The Influences of the Celtic Languages on Present-Day English

Table of Contents

1. Introduction
1.1 Definition of terms

2. The Historic Background
2.1 Written Sources
2.2 Archaeology
2.3 Genetic analyses
2.4 Conclusion on Sources

3. Language Contact
3.1 Rapid Shift with Imperfect Learning
3.2 How to Define Contact Features

4. English and Celtic in Contact
4.1 Application of Framework on Historical Situation
4.2 How Long Did ‘British’ Languages Survive in Britain
4.3 Old English diglossia
4.4 Approach Towards Proposed Features

5. Syntax (in contact)
5.1 The Analyticisation of English
5.2 Clefting
5.3 The Northern Subject Rule
5.4 External vs. Internal Possessors
5.5 Periphrastic do
5.6 The Expanded Form
5.7 Varia
5.7.1 The definite article
5.7.2 Genitival groups
5.7.3 ‘To go’ as copula
5.7.4 Preposition stranding
5.7.5 Pronouns

6. Phonology
6.1 Interdental Fricatives
6.2 Retroflex /r/
6.3 Influence in American English

7. Celtic Words and Names in English
7.1 Loanwords
7.1.1 Dialects of English
7.2 Place-Names
7.2.1 Names for Britain
7.3 Personal Names

8. Conclusion
8.1 Outlook

9. Appendices
9.1 Abbreviations used
9.2 Maps

10. Bibliography

1. Introduction

The perceived lack of Celtic loanwords in English has generally been seen as proof that the Anglo-Saxon invaders made short notice of their Celtic predecessors when they took possession of Britain during the fifth century. Thus, the Celts simply would not have had the chance to leave their mark on the English language as they were either killed, driven into the sea or had to take refuge in the mountainous West and North of Britain. The possibility of any Celtic influence on the very structure of English has been discounted altogether.

In recent years, this view has met mounting opposition from different fields of study. New archaeological evidence as well as a methodological reassessment have called for a examination of the history of the Anglo-Saxon immigration. Besides, new advances in contact linguistics provide tools with which a more detailed look on the history of the English language has become possible.

These developments have lead to a new approach to the question of Celtic influence on the English language. The new argument runs that the dearth of Celtic loanwords in PDE can rather be seen as proof for the rapid shift of the indigenous peoples from Celtic to Anglo-Saxon speech, taking with them hardly any loanwords. Due to their ‘imperfect learning’ of the Anglo-Saxon language the Britons are assumed to have carried over a number of morphosyntactic and phonological features from Brittonic that, found their way into the general spoken language of the people due to the large number of British-influenced speakers of Anglo-Saxon. This linguistic interference is then assumed to have influenced a number of changes in the English language.

Expectably, this view has met (sometimes quite sharp) opposition from scholars who, discount the possibility of any Celtic influences on the English language for a variety of reasons. They attribute the changes of the English language to internal developments or, at most, medieval language contact with speakers of Old Norse in the Danelaw.

The question of whether such influence exists is not without implications, as language and nationality are closely tied together. It may also provide a new point of view on the current debate of national or European identity. So, as has been emphasised by Filppula et al., it is time to reconsider the question of linguistic outcomes of language contact between Celtic and English, particularly during the first centuries after the coming of the Saxons, but also taking into account the possibility of ongoing linguistic contacts with speakers of Celtic languages, e.g. Welsh (Filppula et al. 2002:7).

This paper aims to give an account of the current state of research on the question of language contacts between Celtic and English along with its possible outcomes. The conflicting opinions will be contrasted, taking into account the different disciplines that provide information.

In the first chapter, the historical background of the contact between speakers of English and the Celtic languages will be examined. The question of whether Britons and the British language did survive the coming of the Anglo-Saxons and what their sociocultural situation after the adventus saxonum might have been is of special interest.

For the consideration of the process and outcome of language contact, the approach of Thomason and Kaufman (1988) and Thomason (2001) will taken as the instrumental framework set out. This makes it possible to distinguish between situations of ongoing language contact and situations of rapid language shift. These situations have different linguistic consequences, and it will be illustrated that the Celtic languages are a possible cause of language interference along both ways i.e. by means of substratal influence through the rapid language shift towards Anglo-Saxon by speakers of Brittonic as well as continuing long-term contact with speakers of Celtic languages in the British Isles with the possible result of areal convergence.

