PCIM — Predictive Claim Identification Module
Architecture Ecosystem: Structured Reality Standards™
Architecture Family: PGA™ — Prediction Governance Architecture
Parent Standard:
Prediction Formation Standard (PFOS)
Operational Layer: Prediction Formation Governance Layer
Category:
Governance & Enforcement
Subcategory:
Prediction Governance
Type: Parent Standard Module
Governed Space:
Predictive Claim Identification
Version: 1.0
Status: Canonical · Open Module
Effective Date: 8 September 2026
Compatibility: OOF® Methodology OS™ · GOA™ · ORA™ · VALIDOS™ · INTEGROS® · ADIT® · AGA™ · AIG® · ASGA™ · CLIA® · SIMULOS®
AI-Readable: Yes
Authority: OOF®
Protection:
MIP® — Methodological Intellectual Property
Canonical Language:
English (UCL™)
Governed Space
Predictive Claim Identification
Canonical Definition
Predictive Claim Identification Module (PCIM) defines the governance conditions under which a coherent Prediction Composition is transformed into a specific, explicit, semantically bounded, temporally anchored, condition-aware, and reconstructable Predictive Claim.
PCIM identifies exactly what future-oriented proposition is being asserted by the predictive process.
It transforms:
Governed Prediction Composition
into:
Identified Predictive Claim
by establishing the claim's:
subject; predicted state or event; direction; magnitude or range where applicable; Prediction Horizon; Predictive Boundary; applicable conditions; scenario dependence; comparison baseline; claim precision; semantic meaning; and claim state.
PCIM does not yet establish the final Prediction Output state.
That belongs to:
POFM — Prediction Output Formation Module.
Its foundational distinction is:
Predictive Output ≠ Predictive Claim
Operational Role
PCIM serves as the second internal governance space of the Prediction Formation Standard (PFOS) .
It receives:
Governed Prediction Composition
and determines what specific future-oriented claim that composition supports.
The module asks:
What exactly is being asserted about the future, and is that assertion defined precisely enough to be interpreted, communicated, monitored, and later compared with an outcome?
The governed progression is:
Governed Prediction Composition → Claim Candidate Identification → Claim Subject Definition → Predicted State / Event Definition → Horizon / Boundary Association → Conditionality Association → Claim Precision Assessment → Semantic Validation → Identified Predictive Claim
The primary output is:
Identified Predictive Claim
This becomes a principal input to:
POFM — Prediction Output Formation Module.
Internal Module Flow
Prediction Composition → Predictive Claim Identification → Prediction Output Formation → Prediction State Establishment → Prediction Formation Traceability
Current Position: Predictive Claim Identification
Module Operational Space
PCIM governs:
Predictive Claim identification; claim subject; claim object association; future-state representation; future-event representation; directionality; magnitude; range; threshold; timing; Prediction Horizon; Predictive Boundary;
comparison baseline; conditionality; scenario dependence; claim type; claim precision; claim granularity; semantic clarity; semantic consistency; claim uniqueness; claim ambiguity; compound claim structure; alternative claims; mutually exclusive claims; nested claims; derived claims; implicit claims; explicit claims; qualitative claims; quantitative claims; categorical claims; interval claims; threshold claims; event-occurrence claims; state-continuity claims; directional claims; timing claims; relational claims; claim normalization; claim decomposition; claim conflict; claim incompleteness; claim state; and historical claim integrity.
Module Function
A predictive process may contain extensive methodology, information, assumptions, calculations, and outputs while still failing to state clearly what it actually predicts.
Examples:
“82%”
“Growth expected.”
“High risk.”
“Likely failure.”
“Demand: 12,500.”
These may contain predictive meaning, but they may still be incomplete as governed Predictive Claims.
PCIM turns that latent predictive meaning into an explicit proposition.
For example:
Demand for Product A in Switzerland is predicted to exceed 12,500 units during Q4 2026 under governed Assumptions A–D.
