Version classiqueVersion mobile
OpenEdition Books

Travaux de Diachronie 2

 | 
Jean-Paul Regis

The Genesis of Analytic Structure in English: The Case for a Brittonic substratum1

Gary German

Texte intégral

  • 1 I wish to thank Dr. Hildegard Tristram (University of Freiburg) and Dr. Jean Le Dû (University of (...)
  • 2 Traditionally, the Celtic languages have been divided into two branches: P-Celtic, of which Britto (...)
  • 3 This point was used as proof to argue that the English were, as a consequence, "purely Teutonic" ( (...)
  • 4 Chadwick (1963) convincingly argued that these sources had simply been misunderstood.

1Numerous scholars have been struck by certain morphosyntactic features shared by English and Celtic which do not necessarily characterise other Germanic languages. Nevertheless, the idea of a possible "Celtic" (i.e. Brittomc2 substratum in English has been generally brushed aside for both historical and linguistic reasons or simply explained away as typological universals. Furthermore, until recently, 19th century historical views of the "Anglo-Saxon conquest", on which this conclusion is based, have gone virtually unchallenged, namely that the Britons were wholly "extirpated" or driven out of England (Freeman 1867: 18) by the incoming "Anglo-Saxon" conquerors and that the "tolerated remnant of their predecessors" (i.e. the Britons) were enslaved (Stubbs 1870: 1). Intermarriage would have been forbidden explaining why there are so few Celtic loanwords in English (Freeman ibid.; Stubbs ibid.). There was thus no reason to believe that there was any fusion between the two populations.3 Furthermore, literal interpretations of the oldest historical sources of the period, Gildas and Bede,4 outwardly appeared to reinforce such conclusions. This is still the dominant view to this day (Burchfield 1986, Algeo and Pyle 1993). How could there be a Brittonic substratum in England, if the Britons were killed or driven out of the country? The logic, of course, is circular. Even the most generous accounts proposing that the Britons largely survived the colonisation (Jesperson 1938) argue that the entire Brittonic-speaking population was quickly assimilated without leaving any linguistic trace.

2The fact that there are no existing texts preserving Brittonic and Old English contact vernaculars further complicates matters and would appear to make the entire debate a moot point since it is assumed that, given the dearth of textual sources, the question cannot be answered one way or another.

Sociolinguistic and sociohistorical considerations

3Yet, despite these reservations, we would argue 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.

  • 5 Härke will be showing evidence from some initial DNA data which would support this theory (Potsdam (...)
  • 6 One of the most conservative of these estimates is by Härke (1995): 3 to 1 in the South and Midlan (...)

4Contrary to what was once believed, recent archaeological and historical studies (Chadwick 1963, Alcock 1980, Laing, M. & J. Laing 1990, Higham 1992, Davis 1993, Härke 1995, 20015 of the 5th to the 8th centuries present 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.6 Writing in 1953, even the prudent Jackson expressed the view that: "the 'clean sweep theory' has been pretty well abandoned now, partly on the grounds of general probability and partly on the basis of direct evidence."

  • 7 In his monumental History of Wales Davies (1993: 68) writes: "The greater part of the pre-English (...)

5If true, one wonders why linguists have clung so tenaciously to the linguistic counterpart of this theory, that is, that there is no Brittonic influence on English. Indeed, accepting the large scale survival of Britons would require a revision of the way historical linguists have viewed the development of the English language until now. The importance of this point cannot be understated. Under such conditions, the anglicisation of England occurred, not as the result of ethnic cleansing but, rather, as the consequence of a language shift from Brittonic to Old English, a process which, depending on the geographical region involved and the social level of the Britons, dragged on for hundreds of years. Put another way, the anglicisation currently underway in the Celtic-speaking countries would have to be viewed as the end of a process which began 1,600 years ago with the arrival of the first Old English speakers to Britain.7

6Despite obvious problems with sources, the hypothesis of a possible Brittonic substratum in English has been reinforced in recent times thanks to advances in historical linguistic and sociolinguistic research on language contact which shows how such a language shift could have occurred. Indeed, Thomason and Kaufman (1991) convincingly argue that one of the errors linguists have made in the past in dealing with language interference is to confuse lexical "borrowing" and "substratal influences", two separate processes. For them "borrowing" refers "only to the incorporation of foreign elements into the speaker's native language" (1991: 21). Substratum effects, on the other hand, result 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 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 the original speakers of that language…

7Using numerous case studies, they demonstrate that such substratal influence on a language primarily concerns grammatical structures rather than vocabulary. This would simultaneously seem to answer the frequently asked question of why there appear to be so few Celtic words in English (Mossé 1947: 37) and provide a logical explanation for morphosyntactic parallels between Celtic and English.

8In this paper, we shall concentrate on how Brittonic may have provoked the dramatic shift from a synthetic to an analytic language that characterises the history of the English language from the Old English to Middle English periods. We shall then outline some morphosyntactic parallels between Brittonic and English. Given the complexity of the subject, the information in this article should be considered to be a simple summary of features rather than an in-depth analysis of the process.

  • 8 This point is not necessarily in contradiction with the hypothesis that Breton probably contains a (...)
  • 9 Recent research would suggest that the colonizers of Brittany were primarily southwestern Britons (...)

