The Fundamental Group and Van Kampen's Theorem

  • The fundamental group π₁(X, x₀) classifies loops in a space up to continuous deformation, providing an algebraic invariant that captures topological "holes"
  • Van Kampen's Theorem computes π₁(X) by decomposing X into simpler open sets and combines their fundamental groups via free product with amalgamation
  • For path-connected open sets U, V covering X: π₁(X) ≅ π₁(U) *_{π₁(U∩V)} π₁(V)
  • Direct computation from definitions is hard; Van Kampen's Theorem enables practical calculations by exploiting space decompositions
  • Introduced by Poincaré (1895) and formalized by van Kampen (1933), the theorem remains essential for distinguishing topological spaces algebraically

The Fundamental Group and Van Kampen's Theorem

Overview

The fundamental group is one of the most important algebraic invariants in algebraic topology, capturing information about the "holes" and "loops" in a topological space. Given a space XX and a basepoint x0∈Xx_0 \in X, the fundamental group π1(X,x0)\pi_1(X, x_0) consists of equivalence classes of loops based at x0x_0, where two loops are considered equivalent if one can be continuously deformed into the other. This construction provides a bridge between topology and group theory, allowing us to use algebraic techniques to distinguish topological spaces. Computing fundamental groups directly from the definition can be challenging, even for relatively simple spaces. Van Kampen's Theorem provides a powerful tool for this computation by showing how to construct the fundamental group of a space from the fundamental groups of its subspaces. When a space XX can be decomposed into simpler open sets whose fundamental groups are known, Van Kampen's Theorem gives an explicit description of π1(X)\pi_1(X) in terms of a free product with amalgamation. Historically, the fundamental group was introduced by Henri Poincaré in 1895 as part of his development of algebraic topology. Egbertus van Kampen published his theorem in 1933, though similar results were discovered independently by Herbert Seifert. The theorem has become indispensable in algebraic topology, providing one of the primary computational tools alongside covering space theory and homology theory.

💡 The Core Insight

Van Kampen's Theorem shows that the fundamental group of a union of path-connected open sets is the free product of their fundamental groups, amalgamated over their intersection—transforming topological decompositions into algebraic constructions.

Definition

The Fundamental Group: Let XX be a topological space and x0∈Xx_0 \in X a basepoint. A loop based at x0x_0 is a continuous map γ:[0,1]→X\gamma: [0,1] \to X with γ(0)=γ(1)=x0\gamma(0) = \gamma(1) = x_0. Two loops γ0,γ1\gamma_0, \gamma_1 are homotopic (written γ0≃γ1\gamma_0 \simeq \gamma_1) if there exists a continuous map H:[0,1]×[0,1]→XH: [0,1] \times [0,1] \to X such that:

  • H(s,0)=γ0(s)H(s, 0) = \gamma_0(s) and H(s,1)=γ1(s)H(s, 1) = \gamma_1(s) for all s∈[0,1]s \in [0,1]
  • H(0,t)=H(1,t)=x0H(0, t) = H(1, t) = x_0 for all t∈[0,1]t \in [0,1]

The fundamental group π1(X,x0)\pi_1(X, x_0) is the set of homotopy classes [γ][\gamma] of loops based at x0x_0, with group operation given by concatenation: [γ1]⋅[γ2]=[γ1∗γ2][\gamma_1] \cdot [\gamma_2] = [\gamma_1 * \gamma_2] where

(γ1∗γ2)(s)={γ1(2s)0≤s≤1/2γ2(2s−1)1/2≤s≤1(\gamma_1 * \gamma_2)(s) = \begin{cases} \gamma_1(2s) & 0 \leq s \leq 1/2 \\ \gamma_2(2s-1) & 1/2 \leq s \leq 1 \end{cases}

A space is simply connected if π1(X,x0)\pi_1(X, x_0) is trivial for some (equivalently, every) basepoint.

