Strand 5

By RBJ for The Grice Club

Speranza has made much lately of Grice in his strand 5 on
"central meaning".

I would like to engage, but feel the need to start at an
earlier point in Grice's exposition.

But before even that I feel the need to comment on the
relations between:

1. Forms of Judgement
2. Propositional attitudes, intensional contexts
3. Pragmatics

In Frege we find that an assertion is something which is
formed by applying the (vertical) judgement stroke to some
proposition formed by applying the content stroke
(horizontal) to some expression.
i.e. assertion is not something which a proposition does, it
is something which is done to a proposition in a judgement.

After Frege this is not followed very closely, his two
strokes combine to yield the turnstile, which is used to
mark a formula as a theorem, or as the relationship of
syntactic consequence, or as a sequent constructor in
sequent calculi. Forms of judgement become more complex in
constructive logical systems, but I won't go into that.

Though assertion is more often thought of as something done
to a proposition than done by a proposition, propositions
can also do things to other propositions. We will disregard
the truth functional cases, for they really operate only on
the truth value not on the proposition. The cases where the
proposition itself is involved are those in which
substitutivity of extensionally equivalent propositions
(i.e. ones with the same truth value) "salva veritate" may
fail, and we say that the context in which the proposition
appears is opaque or intensional. The best known of these
are where a propositional attitude is expressed (hope,
believe, assert) and those involving modal operators
(necessarily, possibly, eventually).
In these cases we get propositions which are about
propositions, and semantics has something to say about the
matter.

When we do things with propositions other than thus
incorporating them into some more complex proposition, as in
forming a judgement or asserting or proving them, we then go
beyond semantics into pragmatics.

In Strand 5 of his Retrospective Epilogue (the valedictory),
Grice uses a great variety of novel terminology as well as
some perhaps novel usage of more traditional terms.
By doing this he gives us multiple perspectives on the
distinctions he is hoping to clarify. This is gives us a
richer collection of hints about what he is after, and our
problem is then to understand what he is looking for.
In the course of doing this it may be helpful to explore the
connections between his terms and others which are more
commonly used.

At the very beginning, then, when Grice identifies the
strands but before he comes to discuss them, we have already
two distinctions each articulated in ways which we might
better connect with the terms which I have touched upon
above.

The distinctions are here presented as kinds of meaning:

1. Conventional / non-conventional meaning
2. Assertive / non-assertive meaning

where the "asserted meaning" is understood as what is
asserted, either by the words or the user.
"user" here connects us back with pragmatics, with Carnap's
definition.

There are issues I'd like to raise here about the use of
"meaning", of "asserted".
My own natural reaction is to consider meaning to be at
stake only on the one side of this distinction, and so to
think that Grice's usage of the term here is broader than I
would enthusiastically endorse.
Possibly Grice also has such reservations, because as we
move from the headline to the details Grice drops "meaning"
and talks about "significance". That's good for me, I am
happy with allowing "significance" to be broader than
meaning.

So my inclination would be to talk instead of two kinds of
meaning, of the distinction between meaning (which we might
possibly identify with the conventional) and some broader
notion of significance. In that case I would also be
inclined to characterise the distinction as connected with
the difference between what is asserted, which I might also
call the (explicit?) content, and the significance of its
being asserted in the way it was in those particular
circumstances.

What I think I am doing here is not suggesting that we
should be concerned with some other distinctions than those
which Grice puts forward, but rather than it might help to
describe those distinctions in some other way.
Two questions arise from such a suggestion.
The first is whether my rewording is faithful to Grice's
intentions (words apart) and the second is whether it is any
kind of improvement.

To be explicit, I am suggesting this third initial
characterisation

3. the meaning of the sentence asserted /
the significance of its being asserted
(in the particular context and manner)

and I note that when put this way, the distinction seems
more likely to correspond to that between semantics and
pragmatics.

This does not gel well with the idea that Grice is here
looking for "central meaning", but in fact he seems to talk
mainly about "central significance".
It might be thought to trivialise his problem, for we offer
the idea that "central significance" is just "meaning", but
this is an illusion, for the same problem remains in
different words. The problem then is, how to we distinguish
meaning from other aspects of significance.

The reformulation may just be different words, but the change
of words suggests possibilities which are less conspicuous
in the original wording. They suggest that "central
significance" should be associated exclusively with what is
in the proposition asserted rather than in its context and
manner of assertion. This arises from our having coupled
the terminology with that which defines the boundary between
semantics and pragmatics.

Possibly this is not the line which Grice is after.
If in fact Grice is looking for a narrower conception of
"central significance" then this broader notion of
significance (what we count as belonging to semantics)
provides a domain in which finer discriminations might be
facilitated.
Possibly Grice is looking for something as narrow as, or
closer to, that of "truth conditions", which are arguably
just a part of a reasonable conception of meaning.

Three conceptions of meaning now seem to me relevant to
Grice's discussion.

1. The meaning of a sentence
2. The proposition expressed by a sentence asserted in some
particular context
3. The truth conditions of a proposition.

The first seems to me the best candidate for what is
conventional, which is Grice's headline criterion.
This notion of meaning is one which encompasses all the
propositions which might be expressed by the sentence in
various different contexts, and hence one might argue
speakers meaning for every speaker of the language.

Once we get into the detail we find Grice seriously
interested in the candidacy of propositions, which in the
present context we may describe as the meaning of a sentence
in context.
These are the candidates for the status of "Eternal Truths",
and they have strictly less to them than meanings of type 1,
but strictly more than the mere truth conditions.
Grice talks of the essential characteristic of propositions
as having a truth value, but this is not enough, they need
to have truth conditions. Its not enough for a proposition
to have a truth value, it must have one in every possible
situation.

Context comes into this story in two places, and this is
reflected in some controversy about the nature of pragmatics.
The context of an assertion plays a role in determining the
proposition asserted, and its role in this belongs to
semantics and this encompasses what is entailed by that
proposition (to which only the truth conditions pertain).
There may be significance to the assertion which goes beyond
what may be found in the proposition asserted.
This belongs to pragmatics, and implicature belongs here.

So the question I ask now is, how far have I now diverged
from what Grice intended, or from what Grice might have been
persuaded to accept?

RBJ