4.1. Monoidal category
4.1.1. Definition
Let π° {\mathcal{U}} be the differential strict monoidal category generated by an object e e and a map Ο : e 2 β e 2 \tau:e^{2}\to e^{2} subject to the
relations
(4.1.1)
d β‘ ( Ο ) = 1 , Ο 2 = 0 β Β andΒ β e β Ο β Ο β e β e β Ο = Ο β e β e β Ο β Ο β e . d(\tau)=1,\ \tau^{2}=0\text{ and }e\tau\circ\tau e\circ e\tau=\tau e\circ e\tau\circ\tau e.
There are isomorphisms of differential monoidal categories
opp : π° β βΌ π° opp {\operatorname{opp}\nolimits}:{\mathcal{U}}\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}{\mathcal{U}}^{\operatorname{opp}\nolimits} and rev : π° β βΌ π° rev \mathrm{rev}:{\mathcal{U}}\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}{\mathcal{U}}^{\mathrm{rev}}
given on generators by
e β¦ e e\mapsto e and Ο β¦ Ο \tau\mapsto\tau .
The following result is clear.
0P5F
Proposition 4.1.1 . The objects of the category π° {\mathcal{U}} are the e n e^{n} , n β₯ 0 n\geq 0 . We have
Hom β‘ ( e n , e m ) = 0 \operatorname{Hom}\nolimits(e^{n},e^{m})=0 if n β m n\neq m and there is an isomorphism of differential
algebras
H n β βΌ End β‘ ( e n ) , T i β¦ e i β 1 β Ο β e n β i β 1 . H_{n}\xrightarrow{\sim}\operatorname{End}\nolimits(e^{n}),\ T_{i}\mapsto e^{i-1}\tau e^{n-i-1}.
There is a commutative diagram
H m β H n \textstyle{H_{m}\otimes H_{n}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces} T i β T j β¦ T i β T m + j \scriptstyle{T_{i}\otimes T_{j}\mapsto T_{i}T_{m+j}} can \scriptstyle{{\mathrm{can}}} βΌ \scriptstyle{\sim} H m + n \textstyle{H_{m+n}\ignorespaces\ignorespaces\ignorespaces\ignorespaces} βΌ \scriptstyle{\sim} can \scriptstyle{{\mathrm{can}}} End β‘ ( E m ) β End β‘ ( E n ) \textstyle{\operatorname{End}\nolimits(E^{m})\otimes\operatorname{End}\nolimits(E^{n})\ignorespaces\ignorespaces\ignorespaces\ignorespaces} β \scriptstyle{\otimes} End β‘ ( E m + n ) \textstyle{\operatorname{End}\nolimits(E^{m+n})}
The isomorphism opp : π° β βΌ π° opp {\operatorname{opp}\nolimits}:{\mathcal{U}}\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}{\mathcal{U}}^{\operatorname{opp}\nolimits} gives rise to the
isomorphism of differential algebras
opp : H n β βΌ H n opp , T i β¦ T i . {\operatorname{opp}\nolimits}:H_{n}\xrightarrow{\sim}H_{n}^{\operatorname{opp}\nolimits},\ T_{i}\mapsto T_{i}.
The isomorphism rev : π° β βΌ π° rev \mathrm{rev}:{\mathcal{U}}\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}{\mathcal{U}}^{\mathrm{rev}} gives rise to the
isomorphism of differential algebras
ΞΉ n : H n β βΌ H n , T i β¦ T n β i . \iota_{n}:H_{n}\xrightarrow{\sim}H_{n},\ T_{i}\mapsto T_{n-i}.
The functor β β E n -\otimes E^{n} induces an injective morphism of differential algebras
H r = End β‘ ( E r ) β H r + n = End β‘ ( E r + n ) , T i β¦ T i H_{r}=\operatorname{End}\nolimits(E^{r})\to H_{r+n}=\operatorname{End}\nolimits(E^{r+n}),\ T_{i}\mapsto T_{i} and we will identify H r H_{r} with a subalgebra
of H r + n H_{r+n} via this morphism.
The functor E n β β E^{n}\otimes- induces a morphism of differential algebras
f n : H r = End β‘ ( E r ) β H n + r = End β‘ ( E n + r ) , T i β¦ T n + i . f_{n}:H_{r}=\operatorname{End}\nolimits(E^{r})\to H_{n+r}=\operatorname{End}\nolimits(E^{n+r}),\ T_{i}\mapsto T_{n+i}.
Note that H n H_{n} commutes with f n β ( H r ) f_{n}(H_{r}) and that
f n = ΞΉ n + r β ΞΉ r f_{n}=\iota_{n+r}\circ\iota_{r} .
4.1.2. 2 2 -representations
Let π± {\mathcal{V}} be a differential category.
0P5G
Definition 4.1.2 . A 2 2 -representation on π± {\mathcal{V}} is the data
of a strict monoidal differential functor π° β End β‘ ( π± ) {\mathcal{U}}\to\operatorname{End}\nolimits({\mathcal{V}}) .
