Early English and the Celtic hypothesis

Hickey, Raymond 2012. ‘Early English and the Celtic hypothesis’, in: Terttu Nevalainen and Elizabeth Closs Traugott (eds) The Oxford Handbook of the History of English. Oxford: Oxford University Press, pp. 497-507. Early English and the Celtic hypothesis Raymond Hickey 1. Introduction The aim of the current chapter is to consider Celtic influence on early Old English. This issue has recently been debated in detail, e.g. in Trudgill (2010: 1–35) and Ahlqvist (2010), and it helps to throw light on general contact arguments in language change. The scholarly opinion that Brythonic, the language spoken in England by the Celtic population at the time of the Germanic invasions, had a significant effect on the development of English is known as the “Celtic hypothesis” (Filppula and Klemola 2009). The standard wisdom on contact and transfer has traditionally been that the language with more status influences that with less, i.e. borrowing is from the superstrate into the substrate, as is attested by Latin and French borrowings into English. This is, however, a simplistic view of possible influence in a contact scenario. Vocabulary, as an open class with a high degree of awareness by speakers, is the primary source of borrowing from the superstrate. Again French and Latin are standard examples. Histories of English have not in general concerned themselves with features of early English with a possible Celtic source: for many older scholars there was no Celtic issue.1 Later works sometimes contain a certain awareness of the issue and the scholarship surrounding it, e.g. Hogg and Denison (2006: 8f.) have some references to the Celtic-English interface but no discussion of features; the contributions in Mugglestone (2006) do not discuss Celtic-English contact either. Supporters of the Celtic hypothesis criticise the view taken in many textbooks that because there are only a few loanwords from Celtic there was no other influence (Filppula, Klemola, and Paulasto 2008: 25). However, if contact persists over a longer period and forms the language learning environment for many generations, then the substrate can have a gradual and imperceptible influence on the superstrate, often leading to grammatical change. This scenario may well have been the source of syntactic features in English which the latter has in common with Celtic (Poussa 1990; Hickey 1995; Vennemann 2002). It is especially likely if a section of the population shifted language and transferred features from their outset language in the process (see Hickey on contact, this volume). However, language shift is not the only source of contact features. If speakers of two languages live in close proximity then child language learners in one group can pick up features from the other group, irrespective of which group is superstrate relative to the other. This type of infection through contact applies to speech habits, 1 The general view that the Celts had made no substantial contribution to the history of English prevailed from Henry Sweet, Walter Skeat, Karl Luick, and Otto Jespersen onwards. Early English and the Celtic hypothesis 2 of 10 such as phonetic realisations, prosodic patterns, alternative exponents of identical grammatical categories alongside the use of lexical items from the other group. In the fifth and sixth centuries in England the Germanic invaders probably shared the same environment with the Celts who they had subjugated. Several facts point to this. No records speak of a battle in which the Germanic invaders were victorious over the Celts. The notion that they banished the latter into highland areas in the north and west of England and down to the remoter parts of the south-west is an assumption based on the later distribution of P-Celtic languages – Welsh and Cornish (descendants of Brythonic) – in England. In addition, Old English wealh meant ‘foreigner’ but also ‘Celt’. The word came to be used for ‘servant, slave’ (cf. wielen/wiln ‘female servant, slave’ with the same root) which apparently indicates the status of the Celts vis à vis the Germanic settlers. Lutz (2009: 239–40), drawing on work by the historian David Pelteret, emphasises that the meaning of ‘servant, slave’ was predominantly used in West Saxon, although there were other words with a similar meaning. This would suggest that the subjugation of the Celts was most marked in the south of England where the concentration of the Germanic settlers was greatest. Furthermore, the view that the Celts, left in the south and east of England, would have had to shift to English is subscribed to by scholars investigating the Celtic hypothesis (Lutz 2009: 228). Earlier scholars of English highlighted this fact, e.g. Chadwick (1963), but their work was regarded as peripheral and was not given consideration in “mainstream” works