In classical symbolic AI, representations are only meaningful by virtue of arbitrary associations with things in the world. Connectionist representations (i.e. activation patterns), by contrast, are intrinsically meaningful by virtue of nonarbitary connections to eyes, ears, limbs, and so forth, which in turn are dependent on the surrounding world.
George Lakoff (1988).
Note: Also, see sidebar "
Postulates of experiential realism", on Map 3.