Research in Focus: Ad Putter

Over a series of blogs we’ve been learning about the research of our members and others whose research is pertinent to the networks ambitions. This month we hear from Professor Ad Putter of the University of Bristol.

Putting the ‘multi’ back in multilingualism: Anglo-Dutch

Research on Middle English has become much more conscious of the presence of other languages, but the Big Three (Latin, French and English) have dominated scholarship to such an extent that ‘multilingualism’ in the context of medieval England has become almost coterminous with ‘trilingualism’. We must do more to acknowledge the multiplicity of languages that were current. My first stab at doing so was a volume I co-edited, Multilingualism in Medieval Britain (2013). This puts English in the context of French, Latin, Old Norse, Dutch, Hebrew, Welsh, Cornish and Gaelic. For the front cover we picked a page from a manuscript (a Psalter, c. 1220, now Oxford, Bodl. MS Or. 621, fol. 2r), written in England in Hebrew and Latin.

Image 1.jpg

My current research is on Anglo-Dutch relations in the medieval and early modern period. Since the Low Countries were a crucial trading partner, Dutch was important, and various writers came into contact with Dutch. A fascinating example is William Caxton. Caxton’s bestseller, Reynard the Fox, was a translation from Dutch. Caxton’s fluency in Dutch seems remarkable today, but was not uncommon for someone in his situation. Caxton was a cloth-trader and had spent decades working in the Low Countries.

Literary critics have not been complimentary about the quality of his translation. If we examine a sentence from Reynard, we can understand this point of view:

Yf the fox will telle how it byfel, I wyl gyve hym the fordele thereof, for I can not telle it so wel but he shal beryspe me.

(If the fox will tell me what happened, I will give him the privilege of doing so, for I cannot tell it so well that he would not reproach me.)

The italicized words are Dutch borrowings, like many others in Caxton’s Reynard. Editions from the nineteenth century onwards, including the fine Kelmscott edition by William Morris (1892), carry glossaries listing them.

Image 2

Caxton’s ‘double Dutch’ has encouraged the view that he was translating slavishly from Dutch, but on closer inspection the situation is more complicated. For example, while ‘beryspe’ was taken over verbatim by Caxton from his Dutch source, ‘fordele’ does not occur in the corresponding Dutch, which instead reads ‘vorwaerde’. Where ‘vordele’ does occur in the source, it only ever has the usual sense of the word, ‘advantage’, in which sense the word was borrowed into English in the fifteenth century. The context in which Caxton here uses ‘fordele’ shows, however, that he was sufficiently fluent in Dutch to know that the word had another sense in Middle Dutch, namely ‘privilege, precedence’.

Instead of thinking of Caxton as a slavish translator, I see him as a bilingual speaker, whose English shows influence from the language he spoke and wrote during his years in the Low Countries. In linguistics, such influence is known as interference, and this linguistic perspective provides a more sympathetic way of approaching Caxton’s English. It explains why he was also influenced by Dutch when he was translating from French. Examples of Dutch words that infiltrated his French-based books are ‘butter’ (‘cheat’, from Middle Dutch ‘botter’) in Game and Play of The Chesse (1474) and ‘spynroke’ (‘distaff’) in Book of the Knight of the Tower (1484). It also explains why Dutch not only shaped Caxton’s vocabulary but also other aspects of his language. For instance, in History of Jason (1477), Caxton on a few occasions spells ‘fleece’ in Dutch fashion as ‘vliese’. In Receuyell of the Historyes of Troye, he writes ‘styfemoder’ (Dutch ‘stiefmoeder’) instead of his normal ‘stepmoder’. In short, the real issue with Caxton’s style is not that he translated Dutch literally, but that he could not help himself thinking in Dutch, even when he was translating texts from French into English.

A wide-ranging research project on Anglo-Dutch relations, funded by Leverhulme,, is ongoing, and what I offer here is only a taster of a substantial book, North Sea Crossings: The Literary Heritage of Anglo-Dutch Relations, 1066-1688, co-authored with Sjoerd Levelt. The book will be published by the Bodleian Library to accompany an exhibition in the Weston Library in Oxford, opening in December 2020.


Research in Focus: David Parsons

Over a series of blogs we’ve been learning about the research of our members and others whose research is pertinent to the network’s ambitions. This month we hear from David Parsons of the Centre for Advanced Welsh & Celtic Studies, University of Wales

Local place-names as linguistic data: current work on Welsh and English in Shropshire

Archives, local and national, contain tens of thousands of documents which can cast light on the languages of the people of Britain in the Middle Ages. Vernacular literary texts make up a tiny proportion of these; the vast majority take the form of local administrative records, generally written in Latin, which preserve the names of people going about their daily business, and the names of the places where they lived and worked. Both offer rich sources for the historical linguist.

