Conversation
alreadydone
left a comment
There was a problem hiding this comment.
LGTM but IMO the result could be split into two main facts:
- if T is a nonempty, proper subset of a G-orbit O, then there exists a generator of G that sends an element of T to an element of O\T.
- if a subgroup contains all transpositions in T and an arbitrary transposition (ab) for some
a ∈ T, then it contains all transpositions inT ∪ {b}. It might be easier to prove and use that the permutations onT ∪ {b}is generated by all permutations onTtogether with an arbitrary transposition of the form (ab) witha ∈ T.
|
Thanks! I tried extracting the lemmas (47e8fbe). I was able to extract the second fact, but the statement ended up looking awfully clunky and specialized in my eyes, so I don't know if there's much point keeping it as a separate lemma? The first fact (after adding a symmetry assumption on the generating set) seems tricky to extract since "sending an element of T to an element of O\T" is not quite the same thing as "swapping an element of T with an element of O\T", and going between these would require a bit of case bashing (you would assume sigma = swap x y takes a in T to b not in T, and then show that {x,y}={a,b}). |
|
I'm going to suggest a golf and a helper lemma later tonight. |
|
I think that the following lemma is a nice generalization of your statement: open Subgroup MulAction in
theorem mem_closure_iff_mem_orbit_of_isSwap [Finite α] {S : Set (Perm α)} {σ : Perm α}
(hS : ∀ σ ∈ S, σ.IsSwap) : σ ∈ closure S ↔ ∀ x, σ x ∈ orbit (closure S) x := by
refine ⟨fun hσ x ↦ mem_orbit_iff.2 ⟨⟨σ, hσ⟩, rfl⟩, fun hσ ↦ ?_⟩(I only prove the easy implication here). Unfortunately, I don't see a short proof yet. I'll think about it tomorrow. |
Thanks for trying!
I think it's not too clunky and is worthwhile to extract. Using
The symmetry assumption is unnecessary if T or O\T is finite. And yes we seem to be missing a lemma stating |
Sorry, I don't know what I was thinking. The correct statement should be: if the fixingSubgroup of T' is a subset of H, and H includes a transposition (ab) with a not in T', then the fixingSubgroup of T'{b} is a subset of H. |
|
I managed to get a sorry-free proof of @urkud's generalization. I haven't cleaned it up yet, but before I do, I just wanted to check whether either you had ideas for streamlining/shortcutting the proof at all? (for instance, is there an easier way to get something like |
|
Here is a proof that doesn't use List and also generalizes to action on an infinite type: |
|
This looks wonderful, thanks so much both of you! |
|
I left some lemmas involving |
|
@urkud Does this look good to you now? If so it would be natural that you merge it; I wrote most of the final version so I'm not sure whether I'm supposed to call maintainer mergee. |
|
maintainer merge |
|
🚀 Pull request has been placed on the maintainer queue by alreadydone. |
|
Thanks! bors merge |
…nerated by transpositions must be the whole symmetric group (#10035) This PR proves that a transitive permutation group generated by transpositions must be the whole symmetric group. Co-authored-by: Thomas Browning <tb65536@users.noreply.github.com>
|
Build failed (retrying...): |
|
bors r- |
|
Canceled. |
|
There is an error related to the new style of bors d+ |
|
✌️ tb65536 can now approve this pull request. To approve and merge a pull request, simply reply with |
|
It would be nice if you could add the result that the subgroup generated by all swaps is the subgroup of permutations with finite support. |
|
bors r+ |
…nerated by transpositions must be the whole symmetric group (#10035) This PR proves that a transitive permutation group generated by transpositions must be the whole symmetric group. Co-authored-by: Thomas Browning <tb65536@users.noreply.github.com>
|
Pull request successfully merged into master. Build succeeded: |
…nerated by transpositions must be the whole symmetric group (#10035) This PR proves that a transitive permutation group generated by transpositions must be the whole symmetric group. Co-authored-by: Thomas Browning <tb65536@users.noreply.github.com>
…nerated by transpositions must be the whole symmetric group (#10035) This PR proves that a transitive permutation group generated by transpositions must be the whole symmetric group. Co-authored-by: Thomas Browning <tb65536@users.noreply.github.com>
…nerated by transpositions must be the whole symmetric group (#10035) This PR proves that a transitive permutation group generated by transpositions must be the whole symmetric group. Co-authored-by: Thomas Browning <tb65536@users.noreply.github.com>
This PR proves that a transitive permutation group generated by transpositions must be the whole symmetric group.