The Battle of Brunanburh: The Lanchester Hypothesis

Abstract

The location of the battle of Brunanburh in 937 remains a source of disagreement among investigators. In recent years many places have been identified as Brunanburh. This article interrogates the claims of Andrew Breeze, in several works, to have securely located the battle at Lanchester in County Durham. The methods by which Breeze reaches his conclusion are analyzed, and the arguments he cites for it are examined. Breeze’s main proposals are discussed: that Brunanburh refers to the River Browney in County Durham; that the name We(o)ndun recorded by Symeon of Durham refers to a wen-shaped hill; and that dinges- in the Old English poem on the battle should be emended to dingles-. Alternative interpretations of the material are given, some based on hitherto unexamined evidence, including a new suggestion for the etymology of Dingley in Northamptonshire. It is argued that the Lanchester hypothesis does not stand linguistic and critical analysis.

Introduction

The battle of Brunanburh was fought in 937. West Saxons and Mercians under King Athelstan and his brother Edmund decisively defeated the armies of King Constantine of Alba, King Anlaf of Dublin and Owain of Strathclyde. Some of the details of the battle are obscure and different writers add information to the tradition which may or may not be reliable, for example, from local memory or lost accounts. A text that has caused debate over the years is John of Worcester’s record that the Scandinavian fleet from Dublin landed in the Humber (Darlington and McGurk, 1995: II, 392–3, sub anno 959 [937]). John is the first to add a place of landing for the fleet, and is followed by several medieval writers. This added information has led to controversy, because some recent writers argue that the battle must therefore have taken place in Yorkshire or on the north-east coast of England, notably Wood (2013).¹ Others reject John’s information and suggest that the battle might have taken place elsewhere (Downham, 2021; Livingston, 2021): suggested sites have been listed by Hill (2004: 135–60) and Cockburn (1931).

Sources and Method

The question of the site of the battle of Brunanburh continues to fascinate writers and investigators of all kinds. The battle itself has been of interest for centuries, and well over fifty medieval poets and historians mention, or narrate, or discuss it. Most of these medieval sources have been collected, edited, translated and discussed in Michael Livingston’s Casebook (Livingston, 2011).² Some find the welter of sources daunting. Andrew Breeze, whose work is the focus of discussion in this article, makes no reference to the Casebook. A very positive review of Breeze’s book British battles notes the omission as a ‘mystery’, but then rehearses Breeze’s identification of Lanchester as the site of the battle and remarks, ‘as a potential guide for archaeological exploration, and as proof of method, it beats the heck out of the sort of conclusion that Cavill and Harding can reach, which is roughly that a precise location near Bromborough remains, and is always likely to remain, undeterminable’ (Morillo, 2021).

Some impatience with the detail of the sources and lengthy discussion of them is understandable. But the reviewer has not picked up from Breeze that the Old English poem, The Battle of Brunanburh, the most reliable source for the battle, gives such significant and uncontroversial information as that the battle took place ymbe brunanburh ‘around Brunanburh’ (5a, naturally making a precise location difficult), that the English pursued the fleeing forces ondlongne dæg ‘for the whole day’ (21a, so it must have spread over a wide area), that Anlaf fled from the battle by sea in a ship (32a–6, so not from Lanchester or anywhere near it, as it is seventeen miles from the nearest coast; the Wear at Durham is over seven miles away), and much more (Campbell, 1938: cited by line-number of the edited text of the poem). These and similar details are crucial to the interpretation of the site of the battle.

Breeze has made strong and repeated assertions about Brunanburh. Without qualification, he identifies the battle as having taken place at Lanchester in County Durham. The article which launched this identification most seriously was ‘Brunanburh located’ (Breeze, 2018a), a title which leaves the reader in no doubt about the conclusion proposed. The assertions are repeated in other works (Breeze, 2016, 2018b: 89–90, under the River Went, 2021: 121–127). The article is a selective history of Brunanburh criticism with Breeze’s pithy judgements added: a ‘move in the wrong direction’ (2018a: 63), ‘We shall turn the argument upside down’ (2018a: 66), ‘misplaced enthusiasm’, ‘misplaced confidence’ (2018a: 69), ‘The interpretation is out of the question’ (2018a: 70), ‘This is false’ (2018a: 72), ‘far-fetched’ (2018a: 74) and much more. The criterion of evaluation is whether or not the writers agree with Breeze’s view about where Brunanburh took place: so, for example, Plummer’s scholarship ‘had an unfortunate influence on all later writers’ (2018a: 62), because he thought John of Worcester was mistaken and a west-coast location was more plausible (Plummer, 1952: II, 139–140).³

The confidence of the assertions is not in doubt, but Breeze chooses largely to ignore the detail of the evidence. He discusses selected words of Latin, Welsh, Old or Middle English, and dismisses arguments contrary to his own rather than offer cogent criticism. Breeze picks out the names (Brunanburh, dinges mere) from the Old English poem on the battle, but does not discuss its evidence except to correct Longfellow’s translation (2018a: 72). Below I will examine Breeze’s arguments and explore the evidence which Breeze does not use. It will become evident that Breeze’s arguments do not hold up under scrutiny. It will also become evident that scholars in the Casebook (Livingston, 2011) have articulated arguments and discussed the evidence relating to a case for Bromborough on the Wirral as the site of Brunanburh. The point of this article, however, is not to advance that specific case in a direct fashion, but simply to analyze the proposal advanced by Breeze for Lanchester.

