I'm not sure that most people on this list would be interested in taking a
discursive or textual route in rational choice (and I'm not sure I'm crazy about
it either), but some social theorists have made efforts to examine rational
choice as being a subset of hermeneutics, or the study of social meaning. In
short, as I understand the argument (it is developed most fully by Charles
Taylor in one of the essays in "Philosophy and the Human Sciences," Cambridge
University Press, sometime in the 1980s), only by undertaking an examination of
the meaning that social action has for participants (through analysis of
symbols, texts, etc.) can one come to a clear understanding of the person's
interests. In doing so, one can sidestep the problem of tautology by building
interest-based explanations not on prior revelation of interests but rather on a
critical reading of how that person is likely to interpret her social
environment and to ascribe meaning and consequence to her own actions. I think
this is similar to what Geertz has in mind in studying "Thick Description,"
which is probably supposed to replace the "Thick Imputation" that we would
likely find in rational choice analyses based on revealed preferences.

There are a number of particularly troublesome elements about this line of
theory too. First, no one has yet created a really clear way of determining what
it is that we are supposed to describe thickly. Second, it's not always
intuitively obvious how hermeneutics can be done rigorously (I actually think it
can be done quasi-scientifically, but I'm kind of in the minority on this one,
and tend to take a lot of flak from both sides in this debate), since it's not
possible to get into anyone else's head, and frequently hermeneutics reduces to
a sort of armchair phenomenology, where we sit around thinking, "Now what would
I do if I were in that guy's shoes?"

In any case, much as one might fiddle with and fine-tune one's assumptions about
preferences, based on behavior, one can "test" interpretations of meaning by
whether or not a given set of interpreted, rather than revealed, preferences
provides a good explanation for action. This doesn't guarantee anything about
getting good explanations; it simply provides another way of getting at
preferences that allows one to pay attention to symbols or discourse, phenomena
that have no independent existence outside of people's attribution of meaning to
them. But it allows for the possibility that one can create good
preference-based explanations that are not based on tautology.

Jurgen Habermas goes a bit further and claims that the only category of social
action that can explained without an analysis of social meaning is "strategic
action," which I read to mean "economic action," since that's about all he
discusses in this category, showing his Marxist orientation. I suspect -- and I
have to admit to some problems in trying to understand Habermas -- that this is
so because the market is a structure that determines action in a way that does
not rely much on symbols or social meaning; it *does* have an independent
existence. My guess is that he would claim that political action is not based on
anything like a structure, but rather on human institutions and artifice, making
symbols, meaning, and especially communication central to any action. Hence, the
need for attention to these issues in trying to determine preferences.

In any case, I bring this up not because I think it's going to help anyone get
to the bottom of the bureaucracy vs. pols debate. And I agree with Professor
McKean (and others), that rational choice has a lot of heuristic value;
irrationality isn't a particularly useful analytical category. But for those
people who are interested in issues areas where matters like identity and
discourse are important (say, gender politics, immigration policy, or, in
America, prayer in the schools), I think it's important to bear in mind that one
need not eschew intentional action simply because one wants to pay attention to
symbols and texts.

David Leheny
Department of Government
Cornell University