That statement can later be:
interpreted; communicated; monitored;
revised; compared with an outcome; assessed for predictive performance.
Predictive Claim
A Predictive Claim is a specific future-oriented proposition established by the predictive process.
It defines what the prediction asserts is expected, probable, likely, projected, or otherwise predicted to occur within a governed predictive context.
A Predictive Claim may concern:
occurrence; non-occurrence; state; change; direction; magnitude; range; threshold crossing; timing; persistence; relationship.
Claim Subject
The Claim Subject identifies the entity, process, event, system, population, market, object, or governed phenomenon to which the Predictive Claim applies.
Examples:
demand for Product A; system availability; unemployment rate; probability of default; flood occurrence; project completion; patient admissions; supply interruption.
Claim Subject ≠ General Topic
“Energy market” is a topic.
“Wholesale electricity price in Market X” may be a sufficiently bounded Claim Subject.
Claim Object Association
The Claim Subject must remain associated with the Prediction Object established under POPS.
A predictive claim should not silently change the object being predicted.
Example:
Prediction Object:
Annual revenue of Business Unit A
Claim:
Corporate group revenue will grow 8%.
This creates an object mismatch unless explicitly governed.
Predicted State
A Predictive Claim may concern a future state.
Example:
Inventory will remain above 20,000 units on 31 December 2026.
The state must be sufficiently defined to permit later interpretation.
Predicted Event
A Predictive Claim may concern whether an event will occur.
Example:
A system outage lasting more than one hour will occur during the next 30 days.
The event definition must remain sufficiently bounded.
Undefined Event ≠ Governable Prediction Target
Occurrence Claim
An Occurrence Claim asserts that a defined event is predicted to occur within a governed period.
Non-Occurrence Claim
A Non-Occurrence Claim predicts that a defined event will not occur within the Prediction Horizon.
Non-occurrence claims require sufficiently bounded event and horizon definitions.
Directional Claim
A Directional Claim predicts movement in a defined direction.
Examples:
increase
decrease
remain stable
Direction alone may not establish magnitude.
Magnitude Claim
A Magnitude Claim predicts the amount or scale of a future state or change.
Example:
Revenue will increase by 8%.
The applicable baseline must remain identifiable.
Range Claim
A Range Claim predicts that an outcome will fall within a defined interval.
Example:
Demand is predicted to fall between 9,500 and 10,500 units.
The range is part of the Predictive Claim.
Threshold Claim
A Threshold Claim predicts whether a defined threshold will be crossed.
Example:
Demand will exceed 10,000 units during Q4.
The threshold and direction must remain explicit.
Timing Claim
A Timing Claim predicts when an event or state transition will occur.
Example:
The project is predicted to reach operational readiness by 31 March 2027.
Persistence Claim
A Persistence Claim predicts that a condition will continue through a defined period.
Example:
System availability is predicted to remain above 99.9% during Q1.
Relational Claim
A Relational Claim predicts a future relationship between governed variables or states.
Example:
Demand growth is predicted to remain above population growth during the next three years.
The compared elements and period must remain identifiable.
Quantitative Predictive Claim
A Quantitative Predictive Claim expresses the predicted future state numerically.
Examples:
12,500 units
4.2% growth
£8 million revenue
The units and applicable baseline must remain identifiable.
Qualitative Predictive Claim
A prediction may legitimately be qualitative where numerical precision is not justified.
Examples:
Demand is predicted to decline materially during Q4.
System reliability is predicted to deteriorate over the next six months.
Qualitative does not mean ungoverned.
Qualitative Claim ≠ Undefined Claim
Categorical Predictive Claim
A Categorical Claim predicts a future state represented by a defined category.
Example:
Risk state is predicted to move from Moderate to High within 90 days.
The categories must have governed meaning.
Claim Direction
Where the claim concerns change, direction should remain explicit where material.
