Let $V_1, V_2, ... V_{n-1} \in R^n$ be independent vectors. I want to prove that their cross product $V_1 \times V_2 \times ...\times V_{n-1} \ne 0$.

I know that the cross product is equal to the determinant of: $$ \begin{pmatrix} e_1 & e_2 & \cdots & e_n \\ V_{1,1} & V_{1,2} & \cdots & V_{1,n} \\ \vdots & \vdots & \ddots \\ V_{n-1,1} & V_{n-1,2} & \cdots & V_{n-1,n} \\ \end{pmatrix} $$ which is then equal to $$e_1\begin{vmatrix} V_{1,2} & V_{1,3} & \cdots & V_{1,n} \\ \vdots & \vdots & \ddots \\ V_{n-1,2} & V_{n-1,3} & \cdots & V_{n-1,n} \\ \end{vmatrix} + \space... \space + e_n\begin{vmatrix} V_{1,1} & V_{1,2} & \cdots & V_{1,n-1} \\ \vdots & \vdots & \ddots \\ V_{n-1,1} & V_{n-1,2} & \cdots & V_{n-1,n-1} \\ \end{vmatrix}$$

So it is enough to prove that one of those sub-determinants is not zero.

However when I to assume they all zero, so the rows are linearly dependent and to get to a contradiction I got stuck.

Any help here?

  • $\begingroup$ I might try the contrapositive. $\endgroup$ – David Peterson Apr 7 '19 at 23:03
  • $\begingroup$ what do you mean? $\endgroup$ – Gabi G Apr 7 '19 at 23:05
  • $\begingroup$ Assume that cross product is zero, then show that as a result, the vectors cannot be linearly independent. $\endgroup$ – D.B. Apr 7 '19 at 23:09
  • $\begingroup$ That is what I tried, but it got messed up with all the equations $\endgroup$ – Gabi G Apr 7 '19 at 23:13

If $V_1, V_2, \dots, V_{n-1}$ are independent in $\mathbb R^n$, there is an $n^{\text{th}}$ vector $V_n \in \mathbb R^n$ forming a basis with them.

Then we have $$ (V_1 \times V_2 \times \dots \times V_{n-1}) \mathbin{\boldsymbol{\cdot}} V_n = \det\begin{bmatrix}V_{n1} & V_{n2} & \cdots & V_{nn} \\ V_{11} & V_{12} & \cdots & V_{1n} \\ \vdots & \vdots & \ddots & \vdots \\ V_{n-1,1} & V_{n-1,2} & \cdots & V_{n-1,n}\end{bmatrix} $$ and this determinant is nonzero because $V_1, V_2, \dots, V_n$ are linearly independent. So $V_1 \times V_2 \times \cdots \times V_{n-1}$ can't be the zero vector, otherwise it could not have a nonzero dot product with $V_n$.

If you're not convinced that the dot product above is equal to the determinant, expand the cross product as you have done, then take the dot product with $V_n$. Since $$(a_1 e_1 + \dots + a_n e_n) \mathbin{\boldsymbol{\cdot}} b = (a_1 b_1 + \dots + a_n b_n),$$ you get the same expansion, but for the determinant with $V_n$ in it.

  • $\begingroup$ It's a pretty smart answer, thank you! $\endgroup$ – Gabi G Apr 7 '19 at 23:27

The other answer covers it, but here is an approach that uses the Riesz Representation theorem: define the linear functional

$$\phi(x)=\det(V_1,\cdots, V_{n-1},x)$$

We apply Riesz to get a nonzero $z\in \mathbb R^n$ such that

$$\phi(x)=\langle x,z\rangle$$


$$\det(V_1,\cdots, V_{n-1},x)=\langle x,z\rangle$$


$$\langle e_i,z\rangle=\det(V_1,\cdots, V_{n-1},e_i)$$

is the $i^{\text{th}}$ subdeterminant in your post. If each of these were zero then,

$$\langle e_i,z\rangle =0;\ 1\le i\le n$$

and so $z=0$, a contradiction.


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.