I still don't understand how such a tiny amount of water in the airlock could provide meaningful back-pressure compared to the hydrostatic pressure of a foot or two of wort in the fermentor. I buy commercial scale brewers having to deal with fermentor geometry issues, it just doesn't make sense for homebrewers where our fermenting beers have so little pressure regardless of fermentor (unless you are using a spunding valve).
It's odd, I agree. At first glance I was suspect as well, but I have to admit something about not using an airlock during primary fermentation helps to prevent the dreaded stall.
I know at my work, as little as 0.5 inH2O can make a difference, and that's about what an airlock would supply as far as back pressure. However, I agree that that little amount should not be felt by the yeast, which leans me towards the dissolved CO2 explanation. However, even there that little back pressure shouldn't drastically change the amount of CO2 able to stay in solution.
Honestly, I'm at a loss for an explanation. All I can point to is that every time I've done it, I haven't had a stuck fermentation, even using the notorious Dupont strain. Others have had the same experience (Drew Beechum, and that one Brulosophy experiment come to mind). It's frustrating for me as an engineer to not have a reason, but it is what it is.
TL;DR: I don't know why, I could make something science-sounding up but I have no proof, but it works for me and others.