Type: Node / Governed Examination
Placement note: Review belongs in Division VII because it is not merely looking at something again. It is an attributable, governed examination whose reviewer, purpose, grounds, scope, and outcome can be held institutionally answerable. Process supplies its temporal genus; Procedure, Rule, Role, Standard, or another governing basis supplies the conditions under which a particular Review operates.
Working Definition: A review is a governed examination of an identified existing object by an attributable reviewer, for a stated purpose, within a stated scope, against disclosed grounds or criteria, producing a traceable outcome. Review may inform later institutional acts, but it does not by itself decide, judge, recognize, publish, authorize, or otherwise give its outcome binding force.
Function in the Grammar: Review makes institutional re-examination possible without collapsing scrutiny into disposition. It creates a governed place where a Claim, Evidence item, Record, Process, Decision, Judgment, Rule, Procedure, artifact, or prior Review can be examined in light of a purpose and disclosed basis. The outcome may identify findings, questions, limitations, disagreement, a Recommendation, a request for revision, or no warranted conclusion. Those outputs remain available for contestation and later use.
The reviewed object must exist independently of the Review. A manuscript is not its peer review. A Decision is not the appellate Review that later examines it. A Process is not the dependency Review that assesses its effects. This does not require the object to have been reviewed before. It means Review takes an already existing object as its subject rather than creating that subject through the act of examination.
Review is a specialized Process, not a Procedure. It unfolds through attributable Actions, may involve several reviewers or iterations, and may remain incomplete. A Procedure may govern how the Review is initiated, who may participate, which materials are available, what Standards apply, how conflicts are handled, and what Records are required. The Procedure specifies possible form. The Review is what actually occurs.
The act, outcome, Record, and review stage must also remain distinct. Completing a Review may justify changing an object's review_stage from under-review to reviewed, but the metadata is not the examination. A Record may represent what was examined and concluded, but it is not the Review itself. A review outcome may later ground a Decision, Judgment, Recognition, or Recommendation without becoming any of them.
Formal Grammar Representation
<review> ::= Review(
id: <review-id>,
object: <object-reference>,
object_version: <version-or-state-reference>,
reviewer: <agent-or-agent-set>,
forum: <forum-reference-or-none>,
purpose: <review-purpose>,
governed_by: <governing-basis>,
procedure: <procedure-reference-or-none>,
grounds: <ground-set>,
scope: <scope>,
interval: <time-interval>,
outcome: <review-outcome-or-pending>
)
<object-reference> ::= Claim | Evidence | Interpretation | Record | Process
| Decision | Judgment | Rule | Procedure | Artifact
| Review | other identified object
<ground-set> ::= disclosed Standard, Rule, criteria, questions,
Evidence requirements, or other stated bases of examination
<review-outcome> ::= a traceable statement of what the examination found,
questioned, could not establish, or proposes for later action
Review(r) is well-formed when:
object(r) = x, where x and its reviewed version or state are identifiable
reviewer(r) = a or A, where each reviewing Agent is attributable and any
collective result discloses its aggregation basis
forum(r) = f or none, where a constituted Forum is identified when relevant
purpose(r) = P and scope(r) = S
governed_by(r) = g, where the basis for this Review is disclosed
grounds(r) = G, where the terms of examination are inspectable
interval(r) = T, where initiation, completion, or pending state can be located
outcome(r) = o when complete, including an explicit no-conclusion outcome
Semantic Constraints
The expression Review(r, x, a, f, P, g, G, S, T, o) is well-formed when the object, reviewer, optional Forum, purpose, governing basis, grounds, scope, interval, and outcome or pending state are identifiable. The syntax does not establish that the reviewer was competent, independent, attentive, adequately informed, free from conflict, faithful to the governing Standard, or correct.
A Forum may receive a Review, constitute the setting in which it occurs, or recognize its collective outcome. It does not literally examine the object. The reviewing Agents, their Roles, and any aggregation Rule must remain attributable, including when ordinary language says that a court, board, committee, or institution "reviewed" something.
