Ethnomethodology at Xerox PARC: When the Study of Ordinary Action Entered the Cathedral of High Technology
Wayne Martin Mellinger, Ph.D.
Revised 9-26
Abstract
This essay revisits the ethnomethodology/conversation analysis (EM/CA) program at Xerox PARC and its sister laboratory, Rank Xerox EuroPARC, to show how workplace studies helped remake human-computer interaction from the shop floor up — and to complicate that familiar story by examining what happened to ethnomethodology itself when it acquired authority inside a corporate research institution. The essay contrasts cognitive, plan-centered models of action with situated action: the ways people accomplish tasks through locally organized practices, talk, documents, technologies, and material environments. Microanalytic studies of photocopiers, service work, repair technicians, documents, and communication systems demonstrated that coordination, troubleshooting, and accountability are not simply designed into formal systems but continuously produced through interaction.
Alan Blackwell, Mark Blythe, and Jofish Kaye's 2017 reconstruction of ethnomethodology at EuroPARC considerably complicates this success story. Drawing on interviews with more than twenty participants, they ask what kind of intellectual and institutional authority ethnomethodology acquired within Xerox and use a deliberately provocative High Church/Low Church analogy to describe the resulting tensions. Ethnomethodology performed a “Low Church” mission by directing attention toward ordinary users and situated work, while sometimes acquiring “High Church” characteristics of doctrinal guardianship, ritualized analytic practice, difficult canonical texts, and disciplinary boundary policing. The history therefore has a double legacy. EM/CA became a powerful form of practical intelligence for technology design, but its Xerox career also provides a cautionary sociology of knowledge: even an intellectual movement founded upon resistance to formalism and externally imposed categories can develop its own forms of hierarchy, boundary maintenance, and exclusion. Both legacies remain highly relevant in the era of artificial intelligence.
Keywords: ethnomethodology; conversation analysis; Xerox PARC; EuroPARC; Lucy Suchman; situated action; workplace studies; CSCW; HCI; artificial intelligence; work practice; sociology of knowledge
I. An Unlikely Encounter
There is something wonderfully improbable about the history of ethnomethodology at Xerox PARC.
Ethnomethodology emerged from Harold Garfinkel's sustained dissatisfaction with the explanatory practices of mainstream sociology. Where conventional sociology often began with social structures, roles, norms, or theoretical categories, Garfinkel insisted upon returning to the practical methods through which members themselves produce recognizable social worlds. Harvey Sacks subsequently demonstrated, through meticulous examination of recorded interaction, how much social organization could be discovered by looking closely at what people actually did rather than beginning from an analyst's representation of what they must have been doing.
Neither intellectual project would initially have seemed destined to become important to the design of computers.
Yet during the 1980s and 1990s, this radically empirical sociology entered one of the most celebrated technological environments in the world: Xerox's Palo Alto Research Center. It subsequently became deeply entangled with Xerox's European laboratory, EuroPARC, in Cambridge, England. These were institutions populated by computer scientists, artificial-intelligence researchers, cognitive psychologists, engineers, designers, anthropologists, sociologists, and eventually a distinctive collection of researchers concerned with what Xerox came to call “Work Practices.”
The encounter mattered because these communities began from very different conceptions of intelligent action. Artificial intelligence and cognitive science frequently represented human conduct through plans, rules, goals, internal representations, and information-processing systems. Ethnomethodology began somewhere else. It asked how intelligible action was actually assembled within particular circumstances. The difference was not merely methodological. It concerned what human action is.
This is why the history of ethnomethodology at Xerox should not be reduced to the familiar claim that ethnographers taught engineers to observe users. Something much more interesting occurred. A social science skeptical of abstract representations of human conduct entered institutions whose central task was precisely to build formal representations of human conduct. The collision changed both sides.
II. Lucy Suchman and the Problem of the Plan
Lucy Suchman stands at the center of this history.
When Suchman began working at PARC, artificial intelligence was strongly influenced by models in which intelligent action involved planning, symbolic representation, problem solving, and rule-governed execution. The attraction of such models for computer science is obvious. If intelligent action can be represented formally, perhaps it can also be computationally reproduced.
Suchman did something characteristically ethnomethodological. She watched what people actually did.
Her celebrated study involved videotapes of people attempting to use an advanced Xerox photocopier. The machine had been presented as sophisticated but straightforward, advertised with a figure in “the white lab coat of the scientist/engineer” reassuring users that all that was required was to “press the green [start] button.” Yet actual users repeatedly became confused, complaining that the machine was “too complicated.” The conventional explanation might have located the problem inside the user: insufficient technical knowledge, an inadequate mental model, failure to follow directions. Suchman instead examined the unfolding encounter.
Users did not arrive carrying a detailed internal program that they mechanically executed. They approached with general purposes. They pressed something. The machine responded. They interpreted what had happened. Sometimes they hesitated. Sometimes they misunderstood. They reformulated what they were doing in light of the machine's response. Action developed sequentially and contingently, on the fly rather than according to plan.
This became the foundation for her distinction between plans and situated actions. The argument is sometimes misunderstood as saying that people do not plan — of course they do. The deeper claim is that plans do not determine situated conduct in anything like the way a computer program determines execution. Plans are themselves practical resources. Their relevance, interpretation, adequacy, and revision emerge within circumstances that cannot be completely specified beforehand.
Anyone familiar with conversation analysis will recognize the parallel. A conversation is not ordinarily produced by two speakers independently executing complete internal scripts. One action establishes an environment for another. Understandings are displayed through subsequent turns. Misunderstandings become observable through their consequences. Participants initiate and accomplish repair. Meaning is inseparable from sequential position.
Suchman extended this sensibility to human-machine interaction. The photocopier could process inputs but could not participate competently in the rich interactional procedures through which human beings ordinarily establish mutual intelligibility. It could not adequately discover what kind of misunderstanding had occurred. It could not reformulate itself in light of the precise trouble emerging in the sequence. It could not use gaze, gesture, hesitation, history, shared background knowledge, or the enormous range of contextual resources through which human beings ordinarily repair interaction.
The problem was therefore not simply interface design. It was an interactional problem — and, as Blackwell, Blythe, and Kaye note, it was already circulating as a live controversy at PARC just as EuroPARC was being founded in 1987, so the new European laboratory inherited it as unfinished business rather than settled history.
III. From “The User” to Work Practice
Once one begins from situated action, the familiar figure of “the user” becomes unstable. Technology design often constructs an abstract user in advance: someone with identifiable needs, tasks, capabilities, goals, and predictable sequences of action. Workplace studies asked instead what actual people did with technologies in actual environments. The difference sounds modest. It is not.
Formal descriptions of work almost invariably leave out much of the practical activity that makes the work possible. Employees interrupt one procedure to solve another problem. They ask colleagues questions. They maintain scraps of paper beside computers. They develop unofficial categories. They keep information visible on desks. They put callers on hold. They move between databases. They ignore instructions that do not fit the current case. They develop shortcuts, remember precedents, translate classifications into particular circumstances, and repair incompatibilities between systems. From the standpoint of formal management, these may appear to be deviations from proper procedure. From an ethnomethodological standpoint, they may be the procedure by which the work actually gets done.
This insight became fundamental to workplace studies at Xerox. Jack Whalen, Marilyn Whalen, Peggy Szymanski, Jeanette Blomberg, Julian Orr, Lucy Suchman, Richard Harper, Graham Button, Paul Luff, Bob Anderson, and others differed in training, location, and analytic emphasis, but collectively helped establish technologically mediated work as an object worthy of close sociological investigation.
There is an important UCSB genealogy within this larger history. Jack and Marilyn Whalen were trained in the UC Santa Barbara EM/CA tradition shaped by Don Zimmerman, and Peggy Szymanski emerged from the broader Language, Interaction and Social Organization environment at UCSB that Zimmerman helped establish. This connection is not merely inferred from shared training. In their own introduction to Making Work Visible, Szymanski and Whalen name both Harold Garfinkel and Don Zimmerman among the scholars who “collaborated with PARC and XRCE researchers or spent significant time in these labs,” alongside Chuck and Candy Goodwin, Paul Drew, Geoff Raymond, and Ron and Suzie Wong Scollon. Blackwell, Blythe, and Kaye's reconstruction of EuroPARC — built from more than twenty interviews and archival research — does not mention Zimmerman, a reminder that no single source captures the whole network. But Szymanski and Whalen's own account, written from inside the Xerox workplace-studies tradition, establishes a direct and documented link between Zimmerman, Garfinkel, and the Xerox labs — one distinct from, and considerably more secure than, any claim that Zimmerman personally trained the Xerox ethnographers.
Intellectual history rarely travels through one straight genealogical line. The Xerox development involved UCLA ethnomethodology, conversation analysis, anthropology, British ethnomethodology, participatory design, cognitive science, corporate management, engineering, and highly contingent networks of personal recruitment and institutional sponsorship. There was no master plan for the rise of ethnomethodology at Xerox. That fact is itself rather ethnomethodological.
IV. Making Work Visible
The workplace studies movement turned the analytic machinery of EM/CA toward organizational environments saturated with technology.
Whalen and his colleagues, for example, studied service representatives navigating telephone conversations while simultaneously working across computer screens, paper records, organizational categories, and colleagues located nearby — the subject of their 2002 study with Kathryn Henderson, “Improvisational Choreography in Teleservice Work.” The resulting workscape was neither purely technological nor purely conversational; it was an ecology of coordinated practice. A customer says something. A representative searches a database. The available category does not quite fit. The representative asks a colleague. One computer system contains information that another lacks. The customer remains on the line. An organizational rule becomes relevant. The representative translates between the language of the customer and the categories through which the organization recognizes the case. None of these operations is incidental — together they constitute the service encounter.
This perspective has profound implications for design. Workflow systems are commonly produced from idealized descriptions of how work should proceed. Yet competent workers constantly encounter contingencies that no workflow can completely anticipate. A system that attempts to eliminate improvisation may therefore eliminate the very practical intelligence that keeps an organization functioning. The design question changes accordingly: instead of asking how technology can make workers follow the correct sequence, one asks how technology can provide resources through which competent workers can manage contingencies. That is a very different philosophy of technological design.
V. Julian Orr and the Narrative Order of Technical Work
Julian Orr's research on Xerox copier technicians, conducted through the Institute for Research on Learning, remains one of the clearest examples. The formal organization possessed service manuals, diagnostic procedures, specifications, training programs, and technical databases. Technicians nevertheless encountered particular machines in particular states presenting particular combinations of trouble. The map was not the territory.
Orr followed technicians into the field and discovered that technical competence was inseparable from social interaction. Technicians talked about machines. They exchanged stories about especially difficult repairs. They reconstructed what they had tried, what had failed, what had finally worked, and what peculiar combination of symptoms might matter in another case. These “war stories,” published in Talking About Machines (1996), were not colorful additions to the actual technical work. They were part of the technical work. Stories circulated diagnostic knowledge unavailable in formal documentation; they allowed technicians to compare cases, established occupational competence, transmitted experience, and preserved unusual solutions. Organizational knowledge did not reside exclusively in the organization's official records — it circulated through a community of practice, a system the later Eureka database, built by Whalen and colleagues, deliberately preserved rather than replaced.
The implication reaches well beyond photocopier repair. Every organization contains a gap between its official representation of work and the situated practices through which members make that work possible. Technology can either recognize that gap or attempt unsuccessfully to abolish it.
