Braided, Symmetric, and Ribbon Monoidal Categories

Intuition Beginner

Some operations care about order. If you swap two objects and then swap them again, you might expect to be back where you started. For ordinary pairs of things this is exactly what happens: two swaps cancel. But there is a richer world where each swap is a genuine crossing, the way one strand of a braid passes over another. In that world, two crossings in the same direction do not cancel; they form a twist. A braided monoidal category is the setting that combines the idea of putting objects together with this richer, crossing-aware kind of swap.

Picture two ribbons hanging side by side. To exchange their positions you must let one pass in front of the other. A second exchange, in the same direction, adds a second pass on top of the first; the ribbons now wind around each other instead of returning to the plain starting picture. Symmetric swaps are the special case where a pass in front and a pass behind are treated as the same, so any two exchanges undo. Braided swaps refuse this: the over-pass and the under-pass are different, and the category keeps a record of every crossing.

Add two ideas and the picture deepens. First, every strand can carry a twist, like a ribbon given a half-turn; this is the ribbon structure. Second, every object can have a dual partner, like a direction and its reverse, so that a strand can turn back on itself. Together these let you take an open braid, cap its ends with cups and caps coming from the duals, and close it into a loop. The category then assigns to that loop a number. That number is a knot invariant, and this is the doorway from category theory to the geometry of knots and three-dimensional space.

Visual Beginner

Picture two panels. On the left panel two straight strands cross once, and crossing them again returns the picture to two straight strands: a crossing equals its mirror, and two crossings vanish. On the right panel the crossing records which strand goes over and which goes under. Crossing twice in the same direction leaves a visible full twist of one strand around the other, a genuine braid rather than the empty one.

feature	symmetric crossing	braided crossing
over vs under	treated as the same	recorded separately
crossing twice	returns to the start	leaves a full twist
three crossings around a triangle	a permutation	obeys the braid relation
closing a braid into a loop	gives a permutation	gives a knot or link

The deep point is that the braided panel still satisfies a rule: three crossings arranged around a triangle obey the braid relation, where sliding one crossing past the other two does not change the result. That relation is the heart of the Yang-Baxter equation, and it is what lets a braided monoidal category describe both the algebra of braids and the geometry of knots.

Worked example Beginner

Work with three strands hanging from three fixed top pegs to three fixed bottom pegs, labelled left, middle, right. Two moves are allowed. Move ${\sigma }_1$ crosses the left and middle strands, with the left strand passing in front. Move ${\sigma }_2$ crosses the middle and right strands, with the middle strand passing in front. Every braid on these three pegs is a sequence of these two moves.

Step 1. Do move ${\sigma }_1$ twice. After the first crossing the left and middle strands have swapped pegs; after the second crossing they swap again and return to their starting pegs. Read only by which strand ends at which peg, the result looks like doing nothing. But the diagram on the page holds two in-front crossings stacked on top of each other, a full twist of the two strands. This full twist is a genuine braid, different from the straight-through braid with no crossings.

Step 2. Collapse to the symmetric version, where an in-front crossing and an in-behind crossing count as the same move. In that version the move ${\sigma }_1$ undoes itself, so doing it twice gives the identity, the straight-through picture. The braid group on three strands drops to the symmetric group on three letters, where every swap squares to nothing.

Step 3. Compare two braids each with two crossings. The braid ${\sigma }_1$ then ${\sigma }_2$ crosses the left pair and then the right pair. The braid ${\sigma }_2$ then ${\sigma }_1$ crosses the right pair first and then the left pair. Both have two crossings, but as diagrams they are different braids: there is no sequence of braid rules that turns one into the other.

What this tells us: braids remember crossings, symmetric swaps forget them, and the braid relation is the only rule that lets three crossings trade places. The jump from symmetric to braided is exactly the jump from counting permutations to recording how strands weave, and that recorded weaving is what a knot invariant later reads off.

