What Is a Witness? Encounter, Memory, and the Making of Past Presence (Witnesses 1/4)

Abstract

What is a witness? Ordinary usage calls any observation, memory or report witnessing, and the breadth hides a distinctive human achievement. This article defines witnessing as the consequential interrecursive reopening, through symbolic articulation, of an actual firsthand encounter with a past occurrence, in response to another person’s questioning. Encountering, remembering, narrating, expert reconstruction, recording and false claims are distinguished from it. Evidence from animal cognition, a 67,800-year-old hand stencil, Genesis and legal history locates its conditions. Witnessing emerges as a poietic achievement, sometimes compulsory, through which a past encounter becomes answerable in the present.

I. "Were You There?"

Few questions are asked more often, or with less sense of their difficulty, than "were you there?" A parent asks it of a child standing beside a broken vase. A colleague asks it about a meeting at which something was decided that no one now admits to having agreed. A police officer asks it at the roadside, a judge in court, a family at a funeral when the circumstances of a death are disputed. Under the right circumstances the question performs an extraordinary operation. It takes an occurrence that is over and cannot be changed, finds a living person who encountered some part of it, and makes that person's past encounter consequential now. The person may be required to reconstruct what they met with, to separate it from what they inferred or were told, to answer another's doubts, and to stand behind an account. They may also refuse, evade or lie.

Ordinary language calls many things witnessing: watching a sunset, seeing an accident, reading about a war, being present at a wedding, giving expert evidence in court. This breadth is the reason a stricter concept is needed. If every observation, memory and report counts as witnessing, the word names nothing that observation, memory and reporting do not already name, and the distinctive achievement hidden in the question "were you there?" disappears from view. This article therefore proposes a narrow definition and holds to it throughout.

Witnessing is the consequential interrecursive reopening, at L3 or above, of a past occurrence through the articulated firsthand encounter of someone who actually experienced it, in response to another person's questioning or challenge about what happened.

Six conditions are jointly necessary. There must be an actual past occurrence: something happened. There must be an actual firsthand encounter: the person questioned was situated within sensory or practical access to the particular occurrence at issue. There must be temporal separation: the encounter has already taken place, and its relevance is being reopened in the present. There must be interrecursive reopening: another person brings that past encounter into explicit question, rather than merely receiving an unsolicited narrative. There must be articulation at L3 or above: what was encountered becomes symbolically articulable and available for further questioning. And there must be a vital stake: something in present coordination depends on establishing, contesting or accepting what happened. None of these conditions can substitute for another.

On this definition, living beings encounter events and retain their consequences all the time, and some animals track what other individuals have previously perceived. None of this is witnessing. Witnessing is a specific event within human living coordination, a poietic achievement through which one person's actual past encounter is explicitly reopened within a present interrecursive engagement because something depends on establishing what happened. The article argues that this achievement is, provisionally, distinctively human, and that its realisation depends on all five mediations of living and on the recursive jurisdiction through which people can be called upon to account for what they experienced.

Several propositions on which the argument builds are already established in Living Value Theory (LVT): the five mediations, the three types of recursive counterpart and the five recursivity levels (Ecks 2026i); memory as reconstruction rather than storage (Ecks 2026a; 2026c); recursive opening and reopening, the distinction between checking and opening, and recursive jurisdiction (Ecks 2026f); notation as recoverable invariance (Ecks 2026e); sedimentation and the dependence of deposits on uptake (Ecks 2026g); the three modes of multisymbolization (Ecks 2026h); and the comparative treatment of interrecursivity in other animals (Ecks 2026b; 2026d). A study of how painting can make witnessing askable has also been developed (Ecks 2026j). What is new here is the strict definition of witnessing as an interrecursive event, the separation of its conditions from witnessing proper, the distinctions between elicitation, witnessing and obstruction, the analysis of compulsory witnessing as a recursive burden, and the account of what witnessing added to the evolutionary repertoire of living coordination.

II. Six Things That Are Not the Same

Discussions of witnesses usually move without pause between the event, the person who saw it, what that person remembers, what they say, and whether they are believed. Six terms need to be separated, and witnessing proper needs to be located precisely among them.

The first is the occurrence: what actually happened. A vase fell; a contract was agreed; a man was killed. The occurrence is not itself witnessing. Once it has happened it is, in LVT's terms, "recursively closed, absolutely and without exception" (Ecks 2026a). Nothing further can be done to it. Witnessing cannot change what occurred; it changes how the occurrence becomes available and consequential in present coordination.

The second is the encounter: a living being's actual, partial, situated firsthand access to some portion of the occurrence. No one encounters an occurrence whole. A person in the next room hears the crash but not the push; a person in the doorway sees the push but not the face. An encounter is necessary for witnessing and insufficient for it.

The third is retention and reconstruction. The encounter continues to affect later coordination, and when it is called upon it is rebuilt under present constraints. Memory is "a live, continuing negotiation between a present self and a past that can no longer talk back" (Ecks 2026a), and remembering is "constrained reconstruction" (Ecks 2026c). A person carries forward no sealed container of the occurrence, only an altered capacity to reconstruct their encounter with it. Retention is not inherently witnessing either.

