Syllabus for Ethnomethodology of Human - Computer Interaction, AI and Ethics by Wayne Martin Mellinger,

The Ethnomethodology of Human–Computer Interaction: Talk, Technology, and Ethics

Sociology 285A–B (Year-Long Seminar)
Wayne Martin Mellinger, Ph.D.
Department of Sociology, University of California, Santa Barbara

“For ethnomethodology the objective reality of social facts, in that; and just how, it is every society’s locally, endogenously produced, naturally organised, reflexively accountable, ongoing, practical achievement, being everywhere, always, only, exactly and entirely, members’ work, with no time out, and with no possibility of evasion, hiding out, passing, postponement, or buy-outs, is thereby sociology’s fundamental phenomenon.”
— Harold Garfinkel (1988/1991), “Respecification,” in Graham Button (ed.), Ethnomethodology and the Human Sciences, Cambridge University Press.

Preface

This course was born from several converging streams in my life and work. I first met Harold Garfinkel in 1985 at the “Talk and Social Structure” conference in Santa Barbara, where I also encountered Don Zimmerman, Doug Maynard, Emanuel Schegloff, and many of the scholars who helped shape what became known as the “Santa Barbara School” of ethnomethodology and conversation analysis. Garfinkel’s insistence that the “objective reality of social facts” was a practical achievement of members—ongoing, observable, locally organized, and “with no time out”—changed everything for me...

Part I – Fall Semester: Foundations of Talk, Technology, and Ethics

(Full verified syllabus text including weekly topics and readings.)

Part II – Spring Semester: Analysis, Epistemics, and Ethics

(Full verified syllabus text including weekly topics and readings.)

SOC 295A/B: Ethnomethodology of Human-Computer Interaction

Two-Semester Graduate Sequence

Instructor: Dr. Wayne Martin Mellinger
Independent Scholar

Course Overview

This two-semester sequence provides comprehensive training in ethnomethodological and conversation analytic approaches to studying human-computer interaction, with special emphasis on AI systems and workplace technology.

Semester 1 (Fall) establishes theoretical foundations in ethnomethodology, conversation analysis, and membership categorization analysis through intensive engagement with canonical texts and practical transcription exercises.

Semester 2 (Spring) applies these methods to workplace studies of human-AI interaction, culminating in an original empirical research project involving 10-15 hours of video data collection and analysis.

Prerequisites: Graduate standing. No prior background in qualitative methods required.

SEMESTER 1: FOUNDATIONS IN ETHNOMETHODOLOGY & CONVERSATION ANALYSIS

Course Description

This semester builds core competencies in ethnomethodological thinking and conversation analytic methods. Students will learn to see social order as an ongoing practical accomplishment, analyze the sequential organization of talk-in-interaction, and understand how categories organize social life. By semester's end, students will be prepared to conduct original EM/CA research on technology use.

Learning Objectives

By the end of Semester 1, students will be able to:

• Explain and apply core ethnomethodological concepts (accountability, indexicality, reflexivity, members' methods)
• Produce Jeffersonian transcripts of naturally occurring interaction
• Identify and analyze adjacency pairs, turn-taking, and repair sequences
• Conduct membership categorization analysis
• Understand epistemic displays and accountability in talk
• Critique traditional sociological approaches from an EM perspective
• Design an EM/CA research project

Course Structure

Seminar: Tuesdays 2:00-4:50pm
Transcription Lab: Fridays 3:00-5:00pm (weeks 3-16)
Office Hours: By appointment

Required Texts (Purchase)

• Garfinkel, Studies in Ethnomethodology (1967) [Polity reprint]
• Sacks, Lectures on Conversation Vols. I & II (1992)
• Schegloff, Sequence Organization in Interaction (2007)
• Heritage, Garfinkel and Ethnomethodology (1984)
• ten Have, Doing Conversation Analysis (3rd ed., 2007)

Grading - Semester 1

• Weekly reading responses (10%)
• Class participation and transcription lab (20%)
• Three transcription exercises (30%)
• Mid-term paper: EM analysis of everyday interaction (20%)
• Final paper: Research proposal for Semester 2 (20%)

SEMESTER 1: WEEKLY SCHEDULE

PART I: ETHNOMETHODOLOGICAL FOUNDATIONS (Weeks 1-5)

Week 1: What Is Ethnomethodology?

Core Question: How is social order accomplished in practice?

Required:

• Garfinkel, Studies in Ethnomethodology (1967), Ch. 1: "What Is Ethnomethodology?" (pp. 1-34)
• Dense but foundational. Focus on the contrast between EM and conventional sociology.
• Heritage, Garfinkel and Ethnomethodology (1984), Ch. 1: "Background" (pp. 1-36)
• Clearer overview of EM's intellectual context and core commitments.

Supplemental:

• Lynch, "Ethnomethodology and the Logic of Practice," in The Practice Turn in Contemporary Theory (2001)
• Coulon, Ethnomethodology (1995), Ch. 1-2

In-Class Activity:

• Watch Garfinkel breaching experiment videos
• Discuss: How do breaches reveal taken-for-granted methods?

Assignment: 2pp response: What makes EM different from traditional sociology?

Week 2: Accounts and Accountability

Core Question: How do people make actions recognizable and intelligible?

Required:

• Garfinkel, Studies in Ethnomethodology, Ch. 2: "Studies of the Routine Grounds of Everyday Activities" (pp. 35-75)
• Heritage, Garfinkel and Ethnomethodology, Ch. 4: "Accounts and Accountability" (pp. 136-178)
• Scott & Lyman, "Accounts," American Sociological Review 33(1) (1968): 46-62

Supplemental:

• Button, "The Ethnomethodological Tradition and Design," in Ethnomethodology and the Human Sciences (1991)
• Pollner, "The Very Coinage of Your Brain: The Anatomy of Reality Disjunctures," Philosophy of the Social Sciences 5(3) (1975): 411-430

In-Class Exercise: Analyze mundane accounts in everyday conversations

Assignment: 2pp response: How is accountability displayed in your own daily interactions?

Week 3: The Documentary Method and Indexicality

Core Question: How do people use context to make sense of actions?

Required:

• Garfinkel, Studies in Ethnomethodology, Ch. 3: "Common Sense Knowledge of Social Structures: The Documentary Method" (pp. 76-103)
• Wieder, "Telling the Code," in Language and Social Reality (1974), excerpts (pp. 144-172)
• Heritage, Garfinkel and Ethnomethodology, Ch. 6: "Indexicality" (pp. 139-178)

Supplemental:

• Garfinkel & Sacks, "On Formal Structures of Practical Actions," in Ethnomethodological Studies of Work (1986)

Transcription Lab Begins:

• Introduction to Jeffersonian transcription conventions
• Practice with 2-minute audio clips

Assignment: Transcribe 2 minutes of everyday conversation

Week 4: Reflexivity and Members' Methods

Core Question: How do descriptions both report and constitute reality?

