How a coalescing set of about twenty people becomes “the group.”
The starting point is neither twenty strangers nor a fully institutionalized organization. It is a nascent collective: members already recognize a shared concern, have begun recurring interaction, and can roughly identify one another, but membership, roles, memory, authority, and continuity remain only partly stabilized.
For Actor-Network Theory, group formation is the progressive stabilization of heterogeneous relations until participants and outsiders can treat a distributed network as one actor. The group is not simply the people. It is the recurring association of people, meetings, channels, documents, routines, roles, objects, places, and representations that makes a durable “we” possible.
Actor-Network Theory: the minimum toolkit
ANT begins with associations rather than a pre-given social container. It asks how people, artifacts, spaces, texts, rules, software, representations, and absent actors become linked strongly enough to produce an observable course of action. Action is distributed: a person rarely acts alone when a calendar, agenda, room, message history, permission system, prior decision, and other participants make the act effective. Mediators matter because they transform what they carry. A summary does not merely transport a meeting; it selects, compresses, and authorizes one version of it. Durability is an empirical achievement. When a heterogeneous network becomes reliable enough to be treated as one actor, ANT describes the simplification as punctualization.
Interaction-Order Sociology: the minimum toolkit
This explainer uses Interaction-Order Sociology as a compact label for the Goffmanian microsociological tradition; it is broader than dramaturgy alone. It treats the face-to-face or technologically mediated encounter as a distinctive domain of social order. Participants sustain a definition of the situation, allocate participation statuses, shift footing as speakers and hearers, protect or threaten face, manage embarrassment, and repair disruptions. These processes explain how a particular meeting remains intelligible and orderly. ANT enters one level sideways: it asks what objects, records, infrastructures, and transported absences allow that local interaction order to accumulate into a group that persists beyond the encounter.
There is enough recurrence and mutual recognition that the analyst can follow an emerging collective rather than explain first contact.
Large enough that the whole collective cannot remain continuously present through unmediated conversation; small enough that mutual recognition and direct interaction remain plausible.
How does a partially assembled network acquire continuity, boundaries, internal differentiation, memory, and the capacity to act as one?
Latour’s first uncertainty is almost perfectly fitted to a nascent twenty-person collective.
Reassembling the Social explicitly begins with “No Group, Only Group Formation.” The methodological instruction is to trace the work by which an emerging “we” is repeatedly made rather than treating the group as a container that already exists.
Membership is unfinished
Some people are clearly core, some peripheral, some prospective, some absent but still counted, and some present without yet being recognized as members.
Purpose is under translation
Participants may agree that “something” matters while still translating the common concern differently into projects, priorities, or expected outcomes.
Roles are emerging
Conveners, interpreters, organizers, skeptics, scribes, bridge figures, and experts become recurrent without necessarily being formally appointed.
Continuity is fragile
The group may depend heavily on one recurring meeting, one chat, one organizer, one document, or one venue. Remove the carrier and the collective may partially disappear.
Assume that coalescence has already begun, but punctualization has not.
A useful baseline is a collective that has met several times, contains recognizable recurring participants, and possesses a preliminary common concern, yet still requires active work to remain one thing from encounter to encounter.
Formation is better represented as several stabilization dimensions than as a fixed sequence of stages.
A Latourian account is suspicious of developmental stage models that imply an internally driven progression. Different dimensions can stabilize at different rates, reverse, or be delegated to different mediators.
Recurrence
Do interactions happen often enough that prior encounters constrain later ones?
Boundary
Can participants practically distinguish member, guest, affiliate, and outsider?
Concern
Is there a sufficiently durable matter around which heterogeneous interests can keep associating?
Infrastructure
Do calendars, channels, venues, documents, and routines carry the group between encounters?
Differentiation
Have roles, competencies, recurring subgroups, or divisions of work become recognizable?
Representation
Can a spokesperson, document, decision, or artifact legitimately stand for the collective?
A group need not share one mental representation; it needs a concern that can survive repeated translation.
ANT shifts attention from consensus inside heads to the practical carriers through which a common matter remains discussable across heterogeneous participants.
The boundary is produced by mundane devices long before it becomes a constitution.
