/
2019-02-13-homological-algebra.tex
290 lines (244 loc) · 11.3 KB
/
2019-02-13-homological-algebra.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
\documentclass[10pt]{amsart}
\usepackage{fullpage}
% https://tex.stackexchange.com/questions/174073
\usepackage{parskip}
% macros
\usepackage{ccg-macros}
\usepackage{amscd, amssymb}
\usepackage{physics}
% pandoc is utf8 native
\usepackage[utf8]{inputenc}
% pandoc transfers h2 to subsections, I prefer h2 for sections
\let\subsubsection\subsection
\let\subsection\section
\let\section\chapter
\let\chapter\part
% pandoc section numbering
\setcounter{secnumdepth}{5}
% pandoc urls as footnotes
\usepackage[unicode=true]{hyperref}
\renewcommand{\href}[2]{#2\footnote{\url{#1}}}
% pandoc for upquote if available
\IfFileExists{upquote.sty}{\usepackage{upquote}}{}
% tables, pandoc manges export
\usepackage[all]{xy}
% pandoc prevent overfull lines
\setlength{\emergencystretch}{3em}
\providecommand{\tightlist}{%
\setlength{\itemsep}{0pt}\setlength{\parskip}{0pt}}
% pandoc default figure placement to htbp
\makeatletter
\def\fps@figure{htbp}
\makeatother
\title{First look at the Universal Coefficient Theorem}
\author{Colton Grainger (scribe) \and Marvin Qi (presenter)}
\date{2019-02-13}
\begin{document}
\maketitle
This problem is set from Bredon {[}1, No. IV.5.1{]}; it works out
techniques that might later play into a proof of the universal
coefficient theorem.
\emph{Given}. Let \(p \colon \ZZ \to \ZZ\) be multiplication by the
prime \(p\), and consider the short exact sequence \begin{equation}
0 \to \ZZ \xrightarrow{p} \ZZ \xrightarrow{\pi} \ZZ_p \to 0.
\label{eq:p}
\end{equation}
Say \(X\) is a topological space. We assume
\(\Delta_*(-) \colon \mathsf{Top} \to \mathsf{Comp}\) is the singular
chain functor from topological spaces to chain complexes of abelian
groups, with \(H_*(-) \colon \mathsf{Comp} \to \mathsf{GradedAb}\) the
homology functor on chain complexes. For an abelian group \(G\), denote
by \(H_*(C_*; G)\) the \emph{homology of \(C_*\) with coefficients in
\(G\)}, which is the homology of the chain complex \(G \otimes C_*\)
with boundary operator \(\id_G \otimes \partial\).
\emph{To prove.} There is an exact sequence \begin{equation}
0 \to \ZZ_p \otimes H_n(X) \to H_n(X;\ZZ_p) \to \Tor(H_{n-1}(X), \ZZ_p) \to 0,
\label{eq:wts}
\end{equation} with a homomorphism
\(H_n(X; \ZZ_p) \to \ZZ_p \otimes H_n(X)\) such that \eqref{eq:wts}
splits.
\emph{Proof.} We'll construct the exact sequence \eqref{eq:wts}, then
the splitting homomorphism.
