Grice in the Lab

by J. L. Speranza
for the Grice Club

R. M. HARNISH, a philosopher, reviews Grice's 'experimental' work in the Philosophy in Review 30 -- "Semantics and Pragmatics: From Experiment to Theory", a new book with Palgrave Macmillan.

The story is the usual one.

Not long ago, Humbolt University, in Berlin, held a conference on experimental pragmatics, which formed the basis of this volume.

The editors wisely chose between invited talks and submitted papers, organizing the contributions around three parts:

I. Implicature.

II. Negation, and

III. Presupposition.

The result is a much more coherent volume than typical conference proceedings.

As the editors note, the three topics are conceptually connected:

--- implicature (scalar at least) are sensitive to negative polarity items and presupposition, unlike entailment, is typically invariant under negation.

And both presupposition and implicatures interact with projection and compositionality.

Part I: ‘Implicatures’.

M. Krifka,
Implicatures.
(ch. i)

Krifka provides an excellent brief introduction and overview of implicature work, concentrating on the study of so-called scalar implicatures.

The four remaining chapters, as well as ch. viii of Part II, review experiments related to either

(i) the categorization of material as a part of what is said or what is implicated

M. Larson, R. Doran, Y. Mcnabb, R. Baker, M.Berends, A. Djalali, and G. Ward

Distinguishing the said from the implicated
using a novel experimental paradigm.

ch. 5

-- or

(ii) uncovering evidence for or against various theories of the mechanisms underlying the comprehension of implicature.

These latter discussions, dealing with implicature, include

L. Bott.
Changes in activation levels with scalar implicature.
(ch. ii)

P. Hendriks, J. Hoeks, H. De Hoop, I. Krämer, E-J. Smits, J. Spenader and H. De Swart;

A large scale investigation of implicature
ch. iii.

--

N. Katsos,

Evaluating under-informative utterances with context-independent scales: experimental and theoretical implications.

(ch. 4)


A. Zondervan,

Experiments on QUD and focus as a contextual constraint on
scalar implicature calculation.
(ch. vi)

Part II: ‘Negation’.

I. A. Noveck

Meaning and inference linked to negation:
an experimental pragmatic approach
(ch. vii)

surveys some relevant work on negation.

Mainly this has to do with the issue of the scope of negation raised in such sentences as

The King of France is not bald.

Does ‘not’ have narrow scope, i.e. over just the predicate ‘bald’, or wide scope, i.e. over the whole sentence?

Or is it indeterminate until uttered in a context?

Noveck also reviews the contributions to the present part of the
volume which are

A. Bezuidenhout, R. Morris, and C. Widmann
The De-blocking hypothesis:
the role of grammar in scalar reasoning
(ch. viii)

A. Gualmini

Experimental pragmatics and parsimony:
the case of scopally ambiguous sentences containing negation
(ch. ix)

B. Kaup,

How are pragmatic differences between positive and negative sentences captured in the processes and representations in language comprehension?
(ch. x)

J. E. Drury & K. Steinhauer.
Brain potentials for logical semantics/pragmatics
ch. xi

Part III: ‘Presupposition’.

Sauerland’s

Presupposition: From Theory to Experiment
(ch. xii)

reviews some linguistic points regarding presupposition, relates that notion to the notions of ‘common ground’ and ‘accommodation’, and previews the contributions to this part.

These contributions are

D. Heller, D. Grodner, and M. Tanenhaus,

The real-time use of information about common ground in restricting domains of reference
(ch. xiii)

E. Chemla

An experimental approach to adverbial modification
ch. xiv

N. M. Klein, W. M. Gegg-Harrison, R. S. Sussman, G. N. Carlson, and M. K. Tanenhaus,

Weak definite noun phrases: rich but not strong, special but not unique.
ch. xv.

and

J. Berkum,

The neuropragmatics of “simple” utterance comprehension:
an ERP review’
(ch. xvi)

***********************************************************
How is any of this relevant to the philosophy of language?
***********************************************************

Good question.

The above studies contain mostly linguistics and psycholinguistics.

Linguistic analysis is typically language internal and concerned with linguistic units, levels and principles.

Psycholinguistics is typically concerned with the acquisition and processing of those units.

Philosophy of language, on the other hand, is typically concerned with word-world
relations, such as are involved in reference, truth, felicitous speech acts and implicature.

In these areas the world and beliefs about the world loom large.

There is, of course, overlap, but so far less on the experimental side.

Implicature and presupposition, at least, are topics one can legitimately cover in a philosophy of language course, though the time-course of understanding, and the mental representations and memory mechanisms deployed are not legitimate.

