Special issue on Re-evaluating the Celtic hypothesis

Introduction¹

Present-day historians of English are widely agreed that, throughout its recorded history, the English language has absorbed linguistic influences from other languages, most notably Latin, Scandinavian, and French. What may give rise to differing views is the nature and extent of these influences, not the existence of them. Against the backdrop of this unanimity, it seems remarkable that there is one group of languages for which no such consensus exists, despite a close coexistence between English and these languages in the British Isles spanning more than one and a half millennia. This group is, of course, the Insular Celtic languages, comprising the Brittonic subgroup of Welsh and Cornish and the Goidelic one comprising Irish, Manx, and Scottish Gaelic. The standard wisdom, repeated in textbooks on the history of English such as Baugh and Cable (1993), Pyles & Algeo (1993), and Strang (1970), holds that contact influences from Celtic have always been minimal and are mainly limited to Celtic-origin place names and river names and a mere handful of other words. Thus, Baugh & Cable (1993: 85) state that ‘outside of place-names the influence of Celtic upon the English language is almost negligible’; in a similar vein, Strang (1970) writes that ‘the extensive influence of Celtic can only be traced in place-names’ (1970: 391).

The usual explanation for the impermeability of English against Celtic influences rests not so much on any linguistic properties of English or Celtic but on the sociopolitical and cultural factors surrounding the relationships between the English and the Celtic populations, starting from the arrival and settlement of the Anglo-Saxons in Britain from the mid-fifth century onwards and extending up to the present day. Throughout the history of their encounters, the Celts have been considered the underdogs from a political, military, and also cultural point of view, and it is this hegemony of the English which is commonly believed to have blocked any significant linguistic influences from the Celtic languages upon English. The small number of Celtic loanwords in English is usually cited as definitive proof of this; the conquerors have never, as the argument goes, had any practical need to borrow words from the language of the conquered. This view can be traced back to the early twentieth-century Danish Anglicist Otto Jespersen, whose authoritative statement on this subject is echoed in some form or another in almost all subsequent treatments of Celtic–English contacts:

We now see why so few Celtic words were taken over into English. There was nothing to induce the ruling classes to learn the language of the inferior natives; it could never be fashionable for them to show an acquaintance with that despised tongue by using now and then a Celtic word. On the other hand the Celt would have to learn the language of his masters, and learn it well; he could not think of addressing his superiors in his own unintelligible gibberish, and if the first generation did not learn good English, the second or third would, while the influence they themselves exercised on English would be infinitesimal.

(Jespersen 1905: 39)

Although Jespersen's account has dominated the field of English historical linguistics for so long, there have from early on been those who have argued for a different approach, which assigns a more prominent place to contact influences between the Celtic languages and English. Indeed, such views have been expressed from time to time for well over a century now, and what is more, with increasing intensity over the last couple of decades (see, e.g., Filppula, Klemola & Pitkänen 2002 for discussion). The recent interest in what we have here chosen to call the ‘Celtic hypothesis’ (CH) goes to show that the writing of the early linguistic history of Britain may not be completed yet, and that the ‘old’ facts about the outcomes of the Celtic−English contacts may not have stood the test of time so well as has sometimes been assumed. The present volume can be seen as yet another in a long line of previous efforts to reassess the question of Celtic influences upon English in the light of the latest linguistic and other kinds of evidence.

One can discern at least four factors which explain the upsurge of interest in the CH in recent years. First of all, fresh archaeological and historical evidence is now available about the relationships of the Celts and the Anglo-Saxons in the first few centuries following the arrival of the Anglo-Saxons. This evidence sheds new light on the relative sociopolitical and cultural positions of these populations, on population movements across time and in different parts of the British Isles, and on the crucial question of the fate of the British Celtic population in the aftermath of the Anglo-Saxon settlements. Although different views on these matters still exist, it is no exaggeration to say that current thinking holds that, rather than being exterminated or banished from their homeland by the Anglo-Saxons, the Celtic-speaking population continued to live side by side with their new rulers in many areas and, after a period of extensive bilingualism, were gradually absorbed into them both linguistically and culturally (see, e.g., Hines 1990; Higham 1992, 1994; Härke 2003, 2007). This is also supported by the latest population genetic studies, which point to a significant degree of continuity of the indigenous Celtic-speaking population even in the southern parts of England where the Germanic intrusion was at its strongest (see, e.g., Capelli et al. 2003). Taken together, all this evidence has important repercussions on the question of the linguistic outcomes of the Celtic–English contacts.

