Tuesday 21 September 2010

What I don't believe

Had another chat the other day with BS regarding the perennial topic of free will. BS thought that he'd worked out another way to tackle the problem and to get around my obstinate neo-dualism. Possibly since alcohol was involved at the same time, I couldn't get my head around how his argument moved things forward. Actually, the lack of traction possibly had more to do with simple intransigence on my part.

I was hoping to re-articulate his viewpoint here to clear things up for myself, but I’m drawing a blank. Instead, I’ll re-articulate my own interpretation of the conventional viewpoint, and see if that can intrigue him enough to repeat (and rephrase?) his point. One point of confusion seemed to be what we meant by deterministic, chaotic and stochastic processes, and what each meant for considerations of free will. My argument runs as follows …

If we assume that there are no stochastic processes (i.e. that quantum mechanical randomness, is really a manifestation the operation of chaotic hidden variables), then all that’s left are particles and fields that interact in completely deterministic ways. And if we know what the complete set of rules (= physical laws) are, then the resulting system is completely predictable.

Now, by “completely predictable” I mean that, were we to know the state of a system perfectly (which is obviously not possible for any reasonably interesting system), we’d be able to predict its time-evolution completely. And when I say “perfectly” here, I’m obviously assuming absolute perfection in measurement, such that chaotic effects can be ignored (and, as already noted, a system in which quantum systems aren’t what they appear to be).

Now, even if we can’t achieve perfection in measurement, I’d argue that such a system, while perhaps not being predictable, is still determined for any future point by its current state. Further, I’d argue that this is true even if the system includes machines such as ourselves which have a vast number of interacting internal elements that can convolute near-identical inputs into myriad diverse outputs.

If we replace this pure deterministic system with stochastic (say, quantum) elements, the result is something that is unpredictable both in principle and in practice. It might still be largely predictable if the stochastic elements are swamped by deterministic ones. The world in which we live is widely believed to be in this mould.

So what has any of this to do with free will? The point the other day seemed to be about whether the deterministic system described above engages in “decisions”. My argument is that it does not, since its future behaviour, however complex, is completely determined in advance by its current state. No matter how tortuous the path by which, say, a stimulus is connected to the behaviour of a human-shaped robot, it is preordained – as well as completely predictable if one is in the god-like position to know everything there is to know about the world.

While the situation is qualitatively different in the case of a system contains stochastic elements, there still aren’t any decisions being made. A particular outcome might only have a particular likelihood associated with it, but that’s not because of anything decision-flavoured.

I think BS’s point may have been that a human-shaped robot has neural pathways that process stimuli through a complex net of memories, instincts and heuristic algorithms into particular, decision-like outcomes. Which is, of course, true. But I’d suggest that said neural pathways got to where they are by earlier boring old deterministic (or deterministic + stochastic) processes.

In this view, “decisions” are simply the result of a complex chain of predictable deterministic algorithms that themselves are the result of an earlier-occurring and slightly simpler chain of predictable deterministic algorithms. And so on backwards till we get to the base neural pathways built by the decidedly non-decision-like interactions between genes and their serf-like protein products.

Anyway, all of the above serves just to paint “decisions” in the world of mainstream physics as something of far less interest than that which we normally attribute to them. Basically to make my faux neo-dualism more appealing!

1 comment:

BS said...

I will reply in the fullness of time...