Learning in Leaps: The Development of Unprompted Integration and Self-Derivation

• In Event: Poster Session 08
  In Poster Session: PS 08 Topic - Attention, Learning, Memory

Abstract

What are the boundary conditions that may limit semantic knowledge extension across development? Semantic knowledge can be extended directly, through experiences such as classroom instruction. Importantly, semantic knowledge also can be extended indirectly, through processes including associative inference, analogy, and self-derivation. These productive processes allow learners to self-generate novel knowledge. Although substantial research has investigated these processes (cf. Bauer & San Souci, 2010; Gentner, 1983; Schlicting & Preston, 2015), the boundary conditions that limit productive processing, particularly in childhood, are not well-understood. Accordingly, in the present work, we aimed to elucidate the conditions under which memory integration and subsequent self-derivation occurs across development.

Memory integration and subsequent self-derivation is the integration of multiple, related episodes of learning to generate novel facts. For example, one might learn that hematopoiesis is the cellular formation of blood. Later on, one might learn hematopoiesis occurs in bone marrow. Integrating these facts supports self-derivation of the novel fact the cellular formation of blood occurs in bone marrow. Thus, self-derivation supports generation of novel, never directly-taught information. But under what conditions do children and adults engage this process? One potentially salient condition may be the presence or absence of a prompt to integrate. There is preliminary evidence that adults begin to integrate at the time they encode a second, related fact (Varga & Bauer, 2017), whereas children may not integrate until they are prompted to self-derive (e.g., Bauer et al., 2015). Thus, children may require a prompt to integrate and self-derive, whereas adults may engage the process independent of a prompt.

In this study, we introduce a novel behavioral paradigm—recursive integration—to examine prompted and unprompted integration and subsequent self-derivation in both children (7-9 years old) and adults (18-22 years old). Participants were taught novel, true facts that could be integrated to self-derive new facts (as in the above example). Half the trials included a prompt to integrate, whereas half did not. Performance on final self-derivation questions (both open-ended and forced-choice) was measured. Overall, we found that both children and adults benefitted from a prompt to integrate; both age groups engaged the process of self-derivation even unprompted. However, our results also suggest developmental differences. Whereas there was a significant difference between prompted and unprompted trials in open-ended questioning for adults, children did not demonstrate a significant difference between prompted and unprompted trials until open-ended and forced-choice performance were considered together (Figure 1). Further, the relation between prompted and unprompted performance was stronger in adults than in children. Thus, prompting may confer a benefit for children primarily in forced-choice questioning. Additionally, adults may rely on similar mechanisms when integrating both prompted and unprompted, but children’s integration may depend on differing mechanisms when prompted versus unprompted.

The data demonstrate that both children and adults self-engage processes by which knowledge is productively extended, but prompts to integrate facilitate performance. Further, there may be developmental change in mechanisms underlying productive processing. These findings inform our understanding of how and when the knowledge base is built across development.

Authors

• Julia Wilson, Emory University
  Presenting Author

• Patricia J Bauer, Emory University
  Non-Presenting Author