Examples:
increase
decrease
remain unchanged
cross above
fall below
recover
deteriorate
Claim Magnitude
Where magnitude forms part of the claim, it should remain identifiable.
A claim such as:
Sales will increase.
is different from:
Sales will increase by 10%.
Both may be legitimate claims, but they possess different precision and later performance-assessment potential.
Comparison Baseline
Where a claim concerns change, the applicable comparison baseline must remain identifiable.
Possible baselines include:
previous period; current value; historical average; defined benchmark; scenario baseline; target value.
Predicted Change Without Baseline ≠ Fully Defined Change Claim
Prediction Horizon Association
Every Predictive Claim must remain associated with a sufficiently defined Prediction Horizon.
Examples:
within 24 hours
during Q4 2026
by 31 December 2027
within five years
A claim without a horizon may be impossible to evaluate meaningfully.
Future Event Without Governed Horizon ≠ Sufficient Predictive Claim
Point Horizon
A prediction may refer to a defined future point.
Example:
Inflation is predicted to be 2.5% on 31 December 2026.
Interval Horizon
A prediction may refer to a defined future interval.
Example:
At least one system failure is predicted during the next 90 days.
Open-Ended Horizon
Open-ended language such as:
“eventually”
may be insufficient for governed Prediction Formation unless the use case legitimately supports such a horizon and its limits are explicit.
Predictive Boundary Association
The claim must remain bounded to the relevant:
geography; population; organization; system; market; product; infrastructure; scenario; operational environment.
Same Claim + Different Boundary ≠ Same Prediction
Conditionality
A Predictive Claim may depend upon conditions.
Example:
If supplier capacity remains above 80%, production is predicted to exceed 50,000 units during Q4.
The conditional structure forms part of the claim.
Condition Removed ≠ Same Predictive Claim
Condition Placement
Conditions should be associated with the part of the claim they govern.
In compound predictions, one condition may apply to one subclaim but not another.
PCIM preserves this relationship.
Scenario Dependence
A claim may be valid only under a defined scenario.
Example:
Under the high-demand scenario, revenue is predicted to exceed £20 million.
Scenario dependence must remain explicit.
Scenario-Dependent Prediction ≠ Unconditional Prediction
Claim Precision
Claim Precision represents how clearly the Predictive Claim defines the future state or event it asserts.
Precision may concern:
subject;
event; state; magnitude; direction; horizon; boundary; conditions; baseline.
Precision should be sufficient for the Prediction Purpose.
Precision ≠ Accuracy
A highly precise claim may be wrong.
A vague claim may later appear “correct” only because it was insufficiently bounded.
Therefore:
Claim Precision ≠ Predictive Accuracy
Precision ≠ Certainty
Greater semantic precision does not imply greater certainty about the future.
A prediction may be precisely stated while carrying high uncertainty.
Claim Precision ≠ Predictive Certainty
False Precision
PCIM should not require numerical specificity that the methodology cannot legitimately support.
Example:
Methodology supports:
approximately 10,000–12,000 units
but the final claim becomes:
10,843 units.
This may create False Precision.
More Decimal Places ≠ Greater Predictive Legitimacy
Claim Granularity
Granularity determines the level of detail in the Predictive Claim.
Example:
National electricity demand will rise.
versus:
Residential electricity demand in Region A will rise by 6–8% during January 2027.
The appropriate granularity depends upon Prediction Purpose and available governance.
Overly Broad Claim
A claim may be too broad to remain meaningfully governed.
Example:
Technology will improve.
Such a statement lacks sufficient bounded predictive meaning unless further defined.
Overly Narrow Claim
A claim may be more specific than the methodology or Prediction Basis legitimately supports.
PCIM must therefore balance:
governability
with:
methodological support.
Semantic Clarity
The words used in the claim must possess sufficiently governed meaning.
Examples of potentially ambiguous terms include:
failure; success; growth; recovery; recession; high risk; significant; likely; stable.
