Here is a cute observation: Let $F,G : \mathcal{C} \to \mathcal{D}$ be a symmetric monoidal functors between symmetric monoidal categories, and let $\eta : F \to G$ be a monoidal transformation. Then for every dualizable object $V \in \mathcal{C}$ the morphism $\eta_V : F(V) \to G(V)$ is actually an isomorphism! The inverse is given by

$G(V) \cong G(V^*)^* \xrightarrow{\eta_{V^*}*} F(V^*)^* \cong F(V).$

Of course one has to check that this is, in fact, inverse to $\eta_V$. A nice corollary: If $F,G$ are cocontinuous, and every object of $\mathcal{C}$ is a colimit of dualizable objects (which happens quite often), then every monoidal transformation $F \to G$ is an isomorphism, i.e. $\mathrm{Hom}_{c\otimes}(F,G)$ is a groupoid.

Is this well-known? Does somebody know a reference for this observation? The proof is not so hard, but this is one of dozens of lemmas which I need in my thesis and with which I don't want to waste time / place with the proof. In an example this appears in Remark 2.4.7 in Lurie's On the classification of topological field theories.


1 Answer 1


This goes back at least to Saavedra-Rivano "Categories Tannakiennes." In fact, this has a generalisation to Frobenius functors, in Day-Pastro "Note on Frobenius monoidal functors," and to more general settings, as monoidal bicategories: Lopez Franco, Street, Wood; "Duals Invert," (App. Categor. Str. Vol 19).

  • 2
    $\begingroup$ Thank you. In Categories Tannakiennes this is noted (with a sketchy proof, though) in the special case of rigid $\otimes$-categories in Chapter I, Proposition 5.2.3. In Note on Frobenius monoidal functors it is Proposition 7. In Duals Invert, it is Theorem 3.2. $\endgroup$ Sep 13, 2013 at 8:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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