PI §128 is a tricky one; it keeps going in and out of focus for me. My general advice is to key on the sections (i.e. in the neighborhood of §§109-133) that you understand and read the others in that context (I guess that's obvious, isn't it?). For example, look at 128's immediate neighbors:
127: The work of the philosopher consists in assembling reminders for a particular purpose.
129: The aspects [Aspekte] of things that are most important for us are hidden because of their simplicity and familiarity. (One is unable to notice something—because it is always before one's eyes.) [...]: we fail to be struck by what, once seen, is most striking and powerful.
So when the philosopher says "[P]", this is to be taken as a reminder of something we already know, not a thesis to be proved in the face of ignorance or doubt. We already know it, so an argument for it would be out of place, as this would make it look like its truth, rather than its significance in the context, is what we are unclear on. We would say "Well, duh! But so what?", asking for a further, substantive claim instead of looking for the "hidden" aspect of [P] which was the real philosophical point. Compare §126, in which he says that "since everything lies open to view, there is nothing [for philosophy] to explain. For what is hidden, for example, is of no interest to us." There are no hidden philosophical truths; instead, philosophy deals with the "hidden" significance (to the one suffering from philosophical confusion) of truths which are not hidden at all, and thus do not need "philosophers" to unearth them, but instead lie open to view. What is "hidden" here is hidden in plain sight.
All this is pretty straightforward itself, as an interpretation. But what does it mean? Remember the joke about the plumber: when he bangs the pipe with his wrench and the water flows again, he bills you $505 – $5 for banging, and $500 for knowing where to bang. Here too, one can hardly be effective as a philosopher by pointing out obvious facts at random. It is for a particular purpose that these "reminders" are "assembled." Our next question is: what is that purpose? We already know the answer – it is to dispel philosophical illusion, of the sort that makes it seem as if a philosophical doctrine is required to unravel an apparent mystery.
But let's leave that for later. Right now I'll close with a few paragraphs from a recent collection on the topic, Wittgenstein at Work: Method in the Philosophical Investigations, edited by E. Ammereller and E. Fischer (Routledge, 2004). This is from Anthony Kenny's contribution, "'Philosophy only states what everyone admits'" (the quotation is from PI §599). Here (pp. 175-6) Kenny is asking the natural question – to what extent does Wittgenstein's own procedure in PI match fit his description of philosophy, in particular his rejection, in these sections, of the advancing of "theses"?
A reason why one might think that there is room for argument even in Wittgensteinian therapy is that the philosophical treatment of a problem may well involve the use of words like 'so,' 'therefore,' and 'because' which are characteristic of genuine inference. But that is due to the demands of the therapeutic procedure. The misguided philosopher believes that his dogma is a genuine proposition. To cure him of that illusion we have to humour him: we have to take his pseudo-proposition seriously by treating it as if it was a genuine proposition and drawing consequences from it. Of course these consequences will themselves be pseudo-propositions and only pseudo-consequences.
The purpose of this operation is not to lead the patient to a conclusion that he will recognize as false, so that he will have to give up his premise. It is, rather, to bring him to realize the illusory nature of his original claim, and thus make him cease to wish to persevere with it. It is literally a reductio ad absurdum, not the reduction to self-contradiction that goes by that name in logic textbooks.
The therapeutic procedure is not, however, a mere incantation. It must obey the laws of logic. What 'follows from' the pseudo-proposition must be what would really follow from it if it were a genuine proposition. To the non-Wittgensteinian philosopher – and in particular to the philosopher whose intellectual malaise is being treated – it does, indeed, appear to be an argument. A commentator, therefore, who uses expressions such as 'the private language argument' need not necessarily be in error about Wittgenstein's therapeutic conception of philosophy.
However, Kenny ends the article by telling us that he does "not believe that it is, in the end, possible to reconcile Wittgenstein's account of philosophy with the entirety of his philosophical activity in the Investigations" (p. 181). I don't either – I think he's taking a hard line in sections like 128 in order to keep us from misunderstanding his intentions (all too easy to do even so), where his, and our, actual procedure is more nuanced (versatile, pluralistic, continuous with the philosophical tradition, etc.). Ultimately, it is we, his patients, who must decide what we will do – use philosophical arguments, or abjure their use, as needed – which is how it should be, even, as I will argue, on Wittgensteinian grounds.