a dominant and pervasive assumption that schools, at least at the middle school level, should provide all students with equal opportunities for learning valued competencies and fairly position them vis-a-vis each other in situations of formal assessment of these competencies. Traditional classroom practices avoid this collision by prescribing uniformity of learning and assessment experiences, but uniform assessment practices in places like Ms. Leoni's project-based, collaborative learning classroom would likely create even more perilous collisions. Where then does this leave the designer of educational experiences?
With collaboration in complex, temporally- distributed projects comes spontaneous divisions of labor. And with these divisions of labor come the possibility of distinct forms of individual competence. In schools, this presents creative challenges to educators, two of which I sketch briefly here.16 First, educators are challenged to design and implement new activity structures for classrooms (e.g. reciprocal teaching, jigsaw) that take advantage of emergent, practical divisions of labor and learning so that students can participate with each other in ways that distribute what they've learned to others.17 Second, educators are also challenged to implement new assessment practices that recognize and register productive diversities of competence and, more generally, become appropriate instruments for the assessment of learning in classrooms that are increasingly less amenable to traditional uniform testing procedures. ( Hall, Knudsen, & Greeno, 1995/96) In summary, an implication for CSCL designers is that for every technical design endeavor, putting systems into use will involve a accompanying set of design problems that lie "beyond the interface" ( Bannon & Bødker, 1990).
A century ago Durkheim forcefully argued that spontaneous, relatively durable divisions of labor would create a cohesive, diverse and unrepressive society composed of well-developed individuals. Current interest in distributed cognition and in collaborative, project-based educational initiatives invite a re-inspection of this hypothesis, if not at Durkheim's macro- societal level but (pace Strauss) within and between concrete institutional settings. Based on this comparative case analysis, a tentative response to Durkheim's hypothesis is that divisions of labor are in themselves neither favorable or unfavorable for people who am working and learning in institutional settings. The value of divisions of labor for learners depends on how people divided come together, how different types of work divided among people are made visible and accounted for, and on how divisions of labor enable or constrain changing forms of participation and the development of new competencies.
I conclude with two directions for extending this analysis that should widen its scope and deepen the account of how divisions of labor appear, stabilize, and change.
Widening the analytic focus to include the entire arc of project work in both settings. This widening will bring other types of work that are divided among persons under analytic scrutiny, including mathematical practices, writing, document management, and graphic design. Computers support many of these types of work as well and widening the scope will provide a more complete view of how these technologies enable and constrain project development and how they shape different kinds of learning.
Exploring the interactional, negotiated organization of divisions of labor. While this paper has found similarities in divisions of labor in two settings, it has not analyzed the "articulation work" involved in dividing the labor among persons or fitting it together once divided. Much of this fitting together work will be understood through analyses of talk-in- interaction. (e.g. Drew & Heritage, 1993; Goodwin & Heritage, 1990)) However, in design projects, fitting together also has a decidedly material character and further analyses will look closely at how persons transform representations from one media state to another (e.g. from paper to digital form or from a collection of disparate papers to a single poster) and at what people learn in the process of making these transformations ( Hutchins, 1995). Documenting both these types of articulation work should____________________