Slide Slam S9 Sandbox Series
The influence of contextual variability on learning novel words: Does the type of variability matter?
Raphaël Fargier1, Andreas Falck2,3, Janne von Koss Torkildsen1; 1University of Oslo (Norway), 2Lund University (Sweden), 3Institut Jean Nicod, ENS-EHESS-CNRS, PSL University (France)
Adults predominantly learn new vocabulary incidentally, from reading. Several studies indicate that contextual variability benefits such learning. However, we do not know what features of variability underlie this facilitation. Often, context variability is operationalized as the number of unique documents a new word appears in, or as the number of different topics covered by the texts. In addition to studies on textual variability, studies with young children suggest that visual variability benefits learning of object words. In particular, variability in irrelevant object features help children determine the core features of objects, which may support generalization and retention. In the current study, we examine the effects of two types of variability on the learning and retention of novel object words in narrative contexts. Specifically, we manipulate variability in situational contexts in which new objects are experienced (location, characters, activities) and variability in irrelevant object-features (color, size, texture). Narratives are optimal for comparing situation-related and object-related variability manipulations, as they invite the reader to build a mental simulation of the contexts. Web-based behavioral experiments will be conducted with English-speaking adults, across two sessions. In session 1, participants will encounter 16 novel words in learning blocks of three consecutive short fictional narratives of approximately 50 words each. Stories will either display high variability in non-definitional object features (e.g. color, size; Condition A), high variability in (non object-related) situational features (e.g. people, location; Condition B) or high variability in both object and situational features (Condition C). In a control condition, the same story will be repeated three times (Condition D). A fourth of the target words will be randomly allocated to each condition. Immediately after reading the three narratives, participants will be asked to define the word. The semantic features given in these definitions will be the outcome of interest. Ratings of participants’ engagement in each narrative will also be collected. In session 2 (follow-up), two weeks later, participants will complete three additional tasks: a lexical decision task, an open-ended definition task (identical to session 1) and a multiple-choice task where the participants will be asked to (a) select the words’ core semantic features, and (b) judge the acceptability of non-core semantic features. The analyses will examine our main hypotheses: 1) The variability conditions (A-C) will lead to improved retention of both word form and meaning, as evidenced by increased accuracy in lexical decision and greater definition knowledge. 2) Words learned in the object variability conditions (B + C) will lead to recall of a larger number of core features, and higher acceptance of irrelevant features (generalization) in the multiple choice task at follow-up. 3) Words learned in the situational variability conditions (A + C) will foster decontextualized knowledge, and will be associated with higher engagement at immediate post-test, in turn leading to better memory performance at follow-up. We will present pilot data from the behavioral experiment, which will serve as the foundation for an electroencephalography experiment aiming to track the neural correlates of word learning under the influence of contextual variability.