HOME

Our research approaches language from a social networks perspective. We show how network structure can influence how good individuals are at understanding others and expressing themselves, why languages spoken by different communities differ from each other, and how individuals’ language use influences the language of the community. We examine these questions using a combination of individual differences, experimental, and computational methods and across different linguistic levels.

PhD studenthip

The lab is recruiting a PhD student. If you are interested, email shiri.lev-ari AT rhul.ac.uk ASAP.

LATEST NEWS

Read recent news articles about our research:

Check out our latest papers:

  • Boduch-Grabka, K. & Lev-Ari, S. (in press) Exposing individuals to foreign accent increases their trust in what non-native speakers say. Cognitive Science

TLDR People believe information less when it is delivered in a foreign accent even if they are not prejudiced. They doubt it because it is harder to process foreign-accented speech and people implicitly infer form the difficulty that the information is wrong. This bias against believing non-native speakers, however, can be reduced by exposing listeners to foreign-accented speech, because exposure trains listeners to become better at processing foreign-accented speech.

TLDR Languages that are spoken by larger communities (i.e., millions of speakers) are more sound symbolic than languages spoken by smaller communities (e.g., thousands of speakers). We demonstrate it by showing that people are better at guessing the meaning of words in unfamiliar languages if those languages are spoken by large rather than smaller communities. The reason for that is that communication is harder in larger communities, so the languages of large communities adapt and overcome the greater challenges by relying on sound symbolism, that is, having words that sound like what they mean.

TLDR People with larger social networks show poorer voice recognition, potentially because they attend less to speaker-specific information.

TLDR Real world social networks are interconnected, so information we receive from several people might originate from the same person. We show that people tend to neglect to consider this interconnectivity when learning information, and even more so when the information is linguistic (how to call something) compared to opinions. We also show that this neglect facilitates the spread of language and trends.

Upcoming presentations:

Zoom times!
27/07/2021: CogSci 2021
09/2021: ProtoLang