> > Do any nat or conlangs express this type of ideas by means of a
> > specific conjugation of verbs such as "guess", "believe",
> > "deny", "affirms", etc., instead of using "might",
> > "probably", "not", "yes", etc.?
> I am sure the answer to this would be "yes"! ;)
> I think what you're asking about here is evidentiality:
> Padraic
> > Leonardo

As for polarity ("yes"/"no"), indicating negation with a bound affix on the
verb is extremely common in languages around the world. I guess the goto
reference here is Turkish.

At least Yukaghir extends this system by having a special affirmative affix
that has similar morphology than the negative affix and that contrasts with
the unmarked "non-negative" verb forms. While the unmarked forms simply
indicate non-contrastive declarations, or inferences if evidential
morphology is present, the marked affirmative forms place contrastive focus
on the verb or the whole clause. So in a sentence like

"I got a lot to do for today, but managed to finish it all."

you'd likely leave "got" in the first clause unmarked for polarity but mark
"managed" in the second clause with the affirmative affix to highlight the
contrast between the information contents of the two clauses.