GENERAL ARTICLE

Freeman Dyson, the Mathematician∗

B Sury

A mathematician of the highest rank

with ideas ever-brimming with a full tank.
He designed special lenses
to ‘see’ Ramanujan’s congruences
partitioned through the yet-invisible crank!

The one person who comes closest to the legacy of Hermann B Sury is with the Indian
Statistical Institute in
Weyl was Freeman Dyson, who contributed enormously to
Bengaluru. He is the National
both Physics and Mathematics. His books and talks are tes- Co-ordinator for the
taments to his prolificity in writing widely on the world at Mathematics Olympiad
large and on science and mathematics in particular. To name Programme in India, and also
the President of the Indian
a few of his books, he has (talked and) written on ‘Bombs and
Mathematical Society.
Poetry’, ‘Imagined Worlds’, ‘Origins of Life’ and ‘Birds and
Frogs’. Remarkably, Dyson contributed handsomely to what
is termed ‘pure mathematics’. One would expect a physicist-
mathematician to interest himself mainly in problems of an
‘applied’ nature. Here, we take a necessarily brief peek into
some of his ‘purely mathematical’ work. The suggested read-
ing at the end can be referred to for more details.

1. Statistical Mechanics – Dyson’s Conjecture

In an enormously influential paper, ‘Statistical theory of energy Keywords

levels of complex systems’, I. J. Math. Phys., 3:140–156, 1962, Congruences for partitions, rank
and crank, random matrix the-
Dyson proposed a new type of ensemble transforming the whole
ory, Riemann zeta function, Dyson
subject into purely group-theoretic language. In the paper, Dyson transform, quasi-crystals.
studied new types of statistical ensembles, demonstrating a math-
ematical idealisation of the concept of considering “all physical

∗
Vol.25, No.10, DOI: https://doi.org/10.1007/s12045-020-1055-2

RESONANCE | October 2020 1339

 GENERAL ARTICLE

In an enormously systems with equal probability”. He studied three such ensem-

influential paper, bles based on the orthogonal, unitary and symplectic groups. He
‘Statistical theory of establishes an exact mathematical correspondence between the
energy levels of complex
systems’, Dyson distribution of eigenvalues of a random matrix and the distribu-
proposed a new type of tion of positions of charges in a finite Coulomb gas at a finite
ensemble transforming temperature T . The corresponding temperature T is 1, 1/2 or 1/4
the whole subject into
according to as to whether the matrix is taken from an orthogonal,
purely group-theoretic
language. unitary or symplectic ensemble.
In this paper, Dyson arrives at some purely mathematical conjec-
tures one of which is the so-called ‘Dyson conjecture’ (Conjec-
ture C in the paper), asserting that for positive integers a1 , · · · , aN ,
function i, j (1 − zi /z j )ai is the
Q
the constant term of the rational
P
( a )!
multinomial coefficient Qii(aii )! . The conjecture was proved soon
afterwards but the richness of the idea can be divined from the
fact that the conjecture was generalized by Macdonald to all root
systems (Dyson’s conjecture corresponding to the type An ) and
proved later by Cherednik using double affine Hecke algebras. As
a matter of fact, the conjecture is a mathematical reformulation of
another conjecture (Conjecture A) which specifies precisely the
statistical properties of a Coulomb gas of N particles.

1.1 A Simple Proof of Dyson’s Conjecture

Here is a ‘good’ proof of the conjecture. The proof by I J Good

uses just Lagrange’s interpolation formula.
Write f (x, a1 , · · · , an ) = i, j (1 − xi /x j )ai .
Q

Now, interpolation formula gives the function that equals 1 at all

xi ’s to be
X Y t − xi
.
x − xi
j i:i, j j

Putting t = 0, we have

XY
1= (1 − x j /xi )−1 .
j i:i, j

1340 RESONANCE | October 2020

 GENERAL ARTICLE

Multiplying f (x, a1 , · · · , an ) = i, j (1 − xi /x j )ai by the above

Q

expression for 1, we get

X
f (x, a1 , · · · , an ) = f (x, a1 , · · · , a j − 1, · · · , an )
j

where the right hand side expression has terms where the variable
a j is reduced by 1 and other variables the same. If the sought for
constant term of f (x, a) is denoted by g(a), then
X
g(a) = g(a1 , · · · , a j − 1, · · · , an ) · · · (♠)
j

Now, clearly if a j = 0, the term x j occurs with a strictly nega-

tive power in f (x, a), and would, therefore, not contribute to the
constant terms. In other words,

g(a1 , · · · , an ) = g(a1 , · · · , a j−1 , a j+1 , · · · , an ) i f a j = 0.

Clearly g(0, · · · , 0) = 1. Thus, (♠) along with the last two equal-
ities determine the general value g(a) recursively. The multino-
mial coefficient
P
( ai )!
M(a1 , · · · , an ) = Q i
i (ai )!

satisfies the same recursion and initial value. Thus, g = M.

2. Random Matrices and Zeta Zeroes

The story is supposed to have started in 1972 when the analytic

number theorist Hugh Montgomery met Dyson at the Institute for
Advanced Study—this conversation is sometimes called ”one of
the most celebrated denouements in mathematics in recent years.”
In Montgomery’s words,
“I [Montgomery] took afternoon tea that day in Fuld Hall with
[Sarvadaman] Chowla. Freeman Dyson was standing across the
room. I had spent the previous year at the Institute and I knew him
perfectly well by sight, but I had never spoken to him. Chowla
said: ”Have you met Dyson?” I said no, I hadn’t. He said: ”I’ll
introduce you.” I said no, I didn’t feel that I had to meet Dyson.

RESONANCE | October 2020 1341

 GENERAL ARTICLE

Chowla insisted, and so I was dragged reluctantly across the room

to meet Dyson. He was very polite, and asked me what I was
working on. I told him I was working on the differences between
the non-trivial zeros of Riemann’s zeta-function, and that I had
developed a conjecture that the distribution function for those dif-
ferences had integrand 1−sin2 (πu)/π2 u2 . He got very excited. He
said: ”That’s the form factor for the pair correlation of eigenval-
ues of random Hermitian matrices!”

When Montgomery In other words, when Montgomery described his observations on

described his the pair correlations of zeroes of the Riemann zeta function seem-
observations on the pair ing to follow a certain distribution, Dyson immediately responded
correlations of zeroes of
the Riemann zeta by saying that it was exactly the pair-correlation function for the
function seeming to eigenvalues of a random Hermitian matrix, and also a model of
follow a certain the energy levels in a heavy nucleus like Uranium 238. This
distribution, Dyson
amazing ‘observation’ is contingent upon assuming the Riemann
immediately responded
by saying that it was hypothesis and could perhaps be turned around to attack1 the Rie-
exactly the mann hypothesis? The idea was already proposed by Hilbert and
pair-correlation function Polya that the zeroes of the Riemann zeta function could perhaps
for the eigenvalues of a
be the same as the eigenvalues of an appropriate Hermitian oper-
random Hermitian
matrix, and also a model ator.
of the energy levels in a
heavy nucleus like
Uranium 238. 3. Birds and Frogs

In a very incisive article titled ‘Birds and Frogs’, Dyson talks

about two types of mathematicians. He says, “ Birds fly high in
the air and survey broad vistas of mathematics out to the far hori-
zon. They delight in concepts that unify our thinking and bring
together diverse problems from different parts of the landscape.
1 As a mathematician, I use the Frogs live in the mud below and see only the flowers that grow
word ‘attack’ in the sense of nearby. They delight in the details of particular objects, and they
trying to prove it or rather than
solve problems one at a time. I happen to be a frog.”
attack its proposed veracity!
“The world of mathematics is both broad and deep, and we need
birds and frogs working together to explore it.”
“Rene Descartes was a bird, and Francis Bacon is a frog.”
In ‘Birds and Frogs’, Dyson writes of a viable method to attack
the Riemann hypothesis:

