With braiding
With duals for objects
category with duals (list of them)
dualizable object (what they have)
ribbon category, a.k.a. tortile category
With duals for morphisms
monoidal dagger-category?
With traces
Closed structure
Special sorts of products
Semisimplicity
Morphisms
Internal monoids
Examples
Theorems
In higher category theory
abstract duality: opposite category,
concrete duality: dual object, dualizable object, fully dualizable object, dualizing object
Examples
between higher geometry/higher algebra
Langlands duality, geometric Langlands duality, quantum geometric Langlands duality
In QFT and String theory
A (left/right) dual to an object in a monoidal category $\mathcal{C}$ is a left/right adjoint to the object regarded as a morphism in the delooping 2-category $\mathbf{B}\mathcal{C}$. If a dual exists, the object is called dualizable.
Being dualizable may often be thought of as a category-theoretic notion of finiteness [Pareigis 1976 p. 113] for objects in a monoidal category. For instance:
a vector space is dualizable in Vect with its standard tensor product of vector spaces precisely if it is a finite-dimensional vector space (cf. Exp. ),
a spectrum is dualizable in the stable homotopy category with its smash product precisely if it is a finite spectrum.
A more precise intuition is that an object is dualizable if its βsizeβ is no larger than the βadditivityβ of the monoidal category. For instance, since Vect and the stable homotopy category are finitely additive, but not infinitely so, dualizability there is indeed a notion of plain finiteness, as is the case for many monoidal categories in which one considers dualizability.
However, in a monoidal category which is not additive at all, such as Set (or any cartesian monoidal category), only the terminal object is dualizable β whereas in an βinfinitely additiveβ monoidal category such as Rel or SupLat, many βinfiniteβ objects are dualizable. (In $Rel$, all objects are dualizable.)
Beware that there are other notions of βdual objectβ, distinct from this one. See for example dual object in a closed category (4), and also the discussion at category with duals.
An object $A$ in a monoidal category $(\mathcal{C}, \otimes, 1)$ is dualizable if it has an adjoint when regarded as a morphism in the one-object delooping bicategory $\mathbf{B}\mathcal{C}$ corresponding to $\mathcal{C}$. Its adjoint in $\mathbf{B}\mathcal{C}$ is called its dual in $C$ and often written as $A^*$.
(This notion, though not the terminology, is due to Lindner 1978.)
If $C$ is braided then left and right adjoints in $\mathbf{B}C$ are equivalent; otherwise one speaks of $A$ being left dualizable or right dualizable.
Explicitly this means the following:
A right duality between objects $A, A^\ast \in (\mathcal{C}, \otimes, \mathbb{1})$
consists of
a morphism of the form
called the counit of the duality, or the evaluation map;
a morphism of the form
called the unit or coevaluation map
such that
(triangle identity) the following diagrams commute
and
where $\alpha$ denotes the associator of the monoidal category $\mathcal{C}$, and $\ell$ and $r$ denote the left and right unitors, respectively.
(Terminology)
Unfortunately, conventions on left and right vary and sometimes contradict their use for adjoints. A common convention is that a right dual of $A$ is an object $A^*$ equipped with a unit (or coevaluation)
and counit (or evaluation)
satisfying the βtriangle identitiesβ familiar from adjunctions.
With this convention, if $\otimes$ in $C$ is interpreted as composition in $\mathbf{B} C$ in diagrammatic order, then right duals in $C$ are the same as right adjoints in $\mathbf{B}C$ β whereas if $\otimes$ in $C$ is interpreted as composition in $\mathbf{B} C$ in classical βLeibnizianβ order, then right duals in $C$ are the same as left adjoints in $\mathbf{B} C$.
Of course, in a symmetric monoidal category, there is no difference between left and right duals.
There are various equivalent definitions of dualizability, some of which are apparently weaker than the explicit definition in terms of both unit and counit, or which assume only one of them together with a universal property for it.
Regarding the equivalent characterization via adjointable tensor products, beware Rem. below.
A dualizable object $A$, def. , for which the structure unit/counit maps between $A \otimes A^\ast$ and the unit object are isomorphisms is called an invertible object.
If every object of $C$ has a left and right dual, then $C$ is called a rigid monoidal category or an autonomous monoidal category. If moreover it is symmetric, it is called a compact closed category.
See category with duals for more discussion.
Given a morphism $f \colon X \to Y$ between two dualizable objects in a symmetric monoidal category, the corresponding dual morphism
is the one obtained by $f$ by composing the duality unit, the counit and the braidingβ¦
(finite-dimensional vector spaces)
Let $V$ be a finite-dimensional vector space over a field $k$, and let $V^* = Hom(V,k)$ be its usual dual vector space. We can define $\varepsilon\colon V^* \otimes V \to k$ to be the obvious pairing. If we also choose a finite basis $\{v_i\}$ of $V$, and let $\{v_i^*\}$ be the dual basis of $V^*$, then we can define $\eta\colon k \to V\otimes V^*$ by sending $1$ to $\sum_i v_i \otimes v_i^*$. It is easy to check the triangle identities, so $V^*$ is a dual of $V$ in $Vect_k$.