This work would eventually figure in the ethnomethodological canon itself. Garfinkel, in a self-published photobook titled Garfinkel's Study, once circulated a bibliography of thirty-two texts he presented as a “curriculum of ethnomethodology.” Eight of them were by Xerox employees — Orr and Suchman at PARC, and Anderson, Button, Luff, and Dourish at EuroPARC. That is an extraordinary institutional fact, one this essay returns to below: Xerox was no longer simply borrowing ethnomethodology. For a period, Xerox had become one of the places where ethnomethodology was being produced.
VI. EuroPARC: The Experiment Crosses the Atlantic
The story becomes more institutionally complicated at Rank Xerox EuroPARC, founded in Cambridge, England, in 1987 — for political and financial reasons as much as strategic ones; a 1985 economic consultancy report had identified “the Cambridge Phenomenon,” and Xerox wanted a European R&D foothold in an English-speaking market. Tom Moran was dispatched from PARC to find premises. EuroPARC initially participated in the emerging technological vision of ubiquitous computing: projects such as AMODEUS and RAVE brought together cognitive science, computer science, interactive media, and experimental communication systems, alongside the Active Badge location system adapted from nearby Olivetti Research and the Digital Desk.
Suchman's critique entered this setting while these questions were still unsettled. John Seely Brown, an early advocate of that critique, was instrumental in sending Suchman to the UK to recruit like-minded researchers, and her photocopier-study collaborator, Austin Henderson, became a regular envoy between the two laboratories. Over time, ethnomethodologists became unusually influential within EuroPARC — a shift formalized institutionally in 1992, when the laboratory's then-director, ethnomethodologist Bob Anderson, hired fellow ethnomethodologist Graham Button, of the Manchester school, to establish a dedicated Studies of Technology, Organizations, and Work group. Button would go on to succeed Anderson as director.
And here Blackwell, Blythe, and Kaye's history becomes indispensable. Their article does not portray EuroPARC as a clean intellectual succession in which superior empirical sociology defeated naïve cognitive science. It was messier. There were doctrinal battles, struggles over rigor, arguments about what counted as science, conflicts between ethnomethodologists and cognitive researchers, misunderstandings about what ethnomethodology even meant, corporate pressures concerning the practical usefulness of research, and highly consequential personalities. Eventually ethnomethodologists themselves became directors of EuroPARC — rather than the cognitive scientists, or the participatory-design researchers with whom the laboratory also maintained close ties through Liam Bannon's group at Aarhus.
That development presents an extraordinary sociological problem: how did one of sociology's most persistently marginal and anti-establishment movements acquire something approaching establishment authority inside an elite corporate research laboratory? And the answer, on Blackwell, Blythe, and Kaye's evidence, is bounded by an ending. Rank Xerox EuroPARC was closed in the late 1990s, its remaining staff transferred to a Xerox research center in Grenoble, France, later renamed Xerox Research Centre Europe. According to the Grenoble laboratory's own historical archive, as Blackwell, Blythe, and Kaye read it, EuroPARC's most substantial acknowledged legacy was precisely the Work Practices research conducted through ethnomethodology — the RAVE, Active Badge, and AMODEUS projects that had once defined the lab's ubicomp ambitions receive comparatively little mention by comparison. Ethnomethodology outlasted the very agendas it had been recruited to serve. But the laboratory that had housed its authority did not survive, and the authority did not automatically travel with the people who once held it.
VII. High Church and Low Church
Blackwell, Blythe, and Kaye answer the puzzle of ethnomethodology's rise at EuroPARC through an intentionally playful but surprisingly productive analogy: the distinction between High Church and Low Church Anglicanism. They are not suggesting that ethnomethodologists at Xerox were literally religious. Their interest is organizational. High Church represents doctrinal authority, hierarchy, disciplined interpretation, ritual, and institutional legitimacy. Low Church represents outward engagement, ordinary people, practical experience, and comparatively decentralized authority.
