‘A Word Factory was wanted’: Organizational objectivity in the making of the Oxford English Dictionary

Emergence of the New English Dictionary

What would become the Oxford English Dictionary was for over half a century known by its creators as the New English Dictionary (NED). It began in 1857 when three members of the London-based Philological Society – the Rev. Richard Chenevix Trench (then the Dean of Westminster), Fredrick Furnivall, and Herbert Coleridge – formed an ‘Unregistered Words Committee’ to inventory all English words not previously recorded by the dictionaries of Samuel Johnson and Charles Richardson. Members of the Society and the general public were called upon to voluntarily scour printed material for such words and extract quotations containing them. The initial response proved so fruitful that the compilation of an entirely new kind of English dictionary was suggested in two lectures that Trench (1860 [1857]) presented to the Society and later published, entitled On Some Deficiencies in Our English Dictionaries. The Society voted in favor of the project in early 1858.

In addition to its intended practical use, the dictionary’s creation was motivated by the lofty goals of comparative philologists to scientifically illuminate general processes of social change, variation, and cultural unity. ³ As Trench (1927 [1851]) states in an earlier work: ‘[t]o study a people’s language will be to study them, and to study them at a best advantage; there where they present themselves to us under fewest disguises, most nearly as they are’ (p. 40, italics in original).)’ The Society’s plan for an English dictionary was also advanced as a means to rectify the poor reputation and amateur state of philological scholarship in Britain. Historical dictionaries in Germany and France were already well underway and were cited as sources of inspiration and guidance. The recent expansion of train networks and postal service allowed the project’s facilitators to cast a wide net in garnering voluntary assistance from all over Britain and abroad. Citing the collaborative precedent set by the Grimm brothers in constructing Germany’s first historical dictionary, Trench (1860 [1857]) expressed the view that it was

only by such combined action, by such joining of hand in hand on the part of as many as are willing to take their share in this toil, that we can hope the innumerable words which have escaped us hitherto, … will ever be brought within our net. (pp. 69–70)

With this, the Philological Society held firm to the notion that everyone was equally qualified to contribute to their project, while seeing their role as assemblers rather than arbiters.

From the outset, however, there were doubts about this scheme. For instance, a reviewer of the project proposal said that

Their design is a magnificent one, but they will have difficulties of corresponding magnitude to contend with in carrying it out. … [B]eyond the collecting and arrangement of materials their prospect is at present, rather hazy. … – where is the wise master-builder? … There must somewhere lie a power of arbitration. From the moment that the building begins, the republic must give place to a dictator. (Marsden, 1859: 386)

Likewise, even one of their own Society members voiced his concern that a leader with ‘cultivated instinct’ would be needed to adjudicate various decisions ‘under the guidance of his own observation, or more commonly of that life-long, unconscious induction, which amounts in a highly-cultivated native speaker’ (Coleridge, 1860: 155).

Nevertheless, the dictionary plans pressed on, even as disagreement mounted over the project’s details. There was the question of whether authors or texts would first need to be, as the project’s proposal suggested, ‘admitted to the rank of a Dictionary authority’ before excerpting (Philological Society, Initial Proposal, 1857, OED/B/1/1/1). Debate also ensued around the inclusion of provincial words, the time span encompassing the English language and the distinct eras of its development, as well as the competence of volunteer contributors and the extent to which they should be involved in the project. Although the original leaders vowed that their dictionary would not instate ‘a self-made dictator, or 40’ at the helm of the English language (‘40’ being a reference to the for-life appointees overseeing the content of France’s Dictionnaire de l’Académie française), their task was not a self-evident one (Trench, 1860 [1857]: 5). Even plainly stating the aim, as the final proposal would, to ‘contain every word occurring in the literature of the language it professes to illustrate’, still required settling sticky points common to every lexicographical endeavor: What is the language? What is literature? What is a word? (Philological Society, 1859: 3).

Over the course of its first three years, the project proceeded to attract considerable interest and numerous willing volunteers. Coleridge became the de facto editor and administrator of the reading program, but in 1860 he died of tuberculosis. Soon afterward, Trench left London for his new appointment as the Archbishop of Dublin, and sole oversight for the project then fell into the hands of Frederick Furnivall. Furnivall implemented a new tier in the project’s division of labor, welcoming volunteers not only to the reading and extraction program, but also to sub-editor roles. The sub-editors were requested, to the extent of their ability, to demarcate the materials by parts of speech, sense of meaning, and historical development, and to provide information on etymology, spelling, and pronunciation among other details and denotations.

But rather than furthering the project, from this point the project would dwindle to a standstill. Faced with the real complexities of their task, many sub-editors – who typically possessed only a hobbyist’s skill, interest, and free-time – admitted feeling overwhelmed or under-qualified. Some simply dropped out or quietly abandoned their work (see examples in Munro, 1911). Some also feared the variability of treatment across the sub-editors and demanded more oversight for their work. For instance, one of the few eminent members of the Society to contribute to the dictionary, Walter Skeat, pleaded with Furnivall to

take exact stock of the amount actually done, and examine if the sub-editors have up to this point done their work in a way which is worth their while to pursue … The result can be nothing but unmixed food. (Skeat to Furnivall, 17 November 1865, MP1)

As Furnivall increasingly issued calls for replacement sub-editors, others ceased working as they saw the prospects of a completed work fading despite their best efforts. Furnivall himself seemed to lose interest in the project, and his time became occupied by other interests. Some seventeen years after the project began, the then-President of the Society, Alexander Ellis (1874), stated that he was ‘inclined to think that a Society is less fitted to compile a dictionary than to get the materials collected’. Nonetheless, it was only five years later that the dictionary would be resurrected by a Scottish schoolmaster, James Murray, along with a publishing contract with Oxford University Press.

Resurrection and relations with the press

Murray had been a member of the Philological Society since 1868, and with his 1873 text, The Dialect of the Southern Counties of Scotland, had only recently solidified his position as a prominent scholar of the English language. The son of a tailor, his formal schooling ended at the age of 14, and he never attained the appropriate credentials for an academic career, but this text helped him to obtain an honorary doctorate in English from the University of Edinburgh in 1874. With his newfound status, he was contacted by the Macmillan publishing house in London, who, in partnership with Harper publishers in America, were seeking a suitable editor for an English dictionary to take on the market dominated by Webster. Murray was intrigued by the proposal, but only if it would allow him to make a historical dictionary using the materials already collected by the Philological Society. The Harper Press was ultimately unwilling to meet Murray’s requests regarding the estimated size of the dictionary and the costs it would incur. All the while, Furnivall and the phonetist Henry Sweet had begun seeking offers from other major publishers.

