Monday, July 18, 2011

Causal Regularity is not Universal but Rare

The picture of our universe as a machine governed by deterministic laws is hard to shake. Physics has profoundly undermined this vision of course, but even before this was known, it is a bit surprising that philosophers and scientists were inspired by everyday macroscopic experience to form such a conception. Was Hume so talented at billiards that he experienced constant conjunction? I can’t come close! Certainly to build and maintain a machine which achieves any determined outcome with regularity takes a lot of human effort.

This last point was made by philosopher John Dupré in a 1995 paper called “A Solution to the Problem of the Freedom of the Will” (hat tip: tweet by Rani Lill Anjum). This paper has a number of thought-provoking insights: contrary to the title, it’s doesn’t put forth a full theory of free will, but argues that the way the world works makes human autonomy unsurprising. (Dupré, a philosopher of biology among other research interests, took part in an interesting debate regarding reduction and emergence on Philosophy TV here). Most assume the world is governed by global microphysical laws, such that autonomy would require an exception to these laws. Dupré argues that we actually have no reason to think the world is governed by such globally applicable laws. Using his terminology, the world is far from causally complete.

Today, our most fundamental physical law, quantum mechanics, is indeterministic -- and I’m continually surprised that well-informed people still assume determinism. But Dupré says it’s not enough to simply argue indeterminism vs. determinism, the argument is targeted against completeness. (He doesn’t get into the details of QM, but I think some of them are supportive of his point: for instance, the fact that quantum probabilities are not objective laws of particle behavior, but rather apply strictly relative to their measurement context.)

Conventional wisdom takes examples of causal regularities, and extrapolates from these to an assumption of causal completeness. Dupré argues that this extrapolation is unwarranted. The evidence, rather, is that causal regularities are the exception rather than the rule in nature, applying only in specialized situations. In fact, “…humans, far from being putative exceptions to an otherwise seamless web of causal connection, are in fact dense concentrations of causal power in a world where this is in short supply.”

The alternative picture, then, is one of objects or systems, at a variety of structural levels, having causal powers. These causal powers do not result in uniform outcomes, because other objects at various levels often interfere with the powers. As discussed in prior posts (e.g. here) on the topic of powers/dispositions, these properties tend toward manifestations, but not of necessity.

Whence the intuition that the world is deterministic, or even subject to universally applicable laws of any kind? Science reveals impressive regularities only in highly controlled settings, and cannot precisely explain complex systems (see Newtonian mechanics and the 3-body problem). Dupré goes on to discuss the rareness of reliable regularities in everyday experience (in contrast to the conclusions of philosophers like Hume and Mill). He then discusses machines, and their inability to maintain reliable regularity without great effort. Machines are not sources of causal autonomy; they are subject to our creation and control. Organisms, on the other hand, do behave autonomously, and are loci of causal order. Of these, humans seem to have the most impressive causal powers for imposing regularities on a world where these are in fact rare.


Allen said...

So how do you think this compares to Meillassoux's views on the necessity of contingency?

Steve Esser said...

I've been thinking about the following: The philosophers advocating this "causal powers" view (see here) advance this idea that they are characterized by having a modal strength in-between pure contingency on the one hand and necessity on the other.

Power properties dispose toward certain outcomes, but the outcomes are not guaranteed.

Meillassoux (and most others) don't have anything like this in mind as an option.

Could the ground of being be something pushing toward a certain kind of reality, rather than being simply hyper-chaos?

Allen said...

Why would the ground of being push towards one kind of reality instead of some other kind?

Why would power properties dispose towards certain outcomes instead of others?

It would seem to me that contingency would still have the final word: ultimately there is no reason, it just is that way.

And since there's no reason that it's that way, there's no reason it why it won't be some other way tomorrow.


Steve Esser said...

If at some point your explanations end with a brute fact or facts about reality (i.e there's no further "reason that it's that way"), I don't see why it follows that this means it could be something different tomorrow.

Allen said...

A brute fact *is* a contingent fact, isn't it? The two terms are synonymous...I would think?

Also, I didn't say that it "could be something different tomorrow". I said that there's no reason why it won't be some other way tomorrow.

A huge difference.

Maybe it will be some other way tomorrow, maybe it won't - but regardless there would be no reason for that difference or sameness.

The way things are tomorrow is (will be?) a brute fact. A contingent fact.

Which was Meillassoux's point with his "principle of unreason" - and thus the fundamental "necessity of contingency".

Steve Esser said...

I take your point. But we continue to experience a (partially) ordered world, filled with sustained patterns of events.

Thoughts said...

There seems to be no discussion of the temporal location of the observer in this discussion. This is crucial in any discussion of free will because if the observer is located in the present instant it can only know it has made a decision after it has made it.

It always amazes me that editors publish papers such as your example because of this obvious flaw. The paper holds that there is a series of events unfolding over time and that the observer can either freely affect or not freely affect the events. If the observer is always after the event then the argument is a foregone conclusion: there is no free will. Indeed, even if constructive indeterminism were possible if "free will" is not a property of the observer then it is not our free will that is being considered.

In the absence of a model for the observer the free will debate is pointless.