What is a sensible definition of the term language to use when interpreting private language arguments in general (if possible), or Wittgenstein in particular?
Conversely, is rejecting the idea of private languages about the definition of language?
Depending on the definition of language, e.g. if it includes an aspect of being embodied amongst a community, then the absence of private languages might be purely definitional.
Conversely, if one takes a very loose definition of the term then I can conceive of things that at least "skirt the edges" of a private language.
To set some background, suppose there is a written language shared only by Alice and Bob, e.g. they made up their own pairwise language. I'm assuming that there is no conceptual problem with such a just pairwise language under (basically) any definition of the term.
Suppose that Bob writes notes to himself, in a conventional language, in order supplement his memory (I do not think that he needs to be mitigating severe amnesia in order for this to make sense, though assuming such doesn't seem a problem). I'm assuming that there no problem with considering this as language use under any definition of the term.
Now, suppose that Bob makes up a custom written language for keeping his notes. It seems to me that given the previous point, this case shares a "family resemblance" to that example of language use, and though not a public language in practice, it is (typically) a public language in principle.
Some experiences are characterized as being un-communicatable, and thus intrinsically private (one might say they are ineffable in public languages), and yet are often characterized as being personally meaningful.
Finally, if Bob is able to make a set of signs, that helps him (or causes him) to remember/reconstruct these kinds of ineffable experiences, then couldn't it be a private in that it is not in principle translatable due to the fact that his experiences are his alone?
It seems to me that at important link in this chain is the rather loose use of the term language in step 5, and thus, can be easily severed with a suitable definition of the term "language"; but if you define language too strictly, the private language argument is just an aspect of the definition of the term. Thus, it seems to me, that being careful about the definition of the term "language" at the outset is critical to understanding private language arguments.
This question is closely related to this one. Indeed, under very broad usage of the term "language" mystic experiences themselves, being meaningful and significant yet intrinsically private, might be considered as involving a language, in the broad sense.