Bodong Chen

Crisscross Landscapes

Notes: Higher Levels of Agency for Children in Knowledge Building



Citekey: @scardamalia1991

Scardamalia, M., & Bereiter, C. (1991). Higher Levels of Agency for Children in Knowledge Building: A Challenge for the Design of New Knowledge Media. Journal of the Learning Sciences, 1(1), 37–68. doi:10.1207/s15327809jls0101_3


This paper is particularly useful: * It connects agency with ZPD, and makes it clear KB turns more agency to students and lets students decide ZPD. * It reports a set of interesting studies on student questioning, especially the one examining whether students can pick promising questions from those generated by themselves. * Although it doesn’t explicitly mention epistemology, it “makes clear that we need an epistemology more in keeping with group processes and with students’ abilities to focus on promising ideas. Thus the world 3 epistemology elaborated in the following article.” (quote from Marlene)

All points are useful for my thesis.


Abstract: Although adults and children both have zones of proximal development in which more knowledgeable others play essential roles, there is a difference in executive control that is most salient in question-answer dialogue. Adult learners typically ask questions based on their perceived knowledge needs, whereas with school children, questions are typically asked by the teacher, based on the teacher’s perception of the child’s needs. Evidence shows that children can produce and recognize educationally productive questions and can adapt them to their knowledge needs. The challenge is to design environments in which students can use such questions to guide their building of knowledge, thus assuming a higher level of agency in learning. Computer Supported Intentional Learning Environments (CSILE), a computer- supported knowledge medium designed to support intentional learning, is described, with illustrations of children’s use of it in cooperative knowledge building.

The difference between the Teacher B and Teacher C models becomes, in these terms, a difference in the control structure of activities in the zone of proximal development.


Second is growth toward children’s takeover of executive functions -what amounts to control of the zone of proximal development itself. The Teacher B model emphasizes “growing into” more mature competence; the Teacher C model adds a deliberate effort to promote “growing out o f ‘ dependency.

Much of what we learn as adults is through interaction with peers, where the zone of proximal development is defined by activities the participants can handle cooperatively better than they can alone. In such cases, control of the zone is mutual. When we seek the help of someone more knowledge- able, it is usually on our own terms. Although dependent on the other’s competence, we retain the regulating role. We know what we want.

We question the maturity of adults who surrender control of their zones of proximal development to a mentor or cult leader. But with children such surrender is taken for granted.


The two principles illustrated here that we sought to generalize through CSILE were (a) providing external supports for higher level cognitive processes and (b) making metacognitive activity, which is nor- mally hidden and private, overt and a subject for public consideration. This attention to both the private and the public aspects of cognitive activity remains characteristic.


We are focusing on question asking in this article because, as traditionally carried out in schools, it epitomizes the teacher’s control over the zone of proximal development and the unidirectional flow of information. Teachers ask the questions, thereby determining the goals of learning activities, and they evaluate the answers, thereby determining what is to be accepted as knowledge (Sinclair & Coulthard, 1975)

The typical question-answer discourse of schooling provides no means for the child to grow into the competencies that underlie the asking of good questions. Reciprocal teaching (Brown & Palincsar, 1989), by contrast, creates a zone of proximal development in which children can grow into the asking of text-based questions.

Collins and Stevens (1982) found the questions of skilled inquiry teachers to perform functions such as the following: * Ask for sufficient factors. * Ask for similarities in factors between similar cases. * Ask for the formulation of an alternative rule. * Ask for predictions using hypothetical cases based on irrelevant or insufficient factors. * Ask questions that trap students into revealing their misconceptions. * Ask series of questions that lead students into a contradiction.

We have seen that question asking is a means by which teachers retain control over the zone of proximal development. As Dillon (1982) put it: “The dynamics of a question-answer relationship establish respondents in a passive, reactive role, fostering dependency and removing a sense of responsibility, initiative, and a kind of energy” (p. 160). Letting the children ask the questions might therefore seem to be an obvious way of giving them ownership over their zones of proximal development.



This was what we call the knowledge-based condition, because the students’ prior knowledge was the only basis on which their questions could be formed.

