Simplicial and Singular Homology

Homology provides a powerful, algebraic language for distinguishing topological spaces by quantifying their "holes." While two spaces might look different, homology groups—computable topological invariants—can often prove they are fundamentally distinct or reveal when one can be continuously deformed into another. You will learn two primary methods for defining these groups: the combinatorial, concrete approach of simplicial homology built on triangulations, and the flexible, general theory of singular homology built on continuous maps. Mastering their computation and the tools that connect them, like the Mayer-Vietoris sequence, unlocks the ability to analyze shapes from simple spheres to complex manifolds.

From Triangles to Chains: Simplicial Homology

Simplicial homology offers a computable entry point by working with spaces built from simple building blocks. A simplicial complex $K$ is a space constructed by gluing together points, line segments, triangles, tetrahedra, and their higher-dimensional analogs called simplices, according to specific rules about how they intersect. A 0-simplex is a point, a 1-simplex is a line segment, a 2-simplex is a solid triangle, and so on. This structure is a triangulation of the underlying topological space.

The algebraic machinery begins by forming groups of chains. For each dimension $n$, you create the free abelian group $C_n(K)$ whose basis elements are the oriented $n$-simplices of $K$. An orientation assigns a direction or order to the vertices of a simplex. A 1-simplex $[v_0,v_1]$ is oriented from $v_0$ to $v_1$, while a 2-simplex $[v_0,v_1,v_2]$ has a clockwise or counterclockwise ordering. A general $n$-chain is a formal integer sum of these oriented $n$-simplices, like $3{\sigma }_1-2{\sigma }_2$.

The crucial map connecting these groups is the boundary operator ${\partial }_n:C_n(K)\to C_{n-1}(K)$. It is a homomorphism defined linearly on each simplex by taking its oriented geometric boundary. For a 1-simplex (edge) $[v_0,v_1]$, the boundary is ${\partial }_1([v_0,v_1])=[v_1]-[v_0]$. For a 2-simplex (triangle) $[v_0,v_1,v_2]$, the boundary is: $${\partial }_2([v_0,v_1,v_2])=[v_1,v_2]-[v_0,v_2]+[v_0,v_1]$$ A fundamental property is that applying the boundary twice yields zero: ${\partial }_n\circ {\partial }_{n+1}=0$. This condition allows us to define the core objects of homology: cycles and boundaries. An $n$-cycle is a chain whose boundary is zero; the group of $n$-cycles is $Z_n(K)=\ker ({\partial }_n)$. An $n$-boundary is a chain that is itself the boundary of some higher-dimensional chain; the group of $n$-boundaries is $B_n(K)=\mathrm{im}({\partial }_{n+1})$. Since every boundary is a cycle ($B_n\subseteq Z_n$), we can form the quotient.

The $n$-th simplicial homology group of the complex $K$ is this quotient: $$H_n^{\text{Simp}}(K)=Z_n(K)/B_n(K)$$ Its elements are homology classes, where two cycles are in the same class if their difference is a boundary. The group structure (whether it is $\mathbb{Z}$, a finite cyclic group $\mathbb{Z}/m\mathbb{Z}$, or a direct sum of these) and its rank (the Betti number) give topological information. The rank of $H_0$ counts path-connected components, while the rank of $H_1$ counts "1-dimensional holes" or non-contractible loops.

Continuous Maps as Building Blocks: Singular Homology

While simplicial homology is intuitive, it depends on a specific triangulation. Singular homology generalizes this idea to any topological space $X$, without requiring a combinatorial structure. The basic building block is a singular $n$-simplex, defined as any continuous map $\sigma :{\Delta }^n\to X$, where ${\Delta }^n$ is the standard geometric $n$-simplex in ${\mathbb{R}}^{n+1}$. Unlike a simplicial simplex, which is an embedding, a singular simplex can be wildly degenerate—it can be constant, crumpled, or stretched.

You then define the singular chain group $C_n(X)$ as the free abelian group generated by all singular $n$-simplices in $X$. These groups are enormous, but the algebraic definitions proceed analogously. The singular boundary operator ${\partial }_n:C_n(X)\to C_{n-1}(X)$ is defined on a generator $\sigma$ by restricting $\sigma$ to the faces of ${\Delta }^n$. Formally, if $\sigma :[v_0,...,v_n]\to X$, then: $${\partial }_n(\sigma )=\sum _{i=0}^n(-1{)}^i\sigma {|}_{[v_0,...,\widehat{v_i},...,v_n]}$$ where the hat indicates the omitted vertex. This satisfies ${\partial }^2=0$. The groups of singular cycles $Z_n(X)$ and singular boundaries $B_n(X)$ are defined as kernels and images, leading to the $n$-th singular homology group: $$H_n^{\text{Sing}}(X)=Z_n(X)/B_n(X)$$ A major theorem states that for a triangulable space (like those studied in simplicial homology), its simplicial and singular homology groups are naturally isomorphic. This justifies using the simpler notation $H_n(X)$ and computing with whichever theory is more convenient. Singular homology's strength is its functoriality: any continuous map $f:X\to Y$ induces a homomorphism $f_{\ast }:H_n(X)\to H_n(Y)$ on homology, a property that is less straightforward in the simplicial setting without a map between complexes.

Computing Homology: Key Examples and Results

Computing these groups directly from the definition can be cumbersome. However, by leveraging geometric intuition and algebraic tools, you can determine the homology of fundamental shapes.