Check your understanding Beginner

Formal definition Intermediate+

A braiding on a monoidal category $(\mathcal{C},\otimes ,I,\alpha ,\lambda ,\rho )$ (the structure of 41.07.01) is a natural isomorphism $c_{X,Y}:X\otimes Y\stackrel{\sim }{\to }Y\otimes X$ satisfying the two hexagon axioms, which express compatibility of the braiding with the associator on a triple tensor product ^{[Mac Lane 1998]}:

$$c_{X,Y\otimes Z}={\alpha }_{Y,Z,X}\circ ({\mathrm{id}}_Y\otimes c_{X,Z})\circ {\alpha }_{Y,X,Z}^{-1}\circ (c_{X,Y}\otimes {\mathrm{id}}_Z)\circ {\alpha }_{X,Y,Z},$$ $$c_{X\otimes Y,Z}={\alpha }_{Z,X,Y}^{-1}\circ (c_{X,Z}\otimes {\mathrm{id}}_Y)\circ {\alpha }_{Z,X,Y}\circ ({\mathrm{id}}_X\otimes c_{Y,Z})\circ {\alpha }_{X,Y,Z}^{-1}.$$

A braided monoidal category is a monoidal category equipped with a braiding. A braided monoidal functor $F:\mathcal{C}\to \mathcal{D}$ is a strong monoidal functor (in the sense of 41.07.01) with $F(c_{X,Y}^{\mathcal{C}})\circ {\varphi }_{X,Y}={\varphi }_{Y,X}\circ c_{FX,FY}^{\mathcal{D}}$.

Definition (symmetric monoidal category). A braiding is a symmetry, and $\mathcal{C}$ is symmetric monoidal, when $c_{Y,X}\circ c_{X,Y}={\mathrm{id}}_{X\otimes Y}$ for all $X,Y$. The category $({\mathbf{Vect}}_k,{\otimes }_k)$ is symmetric via $v\otimes w\mapsto w\otimes v$; the category $\mathbf{Rep}(U_q\mathfrak{g})$ of representations of a quantum group at generic $q$ is braided but not symmetric, because its braiding and its inverse are genuinely different maps.

Definition (braid group; braid groupoid). The braid group $B_n$ on $n$ strands is the group on generators ${\sigma }_1,\dots ,{\sigma }_{n-1}$ with relations ${\sigma }_i{\sigma }_j={\sigma }_j{\sigma }_i$ for $|i-j|\ge 2$ and the braid relation ${\sigma }_i{\sigma }_{i+1}{\sigma }_i={\sigma }_{i+1}{\sigma }_i{\sigma }_{i+1}$. The braid groupoid (or braid category) $\mathcal{B}$ has objects the natural numbers, ${\mathrm{Hom}}_{\mathcal{B}}(m,n)=\varnothing$ for $m\ne n$, and ${\mathrm{Hom}}_{\mathcal{B}}(n,n)=B_n$, with tensor product given by addition of objects and juxtaposition of braids. Adding the relations ${\sigma }_i^2=1$ to $B_n$ yields the symmetric group $S_n$, so a symmetric monoidal category is one whose braiding factors through this quotient — the symmetric case is the braid case with the over-and-under data erased ^{[Kassel 1995]}.

Definition (Yang-Baxter equation). An $R$-matrix on a vector space $V$ is an invertible linear map $R:V\otimes V\to V\otimes V$ satisfying the Yang-Baxter equation

$$(R\otimes {\mathrm{id}}_V)({\mathrm{id}}_V\otimes R)(R\otimes {\mathrm{id}}_V)=({\mathrm{id}}_V\otimes R)(R\otimes {\mathrm{id}}_V)({\mathrm{id}}_V\otimes R).$$

In any braided monoidal category the braiding $c$ satisfies the braid form of this equation on $X\otimes Y\otimes Z$ (suppressing associators after strictification of 41.07.01):

$$(c_{Y,Z}\otimes {\mathrm{id}}_X)({\mathrm{id}}_Y\otimes c_{X,Z})(c_{X,Y}\otimes {\mathrm{id}}_Z)=({\mathrm{id}}_Z\otimes c_{X,Y})(c_{X,Z}\otimes {\mathrm{id}}_Y)({\mathrm{id}}_X\otimes c_{Y,Z}).$$

