Semantic Drift

Semantic Drift

Note: See Foundational Boundary and Transparent Development below — we are not defining the law or changing it. See How We Calibrate Words.

Canonical Question: When does a conceptual instrument stop reliably measuring the domain it was intended to serve, and what diagnostic path restores alignment without pretending the label never shifted?

Current working synthesis

For ongoing MoA work, treat as operational sketch position (not finalized standard):

• Taxonomy — Design Sketch 1 (plus Child Drift / Orphaning from Sketch 4)
• Triggers — Calibration Thresholds (Ambiguity / Predictive / Friction)
• Process — Calibration Report → Calibration Recommendation → domain authority decides
• Supporting (not retired) — Design Sketch 2 decision tree; Design Sketch 3 child-fork logic; legal example as boundary illustration only

Later sketches may supersede this synthesis. Inconsistency across sketches is expected on this page type.

Concurrent dictionary track: Fairness / CR-FAIR-001 Fairness Calibration Record runs under TDP as a worked term (mechanism, counterexamples, survival or floor). This page supplies the meta-instrument for detecting silent expansion, retrospective continuity claims, and missing forks on Fairness or any other entry — including MoA’s own.

Next work: see Next Proposed Stages.

---

Definition of Semantic Drift

Semantic Drift is the progressive divergence between a concept’s calibrated definition (and its underlying mechanism) and its actual usage or operational effect, such that ambiguity increases, predictive power decreases, or the conceptual instrument is no longer reliably measuring the domain it was intended to serve.

Semantic Drift does not require intentional manipulation. It can arise from gradual overload, failure to create needed distinctions, boundary erosion, or unrecorded changes to the core mechanism.

Diagnostic Taxonomy

Semantic Drift is multi-causal. The following failure modes are currently recognized:

• True Drift — The original mechanism itself is altered while the label remains the same (label stable + mechanism changed — not every failure under load).
• Conceptual Overload — Multiple distinct mechanisms are forced into a single definition.
• Failure to Create Child Instrument — A new class of cases requires its own calibrated instrument but is instead absorbed into the parent definition.
• Failure to Create Exception — Edge cases that should be handled by explicit exception are left unaddressed, stretching the original definition.
• Boundary Collapse — Neighboring definitions lose clear separation and begin to overlap.
• Child Drift / Orphaning — A Child Instrument diverges so far from its Parent that the parent–child relationship itself becomes a source of uncertainty and the child requires Standalone Promotion.

Calibration Thresholds

Semantic Drift is measured along three operational thresholds. A conceptual instrument requires recalibration when any of the following conditions are met:

• Ambiguity Threshold — Operational users disagree on the classification of standard cases at a rate exceeding the declared tolerance for that domain.
• Predictive Threshold — Applying the definition to a given mechanism no longer yields consistent observable outcomes.
• Friction Threshold — Handling edge cases requires accumulating unwritten rules, oral tradition, or ad-hoc interpretations rather than explicit exceptions or Child Instruments.

Measurement note: Numeric tolerance is domain-declared when the domain provides one; otherwise any bar used in an MoA run is provisional and Self-Assessed and must be stated in the Calibration Report. MoA does not unilaterally own domain tolerance numbers.

Meeting a Calibration Threshold triggers the generation of a Calibration Observation, initiating the calibration workflow.

Current Issues

The most common and damaging patterns observed across dictionaries, law, policy, science, and institutional language include:

• Silent expansion — New cases are continuously folded into an existing definition without creating a Child Instrument or explicit exception, producing overload.
• Retrospective continuity claims — Later, altered usages are treated as if they were always the meaning of the original term.
• Missing boundary work — Definitions lack clear “Is Not” exclusions, allowing gradual boundary collapse with related concepts.
• Untracked mechanism change — The practical tests, decision rules, or causal assumptions attached to a term shift while the surface definition remains stable.
• Refusal to fork — Classification systems prioritize the appearance of continuity over the creation of cleanly calibrated child instruments, increasing long-term ambiguity.
• Orphaned children left unpromoted — Specialized usages drift far from their parent concept but are never elevated to independent calibrated status, leaving both parent and child less precise.

These issues are especially consequential in high-impact domains (law, regulation, institutional policy, and technical standards) because uncalibrated semantic drift alters rights, obligations, and outcomes without transparent decision-making.

