Journal of Research in Science Teaching, scientific discourse, Journal of Classroom Interaction, observations, Chen, ethnographic research, knowledge construction, teachers and students, science and technology studies, University of Chicago Press, discourse processes, rubber bands, technical paper, lines, interpretation, related studies, student papers, National Association for Research in Science Teaching, Science Education, Harvard University Press, Classroom Interaction, University of Wisconsin Press, writing activities, classroom teacher, musical instruments project, Center for Applied Linguistics, pp, Princeton University Press, Curriculum Inquiry, Classroom discourse, science teaching, science studies, learning science, student groups, scientific discourse processes, student group, interpretations, scientific writing, classroom communities, research methods, discourse practices, Charles Bazerman, Judith Green, Santa Barbara Classroom Discourse Group, scientific knowledge, Cambridge University Press, group processes, studies show, Ethnographic Research Cycle, American Educational Research Association
JOURNAL OF RESEARCH IN SCIENCE TEACHING
VOL. 36, NO. 8, PP. 883915 (1999)
The Sound of Music: Constructing Science as Sociocultural Practices through Oral and Written Discourse
Gregory J. Kelly, Catherine Chen Graduate School of Education, University of California, Santa Barbara, California 93106 Received 2 June 1998; accepted 28 May 1999 Abstract: In this article, we examine the oral and written discourse processes in a high school physics class and how these discourse processes are related to sociocultural practices in scientific communities. Our theoretical framework is based on sociological and anthropological studies of scientific communities and ethnographies of classroom life. We review the use of discourse analysis as a methodological orientation in science education and provide a logic-of-inquiry framing how we used discourse analysis in our ethnographic research. Our ethnographic analysis showed that, through students' participation in creating scientific papers on the physics of sound, their appropriation of scientific discourse was related to the framing activities of the teachers and the social practices established over time in the classroom. Our textual analysis of the student papers focused on how they used evidence to make claims. We explore the lessons learned from participating in the classroom of these students. © 1999 John Wiley & Sons
, Inc. J Res Sci Teach 36: 883 915, 1999 Classroom discourse is increasingly becoming a prominent area of research in education (Gee & Green, 1998; Hicks, 1995) and in science education in particular (Kelly & Green, 1997; Klaassen & Lijnse, 1996). Concurrent with this trend are increasingly more detailed and specified pictures of the workings of scientific practices studied through the multidisciplinary lens that have come to be known as science studies (Roth, McGinn, & Bowen, 1996). Sociological and anthropological studies of scientific practice and knowledge offer educators access to the discourse processes leading to scientific knowledge in the particular communities that comprise scientific fields (Kelly, Carlsen, & Cunningham, 1993). In this article, we explore discourse processes in school science through the lens of science studies and educational ethnography. Our examination of oral and written discourse in a high school physics classroom responds to the call in recent reviews for research that investigates how disciplinary knowledge
is accomplished through classroom communication (Hicks, 1995; Klaassen & Lijnse, 1996). Anthropological studies of scientific practices (Traweek, 1988; Knorr-Cetina, 1995) and ethnographies of science in schools (Lemke, 1990; Moje, 1995) document the importance of studying how what counts as science is interactionally established by members within given communities. In this study of a high school physics class, we adapt the methodological orien-
Correspondence to: G. J. Kelly Contract grant sponsor: Spencer Foundation © 1999 John Wiley & Sons, Inc.
KELLY AND CHEN
tation of anthropological studies of scientific practices to research how what counts as science is interactionally constructed, defined, acknowledged, and/or appropriated by members within this particular community (Kelly, Chen, & Crawford, 1998a). Thus, we examine school sciencein-the-making (Latour, 1987)--that is, the developing and evolving social processes of members of a classroom as they construct situationally defined notions of science, experiment, text, and evidence, among others. Through a close examination of classroom practices, we identify through the study of social interaction (e.g., deciding experimental protocols, interpreting inscriptions) the cultural practices (e.g., presenting experimental results, writing in conventional genres, applying exemplars to unique problems) that constitute membership in a community. The theory underlying our work is that as members of a community (e.g., scientists in a particular field, members of a classroom) affiliate over time, they create through social interaction particular ways of talking, thinking, acting, and interacting (Green & Dixon, 1993). Thus, to examine how what counts as science is established in a high school physics class, we focused on the communicative processes of oral and written discourse. In a recent review of classroom discourse, Hicks (1995) provided insights into how knowledge and practices of academic disciplines are shaped in and through the everyday talk and actions of teachers and students. Studies of classroom interaction show how discourse is a mediator for student learning (Cazden, 1988; Hicks, 1995; Mehan, 1979) and how particular teaching practices shape student opportunities for learning (Green & Dixon, 1993; Tuyay, Jennings, & Dixon, 1995). For example, classroom practices that provide students opportunities to engage in scientific ways of questioning, investigating, and knowing afford uniquely different opportunities from those that focus on disciplining students into particular semantic relations
hips constituting propositional knowledge (Carlsen, 1992; Lemke, 1990). Thus, what counts as disciplinary knowledge and practices of, and relevant to, a particular community can be viewed as constructed through the conventionalization and formalization of discourse processes as group members affiliate over time and build common knowledge (Edwards & Mercer, 1987; Kelly & Green, 1998). In this article, we review studies of scientific discourse as well as studies of discourse processes in classrooms. This review frames our analyses of the oral and written discourse processes in a high school physics class. Through examination of one genre of scientific discourse, that of an experimental article, we identify ways students used evidence in their writing. We argue that creating a scientific argument is tied to particular and situated social practices established by members of the classroom. Finally, we discuss the lessons learned from these analyses.
Viewing the Communicative System of the Classroom through a Science Studies Lens To establish the theoretical framework informing our work, we review the ways scholars of science have identified the discursive shaping of disciplinary knowledge. Our review is not comprehensive; rather, we offer some of the ways for thinking about science and discourse that influenced our analysis and writing. (For reviews of what science studies might offer science education, see Kelly et al., 1993, 1998a; Millar, 1989; Roth et al., 1996. An overview of the field of science and technology studies can be found in Jasanoff, Markle, Peterson, & Pinch, 1995.) In A fragile power: Scientists and the state, Mukerji (1989) examined how scientists negotiate away aspects of their intellectual authority in the process of trying to maintain intellectual autonomy while being fiscally dependent on state funding agencies. To accomplish the work of doing their research, the scientists in this study (oceanographers) needed to partially turn over their autonomy and participate in the complex social world of making science function as an in-
SCIENCE AS SOCIOCULTURAL PRACTICES
stitution with sets of intradisciplinary and interdisciplinary rivalries that enter into the negotiations of creating viable mechanisms and contexts for doing research. In particular, oceanographers and other scientists need to engage in various forms of discourse tempered and mediated appropriately to given audiences. In examining how scientific discourse is directed, Mukerji identified a broad range of discursive processes necessary for scientists to be successful: Scientists provide expertise to the state in ways that maintained their credibility without trivializing the complexities of the topical debates in a given field; they need to find ways to change the substance of a scientific debate to direct the need to their line of work; they often discredit rival social groups that compete for the same research funding and geographical space; they read and write to journals with specialized and stylized discourse procedures; they present their science to the mass media; they use persuasion to recruit materials and personnel to their particular laboratories and projects; and they find ways to collaborate with colleagues, both within and across disciplines. Nevertheless, while the final products of science can only be achieved through a variety of discourses drawing from the social and political dimensions of scientists' repertoires, the final products are typically represented in written forms following the restraints of particular genres. Mukerji's study of the processes and products of oceanographers show the two faces of science identified by Latour (1987), ready-made science (science in its compressed, formalized, abstracted forms of product) and science-in-the-making (science as it is being brought into existence with social, experimental, and epistemological contingencies). We now turn to the rhetorical form of formalized scientific writing and the sociocultural conditions that shape the construction of these texts. In Shaping written knowledge, Bazerman (1988) examined the genre of the experimental research
article as a cultural form. Through a series of textual analyses, Bazerman identified restraints made on the demands for communication over the history of writing in science. For example, the agonistic forum of the experimental journal article can be seen as a response to the rhetorical demands of the scientific communities that have evolved over time. This historical view suggests that knowledge is entered into science through the use of persuasion and survives through the communication systems and lived practices of the respective communities. Bazerman viewed these communities as having particular conventionalized practices that shape the relationships of text and audience, making persuasion "a lengthy process of negotiation, transformation, and growth of the central formulations and related arguments" (p. 309). Communication presupposes shared knowledge; Bazerman offered two kinds of situations in which this knowledge may be examined: the neophyte becoming familiar with the shared knowledge of the community and the establishment or reestablishment of the shared understanding in times of change, growth, or instability. Bazerman explained the first situation drawing from a Vygotskian perspective. He identified how through processes of participation and scaffolding "gradually the neophyte becomes socialized into the semiotic-behavioral-perceptual system of a community with language taking a major and multivalent role in the organization of that system" (p. 307). As in the study by Mukerji, Bazerman pointed to the commitment of scientists to new formulations, new knowledge, and how such contributions must be understood as promising to be more useful or productive for the relevant community. Proposers of new knowledge must be willing to hold their assertions up to public scrutiny. By entering their ideas into the nexus of discourses, behaviors, and formulations, these candidates for knowledge may come to count as, or be rejected as, science through dialogical and dialectical processes. To understand how the multiple discourse practices among members of and within scientific communities lead to legitimated knowledge in the compressed and compact form of the experimental article, we review Latour's (1987) analysis of the processes of fact production. While science is often entered into schools in its ready-made form (e.g., in the written form of text-
KELLY AND CHEN
books), the sociological processes leading to such formulations need to be examined. By studying science-in-the-making, Latour investigated ways scientists seek to establish assertions as facts. This fact production of the particular communities occurs during periods of instability in the shared knowledge, the second situation identified by Bazerman. Latour analyzed how scientists position themselves and others in their written texts by "bringing in friends" through citation and by using other texts in strategic ways. Latour identified one goal of scientific text production as the generation of high-inference claims (e.g., about the properties of mammal countercurrent structure in the kidney) rather than highly contingent, qualified claims referring to particulars (e.g., slices of flesh in a particular laboratory). Thus, in the process of fact producing, scientists seek to establish facts about constructs like kidney structure from a body of evidence that starts with "slices of flesh." Pictures, figures, numbers, and inscriptions of other sorts can be used by authors to fortify their claims. However, these are potentially dangerous as they also provide readers (particularly critics) ways to unravel the highly generalized assertions to a set of contingent, highly problematic, and perhaps isolated facts about particular physical entities. Thus, the rhetorical demands on scientists include ways of moving from the contingencies of science-in-the-making into the concretized facts that come to be presented as ready-made science. These three studies by Mukerji, Bazerman, and Latour show the range of discourses employed in the construction of scientific knowledge and expertise, the role written texts play in shaping what counts as knowledge, and the rhetorical and textual strategies of producing facts in science. From a research perspective, the treatment of the discourse of science as a unified notion is potentially problematic as scientific discourse processes serve multiple purposes and are embedded in diverse activity systems. These science studies show that scientific practices are complex and embedded in multiple activity systems. Furthermore, they suggest that those considering science and discourse in schools need to examine how science is invoked, appropriated, positioned, understood, and taken up by participants.
