The Role of Relatedness in Word Meanings in Typical and Atypical Populations

Abstract

Many words in English and other languages are polysemous, with multiple distinct, but related meanings. Previous research has shown that learners have a polysemy over ambiguity advantage in which related meanings of words are easier to learn than unrelated meanings of words (polysemy vs. ambiguity). This is true for neurotypical children, but not necessarily for children with Autism Spectrum Disorder (ASD). The current work aims to understand the underlying reason for this difficulty in learning polysemous words. I hypothesize that it is due to a hyper-focus on specifics: individuals on the Autism spectrum are more likely to focus on differences in individual exemplars, rather than generalizing instances. The question is highly relevant, given the known delays in language learning in the ASD population. In Experiment 1, we tested adults’ and children’s’ ability to notice the abstract relationships between symbol and image pairings. We hypothesized that performance on this experiment and how well participants are able to recall symbol-image pairings will correlate with performance on Experiment 2, because the ability to learn polysemous words is grounded in recognizing semantic relatedness rather than language difficulty. In Experiment 2, we explicitly named the similarities across the polysemous meanings. We hypothesized that this strategy will increase children’s performance from pre- to post-intervention testing. However, results showed that for neurotypical children, there was about equal learning between the control and intervention condition groups. We discuss possible explanations, namely that neurotypical children may be able to spontaneously observe similarities between words and generalize among them, resulting in near ceiling performance regardless of the intervention. Hypothetical results are also discussed.

Keywords: polysemy, Autism spectrum disorder, neurotypical, generalization, word learning