Why FA+FB, rather than say FA+2FB? A partial answer is that if it would turn out that the latter formula gives us better predictions about results of experiments, then it would suggest that some sort of fundamental asymmetry is making things difficult for us. Maybe the result of the experiment depends on who's performing it. This would pretty much make it impossible to do science. Maybe there's something fundamentally different about the direction in space in which Bob is pushing, which makes his push "count" twice as much. This would contradict the principle of isotropy of space.
Ye old tug pushing a barge against/with the flow of water in a river comes to mind of an anisotropic space. Pushing in either direction will of course give the barge the same velocity, or acceleration, the case may be, in the water "ether", but certainly not with respect to the distant shoreline.
Being in a gravitational field is another situation, where if one did not know they were in such as field (ie you had your back to the planet or star ), engaging your rockets to or away from the gravitating body would give the same acceleration with respect to the "space ether" but not wrt to the distant stars.
In fact, in both situations, even by not providing a force by tugging or using rocket propulsion, one would notice a drift wrt the distance shore or distant stars.
Of course we know how to deal with such situations, by determining the flow of the river or by calculating the gravitational attraction of the planet or star, and thereby break the problem up into, what we consider to be, its separate calculable physical constituent parts.
The "ether" problem was done away with earlier in the century by Michel-Morrison ( isotropic space ) and Einstein (there is no either ).
And we are left with what we have now, which seems to work, so far so good,