Background and history

(Source: Margaret Laing, Introduction to A Linguistic Atlas of Early Middle English, Preface, http://www.lel.ed.ac.uk/ihd/laeme1/laeme_intro_preface.html)

Much of the thinking behind A Linguistic Atlas of Early Middle English (LAEME) evolved gradually over the twenty years of its making. Given the very small amount of data available for early Middle English compared to that usable in its great predecessor A Linguistic Atlas of Late Mediaeval English (LALME), I expected the LAEME project to take a far shorter time than it has done. In fact, for a number of reasons LAEME and its associated ‘daughter’ atlas A Linguistic Atlas of Older Scots (LAOS) have become very different kinds of works from LALME, and their histories explain why. They show how necessity drove us to develop a methodology that makes the new atlases more powerful research tools than we initially envisaged them. We hope, indeed, that they may (perhaps rather far in the future) become models for an entire reworking of LALME itself along the same lines. This particular history also offers an apologia for what LAEME does not do, and what it has not (so far at least) achieved that it might have promised or been expected to achieve.

In 1987, after the publication of LALME, a decision had to be made as to what would happen next. Of the three main compilers of LALME, two were heavily engaged in other projects. As early as 1964, M.L. Samuels had initiated The Historical Thesaurus of English at the University of Glasgow. With LALME safely published, he was able, with his colleagues, to devote his full attention to this other very large-scale investigation. [1] In 1983, because of the precarious financial situation facing the Middle English Dialect Project during the final years of the compilation of LALME, Michael Benskin had left the University of Edinburgh and taken up a post in the University of Oslo. From there, as Professor of Older English, he had written much of the introductory material for LALME. Since its publication, he has continued to do a great deal of further work on late Middle English in local archives, augmenting and correcting the LALME materials towards a second edition, the necessity of which in due course was always envisaged. [2] He has also continued his own research work on medieval Hiberno-English and on the beginnings of Standard English.

In 1979, Angus McIntosh had retired from the Forbes Chair of English Language and from head of the Department of English Language at the University of Edinburgh in order to be able to spend more time on the completion of LALME. In 1987 he was still academically active, and became Honorary Consultant to what was to become the Institute for Historical Dialectology, at that stage part of the Department of English Language. [3] The Institute had a small endowment (made up of a personal gift from Lt Col Gayre of Gayre and Nigg), which had been generously supplemented by the University of Edinburgh from two other legacies. It had two Research Fellows: Keith Williamson and me. We had both been post-doctoral fellows helping, since 1979 and 1982 respectively, with the completion of LALME. After the death of James Thorne, Angus McIntosh’s successor in the Forbes Chair, we were fortunate to secure Derek Britton, the English Language Department’s Middle English specialist, as interim Director of the Institute.

In June 1987, I had just returned from a two-year stay with my husband and two young children in Boston, Massachusetts. Keith was recovering from what turned out to be the massive task of getting LALME into camera-ready form for publication. At that stage, he and I both decided that we would stay on at the Institute and continue its work. Because of family commitments I would only work half time. Even before LALME's publication, some thought had been put into a further generation of linguistic atlases. The most obviously desirable and practicable ‘daughter‘ atlases were deemed to be those for early Middle English and for Older Scots.

It was known that for the earlier period of Middle English there would not be as full geographical coverage of surviving material as we had for LALME, but many literary texts from the late 12th and the 13th century were well known, and had already been edited and extensively discussed in the scholarly literature. It was clear that these early texts also displayed signs of regional variation and it could be assumed that a detailed investigation of the earlier period, along the lines of the LALME project, would yield interesting results. The Older Scots sources had only been given cursory attention in LALME. The documentary sources were largely inedited or existed (for use mainly by historians) only in linguistically normalised editions. There was also a widespread assumption, which badly needed testing, that the literary texts were to a great extent in a kind of literary standard. Because of our respective academic backgrounds in Scots and in Middle English, Keith elected to take on LAOS, and I LAEME. The hubristic lunacy of undertaking two large-scale projects with limited funds and one and a half people did not occur to us.