9In searching for Brittonic parallels with English, we shall focus primarily on Breton, since it is the only living form of P-Celtic to have evolved from the varieties once spoken in the southwest of England. These varieties were presumably closest to the southeastern and eastern varieties of Brittonic with which the "Anglo-Saxons" would have been in direct contact.8 Furthermore, unlike the insular Celtic languages, Breton has never been subjected to direct English influence. For this reason, typological similarities between Breton, Cornish and Welsh probably reflect the Brittonic vernaculars prior to the Brittonic immigration from southwestern England to Brittany which ended at the beginning of the 8th century (even if there is no textual proof of this).9

Survival of Brittonic

  • 10 This could explain the existence of Brittonic sheep-counting numerals in Cambridgeshire, Lincolnsh (...)
  • 11 Cognate with Welsh "Cymru" (< *cum-brogi "the fellow countrymen").
  • 12 When one considers that the Welsh aristocracy was largely anglicised by the 16th century and that (...)

10If one accepts the idea that the Britons were not exterminated, the entire question of a Brittonic substratum hinges on how long the Brittonic language survived, who actually spoke it and the degree to which it exerted influence sociolinguistically. As the extra-linguistic details concerning the survival of Brittonic and the manner in which the shift may have occurred have been discussed in greater detail elsewhere (German 2000), we shall not discuss the matter further. Suffice it to say that it is clear that it could not have died out immediately as has often been supposed. In fact it would not be unreasonable to posit that Brittonic lingered on in pockets among the peasantry until the 10th - 12th centuries, especially in the regions peripheral to the English Midlands: along the Welsh border, and in scattered areas of the North-West and South-West.10 Even the Anglo-Saxonist Freeman (1867: 35) accepted that the inhabitants of the Southwest of England were simply "naturalised Englishmen" wholly Brittonic "in blood". He further adds that their language was still spoken in Exeter during the reign of Aethelstan in the first half of the 10th century. In Cumbria11 the military power of the Britons was not broken until 1092 and parts of Strathclyde were Brittonic-speaking until, at the very least, the 11th century (Jackson 1953: 219). Indeed, Cornish did not die out in Cornwall until the late 18th century and Welsh was still spoken in Hereford until the 20th century. Moreover, simply because the Brittonic aristocracy may have been exterminated, knocked from power or simply anglicized in a given area does not in any way mean that the peasantry (i.e. the mass of the population) followed suit immediately.12

11This implies that, after the arrival of the first Old English speakers to Britain, the language shift would have taken approximately 400 to 700 years to affect the peasantry in the more isolated regions of England and Scotland leaving more than a chance for Brittonic features to have crept into the language. They did so either through direct Brittonic influence or, as we shall argue, through the medium of what we shall call Anglo-Brittonic dialects, that is to say, Brittonic-influenced basilectal forms of English that probably lingered on and made themselves felt within the Old English linguistic community long after the language shift was complete. This could explain why some of these features do not appear in writing until the Middle English period, probably as a direct result of the collapse of the Anglo-Saxon social structure.

Anglo-Brittonic basilectal dialects

12That such basilectal varieties existed is highly probable. Burchfield (1986: 4) has posited the existence of what he calls "Vulgar Old English" which he describes as "an unrecorded species of ancient spoken English" and claims it "must have been substantially different from the language recorded by the scribes", a Schriftsprache reflecting the speech of the Anglo-Saxon ruling elite from which the clergy was probably issued. If the peasantry were largely descended from the Brittonic inhabitants, it therefore seems natural to suppose that the speech of the peasantry contained the greatest amount of Brittonic influence. Given what is known about creolization and language contact (cf. Gumperz and Winston 1971; Thomason and Kaufman 1991), we would expect these basilectal forms of Anglo-Brittonic to be characterised by a number of features such as:

  1. morphological simplification
  2. morphosyntactic calques on Brittonic
  3. lexical borrowing (but only during the first stage of the shift while Brittonic was still a living language13
  4. phonological interference.
  • 14 He argues that less than 5% of the population in England ever spoke French at the peak of French i (...)

13This kind of evidence would seem to substantiate Blake's (1996: 107) view that the dramatic shift from Old English to Middle English is more apparent than real.14

Early linguistic evidence: the loss of nominal case endings in Brittonic

14One of the most striking and significant features of late Brittonic is that, by at least the end of the 6th century, "all the British case-terminations vanished, thus radically changing the morphological and syntactical character of the language" (Jackson 1953: 618).

  • 15 The loss of case endings is what led to the mutation system in both the Brittonic and Goidelic lan (...)

1st century:

Sindos caítos biccanos

Sinda merka biccana

6th century

/hindëh kattëh bi'xanëh/

/hindë verxë vi'xanë/

Modem Breton

An/An c'haz bihan

An/ar verc'h vihan 15

The small cat

The small girl

15Thus by the Middle Welsh period, old Brittonic names recorded by Gildas, such as Maglocunus or Constantinus, are rendered respectively Maelgwn and Custennin. The loss of nominal inflexion in Brittonic opened the door to a series of changes which led Brittonic to become an analytic language. This is also one of the hallmarks of English and is highly significant to our argument. As Crépin (1994: 60) puts it: "l'anglais a simplifié à l'extrême ses flexions, il est devenu surtout analytique." Why did this change occur first in Brittonic? How long after did it occur in English and why?

  • 16 In this writer's view, this is not necessarily a sign of "language death" as some have maintained.

16We believe that there may be two interrelated causes which initially prompted this change in English. The first is that Brittonic speakers may have imported their stress system into Old English, perhaps of the kind still heard in modern Cornouaillais Breton and which is characterised by such force that atonic syllables weaken and disappear (German 1984).16 Fleuriot (1980: 65) claims that what is happening in this Breton dialect now is a repeat of what once happened in Brittonic. It is the stress system which led to the weakening and loss of final atonic syllables in Brittonic (Jackson 1953) and, in this regard, what Crépin (1994: 35) says for English is especially relevant: "Et parmi les langues germaniques, l'anglais est celle qui a le plus creusé l'écart entre les syllabes fortement accentuées et les autres. La plupart des syllabes inaccentuées ont disparu, et leur disparition a multiplié les monosyllabes" (our italics).