1342 RESONANCE | October 2020

 GENERAL ARTICLE

“A quasi-crystal is a distribution of discrete point masses whose Suppose that we find one
Fourier transform is a distribution of discrete point frequencies. of the quasi-crystals in
Or to say it more briefly, a quasi-crystal is a pure point distri- our enumeration with
properties that identify it
bution that has a pure point spectrum.... Here comes the con- with the zeros of the
nection of the one-dimensional quasi-crystals with the Riemann Riemann zeta-function.
hypothesis. If the Riemann hypothesis is true, then the zeros of Then we have proved the
Riemann Hypothesis and
the zeta-function form a one-dimensional quasi-crystal according
we can wait for the
to the definition. They constitute a distribution of point masses telephone call
on a straight line, and their Fourier transform is likewise a distri- announcing the award of
bution of point masses, one at each of the logarithms of ordinary the Fields Medal.
prime numbers and prime-power numbers.... My suggestion is
the following. Let us pretend that we do not know that the Rie-
mann Hypothesis is true. Let us tackle the problem from the other
end. Let us try to obtain a complete enumeration and classifica-
tion of one-dimensional quasicrystals. That is to say, we enumer-
ate and classify all point distributions that have a discrete point
spectrum...We shall then find the well-known quasi-crystals asso-
ciated with PV2 numbers, and also a whole universe of other qua- 2 Here PV stands for Pisot and

sicrystals, known and unknown. Among the multitude of other Vijayaraghavan.

quasi-crystals we search for one corresponding to the Riemann

zeta-function and one corresponding to each of the other zeta-
functions that resemble the Riemann zeta-function. Suppose that
we find one of the quasi-crystals in our enumeration with proper-
ties that identify it with the zeros of the Riemann zeta-function.
Then we have proved the Riemann Hypothesis and we can wait
for the telephone call announcing the award of the Fields Medal.
These are of course idle dreams. The problem of classifying one-
dimensional quasi-crystals is horrendously difficult, probably at
least as difficult as the problems that Andrew Wiles took seven
years to explore. But if we take a Baconian point of view, the
history of mathematics is a history of horrendously difficult prob-
lems being solved by young people too ignorant to know that they
were impossible. The classification of quasi-crystals is a worthy
goal, and might even turn out to be achievable.”

RESONANCE | October 2020 1343

 GENERAL ARTICLE

4. Partitions and Crank

One of the most riveting short papers by Dyson, titled ‘Some

guesses in the theory of partitions’, was published in the Eureka
“Professor Littlewood, Magazine in 1944. He begins the article thus:
when he makes use of an “Professor Littlewood, when he makes use of an algebraic iden-
algebraic identity,
tity, always saves himself the trouble of proving it; he maintains
always saves himself the
trouble of proving it; he that an identity, if true, can be verified by anybody obtuse enough
maintains that an to feel the need of verification. My object in this following note
identity, if true, can be is to confute this assertion.”
verified by anybody
“After a few preliminaries, I state certain properties of partitions
obtuse enough to feel the
need of verification. My which I am unable to prove; these guesses are then transformed
object in this following into algebraic identities which are also unproved, although there
note is to confute this is conclusive numerical evidence in their support; finally, I in-
assertion.....I indulge in
dulge in some even vaguer guesses concerning the existence of
some even vaguer
guesses concerning the identities which I am not only unable to prove but also unable
existence of identities to state. I think this should be enough to disillusion anyone who
which I am not only takes Professor Littlewood’s innocent view of the difficulties of
unable to prove but also algebra. Needless to say, I strongly recommend my readers to
unable to state. I think
this should be enough to supply the missing proofs, or even better, the missing identities.”
disillusion anyone who
For the partition function p(n) that counts the number of ways to
takes Professor
Littlewood’s innocent partition n into (unordered) sums of positive integers, Ramanu-
view of the difficulties of jan’s famous congruences assert the divisibility properties:
algebra.”
– F Dyson 5|p(5n + 4), 7|p(7n + 5), 11|p(11n + 6) ∀ n.