Review does not authorize or apply itself. A Rule, Procedure, Role, Constitution, agreement, request, or other governing basis must explain why this Review is being performed and what authority or standing the reviewer has. Anyone may examine an object informally, but an institution is not bound to treat that examination as its Review.
reviewed is not a truth, quality, acceptance, or validity score. It records that a relevant examination was completed for a particular object version, purpose, scope, and basis. A reviewed Claim may remain false or challenged. A reviewed manuscript may remain unpublishable. An accepted item may have received defective Review.
A HumanReview requires an attributable human reviewer who performs the substantive examination required by the governing basis. A model-generated report followed by a human click is not HumanReview when the human did not perform the required scrutiny. A machine or hybrid Agent may perform a Review if the governing grammar permits and scopes that participation, but its producer type remains visible and it does not satisfy a specifically human review requirement.
A completed Review does not carry binding force by default. A governing Rule may make Review completion or a stated outcome a condition for a later transition. The Decision, Judgment, Recognition, or constitutive operation that supplies the effect must remain explicit.
Review outcomes do not require a universal taxonomy. "Assessment," "finding," "question," "request," and similar terms may remain ordinary outcome vocabulary unless the admission test later establishes independent concepts. A Review can itself become the object of another Review without creating a mandatory regress. Further Review occurs only when a governing basis, challenge, or practical need calls for it.
Formal Pattern
Review(r, object: x, version: vx, reviewer: a, forum: f, purpose: P,
governed_by: g, procedure: pi, grounds: G,
scope: S, interval: T, outcome: o)
REVIEWS(
Agent(a),
Object(x),
in: Review(r),
for: Purpose(P),
against: Grounds(G),
within: Scope(S)
)
AGGREGATES(AgentSet(A), Review(r), under: AggregationRule(h), for: Forum(f))
GOVERNS(Procedure(pi), Review(r), under: GoverningBasis(g))
when Procedure(pi) is specified
APPLIES_TO(Standard(s), Object(x), in: Review(r))
when Standard(s) is part of Grounds(G)
COMPLETES(Agent(a), Review(r), with: outcome(o), at: t2)
PRODUCES(Review(r), outcome(o))
CREATES(Agent(a), Record(k), representing: Review(r))
ATTESTS(Receipt(z), completion_of: Review(r), object_version: vx)
INFORMS(outcome(o), Recommendation(rec) | Decision(d) |
Judgment(j) | Recognition(n))
INFORMS expresses possible use, not automatic progression. The receiving act must disclose how the Review outcome bears on it, and may depart from the outcome where its own governing basis permits.
Core Relations
| Relation | Notes |
|---|---|
| Process supplies the genus | Review is an actual unfolding of examination through time. It may pause, branch, involve several reviewers, or end without a determinate conclusion. |
| Procedure may govern | Procedure specifies entry conditions, participation, materials, transitions, deadlines, outputs, and Records. It is not evidence that the actual Review conformed. |
| Agent and Role locate responsibility | Review requires an attributable reviewer. Role may establish competence, standing, duties, or limits without proving that the reviewer performed adequate scrutiny. |
| Standard and governing grounds orient | Standards, Rules, questions, or criteria disclose what the Review examines and for what purpose. Their existence does not prove correct application. |
| Evidence and Interpretation inform | Review may examine Evidence and rely on Interpretation. Their provenance, warrant, and limitations remain independently assessable. |
| Recommendation may result | Review may propose a course while leaving authority to decide elsewhere. It may also conclude without recommending anything. |
| Decision, Judgment, and Recognition may follow | Later acts may use a Review outcome to resolve, settle, or confer status. Review itself performs none of those functions merely by being complete. |
| Record and Receipt preserve | A Record represents the Review and its outcome. A Receipt may attest completion under stated conditions. Neither artifact proves substantive adequacy. |
| Epistemic Status records a dimension | Review stage records how far an epistemic object has been examined. It does not replace the Review's reviewer, object version, grounds, scope, or Record. |
| Appeal and Accountability may invoke | Appeal may request Review of a prior Judgment or Decision. Accountability may require Review as a means of answerability. Neither is reducible to Review. |
Typical Questions
- What exact object and version or state were reviewed?
