University of Pennsylvania Working Papers in Linguistics


In this paper, we introduce Worldlikeness, a web-based tool for collecting and sharing cross-linguistic wordlikeness judgments (nonce word acceptability judgments) to facilitate typological psycholinguistic research. Typological psycholinguistic research is essential since crucial factors affecting language processing vary across languages, but these factors often too confounded to tease apart by comparing just two languages at a time. This type of research is nevertheless difficult since it requires testing many speakers from each language, using materials designed with the help of expert native speakers. Worldlikeness aims to make typological psycholinguistics more feasible, by providing tools for separate groups of experimenters to design experiments with text, audio, images, and video for individual languages, collect judgments and reaction times online, and crucially, share their data with each other for typological analysis. We show that Worldlikeness successfully replicated Mandarin wordlikeness judgments collected using traditional lab-based software, and report its first use in a cross-linguistic study collecting wordlikeness judgments from bilingual speakers of Mandarin and Southern Min.