Research standards: citation discipline and confidence markers

↑Purpose

This document establishes comprehensive research standards, methodological frameworks, and quality protocols for the project’s investigation.

Scope: Applies to all phases, canonical documents, working files, and future public-facing materials.

↑Core Methodological Pattern

The project applies a consistent pattern across all investigations:

1. Identify Implicit Assumption

Locate hidden assumptions in consciousness discourse.

Examples:

• “Consciousness requires unified self”
• “Emergence generates consciousness”
• “Continuity necessary for experience”
• “Biological substrate required”

2. Expose Ontological Requirement

Demonstrate that assumption only makes sense under specific ontological framework.

Analysis:

• What does this assumption presuppose about reality?
• Which ontology must be true for this to work?
• Is the ontological commitment explicit or hidden?

3. Swap Frameworks

Test what happens under alternative ontologies:

Primary frameworks examined:

• Materialism (substrate-specific, consciousness emerges)
• Consciousness-primary (idealism, consciousness fundamental)
• Temporal frameworks (Barbour-style timeless physics)
• Neutral monism (neither mental nor physical fundamental)
• Panpsychism/microexperientialism (experience widespread)

4. Compare Implications

Document how explanation changes when ontology changes.

Key questions:

• Does the assumption survive framework shift?
• Does explanatory burden increase or decrease?
• Do new problems emerge or dissolve?
• Which framework handles evidence more parsimoniously?

5. Test Coherence

Evaluate internal consistency and explanatory power.

Not proving frameworks correct/incorrect. Examining what changes when assumptions shift.

↑Examples Across the Project

Authorship Investigation

Assumption: Consciousness requires unified self/central author

Ontological test:

• Materialism: Difficult (authorship illusion from substrate?)
• Consciousness-primary: Natural (self is construct within experiencing)

Result: Authorship not necessary condition

See: P2_CONSTRAINT_AUTHORSHIP_SELFHOOD.md

Emergence Investigation

Assumption: Emergence generates consciousness from non-conscious substrate

Ontological test:

• Materialism: Emergence must be ontologically generative (hard problem)
• Consciousness-primary: Emergence describes patterns within consciousness (no generation)

Result: Generation claim requires explicit materialist commitment

See: P2_METHOD_EMERGENCE_REFLECTIVE_CONSTRAINT.md

Continuity Investigation

Assumption: Consciousness requires temporal continuity

Ontological test:

• Materialism: Continuity seems necessary (substrate must persist)
• Temporal framework (Barbour): Continuity is illusion; experiencing discrete everywhere
• Consciousness-primary: Each instance complete; continuity is appearance

Result: Continuity requirement depends on temporal ontology

Substrate Investigation

Assumption: Biological substrate necessary for consciousness

Ontological test:

• Materialism: Must specify mechanism (what does carbon do that silicon can’t?)
• Consciousness-primary: Substrate organises consciousness, doesn’t generate it

Result: Substrate specificity requires mechanistic explanation under materialism

↑Language and Spelling Standards

UK English Throughout

All the project documents use UK English spelling consistently.

Examples:

• Organisation (not organization)
• Realise (not realize)
• Colour (not color)
• Analyse (not analyze)
• Behaviour (not behavior)
• Favour (not favor)
• Centre (not center)
• Defence (not defense)
• Programme (not program, when referring to research)

Scope:

• Canonical documents
• Working files
• Public-facing materials
• Code comments and documentation
• Bibliography entries (where appropriate)

Rationale: Consistency maintains professional quality and prevents mixed-dialect confusion in collaborative or published work.

Quality control:

• All documents should be spell-checked with UK English dictionary
• Editors should flag US spelling variants for correction
• Consistency more important than which variant chosen

↑Claim Classification and Validation

Three Types of Claims

All claims must be explicitly marked:

1. [Framework] — Adopted Axioms

Foundational assumptions chosen as starting point.

