How Do We Ground ‘Grounding’? Why the Vector Grounding Problem Remains Unsolved
How Do We Ground ‘Grounding’? Why the Vector Grounding Problem Remains Unsolved

Summary/Abstract: On the view I develop, grounding is a phenomenological achievement of organisms structured, concern laden encounter with the world, observable in basic minds that connect to their environment without representational content. Language coordinates with this connection; it does not create nor implement it. I develop these views within the Pattern Recognition Unity (framework,distinguishing experiential Pattern Constellations from Linguistic Pattern Constellations and exploring how they coordinate. Contra Mollo and Milliere’s (view, I further argue that teleosemantics a theory that earned its credibility on pre-verbal organisms selected against grounded functions like feeding and fleeing loses its explanatory force when applied to systems selected against linguistic adequacy. The selection history of LLMs borrows its world connection from human raters rather than constituting its own; its causal relation to worldly features is inherited ancestry, not the proximal coupling teleosemantics describes in the natural world. It follows that no evaluation conducted within the linguistic register can distinguish genuine grounding from sufficiently powerful linguistic mastery.

• Details
• Contents