The data of a
2 2 -representation on π± {\mathcal{V}} is the same as the data of a
differential endofunctor E E of π± {\mathcal{V}} and of Ο = Ο E β End β‘ ( E 2 ) \tau=\tau_{E}\in\operatorname{End}\nolimits(E^{2})
satisfying
(4.1.1 ).
Note that a 2 2 -representation on π± {\mathcal{V}} extends to a 2 2 -representation
on π± Β― \bar{{\mathcal{V}}} and on π± i {\mathcal{V}}^{i}
(uniquely up to an equivalence unique up to isomorphism).
A morphism of 2 2 -representations
( π± , E , Ο ) β ( π± β² , E β² , Ο ) ({\mathcal{V}},E,\tau)\to({\mathcal{V}}^{\prime},E^{\prime},\tau) is the data of a
differential functor Ξ¦ : π± β π± β² \Phi:{\mathcal{V}}\to{\mathcal{V}}^{\prime} and of an isomorphism
of functors Ο : Ξ¦ β E β βΌ E β² β Ξ¦ \varphi:\Phi E\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}E^{\prime}\Phi (with d β‘ ( Ο ) = 0 d(\varphi)=0 ) such that
Ο β² β Ξ¦ β E β² β Ο β Ο β E = E β² β Ο β Ο β E β Ξ¦ β Ο : Ξ¦ β E 2 β E β² 2 β Ξ¦ \tau^{\prime}\Phi\circ E^{\prime}\varphi\circ\varphi E=E^{\prime}\varphi\circ\varphi E\circ\Phi\tau:\Phi E^{2}\to E^{\prime 2}\Phi .
0P5H
Example 4.1.3 . Let π± = k β β diff {\mathcal{V}}=k\operatorname{\!-diff}\nolimits and E = Ο = 0 E=\tau=0 . This is the βtrivialβ
2 2 -representation.
Let π± {\mathcal{V}} be a 2 2 -representation. The opposite 2 2 -representation
is ( π± β β diff , E β² , Ο β² ) ({\mathcal{V}}\operatorname{\!-diff}\nolimits,E^{\prime},\tau^{\prime}) , where
E β² β ( ΞΆ ) = ΞΆ β E E^{\prime}(\zeta)=\zeta E and Ο β² β ( ΞΆ ) = ΞΆ β Ο β End β‘ ( E β² 2 β ( ΞΆ ) ) \tau^{\prime}(\zeta)=\zeta\tau\in\operatorname{End}\nolimits(E^{\prime 2}(\zeta))
for ΞΆ β π± β β diff \zeta\in{\mathcal{V}}\operatorname{\!-diff}\nolimits . Note that
the canonical functor π± β ( π± β β diff ) β β diff , v β¦ ( ΞΆ β¦ ΞΆ β‘ ( v ) ) {\mathcal{V}}\to({\mathcal{V}}\operatorname{\!-diff}\nolimits)\operatorname{\!-diff}\nolimits,\ v\mapsto(\zeta\mapsto\zeta(v)) is a fully faithful morphism of 2 2 -representations.
Assume E E has a left adjoint E β¨ E^{\vee} . We still denote by Ο \tau the endomorphism of
( E β¨ ) 2 (E^{\vee})^{2} corresponding to Ο \tau (cf Β§2.1.1 ). The pair ( E β¨ , Ο ) (E^{\vee},\tau) defines the
left dual 2 2 -representation of ( E , Ο ) (E,\tau) .
Similarly, if E E has a right adjoint β¨ E {{}^{\vee}E} , we obtain a
right dual 2 2 -representation
( E β¨ , Ο ) ({{}^{\vee}E},\tau) of ( E , Ο ) (E,\tau) .
4.1.3. Pointed case
We denote by π° β {\mathcal{U}}^{\bullet} the strict monoidal differential pointed category generated
by an object e e and a map Ο β End β‘ ( e 2 ) \tau\in\operatorname{End}\nolimits(e^{2}) subject to the relations
(4.1.1 ). Its objects are the e n e^{n} , n β₯ 0 n\geq 0 ,
Hom β‘ ( e n , e m ) = 0 \operatorname{Hom}\nolimits(e^{n},e^{m})=0 for m β n m\neq n and End β‘ ( e n ) = H n β \operatorname{End}\nolimits(e^{n})=H_{n}^{\bullet} .
Let π± {\mathcal{V}} be a differential pointed category.
A 2 2 -representation on π± {\mathcal{V}} is the data of a strict monoidal
differential pointed functor π° β β End β‘ ( π± ) {\mathcal{U}}^{\bullet}\to\operatorname{End}\nolimits({\mathcal{V}}) .
This is equivalent to the data of an endofunctor E E of the differential pointed category
π± {\mathcal{V}} and Ο β End β‘ ( E 2 ) \tau\in\operatorname{End}\nolimits(E^{2}) such that
( E , Ο ) (E,\tau) induce a 2 2 -representation on k β‘ [ π± ] k[{\mathcal{V}}] .