Although they were proved analytically, one would like natural

counting functions that would count p(5n + 4)/5, p(7n + 5)/7, and
P(11n + 6)/11. In the Eureka paper, Dyson defined the notion
of the rank of a partition to be the largest part minus the num-
ber of parts. Dyson conjectured that the number of partitions of
5n + 4 with rank equal to each given congruence class mod 5 is
equally numerous (that is, independent of the class), thereby ex-
plaining the congruence 5|p(5n+4). He made a similar conjecture
about the rank belonging to the 7 congruence classes mod 7 for
partitions of 7n + 5. Both these were proved later by Atkin and
Swinnerton-Dyer. Dyson observed that the rank does not sepa-
rate the classes mod 6 while considering the partitions of 11n + 6.

1344 RESONANCE | October 2020

 GENERAL ARTICLE

Here, Dyson came up with a brilliant idea. He looked at vari- Dyson looked at various
ous tables of partition functions and “guessed” the existence of a tables of partition
more esoteric hypothetical function he christens the ‘crank’ of a functions and “guessed”
the existence of a more
partition, which would explain the congruences for p(11n + 6). esoteric hypothetical
In 1988, George Andrews and Frank Garvan defined the crank of function he christened
a partition as follows. It is defined to be the largest part if the the ‘crank’ of a partition,
which would explain the
partition has no 1’s. If there are ω > 0 one’s in the partition,
congruences for
then the difference “(the number of its parts bigger than ω)—ω p(11n + 6).
is defined to be the crank. They showed that with this notion of
the crank, we have the correct analogue for p(11n + 6)/11 of the
above results with rank for p(5n + 4)/5 and p(7n + 5)/7.

5. Missed Opportunities

In a wonderful talk (the Gibbs lecture published in the Bulletin of

the American Mathematical Society, 1972) titled ‘Missed Oppor-
tunities’, Dyson says wistfully:
“I shall examine in detail some examples of missed opportunities,
occasions on which mathematicians and physicists lost chances
of making discoveries by neglecting to talk to each other. My
purpose in calling attention to such incidents is not to blame the I have learned from
mathematicians or to excuse the physicists for our failure in the Hilbert and Minkowski
that one does not
last twenty years to equal the great achievements of the past. My
influence people by
purpose is not to lament the past but to mould the future. It is ob- talking in
viously absurd for me to imagine that I can mould the future with generalities....the
a one-hour lecture. The fact that Hilbert in 1900 and Minkowski progress of both
mathematics and physics
in 1908 succeeded in doing it does not give me any confidence
has in the past been
that I can do it too. But at least I have learned from Hilbert and seriously retarded by our
Minkowski that one does not influence people by talking in gener- unwillingness to listen to
alities. Hilbert and Minkowski gave specific suggestions of things one another.”
that mathematicians and physicists could profitably think about; – F Dyson
I shall try to follow their style. I shall try to convince you by ex-
amining actual cases that the progress of both mathematics and
physics has in the past been seriously retarded by our unwilling-
ness to listen to one another. And I will end with an attempt to
identify some areas in which opportunities for future discoveries
are now being missed.” We will describe one of the examples

RESONANCE | October 2020 1345

 GENERAL ARTICLE

Dyson gives to demonstrate the kind of things these ‘missed op-

portunities’ entail.
Quoting him, “Long after I became a physicist, I retained a senti-
mental attachment to the (Ramanujan) τ-function, and as a relief
from the serious business of physics I would from time to time go
back to Ramanujan’s papers and meditate on the many intrigu-
ing problems that he left unsolved. Four years ago, during one
of these holidays from physics, I found a new formula for the τ-
function, so elegant that it is rather surprising that Ramanujan did
not think of it himself. The formula is:”
X Q1≤i< j≤5 (ai − a j )
τ(n) =
1!2!3!4!
where the sum is over all ai ≡ i mod 5 with

a1 + a2 + a3 + a4 + a5 = 0;

a21 + a22 + a23 + a24 + a25 = 10n.