(dualizable modules)
More generally, in the symmetric monoidal category of modules over a commutative ring, dualizable objects are precisely finitely generated projective modules. See at dualizable module for more.
(Spanier-Whitehead duality)
Let $M$ be a finite-dimensional smooth manifold, choose an embedding $M\hookrightarrow \mathbb{R}^n$ for some $n$, and let $Th(N X)$ be the Thom spectrum of the normal bundle of this embedding. Then the Thom collapse map defines an $\eta$ which exhibits $Th(N X)$ as a dual of $\Sigma_+^\infty M$ in the stable homotopy category. This is a version of Spanier-Whitehead duality.
(adjoint endofunctors)
Given any category $\mathscr{C}$, the endo-functor category $End(\mathcal{C}) \,\coloneqq\, \text{Fun}(\mathscr{C},\mathscr{C})$ canonically becomes a monoidal category, with tensor product given by composition of functors and horizontal composition of natural transformations (whiskering).
Under this identification, an object in $End(\mathcal{C})$ is dualizable iff it is an adjoint functor, with evaluation and co-evaluation given by the counit of the adjunction and the unit of the adjunction, respectively.
(PoincarΓ© duality algebras)
A C*-algebra is a PoincarΓ© duality algebra if it is a dualizable object in the symmetric monoidal category KK with dual its opposite algebra.
See at KK-theory β Poincare duality.
For $E$ an E-β ring, then in the (β,1)-category of (β,1)-modules $E Mod$ the dualizable objects coincide with the compact objects and the perfect objects.
See at (β,1)-category of (β,1)-modules β Compact generation for more.
In a symmetric monoidal category, the following are equivalent:
an object $A$ is dual to $A^\ast$ in the sense of Def. , with evaluation map $ev_A \,\colon\, A^\ast \otimes A \to \mathbb{1}$ (1),
the map
is an isomorphism (a bijection of hom sets) for all objects $X$, $Y$.
Prop. says in particular that for dual objects $A$, $A^\ast$ the map (3) exhibits the tensor product functor $(\text{-}) \otimes A^\ast$ as a right adjoint to $(\text{-}) \otimes A$ (hence: an internal hom-functor):
If the ambient category is indeed closed monoida with internal hom denoted $[-,-]$, this means that $(\text{-}) \otimes A^\ast \,\simeq\, [A,\text{-}]$ (by essential uniqueness of adjoints) and hence in particular that:
But since (by symmetry) the object $A \,\simeq\, (A^{\ast})^\ast$ is also the dual object to $A^\ast$, there is
a further right adjoint to the internal hom (an βamazing right adjointβ)
which coincides with the left adjoint to make an ambidextrous adjunction:
(Tensor-adjuncatibility does not imply dualizability)
Prop. does not claim that for $A'$ to be a dual object of $A$ it is sufficient that $(\text{-}) \otimes A$ has a right adjoint of the form $(\text{-}) \otimes A'$, instead it claims that the hom-isomorphism for this adjunction must be of the particular form (3).
A counterexample to sufficiency of the weaker condition is indicated by Noah Snyder in math.SE:a/692318, referring to Exp. 2.20 in arXiv:1406.4204.
Dualizable objects support a good abstract notion of trace. (β¦)
Dualizable objects in an symmetric monoidal (β,1)-category are already fully dualizable objects. The cobordism hypothesis implies that there is a canonical $O(1) \simeq \mathbb{Z}/2\mathbb{Z}$-action on the β-groupoid of dualizable objects, and this is just the dualizing operation. See at cobordism hypothesis β Framed version β Implications: Canonical O(n)-action.
In a closed category $(\mathcal{C}, [-,-], 1)$ the (weak) dual to an object $X \in \mathcal{C}$ is defined to be the internal hom into the unit object
In a closed monoidal category $\mathbb{D}X$ (4) is also called the weak dual of $X$ [Becker & Gottlieb, p. 5] (in contrast with the monoidal dual of Def. , which would then be called the strong dual [Dold & Puppe 1984 Def. 1.2]).
(reflexive dualizable object)
If the adjunct
of
is an isomorphism (identifying $X$ with its double dual) then $X$ is called reflexive as a weak dualizable object. [Deligne & Milne 1982 p 111, Dold & Puppe 1984 Def. 1.2]
If $\mathcal{C}$ is a compact closed category, def. , then the weak dual $\mathbb{D}X$ is also the strong dual object $X^\ast$ to $X$ in the above monoidal sense. Here dualization exhibits $\mathcal{C}$ as a star-autonomous category ($\mathbb{D}(-) = (-)^\ast$ is the star-operation).
