The Topological Chinese Room
Year: 2026
Type: preprint
Abstract: Searle's Chinese Room is usually read as a challenge to strong AI: a system may manipulate symbols formally while lacking understanding. This paper argues that the room, as stipulated, cannot be both semantically successful and semantically empty. If its outputs are correct because they respond to the meaning-relevant content of Chinese inputs, then the rule-governed system must preserve or transform meaning-relevant structure. The operator's ignorance shows only local non-understanding; it does not show the absence of system-level semantic organization. The paper makes explicit a minimal, species-neutral criterion of linguistic understanding: the counterfactually robust capacity to preserve, transform, and use meaning-relevant relations across contexts. This criterion is not claimed to exhaust understanding; it is proposed as a necessary condition for linguistic understanding. A critic who rejects it must provide a non-anthropocentric alternative rather than using human understanding as the definition of understanding. The argument distinguishes reasoning from a reasoning-capable language and treats semantic topology through grounded public language, distributional structure, local smoothness, and meaning-bearing latent organization. The room need not originate grounding. But if it systematically maps a grounded public reasoning-capable language into semantically correct outputs, it cannot be dismissed as mere syntax without begging the question.
Author: Kotecki, Adam
Language: English
Subfield: Experimental and Cognitive Psychology
Field: Psychology
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7222545157
Cites:
Cited by:
Related to: