ThreeKindsOfNaturalConstraint

If we are concerned with claims that a particular trait of human behaviour is natural it's probably worth trying to distinguish one of three kinds of nature we might be thinking about.

• adapted function

• spandrel

• situational constraints

Adapted function

This is the one most people think of today. If you say behaviour X is human nature, it seems like a claim that humans have evolved to have a tendency towards doing X because it's contributed to the fitness of those with this tendency, and has been actively selected for. The expectation is that, were we to decode the genotype fully we'd discover a gene (or a complex of genes) who's semantics is for "doing" X.

Example : humans probably have an inate tendency to avoid most kinds of pain. Even if we can't find one unified "concept" of pain represented in the genotype, we can probably find the disjunct set of references to various painful experiences which all trigger "withdrawing" / "learn to avoid" reflexes and patterns of behaviour.

Spandrel

Here the claim is different. The claim is that the architecture of the human body / brain does indeed bias human behaviour towards X, but not because behavior X was selected for, or because it contributes to the fitness of it's possessor, but because it is a side-effect of other architectural properties which were selected for. We do not expect to find a gene or gene complex coding "for" it or representing the behavior. It emerges from other behaviors, physical properties.

Example : there may be no genetic predisposition for men to be more violent towards women than vice-versa. Nevertheless males are on average larger than women (something for which there is a genetic disposition) and therefore more likely to win a physical fight. Hence, on average, men's experiments with hitting women are more likely to "work" than women's experiments with hitting men. And so will get more positive re-enforcement.

Situational Constraints

Some constraints are not really represented at all in the architecture of body or brain but are represented in the overall situation that humans find themselves in. Of course, animals without the requisite flexibility might never have found themselves in such situations - the canalization of their evolutionary history prevented them. But humans, as examples of animals with sufficient plasticity are likely to have fallen into these patterns.

My example here is : OneWayLinks. Any network-shaped human behaviour is likely to find it grows faster if it needs one-way as opposed to two-way links. This isn't coded in the genes or even a spandrel. Humans only need to be sufficiently flexible for human behaviours to end-up constained by this patterning.