I'm looking for a theorem about quotient of quasi-isomorphic cdga's:

Let $A, B$ be two cdga's (commutative differential $\mathbb Z$-graded algebra) of nonpositive degrees, and $\mathfrak m \subset A, \mathfrak n \subset B$ be maximal ideals closed under differentials. Suppose $\phi: A \to B$ is a quasi-isomorphism of cdga's and $\phi(\mathfrak m) \subset \mathfrak n$. Is the induced map $\phi: A/ \mathfrak m^2 \to B/\mathfrak n^2$ still a quasi-isomorphism?

I guess more assumptions on $A,B, \mathfrak m, \mathfrak n$ might be needed. Any theorems of this type would be helpful.

  • $\begingroup$ Are your cdgas $\mathbb{N}$-graded, $\mathbb{Z}$-graded...? $\endgroup$ – Najib Idrissi Nov 8 '18 at 8:24
  • $\begingroup$ @NajibIdrissi Z-graded $\endgroup$ – Hsuan-Yi Liao Nov 9 '18 at 3:05
  • 4
    $\begingroup$ It seems that you have changed the question in a rather drastic way. It would be better to post a new question. $\endgroup$ – Mike Miller Nov 9 '18 at 3:06

Let $A$ be the complex $\Bbb Q$ concentrated in degree zero, and let $B$ be the CDGA $(\Bbb Q[t] \to \Bbb Q[t] dt)$ concentrated in (cohomological) degrees $0$ and $1$, with de Rham differential $d f(t) = f'(t) dt$. Then $B$ is a commutative DGA under the wedge product, and the natural inclusion $\Bbb Q \to \Bbb Q[t] \subset B$ becomes a quasi-isomorphism of CDGAs.

The only maximal ideal $\mathfrak m \subset A$ is zero, whereas there are a lot of maximal ideals of $B$ of the form $\mathfrak n = ((f(t)) \to \Bbb Q[t] dt)$ for $(f(t)) \subset \Bbb Q[t]$ a maximal ideal of $\Bbb Q[t]$ generated by a monic irreducible polynomial $f(t)$. The resulting map on quotients is $\Bbb Q \to \Bbb Q[t] / (f(t))$, which is an isomorphism if and only if $f(t)$ is linear; for example, we could take $f(t) = (t^2 + 1)$.

For a positive result, you could ask about CDGAs concentrated in nonpositive (cohomological) degree such that $\mathfrak m$ and $\mathfrak n$ are closed under the differential; then any maximal ideal automatically contains all elements in nonzero degree.


Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

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