Read the Masters!

Harold M. Edwards
It appears to me that if one wants to make
progress in mathematics one should study the
masters and not the pupils.
- N.H. Abel (1802-1829), quoted from
an unpublished source by O. Ore in
Niels Henrik Abel, Mathematician
Extraordinary, p. 138.

It is as good an idea to read the masters now as it was in Abel's time. The
best mathematicians know this and do it all the time. Unfortunately,
students of mathematics normally spend their early years using textbooks
(which may be, but usually aren't, written by masters) and taking lecture
courses which are self-contained and make little or no reference to the
primary literature of the subject. The students are left to discover on their
own the wisdom of Abel's advice. In this they are being cheated.
The phrase "read the masters" can be interpreted in two ways. There are
mathematical specialties today which scarcely existed 50 years ago, and in
these specialties "reading the masters" would mean reading the works of
the most important contributors. This is indeed a very worthwhile activity;
it is pursued by the best workers in such fields, but even here students are
often not directed to the primary literature. However, what I have in mind
is a broader view of mathematics as a unified subject that goes back at least
as far as Newton and Leibniz, and that numbers among its masters Euler,
Gauss, Abel himself, Galois, Riemann, Poincare, Hilbert, and others of
past centuries.
It may only be my own prejudice as to what is first rate, but I am
convinced that most first rate mathematicians have this broad view of
mathematics and do read the masters. An outstanding example of this is
provided by Carl Ludwig Siegel's essay on the history of the Frankfurt
Mathematics Seminar [1] where he describes the participation of faculty

Harold M. Edwards is Professor of Mathematics at New York University. He received a B.A.

from the University of Wisconsin in 1956, an M.A. from Columbia University in 1957, and a
Ph.D. from Harvard University in 1961, and has been on the faculties of both Harvard and
Columbia. Edwards's primary interests are number theory and the history of mathematics; his
two recent books-Riemann's Zeta Function (1974) and Fermat's lAst Theorem (l977}-for
which he was awarded the Steele Prize for Mathematical Exposition of the American
Mathematical Society in 1980, explain topics in number theory in terms of their historical
development. He is now at work on a third book on history and number theory.

L. A. Steen (ed.), Mathematics Tomorrow

© Springer-Verlag New York Inc. 1981
106 Harold M. Edwards

and students in Frankfurt in the 1920's and early 1930's in a seminar on the
history of mathematics.
It was the rule in those seminars to study the more important mathematical
discoveries from all epochs in the original: everyone involved was expected to
have studied the text at hand in advance and to be able to lead the discussion
after a group reading. In this manner we studied ancient mathematicians and
for many semesters devoted ourselves to a detailed study of Euclid and
Archimedes. Another time we spent several semesters on the development of
algebra and geometry from the Middle Ages to the mid-17th century, in the
course of which we became thoroughly familiar with the works of Leonardo
Pisano, Vieta, Cardano, Descartes and Desargues. Our joint study of the
ideas from which infinitesimal calculus developed in the 17th century was
also rewarding. Here we dealt with the discoveries of Kepler, Huygens,
Stevin, Fermat, Gregory, and Barrow, among others.

To most of us, the idea of reading Leonardo Pisano (alias Fibonacci,

1180?-1250?) may seem to carry the idea of reading the masters too far, but
the lasting value of the activities of this history seminar is attested to by
more than Siegel's tribute to it as having afforded him "some of the
happiest memories of my life."
The guiding spirit of the history seminar was Max Dehn, one of the
outstanding mathematicians of the 20th century, who made fundamental
contributions to combinatorial group theory, combinatorial topology, and
the foundations of geometry. Dehn's work is characterized by its imagina-
tiveness and the fertility of its ideas. Those who knew him see a direct
connection between these qualities and Dehn's profound interest in the
history of mathematics. His seminar dealt specifically only with mathemat-
ics before the 18th century-perhaps because the mathematics of these
periods belongs equally to mathematicians in all the modern specialties, so
that geometers, algebraists, analysts, topologists, and all could take part-
but for Dehn and his colleagues, the method, the approach, and the belief
in the value of this type of study went far beyond the 18th century.
Perhaps the greatest example of a combination of historical scholarship
and mathematical research is Siegel's discovery of what are now known as
the Riemann-Siegel formulas in the theory of the zeta function. Around
1930, Siegel undertook the study of Riemann's Nachlass in analytic number
theory, the disconnected jottings of formulas on loose sheets which were in
Riemann's papers when he died in 1866, and which are preserved in the
Archives of the Gottingen State and University Library. Because of the
great interest in the zeta function and in Riemann's hypothesis that its zeros
(other than the ones at the negative even integers) all have real part equal
to 1/2, there was great curiosity to know what unpublished information
Riemann might have had about the zeta function. Others had tried before
Siegel to make some sense out of Riemann's disconnected notes, but the
Read the Masters! 107