As the Dedekind eta function η(x) = x1/24 ∞ n
Q
n=1 (1 − x ) satisfies
X
η(x)24 = τ(n)xn
n≥1

the above expression of Dyson for τ gives a formula for the 24-
th power of the eta function. Dyson discovered that there were
analogous formulae for η(x)d for

d = 3, 8, 10, 14, 15, 21, 24, 26, 28, 35, 36, · · ·

Then Dyson goes on to say:

“I stared for a little while at this queer list of numbers (3). As I
was, for the time being, a number theorist, they made no sense
to me. My mind was so well compartmentalized that I did not
remember that I had met these same numbers many times in my
life as a physicist. If the numbers had appeared in the context of a
problem in physics, I would certainly have recognized them as the
dimensions of the finite-dimensional simple Lie algebras. Except
for 26. Why 26 is there I still do not know. The others all cor-
respond to simple algebras: Al , A2 , B2 , G2 , A3 , B3 , A4 , D4 , A5 , B4

1346 RESONANCE | October 2020

 GENERAL ARTICLE

and so on. For example, d = 24 (above) corresponds to the al-

gebra A4 and in the structure of the formula (above) you can see
the root system of A4 . So I missed the opportunity of discover-
ing a deeper connection between modular forms and Lie algebras,
just because the number theorist Dyson and the physicist Dyson
were not speaking to each other. This story has a happy ending.
Unknown to me the English geometer, Ian MacDonald, had dis-
covered these same formulae as special cases of a much more
general theory. In his theory, the Lie algebras were incorporated
from the beginning, and it was the connection with modular forms
which came as a surprise. Anyhow, MacDonald established the
connection and so picked up the opportunity which I missed.”

6. Tidbits

6.1 Counterfeit Coin

The famous problem asks for identification of one ‘defective’

coin among (3n −3)/2 coins of which all others have equal weights
if we are allowed n weighings. The weighing is done on a two-
sided pan without any weights provided, and we may place coins
on either side. We are asked also to determine if the defective
coin is heavier or lighter; if we are asked only to determine which
coin is defective among (3n − 1)/2 coins, it turns out we can do so
in n weighings. The solutions are usually given by a case-by-case
branching method where the weighing at the ith stage depends on
the outcomes of the weighing in the previous step. Dyson gave a Dyson gave a delightful
delightful new solution where a pre-determined algorithm right at new solution where a
the beginning could be given which, when followed, determines pre-determined
algorithm right at the
the bad coin and also whether it is heavier or lighter! His proof beginning could be given
also shows that n weighings will not suffice if we have more than which, when followed,
(3n − 3)/2 coins. The algorithm exists for any number of coins determines the bad coin
not exceeding (3n − 3)/2 but the extreme case of (3n − 3)/2 coins and also whether it is
heavier or lighter!
is the most interesting. As the solution is easy to describe, we do
so here.
The algorithm only requires labelling the (3n − 3)/2 coins in the
base 3 as follows.