Discourse Processes in Science Education Discourse processes, both oral and written, have become the subject of study among educational researchers concerned with the ways such processes support or constrain access to scientific knowledge (Chen & Crawford, 1998; Crawford & Chen, 1997; Kelly & Crawford, 1997), the ways science is presented to and positioned for students (Moje, 1995; Lemke, 1988, 1990), the ways students can be seen as constructing knowledge (Roth & Lucas, 1997), the ways arguments are made (Kelly, Druker, & Chen, 1998b), and the ways authority is invoked (Carlsen, 1997), among others. As our review of science studies showed that scientific practices are "highly literate" (Mukerji, 1989) activities involving both oral and written discourse, we begin by reviewing research in science and writing and then situate this research in a larger field concerned with discourse processes more generally. A recent review of writing in secondary science by Prain and Hand (1996) considered a variety of issues involved in writing science in schools, including conceptions of the purposes of science education and conceptions and perspectives on language from modernist and postmodernist perspectives. In this review, the authors explored the tensions among those who advocate initiating students into the current discourse practices of science, and those, following constructivist perspectives, who advocate a consideration of students' personal understandings and explorations through writing, as well as those postmodernists advocating border crossings and mixed genres. Prain and Hand demonstrated through this review that there are a wide variety of prescriptions for the use of writing in science and that there is considerable disagreement about
SCIENCE AS SOCIOCULTURAL PRACTICES
the purposes of learning science, a view evident in a recent Special Issue of the Journal of Research in Science Teaching focused on the reading-science learning-writing connection (Yore, Holliday, & Alvermann, 1994). However, the model offered by Prain and Hand (1996) suggested a multiplicity of genres and purposes for using writing so that students can understand how to write and critique current representations in and of science. The model they advocate offers five types of elements for writing to learn in science: "writing types; writing purposes; audience or readership; topic structure including concept clusters; and method of text production, including how drafts are produced both in terms of the technologies used as well as variations between individual and composite authorship processes" (p. 618). Our review of science and discourse suggested that the production of written texts was the consequence of particular social actions and oral discourse processes. Thus, written discourses represent only some of the interconnected discourse processes of classroom life. We therefore need to consider the range of discourse processes shaping science in schools. Studies of classroom interaction in science have focused on the ways that science and authority are shaped by teachers through oral discourse, although there is a growing body of work that examines student discourse, particularly in small groups settings (Bianchini, 1997; Finkel, 1996; Kelly & Crawford, 1996; Richmond & Striley, 1996; Roth, 1996). We will consider both here. Discourse analysis centered on the ways teachers shape disciplinary knowledge paints a picture of classroom interaction that is considerably less multidimensional than the activities of scientists identified in science studies. Carlsen (1991, 1992) showed that insufficient subject matter knowledge led teachers to control class
room conversations by privileging facts rather than treating concepts in a dialogic and interactive manner. When they were talking about an area that was less familiar, these teachers generally stayed closer to the textbook inscription by orally reproducing what was written as science. They were also more likely to ask factual, rather than provocative questions, thus invoking the authority of scientific facts. In a later study, Carlsen (1997) took on the role of teacher and analyzed his own classroom discourse. When teaching an area of science in which the teacher had practical and academic knowledge (biology), he entertained and asked more complex questions than when teaching a subject in which he had less background knowledge (chemistry). Through the examination of his discourse processes from an argumentation perspective, he found that his own arguments were more philosophically problematic when teaching the less familiar subject matter. Lemke (1990) showed that through particular discourse practices teachers invoked the authority of science in an ideological manner by focusing on the propositional knowledge of the subject matter without providing the relevant theoretical backing and justification. He suggested that the processes of learning science are connected to the learners' understanding of the genres, formats for reasoning, speaking, and writing in the community that constitutes the discipline of study. However, unlike the studies by Mukerji and other sociologists and anthropologists of science, the teachers in Lemke's studies conceived of and evoked the discourse of science in narrow ways, focusing mainly on the semantic relationships that form the theoretical knowledge of ready-made science. Thus, science was presented to students in its compressed, dense forms, making it less readily accessible to students. Through these discourse processes, the teachers positioned science as a discipline in particular ways that included a distorted view of the methods of knowledge construction (see also Moje, 1995). The resulting portrayal created a "mystique of science" that ideologically represented science as particularly authoritative and difficult. Lemke's (1990) studies of science classrooms led him to suggest that students should be offered opportunities to talk science in a variety of contexts. One approach to providing students such opportunities is the use of group work that allows students to conduct experiments in social situations where they are expected to articulate their ideas.
KELLY AND CHEN
While many studies show that discourse practices of teachers shape the views of science made available in schools, little is known about the extent to which students take up these views or how these views change the concepts held by students. However, recently a series of studies focused on student discourse has emerged aimed at understanding group processes. This research on small-group work shows that access to scientific discourse processes varies within, and certainly across settings and contexts. By focusing on the student discourse in small-group settings, researchers are beginning to get a picture of the complexities of achieving normative goals sought by educators (e.g., uses of evidence, consideration of others' points of view). These studies document the ways students interact, provide initial insights into students' appropriation and use of scientific discourse, and identify the problematic nature of small-group interaction. In a series of studies, researchers Warren, Roseberry, and Conant (1994) offered a view of classroom communities quite different from those studied by Lemke (1990). Working with teachers and students, these researchers based their work on a view of science dependent on argumentation and persuasion, a view informed by science studies. Warren and Roseberry (1995) included in their goals the creation of classroom communities
in which students appropriate the discourse of science: a set of sociohistorically constituted practices for constructing facts, for integrating facts into explanations, for defending and challenging claims, for interpreting evidence, for using and developing models, for transforming observations into findings, for arguing theories (p. 5).
Through these processes, learning science is conceived of as "the appropriation of a particular way of making sense of the world: of conceptualizing, evaluating, and representing the world" (p. 5). In these and related studies, these authors provided examples of students working on meaningful problems (e.g., assessing water quality through sampling of pond, home, and school sources; snails in aquatic environments) for which they are able to engage in the discourse processes of scientific practices, like those described by Latour and Woolgar (1986). The heterogeneous nature of access to scientific discourse has been documented in studies focused on student discourse. Bianchini (1997) studied a middle school
science course that employed Complex Instruction (a model of group work) as an instructional strategy for learning human biology. Through research interviewing, videotape analysis of student discourse, and quantitative analyses, this study examined the relationship of student status, participation patterns, and their relationship to science learning. Drawing from examples of student discourse, Bianchini found that even under the specified conditions of Complex Instruction, students often strayed from the curricular goal of talking science and sought to accomplish other goals such as social positioning. Richmond and Striley (1996) analyzed students' use of arguments in 10thgrade integrated science and found that while students made strides toward improved use of argumentation, these results similarly varied across student groups. The differential opportunities to learn science were constructed in part by the emergence of the social roles of the student group members. In particular, the groups' leadership roles were analyzed in detail, showing how student group leaders shaped the construction of knowledge through their role of mediator and distributor of talk. Across subject matter and grade level, these studies, while generally favorable to the potential of small-group work, suggest that a close examination of the relationships of the established practices of classroom life with the processes and products of student work needs further investigation. Such examination requires long-term data to situate discourse process events in the patterned activities of the members of the classroom established over time through collective activity.