Definition (rigid, pivotal, ribbon). A monoidal category is rigid (or autonomous) when every object $X$ has a left dual $X^{\ast }$ with evaluation ${\mathrm{ev}}_X:X^{\ast }\otimes X\to I$ and coevaluation ${\mathrm{coev}}_X:I\to X\otimes X^{\ast }$ satisfying the two zigzag identities of 41.07.01. A pivotal (or sovereign) category is a rigid monoidal category equipped with a monoidal natural isomorphism ${\omega }_X:X\stackrel{\sim }{\to }{X}^{\ast \ast }$, the pivotal structure, which identifies left and right duals and lets left and right traces be compared. A twist (or balancing) on a braided category is a natural isomorphism ${\theta }_X:X\stackrel{\sim }{\to }X$ with ${\theta }_I={\mathrm{id}}_I$ and ${\theta }_{X\otimes Y}=({\theta }_X\otimes {\theta }_Y)\circ c_{Y,X}\circ c_{X,Y}$. A ribbon (or tortile) category is a braided rigid category with a twist satisfying the ribbon axiom ${\theta }_{X^{\ast }}=({\theta }_X{)}^{\ast }$ ^{[Turaev 1994]}.

Definition (categorical trace). In a rigid category the left and right categorical traces of an endomorphism $f:X\to X$ are the endomorphisms of the unit obtained by closing the dual diagrams,

$${\underline{\mathrm{tr}}}_L(f)={\mathrm{ev}}_X\circ ({\mathrm{id}}_{X^{\ast }}\otimes f)\circ {\widetilde{\mathrm{coev}}}_X,{\underline{\mathrm{tr}}}_R(f)={\widetilde{\mathrm{ev}}}_X\circ (f\otimes {\mathrm{id}}_{X^{\ast }})\circ {\mathrm{coev}}_X,$$

where the tilded maps use the dual data on the other side. In a pivotal category the pivotal structure forces ${\underline{\mathrm{tr}}}_L={\underline{\mathrm{tr}}}_R=:\underline{\mathrm{tr}}$; in a ribbon category the twist makes $\underline{\mathrm{tr}}$ cyclic, $\underline{\mathrm{tr}}(f\circ g)=\underline{\mathrm{tr}}(g\circ f)$ whenever both composites are defined, the property that turns the trace into a link invariant ^{[Kassel 1995]}.

Counterexamples to common slips Intermediate+

• Braided is strictly weaker than symmetric. A braiding need not satisfy $c_{Y,X}\circ c_{X,Y}=\mathrm{id}$; in $\mathbf{Rep}(U_q\mathfrak{g})$ the two composites differ by a scalar depending on the quantum parameter, which is the data the Reshetikhin-Turaev invariants detect. Imposing symmetry collapses $B_n$ onto $S_n$ and erases every link-invariant signal.

• A twist is not automatic from a braiding. A braided category admits many twists or none; the twist is extra structure recording a half-turn of each strand, and the ribbon axiom ${\theta }_{X^{\ast }}=({\theta }_X{)}^{\ast }$ ties the twist to duality. Dropping the twist leaves a braided rigid category in which the trace need not be cyclic, so closed diagrams can depend on a chosen framing.

• The Yang-Baxter equation is a theorem of the braided setting, not a separate axiom. It is the braid relation read off the threefold tensor, and it follows from the two hexagons plus naturality. Conversely, an $R$-matrix satisfying the equation builds a braided monoidal category of $R$-modules, so the equation and the braiding axiom are two readings of one condition.

Key theorem with proof Intermediate+

The signature result is the braided coherence theorem of Joyal and Street: in the free braided monoidal category, the only equations between braidings are the equations that hold in the braid group.

