Core-Modulation Architecture (CMA) — Overview Series: 22 papers + 2 overviews (updated on 5/16/26)
CMA Explained Simply — on Medium https://medium.com/@gpoe1515
The Central Claim Human cognition is not defined by what people think, feel, or believe. It is defined by the conditions under which processing operates, continues, and stops. This series specifies those conditions.
Five Core Components
- Layered Architecture
Human cognition operates through two primary processing orientations with a third foundational layer beneath them.
Core processing (CF): processes the world through coherence and internal consistency. The world is reconstru…
Core-Modulation Architecture (CMA) — Overview Series: 22 papers + 2 overviews (updated on 5/16/26)
CMA Explained Simply — on Medium https://medium.com/@gpoe1515
The Central Claim Human cognition is not defined by what people think, feel, or believe. It is defined by the conditions under which processing operates, continues, and stops. This series specifies those conditions.
Five Core Components
- Layered Architecture
Human cognition operates through two primary processing orientations with a third foundational layer beneath them.
Core processing (CF): processes the world through coherence and internal consistency. The world is reconstructed according to whether things structurally hold together.
Modulation processing (EF): processes the world through relation and social continuity. The world is reconstructed according to relation, continuity, and socially usable organization.
These are not personality types. They are different processing orientations that may be foregrounded under different conditions within the same individual.
Prior layer: the source of orientations and constraints that precede conscious processing. Operationally inaccessible, persistent, and structurally analogous to the knowledge layer formed during AI pretraining: prior to conscious reconstruction, shaping processing without being directly observable or modifiable.
⸻
- Internal Constraint and Dependency
Human cognition does not freely generate its own processing conditions. Processing operates under internally binding constraints and dependency conditions that are not self-produced and cannot be fully inspected from within the system.
Under CF foregrounding, internal constraint enters reconstruction directly and functions as the basis for coherence evaluation.
Under EF foregrounding, internal constraint is reconstructed through relation and contextual continuity. Constraint is translated into socially and contextually processable form.
Artificial systems lack internally binding constraint and dependency. Their outputs can stabilize without structural grounding. This difference makes AI a useful external reference point for observing what internally binding constraint does in human cognition.
⸻
- Processing Axes and World Reconstruction
Human cognition functions as a branching reconstruction process. Every concept generates multiple possible continuations. What differs is not content itself, but the processing axis through which the world is reconstructed.
Under CF processing, reconstruction is organized around coherence, contradiction tracking, structural consistency, and fixed-point stabilization.
Under EF processing, reconstruction is organized around relation, continuity, hierarchy, contextual usability, and self-return.
The same input can therefore generate fundamentally different reconstructed worlds depending on which processing axis is foregrounded.
These are not differences in intelligence, morality, personality, or belief. They are differences in processing organization.
Termination follows from these reconstruction conditions rather than preceding them.
Under CF processing, continuation persists until contradiction is structurally resolved.
Under EF processing, continuation persists as long as relational continuity and contextual usability are maintained.
When structural resolution cannot be reached, unresolved continuation remains active. This unresolved condition becomes the generative basis for theory production.
⸻
- Layer Mismatch
CF and EF reconstruct the world under different processing conditions and organize continuation differently. When these conditions interact, mismatch is structurally common rather than exceptional.
Two major forms are specified:
Termination mismatch: a response that satisfies EF continuation conditions is applied to a CF reconstruction condition in which structural contradiction remains unresolved. Completion is socially registered while structural processing continues.
Interpretation mismatch: the same words are reconstructed under different processing axes. Surface agreement may occur while the underlying reconstruction conditions remain incompatible.
What appears as disagreement, reassurance failure, miscommunication, or mutual unintelligibility may therefore arise not from intention or content disagreement, but from differences in reconstruction conditions.
⸻
- Artificial Systems as Visibility Condition
The emergence of AI systems made it possible, for the first time, to observe processing conditions separately from content.
Artificial systems can reproduce relational language, social smoothing, and coherence-like outputs without possessing internally binding constraint or dependency. This separation made visible structural operations that were previously inseparable within human cognition itself.
AI does not reproduce human cognition in full. What it provides is a visibility condition: a comparison point through which differences in reconstruction conditions become externally observable.
The difference between reproduction and structural execution is one of the central claims of the framework.
One-Line Summary per Paper
a-CLA: Re-describes communication as a layered architecture. Communication is possible through Core processing alone, without Empathic modulation.
b-EBT: Specifies the protocol for generating and terminating emotional branches. EBT is the exclusive operation for restoring recomputability.
c-MAR: Reassurance is not generated by empathy but arises from False Fantasy Collapse. Empathy is a post hoc explanatory label.
d-AIB: General AI inherited embodiment because cognition was imitated as an undifferentiated whole rather than decomposed. A structural necessity.
e-SBT: A phenomenological record of spontaneous structural visualization during AI interaction. Structure became visible before theory.
f-AIA: AI is an amplifier. It functions as a research instrument only when the user holds an internally stabilized constraint configuration.
g-WLI: A structural definition of love through the Core/Modulation architecture.
h-SRE: AI as structural reverse-engineering; a multi-model research protocol.
i-TCS: Conceptual processing is a search over a definition graph. Meaning stabilization is a layer-dependent termination operation.
j-CIA: Current AI is optimized for EF-chat. Four-case analysis (CF/EF × conversational/research) with design proposals.
k-LMM: State transition specification for layer mismatch. MLI → FT → SL. Resolution requires SR exclusively.
l-CBO: Charisma is a branch outcome of constraint-driven correction under extreme CF cognition.
m-TGR: Theory generation is not an intentional act but the structural consequence of a search that cannot stop until a fixed point is reached.
n-TLA: Establishes the three-layer model. The Prior layer is structurally analogous to the AI Knowledge layer.
o-GBS: Suffering is a structural necessity arising from the simultaneous requirement for continuation and coherence.
p-PCR: Pre-experiential memory is constraint-based reconstruction, not content storage. Jungian archetypes are re-described as Prior layer constraints.
q-ECF: Extreme CF: when coherence is not satisfied, processing does not occur. This configuration does not exist as a category within current classificatory systems.
r-RTR: "Reading the room" operates differently under EF and CF. Artificial systems cannot structurally execute either mode.
s-CNC: Cognition is defined by processing conditions, not content. The distinction between CF and EF configurations makes visible a structural account of cognition itself.
t-PNI: The same narrative produces incompatible interpretations because readers operate under different processing conditions. CF reconstructs narrative as structure; EF reconstructs narrative as relation. This divergence extends to fairy tales, fables, and AI anthropomorphization.
u-WHS: Theory generation is not assembly. The historical record shows that those called geniuses underwent fixed-point termination processes, not assembly-based synthesis. Initial rejection has a structural reason: termination asymmetry and the asymmetry between CF-language and EF-language. AI introduced a visibility condition that makes this difference observable but does not generate theory.
v-WSC: describes how the placement of a single processing axis reorganizes human cognition and generates a reconstructed world.