---- Let $(A, X)$ be a [[topological pair]]. That is, $X$ is a [[topological space]] and $A \subset X$ is a [[subspace topology|subspace]]. Write $\iota:A \hookrightarrow X$ for the [[inclusion map]]. > [!definition] Definition. ([[relative singular homology]]) > > > [[singular (co)chain map and homomorphism induced by a continuous map|The map]] $\iota_{n}:C_{n}(A) \hookrightarrow C_{n}(X)$ is evidently [[injection|injective]], giving a left-[[exact sequence]] $0 \to C_{\bullet}(A) \to C_{\bullet}(X)$. The $i$th [[relative homology of an embedding of chain complexes|relative homology]] of this embedding of [[chain complex of modules|chain complexes]] is called the **$i$th relative (singular) homology of the pair $(X,A)$** and denoted $H_{i}(X,A)$. > $H_{i}(-, -)$ is a [[covariant functor]] $\mathsf{TopPairs} \to \mathsf{Ab}$, obtained by composing the [[singular map of pairs functor]] $\mathsf{TopPairs} \to \mathsf{Chain Pairs}$ with the general [[relative homology of an embedding of chain complexes|relative homology functor]] $\mathsf{ChainPairs} \to \mathsf{Ab}$. [[relative homology of an embedding of chain complexes|Explicitly]], the functor $H_{i}(-,-)$ sends a [[topological pair|map of pairs]] $(X,A) \xrightarrow{f}(Y,B)$ to the [[linear map|homomorphism]] $\begin{align} f_{*}:H_{i}(X,A) &\to H_{i}(Y,B) \\ \big[ [x] \big] & \mapsto \big[ [f_{i}(x)] \big]. \end{align}$ - [ ] relative singular cohomology and its (contravariant!) functor > [!definition] Definition. ([[relative singular homology]] — unpacked) > Here is an 'unpacked' version of the above definition. [[singular (co)chain map and homomorphism induced by a continuous map|The map]] $\iota_{n}:C_{n}(A) \hookrightarrow C_{n}(X)$ is evidently [[injection|injective]], giving a [[short exact sequence]] $0 \to C_{n}(A) \xhookrightarrow{\iota_{n}}C_{n}(X) \twoheadrightarrow \underbrace{ \frac{C_{n}(X)}{C_{n}(A)} }_{ := C_{n}(X, A) } \to 0.$ > The [[singular homology|differential]] $d_{n}:C_{n}(X) \to C_{n-1}(X)$ restricts to a map $C_{n}(A) \to C_{n-1}(A)$, and thus descends[^1] to a [[well-defined]] differential $d_{n}=\overline{d_{n}}:C_{n}(X,A) \to C_{n-1}(X,A)$ sending $[c] \mapsto [d_{n}(c)]$. > > ```tikz > \usepackage{tikz-cd} > \usepackage{amsmath} > \footnotesize > \begin{document} > % https://tikzcd.yichuanshen.de/#N4Igdg9gJgpgziAXAbVABwnAlgFyxMJZABgBpiBdUkANwEMAbAVxiRAGEB9MACgA0AlCAC+pdJlz5CKAIzkqtRizZdgYALQzh-IaPHY8BIgCZ51es1aIOnNZu19SAQQEBeADruAZgCc6AY2BVDS0dYSDbEO0XYRExEAwDKSIyGQULZWtPXwCItQcBcODowtcuXkcXEQUYKABzeCJQXwgAWyQAZmocCCRTRUs2TxgADyw4HDgAAgBCKc8IGhgfBiwwGGAobljqBjoAIxgGAAUJQ2kQHyw6gAscOOafNqQyEB6kOQHMkC3CPRAWu1EJ93oh+jcYHQoGwcAB3CAQqEIf6Al7dXqILogRHQ6xwhGQqDIijCIA > \begin{tikzcd} > C_n(X) \arrow[r, "d_n"] \arrow[d, two heads] & C_{n-1}(X) \arrow[r, two heads] & {C_{n-1}(X,A)=\frac{C_{n-1}(X)}{C_{n-1}(A)}} \\ > {\frac{C_{n}(X)}{C_{n}(A)}=C_n(X,A)} \arrow[rru, "\exists ! \overline{d_n}"'] & & > \end{tikzcd} > \end{document} > ``` > > The $i$th [[(co)homology of a complex|homology]] $H_{i}\big(C_{\bullet}(X,A)\big)$ of the [[chain complex of modules|chain complex]] $C_{\bullet}(X, A)$ is called the **$i$th relative homology** of the [[topological pair|pair]] $(X,A)$ and denotd $H_{i}(X,A)$. One says relative homology is given by **relative cycles** — chains $c$ whose boundary $d_{n}(c)$ is a chain in $A$ (so that $\overline{d_{n}}(c)=[d_{n}(c)]=0$) — modulo relative boundaries — ... $C_{n}(X,A)=\frac{C_{n}(X)}{C_{n}(A)}$ is 'chains in $X$, but identify any two chains which differ by a chain in $A. Intuitively, seeing $C_{n}(X,A)$ should read like 'chains in $X$, ignoring what happens in $A. > [!basicproperties] > - [[long exact sequence for relative singular homology]] ([[long exact sequence for relative singular homology|long exact sequence for a pair]]) > - [[detecting isomorphisms with relative homology]] ^properties ---- #### [^1]: E.g., [[characterization of quotienting a group|because]] if $\pi:C_{n-1}(X) \to C_{n-1}(X,A)$ is the canonical projection, then the induced map $\overline{d_{n}}$ is [[well-defined]] provided $A \subset \operatorname{ker }(\pi \circ d_{n})$, i.e., provided $d_{n}(A) \subset C_{n-1}(A)$. This is why we care about whether $d_{n}$ restricts to a map $C_{n}(A) \to C_{n-1}(A)$. > [!basicexample] > Let $\Sigma_{2}$ denote the two-torus $\mathbb{T}^{2} \# \mathbb{T}^{2}$. Compute $H_{\bullet}(\Sigma _{2}, A)$, where $A$ is a simple closed curve which: > > **(a) separates $\Sigma_{2}$ into two genus one pieces with one boundary component each.** > > **(b) Is a non-separating simple closed curve cutting along which gives a genus one surface with two holes.** > > **(c) bounds an embedded disc.** > > Note that in all $3$ cases it is easy to see that $X,A$ is a [[good pair]]. > > As such, will liberally employ [[relative homology for a good pair is reduced homology of the quotient]]. We will also use the result [[homology of the closed orientable genus-g surface]]. And we will invoke the following lemma: > > ![[reduced homology of a good sum of wedges is direct sum of reduced homologies#^proposition]] > > ![[reduced homology of a good sum of wedges is direct sum of reduced homologies#^4af639]] > > > > **(a) separates $\Sigma_{2}$ into two genus one pieces with one boundary component each.** In this case $\Sigma_{2} / A$ is [[homotopy equivalent]] to a [[wedge sum]] $\mathbb{T}^{2} \vee \mathbb{T}^{2}$ of Torii. By [[relative homology for a good pair is reduced homology of the quotient]] it suffices to compute the [[reduced homology]] of the space $\mathbb{T}^{2} \vee \mathbb{T}^{2}$. Using the lemma, it corresponds to the [[direct sum]] of $\widetilde{H}_{\bullet}(\mathbb{T}^{2})$ with itself. Thus, > > $H_{n}(\Sigma_{2}, A)=\begin{cases} > 0 & n=0 \\ > \mathbb{Z}^{2} \oplus \mathbb{Z}^{2} & n=1 \\ > \mathbb{Z} \oplus \mathbb{Z} & n=2 \\ > 0 & n \geq 3. > \end{cases}$ > > **(b) Is a non-separating simple closed curve cutting along which gives a genus one surface with two holes.** > > This space looks like $\mathbb{T}^{2} \vee \mathbb{S}^{2}$. So $H_{n}(\Sigma_{2},A)=\begin{cases} > 0 & n=0 \\ > \mathbb{Z}^{2} & n=1 \\ > \mathbb{Z} \oplus \mathbb{Z} & n=2 \\ > 0 & n \geq 3. > \end{cases}$ > > **(c) bounds an embedded disc.** > In this case, the space $\Sigma_{2} / A$ is [[homotopy equivalent]] to $\Sigma_{2}$ itself. And so we may appeal immediately to [[homology of the closed orientable genus-g surface]] to obtain $H_{n}(\Sigma_{2}, A)=\begin{cases} > 0 & n=0 \\ > \mathbb{Z}^{4} & n=1 \\ > \mathbb{Z} & n = 2 \\ > 0 & n \geq 3. > \end{cases}$ ^basic-example - [ ] there is an error here (see his feedback) ---- #### References > [!backlink] > ```dataview > TABLE rows.file.link as "Further Reading" > FROM [[]] > FLATTEN file.tags as Tag > WHERE Tag = "#definition" OR Tag = "#theorem" OR Tag = "#MOC" OR Tag = "#proposition" OR Tag = "#axiom" > GROUP BY Tag > ``` > [!frontlink] > ```dataview > TABLE rows.file.link as "Further Reading" > FROM outgoing([[]]) > FLATTEN file.tags as Tag > WHERE Tag = "#definition" OR Tag = "#theorem" OR Tag = "#MOC" OR Tag = "#proposition" OR Tag = "#axiom" > GROUP BY Tag > ```