Conversation and dialogue

Conversation is not simply two people taking turns to transmit finished messages. It is a reciprocal coordination system in which participants continuously produce, test, interpret, correct and revise meaning together.

Spoken language gave humans a flexible way to encode complex meaning in sound. Conversation added something different and equally consequential: feedback. A recipient could interrupt, question, reject, clarify, confirm, reinterpret or extend what had just been said. Information no longer had to travel as a one-way package whose defects were discovered only after the damage had reached the village.

Conversation therefore reduced several constraints at once:

• uncertainty about whether a message had been received;
• uncertainty about whether it had been understood;
• inability to correct errors immediately;
• difficulty coordinating intentions in changing situations;
• difficulty building shared knowledge between participants;
• difficulty negotiating commitments, roles and joint action.

Its core machinery includes:

• turn-taking, which coordinates access to the communicative floor;
• sequence organisation, through which one action makes another relevant, such as a question inviting an answer;
• repair, through which problems in speaking, hearing or understanding are detected and corrected;
• grounding, through which participants accumulate evidence that they understand one another well enough for present purposes;
• alignment, through which speakers partially converge on words, structures, rhythms, references and situation models;
• pragmatic inference, through which listeners infer more than the literal words state;
• embodied signalling, including gaze, gesture, facial expression, posture, timing and prosody.

Conversation is ancient, but it has no defensible invention date. Its origins lie in deep prehistory and cannot be directly observed. Comparative studies find turn-taking patterns in several non-human species, while human infants participate in conversation-like exchanges before they possess mature language. These findings suggest that parts of the interactional machinery may be older than fully developed spoken language, but they do not establish a complete evolutionary sequence.[10][11][12]

The practice is also not limited to speech. Signed conversations display tightly coordinated turn-taking, and later technologies allowed conversation through writing, telephony, video, instant messaging and human-machine interfaces.[9] Conversation is therefore best understood as an interaction protocol rather than a particular sensory channel.

Its civilisational contribution is difficult to overstate. Conversation supports:

• cooperative work;
• teaching and apprenticeship;
• negotiation and exchange;
• dispute resolution;
• family formation and social bonding;
• governance and deliberation;
• diagnosis and care;
• planning and command;
• scientific criticism;
• storytelling and collective interpretation.

Yet conversation is not naturally equal, truthful or benevolent. The ability to control turns, topics, questions, interruptions and acceptable responses can reproduce institutional power. Conversation can also spread rumour, manipulate memory, enforce conformity, exhaust participants, obscure accountability and convert unequal relationships into performances of consent.[14][15]

The most important analytical conclusion is:

Speech makes complex meaning transmissible. Conversation makes meaning corrigible, negotiable and jointly actionable.

The big idea

| Dimension | Finding | |---|---| | Subject | Conversation and dialogue | | Register name | Conversation and dialogue | | Topic type | Practice and interaction protocol | | Current primary category | Distribution and amplification | | Recommended analytical category | Interaction, feedback and coordination | | Main breakthrough | Reciprocal exchange with immediate feedback and repair | | Main problem addressed | Uncertainty about understanding and coordinated action | | Emergence | Deep prehistory; no defensible single date | | Origin confidence | Low | | Mechanism confidence | High | | Civilisational significance | Foundational | | Unaided persistence | Very low | | Unaided reach | Very low | | Interactivity | Extremely high | | Repairability | Extremely high during live interaction | | Scalability | Low without organisations or technology | | Principal predecessors | Embodied communication; spoken language; biological memory; shared attention | | Principal descendants | Teaching; negotiation; interviews; deliberation; telephony; chat; forums; conversational AI | | Present status | Active, ubiquitous and continually remediated |

| Field | Value | |---|---| | Subject | Conversation and dialogue | | Register name | Conversation and dialogue | | Topic type | Practice and interaction protocol | | Primary category in register v0.1 | Distribution and amplification | | Secondary categories | Feedback; interpretation; encoding; governance; trust | | Era | Era I: Embodied and Oral Communication | | Suggested confidence label | Established mechanism; uncertain origin |

2.1 Taxonomy pressure point

The register places conversation under Distribution and amplification. That classification is understandable because conversation circulates information among participants. It is not, however, a perfect fit.

Conversation usually does not amplify a message to a larger audience. Its distinctive contribution is that it turns recipients into participants who can shape what happens next. The topic therefore exposes a gap in the Master Specification: interaction and coordination are currently spread across distribution, interpretation and feedback rather than represented as a first-class layer.

Three options should be considered for v0.2:

1. retain Conversation and dialogue for identifier stability but add Interaction and coordination as a formal secondary category;
2. create a new taxonomy layer while retaining the existing ID as a historical alias;
3. broaden Distribution and amplification so that it explicitly includes reciprocal routing, participation and feedback.

The first option is recommended until several more dossiers test whether the gap recurs.

2.2 Conversation, dialogue and related terms

These terms overlap but are not identical.

Conversation

A flexible, usually reciprocal interaction in which participants coordinate turns and actions. It may be informal, social, practical, ceremonial or institutional.

Dialogue

A reciprocal exchange that often carries a stronger implication of sustained engagement, inquiry, negotiation or structured interaction. Dialogue may be spontaneous, formally organised, literary or technologically mediated.

Discussion

Conversation organised around examining a topic, problem or decision.

Debate

An adversarial or contrastive exchange in which participants defend competing positions, often before an audience or according to formal rules.

Interview

An asymmetrical dialogue in which one participant is institutionally authorised to ask most of the questions.

Interrogation

A highly asymmetrical exchange organised around extracting information, testing claims or securing admissions.

Chat

A broad modern label for informal conversation, especially when mediated by text or digital platforms.

This research notes uses conversation and dialogue as an umbrella for reciprocal, sequential interaction. It does not assume that the exchange is friendly, equal, truthful or oriented towards consensus.

2.3 Modality correction

Conversation is not inherently spoken.

It can occur through:

• speech;
• natural sign language;
• gesture and gaze;
• written correspondence;
• text chat;
• telephone;
• video communication;
• mixed human-machine interfaces.

The earliest human conversations were probably embodied and vocal, but the topic concerns the organisation of reciprocal interaction, not the physical medium alone.

2.4 Recommended topic boundary

Conversation and dialogue includes:

• reciprocal exchange;
• turn allocation;
• sequence organisation;
• feedback;
• clarification;
• repair;
• grounding;
• conversational alignment;
• topic entry, continuation and closure;
• negotiation of joint action.

It does not absorb:

• Vocalisation, prosody and spoken language Spoken language, which supplies a major encoding channel;
• Oral tradition and storytelling Oral tradition, which specialises in intergenerational distribution and collective memory;
• Imitation, repetition and apprenticeship Imitation and apprenticeship, which reproduces skills and behaviour;
• formal education as an institution;
• public debate, parliament, courts or interviews as separate institutional descendants;
• telephone, messaging or conversational AI as technological extensions.

Conversation and dialogue are reciprocal, sequential forms of joint action in which two or more participants coordinate attention, turns, interpretations and responses in order to exchange information, manage relationships, perform social actions and pursue shared or competing goals.

A minimal conversational event requires:

1. at least two recognisable participants;
2. a channel through which signals can pass;
3. some shared conventions or interpretable cues;
4. alternating or otherwise coordinated contributions;
5. evidence that each contribution is responsive to previous conduct;
6. an opportunity for feedback, uptake or repair.

Conversation is not defined merely by the presence of multiple messages. A row of speeches delivered one after another may remain a sequence of monologues if later speakers do not orient to what preceded them.

3.1 Conversation as joint action

Herbert Clark's joint-action account argues that language use is not adequately described as one person sending and another person receiving. Participants coordinate individual actions into a joint activity.[6]

That framing is particularly useful for this map. Conversation is less like posting two parcels and more like carrying a table through a narrow doorway. Each participant must continuously adjust to the movement, knowledge and intentions of the other. One person can still drop the table, dominate the route or pretend not to hear instructions, but the activity remains interactionally coupled.

