaps—a metaphor may assert.
124
RICHMOND KWESI
the speaker or the metaphor-maker is caused to see one thing as another thing. In
fact, the seeing-as experience happens prior to the causes and effects that take
place. The metaphor-maker is not caused to see anything by the metaphor and
neither does the metaphor have any effect on the metaphor-maker; the metaphor
rather reflects what the metaphor-maker has already seen or experienced.
12
A metaphor is like “a bump on the head” Davidson says, but on whose head? It
cannot be on the head of the metaphor-maker. A metaphor is like “a joke” but
who is to get the joke? It is not about whether the joke is funny or not—a come-
dian gets his own joke as he is the one making it. A causal account cannot ex-
plain the making of novel metaphors even if it can explain the reception of novel
metaphors. Thinking of metaphors in terms of effects leaves out the production
of metaphors even if the metaphor-maker is using the metaphor to bring out
certain effects in others. A cause-effect approach to the understanding of meta-
phor cannot extend to the making, conception, and evaluation of metaphor.
13
What causes and effects could there be when the metaphor-maker uses a met-
aphor in a soliloquy? We can make sense of how a metaphor-maker may attempt
to bring about certain effects in his audience, and perhaps, where there is no
audience, the intended effect may not be successful or applicable. But this pre-
sents a problem in the case of soliloquies where the metaphor-maker is his own
audience, that is, in this case, the metaphor-maker utters the metaphor to himself
rather than to a perceived audience. In this case, it does not seem right to suppose
that the metaphor-maker utters a metaphor to bring about some effects in him or
to cause himself to see certain insights. The making of a metaphor is an inten-
tional action and it is not clear how a metaphor-maker nudges himself into notic-
ing things when he utters a metaphor to himself. The causal theorist could ex-
12
An interesting way of making this point is to take seriously the is (identity) of the
metaphorical “x is y”. The seeing-as view conceives of metaphor as a figure of speech in
which the is metamorphosis into an as so that when the metaphor says that “x is y”, it can
be understood as seeing x as y. But I as defend in Kwesi (2018b), in metaphor, the meta-
morphosis is reversed: seeing x as y, that is, perceiving or creating a resemblance between
x and y, involves a transformative process that changes an as into an is so that the meta-
phorical assertion of “x is y” is a resultant of the transformative process. The “x is y” is an
identity statement, a fusion of the x and y into a new reality. The metaphor-maker asserts
that “x is y” not consequently to be caused to see x as y, but rather the assertion represents
an antecedent seeing of x as y that has now transfigured into the claim that x is y.
13
Davidson could respond to the criticism in this paragraph by saying that although
the speaker is not caused to do anything, the speaker uses a metaphor with the primary
intention of producing certain effects in his hearers. And hence, the making, conception,
and evaluation of metaphor can be understood in terms of the intention to produce effects
and the subsequent success or effectiveness of those effects on hearers. However, this
response is not satisfactory: it merely shifts the locus of the criticism to the effects on
hearers. And, as I go on to argue in the next paragraph, this primary intention to produce
effects is not applicable in cases where the speaker is speaking to himself alone. In solilo-
quies, it is not only that the speaker is not caused to make a metaphor, he also does not use
a metaphor with the intention of producing an effect in himself.
SEMANTIC MEANING AND CONTENT…
125
plain that the soliloquist uses a metaphor as if there was an audience and that the
absence of an existing audience does not imply that there are no causes and ef-
fects of the metaphor; it only shows that these effects do not act on anyone, but
would if there were indeed existing audiences. This explanation may seem plau-
sible in a particular kind of soliloquy. We can distinguish between two senses of
soliloquies: in the one sense, a speaker makes a speech to himself with an audi-
ence in mind like in the cases of practicing a speech one is to give at a later date
or when an actor on stage gives a monologue; in the other sense, the speaker has
no intended or perceived audiences other than himself as in the cases of thinking
out loud or making a note to oneself. It is the second of these senses that causes
a problem for the causal account. Where there is a perceived audience, the maker
of the metaphor may have certain beliefs about how his metaphorical utterance
will affect his audience or the various effects his utterance might have. But the
causal account is not able to explain where these beliefs come from and how they
are generated. Where there are invisible or perceived audiences, and where there
are no audiences at all (perhaps, other than the speaker himself) from the speak-
er’s perspective, the causal account cannot satisfactorily explain how speakers
deliberately utter metaphors to themselves.
