Why Polysemy Supports Radical Contextualism

François Recanati

To cite this version:

François Recanati. Why Polysemy Supports Radical Contextualism. Patrick Brézillon; Roy Turner;
Carlo Penco. Modeling and Using Context, Springer, pp.216-222, 2019, �10.1007/978-3-030-34974-
5_18�. �hal-02932361�

HAL Id: hal-02932361

https://hal.science/hal-02932361
Submitted on 7 Sep 2020

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
 Why Polysemy Supports Radical Contextualism

François Recanati
Collège de France

In G. Bella and P. Bouquet (eds.), Context 2019, Berlin: Springer, 2019, pp. 1-7

Meaning and content

I take content to be fundamentally a property of mental states or acts (e.g. belief or

judgment) and derivatively of speech acts. Consider someone who believes that elephants
have wings, and expresses that belief by saying that elephants have wings. The proposition
that elephants have wings is the content of her belief, as well as the content of the assertion
she makes when she expresses that belief linguistically.

I take meaning to be a property of linguistic expressions (considered as types). The sentence

‘Elephants have wings’ has a certain meaning, and the words in that sentence do as well.

I take the debate between Literalism and Contextualism to bear on the relation between
meaning and content. Literalism holds that they are the same thing. It accepts what I call the
basic equation:

The basic equation:

meaning = content

Contextualism is the opposite view. It rejects the basic equation.

Because it accepts the basic equation, Literalism takes the meaning of the sentence-type
‘Elephants have wings’ to be the proposition that elephants have wings. The meaning of sub-
sentential constituents is taken to be their contribution to the meaning/content of the
sentences in which they occur, i.e. objects, properties, relations etc. or modes of
presentation thereof.

Indexicals

Indexicals constitute an obvious counter-example to the basic equation. Their linguistic

meaning is not the same thing as their content. Their linguistic meaning is invariant, while
their content is contextually variable (whether we take that content to be an object or a
Fregean sense). In Kaplan’s influential framework, the meaning of an indexical is a
‘character’ that determines the content carried by the indexical in context.

Indexicals are not sufficient to arbitrate the debate between Literalism and Contextualism,
however. They are not sufficient because Literalists acknowledge that indexicals constitute
an exception to the Basic Equation. In the case of indexicals, meaning ≠ content. Literalists
accept that. Still, they maintain the basic equation as the default, while Contextualists reject
the basic equation, even construed as the default.
 From Literalism to Methodological Contextualism

Contextualism comes in several varieties. I distinguish between methodological and

substantial forms of Contextualism, and between two substantial forms of Contextualism
(Figure 1).

Contextualism

Methodological Substantial

Moderate Radical

Figure 1. Varieties of Contextualism

The weakest form of Contextualism is Methodological Contextualism. It contrasts with

Literalism in the following manner. Literalists take the indexical exception to be well-
circumscribed: there is a list of expressions known to be indexical (the so-called ‘basic set’),1
and for the expressions that are not in that list the basic equation holds. So the default is:
meaning = content (unless we are dealing with an expression in the list). According to
Methodological Contextualism, however, we don’t know in advance which expressions are
indexical and which aren’t. Ahead of inquiry, we should assume (by default) that meaning ≠
content, for indexicality is always a possibility. So Methodological Contextualism reverses
what Literalism takes to be the default.2

Substantial forms of Contextualism

In its substantial forms, Contextualism considers that indexicals are not an ‘exception’:
context-sensitivity generalizes to all expressions (whether indexical or not). All expressions
are such that the content they contribute depends upon the context, in contrast to the
(invariant) linguistic meaning of the expression.

There are two forms of Contextualism that count as substantial by my characterization. One
is moderate, the other radical. Each appeals to a particular phenomenon. Moderate
Contextualism appeals to the phenomenon of modulation, while Radical Contextualism to
the phenomenon of polysemy.

Moderate Contextualism

The meaning of an indexical is gappy and calls for a contextual process of saturation (e.g. an
assignment of values to free variables in logical form). That process is mandatory: without
saturation, no content can be assigned to an indexical expression. According to Moderate

1
The expression ‘basic set’ comes from Cappelen and Lepore 2005.
2
On Methodological Contextualism, see Recanati 1994 and 2004:00.
contextualism, however, there is another contextual process that takes place on the way
from meaning to content: modulation.

Modulation covers processes of sense extension (loosening/broadening) and sense

narrowing (enrichment) as well as semantic transfer (metonymy) and possibly other
phenomena (see Recanati 2004 for an overview). It is hard to deny that a sentence like ‘The
ham sandwich stinks’ carries distinct truth-conditional contents depending on whether the
description is taken literally as referring to the sandwich or metonymically as referring to the
person who ordered it. Similarly, ‘John is crazy’ carries distinct truth-conditions when ‘crazy’
is taken literally and when it is a hyperbole. So context-sensitivity generalizes: Just as the
content of an indexical depends upon the context of use, the content actually carried by an
ordinary, non indexical expression also depends upon the context: it depends on whether,
and how, the literal meaning of the expression is ‘modulated’ in context.