17The second possibility is bound to the first. Since they had already lost their own case endings, Britons would have had great difficulty trying to master the highly inflected Old English language, particularly in an uncontrolled learning context. Moreover, given that morphological reduction is a widespread consequence of language contact (Thomason and Kaufman ibid.), it seems logical to suppose that reduced morphological complexity would have been one of the most apparent traits of the Anglo-Brittonic varieties (i.e. Old English spoken by Brittonic learners).

18Indirect supporting evidence for such morphological reduction in English could come from Brittonic versions of Gildas' Lorica showing an inability on the part of Brittonic scribes to master the use of Latin nominal inflection, a concept which they had, by now, lost in their own language. If learned clerics had such difficulties in Latin, it is not surprising that the Brittonic peasantry would have faced similar difficulties with regard to Old English. Crépin (ibid.) adds a very pertinent observation which goes in the direction of what has just been said:

L'affaiblissement des suffixes de désinences dès la fin du vieil-anglais s'est accompagné de l'extention de l'emploi des prépositions, la disparition de la négation multiple et le peu de volume du négatif simple ont favorisé l'essor de l'auxiliaire DO.

19Here too, we observe a similar evolution in the Brittonic languages, namely the development of periphrastic DO (Welsh gwneuthur, Cornish gruthyl, Breton (g)ober) and the extended analytic use of prepositions. Based on what has been said, we could deduce that such changes occurred first in Brittonic. During a second stage, indirect Anglo-Brittonic influence from within the Old English linguistic community thus resulted in the resyntactification of basilectal varieties of Old English (Tristram 1995, 1997, 2000, German 1996, 2000). Considering the stigmatised nature of Brittonic, such influence would not have been immediately apparent in the literary language for centuries to come. This possibly spawned many of the typological similarities that can be observed today between the surviving Brittonic languages and some English English, Celtic English and Breton French varieties today. Tristram (2000) eloquently expresses a similar point of view:

In many aspects of its morphosyntax, present day English is a "Brythonised" dialect of West Germanic. I would say that the very vital contribution of the speakers of the Brythonic languages to the creation of the English language lay in triggering the initial typological change from a predominantly synthetic language to a predominantly analytical language. Therefore, this contact determined that all subsequent changes would tend towards analyticity. Since Brythonic has been classified by scholars as belonging to the family of Insular "Celtic" languages, English, therefore, may be called a "Celticised" West-Germanic language.

20By the 9th century, and where it had not already occurred as a result of Brittonic influence, it is probable that later contacts with Scandinavian languages (Milroy 1991) accelerated a process that was probably already under way. Thomason and Kaufman (1991: 303) share this view:

  • 17 Crépin (1994: 75) demonstrates that the definite article the begins to take over the functions of (...)

The Norse influence on English was pervasive, in the sense that its results are found in all parts of the language; but it was not deep, except in the lexicon. Norse influence could not have modified the basic typology of English because the two were highly similar in the first place. Norse did not stimulate simplification in English, since the simplifications we see in ME when compared to OE probably were taking place in English before Norse influence became relevant, (our italics)17

21Yet, even later ME simplification in the Northern dialects, they claim, did not "correlate with anything in the structure of Norse" (ibid.). If this is true, what prompted these changes in the first place? Despite their conclusion, the authors do not consider the possibility of Brittonic influence, perhaps because of the reasons outlined in our introduction. Nevertheless, the two hypotheses need not be mutually exclusive but rather should be seen as complementary processes. The arrival of the French-speaking Normans simply completed, and perhaps masked, developments that had begun centuries before.

Morphosyntax. DO periphrasis in English and the Brittonic languages (gwneud/ober/ gruthyl)

  • 18 Cf Preusler 1956, Teynière 1959, Poussa 1990, Gachelin 1990, German 1996, Tristram 1997, van der A (...)

22Linguists have long considered the possible link between Brittonic and English periphrastic DO.18 Teynière (1959: 211), for instance, writes: "on sait que le brittonique de Grande-Bretagne est un des substrats qui a pu agir à distance sur l'anglais". If so, what are the underlying Brittonic structures which could have, either through direct Brittonic influence or indirect, but longer-lasting, Anglo-Brittonic influence, led to the adoption of DO periphrasis in English? We believe the question has far-reaching consequences and is linked to Celtic cleft constructions. Furthermore, the morphological simplification of the verb in Brittonic went hand-in-hand with this process.

23We posit that there were probably two ways to express DO periphrasis in Brittonic, both involving the fronting of the atonic copula IS. The first structure presented below is attested in Middle Welsh, Middle Comish and is still common in modern Breton, but without the copula. The fact that the relative particle a remains, however, would seem to support the hypothesis that, in diachronic terms, the copula was indeed present.

a) (IS?) + thematised INF. + relative particle A (who/that: provoking lenition) + conjugated GWNEUTHUR (DO) + SUBJ

  • 19 Komz a raent (speak[ing] they did).
  • 20 One wonders whether it clefting in Breton French recalls the older Brittonic structure. In other w (...)

