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

Suppose $k$ is an algebraically closed field of characteristic zero and $A$ is a finitely generated commutative associative reduced $k$-algebra.

Suppose the group $\mathbb{Z}_2$ acts on $A$ in such a way that the induced action on maxSpec $A$ is free. (Side note: This is equivalent to the induced grading $A = A_0 \oplus A_1$ being a strong grading, i.e., $A_1^2 = A_0$.)

Let $S^2 A = (A \otimes A)^{S_2}$, where $S_2$ acts by swapping the factors in the tensor product. The action of $\mathbb{Z}_2$ on $A$ induces a diagonal action of $\mathbb{Z}_2$ on $A \otimes A$. Since this action commutes with the $S_2$ action, we have an induced action of $\mathbb{Z}_2$ on $S^2 A$.

Now let $(S^2 A)_{\mathbb{Z}_2}$ be the corresponding algebra of coinvariants. By definition, this is the quotient of $S^2 A$ by the ideal generated by elements of the form $(b - g \cdot b)$ for $g \in \mathbb{Z}_2$ and $b \in S^2A$.

My question is: Is $(S^2 A)_{\mathbb{Z}_2}$ isomorphic to $A^{\mathbb{Z}_2}$ (as an algebra)?

My reason for thinking this is true is the following intuition:

  • $S^2 A$ is the coordinate ring of the set of unordered pairs of points of $A$.

  • Thus $(S^2 A)_{\mathbb{Z}_2}$ is the coordinate ring of set of unordered pairs of points of $A$ that are invariant under the $\mathbb{Z}_2$-action.

  • Since the action of $\mathbb{Z}_2$ is free, unordered pairs of points of $A$ that are invariant under the $\mathbb{Z}_2$ action are just $\mathbb{Z}_2$-orbits.

  • Thus $(S^2 A)_{\mathbb{Z}_2}$ should be isomorphic to $A^{\mathbb{Z}_2}$, which is precisely the coordinate ring of the variety of $\mathbb{Z}_2$-orbits.

I think the above reasoning is rigorous on the level of maximal ideals. But I'd like to know that the rings are actually isomorphic.

share|cite|improve this question

You are right, they are isomorphic.

Denote by $g$ the generator of $\Bbb Z_2$. The map given by $x\otimes y\mapsto x^gy$ is obviously a homomorphism $h:A\otimes A\to A$ of algebras. Since $S^2A=S^2A_0+S^2A_1+K$, where $K$ is spanned by $a_0\otimes a_1+a_1\otimes a_0$, $a_i\in A_i$, and $A_1^2=A_0$, we conclude that $h(S^2A)=A_0$ (even $h(S^2A_1)=A_0$) and that $K$ generates in $S^2A$ the ideal $I$ such that $S^2A/I$ is the algebra of coinvariants.

In order to show that the kernel of $h$ on $S^2A$ equals $I$, we observe that, providing $\sum_ia_{0i}a_{1i}b_{1i}=\sum_jc_{1j}d_{1j}$, we have $$\sum_i(a_{0i}\otimes a_{1i}b_{1i}+a_{1i}b_{1i}\otimes a_{0i})+\sum_j(c_{1j}\otimes d_{1j}+d_{1j}\otimes c_{1j})=$$ $$=\sum_i(a_{0i}\otimes a_{1i}+a_{1i}\otimes a_{0i})(1\otimes b_{1i}+b_{1i}\otimes1)-$$ $$-\sum_i(1\otimes a_{0i}b_{1i}+a_{0i}b_{1i}\otimes1)(1\otimes a_{1i}+a_{1i}\otimes1)+$$ $$+\sum_j(1\otimes c_{1j}+c_{1j}\otimes1)(1\otimes d_{1j}+d_{1j}\otimes1)\in I,$$ where $a_{0i}\in A_0$ and $a_{1i},b_{1i},c_{1j},d_{1j}\in A_1$.

Let $s_i\in S^2A_i$ be such that $h(s_0+s_1)=0$. We can write $s_0=\sum_i(a_{0i}\otimes a_{1i}b_{1i}+a_{1i}b_{1i}\otimes a_{0i})$ and $s_1=\sum_j(c_{1j}\otimes d_{1j}+d_{1j}\otimes c_{1j})$ for suitable $a_{0i}\in A_0$ and $a_{1i},b_{1i},c_{1j},d_{1j}\in A_1$ because $A_0=A_1^2$. From $h(s_0+s_1)=0$, we obtain $\sum_ia_{0i}a_{1i}b_{1i}=\sum_jc_{1j}d_{1j}$ (the characteristic of $k$ is different from $2$). By the above idenity, $s_0+s_1\in I$.

share|cite|improve this answer

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.