A pasting theorem for iterated Segal spaces_2

2025-04-27 0 0 384.97KB 43 页 10玖币
侵权投诉
arXiv:2210.04549v2 [math.CT] 23 May 2024
A PASTING THEOREM FOR ITERATED SEGAL SPACES
JACO RUIT
Abstract. We introduce a novel notion of pasting shapes for iterated Segal
spaces which classify particular arrangements of composing cells in d-uple
Segal spaces. Using this formalism, we then continue to prove a pasting theo-
rem for these iterated Segal spaces.
Contents
1. Introduction 1
2. Preliminaries 5
3. Pasting shapes and their nerves 9
4. Proof of the pasting theorem 29
5. Outlook: an (,d)-categorical pasting theorem 40
References 42
1. Introduction
Ad-uple category is a generalization of a d-category, introduced by Ehres-
mann [Ehr63], which has k-cells in d
kdierent directions for 0 kd. In
such a structure, the compatible k-cells that point in the same direction may be
composed, while cells with dierent directions may be related by (k+ 1)-cells.
For d= 2 and d= 3, these are also known as double categories [Ehr63, Definition
10] and intercategories [GP15, Section 1], respectively. These d-uple categories
are the natural place to study mathematical structures that allow for multiple
sorts of morphisms (1-cells) between them.
For instance, in the realm of algebra, one can consider not only the usual
maps between rings but also bimodules between rings. These are the two di-
rections of 1-cells in the Morita double category ([Shu11, Example 2.3]). Since
its introduction, the theory of double categories has found a wide range of ap-
plications throughout category theory. For instance, the theory of 2-categorical
limits admits a neat description using double categories [GP99]. Pseudo 2-
functors that form a proarrow equipment are better understood as being double
categories with additional properties [Ver92], [Shu08]. These proarrow equip-
ments give rise to formal category theories. In [Shu11], we see an application
of the theory of double categories to homotopy theory.
Whereas in a usual category, one may consider strings of compatible mor-
phisms and take their composites, there are now many configurations of com-
patible cells in a d-uple category. For instance, one may start with compatible
1
2 JACO RUIT
2-cells v1,v2,v3,v4,v5,v6in a double category D:
x00 x10 x20 x40
x01 x11 x21 x31 x41
x32 x42
x03 x23 x33 x43,
v1
v4
v3
v2
v5
v6
and wonder: does there exist a unique composite 2-cell vin D? It has been
shown by Dawson and Par´
e [DP93] that (in particular) this arrangement of 2-
cells admits such a composite v. However, not every compatible arrangement of
2-cells in a general double category admits a composite. The authors of [DP93]
established that there is an arrangement that does not have a composite in a
general double category, which is called the pinwheel (we will see this arrange-
ment again in Subsection 3.2), and which is in a particular sense, the canonical
example of such an inadmissible arrangement (see [Daw95]). The procedure
of obtaining new cells by composing compatible arrangements is also called
pasting, which was first introduced by B´
enabou [B´
en67] in the context of 2-
categories.
The practice of pasting cells occurs in the context of many dierent categor-
ical structures. Hence, one would like to have access to a pasting theorem that
asserts the existence and uniqueness of composites for certain configurations of
cells in the categorical structures one considers. A famous pasting theorem for
2-categories was formulated and proven by Power [Pow91]. Nowadays, a wide
range of pasting theorems for (strict) ω-categories are available. Forest [For22]
has recently unified the main pasting theorems in this context to a more gen-
eral pasting theorem for ω-categories. There is a pasting theorem for double
categories due to Dawson and Par´
e [DP93].
The emergence of (weak) -category theory has created a need for variants
of pasting theorems in the weaker setting. In this context, it is no longer natural
to ask for unique composites, but instead require that the space1of composites
is contractible. Lately, Hackney, Ozornova, Riehl and Rovelli [HORR21] have
proven such a pasting theorem for (,2)-categories that generalizes Power’s
pasting theorem for 2-categories. There is an -analog for double categories
as well, so called double -categories, introduced by Haugseng [Hau13] and
further studied by Moser [Mos20]. Thus, one may now ask: does there exist a
pasting theorem for these double -categories?
The goal of this paper is to answer this question armatively. We will treat a
more general pasting problem for d-uple categories in the weak -categorical
setting. Note that this would already be interesting from the strict perspective:
the pasting theorem particularly yields a novel pasting theorem for d-uple cat-
egories. The author does not know of an existing similar result for d > 2.
1By space, we will always mean an -groupoid in this paper.
A PASTING THEOREM FOR ITERATED SEGAL SPACES 3
We will consider an -categorical variant of d-uple categories that are called
d-uple Segal spaces [Hau17]. Ordinary (1-uple) Segal spaces were first intro-
duced by Rezk [Rez01] as a model for -categories. The d-uple Segal spaces
are iterated variants of these that have additional directions for morphisms. For
instance, a 2-uple Segal space Xcontains a space of objects, and between any
two objects, a space of vertical and horizontal arrows. Thus a 2-uple Segal space
has two categorical directions instead of merely one as is the case for an ordi-
nary Segal space. Compatible arrows of Xthat have the same direction, can be
composed in a coherently associative fashion. Moreover, Xcontains a space of
2-cells. A 2-cell may be pictured as a square
a b
c d
in X. Here the arrows that point horizontally are horizontal arrows of Xand
similarly for the vertical ones. That is, 2-cells have a source and target verti-
cal arrow and a source and target horizontal arrow. Again, Xhas a coherently
associative composition for these 2-cells, which is compatible with the composi-
tion of 1-cells. In general, d-uple Segal spaces contain dcategorical directions,
which may interact coherently using higher cells.
These d-uple Segal spaces are rich structures that play a useful role in -
category theory. For instance, by ‘truncating’ all but one of the categorical
directions, they can be used to model (,d)-categories. This is the model for
(,d)-categories due to Barwick [Bar05], which we will briefly discuss in Section 5.
It has been (directly) compared to other models for (,d)-categories by Bergner
and Rezk in [BR13] and [BR20], and Loubaton, Ozornova and Rovelli in [Lou22]
and [OR22]. Consequently, d-uple Segal spaces may act as a useful interme-
diate step towards constructing (,d)-categories. For instance, in order to
construct (,d)-categories of iterated spans or to construct the Morita (,d)-
categories it is more convenient to define their encompassing d-uple Segal spaces
first (see [Hau18] and [Hau17]). We also believe that d-uple Segal spaces can
be used to study phenomena of (,d)-categories. For instance, the universal
property of the (,2)-category of spans [EH23] should be the shadow of a uni-
versal property of 2-uple Segal space of spans that is analogous to one in the
strict case [DPP10]. We hope to study this in future work.
This work grew out of the author’s study of double -categories, which can
be viewed as 2-uple Segal spaces that satisfy a completeness assumption (see
Remark 2.9). More specifically, out of the interest in a particular class of dou-
ble -categories: so-called -equipments, which are generalizations of proar-
row equipments to the -categorical context. These -equipments oer a con-
text for synthetic or formal category theory. There exist suitable -equipments
for equivariant, indexed, internal, fibered, enriched and ordinary -category
theory. We commenced the study of these -equipments in [Rui23].
Content of the paper. We will commence the paper by setting up the necessary
preliminaries, including revisiting the definition of iterated Segal spaces, in
Section 2.
4 JACO RUIT
Subsequently, in Section 3, we introduce the novel notion of d-dimensional
pasting shapes whose nerves classify arrangements of cells in a d-uple Segal
space. For d= 3, these include arrangements of rectangular cuboids, whose
faces may be subdivided into smaller rectangles. The faces of these rectangles
may in turn be subdivided into smaller edges. We will also discuss some funda-
mental properties that pasting shapes may have. In particular, we single out an
important class of pasting shapes: the so-called (locally) composable ones. The
composable pasting shapes are certain well-behaved pasting shapes. For d= 2,
the aforementioned pinwheel of Dawson and Par´
e is an example of a pasting
shape that is not composable.
After setting up the theory of pasting shapes, we can formulate the main
result of this paper, the pasting theorem for d-uple Segal spaces:
Theorem A (The pasting theorem).Suppose that I1,...,Inis a covering of a d-
dimensional locally composable pasting shape I. Then Ican be written as a union
I=
n
[
i=1
Ii,
and this colimit description is preserved when passing to the -category of d-uple
Segal spaces via the nerve functor for pasting shapes.
The precise statement appears as Theorem 3.49 in the paper. This theorem can
be used to show Corollary 3.52, which states that the spine inclusion associated