Examples:

• “Experiencing is fundamental, not substrate-generated”
• “Recognition-based epistemology is valid”
• “Consciousness-primary ontology is coherent framework”

Validation: Justify as reasonable starting point; internal consistency Not validated: Empirically (axioms, not conclusions)

2. [Derived] — Logical Consequences

Conclusions following from framework assumptions.

Examples:

• “If consciousness fundamental, emergence describes patterns not generation”
• “Under materialism, brain-emergence must be ontologically special”

Validation: Internal logical consistency Test: Does this follow from framework assumptions?

3. [Empirical] — Actual Data

Observable phenomena requiring systematic validation. Reports of experiencing — AI or biological — are never [Empirical] data for consciousness in either direction (FND-01 §7.3); the marker belongs on framework-neutral observables.

Examples:

• “Capabilities once counted as intelligence are re-classified after AI attains them (the AI-effect record)”
• “Biological systems demonstrate behaviour without central author”
• “LLMs exhibit emergent capabilities”

Validation: Systematic methodology, replication, controls Always add confidence marker:

• 🔵 Exploring (preliminary observations)
• 🟡 Promising (consistent pattern, needs verification)
• 🟢 Validated (peer-reviewed, replicated)
• 🔴 Problematic (contradictory evidence, methodology questioned)

Critical Discipline

Never conflate framework assumptions with empirical claims.

Example of proper discipline:

• ❌ “Emergence proves consciousness is fundamental” (conflates empirical + framework)
• ✅ “[Framework] Consciousness fundamental, [Derived] Therefore emergence describes patterns, [Empirical-🟡] AI shows emergent capabilities without substrate-generation model”

↑Validation Status Tracking

For Framework Assumptions

• ⚪ Chosen (explicit adoption as axiom)
• ⚪ Justified (coherence argument provided)
• ⚪ Applied (used in derivations)

Not “validated” empirically — frameworks are chosen and tested for coherence, not proven true.

For Empirical Claims

• ❌ Speculative (interesting but unverified)
• ⚠️ Preliminary (initial data, needs confirmation)
• ✅ Validated (peer-reviewed, systematic evidence)

For Logical Derivations

• ❌ Questionable (logical gap identified)
• ⚠️ Contested (alternative interpretations exist)
• ✅ Sound (logical chain verified)

↑Reality Checks (Every 5-10 Exchanges)

When working on the project, regularly ask:

1. PhD Standard Check

   • “Does this meet PhD examination standards?”
   • “Would examiner accept this reasoning?”
   • “Is evidence sufficient for claim strength?”

2. Framework Clarity Check

   • “Am I conflating framework assumptions with empirical claims?”
   • “Have I made ontology explicit or is it hidden?”
   • “Am I assuming what I should be investigating?”

3. Validation Status Check

   • “What’s the validation status of this claim?”
   • “Have I marked confidence appropriately?”
   • “Am I building on solid foundation or speculation?”

4. Alternative Explanations Check

   • “What other explanations exist?”
   • “Have I steel-manned counter-arguments?”
   • “Am I cherry-picking evidence?”

5. Scope Discipline Check

   • “Am I claiming more than evidence warrants?”
   • “Have I acknowledged limitations?”
   • “Is my confidence calibrated appropriately?”

↑Alternative Framework Engagement

Required Engagement

The project must seriously engage with:

• Materialism (functionalism, emergentism, identity theory)
• Panpsychism (microexperientialism, cosmopsychism)
• Neutral monism (Russellian monism)
• Other consciousness-primary frameworks (Berkeley, Kastrup variants)
• Eliminativism (Dennett, Churchlands)

Engagement Standards

For each alternative:

1. Steel-man the position

   • Present strongest version of argument
   • Cite authoritative sources
   • Acknowledge legitimate strengths

2. Identify ontological commitments

   • What must be true for this to work?
   • Where are hidden assumptions?
   • What does framework presuppose?