Theorem (Joyal-Street; braided coherence). Let $\mathcal{B}$ be the free braided monoidal category on one generating object $X$. Then $\mathcal{B}$ is strict monoidal; its objects are the natural numbers, with $n$ corresponding to $X^{\otimes n}$ and $0$ to $I$; the hom-set $\mathcal{B}(m,n)$ is empty for $m\ne n$; and $\mathcal{B}(n,n)$ is in bijection with the braid group $B_n$, composition being multiplication in $B_n$ and the tensor product being addition of strand numbers with juxtaposition of braids ^{[Joyal-Street 1993]}. Moreover, two formal diagrams built from the braiding, associator, and unitors in any braided monoidal category are equal precisely when the braids they depict are equal.

Proof. Construct $\mathcal{B}$ syntactically. Objects are generated by $X$ and $I$ under $\otimes$ modulo the strict monoidal relations, so every object is $X^{\otimes n}$ for a unique $n\in \mathbb{N}$. Morphisms are formal composites of instances of $c_{X,X}$ and identities, tensored and composed, modulo the relations expressing naturality of $c$, the two hexagons, and the pentagon and triangle of 41.07.01. On $X^{\otimes n}$ the generator $c$ acts on a pair of adjacent factors; write ${\sigma }_i$ for the braid crossing the $i$-th and $(i+1)$-st factors by $c$. The hexagon on three factors unwinds to the braid relation ${\sigma }_i{\sigma }_{i+1}{\sigma }_i={\sigma }_{i+1}{\sigma }_i{\sigma }_{i+1}$, and naturality of $c$ on disjoint pairs of factors gives ${\sigma }_i{\sigma }_j={\sigma }_j{\sigma }_i$ for $|i-j|\ge 2$. These are exactly the defining relations of $B_n$, so the evaluation map $e:\mathcal{B}(n,n)\to B_n$ sending ${\sigma }_i\mapsto {\sigma }_i$ is well-defined and surjective.

Injectivity of $e$ is the content of the theorem. Equip braid words with the Garside rewriting system, in which every word in the ${\sigma }_i^{\pm 1}$ admits a unique normal form as a power of the central half-twist $\Delta$ times a positive word drawn from a fixed set of divisors of $\Delta$. The Garside system is terminating and confluent, so by Newman's lemma every braid word has a unique normal form, and two words are equal in $B_n$ exactly when their normal forms coincide. The categorical relations of $\mathcal{B}$ mirror this rewriting step by step — the hexagon is the only three-strand relation and it is the Garside move in disguise — so two formal diagrams agree in $\mathcal{B}(n,n)$ exactly when their braid words have the same Garside normal form. Hence $e$ is injective, and $\mathcal{B}(n,n)\cong B_n$. The claim for general diagrams in an arbitrary braided monoidal category reduces to this case by strictification of 41.07.01: strictify the ambient monoidal structure, after which every formal diagram is a braid word, equal precisely when the underlying braids agree. $\square$

Bridge. This theorem builds toward the identification of categorical braidings with solutions of the Yang-Baxter equation and appears again in the Master tier, where the free ribbon category is recognised as the category of framed tangles and the Drinfeld centre supplies braided structure universally. The foundational reason the braid group and not the symmetric group appears is that the hexagon axiom is precisely the braid relation read off the threefold tensor, so the action of $B_n$ on $X^{\otimes n}$ is forced the moment the braiding is natural and hexagon-compatible; this is exactly the mechanism that turns a categorical braiding into a link invariant, and the bridge is that the single hexagon diagram simultaneously controls coherence, the braid group, and the geometry of knots.

Exercises Intermediate+

Advanced results Master

Theorem (Yang-Baxter from the hexagon). In any braided monoidal category the braiding satisfies the braid form of the Yang-Baxter equation on $X\otimes Y\otimes Z$, and this equation is precisely the braid relation ${\sigma }_1{\sigma }_2{\sigma }_1={\sigma }_2{\sigma }_1{\sigma }_2$ read off the threefold tensor ^{[Kassel 1995]}. Conversely, any invertible solution $R\in \mathrm{End}(V\otimes V)$ of the Yang-Baxter equation equips the induced category of $R$-modules with a braiding, so braided monoidal categories and representations of the Yang-Baxter equation are equivalent structures on three strands.