Radical Contextualism

What makes moderate contextualism moderate is the fact that, in contrast to saturation,
modulation is optional: it may or may not take place. Whether or not it takes place depends
upon the context, so the possibility of ‘zero-modulation’ (Recanati 2010) is compatible with
the generalization of context-sensitivity characteristic of substantial contextualism.

Cases of zero-modulation correspond to literal language use. In such cases, the basic
equation still holds: meaning = content. According to radical contextualism, however,
meaning is never identical to content. Lexical meaning is constitutively unable to figure as a
constituent of content ; it does not have the proper format for that (‘wrong format view’).
This is where polysemy comes into the picture.

Polysemy as ambiguity

As soon as an expression comes into public use, it becomes polysemous – the more frequent
its use, the more polysemous it is. The senses of a polysemous expression result from
pragmatic modulation (one sense is a modulation of another) but these modulations have
become conventionalized and the senses of a polysemous expression are stored in the
memory of language users.

Since the senses of a polysemous expression are conventionalized (in contrast to novel
instances of modulation), it is tempting to construe polysemous expressions as straightfor-
wardly ambiguous (Fig. 2).

Expression

meaning1 meaning2

content1 content2

Figure 2. The ambiguity model

Ambiguous expressions contribute different contents in different contexts, but this does not
threaten the literalist equation of meaning and content (since ambiguous expressions
possess distinct meanings).

Objection: Polysemy vs homonymy

Two homonymous expressions (e.g. ‘bank’ and ‘bank’) are different expressions, with the
same phonological realization but distinct meanings. A polysemous expression is supposed
to be something else. A polysemous expression admittedly carries distinct senses, but these
senses are felt as related : they form a family of senses. So instead of two different
expressions with the same shape but distinct meanings (homonymy), what we seem to have
is a single expression, i.e. a semantic as well as a phonological unit (polysemy): The
expression has a single meaning which (depending on one’s theory) either accounts for, or
supervenes on, the diversity of its conventional uses. If we don’t allow polysemous
expressions such an inherent meaning, distinct from the various senses they contribute in
context, we are bound to deny that there is a difference between polysemy and homonymy.
In other words : either polysemy does not exist (as a phenomenon distinct from homonymy),
or, if it exists, it cannot be accounted for along the lines of the ambiguity model.

I call the alternative model we need the ‘context-sensitivity model’ because it posits a single
meaning to which there correspond different contents in different contexts (as in the case of
indexicals).

Expression

meaning

sense1 sense2…

Figure 3. The context-sensitivity model

What is the unitary meaning of a polysemous expression?

For Ruhl (1989), polysemous words possess a highly abstract (and underspecified) meaning
which they carry in all their occurrences and which is responsible for the various senses they
contextually express. Because the underspecified meaning lies below the level of
consciousness, what intuitions reveal (and dictionaries record) are the expressed senses.
These senses depend on context (both linguistic and extralinguistic) , while ‘a word’s
semantics should concern what it contributes in all contexts’ (Ruhl 1989 : 87). The task of the
theorist is to discover lexical meaning by extracting from the data some abstract, unitary
schema which all the uses fit.

I see two problems with this approach. First, it’s not clear how it handles metonymies (which
Ruhl hardly mentions). Second, even though polysemous expressions are conventionally
associated with determinate senses which they regularly convey, these senses are not an
aspect of the linguistic meaning of the expression, in Ruhl’s framework. (The linguistic
meaning is more abstract than these senses.) This is similar to the idea, floated in the
seventies, that there are ‘conventions of use’ that are not ‘meaning conventions’ : e.g. the
convention that ‘Can you pass the salt ?’ is a request that should be complied with rather
than a question that should be answered.3 But this construal of ‘meaning’ is overly narrow.
As Langacker emphasizes, our goal as meaning theorists should be ‘to properly characterize
a speaker’s knowledge of linguistic convention’. Now,

A lexical item of frequent occurrence displays a substantial, often impressive

variety of interrelated senses and conventionally sanctioned usages… Even
when all its attested values are plausibly analysed as instantiations of a single
abstract schema, or as extensions from a single prototype, there is no way to
predict from the schema or prototype alone precisely which array of
instantiations or extensions — out of all the conceivable ones – happen to be
conventionally exploited within the speech community. (Langacker 1987 : 370)

Following Langacker,4 therefore, I take the meaning of a polysemous expression to be

neither a ‘prototype’ nor an ‘all-subsuming superschema’, but the network of senses the
expression is conventionally associated with (including the prototype and/or the
superschema, should there be any, as well as the modulation relations between the senses).

