Marlyse Baptista
Department of Linguistics
University of Pennsylvania

An experimental approach to congruence in language contact with implications for Creole genesis

In Baptista (2020), I provided a survey of congruent forms in 20 contact languages across 19 grammatical domains (including negation,Tense, Mood, Aspect markers, complementizers…), as they were reported in the scholarly literature (Kihm, 1990; Corne, 1999; Mufwene, 2008; Aboh, 2015). This empirical survey led to the design of a Pattern-Matter-Mapping (PMM) model which rests on the assumption that in the context of Creole formation, the linguistic properties that speakers perceive to be shared (congruent) among the input languages maybe more easily learned and are probabilistically more likely to emerge in a given Creole.

This supports linguists’ assumption that feature similarity across languages in contact can correspond to facilitation in acquisition and language emergence (as in the case of Creole languages). However, because feature similarity is often confounded with other variables (e.g., feature frequency, language type, transparency, perceptual salience), it remains unclear whether congruence per se benefits learners. In this project, our team (Danielle Labotka, Emily Sabo, Rawan Bonais, Susan Gelman and myself) abstracts away from such variables. Following up on a prior study (Baptista, Gelman and Beck, 2016), we provide an experimental test of the effects of congruence on acquisition through an artificial language-learning experiment involving English (L1) and two artificial languages (Flugerdu and Zamperese). English-speakers (N=163) were randomly assigned to one of four conditions, varying which of the three languages expressed negation with converging forms: all three languages (English, Flugerdu and Zamperese); only Flugerdu and Zamperese; only English and Flugerdu; or none. Our findings show that participants better acquired the negation morpheme when the form converged with negation in English but not when the forms in Zamperese and Flugerdu converged with each other.

We likewise found unanticipated spillover effects in which participants better acquired the vocabulary and grammar of the artificial languages when all three languages had converging negation forms. These findings provide insight into issues surrounding language acquisition in multilingual environments and Creole language formation.