ScalingStacks

3. The Center Construction

Let (π’ž,βŠ—,1)(\mathcal{C},\otimes,1) be a semistrict monoidal 2-category. The center 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}) would be easy to construct if we had a properly functioning theory of semistrict weak 4-categories. As it stands, all we can do is use our limited insight into 4-categories to guess the right answer, and then try to justify it by proving that we obtain a braided monoidal 2-category with good properties. We proceed in several stages. First we describe 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}) as a 2-category. Then we describe the monoidal structure, and then the braiding.

3.1. 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}) as a 2-Category

As noted in SectionΒ 1.1, the center construction applied to a monoid yields its usual center, because a certain square must commute. However, as one would expect from the weakening principle, when π’ž\mathcal{C} is a monoidal category the corresponding square need only commute up to a specified natural isomorphism. An object of 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}) thus turns out to be an object Aβˆˆπ’žA\in\mathcal{C} equipped with a natural isomorphism RA,βˆ’:AβŠ—βˆ’β‡’βˆ’βŠ—AR_{A,-}\colon A\otimes-\Rightarrow-\otimes A satisfying various coherence laws, such as the commutativity of following diagram:

AβŠ—XβŠ—Y{\lx@inpgf@ignorespaces A\otimes X\otimes Y}XβŠ—YβŠ—A{\lx@inpgf@ignorespaces X\otimes Y\otimes A}XβŠ—AβŠ—Y{\lx@inpgf@ignorespaces X\otimes A\otimes Y}RA,XβŠ—Y\scriptstyle{\lx@inpgf@ignorespaces R_{A,X\otimes Y}}RA,XβŠ—Y\scriptstyle{\lx@inpgf@ignorespaces R_{A,X}\otimes Y}XβŠ—RA,Y\scriptstyle{\lx@inpgf@ignorespaces X\otimes R_{A,Y}}

for any objects X,Yβˆˆπ’žX,Y\in\mathcal{C}. Of course, this diagram is part of the definition of a braided monoidal category. Similarly, the morphisms in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}) also work out to have properties that form part of the definition of a braided monoidal category.

Heuristic 4-categorical computations suggest how these patterns should continue when π’ž\mathcal{C} is a monoidal 2-category. We thus define 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}) as follows.

Objects in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}):

An object of 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}) is a triple (A,RA,βˆ’,R~(A|βˆ’,βˆ’))(A,R_{A,-},\tilde{R}_{(A|-,-)}) consisting of:

  1. (1)

    an object Aβˆˆπ’žA\in\mathcal{C}

  2. (2)

    a pseudonatural equivalence RA,βˆ’:AβŠ—βˆ’β‡’βˆ’βŠ—AR_{A,-}\colon A\otimes-\Rightarrow-\otimes A

  3. (3)

    an invertible modification R~(A|βˆ’,βˆ’)\tilde{R}_{(A|-,-)}, giving for any objects X,Yβˆˆπ’žX,Y\in\mathcal{C} a 2-isomorphism

    AβŠ—XβŠ—Y{\lx@inpgf@ignorespaces A\otimes X\otimes Y}  XβŠ—YβŠ—A{\lx@inpgf@ignorespaces X\otimes Y\otimes A}XβŠ—AβŠ—Y{\lx@inpgf@ignorespaces X\otimes A\otimes Y}RA,XβŠ—Y\scriptstyle{\lx@inpgf@ignorespaces R_{A,X\otimes Y}}RA,XβŠ—Y\scriptstyle{\lx@inpgf@ignorespaces R_{A,X}\otimes Y}XβŠ—RA,Y\scriptstyle{\lx@inpgf@ignorespaces X\otimes R_{A,Y}}⇑R~(A|X,Y){\lx@inpgf@ignorespaces\Uparrow\tilde{R}_{(A|X,Y)}}

such that for any objects X,Y,Zβˆˆπ’žX,Y,Z\in\mathcal{C}, the tetrahedron (βˆ™βŠ—(βˆ™βŠ—βˆ™βŠ—βˆ™))(\bullet\otimes(\bullet\otimes\bullet\otimes\bullet)) commutes.

Here we mean that the diagram (βˆ™βŠ—(βˆ™βŠ—βˆ™βŠ—βˆ™))(\bullet\otimes(\bullet\otimes\bullet\otimes\bullet)) commutes with objects A,X,Y,ZA,X,Y,Z, and with the modification R~(βˆ’|βˆ’,βˆ’)\tilde{R}_{(-|-,-)} in the definition of a braided monoidal 2-category replaced by the above R~(A|βˆ’,βˆ’)\tilde{R}_{(A|-,-)}. Throughout the following we use the hieroglyphical notation in this way. Also, we use letters near the beginning of the alphabet to denote objects of π’ž\mathcal{C} underlying objects in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}), and letters near the end to denote objects of π’ž\mathcal{C} being used as such.

0N87

Remark 9. The fact that RA,βˆ’R_{A,-} is a pseudonatural equivalence can be expressed equivalently as follows: for any object Xβˆˆπ’žX\in\mathcal{C}, there exists an equivalence RA,X:AβŠ—Xβ†’XβŠ—AR_{A,X}\colon A\otimes X\to X\otimes A, and for any morphism f:Xβ†’Yf\colon X\to Y in π’ž\mathcal{C}, there exists a 2-isomorphism RA,f:(AβŠ—f)∘RA,Yβ‡’RA,X∘(fβŠ—A)R_{A,f}\colon(A\otimes f)\circ R_{A,Y}\Rightarrow R_{A,X}\circ(f\otimes A):

AβŠ—X{\lx@inpgf@ignorespaces A\otimes X}XβŠ—A{\lx@inpgf@ignorespaces X\otimes A}AβŠ—Y{\lx@inpgf@ignorespaces A\otimes Y}YβŠ—A{\lx@inpgf@ignorespaces Y\otimes A}RA,X\scriptstyle{\lx@inpgf@ignorespaces R_{A,X}}AβŠ—f\scriptstyle{\lx@inpgf@ignorespaces A\otimes f}⇑RA,f{\lx@inpgf@ignorespaces\Uparrow R_{A,f}}fβŠ—A\scriptstyle{\lx@inpgf@ignorespaces f\otimes A}RA,Y\scriptstyle{\lx@inpgf@ignorespaces R_{A,Y}}

such that (βˆ™βŠ—β†’β†’)(\bullet\otimes{\to}\to) and (βˆ™βŠ—β‡“)(\bullet\otimes{\Downarrow}) commute.

Similarly, the fact that R~(A|βˆ’,βˆ’)\>\tilde{R}_{(A|-,-)} is a modification means that the diagrams (βˆ™βŠ—(β†’βŠ—βˆ™))(\bullet\otimes({\to}\otimes\bullet)) and (βˆ™βŠ—(βˆ™βŠ—β†’))(\bullet\otimes(\bullet\otimes{\to})) commute.

Morphisms in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}):

A morphism in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}) from (A,RA,βˆ’,R~(A|βˆ’,βˆ’))(A,R_{A,-},\tilde{R}_{(A|-,-)}) to (B,RB,βˆ’,R~(B|βˆ’,βˆ’))(B,R_{B,-},\tilde{R}_{(B|-,-)}) is a pair (f,Rf,βˆ’)(f,R_{f,-}) consisting of:

  1. (1)

    a morphism f:A→Bf\colon A\to B

  2. (2)

    an invertible modification Rf,βˆ’R_{f,-}, giving for any object Xβˆˆπ’žX\in\mathcal{C} a 2-isomorphism

    AβŠ—X{\lx@inpgf@ignorespaces A\otimes X}BβŠ—X{\lx@inpgf@ignorespaces B\otimes X}XβŠ—A{\lx@inpgf@ignorespaces X\otimes A}XβŠ—B{\lx@inpgf@ignorespaces X\otimes B}fβŠ—X\scriptstyle{\lx@inpgf@ignorespaces f\otimes X}RA,X\scriptstyle{\lx@inpgf@ignorespaces R_{A,X}}⇓Rf,X{\lx@inpgf@ignorespaces\Downarrow R_{f,X}}RB,X\scriptstyle{\lx@inpgf@ignorespaces R_{B,X}}XβŠ—f\scriptstyle{\lx@inpgf@ignorespaces X\otimes f}

such that the prism (β†’βŠ—(βˆ™βŠ—βˆ™))({\to}\otimes(\bullet\otimes\bullet)) commutes.

0N88

Remark 10. The fact that Rf,βˆ’R_{f,-} is a modification can be expressed equivalently by saying that (β†’βŠ—β†’)({\to}\otimes{\to}) commutes. (Note that (fβŠ—βˆ’)RB,βˆ’(f\otimes-)R_{B,-} and RA,βˆ’(βˆ’βŠ—f)R_{A,-}(-\otimes f) are pseudonatural transformations in an obvious way.)

