§48

Wittgenstein introduces a new language game in §48. He refers to the 'method of 2' here. I assume he is talking about using a language game (i.e. a 'primitive langauge') to highlight something about language/meaning.

In this language/language game words ('R', 'G', 'W', 'B') correspond to elements - coloured squares. I assume that he is constructing a language like this because he wants to come up with a language that approximates to the account of language given in the Tractatus - with names corresponding to (simple) elements. - This is a bit like the language constructed in §2, which was supposed to 'fit' with the Augustinian picture.

It seems natural to say, in this case, that a coloured square is a simple element BUT you could say that a square consists of two rectangles or that it consists of a colour and a shape as elements. It could be that something smaller was composed of something greater (e.g. that a square was composed of two squares minus one square). - This might sound strange but there are cases where we talk in this way (e.g. about the 'composition' of forces). - What is 'simple' is relative to the kind of thing that we are talking about and our purposes in talking about that kind of thing.

It doesn't matter how we think about how things are made up of other things as long as we avoid misunderstanding.

(What applies to the language game of §48 presumably applies to a greater extent to our own language).