SCIENCE AS SOCIOCULTURAL PRACTICES
The overall ethnographic research project from which this study was conducted spanned 3 academic years. The study took place in a conceptual physics (Hewitt, 1992) course at a public high school in southern California. The first two academic years were cotaught by the first author. The data analyzed for this article were collected midterm in Year 2 by our research team, which consisted of a university professor and two graduate assistants (Chen, 1997; Crawford, Chen, & Kelly, 1997; Kelly & Crawford, 1997). Besides the coteaching by a university researcher, structural features made this course unique. The course was populated by students from all 4 years of high school (9th through 12th grade) and met 90 min a day, 5 days a week. The school is situated geographically between two small cities and draws students from a wide range of socioeconomic backgrounds. The ethnic background of the school is approximately 50% White and 44% Hispanic, with much smaller percentages from Black, Asian/Pacific Islander, and American Indian population
s. [These ethnic/cultural/racial terms are taken from the school's records and are thus "folk" terms (Spradley, 1980)]. However, representation in the physics course did not reflect these school percentages. In Year 2, the ethnic breakdown was as follows: 61% White, 29% Hispanic, and 10% Asian/Pacific Islander. The gender distribution was 51% male and 49% female, consistent with roughly equal numbers of male and female students throughout the 3 academic years. This conceptual physics course was the site for a number of pedagogical innovations over the course of the 3-academic-year project. For example, students used microcomputers to acquire and analyze data, search for historical and technical information on the Internet, and write technical papers. Working in collaborative teams, students spent 24 weeks designing, testing, refining, and presenting scientific projects of their own choosing, such as functional mechanical and thermodynamic devices. The rationale for our choice of participation as participant-observers was derived from ethical, epistemological, and pedagogical considerations. The first author chose to be coteacher of the course to get a close look and develop relationships with the classroom teacher and students; to offer labor to the processes of reconstructing the physics course; to provide teaching assistance and support for student learning; and to use his and his fellow researchers' observations to make suggestions for improving the course. The choice was strategic from a research point of view as, through his role as teacher/researcher, the first author had a reason to be in the conversations that the classroom teacher and students were having about science. In addition, we sought to put the researchers literally and metaphorically in front of as well as behind the video cameras. Rather than conducting research on a teacher and critiquing her or his methods in a judgmental manner, as is so often and easily done to other's teaching methods as noted in a recent editorial (Abell & Flick, 1997), we sought to balance the scrutiny on both the researcher and teacher. Thus, our choice of participation was partially aimed at creating a model of classroom research that disturbed the roles of the researcher as theorist and teacher as practitioner with the former judging the quality of the latter [for studies with similar methodological and participation strategies, see studies by Roth (1995) and Carlsen (1997)]. This choice raises a number of problems, however, not the least of which is the documentation and interpretation of the range of activities in the classroom. To capture the classroom interactions, we included in our research team graduate student ethnographers who made daily observations, recorded written fieldnotes, operated the video equipment, and conducted informal interviews (Spradley, 1980). Data were gathered from multiple sources over the entire ethnography, including videotaped records (lectures, group work, and presentations), student artifacts (written assignments and exhibit displays), ethnographic in-
KELLY AND CHEN
terviews (Spradley, 1979) and fieldnotes [for related studies, see Chen (1997), Crawford et al. (1997), Kelly and Crawford (1997), and Kelly et al. (1998b)].
Methods and Analyses Our study of the oral and written discourse processes of a cycle of activity--an interactionally bound set of lessons and activities centered around a specific theme (Green & Meyer, 1991)--is based on an ethnographic approach to researching human activity (Spradley, 1980). This approach focuses on examining cultural actions, artifacts, and discourse processes through which group members construct social situations. Spradley's ethnographic research cycle suggests a gradual change in the scope of ethnographic observations, from descriptive to focused, and finally, to selective observations. Thus, over the course of our analyses, we zoomed in to look at specific actions and zoomed out to look across groups and over time. Our analysis sought to identify the ties among classroom practices. We began at a general level of observation and through a set of questions posed as part of the Ethnographic Research Cycle gradually focused on how patterns of activity were accomplished through discourse (Kelly, Crawford, & Green, 1997). The Ethnographic Research Cycle consists of asking questions, collecting data, making an ethnographic record, and analyzing these data, through multiple iterative cycles (Spradley, 1980, p. 29). This iterative research process enabled us to examine a range of cultural practices and to explore how these practices, in turn, shaped what was interactionally accomplished. Thus, in the section that follows, we describe our research methods as we unfold the logic-of-inquiry (Gee & Green, 1998) that led us to both our initial interpretations and our next step in analysis. In this way, the later research methods, derived from both questions previously posed and initial interpretations, were contingent on previous iterations of the research cycle. Consistent with ethnographic perspectives (Emerson, Fretz, & Shaw, 1995), there is no clear distinction between methods and findings: "substance (`data,' `findings,' `facts') are products of the methods used, substance cannot be considered independently of method; what the ethnographer finds out is inherently connected with how she finds it out" (emphasis in original, p. 11). Thus, we describe the research methods we used in conjunction with the substantive findings of the study. The oral and written discourse processes of the teachers and students examined in this study, the types and range of data sources, and the types of analyses are presented in Figure 1. This cognitive map shows how the research processes of the ethnographic research cycle include cycles of focusing to specific events and of situating these events in broader social actions. Our logic-of-inquiry can be traced through this representation.
Identifying and Investigating a Cycle of Activity After collecting our ethnographic data, one of the first interpretive decisions made was the choice of a unit of analyses. Since the ethnographic data include videotape records that span 3 academic years, we unable to conduct a discourse analysis of all the oral and written language used in the classroom. Rather, we had to make selection decisions about what and how to represent our data. Green and Meyer (1991) use the term "cycle of activity" to denote a set of intercontextually tied activities initiated, enacted, and bound interactively by the participants with common thematic content (Floriani, 1993). We decided to consider a 2-week cycle of activity concerning the study of wave motion. This choice was based on our initial ethnographic information and experience through which we identified a key discourse event, the cumulative discourse activity of writing a scientific paper. The scientific paper was part of a teacher-assigned student project in which students designed, built, and tested musical instruments of their choosing.
SCIENCE AS SOCIOCULTURAL PRACTICES
Figure 1. "Physics of sound" cycle of activity: sources and types of educational research data and analysis.
KELLY AND CHEN
The cycle of activity, analytically named "physics of sound," spanned 2 weeks (November 6, 1995, through November 21, 1995) and was embedded in the larger classroom history and ethnographic study (Figure 2). Events and topics listed on the timeline were specific to the thematic content of the physics of sound. Topics involved in and comprising the cycle of activity in the class during this time included: simple harmonic motion (lecture and pendulum lab); wave theory (lecture, film, and demonstrations); graphing wave forms (lecture and labs); sound (lectures and sound labs: tuning fork and human voice; consonants and vowels; and fundamentals and harmonics); light (e.g., optics lab); and the musical instruments project. As shown in Figure 2, there was a range of instructional strategies (e.g., lectures, discussions, demonstrations, group work, use of media, laboratory experiments, and presentations by students to class) and interactional spaces (Heras, 1993) (e.g., whole class, small groups). Two key events (Gumperz & Cook-Gumperz, 1982) within the cycle of activity were noted as important, as they framed the events that counted as the cumulative task for the cycle of activity: the production of a musical instrument; and a corresponding technical paper written to describe the experiments conducted with the instrument. The next level of analysis involved creating structuration maps of the moment-to-moment interactions of the participants for each of the days in the cycle of activity. These maps were created by identifying the ways that the members of the classroom oriented to the topics and each other, and by noting the episodic nature of the instructional conversations marked interactionally by the members of the classroom (Green & Wallat, 1981; Lemke, 1990; Mehan, 1979). Following a methodology developed by Green and colleagues (Green & Dixon, 1993; Green & Meyer, 1991; Green & Wallat, 1981), we identified the different phases of activity for each event on each day. Phases of activity representing the ebb and flow of concerted and coordinated action among participants, and reflecting a common content focus of the group, were identified by examining the actors' talk. For example, the phase unit labeled "Introduction of Musical Instrument Project" is written in bold on Figure 2 and is the third of the four phases of activity comprising the key event (i.e., class on November 15, 1995). Within each phase, the participants structure the conversations and cue each other through their interactions by marking cohesive or thematically tied interactions to form a sequence unit. For example, the sequence units comprising the phase unit "Introduction of Musical Instrument Project" are represented in Figure 3. The sequencing of the talk and actions represented in Figure 3 show how through a set of discourse processes this teacher/researcher (first author) constructed a lecture that came to be an "introduction of musical instruments project." In this case, the lecture began with a sequence labeled "Passing out Project Instructions." This is one of many examples of how the oral discourse reinforced and was connected to written discourse. Analytically, this signaled to us the importance of both discourse forms and led us to consider both in subsequent analysis. For example, the relationship between oral and written texts in a subsequent event was visible when we examined the developing activity across sequence units. For the case represented in Figure 3, the lecture proceeded through a series of steps that oriented the students to the project in a particular way. After offering a rationale for the sequencing of the curriculum (sequence starting at 00:56:38), the teacher introduced the first suggested activity: The student groups were instructed to keep a record of their activities. As is often the case, the sequence labels were not sufficiently descriptive to be analytically exhaustive. We therefore included a record of notes and comments.
Discursively Accomplished Nature of Everyday Life The analysis of the classroom interactions at the phase and sequence level allowed us to identify a range of discourse processes, how these processes were connected with thematic de-
SCIENCE AS SOCIOCULTURAL PRACTICES
Figure 2. The "musical instruments" project situated in and across time.