4. THE SOCIAL PRACTICES OF USING METAPHORS
There are certain features of our use of metaphor that give us good reason to
assume that metaphor has meaning and content rather than mere effects on us.
Our shared communal practice of employing similar metaphors in everyday
discourse attests to the fact that there is a meaning that is grasped and shared by
all. Rarely are live metaphors confined to individual speakers in a community.
The same active metaphors may be used by a number of speakers or writers in
a particular linguistic community. An effect-based approach to metaphor only
assumes that one is dazzled upon hearing a metaphor, that one is directed to
notice certain similarities between two things. But even if we grant that this is
the only business of metaphor—directing one’s attention to notice similarities—
the ability of two or more people using the same metaphor to enable others to
notice the same similarities presupposes that there is a meaning and content of
the metaphor that is shared by them.
If Davidson is right that what many people refer to as the content of a meta-
phor is merely an effect metaphor has on hearers, how can we predict that the
same or a similar effect can occur each time a particular metaphor is used? How
is the hearer able to grasp a metaphor, exploit it, and use it to produce similar
effects on others? How can we judge which effect is appropriate or inappropriate
to have in each context of use of the metaphor? If someone is banged on the head
but feels no pain, he has a deviant reaction, yet we don’t criticize him. But if
a hearer fails to get the point of a metaphor—treating it as only having literal
content or getting the wrong metaphorical interpretation—then he is apt for criti-
cism. Causal patterns only have deviant instances; and causal deviance doesn’t
126
RICHMOND KWESI
warrant censure. Since receipt of a metaphor can, on occasion, warrant censure,
it is not merely a causal phenomenon. The censure or criticism that is associated
with metaphors is even more salient with respect to the making of metaphors.
A principal feature of metaphors that was highlighted primarily by rhetoricians is
the aesthetic or ornamental value of metaphors: metaphors are useful for embel-
lishing speeches. Hence, rhetoricians developed rules and guidelines for making
apt and poetic metaphors that will make speech pleasant. Hackneyed and trite
metaphors, and metaphors that involved obscene language were criticized for
being unpleasant to the ear, and the makers of such metaphors were seen to lack
the artistic skills of making figures of speech. If metaphors are like “bumps on
the head” as Davidson argued, the criticisms associated with metaphors will not
be applicable; indeed, talk of using metaphors to embellish speech or appreciat-
ing metaphors for their aesthetic value will be meaningless. To the extent that
some metaphors can be appraised as live, vivid, insightful, astute, and to the
extent that some metaphors can be criticized as being banal, pale, unimaginative,
metaphors are not merely causal prods; for nudges and pokes and prods are not
inherently praised or criticized.
If we can meaningfully talk about grasping or understanding a metaphor,
what is it that we grasp and understand? The effect? The content? Isn’t the ability
to grasp a particular metaphor and effectively use that metaphor in other contexts
with the expectation that others understand and utilize that metaphor an indica-
tion of something more than effects at play? If metaphor only has a point or if it
merely intimates one to see something in a certain way, we cannot conclude from
this that grasping the point of a metaphor or being nudged to perceive certain
similarities will result in one using the same metaphor to put across the same
point or to nudge others to perceive the same similarities. It is very mysterious
how one becomes aware of the effect of an utterance on him and whether that
was the intended effect of the utterance, and that in using the same utterance he
will be bringing about the same effect. Also, if all there is to metaphor is the
effects it has on one, it is not clear whether the effects include the ability to pass
on the same metaphor to achieve similar effects in others. And similarly, an ef-
fect-based account of metaphor cannot explain one’s ability to teach and explain
metaphors to others, for in teaching and explicating metaphors to others we do
not just indicate what the causes and effects of metaphors are. Sameness of the
effects of the metaphor on two people cannot account for their ability to use the
metaphor to produce the same or similar effects. Meanings are such things that
are transferrable; effects are generally not.