3.2 Conversation as social action

Conversational turns do more than describe facts. They can:

• request;
• offer;
• promise;
• accuse;
• apologise;
• invite;
• refuse;
• diagnose;
• command;
• joke;
• console;
• threaten;
• grant permission;
• create or terminate commitments.

Conversation therefore transmits information and changes social reality.

| Attribute | Classification | |---|---| | Participant structure | Primarily one-to-one or many-to-many | | Direction | Bidirectional or multidirectional | | Time relationship | Primarily synchronous; later extended to asynchronous dialogue | | Spatial relationship | Originally co-located; later remote | | Persistence | Ephemeral unless remembered, transcribed or recorded | | Feedback | Immediate or near-immediate | | Turn structure | Alternating, overlapping or collaboratively completed | | Primary modes | Vocal-auditory, visual-manual and multimodal | | Typical authentication | Face, voice, bodily presence, social context, account identity or device credentials | | Typical noise | Physical interference, ambiguity, attention failure, mismatched assumptions, power pressure and hostile inference | | Typical audience | Small, although organisations and media can add overhearers or mass audiences | | Typical control structure | Locally negotiated but socially and institutionally constrained |

4.1 Dyadic and multiparty conversation

Two-person conversation permits relatively simple turn coordination. Multiparty conversation introduces competition for the floor, shifting address, side sequences, coalition formation and uncertainty about who should respond.

The underlying system remains flexible. Participants can nominate the next speaker, self-select, yield, interrupt or remain silent. Group size changes the probability that several people prepare responses at once and may shorten transition times because the fastest entrant captures the floor.[13]

4.2 Synchronous and asynchronous dialogue

Face-to-face conversation is synchronous. Written correspondence introduced slower reciprocal dialogue. Digital systems later created intermediate forms:

• live chat with near-immediate response;
• threaded forums with delayed turns;
• voice notes with asynchronous speech;
• collaborative documents with embedded discussion;
• AI chat with machine-generated response.

The interactional logic survives even when the time scale changes.

4.3 Multimodal organisation

Conversation is assembled from more than words. Participants use:

• gaze to display attention or select recipients;
• gesture to depict, point or regulate turns;
• facial expression to signal stance and understanding;
• posture and orientation to enter or withdraw;
• prosody to project continuation, completion, emphasis or resistance;
• silence and timing to communicate hesitation, refusal, deference or trouble.

Conversation analysis increasingly treats these resources as integrated packages rather than decorative additions to verbal content.[16][17]

Participant A intention, perception or social action ↓
Turn design using language, prosody and embodied cues ↓
Signal enters shared perceptual environment ↓
Participant B interprets content, action and stance ↓
B displays uptake, uncertainty, agreement, resistance or silence ↓
A interprets B's response as evidence about understanding ↓
Participants update common ground, relationship and task state ↓
Next turn is selected or self-selected ↓
Sequence continues, branches, repairs or closes

This loop distinguishes conversation from one-way transmission.

5.1 Feedback is evidence, not certainty

A nod, “yes,” repeated phrase or relevant next action can provide evidence that a message was understood. It does not prove that the internal interpretation is identical.

Conversation therefore operates through sufficient grounding, not perfect mind-reading. Participants continue when they have enough evidence for the current purpose.[7][8]

5.2 Shared state

Conversation accumulates a locally shared history:

• what has been said;
• which terms have been adopted;
• what has been accepted or disputed;
• who knows what;
• which actions are pending;
• which commitments have been made.

This resembles a shared state cache, but the analogy has limits. Human memory is selective, interpretations diverge, and participants may strategically misrepresent what they accepted.

5.3 Mixed initiative

Either participant may redirect the interaction, ask a question, introduce a new goal or repair an earlier assumption. This mixed initiative is a defining strength of human conversation and a persistent challenge for automated dialogue systems.[18]

6.1 Dating conclusion

There is no defensible invention date for conversation.

Conversation leaves no direct fossil. Archaeology can reveal social complexity, symbolic artefacts and cooperative technology, but not the exact organisation of prehistoric turns. Any precise date would be a modern label pinned to a vanished exchange.

The safest claim is:

Conversation emerged in deep prehistory through the interaction of embodied signalling, vocal communication, shared attention, social cognition and increasingly flexible language.

6.2 Evidence streams

A. Comparative turn-taking

Turn-taking or turn-like coordination appears in several animal taxa. Reviews have examined reciprocal timing in birds, mammals, insects and primates, while studies of great apes and marmosets report structured exchanges with some parallels to human interaction.[11][19][20]

These findings support the possibility that parts of the interactional infrastructure are evolutionarily old. They do not show that non-human species possess human conversation with its open-ended symbolic content, common-ground reasoning and complex social actions.

B. Infant protoconversation

Human infants participate in alternating vocal and embodied exchanges before mature speech develops. Caregivers adjust timing, exaggerate responses and treat infant behaviour as meaningful participation. These exchanges are commonly described as protoconversations.[10][21][22]

This indicates that interactional timing and reciprocal expectation develop early and may scaffold language acquisition.

C. Cross-linguistic turn-taking

A study across ten languages found broad similarities in the organisation of response timing, despite measurable cultural variation.[2] This supports a widespread human system oriented towards limiting both excessive overlap and excessive silence.

D. Signed conversation

Signed languages display coordinated turn timing despite using a visual-manual channel. This shows that conversation is not a by-product of breathing pauses or vocal anatomy alone.[9]

E. Social cooperation

Human conversation depends on capacities for shared attention, perspective-taking, intention recognition and cooperative or strategically coordinated action. These capacities are difficult to date independently, but they form part of the social infrastructure needed for sustained dialogue.[23]

6.3 Provisional emergence sequence

| Phase | Description | Confidence | |---|---|---| | Reciprocal biological signalling | Alternating calls, gestures and attention displays | Broadly accepted | | Proto-turn-taking | More regular timing and role alternation | Broadly accepted as a comparative possibility | | Shared-attention exchanges | Participants coordinate reference to people, objects and actions | Broadly accepted | | Linguistically structured conversation | Open-ended turns encode complex symbolic meaning | Date unknown | | Institutional dialogue | Recurrent roles and rules organise teaching, ritual, trade and governance | Inferred from later societies | | Recorded dialogue | Writing preserves stylised or reported exchanges | Historically visible | | Remote conversation | Telephone, radio links, digital chat and video remove co-presence | Historically documented | | Machine-mediated conversation | Systems transcribe, translate, rank or generate turns | Contemporary and documented |

6.4 Conversation may predate full language

Some scholars propose that turn-taking and interactional coordination formed an ancient platform upon which language developed.[3][11] This is plausible, but the relationship should not be presented as settled.

A cautious formulation is:

Elements of reciprocal coordination probably predate fully elaborated spoken language, while human conversation in its modern sense depends on linguistic and social capacities that emerged through a long, entangled process.

7.1 Biological prerequisites

• sensory systems capable of detecting partners and signals;
• motor control for producing vocal, facial or gestural responses;
• attention that can shift between partner, signal and shared environment;
• memory sufficient to retain prior turns;
• arousal regulation that permits waiting, responding and inhibiting interruption.

7.2 Cognitive prerequisites

• recognition of agents and communicative intent;
• expectation that another participant can respond;
• prediction of possible turn completion;
• perspective-taking;
• reference tracking;
• inference from context;
• updating of shared assumptions;
• planning while listening.

Turn transitions are often faster than the time normally required to plan an utterance from scratch. This suggests that listeners begin preparing responses before the current speaker has finished, using content and timing cues to predict when a turn may end.[3][24][25]

7.3 Social prerequisites

• repeated interaction;
• enough mutual tolerance to permit exchange;
• conventions for attention and turn rights;
• incentives to coordinate;
• sanctions for persistent disruption;
• trust or at least predictable self-interest;
• culturally learned norms for politeness, deference and disagreement.

