The effect of varying channel capacities on artificial grammar learning: A comparison between people with and without dyslexia

Abstract

In their language development, children are able to extract rules from the limited input they hear extremely fast. They are able to apply those rules to new strings of words, which they never heard before: a process known as ‘the logical problem of language acquisition’. The entropy model for linguistic generalizations addresses this problem and states that the process of making generalizations is influenced by the input complexity (entropy) and the channel capacity. Whenever the channel capacity is exceeded by the input complexity, the process of making generalizations is thought to increase. In order to learn more about the point at which children start to extract rules from the input, a previous artificial grammar study is used as a model for language acquisition in children (Rădulescu, Wijnen, & Avrutin, 2014). A tendency of making more generalizations when the entropy increased and channel capacity was kept constant, is reported. The present study investigates the effect of group (dyslexics vs. non-dyslexics) on the process of making generalizations, when entropy was kept constant, in order to learn more about the effect of varying channel capacities. The low entropy condition of the original experiment was used in this follow-up study, as dyslexics were expected to make more generalizations than non-dyslexics when confronted with a low amount of entropy. This is due to the assumption that participants with dyslexia are hypothesized to have smaller channel capacities than participants without dyslexia, due to either weaker working memories or problems in procedural learning. Dyslexics’ channel capacities were thus expected to be exceeded by the input and dyslexics were expected to abstract more generalized rules from the input as a consequence. Results indicated a tendency in dyslexics of making more generalizations than non-dyslexics, but no significant differences between groups were found. This is probably due to power issues of the results, as the sample sizes in the present study were relatively small. Therefore, people with dyslexia are still hypothesized to have smaller channel capacities than people without dyslexia.