In Chapter 3, this methodological framework will be applied to the specific situation of English-Celtic language contact, whereby the different conclusions that have been drawn from the historical evidence have to be considered. As Thomason and Kaufman point out, it is important to consider the whole language, not separate subsystems alone, when assessing the potential effects of language contact interference in a language (Thomason and Kaufman 1988:60).

Hence, the general development of the English language in its evolution from Old English until today will be depicted, pointing out where Celtic influence can possibly be suspected. Then a number of specific features found in PDE that might be attributed to Celtic language interference will be discussed. This includes features of English morphosyntax as well as of its phonology. In a last step, the lexis of PDE will be examined for words that have been borrowed from Celtic languages, including a (necessarily brief) look on place-names.

1.1 Definition of terms

The language spoken by the Britons at the time of the Roman occupation has come to be termed ‘British’. This language went through a number of substantial changes around the fifth and sixth century, after which it is called ‘Brittonic’ (Coates fc:1). These languages, along with their descendants, Welsh, Cumbric, Cornish and Breton belong to the Brythonic branch of Insular Celtic languages, as opposed to the Goidelic branch that encompasses Irish, Scottish and Manx Gaelic. All of these languages are part of the Celtic branch of the Indo-European languages and are thus distantly related to the other IE branches, e.g., Germanic or Italic. The term ‘Celtic’ and ‘Celts’ will be used in this paper to refer to Celtic languages and their speakers respectively. For a recent discussion of the validity of the term ‘Celtic’ see, e.g., Sims-Williams (1998).

The definition of a Standard English language is somewhat problematic, giving rise to such description as ‘Queen’s English’ or ‘BBC English’. One could ask whether this standard language is spoken at all or if it has any ‘native speakers’ at all. For the present purpose, Standard English will nevertheless be defined as a general, regionally and socially unmarked language (that arguably is based on the regional dialects of the South East). But, attention is also paid to more regional variants because, as Filppula et al point out, the “traditional regional dialects of English English […] provide a more realistic point of reference than standard English for assessing the impact of the Celtic languages on the development of the British Isles Englishes” (Filppula et al. fc.:3).

The term ‘Celtic Englishes’ is applied to the regional varieties of English spoken in (formerly) Celtic areas, i.e. Ireland, Scotland, Wales, the Isle of Man and Cornwall. In these areas, a Celtic language was formerly, (or is even today) spoken by a sizeable portion of the population. Celtic influence on these varieties can be shown by linguistic and extra-linguistic evidence to be “beyond reasonable doubt” (Filppula et al. fc.:4). Some degree of Celtic influence has also been proposed for a number of regional Englishes where it is linked with more recent immigration from Celtic speaking countries, e.g. Bretons in Canada, Scots in New Zealand, Welsh in Patagonia and Irish in the United States.

The Celtic Englishes have been the focus of a series of Colloquia held at Potsdam University that have generated a renewed interest in the question of possible Celtic influences even outside the ‘classic’ Celtic Englishes. While these Celtic Englishes are themselves not the focus of this paper, they will occasionally be drawn upon for reference as they can serve as examples for processes of language contact interference between English and Celtic languages. As Markku Filppula points out, “the linguistic characteristics of the so-called ‘Celtic- Englishes’ that have emerged in the modern period provide yet another important source of indirect evidence supporting the Celtic Hypothesis with regard to medieval contacts” (Filppula 2006:1). Some of the features that will be discussed below are also part of the structure of Celtic Englishes and where a proposed contact feature in Standard English occurs in even higher frequency in the Celtic Englishes this may be a potential indicator for Celtic influence.

2. The Historic Background

In order to assess the linguistic outcome of Celtic-English language contacts, it is necessary to pay attention to the sociocultural environment in which these contacts occurred. A number of different disciplines provide us wit hsometimes conflicting information on the adventus saxonum and the events of the following centuries. These include archaeology, linguistics and genetics.

Since this area is by no means free of discussion, this chapter will outline not only the historical developments themselves, but also, how they were treated in the academic debate.