7.4 Linguistic prerequisites

Conversation can occur with limited conventional signals, but complex dialogue benefits from:

• a shared language or translation mechanism;
• lexical and grammatical resources;
• pragmatic conventions;
• devices for reference, tense, modality and stance;
• metacommunicative expressions such as “What do you mean?” or “Let me rephrase that.”

7.5 Environmental prerequisites

• a channel with manageable noise;
• enough mutual access to signals;
• temporal availability;
• a setting in which responses are possible;
• privacy where sensitive exchange requires it.

Conversation appears casual because skilled participants perform a large amount of coordination automatically. Underneath the surface lies an intricate system that would make a badly documented network protocol blush.

8.1 Turn-taking

Sacks, Schegloff and Jefferson described a flexible system for organising who speaks next. At possible completion points:

1. the current speaker may select the next speaker;
2. if nobody is selected, another participant may self-select;
3. if nobody enters, the current speaker may continue.[1]

The system is designed to minimise long gaps and extensive overlap without requiring a central controller.

Turn-construction units

Participants project possible turn completion from combinations of:

• syntax;
• prosody;
• meaning;
• gesture;
• gaze;
• activity context.

Transition relevance

A possible completion creates an opportunity for another participant to enter. This does not mean every transition is neat. Overlap, interruption and silence are normal resources whose meaning depends on context.

8.2 Sequence organisation

Turns occur in organised sequences rather than as isolated sentences.

Common paired actions include:

• greeting → greeting;
• question → answer;
• request → grant or refusal;
• invitation → acceptance or rejection;
• accusation → denial, admission or counter-accusation;
• assessment → agreement or disagreement;
• farewell → farewell.

These structures are often called adjacency pairs. The second action is not mechanically forced, but the first makes a class of responses socially relevant.[26]

Conditional relevance

After a question, silence is noticeable. After an invitation, a sudden discussion of weather may be interpreted as avoidance. Conversation therefore creates local expectations that help participants detect missing, delayed or resistant responses.

8.3 Repair

Repair is the machinery through which participants address problems in:

• speaking;
• hearing;
• word selection;
• reference;
• understanding;
• factual formulation;
• action recognition.

Examples:

• “Tuesday, sorry, Wednesday.”
• “Who do you mean?”
• “Did you say fifteen or fifty?”
• “No, I meant the other account.”
• “So you are saying the contract ends in May?”

Research in conversation analysis distinguishes who initiates repair and who completes it. A recurring pattern favours giving the original speaker an opportunity to correct their own trouble rather than immediately replacing their formulation.[4]

Cross-linguistic work has identified widely shared systems for initiating repair, including brief forms equivalent to “huh?” or “what?” and more specific requests for clarification.[5]

Repair is conversation's error-correction layer. Unlike a perfect checksum, it relies on fallible humans who may not notice the corruption, may repair the wrong thing or may strategically refuse correction.

8.4 Grounding and common ground

Participants need enough evidence that their contributions have been understood for current purposes. They establish this through:

• acknowledgements;
• relevant responses;
• repetition;
• paraphrase;
• completion of requested action;
• correction;
• continued use of an agreed term.

Common ground includes information participants treat as mutually available. It can derive from shared culture, shared perception, previous interaction or the current conversation.[7][8]

Common ground is not a perfectly synchronised database. Participants can disagree about what was established, forget earlier turns or falsely claim that “we already agreed.”

8.5 Alignment and entrainment

Conversational partners often begin to reuse one another's:

• vocabulary;
• syntactic patterns;
• pronunciation features;
• timing;
• gestures;
• descriptions;
• conceptual framing.

The Interactive Alignment account proposes that partial alignment across multiple levels helps dialogue proceed efficiently.[27][28][29]

Alignment can reduce coordination cost, but it is not always evidence of agreement. A person can reuse another's wording while challenging the underlying claim.

8.6 Pragmatic inference

Literal wording rarely carries the whole message.

If someone asks, “Are you coming to the meeting?” and receives “I have a hospital appointment,” the relevance of the response is inferred rather than explicitly encoded.

Grice's cooperative principle and conversational maxims remain influential tools for explaining how participants infer unstated meaning from expectations about relevance, quantity, quality and manner.[30]

The model should not be mistaken for a moral claim that speakers are always cooperative. Participants may lie, joke, evade, threaten or manipulate while still exploiting shared expectations about how conversation normally works.

8.7 Preference organisation

Some responses are interactionally easier to deliver than others.

Agreement, acceptance and straightforward answers are often produced promptly. Disagreement, refusal and other socially delicate responses may be delayed, prefaced, qualified or accompanied by gaze aversion.[31][32]

“Preference” in conversation analysis refers to structural organisation, not necessarily private desire. A person may desperately want to say no and still wrap the refusal in three cushions, a weather report and an apology.

8.8 Topic management

Participants negotiate:

• when a topic begins;
• whether it is relevant;
• how long it continues;
• when a side sequence is permitted;
• whether an earlier topic is resumed;
• when the conversation may close.

Topic control is both a coordination mechanism and a source of power.

8.9 Openings and closings

Conversation requires entry and exit work.

Openings establish:

• availability;
• identity;
• attention;
• relationship;
• reason for contact.

Closings establish that no unresolved business prevents departure. Telephone research showed that even apparently simple openings are sequentially organised rather than improvised from nothing.[33]

8.10 Silence

Silence can indicate:

• thinking;
• refusal;
• shock;
• deference;
• hostility;
• emotional overwhelm;
• lack of access;
• technical failure;
• a culturally appropriate pause.

Its interpretation depends on sequence, relationship and context. Silence is not the absence of communication. Sometimes it arrives carrying a suitcase.

9.1 The one-way transmission problem

A complex message can be spoken without becoming mutually understood.

One-way speech leaves several unresolved questions:

• Was the signal heard?
• Was the intended person listening?
• Did the listener recognise the words?
• Did both parties attach the same meaning?
• Does the listener accept the premise?
• Can the instruction be followed?
• Has the situation changed?
• Is further detail required?

9.2 Constraint reduced

Conversation reduces the coordination and verification constraint.

It allows participants to:

• request missing information;
• test interpretations;
• challenge assumptions;
• correct errors;
• update plans;
• negotiate conflicting interests;
• establish commitments;
• respond to changing conditions.

9.3 Information-theoretic interpretation

Conversation adds a human feedback channel.

A simplified model is:

Message → Response → Confirmation or repair → Updated message

This resembles acknowledgment, retransmission and error correction in engineered systems. The analogy is useful but incomplete:

• humans do not share perfectly specified protocols;
• meaning is context-dependent;
• social power affects who may request repair;
• participants can deceive;
• emotional and relational outcomes matter alongside task accuracy.

9.4 Constraints not solved

Conversation does not automatically solve:

• long-distance communication;
• durable storage;
• large-scale distribution;
• objective truth;
• equality of participation;
• accountability;
• translation across languages;
• protection from coercion;
• exact preservation of wording.

