nLab Segal's category

Redirected from "Segal Gamma-category".
Contents

This entry is about the small category named after Graeme Segal. Not to be confused with the model in higher category theory called Segal categories.

Contents

Definition

Definition

Segal’s category, denoted by Γ\Gamma, is the category opposite to the skeleton of the category FinSet */FinSet^{*/} of pointed finite sets:

Γ opFinSet */.\Gamma^{op} \simeq FinSet^{*/}.

Equivalently, Γ\Gamma is the category opposite to the category FinSetPartFinSetPart of finite sets and partially defined maps. Objects are finite sets. Morphisms ABA\to B are pairs (A,f)(A',f), where AAA'\subset A and f:ABf\colon A'\to B is a map of sets. Identity morphisms are specified by id A=(A,id A)id_A=(A,id_A). Composition is defined as follows:

(A,f):AB,(B,g):BC,(B,g)(A,f)=(Af 1B,gf| Af 1B).(A',f)\colon A\to B,\qquad (B',g)\colon B\to C,\qquad(B',g)\circ(A',f)=(A'\cap f^{-1}B',g\circ f|_{A'\cap f^{-1}B'}).

Equivalently (Segal, Definition 1.1), Γ\Gamma admits the following covariant description. Objects are finite sets. Morphisms ABA\to B are maps f:AP(B)f\colon A\to P(B) such that a 1a 2a_1\ne a_2 implies f(a 1)f(a 2)=f(a_1)\cap f(a_2)=\emptyset. Identity morphisms are specified by id A(a)={a}id_A(a)=\{a\}. Composition is defined as follows:

f:AP(B),g:BP(C),(gf)(a)= bf(a)g(b).f\colon A\to P(B),\qquad g\colon B\to P(C),\qquad (g\circ f)(a)=\bigcup_{b\in f(a)}g(b).
Remark

The category Γ\Gamma is related to (infinity,1)-operads in a way similar to how the simplex category (non-empty and linearly ordered finite sets) is related to (∞,1)-categories.

Remark

The categories FinSetPartFinSetPart and FinSet *FinSet_* are equivalent via the following pair of functors. The functor

FinSetPartFinSet *FinSetPart\to FinSet_*

sends

AA{*},((A,f):AB)(g:A{*}B{*}),g| A=f,g| A¯(x)=*.A\mapsto A\sqcup\{*\},\qquad ((A',f)\colon A\to B)\mapsto(g\colon A\sqcup\{*\}\to B\sqcup\{*\}),\qquad g|_{A'}=f,\qquad g|_{\bar A'}(x)=*.

The functor

FinSet *FinSetPartFinSet_* \to FinSetPart

sends

AA{*},(g:AB)(g 1(B{*}),g| A{*}).A\mapsto A\setminus\{*\},\qquad (g\colon A\to B)\mapsto(g^{-1}(B\setminus\{*\}),g|_{A\setminus\{*\}}).

Comparison to the simplex category

Recall the simplex category Δ\Delta, whose objects are finite inhabited totally ordered sets and morphisms are order-preserving maps of sets. We set

[m]={0<<m}.[m]=\{0\lt\cdots\lt m\}.

Interpreting a simplex AΔA\in\Delta as a (posetal) category, a spine edge in AA is a morphism in AA that is not an identity morphism and cannot be decomposed as the composition of two nonidentity morphisms. That is to say, in a standard simplex [m][m], a spine edge is a pair of the form (i1,i)(i-1,i), 0<im0\lt i\le m, which encodes the morphism (i1)i(i-1)\to i.

The canonical functor

ι:ΔΓ\iota\colon\Delta\to\Gamma

sends a simplex to its set of spine edges:

[m]={0<1<2<<m}{(0,1),(1,2),,(m1,m)}=ι[m].[m]=\{0\lt1\lt2\lt\cdots\lt m\}\mapsto \{(0,1),(1,2),\ldots,(m-1,m)\}=\iota[m].

To define ι\iota on morphisms, observe that a morphism of simplices f:[m][n]f\colon[m]\to[n] can be interpreted as a functor between posetal categories. This functor sends a spine edge in [m][m] to the composition of a set of spine edges in [n][n]. More specifically, ff sends a spine edge i1ii-1\to i to the composition of spine edges

f(i1)f(i1)+1f(i)1f(i).f(i-1)\to f(i-1)+1\to \cdots \to f(i)-1\to f(i).

Using Segal’s covariant description of Γ\Gamma, we can define a morphism ι(f):ι[m]ι[n]\iota(f)\colon\iota[m]\to\iota[n] by sending a spine edge ee in [m][m] to the set of spine edges in [n][n] whose composition yields the image of ee under the functor f:[m][n]f\colon[m]\to [n].

In terms of the category FinSetPart, we can describe ι(f)\iota(f) as a partial map of finite sets ι[n]ι[m]\iota[n]\to\iota[m] that sends (j1,j)ι[n](j-1,j)\in\iota[n] to the unique pair (i1,i)ι[m](i-1,i)\in\iota[m] such that f(i1)<jf(i)f(i-1)\lt j\le f(i). (If no such pair exists, then ι(f)\iota(f) is undefined on (j1,j)(j-1,j).)