Of particular interest for this paper is the question of the relations of the Britons with the Anglo-Saxons, including the linguistic situation at the time. Is there any evidence for a survival of a (sizeable) British population or even for a degree of survival of the Brittonic language? Which parts of society, if any, were literate, and if so, in which language(s)? Different scenarios of language contact are known to have radically different outcomes. A variety of factors determine not only the degree of interferences languages can have upon another, but also the fields that influence is exerted upon. So, much of the linguistic argumentation rests on particular interpretations of historical and archaeological evidence relating to the earliest invasions and settlement of Germanic tribes in Britain (Filppula et al. 2002:1).

The last century evidenced a paradigm shift in scientific assessment of the adventus saxonum, ‘the coming of the Saxons’. With the scientific consensus moving from the so-termed ‘double-X theory’, with expulsion and extermination effecting a population replacement, to so called ‘elite replacement’ theories. The former approach was largely based on the few textual sources describing the events, but improvements in archaeological methods as well as a critical reassessment of other available data led to the abandonment of the theory of ‘ethnic-cleansing’. Instead, a large degree of continuity of the population of Britain has come to be assumed, with the Anglo-Saxons simply taking over the Post-Roman society ‘from the top’ in a form of elite take-over (Tristram 2004:100). This so-called ‘new debate’ was “stimulated by theoretical reconsiderations as well as by some new evidence” (Härke 2003:1). In 2002, Markku Filppula summarises these developments:

It seems safe to conclude that the last decade or so has seen us enter a new phase in the history of research on the early Celtic–English contacts: a substantial amount of new research has been undertaken, or is under way, on a wide range of problems covering the general historical and archaeological background to these contacts and the linguistic outcomes in all domains of language (Filppula et al. 2002:22).

Later contacts of speakers of English and Celtic languages are considerably better documented and do not involve as much controversies as the period after the coming of the Saxons to Britain. There continued to be considerable British military opposition to the Saxons, with Wales only coming under English control after the defeat of Llywelyn ap Gruffydd by Edward I in 1282. Indeed, Welsh is still, the most widely used surviving Celtic language. Continuing waves of Viking raiders and settlers that had begun to arrive on English shores in the eighth century led to the establishment of the Danelaw, an area encompassing roughly half of England under the control of the Vikings. It was not until the 10th century that these areas were brought back under English rule. The end of the Anglo- Saxon aristocracy came in 1066 with the Norman invasion of England. Still, the Irish language in Ireland and Scottish Gaelic in Scotland survived. These languages came into stronger contact with English when England brought its neighbouring countries under its control, effectively establishing the English language in Ireland, Scotland, Wales and Cornwall. The subsequent Anglicisation had profound effects on these areas, giving rise to distinct varieties of English that have come to be called ‘Celtic Englishes’ as they display obvious influences of the original languages of the areas. Economic, social and confessional pressure led to mass emigrations into all parts of the World, particularly after the Industrial Revolution and the social upheaval it entailed along with an increase of personal mobility.

2.1 Written Sources

As no contemporary records of the period of invasion exist, we must rely on later sources for information. The most influential textual sources were Gildas’ De excidio et conquestu Britanniae (written c. 500 AD) and the accounts given by Bede (Beda Venerabilis) in his Historia Ecclesiastica Gentis Anglorum as well as the Anglo-Saxon Chronicle (Härke 2003:1). The texts imply that the incoming Saxons performed nothing less than ‘ethnic-cleansing’ in either killing the Britons or driving them over the sea, before they took possession of the vacated island. In his historiographical account, Bede tells of a letter purportedly written by the inhabitants of Britain to the consul of Rome, asking for Roman military support against the invading Saxons.

To Aëtius, thrice consul, the groans of the Britains. … The barbarians drive us to the sea. The sea drives us back towards the barbarians. Between them we are exposed to two sorts of death: we are either slain or drowned (Tristram 1999 3f).

Until the emergence of the so called “New Archaeology” during the 1980s these descriptions were taken at face value, thus giving rise to the theory of ‘extermination and expulsion’. Richard Coates cautions against “the possibility in these sources of rhetorical, politically-motivated exaggeration of the severity of what happened to the Britons, and one must also allow that some of the principal sources were written (in their current form) over 300 years after the events they purport to describe” (Coates fc.:18).

Concerning the limitations and obvious inaccuracies exhibited by these texts, Gary German points out that “[…] the exploitation of original written sources, though of critical importance to our understanding of the languages involved, is only one element among others forming an intricate multidimensional mosaic” (German 2001:126). Of course, other areas of research greatly contribute towards our assessment of textual sources.