For nearly a century the Survey of English Place-Names has been gathering and publishing the evidence of local place-names, initially with an accent on the ‘major’ names of towns and villages, but in the last fifty years with much more emphasis on the huge body of ‘minor’ names. This has encouraged several (though nowhere near enough) studies of intensely local medieval dialect, especially in areas of historic Scandinavian settlement, where interest lies particularly in the nature and degree of Norse influence on Middle English.

Nineteenth-century field-names in Sweeney, south of Oswestry ©Shropshire Archives

In current work on the Survey of Shropshire we are now tackling an area where linguistic contacts are of a different kind. In the north-west of the county, around Oswestry, Welsh and English were spoken side by side throughout the medieval and into the modern period. Collecting, presenting and analysing the place-names of the area offers challenges, not least because much of the material has passed through the hands of scribes – and sometimes modern editors! – with no knowledge at all of Welsh. (Though this can of course be a blessing, because there tends to be no question of ‘correct’ orthography obscuring phonological developments – those interested in the history of consonantal mutations in Welsh should take note.) When a baffling eron erhorcet goh’ epanhit proves, with the help of overlapping records, to resolve into a wholly intelligible erw’r orsedd goch y pant ‘acre of the red mound in the dip’ we are in interesting territory.

The fourteenth-century will of Dyddgu, daughter of Ieuan ap Madoc of Bronygarth, near St Martin’s to the north of Oswestry ©Shropshire Archives

That example comes from Aston, two miles south-east of Oswestry, an English-named Domesday vill with a rich mix of linguistic forms in its nomenclature, typical of the wider area but especially well documented in the surviving records of Haughmond Abbey and the Aston Hall estate. These show extensive use of both English and Welsh and clear influence of each on the other. In Henrekessurdir, for instance, the personal name (ultimately Continental Germanic) Henric is bound by Middle English genitive ‑es to the recurrent Welsh compound surdir ‘sour land’, which itself involves sur from English. In Y bwttie byrrion ‘the short butts’, we have a Welsh version of Middle English butte ‘short strip of arable land’, given a Welsh plural ending and matching concord with the inflected adjective. The latter type is a good deal more common than the former: it is evident that then as now Welsh was much more porous to English than vice versa. But one of the attractions of the material is that it is copious and can be analysed quantitatively. When it is remembered that most of it is also closely dated – albeit that the dates are termini ante quem for the coinages – then the potential for detailed analysis of aspects of a medieval linguistic community that has left us no vernacular documents should be clear. The Survey volume, on The Hundred of Oswestry, is due to be published in 2020.

Research in Focus: Paul Dryburgh

Over a series of blogs we’ve been learning about the research of our members and others whose research is pertinent to the network’s ambitions. This month we hear from Paul Dryburgh of The National Archives.


The National Archives (UK) has one of the largest collections of medieval records anywhere in the world, which it holds as the successor archive for English, and later British, royal government. (For our purposes ‘medieval’ constitutes everything between Domesday Book, the first ‘public record’ (created by or for the crown), and the creation of the State Paper Office in 1509.) Broken down, say, to the level of the individual writ or warrant – those tiny strips of parchment on which royal orders and instructions were written, the engine room of government – the medieval archive, constituted principally of the records of Chancery (the secretariat), Exchequer (financial office), and central and local law courts, dwarfs all other non-digital collections. The collection’s richness and diversity of content is such that it can inform almost any aspect of research into the Middle Ages to some degree. One of the aspects that exercises me and my colleagues within the Medieval Records team is what we often call ‘the mechanics of government’. This takes us beyond the informational content of the record and examines the processes, linguistic and material, behind seeking, expressing and executing the royal will. Language played a critical role in administrative process; government in medieval England was therefore at the core of the nation’s multilingual culture as more sections of society came into contact with the centre. However, the role of language, its continuity and mutability, is not always prominent in studies of medieval administration.

E 212/79: Settlement, indented, by Thomas Chaucer, esquire, of the reversions of the manors of Hook Norton and Kidlington (Cudlyngton) and lands in Ewelme, Swincombe, Nuffield (Tuffeld), Benson, Newnham Murren, Mongewell, Warborough (Oxfordshire), on Thomas Chaucer and Maud, his wife, and of the manor of Dorton (Buckinghamshire), to the use of the said Thomas. 20 May 1409. English.