Alternative description of the comparison functor

Given that the category Δ\Delta can be used to encode homotopy coherent associative monoids in the same manner that the category Γ\Gamma can be used to encode homotopy coherent commutative monoids, it is natural to wonder why the description of the comparison functor ι:ΔΓ\iota\colon \Delta\to\Gamma appears to be so asymmetric. The primary reason for this asymmetry is that the traditional description of the category Δ\Delta relies on the fact that elements are composed in a prescribed order, just like morphisms in a category. This enables us to define Δ\Delta as a full subcategory of the category of small categories. There is no analogue of such categorical structure for the category Γ\Gamma.

However, it is possible to give an alternative definition of the category Δ\Delta resembling Segal’s covariant formulation and the category FinSetPart. In these formulations, objects encode sets of spine edges (as defined above) instead of sets of vertices.

In the first formulation, the category Δ\Delta is defined as the category opposite to the following category. Objects are finite totally ordered sets. Morphisms ABA\to B are pairs (I,f)(I,f), where II is an interval in AA, defined as a triple (X,Y,Z)(X,Y,Z) of disjoint subsets of AA whose union equals AA and such that X<Y<ZX\lt Y\lt Z, and f:YBf\colon Y\to B is an order-preserving map. The set YY uniquely determines XX and ZZ unless Y=Y=\emptyset. The identity morphism on AA is given by the pair (I,id A)(I,id_A), where I=(,A,)I=(\emptyset,A,\emptyset). Given morphisms

(I 2,f 2):A 1A 2,(I 1,f 1):A 0A 1,(I_2,f_2)\colon A_1\to A_2,\qquad (I_1,f_1)\colon A_0\to A_1,

we define their composition as

(I 2,f 2)(I 1,f 1)=(I,f 2f 1| Y 1f 1 1Y 2),(I_2,f_2)\circ(I_1,f_1)=(I,f_2\circ f_1|_{Y_1\cap f_1^{-1}Y_2}),

where

I=(X 1f 1 1X 2,Y 1f 1 1Y 2,Z 1f 1 1Z 2).I=(X_1\cup f_1^{-1} X_2,Y_1\cap f_1^{-1} Y_2,Z_1\cup f_1^{-1} Z_2).

Now the forgetful functor ΔΓ=FinSetPart\Delta\to\Gamma=FinSetPart can be described as the functor that discards the total order on objects and discards the interval II on morphisms.

There is also an analog of the covariant description of Γ\Gamma for the category Δ\Delta. In this setting, Δ\Delta is the following category. Objects are finite totally ordered sets. Morphisms ABA\to B are maps of sets f:{}A{}P(B)f\colon \{-\infty\}\sqcup A\sqcup\{\infty\}\to P(B) such that a 1<a 2a_1\lt a_2 implies f(a 1)<f(a 2)f(a_1)\lt f(a_2) and we have a{}A{}f(a)=B\bigcup_{a\in \{-\infty\}\sqcup A\sqcup\{\infty\}}f(a)=B. The sets f()f(-\infty) and f()f(\infty) are uniquely determined by aAf(a)\bigcup_{a\in A}f(a) unless the latter is empty. Morphisms are composed like in the category Γ\Gamma.

Now the forgetful functor ΔΓ\Delta\to\Gamma can be described as the functor that discards the total order on objects and discards the data of f()f(-\infty) and f()f(\infty) on morphisms.

Monoidal structure

The category Γ\Gamma is a symmetric monoidal category, induced from the cartesian monoidal structure on the category Set:

A 1A 2=A 1×A 2,(A 1,f 1)(A 2,f 2)=(A 1×A 2,f 1×f 2).A_1\otimes A_2=A_1\times A_2,\qquad (A'_1,f_1)\otimes(A'_2,f_2)=(A'_1\times A'_2,f_1\times f_2).

The reader should keep in mind that ×\times is not a product in the category FinSetPart, but only in Set.

In terms of the category FinSet *FinSet_*, the monoidal structure is given by the smash product of pointed sets:

A 1A 2=(A 1×A 2)/({*}×A 2A 1×{*}).A_1\otimes A_2=(A_1\times A_2)/(\{*\}\times A_2\cup A_1\times \{*\}).

Applications

A Γ-object in a cartesian monoidal category (more generally, cartesian monoidal (∞,1)-category) CC is a functor

X:Γ opCX\colon \Gamma^{op} \to C

(equivalently, a functor FinSetPartCFinSetPart \to C) such that for all A,BFinSetPartA,B\in FinSetPart the map

X ABX A×X BX_{A\sqcup B}\to X_A\times X_B

induced by the morphisms

(A,id A):ABA,(B,id B):ABB(A,id_A)\colon A\sqcup B\to A, \qquad (B,id_B)\colon A\sqcup B\to B

is an equivalence, as is the map X 1X_\emptyset\to1. Here ABA\sqcup B denotes the coproduct in the category Set (i.e., disjoint union). It does not denote the coproduct in the category FinSetPart.

Γ-objects in CC encode E \mathrm{E}_\infty-monoids in CC, i.e., homotopy coherent commutative monoids. See the article Γ-space for more information.

References

Original reference:

For instance notation 2.0.0.2 in

Last revised on June 12, 2025 at 02:01:42. See the history of this page for a list of all contributions to it.