The difference in the abilities of two people to use the same metaphor can be
attributed to their understanding—grasp of meaning—of the metaphor. To be
able to use a metaphor in multiple contexts, to be able to use a metaphor to in-
tend to achieve a different effect, marks the presence of understanding of the
metaphor, such an understanding involves both the grasp of meaning and the
ability to use the metaphor (Kwesi, 2019b). Meaning is, therefore, at stake in
SEMANTIC MEANING AND CONTENT…
127
both the ability to use a metaphor and the inability to use a metaphor in situations
where one is aware of the effect of the metaphor.
In practice, hearers rarely ask for the meaning or interpretation of metaphors
they freshly encounter; yet, they work out the meaning of those metaphors and
employ the same metaphors in their own discourses with the expectation that
other hearers will be able to work out what the metaphors mean. The capacity to
work out what a novel metaphor means, unaided by the metaphor-maker, in-
volves, at the very least, a kind of reflective comprehension of the effects of the
metaphor. He ar ers can b ec ome us er s of certain metaphors not only
because of their ability to appreciate the point of metaphors or the similarities
they are directed to perceive, but, more importantly, because they can reflect on,
and understand, the content of metaphors. This observation is common to both
literal and metaphorical uses of language. The crude causal account cannot ade-
quately explain how hearers of metaphors can be become effective users of met-
aphors.
Another crucial feature of our practices of using metaphors is our capacity to
use metaphors in arguments and engage in drawing certain inferences and impli-
cations from metaphorical sentences (Kwesi, 2018a; 2019a; 2019b). Consider
these two arguments from Martinich (1996, p. 431, 435):
(1). “My love is a red rose.
A red rose is beautiful, or sweet smelling, or highly valued…
Therefore, my love is beautiful, or sweet smelling, or highly valued…”
14
(2). “No man is an island
Every island is separated from every other thing of its own kind, does not
depend upon any other thing of its own kind for its existence or well-being,
and is not diminished by the destruction of any other of its own kind; …
Therefore, no man is separated from every other thing of its own kind, does
not depend upon any other thing of its own kind for its existence or well-
being, and is not diminished by the destruction of any other of its own kind”
(Martinich, 1996, p. 435).
15
Martinich considers (1) a valid argument and (2) an invalid argument. We
need not worry about the validity of the arguments containing metaphors; it is
enough to see that metaphorical sentences can serve as premises in arguments;
14
The ellipsis is intended to show the open-endedness of the metaphor
15
Martinich, however, thinks that “no man is an island” is not a metaphor. For accord-
ing to him, “every metaphorical proposition is false” (1996, p. 430) and “it is true and not
false that no man is an island” (p. 435) although he concedes that Donne’s line is a figure
of speech. I regard it as a metaphor because I do not subscribe to the view that the identi-
fying mark of a metaphor is literal falsity.
128
RICHMOND KWESI
metaphorical sentences play a role in reasoning. As premises of arguments, they
can serve as reasons and justifications for conclusions, and stand in need of rea-
sons and justifications (Kwesi, 2019a). The observation that metaphors can serve
as premises and conclusions of arguments and that we can draw inferences from
the metaphors we put forward suggests that metaphors must have meanings and
contents. In reasoning with metaphors, users and their audiences are able to
make inferences from the metaphors and provide other statements (metaphorical
and literal) that tend to extend and explicate further the meaning and import of
the metaphors. Consider the popular Psalm 23 from the Bible:
The Lord is my shepherd, I shall not want.
