Top sources in OED3

This page should be read in conjunction with the previous page on Top sources in the OED up to 2000. Here we discuss the transformation now being made to the existing record by OED3, so far as can be discerned from the OED Online website. These are firstly the new predominance of multi-authored newspapers, journals and periodicals as quotation sources, and secondly OED3’s Changes to top sources in OED1/OED2.

As we conclude at the foot of the page, OED evidence tells us about the (changing) methodology of the lexicographers as well as about the history and development of the English language.

Top sources in OED3 – new predominance of newspapers, journals and periodicals

The OED3 revision has been underway since the late 1990s and is approaching the half-way point in re-writing its original entries. One of the tricky issues the lexicographers will have had to consider is the question how to re-balance the heavy quotation from individual sources in OED1, clearly chosen for (what we would now see as) cultural rather than linguistic reasons. See their own account of this in their pages on Documentation in Preface to the Third Edition, 2000 [accessed July 2019], which we quote and briefly discuss at Policy and practice.

Since December 2010, with the transfer to a new website, OED Online has published a list of ‘The top one thousand authors and works cited in the OED’, reached by clicking on the ‘Sources’ link on the front page. The list changes every quarter as a series of revisions and new entries is uploaded to the website.1

It is fascinating to look at the results so far. The screenshot below shows the first 18 of the top 1000 sources as of June 2019 (screenshot taken 5 August 2019).2 As always, OED Online data represents a combination of revised with unrevised entries, so that we cannot look at OED3 results separately. A clear pattern emerges nevertheless.

First 18 items on the list of ‘Top 1000 sources in the OED’ after the June 2019 update. Source: screenshot from OED Online 5 August 2019

It can be seen instantly that the main change in most quoted sources is the massive increase from multi-authored newspapers, periodicals and journals. Dates of publication stretch from the late 17th century to the present for Philosophical Transactions, from the late 18th century onwards for The Times, and from the mid-19th century onwards for the other sources of this kind, so the main impact of these new quotations will have been on the documentation of 18th-century vocabulary (Philosophical Transactions; see discussion at 1700-1799 in OED3) and on 19th and 20th-century documentation (remaining journals).

Digital access to these sources will have enabled the revisers to comb systematically, database by database, through a wide range of linguistic domains, with a degree of thoroughness and consistency not possible to their predecessors (obliged to work through individual texts by eye alone). The newly visible sources of this type here are The Times (1788-), Encyclopedia Britannica (1768-), New York Times (1857-), Daily Telegraph (1855-), Nature (1869-), and Guardian (1959-); on screen the list continues with the Daily News (1846-), Westminster Gazette (1893-1928) and Blackwoods Magazine (1817-1980), Science (1883), the Independent (1986), Harpers Magazine (1850-), Washington Post (1877-), and the Listener (1929-1991), London Gazette (1666-), and Lancet (1823-).

Many of these sources were also quoted in OED1, but in much lower numbers. One of the most striking changes in rate of citation is from The Times, quoted 4,085 times in OED1 (according to Willinsky 1994: 214). That figure has now risen nearly 10-fold to 43,073, a staggering increase.

Notwithstanding these huge additions from multi-authored sources of the (mostly) 1800s onwards, the continued predominance of OED1’s favourite main authors in today’s OED has not shifted. It is certainly the case that the OED3 revisers have quoted extensively from many other individual sources, each in much lower numbers.3 Nearly half-way through the OED3 revision, however, it still looks as if the Victorian cultural canon is dictating the outline shape of documentation in the OED. To some degree, this may be due to the easy availability of the relevant texts on Oxford Scholarly Editions Online, as described at Links on OED Online.

Changes to top sources in OED1/OED2

Chart 27 below shows the changes so far being made by the revision to OED1 and OED2’s representation of top sources in the Dictionary. Numbers in brackets after the name of the source indicate the respective position on the OED Online list of top sources after the quarterly update in June 2019.