Ethnomethodology at Xerox occupied both positions simultaneously. Its substantive orientation was remarkably Low Church: go to the workplace, watch ordinary people, do not privilege the designer's account, do not assume that official categories describe the situation adequately, examine how members themselves make circumstances intelligible, study users rather than constructing them theoretically from above. Yet the intellectual culture surrounding ethnomethodology could sometimes become distinctly High Church. There were canonical texts, demanding procedures for competent reading, arguments concerning who properly understood Garfinkel, sharp distinctions between ethnomethodology and allegedly conventional sociology. Video-analysis sessions could appear ritualized to outsiders. Garfinkel's famously difficult prose became something through which disciplinary competence itself could be demonstrated — Graham Button reportedly welcomed the difficulty, since demanding texts required discipline and prolonged close reading, and one of Garfinkel's own students told Blackwell, Blythe, and Kaye that it took years before he could read the first paragraph of Studies in Ethnomethodology “with a fairly good sense of comprehension.” Steve Woolgar apparently captured the phenomenon succinctly by describing Button as a “high-church ethnomethodologist.” Decades earlier, the philosopher Ernest Gellner had made a related, considerably less generous argument, dismissing ethnomethodology altogether as “the re-enchantment industry” and a symptom of Californian “subjectivism.”
The paradox is delicious: a sociology founded upon studying ordinary people's practical reasoning developed unusually demanding procedures for determining whether professional analysts were themselves reasoning properly. But the paradox deserves more than a joke. Ethnomethodology has always faced a genuine danger of assimilation. “Study users.” “Do qualitative research.” “Watch people at work.” “Do ethnography.” All may be useful. None necessarily constitutes ethnomethodology. The distinctive analytic question remains: by what practical methods are members producing this scene as the recognizable social phenomenon that it is? That question is considerably harder than simply observing users.
VIII. Authority, Conflict, and the Price of Intellectual Culture
The High Church analogy also prevents us from romanticizing the Xerox experiment.
Blackwell, Blythe, and Kaye report vigorous intellectual combat. Suchman's presentations were remembered as something of a “moral crusade,” in which she pressed not merely for agreement but for an acknowledgment of moral rightness — and attracted vigorous counter-attacks in turn. EuroPARC researchers could “lay into” one another, and founding PARC manager George Pake described the culture with magnificent managerial understatement: researchers could be “rather intemperate in their criticisms of each other's ideas.” Some of this intellectual combat undoubtedly produced excellent work. But intensity has costs.
An especially disturbing detail in the Blackwell account concerns several younger women in the laboratory. Their reconstruction of EuroPARC design discussions found that a young female physicist sometimes emerged as the strongest advocate for users — and that she was also one of several young women who eventually “burned out,” leaving both EuroPARC's “macho” culture of argument and research itself. That detail belongs in the history, because an intellectual tradition committed to recovering the overlooked competencies of ordinary members should be especially willing to examine the interactional organization of its own authority: whose intellectual style gets recognized as rigor, and who does not survive the argument required to demonstrate it. Ethnomethodology should not be exempt from ethnomethodological scrutiny; exemption would contradict its most important commitments.
IX. Documents, Materiality, and Organizational Intelligence
The EuroPARC tradition also extended attention beyond talk and human-machine interaction into documents and material infrastructures.
Richard Harper's Inside the IMF (1998) — Harper also spent a year at UC Santa Barbara working with Zimmerman — examined an institution commonly imagined as the epitome of abstract quantitative rationality. Yet its actual organizational life depended upon concrete documentary practices. Documents circulated. People annotated them. Lists established distributions. Reports moved through sequences of review. Documents displayed who knew what, who was responsible, who possessed authority, and what stage of an organizational process had been reached. A document was not merely a container carrying information from one mind to another. It was an organizational object.
This insight became especially important in Harper and Abigail Sellen's The Myth of the Paperless Office (2002). Predictions of the disappearance of paper had assumed that paper was simply an inferior medium for storing information. But workers did much more with paper: they spread documents across desks, stacked them, placed one beside another, marked them, left them visible as prospective reminders, carried them into meetings, spatially organized concurrent projects. The material characteristics of paper were embedded in practical routines. The persistence of paper therefore represented neither irrationality nor resistance to technological progress. Designers had misunderstood the work paper was doing. Again the ethnomethodological lesson was simple but radical: look before explaining.
