Thursday, March 13, 2014

Justification of induction and Leibniz on free will

To: Alistair L.
From: Geoffrey Klempner
Subject: Justification of induction and Leibniz on free will
Date: 29th November 2011 12:38

Dear Alistair,

Thank you for your email of 19 October with your essay for the University of London BA Methodology module, in response to the question, 'What, if anything, can explain the rationality of reasoning according to inductive principles?', and your email of 25 November, with your essay for the Spinoza, Leibniz, Kant module, in response to the question, 'Can Leibniz sustain his claim that Caesar freely crossed the Rubicon?'

Once again, apologies for overlooking your Methodology essay. I have a system, which usually works very well, of printing off the email and also making an icon on my desktop with the student's name and date, at same the time when I email my acknowledgement.

I think what must have happened is that I overlooked the fact that I had two quite separate actions to perform -- record your payment and also queue your essay -- at the time when I sent my acknowledgement. Must do better next time.

Do please send a reminder if I take longer than two weeks to respond. My deadline is 10 days, and I usually succeed in replying within that time.

Rationality of inductive reasoning

This is a very thorough essay which most of the angles in the debate over the rationality of inductive reasoning. You wisely avoid the question of Popper's claim that the method of science is not inductivist but rather conjecture and refutation. Even if that were accepted, it remains true that in a myriad ways, we rely on inductive inferences all the time. Every purposive human action takes place against the background of an assumed predictability in the way the world (and our own body) works.

You also briefly mention Goodman's 'new riddle', although I found the idea that this involves the claim that 'induction cannot possibly be reliable' difficult to follow. If anything, Goodman's own 'solution' to the grue paradox -- the theory of 'entrenched' predicates -- looks like a route to the solution of the traditional problem of induction. The 'rationality' of certain basic practices consists wholly in the fact that 'this is what we do'. This is the bedrock, or the background against which significant departures from rationality are measured.

What is it to 'rely' on something, or some process? I rely on the fact that when I hit a key on my computer keyboard (the heavy classic Mac beige keyboard which I have fixed up to work with my G5) a certain pressure, and no more, is required to produce a character. My previous keyboard, after years of sterling service, developed a problem with some of the keys. Not knowing how hard you have to hit the keys means that you end up hitting them extra-hard each time. You could have fun running a Goodman-style argument for computer keys (which need to be hit with a force of 100 psi after time t) and imagining creatures who (by accident or design?) have evolved to anticipate this very event.

That we rely on a certain thing or process isn't a 'reason', in the sense of something you could use to persuade a doubter to put their reliance on that thing or process, if for any 'reason' they weren't prepared to. 'Come in the water's lovely.' 'It may be for you!'

But let's try to imagine someone who really would count as a 'practical sceptic' (not just the philosophical, armchair style) about induction. Say, someone who believes in 'Sod's Law'. According to Sod's Law, things we rely on are reliable, except at the precise time when failure has the worst possible consequences. Your hard drive will have a catastrophic breakdown just as you are in the process of transferring your precious work to a new computer. Your brakes will fail just as you are driving round the bend in the mountain pass.

A pragmatic argument could be constructed in favour of belief in Sod's Law, not exactly along the lines that -- as you show -- Reichenbach argues pragmatically in favour of induction, but still in the same ball park. It is pragmatically useful to rely on induction. But agents who believe in Sod's Law fare better in the long run because they take extra precautions. Admittedly, this isn't to disbelieve in induction, and therefore not an argument against the claim that if any method works, induction will work. But at least it gives the semblance of a genuine choice: should I be a simple inductivist, or would it be better to embrace inductivism modified by Sod's Law?

We are skirting on the edge of a grey area of human psychology, where issues such as self-fulfilling prophecy arise. Expecting things to go wrong can often be a cause of their going wrong. Not everyone is the same. We don't all put the same faith in induction. Why should there be an 'ideal' position to take? If not, what does that show about the 'rationality' of induction?

Hume, it should be noted, was perfectly satisfied with his 'human nature' account of inductive and causal reasoning. It isn't a paradox for him that there is no such thing as a 'rational justification', whereas the continued and distinct existence of objects ('On Scepticism With Regard to the Senses') is seen as paradoxical, requiring an extra 'fix' (the theory of fictions).

Leibniz on free will

You offer a good account of Leibniz's defence of free will, in terms of his theory of monads. I agree with you that, on the face of it, the defence doesn't look as if it would convince a sceptic. But what exactly is Leibniz setting out to do?

As you yourself admit, there are two things in Leibniz's theory -- the foreknowledge of an omniscient Deity and the principle of causality (for Leibniz, the Principle of Sufficient Reason as applied to the world of phenomena) -- which pose problems quite apart from any problems entailed specifically by the monad theory.

There is a problem raised for free will in the thought that at every moment God knows what I am going to do, and also a problem raised for free will in the thought that every action that I do is the causal result of the total state of the universe, including the state of my own body etc.

Given either of these things, there is no alternative possibility, e.g., to Caesar's crossing the Rubicon. God knew what Caesar was going to do, and it was impossible for Caesar not to cross the Rubicon given the state of the universe at the time when he pondered his decision.

The theory of monads is a metaphysical theory, an attempt to describe reality at the ultimate level. It's justification is, as you indicate, that alternative descriptions (such as the one offered by Descartes in his simple dualism of immaterial souls and physical stuff) are logically flawed. As in Berkeley's immaterialism, the phenomena are saved. Billiard balls continue to collide in predictable ways, the Earth continues in its orbit around the Sun. Wouldn't it be rather surprising if the thing or capacity which we rely on in our phenomenal existence -- our ability to reason from a practical perspective, make choices and act on them -- was undermined by the purely metaphysical description?

Like other philosophers (including Spinoza) Leibniz quickly dismisses the notion of 'freedom of indifference'. Leibniz's positive account of freedom, as you admit, in practice sounds very much like the standard analysis offered by compatibilism. My 'free' actions are those which issue from my own deliberative process, rather than things I do under external compulsion. All the distinctions we standardly make, the reasons for or against holding an agent 'responsible', can still be made with the theory of monads as the ultimate metaphysical background.

However, it could be argued that that point is at least not blindingly obvious. At first glance, the picture of you and I as 'windowless monads', pre-programmed to make all the decisions, undertake all the actions we will do in our entire lives looks the very picture of unfreedom. It's the 'freedom' of a clock to show 27 minutes past 12 one minute after showing 26 minutes past twelve.

Leibniz's task, as I would interpret it, is not to 'solve' the problem of causality or God's foreknowledge but simply to dispel this superficial reading of his own theory. As a bonus, he can argue that alternative metaphysical views make it more, rather than less difficult to draw the distinctions necessary to allow that, e.g., the things that God knows that I am going to do are my actions nonetheless.

All the best,