Although some written sources indicate that substantial numbers of Britons survived the coming of the Saxons, these were, more or less consciously, ignored as they did not fit in with the traditional model of ‘expulsion and extermination’.

For example, the Law codes of King Ine of Wessex imply the survival of a distinctly British part of the population in the late seventh century. They list, in some detail, the w eregilds ‘honour-prices’ for both Saxons and Britons. While these weregilds are considerably lower for Britons than for Saxons, thus putting the native population at a distinct economic and legal disadvantage. They indicate, however, a certain amount of land-ownership by free Britons (Härke 2003:1). Richard Coates points out that, on the whole, Britons appear to have constituted a recognisable ethnic identity until the 10th century (Coates fc.:19).

2.2 Archaeology

Until the reassessments caused by the ‘new debate’, Anglo-Saxon archaeology was mainly concerned with finding proof for the ‘historical facts’ as portrayed by the written sources instead of researching independently. Härke notes that “a circulus vitiosus was established in which the disciplines confirmed one another by adopting each other’s results as underlying assumptions for their own work” (Härke 2003:2).

In his 1983 study, Christopher Taylor remarks that, considering the numbers of immigrants in relation to the native population: “The Saxon invasions and settlement appear more as the political takeover of a disintegrating society rather than a mass replacement of population” (cited in Viereck 2000:391).

Heinrich Härke notes that recent estimations for Post-Roman population numbers for Britain (ca. 3.7 million inhabitants at the beginning of the 4th century) are roughly the same as those given by the Norman Domesday census for the 11th century. As he sees no reliable archaeological evidence for a dramatic change in population density in Post-Roman Britain, nor any archaeological evidence for plague, famine or slaughter, he argues for a relative stability of the British population (Härke 2003:3). Gary German, agreeing to this model, stresses the existence of “convincing evidence that the Brittonic peasantry largely outnumbered the incoming Germanic-speaking foederatii and their followers who formed a dominant social, economic and military elite” (German 2001:126).

Although in Bede’s accounts the year 449 AD is given for the adventus saxonum, recent archaeological evidence points to a rather earlier date, probably shortly after AD 400, implying a longer duration of less organised immigrations into Britain (Viereck 2000:390).

In her 1999 paper, Hildegard Tristram highlights that a number of different settlement patterns are visible with each area exhibiting its own pattern. They range from complete kinship settlements and settlements by male war-bands to settlements of individuals and their followers with successive reinforcements. Rural areas appear to have been settled first, and, most importantly, there are distinct variations in settlement density. The South-East of Britain was settled most densely by the Anglo-Saxons, whereas in the North and the West of the island the density of Anglo-Saxon settlement was lowest. Additional evidence for cultural continuity is given by the fact that there appears to have been no change in the pattern of land-ownership after the coming of the Anglo-Saxons. Summarising, Tristram argues that “[t]he archaeological evidence can be interpreted at its simplest as showing a smooth assimilation of the two cultures” (Tristram 1999:12f).

Information on the social structure of early Anglo-Saxon England may be derived from comparison of burials. An indicator of status differences is the amount and value of grave goods, which vary considerably but apparently do so in relation to the ethnicity of the buried [INT 4:5].

Until the seventh century, two groups of male burials can be distinguished. One group, constituting 47% of grown men, are buried with weapons while the remainder are buried without [INT 4:4]. Examinations of the graves and skeletons allow to draw the conclusion that only the immigrants and their descendants were buried with weapons, while native Britons make up a significant portion of the other group. It is only towards the seventh century that this distinction ceases to be visible, pointing at large scale assimilation by that time (Härke 2003:9). As certain large burial sites exhibit a discontinuity in physical appearance, they indicate a replacement of native inhabitants. Other sites point at two separate populations living alongside each other, but without intermarrying (Härke 2003:6).

The exact numbers for immigrant and native population are still under debate, with evidence coming from all areas of archaeology, such as e.g. analyses for the amount of forest regrowth after the Roman occupation, analyses of graves etc. Heinrich Härke cites recent estimates for the Romano-British population as ranging between 2 and 4 million inhabitants (Härke 2002:147). He points out that there is “less clear, but still persuasive, evidence of substantial survival of a large native population” (Härke 2002:148f).

