- #1
Aerostd
- 18
- 0
Homework Statement
There is a theorem in Folland's book (Real Analysis) in which he gives three statements that are equivalent. In the proof he decides to take the first statement and prove the other two, but in the end he derives a statement that looks different than what is given in the theorem. This means that the statement he derived is the same as the statement in the proof, but I am not able to go from one to the other.
Theorem: Let [tex]E \in R[/tex](reals). Then the following are equivalent.
(a) [tex]E[/tex] is measurable.
(b) [tex]E = V[/tex] [tex]- N_{1} [/tex] where [tex] V [/tex] is a [tex]G_{\delta}[/tex] set and [tex]\mu(N_{1})=0[/tex].
There is a (c) but let us ignore that for now. The measure is the Lebesgue measure and the domain I think is the extended Borel sigma algebra, so both [tex]V[/tex] and [tex] N_{1} [/tex] are measurable.
In the book, the author assumes (a) and then shows that:
(d) there exists a [tex]G_{\delta}[/tex] set [tex]V \subset E[/tex] such that [tex]\mu(V-E)=0[/tex].
I am trying to either go from this statement to (b) or from (b) to (d), but I don't know how.
2. The attempt at a solutionI don't know how to go from (d) to (b) because I don't know how one can conclude that sets are equal based on their measures. So I tried to go from (b) to (d). I have to show that [tex] E \in V [/tex] and [tex]\mu(E-V)[/tex] is zero.
If [tex] E = V-N_{1}[/tex], then
[tex] \mu(E) = \mu(V) - \mu(V \cap N_{1})[/tex],
[tex] \mu(E) = \mu(V)[/tex].
I don't know what to do next.