Fundamental group - Foundations and Intuition

The Fundamental Group: A Topological Record of Loops Introduction One of the most important invariants in topology is the fundamental group. At its core, it answers a simple but powerful question: how many topologically distinct ways can a loop wrap around a space? Rather than studying individual loops, we collect all loops together into equivalence classes and give them an algebraic structure. This algebraic object captures essential information about the "holes" or "obstructions" in a space, allowing us to compare spaces topologically and detect whether they have different shapes. Loops and the Base Point The fundamental group begins with loops. A loop in a topological space $X$ is a continuous path that starts and ends at the same point. To make this concrete, we fix a particular point $x0 \in X$ called the base point, and consider loops $\ell: [0,1] \to X$ with $\ell(0) = \ell(1) = x0$. Why fix a base point? Because the fundamental group is fundamentally about tracking how paths return to where they started. The space might have different topological features at different locations, so our choice of base point matters for understanding the local topology around that point. Not all loops are considered different—two loops are treated as equivalent if one can be continuously deformed into the other without breaking the path and while keeping the base point fixed at all times. This notion of continuous deformation is called homotopy. Homotopy: Continuous Deformation of Loops To formalize what we mean by "continuous deformation," we use the concept of homotopy. Two loops $f$ and $g$ based at $x0$ are homotopic (written $f \simeq g$) if there exists a continuous map called a homotopy: $$H: [0,1] \times [0,1] \to X$$ that satisfies: $H(s, 0) = f(s)$ (at one end, we have the first loop) $H(s, 1) = g(s)$ (at the other end, we have the second loop) $H(0, t) = H(1, t) = x0$ for all $t$ (the base point never moves) Think of $H$ as recording a continuous family of loops, where as you move through the homotopy parameter $t$ from $0$ to $1$, you gradually deform loop $f$ into loop $g$. The condition that $H(0,t) = H(1,t) = x0$ ensures that every intermediate loop in this family also returns to the base point. The torus (shown above) has a hole in the middle. Some loops wrap around this hole and cannot be continuously shrunk to a point, while others can. This topological difference is captured by the fundamental group. Equivalence Classes and the Fundamental Group The homotopy relation partitions all loops at $x0$ into equivalence classes. Two loops belong to the same class if they are homotopic to each other. Each equivalence class represents a distinct topological type of loop, and the collection of all these classes is the fundamental group: $$\pi1(X, x0) = \{\text{homotopy classes of loops based at } x0\}$$ We denote the homotopy class of a loop $f$ by $[f]$. The key insight is that while there may be infinitely many loops in a space, they naturally fall into a much smaller number of equivalence classes. These classes capture all topologically distinct ways a loop can wrap around the space. Group Structure: Concatenation The fundamental group is not just a set—it has an algebraic structure given by concatenation. Given two loops $f$ and $g$ based at $x0$, we can "multiply" them by first traversing $f$ and then traversing $g$. This product $(fg)$ is defined by: $$(fg)(s) = \begin{cases} f(2s) & \text{if } 0 \le s \le \tfrac{1}{2} \\[4pt] g(2s-1) & \text{if } \tfrac{1}{2} \le s \le 1 \end{cases}$$ By speeding up each loop by a factor of 2, we ensure that the concatenated path is still traversed in the time interval $[0,1]$. The result is a new loop that starts at $x0$, follows $f$ for half the time, then follows $g$ for the remaining half time, and ends back at $x0$. Crucially, concatenation respects homotopy: if $f1 \simeq f2$ and $g1 \simeq g2$, then $(f1 g1) \simeq (f2 g2)$. This means we can define a well-defined group operation on equivalence classes: $$[f] [g] = [f g]$$ Group Axioms Concatenation gives $\pi1(X, x0)$ the structure of a group, meaning it satisfies: Identity element: The constant loop $c{x0}(s) = x0$ (which doesn't move at all) serves as the identity. Concatenating any loop with the constant loop is homotopic to the original loop: $[f] [c] = [f]$. Inverse elements: For