11 Comments
User's avatar
Wyrd Smythe's avatar

I have a question about the condition (a⁻¹x ∈ H). I tried that with the exercise 2 subgroup H={e, s}, and found that multiplying them one way always gave me s (which satisfies s ∈ H), but in some cases the multiplication didn't commute, and I got, in all four exceptions, sr² (which is not in H).

Not that my answers for exercise 2 are necessarily correct, but is this result because the subgroup isn't normal, or is it sufficient that there exists an ab ∈ H?

FWIW, I've decided to start a "homework and notes" page on my blog here. I may continue to also post answers in a restack, but I may just start adding to that page. (I find the difference between Notes text and post/page text downright weird here.)

https://logosconcarne.substack.com/p/group-theory-notes-and-homework

Expand full comment
David's avatar

It isn’t clear to me what “multiplying THEM” is referring to in your second sentence; could you expound a bit more?

Expand full comment
Wyrd Smythe's avatar

OTOH, multiplying any two elements of the two cosets of H={e, r, r², r³} always results in x ∈ H.

Expand full comment
David's avatar

I guess I’m also unsure what is meant by “or is it sufficient that ab in H”? Sufficient for what?

Expand full comment
Wyrd Smythe's avatar

In the case of rH, which has {r, sr³} as members, there is sr³⋅r=s, and s is in H, but r⋅sr³=sr², and sr² is not in H. So, my question was whether the first case was sufficient. As mentioned in my other comment, with the two cosets of H={e, r, r², r³}, a operation between any two members results in x ∈ H.

Expand full comment
David's avatar

I think what you’re seeing is that the cosets rH and Hr are not the same; we require that the cosets line up exactly to be considered the same. The issue is that, while both the left and right cosets of G partition G, unless G is normal, the partitions do not have to be the same.

Expand full comment
Wyrd Smythe's avatar

They're definitely not the same. rH = {r, sr³} and Hr = {r, sr}. In both of them, ab≠ba, when multiplying the two members of the respective sets.

And r³H = {r³, sr} and Hr³ = {r³, sr³} also don't equal each other, and they are where the other case of ab≠ba is. In all four cases, the answer is sr² rather than s.

So, yeah, it seems when aH≠Ha, then it is also not the case the multiplying any two, in any order, results in x ∈ H.

Expand full comment
Wyrd Smythe's avatar

Sure, here's my actual math:

H: e⋅s=s; s⋅e=s.

rH: r⋅sr³=sr²; sr³⋅r=s.

r²H: r²⋅sr²=s; sr²⋅r²=s.

r³H: r³⋅sr=sr²; sr⋅r3=s.

Hr: r⋅sr=s; sr⋅r=sr².

Hr²: r²⋅sr²=s; sr²⋅r²=s.

Hr³: r³⋅sr³=s; sr³⋅r³=sr².

In most cases, the result is s, but in four cases, it's sr².

Expand full comment
Wyrd Smythe's avatar

Ha! So, in my answer to #3 for last time, I stumbled on the three-member cosets of D₆ without knowing it.

FWIW, the problem with x⁻¹ seems to be font related. Substack uses the "Spectral" font in the post's text, and it apparently doesn't have a good implementation of the Unicode "superscript minus" character (U+207B). Post comments use a different font entirely, so it works fine here.

Expand full comment
David's avatar

Interesting; I’ll mess around with the fonts on the backend and see if that helps

Expand full comment
Wyrd Smythe's avatar

I've got a full weekend so won't get to the lesson until Monday but taking a quick glance I laughed when you mentioned "arborescence". Synchronicity strikes again. I just ran into that recently in the comments on a video about an interesting maze generation algorithm.

Have a good weekend!

Expand full comment