See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/243680991

Learning communities in classrooms: A reconceptualization of educational

practice

Article · January 1999

CITATIONS READS

424 1,193

2 authors, including:

Kate Bielaczyc
Clark University
37 PUBLICATIONS   2,973 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Knowledge Building View project

All content following this page was uploaded by Kate Bielaczyc on 29 May 2017.

The user has requested enhancement of the downloaded file.

 Learni ng Communi ti es i n Cl assrooms:
A Reconceptual i zati on of Educati onal Practi ce

Katerine Bielaczyc
Boston College
OISE, University of Toronto

Allan Collins
Northwestern University
Boston College

From C. M . Reigeluth (Ed.): Instructional design theories and models, Vol. II. M ahw ah N J:
Law rence Erlbaum A ssociates.
 I ntroducti on

In recent years in A merica there has developed a “ learning-communities” approach to

education. In a learning community the goal is to advance the collective know ledge and
in that w ay to support the grow th of individual know ledge (Scardamalia & Bereiter,
1994). The defining quality of a learning community is that there is a culture of learning,
in w hich everyone is involved in a collective effort of understanding.

There are four characteristics that such a culture must have: (1) diversity of expertise
among its members, w ho are valued for their contributions and given support to
develop, (2) a shared objective of continually advancing the collective know ledge and
skills, (3) an emphasis on learning how to learn, and (4) mechanisms for sharing w hat is
learned. If a learning community is presented w ith a problem, then the learning
community can bring its collective know ledge to bear on the problem. It is not
necessary that each member assimilate everything that the community know s, but each
should know w ho w ithin the community has relevant expertise to address any
problem. This is a radical departure from the traditional view of schooling, w ith its
emphasis on individual know ledge and performance, and the expectation that students
w ill acquire the same body of know ledge at the same time.

Why Learni ng Communi ti es?

A s the w orld becomes more complex, students find themselves unprepared for the
challenges, both personal and social. The new demands that society is placing on young
people are reflected in a w ide variety of reports on education, such as the U. S.
Department of Labor’s SCA N S report (1991) and a recent book by M urnane and Levy
(1996), w hich address the question of w hat skills and know ledge w ill be needed for
w ork in the tw enty-first century. To summarize their findings, students need to be able
to direct their ow n learning, w ork w ith and listen to others, and develop w ays of
dealing w ith complex issues and problems that require different kinds of expertise.
These, for the most part. are not skills that are currently taught in schools.

So w hy should w e redesign education around learning communities? There are at least

three arguments as to w hy it w ould be good to do so:

Social-constructivist argument . The "social-constructivist" view of education,

characteristic of Dew ey and Vygotsky, holds that the theory of individual learning,
w hich pervades schools, is flaw ed. The constructivist view is that people learn best, not
by assimilating w hat they are told, but rather by a know ledge-construction process. In
order for individuals to learn how to construct know ledge, it is necessary that the
process be modeled and supported in the surrounding community. This is w hat occurs
in a learning community.

Bielaczyc and Collins p. 2

Learning-to-learn argument. Frank Smith (1988) argues that children w ill learn to read
and w rite if the people they admire read and w rite. That is, they w ill w ant to join the
"literacy club" and w ill w ork hard to become members. Brow n, Ellery, and Campione
(1998) argue that there has been a change in the demand on schools, tow ard a goal of
producing expert learners or "intelligent novices." This change has been brought on by
(1) increasing know ledge, such that no one can absorb in school everything they w ill
need to know in life, and by (2) the changing demands of w ork, w here technology can
carry out low -level tasks, requiring w orkers w ho can think abstractly and learn new
skills. So given that w e w ant people w ho know how to learn, it follow s from Smith's
argument that children w ill learn to be learners by joining a "learning club."

M ulti-cultural argument . The w orld is becoming more closely integrated through the
advent of new communication technologies, and societies are becoming increasingly
diverse through mixing of people from different cultures. This requires people to
interact and w ork w ith people from different backgrounds. To prepare people to live
and w ork amid such cultural diversity, schooling needs to construct a learning
environment that fosters students’ abilities to w ork and learn w ith other people. Each
person’s contributions must be respected, and the community must synthesize diverse
view s. This is the type of learning environment that a learning-communities approach
promotes.

In summary, the learning-communities approach addresses the needs for students to

deal w ith complex issues, figure things out for themselves, communicate and w ork w ith
people from diverse backgrounds and view s, and share w hat they learn w ith others.
Therefore educational researchers in A merica have begun to experiment w ith different
models of learning communities to determine w hich w ays of organizing learning
communities are most effective (Brow n & Campione, 1994; Collins & Bielaczy c, 1997;
Lampert, Rittenhouse, & Crumbaugh, 1996; Rogoff, 1994; Scardamalia & Bereiter, 1994;
Wineburg & Grossman, in press).

A Framew ork f or V i ew i ng Learni ng Communi ti es

The learning-communities approach raises a number of issues about the design of

learning environments. We treat each issue as a dimension, along w hich w e w ill
contrast a learning-communities approach w ith the approach of teachers w ho
emphasize the individual mind and how it develops. Classrooms have changed over
the years to involve more social interaction, but classrooms organized as learning
communities still differ from most classrooms along these dimensions. These eight
dimensions provide a framew ork w e w ill use to examine three examples of classroom-
based learning communities described in the next section.

Goals of the community: In a learning-communities approach the goal is to foster a

culture of learning, w here both individuals and the community as a w hole are learning
how to learn. Further, members of the community share their individual efforts

Bielaczyc and Collins p. 3

tow ards a deeper understanding of the subject matter under study. Students learn to
synthesize multiple perspectives, to solve problems in a variety of w ays, and to use
each other's diverse know ledge and skills as resources to collaboratively solve problems
and advance their understanding. The intent is for members to come to respect and
value differences w ithin the community. In contrast, most classrooms tend to foster a
culture in w hich students are expected to acquire the same body of know ledge at the
same time. Rather than an emphasis on diverse expertise and problem solving, there
tends to be an emphasis on conformity and on learning particular subjects.