2-Morphisms in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}):

A 2-morphism Ξ±\alpha in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}) from (f,Rf,βˆ’)(f,R_{f,-}) to (g,Rg,βˆ’)(g,R_{g,-}) is

  1. (1)

    a 2-morphism Ξ±:fβ‡’g\alpha\colon f\Rightarrow g in π’ž\mathcal{C}

such that (β‡“βŠ—βˆ™)({\Downarrow}\otimes\bullet) commutes.

We define the composition operations in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}) as follows. Composition of morphisms is defined by:

(f,Rf,βˆ’)∘(g,Rg,βˆ’):=(f∘g,((fβŠ—βˆ’)∘Rg,βˆ’)β‹…(Rf,βˆ’βˆ˜(βˆ’βŠ—g)))(f,R_{f,-})\circ(g,R_{g,-}):=(f\circ g,((f\otimes-)\circ R_{g,-})\cdot(R_{f,-}\circ(-\otimes g)))

where f:Aβ†’Bf\colon A\to B and g:Bβ†’Cg\colon B\to C are the underlying 1-morphisms in π’ž\mathcal{C}. Note that for any object Xβˆˆπ’žX\in\mathcal{C}, the 2-morphism ((fβŠ—X)∘Rg,X)β‹…(Rf,X∘(XβŠ—g))((f\otimes X)\circ R_{g,X})\cdot(R_{f,X}\circ(X\otimes g)) equals the back of the following diagram:

A​X{\lx@inpgf@ignorespaces AX}C​X{\lx@inpgf@ignorespaces CX}B​X{\lx@inpgf@ignorespaces BX} X​A{\lx@inpgf@ignorespaces XA}X​C{\lx@inpgf@ignorespaces XC}X​B{\lx@inpgf@ignorespaces XB} (f∘g)βŠ—X\scriptstyle{\lx@inpgf@ignorespaces(f\circ g)\otimes X}RA,X\scriptstyle{\lx@inpgf@ignorespaces R_{A,X}}fβŠ—X\scriptstyle{\lx@inpgf@ignorespaces f\otimes X}⇓Rf,X{\lx@inpgf@ignorespaces\Downarrow R_{f,X}}⇑id{\lx@inpgf@ignorespaces\Uparrow\textrm{id}}RC,X\scriptstyle{\lx@inpgf@ignorespaces R_{C,X}}gβŠ—X\scriptstyle{\lx@inpgf@ignorespaces g\otimes X}⇓Rg,X{\lx@inpgf@ignorespaces\Downarrow R_{g,X}}XβŠ—f\scriptstyle{\lx@inpgf@ignorespaces X\otimes f}⇑id{\lx@inpgf@ignorespaces\Uparrow\textrm{id}}XβŠ—g\scriptstyle{\lx@inpgf@ignorespaces X\otimes g}
0N89

Remark 11. Eventually this will imply that the braiding in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}) satisfies (β†’β†’βŠ—βˆ™)(\to{\to}\otimes\bullet).

To show that the composite of morphisms in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}) is again a morphism, we have to check that (β†’βŠ—β†’)({\to}\otimes{\to}) and (β†’βŠ—(βˆ™βŠ—βˆ™))({\to}\otimes(\bullet\otimes\bullet)) commute. These can be seen by pasting together two diagrams of the form (β†’βŠ—β†’)({\to}\otimes{\to}) and (β†’βŠ—(βˆ™βŠ—βˆ™))({\to}\otimes(\bullet\otimes\bullet)), respectively.

Vertical and horizontal composition of 2-morphisms is defined the same as in π’ž\mathcal{C}; one can check that these composites again satisfy (β‡“βŠ—βˆ™)({\Downarrow}\otimes\bullet) by pasting together two diagrams of this form.

3.2. The Monoidal Structure

We have to show that 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}) bears a monoidal structure (𝒡(π’ž),βŠ—π’΅β‘(π’ž),I)(\mathcal{Z}(\mathcal{C}),\otimes_{\mathcal{Z}(\mathcal{C})},I), such that all the requirements for a monoidal category given in Definition 4 are satisfied.

(Ad 4.1): The object Iβˆˆπ’΅β‘(π’ž)I\in\mathcal{Z}(\mathcal{C}) is (I,1βˆ’,11(βˆ’βŠ—βˆ’))(I,1_{-},1_{1_{(-\otimes-)}}).

The tensor product of objects: (Ad 4.2): The tensor product of two objects (A,RA,βˆ’,R~(A|βˆ’,βˆ’))βŠ—π’΅β‘(π’ž)(B,RB,βˆ’,R~(B|βˆ’,βˆ’))(A,R_{A,-},\tilde{R}_{(A|-,-)})\otimes_{\mathcal{Z}(\mathcal{C})}(B,R_{B,-},\tilde{R}_{(B|-,-)}) is defined to be the triple (AβŠ—B,(RAβŠ—RB)βˆ’,(R~AβŠ—R~B)(βˆ’,βˆ’))(A\otimes B,(R_{A}\otimes R_{B})_{-},(\tilde{R}_{A}\otimes\tilde{R}_{B})_{(-,-)}), where:

  1. (1)

    The underlying π’ž\mathcal{C}-object is the tensor product AβŠ—BA\otimes B in π’ž\mathcal{C}.

  2. (2)

    By RemarkΒ (9), the underlying pseudonatural equivalence (RAβŠ—RB)βˆ’:(AβŠ—B)βŠ—βˆ’β‡’βˆ’βŠ—(AβŠ—B)(R_{A}\otimes R_{B})_{-}:(A\otimes B)\otimes-\Rightarrow-\otimes(A\otimes B) assigns a 1-morphism (RAβŠ—RB)X(R_{A}\otimes R_{B})_{X} to any object Xβˆˆπ’žX\in\mathcal{C} and a 2-morphism (RAβŠ—RB)f(R_{A}\otimes R_{B})_{f} to any 1-morphism f:Xβ†’Yf\colon X\to Y. These are given as follows:

    (RAβŠ—RB)X=(AβŠ—RB,X)​(RA,XβŠ—B),(R_{A}\otimes R_{B})_{X}=(A\otimes R_{B,X})(R_{A,X}\otimes B),
    OPEN(RAβŠ—RB)f=((AβŠ—RB,f)∘(RA,YβŠ—B))β‹…(AβŠ—RB,X)∘(RA,fβŠ—B)),(R_{A}\otimes R_{B})_{f}=((A\otimes R_{B,f})\circ(R_{A,Y}\otimes B))\cdot(A\otimes R_{B,X})\circ(R_{A,f}\otimes B)),

    or in terms of a diagram:

    A​B​X{\lx@inpgf@ignorespaces ABX}A​X​B{\lx@inpgf@ignorespaces AXB}X​A​B{\lx@inpgf@ignorespaces XAB}A​B​Y{\lx@inpgf@ignorespaces ABY}A​Y​B{\lx@inpgf@ignorespaces AYB}Y​A​B{\lx@inpgf@ignorespaces YAB}A​BβŠ—f\scriptstyle{\lx@inpgf@ignorespaces AB\otimes f}AβŠ—RB,X\scriptstyle{\lx@inpgf@ignorespaces A\otimes R_{B,X}}⇑AβŠ—RB,f{\lx@inpgf@ignorespaces\Uparrow A\otimes R_{B,f}}RA,XβŠ—B\scriptstyle{\lx@inpgf@ignorespaces R_{A,X}\otimes B}⇑RA,fβŠ—B{\lx@inpgf@ignorespaces\Uparrow R_{A,f}\otimes B}X​AβŠ—f\scriptstyle{\lx@inpgf@ignorespaces XA\otimes f}AβŠ—RB,Y\scriptstyle{\lx@inpgf@ignorespaces A\otimes R_{B,Y}}RA,YβŠ—B\scriptstyle{\lx@inpgf@ignorespaces R_{A,Y}\otimes B}
    0N8A

    Remark 12. This will imply that the braiding in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}) satisfies ((βˆ™βŠ—βˆ™)βŠ—β†’)((\bullet\otimes\bullet)\otimes{\to}).

    To show that these data constitute a pseudonatural equivalence, we have to show that (βˆ™βŠ—β†’β†’)(\bullet\otimes{\to}\to) and (βˆ™βŠ—β‡“)(\bullet\otimes{\Downarrow}) hold. This can be done easily by pasting together the corresponding diagrams for RA,βˆ’R_{A,-} and RB,βˆ’R_{B,-}.