For a dictionary that now carries the Oxford name in its title, the contract the Press made with Murray and the Philological Society resulted initially in an unlikely pairing. In the 1830s, Oxford scholars sparred with John Mitchell Kemble, an original member of the Philological Society, across the editorial pages of popular magazines, in what would be called the ‘Anglo-Saxon Controversy’ (see Wiley, 1990). The same period saw the beginning of the British Association for the Advancement of Science (BAAS) and a concatenate rise in the founding of new universities, styled in part after the new German research universities and oriented to modern languages and the interests of the middle class and industry (Ben-David and Zloczower, 1965; Morrell and Thackray, 1981). These trends were facilitated by and facilitated the growth of a popular scientific movement which represented a threat to the basis for knowledge authority and status of the Oxbridge elite. Cannon (1978) has characterized this period by its ‘disintegration of the truth-complex’, in which the image of the scholar as a possessor and professor of absolute truths was challenged by the image of the scientist as an explorer or pioneer (p. 3). In fact, she claims that the real enemy of science in Britain was not the religious structure of the old universities; rather, the enemy of both science and religion was the Greek and Roman classics. Henry Sweet, who was a reader at Oxford and highly critical of its culture, ⁴ quipped that ‘[t]hey are hardly aware of the existence of Shakespeare & Chaucer yet’ (Sweet to Murray, 4 April 1970, as cited in Murray, 1977: 247).

Nonetheless, the BAAS proved to be a particularly potent force for solidifying public, and then government, support for the sciences in Britain. With this, Oxford would be forced to make considerable reforms in line with the Royal Commission of 1850 (Ben-David and Zloczower, 1965; Soares, 1999). These changes made conditions far more favorable for a project like the dictionary, but amid lingering tensions, animosity, and suspicion, Murray, Furnivall, and Sweet were cautious and disingenuously deferential in their dealings with the Press. For instance, following the initial proposal to the Delegates of Oxford University Press, Sweet wrote to Murray that

They made a variety of other criticisms, some of which it would of course be impracticable to carry out. It is, however, very important that we should show every deference to their suggestions and consider them fully. When we have once got over their chief objections they will leave the Editor entirely to himself. They naturally like to make as many criticisms as possible in order to show their sharpness and wideawakeness [sic]. (Sweet to Murray, 29 June 1877, MP3)

The three also stressed that the project was a good business decision, promising to be exceptionally and easily lucrative. But the rejection of the project by Harper/Macmillan should have provided doubt of the work’s profit-potential, as Alexander Macmillan himself would later advise the Press: ‘[t]he terms may be such as the delegates in the interests of scientific philology might accept but they certainly are not business terms’ (Macmillan to Price, 1 May 1877, OED/B/3/1/1). The completeness of the materials under the previous sub-editors was also overstated, as was the ease of assembling them and the quality and consistency of the work. Sweet showed no qualms about misleading the Delegates. Believing their main objective was that ‘of depriving the Society of all control of the material and the way in which it is worked up,’ he warned Murray:

[y]ou must be prepared for a good deal of vexations, interference and dictation hereafter, liable to be enforced at any moment by summary dismissal. … I know something of Oxford, and of its low state of morality as regards jobbery and personal interest. (Sweet to Murray, 7 July 1878, MP3)

This vilification and distrust of the Oxford Press would even be extended to Murray. A small number of Society members (including Furnivall and Sweet’s uncle) feared that Murray was in cahoots with the Delegates and attempted to block the Society from contracting with the Press. When privately explaining the incident to the Secretary of the Press, a vexed Murray would exclaim that ‘hostility to Oxford has become a species of monomania’ within the Society (Murray to Price, 25 January 1879, OED/B/3/1/3). But Murray, too, showed some leeriness about working with the Press; he would later reflect on the negotiation discussions that ‘many of the Delegates cared nothing about it, and rather grumbled at the funds of the press being diverted, as they think, from their legitimate purposes, to so large an “outside object”. I am also an “outsider” of them’ (Murray to Gibbs, 25 February 1882, MP5).

Relaunching production

A contract was finalized in March 1879, and Murray assumed he would simply build on the previous decades’ materials. But, as he began calling the materials in from various parts of the world, earlier fears of variation were validated. He would call the materials ‘an incubus of rubbish and error’ and in April of 1879 a new appeal for volunteer readers was circulated (Murray to Hall, 11 April 1899, MP14). With the appeal, readers were offered one page with 12 simply stated rules, followed by three specimen quotation slips. Where the previous directions required readers to consult ‘Basis of Comparison’ lists and offered paragraphs of fairly abstract and technical guidance, Murray simply instructed readers to ‘[m]ake a quotation for every word that strikes you as rare, obsolete, old-fashioned, new, peculiar, or used in a peculiar way’, and also to ‘[m]ake 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’ (Murray ‘Directions to Readers’ c.1879, OED/B/4/1/1). Murray maintained an earlier directive that readers submit individual quotes on a 6 ⅝ × 4 ¼ slip (a half-sheet of the standard note paper).

A total of 2000 of the new appeals made their way around universities, libraries, bookshops, and scholarly societies; news of it was further circulated by the British and foreign press. By the middle of May 1879, Murray already had a total of 165 readers, then 400 by September, and a year later there were 754 readers, 1568 books read or being read, and already 361,670 slips; by 1882, a total of 1 million additional slips had been produced (Mugglestone, 2005). Prospective volunteers would write in and offer to read one or more texts, which in some cases could be loaned out to them. ⁵ Other readers were more irregularly involved, such as those responding to Murray’s desiderata word lists that he annually published in popular journals such as Notes & Queries.

Regardless of their form of recruitment into the project, each volunteer’s slips would first come into the Scriptorium where Murray based his operation (first in Mill Hill near London, later in Oxford). The Scriptorium was the physical centerpiece of operations, where materials were passed in ever-cumulative degrees of synthesis. It was a 30′ × 15′ poorly ventilated structure made of corrugated iron, which sat in front of the Murray family home – being reconstructed after their move to Oxford in 1895. Murray referred to it as his ‘lexicographical laboratory’, where the slips would first be alphabetized, cross-checked against their sources, and assembled with existing slips for each word (Anonymous, 1882: 245). These tasks were undertaken by low-level assistants, female clerical staff, and sometimes one of Murray’s 11 children. A mid-level assistant would then order the slips for each word into different parts of speech and senses and arrange the quotations chronologically. A packet of these ordered slips for a particular section of the alphabet would then go out to a sub-editor or more skilled assistant, who would divide each sense of a word into the different shades of meaning, attempt preliminary definitions, highlight the quotations which seemed to best exemplify each meaning, and arrange the senses in the perceived order of their historical development. This work might then be sent to a re-sub-editor who would comment on the previous work and incorporate newly acquired quotations before passing the bundled materials back to the Scriptorium for further processing.