Learning activities Because the goals focus on fostering a culture of learning, the
activities of learning communities must provide a means for (1) both individual
development and collaborative construction of know ledge, (2) sharing know ledge and
skills among members of the community, and (3) making learning processes visible and
articulated. A learning-communities approach tends to use a variety of learning
activities, including individual and group research; class discussions; cross-age
tutoring; w orking together to create artifacts or presentations that make public both
w hat is learned and w ays of learning; and collaborative problem solving w here
students take on particular roles tow ard a common end.

It should be noted that the learning activities described in a learning-communities

approach and those found in most classrooms may share some similarities. H ow ever,
because the learning activities are used tow ard different ends, differences arise. For
instance, social learning techniques such as cooperative learning and collaborative
learning (Cohen, 1985; Damon & Phelps, 1989; Slavin, 1986) can be used to support a
learning community’s goals, but they can equally w ell support more traditional
learning aimed at inculcating particular know ledge among students. Brow n and
Campione (1996) contrast the learning activities in most classrooms w ith those of
learning communities in terms of tw o other factors: that the activities in a learning
community operate as a system and their underlying objectives are articulated:
There are by now many procedures available that w ere designed to foster thinking.
These procedures are part of the teacher's tool box. But the procedures are
understood as unrelated tools, not as systems of interdependent activities....
Teachers may, for example, decide to include forms of cooperative learning, the use
of long-term projects, a w riter's w orkbench approach, etc. The problem w e see is
that such an approach ignores the potential pow er of creating a classroom system
of activities that mutually influence and reinforce each other.... There is a purpose
for every activity, and nothing exists w ithout a purpose. A ll members of the
community -- students, teachers, parents, and researchers alike -- should be aw are
of this (Brow n & Campione, 1996, pp. 292, 314).

Teacher roles and power relationships: In a learning-communities approach, the teacher

takes on roles of organizing and facilitating student-directed activities, w hereas in most
classrooms the teacher tends to direct the activities. The pow er relationships shift as
students become responsible for their ow n learning and the learning of others. Students
also develop w ays to assess their ow n progress and w ork w ith others to assess the

Bielaczyc and Collins p. 4

community's progress. In contrast, in most classrooms the teacher is the authority,
determining w hat is studied and assessing the quality of students' w ork.

Centrality/peripherality and identity: The degree to w hich people play a central role and
are respected by other members of a community determines their sense of identity
(Lave & Wenger, 1991). In a learning-communities approach the central roles are those
that most directly contribute to the collective activities and know ledge of the
community. H ow ever, opportunities exist for all community members to participate to
w hatever extent is possible and students w orking in peripheral roles are also valued for
their contributions. Centrality and peripherality are context-dependent. Certain
students may have more to contribute at a given time, so a student’s centrality can
change over time. A s members of a learning community take on different roles and
pursue individual interests tow ard common goals, students develop individual
expertise and identities. Because diversity is important, an atmosphere in w hich
students respect each other’s differences needs to develop.

In contrast, in most classrooms students w ork on the same things and are all expected
to reach a base level of understanding. Students tend to form their identity through
being measured or by measuring themselves against this base level. Centrality tends to
mean those w ho meet and exceed this base level -- those w ho "get it." Schofield (1995)
notes the benefits of such centrality in that teachers typically spend most of their time
interacting w ith the better students. Students on the periphery are then those students
needing remediation and extra help -- those that "aren't quite there yet," w hich
diminishes their value to others in the classroom.

In a learning-communities approach there is also the notion of a community identity. By

w orking tow ard common goals and developing a collective aw areness of the expertise
available among the members of the community, a sense of "w ho w e are" develops. In
the absence of a learning culture that builds a collective understanding and view s its
members as learning resources, most classrooms fail to develop a strong sense of
community identity.

Resources: Both a learning-communities approach and many classrooms use resources

outside of the classroom, including disciplinary experts, telementors, the Web, etc.
H ow ever, in learning communities both the content learned and the processes of
learning from the outside resources are shared more among members of the community
and become part of the collective understanding. A further distinction betw een
learning communities and most classrooms is that in learning communities, both the
members themselves and the collective know ledge and skills of the community are
view ed as important resources.

Discourse: In the learning-communities approach the language for describing ideas and
practices in the community emerges through interaction w ith different know ledge
sources and through co-construction and negotiation among the members of the

Bielaczyc and Collins p. 5

community. A lso, learning communities develop a common language for more than
just content know ledge and skills. The community develops w ays to articulate learning
processes, plans, goals, assumptions, etc. In contrast, in most classrooms the teacher
and texts tend to promulgate the formal language to be learned.

Discourse functions in a learning community as a medium for formulating and

exchanging ideas. It serves to motivate the research and reflection in the community by
raising new questions and hypotheses, w hich give rise to further research and
understanding (Bereiter & Scardamalia, 1993). Students are expected to provide
feedback to each other, and are supported in doing so. In contrast, in most classrooms
communication occurs principally betw een the teacher and students. The discourse
functions as a medium for conveying know ledge to students and asking students
questions to test their know ledge (Schofield, 1995).

Knowledge: In learning communities the development of both diverse individual

expertise and collective know ledge is emphasized. In order for students to develop
expertise, they must develop an in-depth understanding about the topics that they
investigate. Rich subject matter is important. The topics are not randomly chosen, but
rather the depth centers on key principles or ideas in a domain that are generative for
understanding a broad array of topics. There is also a circular grow th of know ledge,
w herein discussion w ithin the community about w hat individuals have learned leads
individuals to seek out further know ledge that they then share w ith the community.
Thus, there is an interplay betw een the grow th of collective know ledge and of
individual know ledge, w ith each supporting the other. In most classrooms the goals
tend tow ards covering all the topics in the curriculum (breadth over depth) and for
everyone to learn the same thing.

Products: Dw eck (1986) has show n how students w ho adopt performance goals put
their energy into looking good and tend to give up w hen they fail. But those students
w ho adopt learning goals learn more from their mistakes and pursue learning in the
face of failure. One concern is that an emphasis on products may lead students to adopt
performance goals, and focus on production values rather than meaningful learning.
But, as Bruner (1996) points out, a culminating event or product can act to focus the
energy of the entire class on a joint effort, w hich helps to build community.