  3. (3)

    The underlying modification (R~AβŠ—R~B)(X,Y):(AβŠ—RB,XβŠ—Y)​(RA,XβŠ—BβŠ—Y)​(XβŠ—AβŠ—RB,Y)​(XβŠ—RA,YβŠ—B)β‡’(AβŠ—RB,XβŠ—Y)​(RA,XβŠ—YβŠ—B)(\tilde{R}_{A}\otimes\tilde{R}_{B})_{(X,Y)}:(A\otimes R_{B,X}\otimes Y)(R_{A,X}\otimes B\otimes Y)(X\otimes A\otimes R_{B,Y})(X\otimes R_{A,Y}\otimes B)\Rightarrow(A\otimes R_{B,X\otimes Y})(R_{A,X\otimes Y}\otimes B) is defined to be the pasting:

    A​B​X​Y{\lx@inpgf@ignorespaces ABXY}A​X​Y​B{\lx@inpgf@ignorespaces AXYB}A​X​B​Y{\lx@inpgf@ignorespaces AXBY}    X​Y​A​B{\lx@inpgf@ignorespaces XYAB}X​A​Y​B{\lx@inpgf@ignorespaces XAYB}X​A​B​Y{\lx@inpgf@ignorespaces XABY}AβŠ—RB,X​Y\scriptstyle{\lx@inpgf@ignorespaces A\otimes R_{B,XY}}⇑AβŠ—R~(B|X,Y){\lx@inpgf@ignorespaces\Uparrow A\otimes\tilde{R}_{(B|X,Y)}}RA,X​YβŠ—B\scriptstyle{\lx@inpgf@ignorespaces R_{A,XY}\otimes B}β‡‘βŠ—RA,X,RB,Yβˆ’1{\lx@inpgf@ignorespaces\Uparrow\otimes_{R_{A,X},R_{B,Y}}^{-1}}⇑R~(A|X,Y)βŠ—B{\lx@inpgf@ignorespaces\Uparrow\tilde{R}_{(A|X,Y)}\otimes B}

    Again it is easy to verify that this satisfies (βˆ™βŠ—(β†’βŠ—βˆ™))(\bullet\otimes({\to}\otimes\bullet)) and (βˆ™βŠ—(βˆ™βŠ—β†’))(\bullet\otimes(\bullet\otimes{\to})) and hence is a modification.

To show that this definition gives in fact an object in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}), we have to verify that (βˆ™βŠ—(βˆ™βŠ—βˆ™βŠ—βˆ™))(\bullet\otimes(\bullet\otimes\bullet\otimes\bullet)) is satisfied. The following picture shows the tetrahedron. Those vertices in the picture that are vertices of the tetrahedron are written in big capitals. The remaining vertices occur since they are needed for the decomposition.

X​Y​Z​A​B{\lx@inpgf@ignorespaces XYZAB}  A​X​Y​Z​B{\lx@inpgf@ignorespaces\scriptstyle{AXYZB}}X​Y​A​Z​B{\lx@inpgf@ignorespaces\scriptstyle{XYAZB}}  X​A​Y​Z​B{\lx@inpgf@ignorespaces\scriptstyle{XAYZB}}A​B​X​Y​Z{\lx@inpgf@ignorespaces ABXYZ}A​X​Y​B​Z{\lx@inpgf@ignorespaces\scriptstyle{AXYBZ}}X​Y​A​B​Z{\lx@inpgf@ignorespaces XYABZ}A​X​B​Y​Z{\lx@inpgf@ignorespaces\scriptstyle{AXBYZ}}X​A​Y​B​Z{\lx@inpgf@ignorespaces\scriptstyle{XAYBZ}}X​A​B​Y​Z{\lx@inpgf@ignorespaces XABYZ}

The following picture gives a decomposition of the tetrahedron into four smaller commutative diagrams.

A​X​Y​Z​BAXYZBX​A​Y​Z​BXAYZBX​Y​A​Z​BXYAZBX​Y​Z​A​BXYZABA​B​X​Y​ZABXYZA​X​B​Y​ZAXBYZA​X​Y​B​ZAXYBZA​X​Y​Z​BAXYZBA​X​B​Y​ZAXBYZX​A​B​Y​ZXABYZA​X​Y​B​ZAXYBZX​A​Y​B​ZXAYBZA​X​Y​Z​BAXYZBX​A​Y​Z​BXAYZBA​X​Y​B​ZAXYBZX​A​Y​B​ZXAYBZA​X​Y​Z​BAXYZBX​A​Y​Z​BXAYZBX​Y​A​B​ZXYABZX​Y​A​Z​BXYAZB

Two of them are tetrahedra of the form (βˆ™βŠ—(βˆ™βŠ—βˆ™βŠ—βˆ™))(\bullet\otimes(\bullet\otimes\bullet\otimes\bullet)), tensored by an object from the left and the right, respectively. The upper of the two triangular prisms commutes by the axioms 4.(v​i​i)4.(vii) with Ξ±=R~(A|X,Y)\alpha=\tilde{R}_{(A|X,Y)} and g=RZ,Bg=R_{Z,B}, together with 4.(v​i​i​i)4.(viii). The lower commutes by 4.(v​i)4.(vi), applied to Ξ²=R~(B|Y,Z)\beta=\tilde{R}_{(B|Y,Z)} and f=RA,Xf=R_{A,X} together with 4.(v​i​i​i)4.(viii). One can verify that this tensor product is in fact associative.

We shall often write (AβŠ—B,RAβŠ—RB,R~AβŠ—R~B)(A\otimes B,R_{A}\otimes R_{B},\tilde{R}_{A}\otimes\tilde{R}_{B}) as a shorthand symbol for the tensor product of objects in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}).

The tensor product of an object and a morphism:

(Ad 4.3): Let (f,Rf,βˆ’):(A,RA,βˆ’,R~(A|βˆ’,βˆ’))β†’(Aβ€²,RAβ€²,βˆ’,R~(Aβ€²|βˆ’,βˆ’))(f,R_{f,-})\colon(A,R_{A,-},\tilde{R}_{(A|-,-)})\to(A^{\prime},R_{A^{\prime},-},\tilde{R}_{(A^{\prime}|-,-)}) be a morphism in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}) and let (B,RB,βˆ’,R~(B|βˆ’,βˆ’))(B,R_{B,-},\tilde{R}_{(B|-,-)}) be an object. Their tensor product is the morphism given by the pair

(fβŠ—B,(βŠ—(f,RB,βˆ’)∘(RAβ€²,βˆ’βŠ—B))β‹…((AβŠ—RB,βˆ’)∘(Rf,βˆ’βŠ—B))),(f\otimes B,(\otimes_{(f,R_{B,-})}\circ(R_{A^{\prime},-}\otimes B))\cdot((A\otimes R_{B,-})\circ(R_{f,-}\otimes B))),

or in terms of a diagram:

A​B​X{\lx@inpgf@ignorespaces ABX}A​X​B{\lx@inpgf@ignorespaces AXB}X​A​B{\lx@inpgf@ignorespaces XAB}A′​B​X{\lx@inpgf@ignorespaces A^{\prime}BX}A′​X​B{\lx@inpgf@ignorespaces A^{\prime}XB}X​A′​B{\lx@inpgf@ignorespaces XA^{\prime}B}fβŠ—B​X\scriptstyle{\lx@inpgf@ignorespaces f\otimes BX}AβŠ—RB,X\scriptstyle{\lx@inpgf@ignorespaces A\otimes R_{B,X}}β‡‘βŠ—f,RB,X{\lx@inpgf@ignorespaces\Uparrow\otimes_{f,R_{B,X}}}RA,XβŠ—B\scriptstyle{\lx@inpgf@ignorespaces R_{A,X}\otimes B}⇑Rf,XβŠ—B{\lx@inpgf@ignorespaces\Uparrow R_{f,X}\otimes B}XβŠ—fβŠ—B\scriptstyle{\lx@inpgf@ignorespaces X\otimes f\otimes B}Aβ€²βŠ—RB,X\scriptstyle{\lx@inpgf@ignorespaces A^{\prime}\otimes R_{B,X}}RAβ€²,XβŠ—B\scriptstyle{\lx@inpgf@ignorespaces R_{A^{\prime},X}\otimes B}
0N8B

Remark 13. This will imply that the braiding in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}) satisfies ((β†’βŠ—βˆ™)βŠ—βˆ™)(({\to}\otimes\bullet)\otimes\bullet).

