<?xml version="1.0" standalone="yes"?>
<Paper uid="C04-1195">
  <Title>Tony.Veale@UCD.ie</Title>
  <Section position="7" start_page="2" end_page="2" type="concl">
    <SectionTitle>
6 Conclusions and Future Work
</SectionTitle>
    <Paragraph position="0"> We are currently considering ways of broadening the scope of internal validation while maintaining its conceptual rigour. This should counter-balance the high rejection rate caused by an overly conservative external validation process, and thereby allow us to identify a higher percentage of H-creative products. As shown with the &amp;quot;pizza&amp;quot; examples of section 4.3, we have already begun to explore the possibilities of validation latent in the WordNet ontology itself. So while the use of web content for external validation suggests that creative discovery has a role to play in producing and expanding web queries, internal validation remains our central thrust, leading to what we hope will be a new, more creative model of the thesaurus.</Paragraph>
    <Paragraph position="1"> In grounding our discussion in the creative framework of Boden (1990) and its formalization by Wiggins (2003), we have placed particular emphasis on the labels P-Creative and H-Creative.</Paragraph>
    <Paragraph position="2"> However, the empirical results of section 5 suggest that this binary categorization may be overly reductive, and that a more gradated system of labels is needed. For instance, the novel compounds computer-consultant and handwriting-consultant are both created from a yoking of the domains expert and consultant, and because each is externally validated, each is considered P-Creative. However, while only a handful of documents can be marshalled to support handwriting-consultant, the amount of web evidence available to support computer-consultant is vast. So it is wrongheaded to consider both as equally P-Creative and lacking in H-Creativity, since the dearth of existing uses suggests handwriting-consultant has far greater novelty. Perhaps what is needed then is not a binary categorization but a continuous one, a numeric scale with P- and H-Creativity as its poles.</Paragraph>
    <Paragraph position="3"> This scale would function much like the continuum used by (MacCormac, 1985) to separate banal metaphors (which he dubbed epiphors) from creative ones (or diaphors).</Paragraph>
  </Section>
class="xml-element"></Paper>