Core Theory

TheoremVan Kampen's Theorem

Theorem [1]: Let XX be a path-connected space and let U,V⊆XU, V \subseteq X be path-connected open sets such that U∪V=XU \cup V = X and U∩VU \cap V is path-connected and non-empty. Let x0∈U∩Vx_0 \in U \cap V be a basepoint. Let iU:U∩V→Ui_U: U \cap V \to U, iV:U∩V→Vi_V: U \cap V \to V, jU:U→Xj_U: U \to X, and jV:V→Xj_V: V \to X denote the inclusion maps. Then: 1. The homomorphisms (jU)∗,(jV)∗:π1(U,x0),π1(V,x0)→π1(X,x0)(j_U)_*, (j_V)_*: \pi_1(U, x_0), \pi_1(V, x_0) \to \pi_1(X, x_0) induced by the inclusions satisfy (jU)∗∘(iU)∗=(jV)∗∘(iV)∗(j_U)_* \circ (i_U)_* = (j_V)_* \circ (i_V)_*. 2. The fundamental group π1(X,x0)\pi_1(X, x_0) is isomorphic to the amalgamated free product (also called pushout):

π1(X,x0)≅π1(U,x0)∗π1(U∩V,x0)π1(V,x0)\pi_1(X, x_0) \cong \pi_1(U, x_0) *_{\pi_1(U \cap V, x_0)} \pi_1(V, x_0)

3. More precisely, π1(X,x0)\pi_1(X, x_0) is the quotient group of the free product π1(U,x0)∗π1(V,x0)\pi_1(U, x_0) * \pi_1(V, x_0) by the normal subgroup generated by elements of the form (iU)∗(α)(iV)∗(α)−1(i_U)_*(\alpha) (i_V)_*(\alpha)^{-1} for all α∈π1(U∩V,x0)\alpha \in \pi_1(U \cap V, x_0).

Proof

Proof of Van Kampen's Theorem: We must show that the diagram ``` π₁(U∩V, x₀) → π₁(U, x₀) ↓ ↓ π₁(V, x₀) → π₁(X, x₀) ``` is a pushout in the category of groups.

Step 1: Surjectivity of the natural map. Let Φ:π1(U,x0)∗π1(V,x0)→π1(X,x0)\Phi: \pi_1(U, x_0) * \pi_1(V, x_0) \to \pi_1(X, x_0) be the homomorphism extending (jU)∗(j_U)_* and (jV)∗(j_V)_*. We show Φ\Phi is surjective. Let [γ]∈π1(X,x0)[\gamma] \in \pi_1(X, x_0). Since γ([0,1])\gamma([0,1]) is compact and covered by the open sets UU and VV, by the Lebesgue number lemma there exists a partition 0=s0<s1<⋯<sn=10 = s_0 < s_1 < \cdots < s_n = 1 such that for each ii, γ([si−1,si])\gamma([s_{i-1}, s_i]) lies entirely in either UU or VV. Let γi\gamma_i denote the restriction of γ\gamma to [si−1,si][s_{i-1}, s_i], reparametrized to [0,1][0,1]. Each γi\gamma_i is a path in either UU or VV from γ(si−1)\gamma(s_{i-1}) to γ(si)\gamma(s_i). Since UU and VV are path-connected, we can choose paths τi\tau_i in UU or VV (whichever contains γi\gamma_i) from x0x_0 to γ(si)\gamma(s_i). Define δi=τi−1−1∗γi∗τi\delta_i = \tau_{i-1}^{-1} * \gamma_i * \tau_i, which is a loop based at x0x_0 in either UU or VV. Then γ≃δ1∗δ2∗⋯∗δn\gamma \simeq \delta_1 * \delta_2 * \cdots * \delta_n (by homotopy using the paths τi\tau_i), so [γ]=[δ1]⋯[δn][\gamma] = [\delta_1] \cdots [\delta_n] where each [δi][\delta_i] is in the image of either (jU)∗(j_U)_* or (jV)∗(j_V)_*. Thus Φ\Phi is surjective.