• Spheres ($S^n$): The homology of the $n$-sphere captures the idea of a single, $n$-dimensional void. For $n\ge 1$:

$$H_k(S^n)\cong \{\begin{array}{ll}\mathbb{Z}& \text{if }k=0\text{ or }k=n\\ 0& \text{otherwise}\end{array}$$ The $\mathbb{Z}$ in dimension $0$ represents the single connected component. The $\mathbb{Z}$ in dimension $n$ corresponds to the $n$-dimensional "hole" enclosed by the sphere. The calculation involves analyzing cycles that are not boundaries: in $S^1$, the 1-cycle that loops around the circle is not the boundary of any 2-chain within the circle itself.

• Tori ($T^2=S^1\times S^1$): The torus has more structure, reflected in its homology. Its homology groups are:

$$H_0(T^2)\cong \mathbb{Z},H_1(T^2)\cong \mathbb{Z}\oplus \mathbb{Z},H_2(T^2)\cong \mathbb{Z}$$ The rank of $H_1$ is 2, generated by two independent non-contractible loops: one going around the "hole" of the doughnut (meridional) and one going through its center (longitudinal). The $H_2$ generator corresponds to the entire 2-cycle that is the surface itself.

• Real Projective Plane ($\mathbb{R}{P}^2$): This non-orientable surface introduces torsion into homology. Its homology with integer coefficients is:

$$H_0(\mathbb{R}{P}^2)\cong \mathbb{Z},H_1(\mathbb{R}{P}^2)\cong \mathbb{Z}/2\mathbb{Z},H_k(\mathbb{R}{P}^2)=0\text{ for }k\ge 2$$ The $H_1$ group being $\mathbb{Z}/2\mathbb{Z}$ is the key invariant. It arises because the central loop in $\mathbb{R}{P}^2$ (like a line through the origin in the model where antipodal points are identified) is not a boundary, but traversing it twice is a boundary. This "twice equals zero" phenomenon creates a finite cyclic component in the homology group.

The Power of Exact Sequences and Mayer-Vietoris

Direct computation from chains becomes intractable for complex spaces. This is where algebraic tools like exact sequences become indispensable. A sequence of homomorphisms between groups $...\to A\stackrel{f}{\to }B\stackrel{g}{\to }C\to ...$ is exact at $B$ if $\mathrm{im}(f)=\ker (g)$. A short exact sequence $0\to A\to B\to C\to 0$ implies $B$ is built in a specific way from $A$ and $C$.

The Mayer-Vietoris sequence is a powerful long exact sequence that lets you compute the homology of a space $X$ by decomposing it into two smaller, overlapping subspaces $U$ and $V$ whose union is $X$. The sequence relates their homologies: $$\cdots \to H_n(U\cap V)\to H_n(U)\oplus H_n(V)\to H_n(X)\stackrel{{\partial }_{\ast }}{\to }{H}_{n-1}(U\cap V)\to \cdots$$ The maps are induced by inclusions, and the connecting homomorphism ${\partial }_{\ast }$ carries geometric information about how cycles in $X$ can be expressed as a union of cycles in $U$ and $V$. For example, to compute $H_n(S^n)$, you can decompose the sphere into two overlapping hemispheres (contractible sets $U$ and $V$) whose intersection is homotopy equivalent to $S^{n-1}$. The Mayer-Vietoris sequence then reduces the problem to the homology of $S^{n-1}$, setting up an inductive proof. It is the workhorse for breaking complicated spaces into manageable pieces.

Common Pitfalls

1. Confusing Simplicial and Singular Contexts: A singular simplex is not a "piece" of the space; it is a continuous map that can be highly degenerate. A common error is to treat singular chains as if they form a nice, finite triangulation. Remember, singular theory uses all possible such maps, which makes the chain groups large but ensures the theory applies universally.
2. Misinterpreting Torsion: Finding $H_1(\mathbb{R}{P}^2)\cong \mathbb{Z}/2\mathbb{Z}$ does not mean there is a 2-torsion "hole." It means there exists a 1-cycle (a loop) that is not itself a boundary, but when added to itself (traversed twice), it becomes a boundary. The homology class of this cycle has order 2 in the group.
3. Ignoring Orientation in Boundaries: The signs in the boundary formula $\partial ([v_0,v_1,v_2])=[v_1,v_2]-[v_0,v_2]+[v_0,v_1]$ are crucial for ensuring ${\partial }^2=0$. Omitting them or misordering vertices will break the algebra and lead to incorrect kernel and image calculations.
4. Overlooking the Functorial Nature: Homology is not just a number; it's a functor. The induced maps $f_{\ast }$ are as important as the groups themselves. Two spaces may have isomorphic homology groups, but if a proposed homeomorphism between them induces different maps on homology, it cannot be continuous in the required way.

Summary

• Simplicial homology provides a combinatorial, computable theory by building a space from oriented simplices and studying cycles and boundaries within that structure.
• Singular homology generalizes this to any topological space by using continuous maps from standard simplices as generators, yielding isomorphic results for triangulable spaces and superior functorial properties.
• Homology groups $H_n(X)$ are quotient groups (cycles/boundaries) whose ranks (Betti numbers) and torsion subgroups encode the number of $n$-dimensional "holes" and more subtle topological features.
• Key examples include spheres ($H_n(S^n)=\mathbb{Z}$), tori ($H_1(T^2)=\mathbb{Z}\oplus \mathbb{Z}$), and projective spaces, which can exhibit torsion ($H_1(\mathbb{R}{P}^2)=\mathbb{Z}/2\mathbb{Z}$).
• The Mayer-Vietoris sequence is an essential computational tool—a long exact sequence that allows you to compute the homology of a space from the homologies of overlapping subspaces that cover it.