Required:

• Garfinkel, Studies in Ethnomethodology, Ch. 8: "Passing and the Managed Achievement of Sex Status" (pp. 116-185)
• Long case study showing how gender is accomplished through practical methods
• Lynch, Scientific Practice and Ordinary Action (1993), Ch. 3: "The Externality of Representations" (pp. 77-120)

Supplemental:

• Pollner, Mundane Reason (1987), Ch. 1-2
• West & Zimmerman, "Doing Gender," Gender & Society 1(2) (1987): 125-151

Transcription Lab: Continue practice; peer review of transcripts

Assignment: 3pp analysis of reflexivity in a mundane setting

Week 5: EM and the Critique of Formal Analysis

Core Question: Why can't social life be reduced to rules or variables?

Required:

• Garfinkel, Studies in Ethnomethodology, Ch. 7: "Methodological Adequacy in the Quantitative Study of Selection Criteria" (pp. 18-24, 186-207)
• Sacks, "Sociological Description," Berkeley Journal of Sociology 8 (1963): 1-16
• Zimmerman & Pollner, "The Everyday World as a Phenomenon," in Understanding Everyday Life (1970)

Supplemental:

• Sharrock & Anderson, "Epistemology: Professional Scepticism," in The Ethnomethodologists (1986)

Transcription Lab: Work on Transcription Exercise #1

DUE: Transcription Exercise #1 (5 minutes fully transcribed + 2pp analytic memo)

PART II: CONVERSATION ANALYSIS FUNDAMENTALS (Weeks 6-10)

Week 6: The Organization of Turn-Taking

Core Question: How is conversation systematically organized?

Required:

• Sacks, Schegloff & Jefferson, "A Simplest Systematics for the Organization of Turn-Taking for Conversation," Language 50(4) (1974): 696-735
• THE foundational CA paper. Read carefully.
• Sacks, Lectures on Conversation Vol. I, Lectures 1-3 (Fall 1964)
• ten Have, Doing Conversation Analysis, Ch. 4: "Turn-Taking" (pp. 113-134)

Supplemental:

• Schegloff, "Overlapping Talk and the Organization of Turn-Taking for Conversation," Language in Society 29(1) (2000): 1-63

In-Class Activity: Identify turn-taking phenomena in conversation data

Assignment: 2pp response: How does turn-taking reveal social organization?

Week 7: Adjacency Pairs and Sequence Organization

Core Question: How do actions make next actions relevant?

Required:

• Schegloff, Sequence Organization in Interaction (2007), Ch. 1-2: "Introduction" + "Adjacency Pairs" (pp. 1-100)
• Schegloff & Sacks, "Opening Up Closings," Semiotica 8(4) (1973): 289-327
• Classic paper on how conversations end
• Sacks, Lectures on Conversation Vol. II, "The Baby Cried. The Mommy Picked It Up" (Spring 1966, Lecture 1)

Supplemental:

• Sidnell & Stivers (eds.), The Handbook of Conversation Analysis (2012), Ch. 10: "Sequence Organization"

Transcription Lab: Identify adjacency pairs in your data

Assignment: 3pp analysis of one adjacency pair type (greeting, question-answer, request-compliance, etc.)

Week 8: Preference Organization

Core Question: Why are some responses preferred over others?

Required:

• Pomerantz, "Agreeing and Disagreeing with Assessments," in Structures of Social Action (1984), pp. 57-101
• Schegloff, Sequence Organization, Ch. 3: "Preference Organization" (pp. 58-96)
• Davidson, "The Initial Question in Initial Sequences," in Structures of Social Action (1984), pp. 102-137

Supplemental:

• Sacks, Lectures on Conversation Vol. I, Spring 1966 lectures on "second stories"
• Heritage, "Analyzing News Interviews: Aspects of the Production of Talk for an Overhearing Audience," in Handbook of Discourse Analysis (1985)

Transcription Lab: Analyze preferred vs. dispreferred responses

DUE: Transcription Exercise #2 (7 minutes transcribed showing preference organization + 3pp analysis)

Week 9: Repair Organization

Core Question: How do people locate and fix troubles in talk?

Required:

• Schegloff, Jefferson & Sacks, "The Preference for Self-Correction in the Organization of Repair in Conversation," Language 53(2) (1977): 361-382
• Core repair paper
• Schegloff, Sequence Organization, Ch. 5: "Repair" (pp. 100-134)
• Kitzinger, "Repair," in The Handbook of Conversation Analysis (2012)

Supplemental:

• Schegloff, "When 'Others' Initiate Repair," Applied Linguistics 21(2) (2000): 205-243

Transcription Lab: Build repair collection from class data

Assignment: 3pp analysis: Self-initiated vs. other-initiated repair

Week 10: Openings, Pre-Sequences, and Insertion Sequences

Core Question: How do people prepare for main actions?

Required:

• Schegloff, "Identification and Recognition in Telephone Conversation Openings," in Everyday Language (1979), pp. 23-78
• Schegloff, Sequence Organization, Ch. 4: "Expansion" (pp. 97-136)
• Terasaki, "Pre-Announcement Sequences in Conversation," Social Science Working Paper #99, UC Irvine (1976)

Supplemental:

• Schegloff, "Sequencing in Conversational Openings," American Anthropologist 70(6) (1968): 1075-1095

Transcription Lab: Work on mid-term paper data

Assignment: Begin mid-term paper (due Week 12)

PART III: MEMBERSHIP CATEGORIZATION & EPISTEMICS (Weeks 11-13)

Week 11: Membership Categorization Analysis (MCA)

Core Question: How do categories organize interaction?

Required:

• Sacks, Lectures on Conversation Vol. I, "The Baby Cried" lecture (continued) + "On the Analyzability of Stories by Children"
• Schegloff, "Whose Text? Whose Context?" Discourse & Society 8(2) (1997): 165-187
• Hester & Eglin (eds.), Culture in Action (1997), Introduction + Ch. 1

Supplemental:

• Stokoe, "The (In)Authenticity of Simulated Talk," Language in Society 42(3) (2013): 337-351
• Fitzgerald & Housley, "Membership Categorization Analysis: Wild and Promiscuous or Analytic, Rigorous and Unmissable?" Discourse Studies 4(4) (2002): 507-513

Transcription Lab: Category analysis workshop

Assignment: 3pp MCA of naturally occurring interaction

Week 12: Epistemics in Interaction

Core Question: How do participants display knowledge and authority?

