5.156 Hence probability is a generalization.

It involves a general description of a propositional form.

Only in the absence of certainty do we need probability. – If we are indeed not completely acquainted with a fact but do know something about its form.

(A proposition can indeed be an incomplete picture of a certain state of things, but it is always a complete picture [i.e. a complete picture of something, or a complete picture in some sense].)

A proposition of probability is, as it were, an extract from other propositions.

Why talk about needing probability? And how can a proposition be an incomplete picture of something other than that of which it is a complete picture? And why “as it were”? Otherwise this seems straightforward.

Black’s conclusion (p. 256): “Wittgenstein’s definition [of probability] may tell us how probability might be computed in a language that does not exist; applied to any language actually available, it would compel us to treat all probability measures as unknown.”