Thought about it some more ... realized that the voltage at either end of the secondary winding would be referenced to chassis ground (i. e. the negative battery terminal). I (for no good reason) assumed 40 kV across the secondary winding, but realized it would be split, with one side 20 kV about ground, and one side 20 kV below. One side would be positive, and one would be negative, w.r.t. ground, and therefore current flow would be in the same direction, which means the two plugs would experience opposite current flow (one ground electrode to centre electrode, and the other opposite). Please see my rough sketch.
View attachment 38354
I realize that the voltage would likely not be a 50-50 split - the plug wires have a significant resistance, and the coil is typically close to one cylinder and farther from the other, so a voltage divider situation applies. Plus the plug on the compression side takes a fair bit more voltage to fire.
But regardless, I now understand why in a waste-spark system, the current flows oppositely through the two spark plugs.
I have one more question; I assume that the electron flow (i. e. the opposite of conventional current) carries material away from the one electrode toward the other. That is, if the electron current is from the centre electrode to the ground electrode, there will be wear on the centre electrode (and build-up on the ground electrode?), and vice versa. This makes me think of the peaks and valleys on old breaker points. Now I'm dating myself!