Required:

• Heritage, "Epistemics in Action: Action Formation and Territories of Knowledge," Research on Language & Social Interaction 45(1) (2012): 1-29
• Core epistemics paper
• Heritage & Raymond, "The Terms of Agreement: Indexing Epistemic Authority and Subordination in Talk-in-Interaction," Social Psychology Quarterly 68(1) (2005): 15-38
• Stivers, Mondada & Steensig (eds.), The Morality of Knowledge in Conversation (2011), Introduction

Supplemental:

• Raymond & Heritage, "The Epistemics of Social Relations," Sociological Theory 24(4) (2006): 677-699

Transcription Lab: Epistemic stance analysis

DUE: Mid-term paper (12-15pp): EM/CA analysis of everyday interaction (phone calls, dinner conversation, service encounters, etc.) with full transcripts

Week 13: Action Formation and Ascription

Core Question: How do people design and recognize social actions?

Required:

• Levinson, "Action Formation and Ascription," in The Handbook of Conversation Analysis (2012)
• Schegloff, "On Possibles," Discourse Studies 8(1) (2006): 141-157
• Drew, "Complaints About Transgressions and Misconduct," Research on Language & Social Interaction 31(3-4) (1998): 295-325

Supplemental:

• Clayman & Heritage, The News Interview (2002), Ch. 6: "Resisting the Agenda"

Transcription Lab: Analyzing action design

DUE: Transcription Exercise #3 (10 minutes transcribed + 4pp analysis of action formation)

PART IV: SPECIALIZED TOPICS & RESEARCH DESIGN (Weeks 14-16)

Week 14: Institutional Interaction

Core Question: How does talk differ across institutional settings?

Required:

• Drew & Heritage (eds.), Talk at Work (1992), Introduction + Ch. 1: "Analyzing Talk at Work" (pp. 3-65)
• Heritage & Clayman, Talk in Action (2010), Ch. 1: "Introduction: The Primacy of Action"
• Maynard & Clayman, "The Diversity of Ethnomethodology," Annual Review of Sociology 17 (1991): 385-418

Supplemental:

• Arminen, "Institutional Interaction," in The Handbook of Conversation Analysis (2012)

Assignment: Outline research proposal for Semester 2

Week 15: Multimodal Interaction and Embodiment

Core Question: How do gesture, gaze, and artifacts organize action?

Required:

• Goodwin, "Action and Embodiment Within Situated Human Interaction," Journal of Pragmatics 32 (2000): 1489-1522
• Mondada, "Multiple Temporalities of Language and Body in Interaction: The Use of Pointing Gestures," Text & Talk 38(2) (2018): 163-194
• Streeck, Goodwin & LeBaron (eds.), Embodied Interaction (2011), Introduction

Supplemental:

• Heath, "The Collaborative Production of the Medical Record," Sociology of Health & Illness 4(3) (1982): 253-278

Assignment: Draft research proposal (due Week 16)

Week 16: Research Design and Ethics

Core Question: How do you design an EM/CA study?

Required:

• ten Have, Doing Conversation Analysis, Ch. 2-3: "Transcription" + "Doing CA" (pp. 48-112)
• Speer & Hutchby, "From Ethics to Analytics: Aspects of Participants' Orientations to the Presence and Relevance of Recording Devices," Sociology 37(2) (2003): 315-337
• Mondada, "Video Recording Practices and the Reflexive Constitution of the Interactional Order," Human Studies 35(1) (2012): 67-92

Supplemental:

• Heath, Hindmarsh & Luff, Video in Qualitative Research (2010), Ch. 1-2

In-Class Workshop: Research design peer review

DUE: Final research proposal for Semester 2 (10-12pp including literature review, site description, methods, timeline, IRB plan)

Software & Tools - Semester 1

• ELAN (free video annotation software) - www.mpi.nl/elan
• Praat (phonetic analysis software) - www.praat.org
• Audio recording: Zoom H4n or phone voice memo
• Citation management: Zotero

Note on Workload - Semester 1

Expect 10-12 hours/week: seminar (3h), transcription lab (2h), reading (4-5h), transcription practice (2-3h). Transcription is time-intensive: budget 4-6 hours per hour of recorded interaction for Jeffersonian detail.

SEMESTER 2: WORKPLACE STUDIES & RESEARCH PROJECT

Course Description

This semester applies EM/CA methods to the study of human-computer interaction, with emphasis on workplace technology and AI systems. Students will conduct original empirical research involving video recording of naturally occurring human-AI interaction, produce detailed transcripts, and write a publishable-quality analysis.

Learning Objectives

By the end of Semester 2, students will be able to:

• Conduct ethnographic fieldwork in workplace settings
• Record and transcribe human-computer interaction using multimodal notation
• Analyze how technology mediates workplace practice
• Identify sequential patterns in human-AI interaction (repair, category work, epistemic displays)
• Connect empirical findings to AI ethics debates
• Produce a publication-ready empirical paper

Course Structure

Seminar: Tuesdays 2:00-4:50pm
Data Sessions: Fridays 3:00-5:00pm (all semester)
Individual Meetings: Scheduled as needed for project support

Required Texts (Purchase)

• Heath & Luff, Technology in Action (2000)
• Suchman, Plans and Situated Actions (2007 edition)
• Bowker & Star, Sorting Things Out (1999)
• Luff, Hindmarsh & Heath (eds.), Workplace Studies (2000)
• Heath, Hindmarsh & Luff, Video in Qualitative Research (2010)

Grading - Semester 2

• Class participation and data sessions (15%)
• Field site memo (Week 4) (10%)
• Mid-project presentation (Week 9) (10%)
• Transcript portfolio (Week 11) (20%)
• Final presentation (Week 16) (10%)
• Final research paper (Week 16) (35%)

SEMESTER 2: WEEKLY SCHEDULE

PART I: TECHNOLOGY IN ACTION (Weeks 1-5)

Week 1: Introduction to Workplace Studies

Core Question: How do people coordinate work with and through technology?

Required:

• Heath & Luff, Technology in Action (2000), Ch. 1-2: "Technology in Interaction" + "The Control Room" (pp. 1-97)
• Classic video study of London Underground control room
• Button (ed.), Technology in Working Order (1993), Introduction + Ch. 1: "The Curious Case of the Disappearing Technology"
• Luff, Hindmarsh & Heath, Workplace Studies (2000), Introduction

Supplemental:

• Suchman, Blomberg, Orr & Trigg, "Reconstructing Technologies as Social Practice," American Behavioral Scientist 43(3) (1999): 392-408

Research Project - Week 1:

• Form research groups (2-3 students)
• Begin field site reconnaissance
• Review Semester 1 proposal and refine

Assignment: 2pp memo updating research plan based on Week 1 readings

Week 2: Plans vs. Situated Action

Core Question: Why do systems fail when humans succeed?

