For a ring $A$ and an ideal $\mathfrak{a}$ of $A$, Atiyah-Macdonald define $$A^*=\bigoplus_{n=0}^\infty \mathfrak{a}^n$$ and claim that it is a graded ring on p. 107 of their commutative algebra book. My confusion is that I cannot see how $A^*$ is a ring. Does the direct sum represent the internal direct sum or the external direct sum?

Since $\mathfrak{a}^0\supseteq\mathfrak{a}^1\supseteq\cdots$, I think the internal direct sum is not possible. If it is the external direct sum, then what would be the multiplicative identity of $A^*$? Maybe I am missing something completely obvious.


It's the external direct sum, yes (as you remark the internal direct sum wouldn't make sense). The product is such that $\mathfrak{a}^i \cdot \mathfrak{a}^j \subset \mathfrak{a}^{i+j}$. The unit is the unit of $A = \mathfrak{a}^0$.

  • $\begingroup$ If it is the external direct sum, its elements be of the form $(a_0, a_1, \cdots)$, $a_i\in\mathfrak{a}_i$ and $a_i=0$ for all but finitely many $i$ and the product would be componentwise. Am I correct? $\endgroup$ – PJK Oct 7 '14 at 8:33
  • 1
    $\begingroup$ The addition would be component-wise. The product, on the other hand, would be defined by $(a_0, a_1...) \cdot (b_0, b_1...) = (c_0, c_1...)$ where $c_n = \sum_{i+j=n} a_i b_j$. That's what I meant by $\mathfrak{a}^i \cdot \mathfrak{a}^j \subset \mathfrak{a}^{i+j}$. $\endgroup$ – Najib Idrissi Oct 7 '14 at 8:36
  • $\begingroup$ I didn't think of multiplication as such. It seems to be pretty clear now. Thank you. $\endgroup$ – PJK Oct 7 '14 at 8:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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