Newest estimates of the Anglo-Saxon migration vary considerably more, ranging from less than 10,000 up to 200,000 immigrants [INT 4:1]. This, along with the unequal distribution of the incomers, who spread only thinly in the West and North areas of Britain leads to a number of different estimations for the relative proportions of Anglo-Saxons and Britons. Laing et al. give a ratio of 20 Britons on 1 Anglo-Saxon for the south east and as little as 50:1 for the Anglian north. Härke suggests more conservative ratios, 3:1 in the south east and 5:1 in the north (both cited in Tristram 1999:13).

Concerning modern reassessments of archaeological theories he cautions that “it is also worth bearing in mind that we are as influenced by the zeitgeist and our own expectations as the Victorians were a century ago” (Härke 2003:9).

2.3 Genetic analyses

Recent genetic analyses of Y-chromosome distribution indicate that the Anglo- Saxon contribution to the modern English gene-pool lies between 50% and 100%, also finding significant dissimilarity between the distribution of certain Y- chromosome haplotypes between Central England and North Wales. In order to explain this level of influence solely with mass immigration, Thomas et al. estimate a necessary influx of approx. 500,000 people. Since no movement of this scale is attested by the archaeological data, they set out for alternative models to explain the modern genetic distribution. They argue that “[a]n alternative explanation would be provided by an apartheid-like situation […] in which elevated social and economic status grant higher reproductive success to the immigrants when compared to the native population and a degree of post- migration reproductive isolation is maintained among ethnic groups for several generations” [INT 5:1].

Evidence for such division of Anglo-Saxon social structure along ethnic lines is present, e.g., in the significant differences in legal status assigned to Britons and Anglo-Saxons in the Laws of Ine [INT 4:2].

Thomason et al. see the imposition of an apartheid-like social structure as a means of securing political and military control by a small immigrant population that would otherwise risk assimilation with subsequent loss of power. This model also provides an explanation for the long span of skeletal distinctiveness, as in an apartheid-like system a low degree of intermarriage between the incomers and the natives would be expected, precluding an assimilation of physical features [INT 4:4].

This distinction appears to have been upheld at least until the seventh century, when the two groups cease to be distinguishable archaeologically. No ethnic distinctions are made any more in the Laws of Alfred the Great (c. 890 AD), so Thomas et al. assume a maximum of fifteen generations of ethnic division after the coming of the Saxons. [INT 5:2f].

Calculating different rates of population development in different theoretical social environments, they conclude that “the genetic contribution of an immigrant population can rise from less than 10% to more than 50% in as little as five generations, and certainly less than fifteen generations” [INT 5:6].

Correspondingly, Bryan Ward-Perkins sees a continued assertion of alterity, ranging from the earliest sixth century sources until the tenth century where again and again the natural, seemingly innate, differences between Anglo-Saxons and Britons are stressed. He adds that “the broader evidence of failed contacts, in religion and in language, provides strong support for the idea that this perception of difference was no mere literary construct, but was felt (and lived) throughout society” (Ward-Perkins 2000:2).

2.4 Conclusion on Sources

On the whole, the historical evidence suggests an ‘elite-transmission’ with subsequent cultural assimilation rather than, a ‘clean sweep’ with large scale population replacement, as previously favoured and still upheld by a number of scholars.

The ‘New Debate’ in archaeology concerns itself with the ethnogenesis of Anglo-Saxon England. The works of Lloyd Laing, Nicholas Higham and Heinrich Härke showed “that the nineteenth century ‘Anglo-Saxonist’ ideology of the Germanic racial ‘purity’ of the Anglo-Saxon society cannot be maintained in the light of recent archaeological research” (Tristram 2004:100). Härke notes that the question of racial purity had a tangible political background that prohibited the notion of a Celtic element in the English population. With Ireland demanding Home Rule, the question was: “were the Celts able to govern themselves, or did they need English masters to look after them?” (Härke 2003:2). This also reflected on British attitudes towards the Celtic languages that were seen as impeding economic progress in the Celtic areas. While they were sometimes admitted to possess a certain ‘aesthetic value’, the overall attitude towards them remained negative. So, the prevailing mood of Anglo-Saxonism continued to have not only influence on the interpretation of data and sources, but also on the direction of research, with the possibility of any Celtic influence on the English language only emerging in discussion rather recently.

