<?xml version="1.0" standalone="yes"?>
<Paper uid="W06-2713">
  <Title>Representing and Accessing Multilevel Linguistic Annotation using the MEANING Format</Title>
  <Section position="3" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> It is well known that when using XML-based annotation schemes to represent multi layer annotations, it can be difficult to handle partially overlapping annotations. Annotating discontinuous elements may be considered as a variant of the same problem (Pianta and Bentivogli, 2004).</Paragraph>
    <Paragraph position="1"> Other difficulties can arise from the necessity of integrating manual and automatic annotations, as we will show in this paper.</Paragraph>
    <Paragraph position="2"> One of the most effective solutions to the above mentioned problems is the so called stand-off annotation, based on the separation between textual data and annotations, and between various types of annotation, possibly pointing to same text. This approach has been systematically adopted in the design of MAF, a multilayer XML format developed for the EU-funded MEANING project, in the context of the creation of the Italian MEANING Corpus (Bentivogli et al., 2003).</Paragraph>
    <Paragraph position="3"> In this paper we will describe our experience in the use of MAF, with special emphasis on how we solved issues related to representing annotation levels which come from different sources, and can possibly overlap. We will also give details about the solutions we adopted to allow for efficient access and human browsing of MAF standoff annotations.</Paragraph>
    <Paragraph position="4"> The rest of the paper is organized as follows.</Paragraph>
    <Paragraph position="5"> Section 2 describes MAF and the types of annotations which have been represented with it. Section 3 reports on the integration into MAF of linguistic annotations coming from different sources. Section 4 illustrates the strategies adopted to make the information encoded in MAF quickly accessible. Finally, Section 5 presents the MEANING Browser, a tool for accessing and navigating corpora linguistically annotated with MAF.</Paragraph>
  </Section>
class="xml-element"></Paper>