you know what, I've been thinking about this, and I think you are misunderstanding how it it works, even tho it may work as explained, i think you have the wrong idea in your head.
(http://www.stanford.edu/class/me220/data/lectures/lect08/fig4_2.gif)
here is a diagram of some parallel plates with an electric field in between them, q+ and q-, and as you see there is a gradient for the voltage perpendicular to the arrows, while the field lines are in the direction of the arrows, and only in one single line straight down the center is there "zero" volts, and everywhere else there is a + or - voltage.
so 20,000+ volts, and 20,000- volts will not cancle out for zero volts, that just isnt how it works, there is a voltage gradient and only at the center between the plates will you have zero volts.
picture this, you have a rock 20 feet above the ground, and you drop it, when it hits the ground, you are at zero
now picture you dig a hole 20 feet down, and you have the same rock 20 feet above the ground, and drop it, it will fall 20 feet to ground level, and then 20 feet down the hole, (-20 feet), but the rock fell a total of 40 feet
so 20+ volts PLUS 20- volts is 40 volts, not zero volts. the voltage does not cancel, it adds.
there is only one spot on the voltage gradient where you have zero volts, and that would be the same as in my explanation, of the ground level.
If you want to talk about electrical stress, then sure, the oxygen is being pulled one way, and the hydrogen the other way, but the voltage does not cancel.