Later topics address these weaknesses only partially.

| Dimension | Unaided face-to-face conversation | Notes | |---|---|---| | Reach | Very low | Limited by physical presence, sight and hearing | | Latency | Very low | Responses can follow within fractions of a second | | Bandwidth | Moderate to high | Words, prosody, gesture, gaze and facial expression combine | | Fidelity | Moderate | Immediate repair helps, but ambiguity and memory remain | | Persistence | Very low | Sound and gesture vanish unless remembered or recorded | | Replication cost | Low per retelling, high in fidelity loss | Repetition requires time and another performance | | Distribution cost | Low for small groups | Cost rises sharply with audience size and distance | | Accessibility | High but unequal | Requires shared code and sensory or technological access | | Portability | Very high | Biological system travels with participants | | Interactivity | Extremely high | Recipients can become speakers immediately | | Feedback quality | Very high | Includes verbal and embodied evidence | | Repairability | Extremely high | Problems can be addressed as they emerge | | Searchability | Very low | Unrecorded talk cannot be searched reliably | | Editability | High during exchange; low after disappearance | Speakers can reformulate but cannot erase memory | | Scalability | Low | Turn competition rises with group size | | Authentication | High in familiar face-to-face settings | Lower with strangers, disguise or mediated identity | | Privacy | Variable | Private in small settings, vulnerable to overhearing and recording | | Censorship resistance | Moderate locally | Private talk is difficult to eliminate, but participants can be monitored or punished | | Infrastructure dependence | Very low unaided | Remote versions depend on networks and devices | | Energy dependence | Low | Biological attention and vocal or motor effort | | Interpretive burden | Moderate to high | Shared language, context and social inference are required | | Attention demand | High | Skilled participation requires continuous monitoring | | Participation symmetry | Variable | Can range from reciprocal equality to severe domination | | Context sensitivity | Very high | Meaning changes with sequence, relationship and setting |

11.1 Immediate feedback

Participants can display understanding or trouble before the interaction moves too far.

11.2 Local error correction

A misunderstood word can be repaired without retransmitting the entire conversation.

11.3 Adaptive bandwidth

Speakers can expand, compress, exemplify or simplify a message in response to visible need.

11.4 Joint problem-solving

Participants can divide reasoning across several minds, propose alternatives and test consequences.

11.5 Negotiation

Conversation allows interests, prices, duties, boundaries and commitments to be adjusted rather than merely announced.

11.6 Emotional calibration

Tone, pace, gaze and response permit participants to alter their approach when another person appears confused, angry, frightened or unconvinced.

11.7 Relationship maintenance

Much conversation exists not to transfer novel facts but to maintain affiliation, trust, identity and mutual attention.

11.8 Low material cost

Unaided conversation requires no writing surface, machine or permanent infrastructure.

11.9 Flexibility

The same interactional system can support gossip, diagnosis, prayer, command, flirtation, bargaining, teaching and scientific criticism.

11.10 Distributed cognition

Participants can externalise partial thoughts, receive prompts and construct conclusions that neither held fully at the beginning.

12.1 Survival

Conversation improves adaptive coordination:

• identifying danger;
• revising plans;
• assigning roles;
• checking whether instructions were understood;
• reporting changing conditions;
• integrating observations from several people.

12.2 Cooperation

Conversation lets groups coordinate actions whose details cannot be fixed in advance. It supports planning, monitoring, correction and renegotiation during collective work.

12.3 Family and social organisation

Conversation supports:

• courtship;
• caregiving;
• conflict repair;
• transmission of household expectations;
• identity formation;
• emotional reassurance;
• negotiation of boundaries.

Infant-caregiver protoconversation also provides an early environment for social and communicative development.[10][21]

12.4 Teaching and apprenticeship

A learner can:

• ask why;
• request demonstration;
• expose misunderstanding;
• receive tailored explanation;
• attempt a task and obtain correction.

This makes dialogue a powerful complement to imitation.

12.5 Trade and exchange

Conversation enables bargaining, clarification of quantities, testing of trust and formation of agreements. Before durable contracts, repeated social interaction and witnessed dialogue helped sustain commitments, although with weaker accountability.

12.6 Governance

Councils, audiences, assemblies and negotiations depend on organised talk. Conversation can surface local knowledge and objections that one-way proclamation suppresses.

It can also merely stage participation while decision power remains elsewhere.

12.7 Law and dispute resolution

Questions, testimony, challenge and response help organisations evaluate competing accounts. The same tools can produce insight or coercion depending on rules and power.

12.8 Medicine and care

Clinical interaction gathers symptoms, explains options, delivers diagnoses and negotiates treatment. Conversation analysis shows that turn design affects patient participation and response.[15][34]

12.9 Science and engineering

Dialogue supports:

• criticism;
• hypothesis refinement;
• troubleshooting;
• peer questioning;
• collaborative design;
• tacit knowledge transfer.

Written publication preserves conclusions, but conversation often reveals the assumptions and doubts that polished documents quietly bury under the carpet.

12.10 Religion and philosophy

Question-and-answer formats, commentary, confession, instruction and disputation allow doctrines and moral claims to be examined, transmitted or enforced.

12.11 Warfare and diplomacy

Conversation supports command, intelligence gathering, interrogation, negotiation, alliance formation and surrender. It can reduce violence through settlement or prepare violence with chilling efficiency.

12.12 Culture and entertainment

Conversation generates stories, jokes, dramatic scenes, interviews, commentary and public personalities. Recorded conversation later became a major media genre.

12.13 Social bonding and well-being

Conversation is one mechanism through which social relationships are created and maintained. Broader research associates social connection with important psychological and physical outcomes, though conversation alone should not be treated as a medical cure.[35]

13.1 Roles created or strengthened

Conversation supports specialised roles such as:

• teacher;
• negotiator;
• mediator;
• counsellor;
• interviewer;
• judge;
• advocate;
• diplomat;
• salesperson;
• therapist;
• journalist;
• moderator;
• interpreter;
• call-centre operator;
• conversational system designer.

13.2 Control of the floor

Power can appear through:

• speaking longer;
• interrupting without sanction;
• selecting the next speaker;
• refusing another participant's topic;
• defining what counts as relevant;
• demanding answers while withholding one's own;
• closing the interaction;
• controlling the record.

13.3 Epistemic authority

Participants negotiate who has the right to know, describe or decide.

A doctor, elder, witness, technician, parent or official may possess domain authority. That authority can be legitimate, exaggerated or contested turn by turn.[14]

13.4 Institutional talk

Institutional conversations differ from casual talk because roles and goals constrain participation.

Examples:

• doctor and patient;
• police officer and suspect;
• teacher and pupil;
• interviewer and applicant;
• judge and witness;
• customer and service agent.

The institution may determine:

• who asks questions;
• which answers are acceptable;
• how long turns may last;
• what becomes part of the official record;
• what consequences follow.

13.5 Politeness and deference

Norms of politeness can reduce friction, preserve dignity and make disagreement possible. They can also hide domination by requiring weaker participants to package urgent truth as a delicate suggestion.

13.6 Language, accent and status

Participants may be judged by:

• vocabulary;
• accent;
• fluency;
• speed;
• confidence;
• familiarity with institutional formats.

Conversational competence is therefore partly a social resource. People who do not know the hidden script of an interview, hearing or boardroom may appear less capable than they are.

13.7 Moderation

Multiparty dialogue often creates a moderator role to allocate turns, enforce scope and prevent escalation. Moderation can protect participation or become a gatekeeping mechanism.

14.1 Ephemerality

Unrecorded conversation disappears as a signal and survives only through memory, retelling or consequence.

14.2 Limited reach

Unaided conversation requires proximity or relay.

14.3 Low scalability

As group size grows, turn competition and coordination cost increase.

14.4 Sequential bottleneck

Only a small number of people can speak intelligibly at once. Parallel ideas must queue, split into side conversations or become lost.

14.5 Memory dependence

Participants may forget wording, sequence, commitments or qualifications.

14.6 Ambiguity

Meaning depends on context, inference and shared assumptions. Two people can leave the same exchange carrying incompatible versions of what happened.

14.7 Unequal participation

Status, personality, language ability, disability, gender norms and institutional power can shape who speaks and who is heard.

14.8 Attention cost

Conversation demands ongoing monitoring. Distraction can break reference, timing and trust.

14.9 Emotional volatility

Immediate feedback permits rapid repair, but it also permits rapid escalation.

14.10 Weak auditability

Without a record, participants may dispute what was said.

14.11 Context dependence

A transcript can omit tone, gaze, relationship and situational knowledge needed to interpret the exchange.

14.12 Language barriers

Conversation requires shared codes or interpretation.

14.13 Accessibility barriers

Hearing, speech, vision, motor, cognitive and sensory differences can affect access unless alternative modes or supports exist.