The Scriptorium was a designed as a unique structure to direct this flow of production. Upon visiting the site, one journalist described it as

A Word Factory … where the philological raw material, spread abundantly and heterogeneously, could be ground out by division and sub-division, mechanically, and afterwards with reason; where there could be operatives and overseers parceled, appropriated … Entering the slight door of it, there is entrance upon plain deal. There is an aspect of unvarying, unincident-giving, sawn and planed white-wood. (Humphreys, 1882: 445–447)

Below its skylights, Murray and his top assistants were often positioned on a daïs raised about a foot higher than the rest of the production. To sort out and store what would be millions of word slips, much of the wall-space was lined with pigeon holes sized appropriately to the 6 ⅝ × 4 ¼ dimensions of the slips. Humphreys (1882) claimed that there were over 1100 pigeon holes ‘higher than the arm can reach; going down so low there is need to stoop’ (p. 447). Other wall-space was filled with shelving for reference books, and also for filing correspondence. The latter was not a marginal affair in the dictionary’s production; Murray himself calculated that at the height of the reading program ‘I had nearly 2000 Readers actually at work, writing letters to me at the rate of 40 or 50 in a day’ (Murray ‘Dictionary Evening’, 1910, MP33). The Scriptorium as a tangible place of research was still only one node in the full operation – a kind of ‘counting-house’, as Humphreys (1882: 445) described it. The red pillar post box that was specially installed in front of Murray’s Oxford property was one indication of the loose network structuring the dictionary’s global operations. Even Murray’s later co-editors would work remotely. ⁶

Once the sub-editing process was complete, the bundles of ordered slips would then be available for Murray or another editor, who would check the work, further sub-divide or condense divisions of meaning, and proceed to draft definitions, etymological notes, and pronunciation. Each bundle of slips was then numbered and sent to the press. The press would prepare draft proof sheets of the material, which would incur intense scrutiny and modification at the hands of the editors, high-ranking assistants, and a small contingent of reputable scholars and invited volunteers who had shown special proficiency as readers or sub-editors.

Thus, iterative corroborative processes were pervasive at all levels of the project’s production. Mugglestone (2005) has called the making of the dictionary a ‘densely accretive process’ (p. 38) and former editor William Craigie (1928 as cited in Gilliver, 2004) speaks of ‘how the long stretches of raw material on the shelves have gradually passed through the various hands, until they reached the stage of printer’s copy, proofs, revises, and finished sheets’ creating altogether a ‘mass of patient, honest, careful work’ (p. 63). Once in circulation, it could still be a considerable amount of time, often months, before a page reached its final publishable state. The extent of this deliberative process produced what a later editor at the Press conjectured were ‘possibly the most heavily corrected proofs ever known’ (Sutcliff, 1978: 61). In various works, Mugglestone (2000b, 2005) has examined the proofs for what they reveal about ‘the controversies and conflicts, the quibbles over labels, citations, canonicity, and legitimacy as they took place’ (Mugglestone, 2005: xix).

Whereas Furnivall’s failed model of production delegated multiple facets of the work to sub-editors, Murray more minutely divided the tasks to be undertaken along a network characterized by high degrees of standardization, an ad hoc division of labor, iterated corroboration and deliberation, and a centralized checkpoint for operations. Under these circuitous processes, no individual was to have full discretion over any part of the final product.

Historical significance of the NED

Covering A-Ant, the dictionary’s first fascicle was published in January of 1884. At the time, it was already 5 years after the commencement of the contract – which initially supposed the entire dictionary would be completed by 1892. Murray hoped that upon this first published offering, the Press might finally realize the grandeur of the work, thus impelling them to loosen the purse strings in order to ensure its completion. However, the Delegates, like many reviewers of the first fascicle, were by no means overjoyed by the content of the work. It was the subject of mixed reviews, or, more specifically, ‘many philologists on the Continent and in the United States gave it enthusiastic reviews. In England there were few scholars qualified to appreciate the work’ (Murray, 1977: 234). The British media outright rejected the descriptive approach and lambasted its use of ‘non-authoritative’ sources or seemed preoccupied with incidental details (see Bivens, 1980–1981 for an annotated bibliography covering 19th-century reactions to the dictionary). For instance, one reviewer wrote,

The truth is that the editors have been extremely lax in their choice of authorities. Some of the best are omitted, some of the worst are adopted. … the most recent examples are taken from ephemeral publications, newspapers, magazines, and trashy novels. … quotations are made from the most worthless, careless, and ignorant publications of the day. (Reeve, 1889: 346–347)

Echoing earlier doubts about the production scheme (see above quotes from Marsden and Coleridge), another complained,

If the Editor goes on upon this plan of tabulating every case, giving all the dates and all the spellings and every minute fact that can be rummaged up, whether it is worth anything or nothing … I (who am an old man) shall certainly never live to see it. … The Editors of the Philological Society’s Dictionary seem to be Legion, and no central authority to control them. (Henry Norton to [unclear, likely Henry Frowde of the Philological Society], 30 January 1884, MP6)

To better appreciate the historical significance of this dictionary style and why it provoked such dissatisfaction and hostility, consider the prior standard-bearer in British lexicography, Samuel Johnson. Initially published in 1755, Johnson’s dictionary was a work that was progressive in its meticulousness, comprehensiveness, and the inclusion of word evidences from print. But, as one 20th-century critic put it, Johnson

made many mistakes; he put into his work expressions of his own likes and hates, epigrammatic and mordant flashes from his own mind, such as – ‘Patriotism. – Love of country; the last refuge of a scoundrel. Oats. – A grain which in England is generally given to horses, but in Scotland supports the people’. (Renwick, 1913)

Murray finds Johnson’s work to be that of an author rather than an editor, showcasing ‘the prescientific age’ of philology, when ‘a priori reasoning was applied to problems which could only be solved by patient induction’ (Murray, n.d. [c.1883], Notice of Publication of a New English Dictionary on a Historical Basis [early draft], MP29; see also Murray n.d. [c.1900], ‘Editor or Author’, MP32).

From the inception of the project in the 1850s, it was clearly never the intention of the new dictionary’s founders to follow in the tradition of Johnson. The Philological Society’s approach sought to work from the ground up, with the raw information of word-usages being the foundations on which a holistic knowledge of the English language could be built. As one reviewer put it,

Earlier dictionary makers seem either to have despised illustration altogether or to have simply angled in the waters of literature for illustrations suitable to a preconceived set of meanings; but the modern dictionary-maker sweeps the whole of those waters with a drag-net, and then proceeds to classify and catalogue his spoil with the minute completeness of a naturalist. (R. McLinktock, 29 April 1889, MP34)

The work produced throughout the first edition would also be considerably distinguished from the German and French historical works that originally inspired the dictionary’s creation. Despite the involvement of thousands of contributors and stakeholders, and notwithstanding Murray’s death thirteen years before completion of the first volume, one would find ‘wasp and bee, yellow and brown, brother and sister’, each treated in the same manner (Osselton, 2000: 73). Not even the French dictionary produced exclusively by Émile Littré sustained a similar consistency. In his entries for the Deutsche Wörterbuch, Jacob Grimm ‘battens on that single point in historical development which he judges to be the most interesting – in one case, morphology, in the other, syllable stress’ (Osselton, 2000: 63). The very skeletal, succinct presentation of entries in the English dictionary was in considerable contrast to Grimm’s tendency to indulge in readable prose, as well as Littré’s tendency to offer up lengthy quotations of word usage. The English dictionary also refrained from the more-or-less subtle tendencies of these continental lexicographers to steer their countrymen’s vocabulary, spelling, and pronunciation (see Osselton, 2000; Zgusta, 1986, 1989). Given these distinctions, as well as internal disagreement over vision and method across even the originators and leaders of NED, it is clear that the shared philological and scientific intentions across each of these innovative projects were open to multiple interpretations. The makers of the NED were not simply better able to realize their vision; they faced far more obstacles than the Grimms or Littré, who had more resources at their disposal and more direct control over their work.

