Type: Node / Selective Allocation of Processing Capacity
Working Definition: Attention is the selective allocation of an Agent's bounded processing capacity to an object during a time window, relative to other objects eligible or able to receive that capacity. It is a practical precondition for Interpretation, Review, and other acts insofar as they require examination of that object, but it is not sufficient for any of them: capacity may be allocated too briefly, shallowly, or unsuccessfully for the presupposed act to occur. Attention may be allocated independently of whether a governing basis requires the examination or Authority permits an ensuing act.
Function in the Grammar: Attention is what makes Interpretation, Review, Decision, and every other act that presupposes examining something possible in practice, without itself producing any of them. Its absence is a distinct failure from the absence of Authority to act or a Rule requiring the act, and its presence is not proof that a presupposed act succeeded. An object can be fully within an Agent's Authority to examine, fully required by a Rule or Procedure to be examined by a stated deadline, and still never actually get examined, because none of the Agent's own bounded processing capacity was ever allocated to it in the relevant window. This sharpens the parallel with Power: Power is a capacity to produce effects; the underlying resource here is processing capacity, a capacity to examine; Attention is the selective allocation of that processing capacity to one object rather than another. Attention does not require a governing basis to exist or matter; an Agent's attention can be allocated to something no Rule ever mentions, and can fail to reach something a Rule explicitly requires.
Significance, where it has been formally classified, is among the things that may bias allocation toward an object; a highly significant object is likelier to receive Attention than the same object buried in a large enough set of competing claims, but Significance does not allocate Attention by itself. A Concern Hypothesis can satisfy every required element, including a well-formed ReviewPath and ResolutionDeadline, and still receive no Review because no reviewing Agent's processing capacity is allocated to it before the deadline closes. This is not an explanation of that entry's existing "Missing review path" or "Unresolved into permanence" failure modes, which name a missing structure and an unresolved outcome respectively; it is a distinct failure this project's grammar had not yet named, a fully specified path that still goes unattended. Danger, drafted separately, may bias the allocation, intensity, or persistence of Attention, including through capture or avoidance. The relationship is recursive rather than sequential: Attention to a source or to Evidence makes explicit appraisal-as-dangerous possible, and the resulting appraisal can reallocate further Attention, toward vigilance or away through avoidance and normalization, one act later rather than within the same act. Danger is not a kind or magnitude of Significance. A Danger appraisal is an Interpretation that produces a threat-directed Meaning to which Significance may subsequently be assigned. That appraisal or Significance assessment may then bias a later allocation of Attention.
Formal Pattern
Attention(att, agent: a, object: o, window: w) iff
a possesses bounded processing capacity during w
o is a member of CompetingSet(O, a, w): the objects eligible or able
to receive a's capacity during w
some portion q of that capacity is allocated to o during w
ALLOCATES(a, capacity: q, to: o, within: w)
NOT_ALLOCATED(a, o, w) iff
o is a member of CompetingSet(O, a, w)
and no portion of a's processing capacity is allocated to o during w
Attention(att) does not entail:
INTERPRETS(a, o)
REVIEWS(a, o)
sufficient examination for any particular act
Authority(a, act, o)
a governing basis requiring attention to o
ALLOCATES identifies where an Agent's processing capacity is directed; it does not entail that the Agent deliberately chose the object. Allocation may result from deliberate selection, salience, capture, habit, imposed sequencing, or other causal conditions.
NOT_ALLOCATED describes an actual absence of allocation, not merely an absence of evidence that allocation occurred. Assignment records, access logs, elapsed time, testimony, and work products may support an Attention claim, but no single proxy necessarily proves or disproves actual Attention. A record of Attention is not Attention.
Examples
| Domain | Object competing for Attention | What determines allocation | What happens without it |
|---|---|---|---|
| Digital governance | A flagged post in a 40,000-item moderation queue | Queue ordering, staffing levels, competing flagged items | A deadline-bound Concern Hypothesis passes its ResolutionDeadline unreviewed: structurally well-formed but procedurally unfulfilled |
| Law | A single motion in a judge's crowded docket | Docket position, filing volume, case urgency | A timely, properly filed motion sits unread past a hearing date no one revisits |
| Ordinary life | One child's quiet distress next to a sibling's loud tantrum | Volume and salience of competing demands, not comparative need | The quieter distress goes unaddressed, not because raising it was unauthorized, but because no attention reached it |
| Science / publishing | One manuscript in a reviewer's overcommitted stack | Reviewer's own remaining capacity, other overdue commitments | A submission with a valid review assignment and a real deadline gets no substantive review before it lapses |
Distinctions
Attention ≠ Authority. Authority is the recognized right to act; Attention is the allocation of processing capacity to an object regardless of any right to do so. A fully authorized reviewer with no capacity left to allocate produces no Review. An Agent with no authority to act on an object can still attend closely to it. The two can align or diverge independently, and neither implies the other.
Attention ≠ Power. Power is the causal capacity to produce an effect once an Agent acts, typically within an institutional or social field. Attention is the allocation of processing capacity to an object, prior to any effect being produced from it. An Agent may hold ample Power to act decisively on something it never actually attends to, and may attend closely to something it has no power to change.
