In French, c’est à savoir originally signified (1) 'that is to know', then (2) 'that is to say'.
The English verb 'wit' underwent the same semantic shift; see Etymonline:
The phrase to wit, almost the only surviving use of the verb, is first recorded 1570s, from earlier that is to wit (mid-14c.), probably a loan-translation of Anglo-French cestasavoir, used to render Latin videlicet (see viz.).
So what semantic notions underlie (1) and (2)?
Which category is this semantic shift? I'm guessing Metonymy?