Anchored planar algebras and 3-categorical graphical calculus
Brett Hungar · 2025
Read Anchored planar algebras and 3-categorical graphical calculus by Brett Hungar (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-00995.txt.
Introduction
Brett Hungar, writing alone, compares two diagram languages in higher category theory. Anchored planar algebras generalize Vaughan Jones' planar algebras. They have seen use in analysis and in field theory. They carry a natural 3D graphical calculus. Tricategories carry a similar calculus of their own.
Hungar shows these algebras live in a particular tricategory. That single fact has several consequences. More general tools can now be applied to them. The types of diagrams they can interpret also expand. The result bridges two ways of drawing higher structures. Results proved in one setting can transfer to the other. It clarifies where these algebras sit in the broader theory.
Catalog record
| CREATOR | Brett Hungar |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 290 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-00995
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Anchored planar algebras and 3-categorical graphical calculus Brett Hungar Abstract . Anchored planar algebras, a generalized notion of Vaughan Jones’ planar algebras, have recently seen use in higher category theory, functional analysis, and TQFT applications. These algebras are equipped with a natural 3-dimensional graphical calculus. We compare this graphical calculus with the 3-dimensional graphical calculus associated to tricategories, and we show that anchored planar algebras can be thought of as living in a particular tricategory. This allows for more general techniques to be applied to anchored planar algebras and expands the types of diagrams that anchored planar algebras can interpret. 1. Introduction Anchored planar algebras were first introduced in [ HPT16 ] , building off Vaughan Jones’ notion of planar algebras [ Jon99 ] . These algebras have already seen use in higher category theory and functional analysis as in [ HPT23 ] , and related graphical calculi have applications in physics [ HBJP23 , KJS+24 ] . They are also related to notions of categorified trace, such as the categorified traces discussed in [ HPT15 ] and [ PS13 ] . Anchored planar algebras, as algebras over an operad, have a natural graphical calculus arising from that operad. This graphical calculus can be seen to be 3-dimensional by viewing anchored planar tangles as strings on tubes that are allowed to merge and twist. The prototypical example of 3-dimensional graphical calculus comes from 3-categories themselves, viewing objects as regions, 1-cells as sheets, 2-cells as strings, and 3-cells as junctions. It is reasonable to wonder whether there is a connection between these two calculi. In this work, we show that we can interpret the operadic graphical calculus within the graphical calculus of 3-categories. Theorem A . For every planar pivotal 3-category {\mathcal{C}} , objects A,B , morphisms M:A\to B and M^{*}:B\to A , and adjoint pair F:M\boxtimes_{B}M^{*}\to A , F^{*}:A\to M\boxtimes_{B}M^{*} with isomorphisms \tau^{\ell}:(\operatorname{id}_{M}\boxtimes-)\circ F\to(-^{*}\boxtimes\operatorname{id}_{M}^{*})\circ F,\ \ \tau^{r}:F^{*}\circ(\operatorname{id}_{M}\boxtimes-)\to F^{*}\circ(-^{*}\boxtimes\operatorname{id}_{M}^{*}), satisfying certain axioms, we can construct an anchored planar algebra internal to \operatorname{End}(1_{A}) which interprets anchored planar tangles via the graphical calculus of {\mathcal{C}} . Moreover, all anchored planar algebras in braided fusion categories arise this way. Theorem B . For every anchored planar algebra {\mathcal{P}} , there is a 3-category {\mathcal{C}} and objects as in Theorem A such that the anchored planar algebra internal to the graphical calculus of {\mathcal{C}} is equivalent to {\mathcal{P}} . Morally, we should think of this theorem as lending evidence to the idea that all 3D graphical calculi are subexamples of the 3D graphical calculus of a 3-category. In Section 2 , we recall the graphical definition of anchored planar algebras and describe the 3-category in which our construction in Theorem A takes place. Our proof proceeds by taking a pivotal module fusion category for a braided pivotal fusion category and showing that it naturally gives rise to certain maps inside a particular 3-category that define an anchored planar algebra. Moreover, we show that this anchored planar algebra is equivalent to the one given by the construction in [ HPT16 ] . Since it is already known that every anchored planar algebra arises this way (shown in [ HPT16 ] ), this also shows that every anchored planar algebra can be interpreted internal to the graphical calculus of some 3-category. We discuss this construction and prove Theorem B in Section 3 . Since the graphical calculus of 3-categories can interpret more geometric shapes than the anchored planar operad, this also enriches the algebraic structure of anchored planar algebras. We are able to interpret surfaces of higher genus and comultiplication-like operations, in addition to the generating operations of the anchored planar operad. In Section 3.3 , we use this structure to answer a question from [ HPT24 , 1.1.3] . 