Danger

Type: Node / Prospective Harm Relation

Working Definition: Danger is a prospective, relational condition: a source or situation stands in a Danger relation to a subject at risk when it could bring about some possible harm to that subject, through some pathway, under some enabling conditions, within some time horizon. Nothing is dangerous simpliciter — the relation always holds to some subject, with respect to some possible harm, under some conditions, through some pathway, within some horizon, never as a free-floating property of the source alone. The relation can hold before any Agent notices it, appraises it, assigns it Significance, or allocates Attention to it.

Function in the Grammar: Danger names the specific, prospective-harm-directed relation that other concepts in this division presuppose but do not themselves supply. Its relationship to the surrounding grammar is a set of dependencies with feedback, not a guaranteed temporal order. A Danger relation may hold without being attended to or appraised at all. When an Agent encounters relevant Evidence or a candidate source, Attention to it makes explicit appraisal possible — appraisal is a species of Interpretation, and Interpretation, like every other act that presupposes examining something, has Attention as an unstated precondition. The resulting appraisal may then reallocate further Attention, toward vigilance and investigation or away from it through avoidance, denial, and normalization. An appraisal may be externalized as a Claim. Once that Claim is itself interpreted and carries Meaning, Significance may be assigned to it. That Significance assessment may then be externalized as a further Claim, entering a subsequent interpretive cycle — the same reciprocal coupling that holds between Attention and appraisal, one act removed from it in time rather than simultaneous with it. Review may inform a later Decision or Recognition; it does not by itself confer status. None of these steps is guaranteed by the one before it, and a later step can act back on an earlier one — an appraisal reshapes the very Attention that made it possible in the first place.

This entry sits in Division IV because its entire reason for existing, per the admission evidence behind it, is compositional work at the appraisal layer, not because the relation itself is an act of meaning-making. The relation can and typically does hold as an appraisal-independent and recognition-independent condition — a fraying cable is dangerous to anyone who might touch it whether or not anyone has appraised it yet — and that independence is deliberate, not a defect to be reasoned away. What earns Danger a place among Claim, Context, and Significance rather than among Event, Process, and Act is that its grammatical work is entirely in service of how that relation gets appraised, claimed, recognized, and acted on: a source's causal potential to harm someone is, on its own, already fully described by ordinary physical or situational vocabulary, and needs no encyclopedia entry. What needs one is the recurring conflation this project has found in its own corpus between how much something matters and which direction it matters in — see Distinctions, below — and the recurring gap between a source's Danger relation holding and anyone's Attention, appraisal, or institutional response actually reaching it. Both of those are appraisal-layer problems. Danger is, in this sense, a boundary concept: the underlying relation is not itself interpretive, but the entry's demonstrated encyclopedia work — organizing appraisal, Claim, Significance, Attention, and Recognition around that relation — is. If a future revision finds that the entry's center of gravity has shifted toward the underlying relation apart from any appraisal, Claim, or Recognition act — for instance, if a formal treatment of physical or structural risk assessment comes to depend on Danger without ever passing through one — this placement should be revisited; nothing in the relation's own definition requires Division IV, only the entry's demonstrated compositional yield does.

Significance and Danger are frequently, and wrongly, treated as the same axis at different intensities. Significance is a magnitude of mattering, direction-agnostic by its own formal definition: an opportunity, an obligation, and a threat can all carry high Significance under the right frame. Danger is not high Significance under a different name; it is the relation that may ground prospective-harm-directed Significance, not a subtype of mattering itself. An already-published example shows the conflation happening in practice rather than merely in theory: describing an appraisal as assigning "catastrophic Significance" to a call signals or smuggles an unarticulated Danger appraisal into Significance vocabulary. It supplies adverse direction but leaves the relational structure required for an examinable Danger Claim — source, subject, harm, pathway, conditions, and horizon — unstated. Formalizing Danger gives that sentence, and others like it, a home that does not require Significance to secretly carry valence it was never built to carry — Significance tells us how much something matters within a Frame; Danger identifies a prospective-harm structure that can give that mattering an adverse direction.

Attention and Danger are related recursively, not through a conflation risk running in only one direction. A Danger relation may hold without ever being attended to, and Attention may be allocated somewhere no Danger relation actually holds. Where an Agent does encounter relevant Evidence or a candidate source, some Attention to it is itself a precondition for the explicit appraisal that might follow — Attention makes appraisal possible; it does not constitute it. The resulting appraisal can then reallocate further Attention, toward vigilance and investigation or away from it through avoidance, denial, and normalization. Where Attention can be connected to a Danger appraisal, it may evidence a response to perceived Danger; by itself, Attention establishes neither that Danger was perceived nor that the underlying relation held.