task required someone who was at the same time a mathematician with
great sophistication and technical ability, and an historian with the dedica-
tion and patience to analyse an original text such as Riemann's. The result
of Siegel's efforts was the astonishing discovery that Riemann possessed
two formulas for the zeta function, each of which was a substantial
contribution to the theory of this function in 1932, seventy-five years after
Riemann's death! It is entirely conceivable that, were it not for Siegel, these
formulas would still be unknown today.
Another major mathematician of the 20th century who is an avowed
reader of the masters is Andre Weil.
As a young normalien I had studied Riemann, and later Fermat; I was
persuaded very early that diligent attention to the great mathematicians of
the past is a source of inspiration no less fertile than the reading of the
fashionable authors of the day.* [2, v. 1, p. 520]
Weil visited Dehn's seminar in Frankfurt in 1926 and "as often as possible"
from then on. He has written that in the sessions of the seminar "I felt I
was taking part in an incomparable intellectual festival"** [2, v. 3, p. 460],
and,
having had the benefit of such an experience, I naturally found myself led, at
the time Bourbaki began his works, to propose that these include historical
commentaries to put in proper perspective the expositions, which were in
danger of falling into excessive dogmatism. For a while, the care of these was
placed mainly in my hands, and the preliminary plans of this nature that I
submitted to Bourbaki were generally approved with a minimum of revisions,
contrary to what always happened with the properly mathematical texts that
Bourbaki received from his collaborators. .. Little by little, Bourba\<i's other
collaborators acquired a taste for this type of work, and my own participation
became more and more sporadic ... *** [2, v. 3, p. 460-461]
Thus, through Weil, the collaborators of Bourbaki-perhaps even Bourbaki
himself-became readers of the masters.

• Jeune normalien, j'avais etudie Riemann, puis Fermat; je m'/uais tot persuade que la
frequentation assidue des grands mathematiciens du passe est une source d'inspiration non moins
a
feconde que la lecture des auteurs la mode du jour.
• * ... j' eus Ie sentiment de prendre part a une incomparable fete intellectuelle.
*.* Ayant eu Ie benefice d'une pareille experience, je me trouvai naturellement porte, lorsque
Bourbaki commenc,a ses travaux, aproposer d'y faire figurer des commentaires historiques pour
replacer dans une juste perspective des exposes qui risquaient de tomber dans un dogmatisme
a
excessif. Pendant un certain temps, c'est moi surtout qu'en incomba Ie soin, etles avant-projets
a
de celie nature que je soumis Bourbaki furent generalement approuves avec un minimum de
a
retouches, contrairement ce qui s' est toujours passe pOUl' les redactions proprement mathemati-
ques que Bourbaki recevait de ses collaborateurs. . .. Peu a peu, d' autres collaborateurs de
Bourbaki prirent gout a ce genre d' exposes, et ma participation devint de plus en plus sporadi-
que ....
108 Harold M. Edwards

There are several reasons that the best mathematicians are attracted to
the classics. For one thing, they are, by nature, synthesizers. The essence of
mathematics is the perception of logical relationships between seemingly
disparate ideas; the most successful mathematicians are those whose range
is widest and whose ability to see analogies and connect ideas is greatest.
The role of history in such activity is obvious. Often the analogies have
been recognized and developed for generations, and often the connections
are most easily made via a common historical antecedent. More often than
we like to imagine, advances result not from new ideas, but from the
realization that old ideas apply in new circumstances.
A colleague of mine, who agreed that the best mathematicians study
history, suggested yet another reason. "Good mathematicians work on good
problems," he said, "and good problems have histories." He might have
added that it is often the history that makes it a good problem, because it
establishes the connections with the solutions of other problems and with
applications both inside and outside mathematics.
Of course, all the usual reasons for studying history of any kind apply
equally well to mathematics. History is the accumulated experience of
mankind, and no intelligent person ignores it. It has shaped what we are,
and it is our best source of information as to what is possible-or impossi-
ble. Countless aspects of our complicated and mystifying world are ren-
dered intelligible only by an understanding of how they came to be as they
are.
In mathematics, however, the importance of history is even greater.
Mathematicians talk about mathematical "objects" and like to think that
there is an objective truth to the theorems that they prove. But no one has
ever seen one of these "objects" and in the last analysis the only way one
can learn about them is to read about them in the works of others. For this,