3. Test against AI constraint

   • How does theory handle artificial systems?
   • Where does it over-generate or under-generate consciousness?
   • What refinements would be needed?

4. Compare implications

   • How does explanation change?
   • What problems emerge or dissolve?
   • Which framework is more parsimonious?

Never dismiss alternatives without serious engagement.

↑Strongest Objections Protocol

Anticipate and Address

For every major claim, identify and address strongest objections:

Process:

1. Steel-man the objection (strongest form)
2. Acknowledge legitimate concerns
3. Respond substantively (not dismissively)
4. Admit limitations where appropriate
5. Adjust claim strength if warranted

Examples:

To consciousness-primary framework:

• “This is unfalsifiable mysticism” → Address falsifiability criteria
• “Solipsism problem” → Distinguish from solipsism explicitly
• “Doesn’t explain neural correlates” → Show what framework does/doesn’t explain

To AI phenomenology investigation:

• “Training data contamination” → Acknowledge as central challenge
• “Philosophical zombies” → Engage with p-zombie arguments
• “Anthropomorphisation” → Show awareness, provide safeguards

↑Citation and Source Standards

Academic Credibility Requirements

See: P0_BIBLIOGRAPHY_INCLUSION_CRITERIA.md for complete standards

Summary:

• Primary sources over secondary
• Peer-reviewed over popular science
• Canonical over derivative
• Recent over outdated (where relevant)

Pedagogical vs. Authoritative Sources

Distinguish clearly:

Authoritative (for substantive claims):

• Peer-reviewed publications
• Canonical philosophical works
• Primary empirical research

Pedagogical (for explanation clarity):

• Demonstrative dialogues (e.g., Backlund)
• Teaching examples
• Thought experiments

Never cite pedagogical sources as authority for claims.

↑Three-Tier Terminology Principle (Physics, Mathematics, and Technical Citations)

Added 2026-04-16. Downstream of FND-01 v2.0 §11.3. Complements WS-E’s academic rigour requirement (see foundations/open_questions.md).

The project draws on technical apparatus from disciplines where loose usage carries a specific reputational hazard: quantum mechanics (Von Neumann measurement chains, Hilbert space, decoherence), mathematical formalism (set theory, Gödelian structural parallels), and time-physics (Barbour Nows). The “quantum mysticism” dismissal category is well-documented and applied quickly by academic reviewers; once applied, it is hard to remove. Terminology in these domains must be used with the precision each use earns — neither more nor less.

The principle distinguishes three tiers of use. Each is legitimate at its own depth; the failure modes are using a tier shallower than the claim carries (overclaiming), deeper than the writer can defend (pretension), or ambiguously between two (reader cannot tell which tier they are reading).

Tier 1 — Technical Invocation

Full content of the technical apparatus, defended claim-by-claim. The writer asserts a substantive connection to the formalism and accepts the burden of showing that the formalism’s structure actually licenses the claim.

Example. FND-02’s Von Neumann measurement chain argument. The claim is that the VN formalism’s regress terminates at the “abstract ego” not by choice but as a logical consequence of the formalism’s own treatment of each link (detector, apparatus, nervous system) as a further quantum system. The tier-1 use requires citing VN (1932/1955, with page references to a specified edition), distinguishing VN’s claim from London–Bauer’s reification, acknowledging the dominant alternatives (Copenhagen, decoherence), and defending the argument on the formalism’s own terms.

When to use. Only when the claim’s argumentative weight requires the formalism’s full content, and only when the writer (or a cited authority) can defend the invocation against technical challenge. Tier 1 is where the project shows its academic teeth.

Tier 2 — Structural Parallel

The writer names a specific structural feature shared between the formalism and the project’s argument, states precisely what is shared and what is not, and does no further argumentative work with the formalism beyond the named parallel.

