Assume that $V$ is a finite dimensional real vector space of dimension $n$. Is there a $\mathbb{R} -$ valued $k$- linear map $T$ on $V$ which is not an alternative form but it vanish on all $k$- tuple $(x_1,x_2,\ldots,x_k)$ with $\sum_{i=1}^k x_i =0$?


No. Notice that by linearity we should have $$ f(v_1,\dots,v_{k-1},\alpha_1v_1+\dots+\alpha_{k-1}v_{k-1}) =\frac{f(\alpha_1v_1,\dots,\alpha_{k-1}v_{k-1},-\alpha_1v_1-\dots-\alpha_{k-1}v_{k-1})}{(-1)^k\alpha_1\cdots\alpha_{k-1}} =0 $$ for all nonzero $\alpha_1,\dots,\alpha_{k-1}$. Since each linear combination is a sum of two linear combinations with nonzero coefficients, this also holds for all $\alpha_i$ [this argument works also for all fields with at least 3 elements]. Thus, the linear map $f$ vanishes on all linearly dependent $k$-tuples, in which case it is well known to be alternative.

  • $\begingroup$ Thanks for your answer. How does this argument work for a triple $f(v,v,w)$, for example? $\endgroup$ – Ali Taghavi Sep 27 '17 at 14:52
  • $\begingroup$ You may perform the same for the last two vectors instead of the first two. $\endgroup$ – Ilya Bogdanov Sep 27 '17 at 15:05
  • $\begingroup$ Yes but you implicitly assume that all $\alpha_i$ are non zero. i can imagine a resolution for three vector: $2f(v,v,w)=f(v,v.2w+2v)+f(v,v,-2v)=0$. I guess that the general case can be solved, similarly. Yes? $\endgroup$ – Ali Taghavi Sep 27 '17 at 17:18
  • $\begingroup$ No, I first explicitly tell that all $\alpha_i$ are nonzero, and second tell that by summing two such expression we can get an arbitrary linear combination, with each coefficient being zero or nonzero on our choice. $\endgroup$ – Ilya Bogdanov Sep 27 '17 at 19:17
  • 1
    $\begingroup$ $f(v,v,w)=f(1\cdot v+0\cdot w,v,w)=f(2v+w,v,w)-f(v+w,v,w)=0$. $\endgroup$ – Ilya Bogdanov Sep 28 '17 at 4:43

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.