-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathcart-models.tex
566 lines (501 loc) · 33 KB
/
cart-models.tex
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
\documentclass[reqno]{amsart}
\usepackage{amssymb}
\usepackage{hyperref}
\usepackage{mathtools}
\usepackage[all]{xy}
\usepackage{ifthen}
\usepackage{xargs}
\usepackage{bussproofs}
\usepackage{verbatim}
\usepackage{turnstile}
\hypersetup{colorlinks=true,linkcolor=blue}
\renewcommand{\turnstile}[6][s]
{\ifthenelse{\equal{#1}{d}}
{\sbox{\first}{$\displaystyle{#4}$}
\sbox{\second}{$\displaystyle{#5}$}}{}
\ifthenelse{\equal{#1}{t}}
{\sbox{\first}{$\textstyle{#4}$}
\sbox{\second}{$\textstyle{#5}$}}{}
\ifthenelse{\equal{#1}{s}}
{\sbox{\first}{$\scriptstyle{#4}$}
\sbox{\second}{$\scriptstyle{#5}$}}{}
\ifthenelse{\equal{#1}{ss}}
{\sbox{\first}{$\scriptscriptstyle{#4}$}
\sbox{\second}{$\scriptscriptstyle{#5}$}}{}
\setlength{\dashthickness}{0.111ex}
\setlength{\ddashthickness}{0.35ex}
\setlength{\leasturnstilewidth}{2em}
\setlength{\extrawidth}{0.2em}
\ifthenelse{%
\equal{#3}{n}}{\setlength{\tinyverdistance}{0ex}}{}
\ifthenelse{%
\equal{#3}{s}}{\setlength{\tinyverdistance}{0.5\dashthickness}}{}
\ifthenelse{%
\equal{#3}{d}}{\setlength{\tinyverdistance}{0.5\ddashthickness}
\addtolength{\tinyverdistance}{\dashthickness}}{}
\ifthenelse{%
\equal{#3}{t}}{\setlength{\tinyverdistance}{1.5\dashthickness}
\addtolength{\tinyverdistance}{\ddashthickness}}{}
\setlength{\verdistance}{0.4ex}
\settoheight{\lengthvar}{\usebox{\first}}
\setlength{\raisedown}{-\lengthvar}
\addtolength{\raisedown}{-\tinyverdistance}
\addtolength{\raisedown}{-\verdistance}
\settodepth{\raiseup}{\usebox{\second}}
\addtolength{\raiseup}{\tinyverdistance}
\addtolength{\raiseup}{\verdistance}
\setlength{\lift}{0.8ex}
\settowidth{\firstwidth}{\usebox{\first}}
\settowidth{\secondwidth}{\usebox{\second}}
\ifthenelse{\lengthtest{\firstwidth = 0ex}
\and
\lengthtest{\secondwidth = 0ex}}
{\setlength{\turnstilewidth}{\leasturnstilewidth}}
{\setlength{\turnstilewidth}{2\extrawidth}
\ifthenelse{\lengthtest{\firstwidth < \secondwidth}}
{\addtolength{\turnstilewidth}{\secondwidth}}
{\addtolength{\turnstilewidth}{\firstwidth}}}
\ifthenelse{\lengthtest{\turnstilewidth < \leasturnstilewidth}}{\setlength{\turnstilewidth}{\leasturnstilewidth}}{}
\setlength{\turnstileheight}{1.5ex}
\sbox{\turnstilebox}
{\raisebox{\lift}{\ensuremath{
\makever{#2}{\dashthickness}{\turnstileheight}{\ddashthickness}
\makehor{#3}{\dashthickness}{\turnstilewidth}{\ddashthickness}
\hspace{-\turnstilewidth}
\raisebox{\raisedown}
{\makebox[\turnstilewidth]{\usebox{\first}}}
\hspace{-\turnstilewidth}
\raisebox{\raiseup}
{\makebox[\turnstilewidth]{\usebox{\second}}}
\makever{#6}{\dashthickness}{\turnstileheight}{\ddashthickness}}}}
\mathrel{\usebox{\turnstilebox}}}
\newcommand{\newref}[4][]{
\ifthenelse{\equal{#1}{}}{\newtheorem{h#2}[hthm]{#4}}{\newtheorem{h#2}{#4}[#1]}
\expandafter\newcommand\csname r#2\endcsname[1]{#3~\ref{#2:##1}}
\expandafter\newcommand\csname R#2\endcsname[1]{#4~\ref{#2:##1}}
\expandafter\newcommand\csname n#2\endcsname[1]{\ref{#2:##1}}
\newenvironmentx{#2}[2][1=,2=]{
\ifthenelse{\equal{##2}{}}{\begin{h#2}}{\begin{h#2}[##2]}
\ifthenelse{\equal{##1}{}}{}{\label{#2:##1}}
}{\end{h#2}}
}
\newref[section]{thm}{Theorem}{Theorem}
\newref{lem}{Lemma}{Lemma}
\newref{prop}{Proposition}{Proposition}
\newref{cor}{Corollary}{Corollary}
\newref{cond}{Condition}{Condition}
\newref{conj}{Conjecture}{Conjecture}
\theoremstyle{definition}
\newref{defn}{Definition}{Definition}
\newref{example}{Example}{Example}
\theoremstyle{remark}
\newref{rem}{Remark}{Remark}
\newcommand{\red}{\Rightarrow}
\newcommand{\deq}{\equiv}
\newcommand{\repl}{:=}
\newcommand{\idtype}{\rightsquigarrow}
\newcommand{\Id}{\mathrm{Id}}
\newcommand{\cat}[1]{\mathbf{#1}}
\newcommand{\C}{\cat{C}}
\newcommand{\D}{\cat{D}}
\newcommand{\Cat}{\cat{Cat}}
\newcommand{\Mod}[1]{#1\text{-}\cat{Mod}}
\newcommand{\Th}{\cat{Th}}
