First off, to solve this you do exactly what you've been doing in all of your other problems. Since you weren't sure of your answer, I went ahead and worked it out for you.
The augmented matrix equation that you're trying to solve is this one:
$$\left[\begin{array}{cccc|c} 1 & -1 & 2 & 3 & 0 \\ 4 & 2 & -1 & 3 & 0\end{array}\right]$$
Now I'll do Gaussian elimination to solve it. Note that I'm going to leave off the final column of $0$'s because no elementary row operation will actually change them. So just imagine that column still being there:
$$\begin{align}\begin{bmatrix} 1 & -1 & 2 & 3 \\ 4 & 2 & -1 & 3\end{bmatrix} &\sim_{R_2 \to R_2-4R_1} \begin{bmatrix} 1 & -1 & 2 & 3 \\ 0 & 6 & -9 & -9\end{bmatrix} \\ &\sim_{R_2\to \frac 16R_2} \begin{bmatrix} 1 & -1 & 2 & 3 \\ 0 & 1 & -\frac 32 & -\frac 32\end{bmatrix} \\ &\sim_{R_1\to R_1+R_2} \begin{bmatrix} 1 & 0 & \frac 12 & \frac 32 \\ 0 & 1 & -\frac 32 & -\frac 32\end{bmatrix}\end{align}$$
This is the RREF of your matrix. Now we see that columns $3$ and $4$ don't have pivots so we set $z=s$, $s\in \Bbb R$ and $w=t$, $t\in \Bbb R$. Then we write down the equations that the above matrix represents and substitute these new variables:
$$\begin{cases} x +\frac 12z+\frac 32w=0 \\ y-\frac 32z-\frac 32w=0\end{cases} \iff \begin{cases} x= -\frac 12s-\frac 32t \\ y=\frac 32s+\frac 32 t\end{cases}$$
So each element of the solution set is of the form $$\begin{bmatrix} x \\ y \\ z \\ w\end{bmatrix} = \begin{bmatrix} -\frac 12s-\frac 32t \\ \frac 32s+\frac 32 t \\ s \\ t\end{bmatrix} = s\begin{bmatrix} -\frac 12 \\ \frac 32 \\ 1 \\ 0\end{bmatrix} + t\begin{bmatrix} -\frac 32 \\ \frac 32 \\ 0 \\ 1\end{bmatrix},\quad s,t\in\Bbb R$$
Thus the set $\{(-\frac 12, \frac 32, 1, 0),(-\frac 32,\frac 32, 0, 1)\}$ spans the space. But fractions are a little annoying. So what we can do is multiply each vector by $2$. Really what we're doing here is defining new variables $s'=\frac 12s$ and $t'=\frac 12t$. Then plugging these into the above and moving the constant inside the vectors we get $$\bbox[5px,border:2px solid red]{\{(-1,3,2,0),(-3,3,0,2)\}}$$
Note that these vectors are not unique. There are an infinite number of other pairs of vectors which will span this space. For instance, Vineet's solution is another perfectly valid pair.
Now that we see this exercise just asked to you solve it the same way as all the others you've done so far, I'll tackle your question about what the word span means. Let $S=\{\mathbf u_1, \dots, \mathbf u_k\}\subset \Bbb R^n$.
Definition: A linear combination of $S$ is a vector $\mathbf w$ such that $$\mathbf w = w_1\mathbf u_1 + \cdots + w_k\mathbf u_k$$ for some $w_1, \dots, w_k \in \Bbb R$.
Example: The vector $(-7,-2,3)$ is a linear combination of $\{(1,2,3),(3,2,1)\}$ because $$(-7,-2,3) = 2(1,2,3)-3(3,2,1)$$
However $(1,0,0)$ is not a linear combination of $\{(1,2,3),(3,2,1)\}$ because there are no numbers $a,b$ such that $$(1,0,0) = a(1,2,3)+b(3,2,1)$$
Definition: The span of $S$ is the set of all linear combinations of $S$. I.e. $$\operatorname{span}(S) = \{\mathbf w \in \Bbb R^n \mid\mathbf w = w_1\mathbf u_1 + \cdots + w_k\mathbf u_k\}$$
Examples:
$$\begin{align}(1)\quad &\operatorname{span}\{(1,0,0),(0,1,0)\} = \operatorname{span}\{(10,9,0),(e,-\pi,0)\} = \text{the $xy$-plane} \\ (2)\quad &\operatorname{span}\{(1,2,3)\} = \text{the line containing $O$ parallel to $(1,2,3)$}\end{align}$$
Lemma: The span of any set of vectors in $\Bbb R^n$ is a subspace of $\Bbb R^n$. (Note: I'm not going to prove this lemma here, it's just important to know.)
So this question is asking you to find a set of two vectors whose span is the solution set to your system of linear equations.
Best Answer
As to your first question, the span of vectors is always a vector space, so it doesn't make sense to write that a set of vectors span something other than a vector space.
Now, as to your second question, yes. Think of it, what it means to span is that every vector in the space can be represented as a linear combination of vectors in the space. Thus the span of all the basis vectors also span all of the subspaces up to the space that you are interested in. I.e you could take a subset of your space that would span a subspace of V.