In his entry on "Analysis" at Stanford, Fregean scholar M. Beaney tells the story with some continuity. His section 7 is Carnap. His section 8 is Grice. Some commentary. From the site at:
http://plato.stanford.edu/entries/analysis/s6.html#7
Section 7 is thus entitled, "Carnap".
Some fragments with commentary:
"influenced not only by Frege, Russell and Wittgenstein but also by neo-Kantianism (see Friedman 2000, Richardson 1998)."
I was surprised to read in the "Finding Aid to Grice" that he has joint work with this M. Friedman on universalia!
---
"to be ‘constructed’ by quasi-analysis, a method that mimics analysis in yielding ‘quasi-constituents’, but which proceeds ‘synthetically’ rather than ‘analytically’ (1928, §§ 69, 74)."
This seems to trade on "analysis" as a mathematical practice, and Grice would have favoured that. Perhaps Aristotle is given too much credit with the invention of "analysis" -- his "Analytica priora and posteriora", but perhaps, analysis, qua analysis, is after all, a mathematical conception (Descartes).
Beaney quotes from Carnap 1936:143:
“The logical analysis of a particular expression
consists in the setting-up of a linguistic system
and the placing of that expression in this system.”
He also quotes from the later, "Meaning and Necessity" (1947):
"The task of making more exact a vague
or not quite exact concept used in everyday life
or in an earlier stage of scientific or logical
development, or rather of replacing it by a newly
constructed, more exact concept, belongs among the most
important tasks of logical analysis and logical construction.
We call this the task of explicating, or of giving an
explication for, the earlier concept."
(1947, 7-8)
Beaney concludes his discussion of Carnap with quotations from the still later, "Logical Foundations of Probability" and the idea that temperature is an elucidation of warmth. He notes the longitudinal unity of this conception of analysis which Beaney traces to the Greeks and Descartes on geometry.
The next section 8, is "Oxford Linguistic Philosophy"
Alas there is no explicit mention of Grice, but if Beaney is reading all the profusive quotes by Frege that Horn is bringing to the forum in his discussion of F-implicature, he SHOULD start quoting Grice more often!
Beaney dedicates a whole passage to Ryle, which Grice saw as too senior for consideration. Ryle had been born in 1900, and the post-war philosophers of Austin´s group -- to which Grice belonged -- found his figure too "fatherly". THEY were supposed to be doing philosophy, not aplying it.
Beaney discusses Austin, whom Jones-Speranza regard as ´prototype´ for the type of analysis from which Grice departs and refines.
Beaney notes:
"J. L. Austin ... emphasized the need to pay careful attention to our ordinary use of language, although he has been criticized for valuing subtle linguistic distinctions for their own sake."
Whose sake?
"He was influential in the creation of speech-act theory, with such distinctions as that between locutionary, illocutionary and perlocutionary acts (Austin 1962a)."
Grice does use the idea of a "central speech act" (sic, in WoW:vi) but without the dogmas. He was of course more onto the assertion-implication distintion, which he was attracted to, as Austin was with his own theory, for methodological rather than substantive reasons.
Beaney jumps from Austin to Strawson -- who was Grice´s student.
"P.F. Strawson, whose critique of Russell’s theory of descriptions in his own seminal paper of 1950, ‘On Referring’, and his Introduction to Logical Theory of 1952 had also helped establish ordinary language philosophy as a counterweight to the tradition of ... Carnap."
--- with caveats. Grice gets TWO mentions in that book. In the preface, and in a pretty famous (among Griceians) footnote, where Strawson discusses the "pragmatic rules" of ´implicature´. In fact in his earlier "On referring", Strawson HAD used "imply" to mark this important non-logical relation, which he later re-baptised "presupposing". The important difference is of course metaphysical: truth-value gaps are a no-no for Grice.
Beaney notes:
"The appearance of Individuals in 1959 and The Bounds of Sense in 1966 signalled a return to metaphysics, but it was a metaphysics that Strawson called ‘descriptive’ (as opposed to ‘revisionary’) metaphysics, aimed at clarifying our fundamental conceptual frameworks."
Grice suggests that he regrets Strawson never cared to mention him in THAT book. Grice always kept the notes on "Categories" -- some archival material cited by Chapman in her biography of Grice --. So, this is indeed a Griceian-Strawsonian joint thing. Grice became less and less descriptive as time went by an self-confessed revisionary.
Beaney:
"It is here that we can see how ‘connective’ analysis has replaced ‘reductive’ analysis; and this shift was explicitly discussed in the work Strawson published shortly after he retired, Analysis and Metaphysics (1992)."
Strawson perhaps published one book too many!
