You have successfully developed an intuition for the conservation of linear momentum. As the case of a single particle is a very specialized and ideal case (every particle can be zoomed into and resolved into multiple parts) let us only consider groups of particles:

An isolated group of particles, too, keep drifting as a whole in an average direction unless forced to change.

That is the conservation of linear momentum.

Now, forget about the net (average) direction of the whole and start looking at the parts. (i.e., go to the center of mass frame). And suppose that you have fine threads in your hands, attached to each of the parts so that they don't zoom away. (These are the internal forces). It does not matter how fast you pull at the parts or even if you don't pull at the parts at all.

Now, in general, your threads cannot change the rotations of the parts around you. You can pull them closer, making their orbits tighter, but then they will get faster, as you might have observed if you ever spin when ice-skating or try to do a somersault.

So, in net, something is being conserved. And you can think of that "something" as the area being swept out by the parts per unit time (only works if all motions are in a plane).

As long, as there is no external torque - which means nothing external attempts to twist your group of particles in some way, the parts in your system can collide and do whatever and you can pull at them by threads howsoever, that "something" will remain conserved.

Hope this helps.

You have successfully developed an intuition for the conservation of linear momentum. As the case of a single particle is a very specialized and ideal case (every particle can be zoomed into and resolved into multiple parts) let us only consider groups of particles:

An isolated group of particles, too, keep drifting as a whole in an average direction unless forced to change.

That is the conservation of linear momentum.

Now, forget about the net (average) direction of the whole and start looking at the parts. (i.e., go to the center of mass frame). And suppose that you have fine threads in your hands, attached to each of the parts so that they don't zoom away. (These are the internal forces). It does not matter how fast you pull at the parts or even if you don't pull at the parts at all.

Now, in general, your threads cannot change the rotations of the parts around you. You can pull them closer, making their orbits tighter, but then they will get faster, as you might have observed if you ever spin when ice-skating or try to do a somersault.

So, in net, something is being conserved. And you can think of that "something" as the area being swept out by the parts per unit time (only works if all motions are on a plane - in the general case, you can take the projection of motions on any plane and the area-rate will remain conserved, since, for a conserved vector all its projections remain conserved).

As long, as there is no external torque - which means nothing external attempts to twist your group of particles in some way, the parts in your system can collide and do whatever and you can pull at them by threads howsoever, that "something" will remain conserved.

Hope this helps.