9
$\begingroup$

Let $0 \to A \to B \to C \to 0$ be an exact sequence of vector spaces. I want to show that I have a canonical isomorphism

$$\det(B)= \det(A) \otimes \det(C).$$

Here, "det" refers to the $n$-th exterior product where $n$ is the dimension of the vector space.

Although it's easy to see that there is an isomorphism by fixing a basis, what is important to me is to show that this isomorphism is canonical.

$\endgroup$
  • $\begingroup$ Can you show that the isomorphism is independent of the basis? $\endgroup$ – Travis Willse Dec 12 '14 at 21:29
  • $\begingroup$ Well... this is basically what I'm struggling with. $\endgroup$ – user142700 Dec 12 '14 at 21:45
6
$\begingroup$

Denote $p := \text{dim}\ A$, $q := \text{dim}\ C$ and $n := p+q = \text{dim}\ B$. Then I claim the following makes sense$$(\ddagger)\quad {\bigwedge}^p A\otimes {\bigwedge}^q C\to{\bigwedge}^n B,\quad (a_1\wedge ...\wedge a_p)\otimes (c_1\wedge ...\wedge c_q)\longmapsto a_1\wedge ...\wedge a_p\wedge \tilde{c}_1\wedge ...\wedge \tilde{c}_q$$ for arbitrary preimages $\tilde{c}_i$ of the $c_i$ under $B\twoheadrightarrow C$.

First, for fixed $a_i$ and $c_j$ the r.h.s. is independent of the choice of the $\tilde{c}_i$:

  • If the $a_i$ do not form a basis of $A$, the right hand side vanishes independently of the choice of $\tilde{c}_i$.
  • If the $a_i$ do form a basis of $A$, any other choice $\tilde{c}_j^{\prime}$ of preimage for $c_j$ differs from $\tilde{c}_j$ by a linear combination of the $a_i$, which cancels in the r.h.s. of $(\ddagger)$.

Hence $(\ddagger)$ makes sense as a map $A^p\times C^q\to {\bigwedge}^n B$, and since it is antisymmetric and multilinear in the $A$- and $C$-variables, it descends to a morphism as indicated.

Finally, if $(a_1,...,a_p)$ is a basis of $A$ and $(c_1,...,c_q)$ is a basis of $B$, the morphism $(\ddagger)$ sends the generator $(a_1\wedge ...\wedge a_p)\otimes (c_1\wedge ...\wedge c_q)$ of the l.h.s. to the generator $a_1\wedge ...\wedge a_p\wedge \tilde{c}_1\wedge ...\wedge \tilde{c}_q$ of the r.h.s. corresponding to the basis $(a_1,...,a_p,\tilde{c}_1,...,\tilde{c}_q)$ of $B$, hence is an isomorphism.

$\endgroup$
3
$\begingroup$

Daniel Murfet's notes, Tensor, Symmetric and exterior algebras, Section 3.3.

$\endgroup$
  • $\begingroup$ Much better, as expected... :) $\endgroup$ – Hanno Dec 12 '14 at 22:04
  • $\begingroup$ The proof is the same as yours. And I even like your exposition more. $\endgroup$ – Martin Brandenburg Dec 12 '14 at 22:05
  • $\begingroup$ Just liked the formulation of Prop. 32 $\endgroup$ – Hanno Dec 12 '14 at 22:08

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.