For Grice's Sake

By J. L. Speranza
-- for the Grice Circle

IT SEEMS TOME THAT UNNECESSARY (can fuss be otiose?) fuss is made about what or who implicates what or who. To what extent does the mere statement of one's _analysandum_ may involve a categorial mistake.

So, the correct title for this post should be:

What Implicates What? On Trying To Get Rid Of Conversational Implicature's Categorial Mistakes

Let me explain. Much of what many may see as "verbal" disputes in Gricean
theory concern things like "what is the difference between an implicature
and an explicature?", "what is the true taxonomy of utterer's meaning?",
"where does implicature fit in in the utterer's meaning taxonomy?".

Part of the problem is due to Grice himself. In 'Logic and Conversation',
he writes of "implicature", or more importantly (for our purposes, see
below), of "implicate" re the example:

(1) A: How is C getting on in his new job at the bank?
B: Oh, quite well, I think; he likes his colleagues
and he hasn't been to prison yet.

"Whatever B _implied_, or _suggested_, or even _meant_ seems to be
_distinct_ from what he explicitly conveyed, which is, simply, inter alia
(excluding the interjection, "oh", etc) that C has not been to prison yet.
Now, I wish to introduce, as a term of art, the verb "implicate". The point
of the manoeuvre is to _avoid having to choose between this or that member
of the family of verbs ("imply", "suggest", "mean", "hint", etc) for which
my "implicate" will do general duty".

So far so good.

What strikes me as interesting in that passage, and also struck Davis and Saul, is the reference to "mean". "Implicate" is to do genearal duty for "mean" on context.

Now, we're talking Grice 1967 here. But if we flash-back to Grice 1948 --
his essay on 'Meaning' -- we see he was _all_ concerned about the analysis
of this _other_ lexeme, then, "mean", proposing in fact a very
sophisticated analysis of it in terms of "intends".

You'll see that so far, all the verbs Grice is concerned with are what we
may call propositional-attitude verbs. Quoting from S. C. Levinson's book
on _Conversational Implicature_ (MIT), Grice's "analysis" of "mean" in
terms of "intends" boils down to:

(2) By uttering x (x: utterance), U (U: utterer)
means that p iff
i. Exhibitiveness:
U intends A (addressee) to believe
that U believes that p.
ii."Grice way":
U intends A to recognise that U intends that (i)
iii.Anti-Sneak Clause ("Don't Disgrice if you Can Grice"):
There's no inference-element involved
in (2) such that U intends that A rely on
to arrive at a conclusion as to what U means.

Now, it is obvious that if "implicate" is to do general duty for "mean" (on
occasion) and "mean" is a complex form of "intend", then, via substitution,
"implicate" boils down, too, to "intend". Surely, some qualifications are
in order. But "implicate" _is_ a form of _intend_, too.

Now, much of the fuss about the exegetical work by authors like Davis and
Saul, or indeed Levinson, or, say, Schiffer -- in _Remnants of Meaning_ has
to do with what with Grice we could call "expression meaning". Since the
early 1948 attempt, it's pretty obvious that Grice was interested in
providing an analysis of

(3) x means that p.

-- where x is utterance-token (Grice would use "X" (capital x) to mean
utterance-_type_, see Studies, Lecture VI). Philosophers have presupposed
that, qua philosopher, Grice was like "bound" to have an interest in
"expression" meaning, since who but a very neo-behaviouristic "soul" like,
say, Stevenson -- the author that Grice criticises in 'Meaning'-- would _not_?

Now, as Levinson notes, things complicate when we even proceed from (3) to

(4) X means that p.

Since it's _this_, Levinson claims, the real terrain for _generalised_
conversational implicature, which, he claims, is Grice's favoured notion to
deal with a couple of philosophical puzzles. As for (3), one would be
tempted to think that it can easily be _explicated_ (oops) in terms of
utterer's meaning (i.e. our (2) above), via:

(5) x means that p iff
by uttering x, U means that p.

Note that I'm taking a general overview where "means" is doing "general
duty" for, say, both "explicitly conveying something" and "implicitly
conveying something", too. Thus we could have, as per substitution of (1):

(6) "C has not been to prison yet" means
that C is potentially trecherous
iff by uttering "C has not been to prison yet"
U means that C is potentially trecherous.

