your claim that there's a spectrum is speculative, and i'd say it's at odds with my experience. there are people who would like to be teachers, and people who wouldn't, and no amount of money is going to allow the second group to survive the first week. there's the claim that some amount of money will allow more of the first group to actually do it, but this is at odds with the fact that teachers actually get paid pretty well. when we talk about teacher pay, we're not talking about mcdonalds here -- maine pays quite poorly, and i still make more than enough to get by.
i'm not sure what your point is about the second quote. it doesn't say anyhing at all about raising salaries, except that it might not solve anything. whatever.
as to the phd scenario, you suppose several things that aren't really supported by fact -- that a phd in math will help you teach 7th greaders, that didactic skills are more important than other related skills, that the job will remain interesting to people with this sort of training, that there's consensus on how abstract thinking skills develop. i'm not sure where you're getting this from, and it would be 'nice to think about' if it were true, but i have no reason to believe that any of these things are.
finally, in re: 'a few really, really talented people,' we don't need a few really really talented people. we need a whole lot of passable people who can go to work and be teachers. the people who claim they 'can't justify doing it now', they haven't really looked at the math -- or they've committed themselves financially to a level that's well beyond that of the median american household which, again, is their prerogative. but let's not make this out like there's a huge class of noble geniuses out there just waiting to bail out american learners if only they could afford to help. it ain't like that. the right answer can't be money, because money won't solve the problem. if there's an answer, that's not it.