14.14 Vulnerability to noise

Physical sound, poor connection, cross-talk, emotional arousal and prejudice can all corrupt uptake.

15.1 Deception

Feedback helps a liar adapt the story to the listener's doubts.

15.2 Manipulation

Skilled speakers can exploit timing, framing, emotional pressure and social norms.

15.3 Coercion

Interrogation and unequal negotiation can convert conversational form into compelled participation.

15.4 Rumour

Conversation can reproduce unverified claims rapidly through trusted networks.

15.5 Group conformity

Immediate social feedback can punish dissent before it is fully articulated.

15.6 Polarisation

Conversation may improve understanding, but hostile exchanges can harden identity and reward performative certainty.

15.7 Harassment and abuse

Reciprocal access provides repeated opportunities for insult, intimidation and unwanted contact.

15.8 False consensus

Silence, politeness or powerlessness may be interpreted as agreement.

15.9 Conversational exhaustion

Constant availability creates emotional and cognitive load, especially in care, service and digital work.

15.10 Privacy loss

Conversation can be overheard, recorded or reported. Technological mediation often expands the number of invisible listeners.

15.11 Memory contamination

Later conversation can reshape recollection. Participants may converge on a shared account that is coherent but inaccurate.

15.12 Accountability theatre

Organisations may invite dialogue without allowing responses to alter decisions.

15.13 Persuasion outrunning truth

Fluency, confidence and charisma can make weak claims feel stronger than well-supported but awkwardly expressed ones.

| Topic or capability | Relationship | Contribution | |---|---|---| | Embodied non-verbal communication Embodied non-verbal communication | Direct predecessor and continuing complement | Gaze, gesture, facial expression, stance and turn regulation | | Vocalisation, prosody and spoken language Vocalisation, prosody and spoken language | Primary encoding predecessor | Complex symbolic content and vocal timing | | Neural and cognitive memory Biological memory | Prerequisite | Retains prior turns and shared history | | Shared attention | Cognitive prerequisite | Establishes common reference | | Social cognition | Cognitive prerequisite | Supports intention and perspective inference | | Reciprocal signalling | Evolutionary precursor | Alternation and response expectation | | Group norms | Social prerequisite | Regulates interruption, relevance and acceptable action |

Conversation should not be depicted as replacing embodied or vocal communication. It organises them into a reciprocal system.

| Descendant or extension | Relationship | Limitation addressed or capability added | |---|---|---| | Imitation, repetition and apprenticeship Imitation, repetition and apprenticeship | Complement and specialisation | Converts dialogue into guided skill transfer | | Oral tradition and storytelling Oral tradition and storytelling | Extension | Carries conversationally shaped knowledge across generations | | Formal teaching | Institutional specialisation | Organises questions, explanation and evaluation | | Negotiation and mediation | Institutional specialisation | Structures conflicting interests and settlement | | Councils and deliberative assemblies | Scaling institution | Extends dialogue to collective governance | | Interviews and journalism | Role specialisation | Organises questioning for public information | | Courts and hearings | Governance specialisation | Formalises testimony, challenge and decision | | Written correspondence | Modal extension | Preserves delayed reciprocal exchange | | Literary dialogue | Representational extension | Models conversation in durable text | | Telephone | Transmission extension | Removes co-location while preserving live voice feedback | | Radio call-ins and remote links | Broadcast convergence | Adds public audiences to dialogue | | Bulletin-board systems, Usenet and Internet chat BBS, Usenet and Internet chat | Network extension | Enables persistent and remote many-to-many dialogue | | Instant Messaging and Chat Applications Instant messaging and chat applications | Mobile extension | Combines presence, history and rapid response | | Video conferencing | Multimodal remote extension | Restores some gaze and gesture cues | | Forums and comment threads | Asynchronous scaling | Permits delayed, persistent multi-party dialogue | | Machine Translation Machine translation | Interpretive mediation | Reduces language barriers | | Speech Recognition and Automated Transcription Speech recognition and transcription | Persistence and discovery extension | Converts conversation into searchable text | | Conversational AI Assistants and Retrieval-Augmented Generation Conversational AI | Machine participant simulation | Generates responsive turns and mixed-initiative interfaces |

17.1 Successor logic

Later systems usually preserve one strength of conversation while sacrificing another.

• Writing increases persistence but slows feedback.
• Telephone preserves immediacy but removes much embodied context.
• Forums scale participation but fragment timing and common ground.
• Messaging preserves history but increases availability pressure.
• AI offers rapid response but may simulate understanding without possessing human commitments, experience or accountability.

Conversation survived every later medium because no successor combines all of its strengths.

It remains valuable for:

• ambiguous problems;
• emotionally sensitive decisions;
• rapidly changing situations;
• negotiation;
• diagnosis;
• trust formation;
• tacit knowledge;
• collaborative reasoning;
• conflict repair;
• personalised teaching.

Documents excel at stable reference. Broadcast excels at scale. Databases excel at retrieval. Conversation excels when the problem itself changes as participants learn what the other person means.

19.1 Evidence limitation

The first conversations are inaccessible. The following scenarios illustrate capabilities rather than claiming direct transcripts from prehistory.

19.2 Coordinating a hunt or collective task

A participant proposes a plan. Another identifies a hazard. A third supplies missing local knowledge. The group revises roles before acting.

The important advance is not the original proposal. It is the ability to alter the plan through response.

19.3 Teaching a difficult skill

A learner imitates a movement, receives correction, asks a question and tries again. Dialogue converts demonstration into adaptive instruction.

19.4 Negotiating exchange

Two parties clarify quantity, timing, quality and obligation. Each tests the other's willingness to proceed.

19.5 Repairing misunderstanding

A name is misheard. The recipient requests clarification. The speaker repeats it with additional context. The interaction resumes without discarding the entire exchange.

19.6 Plato's dialogues

Ancient philosophical dialogues preserve a literary model in which questioning, contradiction and revision become instruments of inquiry. They are not transcripts of the first dialogues, and their literary form may idealise interaction, but they reveal the enduring association between dialogue and reasoning.

19.7 Telephone conversation

The telephone removed shared physical space while preserving rapid turn exchange. It also made openings more important because participants could no longer rely on visible approach, gaze or shared setting.[33]

19.8 Clinical consultation

A patient's description, a clinician's questions and subsequent clarification demonstrate how institutional dialogue converts experience into a working diagnosis. The sequence also reveals asymmetry: one participant typically controls questions, categories and official conclusions.[15]

19.9 Digital group chat

Multiple participants contribute across time, quote earlier turns, attach media and create persistent history. The system scales conversation but introduces notification overload, ambiguous tone and fragmented attention.

19.10 Conversational AI

A user asks a question, receives an answer, challenges an assumption and requests revision. The surface resembles dialogue. The deeper question is whether the system possesses common ground, commitments and understanding comparable to human partners, or produces sufficiently responsive language without them.[18][36]

| Topic | Primary achievement | What it lacks compared with conversation | |---|---|---| | Embodied non-verbal communication Embodied non-verbal communication | Expresses attention, emotion and intent | Limited open-ended symbolic detail | | Vocalisation, prosody and spoken language Spoken language | Encodes complex meaning in sound | Does not by itself require reciprocal feedback | | Neural and cognitive memory Biological memory | Preserves experience within people | Does not distribute or verify it | | Imitation, repetition and apprenticeship Imitation and apprenticeship | Reproduces behaviour and skill | Can operate with limited explicit negotiation | | Oral tradition and storytelling Oral tradition | Preserves and distributes knowledge across generations | Often privileges stable performance over immediate co-construction | | Song, music and chant Song, music and chant | Increases memorability and emotional synchrony | Less suited to unrestricted propositional exchange | | Long-distance acoustic and visual signals Long-distance signals | Extends simple alerts beyond normal range | Low semantic complexity and weak repair |

20.1 Key distinction from spoken language

Spoken language asks:
How can complex meaning be encoded in sound? Conversation asks:
How can participants coordinate meaning and action through responsive turns?

