1
$\begingroup$

Suppose that we defined a tensor product of vector spaces $U$ and $V$ as a quotient of a vector space with basis $V \times W$ by the vector space spanned by $-(u_1+u_2,v)+(u_1,v)+(u_2,v), -(u,v_1+v_2)+(u,v_1)+(u,v_2),-(au,v)+(u,av)$.

I have proved the universal property for two vector spaces. But how then extend it to a tensor product of $n$ vector spaces?

$\endgroup$
  • $\begingroup$ What goes wrong with trying to do the same thing? $\endgroup$ – Hurkyl Dec 2 '13 at 9:59
0
$\begingroup$

You can define it recursively by $V_1 \otimes \cdots \otimes V_n = \left( V_1 \otimes \cdots \otimes V_{n-1} \right) \otimes V_n$, but you want to check associativity.

See this answer.

$\endgroup$
  • $\begingroup$ Well, I proved that for vector spaces $U,V,W$, $U \otimes (V \otimes W) \cong (U \otimes V) \otimes W$. But how can we now "recursively" say that for any multilinear map $V_1 \times ... \times V_n \rightarrow U$ we have a unique linear transformation $V_1 \otimes ... \otimes V_n \rightarrow U$? $\endgroup$ – AlexCon Dec 2 '13 at 10:23

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.