\newcommand{\emptyCtx}{\mathbf{1}}
\newcommand{\Set}{\cat{Set}}
\newcommand{\sSet}{\cat{sSet}}
\newcommand{\qcat}{\cat{qcat}}
\newcommand{\K}{$\mathcal{K}$}
\newcommand{\join}{\star}
\newcommand{\Hom}{\mathrm{Hom}}
\newcommand{\cmap}[1]{\mathrm{c}_{#1}}
\newcommand{\nf}{\mathrm{nf}}
\newcommand{\M}{H}
\newcommand{\we}{\mathcal{W}}
\newcommand{\I}{\mathrm{I}}
\newcommand{\J}{\mathrm{J}}
\newcommand{\class}[2]{#1\text{-}\mathrm{#2}}
\newcommand{\Icell}[1][\I]{\class{#1}{cell}}
\newcommand{\Icof}[1][\I]{\class{#1}{cof}}
\newcommand{\Iinj}[1][\I]{\class{#1}{inj}}
\newcommand{\Jinj}[1][]{\Iinj[\J#1]}
\newcommand{\Jcell}[1][]{\Icell[\J#1]}
\newcommand{\Jcof}[1][]{\Icof[\J#1]}
\numberwithin{figure}{section}
\newcommand{\pb}[1][dr]{\save*!/#1-1.2pc/#1:(-1,1)@^{|-}\restore}
\newcommand{\po}[1][dr]{\save*!/#1+1.2pc/#1:(1,-1)@^{|-}\restore}
\begin{document}
\title[Equivalence between quasicategories and models of type theory]{Equivalence between finitely complete quasicategories and models of simplicial homotopy type theory}
\author{Valery Isaev}
\begin{abstract}
Abstract.
\end{abstract}
\maketitle
\section{Introduction}
\section{Functors between $\sSet$ and $\Mod{T}$}
\label{sec:nerve}
In this section we define a Quillen adjunction between $\sSet$ and $\Mod{T}$ for every appropriate algebraic dependent type theory $T$.
Recall that $sq_l$ is the theory with the following axioms:
\medskip
\begin{center}
\AxiomC{$\Gamma \vdash i : I$}
\AxiomC{$\Gamma \vdash j : I$}
\BinaryInfC{$\Gamma \vdash sq_l(i,j) : I$}
\DisplayProof
\end{center}
\begin{align*}
sq_l(left,j) & = left \\
sq_l(right,j) & = j \\
sq_l(i,left) & = left \\
sq_l(i,right) & = i
\end{align*}
To define functors between $\sSet$ and $\Mod{T}$, we also need to assume that $sq_l$ is associative.
Let $sq^a_l$ be the theory which has the same axioms as $sq_l$ together with the following axiom:
\[ sq_l(sq_l(i,j),k) = sq_l(i,sq_l(j,k)) \]
Let $T$ be any theory under $sq^a_l$.
For every finite nonempty linearly ordered set $J$, we define a model $\mathfrak{C}(\Delta^J)$ of $T$.
It is generated by the generators and relations described below.
Generators of $\mathfrak{C}(\Delta^J)$ are $O_j : (ty,0)$ for every $j \in J$, and $\M_{J'} : (tm,|J'|-1)$ for every subset $J'$ of $J$ such that $|J'| \geq 2$.
For every $j \in J$, $\mathfrak{C}(\Delta^J)$ has relation $\vdash O_j\ type$.
For every $J' = \{ j_1 < \ldots < j_n \} \subseteq J$ such that $n \geq 2$, it has the following relation:
\[ x : O_{j_1}, x_{j_2} : I, \ldots x_{j_{n-1}} : I \vdash \M_{J'} : O_{j_n} \]
For every $j \notin J'$ such that $j_1 < j < j_n$, it has the following relations:
\begin{align}
x : O_{j_1}, x_{j_2} : I, \ldots x_{j_{n-1}} : I & \vdash \M_{J' \cup \{j\}}[x_j \repl left] \deq \M_{\{ j, \ldots j_n \}}[x \repl \M_{\{ j_1, \ldots j \}}] \label{rel:left} \\
x : O_{j_1}, x_{j_2} : I, \ldots x_{j_{n-1}} : I & \vdash \M_{J' \cup \{j\}}[x_j \repl right] \deq \M_{J'} \label{rel:right}
\end{align}
The idea of this definition is similar to \cite[Definition~1.1.5.1]{lurie-topos}.
For every vertex $j$ of $\Delta^J$, we have a type $O_j$,
and for every $(n+1)$-dimensional face $J'$ of $\Delta^J$, we have an $n$-dimensional cube $\M_{J'}$.
Relations describe faces of these cubes.
Let $sq_l(x_1, \ldots x_n) = sq_l(x_1, \ldots sq_l(x_{n-1},x_n) \ldots )$.
In particular, $sq_l(x) = x$ and $sq_l() = right$.
Now, we can extend $\mathfrak{C}$ to a functor $\Delta \to \Mod{T}$.
Let $f : J \to K$ be a monotone map between linearly ordered sets.
Then we define $\mathfrak{C}(f)$ as follows:
\begin{align*}
\mathfrak{C}(f)(O_j) & = O_{f(j)} \\
\mathfrak{C}(f)(\M_{J'}) & =
\begin{cases}
x & \text{if } |f(J')| = 1 \\
\M_{f(J')}[\ \ldots\ x_{k_i} \repl sq_l(x_{f^{-1}(k_i)})\ \ldots\ ] & \text{if } |f(J')| > 1 \\
\end{cases}
\end{align*}
where $k_i$ runs from $k_2$ to $k_{m-1}$, $\{ k_1 < \ldots < k_m \} = f(J')$,
and $x_{f^{-1}(k_i)} = x_{j_1}, \ldots x_{j_n}$, $\{ j_1 < \ldots < j_n \} = f^{-1}(k_i)$.
