---- By J. L. Speranza, the Grice Club,
------------ for the Carnap Corner.
---- TO CONSIDER SOMEWHAT seriously, from an online source: "Beall and Restall [2000], [2001] and [2006] advocate a comprehensive pluralist approach to logic, which they call Logical Pluralism, according to which there is not one true logic but many equally acceptable logical systems. They maintain that Logical Pluralism is
compatible
with
monism about metaphysical modality."
-- Restall, whom I know, is explicitly neo-Carnapian. Not oddly, Grice has defended monism about metaphysical modality too (i.e. that necessarily 'must' and analytically 'must' must be the same must) in what he calls the Aequi-Vocality Thesis.
I think it should be pretty serious to consider here the epithets:
logical
vs.
ontological
cfr.
physical
vs.
metaphysical
--- &c.
And also the
monism-pluralism distinction.
Monism since strictly the possibly only good alternative to 'pluralism' of the sort that Carnap defends explicitly elsewhere. To consider as demons, perilous and malevolent (betes noires) or other.
Subscribe to:
Post Comments (Atom)
I have taken issue with Restall (who I met in Cambridge a few years back) over his claim that he is more pluralistic than Carnap was, which he says because he admits that a language might have more than one deductive system. I pointed out that he admitted no possibility excluded by Carnap but possibly described matters differently.
ReplyDeleteInstead of talking about one language with two inference systems, Carnap would probably talk of two languages with the same grammar, but exactly the same linguistic/logical possibilities would be acceptable.
At the time Greg seemed to accept this, but he has nevertheless continued to make the same claim.
The phrase "monism about metaphysical modality" seems odd to me, for it seems to me to be just the thesis that there is no metaphysical modality, that all necessity is de dicto, which is itself a logical consquence of a correct characterisation of the concept of analyticity (in which the semantics relative to which analyticity is judged is required to contain all that is definite about truth conditions).
I note that logical and ontological are not necessarily as distinct as physical and metaphysical.
I am tempted to regard ontology as a part of logic.
It's hard to judge what Carnap would say: if you want to carve the "one language, two logics" examples as two different languages — so as to agree, with Carnap, that each language has its own single logic, then either this is very very different from my pluralism, or you've got to say that the one and the same concrete argument (in which particular utterances form the premises and conclusion, and the logical structure thereof is completely specified) can be correctly analysed in more than one language. Take a concrete example: my argument here:
ReplyDeleteThe dog went left or the dog went right.
The dog didn't go left.
Therefore the dog went right.
For a Carnapian to follow the Beall-Restall pluralist analysis, they need to say that this argument (of the form of disjunctive syllogism) is at the one and the same time in Language 1 (according to which it is valid -- as the logic of language 1 is classical) and Language 2 (according to which it is invalid -- as the logic of language 2 is relevant). So "the dog went left or the dog went right" as I expressed it, is both a sentence of Language 1, and a sentence of Language 2.
I don't have enough sense of what's appropriate to call Carnapian, but if anyone could point at wherever he explicitly addresses how to characterise languages: and whether the one and the same expression could be a part of two different languages (with two different logics governing them), please let me know.
Thanks to Roger Bishop Jones and Greg Restall. I wouldn't know! I guess I could consult, for starters, Jennings, The Genealogy of Disjunction as there may be more to 'or' than meets they 'vel', and quite a gamut of expectation when it comes to different logical systems (what Restall calls 'classical' and relevant') to take into account.
ReplyDeleteBut it's very good to be able to focus, and I thank G. Restall for his apt example: "the dog scenario", in the literature:
I see S. Read has written on the figure of the 'disjunctive syllogism' for Analysis 1981, and that Routley has commented for the J. Canadian Philosophy in an essay bearing 'disjunctive syllogism' in the title too.
Grice at this point would be of little use, since we are considering metalogical complex utterances with embedded operators, turning his his point about any 'implicatum' of 'or' a rather moot one, if well-meant!
.
(p v q, -p) . . p
And I hope Carnap has something to say that relates. I'm sure he will, if pressed. Shouldn't a sentence, to count as being in-one-language come compleat [sic. JLS] with its 'method of verification' or truth-evaluation? (Call me Schlickian before St. Patrick's Day).
My entry point into this (back in 2007) was the question "How is it possible to be more pluralistic than Carnap", which possibly Greg didn't understand (he was giving a talk in the Centre for Mathematical Sciences to a bunch of mathematicians and probably wasn't expecting to be taken up on a completely unrelated philosophical question!).
ReplyDeleteAnyway, my point was, that Carnap's criteria for "acceptance" of a "language framework" (see "Empiricism Semantics and Ontology) are completely pragmatic, and exclude nothing.
To be "more pluralistic" Greg would have to admit some possibility which Carnap excluded, but if Carnap excludes nothing then that is not possible.
Greg did not then, and has not since, offered me any scenario which Carnap would have rejected, though there may be some divergence in the way he and Greg would describe the scenarios.
The headline claim "more pluralistic than Carnap", I think can only properly be made if Greg accepts languages or combinations of languages which Carnap would reject. To establish this claim Greg would have to offer some textual evidence that Carnap would ever reject anything!
I did not then actually read Greg's paper, nor have I now, but I have had a look, and it looks to me as if Greg's argument is to the effect that Carnap would not count the two derivations as being in the same language, and it is because he does that he considers himself to be more pluralistic.
This seems to me incorrect, I would not myself count that as a greater pluralism.
Furthermore, I'm not even convinced that there is a relevant disagreement about the use ot the term "language", for Carnap talks at least in ESO of "language frameworks", which he clearly intends to encompass the deductive system. So wouldn't need to say that the two languages were different, he would just say that the two language frameworks were different, and no disagreement (not even a verbal one) would arise.
Considering the central issue more directly, would
Carnap accept that the same "language" (meaning by that: sentential syntax), might have more than one deductive system?
Certainly he would.
In formal logic such situations are commonplace and uncontroversial e.g. ZF and NF or any number of other variant deductive systems for the language of first order set theory.
Does it make a difference if we are talking about natural languages?
Certainly it does, for Carnap the semantics of natural languages are contingent, and consideration of what are sound derivations in such languages do no belong to philosophy as he conceived it. So I don't think there is any possibility of him proscribing the choices offered in the example.
So, in my opinion there are only two kinds of difference at stake between Greg and Carnap.
The first is a difference in the way languages and their deductive systems are described.
The second is a possible difference in what "pluralism" in this context means.
For Greg it sounds as if one can be more pluralistic just by describing a situation differently, without actually admitting anything which is substantively different.
RBJ
For [Restall] it sounds as if one can be more pluralistic just by describing a situation differently, without actually admitting anything which is substantively different.
ReplyDeleteOr one can well subscribe to The Monist. :).
I will consider any deeper thought on this under a different thread because it defeats me nicely!
I will revise the semantics for 'or' in Grice's System G -- and may come to think of the disjunctive syllogism -- and the dog scenario.
I like the idea of Carnap's phrase, 'language framework' and Jones's qualification about 'sentential' syntax, etc. --.
I would submit that it yields with topic then of 'semantic' consequence (/-), complete with truth-value assignation, and mere syntactic derivation (with Gentzen type intro and elimination rules for 'v').
It does seem like two different languages to me if their syntax and semantics differ for 'or', but then as Jones has elsewhere remarked, the very idea of a 'valid' derivation trades on the 'pragmatic' connection between the syntax and the semantics. Etc. And no, I'm not sure I understand myself, etc. But try!