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.

Here is another problem from Golan.

Problem: Let $F$ be a finite field. Show there exists a symmetric $2\times 2$ matrix over $F$ with no eigenvalues in $F$.

share|improve this question
    
No, if you have $a = 0$ then the matrix has eigenvalues iff $b$ is a square. –  Tobias Kildetoft Jun 14 '12 at 8:06
    
I think you have to take care of the fact that your entries are from finite field. –  Theorem Jun 14 '12 at 8:08
3  
Why do you have $a$ for both diagonal elements? A general symmetric $2\times2$ matrix has the form $$\pmatrix{a&b\\b&c}\;.$$ –  joriki Jun 14 '12 at 8:14
    
@joriki Thanks for pointing that out! –  Potato Jun 14 '12 at 8:18
    
in the question if your field is algebraically closed this could not happen. Because every polyniomial can be written as a product of first degree polynomials..right? –  clark Jun 14 '12 at 8:45

1 Answer 1

up vote 3 down vote accepted

The solution is necessarily split into two cases, because the theory of quadratic equations has a different appearance in characteristic two as opposed to odd characteristic.

Let $p=\mathrm{char}\, F$. Assume first that $p>2$. Consider the matrix $$ M=\pmatrix{a&b\cr b&c\cr}. $$ Its characteristic equation is $$ \lambda^2-(a+c)\lambda-(ac-b^2)=0.\tag{1} $$ The discriminant of this equation is $$ D=(a+c)^2-4(ac-b^2)=(a-c)^2+(2b)^2. $$ By choosing $a,c,b$ cleverly we see that we can arrange the quantities $a-c$ and $2b$ to have any value that we wish. It is a well-known fact that in a finite field of odd characteristic, any element can be written as a sum of two squares. Therefore we can arrange $D$ to be a non-square proving the claim in this case.

If $p=2$, then equation $(1)$ has roots in $F$, if and only if $tr((ac-b^2)/(a+c)^2)=0.$ By selecting $a$ and $c$ to be any distinct elements of $F$, we can then select $b$ in such a way that this trace condition is not met, and the claim follows in this case also.

share|improve this answer
    
Admittedly my answer assumes some familiarity with finite fields. I was running on autopilot here, sorry. If I come up with something even more elementary, I will edit. –  Jyrki Lahtonen Jun 14 '12 at 8:20
    
Is it not that the field must be of prime characteristic ? Just a bit confused although primes are definitely odd. –  Theorem Jun 14 '12 at 8:25
    
Interesting solution :) . –  Theorem Jun 14 '12 at 8:27
    
@Theorem If the characteristic was $ab$ for $a,b>1$, then the elements $a$ and $b$ (modelling the "field" as numbers from $0$ to $ab-1$ as usual) would be non-zero but multiply to $0$, and fields can't have zero divisors. –  Matt Pressland Jun 14 '12 at 9:49

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.