Objections to this new approach have been voiced, e.g., by Richard Coates, who argues for cultural annihilation by means of enslavement rather than large scale survival of free Britons or extermination at the hands of the Saxons. While he concedes, e.g., the possibility of ‘slave-coloured’ variety of English emerging in Brittonic communities, illiteracy on part of the slaves prevented its documentation and it could not have had great influence on the standard language as it was a severely stigmatised variant (Coates fc.:19).

3. Language Contact

Before proceeding to the contact(s) between English and the Celtic languages, a theoretical background for establishing the mechanisms of language contact is necessary. In this paper, the approach of Thomason and Kaufman (1988) and Thomason (2001) will be followed, so their main theories concerning the differentiation between situations of borrowing and language shift, as well as their systematics for determining contact interference for a given feature will be presented.

As the basis for a systematic approach, the discipline of contact-linguistics provides us with a theoretical framework which not only helps in the analysis of a given contact situation or phenomenon, but comparison with similar situations may also grant us insight which is otherwise unavailable. Light will be thrown on the sociocultural background of a language-contact situation from what we know about its outcome. Conversely, the observation of language contact phenomena, as Pieter Muyshen points out, contributes to our general understanding of syntactic structures and their roles in the behaviour of a language (Muyshen 1996:117).

Most importantly, Thomason and Kaufman distinguish between language contact situations involving borrowing and those involving language shift. When two languages come into contact with each other, the typical result is borrowing of some material from one language into another, usually starting with lexical items in form of loanwords. If the contact between these languages is strong enough and cultural pressure is made on the speakers of the receiving language, eventually anything may be borrowed, including morphosyntactic features (Thomason and Kaufman 1988:37ff).

This matches closely with the traditional approach towards language contact, which postulates that contact influence first and foremost takes the form of lexical influence. While the transfer of lexical items may be easiest to prove, this form of contact is by no means the only one (Thomason 2001:64). In a situation where a large number of speakers abandon their native language, thereby ‘shifting’ towards an other language, the outcome is almost the exact opposite: contact interference then starts with features of phonology and syntax and only a small numbers of lexical items are transmitted, if any at all (Thomason 2001:75). This, of course, makes it far easier to spot historical language contact that has taken the form of borrowing, especially when information on the overall social and linguistical situation at the proposed time of contact is scarce.

3.1 Rapid Shift with Imperfect Learning

Concerning their approach towards what has traditionally been termed ‘substratum interference’, Thomason and Kaufman state that:

Substratum interference is a subtype of interference that results from imperfect group learning during a process of language shift. That is, in this kind of interference a group of speakers shifting to a target language fails to learn the target language (TL) perfectly. The errors made by members of the shifting group in speaking the TL then spread to the TL as a whole when they are imitated by original speakers of that language (Thomason and Kaufman 1988:38f).

These learners now carry features from their original language into their version of the TL (then called TL2). If this group of speakers is then integrated into a shared speech community consisting of both speakers of TL1 and TL2, this will lead to a shared variety, TL3, emerging because the original speakers of TL1 will take over some of the distinct features of TL2 into their language. Thomason here speaks of ‘negotiation of a shared version’ (Thomason 2001:75).

She stresses the importance of markedness in shifting situations, as marked features of a TL are not only less likely to be learned, and thus less likely to appear in the TL2 of the shifting speakers, but are also less likely to be taken over from this TL2 into the shared TL3 by the original speakers of a language. (Thomason 2001:76).

In addition, crucial importance is attributed to the relative sizes of speaker communities. If the shifting population is numerically larger than the amount of original speakers, this improves the chance of at least some of the shift interference being taken over into the ‘new’ community language. An example given by Thomason is the presence of Irish Gaelic features in the English of Ireland due to large numbers of shifters compared to the incoming ‘native’ speakers of English (Thomason 2001:78f).

Thomason and Kaufman point out that “[i]n changes resulting from imperfect learning of a second language, the TL is not so much accepting the changes as giving in to them, since it is the shifting speaker, not the original TL speakers, who initiate the changes” (Thomason and Kaufman 1988:43). They also note the ineffectiveness of strong attitudes towards this influence if the shifting speakers are numerous enough to ‘impose’ their variety upon the community as a whole.

Thomason criticises the common tendency to judge the probability of language contact by the amount of loanwords from a language, and assumes that the absence of lexical interference precludes any influence in any other area of the language, or even indicates lack of language contact (Thomason 2001: 80).