We are all aware that throughout the medieval period – and, indeed, as far as 1733 – the principal language of record in government and the church was Latin. As written record culture matured and the communities by whom government was staffed and with whom it interacted diversified, Anglo-Norman French, the vernacular of the nobility, gentry and urban elites, proliferated from c. 1220. The language of correspondence and of legal pleading (if not initially recording), French expanded over a century or so not only to become the language of supplication and of parliamentary debate but also embedded itself into the internal written culture of government. A vast number of warrants in Chancery, mandating the composition of letters, for example, under the great or privy seals (the ultimate royal authority for grants, awards and commissions to office or to try legal cases) were couched in the vernacular. After the 1362 Statute of Pleadings, which permitted litigants to plead in English in courts of law, English gradually seeped into bureaucratic language. Indeed, the use of English to record the proceedings of the court of Chancery, an equitable jurisdiction based on evidence and fairness, brought wider comprehensibility to the legal corpus. By the reign of Richard II (1377-1399), and certainly for much of the early fifteenth century, the administration of English royal government was operating in a complex multilingual world, where linguistic choice mixed with formal (not to say formulaic and legalistic) modes of expression.

PSO 1/2, no. 75: Itemised list of artillery and gunpowder (anglice gonne pouder) prepared for Henry IV. [April 1405]. Latin.
Language use and language choice in context can add an extra layer to a record’s history and interpretation. But for me as a historian and archivist rather than, say, a historical linguist or lexicographer, and being the primary member of the Medieval Records team at TNA to deal with French in our collections, gaining a better understanding of linguistic and orthographic choices, the politicisation of language, the mixing of languages in single documents, the adaptation of vocabulary and the use of loan words from other languages within formulaic text and the adaption of formula into different languages will sharpen my knowledge of scribal and legal culture and practice within government.

E 28/4, no. 14: Warrant under the privy seal to [Thomas Arundel], archbishop of York, chancellor of England, to draw up letters patent confirming the license granted by Richard II to the prior of Caldwell to appropriate the church of Sandy (Sondey) [Bedfordshire], 7 July 1393; with an accompanying memorandum concerning payments to poor parishioners and the incumbent. French.
My interest in the Medieval Multilingualism network, then, is twofold: firstly, to bring innovations from other disciplines into my own research, and to share my knowledge of the archive and of record making and record keeping; and, secondly, to broaden the disciplinary base of the research community using government records. It is becoming evident within the wider archive sector that the skills required to interpret and describe pre-modern records are, for reasons of resource and the multiplicity of demands made of professionals, somewhat on the wane. Similarly, the teaching of the traditional skills of linguistic and palaeographic analysis within the academic sector does not always fit graduates with the breadth (if not depth) of knowledge and experience to cope with the range of material they might encounter within the archive. At The National Archives, we have developed a programme entitled Postgraduate Archival Skills Training (PAST) which aims, over three stages, to: (1) introduce academic researchers of any discipline to archival collections, archival research and navigating online and legacy catalogues; (2) give instruction in some of the skills and methodologies vital to researchers; (3) go more in-depth on a particular theme (Legal Records, Material Culture, Religion and the Church, etc.) Check out the webpages and do sign up! While there is some focus on language and linguistic culture, there is, I’m sure, more we can do.

E 28/4, no. 13: memorandum of confession by four Cornishmen for exporting tin to Flanders without passing through the staple, 16 RIchard II [1392-3]. French.
The opportunity to collaborate with scholars across a number of fields to open up our collections in exciting new ways is a key aspect of the Multilingualism network. Do please get in touch and let’s explore how we can work together.


Research in Focus: Peter Stokes

Over a series of blogs we’ve been learning about the research of our members and other whose research is pertinent to the network’s ambitions. This month we hear from Peter Stokes of King’s College London.

Multilinguism – Multigraphism

My focus is a little different from that of most of the other members of the network. Rather than focusing on multilingualism, namely the use of different languages in a given culture, my research is on multigraphism, that is, one culture using different styles of writing or even entirely different alphabets. This is common in today’s multicultural society: if you look at street signs or newspapers in different areas of large cities, you might find writing in the Latin alphabet (used for English and most other Western European languages) but also in Chinese, or Japanese; in some countries you can find Cyrillic and Latin alongside each other; or you might see Hebrew and Arabic; Greek and Latin; and so on. However, this is not a new phenomenon. Multiculturalism is as old as human culture itself, and examples of multigraphism can be found throughout history as well:

  • We find Greek, Latin, Arabic and Hebrew being used together in places like medieval Sicily, such as the famous quadrilingual tombstone from Palermo discussed and illustrated here.
  • We find manuscripts in Greek, Latin and Hebrew written possibly in France as shown here; this particular manuscript is also discussed in an article by Judith Olszowy-Schlanger and Patricia Stirneman.
  • We find many different mixes of writing systems in the materials from Dunhuang, where many different people met along the Silk Road. Examples are available from the International Dunhuang Project, where you can search for manuscripts by language.
  • There are manuscripts in Hebrew and Chinese from the Kaifeng Jewish community in Henan province, China, for instance Ohio, Hebrew Union College MS 926 or the Chinese Torah discussed in a blog post from the British Library.

