Gilbert Harman on logic and reasoning

Ok – beginning an if-possible-even-looser-than-previously practice here of just putting up notes on and associations around things as I read them, I want to post briefly about Gilbert Harman’s article 1984 ‘Logic and reasoning’.  This is an article I read because Brandom frequently cites it.  The point Brandom cites the article for is a good one, which runs something like as follows:

When we think about principles of logic, it is tempting to think of them as rules of belief revision.  But this is a mistake.  In Harman’s words:

Logical principles are not directly rules of belief revision.  They are not particularly about belief at all.

For example:

It is not always true that, if one believes p and believes if p then q, one may infer q.  The proposition q may be absurd or otherwise unacceptable in the light of one’s other beliefs, so that one should give up either one’s belief in p or one’s belief in if p then q rather than believe q.

As Brandom frames this point, all that this inferential space (expressed by logical rules of inference) tells us is what beliefs we are, or are not, jointly entitled to.  The task of then modifying our beliefs in order to render those beliefs consistent is a substantive task that cannot be determined by the logical principles in question.  In Harman’s vocabulary, the practice of reasoning – actually revising beliefs in response to incompatibilities or implications – is separate from the constraints on entitlements imposed by the inferential relations expressed in logic.

Moreover, this argument applies even to the normative injunction to avoid inconsistency – something that is the core task responsibility of the subject as subject – or agent as agent – per the Kantian account of the subject as constituted by the synthetic unity of apperception.  In Harman’s words:

even the rule “Avoid inconsistency!” has exceptions, if it requires one not to believe things one knows to be jointly inconsistent.  On discovering one has inconsistent beliefs, one might not see any easy way to modify one’s beliefs so as to avoid the inconsistency, and one may not have the time or ability to figure out the best response.  In that case, one should (at least sometimes) simply acquiesce in the contradiction while trying to keep it fairly isolated.

I think Harman is here very usefully distinguishing between logical claims about inferential relations, and the practices of reasoning via which one responds to those claims about inferential relations.  One can even potentially say that there are two importantly different senses of ‘rational’ in play here.  (I’m not committing myself to this idea, to be clear – I’m just trying to draw out what I take to be Harman’s point.)  Thus one can (easily) imagine a scenario in which an inferential principle tells us that we are not jointly entitled to hold three of our commitments simultaneously.  It is irrational to hold all three of these commitments together.  How are we to respond to this flaw in our belief system?  Well, it might in fact be that the most rational response to this irrationality is to maintain the irrationality, rather than to over-hastily jettison one of three well-founded commitments.  At the level of reasoning practice, the rational thing to do may be to ignore the ‘first order’ irrationality of the incompatibility of commitments – at least for now, but in principle or conceivably indefinitely.  Thus, for example: It may be rational to be a bit irrational, rather than dogmatically follow an inferential principle in a way that weakens our overall understanding in the service of shallow consistency.

In fact, I think this is a recognisable psychological phenomenon.  One of the reasons I personally found formal logic so alienating when I studied it at university was the feeling that I was being sold false coin.  Logicians were strongly of the view that they were teaching us principles of good reasoning – and yet these logicians did not seem to be notably good reasoners.  If anything, they often seemed narrow, shallow, dogmatic, blinkered, etc. in their cognitive practices.  I think this distinction that Harman highlights in this paper goes some way towards explaining this phenomenon.  This ‘archetypal’ logician is someone who has gone all in on the ‘narrower’ kind of rationality associated with following specific inferential principles to the letter, rather than the ‘broader’ kind of rationality which can respond to inconsistency of commitments by wisely refusing an overhasty jettisoning of commitments, but can rather “live with” the tension of inconsistencies while waiting to see what, if any, is the wisest route to their resolution.

In a post on the blog a couple of years ago, I made this point as follows:

because sustaining incompatible commitments is in fact a core part of a normatively desirable process for rationally addressing incompatibilities within the ‘synthetic unity of apperception’, psychological mechanisms for sustaining inconsistent commitments are a crucial, inescapable part of the machinery of reason. 

I connected this point to the Freudian account of the machinery of the psyche – and argued that the necessity of holding incompatible commitments together as part of rational thought means that:

we can’t a priori say that any of the Freudian categories that capture such inconsistencies of commitment involve ‘irrationality’. 

I stand by this take, and now think that this point can be articulated using Harman’s distinction between the principles of logic and actual practices of belief revision – two different kinds of ‘rationality’.

Ok – but then we need an account of those actual reasoning practices – or, at least, such an account would be nice to have.  (As usual, I’m sure countless people have already done this work, and I simply haven’t read them.)  I don’t think Brandom’s apparatus provides that account (nor is it trying to) except at a very high level of abstraction indeed.

Moreover, this kind of thinking applies to practical, not just theoretical, reasoning.  The practical equivalent or dimension of the synthetic unity of apperception is the obligation for our actions to be consistent – for us to have coherent goals not just within some sub-component of the psyche, but within the psyche as a whole.  On this broadly Kantian account (which Christine Korsgaard seems to have elaborated in considerable depth – but I’ve still not read much Korsgaard), what makes something an agent is a practice of synthesising actions such that those actions form a coherent whole – a self, not just a collection of individual acts.  But the task responsibility to repair the self’s inconsistent actions does not mean that the best way to respond to inconsistency of action (or commitment) is to immediately do whatever one can to eliminate the contradiction.  At a ‘higher’ level of reasoning practice, the ‘best’ practice may (or may not!) be to sit with inconsistency, of word or deed.  This is separate from the questions of a) whether there is such inconsistency, and b) whether one has an obligation to attempt to reduce that inconsistency.  One may indeed have such an obligation, without that obligation overriding others.  Making such calls is part of the work of self-constitution.