citizens are left without incentive to stay informed regarding their political choices ( Downs 1957). Such legal and philosophical concepts that link what we can hold people responsible for with their psychological state at the time of choice (e.g., mens rea) will be affected by showing that what we can reasonably expect from rational individuals is ignorance and lack of disposition to be involved. Hence, this generates arguments for a lessened individual ethical responsibility with regard to political matters ( Oppenheimer 1985). If one can only marginally affect the outcome, then (given the modern philosophical adage "no ought without a can") the contours of the modern political philosophy scene must flow around the findings of the rational choice theories just as they flow around the findings of other modern sciences.
Our extended (though admittedly partial) tour of the findings of formal theory in political science puts the reader in a position to assess the current state of the field and some of the controversies within, and surrounding, the field. We temporize a bit about this task because controversies do get resolved and, in a developing field, the current state is passed before one can comment on it. We also offer some prognosis for the future of a science of politics and in what way formal theory conforms to the scientific standards we alluded to above.
If we conceptualize the field of formal theory as the combination of deductive methods and the basic conjectures of rationality, we might wonder how far such a thin reed can be pushed. Work continues at the foundational level and in expanding the range of applications. One of the healthy signs that the field will continue to grow and develop is the displayed willingness of its practitioners to explore the possibility of reformulating the foundational assumptions. These reformulation efforts have so far been pushed in two areas: choice under risk and self-interest.
Over the last fifteen years or so, how people behave in the face of risk has become a topic of intense experimental research. Major, replicable discrepancies to the ( von Neumann-Morgenstern) expected utility assumptions have shown up in the laboratory. These negative findings have implications, not only for the status of the expected utility theory as a valid descriptive theory of choice, but also for the behavioral consequences of many formulations of rationality. Although a number of serious theoretical reformulations have been proposed, they have proved to be neither a reasonable normative, nor a reliable descriptive, alternative to the standard expected utility theory. 42
Like the rationality conjecture, narrow self- interest certainly is part of the foundational conjectures of economics, and has stood the test of time well in market contexts. On the other hand, the display of pure self- interest seems to be context dependent. In non-market contexts, significant non-self-interested behavior such as altruism has often been observed. 43 Such displays, often thought to have some political role, call for a theoretical explanation.
The theory of rationality does not restrict the values which an individual may have and which generate his or her preferences. But many models of rational behavior do stipulate narrowly self-interested preferences, assuming that preferences of an individual are independent of the preferences of others. And to consider dropping or changing that assumption leaves us with a problem. If preferences are not independent, then what is the specific form of their interdependence? Beyond self-interest lies an infinity of alternative conjectures. Behaviorally, although at times significant, this "deviant," non-self-interested behavior has not been found to be predominant (see the experiments reported in Frohlich et al. 1984). Simple "other-regarding" alternatives to theories of self-interest, such as a specific form of altruism, have theoretical problems themselves (see Margolis 1982) even though they do cover a considerable amount of the observations of non-self- interested behavior. Further, there is some evidence that individuals who hold these other-regarding values are quite consistently acting in a rational way ( Frohlich and Oppenheimer 1984; Goetze and Galderisi 1986). Trying to understand the role of altruism has, however, led to the development of both non-self-interested models and tests of such arguments.
Perhaps most central to the evaluation of the field should be its record in generating a trail of claims to knowledge -- claims that are made so as to be correctable. Not only is it clear that a large number of topics central to the concerns of the discipline are now theoretically "understood," but the knowledge claimed about these topics is increasing. The field of formal