2. Background 2.1. 3D graphical calculus of {\mathcal{A}}-\mathsf{FusCat} Our general setting is the 3-category {\mathcal{A}}-\mathsf{FusCat} . We begin with a brief discussion of the structure of this 3-category and the various graphical calculi we will use within {\mathcal{A}}-\mathsf{FusCat} . Our construction of fusion and braided fusion categories, especially in the enriched setting, is based on [ EGNO15 ] , [ JMPP21 ] , [ BJS21 ] , and [ KZ18 ] . Definition 2.1 ( [ BJS21 ] , [ JMPP21 ] ) . Let {\mathcal{A}} be a braided fusion category. The 3-category {\mathcal{A}}-\mathsf{FusCat} has objects {\mathcal{A}} -enriched fusion categories (that is, fusion categories {\mathcal{X}} with a braided monoidal functor {\mathcal{A}}\to{\mathcal{Z}}({\mathcal{X}}) ); a 1-cell {}_{\mathcal{X}}{\mathcal{M}}_{\mathcal{Y}} is a (finitely semisimple) {\mathcal{X}}-{\mathcal{Y}} bimodule category {\mathcal{M}} equipped with an {\mathcal{A}} -centered structure; 2-cells are given by bimodule functors; and 3-cells are bimodule natural transformations. Composition of 1-cells is given by relative Deligne product. By a {\mathcal{A}} -centered structure on {}_{\mathcal{X}}{\mathcal{M}}_{\mathcal{Y}} , we mean a natural isomorphism between the left {\mathcal{A}} -action given by {\mathcal{A}}\to{\mathcal{Z}}({\mathcal{X}}) and the right action given by {\mathcal{A}}\to{\mathcal{Z}}({\mathcal{Y}}) . This isomorphism must satisfy appropriate hexagon identities (with the action associators) and must intertwine the tensorators of the {\mathcal{X}} and {\mathcal{Y}} actions. Lemma 2.2 ( [ BJS21 ] , [ KZ18 ] ) . Given a braided fusion category {\mathcal{B}} and a {\mathcal{B}} -enriched fusion category {\mathcal{Y}} , there exists a braided fusion category {\mathcal{A}} , object {\mathcal{X}}\in{\mathcal{A}}-\mathsf{FusCat} , and 1-cell {}_{\mathcal{A}}{\mathcal{M}}_{\mathcal{X}} with monoidal (resp. braided) equivalences \operatorname{End}^{\mathcal{A}}({\mathcal{M}}_{\mathcal{X}})=\operatorname{End}_{{\mathcal{A}}-\mathsf{FusCat}}({\mathcal{M}}_{\mathcal{X}})\cong{\mathcal{Y}} and {\mathcal{Z}}^{\mathcal{A}}({\mathcal{X}})=\operatorname{End}_{{\mathcal{A}}-\mathsf{FusCat}}({\mathcal{X}})\cong{\mathcal{B}} . Proof. This is well-known; we briefly recall the construction given in [ KZ18 ] . Choose {\mathcal{A}}={\mathcal{B}} and {\mathcal{X}}={\mathcal{B}} . Certainly we have a braided monoidal embedding {\mathcal{B}}\hookrightarrow{\mathcal{Z}}^{\mathcal{B}}({\mathcal{B}}) ; we just need to show this is essentially surjective. Given an object b and enriched half-braiding \theta for b , notice that enriched naturality means the diagram {\lx@inpgf@ignorespaces{\left[x,y\right]_{\mathcal{B}}}} {\lx@inpgf@ignorespaces{\left[x\otimes b,y\otimes b\right]_{\mathcal{B}}}} {\lx@inpgf@ignorespaces{\left[b\otimes x,b\otimes y\right]_{\mathcal{B}}}} {\lx@inpgf@ignorespaces{\left[b\otimes x,y\otimes b\right]}} \scriptstyle{\lx@inpgf@ignorespaces-\otimes\operatorname{id}} \scriptstyle{\lx@inpgf@ignorespaces\operatorname{id}\otimes-} \scriptstyle{\lx@inpgf@ignorespaces\theta_{x}} \scriptstyle{\lx@inpgf@ignorespaces\theta_{y}} commutes for all x,y\in{\mathcal{B}} . Recall that in a fusion category, internal hom is given by [x,y]_{\mathcal{B}}\cong x\otimes y^{*} . Then setting x=1_{\mathcal{B}} , we have that the diagram {\lx@inpgf@ignorespaces y} {\lx@inpgf@ignorespaces y\otimes b\otimes b^{*}} {\lx@inpgf@ignorespaces b\otimes y\otimes b^{*}} {\lx@inpgf@ignorespaces y\otimes b\otimes b^{*}} \scriptstyle{\lx@inpgf@ignorespaces\eta_{b}} \scriptstyle{\lx@inpgf@ignorespaces\beta_{by}\circ\eta_{b}} \scriptstyle{\lx@inpgf@ignorespaces\theta_{1}=\operatorname{id}} \scriptstyle{\lx@inpgf@ignorespaces\theta_{y}} commutes for all y , so \theta_{y}=\beta_{b,y} (where \beta is the braiding in {\mathcal{B}} ). Thus an enriched half-braiding must be the same as the original braiding, so the map {\mathcal{B}}\to{\mathcal{Z}}^{\mathcal{B}}({\mathcal{B}}) is essentially surjective and an equivalence. For {\mathcal{M}} , choose an indecomposable {\mathcal{B}} -module; it follows that \operatorname{End}({\mathcal{M}}_{\mathcal{B}})\cong{\mathcal{B}} . ∎ In the arguments that follow, we frequently switch between two different graphical calculi for Gray categories, using the coherence result in [ GPS95 ] that every 3-category is triequivalent to a Gray category. The first graphical calculus is the 3D graphical calculus found in [ BMS24 ] . In our graphical calculations, the Deligne tensor product should be read back to front, functor composition read left to right, and natural transformation composition bottom to top. The second graphical calculus is written using string diagrams plus arrows representing transformations of string diagrams. In this graphical calculus, Deligne tensor product is read left to right, functor composition bottom to top, and natural transformation composition is in the time direction. Both of these graphical calculi are coherent for general 3-categories by [ GPS95 ] , [ BMS24 ] , and [ Gut19 ] . We illustrate the two calculi in Figure 1 . \scriptstyle\mathcal{C} \scriptstyle\mathcal{D} \scriptstyle F \scriptstyle G \scriptstyle\alpha \scriptstyle\mathcal{E} \scriptstyle\mathcal{E}