Let $I$ be a directed order set. And let $\{ (M_i,\mu_{ij}) | i\leq j\} $ be a directed system of modules with $M = \lim M_i$ the direct limit. With maps $\mu_i :M_i\to M$ satisfying the required properties.

Fix a module $N$. We also have that the system $(M_i\times N),\mu_{ij}\times 1$ is directed and furthermore $\lim (M_i\times N) = M\times N$ with the limiting maps $M_i\times N\to M\times N$ given by $\mu_i\times 1$.

The system $(M_i\otimes N,\mu_{ij}\otimes 1)$ is also directed. Let $P = \lim (M_i\otimes N)$ (do not assume that $P=M\otimes N$, because I am trying to prove this). Passing to the limit we have a map of module $g:M\times N\to P$.

My book says to prove that $g$ is bilinear. This makes no sense. How can it both be linear and bilinear (unless it is trivial map)?

  • 2
    $\begingroup$ Um, what's with the algebra-precalculus tag? $\endgroup$ – Matt Samuel May 8 '15 at 19:14
  • $\begingroup$ Well I didn't learn it until graduate school. Guess that means I'm slow, or something. $\endgroup$ – Matt Samuel May 8 '15 at 19:17

$g$ is not assumed to be linear. The purpose is to eventually show that $g:M\times N\to P$ is the universal bilinear map characterizing $P$ as the tensor product of $M$ and $N$.

  • $\begingroup$ Let $M_i$ and $N_i$ be modules over the same directed set. If we have a family of maps $g_i:M_i\to N_i$ which makes the diagrams commute then we have a map $g:M\to N$, the direct limit, which is a homomorphism of modules and makes the appropriate diagrams commute. $\endgroup$ – Nicolas Bourbaki May 8 '15 at 19:23
  • $\begingroup$ What are your maps? Are they homomorphisms of modules or are they bilinear? I've never heard of using a homomorphism of modules $M\times N\to M\otimes N$. $\endgroup$ – Matt Samuel May 8 '15 at 19:25
  • $\begingroup$ If you have Atiyah-Macdonald the construction comes from page 33 exercise 18. $\endgroup$ – Nicolas Bourbaki May 8 '15 at 19:28
  • $\begingroup$ Sorry I don't. But if you have maps $M_i\times N\to M_i\otimes N$, I'd bet a lot of money that they're supposed to be bilinear, not homomorphisms of modules. $\endgroup$ – Matt Samuel May 8 '15 at 19:29
  • $\begingroup$ Let $(M_i)$ be a directed system of modules. If we multiply through by $N$, we get a directed system $M_i\times N$. The maps within this new directed system are not bilinear, they are linear maps. $\endgroup$ – Nicolas Bourbaki May 8 '15 at 20:13

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.