Theorem (Drinfeld centre; universal braiding). For any monoidal category $\mathcal{C}$ the Drinfeld centre $Z(\mathcal{C})$ is the category of pairs $(X,{\gamma }_X)$ with ${\gamma }_X$ a natural half-braiding $V\otimes X\stackrel{\sim }{\to }X\otimes V$; it is braided monoidal via $c_{(X,{\gamma }_X),(Y,{\gamma }_Y)}={\gamma }_X(Y)$, and if $\mathcal{C}$ is rigid then $Z(\mathcal{C})$ is rigid as well ^{[EGNO 2015]}. The centre is the universal recipient of a braiding for $\mathcal{C}$: braided monoidal functors $Z(\mathcal{C})\to \mathcal{D}$ correspond to monoidal functors $\mathcal{C}\to \mathcal{D}$, and when $\mathcal{C}$ is fusion the centre is a modular-like nondegenerate braided fusion category.

Theorem (Shum; the free ribbon category is tangles). The category ${\mathrm{Tang}}_{\mathrm{fr}}$ of framed tangles in the $3$-ball, modulo ambient isotopy fixing the boundary, is the free ribbon (tortile) category on one self-dual generator ^{[Shum 1994]}. Consequently every ribbon category $\mathcal{C}$ and self-dual object $X\in \mathcal{C}$ determine a unique-up-to-ribbon-isomorphism ribbon functor ${\mathrm{Tang}}_{\mathrm{fr}}\to \mathcal{C}$ sending the generator to $X$, and the categorical trace evaluates each closed framed tangle to an element of $\mathrm{End}(I)$. The graphical calculus of cups, caps, crossings, and twists is therefore a faithful syntax for ribbon categories, a theorem of Freyd and Yetter making the diagrammatic reasoning rigorous ^{[Freyd-Yetter 1989]}.

Theorem (Reshetikhin-Turaev and Turaev-Viro invariants). A modular tensor category — a ribbon fusion category whose braiding is nondegenerate, in that the Müger centre of transparent objects reduces to the unit — produces, by colouring the strands of a framed link diagram with simple objects and evaluating through the braiding, twist, duals, and the categorical trace, the Reshetikhin-Turaev invariant of the link and, by surgery along framed links, of every closed oriented $3$-manifold ^{[Reshetikhin-Turaev 1991]}. Dually, a spherical fusion category produces the Turaev-Viro state-sum invariant by labelling the edges of a triangulation and multiplying the quantum $6j$-symbols over tetrahedra ^{[Turaev-Viro 1992]}. These are the categorical engines of three-dimensional topological quantum field theory.

Synthesis. Putting these together, a braided monoidal category is a monoidal category whose coherence theorem is the braid group rather than the symmetric group, and the hexagon axiom is the foundational reason: it is exactly the braid relation on three strands, generalises Mac Lane's coherence from 41.07.01 to the setting where swaps carry over-and-under data, and is dual to the symmetric case in that the braid group $B_n$ surjects onto $S_n$ by imposing ${\sigma }_i^2=1$. This builds toward the ribbon and pivotal refinements, which appear again in the categorical trace and the Reshetikhin-Turaev construction: a modular tensor category is a nondegenerate ribbon fusion category, and the bridge is that colouring a framed link diagram by its simple objects and evaluating via the braiding, twist, and duals returns a number invariant under ambient isotopy — the categorical machine that converts the geometry of knots into algebra.

Full proof set Master

Proposition 1 (the Yang-Baxter equation follows from the hexagon). In any braided monoidal category, the braiding satisfies the braid form of the Yang-Baxter equation on $X\otimes Y\otimes Z$ (suppressing associators after strictification of 41.07.01).

