The tensor product of function semimodules. (Q387619)

The author is ``chiefly motivated by the problem of developing a correct formulation of topological quantum field theories over semirings'' required in certain constructions and applications. \textit{M. Atiyah} has axiomatized such theories over rings in a paper since 1988 [Publ. Math., Inst. Hautes Étud. Sci. 68, 175-186 (1988; Zbl 0692.53053)] and tensor products play a central role in his axioms. But Atiyah's axioms cannot be translated to the case of semirings, as the paper has shown. Given a semiring \(S\) and two infinite sets \(A\) and \(B\), one can construct first the semimodules of functions from \(A\) to \(S\) and from \(B\) to \(S\), then the tensor product between them and a canonical map from this tensor product to the semimodule of functions from \(A\times B\) to \(S\). In contrast with the case of a field, in the case of a semiring, this canonical map is not injective in general. The author constructs such a situation and proves (since the proof cannot be translated from the rings) that the surjectivity of this canonical map fails. An interesting discussion in the case of a b-complete commutative semiring \(S\) and a tensor product constructed by \textit{G. L. Litvinov} et al. [Math. Notes 65, No. 4, 479-489 (1999); translation from Mat. Zametki 65, No. 4, 572-585 (1999; Zbl 0974.46057)] is given in the last section of the paper.