Welsh and Latin

Breeze opens his article, ‘Brunanburh located’ thus, with reference to an entry in the Annales Cambriae:

We begin not with the English poem but a Welsh annal, which (for the year 938) has Bellum Brune, where bellum is ‘battle’. This might seem bald and unhelpful. Yet it shows that the form is a toponym, not a personal name. We may compare Old Welsh Gueit Conguoy for the year 880, where gueit is also ‘battle’ and Conguoy is the River Conway of North Wales. So we can be certain that Brune and Brunanburh are not called after some Anglo-Saxon, as has been thought. (2018a: 61)

The initial puzzle as to who might have thought Brune was a personal name here, and how these Latin and Welsh words prove anything quite so certain about the Anglo-Saxon name Brunanburh is partly explained a few pages later.⁴ Breeze comments on the phrase ymlad y Brune in ‘native chronicles’ (2018a: 64),⁵

The y before Brune shows the form was in later Welsh chronicles unrecognized as a hydronym, because Welsh river-names (unlike English ones) appear without the definite article. Even Welsh grammar thus supports a river-side location for the English victory.

This latter argument is repeated later (2018a: 70):

In a later chronicle in Welsh it is ‘The Battle of (the) Brune,’ where the parenthesis reflects one version of the text. Because hydronyms in Welsh do not have the article, the article proves that Brune was no longer perceived as a river-name.

Breeze’s overall argument is that these entries and the Old English name Brunanburh refer to the River Browney in County Durham. But a moment’s reflection will reveal this argument from Welsh sources to be quite extraordinary and contradictory. The Annales Cambriae may be a Welsh chronicle, as its modern title suggests, but it is written in Latin.⁶ Dumville lists eight or (if one counts the word in as Welsh) nine Welsh words apart from personal and place-names in the text from which Bellum Brune is taken (2002: 23).⁷ The lack of the Welsh definite article in a Latin annal proves nothing, certainly not that Brune was a river name. In fact, the Welsh article y is not recorded at all in the A text of the Annales Cambriae 682–954, London, British Library, MS Harley 3859, fol. 190r–193r, the only text to include the entry for 937 Bellum Brune.

Two of the Welsh terms are the battle words gueit and cat and these occur frequently in the Annales Cambriae text, always without a definite article: Gueith Gart Mailauc 722, Cat Pencon 722, Guei[th] Mocetauc 750, Gueith Hirford 760, Gueith Cetill 844, Gueit Finnant 848, Cat Brinonnen 870, Gueith Ban\n/guolu 874, Gueith diu Sul in Mon 877, Gueith Dinmeir 906, Gueith Dina\s/ Neguid 921. Not all of these places have been identified, but several demonstrably are not references to rivers: Hirford 760 is probably Hereford and Cat Brionnen refers to the battle of Ashdown (Dumville, 2002); ‘Sunday’s battle in Anglesey’ 877 similarly does not contain a river name. Elements such as garth ‘ridge, hill’, and dinas ‘fort’ likewise are obviously not river names (Owen & Morgan, 2007). Indeed, Dumville translates Gueit Conguoy, Breeze’s chosen comparandum for Bellum Brune as containing a river name, as ‘The battle of Conwy’, apparently the place rather than the river (2002: 12).⁸

The presence of the definite article in reference to the battle in the Brenhinoedd y Saeson is emphatically repeated by Breeze to demonstrate that Brune was not (then) perceived as a river name. But by some prestidigitation it therefore ‘supports’ the notion that Brune actually was a river name, because a correct construction of the element was in this case ‘unrecognized’. The evidence taken at face value and sensibly interpreted shows the very opposite of what Breeze claims. Welsh grammar, insofar as it applies to the Annales Cambriae at all, shows that there is no specific reason to suppose that Brune was a river name; and where Welsh grammar clearly applies in the Brenhineodd y Saeson, it shows that Brune was emphatically not a river name.

Breeze puts a great deal of emphasis on these Welsh sources as determining the meaning of the Old English Brunanburh. The extant Welsh sources are presented, translated and discussed in the Casebook by John Bollard and Marged Haycock (Livingston, 2011: 88–89, 216–219, Bollard & Haycock 2011: 245–268). They include, in Bollard’s and Haycock’s texts and translations: Ac y bu ryfel Brun ‘And there was the battle of Brun’ in the late thirteenth-century Brut y Tywysogion (Livingston, 2011: 88–89);⁹ and y bu ymlad y Brune ‘There was the battle of Brune’ in the early thirteenth-century Brenhinoedd y Saeson has already been discussed. Both of these texts are chronicles, and the manuscript dating anchors the references to the battle of Brunanburh.¹⁰

There are other possible examples that are rather obscure and may or may not refer to the battle. These are the references to *kattybrunawc ‘the battle for the settlement in Brun’s region’ in the late tenth-century Welsh Glaswawt Taliessin (Livingston, 2011: 48–49);¹¹ and Cad Dybrunawc ‘the battle of Brunanburh(?)’ of the late twelfth-century Canu y Dewi (Livingston, 2011: 66–67). The consensus here is that this brun- might have referred to the English Brunanburh; the element is not obviously a meaningful Celtic term. Brun and -brun- might have been toponymical in these non-chronicle texts, to be sure, as they appear to be in the Welsh chronicle texts. But they might also have been reflexes of a personal name. Brun(e) was not obviously a river name in the chronicle texts, and there is no particular reason to suppose it was a river name in the non-chronicle texts.

A question arises as to why there was any Welsh interest in this battle at all. Numerous battles are recorded in the text of Annales Cambriae 682–954, but the only ones that took place on English soil, apart from Bellum Brune, were the battle of Hereford in 760 between the English and the Welsh, not recorded in the Anglo-Saxon Chronicle, and the battle of Brinnonen or Ashdown in 870. In the one, the Welsh fought the English in the engagement; after the other, Alfred became king, and Asser, his Welsh bishop from St David’s and biographer, was very interested in the battle.¹² Neither of these battles is recorded in the C text of the Annales, London, British Library, MS Cotton Domitian A. i, fol. 138r–155r; and indeed Bellum Brune is not recorded in the C text, or the B text, London, National Archives, MS E.164/1, pp. 2–26 either. If, as Breeze believes, Brunanburh took place in County Durham, it seems extraordinary that the Welsh chroniclers should show any interest at all since the Welsh were neither involved, nor, on this account, anywhere near the place of engagement. A site for the battle close to Wales and within the orbit of Welsh influence and concern, such as the Wirral, is very much more plausible simply on the basis of proximity and immediacy.¹³

The Community of St Cuthbert

Medieval English sources relating to north-east England are available. Lanchester is about seven or eight miles distant from Chester-le-Street where the community of St Cuthbert was based in 937, and on the boundary of its enormous territories, the Roman road Dere Street. One of the community’s histories, the Historia de sancto Cuthberto, gives lavish details of King Athelstan’s visit to the monastery in 934, as he made his way north to ravage Scotland, including a record of a charter Athelstan granted the community (South, 2002: 64–67, ch. 26–27). The course of the River Browney runs from the west of modern Consett, through Lanchester, through the ancient lands of the community of St Cuthbert towards Durham, where it joins the Wear. If the battle of Brunanburh took place anywhere near the Browney it would have been of vital interest to the community. But no mention is made of Brunanburh in the Historia de sancto Cuthberto.¹⁴ The Cronica monasterii Dunelmensis in the Red Book of Durham, a probably eleventh-century record surviving in a manuscript from the early fifteenth century, records Athelstan’s visit in 935 (sic) on the way to Scotland, and his benefactions to the community and prayers to St Cuthbert, but nothing of a battle in 937 (Craster, 1925: 525–526).¹⁵ Similarly, the Annales Lindisfarnenses et Dunelmenses, possibly from the early twelfth century, record Athelstan’s expedition to Scotland in 924 (sic.), his acquisition of totius Angliae monarchiam ‘rule over all England’, and his death in 939, but again, no Brunanburh (Levison, 1961: 485).¹⁶

When we turn to the other indispensable source for the history of the community of St Cuthbert, Symeon of Durham’s own work, we find that Symeon is reduced to adopting the rather implausible account of the 615 enemy ships that came to the battle also found in the Historia regum (Arnold, 1885: II, 93, § 83) into his history, the Libellus de exordio (Rollason, 2000: 138–139, ii. 18).¹⁷ Symeon adds that Athelstan sancti Cuthberti patrocino confisus ‘trusted in the protection of St Cuthbert’ for the battle, and gives the name of the place as Weondun as in the Historia regum, but adds, quod alio nomine Aet Brunnanwerc uel Brunnanbyrig appellatur ‘Weondun which is called by another name Æt Brunnanwerc or Brunnanbyrig’. He gives no details either of the fighting, or of anything that would locate the battle in the community’s territory (or anywhere else, beyond the idea that he reges illos de regno suo propulit ‘drove those kings from his kingdom’).¹⁸

