<?xml version="1.0" standalone="yes"?>
<Paper uid="N06-2031">
  <Title>Computational Modelling of Structural Priming in Dialogue</Title>
  <Section position="4" start_page="122" end_page="123" type="metho">
    <SectionTitle>
3 Results
</SectionTitle>
    <Paragraph position="0"/>
    <Section position="1" start_page="122" end_page="122" type="sub_section">
      <SectionTitle>
3.1 Task-oriented and spontaneous dialogue
</SectionTitle>
      <Paragraph position="0"> Structural repetition between speakers occured in both corpora and its probability decreases logarithmically with the distance between prime and target. Figure 1 provides the model for the influence of the four factorial combinations of ROLE and SOURCE on priming (left) and the development of repetition probability at increasing distance (right).</Paragraph>
      <Paragraph position="1"> SOURCE=Map Task has an interaction effect on the priming decay ln(DIST), both for PP priming (b = [?]0.024,t = [?]2.0, p &lt; 0.05) and for CP priming</Paragraph>
      <Paragraph position="3"> ficients indicate more decay, hence more priming.) data points per corpus.</Paragraph>
      <Paragraph position="4"> In both corpora, we find positive priming effects. However, PP priming is stronger, and CP priming is much stronger in Map Task.</Paragraph>
      <Paragraph position="5"> The choice of corpus exhibits a marked interaction with priming effect. Spontaneous conversation shows significantly less priming than task-oriented dialogue. We believe this is not a side-effect of varying grammar size or a different syntactic entropy in the two types of dialogue, since we examine the decay of repetition probability with increasing distance (interactions with DIST), and not the overall probability of chance repetition (intercepts / main effects except DIST).</Paragraph>
    </Section>
    <Section position="2" start_page="122" end_page="123" type="sub_section">
      <SectionTitle>
3.2 Frequency effects
</SectionTitle>
      <Paragraph position="0"> An additional model was built which included ln(FREQ) as a predictor that may interact with the effect coefficient for ln(DIST).</Paragraph>
      <Paragraph position="1"> ln(FREQ) is inversely correlated with the priming effect (Paraphrase: blnDist =</Paragraph>
      <Paragraph position="3"> p &lt; 0.001). Priming weakens with higher (logarithmic) frequency of a syntactic rule.</Paragraph>
    </Section>
  </Section>
  <Section position="5" start_page="123" end_page="123" type="metho">
    <SectionTitle>
4 Discussion
</SectionTitle>
    <Paragraph position="0"> Evidence from Wizard-of-Oz experiments (with systems simulated by human operators) have shown that users of dialogue systems strongly align their syntax with that of a (simulated) computer (Branigan et al., 2003). Such an effect can be leveraged in an application, provided there is a priming model interfacing syntactic processing.</Paragraph>
    <Paragraph position="1"> We found evidence of priming in general, that is, when we assume priming of each phrase structure rule. The priming effects decay quickly and nonlinearly, which means that a dialogue system would best only take a relatively short preceding context into account, e.g., the previous few utterances.</Paragraph>
    <Paragraph position="2"> An important consideration in the context of dialogue systems is whether user and system collaborate on solving a task, such as booking tickets or retrieving information. Here, syntactic priming between human speakers is strong, so a system should implement it. In other situations, systems do not have to use a unified syntactic architecture for parsing and generation, but bias their output on previous system utterances, and possibly improve parsing by looking at previously recognized inputs.</Paragraph>
    <Paragraph position="3"> The fact that priming is more pronounced within (PP) a speaker suggests that optimizing parsing and generation separately is the most promising avenue in either type of dialogue system.</Paragraph>
    <Paragraph position="4"> One explanation for this lies in a reduced cognitive load of spontaneous, everyday conversation.</Paragraph>
    <Paragraph position="5"> Consequently, the more accessible, highly-frequent rules prime less.</Paragraph>
    <Paragraph position="6"> In task-oriented dialogue, speakers need to produce a common situation model. Interactive Alignment Model argues that this process is aided by syntactic priming. In support of this model, we find more priming in task-oriented dialogue.3</Paragraph>
  </Section>
class="xml-element"></Paper>