Cognitive Memoisation: A Framework for Human Cognition

Author: Ralph B Holland Date: 2025-12-23

Introductory Position

This paper serves as the primary introduction and conceptual anchor for the Cognitive Memoisation (CM) corpus.

Cognitive Memoisation is a human-governed knowledge-engineering pattern that preserves conceptual memory across interactions with stateless Large Language Models (LLMs). CM operates entirely outside model-internal memory, respecting architectural safety constraints while enabling humans to avoid repeated rediscovery (“Groundhog Day”) and to carry forward both resolved knowledge and deliberately unresolved cognition.

This document establishes the rationale, scope, and interpretive frame required to correctly understand the remainder of the CM papers.

Normative CM Paper References

The following documents constitute the authoritative CM corpus. Titles are normative MediaWiki page names and must not be paraphrased.

Dimensions Addressed in This Paper

The following dimensions are key to understanding the problems that Cognitive Memoisation (CM) addresses, especially in the context of stateless Large Language Models (LLMs) and the human-managed preservation of cognitive state:

1. Statelessness and Memory Management in LLMs

Core Concept: Addressing the statelessness of LLMs and the challenge of managing conceptual memory externally. Dimension Addressed: How can cognitive memory be maintained outside the LLM model to overcome statelessness, and how does CM provide this functionality while respecting LLM safety constraints? Relevant Papers:

2. Externalisation of Cognitive Artefacts

Core Concept: The process of externalising concepts, facts, inferences, and unresolved cognition into structured, durable formats. Dimension Addressed: How can cognitive content be externalised and stored in a manner that ensures its continued use across sessions, without being lost due to session termination or model limitations? Relevant Papers:

3. Round-Trip Knowledge Engineering (RTKE)

Core Concept: The cyclical process of taking externalised cognitive artefacts, reintegrating them into reasoning processes, and ensuring that knowledge evolves without loss. Dimension Addressed: How can externalised knowledge be reused, refined, and preserved over time through iterative processes, and how does CM facilitate this while maintaining consistency? Relevant Papers:

4. Dangling Cognates and Unresolved Cognition

Core Concept: Managing cognitive elements that are under construction or incomplete, allowing them to participate in reasoning without forcing premature resolution. Dimension Addressed: How can unresolved cognitive elements (Dangling Cognates) be preserved, tracked, and used safely in ongoing reasoning, without prematurely solidifying them? Relevant Papers:

5. Constraints and Knowledge Integrity

Core Concept: Defining and applying constraints to preserve the integrity of cognitive memory and prevent “Groundhog Day” rediscovery. Dimension Addressed: How can constraints be implemented to ensure that knowledge persists across sessions without redundancy, and how can it be efficiently reused? Relevant Papers:

6. Human Curated Knowledge vs. Model State

Core Concept: Differentiating between human-curated knowledge and LLM model state, ensuring that cognitive memory and decision-making remain under human control. Dimension Addressed: How can the human maintain full authority over cognitive content while ensuring that the stateless nature of LLMs is respected? Relevant Papers:

Role of This Paper in the Corpus

This paper is authoritative for the following assertions:

Cognitive Memoisation preserves conceptual-level memory, not dialogue or model state.
CM exists specifically because LLMs are stateless by architecture.
CM externalises cognition into durable artefacts to enable Round-Trip Knowledge Engineering.
CM supports the deliberate carriage of unresolved cognition (Dangling Cognates).
CM places authority, curation, and provenance entirely with the human.

All listed papers assume this framing and should be interpreted in relation to it.

Main Paper

The primary paper for understanding the full scope and technical details of Cognitive Memoisation is:

Cognitive Memoisation and LLMs: A Method for Exploratory Modelling Before Formalisation

If you want next, I can:

Emit this as a strict MWDUMP-compliant page
Add a categories block aligned to your wiki taxonomy
Produce a one-paragraph abstract for publication landing pages
Design a front-matter template reused across all CM papers
Cross-check internal consistency of terminology across the listed documents