Friday, June 7, 2013

Goodman's new riddle of induction

To: Chris M.
From: Geoffrey Klempner
Subject: Goodman's new riddle of induction
Date: 12th March 2010 14:00

Dear Christian,

Thank you for your email of 7 March, with your essay for the University of London BA Logic module, in response to the question, 'Is there a satisfactory solution to Goodman's new riddle of induction? If there is, what is it? If there is not, what are the consequences?

This is a very good answer to the question. You have successfully got to grips with the main issues around Goodman's paradox, in particular the idea of 'kinking' (as you describe it,) and the implication that the kinking would have to go 'all the way down' in order for the putative symmetry between normal predicates and grue-like predicates to hold. As you observe, you would need a story about powers and supervenience of phenomenal qualities on underlying structure etc. etc.

The point to make about disjunctive predicates generally, is that you are still relying on a given language in order to define the kinked version. So we have to take a step further and imagine an untranslatable language (Davidson questions this in 'On the Very Idea of a Conceptual Scheme', but all he shows is that nothing could count as 'discovering' the existence of such a language).

Let's keep the discussion on the level of human and alien science. Because, as hypothesised, the kinking goes 'all the way down', there can be no communication between humans and aliens. Our conceptual schemes are too radically different.

However, there is one aspect which one needs to take into consideration, which is how the world 'really is'.

Consider first the original problem of induction. Why should the future resemble the past? Are justified in believing this on inductive grounds? or deductively? or are we just guessing and hoping?

Descartes, in effect, posed this question before Hume with his hypothesis of an evil demon. Leave aside the radical possibility that there is no spatial world at all (just me and the evil demon). The evil demon can make the world any way he likes. If he is good (i.e. godlike) he would make a world in which the responsible use of inductive argument is more likely to result in true beliefs. If he is bad, then the world *looks* like a world in which the responsible use of inductive argument is more likely to result in true beliefs, but in reality is cleverly designed so that we will end up believing all sorts of fantastical things, on 'good' inductive grounds, which aren't in fact the case.

Now, let's look at the same question from the point of view of Quine's 'naturalized epistemology'. There is no a priori proof of a benevolent god, or that the universe is fit for science (as we understand it). All we have is our capacity for empirical research, which, as it happens, thanks to Darwin can provide a very good explanation of how human beings have evolved with dispositions appropriate for noticing relevant properties and projecting them.

However, a priori, you still can't rule out the evil demon or evil scientist hypothesis. Let's say the evil demon has created the human race and also the alien race, and given the aliens the innate ability to project the properties which really correspond to the 'kinked' way the universe has been created, while human beings are blissfully unaware that inductive practices which they find natural and reasonable are leading them further and further away from a true picture of this warped universe, or mini-universe.

As an empirical theory, the 'new' evil demon is a non-starter. There is absolutely no reason to believe it. But that's precisely the point: the game of science is about looking for the best explanation (as we've already discussed). Only philosophers concern themselves with crazy, science fiction hypotheses which have no bearing on the way science is actually practised.

I haven't said anything here with which Goodman would disagree. We can't justify our inductive practice because (short of proof of the non-existence of an evil demon) all we have to go on in determining how the world is, is our inductive practice. We do what we do. The predicates we project are the ones which are projectible. That is all rationality amounts to. There is 'no first philosophy' (Quine), no higher court of appeal.

I'm not sure how to best describe that conclusion in terms of the alternative offered by the exam question of a 'satisfactory solution' or 'no satisfactory solution'. To talk about solution implies a problem, something we have to do something about (as, e.g. quantum mechanics poses a problem for the traditional view of causality). I don't think we have to 'do' anything about it. It's an intriguing puzzle, a 'riddle', as Goodman says, but one which nevertheless shows something profound about the nature of reality, and the limits of human knowledge.

All the best,