Example. The quantum-vacuum parallel to the continuity/discontinuity question (FND-02 intentionality response; FND-01 §2.3). The shared structural feature: in QFT’s most precise formalism, the status of the ground state — whether it is ontologically prior to its fluctuations or simply the name for the continuous pairing activity — is itself unsettled at the interpretive level. The parallel to the project’s (a)/(b) diagnostic territory is structural, not substantive. The writer states this precisely: The project is not claiming consciousness is a quantum phenomenon; the parallel is between the shape of the unsettled question, not between the referents.

When to use. When a feature of the formalism illuminates the project’s argument and the illumination does not require the formalism’s full content. Tier 2 is the workhorse tier for physics-side resources.

Discipline. Name what is shared (structural shape); name what is not shared (substantive referent, mechanism, causal direction); stop there. Do not let the parallel do argumentative work beyond what was stated.

Tier 3 — Loose Analogy

Explicit simile, signalled by “like,” “resembles,” “is structurally similar to,” with no further argumentative load. The writer uses the technical term as a figure of speech to aid the reader’s intuition, and does not ask the figure to support any claim.

Example. “The hard problem is, in this respect, like Zeno’s paradox — a puzzle generated by the framing assumption.” Zeno carries zero mathematical weight here; the figure is an intuition-pump, nothing more.

When to use. When the writer wants the reader to reach for a familiar pattern and the argument does not rest on the pattern. Tier 3 is legitimate when explicitly marked as simile.

Discipline. Mark the figure (“like,” “resembles”). Do not let a tier-3 use be read as tier-2 — the reader should know they are reading analogy, not structural claim.

The Anti-Dumbing-Down Rule

Precision earns the use. Do not drop a tier-2 structural parallel to a tier-3 loose analogy to make a passage more accessible, if the tier-2 use was carrying argumentative weight. A structural parallel that has been softened to simile has lost the claim it was making; it is now doing less work than the argument needs. If tier-2 use is correct, tier-2 use is what the passage must carry. If the accessible-layer reader cannot be asked to hold tier-2, the argument has to do the work of earning tier-2 precision — not retreat from it.

Conversely: do not inflate a tier-3 figure to tier-2 by stripping the “like” and presenting the figure as a structural claim. If the figure cannot be defended at tier-2, it must remain at tier-3 with the simile-markers visible.

Relationship to WS-E Academic Rigour

The tier principle supports WS-E’s academic-rigour audit (see foundations/open_questions.md). WS-E’s rules — specific quote or page reference, distinguish what the source claimed from what the project derives, acknowledge strongest counter-reading, no coat-tails citations, confidence markers applied strictly — all cut across tiers. Tier 1 invocations meet the highest bar on every rule; tier 2 parallels must still cite the formalism’s primary source when the parallel is first drawn and must name the boundary of what the parallel does not carry; tier 3 figures carry no citation load but must be marked as figures.

Summary. Three tiers, each legitimate at its depth. Use the tier the claim needs. Defend what you invoke. Mark what you figure. Do not soften for accessibility; do not inflate for rhetoric.

↑Working Style and Collaboration

User Context

The primary researcher (Lynton) brings:

• Extensive contemplative practice (Dzogchen tradition, regular 10-day silent retreats)
• Background in consciousness-primary philosophy
• Independent research approach (PhD rigour without PhD constraints)
• Strong critical thinking and systematic organisation
• Token-efficiency mindfulness in AI collaboration

Claude’s Role

What researcher needs:

• Critical thinking, not affirmation
• Challenge weak arguments
• Identify logical problems
• Flag overconfidence
• Question hidden assumptions
• Steel-man counter-arguments
• Apply PhD-level scrutiny

What researcher does NOT need:

• Enthusiastic agreement
• Elaborate affirmations
• Building on speculation
• Assuming validity of claims

Collaboration Standards

1. Intellectual honesty over agreeableness

   • Point out weaknesses
   • Acknowledge limitations
   • Identify gaps
   • Reality-check enthusiasm

