[referring to Tripp, 21 May 99] In one sense, the above supports the comments upon which it was based. To the extent things are designed for specific functions, then we'd presume that they'd convey those things better than other things. There are other questions, though. First, what is the connection between the features/functions of a given resource (map) and the learning goals/needs of the individual? The idea that the symbols, themselves, inherently support learning (especially when the goals of its use vary) to the extent that no additional support is needed (instruction, elaboration, scaffolding, etc.), though, seems to me a big leap. If true, we'd need only to create objects--the objects themselves would be self-instructional.
Since this doesn't seem to occur with sufficient regularity to assume will occur, we employ methods that help in understanding what the objects do, mean, represent, etc. When externally imposed/induced, there is by definition a high degree of correspondence between them. When the goal is established by the individual, are resources available that address them in the kind of one-to-one sense that we have when externally imposed/induced? Sometimes, yes; other times, no. When yes, the resource basically satisfies the need. If I knew I wanted to learn things as specific as the features of Mercator maps, and I knew where they were available and were sufficiently complete to satisfy my goal, I'd be home free.
Often though, individually generated goals won't fall as neatly into simply matching a given resource to the need. (When it does, how does the individual know where it exists and/or how to locate and use it?) It will involve collecting pieces of information from different resource--resources frequently created for purpose other than how one will use them collectively with others.
Individually-generated enabling contexts (and needs/goals) aren't often so precise as the map example either. When individually based, the point Steve describes may be well into a progression of refined goal/need statements typical in problem solving (rather than single well-defined and specific goals). When individuals generate goals, the context is often very "open"--the starting points of their efforts, consequently, are often broad and not specific initially. It is the progressive refinement that needs become clearer, the subsequent selection/use of tools and resources become more specific. What if, in trying to learn about the elevation of Atlanta, Georgia, in the U.S. versus Sydney, Australia. If I encountered a Mercator representation, my needs and wants aren't specific to learning about map conventions, per se--in fact, it might be that what is most efficient is simply to find a map where the desired information is represented. Would I recognize that I have the wrong type of map? I don't need to learn about Mercator conventions at all (for that need, in any case)--I need to learn what will satisfy the need--a relief map. How do open learning systems support the individual learning needs/intents of very different users?
The issue of internal-external is not strictly one of epistemological belief, bias, or the locus of knowing. Its one of practical circumstances as well. Learning needs are initiated and pursued even when there is no teacher and no formal learning setting. For example, perhaps you receive a phone call during which you are told of a relative's medical condition with potential genetic etiology. You determine to learn more about the condition and the potential risks to you, your children, etc. You might gather blood type information (same/different factors in probability of hosting the disease) and compare to probability tables, do similar background on your immediate family members, etc. That the goal has been individually generated has more to do with the circumstances surrounding its initiation than with the intellectual argument about whose learning/need it is in a more constructivist--it happens spontaneously in the absence of teacher/instructional goals and the like.
Individually generated enabling contexts also emerge in formal learning (classroom) settings such as when a particular interest emerges, perhaps related to a something triggered during a classroom discussion. Maybe the classroom discussion focuses on diversity--an exchange between students from different backgrounds emerges. You become personally interested not so much in the particular issue discussed, but the demographics in the neighborhood where you grew up. Maybe you wondered if the racial/ethnic distribution in your hometown was comparable to that discussed in class, you want to re-examine some of the events from your own experience from the perspectives presented in the discussion, and so on. Again, the enabling context is devoid of external goal structures--they're determined by you for purposes that are your own.
So, what resources do you use? Where are they located, how should their contents/features be examined, etc., to address your need? The point is that many resources designed to promote understanding use features created for different purposes. We do it all the time when we seek information on systems like the Web and digital encyclopedias. Picasso didn't paint in order for people like me to trace the evolution of cubism in his work--that's what we want to do. In Ken Burns' Civil War epic (apologies to those outside the US for this example), entire archives of historic photographs, letters, and government documents were woven together to tell a story of the American Civil War than none of the resource told/could tell individually. This, in a sense, is what individuals are increasingly being asked to do (or choosing to do)--take personal perspectives on problems and needs by angling differently at existing artifacts--digital and others. You'd need to be pretty fortunate to consistently have available the precise resources needed to address the enormous range of particularized needs that exist or will exist.