A scarcity of loanwords is to be expected, since the shifting speakers may see no reason to preserve their original language due to the strength of economical, political or other pressures that led them to abandon their language in the first place, quickly acquiring the new lexicon (Thomason and Kaufman 1988:117).

3.2 How to Define Contact Features

Thomason, gives the definition for contact-induced language change as follows: “any linguistic change that would have been less likely to occur outside a particular contact situation is due at least in part to language contact” (Thomason 2001:62). These changes may take the form of direct transfer from one language to another, but also more indirect influences (Thomason 2001:62). Concerning the search for the source of a certain change in a language, Thomason and Kaufman maintain that:

[A] successful criterion for establishing external causation is possible only when we consider a language as a complex whole–a system of systems, of interrelated lexical, phonological, morphosyntactic, and semantic structures. Instead of looking at each subsystem separately, we need to look at the whole language (Thomason and Kaufman 1988:60).

Firstly, in order to establish language contact as a cause for a given feature a source language with which the language evidently has been in contact must be identified. This is naturally more problematic in shift situations (which may well lead to the eventual death of the shifting speakers’ original language) than in those of borrowing situations. Secondly, ‘shared structural features’ have to be identified in both languages. One must be aware, that in order to establish a shared feature, a one to one identity is not necessary, even unlikely. One also has to prove that the proposed features were not part of the system of the receiving language, and that they were present in the donating language before the contact occurred. Finally, plausible internal motivations for any change have to be considered as well. Here, Thomason points out the possibility of ‘universal structural tendencies’ for language evolution as well as the possibility of multiple causation (Thomason 2001:93f).

She also draws attention to the unlikeliness of solitary contact features occurring, noting that “an argument for a contact origin will only be convincing if it is supported by evidence of interference elsewhere in the language’s structure as well” (Thomason 2001:93).

Concerning the identification of a source for a given change in a language, Thomason stresses the difficulty of distinguishing between contact-induced and internally motivated changes. In either case, features may be lost from a language, a language may gain certain features or native features may be replaced by new ones. (Thomason 2001:86f). Furthermore, a clear reason for a given feature may not always be obtainable for a given feature and “the possibility of multiple causation should always be considered and […] it often happens that internal motivation combines with an external motivation to produce a change” (Thomason 2001:91). Thus, the possibility of multiple causation should be borne in mind when establishing the cause of a given feature.

Filppula et al. note that the complete identity of features is not a necessary “especially [as] syntactic parallels between the substrate and the emerging contact variety are often only partial in nature” (Filppula et al. fc.:2). In addition, speakers may overgeneralise on features that resemble those found in their native language, or develop uses for them that were not previously part of either language. Thomason and Kaufman stress that “many interference features will in fact not be exactly the same as the source-language features that motivated the innovations. Lack of ‘point-by-point-identity’ must therefore not be taken to mean that an innovation is not due to foreign influence” (Thomason and Kaufman 1988:62). Also, they point out, the fact that an internal motivation can be determined for a change in one language, does not have to be a valid explanation for the same change occurring in an other language (Thomason and Kaufman 1988:59).

A further complicating aspect of interference through shift is that it most often results in grammatical simplification, thus making it hard to distinguish from any internal simplifications in a language (Thomason and Kaufman 1988:114).

Markku Filppula explicitly discounts the ‘principle of parsimony’ that has been voiced by Roger Lass when regarding the problematic issue of distinguishing between internal and external causation for a feature. While this principle states that in cases where both an external and an internal explanation are available, endogeny is always preferable, Filppula argues for possible external influence even in features where an internal explanation is possible as well (Filppula 2003a:161). He admits that the burden of proof lies with those wanting to establish contact influence rather than with those arguing for internal causation, but stresses that “the quest must always be for the best explanations whether more or less parsimonious” (Filppula 2003a:170).

Additionally, the idea that substratum and superstratum form identifiable, discrete layers in a language has come under criticism, e.g. by Markku Filppula who instead argues for “intricate patterns of variation which exist in contact vernaculars both at the inter- and intraindividual level” (Filppula 2000: 322), further complicating the issue of identifying the source of a given feature. Still, he warns against accepting multicausation as a default solution, stressing instead the need for careful search for evidence pointing out a feature’s most likely source (Filppula 2001:23).

