Programme: 

12:00: Stereotypes may be beneficial in language comprehension
Speaker: Prof. Holger Mitterer, Hosted by: Department of Cognitive Science, MAKS

13:00: Q & A session/informal discussion

Admission is free, but kindly reserve a place by sending an email.

Stereotypes may be beneficial in language comprehension

Abstract:

The assumption of Universal Grammar assumes that we, as humans, have a specific, genetically endowed language faculty which allows us to acquire any given language. As a counterpoint, statistical learning has been hailed as the foundation on which language is acquired (Aslin, 2017).

Research over the last decades has shown how powerful statistical learning is and how the brain makes use of every co-variation that exists in language (Hagoort & Van Berkum, 2007), even in syntactic core operations in language processing (Hanulíková et al., 2012). An interesting detail of the last finding is that it is based on the correlation of non-native accents and grammatical errors, which is only probabilistically true (i.e., there are speakers with an accent who will not make grammatical errors). We further investigated how probabilistic aspects related to speakers’ group identity influence language processing using eye-tracking in the Semitic language Maltese, exploiting the fact that many verb roots carry multiple meanings.

For instance, q-s-m has the meaning “to break” but can also be used for crossing the street. We found 48 verbs that have multiple meaning that are differentially associated with different types of speakers, with distinctions by age (e.g., l-għ-b, which is likely to mean “to play” if used by a child but “to gamble” if used by an adult) and gender (e.g., r-fgħ, which is more likely to be associated with heavy lifting with male speakers but stowing away with a female speaker). Results showed that listeners are able to exploit such associations. This raises the question whether these are based on stereotypes or statistical learning. However, as I will explain, this juxtaposition is a fallacy since stereotypes may actually rely on statistical learning. Further research shows that such effects are indeed strongly tied to statistical regularities since they reflect ecological effect sizes.

Speaker’s Profile:

After studying psychology at the Universities of Bielefeld (Germany) and Leiden (The Netherlands), Holger Mitterer received his PhD (cum laude) at the University of Maastricht in 2003. From there, he moved to the Max Planck Institute of Psycholinguistics in Nijmegen (Netherlands), where he worked for nearly 9 years.

During this time, he published widely in a wide range of journals in Linguistics, Experimental Psychology, and Cognitive Neuroscience, mostly on topics related to spoken-word recognition. One of his most cited papers shows how subtitles can be a resource in second-language acquisition. In 2013, he moved to the University of Malta and served as an Associate Editor on the journal Laboratory Phonology from 2014-2019 and in 2019 become joint Editor-in-Chief of the interdisciplinary journal Language and Speech since 2019. Under his leadership, the journal was the first in Linguistics to introduce registered reports to combat the replication crisis.