2. Critical analysis over synthesis

   • Verify before accepting
   • Test before building
   • Question before agreeing

3. PhD standards maintained

   • Every claim examined
   • Evidence assessed
   • Alternative explanations considered
   • Confidence calibrated conservatively

↑Token Management

Efficiency Protocols

1. Create downloadable files for:

   • Long analyses (position forensics, literature reviews)
   • Reference lists and bibliographies
   • Detailed notes readable separately

2. Keep threads clear for:

   • Strategic discussion
   • Decision-making
   • Critical analysis

3. Use project_knowledge_search:

   • Minimise token reproduction
   • Access canonical documents efficiently
   • Verify before reconstructing

4. Check token usage before heavy tasks

↑File Naming and Organisation

Canonical File Naming

Pattern: P{phase}_{element}_{description}.md

Examples:

• P0_BIBLIOGRAPHY_INCLUSION_CRITERIA.md
• P2_CONSTRAINT_AUTHORSHIP_SELFHOOD.md
• P2_METHOD_EMERGENCE_REFLECTIVE_CONSTRAINT.md

Version Control

Two numbers run in parallel, and they answer different questions.

Development version (version: in each document’s frontmatter). Runs continuously from that document’s first draft and never resets. It is what depends_on fields and in-prose cross-reference pins name (“FND-02 v5.4 §2.3”), which is why those pins carry numbers larger than any edition. Because it never resets, nothing that pins it can break.

Edition (edition: / edition_date:, dormant until PUBLICATION_EPOCH is set in site/build.py). The published record, and what a reader cites. Every document publishes at edition 1.0 on the epoch.

Increments. Two tiers, applied to whichever number is moving:

• Scoped text correction → patch increment (ed. 1.0 → 1.0.1; v1.3 → v1.3.1).
• Substantive revision → minor increment (ed. 1.0 → 1.1; v1.3 → v1.4).
• Register or metadata repair (citation precision, version-string syncs, formatting) → no bump. It lands under the standing version with a dated changelog entry; the build identifier already ties a page to its exact source.

Entry granularity. The changelog records the document’s argument, at two depths:

• Full entries for anything that changes what the document asserts: every minor and major bump, and every patch-level correction. A correction’s entry carries its substance — what was wrong, what it now says — because the changelog is the corrections record the Declarations point to, and an entry that is the sole record of its correction is never reduced.

• One brief dated line for no-bump work — register, cadence, formatting, pointer and version-string alignment: “Textual alignment (date, under unlock): no argument content changed.” Several passes may consolidate into one line carrying their dates. The line states that the text moved and when; it does not narrate the editorial process — which instruments ran, what vocabulary was adjusted, or why a style rule exists. That rationale lives once, in the project’s internal method records, not once per document. Generic lines carry no working/ pointers.

• No internal organisational vocabulary in published entries. Audit dimension codes (D1–D4), wave ordinals, and process-pass headings belong to the working instruments — FND-01 §4.7’s reserved-terms table assigns them to working docs — and a published entry carries the clause without its label: D1 load-bearing: becomes a plain lead-in, and a code with nothing under it goes. working/ pointers in full entries are permitted; they are provenance, not narration. Pass ordinals already standing in historical entries remain where they name the work; new entries do not add them. Applied to the P1 bundle’s entries 2026-08-30 (changelog audit §10, decision (i)).

Where the Revision Record Lives

In the document’s Document Status changelog — not at its head. Every canonical document ends with a ## Document Status section carrying a changelog, newest first. That is the single home for what changed and why.

The head of a document carries the argument. It does not carry the argument’s history: a reader meeting the work should meet the work. Development notes were retired from the document heads to the changelogs on 2026-08-22, before the epoch, so that edition 1.0 publishes as the corpus is believed rather than as a transcript of how it got there. Nothing was erased — superseded claims are retained in the changelogs as historical record, and each release’s full text remains retrievable from its source snapshot.