I could go on. In fact, just like multilingualism, once we start looking for multigraphism, we find it everywhere.

It raises lots of interesting questions, too. As a palaeographer, my research is in the history and development of handwriting. I am interested in questions like these:

  • Why are letters shaped the way that they are? When did different alphabets and different styles of writing develop? How can we use that to help understand where and when a given document was written?
  • Why do different cultures use different alphabets or writing systems? What do these choices tell us about these people, their beliefs and their interactions with other cultures?
  • Is it possible to tell from somebody’s handwriting whether that person learned to write first in Chinese or in Mongolian? In Latin or in Devanagari? (There’s a very interesting video addressing this last case for typefaces here.)
  • Is it possible to detect ‘accents’ in people’s handwriting, in the way that we can in their spoken or even written language?

Why are these interesting questions? Because they allow us to learn new things about history, and, most importanty, because they tell us about ourselves. Writing is everywhere, whether in books or newspapers, on street signs or on food packaging, on our phones or on our computers. It’s fundamental to much of our communication, and most of what we know about history, but also the way that we write influences what and how we communicate, and it also says things about us. The style of writing I use is partly based on where and when I learned to write, but it also has a political element as something like German Blackletter shows.

To give an example, look at this sample of a manuscript:

Cambridge, Trinity College, O.2.41, 87r (detail). See the full manuscript online here.
This image is copyright the Master and Fellows of Trinity College, Cambridge and is reproduced here under a Creative Commons Attribution-NonCommercial 4.0 International Licence.

As a palaeographer, I can see immediately that the style of writing here is very distinctive and was used only around the end of the eleventh or start of the twelfth century. I can also see that it is particular to people who learned to write in England. In fact, the sample shows writing in two different languages, Latin and Old English. This is what it says, if I transcribe it into modern type:

hit gesealde ædgare cyninge un besacan eal
swa hit his yldron ær mid feo gebohten
and se fore specena cyningc hit gesealde go-
de and s’c’e Æðeldriðe his saule to alysetnesse.
Anno d’nicȩ incarnationis n’gentesimo septua-
gesimo sc’ipta e’ hȩc carta his testib’ consentienti-
bus quor’ inferius nomina caraxantur.
Ego Eadgar rex p’fatam donacione’ concessi.
Ego Dunstan dorou’nensis ȩccl’ȩ archiep’s c’signaui.

If you look carefully, you can probably make out the letters from Anno d’nicȩ onwards fairly easily. However, notice that in the first four lines, the scribe wrote some letters in a different way compared to the last five lines. For instance, compare the s in gesealde (line 1) with the one in incarnationis (line 5). Similarly, the r in ædgare (line 1) is very different to the way we write this letter today, whereas the one in incarnationis is very familiar to us. There are several other letters that change in this way: look in particular at the different shapes of d, g, f, and h (you will have to look closely for this last one!).

There is also another difference between the two parts. Look closely at the spaces between the words and between letters. Notice that the words in the first four lines have very large spaces compared to the second part. Notice also that these four lines are less regular and consistent in the spacing. In fact the scribe who wrote this also put some spaces in the wrong places (gesealde is written as ges ealde in line 1, for example). But again these irregularities and errors in spacing are all confined to the first four lines.

So, what does this tell us? One possibility is that the sample was written by two different people, with the first four lines by one person, perhaps not a very experienced scribe, and the rest by someone else. However, the details of the script are too consistent, and it seems clear to me that only one person was involved. How do we explain this division, then? Well, there is another aspect here: language. As mentioned above, the first four lines are in Old English, and the rest are in Latin. And I’ve told you already that this was written in England early in the twelfth century. In this period it was normal to use different forms of some letters depending on the language, using one set of letter-shapes for English and another for Latin. This scribe has followed this principle very closely. However, it seems clear to me that the scribe was familiar with writing Latin but not very comfortable writing English. I think the irregularity and problems of spacing in the English are because the scribe was writing slowly and was not sure about the text, stopping often to check, and not really able to remember even short phrases, whereas the Latin seems much more fluent. So we have a scribe, in England, early in the twelfth century, happy writing Latin but not happy writing English, maybe not even understanding English very well, but who knew the English style of writing. This seems very likely to be an Anglo-Norman monk, then, and we figured all this out just from this little bit of handwriting. That’s detective work for you!

