03.12.22 · modern-geometry / homotopy

$Δ$ -complex / semi-simplicial set

shipped3 tiersLean: none

Anchor (Master): Hatcher §2.1 (combinatorial form); Rourke-Sanderson 1971 *$\Delta$-sets I/II* (Quart. J. Math. Oxford 22); May 1967 *Simplicial Objects in Algebraic Topology*; Goerss-Jardine *Simplicial Homotopy Theory*

Intuition [Beginner]

A $Δ$ -complex is a recipe for building a topological space from standard simplices — points, line segments, triangles, tetrahedra, and their higher cousins — by gluing entire faces of one simplex to entire faces of another. The data is purely combinatorial: a list of simplices in each dimension, together with the rule for how each simplex's faces sit inside lower-dimensional simplices. Once you specify this data, the actual space is determined.

The semi-simplicial set is the same data presented as a category-theoretic object. For each dimension $n$ you list a set of abstract $n$ -simplices, and for each face inclusion $Δ^{n - 1} ↪ Δ^{n}$ you record where each abstract $n$ -simplex sends that face. There is no requirement that two abstract simplices be embedded the same way or even injectively — you can glue a triangle's three edges into a single circle, and the resulting space is the dunce cap.

Why bother with this if CW complexes already exist? Because the gluing for a $Δ$ -complex is along combinatorial faces, not arbitrary continuous maps. The combinatorial structure means homology can be computed directly from the abstract simplex-counting data, with no extra work to set up cellular boundaries. This combinatorial accessibility is the reason simplicial methods preceded CW methods historically and remain the workhorse for computer-assisted computations.

Visual [Beginner]

A schematic showing a square with its four edges identified pairwise: the top edge with the bottom edge (with matching arrows pointing the same direction) and the left edge with the right edge (also matching). The square is divided by one diagonal into two triangles, and each triangle is labelled as an abstract $2$ -simplex. The two triangles share the diagonal as one of their edges, and the boundary edges get identified into a single horizontal edge and a single vertical edge. The resulting space is the torus.

The picture captures the essence: two $2$ -simplices, three $1$ -simplices (the two boundary edges and the diagonal), and one $0$ -simplex (every corner of the square becomes the same point after identification). The combinatorial data is one vertex, three edges, two triangles — the same numbers you would write down on a piece of paper. Replacing the matching-arrow rule for the top and bottom edges by an opposite-arrow rule gives the Klein bottle instead, with the same vertex-edge-face count.

Worked example [Beginner]

Build the two-sphere $S^{2}$ as a $Δ$ -complex and check the count of simplices.

Step 1. Take two $0$ -simplices $v_{0}$ and $v_{1}$ (the north and south poles), two $1$ -simplices $e_{E}$ and $e_{W}$ (the eastern and western halves of the equator, each running from $v_{0}$ to $v_{1}$ ), and two $2$ -simplices $σ_{N}$ and $σ_{S}$ (the northern and southern hemispheres).

Step 2. Each hemisphere is a triangle with two boundary edges along the equator and the third boundary edge collapsed to the pole opposite to it. So $σ_{N}$ has boundary $e_{E}$ on one side, $e_{W}$ on another, and a constant path at $v_{0}$ on the third; symmetrically for $σ_{S}$ with $v_{1}$ in place of $v_{0}$ .

Step 3. Count the simplices and check the Euler characteristic: $$ \chi(S^2) = (\text{vertices}) - (\text{edges}) + (\text{faces}) = 2 - 2 + 2 = 2, $$ matching the known $χ (S^{2}) = 2$ .

What this tells us: the combinatorial data of a $Δ$ -complex is purely finite — two vertices, two edges, two faces realise a sphere — and the Euler characteristic is the basic check that the data realises the intended space. The torus, Klein bottle, and $S^{n}$ each have a standard simplest $Δ$ -structure, and learning to recognise these standard forms is the first practical skill.

Check your understanding [Beginner]

Formal definition [Intermediate+]

Let $Δ_{inj}$ denote the category whose objects are the finite linearly ordered sets $[n] = {0 < 1 < \dots < n}$ for $n \geq 0$ and whose morphisms are the order-preserving injections. This is the wide subcategory of the simplex category $Δ$ obtained by discarding the order-preserving surjections.

Definition (semi-simplicial set). A semi-simplicial set (equivalently, a $Δ$ -set in the sense of Rourke-Sanderson) is a functor $X_{∙} : Δ_{inj}^{op} \to Set$ . Concretely, this is a sequence of sets $X_{0}, X_{1}, X_{2}, \dots$ together with face maps $d_{i} : X_{n} \to X_{n - 1}$ for $0 \leq i \leq n$ satisfying the simplicial identities $$ d_i d_j = d_{j-1} d_i \quad \text{for } i < j. $$ A morphism of semi-simplicial sets $f : X_{∙} \to Y_{∙}$ is a natural transformation: a family of maps $f_{n} : X_{n} \to Y_{n}$ commuting with all face maps. The resulting category is denoted $sSet_{+}$ .

Definition (geometric realisation). Let $Δ^{n} = {(t_{0}, \dots, t_{n}) \in R^{n + 1} : t_{i} \geq 0, \sum t_{i} = 1}$ denote the standard topological $n$ -simplex, and let $δ_{i} : Δ^{n - 1} \to Δ^{n}$ denote the $i$ -th face inclusion $(t_{0}, \dots, t_{n - 1}) \mapsto (t_{0}, \dots, t_{i - 1}, 0, t_{i}, \dots, t_{n - 1})$ . The geometric realisation of a semi-simplicial set $X_{∙}$ is the topological space $$ |X_\bullet| = \left( \coprod_{n \geq 0} X_n \times \Delta^n \right) \bigg/ \sim $$ where $(d_{i} x, t) \sim (x, δ_{i} t)$ for every $x \in X_{n}$ and $t \in Δ^{n - 1}$ , equipped with the quotient topology from the discrete topology on each $X_{n}$ and the standard topology on each $Δ^{n}$ .

Definition ( $Δ$ -complex structure on a space). A $Δ$ -complex structure on a topological space $X$ is a collection of characteristic maps $σ_{α} : Δ^{n_{α}} \to X$ such that:

(i) The restriction $σ_{α} ∣_{\overset{˚}{Δ}^{n_{α}}}$ to the open interior is injective, and every point of $X$ lies in the image of exactly one such open interior.
(ii) Each face $σ_{α} \circ δ_{i}$ equals a characteristic map $σ_{β}$ for some $β$ of dimension $n_{α} - 1$ .
(iii) A subset $A \subseteq X$ is open if and only if $σ_{α}^{- 1} (A)$ is open in $Δ^{n_{α}}$ for every $α$ .

