No abstract

Miniature language learning is gaining increasing popularity to study biases underlying language universals. However, it is unclear whether learning preferences in these studies are influenced by learners’ native language. We ask whether a previously identified bias to balance production effort against message uncertainty holds across speakers of structurally different languages. We expose English (fixed order language without case) and German (flexible order language with case) speakers to miniature languages with optional case and either fixed or flexible constituent order and study their deviations from the input. We find that English and German speakers restructure the input in the same way: They match the input constituent order proportions and use more case in the flexible order language than in the fixed order language, thus following the bias to balance production effort against message uncertainty. Our findings suggest that this bias and its specific realization are independent of learners’ native language.

Pretrained transformer-based language models achieve state-of-the-art performance in many NLP tasks, but it is an open question whether the knowledge acquired by the models during pretraining resembles the linguistic knowledge of humans. We present both humans and pretrained transformers with descriptions of events, and measure their preference for telic interpretations (the event has a natural endpoint) or atelic interpretations (the event does not have a natural endpoint). To measure these preferences and determine what factors influence them, we design an English test and a novel-word test that include a variety of linguistic cues (noun phrase quantity, resultative structure, contextual information, temporal units) that bias toward certain interpretations. We find that humans' choice of telicity interpretation is reliably influenced by theoretically-motivated cues, transformer models (BERT and RoBERTa) are influenced by some (though not all) of the cues, and transformer models often rely more heavily on temporal units than humans do.