Athelstan carefully cultivated relations with the community of St Cuthbert according to the local records. Symeon records that ante illum nullus regum ecclesiam sancti Cuthberti tantum dilexit ‘no king before [Athelstan] held the church of St Cuthbert in so much affection’ (Rollason, 2000: 134, ii. 17). Given the efforts made by Athelstan to maintain the friendship between the West Saxon kings and the monks, valuable to both parties, this is plausible. It is unimaginable that Athelstan would not have visited the community before or after the battle of Brunanburh if it took place seven or eight miles away from the community’s base, but no visit in 937 is recorded. Breeze, aware of this difficulty, suggests that Athelstan really did visit, and that the Life of St Cuthbert in prose and verse mentioned in the Historia de sancto Cuthberto, generally identified as Cambridge, Corpus Christi College, MS 183, which the charter says he gave to the community, is the ‘outward and visible sign’ of the visit. The argument hangs on the fact that the episcopal lists in the manuscript include Ælfheah of Winchester and Æthelgar of Crediton, who did not accede to their sees until after Athelstan had departed on his Scottish expedition, so the book could not have been given to the community in 934, as has been supposed. Breeze’s alternative view is that the book was given after the victory at Brunanburh ‘when Athelstan will have been in the Chester-le-Street area’, having been prepared as ‘propaganda’ in the summer of 937 (2016: 144–145).¹⁹ Whatever one might make of Breeze’s suggestions about the book, the silence of the narrative and annalistic sources about a triumphal royal visit in 937 is inexplicable if it actually happened.²⁰

The above is sufficient to indicate that there are grounds to be cautious about accepting Breeze’s interpretations of the context of Brunanburh, no matter how forcefully they are asserted. The Welsh sources give no support to the idea that Brune, -brun- or Brun was the river name Browney or necessarily a river name at all. The relative interest shown by the Welsh sources in the battle of Brunanburh, and the contrasting lack of interest and information about Brunanburh in the English sources local to Lanchester and the Browney, suggest very strongly that the battle did not take place in the north-east. The substantive elements of Breeze’s argument will now be considered with a view to assessing their reliability.

Brunanburh

Breeze repeatedly identifies the River Browney in County Durham as the referent of the first element of Brunanburh. He quotes Bosworth’s suggestion (Bosworth, 1898: 129) of ‘a plain between the Rivers Wear and Browney [Brunan ea]’, without acknowledging that Bosworth’s Old English gloss of the name Browney, *Brunan ea, does not exist in any extant text (2018a: 61). It is not an extant spelling, but a linguistic justification for the identification Bosworth makes. Breeze only quotes a form Brune for the river name, and never gives a direct source for that: ‘[t]he Browney near Durham was once called Brune, and there may have been others’, he remarks (2018a: 63), attributing this idea to Campbell. Brune ‘reproduces an Old English weak declension nominative’, he informs the reader at a later stage (2018a: 70). He gives no reference for the spelling of the name, and only gives a secondary source in ‘Names of Yorkshire’s rivers’ (Breeze, 2018b: 70), where he cites as support for the name Brune Ekwall’s Concise Oxford dictionary of English place-names (Ekwall, 1960).²¹

The implication of this kind of assertion is that it is an obvious and settled fact that the Browney always appears as Brune in the sources and that this form represents a weak nominative Old English noun. This would in turn give the inflected genitive form Brunan- in Brunanburh and related names. Standard onomastic sources indicate, however, that the name of the Browney was more frequently Brun, with a second element, deriving from OE ēa ‘river’, later added. Ekwall in his Dictionary, quoting forms ‘Brune c 1190, Brun c 1195 Finchale’, gives ‘OE Brūn or Brūne “the brown one” with later addition of -ey, which may be OE ēa “river”’ (1960: 70). Ekwall’s standard work on river names gives a range of forms, and glosses ‘OE Brūn or Brūne (fem) “the brown one”’ (1928: 55). Victor Watts glosses Browney as ‘“The brown one”. OE brūn + ēa “river”’ (Watts, 2002: 19).

