Tutorials › Topology (Undergraduate) › Maps into Subspaces

Continuous Maps and Homeomorphisms · Tutorial 194 of 799

Maps into Subspaces

Use the inclusion map to test continuity into a subspace, then apply the criterion to familiar subsets of Euclidean space.

Intermediate 9 min read

What You'll Learn

  • State and prove the continuity criterion for a map whose target is a subspace.
  • Relate a map into a subspace to its corestriction from an ambient target.
  • Check continuity into an interval by viewing the map as real-valued.
  • Verify circle- and sphere-valued maps using their Euclidean coordinates.
  • Distinguish openness in a subspace from openness in the ambient space.

Continuity into a Subspace

A function may take values in a subset of a space, but continuity depends on the topology assigned to that subset. For a subset \(A\) of a topological space \(Y\), the subspace topology declares a set open in \(A\) when it has the form \(A\cap V\) for some open set \(V\) in \(Y\). This makes it possible to test continuity of a map into \(A\) by considering the same values in the ambient space \(Y\).

Throughout, let \((X,\mathcal{T}_X)\) and \((Y,\mathcal{T}_Y)\) be topological spaces, and let \(A\subseteq Y\) carry the subspace topology. A map \(f:X\to Y\) is continuous if the preimage of every open subset of \(Y\) is open in \(X\). Write \(i:A\to Y\) for the inclusion map, \(i(a)=a\).

Theorem (Continuity into a subspace): A map \(g:X\to A\) is continuous if and only if the composite \(i\circ g:X\to Y\) is continuous.

Proof. For any open set \(V\subseteq Y\), the set \(i^{-1}(V)\) is \(A\cap V\), which is open in \(A\) by the definition of the subspace topology. If \(g\) is continuous, then \((i\circ g)^{-1}(V)=g^{-1}(i^{-1}(V))\) is open in \(X\), so \(i\circ g\) is continuous.

Conversely, suppose \(i\circ g\) is continuous, and let \(O\) be open in \(A\). By the subspace topology, \(O=A\cap V\) for some open \(V\subseteq Y\). Since \(g(x)\in A\) for every \(x\in X\), we have \(g^{-1}(O)=g^{-1}(A\cap V)=(i\circ g)^{-1}(V)\). This is open in \(X\) by the assumed continuity of \(i\circ g\). Thus \(g\) is continuous. \(\square\)

The fact that \(g\) takes values in \(A\) is part of the statement: it is what makes \(g:X\to A\) a map and gives the preimage identity in the proof. The theorem says that once the target is \(A\), testing continuity there is exactly equivalent to testing the same function as a \(Y\)-valued map.

Corollary (Corestriction criterion): Suppose \(f:X\to Y\) and \(f(X)\subseteq A\). Define \(\widetilde f:X\to A\) by \(\widetilde f(x)=f(x)\). Then \(\widetilde f\) is continuous if and only if \(f\) is continuous.

Proof. The image condition ensures that \(\widetilde f\) is well-defined, and \(f=i\circ\widetilde f\). The theorem therefore says precisely that \(\widetilde f\) is continuous if and only if \(f\) is continuous. \(\square\)

Maps into Intervals

For an interval \(I=[0,1]\), use the subspace topology inherited from \(\mathbb{R}\) with its standard topology. A useful practical consequence is that a map with values in \(I\) can be tested as a real-valued function: prove it is real-valued and continuous, and separately check that its values lie in \(I\). The range check is essential; continuity alone does not make a function \(I\)-valued.

Worked Example: A Map into the Unit Interval

Define \(g:\mathbb{R}\to I\) by \(g(x)=\dfrac{x^2}{1+x^2}\), where \(\mathbb{R}\) has its standard topology and \(I\) has the subspace topology. Verify its range and continuity, and compute \(g(2)\).