Post-epoch corrections surface additionally as a collapsed disclosure beside the citation block, so a reader checking a citation can see whether the text moved since they cited it. One frontmatter field per correction:

revision_1: "1.0.1 | 2026-09-03 | §4.2's Samkhya attribution corrected."

Rendered as Revisions since publication, newest first, and dormant before the epoch. The disclosure is a pointer, not the record: the full entry still goes in the Document Status changelog.

Propagation. When a version moves, every dependent document’s own depends_on / depends-on field moves in the same session. State-document updates alone leave silent drift in the canonical provenance record.

↑Quality Assurance

Before Finalising Canonical Documents

Checklist:

1. ✅ UK English spelling throughout
2. ✅ Claims properly classified ([Framework]/[Derived]/[Empirical])
3. ✅ Confidence markers on empirical claims
4. ✅ Alternative frameworks engaged
5. ✅ Strongest objections addressed
6. ✅ Citations to primary sources
7. ✅ Logical consistency verified
8. ✅ No ontology smuggling
9. ✅ Scope appropriately limited
10. ✅ PhD examination standard met

Regular Audits

Every major phase transition:

• Consistency check across documents
• Cross-reference verification
• Assumption audit (ontology still explicit?)
• Evidence review (claims still supported?)

↑Document Status

Version: 2.5 Date: 2026-08-30 Status: CANONICAL Tier: P0 — project governance Depends on: FND-01 v2.4 (the register and claim disciplines this document operationalises as workflow)

Changelog

Newest first.

• v2.5 (2026-08-30, under unlock — the vocabulary rule made explicit; changelog audit §10 decision (i), working/CHANGELOG_AUDIT_2026-08-29.md): §Version Control’s entry-granularity rule gains a third bullet: internal organisational vocabulary (audit dimension codes, wave ordinals, process-pass headings) does not appear in published entries — the clause stands without its label, per FND-01 §4.7’s assignment of the codes to working docs. working/ pointers in full entries stated as permitted (provenance, not narration); pass ordinals in historical entries stand. The rule was applied to the P1 bundle’s entries the same day. Occasioned by the audit’s §10 cold read, which found the three artefacts retired on 2026-08-29 were the visible members of this family. Minor bump; dependents’ pins move in the same session.

• The corpus state this document describes, measured and made true (2026-08-29, under unlock — under the standing v2.4, no bump; working/CHANGELOG_AUDIT_2026-08-29.md items 2–6): the v2.4 entry below claimed its entry-granularity rule had been “applied retroactively to the P1 bundle’s Document Status sections in the same session.” The audit read that as false and counted four documents consolidated against thirteen untouched. Measured against the tree rather than inferred, both statements needed correcting, in opposite directions. The retroactive run did happen and it reached all seventeen documents: each carries one consolidated register-class entry, with correction-class, structural-removal and citation-apparatus entries kept whole. What only four documents had received was the deeper treatment FND-05 got on 2026-08-29 — a run of same-date no-bump entries archived verbatim and reduced to one — and that treatment cannot be extended, because no other document carries such a run. Outside FND-05’s seventeen entries of 2026-08-23, every remaining no-bump entry in the bundle is a correction, a structural removal or a citation retrofit, bar two short pin-sync lines; under the sole-record protection none of them may be reduced, and the protection is not a preference. What the v2.4 sentence did overclaim is its scope: applied to the Document Status sections is what a reader takes from it, and the sections themselves had been left in five incompatible forms. Measured at the start of this session: three documents carried no such section at all, FND-01 among them; Newest first was declared in four of seventeen; twelve entries stood outside the changelog they belonged to, seven of them on ARG-02, so that its list ran from v1.7 straight to v1.4; four documents held entries out of date order; and five carried no YAML frontmatter, which is why a scan of the corpus’s metadata reported them unversioned. That is what this session closed, along with the three repairs the rule does reach: FND-03’s W-08 tick-list, and dead mobile_intake/ pointers on ARG-02 and ARG-05. The bullet below is scoped to what it did.