Research in Focus: Elisa Ramírez Pérez

Over a series of blogs we’ve been learning about the research of our members and others whose research is pertinent to the network’s ambitions. This month we hear from Elisa Ramírez Pérez of the University of Cardiff.
My research focuses on processes of verbal simplification in the late Northumbrian dialect of Old English. Old English is the term used to refer to the language spoken in Britain from approximately 700 to 1150. It is general consensus to distinguish four dialects in Anglo-Saxon England: West-Saxon, Kentish, Mercian and Northumbrian. Of these four dialects, West-Saxon is the best attested one, since most of the extant Anglo-Saxon texts are written in this variety. It is understandable, therefore, that until quite recently, the study of Old English was largely limited to this variety.
Map from Crowley (1986: 100)

My research, however, aims to broaden the scope of the linguistic study of Old English, focusing instead on the Northumbrian dialect, which was spoken in the kingdom of Northumbria, off the north-eastern coast of Britain. More specifically, I focus on the later stage of this dialect, namely late Northumbrian. Three major texts record the late Northumbrian dialect: the Lindisfarne Gospels, most of the Rushworth Gospels and the Durham Ritual. Although these three texts were originally written in Latin, they were given interlinear, word by word Old English glosses during the 10th century. As a result of this later glossing, we have three medieval bilingual manuscripts, hence my relation to the MEMC network.

A quick look at any Old English grammar book will reveal that the Northumbrian dialect was a rather different variety, contributing to many exceptions to the general grammatical rules of Old English. However, what is most fascinating about the Northumbrian dialect is that it attests early occurrences of many features which are more typical of Middle English grammar. Some examples from the Lindisfarne Gospels gloss include:

  • the first instance of the blanket definite article ðe‘the’:

[1] MtGl (Li) 2.3:

Old English: geherde wiototlice herodes ðe cynig gedroefed węs

Translation: And (the) King Herod hearing this, was troubled.

  • early instances of -s ending for 3rd person singular present indicative:

[2] MtGl (Li) 15.23:

Old English: folet hia forðon cliopas æfter usig

Translation: Send her away, for she cries after us.

My focus is on verbal morphology, that is, I study the structure and form of verbs in the late Northumbrian dialect of Old English. Within verbal morphology, there are also quite a number of Northumbrian features which point towards the direction of Middle English grammar. I specialise particularly on a group of verbs which had a characteristic middle vowel <i> (known as the -i- formative) between the root and the inflectional ending in many of its forms:

[3] Old English: macian > Present-Day-English ‘to make’

Mac       i                      an                               >          make     

root   i form.   infinitival ending                             root

[4] Old English: ic swerie > Present-Day-English ‘I swear’

Ic         swer        i           e                        >                I         swear

pronoun    root    i form.  ending                           pronoun    root

You can clearly see from the examples above that both the -i- formative and the inflectional endings have been lost, since nowadays we only keep the root (with the exception of the 3rd person singular present indicative which adds an -s, as in ‘she makes’). My research is specifically concerned with the loss of the -i- formative, a phenomenon which, although generally classified as a Middle English development, is already widely found in the Northumbrian dialect. Examples [5] and [6] below show the exact same verse from Matthew’s Gospel, the first one taken from the section in the Rushworth Gospels which is not glossed in Northumbrian but rather in Mercian dialect. There you can see that the verb lufigaþ ‘love’ (imperative plural) still maintains the -i- formative. [6], from the Lindisfarne Gospels gloss, has lufas ‘love’ instead, where the -i- has been already dropped.

[5] MtGl (Ru) 5.44, Rushworth Gospels (Old Mercian dialect, 10th c.)

Old English:  lufigaþ eowre fiondas

Translation:   Love your enemies.

[6] MtGl (Li) 5.44, Lindisfarne Gospels (late Northumbrian dialect, 10th c.)

Old English: lufas ge fiondas iurre

Translation:   Love your enemies.

Although my preliminary investigation shows that Northumbrian verbs are more morphologically simplified and, therefore, more similar to Middle English grammar than other Old English dialects, further research is required to understand how advanced this process of verbal morphological simplification was in the Northumbrian dialect. Thus, I am in the process of analysing the language of the glosses to the Lindisfarne Gospels, the Rushworth Gospels (Northumbrian section only) and the Durham Ritual. Furthermore, in order to understand why it is that the -i- formative was lost, I will consider which factors could have triggered this loss. One of these factors involves linguistic contact with Old Norse, the language spoken by the Vikings, where the verbs I am studying did not have an -i- formative anymore. This element of my investigation also ties in with the MEMC network and its research interests.

Durham Ritual
Lindisfarne Gospels, f.27r, beginning of Matthew (foto que yo tomé del facsimile de Senate House)

Research in Focus: Antonio Benítez Burraco

Over a series of blogs we’ve been learning about the research of our members and others whose research is pertinent to the network’s ambitions. This month we hear from Antonio Benítez Burraco of the University of Seville.

Human self-domestication and the nature of prehistoric languages


The comparative method in linguistics has enabled to trace phylogenetic relationship among distant languages and reconstruct extinct languages from the past. Nonetheless, it has limitations and shortcomings, which result, in part, from some of its methodological assumptions (particularly, its heavy reliance on the lexicon), but mostly, from the real nature of language change, as languages do not only change by divergence from a common ancestor, but also as a result of (extensive) contact with non-related languages. At the same time, ongoing research suggests that language change depends not only of the internal dynamics of linguistic systems, but also of factors external to languages, particularly, aspects of human cognition and features of our physical and cultural environments.