The highly descriptive and unassuming nature of the English dictionary extends from the highly impersonal nature of its production processes. While the results of this method were clearly appalling to some, the production of the dictionary was (and increasingly came to be seen as) an intriguing, if not exciting, development in knowledge production. The operation seemed to exploit all the recent advancements in travel, communication, and printing, thus showcasing the potential of the new global information age. Even other collaborative knowledge projects to that point, for instance, the King James Bible, Diderot’s Encyclopédie, and the Encyclopedia Britannica, were developed by closed groups or invited contributors, or made little mention of lower status assistants (see, for example, Blair, 2010; Kafker, 1996; Kafker and Loveland, 2009; Nicolson, 2003). Extending an open invitation to strangers was a fairly remarkable approach. ⁷ With this, the makers of the dictionary offered a model for a more inclusive form of public knowledge. Not only was this distinguished from the exclusivity upheld by Oxford and Cambridge but it was also untethered from the gentlemanly norms that operated as an implicit basis for membership and credibility within Britain’s informal networks of scientific practice, including the BAAS (Morrell and Thackray, 1981).

Following from this intrigue, the Scriptorium became a novelty attracting considerable attention from scholars, journalists, and other visitors from around the world who wished to gawk at the machinery of modern knowledge manufacture in action. It is said that

[a] German professor visiting the Scriptorium proclaimed: ‘It is so English; when a German is going to write a big book he sits down in his garret, with his wife and his tobacco, but the first thing an Englishman does is to build a house to write it in!’ (Hawick Advertiser, 25 December 1903, MP32)

One journalist colorfully proclaimed, ‘Sir James Murray has, indeed, almost become one of the “sights” of the country. American tourists come by the score to shake him by the hand. From Germany squadrons of professors appear in single file’ (Renwick, 1913). Even those who were critical of the outcome showed some fascination with the method. For instance, despite Reeve’s (1889) negative appraisal of the dictionary’s sources (excerpted above), he shows some approval of the new era of encyclopedia and dictionary-making, in which ‘[t]here may be no genius, there may be no originality, but there is an amount of industry and scholarship employed in storing and reproducing the knowledge of the world which has never been surpassed’ (p. 330). Echoing this industrial sentiment, in 1906, another proclaimed that ‘[t]he dictionary … must now be the largest single engine of research working anywhere in the world’ (Cannan, 1906, as cited in Mugglestone, 2005: 188).

The necessity and strategy of impersonal procedure

One can reasonably question the production and presentation of the NED: These processes were not necessarily self-evident, were not guided by market considerations, and involved extensive networks across collaborators; the networks were not one-way channels in a hierarchical structure but instead included circuits of corroboration and deliberation that effectively doubled work and impeded resolve.

It must be recalled that the British comparative philologists, like most British scientists at the time, were not as professionalized as their forebears on the European continent, particularly in Germany. They shared no common education and standards, and no generally accepted measures for assessment of their skills. As Sweet articulated,

[m]ost of us – indeed, nearly all of us – are by force of circumstance compelled to work in a dilettante style; we cannot expect much from a philologist whose whole working day consists, perhaps, of an hour snatched from other labors. Where, again, are we to get our training? We are left to pick it up at random. (Sweet, 1877, as cited in Murray, 1977: 72)

But, in light of this, in contrast to the Grimms’ initial contingent of 83 dictionary collaborators – ‘a dozen professors, a couple preachers, all the rest are philologists’ – the NED contributors came from all walks of life and from all over the world. Perusing Murray’s correspondence records, Humphreys (1882) reports,

here are correspondents brought together in it from such distant spots as Florida, Llandudno, Copenhagen, Illinois, Wrexham, Dublin, Biarritz, Halle, Fife, Japan; from Tours, Iowa, Ceylon, Machynlleth, Taunton, Birmingham, Llandaff, Mauritius, Indianapolis, the Temple, the Universities, Lincoln’s Inn. Here are correspondents who are professors, filling various chairs; who are poets, historians, critics, musicians, inventors; who are canons, archdeacons, army-men, navy-men, ladies, peers; who have attained the distinctions of D.D., M.A., F.R.S, C.E., F.R.I.B.A, F.S.A, M.D. (p. 452)

The diverse composition of this pool of contributors was likewise highly heterogeneous with regard to their motivations for contribution, understandings of the project, and levels of skill. Misunderstandings of the project abounded, with Murray (1977) being frequently commended for finally ‘fixing’ the embarrassingly unwieldy English language. He also courted, as his biographer puts it, ‘irritatingly stupid questions from volunteers upon whose goodwill he depended’ (p. 181). Many made inquiries regarding ‘correctness’ – to which Murray wished to ‘point out the impropriety of applying right and wrong to mere matters of current usage’ (Murray to unnamed, 28 March 1884, MP6).

Variation of vision was even evident at the final stages of production and among the uppermost circle of trusted editors and experts. Mugglestone (2000b) portrays the inner circle of proofreaders as divided between those favoring ‘quantitative’ or ‘qualitative’ norms of usage. The former accepted as valid any linguistic development with a certain number of users, while the latter found more significance in who was doing the using. Walter Skeat expressed a preference for the qualitative norm on the basis of marketability, maintaining, ‘we need not record the dream of every driveller. … If we impose on purchasers our theories and crotchets, they can effectively retort by buying some other Dictionary’ (Skeat to Murray, 30 November 1878, MP24). Henry Hucks Gibbs (later to be Baron Aldenham) was also an outspoken proponent of qualitative norms, particularly when it came to quotations of modern usage, often taken from newspapers. He argued, ‘[t]hat there are many indefensible words used in all periods, but more in this because of the rapidity with which men can now write and print and of the omnivorous greed which makes readers swallow everything’ (Gibbs to Murray, 30 July 1882, MP5). Furnivall, who clipped word-usages from his daily newspapers, obviously took quite another position, and even advocated giving special preference to quotes from female writers (see Gibbs to Murray, 3 May 1883, MP5).

With difficulties compounded by the inability to identify good volunteers from the outset, it was of no help that members of the Philological Society themselves, as well as scholars engaged in relevant subject matter at the British universities, while warmly recruited, were not forthcoming in their contributions to the project. Given the wide-ranging heterogeneity that ensued, there was a two-way channel of doubt between the leading editors and the volunteer collaborators, in which expertise and objectives were uncertain from the outset.