Backed by my experience of working on LALME, I began work on LAEME on three different fronts: (a) the identification and cataloguing of potential sources; (b) the search, as part of this larger task, for local documentary texts from the period that might serve as anchor texts for the early Middle English dialect continuum; (c) the development of a suitable questionnaire with which to conduct the analysis. In the first task, I was fortunate in having a very good head start. McIntosh and Samuels, in their own quest for late Middle English source texts, had come across and made note of a number of earlier texts. Indeed, Samuels, who was responsible for the southern area of survey, had included in LALME a dozen or so of the late 13th-century manuscripts that he considered to be too important to leave out. [4] Patrick Stiles, who was employed by the Institute as a Research Fellow between 1984 and 1986, and who was not engaged (as Keith and I were) in work on LALME, had put the early Middle English card file in order and considerably augmented it. Patrick was responsible for causing the Institute to buy its first PC – in that era an expensive and bulky machine, which we had to share, and which laboriously churned out from its daisy-wheel printer rather pale and unattractive hard copy. By the time he left the Institute, Patrick had keyed the early Middle English source material onto disk. I continued his work, augmenting it still further and turning it into as full an index of potential sources as was then feasible. [5] I would like to thank Patrick for his invaluable work towards the original catalogue. I would also like to thank Richard Beadle, Michael Benskin, Ian Doyle, Alan Fletcher, Helmut Gneuss, Peter Kitson, Christian Liebl and Oliver Pickering for extremely useful comments and contributions to source material since its publication.

The identification of suitable local documents turned out to be a chimera. Though disappointing, this was not really surprising. The lack of local documentary anchor texts in English from before 1350 was one reason why McIntosh had chosen the late Middle English period for his investigation. But McIntosh had already identified in his work on the localisation of the early Middle English Havelok (McIntosh 1976 and cf. Lowe 1992, 1993) a body of Old English documentary material associated with Bury, Suffolk, that had been copied during the 13th century into registers in Bury Abbey. The copying scribes, and probably already to a certain extent their precursors, had updated the original Old English so that their texts could be taken as representing the English of their own period. Ironically, this material seemed not to be in the early Middle English of Bury itself, but rather to belong further north in West Norfolk. It did not constitute therefore the normal kind of anchor texts. Nevertheless, our hope had been that we would be able to find many more late 12th- and 13th-century copies of pre-Conquest vernacular documents and that some at least of them could serve as local anchors. In the event, a detailed trawl through all the documents listed as having versions copied in the early Middle English period in Sawyer (1968) and Pelteret (1990) resulted in only a handful of usable documents. [6]

In the early stages of LAEME (continuing the long-standing collaborative links with the English Language Department of the University of Glasgow) it had been intended that Jeremy Smith, as a Middle English specialist, and Kathryn Lowe, as an advisor on Old English documents, would collaborate with me. In the event, they both (wisely) became engaged on their own projects: Katie in editing the complex documentary material from Bury St Edmunds and Jeremy in writing a large number of books and articles, and in setting up the Glasgow-Stavanger Middle English Grammar Project (to which the LAEME materials also contribute). [7] Katie provided LAEME with initial transcripts of the early Middle English copies of Bury documents and Jeremy has provided, over the years, initial transcripts of a number of the Ancrene Riwle texts and the Lambeth Homilies. I here gratefully acknowledge this very valuable help and their continuing collegial support, as well as that of their Glasgow colleagues Graham Caie, Christian Kay, Carole Hough and Irene Wotherspoon.

At the same time as producing the catalogue of source material, I was also working hard to develop a questionnaire suitable for the dialectal analysis of early Middle English. I had the LALME questionnaire as a model. Because part of the point of the survey of the earlier period was comparison with the later material, I assumed that most of the LALME items would also be included on the LAEME questionnaire. I was aware, however, both of the shortcomings of the LALME questionnaire, and that the variables for early Middle English would be of rather a different kind. The shortcomings were operational rather than theoretical. The LALME questionnaire was hugely more detailed than anything that had been previously applied to the late Middle English sources. For very good methodological reasons (chaque mot a son histoire) it had not included many items eliciting phonological categories. One operational problem was that it was very difficult for the analyst to collect data consistently for these items. For categories such as a/o (the reflexes of OE ā) or wh- (reflexes of OE hw-) one had to record for that item the segments not just from the relevant words that featured already on the questionnaire as items in their own right (for which one also had to record the full spelling in the relevant place), but also for those that were not items in their own right.

For early Middle English, though fully acknowledging lexical specificity, I wanted to include many more phonological items that I thought would be of particular interest at the period, such as the reflexes of OE ȳ / y̆ and of OE o. Of potential interest also for the history of the early Middle English period were orthographical items such as the temporal and regional distributions of ‘æ’, insular ‘g’ <ᵹ> (as distinct from yogh <ȝ>), of wynn <ƿ> and <w>, and the contextual distributions of thorn <þ> and edh <ð>. It was also desirable to record evidence for the retention or levelling of grammatical inflexions at this period. When all these variables were added to the early Middle English questionnaire it became an operational nightmare. The likelihood of inaccuracy and inconsistency in the recording of information was very high in an analytical method that required constant cross-reference by hand.