24This may have given constructions such as Middle Welsh ymdidan a wnaethont, perhaps from older *ys ymdidan a wnaethont (It is speak that they did).19 Although IS is unattested before a verbal noun or infinitive (however, cf. Fleuriot 1985, Old Breton Is bulch: lit. ?it is open(ing)), the use of an identical cleft structure (C'EST + INF + FAIRE) is in common use today in the French of Western Brittany, where it is normally emphatic, and clearly a calque on Breton. Indeed, the above structure could be adapted literally in response to a question such as: "Il pleure?" to which one could easily reply "Oh non, alors! C'est rire qu'il fait!" (< [IS] c'hoarzin a ra; lit. It is laugh that he does).20 The second possibility is more problematical:

b) (IS?) + personal pronoun + relative particle A (who/that: provoking lenition) + conjugated GWNEUTHUR + INF

  • 21 A native Breton speaker from Landunvez, Leon (PC), claims that his father says "me a ra drebi" wit (...)

25Although unattested, it may have yielded a structure such as fe naeth ef mynd (lit. He did-he go; i.e. 'He did go') which is very common in modern spoken Welsh and which perhaps originated from older *ys ef a wnaeth mynd (It is he who did go). This idea seems reinforced by the fact that the structure also exists in Cornish: ef a-wra mones (lit. He who does go) < *(is) ef a-wra mones. It appears that at some point before the separation of Cornish and Welsh, DO (gruthyl/gwneuthur) came to function as a mere tense and person marker for the infinitive (mones/mynd, etc.).21

26There is indeed evidence to reinforce this hypothesis as a similar structure does exist in Middle Welsh. In inital position of a sentence one encounters SEF, a synchretism of both the copula YS + the third person masculine pronoun EF meaning ‘it is he'. What was almost certainly fully productive in earlier Brittonic (ys mi, ys ti, ys ef ys hi, etc.) had, by the Middle Welsh period, become a frozen syntagm, ‘SEF.

Sefa wnaeth Arthur kyuodi a mynet y kymryt kyghor.
(lit. [It is he] who did Arthur arise and go to take counsel)
What Arthur did was to arise and take counsel (translation by Evans 1976: 52)

27SEF was already bereft of its semantic content by the Middle Welsh period (copula + pronoun) and this could explain why the real subject, Arthur, is in post-verbal position here, in keeping with the VSO order that characterizes modern Welsh syntax. The same phrase in Modern Welsh would give fe naeth Arthur cyfodi (lit. pre-verbal particle fe < ef + Did Arthur rise). Considering this modern structure, older Welsh could possibly have had: * Ys Arthur a wnaeth kyuodi a mynet y kymryt kyghor (It is Arthur who did rise and go to take counsel; i.e Arthur did rise and go to take counsel). Considering the Cornish evidence cited above, the hypothesis does not seem rash. Could this be the kind of structure which was adopted into Anglo-Brittonic varieties of English? If so, the English construction would be a word-forword calque on Brittonic (ex. NP + DO + INF). Under such conditions, it becomes easier to understand how DO could have come to be a simple tense and aspectual marker in both the south-western English dialects (see Aspect below) as well as in Cornish and Welsh.

Morphological simplification of the verb in Brittonic and English

  • 22 Lambert mentions that French cleft constructions (ex. C'est Jean qui parle) could be due to Gaulis (...)

28Linked to the question of IS-clefting is the morphological simplification of the verb in Brittonic. Hardie (1949) gives the following Old Breton source for modern Breton: me a gar (me + a + invariable gar: I [who] love[s]) < *Is mi a karam (Lit. Is me who love-I; i.e. It is I who loves). It is also similar to our b) construction above. That this cleft structure once existed in Brittonic is beyond question since it is found in archaic Middle Welsh: Is mi a-eh eirch: (It is I who-her asks; i.e. It is I who seeks her). The atonic copula, IS, is absent, phonetically at least, in later Middle Welsh, Middle Cornish and in Breton, the relative particle a, and the lenition it provokes, being the only trace of its (former) presence. The existence of such structures in Old Irish: Is hé dia as éola indium-sa (It is God who is knowing in me) (Thurneysen 1946:492) and possibly in Gaulish (Lambert 1994: 68)22 proves that it is certainly an ancient and integral part of both Goidelic and Brittonic syntax. Thus what is sometimes interpreted as straightforward SVO/SOV structure in Breton and Middle Welsh and Cornish is in reality more complex than this.

29This situation in Brittonic led Tesnière (ibid.: 138) to observe parallels between English and Breton: "l'indice seul assure la conjugaison personnelle dite impersonnelle en breton, du verbe tandis que le verbe lui même reste complètement invariable." In many English dialects, of course, this process has gone the full route where the third person singular //-s// is often absent. Conversely, //-s// can appear for all persons but, in many varieties, it can have an iterative or habitual function (Shorrocks 1999).

me a gar: lit. I (who) love(s)

ni a gar: we (who) love(s)

te a gar: lit. thou (who) love(s)

c'hwi a gar: you (who) love(s)

hi a gar: lit. she (who) love(s)

int a gar: they (who) love(s)

30Compare this to non-standard English:

  • 23 This is partly due to the loss of southern 3rd pers. sing. //-th//: he loveth > he love.

I love(s)

We love(s)

Thou love(s)

You love(s)

He/she/it love(s)23

They love(s)

31Although the tendency to systematize the use of //–s// or //zero// morphemes varies according to region, it is clearly another example of the shift towards analycity which occurred in both English and the Brittonic languages. Here we are probably dealing with a case of convergence but, as Tristram (2000) points out, the original impetus for the move to analycity in English would presumably have stemmed from an earlier morphosyntactic breakdown in Brittonic.

Northern subject rule

  • 24 Breton offers a parallel structure here: Ema Yann ha Mari o vond (IS [situation] John and Mary ago (...)
  • 25 Maen consists of two morphemes: //mae// "is" + //n// "they", the latter being the third person plu (...)

