metadata

Provenance note

The concepts described here — including Cognitive Memoisation (CM) and Dangling Cognates (DC) — were originally developed and authored by the site owner.

This document represents an AI-assisted redaction and external-facing re-articulation of those ideas as a test generation from the main CM-artefact used to guide the model through the CM-framework. Tbis document was generated without access to the original paper text on CM and DC, and subsequently curated for conceptual fidelity.

An experimental probe into the clarity of the CM and DC concept definition that are held within the CM-master file uploaded as input stimulation and guidence of the LLM, and for continued human conversations across a session.

Curator:	Ralph Holland
Email:	ralph.b.holland at gmail.com
Author:	ChatGPT-5 platform
Version:	1.1
Publication Date :	2025-12-28T06:55Z

Metadata (Normative)

The metadata table immediately preceding this section is CM-defined and constitutes the authoritative provenance record for this MWDUMP artefact.

All fields in that table (including artefact, author, version, date, local timezone, and reason) MUST be treated as normative metadata.

The assisting system MUST NOT infer, normalise, reinterpret, duplicate, or rewrite these fields. If any field is missing, unclear, or later superseded, the change MUST be made explicitly by the human and recorded via version update, not inferred.

Curator Provenance and Licensing Notice

This document predates its open licensing.
As curator and author, I apply the Apache License, Version 2.0, at publication to permit reuse and implementation while preventing enclosure or patent capture. This licensing action does not revise, reinterpret, or supersede any normative content herein.
Authority remains explicitly human; no implementation, system, or platform may assert epistemic authority by virtue of this license.

Dangling Cognates: Preserving Unresolved Knowledge in Cognitive Memoisation

Abstract

Long-running human–machine reasoning systems frequently suffer from epistemic discontinuity: unresolved concepts, questions, and dependencies are lost at session boundaries, leading to repeated rediscovery, implicit assumptions, and false convergence. This paper introduces the notion of Dangling Cognates (CM), a first-class construct within the Cognitive Memoisation (CM) framework, designed to explicitly capture and preserve unresolved conceptual artefacts. By treating acknowledged incompleteness as a durable and governable state, Dangling Cognates provide an anti-amnesia mechanism that enables cumulative reasoning across sessions without forcing premature resolution.

1. Introduction

Interactive reasoning systems—particularly those involving large language models—are increasingly used in extended, multi-session knowledge work. Despite their apparent fluency, such systems exhibit a structural weakness: they struggle to preserve where reasoning stopped. Open questions, undefined terms, and partially explored ideas are often reintroduced in later sessions as if they were new, or worse, silently assumed to have been resolved.

This phenomenon leads to what practitioners frequently experience as “Groundhog Day” reasoning: the same conceptual terrain is revisited repeatedly, not because progress is impossible, but because unresolved artefacts were never made durable.

The Cognitive Memoisation (CM) framework addresses this problem by formalising how knowledge, state, and uncertainty are captured, curated, and rehydrated across sessions. Within CM, a central construct for managing unresolved material is the Dangling Cognate.

2. The Problem of Unresolved Concepts

In conventional documentation systems, unresolved ideas are often handled informally:

assumptions remain implicit,
questions are left unanswered in transcripts,
dependencies are referenced but not defined,
partial reasoning trails are abandoned at session end.

While such artefacts are cognitively salient in the moment, they lack persistence. When a new session begins, participants reconstruct context heuristically, often filling gaps differently than before. This creates drift, duplication of effort, and, in some cases, the illusion of progress without genuine convergence.

Crucially, the failure is not that these concepts are unresolved, but that their unresolved status is unrecorded.

3. Dangling Cognates: Definition and Rationale

A Dangling Cognate is a named and explicitly captured conceptual artefact that has been introduced into discourse but has not yet been fully defined, resolved, or bound to an authoritative state.

The term “dangling” denotes acknowledged incompleteness rather than error. A Dangling Cognate exists precisely to say: this matters, but we are not finished with it.

Unlike assumptions, Dangling Cognates are explicit. Unlike questions, they persist as identifiable artefacts. Unlike hypotheses, they are not required to be testable or fully formed. They occupy a distinct epistemic category: unresolved, yet durable.

The rationale for Dangling Cognates is simple but powerful: unresolved concepts should not disappear merely because a session ends.

4. Lifecycle and Governance

Dangling Cognates are not transient placeholders. Once captured, they persist across sessions and are subject to explicit governance. Over time, a Dangling Cognate may follow one of several paths:

Refinement, where additional structure or partial understanding is added without closure.
Resolution, where the concept is fully defined and promoted to a grounded, authoritative concept.
Abandonment, where the artefact is explicitly retired as irrelevant, incorrect, or superseded.

Importantly, none of these transitions are implicit. In CM, silence is not a valid state transition. If a concept remains unresolved, it remains explicitly marked as such.

This governance model prevents unresolved artefacts from being accidentally treated as settled knowledge through repetition or familiarity.

5. Epistemic Continuity Across Sessions

The primary contribution of Dangling Cognates is the preservation of epistemic continuity. When a new session begins, previously captured Dangling Cognates can be rehydrated alongside grounded knowledge. This allows reasoning to resume from a precise and faithful representation of prior state, including what was not yet known.

As a result, progress becomes cumulative rather than repetitive. The system no longer needs to rediscover unanswered questions or re-identify missing definitions; these are already present as explicit artefacts awaiting attention.

In this sense, Dangling Cognates form an anti-amnesia layer within CM, ensuring that uncertainty itself is preserved as part of the knowledge state.

6. Implications for Human–Machine Reasoning

Treating unresolved concepts as first-class artefacts has several implications:

Uncertainty becomes manageable rather than accidental.
Premature closure is avoided without sacrificing persistence.
Governance mechanisms gain visibility into epistemic gaps.
Long-running collaborations benefit from reduced drift and redundancy.

More broadly, Dangling Cognates challenge the notion that knowledge systems should only preserve what is known. Instead, they argue that what is not yet known, but known to be missing, is equally important.

7. Conclusion

Dangling Cognates formalise a simple but often neglected idea: not knowing yet is a legitimate and valuable state. By explicitly capturing unresolved conceptual artefacts and carrying them forward intact, Cognitive Memoisation enables reasoning systems to progress without forgetting where they stopped.

In doing so, Dangling Cognates transform uncertainty from a transient by-product of conversation into a durable component of cumulative knowledge.