KELLY AND CHEN
Figure 3. Structuration table representing the "Introduction of Musical Instruments Project" segment of activity with time-stamped sequence units and researcher notes and comments. velopment, and how particular events framed later events that occurred within the same cycle of activity [see Floriani (1993) for a related discussion of intercontextuality]. Our ethnographic analysis showed that the musical instruments project consisted of a practical component (e.g., instrument design, testing, analysis of data) and a written component (e.g., writing of a technical paper). Our theoretical position suggests that the framing of these events occurred discursively through the moment-to-moment interactions. Therefore, the next step in our analysis was to examine the ways that the events sequenced in the structuration maps were spoken and acted by the participants. To do this, we created transcripts of the talk and action for selected sequence units following theoretical sampling procedures: That is, we chose to sample those events that were signaled among the participants as important for accomplishing the tasks. The transcripts were created in message and action units. Message units are the smallest unit of linguistic meaning (Bloome & Egan-Robertson, 1993; Green & Wallat; 1981; Kelly & Crawford, 1996), defined by boundaries of utterances or social action that are identified through cues to contextualization, e.g., pitch, stress, intonation, pause structures, physical orientation, proxemic distance, and eye gaze (Gumperz, 1982, 1992). This was done directly from the videotape, as the nonverbal cues are important in identifying the message units. Action units are comprised of one or more message units that show a semantic relationship among message units and
SCIENCE AS SOCIOCULTURAL PRACTICES
represent an observed intended act by a speaker (Kelly & Crawford, 1996). Action units, like message units, are identified post hoc, and as researchers, we again considered the contextualization cues as well as the topical content of the talk. For the transcripts in this article, message units are separated by forward slash marks. Thus, the line numbers reflect the lines of texts for discussion purposes and not individual message or action units. Figure 3 shows that during the teacher/researcher's exposition about student record keeping (starting at time 00:56:56), there was an interruption (at time 00:57:10) involving student concerns about the time constraints of the proposed plan for the project which we labeled as a potential divergence [following Green & Wallat (1981)]. At 00:58:49, the teacher/researcher (TR) reoriented the topic to the keeping of records and then started a series of sequence units referring directly to the musical instruments project. These sequences were transcribed verbatim with researcher notes in italics as represented in Figure 4. In line 105 (Figure 4) the teacher/researcher took back the floor from the classroom teacher and began a series of instructions emphasizing components of the students' musical instruments projects. This emphasis centered on the teachers' instructions for the thematic and social aspects of the intended procedures and goals. The students were instructed to use outside resources e.g., the Internet and library (lines 106107) as they had done in a previous thematic project, i.e., the solar energy project (lines 109111); build an instrument as a group (lines 113115) in cooperation with the members of each small group (lines 115117); and use relatively
Figure 4. Transcript of teacher discourse relating to the suggested processes of the musical instruments project. Time stamp, sequence units, line number, and message units for selected sequences shown (compare with Figure 3).
KELLY AND CHEN
simple materials (lines 117120). In a second series of assertions, the teacher/researcher emphasized some of the epistemological goals through another set of instructions. The students were instructed to conduct experiments with the microcomputer that produces graphs (lines 121127), vary the parameters of the instruments (lines 130136), and conduct other experiments for comparison (line 137). This ended the sequence unit labeled "Changing the Instrument." The teacher/ researcher foreshadowed a discussion of the writing aspects of the project (starting at 01:01:20 in Figure 3), answered a student question about extra credit, and summarized the project tasks. The subsequent sequence was initiated by the classroom teacher (CT). Through our examination of the transcript of teacher discourse for this sequence, we were able to identify the shift in the presentations of the teachers' expectations about the project. The classroom teacher constructed, and was constructed as having, the role of organizer and manager of the classroom activities. This was evident in his talk (Figure 5). The teacher referenced time or due dates seven times; e.g., "there's not/a lot of time" (line 202), "itta/be due/on next/Tuesday" (lines 214
Figure 5. Transcript of teacher discourse reiterating and orienting to task. Time stamp, sequence units, line number, and message units for selected sequences are shown.
SCIENCE AS SOCIOCULTURAL PRACTICES
215), "so let's/let's/look at/go to the time table" (lines 226227). In addition, he repeated the steps of the project reinforcing the description of the teacher/researcher: "select an instrument" (lines 207208), "build it/or make it" (line 209), "test it out" (lines 210211), and "write a paper about it" (line 212). He ended this sequence by suggesting particular strategies for "the first order of business" (lines 230 231), i.e., a way for students to get started on their projects, including meeting in the respective groups and discussing ideas. The role of time manager again surfaced as he instructed the students that they "don't have to to any searching/in the library or Internet" (lines 235237) if the groups had "an idea already," and by suggesting that they take "5 or 6 minutes/or maybe 10 minutes" (lines 242243) to complete these tasks. Our examination of the particular ways the teacher/researcher and classroom teacher talked about the musical instruments project tasks revealed differences in their roles and responsibilities to the students and school. The tension between completing a scientific paper and completing a school task resurfaced later in the ethnographic analysis of the events.
Written Texts as Framing Artifacts To understand how the writing of the technical paper was situated in the over-time practices of the classroom, we examined how the task was constructed by teachers and students and how this particular genre of writing was one of the kinds of writing sanctioned in the conceptual physics course. Before presenting our analysis of how framing activities and texts came to define the technical paper, we identified the range of writing activities that were sanctioned in the course through taxonomic analysis (Spradley, 1980), presented in Figure 6. We classified the particular writing activities into four groups: classwork and assignments, creative writing, presentation of science through class thematic projects, and essays about the physics course. The
Figure 6. Taxonomic analysis of writing activities sanctioned in conceptual physics.
KELLY AND CHEN
writing activities found within each of these groups showed a range of writing genres with varying topics, purposes, types, audiences, and methods of production [cf., the expanded model of elements for writing to learn in science, proposed by Prain and Hand (1996)]. For example, students recorded notes from teacher lectures with accompanying overhead projection slides in their notebooks; they wrote a science fiction story about time travel; and they proposed ways for improving the teaching of the course. For the musical instruments project, the writing of the technical paper was one of two elements for writing to learn in science, as in addition to this task the students were instructed to record their activities while they created their musical instruments. Although all students wrote the technical paper, we have little evidence that students completed the record-keeping task. The school science task of writing a technical paper was modeled after a university science course that incorporates the writing of scientific papers as a central feature of instruction. For this university course, the professor created a text describing how to write a technical paper, which served as a basis for instructions to the students. We modified the instructions prepared for the university students and created a simplified version for the high school conceptual physics students, while maintaining the substantive components of the original text. Through our ethnographic analysis, we identified a key event on November 16, 1995, in the cycle of activity labeled in Figure 2 as "Introduction to Writing Technical Papers." In this event, the teacher/researcher talked through the modified instruction sheet about how to write a technical paper. Figure 7 shows the juxtaposition of the oral and written discourse about the
Figure 7. Framing the writing of the technical paper with oral and written discourse.
SCIENCE AS SOCIOCULTURAL PRACTICES
Figure 7. (continued) writing of these papers. The left-hand column is the set of discourse processes of the teacher/ researcher as he walked and talked the students through the main elements of the written instructions presented in the right-hand column. For analytical purposes, we have kept both the oral and written discourse in the sequences in which it was presented to the students, but in Figure 7 we have aligned the ways the teacher/researcher spoke about the issues with the writing text. Similar to the previous transcripts, the line numbers do not refer to message units or lines in the written document; they label the utterances and lines of texts for discussion purposes. The written instruction was comprised of four sections: orientation; general writing tips; presentation; and headings, including six subtopics--introduction, methods, observations, interpretation, conclusions, and using figures. However, these sections and subsections were not treated equally by the teacher/researcher as he reviewed the issues. We can read the left-hand column of teacher discourse as a way of signaling to the students emphasis on certain issues and
KELLY AND CHEN
not others. In both the oral and written discourse, the students are provided with a rationale for writing in science (lines 400 413). The written discourse mentioned audience and the uses of writing in science (lines 400405), while the oral discourse used contrasts with fiction and biography to make distinctions about writing in science (lines 405, 410411). The next section in the written section concerning "General Writing Tips" referred to the mechanics and regulations of writing in science. This section was given only a brief mention and the students were told that "you can read through that" (line 417). The next section on "Presentation" showed contrasts in the instructions, with the oral discourse qualifying the need for the papers to be "clearly typed or clearly hand written and thoroughly proofread" (lines 434 436 for written discourse) by suggesting that the students "can either type it [the paper] or handwrite(n) is neat (enough), either way is OK" (lines 435438). The section labeled "Headings" (line 440) described a format for the papers, and in doing so epistemologically positioned the students as writers, technical writing as a genre, and science as a discipline. There were five types of headings (introduction, methods, observations, interpretation, and conclusion), as well as a discussion of professional ethics
(lines 476 479) and use of figures. As a comprehensive discourse analysis of the written and oral forms is not possible here, we will review how the section on interpretation was written and talked into being (Green & Dixon, 1993; Tuyay et al., 1995). The interpretation section of the written discourse suggested the use of observations and an explicit invocation of personal "experience, insight, and knowledge" (lines 469471) to explain these observations and to "reason with evidence" (line 473). Emphasis on these issues was similarly evoked in the oral discourse, although in different ways. The teacher/researcher referenced a previous discussion about performing "tests" on the musical instruments to create data in the form of printed inscriptions (lines 466468). Interpretation and its relationship with data and figures were explicitly mentioned in lines 473 484. In this section, the students were instructed that their interpretation was to "discuss what the data means [sic]" and offered the suggestion of referring to their "waveform"--an inscription representing pressuretime relationship (see subsequent discussion). In this case, the teacher/researcher gave further specification by suggesting a possible relationship, that of "frequency/how that's related to pitch/how it's related to the change in your/um/your modification of your instrument" (lines 480482). Thus, in both the written and oral discourse processes, the importance of interpretation was signaled to the students. In both cases, the uses of data were invoked; in the case of the oral discourse, this was connected to the uses of figures--a point described subsequently in the written discourse (lines 485496). The written description of interpretation suggested personal creativity, but pointed to the use of reasoned evidence as "important." This was later suggested as the "main idea" and linked to interpretation in the oral discourse: "And the main idea now is that you're going to be using the data you collect in class and use that as evidence for your interpretations" (lines 490493). The comparison of oral and written discourse showed how certain aspects of the written text were emphasized and how, in the range of epistemological issues, certain interpretations were related to the particular tasks of the musical instruments project. These texts provided a framing for the students of what was to count as scientific practices in this cycle of activity. Consistent with the ethnographic research cycle, we used this analysis to pose a number of questions about how the student chose to engage in the practices signaled through the teachers' discourses. Did the students follow the suggestions of the writing genre recommended to them? In what ways were they reading these discourses and incorporating them into their own writing? How did they use evidence and in what ways? Before turning to the examination of the students' technical papers, we review further the physics of the experimentation suggested to the students.