In an emerging twenty-person group, membership is often visible through traces rather than rules. Who receives the calendar invite? Who is in the chat? Who gets asked for input? Who is introduced as “one of us”? Who may vote, edit, host, invite, or see the archive?
Mailing lists, channel permissions, shared folders, room access, recurring invites.
Introductions, pronouns, invitations to speak, expectations of attendance, references to shared history.
Member lists, notes, photos, archives, participant records, version histories.
The boundary is therefore neither purely subjective nor purely administrative. It is a network effect distributed across social recognition and material access.
The encounter organizes participation; the network makes the encounter consequential beyond itself.
Interaction-Order Sociology analyzes how participants make a particular encounter workable: footing, face-work, participation frameworks, deference, embarrassment, framing, and repair. ANT asks how that moment is materially equipped, how absent events and actors enter it, and what residue is carried forward. The two lenses overlap at the point where an interactional move becomes durable enough to reorganize later interaction.
| Interaction event | Interaction-order question | ANT question | Coupled diagnosis for group formation |
|---|---|---|---|
| The convener opens the meeting | How is the encounter framed, and how is the convener’s footing as legitimate opener ratified by others? | What calendar ownership, agenda, meeting link, room control, prior repetition, and delegated authority make the first words effective? | A role is stabilizing when the local recognition of the opener is repeatedly backed by mediators that make recurrence easier to reproduce. |
| The agenda is accepted without discussion | How does tacit acquiescence define what kind of encounter this is and what counts as relevant participation? | Who authored the agenda, what template or prior meeting shaped it, and what topics disappeared before the interaction began? | Interactional consent can black-box earlier selection work. A seemingly neutral agenda may already encode emerging priorities and role asymmetries. |
| A newcomer asks a basic question | How is novice status displayed, protected, or stigmatized? Does the group perform welcome, impatience, or corrective instruction? | What onboarding document, archive, glossary, invitation path, or absent prior encounter could have supplied the missing context? | Boundary formation becomes visible in the cost of becoming interactionally competent and in whether that cost is carried by people or delegated to artifacts. |
| A member says “we agreed…” | How does the utterance alter footing by positioning challenge as reopening something already settled? | What notes, attendance, decision rule, chat history, recording, or remembered procedure warrants the collective pronoun? | “We” becomes durable only when an interactional claim to closure is supported by a representation chain that later participants accept. |
| A peripheral member remains silent | What participation status does silence acquire given gaze, turn opportunities, expertise, embarrassment, and face concerns? | How do microphone settings, seating, agenda pace, chat affordances, subgroup history, and invitation routines redistribute the practical ability to intervene? | Repeated local silence can become a stable core/periphery topology without anyone formally assigning membership tiers. |
| A disagreement becomes awkward | What face-work, tact, joking, hedging, apology, or topic shift keeps the encounter from becoming interactionally unworkable? | What shared document, voting option, parking-lot list, facilitator role, or asynchronous channel translates conflict into another form? | Repair devices may convert potentially fragmenting disagreement into a durable procedure for carrying disagreement without dissolving the group. |
| A shared document is edited live | Who can interrupt, correct, ratify, claim authorship, or save face while the wording is negotiated? | How does the document transform fleeting speech into versioned, attributable, portable text with permissions and a revision history? | The document couples conversational ratification to material memory. It can become more durable than anyone’s recollection of the interaction that produced it. |
| A subgroup reports back | How do presenters claim the floor and establish that they speak from a different participation footing than ordinary members? | What notes, subgroup membership, slides, liaison role, or mandate lets two people speak for six absent collaborators? | The group begins to tolerate internal differentiation when spokesperson chains allow local work to re-enter the plenary without requiring universal co-presence. |
| The chair summarizes apparent consensus | Do participants display ratification, hesitation, avoidance, or challenge? What interactional cues make silence count as agreement or merely non-opposition? | Does the summary become an inscription? Who can edit it later, and what procedure determines whether it stands for the encounter? | Punctualization often begins in precisely this move: heterogeneous positions are compressed into a portable statement attributed to the collective. |
| A late arrival asks what was decided | How is the newcomer’s participation frame repaired without restarting the encounter for everyone else? | Which notes, chat, agenda markers, recordings, or human summaries compete to translate the absent portion of the meeting? | The ability to reintegrate partial absence tests whether collective memory has moved beyond continuous co-presence. |
| Someone invokes an absent expert or rule | How does the invocation change footing and authority within the local encounter? | What document, credential, policy, citation, screenshot, or relationship allows the absent actor to become present and consequential? | The small group is never interactionally sealed. Its local order is populated by transported absences that can reorganize authority without entering the room as persons. |
| The meeting closes with action items | How are future obligations publicly assigned, acknowledged, hedged, or left ambiguous? | What task list, assignee field, deadline, notification, calendar event, or follow-up routine makes the obligation survive the closing ritual? | A meeting contributes to durable group formation when interactional commitments acquire carriers capable of reappearing in later episodes. |
Proto-roles emerge when recurrent formation work becomes attached to positions, people, and artifacts.
