Rhona Alcorn, Joanna Kopaczyk, Bettelou Los & Benjamin Molineaux (eds.), Historical dialectology in the digital age. Edinburgh: Edinburgh University Press, 2019. Pp. xvi + 274.

The innovative methods devised by Angus McIntosh and his colleagues Michael Samuels and Michael Benskin in creating A Linguistic Atlas of Late Mediaeval English (LALME, McIntosh et al. 1986) revolutionised historical dialect research. The development of the ‘fit technique’, which allowed a text’s provenance to be scientifically identified on the basis of its linguistic features for the first time, and McIntosh’s groundbreaking insights on copying practices were to pioneer the use of variationist methodology in the field of historical dialectology (McIntosh 1973: 92; Benskin 1991). Historical Dialectology in the Digital Age, edited by Rhona Alcorn, Joanna Kopaczyk, Bettelou Los and Benjamin Molineaux, emerges from the First Angus McIntosh Centre Symposium celebrated at The University of Edinburgh in 2016 and is evidence that the Angus McIntosh spirit of harnessing traditional approaches to historical dialectology and cutting edge methodologies is in robust health.

The introductory chapter by the editors provides an overview of the digital repositories hosted by the Angus McIntosh Centre for Historical Linguistics that have emerged in rapid succession in little over a decade, detailing how they have grown, evolved and been repurposed. The years 2008 and 2013 were particularly significant. The ‘daughter’ atlases of LALME, A Linguistic Atlas of Early Middle English (LAEME) and A Linguistic Atlas of Older Scots (LAOS) were both released in 2008. The year 2013 witnessed the publication of the remodelled LALME in electronic format (eLALME, Benskin et al. 2013). The Corpus of Narrative Etymologies from Proto-Old English to Early Middle English (CoNE, Lass et al. 2013) and new versions of LAEME (Laing 2013) and LAOS (Williamson 2013) were also released. The work carried out by the Middle English Scribal Texts programme (MEST) at the University of Stavanger has continued to build on the LALME tradition by extending the digital repositories of late Middle English material. The Middle English Grammar Corpus (MEG-C, Stenroos et al. 2011), the initial stage of MEST, provides tagged and annotated diplomatic transcriptions of 410 LALME texts, while the more recent Corpus of Middle English Local Documents (MELD, Stenroos, Thengs & Bergstrøm 2017) comprises transcriptions of over 2000 fifteenth-century documents.

The three thematically-devised sections that follow showcase these Edinburgh-based, and related, corpora for the study of Middle English (ME) and Older Scots, and ongoing efforts at expanding the range of resources available. A strength of the volume is that it brings together many of the researchers involved in compiling such corpora. Part 1 (‘Creating and mining digital resources’) comprises three chapters that discuss new digital resources. Robert Truswell, Rhona Alcorn, James Donaldson & Joel Wallenberg describe A Parsed Linguistic Atlas of Early Middle English (PLAEME, Truswell et al. 2018), a syntactically parsed version of LAEME, which supplements the patchy Early Middle English coverage of the Penn Parsed Corpora of Historical English (PPCHE, Kroch & Taylor 2000) for the period 1250–1350. The authors’ replication of three recent PPCHE-based studies on ME syntax using PLAEME indicates that the new digital resource allows for a more nuanced understanding of linguistic change during a crucial period in the syntactic development of English.

The following two chapters describe efforts to enlarge and repurpose the digital repositories of Scots. Klaus Hofmann’s chapter reports on the Dunfermline Corpus, a new digital resource which expands LAOS into c. 1550–1700, a period in which written Scots evolved under the influence of written English. Based on a study of markedly Scottish forms, Hofmann shows that the Anglicisation of Scots in local records was notably slower than at the supralocal level. He argues that local networks of clerks with strong professional and personal ties created a ‘community of practice’ that preserved the transmission of shared linguistic features. English orthography only started to creep in with the arrival of new clerks from outside the immediate scribal network. The From Inglis to Scots (FITS) project database comprises texts from A Linguistic Atlas of Older Scots (LAOS) that are grapho-phonologically parsed with the aim of reconstructing the sound-to-spelling mappings of early Scots. This new resource provides the necessary data to trace sound change in Older Scots. Drawing on data from FITS, Benjamin Molineaux, Joanna Kopaczyk, Warren Maguire, Rhona Alcorn, Vasilis Karaiskos & Bettelou Los challenge the traditional view that L-vocalisation in Scots had gone to completion by the early sixteenth century by showing that there is actually no evidence of an increase in L-vocalisation during the fifteenth century and the change is far from completing.

