Professional Documents
Culture Documents
MAZUR
CONJECTURE
A friend of mine recently told me that the Dutch word for Mathematics is
Wiskunde, a word coined in the 16th century by Simon Stevins. My friend
suggested that it was derived from “Wis” meaning “sure” and “kunde”
meaning “the study of”. Carrying this word back into English, we would
come up with something like “Sure-ology”. But even if the subject, Math-
ematics, is named “Sure-ology”, its history would be poorly described by
such a name. For like any Intellectual History, lots of the History of Math-
ematics is simply never captured: its major artifacts are ideas which spend
much of their life in a volatile, unrecorded, state. Their eventual distillation
as written record occurs long after their initial discovery.
The presence of “Conjecture” as a pointedly formal element in the syntax
of recent Mathematics may have the effect of counteracting this a bit, of
bottling some vapours of mathematical ideas at an earlier stage than was
previously done. Might this help in the historiography of Mathematics?
When Aki Kanamori telephoned me asking whether I might be inter-
1
ested in giving a talk in the BU conference, I had been reading a book
(Rapoport 1989) entitled Beilinson’s Conjectures on Special Values of L-
functions. Telephone receiver in hand, I was suddenly struck by the thought
that such a book, broadcasting “conjectures” in its very title, represents a
large change in the way mathematical theories are worked on; this change
has not been all that widely written about. At least I think it hasn’t. Here we
have a book in Pure Mathematics, a whole book, fully devoted to the
exposition of an extensive constellation of conjectures! It, and other math-
ematical works like it, signal that the art of conjecturing has achieved a
formidable, and quite formal, prominence in the mathematical landscape. I
believe that “formalization of conjecture” is a good thing for Mathematics,
and an inevitable thing, as large mathematical theories grow larger.
There is something perplexing and delicious when one finds a rarely
discussed theme which one imagines to be interesting. The temptation to
destroy its deliciousness by actually talking about it is then hard to resist!
2. ARCHITECTURAL CONJECTURES
Having spoken out “for” conjecture, perhaps I should also quote some
cautionary prose about its careless practice. The author here is Andre´
Weil, who in the notes he wrote as addenda to his Collected Works (Weil
1979) gives himself
2
l’occasion de dire mon sentiment sur ce mot [conjecture ] dont on a tant use´ et abuse´. Sans
cesse le mathematicien´ se dit: “Ce serait bien beau” (ou: “Ce serait bien commode”)
200 B. MAZUR
si telle ou telle chose etait´ vraie. Parfois il le verifie´ sans trop de peine; d’autre fois il ne
tarde pas a` se detromper´. Si son intuition a resist´e´ quelques temps a` ses efforts, il
tend a` parler de “conjecture”, memeˆ si la chose a peu d’importance en soi. : : : En tout
cas, s’il m’appartenait de donner un conseil a` qui ne m’en demande point, je
recommanderais d’employer desormais´ le mot de “conjecture” avec un peu plus de
circonspection que dans ces derniers temps.
Rough translation:
: : : the occasion to vent my feelings regarding that word [conjecture] which has been
very much used and abused. The mathematician continually says “it would be beautiful”
(or: “it would be convenient”) if this or that were true. Sometimes he verifies it without
too much trouble; on other occasions he loses no time in correcting his misapprehension.
If the thing resists his efforts for some while, he tends to call it a “conjecture” even if it
has little importance in itself. In any event, if it is up to me to give advice (to those who
never asked for it) I would recommend using the word “conjecture” with a bit more
circumspection than has been done recently.
But perhaps there are other reasons to be wary, as well. During the pan-
el discussion at the BU colloquium, David Kazhdan, responding to the
statement that “conjecture” is a way for present-day researchers to
imprint their intent, their goals, on future researchers, cautioned that if
the practice achieved too much prominence, it, like the “five-year plans”
of the former Soviet Union might eventually have a stultifying effect.
Assuming, for a moment, that “conjecture” has achieved the position
in Mathematics that I claim it has, what are the forces at play which
contribute to the rise of conjecture? Here are some possibilities:
Rough translation:
: : : I already emphasized in discussion of the methods of Mathematics that it is a thoroughly
legitimate procedure to construct some special theory dependent upon admittedly unproven,
CONJECTURE 201
but plausible, theorems (as axioms), provided one is clear about the incomplete nature of
the foundations of that theory.