For ANT, an informal role is not merely a personality trait or a title. It is a recurrent position in a network: certain tasks, artifacts, permissions, expectations, and dependencies repeatedly route through the same person or person-artifact assemblage.
| Emerging function | How it becomes visible | Typical delegated supports | Formation consequence |
|---|---|---|---|
| Convener | Repeatedly initiates recurrence, opens meetings, protects time, and reactivates stalled participation. | Calendar ownership, recurring event, mailing list, venue access, meeting link. | Recurrence becomes easier, but the group may be highly dependent on one bottleneck. |
| Memory keeper | Produces the account later treated as what happened, what was decided, and what remains open. | Notes, recordings, action log, canonical document, version history. | Collective memory becomes externalized; archival control can also become power over the usable past. |
| Boundary keeper | Handles invitations, newcomer introductions, removals, permissions, and questions about who counts. | Roster, access-control list, invitation link, membership form, chat permissions. | A fuzzy social boundary becomes operationally consequential before any formal membership rule exists. |
| Interpreter | Restates technical, ideological, or subgroup-specific language so that heterogeneous participants can continue coordinating. | Glossaries, examples, shared diagrams, summary notes, comparison tables. | Local vocabularies can remain differentiated without immediately fragmenting the wider collective. |
| Bridge / broker | Participates in multiple clusters and becomes the route by which concerns, reputations, and commitments move between them. | Cross-channel access, recurring subgroup attendance, liaison meetings, forwarding practices. | The larger group can remain connected despite sparse all-to-all interaction; dependency on the broker may become a structural vulnerability. |
| Spokesperson | Outsiders increasingly treat one participant’s statements as equivalent to the collective’s position or commitments. | External email thread, public profile, meeting invitation, presentation deck, signature block. | The network becomes externally punctualizable, but representational authority may outrun internal authorization. |
| Procedural keeper | Remembers how decisions are made, when a topic is closed, and what sequence is expected. | Agenda template, facilitation script, voting tool, checklist, standing rules. | Procedure shifts from tacit memory toward reproducible infrastructure and becomes less dependent on the original founders. |
| Repairer / host | Notices disengagement, conflict, confusion, missing access, or broken logistics and restores workable participation. | Backchannel messages, technical support, reminders, alternative meeting modes, conflict routines. | Fragility is absorbed before it becomes collective failure; repair competence can become an informal source of authority. |
How a proto-role stabilizes
The collective becomes more group-like as its past becomes recoverable without relying on the same minds being present.
Around twenty people, attendance will vary. Durable group formation therefore depends on ways of letting absent members, prior decisions, and earlier controversies re-enter the present encounter.
Recurring calendar series, deadlines, ritual openings, periodic reports, scheduled reviews.
Notes, chat history, recordings, version history, shared terminology, decisions, artifacts.
Name, logo, shared vocabulary, canonical examples, origin stories, recurring jokes, recognizable practices.
An inscription is sociologically consequential because it can make an absent encounter present again. A one-line decision record can alter what twenty people are permitted to reopen weeks later.
The group rarely exists as one continuously interacting circle; it exists through a topology of subgroups and bridges.
