art and artifact

Art and Artifact of Children's Designing: A Situated Cognition PerspectiveAuthor(s): Wolff-Michael RothReviewed work(s):Source: The Journal of the Learning Sciences, Vol. 5, No. 2 (1996), pp. 129-166Published by: Taylor & Francis, Ltd.Stable URL: http://www.jstor.org/stable/1466773 .Accessed: 27/11/2012 01:22

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

Taylor & Francis, Ltd. is collaborating with JSTOR to digitize, preserve and extend access to The Journal ofthe Learning Sciences.

http://www.jstor.org

This content downloaded by the authorized user from 192.168.52.64 on Tue, 27 Nov 2012 01:22:39 AMAll use subject to JSTOR Terms and Conditions

THE JOURNAL OF THE LEARNING SCIENCES, 5(2), 129-166 Copyright ? 1996, Lawrence Erlbaum Associates, Inc.

Art and Artifact of Children's Designing: A Situated Cognition

Perspective

Wolff-Michael Roth Faculty of Education

Simon Fraser University

The purpose of this study was to investigate knowing and learning in an engineering design environment within an elementary classroom. Based on extensive ethnographic observation, video recordings, interviews with participants and observers, and children's design artifacts and engineering logbooks, fourth- and fifth-grade students' designing activities were interpreted from the perspective of situated cognition. The results show that children's designing was related to the artifacts, tools, materials, teacher-set constraints, and current trends in the setting. However, these elements cannot be taken as having some absolute ontology but are interpretively flexible. In the course of the engineering unit, the fourth- and fifth-grade students learned to exploit this interpretive flexibility to frame and solve problems. The emerging artifacts had at least two important functions: They were resources that structured the design process by both opening up possibilities and providing constraints, and they served to coordinate discursive and practical actions. The findings have important implications for affordances and constraints of learning environments in which designing is both a goal and a vehicle of instruction and for the evaluation of students' activities in such settings.

Interests and goals are central to people's meaningful everyday activities; understanding how people pursue goals is critical to understanding cognition (Schank, 1993/1994). Design is a form of problem solving in which interests and goals bear on processes and outcomes in essential ways; by their very nature, design processes are about the new, uncertain, and ill-defined (Bucciarelli, 1994). Because design is

Requests for reprints should be sent to Wolff-Michael Roth, Faculty of Education, Simon Fraser University, Burnaby, British Columbia, Canada V5A 1S6. E-mail: [email protected]


130 ROTH

the central activity in technology (Faulkner, 1994), it fits well with an emerging interest in science, technology, and society issues, and in technology as context for learning science (see essays in Layton, 1994).

Designing, the creation of artifacts that specify how a system should be organ- ized, is one of four types of learning environments that make learning in practice a central element (Schank, Fano, Bell, & Jona, 1993/1994). Design activities offer opportunities for learning in school because they are the epitome of open-ended and ill-structured problem solving and allow students to construct and test knowledge by incorporating in the design their ideas and feelings (e.g., Harel, 1991; Harel & Papert, 1991; Kafai, 1994). Papert and his students distinguished between learning through design and professional designing. Learning through design is not exclusively represented in final products (the artifacts in the title), but in the process (art). Although most professional design concentrates on the product as the essential outcome, learning to and through design makes process the central issue of education. Such learning is only partially reflected in the artifacts that result from the design process and which are traditionally the sole basis for the evaluation of learning.

Based on the results of previous studies on children's designing and my own work on student-centered design of science experiments, I identified the Engineer- ing for Children: Structures (EFCS) curriculum that allows students to design artifacts in an open-ended process (Association for the Promotion and Advance- ment of Science Education, 1991). EFCS activities are not designed to transmit legitimated and canonical engineering knowledge. Rather, they provide students with opportunities to (a) explore and experience some critical (but not necessarily prespecified) issues about designing structures, (b) learn to manage the complexity of open-ended and ill-structured design situations, and (c) learn to collectively design, make sense, negotiate, plan, and execute group projects.

The goal of this study was to understand the process of designing during EFCS activities. Specifically, the study was designed to provide empirically based answers to three questions: (a) What elements of the learning environment contribute to the design process and products? (b) What is the ontological status of the identified elements? and (c) What is the role of design artifacts in collective designing? The answers provide an understanding of the processes by means of which students achieve learning outcomes, such as knowing how to design, negotiate differences, use a variety of tools and materials, talk and write engineering design, and exploit complexity and uncertainty of open-ended problems. I assumed that my new understandings would help to construct better design environments that capitalize on the strengths of currently existing curricula and eliminate some of their weaknesses. To make minimal assumptions about the contributions of individual and context to cognition, this study takes as a starting point a theoretical position based on situated learning (Gooding, 1990; Lave, 1988; Suchman, 1987)


CHILDREN'S DESIGNING 131

and symmetric anthropology (Latour, 1993) which do not a priori place cognition into the head of individuals.

THE EMPIRICAL STUDY

This is one of a series of studies designed to understand elementary students' knowing and learning in science classroom environments (McGinn, Roth, Bou- tonn6, & Woszczyna, 1995; Roth, 1995c, in press). All studies in this series are based on the following three attributes. First, children design artifacts as a means to learn science. Premised on the work of Schank and his colleagues and their notion of coincidental learning (Schank et al., 1993/1994), I assumed that children who designed towers, strong arms, bridges, or domes would learn (a) about properties of materials and forces operating in structures and (b) how to modify material properties, structural strength, and stability. By designing complex machines for specific purposes, such as lifting or moving heavy loads, I predicted that students would learn about simple machines such as pulleys, levers, or inclined planes and about associated concepts, such as mechanical advantage, force, energy, and work (McGinn et al., 1995). Second, these studies are designed as "authentic" learning environments to the extent that problems are either ill-defined or defined so loosely that the students can define their own goals and problem frames and experience a sufficient level of uncertainty and ambiguity in finding and achieving solutions. Third, these learning environments allow students to (a) experience themselves as part of a community in which specific practices and resources are shared and scientific and technological knowledge are socially constructed and (b) draw on the expertise of more knowledgeable others, whether they are peers, teachers, or outside experts. Three ethnographic studies have been completed. The data used to illustrate children's designing in this study came from a Grade 4 to 5 classroom in which children engaged in the design of structures. The extent to which this classroom was a community of practice has already been described and theorized (Roth, 1995c, in press). The following paragraphs provide descriptions of curriculum, participants, data, and methods of analysis employed in this study.

Curriculum

The curriculum EFCS was developed to generate opportunities for elementary- school children's construction of engineering knowledge and positive first experiences with engineering. As a practical application of science, EFCS is a vehicle for introducing science concepts, providing ill-defined problem-solving contexts, and fostering positive attitudes in children toward science and technology. Activities


132 ROTH

in the program are set up as engineering design problems and as contexts for learning to work and solve problems in a collaborative manner. Thus, the EFCS curriculum exhibits some of the fundamental properties of courses that take students' goals and interests as starting points for learning and that focus on practice rather than mere acquisition of facts (Schank, 1993/1994; Schank et al., 1993/1994; Schon, 1983). As the core goal, EFCS students constructed bridges that became part of an exhibit staged by the local science museum. This exhibit explained and illustrated an ongoing engineering competition for constructing a link between two sections of the city via a tunnel or bridge.

In this curricular implementation, the unit began with a series of initial problems designed as motivational activities and as an occasion to observe students' problem framing and solution finding prior to the engineering activities. In two subsequent activities, students developed strengthening techniques and joints with given materials like glue, pins, paper clips, masking tape, pipe cleaners, cardboard, paper, popsicle sticks, skewers, string, wool, and so forth. These lessons allowed children to develop resources for later engineering challenges. After a lesson on stabilizing two- and three-dimensional geometric figures such as pentagons, hexagons, and cubes, the children built their first structure, a real or imaginary object or creature. For the remainder of the unit, students (a) constructed towers meeting student-determined specifications (e.g., being earthquake proof, supporting a ball, or featuring a moving elevator), (b) built bridges with a minimum span of 30 cm, and (c) mounted huts from newspaper and newspaper rolls large enough to harbor several children, the so-called "megastructure."