Proof. Both sides of the displayed equation are morphisms $X\otimes Y\otimes Z\to Z\otimes Y\otimes X$ in the strictified category. Expand the left-hand side $(c_{Y,Z}\otimes {\mathrm{id}}_X)({\mathrm{id}}_Y\otimes c_{X,Z})(c_{X,Y}\otimes {\mathrm{id}}_Z)$ using the second hexagon to rewrite each $\mathrm{id}\otimes c$ and $c\otimes \mathrm{id}$ factor in terms of $c$ on the relevant tensor products; the expansion lands in a composite of three braidings arranged cyclically around the hexagon. Applying the first hexagon to the right-hand side $({\mathrm{id}}_Z\otimes c_{X,Y})(c_{X,Z}\otimes {\mathrm{id}}_Y)({\mathrm{id}}_X\otimes c_{Y,Z})$ produces the same cyclic composite, because the two hexagons are the two faces of the single coherence diagram for the triple tensor product. Hence the two sides agree, which on labelling the adjacent braidings ${\sigma }_1,{\sigma }_2$ is exactly ${\sigma }_1{\sigma }_2{\sigma }_1={\sigma }_2{\sigma }_1{\sigma }_2$. The argument uses only the hexagon axioms and naturality of $c$. $\square$

Proposition 2 (symmetry collapses the braid group to the symmetric group). In a symmetric monoidal category the braiding satisfies $c_{Y,X}\circ c_{X,Y}={\mathrm{id}}_{X\otimes Y}$, and the induced action of the braid group $B_n$ on $X^{\otimes n}$ factors through the symmetric group $S_n$.

Proof. The symmetry condition holds by the definition of a symmetric monoidal category. In the free braided category the generator ${\sigma }_i$ acts as the braiding on the $i$-th and $(i+1)$-st factors, so ${\sigma }_i^2$ acts as $c_{Y,X}\circ c_{X,Y}=\mathrm{id}$ on those factors, hence as the identity on $X^{\otimes n}$. The braid group with the additional relations ${\sigma }_i^2=1$ is exactly the presentation of the symmetric group $S_n$, so the action homomorphism $B_n\to \mathrm{Aut}(X^{\otimes n})$ factors through $B_n/⟨{\sigma }_i^2⟩\cong S_n$. Therefore every symmetric monoidal category is braided with the additional property that braids act through permutations. $\square$

Proposition 3 (cyclicity of the ribbon trace). In a ribbon category, the categorical trace is cyclic: for morphisms $f:X\to Y$ and $g:Y\to X$, $\underline{\mathrm{tr}}(f\circ g)=\underline{\mathrm{tr}}(g\circ f)$ as endomorphisms of $I$.

Proof. The trace of an endomorphism $h:X\to X$ is the closed composite $I\stackrel{{\mathrm{coev}}_X}{\to }{X}^{\ast }\otimes X\stackrel{\mathrm{id}\otimes h}{\to }{X}^{\ast }\otimes X\stackrel{{\mathrm{ev}}_X}{\to }I$, with the pivotal structure $\omega$ and the twist $\theta$ inserted so that the diagram is a framed circle in the graphical calculus of ^{[Selinger 2010]}. Expanding $\underline{\mathrm{tr}}(f\circ g)$ on $Y$ and $\underline{\mathrm{tr}}(g\circ f)$ on $X$ and using the zigzag identities to slide $f$ and $g$ along the strands, the two composites become the same closed framed diagram read in opposite orientations; naturality of the braiding gives $c\circ (f\otimes g)=(g\otimes f)\circ c$, and the ribbon axiom ${\theta }_{X^{\ast }}=({\theta }_X{)}^{\ast }$ identifies the twist contributions on the two dual strands, so the two framed circles are ambient-isotopic and evaluate to the same element of $\mathrm{End}(I)$. Hence $\underline{\mathrm{tr}}(f\circ g)=\underline{\mathrm{tr}}(g\circ f)$, the cyclicity that makes the trace a link invariant. $\square$

Connections Master

• Monoidal categories and Mac Lane coherence 41.07.01. This unit is a direct continuation of 41.07.01: a braided monoidal category is a monoidal category with a braiding subject to the hexagons, and every argument here strictifies the underlying monoidal structure using the coherence and strictification theorems of 41.07.01. The pentagon and triangle of 41.07.01 persist underneath the two hexagons, and the identification $B_n\twoheadrightarrow S_n$ of Proposition 2 measures exactly the extra data a braiding carries beyond the symmetric monoidal structure of the sibling.