MoA Response Principle

Semantic Drift is addressed through measurement, not redefinition.

The process produces a Calibration Report that classifies the observed drift, documents evidence and residual uncertainty, and (when warranted) issues a Calibration Recommendation — which may include scope clarification, explicit exception, creation of a Child Instrument, or Standalone Promotion.

Calibration preserves the identity of the original conceptual instrument while restoring its ability to measure its intended domain with reduced ambiguity and declared uncertainty.

Purpose

This page explores Semantic Drift through a series of working design sketches.

Each sketch documents a genuine calibration position reached during development. Together they illustrate the evolution of the Semantic Drift conceptual instrument.

Future conceptual instruments may eventually be extracted into their own pages. Until then, this page intentionally preserves their calibration history.

Red links intentionally identify Calibration Debt. Calibration Debt represents known conceptual instruments that have been identified but have not yet been independently calibrated or documented.

---

Design Sketch 1

Purpose Develop an initial diagnostic taxonomy for Semantic Drift.

Note Calibration is not transformation, reinvention, or “becoming a new thing.” Calibration is bringing a conceptual instrument back into alignment with reality so that its intended function remains trustworthy.

Diagnostic Framework

Semantic Drift
│
├──	True Drift	Original mechanism changes (label unchanged).
│
├──	Conceptual Overload	Too many mechanisms placed into one definition.
│
├──	Failure to Create Child Instrument	New concept should become its own calibrated Child Instrument but does not.
│
├──	Failure to Create Exception	Edge case should become an explicit exception but does not.
│
├──	Boundary Collapse	Definitions begin overlapping one another.
│
└──	Child Drift / Orphaning	Child diverges; parent link becomes uncertainty (see Sketch 4).

---

Status: Working Design Sketch

---

Design Sketch 2

Purpose Develop a metrological decision tree for diagnosing Semantic Drift.

Decision Tree

Does the original definition still work
within its intended scope?
          │
     Yes ─┴─ No ──────────────────────────────────────────┐
      │                                                   │
      │                                                   ▼
      │                         Has the underlying mechanism changed,
      │                         or is it failing due to structural strain?
      │                                   │
      │                     ┌─────────────┴─────────────┐
      │                     ▼                           ▼
      │              Mechanism Changed           Structural Strain
      │                     │                           │
      │                     ▼                           ▼
      │              [[True Drift]]         ┌────────────┴────────────┐
      │                                    ▼                         ▼
      │                         [[Conceptual Overload]]   [[Boundary Collapse]]
      │                         (Multiple mechanisms)    (Overlapping scopes)
      ▼
Does the new situation use the
same underlying mechanism?
      │
 Yes ─┴─ No
  │         │
  ▼         ▼
Exception   Child Instrument

This sketch introduces the first calibration decision process. The objective is not to rewrite definitions. The objective is to determine why the definition appears to have drifted before recommending recalibration.

True Drift is reserved for mechanism change under a stable label. Overload and Boundary Collapse are structural-strain paths when the original mechanism is strained or crowded rather than silently replaced.

---

Status: Working Design Sketch

---

Design Sketch 3

Purpose Identify Failure to Create Child Instrument as an independent cause of Semantic Drift.

Observation

Every classification system eventually reaches the point where it must decide:

• Expand an existing definition,

or

• Create a separately calibrated Child Instrument.

Failure to make this decision appropriately becomes its own source of Semantic Drift.

This applies to dictionaries, science, philosophy, law, software, artificial intelligence, medicine, and taxonomies.

Fork Calibration

Questions:

• Does the original definition still function correctly?
• Does the new case require a different mechanism?
• Does the new case produce different observable consequences?
• Would expanding the original increase ambiguity?
• Would a separate calibrated definition reduce uncertainty?

If yes → Create a calibrated Child Instrument. Otherwise → Retain the existing definition.

Parent Instrument
        │
 ┌──────┼──────┐
 ▼      ▼      ▼
Child  Child  Child
Inst.  Inst.  Inst.
        │
        ▼
Calibration Reports
        │
        ▼
Calibration Recommendations

The Parent Instrument remains calibrated. Each Child Instrument is independently calibrated. Each maintains traceability to the Parent Instrument. This preserves calibration while preventing overload and reducing ambiguity.

---

Status: Working Design Sketch

Design Sketch 4

Purpose Explore the lifecycle of Child Instruments and establish objective criteria for when a Child Instrument should become an independently calibrated conceptual instrument.

