Grounding Meaning: The Symbol Grounding Problem and the Geometry of Understanding

Join us on this Deep Dive as we unpack the symbol grounding problem—how AI can move from manipulating tokens to attaching real meaning to words. We trace the problem from Harnad's iconic three-stage grounding (iconization, categorization, labeling) through early critiques and embodied/intrinsic approaches, to the middle-ground of Peter Gardenfors’ conceptual spaces. We'll see how a three-layer model (sub-symbolic perception, grounded conceptual spaces, and symbolic language) might finally link perception and meaning, with concrete examples like color space and a learn-by-guess grounded language game.

Note: This podcast was AI-generated, and sometimes AI can make mistakes. Please double-check any critical information.

Sponsored by Embersilk LLC