Under these conditions, it would likewise be difficult to self-assess or trust one’s own judgment – or at least, to make authoritative claims regarding one’s knowledge and judgment. Although Murray was the salaried editorial leader of the project, there are several instances where he had difficulty arbitrating between competing views. He would repeatedly seek consolation and support from his friends and colleagues for his decisions. In one instance, Gibbs would feel compelled to remind Murray of his leadership position, advising,

[t]here is a great deal in what you say about it not being the business of the Editor to sit in judgment on words. … But you bear two things in mind. 1. you have appointed a committee [figuratively speaking], of whom I am one, and who happily have no authority, but who I daresay speak their minds as plainly as I do, for you to take or reject what we say. … 2. You do necessarily sit in judgment to some extent, and do and must reject some rubbish which passes for words. (Gibbs to Murray, 30 July 1882, MP5)

On another matter Henry Bradley would reassure Murray that

[i]t is unlikely that there is anyone else, … who has worked out the whole question as you have, and it seems to me that this is one of the few matters in which a dictionary-maker, if properly qualified, may claim to make law instead of merely recording usage. (Bradley to Murray, 12 December 1890, MP9)

Besides his own self-doubt, Murray also faced several challenges to his authority. Strangers wrote to him trying to press their influence or trump his expertise on dictionary matters. Furnivall made roundabout complaints about him to the Delegates and the Society. The Oxford Vice-Chancellor Benjamin Jowett expressed several reservations about Murray’s decisions and, in incidents discussed below, even took it upon himself to alter Murray’s materials. Another challenger to contend with was Max Müller, a Delegate and Oxford’s first chair of comparative philology, who in all other respects was a fortunate ally for the project. Müller would peer over Murray’s shoulder from time to time and offer firmly stated suggestions, which Gibbs chalked up to jealousy (‘I have curiously fancied for some time a little jealousy on his [Müller’s] part’) (Murray, 15 November 1884, MP6). But nonetheless, as Sutcliff (1978) notes,

the idea of Max Müller advising the dictionary editors on etymologies, as he proceeded to do, was hair-raising even then: his own etymologies tended to justify his mythological theories and, despite his reputation, did not conform to the scientific principles of the new philology. (p. 55)

Although both Müller and Murray were representative of the new philology, these incidents depict a noteworthy confrontation between professional authority and the highly skilled amateur.

Heterogeneity and impersonality

Star and Griesemer (1989) argue that knowledge work among heterogeneous actors can proceed without consensus or personal trust. Instead, it can be facilitated by means of standardized methods and boundary objects, both of which are clearly evident within the NED case. For instance, consider the uniform rules Murray issued to participant readers and sub-editors. These were brief, largely sequential, and were written in nontechnical everyday language. Their focus on ‘how, and not what or why’ is consistent with the character of standardized methods emphasized by Star and Griesemer (p. 407). Boundary objects also serve to restrict the extent of participation while still allowing for a diverse body of contributors. These are defined as objects which ‘are both plastic enough to adapt to local needs and the constraints of the several parties employing them, yet robust enough to maintain a common identity across sites’ (Star and Griesemer, 1989: 393). As a type of standardized form, the 6 ⅝ × 4 ¼ slips played this role in the production of the dictionary. Their significance was variously defined throughout the workflow of operations and depending on one’s location in the production process.

The significance of the slips was particularly acute given the diverse material and stylistic means of communication across individuals. As Humphreys (1882) observes of Murray’s correspondences with contributors:

Here are letters in a clear smooth hand; in a weak scrawl; in a double-up thick knottiness, like an attacking fist; … on one sheet, and on three or four, and brimming over on to every margin; and written sideways, and written longways, and written on the inner leaves, and on the outer leaves; and in lilac ink and black ink and blue ink; and on every-conceivable kind of stationery. (p. 452)

This posed particular problems, as Murray explains,

[t]he original method differed a little from mine, in the position of the catch-words, book-titles, and other details; and now the time has come when differences must no longer be. For if we do not know where to find quotations, is there any use in the quotations being here?. (Humphreys (interviewing Murray) 1882: 447–448, MP34)

These expectations not only aided the transfer and synthesis of information contained on the slips but also communicated important information about the contributors themselves. Murray explains that, despite the codified instructions,

[s]lips come, proving to be not English at all, but Arabic (‘aya-el-khan’, and so forth); stray examples come, with catchwords only, and all else omitted; other stray examples show quotations right enough, but book, chapter, and verse forgotten … Nothing shows us better how character will make its mark … And character has broad space for the marking among our good ranks of readers. (Humphreys (interviewing Murray) 1882: 451, MP34)

And marked they were, as Humphreys (1882) revealed. Murray maintained files on each contributor annotated with updates on their involvement and commentary on their skills and habits, such as ‘“[t]hrown up; slips lost;” and “Gone; no address left;” and “Promised by end of year;” and “Will send miscellaneous jottings;” … and, more often than it should be, “No good,” “No good,” “No good,”’ (Humphreys, 1882: 452–453, MP34). Schaffer (1988) has observed a similar association between standardization and evaluation among the growing network of astronomers in the 19th century, in which ‘[t]he observatory became a factory if not a “panopticon”’ (p. 119).

Standardized procedures and their related infrastructural objects thus facilitated communication between project contributors, as well as the commensurability of the knowledge and information they each supplied. These methods allow for some autonomy, which makes it possible to recruit and retain a heterogeneous contingent of contributors who can attach their own interests, values, and meaning to the work and uphold their own motivations for participation. But they also strip away or correct individual idiosyncrasy and restrict the range of action. Even those who were more known to the editors, personally trusted, and able to offer richer information and insights at the final stages of the dictionary’s production were still limited in how they could make their case. As Gibbs remarks, ‘when I speak my mind on the margin of a proof, space demands that my remarks be short and sharp, and not as gentle as they would be if I had more room on which to smooth them out’ (Gibbs to Murray, 30 July 1882, MP5).

The originators and leaders of the NED had few other options but to make use of a large and diverse array of language enthusiasts; Star and Griesemer help explain how this setup could work despite inherent challenges. The relationship between impersonal procedure and heterogeneity, particularly in combination with uncertainty (including weak or contested authority), has been noted across Star’s wealth of research. Her original study with Griesemer focused on the 1907–1939 creation of Berkeley’s Museum of Vertebrate Zoology and its cooperative processes between a professional biologist, amateur specimen collectors, bureaucrats, and patrons. But heterogeneity in knowledge production, along with similar tendencies toward impersonal procedure (and the limitations that come along with it), has been explored in other forms as well. For instance, Vaughan’s (1999a, 1999b) studies of the Challenger space shuttle disaster investigated the nature of contemporary scientific work and the presentation of technical knowledge within formal organizations with specialized modular units. Additionally, Galison (1997) has examined the ‘trading zones’ in which scientists working within different paradigms collaborate and exchange information.

