And the mechanism is …

Why do people fall in love? This ageless question has seen attempts at resolution by minds of the caliber of Aristotle and Shakespeare, and it is one of the last retreats of people who think that surely science cannot explain everything. And yet a recent book by anthropologist Helen Fisher (Why We Love: The Nature and Chemistry of Romantic Love, H. Holt, 2004) compiles an impressive amount of evidence that science can, indeed, infringe onto this most cherished realm of human romanticism. The book isn't actually particularly well written, and the arguments presented are circumstantial at best (there is also an annoying tendency to quote extensively from poets and writers, perhaps to sweeten the rationalistic blow that is the main message of the tome). Nonetheless, some of the findings are compelling, and provide me with an entry point to discuss what we mean when we say that the cause of X is Y.

But let's proceed with order, discussing some of the findings reported in Fishers book. Monogamy isn't very common in the animal world, but male birds of species that exhibit this behavior, and that stick with the progeny to do their part in parenting, tend to have low levels of the hormone testosterone. This is an interesting correlation that suggests an obvious causal link between hormone levels and behavior. When scientists tested such hypothesis by injecting testosterone in these male birds, their behavior changed dramatically: they abandoned their offspring and went off to court other females. So what, you may say, we (Homo sapiens) are a heck of a lot more complicated, both physiologically and behaviorally, than birds, and one cannot willy-nilly translate such findings to the cultural complexities of twenty-first century humanity. Except that one can: It turns out that among humans, males with high levels of testosterone are less likely to marry, more likely to engage in extramarital affairs, and divorce more often (they are also more violent, being more likely to assault their spouse). Makes you pause, doesn't it?

But the broader philosophical point here is that--even considering the obligatory experimental and statistical caveats--can we say that philandering behavior is (at least in part) caused by high levels of a particular hormone? Well, yes, we can, but we should hasten to say that that is by no means the end of the story. There are additional, and equally interesting, levels at which one can analyze the causes of behaviors (or of anything else for that matter). For example, one might ask: okay, we may have evidence pointing to a causal link between testosterone levels and the propensity for philandering, but what causes such link? That is, how did it come about that testosterone, and not another hormone, plays that role? Or, for that matter, how is it that male birds and humans display such a component within their behavioral repertoire to begin with?

Here the quest for an answer ought to be undertaken in a different fashion, more appropriate to the kind of question we are asking. For instance, one may hypothesize (and this is much more difficult to back up empirically) that male philandering evolved by natural selection, to favor the spread of as many copies of the male's genes as possible by mating (and impregnating) as many females as possible. From this (hypothetical) perspective, then, the testosterone-behavior link is itself caused by the process of natural selection in natural populations, and is one but not the cause of male philandering.

My students and I came up with an analogy that might help drive the main point home. Suppose you are the detective in charge of investigating a homicide. You may ask yourself the generic question: Why did person X die? Immediately you realize that the appropriate answer varies with the level at which you wish the question answered. Do you mean what weapon caused X's death? Then the answer may be a gun. But why did a gun cause the death? Because the bullet shot by the gun penetrated the heart. But why does a bullet that enters a heart cause death? And so on, getting deeper and deeper into the mechanistic causes of X's death. You would also be getting further and further away from your police work, however, because the answers become increasingly irrelevant to your chief objective, which is to apprehend the killer.

You could then move your "why" search some levels up from the murder weapon: Why did somebody want to murder X? Because the latter was carrying some valuables that the killer warned. But why did the killer have to actually commit murder to obtain whatever it is that he wanted? Because we live in a society that puts a premium on the possession of certain objects or wealth and that makes it easy for criminals to obtain possession of guns. And so on, again you could expand your search for causes and make it increasingly broader, though once more you would soon reach the point at which whatever answer you find isn't much relevant to the task at hand, solving the murder case.