Grice and Hume

Roger Jones, for the Grice Club.

I am still trying to come up with a book in which Hume's fork has a central place (as in Hume's "Enquiry"), and I have recently been focussed on that central place.

The fork is, of course, the analytic/synthetic distinction, the underdogma to the defence of which, against Quine's attack, Grice and Strawson rallied (only the words, or this particular use of them, are Kant's).

Naturally, at some point in reconsidering this issue the connection with Grice came to mind.

It is a feature of the Grice/Strawson defence that it pretty much ignores the arguments which Quine deploys against the analytic/synthetic distinction.
It is therefore not at all specific to Quine's attack but is instead a kind of defence which might still be of some interest if we should all one day have forgotten about Quine.

How then does their positive case for the distinction relate to the case which is put by Hume?
Well it seems to me that there are some similarities.

The impression I have of Hume, extracted from his chapter in ECHU "on the different species of philosophy", is that he had acquired a good sense of the relevant distinction (possibly from reading Aristotle, the distinction for example, between necessary and contingent truths, but more probably from diverse sources), and that it had become incorporated into his way of evaluating philosophical arguments.  It is valuable in cutting a swathe through complex abstract argumentation when it is connected with the expectation that arguments a priori can establish only analytic conclusions.  The effect is that any argument, however complex and subtle, which purports to demonstrate a matter of fact without being based on factual premises, i.e. an argument a priori to a synthetic or factual conclusion, must be unsound.  There is in such cases no need to examine the detail of the argument, it need not be taken seriously.
So long as you can tell whether or not a proposition expresses a "matter of fact" you have an intuitive test which will often tell you that you need not take an argument seriously without troubling with its details.

In reading philosophers he observed that they frequently offered arguments a priori for propositions concerning matters of fact, that this simple rule about what one can establish a priori seemed not to be understood.
The essence of Hume's new philosophy consisted in taking this rule seriously.


Hume has two notions of metaphysics at play.
At first he uses the term in what he thinks perhaps a lay sense, to refer to that kind of philosophy which consists in reasoned arguments on matters profound.
In this kind of metaphysics he observes some difficulties which he proposes to remedy in a new more rigorous "metaphysics".
However, he also uses the term more narrowly, using it specifically for the pathological elements which are to be excluded from his new kind of philosophy.  This is the more familiar picture, in which the new kind of philosophy is positivism, formed primarily by the eviction of metaphysics (in the narrow sense).

My concern here is not with that new kind of philosophy, but with the special conception of metaphysics which leads Hume to it, and the intuitive understanding of the phrases "relations between ideas" and "matters of fact" which connect with Hume's conviction that the metaphysics is unsound and unsatisfactory.

It seems to me that some philosophers have, and others do not, a good intuitive grasp of when a proposition says something about the world and when it does not.  Hume had, but others do not.  I'm guessing that Hume may have acquired it from Aristotle, for I think it is there to be had.  I think it probable that it came to me through reading Ayer's "Language Truth and Logic" very early in my philosophical education.  But some seem to me not to grasp it, quite possibly despite reading both these sources and more.  Among today's philosophers, few will have been nurtured in an environment sympathetic to this particular insight (and few will think much of this paragraph).
In my limited reading of Critical Rationalism this is the principal difficulty which I find, that despite its serious desire to be a sceptical philosophy, it is shares the rationalistic defect by failing to perceive the limits of what can be established by deductive arguments.
Of course to carry through this critique I should have to get to specific examples which I won't do here, but I will say that the most easily identifiable reason why (despite all the problems one can find in positivistic philosophers) I prefer to call myself a positivist rather than a critical rationalist, is precisely that at the heart of positivism I see this important distinction, and in critical rationalism it has no such special status.

In the existence and relative solidity (at least for philosophy) of these intuitions lies support for the Grice/Strawson response to Quine.  The status of the analytic/synthetic distinction as a source of philosophical insight does not rest upon any detailed formal characterisation of the distinction.
The sympathy between the two is however limited.
There is no suggestion in Hume that his ability to perceive the defects in metaphysics is something he would expect to see in common sense or to draw from ordinary language.
Hume is talking about profound abstract reasonings, not about ordinary discourse.

And then of course, the purpose to which he initially deploys his insights, the elimination of "metaphysics", is not one with which Oxford or Grice could wholeheartedly support.
Hume can hardly be thought of as a "Viennese redneck" but he is a primary source of the arguments with which they swept aside huge swathes of our intellectual history.

Grice accepted the analytic/synthetic distinction, but preserved a respect for at least some kinds of metaphysics.
Does he speak of that specific notion of metaphysics common to Hume and Carnap (inter alia) which consists in a priori demonstration of synthetic claims?

Roger Jones