I would like to know how to show that any linear category is a $\mathrm{Vec}$-module. Here $\mathrm{Vect}$ denotes a category of finite dimensional vector spaces.

More general statement can be found in the paper Dualizable tensor categories, Example 3.2.9


Let $\mathsf{C}$ be a linear category. To give $\mathsf{C}$ the structure of a $\mathsf{Vect}$ (left) module essentially amounts to giving a bifunctor $$\otimes : \mathsf{Vect} \times \mathsf{C} \to \mathsf{C}$$ plus associativity and unitality constraints satisfying a bunch of conditions. I will only give an idea of the definition of $\otimes$, constructing the constraints and checking the conditions is tedious but doable.

On the level of objects, given a vector space $V$ and an object $X \in \mathsf{C}$, we must give something called $V \otimes X$. Since $V$ is finite-dimensional, $V \cong k^n$ for some integer $n$. Intuitively, $k \otimes X$ must be $X$ by the unitality constraint, and $(V \oplus W) \otimes X$ must be $(V \otimes X) \oplus (W \otimes X)$, where now $\oplus$ is the direct sum (biproduct) in the category $\mathsf{C}$. So we define $k^n \otimes X$ to be $X \oplus \dots \oplus X = X^{\oplus n}$.

On the level of morphism, given $f : k^n \to k^m$ a linear map and $g : X \to Y$ a morphism in $\mathsf{C}$, we must give $f \otimes g : k^n \otimes X \to k^m \otimes Y$. The linear map $f$ is represented by a matrix $A = (a_{ij})$, and we have $k^n \otimes X = X^{\oplus n}$, $k^m \otimes Y = Y^{\oplus m}$.

By definition of the biproduct, giving a morphism $X^{\oplus n} \to Y^{\oplus m}$ is the same thing as giving $nm$ morphisms $X \to Y$, in the form of an $m \times n$ matrix. Since $\hom_{\mathsf{C}}(X,Y)$ is a $k$-vector space, we can consider $a_{ij} \cdot g : X \to Y$; these will exactly the entries of the matrix defining the morphism $X^{\oplus n} \to Y^{\oplus m}$.

(I'm not very precise here – I'm using the fact that $\mathsf{Vect}$ is equivalent to its full subcategory spanned by the objects $k^n$.)


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.