The Mason-Stothers Theorem
Jeffrey Paul Wheeler, the University of Pittsburgh∗
A child learns of the nonnegative numbers at an early age. Polynomials,
on the other hand, demand a little more sophistication and are reserved in
a U.S. child’s education for middle school. Those fortunate enough to take
an undergraduate abstract algebra class realize that the chasm between the
integers and polynomials is not so vast. One learns there are similarities:
both integers and polynomials form rings; and that there are analogies: the
integers have prime factors as their basic building blocks, whereas polynomi-
als (over C) have linear factors. Given this, it is not that surprising that we
have the following definitions:
Definition 1 (The Radical of an Integer).
For n ∈ Z+ suppose n = p1 e1 · · · pk ek where the pi ’s are distinct primes and
the ei ’s are positive integers. We the define the radical of n to be:
r(n) = p1 · · · pk with r(1) := 1.
In other words, r(n) is the greatest square-free factor of n or, more simply, the
product of distinct prime factors of n. As an example, r(100) = r(22 · 52 ) =
2 · 5 = 10.
Definition 2 (The Radical of a Polynomial).
Let p(t) be a polynomial whose coefficients belong to an algebraically closed
field of characteristic 0. We put n0 (p(t)) = the number of distinct zeros of
p(t).
(In a ring R, if there exists a n ∈ Z+ such that na = 0 for all a ∈ R,
then the least such positive integer is called the characteristic of the ring.
∗
Department of Mathematics, the University of Pittsburgh, Pittsburgh PA 15260, USA.
E-mail:
[email protected].
1
Algebraically closed just means we are in the right place for all the roots of
the polynomial to exist; think of polynomials with coefficients in C - they
can be written as a product of linear factors if we allow roots from C).
With these definitions we state
Theorem 1 (Mason-Stothers Theorem).
Let a(t), b(t), and c(t) be polynomials whose coefficients belong to an alge-
braically closed field of characteristic 0. Suppose a(t), b(t), and c(t) are rela-
tively prime and that a(t) + b(t) = c(t). Then
max deg{a(t), b(t), c(t)} ≤ n0 (a(t) · b(t) · c(t)) − 1.
This beautiful theorem is easily understood and it’s proof requires just a
little knowledge of abstract algebra. What is intriguing is that the analogous
statement for the integers is still unproven1 . This is the important open
problem in Number Theory known as the abc Conjecture 2 :
Conjecture 1 (The abc Conjecture - Masser’s Version).
Consider a nontrivial triple of integers (a, b, c) such that a + b = c and
gcd(a, b, c) = 1. Then for every > 0 there exists a universal constant
µ() such that
max{|a|, |b|, |c|} ≤ µ()[r(abc)]1+ .
Hence, the remarkable fact that we have a problem for polynomials that is
much easier to establish than the analogous statement for integers. This is
not the only one. We offer
Problem: Use the Mason-Strother’s Theorem to establish a polynomial Fer-
mat’s Last Theorem; that is
Corollary (Polynomial Fermat’s Last Theorem).
Let x(t), y(t), and z(t) be relatively prime polynomials whose coefficients be-
long to an algebraically closed field of characteristic 0 such that at least one
1
We note that Shinichi Mochizuki of Kyoto University has offered a proof in the form
of a series of papers on his website [3]. A review is in process.
2
The conjecture was originally posed in 1985 by David Masser (considering an integer
analog of Mason’s Theorem) [2] and in 1988 by Joseph Oesterlé (considering a conjecture
of Szpiro regarding elliptic curves) [4].
2
of them has degree > 0. Then
x(t)n + y(t)n = z(t)n
has no solution for n ≥ 3.
References
[1] Mason, R. C. (1984), Diophantine Equations over Function Fields, Lon-
don Mathematical Society Lecture Note Series 96, Cambridge, England:
Cambridge University Press.
[2] Masser, D. W. (1985). Open problems. In Chen, W. W. L. Proceedings of
the Symposium on Analytic Number Theory. London: Imperial College
[3] Mochizuki, Shinichi. Personal webpage.
http://www.kurims.kyoto-u.ac.jp/~motizuki/top-english.html
[4] Oesterlé, Joseph (1988), Nouvelles approches du “thorme” de Fermat,
Astrisque, Sminaire Bourbaki exp 694 (161): 165186
[5] Stothers, W. W. (1981), Polynomial identities and hauptmoduln, Quar-
terly J. Math. Oxford, 2 32: 349370.