Meantime Keith was abandoning the idea of using a questionnaire for somewhat different reasons. For the analysis of the Older Scots material, the LALME questionnaire was not a suitable model. Not enough was yet known about the linguistic variation in Older Scots for the inductive method to be used safely. So he began to devise a tagging program so as to taxonomise all the linguistic data from his source texts. It would then be possible, from the resulting text dictionaries, to extrapolate a questionnaire after the enquiry from a full inventory of every scribe’s entire output – or of large samples of any very extensive texts. With my difficulties mounting in the operation of the early Middle English questionnaire, I soon enthusiastically followed his lead (Williamson 1992/3; Laing 1994).

At this point, however, I made a major error in my approach to the new methodology. With my dialectological background still firmly in the tradition of the questionnaire, I saw the process of tagging the early Middle English texts as a means primarily for the ultimate creation of linguistic profiles for dialect mapping. I did not at this stage have in mind that what I would be creating was a corpus of lexico-grammatically tagged texts; nor did I fully consider the much wider possible applications of such a resource. So when I began transcribing the early Middle English texts for tagging, I did not include textual ‘details’ such as punctuation, accompanying Latin tags and quotations, notes of corrections or additions by other hands, or even – at the beginning – manuscript line ends. Gradually, in the course of building up the corpus, I began to rectify these omissions, but as a result of the early failure, I am still, at the time of writing, in the process of going back to the microfilms and adding manuscript punctuation, embedded Latin text and marginal notes to a corpus of nearly 650,000 tagged words.

In 1991 the IHD was very fortunate to find a new home in the School of Scottish Studies whose director and head of department was Alexander Fenton. It was under Sandy’s aegis that in 1993 we gained our first major research grant of £245,290 from The Leverhulme Trust to pay both our salaries for five years. At this stage I moved from half-time to full-time work on LAEME. On Sandy’s retiral, Margaret Mackay became director of IHD and took over as grant holder. After five years, there were a number of major publications showing the progress of the two projects. Those illustrating the LAEME methodology included one on the importance of different text versions for identifying dialectal discriminants (Laing 1992), and two on the detailed analysis required to sort out dialectal stratification in composite texts (Laing and McIntosh 1995a and 1995b; and see further the LAEME bibliography).

At this stage we realised that the projects, while firmly placed in the LALME tradition, were of a very different kind from LALME (Laing 2000b). They were no longer just dialect atlases in the making. We were also creating text corpora with a potentially very powerful element of lexico-grammatical tagging that could facilitate a wide range of study at all linguistic levels (see, for instance, Laing 2002). It was also clear that as a result of adopting this powerful but labour-intensive methodology, far from having nearly finished the projects, we had only just begun. We were very fortunate that the School of Scottish Studies was able to find bridging funds for the IHD for six months [8] before we gained further grants for 1999–2000 from the British Academy (for LAEME) and the Carnegie Trust for the Universities of Scotland (for LAOS). From 2000–2003 and 2003–2006 the work of the IHD was very generously supported by two major research grants (of £309,360 and £311,487 respectively) by the AHRB/AHRC.

In 2002, at the time of major restructuring of the University of Edinburgh, we accepted the invitation of Heinz Giegerich, then Head of Department, to return to English Language. We are immensely grateful for all the support we have had from him, from April McMahon, and from our other colleagues in English Language (now merged with the former department of Theoretical and Applied Linguistics as Linguistics and English Language in the larger School of Philosophy, Psychology and Language Sciences). We are especially delighted that Derek Britton agreed to resume his role as Director of IHD. We owe him a very special debt for his constant, cheerful, and always optimistic support.

Throughout the six years of AHRC funding, the LAEME corpus continued to grow, and as a result of the detailed tagging there emerged a great deal of evidence on the complexity of scribal copying practices and writing systems in early Middle English (Laing 1997, 1998a, b, c, 1999, 2000a, 2004). Roger Lass of the University of Cape Town expressed an early and strong interest in some of these results, and especially in the interface between the complex early Middle English writing systems and the underlying sound substance. His persistent sense of excitement at what I was achieving with the LAEME materials seemed too good an opportunity to miss, and I asked if he were interested in collaboration. It is my great joy and privilege that in February 2002 he agreed to join me in working on LAEME. [9] Our work on the reconstruction of early Middle English sound systems (Laing and Lass 2003, 2005, 2006, Lass and Laing 2005) forms the background to the potestatic maps (see Introduction Chapter 6 forthc. [pdf])

When Roger joined LAEME, he was at first anxious that coming into the project so late in its progress he would not have the opportunity to help as much as a full collaborative role should imply. His solution was characteristically creative and grandiose: he joined LAEME on the understanding that his main contribution would be to create another whole layer to its structure. He offered to provide for LAEME (and it is hoped in due course also for LAOS) a corpus of etymologies that would present a ‘story’ to explain every single spelling recorded in the corpora of tagged texts. The corpus of etymologies and its accompanying corpus of sound changes (for both of which see Introduction, Chapter 8 [pdf]) provide a greatly enhanced historical dimension. The assessment of the processes of change from Old English into Middle English is now an integral part of LAEME and will be covered in a detail not at first envisaged. The corpus of etymologies, which was only started in 2002, is ongoing work but at the time of writing is in part usable. New etymologies will be added as they are completed.

