For a many-sorted signature Sigma= (S,TF,PF,P) a many-sorted model M e Mod(Sigma) is a many-sorted first-order structure consisting of a many-sorted partial algebra:
[AT] So, we are back to models with possibly empty carriers. I don't recall the explicit argument for going back to this, but I do recall that this didn't fit nicely together with the role of variable declarations and their use in the interpretation of axiom declarations. If I understand what is there now, we surround each axiom with universal quantification over all variables declared (see a later section). This was one of the combinations rejected in Edinburgh, I thought. Of course one can live with this, since after all one can simply never declare any variables and use only explicit quantification.
Discharged: Empty carriers are disallowed.
A (weak) many-sorted homomorphism h from M1 to M2, with M1, M2 e Mod(S,TF,PF,P), consists of a function hs:sM1 -> sM2 for each s e S preserving not only the values of functions but also their definedness, as well as the truth of predicates.
[AT] Reducts of models w.r.t. signature morphisms should be mentioned, since they are used in further explanations.
Discharged: Adjust text accordingly.