Foam stoppers actually do allow for gas exchange. If that wasn't the case, at a minimum I should be hearing a "pop" when pulling it out, only to find an extremely carbonated starter, and at worst, the thing would keep flying out like a plugged-up airlock.
So there's a strong indication that at the very least, it permits gas exchange at a rate equal to or greater than the rate of CO2 generation by the yeast in a typical starter. The positive pressure therefore should be pretty minute, and should impair oxygen flow no more than with a loose aluminum foil cap or even an uncovered starter. I prefer to use the foam stopper because, while bacteria may not be able to maneuver around a foil cap, fruit flies certainly can. I used to make traps for them as a kid - with beer, no less - and it quickly became obvious that I needed to tape every possible seam down, because those guys can find their way through little holes and cracks that even seem too small for them.
I'd be interested to see if a foam stopper could handle a 5Gal batch without flying out though... it would definitely demonstrate that it is more than capable of handling both the CO2 outflow and much more than an equivalent inflow for any homebrew-sized starter.
I realize though that it still leaves the question of whether oxygen can even get past the tiny amount of positive pressure created by the constant CO2 generation. If some has access to some sort of O2 meter, it seems like an easy enough experiment.
However, if you're not just being stubborn and argumentative, I highly recommend you pick up Chris White's book, "Yeast." You can probably find it for about $10, and it's definitely a worthwhile read. I don't personally feel the need to second guess him on matters of yeast propagation anymore than I would have felt the need to debates Richard Feynman on a nuclear-physics related matter - not that I am comparing the two, as it's damn near impossible to even hold a dim candle to Feynman. But this is extremely simple science for Chris White, something so fundamental and so easily demonstrated to be wrong or right, that it almost seems like a waste of time to double-check him (and all his colleagues, competitors, and contemporaries) on this particular assertion. But if you disagree, he cites plenty of references in his book if you really need to review them, and also responds to emails from anyone (though it takes a week or two) if you're still not satisfied at that point.