Let us check that $\mathfrak{C}(f)$ preserves relations.
Let $J' = \{ j_1 < \ldots < j_i < j < j_{i+1} < \ldots < j_n \}$ be a subset of $J$, and let $f'$ be the restriction of $f$ to $J'$.
If $|f(J')| = 1$, then $\mathfrak{C}(f)(\M_{J' \cup \{j\}}[x_j \repl left]) = \mathfrak{C}(f)(\M_{\{ j, \ldots j_n \}}[x \repl \M_{\{ j_1, \ldots j \}}]) = \mathfrak{C}(f)(\M_{J' \cup \{j\}}[x_j \repl right]) = \mathfrak{C}(f)(\M_{J'}) = x$.
Thus we may assume that $|f(J')| > 1$.
Let $\rho(x_{f(j_i)}) = sq_l(x_{f'^{-1}(f(j_1))})$ for every $1 \leq i \leq n$ such that $f(j_i) \neq f(j)$.
Let us consider the first relation.
If $f(j) = f(j_1)$, then
\begin{align*}
\mathfrak{C}(f)(\M_{J' \cup \{j\}}[x_j \repl left]) & = \\
\M_{f(J')}[\rho] & = \\
\M_{\{ f(j), \ldots f(j_n) \}}[\rho][x \repl x] & = \\
\mathfrak{C}(f)(\M_{\{ j, \ldots j_n \}}[x \repl \M_{\{ j_1, \ldots j \}}]) & .
\end{align*}
If $f(j) = f(j_n)$, then
\begin{align*}
\mathfrak{C}(f)(\M_{J' \cup \{j\}}[x_j \repl left]) & = \\
\M_{f(J')}[\rho] & = \\
x[x \repl \M_{\{f(j_1), \ldots f(j)\}}[\rho]] & = \\
\mathfrak{C}(f)(\M_{\{ j, \ldots j_n \}}[x \repl \M_{\{ j_1, \ldots j \}}]) & .
\end{align*}
If $f(j_1) < f(j) < f(j_n)$, then
\begin{align*}
\mathfrak{C}(f)(\M_{J' \cup \{j\}}[x_j \repl left]) & = \\
\M_{f(J') \cup \{f(j)\}}[\rho, x_{f(j)} \repl sq_l(x_{f'^{-1}(f(j))}, x_j)][x_j \repl left] & = \\
\M_{f(J') \cup \{f(j)\}}[\rho, x_{f(j)} \repl left] & = \\
\M_{\{f(j), \ldots f(j_n)\}}[x \repl \M_{\{f(j_1), \ldots f(j)\}}][\rho] & = \\
\mathfrak{C}(f)(\M_{\{ j, \ldots j_n \}}[x \repl \M_{\{ j_1, \ldots j \}}]) & .
\end{align*}
Let us consider the second relation.
If either $f(j) = f(j_1)$ or $f(j) = f(j_n)$, then $\mathfrak{C}(f)(\M_{J' \cup \{j\}}[x_j \repl right]) = \M_{f(J')}[\rho] = \mathfrak{C}(f)(\M_{J'})$.
If $f(j_1) < f(j) < f(j_n)$, then
\begin{align*}
\mathfrak{C}(f)(\M_{J' \cup \{j\}}[x_j \repl right]) & = \\
\M_{f(J') \cup \{f(j)\}}[\rho, x_{f(j)} \repl sq_l(x_{f'^{-1}(f(j))}, x_j)][x_j \repl right] & = \\
\M_{f(J') \cup \{f(j)\}}[\rho, x_{f(j)} \repl sq_l(x_{f'^{-1}(f(j))})] & = \\
\mathfrak{C}(f)(\M_{J'}) & ,
\end{align*}
If $f(j_i) < f(j) < f(j_{i+1})$, then the last equality holds by \eqref{rel:right},
otherwise it holds immediately by definition of $\mathfrak{C}(f)(\M_{J'})$.
Let $[n]$ be the set of natural numbers $\{ 0, \ldots n \}$ with the obvious linear order.
Then $\Delta$ is the full subcategory of linearly ordered sets on objects of the form $[n]$.
It is easy to see that $\mathfrak{C}$ preserves identity morphisms,
and the fact that it preserves compositions follows from associativity of $sq_l$.
Thus we have a functor $\mathfrak{C} : \Delta \to \Mod{T}$.
This functor (uniquely) extends to a colimit-preserving functor $\mathfrak{C} : \sSet \to \Mod{T}$,
which has a right adjoint $N : \Mod{T} \to \sSet$ defined by equation $N(X) = \Hom(\mathfrak{C}(-),X)$.
We can explicitly describe $\mathfrak{C} : \sSet \to \Mod{T}$.
Model $\mathfrak{C}(X)$ is generated by symbols $O_a : (ty,0)$ for every $a \in X_0$, and $\M_a : (tm,n)$ for every $a \in X_n$, $n \geq 1$.
For every $a \in X_0$, $\mathfrak{C}(X)$ has relations $\vdash O_a\ type$ and $\M_{s_0(a)} = (x : O_a \vdash x : O_a)$.
