Pupils who are introduced to algebra-as-a-game can appreciate that they have a choice of moves, that some moves will be better than others, that they need to look ahead, that look-ahead depends on their past experience, and that practice is required, not the skill-practice which hones a skill to perfection, but game-practice which develops their understanding of the miniature game-world.
School algebra is wrecked as a creative subject by being taught – like elementary arithmetic – as a collection of algorithms which few pupils understand, instead of being a creative and fascinating exploration of patterns and relationships, smart moves and surprising results. The following example by that great master Euler shows him, at a very high level, playing the game of algebra quite brilliantly. Not surprisingly, his
pentagonal number theorem
also illustrates the roles of perception and scientific induction: all three aspects of mathematics really are difficult to separate in practice.
Euler starts by explaining that,
In considering the partitions of numbers, I examined, a long time ago, the expression
Euler then records that he actually multiplied out a large number of terms to get this series:
This is a remarkable result because the coefficients here are what are called the
generalised
pentagonal numbers. We take the pentagonal numbers [page 179] which are, 1, 5, 12, 22, 35, 51, 70…and their formula
n
(3
n
−1)/2 and we allow
n
to be negative as well as positive.
The formula then gives us for
n
= 0, ±1, ±2, ±3, ±4…the sequence, 0, 1, 2, 5, 7, 12, 15, 22, 26, 35, 40, 51, 57, 70…
So far, so scientific. Euler is confident that the coefficients must be what they seem to be:
For each of us can convince himself of this truth by performing the multiplication as far as he may wish; and it seems impossible that the law which has been discovered to hold for 20 terms, for example, would not be observed in the terms that follow.
However, so far he has only
induced
his conclusion scientifically: a proof is nowhere to be seen – but Euler perseveres, with his usual imagination and flair: he ‘manoeuvred those two expressions in many ways’, he writes, his first move being to ‘[get] rid of the factors in the first by taking logarithms’.
Calling the product
s
, he concludes that,
Next, ‘In order to get rid of the logarithms, I differentiate and obtain the equation’:
Euler, unlike most mathematicians then or now, is fully explaining his actions. His next move is to calculate the same expression (−
x
/
s
)d
s
/d
x
from the second equation, the series with the generalised pentagonal number coefficients, (so that he now has two ways of writing the same function), and then to go back to the expression above and expand every one of the expressions
x
/(1−
x
), 2
x
2
/(1−
x
2
)…as a geometric series. He then makes a remarkable observation which brings in the
divisors
of the exponents:
Here we easily see that each power of x arises as many times as its exponent has divisors, and that each divisor arises as a coefficient of the same power of
x
.
and so on…
In this table,
x
6
(for example) appears four times, with the coefficients 1, 2, 3 and 6, the divisors of 6.
It follows that, ‘if we collect the terms with like powers, the coefficient of each power of x will be the sum of the divisors of the exponent’.
Euler concludes his memoir, (we are jumping forward at this point) by
proving
a remarkable formula for the sum of divisors of an integer which he had previously inducted by scientific observation:
…in which the series continues until an argument is negative, with the proviso – itself remarkable – that if
r
(0) appears at the end of the series, then it is counted as
n
.
Euler, by his original scientific observation followed by a sequence of very powerful game-like moves has found this stunning connection between the generalised pentagonal series and the sums of divisors of the integers.
His simple but powerful
manoeuvres
– as he puts it – were revealed to him by his analysis of his original position, just as any chess player analyses a position, seeking the best sequence of moves. (Unfortunately we do not know what his
unsuccessful
manoeuvres were.)
Euler comments that ‘although we consider here the nature of integers to which the Infinitesimal Calculus does not seem to apply, nevertheless I reached my conclusion by differentiations and other devices.’ The mathematician has the advantage that he or she can switch games or develop the rules of the game, up to a point, in a way that the chess player cannot.
His brilliant game-like combinations, producing the relationship he needs in just a handful of simple powerful moves is worthy of the great Capablanca,
or of Fisher or Kasparov. Without it, or some equivalent game-like sequence, he would not have progressed beyond his scientific pattern spotting.
George Polya claimed that Euler's beautiful memoir on his pentagonal number theorem ‘is entirely devoted to the exposition of an inductive argument’. This is misleading: Polya was prejudiced by his devotion to induction. On the contrary, it is just as much an example of game-like brilliance [Polya
1954
: 96–98] [Wells
1991
b].
Is it a coincidence that Euler here combines his dazzling ability for pattern-spotting with his equal ability to play the game of mathematics brilliantly? No, rather the opposite, the two aspects go hand in hand. The theory of numbers, even more than other branches of mathematics, has been developed over centuries to a highly game-like state, in which it is both extremely easy – because it involves the integers – to generate data for scientific inductions
and
to perform formal manipulations. It is natural, therefore, that amateurs should always have found it a rich field for their modest efforts while it has also exercised the genius of the greatest masters of induction in mathematics, such as Euler, Gauss and Ramanujan, who were also among the greatest formal manipulators.