In a learning-communities approach, members w ork together to produce artifacts or

performances that can be used by the community to further their understanding. There
is sustained inquiry and development of products over months. In contrast, most
classrooms tend tow ard individual or small group assignments w ith little sharing or
collective products. Usually w ork is produced in short periods of time.

A nal ysi s of Learni ng-Communi ty Cl assrooms

Bielaczyc and Collins p. 6

To give a picture of w hat a learning-communities approach implies for schooling, w e
describe three exemplary cases of learning communities that have been set up in
A merican classrooms. A fter briefly describing each of the three cases, w e w ill compare
them w ith respect to the eight issues raised in the Introduction. Then w e w ill attempt to
extract general principles for the design of learning communities.

Scardamal i a and Berei ter’s K now l edge-Bui l di ng Cl assrooms

Scardamalia and Bereiter (1991, 1994) have developed a model they call Know ledge-
Building Communities. CSILE (Computer Supported Intentional Learning
Environments) is the name commonly applied to this model, although strictly speaking
it is the name of the computer softw are they developed, w hich is used in classrooms
that may or may not have adopted the pedagogical model. The essential idea is that
students w ork together to make sense of the w orld around them and w ork tow ards
advancing their ow n state of know ledge and that of the class.

The model involves students investigating problems in different subject areas over a
period of w eeks or months. A s students w ork, they enter their ideas and research
findings as notes in an on-line know ledge base. The softw are (originally called CSILE,
now in a new version called Know ledge Forum) supports students in constructing their
notes through features such as theory-building scaffolds (e.g. "M y Theory," "I N eed to
Understand” ) or debate scaffolds (e.g. “ Evidence For” ). Students can read through the
know ledge base adding text, graphics, questions, links to other notes, and comments on
each other’s w ork. When someone has commented on another student’s w ork, the
system automatically notifies them about it.

The central activity of the community is contributing to the communal know ledge base.
Contributions to CSILE can take the form of (a) individual notes, in w hich students state
problems, advance initial theories, summarize w hat needs to be understood in order to
progress on a problem or to improve their theories, provide a draw ing or diagram, etc.,
(b) views, in w hich students or teachers create graphical organizations of related notes,
(c) build-ons, w hich allow students to connect new notes to existing notes, and (d) “ Rise
Above It “ notes, w hich synthesize notes in the know ledge base. A ny of these kinds of
contributions can be jointly authored.

When students feel a note makes an important contribution to the collective know ledge
base, they can propose the note for publication. A n editorial group and the teacher then
decide w hether to publish the note. A t the end of the school year the class may decide
on a selection of notes to remain in the know ledge base for classes that come after them.
The goal is to engage students in progressive know ledge building, w here they
continually develop their understanding through problem identification, research, and
community discourse. The emphasis is on progress tow ard collective goals of
understanding, rather than individual learning and performance.

Bielaczyc and Collins p. 7

Brow n and Campi one’s FCL Cl assrooms

Brow n and Campione (1994, 1996; Brow n, 1992) have developed a model they call
Fostering a Community of Learners (FCL) for grades 1-8. The model provides w hat is
termed a "developmental corridor," w here the learning community extends not only
horizontally across a classroom, but vertically across grades. This makes it possible for
learning topics to be revisited at increasing levels of disciplinary sophistication. We w ill
focus here on communities in classrooms, rather than across grades.

The FCL approach promotes a diversity of interests and talents, in order to enrich the
know ledge base of the classroom community as a w hole. The current focus of FCL
classrooms is on the subject areas of biology and ecology, w ith central topics such as
endangered species and food chains and w ebs. There is an overall structure of students
(1) carrying out research on the central topics in small groups w here each student
specializes in a particular subtopic area, (2) sharing w hat they learn w ith other students
in their research group and in other groups, and (3) preparing for and participating in
some “ consequential task” (Scardamalia, Bereiter & Fillion, 1981) that requires students
to combine their individual learning, so that all members in the group come to a deeper
understanding of the main topic and subtopics. Teachers orchestrate students’ w ork,
and support students w hen they need help.

There are roughly three research cycles per year. A cycle begins w ith a set of shared
materials meant to build a common know ledge base. Students then break into research
groups that focus on a specific research topic related to the central topic. For example,
if the class is studying food chains, then the class may break into five or six research
groups that each focus on a specific aspect of food chains, such as photosynthesis,
consumers, energy exchange, etc. Students research their subtopic as a group and
individually, w ith individuals "majoring" by follow ing their ow n research agendas
w ithin the limits of the subtopic. Students also engage in "crosstalk," talking across
subtopic groups to explain, ask questions, and refine their understanding. The research
activities include reciprocal teaching (Palincsar & Brow n, 1984), guided w riting and
composing, consultation w ith subject matter experts outside the classroom, and cross-
age tutoring. In the final part of the cycle, a member from each of the subtopic groups
come together to form a “ jigsaw ” group (A ronson, 1978) in order to share learning on
the various subtopics and to w ork together on some consequential task. Thus, in the
jigsaw , all pieces of the puzzl e come together to form a complete understanding.

The consequential task requires the different subtopics to be used together to form a
common product or common understanding. The choice of consequential tasks is
ideally made by the teacher and students together. In some cases the consequential task
might be a bulletin board display, the design of a bio-park to protect an endangered
species, a presentation to the community at large, or in some cases a test of students’
know ledge. These tasks “ bring the research cycle to an end, force students to share

Bielaczyc and Collins p. 8

know ledge across groups, and act as occasions for exhibition and reflection” (Brow n &
Campione, 1996, p. 303).

Bielaczyc and Collins p. 9

Lampert’s M athemati cs Cl assroom

Lampert (1986, 1990; Lampert, Rittenhouse & Crumbaugh, 1996) taught mathematics to
fifth grade students for a number of years, w here she developed an approach to
teaching that reflected her view of an idealized mathematics community. The class
usually starts w ith a problem posed to the students, w hich they w ork on alone or in
groups, developing their solutions in notebooks that retain all their w ork during the
year. A fter 15-20 minutes of w ork the class as a w hole discusses the problem and
various possible solutions. Lampert encourages students to discuss different ideas and
solutions, so that they develop a deep understanding of the mathematical principles
underlying their w ork.