20.2 Key distinction from oral tradition

Oral tradition optimises memory, continuity and group identity. Conversation optimises local adaptation, feedback and negotiation.

A story may be repeated with ritual stability. A conversation may abandon its original route because somebody asks the right question.

| Claim ID | Claim | Confidence | Basis | Publication note | |---|---|---|---|---| | DIS001-C01 | Conversation has no defensible single invention date | High | No direct fossil or recording; origin inferred indirectly | Use “deep prehistory” | | DIS001-C02 | Conversation is an interaction protocol rather than a single medium | High | Spoken, signed, written and digital dialogue share reciprocal organisation | Well supported | | DIS001-C03 | Turn-taking tends to minimise long gaps and extensive overlap | High | Conversation analysis and cross-linguistic timing studies [1][2][3] | Avoid claiming zero overlap | | DIS001-C04 | Broad turn-timing similarities occur across diverse languages | High | Ten-language comparative study [2] | Mention cultural variation | | DIS001-C05 | Participants often begin response planning before the current turn ends | High | Timing and prediction research [3][24][25] | Well supported with citation | | DIS001-C06 | Repair addresses problems in speaking, hearing and understanding | High | Conversation analysis literature [4][5] | Well supported | | DIS001-C07 | Conversational repair often favours self-correction by the original speaker | High | Schegloff, Jefferson and Sacks [4] | Phrase as organisational preference | | DIS001-C08 | Human languages share some basic repair practices | High | Cross-linguistic study [5] | Avoid claiming complete universality | | DIS001-C09 | Infants engage in conversation-like turn-taking before mature language | High | Developmental studies [10][21][22] | Use “protoconversation” | | DIS001-C10 | Signed conversations exhibit coordinated turn timing | High | Sign-language studies [9] | Supports modality-neutral definition | | DIS001-C11 | Turn-taking analogues occur in several non-human species | High | Comparative reviews and studies [11][19][20] | Do not equate with human dialogue | | DIS001-C12 | Language use can be analysed as joint action | Moderate to high | Clark's theoretical framework [6] | Identify as framework, not sole consensus | | DIS001-C13 | Common ground grows through interaction and partner-specific experience | High | Psycholinguistic research [7][8] | Avoid implying perfect mutual knowledge | | DIS001-C14 | Conversational partners often align on linguistic and multimodal features | High | Alignment research [27][28][29] | Alignment does not equal agreement | | DIS001-C15 | Conversation performs social actions, not merely information transfer | High | Conversation analysis and pragmatics [15][34] | Well supported | | DIS001-C16 | Institutional roles create asymmetries in turn rights and topic control | High | Conversation-analysis research [14][15] | Use setting-specific examples | | DIS001-C17 | Conversation's unaided persistence and scalability are low | High | Mechanism-level inference | No numerical score required | | DIS001-C18 | Human-machine dialogue remains challenged by grounding, mixed initiative and co-construction | Moderate to high | Human-agent communication research [18][36] | Rapidly changing field; reverify |

22.1 Evolution

• Which components of human turn-taking are homologous with other primates, and which evolved independently?
• Did reciprocal interaction provide a scaffold for language, or did richer language transform an older turn system?
• How much did cooperative childcare influence conversational timing and motivation?

22.2 Universality and culture

• Which turn-taking patterns are genuinely universal?
• How should cultural variation in silence, overlap, gaze and interruption be represented without caricature?
• How do multilingual communities negotiate conversational norms?

22.3 Cognition

• How do participants plan responses while continuing to listen?
• How much common ground is explicitly represented versus inferred through simpler mechanisms?
• When does alignment improve understanding, and when does it merely create superficial convergence?

22.4 Power

• How should the dataset score participation symmetry?
• Which institutional formats best protect weaker participants?
• How do transcription and recording redistribute power?

22.5 Technology

• Which conversational cues are lost or distorted by telephone, text and video?
• Does persistent chat improve accountability or create self-censorship?
• Which features distinguish genuine conversational competence from fluent response generation?

22.6 Measurement

• Can repair rate serve as a useful comparative metric across media?
• How should latency be measured in asynchronous dialogue?
• Can conversational quality be evaluated without imposing one culture's norms?

1. Add stronger African and Indigenous scholarship on conversational norms and oral interaction.
2. Include dedicated research on multilingual dialogue and interpreter-mediated conversation.
3. Add disability-centred sources covering Deaf interaction, augmentative communication and conversational access.
4. Expand institutional cases beyond healthcare to courts, classrooms, workplaces and political assemblies.
5. Verify the historical development of recorded dialogue and interview formats.
6. Add research on conversation, social bonding and well-being without collapsing social connection into talk alone.
7. Compare face-to-face, telephone, text, video and AI dialogue using a common metric set.
8. Decide whether Interaction and coordination becomes a formal taxonomy category.
9. Review whether Conversation and dialogue should be split into:
   • spontaneous conversation;
   • structured dialogue;
   • institutional interaction.
10. Conduct an independent source and claim audit.

24.1 Conversation feedback loop

Turn → Uptake → Grounding → Repair → Updated turn

Show meaning becoming more reliable through repeated cycles.

24.2 Turn-taking timeline

A millisecond-scale graphic showing:

• current speaker's turn;
• listener prediction;
• response planning;
• transition gap;
• next turn.

24.3 Repair tree

Trouble detected
├── Speaker self-repairs
├── Listener requests clarification
│ ├── Open request: “What?”
│ └── Specific request: “Fifteen or fifty?”
└── Listener proposes correction

24.4 Common-ground pool

Two participants contribute facts, labels and commitments into a shared but imperfect pool. Faded or disputed items show that common ground is not a flawless database.

24.5 Conversation stack

Layers:

1. biological signal;
2. language;
3. turn timing;
4. social action;
5. common ground;
6. relationship and power;
7. institutional or technological rules.

24.6 The first protocol

Translate conversational machinery into a programmer-friendly schematic:

• floor allocation = token passing;
• acknowledgement = ACK;
• clarification = error report;
• repetition = retransmission;
• common ground = shared state;
• topic closure = session termination.

A visible warning should state that this is an analogy, not a claim that humans run on TCP with feelings bolted on.

24.7 Power map

Show how control enters at different points:

• who opens;
• who asks;
• who answers;
• who interrupts;
• who defines relevance;
• who records;
• who closes;
• who imposes consequences.

24.8 Modality comparison

Compare:

• face-to-face;
• sign;
• telephone;
• letters;
• instant messaging;
• video calls;
• AI chat.

Score each for latency, persistence, embodied context, scalability and repair.

24.9 Evolutionary branch graphic

Reciprocal signalling ↓
Proto-turn-taking ↓
Embodied spoken conversation ├── Teaching ├── Negotiation ├── Deliberation ├── Telephone ├── Text chat ├── Video conversation └── Conversational AI

26.1 Concept A: The First Protocol

Anchor: Two human profiles facing one another.
Tension: Speech waves become packet arrows with ACK and ERROR labels.
Text: THE FIRST PROTOCOL
Visual system: Clueless Pundit archival darkness with yellow technical annotations.

26.2 Concept B: Why We Take Turns

Anchor: Two mouths or speech bubbles colliding.
Tension: A traffic-light or token symbol controls who speaks.
Text: YOUR TURN

26.3 Concept C: Human Error Correction

Anchor: A broken message between two faces.
Tension: “WHAT?” repairs the broken signal.
Text: MESSAGE FIXED

26.4 Concept D: Conversation vs AI

Anchor: Human face opposite a machine interface.
Tension: Shared-ground pool on the human side, probabilistic text stream on the machine side.
Text: DOES IT UNDERSTAND?

26.5 Thumbnail restraint

Do not attempt to depict turn-taking, common ground, repair, power, Plato, telephones and artificial intelligence in one rectangle. That path ends with a thumbnail looking like a symposium exploded inside a circuit board.