Another problem in establishing historical language contact as a source for a certain development may be the long latency before any innovation becomes attested in written language. Spoken and written language may differ considerably from one another, the written form usually being more conservative by far, while any contact influenced changes are likely to take place in the spoken variety of a language. This effect becomes especially important in Hildegard Tristram’s approach of assuming a form of diglossia after the shift of large numbers of speakers of British towards Anglo-Saxon (Tristram 1999: 27). This approach will be discussed in detail below.

It should be noted, however, that the debate on the theoretical background of contact linguistics is far from being settled. In particular, the question of what features may be indicative of language contact remains. While Tristram assumes contact influence in the transfer of features from one language to another, she also considers not only shared innovations between adjacent languages but also shared retention of features as indicative of contact influence (Tristram 2002a:260). This is debated by Graham Isaac, who sees only shared innovations as indicative of contact, while shared retention of archaisms may be due to pure coincidence (Isaac 2003:53). Likewise, Muyshen sees language convergence as a distinct form of contact influence as well, with the distinction that it is a bi-directional process affecting not only one language in a contact situation (Muyshen 1996:121). Clearly, an expanded theoretical background is necessary for this matter.

All in all, Filppula et al. summarise that “ascertaining contact influences is more a matter of greater or smaller likelihoods than of achieving definite proof” (Filppula et al. fc.:3).

4. English and Celtic in Contact

In the following chapter, different theories and approaches towards the contact of English with the Celtic languages will be presented. As mentioned above, the traditional view that the Britons were annihilated by the incoming Anglo-Saxons, thus leaving no trace of their original language has come under debate. Analogous with the historical and archaeological reconsideration of the issue, more and more research is undertaken suggesting the survival of distinctly Brittonic features in the English language of today.

In contrast to past approaches that dismissed the possibility of Celtic influences on English on the grounds that no significant lexical loans could be found, advances in contact-linguistics have lead to the theory that in situations of language shift such scarcity is to be expected. The proposed process of rapid language shift with imperfect learning would see these substantial numbers of surviving Britons abandoning their native language, rapidly shifting to Anglo- Saxon and thereby introducing a tangible amount of influence features into their new language. Some of those features would survive to become part of the standard language, making this Celtic influence felt even today. Expectably, the validity of such theories has been debated, with the occasional rise of tempers hinting at the ideological implications such claims entail. Before any proposed features are discussed in detail, the process of how they may have found their way into the English language of today will be examined.

A wide variety of languages were spoken in Britain over the course of its history. When the first speakers of a Celtic language arrived on the British Isles, they encountered a native population, presumably speaking a non-Indo European language. With the Romans came Latin that, besides being the language of the Roman occupation, continued to have a profound influence as a prestige language well through the middle ages, e.g., in the domains of religion as well as science. Continental mercenaries and auxiliaries within the Roman army in Britain spoke a wide variety of languages, among them Germanic dialects. Raiders and settlers from neighbouring Ireland brought with them their Gaelic language, which would later develop into Scottish Gaelic in the area settled by the Irish scotii,. After the Roman retreat and the adventus saxonum we encounter a variety of Germanic dialects, eventually becoming Old English. The Brittonic language previously spoken all over England was pushed back into distinct areas, where it developed to Welsh in the West, Cornish in the South West and Cumbric, spoken in what remained of the British area north of the Anglo-Saxon territory. The Cumbric area was Anglicised in the Old English period, Cornish as well as Manx, spoken in the Isle of Man only died out in the Modern Period. Welsh as well as Irish and Scottish Gaelic survive to the present day, thus being possible contact candidates with their neighbour English for one and a half millennia. Over this time, the different languages went through a number of developments that brought them to their present form. This involved different kinds of contact situations under different socio-cultural influences.

Although Old English can be seen as a thoroughly Germanic language of a distinctly synthetic character and mainly exhibiting the grammatical categories it inherited from Indo-European, it underwent a number of changes over time. In the course of its development over Middle English, Early Modern English to Present Day English it developed a distinct character of its own, setting it aside from its Germanic cousins. Of these, only the later changes are well documented in written form. Changes that took place before the appearance of any number of written documents in the 8th century are largely undocumented. This is problematic insofar as it is some of these changes that play a great role in alienating English from its Germanic source, although later changes would further increase this distance (Meid 1990:112).

[...]