For every $a \in X_n$, $n \geq 1$, it has the following relation:
\[ x : O_{a|\Delta^{\{0\}}}, x_1 : I, \ldots x_{n-1} : I \vdash \M_a : O_{a|\Delta^{\{n\}}} \]
For every $1 \leq j \leq n-1$, it has the following relations:
\begin{align*}
\M_a[x_j \repl left] & = \M_{a|\Delta^{\{ j, \ldots n \}}}[x \repl \M_{a|\Delta^{\{ 0, \ldots j \}}}] \\
\M_a[x_j \repl right] & = \M_{a|\Delta^{\{0, \ldots j-1, j+1, \ldots n\}}}
\end{align*}
For every $0 \leq j \leq n$, it has the following relation:
\[ \M_{s_j(a)} =
\begin{cases}
\M_a[x_1 \repl x_2, \ldots x_{n-1} \repl x_n] & \text{if } j = 0 \\
\M_a[x_j \repl sq_l(x_j,x_{j+1}), x_{j+1} \repl x_{j+2}, \ldots x_{n-1} \repl x_n] & \text{if } 1 \leq j \leq n-1 \\
\M_a & \text{if } j = n \\
\end{cases}
\]
For every morphism $f : X \to Y$ of simplicial sets, let $\mathfrak{C}(f)(O_a) = O_{f(a)}$ and $\mathfrak{C}(f)(\M_a) = \M_{f(a)}$.
It is easy to see that this definition preserves relations, identity morphisms and compositions.
\begin{prop}
Functors $\mathfrak{C} \dashv N$ determine a Quillen adjunction between $\sSet$ with the Joyal model structure and $\Mod{T}$ with the model structure defined in \cite{alg-models}.
\end{prop}
\begin{proof}
Note that $\mathfrak{C}(\partial \Delta^n)$ is isomorphic (over $\mathfrak{C}(\Delta^n)$) to the theory
generated by the same generators and relations as $\mathfrak{C}(\Delta^n)$ except for $\M_{[n]}$.
Similarly, if $i \in [n]$, then $\mathfrak{C}(\Lambda^n_i)$ is generated by the same generators
and relations as $\mathfrak{C}(\partial \Delta^n)$ except for $\M_{\{ 0, \ldots i-1, i+1, \ldots n \}}$.
Let us prove that $\mathfrak{C}$ preserves cofibration.
To do this, it is enough to show that $\mathfrak{C}(\partial \Delta^n) \to \mathfrak{C}(\Delta^n)$ is a cofibration for every $n$.
If $n = 0$, then this map equals to $i_{(ty,0)}$.
Suppose that $n > 0$.
First, let us define objects $\square^k$ and $\partial \square^k$.
Let $\square^k = F(\{ x : A, x_1 : I, \ldots x_k : I \vdash b : B \})$,
and let $\partial \square^k$ be generated by the generators and relations described below.
Generators of $\partial \square^k$ are $A,B : (ty,0)$ and $b_{[i=c]} : (tm,k)$ for every $1 \leq i \leq k$, $c \in \{left,right\}$.
For every $1 \leq i \leq k$ and $c \in \{left,right\}$, it has the following relation:
\[ x : A, x_1 : I, \ldots x_{i-1} : I, x_{i+1} : I, \ldots x_k : I \vdash b_{[i=c]} : B \]
For every $1 \leq i_1 < i_2 \leq k$ and $c,c' \in \{left,right\}$, it has the following relation:
\[ b_{[i_2=c']}[x_{i_1} \repl c] = b_{[i_1=c]}[x_{i_2} \repl c'] \]
Map $i^k : \partial \square^k \to \square^k$ is defined in the obvious way: $i^k(b_{[i=c]}) = b[x_i \repl c]$.
Let $v : \square^{n-1} \to \mathfrak{C}(\Delta^n)$ be defined as follows: $v(A) = O_0$, $v(B) = O_n$, $v(b) = \M_{[n]}$.
Map $v \circ i^k$ factors through $\mathfrak{C}(\partial \Delta^n) \to \mathfrak{C}(\Delta^n)$, and the following square is cocartesian:
\[ \xymatrix{ \partial \square^{n-1} \ar[r] \ar[d]_{i^k} & \mathfrak{C}(\partial \Delta^n) \ar[d] \\
\square^{n-1} \ar[r]_v & \mathfrak{C}(\Delta^n)
} \]
Thus we just need to show that $i^k$ is a cofibration.
But it is a pushout of $i_{(tm,1)} : F(\{ x : A \vdash B\ type \}) \to F(\{ x : A \vdash b : B \})$.
Indeed, let $v' : F(\{ x : A \vdash b : B \}) \to \square^k$ be defined as follows: $v'(b) = path(x_1.\,\ldots path(x_n.\,b) \ldots)$ and $v'(B)$ equals to
\begin{align*}
Path(x_1.\, \ldots Path(x_{n-2}.\, & Path(x_{n-1}.\, Path(x_n.\,B, b[x_n \repl left], b[x_n \repl right]), \\
& \qquad \qquad \quad path(x_n.\,b[x_{n-1} \repl left]), \\
& \qquad \qquad \quad path(x_n.\,b[x_{n-1} \repl right])), \\
& path(x_{n-1}.\, path(x_n.\,b[x_{n-2} \repl left])), \\
& path(x_{n-1}.\, path(x_n.\,b[x_{n-2} \repl right]))) \ldots )
\end{align*}
Then $v' \circ i_{(tm,1)}$ factors through $i^k$, and the following square is cocartesian:
\[ \xymatrix{ F(\{ x : A \vdash B\ type \}) \ar[r] \ar[d]_{i_{(tm,1)}} & \partial \square^k \ar[d]^{i^k} \\
F(\{ x : A \vdash b : B \}) \ar[r]_-{v'} & \square^k
} \]
Now, let us prove that for every $X \in \Mod{T}$, $N(X)$ is fibrant.
