Now, to return to the point that I was trying to prove erroneously.
I still think the "random surfer" model would include the concept that a link higher on a page -- or in a left nav rather than a footer -- would be more likely to be clicked and so might get more PageRank than a link not so well positioned.