Descartes thinks to himself “I’ve done plenty of math in my time, and though most of the time I do it very well, I have often made mistakes, miscalculations. Sometimes I’ve even been sure, upon reviewing my work, that I had done it all correctly, only later, upon reviewing it anew, to find that I had made mistakes which were now open and obvious to me. Perhaps even, on the rare occasion, I returned to review the matter a third time and became convinced once again that I had been right the first time around. All of this makes me think that I can in principle be wrong about even seemingly self-evident mathematical propositions like that 2+3=5. Therefore, I cannot take any mathematical proposition, even though it is clearly analytic, as indubitable.” At the present time, 2+2=4 is self-evident to me. Perhaps I remember doing a mathematical calculation wrong in the past. Should this lead me to conclude that I can be wrong about mathematical truths, or truths of logic? No, perhaps not. Perhaps at best what I can be skeptical of is my memory of ever having miscalculated anything. The evil demon, however powerful, cannot efface from my mind the self-evidence of analytic propositions under my immediate comprehension. Descartes thinks we could logically possibly be mistaken about a mathematical proposition even if it were fully understood (not mis-understood). I think that kind of universal possibilism is incoherent and literally unthinkable.
This analysis may just push the skeptical question back one step, though. After all, how do I know right now that I understand with full comprehension that Pv~P self-evidently? The same answer can be given though: it may just be self-evident to me that it is self-evident. It may be rational to doubt whether I have ever misunderstood an apparently self-evident proposition (i.e., taken a proposition as self-evident without it being so), but not rational to doubt what is now presented to my mind as a clear and distinct self-evident truth. Descartes thinks that it is because we have memories of experiences in which we recall encountering what seemed to us a self-evident proposition which turned out to be false, that we must concede the possibility of being wrong about any proposition which appears to us to be self-evidently true, but this more easily licenses skepticism with respect to those memories and experiences, however recent, than it does skepticism with respect to logical, mathematical, and analytic propositions.