Teleology in Big Systems

In two earlier posts, I summarized John Hawthorne and Daniel Nolan’s analysis of teleological causation (here), and examined the question of what kind of evidence would persuade us that single particles were teleologically directed (here), as, in Aristotle’s system, where heavy particles are teleologically caused to travel to and rest at the centre of the Earth. My conclusion was that unless the dynamics of such a particle’s movement was different from that predicted by contemporary mechanics, there would be no reason to adopt teleology.

What about systems and wholes? Under what circumstances should we say that the Universe or Earth’s ecosystem is teleologically directed, or that an organic system is?

We should note first of all that this question is (as H&N fully realize) completely distinct from that of analysing the meaning of the words ‘function’ and ‘goal’, as scientists use these terms today.

Biologists, for example, talk about homeostatic systems being goal-directed and argue about the function of T-cells in the immune system. They assume that if this kind of talk would be wrongheaded if it assumed any form of causation other than that which is sanctioned by contemporary mechanics. Accordingly, philosophers have spent a lot of effort reducing the scientific use of these terms to physical processes, or at least to terms that don’t posit types of processes additional to those that modern physics admits. And, as I mentioned earlier, they come up with analyses like the following:

A system S is oriented to goal G, if there is a “servomechanism” that (a) detects any deviation of S from the most direct path to G, and (b) exerts a force that returns S to a direct path to G when a deviation is detected.

As should be obvious, this analysis doesn’t necessarily violate contemporary mechanics. Organic systems such as the body’s thermoregulation system have servomechanisms of the stated type: sensors that detect departures from a “goal” temperature and forces such as sweating that come into play when departures are detected.

 In my post regarding particles, I asked when it would be right to posit that a particle was teleologically goal-directed. I replied that this would be justified only if we observed movements that contradicted mechanics—for example, if a particle were regularly to swerve to avoid obstacles that stand between it and its goal. Similarly, one may ask: when is it right to posit goal-direction for systems. And the answer, once again, is: when we observe changes that cannot be explained by mechanics.

Suppose we were to find an organism that maintained its body temperature without a servomechanism of any kind. For example, it maintains its body temperature in higher ambient temperatures, but without radiative loss of heat, or sweating, or anything else. Similarly, in lower ambient temperatures, but without metabolic generation of heat, or seeking out a heat source, or anything else. Such a creature would not just merit the biologist’s use of the word ‘goal-directed,’: it would be goal-directed in the fully teleological sense.

Now, this whole issue arose for me in connection with Tom Nagel’s recent book on Darwinism. Nagel’s claim is that historically there was insufficient genetic variation to generate human mentality, hence human mentality must have been teleologically generated. Let’s, for the sake of argument, grant the claim about insufficient genetic variation. The question is: how did mentality come about? In light of the above discussion, we discern two distinct possibilities that Nagel might have in mind.

First, he might be saying that though it is physically possible (by a fluke series of mutations, for example) for mentality to have come about, it would be better explained by teleology. (Let’s call this the “intelligibility” argument.)

Second, he might be saying that it is physically impossible for mentality to have come about, and that teleology is needed to generate it.  (Let’s call this the “argument from causality.”)

I am inclined to think that Nagel confuses these arguments. Here’s why. Suppose first that there was very little genetic variation five billion years ago, and hence no reason to believe then that consciousness would emerge naturally. Should we conclude now that consciousness emerged by a fluke or by teleology? Nagel might be suggesting that teleology offers us a better explanation, and should therefore be accepted. But is this right? If the emergence of consciousness is physically possible, then why should we prefer an explanation that violates physics? Wouldn’t such an explanation be a bad one? Bad, for the very reason that it posits an extra-physical causes and thus throws the hitherto accepted completeness of physics into doubt without adequate reason? This is why the intelligibility argument fails.

Of course, if it were physically impossible for mentality to have come about from insufficient variation, then Nagel would be justified. The case would then be like that of the strange thermoregulator above. But obviously this is not so. The human genotype exists. It is finitely different from genotypes that existed five billion years ago. Finite differences can be bridged in an instant, given a sufficiently bizarre and unlikely physical setup. Imagine a machine that takes E. coli DNA and generates human DNA. Physically possible? Why not? So why is it physically impossible that the machine’s internal workings should have been replicated by chance in the Earth’s ecosphere at an instant, or over 5 billion years? This is why the argument from causality fails.

I am inclined to think that Nagel doesn’t understand the nature of statistical mechanics. Suppose that a large insulated volume of air spontaneously became thermally unstable.  We could posit that Maxwell’s demon was at work. Or we could say that this is a fluke occurrence that is, after all, consistent with statistical mechanics. If Maxwell’s demon was at work, the earlier probability of instability occurring was high. If not, it was low. Does this make Maxwell’s demon a “better explanation?”