MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

This is a general question about group cohomology. I'm interested in the case when the coefficients are the rational numbers and hence I suppose when my groups are infinite. The question splits into two:

1) Are there any favoured examples that you would recommend a look at? (Recommended references would be just as welcome.)

And the main question:

2) What sort of functors on the category of groups leave the rational cohomology unchanged? In particular is there a projection onto a special subcategory of groups that is in some way the right category to study?

I have a feeling that someone with a good knowledge of rational homotopy theory would be able to answer this question with relative ease.

share|cite|improve this question
DC's answer is very good and I'm going to accept it, but I don't yet understand why there might not be a stronger result. I would welcome further answers which could enlighten me and the rest of MathOverflow. – James Griffin Nov 18 '09 at 9:45
up vote 11 down vote accepted

Stallings showed that if $f:\Gamma \to \Delta$ is a homomorphism between finitely presented groups where $f_i:H_i(\Gamma,Q) \to H_i(\Delta,Q)$ is bijective for $i\le 1$ and surjective for $i=2$ then $f\otimes Q: \Gamma \otimes Q \to \Delta \otimes Q$ is an isomorphism between the $Q$-unipotent completions. So "taking the $Q$-unipotent completion" is a functor with some interesting properties with respect to rational group (co)homology. I'm not sure if this is the kind of thing you're after ...

Stallings' paper is

MR0175956 (31 #232)
Stallings, John
Homology and central series of groups.
J. Algebra 2 1965 170--181.

share|cite|improve this answer
Thank you, a "Q-unipotent completion" sounds likely; I shall get hold of that paper. Would I be right or wrong in guessing that a homomorphism inducing an isomorphism between Q-unipotent completions induces an isomorphism between the homology groups? – James Griffin Nov 11 '09 at 16:39
I assume you mean "induces an isomorphism between homology in dimension 1 and a surjection in dimension 2". Off the top of my head, I don't see why this is true (or false) but maybe it is easy to see. – Danny Calegari Nov 11 '09 at 17:20
Well I've had a look at the paper and it looks like it probably pre-dates the notion (at least in name) of Q-unipotent completion. Can you recommend a reference? I not had any luck on Google. I think that your correction to my question in the comment above is a very sensible one. – James Griffin Nov 11 '09 at 17:29
I'm definitely no expert, so unfortunately I don't know a canonical reference. One reference would be Quillen's "Rational homotopy theory" from 1969 (maybe that doesn't use the name "Q-unipotent completion" either?) or Appendix A to "Fundamental groups of compact Kahler manifolds" by Amoros et. al. (AMS Math. Surveys and Monographs vol. 44). Sorry not to be more help. – Danny Calegari Nov 11 '09 at 17:47

One class of groups whose rational cohomology is particularly easy to handle are the finitely generated torsion-free nilpotent groups. This is because they admit refined Postnikov systems which can be localised, leading to a recipe for computing the Sullivan minimal model of cochains in terms of their central series. This means one can also read off cup and Massey products quite easily.

A good reference (with a crash course in rational homotopy theory included) is the paper

Oprea, John The category of nilmanifolds. Enseign. Math. (2) 38 (1992), no. 1-2, 27–40.

share|cite|improve this answer
Thank you Mark, interestingly enough I've been reading recently about the homology of the group of upper uni-triangular matrices with integer coefficients. – James Griffin Feb 17 '12 at 9:43

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.