The EFCS unit lasted 13 weeks. Due to the regular, part-time teacher's working schedule, the unit was taught twice per week, normally for 90 min. In some instances, however, the students spent all morning (3 hr) on their engineering projects. Two engineers visited the class to talk for about 30 min each about their own work and to interact with the students in the context of their engineering projects. The unit ended with two activities: (a) a field trip to the local science museum to see an exhibition on bridges, which also included the children's own bridge projects completed during the unit; and (b) an individual construction project completed at home, which was used as an occasion to assess individual learning. Students had the opportunity to view several films on bridges, usually during lunch hours, including the classic recordings of the Tacoma Narrows Bridge collapse, the building of wooden bridges in the 1990s, and a spaghetti-bridge competition.

Students spent most of their time on practical work that they completed with a partner of their own choice. The two teachers present went from group to group to talk with the students at length about engineering issues and technical problems arising from their work, to facilitate group work and collaboration, and to dissipate student frustrations. During each lesson, time was set aside to talk with the whole class about the children's work. Teachers pointed out features in children's joining or strengthening techniques that are also used by professional engineers, or students



presented what they had done to date, the problems they had encountered, and how they had solved them. Teachers used these sessions to help students reflect on their own problems and solutions by asking them to compliment their peers, ask questions, or provide suggestions for improvements.

Teachers and Students

This investigation was conducted in a mixed-grade French Immersion classroom with 23 fourth graders (10 boys, 13 girls) and 5 fifth graders (3 boys, 2 girls). Most of these students came from middle-class backgrounds, although the full socioeco- nomic range from working class to middle class was represented. The unit was team taught by one of the two regular classroom teachers (Tammy) and a part-time graduate student (Gitte) who was also the developer of the EFCS curriculum. Tammy taught the class for 2 days per week; another teacher was in charge for the remainder of the week. She had 12 years of experience teaching at the elementary level, 6 years of which were part time. Her previous experience included a unit on bridge building, which she taught once before, and two 3-hr unit-related workshops: one on building bridges and the other one on the EFCS curriculum. Gitte had taught at the same grade level for 3 years and had worked for 4 years as a curriculum developer and workshop presenter. Tammy and Gitte planned the activities together to adjust them to the specific needs of the participating students.

Data Collection

Because of its focus on reasoning in complex ill-defined situations, this investigation was modeled on studies of cognition in everyday activity and of work practices in scientific laboratories (e.g., Hutchins, 1991; Lave, 1988; Suchman, 1987). Direct observations and videotapes of design-related activity, field notes, interviews with students and regular and visiting teachers, and students' engineering logbooks constituted the primary data sources. Data analyses adhered to recommendations by theorists in the domains of interpretive research (Erickson, 1986), discourse and conversational analysis (Edwards & Potter, 1992), and interaction analysis (Jordan & Henderson, 1995). Any assertion constructed was tested in the entire data corpus or used to direct further data collection to seek both confirming and disconfirming evidence.

Two video cameras were used to establish a continuous record of student activities during the EFCS unit. At appropriate moments, I interviewed students in the context of their ongoing work. Video-based data were complemented by ethnographic descriptions of events that could not be captured by the cameras. Data sources included the children's engineering logbooks where they kept plans for constructions, reflections on their learning experiences, a glossary, and photographs of their work. Videotaped meetings with Tammy, Gitte, and other teachers who


134 ROTH

had observed the classroom and interacted with the children constituted another data source.

To provide some indication of children's learning, two types of testing situations were arranged. First, students were asked prior to the unit to note or draw their ideas about engineering (e.g., "Engineering is ?"). At the end of the unit, students responded to the same question by writing a sentence and mapping all the words they associated with the topic. Children's glossaries constituted another data source for analyzing their understanding of engineering design. To obtain a better sense of changes in their designing-related activities, I asked students prior to the unit to build a bridge from a single sheet of paper supporting as much weight as possible. Their bridge project, in which they were to build a bridge as strong as possible and spanning at least 30 cm, provided a comparison case 2 months later, but before the end of the unit.

Data Analysis

My interpretative work began with the assumption that reasoning as socially structured and embodied activity can be observed (Suchman & Trigg, 1993). I treated the transcripts of students' conversations as natural protocols of their efforts to make sense of events, structure their physical and social environments, or interact with teachers. Together with the collected artifacts, protocols provided occasions for construing the design work done by individuals, groups, and the classroom community. Sense, objectivity, fact, and social order in human action were treated as irremediably local and contingently produced phenomena.

All videotapes were transcribed as soon as possible after they were recorded. In daily meetings with my research team, I generated assertions that were subsequently tested in the body of the data sources or used to direct the data collection efforts during future field days. Data sources were searched for evidence supporting or rejecting these assertions and followed up with on-site interviews. On the basis of any new information, I discarded, modified, or retained working assertions.

EMPIRICAL EVIDENCE FOR CHILDREN'S DESIGNING

This study sought answers to three questions: (a) What elements of the learning environment contribute to the design process and product? (b) What is the ontological status of the identified elements? and (c) What is the role of design artifacts in collective designing? Each of the subsequent subsections is framed by one of these research questions and a brief answer; each frame is followed by supporting empirical evidence. An overview of learning outcomes is provided in the subsequent section.



Multiple Actors and Heterogeneous Products

"What elements of the learning environment contribute to the design process and product?" Cognition in designing is situated: It is best understood as the confluence of many different elements (actors) and therefore is a heterogeneous process. Designing is not a cognitive activity that can be reduced to individual mental processes. Decisions taken by principal actors, or owners of projects, cannot be understood apart from the tools, materials, artifacts, teacher interdictions and constraints, or the emergent properties of collective, discursive and tool-related, activity and design artifacts.

Although the children claimed sole authorship for the creation of certain artifacts, there were other aspects of the learning environment that contributed in fundamental ways to the art and artifacts of their designing (see Figure 1). Designing in this classroom did not occur in a vacuum but in a studio-like atmosphere that represented a learning culture and the associated network of relationships among psychological, sociological, and material actors in specific settings. (In symmetric anthropologies, actor stands for any aspect of the setting that could be important in the learning of the system to prevent premature closure in my ordering activities on questions of causes and effects, human and nonhuman factors involved in the constructed phenomena, or individual and social aspects of knowing; Latour, 1987.) The contributions of various actors to an emerging artifact are graphically represented in Figure 1; these contributions find their material expression in the completed artifacts.

A variety of actors contributed to an artifact's development at various stages to give it its final shape. Figure 1 shows how tools, materials, community standards, teacher-set constraints, current state-of-the-design artifacts, individual preferences, and past discursive achievements contributed to the emergence of a specific design artifact, AST's' tower. The heterogeneity of the tower's origins makes the descrip- tion assemblage (Turnbull, 1995) quite appropriate. In spite of the observed heterogeneous origins of each artifact, one student group was always identified as its author. This attribution was asserted by children through accompanying descriptions assigning ideas and work to themselves and by the teachers when they praised the designers. There existed very few instances in which a child or group credited others. In this example, AST's tower emerged as a product of the interaction between these elements, with the three students-the designers-as principle actors (see Figure 1). However, not all tools, artifacts, materials, individuals, and rules that exist in the classroom were of equal importance in accounting for the emergence

Ilnitials are used to refer to a group as a whole. The following abbreviations are used: AST (Andy, Simon, and Tim), CA (Chris and Arlene), JJ (Jeff and John), KK (Kitty and Kathy), and PR (Peter and Ron).