The fourth is the witness-position, and here the definition does its work. Rather than something a person possesses because they were once present, a witness-position is constituted when that person's actual past access is reopened interrecursively, at L3 or above, by someone for whom something depends on it. Before that, there is a retained capacity that may later be reopened. During it, there is an operative witness-position. Witnessing proper is the realisation of that position through the articulation of the encounter within the consequential exchange.

The fifth is testimony: symbolic articulation concerning an occurrence, in speech, gesture, writing, image or formal deposition. Testimony is a much broader category than witnessing. It includes reporting, explaining, giving an opinion, passing on what others said, and making statements within institutional procedures. Some testimony occurs as part of witnessing. Much does not.

The sixth is reception: the further coordination in which an account is believed, rejected, challenged or acted upon. A true account may be ignored; a false one may be believed. Reception is consequential, and it often decides what follows, but it does not decide whether witnessing took place.

These six terms do not describe stages through which every occurrence passes. The vast majority of encounters never become occasions for witnessing. The first three name conditions of witnessing. The fourth names its active constitution. The fifth and sixth name what can happen within and after it.

The temporal condition needs a word of its own. Someone who shouts "look out!" as a car swerves towards a pedestrian is warning, intervening symbolically in an unfolding encounter. The fact that perception involves retention over fractions of a second does not make it witnessing; if it did, the definition would collapse back into a general account of perception and memory. The person who shouted may later become a witness, when they are asked what they saw. An ongoing event can itself become the object of explicit interrecursive questioning, but the paradigmatic case, and the focus here, is the reopening of something that has already happened.

The condition of interrecursive demand also needs care. Someone describing their journey to work, recounting yesterday's weather, or telling friends about a holiday is reporting or narrating. Even when the account is accurate and firsthand, it is not witnessing, because no one has reopened the encounter for a purpose that depends on it. "I saw a fox yesterday" is a report. "You were out in the garden when the hens were killed; what did you actually see?" can begin witnessing. "Tell me about your holiday" invites narration. "You were in the room when that was decided; what was actually agreed?" opens a witnessing situation. Narration is everywhere in human life, and its ubiquity is the reason witnessing should not be identified with it. The paradigm of the reopening is another person's question, but grammar is not the point. What matters is that another person makes the past encounter explicitly questionable and consequential. Someone who steps forward unasked in a dispute that is already under way joins an existing interrecursive reopening; this does not license treating every unsolicited report as witnessing. Nor does every question about a past encounter reopen it. "Did you see my email?" is ordinarily a check, and checking, in LVT's terms, "asks whether the settled thing remains as expected", protecting closure rather than suspending it (Ecks 2026f).

The vital stake should not be read as an intensifier. An occurrence becomes relevant to witnessing when the answer matters for ongoing coordination: for responsibility or liability, for trust, for recognition, for safety, for a decision, for a disputed recollection between people who must go on living together. The stake need not be spectacular. It may concern a small domestic disagreement. But it must be more than an abstract wish for one more piece of information. And the stake need not be the same for everyone. The person questioned may regard the encounter as trivial while the person asking regards it as decisive. The importance of an occurrence for witnessing is not intrinsic to it; it emerges in the present coordination that reopens it.

Firsthand access is non-negotiable. An engineer reconstructing a collapse did not encounter the collapse. Someone repeating a neighbour's account did not encounter what the neighbour encountered. Someone inventing a story about having been present encountered nothing. These are different activities, some of which imitate or institutionally borrow the language of witnessing. Correct interpretation, by contrast, is not required. A person who was present may misremember, misidentify or draw the wrong conclusion. What matters is the actuality and scope of the encounter, which determines whether witnessing is possible and what it can concern, as distinct from the reliability of the subsequent reconstruction, which may always be contested. Someone who heard a collision from another room can witness the sound. They cannot, on that basis, witness which driver caused it. Two consequences follow. A witness can give an inaccurate account, and someone who is not a witness can give an accurate one. Neither accuracy nor credibility establishes witnessing.

Finally, witnessing must be distinguished from the social attribution of witness status. Courts designate expert witnesses. Institutions describe accounts as eyewitness testimony. Communities believe, sometimes wrongly, that a particular person was present. These attributions have consequences, but they do not change what happened, and an L4 classification does not determine whether witnessing took place. Four things must be kept apart throughout: the actual encounter; witnessing proper through interrecursive reopening; claims and attributions of witnessing, correct or incorrect; and institutional witness status.

A single ordinary case brings these distinctions together. At dusk on a village road, a motorcycle strikes a cyclist. The rider encounters the collision through handlebars and brakes; the cyclist through the blow and the fall; a shopkeeper inside his stall through the screech of tyres and a shout; a dog asleep by the road through the noise and the sudden rush of people. Each encounter is partial, and only two include the moment of impact. Over the following days the shopkeeper tells the story many times, and his telling acquires details, the motorcycle's speed, the cyclist's swerve, that he could not have perceived from behind his counter. These conversations are narration, rumour and speculation, not witnessing. Then responsibility is disputed. The rider's family says the cyclist swerved; the cyclist's family says the rider was drunk. A police officer comes to the stall and asks the shopkeeper what he heard, what he saw, where he was standing, and when he came out. At that moment, and only then, his actual sensory access becomes the subject of a consequential interrecursive reopening at L3. He can witness the screech, the shout, and whatever he saw when he reached the road. He cannot witness the moment of impact. His claims about speed and swerving may be inferred, reported, invented or reconstructed, and their reception as eyewitness testimony does not make them firsthand. The dog matters to the case as a comparison rather than as an unrecognised witness: it encountered the collision and may recognise the rider again, without taking part in the human activity of witnessing at all.

