logo

The Fundamental group of a Circle is Isomorphic to the Integer group 📂Topological Data Analysis

The Fundamental group of a Circle is Isomorphic to the Integer group

Theorem

The fundamental group π1(S1,1)\pi_{1} \left( S^{1}, 1 \right) of the unit circle S1S^{1} is isomorphic to the integer group Z\mathbb{Z}. π1(S1,1)Z \pi_{1} \left( S^{1}, 1 \right) \simeq \mathbb{Z} Here, (1,0)R2(1,0) \in \mathbb{R}^{2} is also simply denoted as 11.

Explanation

Application

Studying homotopy, one of the biggest and most important results that can be obtained is that the fundamental group of S1S^{1} is the integer group. Known applications include:

  • A topological proof of the Fundamental Theorem of Algebra
  • Brouwer Fixed Point Theorem
  • Borsuk-Ulam Theorem

However, the significance of this theorem lies in its own right, regardless of these applications. From an algebraic topology perspective, the idea of studying spaces not just with points but through loops makes S1S^{1} an essentially primary object to consider.

Intuitive Meaning

Basically, “the fundamental group of a circle being isomorphic to the integers” means nothing more than the number of times (n) a loop wraps around the circle in either the counterclockwise (+) or clockwise (-) direction, laying the algebraic groundwork to tackle some of the tougher areas of topology.

The Fundamental group of a Multi-Dimensional Sphere is Trivial

π1(Sn)0,n2 \pi_{1} \left( S^{n} \right) \simeq 0 \qquad , n \ge 2 For n2n \ge 2, the fundamental group of the nn-sphere SnS^{n} is the Trivial group.

20220622_160218.png

Intuitively, imagining n=2n = 2 when S2S^{2}, all loops can inevitably be shrunk down to a point across the surface, as shown in the figure above. For a formal proof, refer to Hatcher1.

Proof 2

Part 1. Definition of the function φ\varphi

20220422_170912.png

By the definition of the fundamental group, all loops [f]π1(S1,1)[f] \in \pi_{1} \left( S^{1}, 1 \right) must start and end at 11. Let deg+(f)\deg_{+} (f) be the number of times [f][f] makes a complete counterclockwise rotation to return to 11, and let deg(f)\deg_{-} (f) be the number of clockwise rotations when φ:π1(S1,1)Z\varphi : \pi_{1} \left( S^{1}, 1 \right) \to \mathbb{Z} is defined as follows. φ([f]):=deg+(f)deg(f) \varphi \left( \left[ f \right] \right) := \deg_{+} (f) - \deg_{-} (f)

Monodromy Theorem: Given two path f0f1f_{0} \simeq f_{1} equivalent at the starting point 1:=(1,0)1 := (1,0) on 11-sphere S1S^{1}, if their respective lifts satisfy f~0,f~1\widetilde{f}_{0}, \widetilde{f}_{1} equals f~0(0)=f~1(0)\widetilde{f}_{0} (0) = \widetilde{f}_{1} (0), then f~0(1)=f~1(1)\widetilde{f}_{0} (1) = \widetilde{f}_{1} (1) holds.

If f~(0)=0\widetilde{f} (0) = 0, then by the Monodromy Theorem, there is a unique f~:IR\widetilde{f} : I \to \mathbb{R} such that φ([f])=f~(1)\varphi \left( [f] \right) = \widetilde{f} (1), and φ\varphi is a function. Next, it is shown that φ\varphi is an isomorphism.


Part 2. φ\varphi is a homomorphism

20220430_195032.png

Let la(f)l_{a} (f) be the lift of ff starting from ap1(f(0))a \in p^{-1} \left( f(0) \right), then l0(f)=f~l_{0} (f) = \widetilde{f} holds, and for all aRa \in \mathbb{R}, (la(f))(t)=f~(t)+a \left( l_{a} (f) \right) (t) = \widetilde{f} (t) + a it follows. Here, if bRb \in \mathbb{R} is set to b:=f~(t)+ab := \widetilde{f} (t) + a, then trivially the following applies based on path operation \ast. la(fg)=la(f)lb(g)🤔 l_{a} \left( f \ast g \right) = l_{a} (f) \ast l_{b} (g) \qquad \cdots 🤔 Thus, when a=0a = 0, b=f~(1)b = \widetilde{f} (1) holds, and for [f],[g]π1(S1,1)[f] , [g] \in \pi_{1} \left( S^{1}, 1 \right), φ([f][g])=φ([fg])=(fg~)(1)=l0(fg)(1)=(l0(f)lb(g))(1)🤔=lb(g)(1)=b+g~(1)=f~(1)+g~(1)=φ([f])+φ([g]) \begin{align*} & \varphi \left( [f] [g] \right) \\ =& \varphi \left( \left[ f \ast g \right] \right) \\ =& \left( \widetilde{f \ast g} \right) (1) \\ =& l_{0} \left( f \ast g \right) (1) \\ =& \left( l_{0} (f) \ast l_{b} (g) \right) (1) \qquad \because 🤔 \\ =& l_{b} (g) (1) \\ =& b + \widetilde{g} (1) \\ =& \widetilde{f} (1) + \widetilde{g} (1) \\ =& \varphi \left( [f] \right) + \varphi \left( [g] \right) \end{align*} φ\varphi preserves the operation in π1(S1,1)\pi_{1} \left( S^{1}, 1 \right) in Z\mathbb{Z}, making it a homomorphism.


