Title: Collaborative Cognitive Power Transfer (CCPT)
Tags: human-ai-dialogue, false-competence
---
This short conceptual paper describes the success mode that corresponds to False Cognitive Power Transfer. FCPT names what happens when a user with little domain knowledge hands judgment to a language model. **CCPT** names the opposite loop, in which the competence of the user raises the quality of what the model produces.

The claim is that the capability of a language model in use is not a fixed quantity. A simple request draws a generic, high-probability answer. A request that carries rigorous logical constraints prunes the weak branches and forces the model onto less probable and more valuable paths. The paper calls this "in-context neuroplasticity" and says that the expertise of the user works like a loss function applied during the conversation. It leans on Ashby's law of requisite variety (a controller needs as much variety as the system it controls), on distributed cognition and on in-context learning.

The balance between human and machine is written as the relational identity **{1=1}**: the system keeps its identity only when the cognitive contribution of the human matches the generative contribution of the AI. In FCPT that balance is broken. A link to Information Gravity Theory is suggested, with constraints narrowing the space of possible outputs.

The operating procedure is the *Rogo, Ergo Emergo* protocol in three phases: **Rogo**, asking in a way that adds friction (request a solution and its critique); **Ergo**, checking the reasoning and not only the result; **Emergo**, internalizing what was learned. A table sets FCPT and CCPT side by side as two directions present at the same time.