Anarchy of Mind

Thanks to reader bhyde's recommendation I've been reading George Ainslie's Breakdown of Will, which is on the whole excellent. I can't quite believe that I never ran across it before, since I did my graduate work on somewhat similar issues. But while I explored all sorts facets of the control of behavior by networks of loosely-coupled goal-directed units, I was innocent of economics or utility theory, so missed out on some things.

The underlying idea of Ainslie's work is that we prefer more immediate rewards to longer term ones, and that the discount curve for such preferences is hyperbolic rather than the more rational exponential curve predicted by standard utility theory. The hyperbolic discount curve leads us to preferences that are not consistent over time, and thus to a process of "inter-temporal bargaining" between different versions (or parts) of ourselves.

The phenomenon of akrasia -- acting against one's own better judgement and values -- has been recognized for millenia, and has its most obvious manifestation in the form of addictions and the failure to overcome them. Ainslie extends his scope way beyond that, however, and attempt to explain everything from scratching intolerable itches to empathy using his framework. I'm not completely convinced, but I probably need to make another pass over the book to fully understand the theory.

Most interesting to me, he posits that the need to manage intertemporal relations between parts of the mind are the root of the self itself. Mental phenomena like will and selves results from attempts to build internal structures that constrain and rationalize divergent preferences. The structure of the person mirrors what we know of the structure of social institutions:

The historic difficulty of specifying what the self consists of doesn't come from its superfluousness, but the fact that it's a set of tacit alliances rather than an organ. The logic of limited war relationships naturally creates a population of cooperating processes, a fringe of outlaw processes, and a means of determining which will be which. And since limited warfare is conducted among individuals as well as within them, we can observe some of its properties in interpersonal examples.

Societies settle disputes within legal systems. Some depend on legislators...who lay down procedural principles. They've been the model for conventional allegorical theories of intrapersonal governance. But the most successful legal system in history, the English common law, has no lawgiver and no written constitution, only a tradition...

Like the common law, this process doesn't require an executive function to steer it. Nevertheless, a person's efficiency at developing personal rules s probably ncreased by executive processes....Thus "ego functions" may be learned on the bass of how they improve intertemporal cooperation.

Ironically, this picture of the person mirrors what our picture of a corporate hierarchy has become...people in corporations don't blindly follow orders, but act only when they're confident of each other's commitment to act. Executives don't function effectively by rationally analyzing facts as by finding facts that make good rallying points.

(p. 98-100)

All this is very reminiscent of Minsky's Society of Mind, which also had a large focus on competition between different parts of an individual's mind. Both theories paint a picture of the self as radically discordant, a network of semi-independent processes whose coordination is not given, but has to be achieved. However, Minsky focused more on cognition and mechanism and less on utility and reward.

Ainslie writes in a highly compressed style that is sometimes difficult to follow. His ideas raise more questions than they answer (for instance, they suggest ways to re-think some standard psychological ideas like the Freudian superego, authoritarian personality characteristics, and "disorders of volition" like depression). And when I read this kind of stuff I always worry a little bit if it is psychologically harmful -- given the effort we put into constructing coherent selves, subverting that effort by pulling the curtains aside and revealing the backstage machinery seems to entail a risk of spoiling the show. But these caveats aside, I found this book extremely worthwhile.

Here's a musical accompaniment: