I am working through VI Arnold "Mathematical Methods of Classical Mechanics". One of the first problems, after defining Galilean structures and Newton's equation of motion, it is to prove 'For a mechanical system of two points with initial velocity zero, the points will remain in the line that connected them in the begining'
I tried to prove as follows: denote $\vec{x} = (\vec{x}_1, \vec{x}_2)$ the position of the system, and $\ddot{\vec{x}} = \vec{F}(\vec{x},\dot{\vec{x}})$ the equations of motion. By galilean invariance I can prove that
$$ \ddot{\vec{x}_1} = \vec{F}_1(\vec{x}_1-\vec{x}_2,\dot{\vec{x}_1}-\dot{\vec{x}_2})\\ \ddot{\vec{x}_2} = -\vec{F}_1(\vec{x}_1-\vec{x}_2,\dot{\vec{x}_1}-\dot{\vec{x}_2}), $$
and, thus, proving that the center of mass $\vec{x}_1+\vec{x}_2$ stays static if $\dot{\vec{x}_i}(t_0) = \vec{0}$.
In order to prove that they remain in the line that connected them I thought I could try to prove that for any instant $t$, $\vec{x}_1(t)-\vec{x}_2(t)\propto\vec{x}_1(t_0)-\vec{x}_2(t_0)$. Rearrange the previous equations, calling $\vec{R}=\vec{x}_1+\vec{x}_2$; $\vec{r}=\vec{x}_1-\vec{x}_2$:
$$ \ddot{\vec{R}} = \vec{0}\\ \ddot{\vec{r}} = 2\vec{F}_1(\vec{r},\dot{\vec{r}}), $$
with initial conditions $\vec{R}(t_0)=\vec{R}_0$, $\vec{r}(t_0)=\vec{r}_0$, $\dot{\vec{R}}(t_0)=0$ ,$ \dot{\vec{r}}(t_0)=0$.
How can I prove that $\vec{r}(t)\propto\vec{r}(t_0)$? If it is not possile in this way, how can I prove that the points remain in the same line without invoking angular momentum conservation?
Edit: To include the insights of @Alexander : We can perform a galilean transformation such that the center of mass $R_0 = \vec{0}$, in that way the dynamics simplifies to $\ddot{\vec{r}} = \vec{F}(\vec{r},\dot{\vec{r}}) $
Because the equation must be invariant under the galilean transformations group, it follows that $\vec{F}(M\vec{r},M\dot{\vec{r}}) = M\vec{F}(\vec{r},\dot{\vec{r}})$ where $M$ is an orthogonal matrix (the rest of galilean transformations have been alreagy gauged away by previous choice $\vec{R}=\vec{0}$). This implies (I think) that the force is radial, therefore $\vec{F}(\vec{r},\dot{\vec{r}}) = f(|\vec{r}|,|\dot{\vec{r}}|)\vec{r}$ where $f$ is a scalar function. Thus $\ddot{\vec{r}}\propto\vec{r}$. Given that the initial velocity is zero it is enough to conclude that $\dot{\vec{r}}\propto\vec{r}$, and the bodies remain in the same line.
Is this logic flawed?