Lampert chooses problems that foster deep inquiry and mathematical argumentation by
students. Students are encouraged to present different ideas and methods, and to
discuss w hich are correct and w hy. There is an emphasis on how to resolve
mathematical arguments by appeal to logic and evidence. Participating in the
mathematical discussions, learning how to make mathematical arguments, and learning
the language of mathematics (terms such as “ conjectures” and “ commutativity” ) are the
central activities in the classroom.

Lampert orchestrates the discussion and picks up on certain ideas, revoicing them so
that everybody can understand. She is very much in control and uses various
techniques to make sure students participate in the discussion. She opportunistically
follow s the ideas the students suggest in order to relate them to important
mathematical ideas. The students are on an equal footing in the discussions, offering
their ideas and discussing other students’ ideas and arguments. She carefully
orchestrates the discussion to maximize the participation among the students. H er
technique of asking students to explain other student’s ideas is particularly effective in
making them listen to and respect other students. The discussion involves students in a
w ay that fosters understanding of the ideas and principles that the class is developing.

Compari son of the Three Cases

We w ill compare the three cases in terms of the eight issues outlined in the
Introduction. By looking at the similarities and differences betw een these exemplary
cases of classroom learning communities, w e can more clearly see the essential
characteristics of learning communities.

Goals: A ll three cases foster a culture of learning, w here students come to see
themselves as contributors to their ow n learning and that of the community. The goals
in all three cases are consonant w ith the learning-community goals described in the
Introduction. This includes the goals of students learning how to (1) learn and reflect
on their learning, (2) become critical thinkers w ho know how to frame questions and

Bielaczyc and Collins p. 10

develop a deep understanding of the issues they investigate, and (3) share their learning
and w ork w ith others in the community as resources.

Learning activities: While the three cases share the same goals, they involve different
learning activities and types of support. In CSILE, students investigate problems and
develop theories, contribute w ritten and graphic descriptions about w hat they are
learning to the collective know ledge base, and comment on and respond to other
students’ contributions. This tends to be accompanied by oral discussion. These
learning activities are guided by the softw are through the different scaffolds (e.g., “ M y
Theory,” “ What I Learned” ) and by interactions w ith other students around their ideas.
In FCL, the learning activities center on research, sharing know ledge, and producing
joint products. The different activities (e.g., reciprocal teaching, guided w riting, cross-
age tutoring) each have a structure, w hich serves to guide the students. Lampert’s
classroom focuses on problem solving and mathematical argumentation as learning
activities. Lampert provides guidance throughout the process, by posing provocative
problems and directing discussion tow ard important mathematical issues.

Teacher roles and power relationships: In all three classrooms, the teacher takes the role of
a facilitator. The learning activities and lines of inquiry tend to be driven by student
questions and interests. In Lampert’s classroom, by comparison w ith CSILE and FCL,
the teacher is much more in control of w hat students are doing. By leading students in
w hole class discussion, Lampert supports students in coming to reason and argue
mathematically in the w ays that she has mastered. The teacher’s role in CSILE is not
prescribed and can vary w idely, depending on the teacher’s know ledge and orientation
tow ard a particular unit. A s Scardamalia, Bereiter, and Lamon (1994, p. 209) point out
"CSILE opens up a significant channel for communication in the classroom that is not
mediated through the teacher." FCL falls somew here in betw een. Certain activities,
such as the benchmark lessons, are closely guided by the teacher or by guest experts.
H ow ever, students also direct the community's learning, as in one case w here students
became interested in the question of w hether mosquitoes could transmit A IDS (Brow n
& Campione, 1994).

In both CSILE and FCL classrooms a student may have more expertise in a particular
area than the teacher, changing the typical student-teacher pow er relationship that
exists in most classrooms. Ideally, students benefit from the know ledge of their
teachers and available experts, but at the same time go beyond such know ledge
w henever feasible (Scardamalia & Bereiter, 1991).

Centrality/peripherality and identity: A s stated earlier, the central roles are those that
directly contribute to the collective activities and know ledge of the community. A ll
three cases also provide a means for all community members to participate in
peripheral roles to w hatever extent is possible. In CSILE, students can still participate
in the community w hile engaging in peripheral activities, such as reading notes in the
know ledge base and making comments on other students' notes. Students' roles begin

Bielaczyc and Collins p. 11

to shift as they create notes of their ow n, either as an individual or in a group. Students
begin to play more central roles as they have their notes published or linked to other
students' notes. Centrality increases further w hen a student's notes are chosen to be
saved for future generations of CSILE users. In FCL classrooms, students move into
central roles w hen their expertise is required w ithin their jigsaw group, w hen creating
products, or w hen their individual expertise is called on by the w hole community. In
Lampert’s classrooms, students move into central roles w hen they contribute to the
mathematical discoveries and arguments that the class is engaged in.

Individual identity is developed in CSILE w hen one's contribution is recognized: other

students read, comment, or make links to one's note; one's note is published, or the
w hole class identifies one's note as w orthy of being saved for later generations. In FCL
classrooms, individual identity is developed through being responsible for others'
learning in contexts, such as jigsaw groups or cross-age teaching. Identity is also
developed in contributing one's individual expertise and skills to the public, collective
understanding. In Lampert's classroom, individual identity is developed as one's
contributions to the class discussions are commented on or built upon, and from the
teacher revoicing w hat one has said.

A major effort in these classrooms goes into ensuring that all students are making
contributions to the community and that their contributions are valued by other
students. This is accomplished in different w ays. CSILE encourages students to
investigate issues they care about and so they develop diverse expertise, w hich serves
to make their contributions valuable to other students. In addition, students are taught
how to make effective comments on each other's notes, so that their criticism is
constructive (Woodruff & Brett, 1993). By setting up jigsaw and reciprocal teaching
groups, and jointly-produced consequential tasks, FCL fosters diverse expertise and
interdependence, w hich encourages students to rely on and value other students’ w ork.
Students also learn how to give helpful guidance to each other. For example, before
students w ork w ith each other in cross-age teaching, tutors are trained in tutoring
methods (Brow n & Campione, 1996). Lampert also uses a variety of stratagems to build
a community w here all the students respect each other. This includes revoicing w hat
students say so that their ideas are understood by other students, and asking students
to explain w hat other students are saying before they disagree, so that they must listen
carefully to other students during the discussion. In addition, she gives feedback to
students w hen they do not listen to and respect other students.

