MIT Media Laboratory
20 Ames Street
Cambridge, MA 02139
Proceedings of the International Conference on the Learning
Association for the Advancement of Computing in Education
(accepted: March 1996; published: July 1996)
Collaborating on Constructions
The Internet as Rorschach
This paper introduces the concept of distributed constructionism, building
on previous research on constructionism and on distributed cognition. It
focuses particularly on the use of computer networks to support students
working together on design and construction activities, and it argues that
these types of activities are particularly effective in supporting the
development of knowledge-building communities. The paper describes three
main categories of distributed constructionist activities: discussing
constructions, sharing constructions, and collaborating on constructions. In
each category, it describes ongoing research projects at the MIT Media Lab
and discusses how these projects support new ways of thinking and learning.
At a recent educational-research conference, researchers were debating
various approaches for using computer networks in education. One group of
researchers talked excitedly about new ways of delivering instruction to
students. They explained how lectures by expert scientists could be beamed
down to thousands of schools. They imagined the day when personal
workstations will present problems to students, monitor student progress on
the problems, and automatically download video segments from network servers
at appropriate times during the instruction.
A second group of researchers presented a different vision for computer
networks in education. They dismissed the idea of delivering information to
students across the network. They wanted to turn the tables, putting
students in control of the information. They talked about new tools that
allow students to search through thousands of servers on the network,
locating information that they are interested in.
These two approaches are very different from one another. In one case,
information is delivered to students according to some instruction plan; in
the other case, students search for the information they need, when they
need it. But the two approaches share a common feature: they both focus on
information. Indeed, amidst all the talk about the "information
superhighway" and "information society," the idea of information is often at
the center of discussions about computer networks in education.
This paper describes a third vision of computer networks, strikingly
different from the first two. This vision puts construction (not
information) at the center of the analysis. It views computer networks not
as a channel for information distribution, but primarily as a new medium for
construction, providing new ways for students to learn through construction
activities by embedding the activities within a community.
The paper introduces a theoretical underpinning for this approach, known as
distributed constructionism, then describes a set of research projects at
the MIT Media Laboratory designed to explore the educational possibilities
for using computer networks in support of construction activities. The paper
does not go into great depth on any one project or activity; rather, it aims
to provide a framework for thinking about collaborative construction
activities and the role of computer networks in supporting those activities.
It highlights several new network-based environments that our research group
is developing to enable pre-college students to collaborate on the
construction of dynamic artifacts (such as animations and simulations).
Constructionism (Papert, 1993) is both a theory of learning and a strategy
for education. Constructionism is based on two types of "construction."
First, it asserts that learning is an active process, in which people
actively construct knowledge from their experiences in the world. People
don't get ideas; they make them. (This idea is based on the constructivist
theories of Jean Piaget.) To this, constructionism adds the idea that people
construct new knowledge with particular effectiveness when they are engaged
in constructing personally-meaningful products. They might be constructing
sand castles, poems, LEGO machines (Resnick, 1994), or computer programs
(Harel, 1991; Kafai, 1995). What's important is that they are actively
engaged in creating something that is meaningful to themselves or to others
Distributed constructionism extends constructionist theory, focusing
specifically on situations in which more than one person is involved in the
design and construction activities. It draws on recent research in
"distributed cognition" (Salomon, 1994), recognizing that cognition and
intelligence are not properties of an individual person but rather arise
from interactions of a person with the surrounding environment (including
other people and artifacts). Recent research projects have attempted to use
computer networks can facilitate the development of "knowledge-building
communities" (Scardamalia & Bereiter, 1991), in which groups of people
collectively construct and extend knowledge. In many of these projects,
students share ideas, theories, and experimental results with one another.
Distributed constructionism asserts that a particularly effective way for
knowledge-building communities to form and grow is through collaborative
activities that involve not just the exchange of information but the design
and construction of meaningful artifacts.
Computer networks can be used to support distributed construction activities
in several different ways. The rest of this paper is organized around three
major categories of distributed construction activities (with examples of
projects from our research group in each category):
* Discussing Constructions
* Sharing Constructions
* Collaborating on Constructions
Perhaps the most basic constructionist use of computer networks is as a
forum for discussing construction activities. By using electronic mail,
newsgroups, and bulletin boards, students can exchange ideas, tips, and
strategies about their design and construction activities.
For example, Michele Evard (1996) examined how a computer network allowed
elementary-school students to share ideas with one another while they were
engaged in a project designing their own video games. The students exchanged
questions and answers through a Usenet-style newsgroup. By the end of the
project, the students produced much more sophisticated video games than
students in previous years (who did not have access to network discussion
groups). The students found answers to technical problems more quickly, and
good design ideas spread through the class more rapidly. Interestingly,
Evard found that the students used the video game newsgroups much more
frequently than they used other online newsgroups that were not directly
associated with ongoing construction activities.