Since fibrant objects in $\sSet$ are precisely quasicategories,
it is enough to show that $\mathfrak{C}(\Lambda^n_i) \to \mathfrak{C}(\Delta^n)$ is a trivial cofibration for every inner horn $\Lambda^n_i \to \Delta^n$.
First, for every $1 \leq i \leq k$, let $\sqcap^k_i$ be the object defined by the same generators and relations as $\partial \square^k$ except for $b_{[i=right]}$.
Then we have the following cocartesian square:
\[ \xymatrix{ \sqcap^{n-1}_i \ar[r] \ar[d] & \mathfrak{C}(\Lambda^n_i) \ar[d] \\
\partial \square^{n-1} \ar[r] & \mathfrak{C}(\partial \Delta^n)
} \]
Since $\sqcap^k_i \to \square^k$ is a cofibration, we just need to show that it is a weak equivalence.
Note that $\sqcap^k_i \to \square^k$ is isomorphic to $\sqcap^k_k \to \square^k$ for every $1 \leq i \leq k$.
Indeed, we can define isomorphism $g : \square^k \to \square^k$ as $g(b) = b[x_i \repl x_k, x_k \repl x_i]$
and isomorphism $f : \sqcap^k_i \to \sqcap^k_k$ as $f(b_{[i=left]}) = b_{[k=left]}[x_i \repl x_k]$, $f(b_{[k=c]}) = b_{[i=c]}[x_k \repl x_i]$,
and $f(b_{[j=c]}) = b_{[j=c]}[x_i \repl x_k, x_k \repl x_i]$ for every $j \neq i,k$.
Then $f$ and $g$ commute with $\sqcap^k_i \to \square^k$ and $\sqcap^k_k \to \square^k$.
Thus we just need to prove that $\sqcap^k_k \to \square^k$ is a weak equivalence.
Note that $\square^{k+1}$ is a (relative) cylinder object for $F(\{ A : (ty,0), B : (ty,0) \}) \to \square^k$.
Indeed, $\square^k \amalg_{F(\{ A : (ty,0), B : (ty,0) \})} \square^k \to \square^{k+1}$ is a cofibration,
and $\square^k \to \square^{k+1}$ is a weak equivalence since it is a pushout of
a generating trivial cofibration $F(\{ \Gamma \vdash b : B \}) \to F(\{ \Gamma, y : I \vdash h : B \})$.
Every object has RLP with respect to $\sqcap^k_k \to \square^k$ (see \cite[Section~2.4]{alg-models}).
In particular, $\sqcap^k_k \to \square^k$ has a retract $\square^k \to \sqcap^k_k$.
The composite map $\square^k \to \sqcap^k_k \to \square^k$ is homotopic to $id$.
To show this, we just need to construct an appropriate map $\square^{k+1} \to \square^k$,
which is easy to do using filler operations from \cite[Section~2.4]{alg-models}.
Thus $\sqcap^k_k \to \square^k$ is a homotopy equivalence.
Finally, let us prove that if $f : X \to Y$ is a fibration in $\Mod{T}$, then $N(f)$ is a fibration in $\sSet$.
Since $N(Y)$ is fibrant, by \cite[Corollary~2.4.6.5]{lurie-topos}, we just need to prove the following conditions:
\begin{enumerate}
\item $N(f)$ is an inner fibration.
\item For every object $A$ of $N(X)$ and every equivalence $p : N(f)(A) \to B$ in $N(Y)$, there exists an equivalence $p' : A \to B'$ in $N(X)$ such that $N(f)(p') = p$.
\end{enumerate}
The first condition follows from the fact that maps $\mathfrak{C}(\Lambda^n_i) \to \mathfrak{C}(\Delta^n)$ are trivial cofibrations.
The second condition can be reformulated as follows: for every closed type $A$ of $X$ and every equivalence $p : f(A) \to B$ in $Y$,
there exists an equivalence $p' : A \to B'$ in $X$ such that $f(p') = p$.
But this follows from the fact that $f$ has RLP with respect to $\J_{\I_{ty}}$.
\end{proof}
\section{Slice quasicategories}
Let $T$ be a stable theory with $\Sigma$-types with eta rules.
Then for every model $M$ of $T$ and every $A \in M_{(ty,0)}$, we can define another model $M/A$ of $T$.
Let $(M/A)_{(p,n)} = \{ x \in M_{(p,n+1)}\ |\ ft^n(x) = A \}$.
Since $T$ is stable, the structure of a model of $T$ on $M$ induces this structure on $M/A$.
In this section we construct a categorical equivalence $p : N(M/A) \to N(M)/A$ between the nerve of $M/A$ and the slice quasicategory of $N(M)$ over $A$.
We also prove that $N(M)$ has finite limits.
We begin with terminal objects.
A type $C \in M_{(ty,n)}$ is a \emph{proposition} if there is a term $\Gamma, x : C, y : C \vdash p : x \idtype y$.
It is \emph{contractible} if it is a proposition and there is a term $\Gamma \vdash c : C$.
\begin{prop}[contr-terminal]
Let $M$ be a model of a dependent type theory.
Then $C \in M_{(ty,0)}$ is a contractible type of $M$ if and only if $C$ is a terminal object of $N(M)$.
Moreover, if $C$ is contractible, then, for every type $A$, the weakening of $C$ is a terminal object of $N(M/A)$.
\end{prop}
\begin{proof}
An object $C$ of $N(M)$ is terminal if and only if every map $f : \partial \Delta^n \to N(M)$ such that $f|_{\Delta^{\{n\}}} = C$ extends to a map $\Delta^n \to N(M)$.