Conversation is a reciprocal interaction system through which participants exchange turns, test understanding, repair errors and coordinate action. Its importance lies not simply in allowing two people to speak, but in creating a feedback loop. Questions can expose missing information, responses can confirm or reject interpretations, and misunderstandings can be corrected before they spread further.

The practice probably emerged in deep prehistory from older systems of embodied and vocal coordination, but it has no defensible invention date. Turn-taking analogues appear in other species, and human infants participate in conversation-like exchanges before mature language develops. Modern conversation, however, combines this timing infrastructure with open-ended language, shared attention, pragmatic inference and complex social action.

Conversation remains one of humanity's most adaptable information systems. It supports teaching, negotiation, diagnosis, planning, care, governance and relationship formation. Its weaknesses are equally persistent: it is ephemeral, difficult to scale, dependent on attention and vulnerable to unequal power, manipulation and disputed memory.

Later technologies extended conversation across time and distance, but each altered the balance among immediacy, persistence, embodiment, privacy and scale.

The historical significance of conversation is not that it allowed humans to produce more messages.

It allowed them to interrogate the message while it was still alive.

A spoken statement can carry a plan. Conversation allows another person to ask whether the river has flooded, whether the number was fifteen or fifty, whether the agreement applies tomorrow and whether the speaker has understood the problem at all.

That converts communication from delivery into coordination.

In the long arc of information transmission:

Embodied signals made states visible.
Spoken language made complex meaning expressible.
Conversation made meaning negotiable and repairable.
Writing made it persistent.
Networks made it remote and scalable.
AI began generating the next turn.

The next research challenge is to determine whether machines that generate responsive language have inherited conversation's surface pattern, its deeper coordination machinery or a new system that merely resembles both.

Project sources

[P1] Information Transmission Evolution Map: Initial Topic Register v0.1. Takura Nyagumbo project document, 2026.
Use: Topic identity, relationships, priority and provisional classification.

[P2] Full Topic Research notes Vocalisation, prosody and spoken language: Vocalisation, Prosody and Spoken Language v0.1. Takura Nyagumbo project document, 2026.
Use: Predecessor boundary, spoken-language mechanism and adjacent-topic logic.

Core scholarly sources

[1] Sacks, H., Schegloff, E. A. and Jefferson, G. (1974). “A Simplest Systematics for the Organization of Turn-Taking for Conversation.” Language, 50(4), 696–735. DOI: 10.2307/412243.
https://www.cambridge.org/core/journals/language/article/simplest-systematics-for-the-organization-of-turntaking-for-conversation/3DB6C0378A96BD5E480A2C0A12FE229A
Use: Foundational turn-taking model.

[2] Stivers, T. et al. (2009). “Universals and Cultural Variation in Turn-Taking in Conversation.” Proceedings of the National Academy of Sciences, 106(26), 10587–10592. DOI: 10.1073/pnas.0903616106.
https://pmc.ncbi.nlm.nih.gov/articles/PMC2705608/
Use: Cross-linguistic timing similarities and variation.

[3] Levinson, S. C. and Torreira, F. (2015). “Timing in Turn-Taking and Its Implications for Processing Models of Language.” Frontiers in Psychology, 6, 731. DOI: 10.3389/fpsyg.2015.00731.
https://pmc.ncbi.nlm.nih.gov/articles/PMC4464110/
Use: Turn timing, prediction and response planning.

[4] Schegloff, E. A., Jefferson, G. and Sacks, H. (1977). “The Preference for Self-Correction in the Organization of Repair in Conversation.” Language, 53(2), 361–382. DOI: 10.2307/413107.
https://www.cambridge.org/core/journals/language/article/preference-for-selfcorrection-in-the-organization-of-repair-in-conversation/5549B861FDE7180B75FA5C382821875E
Use: Repair organisation and self-correction.

[5] Dingemanse, M. et al. (2015). “Universal Principles in the Repair of Communication Problems.” PLOS ONE, 10(9), e0136100. DOI: 10.1371/journal.pone.0136100.
https://pmc.ncbi.nlm.nih.gov/articles/PMC4573759/
Use: Cross-linguistic repair practices.

[6] Clark, H. H. (1996). Using Language. Cambridge University Press.
https://web.stanford.edu/~clark/1990s/Using%20language/Clark.Ch1.Using.language.96.pdf
Use: Language use as joint action.

[7] Brown-Schmidt, S. and Duff, M. C. (2016). “Memory and Common Ground Processes in Language Use.” Topics in Cognitive Science, 8(4), 722–736. DOI: 10.1111/tops.12224.
https://pmc.ncbi.nlm.nih.gov/articles/PMC5813694/
Use: Common ground, memory and partner-specific interaction.

[8] Rączaszek-Leonardi, J., Dębska, A. and Sochanowicz, A. (2014). “Pooling the Ground: Understanding and Coordination in Collective Sense Making.” Frontiers in Psychology, 5, 1233. DOI: 10.3389/fpsyg.2014.01233.
https://pmc.ncbi.nlm.nih.gov/articles/PMC4224066/
Use: Common ground as coordination.

[9] de Vos, C., Torreira, F. and Levinson, S. C. (2015). “Turn-Timing in Signed Conversations: Coordinating Stroke-to-Stroke Turn Boundaries.” Frontiers in Psychology, 6, 268. DOI: 10.3389/fpsyg.2015.00268.
https://pmc.ncbi.nlm.nih.gov/articles/PMC4371657/
Use: Turn-taking across language modalities.

[10] Gratier, M. et al. (2015). “Early Development of Turn-Taking in Vocal Interaction between Mothers and Infants.” Frontiers in Psychology, 6, 1167. DOI: 10.3389/fpsyg.2015.01167.
https://pmc.ncbi.nlm.nih.gov/articles/PMC4560030/
Use: Infant protoconversation.

[11] Pika, S. et al. (2018). “Taking Turns: Bridging the Gap between Human and Animal Communication.” Proceedings of the Royal Society B, 285, 20180598. DOI: 10.1098/rspb.2018.0598.
https://pmc.ncbi.nlm.nih.gov/articles/PMC6015850/
Use: Comparative and evolutionary turn-taking.

[12] Chow, C. P., Mitchell, J. F. and Miller, C. T. (2015). “Vocal Turn-Taking in a Non-Human Primate Is Learned during Ontogeny.” Proceedings of the Royal Society B, 282, 20150069. DOI: 10.1098/rspb.2015.0069.
https://pmc.ncbi.nlm.nih.gov/articles/PMC4424641/
Use: Learned vocal turn-taking in marmosets.

[13] Holler, J. et al. (2021). “Competition Reduces Response Times in Multiparty Conversation.” Frontiers in Psychology, 12, 693124. DOI: 10.3389/fpsyg.2021.693124.
https://pmc.ncbi.nlm.nih.gov/articles/PMC8481383/
Use: Multiparty turn transition.

[14] Ekström, M. and Thelander, Å. (2023). “Conversation Analysis and Power: Examining the Descendants and Antecedents of Social Action.” Frontiers in Sociology, 8, 1196672. DOI: 10.3389/fsoc.2023.1196672.
https://pmc.ncbi.nlm.nih.gov/articles/PMC10282763/
Use: Interactional and institutional power.

[15] Drew, P., Chatwin, J. and Collins, S. (2001). “Conversation Analysis: A Method for Research into Interactions between Patients and Health-Care Professionals.” Health Expectations, 4(1), 58–70. DOI: 10.1046/j.1369-6513.2001.00125.x.
https://pmc.ncbi.nlm.nih.gov/articles/PMC5060048/
Use: Social action and institutional medical interaction.

[16] Mondada, L. and Meguerditchian, A. (2022). “Sequence Organization and Embodied Mutual Orientations: Openings of Social Interactions between Baboons.” Philosophical Transactions of the Royal Society B, 377, 20210101. DOI: 10.1098/rstb.2021.0101.
https://pmc.ncbi.nlm.nih.gov/articles/PMC9310171/
Use: Embodied sequence organisation and comparative interaction.