RESONANCE | October 2020 1347

 GENERAL ARTICLE

n
Write N = 3 2−3 for simplicity of notation and we will give two
labels for each coin as follows.
The N coins are numbered 1, 2 · · · , N and the first label of coin
i is its base 3 expansion in n-digits (if the number of digits is
less than n, one puts 0’s, in the beginning, to make it genuinely n
digits). The second label of the coin i is obtained by subtracting
the first label from the number 3n − 1 in base 3 (which consists of
n two’s). Thus, whichever digit in the first label is 0 becomes 2
in the second label and vice versa; the digits 1 remain the same.
In this manner, each coin gets two labels and any n-digit base 3
number (other than the same digit repeated n times) can occur
just once as a label. Among the two labels of any coin, one calls
‘clockwise’, if the first change of digit starting from the left-most
digit takes either 0 to 1 or 1 to 2 or 2 to 0. Otherwise, it is called
anticlockwise. Thus, for any coin, one label is clockwise and the
other is anticlockwise. For each i ≤ n, we divide the set of coins
into 3 sets C(i, 0), C(i, 1), C(i, 2) where C(i, d) is made up of the
set of all those coins whose clockwise label has digit d int he
i-th place form the left. Note that since the cyclic permutation
0 7→ 1, 1 7→ 2, 2 7→ 0 maps C(i, 0) to C(i, 1), C(i, 1) to C(i, 2) and
C(i, 2) to C(i, 0), these sets are equinumerous; that is,
|C(i, d)| = N/3 ∀ i ≤ n, d = 0, 1, 2.
The algorithm is described as follows.
At the i-th weighing, the coins for C(i, 0) and C(i, 2) are placed
on the left side and the right side of the pan respectively. The set
C(i, 1) is kept aside.
We define the number ai to be 0, 1 or 2 according as to whether
the left side is heavier, both are equal or the right is heavier in the
i-th weighing.
Consider the number A = 3n−1 a1 + 3n−2 a2 + · · · + 3an−1 + an .
From the i-th weighing, it is clear then that the defective coin is
heavier and its clockwise label has ai as its i-th digit, or lighter
and its anticlockwise label has ai as its i-th digit.
Therefore, after n weighings, we know all the ai ’s and the coin

1348 RESONANCE | October 2020

 GENERAL ARTICLE

one of whose labels is the number A is the defective coin. Also, it

is heavier or lighter according to whether this label is clockwise
or anticlockwise!

6.2 Stable Table

In 1951, Dyson came up with a very clever argument to prove that In 1951, Dyson came up
corresponding to any continuous function f defined from the 2- with a very clever
dimensional unit sphere S 2 to the real numbers, one can find two argument to prove that
corresponding to any
orthogonal diameters so that f takes the same value at all the four continuous function f
endpoints (which, of course, form the vertices of a square). The defined from the
result turns out actually to be equivalent to the general case when 2-dimensional unit
sphere S 2 to the real
orthogonality of diameters can be replaced with the condition that
numbers, one can find
they meet at any given angle. More precisely, Dyson’s theorem is two orthogonal
equivalent to: diameters so that f takes
For ant continuous function f : S 2 → R, and any fixed positive the same value at all the
four endpoints.
real number r with 0 < r < 2, there are points x, y ∈ S 2 at
distance r so that f takes the same value at x, −x, y, −y.
In 1955, C. Yang generalized Dyson’s theorem to general dimen-
sions; he showed:
For any positive integers n, d and any continuous function from
S dn to Rd , there exist n mutually orthogonal diameters whose 2n
endpoints are mapped to the same point.
The special case n = 1 is the well known Borsuk–Ulam theorem.
The theorem of Dyson on four points on a sphere is directly re-
lated to the ‘stable table’ problem.
A rectangular table placed on a slightly inclined floor (but one
which is continuous) that wobbles as not all legs are resting, can
be brought to a stable position simply by a rotation.
This is a form of the intermediate value theorem but more pre-
cisely follows from a 1954 theorem of Livesay that (generalizes
Dyson’s theorem from the square case and) asserts: For any con-
tinuous function f defined on the unit sphere, we can position a
given rectangular table with all its vertices on the sphere such
that f takes on the same value at all four vertices.

RESONANCE | October 2020 1349

 GENERAL ARTICLE

The above theorem of Dyson on spheres is not related to the con-

cept of ‘Dyson sphere’ which is a thought experiment by him
popularized in 1960. His idea was that if the energy demands
of human civilization increased for a long enough period, a time
would come when it would require all the energy produced by
the Sun. He thought of a system of orbiting “shells” that would
intercept and collect all of the Sun’s energy.

