The Features and Structure of the lower Noun Phrase

UoM administered thesis: Phd

  • Authors:
  • Christopher Hicks

Abstract

This thesis is an investigation into the behaviour of the features that exist, at least in some languages, in the lower reaches of the nominal projection and how they are structured. The features and structure of the lower noun phrase show a remarkable amount of cross-linguistic variation, despite the consistency in what types of information this level of the noun phrase must contribute to the full phrase and the wider grammar. Thus far, the complex and widespread variation in this domain has resulted in analyses which struggle to account for the variability in interpretation and in morphosyntactic marking under a single model, whilst also being something a language-learning child could acquire. This thesis aims to address some of the issues facing the features and structure of the lower noun phrase by taking a closer look at how the semantics, morphology, syntax and phonology interact, thus proposing an approach which considers, at its core, the nature of the acquisition of the grammar at the syntax-semantics interface. Various features and phenomena are discussed, including the count/mass distinction in English and Mandarin Chinese, classifiers and measure words in Mandarin Chinese, gender in Spanish and German, and a construction in Dutch and German which suggests a null noun. For each of these case studies for my approach, I discuss the cross-linguistic variation at the levels of interpretation and marking, and consider the effect this has on the acquisition of any features within the syntax. In all cases, the variation can be attributed to differences in the feature specification of the noun, how the features of the noun shape its denotation, and how the denotation determines and interacts with the functions that are applied during the structure-building process. We find examples of features that are driven by a combination of meaning and marking, some that are primarily driven by meaning, and some that are primarily driven by marking, suggesting that the acquirer is sensitive to systematicity in the input on more than one level. I conclude by arguing that the proposed approach provides the starting point for a model that has the potential to be powerful for a number of reasons. First, the model is functionally-driven. Nothing can exist within the grammar without a full functional motivation. This greatly constrains the system and aims to prevent any theory-motivated devices. Features within the grammar are not universal, they are created by the acquirer based on evidence within the input. This evidence can take various forms. A grammar based on the input will preserve those features that are transparently marked, but features with weaker marking may be candidates for loss or reanalysis. Finally, the model can account for the cross-linguistic variation we observed in the case studies and, in some cases, the differences in acquisition could be attributed to the strength of evidence for the acquirer and its effect on how a feature is created. There is a wide range of variation that still needs to be explored, but the potential of functionally-driven emergent features should be evident.

Details

Original languageEnglish
Awarding Institution
Supervisors/Advisors
Award date1 Aug 2020