Main hypothesis

Our fundamental hypothesis is that the limitations of historical linguistics can be partially alleviated by the consideration of the links between aspects of language structure and aspects of the biological underpinnings of human language, human cognition, and human behaviour. Specifically, we think that research on human self-domestication (that is, the existence in humans of features of domesticated mammals compared to wild extant primates), which seemingly entailed notable physical, cognitive, and behavioural changes in our species, can help illuminate facets of the languages spoken in remote Prehistory, the vast time period during which human beings have lived for longer. We expect that the languages spoken in that epoch exhibit most of the features of the so-called esoteric languages, which are used by present-day, close-knit, small human communities that share a great amount of knowledge about their environment.

Our (ongoing) project

Present-day populations exhibit a significant correlation between social and physical aspects of self-domestication. For instance, higher relative status of women compared to men has been found to correlate with reduced sexual dimorphism. We want to check whether all the features associated to human self- domestication parallel the esoteric-exoteric continuum of languages. Putting this roughly, because self-domestication results in enhanced sociability and because enhanced sociability results in exoteric languages, our prediction is that features of domestication will be attenuated in esoteric human groups and exacerbated in exoteric societies. For testing this possibility, we have planned to rely on the Standard Cross-Cultural Sample and the WALS. However, a good characterization of the languages found in these samples is still pending, particularly, regarding their morphology, that allows to classify them as esoteric or exoteric languages in an accurate way. You are very welcome to contribute to this project!

Pic for the blog entry

Research in Focus: Kari Kinn

Over a series of blogs we have been learning about the research of our members and others whose research is pertinent to the network’s ambitions. This month we hear from  network member Kari Kinn, of the University of Oslo.

My research is not about Medieval English; I currently work on Norwegian as spoken in North America. This is, however, more relevant to the study of Medieval English in a multilingual context than one might think; I will come back to that towards the end of this post.


In the 19th and 20th centuries 800,000 people emigrated from Norway to the USA and Canada; some of the descendants of these people still speak Norwegian as a heritage language. The term heritage language refers to a language which is acquired by children as a first language in the home, but which is not the majority language of the larger society. Heritage speakers are bilinguals and also speak the majority language; in our case, the majority language is English.

Settlers’ houses at Norskedalen Museum, Wisconsin. (Photo: Janne Bondi Johannessen)

Heritage languages exist more or less in isolation from the speech community of the homeland (in our case, Norway), and under intense contact with a majority language. What happens under these circumstances? Differences between heritage languages and their homeland counterparts range from borrowing of words from the majority language to structural differences in the way that words are inflected (morphology) and how phrases and sentences are constructed (syntax). Morphology and syntax tend to be more stable than the vocabulary, but changes happen even here; in what follows, I will focus on syntactic change.

The perhaps least surprising type of change involves cross-linguistic influence, or transfer, from the majority language into the heritage language. In American Norwegian, there are not many examples of syntactic transfer from English, but some speakers use an English-like pattern in constructions describing a person’s profession or nationality. The English pattern involves the indefinite article a(n): He is a teacher. In homeland Norwegian, the indefinite article is not used; the equivalent construction would be Han er lærer, literally ‘He is teacher’. However, in a forthcoming paper, I show that occasionally, some American Norwegian speakers use the (Norwegian) indefinite article in a way that resembles English: Han er en lærer.

Cross-linguistic transfer is a well-known phenomenon. What is perhaps less widely known is that language contact can sometimes also have an almost opposite effect. This effect is referred to as cross-linguistic overcorrection; in American Norwegian, this has been observed for constructions expressing possession (See full article by Anderssen, Lundquist and Westergaard). In possessive constructions such as my bike in English, homeland Norwegian has two possible word orders. One is similar to English, with the possessor before the noun: min sykkel, literally ‘my bike’. This word order is not the most widely used one, though; instead, speakers very often put the possessor after the noun, yielding patterns like sykkelen min, literallybike-the my’. American Norwegian speakers generally stick with the latter pattern which has no clear parallel in English; in fact, they use this option even more than Norwegian speakers in Norway. Thus, it looks like contact with English has yielded a “hyper-Norwegian” result.