Where such terms are material, their meaning must remain bounded.
Same Label ≠ Same Predictive Meaning
Two systems may both predict:
failure
while using materially different failure definitions.
PCIM preserves semantic distinction.
Implicit Predictive Claim
Some systems produce outputs without explicitly stating the underlying claim.
Example:
Risk score: 92
may implicitly be interpreted as:
failure is likely.
PCIM requires the actual Predictive Claim to be surfaced where the output is intended to function as a prediction.
Implicit Predictive Meaning ≠ Governed Predictive Claim
Explicit Predictive Claim
A Predictive Claim becomes explicit when its material semantic components are represented sufficiently for interpretation and reconstruction.
Compound Predictive Claim
A predictive statement may contain multiple subclaims.
Example:
Demand will rise by 10% and remain above supply capacity throughout Q4.
This contains at least:
1. a demand-growth claim; 2. a demand-vs-capacity claim.
PCIM determines whether they should remain one governed compound claim or be decomposed.
Claim Decomposition
Complex claims may need to be decomposed where subclaims have different:
horizons; conditions; objects; probability states; outcome criteria.
One Sentence ≠ One Predictive Claim by Default
Nested Claims
One claim may depend upon another.
Example:
If Candidate A wins, policy X is predicted to be implemented within one year.
This includes:
event condition; downstream prediction.
The structure must remain explicit.
Alternative Claims
A predictive process may produce alternative claims under different conditions.
Example:
Scenario A → Demand grows 8%.
Scenario B → Demand declines 3%.
These should remain distinguishable.
Mutually Exclusive Claims
Some predictive claims cannot simultaneously be true.
Example:
Candidate A wins
versus:
Candidate B wins
where only one can win.
Mutual exclusivity may become relevant later for probability and collective governance.
Competing Claims
Multiple candidate claims may exist during formation.
PCIM identifies them and establishes which claim or claims proceed toward Prediction Output Formation.
Derived Claim
A Predictive Claim may be derived from another predictive output.
Example:
Model predicts:
Revenue = £9 million
and a downstream system derives:
Company will breach debt covenant.
These are distinct predictive claims.
Derived Predictive Claim ≠ Source Predictive Claim
Claim Transformation
Where a predictive output is translated into a different form, the transformation must remain explicit.
Example:
Model output:
0.82
Claim:
Event X is predicted to occur with modeled probability 82%.
The transformation from score to claim must be governed.
Claim Normalization
Claims from different systems may need standardized representation for interoperability.
Normalization may include:
canonical subject; standardized units; explicit horizon; explicit boundary; normalized condition structure.
Normalization must preserve meaning.
Normalized Claim ≠ Changed Claim
Claim Uniqueness
A Predictive Claim should remain distinguishable from other claims associated with the same Prediction Object.
This may require:
unique identifier; claim version; formation reference; scope reference.
Duplicate Claim
Two records may represent the same underlying Predictive Claim.
Duplication should not create the false impression of multiple independent predictions.
Duplicate Records ≠ Multiple Predictions
Claim Conflict
A Claim Conflict occurs where two candidate claims cannot both represent the same intended prediction without clarification.
Examples:
Demand will rise 10%.
and:
Demand will fall 5%.
PCIM does not require forced reconciliation.
It requires the claims to remain distinguishable and governed.
Claim Ambiguity
A Claim Ambiguity exists where materially different interpretations remain possible.
Examples:
“Sales will rise next year.”
Ambiguities may include:
sales volume vs revenue; calendar year vs fiscal year; gross vs net; one market vs all markets.
Ambiguity should be resolved before final output formation where material.
Claim Incompleteness
A claim may be incomplete where material elements are absent.
Potential missing elements include:
subject; predicted state/event; horizon; boundary; baseline; conditions; units.
Claim Identification Gap
A Predictive Claim Identification Gap exists where the predictive composition cannot yet support a sufficiently clear, bounded, and reconstructable claim.
