234 JOHN H. FLAVELL
Keep track of past solution efforts, their outcomes, and the information they
yielded, using external records if it makes sense to do so (and if that perverse
‘adult will let you); actively “remember” to remember, monitor, and update
information and actively bring this information to bear on the problem. If
children and other fallible organisms (¢g,, adults) could somehow be induced to
produce and react appropriately to the functional equivalents of such impere
tives and questions, I believe they would be far better problem solvers. Resnick
and Glasor soem to make a similar point: “A job now ahead is to devise means of
instructing people in the processes we have hypothesized as general to problem
solution, and to evaluate the effects of such instruction across a variety of task
environments [page 229].”
Finally, some priceless suggestions to Resnick and Glaser about the conduct of
future research. Nothing is more conducive to a self-righteous sense of scientific
altruism than to make sage, ex cathedra pronouncements as to how others
should do their research; it also takes remarkably little effort and “intelligence.”
First, follow your own prescriptions about how to solve problems. Make your
research goals as clear and specific as possible and try to make each experiment
tell you something you really wanted to know concerning these goals. 1 think
some of the research moves reported have been at least a bit tangential to the
main objective.
Second, don’t be quite so exclusively oriented toward external behaviors and
external environments. Try to find out what is running through the child’s mind
as he or she wends his or her way through the task. I have found that one can
sometimes get good leads simply by asking them what they are trying to do at
various points. Interrogation procedures are likelier to be productive with older
‘than with younger subjects, of course,
Third, don’t overestimate the correspondence between your information:
processing type flow diagrams (e.g., Fig. 1) and cognitive reality. While I realize
you are well aware of the dangers of doing this, models do have a way of taking
fon an air of reality through sheer use and familiarity. I suspect that a lot of
hhuman thought, even in problem-solving situations, may be erratic and inconsis
tent in direction, subject to multiply embedded interruptions and detours, and
generally replete with vague, difficult-to-model ideas.
Fourth, consider trying to devise problem situations which are still more
naturalistic, even more “ethological-ecological” than the school-type tasks you
have been working with. For example, children are constantly solving soca
interactional as well as nonsocial problems; indeed, some of your subjects were
apparently trying to solve a social problem when you thought they were trying
to solve @ nonsocial one (page 224). Similarly—my idée fixe again—it would be
interesting to see how children would handle problems when given free accesso
external stores of relevant information. One possibility that combines your
inclinations with mine would be to study the way children do homework
problems. The investigator's role in relation to the child in such a study would of