X. The Reciprocal Transformation of Ethnomethodology
We usually ask what ethnomethodology contributed to Xerox. The reciprocal question is at least as interesting: what did Xerox contribute to ethnomethodology?
It placed ethnomethodologists in sustained encounters with engineers, technologies, customers, managers, markets, organizations, screens, documents, communication systems, technical repair, design problems, and corporate demands for practical relevance. That greatly expanded ethnomethodology's empirical territory. And the influence became sufficiently deep to appear inside Garfinkel's own later representation of the field — the eight-of-thirty-two curriculum fact noted above. Few corporations have achieved anything remotely comparable within a social-science tradition. This changes the historical interpretation. Xerox was not simply a company that employed ethnomethodologists. For a period, it became one of the institutional locations in which ethnomethodology itself was being developed. Industrial research and academic sociology became mutually constitutive.
XI. Diffusion Without Doctrine
But institutional success created another problem.
Ethnography became increasingly common within HCI. Ethnomethodology did not necessarily accompany it. Paul Dourish, a EuroPARC alumnus, summarized the historical irony memorably: “HCI needed ethnography, and it got ethnomethodology.” Initially the two became closely associated. But a bibliometric search of the ACM Digital Library that Blackwell, Blythe, and Kaye conducted shows that, over time, ethnographic work expanded while the proportion explicitly grounded in ethnomethodology steadily declined.
The confusion this reflects runs deeper than citation counts. Andy Crabtree's widely used textbook Designing Collaborative Systems — with a cover blurb from EuroPARC director Graham Button himself — barely uses the word “ethnomethodology” inside its own pages, even as its concluding chapter claims to demonstrate “ethnomethodologically informed ethnography in the creative process of design.” Blackwell, Blythe, and Kaye read this slippage as revealing: many HCI practitioners assumed, understandably but incorrectly, that “ethnomethodology” simply names the professional research methodology that “ethno” researchers use — the opposite of Garfinkel's own insistence that ethnomethodology is the study of the ordinary methods lay people use to construct social order, not a methodology for professional researchers at all.
That same confusion produced open institutional conflict over what counted as properly rigorous design ethnography. The CHI 2009 paper “Ethnography Considered Harmful,” by Andy Crabtree, Tom Rodden, Peter Tolmie, and Graham Button, argued — in a direct line of descent from Garfinkel's own polemics against mainstream sociology decades earlier — that generic, undisciplined “ethnography” should be replaced by properly “ethnomethodologically informed ethnography.” Blackwell, Blythe, and Kaye read this less as a self-evidently correct call to rigor than as a rhetorical strategy with a long pedigree: dividing the field into ethnomethodology versus “all sociology which is not ethnomethodology” lets its proponents claim a uniquely privileged relationship to “the world as it is” — repeating exactly the move that had marginalized ethnomethodologists within academic sociology in the first place, a discipline whose graduates, as their own interview sources note, had long struggled to find sociology jobs precisely because ethnomethodology had built its reputation on papers “fundamentally hostile to the entire academic enterprise of Sociology.”
In one sense, all of this is evidence of success: researchers had established that technology design benefits from studying people in actual environments. But something was also lost in the process. Ethnomethodology could become reduced to “the method the ethnographers use.” That is precisely what Garfinkel never meant. If ethnomethodology becomes merely synonymous with observation in natural settings, its theoretical provocation disappears — the field acquires the technique while losing the epistemological challenge.
XII. From Radical Inquiry to Market Research?
Bob Anderson, the first ethnomethodologist to direct EuroPARC, offered an especially sobering retrospective judgment to Blackwell, Blythe, and Kaye: perhaps ethnomethodology's contribution to Xerox should be understood as market research rather than design.
