Coming Soon:

Now Available: Volumes I, II, III, and IV of the Collected Published and Unpublished Papers.



Total Pageviews

Friday, August 26, 2016


In my Autobiography and elsewhere I have written a good deal about my personal relationship with Jack Rawls.  As some of you may recall, he and I were colleagues for a year [1959-60, I believe] when I was an Instructor in the Harvard Philosophy Department and he was a Visiting Professor [he returned in 1961-62, as I recall, as a regular member of that department where he then taught until his retirement.]  Jack published his hauptwerk, A Theory of Justice, in 1971 and six years later I published the first book-length critique, Reading Rawls.  Since Rawls is widely viewed as one of the most important 20th century philosophers to write in English, and is perhaps world-wide the most important political philosopher of the past 150 years, there is some value in adding to the public record any information about his views of the philosophical response to his work.

This morning I was cleaning up my office and throwing out various things that have accumulated when I came across a letter Jack wrote to me in 1977.  I am going to reproduce it here verbatim, for such interest as it may hold to students of his work.  A few words of explanation are called for.

Stephen Strasnick was a Harvard doctoral student who wrote a dissertation on Rawls’ theories under the directorship of a committee consisting of Rawls himself, his philosophy colleague Robert Nozick, and the great economist Kenneth Arrow.  In his dissertation, Strasnick undertook to produce a formal proof of Rawls’ so-called Difference Principle.  He published the proof as an article in the Journal of Philosophy in 1976.  I read the article while I was sitting in an airplane, returning from giving a talk somewhere in Ohio.  Something seemed wrong to me about the proof, and when I got home I took a close look at it.  In December of that year, I published a refutation.  Strasnick’s error was actually rather interesting [at least if you like that sort of thing!]  His idea was to adapt the logical framework of Arrow’s General Possibility Theorem and use it to try to prove the Difference Principle, but he failed to note that Arrow assumes ordinal preference whereas Rawls implicitly assumes cardinal utility functions.  The result was that Strasnick’s premises, when correctly interpreted, reduced to tautologies entailing nothing significant and certainly not the Difference Principle.  I sent a copy of my refutation to Jack when it appeared, and what follows is his handwritten response, dated Mar(ch) 27 (1977).  By the way, Jack’s reference to Bob is to his son, who majored at UMass in an undergraduate interdisciplinary program called Social Thought and Political Economy which I created and was then running.

                                                                                                            Emerson Hall
                                                                                            Cambridge, Massachusetts 02138
                                                                                                                    (617) 495

Dear Bob

            Many thanks for your piece on Strasnick, just received.  I find myself at the moment in a tangle trying to collect myself for a brief trip abroad, and don’t know when I shall be able to look at it.  But hope to, once I get myself in one piece when I get back.  I was away last year and didn’t see S’s final thesis draft until after the oral exam (Arrow & Nozick were the committee), though I knew of the theorems.  I think they are correct, but heavens knows, they are not well presented.  Your essay interests me;   because I am puzzled by these formal proofs; and other proofs that have been formal [or possibly, “that have been found”]  There are 4 or 5 proofs of the DP floating around now – all more or less the same, I think.  Strasnick’s were as early as any.  It’s hard to know what their real significance is.  Anyway

            Thanks for your paper -- & keep flourishing.


PS  Bob is enjoying UMass, for which I’m grateful, and to you for a big part of that.

That was the last I heard from Jack about Strasnick and my article.


wallyverr said...

This post clearly fit into the category of "too good not to google", and in particular I wondered who the other authors of the “4 or 5 proofs of the DP floating around” by 1977 might be.

Amartya Sen’s collection of essays Choice Welfare and Measurement cites the Strasnick paper a number of times, usually in conjunction with journal articles by other scholars. Some names I recognised, with papers published by 1977, include:
Peter Hammond
Eric Maskin
Partha Dasgupta
Kevin Roberts
and Sen himself.

All of these are very highly-regarded mathematical economists (Sen of course is much more as well). So even though Strasnick’s own paper mixed up ordinal and cardinal, I’d be surprised if all the others did as well.

When Rawls wrote you that letter, I was either taking, or had taken the previous semester, Rawls’ own lecture class on Theory of Justice. I think it was the first time he had taught the class after the book’s publication. He had us read Kant’s Groundwork and some Sidgwick, but most of the term was spent on The Book, if not as close a reading as you plan to give to the Critique of Pure Reason. Brian Barry’s book-length critique was on the reading list. I’m not sure if your own book on Rawls had quite been published by then.

