I am currently studying Harmonic Analysis and didn't quite understand part of the proofs for the structure theorems of locally-compact abelia groups (LCA).

Let $A$ be an LCA group such that $\hat{A}/\hat{A}_0$ is compact, where $A_0$ denotes the connected component of the unit in $A$ and $\hat{A}$ is the dualization of $A$.

Now suppose that there is an infinite compact subgroup $E \leq A$. We can write this as an exact sequence

$$ 0 \to E \to A\to A/E\to 0$$

and dualizing it to the exact sequenz $$ 0 \to B \to \hat{A} \to \hat{E}\to 0 .$$ Where $\hat{E}$ is an infinite discrete group and as such non-compact. This is supposedly a contradiction to $\hat{A}/\hat{A}_0$ being compact, but I don't see why. By that, we would know that no such group can exist in $A$.

The map $\Phi$ from $\hat{A} \to \hat{E}$ is surjective and continous by construction, and we can restrict it to a map $$\varphi: \hat{A}/\hat{A}_0 \to \hat{E}/(\hat{A}_0\cap\hat{E}).$$ If we knew that $\hat{A}_0\cap\hat{E}$ is finite (or even better $=\{e\}$) then we would be done, having a surjection from a compact to a non compact group.

Is this the right way to go or am I missing something important? Any help would be greatly appreciated.


If by contradiction $A$ has an infinite compact subgroup $E$, then, as you said, $\hat{A}$ has the infinite discrete quotient $\hat{E}$. Since $\hat{A}_0$ has a trivial image in $\hat{E}$, this is an infinite discrete quotient of $\hat{A}/\hat{A}_0$. But the latter is compact, so its discrete image should be finite. Contradiction.


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.