Saturday, July 14, 2012

Why we still wear pants

Peter Turchin recently published a series of blogs (part 1 and part 2) addressing the question of why westerners wear pants.  It's a story I'd heard before years ago in a comparative Indo-European studies course, but not from an expressly evolutionary perspective and not with Peter's flare.  It's a good story, particularly for introducing cultural evolution, and I don't want to spoil it for those who haven't heard it before.  Nevertheless, not much about this post will make sense if I don't at least mention that wearing pants came about whenever riding horses became an important part of warfare.

As much as I like the story, I made what I thought was a somewhat innocuous comment saying that I would like to see a follow up series of blogs that explored why we still wear pants.  After all, very few people ride horses and, aside from some special exceptions, horse cavalry became obsolete in World War 1.  However, not only has the wearing of pants continued, it has spread, both to societies we've come into contact with and to women within our own society.  This spread has happened without any obvious adaptive advantage.  Moreover, this persistence and spread has occurred despite available alternatives and social movements, such as the androgynous movement of the 1980s, that attempted, and failed, to change these established practices.

Peter's response to this question, while not exactly surprising, was surprisingly dismissive.  His position is that there is no need to explain when cultural traits fail to change.  For him, the lack of change is the null hypothesis.  In his own words:

The null model is dp/dt = 0, where p is the frequency of the trait, and dp/dt it’s (sic) rate of change. If there is change, then dp/dt = c, or dp/dt = cp, depending on the details of how the trait changes. In any case, the greater the rate parameter, c, the faster the trait will change. So if c is high enough you will have yearly changes of fashion, if c is small enough, then it will take decades or even centuries.

The failure of something to change, in Peter's view as I understand it, can simply be dismissed as normal inheritance.  In the absence of variation and/or in the absence of selection pressure to change, we simply go on as we've always done.

I object to this position on a number of fronts.  To begin, it appears to be an assumption, not an empirically established fact.  Moreover, all these equations really establish is that some traits change faster than others, hardly a startling revelation.  What they fail utterly to do is account for why some traits change faster than others.  The salient question, for me, is why dp/dt should ever be zero, or even near zero. 

It is, of course, entirely plausible that Peter is right, that the failure of this particular trait to change might be due to simple inertial processes.  However, this is far from the only possibility.  We need only look to biological systems to see why this is so.  Robustness, the propensity of a system to continue functioning in the face of perturbation, is a general property of complex, evolved systems.  The structure of DNA and the genetic code are robust to mutations.  Proteins similarly tend to perform their functions even with amino acid substitutions.  This robustness is evident all the way up the hierarchy of biological complexity.  Even developmental systems are robust to changes in metabolic pathways. 

I see no logical reason to assume this general property of biological systems, a property that is an adaptation in its own right, should somehow no longer hold true once the level of complexity moves outside the skin of the individual organism.  Moreover, if evolutionary approaches to understanding human culture are to meet their full potential, they must grapple with practical problems in the current world, many of which represent historical patterns from which groups seem almost incapable of escaping.  For instance, Nisbett andCohen have traced modern rates of male violence in the southern United States to Celtic herding cultures, an ancestral pattern that has persisted long after herding faded out as a way of life.  Similarly, Robert Putnam has traced failures of democratic institutions in southern Italy to patterns of social capital building that go back 500 years.

Treating the persistence of these problems as nothing more than null hypotheses drastically reduces the chances of finding real solutions.  Aside from blinding ourselves to the root causes of recalcitrance, robustness paradoxically makes evolvability possible.  A handful of textbook examples notwithstanding, it is rare for a single mutation in biological systems to be adaptive.  Most often, adaptations arise from a series of mutations.  Fragile systems, however, are relatively incapable of building pools of useful variation since any slight change is likely to lead to suboptimal performance.  Robust systems, by contrast, tend to build up pools of variation, some combination of which may prove adaptive when new niches present themselves or existing environmental conditions change.

As for the question that ultimately motivated this post, I have no idea why we are still wearing pants.  If I were to hazard a guess, I would say that once pants became associated with warrior castes they acquired protection by virtue of prestige bias.  This in itself, though, might not have been enough to ensure the practice persisted.  More likely, the wearing of pants started to get tied to concepts in addition to warrior attire.  For instance, pants became associated with markers of gender identity, group heritage, etc.  In other words, it became connected to many other cultural constructs in intricate webs not unlike the way genes often have pleiotropic effects in gene networks.  If so, then wearing pants could become difficult to change even if there were compelling reasons to do so.

Sure, this is all purely conjecture at this point.  However, it hopefully goes some ways toward demonstrating that lack of change is a worthy subject for evolutionary study.