__)

DAY 1

Negotiation S: 'Make cubes'. T: 'Make cubes and then we attach them all together' [9:441

Negotiation T: We are making the bottom part. [Yours] is for the bottom part. [9:50]

Present state S: It's really unshapely, but it's coming T: It really has the shape, 'cause it's gonna shape... [10:12]

Interdiction, constraint tape an pin joint: 'We are not allowed to use [tape]' 'But it makes [the joint] strong' [9:52]

Teacher Tammy: Which shape does it take to make something stable? T: A triangle [10:191

DAY 2

Present state T: Take these ones off, and make the bottom really stable... then reassemble it [9:16]

Present state A: Perfect size, because you can switch it around different sizes [9:26]


"-4

Individual T: 'I am making a

-cone top' [9:16] Negotiation

- S: Make a pyramid A: 'Don't make a triangle', 'the bottom needs to be square' [9:19]

- Teacher Tammy: 'Can you add some triangles there?' [9:351

Teacher Tammy: 'Can you enlarge it in this [vertical] direction?' [9:35]

pin jointI

'welded' joint

Other student, material Tom: 'I can't glue that, it's way too hard. Use tape' [9:43] Tool welded joint: Tim melted and fused straw with his glue gun . A: 'When we attached it with tape, it would come apart' T: 'Simon put these things and I got an idea that it wouldn't wiggle that well.' (Repair, Day 3)

Community flag: (Jeff, Chris/Arlene); A: 'And all the other people have it alike...' T: I am gonna make a flag [10:10]

Teacher Tammy: You can fix that there, to me, there is a little too much movement [10:13]

FIGURE 1 The development of a design artifact, Andy, Simon, and Tim's tower, beginning with some initial idea to a finished product. The design artifact emerged from the interaction of individuals with their peers, materials, setting, teachers, tools, community standards, rules, and the design artifact itself.


138 ROTH

of an artifact. There exists a horizon of events; elements that lie outside of this horizon, although they may exist for other people in the same setting, are not relevant aspects in the design history. For example, although a glue gun was present in the classroom, it did not influence the design of AST's tower in its early stages. Only later, with a growing sense in this classroom community that glue guns make strong joints, did AST also consider the use of this tool in their own activities. The following examples highlight how (a) intragroup negotiations, (b) interactions in the community, (c) tools, and (d) the present state of the artifact shaped design processes and products.

How Intragroup Negotiations Shape Designing and Designs

Design artifacts emerged as material witnesses of the collaborative planning, constructing, and negotiating of team members. During these negotiations, design changes and "master" plans emerged, but they were not the sum total or least common denominator of the key actors' initial input. Rather, the negotiations constituted a changing context such that individual proposals were transformed. New plans emerged with or without bearing any likeness to the originating individual ideas.

Example 1. Tim had first suggested and then constructed a "cone top" for AST's tower which had a triangular base. Simon and Andy objected, claiming that the triangular base could not be fitted to the square of the cubic shapes which they had produced (see Figure 1, bottom left). After Tim's solution to the lack of fit was rejected on both technical and aesthetic grounds, for it consisted of joists across the top floor (see Figure 1, top right), he changed the triangular base into a square base. The resulting top of their tower emerged from their interaction rather than being the idea of any single individual. Neither Tim's pointy and triangular cone top, nor Andy and Simon's "cube top" became the common plan. They designed a much wider pyramidal top that fit to the top level of the existing tower artifact.

Example 2. Chris (C) and Arlene (A) discussed over their artifact, a bridge deck with two piers.

A: Why don't we attach it to cardboard? C: You want to? And put water under it? A: Yeah, and we could attach it so it will stay up, straight. [1210V3p.1]2

2Source codes are used throughout. A source code, such as [1210V3p. 1], indexes the excerpt in terms of the date (December 10th), type and number of recording (V3 = Videotape 3 of that day), and page of the corresponding transcript (p. 1).



Arlene suggested to attach each pier to a small cardboard foundation that would stabilize the "wobbly" bridge. Chris, who thought that Arlene wanted to glue the bridge onto a large piece of cardboard suggested to "put water under it." Their design artifact consequently developed to include both features. Rather than being the sum total of two ideas, the resulting artifact emerged as the second idea contingently followed the first.

How Interactions in the Community Shape Designing and Designs

This classroom showed similarities with design studios in which knowledge and information circulate freely among members (Roth, 1995c, in press). The nature of the classroom as design studio facilitated interactions with other design groups and the emergence of community standards and trends. Such standards and trends contributed to individual designs and could be recognized in resemblances among artifacts belonging to different groups and the associated explanations, such as "everyone else is doing a Canadian flag." Furthermore, students directly interacted with students from other groups. Through these interactions, outsiders to a design project became themselves actors who shaped an emerging artifact.

Example 1. KK's tower showed a number of influences and inspirations some, but not all of which were credited to peers. For example, their initial solution consisted of a compact central tower from bundled straws; in this, it resembled JJ's tower. After a catastrophic collapse, KK relocated their construction site next to AST where they completed the design with an addition that bore great similarity with AST's tower. To make braces for the lower section, they doubled the length of straws, a technique which Kathy credited to Clare. Finally, they adorned their structure with a Canadian flag because "a lot of people have them" [1105V4p.2].

Example 2. Other design artifacts similarly embodied contributions from passing groups and students. In AST's tower, a pinned-sleeve joint for producing double length straws was used for the base module (see Figure 1, bottom center). Tim developed this joint after Tom refused to hot glue the straws and suggested using cellophane tape. When the tape failed to produce a solid joint, Tim added a pin resulting in a pinned-sleeve joint. Tim also constructed a Canadian flag to adorn the tower long after other flags had appeared in the classroom community (see Figure 1, bottom right).

How Tools Shape Designing and Designs

The availability of certain tools changed the design process and affected the artifacts produced. As an increasing number of students brought their glue guns,


140 ROTH

and as even more learned how to operate them, the design culture changed favoring joining techniques on the basis of this tool (Roth, 1995c). The strength that the materials and joints received from this technique was so great that some students found specific engineering techniques unnecessary, in particular some of those that the teachers wanted students to "discover," such as triangular braces. This phenomenon was so pervasive that Gitte indicated she would disallow certain materials and tools when she taught the unit again. In the following examples, the glue gun changed the task of designing. It afforded the construction of artifacts impossible to achieve with other techniques.

Example 1. Tim serendipitously discovered a "welding" technique in which two straws could be joined by softening the plastic with the glue-gun tip. Drawing on this technique, his team designed a very tall tower that drew its stability from Tim's joint (see Figure 1, bottom center). Similarly, Tom's tower project took its shape because of the strength it received from another, serendipitously discovered joint. After he accidentally burned a hole into a straw, Tom recessed a second straw into this hole. The strength of this new joint allowed him to construct a windmill with turning blades.

Example 2. Because Peter and Ron used a glue gun to join their wooden materials, they could design the arch of their bridge on the basis of rectangular and trapezoidal building elements. Glue and skewers provided enough strength and stability to make triangular braces redundant. In this way, Peter and Ron did not use the bracing technique ("triangles") that teachers hoped all children would eventually master and employ on a regular basis.

How the Present State of an Artifact Shapes Designing and Design

Children's ideas literally took form in the emerging design artifacts. As the ideas took form in the artifacts, the latter constrained and limited other design options, but also afforded new and different ideas. In this sense, children's designs became self-reflexive artifacts that embodied past activities and at the same time foreshad- owed affordances and future constraints. The future constituted a horizon that included all thinkable trajectories. However, in the evolutionary process of design, only one of these trajectories was realized. The emerging artifact embodied and enfolded the actual design trajectory, decisions, conversations, and so on. In this way, design projects summed over their own history: Any new idea took the present state of the artifact as starting point (in more than 50 projects there was only one counter example where a design was abandoned in its current form).