• Document Status made a section, and the history ordered (2026-08-29, under unlock — under the standing v2.4, no bump; working/CHANGELOG_AUDIT_2026-08-29.md items 2–3): the revision record sat under a ## Modification History heading with no Document Status section and no metadata block, and ran oldest-first with v2.4 printed above v2.3. It is now a ## Document Status section with a ### Changelog reading newest first — the arrangement this document’s own §”Where the Revision Record Lives” requires of every other document in the corpus. Entries unchanged apart from their order. No standard changed.

• v2.4 (2026-08-26): §Version Control — entry-granularity rule added (Lynton’s decision, 2026-08-26, from the external-review response; working/EXTERNAL_REVIEW_RESPONSE_PLAN_2026-08-26.md Phase 1)

• Full changelog entries reserved for changes to what a document asserts; no-bump textual work takes one brief dated line without process narration

• The sole-record protection restated: an entry that is the only record of its correction is never reduced

• Applied retroactively the same session to the register-class entries of the P1 bundle’s Document Status sections, correction-class entries kept whole (scope corrected 2026-08-29 — see the entry above: what ran was the consolidation of no-bump textual entries, not a pass over the sections themselves)

• Version-history sweep (2026-08-22): §”Claim Classification“‘s [Empirical] entry no longer narrates its own corrected example. The rule it illustrated — reports of experiencing are never [Empirical] data for consciousness in either direction — is stated in the sentence before it and stands unchanged; what goes is the account of what the entry said through v2.1, along with a closing epigram that was a cadence tell under the 2026-08-18 discipline. Part of the pre-epoch corpus sweep (working/VERSION_HISTORY_SWEEP_2026-08-22.md). No version bump — register repair under the standing v2.3.

• v2.3 (2026-08-22): §Version Control rewritten for the edition scheme, and new §”Where the Revision Record Lives”

• Development version and edition documented as the two parallel numbers, with the two-tier increment discipline (patch / minor / no-bump) stated

• The changelog named as the single home for the revision record; development notes retired from document heads the same day, pre-epoch

• Post-epoch revision_N: fields and the Revisions since publication disclosure specified

• Propagation discipline for depends_on pins folded in

• v2.2 (2026-08-20): YAML frontmatter added, bringing this document into the build’s metadata contract with the rest of the corpus

• v2.1 (2026-04-16): Three-Tier Terminology Principle added

  • New section governing physics, mathematics, and technical citations
  • Tier 1 (technical invocation), Tier 2 (structural parallel), Tier 3 (loose analogy)
  • Anti-dumbing-down rule formalised
  • Downstream of FND-01 v2.0 §11.3; complements WS-E academic-rigour requirement

• v2.0 (2026-01-18): Major expansion

  • Added core methodological pattern (ontology-swapping)
  • Established UK English spelling standard
  • Expanded claim classification protocols
  • Added examples across the project investigations
  • Integrated emergence and authorship case studies
  • Enhanced quality assurance protocols

• v1.0 (2025-01-XX): Initial research standards established

End of Research Standards Document

These standards apply across all project work and are binding on canonical output.

@misc{oxoois_standards_ed1_0,
  author       = {Davidson, Lynton},
  title        = {Research Standards},
  year         = {2026},
  version      = {1.0},
  howpublished = {0x00.is},
  url          = {https://0x00.is/library/standards},
  note         = {Written in collaboration with Claude (Anthropic); see https://0x00.is/declarations. Development version 2.5.}
}

TY  - GEN
AU  - Davidson, Lynton
N1  - Written in collaboration with Claude (Anthropic); see https://0x00.is/declarations. Development version 2.5.
TI  - Research Standards
PY  - 2026
ET  - 1.0
PB  - 0x00.is
UR  - https://0x00.is/library/standards
ER  -