Poster #114 - Syntax or something else: what is the source of systematic errors in children’s complex questions?

• In Event: 3-053 - Poster Session 10
  In Poster Session: PS 10 Section - Language, Communication

Integrative Statement

English speaking kindergarteners have been reported to make unique syntactic errors in biclausal questions. In elicitation tasks, children may insert a medial wh-word in their production, as in (1) (Thornton, 1990); they also show evidence of immature comprehension by answering the medial wh-phrase, as in (2) (DeVilliers & Roeper, 1995).

(1) What do you think what is in the box?
(2) Experimenter: How did the boy say what he caught?
Child: A fish!

These errors have been cited as evidence of immature syntactic knowledge, whereby children mistakenly use a Scope-Marking/SM construction that is grammatical in some languages hence universally available, lending support to the claim that young children initially have access to Universal Grammar. However, various methodological issues (e.g. pragmatic biases in stimuli) lead us to question this conclusion. The current study used improved experiment designs to investigate children’s production (Exp 1) and comprehension (Exp 2) of information questions. Exp 3 combines both experiments in a within-subject design.

Experiment 1 used an interactive story to elicit biclausal questions. Children worked together with a fairy to ask characters what they believed should be done at ‘cross-road moments’ in the story. The fairy provided a preamble "What/Who do you believe…" and the child had to repeat the prompt and complete the question. We elicited 12 wh-questions from 30 children (mean age = 5;0). Of the questions produced, 22.5% contained two wh-phrases. Individual production of this structure fell into one of 3 SM patterns: consistent (>67%; n=4), rare (<20%; n=18), and moderate (20-50%; n=8). This widespread distribution is compatible with the idea that language development involves a competition of multiple grammars (Yang, 2002). However, if children are using alternative grammars, it should be apparent in comprehension as well as production.

In Experiment 2 we presented 42 children (mean age = 4;9) with 8 stories, after which a puppet asked a question. Experimental condition questions were biclausal how + what, which could allow an SM interpretation, while control, unambiguously English questions were what + that questions (Table 1). We addressed two methodological issues in previous work. First, story bias: we created stories with equal emphasis on events described by the main and embedded verbs. Second, children might process the embedded clause as an independent question. We designed stimuli that distinguish SM interpretation (i.e., what Jane said she saw) from a monoclausal interpretation of the embedded clause (i.e. what she saw). In our study, 8.3% of responses indicated an SM interpretation of the questions. These responses were less frequent than errors resulting from misremembering the stories (11.5%).

In Experiment 3 (on-going), 14/20 planned children (mean age = 5;4) have so far performed the tasks from experiments 1 & 2. Currently, we have replicated the findings from Exp 1 and 2, and find no correlation between production and comprehension performance (p >.1).

Overall, these results do not suggest the use of an alternative grammar. We interpret the production errors as resulting from children’s immature production mechanisms,
specifically memory and planning.

Authors

• Carolyn J Lutken, Johns Hopkins Univeristy
  Presenting Author

• Geraldine Legendre, Johns Hopkins University
  Non-Presenting Author

• Akira Omaki, University of Washington
  Non-Presenting Author