| OCR Text |
Show Rationality and the Structure of the Self, Volume I: The Humean Conception 137 Allais depends on these laws to ground his claim that for a given field of choice defined by an ordinal preference function (3) S = S(A, B, … C), such that A, B, … C are quantifiable commodity bundles, "it is possible to specify a cardinal preference function (4) S = S(A, B, … C) [in which S is the corresponding expected utility function], determined up to a linear transformation, such that for two increments of utility that are deemed equivalent, the variation ∆S will have the same value (43)." But he does not explain how or why two "increments of utility" are to be "deemed equivalent" any more than Fechner did. He does argue as follows: Hardly anybody would fail to reply 'yes' without any hesitation if asked 'would you prefer to inherit $100 million rather than $10,000 more strongly than you would prefer to inherit $10,000 rather than $1,000?' The absence of hesitation demonstrates without any doubt that the notion of equivalent psychological increments indeed corresponds to a psychological reality (46). That is, Allais reasons that because we would unhesitatingly most prefer a larger amount of money that is ten thousand times more than what we would least prefer, to a smaller amount that is only ten times more than what we would least prefer, we must have a concept of equivalent psychological increments. But in order to infer from our ranking of the proffered monetary sums that we have a concept of equivalent psychological increments, this argument must presuppose that monetary value is equivalent to psychological value. Allais does not defend this presupposition. Instead he offers the suggestion, in a footnote, that "[i]t is worth recalling that the problem of determining cardinal utility is identical to the problem of determining the marginal utility of money (n. 17)." He also simply states in passing that to each monetary value there corresponds such a measurable psychological value, or cardinal utility (45-46). But if Allais' experiments are supposed to yield conclusions merely about agent responses to the probabilities of receiving certain sums of money, then he does not need the concept of psychological value; whereas if they are to have implications for a broader and more complex range of cases of the sort described above, then his equation of monetary with psychological values is inadequately defended. On the basis of this equation, Allais then draws a further analogy between the comparison of (3) and (4) and a second one. Given a field of choice among risky outcomes defined by the function © Adrian Piper Research Archive Foundation Berlin |
