In the section §A.3.1 of Lurie's "Higher Topos Theory", the map $ \beta_{X,S} \colon S \otimes FX \to F(S \otimes X) $ is defined without assuming $ F $ has the structure of $ \mathbf{S} $-enriched functor.

However, in the definition, the morphism $ \operatorname{Map}(X, GY) \to \operatorname{Map}(FX, FGY) $ seems to be used.


Am I wrong? Where did I make a mistake?

  • $\begingroup$ It seems to me that this map should exist as soon as $F$ is a functor, it doesn't need to be $S$-enriched. Am I missing something? $\endgroup$ – Arnaud D. Feb 6 '18 at 9:23
  • $\begingroup$ @ArnaudD. My thought: $F$ as a functor only defines $\operatorname{Hom}(X, GY) \to \operatorname{Hom}(FX, FGY)$, which is a morphism in $\mathsf{Set}$, not in $\mathbf{S}$. $\endgroup$ – B. W. Feb 6 '18 at 9:35
  • $\begingroup$ Ah, that's what I missed. Good point. $\endgroup$ – Arnaud D. Feb 6 '18 at 9:43

You are right that the quoted definition of the natural transformation $\beta$ uses an $\mathbf{S}$-enriched functor structure for $F$, in contradiction with the stated hypotheses. The source of the error is that the natural transformation $\beta$ should point in the opposite direction to what Lurie writes, i.e. it should have components $\beta_{S,X} \colon F(S\otimes X) \to S \otimes FX$. This direction for $\beta$ is in fact what is used in the proof of Proposition A.3.1.10 in HTT.

We may then correctly define $\beta_{S,X} \colon F(S\otimes X) \to S \otimes FX$ as the morphism corresponding under the Yoneda embedding to the composite natural transformation \begin{align} \mathcal{D}(S\otimes FX,Y) &\cong \mathbf{S}(S,\underline{\mathcal{D}}(FX,Y)) \\ &\to \mathbf{S}(S,\underline{\mathcal{C}}(GFX,GY)) \\ &\cong \mathcal{C}(S\otimes GFX,GY) \\ &\to \mathcal{C}(S \otimes X,GY) \\ &\cong \mathcal{D}(F(S\otimes X),Y) \end{align} where the first arrow is given by the $\mathbf{S}$-enrichment of the functor $G$, the second arrow is given by composition with the unit $X \to GFX$ of the adjunction, and the isomorphisms are given by the universal properties of tensoring with the simplicial set $S$ and of the adjunction $F \dashv G$. (Note that $\mathcal{C}(-,-)$ and $\underline{\mathcal{C}}(-,-)$ denote the $\mathbf{Set}$-enriched and $\mathbf{S}$-enriched hom objects respectively.)

This definition can be understood in the context of categories bearing an action of the monoidal category $\mathbf{S}$, called "$\mathbf{S}$-actegories". The $\mathbf{S}$-enrichment of the functor $G$ gives it the structure of a "lax morphism" of $\mathbf{S}$-actegories, consisting of a natural transformation $S \otimes GY \to G(S \otimes Y)$ subject to axioms. By the theory of doctrinal adjunction, this lax morphism structure on $G$ corresponds to an "oplax morphism" structure on its left adjoint $F$, consisting of a natural transformation $F(S \otimes X) \to S \otimes FX$ subject to axioms, which is none other than the natural transformation $\beta$ defined above.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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