And this leads to my apologia. Because LAEME is a very large project, and because funding for it has not been limitless, quite apart from the fact that the corpus of etymologies is incomplete, LAEME does not yet contain everything that it should contain. I have not followed up the work on early Middle English glosses that are listed in the Catalogue of Sources. I have not had time to look at all the post-1300 copies of Anglo-Saxon charters that were also listed there in the hope that their language might represent an intermediate historical stage. I have not even tagged all the texts I ought to have tagged. There are still early versions of the South English Legendary and Cursor Mundi to be transcribed and processed (see further, Introduction Chapter 3, §3.1 [pdf]). But the corpus nevertheless contains nearly 650,000 tagged words of early Middle English. It is hoped that more will be added in the future.

As we came towards the end of our second AHRC grant we began to think seriously about how we would disseminate the huge amount of processed data we had accumulated. The projects had evolved almost beyond recognition from how they had first been envisaged. Our publication strategies had to evolve too. It had become apparent that to publish the materials in the form of books or CDs would not provide a flexible enough medium. Few resources of this size and complexity (like dictionaries) are ever wholly ‘correct’ or completely finished. The option of publishing LAEME and LAOS as interactive websites seemed the answer to a host of problems. We could put the materials into the public domain in stages as they became ready. We could update, modify and correct them in response to comments and criticisms. With suitable software the atlases can become interactive research tools – the software too can be added to and upgraded into the future. Keith Williamson has been responsible for all the thinking and the practicalities of this exciting and liberating decision. His contribution to LAEME has been colossal. [10]

The IHD has been given a great deal of academic and moral support by a very large number of people, many of whom have been mentioned above. We are no less grateful to those who are not individually cited. I would like to give particular and special thanks here to Ian Doyle and Bob Lewis who have been generously and tirelessly supportive through all the years. I end with heartfelt and loving thanks for all they have done, to the late Angus McIntosh without whom LAEME would not exist, to my dear friend and constant mentor Michael Benskin, and to my closest partners in this wonderful enterprise, Keith Williamson and Roger Lass.

Notes

[1] Now under the directorship of Christian Kay, publication of the Historical Thesaurus is expected in 2007.

[2] A revised on-line edition of A Linguisitic Atlas of Late Mediaeval English (eLALME) is now in progress, generously funded by AHRC Resource Enhancement Scheme grant to the University of Edinburgh. The Principal Investigator is Derek Britton, with Keith Williamson and me as main researchers and Michael Benskin (University of Oslo) as Project Partner.

[3] Angus McIntosh remained a much valued and much loved consultant until his death, aged 91, on 26 October 2005.

[4] I am very grateful to Michael Samuels for so generously making over the results of his early searches and for his continued interest and support over the years.

[5] The result was published as Laing 1993.

[6] See further Laing 1991 and Introduction Chapter 1, a7;1.5.3 with references there cited.

[7] The MEG project (under the directorship of Jeremy Smith in Glasgow and Merja Stenroos in Stavanger) is assembling machine-readable and diatopically-ordered corpus of late Middle English texts, using 3,000 word samples transcribed from either the original manuscripts or from microfilms. This corpus uses the same conventions for transcription as LAEME and is designed to be compatible with it. One aim of the MEG project is to produce a ‘diachronic and diatopic interpretative survey of Middle English orthography and phonology, based on published and unpublished material, with the initial aim of producing a replacement of Jordan's Handbuch’. There are close links between the Glasgow/Stavanger project and IHD. During spring/summer 2007 Dr Martti Mäkinen, a Research Fellow on the MEG project in Stavanger, was based at IHD and working on the microfilms in the Edinburgh University Library’s Special Collections. See further at http://www.arts.gla.ac.uk/STELLA/ihsl/projects/MEG/meg.htm.

[8] We are very grateful indeed to Margaret Mackay for this and to both her and Sandy Fenton for continuing generous interest and support.

[9] He and Dr Anneli Meurman Solin of the University of Helsinki are both much-valued Collaborating Scholars of the IHD. Anneli's Corpus of Scottish Correspondence uses the same tagging methodology as LAOS and LAEME. Her concerns with Older Scots mesh more closely with LAOS than with LAEME, but LAEME has also benefited greatly from her insightful and scholarly comments and ideas, and from her passionate enthusiasm for our subject.

[10] I am also very grateful to Ellen Bard and to Jean Carletta for their good offices in securing the invaluable, creative and patient help of Vasilis Karaiskos in preparing the LAEME Index of Sources and the Introduction for web publication.