Talk:Conway group

Latest comment: 6 years ago by Scott Tillinghast, Houston TX in topic Centralizer of an S3

Versions of Robert A. Wilson website

edit

I re-instated version 2 alongside version 3. The nice thing about version 2 is that lists of maximal subgroups have links to pages on the simple groups that are involved. Scott Tillinghast, Houston TX (talk) 21:45, 4 May 2008 (UTC)Reply

Summary of sources, aids to editors

edit

It has been requested of me to summarize the sources of where I get the material, as an aid to others who want to edit. I would welcome suggestions about what else I should include.


Opening of article

edit

I added in some history, which I mainly got from Thomas Thompson's Carus monograph. He talks about the groups he has space to give adequate mathematical treatment. That includes the 3 Conway groups and some basics about M24, but he barely mentions the other 4 soradic groups of the second generation.

Other sporadic groups

edit

The best source is Robert Griess's Twelve Sporadic Groups. Quite a bit of information, but there are many interesting questions he does not touch. He gives very little history on these groups.

For history I have resorted to the book edited by Brauer and Shih, a compilation of lectures from May 1968 conference at Harvard.

I am trying to enumerate all the conjugacy classes of simple subgroups and subquotients in Co0, but this may constitute original research.

The group 212:M24

edit

From Thomas Thompson. He relates how it began Conway's investigation, how he managed to prove this is a proper but maximal subgroup of Co0. He could then deduce |Co0|.

Maximal subgroups of various subgroups

edit

Mainly from Robert A. Wilson's Atlas website. Perhaps it would be better to list the maximal subgroups of Co0, but I don't know of a listing. The interesting thing about them is that some of them are direct products with maximal subgroups of Co1, and others are stem extensions. It is Co0 that is represented in SL(Q,24), as well as Co2 and Co3, but not Co1.

Scott Tillinghast, Houston TX (talk) 04:05, 9 June 2008 (UTC)Reply

I just checked out a new book by Robert A. Wilson: 'The Finite Simple Groups.' He lists all the maximal subgroups of the 3 Conway groups and some maximal subgroups of Co0. One topic of interest is a chain of subgroups involving the Suzuki group at one end and the alternating group A9 at the other. I am drafting a section about this chain. Scott Tillinghast, Houston TX (talk) 23:12, 1 April 2010 (UTC)Reply

Possible split

edit

Scott Tillinghast, Houston TX (talk), at WT:WPM#Separate page, suggested separate articles on the McLaughlin and Suzuki sporadic simple groups. Suzuki sporadic group and McLaughlin group (mathematics) already exist as redirects, but on his behalf I've put some split tags up for discussion. -- Radagast3 (talk) 14:22, 16 April 2010 (UTC)Reply

Personally, I see no problem with turning the redirects into articles, copying text from here to begin with. That would need to be noted in the initial edit summary, of course. -- Radagast3 (talk) 08:09, 19 April 2010 (UTC)Reply

I see that the new articles have been created. I shall be finding things to add on the McLaughlin and sporadic Suzuki groups. I would also like to see an article on the infinite family called the Suzuki groups. They are of interest as the only simple groups of composite order not divisible by 3. They have little to do with the sporadic Suzuki group. Scott Tillinghast, Houston TX (talk) 07:59, 7 November 2010 (UTC)Reply

Rabbit trails

edit

Figure of speech for something seemingly unrelated. What I was saying was that the Conway's work unified some previously known sporadic groups that seemed unrelated. Namely Higman-Sims, Hall-Janko, Suzuki, and McLaughlin. Scott Tillinghast, Houston TX (talk) 23:23, 4 November 2010 (UTC)Reply

Compatibility

edit

The only matrices specified in this article are those of type ζ, introduced as non-monomial matrices. They are built from a 4x4 matrix η and its negative.

The standard subgroup of type 2.A9 in Co0 includes the non-monomial matrix η, -η, η, -η, η, -η. The diagonal dodecads of 24 can be identified with some of the 70 4-subsets of a set of 8. PSL(2,7) (degree 8) splits those 4-subsets into 2 orbits of 14 and an orbit of 42. All 3 orbits include complements. The 2 orbits of 14 both generate groups of order 16 under symmetric difference. The orbit of 42 does not generate a good group. What representation of 24:PSL(2,7) will be compatible, in the sense of generating a subgroup of Co0?

The ζ matrix and 24:PSL(2,7) generate a block sum of 3 identical 8x8 matrix groups. I do not know whether this is 2.A8 or the full group 2.A9. Scott Tillinghast, Houston TX (talk) 23:10, 23 December 2015 (UTC)Reply

The section on the monomial subgroup really ought to define precisely a standard representation of the binary Golay code, thus one of M24. What would be an easiest way to do this? Scott Tillinghast, Houston TX (talk) 21:30, 24 December 2015 (UTC)Reply

The group PSL (2,7) can be represented as the linear fractional group on F7 ∪ {∞}, generated by permutations (0123456) and (0∞)(16)(23)(45). Griess (p. 59) uses the labeling:

∞ 0 |∞ 0 |∞ 0
3 2 |3 2 |3 2
5 1 |5 1 |5 1
6 4 |6 4 |6 4

One dodecad can be represented either by the subset {0,1,3,∞} or by its negative {0,4,6,∞}. Scott Tillinghast, Houston TX (talk) 02:13, 26 December 2015 (UTC)Reply

I have defined a convenient representation in the article on the binary Golay code and made a link. Scott Tillinghast, Houston TX (talk) 17:15, 27 December 2015 (UTC)Reply

Involutions

edit

Perhaps a proof should be included that all involutions of the same trace in Co0 are conjugate. A useful tool: if the product of 2 involutions has odd order n, these involutions are conjugate in a dihedral subgropu of order 2n. Scott Tillinghast, Houston TX (talk) 19:27, 18 January 2016 (UTC)Reply

The matrix η =

 

has with the diagonal matrix {-1,1,1,1} a product of order 3. Hence all matrices of type ζ (see article) are conjugate to diagonal matrices. Scott Tillinghast, Houston TX (talk) 20:18, 18 January 2016 (UTC)Reply

A permutation matrix representing a product of transpositions is easily diagonalized in SL(24,R) but this requires sqrt 2. Diagonalizing in Co0 is more complicated. Scott Tillinghast, Houston TX (talk) 03:18, 21 January 2016 (UTC)Reply

Centralizer of an S3

edit

Co0 has a subgroup isomorphic to S3 permuting 3 subspaces of the Leech lattice, each of dimension 8. That led to the discovery of the Suzuki chain. Let Z be the center of Co0 (order 2) and C0 be the centralizer of the S3. Griess proves that C0 is the double cover of A9 but is not clear about details. He does not prove that C0/Z is simple but perhaps this is similar to proving Co1 is simple. Let t be a diagonal involution in C0. It is not so difficult to show that the centralizer of {±t} in C0/Z has form 23:S4. A survey of simple groups shows only 2 possibilities with an involution with such a centralizer: A8 and A9. Griess is obscure about why a Sylow 3-subgroup of C0 is larger than 32. Because C0 contains a subgroup isomorphic to SL(2,7) it is 2.A9, not the direct product 2xA9. Scott Tillinghast, Houston TX (talk) 15:36, 23 January 2018 (UTC)Reply