A lot of conservation laws are derived from the lagrangian in my book. However, I fail to see why the Lagrangian incorporates action-reaction. Since it works for an arbitrary amount of particles and linear momentum can be show to be conserved from translational invariance it must do so. But deriving it from the principle of stationary action, doesn't really have any assumption that action must have a reaction as far as I can see it. What do I fail to see?