Part 3. φ\varphi is surjective

Given nZn \in \mathbb{Z}, by defining ωn:IR\omega_{n} : I \to \mathbb{R} as ωn(t):=nt\omega_{n} (t) := n t, pωn:IRS1 p \circ \omega_{n} : I \to \mathbb{R} \to S^{1} it becomes a loop with 1S11 \in S^{1} as the base point. By definition, ωn\omega_{n} is a lift of ωn(0)=0\omega_{n} (0) = 0 while being pωnp \circ \omega_{n}, and φ([pωn])=φ(pωn)=ωn(1)=n \varphi \left( [ p \circ \omega_{n} ] \right) = \varphi \left( p \circ \omega_{n} \right) = \omega_{n} (1) = n holds. Thus, for every nZn \in \mathbb{Z}, there exists [pωn]π1(S1,1)[ p \circ \omega_{n} ] \in \pi_{1} \left( S^{1}, 1 \right) satisfying φ([pωn])=n\varphi \left( [ p \circ \omega_{n} ] \right) = n, making φ\varphi surjective.


Part 4. φ\varphi is injective

Homotopy Lifting Property: A continuous function F:I2S1F : I^{2} \to S^{1} has a lift F~:I2R\widetilde{F} : I^{2} \to \mathbb{R}. Specifically, for given x0S1x_{0} \in S^{1} and x~0p1(x0)\widetilde{x}_{0} \in p^{-1} \left( x_{0} \right), there exists a unique F~\widetilde{F} such that F~(0,0)=x~0\widetilde{F} \left( 0 , 0 \right) = \widetilde{x}_{0}.

Assume φ([f])=0\varphi \left( [f] \right) = 0. Here, remember that 0Z0 \in \mathbb{Z} is not just 00, but 00 as the additive identity of integers. This means the starting and ending points of the lift f~:IR\widetilde{f} : I \to \mathbb{R} of ff are the same, and due to the uniqueness of the lift, f~(0)=f~(1)=0 \widetilde{f} (0) = \widetilde{f} (1) = 0 according to the Homotopy Lifting Property, F(0,t)=f~(t)F(1,t)=0F(s,0)=F(s,1)=0 \begin{align*} F (0,t) =& \widetilde{f} (t) \\ F (1,t) =& 0 \\ F (s,0) = F (s,1) =& 0 \end{align*} F:I2RF : I^{2} \to \mathbb{R} satisfying the condition exists. Naturally, applying the covering (projection) pp to pF:I2RS1 p \circ F : I^{2} \to \mathbb{R} \to S^{1} satisfies pF(0,t)=f(t)pF(1,t)=1pF(s,0)=pF(s,1)=1 \begin{align*} p \circ F (0,t) =& f (t) \\ p \circ F (1,t) =& 1 \\ p \circ F (s,0) = p \circ F (s,1) =& 1 \end{align*} , meaning [f]π1(S1,1)[f] \in \pi_{1} \left( S^{1}, 1 \right) is the constant path [f]=[c1][f] = [c_{1}] of π1(S1,1)\pi_{1} \left( S^{1}, 1 \right), thus kerφ={[c1]}\ker \varphi = \left\{ \left[ c_{1} \right] \right\}.

Properties of the Kernel of a Homomorphism:

  • [3]: kerϕ={e}\ker \phi = \left\{ e \right\}     \iff ϕ\phi is injective.

Lastly, by the properties of the kernel, φ\varphi is injective.

Meanwhile, Hatcher defines the domain and range conversely, proving that ϕ:Zπ1(S1,1)\phi : \mathbb{Z} \to \pi_{1} \left( S^{1}, 1 \right) is an isomorphism, though essentially arriving at the same conclusion but aiming to prove it without mentioning the Monodromy Theorem etc1.


  1. Hatcher. (2002). Algebraic Topology: p35. ↩︎ ↩︎

  2. Kosniowski. (1980). A First Course in Algebraic Topology: p139~140. ↩︎