Pervasive metaphysics and other Grice/Carnap tensions

Roger Bishop Jones, for The Grice Club

I don't know why I should have been surprised when Speranza pointed out the pervasiveness of metaphysics in Grice's philosophy, for surely every aspect of our language engages with its own special kinds of entity and thus involves some special metaphysics.

Naturally I ask the question whether this deals a serious blow to the prospects for our Grice/Carnap dialogue.

The simple answer is: "not at all!", for surely this is what Carnap's principle of tolerance is all about, freedom to use languages irrespective of whatever ontology they presuppose (though I don't think that way of putting it is Carnap's), subject only to pragmatic questions (does it serve any purpose?), not the meaningless metaphysical "external questions".

This kind of response, however, leaves us with a puzzle.  If Carnap's positivism is so very accomodating, what is left of his rejection of metaphysics?  Is this something which just melted away?  On the other hand, can we be sure that the pervasive metaphysics in Grice is entirely concerned with questions which Carnap would recognise as "internal", as we might suppose by considering Grice's methods.


I revisited some of the milestones in Carnap's writings on Metaphysics to clarify my thinking on the first of those puzzles, and it is interesting to see that this very question (about what is left of Carnap's proscription of metaphysics after "Empiricism, Semantics and Ontology") is answered in the statement of his position on metaphysics in the Schilpp volume.
Evidently other people had wondered, after "Empiricism, Semantics and Ontology" whether Carnap's objections to Metaphysics had not just dwindled into nothingness (Carnap is responding to Beth).

Carnap's answer there is interesting because it presents a new aspect of what he had introduced as the internal/external distinction.

In relation to Grice's metaphysics it is easy to imagine that as an ordinary language philosopher all his metaphysics is based on the analysis of ordinary language, and that the ontological and conceptual analyses thus obtained are all "internal" in Carnap's terms and hence unobjectionable to Carnap.  Of course, Carnap might have preferred the topics be addressed in formal languages, and then yield necessary conclusions, and would have considered the study of ordinary language to belong to pragmatics (though I myself think him mistaken to exclude the possibility of a semantic study of natural languages, even though the results would be synthetic).

However, the distinction between internal and external questions is particularly difficult when it comes to natural languages, for all Carnap's examples of external questions are couched in natural languages.  Natural languages are often their own metalanguage and the distinction between internal and external is muddied.  External questions are often put using the word "real" or "really".
Do these kinds of entity "really" exist? (Carnap often exemplifies external question using talk of "reality") This works fine when we are talking about the ontology of a formal language and the talk of "reality" is confined to the metalanguage.  But in a natural language, talk of "reality" is internal and all those external questions are internalised.

So its helpful that in the Schilpp volume Carnap gives us an alternative vocabulary for talking about what he means by "metaphysics" and which he still rejects.  As it happens this has been my preferred way of thinking of this for some time.  The alternative account distinguishes between ontological claims which are relative to some language, and those which purport to be absolute, and on this account Carnap's rejection of metaphysics becomes a rejection of absolute ontological claims.  Relative (aka internal) claims (relative to some language) are to be settled by the "rules" of the language, yielding answers only for specific languages which may disagree among themselves (between different languages).

With this clarification in mind we may ask again how Grice's work might have appeared in Canap's eyes.  The question then becomes, is Grice's interest in metaphysics exclusively relative to some language (presumably English), or does he get into more absolute questions?

That's a question for JL perhaps, but I shall speculate a little myself.

It seems to me that Grice's species of ordinary language philosophy is not so exclusively concerned with the analysis of language as would be needed for there not to be an issue here.  For Grice "ordinary language" is not an exclusive subject matter, but rather an ubiquitous source of insight.  In at least some of his metaphysical enterprises the object of his studies does not seem to be language.
In some cases the point might be exegetical, he might be excavating the metaphysics of philosophers (Aristotle perhaps).  This would not fit Carnap's narrow conception of philosophy, but at least it would not be the proscribed metaphysics.

I think however, that even when considering, say, some aspect of Aristotelian metaphysics, Grice is not purely, or even primarily. exegetical, rather he seeks to take up and progress some aspect of the problem which Aristotle was addressing.  Aristotle is of course one of the early sources of just those metaphysical "pseudo-problems" which Carnap criticised, and so it seems likely that Grice may well in this way find himself crossing Carnap's line.

Why should we care whether Carnap would have found the problems which Grice addressed genuine?  What was the point of fantasising about a conversation between these two philosophers.

There is more than one, but in this domain it seems to me that what is happening is that we are stripping away those kinds of metaphysics to which it is easy to give meaning (say, descriptive metaphysics, or the exegesis of Aristotelian metaphyics) and which for that reason do not fall foul of Carnap's critique, and when focussed down on the real metaphysics (perhaps what Strawson called "revisionary" metaphysics, perhaps only a part of that), we can imagine Carnap challenging Grice to give meaning to the enterprise.  To the extent that Grice succeeds in doing so, the scope of Carnap's critique would be narrowed.

There is interplay here between method and meaning.  Carnap wants to see a definite meaning for a metaphysical claim, because in default of that we can have no idea how it can be verified (using that term loosely).  Conversely, if we could say in what way such claims could be verified or refuted then those methods would suffice to give meaning to the claims.

The effect of the dialogue is to extract from Grice more detail about meanings and methods, and from Carnap consequent narrowing of the scope of critique.

Further effects might be hoped for.  From Carnap it seems to me one might hope for two further kinds of concession.  The first is in the use of the term "metaphysics", which for Carnap is used exclusively in a perjorative, proscriptive way.  We could reasonably hope that he might be persuaded to accept a wider use ot the term which embraced questions which he does not consider meaningless, e.g. to encompass descriptive metaphysics.  We might suggest perhaps in the first instance that Carnap reserve the term "absolute metaphysics" for the external questions which he regards as meaningless and allow that internal ontological questions (especially ones internal to natural languages) be spoken of as a kind of meaningful metaphysics.

A second concession which might be easy to extract is the acknowledgement that meaningfulness is not discrete, that it is the business of philosophers and particularly of metaphysicians to probe into just those areas where meaning is hard to grasp, and that one should perhaps in metaphysics accept more a more tenuous grasp on meaning that one might hope for in say, arithmetic.

The concessions here, in relation to metaphysics, seem all on Carnap's side, Grice's part, clarification of meanings and methods seemingly just more of what he is ordinarily engaged in.   It is in the dogmas that we seek concessions, and we have been talking here about Carnap's anti-metaphysical dogma.
The place for Grice's concessions is in his own dogmas, which is what I am here calling his Betes Noire, the various aspects of "minimalism".

There is a symmetry here, for the dogmas of Carnap and those of Grice are both anti-dogmatic.  Carnap rejects external questions as criteria for the acceptability of languages, because he wants to be tolerant about language forms.  Grice rejects minimalism for similar reasons.  He construes minimalism as a set of nominalistic dogmas and he doesn't like being deprived of any of the ontology implicit in our language.  Carnap's minimalism is however a pragmatic rather than a dogmatic enterprise.  Our conversation will progress more fruitfully if Grice(*) would recognise that not all minimalism is abhorrent.

RBJ