Formal Pattern

Danger(source: s, subject: subj, harm: h, pathway: p, conditions: C, horizon: t)
  is well-formed iff each argument is supplied and typed as follows:
  s is a Source(s): an entity, situation, process, or condition capable of
    initiating harm, without itself requiring agency or any institutional
    position
  subj is a Subject_At_Risk(subj): the person, population, system, or thing
    the relation would be dangerous to
  h is a PossibleHarm(h): a possible future instance or continuation of harm,
    characterized by kind and, where support permits, severity
  p is a Pathway(p): the mechanism by which s could bring about h in subj
  C is EnablingConditions(C): the conditions that must obtain for p to be
    traversable — their absence can suspend the relation without eliminating
    the source
  t is a Horizon(t): the interval within which s could bring about h in subj
    via p, given C

  A well-formed Danger statement is not thereby a true one: supplying every
  slot produces a well-formed candidate, not a fact about the world.

Danger(source: s, subject: subj, harm: h, pathway: p, conditions: C, horizon: t)
  holds iff:
  C obtains during the relevant portion of t
  p is causally available under C
  p connects s to subj
  traversal of p under C can produce a future or continuing instance of h
    in subj during t

  This truth condition requires neither Evidence nor an appraisal nor
  Recognition to be satisfied — an unnoticed Danger relation must remain
  possible. Evidence belongs to the appraisal and the Claim it may support,
  not to the relation's own truth condition.

APPRAISES_AS_DANGEROUS(
  agent: a, source: s, subject: subj, harm: h, pathway: p,
  conditions: C, horizon: t, basis: b
)
  -- an Agent's act of Interpretation directed at whether this candidate
     relation holds; presupposes some Attention to s, to relevant Evidence,
     or to the situation; may be externalized as, or support, a Claim, not
     the relation itself, and can be mistaken in either direction

Claim(c, claimant: a, proposition: Danger(source: s, subject: subj, harm: h,
      pathway: p, conditions: C, horizon: t), modality: interpretive, context: Cx)
ASSERTS(Agent(a), Claim(c))
  -- Claim's own EpistemicProfile, InstitutionalDisposition, and
     EvidenceAttachments track how well-supported and how contested c is;
     none of them determines whether c's proposition is true

RECOGNIZES(recognizer: r, claim: c, status: st, context: Cx)
  -- institutional or collective uptake of a Danger Claim as warranting some
     status (e.g. "confirmed," "acknowledged," "unresolved"); a case of
     Recognition's general RECOGNIZES(recognizer, recognized-thing, status,
     context), targeting the Claim, not the relation directly

COUNTS_AS(s, Hazard, Cx, K, I)
  -- s bears Hazard status in context Cx under constitution K for
     institution I; distinct from Recognizing any particular Danger Claim,
     and from the relation's own truth. Count-As names a relation, not an
     act — the Decision, Recognition, or declaration by which an
     institution establishes or records that a source bears this status
     remains a separate question

Well-formed Danger(...) does not entail that Danger(...) holds.

Danger(...) holding does not entail:
  Attention
  appraisal
  a Claim
  Review
  Recognition
  Count-As standing

The existence, evidential support, Review, or Recognition of a
Danger Claim does not entail that its proposition is true.

A source Counting As Hazard within an institution does not entail
that every asserted Danger relation involving that source holds
for every subject, pathway, condition, or horizon.

Examples

Distinctions

Danger ≠ harm. Harm is realized injury or loss; Danger concerns possible future or continuing harm. A live wire can injure one person and remain dangerous to everyone else; a structural fault can cause one failure and continue to threaten further harm; even the same subject can remain exposed to another harm token along the same pathway. Realized harm may provide Evidence that a Danger relation existed and may still exist — it opens an accountability inquiry without necessarily closing the Danger inquiry.

Danger ≠ risk. Risk is not yet an entry in this encyclopedia. For purposes of this entry, risk is provisionally reserved for a quantified or comparative characterization involving possible harm, likelihood, consequence, exposure, or related measures, while Danger is the qualitative relation any such characterization would be of. A future Risk entry may revise this boundary; the contrast drawn here is a working one, not a canonical ontology.