Possible responses include:
clarification; semantic definition; claim decomposition; horizon specification; boundary specification; conditional qualification; restriction; inability to identify a governed claim.
Claim State
A Predictive Claim may occupy states such as:
Candidate
Identified
Conditionally Identified
Ambiguous
Incomplete
Conflicting
Unable to Identify
These states remain distinct from the later overall Prediction Formation State under PSEM.
Historical Claim Integrity
Once a Predictive Claim is established for a prediction version, later wording changes must not silently change the historical claim.
Example:
Original claim:
Demand will exceed 10,000 units.
Later wording:
Demand likely reaches approximately 10,000 units.
These are not necessarily semantically equivalent.
Rewording ≠ Same Predictive Claim by Default
Claim Version
Material changes in:
subject; event; direction; magnitude; range; horizon; boundary; conditions
may create a new Claim Version.
Claim Comparison Readiness
A sufficiently governed claim should be capable of later association with a governed outcome reference.
This does not mean the outcome must already exist.
It means the claim is sufficiently defined to permit meaningful future comparison.
Outcome Testability
Where applicable, the claim should establish criteria under which later comparison can determine whether the predicted event/state occurred.
Examples:
Demand exceeds 10,000 units by 31 December 2026.
is more testable than:
Demand will be strong.
However, not every legitimate prediction must be reducible to a binary test.
Claim Identifiability and Probability
PCIM identifies what is predicted.
PPUS later governs the probability, uncertainty, confidence, and ranges associated with that claim.
Claim Identity ≠ Probability Representation
Claim Identifiability and Feasibility
PCIM establishes the claim.
PFES later determines whether the predicted outcome is realistically achievable under governed conditions.
Claim Identification ≠ Feasibility Determination
Claim Identifiability and Communication
PCIM establishes the canonical predictive meaning.
PCRS later governs how it may be communicated and relied upon.
A communication layer must not silently change the claim.
Claim Identifiability and Outcome Performance
POFPS later compares the Governed Prediction with a governed outcome reference.
Without a sufficiently identified claim, performance assessment may become arbitrary.
Predictive Claim Governance Sufficiency
Predictive Claim Identification is sufficiently governed when:
the Claim Subject is identifiable; the predicted event or state is defined; direction is explicit where material; magnitude/range is explicit where material; units are defined where material; baseline is defined where required; Prediction Horizon is explicit; Predictive Boundary is explicit where relevant; applicable conditions are preserved; scenario dependence remains visible; semantic meaning is sufficiently bounded; false precision is avoided; compound claims are decomposed where necessary; ambiguity and conflicts remain explicit; claim state is established; historical claim versions remain reconstructable.
Governance Requirements
PCIM requires that:
1. every Governed Prediction possess an identifiable Predictive Claim; 2. the Claim Subject remain associated with the governed Prediction Object; 3. the predicted state or event be sufficiently defined; 4. direction be represented where material;
5. magnitude or range be represented where material; 6. units remain identifiable where material; 7. comparison baseline remain identifiable where required; 8. Prediction Horizon remain explicit; 9. Predictive Boundary remain explicit where material; 10. conditionality remain attached to the claim; 11. scenario dependence remain explicit; 12. semantic ambiguity be resolved or preserved as a governance gap; 13. quantitative claims not imply unjustified precision; 14. qualitative claims remain sufficiently bounded; 15. categorical claims use governed categories; 16. threshold claims identify threshold and direction; 17. compound claims be decomposed where materially necessary; 18. nested claim relationships remain explicit; 19. alternative and competing claims remain distinguishable; 20. derived claims remain distinguishable from source claims; 21. transformations from output to claim remain traceable; 22. duplicate records not be treated as independent predictions; 23. Claim Conflicts remain explicit; 24. Claim Identification Gaps remain explicit; 25. claim state remain identifiable; 26. later rewording not silently alter historical claim meaning; 27. material claim changes produce distinguishishable versions or equivalent historical states; 28. claim identity remain distinct from probability representation; 29. claim identity remain distinct from feasibility determination; 30. claim structure support later outcome comparison where applicable.