A space carrying a $Δ$ -complex structure is called a $Δ$ -complex. Equivalently and more efficiently: a $Δ$ -complex on $X$ is a semi-simplicial set $X_{∙}$ together with a homeomorphism $X ≅ ∣ X_{∙} ∣$ .

Examples.

Torus. Take $X_{0} = {v}$ , $X_{1} = {a, b, c}$ , $X_{2} = {U, L}$ . Face maps: $d_{0} U = a$ , $d_{1} U = c$ , $d_{2} U = b$ , $d_{0} L = b$ , $d_{1} L = c$ , $d_{2} L = a$ . All face maps from $X_{1}$ to $X_{0}$ send their argument to $v$ . The geometric realisation is the two-torus $T^{2}$ , presented as a square with opposite edges identified and one diagonal $c$ dividing it into the two triangles $U$ and $L$ .

Klein bottle. The same simplex counts ( $1$ vertex, $3$ edges, $2$ triangles) with one face map flipped: $d_{2} L = a$ replaced by $d_{2} L = a$ but with the opposite orientation, encoded by relabelling. Concretely: $X_{0} = {v}$ , $X_{1} = {a, b, c}$ , $X_{2} = {U, L}$ , with $d_{0} U = a$ , $d_{1} U = c$ , $d_{2} U = b$ , $d_{0} L = a$ , $d_{1} L = c$ , $d_{2} L = b$ . The geometric realisation is the Klein bottle.

Sphere $S^{n}$ (the two-simplex form). Take one $0$ -simplex $v_{0}$ and one $n$ -simplex $σ$ whose entire boundary is collapsed to $v_{0}$ : every face map sends $σ$ to the unique $(n - 1)$ -simplex which is itself a face of some lower-dimensional simplex with everything collapsing to $v_{0}$ . The geometric realisation is the suspension of a point, iteratively — that is, $S^{n}$ .

Real projective space $RP^{n}$ . Take one simplex in each dimension from $0$ to $n$ , with face maps coming from the standard antipodal-identification cell structure. The semi-simplicial set has $X_{k} = {e_{k}}$ for $0 \leq k \leq n$ , with $d_{i} e_{k} = e_{k - 1}$ for every $i$ , and the realisation is $RP^{n}$ .

Counterexamples to common slips

The dunce cap (a $2$ -simplex with all three edges identified to a single $1$ -simplex, all in the same orientation) is a $Δ$ -complex but is contractible — the combinatorial data does not match the naive geometric expectation. Each face map is forced; only the identification of edges differs from the torus example.
A pair of disjoint open simplices identified along a single boundary face is a $Δ$ -complex; a pair of disjoint open simplices glued along a partial face (say half of an edge) is not — $Δ$ -attaching is along entire faces, not partial subsets.
A simplicial complex in the older Munkres sense (vertices form a set, simplices are subsets of vertices) is a special case of a $Δ$ -complex with the extra rigidity that each $n$ -simplex is determined by its set of vertices and is embedded in its realisation. Most $Δ$ -complexes are not simplicial complexes in this older sense: the torus's standard $Δ$ -structure has both $2$ -simplices sharing the same triple of vertices (the unique $0$ -simplex repeated three times).
A $Δ$ -complex carries no degeneracy maps. The induced full simplicial set obtained by formally adjoining degeneracies (the free degeneracy completion) has the same geometric realisation but distinct categorical data.

Key theorem with proof [Intermediate+]

Theorem (geometric realisation is CW; Hatcher §2.1). Let $X_{∙}$ be a semi-simplicial set. The geometric realisation $∣ X_{∙} ∣$ is a CW complex with one $n$ -cell for each element of $X_{n}$ . The interior $σ_{α} (\overset{˚}{Δ}^{n_{α}})$ of each characteristic map is an open $n$ -cell, and the characteristic maps $σ_{α}$ are the attaching maps.

Proof. Define $∣ X_{∙} ∣^{(k)}$ to be the image in $∣ X_{∙} ∣$ of $∐_{n \leq k} X_{n} \times Δ^{n}$ under the quotient map. We claim $∣ X_{∙} ∣^{(k)}$ is the $k$ -skeleton of a CW structure with one $n$ -cell for each element of $X_{n}$ .

Cells and attaching maps. For each $x \in X_{n}$ and $n \geq 0$ , define the characteristic map $$ \sigma_x : \Delta^n \to |X_\bullet|, \qquad t \mapsto [x, t]. $$ The restriction $σ_{x} ∣_{\overset{˚}{Δ}^{n}}$ is injective by the equivalence-relation analysis: two points $(x, t)$ and $(y, s)$ with $t, s$ in open interiors and $[x, t] = [y, s]$ in the quotient must have $x = y$ and $t = s$ , since every generating relation $(d_{i} x^{'}, t^{'}) \sim (x^{'}, δ_{i} t^{'})$ moves the simplex-coordinate from an open interior of dimension $n - 1$ to a boundary point of dimension $n$ , never between two open interiors. The complement $σ_{x} (\partial Δ^{n})$ lies in $∣ X_{∙} ∣^{(n - 1)}$ because $δ_{i} (Δ^{n - 1})$ is the image of the face $d_{i} x \in X_{n - 1}$ paired with $Δ^{n - 1}$ . So each $σ_{x}$ has the structure of a characteristic map for an $n$ -cell with attaching map into $∣ X_{∙} ∣^{(n - 1)}$ .

Disjointness of open cells. The previous paragraph shows distinct $x, y \in X_{n}$ produce disjoint open cells $σ_{x} (\overset{˚}{Δ}^{n})$ and $σ_{y} (\overset{˚}{Δ}^{n})$ in $∣ X_{∙} ∣$ , and that the union of open cells covers $∣ X_{∙} ∣$ . So the cell decomposition is well-defined.

Closure-finiteness (C). Each closed cell $σ_{x} (Δ^{n})$ meets only finitely many open cells: the $n$ -cell $\overset{σ}{˚}_{x}$ itself, plus the open cells of $∣ X_{∙} ∣^{(n - 1)}$ that lie in $σ_{x} (\partial Δ^{n})$ . The boundary $\partial Δ^{n}$ is the union of $n + 1$ closed faces, each contributing finitely many lower-dimensional cells by induction on $n$ . So each closed cell meets finitely many open cells in total.