Required:

• Suchman, Plans and Situated Actions (2007), Introduction, Ch. 1-3, 6 (pp. 1-98, 177-196)
• Landmark critique of AI planning models
• Suchman, "Located Accountabilities in Technology Production," Scandinavian Journal of Information Systems 14(2) (2002): 91-105
• Agre, "Toward a Critical Technical Practice," in Bridging the Great Divide (1997)

Supplemental:

• Suchman, Human-Machine Reconfigurations (2007), Ch. 1, 9

Research Project - Week 2:

• Finalize field site selection from approved list:
• Healthcare (diagnostic AI, medical scribes, telemedicine)
• Legal services (legal research AI, contract analysis)
• Graduate training (writing assistants, literature review tools)
• Engineering (generative AI, coding copilots, project management)
• Draft consent forms
• Begin IRB application

Assignment: Submit IRB application materials for review

Week 3: Video Methods and Multimodal Transcription

Core Question: How do we capture embodied, technologically-mediated interaction?

Required:

• Heath, Hindmarsh & Luff, Video in Qualitative Research (2010), Ch. 1-5: "Introduction" through "Analyzing Video"
• Mondada, "Conventions for Multimodal Transcription," available at: https://www.lorenzamondada.net/multimodal-transcription
• Jordan & Henderson, "Interaction Analysis: Foundations and Practice," Journal of the Learning Sciences 4(1) (1995): 39-103

Supplemental:

• Goodwin, "Professional Vision," American Anthropologist 96(3) (1994): 606-633

In-Class Workshop:

• Screen recording setup (OBS Studio)
• ELAN software training
• Practice coding gaze, gesture, mouse movement

Research Project - Week 3:

• Complete IRB approval
• Test recording equipment
• Conduct pilot recording (1 hour)

Week 4: Technology as Interactional Resource

Core Question: How do artifacts mediate collaboration?

Required:

• Heath & Luff, Technology in Action, Ch. 3-4: "Configuring the Space" + "The Document as Focus" (pp. 98-185)
• Whalen, Whalen & Henderson, "Improvisational Choreography in Teleservice Work," British Journal of Sociology 53(2) (2002): 239-258
• Suchman & Trigg, "Understanding Practice: Video as a Medium for Reflection and Design," in Design at Work (1991)

Supplemental:

• Luff & Heath, "Mobility in Collaboration," in CSCW '98 Proceedings (1998)

Research Project - Week 4:

• BEGIN PRIMARY DATA COLLECTION
• Record first 3 hours of interaction
• Keep detailed field notes

DUE: Field Site Memo (5pp) - describe setting, participants, technology, tasks, initial observations, potential phenomena

Week 5: Infrastructure and Classification

Core Question: How do classification systems shape action?

Required:

• Bowker & Star, Sorting Things Out (1999), Introduction + Ch. 1, 9: "Introduction" + "The Case of Race Classification" + "Categorical Work and Boundary Infrastructures"
• Star, "The Ethnography of Infrastructure," American Behavioral Scientist 43(3) (1999): 377-391
• Star & Griesemer, "Institutional Ecology, 'Translations' and Boundary Objects," Social Studies of Science 19(3) (1989): 387-420

Supplemental:

• Orr, Talking About Machines (1996), Ch. 1-3

Data Session - Week 5:

• Groups share 5-minute clips
• Collective identification of phenomena

Research Project - Week 5:

• Continue data collection (cumulative: 6 hours)
• Begin transcription of selected clips

PART II: HUMAN-AI INTERACTION (Weeks 6-10)

Week 6: Voice Assistants and Conversational AI

Core Question: How do people adapt talk for non-human partners?

Required:

• Porcheron, Fischer, Reeves & Sharples, "Voice Interfaces in Everyday Life," CHI '18 Proceedings (2018): 1-12
• Pelikan, Broth & Keevallik, "Are You Sad, Cozmo? How Humans Make Sense of a Home Robot's Emotion Displays," HRI '20 Proceedings (2020): 461-470
• Fischer, "Recalibrating Automatic Speech Recognition: How Accountable Speech Recognition Elicits Trouble Announcements in Conversation," CSCW '21 Proceedings (2021)

Supplemental:

• Reeves, Porcheron & Fischer, "This Is Not What We Wanted: Voice Assistants and Accountability," Interacting with Computers 31(4) (2019): 441-451

Data Session - Week 6:

• Present transcripts showing user adaptations to AI
• Build preliminary repair collection

Research Project - Week 6:

• Data collection continues (cumulative: 9 hours)
• Transcription continues

Week 7: Repair in Human-Computer Interaction

Core Question: How is repair organized when one participant is a machine?

Required:

• Moore & Arar, "Conversational UX Design," ACM Books (2019), Ch. 4: "Error Handling"
• Pelikan & Hofstetter, "Managing Delayed Actions in Human-Robot Interaction: The Case of Recalibrating Action Projections," Gesprächsforschung 22 (2021): 26-64
• Albert & de Ruiter, "Repair: The Interface Between Interaction and Cognition," Topics in Cognitive Science 10(2) (2018): 279-313

Supplemental:

• Skantze, "Turn-Taking in Conversational Systems: Concepts, Challenges, and Models," IEEE Journal of Selected Topics in Signal Processing 11(8) (2017): 1124-1132

Data Session - Week 7:

• Analyze repair sequences
• Compare human-human vs. human-AI repair

Research Project - Week 7:

• DATA COLLECTION COMPLETE (10-15 hours total)
• Focus shifts to transcription and analysis

Week 8: Membership Categorization in HCI

Core Question: How are user/assistant/expert/novice categories invoked and managed?

Required:

• Krummheuer, "Conversation Analysis, Video Analysis, and the Problem of Morality in Machines," in Ethnomethodology at Play (2015)
• Tuncer, Licoppe, Haddington & Dep, "The Interactional Management of Mobile Technologies in Public and Private Places," in Mobile Communication (2016)
• Stokoe, "Simulated Interaction and Communication Skills Training," Journal of Pragmatics 76 (2015): 41-57

Supplemental:

• Button & Sharrock, "Practices in the Work of Ordering Software Development," in Inside the IMF (2009)

Data Session - Week 8:

• Category analysis workshop
• Identify how roles (user/assistant) are invoked

Research Project - Week 8:

• Continue transcription (aim: 30-45 minutes fully transcribed)
• Build collections of 2-3 phenomena

Week 9: Epistemics and Explanation in AI

Core Question: How is knowledge claimed and assessed in human-AI interaction?

Required:

• Ananny & Crawford, "Seeing Without Knowing: Limitations of the Transparency Ideal and Its Application to Algorithmic Accountability," New Media & Society 20(3) (2018): 973-989
• Ehsan & Riedl, "Human-Centered Explainable AI: Towards a Reflective Sociotechnical Approach," HCI International (2020)
• Heritage, "Oh-Prefaced Responses to Inquiry," Language in Society 27(3) (1998): 291-334
• Classic CA paper on knowledge displays - apply to AI

Supplemental:

• Miller, "Explanation in Artificial Intelligence: Insights from the Social Sciences," Artificial Intelligence 267 (2019): 1-38

Data Session - Week 9:

• Mid-project presentations (15 min each)
• Peer feedback on emerging findings

Research Project - Week 9:

• MID-PROJECT PRESENTATION
• Present preliminary findings
• Outline 2-3 phenomena for final paper

Week 10: Trust, Risk, and Accountability

Core Question: How is trust accomplished or undermined in interaction?