such as Campbell (1959) or Mitchell and Robinson (2007). That the majority population of sixth- and seventh-century West Saxony consisted of Celts who had shifted to English is evident from the numerical relationship of the Celts to the Germanic tribes in the early Old English period. Estimates vary here: the number of Germanic settlers during the fifth century has been put at anything between 10,000 and 200,000 (Filppula, Klemola, and Paulasto 2008: 15). But given a population of Britain just before the Germanic arrivals of approximately one million then the relationship would have been anything from 1:100 to 1:5. The latter ratio is hardly likely as it would have implied a huge movement from the continent to England. A figure somewhere in the middle, say 1:20, would still imply that the Celts greatly outnumbered the Germanic settlers, assumed to be about 50,000 with this ratio. In the generations following the initial Germanic settlements, most Celts in contact with the new settlers would have given up their native Brythonic, speaking shift varieties of the newcomers’ dialects during the transition. It should furthermore be mentioned that many higher-position Celts were also speakers of British Latin (Schrijver 2006). Even approximate numbers of this section of the British population are impossible to ascertain but the trilingualism which they help to form in early England added another dimension to the contact situation which cannot, for reasons of space, be discussed in the present chapter. In section 2 features are discussed which illustrate possible influence from Celtic on English. The arguments for transfer are presented with a summary in the conclusion; detailed treatments of these can be found in Filppula, Klemola, and Paulasto (2008: 30–123). Early English and the Celtic hypothesis 3 of 10 2. Features illustrating possible Celtic influence 2.1 The rise of the internal possessor construction English is the only Germanic language which categorically uses possessive pronouns with instances of inalienable possession, e.g. My head is sore (contrast with German Mir tut der Kopf weh, lit. ‘me-DATIVE does the head hurt’). Structures like *The head is sore to me are not possible in present-day English.2 In Old English the dative of a personal pronoun was found with the head of the noun phrase accompanied by a determiner.3 In the Old English Poem of Judith one can see this use: (1) þæt him þæt heafod wand forð on ða flore lit.: that him-DATIVE that head [...] ‘that his head rolled forth on the floor’ Such structures are labelled “external possessor” constructions because possession is expressed by an oblique case pronoun which is outside the semantically related noun phrase. Nowadays the possessive element is internal to the noun phrase – a determiner modifying the head noun – hence the expression “internal possessor construction”. Why did English abandon the external possessor construction? Such constructions are favoured in most European languages and usually involve a dative-like case for the possessor (in function and commonly in form as well, Haspelmath 1999: 110f.). This is an areal feature in Europe (Haspelmath 1999: 116f.) for, while most languages with this feature are Indo-European, some of these are from outside this family, such as Basque, Hungarian and Maltese. Remarkably, English, Welsh and Cornish only have NP-internal possessors. Put in areal terms, one has internal possessor constructions in the British Isles and external possessor constructions in continental Europe. Only in the extreme southeast are internal possessor constructions found again, in Turkish and Lezgian, a Caucasian language. A contact explanation would assume that the use of possessive pronouns in instances like (1) diffused from language shift varieties of the Germanic dialects into inherited varieties and became established there. This transfer did not affect the existence of a possessor construction, but it changed the exponence of the category, a frequent effect of language contact, especially in language shift situations. 2.2 The twofold paradigm of to be Contact explanations are worth considering with regard to the twofold paradigm of to be in Old English, specifically West Saxon. There is one paradigm, beginning in 2 With prepositional objects the definite article can, however, be used, e.g. She hit him in the face. He was struck on the head. 3 On the dative Mitchell and Robinson (2007: 106) say: “It may express possession, e.g. him on heafod ‘on his head’”. Early English and the Celtic hypothesis 4 of 10 a vowel in the singular and s- in the plural, and a further paradigm with forms in b-; both paradigms are inherited from Indo-European. Table 1. West Saxon present-tense forms of to be (after Campbell 1959: 349) Existential present Singular 1. eom, 2. eart, 3. is Plural sindon, sint