Mathematics, like philosophy, is virtually inseparable from its

history.

one naturally is well advised to read the works of the most clear-sighted
authors, the masters. From this point of view, mathematics, like philosophy,
is virtually inseparable from its history. The central concepts, the problems
to be studied, and the way the theories are organized, are all inextricably
associated with the names and works of the authors who first advanced
them.
Finally, a simple reason for the attraction of the best mathematicians to
history, which may supersede all the others, is that history is so interesting.
First rate mathematicians are, naturally, passionately interested in mathe-
matics, and are consequently interested in the great mathematics of the
past, which is better than a lllajor part of the mathematics of the present.
Read the Masters! 109

And the stimulation of interest does not go in only one direction. Many
times mathematicians whose interest in their own work was flagging have
found new inspiration by reading the classics.
For· all these reasons, reading the masters is an important part of
mathematical life and should, therefore, be an important part of mathemat-
ical education. Unfortunately, it seldom is. The main reason, no doubt, is
lack of time. American graduate students need to learn so much in so little
time and need to start doing research for a dissertation so soon that there is
no time for the classics. In this situation, it is the responsiblity of the faculty
to give students guidance, to explain that it is impossible to learn everything
at once, to make them aware of the value of reading the classics, and to
help them acquire the skills necessary to do so.

One must bring to older writings an understanding of the

mathematics of the period in which they were written.

To read the classics in mathematics is not as easy as it might sound.

Changes in terminology, in style, and in the general mathematical milieu
make mathematics (regarded, ironically, by the man in the street as an
immutable body of facts) a field in which works written only a decade or
two ago often sound archaic to young readers. One must bring to older
writings an understanding of the mathematics of the period in which they
were written-the assumptions their writers made about the reader as
regards basic concepts, known theorems, problems to be considered, cus-
tomary methods of proof, and so forth. It is not easy for students to acquire
this kind of background information. Patient study is the only method.
Courses in the history of mathematics cannot normally give enough atten-
tion to anyone field of modern mathematics to give specialists in that field
all the information they need to read the classics. It is of great help to find
a thesis supervisor who has a good grasp of the classical literature in the
field and who will direct the student's reading. Unfortunately, even supervi-
sors who do know the classics well often fail to see the importance of
opening them up to their students, and, even more unfortunately, there are
many supervisors who have only a superficial knowledge of the classics.
Another skill that students must develop in order to read the masters is
the knowledge of languages. Before WorId War II, very few of the masters
wrote their works in English. If American students are going to study these
works, they will have to study foreign languages, too. This important
component of education-useful in all areas of mathematical research, not
just in reading the classics-is increasingly ignored today, to the long-term
disadvantage of our students. I have knowl!young Ph.D. mathematicians
who specialized in number theory, a field in which at least 80% of the
classical literature is in German, and who were allowed to enter adult life
110 Harold M. Edwards

without even the ability to work their way through a German mathematical
text with the aid of a dictionary. They hope the classics will be translated,
but this is unlikely, because the handful of people who have both the
linguistic and the mathematical ability to make adequate translations are
unlikely to want to spend the needed time and effort for the benefit of the
handful of potential readers.
The continued health and vitality of mathematics depends on leaders in
the mathematical community who are versatile and active in several areas
of mathematics and who are therefore able to oppose the tendency toward
fragmentation that results from there being so many specialties that are
pursued so intensely. There are many such leaders today, and we owe to
them several wonderful syntheses in recent years that have brought together
techniques from separate specialties, thereby enhancing these techniques
and increasing their applicablity. The generation now being educated will
be more likely to produce such leaders if they heed Abel's advice and read
the masters.

References

[I) c.L. Siegel. "Zur Geschichte des Frankfurter Mathematischen Seminars."

Frankfurter Universitiitsreden, 1964, Heft 36; also in Gesammelte Abhandlungen,
V. 3, pp. 462-474. Authorized English translation in The Mathematical Intelli-
gencer 1:4 (1979) 223-230.
[2) A. Weil. Andre Wei!: Oeuvres Scientifiques (Collected Papers). Springer-Verlag,
New York, 1979.