But as a thought experiment, let’s assume that Jesse Jackson had run as strongly in 1984 as he eventually would in 1988. In his first go-round, Jackson didn’t win any primaries and received 3.2 million votes. In 1988, he won 11 primaries and caucuses, receiving nearly 7 million votes.Well, no. If Jesse Jackson had won a lot more vote and delegates in 1984, he would have been taking them from someone. I'm not going to look it up, but my memory is that Walter Mondale did reasonably well with African Americans and Latino voters that year; Jackson's improvement in 1988 was probably partially a result of him winning a higher percentage of those constituencies. On the other hand, Jackson also probably won some support from relatively wealthy, highly educated liberals in 1988, the kind that Gary Hart did well with in 1984. Here's the thing: there's no reason to believe that a stronger Jackson in 1984 takes equally from both of those groups (or that he draws from nonvoters). If not, then either Hart or Mondale is helped in the battle between them in a world in which Jackson does better. And if Mondale gets stronger at all, Hart probably fizzles out entirely somewhere along the way -- and, with just a bit more leeway, if Hart does better than Mondale drops out. For example, on Super Tuesday that year, had Jackson taken just a few votes from Mondale in Georgia then Hart would have won the state, giving him a 4-1 state advantage that day instead of 3-2 -- and perhaps changing the complexion of the race enough to kill off Mondale entirely.
If Jackson had been sufficiently organized and funded in 1984, we would have had a scenario vaguely similar to what we have today: Gary Hart winning prototypical “New Democrats” (“Atari Democrats,” we called them), Mondale running strong with “traditional Democrats,” and Jackson running well with African-Americans. In that scenario, a brokered convention would have been likely.
The problem is that you have to think about these things dynamically: if one candidate improves, everyone else loses vote share, and thus doesn't get the resources going forward (money, organization, positive attention) that they earned from the vote share they actually got. It's just not reasonable to assume that a candidate will do exactly as well in May as he or she did in February; that assumes an equilibrium that just isn't seen very often in nomination fights.
In the current cycle, that means that if Rick Santorum does well enough to split delegates from here out with Mitt Romney, that probably means that Newt Gingrich gets few if any delegates going forward. And if that's the case, either Romney or Santorum would need only a narrow margin in order to win the nomination. Of course, more likely than that is that Romney (or Santorum for that matter) just wins the bulk of the remaining contests.
I think part of the problem here is that what happened with Barack Obama and Hillary Clinton -- where two very strong candidates basically had a stable contest in which demographics, and not campaign dynamics, proved important -- is misleading people. Normally, candidates who fall behind just drop out, and that's that.
*Except Trende calls it a brokered convention, which is, as I've said many times, wrong. What we won't get will be a deadlocked convention, or I'm okay with a "contested" convention, but a brokered convention requires delegates willing to be brokered, and that's just not the case under current procedures.