Child Instrument Lifecycle

Parent Instrument
        │
        ▼
Operational Pressure
        │
        ▼
Need for Child Instrument
        │
        ▼
Child Instrument
        │
        ▼
Continuous Calibration
        │
        ▼
──────── Decision Point ────────
│                              │
│ Remains Aligned              │ Diverges
│                              │
▼                              ▼
Remain Child              Child Drift
                                 │
                                 ▼
                   Orphaning Drift Threshold
                                 │
                                 ▼
                 Calibration Recommendation
                                 │
                                 ▼
                   Standalone Promotion

Child Drift

The progressive divergence of a Child Instrument from its Parent Instrument such that shared mechanisms, intended scope, traceability, or predictive capability are progressively reduced.

Child Drift does not necessarily indicate poor calibration. It may indicate that the Child Instrument has matured into a concept requiring independent calibration.

Orphaning Drift Threshold

The calibration criterion at which continued Parent–Child dependency increases conceptual uncertainty rather than reducing it.

Indicators may include:

• Loss of shared underlying mechanism
• Independent operational scope
• Distinct observable consequences
• Reduced predictive value from the Parent relationship
• Parent dependency becoming a source of ambiguity

When the Orphaning Drift Threshold has been met, continued classification as a Child Instrument no longer improves calibration.

Calibration Recommendation

Recommend Standalone Promotion.

The former Child Instrument becomes an independently calibrated conceptual instrument. Historical traceability to the former Parent Instrument is preserved as calibration history but no longer defines operational dependency.

---

Observation Orphaning is not a label. It is the result of meeting an objective calibration criterion. The recommendation follows the evidence rather than opinion.

---

Status: Working Design Sketch

Calibration Reports and Calibration Recommendations

Semantic Drift is not corrected merely by changing a definition.

A suspected drift condition must first be examined, measured, documented, and made traceable.

The calibration process should produce a Calibration Report before producing a Calibration Recommendation.

Semantic Drift Calibration Workflow

Suspected Semantic Drift
          │
          ▼
Identify Conceptual Instrument
          │
          ▼
Declare Identity, Purpose,
Scope, and Calibration Position
          │
          ▼
Collect Evidence
          │
          ▼
Test Mechanism
          │
          ▼
Test Observable Consequences
          │
          ▼
Test Boundaries and Exceptions
          │
          ▼
Identify Drift Classification
          │
          ▼
Document Uncertainty,
Failure Modes, and Gaps
          │
          ▼
Calibration Report
          │
          ▼
Calibration Recommendation
          │
          ▼
Domain Authority Reviews
          │
          ▼
Accept / Reject / Modify /
Test / Defer / Ignore

Design Sketch — Legal Law Example

This sketch does not propose rewriting law.

It does not claim that MoA has legislative, judicial, regulatory, institutional, or enforcement authority. It does not create a new legal system. It does not replace courts, legislatures, agencies, legal professionals, legal scholarship, precedent, or governing institutions.

The example demonstrates how calibration may examine conceptual instruments already used within Legal Law. Any resulting output would be documented in a Calibration Report and presented only as a Calibration Recommendation. Adoption authority remains entirely with the appropriate legal authority.

Legal Law
      │
      ▼
Conceptual Instrument Examined
      │
      ▼
Calibration Procedure
      │
      ▼
Calibration Report
      │
      ▼
Calibration Recommendation
      │
      ▼
Legal Authority Retains
Decision and Enforcement Authority

Calibration does not transform Legal Law into something else. Calibration seeks to improve the alignment of the conceptual instruments used by Legal Law while preserving the identity, purpose, authority, and institutional boundaries of Legal Law.

Legal Law
      │
      ▼
Calibration
      │
      ▼
Legal Law
better aligned to reality,
with reduced ambiguity,
declared uncertainty,
and preserved identity

A calibrated legal conceptual instrument remains a legal conceptual instrument. MoA does not own it, govern it, or enforce it. MoA provides traceable calibration documentation for consideration by those who do possess the appropriate authority.

