The Co-Construction of Counterfactual Worlds in Parent-Child Conversations

• In Event: How do children reason about hypothetical possibilities? Evidence from language production

Abstract

Episodic counterfactual reasoning (eCFR) – reasoning about hypothetical alternatives to past events – plays a central role in adults’ causal and moral judgments and decision-making (Byrne, 2016). This ability is cognitively demanding, showing protracted development between the ages of 4 and 8 years (Beck & Riggs, 2014; Nyhout & Ganea, 2019). Most of what we know about the development of eCFR comes from studies in which children are asked to reason about counterfactuals in response to conditional questions. In which everyday contexts do children first hear and produce counterfactuals? In this study, we explored parent-child reminiscing as a promising context, examining the conversational contexts in which counterfactuals may emerge and the dynamics of counterfactual co-construction. We propose that children’s early counterfactuals may be scaffolded by caregivers via a co-construction process, with one member of the dyad offering an antecedent (e.g., “if we had remembered an umbrella…”) and the other offering a consequent (e.g., “…we wouldn’t have been wet.”).

Method
We instructed dyads (parents and 3-5-year-olds; N = 63) to reminisce on positive and negative shared experiences to explore the incidence of spontaneous counterfactuals, and later to discuss “what could have happened differently” (prompted counterfactuals). Conversations were video-recorded and later transcribed. We coded transcripts for:

1) Spontaneity: Whether counterfactuals arose in the Spontaneous or Prompted phase.
2) Context: Types of events that evoked counterfactuals – Positive-Routine (e.g., getting ice cream); Negative-Routine (e.g., going to the dentist); Positive-Unexpected (e.g., getting a surprise gift); Negative-Unexpected (e.g., getting hurt).
3) Dynamics: Who initiated & completed each counterfactual.

Results
Spontaneity
Spontaneous counterfactuals arose infrequently (M = 0.1, SD = 0.1), however all dyads produced counterfactuals when prompted (M = 7.00, SD = 3.98). Because of the lack of variance in spontaneous counterfactuals, subsequent results refer to prompted counterfactuals.

Context
Figure 1 presents results by context and dyad member. We conducted two separate generalized estimating equation analyses (GEEs) with context as a predictor and age as a covariate - one with children's production, and one with parents' production as a dependent measure. There were no significant effects of age.
For children, production of counterfactuals did not significantly differ by context.
For parents, the effect of context was significant, Wald X2 (3) = 15.37, p = .002. Specifically, parents produced more counterfactuals when discussing negative-unexpected compared to negative-routine events, p = .007, and positive-routine compared to negative-routine events, p = .005. All other comparisons were non-significant.
Dynamics
Figure 2 presents conversational dynamics. Parents initiated counterfactuals more frequently (70%) than children (30%), p < .001, and were significantly more likely to offer a complete counterfactual (i.e., antecedent + consequent) than children, p < .001. However, children were significantly more likely to complete a co-constructed counterfactual (77%) than parents (23%), p < .001, suggesting early conversational counterfactuals are heavily scaffolded by parents.
Discussion
The results of the current study indicate that counterfactuals arise relatively infrequently in everyday conversations about past events between parents and preschoolers. However, when they do, parents scaffold children’s production by initiating counterfactuals that children can build on.

Authors

• Angela Nyhout, University of Kent
  Presenting Author

• Emily Veall

• Patricia A Ganea, University of Toronto