Each group in \(\Delta_*(X)\) is free abelian, and, in particular, flat
as a \(\ZZ\)-module. Thus tensoring the short exact sequence
\eqref{eq:p} on the right by the \(n\)th singular chain \(\Delta_n(X)\)
group preserves exactness, and produces: \begin{equation}
\xymatrix{
0 \ar[r] & \ZZ\otimes \Delta_n(X) \ar[r]^{p\otimes \id_\Delta} & \ZZ\otimes \Delta_n(X) \ar[r]^{\pi\otimes \id_\Delta} & \ZZ_p \otimes\Delta_n(X) \ar[r]& 0
}
\label{eq:level}
\end{equation}
Lining up copies of \eqref{eq:level} levelwise with the appropriate
differentials, we exhibit the short exact sequence of chain complexes:
\begin{equation}
\xymatrix{
& \vdots \ar[d] & \vdots \ar[d] & \vdots \ar[d] &\\
0 \ar[r] & \ZZ\otimes \Delta_n(X) \ar[d]^{\id_\ZZ \otimes \partial_n} \ar[r]^{p\otimes \id_\Delta} & \ZZ\otimes \Delta_n(X) \ar[d]^{\id_\ZZ \otimes \partial_n} \ar[r]^{\pi\otimes \id_\Delta} & \ZZ_p \otimes\Delta_n(X) \ar[d]^{\id_{\ZZ_p} \otimes \partial_n}\ar[r]& 0\\
0 \ar[r] & \ZZ\otimes \Delta_{n-1}(X) \ar[d] \ar[r]^{p\otimes \id_\Delta} & \ZZ\otimes \Delta_{n-1}(X) \ar[d] \ar[r]^{\pi\otimes \id_\Delta} & \ZZ_p \otimes\Delta_{n-1}(X) \ar[d] \ar[r]& 0\\
& \vdots & \vdots & \vdots &\\
}
\label{eq:ch1}
\end{equation}
As abelian groups, we'll identify
\(\ZZ \otimes \Delta_n(X) = \Delta_n(X)\) and
\(\ZZ_p \otimes \Delta_n(X) = \Delta_n(X)/p\) (the cokernel of the
\(p\)th multiple map). Accordingly, we'll write the homomorphisms
\(p \otimes \id_\Delta = p\), \(\pi \otimes \id_\Delta = \pi\), with all
the differentials \(\id \otimes \partial_n\) as \(\partial\). Given
these abbreviations, the short exact sequence \eqref{eq:ch1} becomes:
\begin{equation}
\xymatrix{
0 \ar[r] & \Delta_*(X) \ar[r]^{p} &\Delta_*(X) \ar[r]^{\pi} & \Delta_*(X)/p \ar[r]& 0
}
\label{eq:ses}
\end{equation}
By inspection of the definition of homology with coefficients along with
the diagram \eqref{eq:ch1}, the long exact sequence of homology groups
induced by \eqref{eq:ses} is: \begin{equation}
\xymatrix{
\cdots \ar[r]^{\delta_*} & H_n(X) \ar[r]^{ p_* } & H_n(X) \ar[r]^{ \pi_* } & H_n(X; \ZZ_p) \ar[r]^{\delta_*} & H_{n-1}(X) \ar[r]^{p_*} & \cdots
}
\label{eq:les}
\end{equation}
We now make an aside to argue the algebraic structure of a long exact
sequence is encoded in a series of intertwined short exact sequences
{[}2, No. 2.5{]}. Start by considering an exact sequence of abelian
groups
\[\cdots \to A_{n+2} \to A_{n+1} \to A_{n} \to A_{n-1} \to A_{n-2} \to \cdots.\]
By definition of exactness at \(A_n\),
\begin{equation}\label{eq:defexact}\im (A_{n+1} \to A_n) = \ker(A_n \to A_{n-1}).\end{equation}
Quotienting by the kernel of the map out of \(A_n\), we have an
isomorphism onto the image in \(A_{n-1}\) \begin{equation}
A_n/\ker (A_n \to A_{n-1}) \cong \im(A_n \to A_{n-1})
\label{eq:isomthm}
\end{equation} Now, the cokernel of an additive homomorphism is the
target group quotiented by the incoming image. Substituting
\eqref{eq:defexact} into \eqref{eq:isomthm} gives an example of such
\[\coker ( A_{n+1} \to A_n) := A_n/\im (A_{n+1} \to A_n) \cong \im(A_n \to A_{n-1}).\]
Putting it all together, there's a group \(C_n\) that is
\begin{itemize}
\tightlist
\item
the kernel of the map \(A_n \to A_{n-1}\),
\item
isomorphic to the image of the map \(A_{n+1} \to A_n\), and
\item
isomorphic to the cokernel of the map \(A_{n+2} \to A_{n+2}\).
