Edited and written by David Gordon, senior fellow of the Mises Institute and author of four books and thousands of essays.

The Facts of Economic Life


(Winter 2002)

The Collapse of the Fact/Value Dichotomy and Other Essays by Hilary Putnam

Harvard University Press, 2002 xi + 190 pgs.


Hilary Putnam ranks as one of the outstanding American philosophers of the twentieth century, but never before now has he shown any interest in economics. It transpires that he has been a friend for 50 years of Vivian Walsh, a philosophically-inclined economist; and owing to Walsh's influence, Putnam has decided to bring to bear his formidable philosophical talent on welfare economics. Putnam hardly inclines to classical liberalism. For him, the economist of the age is Amartya Sen—very much a social democrat of the usual boring sort. Nevertheless, Putnam's book has much to offer Austrians.


As Putnam rightly notes, Lionel Robbins, writing in the 1930s, established the conceptual framework of modern welfare economics: "it was during the depths of the Depression that Lionel Robbins, certainly one of the most influential economists in the world, persuaded the entire economics profession that interpersonal comparisons of utility are 'meaningless'" (p. 53, emphasis removed). The old welfare economics of A.C. Pigou was cast aside.


Pigou famously argued that, owing to the law of diminishing marginal utility, transfers of money from the rich to the poor would increase social welfare, other things being equal. Is it not plausible "that the marginal utility of, say, a thousand dollars to someone at the point of going hungry . . . is greater than the marginal utility of a thousand dollars to, say, Bill Gates?" (p. 53). But if Robbins was right, such judgments lack a scientific basis.


What then could stand in their place? Given Robbins's skepticism about ethics, one would expect him to answer, "nothing." Rational argument could not settle questions of ethics. Putnam, here following Walsh, maintains that the logical positivists crucially influenced Robbins. Judgments of ethics reflect no more than subjective preferences. As Robbins phrased the matter with characteristic flair, "If we disagree about ends, it is a case of thy blood or mine—or live and let live, according to the importance of the difference, or the relative strength of our opponents" (p. 54, quoting Robbins).

Given this view of ethics, must not welfare economics be totally cast out? Since value judgments are of their essence subjective, any discipline purporting to determine objectively how to advance social welfare seems doomed.


Robbins and his colleagues found an ingenious escape: if valid, it would secure the scientific status of welfare economics and, not coincidentally, their own continued employment as expositors of the new wisdom. Suppose that a course of action increases someone's welfare without hurting anyone else's. Can we not say, then, that the action is objectively desirable? We have not claimed to measure interpersonal utility; and in what way does the appeal to the principle suggested, the Pareto criterion, involve us in any subjective value judgments?


In just this way, Putnam responds. An increase in anyone's utility that harms no one else, it is contended, is desirable; but is not this very claim a judgment of value? If so, on the subjectivist reading of ethics that underlies Robbins's analysis, the principle itself is groundless. If, in response, one detaches the principle from ethical subjectivism, declaring it objectively true, an obvious inquiry rises up to threaten the sufficiency of the new welfare economics. If the Pareto criterion is more than an arbitrary preference, then at least one objective ethical judgment is possible. But then what rules out appeal to other principles alleged to be objectively true?


Putnam explains the vital point at issue: "if the reason for favoring Pareto optimality as a criterion is that one approves the underlying value judgment that every agent's right to maximize his or her utility is as important as every other's, then it would seem that Pareto optimality isn't a value neutral criterion of 'optimality' at all. How could there be a value neutral criterion of optimality, anyway?" (p. 56, emphasis removed).


The Pareto criterion et hoc genus omne do not then rescue us from ethical subjectivism. But our author does not write to berate Robbins for insufficient faith in the logical positivist creed. Rather, he contends, contrary to the logical positivists, that ethical judgments are not mere subjective preferences. Like ordinary factual judgments, they are objectively true or false. In pressing his case, Putnam's skill as an outstanding philosopher emerges in full force.


To refute the positivist argument is no easy task. Is it not obvious that factual judgments, e.g., "this table is brown," differ entirely from such value judgments as "I like professional wrestling"? Further, who can deny that factual judgments are objective in a sense that value judgments are not? We can look at a table and see what color it is: we cannot test whether wrestling "really" is good. Logical positivism may be outdated, but is not the fact-value gap as wide as it ever has been?