32Tied to the morphology of the English verb in non-standard English is a feature that is sometimes called "Northern subject rule" where a plural subject is used with a singular verb: John and Mary goes. When the plural subject is replaced by the third person plural pronoun, however, we have: They go. Klemola (2000) suggests that this lack of subject-verb concord may have a Brittonic source and that the distinction between the two constructions was already well-established in the Northern English by the 14th century. Welsh examples of the type: Mae Ieuan a Fair yn mynd (lit. IS John and Mary a-go[ing])24 can thus be translated as either "John and Mary goes" (or "John and Mary is going"). However, when the conjugated form of BOD (i.e. BE) is used we have: Maen25 nhw'n mynd "They go" (or They are a-go[ing]). Eric Hamp (1975: 73), who cites the example "Horses runs" vs "They run" in northern English, shares Klemola s analysis: "This looks for all the world like an independent witness from Cumbrian or Strathclyde substratum syntax."

33Nevertheless, this neat picture could be blurred by taking into consideration the (formerly) cleft structures discussed above, where such distinctions disappear:

Modern Breton:

(Is) An den a wel…

(It is) The man (who) sees…

(Is) Int a wel…

(It is) They (who) sees…

Middle Welsh:

(Ys) Dyn awyl…

(It is) A man (who) sees…

(Ys) Wynt a wyl…

(It is) They (who) sees…

34In both pairs of overt SVO sentences (i.e. with copula loss), the verb forms are invariable. In other words, if Northern subject rule is indeed a calque, it would have to have been based on the Brittonic expanded form (i.e. progressive: BOD(BE) + yn + VN), the use of which became generalized in Welsh to encompass habitual meaning, but not in other Celtic languages (cf. Breton and Irish) which have maintained these aspectual distinctions to this day. Although we have not yet examined the question in detail, given the fact that both overt SVO and VSO structures co-existed in the 12th century Brittonic languages, we would be inclined to think that, if Northern subject rule truly stems from Brittonic substratal influence, the picture here is only a partial one and that, in various regions of Britain, we might find other combinations which take into account both types of Brittonic constructions: My mother and father is from Merthyr (George 1991); Shakespeare's Fluellen: The French is gone off Leeks is good; we is and they is are also attested in Mid Glamorgan and in Gwent (Parry 1977, 1979).

Habitual Aspect: DO and BE

35One of the oft mentioned characteristics of DO is its ability to convey habitual aspect in southwestern English, a feature also possible in Breton and Middle Welsh. Ihalainen (1991) has shown for east Somerset, for example, that DO has a clear habitual function, particularly in the past tense. Gachelin (1990: 238), who has also studied this feature adds: "The permanence of habitual do in southwest England may represent an older Celtic substratum." Tristram (1997) and German (1996) show that DO is also used to mark habitual aspect in southern Welsh English. A Breton structure such as Glao a ra: (rain that [it] do[es]), presumably from *is glao a ra, meaning either "it's raining" or "it rains (everyday)", recalls southwestern dialectal English usage where one hears- It do rain-meaning either "it rains (everyday)" or "it is raining". Gachelin (1997: 38) gives Do rain, don't it which here conveys progressive aspect as was also possible in Early Modern English.

  • 26 Habitual DO BE and BE generalized to all persons may have entered Black American English via the s (...)
  • 27 "He/she bes"is another example of an attempt to translate habitual BE. It is found in both Ireland (...)

36Poussa (1990: 421), furthermore, believes periphrastic DO + BE expressing habitual aspect (Cornwall: The childer do be laffen at me; Oxfordshire: She do be so strict with us gals), normally associated with Hiberno-English and Black American English, may have originally been introduced into southern Ireland by southwestern and western English colonists in the 17th century.26 If true, the habitual aspect of Hiberno-English DO BE could result from multicausation: an attempt to translate Irish Goidelic consuetudinal BE27 as well as 17th century southwestern English vernacular input where it may originally have translated habitual Brittonic B-forms. Moreover, if it is indeed a calque on Celtic, the source would have to be Brittonic since Goidelic does not possess such a construction (NP + DO + INF). This hypothesis would seem to justify Harris' (1991: 208) warning to Irish substratists that archaic English dialects (i.e. from the West and South-West) certainly played a major reinforcing role in maintaining at least some of the Goidelic features which typify Hiberno-English. If the Brittonic substratum is not a factor, one is forced to ask why the link exists between Irish Goidelic, Hiberno-English and Southwestern English in the first place.

37This brings us to the similarity between the Brittonic and Old English use of the verb BE just alluded to, a feature which has intrigued linguists for some time. Crépin (1994: 87) writes: "Le présent vieil-anglais en b- s'emploie pour le futur ou l'atemporel; l'influence du quasi-homonyme gallois bydd a pu renforcer cet emploi." Tolkien (1963: 32) echoes this position:

Now this system is peculiar to Old English. It is not found in any other Germanic language, not even in those most closely related to English.

  • 28 Perhaps because the habitual and future aspectual forms were identical, Breton later turned to the (...)

38He suggests that the Old English system results from a Brittonic substratum and shows that in Old Northumbrian the third person plural was bydun (Welsh byddant) instead of beoth. He writes: "Now this must have been an innovation developed on British soil" (p. 32). Habitual and future B-forms also exist in Cornish, Breton28 as well as in the Goidelic languages and must have been a part and parcel of the Brittonic verbal system during the early period of English colonisation. The morphological similarity between surviving Middle English B-forms and the Brittonic consuetudinal forms is indeed noteworthy:

  • 29 Denison (1993) notes that the preverbal particle was more often "in" in Middle English than "on" ( (...)

