SOCIOLINGUISTIC ARGUMENTS IN FAVOUR OF A DANELAW PROVENANCE FOR BEOWULF
Linguists arguing for an early date of the Beowulf text and a long history of transmission through copying which results in the extant manuscript, despite the loss of all the hypothesised earlier manuscripts, were on relatively safe ground as long as they were looking at the manuscript from a purely linguistic point of view. But since the advent of sociolinguistic research in the 1960s, the arguments can be shown to be worthless against the background of what we now know about language change, linguistic variation, processes of standardisation, dialectal variation and stylistic differences between text genres.
Kiernan suggests that “by far the most persuasive case for an early date rests in the bewildering variety of linguistic forms, of uncertain date and dialect, embedded in the essentially Late West Saxon dialect of the preserved text” (1996: 23). However, this becomes a problem only if one firmly believes in a homogeneous linguistic system spoken by members of a socially homogeneous speech community. Since the appearance of the groundbreaking article “Empirical foundations for a theory of language change”, by Weinreich, Labov, and Herzog (1968) , sociolinguists have assumed the opposite. No linguistic system is ever homogeneous, no speaker (or even writer) is ever in perfect control of such a system and no linguistic system can ever be divorced from the social and cultural conditions of its use by a potentially endless number of speakers (writers). Language always varies from speaker to speaker and from one occasion of its use to the next. That variability is the essence of change in language. There may be variability without the language necessarily having to change, but there could never be change if there were no linguistic variation.
The misguided belief in linguistic homogeneity leads scholars to desperate measures in explaining the profusion of varied forms in the Beowulf manuscript. Many have maintained that those forms which are not West Saxon must be relics copied into the successive manuscripts of Beowulf by somewhat mindless scribes. We have already seen that scribe A of the Nowell codex could certainly show a lackadaisical attitude towards copying if we examine his performance in the three prose texts of the codex, but his work on Beowulf was meticulous, and both he and scribe B provide adequate evidence that they understood perfectly what they were writing out. The only two Anglo-Saxon texts that can be shown to have a long textual history (Csdmon’s hymn and Bede’s death song) do not have traces of other “dialects” or earlier forms in their transition from early Northumbrian to Late West Saxon.
Klaeber (1950: cviii) suggests a syntactic test whereby progressively later texts should show an increase in the use of the old demonstratives used as the new definite determiners, a decrease in the use of weak adjective and noun inflections and an increase in the use of the determiner with a weak adjective and noun. So if the Beowulf manuscript has high proportions of such structures, these must be traces of the earlier manuscripts. The argument crumbles into dust, however, if we accept a predilection on the part of poets to use archaic forms of language in their work. The Battle of Brunanburh and The Battle of Maldon are both tenth-century productions, but they, too, have similar kinds of archaic, formulaic syntactic structure. Even today, songwriters in the folk idiom use archaic structures to evoke an atmosphere of the past. A spurious syntactic argument such as Klaeber’s is based on a disregard for variation and simply takes no account of different styles in different written genres.
A further “linguistic test” for the conjectured long manuscript history of Beowulf is phonetic/metrical. The argument runs as follows: certain half lines do not appear to scan well because they contain contractions. Hence in the “original” text they must have appeared with uncontracted forms; for example, gepeon is said to appear in place of a reconstructed original *gepihan. The argument is spurious, however, since, in accordance with syllable theory in phonology (Roca & Johnson 1999, chap. 9; Zec 2007), both words are trisyllabic (ge- pe- on and ge- pi- han), and we can surely expect both performers and listeners (readers) to have had a good sense of the rhythmic and metrical conventions of their own poetry. In addition, if scribes were sensible enough to change *gepihan to gepeon from one older manuscript to another, why did they fail to insert definite determiners into a large percentage of determinerless noun phrases in Beowulf?
I will not run through the story of the hypothetical “early” form wundini, which resulted from a blind reliance on Zupitza’s facsimile, except to say that it remains the only argument left to those who wish to use linguistic evidence to prove their hypothesis of the long manuscript history of Beowulf. Kiernan makes the eminently commonsense point that if one were to carry out a detailed study of the manuscript rather than the text, the hypothesised word wundini disappears, and the word that can be suggested (and only suggested because of the damage to the vellum at this point in the manuscript) is wunden, which would be perfectly normal for the poetry of the early eleventh century.