(Ad 4.4) Let (f,Rf,βˆ’):(B,RB,βˆ’,R~(B|βˆ’,βˆ’))β†’(Bβ€²,RBβ€²,βˆ’,R~(Bβ€²|βˆ’,βˆ’))(f,R_{f,-}):(B,R_{B,-},\tilde{R}_{(B|-,-)})\to(B^{\prime},R_{B^{\prime},-},\tilde{R}_{(B^{\prime}|-,-)}) be a morphism in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}) and let (A,RA,βˆ’,R~(A|βˆ’,βˆ’))(A,R_{A,-},\tilde{R}_{(A|-,-)}) be an object. Their tensor product is the pair

(AβŠ—f,((AβŠ—Rf,βˆ’)∘(RA,βˆ’βŠ—Bβ€²))β‹…((AβŠ—RB,βˆ’)βˆ˜βŠ—RA,βˆ’,f)),(A\otimes f,((A\otimes R_{f,-})\circ(R_{A,-}\otimes B^{\prime}))\cdot((A\otimes R_{B,-})\circ\otimes_{R_{A,-},f})),

or in terms of a diagram:

A​B​X{\lx@inpgf@ignorespaces ABX}A​X​B{\lx@inpgf@ignorespaces AXB}X​A​B{\lx@inpgf@ignorespaces XAB}A​B′​X{\lx@inpgf@ignorespaces AB^{\prime}X}A​X​Bβ€²{\lx@inpgf@ignorespaces AXB^{\prime}}X​A​Bβ€²{\lx@inpgf@ignorespaces XAB^{\prime}}AβŠ—fβŠ—X\scriptstyle{\lx@inpgf@ignorespaces A\otimes f\otimes X}A​RB,X\scriptstyle{\lx@inpgf@ignorespaces AR_{B,X}}⇑AβŠ—Rf,X{\lx@inpgf@ignorespaces\Uparrow A\otimes R_{f,X}}RA,X​B\scriptstyle{\lx@inpgf@ignorespaces R_{A,X}B}β‡‘βŠ—RA,X,f{\lx@inpgf@ignorespaces\Uparrow\otimes_{R_{A,X},f}}X​AβŠ—f\scriptstyle{\lx@inpgf@ignorespaces XA\otimes f}A​RBβ€²,X\scriptstyle{\lx@inpgf@ignorespaces AR_{B^{\prime},X}}RA,X​Bβ€²\scriptstyle{\lx@inpgf@ignorespaces R_{A,X}B^{\prime}}
0N8C

Remark 14. This will imply that the braiding in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}) satisfies ((βˆ™βŠ—β†’)βŠ—βˆ™)((\bullet\otimes{\to})\otimes\bullet).

To verify that these formulas really define morphisms in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}), one must check that (β†’βŠ—β†’)({\to}\otimes{\to}) and (β†’βŠ—(βˆ™βŠ—βˆ™))({\to}\otimes(\bullet\otimes\bullet)) hold. We only do this for 4.44.4; the other case being similar. To show (β†’βŠ—β†’)({\to}\otimes{\to}) one pastes together two cubes, one being the (β†’βŠ—β†’)({\to}\otimes{\to}) cube for f:Bβ†’Bβ€²f\colon B\to B^{\prime} and g:Xβ†’Yg\colon X\to Y tensored on the left by AA, the other being a special case of 5.(v​i​i)5.(vii). For (β†’βŠ—(βˆ™βŠ—βˆ™))({\to}\otimes(\bullet\otimes\bullet)) we must show the following diagram commutes:

A​B​X​Y{\lx@inpgf@ignorespaces ABXY}A​X​Y​B{\lx@inpgf@ignorespaces AXYB}A​X​B​Y{\lx@inpgf@ignorespaces AXBY}    X​Y​A​B{\lx@inpgf@ignorespaces XYAB}X​A​Y​B{\lx@inpgf@ignorespaces XAYB}A​B′​X​Y{\lx@inpgf@ignorespaces AB^{\prime}XY}X​A​B​Y{\lx@inpgf@ignorespaces XABY}A​X​Y​Bβ€²{\lx@inpgf@ignorespaces AXYB^{\prime}}A​X​B′​Y{\lx@inpgf@ignorespaces AXB^{\prime}Y}X​Y​A​Bβ€²{\lx@inpgf@ignorespaces XYAB^{\prime}}X​A​Y​Bβ€²{\lx@inpgf@ignorespaces XAYB^{\prime}}X​A​B′​Y{\lx@inpgf@ignorespaces XAB^{\prime}Y}AβŠ—RB,X​Y\scriptstyle{\lx@inpgf@ignorespaces A\otimes R_{B,XY}}RA,BβŠ—X​Y\scriptstyle{\lx@inpgf@ignorespaces R_{A,B}\otimes XY}⇑AβŠ—R~(B|X,Y){\lx@inpgf@ignorespaces\Uparrow A\otimes\tilde{R}_{(B|X,Y)}}1.{\lx@inpgf@ignorespaces 1.}RA,X​YβŠ—B\scriptstyle{\lx@inpgf@ignorespaces R_{A,XY}\otimes B}7.{\lx@inpgf@ignorespaces 7.}8.{\lx@inpgf@ignorespaces 8.}β‡‘βŠ—RA,X,RB,Yβˆ’1{\lx@inpgf@ignorespaces\Uparrow\otimes_{R_{A,X},R_{B,Y}}^{-1}}2.{\lx@inpgf@ignorespaces 2.}X​YβŠ—RA,B\scriptstyle{\lx@inpgf@ignorespaces XY\otimes R_{A,B}}⇑R~(A|X,Y)βŠ—B{\lx@inpgf@ignorespaces\Uparrow\tilde{R}_{(A|X,Y)}\otimes B}4.{\lx@inpgf@ignorespaces 4.}AβŠ—RBβ€²,XβŠ—Y\scriptstyle{\lx@inpgf@ignorespaces A\otimes R_{B^{\prime},X}\otimes Y}5.{\lx@inpgf@ignorespaces 5.}RA,XβŠ—B′​Y\scriptstyle{\lx@inpgf@ignorespaces R_{A,X}\otimes B^{\prime}Y}6.{\lx@inpgf@ignorespaces 6.}XβŠ—RA,YβŠ—Bβ€²\scriptstyle{\lx@inpgf@ignorespaces X\otimes R_{A,Y}\otimes B^{\prime}}X​AβŠ—RBβ€²,Y\scriptstyle{\lx@inpgf@ignorespaces XA\otimes R_{B^{\prime},Y}}3.{\lx@inpgf@ignorespaces 3.}
1.=AβŠ—Rf,XβŠ—Y2.=βŠ—RA,X,fβŠ—Y3.=XβŠ—AβŠ—Rf,Y4.=XβŠ—βŠ—RA,Y,f5.=AβŠ—Rf,XβŠ—Y6.=βŠ—RA,XβŠ—Y,f7.=AβŠ—XβŠ—Rf,Y8.=βŠ—RA,X,YβŠ—f\begin{array}[]{llll}1.\>=\>A\otimes R_{f,X}\otimes Y&2.\>=\>\otimes_{R_{A,X},f\otimes Y}&3.\>=\>X\otimes A\otimes R_{f,Y}&4.\>=\>X\otimes\otimes_{R_{A,Y},f}\\ 5.\>=\>A\otimes R_{f,X\otimes Y}&6.\>=\>\otimes_{R_{A,X\otimes Y},f}&7.\>=\>A\otimes X\otimes R_{f,Y}&8.\>=\>\otimes_{R_{A,X},Y\otimes f}\end{array}

We cut it into one rectangular and two triangular prisms. To see that the left triangular prism commutes, we apply (β†’βŠ—(βˆ™βŠ—βˆ™))({\to}\otimes(\bullet\otimes\bullet)) to (f,Rf,βˆ’)(f,R_{f,-}), tensored on the left by AA. The rectangular prism commmutes by 5.(v​i)5.(vi) and 5.(v​i​i​i)5.(viii), applied to the 2-morphism Rf,YR_{f,Y}. The right triangular prism commutes by 5.(i​v)5.(iv) and 5.(v​i​i)5.(vii), applied to the 2-morphism R~(A|X,Y)\tilde{R}_{(A|X,Y)}.

The tensor product of an object and a 2-morphism:

(Ad 4.5): For any object (A,RA,βˆ’,R~(A|βˆ’,βˆ’))(A,R_{A,-},\tilde{R}_{(A|-,-)}) and any 2-morphism Ξ±:(f,Rf,βˆ’)β‡’(fβ€²,Rfβ€²,βˆ’)\alpha\colon(f,R_{f,-})\Rightarrow(f^{\prime},R_{f^{\prime},-}) we have a 2-morphism

AβŠ—Ξ±:(AβŠ—f,…)β‡’(AβŠ—fβ€²,…)A\otimes\alpha:(A\otimes f,\dots)\Rightarrow(A\otimes f^{\prime},\dots)

(Ad 4.6): For any object (B,RB,βˆ’,R~(B|βˆ’,βˆ’))(B,R_{B,-},\tilde{R}_{(B|-,-)}) and any 2-morphism Ξ±:(g,Rg,βˆ’)β‡’(gβ€²,Rgβ€²,βˆ’)\alpha:(g,R_{g,-})\Rightarrow(g^{\prime},R_{g^{\prime},-}) we have a 2-morphism

Ξ±βŠ—B:(gβŠ—B,…)β‡’(gβ€²βŠ—B,…)\alpha\otimes B:(g\otimes B,\dots)\Rightarrow(g^{\prime}\otimes B,\dots)

We must verify that these are 2-morphisms in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}), so we must check (β‡“βŠ—βˆ™)({\Downarrow}\otimes\bullet). We do this only for 4.54.5.

