International Journal of Corpus Linguistics - Volume 23, Issue 1, 2018

The aim of this paper is to identify the effect of register variation in spoken British English on the occurrence of the four principal verb-forming suffixes: ‑ate, ‑en, ‑ify and ‑ize, by building on the work of Biber et al. (1999) , Plag et al. (1999) and Schmid (2011) . Register variation effects were compared between the less formal Demographically-Sampled and the more formal Context-Governed components of the original 1994 version of the British National Corpus. The pattern of ‑ize derivatives revealed the most marked register-based differences with respect to frequency counts and the creation of neologisms, whereas ‑en derivatives varied the least compared with the other three suffixes. Quantitative and qualitative analyses of these suffix profiles in the context of spoken language reveal markers of register formality that have not hitherto been explored; derivative usage patterns provide an additional dimension to previous research on register variation which has mainly focused on grammatical and lexical features of language.

Current syntactic annotation of large-scale learner corpora mainly resorts to “standard parsers” trained on native language data. Understanding how these parsers perform on learner data is important for downstream research and application related to learner language. This study evaluates the performance of multiple standard probabilistic parsers on learner English. Our contributions are three-fold. Firstly, we demonstrate that the common practice of constructing a gold standard – by manually correcting the pre-annotation of a single parser – can introduce bias to parser evaluation. We propose an alternative annotation method which can control for the annotation bias. Secondly, we quantify the influence of learner errors on parsing errors, and identify the learner errors that impact on parsing most. Finally, we compare the performance of the parsers on learner English and native English. Our results have useful implications on how to select a standard parser for learner English.

This paper investigates grammatical variation in the complementation of the verb provide. It describes the distribution of the four possible patterns with two internal arguments and the interaction between pattern choice and lexical choice. The study finds and documents significant differences in the preferred complementation patterns for American and British English as well as for spoken and written news genres. It also establishes the double object construction as a viable option for American English. Methodologically, this study is based on robust automatic syntactic annotation and computerized retrieval from a data-set comprising 2.5 billion words. It is this large amount of data that permits the observation of strong preferences in terms of pattern choice at the interface between grammar and lexis.

This paper studies the relationship between grammar and language use by comparing word association and collocation. Since word association reveals mental semantic knowledge, usage-based approaches expect word association to mirror the relation between words in use, namely collocation. The paragraph is a more apt unit for collocation than the sentence in mirroring word association. Among measures of collocation, (simple) log likelihood and t-score turn out to be more consistent with association, with log likelihood leading by a small margin over MI or MI3. Overall, word association and collocation are quite close, but not perfectly close because of differences in relevant resources and the characteristics of lexical/semantic relations.