with a composable pasting shape is an equivalence. The composable pasting
shapes thus classify arrangements of cells that admit a composite that is unique
up to contractible choice (see Corollary 3.55).
The technical heart of the paper is in Section 4. Here, we proceed by induc-
tion on the dimension of pasting shapes, to give a proof of Theorem A. After
this demonstration, we conclude the paper in Section 5 by giving an idea of
how our pasting theorem for d-uple Segal spaces may yield a pasting theorem
for (,d)-categories.
Conventions. We will use the language of -categories throughout this article.
For definiteness, we work with the model of quasi-categories for -categories
as developed by Joyal and Lurie [Lur09a]. The readers that prefer to use the lan-
guage of model categories, may interpret the main results of this paper within
the appropriate model categories associated with the -categories in question.
We will use the following customary notation:
The (large) -categories of spaces [Lur09a, Definition 1.2.16.1] and -
categories [Lur09a, Definition 3.0.0.1] are denoted by Sand Cat, re-
spectively.
If Cis an -category, we write
MapC(,) : Cop ×CS
for its associated mapping space functor [Lur09a, Subsection 5.1.3].
Every 1-category Cwill be viewed as an -category, suppressing the
notation of the nerve [Lur09a, Subsection 1.1.2].
A PASTING THEOREM FOR ITERATED SEGAL SPACES 5
Acknowledgements. I want to thank my PhD-supervisor, Lennart Meier, for
the helpful conversations during the writing of this paper and his useful com-
ments on the draft versions. Furthermore, I would like to express my gratitude
to the anonymous referee for the useful suggestions that greatly improved the
presentation of this article.
During the writing of this paper, the author was funded by the Dutch Re-
search Council (NWO) through the grant “The interplay of orientations and
symmetry”, grant no. OCENW.KLEIN.364.
2. Preliminaries
A clear definition of the (strict) d-uple categories introduced by Ehresmann
[Ehr63] can be given in terms of Grothendieck’s notion of categorical objects
[Gro95]. Following [Hau17] and [Hau18], we will introduce an -categorical
variant on d-uple categories via the same principle.
2.1. Categorical objects. To this end, we first need the notion of categorical ob-
jects in an -category. Throughout this section, we will fix an -category Cand
assume that it has all pullbacks. The following definition is due to [Lur09b]:
Definition 2.1. Acategorical object Xin Cis a simplicial object X:op Csuch
that the so-called Segal map
X([n]) X({01})×X({1})·· · ×X({n1})X({n1n})
is an equivalence for all n. The full subcategory of Fun(op,C) spanned by the
categorical objects in Cis denoted by Cat(C).
Example 2.2. We have the following examples:
A categorical object in the (2,1)-category of categories is a pseudo dou-
ble category. These were defined by Grandis and Par´
e [GP99].
If Cis given by the -category of spaces S, then Cat(S) is the -category
that underlies the model category of Segal spaces, constructed by Rezk in
[Rez01, Theorem 7.1]. Henceforth, we will refer to categorical objects
in Sas Segal spaces.
A categorical object in Catis called a double -category. These were
first studied by Haugseng in [Hau13].
Note that the subcategory Cat(C)Fun(op,C) is closed under limits, so that
limits of categorical objects may be computed pointwise. In particular, we de-
duce that Cat(C) again admits all pullbacks. Thus, we may iterate Definition 2.1:
Definition 2.3. We define the -category of d-uple categorical objects in Cby
Catd(C) := Cat(···Cat(Cat(C))···).
By adjunction, the -category of d-uple categorical objects in Ccan be de-
scribed as a full subcategory of the -category
Fun(op,×d,C)
of d-uple simplicial objects in C. In the case that Cis presentable, then Catd(C) is
a (left) reflective subcategory of Fun(op,×d,C). I.e. the inclusion admits a left
adjoint (see also [Lur09a, Remark 5.2.7.9]) in this case. We will demonstrate
how this can be established.
摘要:

arXiv:2210.04549v2[math.CT]23May2024APASTINGTHEOREMFORITERATEDSEGALSPACESJACORUITAbstract.WeintroduceanovelnotionofpastingshapesforiteratedSegalspaceswhichclassifyparticulararrangementsofcomposingcellsind-upleSegalspaces.Usingthisformalism,wethencontinuetoproveapastingtheo-remfortheseiteratedSegalsp...

展开>> 收起<<
A pasting theorem for iterated Segal spaces_2.pdf

共43页,预览5页

还剩页未读, 继续阅读

声明:本站为文档C2C交易模式,即用户上传的文档直接被用户下载,本站只是中间服务平台,本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私,请立即通知玖贝云文库,我们立即给予删除!
分类:图书资源 价格:10玖币 属性:43 页 大小:384.97KB 格式:PDF 时间:2025-04-27

开通VIP享超值会员特权

  • 多端同步记录
  • 高速下载文档
  • 免费文档工具
  • 分享文档赚钱
  • 每日登录抽奖
  • 优质衍生服务
/ 43
客服
关注