The following cases show how the definition sorts adjacent situations.

The last case is the most instructive. Being summoned to witness is not the same as witnessing. A person can obstruct or refuse an attempted reopening, and the refusal can itself become consequential, but it does not turn into an articulated account of what occurred.

III. Two Multimediated Coordinations

It is tempting to treat witnessing as perception followed by storage followed by report, a pipeline in which information is first acquired and only later dressed in social meaning. LVT rejects this model on two counts. The encounter is multimediated from the beginning, so there is no stage of pure information. And witnessing involves at least two distinct multimediated coordinations: the original encounter, and the later interrecursive reopening in which that encounter is questioned, reconstructed, articulated and possibly contested. Witnessing emerges in the second, not in the first.

Imagine a collision witnessed years after it happened. The original encounter took place on a dark street in heavy rain. The reopening takes place in a brightly lit courtroom, with photographs, plans, diagrams, documents and competing accounts. The later activity, far from being a return to the earlier encounter, is a new living coordination in which traces of the earlier encounter become consequential. Each coordination has its own bodies, participants, conditions, material arrangements and symbolic forms, and the five mediations can be traced through both.

Multisensorial embodiment shaped what the encounter could include: what could be seen or heard from that position, at that distance, with that degree of attention or fatigue. Seeing is only one possibility; the most famous Christian formula of testimony speaks of "that which we have heard, which we have seen with our eyes, which we have looked upon, and our hands have handled" (1 John 1:1). In the reopening, embodiment returns differently: in the retained bodily familiarity that lets a person recognise a voice or a gait again, in hesitation, in the tremor or steadiness of a voice under questioning, in the face that the questioner watches.

Multispecies being-with gave the encounter its participants. Who else was there? Did the person recognise them? Were they welcomed, concealed, excluded or summoned? Recognising someone is not reducible to perceiving a human figure; it draws on a history of being-with that the moment of perception does not contain. In the reopening, being-with is the questioner and everyone the answer will affect: the people who ask, resist, threaten, believe, doubt or demand clarification, and the relationships that will be changed by what is said.

Multiversal dwelling gave the encounter its conditions in what no one made: daylight or darkness, the hour, the season, rain, the distance at which a voice carries over open ground. The reopening has its own conditions and its own place in time, often far from the encounter. The difference between the two temporal locations is itself one of the most frequent objects of questioning: could the light have shown what is claimed, and how long ago was it?

Multimateriality gave the encounter its arrangements: rooms, doors, windows, roads, vehicles, clothing, tools. Built places belong here. The reopening has material conditions of its own: a stall, a police notebook, a courtroom, a plan of the junction, a photograph. Traces of the original arrangement can constrain the reconstruction. But witnessing does not require that an independent artifact survive. A person can be questioned about a firsthand encounter when no photograph, document or recoverable object remains. The materiality of the original encounter and the material conditions of its reopening are indispensable; an externally preserved piece of evidence is not.

Multisymbolization gave the encounter its categories, the terms through which it was understood at the time, and gives the reopening its questions, distinctions, accounts and contradictions. What a person perceived, what they understood, and what they say are not identical, and the reopening is precisely the occasion on which these differences can be drawn.

Taken together, these considerations yield the article's central proposition about the process. Witnessing is a new multimediated interrecursive coordination in which traces of an actual earlier encounter are symbolically articulated and made consequential under present conditions. "Traces" here must be understood broadly, to include retained bodily and practical consequences, not only objects kept outside the person. The five mediations are not a credibility checklist, and they are not commensurable evidentiary categories that could be added up. They describe what both coordinations are made of, and therefore what the reopening can call into question.

Languages show some of the distinctions that witnessing mobilises, though they cannot settle what witnessing is. Etymology licenses no conclusion about what a word must now mean. Still, the words for the witness are suggestive. The Sanskrit sākṣin is associated with having a matter "before the eyes" (sākṣāt); the Latin testis with the third party who stands by; the Latin superstes with one who has lived through and survived (Benveniste 1973; Agamben 1999); English witness with wit, knowledge. Some languages also have grammatical means of marking the source of information, distinguishing what a speaker experienced directly from what was reported or inferred (Aikhenvald 2004). The Turkish past-tense contrast between -dI and -mIş is often described in these terms, though the second form also covers inference, hearsay and newly realised information, and the system is better described as marking the speaker's relation to knowledge than as a simple seen/unseen opposition. Many Quechua varieties have enclitics conventionally glossed as direct, reportative and conjectural, whose functions also extend to speaker commitment and vary by variety. Such marking is highly relevant to the preconditions of witnessing, since it stabilises the difference between firsthand access and its absence. It does not itself constitute witnessing, since a speaker can mark firsthand knowledge without anyone reopening the encounter. Languages can stabilise distinctions that witnessing subsequently mobilises, without every use of those distinctions constituting an act of witnessing.