SCIENCE AS SOCIOCULTURAL PRACTICES
The Physics of the Sound Project: Material and Semiotic Considerations Figure 8 shows the inscription of a sound wave, represented on the axes Sound (pressure) versus Time (milliseconds), and recorded through the Vernier software interface and software package entitled "Sound." This particular sound wave represents the sound of a commercially produced recorder. Even with an instrument of this tonal quality, the graph shows that there are a number of contingencies that make the decisions about what counts as a periodic wave and what the periodicity might be, as well as what counts as the amplitude of the wave, among other aspects, interpretive. For example, observers such as the high school physics students are faced with a set of problematic decisions to make: If there is periodicity, are there three peaks to a period or one? How would one know and decide what counts as a time period for a sound wave? In addition, the hypertext overlay is the computer's representation of the fast Fourier transform (FFT). This analysis, plotted as amplitude (pressure) versus FFT of sound (Data A), shows how the respective wave form is composed of a fundamental frequency and a set of harmonics. In this case, the arrow points to the fundamental and the computer reports a frequency range of 428442 Hz. The harmonics occur at roughly equal intervals. In this case, we can interpret the peak reported to be in the interval range of 864 877 Hz and the peak reported as 12991312 Hz to be harmonics. There are four other peaks that require further explanation. This type of analysis was introduced to the students through two lectures as noted in Figure 2 as "wave lecture" and "FFT (Fast Fourier Transform) lecture with demonstration," and three preparatory laboratory experiences labeled "sound lab: tuning fork and voice," "sound lab: consonants and vowels," and "sound lab: fundamental and harmonics." These experiences were part of students' preparation for using the recording equipment and programs, observing the wave forms and Fourier transform peaks, and drawing inferences. However, the complexities of the experiments and their interpretation depended on a number of contingencies, including the
Figure 8. Sound wave form inscription with associated fast Fourier transform (FFT) hypertext (created for commercial recorder with Vernier software "Sound," version 4).
KELLY AND CHEN
types of sound waves recorded and the extent to which explanation was sought. Therefore, the task facing the student groups for the analysis of their musical instrument was at once relatively simple (i.e., record a sound and interpret the computer representations) and very sophisticated. The problem space does not have clear limits; the students could have explored a range of possibilities such as amplitude and frequency variation (which many did), error ranges (as one did), and the theory of Fourier transform (which none did).
Analysis of Student Products: Appropriation of Scientific Discourse--Technical Paper as Artifact Our next analyses concerned the student technical papers. From the ethnographic point of view, these were treated as artifacts produced by a community under certain conditions for particular purposes. The logic-of-inquiry for our analysis of the students' writing of the technical papers was as follows. First, we entered each of the students' papers into a computer file, creating a case for each. Second, for each paper we proceeded to consider the central arguments that the authors were making. This provided us with a holistic view of what the students were arguing in their writing and what was being accomplished through their writing. Third, because the teachers indicated through both oral and written discourse the importance of the use of evidence as a central goal of the writing task, this led us to consider how the students were using evidence in their papers. To specify the arguments, we applied an argumentation analysis developed previously (Kelly et al., 1998b; Druker, Chen, & Kelly, 1997) following the Toulmin model (1958). Toulmin's layout of substantive arguments involves the warranting of a move from data to claims. He characterized the components of the argument as follows: Data (D) are the facts the proponent of the argument explicitly appeals to as a foundation for the claim. The claim (C) is the conclusion whose merits are sought to establish. The warrant (W) is the rules, principles, or inference license that demonstrate that the step to the claim from the data is a legitimate one. The strength of the warrant may be indicated by modal qualifiers (Q). The rebuttal (R) indicates the circumstances for which the general authority of the warrant is not merited. The backing (B) establishes the general conditions which give authority to the warrants. We used this general structure to consider the claims and supporting evidence (combination of data and warrants) in the students' papers. All claims referring to substantive issues and corresponding evidence were identified through this analytic procedure. Fourth, we considered only the evidenced claims and marked these claims and the evidence supporting them. As was found in Kelly et al. (1998b), not all claims in a discourse event will typically be supported with evidence. This should not be alarming, even for writing in science. At the propositional level, support for each and every assertion would lead to infinite regress, or at least regress to first principle in each case. Fifth, we jointly reviewed the evidenced claims and inductively generated a taxonomy of kinds of claims and kinds of evidence. We did this through multiple iterations of reviewing the data until we reached mutual agreements on each case. Sixth, following Latour (1987), we then considered the status of the claims. Latour argued that scientists typically argue from the particular contingencies of their actual experiments and try to construct facts at a more generalized level. In this way, they stack the facts, moving from low induction facts using the pictures, figures, and numbers to progressively higher induction, more abstract facts. According to Latour, the trick is to stack facts so that there are no gaps between layers that provide ways to unravel the arguments. To consider the facts presented in the students' papers, we reviewed all the kinds of claims and kinds of evidence and created two stacks, shown in Figures 9 and 10, with illustrative examples. Through examination of these
SCIENCE AS SOCIOCULTURAL PRACTICES
Figure 9. Kinds of evidenced claims, typical examples, and distribution across student papers (n 27).
KELLY AND CHEN
Figure 10. Kinds of evidence, typical examples, and distribution across student papers (n 27).