He maketh me to lie down in green pastures;
He leadeth me beside the still waters.
He restoreth my soul:
He leadeth me in the paths of righteousness, for his name’s sake.
Yeah, though I walk through the valley of the shadow of death,
I will fear no evil: for thou art with me,
Thy rod and thy staff, they comfort me.
What is going on here in the psalm is that the Psalmist starts with the meta-
phor “the lord is my shepherd” and provides inferences that we can draw from
the metaphor: if the lord is my shepherd then I shall not want, he will lead me to
green pastures, his rod and staff with comfort me, etc.
16
Tirrell (1989) has called
this phenomenon the “extending of metaphor”. “The Lord is my shepherd” in our
example is for her the “initiating metaphor” and the other expressions as the
“extensions” of the metaphor which together with the initiating metaphor form
a “metaphorical network or chain”. The Psalmist presents us with an inferential
metaphorical network where we see that his not wanting and being led to green
pastures follow from his initial metaphor that the lord is his shepherd. For Tirrell,
understanding a metaphor amounts to being “able to make appropriate uses of its
extensions” (p. 18). Sometimes, the metaphor-maker herself provides the various
extensions of the metaphor which develop and explain the metaphor in more
detail. An example Tirrell uses is from Shakespeare’s Romeo and Juliet where
Lady Capulet not only tells Juliet to “read o’er the volume of Young Paris’ face”
but extends her metaphor to provide better and further particulars of what is
involved in this kind of reading:
16
This does not suggest that the Psalmist himself is actually making these inferences
and connections; it is enough for his audience and readers of the Bible to draw these
connections as they ponder on the initial metaphor.
SEMANTIC MEANING AND CONTENT…
129
Read o’er the volume of Young Paris’ face,
And find delight writ there with beauty’s pen;
Examine every married lineament,
And see how one another lends content;
And what obscur’d in this fair volume lies
Find written in the margent of his eyes.
This precious book of love, this unbound lover,
To beautify him, only lacks a cover. (Act 1 Sc 3)
Tirrell herself stops short of saying that the extensions of the metaphor serve
as unpacking the meaning and content of the metaphor; her interest is in showing
how extended metaphors impact our understanding of metaphors. However con-
troversial the relation between the original metaphor and its extensions could be,
the possibility of providing extensions to a metaphor suggests that the original
metaphor had a meaning and a content, for it does not sound intuitive to suggest
that the effects of the metaphor were being extended. The meaning of a metaphor
can be extended; its effects cannot be analogously extended.
A final feature about metaphor worth noting is that we can agree and disagree
with, assent and dissent to, certain metaphorical utterances. Such agreements and
disagreements reflect our understanding of metaphors—we cannot agree or disa-
gree on a metaphor if we do not understand it. Also, if we can agree or disagree
over a metaphorical sentence then it implies that the sentence has been recog-
nized or identified as a metaphor, since a genuine disagreement cannot obtain
between two people over a particular sentence if one construes the sentence
metaphorically and the other understands it literally. The two people clearly
agree about the statement as put forward literally, but disagreement only emerges
when the sentence is considered metaphorically. For instance, where Tom asserts
that “the vice-chancellor is a bulldozer” and Harry responds by saying that “No,
that’s not true”, the use of “that” here refers to the proposition expressed by
Tom’s assertion. Harry’s response here expresses his disagreement with the con-
tent of the assertion made by Tom.