There is no known source in the Old English vernacular containing the name. This means we are dependent on Latin sources, mainly post-Conquest charters, for spellings of the river name, in some of which it appears as Brune. Medieval writers had some difficulties with rendering the names of rivers in England in Latin and took different approaches. Bede, for example, tended to give Latin inflections to his river names: so the Thames has various forms ad flumen Tamensim (Colgrave and Mynors, 1969: 22, (i) 2), Tamense fluuio (Colgrave and Mynors, 1969: 142, (ii) 3), Tamensis (Colgrave and Mynors, 1969: 254, iv. 5); the Humber has forms Humbri fluminis (Colgrave and Mynors, 1969: 50, (i) 15), Humbrae fluuio (Colgrave and Mynors, 1969: 148, (ii) 5); and the Wear has fluminis Uiuri (Colgrave and Mynors, 1969: 388, iv. 18).²² Twelfth-century writers vary in their practice. William of Malmesbury writes of Wirae amnis (Winterbottom and Thomson, 2007: 494, I, iv. 186.7).²³ For the Wear and the nearby Tyne, Symeon of Durham gives inter Weor et Tine (Rollason, 2000: 124, ii. 13). John of Worcester regularized river names with the exception of the Thames, giving the -e ending without inflectional variation, iuxta flumen … Tine (Darlington and McGurk, 1995: 304, II, 875), Tine fluminis (Darlington and McGurk, 1995: 602, II, 1066).²⁴

It is difficult to be sure what the significance of the inflection on the river name might be. Ekwall in English River-Names and elsewhere posited that Brune might be an Old English feminine (weak) noun, and Breeze follows this interpretation without question and without reference to the predominant strong (and probably masculine) forms recorded as Brun.²⁵ To arrive at any reliable conclusion about this name, we have to assess the early evidence, presented below. The main sources are charters of Finchale Priory (Raine, 1838, cited as Finchale, by date and page number), the survey of Durham estates, the Feodarium Prioratus Dunelmensis (Greenwell, 1872, cited as FPD, by date and page number), and the Calendar of Charter Rolls (Maxwell Lyte et al., 1906, 1912, cited as CCR, by volume and page number). There is an interesting pattern of usage in the early sources to be noted.

• (a) ultra aquam de Wer usque ad aquam de Brun (12th century, FPD lv)

  [The religious house of Sancta Maria] de Novo Loco super Brun (c.1195 and frequently, Finchale 9, 10, 11, 12, 15)

  usque in Brun, et ita ascendendo de Brun (1268 FPD 187)

  ab aqua de Brun; usque in aquam de Brun (1268 FPD 188)

  ex australi parte aquæ de Brun … usque in aquam de Brun (1270 CCR II: 141, 1300 FPD 187)

  ultra aquam de Wer’ usque ad aquam de Brun (1195 (1335) CCR IV: 323)

• (b) ex occidentali parte Brune fluminis … in ipsum fluvium Brune; ex altera parte predicti fluminis Bruni (c.1190 Finchale 8–9)

• (c) ab aqua de Were … iterum in aquam de Brune (c.1300 FPD 192)

  ab aqua de Were usque ad aquam de Brune; ultra aquam de Brune … ultra aquam de Brune (14th century, FPD 193–194).

The first group of spellings, (a), clearly indicates that the river name was Brun, as Watts (and Ekwall less decisively) proposed. The second group (b) shows how some Latin writers instinctively Latinized the morphology of the name, without necessarily being sure which class of noun (or adjective) it might have belonged to. Alternatively, these forms in (b) are examples of Brun with a Latinized or reduced English form of ēa ‘river’. The third and later group, (c), gives the name in the form Brune, possibly here an ablative of the third declension, following the prepositions de and in, but possibly a generalized ending of no grammatical significance. In the first group of spellings, Wer ‘the River Wear’, like Brun, does not show a Latin inflection; but the same writers who add the -e to Brune also add it to the Wear, as in Were, when it is syntactically appropriate in (c).²⁶

What these groups of spellings indicate, rationally interpreted, is categorically not the presence of a weak Old English noun in the nominative case. It is entirely to be doubted that twelfth-, thirteenth- and fourteenth-century charter scribes in County Durham could have reconstructed the morphology of such an Old English noun or adjective if they had wanted to. The spellings of the River Browney in the early sources indicate that the river name was Brun, occasionally Latinized and given a Latin or generalized inflection. In Old English, which is understood to be the language in which the river was named, Brun as a noun is a strong masculine with a genitive singular in -es, appearing in such names as Brownsall Hundred, Dorset (Bruneselle c.1086, Brūn in the genitive Brunes + hyll ‘hill’, Mills 1989: 275–276),²⁷ and Bruneshurst, Wybunbury Cheshire (Bruneshurst 1275, Brūn in the genitive Brunes + hyrst ‘wooded hill’, Dodgson 1971: 53–54).²⁸ It cannot therefore have given the weak oblique form Brunan- even if a burh were to be associated with it. This evidence in itself is sufficient to discredit Breeze’s claim that Brunanburh refers to the Browney: Breeze’s whole case rests on the supposition that the name of the Browney was a weak feminine Old English Brune, with an oblique inflection Brunan, and on the evidence presented above, the claim cannot be substantiated.