In all three cases a community identity is also developed. In CSILE this comes from all
students building a common know ledge base, and from students w orking together to
examine their collective know ledge base and to decide w hat should be passed on to
future generations. In FCL classrooms, community identity comes from participating in
the creation of joint products and through experiencing how the subgroups of the class
w ork together. In Lampert’s classroom this comes from the w hole class w orking

Bielaczyc and Collins p. 12

together in depth on math problems, engaging in mathematical argumentation w ith
each other, and coming to a common understanding of mathematical principles.

Resources: In all three cases, the students come to view each other as legitimate
resources for learning. A nother resource that is common to all three cases is the
collective know ledge and skills that the community is developing. The teacher is also a
resource, although in Lampert's classrooms the teacher is a more central resource than
in the CSILE and FCL classrooms. Lampert provides a deep understanding of
mathematical issues and skills in mathematical argumentation, w hich she uses in
selecting problems and guiding students’ discussion.

FCL and CSILE classrooms bring in resources from outside the classroom in w hatever
w ays possible. Because students are investigating questions in depth, they often come
upon issues that are beyond the classroom community’s expertise to answ er. Therefore,
FCL encourages students to find resource people that can help w ith the questions they
are investigating, to learn from students in cross-age tutoring interactions, to
communicate w ith telementors about issues, and to find information on the Web. FCL's
benchmark lessons often bring in experts from the community. This resource
contributes not only to the students' learning, but also to the teacher's professional
development. "With increasing exposure to the visitor's lessons, the classroom teachers
learn more about the content area and increasingly come to take over responsibility for
benchmark lessons" (Brow n & Campione, 1996, p. 299).

Scardamalia and Bereiter (1994, 1996) have begun to develop a new type of relationship
betw een students and outside resources in the CSILE model. They envision a
know ledge building society w here both adults and students w ork in a common
know ledge base. To illustrate how they see such a society functioning, suppose that
fifth and sixth graders are w orking on the problem of “ H ow does electricity w ork?” and
that museum curators are w orking on an exhibit about electricity. Then the curators
might investigate the student’s know ledge base to see w hat interests them or confuses
them about electricity. Similarly, the students might follow the ongoing development of
the exhibit discussions in the know ledge base, making comments on w hat they find
interesting and w hat they don’t understand. Further, scientists involved in the exhibit
might contribute to know ledge building by providing useful comments on or links
betw een students' notes. Students might also learn from observing the thinking
processes exhibited in adult discourse in the know ledge base.

Discourse: A ll three cases encourage public discussion of issues among students. This is
one of the central w ays that a learning community expands its know ledge.

The CSILE system emphasizes a discourse of formulating problems, constructing

theories, and bringing questions, comments, and new information to bear on them. The
labels on the notes guide the students in making contributions (e.g., "M y Theory", "I
N eed to Know "). In producing notes about w hat they have learned, students are

Bielaczyc and Collins p. 13

encouraged to use both w ritten and graphical discourse. Communication in CSILE is
asynchronous, w hich may allow students to express themselves more clearly, since they
are not able to rely on immediate context in making their points clear (Bereiter &
Scardamalia, 1993). A n interesting aspect about the use of w ritten discourse in CSILE is
that the basis for determining a student's status tends to be different from that in face-
to-face settings. Because a great deal of classroom interchange is in w ritten form,
qualities such as clarity, persuasiveness, and inventiveness come to dominate over
forcefulness, looks, and popularity, w hich usually determine status in face-to-face
settings. In CSILE, students are also encouraged to critique each other’s w ork by
reading and commenting on their notes. Students are taught how to use these different
genres effectively. For example, in order to get students not to make superficial
comments, students are taught to identify both the strengths and w eaknesses in other
student’s w ork.

Similarly, FCL emphasizes students engaging in questioning, explaining, and

constructive discussion of issues, both in w ritten and oral modes. Students w ork in
groups to formulate questions that they pursue in their research. They w rite up their
findings for the other students and these w ritings form the basis for reciprocal teaching
sessions w ith the other students in the class. In the crosstalk and jigsaw sessions
students discuss w hat they have learned and make comparisons w ith w hat others have
learned. Brow n and Campione (1996, pp. 305, 319) claim that “ dialogue provides the
format for novices to adopt the discourse structure, goals, values, and belief systems of
scientific practice. Over time the community of learners adopts a common voice and
common know ledge base, a shared system of meaning, beliefs, and activity that is as
often implicit as it is explicit... Ideas are seeded in discussion and migrate throughout
the community.”

Lampert emphasizes mathematical argumentation in her classroom, in order to teach

students how to reason mathematically. One challenge to this approach is that students
find this emphasis on argumentation at odds w ith their desire to get along w ith other
students and not to criticize them (Lampert, Rittenhouse & Crumbaugh, 1996). A s
Lampert and her colleagues point out, it is very difficult for adults to separate criticism
of ideas from criticism of the person, so it is not surprising that students find
argumentation uncomfortable. With the teacher’s guidance and w ith experience and
modeling by other students over time, students learn how to argue about ideas
effectively w ithout personalizing the criticisms.

Knowledge: The three cases differ as to how much they encourage people to develop
common know ledge as opposed to diverse know ledge. Lampert strives for common
know ledge among all participants. Lampert encourages students to help each other in
their groups, and in the group discussion of problems she w orks tow ard the goal of
everyone understanding the ideas discussed. With CSILE, on the other hand, students
are encouraged to go off in depth in their ow n direction to develop expertise. Some
students may focus on one aspect of a topic and other students on another aspect.

Bielaczyc and Collins p. 14

Through reading each other’s notes and producing a database for w hich they are
collectively responsible, they form a common understanding.

FCL supports diverse expertise in that it has each research group study a different topic,
and each member of the group become expert in a different aspect of the topic. But FCL
also strives for common know ledge through different mechanisms, such as crosstalk,
classroom discussions, consequential tasks, and students sharing their expertise. The
activities are also structured to ensure that students know w ho has w hat expertise. So, if
a question arises w ith respect to a problem they are w orking on, they know w ho to ask
for help.

