Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Let $R$ be any ring, let $\text{Mod}_R$ be the category of right $R$-modules and let $\text{Ab}$ be the category of abelian groups. There is a classical theorem of Eilenberg (I think) which says that for any right exact functor $F:\text{Mod}_R \to \text{Ab}$ which preserves direct sums, there exists a left module structure on $F(R)$ making $F$ naturally isomorphic to the functor $- \otimes_R F(R)$.

Does anyone know any nice "incarnations" of this theorem? By this I mean "nice", simple and concrete right exact functors from $\text{Mod}_R$ to $\text{Ab}$ (for some $R$) preserving direct sums, for which it is not immediately clear that they are given by a tensor product (= isomorphic to a tensor functor) with a left $R$-module, but for which this left $R$-module can still be constructed in a concrete way.

share|improve this question
MacLane attributes that theorem to Watts. –  Mariano Suárez-Alvarez Oct 30 '11 at 22:42
...also to Watts, I mean. –  Mariano Suárez-Alvarez Oct 30 '11 at 22:49
It does not quite fit this theorem of Watts, but Grothendieck's base change module $\mathscr Q$ (EGA III, 7.7.6) comes to mind. –  user2035 Oct 31 '11 at 6:12
add comment

2 Answers 2

up vote 10 down vote accepted

More generally, the Theorem of Eilenberg-Watts says the following: The category of cocontinuous functors $\mathrm{Mod}(R) \to \mathrm{Mod}(S)$ is equivalent to the category of $(R,S)$-bimodules. A bimodule ${}_R M _{S}$ corresponds to the functor $- \otimes_R M$. There is a recent paper by A. Nyman which deals more generally with cocontinuous functors $\mathrm{Qcoh}(X) \to \mathrm{Qcoh}(Y)$ for nice schemes $X,Y$.

The Eilenberg-Watts theorem over schemes, J. Pure Appl. Algebra, 214 (2010), 1922-1954; online.

As for your question, take a finitely generated projective module $P$. Then $\mathrm{Hom}(P,-) : \mathrm{Mod}(R) \to \mathrm{Mod}(R)$ is right exact (since $P$ is projective) and preserves infinite direct sums (since $P$ is finitely generated), thus cocontinuous. In this case the Theorem shows that $\mathrm{Hom}(P,-) \cong (-) \otimes P^*$, where $P^* = \mathrm{Hom}(P,R)$ is the dual of $P$. More generally, for every vector bundle $V$ on some scheme/manifold we have $\underline{\mathrm{Hom}}(V,-) \cong (-) \otimes V^*$.

share|improve this answer
add comment

There is a paper by Mark Hovey which discusses incarnations of the Eilenberg-Watts theorem in homotopy theory. First he reviews the algebraic version and provides an equivalent formulation wherein the theorem says that the category of abelian groups is left self-contained (note that Eilenberg also proved Ab is right self-contained, though it seems Watts did not prove this). Hovey then finds the correct definition of this concept in homotopy theory and proves a very general theorem about when a model category has this property. His theorem holds for topological spaces, simplicial sets, chain complexes, and all the models of spectra. Note, there is a more recent version of this paper on Hovey's website, but it does not compare the new results to the classical ones as fully.

On page 2 of Hovey's paper above is an open problem which you might be interested in. It asks for conditions such that a closed symmetric monoidal category is self-contained (rather than homotopically self-contained as Hovey's paper answers).

share|improve this answer
add comment

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.