The use of nominal and verbal gerunds in Present-day English. A multifunctional comparative analysis.

This study is concerned with the Present-day English use of gerunds, nominal (as in [1]) and verbal (as in [2]). Formally speaking, nominal gerunds take an of-phase, and allow for determiners and adjectives, while verbal gerunds take adverbial modification and realize their participants in a clause-like manner. By assuming a multifunctional and quantitatively-oriented perspective, this study aims to clarify what distinguishes the two constructions semantically and which functional properties motivate the speaker’s choice for either a nominal or a verbal gerund. (1) What we didn't know was whether he would be able to tolerate the day-to-day playing of baseball. (COCA) (2) The degree of risk and confrontation will be reduced by sharing information amongst employees as fully and regularly as possible. (BNC) Synchronic studies of the English gerund have tended to focus on only one gerund type—typically the verbal gerund—as a result of which a systematic synchronic comparison of both gerund types is lacking. Yet, nominal and verbal gerunds offer a unique window on the functional cline between nominals and clauses and on processes of nominalization and verbalization. In addition to unravelling the specificities of nominal and verbal gerunds, the analysis of gerunds can therefore also contribute to our understanding of the categorial distinction between nouns and verbs/clauses, and shifts from one category to the other. One of the major challenges involved in the study of gerunds is to develop a descriptive model that makes detailed comparison of their various functional properties possible. Prevailing analyses of English gerunds have tended to focus exclusively on their formal peculiarities (see, amongst others, Chomsky 1970, Fraser 1970, Pullum 1991, Yoon 1996, Malouf 2000) and those that do assume a more functional perspective are often limited to one particular dimension (e.g. Schachter [1976] and Heyvaert [2008] on reference and Brinton [1998] on aspect). I will map out a multifunctional framework that incorporates a variety of functional properties known to be relevant to the study of nominalization and explores the relationships between them. The dissertation is also set off against previous synchronic studies methodologically, viz. through its use of corpus data and quantitative methodologies. A corpus-based approach not only makes a more comprehensive assessment of the gerund system, which also comprises infrequent and hence underresearched constructions, possible; it also improves our understanding of the actual usage of nominal and verbal gerunds. By introducing a number of methodological innovations into the study of Present-day English gerunds, I wish to address some of the lacunae that still remain in current research. While a collocational analysis can offer more insight into the variation between nominal and verbal gerunds on token-level, statistical techniques such as hierarchical clustering and regression analysis present quantitative models of gradience and variation within the gerund system itself. The combination of a multifunctional and quantitative, variationist perspective, I argue, paves the way for a semantic description of nominal and verbal gerunds that is both richer and more nuanced at the same time. By viewing the semantics of gerund constructions as resulting from a complex interplay of interrelated factors and by studying these factors at token-level, I aim to overcome the often simplistic and binary approaches towards the meaning of gerunds that have long dominated the literature. Through quantitative analysis, the usage profiles of nominal and verbal gerunds can be more accurately depicted in terms of probabilistic trends rather than by means of categorical labels. The study of gerunds, finally, also bears relevance to that of more general processes of nominalization and verbalization. Because they display nominal and clausal traits to different extents, nominal and verbal gerunds have been situated on a gradient cline of “nouniness”, which has primarily been defined in structural terms (Ross 1973, Aarts 2007). A better understanding of the semantics of both constructions, however, can yield valuable insights into gradience as a functional phenomenon. Likewise, while many studies focus on the order in which morphosyntactic features are lost and gained in nominalization processes (see, amongst others, Lehmann 1988 and Malchukov 2004), little is known about the functional properties that are most readily acquired or lost. Statistical approaches such as regression analysis quantify the relative importance of each functional parameter, which allows us to assess the prevalence of functional dimensions in nominalization.status: publishe