Weak topology (W). By construction, the topology on $∣ X_{∙} ∣$ is the quotient topology from $∐_{n} X_{n} \times Δ^{n}$ . A subset $A \subseteq ∣ X_{∙} ∣$ is closed in this quotient if and only if $σ_{x}^{- 1} (A) \subseteq Δ^{n}$ is closed for every $n$ and every $x \in X_{n}$ . This is exactly the weak-topology condition: $A$ is closed if and only if $A \cap \overline{σ_{x} (\overset{˚}{Δ}^{n})} = A \cap σ_{x} (Δ^{n})$ is closed in each closed cell.

Combining: $∣ X_{∙} ∣$ is a Hausdorff space (the open cells separate points by the open-interior injectivity argument) with a cell decomposition satisfying (C) and (W), hence a CW complex. $□$

Corollary. Every $Δ$ -complex is a CW complex. Singular homology, cellular homology, and simplicial $Δ$ -homology of a $Δ$ -complex all agree.

The first sentence is immediate from the theorem. The second follows from the comparison theorem (Hatcher 2.27) identifying simplicial $Δ$ -homology with singular homology, together with the cellular-equals-singular agreement (Hatcher 2.35) on any CW complex.

Bridge. This theorem builds toward the entire combinatorial-topology infrastructure of the chapter: every space with a $Δ$ -structure inherits a CW structure, and the cellular boundary formula on that CW structure agrees with the alternating-sum face boundary on the underlying semi-simplicial chain complex. The foundational reason this works is that the open interiors of the standard simplices already provide disjoint open cells and the face inclusions already provide the attaching data, so no additional choices need to be made. This is exactly the pattern that appears again in 03.12.13 (cellular homology) where cellular chains are read off directly from the cell decomposition without computing singular chains. Putting these together, the bridge is the recognition that the combinatorial face data of a semi-simplicial set is rich enough to determine the CW structure, the cellular chain complex, and the homology, all without invoking the singular complex. The central insight is that semi-simplicial sets are the universal combinatorial source for CW homology with one cell per simplex: anything one wants to compute from the cell structure factors through the semi-simplicial data, and this same combinatorial-to-homotopical translation appears again in 03.12.12 (simplicial homology) where the comparison chain map identifies $Δ$ -homology with singular homology.

Exercises [Intermediate+]

Exercise 4 (medium, short-answer).

Show that the geometric realisation of a semi-simplicial set is a CW complex by exhibiting the attaching map for an arbitrary $n$ -simplex $x \in X_{n}$ explicitly in terms of the face maps $d_{i} x \in X_{n - 1}$ .

Hint

The attaching map $\partial Δ^{n} \to ∣ X_{∙} ∣^{(n - 1)}$ is determined by where it sends each of the $n + 1$ faces.

Answer

The attaching map for an $n$ -cell labelled $x \in X_{n}$ is the composition $\partial Δ^{n} ⨆ δ_{i} ∐_{i = 0}^{n} Δ^{n - 1} \to ∣ X_{∙} ∣^{(n - 1)}$ , where the second map sends the $i$ -th copy of $Δ^{n - 1}$ to the characteristic map of $d_{i} x \in X_{n - 1}$ . The compatibility across the $(n - 2)$ -faces is exactly the simplicial identity $d_{i} d_{j} = d_{j - 1} d_{i}$ for $i < j$ : a point on the codimension-two face $δ_{i} δ_{j} Δ^{n - 2} = δ_{j - 1} δ_{i} Δ^{n - 2}$ is sent to $d_{i} d_{j} x = d_{j - 1} d_{i} x$ , ensuring the two routes through the boundary agree. Rubric: full credit for the explicit attaching map plus the codimension-two compatibility check via the simplicial identity.

Exercise 5 (medium, short-answer).

Construct a $Δ$ -complex structure on the genus- $2$ orientable surface $Σ_{2}$ . State the simplex counts in each dimension and verify $χ (Σ_{2}) = - 2$ .

Hint

Present $Σ_{2}$ as an octagon with edges identified in the standard pattern $a_{1} b_{1} a_{1}^{- 1} b_{1}^{- 1} a_{2} b_{2} a_{2}^{- 1} b_{2}^{- 1}$ . Subdivide into triangles by diagonals from one vertex.

Answer

The octagon is subdivided into $6$ triangles by adding $5$ diagonals from one vertex. After identifications: one $0$ -simplex (all octagon vertices become a single point), $4 + 5 = 9$ $1$ -simplices (four boundary-edge classes $a_{1}, b_{1}, a_{2}, b_{2}$ plus five diagonals), and $6$ $2$ -simplices. Euler characteristic: $1 - 9 + 6 = - 2 = χ (Σ_{2})$ . The general formula for $Σ_{g}$ : one vertex, $2 g + (4 g - 2) = 6 g - 2$ edges (counting diagonals), $4 g - 2$ triangles, giving $χ (Σ_{g}) = 1 - (6 g - 2) + (4 g - 2) = 1 - 2 g = 2 - 2 g$ . Rubric: full credit for the explicit subdivision and the Euler-characteristic verification.

Exercise 6 (medium, symbolic).

Let $X_{∙}$ be a semi-simplicial set. Define the $Z$ -chain complex $C_{n}^{Δ} = Z [X_{n}]$ (the free abelian group on $X_{n}$ ) with boundary $\partial (x) = \sum_{i = 0}^{n} (- 1)^{i} d_{i} x$ . Verify $\partial \circ \partial = 0$ .

Hint

Expand $\partial (\partial x)$ as a double sum and re-index using the simplicial identity $d_{i} d_{j} = d_{j - 1} d_{i}$ for $i < j$ .

Answer

$\partial (\partial x) = i, j \sum (- 1)^{i + j} d_{i} d_{j} x = i < j \sum (- 1)^{i + j} d_{i} d_{j} x + i \geq j \sum (- 1)^{i + j} d_{i} d_{j} x .$ Split the first sum and apply $d_{i} d_{j} = d_{j - 1} d_{i}$ : $i < j \sum (- 1)^{i + j} d_{i} d_{j} x = i < j \sum (- 1)^{i + j} d_{j - 1} d_{i} x .$ Re-index by setting $i^{'} = j - 1$ , $j^{'} = i$ so that $i^{'} \geq j^{'}$ (since $i < j$ iff $j - 1 \geq i$ ): $= i^{'} \geq j^{'} \sum (- 1)^{i^{'} + j^{'} + 1} d_{i^{'}} d_{j^{'}} x = - i \geq j \sum (- 1)^{i + j} d_{i} d_{j} x .$ This cancels the second sum exactly. So $\partial\partial = 0$ . Rubric: full credit for the explicit re-indexing argument.

Exercise 7 (hard, short-answer).