Example. AST began their tower (see Figure 1) from cubic modules but did not know whether the smaller or larger cube should form the base (see their negotiation in this situation later). They framed their alternative designs as "the earthquake building downtown" (also West Coast Energy) or as the London Tower, which are narrower on the bottom or the top, respectively. After the decision was made to go with the London Tower plan, AST continued with this shape, although they encountered stability and strength problems so that they reframed their project as Tower of Pisa. Later, while they discussed how to attach Tim's cone shape to the existing artifact, Tim suggested several intermediate cubic modules to produce the shape of the Empire State Building. Finally, the group agreed that Tim could attach his cone top if he made a square base for it so that they could make a shape like the Eiffel Tower. At each of the decision points, future design states were discussed in terms of the present artifact and individual prior knowledge, shared prior knowledge, or both. Had the students decided to use the West Coast Energy design with a base narrower than the top, subsequent framing of their design problems as Tower of Pisa, Empire State Building, and Eiffel Tower would have been less likely.

Summary

After an artifact was completed, specific individuals or groups received credit for its design and construction. On the other hand, the presented empirical evidence shows that design processes are situated and their products cannot be reduced to individual, or sum of individual, knowledge and skills. Rather, many different elements (e.g., intragroup negotiations, interactions in the community, tools, and the state of the artifact) came to bear on designing and emerging design artifacts. The important point is that designing is not a homogeneous process by means of which designs are thought of and then implemented in material form as design artifacts. Rather, designing is a heterogeneous, distributed, and situated process that arises from the interaction of different elements. The resulting artifacts have to be understood as heterogeneous assemblages that emerged from situated activity of designers rather than as homogeneous fixations of children's ideas and skills.

Interpretive Flexibility

"What is the ontological status of the identified elements?" The relative importance of tools, materials, community standards, teacher-set constraints, current state-of-the-design artifact, individual preferences, and past discursive achievements is indeterminate because they do not exist in any absolute sense. Their meanings are interpretively flexible and the degree of their


142 ROTH

salience is contingent on specific local developments. Tools, materials, artifacts, teacher interdictions and constraints, history of activity and design artifacts, or plans do not have stable ontologies. Because of the different meanings that they can have for different children, they lead to different design actions and ultimately, learning.

In the engineering design lessons, children learned to recognize and exploit in productive ways the interpretive flexibility of discourse with group members, materials, tools, design artifacts, and teacher interdictions. Rather than being stifled by the open-ended and ill-structured situation, students used the flexibility of their situation in creative ways. Problems did not exist in any absolute sense but emerged as interactive achievements. As a consequence of this interpretive flexibility, situational choices made by the principal actors (members of a design group) appeared to be inherently unpredictable. There is ample evidence in the data that all the elements indicated in Figure 1 were subject to variable ontology. I limit the analysis to factors not previously presented; that is, empirical evidence is provided for the interpretive flexibility of (a) plans, artifacts, and problems and (b) teacher constraints and interdictions.

Interpretive Flexibility of Plans, Artifacts, and Problems

In this classroom, it made sense to understand the artifacts as outcomes of first drafts rather than as the fabrication and implementation of complete and standard architectural plans. From the children's perspective, they treaded novel design territory and engaged in design innovation. Children's final artifacts were products of a designing process, and their "experiments" were more like engineers' sketches rather than their final plans, models, or actual structures. Such activities are of that kind which, according to Anderson (1990), have not yet yielded to modeling efforts in terms of "rational" processes. The failure of such modeling rises in part because problem is frequently used as an ontological category, such as in word problem (Roth & Bowen, 1993/1994, 1995). In this context, the interpretive flexibility of materials, artifacts, tools, or rules and constraints led to the simultaneous charac- terization of situations as problematic and unproblematic. That is, one student may have seen a problem to be solved before the overall design work could resume, whereas a partner may not have seen a problem at all. Being able to interpret situations in a flexible manner, to reframe a situation such that it was unproblematic or afforded an action which would remove a problematic situation, was one of the major aspects of learning.

Example 1. Initially, children's ideas and plans were unspecified so that the meaning of an expression, such as "earthquake-proof tower" (AST), "two towers



with sliding object between" (CA), or "baseball stadium with removable top" (PR), was quite open even among members of a group. As the artifacts emerged (like residues of children's designing activities), meanings converged and stabilized. In the beginning of each project, existing artifacts did not constrain the interpretive flexibility enough for group members to have similar conceptions of how the final artifact would look. The following episode illustrates that students could begin their design process (including construction) without prespecifying exact outcomes so that their plans remained open to interpretation to be settled far into the project.

AST had decided to construct an earthquake-proof tower. They had planned to "make cubes and then attach them all together" [1104Vlp.1]. At the point of the following conversation, Tim (T) and Simon (S) had already constructed two "cubes" of different sizes that they attached in the way shown. The excerpt was part of a discussion about which shape the tower should take.3

1 T: We are making the bottom though (1) like the bottom part, this is for the top ((pointing to Simon's cube)), the top part (1) because we are making the base 'cause we can make like a smaller part and a bigger one, like the one building, like the earthquake building downtown.

2 S: But then you have to turn it over (1) ((turns artifact over)) so it can stay on the ground.

3 T: Yeah (2) 4 S: So I am actually building the bottom ((touching bottom

cube)) 5 T: Yeah, but we are supposed to be building the bottom

just for the change of the ground. (1) So we are going to change (1) and we are testing it on either side (4) [1104Vlp.2]

Simon

Tim

Tim

Simon

Tim thought that he had built the (smaller) base, which would make his tower look like the earthquake-proof West Coast Energy Building downtown (line 1), a landmark which the class had repeatedly discussed (see the upper sketch). Simon, on the other hand, felt that the larger cube should make the foundation to assure the tower's stability. He later admitted to one of the teachers that their differences in interpretation had confused him ("I thought I was going to do the floor, and Tyler

3The following transcription conventions are used:

((pointing)) double parentheses enclose concurrent gestures, actions (10) pause, in seconds (??) unheard words; the number of "?" indicates approximate number words missing

Ion the ground] "/" begins overlap which, for both speakers, ends with "]"


144 ROTH

said, 'do the top'; and now I am confused" [1104Vlp.6]). Tim resolved the issue by suggesting that they could test both designs, thus leaving open what the tower's final shape should be and who had built the "bottom part" (line 5). Each of the three boys had started the earthquake-proof tower with his own idea. Immediately before the conversation, all of these ideas were compatible with their activities of constructing cubes of different sizes. During the episode, it became clear that Tim and Simon interpreted their current artifact in different ways; thus, whether the design was the same depended on the point of view. Children's discussion over and about the artifact allowed the differences to come out and to be discussed. The group accepted Tim's suggestion in line 5 and deferred to a future moment the decision that would remove the interpretive flexibility of the artifact. In this way, the design of their tower was repeatedly reinterpreted as expressed by their labels, including West Coast Energy Building (Vancouver), Eiffel Tower, London Tower, Empire State Building, CN Tower (Toronto), and the Leaning Tower of Pisa.

Example 2. In a similar way, the interpretive flexibility of KK's plans (a tower that supports a ball) allowed them to make significant design changes. After their first tower collapsed, they reinterpreted the artifact at hand to be the center piece of their new tower. With the redesigned tower in the form of an oil rig, they achieved their goal. The old tower was then reinterpreted as a centrally located elevator that carried much of the load placed on top of the tower.