IV. What Animals Know about Who Saw What

The evolutionary question needs to be posed precisely. Animals demonstrably observe, communicate, deceive, remember and track the perceptual access of others, so the question is not whether they do these things but whether there is evidence of consequential interrecursive reopening, at L3 or above, of a past occurrence through the articulated firsthand access of an individual who encountered it. That is a substantially more demanding configuration, and the animal evidence is most illuminating when each case is described positively, in terms of what the animals actually do.

The honeybee's waggle dance conveys the direction and distance of a food source the dancer has visited (von Frisch 1967). Hockett (1960) counted it as displacement, communication about something not present. The dance coordinates the colony with an absent place through an established form of living coordination. Recruits follow it or do not, find the food or fail to. Nothing in this coordination reopens the dancer's past encounter as such.

Vervet monkeys give acoustically distinct alarm calls for leopards, eagles and snakes, and listeners respond appropriately (Seyfarth, Cheney and Marler 1980). The fork-tailed drongo gives deceptive alarm calls to frighten meerkats from their food, and when meerkats habituate to one call, drongos switch to mimicking another (Flower, Gribble and Ridley 2014). This is deceptive alarm signalling, and the meerkats' habituation tracks the reliability of a signal. Neither party is engaged in questioning a claimed past encounter.

Western scrub-jays that were watched by another jay while caching food return alone and move their caches, but only if they have themselves pilfered other birds' caches in the past (Emery and Clayton 2001). They keep track of which individual watched which caches being made, of "who was watching when", in the words of one study's title (Dally, Emery and Clayton 2006). Subordinate chimpanzees preferentially approach food that a dominant competitor did not see being hidden (Hare, Call and Tomasello 2001). Within LVT these behaviours appear as plausible cases of interrecursivity, in which coordination tracks "an inferred internal state of the other, however lightly" (Ecks 2026b). The jay changes its own behaviour in response to another bird's earlier opportunity to observe. It does not treat the other bird as a witness, and there is no continuum running from such tracking to human witnessing through ever stronger witness-positions. Tracking another's past perceptual access can occur at L1 and L2, where coordination "becomes recursively salient, or questionably available". Witnessing is an event at L3 or above, where "salience becomes symbolically articulable" (Ecks 2026i).

The animal evidence therefore establishes the evolutionary capacities on which the human achievement could build: individual recognition, memory of particular episodes, sensitivity to what others could perceive, deception and its detection. The history is one of continuity in these underlying mediations. The human configuration introduces a qualitatively new possibility, which depends on the coupling of multimateriality and multisymbolization that LVT identifies as "the distinctively human threshold" (Ecks 2026d).

What does witnessing make possible that observation, memory, communication and the tracking of others' perception do not? At least five things. First, another being's past encounter becomes explicitly investigable. Its significance is no longer confined to what the individual does as a consequence of having had it. Second, the grounds of the encounter become contestable. One person can ask another whether they had access at all, what aspects they encountered, and whether their reconstruction is sound. Third, the past becomes reopenable through someone else's retained access. This is neither one's own memory nor one's own observation, but a new coordination between people in which one person's past becomes available to another. Fourth, witnessing creates a new form of accountability. A person can be required to articulate what they encountered, including against their wishes. Fifth, the practice can be stabilised symbolically and materially, in questions, formulas, records, procedures and institutional positions that allow later engagements to reopen earlier witnessing in turn. The evolutionary novelty lies less in the emergence of an abstract representation of the past than in the capacity to make someone's actual past presence an explicit, consequential matter for another person.

This is offered as an evolutionary hypothesis. Research has not established the absence of every possible non-human analogue, and the study of animal communication keeps finding more than was expected. The claim is that no established non-human case demonstrates the full configuration this definition requires.

Human development raises a corresponding question rather than answering it. Children come to understand that looking leads to knowing, that the person who looked into a box knows what is in it while the person who only touched it may not, at around three to four years of age (Pratt and Bryant 1990). By four, they prefer informants who have been accurate and are wary of those who have proved ignorant or mistaken (Koenig and Harris 2005). These are crucial prerequisites: the distinction between situated access and its absence, and the evaluation of informants. They are not witnessing proper. The question that standard looking-knowing, false-belief and selective-trust experiments do not address is more specific: at what point can a child take part, as questioner or questioned, in a consequential interrecursive reopening of another person's actual past encounter? That is an open empirical question, and the definition proposed here makes it askable. Tomasello's account of shared intentionality (2014) offers a congenial frame for pursuing it, though it concerns joint attention and collaboration rather than the reopening of past encounters.

V. Presence without Witnessing: The Prehistoric Hand Stencil

In January 2026, Adhi Agus Oktaviana and colleagues reported a hand stencil on the wall of Liang Metanduno, a limestone cave on Muna Island off south-east Sulawesi, with a minimum age of about 67,800 years (Oktaviana et al. 2026). The date comes from uranium-series analysis of mineral crusts that formed over the pigment. The stencil was made by placing a hand against the rock and spraying pigment around it, and its fingers appear to have been narrowed, a feature so far known only from Sulawesi. It is the oldest dated hand stencil reported so far.

