@phdthesis{SmithThesis,
  author = "Eric J.M. Smith",
  title = "Query-based annotation and the Sumerian verbal prefixes",
  school = "Department of Linguistics, University of Toronto",
  year = "2010",
  abstract = "<p>The study of Sumerian has traditionally been carried out in isolation from mainstream linguistics,
              thus limiting our ability to understand the language and to situate it in a cross-linguistic
              context. This dissertation shows how the tools of corpus linguistics and modern syntactic theory
              can be gainfully applied to Sumerian.</p>
              <p>Existing corpora of Sumerian texts are largely lacking in morphological annotation, with
              query facilities consisting only of basic string searches. Two existing corpora (one completely
              unannotated and one tagged for part-of-speech) are given morphological annotation using a
              process of query-based annotation. A query language (based on CQL and XPath) is used to
              query this corpus, and as queries are made, the results are tagged so that the resultant query
              objects can be used as the basis for subsequent queries. In this fashion a morphologicallyannotated
              corpus is built up without having to rely on the services of a skilled annotator.</p>
              <p>This annotated corpus is then used to provide evidence for two important problems in
              Sumerian morphosyntax: the dimensional prefixes and the conjugation prefixes. The dimensional
              prefixes, which have previously been considered to represent concord between the verb
              and the associated nominal phrases, are shown instead to be a system of applicative heads
              which serve to introduce the verb’s arguments. The conjugation prefixes, whose purpose has
              been the subject of a century of debate, are shown to be the manifestation of inner aspect
              features which express the speaker’s perspective on the structure of the event.</p>
              <p>By using a corpus to provide the underlying data and by considering Sumerian morphosyntax
              in light of cross-linguistic evidence and modern syntactic theory, previously misanalysed
              aspects of Sumerian are shown to have analogues in other languages. The dimensional prefixes
              and conjugation prefixes are not oddities specific to Sumerian, but represent variations on
              morphological systems found elsewhere.</p>
  download = "ftp://ftp.cs.toronto.edu/pub/gh/Smith-thesis-2010.pdf"
}