10.1.1.145.4116.pdf

ENRICHMENT AND LOOSENING: COMPLEMENTARY PROCESSES IN

DERIVING THE PROPOSITION EXPRESSED?*

ROBYN CARSTON

-----------------------------------------------------------------

Abstract

Within relevance theory the two local pragmatic processes of enrichment and loosening of

linguistically encoded conceptual material have been given quite distinct treatments.

Enrichments of various sorts, including those which involve a logical strengthening of a lexical

concept, contribute to the proposition expressed by the utterance, hence to its truth-conditions.

Loosenings, including metaphorical uses, do not enter into the proposition expressed by the

utterance or affect its truth-conditions; they stand in a relation of 'interpretive resemblance' with

the linguistically encoded concept used to represent them. This asymmetric treatment is

questioned here, arguments are given for an account which reflects the complementarity of these

processes and several alternative symmetrical treatments are explored.

1 Introduction

One important consequence of the relevance-theoretic view of cognition and communication is

the following: we can think many thoughts that our language cannot encode, and we can

communicate many thoughts that our utterances do not encode. Strictly speaking, virtually no

sentence encodes a complete thought; certain processes of contextual filling-in are required

before anything of a propositional nature emerges at all. However, that more basic point is not

my primary concern in this short paper. The idea is that, even given such processes of

propositional completion, a great many of our thoughts are of a much finer grain than that of the

minimal propositions which result from these processes. It follows that there are many more

concepts (construed as constituents of thoughts) than there are words in the language.

One way of trying to account for this would be to suggest that words are multiply ambiguous,

many of them encoding a vast number of discrete senses. This is not, of course, the way a

relevance-theoretic approach to communication would explain this fact. Most other inferential

pragmatic approaches to communication and interpretation would not take the ambiguity line

either. The relevance theory view (Sperber & Wilson, 1986, 1995) is that our powerful

inferential capabilities enable us to construct ad hoc concepts out of lexically encoded concepts

during our on-line interpretation of utterances, on the hoof as it were. This process is both driven

by and constrained by the inevitable considerations of processing effort and cognitive effects.

The two main varieties of ad hoc concept construction that Dan Sperber and Deirdre Wilson

have discussed in various lectures and seminars over the past decade are the narrowing of

--------------------------------------------

* Many thanks to Deirdre Wilson whose communicated thoughts on these matters have, as always, given me many

(positive) cognitive effects.

a lexically encoded concept and its loosening (leaving aside the echoic or quotational use of

concepts for the time being). Other terms are sometimes used: narrowing is sometimes called

enrichment or strengthening; loosening is sometimes called broadening or weakening. It is these

processes which I want to consider here, particularly loosening or broadening, and their

contribution, if any, to the proposition expressed by an utterance.

2 Enrichment

This term covers a variety of cases, some of which have been discussed extensively elsewhere,

for instance:

(1)

Everyone got drunk.

I've got nothing to wear to the party.

He handed her the scalpel and she made the incision.

The police hit the suspect and she had to go to hospital.

He begged her not to jump.

In the first two examples a domain for the quantifier to range over has to be contextually

inferred, thereby narrowing down the interpretation. In the two conjunction examples, the

encoded assumptions that there is some connection or other between the events is further

specified by the inference of a temporal relation in (1c) and a cause-consequence relation in (1d),

these relations being supplied perhaps by highly accessible general knowledge schemas

concerning relevant ways in which events connect up. Note that in (1c) the second conjunct is

further narrowed by the obvious assumption that the incision was made with the scalpel

mentioned in the first conjunct. Similarly in (1e), a further constituent may be supplied so that,

in an appropriate context, this could be taken to communicate that he begged her not to jump off

the ledge of a high building. I will be focusing on a slightly different type of enrichment in this

paper and won't return to these examples (for detailed discussion of them, see Carston 1988,

1993; Recanati 1989, 1993; Wilson & Sperber 1993).

