Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

What is the Classifying Space of the Discrete Heisenberg Group? Which paper/book contains a detailed proof?

Thank you for your time.

share|improve this question
The Heisenberg manifold, which is the quotient of the real Heisenberg group by the discrete Heisenberg (sub)group –  Fernando Muro Jan 1 '12 at 2:54

2 Answers 2

up vote 8 down vote accepted

As was said by Andy, the classifying space of the discrete Heisenberg group $\Gamma$ is $B\Gamma=G/\Gamma$, where $G$ is the 3-dimensional Heisenberg group over the reals. Due to the central extension $$0\rightarrow\mathbb{Z}\rightarrow\Gamma\rightarrow\mathbb{Z}^2\rightarrow 0$$ you may view $B\Gamma$ as a circle bundle over the 2-torus. Alternatively, viewing $\Gamma$ as the semi-direct product

$\Gamma=\mathbb{Z}^2\rtimes\mathbb{Z}$, where $\mathbb{Z}$ acts by (powers of) $\left(\begin{array}{cc} 1 & 1 \\ 0 & 1 \end{array}\right)$, you can view $B\Gamma$ as the mapping torus of this automorphism of $B\mathbb{Z}^2=\mathbb{T}^2$, i.e. as a 2-torus bundle over the circle.

share|improve this answer
@Alain Valette, you have mentioned that $\Gamma$ can be viewed as the semi-direct of $\mathbb{Z}^2$ and $\mathbb{Z}$; did you mean that $n\in\mathbb{Z}$ is mapped to the element $f_n\in Aut(\mathbb{Z}^2)$, such that $f_n(m_1,m_2)=m_1+nm_2$? Then why is $\Gamma$ isomorphic to the semi-direct product? It seems that the central extension $0\rightarrow\mathbb{Z}\rightarrow\Gamma\rightarrow\mathbb{Z}^2\rightarrow0$ does not split; then is there any nature semi-direct structure from this central extension? Thank you! –  Zuriel Jan 2 '12 at 2:57
@ Zuriel. The central extension does not split. But you have another short exact sequence $0\rightarrow\mathbb{Z}^2\rightarrow\Gamma\rightarrow\mathbb{Z}\rightarrow 0$ (which of course splits) given as follows: give $\Gamma$ the presentation $\Gamma=<x,y,z|z=[x,y],[x,z]=[y,z]=1>$. Then the subgroup generated by $z$ and $y$ is normal and isomorphic to $\mathbb{Z}^2$, and the action of $x$ by conjugation on this subgroup is given by the 2-by-2 matrix mentioned in my answer. –  Alain Valette Jan 2 '12 at 9:21

If $\Gamma$ is a finitely generated torsion-free nilpotent group, then Malcev proved that there is a connected nilpotent Lie group $G$ such that $\Gamma$ is a lattice in $G$. The Lie group $G$ is often called the Malcev completion of $\Gamma$. It is an easy exercise to show that a connected simply-connected nilpotent Lie group is homeomorphic to $\mathbb{R}^n$. It follows that $G/\Gamma$ is a classifying space for $\Gamma$.

The baby example of this is $\Gamma = \mathbb{Z}^n$ and $G = \mathbb{R}^n$, so we obtain the usual $n$-torus $\mathbb{R}^n/\mathbb{Z}^n$ for the classifying space of $\mathbb{Z}^n$.

Of course, the Malcev completion of the discrete Heisenberg group is the nondiscrete Heisenberg group.

share|improve this answer
A question: will $\Gamma$ generally be a uniform (cocompact) lattice in $G$? –  Mark Jan 1 '12 at 2:40
@Mark: yes, the standard definition of the Malcev completion requires that $\Gamma$ is cocompact in $G$. This also makes Malcev completion unique. @Andy, I think you wanted to say that a simply connected nilpotent Lie group is diffeomorphic to $\mathbb R^n$. –  Vitali Kapovitch Jan 1 '12 at 3:06
@Mark: all lattices in nilpotent Lie groups are cocompact. –  Tom Church Jan 1 '12 at 3:57
@Tom Church Oops, I guess the standard definition of a lattice requires finite volume of the quotient. Then cocompactness is of course automatic for nilpotent groups. I've often had to work with discrete noncocompact subgroups of nilpotent Lie groups which I often also call lattices but I guess that's wrong terminology. –  Vitali Kapovitch Jan 1 '12 at 4:20
@Vitali : Thanks for the correction! I've been celebrating the new year, so I'm pretty happy I managed to at least put the word "connected" in all the right places... –  Andy Putman Jan 1 '12 at 5:05

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.