A ll three cases help students to develop meta-know ledge about both the subject matter
and the learning processes they are engaged in. In CSILE the development of higher-
order view s of the community’s w ork, together w ith the "What We H ave Learned"
notes and "Rise A bove It" notes, encourages students to engage in a type of “ meta-
discourse.” That is, students engage in discourse about the discourse in the know ledge
base w here they reflect on their ow n and on the community's progress in
understanding. FCL grow s out of a long line of research on metacognition and the
development of activities that foster reflective learning practices (see Brow n &
Campione, 1996). The learning activities in FCL classrooms are meant to create an
atmosphere of reflection on learning and encourage articulation of learning processes.
For example, in completing a public performance, students reflect on w hat they have
learned and set priorities -- "What is important to know ? What is important to teach?
What of our new found know ledge do w e display?" (Brow n & Campione, 1996, p. 295).
In Lampert’s classrooms, students are frequently asked to explain w hat another student
is thinking or to articulate the idea one is arguing against prior to making a proposal of
one’s ow n. Such activities require students to examine the ideas of the community, to
compare proposals, and talk about know ledge and understanding.

Products: A s mentioned earlier, a concern w ith students developing products is that

such an activity may lead students to adopt performance goals and focus on production
values rather than meaningful learning. Scardamalia and Bereiter (1991) have alw ays
been concerned w ith keeping students focused on learning goals, but they have
incorporated into CSILE the publishing of notes and the handing on of published notes
to future classes. This acts as a kind of culminating product that can serve to focus the
community’s w ork. FCL emphasizes consequential tasks, w here students w ork on a
project that requires all the diverse expertise that different students have acquired. FCL
also has students give presentations and exhibitions for a w ide variety of audiences,
such as parents, community members, and younger students. So there is an emphasis
on products that bring together different strands of the students’ w ork and that reach
outside the classroom. In Lampert's classes, students produce a journal w ith their
individual w ork, how ever the class does not produce any physical collective products.

Bielaczyc and Collins p. 15

One final point concerns the idea of the collective know ledge of the community as a
product. While in all three cases the members of the community share a certain level of
common understanding, differences exist betw een the three cases in terms of the level
to w hich the collective know ledge is objectified. In CSILE there is a complete w ritten
database, w hich embodies the community’s know ledge and forms a repository that can
be inspected and reflected upon. In FCL the collective know ledge is expressed orally,
w hich produces no repository, and also through artifacts and public performances,
w hich provide records of the collective know ledge. In Lampert’s classes, all discussions
are oral, leaving no tangible record of the shared, collective know ledge. The only
w ritten record of w hat w as learned is in the individual students’ journals, w hich do not
reflect the collective understanding and are not shared resources.

Pri nci pl es f or the D esi gn of Ef f ecti ve Learni ng Communi ti es

In considering these different cases, w e have tried to encapsulate w hat w e have learned
from them into a set of principles for the design of effective learning communities.

Communi ty-Grow th Pri nci pl e: The overall goal of the community should be to expand
the community’s know ledge and skills. To maximize its learning, the community needs
to take advantage of the know ledge of all its members and w hat they learn. The goal is
for individuals to constantly gain new know ledge and to share among themselves. By
pooling know ledge from each individual, the community can expand its collective
know ledge.

Emergent-Goal s Pri nci pl e: The learning goals of the community should be co-
constructed w ith the students and come out of the activities and questions that arise, as
students carry out their investigations. The teacher must be sensitive to the needs,
interests, and abilities of the individual students. The goals therefore should reflect
w hat the students know and help them build on both their strengths and w eaknesses.
The students spaw n goals of their ow n as the learning community evolves and they
take over more of the w ork of the community. In this w ay collective goals are emergent.

A rti cul ati on-of -Goal s Pri nci pl e: The teacher and students should articulate the goals
they are pursuing and the terms by w hich they w ill judge their success. This allow s all
members of the community to have a clear idea of the goals and of the criteria by w hich
they can tell if they have reached their goals. A ll the students should develop the ability
to judge if the goals have been met.

M etacogni ti ve Pri nci pl e: M etacognition involves (1) monitoring one’s thinking

processes, (2) being aw are of w hat one know s and doesn’t know , and (3) reflecting on
w hat one has learned (Brow n, Bransford, Ferrara & Campione, 1983). In terms of
monitoring, the community should keep asking itself w hat its goals are and if w hat it is
doing w ill help it reach them. The community should also try to identify at regular
intervals w hat it know s and does not know . Finally, in terms of reflection, the class can

Bielaczyc and Collins p. 16

look back at w hat it has done (for example, at its products and performances), and
evaluate w hat w as learned and how w ell they did.

Beyond-the-Bounds Pri nci pl e: The community should attempt to go beyond the

know ledge and skills w ithin the community and the resources easily available to them.
They should try to make sense of things for themselves and w elcome new approaches
and challenges. They do not w ant simply to regurgitate w hat they find in their
resources. They should seek ideas that challenge w hat they believe, by soliciting diverse
opinions and view s on a topic. They should not just try to find support for their current
beliefs.

Respect-f or-Others Pri nci pl e: Students need to learn to respect other students’
contributions and differences, and to feel safe in speaking up and giving their ow n
ideas. The more everyone is heard, the more sources of know ledge there are for
expanding the community’s know ledge. When only one or tw o students are heard, then
the learning of the community is limited to w hat those few students provide and
develop. The rules for respect should be clearly enforced and articulated.

Fai l ure-Saf e Pri nci pl e: We often learn from failures, so that to the degree a learning
community accepts failures and does not try to assess blame, then it fosters a more
experimental approach that allow s failures to occur as the community learns. Often the
failures w ill be collective failures. There must be a sense that failure is okay, and that
taking risks and an experimental approach w ill lead to more learning. Reflection
w ithout blame, can help to ensure that the community learns from its mistakes.

Structural -D ependence Pri nci pl e: The community should be organized such that
students are dependent on other students’ contributions in some w ay. It is important to
have a valid reason for students to w ork together that makes sense to the students, such
as a common task that requires their joint effort. If students are w orking on a task and
they need another student’s help, it makes that student important to them. This fosters
both respect for the other student and that student’s self esteem. This validation of
differences is lost in traditional schooling, because it tries to ensure that everyone is
alw ays learning the same thing.