This is true if and only if every map $f : \partial \square^n \to M$ such that $f(B) = C$ extends to a map $\square^n \to M$.
Thus, if $C$ is a terminal object of $N(M)$, then we have a term $x : I \vdash b : C$ and hence a term $\vdash c : C$.
Moreover, we have a term $x : C \vdash p : x \idtype c$, which implies that $C$ is contractible.
Conversely, assume that $C \in M_{(ty,n)}$ is a proposition.
Then a standard argument implies that the type $\Gamma, x : C, y : C \vdash x \idtype y$ is also a proposition.
Since $Path(x.\,A,a_1,a_2)$ is equivalent to a type of the form $a_1' \idtype a_2$,
this implies types of the form $Path(x_n.\, \ldots Path(x_1.\,C,c_1,c_1'), \ldots, c_n, c_n')$ are inhabited.
An element of such a type determines an extension of a map $\partial \square^n \to M$ determined by terms $c_1,c_1', \ldots c_n,c_n'$.
Thus, every such map has an extension to a map $\square^n \to M$.
Finally, if $C$ is contractible, then its weakening is contractible type.
Hence, it is a terminal object of $N(M/A)$.
\end{proof}
Let $C$ be a quasicategory, and let $A \in C_{[0]}$ be an object of $C$.
Then $C/A$ is the quasicategory defined as follows: $(C/A)_{[n]} = \{ x \in C_{[n+1]}\ |\ x|_{\Delta^{\{n+1\}}} = A \}$.
For every monotone map $f : [k] \to [n]$, let $(C/A)_f(x) = C_{f'}(x)$, where $f' : [k+1] \to [n+1]$ is the map such that $f'(k) = n$ and $f'(i) = f(i)$ for every $0 \leq i \leq k$.
For a proof that this simplicial set is a quasicategory, see for instance \cite[Proposition~1.2.9.3]{lurie-topos}.
Elements of $N(M)_{[n]}$ can be represented by pairs $(O,H)$, where $O$ is a function that maps elements of $[n]$ to $N(M)_{(ty,0)}$
and $H$ is a function that maps a subset $J = \{ j_1 < \ldots < j_k \}$ of $[n]$ such that $k \geq 2$ to a term of the form $x : O_{j_1}, x_{j_2} : I, \ldots x_{j_{k-1}} : I \vdash H_J : O_{j_k}$.
These functions must satisfy relations described in the previous section.
Let $(O,H)$ be an element of $N(M/A)_{[n]}$.
Then $O_i \in M_{(ty,1)}$ is such that $z : A \vdash O_i\ type$ and $H_J$ can be identified with an element of $M_{(tm,k)}$ such that $z : A, x : O_{j_1}, x_{j_2} : I, \ldots x_{j_{k-1}} : I \vdash H_J : O_{j_k}$.
To define an element $p(O,H) \in (N(M)/A)_{[n]}$, we need to specify functions $O'$ and $H'$ as described above.
Let $O'_{n+1}$ be equal to $A$, and let $O'_i$ be equal to $\Sigma(z : A)\,O_i$.
If $J = \{ j_1 < \ldots < j_k \}$ and $j_k \leq n$, then we define $H'_J$ as follows:
\[ x : \Sigma (z : A)\,O_{j_1}, x_{j_2} : I, \ldots x_{j_{k-1}} : I \vdash (\pi_1(x), H_J[z \repl \pi_1(x), x \repl \pi_2(x)]) : \Sigma (z : A)\,O_{j_k} \]
If $J = \{ j_1 < \ldots < j_k \}$ and $j_k = n+1$, then we define $H'_J$ as follows:
\[ x : \Sigma (z : A)\,O_{j_1}, x_{j_2} : I, \ldots x_{j_{k-1}} : I \vdash \pi_1(x) : A \]
It is easy to see that $p$ respects face maps and, since the theory has eta rules for $\Sigma$-types, $p$ respects degeneracy maps too.
Thus, we defined a map $p : N(M/A) \to N(M)/A$.
\begin{prop}[slice-stability]
Let $M$ be a model of a stable theory.
Then, for every $A \in M_{(ty,0)}$, the map $p : N(M/A) \to N(M)/A$ is a categorical equivalence.
\end{prop}
\begin{proof}
Since $N(M/A)$ and $N(M)/A$ are quasicategories, \cite[Proposition~3.5]{f-model-structures} implies that to prove that $p$ is a categorical equivalence,
we just need to show that for every $n \in \mathbb{N}$ and every commutative square of the form
\[ \xymatrix{ \partial \Delta^n \ar[r]^-f \ar[d] & N(M/A) \ar[d]^p \\
\Delta^n \ar[r]_-g & N(M)/A,
} \]
there is a map $g' : \Delta^n \to N(M/A)$ such that the top triangle commutes and the bottom triangle commutes up to relative homotopy.
Maps $g,g' : \Delta^0 \to X$ are (relatively) homotopic if the objects of $X$ represented by these maps are equivalent.
If $n > 0$, then maps $g,g' : \Delta^n \to X$ are relatively homotopic if there is a map $h : \Delta^{n+1} \to X$
such that $\delta_0(h) = g$, $\delta_0(h) = f$, and $\delta_i(h) = \sigma_0(\delta_{i-1}(f))$ for every $2 \leq i \leq n+1$.
Let us prove that for every object $s$ of $N(M)/A$, there is an object $s'$ of $N(M/A)$ such that $p(s')$ is equivalent to $s$.
An object of $N(M)/A$ is a 1-simplex $s$ of $N(M)$ such that $s|_{\Delta^{\{1\}}} = A$.
Thus, it is just a term of the form $y : C \vdash a : A$.