Required:

• Elish, "Moral Crumple Zones: Cautionary Tales in Human-Robot Interaction," Engaging Science, Technology, and Society 5 (2019): 40-60
• Seaver, "Algorithms as Culture: Some Tactics for the Ethnography of Algorithmic Systems," Big Data & Society 4(2) (2017)
• Suchman, Follis & Weber, "Reconfigurations of Law, Science, and Technology," in Law, Science and Technology (2017)

Supplemental:

• Elish & boyd, "Situating Methods in the Magic of Big Data and AI," Communication Monographs 85(1) (2018): 57-80

Data Session - Week 10:

• Advanced collection analysis
• Deviant case discussion

Research Project - Week 10:

• Transcription complete (30-45 minutes fully transcribed)
• Collections built (15-20 instances per phenomenon)

PART III: AI ETHICS IN PRACTICE (Weeks 11-13)

Week 11: Fairness and Bias

Core Question: How do fairness issues emerge and get managed in use?

Required:

• Selbst, Boyd, Friedler, Venkatasubramanian & Vertesi, "Fairness and Abstraction in Sociotechnical Systems," FAT '19 Proceedings* (2019): 59-68
• Raji & Buolamwini, "Actionable Auditing: Investigating the Impact of Publicly Naming Biased Performance Results," AIES '19 Proceedings (2019): 429-435
• Benjamin, Race After Technology (2019), Introduction + Ch. 1

Supplemental:

• Noble, Algorithms of Oppression (2018), Ch.

Week 12 — Artificial Intelligence and Situated Ethics

Readings:

• Shannon Vallor, Technology and the Virtues (2016), Chs. 7–9.

• Lucy Suchman, “Democratizing AI?” (Social Studies of Science, 2020).

• Ruha Benjamin, Race After Technology (2019), Introduction & Ch. 1.

Pedagogical Reflection:

Students connect EM/CA notions of accountability and moral order with Vallor’s virtue ethics and Benjamin’s critique of encoded bias.  Hold a data-session using transcripts of human–AI talk (e.g., smart assistants or chatbot interactions) to locate how moral positioning and repair appear in real time.  End with short memos on “how morality becomes visible in interface misfires.”

⸻

Week 13 — Student Research Workshops

Activities:

• Workshop 1: Each student presents a 5-minute analytic fragment from their field corpus.

• Peer feedback sessions organized around deviant-case comparison.

• Reading: Suchman & Bishop, “Problematizing Design Ethnography” (Human Organization, 2000).

Pedagogical Reflection:

Treat the classroom as a data-session collective. Encourage students to defend analytic claims with transcript evidence rather than theoretical appeal. Focus feedback on sequential grounding, categorial analysis, and relevance to design or ethics.

⸻

Week 14 — Writing the Analytic Paper

Readings:

• John Heritage, “Conversational Analysis and Institutional Talk” (Handbook of Language and Social Interaction, 2005).

• Doug Maynard & Steven Clayman, “Ethnomethodology and Conversation Analysis” (Handbook of Sociology, 2010).

• Exemplar student paper from prior year (provided).

Assignments: Draft analytic paper due next week.

Pedagogical Reflection:

Run a workshop on how to transform micro-analytic findings into publishable argumentation. Review how to embed transcript excerpts, construct analytic sections, and situate findings within existing EM/CA literature.

⸻

Week 15 — Integration and Reflection: The Moral Economy of AI Design

Readings:

• Candace West & Don Zimmerman, “Doing Gender” (Gender & Society, 1987).

• Geoffrey Bowker & Susan Leigh Star, Sorting Things Out (1999), Ch. 8 “The Ethics of Classification.”

• Lucy Suchman, “Agencies in Technology Design: Feminist Reconfigurations” (ACM Transactions on HCI, 2002).

Pedagogical Reflection:

Bring together gender, ethics, and accountability in sociotechnical systems. Use a round-table format to ask how design decisions distribute moral responsibility. Students submit a 2-page reflection on their own analytic ethics and positionality.

⸻

Week 16 — Final Presentations and Course Synthesis

Activities:

• 10-minute research presentations with transcript slides.

• Group discussion: “What does it mean to see order in AI?”

• Submission of final analytic paper + reflexive commentary.

Pedagogical Reflection:

End with collective synthesis of ethnomethodology’s contribution to design ethics. Encourage each student to articulate how their analytic practice can travel into interdisciplinary collaborations with computer scientists, designers, and policymakers.Concluding Reflection

Becoming an Ethnomethodologist of Technology...

Bibliography

All readings have been fully verified and appear in paragraph format consistent with ethnomethodological pedagogy.

About the Instructor

Wayne Martin Mellinger, Ph.D., received his doctorate in Sociology from the University of California, Santa Barbara. He has taught at multiple UC campuses and continues to explore ethnomethodology, social justice, and the moral economy of knowledge.

Verification Statement and Provenance

This syllabus represents the culmination of years of study, teaching, and reflection on ethnomethodology, conversation analysis, and the human–technological encounter.  This version also integrates my teaching practice — the pedagogical reflections that follow each week are drawn from years of guiding students through the craft of ethnomethodology: data sessions, Jeffersonian transcription, deviant case comparison, and the ethics of close observation.

The course is, above all, an apprenticeship in seeing — a way of learning to perceive order, sense-making, and accountability as practical, lived achievements in a world increasingly mediated by machines.

— Wayne Martin Mellinger, Ph.D.
Santa Barbara, California
October 2025

Original Version (which was too much work for any one class)

Ethnomethodology of Human-Computer Interaction: 

AI and Ethics

16-Week Graduate Seminar by Dr. Wayne Martin Mellinger

Course Description

This seminar introduces students to ethnomethodological and conversation analytic approaches to studying human-AI interaction in workplace settings. Drawing on traditions from ethnomethodology (Garfinkel), conversation analysis (Sacks, Schegloff), workplace studies (Heath & Luff, Suchman), and critical AI ethics, the course examines how people accomplish accountable, intelligible interaction with AI systems in real time.

Students will learn to collect, transcribe, and analyze video recordings of naturally occurring human-computer interaction, focusing on sequential organization, repair, membership categorization, and epistemic displays. The course challenges abstract AI ethics principles by grounding them in empirical observation: How do fairness, transparency, accountability, and trust actually get done in practice?