A​B​X{\lx@inpgf@ignorespaces ABX}⇓AβŠ—Ξ±βŠ—X{\lx@inpgf@ignorespaces\Downarrow A\otimes\alpha\otimes X}A​B′​X{\lx@inpgf@ignorespaces AB^{\prime}X}A​X​B{\lx@inpgf@ignorespaces AXB}⇓A​XβŠ—Ξ±{\lx@inpgf@ignorespaces\Downarrow AX\otimes\alpha}A​X​Bβ€²{\lx@inpgf@ignorespaces AXB^{\prime}}X​A​B{\lx@inpgf@ignorespaces XAB}⇓X​AβŠ—Ξ±{\lx@inpgf@ignorespaces\Downarrow XA\otimes\alpha}X​A​Bβ€²{\lx@inpgf@ignorespaces XAB^{\prime}}2.{\lx@inpgf@ignorespaces 2.}AβŠ—RB,X\scriptstyle{\lx@inpgf@ignorespaces A\otimes R_{B,X}}AβŠ—RBβ€²,X\scriptstyle{\lx@inpgf@ignorespaces A\otimes R_{B^{\prime},X}}RA,XβŠ—B\scriptstyle{\lx@inpgf@ignorespaces R_{A,X}\otimes B}4.{\lx@inpgf@ignorespaces 4.}RA,XβŠ—Bβ€²\scriptstyle{\lx@inpgf@ignorespaces R_{A,X}\otimes B^{\prime}}1.{\lx@inpgf@ignorespaces 1.}3.{\lx@inpgf@ignorespaces 3.}
1.=AβŠ—Rf,X2.=AβŠ—Rg,X3.=βŠ—RA,X,f4.=βŠ—RA,X,g\begin{array}[]{ll}1.\>=\>A\otimes R_{f,X}&2.\>=\>A\otimes R_{g,X}\\ 3.\>=\>\otimes_{R_{A,X},f}&4.\>=\>\otimes_{R_{A,X},g}\end{array}

The upper prism commutes by (β‡“βŠ—βˆ™)({\Downarrow}\otimes\bullet) tensored from the left by AA. The lower prism commutes by an application of the axiom 4.(v​i)4.(vi) for monoidal 22-categories to the 2-morphism Ξ±\alpha.

The tensor product of morphisms:

(Ad 4.7): For any morphisms (f,Rf,βˆ’):(A,RA,R~A)β†’(Aβ€²,RAβ€²,R~Aβ€²)(f,R_{f,-}):(A,R_{A},\tilde{R}_{A})\to(A^{\prime},R_{A^{\prime}},\tilde{R}_{A^{\prime}}) and (g,Rg,βˆ’):(B,RB,R~B)β†’(Bβ€²,RBβ€²,R~Bβ€²)(g,R_{g,-}):(B,R_{B},\tilde{R}_{B})\to(B^{\prime},R_{B^{\prime}},\tilde{R}_{B^{\prime}}) we have a 2-isomorphism:

βŠ—(f,Rf,βˆ’),(g,Rg,βˆ’):=βŠ—f,g\otimes_{(f,R_{f,-}),(g,R_{g,-})}:=\otimes_{f,g}

To verify that this is a 2-morphism in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}), we have to check (β‡“βŠ—βˆ™)({\Downarrow}\otimes\bullet). The following diagram gives the proof.

A​B​X{\lx@inpgf@ignorespaces ABX}β€ƒβ€ƒβ€„β€ŠA′​B​X{\lx@inpgf@ignorespaces A^{\prime}BX}A​B′​X{\lx@inpgf@ignorespaces AB^{\prime}X}A′​B′​X{\lx@inpgf@ignorespaces A^{\prime}B^{\prime}X}A​X​B{\lx@inpgf@ignorespaces AXB}A′​X​B{\lx@inpgf@ignorespaces A^{\prime}XB}A​X​Bβ€²{\lx@inpgf@ignorespaces AXB^{\prime}}A′​X​Bβ€²{\lx@inpgf@ignorespaces A^{\prime}XB^{\prime}}X​A​B{\lx@inpgf@ignorespaces XAB}X​A′​B{\lx@inpgf@ignorespaces XA^{\prime}B}X​A​Bβ€²{\lx@inpgf@ignorespaces XAB^{\prime}}X​A′​Bβ€²{\lx@inpgf@ignorespaces XA^{\prime}B^{\prime}}AβŠ—RB,X\scriptstyle{\lx@inpgf@ignorespaces A\otimes R_{B,X}}fβŠ—B​X\scriptstyle{\lx@inpgf@ignorespaces f\otimes BX}β‡‘βŠ—f,gβŠ—X{\lx@inpgf@ignorespaces\Uparrow\otimes_{f,g}\otimes X}1.{\lx@inpgf@ignorespaces 1.}Aβ€²βŠ—gΓ—X\scriptstyle{\lx@inpgf@ignorespaces A^{\prime}\otimes g\times X}3.{\lx@inpgf@ignorespaces 3.}fβŠ—B′​X\scriptstyle{\lx@inpgf@ignorespaces f\otimes B^{\prime}X}2.{\lx@inpgf@ignorespaces 2.}Aβ€²βŠ—RBβ€²,X\scriptstyle{\lx@inpgf@ignorespaces A^{\prime}\otimes R_{B^{\prime},X}}RA,XβŠ—B\scriptstyle{\lx@inpgf@ignorespaces R_{A,X}\otimes B}5.{\lx@inpgf@ignorespaces 5.}β‡‘βŠ—fβŠ—X,g{\lx@inpgf@ignorespaces\Uparrow\otimes_{f\otimes X,g}}4.{\lx@inpgf@ignorespaces 4.}7.{\lx@inpgf@ignorespaces 7.}fβŠ—X​Bβ€²\scriptstyle{\lx@inpgf@ignorespaces f\otimes XB^{\prime}}6.{\lx@inpgf@ignorespaces 6.}RAβ€²,XβŠ—Bβ€²\scriptstyle{\lx@inpgf@ignorespaces R_{A^{\prime},X}\otimes B^{\prime}}X​AβŠ—g\scriptstyle{\lx@inpgf@ignorespaces XA\otimes g}8.{\lx@inpgf@ignorespaces 8.}XβŠ—fβŠ—Bβ€²\scriptstyle{\lx@inpgf@ignorespaces X\otimes f\otimes B^{\prime}}⇑XβŠ—βŠ—f,g{\lx@inpgf@ignorespaces\Uparrow X\otimes\otimes_{f,g}}
1.=AβŠ—Rg,X2.=βŠ—f,RBβ€²,X3.=Aβ€²βŠ—Rg,X4.=βŠ—f,RB,X5.=βŠ—RA,X,g6.=Rf,XβŠ—Bβ€²7.=βŠ—RAβ€²,X,g8.=Rf,XβŠ—B\begin{array}[]{llll}1.\>=\>A\otimes R_{g,X}&2.\>=\>\otimes_{f,R_{B^{\prime},X}}&3.\>=\>A^{\prime}\otimes R_{g,X}&4.\>=\>\otimes_{f,R_{B,X}}\\ 5.\>=\>\otimes_{R_{A,X},g}&6.\>=\>R_{f,X}\otimes B^{\prime}&7.\>=\>\otimes_{R_{A^{\prime},X},g}&8.\>=\>R_{f,X}\otimes B\end{array}

The top cube commutes by 4.(i​v),(v​i),(v​i​i​i)4.(iv),(vi),(viii), applied to the 2-morphism AβŠ—Rg,XA\otimes R_{g,X}. The bottom cube commutes by 4.(i​v),(v​i​i),(v​i​i​i)4.(iv),(vii),(viii), applied to the 2-morphism Rf,XβŠ—BR_{f,X}\otimes B.

We have to verify that these data satisfy the conditions 4.(i)βˆ’(v​i​i​i)4.(i)-(viii). These follow from the corresponding conditions holding in π’ž\mathcal{C}.