If one had to single out one driving source responsible for important con-
jecturing, the most potent is simply “analogy”. Algebra is “like”
Geometry and perceiving this analogy, one (Weil, in fact) developed the
classical “Weil Conjectures” of Algebraic Geometry which predicted an
arsenal of topological tools, fixed point theorems, etc., to be used to
count numbers of solutions of equations over finite fields.
4
Complex function theory is “like” Arithmetic and perceiving this
analo-gy, one (Paul Vojta, in fact) developed the more recent “Vojta
Conjectures” which give very precise arithmetic predictions concerning
the structure of solutions of equations over the ring of integers, and the
field of rational numbers.
Concerning the romance of analogy in mathematics and its inevitable
end, listen to Weil:
Rien n’est plus fecond,´ tous les mathematiciens´ le savent, que ces obscures analogies, ces
troubles reflets d’une theorie´ a` une autre, ces furtives caresses, ces brouilleries inexplicables;
rien aussi ne donne plus de plaisir au chercheur. Un jour vient ou` l’illusion se dissipe; le
pressentiment se change en certitude; les theories´ jumelles rev´elent` leur source commune
avant de disparaˆitre; comme l’enseigne la Gita on atteint a` la connaissance et a`
l’indifference´ en memeˆ temps. La metaphysique´ est devenue mathematique,ˆ preteˆ a`
former la matiere` d’un traite´ dont la beaute´ froide ne saurait plus nous emouvoir´.
Rough translation:
202 B. MAZUR
Nothing is more fruitful – all mathematicians know it – than those obscure analogies, those
disturbing reflections of one theory on another; those furtive caresses, those inexplicable
discords; nothing also gives more pleasure to the researcher. The day comes when this illusion
dissolves: the presentiment turns into certainty; the yoked theories reveal their common source
before disappearing. As the Gita teaches, one achieves knowledge and indifference at the same
time. Metaphysics has become Mathematics, ready to form the material of some treatise whose
cold beauty has lost the power to move us.
One of the impressive things about the Hilbert Problems is that many of
them bridge the gap between what one might call “problem-oriented
research” and “programmatic research”. Some of them are simply “prob-
lems” of a traditional nature. I read Hilbert’s beautiful and quite open-ended
16th problem about the nestedness of the connected components of a real
algebraic curve, and the limit cycles of differential equations, as being of
that sort. But underlying the formulation of the entire collection of Prob-
lems is the clear aim of setting programmatic directions, of forecasting large
theories and research programs. If you doubt this, just consider the
CONJECTURE 203
Even if one thinks of the Hilbert Problems as giving the “research pro-
grammatic” spirit a serious boost, one can also see this spirit arising with
much energy in a number of different mathematical works at the turn of the
century. Consider, for example, the way in which Poincare´ first asks (in
Poincare´ 1953, his 1895 article ‘Analysis Situs’) the question that eventu-
ally became focussed (and corrected in the fifth complement to that article)
giving rise to what we now call the (three-dimensional) “Poincare´ Conjec-
ture”. Poincares´ article represents the birth of modern algebraic topology.
He had constructed the fundamental group of a manifold earlier in the arti-
cle, and then writes that “it would be interesting to treat” three questions. To
paraphrase them, Poincare’s´ questions are:
Given any group, is it the fundamental group of a manifold in n dimensions? How can
one construct that manifold? Is the manifold determined by its fundamental group?
The work on these famous conjectures dating from the 40’s, parts of which
were resolved by Bernard Dwork, and by Alexander Grothendieck who set
up a formidable apparatus to deal with them, was completed by Pierre
Deligne in 1974. The conjectures provide a bridge between the fields of
algebraic topology and number theory. To talk about this bridge with a
CONJECTURE 205
I mean “by hypothesis” ["’ ‘ o " "!s] what the geometricians often do in dealing with a
question put to them; for example, whether a certain figure is capable of being inscribed
as a [: : : ] in a given circle: they reply- “I cannot yet tell whether it has that capability;
but I think, if I may put it so, that I have a helpful hypothesis for the problem, and it is as
follows: If [: : : ] then it seems to me that we can conclude one thing, but if [ : : : ] then
some other. Hypothesizing thus, I will tell you the conclusion about the inscription of it
in the circle, whether it is impossible or not”.
