Systematicity, motivatedness, and the structure of the lexicon

Abstract

For the majority of the 20th century, one of the central dogmas of linguistics was that, at the level of the lexicon, the relationship between words and meanings is arbitrary: there is nothing about the word ‘dog’ for example that makes it a particularly good label for a dog. However, in recent years it has become increasingly recognized that non-arbitrary associations between words and meanings make up a small, but potentially important portion of the lexicon. This thesis focuses on exploring the effect that non-arbitrary associations between words and meanings have on language learning and the structure of the lexicon. Based on a critical analysis of the existing literature, and the results of a number of experiments presented here, I suggest that the overall prevalence and developmental timing of two forms of non-arbitrariness in the lexicon– systematicity and motivatedness – is shaped by the pressure for languages to be learnable while remaining expressive. The effect of pressures for learnability and expressivity have been recognized to have important implications for the structure of language generally, but have so far not been applied to explain structure at the level of the lexicon. The central claim presented in this dissertation is that features of the perceptual and cognitive organization of humans results in specific types of associations between words and meanings being easier for naïve learners to acquire than others, and that the pressure for languages to be learnable results in lexica that leverage these human biases. Taking advantage of these biases, however, induces constraints on the structure of the lexicon that, left unchecked, might limit its expressivity or penalize subsequent learning. Thus, lexica are structured such that early-acquired words are able to leverage these biases while avoiding the limitations imposed by those biases when they are extended past a certain point.