Beyond simply discussing constructions, students can use computer networks
to share certain types of constructions with one another. In that way,
students can try out one another's constructions, and perhaps copy and reuse
parts of each other's constructions.
On the World Wide Web, students can create pages for others to see. But the
Web is very limited in the types of constructions that it can support.
Currently, it is used primarily for displaying "passive" data forms such as
text, images, and videos. Students can not easily post animations,
simulations, and other process-oriented artifacts that they construct.
Brian Silverman and I developed the LogoWeb to enable students to share
dynamic artifacts with one another. To the student, the LogoWeb software
looks identical to MicroWorlds, the most common commercial version of Logo.
Students can use the LogoWeb software to create animated stories and
simulations, just as they would in MicroWorlds. But instead of saving their
projects on a local hard disk, students can save their projects on the
LogoWeb (a network totally separate from the World Wide Web). Once a project
is saved on the LogoWeb, anyone else with an Internet connection and a copy
of the LogoWeb software can access the students' creations.
The LogoWeb hides all of the messy inner workings of the Internet (such as
URLs and directory hierarchies) from the students. Students need to learn
only two new commands: savelink (to save a project on the LogoWeb) and
linkto (to access a project on the LogoWeb). Students use multiple-word file
names to create an implicit directory structure (for example, using the name
of their school as the first word in the file name).
In an initial trial, two elementary-school classrooms (one in California and
one in New York) used the LogoWeb to post copies of video games that they
had written in Logo. Students at each site were able to try out one
another's games, look at the underlying code, and copy segments that they
found useful for their own projects. In the future, the LogoWeb could serve
as an ever-growing museum of Logo projects, where students go to get ideas
for new projects and to exchange projects with friends.
Of course, it would be much better if these types of projects could be
shared on the World Wide Web itself, rather than on the separate LogoWeb
network with its own software. With the new Java programming language,
people can now create LogoWeb-like dynamic projects on the World Wide Web.
But there is a problem: Java is designed for expert programmers, not for
pre-college students. Java will no doubt lead to a proliferation of dynamic
Web pages, but most students will be only users of those pages, not
In an effort to blur this boundary between designers and users, our research
group is now developing a new environment that we call Cocoa--a type of
"Java for kids." As with LogoWeb, students can create Logo-like animations
and simulations, but they can place them directly on the World Wide Web. A
major challenge is to facilitate the sharing of objects in Cocoa. Our goal
is to make it easy for students to copy and reuse parts of each other's
projects--for example, copying a "flying bird" from one project and
inserting it into a different project. We are developing libraries of "clip
behaviors," analogous to clip art. Students can clip a behavior from one
page and insert it directly into another object.
Collaborating on Constructions
Computer networks can support a more fundamental change when they enable
students not only to share ideas with one another, but to collaborate
directly, in real time, on design and construction projects.
MUDs provide one approach for collaborative construction on the Internet
(Curtis, 1992; Bruckman & Resnick, 1995). MUDs (The acronym MUD initially
stood for "multi-user dungeon," since MUD were originally developed for
multi-player Dungeons and Dragons games. Today, people sometimes refer to
MUDs as "multi-user domains.") are text-based virtual worlds in which
participants literally construct the world in which they live, writing
programs to define the behaviors of objects in the online world. MUDs began
as multi-player game environments, but they have evolved into more
general-purpose meeting places on the Internet, where people gather to enjoy
one another's virtual company and to work together to extend the virtual
world. For example, you can decide to be a purple bird and build yourself a
nest, or you can become a munchkin and join others in building a replica of
MUDs explicitly combine construction and community. When people build new
objects and places on a MUD, there are always other people around to act as
users, consultants, advisers, and critics. As part of her research on the
MediaMOO project, Amy Bruckman (1994a) conducted a study of adults who
learned to program on a MUD. She found that people learned to program more
easily in MUD environments (than in traditional single-user settings). The
reasons: programming became more authentic and motivating since there was an
"audience" for the artifacts that they created; other members of the
community provided technical help, emotional support, and feedback; the MUD
world was full of example objects, providing a sense of what was possible, a
source of ideas, and collection of reusable parts. More recently, Bruckman
created a new MUD called MOOSE Crossing, designed explicitly as a learning
environment for children (Bruckman, 1994b). MOOSE Crossing includes a new
programming language called MOOSE, designed to make it much easier for
nonexpert programmers to join in the construction of the virtual world.
Computer networks also make possible new forms of modeling activities, in
which students collaboratively construct models and simulations--and, in
some cases, participate in the simulations that they construct. Greg
Kimberly (1995) developed an environment called MarketPlace that enables
students to participate in economic simulations over the Internet, playing
the roles of buyers and sellers in a virtual marketplace. MarketPlace
includes online discussion facilities, designed to support not only economic
deal-making among participants but also reflection and analysis of the
economic patterns that arise from the interactions. Kimberly found that
participants gained insights into core economic ideas such as supply and
demand, economies of scale, and externalities.
