Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I would like to get a more concrete understanding of a general isomorphism I have read about. I apologize if this is too basic, but I was not satisfied with the references at my disposal.

Let $K$ be a field of characteristic zero and $V$ a finite dimensional vector space over $K$.

Let $Alt_k(V,K)$ denote the set of k-multilinear alternating forms on $V$ and let $\Lambda ^k (V)$ be the k-th exterior power, defined as usual as a quotient of the tensor algebra $V^{\otimes k}$.

We have an isomorphism $Alt_k(V,K) \cong \Lambda ^k (V^*)$ induced by the analogous one at the level of the tensor algebra.

Therefore it makes sense to take some functionals $\varphi_1,\dots,\varphi_k \in V^*$ and consider $\varphi_1 \wedge\dots \wedge\varphi_k$ as a multilinear alternating form on $V$. So, take $v_1,\dots,v_k \in V$; how should the following object (=number in $K$) be defined?

$$(1) \quad \quad \quad \quad \quad \quad (\varphi_1 \wedge\dots \wedge\varphi_k)(v_1,\dots,v_k) $$

In particular, if I fix some basis $e_1, \dots, e_n$ for $V$, how should I compute the following?

$$(2) \quad \quad \quad \quad \quad \quad (e^*_{i_1} \wedge\dots \wedge e^*_{i_k})(e_{j_1},\dots,e_{j_k})$$

I can carry out the case $k=2$, so I suspect some kind of determinant is involved in (2), maybe the det of the submatrix obtained from the rows of indices $i_1,\dots,i_k$ in the matrix having the $e_{j_1},\dots,e_{j_k}$ in columns. But I cannot check this, since I don't know the right (=coordinates-free) definition for (1).


To explain what I'm looking for, I'll sketch the same things for the case of $V^{\otimes k} \cong \{ $ k-multilinear forms on $ V \}$. In this case if I take $\varphi_1,\dots,\varphi_k \in V^*$ and $v_1,\dots,v_k \in V$ then it's ok to set

$$ (\varphi_1 \otimes \dots \otimes \varphi_k)(v_1,\dots,v_k)= \varphi_1 (v_1)\cdot \textrm{ $\dots$ } \cdot \varphi_k (v_k) $$

and this gives a k-multilinear form. If I fix some basis $e_1, \dots, e_n$ for $V$, then I get

$$ (e^*_{i_1} \otimes \dots \otimes e^*_{i_k})(e_{j_1},\dots,e_{j_k})=\delta_{i_1,j_1} \cdot \textrm{ $\dots$ } \cdot \delta_{i_k,j_k}$$

So, I think one should properly reduce these expression in a way which is compatible with the ideal defining the exterior power, but I don't know how to do this.

share|improve this question
    
Well $\omega \wedge \eta = \frac{(k+l)!}{k!l!}$ Alt($\omega \otimes \eta$). Is this what you were looking for? –  Sandeep Thilakan Feb 12 at 21:00
    
I suppose Alt is defined so that this gives the same formula as the one in Peter Crook's answer below, so yes, thank you. Anyway, are all these coefficients necessary? –  Lor Feb 12 at 23:28
1  
The reason why the coefficients are defined this way is that you get $\epsilon ^I \wedge \epsilon ^J = \epsilon ^{IJ}$ –  Sandeep Thilakan Feb 13 at 6:17

1 Answer 1

up vote 0 down vote accepted

I think you want $$(\varphi_1 \wedge\dots \wedge\varphi_k)(v_1,\dots,v_k)=\frac{1}{k!}\sum_{\sigma\in S_k}sgn(\sigma)\varphi_1(v_{\sigma(1)})\varphi_2(v_{\sigma(2)})\cdots\varphi_k(v_{\sigma(k)}).$$

share|improve this answer
    
yes, thank you! maybe with some sign? I think one should get for example $(e^*_1 \wedge e^*_2)(u,v)=u_1v_2-u_2v_1$, so maybe something like $sign(\sigma)$. –  Lor Feb 12 at 23:22

Your Answer

 
discard

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.