Lanchester

Breeze’s argument, based on the supposition that Brunan- refers to the River Browney, then identifies the burh as Lanchester. Breeze gives little by way of argument for the identification:

The Anglo-Saxons often called a Roman fort a burh. … The stronghold being situated directly north of the Browney, there is no objection to taking it as Brunanburh or Brunanwerc [sic.] ‘fort of the Browney, fortification of the Browney’, and the hill on which it crouches as Brunandune ‘hill of the Browney.’ If it is objected that the form is nowhere elsewhere recorded, we may remember that study of County Durham’s early place-names is problematic. There are few early charters, and the region is absent from Domesday Book. (2018a: 73)

Breeze’s identification of the Roman fort at Lanchester with that burh depends on a complete lack of supporting evidence, as he acknowledges. The same documents that identify the Browney also identify Langcestre 1248 FPD 186, Langchestre c.1300 FPD 192, and we might have expected some nod in the direction of the Brunanburh name if it derived from the river name and was current here. It is true, as Breeze claims, that the Anglo-Saxons sometimes ‘called a Roman fort a burh’, but it is doubtful that this one was ever so called. The Roman fort was Longovicium, and the earliest references in vernacular sources are from the middle of the twelfth century, Langescestre, Langchestre. The second element of the name is OE cæster ‘Roman fort’, and the first element is likely to be a borrowing of British longo- ‘ship’ assimilated to Old English lang ‘long’ (Watts, 2004: 357–358).

Despite the late forms, it seems probable that there is here a tradition of naming that dates to Anglo-Saxon times. Parallel processes of borrowing of British elements occur in the names Winchester (which borrows Venta), Wroxeter (which borrows Virico), Leicester (which borrows Legore) and others (Parsons & Styles, 2000: 158–162). There are examples where a cæster name from Bede is replaced by a burh name: Tiouuulfingacæstir which became Littleborough, Nottinghamshire; and Reptacæstir which became Richborough, Kent (Parsons & Styles, 2000: 159). In both these cases, however, it was the burh name that survived, not the ceaster one. It is unwise to be dogmatic, but the Anglo-Saxons most likely called the Roman fort on the Browney something like ‘Lanchester’, and no evidence exists to suggest that they called it Brunanburh.

We(o)ndun

Mention has already been made of the information borrowed by Symeon of Durham from the source of the Historia regum (Arnold, 1885: 93, II, § 83), which gives alternative names for the battle:

Ethelstanus rex apud Wendune pugnavit, regemque Onlafum cum dc. et xv. navibus, Constantinum quoque regem Scottorum et regem Cumbrorum, cum omni eorum multitudine in fugam vertit. (King Athelstan fought at Wendun and put to flight King Onlaf with 615 ships, and Constantine, king of the Scots, and the king of the Cumbrians and all their host.)²⁹

Symeon writes:

apud Weondune, quod alio nomine Aet Brunnanwerc uel Brunnanbyrig appellatur, pugnauit contra Onlaf Guthredi quondam regis filium, qui sexcenti et quindecim nauibus aduenerat, secum habens contra Aethelstanum auxilia regum prefatorum scilicet Scottorum et Cumbrorum. (at Weondun which is called by another name Æt Brunnanwerc or Brunnanbyrig, he fought against Olaf, son of the former king Guthred, who had come against Æthelstan with 615 ships and had with him the help of the aforesaid kings, that is of the Scots and the Cumbrians. (Rollason, 2000: 138–139))

Wendun is the form found in the Historia regum; Symeon adjusts the spelling to Weondun. The names share the element dūn ‘low rounded hill’ with a name given in Æthelweard’s chronicle (Campbell, 1962) as Brunandun. Breeze identifies Æthelweard’s hill name as that in the northern texts, and goes on to explain the difference in spelling of the first element between the Historia regum and Symeon as follows:

[In] Symeon of Durham’s Weondune or Wendun, … the first element will be Old English wenn ‘tumour,’ used as a hill-name … Mutation of e to -eo-, which in Old Mercian and Old Northumbrian is general before u in a following syllable (if here with intervocalic grouping of consonants), will be due to the back vowel of dun. (2018a: 73)

Breeze gives a reference to Campbell’s Grammar to explain the process of vowel mutation (Campbell, 1959: § 639). Mutation of this kind is ‘general’ in the sense that it can occur with most single consonants, not in the sense that it occurs in every syllable with a back vowel following. Both Campbell (1959: § 639) and Hogg (1992a: §§ 103–112, at §§ 5. 105 (1)) note that back-mutation is rare with a geminate (double consonant, here -nn-) or a consonant cluster, and there is no apparent example of this umlaut or back-mutation with a geminate and a further consonant (here -nn-d-) such as Breeze posits.³⁰ Moreover, back-mutation is triggered by back-vowels in multisyllabic simplex or affixed words, and by morphological markers: so we have woruld > weoruld ‘world’, cweðað > cweoðað ‘they say’. That is to say, it occurs through regular, repeated conjunction of sounds in speech, or as Hogg puts it, ‘breaking and back mutation comprise an instance of the repetitive character over time of many sound change types’ (Hogg, 1992b: 116). It has yet to be demonstrated that back-mutation can be triggered by the vowels in nonce-compounds such as *wenn-dūn (of which no known example survives even unmutated), and neither published grammar has an example of this postulated process.³¹