SCIENCE AS SOCIOCULTURAL PRACTICES
stacks and the respective distribution of kinds of claims and evidence, we began to get a picture of the students' use of evidence. Seventh, we created Table 1, showing the distribution of evidenced claims across technical paper subject headings suggested by the teachers. Five students did not follow the heading format, and although we included their papers in all other analyses, we did not consider the position of their claims in the paper for the creation of this distribution. The merits of their scientific arguments were not viewed negatively because of their formatting choice. Our analysis of the claims and evidence presented in the students' papers revealed a partial engagement with the intended task. The students generally followed the particular scientific genre presented to them; they made a large set of claims, supported some of these claims with evidence, and used the data inscriptions as figures in their papers. Consider the following example taken from the interpretation section of a ninth-grade student:
After comparing all the graphs we went to F.F.T [fast Fourier transform] and compared the highest and lowest peaks from all of them. When we did that for all the sounds we found out that the smallest drum has the highest pitch and the big drum has the lowest pitch. The big drum had 401 423 hertz and 5.9 amps and the smallest drum has 1485 1502 hertz and 7.4 amps. This is what we found out at the end that we wanted to find out in the beginning that the smallest drum has the highest pitch and the biggest drum has the lowest pitch. (Yasmin1)
In this example, the student constructed an argument by relating the physical structure of the set of drums which constituted her group's musical instrument, to the construct of pitch. The support for her central claim of the relationship of drum size to pitch was supported by the evidence of the frequency and amplitude measures of the wave forms. Other types of argument set out to make other sorts of claims. For example, consider the following argument by a 12th-grade student:
Even though all four of the rubber bands sounded the same to me, according to the computer sound program they weren't. I think that the thickness made more of a difference on the graph than the lenth [length] between the two nails. The thinnest rubber band was "B" and the graph showed a higher pressure reading. I think this happened because the thinner rubber band was a higher pitch and therefore created a higher pressure on the graph. When we tested rubber bands "C" and "D" they were very similar on the graph readings, this led me to believe that the lenth [length] between the two really did not mat-
Table 1 Distribution of evidenced claims for all student papers (n 22) across specified paper headings defined by technical writing genre
Heading in Technical Paper
No. of Evidenced Claims
KELLY AND CHEN
ter too much. Rubber band "A" and "D" graphs were similar and "D" was 6.2 inches shorter. So thickness, not lenth [length] is what affects the sound. (Leslie)
For Leslie's example, a number of assertions were made in support of the central argument regarding the relative importance of the length and thickness of the four struck rubber bands (labeled A, B, C, and D by her group) for variations in sound wave forms. In this case, evidence is marshaled to compare characteristics of the instrument, rather than the more direct examination of relationships of constructs (pitch) to wave forms as in Yasmin's example. In addition, she argued that pitch and amplitude ("higher pressure on the graph") were confounded variables. As suggested by ethnographers of interaction (Erickson, 1992), we examined each paper in detail and then compared across the papers for comparison purposes. Figure 9 shows that there was a broad range in the types of claims made by the students, from low induction claims about observations of their instruments (e.g., "We had four different drums, all of different sizes, to make a variance of noises") to claims that tied the inscriptions (e.g., "large spikes, medium peaks, and small waves") to constructs in physics such as frequency. The claim types with the highest frequency were those that involved physics constructs such as a harmonic frequency, tension, sound pitches, and amplitude. However, there were also claims referring to specific measured values, aspects of the instruments, and relationships among aspects of the representations. For the examples of Yasmin and Leslie, Yasmin's claim relating physical structure to a physics construct was classified as high induction (second from the top of the stack in Figure 9); Leslie's assertion concerning characteristics of the instrument was represented as a relatively low induction claim in Figure 9. Two kinds of evidenced claims were not included in the stack: those commenting on the projects and those addressing the section headings of the technical paper (Figure 9, bottom) as they were of a different sort, unrelated to the analysis. An examination of where the students chose to make their evidenced claims partially reveals how they understood the task of creating arguments with observations and interpretations. As suggested in the teacher discourse, the referential and explanatory aspects of the technical paper fell under the headings of observation and interpretation. Perhaps not surprisingly, the students' use of evidenced claims was predominately under these headings (Table 1). Thus, the engagement of the students in the task of using evidence was interpreted by them as aspects of describing observations and interpretations. Interestingly, the distinction between observation and interpretation is not a clean one for scientists. Indeed, Hanson (1958) and Kuhn (1970) suggested that all observation in science is interpretative. Thus, the relatively similar numbers of evidenced claims in the student papers under observations (n 27) and interpretations (n 35) may represent their struggle to identify what counted as observation and interpretation for this exercise and in science generally. The written instructions indicating that for observations one is to "Discuss what you observed . . ." (Figure 7, line 460) and for interpretations one is to "Discuss what the data means . . ." (Figure 7, lines 467468) may not have been readily distinguished by the students. This may indicate that some sophistication (however tacit) among the students that observations, as well as interpretations, need evidential support and are not merely read from data inscriptions. The range and variety of evidence used by the students also show a partial appropriation of the ways scientific writing was described to them. As shown in Figure 10, we considered the kinds of evidence, stacked these kinds from the most grounded in specifics to the most abstract, and then counted the distribution across all student papers. As presented in the teachers' discourse describing writing in science (Figures 5 and 7), the students were instructed to use their data inscriptions to make observations, and through this use offer explanations as their interpretations. The bins with the highest frequencies in Figure 10 are "gestures to the graph," "de-
SCIENCE AS SOCIOCULTURAL PRACTICES
scriptions of graphs," and "numerical values of graphs." While each of these represents different uses of evidence, the students can be seen as appropriately referring to their data sets in making their arguments. This use of evidence can be very grounded in particulars to support a specific claim, such as "The frequency of the fundamental was between 488 and 509 hertz, and its amplitude was 3.4" to rather vague invocations left to the reader to decipher, "The graphs show that the frequency of the rubber bands were almost exactly the same." The levels of evidence in this case are not measures of argument quality, but are intended to illustrate a range of evidence types provided by the students. Our stacking does not suggest that more or less abstract uses of evidence are more scientific or are a measure of student engagement in scientific discourse; all uses of evidence must be considered in the context of use given local conditions and purposes. In the examples of Yasmin and Leslie presented earlier, we classified Yasmin's use of evidence as "numerical values of graphs" as she specifically referred to numerical data, i.e., "The big drum had 401423 hertz and 5.9 amps and the smallest drum has 14851502 hertz and 7.4 amps." Leslie's use of evidence was classified by us as "description of graphs," as she spoke of the graphs but did not ground her description in specifics, e.g., "made more of a difference on the graph," and "they were very similar on the graph readings." Again, comparative strength of argument was not examined by us. In reviewing the students' use of evidence we found that not all students engaged in the writing of science in the same way or with the same argumentation strength. There were instances where we saw the use of the scientific genre in form but lacking much substance. The partial appropriation of the suggested discourse processes is perhaps not surprising, as the students did not experience all the sociocultural practices typically found in scientific communities when facts are constructed in written forms. For example, in this cycle of activity, the students were not asked to complete an informal and formal peer review nor editorial review. However, in scientific communities, the creation of the stacked claims leading to constructed facts is the product of long, rigorous, agonistic struggles. Only after such processes can claims be considered as potential new knowledge. Thus, the students were given some opportunities to engage in the practices of science, and not others. Therefore, not all of their claims nor evidence would necessarily be considered correct from a canonical physics point of view. For a final analysis of the students' papers, we considered some of the ways that the students interpreted their task and how this interpretation, while reasonable given the ways the tasks were framed through the oral and written discourses of the teachers, used a voice different from that typically found in an experimental scientific article. We found that in the introduction and conclusion sections of their papers, the students were most likely to consider audience issues and speak from and about Personal experience
. Consider the introductory remarks of the following student, Maria, as she orients a specific reader, i.e., "you:"
For our forth [fourth] group project we decided to build a wooden guitar. In this paper you will read about the data that we collected from our instruments, what it's [its] limits are, how we got the data, and where we got it. You will also read about our observations. After, we will write about what the data means [mean].
This practice was found in another student's introduction as well. Dennis, like Maria, built a guitar, but was in another student group:
Our group made a guitar for our project. We did the guitar because it was easy yet fun. I am writing this paper so the readers can know our results and how well our little guitar works. I am going to tell you, how we built our project.
KELLY AND CHEN
Patricia, a member of yet another group, again used the second person to address the reader and inform the reader about what follows in the paper.
In this lab/project I have explored the world of sound and the relationship between patterns on paper to the actual noise heard by the human ear. In this paper I will take you through the process of exploring sound and disecting [sic] its scientific meaning. The instrument our group built was simple but effective. We took a scrap of wood and put nails in a wide "V" shape and connecting opposite nails with rubber bands. The sound that was predicted was similar to a harp or a guitar. In the next page or so you will hear our results and difficulties along with modifications done with our instrument in order to explore further into the purpose and intreging [sic] facts about sound.
The conclusion sections were similarly spiced with remarks about the experience of participating in the cycle of activity, about how students felt about the project tasks, and influences on their personal affinity toward the projects. Here are three examples from Laura, as well as from Ken and Eileen, who were members of the same group.
This project having to do with sound was very interesting to me because I do alot [sic] of things involving music and sound. This project was very fun, and it was interesting to see which pitches make which graphs and what frequencies you get with different pitches. (Laura) Over all this project was ok but you did not give us enough time to do it in. You did not give us enough time to type it or to build it. After the first day you assumed us to be done building it, when we finished over the week end. The project was ok just needed more time. (Ken) I have learned that it is very simple to use the graphs on the computer than on paper, and I have also learned that the guitar has a lot of different shapes, and a lot of different pitches. I think this is the best group project we ever did so far because, it was fun making the object and we had a lot of class time to accomplish them. (Eileen)
The reference to how the students related to the task and commentary on the tasks was not signaled in the oral and written instructions to the students. However, while this was a divergence from the particular scientific genre suggested to them, the teachers had established a practice of having the students write about the course (Figure 6, "essays about the physics course"). Thus, the students took the writing of the conclusion as an opportunity to let the teachers know how they felt about the project, what they learned, and the difficulties they faced given the time constraints. Thus, the students were continuing a patterned practice consistent with this classroom community's norms and expectations (Santa Barbara Classroom Discourse Group, 1992; Zaharlick & Green, 1991).