- Who performed the Review, in what Role, with what standing, competence, and disclosed conflicts?
- What purpose, governing basis, scope, and Procedure applied?
- What Standards, Rules, questions, criteria, or Evidence requirements governed the examination?
- Which materials were available, excluded, missing, or contested?
- When did the Review begin, end, pause, or become stale?
- What outcome was produced, including any limits or unresolved questions?
- Where are the Review Record and any required Receipt?
- What later Recommendation, Decision, Judgment, Recognition, or repair action used the outcome?
- Is
reviewedbeing used to overclaim truth, acceptance, adequacy, or human involvement? - Can the Review itself be challenged, and through what forum or Procedure?
Examples
| Domain | Object and purpose | Reviewer and governing basis | Possible outcome without automatic force |
|---|---|---|---|
| Ordinary life | A household budget reviewed for affordability before a move | Partners using bank Records and shared constraints | Questions, revised assumptions, or a Recommendation to delay |
| Law | An administrative Decision reviewed for procedural and legal error | Authorized court or review panel under the applicable appeal Procedure | Findings that may ground a separate Judgment |
| Science | A manuscript reviewed for methodological adequacy and contribution | Peer reviewers under journal criteria and conflict rules | Critique and Recommendation to an editor |
| Publishing | An encyclopedia draft reviewed for publication readiness | Human author under the editorial protocol and rubric | Approval recommendation, revision request, or unresolved concern |
| Markets | A proposed acquisition reviewed for financial and operational risk | Board committee and advisers under delegated scope | Risk findings and conditions proposed to the board |
| Community governance | An election Process reviewed for accessibility and procedural conformance | Independent community panel under adopted terms of reference | Identified deviations and proposed repair |
| Digital systems | A code change reviewed for correctness, security, and maintainability | Human engineer, automated checks, or both under repository rules | Comments, requested changes, or evidence supporting later merge Decision |
Distinctions
Review ≠ Process in general. Every Review is a Process, but not every Process is a Review. A conversation or storm may unfold without an identified object, reviewer, purpose, or grounds of examination. Review adds that governed epistemic structure.
Review ≠ Procedure. Procedure is the normative specification that may govern a Review. Review is the actual examination that unfolds. A peer-review policy may be impeccable while the reviewer reads the wrong manuscript version or performs no substantive examination.
Review ≠ Interpretation. Interpretation states what an input is taken to mean. Review examines an object for a purpose against grounds. A reviewer may interpret a contractual clause while reviewing a Decision, but the interpretation and the broader Review remain distinct.
Review ≠ Judgment. Judgment is an authorized act that settles what claims or evidence establish within jurisdiction and produces a binding Finding. Review produces a traceable outcome that may inform Judgment. A panel report can identify error without itself setting aside the prior Judgment.
Review ≠ Decision. Decision selects what will govern next. Review examines and reports. An editor may receive a completed peer Review and still decide to accept, reject, defer, or seek further Review.
Review ≠ Recommendation. Recommendation proposes a course to a separate decider. Review may produce one, but it may instead identify questions, limitations, disagreement, or no warranted conclusion. Advice can also be given without conducting a Review.
Review ≠ Recognition. Recognition accepts or confers a status within scope. Review may support recognition, but a favorable Review does not itself make a manuscript published, a person eligible, or a Claim accepted.
Review ≠ Appeal. Appeal is a governed challenge to a prior Judgment or Decision that invokes Review by an authorized forum. Review is broader: peer review, dependency review, and ordinary editorial review need not arise from a challenge.