How can we account for the variation of forms that occur in the Beowulf manuscript? First, we should remember that sociolinguists know very little indeed about dialectal variation in pre-Conquest times. The estimated population of England at the turn of the eleventh century was between 1.5 million and 1.8 million. With relatively poor means of transportation overland, goods were more easily carried by boat around the coasts and up the navigable rivers, which led to population clusters in coastal areas and along the courses of such rivers as the Thames, the Severn, the Humber, the Trent, the Yorkshire Ouse and the Avon. So although population in those areas was not necessarily scattered, communication between them was not always easy. We would therefore expect a reasonably broad range of spoken dialects to have evolved from the fifth to the eleventh century, but we have little to go on apart from the written documents that have survived.
There are notorious difficulties in extrapolating from written documents to hypothesised oral usage, which I will not go into here. Suffice it to say that we can very broadly accept a range of dialects reaching from the northern limits of “English”-speaking territory (present-day Northumberland and the southeastern counties of lowland Scotland) down to the River Humber. These have traditionally been called “Northumbrian”, although even within this large area there must have been degrees of variation. A second area stretched from the east coast south of the Humber across to the Welsh borders and as far south as the Thames valley, constituting the varieties that have traditionally been called “Anglian”. Here, however, a language contact situation existed roughly to the east of the Roman road called Watling Street and created by widespread Danish settlement in those areas. We can assume that speakers of Danish in what was called the Danelaw area would have lost their mother tongue within roughly three generations if they were in frequent contact with speakers of Anglo-Saxon, but not before Danish had exerted a considerable lexical and morphological influence on eastern Anglian forms of Anglo- Saxon. The Danish (or it might be more appropriate to talk of “Norse”) influence also extended into the Northumbrian area, and there is evidence that Danish (or Norse) was still spoken in York in the tenth century.
The two remaining dialect areas were Kentish south of the River Thames in the southeastern part of the country as far as the coast, and West Saxon roughly south of the Thames to the coast and stretching west as far as the Bristol Channel coast, Cornwall and the River Severn. The struggle for the control of England after the Danish incursions of the late eighth and ninth centuries led to the hegemony of the West Saxon royal house and, from the time of King Alfred on, the imposition of a proto-standard written variety of West Saxon.
At this point, we need to be very careful about using the term “standard”, which is too often and too glibly used to refer to this written variety in the literature. In chapters 8, 9 and 10, I will focus in more detail on the discourse of the “standard language”, which the Milroys have rightly called “the ideology of the standard”. Linguists in the twentieth century have often failed to see that their concept of a homogeneous standard variety is inapplicable to earlier varieties of English. One of the principles of the “ideology of the standard” is that there should be, as far as possible, no variation, particularly in written genres. Judging by their use of Late West Saxon, many of the scribes involved in producing the manuscripts of this era felt under no compunction to avoid orthographic and lexical variation, much of which reflected the area from which they came and in which they wrote. This is indeed the case in the Beowulf manuscript, and the vast majority of deviations from Late West Saxon are demonstrably of Anglian origin. In his discussion of pertinent examples of variation, Kiernan concludes that “the reversal in the mixture of forms is more likely to occur from conscious variation, or laxity in the spelling tradition, than from remarkably coincidental scribal slips in the course of transmission” (1996: 56). I agree with him in all but one point: there was probably no real “spelling tradition” that the scribe needed to adhere to.
None of the deviations in the Beowulf manuscript constitutes linguistic proof of a long manuscript history, and almost all of them support the sociolinguistic focus on variation and heterogeneity. If we combine these results with the contents of the poem, which are undeniably Danish/Scandinavian, it is a small and logical step to conclude that the manuscript, at least part of which was authored by scribe B, was at just one remove from and very possibly contemporaneous with the original. We can also conclude that the “Beowulf project”, for want of a better term to refer to it, was scribe B’s very own and that it was carried out at a scriptorium somewhere in Danelaw Anglian territory.
-  Witness the birth of the mindless scribe myth, which has severely hampered sensible discussion ofsociolinguistic aspects of medieval manuscripts throughout the Anglo-Saxon and Middle English periods.
-  Irish songwriter Sean Mone recently wrote a song called “New Holland Grove” about Irish immigration to North America in the nineteenth century. The refrain is as follows: “Here’s a health unto you, sweetKeady town, a village of renown / And likewise to New Holland Grove where pleasure could be found. / Wherelads and lasses sport and play on the bright long summer’s day, / But I am resolved to leave all behind, all forAmericay” (archaic, formulaic structures noted in bold type). The distance in time between the fictive experience and the song itself is roughly 150 years.