If, as argued by Davidson and Reimer (2001), a metaphorical assertion like
“the vice chancellor is a bulldozer” only has a literal content or expresses a literal
proposition, Harry’s response will be conversationally infelicitous or inappropri-
ate. For, the metaphorical assertion is literally false and hence, responding to it
by saying that “that’s not true” or “that’s false” is both inappropriate and unin-
formative. But if Harry is warranted in making his response, if he is understood
to be denying the assertion made by Tom, and if his use of “that” refers to the
proposition expressed by Tom’s assertion, then it is plausible to suppose that
there is a propositional content other than the literal content of Tom’s assertion
that Harry rejects here. The intuitive conflict in the dialogue between Tom and
130
RICHMOND KWESI
Harry can be attributed to the content expressed by Tom’s assertion that Harry
disagrees with. In our practices of using metaphors we can have disagreements—
disagreements not merely over the significance or effects of metaphors but the
contents expressed by the metaphors. The notion of disagreement primarily in-
volves an incompatibility in the attitudes of the disagreeing parties towards
a particular proposition. And if disagreements can occur with metaphors then we
can infer that metaphors must have contents for disagreements to be possible.
5. CONCLUSION
The above criticisms of the denial of the meaning and content of metaphors
and the arguments we adduced in favour of metaphors having contents, suggest
the following desiderata for a satisfactory account of metaphorical content:
1. Non-compositionality: The account should explain how the content of
a metaphor is not arrived at by a compositional analysis, although, unlike
that of an idiom, the content is connected to the meanings of the constitu-
ents that make it up.
2. Figurativeness: The account should explain the ways in which the meta-
phorical is distinct from the literal in terms of the derivation of their con-
tents.
3. Disagreement: The account should be compatible with how there can be
genuine disagreements involving metaphors.
4. Assertion and Retraction: The account should show speakers’ ability to put
forward claims and stand by those claims or retract earlier claims. It should
also be able to explain how metaphors can serve as premises and conclu-
sions of arguments.
5. Inference and Extension: The account should explain speakers’ ability to
make inferences from metaphorical claims and be able to extend and expli-
cate original metaphors
6. Use in Soliloquies: The account should make sense of speakers’ use of
metaphors in monologues and in soliloquies where there are no intended
audiences.
7. Hearers’ Uptake: The account should be able to explain hearers’ immediate
understanding of metaphors and their ability to use the metaphors in other
contexts to produce effects on their hearers.
We can conclude that the Davidsonian causal account of metaphor gets it
right by arguing that the words in a metaphor do not have additional or extra
meanings other than their literal meanings. The account also seems plausible in
indicating that metaphor also “nudges”, “provokes”, and “intimates” us to do
things in certain ways because they have effects on us. But for reasons given
SEMANTIC MEANING AND CONTENT…
131
above, the account is not favourable in its further thesis that metaphors having
effects is opposed to their having content, meaning or truth. I have tried to show
that one can accept the central thesis of the causal account of metaphor (when it
is understood in terms of word-literalism) and still posit that metaphors have
content and meaning. The fact that we can use/misuse metaphors, that we can
understand/misunderstand metaphors, that we can agree/disagree with metaphors,
and the fact that we do reason with metaphors in arguments and make inferences
from metaphors—all go to show that associated with a metaphor is a proposi-
tional content that we can grasp and evaluate.
REFERENCES
Beardsley, M. (1962). Metaphorical Twist. Philosophy and Phenomenological
Research, 22(3), 293–307.
Beardsley, M. (1978). Metaphorical Senses. Noûs, 12(1), 3–16.
Black, M. (1979). How Metaphors Work: A Reply to Donald Davidson. In:
S. Sacks (Ed). On Metaphor (pp. 181–192). Chicago: The University of Chi-
cago Press.
Black, M. (1993). More About Metaphor. In A. Ortony (Ed.), Metaphor and
Thought. (2nd ed., pp. 19–41). Cambridge: Cambridge University Press.
Camp, E. (2006a). Metaphor and that Certain “je ne sais quoi”. Philosophical
Studies, 129(1), 1–25.
Camp, E. (2006b). Metaphor, Contextualism, and What Is Said. Mind & Lan-
guage, 21(3), 280–309.
Camp, E. (2006c). Metaphor in the Mind: The Cognition of Metaphor. Philoso-
phy Compass, 1(2), 154–170.