Step 2: Determining the kernel of Φ\Phi. Let NN be the normal subgroup of π1(U,x0)∗π1(V,x0)\pi_1(U, x_0) * \pi_1(V, x_0) generated by {(iU)∗(α)(iV)∗(α)−1:α∈π1(U∩V,x0)}\{(i_U)_*(\alpha) (i_V)_*(\alpha)^{-1} : \alpha \in \pi_1(U \cap V, x_0)\}. We must show ker⁡(Φ)=N\ker(\Phi) = N. Clearly N⊆ker⁡(Φ)N \subseteq \ker(\Phi) since (jU)∗∘(iU)∗=(jV)∗∘(iV)∗(j_U)_* \circ (i_U)_* = (j_V)_* \circ (i_V)_*. For the reverse inclusion, suppose w=g1g2⋯gm∈ker⁡(Φ)w = g_1 g_2 \cdots g_m \in \ker(\Phi) where each gi∈π1(U,x0)∪π1(V,x0)g_i \in \pi_1(U, x_0) \cup \pi_1(V, x_0) (written in reduced form). Then the loop γ=γ1∗γ2∗⋯∗γm\gamma = \gamma_1 * \gamma_2 * \cdots * \gamma_m is null-homotopic in XX, where γi\gamma_i represents gig_i in either UU or VV. Let H:[0,1]×[0,1]→XH: [0,1] \times [0,1] \to X be a null-homotopy of γ\gamma. Since [0,1]×[0,1][0,1] \times [0,1] is compact, by the Lebesgue number lemma applied to the cover {H−1(U),H−1(V)}\{H^{-1}(U), H^{-1}(V)\}, we can subdivide the square into smaller rectangles, each mapped entirely into UU or VV. Using a careful inductive argument on these subdivisions (which we omit for brevity, but follows standard techniques in homotopy theory), we can show that ww can be transformed into the identity by:

  • Inserting or removing elements of the form (iU)∗(α)(iV)∗(α)−1(i_U)_*(\alpha)(i_V)_*(\alpha)^{-1}
  • Using relations in π1(U,x0)\pi_1(U, x_0) or π1(V,x0)\pi_1(V, x_0)

Therefore w∈Nw \in N, giving ker⁡(Φ)=N\ker(\Phi) = N.

Step 3: Universal property. To complete the proof, we verify the universal property of the pushout. Suppose GG is a group with homomorphisms ϕU:π1(U,x0)→G\phi_U: \pi_1(U, x_0) \to G and ϕV:π1(V,x0)→G\phi_V: \pi_1(V, x_0) \to G satisfying ϕU∘(iU)∗=ϕV∘(iV)∗\phi_U \circ (i_U)_* = \phi_V \circ (i_V)_*. By the universal property of the free product, there is a unique homomorphism ψ:π1(U,x0)∗π1(V,x0)→G\psi: \pi_1(U, x_0) * \pi_1(V, x_0) \to G extending ϕU\phi_U and ϕV\phi_V. The compatibility condition ensures N⊆ker⁡(ψ)N \subseteq \ker(\psi), so ψ\psi factors through the quotient π1(X,x0)≅(π1(U,x0)∗π1(V,x0))/N\pi_1(X, x_0) \cong (\pi_1(U, x_0) * \pi_1(V, x_0))/N, giving the unique homomorphism making the required diagram commute. □

∎
⚠️ Common mistake

Common Error: Van Kampen's Theorem requires both UU and VV to be open and their intersection U∩VU \cap V to be path-connected. If U∩VU \cap V is not path-connected but has multiple components, a more general version of the theorem applies involving free products over multiple amalgamations. Without path-connectedness of the intersection, the simple form of the theorem fails completely.

mathematician

Egbertus van Kampen (1933) — Published the seifert-van Kampen theorem providing the fundamental tool for computing fundamental groups via decomposition.

visualization

Van Kampen Decomposition of the Torus U V U ∩ V x₀ α β γ π₁(T²) = π₁(U) *_{π₁(U∩V)} π₁(V) ≅ ℤ × ℤ

