Presentation
Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Lightning Talks
The influence of contextual variability on the learning and retention of novel words: Does the type of variability matter?
There is a Poster PDF for this presentation, but you must be a current member or registered to attend SNL 2023 to view it. Please go to your Account Home page to register.
Poster E20 in Poster Session E, Thursday, October 26, 10:15 am - 12:00 pm CEST, Espace Vieux-Port
Raphael Fargier1, Andreas Falck2, Tine Hovland2, Hakan Bayar2, Janne von Koss Torkildsen2; 1Université Côte d'Azur, CNRS, BCL, France, 2Department of Special Needs Education, University of Oslo, Norway
Adults predominantly learn new vocabulary from reading, and contextual variability benefits such learning. Importantly, contexts are never entirely repeated or entirely changing, and it remains unclear what features of variability lead to better learning of words and their meanings. In this project, we designed a web-based learning experiment, and used short fictional narratives to examine this issue. Participants encountered 8 novel words, each one in a block of 3 successive narrative contexts that could be identical or could vary to different extents. We manipulated variability in non-core object features (e.g. color, size) and variability in situational features of the stories in which target words were encountered (e.g. protagonists, locations, events). For each target word, 2 semantic features were consistent across all stories and were considered core semantic features of the target word (e.g. “can be grasped”, “sticks”). After each story, participants were invited to define the target word through typing. An immediate lexical decision task was performed to investigate word recognition. While learning occurred in a first session, we conducted a follow-up session, 1 day later, which also included the definition task and the lexical decision task. A thorough coding procedure of the definitions was developed and several scores were computed. Scores included notably the number of core features mentioned per definition, and the number of explicit references to variability (e.g. lists of different options of a feature by use of modal verbs and quantifiers). The number of words correctly recognized, operationalized as a lexical decision score, was also used as a dependent variable. The study was pre-registered on OSF (https://osf.io/6gezm) and power analysis was based on pilot data. Final dataset included 280 participants. Definition scores indicated significantly better learning of core semantic features in the conditions with situational variability, including the condition with variability in both object and situational features (i.e. main effect of variability in situational features: F(1, 279) = 82.4, p<0.001). Target words were also better recognized in those conditions (F(1, 279) = 25, p<0.001). Although performance dropped at follow-up up 1 day later, performance remained higher in the conditions with situational variability. Our results suggest that situational variability in narrative contexts support word learning as well as learning and retention of core semantic features of words. We suggest that the simulation and integration of variable contexts allows us to better identify consistent semantic features and to form a unified memory representation that is more resistant to decay. We discuss how the paradigm could be adapted for experiments using electroencephalography to track the neural correlates of lexical-semantic representation under the influence of contextual variability.
Topic Areas: Meaning: Lexical Semantics, Language Development/Acquisition