Conceptual and psycholinguistic factors in the acquisition of conditionals

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

Abstract

Children produce conditionals (if-clauses) later than other complex constructions (e.g., because-clauses, when-clauses) for reasons that are poorly understood (Bowerman, 1986; Kuczaj & Daly, 1979; Reilly, 1986). On a conceptual explanation, children acquire conditionals later because they are cognitively more complex (e.g., hypothetical) than other constructions (ibid.). On a psycholinguistic explanation, children produce conditionals infrequently because their meaning can also be conveyed by alternative constructions (Bowerman, 1986; McCabe et al., 1983). However, relevant evidence comes primarily from naturalistic studies of children’s production, thus making it hard to adjudicate between competing explanations. Here, we revisit children’s production of conditionals by eliciting descriptions of events with simple conditional/causal structure. Unlike prior research, we tested both children and adults and manipulated the hypotheticality of events.

Eighty 3-to 6-year-olds and 20 adults were presented with a causal toy (a box that lights up if you put the “right block” on) and were shown by a teacher-puppet how it works. Then, participants taught a student-puppet how to make the box light up. In low-hypotheticality trials, the student-puppet requested information in the present indicative form; in high-hypotheticality trials, the student-puppet made a failed attempt to light up the box (by using the inactive block) and requested information in the past counterfactual form (Fig.1). Participants’ verbal responses were recorded. On the conceptual explanation, children should use conditionals less than other constructions-especially in high-hypotheticality trials-but no such difference should be found in adults. On the psycholinguistic explanation, children, similarly to adults, may use conditionals as frequently as other constructions.

To assess complexity in participants’ responses, we coded for mention of antecedent (block on box) and consequent propositions (box lights up), and how these related. Inspection of the data demonstrated that children overwhelmingly mentioned antecedents only (main clause), while adults mentioned both propositions (main+subordinate clauses). A mixed effects logistic regression on mention of subordinate clauses confirmed these observations: an effect of age revealed that adults were more likely to use subordinate constructions than children (MAD=.65, MCH=.25). An effect of trial type, qualified by an interaction of trial type and age indicated that 6-year-olds were more likely to use subordination in low- than high-hypotheticality trials (MLOW=.36, MHIGH=.21), but there was no such difference in younger children (MLOW=.24, MHIGH=.22).
To assess the frequency of each type of construction in participants’ responses, we coded the types of subordinate clauses used (Fig.2). Inspection of the data demonstrated systematic use of conditionals after age 5. Log-linear models tested the fixed effect of clause type on counts of mention in each age group and trial type. Results showed that mention of conditionals did not differ from mention of other complex constructions (temporal, causal) at any developmental stage.

In sum, results support the psycholinguistic explanation of children’s late production of conditionals; unlike prior naturalistic studies, our experimental investigation found that conditionals are not delayed compared to other constructions of similar morphosyntactic complexity, in either low- or high-hypotheticality scenarios. The present findings inform our understanding of the factors that affect the order of acquisition of semantically complex constructions.

Authors

• Myrto Grigoroglou, University of Toronto
  Presenting Author

• Patricia A Ganea, University of Toronto