Let $B$ be the object of $N(M/A)$ defined as $x : A \vdash \Sigma (y : C)\,(a \idtype x)$.
It is an easy exercise to prove that $\Sigma (x : A)\,\Sigma (y : C)\,(a \idtype x)$ and $C$ are homotopy equivalent over $A$.
It follows that $p(B)$ is equivalent to $a$.
Let $f : \partial \Delta^n \to N(M/A)$ and $g = (O,H) : \Delta^n \to N(M)/A$ be maps such that $n > 0$ and the square above commutes.
Then $x : \Sigma (z : A)\,O_0, x_1 : I, \ldots x_n : I \vdash H_{[n+1]} : A$ and
$x : \Sigma (z : A)\,O_0, x_1 : I, \ldots x_{n-1} : I \vdash H_{[n]} : \Sigma (z : A)\,O_n$.
To extend $f$ to a map $\Delta^n \to N(M/A)$, we need to define a term $z : A, x : O_0, x_1 : I, \ldots x_{n-1} : I \vdash h' : O_n$
in such a way that relations \eqref{rel:left} and \eqref{rel:right} hold.
Let $h' = coe_0(x_n.\,O_n[z \repl H_{[n+1]}], \pi_2(H_{[n]}))[x \repl (z,x)]$.
Note that $H_{[n+1]}[x_j \repl c] = \pi_1(x)$ for every $1 \leq j \leq n$ and every $c \in \{ left, right \}$.
By $\sigma$-rule, $h'[x_j \repl c] = \pi_2((H_{[n]}[x \repl (z,x), x_j \repl c]))$.
Thus, the relations for $h'$ hold since they hold for $H_{[n]}$.
Let $g' : \Delta^n \to N(M/A)$ be the map corresponding to $h'$.
We need to prove that $p \circ g'$ is relatively homotopic to $g$.
That is, we need to find a map $T : \Delta^{n+2} \to N(M)$ such that $\delta_0(T) = g$, $\delta_1(T) = p \circ g'$,
and $\delta_i(T) = \sigma_0(\delta_{i-1}(g))$ for every $2 \leq i \leq n+1$.
Since all of the faces except the last one are fixed, we just need to specify a term $x : \Sigma (z : A)\,O_0, x_1 : I, \ldots x_n : I \vdash T_{[n+1]} : \Sigma (z : A)\,O_{n+1}$
corresponding to the last face and a term $x : \Sigma (z : A)\,O_0, x_1 : I, \ldots x_{n+1} : I \vdash T_{[n+2]} : A$ corresponding to the interior of the simplex $T$.
We define $T_{[n+1]}$ as $(H_{[n+1]}, coe_1(x_n.O_n[z \repl \pi_1(H_{[n]})], \pi_2(H_{[n]}), x_n))\rho$, where $\rho = [x_1 \repl x_2, \ldots x_{n-1} \repl x_n, x_n \repl x_1]$.
Let $T_{[n+2]} = H_{[n+1]}[x_1 \repl x_2, \ldots x_{n-1} \repl x_n, x_n \repl sq_r(x_1, x_{n+1})]$,
where $sq_r$ is a function such that $sq_r(i,right) = right$, $sq_r(right,j) = right$, $sq_r(i,left) = i$, and $sq_r(left,j) = j$.
It is easy to see that relations \eqref{rel:left} and \eqref{rel:right} hold.
Thus, $T$ determines a homotopy between $g$ and $p \circ g'$.
\end{proof}
The previous proposition shows that stable theories are stable under slicing.
That is if we can prove that, for every model $M$ of a stable theory, the quasicategory $N(M)$ has some property,
then this property holds for every slice quasicategory $N(M)/A$.
For example, if we can prove that $N(M)$ has all finite products, then it also has all finite limits.
Let $C$ be a quasicategory, and let $A$ and $B$ be objects of $C$.
Then we define $C/A,B$ as the slice category $C/p$, where $p : \partial \Delta^1 \to C$ is the map determined by $A$ and $B$.
The set $(C/A,B)_{[n]}$ consists of pairs of simplices $x,y \in C_{[n+1]}$ such that
$x|_{\Delta^{\{n+1\}}} = A$, $y|_{\Delta^{\{n+1\}}} = B$, and $x|_{\Delta^{\{0, \ldots n\}}} = y|_{\Delta^{\{0, \ldots n\}}}$.
Then a product of $A$ and $B$ is a terminal object of $C/A,B$.
\begin{prop}
For every model $M$ of a stable theory with $\Sigma$-types with eta rules, the quasicategory $N(M)$ is finitely complete.
\end{prop}
\begin{proof}
By \rprop{slice-stability}, we just need to prove that $N(M)$ has all finite products.
By \rprop{contr-terminal}, $N(M)$ has a terminal object since the interval is contractible.
To prove that $N(M)$ has binary products, note that $N(M)/A,B$ is isomorphic to $N(M)/(A \times B)$ and the latter has a terminal object.
\end{proof}
The results of \cite{szumilo} and \cite{kapulkin-szumilo} imply that the simplicial localization of $M$ as a category with weak equivalences is finitely complete.
If $M$ is a model of a theory with $\Pi$-types, then it was shown in \cite{kapulkin} that it is also locally Cartesian closed.
We believe that $N(M)$ is equivalent to the simplicial localization of $M$, but we do not have a proof, so these results cannot be applied in our case.
\section{Marked models}
Let $\C$ be a category, let $\mathcal{K}$ be a small category, and let $\mathcal{F} : \mathcal{K} \to \C$ be a functor.
A \emph{\K-marked object} of $\C$ is a pair $(X,\mathcal{E})$, where $X$ is an object of $\C$ and $\mathcal{E} : \mathcal{K}^{op} \to \Set$ is a subfunctor of $\Hom(\mathcal{F}(-),X)$.