Interpretive Flexibility of Constraints and Interdictions

Both teachers contributed to the students' designs to a nonnegligible degree. These contributions were primarily of three types: (a) Teachers suggested possible shapes and forms for the students' artifacts; (b) they provided hints as to how existing artifacts could be strengthened and stabilized; and (c) they set certain constraints which allowed or forbid certain materials. Some of the suggestions or constraints were constant across groups, such as when teachers wanted children to be familiar with specific engineering techniques (triangular braces, testing structures by bringing them to "catastrophic failure"), whereas others were more design specific and changed from group to group. The children usually recapitulated these constraints by tagging each others' ideas: "We are not allowed to ." Whether and how these constraints limited the students' designs was a matter of interpreting a teacher's utterance and the situation at hand. Thus, teacher interdictions and constraints could not be taken in any unequivocal way because their meaning had to be established in each specific situation. Constraints, such as "design a bridge by using only a set of materials provided," meant different things in different contexts. There were situations in which students acted in a way that may be interpreted as "according to the law." In other cases, the same group of students



submitted to an interdiction in one situation, but bypassed the same interdiction a little while later. Finally, students renegotiated the conditions of a project, essen- tially changing it in the process. These observations are consistent with other reports that show rule following as a creative achievement rather than a self-evident act (Suchman, 1987). It is also possible to argue that rules as such do not exist. What a text (pronounced by a teacher, written on a piece of paper) means has to be constructed in each situation so that it makes little sense to speak of a constraint, interdiction, or rule as existing a priori.

Example 1. It was quite difficult to predict how students would interpret teacher constraints because this differed not only across groups but also within groups. For example, each group was allowed only a limited amount of materials for their bridge designs. When Tim and Stan, who worked together on a bridge from cardboard and glue, wanted more material, they negotiated with one of the adults. They convincingly argued that the skewers provided additional strength to their project. Only moments later, they requested more cardboard, but Gitte refused. Immediately afterwards, Tim and Stan cut up the cardboard pad that they had previously used to protect their table top; the pad became a source for additional materials in spite of the interdiction. They reinterpreted "cardboard pad" to mean "cardboard as a source of needed material." The students' actions appeared to be possible (there were no repercussions), for the teachers had not explicitly forbidden more materials.

Example 2. AST did not even ask a teacher to get additional materials. Although they hesitated at first to add another fastening material (tape), Tim's argument that tape would strengthen their structure legitimized their "forbidden" action (see Figure 1, top center). In spite of the teachers' constraints to use only one type of fastening material, the interpretation of the teachers' other constraint, "make it strong," was more important; it was also the grounds for convincing the teachers that changing the rules of the assignment was necessary and thus legitimate. In a similar way, JJ exploited the interpretive flexibility of the constraint that no tower could be taped to a support. They stabilized their tower with stays (from string) that they taped to a support. During a whole-class discussion, Jeff argued successfully that the stays and the central structure were different; although the stays were taped, the tower itself was not taped.

Summary

Interpretive flexibility influenced processes and products of children's designing. Contrary to traditional assumptions about discourse, objects, and events, the data illustrate that these had an ambiguous ontology and were inherently subject to


146 ROTH

interpretive flexibility. That is, what an utterance meant, what activities a specific tool afforded, what actions a plan implied, or how teachers' interdictions and task constraints limited students' designing was subject to local interpretations; these interpretations were contingent and situated achievements. At the same time, students' activities were not random-although inventions were often spontaneous and serendipitous-but students framed and actively used inventions to varying purposes to complete their ideas and goals. Their activities could not be downgraded as "mere playing" and "trial-and-error" processes; rather, the nonrandom nature of children's designing rested in the exploitation of interpretive flexibility and the interpretations of unexpected outcomes as affordances for future activity.

Design Artifacts: Structuring Devices

"What is the role of design artifacts in collective designing?" Artifacts resulting from the children's designing not only bear the marks of their situated, or heterogeneous origins but also have important functions in the design process itself. They reflexively constitute their own context. Design artifacts become tools for the design process, that is, tools to think with. They constitute part of a designing discourse, and become tools for mediating design talk.

Earlier studies revealed the importance of physical objects for high-school physics students' planning of experiments (Roth, 1994, 1995b). By means of material objects, high-school physics students structured phenomena that allowed them to frame meaningful and important research questions. Furthermore, the objects they used constrained the interpretive flexibility of talk in such a way that they could make sense of each other's ideas. Design artifacts in this study had the same functions. They not only were products of children's design activities but also allowed children to structure an initially ill-defined design problem; they allowed children to negotiate the interpretive flexibility of tools, materials, teacher-set constraints, or the present state of the artifact. In this way, design artifacts are reflexive because they are both goals of children's activities and they structure these activities. Two aspects of design artifacts contribute to learning in a significant way: (a) During the design of artifacts, thinking and acting are inextricably intertwined; and (b) design artifacts constrain the interpretive flexibility of students' design talk.

Designing: Inextricability of Thinking and Acting

This and other formal studies of children's designing (e.g., Harel, 1991; Kafai, 1994) attest to the power of design activities to engender various forms of learning. One may ask, what is it about designing artifacts that leads to demonstrable learning outcomes? Is it that design artifacts themselves contribute to children's cognition



and interaction with others during the design process? I contend that the design artifact allows the reintegration of thinking and acting that traditional schooling (based on traditional psychological theory) separated into abstract and concrete modes, and after denigrating the concrete, supported them to different degrees (Roth, 1995b; Wilensky, 1991). This study shows that design artifacts are important aspects of learning because they allow thinking through manipulating; they are objects student use to integrate acting and thinking. (In Double Helix, Watson, 1968, provided an interesting account of how important insights about the DNA structure arose from the manipulation of physical representation of the four bases involved: Material objects were things to think with that integrated the concrete and abstract.) Children's manipulations of materials, tools, and artifacts were part of an integrated designing activity. Thinking and manipulating tools, materials, and the artifact were intricately intertwined. Thus, designing could not be understood as fixation of ideas from the head. Rather, it was an integral activity that had mental, material, practical, and social aspects that make "thinkering" a quite suggestive metaphor. Thinkering decenters the traditional notion of cognition to include, as the following examples show, the manipulation of nonmental entities.

Example 1. Although Chris and Arlene constituted one of the two groups that used drawings (which were not required), their design work was characterized by the union of thinking and acting. While Chris manipulated some straws in designing a tower, she repeatedly uttered, "I am just experimenting" [1104V2p.6]. Out of these experimentations grew the design artifact. She later explained, "I just started out to make an experiment of the frames and then Arlene taped it; and she said to make it more stable" [ 1104V2p.7]. This experimental nature of the designing process, the union of manipulating and thinking was also evident in the following transcription from the first day of Chris (C) and Arlene's (A) first bridge design [1126V1p.7].

1 C: I got an idea too, we can just, this could be the bottom one, we tie it around like this ((begins to tie)), like I might use (??)

2 C: If it's still too wobbly at the end, ((she moves string and structure which "wobble")) //we can put (??)]

3 A: And then we] go around like that ((loops string around post)) and tie it ((ties loop around second post))

4 A: Is this a little far off? We should put it down a bit ((lowers the loop)) 5 C: Yeah, a bit down 6 A: Put it down a bit, right here ((lowers string)) before we tie it.

After deciding to make a bridge from string, they began tying a string around two posts "which could be the bottom one" (line 1). Chris uttered, "if it's still wobbly"


148 ROTH

simultaneously while she tested the structure (line 2); Arlene's planning for the next design move coincided with tying the string into a loop. In line 4 again, Arlene reacted to the current state of the artifact, especially to the result of an earlier move; she tested her question, "Is this a bit far off?," by lowering the loop of string. While Chris worked on her second bridge, her question, "How are we gonna get this span on?" was accompanied by the move of bringing together the pier and platform (span), and resulted in three utterances (C: "Glue?" A: "Tape." C: "Glue and tape." [1209V1p.9]) and the associated actions of hot gluing (Chris) and taping (Arlene) the two pieces.

Example 2. JJ designed a straw bridge in repeated cycles of testing and modifying their emerging artifact. In the process, they spontaneously invented layering of materials and bracing as techniques that increased the strength of the bridge from about 50 to 346 wooden blocks of testing weight. The difference between building a bridge according to a prespecified plan (their drawings) or by algorithmically applying standard formulas and designing an artifact can be seen further from Jeff's recollection of the design process: "First I wanted to make my bridge, but me and John started to argue, so we didn't use our plans, and just made it up as we went along" [1209V1p.2].