For every real \(x\), \(x^2\geq 0\), so \(g(x)\geq 0\). Also \(1+x^2>0\), and \(x^2<1+x^2\), so \(g(x)<1\). Hence \(0\leq g(x)<1\), which verifies \(g(x)\in[0,1]\) for every \(x\). As a real-valued function, \(g\) is a quotient of polynomials with denominator \(1+x^2\), which never vanishes; the standard continuity rules for real functions therefore show that it is continuous into \(\mathbb{R}\). By the corestriction criterion, \(g:\mathbb{R}\to I\) is continuous.

Substituting \(x=2\) gives \(g(2)=\dfrac{2^2}{1+2^2}=\dfrac{4}{5}\). This value lies in \(I\), since \(0\leq \dfrac45<1\). As a further range check, \(g(0)=\dfrac{0^2}{1+0^2}=0\), and for every finite \(x\) the strict inequality \(g(x)<1\) shows that the endpoint \(1\) is not attained.

Maps into the Circle

Let \(S^1=\{(u,v)\in\mathbb{R}^2:u^2+v^2=1\}\), with the subspace topology inherited from the Euclidean topology on \(\mathbb{R}^2\). To define a map into \(S^1\), it is enough to give two continuous coordinate functions and verify that their values satisfy the equation \(u^2+v^2=1\). The continuity criterion then turns continuity as an \(\mathbb{R}^2\)-valued map into continuity as an \(S^1\)-valued map.

Worked Example: A Parametrization of the Circle

Let \(I=[0,1]\) have its subspace topology from \(\mathbb{R}\), and define \(g:I\to\mathbb{R}^2\) by \(g(t)=(\cos(2\pi t),\sin(2\pi t))\). Show that \(g\) defines a continuous map \(I\to S^1\), and evaluate it at \(t=0\), \(t=\dfrac14\), and \(t=1\).

The coordinate functions are continuous, so \(g:I\to\mathbb{R}^2\) is continuous. For every \(t\in I\), the identity \(\cos^2(2\pi t)+\sin^2(2\pi t)=1\) gives \(g(t)\in S^1\). Thus the same formula, regarded as a map \(I\to S^1\), is continuous by the corestriction criterion.

At \(t=0\), \(2\pi t=0\), so \(g(0)=(\cos 0,\sin 0)=(1,0)\). At \(t=\dfrac14\), \(2\pi t=\dfrac{\pi}{2}\), so \(g(\dfrac14)=(\cos(\dfrac{\pi}{2}),\sin(\dfrac{\pi}{2}))=(0,1)\). At \(t=1\), \(2\pi t=2\pi\), so \(g(1)=(\cos(2\pi),\sin(2\pi))=(1,0)\).

Check: At the computed point \(g(\dfrac14)=(0,1)\), the circle equation gives \(0^2+1^2=1\); at \(g(0)=g(1)=(1,0)\), it gives \(1^2+0^2=1\). These evaluations agree with the range verification for every \(t\).

Maps into Spheres

For \(S^2=\{(x,y,z)\in\mathbb{R}^3:x^2+y^2+z^2=1\}\), use the subspace topology from Euclidean \(\mathbb{R}^3\). The same method applies: establish continuity of the three coordinate functions, verify the sphere equation, and then use the subspace criterion. A formula with a denominator also requires checking that the denominator is nonzero throughout its domain.

Worked Example: A Map from the Plane into the Sphere

Give \(\mathbb{R}^2\) its Euclidean topology and define \(F:\mathbb{R}^2\to\mathbb{R}^3\) by \(F(u,v)=\left(\dfrac{2u}{1+u^2+v^2},\dfrac{2v}{1+u^2+v^2},\dfrac{u^2+v^2-1}{1+u^2+v^2}\right)\). Show that its values lie in \(S^2\), prove continuity into \(S^2\), and evaluate it at \((0,0)\) and \((\dfrac34,0)\).