3.3. The Braiding

(βˆ™βŠ—βˆ™)(\bullet\otimes\bullet): For any two objects we have the morphism

(RA,B,RRA,B,βˆ’):(AβŠ—B,RA,βˆ’βŠ—RB,βˆ’,R~AβŠ—R~B)β†’(BβŠ—A,RB,βˆ’βŠ—RA,βˆ’,R~BβŠ—R~A)(R_{A,B},R_{R_{A,B},-}):(A\otimes B,R_{A,-}\otimes R_{B,-},\tilde{R}_{A}\otimes\tilde{R}_{B})\to(B\otimes A,R_{B,-}\otimes R_{A,-},\tilde{R}_{B}\otimes\tilde{R}_{A})

in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}), where the 2-morphism RRA,B,XR_{R_{A,B},X} is defined to be the pasting:

A​B​X{\lx@inpgf@ignorespaces ABX}B​A​X{\lx@inpgf@ignorespaces BAX}A​X​B{\lx@inpgf@ignorespaces AXB}B​X​A{\lx@inpgf@ignorespaces BXA}X​A​B{\lx@inpgf@ignorespaces XAB}X​B​A{\lx@inpgf@ignorespaces XBA}RA,BβŠ—X\scriptstyle{\lx@inpgf@ignorespaces R_{A,B}\otimes X}AβŠ—RB,X\scriptstyle{\lx@inpgf@ignorespaces A\otimes R_{B,X}}BβŠ—RA,X\scriptstyle{\lx@inpgf@ignorespaces B\otimes R_{A,X}}⇓R~(A|B,X){\lx@inpgf@ignorespaces\Downarrow\tilde{R}_{(A|B,X)}}⇓RA,RB,Xβˆ’1{\lx@inpgf@ignorespaces\Downarrow R_{A,R_{B,X}}^{-1}}⇓R~(A|X,B)βˆ’1{\lx@inpgf@ignorespaces\Downarrow\tilde{R}_{(A|X,B)}^{-1}}RA,XβŠ—B\scriptstyle{\lx@inpgf@ignorespaces R_{A,X}\otimes B}RB,XβŠ—A\scriptstyle{\lx@inpgf@ignorespaces R_{B,X}\otimes A}XβŠ—RA,B\scriptstyle{\lx@inpgf@ignorespaces X\otimes R_{A,B}}

First we have to show that RRA,B,βˆ’R_{R_{A,B},-} satisfies (β†’βŠ—β†’)({\to}\otimes{\to}) and hence is a modification. This is shown in the following diagram (or follows from the fact that it is a pasting of modifications).

A​B​X{\lx@inpgf@ignorespaces ABX}β€ƒβ€ƒβ€„β€ŠB​A​X{\lx@inpgf@ignorespaces BAX}A​B​Xβ€²{\lx@inpgf@ignorespaces ABX^{\prime}}B​A​Xβ€²{\lx@inpgf@ignorespaces BAX^{\prime}}A​X​B{\lx@inpgf@ignorespaces AXB}B​X​A{\lx@inpgf@ignorespaces BXA}A​X′​B{\lx@inpgf@ignorespaces AX^{\prime}B}B​X′​A{\lx@inpgf@ignorespaces BX^{\prime}A}X​A​B{\lx@inpgf@ignorespaces XAB}X​B​A{\lx@inpgf@ignorespaces XBA}X′​A​B{\lx@inpgf@ignorespaces X^{\prime}AB}X′​B​A{\lx@inpgf@ignorespaces X^{\prime}BA}AβŠ—RB,X\scriptstyle{\lx@inpgf@ignorespaces A\otimes R_{B,X}}RA,BβŠ—X\scriptstyle{\lx@inpgf@ignorespaces R_{A,B}\otimes X}B​AβŠ—f\scriptstyle{\lx@inpgf@ignorespaces BA\otimes f}RA,BβŠ—X\scriptstyle{\lx@inpgf@ignorespaces R_{A,B}\otimes X}BβŠ—RA,Xβ€²\scriptstyle{\lx@inpgf@ignorespaces B\otimes R_{A,X^{\prime}}}RA,XβŠ—B\scriptstyle{\lx@inpgf@ignorespaces R_{A,X}\otimes B}RB,Xβ€²βŠ—A\scriptstyle{\lx@inpgf@ignorespaces R_{B,X^{\prime}}\otimes A}fβŠ—A​B\scriptstyle{\lx@inpgf@ignorespaces f\otimes AB}Xβ€²βŠ—RA,B\scriptstyle{\lx@inpgf@ignorespaces X^{\prime}\otimes R_{A,B}}

The front and the back side of the cube are the 2-morphisms RRA,B,XR_{R_{A,B},X} and RRA,B,Xβ€²R_{R_{A,B},X^{\prime}}, respectively. The top and the bottom are βŠ—RA,B,f\otimes_{R_{A,B},f} and βŠ—f,RA,B\otimes_{f,R_{A,B}}, respectively. The left and the right side are the 2-morphisms corresponding to the pseudonatural transformations in the tensor product of the objects AA and BB, (RAβŠ—RB)f(R_{A}\otimes R_{B})_{f} and (RBβŠ—RA)f(R_{B}\otimes R_{A})_{f}, respectively.

The top triangular prism commutes by (βˆ™βŠ—(βˆ™βŠ—β†’))(\bullet\otimes(\bullet\otimes{\to})). The bottom triangular prism commutes by (βˆ™βŠ—(β†’βŠ—βˆ™))(\bullet\otimes({\to}\otimes\bullet)). The cube in the middle commutes by (βˆ™βŠ—(βˆ™βŠ—β†’))β€²(\bullet\otimes(\bullet\otimes{\to}))^{\prime}, which is a consequence of (βˆ™βŠ—β‡“)(\bullet\otimes{\Downarrow}) and (βˆ™βŠ—β†’β†’)(\bullet\otimes{\to}\to) as indicated in LemmaΒ 8.

Next, to show that we have really defined a morphism in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}), we have to verify (β†’βŠ—(βˆ™βŠ—βˆ™))({\to}\otimes(\bullet\otimes\bullet)). This means we have to check the commutativity of the following diagram.

A​B​X​Y{\lx@inpgf@ignorespaces ABXY}A​X​Y​B{\lx@inpgf@ignorespaces AXYB}A​X​B​Y{\lx@inpgf@ignorespaces AXBY}    X​Y​A​B{\lx@inpgf@ignorespaces XYAB}X​A​Y​B{\lx@inpgf@ignorespaces XAYB}B​A​X​Y{\lx@inpgf@ignorespaces BAXY}X​A​B​Y{\lx@inpgf@ignorespaces XABY}B​X​Y​A{\lx@inpgf@ignorespaces BXYA}B​X​A​Y{\lx@inpgf@ignorespaces BXAY}X​Y​B​A{\lx@inpgf@ignorespaces XYBA}X​B​Y​A{\lx@inpgf@ignorespaces XBYA}X​B​A​Y{\lx@inpgf@ignorespaces XBAY}AβŠ—RB,X​Y\scriptstyle{\lx@inpgf@ignorespaces A\otimes R_{B,XY}}RA,BβŠ—X​Y\scriptstyle{\lx@inpgf@ignorespaces R_{A,B}\otimes XY}⇑AβŠ—R~(B|X,Y){\lx@inpgf@ignorespaces\Uparrow A\otimes\tilde{R}_{(B|X,Y)}}1.{\lx@inpgf@ignorespaces 1.}RA,X​YβŠ—B\scriptstyle{\lx@inpgf@ignorespaces R_{A,XY}\otimes B}RA,XβŠ—B​Y\scriptstyle{\lx@inpgf@ignorespaces R_{A,X}\otimes BY}β‡‘βŠ—RA,X,RB,Yβˆ’1{\lx@inpgf@ignorespaces\Uparrow\otimes_{R_{A,X},R_{B,Y}}^{-1}}X​YβŠ—RA,B\scriptstyle{\lx@inpgf@ignorespaces XY\otimes R_{A,B}}⇑R~(A|X,Y)βŠ—B{\lx@inpgf@ignorespaces\Uparrow\tilde{R}_{(A|X,Y)}\otimes B}2.{\lx@inpgf@ignorespaces 2.}BβŠ—RA,XβŠ—Y\scriptstyle{\lx@inpgf@ignorespaces B\otimes R_{A,X}\otimes Y}X​AβŠ—RB,Y\scriptstyle{\lx@inpgf@ignorespaces XA\otimes R_{B,Y}}RB,XβŠ—A​Y\scriptstyle{\lx@inpgf@ignorespaces R_{B,X}\otimes AY}3.{\lx@inpgf@ignorespaces 3.}XβŠ—RB,YβŠ—A\scriptstyle{\lx@inpgf@ignorespaces X\otimes R_{B,Y}\otimes A}X​BβŠ—RA,Y\scriptstyle{\lx@inpgf@ignorespaces XB\otimes R_{A,Y}}
1.=RRA,B,XβŠ—Y2.=XβŠ—RRA,B,Y3.=ROPENRA,B,XβŠ—Y)1.\>=\>R_{R_{A,B},X}\otimes Y\qquad 2.\>=\>X\otimes R_{R_{A,B},Y}\qquad 3.\>=\>R_{R_{A,B},X\otimes Y)}

As shown in the diagram below, we decompose this diagram in the following way: 1) Three tetrahedra of the form (βˆ™βŠ—(βˆ™βŠ—βˆ™βŠ—βˆ™))(\bullet\otimes(\bullet\otimes\bullet\otimes\bullet)). 2) One prism of the form (βˆ™βŠ—(βˆ™βŠ—β†’))(\bullet\otimes(\bullet\otimes{\to})), namely (AβŠ—(XβŠ—(B​Yβ†’Y​B)))(A\otimes(X\otimes(BY\to YB))) (second row, right). 3) One prism of the form (βˆ™βŠ—(β†’βŠ—βˆ™))(\bullet\otimes({\to}\otimes\bullet)), namely (AβŠ—((B​Xβ†’X​B)βŠ—Y))(A\otimes((BX\to XB)\otimes Y)) (second row, left). 4) One prism of the form (βˆ™βŠ—β‡“)(\bullet\otimes{\Downarrow}), namely (AβŠ—R~(B|X,Y))(A\otimes\tilde{R}_{(B|X,Y)}) (in the middle of the first row). All of these diagrams commute by our assumptions.

