Limits of Models

Pin It Now!
Modern economics, with its heavy emphasis on mathematical modeling and quantitative methodology, exhibits a particular affinity with a philosophical tradition called logical positivism.  Crudely speaking, logical positivists posited that there were two kinds of statements: the verifiable and the nonsensical.  As an example, "this pen is red" is a verifiable sentence (though it does presuppose a common acceptance of the meaning of "red" and "pen").  However, a sentence such as "lying is immoral" is not verifiable.  In short, moral judgments are not given credence because they are mere subjective preferences not amenable to empirical verification.
This focus on empirical verification reveals the deep affinities modern science has in relation to logical positivists.  Following the methods underlying observable science, modern economics eschews the notion of "other minds" because we cannot penetrate those minds.  What is must be observable, reality is not interpreted or mediated, but is seen.  By observing human actions we know what they value, and so forth: each person acts in ways that reflect his or her utility function, so to speak.  It is clear that this focus shares a closeness with the micro-foundation view in neoclassical economics that is based on a notion of human beings as rational actors bent on maximizing his or her utility.  However, this view fails to explain problems such as the collective action or prisoner dilemma problem.  Indeed, recent psychological experiments suggest that humans possess some notion of "fairness" and "reciprocity" which, if violated, will lead them to choose an action that is not predicted by the utility maximizing theory alone.
The limits (and also the appeal) of this approach to economics is that it does not address the "other minds" problem.  How can we know the contents of another's mind?  How can we compare satisfactions and utilities?  By focusing solely on observable behavior, and positing that such behavior is emblematic of another's utility function, we have done away with one problem but raised another: why do people act the way they do?  And if we do not know why people act the way they do, are we sure that we are not only positing correlations that stop at the boundaries of the events which we seek to explain?  Of course, to subscribe wholly to the view that reality is interpreted and we constantly mediate our relationship to the world via values and norms invites the criticism of being unverifiable, and hence, only an assertion.
I do not set out to resolve these differences.  Indeed, this is a schism that has existed in social inquiry and policy making.  Recent advancements in science (e.g., fMRI imaging) show correlations between regions in the brain and certain thought processes:  they reveal which questions appeal more to our emotions and which appeal more to our rational consciousness.  There is reason to be skeptical of these studies, however, because of the low signal-to-noise ratio.  Indeed, the measurement of the relaxation rate of the hydrogen caused by the coupling of oxygen to iron is indirect, and in fact, because of a feedback mechanism, more oxygen actually flows to the parts of the brain that supposedly "light up."  However, I would also urge one to consider the premises on which scientific inquiry is based and consider whether such premises are equally applicable to social inquiry.  Indeed, by rejecting that we fill in the contents of a narrative vision of ourselves as we negotiate our place in the world, and insisting that the world is composed of rational beings bent on maximizing his or her own utility function, rather limits the explanatory power of modern economics or a positivist based version of social inquiry.  In fact, many laws in the sciences are not Aristotelian in the sense that they do not start from the behavior of a single atom (or if you want to be really precise, boson) and derive their way to a description of large-scale phenomena.  General relativity works quite well even though it shares no particular affinity with quantum mechanics.  A similar problem appears in language: do we, or can we, ever know anything beyond language?