Discussion We have argued in this article that discourse and interpretive processes are central to the creation of knowledge, both for scientists and students. By examining school science-in-the-making, we have identified how through patterned practices and concerted activity the teachers and students signaled situated interpretations of the nature of scientific activities. In this section, we discuss two related theoretical issues derived from our empirical findings: the nature of language use and its relationship to situated activities of a community. We then turn to how our analysis
SCIENCE AS SOCIOCULTURAL PRACTICES
of the students' papers suggested missing elements in the teaching practices of this class and what implications can be drawn for teaching of writing science. Our view of language use has been influenced by the work of the later Wittgenstein (1958). This view suggests that the meaning of a word, symbol, or construct is situationally defined by its use in a particular discourse practice (language game); that is, there is no essence of meaning, only how the signs and symbols fall into place relationally in particular instances of use. This suggests that stipulative definitions of complex constructs such as "observation" and "interpretation" and "evidence" represent only one context of use, and that practice using terms in multiple contexts is central to understanding. Thus, for the students in this study, the writing exercise can be interpreted as an opportunity to use scientific terminology, genres, and participate in social practices associated with science. Coming to understand the uses of data (in this case, sound wave forms) as evidence for a set of assertions (in this case, relationships of sound and physical changes) involves complex uses of language, meanings, and associated social practices. These practices must be learned through participation with more knowledgeable others (Vygotsky, 1962) and requires time and opportunities for both success and failure. The difficulties posed by using domain-specific knowledge in socially appropriate ways are formidable even for experienced members of the relevant communities of practice, as evidenced by the rejection rates of academic journals. In examining the students' work, we attempted to understand what counted as evidence for them, how they used evidence in their writing, and how their claims and evidence resembled that of scientists. Our goal in this study was to examine student understanding, not to assess the adequacy of their arguments. Thus, we did not attempt to pass judgment as to whether the students created logically coherent arguments, nor whether their reasoning was formally consistent, nor did we hold them to any other normative standard with the goal of determining whether they understood the use of evidence or whether they were rational. Learning how to use evidence in particular circumstances is not something we believe occurs in a short time period, nor does one achieve an understanding of evidence use that is timeless and context independent. In each situation, the interlocutor must read the social situation, make judgments about what counts in a particular community given the current situation, and draw from a repertoire of discourse processes and practices to attempt to act in socially appropriate ways (Gumperz, 1986; Heath, 1982; Heap, 1980, 1991). Our analysis of the students' technical papers suggested that there was only limited appropriation of scientific discourse. Some of the student papers maintained the form of the scientific genre suggested to them without using much evidence to support their assertions. Thus, the processes of school science-in-the-making were messy with successes and failures; some student texts created scientific sounding assertions, i.e., they stacked the claims in ways consistent with that of the disciplinary community (Latour, 1987). Other times, students spoke from personal experience, or could be interpreted as engaging in procedural display, i.e., performing in the processes of studenting, rather than the substantive processes of knowledge construction (Bloome, Puro, & Theodorou, 1989). Thus, like the scientists in their laboratories (Knorr-Cetina, 1995), the students faced interpretative flexibilities and opened up the genre of scientific writing to negotiation. Given this range of appropriation of scientific discourse and the ways the students diverted from this discourse, we now turn to the lessons learned by us through this experience and consider implications for the constraints facing students and teachers of science. The variations in the students' appropriation of written scientific discourses suggest that there were missing elements in the instructional practices. For example, there was no instruction explicitly concerning scientific discourse and norms of the scientific community. Thus, while we believe that there is always interpretative flexibility in the appropriation of scientific practices, exposure to and examination of what counts as an empirical claim, a theoretical as-
KELLY AND CHEN
sertion, or a consistent argument, would have made the implicit knowledge of the teachers explicit to the students. The unpacking of scientific norms may be crucial if classroom activities aimed at reproducing authentic scientific contexts are to be successful at affording opportunities for students engaging in scientific discourses. While the two teachers of this class have extensive experience with practicing and studying science, their knowledge of the social practices of science was not a content theme of the course. Therefore, identification of these scientific practices was left to be induced by students and only some students induced these practices, perhaps because of other experiences outside of this particular classroom. Thus, the curricular move away from propositional knowledge of physics (e.g., formulas and definitions characterizing sound waves) to a focus on scientific processes (e.g., creating a scientific argument with data inscriptions produced by students with technologies) still required strategies for making these processes explicit to students. The heterogeneous nature of the discourses of school science constructed by students, social mediators, texts, and technologies contribute to some of the constraints we have identified to learning scientific practices over a set of ethnographic studies (Chen, 1997; Kelly & Crawford, 1997). The lack of community-based accountability in the culture of the classroom has been a recurring theme. This issue is particularly difficult in school science, where the agonistic argumentation of scientific communities may be unacceptable on normative grounds (e.g., such practices are potentially discriminatory) [see Guzzetti (1998) and subsequent commentaries for an analysis of a gendered discourse practice using agonistic argumentation]. Providing a supportive community that is willing to weigh the value of respective ideas while maintaining value of speakers as individuals remains an unfulfilled goal. As described by Bazerman (1988), scientists are required to hold their ideas up to public scrutiny. This scrutiny sometimes leads to highly contentious rhetorical attacks, sometimes expanding beyond epistemological to professional and personal attacks [see Collins (1985), for analysis of the deconstruction of a researcher, and subsequently, his proposed "high fluxes of gravity waves"]. However, if public scrutiny is to be created in classrooms, it needs to be done so that students can maintain their individual integrity and still be able to discuss ideas in their own voices and present the evidence of their positions. Given the diversity in cognitive and Social Development
and the culturally patterned ways of speaking that students bring to the classroom, this becomes a difficult and delicate affair. Furthermore, scientific argument, in the best-case scenario, would be one of many competing goals of adolescents in schools who typically are concerned with peer culture and status, friendship groups, and other issues not directly related to learning science. We do not believe any small set of prescriptions for teaching should be drawn from empirical research, given the diversity of situations and socially constructed contexts for learning. Nevertheless, if we were to teach this course again, we would attempt several strategies to help students learn to use evidence in science writing. One strategy would be to provide a more explicit discussion about our goals as teachers and how we see this task as an opportunity for them to learn about a particular genre of writing--that is, discussion around and about the processes of doing science in school. By better informing the students that we expected scientific writing to include justified claims and that we expected this to be achieved over the course of the academic year, the students might have been better able to engage in these practices. For example, we used Toulmin's layout of argument for analysis purposes, but not pedagogically. This model for creating an argument might have assisted the students in coming to understand what counted as scientific argument. Our second strategy follows from the discussions about learning science and concerns creating a coherent epistemological theme throughout the course. In this study and others, we identified a range of activities with varying opportunities for learning. The change of frame across the various writing (Figure 5) might not have offered sufficient co-
SCIENCE AS SOCIOCULTURAL PRACTICES
herence for students to understand how to write science in some ways under certain conditions. A third strategy would involve more modeling of the argumentation practices. Creating a substantive argument from a set of inscriptions is a complex social activity, and the students would have probably been better served to see and hear a variety of examples from the teachers.
Note 1 Pseudonyms are used throughout, maintaining students' gender.
Author Note The research reported in this article was assisted in part by a grant from the Spencer Foundation. The data presented, the statements made, and the views expressed are solely the responsibility of the authors. An earlier version of this article, entitled, "The sound of music: Experiment, discourse, and writing of science as sociocultural practices," was presented at the annual meeting of the American Educational Research Association
, San Diego, CA, April 13 17, 1998. The authors thank Charles Bazerman, Julie Bianchini, and Judith Green for their helpful comments on an earlier version of the manuscript.
References Abell, S.K., & Flick, L.B. (1997). Who do we think we are anyway? Journal of Research in Science Teaching, 34, 425 427. Bazerman, C. (1988). Shaping written knowledge. Madison, WI: University of Wisconsin Press. Bianchini, J.A. (1997). Where knowledge construction, equity, and context intersect: Student learning of science in small groups. Journal of Research in Science Teaching, 34, 1039 1065. Bloome, D., & Egan-Robertson, A. (1993). The social construction of intertextuality in classroom reading and writing lessons. Reading Research Quarterly, 28, 305333. Bloome, D., Puro, P., & Theodorou, E. (1989). Procedural display and classroom lessons. Curriculum Inquiry, 19, 265291. Carlsen, W.S. (1991). Subject-matter knowledge and science teaching: A pragmatic approach. In J.E. Brophy (Ed.), Advances in research on teaching, Vol. 2 (pp. 115143). Greenwich, CT: JAI Press. Carlsen, W.S. (1992). Closing down the conversation: Discouraging student talk on unfamiliar science content. Journal of Classroom Interaction, 27, 1521. Carlsen, W.S. (1997). Never ask a question if you don't know the answer: The tension in teaching between modeling scientific argument and maintaining law and order. Journal of Classroom Interaction, 32, 1423. Cazden, C. (1988). Classroom discourse: The language of teaching and learning. Portsmouth, NH: Heinemann. Chen, C. (1997, March). Putting the ME in group MEmbership: Negotiating access into a community of high school scientists. Paper presented at the annual meeting of the National Association for Research in Science Teaching, Oak Brook, IL. Chen, C., & Crawford, T. (1998, April). Scientists in the making: An ethnographic investigation of students' access to scientific knowledge in an elementary classroom. Paper presented at the annual meeting of the National Association for Research in Science Teaching, San Diego, CA.