Questions from the knowledge- based condition averaged significantly higher than those from the text- based condition, with a mean of 2.52 versus a mean of 1.58.

More important than mean item quality, however, is the actual number of items generated in a class that are judged to be educationally worthwhile as guides to study and inquiry. In the text-based condition, 4 of the original set of 104 questions (4%) were rated 3 or higher, indicating a judgment that obtaining an answer to the question would produce “an advance in conceptual understanding.” By contrast, 38 questions (46%) produced in the knowledge-based condition were rated at this level.


Lack of domain knowledge, therefore, does not seem to have hampered students in generating questions. Indications of the students’ limited knowledge about fossil fuels do appear, however, in the kinds of questions asked.

More than half of the questions from the fossil fuels condition represent what we may term basic information questions and uneducated guesses.

The remaining 47 questions we may call wonderment questions. They reflect curiosity, puzzlement, skepticism, or a knowledge-based specula- tion, in contrast to a groping for basic orienting information.

Considering jointly the questioning carried out on the topics of fossil fuels and endangered species, it appears that the students appropriately adjust the kinds of questions they ask, according to their level of kn~wledge.~ If they already have a basic understanding of a topic, as with endangered species, they ask questions that have potential to extend their conceptual understanding. If they lack elementary knowledge, as with fossil fuels, they ask questions of a basic “What is it? Where is it? What causes it?” type. But the asking of such essential questions does not preclude their also asking wonderment questions of a challenging and stimulating kind. These findings thus indicate considerablepotential for children to take over greater control of their zones of proximal development.


it becomes important to find out how successfully students can select ques- tions to pursue.

On the topic of endangered species, where students already had some basic understanding, there was a generally high level of agreement among children and adults about what questions are of greater or lesser educational value. Combined student ratings correlated .89 with experimenter ratings of the 22 questions on this topic. The median correlation of individual student ratings with experimenter ratings was .60 (n = 25 students) but with a range of .37 to .90. For fossil fuels questions, however, the pooled- ratings correlation was only 3 6 over 22 questions. The median correlation for 24 students was .40 with a range of .14 to .66. In rating the fossil fuels questions, students and adults seemed to agree in identifying what we have called wonderment questions as being questions whose investigation will lead to significant conceptual advance, but in judging basic information questions, children seemed to have difficulty distinguishing naive questions that are merely ill conceived from naive questions that raise important conceptual issues. This was evident from interviews discussed in the full report of this research (Scardamalia & Bereiter, 1990).

What seems to be required, in order for children successfully to assume executive control in their own zones of proximal development in a classroom setting, is a social process that allows the wisest judgments to work their way forward.


They also indicate two needs, which we are trying to meet in the redesign of CSILE: More ways to link students with common interests. Support for pursuit of convergence and coherence. Correspondingly, CSILE in its current version lends itself nicely to divergent processes but lacks support for convergence. Plans for the new architecture include special knowledge- building environments. One will provide tools for describing and simulating processes, specifically supporting the kind of constructive interaction that Miyake (1986) found to go on when pairs of people are trying to work out an understanding of a process. Another environment will focus on explanatory coherence, providing an interface for organizing networks of facts needing explaining and hypotheses explaining the facts, based on Thagard’s (1989) computer implementation of his theory of explanatory coherence.

Prospects for a Higher Level of Student Agency in Knowledge Building

There is then no question of displacing the teacher. Teachers can be vital participants in the community knowledge-building process, but they need no longer constrain it to what they comfortably know. The zone of proximal development comes to be defined by the knowledge-building activities that the group, as a collective, can profitably engage in. Individual differences, so often a grave problem in school, can fall into place naturally, if the point becomes that everyone should have something to contribute rather than that everyone should try to excel.

We think this model has not had a chance in education because the structure of the classroom has forced teachers to assume control of knowledge building, whether as didactic conveyors of information or as gentle managers of the intellectual life of the classroom. New knowledge media open up new possibilities-both for democratizing the knowledge resources available to schools and for moving knowledge building to the center of the life of the classroom community.