39At another level, the existence of the expanded form (IN/AT/ON > atonic preverbal particle A- + BE + VN) in both the Brittonic languages and in English is also striking. The following Cornouaillais Breton example, me a ve(z) atao o rei bara da ‘lapoused (lit. I BE [hab.] always a-giv[ing] bread to the birds) can be contrasted with me zo atao o rei bara da ‘lapoused (lit. I IS always a-giv[ing] bread to the birds) which can be used to express annoyance, just as in English (i.e. it's up to someone else to feed them). The similarity in the semantic interpretation of habitual BE (Breton BE > VE) versus AM/IS/ARE (Breton ZO), the use of the progressive (involving preverbal particle, i.e. a-prefixing in English)29 is close in many respects to what is found in Middle English (Preusler 1956) and certain modern varieties of British, Irish and North American English (Appalachian English and Black American English).

  • 30 Cf Stephen Hewitt's article on the subject (1986) "Le progressif en breton à la lumière du progres (...)

40Although we cannot discuss the many remarkable parallels between Breton and English progressive here30 Gachelin (1990: 233) expresses the possibility of a Brittonic source for the expanded form as follows:

Ever since the English settled in the British Isles, they have been in contact with Celtic speakers, and EF (expanded form) distinguishes English among Germanic languages. Inside Europe, only Celtic gives as much prominence to similar locative structures with verbal nouns.

Conclusion

  • 31 Cf D. Parry (1979: 146): That's the chap that his uncle was drowned (Cardiganshire); also cf. Muri (...)
  • 32 De la Cruz (1972) contends this construction does not exist in Celtic.

41Of course, this brief sketch probably raises more questions than it answers. To be certain, many interesting parallels have had to be left aside which are linked to the kind of criteria associated with the shift towards analycity: preposition stranding: Aze ‘ma ‘n daol meus lakaet ma leor warnezi (lit. There's the table I put my book onher); certain relatival constructions common in the non-standard English of Scotland, Cumbria, Isle of Man, Dorset, Wales: Hennez eo an den a gan e vab ba ‘n iliz (He's the man that his son sings in the church31 phrasal verbs: dond ‘mêz (come out, be published); bann en traon (throw down; get rid of some one); digas gand (bring back); the fronting of datives to subject position (Yann a oa tennet warnan: John was shot at)32 and so on.

42Taken individually, of course, each of these linguistic parallels can be (and frequently are) brushed aside as typological universals, which they indeed may be. Nevertheless, if one takes the entire picture into account, both extra-linguistic and linguistic, especially the volume and the typological nature of the morphosyntactic parallels involved, we believe there is a reasonable argument in favour of a systematic review of the evidence in detail. Considering the significance this could have on our understanding of the origins of English, it might be well worth the effort.

Bibliographie

Bibliography

Alcock, L. [1971] 1983. Arthur's Britain, History and Archaeology AD 367-634, Harmondsworth: Penguin Books.

Algeo, J. & T. Pyle [1964] 1993. The Origins and Development of the English Language, New York: Harcourt, Brace Jovanovich.

Bailyn, B., 1986. Voyagers to the West, New York: Alfred Knopf Inc.

Blake, N., 1996. A History of the English Language, London: MacMillan.

Burchfield, J., 1986. The English Language, Oxford: Oxford University Press.
Chadwick, N., 1963. "The British or Celtic Part in the Population of England", in: Chadwick, N., (ed.), Angles and Britons. O'Donnell Lectures. Cardiff: University of Wales Press, pp. 111-147.

Crépin, Α., 1994. Deux mille en de langue anglaise, Paris: Nathan.

Davies, J., 1990. A History of Wales, London: Penguin Books.

De la Cruz, J.M., 1972. "A Syntactic Complex of Isoglosses in the North-Western End of Europe" (English, Germanic and Celtic)". Indogermanishe Forschungen 77: pp. 171-180.

Denison, D., 1993. English Historical Syntax, London: Longman.

Evans, D.S., 1964. A Grammar of Middle Welsh. Dublin: Dublin Institute for Advanced Studies.

Fleuriot, F., 1980. Les Origines de la Bretagne, Paris: Payot.

—, 1985. A Dictionary of Old Breton, Toronto: Prepcorp Ltd.

Freeman, Ε.Α., 1867. The History of the Norman Conquest, Its Causes and Failures, Vol. 1, Oxford: Clarendon Press.

Gachelin, J.M., 1990. "Aspects of Non-Standard English". L'Auxiliaire en Question (Travaux Linguistiques du Cerlico II) Rennes: Presses Universitaires de Rennes II, pp. 221-240.

George, C., 1990. Community and Coal: an Investigation of the English Dialect of the Rhondda Valleys, Mid Glamorgan, unpublished Ph.D. thesis, University College of Wales, Swansea.

German, G., 1984. Etude linguistique du Breton de St. Yvi, unpublished Ph.D. thesis, University of Brest.

—, 1996. Etude sociolinguistique de l'anglais du Pays de Galles, unpublished Ph.D thesis, Université du Littoral, Boulogne-sur-Mer.

—, 2000. "Britons, Anglo-Saxons and Scholars: 19th century Attitudes towards the Survival of Britons in Anglo-Saxon England", in: Tristram, H. ed. CelticEnglishes II. Heidelberg: Carl Winter.

Gumperz, J. & R. Wilson, 1971. "Convergence and Creolization: A Case from the Indo-Aryan/Dravidian Border of India", in: Hymes, D. ed., Pidginization and Creolization of Languages. Proceedings of a Conference held at the University of the West Indies, Jamaica, April 1968, Cambridge: Cambridge University Press, pp. 151-167.

