# Trouble understanding finite vector spaces and Gaussian coefficent

I have studied linear algebra for 2 months now and i cannot understand a task that i am currently trying to solve. Basically i am trying to find the amount of bases for n-dimensional vector space over a field with q elements. Let's call our Vector base V. Because it is field with q elements, i know that 0 is in V.

Now i think that if we make a vector out of all basis vectors and because they are linearly independent , from this i gather that every row should have atleast 1 element that is nonzero. Also it should be able to do simple matrix operations so that nonzero elements appear only on main diagonal. From this i should have a matrix that looks like this: \begin{pmatrix} a_1 & 0 & 0 & \cdots & 0 \\ 0 & a_2 & 0 & \cdots & 0 \\ 0 & 0 & a_3 & \cdots & 0 \\ \vdots & \vdots & \vdots & \ddots & \vdots \\ 0 & 0 & 0 & \cdots & a_n \end{pmatrix} So for choosing $a_1$ i have $q^n-1$ choices, because i swap nonzeros with random elements from V, for $a_2$ i have $q^n-q$ choices, and so on until for $a_n$ where i don't have any choices, this gives me $(q^n-1)(q^n-q)(q^n-q^2)*...*(q^n-q^(n-1))$ choices. ( extra question are $a_i$ = 1 or are they just some numbers , i wasn't sure about that.

But from what i researched from here and the web i found this place http://cameroncounts.wordpress.com/2011/07/20/the-field-with-one-element/ , this is wrong, and i should be getting Gaussian coefficent $Gauss(n,k)_q$ ,but i just cannot understand where the denominator part comes from ... Also what is k in this equation ???

For n = 2 and q = {0,1} i tried manually and i got 3 different bases 1) {0,1} and {1,0} 2) {1,1} and {1,0} 3) {1,1} and {0,1} , but my equation gives me answer 6, but it should be 3, i found somewhere also that the denominator should be $n!$, that seems to work for n = 2 and q = 2.

If you answer, try to use beginner like language sinse i've only studied Algebra 2 months( vector bases 1 week ) and never used Gaussian coefficent.

We will find the number $N$ of ordered bases, that is, of sequences $v_1,v_2,\dots,v_n$ of vectors such that $\{v_1,v_2,\dots,v_n\}$ is linearly independent. Then the number of ordinary (unordered) bases is $\frac{N}{n!}$.

We can assume that our vector space is the space $F^n$ of all $n$-tuples with elements in our field $F$, since any $n$-dimensional space over $F$ is isomorphic to $F^n$.

The first vector $v_1$ can be any of the $q^n-1$ non-zero vectors.

The second vector $v_2$ can be any of the $q^n$ non-zero vectors, except for the multiples of $v_1$. There are $q$ multiples of $v_1$, so for each choice of first vector, there are $q^n-q$ choices of second vector. Thus if we our constructing our basis sequentially, there are $(q^n-1)(q^n-q^2)$ ways to choose the first two vectors.

For every choice of $v_1$ and $v_2$, the third vector $v_3$ can be any vector except for the linear combinations of $v_1$ and $v_2$. Since $v_1$ and $v_2$ are linearly independent, there are $q^2$ such linear combinations. So for each choice of $v_1$ and $v_2$, there are $q^n-q^2$ choices for $v_3$.

Continue. At the end, we find that for every choice of $v_1,v_2,\dots, v_{n-1}$ there are $q^n-q^{n-1}$ choices for $v_n$, and therefore $$N=(q^n-1)(q^n-q)(q^n-q^2)\cdots(q^n-q^{n-1}).$$ We conclude that the number of (unordered) bases is $$\frac{(q^n-1)(q^n-q)(q^n-q^2)\cdots(q^n-q^{n-1})}{n!}.$$

• Thanks , very good and clean answer. – La'tel Apr 10 '14 at 6:18
• You are welcome. – André Nicolas Apr 10 '14 at 6:23