Back to contents
The integration of the text of the Supplement into that of the first edition of the OED (referred to below as ‘the OED’ for brevity's sake) was carried out in two stages. The major processes were performed automatically by computer programs specifically developed for the purpose. Printouts of the resulting merged entries were then edited by lexicographers, and the emendations marked on them were entered by keyboarders into the computer.
The guiding principle of both stages of the integration process was that the intentions of the Supplement should be faithfully followed unless there were very good reasons for departing from them.
About 42,000 main entries in the Supplement were new and independent. They were put into their appropriate place in the alphabetical sequence.
A new entry having a headword with the same spelling and part of speech as one or more in the OED already had a different homonym number: it was placed after its OED homonym(s). If there was only one homonym in the OED, this was given the homonym number 1. A new entry having a headword with the same spelling but a different part of speech was placed in an appropriate position among the pre-existing entries, usually towards the end of the sequence (see the principles of entry ordering, below); where only a noun (‘substantive’) or nouns occurred in the OED, they had no explicit part of speech: the part of speech ‘sb.’ was supplied.
Many combinations and derivatives treated, in the first edition, within the entries for the main words on which they are formed were elevated to the status of main words in the Supplement. The usual reason for this was a significant increase in the complexity of the senses of the word or its general acceptance as more than a casual or obvious compound of the root word. A certain number of forms treated as graphic or spoken variants of a main word in the first edition were similarly registered as separate, independent words in the Supplement. When such upgrading occurred, the material relating to the word in the first edition was not reprinted in full in the new entry; instead, cross-references were used to direct the reader to the definition, etymology, or quotation in question, implying that it was now part of the new entry. Accordingly, in the integrated text this original material has been moved (with appropriate editing) to the new position. Much effort has been devoted to detecting overlapping entries of this sort, many of which are extremely well disguised, especially when linked by only a single, inconspicuous quotation.
A certain number of Supplement entries directed the reader to delete the existing Dictionary entry and substitute a new one, the text of which then follows. Not infrequently, these entries also borrowed snatches of text from the former entry, especially the quotations given there. Before the original entries were discarded, the borrowed text was transferred to the new entries.
There were also entries in the Supplement which added further information to a different main word in the first edition, but nevertheless did not require a new entry to be created. There were two principal kinds. First, where the entry for a main word in the first edition had a derivative listed under it and the only addition to that entry in the Supplement related to the derivative, the latter was given, as it were, the temporary status of a main word solely for the purposes of supplementation. In such cases, the new material has been accommodated within the framework of the original main entry and the supplementary one has been discarded. Secondly, a variant form of a main word could be treated in a separate entry with illustrative quotations, because the variant in question either had not been listed and illustrated under the principal form, or had recently become recognized as the predominant spelling. In both cases the intention was that the two entries should be merged, in the latter with a change of spelling in the headword. Again, detecting such pairs of entries (and some forming larger groups) has been very difficult, especially as the spelling variation can bridge widely separated parts of the Dictionary (C and K, or E and Œ, for example), a factor which has occasionally frustrated the fulfilment of the Supplement's intentions.
The remaining 28,000 entries in the Supplement had counterparts in the OED with which they were integrated. Four basic procedures were followed, in line with the main instructions to the user included in each entry: addition, deletion, transference, and substitution. Naturally not every one of the thousands of occasions when any of these changes was made is explicitly indicated in the text, but the intentions are nearly always quite clear.
In the headword section of the entry (the identification), the usual supplementary material consisted of a pronunciation or one or more variant forms or inflexions. Deletions were mainly of status labels (especially ‘obsolete’). Transference of labels into a sense section was common. Substitution of a different spelling of the headword or a modified label occurred from time to time.
Etymologies were occasionally added or modified by deletion or substitution.
In the signification, additions made up the bulk of the changes occasioned by the Supplement. Where a complete new sense section was added, its numbering normally indicated the correct position. If the first edition had only one sense section at that level in the hierarchy, the absent sense number or letter (the first in the series: A, 1, a, or (a)) was supplied. Supplement sense sections introduced by starred numbers were added at the appropriate point in the sequence, and the ensuing sense divisions were renumbered to accommodate them. When a section division (headed by a capital letter), covering the use of a word in a new grammatical category, was added at the end of an entry, the identification was altered to include the new part of speech. Sections containing derivatives were added at the end of the entry.
Supplementary material was frequently added to definitions; this was commonly made plain by the introductory word ‘also’, which sometimes became superfluous and was edited out. Occasions when the new material could merely be added, without modification, at the end of the existing definition were relatively rare. Often the new material had to be inserted within the old, or the whole rewritten into one definition. The domain of a label had sometimes to be restricted to cover only the intended portion of the definition. So multifarious were the modifications that the results can scarcely be distinguished from the deletions, transferences, and substitutions indicated from time to time by the Supplement.
The parts of the Dictionary which the Supplement especially augmented are the sections in which defined and undefined combinations are listed. Considerable labour was required to merge the corresponding lists (usually in alphabetical order). The defined combinations, resembling as they do complete entries in miniature, were susceptible to supplementation, deletion, modification, and substitution in any of their constituent parts. Their alphabetical sorting and rough merging was achieved automatically, but much work remained for editorial attention.
The addition of quotations was very frequent: when an entirely new sense-division was added, with its own group of quotations, when a sense section was supplemented, and as the sole modification to a sense section. In most cases this was a straightforward operation. Paragraphs of quotations that illustrate combinations, however, usually consist of a series, arranged alphabetically to correspond with the combinations, of short chronological sequences of quotations; the accurate fusing of parallel series of these was a complicated operation, requiring careful editing, both on paper and at the screen.
Integration of the Supplement with the main Dictionary text involved an enormous number of changes to the ‘addresses’ at which sections of text were located: namely, to the form of the headword, the part of speech associated with it, the homonym number distinguishing otherwise identical main words, or the number or letter identifying the sense-division under which a definition, combination, or quotation could be found. As a result, there was a danger that many cross-references would become invalid, since the elements to which they pointed would now be differently identified. This was countered by a twofold strategy. A computer program caused every change to the ‘address’ of an element brought about by the process of automatic integration to be applied to the cross-references which cited that address. Editorial staff subsequently ensured that changes made by them were logged and then applied to the corresponding cross-references. Inevitably, a few cross-references escaped both systems, including those in early volumes which were already passed for press before editorial changes that would affect them had been made in later volumes. There are around 600,000 cross-references in the OED, of which well over 20,000 have been adjusted in response to integration, and many more have been corrected or rendered more precise.
|
Back to contents |
|