All about 'persons' and 'people'

My sense here is that, despite a distinction that has been promoted by many usage mavens in the past, people today prevails as the all-purpose plural of person. In his Modern American Usage, Wilson Follett does make a strong argument for the retention of 'persons' in some situations:

A magazine that calls itself People yet chiefly reports on the few who are known to the many ignores the difference between people and persons. When we say persons we are thinking of ones, individuals with identities; whereas when we say people we should have in mind a very large group, an indefinite and anonymous mass. . .Resting its full weight on its misunderstanding, the magazine in question relies on the names of its subjects being well known.

Follett here confidently asserts a useful and material distinction--one that, by the way, I acknowledge and would not like to neglect. But the question of whether this distinction always has to be reflected in contrasting linguistic forms seems a different issue. Despite its logical basis, persons sounds stilted--even "wrong"--to me, and I would not recommend its use in any context. People are still capable of understanding the distinction, I would argue; and, although the loss of one of the contrasting linguistic forms could dilute the dichotomy considerably, a precious lot of people would have to switch back to the dying form persons to render it viable once again.

In the years since Follett wrote his entry, of course, critical opinion has shifted further in the direction of sanctioning people in all cases; while Strunk & White (1959) concur with Follett's advice, Burchfield and many other later critics (as well as the major style manuals--see Chicago and AP) do not. Also, in case you were wondering about the Google count, it currently shows 473 million to 7 billion hits, in favor of people. And, despite my prescriptivist leanings, I hesitate to say that 7 billion people could be wrong on this one.