The examples of primary interest here are of the following sort:

(2)

He wears rabbit.

I want to meet some bachelors.

Mary cut the cake.

She has a brain.

The cinema is some distance from the restaurant.

Something's happened.

What distinguishes these from the previous set is that, rather than adding a conceptual

constituent, the enrichment targets a particular lexical item and strengthens the concept it

encodes. For instance, in (2a) the noun 'rabbit', which encodes something like rabbit stuff , is

narrowed to rabbit fur/skin . One possible narrowing of the bachelor concept in (2b) would take

place in a context in which the speaker had made it clear that she wants to settle down and have

children; then the denotation of the relevant bachelor concept would be a subset of the set of

unmarried men. A crucial component of the narrowed concept would be eligible for marriage .

In the case of (2c), it is not any old severing of the fibres of the cake that would be

communicated in most contexts but rather a particular mode of cutting; comparison with

different objects of cutting makes this apparent, for instance grass , hair , cloth , flesh , etc. The

last three examples have in common that their linguistically encoded content is a truism: all

human beings have a brain, there is inevitably a measurable space between two locations, etc.

Some pragmatic narrowing down is required, of the sort of brain she has in (2d), of the distance

involved in (2e) and of the nature of the event in (2f).

In short, there is a subset relation between the extension of the concept actually

communicated in these examples and the extension of the lexical concept from which it has been

derived, shown schematically in (3), where L is the extension of the lexical concept and C' is the

extension of the narrowed ad hoc concept, the relevant concept in each case.

(3)

Now one of the features of relevance theory which distinguishes it quite sharply from

standard Gricean theory is the view that these strengthenings, of both the types exemplified in (1)

and (2), may contribute to the explicit level of communication, specifically to the propositional

form of the utterance. On the Gricean approach they have the status of implicatures,

communicated assumptions which are independent from, external to, the core proposition

communicated by the utterance ("what is said", for Grice). Relevance theorists favour the former

view because, in many instances at least, these appear to contribute to the truth-conditional

content of the utterance, to what makes it true or false. Various arguments and tests have been

put forward in support of this view (see Carston 1988; Recanati 1989, 1993; Wilson & Sperber

1993). For the purposes of this paper I am going to assume that it is correct, so that the

proposition expressed (and communicated) by (2b) is as in (4), where bachelor' represents the

new narrowed bachelor concept.

(4)

S x wants at t i to meet some bachelors'

It is the interaction of this propositional form with a set of contextual assumptions that will give

rise to contextual effects and some of those contextual assumptions will be derived from the

encyclopedic entry of the narrowed concept, bachelor' , (for instance, that the people in question

should be heterosexual, youngish, interested in marriage, etc). Finally, it should be emphasised

that this narrowing is a local process; it doesn't necessarily follow that the proposition derived

will always be logically stronger than the proposition before that strengthening took place, as is

obvious in the case of narrowings within the scope of negation and certain quantifiers.

3 Loosening

The other, putatively opposite, process of loosening or broadening, is exemplified by the

following, where the loosely used concept is the one encoded by the highlighted lexical item:

(5)

France is hexagonal .

I love bald men.

This steak is raw .

Have you eaten my chocolate heart ?

Here's my new flatmate . [referring to a newly acquired cat]

This relaxing of a linguistically encoded meaning has been pretty much ignored outside the

relevance-theoretic framework, though a general unease with any process of pragmatic loosening

has been expressed. When discussing words which seem to have several related meanings, one

stronger than the other, Grice (1978, 119) says:

If one makes the further assumption that it is more generally feasible to

strengthen one's meaning by achieving a superimposed implicature, than to

make a relaxed use of an expression (and I don't know how this assumption

would be justified), then Modified Occam's Razor would bring in its train the

principle that one should suppose a word to have a less restrictive rather than a

more restrictive meaning, where choice is possible. (my emphasis)

Atlas (1992), who works within a Gricean view of pragmatics, refers to this passage and says:

The "strengthening" assumption can be justified by discovering that there is an

intelligible inference that brings about the strengthening of a speaker's meaning -

intelligible in the sense that such inferences can be formulated and rationalized -

but no intelligible inference that brings about the relaxation of a speaker's

meaning. Loose uses of words don't seem particularly rule-governed.