There _seems_ to be something odd (or "too libertine") about (6) but it is
what follows from Gricean Logic. It's only when we want to speak of the
_type_ of utterance to which the token "C has not been to prison yet"
belongs to that we need a qualification, a la:

(7) A _type_ of the utterance "C has not been
to prison yet" does _not_ mean
that C is potentially trecherous.

since one second's thought will tell us that it's mighty possible to _fail_
if you think that "C has not been to prison yet" (in every occurrence or
realisation) will carry the _meaning_ (or "mean") that C is potentially
trecherous. Here is where truth-conditions probably fit in. Tapper is not
very impressed with (8), but I'd submit Grice would submit:

(8) "C has not been to prison yet" means
that C has not been to prison yet.

(via Davidson's Disquotation). So, let's now turn to the issue of
"taxonomies". Saul considers this in her essay with _Nous_ ('Utterer's
meaning and what is implicated'):

On the version of Grice's theory which
I was taught, "U's conversationally
implicating" is a _species_ of the _genus_
utterer's meaning, to be analysed in terms
of U's intending, and utterer's meaning
divides exhaustively into "U's explicitly
conveying/expressing" and "U's implicating".
This is a common understanding of Grice, and a
natural one, given Grice's obvious interest
in utterer's meaning. A few examples will provide
some indication of the popularity of this reading.
Neale attributes this view to Grice on pages 73-83
of Neale 1990 and in Neale's essay in Ling/Phil.
In the latter, there are two diagrams which show
"U's meaning" dividing exhaustively into
"U's explicitly conveing" and "U's implicating".
The diagrams make the first level of division between
"U's _conventionally meaning" and "U's non-conventionally
meaning", but the division between "U's explicitly
conveying" and "U's implicating" is still exhaustive.
Horn ('The Said', p.165) also provides a diagram
showing utterer's meaning dividing exhaustively into
U's explicitly conveying and U's implicating as does
Levinson [on p.13 of his _Conversational Implicature_].
For both Horn and Levinson, the _first_ level of
division is between U's explicitly conveying and
U's implicating. Bach also apparently takes this
to be the received view of Grice as he argues for
an additional category, U's conversationally
impicIting" in his Language/Mind essay

In Levinson's book, the diagram is expressed in non-nominalistic terms as
follows:

(9) Genera and species of meaning-nn:

TOTAL SIGNIFICATION of an utterance
.
.
. .
. .
what is said what is implicated
.
.
. .
. .
conventionally conversationally
.
.
. .
. .
generalised particularised


And he does this as per an exegesis of Grice -- i.e in the section, 1.1.
Grice's program --, an exegesis which will guide his whole approach.

So, since I'm supposed to conclude something (cfr. Grice's Aspects of
Reason -- we are such _concluding_ critters) I'll say: a nominalist may
stay at the level of what an utterer means (in terms of his intentions).
This is what Davis does in his book. Since he feels Grice would _not_ buy
his (i.e. Davis's) neo-Gricean utterly nominalistic programme, Davis
subtitles his book, _The Failure of Gricean Theory_. Myself, I'm not so
sure that is a nicely implicating title...

I hope there's room for discancellation, too.

References:

DAVIS WA. Implicature: the failure of Gricean theory.
Cambridge Studies in Philosophy.
GRICE HP. Meaning. 1948. Repr. in Studies.
. Logic and conversation. Repr in Studies
. Studies in the Way of Words. Harvard.
HORN LR. The said and the unsaid.
Semantics & Linguistic Theory, II.
LEVINSON SC. Conversational implicature. MIT.
MURPHY MJ. This Forum.
TAPPER LM. This Forum.
SAUL J. Review of Davis. Nous.
SPERANZA JL. (Still) Join the Grice Circle.
WALKER R. Conversational implicature. in S Blackburn,
Meaning, reference and necessity. CUP.
WRIGHT R. Meaning-nn and conversational implicature.
In P Cole & J Morgan, Speech Acts. Academic.