The free degeneracy completion of a semi-simplicial set $X_{∙}$ is the simplicial set $X_{∙}$ with $X_{n} = ∐_{[n] ↠ [k]} X_{k}$ , where the coproduct runs over order-preserving surjections. Show that the geometric realisation $∣ X_{∙} ∣$ in the simplicial-set sense is homeomorphic to the geometric realisation $∣ X_{∙} ∣$ in the semi-simplicial sense.

Hint

The simplicial-set realisation quotients out also along degeneracies $(s_{i} x, t) \sim (x, σ_{i} t)$ where $σ_{i}$ is the $i$ -th degeneracy of the topological simplex. The degeneracies of $X_{∙}$ are inserted formally, and the corresponding equivalence relation collapses them back.

Answer

By construction the elements of $X_{n}$ are formal pairs $(α, x)$ where $α : [n] ↠ [k]$ is an order-preserving surjection and $x \in X_{k}$ . The face maps act on the source $[n]$ via the simplicial-set theory and on $x$ via the original face structure of $X_{∙}$ . The degeneracies act only on $α$ : $s_{i} (α, x) = (α \circ σ_{i}, x)$ where $σ_{i} : [n + 1] ↠ [n]$ is the standard $i$ -th degeneracy of $Δ$ .

In the simplicial-set realisation, the equivalence relation $(s_{i} x^{'}, t) \sim (x^{'}, σ_{i} t)$ identifies the pair $((α \circ σ_{i}, x), t)$ with $((α, x), σ_{i} t)$ . Iterating, the pair $((α, x), t)$ for $α$ a non-identity surjection $[n] ↠ [k]$ is identified with the pair $((id, x), α_{*} t)$ for the corresponding map $α_{*} : Δ^{n} \to Δ^{k}$ . The latter is a $k$ -simplex's data, paired with the image of $t$ under the surjection-induced map. So every point of $∣ X_{∙} ∣$ is represented by a non-degenerate pair $(x, s)$ with $x \in X_{k}$ and $s \in Δ^{k}$ , and the resulting space is exactly $∣ X_{∙} ∣$ .

The other direction is the canonical map: every point of $∣ X_{∙} ∣$ has a representative pair $(x, s)$ in some $X_{k} \times Δ^{k}$ , which lifts to a representative of $∣ X_{∙} ∣$ by using the identity surjection $id : [k] ↠ [k]$ . The two maps are mutually inverse continuous bijections, hence homeomorphisms on the quotient topology. Rubric: full credit for the bijection construction and the topology-matching argument.

Exercise 8 (hard, short-answer).

Show that the singular complex $Sing (X)$ of a topological space $X$ is canonically a semi-simplicial set when one forgets degeneracies, but that the resulting semi-simplicial structure does not generally satisfy the Kan extension condition.

Hint

The singular $n$ -simplices of $X$ are continuous maps $Δ^{n} \to X$ . Face maps are restrictions to faces. The Kan condition requires every horn $Λ_{k}^{n}$ to extend to a full simplex $Δ^{n}$ .

Answer

For any topological space $X$ , define $Sing (X)_{n} = {σ : Δ^{n} \to X continuous}$ . The face map $d_{i} : Sing (X)_{n} \to Sing (X)_{n - 1}$ is restriction along the $i$ -th face inclusion $δ_{i} : Δ^{n - 1} ↪ Δ^{n}$ . The simplicial identity $d_{i} d_{j} = d_{j - 1} d_{i}$ for $i < j$ follows from the corresponding identity for face inclusions in $Δ_{inj}$ . This defines a semi-simplicial set $Sing_{+} (X)$ obtained by forgetting the degeneracies from the usual singular simplicial set $Sing (X)$ .

The Kan condition asks that every horn $Λ_{k}^{n}$ — a configuration of $n$ singular $(n - 1)$ -simplices fitting together along their common faces, missing the $k$ -th face — extends to a singular $n$ -simplex. The usual full simplicial set $Sing (X)$ is a Kan complex (Hurewicz / Moore), with the degeneracies playing an essential role: the extension is constructed by a retraction $Δ^{n} \to Λ_{k}^{n}$ that uses both face and degeneracy data on the source. Without degeneracies, the retraction argument fails: a horn in $Sing_{+} (X)$ may not extend, since the candidate extension obtained by adjoining a degenerate filler is no longer available within the semi-simplicial framework. Explicit counterexample: take $X$ a discrete two-point space, $n = 2$ , $k = 1$ ; the horn consisting of two distinct edges fitting along a vertex has no semi-simplicial filler because there are no $2$ -simplices in $Sing_{+} (X)$ at all (a triangle into a discrete space is degenerate, hence not part of the semi-simplicial set after forgetting degeneracies). Rubric: full credit for the construction plus an explicit Kan-condition counterexample.

Advanced results [Master]

Theorem (combinatorial equivalence of the two pictures; Hatcher §2.1, Rourke-Sanderson 1971). The functor sending a $Δ$ -complex structure on a topological space $X$ to its underlying semi-simplicial set, and the functor sending a semi-simplicial set $X_{∙}$ to its geometric realisation $∣ X_{∙} ∣$ equipped with the canonical $Δ$ -structure, are inverse to each other up to natural homeomorphism on the realisations and natural isomorphism on the semi-simplicial sets.

This is the abstract content of the geometric-realisation construction: $Δ$ -complexes and semi-simplicial sets are two presentations of the same data, and any combinatorial assertion about one transfers to the other. Rourke-Sanderson formulated the modern category-theoretic version; Hatcher works in the geometric language throughout.

Theorem (semi-simplicial homology equals singular homology; Hatcher Theorem 2.27). Let $X$ be a topological space with a $Δ$ -complex structure realised by the semi-simplicial set $X_{∙}$ . The chain map $C_{∙}^{Δ} (X) \to C_{∙}^{sing} (X)$ sending each $n$ -simplex $x \in X_{n}$ to its characteristic map $σ_{x} : Δ^{n} \to X$ , viewed as a singular $n$ -simplex, induces an isomorphism on homology.

The proof packages a relative argument using the skeletal filtration $∣ X_{∙} ∣^{(k)}$ and the long exact sequence of the pair $(∣ X_{∙} ∣^{(k)}, ∣ X_{∙} ∣^{(k - 1)})$ . The induction step uses excision and the fact that $∣ X_{∙} ∣^{(k)} /∣ X_{∙} ∣^{(k - 1)}$ is a wedge of $k$ -spheres, one per $k$ -simplex.

Theorem (failure of the model-structure transfer; Goerss-Jardine §I). The category $sSet_{+}$ of semi-simplicial sets does not admit a model structure for which the geometric-realisation functor $∣ \cdot ∣ : sSet_{+} \to Top$ is a left Quillen functor producing the standard Quillen-Serre model structure on $Top$ . The obstruction is exactly the absence of degeneracies, which are required for the lifting properties of Kan fibrations.

