Is there a measure for every Borel σ-algebra?

  • Context: Graduate 
  • Thread starter Thread starter Fredrik
  • Start date Start date
  • Tags Tags
    Measure
Join the discussion
Registration is free. Ask a follow-up in this thread, or start your own.
4 replies · 3K views
Staff Emeritus
Science Advisor
Homework Helper
Insights Author
Gold Member
Messages
10,876
Reaction score
423
The Borel σ-algebra can be defined on every topological space. Does that mean that every topological space can be turned into a measure space?
 
Physics news on Phys.org
Yes, in a trivial way. Just define [itex]\mu(S)=0[/itex] for all S. Or take the counting measure. Or a point measure.

A better answer is actually Riesz representation theorem. This says that all regular measures on a locally compact Hausdorff space coincide with positive functionals

[tex]T:C_c(X)\rightarrow \mathbb{R}[/tex]

(with [itex]C_c(X)[/itex] the functions to [itex]\mathbb{R}[/itex] with compact support). Such a functional exist because of the Hahn-Banach theorem.

The same thing can be done with [itex]C_0(X)[/itex] actually. In that case: the measures are the positive functionals, the probability measures are the states and the point measures represent the pure states on X.

In general, the existence of non-trivial measures is not a simple problem and requires set theory (see for example measurable cardinals)
 
Last edited:
I wonder if the space has a cardinality greater than the continuum whether there are any measures other than the trivial one.
 
lavinia said:
I wonder if the space has a cardinality greater than the continuum whether there are any measures other than the trivial one.

The first thing to do is to define the word "the trivial one".

This is quite an interesting theory. In general, this is studied in Boolean algebras and the so-called measure algebras.
 
Excellent answer as always. Thanks. I've been meaning to study the Riesz representation theorem. It will be the first thing I do when I'm done with the basics of integration theory.
 
Last edited: