Nagel on Stroud on the possibility of metaphysics

Much of modern philosophy has been hypnotized by the idea that causality, necessity, and value are not features of objective reality but merely features the human mind projects onto reality.  David Hume is, of course, the main hypnotist in this connection.  In his important book Engagement and Metaphysical Dissatisfaction, Barry Stroud holds that the arguments for this skeptical attitude do not succeed, and that causality, necessity, and value in fact cannot coherently be eliminated from our conception of the objective world.  At the same time, he does not think that this yields a proof that they really are features of the objective world.  It may be that they are illusions after all, even if we can’t help but believe in them.  Hence, if metaphysics is concerned with establishing one way or another whether such features really are part of objective reality, it is, in Stroud’s view, an impossible project.

Thomas Nagel’s review of Stroud’s book is reprinted in his collection Analytic Philosophy and Human Life.  Nagel raises an important objection to Stroud’s pessimism.  To understand it, we need first to understand Stroud’s rebuttal to skepticism about causality, necessity, and value.

To begin with causality, Hume famously holds that we have no grounds for believing that one event causally generates another, nor even a clear idea of what it would be for it to do so.  We observe, for example, that when one billiard ball contacts another, the latter moves.  Having experienced this on multiple occasions, we come to believe that any event like the first will be followed by an event like the second.  But what we don’t actually observe, says Hume, is any causal power in the first billiard ball by virtue of which it brings about the motion of the second.  It is only the repetition of the sequence – the “constant conjunction” of events of the first type with events of the second type – that we can strictly claim to experience.  This constant conjunction produces in us an expectation that events of the second type will follow events of the first type, and the mind then projects this expectation onto the world, interpreting what is really only a habit of ours as if it were a feature of the billiard balls themselves.  The expectation and projection are merely psychological facts about us and tell us nothing about objective reality.

The trouble with this sort of analysis, argues Stroud, is that it is incoherent.  Hume claims to cast doubt on the reality of causality by making of it a mere psychological projection born of a conditioned expectation on our part.  But our being conditioned to form this expectation, and then going on to project it onto the world, are themselves causal processes.  Hence, Hume has to make crucial use of the notion of causality in the course of trying to cast doubt on the notion of causality. 

Causal necessity is a species of necessity in general, so let’s turn to that.  We take it that whereas it is only contingently true that there is water on the earth, or that beetles exist, or that the Allies won World War II, other things are true of strict necessity.  That is to say, there are things that not only are the case, but could not possibly have failed to be the case.  For example, it is true of necessity that 2 + 2 = 4.  It is true that, if all men are mortal and Socrates is a man, then it follows of necessity that Socrates is mortal.  And so on.

But some have cast doubt on the idea that anything is, as a matter of objective fact, true of necessity, and Hume is once again representative of this skeptical position.  What we take to be necessary truths are, in his view, merely expressions of the “relations of ideas.”  We hold, for example, that all bachelors are unmarried, and indeed this is true by definition.  But that merely reflects how we have, as a matter of contingent fact, decided to use certain words; and this in turn reflects how we have, as a matter of contingent fact, decided to relate certain ideas (the idea of being unmarried and the idea of being a bachelor).  Once again, what appear to be facts about objective reality turn out, on the Humean analysis, to be facts about human psychology.

But here too, Stroud argues, there is incoherence.  In developing arguments like Hume’s, one has, as with any argument, to apply canons of valid logical inference (such as modus ponens, modus tollens, and the like).  Otherwise one’s arguments will be unsound and thus without force.  But such rules of inference reflect necessary connections between premises and conclusion.  For example, modus ponens tells us that if it is true that If p, then q and it is true that p, then it must of necessity also be the case that q is true.  The Humean skeptic therefore has to presuppose certain kinds of necessity in the course of arguing against the claim that there are any genuine necessities.

Now consider value.  The Humean skeptic about value holds that whenever we judge some action to be the best one to take or some belief to be the best one to hold, that is only because of some desire we happen contingently to have.  Hence, if the desire were different, what would count as the best would be different.