Chart 27: Top quotation sources in OED1/2 compared with OED Online as of June 20194

*See footnote 4 on Bible quotation figure

The range of variation in OED Online’s changes to OED2 totals can be seen more easily in Chart 28:

Chart 28: OED Online’s changes to OED1/2’s top sources as of June 2019

The most striking change to OED1(2) is the increase in quotations from Philosophical Transactions, discussed at 1700-1799 in OED3. The other changes to OED’s representation of the Bible and of the other individual literary authors, however, are also of note – and it’s important to bear in mind that many changes in documentation will have occurred without much alteration to the overall number of quotations counted in electronic searches: the quotations now included in OED3’s revised entries may not be the same as those chosen for OED2. (For a detailed account of this phenomenon in the treatment of Auden and Joyce, see Brewer 2010a: 116-24).

As ever, given the limitations of OED Online website, it is impossible to identify either the newly added or the newly deleted quotations in these revised entries.  Nevertheless, we can hazard some guesses about the causes behind the more obvious changes in the OED record.

Where Chaucer and Cursor Mundi are concerned, both the increase (Chaucer) and decrease (CM) are likely to be due to OED3’s new reliance on the Middle English Dictionary (MED), as discussed at 1150-1499 in OED1/2 and 1150-1499 in OED3. MED especially favoured Chaucer as a quotation source (hence the proliferation of new quotations from his works in OED3), while at the same time covering a much wider range of early Middle English sources than had been possible for OED1 (hence the reduced testimony in OED3 from CM).

OED3’s access to numerous Early Modern English sources via EEBO and other databases also explains the decline in quotations from the Bible: the lexicographers will have found other examples of a word’s use which – for one reason or another – they deemed more suitable than the Bible quotation included (perhaps as a result of concordance searching) by the OED1 lexicographers.5 The same phenomenon must explain the loss of around 370 Shakespeare quotations too, albeit to a lesser degree: while revising, OED3’s lexicographers will have found better examples of a word’s use in sources less easily accessed (and/or less prized) by their predecessors, and decided to use these instead, jettisoning Shakespeare’s ones in the process.

In contrast, look at the remarkable rise in quotations from Dryden (542), Dickens (1,029) and Scott (1,802). Does this tell us something newly interesting about the intrinsic characteristics of these writers’ use of language? On the limited evidence available, the answer is probably no. In all three cases, the bulk of the new quotations had been added by 2010, with far fewer additions between 2010 and June 2019. It is very unlikely that this was something to do with the linguistic evidence offered by these texts: rather, the lexicographers decided at some point to cease searching in the works concerned for new lexical evidence and instead turned their attention elsewhere, probably choosing instead to search the many new digitized sources of non-literary texts coming into existence from the mid 2000s onwards.

The relevant quotation totals for these three authors in OED2, in OED Online as of December 2010, and in OED Online as of June 2019, are as follows:6

Quotation source (bracketed number = June 2019 ranking)OED2OEDO Dec 2010OEDO June 2019Changes from OED2 to OEDO Dec 10Changes from OEDO Dec 10 to OEDO Jun 19
Walter Scott (3)1531516982171171667
Dryden (15)880592609347455
Dickens (16)8200920592291005

The conclusion to draw from these patterns of quotation citation is a cautionary one. In some respects, OED’s evidence tells us as much about the methodology of its makers as about the history and development of the English language.

Last updated on


  1. No archived copies of past versions are viewable. The author has kept copies of every such list from December 2010 onwards, cut and pasted from the OED Online website into excel spreadsheets, which she would be happy to share with any researcher.
  2. Also discussed at Policy and practice.
  3. Reported from conversation with John Simpson, pre-2013.
  4. The figure for OED Online’s quotations from the Bible, 21,315, has been arrived at by adding together the six separate totals provided on the OED Online website’s list of 1,000 top sources as of June 2019 for individual versions of the Bible, as follows: the Wycliffe Early Version (a 1382, 8,333 quotations), Wycliffe Late Version (c 1425, 1505 quotations), Tyndale (1526-1534, 1,989 quotations), Coverdale (1535, 4327 quotations) Douay (1609-1610, 650 quotations), and King James (1611, 4,511 quotations).
  5. On OED1’s habitual reliance on concordances, see Onions 1928,‘How the Dictionary is Made’.
  6. See footnote 1 for the source of these figures.