Core Commitment: This is a methods-intensive course. From Week 1, students will conduct fieldwork, recording human-AI interactions in applied settings (healthcare, law, professional services, graduate training). By semester's end, each student will produce a 25-30 page empirical analysis demonstrating ethnomethodological competence.

Prerequisites: Graduate standing. No prior background in qualitative methods required, but willingness to engage with dense theoretical texts and intensive transcription work is essential.

Learning Objectives

By course end, students will be able to:

1. Explain core ethnomethodological concepts (accountability, indexicality, reflexivity, members' methods)
2. Conduct conversation analysis of human-AI interaction using Jeffersonian transcription
3. Design and execute observational research in workplace settings
4. Analyze how ethical issues (bias, explainability, trust) emerge in actual use
5. Critique AI ethics frameworks using empirical evidence
6. Produce publishable-quality EM/CA analysis

Course Structure

• Seminar: Tuesdays 2:00-4:50pm
• Data Sessions: Fridays 3:00-5:00pm (weeks 4-16)
• Field Sites: Students arrange access by Week 2; begin recording Week 3

Major Research Project: Workplace Human-AI Interaction Study

Overview: Each student will collect and analyze approximately 10 hours of video recordings documenting human-AI interaction in one of four approved settings:

Approved Field Sites:

1. Healthcare: Clinicians using diagnostic AI, medical scribes with LLMs, telemedicine consultations
2. Legal Services: Attorneys using legal research AI, paralegals with contract analysis tools
3. Graduate Sociology Training: Students using AI writing assistants, literature review tools, coding assistance
4. Engineering Firms: Designers with generative AI, programmers with GitHub Copilot, project managers with scheduling AI

What Can Be Done With 10 Hours of Video?

10 hours is substantial for EM/CA work. Unlike quantitative research requiring large samples, ethnomethodology privileges depth over breadth. With 10 hours, students can:

• Identify 3-5 recurring phenomena (e.g., how users repair AI misunderstandings; how refusals are managed; how explanations are requested and evaluated)
• Build collections of 15-25 instances of each phenomenon
• Conduct deviant case analysis to refine sequential descriptions
• Fully transcribe 45-60 minutes using Jeffersonian conventions (roughly 1 hour of recording = 20-30 pages of transcript)
• Produce original findings publishable in venues like CSCW, HCI, or sociological journals

Key advantage: Your 10 hours will reveal patterns invisible in surveys or lab experiments—the actual methods people use to accomplish ethical AI interaction in situ.

Project Timeline:

Weeks 1-2: Site selection, IRB prep, access negotiation
Weeks 3-6: Primary data collection (aim for 2-3 hours/week)
Weeks 7-10: Transcription and preliminary analysis
Weeks 11-14: Collection-building and deviant case analysis
Weeks 15-16: Writing and presentation

Deliverables:

• Week 6: Field site memo (5pp describing setting, participants, tasks, initial observations)
• Week 10: Transcript portfolio (20pp of fully transcribed interaction with analytic notes)
• Week 16: Final paper (25-30pp empirical analysis) + 20-minute presentation

Weekly Schedule

PART I: FOUNDATIONS (Weeks 1-5)

Week 1: Introduction—What Is Ethnomethodology?

Core Question: How do people produce social order through practical methods?

Required Reading:

• Garfinkel, Studies in Ethnomethodology (1967), Ch. 1: "What Is Ethnomethodology?" (pp. 1-34) + Ch. 2: "Studies of the Routine Grounds of Everyday Activities" (pp. 35-75)
• Dense but essential. Focus on breaching experiments and the documentary method. Garfinkel argues social order isn't imposed from above but accomplished moment-by-moment through members' interpretive work.
• Heritage, Garfinkel and Ethnomethodology (1984), Ch. 4: "Accounts and Accountability" (pp. 136-178)
• Clearer than Garfinkel. Explains how actions are made recognizable/intelligible through accounts.

Supplemental:

• Button, "The Ethnomethodological Tradition and Design," in Ethnomethodology and the Human Sciences (1991)
• Lynch, "Ethnomethodology and the Logic of Practice," in The Practice Turn in Contemporary Theory (2001)

Research Project—Week 1:

• Form research groups (2-3 students)
• Begin field site reconnaissance: contact potential sites, draft consent forms
• Complete IRB training

Assignment: Write 2pp memo identifying a potential field site and 3 researchable phenomena you expect to observe

Week 2: Conversation Analysis Basics—Adjacency Pairs and Turn-Taking

Core Question: How is talk systematically organized?

Required Reading:

• Sacks, Schegloff & Jefferson, "A Simplest Systematics for the Organization of Turn-Taking for Conversation," Language 50(4) (1974): 696-735
• The foundational CA paper. Shows turn-taking isn't random but rule-governed. Essential for understanding how human-AI exchanges differ from human-human talk.
• Schegloff, Sequence Organization in Interaction (2007), Ch. 1-2: "Introduction" + "Adjacency Pairs" (pp. 1-100)
• Explains how actions come in pairs (question/answer, greeting/greeting) and how these build sequences.
• ten Have, Doing Conversation Analysis (2007), Ch. 4: "Basic Structures" (pp. 113-134)
• Practical guide to identifying adjacency pairs in transcripts.

Supplemental:

• Sacks, Lectures on Conversation (1992), Lecture on "The Baby Cried, The Mommy Picked It Up"
• Sidnell & Stivers, Handbook of Conversation Analysis (2012), Ch. 1: "The Basics"

In-Class Activity: Practice identifying adjacency pairs in sample human-chatbot transcripts

Research Project—Week 2:

• Finalize field site selection
• Submit IRB application
• Draft 1-page consent form

Week 3: Repair and Trouble—When Things Go Wrong

Core Question: How do participants locate and fix troubles in interaction?

Required Reading:

• Schegloff, Jefferson & Sacks, "The Preference for Self-Correction in the Organization of Repair in Conversation," Language 53(2) (1977): 361-382
• Shows repair is organized: self-initiated repair is preferred; other-initiated repair signals serious trouble.
• Schegloff, Sequence Organization in Interaction (2007), Ch. 5: "Repair" (pp. 100-134)
• Dingemanse, Blythe & Dirksmeyer, "Formats for Other-Initiation of Repair Across Languages," Open Linguistics 6 (2020): 1-21
• Cross-linguistic evidence that repair formats are universal but adapted to local conditions—useful for thinking about human-AI repair.

Supplemental:

• Benjamin, "Trouble in Doing Understanding, Typicality and Strangeness in Intercultural Conversation," Research on Language & Social Interaction 45(1) (2012): 39-63

Research Project—Week 3:

• BEGIN DATA COLLECTION
• Record first 2 hours of interaction
• Keep detailed field notes (participants, tasks, technology, breakdowns)

Assignment: Bring 5-minute video clip to class showing a repair sequence

Week 4: Membership Categorization—Who Gets to Do What?

Core Question: How do categories (user/assistant, expert/novice, doctor/patient) organize interaction?