The full simplicial set theory has degeneracies precisely so that the Kan-complex / Kan-fibration framework supports a Quillen model structure (Quillen 1967 Homotopical Algebra SLN 43). Semi-simplicial sets are adequate for chain-level constructions and singular-homology-style computations, but not for the higher-categorical machinery. The free degeneracy completion $sSet_{+} \to sSet$ is the left adjoint that bridges the two settings whenever the model-theoretic structure is needed.

Theorem (nerve characterisation). Let $Δ_{inj}$ act as a category on its own; the resulting functor category $Fun (Δ_{inj}^{op}, Set) = sSet_{+}$ . The Yoneda embedding $Δ_{inj} ↪ sSet_{+}$ sends $[n]$ to the representable presheaf $Δ_{+}^{n}$ whose geometric realisation is the standard topological $n$ -simplex. The realisation functor extends uniquely (up to natural isomorphism) to the colimit-preserving functor $sSet_{+} \to Top$ classified by the nerve adjunction.

This is the abstract justification for the explicit coend formula $∣ X_{∙} ∣ = \int^{n} X_{n} \times Δ^{n}$ , identifying the geometric realisation as the left Kan extension of $[n] \mapsto Δ^{n}$ along the Yoneda embedding.

Theorem (semi-simplicial replacement of $Sing$ ). For every topological space $X$ , the forgetful functor $U : sSet \to sSet_{+}$ sends the singular simplicial set $Sing (X)$ to a semi-simplicial set $Sing_{+} (X)$ whose chain complex, with the alternating-sum boundary, recovers the singular chain complex of $X$ exactly. Consequently the chain-level homology of $Sing_{+} (X)$ is singular homology.

The proof is direct: the chain group $C_{n}^{Δ} (Sing_{+} (X)) = Z [Sing (X)_{n}]$ equals the singular chain group $C_{n} (X)$ , and the boundary maps coincide on the nose. The point is that the chain-level construction does not see degeneracies: the alternating-sum boundary uses only face maps. This is the deepest reason that semi-simplicial sets suffice for ordinary homology.

Theorem (degeneracy completion preserves realisation). Let $L : sSet_{+} \to sSet$ denote the left adjoint to the forgetful functor $U : sSet \to sSet_{+}$ . For every semi-simplicial set $X_{∙}$ , the canonical map $∣ X_{∙} ∣ \to ∣ L (X_{∙}) ∣$ is a homeomorphism.

The degeneracies introduced by $L$ are formal and collapse under realisation, by the degeneracy-quotient relation $(s_{i} x, t) \sim (x, σ_{i} t)$ in the simplicial-set realisation. This makes the bridge between semi-simplicial and full simplicial pictures coherent on realisations while differing categorically.

Theorem (compatibility with the standard $Δ$ -structures). The standard $Δ$ -structures on $S^{n}$ , $RP^{n}$ , the orientable genus- $g$ surface $Σ_{g}$ , and the non-orientable genus- $g$ surface $N_{g}$ are explicit semi-simplicial sets whose chain complexes recover the known integer homologies. For $S^{n}$ : one $0$ -simplex, one $n$ -simplex, zero in between, giving $H_{k} = Z$ for $k = 0, n$ and zero otherwise. For $RP^{n}$ : one simplex per dimension, with boundary $\partial e_{k} = (1 + (- 1)^{k}) e_{k - 1}$ , giving the standard projective homology. For $Σ_{g}$ : the simplex counts from the octagon subdivision with the prescribed face maps, giving $H_{0} = H_{2} = Z$ and $H_{1} = Z^{2 g}$ .

These computations are direct: write the chain complex, compute kernel and image of each boundary, read off the homology. The point is that the semi-simplicial framework reduces the computation of homology of any space with a $Δ$ -structure to a finite-rank linear-algebra problem over $Z$ .

Synthesis. The $Δ$ -complex / semi-simplicial set picture is the foundational reason that singular homology factors through a combinatorial intermediate: any space with a $Δ$ -structure has its homology determined by the face data of finitely many simplices, and this data is exactly what the semi-simplicial framework records. The central insight is that the chain-level boundary uses only face maps, so the entire degeneracy machinery of the full simplicial-set theory is overhead for ordinary homology. Putting these together, the bridge from combinatorial face data to topological homology runs through three identifications: semi-simplicial chain homology equals simplicial $Δ$ -homology, $Δ$ -homology equals singular homology (Hatcher 2.27), and singular homology equals cellular homology on the CW structure inherited from the $Δ$ -decomposition (Hatcher 2.35). The construction here is dual to the full simplicial-set theory in the categorical sense that the forgetful functor $U : sSet \to sSet_{+}$ has the free-degeneracy-completion functor $L$ as its left adjoint, with $∣ L (X_{∙}) ∣$ homeomorphic to $∣ X_{∙} ∣$ . This same combinatorial-to-homotopical mechanism appears again in 03.12.13 (cellular homology), where the cellular chain complex is read off directly from the cell structure with no recourse to singular chains — generalising the $Δ$ -picture to arbitrary cell decompositions. The bridge is the recognition that combinatorial chain data is the right intermediate between geometry and homological algebra, and the appropriate categorical level depends on what one wants to compute: homology requires only faces, homotopy requires faces and degeneracies — the Kan-complex / Quillen-model story in 03.12.10 (CW complex) and beyond.

Full proof set [Master]

Proposition (simplicial identities are necessary and sufficient). A functor $X : Δ_{inj}^{op} \to Set$ is the same data as a sequence of sets $X_{n}$ for $n \geq 0$ together with face maps $d_{i} : X_{n} \to X_{n - 1}$ for $0 \leq i \leq n$ satisfying $d_{i} d_{j} = d_{j - 1} d_{i}$ for $i < j$ .

Proof. Every morphism in $Δ_{inj}$ is a composition of the elementary face inclusions $δ_{i} : [n - 1] ↪ [n]$ skipping the $i$ -th element. The defining relations among compositions of $δ_{i}$ 's are exactly $δ_{j} δ_{i} = δ_{i} δ_{j - 1}$ for $i < j$ , which dualise on $Set$ to $d_{i} d_{j} = d_{j - 1} d_{i}$ for $i < j$ . A functor on $Δ_{inj}^{op}$ is determined by its action on each $[n]$ and on each $δ_{i}$ , subject to the dual relation. So the functor and the face-map data are interchangeable. $□$

Proposition (geometric realisation is functorial). The construction $X_{∙} \mapsto ∣ X_{∙} ∣$ extends to a functor $∣ \cdot ∣ : sSet_{+} \to Top$ . For every morphism of semi-simplicial sets $f : X_{∙} \to Y_{∙}$ , the induced map $∣ f ∣ : ∣ X_{∙} ∣ \to ∣ Y_{∙} ∣$ is continuous and natural.

