Issues and problems (OED1)
Tendency to record unusual rather than ordinary words, and words beginning with the same letter
The instructions given to readers in 1859 had asked them to compare their books against the ‘Bases of Comparison’ and only record items of vocabulary not already included in these texts. This had the effect of indicating that obsolete and unusual words were particularly desirable, and this meant, in turn, that such words were recorded at the expense of more ordinary ones.
Appointed editor in 1879, J. A. H. Murray swiftly recognized the mistake made by his predecessors. As he told the Philological Society in one of his regular reports on the Dictionary:
In my opinion, the Bases of Comparison formerly issued by the Society were a mistake, and detrimental to the work which they were designed to serve. Their most obvious result…was that while rare, curious, and odd words, are well represented, ordinary words are often most meagrely present; and the editor or his assistants have to search for precious hours for examples of common words, which readers passed by because they happened to find them put down in their ‘Basis’ as occurring in the Bible or in Burke. Thus of Abusion, we found in the slips about 50 instances: of Abuse not five…’ (Murray 1879b: 571-2)
Murray carefully re-wrote the original instructions to correct this bias (see his Directions to readers); no. 5 asks them to ‘make a quotation for every word that strikes you as rare, obsolete, old-fashioned, new, peculiar, or used in a peculiar way’, but no. 7 says ‘make as many quotations as you can for ordinary words, especially when they are used significantly, and tend by the context to explain or suggest their own meaning’.
As Elizabeth Knowles describes, another early problem had been Furnivall’s encouragement of ‘the silly practice by which sub-editors were urged to get their friends to read books for their own letter only – so as to provoke a rivalry of good works between them’ (Knowles 2000: 32). This resulted in uneven source coverage, which Murray had no means of identifying ‘until [and, one might add, unless] some accident reveals the fact that of late we have not seen any quotes from the book in question’; Murray complained at a Philological Society Dictionary evening in 1892 that this was ‘a most pernicious and deceptive practice’ (Murray 1892: 275).1
Disparities between readers and texts
Assigning a book to a reader did not necessarily result in a sheaf of usable quotations. The reader might have returned very few slips, or the slips returned might not have been deemed useful by the lexicographers when it came to the point of choosing which quotations to print in the dictionary. Material in the OED archives, as well as the ‘List of Readers and Books Read by them for the Dictionary, 1879-1884, with Approximate Number of Quotations Supplied’ (Murray 1884a), shows that some readers were phenomenally productive of slips, others not. Thus J. H. Chamberlain of New Haven, Connecticut, returned 60 slips from Whitney’s Life and Growth of Language, while H. P. Bull of Hereford returned 1,200 from Disraeli’s Vivian Grey. The former work was eventually quoted around 300 times (which meant that Bull’s original offering was supplemented either by another reader or by the lexicographers themselves); the latter over 460 times. It is impossible to know, without further research on the two texts, whether the Whitney work might not have yielded more useful data had it been read intensively in the first place.
These sorts of disparities mean that it would have been impossible for Murray to regulate strictly the proportions of material coming in from one period rather than another, or one type of source rather than another: although he certainly tried to do this. For example, Knowles describes how he particularly encouraged the reading of Lydgate by one volunteer, motivated by the belief that ‘he was in some respects the Dr Johnson of his century; in his free use of words from Latin (directly or through French), and being immensely read in the 16th c (far more than Chaucer…)’ (Knowles 2000: 35).
Individual preferences
It may also be that Murray was susceptible to strong pressure from individual readers to include certain sources. Letters from his old friend Robinson Ellis (the distinguished editor and translator of Catullus) indicate that he read every fascicle through as it appeared, counting up the instances his translation was cited and reporting his views on the wisdom of Murray’s choice of sources (Ellis is quoted around 565 times altogether, 542 times from his Catullus translation). Thus on 17 January 1906 (in a letter now preserved in the Murray Papers), Ellis wrote to Murray as follows:
I am pleased to find two quotations from Catullus in the new part of the Dict….If I were inclined to magnify this success, I might triumph in the thought that I had twice the whole number of quotations in this part which I had in the former…Swinburne is extraordinarily successful; Jowett too much; and I think there are too many quots from J. H. Newman. I am glad to see [James] Bryce [the radical politician, jurist and historian, also a correspondent of Murray’s] and R[obert] Bridges, occasionally, but not frequently.2
The influence of individual readers on the intensity with which certain sources were read (and hence the likelihood of their being quoted in the OED) continued into the 20th-century Supplement to the Dictionary. Its editor, R. W. Burchfield, wrote in 1989 that ‘the pattern of admission 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: 89; cf. similar remarks on pp. 13, 84).
