Intersubiektywne porównania użyteczności


When we think about equality, one thing (not necessarily the only thing) we have in mind is equality of welfare. How should we carry on that sort of thinking? One answer is that we should compare people’s utilities. That answer is attacked for many reasons, but the attack that I consider in this paper is the one that says: interpersonal comparisons of utilities are either 1) impossible, or 2) if possible, so unreliable as to be unusable in actual social situations, or 3) merely expressions of one’s own personal values, so have no standing except with people who share one’s moral outlook.

I begin by asking what are the live issues about interpersonal comparisons. One issue is this: economists have often condemned these comparisons as “meaningless”, but I argue that this is not a serious issue. Another is this: economists have often argued less drastically that interpersonal comparisons either are, or depend upon, value judgements. I consider the versions of this argument advanced by Lionel Robbins and by John Rawls, and I argue that neither succeeds. I conclude that the live issues about interpersonal comparisons are, first, a theoretical issue (what are they like?) and, second, a practical issue (how reliable they are?).

On the theoretical question, I begin by considering an answer proposed by John Harsanyi, Armatya Sen, and Kenneth Arrow, viz. that interpersonal comparisons of utility express the preference of the person making the judgement as between entering another’s state with his probably different tastes, nature, etc. I argue against this proposal. Interpersonal comparisons cannot, as Harsanyi, Sen and Arrow agree, rest upon the judger’s own tastes and values, therefore they cannot, I argue, be expressions of the judger’s preference. So I make a proposal of my own: interpersonal comparisons are factual judgements, based upon judgements about how much different people want, or would want, certain things; they are not expressions of preference, and therefore interpersonal comparisons cannot be reduced to less problematic intrapersonal comparisons.

On the practical question, I admit that my account of the nature of interpersonal comparisons makes their use on the social scale in certain ways especially difficult. But I suggest ways of meeting the difficulty, without resort to value judgements.