KELLY AND CHEN
Collins, H.M. (1985). Changing order: Replication and induction in scientific practice. London: Sage. Crawford, T., & Chen, C. (1997, Dec). Science for all Americans: Who counts as all? An investigation of literate practices and access to scientific knowledge. Paper presented at the annual meeting of the National Reading Conference, Scottsdale, AZ. Crawford, T., Chen, C., & Kelly, G.J. (1997). Creating authentic opportunities for presenting science: The influence of audience on student talk. Journal of Classroom Interaction, 32, 113. Druker, S., Chen, C., & Kelly, G.J. (1997, March). Introducing content to the Toulmin model of argumentation via error analysis. Paper presented at the annual meeting of the National Association for Research in Science Teaching, Oak Brook, IL. Edwards, A.D., & Mercer, N. (1987). Common knowledge. London: Methuen. Erickson, F. (1992). Ethnographic microanalysis of interaction. In M.D. LeCompte, W.L. Milroy, & J. Preissle (Eds.), The handbook of qualitative research in education (pp. 202224). San Diego: Academic. Emerson, R.M., Fretz, R.I., & Shaw, L.L. (1995). Writing ethnographic fieldnotes. Chicago: University of Chicago Press. Finkel, E. (1996). Making sense of genetics: Students' knowledge use during problem solving in a high school genetics class. Journal of Research in Science Teaching, 33, 345368. Floriani, A. (1993). The construction of understanding in a sixth-grade bilingual classroom. Linguistics & Education, 5, 241274. Gee, J., & Green, J. (1998). Discourse analysis, learning, and social practice: A methodological study. Review of Research in Education, 23, 119169. Green, J., & Dixon, C. (Eds.). (1993). Santa Barbara Classroom Discourse Group [Special issue]. Linguistics & Education, 5(3&4). Green, J., & Meyer, L. (1991). The embeddedness of reading in classroom life: Reading as a situated process. In C. Baker & A. Luke (Eds.), Toward a critical sociology of reading pedagogy (pp. 141160). Amsterdam: John Benjamins
. Green, J., & Wallat, C. (1981). Mapping instructional conversations: A sociolinguistic ethnography. In J. Green & C. Wallat (Eds.), Ethnography and language in educational settings (pp. 161205). Norwood, NJ: Ablex. Gumperz, J.J. (1982). Discourse strategies. Cambridge: Cambridge University Press. Gumperz, J.J. (1986). Interactive sociolinguistics in study of schooling. In J. CookGumperz (Ed.), The social construction of literacy (pp. 4568). London: Cambridge University Press. Gumperz, J.J. (1992). Contextualization and understanding. In A. Duranti & C. Goodwin (Eds.), Rethinking context (pp. 229252). Cambridge, UK: Cambridge University Press. Gumperz, J.J., & Cook-Gumperz, J. (1982). Introduction: Language and the communication of social identity. In J.J. Gumperz (Ed.), Language and social identity. Cambridge, UK: Cambridge University Press. Guzzetti, B. (1998). Case studies in conceptual change: Contextual influences on learning with text. In B. Guzzetti & C. Hynd (Eds.), Perspectives on conceptual change: Multiple ways to understand knowing and learning in a complex world (pp. 36). Mahwah, NJ: Erlbaum. Hanson, N.R. (1958). Patterns of discovery. Cambridge, UK: Cambridge University Press. Heap, J.L. (1980). What counts as reading: Limits to certainty in assessment. Curriculum Inquiry, 10, 265292. Heap, J. (1991). A situated perspective on what counts as reading. In C. Baker & A. Luke
SCIENCE AS SOCIOCULTURAL PRACTICES
(Eds.), Towards a critical sociology of reading pedagogy (pp. 103140). Philadelphia, PA: John Benjamins. Heath, S. (1982). Ethnography in education: Defining the essential. In P. Gilmore & A. Glatthorn (Eds.), Children in and out of school (pp. 3358). Washington, DC: Center for Applied Linguistics. Heras, A.I. (1993). The construction of understanding in a sixth-grade bilingual classroom. Linguistics & Education, 5, 275299. Hewitt, P.G. (1992). Conceptual physics: The high school physics program (2nd ed.). Menlo Park, CA: Addison-Wesley. Hicks, D. (1995). Discourse, learning, and teaching. Review of Research in Education, 21, 4995. Jasanoff, S. Markle, G.E., Peterson, J.C., & T. Pinch, T. (Eds.). (1995). Handbook of science and technology studies. Thousand Oaks: Sage. Kelly, G.J., Carlsen, W.S., & Cunningham, C.M. (1993). Science education in sociocultural context: Perspectives from the sociology of science. Science Education, 77, 207220. Kelly, G.J., Chen, C., & Crawford, T. (1998a). Methodological considerations for studying science-in-the-making in educational settings. Research in Science Education, 28, 2349. Kelly, G.J., & Crawford, T. (1996). Students' interaction with computer representations: Analysis of discourse in laboratory groups. Journal of Research in Science Teaching, 33, 693707. Kelly, G.J., & Crawford, T. (1997). An ethnographic investigation of the discourse processes of school science. Science Education, 81, 533559. Kelly, G.J., Crawford, T., & Green, J. (1997). Common task and uncommon knowledge: Dissenting voices in the discursive construction of physics across small laboratory groups. Manuscript submitted for publication. Kelly, G.J., Druker, S., & Chen, C. (1998b). Students' reasoning about electricity: Combining performance assessments with argumentation analysis. International Journal
of Science Education, 20, 849871. Kelly, G.J., & Green, J. (Eds.). (1997). Journal of Classroom Interaction, 32(2). [Special issue]. Kelly, G.J., & Green, J. (1998). The social nature of knowing: Toward a sociocultural perspective on conceptual change and knowledge construction. In B. Guzzetti & C. Hynd (Eds.), Perspectives on conceptual change: Multiple ways to understand knowing and learning in a complex world (pp. 145181). Mahwah, NJ: Erlbaum. Klaassen, C.W.J.M., & Lijnse, P.L. (1996). Interpreting students' and teachers' discourse in science classes: An underestimated problem? Journal of Research in Science Teaching, 33, 115 134. Knorr-Cetina, K. (1995). Laboratory studies: The cultural approach to the study of science. In S. Jasanoff, G.E. Markle, J.C. Peterson, & T. Pinch (Eds.), Handbook of science and technology studies (pp. 140166). Thousand Oaks, CA: Sage. Kuhn, T.S. (1970). The structure of scientific revolutions (2nd ed.). Chicago: University of Chicago Press. Latour, B. (1987). Science in action: How to follow scientists and engineers through society. Cambridge, MA: Harvard University Press. Latour, B., & Woolgar, S. (1986). Laboratory life: The construction of scientific facts. Princeton, NJ: Princeton University
Press. Lemke, J.L. (1988). Genres, semantics, and classroom education. Linguistics and Education, 1, 8199. Lemke, J.L. (1990). Talking science: Language, learning and values. Norwood, NJ: Ablex.
KELLY AND CHEN
Mehan, H. (1979). Learning lessons: Social organization in the classroom. Harvard University Press: Cambridge, MA. Millar
, R. (Ed.). (1989). Doing Science: Images of science in science education. London: Falmer. Moje, E.B. (1995). Talking about science: An interpretation of the effects of teacher talk in a high school science classroom. Journal of Research in Science Teaching, 32, 349371. Mukerji, C. (1989). A fragile power: Scientists and the state. Princeton, NJ: Princeton University Press. Prain, V., & Hand, B. (1996). Writing for learning in secondary science: Rethinking practices. Teaching and Teacher Education, 12, 609626. Richmond, G., & Striley, J. (1996). Making meaning in classrooms: Social processes in small-group discourse and scientific knowledge building. Journal of Research in Science Teaching, 33, 839858. Roth, W.-M. (1995). Inventors, copycats, and everyone else: The emergence of shared (arti)facts and concepts as defining aspects of classroom communities. Science Education, 79, 475502. Roth, W.-M. (1996). Knowledge diffusion in a grade 4 5 classroom during a unit on civil engineer
ing: An analysis of a classroom community in terms of its changing resources and practices. Cognition & Instruction, 14, 179220. Roth, W.-M., & Lucas, K.B. (1997). From "truth" to "invented reality": A discourse analysis of high school physics students' talk about scientific knowledge. Journal of Research in Science Teaching, 34, 145179. Roth, W.-M., McGinn, M.K., & Bowen, G.M. (1996). Applications of science and technology studies: Effecting change in science education. Science, Technology, & Human Values, 21, 454484. Santa Barbara Classroom Discourse Group. (1992). Constructing literacy in classrooms: Literate action as social accomplishment. In H. Marshall (Ed.), Redefining student learning: Roots of educational change (pp. 119150). Norwood, NJ: Ablex. Spradley, J.P. (1979). The ethnographic interview. Fort Worth, CO: Harcourt Brace Jovanovich. Spradley, J.P. (1980). Participant observation. New York: Holt, Rinehart, & Winston. Toulmin, S. (1958). The uses of argument. Cambridge, MA: Harvard University Press. Traweek, S. (1988). Beamtimes and lifetimes: The world of high energy physicists. Cambridge, MA: Harvard University Press. Tuyay, S.; Jennings, L., & Dixon, C. (1995). Classroom discourse and opportunities to learn: An ethnographic study of knowledge construction in a bilingual third-grade classroom. Discourse Processes, 19, 75110. Vygotsky, L. (1962). Thought and language. (A. Kozulin, Trans.). Cambridge, MA: MIT Press. Warren, B., Roseberry, A., & Conant, F. (1994). Discourse and social practice: Learning science in language minority classrooms. In D. Spencer (Ed.), Adult biliteracy in the United States. Washington, DC: Center for Applied Linguistics and Delta Systems. Warren, B., & Roseberry, A.S. (1995). "This question is just too, too easy!" Perspectives from the classroom on accountability in science. (Rep. No. NCRCDSLL/CALRR14). East Lansing, MI: National Center for Research on Teacher Learning. (ERIC Document Reproduction Service No. ED 390658) Wittgenstein, L. (1958). Philosophical investigations (3rd ed.). (G.E.M. Anscombe, Trans.). New York: Macmillan.
SCIENCE AS SOCIOCULTURAL PRACTICES
Yore, L.D., Holliday, W.G., & Alvermann, D.E. (Eds.). (1994). The reading-science learning-writing connection. [Special issue]. Journal of Research in Science Teaching, 31(9). Zaharlick, A., & Green, J. (1991). Ethnographic research. In J. Flood, J. Jensen, D. Lapp, & J. Squires (Eds.), Handbook for research in teaching of the English language arts (pp. 205 225). New York: Macmillan.