Lemma 2
Tag 0BOQ, permanent. Cite it as [Bourbaki, Tag 0BOQ].
General Topology, chapter III, § 7. Inverse limits of topological groups and rings
Let be a Hausdorff topological group and let be a compact normal subgroup of , and the canonical mapping . If is the identity component of , then is the identity component of .
Once this lemma is established, we shall have for each , and since is a closed subgroup of (§ 2, no. 2, Proposition 7) it suffices to apply Proposition 3 a).
To prove the lemma, notice first that if is the component of the identity element in , then since is connected. Suppose that . Since is a closed normal subgroup of (§ 2, no. 2, Proposition 7), is a normal subgroup of ; if is the canonical mapping , then is connected and does not consist of the identity element alone; hence the identity component of does not consist of the identity element alone. But is isomorphic to , hence to , and consequently also to (§ 2, no. 7, Corollary to Proposition 22, and Proposition 20). Now, is Hausdorff and totally disconnected (Chapter I, § 11, no. 5, Proposition 9), and , being the canonical image of the compact normal subgroup of , is a compact subgroup of . It is therefore sufficient to prove Lemma 2 under the additional hypothesis that is totally disconnected, i.e. .
Suppose then that ; replacing by its subgroup , which is totally disconnected and contains , we may suppose that is connected and does not consist of a single point.
Let be the set of closed subgroups of such that . We shall show that the set , ordered by the relation , is inductive. Indeed, if is a linearly ordered subset of , then for each the set of sets with is a filter base formed of closed sets in the compact space ; hence the intersection of these sets is not empty, which shows that the intersection of the subgroups belongs to . Applying Zorn's lemma, we see therefore that has a minimal element . Since is compact, is isomorphic to (§ 4, no. 1, Proposition 1, Corollary 3); since is totally disconnected and is compact, we see that we may replace by ; in other words, we may suppose in addition that there is no closed subgroup such that .
Now let be the intersection of the neighbourhoods of the identity element which are both open and closed in , and let us show that is a closed subgroup of . Clearly is closed in ; hence it is enough to show that . But if and if is an open and closed neighbourhood of in , then so is ; for otherwise would belong to the complement of in , which is again open and closed, and we should have ; hence , contrary to hypothesis. It follows that , the intersection of the sets as runs through the open and closed neighbourhoods of in , contains ; in other words, , which proves our assertion. Since is totally disconnected and , we have . But if is an open and closed neighbourhood of in , then is also open and closed in (§ 4, no. 1, Corollary 1 to Proposition 1), hence is both open and closed in , and this implies that by virtue of the hypothesis. We shall conclude from this that , which will give us the desired contradiction and complete the proof of the lemma. Indeed, for each , meets every neighbourhood of which is both open and closed, hence also meets the intersection of these neighbourhoods, since the sets form a filter base consisting of closed sets in the compact space .
Q.E.D.
Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.
Cited by
Nothing.