Examples and Applications

Example1

Computing π1(S1)\pi_1(S^1), the circle. Decompose S1=U∪VS^1 = U \cup V where:

  • U=S1∖{−1}U = S^1 \setminus \{-1\} (the circle with the left-most point removed)
  • V=S1∖{1}V = S^1 \setminus \{1\} (the circle with the right-most point removed)

Both UU and VV are homeomorphic to (0,1)(0,1), hence contractible, so π1(U,x0)=π1(V,x0)={e}\pi_1(U, x_0) = \pi_1(V, x_0) = \{e\} (the trivial group). The intersection U∩VU \cap V consists of two connected components (top and bottom arcs), each contractible. However, we can instead use the decomposition where UU and VV are slightly larger open sets, so U∩VU \cap V is path-connected (a single arc). With this adjustment, U∩VU \cap V is contractible, so π1(U∩V,x0)={e}\pi_1(U \cap V, x_0) = \{e\}. By Van Kampen's Theorem:

π1(S1,x0)≅{e}∗{e}{e}={e}∗{e}={e}\pi_1(S^1, x_0) \cong \{e\} *_{\{e\}} \{e\} = \{e\} * \{e\} = \{e\}

This approach fails! The issue is that our decomposition doesn't capture the essential loop. Instead, we need a different method (like covering space theory) to show π1(S1)≅Z\pi_1(S^1) \cong \mathbb{Z}.

Correct application: Let X=S1∨S1X = S^1 \vee S^1 (two circles joined at a point). Decompose as U=U = (slightly thickened first circle) and V=V = (slightly thickened second circle). Then U∩VU \cap V is contractible, π1(U)≅Z\pi_1(U) \cong \mathbb{Z}, π1(V)≅Z\pi_1(V) \cong \mathbb{Z}, and:

π1(S1∨S1)≅Z∗{e}Z≅Z∗Z=F2\pi_1(S^1 \vee S^1) \cong \mathbb{Z} *_{\{e\}} \mathbb{Z} \cong \mathbb{Z} * \mathbb{Z} = F_2

the free group on two generators.

Example2

Computing π1(R2∖{p1,p2})\pi_1(\mathbb{R}^2 \setminus \{p_1, p_2\}), the plane with two points removed. Let p1,p2p_1, p_2 be two distinct points in R2\mathbb{R}^2. Define:

  • U=R2∖{p1}U = \mathbb{R}^2 \setminus \{p_1\} (plane with first point removed)
  • V=R2∖{p2}V = \mathbb{R}^2 \setminus \{p_2\} (plane with second point removed)

Then U∪V=R2∖{p1,p2}U \cup V = \mathbb{R}^2 \setminus \{p_1, p_2\} and U∩V=R2∖{p1,p2}U \cap V = \mathbb{R}^2 \setminus \{p_1, p_2\} union an open disk around one of the points—wait, this is incorrect.

Correct setup: Choose UU and VV as open sets that are deformation retracts of circles around each point:

  • UU deformation retracts to a circle around p1p_1, so π1(U)≅Z\pi_1(U) \cong \mathbb{Z}, generated by [α][\alpha]
  • VV deformation retracts to a circle around p2p_2, so π1(V)≅Z\pi_1(V) \cong \mathbb{Z}, generated by [β][\beta]
  • U∩VU \cap V deformation retracts to a wedge of two circles, so π1(U∩V)≅Z∗Z\pi_1(U \cap V) \cong \mathbb{Z} * \mathbb{Z}

Actually, the standard approach uses homotopy equivalence: R2∖{p1,p2}\mathbb{R}^2 \setminus \{p_1, p_2\} deformation retracts to a wedge S1∨S1S^1 \vee S^1, giving π1(R2∖{p1,p2})≅F2\pi_1(\mathbb{R}^2 \setminus \{p_1, p_2\}) \cong F_2 as computed in Example 1. More generally, π1(R2∖{p1,…,pn})≅Fn\pi_1(\mathbb{R}^2 \setminus \{p_1, \ldots, p_n\}) \cong F_n, the free group on nn generators.