At this size, some participants interact more often, share more vocabulary, or perform specialized work together. ANT treats those subgroups neither as noise nor automatically as threats. The question is how their partial networks are connected and how differences are translated.
Affinity clusters
People with similar interests or styles interact more frequently and can develop local shorthand.
Task clusters
Small working sets temporarily become more tightly associated around a document, event, project, or problem.
Bridges
Particular people, shared artifacts, and recurring plenary meetings connect clusters that would otherwise drift apart.
Translation costs
As local vocabularies and histories diverge, more work is required to make one subgroup’s concerns actionable to another.
A network becomes a group-actor when its internal heterogeneity can be provisionally compressed for a particular operation.
Punctualization is therefore not the same as cohesion, consensus, identity, formal organization, or ontological unity. Twenty participants may disagree intensely and still be punctualized as “the group” for scheduling a meeting. The same network may fail to be punctualizable for issuing a statement that purports to bind every member.
| Kind of punctualization | What can now be compressed into “the group”? | Evidence | Typical failure |
|---|---|---|---|
| Referential | A name reliably points to roughly the same emerging collective across encounters. | Members and outsiders understand invitations, messages, and references to the group name. | The name masks radically different rosters or meanings across subgroups. |
| Recurrent | The collective can reconvene without reconstructing contact and purpose from zero. | Recurring schedule, known channel, expected convener, recognizable meeting form. | Remove one organizer and recurrence stops. |
| Mnemonic | The group can be said to remember. | Notes, archives, version histories, origin stories, action logs are routinely accepted as collective memory. | Multiple incompatible archives or disputed summaries reopen the past. |
| Productive | An artifact or event can be attributed to the collective rather than to isolated contributors. | Shared authorship, recognized review process, stable publication or event routine. | Contributors dispute whether the output actually represents the whole group. |
| Representational | A spokesperson or inscription can travel as “the group’s” voice. | Mandate, accepted spokesperson, decision procedure, ratified statement. | Absent members reject the representation or the spokesperson exceeds the mandate. |
| Binding | A decision can create obligations for participants who were not present at the moment of decision. | Accepted authority rule, membership boundary, notification and contestation procedure. | Participants treat the alleged obligation as only one subgroup’s or one convener’s preference. |
Five stress tests
A group becomes externally consequential when representatives can carry it across encounters.
Twenty people cannot all be present in every interaction with outsiders. The collective therefore begins to depend on spokespersons and inscriptions that stand in for absent members.
A convener, delegate, liaison, or subject-matter lead speaks to another group or institution.
A statement, minutes, charter, proposal, vote result, roster, or public page is treated as expressing the collective.
A prototype, demonstration, event, or shared artifact can come to represent “what this group does.”
The ANT question is not simply whether representation is legitimate in the abstract. It is: what chain allows the representative to claim the group, and what mechanisms allow absent participants to correct, reject, ratify, or revise that claim?
Influence becomes durable when some participants can mobilize longer, more reliable chains of association.
Latour’s “Powers of Association” treats power as an effect of successfully enrolling, aligning, and transporting action rather than as a substance possessed prior to the network. In a coalescing twenty-person group, this directs attention toward unequal access to the mediators through which the collective recurs, remembers, translates, and represents itself.
