Old northumbrian verbal morphology in the glosses to the lindisfarne gospels

Abstract

In considering a text such as the Lindisfarne Gospels, one is very much aware of the vast philological attention the manuscript has received since the first contribution made to its study by George Hickes in 1705. Since then, scholars of the stature of Bouterwek (1857), Skeat (1871-87), Lindelöf (1901), Holmqvist (1922), Berndt (1956) and Ross, Stanley & Brown (1960) have advanced the subject (see Ross 1937:17-25 for a detailed summary of early studies on Lindisfarne). This Latin Gospelbook written in the North of England in the early eight century constitutes a major landmark of human cultural, intellectual, spiritual and artistic achievement. While the Latin text of the Lindisfarne Gospels is a valuable early witness to St Jerome’s ‘Vulgate’, it is the carefully inserted interlinear gloss to the Latin, written in Old Northumbrian and added around the 950s960s, and the linguistic importance this gloss holds as one of the most substantial earliest surviving renderings of early northern dialect that will concern us in this study, and more concretely the distribution of verbal morphology found therein. Old and Middle English verbal morphology in the northern dialects diverged most remarkably from that of the southern dialects in two main areas. Crucially, the tenth-century Northumbrian texts bear witness to the replacement of the inherited present-indicative -ð suffixes with -s forms, and by the Middle English period, presentindicative plural verbal morphology in northern dialects was governed by a grammatical constraint commonly referred to as the Northern Subject Rule (NSR) that conditioned verbal morphology according to the type and position of the subject. The plural marker was -s unless the verb had an immediately adjacent personal pronoun subject in which case the marker was the reduced -e or the zero morpheme, giving a system whereby They play occurred in juxtaposition to The children plays, They who plays, They eat and plays. It has tacitly been assumed in the literature that the reduced forms at the crux of the NSR, and the constraint that triggers them, must have emerged in the northern dialects during the early Middle English period, as there is little indication of the pattern existing in extant Northumbrian texts from the tenth century, and by the time northern textual evidence is once again available from c.1300, the NSR is clearly prevalent (Pietsch 2005; de Haas 2008; de Haas & van Kemenade 2009). Nevertheless, the assumption that the NSR was entirely lacking in Old Northumbrian stands on shaky grounds without further detailed analysis of the tenth-century northern writings, as has been pointed out in the literature (Benskin 2011:170). As might well be imagined, such an endeavour is hindered by the fact that extant textual evidence from the period is far from abundant, and that which remains is limited in nature: the only substantial Northumbrian texts passed down to us are the interlinear glosses to the Latin manuscripts of the Lindisfarne Gospels and the Durham Ritual supposedly written by the same scribe, Aldred, in the second half of the tenth-century, as well as the Northumbrian part of the Rushworth Gospels gloss (Rushworth 2 ), written by a scribe called Owun in the late tenth-century and heavily reliant on the Lindisfarne gloss. Yet despite their limitations, the glosses constitute a substantial record of late ONrth verbal morphology that provides important insights into the mechanisms of linguistic change. Although the study of the Northern Subject Rule in the early northern writings has barely been touched upon in the literature (as far as I am aware the matter has only been cursorily considered by de Haas 2008), morphological variation between -s as opposed to -ð in the late Northumbrian texts has been the object of numerous quantitative analyses (most famously Holmqvist 1922; Ross 1934; Blakeley 1949/50 and Berndt 1956). It is striking, however, that the vast majority of these studies were written well over fifty years ago and the matter has not been thoroughly considered since. A reconsideration of present-tense marking patterns in Old Northumbrian that draws from the insights of recent research into variation and benefits from the application of modern statistical methodology is clearly long overdue. Furthermore, certain potentially relevant factors remain unexplored. For instance, while grammatical person and number have been identified as important factors in conditioning variation between the interdental and alveolar variants, the effect of subject type and adjacency on morphological variation in Old Northumbrian has hitherto been disregarded. This is despite the fact that research indicates that subject effects are a crucial factor in determining the selection of verbal morphology, not just in non-standard varieties of present-day English (cf. Chambers 2004; Tagliamonte 2009) and in varieties of EModE, as discussed above, but also most notably in Middle English northern dialect itself (McIntosh 1989; Montgomery 1994; de Haas & van Kemenade 2009; de Haas 2011). Using data drawn from the standard edition of the Lindisfarne gloss (Skeat 1871-87) collated with the facsimile copy of the manuscript (Kendrick, T. D. et al., 1960), this dissertation carries out a detailed study of the replacement of the interdental fricative by the alveolar fricative which differs both methodologically and in perspective from previous studies in several crucial ways. It constitutes the first study to simultaneously examine the effects of all relevant phonetic, lexical and syntactic variables on the process of change using statistical quantitative methodology. The study approaches the issue from an innovative hitherto disregarded perspective and considers factors such as lexical conditioning and morphosyntactic priming and pays particular reference to the subject and adjacency effects of the so-called Northern Subject Rule. By analysing the full breadth of possible language-internal explanatory variables on the development of the alveolar fricative ending in late Old Northumbrian and by applying statistical methodology, the study aims to elaborate and refine the overall view presented in early studies and set the Northumbrian developments within a broader framework of diachronic variation that will aid the verification of crosslinguistic

generalisations and further our understanding of regularisation processes. It will be shown that the distribution of ONrth verbal morphology constitutes the first attested manifestation of a tendency in English for subject type to compete with person and number features for the function of grammatical material. In addition to a variationist study of -ð and -s forms, this dissertation also carries out a contextual and quantitative analysis of reduced morphology in the Old Northumbrian interlinear gloss to the Lindisfarne Gospels. It looks in detail at reduced forms in the Lindisfarne gloss and considers to what extent the nature and distribution of these forms are indicative of the incipient development of the ME -s versus -e/Ø NSR pattern in late Old Northumbrian. I also assess to what extent inflectional morphology already present in the northern dialects constitutes the historical source for the occurrence of -e/Ø/n in the present indicative. To this end, I posit that, not only present-subjunctive morphology, but also preterite-present and preterite-indicative verbal morphology played an important role in perpetuating the levelling of reduced forms and -n into the present indicative. I show that the subject and adjacency effects at the heart of the NSR appear not only to govern the occurrence of reduced morphology in the present indicative as a low frequency variant but also conditions the distribution of reduced verbal morphology in the preterite. A further question that will be examined in this dissertation involves the contentious issue of the authorship of the glosses to Lindisfarne and whether or not the interlinear gloss of the Lindisfarne Gospels was the work of a single hand, Aldred (Ross, Stanley & Brown 1960; Brunner 1947/48; van Bergen 2008). To this end, I will consider the utility of language variation as a diagnostic for determining the authorship and more specifically, what light is shed upon this unresolved problem of Old English philology by the distribution of variants verbal forms in Li. Another aspect under consideration relates to methodology and the unreliability of the text editions of medieval sources for linguistic research. In general, editions are unsuitable as sources unless they are collated with the raw data of the original manuscript because, as van der Hoek (2010) points out, they tend to involve “a reconstruction of a non-extant version of the text in question by selecting and altering from among the different surviving versions, in the attempt to arrive at a text that is purer from either a literary or philological point of view.” The edition in question, in the case of the Lindisfarne Gospels, is that of Skeat (1871-87) which relies on the sole version of Li. but whose language and grammar have nevertheless been subjected to editorial interpretation and alteration.