Historical dialectology in the digital age presents a series of studies, most of which are based on papers originally presented at the First Angus McIntosh Centre Symposium on Historical Dialectology, held at the University of Edinburgh in June 2016. As such it celebrates the legacy of Angus McIntosh, demonstrates the continuity of research in this field and highlights the potential of digital technology for opening up new approaches and facilitating new insights. The editors define historical dialectology as ‘the study of diachronic, diatopic and social variation in the historical record of languages’ (p. 3). Like its synchronic sister, historical dialectology, far from being superseded by sociolinguistics and corpus linguistics, has embraced and been revitalized by these approaches and has also benefited from interdisciplinary research involving disciplines such as geography and paleography. The chapters in this volume provide a fascinating insight into the state of the art of historical dialectology, and, in the span of age and experience represented by the authors, ranging from Emeritus professors to doctoral candidates, reassure us that the future of this discipline and the legacy of Angus McIntosh are in safe hands.
Following an introduction by the editors, Rhona Alcorn, Joanna Kopaczyk, Bettelou Los and Benjamin Molineaux, the volume is divided into three parts: ‘Creating and mining digital resources’, ‘Segmental histories’ and ‘Placing features in context’. The three chapters in part 1 provide accounts of new digital resources: the Parsed Linguistic Atlas of Early Middle English (PLAEME), the Dunfermline Corpus, and the From Inglis to Scots (FITS) database. Each of these chapters combines a description of the corpus or database concerned with reports of studies illustrating how they can be used to answer research questions. All give an exciting sense of the potential of these new resources for further research, suggesting that there is still much work to be done in the field of historical dialectology. In chapter 2, Robert Truswell, Rhona Alcorn, James Donaldson and Joel Wallenberg explain that the PLAEME corpus has been constructed to facilitate morphosyntactic analysis of texts from the LAEME corpus, concentrating on the period 1250-1350, for which there is a gap in the Penn Parsed Corpora of Historical English (PPCHE), hitherto the main source of data for such studies. Because of this gap, they argue, ‘much of the action happened while our backs were turned’ (p. 29). Klaus Hofmann argues in chapter 3 that ‘existing studies have barely scratched the surface’ (p. 39) of the effects of contact between Scots and English in the period between the Scottish Reformation (1560) and the Union of the Parliaments (1707) and later makes the intriguing point that ‘a comprehensive study of the linguistic consequences of the British Civil War is still outstanding’ (p. 53).
Part 2 consists of four chapters dealing with spellings and sounds in Middle English and the challenges involved in deriving phonological information from manuscripts of this period. Chapter 5, by Margaret Laing and Roger Lass, reports on research using four electronic resources – the Linguistic Atlas of Early Middle English (LAEME), the electronic version of McIntosh's Linguistic Atlas of Late Middle English (eLALME), the Linguistic Atlas of Older Scots (LAOS) and the Corpus of Narrative Etymologies from Proto-Old-English to Early Middle English and accompanying Corpus of Changes (CoNE) – to investigate variation in spellings for Old English <hw> and what these resources can tell us about the correspondences between the various spellings and the sound changes for which they provide evidence. Gjertrud Stenbrenden in chapter 6 draws on evidence from the Survey of English Dialects (SED), LAEME and LALME for spellings of the reflexes of æ1 and ǣ2 in Middle English (ME) dialects. In chapter 7, Merja Stenroos makes use of the Middle English Grammar Corpus (MEG-C) to address the question of whether a systematic distinction can be found between reflexes of Old English e/ ē and eo/ ēo. The final chapter in this part sees Donka Minkova examining evidence for affricates in Middle English and addressing the thorny issue of whether these consist of single segments or combinations of stop + fricative.