Heterogeneity can also be seen as historically variable. It is implicated in the work of Porter (1994, 1995) and Daston (1992) as underlying significant shifts in scientific thought and practice during the 19th century. This included the very commencement of ‘objectivity’ as a scientific ideal – particularly a sense of objectivity emphasizing impersonal processing and presentation of information. Before the 19th century ‘[f]ar from embracing the ideal of the interchangeable observer, seventeenth- and eighteenth-century scientists carefully weighted observation reports by the skill and integrity of the observer’(Daston, 1992: 610). And as many studies of early modern science have shown (among them, Dear, 1985; Secord, 2000; Shapin, 1994; Shapin and Schaffer, 1985), these assessments generally followed from established cultural biases and status hierarchies.

Therefore, it is not surprising, as Porter (1995) notes, that overarching social trends toward democratization proceeded in step with new appeals to objectivity. He sees this manifest in the sudden proliferation of quantitative forms of social measurement and statistical analysis during the 19th century. Daston (1991, 1992) instead emphasizes the internationalization of knowledge exchange as a source of scientific expansion leading into the 19th century, but she likewise notes the communication problems this entailed, bringing about an ‘aperspectival’ imperative to standardize observational inputs, while ‘narrowing the range of genuine knowledge to coincide with that of public knowledge’ (Daston, 1992: 600).

The NED case offers an empirical parallel to the statistical tools noted by Porter and the standardized techniques and mechanized apparatuses of 19th-century science studied by Daston – including her work with Galison (1992, 2007). I find it advantageous to use Megill’s (1994) ‘procedural’ sense of objectivity as a categorical concept encompassing these forms. Under this rubric, the NED case reflects an ‘organizational’ model in which collaborators are configured or portrayed as interchangeable cogs in a rationally arranged division of labor. Inputs are standardized, discretion is neutralized, and outputs are assumed to be the product of deliberation and eventual consensus. More contemporary instantiations, including ‘regulatory’ forms of objectivity (see Cambrosio et al., 2009), also fit within the scope of procedural objectivity. These highlight the potential for comparative research considering the instances in which different forms of objectivity are appealed to, and also their different effects on the organization of knowledge production and the knowledge produced.

Rhetorical elevation of impersonality

While the dictionary case proves consistent with existing explanations of the practical basis for impersonal procedure, these explanations – as well as the psychological motivations perpetuating the ethic of self-control manifest in ‘mechanical objectivity’ according to Daston and Galison (1992, 2007) – do not sufficiently capture the whole of the NED narrative. The impersonalized processes of production also appear to be rhetorically leveraged. Slogans of objectivity and impersonality were particularly touted when the dictionary staff was dealing with the Press, the general public, and reviewers. For instance, Murray’s draft of a public announcement for the dictionary stated,

[t]he Editor has sought to give such results only as are beyond dispute, avoiding rash speculation and all dogmatism on doubtful points. His object has been to allow each word, as far as possible, to tell its own story’. (Proof of Notice of Publication … shorter leaflet, MP 29)

And to an American journalist seeking biographical information about him, Murray would state,

[f]or myself I have nothing to reveal, and nothing, (luckily), to conceal; people will do me the greatest favour by forgetting me in my work, and treating my quite impersonally … I have persistently refused to answer the whole buzzing swarm of biographers, saying simply ‘I am nobody – if you have anything to say about the Dictionary, there it is at your will – but treat me as a solar myth, or an echo, or an irrational quantity, or ignore me altogether’. It was unfortunately not practicable to edit the dictionary anonymously, else I should certainly have done so. (Murray to unnamed American journalist, 20 April 1886, MP7)

Emphasis on the impersonal aspects of abstract organizational processes also seemed to be a strategy for neutralizing interference from the Press. From the outset, some Delegates expressed doubts about Murray’s ‘literary taste’, worried about his potential ‘provincial bias’, and in reviewing his specimens, sought some ‘reason why particular authors are regarded as authoritative’ (Liddell to Price, 10 May 1877, OED/B/3/1/1; Shalb to Price 11 June 1877, OED/B/3/1/1). By emphasizing the impersonal and methodological processes in their endeavor, and by deferring to the data on hand, the philologists could evade a direct confrontation on the grounds of ‘taste’ and avoid debates becoming a matter of one’s word against another.

In fact, surprisingly little effort was made to have the Delegates understand the underlying philology and scientific method guiding the work. Instead, it was common for Murray’s counter-arguments to humor the Press’ concerns and reasoning, while inverting their logic or arguing that honoring their preferences would undermine the speed or sale of the project. In one instance, among the many where Murray was instructed to avoid the use of newspapers for quotation evidence, Furnivall wrote directly to the Secretary of the Delegates that it had always been the intention to include newspapers as a source. While this was true, and there was a valid philological reason for using newspapers as sources, Murray was nonetheless compelled to apologize to the Delegates for ‘the imprudence and meddlesomeness of Mr. Furnivall’ while claiming that the inclusion of newspaper citations ‘was to me really a most important practical one – for if I was to leave out newspaper quotations as your letter directed, when I had nothing to substitute for them, what was I to do?’(Murray to Price, 15 June 1882, OED/B/3/1/5). Efficiency and the avoidance of subjective judgment are also put forward when Murray explained why he could not simply, as the Press preferred, omit scientific and technical words unless they had attained a literary sense. Murray explained, ‘omission is not always too often a shorter process than admission … I confess inability to say what is literature, and utter inability to say whether a given “scientific” word has or has not been used in literature’ (Murray – apparent draft of a response to the Delegates’ Suggestions, August, 1883, MP5).

As the project went well over its expected time-frame and budget, and the Delegates became firmer in their demands, again and again Murray would reiterate the claim that their attempts to constrain the work were only holding up its production. In October of 1883, he wrote to the Secretary of the Delegates, ‘if I must discard these and search for others from better writers, the building will stand still while the builder go searching for stones’ (Murray to Price, 18 October 1883, OED/B/2/2/1). Similar reasoning was employed nearly 13 years later, when Murray wrote to the Delegates that ‘[t]he experience of the last two months in trying to apply the Delegates’ instructions has shown that they increase my labours and retard my progress very notably’ (Murray to Gell, 23 July 1896, OED/B/2/2/1).

I claim that these appeals to impersonal processes and organizational efficiency were rhetorically leveraged because the dictionary was not, and couldn’t realistically be, a thoroughly impersonal and data-driven word factory. And yet, this image was played up even though it was at times problematic for Murray and his collaborators. For instance, it is somewhat ironic that Murray said he wished he could have edited the dictionary anonymously (quoted above), for he would at a future date threaten to quit the project when his name was neglected in an advertisement for the work (Murray to Gell, 29 October 1897, OED/B/3/1/11). Despite early inhibitions, he would come to recognize the role he played and the necessity for some degree of adjudication, and this became a point of tension in his relations with his collaborators, the Press, and the general public.