D epth-over-Breadth Pri nci pl e: The students have sufficient time to investigate topics in
enough depth that they gain real expertise in the topics. This is necessary to foster a
sense of their ow n expertise and to support meaningful discourse among the students.
Ideally the depth should center on important ideas that are generative for
understanding a broad array of topics. It is critical for students to get beyond
memorizi ng know ledge and procedures, in order for students to care about w hat they
are learning and to develop a sense of how to learn.

D i verse-Ex perti se Pri nci pl e: Students develop the areas in w hich they are most
interested and capable, w ith the responsibility that they share their expertise w ith the

Bielaczyc and Collins p. 17

other students and the teacher. By developing diverse expertise, the community can
deal w ith problems and issues that are too difficult for any individual to handle. A
learning community continually discusses ideas and examines its progress in
understanding, so that w hat an individual learns is not just from the activities that they
themselves carry out, but from all the activities that different members of the learning
community engage in. This is fundamentally different from “ learning by doing,” as it is
commonly understood. To most people, the phrase implies that individuals learn from
w hat they themselves do. What occurs in a learning community is “ collective learning
by doing,” w here participants also learn from w hat others do. What is learned by
individuals is w hat “ gets into the air” of the community.

M ul ti pl e-Ways-to Parti ci pate Pri nci pl e: In order to advance its collective

understanding, a learning community has a variety of jobs it needs done. Students may
be more or less interested and adept at different activities, so that there should be a
range of activities in w hich they participate. The different activities should support the
multiple learning goals of the community, such as formulating questions, gathering
know ledge, sharing know ledge w ithin the community, presenting their know ledge to
the outside w orld, and reflecting on w hat they have learned. Students w ill take on
different roles in the various activities, such as researcher, expert, co-investigator,
monitor, interpreter, moderator, etc. The community needs to value all roles and their
contributions, and not regard some roles as inferior.

Shari ng Pri nci pl e: There needs to be a mechanism w hereby know ledge and skills
gained by different individuals is shared throughout the community, so that each
student is both a learner and contributor to the community know ledge. Unless
something enters into the collective know ledge of the community, it does not serve the
common good. M any communities lack adequate w ays of sharing know ledge and
practices, so that members often end up doing a poor job, because they did not know
that some other member had the expertise they could benefit from. So it is important to
share know ledge, not simply so that everyone profits from w hat each individual learns,
but also so individuals know w ho to go to w hen a difficult problem arises.

N egoti ati on Pri nci pl e: Ideas, theories, procedures, etc. are constructed by a negotiation
process among members of the community, and arguments among them are resolved
by logic and evidence (Collins, 1998). A rgumentation is necessary for finding better
solutions or understandings, because the learning community needs to identify errors
and misconceptions that inevitably arise. But students usually do not like to participate
in argumentation, since it makes them uncomfortable to criticize others (Lampert,
Rittenhouse & Crumbaugh, 1996). There needs to be w ays to model and to coach
participants on how to critique other people’s ideas w ithout personalizing the critique,
by trying to separate the ideas from the person. There is a variety of stratagems for
depersonalizi ng critiques (e.g., focusing on the strong aspects of w ork as w ell as the
w eak aspects; couching comments in terms of w hat to change, rather than w hat is

Bielaczyc and Collins p. 18

w rong, etc.) that students need to develop to make interactions in a learning
community effective.

Qual i ty-of -Products Pri nci pl e: The quality of the products produced by the
community should be valued both by the community itself and by outsiders to the
community. In particular the students need to think highly of the goals they are
pursuing, and the know ledge and products they are producing. There must be
standards that the community agrees upon as to w hat makes for good quality w ork,
and these standards must be tested against the outside w orld. One w ay to do that is to
bring in different audiences to judge the w ork, such as parents, community members,
and other students.

Concl usi on

The idea of learning communities in classrooms w ill grow as w e try to address the
needs of being able to reason through complex issues and problems, direct one’s ow n
learning, communicate and w ork w ith people from diverse backgrounds and view s,
and share w hat one learns w ith others.

A key idea in the learning-communities approach is to advance the collective

know ledge of the community, and in that w ay to help individual students learn. This is
directly opposed to the approaches found in most schools, w here learning is view ed as
an individual pursuit and the goal is to transmit the textbook’s and teacher’s know ledge
to students. The culture of schools all too often discourages sharing of know ledge, by
inhibiting students talking, w orking on problems or projects together, and sharing or
discussing their ideas. Testing and grading are administered individually. When taking
tests, students are prevented from relying onother resources, such as students, books, or
computers. The w hole approach is aimed at ensuring that students have all the
know ledge in their heads that is included in the curriculum. Thus the learning-
community approach is a radical departure from the theory of learning and know ledge
underlying schooling.

The development of the learning-communities approach reflects a more w idespread

change in our understanding of education. Educational thinking in A merica has long
been dominated by the psychologist's point of view , starting w ith Thorndike and
coming dow n to us through behaviorists, such as Skinner and Gagné, and more
recently cognitive psychologists, such as Bruner and John A nderson. This dominance is
beginning to change as anthropological and Vygotskian influences have begun to
“ enculturate” educational thinking. The cultural view emphasizes that learning is a
social and cultural enterprise, expressed in Smith's (1988) terms as, "We learn from the
company w e keep." Learning in this view is coming to belong to a community of
practice (Brow n, Collins, & Duguid, 1989; Lave & Wenger, 1991). The idea of learning
communities is the culmination of this view as applied to schooling.

Bielaczyc and Collins p. 19

The learning-communities approach also fits w ith the grow ing emphasis on lifelong
learning. While in this paper w e have emphasized the role of learning communities in
classrooms, this view of learning can naturally be extended beyond the classroom w alls.
In addition to children, there is every reason for a learning community to include
parents and members of the w ider society. With the addition of computer netw orks it
becomes possible to include scientists and other professionals, as w ell as students
around the globe, in communities that are trying to understand and deal w ith social
and political ideas and issues (Collins & Bielaczy c, 1997; Scardamalia & Bereiter, 1996).
Children w ill benefit greatly from interacting w ith more adults in learning situations.
Thus, the learning-communities approach offers a w ay to end the isolation of children
in schools and integrate their learning w ith that of the w ider society.