Example 3. Similarly, Stan designed a hinge for his coffin from a cone- shaped cardboard roll as he went along. First, he used two pins to fix the moveable cover plate. However, as he opened and closed the cover plate, the pins came out of their position. In response, Stan added a drop of glue on the tips of the inserted pins preventing them from leaving their position. He realized that he had constructed a hinge in a new way and without using a lot of metal ("I just kept doing like that, [and the pins] went in and then came out. And then I glued them, with the glue gun ... and when I closed it up and pulled it up, it hardly broke" [1029V1p.9]).

How Artifacts Constrain Interpretive Flexibility

Emerging design artifacts had important social functions also attributed to adult designing: They make designing as collective activity possible (Henderson, 1991; Suchman & Trigg, 1993). Design artifacts acted as conscription devices in the sense that they (a) permitted students to engage one another by making direct reference to the artifact (pointing, gesturing), (b) focused participants' attention and communication, and (c) served to represent the knowledge negotiated and constructed in the pursuit of some goal. In this way, design artifacts provided a setting and a backdrop to students' design talk. The presence of the artifacts reduced the interpretive flexibility of students' utterances about their design ideas. This con- scriptive quality of the artifacts encouraged and facilitated extended student con-



versations and was an important aspect of the activity structure that supported the emergence of meaningful engineering design language.

Design artifacts also provided a context in which students could begin to incorporate new elements into their language without loss of intersubjectivity when they used them in inappropriate and ambiguous ways. Sometimes they appropriated a word, such as span, from a teacher's discourse on the span of bridges. At first, they often used these words incorrectly: They had not yet populated their discourse with their own intentions (Bakhtin, 1981). Nevertheless, in spite of the incorrect and ambiguous use (e.g., using span to refer to the piers of a bridge), others understood what they said. Their manipulations of and gestures to the artifact helped listeners to make sense of their partner's utterances. It was only over time that students developed a consistent language (e.g., to refer to the piers of their bridge as piers or posts). The following two examples illustrate how design artifacts facilitated communication by constraining the interpretive flexibility of design talk.

Example 1. This example clearly shows how the artifacts supported students' talk. Tim (T) and Renata's communication could not be understood separate from the artifacts that were thus part of the conversational activities.

1 T: If you make one giant triangle ((points to her structure))

2 it only takes two sticks to make a giant triangle.

3 Like you see, I got the shape and then I got two pieces of spaghetti ((gestures their placement over his pentagon))

4 so its a giant triangle ((outlines the triangular shape over his pentagon)) [1028V1p.2].

(3)

(4)

(3'

Renata had wondered how to stabilize and make a pentagonal shape rigid. Because she was not successful in her first trials, Tim offered to explain what to do. In his attempts prior to the aforementioned transcript, Tim had tried without success to explain with words and gestures in the air how to stabilize a pentagon ("You just take, you only have to put one on the top" and "If you make a triangle, if you make a giant triangle, it keeps it stabler" [ 1028V1p.2]). He then used his own pentagon from spaghetti and mini marshmallows to outline where to put the two pieces (line 3), and how this would yield the triangle he talked about (line 4). The presence of the artifact and Tim's gestures allowed Renata to construct the meaning of "put one on the top" and "make a giant triangle." Pointing to and gesturing the


150 ROTH

placement of the two pieces of spaghetti and outlining where the two "giant triangles" could be seen were important and integral aspects of the communication. This integration of discourse and gesture was made possible by the design artifacts that provided a backdrop against which both were rendered meaningful.

Example 2. In the earlier cited episode during which AST discussed the shape of their tower, the students were able to make sense of their respective utterances by referring to and manipulating the artifact. By standing the artifact (see Figure 1, top left) on the smaller side and pointing to the respective cubes, Tim elaborated the meaning of "bottom part," "top part," the shape of the earthquake- proof tower he wanted to build similar to the "earthquake building downtown." In turn, Simon elaborated his utterances by gesturing and manipulating the artifact. In this way, his comment, "so it can stay on the ground," referred to stability of the building, and his cube became the bottom.

Summary

There are two important functions of artifacts to the process of designing: (a) Design artifacts afford the integration of thinking (abstract) and acting (concrete), the process of designing; and (b) emergent design artifacts support communicative processes by reducing interpretive flexibility. That is, the design of artifacts is not merely a context in which children learn to design and learn about the materials they use, but these artifacts reflexively influence the way children learn and interact. Children do not just design, but the design artifacts structure designing; children do not just talk about their designs, but the design artifacts shape the interactions from which they emerge.

LEARNING OUTCOMES

The purpose of this study was to understand the processes of children's designing. Many stakeholders in the educational arena, however, are interested in outcomes and are likely to ask questions, such as, "What did your students learn through engineering design?" At this point, the answers have to be tentative, for the study of situated cognition in schools is starting. Some of the domain's current assumptions make it difficult to compare performance across contexts and time. Human (practical and intellectual) activities are always situated, tool mediated, and object oriented; tools and objects are structuring resources for knowing and learning such that a change in tools and objects constitutes a change in the task environment rather than simply a change in human capacities (Callon, 1994; Lave, 1988; Sayeki, Ueno,



& Nagasaka, 1991). A problem for assessing children's learning resides in the question, "What constitutes design-related knowledge?" Pea (1993) suggested that

expertise is defined dynamically through continuing participation in the discourse of a community, not primarily through the set of problem-solving skills and conceptual structures. Achieving expertise is becoming indistinguishable in your actions and uses of representations in the language games at play from other members of a community of practice. (p. 271)

Surely then, expertise has many dimensions across which the performance of individuals varies. For example, some children in this study competently talked about braces as an important feature in the stabilization of structures or included them in design drawings. At the same time, because they did not include braces to stabilize their design artifacts, they did not demonstrate the same level of competence in the constructions. Different forms of knowledge assessment may therefore yield different levels of competence. We observed similar variations in a class of sixth- and seventh-grade children engaged in the design of machines (McGinn et al., 1995). Some children correctly analyzed diagrams of pulleys and predicted the tensions in various parts of the apparatus. At the same time, they could not set up appropriate pulley systems when provided with materials.

To assess learning, I took teachers' top-level goals as a starting point. There were three such top-level goals teachers had for children:

Goal A: To explore and experience some critical, but not necessarily prespecified, issues involved in designing structures such as learning to strengthen individual materials, build and reinforcejoints, or stabilize structures.

Goal B: To learn to manage the complexity of open-ended and ill-structured design situations.

Goal C: To learn to collectively design, make sense, negotiate, and plan and execute group projects where "collective" implies that children learn to talk about the objects of their activity and the processes by means of which to attain them.

On the basis of these considerations, I identified six dimensions of children's learning in the engineering design context that are also characteristic for competent design engineers (Bucciarelli, 1994): (a) coping with complex design tasks (Goal B); (b) exploiting interpretive flexibility for creating innovative designs (Goal B); (c) knowing important aspects about design and constructing structures (Goal A); (d) negotiating differences between individual plans and understandings, courses of actions, and the organization of collective work (Goal C); (e) using a variety of tools and materials and using their properties in new and innovative ways (Goals


152 ROTH

A and B); and (f) talking and writing engineering design (Goal C).4 Each of these dimensions is illustrated and discussed in the following paragraphs.