Proof. Define $∣ f ∣$ on representatives by $[x, t] \mapsto [f_{n} (x), t]$ for $x \in X_{n}$ and $t \in Δ^{n}$ . This descends to the quotient because $f$ commutes with face maps: if $(d_{i} x, t) \sim (x, δ_{i} t)$ , then $(d_{i} f (x), t) = (f (d_{i} x), t) \sim (f (x), δ_{i} t)$ , so the relation is preserved. Continuity on the quotient follows from continuity on $∐_{n} X_{n} \times Δ^{n}$ , which is automatic since $X_{n}$ carries the discrete topology and $f_{n}$ is a function. Naturality and functoriality are direct from the formula. $□$

Proposition (realisation commutes with colimits). The realisation functor $∣ \cdot ∣ : sSet_{+} \to Top$ preserves all small colimits.

Proof. A diagram of semi-simplicial sets has colimit computed levelwise: $(colim_{i} X_{∙}^{(i)})_{n} = colim_{i} X_{n}^{(i)}$ . The realisation of this colimit is $(∐_{n} (colim_{i} X_{n}^{(i)}) \times Δ^{n}) / \sim$ . By the universal property of colimits, $colim_{i} X_{n}^{(i)} \times Δ^{n} = colim_{i} (X_{n}^{(i)} \times Δ^{n})$ (using that $- \times Δ^{n}$ is left adjoint to the function-space functor on $Top$ for $Δ^{n}$ compact Hausdorff, hence preserves all colimits). The quotient by the equivalence relation commutes with colimits because the relation is generated by levelwise compatible identifications. Combining, $∣ colim_{i} X_{∙}^{(i)} ∣ = colim_{i} ∣ X_{∙}^{(i)} ∣$ . $□$

Proposition (singular complex factorisation). The singular complex functor $Sing : Top \to sSet$ composed with the forgetful functor $U : sSet \to sSet_{+}$ produces a functor $Sing_{+} : Top \to sSet_{+}$ whose chain complex with the alternating-sum boundary coincides with the singular chain complex.

Proof. The chain group $C_{n} (Sing_{+} (X)) = Z [Sing (X)_{n}]$ equals by definition the singular chain group $C_{n} (X)$ . The semi-simplicial boundary is $\partial (σ) = \sum_{i = 0}^{n} (- 1)^{i} d_{i} σ$ where $d_{i} σ$ is the restriction of $σ$ to the $i$ -th face. This is the same formula as the singular boundary. So the two chain complexes coincide on the nose, and their homologies — semi-simplicial chain homology of $Sing_{+} (X)$ and singular homology of $X$ — are identical. $□$

Proposition (Hatcher Theorem 2.27, sketch). Let $X$ be a topological space with a $Δ$ -structure $X_{∙}$ . The chain map $i : C_{∙}^{Δ} (X) \to C_{∙}^{sing} (X)$ sending each $n$ -simplex $x$ to its characteristic map $σ_{x}$ induces an isomorphism $i_ : H_n^\Delta(X) \to H_n^{\mathrm{sing}}(X) $f or e v er y$ n$.*

Proof sketch. Filter $X$ by its skeleta $X^{(k)} = ∣ X_{∙} ∣^{(k)}$ . Apply the long exact sequence of the pair $(X^{(k)}, X^{(k - 1)})$ in both $Δ$ -homology and singular homology, together with the chain map $i$ . The relative chain groups $C_{n}^{Δ} (X^{(k)}, X^{(k - 1)})$ are zero for $n \neq = k$ and equal to $Z [X_{k}]$ for $n = k$ . The relative singular chain groups $H_{n} (X^{(k)}, X^{(k - 1)})$ are zero for $n \neq = k$ and equal to $⨁_{x \in X_{k}} H_{k} (Δ^{k}, \partial Δ^{k}) = ⨁_{x \in X_{k}} Z$ for $n = k$ , by excision and the standard computation of relative singular homology of a sphere pair. The chain map $i$ identifies these relative groups, so by induction on $k$ and the five lemma, $i_{*}$ is an isomorphism on $H_{n}^{Δ} (X^{(k)})$ for every $k$ and every $n$ . Passing to the colimit over $k$ gives the result on $X$ . $□$

Proposition (free degeneracy completion exists and is left adjoint to $U$ ). The forgetful functor $U : sSet \to sSet_{+}$ has a left adjoint $L : sSet_{+} \to sSet$ , called the free degeneracy completion, given on $n$ -simplices by $L (X_{∙})_{n} = ∐_{[n] ↠ [k]} X_{k}$ .

Proof. Define $L (X_{∙})_{n}$ as stated. A simplicial-set face map $d_{i}$ on $L (X_{∙})_{n}$ is determined as follows: for a pair $(α : [n] ↠ [k], x \in X_{k})$ , the composition $α \circ δ_{i}^{n} : [n - 1] \to [k]$ factors uniquely as an injection followed by a surjection, $α \circ δ_{i}^{n} = β \circ α^{'}$ where $β \in Δ_{inj}$ and $α^{'} : [n - 1] ↠ [k^{'}]$ . Set $d_{i} (α, x) = (α^{'}, β^{*} x)$ , where $β^{*} x$ is the image of $x$ under the appropriate face map. The simplicial degeneracy $s_{i} : L (X_{∙})_{n} \to L (X_{∙})_{n + 1}$ sends $(α, x)$ to $(α \circ σ_{i}, x)$ . Checking the simplicial identities is a finite verification; the universal property is the statement that simplicial-set morphisms $L (X_{∙}) \to Y_{∙}$ are in natural bijection with semi-simplicial morphisms $X_{∙} \to U (Y_{∙})$ , by sending $f : L (X_{∙}) \to Y_{∙}$ to the restriction $X_{∙} = {(id, x)} \to U (Y_{∙})$ . $□$

Connections [Master]

Simplicial homology 03.12.12. The chain complex $C_{∙}^{Δ} (X_{∙}) = Z [X_{∙}]$ of a semi-simplicial set with alternating-sum boundary is the same object as the simplicial $Δ$ -homology chain complex of the underlying $Δ$ -complex. The two units describe the same data from two perspectives: the present unit constructs the combinatorial object and its realisation, while the simplicial-homology unit develops the chain-complex theory and computes the resulting invariants. The comparison theorem (Hatcher 2.27) bridging $Δ$ -homology and singular homology lives in the simplicial-homology unit; the present unit establishes the underlying combinatorial framework.
CW complex 03.12.10. Every $Δ$ -complex is canonically a CW complex via the geometric-realisation theorem above. The converse fails: a generic CW complex with arbitrary attaching maps need not carry any $Δ$ -structure. The class of spaces admitting a $Δ$ -structure is precisely the class of CW complexes whose attaching maps factor through the standard simplex boundary inclusions — a strictly smaller class than all CW complexes. Most spaces of interest in algebraic topology (compact manifolds, finite CW complexes, classifying spaces of discrete groups) admit $Δ$ -structures; some pathological CW complexes do not.
Cellular homology 03.12.13. When a space $X$ carries both a $Δ$ -structure and the induced CW structure, the cellular chain complex and the $Δ$ -chain complex coincide as chain complexes of free abelian groups. The cellular boundary is the alternating sum of face inclusions, which is the $Δ$ -boundary; the cells are the simplices of $X_{∙}$ . So cellular homology on the induced CW structure equals $Δ$ -homology equals singular homology, and the computation is the same in all three pictures. The cellular-homology unit develops the more general case where the cells need not be standard simplices.
Singular homology 03.12.11. The singular complex $Sing (X)$ of any topological space is a semi-simplicial set after forgetting degeneracies, and the resulting chain complex with alternating-sum boundary is the singular chain complex of $X$ . This is the deepest reason that singular homology is well-defined: it factors through the semi-simplicial category, and the chain-level boundary uses no degeneracy data. The full simplicial-set structure on $Sing (X)$ is needed only for the Kan-complex / Quillen-model framework, not for ordinary homology.
Eilenberg-Zilber theorem. The semi-simplicial framework supports the Eilenberg-Zilber theorem on the chain complex of a product: $C_{∙} (X \times Y) ≃ C_{∙} (X) \otimes C_{∙} (Y)$ as chain complexes up to chain homotopy. This is the algebraic input to the Künneth formula on the product of two spaces, both presented as $Δ$ -complexes. The combinatorial pairing of simplices in $X$ with simplices in $Y$ to produce simplices in $X \times Y$ uses an explicit shuffle formula, originating in the 1950 Eilenberg-Zilber paper on semi-simplicial complexes.
Eilenberg-MacLane spaces and the iterated bar construction 03.12.05, 03.08.04. The Eilenberg-MacLane space $K (π, n)$ admits a canonical simplicial model: as the geometric realisation $∣ B^{n} π ∣$ of the iterated bar construction on an abelian group $π$ , with $B π$ the standard semi-simplicial classifying space of the group regarded as a one-object category, and $B^{n} π$ obtained by $n$ -fold iteration. The construction is the foundational reason that $K (π, n) = B^{n} (π)$ as a topological space — every $n$ -simplex of $B^{n} π$ is a coherent tuple of group elements, and the face maps are the bar-construction face maps inherited from May's Simplicial Objects in Algebraic Topology. This identifies the cohomological invariants of $K (π, n)$ as simplicial cochains on a specific semi-simplicial set, and the classifying-space picture in 03.08.04 is the discrete-group instance ( $n = 1$ ) of the same iterated-bar machine.
Simplicial sets and geometric realisation 03.12.25. The full simplicial-set framework — adjoining the degeneracy maps $s_{i}$ to the present face-only data — is developed in 03.12.25 as the upgrade required for the Kan extension condition, the $∣ - ∣ ⊣ Sing$ adjunction, and the Quillen model structure on $sSet$ . The forgetful functor $U : sSet \to sSet_{+}$ of the present unit's Exercise 7 has the free-degeneracy completion $L$ as its left adjoint, and the degeneracy-completion theorem above states $∣ L (X_{∙}) ∣ ≅ ∣ X_{∙} ∣$ : the topology is unchanged, but the categorical input gains exactly the structure absent here. The relationship is bidirectional: every assertion at the chain-complex level transfers between the two frameworks because the alternating-sum boundary uses only face data, while assertions about Kan fibrations, horn-fillers, and higher homotopy live only on the simplicial-set side. The semi-simplicial framework is the chain-level shadow appropriate for ordinary homology; the simplicial-set framework is the universal home for combinatorial homotopy theory. The failure-of-model-structure-transfer theorem above identifies the structural reason this split is unavoidable.
Quillen model category 03.12.31. The failure-of-model-structure-transfer theorem alluded to above is sharpened in 03.12.31 into a positive statement: the Kan-Quillen model structure on $sSet$ lifts the present semi-simplicial framework to the full categorical home needed for the Quillen calculus, and the Reedy-type model structure on semi-simplicial sets sits one level beneath it. The small-object argument used to factorise maps as cofibrations followed by acyclic fibrations requires the degeneracy data, and is the structural reason the model-categorical machinery is cleanest on the full $sSet$ . The present face-only framework is therefore the natural setting for ordinary-homology computations; the upgrade to model-categorical machinery happens at 03.12.25 and 03.12.31.

Notation crosswalk. May 1967 Simplicial Objects in Algebraic Topology uses $Δ^{*}$ for the simplex category, $T (K)$ for the geometric realisation, $S (X)$ for the total singular complex, and $\partial_{i}$ for the face maps. The modern Hatcher / Goerss-Jardine notation used throughout this unit is $Δ$ (or $Δ_{inj}$ for the semi-simplicial subcategory), $∣ X_{∙} ∣$ for geometric realisation, $Sing (X)$ for the total singular complex, and $d_{i}$ for face maps. The translation is direct and notation-only; no content is gained or lost in the change.

Historical & philosophical context [Master]

The combinatorial picture of a space built from simplices with glued faces goes back to Poincaré's 1895 Analysis Situs and its supplements (1899-1904) ^{[Poincaré 1895]}, where simplicial subdivisions and their associated chain complexes were introduced as the basic tool for computing homology. The terminology simplicial complex in the modern sense — a set of vertices together with a collection of subsets ("simplices") closed under taking subsets — was codified by Alexandroff in the 1920s and by Lefschetz in his 1930 monograph Topology (AMS Colloq. Publ. 12).

The semi-simplicial framework — face maps without degeneracies — emerged in the 1940s and 1950s as the combinatorial substrate of singular homology. Eilenberg and Zilber's 1950 paper Semi-simplicial complexes and singular homology (Ann. of Math. (2) 51, 499-513) ^{[Eilenberg-Zilber 1950]} formalised the semi-simplicial object and proved the Eilenberg-Zilber theorem on the chain complex of a product. The companion 1952 monograph Foundations of Algebraic Topology by Eilenberg and Steenrod (Princeton) ^{[Eilenberg-Steenrod 1952]} developed the singular-homology theory in this framework, treating $Δ$ -complexes implicitly as the combinatorial source for singular chains.

