What you describe could well be the explanation, but it's not the same mechanism that Alexander is suggesting. You're suggesting the therapy has a dose-respons

What you describe could well be the explanation, but it's not the same mechanism that Alexander is suggesting. You're suggesting the therapy has a dose-response, and as perception of the therapy changes, patients are motivated less and so they get smaller doses. What Alexander is suggesting there is that for a given level of work that a patient put in (i.e. for a constant dose), the placebo effect has been reducing due to a "miracle effect" halo that is fading over time. I.e. he is explicitly talking about just the placebo effect here.