6.3 Algebraic Topology

Dyson’s work on number Dyson’s work on number theory often led him to discover as
theory often led him to byproducts results of totally different flavours. One such discov-
discover as byproducts
ery was a theorem in algebraic topology of independent interest
results of totally different
flavours. One such that stems from his work on the product of four inhomogeneous
discovery was a theorem forms. This topological result, which addresses simplicial homol-
in algebraic topology of ogy groups is too technical to recall here in its full generality, but
independent interest that
it is a generalization of an easier to state theorem due to Lebesgue.
stems from his work on
the product of four Lebesgue’s result asserts that if an open subset of Rn is covered
inhomogeneous forms. by a finite number of closed sets of sufficiently small diameter,
then some n + 1 of them must have a common point.

6.4 Sixth Fermat Number

One knows that the largest prime known of the form 2k + 1 is

5
216 + 1. The fifth Fermat number 22 + 1 was shown in an elemen-
tary manner to be divisible by 641 by Euler. A similar elemen-
6
tary proof was given by Dyson to prove that 22 + 1 is divisible
by 274177, and therefore, cannot be prime. Basically, Dyson’s
observation amounts to the following:
If f is odd and 2n ≡ 1 mod f , and if the number q = 1 + 2r f
 2
n
divides 1 + 22m 2 f−1 − 2r , the q also divides 1 + 4m+n+r .
This enables him to not only obtain Euler’s result by checking
that q = 641 = 1 + 27 × 5 = 24 + 54 divides 1 + 210 56 (corresponds
to m = 5, n = 4 in Dyson’s observation) but also shows that since
q = 274177 = 1+ 28 × 32 × 7× 17 divides 1+ 154092 (corresponds
6
to m = 0, n = 24), it divides also 1 + 432 = 1 + 22 .

1350 RESONANCE | October 2020

 GENERAL ARTICLE

6.5 Dyson Transform for Ramare

In 1995, Olivier Ramare proved that every even integer is a sum In 1995, Olivier Ramare
of at the most 6 prime numbers. A crucial lemma that went into proved that every even
integer is a sum of at the
his proof was due to Dyson, and is now known as the Dyson trans-
most 6 prime numbers.
form; it permits us to ‘transform’ elements from one summand to A crucial lemma that
another. More precisely, this is the following Lemma: went into his proof was
Let A = {a1 < a2 < · · · }, B = {0 = b1 < b2 < · · · } be sequences of due to Dyson, and is now
known as the Dyson
non-negative integers. For any e ∈ A, define
transform.
A′ = A ∪ (B + e), B′ = B ∩ (A − e).

Then, we have A′ + B′ ⊂ A + B, e + B′ ⊂ A′ , 0 ∈ B′ and, for each

m,

(A ∩ [1, m]) + (B ∩ [1, m − e]) = (A′ ∩ [1, m]) + (B′ ∩ [1, m − e]).

Suggested Reading
[1] F J Dyson, Some guesses in the theory of partitions, Eureka, (Cambridge) Vol.8,
pp.10–15, 1944.
[2] F J Dyson, Mappings and symmetries of partitions, J. Combin. Theory Ser. A,
Vol.51, pp.169–180, 1989.
[3] F J Dyson, Birds and Frogs, Notices of the Amer. Math. Soc., Vol.56, No.2,
pp.212–223, 2009.
[4] F J Dyson, Statistical theory of the energy levels of complex systems – I, Journal Address for Correspondence
of Mathematical Physics, Vol.3, pp.140–156, 1962. B Sury
[5] F J Dyson, Missed opportunities, Bull. Amer. Math. Soc., Vol.78, pp.635–652, Stat-Math Unit
1972. Indian Statistical Institute
[6] F G Garvan, New combinatorial interpretations of Ramanujan’s partition con- 8th Mile Mysore Road
gruences mod 5, 7 and 11, Trans. Amer. Math. Soc., Vol.305, pp.47–77, 1988. Bengaluru 560 059 India.
[7] F G Garvan, The crank of partitions mod 8, 9 and 10, Trans. Amer. Math. Soc., Email: [email protected]
Vol.322, pp.79–94, 1990.

RESONANCE | October 2020 1351