\end{itemize}
Moreover, there are groups \(C_{n+k}\) for \(k \in \ZZ\) satisfying
analogously conditions. We may now decorate our long exact sequence with
inclusions and natural projections so that the diagram \begin{equation}
\xymatrix @=2pt@! { %We want equal spacing, and the entries close together
&&0\ar[dr]&&0&&0\ar[dr]&&0&&&&0\\
&&&C_{n+1}\ar[ur]\ar[dr]&&&&C_{n-1}\ar[ur]\ar[dr]&&&&C_{n-3}\ar[ur]\\
\cdots \ar[rr]&&A_{n+2}\ar[rr]\ar[ur]&&A_{n+1}\ar[rr]\ar[dr]&&A_{n}\ar[rr]\ar[ur]&&A_{n-1}\ar[rr]\ar[dr]&&A_{n-2}\ar[ur]\ar[rr]&&\cdots\\
&C_{n-2}\ar[ur]&&&&C_{n}\ar[ur]\ar[dr]&&&&C_{n-2}\ar[ur]\ar[dr]&&&&\\
0\ar[ur]&&&&0\ar[ur]&&0&&0\ar[ur]&&0&&&&
}
\label{eq:knit}
\end{equation} is exact. We emphasize that the crossing maps are either
\emph{inclusions} or \emph{natural projections}. The crossing sequences
are short exact because, e.g., for
\(0 \to C_n \to A_n \to C_{n-1} \to 0\),
\begin{itemize}
\tightlist
\item
\(C_n\) is the kernel, so injects and passes through \(A_n\) to \(0\)
in \(C_{n-1}\),
\item
\(A_n\) surjects onto \(C_{n-1}\) and passes to \(0\),
\item
any element in \(A_n\) that passes to \(0\) in \(C_{n-1}\) must be in
the image of \(C_n\) in \(A_n\), because \(A_n \to C_{n-1}\) is the
natural projection onto \(A_n / \ker(A_n \to A_{n-1}) = A_n /C_n\).
\end{itemize}
We now construct the desired short exact sequence \eqref{eq:wts}.
Placing the long exact sequence of homology groups \eqref{eq:les} into
the position of the long exact sequence in the cross stitched diagram
\eqref{eq:knit}, one visibly obtains the short exact sequence:
\begin{equation}
\xymatrix{
0 \ar[r] & \coker\big(H_n(X) \xrightarrow{p_*} H_n(X)\big) \ar[r] & H_n(X; \ZZ_p) \ar[r] & \ker\big( H_{n-1}(X) \xrightarrow{p_*} H_{n-1}(X)\big) \ar[r] & 0
}
\label{eq:sestosplit}
\end{equation}
This sequence \eqref{eq:sestosplit} is isomorphic to the one we wanted
\eqref{eq:wts}, as
\[\coker\big(H_n(X) \xrightarrow{p_*} H_n(X)\big) \cong \ZZ_p \otimes H_n(X)\]
and
\[\ker\big( H_{n-1}(X) \xrightarrow{p_*} H_{n-1}(X)\big) =: \Tor(H_{n-1}(X), \ZZ_p).\]
What remains to be proven is that there is a homomorphism from
\(H_n(X, \ZZ_p)\) down to \(\ZZ_p \otimes H_n(X)\) that (left) inverts
the inclusion of \(\ZZ_p \otimes H_n(X)\) up into \(H_n(X, \ZZ_p)\). My
argument here closely follows {[}2, No. 3.21{]}.
