<?xml version="1.0" standalone="yes"?>
<Paper uid="W04-0404">
  <Title>Translation by Machine of Complex Nominals: Getting it Right</Title>
  <Section position="2" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> Noun-noun (NN) compounds (e.g. web server, a1 a2a4a3a6a5a8a7 kikaia3hoNyaku &amp;quot;machine translation&amp;quot;,1 the elements of which we will refer to as Na9 and Na10 in linear order of occurrence) are a very real problem for both machine translation (MT) systems and human translators due to: constructional variability in the translations: a1a11a2a12a3a13a5a14a7 kikaia3hoNyaku &amp;quot;machine translation&amp;quot; (N-N) vs. a15a17a16 a3a19a18a21a20 miNkaNa3kigyou &amp;quot;private company&amp;quot; (Adj-N) vs. a22a8a23 a3a25a24a8a26 kaNkeia3kaizeN &amp;quot;improvement in relations&amp;quot; (N in N); lexical divergences in Japanese and English: a27a29a28 a3a30a21a31 haifua3keikaku &amp;quot;distribution schedule&amp;quot; vs. a32a34a33 a3a30a21a31 keizaia3keikaku &amp;quot;economic plan/programme&amp;quot; vs. a35a17a36 a3a30a21a31 shuyoua3keikaku &amp;quot;major project&amp;quot;; semantic underspecification: compounds generally have multiple interpretations, and can only be reliably interpreted in context (Levi, 1978); the existence of non-compositional NN compounds: a37a11a38a40a39a41a3a13a42a40a43 idobataa3kaigi &amp;quot;(lit.) well-side meeting&amp;quot;, which translates most naturally into English as &amp;quot;idle gossip&amp;quot;; high productivity and frequency In order to quantify the high productivity and frequency of NN compounds, we carried out a</Paragraph>
  </Section>
class="xml-element"></Paper>