Expression

meaning [sense1  sense2]

content sense1 sense2…

Figure 4. The network model

3
See Searle 1975 and especially Morgan 1978.
4
“A strict reductionist approach would seek maximum economy by positing a single
structure to represent the meaning of a lexical category. However, if our goal is to properly
characterize a speaker’s knowledge of linguistic convention, any such account is unworkable.
From neither the category prototype alone, nor from an all-subsuming superschema (should
there be one), is it possible to predict the exact array of extended or specialised values
conventionally associated with a lexeme (out af all those values that are cognitively
plausible). A speaker must learn specifically, for instance, that run is predicated of people,
animals, engines, water, hosiery, noses, and candidates for political office ; the conventions
of English might well be different. Equally deficient is the atomistic approach of treating the
individual senses as distinct and unrelated lexical items. The claim of massive homonymy
implied by such an analysis is simply unwarranted — it is not by accident, but rather by
virtue of intuitively evident relationships, that the meanings are symbolized by the same
form. A network representation provides all the necessary information : an inventory of
senses describing the expression’s conventional range of usage ; the relationships these
senses bear to one another ; schemas expressing the generalizations supported by a given
range of values ; and specifications of distance and cognitive salience.” (Langacker 1991 :
268)
As can be seen by comparing Figure 4 with Figures 2 and 3, the network model blends
features from the ambiguity model and the context-sensitivity model.

Conversion into sense

In language use, senses multiply and diversify through modulation operations, which are
optional in the sense of context-driven. Think of the first time the word 'swallow' was used
to refer to what an ATM sometimes does with credit cards. The sense of 'swallow' was then
creatively extended so as to exploit the similarity between the ATM situation and ordinary
swallowing-situations. What was extended (the input to modulation) was the standard sense
of 'swallow' as it applies to living organisms with a digestive system. The output of
modulation was the (broadened) sense in which an ATM can be said to swallow a credit
card. As a result of conventionalization, the extended sense has become part of the network
of senses which makes up the lexical meaning of 'swallow', but the modulation relation
between the extended sense and the prototypical sense is still alive in the consciousness of
the language users. So we must distinguish between three things:

(i) the lexical meaning of ‘swallow’, which has the wrong format for being a
constituent of content (it is or comprises a network of senses);
(ii) the standard/literal/prototypical sense of ‘swallow’ (with respect to living
organisms), which was the input to modulation in the ATM example; and
(iii) the extended sense relevant to ATMs, which was, and is still perceived as, the
output of modulation.

On that view there is a principled difference between the linguistic meaning of a

polysemous expression and the sense the expression contributes when used in context
(even the standard or prototypical sense, should there be any). Context-sensitivity thus
generalizes in a way which supports Radical Contextualism. An expression cannot directly
contribute its lexical meaning, which has the ‘wrong format’ for being a constituent of
content. The lexical meaning must be contextually converted into an appropriate sense
through various context-sensitive operations (typically a mixture of sense selection and
modulation).

Conclusion: three types of contextual process

We must distinguish the relation between the lexical meaning of ‘swallow’ and the extended
sense the word takes in ATM-situations, namely a special case of conversion into sense, from
the relation between that extended sense and the standard, prototypical sense of ‘swallow’,
namely modulation. That distinction tends to be neglected because it is often (wrongly)
assumed that the lexical meaning of ‘swallow’ is its standard/prototypical sense. But the
distinction between modulation and conversion into sense is important because conversion
into sense is mandatory, while modulation is optional. This makes conversion into sense
similar to saturation (and polysemy similar to indexicality). Still, conversion into sense
concerns all expressions (to the extent that, to a greater or lesser degree, all expressions are
polysemous). This makes it similar to modulation (which may affect any expression), while
saturation only concerns indexical expressions.
In terms of these two features — mandatoriness and universality — we can characterize the
three contextual processes that map meaning to content:

Mandatory Universal

Saturation + -
Modulation - +
Conversion into sense + +

Figure 5. Saturation, modulation, and conversion into sense

References

Cappelen, H. and Lepore, E. (2005) Insensitive Semantics. Oxford: Blackwell.

Benveniste, Emile (1966) Problèmes de Linguistique Générale, vol. I. Paris : Gallimard.
Benveniste, Emile (1974) Problèmes de Linguistique Générale, vol. II. Paris : Gallimard.
Kaplan, D. (1989) Demonstratives
Langacker, Ronald (1987) Foundations of Cognitive Grammar, vol. 1. Palo Alto : Stanford
University Press.
Langacker, Ronald (1991) Concept, Image and Symbol. Berlin : Mouton De Gruyter.
Morgan, Jerry (1978). Two Types of Convention in Indirect Speech Acts. Syntax and
Semantics 9: 261-80
Recanati, François (1994) Contextualism and Anti-Contextualism
Recanati, François (2004) Literal Meaning. Cambridge : Cambridge University Press.
Recanati, François (2010) Truth-Conditional Pragmatics. Oxford : Oxford University Press.
Ruhl, Charles (1989) On Monosemy: A Study in Linguistic Semantics. Albany, N.Y. : State
University of New York Press.
Searle, John (1975) Indirect Speech Acts. Syntax and Semantics 3: 59-82.