Habitual present4 Singular 1. bēo, 2. bist, 3. biþ Plural bēoþ However, West Germanic languages, apart from English, have combined the two paradigms to one, cf. German ich bin, du bist, er ist, wir sind, etc. The West Saxon double paradigm is thus remarkable in the Germanic context, but not when considering Brythonic with which it co-existed (Tolkien 1963). Here one also finds two paradigms which have a similar syntax and semantics.5 Table 2. Middle Welsh present-tense forms of to be (Evans 1976: 136)6 Existential present tense Singular 1. wyf, oef 2. wyt 3. yw, (y) mae, (y) taw, oes Habitual present tense Singular 1. bydaf 2. bydy 3. byd, bit Plural ym ywch ynt, (y) maent, y maen Plural bydwn bydwch bydawnt, bwyant, bint The Old English double paradigm is attested from the eighth century which, as Lutz (2009: 233) rightly highlights, was almost three centuries after the coming of the Germanic settlers. This suggests that both be paradigms were entrenched in Old English from earlier transfer, probably by Celtic speakers shifting to the language of the invaders.7 The presence of a distinction in the outset language triggers its appearance in the target language due to the search for categorial equivalence by the 4 The b-forms were used for the future, but also had durative and iterative uses (Lutz 2009: 233) which are subsumed here under the heading “habitual”. 5 Evans (1976) has a detailed section on the syntax of bot (habitual) and ys (existential) (see 1976: 136–45). 6 The Middle Welsh forms stem from the period 1100–1400 but are derived from older forms which existed in Brythonic during contact with the Germanic settlers. 7 This is a suggestion already put forward by Wolfgang Keller (1925) but not picked up by later scholars. Early English and the Celtic hypothesis 5 of 10 shifting population (a common phenomenon in other contact scenarios; see Hickey 2007: Chapter 4 for details). 2.3 The development of the progressive There are basically three views on the development of the progressive in English: (i) it developed independently (Visser 1963–73), (ii) it arose under Latin influence, perhaps via French (Mossé 1938), and (iii) it resulted from contact with Celtic (Keller 1925; Dal 1952; Preußler 1956; Wagner 1959; Braaten 1967). A type of progressive structure in which a gerund was governed by a preposition existed sporadically in Old English: ic wæs on huntunge ‘I was hunting’ (Braten 1967: 173). This also occurs in vernacular German, with an infinitive, as in Ich bin am Schreiben, lit. ‘I am at write-INFINITIVE’, ‘I am writing’, and in Dutch Hij was aan het schrijven, lit. ‘he was on the write-INFINITIVE’ (Olga Fischer, personal communication). In this context a further, typological consideration is necessary (Mittendorf and Poppe 2000: 120–22). Progressive aspect is frequently expressed – in many unrelated languages – via a locative structure meaning to be ‘at’ or ‘in’ an activity. Furthermore, the step from structures like ic wæs on huntunge to I was hunting is small, involving only preposition deletion and the shift of gerund to non-finite verb form. The fully developed progressive appears in Middle English, but the apparent time delay between contact with Brythonic and the later progressive can be attributed to the strong written tradition in Old English (Dal 1952: 113). The progressive is found in all Celtic languages and is represented by a Modern Welsh sentence like Mae John yn torri coed, lit. ‘is John in cut-VERBAL NOUN wood’, ‘John is cutting wood’ (Jones and Thomas 1977: 63). This shows a prepositional expression for the progressive aspect and is structurally parallel to Old English ic wæs on huntunge. In summary one can say that in both Old English and Brythonic the semantic category of progressive existed. Both languages maintained this category; English lost the locative preposition, increasing the syntactic flexibility and range of the structure, perhaps under the supportive influence of Celtic contact. Another pertinent point is that with habitual and progressive verb forms both English and Brythonic had aspect structures of a process-oriented nature (an areal feature in Britain and Ireland) in contrast to the goal-oriented nature of aspect types in other Germanic languages, such as German, which are largely telic in nature. 2.4 The rise of periphrastic do The syntax and semantics of the verb do is one of the most researched matters in the history of English and there are several opinions concerning why its development took the course it did (Garrett 1998; Klemola 2002). The consensus among the different views is that English developed a causative use of do involving a direct object followed by an infinitive and that this structure was not inherited from earlier forms of (West) Germanic.8 8 Van der Auwera and Genee (2002: 292) show clearly that English sides with Celtic in the frequency of do periphrasis and less with the other West Germanic languages. Early English and the Celtic hypothesis 6 of 10 (2) Þe biscop of Wincestre … did heom cumen þider. ‘The bishop of Winchester … had them come thither.’ In the course of the Middle English period variants of this structure developed which had no direct object. There is some dispute about whether the object-less structure was causative or not (van der Auwera and Genee 2002: 293) but agreement about the resulting periphrastic construction involving do + lexical verb, a developmental path found in German and Dutch as well as in French.9 Contact was already appealed to by those supporting the above development, e.g. Ellegård (1953) and Denison (1993), but it is contact with Latin and possibly with French faire. For other accounts geography plays a role. While causative uses of do are common in eastern texts, periphrastic uses are first observed in western texts. In the east (and south-east) of England French survived longest and in the west (and south-west) Celtic was spoke most intensively (Welsh and Cornish, both from Brythonic). Do periphrasis is a common feature of Welsh, Cornish and Irish (see Evans 1976: 130–31 for the Middle Welsh forms). It may well have been established in early insular Celtic, i.e. before the advent of the Germanic settlers. But the textual record is not sufficient here to determine this clearly. Hence it is very difficult to provide a cast-iron case for a Celtic origin of English periphrastic do. One line of argument is that contact situations per se tend to give rise to auxiliaries and so contact-intensive areas of England (west/south-west) would be where periphrastic do would be expected to surface (Poussa 1990: 412f.; Tristram 1997: 415). Furthermore, do can be used with nouns and so does not require that speakers know sets of corresponding lexical verbs. In a language contact situation, consisting in its early stages of adults shifting from Celtic, this device would improve communication despite being inflectionally less complex. Furthermore, this use of do + noun as equivalent to a lexical verb is an established feature of Celtic, e.g. Irish Rinne sí iarracht é a sheachaint, lit. ‘did she try-NOUN him to avoid’. In addition, using do for emphasis may well have been part of the pragmatic mode of adult Celtic second language learners. With the removal of stress from do the periphrastic use would remain. The Celtic-English contact situation involves the issue of directionality. It is not certain that Celtic first had periphrastic do and thus supplied the model for English. So the question of which direction the transfer went is open. One can get around this issue by assuming mutual influence (Tristram 1997) and to view the contact as an areal phenomenon with bidirectional transfer. There are other structures in Celtic and English which are now areal features, e.g. the absence of external possessor constructions, the progressive, habitual aspect (in non-standard varieties of English). However, the use of do as a type of auxiliary is found in other West Germanic languages so that there could well have been language-internal input to the later English situation. Bearing all such factors in mind can help avoid monocausal contact assumptions such as that by McWhorter who has recently suggested that periphrastic do came from Cornish (McWhorter 2009: 168). This single-view approach ignores the frequent cases in West Germanic languages where do is a 9 Periphrastic uses of do are recorded for various dialects of German, for Low German and for northern dialects of Dutch so that the use is not exclusively found in English (van der Auwera and Genee 2002: 286–88). Early English and the Celtic hypothesis 7 of 10 tense carrier used with a lexical verb in non-finite form, e.g. in West German and Low German dialects (van der Auwera and Genee 2002: 286f.) as in Ich tue dir das morgen bringen, lit. ‘I do you-DATIVE that tomorrow bring-INF’. The inherent semantics of do as a verb denoting direct action makes it likely to be employed as tense carrier or aspect marker. 