We'll work levelwise for a while in the singular chain complex
\(\Delta_*(X)\). Let \(Z_n\) be the cycle subgroup of \(\Delta_n(X)\)
and \(B_{n-1}\) the boundary subgroup \(\delta \Delta_n(X)\) in
\(\Delta_{n-1}(X)\). Observe that \(Z_n\) includes into \(\Delta_n(X)\)
which projects onto \(B_{n-1}\). So we have the short exact sequence of
free abelian groups \begin{equation}
\xymatrix{
0 \ar[r] & Z_n \ar[r]^f & \Delta_n(X) \ar@/^1pc/ @{-->} [l]^g \ar[r] & B_{n-1} \ar[r] & 0,
}
\end{equation} that splits with \(g \circ f = \id_{Z_n}\). Extend \(g\)
to the quotient \(H_n(X)\) so that the following diagram commutes:
\begin{equation}
\label{eq:extend}
\xymatrix{
\Delta_n(X)\ar[d]^g \ar@{-->}[dr]^{g'} & \\
Z_n \ar[r] & H_n(X)
}\end{equation}
Consider the graded abelian group \(\{H_k(X)\}_{k \in \ZZ}\) in a chain
with trivial differentials
\[\ldots \xrightarrow{0} H_{n+1}(X) \xrightarrow{0} H_n(X) \xrightarrow{0} H_{n-1}(X) \xrightarrow{0} \ldots.\]
Call this chain complex \(H_*(X)\). We induce a chain map \(g_*\) from
\(\Delta_*(X)\) to \(H_*(X)\) with \(g'\):
\begin{minipage}{.5\textwidth}
\begin{equation}
\xymatrix{
\vdots \ar[d]^\partial & \vdots \ar[d]^0\\
\Delta_{n+1}(X) \ar[r]^{g'} \ar[d]^\partial & H_{n+1}(X) \ar[d]^0\\
\Delta_{n}(X) \ar[r]^{g'} \ar[d]^\partial& H_{n}(X) \ar[d]^0\\
\Delta_{n-1}(X) \ar[r]^{g'} \ar[d]^\partial& H_{n-1}(X) \ar[d]^0\\
\vdots & \vdots\\
}\label{eq:trivialchain}\end{equation}
\end{minipage}
\begin{minipage}{.5\textwidth}
\begin{equation}
\xymatrix{
\vdots \ar[d]^{\id\otimes\partial} & \vdots \ar[d]^0\\
\ZZ_p\otimes\Delta_{n+1}(X) \ar[r]^{\id \otimes g'} \ar[d]^{\id \otimes \partial} & \ZZ_p\otimes H_{n+1}(X) \ar[d]^0\\
\ZZ_p\otimes\Delta_{n}(X) \ar[r]^{\id \otimes g'} \ar[d]^{\id \otimes \partial} & \ZZ_p\otimes H_{n}(X) \ar[d]^0\\
\ZZ_p\otimes\Delta_{n-1}(X) \ar[r]^{\id \otimes g'} \ar[d]^{\id \otimes \partial} & \ZZ_p\otimes H_{n-1}(X) \ar[d]^0\\
\vdots & \vdots\\
}\end{equation}
\end{minipage}
Observe that
\(\partial(\Delta_{n+1}(X)) = B_n(X) = \ker\left(Z_n \to H_n(X)\right)\)
so that each square in \eqref{eq:trivialchain} commutes. We may then
tensor both chains by \(\ZZ_p\) on the right, obtaining
\[\ZZ_p \otimes \Delta_*(X) \xrightarrow{\id \otimes g_*} \ZZ_p \otimes H_n(X).\]
When we take homology on \(\ZZ_p \otimes \Delta_*(X)\) by definition we
have \(H_*(X, \ZZ_p)\). However, when computing homology in the chain
complex \(\ZZ_p \otimes H_n(X)\), the image of each map in the complex
is \emph{trivial}, so the homology of the complex is
\(\ZZ_p \otimes H_*(X)\).
The map \({\id \otimes g_*}\) of chain complexes induces a homomorphism
\(H_*(\id\otimes g')\) on homology from
\(H_*(X, \ZZ_p) \to \ZZ_p \otimes H_*(X)\). In particular, at the
\(n\)th level,
\begin{equation}\label{eq:finish}H_n(X, \ZZ_p) \xrightarrow{H_*(\id \otimes g')} \ZZ_p \otimes H_n(X).\end{equation}
By construction in \eqref{eq:extend} of \(g'\) as the extension of the
splitting map \(g\) (where \(g \circ f = \id_{Z_n}\)), the induced map
\(H_*(\id \otimes g')\) in \eqref{eq:finish} is a left inverse to the
inclusion of \(\ZZ_p \otimes H_n(X)\) into \(H_n(X, \ZZ_p)\). \qedsymbol
\subsection*{References}
\addcontentsline{toc}{subsection}{References}
\hypertarget{refs}{}
\leavevmode\hypertarget{ref-Bre93}{}%
{[}1{]} G. E. Bredon, \emph{Topology and Geometry}. New York:
Springer-Verlag, 1993.
\leavevmode\hypertarget{ref-Che09}{}%
{[}2{]} J. L. Chen, ``Universal Coefficient Theorem for Homology,''
2009.
\end{document}