Hamp, E.P., 1975-1976. "Miscellanea Celtica I, II, II, IV" SC10-11: 54-73.

Hardie, D. A A, 1948. Handbook of Modern Breton (Armorican), Cardiff: University of Wales Press.

Härke, H., 1995. "Immigrants and Natives: A Provisional Model of Anglo-Saxon Ethnogenesis" unpublished typescript: 1-18.

—, (forthcoming 2001) "Population Replacement or Acculturation? An Archaeological Perspective on Population and Migration in Post-Roman Britain" in: Tristram, H. ed. Celtic-Englishes II. Heidelberg: Carl Winter.

Hemon, R., 1975. A Historical Morphology and Syntax of Breton, Dublin: Dublin Institute for Advanced Studies.

Hewitt, S., 1986. "Le progressif en breton à la lumière de du progressif anglais", in: Le Dû, J & Y. Le Berre, eds. La Bretagne Linguistique. Vol.2, Brest: CRBC, pp. 133-147.

Higham, N., 1992. Rome, Britain and the Anglo-Saxons, London: Seaby.

Ihalainen, O., 1991. "Periphrastic Do in Affirmative Sentences in the Dialect of East Somerset", in: Trudgill, P. & Chambers, J, eds., Dialects of English, London & New York: Longman, pp. 148-160.

Jackson, K., 1953. Language and History in Early Britain. Edinburgh: Edinburgh University Press.

Jespersen, O., [1938] 1982. Growth and Structure of the English Language, Chicago: University of Chicago.

Klemola, J., 2000. "The Origins of Northern Subject Rule: a Case of Early Contact?", in: Tristram, H., ed., Celtic-Englishes II, Heidelberg: Carl Winter, pp. 329-346.

Laing, M. & J. Laing, 1990. Celtic Britain and Ireland, AD 200-800: The Myth of the Dark Ages, Dublin: Irish Academic Press.

Lambert, P.Y., 1994. La langue gauloise, Paris: Editions Errance.

Mossé, F., [1947] 1958. Esquisse d'une histoire de la langue anglaise, Lyon & Paris: IAC.

Paddock, H., 1991. "The Actuation Problem for Gender Change in Wessex versus Newfoundland", in: Trudgill, P. & Chambers, J., eds., Dialects of English, London & New York: Longman, pp. 29-48.

Parry, D., 1977. Survey of Anglo-Welsh Dialects: The South-East, Swansea: University of Swansea.

—, 1979. Survey of Anglo-Welsh Dialects: The South-West, Swansea: University of Swansea.

Poppe, E. & I. Mittendorf, 2000. "Celtic Contacts of the English Progressive", in: Tristram, H.ed. Celtic-Englishes II. Heidelberg: Carl Winter, pp. 177-145.

Poussa, P., 1990. "A Contact-Universal Origin for Periphrastic 'do' with special Consideration of Old English-Celtic Contact, in: Adamson et al. Papers from the 5th International Conferences on English Historical Linguistics, CILT 65, Amsterdam: Benjamins, pp. 407-434.

Preusler, W., 1956. "Keltischer Einfluss im Englischen" RVL 22: 322-350.

Shorrocks, G., 1999. A Grammar of the Dialect of the Bolton Area, Part II, Morphology and Syntax, Frankfurt am Main: Peter Lang.

Stubbs, W., 1870. The Constitutional History of England, Oxford: Clarendon Press.

Tesnières, L., [1959] 1988. Eléments de Syntaxes Structurale, Paris: Editions Klinksieck.

Thomason, T.G. & T. Kaufman, [1988] 1991. Language Contact, Creolization, and Genetic Linguistics, Berkeley, Los Angeles & London: University of California.

Thurneysen, R., [1946] 1975, A Grammar of Old Irish, Dublin: Dublin Institute for Advanced Studies.

Tolkien, J., 1963. "English and Welsh", in: Chadwick, N., ed. Angles and Britons, O'Donnel Lectures, Cardiff: University of Wales Press, pp. 1-41.

Tristram, H., 1999. "How Celtic is Standard English?" Saint Petersburg: Nauka, 45 p.

—, 1995. "Aspect in Contact ", in: W. Riele (ed.), Anglistentag 1994 Graz, Tübingen: Niemeyer, pp. 269-294.

—, 1997. "DO in Contact?" in: H. Ramisch & K. Wynne (eds.), Language Time and Space. Festschrift für Wolfgang Viereck, Stuttgart: Franz Steiner, pp. 401-417.

Van der Auwera, J. & I. Genee, 1998. "Periphrastic-DO: On the Convergence of Languages and Linguists". Paper given at the 10. Wuppertaler Linguistisches Kolloquium, 13-14 November 1998 – Sprachkontact.

Notes

1 I wish to thank Dr. Hildegard Tristram (University of Freiburg) and Dr. Jean Le Dû (University of Western Brittany, Brest) for having read an earlier draft of this paper and for their very helpful observations. I also wish to thank Professor Bourquin, who has long been convinced of the existence of a Brittonic substratum in English, for his encouragement. Any errors herein are, of course, my own.

2 Traditionally, the Celtic languages have been divided into two branches: P-Celtic, of which Brittonic (or Brythonic) and Gaulish are primary examples, and Q-Celtic or Goidelic. The former have Breton, Welsh and Cornish as descendants and the latter, Irish, Scottish Gaelic and Manx.

