Scott, so called "accelerating through" has essentially nothing to do with increasing dwell time, although many believe this is what gives them more cue speed. You are right to point out that this is wrong. But there is a more subtle effect that accrues from it, which can yield more cue speed.
The force we apply to accelerate the cue evolves during a stroke. It goes from zero to some peak, then back down to zero, and then becomes negative (deceleration) for the most part, to slow the cue down and bring it to a stop. If you imagine a graph of that force plotted against time or distance, different shapes yield different cue speeds at impact.
Generally speaking, you get more cue speed when the peak force occurs later in the stroke,ie, the curve is skewed to the right...up to a point. This is naturally accompanied by a longer follow through, though not necessarily so. In that sense, loosely speaking, a longer follow through results in more cue speed.
This is not voodoo physics. It's just less obvious because most players, and people in general, have not been exposed to the concept of an integral.
Jim