MathLabs
TheoremProved

Hahn–Banach extension theorem

Statement

Let XX be a real normed vector space, YY a linear subspace of XX, and φ\varphi a bounded linear functional on YY with ∣φ(y)∣≤M∥y∥|\varphi(y)|\le M\|y\| for all y∈Yy\in Y. Then there exists a bounded linear functional Φ\Phi on all of XX such that Φ(y)=φ(y)\Phi(y)=\varphi(y) for every y∈Yy\in Y, and ∣Φ(x)∣≤M∥x∥|\Phi(x)|\le M\|x\| for every x∈Xx\in X.

Why is it true?

It guarantees that a linear functional defined only on a small subspace — for instance, only knowing "the value of a signal at a few sample points" — can always be extended to the whole space without growing its bound; nothing ever forces us to be stuck working on a subspace.

Proof sketch

First extend by one dimension. Pick x0∉Yx_0\notin Y and let Y1=Y⊕Rx0Y_1=Y\oplus\mathbb{R}x_0. We must choose the value c=Φ(x0)c=\Phi(x_0) so that ∣φ(y)+tc∣≤M∥y+tx0∥|\varphi(y)+tc|\le M\|y+tx_0\| for all y∈Y,t∈Ry\in Y,t\in\mathbb{R}; dividing by t≠0t\ne0 and substituting y/t→yy/t\to y, this reduces to needing cc with φ(y)−M∥y−x0∥≤c≤M∥y+x0∥−φ(y)\varphi(y)-M\|y-x_0\|\le c\le M\|y+x_0\|-\varphi(y) for all y∈Yy\in Y. Using φ(y1)−φ(y2)=φ(y1−y2)≤M∥y1−y2∥≤M∥y1+x0∥+M∥y2−x0∥\varphi(y_1)-\varphi(y_2)=\varphi(y_1-y_2)\le M\|y_1-y_2\|\le M\|y_1+x_0\|+M\|y_2-x_0\|, one checks the supremum of the left expressions over y1y_1 never exceeds the infimum of the right expressions over y2y_2, so a valid cc in that interval exists; this defines Φ\Phi on Y1Y_1 with the same bound MM.

Next, extend to all of XX using Zorn's Lemma. Consider the set of all pairs (Z,Ψ)(Z,\Psi) where ZZ is a subspace with Y⊆Z⊆XY\subseteq Z\subseteq X and Ψ\Psi extends φ\varphi to ZZ with bound MM, partially ordered by (Z1,Ψ1)≤(Z2,Ψ2)(Z_1,\Psi_1)\le(Z_2,\Psi_2) iff Z1⊆Z2Z_1\subseteq Z_2 and Ψ2∣Z1=Ψ1\Psi_2|_{Z_1}=\Psi_1. Every chain has an upper bound (take the union of the subspaces and the functional agreeing with each on its domain), so Zorn's Lemma gives a maximal element (Z∗,Φ)(Z^*,\Phi).

Finally, if Z∗≠XZ^*\ne X, the one-dimension extension step above applied to Z∗Z^* and any x0∈X∖Z∗x_0\in X\setminus Z^* would produce a strictly larger admissible pair, contradicting maximality of (Z∗,Φ)(Z^*,\Phi). Hence Z∗=XZ^*=X, and Φ\Phi is the desired bound-preserving extension to all of XX.

Topics that use this theorem

Step-by-step proofs

No step-by-step proof yet for this theorem.

References

  1. Walter Rudin (1991). Functional Analysis
  2. John B. Conway (2007). A Course in Functional Analysis
  3. Assaf Naor (2012). An introduction to the Ribe program · arXiv:1205.5993