According to the editors, the three chapters in part 3 ‘look at the minutiae of the use of certain features in geographically-defined subsets of texts’ (p. 11). At first glance, this seems an arbitrary classification as all studies in historical dialectology do this, but perhaps the justification for this grouping lies in the word ‘minutiae’. These three chapters all zoom in either on a particular text or on a very specific feature. In chapter 9, Daisy Smith focuses on a particular abbreviation marking the plural in order to investigate the status of <is/ ys> spellings for the plural morpheme in Older Scots. Ad Putter in chapter 10 deals with a single manuscript, demonstrating how electronic resources can be used to challenge previous accounts of its place and date of origin. Finally, Trinidad Guzmán-González in chapter 11 draws on evidence from texts localised as from the South-West of England in LAEME, the Helsinki Corpus of English Texts (HC) and MEG-C to investigate whether the gender system recognized by dialectologists as characteristic of south-western English dialects can be traced back to Middle English. Her research answers this question in the negative, thus challenging the notion of traditional dialects as fossils of earlier stages in the language.
The above outline gives a flavour of the variety of topics and approaches covered in this volume. To go into detail about the content of each chapter would be beyond the scope of this review, so I shall instead draw attention to the ways in which the volume provides new insights and highlights new approaches. Labov is often quoted as stating that historical linguistics involves ‘making the best use of bad data’ (Reference Labov1994: 11). This, of course, raises the question as to what constitutes ‘good’ data, but his point is that we have to deal with the resources available rather than being in the position of the synchronic linguist who can collect data according to scientific principles. Historical linguists have to make the best use of the data available. This may mean discovering or adapting resources previously little used by linguists, such as the Old Bailey Proceedings, digitized as a historical resource but adapted into the Old Bailey Corpus (Huber Reference Huber, Meurman-Solin and Nurmi2007) for historical linguistics. The further back in time we go, the less likely it is that ‘new’ data will be found, and the work in this volume demonstrates how advances in corpus methodology, the creation of new corpora from ‘old’ resources, and insights from allied disciplines can wring new discoveries and insights from this ‘bad’ data.
One challenge faced by historical dialectologists, especially those dealing with earlier stages of the language from which less written material and no spoken material survives, is the lack of resources. In terms of material resources, there is simply less available, and in terms of financial and human resources, it takes much longer to process. This point is made clearly by Smith in chapter 9. Smith here argues for the use of ‘sophisticated statistical methodology’ (p. 187) such as mixed effects regression modelling (Gries Reference Gries2015). She points out, however, that the ‘big data’ to which these methods normally apply is hard for historical dialectologists to find and process. A footnote from Stenroos explains that one text which would have been useful for her research could not be included in MEG-C ‘because it is a nightmare to transcribe’ (p. 152 fn 10). Smith compares the Edinburgh Twitter Corpus of over 2 billion words with LAEME, consisting of approximately 650,000 words, and tells us that the former took two months to compile whilst the latter took twenty years. In other words ‘the Twitter corpus could be compiled in one-twentieth of the time it took to compile LAEME, and in doing so contain[s] around 150 times as many words’ (p. 188). Historical dialectology involves painstaking manual transcription of manuscripts that are often hard to decipher: this work simply cannot be outsourced. Thus Truswell et al. in chapter 2 note ‘we are presently unable to parse the whole of LAEME for lack of resources’ (p. 25): they have had to make the best use of the data they have and of the resources available for the exploitation of this data. Nevertheless, the chapters in this volume prove that the data currently available is rich, and that the electronic resources and statistical methods now available allow us to make sense of variation that might otherwise appear chaotic. Stenroos argues for the systematicity of Middle English spelling by examining a group of West Midland texts which maintain the distinction between Old English eo/ēo and e/ē ‘to a remarkable extent, but always with exceptions’ (p. 133). She notes the challenges involved in such a study: for later periods spelling mistakes or back-spellings may be used as evidence of mergers, but as Stenroos points out, when dealing with Middle English ‘the absence of a standard model makes the concept of “spelling mistake” problematic’ (p. 136). By using a concordancing programme (AntConc 3.2.1) with texts from MEG-C localized to the south-west of England, Stenroos demonstrates that the use of <eo> spellings in these texts is not random, but ‘partly geographically restricted and remarkably consistent’ (p. 151). She suggests that the persistence of these spellings in this area may not necessarily reflect the survival of a phonological distinction, but may have ‘signalled a specific western identity’ (p. 153). This suggestion that the use of orthographic variants could be used to signal identity by Middle English scribes just as they are by twenty-first-century users of social media provides a fascinating example of the uniformitarian principle at work. Stenroos concludes that the spelling of these scribes was ‘variable but certainly not chaotic’ (p. 153), reminding us of the ‘orderly heterogeneity’ (Weinreich et al. Reference Weinreich, Labov, Herzog, Lehmann and Malkiel1968:100) behind the apparent chaos of linguistic variation.