In Murray’s preface, he articulated the kind of personal discretion he had to employ, comparing himself to the natural scientist who must of necessity define forms and kinds where no such natural boundary exists. Murray’s explained,

[i]n its constitution it [the English language] may be compared to one of those natural groups of the zoologist or the botanist, wherein typical species, forming the characteristic nucleus of the orders, are linked on every side to other species, in which the typical character is less and less distinctly apparent, … The lexicographer, like the naturalist, must ‘draw the line somewhere’. (Murray, 1888, as cited in Reeve, 1889)

Murray and his collaborators were also involved in the delicate intellectual matter of ordering the senses of words by their perceived logical–chronological development – regardless of whether those chronologies synced up with the dates of their respective evidence in an entry. Murray, being re-quoted by an assistant, claimed that ‘working out into the chain of ideas in the senses, with the aid of a by-no-means-complete chronological series of examples’, was ‘the greatest and most difficult of all the processes’ (Hallam to Murray, 17 November 1882, MP5). The task required sharply attuned skill and imagination in conceiving how and why one sense of meaning might have given way to another. Related to this process, highly experienced judgment needed to be levied in determining how many unique senses of meaning existed for a word, and in deciding whether additional research was needed to capture known usages absent from the existing materials.

Murray’s character and reputation were also important in building relationships and strengthening ties with the many unseen volunteers. His co-editor, Henry Bradley, recognized that ‘the large amount of outside help which adds so much to the value of the dictionary would never have been forthcoming but for the editor’s energy and personal influence, and for the confidence inspired by his ability’ (Bradley to Gell, Nov. 8, 1887, OED/B/3/1/8). Furthermore, Murray’s extensive personal letter writing to volunteers kept them engaged and on track. But although the project uncovered and made use of some otherwise untapped talent and hard work, Murray quickly recognized the limitations of the volunteers and wished to keep most of them limited to the more menial tasks. For instance, he wrote to a correspondent:

I will send you some specimens: but I should be loath to have 100 printed to distribute. My experience is that the suggestion of friends give much trouble and no help, as necessarily no one does or can consider every point with the patience and in the many lights that I must do. (Murray to Pitman, 7 October 1882, MP5)

And in a speech for one of the Society’s intermittent dictionary evenings in 1910, Murray wrote,

On the whole, the volunteer sub-editing, tho’ done with the greatest good-will, and immense diligence, has not been a great help. … I have come to the conclusion that practically the only valuable work that can be done by the average amateur, and out of the Scriptorium, is that of reading books and extracting quotations. (Murray, Dictionary Evening, 1910, MP33)

Because maintaining an aura of impersonalized process was integral to sustaining the trust of collaborators and neutralizing outside interference, the Press lacked an accurate depiction of how skill, experience, discretion, and trial and error actually figured into the dictionary’s making. They thus felt justified in simply adding more assistants and editors in their hope to speed up production, once complaining to Murray: ‘the fact that perplexes the Delegates is that it is taking almost the same time to produce Part III as it took to produce Part II, and that there is no adequate result apparent for the £1250 placed at your disposal for assistants’ (Gell to Murray, 16 November 1886, MP7). And while turnover was high among Murray’s assistants, the Press was not forthcoming with any means to raise salaries, despite Murray’s insistence that ‘[s]o much has to be carried in the head, that some amount of permanency in my staff is a necessity’ (Murray to Price, 18 June 1884, MP6).

Murray thus found himself in the difficult position of needing to outwardly idealize and reinforce the impersonal nature of the processes underlying the dictionary’s production, while at the same time, for practical and, one presumes, psychological reasons, wishing the Delegates of the Press would recognize the extent of cultivated intellectual expertise that he and each individual brought to the work. The Delegates’ failure to appreciate this fact is further evidenced by an incident in which Oxford Vice-Chancellor Benjamin Jowett took it upon himself to significantly alter a preface, advertising prospectus, and title page that Murray had drafted for the first fascicle. In an emotionally charged first draft of his response, Murray exclaimed that he found the changes to be an ‘intentional slap to remind me that I am only a poor casual Editor, whose work is due not to himself, but to his time’ (Murray to Price (likely a draft), 23 October 1883, MP5).

In the final draft of the letter, Murray mentions that Gibbs had wished to see the dictionary titled as being ‘by’ Murray, but that he is content with simply being named as its editor, so long as the Delegates ‘remember that my work is something very different from editing or “compiling” in the ordinary sense of those words, and that so long as I continue as “Editor”, I expect them to recognize what that work is’ ⁸ (Murray to Price, 27 October 1883, MP5). The revisions to the title page even slighted the great amount of volunteer assistance given to the task, to which Murray declared, ‘I have a strong personal desire, that the two lines acknowledging assistance should appear on the Title-Page; it is both due and expedient’ (Murray to Price, 27 October 1883, MP5).

Even as Murray began accumulating numerous personal honors, he would still continually need to fight for the Delegates’ recognition of the knowledge, skill, and tactful management that he brought to project. In 1897, he was informed that the Delegates had, without his knowledge, appointed the little-known etymologist William Craigie to become an editor on the project. Murray was understandably furious not to have been consulted on the matter, exclaiming, ‘[t]his want of confidence, is either an intentional slight put upon me or it is a failure to recognize my original and factual relation to the work, to which I cannot submit’ (Murray to Gell [possibly a first draft], 15 June 1897, MP12).

Opportunity and invisibility

While procedural objectivity may have roots in practical considerations regarding communication challenges, Murray’s tactical two-sidedness suggests that it served a rhetorical purpose beyond the practical. Therefore, its historical idealization as an epistemological ethic may have been driven by the opportunistic pursuit of legitimacy. This explanation differs from Daston and Galison’s (2007) claim that practical considerations gave way to internalized psychological pressures to fear and deny the self – though it may suggest an intermediary stage in this process. Those who lacked social standing or propriety had reasonable motivation to insist on other grounds for their legitimacy and participation in knowledge production. Appeals to impersonal procedure thus presented a path to legitimacy and control on the only terms attainable to the dictionary’s editor and his motley assortment of amateurs lacking resources, concordance, and power. It was also a path to legitimacy that, at the time, did not have to compete with a well-established and bounded professional realm of British philology. Murray’s (1977) granddaughter commented that ‘[i]n an age when paper qualification, however useful, were not yet the essential passport to advancement which they have now become, perhaps in some ways the opportunities were greater than today’ (pp. 339–340).

But it wasn’t as if social characteristics suddenly didn’t matter, and circumventing such character assessments required touting the impersonal aspects of production beyond the actual extent of their application – even this idealization, as indicated above, left Murray in a bind and unable to gain private recognition for the personal talent and sacrifice he and others contributed to the work. A parallel can be found in Oreskes’ (1996) research on the invisibility of women in science, who ‘have been characteristically employed not in jobs that required a high degree of emotional involvement or contextual judgment, but precisely the opposite’ (p. 89).