Secondly, the standard wisdom about the lack of evidence for Celtic contact influences rests on grounds that cannot be sustained in the light of our present-day knowledge about language contacts and their typical outcomes globally. The nature of contact influences has been found to vary primarily depending on the type of sociohistorical conditions in a given contact situation (see, e.g., Thomason & Kaufman 1988; Winford 2003). Thus, in conditions of relatively intense language shift involving ‘imperfect learning’ (i.e. limited access to the target language because of the lack of formal instruction) and a large population – such as those which characterised the Celtic−English interface in the early mediaeval period and also in the later centuries in many parts of the British Isles – contact influences can be expected to be found in the domains of phonology and syntax rather than lexicon. Attempts to brush aside the Celtic substratum on the basis of lexical evidence only are therefore seriously misguided and have held back research into possible contact effects in the other domains of English grammar. This situation has now changed as a result of an increasing number of studies in the past couple of decades which have sought to unravel the possible Celtic roots of many phonological and syntactic traits of English (see, e.g., Poussa 1990; Hickey 1995; Tristram 1999; Vennemann 2001; Poppe 2003; Filppula, Klemola & Paulasto 2008). The contributions to this volume continue to redress the balance in this area by drawing on the recent advances in the general theory of language contacts, typology and areal linguistics.

Thirdly, there is also new evidence about the history and later stages of both English and the Celtic languages, as well as other relevant languages, which can be brought to bear on this issue and which was not available when the early twentieth-century philologists like Jespersen formulated their views. Fourthly, it is often forgotten that the prevailing view on the paucity of Celtic influences in English has never been endorsed by all of the scholars working on historical and linguistic contacts between English and Celtic. From very early on, there have been ‘dissident’ voices, which have not, however, received the attention they would have deserved but which merit to be re-heard now. Indeed, there is reason to believe that the traditional views on the nature and outcomes of the English–Celtic contacts have at least partially been, or were in the past, inspired by other than purely linguistic agendas (for discussion see, e.g., van der Auwera & Genee 2002; Filppula, Klemola & Pitkänen 2002). We are here referring to an ideological stand known as ‘Anglo-Saxonism’ (see, e.g., Frantzen & Niles 1997), which in the nineteenth century, in particular, informed the views of many influential scholars writing on these issues. It is true that extreme views have also been expressed on the part of those who have defended the CH. ‘Substrato-maniacs’ or ‘Celto-maniacs’ are the terms which have sometimes been used for representatives of this position by those who want to deny any Celtic influences in English. The existence of these kinds of views is yet another factor which underlines the need for a new, open discussion on the exact nature of the English–Celtic contacts and their linguistic outcomes.

The contributions to the present volume build to varying degrees and from different perspectives on the aforementioned kinds of new evidence to produce either novel interpretations or reassessments of the earlier ‘truths’ about the nature and extent of Celtic influence on English, or discover entirely new areas of English phonology or syntax which have not been treated in previous works. In the opening article, John McWhorter discusses the emergence of periphrastic do in English. According to him, the fact that in English do has developed into a semantically empty auxiliary that can be used with almost all verbs underlines its cross-linguistically special nature not only among the Germanic group, but worldwide. As McWhorter points out, the Celtic languages are among the very few languages of the world that exhibit this feature. He considers a number of potential objections to the Celtic hypothesis, but concludes that transfer from Cornish must be considered the most likely source of the construction in English.

The next two articles focus on the question of Celtic influence in Old English phonetics and phonology. Peter Schrijver argues that the changes that the West Germanic sound system underwent during its development into Old English can best be explained by assuming a common substratum for all the Old English dialects. Schrijver identifies this substratum as a form of Primitive Irish Celtic spoken in the Lowland Zone of Britain. He concedes that the evidence for Lowland British is very limited, but argues that the limited evidence supports the conclusion that Lowland British Celtic is closer to the phonetics of Old Irish than it is to the phonetics of Highland British Celtic, which is the variety ancestral to the attested Welsh, Cornish, and Breton languages. In the other article focusing on phonological contacts, Stephen Laker addresses the question of phonemicisation of voiced fricatives in Old English. He reviews a number of earlier accounts, and then presents his proposal, according to which the phonemicisation of a previous allophonic voice alternation in fricatives can best be explained through language contact with Brittonic, where a phonemic contrast between voiceless and voiced fricatives existed at the time of contact.