2.5 The Northern Subject Rule The “Northern Subject Rule” is basically a co-occurrence pattern between verbs and preceding subjects: -s is absent when the verb is immediately preceded by a pronoun, but not otherwise.10 An illustration would be We meet and talks together. This agreement pattern is well-attested in northern Middle English and Middle Scots. In middle and southern England the distribution is uneven (Klemola 2000: 336). Klemola (2000) has considered the possibility of early Celtic influence in the far north, leading to the cross-linguistically rare distribution of verbal -s just outlined Klemola (2000: 337). However, in Welsh there is an agreement rule whereby plural forms of the present are only used with verbs when the pronoun nhw ‘they’ follows immediately. In all other cases the singular is found. In essence this is the Northern Subject Rule: the plural forms of Welsh (maen ‘is’-PL) correspond to the s-less forms of English, the Welsh verbal singular (mae ‘is’-SG) is the equivalent to the s-full forms: Maen nhw’n dysgu Cymraeg, lit. ‘are they on learning Welsh’; Mae Trevor a Sian yn dysgu Cymraeg, lit. ‘is Trevor and Sian on learning Welsh’. Klemola stresses that the Northern Subject Rule is most widespread in regions with a bilingual Brythonic and Anglian population in the Old English period. However, non-contact explanations are also possible: the decline in inflectional morphology meant that children in the Middle English period had difficulties with the remaining inflections and evolved new systemic interpretations, one of which demanded inflectional -s except with an immediately preceding pronoun, the classic Northern Subject Rule. 2.6 Dental fricatives in the history of English A central part of the Germanic Sound Shift is the change of a strongly aspirated /th/ into a dental fricative /θ/, e.g. thin /θın/ from an earlier *thin, in stressed onsets (not preceded by /s/). The individual Germanic languages all lost this fricative later, except Gothic, Icelandic and English. Gothic did not survive long enough to be relevant here. Icelandic has changed little over time, so inertia is the major force maintaining dental fricatives there. English, however, has experienced great phonological change over the centuries, e.g. it has lost consonantal length, has acquired phonemic voiced fricatives and has developed contrastive word stress under Romance influence. So why does a language with so much phonological change still show dental fricatives? Especially given that these are relatively rare 10 The likelihood of verbal -s occurring may vary between dialects, e.g. some have this ending with third person subjects, singular and plural (as in Irish English) while others show a greater range of forms with verbal -s. Early English and the Celtic hypothesis 8 of 10 cross-linguistically: the friction of dental fricatives is much less prominent than with /s/, for instance. A contributory factor could be the existence of dental fricatives in Brythonic (still found in Welsh). Assuming that much of the Old English population consisted of Celts who shifted from Brythonic, dental fricatives would have been natural to them. This helps explain why Welsh, English and Scots still have dental fricatives. Contact may also have contributed to the unrounding of English front vowels. The Celts who shifted to English would not have had front rounded vowels from Brythonic and would in all likelihood not have used rounded realisations of Old English rounded vowels, much as say Slavic speakers of German today have /i, e/ for /y, ø/. 3. Summary and conclusion Table 3. Possible transfer features from Celtic (Brythonic) to English11 1. Lack of external possessor construction 2. Twofold paradigm of to be 3. Progressive verb forms 4. Periphrastic do 5. Northern subject rule 6. Retention of dental fricatives 7. Unrounding of front vowels Morphosyntactic Phonological These features vary in the extent and the timing of their documentation and listing them in a single table does not imply that they have an equal status as transfer features. Furthermore, for features 3 to 5 there are cogent internal arguments for their development so that contact can have provided support for their development but can hardly have been the sole origin. Paradoxically, the consideration of internal factors may strengthen the case for contact accounts as the operation of two factors in principle provides more evidence for the genesis of specific features. So while contact alone may, in an instance like the Northern Subject Rule, be viewed by scholars as too weak a source, the additional reinterpretation of decaying present-tense verbal inflections in medieval English by language learners offers, in tandem with possible transfer from Brythonic, a more plausible scenario for this feature. The Celtic hypothesis now goes back over a century but has occupied the periphery of English historical scholarship, often that practised by German and Scandinavian, including Finnish, scholars (van der Auwera and Genee 2002: 299– 302). The insights it offers can be fruitful for research within the mainstream English tradition and, combined with language-internal motivation for change, can contribute to mature and balanced scholarship without preconceptions about how such change was triggered and continued through time. 11 Other transfer features in shift varieties of Germanic dialects may simply not have made it into the textual record. Comparison with later forms of English in Ireland and Scotland suggest this (Hickey 2007: Chapter 4). Early English and the Celtic hypothesis 9 of 10 References Ahlqvist, Anders. 