logo

Fundamental group in Algebraic Topology 📂Topological Data Analysis

Fundamental group in Algebraic Topology

Definition 1

Given a topological space XX and a unit interval I=[0,1]I = [0,1],

  1. For paths f,g:IXf, g : I \to X in XX, when f(1)=g(0)f (1) = g(0), the product or composition fgf \cdot g of two paths is defined as: fg(s):={f(2s),if s[0,1/2]g(2s1),if s[1/2,1] f \cdot g (s) := \begin{cases} f \left( 2s \right) & , \text{if } s \in [0, 1/2] \\ g \left( 2s - 1 \right) & , \text{if } s \in [1/2, 1] \end{cases}
  2. For a path f:IXf : I \to X, a path defined as f:IX\overline{f} : I \to X when f(s):=f(1s)\overline{f} (s) := f (1-s) is called the inverse path of ff.
  3. For all s1,s2Is_{1} , s_{2} \in I, a path cx0c_{x_{0}} such that cx0(s1)=cx0(s2)=x0c_{x_{0}} \left( s_{1} \right) = c_{x_{0}} \left( s_{2} \right) = x_{0}, i.e., a constant function, is called a constant path.
  4. When f(0)=f(1)=x0Xf (0) = f(1) = x_{0} \in X, i.e., the initial point and the terminal point of the path ff are the same, it is called a loop, and x0Xx_{0} \in X is called the basepoint.
  5. The set of all homotopy classes of loops ff based at x0Xx_{0} \in X is denoted as π1(X,x0)\pi_{1} \left( X , x_{0} \right). When defining a binary operation \ast on homotopy classes of two loops [f],[g]π1(X,x0)[f] , [g] \in \pi_{1} \left( X , x_{0} \right) as [f][g]:=[fg] [f] \ast [g] := \left[ f \cdot g \right] then the group π1((X,x0))\pi_{1} \left( \left( X , x_{0} \right) \right) is called the fundamental group. Usually, \ast is not even mentioned and is simply written as [f][g]=[fg][f] [g] = [f \cdot g].

Explanation

In mathematics, anything with “Fundamental” attached to it is of great importance. Initially, one may feel reluctant about how loops could lead to creating a group, but when considering their homotopy instead of just a set of loops, it becomes relatively easy to imagine its significance as it primarily concerns the properties of the topological space.

Product of Paths

20220412_130812.png

The product of paths fgf \cdot g, as evident from the equation, follows path ff until 1/2s11/2 \le s \le 1 and then follows path gg from then on. That such a product remains continuous is guaranteed by the pasting lemma.

Pasting Lemma: For a topological space X,YX,Y, given two closed sets A,BXA,B \subset X satisfying AB=XA \cup B = X, and two continuous functions f:AYf : A \to Y and g:BYg : B \to Y such that for all xABx \in A \cap B, f(x)=g(x)f(x) = g(x), then a function defined as hh is a continuous function. h(x):={f(x),xAg(x),xB h(x) : = \begin{cases} f(x), & x \in A \\ g(x), & x \in B \end{cases}

Although the concept is not difficult, the terminology may feel awkward. Since paths are essentially functions, referring to their composition can be confusing due to potential confusion with composite functions, and the term product might also be confusing if there’s any operation in XX that could be considered multiplication. However, contrary to concerns, when studying, the operation [f][g]=[fg][f] [g] = [f \cdot g] within the homotopy classes is mainly mentioned, and describing fgf \cdot g in words is rare.

Inverses and Identity Elements in Fundamental groups

20220412_130817.png 20220412_130825.png

An inverse path can be perceived as erasing the path taken in the product of paths. To intuitively understand this, consider f\overline{f} as simply being path ff but in the opposite direction.

20220412_130829.png

Then, the product fff \cdot \overline{f} with f\overline{f} for any given path ff results in a loop since its starting and ending points are the same. Important to note is that the endpoint of ff and the starting point of f\overline{f}, x1x_{1}, either precisely hits or simply remains static at x0x_{0}, which is a constant path cx0c_{x_{0}}, or is homotopically equivalent. Since π1(X,x0)\pi_{1} \left( X, x_{0} \right) is a set of homotopy classes, for all ff, ffcx0 f \cdot \overline{f} \simeq c_{x_{0}} would hold true. Seeing this, regardless of what ff is, its homotopy class [f][f], through operation with the homotopy class of f\overline{f}, [f]\left[ \overline{f} \right], always results in [cx0]\left[ c_{x_{0}} \right], and cx0c_{x_{0}} is evidently the identity element of π1(X,x0)\pi_{1} \left( X , x_{0} \right) from its definition.

Simply Connected Spaces

From the discussions so far, one might wonder about the necessity of x0x_{0} in fundamental groups.

20220412_130836.png

For example, as shown in the picture above, it’s conceivable to think of a loop that ’erases’ the path before reaching a new point x1Xx_{1} \in X, making it seem indifferent to the choice of any basepoint x0x_{0}.

Change of Basepoint in Fundamental groups: Given a topological space XX, let h:IXh : I \to X be a path from x0x_{0} to x1x_{1}. The function defined as βh:π1(X,x1)π0(X1,x0)\beta_{h} : \pi_{1} \left( X , x_{1} \right) \to \pi_{0} \left( X_{1} , x_{0} \right) according to βh[f]:=[hfh]\beta_{h} [f] := \left[ h \cdot f \cdot \overline{h} \right] is an isomorphism, known as the Change of Basepoint.

According to the above theorem, if XX is path-connected, then π1(X,x)\pi_{1} \left( X , x \right) is isomorphic regardless of the choice of basepoint xx, hence, it is sometimes represented as π1(X)\pi_{1} \left( X \right) or more concisely as just π1X\pi_{1} X.

Especially, if a space is path-connected and its fundamental group π1X\pi_{1} X is a trivial group, i.e., isomorphic to a finite group with only the identity element ee such that π1X{e}\pi_{1} X \simeq \left\{ e \right\}, then XX is considered a simply-connected space.

Conversely, this implies that the properties of a fundamental group can significantly vary with the choice of basepoint x0x_{0}, and even if it is path-connected, it is necessary to examine separately the algebraic properties it might ha


  1. Hatcher. (2002). Algebraic Topology: p26~28. ↩︎