MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

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

It is well known that given a short exact sequence $1\to H \to G \to G/H \to 1$ the transgression map $$ H^{p-1}(G/H, H^1(H,A)) \to H^{p+1}(G/H,A^H) $$ in the inflation-restriction sequence is in fact the cup product with the opposite of the class of the extension $1 \to H^{ab} \to G/H' \to G/H \to 1$ in $H^2(G/H, H^{ab})$, given that the action of $H$ on the module $A$ is trivial.

What is known about the transgression map in case $A$ is a non-trivial module? Is it possible then to describe transgression in similar terms?

share|cite|improve this question
Is the upper $G/H$ in $H^{p-1}(...)^{G/H}$ correct ? – Ralph May 28 '13 at 21:57
yes, you are right, shouldn't be there (I started writing this down for p=1 and then forgot to erase $G/H$) – Dima Sustretov May 29 '13 at 7:32
What does $H^*(G/H,A)$ mean if $H$ doesn't act trivially on $A$? – M T May 29 '13 at 16:34
oops, it should be $A^H$ (which obviously is $A$ when the action is trivial) – Dima Sustretov May 29 '13 at 17:36
up vote 2 down vote accepted

I can offer the following generalization for non-trivial coefficients $A$: Write $Q = G/H$. The cap product $$H_1(H;\mathbb{Z}) \otimes H^1(H;A) \to \mathbb{Z}\otimes_H A = A_H$$ is $G$-linear with trivial $H$-action and induces a cup product $$\cup: H^p(Q;H_1(H;\mathbb{Z})) \otimes H^q(Q;H^1(H;A))\xrightarrow{} H^{p+q}(Q;A_H).$$ Let $u \in H^2(Q;H_1(H;\mathbb{Z}))$ be the class that corresponds to the extension $$1 \to H^{ab} \to G/H' \to Q \to 1$$ and let $\kappa: A^H \to A_H$ be the canonical homomorphism.

Theorem: The compostion $$H^p(Q;H^1(H;A)) \xrightarrow{d_2}H^{p+2}(Q;A^H)\xrightarrow{\kappa^\ast}H^{p+2}(Q;A_H)$$ is (up to sign) cup product with $u$, i.e. $(\kappa^\ast \circ d_2^{p,1})(x) = - u \cup x$.

Proof: By abuse of notation let $\kappa: A \to A_H$. Since $\kappa$ is $G$-linear, it induces a map of spectral sequences $\kappa_r^{pq}: E_r^{pq}(A) \to E_r^{pq}(A_H)$. In particular, $\kappa_2^{p+2,0}\circ d_2^{p,1}(A)=d_2^{p,1}(A_H)\circ \kappa_2^{p,1}$. Clearly, $\kappa_2^{p+2,0}$ is the map $\kappa^\ast$ in the theorem and since $H$ acts trivially on $A_H$, we obtain $(\kappa^\ast \circ d_2^{p,1}(A))(x) = d_2^{p,1}(A_H)(\kappa_2^{p,1}x)=-u \cup \kappa_2^{p,1}x$. Hence, it remains to show $u \cup x = u \cup \kappa_2^{p,1}x$ (note the cup products are w.r.t. different pairings). But this follows immediately by applying $H^\ast(Q;-)$ to the following commutative diagram of pairings: $$\begin{array}{ccc} H_1(H;\mathbb{Z}) \otimes H^1(H;A) \;\; & \xrightarrow{\cap} & A_H \newline \scriptstyle id\otimes \kappa^\ast \displaystyle\downarrow\qquad\quad & & \downarrow\scriptstyle id \newline H_1(H;\mathbb{Z}) \otimes H^1(H;A_H) & \xrightarrow{\cap} & A_H. \end{array}$$

share|cite|improve this answer
@Ralph: I fixed a typo by changing an $H$ to a $Q$, I hope it's OK. – Mark Grant Jun 1 '13 at 12:08
Yes, it's OK. Thank you Mark. – Ralph Jun 3 '13 at 9:49
Thanks, Ralph. This seems to be the closest to what I have been hoping for. It's a pity that there is no natural pairing $H^{ab} \times H^1(H, A) \to A$, but alas, this how the things are. – Dima Sustretov Jun 3 '13 at 10:46

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.