Danger ≠ fear. Fear is a subject's or observer's affective response, which may follow an appraisal of Danger but is not required by it and does not require it in return. An unoccupied structurally unsound building may remain dangerous to future entrants, first responders, neighboring persons, or adjacent property even though no one is present to feel Fear; a false alarm can produce real fear with no Danger relation established anywhere.

Danger ≠ threat. For purposes of this entry, threat is treated as either an Agent's communicated intention to cause harm, or a source paired with an active, imminent pathway — a provisional working contrast, not a settled definition. Danger requires neither intent nor imminence: a dormant fault line is dangerous long before conditions make it threatening in the everyday sense.

Danger ≠ hazard. For purposes of this entry, hazard is treated as the source considered in isolation — a chemical, a steep grade, a live wire — apart from any particular subject, pathway, or horizon; again a provisional working contrast rather than a settled definition. Danger is the full relation: the same hazard can be highly dangerous to one subject with direct exposure and pose no Danger at all to another entirely outside the pathway's reach.

Danger ≠ Significance. Significance is a magnitude of mattering, direction-agnostic by its own formal definition — an opportunity, an obligation, and a threat can each carry high Significance under the right frame. Danger is not high Significance under a different name: it is the relation that may ground prospective-harm-directed Significance, not a subtype of mattering itself. A wedding can carry very high Significance without being dangerous. A minor exposure can instantiate a genuine Danger relation without receiving maximal Significance.

Danger ≠ Attention. Attention is the allocation of an Agent's bounded processing capacity to an object; Danger is a relation that may hold whether or not any capacity ever reaches it, and Attention may be allocated somewhere no Danger relation actually holds. Where Attention can be connected to a Danger appraisal, it may evidence a response to perceived Danger; by itself, Attention establishes neither that Danger was perceived nor that the underlying relation held.

Danger ≠ Power or aggression. Power is an Agent's capacity to produce effects within institutional or social fields; aggression is a disposition or behavior an Agent enacts. Danger requires neither agency nor any institutional position — a structural defect or a pathogen is dangerous without any Agent, Power, or aggressive intent behind it. An Agent can hold substantial Power while presenting no Danger to a given subject absent any pathway connecting them, and a source can be highly dangerous while wielding no Power at all.

Danger ≠ safety. This distinction needs the most careful handling of the nine. Safety is not simple negation — not merely "no Danger relation holds" — but the absence, mitigation, or governed management of Danger within some specified scope. A subject can be safe with respect to one Danger relation (a fenced pool) while remaining fully exposed to another in the same setting (an unfenced road on the same property). Treating a safety claim as unscoped negation loses exactly the relational structure that makes it auditable: a "safe" building is safe with respect to specified hazards, under specified conditions, inspected against a specified standard — never safe simpliciter.

Common Failure Modes

Minimum Viable Test Case

Case A — Danger without Attention or appraisal:

s    = Ventilation_System_4B
subj = Occupants(Building_12)
h    = PossibleHarm(kind: respiratory harm, severity: moderate-to-severe)
p    = Pathway(contaminant circulated through shared ductwork)
C    = EnablingConditions(filtration below threshold, contaminant source
         active, system not yet serviced)
t    = Horizon(until next scheduled service or independent detection)

Well-formedness:
  Danger(source: s, subject: subj, harm: h, pathway: p, conditions: C,
         horizon: t) is well-formed: every argument is supplied and typed.

Truth (checked independently of well-formedness):
  C obtains during t                                        -- yes
  p is causally available under C                            -- yes
  p connects s to subj                                        -- yes, shared ductwork
  traversal of p can produce a future instance of h in subj during t  -- yes
  → Danger(...) holds.

No appraisal, no Claim, no Attention:
  Relevant_Agents =
    Members(Occupants(Building_12)) ∪ {Facilities_Inspector_12,
      Building_Operator_12}   -- the bounded set of individual Agents
      positioned to encounter s; not an indefinite population
  for each a in Relevant_Agents:
    APPRAISES_AS_DANGEROUS(a, s, subj, h, p, C, t, basis: b) = false
    NOT_ALLOCATED(a, s, t)   -- per Attention's own predicate; Agents
      quantified individually, not modeled as a single collective Agent

Result:
  The Danger relation is both well-formed and true throughout t. No
  appraisal is ever made, no Claim is ever raised, and no Attention is ever
  allocated. The relation's truth depends on none of these. Deliberate investigation, governed response,
  and accountable mitigation ordinarily require some route by which the
  condition receives Attention and appraisal.

