Blog Archive

Sunday, April 7, 2024

A Categorical Perspective on the Riesz Representation Theorem and Adjoint Maps

This blog is a rehash of the Riesz representation theorem in linear algebra, from a categorical point of view.

You will see why the adjoint map T∗ is defined by ⟨Tv,w⟩=⟨v,T∗w⟩.

To begin with, let us define a functor from the category of finite-dimensional inner product spaces to its dual category.

i.e.,

G:FinF→FinFop

For each object, G:V⟼V. For each morphism T:V→W, G:T⟼T∗, where T∗:W→V.

Another functor with which we are really familiar is the dual space functor HomFinF(−,F), denoted as F.

Riesz's representation theorem claims that there exists a natural isomorphism η:G→F such that the following diagram commutes:

W→ηWW∗T∗↓↓T′V→ηVV∗

Indeed, the adjoint map T∗ is defined by this diagram. T∗ is the unique map that makes this diagram commute.

Firstly, let us prove the Riesz representation theorem.

Riesz representation theorem. Suppose V∈Ob(FinF) and φ is a linear functional on V. Then there exists a unique vector v∈V such that

φ(u)=⟨u,v⟩

for every u∈V.

Proof. First, we prove that there exists a v∈V such that φ(u)=⟨u,v⟩.

Let e1,e2,…,en be the standard orthogonal basis of V. Then

φ(u)=φ(∑i=1n⟨u,ei⟩ei)=∑i=1n⟨u,ei⟩φ(ei)=⟨u,∑i=1nφ(ei)―ei⟩

Hence,

v=∑i=1nφ(ei)―ei

To see v is unique, consider

∀u∈V,⟨u,v1⟩=⟨u,v2⟩⟹∀u∈V,⟨u,v1−v2⟩=0⟹v1−v2=0⟹v1=v2

Hence, the Riesz representation theorem gives us an isomorphism ηV:V→V∗ for each V. The isomorphism is natural.

To make the diagram commute,

W→ηWW∗T∗↓↓T′V→ηVV∗

We have to define T∗:=ηV−1∘T′∘ηW.

From the diagram, we can see that ηV∘T∗=T′∘ηW.

Putting a vector w∈W on both sides, we can see that

ηV∘T∗(w)=T′∘ηW(w)⟺⟨−,T∗w⟩=T′(ηW(w))=ηW(w)∘T=⟨T(−),w⟩

Hence, we get the usual definition of the adjoint map.

⟨−,T∗w⟩=⟨T(−),w⟩

Readers might feel more familiar with this from

⟨v,T∗w⟩=⟨Tv,w⟩

No comments:

Post a Comment

Popular Posts