Stroud objects that this is simply not how such judgments are actually formed.  If someone believes that p is true and also believes that if p is true, then q is also true, then if he goes on to form the belief that q is true, that has nothing to do with his having a desire of some sort.  He simply notes that the premises are a reason to believe the conclusion.  Similarly, if someone decides to help a friend in distress, that can be simply because he judges such distress to be a good reason to help a friend, rather because some additional factor – his having a desire to help the friend – plays any role in his judgment.

Note also that the skeptic’s position ends up being self-undermining if he takes his analysis of such judgments to cast doubt on their validity.  For that would entail that the skeptic’s own belief that his analysis is the correct one has no more connection to the way things objectively really are than the judgments he criticizes do.

Do arguments like these show that causality, necessity, and value really are features of objective reality?  Again, Stroud resists this conclusion.  The most we can say, he thinks, is that we cannot help but conceive of objective reality as having these features.  But for all that, it may be that these features are nevertheless not really out there in the world.  Stroud draws an analogy with G. E. Moore’s famous paradox.  The statement “I believe it is raining, but it isn’t” is not one that anyone could coherently affirm.  If you sincerely say that you believe it is raining, you cannot consistently go on to deny that it is raining.  Still, it is possible for the statement to be true.  That is to say, it can be true both that you sincerely believe that it is raining, and also that it isn’t in fact raining.  Your belief could be false even if you can’t coherently think that it is.  Similarly, Stroud says, while we cannot help but believe in the reality of causality, necessity, and value, it is still possible that our belief in them is false.

I think Stroud’s position goes wrong at this point, and so does Nagel.  As Nagel points out, the analogy Stroud draws with Moore’s paradox fails in a crucial respect.  Moore’s paradoxical statement has the form “I believe that p, but not-p.”  Where p is “It is raining,” while I cannot coherently believe the statement in question, I can certainly conceive of a scenario where it is true.  That is to say, I can coherently conceive of a situation where I believe that it is raining and yet it isn’t raining.

But suppose instead that p is “There are necessary truths.”  In this case, not only can I not coherently believe the statement, but I also cannot conceive of a scenario where it is true.  That is to say, I cannot conceive of a scenario where I believe that there are necessary truths and yet there are no necessary truths.  For the whole point of Stroud’s critique of Humean skepticism is that I can’t coherently doubt the reality of necessary truths (or of causality, or of value).  “Not-p” is conceivable where p is “It is raining,” but not where p is “There are necessary truths.”  Hence the purported parallel with Moore’s paradox is bogus.

I think this is correct, and that Stroud can and should have drawn from his arguments a more robustly anti-skeptical conclusion than he was willing to draw.  He argues persuasively that we cannot coherently doubt the reality of causality, necessity, and value, but stops short of concluding that this shows that these are features of objective reality.  He should not have stopped short.

Stroud’s rebuttal of Humean skepticism deploys what are sometimes called “retorsion arguments.”  The strategy of such an argument is to defend some claim p by showing that anyone who denies p is led thereby into a performative self-contradiction.  Critics of retorsion arguments sometimes suggest that all they show is that the skeptic cannot coherently reject p, but not that p is actually true.  Stroud seems to be conceding this sort of objection, which is why he draws a more modest conclusion than, in my view, he ought to have.

But he should not have conceded it.  At pp. 80-84 of my book Aristotle’s Revenge, I defend retorsion arguments against this sort of objection.  As I there argue, the right way to understand a retorsion argument is as a kind of reductio ad absurdum argument.  That is to say, it defends p by showing that to deny p leads to a contradiction.  And if a reductio argument succeeds, it doesn’t show merely that we must believe p (where this could be the case even if p were false).  It shows that p is actually true.

(As it happens, at pp. 346-50 of Aristotle’s Revenge I also discuss and defend an argument Stroud develops in his book The Quest for Reality which rebuts skepticism about the objective reality of colors.  Here too, Stroud stops short of concluding that his argument actually establishes the truth of realism, as opposed to merely rebutting arguments against it.  And here too, I argue that Stroud is too modest and that his argument does in fact establish the stronger conclusion.)

Related reading:

Stroud on Hume

Roundup of posts on Nagel’s Mind and Cosmos