----
Note that he also reprinted his contribution to the Grice festschrift, "If and -->" in his own "Identity and Entity". What´s the good of a festschrift if you are going to publish the "contribution" elsewhere? At least on principle he never let himself published his joint "Defense of a dogma" with Grice in his own publications!
Beaney:
"Strawson notes that analysis has often been thought of as “a kind of breaking down or decomposing of something” (1992, 2), but points out that it also has a more comprehensive sense (1992, 19), which he draws on in offering a ‘connective model’ of analysis to contrast with the ‘reductive or atomistic model’ (1992, 21). Our most basic concepts, on this view, are ‘irreducible’, but not ‘simple’."
--- This is so complex, and one wonders. Strawson is influential for BEANEY, because Beaney is a Brit and Strawson is a Brit. Grice had become an American by then, but give me Grice anyday!
Beaney continues:
"A concept may be complex, in the sense that its philosophical elucidation requires the establishing of its connections with other concepts, and yet at the same time irreducible, in the sense that it cannot be defined away, without circularity, in terms of those other concepts to which it is necessarily related. (1992, 22-3)
Such a view is not new."
I rather prefer Grice´s clearer reflections on reductive versus reductionist analysis in his WoW: Retrospective Epilogue.
----
Plus, Grice, unlike Strawson, had a sense of humour!
----
Beaney refers, "for further discussion". to Baldwin 2001; Beaney 2007b; Hacker 1996, ch. 6; Lyons 1980; Passmore 1966, ch. 18; Rorty 1967; Stroll 2000, ch. 6; Warnock 1989."
--- of which Warnock, the best. Grice´s English teacher biographer, Siobhan Chapman, keeps quoting Warnock as G. C., which confuses me! He was Sir Geoffrey James, and Vice-Chancelor of the only university that should have one: Oxford.
Hacker succeeded Gordon Baker (festschriftist for Hart and Grice) as tutorial fellow at St. John´s, so the least thing he can do is quote from Grice. But Hacker, like, alas, D. F. Pears, has turned into a Wittersian, leaving all traces of Griceanism implicaturish wit behind! (He has Witt, rather!).
---
Passmore is a good one. A total Colonial. I loved that footnote in "Hundred Years": "There´s this ingenious fellow, Grice, who has written so little it scares". Or words.
Subscribe to:
Post Comments (Atom)
Good stuff, JL.
ReplyDeleteOn the matter of analysis, my own usage generally does not sustain it as opposing but as essentially involving synthesis or construction. It is only in the analytic/synthetic usage that there is a contrast.
They are used in ancient Greece to refer to different kinds of proof (which we would now call, at least in computing, forward and backward proof corresponding to the greek synthetic and analytic proof), but this does not correspond to any distinction in the thing proven, not at least the modern distinction, though possibly a closer look at the Greek usage would be necessary to be sure.
Anyway, I like these days the term "nomologico-deductive analysis" and in this term the "nomologico" (rule based) bit is best done by the construction or synthesis of some kind of model, for preference a logical or mathematical model so that you have something nicely enough defined to provide a good subject for deductive reasoning, which is where the more obviously analytic part begins.
Analysis and Synthesis are then like two sides of the same coin, unity in diversity, different perspectives on or aspects of one method. And I also admit that not all analysis is nomologico-deductive, however broadly one construes it, but that even the less formal kinds of analysis (with which one will begin on tough philosophical problems, and perhaps never progress beyond) still blend analysis and synthesis, but do so in an insufficiently definite way to support rigorous deduction.
RBJ
Good. Indeed, perhaps Stanford should commision someone to write an entry for 'synthesis'.
ReplyDeleteBut I do like the opposition of analysis and synthesis that you propose. Indeed, first comes synthesis (we join things together -- syn, with, thesis, to put). Only THEN can we analyse, i.e. ana-luein, se-parate.
One cannot separate the elementary. Not just in chemstry, but in, say, philosophy of perception. These are the 'givens' -- the colour 'yellow', say, or this 'this'.
-----
I like your 'nomologico-deductive'. To be contrasted perhaps with 'hypothetico-deductive' with which I am more familiar with, and which I associate with Popper. That sort of abduction is, when liberally interpreted, present in something that Grice calls nonlogical, though. So one has to be careful.
E.g.
"He said p."
"He wouldn't be saying p unless he believed q"
"Therefore he believes q"
Etc.
"Nomo-" indeed, from 'nomos', law. A regularity. The idea, when used in hypothetico-deductive, is that we do need something like a 'major' premise. But as you use it, is indeed more like the idea of having a "system" which allows us to 'deduce'. There is the pattern of 'satisfactoriness'. The system should provide the canons for a passage from one statement in the system to another.
The nomo- then refers to the items (premise, conclusion) as belonging to the system. And the deductive then must work in some way or other. For Grice, in "Vacuous Names" (his System Q, renamed system G, by Myro and myself) it is via 'inference rules' alla Gentzen. Without 'rules of inference', no sense of talking 'deduction' ('natural deduction'). These, while not really just syntactic (although I do hold them as syntactic, cfr. Belnap, Tonk, Plonk, and Plink) are not semantic and that's what matters. But they do seem to 'rescue' or 'save the phenomena' -- the intuitive inferences we intuitive grasp as 'valid' --.
There are questions, specific, as to how a general characterisation of nomonlogico-deductive yields PHILOSOPHICAL analysis, etc. -- a minor one. Also cfr. Beaney's cursory remarks on Carnap on temperature. On the face of it, I doubt if 'warmth' IS a philosophical notion, and I would think that 'cold' is elucidaded by 'temperature' as much as 'warm' is!
----- Grice's idea of analysis varied with time, too, and the focus on English was paramount, and his following his intuitions, etc. So one wonders why Beaney is not quoting him again and again!
On the ordering, this is a chicken and egg thing maybe.
ReplyDeleteWhat I described was an ordering on the formal modelling, but before you come to that, you need an informal analysis (at least in your head).
You take the subject matter to pieces first, then reconstruct formally, then use deduction to reason about reality as captured in the formal model.
On my preference for nomo- rather than hypo- first note that I associate the hypo- thing with Popper and with Poppers falsificationism, and I don't subscribe either to that or to the verification principle.
I think in terms of models, preferably formal, and these are not the same as hypotheses, there is no need for them to be "true" a model can be more or less like the reality, and so they aren't the kind of thing which one hypothesises. They are the kind of thing which you compare with reality, and assess fidelity and utility in a more fine-grained way than looking for a true/false verdict. This is a kind of epistemic retreat.
A second very important element of my conception of nomologico-deductive method is that the construction of a model (when done properly!) provides a coherent context in which deductive reasoning can take place.
You have something definite about which you are reasoning, and cannot therefore be lead into contradiction (the model can be defined by "conservative extension" in an accepted logical context, which is a fancy way of saying by definitions rather than by postulates).
This plugs perhaps the most important hole in the use of reason in analytic philosophy, which is a complete lack of any method for ensuring that the context in which deduction is conducted is coherent.
There is a separation between the logical system and the model. The system will be something independent of the particulat application, perhaps set theory, which provides a rich enough logical context for the model to be constructed by definition, thus ensuring that the definition is at least coherent.
This is like Frege's logicist conception of mathematics:
maths = logic + definitions
which we accomodate to those who think narrowly of logic by the Quinean reformulation:
maths = set theory + definitions
which really was the dominant foundational conception of mathematics in the 20th century (at least for mathematicians if not for anti-foundationalist philosophers).
So really the "nomo" isn't strictly appropriate for what I have in mind, because the definition of a model does not involve introducing any new rules.
But I like it anyway, because I think of this as just a better way of formulating "laws of nature", you retreat (epistemically) to the construction and evaluation of models, and there aren't any better terms for what I have in mind.
In doing logical analysis, it is desirable to have a good logical system in which to conduct the analysis which is good enough for a wide range of applications.
In science that includes not only the logic, but also the mathematics, for in most scientific applications you are talking not just about logical models but about mathematical models, and neither the scientist nor the philosopher want's to have to re-invent the logic or the mathematics for every new application.
If what you are modelling is actually a logical system, it may still be pragmatically best to do that in the context of some richer logical system (which might then feel like a metalanguage).
This is exemplified by the work I did on Aristotle, which is all done in higher order set theory, and conducts the analysis by constructing various semantic models of Aristotle's logic and category theory.
As to how this might work out with system Q/G I can't really comment because I don't have enough detail.
RBJ
Thanks, and for your elaborations and notes on your choice of 'nomo-', etc. I will consider.
ReplyDeleteI like it! In a way it reminded me of Nancy Cartwright. In her contribution to the Grice festschrift she wrote, "In next chapter all this will become clear". But the next chapter is the essay by J. Baker, so she couldn't mean that. It was a typo indicating that the whole thing was to be part of Cartwright's book which became, "How the laws of physics lie", i.e. do not say the truth. This had to do with Cartwright's attending seminars on 'as if', with Grice ("Hands across the Bay" -- Cartwright at Stanford, Grice at Berkeley).
----- I like models a lot. Max Black was a good one at that -- and indeed models are like THEORIES. For Grice, sometimes, a theoretical approach (which he thinks is NEVER intuitive) is different from an 'analytical' approach, which in his case, always runs via intuitions.
---- Yes, System G is something to consider. And I follow you in that one interpretation of the System is just THAT, and that various models can be provided for the same statements of the system.
Thanks for your insight into the larger issues, as to the place of reason in analytic methods and the need for a system with coherence and decidability.