American Norwegian speaker serving home-made cake. (Photo: Janne Bondi Johannessen)

Transfer and cross-linguistic overcorrection involve extension of patterns that already exist in either the homeland language or the majority language. However, changes in heritage languages can also be independent of this. For example, many speakers of American Norwegian under certain circumstances put the negation word ikke ‘not’ in a different position than homeland Norwegian speakers. This is neither cross-linguistic transfer nor overcorrection; instead, heritage speakers have created a system that differs both from English and (standard) homeland Norwegian. Their patterns are, however, similar to patterns found in the language of children in Norway, which suggests that the change has arisen during the acquisition process: heritage speakers have not fully generalised a syntactic rule that applies in homeland Norwegian  (see full article by Larsson and Johannessen).

Now, how does all of this relate to the Medieval English and the MEMC network? On a general level, the findings from American Norwegian and other heritage languages highlight the fact that language contact can lead to a number of different outcomes; cross-linguistic transfer is only one of them. This insight is not new, but it is an important one when starting new research on any language in a multilingual context. Furthermore, there are some more direct parallels between modern heritage languages and the languages spoken on the British Isles in the past. In collaboration with George Walkden, another member of the MEMC network, I plan to study one such case from a heritage language perspective, namely Norn. Norn was spoken in Shetland and Orkney until the mid/late 1700s. Like Norwegian, Norn descends from Old Norse (it is not quite clear when it developed into a distinct variety). And like American Norwegian, Norn co-existed with another language, in this case Scots, which gradually gained status as the majority language of the larger society. Recent research on heritage languages gives us a new lens through which we can look at the remaining sources of Norn – we are looking forward to seeing where this will take us.

Textual manifestations workshop (University of Bristol, 15th July 2019)

The third workshop of the network, focusing on the textual manifestations of multilingualism, took place at the University of Bristol on Monday 15th of July 2019. It brought together a group of international academics working on a wide range of areas, all of them extremely important to understand the effects of medieval multilingualism on the page. These areas included historical lexicology and lexicography; code-switching; English, (Anglo-)French, Scots and Welsh literature; palaeography; and cultural studies (historical pedagogical practices, archival work, etc.).

Our discussions centred around the following topics:

  1. When thinking about medieval multilingualism, we need to consider the difficulties involved in separating the various codes in use, as well as the effects that their contact had on the lexis of medieval English, in terms of the stratification of the vocabulary, processes of semantic shift, the development of different registers, etc.


  1. It is important not to assume that one can take a single approach to the study of code-switching in medieval times, for (as noted above) in some contexts it is very difficult to establish distinctions between the various codes involved, but in others this distinction is clearer. It is also fundamental to consider code-switching in relation to various levels of analysis, including letter forms, individual words, sentences, whole texts, etc. We need to understand the different ways in which various languages interacted on the same page.


  1. Because language and culture are closely linked, we need to gain a better understanding of the process of cultural adaptation that went hand in hand with the translation of medieval texts into other languages, and their geographical circulation and transmission.


  1. It is extremely difficult to agree on the best way to edit multilingual texts, as different audiences have very different (and, at times, incompatible) needs.

If you are interested in any of these topics and would like to know more about our discussions and our work, please get in touch:

University of Bristol from Cabot Tower (Adrian Pingstone, Public Domain)

Research in Focus: Mareike Keller

Over a series of blogs we have been learning about the research of our members and others whose research is pertinent to the network’s ambitions. This month we hear from  network member Mareike Keller, of the University of Mannheim in Germany


Mixing languages in medieval England

In a bilingual community you can often observe speakers using both their languages in the same conversation. They may switch back and forth between sentences, between words, and sometimes even within one word. Maybe, if you are fluent in more than one language, you do this yourself, or you have heard bilinguals around you talk to each other like this:

Why make Carol SENTARSE ATRAS PA’QUE everybody has to move PA’ SE SALGA?

Why make Carola sit in the back so everybody has to move for her to get out?

(adapted from Poplack 1980: 589)

In linguistics this kind of language mixing is most often referred to as code-switching. You might be tempted to think that code-switching is a recent development, maybe related to our increasingly multicultural societies. However, this is not so: we have documents from different historical periods and countries showing that code-switching was also practised many centuries ago.

Mareike Keller image for MEMC blg entry

Writing, of course, is different from speaking – but in historical linguistics, written evidence is the only evidence we have. And some of it is closely related to speech, for example sermons written down in Latin and Middle English, which were most probably also intended for preaching to a bilingual congregation. The question that interests me most concerns the differences and even more the similarities between modern and historical code-switching. After about forty years of careful observation of conversations and writing in modern bilingual communities, it is clear now that code-switching is not random, and that there are rules, just as there are for any language. There is no agreement yet on what exactly these rules are, but the most frequent mixing patterns that linguists have discovered hold as well for written texts from 500 years ago.

Just as in modern code-switching, we find that single words can be adapted to the grammar of the language they are inserted to (1), or they can appear as bare forms (2), without any grammatical marking. And we find that, like modern code-switchers, people in the middle ages did not only use words from another language because they had no good word available in the language they were using for the rest of the sentence (3).

