Proofs

The sum of the first nn squares

Add up the square numbers:

12+22+32+⋯+n2=n(n+1)(2n+1)61^2 + 2^2 + 3^2 + \dots + n^2 = \frac{n(n + 1)(2n + 1)}{6}

Check it with n=4n = 4. The left side is 1+4+9+16=301 + 4 + 9 + 16 = 30, and the right side is 4×5×96=30\dfrac{4 \times 5 \times 9}{6} = 30. It works.

This page proves it for every nn at once. The method is proof by induction, which is how a statement about all the counting numbers is usually proved.

The shape of an induction proof

Induction has two steps, and you need both:

  1. Show it works for the first case. Here, n=1n = 1.
  2. Show that if it works for one case, it must work for the next. That is: if the formula is true for some nn, then it is true for n+1n + 1.

Together those cover everything. Step 1 gets you to n=1n = 1. Step 2 then carries you from 1 to 2, from 2 to 3, from 3 to 4, and onward without end. It is worth seeing why one step without the other proves nothing: a formula that is true only for n=1n = 1 passes step 1 and fails step 2, and a formula that is always one step behind passes step 2 and fails step 1.

Step 1: the first case

We need

12=1×2×361^2 = \frac{1 \times 2 \times 3}{6}

The right side is 66=1\dfrac{6}{6} = 1. So the statement is true when n=1n = 1.

Step 2: the inductive step

Now suppose the formula is true for some particular nn. That is our assumption:

12+22+⋯+n2=n(n+1)(2n+1)61^2 + 2^2 + \dots + n^2 = \frac{n(n + 1)(2n + 1)}{6}

We have to show the formula holds for n+1n + 1, which means showing

12+22+⋯+n2+(n+1)2=(n+1)(n+2)(2n+3)61^2 + 2^2 + \dots + n^2 + (n + 1)^2 = \frac{(n + 1)(n + 2)(2n + 3)}{6}

Start with the left side and use the assumption to replace the long sum:

12+⋯+n2+(n+1)2=n(n+1)(2n+1)6+(n+1)21^2 + \dots + n^2 + (n + 1)^2 = \frac{n(n + 1)(2n + 1)}{6} + (n + 1)^2

Now take out the common factor (n+1)(n + 1) from both terms:

=(n+1)[n(2n+1)6+(n+1)]= (n + 1)\left[\frac{n(2n + 1)}{6} + (n + 1)\right]

Put the second term over 66 so the bracket is a single fraction:

=(n+1)[n(2n+1)+6(n+1)6]= (n + 1)\left[\frac{n(2n + 1) + 6(n + 1)}{6}\right]

Expand the top:

=(n+1)[2n2+n+6n+66]=(n+1)[2n2+7n+66]= (n + 1)\left[\frac{2n^2 + n + 6n + 6}{6}\right] = (n + 1)\left[\frac{2n^2 + 7n + 6}{6}\right]

The quadratic on top factorises, because 2n2+7n+6=(2n+3)(n+2)2n^2 + 7n + 6 = (2n + 3)(n + 2):

=(n+1)(n+2)(2n+3)6= \frac{(n + 1)(n + 2)(2n + 3)}{6}

That is exactly the formula with n+1n + 1 in place of nn, because 2(n+1)+1=2n+32(n + 1) + 1 = 2n + 3.

Both steps are done. The statement is true for n=1n = 1, and true for n+1n+1 whenever it is true for nn. So it is true for every counting number.

What each step is doing, and what it is not

The inductive step does not prove the formula for any particular number. It proves a link between one case and the next. On its own it would be satisfied by a formula that is wrong everywhere, as long as the wrongness matched up.

That is why step 1 is not a formality. It is the anchor: it is the one place the formula is checked against reality, and the link in step 2 carries that check onward through every number.

A common error

Watch for this one, because it is easy to do in an exam and it costs the mark.

When you write "suppose the statement is true for nn", you must then use that supposition. A proof that subtracts the two sides and gets zero has usually assumed the very thing it was asked to prove.

The safe structure is the one above: start from the left side, put the assumption in, and arrive at the right side of the n+1n+1 statement. The direction of the work is what makes it a proof rather than a check.

Check your understanding

  1. Use the formula to work out 12+22+⋯+1021^2 + 2^2 + \dots + 10^2.
  2. In step 1, why is it enough to check n=1n = 1 rather than n=0n = 0?
  3. Show that if n(n+1)(2n+1)6\dfrac{n(n+1)(2n+1)}{6} is a whole number, then so is (n+1)(n+2)(2n+3)6\dfrac{(n+1)(n+2)(2n+3)}{6}. What does that tell you about every value of this sum?
Answers
  1. 10×11×216=23106=385\dfrac{10 \times 11 \times 21}{6} = \dfrac{2310}{6} = 385.
  2. The statement is about 12+22+⋯+n21^2 + 2^2 + \dots + n^2, which begins at 11 and has no 020^2 term, so n=1n = 1 is its first case. Starting at 00 would prove something slightly different.
  3. From step 2 the two fractions are equal, so the second is a whole number exactly when the first is. Starting from n=1n = 1, which is a whole number, the step carries that along: the sum of the first nn squares is always a whole number.

Settings

Colour scheme

Applies to every book you read here.

Reading theme

Applies to every book you read here.