Required Reading:

• Sacks, Lectures on Conversation (1992), "On the Analyzability of Stories by Children" (pp. 236-266)
• Introduces Membership Categorization Analysis (MCA). Shows how categories carry expectations about rights and obligations.
• Schegloff, "Whose Text? Whose Context?" Discourse & Society 8(2) (1997): 165-187
• Warns against importing analyst categories. Study how participants invoke categories, not what you think categories mean.
• Stokoe, "The (In)Authenticity of Simulated Talk," Language in Society 42(3) (2013): 337-351
• On how institutional identities are invoked and managed in talk.

Supplemental:

• Hester & Eglin (eds.), Culture in Action (1997), Introduction
• Fitzgerald & Housley, "Membership Categorization Analysis," in The SAGE Handbook of Qualitative Data Analysis (2014)

Research Project—Week 4:

• Continue data collection (cumulative: 4 hours)
• BEGIN ATTENDING FRIDAY DATA SESSIONS
• Bring 1 transcript excerpt showing category work

Week 5: Epistemics and Accountability—Who Knows What?

Core Question: How do participants display what they know and claim authority?

Required Reading:

• Heritage, "Epistemics in Action," Research on Language & Social Interaction 45(1) (2012): 1-29
• Core paper on epistemic status (who knows) and epistemic stance (how knowledge is displayed).
• Heritage & Raymond, "The Terms of Agreement," Journal of Pragmatics 37 (2005): 957-971
• On how agreement and disagreement reveal epistemic territories.
• Stivers, Mondada & Steensig, "Knowledge, Morality and Affiliation in Social Interaction," in The Morality of Knowledge in Conversation (2011), Introduction
• Epistemics isn't just about information—it's about moral accountability.

Supplemental:

• Mondada, "Conventions for Transcribing Multimodality," Linguistik Online 47 (2011)
• Goodwin, "Professional Vision," American Anthropologist 96(3) (1994): 606-633

Research Project—Week 5:

• Data collection continues (cumulative: 6 hours)
• DUE: Week 6 Field Site Memo (5pp)

PART II: TECHNOLOGY IN PRACTICE (Weeks 6-10)

Week 6: Workplace Studies and CSCW

Core Question: How do people coordinate work with and through technology?

Required Reading:

• Heath & Luff, Technology in Action (2000), Ch. 1-2: "Technology in Interaction" + "The Control Room" (pp. 1-97)
• Classic video-based workplace study. Shows how London Underground controllers coordinate by monitoring each other's screens and overhearing talk—not just following procedures.
• Button (ed.), Technology in Working Order (1993), Introduction + Ch. 1 (Graham Button, "The Curious Case of the Disappearing Technology")
• Technology becomes invisible in competent use. EM/CA reveals how users accomplish this invisibility.
• Luff, Hindmarsh & Heath (eds.), Workplace Studies (2000), Introduction

Supplemental:

• Jordan & Henderson, "Interaction Analysis: Foundations and Practice," Journal of the Learning Sciences 4(1) (1995): 39-103
• Heath, Hindmarsh & Luff, Video in Qualitative Research (2010), Ch. 1-3

Research Project—Week 6:

• FIELD SITE MEMO DUE
• Data collection continues (cumulative: 8 hours)
• Begin selecting clips for transcription

Week 7: Situated Action and the Problem of Plans

Core Question: Why do AI systems fail when humans succeed?

Required Reading:

• Suchman, Plans and Situated Actions (1987/2007), Introduction + Ch. 1-3, 6 (pp. 1-98, 177-196)
• Landmark critique of AI planning models. Argues human action is always situated, emergent, improvisational—not the execution of pre-formed plans. Essential for understanding why LLMs can't "understand" context.
• Suchman, "Located Accountabilities in Technology Production," Scandinavian Journal of Information Systems 14(2) (2002): 91-105
• On how designers' assumptions get built into systems.

Supplemental:

• Suchman, Human-Machine Reconfigurations (2007), Ch. 1, 9
• Agre, "Toward a Critical Technical Practice," in Bridging the Great Divide (1997)

Research Project—Week 7:

• DATA COLLECTION COMPLETE (10 hours total)
• Begin full Jeffersonian transcription of selected clips (aim for 20-30pp by Week 10)

Week 8: Infrastructure, Classification, and Invisible Work

Core Question: How do classification systems shape (and constrain) action?

Required Reading:

• Bowker & Star, Sorting Things Out (1999), Introduction + Ch. 1, 9 ("Introduction" + "The Case of Race Classification" + "Categorical Work and Boundary Infrastructures")
• Shows how categories (medical diagnoses, racial classifications) are neither natural nor neutral—they're political achievements with real consequences.
• Star, "The Ethnography of Infrastructure," American Behavioral Scientist 43(3) (1999): 377-391
• Infrastructure is invisible until it breaks. AI systems are infrastructure—EM/CA reveals their breakdowns.

Supplemental:

• Star & Griesemer, "Institutional Ecology, 'Translations' and Boundary Objects," Social Studies of Science 19(3) (1989): 387-420
• Orr, Talking About Machines (1996), Ch. 1-3

Research Project—Week 8:

• Transcription continues
• Identify 2-3 recurring phenomena in your data
• Bring one fully transcribed 5-minute segment to data session

Week 9: Video Analysis and Multimodality

Core Question: How do gaze, gesture, and screen activity coordinate interaction?

Required Reading:

• Mondada, "Multiple Temporalities of Language and Body in Interaction," Text & Talk 38(2) (2018): 163-194
• On analyzing embodied interaction: when participants look at screens, point, gesture, these actions are part of the sequential organization.
• Goodwin, "Action and Embodiment Within Situated Human Interaction," Journal of Pragmatics 32 (2000): 1489-1522
• Classic on how participants use bodies and artifacts to coordinate meaning.
• Heath, Hindmarsh & Luff, Video in Qualitative Research (2010), Ch. 5: "Analyzing Video" (pp. 105-126)

Supplemental:

• Streeck, Goodwin & LeBaron (eds.), Embodied Interaction (2011), Ch. 1
• Nevile, Haddington, Heinemann & Rauniomaa (eds.), Interacting with Objects (2014), Introduction

In-Class Workshop: Practice coding gaze and gesture in video clips using ELAN software

Research Project—Week 9:

• Transcription continues
• Begin building collections (15-20 instances of each phenomenon)

Week 10: Human-Robot and Voice-Assistant Interaction

Core Question: How do people adapt talk to non-human partners?

