Murray's filing system (OUP Museum)
Enter Keywords:
Thursday, 31 January 2013
Home arrow OED editions, updates and revisions arrow Case study: terms for lesbian(ism)
New search tools
New search tools on OED Online
The new search mechanisms on the re-launched OED Online facilitate a truly remarkable range of ways of accessing the vast quantity of data this Dictionary contains. Users should however be aware that the results returned in searches represent the content of very different stages of compilation of OED - i.e., they turn up a combination of old and new scholarship. Moreover, the data searched is not stable: every quarter, the identical search will produce a different set of results, as the lexicographers upload a new batch of revised entries to the Dictionary and remove the corresponding unrevised ones. 

No warning about either of these features is currently offered on OED Online itself.

Some of the associated problems with the new tools can be illustrated by looking in more detail at two of the search mechanisms prominently displayed on the front page of the OED's website, Timelines and Sources.

If we accept the website's front-page invitation to 'discover when words entered the English language' and click on Timelines, we see a graph which merges the (currently) 63% uncorrected datings of OED2 with the 37% corrected datings of OED3.[1] The user has no means of working out which is which, and no means therefore of making any sensible judgement about what the graph tells us: it represents an undistinguishable mixture of old with new scholarship. For example, we know from other sources that the original OED underquoted the 18th century and over-quoted the 16th, and that the new OED, in its revision of entries, is seeking to remedy these sorts of unevennesses (see our pages on Period coverage and the eighteenth century). But the material under Timelines makes no mention of this, and gives us no information either about the original biases or about how and to what extent OED3's revisions are bumping up the 18th-century deficiency (and perhaps correcting the 16th-century over-supply). The same problem applies to all the sub-categories under which one can search the new OED using the Timelines resource, whether by subject, region, or language of origin: the search is applied to a database which mixes old scholarship with new. The results, rich with significance as they appear, can only be interpreted if one has access to information (on what data are new and what old) which OED at present does not supply.

As it happens, EOED provides some indication of the direction of travel being taken by the revision where period coverage is concerned, since in 2005 we searched the online OED2 and OED3 and compared the number of new with old quotations for the 16th, 17th, 18th and 19th centuries (see graphs and discussion in our page on OED3's treatment of 1500-1899 here). Since this research was conducted, however, OED3 has probably doubled the number of revised entries. But at the moment it is impossible for us or anyone else (outside OED) to repeat our 2005 searches, and see what further changes OED3 has made to the chronological record over the last six years, since the online OED2 has been taken down from the website.

Similar issues arise when one looks at the results of searches for individual authors or texts. The fascinating material lodged under Sources, one of the OED Online's front-page search buttons, leads one to suppose that one may easily generate a list of words and/or senses first used by Shakespeare (and by Chaucer, Milton, Walter Scott, and so on). Again, nothing on the site warns us that these lists merge the search results from unrevised OED2 with revised OED3, producing a hybrid mish-mash.

Ironically, the problem is exacerbated by the fact that OED3's admirable new research on sources has in some cases completely overhauled the out-dated scholarship represented on OED2 (deriving from the Victorian period). Take Shakespeare, for example: his plays have been re-dated, his texts re-scrutinized, entries quoting his works have been re-configured (words previously treated in the body of one entry have been pulled out and assigned new independent entries), new words have been identified (e.g. compounds like new snow or Life-in-Death), authorship newly attributed (Two Noble Kinsmen is now identified as part of Shakespeare's oeuvre) - and of course many words which OED2 credited to Shakespeare as first user have now been antedated and ascribed to other, earlier sources (around in one in three; for an account of OED3's treatment of Shakespeare and the effect on our understanding of his lexical creativity, see Brewer 2012 forthcoming or download podcast on the topic from the Oxford University website here).

Running together the Shakespeare first citations as recorded in OED3 with those recorded in OED2 gives us data that is neither fish nor fowl, difficult to work with in any way that produces meaningful analysis either of OED's original account of Shakespeare or of the new edition's vastly changed one.

In particular, the removal of OED2 from the website (see discussion on previous page), as an independently searchable entity, has deprived users of any means of differentiating between new and old material in search results. It has thus prevented us from being able to appreciate the quality and consequences of OED3's new scholarship.

[1]  As of December 2011. The proportion changes very slightly every quarter, as the lexicographers upload a new tranche of revised entries to OED Online and remove the corresponding old entries.
Last Updated ( Wednesday, 28 December 2011 )
< Previous   Next >

Built with Mambo. Any comments or feedback are welcome.
All responsibility for views and data published on this site is that of the author, Charlotte Brewer.
Copyright © 2005-13 Charlotte Brewer. All rights reserved.