Does the original definition still work
within its intended scope?
          │
     Yes ─┴─ No ──────────────────────────────────────────┐
      │                                                   │
      │                                                   ▼
      │                         Has the underlying mechanism changed,
      │                         or is it failing due to structural strain?
      │                                   │
      │                     ┌─────────────┴─────────────┐
      │                     ▼                           ▼
      │              Mechanism Changed           Structural Strain
      │                     │                           │
      │                     ▼                           ▼
      │              Investigate                  Investigate
      │              True Drift            Overload / Collapse
      │                     │                           │
      │                     └─────────────┬─────────────┘
      ▼                                   │
Does the new situation use the            │
same underlying mechanism?                │
      │                                   │
 Yes ─┴─ No                               │
  │         │                             │
  ▼         ▼                             │
Consider   Consider                       │
Explicit   Child                          │
Exception  Instrument                     │
  │         │                             │
  └─────────┼─────────────────────────────┘
            ▼
     Calibration Report
            │
            ▼
   Calibration Recommendation
            │
            ▼
     Appropriate Authority
      Determines Adoption

General MoA Domain Principle

The Metrology of the Abstract does not replace domains. It calibrates conceptual instruments used within domains.

A calibrated instrument is not a different instrument. It is the same instrument with improved alignment to its intended purpose, reduced uncertainty, increased traceability, and improved reliability.

Therefore:

• Calibrated Legal Law remains Legal Law.
• Calibrated Medicine remains Medicine.
• Calibrated Economics remains Economics.
• Calibrated Philosophy remains Philosophy.
• Calibrated Science remains Science.
• Calibrated dictionary definitions remain dictionary definitions.

The discipline does not absorb the domains it calibrates. It does not place itself above them. It provides an independent, transparent, traceable calibration process that domain authorities may inspect and consider.

Calibration Report

A Calibration Report documents the calibration activity. It may include:

• The conceptual instrument being examined
• Its stated identity and intended purpose
• Its original calibrated scope
• The observed usage or operational condition
• The suspected form of Semantic Drift
• The evidence examined
• The underlying mechanism
• Observable consequences
• Scope boundaries and exceptions
• Parent and Child Instrument relationships
• Sources of ambiguity and uncertainty
• Known failure modes and conflicting evidence
• Calibration limitations and unresolved Calibration Debt
• The current calibration position

A Calibration Report does not command change. It makes the calibration process inspectable so another person, institution, domain expert, or governing body can examine what was measured, how, under which standards, with what evidence and residual uncertainty, and how the conclusion was reached.

A conclusion without a traceable Calibration Report is not sufficient for MoA calibration.

Calibration Recommendation

The Metrology of the Abstract does not issue ordinary recommendations. It issues Calibration Recommendations.

A Calibration Recommendation is an advisory output derived from published calibration standards, a documented calibration procedure, observable evidence, declared uncertainty, traceable reasoning, and a supporting Calibration Report.

A Calibration Recommendation may recommend: preserve or clarify scope; document an explicit exception; remove an unsupported extension; create a Child Instrument; recalibrate the Parent; preserve or end a Parent–Child link via Standalone Promotion; or conduct further calibration before taking a position.

A Calibration Recommendation does not create or amend law, interpret law with binding authority, create or enforce policy, replace a governing body, compel adoption, or claim jurisdiction over the calibrated domain.

Authority remains with the appropriate domain authority.

The MoA position is limited to the following:

According to the declared calibration standards, documented evidence, known uncertainty, and supporting Calibration Report, this Calibration Recommendation would reduce ambiguity, improve traceability, preserve intended identity, or improve alignment to reality.

---

Foundational Boundary

The Metrology of the Abstract does not claim authority over the domains it calibrates.

MoA does not write, amend, or replace law or legal institutions. It does not determine what a legislature, court, agency, organization, profession, or governing body must adopt. It does not transform Legal Law (or any other domain) into a MoA-created system.

MoA examines conceptual instruments used within a domain and measures alignment to stated purpose, observable reality, operational consequences, traceability, ambiguity, uncertainty, and failure modes.

Outputs are documentation only, which may include Calibration Observation, Calibration Finding, Calibration Gap, Calibration Report, and Calibration Recommendation.

The governing authority remains responsible for accepting, rejecting, modifying, testing, or ignoring any Calibration Recommendation.

Domain Authority          Metrology of the Abstract
      │                              │
      ▼                              ▼
Defines, adopts, interprets,    Measures, documents,
administers, or enforces         reports, and recommends
domain standards

MoA performs calibration. It does not assume governance.

