The Middle English Dialect Project
The Middle English Dialect Project (MEDP) was the original name under which was carried out the work which led to the publication in 1986 of A Linguistic Atlas of Late Mediaeval English, 1350–1450 (LALME). The principal authors were Angus McIntosh, M.L. Samuels and Michael Benskin, the work being carried out in Edinburgh, Glasgow and, latterly, also Oslo. A number of other scholars and students assisted during the early decades of what was a considerable enterprise for its time.
On the eve of publication of LALME, fresh funding from both outside and within the University of Edinburgh was provided. The MEDP was re-formed as the Institute for Historical Dialectology, with a brief to extend and develop the work begun by McIntosh and his colleagues. These two research units are the immediate forerunners of the Angus McIntosh Centre for Historical Linguistics within the School of Philosophy, Psychology and Language Sciences. Where the MEDP and the IHD focussed on medieval varieties of English and Scots, the Angus McIntosh Centre aims to widen even further the historical investigations of linguistic change through time and across space.
Over the decades involved in the MEDP, Angus McIntosh corresponded about medieval texts and dialectology with many of the top historical linguists and medievalists of the day. He also consulted mathematicians and computer scientists as he sought innovative methods to process and analyse the large quantities of complex data required for the project. This correspondence is still on file in the Angus McIntosh archive, though it has yet to be catalogued.
Angus McIntosh started the MEDP in 1952 four years after becoming Forbes Professor of English Language and General Linguistics in the University of Edinburgh. He was joined the following year by M.L. Samuels, then a lecturer in the department, later to become Professor of English Language in Glasgow. From then on, McIntosh took responsibility for the northern area of the Middle English survey and Samuels for the southern area. They adopted the same methodology as used for dialect surveys of modern languages, developing detailed questionnaires in which were recorded the orthographic forms of each scribal language. The responses to the questionnaire — termed ‘Linguistic Profiles’ (LPs) — provided the core data for the Atlas. LALME displays these data in different formats. There are the LPs themselves (vol. 3); lists of forms for each questionnaire item arranged geographically (County Dictionary, vol. 4); a set of ‘Item Maps’, showing the distributions of all the forms for an item printed on the map (vol. 2); a set of ‘Dot Maps’, displaying the geographical distributions of specific forms or features (vol. 1). The ‘Index of Sources’ (vol. 1) not only describes all the manuscript sources from which the Atlas linguistic data were drawn, but also many more manuscript and printed sources which were consulted for the making of the Atlas.
From 1969, Michael Benskin was a PhD student of Samuels in Glasgow, working on the dialectology of medieval Hiberno-English. He joined the MEPD in Edinburgh in 1973 and continued to work towards the publication of LALME in Norway, after his appointment as Docent and then Professor of Older English at the University of Oslo. The research assistants in MEPD for the latter stages of LALME were Margaret Laing, who came as a PhD student in 1974 and whose work on the medieval dialect material of Lincolnshire was incorporated into the atlas, and Keith Williamson, who came in 1982 to give two week’s editorial help, but stayed on to help co-ordinate the work of getting LALME into publishable state.
Angus McIntosh’s work for the making of LALME was ground-breaking in two ways:
(a) He observed (against the prevailing opinion of the time) that not all ‘copied’ Middle English texts display linguistic mixture – many scribes ‘translated’, that is converted, the language of their exemplar into their own kind of local language. From this it follows that a copying scribe can be as important a witness of a contemporary local written language variety as an original author is.
(b) Faced with the problem of large numbers of anonymous and unlocalised scribal texts, he assumed the general principle, drawn from the results of surveys of modern dialects, that regional dialect tends to change in an orderly way across space. That being so, if one could establish a set of maps for the language of texts which do have identifiable local associations (viz. local documents), then it should be possible to use these maps to establish the provenance of scribal languages not otherwise localisable. McIntosh devised a process to do this. It entails systematic comparison of an assemblage of features in forms recorded for a set of items in the text to be localised with the distributions of features in forms for the same items established on the maps. Areas of the map where the features recorded in the text to be localised do not occur are progressively eliminated. The tendency to orderliness of the distributions predicts that some area of the map should remain where the particular assemblage of the text to be localised should ‘fit’. The procedure became known as the ‘fit’-technique.
For a much fuller account of MEPD (up to 1981) see Michael Benskin, ‘The Middle English Dialect Atlas’, in Michael Benskin and M.L. Samuels (eds.) So meny people longages and tonges; philological essays in Scots and Middle English presented to Angus McIntosh (Edinburgh: The Editors, 1981), pp. xxvii-xli. This includes many important acknowledgements as well as a fascinating account of the early efforts to harness computational techniques to solve fitting problems in the northern area of the survey.