If you are puzzled by exactly what the activity of “hypothesizing” is, that
is described in the above quotation, you are not alone. There is a vast
9
secondary literature related to this paragraph:
That “hypothesis” – however it was employed by the mathematicians
– was not a thoroughly harmless device in the eyes of Plato’s Socrates of
10
Book VI of The Republic is clear from the pronouncement (511 D):
And though it is true that [ : : : the mathematicians : : : ] are compelled to use their under-
standing and not their senses, yet because they do not go back to the beginning [ "’ ’ ] [
: : : ] but start from hypotheses you do not think they possess true understanding : : : .
in general colloquial use through the ages. But it seems to have entered the
scientific or mathematical vocabulary only recently. Since I am not a his-
torian of Mathematics I dare not make any serious pronouncements about
the historical use of the term, but I have not come across any appearance of
11
the word Conjecture or its equivalent in other languages with the above
12
meaning in mathematical literature except in the twentieth century. The
earliest use of the noun conjecture in mathematical writing that I have
encountered is in Hilbert’s 1900 address, where it is used exactly once, in
reference to Kronecker’s Jugendtraum.
208 B. MAZUR
Why not? I think it was that even in Riemann’s time there was too little agreement on
standards of proof. Even the handful of professors of math in Germany could not agree
with each other on which purported proofs of Riemann–Roch were good and which were
useless.
To distinguish conjectures from proofs you need generally accepted standards. These
need not be “right” in any given sense, certainly need not be formal. But they need to be
generally accepted to give a meaningful distinction.
Hilbert could talk about “conjectures” because the Gottingen¨ school at his time was
developing a shared standard and intended to impose it on the profession – and in fact
their standard evolved in a fairly continuous line through Noether and van der Waerden
to Bourbaki.
So this brings me back to your Gaudiesque cathedrals. Shared standards of rigor can only
arise where there are large shared projects, and vice versa. I come to think of “cathedrals” and
“rigor” as the social and theoretical sides (respectively) of one development”.
NOTES
1 For which this article is the written text.
2 Anything occurring in square brackets [: : : ] within some text I am quoting is either an
indication that I have omitted some words, or else is a comment of mine.
3 I am thankful to Leo Corry for mentioning this to me, and for enlightening observations
about it.
4 Specifically, the theory of Nevanlinna.
5 As in the fragment of the ancient poem by Archilocus of Paros: “The fox knows many
things. But the hedgehog knows one big thing”.
6 As Robert Kaplan pointed out to me, there are a few other labels, some no longer in
use, like the Greek porism (which, according to the footnote on page 9 of (Lakatos 1981)
refers to a result which was incidentally, or accidentally, arrived at on the way to
something else). Kaplan suggests that we might distinguish two strains in the tradition of
mathemat-ical labelling: one strain being distinctly subjective, or “biographical” ( as in
labels like porism which commemorate historical events in the theory’s development, or
as in the discriminating use of the word axiom as opposed to postulate) and the other
strain being “architectural” (as in the standard use of lemma). The most telling example
of “historical” labelling that I can think of occurs in the Gettysburg Address, where the
self-evident truth of the Declaration of Independence (that all men are created equal) is
re-asserted as a Proposition (Lincoln’s grim choice of words emphasizing that he is,
indeed, putting it to the proof).
7 An English translation of this treatise occurs as the appendix to (Klein 1968).
8 The English translation given is by W. R. M. Lamb (Plato 1927).
9 I want to thank Robert Kaplan for his unpublished manuscript “A short history of the
unknown” which contains a discussion of this paragraph in a broader context.
10 The English translation given is by Paul Shorey (Plato 1935).
11 For example, in German, Vermutung: What is the earliest occurrence of the
label “Rie-mannsche Vermutung” for the Riemann Hypothesis, for example? As for
Greek, it seems that there is no word which cleanly corresponds to conjecture. If, in the
crudest way, one were to try to haul the word conjecture into Greek by force, one might
expect to end up with a nominalized form of the verb " (?). Does such a word occur in
the Greek mathematical literature?
210 B. MAZUR
12 Nicolas of Oresme does use the term “conjectura” in his treatise ‘The
Geometry of Qual-ities of Motion’ but when he uses it, he means “prophesying”.
13 Which I am reproducing with his approval.
REFERENCES