A ck now l edgments
This paper w as partially funded by a fellow ship from the M cDonnell Foundation to the
first author. We thank M arlene Scardamalia, Carl Bereiter, M agdalene Lampert, N ick
H addad, A nn Kaufman, M ichael Reynolds, and Charles Reigeluth for their suggested
revisions to a previous draft.

Ref erences
A ronson, E. (1978). The jigsaw classroom. Beverly H ills, CA : Sage.
Bereiter, C., & Scardamalia, M . (1993). Surpassing ourselves: A n Inquiry into the nature and implications
of expertise. La Salle IL: Open Court.
Brow n, A . L. (1992). Design experiments: Theoretical and methodological challenges in creating complex
interventions. Journal of the Learning Sciences, 2 (2), 141-178.
Brow n, A . L., Bransford, J. D., Ferrara, R. A ., & Campione, J. C. (1983). Learning, remembering and
understanding. In J. H . Flavell & E. M . M arkman (Eds.), Handbook of child psychology (4th ed.), Cognitive
development (vol. 3, pp. 77-166). N ew York: Wiley.
Brow n, A . L., & Campione, J. C. (1994). Guided discovery in a community of learners. In K. M cGilly
(Ed.), Classroom lessons: Integrating cognitive theory and classroom practice (pp. 229-270). Cambridge, M A :
M IT Press/ Bradford Books.
Brow n, A ., & Campione, J. (1996). Psychological theory and the design of innovative learning
environments: On procedures, principles, and systems. In L. Schauble & R. Glaser (Eds.)Innovations in
learning: New environments for education (pp. 289-325). M ahw ah N J: Law rence Erlbaum A ssociates.
Brow n, A ., Ellery, S. & Campione, J. (1998) Creating zones of proximal development electronically. In J. G.
Greeno and S. Goldman (Eds.) Thinking practices. M ahw ah N J: Erlbaum.
Brow n, J. S., Collins, A ., & Duguid, P. (1989). Situated cognition and the culture of learning. Educational
Researcher, 18(1), 32-42.
Bruner, J. S. (1996).The Culture of education. Cambridge, M A : H arvard University Press.
Cohen, E.G. (1986) Designing groupwork: Strategies for the heterogeneous classroom. N ew York: Teachers
College Press.
Collins, A . (1998) Learning communities: Comments on papers by Brow n, Campione and Ellery and by
Riel.” In J. G. Greeno and S. Goldman (Eds.) Thinking practices. M ahw ah N J: Erlbaum.
Collins, A . & Bielaczyc, K. (1997) Dreams of technology-supported learning communities. In Proceedings of
the Sixth International Conference on Computer-Assisted Instruction. Taipei, Taiw an.
Damon, W. & Phelps, E. (1989). Critical distinctions among three forms of peer leaning. International
Journal of Educational Research, 13(1), 9-19.
Dw eck, C. (1986) M otivational processes affecting leaning. American Psychologist, 41(10), 1040-1048.
Lampert, M . (1986). Know ing, doing, and teaching multiplication. Cognition and Instruction, 3( 4), 305-342.

Bielaczyc and Collins p. 20

 Lampert M . (1990) When the problem is not the question and the solution is not the answ er:
M athematical know ing and teaching. American Educational Research Journal. 27(1): 29-63.
Lampert, M ., Rittenhouse, P. & Crumbaugh, C. (1996) A greeing to disagree: Developing sociable
mathematical discourse. In D. Olson & N . Torrance (Eds.) Handbook of Education and Human Development.
Oxford: Blackw ell’s Press, pp. 731-764.
Lave, J. & Wenger, E. (1991) Situated learning: Legitimate peripheral participation. N ew York: Cambridge
University Press.
M urnane, R. J. & Levy, F. (1996) Teaching the new basic skills. N ew York: Free Press.
Ong, W. J. (1982) Orality and literacy: The technologizing of the word. London: Routledge.
Palincsar, A . S., & Brow n, A . L. (1984). Reciprocal teaching of comprehension-fostering and monitoring
activities. Cognition and Instruction, 1(2), 117-175.
Postman, N . (1982). The disappearance of childhood. N ew York: Delacorte.
Rogoff, B. (1994) Developing understanding of the idea of communities of learners. M ind, Culture, and
Activity, 1(4), 209-229.
SCA N S Commission (1991). What work requires of schools: A SCANS Report for America 2000. Washington,
DC: The Secretary's Commission on A chieving N ecessary Skills, U. S. Department of Labor.
Scardamalia, M . & Bereiter, C (1991). H igher levels of agency for children in know ledge building: A
challenge for the design of new know ledge media. Journal of the Learning Sciences., 1(1), 37-68.
Scardamalia, M . & Bereiter, C (1994). Computer support for know ledge-building communities. Journal of
the Learning Sciences., 3(3), 265-283.
Scardamalia, M . & Bereiter, C (1996) Engaging students in a know ledge society. Educational Leadership,
54(3), 6-10.
Scardamalia, M ., Bereiter, C., & Lamon, M . (1994). CSILE: Trying to bring students in w orld 3. In K.
M cGilley (Ed.) Classroom lessons: Integrating cognitive theory and classroom practice (pp. 201-228).
Cambridge M A : M IT Press.
Scardamalia, M ., Bereiter, C., & Fillion, B. (1981). Writing for results: A sourcebook of consequential
composition activities. Toronto: OISE Press/ La Salle, IL: Open Court.
Schofield, J. W. (1995). Computers and classroom culture. Cambridge, UK: Cambridge University Press.
Slavin, R. (1986). Learning together. American Educator , Summer, 6-13.
Smith, F. (1988) Joining the literacy club. Portsmouth N H : H einemann .
Wineburg, S. & Grossman, P. (in press) Phi Delta Kappan.
Woodruff, E. & Brett, C. (1993, N ovember) Fostering scholarly collaboration in young children through
the development of electronic commenting. Research in Education, N o. 50, 83-95.

Bielaczyc and Collins p. 21

View publication stats