the brain, old and young, and education, college and otherwise

"it takes some rote learning just to get to the point where one is capable of understanding an idea, much less incorporating it into one's own thinking"

Yep, no argument at all.  And I like very much the implicit suggestion that one might equate "facts (and procedures and algorithms ...)" with the hoped for products of "rote learning."  That's not at all to denigrate either; both are essential components of learning processes.  Without them, one can't, as you say, have understandings with which alternate understandings can conflict. 

So, indeed, the problem becomes how to integrate "facts (and procedures and algorithms") into a coherent/practical pedagogical strategy.  I very much wish I had a solution to that problem; my own teaching would be more successful if I did.  Maybe though there's progress in more clearly defining the problem along the lines we have.  Among other things, it at least suggests some general principles that one might start with:

1. Don't mislead students (or be misled oneself) into believing that the primary classroom task is the acqusition of "facts (and procedures and algorithms)."  Be clear that the primary task for students (and teachers) is to acquire greater skill in themselves challenging their own understandings, wherever and however achieved, and creating new ones.
2. Don't mislead students (or be misled oneself) into believing that existing "facts (and procedures and algorithms)" will suffice to effectively challenge existing understandings and create new ones.  New observations that don't yet "make sense" to students (including apparently arbitrary "facts (and procedures and algorithms)" are essential grist for the primary task.
3. In the service of both 1 and 2, choose and make available new observations (including "facts" (and procedures and algorithms) in the context of the primary task.  Be sure students are aware of the problems they have trouble coping with in terms of existing understandings and so experience new observations (including "facts" etc) as tools potentially relevant to their own development of new understandings, rather than as either answers or things to master in their own right.  "facts etc" are actually more effectively assimilated this way than they are by "rote" learning. 

Hmmm.  I've never  tried to make this quite so explicit.  Having done so, its clearer to me how/why my own teaching has been less successful than it might be.  I'm thinking particularly of a class where in fact students needed to acquire some mastery of a computing language and never quite did it to the level needed to work effectively with it.  The problem, now that I think back on it, was that the problems I invited the students to wrestle with didn't immediately motivate acquiring mastery of the language and the time/effort needed to acquire mastery of the language was seen by students as detracting from their wrestling with the problems in other (for them easier? more obvious?) ways.  To be more successful, I would have needed either to scale back  the problems or change to a more readily accessible language (or both).

The bottom line?  Let's not give up either growth by challenge to existing understandings or acquiring/using new information that doesn't yet make sense.  The challenge, it seems to me, is to effectively integrate the two, so that "facts (and procedures and algorithms)" are acquired not "by rote" or as an end in themselves but rather as an integral and natural part of the process of achieving new understandings.  We know it can be done.  That is, of course, how we all learn to speak, to ride a bike, to kayak, and ... to become both effective scientists and effective teachers ourselves (to whatever extent we do).   My guess is that when it doesn't work in a particular classroom context its not because we need not to move to one extreme or the other but rather because we're already to one side or the other of an optimal integration and so need to rethink what we're trying to do.