I his reply to my previous comment prof. Busemeyer remarked:
"Beginning with Thurstone 1927, and later signal detection theory 1954, and random utility models (McFadden 1970’s), most psychologists and economists think that evidence or preference is approximately continuously distributed, and a binary choice is a partition on this continuum (e.g., probability of choosing A over B is the probability that a continuously distributed random utility A exceeds B). In fact, a long history of experimental evidence supports the continuous (or approximately continuous) rather than discrete (small number) state models for binary choices (like signal detection – choose signal vs. noise)."
This is very important remark and it seems that in fact the real mental state space is INFINITE DIMENSIONAL! Would you agree?