What I was trying to express is that IMHO the excuses you mentioned earlier ("yeah, but I saw he was beginning to attack") are *never* valid, because the result of incorrect timing will always speak for itself; this just becomes particularly clear during freestyle practice because then nage can no longer hide behind what his partner is supposed to do (according to nage) - but it's really just the same during formal practice with "preset" attacks and defenses.
Whether you do an exercise slowly and carefully, or quickly and vigorously, you can't do it right unless you learn to "tune in" to your partner and move *together*, instead of just reacting to the other, or trying to act preemptively. Even during an exercise where you know exactly what your partner is going to do and how you should react, what really decides the result (and in a real conflict, possibly life or death) is whether you can grasp that exact moment when the attack starts and harmonize with it - it's the difference between blending with uke or blocking/trying to catch up with him. Even when you know something *will* be coming, and *how* it will be coming, you can't say exactly *when* it'll come, after all. Whether it's just practice or "for real" is irrelevant.
Well, that's more or less how sensei usually explains it - or at least that's how I interpret it; I certainly can't and won't pretend to be an authority on the matter. And I don't pretend to be making sense here either - it's something you can feel, but not analyze, so explaining it is difficult. I think...
Anyway, trying to get my train of thought back on track: IMO timing (=blending, seizing the right moment, ...) is an essential factor, always and everywhere. Which method for learning it is better is something I'll let everyone try and decide for themselves. So no, I don't think the sense of timing in an environment that has assigned roles can be said to be "false" - it's either there, or it isn't.
Unless my definition of "timing" is different from yours of course ("every discussion, if continued long enough, ends in semantics
") - but I don't really think so.