Set \(r^2=u^2+v^2\), so \(r^2\geq0\) and the denominator \(1+r^2\) is strictly positive. Each coordinate is a quotient of continuous polynomial functions with nonzero denominator, so \(F:\mathbb{R}^2\to\mathbb{R}^3\) is continuous. To check its range, compute the sum of the squares of its coordinates: \[ \frac{4u^2+4v^2+(r^2-1)^2}{(1+r^2)^2} =\frac{4r^2+(r^4-2r^2+1)}{(1+r^2)^2} =\frac{r^4+2r^2+1}{(1+r^2)^2} =\frac{(1+r^2)^2}{(1+r^2)^2} =1. \] Thus \(F(u,v)\in S^2\) for every \((u,v)\), and the corestriction \(\mathbb{R}^2\to S^2\) is continuous.

At \((0,0)\), \(r^2=0\) and the denominator is \(1+0=1\), so \(F(0,0)=(0,0,-1)\). Its squared coordinate sum is \(0^2+0^2+(-1)^2=1\). At \((\dfrac34,0)\), \(r^2=(\dfrac34)^2+0^2=\dfrac9{16}\), and the denominator is \(1+\dfrac9{16}=\dfrac{25}{16}\). The first coordinate is \(\dfrac{2(\frac34)}{\frac{25}{16}}=\dfrac{3}{2}\cdot\dfrac{16}{25}=\dfrac{24}{25}\); the second is \(0\); the third is \(\dfrac{\frac9{16}-1}{\frac{25}{16}}=\dfrac{-\frac7{16}}{\frac{25}{16}}=-\dfrac7{25}\). Therefore \(F(\dfrac34,0)=(\dfrac{24}{25},0,-\dfrac7{25})\).

Check: The squared coordinate sum of this last value is \((\dfrac{24}{25})^2+0^2+(-\dfrac7{25})^2=\dfrac{576}{625}+\dfrac{49}{625}=\dfrac{625}{625}=1\), as required.

A Common Error About Relative Openness

Continuity into a subspace does not require preimages of subspace-open sets to be open in the ambient target. They must be open in the domain, and a set open in a subspace need not be open in the ambient space. Confusing these two meanings can lead to a false rejection of a continuous map.

Common pitfall. Consider the inclusion \(j:I\to\mathbb{R}\), where \(I=[0,1]\) has the subspace topology. A mistaken calculation takes the open set \(V=(-1,1)\subseteq\mathbb{R}\), obtains \(j^{-1}(V)=[0,1)\), observes that \([0,1)\) is not open in \(\mathbb{R}\), and concludes that \(j\) is not continuous. The error is the final test: \([0,1)\) is open in \(I\), because \([0,1)=I\cap(-1,1)\), with \((-1,1)\) open in \(\mathbb{R}\). More generally, for any open \(V\subseteq\mathbb{R}\), \(j^{-1}(V)=I\cap V\), which is open in \(I\). Thus \(j\) is continuous.

The reverse issue also matters: a formula may be continuous into the ambient space but fail to define a map into the proposed subspace if some value lies outside it. For example, the constant real-valued function \(h(x)=2\) is continuous as a map into \(\mathbb{R}\), but it does not define a map into \(I\), because \(h(x)=2\notin[0,1]\) for every \(x\). The range condition and the continuity test are separate checks.

Check Your Understanding

For each question, distinguish the ambient topology from the subspace topology where needed.

  1. Let \(A=[-1,1]\subseteq\mathbb{R}\), and define \(g:\mathbb{R}\to A\) by \(g(x)=\dfrac{x}{1+|x|}\). Verify the range condition and explain how continuity into \(A\) can be tested.
  2. Suppose \(g:X\to A\subseteq Y\). Write the preimage identity that relates \(g^{-1}(O)\) to \((i\circ g)^{-1}(V)\) when \(O=A\cap V\).
  3. For \(j:[0,1]\to\mathbb{R}\) the inclusion, compute \(j^{-1}((0,2))\) and state whether it is open in \([0,1]\) and in \(\mathbb{R}\).
  4. Define a continuous map from \(\mathbb{R}\) into \(S^1\) using two coordinate functions, and verify the circle equation for all inputs.
  5. Why must one check that a proposed ambient-valued map has image contained in \(A\) before calling it a map into the subspace \(A\)?