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.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Assume $A_1,A_2,...,A_n\in M_{m×m}(F)$ (where $F$ is a field) such that

  1. $A_jA_i=A_iA_j$

  2. $A_i^2=0, \;\;\forall 1\leq i \leq n.$

If $m\lt2^n$ then how to prove that $A_1A_2...A_n=0.$

Thanks in advance.

share|cite|improve this question
Source?${}{}{}$ – Gerry Myerson Feb 13 '13 at 3:13
its annual contest math of Iran – Maisam Hedyelloo Feb 13 '13 at 3:14
Is it online$?$ – Gerry Myerson Feb 13 '13 at 3:46
@Gerry Myerson: no. contest-math question are in Sharif university website but these language are Persian. if you want i try find them for you with English language mail me your email address – Maisam Hedyelloo Feb 13 '13 at 16:15

Lemma: Suppose $T:V\rightarrow V$ satisfies $T^2 = 0$ on an $n$ dimensional vector space $V$. Then $\mathrm{rank}(T) \le \frac{n}{2}$.

Proof: Suppose that $\mathrm{rank}(T)>\frac{n}{2}$. Then $\dim\ker(T) < \frac{n}{2}$. Therefore there must exist at least one vector in the image which is not in the kernel, for otherwise the kernel contains the image. This contradicts the fact that $T^2 = 0$. $\square$

Now let $W_i$ denote the image of the product of the first $i$ matrices. That is, $$W_i = \mathrm{Im}(A_i \cdots A_1)$$ Clearly each $W_i$ is an invariant subspace of $A_{i+1}$ for we have $$A_{i+1}\left(A_i \cdots A_1\mathbf{x}\right) = \left(A_i \cdots A_1\right)A_{i+1}\mathbf{x}$$ by commutativity. Therefore there exists a valid restrict of each $A_{i+1}$ to $W_i$. The claim now is that $$\mathrm{rank}\left(A_{i+1}\big|_{W_i}\right)\le \frac{m}{2^{i+1}}$$

We proceed by induction.

By the previous lemma, the induction hypothesis is true for $A_1:F^m \rightarrow F^m$ with $W_0 = F^m$ since $A_1^2 = 0$. Now suppose the proposition holds for $i\ge 1$. Then consider $$A_{i+1}\big|_{W_{i}}: W_i \rightarrow W_i$$ By the induction hypothesis, $$\dim W_{i} = \mathrm{rank}\left(A_{i}\big|_{W_{i-1}}\right) \le \frac{m}{2^i}$$ Since originally $A_{i+1}^2 = 0$ on $F^m$, it follows that the restriction of the map to $W_i$ is also nilpotent. By the lemma, $$\mathrm{rank}\left(A_{i+1}\big|_{W_i}\right) \le \frac{\dim(W_{i})}{2} \le \frac{m}{2^{i}}$$ as required. All together, the product of the matrices can be seen as $$A_n\big|_{W_{n-1}}:W_{n-1}\rightarrow W_{n-1}$$ which has rank $\frac{m}{2^n} < 1$. This means that the mapping is $0$ as required.

share|cite|improve this answer

Your Answer


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.