The meaning of We(o)ndun is difficult to pin down. Neither Breeze’s intriguing suggestion, nor Wood’s view that it refers to Went Hill in south Yorkshire (2013: 155–158) carry conviction with regard to the phonology and spellings of the forms.³² Smith’s proposal that the specific is the weak oblique adjective *wēon ‘holy’ (1956: II, 254) was adopted by Cavill in the Casebook (Cavill, 2011: 348) and the existence of a minor name apparently containing the element in the vicinity of Bromborough may give support to the general thrust of the Casebook that Brunanburh may be identified with Bromborough on the Wirral. Richard Coates observed that Rice Wood, the Welondrys (1357 ‘scrubland of the shrine’, OE wēoh, land and hrīs) near Bromborough Court House, could be a reference to land associated with an earlier shrine (Coates, 1998: 288–289).

Breeze identifies ‘the hill on which [the fort at Lanchester] crouches’ with Brunandun ‘hill of the Browney’, and thence with the We(o)ndun of the Historia regum and Symeon of Durham. He then interprets We(o)ndun as *wenn-dūn ‘hill shaped like a tumour’. However, the river name Brun cannot give a form Brunandun; and back mutation almost certainly did not occur in the posited form *wenn-dūn and cannot therefore explain a spelling Weondun. There is, besides, no evidence given to support the existence of a name We(o)ndun or the like, or Brunandun or the like, in the vicinity of Lanchester.

Dingesmere

The interpretation of dingesmere offered in Cavill et al. (2004) as ‘mere or wetland of the Thing’, Old English or Old Norse þing ‘assembly’ with OE mere ‘wetland’, relating probably to an area on the English side of the Dee estuary, is dismissed by Breeze. He writes ‘An attempt has been made to relate Dinges mere to the Wirral village of Thingwall …, with the Irish Sea or Mersey estuary being taken as the “sea of the thing or legal gathering,” despite the difficulties of phonology and sense’ (2018a: 70). He later calls the argument ‘far-fetched’ (2018a: 74). He does not specify in what particulars these supposed difficulties of phonology and sense may reside, or why even the garbled account he gives of the proposal is ‘far-fetched’.

The conundrum of the meaning of dinges mere can be solved, he suggests, by emending dinge- to dingle, as in the once-attested Middle English compound sea-dingle.

There is a far simpler etymology, suggested by the poem’s ‘over deep water’ immediately after [the phrase on dinges mere]. The Battle of Brunanburh is notorious for tautology, and so here. We can therefore emend to on dingles mere ‘into the sea of the dingle, onto the sea of the abyss.’ It will be a unique Old English attestation of the noun giving Middle and Modern English dingle. The thirteenth-century prose text Sawles Warde … thus says that God’s judgements are ‘secret, and deeper than any abyss of the sea (then eni sea-dingle).’ It echoes Psalm 36:6 (‘thy judgements are like a great deep’) … English dingle, now archaic and dialectal, is cited to explain Dingley in Northamptonshire … (also The Dingle, on Liverpool’s waterfront). (2018a: 74)

There are many possible objections to Breeze’s procedure here. A significant one is that no Anglo-Saxon or later scribe thought to make this emendation, even when the place-name element dingle was common from the thirteenth century. The variant spellings of the word in the Old English poem manuscripts are as follows:

Anglo-Saxon Chronicle ‘A’, Cambridge, Corpus Christ College, MS 173, fol. 26v: dinges mere

Anglo-Saxon Chronicle ‘B’, London, British Library, MS Cotton Tiberius A. vi, fol. 32r: dyngesmere

Anglo-Saxon Chronicle ‘C’, London, British Library, MS Cotton Tiberius B. i, fol. 141v: dinges mere

Anglo-Saxon Chronicle ‘D’, London, British Library, MS Cotton Tiberius B. iv, fol. 49v: dynigesmere

The spelling dinnesmere was reconstructed by Laurence Nowell from a now-lost copy of the Anglo-Saxon Chronicle *O, London, British Library, MS Cotton Otho B. xi, in the sixteenth century. Nowell’s collated spellings from the other manuscripts in the margin of London, British Library, MS Add. 43,703, fol. 229r, are Dyngesmere alias Dinnesmere. dyniges mere al. It is possible, as Campbell suggested, that the spelling in *O was ‘an alteration by the scribe’ (Campbell, 1938: 115), but searching for sense as he may have been, he did not light upon dingle as a possible emendation. Dingle, from the thirteenth century onwards, is a common place-name element. This evidence from the spellings in the poem manuscripts strongly suggests that the pre-Conquest scribes did not know the word dingle and the post-Conquest scribes did not think it was relevant.

