Summative imitation as a mechanism for generative learning and innovation in both preschoolers and adults

• In Event: Enhance cognitive functions across development: factors, strategies or interventions

Abstract

Introduction: A unique feature of human culture is that it evolves overtime (Henrich, 2018; Laland, 2016). At the heart of cultural evolution is innovation, widely believed to be driven by asocial (individual) learning or insights, with social learning playing only a major role in their dispersal and propagation (Cosmides & Tooby, 2002; Legare & Nielsen, 2016; Pinker, 2010). However, Subiaul & Stanton (2020) have demonstrated that novel artifacts—innovation— (e.g., Tower Building) may be produced by combining two distinct elements (tower base, apex) learned by imitation, where the novel and unobserved individual ‘insight’ (placing apex atop base) may also be driven by social learning. To address this question, we used this same paradigm to explore whether cumulative cultural innovations can arise from generative social learning; that is, applying observed actions to novel exemplars. Specifically, participants were presented with twice as many tower elements as they observed during the demonstration and, then, measured whether participants extended the actions saw to the new objects.

Methods: Experiment 1 (N = 101, Mage = 23.27, SD = 7.63; females = 51) included three independent conditions: Imitation, Emulation, Invention (Baseline). The imitation group saw a video of a model stacking two cubes and another combining two squares. The emulation group was identical except that the video only showed the final products (e.g., 2 stacked cubes, 2 combined squared) omitting any stacking actions. The Invention group was not shown any videos. All groups where then given 6 different colored cubes and 6 flat squares of two different colors. Experiment 2 (N = 59, Mage = 5.03, SD = .12; females = 32) included two learning conditions: Imitation and independent invention. Procedures mirrored those used in Exp. 1 except that children observed live demonstrations and given 8 tower elements.

Measures: Our measures included (1) tower height and (2) generative learning: building of the optimal tower that included (a) stacking all cubes (generative stacking) (b) combining all squares (generative combining of squares) and placing them on atop of the other (summative imitation),

Results: In Experiment 1 and 2 learning groups (excluding emulation) did not differ in terms of Tower Height (Exp. 1: Adults, p = .12; Exp. 2: Children, p = .11). However, imitation groups differed from Invention when building the optimal tower (Exp 1: p = .03, Exp. 2: p = .06). Both adults and children in the imitation group evidenced significantly more generative social learning (i.e., generative stacking and combining of squares) than those in the Invention group (All ps < .05).

Conclusions: These results show that from an early age, at least some “individual” insights that result in innovations are mediated by social rather than asocial learning and inferences. Of course, some innovations arise from a complex mixture of social and asocial insights. However, the present study demonstrates that these two forms of innovations are dissociable. A question for future research is which type of innovation (i.e., social- vs. asocial insights) is the main driver of cultural evolution in humans.

Authors

• Adrian Istvan Varallyay, Columbia University
  Non-Presenting Author

• Nathalia Beller, George Washington University
  Non-Presenting Author

• Francys Subiaul, George Washington University
  Presenting Author