Renear et al. examine the thesis that "text is an ordered hierarchy of content objects," and they pursue this thesis through pragmatic, empirical, and theoretical arguments. The authors trace what is framed as a technical encoding problem to "a deficiency in our understanding of just what we are doing when we prepare an encoded text," and they make explicit some of the assumptions behind current encoding practices.
This paper discusses Standard Generalized Markup Language (SGML) determinations of document types and objects, logical and physical structures, and the perspectives of the Text Encoding Initiative (TEI) for determining text objects, as well as how SGML and TEI contribute to the project of seeing and using hierarchies to encode machine-readable text. The authors determine that because of the SGML feature CONCUR, the problem of "overlapping" (or less-than-unique strategies for encoding text) is ultimately not a practical problem for encoding projects.
Their devotion to an even-handed discussion includes
"counterexamples," refinements of the initial thesis, a careful
discussion of terminology, and two interesting considerations of
enjambment and strikeout as possible contradictions to their
thesis. They conclude that theories, methodologies, and
analytical practices, termed "perspectives," are at least as
important as genre in the identification of text objects; that
perspectives frequently determine hierarchies of objects; and
that nonhierarchical perspectives can often be decomposed into
hierarchical subperspectives. (Jean Jacobson.)
Michael Hancher Department of English, University of Minnesota URL: http://umn.edu/home/mh/ebibjj5.html Comments to: mh@umn.edu Created 13 June 1995 Last revised 17 September 1996