Event Details

This event is co-organized by the Multilingualism and Cognition Research Group and the Heritage Language 2 Consortium (HL2C).

Beyond Strength: How Weaker Associations Impact (Statistical) Word Learning

Tanja Roembke (University of Aachen)

When we learn a new word, an association is formed between the word and its meaning. Typically, it is assumed that a strong association between a word and its meaning is the end goal, signifying that a word has been learned well. In this talk I will present two projects that challenge this assumption: In the first half of the talk, I will present some data from two eye-tracking experiments on the role that pruning of weaker associations plays during word learning. Here, we manipulated the trial structure during training so that a secondary meaning of a word co-occurred with it for some words, but not for others. While we observed pruning both in the absence and presence of co-occurrence, how well these secondary associations were pruned depended on their initial strength. In the second part of the talk, I will present data from three experiments in which we investigated how easily words with a single meaning vs. words with multiple meanings are remapped to a new meaning. We hypothesized that words with multiple meanings, where each individual meaning may be less strongly associated with the word, would be more easily remapped to a new meaning. As predicted, we found that while words with multiple meanings were harder to learn, they were also more easily remapped. Taken together, these projects suggest that we need to consider not only how well a word is associated with its meaning, but also additional (secondary) associations to better understand the dynamics of word learning.

How to join this event:

Please join us in County South C89 or remotely via Teams. The link will be circulated via our mailing list.