Morphisms $f : \mathcal{F}(K) \to X$ that belong to $\mathcal{E}$ will be called \emph{marked}.
A morphism of marked objects is a morphism of the underlying objects that preserves marked morphisms.
The category of marked objects will be denoted by $\C^m$.
The forgetful functor $U : \C^m \to \C$ has a left adjoint $(-)^\flat : \C \to \C^m$ and a right adjoint $(-)^\sharp : \C \to \C^m$.
Let $\mathcal{K} = \{ \Delta^0, \Delta^1 \times \Delta^1 \}$.
Note that $\mathcal{K} = \{ \Delta^0 \join L\ |\ L \in \mathcal{L} \}$, where $\mathcal{L} = \{ \varnothing, \Lambda^2_2 \}$.
The Joyal model structure induces a model structure on the category of \K-marked simplicial sets.
It was shown in \cite{marked-obj} that there is a model structure on the category of marked simplicial sets with the following properties:
\begin{itemize}
\item A map of marked simplicial sets is a cofibration if and only if its underlying map is a monomorphism
\item A makred simplicial set $X$ is fibrant if and only if its underlying simplicial set is a finitely complete quasicategory,
a map $\Delta^0 \to X$ is marked if and only if it corresponds to a terminal object of $X$, and
a map $\Delta^1 \times \Delta^1 \to X$ is marked if and only if it corresponds to a Cartesian square.
\item A map of fibrant marked simplicial sets is a weak equivalence if and only if its underlying map is a homotopy equivalence of quasicategories.
\end{itemize}
In this section we define a model category $\Mod{T}^m$ of marked models of a theory $T$ with $\Sigma$-types.
We construct a Quillen adjunction $\mathfrak{C}^m \dashv N^m$ between $\sSet^m$ and $\Mod{T}^m$,
so that the following square of left Quillen functors commutes up to a natural isomorphism:
\[ \xymatrix{ \sSet \ar[d]_{(-)^\flat} \ar[r]^{\mathfrak{C}} & \Mod{T} \ar[d]^{(-)^\flat} \\
\sSet^m \ar[r]_{\mathfrak{C}^m} & \Mod{T}^m
} \]
We prove that $(-)^\flat : \Mod{T} \to \Mod{T}^m$ is a Quillen equivalence.
Moreover, if $T = T_\Sigma$, then $\mathfrak{C}^m : \sSet^m \to \Mod{T}^m$ is also a Quillen equivalence.
Thus, the model categories of models of $T_\Sigma$ and finitely complete quasicategories are Quillen equivalent.
Let $\Mod{T}^m$ be the category of $\mathfrak{C}(\mathcal{K})$-marked models of $T$.
To define a model structure on this category, we need to define a functor $R : \Mod{T}^m \to \Mod{T}$.
Let $\C$ and $\D$ be categories, and let $F \dashv G$ be an adjunction between $\C$ and $\D$.
If $\mathcal{K}$ is a set of objects of $\C$, then there is an adjunction $F^m \dashv G^m$ between
the categories $\C^m$ of \K-marked objects and $\D^m$ of $F(\mathcal{K})$-marked objects.
We define $G^m(X)$ as $G(X)$ in which a map $K \to G(X)$ is marked if and only if the adjoint map $F(K) \to X$ is marked in $X$.
Let $F^m(X,\mathcal{E}) = (F(X),F(\mathcal{E}))$.
Then it is easy to see that $F^m$ is left adjoint to $G^m$ and that the following diagram commutes up to natural isomorphisms:
\[ \xymatrix{ \C \ar[r]^{(-)^\flat} \ar[d]^F & \C^m \ar[d]^{F^m} \ar[r]^U & \C \ar[d]^F \\
\D \ar[r]_{(-)^\flat} & \D^m \ar[r]_U & \D
} \]
\begin{lem}
Let $\C$ and $\D$ be model categories, and let $F \dashv G$ be a Quillen adjunction.
Then, for every functor $\mathcal{F} : \mathcal{K} \to \C$ and every set of marked objects $\mathcal{J}$ of $\C^m$,
the adjunction $F^m \dashv G^m$ is a Quillen adjunction between $\C^m_\mathcal{J}$ and $\D^m_{F^m(\mathcal{J})}$.
\end{lem}
\begin{proof}
TODO
\end{proof}
Let us define a set $\mathcal{J}$ of $\mathfrak{C}(\mathcal{K})$-marked models of a dependent type theory $T$.
This set consists of the following objects:
\begin{enumerate}
\item The initial model in which the only marked type is the interval type.
\item The model generated by types $A,B,C : (ty,0)$ and terms $x : A \vdash f : C$ and $y : B \vdash g : C$.
The only marked square is the pullback square of $f$ and $g$:
\[ \xymatrix{ \Sigma (x : A)\,(y : B)\,(f \idtype g) \ar[r]^-{\pi_2} \ar[d]_{\pi_1} \ar[dr]_{f \circ \pi_1} & B \ar[d]^g \\
A \ar[r]_f & C
} \]
The bottom triangle commutes strictly and the homotopy corresponding to the top triangle is determined by the third projection.
\end{enumerate}
We define $\Mod{T}^m$ as the category of $\mathfrak{C}(\mathcal{K})$-marked objects localized at $\mathcal{J}$.
\begin{thm}
The adjunction $\mathfrak{C}^m \dashv N^m$ is a Quillen equivalence between $\sSet^m$ and $\Mod{T}^m$.
\end{thm}
\begin{proof}
TODO
\end{proof}
\bibliographystyle{amsplain}
\bibliography{ref}
\end{document}