[17] Nota, N., Trujillo, J. P. and Holler, J. (2023). “Specific Facial Signals Associate with Categories of Social Actions Conveyed through Questions.” PLOS ONE, 18(7), e0288104. DOI: 10.1371/journal.pone.0288104.
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0288104
Use: Facial signals and conversational action.

[18] Kopp, S. and Krämer, N. C. (2021). “Revisiting Human-Agent Communication: The Importance of Joint Co-Construction and Understanding Mental States.” Frontiers in Psychology, 12, 580955. DOI: 10.3389/fpsyg.2021.580955.
https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2021.580955/full
Use: Human-agent conversation and co-construction.

[19] Fröhlich, M. et al. (2016). “Unpeeling the Layers of Language: Bonobos and Chimpanzees Engage in Cooperative Turn-Taking Sequences.” Scientific Reports, 6, 25887. DOI: 10.1038/srep25887.
https://pmc.ncbi.nlm.nih.gov/articles/PMC4876478/
Use: Great-ape communicative exchanges.

[20] Takahashi, D. Y. et al. (2016). “Early Development of Turn-Taking with Parents Shapes Vocal Acoustics in Infant Marmoset Monkeys.” Philosophical Transactions of the Royal Society B, 371, 20150370. DOI: 10.1098/rstb.2015.0370.
https://pmc.ncbi.nlm.nih.gov/articles/PMC4843608/
Use: Comparative development of vocal turn-taking.

[21] Yoo, H., Bowman, D. A. and Oller, D. K. (2018). “The Origin of Protoconversation: An Examination of Caregiver Responses to Cry and Speech-Like Vocalizations.” Frontiers in Psychology, 9, 1510. DOI: 10.3389/fpsyg.2018.01510.
https://pmc.ncbi.nlm.nih.gov/articles/PMC6117422/
Use: Caregiver scaffolding of infant interaction.

[22] Hilbrink, E. E., Gattis, M. and Levinson, S. C. (2015). “Early Developmental Changes in the Timing of Turn-Taking: A Longitudinal Study of Mother-Infant Interaction.” Frontiers in Psychology, 6, 1492. DOI: 10.3389/fpsyg.2015.01492.
https://pmc.ncbi.nlm.nih.gov/articles/PMC4586330/
Use: Development of conversational timing.

[23] Tomasello, M. (2018). “How Children Come to Understand False Beliefs: A Shared Intentionality Account.” Proceedings of the National Academy of Sciences, 115(34), 8491–8498. DOI: 10.1073/pnas.1804761115.
https://www.pnas.org/doi/10.1073/pnas.1804761115
Use: Shared intentionality and social coordination.

[24] Garrod, S. and Pickering, M. J. (2015). “The Use of Content and Timing to Predict Turn Transitions.” Frontiers in Psychology, 6, 751. DOI: 10.3389/fpsyg.2015.00751.
https://pmc.ncbi.nlm.nih.gov/articles/PMC4463931/
Use: Prediction of turn transitions.

[25] Meyer, A. S., Alday, P. M., Decuyper, C. and Knudsen, B. (2023). “Timing in Conversation.” Journal of Cognition, 6(1), 21. DOI: 10.5334/joc.267.
https://pmc.ncbi.nlm.nih.gov/articles/PMC10077995/
Use: Contemporary review of rapid turn timing.

[26] Roberts, S. G., Torreira, F. and Levinson, S. C. (2015). “The Effects of Processing and Sequence Organization on the Timing of Turn Taking.” Frontiers in Psychology, 6, 509. DOI: 10.3389/fpsyg.2015.00509.
https://pmc.ncbi.nlm.nih.gov/articles/PMC4429583/
Use: Adjacency pairs, sequence organisation and timing.

[27] Menenti, L., Pickering, M. J. and Garrod, S. C. (2012). “Toward a Neural Basis of Interactive Alignment in Conversation.” Frontiers in Human Neuroscience, 6, 185. DOI: 10.3389/fnhum.2012.00185.
https://pmc.ncbi.nlm.nih.gov/articles/PMC3384290/
Use: Interactive alignment framework.

[28] Rasenberg, M., Özyürek, A. and Dingemanse, M. (2020). “Alignment in Multimodal Interaction: An Integrative Framework.” Cognitive Science, 44(11), e12911. DOI: 10.1111/cogs.12911.
https://pmc.ncbi.nlm.nih.gov/articles/PMC7685147/
Use: Multimodal alignment.

[29] Ostrand, R. and Chodroff, E. (2021). “Speakers Align on Some Features but Not Others within a Dialogue.” Journal of Memory and Language, 117, 104195. DOI: 10.1016/j.jml.2020.104195.
https://pmc.ncbi.nlm.nih.gov/articles/PMC8345023/
Use: Limits and selectivity of alignment.

[30] Grice, H. P. (1975). “Logic and Conversation.” In P. Cole and J. L. Morgan (eds.), Syntax and Semantics, Volume 3: Speech Acts, 41–58. Academic Press.
Use: Cooperative principle and conversational implicature.

[31] Kendrick, K. H. and Torreira, F. (2015). “The Timing and Construction of Preference: A Quantitative Study.” Discourse Processes, 52(4), 255–289. DOI: 10.1080/0163853X.2014.955997.
https://pmc.ncbi.nlm.nih.gov/articles/PMC4357221/
Use: Timing of preferred and dispreferred responses.

[32] Pekarek Doehler, S. et al. (2021). “Multimodal Assemblies for Prefacing a Dispreferred Response: A Cross-Linguistic Analysis.” Frontiers in Psychology, 12, 689275. DOI: 10.3389/fpsyg.2021.689275.
https://pmc.ncbi.nlm.nih.gov/articles/PMC8504554/
Use: Cross-linguistic gaze and response design.

[33] Schegloff, E. A. (1968). “Sequencing in Conversational Openings.” American Anthropologist, 70(6), 1075–1095. DOI: 10.1525/aa.1968.70.6.02a00030.
https://anthrosource.onlinelibrary.wiley.com/doi/pdf/10.1525/aa.1968.70.6.02a00030
Use: Telephone openings and sequence organisation.

[34] Sidnell, J. and Stivers, T., eds. (2013). The Handbook of Conversation Analysis. Wiley-Blackwell.
Use: General conversation-analysis framework and social action.

[35] Bzdok, D. and Dunbar, R. I. M. (2020). “The Neurobiology of Social Distance.” Trends in Cognitive Sciences, 24(9), 717–733. DOI: 10.1016/j.tics.2020.05.016.
https://pmc.ncbi.nlm.nih.gov/articles/PMC7266757/
Use: Broader significance of social connection. Not evidence that conversation alone determines health.

[36] Stokoe, E., Albert, S. and Pearl, C. (2026). “What Is ‘Conversational’ about Conversational Technologies, Products, and Services? Insights from Conversation Analysis.” Annual Review of Applied Linguistics, First View, 1–16. DOI: 10.1017/S0267190526100233.
https://www.cambridge.org/core/journals/annual-review-of-applied-linguistics/article/what-is-conversational-about-conversational-technologies-products-and-services-insights-from-conversation-analysis/E9B905F354920A46EBE73985AD8C9D5C
Use: Evaluating conversational technologies against human interaction.

v0.1

Created 29 July 2026.

This version:

• defines conversation as a modality-neutral interaction protocol;
• distinguishes conversation from spoken language, oral tradition, discussion and debate;
• identifies immediate feedback and repair as the central breakthrough;
• documents turn-taking, sequence organisation, common ground, alignment and pragmatic inference;
• flags a taxonomy gap around interaction and coordination;
• adds an evaluation matrix, claim register and provisional dataset record;
• identifies 36 core sources;
• proposes visual and content directions;
• remains pending independent factual review and broader geographic sourcing.