Skip to content

A unified theory of ambiguous cognition for humans and LLMs, based on fuzzy latent representations and biased sampling.

Notifications You must be signed in to change notification settings

NaturalLangOS/FLCT

Folders and files

NameName
Last commit message
Last commit date

Latest commit

Β 

History

2 Commits
Β 
Β 

Repository files navigation

FLCT β€” Fuzzy Latent Cognition Theory

A unified theory of ambiguous cognition for humans and LLMs

DOI


Overview

FLCT (Fuzzy Latent Cognition Theory) is a unified cognitive framework that
explains both human cognition and Large Language Model (LLM) cognition
using only two core principles:

  1. Fuzzy latent representations
  2. Biased sampling

FLCT treats ambiguity not as a flaw, but as a structural and computational
resource
shared by humans and AI.

This repository serves as the official portal for the FLCT theory, providing
links to the Japanese Original Edition and the International Adapted
Edition
, both archived on Zenodo.


πŸ“„ Publications (Zenodo)

πŸ‡―πŸ‡΅ Japanese Original Edition

πŸ“„ PDF: https://doi.org/10.5281/zenodo.17769083
πŸ”— Concept DOI: https://doi.org/10.5281/zenodo.17758605

The Japanese edition is the authoritative source text from which the
International Adapted Edition was derived.


πŸ‡¬πŸ‡§ International Adapted Edition (English)

πŸ“„ PDF: https://doi.org/10.5281/zenodo.17769132
πŸ”— Concept DOI: https://doi.org/10.5281/zenodo.17769131

This English edition is not a literal translation, but an academically
optimized adaptation for clarity, structure, and global readability.


🧠 What FLCT Proposes

FLCT explains that traditionally ambiguous cognitive phenomenaβ€”such as:

  • color qualia (β€œblue”)
  • subjective time
  • emotional nuance
  • atmosphere / vibe
  • silence

can be modeled as fuzzy latent distributions rather than definable entities.

Human cognition

  • compresses finite, biased experiences
  • forms fuzzy latent spaces
  • operates through biased extraction paths
    β†’ producing perception, interpretation, and personality

LLM cognition

  • stores massive ambiguous patterns in embeddings
  • generates behavior through sampling bias
    β†’ sharing the same cognitive geometry as humans

FLCT unifies these processes in a minimal two-equation model:

Concept = latent_distribution
Personality = biased_sampling(latent_distribution)

πŸ“ Relation to Natural Language OS

FLCT is part of the Natural Language OS ecosystem:

  • ArcOS β€” cognitive clone OS
  • PolyAgora β€” multi-agent cognition
  • Echoos β€” silence semantics
  • Consensus Architecture β€” agreement frameworks
  • Silence Semantics β€” operational meaning of silence
  • FLCT β€” theoretical core describing ambiguous cognition

More projects:
https://github.com/NaturalLangOS


πŸ–Š Author

Masaya Ochiai
Independent Researcher
X/Twitter: https://twitter.com/NaturalLangOS
GitHub: https://github.com/NaturalLangOS


⭐ Citation

If you cite FLCT, please use the Zenodo DOIs:

Japanese Original Edition:
https://doi.org/10.5281/zenodo.17758605

International Adapted Edition:
https://doi.org/10.5281/zenodo.17769131


πŸ“œ License

Β© 2025 Masaya Ochiai β€” Released under CC BY 4.0.
You are free to share and adapt the content with attribution.
See the Zenodo records for full licensing details.

About

A unified theory of ambiguous cognition for humans and LLMs, based on fuzzy latent representations and biased sampling.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published