| The skill of bidirectional optimization (i.e., taking into account the perspective of the conversational partner) seems to be necessary for several linguistic phenomena, including for conversational implicatures and the binding principles. It is proposed that children display difficulties with these phenomena, because they cannot take into account the other s perspective (Hendriks & Spenader, 2005/2006). In this project the reason will be investigated why children cannot apply bidirectional optimization. Because much research is done on the Binding Theory, the Pronoun Interpretation Problem is a logical starting point to study bidirectional reasoning. According to the Binding Theory the binding principles A and B specify to which antecedents reflexives like himself and pronouns like him refer. Up to the age of 6;6 children show violations of Principle B by incorrectly allowing pronouns to corefer with the local subject about half the time. From the age of 3;0 on children have learned to use Principle A correctly by taking the local subject as antecedent of a reflexive (e.g., Chien & Wexler, 1990). The difference in the comprehension of pronouns and reflexives is referred to as the Delay of Principle B-effect (DPBE). Hendriks and Spenader (2005/2006) explain the DPBE by the hypothesis that children lack the skills for taking into account the perspective of the conversational partner, operationalized in the idea of bidirectional optimization within the Optimality Theory framework (OT). Hendriks, van Rijn and Valkenier (2007) predict that children are unable to apply bidirectional optimization because of their limited speed of linguistic processing. Their prediction is based on a computational simulation of the acquisition of pronoun comprehension. The proposed main research question can be formulated as follows: * What is the reason that children cannot apply bidirectional optimization? Subquestions are: * How does the transition of unidirectional to bidirectional optimization take place? * Can the inability of children to optimize bidirectionally be attributed to other factors than a limited speed of linguistic processing? It is difficult to address questions about language acquisition like these with only psycholinguistic experiments. Cognitive modeling can lead to insights in language acquisition by explaining human behavior and cognition and formulating specific predictions. These predictions can be helpful in testing theories. Therefore the first year of the project will be focused on combining existing computational models. In the second year, the predictions of the computational models can be empirically tested by performing experimental studies. The precise method of these experiments will depend on the derived predictions, but commonly used methods to study the Pronoun Interpretation Problem are the Picture Truth Value Judgement Task (TVJT) and the Picture Selection Task. |
