Alternations in contact and non-contact varieties

The Multifactorial Prediction and Deviation Analysis (MuPDAR) method (Gries & Deshors 2014) represents an influential methodological advance in studying variation in contexts where linguistic choices in a “peripheral” variety (learner language, New Englishes) are studied in relation to the “central” variety. In this article we demonstrate how the method may be extended to study how varieties produced in settings of language contact (including translation) differ from non-contact varieties, particularly with respect to the degree of lexicogrammatical explicitness. We use the method to determine how (dis)similar the factors governing that-omission are in two different types of contact varieties, namely South African translated (trans-SAE) and South African non-translated English (SAE), in relation to British (GBE) English. The results show that the choices made in the contact varieties can be predicted to a reasonable extent, although South African translators and South African non-translators have a higher and lower inclination respectively to use explicit that compared to GBE non-translators. Based on the findings, we re-evaluate the explanations proposed for the increased explicitness of translated language through the frame of language contact, outlining the advantages of multifactorial methods over the frequency-based methods favoured in earlier studies.