Orð og tunga - 01.06.2001, Síða 64

54
Orð og tunga
inherent tendencies with regard to how long they take to edit and how much space they
occupy on the printed page. Verbs and prepositions/adverbs notoriously take longer and
are harder to reduce to comprehensible and useful dictionary articles than a great many
nouns. Both nouns and verbs that are represented by very few tokens, can, however,
be exceptionally time-consuming in terms of arriving at a semantic description, simply
because there are “too few” citations — insufficient empirical material — on which to
base an immediately convincing definition or explanation. Factors of this type need to
be taken into consideration when planning the future production of the dictionary, year
by year and volume by volume. The dictionary is regrettably produced in alphabetical
order, so it was also evident that a list of this type would be a useful tool in enabling
editors to relate words that are semantically or in other ways connected to one another,
even though they are widely separated in terms of their position in the alphabet; most
specifically cross-references could be incorporated in the dictionary in a reasonably
consistent manner.
A further benefit is the advantages that the list could provide for anybody engaged
in research on early Nordic languages and culture (see §8 below).
Because of a strict production schedule it was evident that the regular dictionary
staff did not have the time available to produce a word-list of the type described above.
The Arnamagnæan Commission applied to Lýðveldissjóður in Reykjavík for additional
funds and in 1998 received a generous grant which provided the stimulus for work to
be started. Further grants were made in 1999 and at the beginning of the year 2000,
such that all the costs involved were generously met by Lýðveldissjóður. The project
description entailed a full year’s work for a person well-qualified in Old Icelandic and
0.85 man-years of work for a secretarial assistant. Jóhannes Bjarni Sigtryggsson was
appointed from amongst a number of applicants for the post as the scholar responsible
for the project and he commenced work in Copenhagen on September 15th 1998; Alex
Speed Kjeldsen was employed as the principal student assistant and other assistants
were subsequently brought into the project on the secretarial level.
3 The word-list and ONP’s computer system
When the project started, all of the letter a and almost all of the letter b had been edited
in final form, so for them no new word-list was needed. To understand further how
the word-list was produced it is necessary to have some insight into the way in which
the dictionary’s database system operates. It is also necessary to appreciate that only
a small percentage of the dictionary’s total collection of citations has been keyed in,
this process having only got as far as the word en. All of the dictionary’s materials
are in reality independent, but structurally correlated, database files.4 The database
system incorporates a program that can, at any stage of the editorial process, produce
printouts for assessment and proofreading. When an operator instructs the database to
produce the edited copy for a given sequence of words, the database first consults its
4Cf. Degnbol et. al. (1992:385-389).