It is a bittersweet consolation to imagine that the idealization of procedural objectivity may have provided a means for traditionally less visible and less empowered groups to play a role in knowledge production, even as it would obscure their contributions. But in this sense, freedom for participation can be reconciled with the ideal of freedom from authorial dictate. The former is emphasized in Tresch’s (2010) notion of ‘multi-perspectival’ objectivity as a value articulated by Alexander von Humboldt, who organized several large-scale collaborative research projects around the turn of the 19th century. The latter is expressed by Daston and Galison (1992) when they speak of mechanical objectivity as ‘shifting the interpretive eye to the reader’ (p. 107). It was also stated quite precisely by a 1911 reviewer of the dictionary, who enthused ‘in no other book I know of is such freedom from mental oppression to be found: here there is no author’s arbitrary handling of the material of life to irk the reader’ (cited in Brewer, 2007: 97). Such praise was a significant change from the early critics who admonished the dictionary’s descriptive approach and wealth of minutia. Porter (1995, 2004), too, is cognizant of the paradoxical entanglement of freedom and egalitarianism with oppressively rigid standardization.

Advantages and limitations

The ‘legitimacy-opportunity’ thesis is speculative and requires further substantiation, but understanding why and when scientific actors deploy different ‘tactics of credibility management’ can have important implications for recognizing the strengths and weaknesses of various forms (Shapin, 1995). On the one hand, the practices of knowledge production have clear implications for the content of knowledge. Collaborative production models similar to that of the NED seem to be instrumental in generating abundant, aggregate information that can be used for varied purposes over generations. This is apparent in varied contemporary ‘crowdsourced’ models of knowledge production, such as Wikipedia, Project Gutenberg, Galaxy Zoo, and the Encyclopedia of Life. In forthcoming work I argue that the Internet has built in or easily developed infrastructure that mirrors the kinds of standardization, boundary objects, and evaluation processes set up by Murray (Kistner, forthcoming). Therefore, nowadays it is easier to utilize the knowledge of diverse contributors and the benefits that come from multiple sets of eyes. The OED itself is again jumping aboard on this trend (see http://public.oed.com/appeals/).

But, while this kind and amount of information is valuable, it is also ‘thin’, as noted by Porter (2009). Skills may thus get overshadowed and devalued; underlying narratives, theory, and purpose can get lost. This was, in fact, evident in the case study underlying the work by Star and Griesemer (1989). Therefore, while such approaches may facilitate being heard and being a part of the program, they are intolerant of speculation, lofty analysis, or the identification of problems and paths to their solutions. As Peter Sutcliff (1978) says of the OED editors,

Murray and his successors did not dwell on the romantic aspects of lexicography, on the mysteries of language and the beckoning enticements of etymology that lured them on. Lexicography had mostly to do with slips of paper measuring, preferably, six inches by four. (p. 57)

This is a predicament familiar nowadays in citizen science collaborations between scientific professionals and public volunteers. Citizens wishing to express their local knowledge, experiences, and concerns often find that their participation requires adherence to procedures that effectively erase their unique perspective and restrict their role in decision making (e.g. Cornwell and Campbell, 2012; Ellis and Waterton, 2005; Ottinger, 2010). However, similar tendencies are also seen among today’s professional scientists. Restivo (1988) criticizes the ways in which scientific knowledge work has become routinized and stylized in ways alienating it from its producers. Porter (2009) and Epstein (1995) have also noted that scientific work can be alienating to potential consumers, contributing to mutual distrust and misunderstanding between scientists and the public.

At the same time, personal influence and judgment cannot be fully expelled from scientific process, and supposing that there is no subjective component to knowledge production can undermine the responsible use of scientific claims. In her studies on the Challenger disaster, Vaughan (1999a) found that expertise and insight do not transfer well outside of local contexts and can leave dangerous gaps in understanding, which are further exacerbated by the tendency of procedural protocols to effectively make uncertainty invisible. In a similar vein, Frickel and Vincent (2007) found that conventional measures of environmental contamination and risk assessment minimize the relevance of distinct contextual features, leading to a kind of persistent ‘organized ignorance’. Furthermore, feminist scholars such as Keller (1985) note that obscurantist practices of and contemporary appeals to objectivity can conceal prejudicial tendencies and biases in scientific practice.

Conclusion

In more recent times, scholars have examined the extent to which the OED falls short of its aim to be an unbiased, complete, and accurate historical record of the English lexicon and its usage (e.g. Brewer, 2005, 2012; Schäfer, 1980; Willinsky, 1994; throughout Mugglestone, 2000a, 2005). They offer important critical assessments and corrective responses to claims about the objectivity of the dictionary; I would entirely agree that the makers and methods of the first edition of the OED were unable to evade systematic or incidental bias. In fact, this likelihood is implicated in the narrative I have offered. However, the problems and questions to which these researchers are responding are motivated by contemporary epistemic values and expectations. Some of these critiques appear simultaneously to discount the possibility of producing a completely unbiased text, while still upholding impersonality as ideal. In contrast, this article has examined why the dictionary was made in a manner that, at the time, was historically unprecedented, not self-evident, and certainly not smooth-sailing.

I have argued that the organizational model and features of production were partly responses to communication, coordination, and control problems within the heterogeneous context of amateur philology in Britain. This supports existing work in STS and provides an example outside the realm of the natural sciences and professional institutions. As contemporary knowledge production increasingly crosses boundaries – for instance, across disciplines, in conjunction with varied administrative and policy channels, and in citizen science movements and public collaborations – it is important to understand why varied organizational forms take shape and what their advantages and limitations are.

But the case also sheds light on an entire era, one that was shaped by uncertainty, but in which individuals did not necessarily succumb to fear of themselves (as suggested by Daston and Galison, 2007). Instead, by contributing to public knowledge projects, it was possible for individuals in this age to become a part of something larger than themselves. This is not without its gains and losses. I am reminded of Hannah Arendt’s (1998 [1958]) remarks on the transformation of humans from workers (as in craft work) to laborers: the latter engaged in what

would appear not as activities of any kind but as processes, so that, as a scientist recently put it, modern motorization would appear like a process of biological mutation in which human bodies gradually begin to be covered by shells of steel. (pp. 322–323)

In light of my research I’ve come to imagine the shell of steel, (which is a more literal translation of Max Weber’s stahlhartes Gehäuse than Talcott Parson’s ‘iron cage’), as a coat of armor encasing the individual – it offers strength and protection that is impenetrable to outsiders, but also limits one’s range of motion, obscures the human within, and yet, being steel, is entirely man made.

Upon Murray’s death in 1915, Bradley (1915) proclaimed that ‘The great English dictionary will always be known chiefly by his name, with far stronger reason than the great German dictionary bears the name of Grimm’ (p. 7). Nevertheless, any degree of recognition of the unique personal qualities and skills of Murray, as well as the rest of the dictionary staff and collaborators, would fall into abeyance over time. The Press would stop printing Murray’s preface after the completion of the first volume in 1928, and a new dust jacket of the work would proclaim it the Oxford English Dictionary, rather than the words originally covering its title page: ‘A New English Dictionary on Historical Principles; Founded Mainly on the Materials Collected by the Philological Society. Edited by James A. H. Murray, L.L.D, with the assistance of many scholars and men of science’.