Communicative efficiency and probabilistic grammar: Bayesian mixed-effect regression models of help + (to) Infinitive in varieties of web-based English

This study focuses on help followed by the bare or to-infinitive in seven varieties of web-based English from Australia, Ghana, Great Britain, Hong Kong, India, Jamaica and the USA. It investigates the role of information content, which depends on the predictability of the infinitive given help and the other way round, in the user’s choice between the constructional variants, in addition to various other factors known from the literature, such as register, minimization of cognitive complexity and avoidance of identity (horror aequi). The probabilistic constraints are tested in a series of mixed-effects Bayesian logistic regression models. The results indicate that the to-infinitive is particularly well represented in contexts with high information content. More specifically, if the expectation that a given infinitive is used with help, and not in another construction, is low, there are greater chances that the speaker will prefer the marked form of the infinitive. This tendency, which can be interpreted as communicatively efficient behaviour, is observed in all seven varieties.