Putnam deftly turns these points aside. There are indeed statements that consist purely of factual terms; others are entirely valuational. But many statements fall into neither class. Consider the judgment: "A free market is the only workable social system." (This example is mine: I doubt that Putnam himself would accept it.) Its truth can be established by rational argument, as readers of Mises and Rothbard will know. Yet surely "workable" is a value term. In sum, values are not always a matter of choice. Values are inextricably bound up with our ordinary factual concepts.


As Putnam knows full well, defenders of the gap will not at once retire from the field of battle. True enough, defenders of the gap like R.M. Hare will concede, some terms have both factual and valuational aspects, but must philosophical analysis stop here? Rather, can one not always separate a "mixed" judgment into descriptive and valuational parts? If so, the fact-value gap returns.


Putnam finds this response unconvincing. "The attempt of noncognitivists to split thick ethical concepts into a 'descriptive meaning component' and a 'prescriptive meaning component' founders on the impossibility of saying what the 'descriptive meaning' of, say, 'cruel' is without using the word 'cruel' or a synonym" (p. 38).


If Putnam is right, economists cannot dismiss ethics as subjective. What, then, is the next step for economists who accept Putnam's arguments? How exactly can an objectively true welfare economics be achieved? I cannot think that Putnam's suggestions are of much help.


He maintains that, contrary to Robbins, interpersonal utility comparisons can validly be made. How can Putnam say this? He himself recognizes that "the idea that the amount of satisfaction different people get from various goods and services (and from such intangibles as opportunities) can be linearly ordered . . . seems absurd" (p. 55). Is this not the essence of the case against interpersonal comparisons?


To Putnam, this consideration does not suffice to dispatch all such comparisons. We cannot place in a single order all rankings of goods, but can we not arrive at some valid rankings? To return to an example previously mentioned, is it not obvious that a starving man values one thousand dollars more than Bill Gates does?  


Putnam is entirely right that people in ordinary life frequently make such comparisons. It hardly follows from this, though, that in doing so people judge that the starving man obtains more fixed units of utility from the money than does Gates. Quite the contrary, such comparative judgments are rough and ready: how then can they form part of a scientific welfare economics?


To his credit, Putnam does not attempt to elevate his interpersonal comparisons into full-fledged scientific assessments. Rather, he contends that such commonsense judgments are all we need. He cites with approval Amartya Sen: "In arriving at an 'agreed' range for social evaluation . . . there has to be some kind of a reasoned 'consensus' on weights, or at least on a range of weights. This is a 'social choice' exercise, and it requires public discussion and a democratic understanding and acceptance" (p. 55, emphasis removed, quoting Sen).


Here Putnam falls into just the arbitrary subjectivism against which he has so effectively argued. Let us grant him the "commonsense" comparisons of utility he wants. Putnam has done nothing to show that these comparisons have any relevance for welfare economics. If a starving man values one thousand dollars "more" than Bill Gates does, how is it supposed to follow that money should be taken from Gates? Surely some set of ethical principles needs to be offered to support redistribution, but Putnam fails to provide any.


Our author would, I suspect, reply in two ways. First, have I not begged the question against him? His contention is that issues of social evaluation need to be settled through democratic discussion, leading to consensus. In asking for principles to justify despoiling Gates, do I not implicitly assume that democracy is not enough?


By no means. If there are objectively true principles that mandate democratic consensus, I have done nothing to exclude them without a hearing. Rather, I simply want to have a look at them. It surely does not suffice for Putnam to postulate democratic consensus without argument, as he appears to do.


Putnam might reply in a different way. Perhaps Sen's "equal capabilities" view is supposed to provide just the grounding in principles for consensus that I have demanded. But neither Putnam nor Sen shows that the equal capabilities approach leads to decision by consensus.


Putnam's analysis of welfare economics thus begins well but ends on an unpromising note. Putnam makes excellent points against the ethical subjectivism that underlies much of standard welfare economics. But he provides us with no adequate replacement.1



1 I confine to a note this more technical point. Putnam, following Walsh and Sen, claims that the standard assumption of transitivity of preferences is controversial (pp.163ff., n.3). Misesian praxeology does not assume transitivity (Human Action, p.103), and thus seems to have an advantage over standard neoclassical models.



Close Window