A​B​X​YABXYB​A​X​YBAXYB​X​A​YBXAYB​X​Y​ABXYAA​B​X​YABXYA​X​Y​BAXYBA​X​B​YAXBYB​X​Y​ABXYAX​Y​B​AXYBAX​B​Y​AXBYAA​X​Y​BAXYBX​Y​A​BXYABX​A​Y​BXAYBX​Y​B​AXYBAA​B​X​YABXYA​X​B​YAXBYB​X​A​YBXAYB​X​Y​ABXYAX​B​A​YXBAYX​B​Y​AXBYAA​X​B​YAXBYX​A​B​YXABYX​B​A​YXBAYX​B​Y​AXBYAA​X​B​YAXBYA​X​Y​BAXYBX​A​B​YXABYX​A​Y​BXAYBX​B​Y​AXBYAX​Y​B​AXYBA

(β†’βŠ—βˆ™)({\to}\otimes\bullet): For any 1-morphism (f,Rf,βˆ’):(A,RA,R~A)β†’(Aβ€²,RAβ€²,R~Aβ€²)(f,R_{f,-})\colon(A,R_{A},\tilde{R}_{A})\to(A^{\prime},R_{A^{\prime}},\tilde{R}_{A^{\prime}}) and any object (B,RB,R~B)βˆˆπ’΅β‘(π’ž)(B,R_{B},\tilde{R}_{B})\in\mathcal{Z}(\mathcal{C}) we have a 2-isomorphism

Rf,B:(fβŠ—B)​RAβ€²,Bβ‡’RA,B​(BβŠ—f)R_{f,B}:(f\otimes B)R_{A^{\prime},B}\Rightarrow R_{A,B}(B\otimes f)

The following diagram shows that Rf,BR_{f,B} satisfies (β‡“βŠ—βˆ™)({\Downarrow}\otimes\bullet) and is therefore a 2-morphism in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}).

A​B​X{\lx@inpgf@ignorespaces ABX}A′​B​X{\lx@inpgf@ignorespaces A^{\prime}BX}B​A​X{\lx@inpgf@ignorespaces BAX}B​A′​X{\lx@inpgf@ignorespaces BA^{\prime}X}(β†’βŠ—(βˆ™βŠ—βˆ™)){\lx@inpgf@ignorespaces({\to}\otimes(\bullet\otimes\bullet))}A​X​B{\lx@inpgf@ignorespaces AXB}A′​X​B{\lx@inpgf@ignorespaces A^{\prime}XB}B​X​A{\lx@inpgf@ignorespaces BXA}B​X​Aβ€²{\lx@inpgf@ignorespaces BXA^{\prime}}(β†’βŠ—β†’){\lx@inpgf@ignorespaces({\to}\otimes{\to})}X​A​B{\lx@inpgf@ignorespaces XAB}X​A′​B{\lx@inpgf@ignorespaces XA^{\prime}B}X​B​A{\lx@inpgf@ignorespaces XBA}X​B​Aβ€²{\lx@inpgf@ignorespaces XBA^{\prime}}(β†’βŠ—(βˆ™βŠ—βˆ™)){\lx@inpgf@ignorespaces({\to}\otimes(\bullet\otimes\bullet))}fβŠ—B​X\scriptstyle{\lx@inpgf@ignorespaces f\otimes BX}AβŠ—RB,X\scriptstyle{\lx@inpgf@ignorespaces A\otimes R_{B,X}}⇓Rf,BβŠ—X{\lx@inpgf@ignorespaces\Downarrow R_{f,B}\otimes X}RAβ€²,BβŠ—X\scriptstyle{\lx@inpgf@ignorespaces R_{A^{\prime},B}\otimes X}BβŠ—fβŠ—X\scriptstyle{\lx@inpgf@ignorespaces B\otimes f\otimes X}BβŠ—RAβ€²,X\scriptstyle{\lx@inpgf@ignorespaces B\otimes R_{A^{\prime},X}}RA,XβŠ—B\scriptstyle{\lx@inpgf@ignorespaces R_{A,X}\otimes B}RB,XβŠ—Aβ€²\scriptstyle{\lx@inpgf@ignorespaces R_{B,X}\otimes A^{\prime}}XβŠ—RA,B\scriptstyle{\lx@inpgf@ignorespaces X\otimes R_{A,B}}X​BβŠ—f\scriptstyle{\lx@inpgf@ignorespaces XB\otimes f}⇓XβŠ—Rf,B{\lx@inpgf@ignorespaces\Downarrow X\otimes R_{f,B}}

The left and right sides are the 2-morphisms RRA,B,XR_{R_{A,B},X} and RRAβ€²,B,XR_{R_{A^{\prime},B},X}, respectively. The front and the back sides are pastings as in our treatment in SectionΒ 3.2 of the tensor product of an object and a morphism in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}).

We decompose this cube into two commutative triangular prisms of the form (β†’βŠ—(βˆ™βŠ—βˆ™))({\to}\otimes(\bullet\otimes\bullet)), correspoding to (fβŠ—(BβŠ—X))(f\otimes(B\otimes X)) and (fβŠ—(XβŠ—B))(f\otimes(X\otimes B)), and one cube of the form (β†’βŠ—β†’)({\to}\otimes{\to}), namely (Aβ†’Aβ€²βŠ—B​Xβ†’X​B)(A\to A^{\prime}\otimes BX\to XB).

(βˆ™βŠ—β†’)(\bullet\otimes{\to}): For any 1-morphism (g,Rg,βˆ’):(B,RB,R~B)β†’(Bβ€²,RBβ€²,R~Bβ€²)(g,R_{g,-}):(B,R_{B},\tilde{R}_{B})\to(B^{\prime},R_{B^{\prime}},\tilde{R}_{B^{\prime}}) and any object (A,RA,R~A)βˆˆπ’΅β‘(π’ž)(A,R_{A},\tilde{R}_{A})\in\mathcal{Z}(\mathcal{C}), we have a 2-iso

RA,g:(AβŠ—g)​RA,Bβ€²β‡’RA,B​(gβŠ—A)R_{A,g}:(A\otimes g)R_{A,B^{\prime}}\Rightarrow R_{A,B}(g\otimes A)

The following diagram shows that RA,gR_{A,g} satisfies (β‡“βŠ—βˆ™)({\Downarrow}\otimes\bullet) and is thus a 2-morphism in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}).

A​B​X{\lx@inpgf@ignorespaces ABX}A​B′​X{\lx@inpgf@ignorespaces AB^{\prime}X}B​A​X{\lx@inpgf@ignorespaces BAX}B′​A​X{\lx@inpgf@ignorespaces B^{\prime}AX}(βˆ™βŠ—(β†’βŠ—βˆ™)){\lx@inpgf@ignorespaces(\bullet\otimes({\to}\otimes\bullet))}A​X​B{\lx@inpgf@ignorespaces AXB}A​X​Bβ€²{\lx@inpgf@ignorespaces AXB^{\prime}}B​X​A{\lx@inpgf@ignorespaces BXA}B′​X​A{\lx@inpgf@ignorespaces B^{\prime}XA}(βˆ™βŠ—(β†’βŠ—βˆ™)β€²){\lx@inpgf@ignorespaces(\bullet\otimes({\to}\otimes\bullet)^{\prime})}X​A​B{\lx@inpgf@ignorespaces XAB}X​A​Bβ€²{\lx@inpgf@ignorespaces XAB^{\prime}}X​B​A{\lx@inpgf@ignorespaces XBA}X​B′​A{\lx@inpgf@ignorespaces XB^{\prime}A}(βˆ™βŠ—(β†’βŠ—βˆ™)){\lx@inpgf@ignorespaces(\bullet\otimes({\to}\otimes\bullet))}AβŠ—gβŠ—X\scriptstyle{\lx@inpgf@ignorespaces A\otimes g\otimes X}AβŠ—RB,X\scriptstyle{\lx@inpgf@ignorespaces A\otimes R_{B,X}}⇓RA,gβŠ—X{\lx@inpgf@ignorespaces\Downarrow R_{A,g}\otimes X}RA,Bβ€²βŠ—X\scriptstyle{\lx@inpgf@ignorespaces R_{A,B^{\prime}}\otimes X}gβŠ—A​X\scriptstyle{\lx@inpgf@ignorespaces g\otimes AX}Bβ€²βŠ—RA,X\scriptstyle{\lx@inpgf@ignorespaces B^{\prime}\otimes R_{A,X}}RA,XβŠ—B\scriptstyle{\lx@inpgf@ignorespaces R_{A,X}\otimes B}RBβ€²,XβŠ—A\scriptstyle{\lx@inpgf@ignorespaces R_{B^{\prime},X}\otimes A}X​RA,B\scriptstyle{\lx@inpgf@ignorespaces XR_{A,B}}XβŠ—gβŠ—A\scriptstyle{\lx@inpgf@ignorespaces X\otimes g\otimes A}⇓XβŠ—RA,g{\lx@inpgf@ignorespaces\Downarrow X\otimes R_{A,g}}

The decomposition is similar to the one before.