Attention ≠ Significance. Significance assigns an interpretive magnitude or weight of mattering within a Frame. Significance does not allocate Attention by itself: an object can carry high Significance and still never receive Attention, exactly as a routine object can. Conversely, an object can receive sustained Attention disproportionate to its Significance, captured by volume, familiarity, or institutional habit rather than by what actually matters.
Attention ≠ Interpretation or Review. Attention allocates processing capacity to an object; Interpretation produces Meaning from it, while Review examines it according to a governed basis and produces an assessment or recommendation. Attention is necessary for those acts insofar as they require examination, but neither a momentary allocation nor sustained fixation on an object proves that either act actually occurred.
Common Failure Modes
| Mode | Description |
|---|---|
| Specified but unattended | Every structural prerequisite is present, a ReviewPath, deadline, assigned Authority, and eligible reviewer, but the required examination does not occur, because no portion of any eligible Agent's processing capacity was ever allocated to the object before the window closed. Distinct from "Missing review path" (no structure exists) and only one possible mechanism behind "Unresolved into permanence" (which may equally result from a missing disposition rule, procedural blockage, failed escalation, or institutional refusal). Where to look: objects with a fully specified review or decision requirement, for which assignment and routing records exist but no evidence beyond assignment supports an inference that processing capacity reached the object. Absence of a record is an investigative lead, not proof of NOT_ALLOCATED. |
| Allocation mistaken for examination | Some processing capacity reaches the object, a brief opening, a glance, a notification acknowledged, a momentary fixation, and this is treated as proof that Interpretation or Review occurred. Attention was allocated, but not necessarily sufficiently or successfully for the downstream act it was supposed to enable. Where to look: a recorded touch, view, or acknowledgment cited as though it were itself the completed Interpretation or Review. |
| Volume capture | A high volume of routine or low-significance objects consumes bounded capacity before a smaller number of more significant objects receive it, not because a comparative judgment determined that the routine objects mattered more, but because aggregate volume and position controlled allocation. This can occur inside a rule-governed queue. Where to look: a genuinely important item that received no more scrutiny than routine ones in the same queue. |
| Role mistaken for allocation | An Agent is assigned a role that presupposes examining something, "reviewer of record," "on-call," "responsible party," and that assignment, which may confer Authority, Obligation, or responsibility, is treated as though it guarantees Attention was actually given, rather than only that examination was authorized or owed. Where to look: outcomes attributed to a named role-holder's judgment, for which the role assignment is recorded but no evidence beyond that assignment supports an inference that the role-holder's capacity was ever actually allocated to the object. Absence of a record is an investigative lead, not proof of NOT_ALLOCATED. |
| Disproportionate capture | An object continues to receive Attention past what its own Significance warrants, crowding out other competing claims, through habit, familiarity, or institutional momentum rather than any continued warrant. Where to look: an object still consuming review capacity long after the condition that first justified the attention has changed or resolved. |
Minimum Viable Test Case
a = Moderator_17
o = FlaggedPost_88214
w = review_window(deadline: t0 + 4h)
O = CompetingSet(items assigned or available to Moderator_17, within: w)
REVIEW_REQUIRED(a, o, by: end(w))
AUTHORIZED(a, review, o)
PRESENT_IN_QUEUE(o, O, w)
ALLOCATES(a, capacity: q, to: o, within: w) = false
-- higher-volume competing traffic in O exhausted a's capacity
before any portion reached o within w
NOT_ALLOCATED(a, o, w)
Result:
o's Concern Hypothesis remains structurally well-formed. Its ReviewPath
exists, its ResolutionDeadline is specified, and a reviewer is
authorized. Nevertheless, REVIEWS(a, o) does not occur. The review
obligation is procedurally unfulfilled because no processing capacity
is allocated to o before w closes, not because any structural element
was missing.
Failure case, contrast:
A second item, FlaggedPost_88215, receives an allocation of a's
capacity within w because it happens to surface earlier in the queue,
despite carrying materially lower Significance than FlaggedPost_88214.
→ demonstrates that allocation tracks competing volume and position,
not Significance, confirming that Attention ≠ Significance is not
merely definitional but produces different real outcomes for
otherwise comparable cases. It does not by itself demonstrate that
FlaggedPost_88215 was actually interpreted or reviewed, only that
some capacity reached it.
Cross-References
- Power — the nearest structural parallel: effect-producing capacity vs. selective allocation of processing capacity
- Authority — normative permission that Attention does not require and does not supply
- Interpretation — Interpretation of an object presupposes some allocation of processing capacity to it; Attention alone does not entail Interpretation
- Review — a governed act that can be fully specified and still fail for want of Attention
- Concern Hypothesis — a fully specified ReviewPath and ResolutionDeadline can still go unattended; distinct from that entry's existing "Missing review path" and "Unresolved into permanence" failure modes, and evidence toward a future amendment adding "specified but unattended" there directly
- Significance — may bias allocation toward an object without performing that allocation
- Danger — related recursively, not sequentially: Attention to a source or to Evidence makes explicit appraisal-as-dangerous possible, and the resulting appraisal can reallocate further Attention toward vigilance or away through avoidance and normalization; neither Attention's presence nor its absence establishes whether the Danger relation holds