EPAM-like Models of Recognition and Learning

A description is provided of EPAM-III. a theory in the form of a computer progrom for simulating humon verbal leorning, olong with o summary of the empiricol evidence for its volidity. Criticisms leveled ogoinst the theory in o recent paper by Borsolou ond Bower ore shown to derive largely from their misconception that EPAM-III employed a binary. rother than n-ory branching discrimination net. It is shown thot Borsalou and Bower olsa foiled to understond how the recursive structure of EPA&III eliminates the need to duplicate test nodes thot are used to recognize subobjects, ond how the possibility of redundont recognition paths controls the sensitivity of EPAM to noticing order. EPAM is also compored briefly with other theories of humon discrimination and discrimination learning, including PANDEMONIUM-like systems and dotaflow nets.