Grice as a Philosopher of Rationality

-- by JLS
---- for the GC

---- IN A RECENT POST TO HIST-ANALYTIC listserv R. B. Jones was wondering about Grice's project, as to whether, say, his appeal to psychological attitude reports involved a desire to elucidate the notion of rationality. Seeing that Jones is having a broader project of his own involving such a notion, I will say: yes, that _was_ Grice's project.

GRICE DID NOT PROBABLY SEE himself as a philosopher of rationality back _then_ -- when he was talking of rational constraints for, say, dialogue, back in 1952, when Strawson was already reporting on his views in the memorable footnote of Strawson's vademecum in logic. But I would say it developed with time, Grice's notion of rationality.

AS A PHILOSPHER MYSELF I too have come to think -- especially on weekdays -- that rationality IS the philosophica notion par excellence. And I especially have come to value Grice in his trying to get to the core of this notion. This core, as he would have it in Aspects of Reason involves:

ψA(p)
ψA(p ->q)
---------
∴ ψA(q)

This above is just _one_ format of one specific step in reasoning -- Modus Ponendo Ponens, as it happens. But it already shows some of the main features of the notion -- of "reasoning" qua verb:

i. It involves a transitional, of the psychological kind. Where "A" is the reasoner, the scheme shows that the premises (A believes that p, and A believes that if p, q) yield, CAUSALLY, and not just transitionally, and this much via the link with desire and intention, the conclusion (that A believes that q).

ii. Desire or intention. Grice was still working on these notions by the time he replied to the 'introduction' to his festschrift in 1986 ("Reply to Richards" proper). He refers back in 1986 (written circa 1984) to his 1977 Kant Lectures (at Stanford) where he had elaborated a first shot at the sufficient and necessary clauses involving "reasoning" -- as he had done vis a vis "meaning". It was obvious to him that a 'desire' -- or an 'intention', even -- that there IS a logical supplementation on there being a 'link' which would be more than 'arbitrary' for the conclusion being yielded by the premises.

iii. The openness of the psychological attitude involved. Grice is onto what he calls the aequi-vocality of 'reason'. There is not one theoretical reason -- the realm of physics, say -- and a 'practical' reason attached to it -- Kant's realm of ends --. Both spring from neutral considerations.

iv. "Reasoning" is founded on psychological attitudes as such. An animal may be held to hold such attitudes -- and thus, his account requires that we explain attitude report in terms OTHER than a connection with a 'sentence' -- since, after all, monkeys cannot ceteris paeribus talk (WoW:iii). Even in homo sapiens, there's such thing as irrational action, self-deception, and 'motivated' akrasia, to use Pears's phrase --. Grice did not see this as detrimental to the notion of reasoning, but quite the contrary, as a safe foundation for it in the realm of general ethology. We need an agent to _behave_ first, in order to have him behaving _rationally_ at a second stage.

v. Non-monotonic. While Grice would distinguish between, roughly, 'logical' and 'pragmatic' inferences -- the latter should also be counted as some sort of manifestation of the faculty of 'reason' or 'rationality'. This would entail allowing for defeasibility -- caeteris paribus character -- and a broad range of variability -- 'probability', 'desirability' -- attached to the idea of 'satisfactoriness' that lies behind the philosophical recontruction of reason.

vi. Etc.