Camp, E. (2008). Showing, Telling, and Seeing: Metaphor and “Poetic” Lan-
guage, The Baltic International Yearbook of Cognition, Logic, and Commu-
nication: A Figure of Speech: Metaphor, 3, 1–24.
Carston, R. (1988). Implicature, Explicature and Truth-Theoretic Semantics. In:
R. M. Kempson (Ed.), Mental Representations: The Interface Between Lan-
guage and Reality (pp. 155–81). Cambridge: Cambridge University Press.
Carston, R. (2002). Thoughts and Utterances: The Pragmatics of Explicit Com-
munication. Oxford: Blackwell.
Chomsky, N. (1965). Aspects of the Theory of Syntax. Cambridge, MA: MIT
Press.
Chomsky, N. (1980). Rules and Representations. New York: Columbia Universi-
ty Press.
Cobreros, P. (2008). Supervaluationism and Logical Consequence: A Third Way.
Studia Logica, 90(3), 291–312.
132
RICHMOND KWESI
Cohen, T. (1975). Figurative Speech and Figurative Acts. Journal of Philosophy,
72(19), 669–684.
Cooper, D. (1986). Metaphor. New York: Basil Blackwell.
Davidson, D. (1967). Truth and Meaning. In: D. Davidson (Ed.), Inquiries into
Truth and Interpretation (pp. 17–42). Oxford: Clarendon Press.
Davidson, D. (1979). What Metaphors Mean. In: S. Sacks (Ed), On Metaphor
(pp. 29–45). Chicago: The University of Chicago Press.
Davidson, D. (1984). Inquiries into Truth and Interpretation. Oxford: Oxford
University Press.
Davies, M. (1982). Idiom and Metaphor. Proceedings of the Aristotelian Society,
New Series, 83(1), 67–85.
Davies, S. (1984). Truth-Values and Metaphors. The Journal of Aesthetics and
Art Criticism, 42(3), 291–302.
Dever, J. (2008). Compositionality. In E. Lepore, B. Smith (Eds.), The Oxford
Handbook of Philosophy of Language. Oxford: Oxford University Press.
Farrell, F. (1987). Metaphor and Davidsonian Theories of Meaning. Canadian
Journal of Philosophy, 17(3), 625–642.
Fine, K. (1975). Vagueness, Truth and Logic. Synthese, 30(3/4), 265–300.
Fodor, J., Lepore, E. (2002). The Compositionality Papers. Oxford: Oxford Uni-
versity Press.
Goodman, N. (1968). Languages of Art: An Approach to a Theory of Symbols.
Indianapolis: The Bobbs-Merrill Company Inc.
Goodman, N. (1979). Metaphor as Moonlighting. Critical Inquiry, 6(1), 125–30.
Hall, A. (2009). Semantic Compositionality and Truth-Conditional Content.
Proceedings of the Aristotelian Society, New Series, 109(1pt3), 353–364.
Hesse, M. (1987). Unfamiliar Noises II: Tropical Talk: The Myth of the Literal.
Proceedings of the Aristotelian Society, 61, 283–311.
Hesse, M. (1988). The Cognitive Claims of Metaphor. The Journal of Specula-
tive Philosophy, 2(1), 1–16.
Johnson, M. (2008). Philosophy’s Debt to Metaphor. In: R. Gibbs (Ed.), The
Cambridge Handbook of Metaphor and Thought (pp. 39–52). Cambridge:
Cambridge University Press.
Katz, J. (1973). Compositionality, Idiomaticity, and Lexical Substitution. In: S. R.
Anderson, P. Kiparsky (Eds.), A festschrift for Morris Halle (pp. 357–376).
New York: Holt, Rinehart & Winston.
Keefe, R. (2000). Theories of Vagueness. Cambridge: Cambridge University
Press.
Keefe, R. (2008). Vagueness: Supervaluationism. Philosophy Compass, 3(2),
315–324.
King, J., Stanley, J. (2005). Semantics, Pragmatics, and the Role of Semantic
Content. In: Z. Szabo (Ed.), Semantics Versus Pragmatics (pp. 111–64). Ox-
ford: Oxford University Press.
