Dear Ken,
The Lagrangian schema has some advantages over the Newtonian schema, or in general over the dynamical system schema. Your essay pointed out this very well, with important insights in the problems of quantum mechanics. I can see how well this relates to your previous essays. My own view on quantum mechanics is again, in my opinion, close to yours (as in the essays on the nature of time). But instead of Newtonian vs. Lagrangian, I identify the problem as local vs. global (please see these slides of a talk I will give in few days, "Global and local aspects of causality").
The importance of the Lagrangian view is related also with the problem of singularities in general relativity, which is in fact the subject of my essay. As you know, the action principle is given by the Lagrangian density
[math]R\sqrt{-\det g}[/math]
There are singularities for which this Lagrangian density remains smooth or even analytic. It is customary to consider only R, and view the square root of the metric as auxiliary. In fact, from geometric viewpoint, the natural quantity should contain it too, since we don't integrate scalars, but densities or 4-forms. For such singularities R may be divergent, but metric's determinant vanishes and compensates this. This allows the writing of a densitized version of Einstein's equation, whose terms remain smooth at singularities. I explain this in my essay, "Did God Divide by Zero?", and references therein. So we can see that the Lagrangian view is superior in GR too, at least in the case of singularities.
Best wishes,