Our research group is currently developing a more general-purpose modeling
environment for the Internet, called the Network Clubhouse. The
decentralized nature of the Internet makes it particularly well suited for
modeling and exploring the workings of decentralized systems. People
encounter many decentralized systems in their everyday lives (such as ant
colonies, traffic jams, and market economies), but they generally have great
difficulty understanding the workings of such systems, often assuming
centralized control where none exists (Resnick, 1994). Research has shown
that modeling activities can help people develop better intuitions about
decentralized systems. The Network Clubhouse is based on the belief that
collaborative modeling activities will provide a new avenue for people to
move beyond the "centralized mindset." For example, students can use the
Network Clubhouse to collaboratively create an ocean ecosystem on the
Internet, with each student programming the behavior of an "artificial
fish"--then discussing with one another the systems-level phenomena that
arise from the interactions. It is expected that students, through these
activities, will be able to develop an understanding certain scientific
phenomena (such as feedback, homeostasis, and self-organization) that are
usually studied only at the university level, using advanced mathematical
With this type of activity, the Internet can support changes not only in the
process of learning (bringing students together into collaborative projects)
but also in the content of what is learned (providing a natural
infrastructure for modeling and exploring decentralized phenomena). Too
often, educational innovations focus only on how students learn, without
enough attention to what students learn. Many of the representations and
activities used in today's schools were developed in the context of (and are
most appropriate for) pencil-and-paper technology. New media (such as the
Network Clubhouse) make possible new representations and formulations of
scientific knowledge--making that knowledge accessible to more people (and
at younger ages) than previously possible.
The Internet as Rorschach
The Internet acts as a type of Rorschach test for educational philosophy.
When some people look at the Internet, they see it as a new way to deliver
instruction. When other people look at it, they see a huge database for
students to explore. When I look at the Internet, I see a new medium for
construction, a new opportunity for students to discuss, share, and
collaborate on constructions.
Most of the projects described in this paper are still in their infancy.
They lay a foundation for the study of distributed constructionism, bringing
together ideas of construction and community. But much work needs to be done
to more fully explore the ways in which construction and community can
influence and enrich each other.
Many members of the Epistemology and Learning Group contributed to the
design and implementation of the projects described in this paper. Brian
Silverman and Andy Begel have been responsible the technical underpinnings
for several of the projects (LogoWeb, Cocoa, and the Network Clubhouse).
Discussions with Alex Repenning have helpful in developing ideas about
distributed constructionism. Cocoa and Network Clubhouse are trademarks of
Bruckman, A. (1994a). Programming for Fun: MUDs as a Context for
Collaborative Learning. National Educational Computing Conference, Boston,
MA. Available via anonymous ftp from media.mit.edu in pub/asb/papers/necc94.
Bruckman, A. (1994b). MOOSE Crossing: Creating a Learning Culture. PhD
dissertation proposal. Cambridge, MA: MIT Media Lab.
Bruckman, A., and Resnick, M. (1995). The MediaMOO Project: Constructionism and Professional Community. Convergence, vol. 1, no. 1, pp. 94-109.
Curtis, P. (1992). MUDding: Social Phenomena in Text-Based Virtual
Realities. Directions in Advanced Computing (DIAC) conference. Berkeley, CA.
Evard, M. (1996). A Community of Designers: Learning through Exchanging
Questions and Answers. In Constructionism in Practice, Y. Kafai & M. Resnick
(eds.). Mahwah, NJ: Lawrence Erlbaum.
Harel, I. (1991). Children Designers: Interdisciplinary Constructions for
Learning and Knowing Mathematics in a Computer-Rich School. Norwood, NJ:
Kafai, Y. (1995). Minds in Play. Mahway, NJ: Lawrence Erlbaum.
Kafai, Y., & Resnick, M., eds. (1996). Constructionism in Practice:
Designing, Thinking, and Learning in a Digital World. Mahwah, NJ: Lawrence
Kimberly, G. (1995). Learning in MarketPlace: Economic Objects to Think With
and Talk About. Masters thesis. Cambridge, MA: MIT Media Laboratory.
Papert, S. (1991). Situating Construction. Constructionism, I. Harel & S.
Papert (eds.) Norwood, NJ: Ablex Publishing.
Papert, S. (1993). The Children's Machine: Rethinking School in the Age of
the Computer. New York: Basic Books.
Resnick, M. (1993). Behavior Construction Kits. Communications of the ACM,
vol. 36, no. 7, pp. 64-71.
Resnick, M. (1994). Turtles, Termites, and Traffic Jams: Explorations in
Massively Parallel Microworlds. Cambridge, MA: MIT Press.
Salomon, G., ed. (1994). Distributed Cognition. Cambridge, UK: Cambridge
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, vol. 1, no. 1, pp. 37-68.