An Evaluation of the Celtic Hypothesis for Brythonic Celtic influence on Early English

The Celtic Hypothesis attributes some of the major linguistic changes in Old and Middle English to influence from the Brythonic languages that were spoken in Britain at the time of the Anglo-Saxon immigrations beginning in the fifth century. The hypothesis focuses on features of English that do not exist, or are not common, in the other Germanic languages but resemble features in the Celtic languages. From the evidence we have of the socio-political relationships between the Britons and the Anglo-Saxons, the likely language contact situations are compatible with Thomason and Kaufman’s (1988) ‘substratum interference’ and van Coetsem’s (1988) ‘imposition’, by which morpho-syntactic features are transferred from one language (L1) to another (L2) through imperfect second-language acquisition. The fact that the social situation was compatible with Brythonic influence on English does not mean, however, that the linguistic features in early English claimed by the proponents of the Celtic Hypothesis as showing Brythonic influence were actually influenced in this way. My purpose is to evaluate the Celtic Hypothesis in the light of the evidence and modern theories of language change due to contact. This thesis focuses on three features that have played a prominent role in the Celtic Hypothesis: (1) the dual paradigm of be (bēon and wesan) in Old English, (2) the periphrastic construction do + infinitive and (3) the periphrastic progressive construction be + -ing, the last two of which began to be grammaticalised in Middle English. I collect independent evidence from a selection of Middle Welsh texts of the parallel constructions: (1) the dual paradigm of bot ‘be’, (2) the periphrastic construction gwneuthur ‘do’ + verbal noun and (3) the periphrastic construction bot ‘be’ + particle + verbal noun. While the proponents of the Celtic Hypothesis provide examples of these constructions from several Brythonic languages including Middle Welsh, they give few examples and do not discuss the variability of the evidence according to date, region or genre. My own research confirms that the dual paradigms of be and bot do form a close parallel, but it also shows that the Old English dual paradigm is unlikely to have arisen due to Brythonic influence. My findings also show that evidence for the construction of gwneuthur ‘do’ + verbal noun is problematic: while it is very common in Middle Welsh prose narratives, it is very rare in the early prose annals and the earliest poems. Evidence for the progressive construction in early Welsh is similarly problematic: while it is regularly used in Colloquial Modern Welsh as bod ‘be’ + particle + verbal noun, it is by no means common in Middle Welsh. By looking at a wider range of Middle Welsh evidence, I demonstrate the limitations of the evidence relied on by proponents of the Celtic Hypothesis. This may lead to better substantiated arguments for the hypothesis in the future