(1)        emenda tuum CLOCK-um

             correct your clock

(2)        tuum CLOCK false vadit

             your clock moves incorrectly

(3)        rota huius HOROLOGII est tua miserablilis vita

             the cogwheel of this clock is your miserable life

             (MS Bodley 649, folio 3)

We know that languages themselves are constantly changing. And if you are not a trained medievalist, it might be difficult for you to understand a text that was written in the middle ages. The words and the grammar have changed so much that it needs some practice to read a Middle English text, even though it is, in fact, English. This makes it even more surprising that, from a linguistic point of view, the way languages are mixed together seems to have remained unchanged. In the light of this observation, the current goal of my research project is to describe the grammar of code-switching across time, establishing one more link between multilingual speakers in medieval England and multilingual speakers today.

Research in Focus: George Walkden

Over a series of blogs we have been learning about the research of our members and others whose research is pertinent to the network’s ambitions. This month we hear from  network member George Walkden of the Universität Konstanz

When speakers of two languages come into contact with one another, many things can happen. One well-known outcome is that languages “borrow” words from one another. Words aren’t the only part of a language that’s affected by language contact, however. Languages can also influence each other’s grammatical structure: pronunciation, morphology (how words are put together) and syntax (how sentences are put together). This post is about one syntactic change that might be due in part to language contact: how English expresses negation.

All languages have a way of turning a positive sentence – like I will see you soon – into a negative sentence (I will not see you soon). English uses the particle not for negation, but it wasn’t always this way. In Old English, sentences were negated by sticking the particle ne in front of the verb, with the word not nowhere to be seen. During the early Middle English period (1150-1325), ne was joined by not, and both words were used together to express negation. Soon after that, ne started to disappear, and not was left on its own. There are thus three stages to this change:

Stage 1: ne will see you soon (Old English to Middle English)

Stage 2: ne will not see you soon (Middle English)

Stage 3: I will not see you soon (Middle English onwards)

These changes don’t happen overnight. All the interesting action happens during the Middle English period, though, particularly Early Middle English. In this period we find texts that display all three variants – Stages 1, 2 and 3 – alongside one another, and that’s what I and my co-author Donald Morrison wanted to investigate. In our published article on this change, we try to answer two questions. First, where did Stage 3 emerge? In other words, where was ne lost first? Secondly, why did this change happen (where and when it did)?

The maps in Figures A-D show the course of the change over windows of 50 years. In the pies, Stage 1 is blue, Stage 2 is red, and Stage 3 is green. Each pie represents a single text, and the bigger the pie, the more negative sentences the text contains. What can be seen from these maps is that there’s very little Stage 3 before 1300. In Figure D, however, representing the period 1300-1350, we suddenly see quite a lot of Stage 3. Figure D is different in another way, though: it’s the first map in which we see texts from the northeast of England. Since there’s clearly quite a lot more Stage 3 (i.e. quite a lot less use of ne) in the north and east, we can safely assume this is where the change began.

Figure A: England 1150-1199. Here we mainly see only Stage 1 (blue) and Stage 2 (red), with little sign of Stage 3 (green).

What’s special about the north and east of England, though? This is where language contact comes in. It’s by now fairly well established, based mainly on archaeological and linguistic evidence, that a lot of Norse speakers from Scandinavia settled in this area of England between about 865 and 1042. They brought their language with them, so that Norse would have been spoken in communities across wide areas of the east and north of the country – especially in what is now Lincolnshire, Nottinghamshire and Yorkshire. It’s hard to know much about the language these people spoke, because they didn’t have a tradition of writing things down. However, based on the Norse written in Scandinavia we can be fairly confident that this language already used Stage 3. The only difference is that they didn’t use the word not – but they used a particle, ekki, which behaves syntactically just like not. (For instance, it follows auxiliaries like will, must and am, rather than preceding them like ne did.)

Figure B: England 1200-1249. There are more texts from this period, but still very little Stage 3.


Our theory is that the population of Norse speakers in the northeast helped to catalyse the change towards Stage 3. When people speak a second language, especially if it’s one they’ve learned as an adult, features of their first language are often subconsciously transferred into it. Norse speakers of English – before Norse in England died out entirely, sometime after 1100 – would have used not rather than ne because it fitted better with what they were used to doing in their native language.

Figure C: England 1250-1299. More texts now show examples of Stage 3, but the overall number is still very small


Of course, the theory can’t be proven: we just don’t know enough about what was going on in England at the time, especially in the northeast. Still, all the pieces of the puzzle (archaeological, linguistic, textual) fit together. And if we’re right, then we have yet another example of how English throughout its history has been shaped by the mouths and minds of multilingual speakers.

Figure D: England 1300-1349. Stage 3 (green) is substantially represented for the first time, especially in the North.