Documenting multilingualism through crowdsourcing and citizen science

Crowdsourcing, a data collection method which broadly concerns gathering data from a large collection of individuals, has become a prominent method across a plethora of fields. In cases that concern the attitudes and knowledge of a specific population, using crowdsourcing methods provides both access to that population and data comparable in quality to that of more traditional methods. We extend this methodology to the study of language use in an effort to capture more effectively the richness of the diverse languages spoken in a multilingual population. We test this on a college campus, an ideal testbed exemplifying the type of diverse language use representative at a larger scale. The present study uses crowdsourcing and citizen science to gauge linguistic diversity on a college campus, targeting different and diverse forms of language knowledge, including informal and unconventional forms of language exposure. Different open-ended prompts inquiring about language use were posted at ten different on-campus locations easily accessible to participants (in this case, college students) in an effort to lower the barrier to entry, compared to more traditional formats of language assessment e.g., questionnaires or surveys that potentially result in artificially strict self-assessment of one’s language proficiency. Prompts were given either on poster paper on which participants could write a response (e.g., a specific language) and/or more detail (e.g., a specific variety of the language, or context, or interlocutors), or submit their response on slips of paper to a designated box. Collecting data within a setting wherein every participant is viewed as a citizen scientist, and every input is valued, captures an empirical depth that is otherwise difficult in more traditional laboratory settings. Our results underscore the strengths of crowdsourcing as a valid, novel, non-traditional form of linguistic data collection that elicits robust data and attracts a wider range, providing an important window into the true diversity of language knowledge present on an undergraduate college campus, with 52 languages being represented in our data, including mixed languages and otherwise underrepresented languages.