Coping With Complexity

During this engineering unit, students learned to cope with the complexities of open-ended design tasks. On the engineering pretest, which asked students to make a strong bridge from just one sheet of paper, they quickly abandoned the task. Some students tried to support the load without manipulating the sheet of paper, others folded it once or twice, and very few attempted to search for stronger solutions. Two months later, the children took the design problem (a bridge spanning at least 30 cm and supporting as much weight as possible) as an extended challenge that they pursued with great persistence over a 3-week period. They considered several design alternatives, actively searched for solutions which would enhance strength and stability, and exploited collaboration to achieve better solutions. Before the unit, students appeared to be at a loss, they walked around the classroom, seemingly not knowing what to do next. Two months into the unit, their behavior during the bridge-building project was characterized by a great deal of self-confidence that permitted them to deal with the challenges of ill-defined design tasks.

Exploiting Interpretive Flexibility

Students learned to exploit the interpretive flexibility of tools, artifacts, materials, talk, and so on. Rather than accepting limited meanings that may be associated with a specific tool or material, children generated new meanings which allowed them to put these things to new use. A case in point are the glue guns. Ordinarily, in what may be interpreted as the canonical use, glue guns are employed to heat and liquefy solid glue sticks by means of electrical energy from AC outlets. In this classroom, glue guns were creatively reinterpreted and successfully used, but frequently in ways for which they were not designed. Children produced new and very stable joints from drinking straws by burning holes in straws into which they placed other straws that they carefully attached with glue. Rather than using the glue gun to exude hot glue, they interpreted it as a tool for burning holes into straws. They created another type of joint by bringing together the heated areas of two straws to form a solid bond. Finally, after serendipitously discovering that glue guns could

4With the notion of "talking and writing engineering design," I follow the equivalent expressions of "talking science" (Lemke, 1990) and "wrighting sociology" (Ashmore, 1989).



be operated even after they were unplugged for a few minutes, they invented the "portable" glue gun.

Designing

Students increased their understanding of design and they intentionally applied this understanding in their subsequent work. This was particularly evident for "design knowledge related to the natural world," one of Faulkner's (1994, p. 447) five categories of design knowledge. Toward the end of the unit, most children strengthened their materials by, for example, laminating sheets of cardboard, bundling straws, spaghetti, and skewers, or producing tight newspaper rolls. At an intermediate level of complexity, they used braces, cantilevers, stays, and arches for stabilizing and strengthening their structures, all of which are examples of knowledge related to design practice in Faulkner's scheme. There is also evidence that children learned when they could frame problems such that previously successful solutions could be reapplied. In such situations, they intentionally applied design strategies that evolved from solutions to earlier problems.

Negotiating Differences

There were many opportunities for participating in collective activity. Through this participation, children learned skills that allowed them to negotiate alternative understandings, plans, and actions. This ability is well-illustrated in the following example from a worst-case scenario: the collective work of two students with previous problems relating to others. A video analysis of Stan and Tim's collective bridge project showed that in the course of its 9-hr duration, the two negotiated all their plans and actions. There was not one situation in which either student was unwilling to negotiate. Neither Tim nor Stan proved to be intransigent or followed the other without reflection. When there were two options to be considered, both demonstrated flexibility by allowing for a comparison between the alternative ideas or solutions and helped each other prepare the second solution. They then based their decision on the relative merits of the two solutions.

Using Tools and Materials

Students developed competent practices related to tools and materials. Thick descriptions of these practices and the development of a design culture, where students learned from each other many tool-related engineering design practices, are provided elsewhere (Roth, 1995c, in press). I showed that children developed great skill in strengthening a variety of materials, stabilizing structures, handling


154 ROTH

tools such as glue guns, and working with a variety of other materials. Many of the resources and practices deemed relevant by the children were rapidly shared at the classroom level, which contrasted with a much slower appropriation of discursive practices championed by the teachers.

Talking and Writing Engineering Design

Central to successful engineering design and science is the ability to communicate in talking and writing (Bucciarelli, 1994; Latour, 1990). The growing competence with communicating design is therefore an important if not the main aspect of learning from design activities. The children's presentation of their work, conversations during construction, and their glossaries prepared at the end of the unit provided evidence of the engineering discourse constructed in this class (Roth, in press). This discourse was rich and appropriate for designing, negotiating, critiqu- ing, and framing problems and solutions in ill-defined design situations. It is evident from the data that the children had populated this discourse with their own intentions and purposes.

Children's expertise in talking engineering design is exemplified in the following episode. Jeff had presented his bridge and described that he included particular braces to distribute the forces acting in the structure: "It brings the weight down on the end. And it would be easier for a force to go across here, the weight goes here, and the tension brings it down" [1216V1p.3]. He accompanied this explanation by gestures that correctly indicated the direction of the forces acting in his structure. The fact that he and his partner John had tested their bridge both in its normal position and upside down gave rise to several exchanges. In one, their classmate Ron (R) explicitly questioned Jeff (J) about the symmetry of the construction.

1 R: If you took the legs off, if you cut them off around the bottom of the bridge, do you think it would hold the same as upside down. Is it built the same?

2 J: Well actually, if you cut the legs off, you might try it putting it down like this, without the legs, to see how strong it would be, you might try that, but we don't think it would hold any more.

3 R: Is the bottom and the top built the same? 4 J: Yes, originally this was it, and then we built this under the bottom, to

make it stronger.

In situations such as this, students demonstrated their design expertise. Much like Jeff here, they were able to explain and defend design choices, and, like Ron, question those of other groups in the community. Jeff explained the distribution of forces along various braces. When Ron asked about the symmetry of the bridge,



which could have led to different performance during the test for strength, Jeff pointed to the specific parts of the bridge that they had added to strengthen the design and increase the maximum carrying capacity.

Students' expertise was also expressed in their written work. Initially, children demonstrated rather vague conceptions of engineering design. At the end of the unit, their associations were rich and varied. The analysis of students' posttests provides evidence for five aspects of learning: (a) Children made rich associations after this unit, (b) they commanded an extended civil engineering-related vocabu- lary, (c) children integrated their experiences and ideas they had prior to the unit with those they added during the unit, (d) they integrated the experiences of various activities within the unit into their understandings, and (e) children made meaningful connections.

Further evidence for children's understanding of various aspects of engineering design comes from the analysis of their glossaries. They demonstrated their understanding through drawings and text. The two glossary entries in Figure 2 illustrate that children expressed both static and dynamic aspects of engineering design. Chris's explanation of triangles (braces) shows that she understood some of the typological features of bridge building; that is, naming parts of the bridge appropriately (in terms of the design language shared in the classroom) and showing braces as fundamental design elements for strengthening structures. The entry under "compression" expresses dynamic features of engineering design. Rather than explaining what compression is, Chris explained the term as something that could be felt.

DISCUSSION

This study was designed to answer three questions: (a) What elements of the learning environment contribute to the design process and product? (b) What is the ontological status of the identified elements? and (c) What is the role of design artifacts in collective designing? In answering these questions, I followed recommendations of interaction analysis to track artifacts, their appearance and disap- pearance from the horizon of human actors, and the function of artifacts in structuring interactions (Jordan & Henderson, 1995). My answers emphasize the heterogeneous (situated) nature of cognition:

* Designing is not simply a psychological, social-psychological, or sociological phenomenon as radical constructivist theorists, social constructivist theorists, or both want to have it, but has important material aspects. These material aspects do not only lie in the artifactual nature of students' products in engineering design, but much of their individual and collaborative thinking and sense making are constituted by manipulating materials, tools, and artifacts.


01 0)

L o~fesiQoL -W'e ,

\OI'C~ es

r~i';a~-j8cl~a-? ,~b93(

fFto l, 1

I,,,d.,,.9on 0 f b?Ajkvsb

'0A6 A1A44~0

ki, it AA" 2W ,o AWAAAA6

FIGURE 2 Two-page entry in Chris's glossary. Her entries illustrate an understanding of the typological (naming design features and techniques, and parts of structures) and dynamic features of design (forces and their actions).



* The elements of a setting (discourse, materials, tools, artifacts, interdiction, community standards, etc.) do not exist in any absolute sense but have a flexible ontology such that their meanings and applications have to be determined and elaborated in the context of their use.

