Economics is supposed to be the most advanced of the social sciences. But this is because it is hardly a social science, and those sectors in which it has made the greatest advances are precisely those least concerned with the human. Economics stands at the point of man's interaction with nature, not with other men; its powers of prediction derive from the inanimate side of its subject matter. Increasing returns to scale, diminishing returns when one input is fixed, the inexorability of material balances in an input/output table, the rate of technical progress, the fact of scarcity-in a word, the logic of commodities: no wonder Marx thought of man's struggle with nature as the predictable part of man's history. No wonder either, that he hoped it determined the superstructure. But that is the trouble: it does not determine, it only constrains. And this is what makes a successful political economy so improbable. Moreover, when economics turns to consider man, it takes a most inhuman view of him: he is a Benthamite, a maximizer of a simple and constant psychological function. The problems he faces, too -those of exchange and choice between alternatives are the same every morning, and his unique faculty of memory alters nothing fundamental in his behavior. It only makes him a better adjuster of the same kind. Indeed, the very rate at which he increases his efficiency is a matter of experience, and can be generalized upon (the learning curve). There is, then, no history, but only a chronology of demographic and technological change and capital accumulation. Ideas may develop, but the motivation is always the same. Political power may change hands, but unless it expresses itself in adequately enforced legislation, men's behavior is always everywhere the same. The astounding thing is that this assumption works. Not perfectly, of course. It does not well describe the behavior of most primitive tribes; nor of complicated exchange situations involving trade unions and large capitalist corporations; and least of all governments, which are fundamentally not in this kind of situation at all. But for the rest, including individuals in a communist society, Benthamism is an evidently workable assumption about economic behavior, fruitful of correct predictions. It may be that sociology,