« The joys of unfettered research | Main | Reality as just another kind of media »

May 14, 2005

The utility of irrationality

I have long been a proponent of rationality, yet this important mode of thinking is not the most natural for the human brain. It is our irrationality that distinguishes us from purely computational beings. Were we perfectly rational thinkers, there would be no impulse buys, no procrastination, no pleasant diversions and no megalomaniacal dictators. Indeed, being perfectly rational is so far from a good approximation of how humans think, it's laughable that economists ever considered it a reasonable model for human economic behavior (neoclassical microeconomics assumed this, although lately ideas are becoming more reasonable).

Perfect rationality, or the assumption that someone will always follow the most rational choice given the available information, is at least part of what makes it inherently difficult for computers to solve certain kinds of tasks in the complex world we inhabit (e.g., driving cars). That is, in order to make an immediate decision, when you have wholly insufficient knowledge about past, present and future, you need something else to drive you toward a particular solution. For humans, these driving forces are emotions, bodily needs and a fundamental failure to be completely rational, and they almost always tip the balance of indecision toward some action. Yet, irrationality serves a greater purpose than simply helping us to quickly make up our minds. It is also what gives us the visceral pleasures of art, music and relaxing afternoons in the park. The particularly pathological ways in which we are irrational are what makes us humans, rather than something else. Perhaps, if we ever encounter an extraterrestrial culture or learn to communicate with dolphins, we will, as a species, come to appreciate the origins of our uniqueness by comparing our irrationalities with theirs.

Being irrational seems to be deeply rooted in the way we operate in the real world. I recall a particularly interesting case study from my freshman psychology course at Swarthmore College: a successful financial investor had a brain lesion on the structure of the brain that is associated with emotion. The removal of this structure resulted in a perfectly normal man who happened to also be horrible at investing. Why? Apparently, because the brain normally stores a great deal of information about past decisions in the form of emotional associations, previous bad investments recalled a subconscious negative emotional response when jogged by similar characteristics of a present situation (and vice versa). Emotion, then, is a fundamental tool for representing the past, i.e., it is the basis of memory, and, as such, is both irrational and mutable. In fact, I could spend the rest of the entry musing on the utility of irrationality and its functional role in the brain (e.g., creativity in young songbirds). However, what is more interesting to me at this moment is the observation that we are first and foremost irrational beings, and only secondarily rational ones. Indeed, being rational is so difficult that it requires a particularly painful kind of conditioning in order to draw it out of the mental darkness that normally obscures it. That is, it requires education that emphasizes the principles of rational inquiry, skepticism and empirical validation. Sadly, I find none of these to be taught with much reliability in undergraduate Computer Science education (a topic about which I will likely blog in the future).

This month's Scientific American "Skeptic" column treats just this topic: the difficulty of being rational. In his usual concise yet edifying style, Shermer describes the tendency of humans to look for patterns in the tidal waves of information constantly washing over us, and that although it is completely natural for the human brain, evolved for this very purpose, to discover correlations in that information, it takes mental rigor to distinguish the true correlations from the false:

We evolved as a social primate species whose language ability facilitated the exchange of such association anecdotes. The problem is that although true pattern recognition helps us survive, false pattern recognition does not necessarily get us killed, and so the overall phenomenon has endured the winnowing process of natural selection. [...] Anecdotal thinking comes naturally; science requires training.

Thinking rationally requires practice, disciplined caution and a willingness to admit to being wrong. These are the things that do not come naturally to us. The human brain is so powerfully engineered to discover correlations and believe in their truth, that, for instance, even after years of rigorous training in physics, undergraduates routinely still believe their eyes and youthful assumptions about the conservation of momentum, over what their expensive college education has taught them (one of my fellow Haverford graduates Howard Glasser '00 studied exactly this topic for his honors thesis). That is, we are more likely to trust our intuition than to trust textbooks. The dangers of this kind of behavior taken to the extreme are, unfortunately, well documented.

Yet, despite this hard line against irrationality and our predilection toward finding false correlations in the world, this behavior has a utilitarian purpose beyond those described above; one that is completely determined by the particular characteristics and idiosyncrasies of being human and has implications for the creative process that is Science. For instance, tarot cards, astrology and other metaphysical phenomenon (about which I've blogged before) may completely fail the test of scientific validation for predicting the future, yet they serve the utilitarian purpose of stimulating our minds to be introspective. These devices are designed to engage the brain's pattern recognition centers, encouraging you to think about the prediction's meaning in your life rather than thinking about it objectively. Indeed, this is their only value: with so much information, both about self and others, to consider at each moment and in each decision that must be made, the utility of any such device is in focusing on interesting aspects which have meaning to the considerer.

Naturally, one might use this argument to justify a wide variety of completely irrational behavior, and indeed, anything that stimulates the observer in ways that go beyond their normal modes of thinking has some utility. However, the danger in this line of argument lies in confusing the tool with the mechanism; tools are merely descriptive, while mechanisms have explanatory power. This is the fundamental difference, for instance, between physics and statistics. The former is the pursuit of natural mechanisms that explain the existence of the real structure and regularity observed by the latter; both are essential elements of scientific inquiry. As such, Irrationality, the jester which produces an incessant and uncountable number of interesting correlations, provides the material through which, wielding the scepter of empirical validation according to the writ of scientific inquiry, Rationality sorts in an effort to find Truth. Without the one, the other is unfocused and mired in detail, while without the other, the one is frivolous and false.

posted May 14, 2005 06:07 AM in Thinking Aloud | permalink