Interactions

Last week I was charmed by the connectedness of everything; this week I've discovered a new and fascinating application of this principle: the, for lack of a better word (or non-word), interactiness of everything! Seriously though, amid my growing frustration with the refusal of all of our various readings this semester to satiate my desperate need for instant gratification by laying out, in one fell swoop, the causal mechanisms underlying all things, I feel as though I've been offered a glimpse at the gratifying light at the end of this very long scholarly tunnel: as both the Coyne article and the Kendler et al article seem to suggest, whether through explicit intention or just incidentally, a key to identifying the latent causal factor(s) that produce depression (and, presumably, other psychological phenomena too) is not necessarily to look at any one variable in isolation (as the members of each respective "camp" - the "cognitive" camp, the "attachment" camp, etc. - seem prone to doing, with all the ensuing methodological problems of third variables and unseen mediators that Coyne cites compelling examples of them running into), but rather to look for the critical interactions between multiple variables. So, for instance, the role of invalidating or abusive interpersonal relationships in precipitating depressive episodes and triggering relapses is substantial, though it apparently gets ignored wholesale by many of the depression researchers and theorists who are presumably too wedded to their pet hypothesis (be it "learned helplessness" or "ruminative thinking style" or "negative attributions," etc.) to ask what the other people in a depressed person's life might be contributing to the equation; on the other hand, as Coyne himself points out, interpersonal patterns repeat themselves and "play themselves out" in an eminently non-random fashion, and such internal characteristics as neuroticism, implicit expectancies of failure and personal inadequacy in one's relationships, deeply conditioned responses to external distress, and, perhaps most prominently, a history of past depression no doubt contribute to the seemingly inexorable cycle of depression-triggering interpersonal contexts in which a person with depression often repeatedly finds him/herself. If one wants to move toward a better understanding of what causes depression, then, it seems ill-advised to select a single variable a priori and then cling to it, while tacitly dismissing any other possible contributing factors by simply neglecting to measure them (a la the studies Coyne described in which college students completed a self-report measure of attachment style, for example, which classified them according to which of four one-sentence descriptions they chose as most reflective of them - without any data collected on the actual characteristics of their past or current attachment figures or the quality of their relationships).

In light of all this, the approach taken by Kendler et al in their study of the interactions between neuroticism, sex, and stressful life events uniquely appealed to me, and even inspired a renewed hope that maybe I won't have to wait too long for the gratification of my causal-mechanism-lust after all. Why? Because, instead of formulating their hypothesis in terms of a specific variable they were particularly attached to (as it were), they posed a theoretical question that could turn out in at least 2 different ways depending on the nature of the interaction between several different variables previously shown to be implicated in depression. There is something about this approach that makes profound intuitive sense to me as a means of getting at causality; perhaps this is actually utterly irrelevant and just seems associated due to some superficial similarity that will eventually turn out to be moot, but - it seems to me like causality, by its very nature, is interactive. ("No duh," you say. "So what's your point?" But, hear me out here.) One thing causes another thing to behave in a certain way given the specific nature and characteristics of both "parties," not just the one doing the causing. It makes no sense to say, for example, that "fire causes boiling"; if that were the case, then I should be able to apply fire to shirts and shoes and cabbages and lines and tigers and bears (why not) and make them all "boil" alike. But in fact, fire causes water to boil - because of certain properties of water that interact in a specific way with certain properties of fire. Fire causes a very different phenomenon when applied to, say, paper, and a different one still when applied to gasoline, etc.  Similarly, I don't think it makes much sense to say "Stressful life events cause depression," any more than it makes sense to say "Cognitive distortions cause depression" or anything else in that form. This seems analogous to the nonsensical "Fire causes boiling" type of claim. Rather it makes a lot more sense to me that, for instance, "Stressful life events of a certain kind (say, rejection by a loved one) cause individuals with vulnerabilities to certain kinds of stressful life events (say, interpersonal events that reinforce their feelings of inadequacy) to get depressed." This is just a vague and overly wordy template, of course, and I'm not about to claim that it epitomizes the kind of causal explanation that might someday be reached for the characterization of depression; nor can I claim, for that matter, that depression can necessarily be boiled down to any single causal explanation. But by way of contrast, I have a hunch that this type of statement gets closer to approximating the form of an effective causal explanation than any of the "univariate" explanations do.