The shift to the full simplicial-set theory — adding degeneracies to obtain a Kan-complex / Quillen-model framework — was carried out by Kan in a series of papers from 1957 onward and packaged in May's 1967 monograph Simplicial Objects in Algebraic Topology (University of Chicago Press) ^{[May 1967]}. The full simplicial-set theory subsumes the semi-simplicial theory via the free-degeneracy-completion adjunction, and provides the model-theoretic framework needed for higher homotopy theory. Quillen's 1967 Homotopical Algebra (Springer Lecture Notes 43) ^{[Quillen 1967]} established the Kan-Quillen model structure on simplicial sets, and the equivalence with the standard model structure on topological spaces.

The modern semi-simplicial framework was revived by Rourke and Sanderson in their 1971 papers $Δ$ -sets I/II (Quart. J. Math. Oxford (2) 22, 321-338 and 465-485) ^{[Rourke-Sanderson 1971]}, which reformulated the theory in categorical language and developed the homotopy theory of semi-simplicial sets directly without passing through the full simplicial-set theory. Hatcher's 2002 Algebraic Topology introduced the geometric $Δ$ -complex terminology for the realisation, and this is the language most graduate students encounter today.

The distinction between semi-simplicial and full simplicial — face maps with or without degeneracies — reflects the distinction between the chain-level and homotopy-level theories. Chain-level constructions (singular homology, ordinary cohomology, the universal coefficient theorem) use only the face data; homotopy-level constructions (Kan complexes, fibrations, the Quillen model structure, $\infty$ -categories) require the degeneracies. The semi-simplicial framework is the right level for the chapter on homology; the full simplicial-set framework is the right level for higher homotopy theory.

Bibliography [Master]

@article{Poincare1895,
  author  = {Poincar{\'e}, Henri},
  title   = {Analysis Situs},
  journal = {Journal de l'{\'E}cole Polytechnique},
  volume  = {(2) 1},
  year    = {1895},
  pages   = {1--121}
}

@article{EilenbergZilber1950,
  author  = {Eilenberg, Samuel and Zilber, Joseph A.},
  title   = {Semi-simplicial complexes and singular homology},
  journal = {Ann. of Math. (2)},
  volume  = {51},
  year    = {1950},
  pages   = {499--513}
}

@book{EilenbergSteenrod1952,
  author    = {Eilenberg, Samuel and Steenrod, Norman E.},
  title     = {Foundations of Algebraic Topology},
  publisher = {Princeton University Press},
  year      = {1952}
}

@book{May1967,
  author    = {May, J. Peter},
  title     = {Simplicial Objects in Algebraic Topology},
  publisher = {University of Chicago Press},
  year      = {1967},
  note      = {Reprinted in the Chicago Lectures in Mathematics series, 1992.}
}

@book{Quillen1967,
  author    = {Quillen, Daniel G.},
  title     = {Homotopical Algebra},
  series    = {Lecture Notes in Mathematics},
  volume    = {43},
  publisher = {Springer-Verlag},
  year      = {1967}
}

@article{RourkeSanderson1971a,
  author  = {Rourke, Colin P. and Sanderson, Brian J.},
  title   = {$\Delta$-Sets I: Homotopy Theory},
  journal = {Quart. J. Math. Oxford Ser. (2)},
  volume  = {22},
  year    = {1971},
  pages   = {321--338}
}

@article{RourkeSanderson1971b,
  author  = {Rourke, Colin P. and Sanderson, Brian J.},
  title   = {$\Delta$-Sets II: Block Bundles and Block Fibrations},
  journal = {Quart. J. Math. Oxford Ser. (2)},
  volume  = {22},
  year    = {1971},
  pages   = {465--485}
}

@book{HatcherAlgebraicTopology,
  author    = {Hatcher, Allen},
  title     = {Algebraic Topology},
  publisher = {Cambridge University Press},
  year      = {2002}
}

@book{GoerssJardine1999,
  author    = {Goerss, Paul G. and Jardine, John F.},
  title     = {Simplicial Homotopy Theory},
  series    = {Progress in Mathematics},
  volume    = {174},
  publisher = {Birkh{\"a}user},
  year      = {1999}
}

@book{tomDieckAlgebraicTopology,
  author    = {tom Dieck, Tammo},
  title     = {Algebraic Topology},
  publisher = {European Mathematical Society},
  year      = {2008}
}

@book{Lefschetz1930,
  author    = {Lefschetz, Solomon},
  title     = {Topology},
  series    = {AMS Colloquium Publications},
  volume    = {12},
  publisher = {American Mathematical Society},
  year      = {1930}
}

Prerequisites

02.01.01
02.01.02
02.01.06
03.12.10

Tier anchors

beginner: Hatcher *Algebraic Topology* §2.1 informal opening on $\Delta$-complex pictures of the torus and Klein bottle
intermediate: Hatcher §2.1; tom Dieck *Algebraic Topology* §3 (CW versus $\Delta$ framings)
master: Hatcher §2.1 (combinatorial form); Rourke-Sanderson 1971 *$\Delta$-sets I/II* (Quart. J. Math. Oxford 22); May 1967 *Simplicial Objects in Algebraic Topology*; Goerss-Jardine *Simplicial Homotopy Theory*

References

TODO_REF
Hatcher — Algebraic Topology · §2.1 (definition of $\Delta$-complex, examples of torus / Klein bottle / $\mathbb{RP}^n$, Proposition 2.1 quotient description)
TODO_REF
Eilenberg-Zilber — Semi-simplicial complexes and singular homology · Ann. of Math. (2) 51 (1950) 499-513; originator paper introducing the semi-simplicial framework as the algebraic shadow of the singular complex
TODO_REF
Rourke-Sanderson — $\Delta$-sets I/II · Quart. J. Math. Oxford Ser. (2) 22 (1971) 321-338 and 465-485; modern homotopy theory of semi-simplicial sets
TODO_REF
May — Simplicial Objects in Algebraic Topology · University of Chicago Press 1967 / 1992 reprint; foundational text on the full simplicial-set theory with degeneracies
TODO_REF
Goerss-Jardine — Simplicial Homotopy Theory · Birkhäuser PMS 174 (1999); Quillen model structure on simplicial sets and Kan complexes
TODO_REF
tom Dieck — Algebraic Topology · European Mathematical Society 2008; §3 CW versus $\Delta$-complex framings

Reviewer

TBD

Estimated time

beginner: 15m
intermediate: 35m
master: 70m