Supervenient Grice

--- by J. L. Speranza
------ for the Grice Club, etc.

READING all this innovative material by R. E. Dale can be inspiring. It brings back discussions of central interest in what possibly is the 'background' of Grice's theory (or anaysis, as I sometimes prefer) of meaning: philosophical psychology.

Dale writes in his commentary on Markerese:

"Suppose we reduce expression meaning to regularities of some sort of Gricean notion of speaker meaning with in a community."

BIG supposition! Oddly, I once participated in a colloquium, if you can believe it, with Searle and Davidson and Biro and who knows who else! What I did (it was a conference in such a big venue that nobody really knew what the others were talking about) all I expanded on was on Grice's account of 'expression meaning' (I think I used the expression -- the thing was published in the proceedings) in terms of value in its 'optimality' guise.

I was trying to be innovative and explore this 'mystery package' in "Meaning Revisited" where he notes that

expression meaning

is best seen as that an expression optimally means. He concludes that section with a bit of a Platonic diversion where meaning comes out as an 'ideal' concept. I loved Grice for that.

---

But, yes, suppose otherwise. Dale indeed surveys all the convention-based accounts of expression meaning, including those which are not perhaps really 'convention-based' but use an interesting 'actual-language' relation. I read and re-read Grice and -- even his very early papers now repr. in WoW, with examples like:

"Bring me a paper tomorrow"

thereby meaning that it is raining (WoW).

-- and it's difficult to see Grice as aiming at expression meaning seriously. In WoW:VI he talks of 'idiosyncratic procedures' I think he calls them. Within the idio-lect, as it were. And in "Meaning Revisited" it is again about Deutero-Esperanto.

(INTERLUDE with Schiffer. I checked the first page to his "Actual-language" relation paper, and he starts by considering Esperanto -- "a language nobody speaks". I think Grice -- and indeed Lady Welby -- Volapuek? -- would disagree. In any case, for Grice, even "Deutero-Esperanto" posits problems that any account of expression-meaning, he thinks, must provide for).

Dale goes on:

"Now, if what we are looking for is a complete physicalistic reduction of meaning, what is left is to give an account of speaker meaning in terms of physicalistic notions."

I agree, and enjoyed Dale's discussion of that book by Avramides dwelling on Grice's distinction between

reductionist

and

reductive.

--- (I am amazed at how Avramides could fill her full DPhil under Strawson on such an abstract issue!)

--

THIS SAID, it was through reading the final section in "Method in philosophical psychology" -- in "Conception of Value" which I was able to quote in the reference list to my PhD dissertation -- that I was able to detect Grice's deep antipathy for physicalism.

He even speak of the Devil of Scientism, to boot, in that section (a remark whose tenor he confesses he got from Myro).

That section is all about nomological irregularities. The idea that a physicalist-level (to which, allegedly, the psychological-level reduces) will ACCOUNT or 'save the phenomena' seems pretty outrageous to Grice. This he delivered in 1975 -- Presidential Adress, APA --. The type of eliminationism alla Churchlands was not something that Grice would render coherent, I would think.

Dale continues:

"So, since our account of speaker meaning is Gricean, it contains all sorts of intentions notions (intention, recognition, belief). These now have to be reduced to physicalistic notions. That is, a physicalistic account of propositional attitudes must be provided. The Language of Thought (LOT) hypothesis of Fodor is one way to do this (though note that I argue in my dissertation that the LOT hypothesis can be construed in such a broad way that its truth can be seen as a mere platitude; it can be seen as trivially true). But, the LOT hypothesis only takes us from public language expressions to the LOT."

I was fascinated when I got hold of Block's popular work in the philosophy of mind, and found out:

(i) that he credits Grice's "Method" as a functionalist credo. In seminars I gave I called it Grice's 'functionalist' gospel, even.

(ii) that he cares to reprint an old chapter form an old book by Geach, "Mental acts" where one sees that all was basically said, in a language I oddly find clearer, by Occam -- sermo interioris. He said some pretty genial things, this Occam. E.g. Consider Carnap on

Homo habet canem.

(discussed by Carnap, Aufbau, discussed by Dale). Now, Occam is saying that 'sermo interioris' can't contain case-markers! This would mean that the type of 'general fucntion' Carnap is attempting to provide would be unphilosophical. Qua grammarian one may care for that, but qua logician one SHOULDN'T, Occam held).

Dale continues:

"It doesn't show how content in the LOT is physicalistically implemented. That is, the LOT needs to be interpreted, and interpreted in a physicalistically reductive way. So, the LOT can be seen as a version of markerese by itself, at least in the sense that it is merely an uninterpreted set of representations."

I wonder what Dale makes, since I did, and a lot -- of Grice's VERY brief and schematic 'general theory of representation' in Retrospective Epilogue. I think it contains three or four tenets. I was always fascinated by it. Grice is clear that he seems to be having Peirce all back again to him. He notes, Grice does, that the prior type of representations are always iconic (natural). One may think that a physicalist reduction of an alleged 'content' of a psychological attitude will NOT be iconic like that.

I suppose that reductionism of the type that Grice disfavoured (in "Conception of Value") and connectionism go hand in hand -- which is NOT Grice's hand, but, he would say, 'The Devil's (The devil of scientism)!

---

Dale goes on:

"Merely translating public language to the LOT, by itself, does not achieve the reduction we wanted. So, the LOT could be markerese for all we care. I believe both Fodor and Schiffer both make such a point in one or more places. Perhaps others do as well. Mind you: so-called "markerese" as introduced in Katz and Fodor (1963) was intended to do a number of things that Fodor's LOT is NOT AT ALL intended to do, like account for expression synonymy, ambiguity, and so forth (see Katz and Fodor (1963) where they give a long list of the semantic properties that they claim can be accounted for with their "semantic markers" (what came to be called "markerese"))."

