nLab
star-autonomous category

Contents

Definition

A *-autonomous category [1] is a symmetric closed monoidal category C,,I with a dualizing object: an object such that the canonical morphism

d A:A(A),d_A: A \to (A \multimap \bot) \multimap \bot ,

which is the transpose of the evaluation map

ev A,:(A)A,ev_{A,\bot}: (A \multimap \bot) \otimes A \to \bot ,

is an isomorphism for all A. (Here, denotes the internal hom.)

Define a functor () *=(). The map d A is natural in A, so that there is a natural isomorphism d:1() **. We also have

hom(AB,C *) =hom(AB,C) hom((AB)C,) hom(A(BC),) hom(A,(BC)) =hom(A,(BC) *)\begin{aligned} \hom(A\otimes B,C^*)& = \hom(A\otimes B, C\multimap\bot) \\ & \cong \hom((A\otimes B)\otimes C,\bot) \\ & \cong \hom(A\otimes(B\otimes C), \bot) \\ & \cong \hom(A,(B\otimes C)\multimap\bot) \\ & = \hom(A,(B\otimes C)^*) \end{aligned}

The functor () *, together with these two isomorphisms, can be taken as an alternative definition of a *-autonomous category. The dualizing object is then defined as I *.

Examples

  • A simple example of a *-autonomous category is the category of finite-dimensional vector spaces over some field k. In this case k itself plays the role of the dualizing object, so that for an f.d. vector space V, V * is the usual dual space of linear maps into k.

  • More generally, any compact closed category is *-autonomous with the unit I as the dualizing object.

Mike: Can someone fill in some examples of *-autonomous categories that are not compact closed?

Finn: Blute and Scott in ‘Category theory for linear logicians’ (from here) give an example: reflexive topological vector spaces where the topologies are ‘linear’, i.e. Hausdorff and with 0 having a neighbourhood basis of open linear subspaces; ‘reflexive’ meaning that the map d V as above is an isomorphism. It seems this category is *-autonomous but not compact. I don’t know enough topology to make much sense of it, though.

Todd: Finn, I expect that example is in Barr’s book, which would then probably have a lot of details. But I must admit I have not studied that book carefully. Also, the Chu construction was first given as an appendix to that book.

John: I get the impression that a lot of really important examples of *-autonomous categories — important for logicians, anyway — are of a more ‘syntactical’ nature, i.e., defined by generators and relations.

  • A more interesting example of a *-autonomous category is the category of sup-lattices and sup-preserving maps (= left adjoints). Clearly the poset of sup-preserving maps hom(A,B) is itself a sup-lattice, so this category is closed. The free sup-lattice on a poset X is the internal hom of posets [X op,Ω]; in particular the poset of truth values Ω is a unit for the closed structure. Define a duality () * on sup-lattices, where X *=X op is the opposite poset (inf-lattices are sup-lattices), and where f *:Y *X * is the left adjoint of f op:X opY op. In particular, take as dualizing object D=Ω op. Some simple calculations show that under the tensor product defined by the formula (XY *) *, the category of sup-lattices becomes a *-autonomous category.

  • Another interesting example is due to Yuri Manin: the category of quadratic algebras. A quadratic algebra over a field k is a graded algebra A=T(V)/I, where V is a finite-dimensional vector space in degree 1, T(V) is the tensor algebra (the free k-algebra generated by V), and I is a graded ideal generated by a subspace RVV in degree 2; so R=I 2, and A determines the pair (V,R). A morphism of quadratic algebras is a morphism of graded algebras; alternatively, a morphism (V,R)(W,S) is a linear map f:VW such that (ff)(R)S. Define the dual A * of a quadratic algebra given by a pair (V,R) to be that given by (V *,R ) where R V *V * is the kernel of the transpose of the inclusion i:RVV, i.e., there is an exact sequence

    0R V *V *i *R *00 \to R^\perp \to V^* \otimes V^* \overset{i^*}{\to} R^* \to 0

    Define a tensor product by the formula

    (V,R)(W,S)=(VW,(1 Vσ1 W)(RS))(V, R) \otimes (W, S) = (V \otimes W, (1_V \otimes \sigma \otimes 1_W)(R \otimes S))

    where σ:VWWV is the symmetry. The unit is the tensor algebra on a 1-dimensional space. The hom that is adjoint to the tensor product is given by the formula AB=(AB *) *, and the result is a *-autonomous category.

  • In a similar vein, I am told that there is a *-autonomous category of quadratic operad?s.

  • Girard’s coherence space?s, developed as models of linear logic, give an historically important example of a *-autonomous category.

  • Hyland and Ong have given a completeness theorem for multiplicative linear logic in terms of a *-autonomous category of fair games, part of a series of work on game semantics for closed category theory (compare Joyal’s interpretation of Conway games as forming a compact closed category).

  • The Chu construction can be used to form many more examples of *-autonomous categories.

References

  • Barr, Michael, *-Autonomous Categories. LNM 752, Springer-Verlag 1979.

More references to come…