any loop $f$, its reverse or inverse $f^{-1}$ is defined by $f^{-1}(s) = f(1-s)$, which traverses the same path but in the opposite direction. The concatenation $[f] [f^{-1}]$ is homotopic to the constant loop: $[f] [f^{-1}] = [c]$. Associativity: While concatenation at the level of loops is not strictly associative (the three different ways to parenthesize three loops involve different reparametrizations), once we pass to homotopy classes, associativity holds in the group: $([f] [g]) [h] = [f] ([g] [h])$. This is because the different parenthesizations are homotopic to each other. These properties confirm that $\pi1(X, x0)$ is indeed a group. Dependence on the Base Point A natural question arises: does the fundamental group depend on our choice of base point? For path-connected spaces (spaces where any two points can be connected by a continuous path), the answer is no—up to isomorphism. If $x0$ and $x1$ are two points in a path-connected space $X$, and $\gamma$ is any path connecting them, then we can construct an isomorphism: $$\phi: \pi1(X, x0) \to \pi1(X, x1)$$ by conjugating loops with $\gamma$. Intuitively, a loop based at $x0$ can be translated to a loop based at $x1$ by first following $\gamma$ to $x1$, then traversing the loop, then returning along $\gamma$ in reverse. In practice, this means we often write $\pi1(X)$ for a path-connected space without specifying a base point, understanding that the group is well-defined up to isomorphism. Homotopy Invariance: A Key Property One of the most important properties of the fundamental group is that it is a topological invariant—it depends only on the topological structure of a space, not on specific geometric details. Specifically, if two spaces $X$ and $Y$ are homeomorphic (topologically identical via a continuous bijection with continuous inverse) or homotopy equivalent (intuitively, have the same "shape" up to continuous deformations), then their fundamental groups are isomorphic: $$X \simeq Y \implies \pi1(X) \cong \pi1(Y)$$ This is powerful because it means we can often compute the fundamental group for a "simple" space homotopy equivalent to our original space, yet still learn about the original space's topology. Examples of Fundamental Groups To build intuition, consider these fundamental groups: Euclidean space $\mathbb{R}^n$ is contractible (can be continuously shrunk to a point), so $\pi1(\mathbb{R}^n) = \{e\}$ is the trivial group. The circle $S^1$ has $\pi1(S^1) = \mathbb{Z}$. Each loop is classified by the number of times it winds around the circle (positive for counterclockwise, negative for clockwise). The sphere $S^2$ has $\pi1(S^2) = \{e\}$ because any loop on a sphere can be continuously shrunk to a point. The torus $T^2$ (the surface of a donut) has $\pi1(T^2) = \mathbb{Z} \times \mathbb{Z}$. A loop on a torus can wrap around both "directions" of the torus independently, giving two independent winding numbers. <extrainfo> Historical Context: Monodromy and Surface Classification The concept of the fundamental group emerged from Henri Poincaré's investigation of complex analysis and algebraic topology in the late 19th century. One motivation came from monodromy: the study of how complex-valued functions change when analytically continued along loops in the complex plane. The values obtained after returning to the starting point encode topological information about singularities and branch points, captured algebraically by the fundamental group. Another major application was the classification of closed surfaces. Poincaré and his successors used the fundamental group as a key invariant to classify all compact surfaces without boundary, showing that the fundamental group (along with one bit of orientation information) completely determines the topological type of a closed surface. </extrainfo> Summary The fundamental group $\pi1(X, x0)$ is a fundamental topological invariant that: Captures topologically distinct ways loops can wrap around a space Is defined as equivalence classes of loops under homotopy Carries a group structure via concatenation Is independent of base point choice in path-connected spaces Is preserved under homeomorphism and homotopy equivalence It transforms a geometric question ("how can loops wind around this space?") into an algebraic one ("what is the structure of this group?"), making the power of algebra available for studying topology.