Excellent point.

And yes, I think they are right in thinking that things like 'synonymy' ("means the same as", in Grice's colloquial rephrase in "Meaning") and perhaps 'entailment' (rather than 'ambiguity', but I would love to discuss that) look like desiderata in any semantic theory worth its name.

I wonder if Dennett, who can ALSO be described as a Griceian, may not provide an alternative here to the type of 'content-ful' type of LoT that Fodor is advocating, incidentally.

Dale continues:

"So, be careful on this point: the LOT is not simply the same thing as "markerese". My point here is merely that, by itself, without an physicalistically acceptable interpretation, it may as well be markerese, that is, an uninterpreted set of representations."

Point taken, thanks.

"Markerese needs to be interpreted, and interpreted in a physicalistically meaningful (reductive) way. Now, when I say "reductive" (and cognates) above, I play a bit fast and loose. There are less strenuous notions that might satisfy our physicalistic motives than complete reduction. Supervenience is the the main alternative concept. And it is the supervenience concept that places the central role in understanding this issue in Schiffer's paper "Does Mentalese Have a Compositional Semantics?" in Loewer and Rey (1991), Meaning in Mind: Fodor and His Critics."

-----

I forget what notion plays a similar role in a favourite with me: Loar's Mind and meaning (Odd that Avramides will have a book with the same truth-functional title). I for one admire Loar bunches, and am very motivated to criticise him when, to stretch his analysis of things, refers to something as basic as a Gricean conversational maxim such as

be truthful

as a 'contigent generalisation over functional states' -- footnote! I wouldn't think Loar is committed to supervenience. I tend to interpret him as a 'naturalist' functionaist, if such a thing exists!

Dale goes on:

"So, "markerese" is not, as such, picked up on within the Gricean tradition. There is no intrinsic need for it, or for an account of synonymy, etc., as Katz and Fodor were seeking in 1963."

True. NOW:

Grice mentions three notions in that 'gloss' after 'timeless meaning':

ENTAILMENT, as we saw. Which is a bit of a stretch. Why should what we explore on a seminar on "Meaning" account for a rather artificial notion coined by G. E. Moore?

SYNONYMY (we agree)

and...

UNDERSTANDING. Here I think it was just natural then that Strawson picked THAT up, when in "Intention and Convention in speech acts" he just defined, and rightly, "A(ddresee) understands U" if "A gets to know what U means"! Better said that done!

---

Dale goes on:

"But, we can speak of [markerese] vaguely re-entering the discussion as the LOT, that is as mentalese, insofar as the LOT is part of an account of the intentional or propositional attitude notions that are at the foundation of a reductive Gricean project."

In Reply to Richards, Grice approaches one type of answer to Grandy/Warner's point that Grice is taking 'p' for tranted:

U means that p iff among other things U intends A to believe that U belives that p.

For all we know, 'p' may just be a gap sign. Let me quote from Grice. Because unless you realise that Grice never found it much of a problem -- to give a principled account of 'p' -- one may be diverted from issues that interested Grice. I think it is Schiffer in "Meaning" that refers to a 'loop'.

The paradox, as I love to state it, would be: the semantic reduces, after all, to the psychological. So, it would be more than otiose to think that an account of the psychological (compositional semantics for mentalese, say -- whatever that is) should rely on public semantics!

This is Grice, then:

"A perfectly sound ... reply", Grice writes, "Reply to Richards", p. 74,

"to the objection as it is presented would be taht in any definition of meaning which I would be willing to countenance, the letters 'p', 'q', etc. operate simply as 'gap signs' if they appear in a definiendum they will reappear int he corresponding definiens. If someone were to advance the not wholly plausible thesis that to feel F is just to have a Rylean agitation which is caused by the thought that one is or might be F, it would surely be ridiculous to criticise him on the grounds that he had saddled himself with an ontological commitment to FEELINGS, or to _modes of feeling_. If quantifiers are covertly involved at all, they will only be universal quantifiers which would in such a case as this be adquately handled by a substitutional account of quantification. My situation vis-a-vis propositions is in no way different."

--- Part of the graciousness I find in Grice's discussion there is reinforced by my having sort of digested, say, Martinich's response to M. Black's anti-Gricean critique (in "Literary History"). The account of Gricean utterer's meaning is indeed psychologically convoluted. But much of it does not touch on "'p', 'q', etc", as Grice puts it above, which are supposed to be, ultimately, and to echo Schiffer, 'the things we mean'!

---

Dale goes on:

"That is, markerese can be seen as re-entering the conversation as mentalese (the LOT). Of course, all this is stretching things."

It's not! (My motto: if it does not stretch, it is not philosophy -- _or_ breaks).

Dale goes on:

"As I said, markerese was meant to fill a certain function that simply doesn't exist in the Gricean program."

Grice is not addressing synonymy, say, in "Meaning" ('means the same as') but indeed, he would go on to say, apparently, that 'x' means the same as 'y' if the timeless meaning for both x and y co-incide, i.e. in terms of the contents of the communicative or m-intentions involved.

Dale notes:

"All markerese and mentalese share is that there is a perspective from which they both can be looked at as uninterpreted representation systems, or, better, systems requiring an account of how they represent. It is probably best to let sleeping dogs lie with respect to markerese. But, since Speranza asked the question, I thought I would address how it might be possible to find a place for markerese in the Gricean program as a whole."

Thanks!