How would one explicitly write the elements of $PSL(n, q)$? Say for $PSL(2,3)$. I know that it is isomorphic to $A_4$, but in this case I am trying to prove this fact, so I can't use those elements.
I know that $$PSL(n,q) = SL(n,q)/Z(SL(n,q))$$
where $Z$ denotes the centre of the group and $SL(n,q)$ is the group of $n \times n$ matrices with elements in $F_q$ having determinant $1$. Is there a (relatively) simple to actually write down the explicit elements of this group? I'm not sure how they will be written due to the quotient.
Here is an explicit isomorphism between the two groups using an intermediate group:
$$\begin{array}{|l|l|l|} \hline PSL(2,\mathbb{Z_3})& \ \ \ Homog(2,\mathbb{Z_3})& \ \ \ \ \ \ \ \ \ \ \ \ \ A_4\\ \hline \pmatrix{1&0\\0&1}&h_{1}(z)=z&\pmatrix{0 & 1 & 2 & \infty\\0 & 1 & 2 & \infty}\\ \hline \pmatrix{1&0\\1&1}&h_{2}(z)=\dfrac{z}{z+1}&\pmatrix{0 & 1 & 2 & \infty\\0 & 2 & \infty & 1}\\ \hline \pmatrix{1&1\\0&1}&h_{3}(z)=z+1&\pmatrix{0 & 1 & 2 & \infty\\1 & 2 & 0 & \infty}\\ \hline \pmatrix{1&0\\2&1}& h_{4}(z)=\dfrac{z}{2z+1}&\pmatrix{0 & 1 & 2 & \infty\\0 & \infty & 1 & 2}\\ \hline \pmatrix{1&1\\1&2}&h_{5}(z)=\dfrac{z+1}{z+2}&\pmatrix{0 & 1 & 2 & \infty\\ 2 & \infty & 0 & 1}\\ \hline \pmatrix{1&1\\2 & 0}&h_{6}(z)=\dfrac{z+1}{2z}&\pmatrix{0 & 1 & 2 & \infty\\ \infty & 1 & 0 & 2}\\ \hline \pmatrix{1&2\\0 & 1}&h_{7}(z)=z+2&\pmatrix{0 & 1 & 2 & \infty\\ 2 & 0 & 1 & \infty}\\ \hline \pmatrix{1&2\\1 & 0}&h_{8}(z)=\dfrac{z+2}{z}&\pmatrix{0 & 1 & 2 & \infty\\ \infty & 0 & 2 & 1}\\ \hline \pmatrix{2&1\\1 & 1}&h_{9}(z)=\dfrac{2z+1}{z+1}&\pmatrix{0 & 1 & 2 & \infty\\ 1 & 0 & \infty & 2}\\ \hline \pmatrix{0&2\\1 & 0}&h_{10}(z)=\dfrac{2}{z}&\pmatrix{0 & 1 & 2 & \infty\\ \infty & 2 & 1 & 0}\\ \hline \pmatrix{0&2\\1 & 1}&h_{11}(z)=\dfrac{2}{z+1}&\pmatrix{0 & 1 & 2 & \infty\\ 2 & 1 & \infty & 0}\\ \hline \pmatrix{0&1\\2 & 1}&h_{12}(z)=\dfrac{1}{2z+1}&\pmatrix{0 & 1 & 2 & \infty\\ 1 & \infty & 2 & 0}\\ \hline \end{array}$$
i.e., we are going to establish a combined isomorphism between 3 groups, the middle one making the connection between the two extreme ones:
The group $PSL(2,\mathbb{Z_3})$ for matrix multiplication,
The group of homographic functions on $\mathbb{Z_3}$ extended to $\mathbb{Z_3} \cup \{ \infty\}$ for the composition of functions (see below), and
The group $A_4$ of even permutations on 4 objects $\mathbb{Z_3} \cup \{\infty\}$, for the composition of bijections.
This "Rosetta stone" deserves of course a detailed explanation.
