A brief thought.

The argument over private language and the picture theory of ideas relate. Take for instance Locke's concern with "determinate ideas" in the context of the question whether it is possible to express (communicate) the entirety of one's thought about something--not itself purely linguistic--with one's language. It is certainly the case that we do not put all of a thought into the utterances we use to communicate, at least normally. Yet we get by. However this would seem to imply that it is possible that there are ideas (Locke's sense) that are never expressed by words. That is, that there might be private concepts--you could find one of these and make a "private language" by naming that feature of your complex idea of .... or ____. On this conception you'd have some kind of certainty that your neologism applies to (all and) only instances of the idea--in other words it is used to express the presence of this idea, call it P, in one's experience--or, better, in objects/structures in the world. That certainty of application, however, is a bedrock assumption of post-Cartesian philosophy; the assumption criticized in the Philosophical Investigations. (If Kripke can be believed, some want to say.) That assumption is tied up with arguments to the effect that there is something "it is like" to experience ____ distinct from the neural activity in the brain. But this is just the problem, as I see it: how can we plausibly talk about "what it is like" to have the idea of P when all of our arguments conclude that there must be such a thing, yet there is no satisfactory method of saying what P is like?

See also this:
A) I have been delivered a box in the post.
B) What was in it?
A) Something.
B) What was it like?
A) I can't really say. But:

1. No one sends a box without contents.
2. Someone sent me a box.
.: 3. This box has contents.

B) Makes sense.