Volume 9, Issue 2

This paper examines parallels and differences between South Asian Englishes and British English with regard to various factors driving the selection of verb-complementation patterns. Focusing on the prototypical ditransitive verb give and its complementation, we use large web-derived corpora and distinguish between two possible response cases, one based on the dative and prepositional construction (i.e. the dative alternation), the other including monotransitive complementation. Our data has been additionally coded for a number of potential driving factors, such as pronominality and discourse accessibility of the participants in the constructions. Applying a model-exploration technique we isolate the main driving factors for the varieties under scrutiny (Indian English, Pakistani English and British English) and analyze their influence on pattern selection based on a multinomial logistic regression formulation. Our findings show that, while there is a large area of overlap between the varieties, Pakistani English is closer to British English with regard to relevant driving factors than Indian English. Furthermore, we reveal interesting parallels between all three varieties in the use of monotransitive complementation.

In existing research on syntactic alternations such as the dative alternation, ( give her the apple vs. give the apple to her ), the linguistic data is often analysed with the help of logistic regression models. In this article, we evaluate the use of logistic regression for this type of research, and present two different approaches: Bayesian Networks and Memory-based learning. For the Bayesian Network, we use the higher-level semantic features suggested in the literature, while we limit ourselves to lexical items in the memory-based approach. We evaluate the suitability of the three approaches by applying them to a large data set (>11,000 instances) extracted from the British National Corpus, and comparing their quality in terms of classification accuracy, their interpretability in the context of linguistic research, and their actual classification of individual cases. Our main finding is that the classifications are very similar across the three approaches, also when employing lexical items instead of the higher-level features, because most of the alternation is determined by the verb and the length of the two objects (here: her and the apple ).

This paper is concerned with the ability of lexical semantic classes to predict the behavior of argument structures in productively selecting novel lexical material. It will be argued that while lexical semantic classes can explain a great deal about the spectrum of potential arguments and resulting entailments for those arguments, constructions which are synonymous from a formal semantic perspective exhibit significant differences in their empirically observed propensity for argument realization. Using corpus data and methods from Baayen's morphological productivity paradigm, I will show that such differences are not restricted to lexicalized, collocational preferences, but also extend to the readiness with which different constructions admit novel arguments.