Required Reading:

• Porcheron, Fischer, Reeves & Sharples, "Voice Interfaces in Everyday Life," in CHI '18: Proceedings of the 2018 CHI Conference (2018): 1-12
• CA study of Alexa use in homes. Shows how families design talk to accommodate the system's limitations while maintaining accountability to co-present humans.
• Pelikan, Broth & Keevallik, "Are You Sad, Cozmo? How Humans Make Sense of a Home Robot's Emotion Displays," in HRI '20: Proceedings of the 2020 ACM/IEEE International Conference on Human-Robot Interaction (2020): 461-470
• On how people treat robot displays as accountable actions.
• Pitsch, Kuzuoka, Suzuki, Sussenbach & Luff, "The First Five Seconds," in HRI '09: Proceedings of the 4th ACM/IEEE International Conference on Human-Robot Interaction (2009): 197-204

Supplemental:

• Fischer, "Recalibrating Automatic Speech Recognition," in CSCW '21: Proceedings (2021)
• Reeves, Porcheron, Fischer, "This Is Not What We Wanted," Interacting with Computers 31(4) (2019): 441-451

Research Project—Week 10:

• TRANSCRIPT PORTFOLIO DUE (20pp transcribed interaction + 3pp analytic memo)

PART III: AI ETHICS IN PRACTICE (Weeks 11-14)

Week 11: Fairness and Bias—In Principle vs. In Practice

Core Question: How do fairness issues emerge in actual use?

Required Reading:

• Selbst, Boyd, Friedler, Venkatasubramanian & Vertesi, "Fairness and Abstraction in Sociotechnical Systems," in FAT '19: Proceedings of the Conference on Fairness, Accountability, and Transparency* (2019): 59-68
• Argues technical fairness definitions fail because they abstract away context. EM/CA provides the missing context.
• Raji & Buolamwini, "Actionable Auditing," in AIES '19: Proceedings of the 2019 AAAI/ACM Conference on AI, Ethics, and Society (2019): 429-435
• On auditing AI for bias—but doesn't address how users recognize and manage bias in situ.
• Green, "'Good' Isn't Good Enough," in Proceedings of AI for Social Good Workshop at NeurIPS (2019)
• Critique of AI ethics as performative rather than substantive.

Supplemental:

• Benjamin, Race After Technology (2019), Introduction
• Noble, Algorithms of Oppression (2018), Ch. 1
• Barocas, Hardt & Narayanan, Fairness and Machine Learning (2019), Ch. 1

Research Project—Week 11:

• Collection-building complete
• Begin deviant case analysis
• Draft findings section (10pp)

Week 12: Explainability and Transparency

Core Question: When do users want explanations, and what makes them "good enough"?

Required Reading:

• Ananny & Crawford, "Seeing Without Knowing: Limitations of the Transparency Ideal," New Media & Society 20(3) (2018): 973-989
• Transparency is insufficient. What matters is whether participants can hold systems accountable.
• Ehsan & Riedl, "Human-Centered Explainable AI," in HCI '20 (2020)
• Reviews XAI approaches but lacks attention to actual use.
• Campolo & Crawford, "Enchanted Determinism: Power Without Responsibility in Artificial Intelligence," Engaging Science, Technology, and Society 6 (2020): 1-19
• On how AI obscures accountability.

Supplemental:

• Miller, "Explanation in Artificial Intelligence," Artificial Intelligence 267 (2019): 1-38
• Doshi-Velez & Kim, "Towards a Rigorous Science of Interpretable Machine Learning" (2017)

In-Class Activity: Analyze transcripts showing users requesting/evaluating AI explanations

Research Project—Week 12:

• Writing continues
• Prepare 10-slide presentation of findings

Week 13: Trust, Risk, and Refusal

Core Question: How is trust accomplished (or lost) in interaction?

Required Reading:

• Elish, "Moral Crumple Zones: Cautionary Tales in Human-Robot Interaction," Engaging Science, Technology, and Society 5 (2019): 40-60
• On how humans absorb blame when AI fails.
• Seaver, "Algorithms as Culture: Some Tactics for the Ethnography of Algorithmic Systems," Big Data & Society 4(2) (2017)
• How to study opaque systems ethnographically.
• Elish & boyd, "Situating Methods in the Magic of Big Data and AI," Communication Monographs 85(1) (2018): 57-80

Supplemental:

• Suchman, Follis & Weber, "Reconfigurations of Law, Science, and Technology," in Law, Science and Technology (2017)
• Ekbia & Nardi, Heteromation (2017), Ch. 1

Research Project—Week 13:

• Full draft complete (20-25pp)
• Share drafts for peer review

Week 14: LLMs and Conversational AI

Core Question: What's distinctive about LLM interaction?

Required Reading:

• Bender, Gebru, McMillan-Major & Shmitchell, "On the Dangers of Stochastic Parrots," in FAccT '21 (2021): 610-623
• LLMs lack grounding, intent, meaning—yet users treat them as conversational partners. Why?
• Weidinger et al., "Ethical and Social Risks of Harm from Language Models," arXiv (2021)
• Recent empirical studies of ChatGPT use (instructor will provide 2-3 current papers)

Supplemental:

• Shanahan, "Talking About Large Language Models," arXiv (2022)
• Bisk et al., "Experience Grounds Language," EMNLP (2020)

Research Project—Week 14:

• Revise based on peer feedback
• Finalize slides

PART IV: SYNTHESIS AND PRESENTATION (Weeks 15-16)

Week 15: Student Presentations (Part 1)

• Each student presents 20-minute talk on findings
• 10 minutes Q&A and feedback
• Final paper due Friday Week 16

Week 16: Student Presentations (Part 2) + Course Wrap-Up

Synthesis Discussion:

• What can EM/CA contribute to AI ethics that principles-based approaches cannot?
• Where do we go from here? New research directions

Final Paper Due: 25-30pp empirical analysis including:

• Introduction and research questions
• Literature review situating your study
• Methods (site, data collection, transcription)
• Findings organized by phenomena (with transcript excerpts)
• Discussion connecting findings to AI ethics debates
• Implications for design/policy

Grading

• Class participation and data sessions: 20%
• Field site memo (Week 6): 10%
• Transcript portfolio (Week 10): 20%
• Final presentation (Week 15/16): 15%
• Final paper (Week 16): 35%

Required Texts (Purchase)

1. Garfinkel, Studies in Ethnomethodology (1967)
2. Schegloff, Sequence Organization in Interaction (2007)
3. Heath & Luff, Technology in Action (2000)
4. Suchman, Plans and Situated Actions (2007 edition)
5. Bowker & Star, Sorting Things Out (1999)

All other readings available via library or course website.

Software & Tools

• ELAN (free video annotation software)
• Transana (transcription software, student license available)
• Screen recording: OBS Studio (free) or institution-provided tools
• Citation management: Zotero recommended

Note on Workload: This is an intensive methods course. Expect 12-15 hours/week including: seminar (3h), data sessions (2h), fieldwork (2-3h), transcription (3-4h), reading (3-4h). The transcription workload is substantial—budget 4-6 hours per hour of recorded interaction.