Counterexample

When the intersection is not path-connected. Let XX be the union of two circles tangent at a point (the "figure-eight" or S1∨S1S^1 \vee S^1). Suppose we try to apply Van Kampen naively with UU and VV being small open neighborhoods of each circle, where U∩VU \cap V consists of two disjoint arcs near the tangent point. Since U∩VU \cap V has two path components, the standard Van Kampen theorem does not apply directly. Each component is contractible, but we cannot treat π1(U∩V)\pi_1(U \cap V) as a single group. The generalized Van Kampen theorem for disconnected intersections shows that:

π1(X)≅π1(U)∗π1(V)\pi_1(X) \cong \pi_1(U) * \pi_1(V)

since the amalgamation is trivial over the multiple components. However, using the wrong basepoint (one not in every component of U∩VU \cap V) can lead to incorrect conclusions.

Lesson: The path-connectedness hypothesis ensures a well-defined amalgamation. Without it, care must be taken with basepoints and the generalized version of the theorem.

🧠 Intuition

Geometric intuition: Imagine building a space XX by gluing together two simpler pieces UU and VV along their intersection U∩VU \cap V. A loop in XX can be "pushed" into pieces that lie mostly in UU or mostly in VV, crossing the overlap region only when necessary. Van Kampen's Theorem says that any loop in XX can be decomposed into "words" alternating between loops in UU and loops in VV. When a loop passes through the intersection U∩VU \cap V, we get a relation: a loop in U∩VU \cap V can be viewed either as a loop in UU or as a loop in VV, and these two viewpoints must be identified. The amalgamated free product precisely captures this: we take all formal words built from π1(U)\pi_1(U) and π1(V)\pi_1(V) (the free product), then impose exactly those relations that identify the same loop viewed from the UU side and the VV side (the amalgamation). Think of it as a "dictionary" where loops in the "shared region" have synonyms in both UU-language and VV-language.

Historical Development

The fundamental group was introduced by Henri Poincaré in his seminal 1895 paper "Analysis Situs," which laid the foundations for algebraic topology. Poincaré recognized that loops and their homotopy classes provided algebraic invariants for topological spaces, though his treatment was not fully rigorous by modern standards.

Egbertus van Kampen published his theorem in 1933 in the paper "On the connection between the fundamental groups of some related spaces" (American Journal of Mathematics). Independently, Herbert Seifert discovered similar results around the same time in his work on three-manifolds. For this reason, the theorem is sometimes called the Seifert-van Kampen theorem, particularly in the European literature. The modern formulation using category theory and the language of pushouts emerged in the 1940s and 1950s with the work of Saunders Mac Lane and Samuel Eilenberg, who developed categorical algebra and recognized that Van Kampen's theorem expresses the fundamental group as a functor that preserves certain colimits. This perspective revealed deep connections between topology and algebra. The theorem has been generalized in numerous ways, including versions for higher homotopy groups, fundamental groupoids, and in the context of algebraic geometry and schemes.

Connections to Other Areas

connection

Connection to Graph Theory and Combinatorial Group Theory: Via Van Kampen's theorem, computing the fundamental group of a CW complex reduces to understanding the Cayley graph structure of free products. For a 2-dimensional CW complex with one 0-cell, kk 1-cells (edges), and mm 2-cells (faces), the fundamental group has a group presentation with kk generators and mm relations. This connects directly to combinatorial group theory, where understanding word problems and normal forms in amalgamated free products is central. The theorem thus transforms topological questions into purely algebraic problems about groups defined by generators and relations, bridging geometric topology and computational group theory.

What You Need First

What This Opens Up

See Also

Homotopy Theory · Covering Spaces · Free Product · Fundamental Groupoid · CW Complexes · Seifert-van Kampen Theorem