COHOMOLOGY WITH & Sunil Chebolu Illinois State University ISMAA Annual meeting, Bradley University, April 3-4 2009. Plan of my talk Motivation Cohomology Cohomology with stones and sticks Applications 1. Groups acting on spheres 2. Generating hypothesis Plan of my talk Motivation Cohomology Cohomology with stones and sticks Applications 1. Groups acting on spheres 2. Generating hypothesis Motivation k - Field of characteristic p (p > 0 prime) G - Finite group kG - Group ring = { g 2 G αg g | αg 2 k } Ultimate Goal: Explore the hidden secrets of finite groups. A traditional approach: One studies finite groups by examining how they act on k - vector spaces. A simple example C2 = h ¾ | ¾2 = 1 i acts on k2 = k © k ¾ (v) = -v ¾2(v) = ¾ (¾ (v)) V = ¾(-v) = -(-v) -V =v Therefore ¾2 = 1 This is a 2-dimensional representation of C2 Definition: An n-dimensional representation of G is a homomorphism of groups ½ : G ! GLn(k). GLn(k) = group of all invertible n £ n matrices over k. Example: C2 ! GL2(k) ¾ – I2 1 I2 Thus, an n-dimensional representation of G is an n-dimensional vector space V over k on which G acts by linear automorphisms. || n-dimensional kG - module Fundamental problem: Given G, classify all of its representations. This problem is notoriously hard ! To study finite groups and their representations, a very powerful algebraic machine was introduced in the early 20th century. This is Group Cohomology --- one of the biggest inventions in the 20th century. G finite group + RG cohomology ring +, x The cohomology ring RG is a very useful object to study invariants of G Henri Poincaré David Hilbert The history of cohomology is long and rich – Early 1900s: Inception – Late 1940s: Matured – Thereafter: active area of research Three avatars of cohomology – Representation theory (G = finite group) Group cohomology – Topology (G = topological group) Continuous cohomology – Number theory (G = Galois group) Galois cohomology Plan of my talk Motivation Cohomology Cohomology with stones and sticks Applications 1. Groups acting on spheres 2. Generating hypothesis Cohomology RG = H*(G, k) = ©n ≥ 0 Hn(G,k) is a ring. char(k) divides |G| Modular case Towards the definition of cohomology: kG linear map: This is a map Á : M ! N between representations M and N – Á is a linear transformation – Á (g ¢ m) = g Á(m) for all g in G and m in M. HomG(M, N) = all kG-linear maps from M to N. Examples of representations – Trivial representation of G is k with trivial G action, i.e., gx = x – Regular representation of G is kG kG = { g 2 G αg g | αg 2 k } – Free representation of G is kG © kG © © kG – Sygizies of a representation M are 1 M = ker( F1 ³ M) 2 M = ker( F2 ³ 1 M) n M = ker( Fn ³ n-1 M) for n > 1. Definition of Hn(G,M) Hn(G,M) = HomG(n k, M)/~ f,g : n k ! M are homotopic (f~g) iff f – g: n k M F H*(G, M) = ©n ≥ 0 Hn(G,M) graded k-vector space. Notation: Hom(A, B) = HomG(A, B)/~ Important special case: M = k (trivial rep.) H*(G, k) = ©n ≥ 0 Hn(G,k) graded ring Ring structure: ® 2 H2(G, k) ) ®¯ 2 H5(G,k) ¯ 2 H3(G, k) ¯ ® 2 3 k k k ®¯ : 5 k 3® 3 k = 3 k ¯ k k Plan of my talk Motivation Cohomology Cohomology with stones and sticks Applications 1. Groups acting on spheres 2. Generating hypothesis Cohomology with stones and sticks C2 = h ¾ | ¾2 = 1 i Char(k) = 2 Freshman’s dream kC2 = k[¾]/(¾2 – 1) = k[¾]/(¾ -1)2 = k[x]/(x2) k= kC2 = x k = ker( kC2 ! k) = ker ) = i k = k for all i. = k Hi(C2, k) = Hom(i k, k) = Hom(k, k) = kh xii H*(C2, k) = ©i≥0 kh xi i x1 x1 = x12 2 H2(C2,k) = h x2 i Products: 1 k x2 : k x1 x1 1 k x1 k k = k x1 x 12 = x 2 Similarly, x1n = xn. ) H*(C2, k) k[x1] k k Klein four group V4 V4 = C2 © C2 = h ¾, ¿ | ¾2 = 1 = ¿2 = 1, ¾ ¿ = ¿ ¾ i This is our favourite group ! kV4 = k[x, y]/(x2, y2) = kh1, x, y, xy i k= x = ¾ -1, y = ¿ -1 kV4 = x y y x Sygizies of k 1 k = ker (kV4 ³ k) = ker = 2 k = ker (kV4 © kV4 ³ 1 k) = ker © = Similarly, we have n k = n copies -n k = n copies Computing H*(V4, k) H1(V4, k) = Hom(1 k, k) = Hom =k , u =khui©khvi ©k v H2(V4, k) = Hom(2 k, k) = Hom , =k l © k m © k r =khli©khmi©khri © Products Educated guess: l = u2, m = uv, r = v2 ! u2 = l u u u.u = u = u2 = uv = m u.v = v u v = = uv v2 = r v v vv= v = = v2 Hn(V4, k) = Hom(n k, k) = Hom , = k-span h un, un-1v, , uvn-1, vn i . Combining all these,we have: H*(V4, k) = © Hn(V4, k) = © k-span h un, un-1v, , uvn-1, vn i = k [u, v] Plan of my talk Motivation Cohomology Cohomology with stones and sticks Applications 1. Groups acting on spheres 2. Generating hypothesis Applications §1. Groups acting freely on spheres. Problem: Given a finite group G, can it act freely on some sphere (Sn, for some n). Definition: An action Á : G £ Sn ! Sn is free if Á(g, x) = x ) g = e . Theorem: If G acts freely on Sn then H*(G,k) is periodic with period dividing n. Special case: G = V4 Recall: H*(V4, k) k[u, v]. In particular, Hn(V4, k) is a k-vector space of dimension n+1 with basis { un, un-1v, , vn }. ) limn ! 1 dimk Hn(V4, k) = limn ! 1 n+1 = 1 Conclusion: V4 cannot act freely on any sphere. !!! YAY !!! §2. Generating hypothesis -- joint work with Carlson and Mináč. Jan Mináč Jon Carlson Induced map in cohomology G finite group and M, N kG –modules Á : M ! N be a kG-linear map. Á induces a map in cohomology: Hi(G, Á) : Hi(G, M) ! Hi(G, N) ®Á±® Á±®: i k ® M Á N The problem of generating hypothesis Suppose Hi(G, Á) : Hi(G, M) ! Hi(G, N) is 0 for all i>0. Does this imply that Á : M ! N is null-homotopic? i.e., does Á factor through a free representation? Remark: The converse of the above problem is true for a trivial reason. M Á F N ) Hi(G, Á) = 0 because Hi(G, F) = 0 8 i > 0 Theorem: (Carlson, C, Mináč) Generating hypothesis holds for G The p-Sylow subgroup of G = C2 or C3 A map h of kV4 modules h is not null homotopic h is zero in cohomology Ars Longa Vita Brevis Thank you