The stencil is a powerful negative case for the theory of witnessing. An embodied, situated gesture produced a durable trace, and the person departed. In LVT's terms the stencil is a deposit, "a form that carries the settled product of what was once an open process and that can no longer participate in that openness" (Ecks 2026a). Its form makes this unusually vivid. A stencil is a negative image: the pigment marks the rock around the hand, and the hand appears as the unpainted shape left when it is withdrawn. What persists is the outline of a body that is no longer there.

The stencil shows how embodied presence can leave a trace that remains available long after the event. That is a condition enabling later reconstruction of presence. It is not witnessing. A later visitor can infer that someone placed a hand against the rock. In the terms of LVT's analysis of notation, an invariance is recoverable "when a competent later party who witnessed nothing can re-identify the referent" (Ecks 2026e), and here the referent recoverable is a hand, not a person and not an encounter. No later visitor can interrogate the maker about what that person met with in the cave or outside it, whether they meant anything by the mark, or what was at stake for them. A durable trace of presence can persist for tens of thousands of years without anyone's past encounter ever being reopened through it.

The same holds, perhaps surprisingly, for explicit statements of presence. When a fifteenth-century painter wrote on a panel that he "was here" (Ecks 2026j), he asserted presence. An assertion of presence is not an interrecursive reopening in which the writer's firsthand access is consequentially questioned. It can become the occasion for such questioning, if someone with something at stake can ask the writer what he encountered. Once the writer is dead, that possibility is closed, and the inscription joins the stencil as a trace of presence without witnessing.

The cave adds one further caution. A second, darker stencil near the first has yielded a minimum age of only about 32,800 years. Minimum ages from overlying crusts give no upper bound, so the two dates cannot establish the interval between the stencils, or even their order. What can be said is only that the site was used for image-making over a very long period, and that later bodies came to stand where earlier bodies had left traces. Material evidence of presence is abundant and ancient. Witnessing is something else.

VI. Genesis: Reopening, Refusal, and the Ground That Speaks

The opening chapters of Genesis are often read as stories of temptation and transgression. Read with the strict definition in hand, they also discriminate with remarkable precision between the reopening of a past utterance and the attempted reopening of a person's firsthand encounter.

The prohibition is given in the second chapter: "The LORD God commanded the man, saying, Of every tree of the garden thou mayest freely eat: But of the tree of the knowledge of good and evil, thou shalt not eat of it" (Genesis 2:16–17). The woman is made afterwards (2:21–22). When the serpent asks her, "Yea, hath God said, Ye shall not eat of every tree of the garden?" (3:1), the text does not say whether she later heard the command herself. If her knowledge derives from the man, she is transmitting another person's account, which is not witnessing. If God repeated the command to her, she would have firsthand access, but the narrative does not say so. There is no textual basis for treating her as a witness to the original command, and the uncertainty should be kept open rather than silently resolved.

The serpent's question reopens a past utterance's content and authority. The woman answers by articulating her understanding of it, and in doing so she adds a clause: "God hath said, Ye shall not eat of it, neither shall ye touch it, lest ye die" (3:3). The serpent then contradicts the threatened consequence (3:4–5). The order matters. The addition comes in her answer, after the serpent's question has already opened the matter, and does not cause that opening; it shows what the opening brings to light, a version of the prohibition already altered somewhere between the command and her reply. The serpent's question has been analysed in LVT as a recursive opening that makes a sedimented relation to the prohibition available for examination (Ecks 2026f). The strict definition adds a further distinction. Not every interrecursive reopening is witnessing. Genesis 3 is an interrecursive reopening of a past utterance; witnessing would additionally require that the reopening turn on the actual firsthand access of the one questioned. The passage should not be forced into that second category. What it shows is that the human capacity to question the past had become narratively ordinary, so ordinary that the text never pauses to explain how an earlier utterance can be recovered, reported by someone else, and reopened.

The fourth chapter supplies the stronger case. Cain kills Abel in the field. God then asks him: "Where is Abel thy brother?" Cain answers: "I know not: Am I my brother's keeper?" (Genesis 4:9). Every condition for witnessing is present except its realisation. There is an actual past occurrence. There is a participant with firsthand access, the fullest access anyone could have, since he committed the act. There is consequential interrecursive questioning by the one with the greatest stake in the answer. What follows is an explicit refusal to acknowledge the encounter. Cain's denial, far from being witnessing, is an attempt to prevent a witness-position from becoming operative. The murderer becomes the person whose access matters most, precisely because that access is being concealed.

God's reply turns elsewhere: "the voice of thy brother's blood crieth unto me from the ground" (4:10). The ground is no witness: it carries traces, and it does not enter an interrecursive exchange as someone who experienced the killing and can account for it. The image is a striking symbolic attribution of speech to nonrecursive material consequences, and it is revealing because it imaginatively crosses a distinction the ontology must preserve. Genesis makes the ground speak precisely where the person who was actually there refuses to do so.

Read together, the two chapters distinguish three things that the strict definition requires: reopening a past utterance (the serpent and the woman), attempting to reopen a firsthand encounter (God and Cain), and the turn to material traces when that attempt is obstructed (the blood and the ground). The text presupposes all three as intelligible to its audience.

