1

Taken from pg-53 of Roger Penrose's road to reality,

Suppose $\sqrt{2}$ is rational, then:

$$ \sqrt{2} = \frac{p}{q}$$

Where $p$ and $q$ are some integers with $ q \neq 0$:

Squaring and rearrangnig:

$$ 2 q^2 = p^2$$

This means $p$ is even, and hence, $p=2p_1$, meaning that:

$$ q^2 = 2p_1^2$$

This means that $q$ is even, we can put $q=2q_1$, where $q_1$ is some integer, meaning that:

$$ 2q_1^2 = p_1^2$$

See that this looks same as the equation we begun with $2q^2 =p^2$, repeating the procedure we done it again we can further find that $p_1$ and $q_1$ is again even writing $p_1 = 2p_2$ and $q_1=2q_2$, after simplifications we will find again:

$$ 2q_2^2 = p_2^2$$

Repeating the procedure many times, we could repeat this as many times as we want , we get a recurrence for $q_i$s as:

$$q_i = 2q_{i+1}$$

Or, $$\frac{q_i}{2} = q_{i+1}$$

Obeying:

$$q>q_1>q_2...$$

Penrose writes:

"..All of these integers being positive. But any decreasing sequence of positive integers must come to an end, contradicting the fact that this sequence is unending. This provides us with a contradiction to what has been supposed , namely that there is a rational number which squares to $2$"

Firstly where was it assumed that the sequence is 'unending' and what exactly does it mean for a sequence to end? I thought sequence could go on forever..

P.S: I understand the square root of two's irrationality proof already, the point is not to prove the statement in itself but understand this particular proof.

Bill Dubuque
  • 272,048

1 Answers1

2

Firstly where was it assumed that the sequence is 'unending'

Here: $$q>q_1>q_2>\ldots$$

He means that it is an infinite strictly decreasing sequence: each later term less than the previous one. This infinite sequence is constructed inductively. He uses the ellipsis notation to suggest this.

what exactly does it mean for a sequence to end? I thought sequence could go on forever..

Yes, a sequence can go on forever. But an infinite sequence of positive integers cannot strictly decrease forever. That is what the well-ordering principle says.

Compare this to the case of rational numbers, where you can have $1/2 > 1/4 > 1/8 > \ldots > 1/2^n > \ldots $ forever.

rschwieb
  • 153,510