1. TEMPLATE DESIGN © 2008 www.PosterPresentations.com Learning Words and Rules Abstract Knowledge of Word Order in Early Sentence Comprehension Yael Gertner and Cynthia Fisher (University of Illinois), and Julie Eisengart (Northwestern University Feinberg School of Medicine) ‏ Presented by Tatiana Gillespie and Isabel McLoughlin ABSTRACT EXPERIMENT 3 RESULTS DISCUSSION EXPERIMENT 2EXPERIMENT 4 Hypothesis: If the child had already abstracted a word-order rule for English transitive sentences they would look longer at the matching sentence – the one which the patient was the direct object of the sentence. Participants & Method: Twenty-four 2-year-olds (12 girls) – five eliminated, with a median vocabulary score of 62, watched the same videos and heard sentences with the subject as a pronoun, to see if the children knew that the object names the patient: “He is gorping the bunny”). Coding: 98% agreement. Results: Showed that children had a robust preference for the matching event in which the direct-object referent played a patient’s role. 2-year- olds use word order to comprehend transitive sentences with novel verbs, they inferred that “gorp” had to refer to an event in which the direct-object played a patient’s role. Discussion: Since the children only heard one character name it may be that they interpreted the first name they heard as referring to the more potent participant in the event, so they would have looked at the wrong event. Hypothesis: Younger children’s vocabulary is limited therefore little basis for abstraction, a lexical account would predict that at this age children would fail this task. Participants: 16 younger children, 21-month-olds, with the median vocabulary score of 34, Method: Videos of people with ordinary clothing, referred to as boy and girl. New actions by boy & girl Procedure: Similar to Experiment 1 but video with different events, and children were shown a preview of the videos Coding: agreed on 97% of the frames. Results: showed that the children had a strong preference for videos with the subject of a transitive verb naming the agent. Thus 21-month-olds use word order to understand novel words. Hypothesis: The children will look at the sentence in which the patient is the object. Participants & Method: 16 21-month-olds (7 girls), median vocabulary score of 29, anticipated. Coding: agreement on 99% Procedure: The sound track was similar to Experiment 3, with a word who added. Results: The children were looking longer than by chance at the matching video, where the patient was the direct object of the sentence. Thus 21-month- olds knew that the direct object named the character undergoing the action. Does early syntactic knowledge involve knowledge of words or rules? Early language comprehension and production are based on verb-specific knowledge. 25 and 21 month year olds extend their knowledge of English word order appropriately to interpret transitive sentences containing novel verbs. They represent language experience in terms of an abstract mental vocabulary. BACKGROUND This article was published in Psychological Science journal in 2006. Yael Gertner and Cynthia Fisher are from the University of Illinois, and Julie Eisengart is from the Northwestern University Feinberg School of Medicine. Their study builds on research of Hirsh-Pasek and Golinkoff (1996), as well as other supporters of the early-abstraction theory of language acquisition, to continue a debate with the lexical theory account (Braine 1963, Tomasello 2003). The argument is that differentiation of subjects and objects of transitive sentences by children depends on their innate ability to have an abstract format of word order along with lexical learning, instead of acquiring verb-specific knowledge first and slowly forming rules via comparison. Hypothesis: If the children detected the abstract word-order pattern of English transitive sentences then they would look longer at the event that matched the sentence which the agent was the subject of the sentence. Participants: Twenty-four 2-year-olds (11 girls) – one eliminated, native learners of English, with productive vocabulary ranging 27-99 words. In four experiments, young children used word order to interpret transitive sentences containing novel verbs. In all cases, the preference for the matching video appeared as soon as they heard the entire test sentence once, and as they watched the events together for the first time. This pattern directly contradicts the claim of the lexical account. The present findings are consistent with evidence that young children use early-developing syntactic knowledge to learn the meaning of verbs, syntactic bootstrapping. With very little vocabulary, 24 and 21-month-olds are constrained to represent language experience in an abstract mental vocabulary. These representations allow children to rapidly learn the rules as well as the words from the start. EXPERIMENT 1 Mean Proportion of Looking Time to the Matching Screen Within the Four 2-s Intervals of the First Test Trial in Experiments 1 Through 4 Experiment Interval 0–2 s 2–4 s 4–6 s 6–8 s Experiment 1: 25-month-olds (e.g., ‘‘The bunny is gorping the duck!’’).67 (.08).61 (.09).57 (.09).58 (.09) ‏ Experiment 2: 25-month-olds (e.g., ‘‘He is gorping the duck!’’).66 (.09).63 (.09).48 (.10).51 (.09) ‏ Experiment 3: 21-month-olds (e.g., ‘‘The girl is gorping the boy!’’).74 (.10).73 (.10).68 (.11).59 (.12) ‏ Experiment 4: 21-month-olds (e.g., ‘‘Who’s gorping the boy?’’).71 (.11).66 (.12).72 (.11).72 (.10) ‏ Note. Standard errors of the mean are shown in parentheses. Apparatus: The children sat on a parent’s lap facing two monitors and listened to a sound track presented centrally. Procedure: Each character was introduced (bunny and duck) then a transitive sentence introduced a novel word: for half the children, “The duck is gorping the bunny!”; for the other half, “The bunny is gorping the duck!” Coding: agreeing on 94% of video frames. Results: The children looked longer than expected by chance at the matching video, where agent was the subject of the sentence Meaning children used knowledge of word order to infer that whatever the non-word meant, in involved the action of the subject on the object EXPERIMENT 1 CONTINUED