The definition of momentum actually comes from the definition of mass. Once you've defined mass, conservation of momentum is right around the corner. In fact, they're connected to the extent that they're basically the same idea. The only reason mass even enters Newtonian mechanics, the only reason all objects can be ascribed this unchanging constant called 'mass', is because momentum is conserved.
In an isolated system of $n$ particles evolving over time, the following identity holds for some constants $c_i$:
$$\sum c_i v_i(t)=constant$$
The above law can be taken to define the mass of the $i^{th}$ particle. The unique constant $c_i$ which satisfies the above equation is defined to be the mass of the $i^{th}$ particle.
The traditional way mass is defined in textbooks is different from the above, but it's equivalent to the above. The traditional way mass is defined is : 'Mass is an object's resistance to change in velocity'. Or more precisely, the changes in velocities of two particles in an isolated system over time is inversely proportional to their masses: $\frac{dv_1}{dv_2}=\frac{-m_2}{m1}$, where $dv1$ and $dv2$ are the changes in velocities of the two particles. The minus sign indicates that the changes are in opposite directions.
The way we defined this mass above naturally gives us another really convenient quantity to work with, called momentum:
Consider the quantity $m1v1+m2v2$ for a system of two particles before collision as well as after collision. Before collision, this quantity's value is $m_1u_1+m_2u_2$, where $u_1,u_2$ are the initial velocities. After collision, its value is $m_1(u_1+du_1)+m_2(u_2+du_2)=m_1u_1+m_2u_2+m_1du_1+m_2du_2=m_1u_1+m_2u_2+0=m_1u_1+m_2u_2$
The quantity $m_1du_1+m_2du_2$ is zero because of the definition of mass. As change in velocity is in the inverse ratio of masses (mass resists change in velocity), $\frac{du_1}{du_2}=-\frac{m_2}{m_1}$, implying, $m_1du_1=-m_2du_2$,.
This means that the momentum of a system of particles is conserved as long as only internal interactions are involved (no net interactions from outside the system). This quantity $mv$ is like a currency which simply gets exchanged in 2-particle interactions. The conservation of momentum holds for n-particle systems too, as the interactions between $n$ particles simply comprise of a bunch of two-particle interactions (all of which are just momentum exchanges).
So the reason it's natural to work with $mv$ instead of $m^2v$ is because $mv$ is where $m$ comes from in the first place.
But that's not true
- It's absolutely true that if you start with $dp/dt=ma$ and integrate both sides you recover $p=mv$. This might not be particularly helpful or the answer that the op was looking for, but it's certainly notstupid
. – aquirdturtle Sep 06 '20 at 22:09