In sociolinguistic research, the connection between linguistic variation and identity is perhaps most overt in ‘third wave’ (Eckert Reference Eckert2012) studies in which speakers are observed operating within communities of practice. Hofmann explains that the Dunfermline Corpus ‘is narrowly focused on scribal idiolects of individual town clerks and scribes, conceptualised here as a community of practice’ (p. 39). Quite apart from the fact that it introduces a new corpus, Hofmann's study is novel in two ways. Hofmann does use the term ‘Anglicisation’ to describe the process of replacing characteristically Old Scots variants with Southern English ones during the period covered by the corpus. However, following Kopaczyk (Reference Kopaczyk, Millar and Cruickshank2013), he rejects the use of this term ‘as a wistful conceptualisation of this period as one of decline for the Scots language’ instead seeing it as ‘a time characterised by striking dynamism due to increased contact between Scots and English standardising usage’ (p. 41). Having shed this ideological baggage, Hofmann is able to concentrate on the linguistic and extra-linguistic information provided in the Dunfermline Corpus to discover how the relationships and practices of the community of scribes influence the choice and transmission of variants. He discovers that, when there was continuity of transmission between clerks and their assistants, the introduction of anglicised variants was slowed down, but when a new clerk was introduced from outside this network, this was accompanied by ‘significant strides towards an English-looking orthography’ (p. 53). Although this seems at first blush to echo the patterns found in social networks research, Hofmann points out that such an approach considers only the strength and density of ties, whereas the community of practice method used here takes account of the nature of those ties: in this case, the relationship between master and apprentice which is disrupted by the arrival of a scribe trained elsewhere. He further notes the breakdown of this transmission of practice by 1700, by which time other means for the implementation of standard practice, such as schools and dictionaries, had taken over.
So far, I have emphasized the innovative nature of the studies in this volume, but it is clear that the authors, whilst making use of the digital resources available to perform sophisticated statistical analysis and to produce clear maps, graphs and tables, never lose sight of the original manuscripts as important artifacts in their own right. A sterling example of this is chapter 9. Here Smith, having discovered via statistical modelling that the symbol under investigation is more likely to appear after certain stem-final letters (SFL) than others, notes that ‘the link between the SFL which make up these two groups is clear only on inspection of the manuscript forms themselves’ (p. 208). She provides illustrations of the paths of pen-strokes to explain how this pattern is due to the physical dynamics of moving from one letter to another, something the digitized corpus could not have revealed, at least in its present form. That some means might be found in future for incorporating this kind of information is not inconceivable.
Historical dialectology in the digital age thus gives the reader a greater understanding of the past and an intriguing glimpse into the future of historical dialectology. It will be of interest not only to historical linguists, but to those working with linguistic corpora of all kinds. The presentation is clear and accessible, but, given the predilection of research groups, and especially those working with corpora, for acronyms, a list of these would have been useful. Having read and enjoyed this volume, I still do not know what an LP is (at least other than a long-playing vinyl record). This is a minor cavil: the volume is a credit to the Angus McIntosh Centre for Historical Linguistics and proof of the vitality of this discipline.