VII. Witnessing as Interrecursive Reopening

Witnessing is a particular kind of interrecursive opening of an earlier encounter, rather than merely an account that may later be reopened. LVT defines an opening as the moment when "a previously inactive selfrecursive or interrecursive relation becomes operative", and a reopening as the reactivation of a relation that has already been closed or sedimented (Ecks 2026f). Both apply. The occurrence was closed when it happened, and witnessing reopens it for present interrecursive consideration. The person's relation to their own encounter may never have been examined before, and what is opened is its previously unexamined relevance. A person may have lived for years with the memory of what they saw on the road without ever being asked; the question opens that relation and reopens the occurrence at once.

The recursivity levels clarify where witnessing sits and what sustains it. At L1, "living coordination proceeds in operative closure" (Ecks 2026i): earlier encounters are tacitly remembered, recognised and used without becoming witnessing. A person knows which neighbour borrowed the ladder because they saw it happen and acts accordingly. Most encounters and most remembering remain here. At L2, coordination "becomes recursively salient, or questionably available": a discrepancy makes an earlier encounter felt as potentially questionable. The ladder has not come back, the neighbour says he never took it, and the memory of seeing him carry it acquires an edge. This is not yet witnessing. At L3, "salience becomes symbolically articulable": a consequential question makes someone's actual past access articulate, and witnessing becomes possible and is realised in the response. At L4, "articulations become portable stabilisations": witnessing is stabilised in roles, procedures, documents and rules of evidence. At L5, "stabilisations become recursively repositionable": the criteria, legitimacy or limits of those stabilisations can themselves be reconsidered.

The level of the witnessing operation must be distinguished from the many L1 and L2 processes that sustain it. A person answering a question about what they saw is not reflectively articulating everything; their recognition of faces, their bodily orientation in remembered space and much of their reconstruction proceed at L1. What is required is that the encounter's firsthand status become explicitly available in the exchange. Equally, witnessing does not require L4. A parent questioning a child about a broken vase, two friends contesting what was said in an earlier conversation, and colleagues trying to reconstruct a consequential decision can all involve witnessing proper without any court, record or institutional category.

The three types of recursive counterpart show why witnessing is an interrecursive event even though it depends on other kinds of coordination. Nonrecursive counterparts "do not alter their course because they have been engaged" (Ecks 2026i): a road, a broken vase, a footprint, a recording. They constrain what can be claimed and are often used to test accounts, but a material trace does not witness. Selfrecursive processes "respond to their own changing state": recollection, hesitation, reinterpretation, deliberation about whether to speak. A selfrecursive recollection, however elaborate, is not witnessing. Interrecursive relations arise "when two or more recursive beings respond to each other's responses": being asked, believed, pressed, doubted, contradicted. Only when a person's actual past access becomes the subject of a consequential interrecursive reopening at L3 or above does the defining process occur.

Three processes surround witnessing proper and must not be confused with it. Elicitation is an attempt to reopen someone's actual firsthand access. Witnessing proper is the articulation of that access within the consequential interrecursive engagement. Obstruction is refusal, intimidation, concealment or denial that prevents the articulation. These are less three varieties of witnessing than the conditions, the realisation and the blocking of a single process. The distinction explains why an occurrence can remain inaccessible even when the person who encountered it is known, as Cain's case already shows.

Not everyone may demand an account from everyone else in every circumstance. Who is permitted to ask? Who must answer? Who may refuse? Who decides whether an answer has settled the matter? LVT calls the governance of these questions recursive jurisdiction, the regulation of who may open, refuse or reclose a matter (Ecks 2026f). The questions arise in families, workplaces, friendships, religious communities and research, not only in law. A child may be asked by one parent what the other said; a junior colleague may be asked by a manager what a senior colleague did. Witnessing depends on a human capacity to make another person's encounter consequential through explicit questioning, and that capacity is socially contested and institutionally regulated.

This brings into view one of the most significant consequences of the strict definition: witnessing can be compulsory. Someone may have been present at an event without ever wanting that experience to become a matter of explicit engagement. Others can nevertheless insist that they account for it. A person pressed to recount an assault, a child asked to choose between contradictory parental accounts, a colleague questioned about conduct they saw at work: in each case presence becomes an unwanted responsibility. In LVT's terms, this is a distinctive form of recursive burden, in which a past encounter is forced back into present coordination by the questions of others. Witnessing should therefore not be romanticised as an inherently emancipatory act of speaking truth. It can be threatening, exhausting, humiliating or dangerous. It can also bring relief, vindication, accountability and recognition. The reopening itself becomes a vital stake, sometimes independently of the original occurrence.

Legal history illustrates the institutional regulation of witnessing; it does not define witnessing. Legal witness status and witnessing in the sense defined here are not coextensive. Legal institutions admit expert opinion, hearsay under specified conditions, technical evidence and records, and they have excluded people who actually encountered the relevant occurrence. The laws of Hammurabi treated the purchase of goods "without witnesses or contract" as grounds for treating the buyer as a thief (Roth 1997: §7) and punished a false witness in a capital case with death (§3). Deuteronomy required that "at the mouth of two witnesses, or at the mouth of three witnesses, shall the matter be established" (Deuteronomy 19:15). The eighth chapter of the Mānava Dharmaśāstra sets out at length who may and may not serve as a witness and how witnesses are to be examined (Olivelle 2005). English law long barred the accused, the person most certainly present at an alleged crime, from giving sworn evidence in their own defence; parties to civil suits gained the right to testify in the mid-nineteenth century, and defendants in criminal trials only with the Criminal Evidence Act of 1898. These reforms changed L4 categories of competence. Whether a given reform also involved L5, a reconsideration of the grounds on which such categories rest, is a historical question in each case and should not be assumed. And a person excluded from giving evidence by a legal order could still take part in witnessing outside it, wherever someone with a stake asked them what they had encountered.