| Emerging power | Carrier | Mechanism | What to observe |
|---|---|---|---|
| Temporal power | Calendar, agenda, deadlines | Controls when collective attention recurs and which issues receive scarce synchronous time. | Who can schedule, cancel, reorder, or defer without negotiation? |
| Archival power | Minutes, summaries, canonical files | Shapes which version of the past becomes available to constrain later interaction. | Whose summary gets cited? Who can correct it? Which disagreements disappear? |
| Boundary power | Roster, invitation links, permissions | Changes who can practically participate and who remains visible as a member. | Who can admit, exclude, mute, remove, or define “inactive”? |
| Classificatory power | Categories, tags, agenda headings, forms | Defines what kinds of issues and participants become legible to the collective. | Who chooses categories such as “decision,” “discussion,” “member,” “observer,” or “resolved”? |
| Brokerage power | Bridge relations across subgroups | Controls or selectively transforms what moves between otherwise weakly connected clusters. | Are there concerns that reach the plenary only through one person? |
| Interpretive power | Specialized expertise, summaries, diagrams | Makes one participant the recurrent translator of difficult material and therefore the shaper of what others can coordinate around. | Can other members independently reconstruct the issue, or must they borrow one person’s translation? |
| Representational power | External email, public statement, presentation | Allows one actor to make claims that outsiders treat as equivalent to the whole collective. | What mandate exists, and how often do external actors bypass the group and contact the representative directly? |
| Repair power | Technical access, conflict routines, social brokerage | Makes one participant indispensable when interactions break down. | Who gets summoned when the network stops working, and what new authority accumulates from successful repairs? |
Power can compound without becoming formal office
A participant who writes the summaries may become the person others consult about what happened. Being consulted produces more contextual knowledge; that knowledge makes future summaries easier to write and more authoritative. Likewise, a bridge figure who repeatedly translates between subgroups becomes more central because each successful translation generates additional ties, expectations, and privileged information. ANT therefore looks for positive feedback in association chains, not merely prestige or explicit command.
Breakdowns reopen black boxes and expose the dependencies that successful interaction had made invisible.
When the collective is working, many carriers of order recede from attention. Breakdown is therefore not only failure; it is a diagnostic event. It reveals which association had been performing recurrence, memory, membership, authority, translation, or representation.
| Breakdown | Dependency exposed | Typical repair work | What successful repair would stabilize |
|---|---|---|---|
| The organizer stops convening | Recurrence had been concentrated in one human-calendar assemblage rather than generalized across the collective. | Recover access, assign a successor, duplicate hosting capability, reconstruct schedule expectations. | Recurrence becomes a group capacity rather than a founder-specific capacity. |
| The meeting link or room becomes unavailable | Co-presence depended on infrastructure that had been treated as background. | Alternative channel, rapid notification, new access path, altered participation format. | The group acquires redundancy in the material conditions of encounter. |
| A subgroup opens a private channel | The plenary’s appearance of informational unity masked a more differentiated topology of trust and work. | Negotiate what must be shared, appoint bridges, create summary routines, or accept partial separation. | The larger group learns which internal differences can remain local and which translations are necessary for continued unity. |
| Six regulars stop attending | The nominal roster and active actor-network diverge. | Contact absentees, reclassify membership, recruit replacements, alter meeting form, or accept a smaller core. | Membership categories become linked to actual participation rather than inherited lists. |
| A disputed “decision” is cited later | Collective memory and authority depended on weak inscriptions or ambiguous ratification. | Reconstruct the episode, compare notes, define decision status, reopen or ratify the issue. | A more explicit chain from discussion → closure → inscription → later citation. |
| A newcomer cannot follow the discussion | Interactional competence depended on tacit history carried mainly in regular members’ memories. | Context brief, glossary, archive pointers, mentoring, slower framing of the issue. | More of the group’s background becomes transportable beyond continuous membership. |
| The bridge figure leaves | Subgroup coordination was centralized in one broker. | Direct cross-links, rotating liaisons, shared artifacts, new joint meetings. | Inter-cluster translation becomes redundant rather than person-dependent. |
| An outsider asks for an official position | The group lacks or contests a spokesperson chain and may not have a clear binding decision procedure. | Clarify membership, define mandate, deliberate, ratify wording, authorize a representative. | Representational punctualization: the collective can sometimes speak as one actor beyond the meeting. |
Repair can strengthen, merely restore, or fragment
Coalescence Profile: which stabilizations are carrying the group?
Toggle the conditions that are already true. The result is a heuristic ANT profile, not a validated psychometric scale.
Remove one mediator and watch which part of the collective becomes expensive again.
Choose an infrastructure element to remove from an already coalescing group.
Boundary Repair: trace who or what is allowed to redefine membership.
Select a disturbance. The lab identifies the mediators through which “who we are” would have to be repaired.
Can this collective legitimately be compressed into “the group”?
Choose an externally visible claim. The lab stress-tests what must be true before the singular noun is analytically safe.
Twenty people can form one group through very different internal topologies.