The statement deserves to be taken seriously. There is a perennial tension surrounding applied social research inside corporations. Researchers may understand themselves as revealing the practical intelligence of users, challenging technological rationalization, or transforming the conceptual foundations of design. The corporation may understand the same activity as a more sophisticated way of discovering what customers want. Both descriptions can simultaneously be true. Ethnomethodology's presence inside Xerox therefore raises a political question the original intellectual program was never especially well equipped to answer: what becomes of radical social inquiry when it becomes organizationally useful? Does institutional adoption demonstrate the power of the critique, or has the institution successfully domesticated it? Probably both.
XIII. From the Photocopier to Artificial Intelligence
The Xerox history matters again because we are living through another extraordinary period of artificial intelligence.
Large language models now perform forms of linguistic interaction that would have seemed almost fantastical when Suchman observed people struggling with a photocopier. Yet her central problem has not disappeared — it has become more interesting. Human beings inhabit radically indexical worlds. We leave things unsaid, depend upon histories, imply, joke, hesitate, use categories differently according to circumstances, revise what we mean while speaking, rely upon shared knowledge that may never be explicitly stated, and repair misunderstandings. Interaction cannot be reduced to information transfer. It is a situated achievement.
This makes contemporary human-AI interaction an extraordinary domain for renewed EM/CA research. The central question is not merely whether the system produces a correct response, but what happens after it: how a human displays acceptance or discovers misunderstanding, who initiates repair, and how a reformulated prompt or a generated answer becomes a resource shaping the next turn. A second, closely related set of questions concerns how epistemic and interactional rights get distributed between human and machine — when a user treats AI output as information, advice, speculation, authorship, companionship, or mere noise, and how people learn, over time and through trouble, what kind of conversational object they are actually dealing with. These are conversation-analytic questions. They are ethnomethodological questions. They are Suchman's questions.
XIV. A Caution About the New High Churches
Blackwell, Blythe, and Kaye's institutional argument may also travel surprisingly well into contemporary AI research. Technological laboratories again organize themselves around specialized forms of epistemic authority. Terms such as alignment, interpretability, safety, reasoning, agency, and intelligence do not merely describe neutral research objects; they become parts of institutional jurisdictions through which researchers establish what problems matter, which forms of expertise possess authority, and who gets to define technological reality. One should be cautious about stretching the church analogy too far, but its sociological question is excellent: how does a particular form of knowledge come to speak authoritatively inside an organization? That question belongs beside the technical ones. The history of EuroPARC reminds us that intellectual authority is never produced solely by the intrinsic correctness of ideas. It also depends upon recruitment, institutional patronage, personalities, budgets, organizational crises, professional networks, rhetorical styles, technologies, and historical opportunities. Knowledge has an interaction order too.
XV. The Continuing Provocation
The ethnomethodological imagination that flourished at Xerox PARC and EuroPARC was therefore more than a methodological fashion. It was a sustained argument about what human action is and how technology should be studied. Its central provocation remains powerful: designers systematically underestimate the situated character of the worlds for which they design. They imagine stable users, determinate tasks, formalizable contexts, and predictable sequences. Actual social life presents improvisational participants, emergent goals, local histories, repairable misunderstandings, shifting relevancies, and situations whose meaningful features cannot be exhaustively specified in advance. The practical intelligence necessary to make technologies work therefore resides partly outside the technology — in users, workers, colleagues, documents, stories, gestures, workarounds, memories, categories, bodies, situations, interaction. That was the great discovery of Xerox workplace studies.
But there is a second lesson. The researchers themselves inhabited an organization. They too produced hierarchies, categories, boundaries, exclusions, reputations, ritual practices, and accountable identities. The ethnomethodological gaze that revealed the invisible competencies of copier technicians and service representatives can also be turned toward ethnomethodologists. That reflexive turn strengthens rather than weakens the tradition.