VIII. Testimony Is Not Witnessing

Testimony and witnessing are different concepts. Testimony is broad: it includes reporting, explaining, giving an opinion, conveying secondhand knowledge, and making statements within institutional procedures. Witnessing is narrow: the consequential reopening, at L3 or above, of something actually encountered firsthand. Some testimony occurs as part of witnessing. Much does not.

Within witnessing, the account given is never a copy of the earlier encounter but a new multimediated coordination in which a past encounter becomes consequential again, with its own bodies, participants, conditions, materials and symbolic forms. This is what gives witnessing its distinctive risks and possibilities: the encounter can be reconstructed under the pressure of the questioner's doubts, shaped by the categories available in the reopening, and transformed by the stakes of those who will act on it. In terms of the three modes of multisymbolization, an account of a past encounter ordinarily works in reference-to-absence mode, making "something not presently available matter to current coordination", and it shades into re-presentation when it claims to make the encounter present again for those who were not there (Ecks 2026h). Neither mode, by itself, establishes what the encounter was.

The legal category of the expert witness should be named plainly. The forensic pathologist who did not see the killing is not a witness to the killing. The engineer who reconstructs a structural failure from photographs and measurements is not a witness to the collapse. The psychologist assessing someone's behaviour is not thereby a witness to the events in dispute. These people may provide indispensable expert analysis, but legal usage has extended the language of witnessing to activities that are not witnessing in the strict sense, and that designation should not determine LVT's ontology. An expert can, of course, witness a particular examination or experiment they personally carried out, if that encounter is later consequentially reopened by questioning. The status must always be assessed in relation to a specific occurrence: it is either witnessing of something the expert encountered or expert interpretation of matters they did not.

Fabricated claims must be treated with the same clarity. A liar can claim to have seen something. Others can accept the claim, and the false account can have enormous consequences. None of this establishes witnessing. What has happened is the production and reception of a false symbolic attribution of firsthand access. Holding this line prevents the theory from defining reality by what people happen to believe.

The scholarship on testimony deserves a precise acknowledgement of where this account differs from it. Hume weighed testimony against the uniformity of experience in his discussion of miracles (Hume 2007 [1748]: §10). Coady (1992) argued at length that testimony is a fundamental source of knowledge, irreducible to individual perception and inference, and examined the social practices through which it is given and received. These are contributions to the epistemology of testimony, concerned with when and why we are warranted in believing what others tell us. LVT is developing something different: an ontology of a specific event, the consequential interrecursive reopening of a past encounter, which testimony may or may not form part of. The scholarship on witnessing catastrophe has also long distinguished between experiencing events and bearing witness to them. Felman and Laub (1992) explored how testimony to catastrophe struggles to make experience communicable; Peters (2001) distinguished the sensory experience of seeing from the discursive act of attesting; Agamben (1999) drew on the Latin testis and superstes to argue that the complete witness of the camps is the one who did not survive. Rather than claiming to be the first to recognise the gap between encounter and testimony, the present article contributes a cross-domain definition grounded in actual firsthand access, vital stakes and interrecursive reopening at L3 or above.

Anthropology has made "being there" central to debates about authority. Geertz argued that ethnographic authority rests on the writer's capacity to convince readers that he or she has actually penetrated another form of life (Geertz 1988: 1–24). Under the strict definition, the ethnographic monograph is ordinarily narration and scholarly representation based partly on firsthand encounters. It is not, as such, witnessing. An ethnographer can become a witness, however, when a specific past encounter is reopened through someone else's consequential questioning: when a researcher who observed a ritual, a dispute or an act of violence is later asked, by people with something at stake, to confirm or challenge an established account of it. The discipline has long debated the authority of having been there without sufficiently distinguishing having been there from being called upon to witness.

The history of experimental science calls for the same distinction. Shapin and Schaffer (1985) describe how Robert Boyle's air-pump experiments were performed before assembled observers of standing, and how their reports were written in circumstantial detail, with engravings of the apparatus, so that readers could become what Shapin and Schaffer call "virtual witnesses". Those present at a demonstration could later become witnesses when their firsthand access was consequentially questioned. Readers of the printed report did not thereby become witnesses to the experiment. "Virtual witnessing" is a historically specific extension of the language of witnessing, and an illuminating one, but a written account can only make an earlier event intelligible and credible to people who were absent. It cannot give them firsthand access to the original occurrence retrospectively. This takes nothing away from Shapin and Schaffer's historical analysis. It marks the difference between their historical category and the analytical concept developed here.

Recordings and synthetic media follow the same rule. A camera records; it registers light and produces a deposit. It is not a witness. A person who watches a video of an assault has encountered the video, not the assault. They may later witness what they saw on the recording, if they are questioned about their own viewing, but their position is not that of someone present at the original event. A recording enters coordination as a nonrecursive deposit requiring investigation and interpretation, whether through live testimony from those who made or kept it, through technical provenance, or through other documentary evidence. Synthetic media sharpen the point without changing it. A fabricated recording can generate false claims about an event and create the appearance that someone or something had firsthand access. It cannot produce firsthand witnessing of an occurrence that never happened. Material and symbolic deposits never occupy witness-positions themselves; they become consequential only when living people take them up.

