The effects of frequency, distribution, mode of presentation, and first language on learning an artificial language

Abstract

This dissertation presents results from a series of experiments investigating adult learning of an artificial language and the effects that input frequency (high vs. low token frequency), frequency distribution (skewed vs. balanced), presentation mode (structured vs. scrambled), and first language (English vs. Japanese) have on such learning. Motivated by cognitive and usage-based accounts of language and learning, the research aims to contribute to theoretical debates concerning the influence of input properties and existing knowledge in language learning. The artificial language used in the experiments is focused on the learning of noun classes modeled on Nilo-Saharan languages. Two artificial noun classes, each with distinct morphological features, were devised based on a semantic contrast between entities that are typically encountered as individuals and those typically encountered as groups, sets, pairs or masses. A total of 150 subjects, college students and young adult native speakers of English and Japanese with no more than limited knowledge of the other language, participated in these experiments. In each experiment, subjects were exposed to words and pictures representing the two noun classes. The learning phase was followed by a testing phase to assess their learning with respect to both trained and previously unseen exemplars of each class. A two-factor factorial analysis of variance design was used to analyze the results. The results show that presentation mode had the largest effect on learning, followed by token frequency and frequency distribution. The results also show a constant effect of L1 knowledge: participants were better at learning morphological features similar to their L1 than dissimilar ones. These findings contribute new knowledge to our understanding of the learning of functional morphology--which has been viewed as a major theoretical challenge by researchers working within such diverse perspectives as the processing instructional paradigm and generative SLA--and leads to pedagogical implications that may benefit learners.