Is Direct Drive Really Better?

Zaikesman,
I think I understand your point. Recreation of the original performance requires retracing the original steps at the original pace. Or in other words the duplicate must be exactly symmetrical to the original in order to precisely mimic it in reverse -- mistakes included.
The concept is simple enough but saying it so that others can understand my meaning isn't quite so simple.

MJ and V: Obviously we can't achieve a correct 1:1 correspondence, or probably even seek one realistically speaking. (Although if that were the goal, then you could make the argument that going with a Technics SP which used the same motor/controller as a Matsushita disk-cutter would be a reasonable strategy. However, lacquers aren't the same as vinyl, cutter-heads aren't the same as cartridges, and of course inscribing a groove isn't the same as following one.) The real point may be that, even were we able to achieve "perfect" playback with zero speed distortion (which we can't), if the record weren't cut the same way, you'd still wind up with speed distortion. Boiled down: You'll always have speed distortion. But we already knew that.

As to Viridian's observation, if our theoretical turntable suffers less dynamic drag on transients than the cutting lathe (maybe not as unreasonable an assumption as it might seem, given the presumably much greater frictional resistance encountered by a cutting-head in inscribing the groove as opposed to a stylus merely reading it), then pitches will go slightly sharp on transients. If the turntable suffers more dynamic drag than the lathe on the other hand, then transient pitches will go slightly flat. All in all though, I don't think it's completely unreasonable to speculate that these speed distortions at the mastering and playback ends might indeed have some rough natural correspondence which tends to mitigate the problem more than exacerbate it.

Expounding a bit: Viridian, your last question I think may really concern phase. You might be thinking of "flanging" as an analogy, or the way loudspeakers behave regarding crossovers and multiple drivers or reflective room surfaces. If so, that would not apply to this situation, it seems to me -- that sort of reinforcement/cancellation phenomana requires mixing of more than one sound source, which I think the encode/decode of vinyl playback doesn't qualify as. You are right that pitches are affected, and the resulting pitch distortions will derive from addition and subtraction, but I don't believe there is any constructive/destructive reinforcement going on as I understand the concept. For instance, to "flange" as they originally did in the recording studio, you need two reels of tape both with the same signal recorded on them, played back in synchrony, which is then modified by applying friction via the engineer's hands to the reels' flanges. But if you recorded one reel of tape while altering the pitch by applying drag to the flange, and then played that pitch-altered reel back while doing the same thing again, the "flanging" effect would not be the result, just further pitch distortion.

This is good!