Orð og tunga - 01.06.2012, Page 11

Matthew Whelpton
From human-oriented dictionaries
to computer-oriented lexical
resources - trying to pin down
words
1 Introduction
Dictionaries are one of the most familiar linguistic resources to which
an ordinary native speaker of a (standardised) language is likely to
have access; indeed the process of dictionary creation has served a
crucial role historically in the standardisation of European languages
and represents an important activity in the creation and maintainance
of standards today. In the age of information technology, the succes-
sors of the paper dictionary have continued to exert great influence,
playing a central role in the field of natural language processing (nlp),
supplying text and speech processors with essential information on
word form, category, and meaning in activities as diverse as gram-
matical parsing, information extraction, and machine translation.
Although some of these nlp resources are essentially electronic
versions of paper dictionaries, the demands of computer applications
are in many ways much greater than those of human users: comput-
ers require that information be presented in a manageable format for
algorithmic manipulation (e.g. in a relational database where each
piece of information is classified and linked explicitly to others) and
that the information itself be systematic and absolutely explicit. The
Orð og tunga 14 (2012), 1-18. © Stofnun Arna Magnússonar í íslenskum fræðum,
Reykjavík.