Review ≠ Review stage or Review Record. Review is the examination. under-review and reviewed are metadata conditions of an object. A Review Record represents what occurred. A field or file can falsely claim completion. A genuine examination may occur without an adequate Record, but later actors cannot verify it as an accountable Review.
Common Failure Modes
| Mode | Description |
|---|---|
| Phantom or checkbox review | A status, signature, or form claims that Review occurred without evidence of substantive examination. Where to look: reviewed fields with no reviewer, object version, grounds, comments, outcome, timestamp, or supporting Record. |
| Human-review laundering | Automated analysis does the substantive work while a human performs only a nominal approval, yet the result is represented as HumanReview. Where to look: human attestations made too quickly to inspect the object, or identical approvals generated from model summaries without access to the source. |
| Object or version mismatch | The reviewer examines a different object, version, time period, or scope from the one later described as reviewed. Where to look: publication or approval Records whose object hash, revision identifier, attachments, or effective dates differ from the reviewed materials. |
| Grounds opacity or drift | Criteria are undisclosed, changed during Review, or applied selectively. Where to look: outcomes justified by requirements absent from the governing Standard, or materially similar objects assessed under different unstated criteria. |
| Reviewer capture or incapacity | The reviewer lacks relevant competence, access, independence, or freedom from conflict. Where to look: Reviews controlled by the party whose conduct is examined, or assigned to a reviewer denied necessary Evidence. |
| Outcome laundering | A Review outcome or Recommendation is treated as a Decision, Judgment, Recognition, or proof of validity without the further authorized act. Where to look: status changes and enforcement actions citing only a review report. |
| Indefinite review as veto | Review remains pending without a completion condition, responsible reviewer, escalation path, or justified need for further examination. Where to look: objects held in under-review status after deadlines or evidentiary questions have ceased to change. |
Minimum Viable Test Case
x = encyclopedia_entry_standard, version: v3
a = human_author
s = encyclopedia_keystone_rubric_v1
pi = editorial_review_procedure_v1
r1 = Review(
object: x,
object_version: v3,
reviewer: a,
purpose: publication_readiness,
governed_by: editorial_protocol,
procedure: pi,
grounds: s,
scope: conceptual_and_editorial_adequacy,
interval: t1..t2,
outcome: revise_two_relations
)
REVIEWS(
human_author,
x,
in: r1,
for: publication_readiness,
against: s,
within: conceptual_and_editorial_adequacy
)
COMPLETES(human_author, r1, with: revise_two_relations, at: t2)
CREATES(human_author, review_record_k, representing: r1)
q1 = Recommendation(return_x_for_revision)
PRODUCES(r1, q1)
Result:
r1 satisfies the human Review requirement for x at version v3.
q1 remains advisory.
A separate human Decision determines whether x is revised, deferred,
rejected, or later published.
Failure case:
If model_assistant produces the report and sets human_reviewed: true
while human_author performs no substantive examination, a model Review
may have occurred, but the required HumanReview did not.
Cross-References
Required
- Process: the temporal genus of the actual examination
- Procedure: may govern initiation, participation, transitions, outputs, and Records
- Agent: the attributable reviewer whose participation and capacity remain inspectable
- Role: may establish standing, competence, duties, and limits for the reviewer
- Standard: may supply disclosed criteria for the examination
Consequential
- Recommendation: may propose a course on the basis of a Review outcome
- Decision: may adopt, reject, modify, or disregard what the Review proposes
- Judgment: may use Review findings in settling a matter with binding force
- Recognition: may confer a status for which Review is a condition
- Accountability: may require examination as one means of answerability without being exhausted by a Review
- Appeal: may contain a Review of a prior disposition, but remains a broader governed challenge Process
Evidential
- Evidence: may be examined or used as material within Review
- Interpretation: may inform Review without becoming the whole examination
- Record: represents the Review, its object version, grounds, and outcome
- Receipt: may attest that a governed Review was completed
- Epistemic Status: records review stage as one dimension of an object's profile