I am puzzled over something I read in Quillen's On the Cohomology and K-Theory of the General Linear Groups Over a Finite Field.

On page 557, when computing the $E_2$ page of a case of the Eilenberg-Moore spectral sequence, he first shows that

$$E_2 \cong \text{Tor}^{A_1}(k,k) \otimes k \otimes k \otimes A_4$$

In this formula,

  • $\otimes = \otimes_k$
  • $A_1 = k[x_1, x_2, \dots]$ with $x_i$ in degree $2i$
  • $A_4 = k[y_1, y_2, \dots]$ with $y_i$ in degree $2i$.
  • The isomorphism is an isomorphism of algebras.

Based on the isomorphism above he claims that $$E_2 \cong P[y_1, y_2, \dots] \otimes \Lambda_k[e_1, e_2, \dots]$$


  • $\Lambda_R$ means "exterior algebra over $R$"
  • $y_i \in E_2^{0,2i}$
  • $e_i \in E_2^{-1,2i-1}$.

Question: I am confused about the last point, that $e_i \in E_2^{-1,2i-1}$. My understanding of the "bigrading" on $\text{Tor}^{A_1}(k,k)$ is that I can take a graded projective resolution $P^{\bullet} \rightarrow k$, and then the second index comes from the fact that the (co)homology objects of $P^{\bullet} \otimes_{A_1} k$ inherit a grading.

Of course, I'm going to use the Koszul resolution. But then,

  1. Shouldn't the generators $e_i$ live in degree $2i$ just like the $x_i$'s they're mapped to under $\Lambda_{A_1}[e_i] \rightarrow A_1$?
  2. If so, doesn't that imply that $e_i \in E_2^{-1,2i}$ ?

1 Answer 1


Different people use different notation on gradings, for example I would have called the bigrdading of $e_i$ $E_2^{1,2i}$. Supposing that this is not a typo, Quillen meant by $k$ in $E_s^{j,k}$ the total degree, not the internal degree of the element.
In other words, he has a spectral sequence $E_2^{s,t}\Rightarrow E_{\infty }^{t}$.

So you are right in 1, but in 2, the above definition of indices gives the shift.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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