Abstract

Learning adjectives is considered to be a rather difficult cognitive task. In this paper we present the results of a simulation of adjective acquisition through a biologically realistic neural network. We begin by discussing theoretical explanations on why learning adjectives appears to be so difficult. Then we present the model and the experiment we conducted on adjective acquisition. Lastly, we propose some general considerations on what exactly our model is able to capture: Is it merely perceptual categorization, or rather, a form of conceptualization? The answer, we propose, depends on the sense we give to the term "conceptualization".