PlanetMath (more info)
 Math for the people, by the people.
Encyclopedia | Requests | Forums | Docs | Wiki | Random | RSS  
Login
create new user
name:
pass:
forget your password?
Main Menu
Owner confidence rating: High Entry average rating: Very high
spectral permanence theorem (Theorem)

Let $ \mathcal{A}$ be a unital complex Banach algebra and $ \mathcal{B} \subseteq \mathcal{A}$ a Banach subalgebra that contains the identity of $ \mathcal{A}$.

For every element $ x \in \mathcal{B}$ it makes sense to speak of the spectrum $ \sigma_{\mathcal{B}}(x)$ of $ x$ relative to $ \mathcal{B}$ as well as the spectrum $ \sigma_{\mathcal{A}}(x)$ of $ x$ relative to $ \mathcal{A}$.

We provide here three results of increasing sophistication which relate both these spectrums, $ \sigma_{\mathcal{B}}(x)$ and $ \sigma_{\mathcal{A}}(x)$. Any of the last two is usually refered to as the spectral permanence theorem.

Proposition - Let $ \mathcal{B} \subseteq \mathcal{A}$ be as above. For every element $ x \in \mathcal{B}$ we have

$\displaystyle \sigma_{\mathcal{A}}(x) \subseteq \sigma_{\mathcal{B}}(x). $

This first result is purely algebraic. It is a straightforward consequence of the fact that invertible elements in $ \mathcal{B}$ are also invertible in $ \mathcal{A}$.

The other inclusion, $ \sigma_{\mathcal{B}}(x) \subseteq \sigma_{\mathcal{A}}(x)$, is not necessarily true. It is true, however, if one considers the boundary $ \partial \sigma_{\mathcal{B}}(x)$ instead.

Theorem - Let $ \mathcal{B} \subseteq \mathcal{A}$ be as above. For every element $ x \in \mathcal{B}$ we have

$\displaystyle \partial \sigma_{\mathcal{B}}(x) \subseteq \sigma_{\mathcal{A}}(x). $

Since the spectrum is a non-empty compact set in $ \mathbb{C}$, one can decompose $ \mathbb{C} - \sigma_{\mathcal{A}}(x)$ into its connected components, obtaining an unbounded component $ \Omega_{\infty}$ together with a sequence of bounded components $ \Omega_1, \Omega_2, \dots$,

$\displaystyle \mathbb{C}-\sigma_{\mathcal{A}}(x) = \Omega_{\infty} \cup \Omega_{1} \cup \Omega_{2} \cup \cdots $
Of course there may be only a finite number of bounded components or even none.

Theorem - Let $ x \in \mathcal{B} \subseteq \mathcal{A}$ be as above. Then $ \sigma_{\mathcal{B}}(x)$ is obtained from $ \sigma_{\mathcal{A}}(x)$ by adjoining to it some (possibly none) bounded components of $ \mathbb{C}-\sigma_{\mathcal{A}}(x)$.

As an example, if $ \sigma_{\mathcal{A}}(x)$ is the unit circle, then $ \sigma_{\mathcal{B}}(x)$ can only possibly be the unit circle or the closed unit disk.



Anyone with an account can edit this entry. Please help improve it!

"spectral permanence theorem" is owned by asteroid.
(view preamble)

View style:


Attachments:
spectral invariance theorem (for $C^*$-algebras) (Theorem) by asteroid
Log in to rate this entry.
(view current ratings)

Cross-references: closed unit disk, unit circle, number, finite, bounded, sequence, component, unbounded, connected components, spectrum is a non-empty compact set, boundary, inclusion, invertible, consequence, increasing, spectrum, identity, contains, subalgebra, Banach algebra, complex, unital
There is 1 reference to this entry.

This is version 2 of spectral permanence theorem, born on 2007-08-23, modified 2007-08-23.
Object id is 9885, canonical name is SpectralPermanenceTheorem.
Accessed 404 times total.

Classification:
AMS MSC46H05 (Functional analysis :: Topological algebras, normed rings and algebras, Banach algebras :: General theory of topological algebras)
 46H10 (Functional analysis :: Topological algebras, normed rings and algebras, Banach algebras :: Ideals and subalgebras)

Pending Errata and Addenda
None.
Discussion
Style: Expand: Order:
forum policy

No messages.

Interact
post | correct | update request | prove | add result | add corollary | add example | add (any)