2018 Sat Poster 6690

Saturday, November 3, 2018 | Poster Session II, Metcalf Small | 3:15pm

With or without too: reasoning about people’s questions and their presuppositions
V. Hacquard, R. Dudley, J. Lidz

Young children don’t always demonstrate sensitivity to aspects of what is meant beyond the contributions of asserted content[1-3]. But interlocutors convey very rich meaning by appealing to alternatives (as in conversational implicature) and focus-sensitive presuppositions also appeal to alternatives (1). We ask whether children demonstrate sensitivity to alternatives in understanding implicatures and presuppositions, and how such sensitivity changes over time. We find that preschoolers show emerging but sophisticated sensitivity to what speakers convey beyond asserted content through their questions and presuppositions.

In previous studies, children performed well in reward-based tasks at 3 years[4], but these may overestimate children’s knowledge[5]. Studies with game-based tasks have not found success until almost 5 years[3], but involved potentially confusing sets of alternatives. We use a game- based task with intuitive sets of alternatives. Participants must find a toy hidden in one of three boxes (red, green, yellow) and test-items are clues. Given this task structure, there are three salient alternatives (2). We assess children’s understanding of how speakers convey more than what is literally said, by appealing to these alternatives. After one experimenter (E1) hides a toy, another (E2) peeks into one box and provides a clue (Table 1) based on what was gleaned through peeking, as an assertion (E-F) or a request for information from E1 (A- D). For A-D, a box is either one E2 peeks into (“peeked”), one that E2 mentions in their clue (“mentioned”), or one that isn’t made salient (“other”).

Given these items, information about the location can come from different sources: asserted content which does not require the consideration of alternatives (“mentioned”: it is or isn’t empty); or by appealing to alternatives, via conversational implicature (“peeked” in A-D: if it weren’t empty then E2 would have said so) or via presupposition accommodation (“peeked” in B,D: if this box weren’t empty then too’s presupposition would be unsupported); or reasoning by elimination (“other”, given information gleaned about “mentioned”/“peeked”). E-F are control-items. A-D are test-items based on a 2x2x2 within-subjects design, manipulating clue-type(affirmative(+)/negative(–)), trigger-type(none/too), and age- group(young=3 years/old=4 years). Results are in Table 2.

We compared the effects of clue-type, trigger-type, age, and the interaction between clue-type and trigger-type on children’s searching behaviors. If children are sensitive to information provided by non-asserted content, we expect a main effect of clue-type. If children are sensitive to the way that presupposition triggers appeal to alternatives, we expect a main effect of trigger-type. If the mechanisms that underlie children’s understanding of implicature and presupposition differ, we expect an interaction between clue-type and trigger-type. If children’s sensitivity changes over time, we expect a main effect of age. We find a significant effect of age-group (F(1,223)=10.813,p<.01) and clue-type (F(1,223)=54.554,p<.0001), but not trigger-type (F(1,223)=0.047,p>.80) and no interaction (F(1,223)=0.014,p>.90).

These results suggest children have sophisticated understanding of alternatives and how speakers may appeal to them, which emerges between 3-4. However, we cannot conclude that the mechanism which allows children to understand presuppositional sentences (1) differs from that which enables them to understand conversational implicatures.

References

[1] Robinson, Goelman, Olson (1983) Journal of Developmental Psychology;

[2] Noveck (2001) Cognition;

[3] Jasbi (2016) Boston University Conference on Language Development;

[4] Berger & Hohle (2012) Journal of Child Language; [5] Katsos & Bishop (2011)

Cognition