* The artifacts created by students are not just ends in themselves but are important structuring resources in children's designing. They are also important mediating tools for making sense through collective activity.

Children who participated in these open-ended design activities increased their competence along several dimensions also characteristic for design engineers. I briefly documented that students learned to cope with complex design tasks, exploit interpretive flexibility, design structures, negotiate differences, use a variety of tools and materials, and talk and write engineering design.

The view of children's designing described in this article advances knowledge in at least three ways. First, although there are a small number of descriptions of learning in complex classroom settings (Harel, 1991; Kafai, 1994), none have attempted to conceptualize learning (designing in this study) as situated activity equivalent to existing studies on adult and out-of-school cognition. This study shows that learning through design is integrally related to those aspects of the setting that student-actors constructed as important. Second, other studies related to school learning take discourse, artifacts, settings, and tools to have stable ontologies with fixed inherent meanings. In contrast, I showed that these items are ontologically unstable and thus in essential ways interpretively flexible. Finally, the study views design artifacts in two ways: (a) as integral aspects of students' cognitive activity during design and (b) as tools to facilitate negotiating, constructing collective meaning, thinking, and planning in groups. In this effort, I address situated learning in a classroom, an issue explicitly eschewed by Lave and Wenger (1991) in their influential monograph.

Engineering design constitutes an interesting learning environment distinctly different from traditional ones. The major educational goal in engineering design is so that students can develop two important kinds of knowledge necessary for making increasingly intelligent choices and decisions: (a) deep familiarity within a specific domain and (b) strategies for structuring complex and ill-defined problem settings. I documented that students achieved such knowledge as outcome of their learning. In the process, student-produced artifacts have important functions in that they serve as and support the structuring of the learning environment and problems themselves; they become tools for designing by indissolubly integrating thinking and acting. At the same time, the emerging artifacts constitute a focus and backdrop for students' discursive activities of talking, pointing, and gesturing, that allow them to make sense of each other's utterances and to negotiate shared meanings in the face of ambiguity. Emergent design artifacts embody these negotiations and thus contribute to the convergence of meanings in the collectivities. In their progression,


158 ROTH

and with the convergence of meanings, the interpretive flexibility of plans, ideas, and the artifacts themselves are increasingly constrained. In these respects, the engineering design artifacts have similar functions in the sense-making processes of students as those observed in settings where students used computer-based microworlds to construct ray diagrams and Newtonian motion experiments (Pea, Sipusic, & Allen, in press; Roschelle, 1992; Roth, 1995a, 1995c), or design-learning environments for other, younger students (Harel, 1991; Kafai, 1994).

Tools, artifacts, materials, and spoken and written text are interpretively flexible rather than embodying and embedding specific meanings and applications. Stu- dents developed a remarkable competence to recognize and capitalize on the ontological ambiguities of their settings, a knowledge dimension important in engineering design. Interpretive flexibility is a crucial attribute of learning environments intended to foster student innovation, creativity, and negotiation, all of which characterize successful and adaptive technological change (e.g., Bijker, Hughes, & Pinch, 1987; Bucciarelli, 1994). Being able to thrive or capitalize on interpretive flexibility allows engineers to frame and reframe problems so that they become suitable for resolution, to redefine needs in a society, or to negotiate or renogotiate customers' demands and thus to become successful entrepreneurs.

Interpretive flexibility is also an important concept for understanding the negotiation of meaning, and for the construction of design languages, which, although not necessarily those of engineers, are important and productive means for communicating and designing within the elementary students' design studio. Students experience first hand the social construction and maintenance of their own design language without the concomitant emergence of an "anything goes" attitude that teachers often associate with constructivist teaching. In the process of learning through design, teachers, engineers, films, or a science museum exhibit become important intermediaries that allow children to link their own design languages to canonical discourses; the extent to which this was achieved by children in this study has been sketched in the section on learning outcomes. Although I was able to observe the influence of the canonical discourses on children's design language, the teachers in this classroom experienced enormous resistance whenever they tried to force a change in students' language (Roth, 1995c, in press). Rather, such changes came about when students felt that a particular notion was relevant and meaningful in their own and their peers' work.

The view of elementary students' designing developed in this article acknow- ledges the situated nature of cognition by accounting for the contributions of multiple actors in the construction of design artifacts. Thus, although individual actors claim their authorship, the children's artifacts have to be seen as heterogeneous objects whose origins cannot be traced to specific individuals and times. As in the world of adult designing, children's designing is not a linear process of applying individual ideas, the sum total of which will lead to some end product, the design artifact. Rather, designing changes aspects of the setting; changing the



setting changes cognition in fundamental and irreversible ways. An appropriate metaphor for this process of change is evolution. We do not simply find an organism of interest adapting to a stable environment but a complex system in which all the parts continually adapt to each other. Because of the complexity of the system, the trajectories of individual organisms over longer time scales are unpredictable. In this classroom, for example, as soon as children introduced glue guns in increasing ways, the task and artifacts were changed in fundamental ways (Roth, in press). The tool permitted new actions (e.g., building new types of joints) and was reconstructed by the children in such a way that it afforded activities for which it was not designed (melting straws, portability). The availability of new tools, joints, or engineering techniques for stabilizing structures changed designing itself. In a similar way, new interpretations of design artifacts, tools, interdictions, and materials changed the design environment and provided new opportunities and constraints.

Design-learning environments are consistent with the findings of situated cognition (e.g., Lave, 1988). Students learn as they are challenged by designs of considerable complexity. Rather than learning an odd collection of decontextual- ized principles, they contextualize their knowledge in sets of experiences that make learning more meaningful. In this, design-learning environments share similarities with the goal-based scenario approach that specifies designing as one of four categories for organizing learning environments (Schank, 1993/1994). Specifi- cally, my students engaged in the explication of the design aspects of their tasks. Questions such as, "What was your major problem and how did you solve it?" "What is it about the glue that's really good compared to pins and tape?" and "Which part would go first if you were testing it?" asked children to critically reflect on the decisions they made during the design process.

Designing and Students' Goals

There is a tension between the notion of design as open-ended activity and the fact that the goals in the activities-building towers, bridges, and megastructures-were set by teachers. At a second level, the goals of individual students were subject to group processes in which one collective goal was negotiated. In schools, the specification of top-level goals by the teacher has the purpose of specifying pedagogical goals to be achieved through the design activity. To work at all, however, top-level goals need to be accessible to learners. For students in this study, the goals included the provision of opportunities for exploring and experiencing critical engineering designing issues, managing complex open-ended design situations, and learning to collectively design, make sense, negotiate, and plan and execute group projects. The learning outcomes documented that the curriculum achieved these goals. The learning environment shares a characteristic feature with


160 ROTH

Schank et al.'s (1993/1994) goal-based scenarios and Harel's (1991) software-design environment: After some top-level goal has been established by the teacher, students define their own goals and subgoals and subsequently complete the tasks on their own behalf. Gitte and Tammy specified the construction of creatures, towers, bridges, and megastructures. The specifications for each of these projects was then determined by the students: Groups chose to make earthquake-proof towers, wind mills, cranes, towers connected by gondolas, or strong arms (towers sticking out from wall).

Important aspects not discussed by Schank et al. (1993/1994) are the possibilities emerging from collective designing where group goals arise from negotiations between individual members. That is, for example, children interested in trucks were not in the position to convert the curriculum into a "truck curriculum" but had to negotiate their goals with teachers and peers. However, students learned to flexibly interpret the overall goal so that they found ways to include their personal goals in the collective goals. They included trucks, other "micromachines" (various toy machines), and dolls as decorative pieces to towers and bridges.5 Furthermore, pursuin

art and artifact

Documents

design artifacts

design processes

jstor terms

jstor archive

engineering design environment

grade students

terms conditions of

cognition schank