beanut putter and jelly

So I believe that Egypt is a ripe proving ground for my latest audio-linguistic theory:

As some of you may know, Egyptians, especially of a lower socio-economic status, have dificulty distinguishing between the voiced and unvoiced bilabial plosive (/p/ & /b/) usually this results in many English /p/ words sounding like /kombooter/ (for computer) or /bleez/ for please.

Well, here's the interesting part (to me, at least): Since I've gotten to Egypt, I've noticed that, when I type, I occassionally transpose these two letters! They are no where near each other on the keyboard, and, although they do mildly resemble each other, it is not like the f and p, or the i and l. So where does this mistake come from?

Some proffessional linguists assure me that there is no known linguistic theory to account for this. They were perfectly willing to offer up some non-linguistic explanations (i.e. that I'm bonkers, or just plain wrong).

So here is my challenge: somebody come up with a cogent theory to explain this phenomenon, and I'll test it here in Egypt.

Any takers?