I wrote in a previous post (Does Truth Exist?) that we assume that truth exists in all our speech and thoughts, and that it would be impossible to truly speak or think otherwise: if I think, “truth does not exist,” I really mean, “truly, truth does not exist”—a logical contradiction. No functioning human—or sentient being for that matter—can reasonably live on the opposite assumption.
There is also this: it is impossible to believe that everything is an illusion. “Everything” here means all that you think you know. In that case, the thought “everything is an illusion” is an illusion, and that thought is an illusion, and so on ad infinitum. An infinite regress of illusions is impossible. A similar problem arises with truth. If everything is true, then the thought “everything is an illusion” is also true. And, as Democritus says of Protagoras’ “man is the measure of all things”, if everything is true, then among truths is the thought that not everything is true. So neither “everything is true” nor “everything is an illusion” are correct.
But it may be objected that saying “everything is an illusion” is not quite the same as saying “nothing is true”. However, presumably the thought “nothing is true” is meant to be true, therefore negating itself. Maybe it depends on what exactly is meant by “nothing is true”. If it is taken to mean that “nothing that I know, apart from this thought, is true” then that seems to be perfectly acceptable; but it is logically impossible to make a generalisation out of it, because that would mean that “it is true that nothing can be known which is true”—again a logical contradiction.
The only weakness I can see in the above is that a distinction needs to be drawn between two kinds of thought: that is, thoughts about the nature and truth of things, and thoughts about the nature and truth of truth. The former might be said to be an epistemic judgement, and the latter a metaepistemic judgement. But is a metaepistemic judgement internal or external? That is, does it concern doubt, or does it concern truth as a thing utterly independent of any observers? If this is a useful distinction to make, it would certainly merit further exploration.