I'm looking for subset $A$ of $\mathbb R$ such that $A$ is a Borel set but $A$ is neither $F_\sigma$ nor $G_\delta$.


There are many examples. Here's one:

Observe first that the rational numbers $\mathbb{Q}$ are an $F_{\sigma}$. This is because they are a countable union of points. The irrational numbers $\mathbb{R} \smallsetminus \mathbb{Q} = \bigcap_{q \in \mathbb{Q}} \mathbb{R} \smallsetminus \{q\}$ are thus a $G_{\delta}$. Since both $\mathbb{Q}$ and $\mathbb{R} \smallsetminus \mathbb{Q}$ are dense and disjoint it follows from the Baire category theorem that $\mathbb{Q}$ cannot be a $G_{\delta}$. [Edit: See also this thread here containing several proofs that $\mathbb{Q}$ can't be a $G_{\delta}$ in $\mathbb{R}$. These proofs explicitly avoid Baire].

The same reasoning shows that $F = \mathbb{Q}_{\geq 0}$ is an $F_{\sigma}$ in $[0,\infty)$, but isn't a $G_{\delta}$, and that $G= \mathbb{R}_{\leq 0} \smallsetminus \mathbb{Q}_{\leq 0}$ is a $G_{\delta}$ in $(-\infty,0]$, but isn't an $F_{\sigma}$. Their union $F \cup G$ is then an example of a Borel subset of $\mathbb{R}$ which is neither an $F_{\sigma}$ nor a $G_{\delta}$ because if it were an $F_{\sigma}$ then the same would hold for $G = (F \cup G) \cap (-\infty,0)$, for example. I leave it as an exercise to show that $F \cup G$ is both an $F_{\sigma\delta}$ and a $G_{\delta\sigma}$.

That's probably the easiest example. A few more (both more interesting but also more involved ones) can be found in this MO thread.

For a much more in-depth discussion of such ideas, I recommend looking into one of the following books:

Specifically, look up the sections on the Borel hierarchy.

  • 3
    $\begingroup$ Very nice example! I had in mind some countable union of copies of the irrationals or something like that. $\endgroup$ – Asaf Karagila Oct 18 '11 at 16:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.