Scribd
Upload
48 views

Infinite Geometric Groups of Rank 4: Peter J

The document summarizes Peter J. Cameron's proof that there is no infinite geometric group of rank 4 and finite type. A geometric group is a permutation group where stabilizers of point sequences act transitively on non-fixed points. Cameron provides an elementary proof by contradiction, showing that the stabilizer of a basis in such a group would have to be non-trivial. This disproves the possibility of an infinite group with this structure.

Uploaded by

Afghanistan
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
48 views

Infinite Geometric Groups of Rank 4: Peter J

The document summarizes Peter J. Cameron's proof that there is no infinite geometric group of rank 4 and finite type. A geometric group is a permutation group where stabilizers of point sequences act transitively on non-fixed points. Cameron provides an elementary proof by contradiction, showing that the stabilizer of a basis in such a group would have to be non-trivial. This disproves the possibility of an infinite group with this structure.

Uploaded by

Afghanistan
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Europ. J.

Combinatorics (1992) 13, 87-88

Infinite Geometric Groups of Rank 4

PETER J. CAMERON

An elementary proof is given of the nonexistence of infinite geometric groups of finite type
and rank 4.

A permutation group G is called geometric if the pointwise stabiliser of any sequence


of points acts transitively on the set of points it doesn’t fix (if there are any). If there is
a natural number r for which the stabilier of some r points is the identity, the least such
r is called the rank of G. The fYPe of a geometric group of rank I is the r-tuple
(lo. . , L,), w h ere li is the (constant) number of fixed points of the stabiliser of i
independent points. (A tuple of points is independent if no point is fixed by the
stabihser of its predecessors. An easy induction from the definition shows that a
geometric group acts transitively on independent i-tuples for every i 6 r). The group is
said to have finite type of 1,-i is finite. For example, a group is sharply r-transitive if
and only if it is geometric of rank r and type (0, . . , r - 1). Tits [3] showed that there
is no sharply 4-transitive infinite permutation group. In this note, Tits’ result is
generalised; I show that there is no infinite geometric group of rank 4 and finite type.
Of course, the non-existence for larger finite rank follows immediately by induction.
The existence of sharply 3-transitive groups (e.g. PGL(2, k), k an infinite field), and of
geometric groups of infinite rank or of finite rank but not of finite type (e.g. affine
groups) shows that the result is best possible.
In 1988, I announced a proof of this theorem. The proof then depended on the
determination of all finite geometric groups of rank at least 2 by Tracey Maund [2] in
her Oxford D. Phil. thesis; her arguments relied heavily on the classification of finite
simple groups. By contrast, the proof given here is completely elementary.

THEOREM . There i.s no infinite geometric group of rank 4 and finite type.

PROOF. We begin with a small amount of theory of geometric groups, specialised to


our situation: see Cameron, Deza and Singhi [l]. Points fixed by G may be deleted
without loss; so we may assume that the type is (0, k, 1, m) with 0 <k <I < m.
Corollary 2 of Cameron et al. asserts that k, 1 - k and m - 1 are all odd. There is a
matroid of rank 4 on the point set, whose flats of rank 1, 2, 3 have cardinality k, l, m
respectively. (I call these flats Points, Lines and Planes respectively.) The group G acts
sharply transitively on (ordered) bases of the matroid, and the fixed point set of any
element or subgroup of G is a flat.
It follows from these remarks that, if an involution t fixes a Plane II setwise but not
pointwise, then the fixed points of t in II form a Point.
Now let (x,, x2, x3, x4) be a basis, and t the (unique) element of G mapping this
basis to (xi, x2, x4, x3). Then t* fixes a basis, so is the identity; that is, t is an
involution. Let F be the fixed point set of t, and C = {x3, x4}. Then F U C contains a
basis, so its pointwise stabiliser is trivial. We reach a contradiction by showing that this
stabiliser is non-trivial.
87
0195~6698/92/020087 + 02 %02.00/O 0 1992 Academic Press Limited
There are only finitely many cycles C’ of t for which C U C’ is dependent. For. it
C U C’ is dependent, then either
(a) rank(C U C’) = 2, whence C’ c (C); or
(b) rank(C U C’) = 3, whence (C U C’) = II is a Plane. In this case, t fixes I7 setwise
but not pointwise, so it fixes a point x E II; then C’ c n = (C U {x}), with x E F. In
either case, the number of possibilities is finite.
Let ‘%be the set of cycles C’ of c for which C U C’ is a basis. Let H be the stabiliser
of C in the centraliser of t. Then H acts transitively on %. For, given c”, C”’ E %‘, there
is g E G with Cg = C and C’g = c”‘, by basis-transitivity; since I is the unique element
of G having C and c” (or C and C”‘) as cycles, g commutes with 1. Now H fixes C U F
setwise, so the pointwise stabiliser of this set has finite index in H and so is non-trivial
(even infinite). Thus we have reached our contradiction.

REFERENCES

1. P. J. Cameron, M. Deza and N. M. Singhi, Infinite geometric groups and sets, pp. 54-61 in Algebric,
Extremal and Metric Combinatorics 1986 (ed. M.-M. Deza, P. Frank1 and I. G. Rosenberg), London
Math. Sot. Lecture Note Ser. 131 (1988), Cambridge University Press, Cambridge.
2. T. Maund, Thesis, Oxford University, (1988).
3. J. Tits, GCnCralisation des groupes projectifs bas&e sur leurs propriCtts de transitivitk. Acad. Roy.
Belgique Cl. Sci. Mem. 27 (1952).

Received I 1 March 1990 and accepted 2 July 1990

PETER J. CAMERON
School of Mathematical
Sciences, Queen Mary
and Westjield College,
Mile End Road,
London El 4NS. UK

You might also like