((βˆ™βŠ—βˆ™)βŠ—βˆ™)((\bullet\otimes\bullet)\otimes\bullet): For any objects (A,RA,R~A),(B,RB,R~B),(C,RC,R~C)βˆˆπ’΅β‘(π’ž)(A,R_{A},\tilde{R}_{A}),(B,R_{B},\tilde{R}_{B}),(C,R_{C},\tilde{R}_{C})\in\mathcal{Z}(\mathcal{C}) we have the 2-isomorphism R~(A,B|C):=1(RAβŠ—RB)C\tilde{R}_{(A,B|C)}:=1_{(R_{A}\otimes R_{B})_{C}}:

AβŠ—BβŠ—C{\lx@inpgf@ignorespaces A\otimes B\otimes C} CβŠ—AβŠ—B{\lx@inpgf@ignorespaces C\otimes A\otimes B}AβŠ—CβŠ—B{\lx@inpgf@ignorespaces A\otimes C\otimes B}(RAβŠ—RB)C\scriptstyle{\lx@inpgf@ignorespaces(R_{A}\otimes R_{B})_{C}}AβŠ—RB,C\scriptstyle{\lx@inpgf@ignorespaces A\otimes R_{B,C}}⇑1{\lx@inpgf@ignorespaces\Uparrow 1}RA,CβŠ—B\scriptstyle{\lx@inpgf@ignorespaces R_{A,C}\otimes B}

((βˆ™βŠ—(βˆ™βŠ—βˆ™))((\bullet\otimes(\bullet\otimes\bullet)): For any objects (A,RA,R~A),(B,RB,R~B),(C,RC,R~C)βˆˆπ’΅β‘(π’ž)(A,R_{A},\tilde{R}_{A}),(B,R_{B},\tilde{R}_{B}),(C,R_{C},\tilde{R}_{C})\in\mathcal{Z}(\mathcal{C}) we have the 2-isomorphism R~(A|B,C)\tilde{R}_{(A|B,C)}:

AβŠ—BβŠ—C{\lx@inpgf@ignorespaces A\otimes B\otimes C} BβŠ—CβŠ—A{\lx@inpgf@ignorespaces B\otimes C\otimes A}BβŠ—AβŠ—C{\lx@inpgf@ignorespaces B\otimes A\otimes C}RA,(BβŠ—C)\scriptstyle{\lx@inpgf@ignorespaces R_{A,(B\otimes C)}}RA,BβŠ—C\scriptstyle{\lx@inpgf@ignorespaces R_{A,B}\otimes C}⇑R~(A|B,C){\lx@inpgf@ignorespaces\Uparrow\tilde{R}_{(A|B,C)}}BβŠ—RA,C\scriptstyle{\lx@inpgf@ignorespaces B\otimes R_{A,C}}

To verify that R~(A|βˆ’,βˆ’)\tilde{R}_{(A|-,-)} is a 2-morphism in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}), we have to check (β‡“βŠ—βˆ™)({\Downarrow}\otimes\bullet). The next diagram gives the proof.

A​B​C​X{\lx@inpgf@ignorespaces ABCX}B​C​A​X{\lx@inpgf@ignorespaces BCAX}B​A​C​X{\lx@inpgf@ignorespaces BACX}(βˆ™βŠ—(βˆ™βŠ—βˆ™βŠ—βˆ™)){\lx@inpgf@ignorespaces(\bullet\otimes(\bullet\otimes\bullet\otimes\bullet))}A​B​X​C{\lx@inpgf@ignorespaces ABXC}B​C​X​A{\lx@inpgf@ignorespaces BCXA}B​A​X​C{\lx@inpgf@ignorespaces BAXC}(βˆ™βŠ—(βˆ™βŠ—β†’)=(AβŠ—(BβŠ—(CXβ†’XC))){\lx@inpgf@ignorespaces(\bullet\otimes(\bullet\otimes{\to})=(A\otimes(B\otimes(CX\to XC)))}A​X​B​C{\lx@inpgf@ignorespaces AXBC}B​X​C​A{\lx@inpgf@ignorespaces BXCA}(βˆ™βŠ—(βˆ™βŠ—βˆ™βŠ—βˆ™)){\lx@inpgf@ignorespaces(\bullet\otimes(\bullet\otimes\bullet\otimes\bullet))}B​X​A​C{\lx@inpgf@ignorespaces BXAC}(βˆ™βŠ—(β†’βŠ—βˆ™))=(AβŠ—(BXβ†’XB)βŠ—C){\lx@inpgf@ignorespaces(\bullet\otimes({\to}\otimes\bullet))=(A\otimes(BX\to XB)\otimes C)}X​A​B​C{\lx@inpgf@ignorespaces XABC}X​B​C​A{\lx@inpgf@ignorespaces XBCA}(βˆ™βŠ—(βˆ™βŠ—βˆ™βŠ—βˆ™)){\lx@inpgf@ignorespaces(\bullet\otimes(\bullet\otimes\bullet\otimes\bullet))}X​B​A​C{\lx@inpgf@ignorespaces XBAC}RA,B​CβŠ—X\scriptstyle{\lx@inpgf@ignorespaces R_{A,BC}\otimes X}A​BβŠ—RC,X\scriptstyle{\lx@inpgf@ignorespaces AB\otimes R_{C,X}}B​CβŠ—RA,X\scriptstyle{\lx@inpgf@ignorespaces BC\otimes R_{A,X}}AβŠ—RB,XβŠ—C\scriptstyle{\lx@inpgf@ignorespaces A\otimes R_{B,X}\otimes C}BβŠ—RC,XβŠ—A\scriptstyle{\lx@inpgf@ignorespaces B\otimes R_{C,X}\otimes A}RA,XβŠ—B​C\scriptstyle{\lx@inpgf@ignorespaces R_{A,X}\otimes BC}RB,XβŠ—C​A\scriptstyle{\lx@inpgf@ignorespaces R_{B,X}\otimes CA}

The top triangle corresponds to the 2-morphism R~(A|B,C)βŠ—X\tilde{R}_{(A|B,C)}\otimes X. The bottom triangle corresponds to the 2-morphism XβŠ—R~(A|B,C)X\otimes\tilde{R}_{(A|B,C)}. The back side is RRA,BβŠ—C,XR_{R_{A,B\otimes C},X}, the left front side is RRA,BβŠ—C,XR_{R_{A,B}\otimes C,X} and the right front side is RBβŠ—RA,C,XR_{B\otimes R_{A,C},X}. The decomposition is indicated in the diagram.

Now we have to verify that these data satisfy all the axioms of a braided monoidal 2-category. The tetrahedron (βˆ™βŠ—(βˆ™βŠ—βˆ™βŠ—βˆ™))(\bullet\otimes(\bullet\otimes\bullet\otimes\bullet)) commutes by the definition of the objects of 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}). The diagram ((βˆ™βŠ—βˆ™)βŠ—(βˆ™βŠ—βˆ™))((\bullet\otimes\bullet)\otimes(\bullet\otimes\bullet)) commutes by the definition of the tensor product of two objects in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}). By the same definition can be shown that ((βˆ™βŠ—βˆ™βŠ—βˆ™)βŠ—βˆ™)((\bullet\otimes\bullet\otimes\bullet)\otimes\bullet) commutes in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}). Note that because of our special choice of the 2-morphism RRA,B,βˆ’R_{R_{A,B},-} that completes the morphism RA,BR_{A,B} to a morphism in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}), the two 2-morphisms S+S^{+} and Sβˆ’S^{-} are equal in 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}).

The other axioms of a braided monoidal 2-category are either part of our definitions, or else we have indicated within our Remarks which definitions imply them. We may summarize by stating:

0N8D

Theorem 15. Given any semistrict monoidal category π’ž\mathcal{C}, the center 𝒡⁑(π’ž)\mathcal{Z}(\mathcal{C}) is semistrict braided monoidal 2-category.

Original mathematics by the credited authors. Source collection and HTML conversion remain in progress.

John C. Baez, Martin Neuchl

Original source: arXiv:q-alg/9511013v2