Garfinkel's enduring question was how recognizable social order becomes possible as an ongoing practical accomplishment. At Xerox, that question acquired a technological formulation: how do people make technologies usable? Today it acquires another: how are humans and increasingly conversational machines together producing intelligible courses of action? We will not answer that question solely by reading technical specifications, studying model architectures, interviewing designers, or accepting corporate descriptions of what artificial intelligence supposedly does. We will have to look closely — at actual occasions of use, at breakdown, at repair, at misunderstanding, at coordination, at work — and, importantly, at ourselves while we are doing the looking.
Selected References
Blackwell, Alan F., Mark Blythe, and Jofish Kaye. 2017. “Undisciplined Disciples: Everything You Always Wanted to Know about Ethnomethodology but Were Afraid to Ask Yoda.” Personal and Ubiquitous Computing 21: 571–592.
Blomberg, Jeanette, Jean Giacomi, Andrea Mosher, and Patricia Swenton-Wall. 1993. “Ethnographic Field Methods and Their Relation to Design.” In Participatory Design: Principles and Practices, edited by Douglas Schuler and Aki Namioka.
Brown, John Seely, and Paul Duguid. 2000. The Social Life of Information. Harvard Business School Press.
Button, Graham, ed. 1991. Ethnomethodology and the Human Sciences. Cambridge University Press.
Crabtree, Andy. 2003. Designing Collaborative Systems: A Practical Guide to Ethnography. Springer.
Crabtree, Andy, Tom Rodden, Peter Tolmie, and Graham Button. 2009. “Ethnography Considered Harmful.” Proceedings of CHI 2009, 879–888.
Dourish, Paul. 2001. Where the Action Is: The Foundations of Embodied Interaction. MIT Press.
Garfinkel, Harold. 1967. Studies in Ethnomethodology. Prentice-Hall.
Gellner, Ernest. 1975. “Ethnomethodology: The Re-Enchantment Industry or the Californian Way of Subjectivity.” Philosophy of the Social Sciences 5: 431–450.
Harper, Richard. 1998. Inside the IMF: An Ethnography of Documents, Technology and Organisational Action. Academic Press.
Lave, Jean, and Etienne Wenger. 1991. Situated Learning: Legitimate Peripheral Participation. Cambridge University Press.
Orr, Julian E. 1996. Talking About Machines: An Ethnography of a Modern Job. Cornell University Press.
Sacks, Harvey. 1992. Lectures on Conversation. 2 vols. Edited by Gail Jefferson. Blackwell.
Sellen, Abigail J., and Richard H. R. Harper. 2002. The Myth of the Paperless Office. MIT Press.
Suchman, Lucy. 1987. Plans and Situated Actions: The Problem of Human-Machine Communication. Cambridge University Press.
Suchman, Lucy. 2007. Human-Machine Reconfigurations: Plans and Situated Actions. 2nd ed. Cambridge University Press.
Szymanski, Margaret H., and Jack Whalen. 2011. “Work Practice Analysis at Xerox.” Introduction to Making Work Visible: Ethnographically Grounded Case Studies of Work Practice, edited by Margaret H. Szymanski and Jack Whalen. Cambridge University Press.
Szymanski, Margaret H., and Jack Whalen, eds. 2011. Making Work Visible: Ethnographically Grounded Case Studies of Work Practice. Cambridge University Press.
Want, Roy, Andy Hopper, Veronica Falcão, and Jonathan Gibbons. 1992. “The Active Badge Location System.” ACM Transactions on Information Systems 10(1): 91–102.
Whalen, Jack, Marilyn Whalen, and Kathryn Henderson. 2002. “Improvisational Choreography in Teleservice Work.” British Journal of Sociology 53(2): 239–258.
Zimmerman, Don H. 1970. “The Practicalities of Rule Use.” In Understanding Everyday Life, edited by Jack D. Douglas. Aldine.
Zimmerman, Don H. 1984. “Talk and Its Occasion: The Case of Calling the Police.” In Meaning, Form, and Use in Context: Linguistic Applications, edited by Deborah Schiffrin. Georgetown University Press.
Comments
Post a Comment