Case B — Attention and appraisal without an established relation:

e    = FlaggedPost_9931                            -- observed Evidence
s    = SuspectedCrisisCondition(author)             -- hypothesized source
subj = Author(FlaggedPost_9931)
h    = PossibleHarm(kind: self-injury)
p    = HypothesizedPathway(intent-to-action)
C    = HypothesizedEnablingConditions(asserted by the appraisal)
t    = HypothesizedHarmHorizon(asserted by the appraisal)
w    = ModerationReviewWindow(configured by procedure)

ALLOCATES(Moderator_a, capacity: q, to: e, within: w)
  -- Attention, per Attention's own predicate: some processing capacity is
     allocated to the Evidence; ALLOCATES does not require the allocation
     to be total

APPRAISES_AS_DANGEROUS(
  agent: Moderator_a, source: s, subject: subj, harm: h, pathway: p,
  conditions: C, horizon: t, basis: automated_pattern_match(e)
)

Claim(c, claimant: Moderator_a, proposition: Danger(source: s, subject: subj,
      harm: h, pathway: p, conditions: C, horizon: t), modality: interpretive,
      context: ModerationQueue_40k)
ASSERTS(Agent(Moderator_a), Claim(c))
  -- a Concern Hypothesis is raised around this proposition

Review, within w, finds:
  s, p, and C are not supported once the post is read in context — no
  suspected crisis condition, no viable pathway, and no enabling condition
  can be substantiated

Result:
  The appraisal and the Claim existed, grounded in real Evidence (the post)
  and made in good faith. The Danger proposition they asserted was never
  established. Attention and appraisal are evidence that someone treated a
  candidate relation as worth checking; they are not evidence that the
  relation held. Checking showed that the proposition was not adequately
  supported, not that it was false.

Cross-References

• Significance — the nearest and most commonly conflated neighbor: a direction-agnostic magnitude of mattering, where Danger is the relation that may ground prospective-harm-directed Significance, not a subtype of mattering itself
• Attention — dependencies run both ways with feedback: Attention to a source or to Evidence makes explicit appraisal of Danger possible, and a resulting appraisal can reallocate further Attention toward vigilance or away through avoidance and normalization; neither Attention's presence nor its absence by itself establishes whether the Danger relation holds
• Interpretation — an appraisal of Danger is a species of Interpretation directed at this specific relation, and, per Attention's own account, presupposes some Attention to the source, the Evidence, or the situation as an unstated precondition
• Meaning — Interpretation's canonical output and the required bearer of any subsequent Significance assignment; a Danger appraisal produces a prospective-harm-directed Meaning before that Meaning can be weighted
• Claim — an appraisal of Danger may be externalized as, or support, a Claim whose proposition is the candidate Danger relation itself; the Claim can exist, carry a basis, and be evidentially supported or contested whether or not that proposition is true
• Recognition — institutional or collective uptake of a Danger Claim, conferring a status such as "confirmed" within a context; targets the Claim, not the relation directly, and is distinct from the relation's own truth
• Count-As — where a source bears Hazard status within an institution, that classification is expressed by a Count-As relation; the act by which the institution establishes or records that standing remains distinct, and is separate from Recognizing any particular Danger Claim
• Review — a Danger Claim may be reviewed on its way toward a later Decision or Recognition; Review may inform that outcome but does not by itself confer status, classify a source, or establish the relation's truth
• Concern Hypothesis — where observed Evidence about a person supports a possible Danger interpretation, Concern Hypothesis may govern how that interpretation is bounded, reviewed, resolved, or retired without becoming an unreviewable personal status
• Power — a source can be highly dangerous while its owner or operator wields no institutional Power at all, and an Agent can hold substantial Power while presenting no Danger to a given subject absent a pathway
• Evidence — bears on whether the asserted source, subject, pathway, conditions, and possible harm are adequately supported; evidential support affects the warrant of a Danger Claim, is not the Claim's Significance, and does not itself determine institutional Recognition
• Standard — may specify criteria for evaluating pathways, enabling conditions, exposure, or acceptable thresholds; applying that Standard and conferring a Recognized status require further acts
• harm (possible) — the realized counterpart to possible harm; a realized instance may provide Evidence that Danger existed and may still exist, without extinguishing the relation while the pathway and enabling conditions remain
• risk (possible) — provisionally used in this entry for a quantified or comparative characterization of possible harm; a future Risk entry may revise this boundary rather than inherit it