IX. Conclusion: Making the Past Answerable

The past can become attributable in many ways that are not witnessing. People construct historical explanations, repeat family stories, read documents, infer events from traces, and identify who once occupied a place. Living beings retain the consequences of earlier encounters in their bodies, habits and surroundings. Some animals track what others have previously perceived. All of this is evolutionarily deep, and all of it supplies conditions on which witnessing builds.

Witnessing is narrower and stranger. Someone who actually encountered an earlier occurrence can be called upon, within a consequential interrecursive exchange, to articulate that encounter and have its grounds questioned. The occurrence cannot participate in that exchange. The past cannot answer back, but someone who encountered it can, and other people can insist that they do. This is a poietic achievement, enabled by multimaterial and multisymbolic coordination and embedded in embodiment, being-with and dwelling. Its human configuration allows another person's actual earlier encounter to become the subject of articulated, consequential, interrecursive questioning.

The question with which the article began can now be heard more exactly. "Were you there?" is more than a request for information. Under the right circumstances it makes another person's actual past encounter consequential now. It creates the possibility that the person must reconstruct what they encountered, distinguish it from what they inferred or were told, respond to another's doubts, and accept responsibility for an account. The same question can be refused, evaded, answered falsely or turned into a dispute of its own. Witnessing is present in no ordinary memory or report as such; it emerges when past presence is explicitly and consequentially reopened between living beings.

A witness is someone whose actual past encounter becomes consequential through its explicit interrecursive reopening by others. Witnessing is the distinctively human achievement of making that encounter answerable in the present.

Selected References

Agamben, Giorgio. 1999. Remnants of Auschwitz: The Witness and the Archive. Translated by Daniel Heller-Roazen. New York: Zone Books.

Aikhenvald, Alexandra Y. 2004. Evidentiality. Oxford: Oxford University Press.

Benveniste, Émile. 1973. Indo-European Language and Society. Translated by Elizabeth Palmer. London: Faber and Faber.

Coady, C. A. J. 1992. Testimony: A Philosophical Study. Oxford: Clarendon Press.

Dally, Joanna M., Nathan J. Emery, and Nicola S. Clayton. 2006. "Food-Caching Western Scrub-Jays Keep Track of Who Was Watching When." Science 312 (5780): 1662–1665.

Emery, Nathan J., and Nicola S. Clayton. 2001. "Effects of Experience and Social Context on Prospective Caching Strategies by Scrub Jays." Nature 414: 443–446.

Felman, Shoshana, and Dori Laub. 1992. Testimony: Crises of Witnessing in Literature, Psychoanalysis, and History. New York: Routledge.

Flower, Tom P., Matthew Gribble, and Amanda R. Ridley. 2014. "Deception by Flexible Alarm Mimicry in an African Bird." Science 344 (6183): 513–516.

Geertz, Clifford. 1988. Works and Lives: The Anthropologist as Author. Stanford: Stanford University Press.

Hare, Brian, Josep Call, and Michael Tomasello. 2001. "Do Chimpanzees Know What Conspecifics Know?" Animal Behaviour 61: 139–151.

Hockett, Charles F. 1960. "The Origin of Speech." Scientific American 203 (3): 88–96.

Hume, David. 2007 [1748]. An Enquiry concerning Human Understanding. Edited by Peter Millican. Oxford: Oxford University Press.

Koenig, Melissa A., and Paul L. Harris. 2005. "Preschoolers Mistrust Ignorant and Inaccurate Speakers." Child Development 76 (6): 1261–1277.

Oktaviana, Adhi Agus, Renaud Joannes-Boyau, Budianto Hakim, et al. 2026. "Rock Art from at Least 67,800 Years Ago in Sulawesi." Nature. doi:10.1038/s41586-025-09968-y.

Olivelle, Patrick. 2005. Manu's Code of Law: A Critical Edition and Translation of the Mānava-Dharmaśāstra. New York: Oxford University Press.

Peters, John Durham. 2001. "Witnessing." Media, Culture & Society 23 (6): 707–723.

Pratt, Chris, and Peter Bryant. 1990. "Young Children Understand That Looking Leads to Knowing (So Long as They Are Looking into a Single Barrel)." Child Development 61 (4): 973–982.

Roth, Martha T. 1997. Law Collections from Mesopotamia and Asia Minor. 2nd ed. Atlanta: Scholars Press.

Seyfarth, Robert M., Dorothy L. Cheney, and Peter Marler. 1980. "Monkey Responses to Three Different Alarm Calls: Evidence of Predator Classification and Semantic Communication." Science 210 (4471): 801–803.

Shapin, Steven, and Simon Schaffer. 1985. Leviathan and the Air-Pump: Hobbes, Boyle, and the Experimental Life. Princeton: Princeton University Press.

Tomasello, Michael. 2014. A Natural History of Human Thinking. Cambridge, MA: Harvard University Press.

von Frisch, Karl. 1967. The Dance Language and Orientation of Bees. Cambridge, MA: Harvard University Press.