The diagrams are schematic. Select a topology to see what ANT would predict becomes consequential for formation.
Interaction Microscope: one event, two analytical resolutions.
This lab keeps the Interaction-Order Sociology × ANT synthesis narrow: interactional form plus the mediators that let the local event contribute to group formation.
A twenty-person discussion-and-practice collective during its first eight weeks.
An open invitation around a shared intellectual/practical concern produces twenty-two channel members. Fifteen attend a typical meeting; eighteen have attended at least twice. Two people convene, three recurring subgroups are beginning to form, no constitution exists, and participants describe the collective as informal and horizontal.
Initial network inventory
Formation sequence
Purpose-relative punctualization at week eight
| Operation | Can the collective be treated as one actor? | Why |
|---|---|---|
| Reconvene | Yes, strongly | Schedule, channel, known participants, facilitation expectations, and backup access now carry recurrence. |
| Remember prior work | Mostly | Archives and status-tagged notes survive changing attendance, though interpretation still depends on regulars. |
| Produce a shared document | Yes, conditionally | Editing and review practices exist, but the meaning of authorship still varies by subgroup. |
| Issue the specific external statement | Yes | The decision procedure, wording, and spokesperson were explicitly authorized for this case. |
| Speak on arbitrary future issues | No | No general spokesperson mandate or general decision rule exists. |
| Spend collective money | No | No treasury, ownership structure, authorization chain, or accepted liability boundary exists. |
ANT contributes a distinctive account of how an emerging group acquires durability and agency.
| Tradition | Strong question | What ANT adds for this case |
|---|---|---|
| Interaction-Order Sociology | How is each encounter ordered and how are participation, footing, face, and situation managed? | How do artifacts, records, schedules, absent authorities, and delegated routines carry that interaction order across encounters? |
| Mead / pragmatism | How do selves, roles, common objects, and coordinated action emerge through interaction? | How are those relations materially extended through heterogeneous nonhuman mediators? |
| Social Network Analysis | Who is connected to whom, with what density, centrality, clustering, or brokerage? | What do the ties consist of, what transformations occur along them, and how do documents and objects participate? |
| Group dynamics | How do cohesion, norms, roles, conflict, leadership, and performance change over time? | Which concrete carriers make those abstractions durable enough to act, travel, and survive changing attendance? |
| CHAT | How is collective activity organized around an object through tools, rules, community, and division of labor? | How does the very boundary and identity of “the activity system” become assembled rather than assumed? |
Study formation as carry-over across encounters, artifacts, boundaries, and claims made in the collective name.
The protocol should capture both interaction-order detail and ANT-style material extension. A transcript alone misses the calendar, permissions, notes, version history, absentees, subgroup channels, and later reuse through which an encounter becomes consequential.
Recommended observation corpus
Event-level coding scheme
| Code family | Record | Analytic question |
|---|---|---|
| Episode | Time, setting, participants, subgroup/plenary status, prior event invoked. | What interactional situation is being produced here? |
| Interaction order | Footing shifts, participation status, face-work, framing, ratification, embarrassment, repair. | How is local participation made orderly enough for the episode to continue? |
| Actors / actants | People, artifacts, spaces, software, texts, absent authorities, categories. | What actually makes a difference to the course of action? |
| Mediation | Input, transformation, output for each consequential mediator. | What changes as action passes through this entity rather than simply being transported? |
| Inscription | Notes, decisions, status markers, charts, posts, files, summaries. | What fleeting event is being made durable and portable, and with what loss or compression? |
| Membership claim | “Member,” “regular,” “core,” “newcomer,” “inactive,” access changes. | Which devices and performances make the boundary consequential? |
| Representation claim | “We,” “the group,” “everyone,” “the decision,” spokesperson acts. | What chain warrants the claim that heterogeneous participants can be represented as one? |
| Differentiation | Recurring functions, bottlenecks, broker positions, permissions, expertise routing. | Which proto-roles are becoming durable despite the absence of formal titles? |
| Breakdown / repair | Failure, exposed dependency, repair actors, changed artifact or procedure. | What hidden association has become visible, and does repair redistribute or reconcentrate it? |
| Carry-over | Later episode in which this event, artifact, claim, or obligation reappears. | Did the episode sediment into group formation, or disappear with the encounter? |
Minimal field-note record
Analytic outputs
ANT is strongest on assembly, mediation, and durability; Interaction-Order Sociology is strongest on the fine organization of the encounter.