In her contribution, Angelika Lutz focuses on two features of Old English, one syntactic and the other lexical. She argues that the twofold paradigm of ‘to be’ in OE, first discussed by Keller (1925), marks a significant formal and functional difference between Old English and the other Germanic languages, and is most probably – as Keller had suggested – due to substratal influence from British Celtic, which had a similar distinction. She also takes issue with a recent proposal by Schumacher (2007), who explains this distinction as a feature preserved in OE from much earlier, continental, contacts of Celtic with West Germanic. Lutz, however, finds the evidence for Insular Celtic contacts more persuasive than that for continental contacts and therefore sides with Keller's account. In the second part of her article, Lutz uses lexical evidence, especially the OE words meaning ‘slave’, to support her argument that members of the Saxon upper classes were in all likelihood in close contact with the British Celts and hence with Celticised forms of Saxon.

Erich Poppe, in turn, uses the notion of Standard Average European (SAE) as his starting point, against which he then examines the degree of ‘Celticity’ of Standard English. He focuses on two ‘test cases’, which are the rise of identical reflexives and intensifiers and the frequency of ‘labile’ verbs, such as to break, which can be used in both the inchoative and the causative sense. Reflexives and intensifiers are identical in English as well as in Insular Celtic, but separate categories in SAE. Similarly, labile verbs are a salient feature of both English and Insular Celtic, which would seem to suggest that both of these features are due to Celtic influence on English. However, on closer examination they turn out to be typologically related, and do not therefore provide independent evidence for Celtic influence. Relying on the typological generalisation formulated by König & Siemund (2000), Poppe concludes that the typological similarities between Welsh, Irish, and English with regard to labile verbs are a secondary development in English, following on from the emergence of new complex reflexives in Standard English. This leaves the latter as a feature which, according to Poppe, may well have arisen in English as a result of Celtic substratal influence.

Also using SAE as his starting point, Markku Filppula investigates areal, typological, and contact-linguistic aspects of the cleft construction. As in Poppe's study, this is a feature which is not one of the defining characteristics of SAE, but is part of the grammar of English, and furthermore plays a particularly prominent position in the Celtic languages as well as in French and some other Romance languages. Filppula's special focus is on the rise of the so-called it-cleft construction in English, which he traces back to OE. From relatively sporadic uses at that stage, the rate of incidence of it-clefts gradually increases, and already in ME they become syntactically and functionally more versatile and more grammaticalised than in OE. Filppula argues for a contact-based explanation because of the chronological precedence of the Celtic cleft constructions, the prominence of the cleft construction in the modern-period ‘Celtic Englishes’, and close parallels between English and the Celtic languages with respect to several other syntactic features.

Juhani Klemola's article deals with a syntactic feature which represents a hitherto little-investigated area of Celtic influence in English dialects: unusual adverb + infinitive constructions in southwestern and west Midlands dialects of English. The most frequently reported form of this construction is the relatively formulaic phrase away to go ‘away he went’, where the to-infinitive must be interpreted as a kind of a historic infinitive. Klemola reviews the history of the construction type in English, and argues that it appears to be marginal at best in earlier varieties of English. Comparable historic infinitive constructions are, however, a well-established feature of the British Celtic languages (i.e. Welsh, Cornish, and Breton). This leads Klemola to argue that transfer from the British Celtic languages offers a likely source for the away to go construction in southwestern and west Midlands dialects of English.

In the final article of this volume, Theo Vennemann also brings into the discussion a hitherto little-known area of early Celtic influence in English, viz. answers to Yes/No-questions. He notes that, compared to speakers of German, speakers of English reply to this type of question more often with a short affirming or negating sentence, instead of a simple yes or no or in addition to either of these. On the basis of selected data from Modern to Early Modern and Middle English, he shows that this tendency, which he characterises as ‘un-Germanic’, was established in the two centuries between Chaucer and Shakespeare and is thus a Late Middle English and Early Modern English innovation. As to its source, he points out the close parallels in Irish and Welsh, in both of which this feature was established at an early stage. Vennemann's conclusion is that, just as in present-day Irish English where the same avoidance of a simple yes or no has been documented, speakers of Brittonic shifting to Anglo-Saxon carried this response type into their new language. First a feature of nonstandard language, it gradually worked its way up and eventually reached the level of educated discourse towards the end of the Middle English period.

Though necessarily selective in their approach and coverage of linguistic features, we believe that, taken together, the articles in this special issue make a significant new contribution to our knowledge of the history of English and especially of the contacts between English and its closest neighbours, the Celtic languages. We also hope that these studies serve the purpose of stimulating further research into this area, which has been on the margins of mainstream scholarship for so long.

Markku Filppula

Juhani Klemola

April 2009