• Natural transformations and functor categories 41.01.02. The braiding $c_{X,Y}$, the half-braidings ${\gamma }_X$, the pivotal structure ${\omega }_X$, and the twist ${\theta }_X$ are all natural isomorphisms in the sense built in 41.01.02, and the hexagon axioms are equations between pasted naturality squares. The Drinfeld centre $Z(\mathcal{C})$ is itself a functor-category-style construction whose objects carry a natural family ${\gamma }_X$, so the rigidity of the braided theory rests on the naturality language of 41.01.02.

• Adjunctions and the duality adjunction 41.03.01. A left dual $X^{\ast }$ with evaluation and coevaluation is exactly an adjunction $X⊣X^{\ast }$ internal to the monoidal structure, with ${\mathrm{coev}}_X$ as unit and ${\mathrm{ev}}_X$ as counit; the zigzag identities are the triangle identities of 41.03.01. The categorical trace is the composite of this internal adjunction's unit and counit, so the ribbon trace and the Reshetikhin-Turaev invariants are applications of adjoint-functor theory to the self-dual objects of a braided category.

• Monads and the distribution law 41.05.01. The braid relation ${\sigma }_i{\sigma }_{i+1}{\sigma }_i={\sigma }_{i+1}{\sigma }_i{\sigma }_{i+1}$ is the categorical analogue of a distributive law between monads of 41.05.01: a braiding on a category of monoids lets two monoid structures commute up to a coherent natural isomorphism, exactly as a distributive law composes two monads. The Drinfeld centre construction is the braided echo of the Eilenberg-Moore setting of 41.05.01.

• Ends, coends, and the graphical calculus 41.06.01. The categorical trace and the Reshetikhin-Turaev invariant are compactly expressed as coends over the ambient category, and the Turaev-Viro state rewrites the partition function as a coend of $6j$-symbols; the coend calculus of 41.06.01 is the algebraic shadow of the graphical string-diagram reasoning that the Freyd-Yetter theorem makes rigorous for ribbon categories.

Historical & philosophical context Master

The braided refinement of a monoidal category was introduced by André Joyal and Ross Street in Braided tensor categories (circulated 1986, published 1993), where the hexagon axioms, the braided coherence theorem identifying endomorphisms of the free braided monoidal category with the braid groups $B_n$, and the connection to the Yang-Baxter equation first appear ^{[Joyal-Street 1993]}. The Yang-Baxter equation itself predates category theory: C. N. Yang wrote it down in 1967 for the delta-function Bose gas, and R. J. Baxter rediscovered it in 1971-1972 as the star-triangle relation solving two-dimensional lattice models in statistical mechanics. The categorical recognition that a solution of this equation is the same object as a braiding on a tensor category unified the two threads.

The subject crystallised around Vladimir Drinfeld's and M. Jimbo's quantum groups (1985-1986), whose representation categories are braided but not symmetric; the non-symmetric braiding is the source of the link and $3$-manifold invariants constructed by Reshetikhin and Turaev in 1990-1991 from modular tensor categories ^{[Reshetikhin-Turaev 1991]}, and of the state-sum invariants of Turaev and Viro from spherical fusion categories ^{[Turaev-Viro 1992]}. The pivotal and ribbon (tortile) axioms were isolated by Freyd, Yetter, Shum, and Turaev in 1989-1994: Freyd and Yetter made the graphical calculus of cups, caps, crossings, and twists into a rigorous coherence theorem ^{[Freyd-Yetter 1989]}, and Shum proved that the category of framed tangles is the free ribbon category on one object ^{[Shum 1994]}. Turaev's monograph Quantum Invariants of Knots and 3-Manifolds (1994; revised 2016) consolidated the ribbon-category machinery as the standard language of three-dimensional topological quantum field theory ^{[Turaev 1994]}.