Governance Logic
PCIM applies the following governance logic:
Governed Prediction Composition
↓
Candidate Predictive Meaning Identified
↓
Claim Subject Established
↓
Predicted Event / State Established
↓
Direction / Magnitude / Range / Threshold Defined where applicable
↓
Units / Baseline Defined where applicable
↓
Prediction Horizon Associated
↓
Predictive Boundary Associated
↓
Conditions / Scenario Dependence Associated
↓
Semantic Meaning Validated
↓
Claim Precision Assessed
↓
False Precision Avoided
↓
Compound / Nested / Alternative Claims Distinguished
↓
Ambiguity / Conflict Assessed
↓
Claim State Established
↓
Historical Claim Identity Preserved
↓
Identified Predictive Claim
↓
POFM — Prediction Output Formation Module
Where sufficient claim identification cannot be established:
Predictive Claim Identification Gap → Clarification / Definition / Decomposition / Qualification / Restriction / Unable to Identify
Minimum Implementation Framework
A minimum implementation of PCIM must establish:
A. Prediction Reference
The prediction formation process to which the claim belongs.
B. Claim Identifier
A unique or sufficiently distinguishable Predictive Claim reference.
C. Claim Subject
The entity, state, event, process, population, system, market, or object being predicted.
D. Claim Type
At minimum, where applicable:
Occurrence / Non-Occurrence / State / Direction / Magnitude / Range / Threshold / Timing / Persistence / Relational / Other
E. Predicted State or Event
The specific future-oriented proposition.
F. Direction
Where applicable:
Increase / Decrease / Remain / Cross Above / Fall Below / Occur / Not Occur / Other
G. Magnitude / Range / Threshold
Where applicable.
H. Unit
Where applicable.
I. Comparison Baseline
Where applicable.
J. Prediction Horizon
The applicable future time point or interval.
K. Predictive Boundary
The applicable geographic, organizational, system, population, market, scenario, or other boundary.
L. Conditionality
Any conditions materially governing the claim.
M. Scenario Reference
Where the claim is scenario-dependent.
N. Claim Precision State
At minimum:
Sufficient / Conditionally Sufficient / Ambiguous / Overly Broad / Overly Precise / Incomplete / Unable to Determine
O. Claim State
At minimum:
Candidate / Identified / Conditionally Identified / Ambiguous / Incomplete / Conflicting / Unable to Identify
P. Claim Version / Historical State
The applicable claim version or equivalent historical record.
Governance Outputs
PCIM may produce:
Identified Predictive Claim;
Predictive Claim Record; Claim Identifier; Claim Subject Record; Claim Type; predicted-state record; predicted-event record; direction record; magnitude record; range record; threshold record; unit record; comparison-baseline record; Prediction Horizon association; Predictive Boundary association; conditionality record; scenario reference; claim precision profile; semantic-definition record; compound-claim decomposition; nested-claim relationship; alternative-claim record; competing-claim record; derived-claim record; claim-transformation record; Claim Conflict Record; Predictive Claim Identification Gap; Claim State; Claim Version; historical claim record.
The primary governed output is:
Identified Predictive Claim
which becomes a principal input to:
POFM — Prediction Output Formation Module.