3 This point was used as proof to argue that the English were, as a consequence, "purely Teutonic" (Stubbs ibid.).

4 Chadwick (1963) convincingly argued that these sources had simply been misunderstood.

5 Härke will be showing evidence from some initial DNA data which would support this theory (Potsdam 2001).

6 One of the most conservative of these estimates is by Härke (1995): 3 to 1 in the South and Midlands and 5 to 1 in the North of England.

7 In his monumental History of Wales Davies (1993: 68) writes: "The greater part of the pre-English inhabitants of England survived in their localities and a substantial portion of the present-day population of England may be accounted among their descendants… it is therefore unlikely that there is any racial distinction of substance between the English and Welsh."

8 This point is not necessarily in contradiction with the hypothesis that Breton probably contains a significant Gaulish substratum. Interestingly, Welsh is often typologically different from Breton, the latter being closer to English in many respects. This may be due to the fact that Irish was spoken in Gwynedd and Dyfed for centuries. For this reason, significant Irish adstratal influence on Welsh cannot be ruled out.

9 Recent research would suggest that the colonizers of Brittany were primarily southwestern Britons (Dorset, Somerset, Devon and Cornwall) regions that did not have significant dealings with the Anglo-Saxons until the second stage of their expansion (cf Jackson 1953). The suggestion that the Brittonic spoken by these colonists may have already been influenced by Old English before their departure therefore seems unlikely. Fleuriot (1980) also contends that since the Britons began migrating to Brittany before the arrival of the Anglo-Saxons, they could not have been fleeing them as is generally argued.

10 This could explain the existence of Brittonic sheep-counting numerals in Cambridgeshire, Lincolnshire and Cumbria until the 20th century as well as the survival of Brittonic laws (Kylch, Butreth) throughout the West and North of England until the 12th and 13th centuries. In Lancashire, Brittonic festivals (Beltan) and personal names (Madoc, Einion, Angharad) lingered on until this same period (Rees 1963).

11 Cognate with Welsh "Cymru" (< *cum-brogi "the fellow countrymen").

12 When one considers that the Welsh aristocracy was largely anglicised by the 16th century and that the Breton aristocracy turned to French about 800 years ago, this hypothesis is not at all extreme.

13 If Breton French and Welsh English are anything to go by, the native vocabulary borrowed into English probably disappeared after the shift was complete.

14 He argues that less than 5% of the population in England ever spoke French at the peak of French influence. The move toward analytic structure thus had to have been well under way.

15 The loss of case endings is what led to the mutation system in both the Brittonic and Goidelic languages.

16 In this writer's view, this is not necessarily a sign of "language death" as some have maintained.

17 Crépin (1994: 75) demonstrates that the definite article the begins to take over the functions of other nominal singular forms such as se and seo… and is attested in Anglian dialect as early as the 10th century. Once again, the simplification of the definite article had already occurred in Brittonic (cf Middle Cornish and Middle Breton) > an.

18 Cf Preusler 1956, Teynière 1959, Poussa 1990, Gachelin 1990, German 1996, Tristram 1997, van der Auwera 1998.

19 Komz a raent (speak[ing] they did).

20 One wonders whether it clefting in Breton French recalls the older Brittonic structure. In other words, could native Breton speakers still sense the semantic presence of the old unstressed copula, even though it is not lealized phonetically? Concerning the possible Gaulish origin of French cleft constructions see footnote 22 below.

21 A native Breton speaker from Landunvez, Leon (PC), claims that his father says "me a ra drebi" with emphatic meaning - I am eating (i.e. don't bother me!). We have never heard this elsewhere and prefer to verify this information. What is interesting, however, is that the speaker does not find the structure ungrammatical.

22 Lambert mentions that French cleft constructions (ex. C'est Jean qui parle) could be due to Gaulish influence but signals that there is not yet any concrete proof of this.

23 This is partly due to the loss of southern 3rd pers. sing. //-th//: he loveth > he love.

24 Breton offers a parallel structure here: Ema Yann ha Mari o vond (IS [situation] John and Mary ago[ing]: "John and Mary are going") vs Emaint o vond (lit. Are-they a-go(ing)) but with progressive aspect.

25 Maen consists of two morphemes: //mae// "is" + //n// "they", the latter being the third person plural marker (<int) "they" (Middle Welsh ymaent) as in Modern Breton: emaint < //ema// + //int//. In Cornouaillais Breton one also has Ema hent (=int) where the third person pronoun is unbound.

26 Habitual DO BE and BE generalized to all persons may have entered Black American English via the same route. The first Africans arrived in Virginia in the early 17th century at a time when the majority of the English were from the West Country (Bailyn 1986).

27 "He/she bes"is another example of an attempt to translate habitual BE. It is found in both Ireland and Newfoundland (cf. Paddock 1991).

28 Perhaps because the habitual and future aspectual forms were identical, Breton later turned to the subjunctive to express the future.

29 Denison (1993) notes that the preverbal particle was more often "in" in Middle English than "on" (and sometimes "at" as in Hiberno-English).

30 Cf Stephen Hewitt's article on the subject (1986) "Le progressif en breton à la lumière du progressif anglais". See also Poppe and Mittendorfs (2000) "Celtic Contacts of the English Progressive".

31 Cf D. Parry (1979: 146): That's the chap that his uncle was drowned (Cardiganshire); also cf. Murison (1977: 39) The woman that her bairn's no weel (Scotland).

32 De la Cruz (1972) contends this construction does not exist in Celtic.

Table des illustrations

URL http://books.openedition.org/pufr/docannexe/image/7276/img-1.jpg
Fichier image/jpeg, 66k

Auteur

Université de Brest

© Presses universitaires François-Rabelais, 2001

Conditions d’utilisation : http://www.openedition.org/6540

Acheter