Following on Anthony Waterman's comment, I tried to get in touch with
Daniel Ellsberg. He was kind enough to let me make public the following
e-mail message. It does not help much in clarifying if there was any
peculiar Harward's practice in the acceptance of dissertations, but
provides interesting autobiographical remarks and indicates who was
Ellsberg's supervisor.

Carlo Zappia

----------------------------

Dear Carlo: John Chipman was my thesis supervisor for my undergraduate
honors thesis, Theories of Rational Choice Under Uncertainty: The
Contributions of von Neumann and Morgenstern.  John had done some informal
experiments on choice himself, which I acknowledged in the thesis and
probably (though I don't recall) in the article which followed from it, on
utility theory, in the Economic Journal.  (Another article, Theory of the
Reluctant Duellist, in the AER was also drawn from the thesis).  For my
Ph.D. thesis, written while I was at RAND in 1962, my nominal supervisor
was my friend (and RAND consultant) Tom Schelling.  However, I don't
recall any discussion of my work with him while I was writing it (in
California) or even after it was written, although he read it.  It was
outside his line of theoretical interest.  His influence on my thinking
about bargaining theory, especially from his first article on the subject,
was very great; our minds worked in parallel paths (though at that time he
had almost no knowledge of the work of von Neumann and Morgenstern, i.e.,
"classical" game theory) but his extremely original ideas stimulated and
influenced my thinking very much.  But that was in the area of my lectures
on The Art of Coercion for the Lowell Institute in 1959 (still
unpublished, though several of the lectures have been reprinted) and later
work, still unpublished.  It didn't bear on my thesis.

Come to think of it, Schelling did have an influence on my publishing of
my work on ambiguity.  He had read an early version of this paper and
wanted to include a piece on this subject in an issue of the Quarterly
Journal of Economics for which he would be the guest editor.  I agreed,
but, as usual, I didn't get around to it, till the deadline was close. (I
should say, as usual for my later life. My practice of procrastination
about publishing  started just about then--not earlier, as my earlier
articles indicated.  In fact, I could say it startedwith a decision to put
off (indefinitely, it turned out) editing and publishing my lectures on
the Art of Coercion.  I returned an advance I had received for that from
McGraw-Hill, in early 1961.  I thought what I was doing on nuclear war
planning at RAND took priority over editing these).

Tom, the great bargaining theorist, then told me: "Don't worry about it,
Dan.  I'll just publish what I have, the piece you gave me."  That sent me
into a panic--that draft wasn't adequate!--and I went into a
round-the-clock crash effort to turn out the published piece, Risk,
Ambiguity and the Savage Axioms, in time for his issue.  Then, in the
beginning of 1962, I devoted myself full-time--with the generous
encouragement and support of RAND--to absorbing the literature in my
thesis bibliography and writing the thesis, again around-the-clock, in
just four months. (I would have thought it was somewhat longer, but I know
I started at the first of the year, and I see--looking just now at a bound
Harvard Library copy which has somehow found its way onto my shelves--that
it's dated April, 1962.)  The deadline here was my 10th Class Reunion at
Harvard in June.  Following Schelling's theories of commitment (really,
not just rhetorically), I announced in my class notes distributed to the
class in the spring that I  would receive my degree in June, in time for
the Reunion.  I explained in the notes that, by announcing this, I was
assuring myself of unspeakable humiliation if I failed to carry it out. 
That wasn't just a joke. And by  truly vast effort (the reading alone took
about four months, as one can tell from the bibliography) in the final
stages, working all night every night at RAND, I did send it in: to
Harvard, and to Schelling (though I don't recall any comment from him on
it, except that it looked good).   Just before commencement, as I recall,
in late May or June, I defended the thesis before a panel of, I think,
four readers (not including Schelling).  They accepted it, and I did march
in Commencement with my Ph.D. robes, with the congratulations of my
classmates.

One odd thing I've never been clear on: I've always regarded myself as
receiving the Ph.D. in June, 1962, and I certainly did rent (or buy?) a
Ph.D. robe for the ceremony, and had certainly passed all the
requirements; but at some point years after that I got some indication
that Harvard carried me in their records as receiving it in early 1963
(though I was back at RAND after June and was not aware of any delay).  It
doesn't matter, of course, but I've occasionally worried that someone
would contradict me, on the basis of a record search, when I give 1962 as
the date of my degree, as I always do.  (I probably have my sheepskin
somewhere; I could look that up and see what it says).

Does that answer the question?  In sum, I did have a supervisor, formally,
Schelling; but I see that I didn't mention this in the thesis or give him
any special acknowledgement, since in this particular area he didn't
influence my thinking.  Nor did Chipman, at all, in this area: I hadn't
been in touch with him for a number of years, before I started this line
of work (in 1957-1958, while I was in the Society of Fellows).  In my 
acknowledgments, I see I listed Schelling as one of ten named colleagues
with whom I had had useful discussions: alongside longer acknowledgments
to others.

One other thought, since I mentioned my unfortunate "publication-block"
earlier: my failure to publish this thesis later in 1962 or 1963 was
regrettable.  As I mentioned in the preface, only about 10% of it was
represented in my earlier article.  I've learned since that most
specialists who had heard I had written a thesis made no effort to see it,
because they assumed that it had been written before my article and that
the article was a summary of it.  Too bad; I have the impression that it
would have been quite useful in the '60's and even 70's (and really,
unless I'm  mistaken, I think there's still analysis and propositions in
it that would be helpful even at this late date, that makes it worth
reading.  I regret that there hasn't been any review of it, to my
knowledge, in the professional literature).  As I hold the two books, the
Garland and the original thesis, in my hands now, I recall my reason for
not publishing it earlier.  The Harvard manuscript is 379 pages (of
typescript, which makes it thicker).  I thought that was too long to
publish, and I intended to cut it down.  But when Garland published it,
without any editing or bringing up to date (Isaac Levi contributed a
commentary, and Mark Machina generously provided an updated bibliography),
my text runs only 284 thin printed pages, a nice size.  Rats!  Oh well...

Daniel Ellsberg