Records
Record-keeping – and perhaps sufficient communication between the separate editors – was also a recurrent concern. In 1893, Henry Bradley, the first of Murray’s fellow-editors, wrote to him:
I am vexed to find that the expression flaunting Fabian occurs in Nash’s Lenten Stuff, a book that is supposed to have been read for the Dictionary. It is I suppose now too late to insert the quotation. I wish I knew absolutely what books one might feel sure had been exhaustively read. (Murray Papers: 1 July 1893)3
This experience has been replicated by modern scholars: McConchie shows time and again that ‘sources already scrutinized, and even relatively thoroughly excerpted, may nevertheless be productive of much more material’; ‘the fact of a book’s having already been read is simply no guide to what useful data might still be found in it, unless it can be shown to have been exhaustively excerpted as in the case of Shakespeare’ (McConchie 1997: 155, 177-8).
Technical problems
There were (and are) inevitable problems involved in reading a text ‘manually’ in search of items suitable for inclusion in the Dictionary. However conscientious and skilled a reader, there is a natural tendency for the eye to slide over familiar words and pick up unfamiliar ones. In addition, it is sometimes very difficult to identify unusual uses of words where there is no corresponding peculiarity of morphology.
An example is the verb mirror. OED1 (and OED2) dated its first occurrence 1820, quoting Keats’s Lamia, which might seem rather late for such an obvious derivative use of the noun mirror (itself attested from the Middle English period onwards). As observed by P. J. Wexler, a major contributor to both the Supplement and to the new OED, ‘It is a commonplace that a dictionary often records long gaps between the appearance of a root-form and its derivative – say between an adjective and an adverb; but is this a fact about language, or about lexicographers?’ (Wexler 1981). In this case, OED1’s record told us about (the first edition) lexicographers, not about language. When OED3 revisited this entry in 2002 they found earlier examples of the verb mirror not only in Nashe (1592) but also in Hoccleve (around 1410), in both cases in works read and quoted elsewhere in the Dictionary by OED1 lexicographers.
Murray was again well aware of the issue. On the difficulty of picking up the first occurrence of a word, he declared, ‘Earlier instances will, I doubt not, be found of three-fourths of all the words recorded, above all, of the words introduced from Latin since the Renascence’ (Murray 1884b: 517).
Last updated on 8 October 2019
Footnotes
- In the past Murray had sometimes been less despairing about the variable standard of reading, and his ability to correct oversights; for instance he recorded in 1880 that ‘As a whole the books have been satisfactorily read; and when, as has occasionally happened, the work of a reader of an important book has appeared to us inadequate, we have not hesitated to ask another reader to go through it again. This plan we have found it desirable to take with many of the books undertaken in former years, as to the completion of which we had no satisfactory evidence’ (Murray 1880: 122).
- Swinburne was eventually cited c. 1,050 times in OED1, Jowett just under 2,150 – over 2,000 from his translation of Plato alone, Newman c. 1,650, Bryce just over 700, Bridges c. 600.
- In fact Bradley was able after all to include Nashe’s quotation (see OED s.v. Fabian). He wrote again to Murray on 6 July 1893 to say, ‘I find the page with Fabian has not been printed off, so that I can put in the Nash quotation…Perhaps this late discovery may prove lucky, for it caused me to make another effort to find out the origin of the phrase, and I hope I have succeeded. It has struck me that it may be a translation of Propertius’ licens Fabius, referring to the Fabii or Fabiani Superei. Naturally the phrase might be caught up as a mere alliterative jingle like Lazy Lawrence…’; Bradley’s suggestion as to the phrase’s Propertian origin also found its way into the OED’s entry.