Preview

PDF, English - main document Download (6MB) | Lizenz: Rights reserved - Free Access

Abstract

This doctoral project identifies formulaic sequences (hereinafter FS and the plural form FSs) in Early Modern English (hereinafter EModE) and intends to investigate the functions they serve in communication and different text types, namely EModE dialogues and letters. Main contributions of the study include, firstly, the study provides solid arguments and further evidence that FSs are constructions in the Construction Grammar instead of exceptions in the traditional grammar-dictionary model. Within this theoreticall framework, I proposed a new working definition of FSs that is inclusive, descriptive, and methodologically neutral. The study also argues that there are fundamental differences between FSs and lexical bundles (LBs), although the latter often treated as an alternative term of FSs or sub-groups of FSs. Nevertheless, after a thorogh review of the characteristics of the two mult-word units, the study argues that despite of the differences, LBs can be upgrated to FSs as long as they fulfill certail sematic, syntactic, and pragmatic criteria. THis forms the fundation of the methodology design of the study. Secondly, the study enhanced the corpus-assisted approach to the identification of FSs, esp. in EModE texts. The approach consists of three steps: preparation, identification, and generalisation. The identification step was further conducted within two phases: automatic generation of LBs for a corpus and manual identification of FSs from LBs. Specifically, in the preparation step, the dissertation critically discussed how spelling variation in EModE texts shall be dealt with in investigations on FSs. I designed a series of criteria for the two-phase identification of FSs. For one thing, I disagree with previous research that two-word LBs shall be excluded from examination by arguing that many of them are formulaic and cannot be captured from longer LBs and the workload of processing the massive number of two-word LBs is actually manageable. For another, the study contributes an easy-to-follow flow chart demonstrating the procedure of the manual identification of FSs from LBs and listing the criteria that guide the decision-making process. Thirdly, the study provides systematic and comprehensive accounts of FSs in EModE dialogues and letters, esp. how their forms are conventionally mapped to their functions. Data analysis were conducted from aspects such as degree of fixedness, grammatical structures, distribution across function categories, multi-functional FSs, genre-specific FSs, etc. General findings suggest that EModE dialogues and letters actually have many similarities regarding the form and function of FSs and general trends of distribution across function categories. However, outstanding differences between the two text types can be observed too. From the perspective of form, the distinction lies in word choice in realisations of certain FSs. From the perspective of meaning/function, the distinction lies in the kinds of functions that need FSs the most or the least and common function combinations. More importantly, the study observed two types of relationships among FSs themselves and the discourse, including horizonal networks and vertical networks, which reflects the complexity of FSs and their identity as constructions. Specifically, three types of horizontal networks of FSs are embedding, attaching, and joining. A pair of new concepts is proposed to describe the vertical networks: superordinate FSs and subordinate FSs. As a result of the vertical networks, three types of functional diviation are observed: function extension, shifting, and specification.