<?xml version="1.0" standalone="yes"?>
<Paper uid="W97-1103">
  <Title>Self Organisation in Vowel Systems through Imitation</Title>
  <Section position="4" start_page="0" end_page="0" type="intro">
    <SectionTitle>
2 Self-Organisation
</SectionTitle>
    <Paragraph position="0"> Quite often spontaneous order can emerge in systems that are not controlled centrally. An example of this is the construction of a honeycomb. No isingle bee (not even the queen) has control over the building behaviour of the whole swarm.i Still, a very regular pattern of hexagons eme'rges. This happens because bees start to build cells at a certain distance from other bees that build cells. After a while they will encounter the neighbouring cells. Thus a pattern of hexagons emerges. Other examples of the outcomes of self-organising processes are termites' nests, sand dunes and the formation of paths.</Paragraph>
    <Paragraph position="1"> All self-organising systems have a large number of constituent parts that interact on a small scale. Order emerges on a large scale. This order is obtained from initial random behaviour of the constituer~t parts through positive feedback processes. Th'ese feedback processes cause the constituent parts to settle collectively in a certain state, once an accidental majority of them happens to be in that state. The field of &amp;quot;artificial life&amp;quot; is concerned with the investigation of self-organisilng processes that are inspired by living systems:through computer simulations.</Paragraph>
    <Paragraph position="2"> The approach that is followed in this paper and that was introduced by Steels (Steels, 1997b) is an artificial life approach. Language is a self-organisi~ig process. It exists in a community of speakers, and persists through the interactions of the speakers. No individual has central control over the language and no individual speaker is necessary for the persistence of the language. They are born and they die and still the language remains more or less continuous over time and throughout a population.</Paragraph>
    <Paragraph position="3"> The computer simulations that are presented here model linguistic interactions in an artificial life way. This means that the emergence of order in a population of agents (small computer programs that can operate autonomously) is studied. The agents are able to produce and perceive speech sounds in an approximately human way, they have only local knowledge (i.e.</Paragraph>
    <Paragraph position="4"> about their own speech sounds) and engage in local interactions with only one other agent at a time. It will be shown that phenomena that are also found in human vowel systems emerge.</Paragraph>
  </Section>
class="xml-element"></Paper>