Indexes and inconsistencies

Indexes of quotation sources have been suggested at various points in the OED’s history, though it seems the idea has never been put into practice in any significant way. This page discusses mooted indexing for compiling the OED1 and the Supplement, followed by a tentative consideration of such possibilities in OED3.1

OED1 and the Supplement

Looking back in 1989 at the early days of compiling his Supplement, embarked on in the late 1950s, Burchfield wrote:

the entire works of writers like Eliot, Auden, Joyce, Lawrence, and many others, needed to be indexed in the manner that the readers of sources drawn on for the OED had indexed the works of Chaucer, Malory, Marlowe, Shakespeare, Milton, Johnson, and all the other famous writers of the past.

Burchfield 1989: 8

Elsewhere he picks out Edith Sitwell, Edward Thomas, Dylan Thomas and Roy Fuller as such authors (e.g. OEDS4 1986: x).

Setting aside the question of how such writers should be identified (see The canon), indexing authors in this way was specifically disclaimed by Murray: ‘forming complete verbal indexes to all books…is not only impossible, but the results would be unmanageable’, as Murray told the Philological Society in 1884, although his assistants had made a start on some books in 1879.2

It is possible that the first OED editors established an index for Shakespeare’s writings, which are almost exhaustively cited in OED – not as difficult as it sounds, given the concordances and glossaries available – but there is no direct evidence that they did so, and no evidence now survives of indexes to any writer or work, though one had originally been planned for the writings of Burke (see account at Initial practice). Understandably, the lexicographers seem to have fallen well short of a complete record of any other writer, even those they quoted most, such as Chaucer and Milton. The editors did on the other hand make regular use of concordances to ‘the Bible and other poets’, as Onions describes in ‘How the Dictionary is Made’, published in The Periodical in 1928 (Onions 1928a; click on the link to read). Naturally, the writers for which such aids were available were the major canonical ones (already, therefore, heavily excerpted by readers and likely to be cited in the Dictionary): Shakespeare, Milton, Chaucer, Pope, Cowper, Walter Scott, Tennyson.3

Clearly, indexing would have been equally impracticable for the second Supplement. There is no indication that it occurred; indeed, in the Introduction to Volume 1, Burchfield apologizes for his policy of ‘liberally representing the vocabulary of such writers as Kipling, James Joyce, and Dylan Thomas’, as against the OED’s ‘policy of total literary inclusiveness for the earlier centuries…[of] all the vocabulary, including hapax legomena, of such authors as Chaucer, Gower and Shakespeare’. This last remark is baffling. Any regular user of the parent Dictionary can attest that OED did not, and could not, achieve ‘total literary inclusiveness’.

Inconsistency was unavoidable in a policy such as this, and Burchfield freely acknowledged that he changed his mind as he went along, for example deciding in 1973 – after the publication in 1972 of his first volume, treating the letters a-h – to include all rather than just some Jabberwocky coinages, with the consequence that, of this group of words, only borogove, callay, callooh, frumious, and gimble are omitted from the Supplement (Burchfield 1974: 13). He also explained that ‘the pattern of admission’ of such words

was governed as much by the choice made by the readers as by any abstract principles adopted by the editors. If a reader made a slip for such an item it was likely to be included, with small regard for consistency in comparable words, or in words drawn from other writers, in other parts of the Dictionary. Conversely a word that was not copied by a reader had little chance of inclusion since the editorial staff would almost certainly be unaware of its existence.

Burchfield 1989: 894

The results can be seen in any Supplement author whose works are checked in detail against the Dictionary itself (e.g. Auden). Similarly, the usage of earlier poets and writers is in many cases inconsistently recorded in the first edition of OED.


The Preface to OED3 published online in 2000 (Simpson 2000) made no reference to the creation of indexes or use of concordances, though naturally the revisers will have used all manner of historical glossaries and indexed glossarial resources for every period of the language. In the early days of the electronic revolution (up to at least 2005), literary texts were particularly likely to be digitized, hence readily responsive to searches for lexicographical purposes, and this may explain the large number of new quotations that rapidly appeared in OED3 for canonical authors such as Joyce and Austen. See discussion of OED3 in The canon, also Online sources in the section on OED3 quotation collection.

Last updated on 8 October 2019


  1. The material on OED1 and the Supplement is drawn from Brewer 2007b: 128, 187-8.
  2. Murray 1884b: 516; 1879b: 572.
  3. For example, for Shakespeare, Becket (1787), Twiss (1805), Adams (1886), Bartlett (1894), Schmidt (1874); for Milton, Cleveland (1867) and Bradshaw (1894); for Pope, Abbott and Abbott (1875); for Cowper, Neve (1887); for Tennyson, Brightwell (1869) and Langley (1870); for Walter Scott, A Complete Glossary for Sir W. Scott’s Novels and Romances, 1974 (1833). On Shakespeare concordances in particular, see Karpova 2004.
  4. Cf. similar remarks in Burchfield 1989: 13, 84, and our pages on Reading and readers.