§143

Wittgenstein tells us not to balk at the expression 'series of numbers' in his example. The point, I assume, is that it isn't incorrect to say 'numbers' rather than 'numerals' here. In fact it might be more misleading. Those defending the idea that 'numeral' is correct are in thrall to something like the Augustinian picture (thinking that the number is an object referred to by the numeral).

In his example B has to write down a series of signs according to a formation rule when A gives an order. The series of signs is the series of natural numbers - 1, 2, 3, 4, 5, 6, ...

B might come to learn the series by copying it up to 9. B would then need to be able to write them down himself in order for instruction to go on. When would we say B has understood? - This isn't completely clear. Wittgenstein makes the point  that there isn't a sharp distinction between a random and a systematic mistake.

Conceptual abilities rest on training. Training can only take place given standard natural reactions of trainees.