One of the things I'm wont to speculate about wildly after imbibing a few too many beers is the possibility that the Stockholm Syndrome is in fact an evolved psychological adaptation. The logic is simple: women, especially young women, seem to be most likely to develop the syndrome and if coming to identify with one's captors is an effective survival strategy and if being kidnapped was a recurrent survival problem in the environment of evolutionary adaptiveness, then it's likely evolution would have favored such a psychological mechanism. The problem, of course, is that data about the Stockholm Syndrome is extremely difficult to come by for the simple reason that doing experiments would be deeply unethical. Indeed, there is some reason to think the syndrome does not exist, that it is an urban myth. Moreover, determining whether Stockholm syndrome is an adaptation is even more difficult: we don't have nearly enough data to settle the issue one way or the other, and I've tried and tried but failed to come up with an ethical way of approaching this topic experimentally. (I considered doing by MA thesis on this but the lack of data and the impossibility of doing experiments dissuaded me. I doubt my university's ethics committee would have looked kindly upon a proposal to kidnap a bunch of doe-eyed first years...). Unless someone comes up with a novel experimental technique, in other words, I doubt we could ever settle this issue.
Nonetheless, evidence is accumulating about one aspect of this problem: whether being kidnapped was a recurrent event in human evolutionary history. We already know women are regularly subject to capture in contemporary non-state societies (Keeley, 1996), but is the same true through history? A paper recently published in the journal Antiquity provides an interesting way to start answering this question: simply determine whether adult female skeletons are represented proportionally in ancient mass graves. (Note: I haven't been able to get access to the full paper, my institution's electronic access to journals is pitiful. I'm relying entirely on the abstract and a single news report). At a Linearbandkeramik site in Talheim, Germany the authors of this paper found that adult women were systematically underrepresented in a mass grave and this, the authors argue, suggests the women were selectively taken. Indeed, a very similar result was obtained at the site of the Crow Creek massacre. If these findings hold up, and if many more similar cases are found, we would have pretty solid (if entirely circumstantial) evidence that being kidnapped was indeed a recurrent problem in human history and that there thus might have been selection pressure for some kind of adaptation.
This kind of data, I must emphasize, is at best suggestive: as I said, I doubt we will ever be able to say for certain whether the Stockholm syndrome is an adaptation or not (or even whether it exists or not). It's still an interesting topic to speculate about over a few beers, however.
(See also: "Sex, Drugs and Cults" by Keith Henson).