Bibliography Master

1. Mac Lane, S. — Categories for the Working Mathematician, 2nd ed., Springer GTM 5, 1998, Ch. XI §1-3.
2. Joyal, A. and Street, R. — Braided tensor categories, Advances in Mathematics 102, 1993.
3. Turaev, V. — Quantum Invariants of Knots and 3-Manifolds, de Gruyter Studies in Mathematics 18, 1994 (revised 2016).
4. Kassel, C. — Quantum Groups, Springer GTM 155, 1995.
5. Etingof, P., Gelaki, S., Nikshych, D., and Ostrik, V. — Tensor Categories, AMS Mathematical Surveys and Monographs 205, 2015.
6. Reshetikhin, N. and Turaev, V. — Invariants of 3-manifolds via link polynomials and quantum groups, Inventiones Mathematicae 103, 1991.
7. Turaev, V. and Viro, O. — State sum invariants of 3-manifolds and quantum 6j-symbols, Topology 31, 1992.
8. Shum, M.-C. — Tortile yang-baxter category, coherence of the category of tangles, J. Pure Appl. Algebra 90, 1994.
9. Freyd, P. and Yetter, D. — Coherence theorems via categorical diagrams, J. Pure Appl. Algebra 60, 1989.

@book{MacLane1998,
  author    = {Mac Lane, Saunders},
  title     = {Categories for the Working Mathematician},
  edition   = {2},
  publisher = {Springer},
  series    = {Graduate Texts in Mathematics 5},
  year      = {1998}
}

@article{JoyalStreet1993,
  author  = {Joyal, Andr\'e and Street, Ross},
  title   = {Braided tensor categories},
  journal = {Advances in Mathematics},
  volume  = {102},
  number  = {1},
  year    = {1993},
  pages   = {20--78}
}

@book{Turaev1994,
  author    = {Turaev, Vladimir},
  title     = {Quantum Invariants of Knots and 3-Manifolds},
  publisher = {Walter de Gruyter},
  series    = {de Gruyter Studies in Mathematics 18},
  year      = {1994}
}

@book{Kassel1995,
  author    = {Kassel, Christian},
  title     = {Quantum Groups},
  publisher = {Springer},
  series    = {Graduate Texts in Mathematics 155},
  year      = {1995}
}

@book{EGNO2015,
  author    = {Etingof, Pavel and Gelaki, Shlomo and Nikshych, Dmitri and Ostrik, Victor},
  title     = {Tensor Categories},
  publisher = {American Mathematical Society},
  series    = {Mathematical Surveys and Monographs 205},
  year      = {2015}
}

@article{ReshetikhinTuraev1991,
  author  = {Reshetikhin, Nicolai and Turaev, Vladimir},
  title   = {Invariants of 3-manifolds via link polynomials and quantum groups},
  journal = {Inventiones Mathematicae},
  volume  = {103},
  year    = {1991},
  pages   = {547--597}
}

@article{TuraevViro1992,
  author  = {Turaev, Vladimir and Viro, Oleg},
  title   = {State sum invariants of 3-manifolds and quantum 6j-symbols},
  journal = {Topology},
  volume  = {31},
  year    = {1992},
  pages   = {865--902}
}

@article{Shum1994,
  author  = {Shum, Mei-CheE},
  title   = {Tortile yang-baxter category, coherence of the category of tangles},
  journal = {Journal of Pure and Applied Algebra},
  volume  = {90},
  year    = {1994}
}

@article{FreydYetter1989,
  author  = {Freyd, Peter and Yetter, David},
  title   = {Coherence theorems via categorical diagrams},
  journal = {Journal of Pure and Applied Algebra},
  volume  = {60},
  year    = {1989}
}