But, then, what about the examples in (5)? It seems pretty clear that we do not want the concepts

encoded by the lexical items 'heart' and 'flatmate' to include in their extension confectionery in

the one case, nonhumans in the other. The same holds for 'hexagonal', 'bald' and 'raw', though

this might need more argument to convince everyone (not, however, Grice or Atlas, who keep

their semantics as minimalist as possible). Loose use is a fact and has to be accounted for by an

adequate pragmatic theory. With a few notable exceptions, neo-Griceans have tended to steer

clear of it. 1

The standard relevance theory account of loose talk, including metaphorical talk, has

been around for some time (see Sperber & Wilson 1985/6), so I'll give just its bare outline here

by way of reminder. The idea is that in some instances a speaker chooses to produce an

utterance which is a less-than-literal (that is, loose) interpretation of the thought she intends to

communicate. This will arise when she judges that communication of her thought is facilitated

by such a non-literal utterance in that it makes that thought more accessible to the hearer than a

literal one would. The process of interpreting loose uses is as follows: the hearer decodes the

lexically encoded concept, thereby gaining access to certain logical and encyclopedic properties;

he treats the utterance as a rough guide to what the speaker intends to communicate, and, in

effect, sorts through the available properties, rejecting those that are not relevant in the particular

context and accepting those that are, as reflections of the speaker's view. For instance, in the

case of 'raw' in (5c), the definitional property of not cooked would be rejected while the

encyclopedic property of, say, difficult to eat , when applied to meat, would be maintained. The

idea is that the lexical concept raw is in a relation of non-identical resemblance with the

concept that figures in the speaker's thought regarding the state of the steak; that is, they share

some logical and contextual implications. So also for a metaphorical statement such as 'Bill is a

bulldozer', where the lexical concept bulldozer is used to represent the non-lexicalised concept

that figures in the speaker's thought about Bill; it represents it by non-identical resemblance.

As is well known, the relevance theory account of metaphor is very different from the

Gricean account - differences that I won't go into here - and considerably more explanatory.

However,

------------------------------------------------------

1 The exceptions are Bach (1994a, 1994b) and Recanati (1995), who both discuss a range of cases of non-

literalness, within broadly Gricean frameworks. A comparison of their accounts with one another and with the relevance

theory account remains to be done.

there is one respect in which it stays close to the Gricean account, at least in the existing

published work, and that is that utterances involving metaphorical uses of words and, in fact,

loose uses quite generally, do not communicate the proposition they express. The propositional

form is not an explicature of the utterance but just a vehicle for the communication of a range of

implicatures. The same is so for Grice. When he wants to maintain that "what is said" has, as

part of its definition, that it must be meant by the speaker (in his technical sense of speaker

meaning), he moves to a different term altogether in discussing metaphorical utterances. He

writes of "what a speaker makes as if to say", precisely because the proposition literally

expressed in a metaphorical case is not part of speaker meaning; only the implicatures of the

utterance are meant (communicated, in relevance theory terms).

What I want to question here is the prevailing relevance theory adherence to this position.

It is reflected in the upper part of the diagram which summarises the Sperber/Wilson view on the

descriptive and interpretive dimensions of language use:

(6)

The propositional form of an utterance

is an interpretation of

a thought of the speaker

which can be

/ \

..... .....

(Sperber & Wilson 1986/1995, 232)

In this diagram the concept of "interpretation" (or interpretive resemblance) is intended to

accommodate not only literal interpretations but also the cases where a concept in the thought the

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: