>at which point the force of attraction would be 0 and the marble's potential energy would be zero That's wrong. If you actually stick to newtonian physics (be

>at which point the force of attraction would be 0 and the marble's potential energy would be zero That's wrong. If you actually stick to newtonian physics (because the picture changes completely in GR), then you can define the potential arbitrarily at infinite distance. Since a body with finite density will see a finite variation in gravitational potential, you could just define the potential at infinity such that it is zero at its lowest point by adding a constant. That's because forces in physics arise from the potential field's gradient - not the field itself. So any constant will just disappear in the derivative. Now, if you go to GR, the story becomes a lot more complicated, since gravitational potential energy isn't even well defined.