Similarly to Molineaux et al.  the four chapters in Part 2 (‘Segmental histories’) collectively illustrate how the broad range of spelling forms gleaned from electronic dialect atlases further our understanding of sound-to-spelling mappings and our ability to provide more accurate answers to old problems, like the development of Old English (OE) (in the chapter by Gjertrud F. Stenbrenden) and OE eo/ēo (in the chapter by Merja Stenroos). Margaret Laing & Roger Lass’s chapter on interpreting the phonetic realisations of the (no fewer than fifty-seven) ME spellings for OE hw- and Donka Minkova’s chapter on the development of and , and the complex relationship between palatalisation and the associated processes of affrication and assibilation, demonstrate how digital resources, such as LAEME, (E)LALME and LAOS, can be harnessed to provide the data to create CoNe type narrative etymologies, in this case for OE hw-, or to flesh out the details of these etymologies by contributing to our understanding of sound change processes, such as velar palatalisation.

Part 3 (‘Placing features in context’) narrows the broader geographical scope of the studies in Part 2 to consider features in particular regions. Using LAOS, Daisy Smith looks at the distribution of the abbreviation used to denote the Older Scots plural noun {S} morpheme. The abbreviation is generally assumed to be functionally identical to <is/ys>, a distinctly Scots’ feature, rather than <(e)s>. Smith finds that the abbreviation is in fact motivated by palaeographical convenience and occurs following stem-final letters terminating in a horizontal stroke. Both Smith’s chapter and that which follows, by Ad Putter, further illustrate how in-depth corpus-based engagement with historical material makes it possible to test earlier assumptions. The approach allows Putter to show that The Court of Love in Cambridge, Trinity College, MS R.3.19 is probably a mid-fifteenth century text rather than the pseudo-medieval text Skeat claimed. The ‘false grammar’ in the poem noted by Skeat actually reflects linguistic differences between the London scribe and the East Anglian poet. The final paper by Guzmán-González illustrates the use of digital repositories of Middle English to explore the time depth of present-day dialect features, in this case, the use of masculine pronoun anaphors to refer to inanimate countable referents in modern Southwest dialect of British English. Trinidad Guzmán-González finds no substantial evidence of a ME origin for ‘he’ as the default pronoun for inanimate countable referents (but compare Stenroos 2008, who concludes that the development of gender assignment from Old to postmedieval English evolves along a scale of individuation).

The studies in this volume elegantly illustrate how the continued expansion and improvement of diachronic corpus databases and methodologies further our understanding of language variation and change by allowing historical linguists to chart linguistic change with greater precision. Yet the volume does not shy away from acknowledging the ongoing limitations of some of the digital resources. Guzmán-González’s study on anaphoric reference in the Southwest dialects makes clear that manual data collection is still at times necessary as search engines cannot handle the very long strings involved in the study of anaphoric reference. LAEME and LAOS both employ a model of semanto-grammatical tagging that is suited to the original purposes of the resources but is out of sync with other historical linguistic corpora that syntactically parse and tag for parts of speech. The syntactically-parsed PLAEME remedies this limitation in the case of LAEME for the period 1250–1350, and future plans include the expansion of PLAEME to cover the period 1150–1250 and a syntactically-parsed version of LAOS.

From the days of manually scanning microfiche or manuscript facsimiles housed in far-flung libraries and archives, the methods and resources available to historical linguistics including readily available digital diplomatic transcriptions, and tagged, parsed corpora have come a long way. Historical Dialectology in the Digital Age illustrates how historical dialectology has profited from the field’s speedy willingness to embrace and develop new digital technologies, and much of that initiative has come from the Angus McIntosh Centre for Historical Linguistics. Despite its focus on English and Scots, researchers working on the historical dialectology of other languages will benefit greatly from reading this volume.