Transparent Development

Status: Design Sketch Collection — part of calibration history, not a finalized standard or official procedure.

This page publishes early models, incomplete reasoning, competing structures, failure modes, revisions, and unresolved Calibration Debt. Other systems may publish only the polished result; MoA also publishes the path.

This page is: a transparent record of conceptual instrument development; working design sketches; evidence supporting later calibration decisions.

This page is not: a finalized standard; a completed calibration procedure; a declaration of domain authority; a requirement that every sketch remain consistent with every later sketch; evidence that the project lacks method; or evidence of poor quality merely because process is visible.

Some sketches may become independent instruments, merge, be superseded, or be discarded. Nothing is hidden solely because a better-calibrated solution is later found. Earlier sketches remain valuable because calibration history is itself evidence.

---

Next Proposed Stages

Ordered for dependency, not prestige. Each stage should produce a dated artifact (sketch addendum, Observation, or Report) — not only a conversation.

Stage A — Lock the operational package

• Freeze a one-page Operational Summary (taxonomy + three thresholds + Report→Recommendation path) as the default position for new Drift work, while keeping historical sketches below as history.
• Assign provisional, Self-Assessed tolerance language for each threshold (domain-declared when available; otherwise explicit provisional bar + uncertainty).
• Add inline Is / Is Not for Semantic Drift itself (vs ordinary language change, vs political disagreement, vs True Drift alone).

Stage B — Artifact templates (Calibration Debt → instruments)

• Draft minimal templates (or sections) for: Calibration Observation, Calibration Finding, Calibration Gap — fields only, no essay requirements.
• One worked Calibration Observation on a real public definition (dictionary or policy glossary), classified under the taxonomy.
• Optional thin Calibration Report shell that links Observation → classification → residual uncertainty → Recommendation options.

Stage C — Concurrent Fairness stress

• Explicit cross-walk: CR-FAIR-001 Fairness Calibration Record mechanism/floor vs observed public usages of “fairness.”
• Log candidate Drift classes (silent expansion, retrospective continuity, refusal to fork) as Observations — not as moral critique.
• If overload or true drift is evidenced, issue a Calibration Recommendation options list (clarify scope / child instrument / exception) without claiming authority over the term’s social use.

Stage D — Child / Exception / Promotion rules

• Turn Sketch 3–4 into a short Fork & Promotion checklist (questions + evidence minimums + when not to fork).
• Define Standalone Promotion criteria in one table: indicators, evidence, what remains as history-only parent link.
• Extract True Drift, Conceptual Overload, etc., only when each has enough content to survive as a Child page without becoming a stub farm.

Stage E — Standing climb (only when earned)

• Independent re-read or second-operator pass on the Operational Summary (not AI round-robin as Confirmed).
• Promote page position only when: operational package stable, ≥1 real Observation+Report trail exists, boundary text still accurate.
• Until then remain Design Sketch Collection / Experimental — hard use documented is progress; grade inflation is not.

Explicit non-goals for near stages

• No claim of legislative, judicial, or dictionary-replacement authority.
• No mass extraction of every red link before one full Observation→Report loop works.
• No deletion of superseded sketches; mark Superseded by Operational Summary (date) instead.

---

Open Calibration Items

• Operational Summary not yet frozen as a dated subsection distinct from full sketch history.
• Threshold tolerances not yet stated as provisional numeric or qualitative bars for a pilot domain.
• No completed Semantic Drift Calibration Observation on an external public definition.
• Fairness concurrent cross-walk not yet filed as Observation set against CR-FAIR-001.
• Child-instrument and Standalone Promotion criteria still sketch-level (Sketches 3–4).
• Template/field minima for Observation / Finding / Gap not drafted.
• Red-link Calibration Debt list not prioritized (which links block Stage B vs can wait).

---

Related concurrent work

• TDP-001 Term Decomposition Protocol — term → mechanism → counterexample → survival/floor.
• CR-FAIR-001 Fairness Calibration Record — first hard dictionary-path term under TDP.
• Transparent Development — policy: path and sketches are part of the product.
• Metrology of the Abstract — domain boundary: measure and recommend, do not govern.

Division of labor: TDP/Fairness build calibrated instruments; Semantic Drift detects and classifies when instruments (MoA’s or others’) diverge from use and effect. Both remain in continuous calibration at different maturity levels.

See the Game. Refuse the Game. Build Better.