This suggestion, whether right-headed or wrong-headed, certainly reflects van Inwagen's proclivity for metaphysical/epistemological pessimism (skepticism). For instance, in The Problem of Evil: The Gifford Lectures Delivered in the University of St Andrews in 2003 (Oxford: Clarendon Press, 2006), we read:
Philosophical arguments are not best thought of as
free-floating bits of text--as mathematical proofs can
perhaps be thought of. A proper mathematical proof,
whatever else it may be, is an argument that should
convince anyone who can follow it of the truth of its
conclusion. We cannot think of philosophical arguments
as being like that (p. 37).
We may conclude from van Inwagen's analysis that "it seems reasonable to believe that no non-theological philosophical argument for a substantive conclusion is a [philosophical] success" (ibid, p. 54).
Of course, this claim must be understood within its proper context. van Inwagen wants to persuade us that the argument (or set of arguments) for the logical problem of evil is a philosophical failure like all other non-theological contentions since "the problem" evidently does not have the power to bring it about that philosophers of all stripes agree with its substantive conclusions regarding ultimate reality. In order to fully appreciate the context for his statements, one must read the chapter in The Problem of Evil titled "Philosophical Failure." It is difficult to summarize the chapter's contents and general thesis in a short blog post.
Finally, Saul Kripke also makes this claim in Naming and Necessity (Cambridge, MA: Harvard University Press, 1980 ), p. 64:
Let me state then what the cluster concept theory of
names is. (It really is a nice theory. The only defect
I think it has is probably common to all philosophical
theories. It's wrong. You may suspect me of proposing
another theory in its place; but I hope not, because
I'm sure it's wrong too if it is a theory).