(Unless you hold one of those combined posts -- "philosphy and linguistics")

Grice introduces the concept of implicatum and implicature into the philosophy of language -- back in 1966, "Logic and Convrsation", the Oxford lectures -- primarily as a way of deflating some philosophical analyses, as he had presented in "The causal theory of perception" (1961) and even earlier in discussion with P. F. Strawson in "Introduction to logical theory" (1952).

According to Grice, the philosopher would suggest
that concept X involved concept Y (or that
a term is ambiguous between X and Y).

But Grice provides an alternative explanation
for the intuitions behind these suggestions
that involved ‘pragmatic’ inferences rather
than ‘semantic’ structures.

Grice’s proposal was broadly speaking epistemological.

We might call it ‘linguistic epistemology’, as it provided an alternative justification for a class of intuitions.

It was not officially psychological, though some of so-called "successors" quickly took his proposed inferences to constitute psychological mechanisms recruited in communication.

One still sees remnants of this conflation in the literature and in the papers in Part I ("Implicature") of this collection.

However, Grice did appeal to intuitions regarding what was said, meant, and implicated, even though his theory recognized how hard it could be for intuition to separate these factors.

Chapter 5 seeks to supplement introspection with experiment.

Larson et al. first review the weaknesses of earlier categorization experiments, e.g., one-sided diet of materials, confusing subject instructions, and theoretical narrowness.

They then propose to answer three related questions.

i.

Can speakers access a level
of meaning corresponding to
the Gricean notion of what is said?

ii.

Do different types of generalized conversational
implicatures behave similarly with respect
to truth-conditional meaning?

iii.

Can previous methodologies be improved on?

Their experiments used 4 blocks of 22 stimuli for each of 72 subjects.

The results suggest that, yes, subjects do have access to something like truth-conditions as opposed to generalised conversational implicature.

And, yes, different generalised conversational implicatures do behave differently.

Finally, by distinguishing an explicitly ‘literal’ observer (‘Literal Lucy’) the authors were able to magnify the resulting chances of getting truth-conditions separated from generalised conversational implicature, improving on previous methodologies.

Grice also distinguished between particularized conversational implicatures, which
depend on the particularities of the context of utterance, and generalized conversational implicatures, which tend to go with the words uttered and are carried in a wide variety of contexts.

Katsos explores aspects of the experimental side of this distinction.

Katsos's main purpose is to ‘explore whether the distinction between generalised and particularised conversational implicature is psychologically real, and in particular whether it is reflected in the pattern of child language acquisition’ (34).

Katsos is also interested in the consequences for the dispute between ‘unitary’ and ‘default heuristic’ pragmatic theories.

Katsos finds experimental evidence that children do not favor context-independent scalar implicatures over context-dependent ones, consistent with unitary theories, though adults reject (with revisions) false context-independent implicature more than a false context-dependent one, in accordance with default theory.

Katsos suggests we may have both by assuming that context-independent implicature are more frequent, give rise to a privileged psychological status, and thus are more prone to revision and rejection when false.

As for presupposition, which Grice clearly saw as subsuming under implicature (vide "Presupposition and conversational implicature", in WoW) Frege, and later Strawson (Grice's pupil at St. John's, Oxford) introduced the idea that (as least some) sentences with singular terms presuppose rather than entail—some say assert -- the existence of their purported referent, in the sense that the falsity of the presupposition precludes the sentence from having a truth value (while the falsity of an entailment falsifies the original sentence).

This notion, familiar in the philosophical literature, does not make, alas, an appearance in Part III ("Presupposition").

Rather, what does appear is variously called ‘linguistic’ or ‘pragmatic’ presupposition.

The discussion is, therefore, often confusing.

For example, in Chapter 12 Sauerland gives

Sue knows that Joe is asleep.

as presupposing

Joe is asleep.

while entailing that

Sue believes that Joe is asleep.

presumably on the ground that the negation

Sue doesn’t know that Joe is asleep.

also carries the ‘presupposition’ that

Joe is asleep

(219).

But for philosophers

X knows that p.

typically entails both

p

and

X believes that p.

and the supposed presupposition would be explained by the scope of negation, and maybe conversational implicature.

The upshot of this discussion is that experimental work on semantics and pragmatics
impacts philosophy mainly either when philosophers rely on linguistic intuitions for data, or appeal to particular psychological mechanisms in explanations.

This volume is a good survey of the state of the art at present with respect to both intuitions and mechanisms.

But serious concerns appear when we take Grice _out_ of the lab (for a cuppa char, or something).