Abstract

In the article we argue that Descartes while considering sign version of cogito does use some hidden premises in spite of well-known declaration there are none. They are:1) there’s such a length of time Δt over which I cannot be wrong on the matter of what happened over Δt ; 2) there’s such time extent Δt upon which I cannot be wrong. We investigate the advantages the evil demon can take from that. This is a sign manipulation we propose. Cogito as a sign object has to have a) a sense-data component and b) can be built only through time by consecutive generation one elementary sign after another. Then an sign manipulation is the following: I is [not am] building the sign si, I thinks wrongly that the previous signs just built are s1, …, si-1, I is [not am] going to built signs si+1, …, sn. For example, one thinks of self as thinking cogito, while in fact the real event is building “t” ). We investigate the source which makes the manipulation possible and the gravity of the destruction of such sort. The crucial reason for the possibility of such misconception turns out to be the sense-data component in any linguistic constructions.