2010. ‘Early Celtic and English’. Australian Celtic Journal 9: 41– 71. Braaten, Björn. 1967. ‘Notes on Continuous Tenses in English’. Norsk Tidskrift for Sprogvidenskap 21: 167–80. Campbell, Alistair. 1959. Old English Grammar. Oxford: Clarendon Press. Chadwick, Nora. 1963. ‘The British or Celtic Part in the Population of England’. In Angles and Britons. The O’Donnell Lectures, ed. Henry Lewis, 111–47. Cardiff: University of Wales Press. Dal, Ingerid. 1952. ‘Zur Entstehung des Englischen Participium Praesentis auf -ing’. Norsk Tidskrift for Sprogvidenskap 16: 5–116. Denison, David. 1993. English Historical Syntax. London: Longman. Ellegård, Alvar. 1953. The Auxiliary Do. The Establishment and Regulation of Its Use in English. Gothenburg Studies in English 2. Stockholm: Almqvist and Wiksell. Evans, Simon D. 1976. A Grammar of Middle Welsh. Dublin: Dublin Institute for Advanced Studies. Filppula, Markku, Juhani Klemola, and Heli Paulasto. 2008. English and Celtic in Contact. London: Routledge. Garrett, Andrew. 1998. ‘On the Origin of Auxiliary Do’. English Language and Linguistics 2: 283–330. Haspelmath, Martin. 1999. ‘External Possession in a European Areal Perspective’. In External Possession, ed. Doris Wayne and Immanuel Barshi, 109–35. Amsterdam: Benjamins. Hickey, Raymond. 1995. ‘Early Contact and Parallels between English and Celtic’. Vienna English Working Papers 4: 87–119. ––– 2007. Irish English. History and Present-Day Forms. Cambridge: Cambridge University Press. Hogg, Richard M. and David Denison, eds. 2006. A History of the English Language. Cambridge: Cambridge University Press. Jones, Morris and Alan R. Thomas. 1977. The Welsh Language. Studies in Its Syntax and Semantics. Cardiff: University of Wales Press. Keller, Wolfgang. 1925. ‘Keltisches im englischen Verbum’. In Anglica: Untersuchugen zur englischen Philologie, Vol. 1: Sprache und Kulturgeschichte, 55–66. Leipzig: Mayer und Müller. Klemola, Juhani. 2000. ‘The Origins of the Northern Subject Rule: A Case of Early Contact?’ In The Celtic Englishes II, ed. Hildegard L. C. Tristram, 329–46. Heidelberg: Carl Winter. ––– 2002. ‘Periphrastic DO: Dialectal Distributions and Origins’. In The Celtic Roots of English. Studies in Language, Vol. 37. Markku Filppula, Juhani Klemola, and Heli Pitkänen, eds., 199–210. Joensuu: University of Joensuu, Faculty of Humanities. Lutz, Angelika. 2009. ‘Celtic Influence on Old English and West Germanic’. English Language and Linguistics 13.2 (special issue, Re-Evaluating the Celtic Hypothesis), ed. Markku Filppula and Juhani Klemola: 227–49. McWhorter, John H. 2009. ‘What Else Happened to English? A Brief for the Celtic Hypothesis’. English Language and Linguistics 13.2 (special issue, ReEvaluating the Celtic Hypothesis), ed. Markku Filppula and Juhani Klemola: 163–91. Early English and the Celtic hypothesis 10 of 10 Mitchell, Bruce and Fred Robinson. 2007. A Guide to Old English. 7th edn. Oxford: Blackwell. Mittendorf, Ingo and Erich Poppe. 2000. ‘Celtic Contacts of the English Progressive?’ In The Celtic Englishes II, ed. Hildegard L. C. Tristram, 117– 45. Heidelberg: Carl Winter. Mugglestone, Lynda, ed. 2006. The Oxford History of English. Oxford: Oxford University Press. Poussa, Patricia. 1990. ‘A Contact-Universals Origin for Periphrastic Do, with Special Consideration of OE-Celtic Contact’. In Papers from the Fifth International conference on English Historical Linguistics, ed. Sylvia Adamson, Vivien Law, Nigel Vincent, and Susan Wright, 407–34. Amsterdam: Benjamins. Preußler, Walter. 1938. ‘Keltischer Einfluß im Englischen’. Indogermanische Forschungen 56: 178–91. Schrijver, Peter. 2006. ‘What Britons Spoke around 400 AD’. In Britons in AngloSaxon England, ed. Nick J. Higham, 165–71. Woodbridge: Boydell. Tolkien, J. R. R. 1963. ‘English and Welsh’. In Angles and Britons. The O’Donnell Lectures, ed. Henry Lewis, 1–41. Cardiff: University of Wales Press. Tristram, Hildegard L. C. 1997. ‘Do-Periphrasis in Contact?’ In Language in Time and Space. Studies in Honour of Wolfgang Viereck on the Occasion of his 60th Birthday, ed. Heinrich Ramisch and Kenneth J. Wynne, 401–17. Stuttgart: Franz Steiner. Trudgill, Peter. 2010. Investigations in Sociohistorical Linguistics. Cambridge: Cambridge University Press. van der Auwera, Johan and Inge Genee. 2002. ‘English Do: On the Convergence of Languages and Linguists’. English Language and Linguistics 6: 283–307. Vennemann, Theo. 2002. ‘On the Rise of “Celtic” Syntax in Middle English’. In Middle English from Tongue to Text, ed. Peter J. and Angela M. Lucas, 203– 34. Frankfurt: Peter Lang. Visser, Federicus Th. 1963–73. An Historical Syntax of the English Language. 4 vols. Leiden: E. J. Brill. Wagner, Heinrich. 1959. Das Verbum in den Sprachen der britischen Inseln. Tübingen: Niemeyer.