Kracht, M. (2011). Interpreted Languages and Compositionality. Studies in Lin-
guistics and Philosophy, 89, 1–223.
SEMANTIC MEANING AND CONTENT…
133
Kwesi, R. (2018a). Metaphor, Truth, and Representation. In P. Stalmaszczyk
(Ed.), Objects of Inquiry in Philosophy of Language and Linguistics (pp.
117–146). Berlin: Peter Lang.
Kwesi, R. (2018b). Resemblance and Identity in Wallace Stevens’ Conception of
Metaphor. In K. Bartczak & J. Macha (Eds.). Wallace Stevens: Poetry, Phi-
losophy, and Figurative Language (pp. 113–137). Berlin: Peter Lang.
Kwesi, R. (2019a). An Inferential Articulation of Metaphorical Assertions. RIFL,
13(1), 116–132.
Kwesi, R. (2019b). Davidson's Phenomenological Argument Against the Cogni-
tive Claims of Metaphor. Axiomathes, 1–24.
Lasersohn, P. (2012). Contextualism and Compositionality. Linguistics and Phi-
losophy, 35(2), 171–189.
Leddy, T. (1983). Davidson’s Rejection of Metaphorical Meaning. Philosophy
& Rhetoric, 16(2), 63–78.
Lepore, E., Stone, M. (2010). Against Metaphorical Meaning. Topoi, 29(2), 165–
180.
Lepore, E., Stone, M. (2015). Imagination and Convention: Distinguishing
Grammar and Inference in Language. Oxford: Oxford University Press.
Martinich, A. (1996). The Philosophy of Language. Oxford: Oxford University
Press.
Moran, R. (1989). Seeing and Believing: Metaphor, Image and Force. Critical
Inquiry, 16(1), 87–112.
Nunberg, G., Sag, I. A., Wasow, T. (1994). Idioms. Language, 70(3), 491–538.
Pagin, P., Westerstahl, D. (2010a). Compositionality I: Definitions and Variants.
Philosophy Compass, 5(3), 250–264.
Pagin, P., Westerstahl, D. (2010b). Compositionality II: Arguments and Problems.
Philosophy Compass, 5(3), 265–282.
Recanati, F. (2004). Literal Meaning. Cambridge: Cambridge University Press.
Reimer, M. (2001). Davidson on Metaphor. In: P. A. French, H. K. Wettstein
(Eds.), Midwest Studies in Philosophy, Volume 25: Figurative Language (pp.
142–55). Boston, Mass, Oxford, UK: Blackwell.
Rorty, R. (1979). Philosophy and the Mirror of Nature. Princeton: Princeton
University Press.
Rorty, R. (1987). Unfamiliar Noises I: Hesse and Davidson on Metaphor. Pro-
ceedings of the Aristotelian Society, Supplementary Volumes, 61, 283–311.
Rorty, R. (1989). Contingency, Irony, and Solidarity. Cambridge: Cambridge
University Press.
Sperber, D., Wilson, D. (1995). Relevance: Communication and Cognition. Ox-
ford: Blackwell.
Stanley, J. (2000). Context and Logical Form. Linguistics and Philosophy, 23(4),
391–434.
Stanley, J. (2002). Making it Articulated. Mind and Language, 17(1–2), 149–168.
Szabo, Z. (2010). The Determination of Content. Philosophical Studies: An In-
ternational Journal for Philosophy in the Analytic Tradition, 148(2), 253–272.
134
RICHMOND KWESI
Tirrell, L. (1989). Extending: The Structure of Metaphor. Noûs, 23(1), 17–34.
Titone, D. A., Connine, C. M. (1999). On the Compositional and Noncomposi-
tional Nature of Idiomatic Expressions. Journal of Pragmatics, 31, 1655–
1674.
Williamson, T. (1994). Vagueness. London: Routledge.