The synthesis is useful precisely because neither vocabulary should absorb the other. The interaction order cannot be reduced to objects and networks, and group durability cannot be reconstructed from conversational performance alone.
| Limit | What ANT contributes | What remains underdescribed | Complement |
|---|---|---|---|
| Turn-by-turn interaction | Shows what artifacts and absences materially equip the episode. | Sequence organization, timing, turn allocation, embodied conduct, subtle repair. | Conversation Analysis; Interaction-Order Sociology. |
| Face, embarrassment, and ritual | Shows how interactional consequences can be delegated and carried forward. | Why particular breaches threaten face or why particular remedial rituals restore workable relations. | Interaction-Order Sociology. |
| Motivation and identification | Traces attachments, participation, and material investments. | Individual desire, personality, affective commitment, identification, and subjective meaning. | Social psychology; pragmatist or phenomenological approaches depending the question. |
| Normative evaluation | Shows how authority, representation, and exclusions are assembled. | Whether the arrangement is fair, legitimate, democratic, or desirable. | Political theory, ethics, deliberative theory, or explicit normative criteria. |
| Structural inequality | Demands a mechanism showing how durable asymmetry is reproduced through associations. | Can understate historically entrenched class, race, gender, organizational, or institutional structures if tracing remains narrowly local. | Bourdieu, feminist STS, institutional analysis, critical sociology. |
| Network infinity | Encourages following consequential associations beyond conventional boundaries. | Every actor connects to further actors, so “follow the network” supplies no automatic stopping rule. | Question-relative sampling and explicit analytic boundaries. |
| Twenty-person scope | Makes mediation visible at a scale where direct mutual recognition still exists. | Does not establish a universal threshold for group formation or generalize automatically to crowds, firms, or mass publics. | Group-size research, organizational sociology, network science. |
| Observer effects and privacy | Encourages attention to the researcher’s own inscriptions and categories. | A dense field protocol can itself alter a small group and make anonymization difficult. | Reflexive ethnography and rigorous research ethics. |
Core ANT sources used here
Bruno Latour (2005), Reassembling the Social. The primary mature-ANT source: especially “No Group, Only Group Formation,” distributed action, objects as participants, and the methodological injunction to trace associations rather than invoke a pre-assembled social domain.
Bruno Latour (1996), “On Interobjectivity.” The crucial bridge to microsociology: objects are reintroduced into the definition of interaction in order to rethink the relation between micro and macro social order. Latour archive
Bruno Latour (1986), “The Powers of Association.” Power is treated as an effect of association, translation, and the ability to mobilize durable resources rather than as an unexplained substance.
Bruno Latour (1992), “Where Are the Missing Masses?” Supplies the vocabulary of delegation and prescription into mundane artifacts.
John Law (1992), “Notes on the Theory of the Actor-Network.” The clearest source here for heterogeneous ordering, relational power, and punctualization: a network can be simplified into something that passes as one actor when its heterogeneous elements are held together. Lancaster PDF
Interaction-Order Sociology: complementary sources
Erving Goffman (1959), The Presentation of Self in Everyday Life. Dramaturgical performance, teams, regions, impression management. Useful here, but only one part of the broader interaction-order tradition.
Erving Goffman (1967), Interaction Ritual. Face-work, deference, demeanor, embarrassment, and the ritual maintenance of workable encounters.
Erving Goffman (1974), Frame Analysis. How participants organize experience by identifying what kind of activity is occurring and what interpretive frame applies.
Erving Goffman (1981), Forms of Talk. Footing and participation frameworks: speaking is decomposed into multiple production and reception roles rather than a simple speaker/hearer pair.
Erving Goffman (1983), “The Interaction Order.” The strongest umbrella statement for treating situated interaction as a substantive domain with its own enabling conventions and vulnerabilities.