<?xml version="1.0" standalone="yes"?>
<Paper uid="W02-0607">
  <Title>Modeling English Past Tense Intuitions with Minimal Generalization</Title>
  <Section position="2" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> In recent years, linguists have explored theoretical models of how speakers discover the rules of their language. Automated learning systems can be of great use in developing such models. The predictions of a theoretical model can be fully tested only when it is formalized explicitly enough to be implemented. null In our research, we have developed and implemented a model for discovering rules of morphology and phonology. The model is trained on pairs of morphologically related words, and learns the rules by which one form of a pair can be derived from the other. We have tested our model by comparing its predictions against intuitions gathered experimentally from human speakers.</Paragraph>
  </Section>
class="xml-element"></Paper>