Critical Distinctions
Predictive Output ≠ Predictive Claim
Claim Subject ≠ General Topic
Undefined Event ≠ Governable Prediction Target
Qualitative Claim ≠ Undefined Claim
Predicted Change Without Baseline ≠ Fully Defined Change Claim
Future Event Without Governed Horizon ≠ Sufficient Predictive Claim
Same Claim + Different Boundary ≠ Same Prediction
Condition Removed ≠ Same Predictive Claim
Scenario-Dependent Prediction ≠ Unconditional Prediction
Claim Precision ≠ Predictive Accuracy
Claim Precision ≠ Predictive Certainty
More Decimal Places ≠ Greater Predictive Legitimacy
Same Label ≠ Same Predictive Meaning
Implicit Predictive Meaning ≠ Governed Predictive Claim
One Sentence ≠ One Predictive Claim by Default
Derived Predictive Claim ≠ Source Predictive Claim
Normalized Claim ≠ Changed Claim
Duplicate Records ≠ Multiple Predictions
Rewording ≠ Same Predictive Claim by Default
Claim Identity ≠ Probability Representation
Claim Identification ≠ Feasibility Determination
Enterprise Application
PCIM enables organizations to move from predictive outputs to claims that can actually be governed.
This matters especially in environments where users see:
dashboards; percentages; model scores; risk labels; ranges; forecasts; expert statements;
and naturally infer a predictive meaning that the producing system may never have formally established.
PCIM prevents organizations from discovering only after a decision, dispute, audit, or failed forecast that different people understood the same predictive output as different claims.
It creates a canonical meaning layer between:
Prediction Composition
and:
Prediction Output Formation.
Use Case 1 — Everyone Agrees on the Number but Not on What the Number Predicts
An AI maintenance system displays:
87%
for a critical industrial machine.
Operations interprets this as:
87% probability that the machine will fail within 30 days.
Maintenance interprets it as:
87% probability that intervention will be required within 30 days.
The model documentation reveals that the score actually predicts:
the probability that vibration measurements will cross a defined anomaly threshold within 30 days.
All three interpretations concern the same machine.
They are not the same Predictive Claim.
The critical governance question is:
Can a predictive output be legitimately relied upon when different users attach materially different future-oriented claims to the same number?
PCIM requires identification of:
the Claim Subject; exact predicted event; Prediction Horizon; applicable threshold; conditions; semantic meaning.
The result is one explicit Predictive Claim rather than several competing interpretations.
Predictive Output ≠ Predictive Claim
Use Case 2 — A Forecast Is Declared Correct Because the Original Claim Was Too Vague to Fail
A strategy team publishes:
“The market is expected to grow strongly next year.”
One year later, total market revenue grows by 1%, transaction volume declines by 4%, and one subsegment grows by 18%.
The team argues that the prediction was correct because the subsegment grew strongly.
No record establishes:
whether “market” meant total market or one segment; whether “growth” meant revenue or volume; what “strongly” meant; which period represented “next year.”
The critical governance question is:
Can predictive performance be legitimately assessed when the original claim was broad enough to support multiple incompatible interpretations after the outcome became known?
PCIM requires the Predictive Claim to be bounded before the future unfolds.
For example:
Total market revenue in Region A is predicted to increase by more than 5% during calendar year 2027 compared with calendar year 2026.
The claim can now later be assessed without rewriting its meaning.
A prediction should not become more precise only after the outcome is known.
Canonical Boundary
PCIM governs the identification, semantic definition, subject association, event/state representation, direction, magnitude, range, threshold, baseline, Prediction Horizon, Predictive Boundary, conditionality, scenario dependence, precision, decomposition, conflict, state, versioning, and historical preservation of the Predictive Claim. It does not govern Prediction Composition itself, final Prediction Output Formation, overall Prediction Formation State, probability and uncertainty representation, Prediction Feasibility, Collective Prediction composition, communication and reliance, prediction-level revision, outcome determination, or predictive performance assessment.
PCIM establishes:
what exactly the prediction claims about the future.
It does not yet determine:
the final governed output state through which that claim becomes the formed prediction.
Canonical Closing Statement
A prediction cannot be governed merely because its producer knows what an output was intended to mean. The predictive claim itself must carry that meaning.
PCIM converts predictive output meaning into an explicit, bounded, and historically stable claim before that claim can become the final prediction output.
If the future-facing claim cannot be identified, the prediction itself has not yet been fully formed.