The property of $X$ being dualizable can be expressed as a property of the weak dual, namely that the induced map $\mathbb{D}X \otimes X \to [X,X]$ is an isomorphsim.
An object in a symmetric monoidal (β,n)-category $C$ is called dualizable if it is so as an object in the ordinary symmetric monoidal homotopy category $Ho(C)$.
This appears as (Lurie, def. 2.3.5).
This means that an object in $C$ is dualizable if there exists unit and counit 1-morphism that satisfy the triangle identity up to homotopy. The definition does not demand that this homotopy is coherent (that it satisfies itself higher order relations up to higher order k-morphisms).
If the structure morphisms of the adjunction of a dualizable object have themselves all adjoints, then the object is called a fully dualizable object.
As before, we may equivalently state this after delooping the monoidal structure and passing to the $(\infty,n+1)$-category $\mathbf{B}C$. Then $C$ has duals for objects precisely if $\mathbf{B}C$ has all adjoints.
In a linearly distributive category, duality is naturally defined by mixing the two tensors $(\otimes,\top)$ and $(\parr,\bot)$: the unit is $i : \top \to A \parr A^*$ and the counit is $ev:A^* \otimes A \to \bot$. The triangle identities make sense by inserting the linear distributivities; they assert that the following composites are identities:
A symmetric linearly distributive category is (symmetric) star-autonomous if and only if all objects have duals in this sense. The same is true in the non-symmetric case if we require both left and right duals.
This notion of duality generalizes to that of linear adjoints in a linear bicategory, and also to dual objects in a polycategory.
geometry | monoidal category theory | category theory |
---|---|---|
perfect module | (fully-)dualizable object | compact object |
The strong-duality of finite-dimensional vector spaces is the lead-in example used in
to motivate category theory in the first place.
Further original articles with dedicated discussion of the notion of dualizable objects in monoidal categories:
Bodo Pareigis, p. 113 in: Non-additive ring and module theory IV: The Brauer group of a symmetric monoidal category, Lecture Notes in Mathematics 549 (1976) [doi:10.1007/BFb0077339]
(calls dualizable objects βfinite objectsβ)
Thomas S. Ligon, Galois-Theorie in monoidalen Kategorien, Munich (1978) [pdf, doi:10.5282/edoc.14958]
transl: Galois theory in monoidal categories (2019) [pdf, doi:10.5282/edoc.24952]
(follows Pareigis 1976)
Harald Lindner, Adjunctions in monoidal categories, Manuscripta Mathematica 26 1-2 (1978) 123-139 [doi:10.1007/BF01167969]
(discusses dual objects in a monoidal category as adjunctions in its delooping 2-category)
Pierre Deligne, James Milne: Β§1 pp. 110 of: Tannakian categories, in: Hodge Cycles, Motives, and Shimura Varieties, Lecture Notes in Mathematics 900, Springer (1982) [doi:10.1007/978-3-540-38955-2_4, webpage]
Albrecht Dold, Dieter Puppe, Β§1 of: Duality, Trace and Transfer, Proceedings of the Steklov Institute of Mathematics, 154 (1984) 85β103 [mathnet:tm2435, pdf]
Early history with an eye towards formulating Becker-Gottlieb transfer:
Further developments:
Mark Hovey, John Palmieri, Neil Strickland: Def. 1.1.2 and Thm. 2.1.3 in: Axiomatic stable homotopy theory, Memoirs Amer. Math. Soc. 610 (1997) [ISBN:978-1-4704-0195-5, pdf]
L. Gaunce Lewis, Peter May, Mark Steinberger (with contributions by J.E. McClure): Β§III.1 in: Equivariant stable homotopy theory, Springer Lecture Notes in Mathematics 1213 (1986) [pdf, doi:10.1007/BFb0075778]
Review:
Peter May, Β§2 in: Picard Groups, Grothendieck Rings, and Burnside Rings of Categories, Advances in Mathematics 163 1 (2001) 1-16 [pdf, doi:10.1006/aima.2001.1996]
Dominic Culver, Mitchell Faulk, Duality Notes, lecture notes for West Coast Algebraic Topology Summer School (2014) [pdf, pdf mathtube]
(towards fully dualizable objects)
See also:
Further examples:
On self-dual objects and the corresponding inner products and dagger-structure:
Monoidal categories with freely adjoint duals:
The notion of fully dualizable objects in a symmetric monoidal $(\infty,n)$-category:
Last revised on September 27, 2023 at 08:31:09. See the history of this page for a list of all contributions to it.