J. W. F. said...

I think I understand why many believe that the difference principle implicitly relies on cardinal utility, but I would emphasize that Rawls is pretty explicit about his ordinalism:

"The difference principle meets some of the difficulties in making interpersonal comparisons. This it does in two ways. First of all, as long as we can identify the least advantaged representative man, any ordinal judgments of well-being are required from then on. We know from what position the social system is to be judged. It does not matter how much worse off this representative individual is than the others. if positions can be ranked as better or worse, the lowest can be found. The further difficulties of cardinal measurement do not arise since no other interpersonal comparisons are necessary. And, of course, in maximizing with respect to the last favored representative man, we need not go beyond ordinal judgments. If we can decide whether a change in the basic structure makes him better or worse off, we can determine his best situation. We do not have to know how much he prefers one situation to another. The difference principle, then, asks less of our judgments of welfare. We never have to calculate a sum of advantages involving a cardinal measure. While qualitative interpersonal comparisons are made in finding the bottom position, for the rest the ordinal judgments of one representative man suffice." Rawls, A Theory of Justice, 91-92

In fact, as I understand Rawls' argument, cardinality is only plausible so long as one accepts hedonism. This is the important, and insufficiently discussed critique that Rawls makes (or so I claim in my dissertation anyway) of what he calls "dominant end" conceptions of the good and their failure to account for the "unity of the self." This comes in sections 83-85, ostensibly directed at Loyola and Sidgwick.

Robert Paul Wolff said...

Rawls may say it, but that doesn't make it so. For example, the concept of an index of primary goods, which is crucial to such argument as Rawls has, cannot be made sense of without cardinal utility. I honestly do not think Rawls understood the technical side of these things very well. If anyone is genuinely interested, I will write about this at length. Otherwise, just look at my book.

Of course, if you do not take Rawls seriously when he says he is giving an argument, and instead simply construe that big book as a fruit salad of tasty assertions vouchsafed by Rawls' intuitions, then anything goes. But I chose, all those years ago, actually to take him at his word when he spoke of sketching a theorem, i.e., a deduction of conclusions from premises by rigorous argument. I am not sure many readers do that.

howie b said...

Aren't some arguments so pretty and nice that mere proof is a sheer afterthought?
It sounds like Rawls just liked his ideas, and felt them practical and of some use.
This is a slightly different topic, but Godel was a Platonist, his proof notwithstanding.
I mean did Marcuse offer logical proofs, held to the same exacting standards you hold Rawls?

Robert Paul Wolff said...

Marcuse did not claim to be proving a theorem. Rawls' justification of the Difference Principle is either that it is a theorem in bargaining theory or else that it is the articulation of his intuitions, in which case why all the appearance of argument? I did write a whole book about this, you know.

s. wallerstein said...

Marcuse is basically a social critic, unlike Rawls. A social critic doesn't work with theorems.

LFC said...

It seems to me that one way to think of the argument in A Theory of Justice might be as follows:

1) Most people want to act justly: they have a 'sense of justice' and at least some
desire to act in accordance w/ it.

2) But most people are too busy in their daily lives to have thought in a systematic way about what their largely intuitive sense of justice actually leads to or means for the way in which society should be set up.

3) The hypothetical contract situation of the original position, although presented in parts of the book as an exercise in bargaining theory, is actually a mechanism or a means for getting the reader to think more carefully about what his/her intuitions about justice (and desert) require or lead to.

4) So the argument assumes the reader starts w/ certain intuitions and that those intuitions can be clarified and systematized w/ the help of the thought experiment that is the original position. There is a passage toward the beginning (p.50, '71 edition) where Rawls says that "everyone has in himself the whole form of a moral conception." The suggestion perhaps is that, to put it metaphorically, there is a moral philosophy in embryo in everyone waiting to emerge w the help of the author-as-guide. Here's a bit more of the passage:

"...if we can describe one person's sense of grammar we shall surely know many things about the general structure of language. Similarly, if we should be able to characterize one (educated) person's sense of justice, we would have a good beginning toward a theory of justice. We may suppose that everyone has in himself the whole form of a moral conception. So for the purposes of this book, the views of the reader and the author are the only ones that count. The opinions of others are used only to clear our own heads."

One might wonder why, if R. were concerned to establish this direct, sort of intimate exchange with the reader, he proceeded to write 600 often dense pages. But perhaps this is one reason why he felt the need, much later, to publish the much shorter Justice as Fairness: A Restatement (which I haven't read).