<?xml version="1.0" standalone="yes"?>
<Paper uid="W03-0612">
  <Title>Population Testing: Extracting Semantic Information On Near-Synonymy From Native Speakers</Title>
  <Section position="4" start_page="0" end_page="0" type="concl">
    <SectionTitle>
4 Discussion
</SectionTitle>
    <Paragraph position="0"> We are primarily researcher and practitioners in the field of foreign language studies, and our interests are focused on the design and implementation of electronic dictionaries and thesauri for human users who are studying a second language. Nevertheless, we have benefited greatly from exposure to research done in computational linguistics, and look forward to the exchange of ideas that we hope would benefit both of our fields.</Paragraph>
    <Paragraph position="1"> In this paper we presented an approach towards solving the problem of building a large scale lexical database with specific emphasis on near-synonymy. As we are still at a very early stage of our investigation, many unanswered questions remain.</Paragraph>
    <Paragraph position="2"> First of all, while we have confidence that our 'informant-friendly' intersubjective approach can extract good semantic information, we are less sure that all the information thus collected can always be easily converted into some numeric format, or be intuitively representable in some visualized form.</Paragraph>
    <Paragraph position="3"> Another concern has to do with the assumption that the informant would intuitively perceive visual similarity as semantic similarity. As this has not yet been tested, we simply do not know if it will work as hoped.</Paragraph>
    <Paragraph position="4"> We also wonder if purely visual design factor (e.g. color clashes, compositional imbalance) could inadvertently skew an informant's judgment on a particular display's semantic &amp;quot;naturalness&amp;quot;.</Paragraph>
    <Paragraph position="5"> Lastly, doing survey of any sort on the Internet involves a whole set of issues that we are aware of, but have not yet seriously investigated.</Paragraph>
    <Paragraph position="6"> Despite these uncertainties, we are in general very optimistic about the direction we are heading. We envision the next phase of our research to involve scaling up the testing, to include both more word groups and many more (in the thousands, ideally) informants, possibly via the Internet but more probably a large university's internal network, in our first venture into the Web-based survey world.</Paragraph>
    <Paragraph position="7"> We would also like to do some more experiments with the visualization tool, e.g. to try out different schemes for calculating semantic distances, to use data from other databases (and in other languages, e.g. English), or to create a more appealing, 3-D game like user interface. Perhaps even a &amp;quot;space war&amp;quot; type game for near-synonymy. Maybe.</Paragraph>
    <Paragraph position="8">  Appendix A. &amp;quot;Nag&amp;quot; Verbs in Finnish jupista mutter, mumble; grumble jurnutta (colloq.) annoy, vex jakattaa (colloq.) [yakety-] yak; nag marista whine, whimper, fret, grumble marmattaa grumble motkottaa carp, nag mukista grumble, grouse nalkuttaa nag, carp napista grumble, gripe, murmur nurista grumble purnatta grouse, grumble ruikuttaa whine, whimper, complain, (colloq.) moan, wail, (colloq.) pester urputtaa -- N/A --*  vaikeroida moan, groan, wail, lament, bemoan null valittaa groan, moan, wail, lament, complain null voihkia groan, moan voivotella moan, whine, bewail (edited from Finnish-English General Dictionary 1984) * The word urputtaa is not yet found in current Finnish-English dictionaries, though it has been collected into the more recent monolingual Finnish dictionaries.</Paragraph>
  </Section>
class="xml-element"></Paper>