Let us begin by recalling what is $PSL(2,\mathbb{Z_3})$ by using a kind of etymological analysis:
(1) P is as "Projective", [see below],
(2) SL2 as "Special Linear", $2$-dimensional, more precisely: "L2" is for $2 \times 2$ matrices and "S" (special) for determinant equal to 1 (mod. 3),
(3) the entries of these matrices being in $\mathbb{Z_3}=\{\bar{0},\bar{1},\bar{2}\}$ with the usual addition and multiplication mod. 3, with the following rules: $3 \equiv 0$, $2 \equiv -1$, $2 \times 2 \equiv 1.$
About the term "projective": two matrices are in the same equivalence class (thus are considered as a same projective "object") iff they obey the equivalence rule:
$$\tag{*}\exists k \neq 0 \ \ \text{such that} \ \ \pmatrix{a&c\\b&d}=k \pmatrix{a'&c'\\b'&d'}.$$
In fact, the unique non trivial case is with $k \equiv 2$ (mod $3$).
Therefore, every equivalence class, i.e., every coset, has 2 elements.
For example, the two following matrices are considered as a same element:
$$\pmatrix{2&0\\1&2} \sim \pmatrix{1&0\\2&1},$$
because the first one is twice the second and reciprocally (only one of them is present in the first column of the table upwards).
Remark: The equivalence relationship (*) is the same as yours because the center of group $GL(2,\mathbb{Z_3})$ is the group with two elements:
$$I=\pmatrix{1&0\\0&1} \ \ \text{and} \ \ 2 I=\pmatrix{2&0\\0&2}$$
constituting the neutral class (the "coset" of $I$).
It's time now to explain the isomorphism between $PGL(2,\mathbb{Z_3})$ and $A_4$, that has been explicited in the table at the beginning.
Every matrix $M=\pmatrix{a&c\\b&d} \in PGL(2,\mathbb{Z_3})$, can be associated with a unique "homography":
$$z \rightarrow h_M(z)=\frac{az+c}{bz+d}$$
Why this uniqueness ? Because equivalence relationship (*) "up to a non-zero multiplicative factor $k$" is associated with a cancellation when a same factor $k$ is present in the numerator and in the denominator of a fractional expression.
Let us now consider the images of $(0,1,2,\infty)$ by $h_M.$
Remark: the presence of $\infty$ isn't a true surprise in such a context because, in general, projective geometry uses infinity almost constantly. Without giving a fully rigorous explanation, it suffices to say that the rules governing the special symbol $\infty$ are the same as the rules dealing with limits "when the variable tends to $\infty$" in real analysis.
Let us consider the example of $M=\pmatrix{1&1\\1&2}$, and its associated homography $h(z)=\frac{1z+1}{1z+2}=1-\frac{1}{z+2}$. Using the upsaid rules, we have:
$$\begin{cases}h(0)&=&1/2=1.2^{-1}=2,\\h(1)&=&"1/0"=\infty,\\h(2)&=&0/4=0,\\h(\infty)&=&1 \ - "1/\infty"=1-0=1. \end{cases}$$
and the fact is that this permutation $\pmatrix{0 & 1 & 2 & \infty\\2 & \infty & 0 & 1}$ on the 4 objects $(0,1,2,\infty)$ is even...
A "rationale" for this evenness : The only possible subgroup with 12 elements of the symmetrical group with $4!=24$ elements is the group of even permutations $A_4$ ($A_n$ is the only subgroup of $S_n$ of index $2$.).
Remark : for projective geometry and homographies on finite fields or rings, see (https://en.wikipedia.org/wiki/Projective_line_over_a_ring).
Edit: using SAGE (calling sagecell.sagemath.org), one can have further informations, for example :
G=PSL(2,3);GAnswer : Permutation Group with generators [(2,3,4), (1,2)(3,4)]
With the request :
G.list()one gets the representation
(under the form of permutations, presented as composition of cycles). A good and simple exercise is to establish a correspondence between letters $a,b,\cdots l$ and the last column of the array above. The following request :
G.cayley_table()gives :
If furthermore, you ask :
G.is_isomorphic(AlternatingGroup(4))the answer is :
True...