logo

Proof that All Isometric Mappings are Embeddings 📂Banach Space

Proof that All Isometric Mappings are Embeddings

Theorem

Let’s call (X,X),(Y,Y)(X, \left\| \cdot \right\|_{X}), (Y, \left\| \cdot \right\|_{Y}) a norm space. And let’s say f:XYf : X \to Y is an isometry. Then, ff is an embedding. In other words, ff satisfies the following two conditions.

(a) f(X)Yf(X) \subset Y

(b) f:Xf(X)f : X \to f(X) is a homeomorphism.

Proof

Strategy: Prove (b)(b) first, then prove (a)(a). Although there isn’t anything particularly difficult in each proof process, it may seem challenging due to the use of various definitions.

(b)

Lemma

Let there be two topological spaces XX and YY. For a bijective function f:XYf : X \to Y, the following three conditions are equivalent.

To show that f:Xf(X)f : X \to f(X) is a homeomorphism, we need to show that ff is bijective, continuous, and f1f^{-1} is continuous.

  • Part 1. f:Xf(X)f : X \to f(X) is bijective.

    It is trivial that f:Xf(X)f : X \to f(X) is surjective. Assume f(x1)=f(x1)f(x_{1})=f(x_{1}). Then, dy(f(x1), f(x2))=0d_{y} \big( f(x_{1}),\ f(x_{2}) \big)=0. Since ff preserves distance, dx(x1,x2)=0d_{x}(x_{1}, x_{2})=0. Therefore, x1=x2x_{1}=x_{2} and thus ff is injective.

  • Part 2. f:Xf(X)f : X \to f(X) is continuous.

    Since ff is bijective, for any yf(X)y\in f(X), there exists a unique xXx\in X satisfying f(x)=yf(x)=y. Hence, because ff is an isometry, for any positive r>0r>0, the following formula holds.

    f1(BdY(y,r))=BdX(x,r) f^{-1}\big(B_{d_{Y}}(y,r) \big)=B_{d_{X}}(x,r)

    For any open set Vf(X)V\subset f(X), since f1(V)f^{-1}(V) is an open set, ff is continuous.

  • Part 3. f1f^{-1} is continuous.

    By the same reasoning, for any xXx \in X and r>0r>0,

    f(BdX(x,r))=BdY(f(x),r) f \big( B_{d_{X}}(x,r) \big) = B_{d_{Y}}\big( f(x),r\big)

    Therefore, ff is an open mapping. Consequently, according to the lemma, f1f^{-1} is continuous.

(a)

Let us say x1,x2Xx_{1},x_{2} \in X. Then, f(x1),f(x2)f(X)f(x_{1}),f(x_{2}) \in f(X). Since XX is a norm space and thus a vector space, x1+x2=xXx_{1}+x_{2}=x\in X. The isometric map ff is linear, so f(x)=f(x1+x2)=f(x1)+f(x2)f(x)=f(x_{1}+x_{2})=f(x_{1})+f(x_{2}). Also, because of xXx\in X, f(x)f(X)f(x) \in f(X). Therefore, whenever f(x1),f(x2)f(X)f(x_{1}), f(x_{2}) \in f(X), f(x1)+f(x2)=f(x)f(X)f(x_{1})+f(x_{2})=f(x)\in f(X), so it is closed under addition. The same logic can be used to show that it is closed under multiplication. Therefore, f(X)f(X) is a subspace of YY.