The Oxford English Dictionary (Simpson et al., 2000–) reports dingle to be of uncertain etymology, with the earliest attestation from Sawles Warde, c.1240, as Breeze notes. In his edition of Sawles Warde, Wilson has the following to say about the term (sea) dingle:

Translating Latin abyssus, the meaning evidently being ‘depth, hollow;’ cf. the modern dialectal meaning ‘deep narrow cleft between hills.’ This is the first recorded occurrence of the word, which does not appear again in literature until used by Drayton who, perhaps significantly, was born in Warwickshire. There are rare examples of the word in place-names; it is known for example as a field name in Worcestershire; see la Dingle (1275), la Dyngle (1299) … It is also known as the name of a district in Liverpool, de Dingyll (1296) … Professor Mawer also suggests that the word may be the first element of Dingley in Northants … It is worth noting that with the possible exception of the Northants example the word seems to be definitely West Midland and may be another example of a specifically West Midland word. (Wilson, 1938: 74–75)

The suggestion here about the distribution of the element is confirmed by the range of field-names listed in A new dictionary of English field-names, in counties north–south from Cheshire to Gloucestershire, and as far east as Derbyshire (Cavill, 2018: 112). Leaving difficulties of chronology aside, this distribution of the element dingle would raise questions over its use by a West-Saxon poet, or as a name or description of a place existing in the far north-east.

The anomaly among the names noted by Wilson is Dingley in Northamptonshire, where the chronological difficulty (it is earlier than any other record, occurring in Domesday Book) and the normal West Midland distribution of the element coincide (Northamptonshire is further east than any other example). The uncertainty of the etymology given by Gover et al. (1933: 164, ‘this name may be a compound of dingle and leah’) is noted in recent scholarship: the Oxford English Dictionary refers to the identification as ‘uncertain and disputed’ (Simpson et al., 2000–). Watts prefers an alternative: ‘Possibly ‘Dynni’s clearing’, OE pers.n. Dynni + ing⁴ + lēah with the same loss of syllable as in king < OE cyning’ (Watts, 2004: 187). Mills gives ‘Possibly “woodland clearing with hollows”. OE lēah with ME dingle’ (Mills, 2003: 155, cited by Breeze, 2018a: 74, n. 81).

In previous work it has been shown that a wide range of undoubted þing names appear in spellings with initial D-, T- and F- as well as Th- (Pantos, 2004: 194; Cavill, 2018: 140 under Fingerfield, and 428 under Tinkfield). The initial consonant evolved to be pronounced differently by different communities, and French-trained scribes, and indeed some earlier and many later ones, struggled to hear and represent the sound /θ/. The evidence for this does not need to be repeated fully here. The D- spelling occurs particularly in Scandinavian-influenced areas. Scandinavian þing-völlr ‘place where the Thing met’ becomes Dingwall in Inverness and Dingbell in Northumberland; and in minor names we find Dingil gate Somerby, Leicestershire (1247, ‘road to the Thing hill’, hyll ‘hill’, gata ‘road’, Cox, 2002: 229); Dings, Hoby, Leicestershire (Abovedingesti, Underdingesti 1322, ‘upper/lower steep path to the Thing’, the Dinges 1601, stīg ‘steep path’, Cox, 2004: 115).

It seems at least possible that Dingley in Northamptonshire could be added to this list of þing names. As recorded by Gover et al. (1933: 164) the principal early spellings are:³³

Dinglei 1086, -leg 1241, Dyngle 1428

Tinglea 1086

Dingele, -y-, -leye, -lea 1166 et passim, Dinggele c. 1214, Dyngyle 1274

Dynelay 1348

Gover, Mawer and Stenton comment, ‘[t]he ground here is much broken, and this name may be a compound of dingle and leah… Hence, “leah marked by one or more valleys”’ (1933: 164). Though there are several pools in the parish and a gentle valley of the River Welland, there is nothing remotely like the Latin abyssus, or ‘deep narrow cleft between hills’, that the element dingle refers to in the welter of field-names such as we find in Shropshire: the land in Northamptonshire is partially wooded and undulating. Like the above-mentioned names which in Old English or Scandinavian begin with /θ/, English <Ð, Þ, Th>, Dingley appears with alternating forms D- or T- in Domesday Book 1086, here Dinglei and Tinglea. The early spellings do not readily indicate a *dingle-lēah. One might note the similarity of spellings for Thingley in Wiltshire, Thingele 1275, Tyngle 1289, Thyngele 1289 (Gover et al., 1939: 97) and the spellings for Dingley, Dingele and Dyngle. The etymology, ‘open woodland of the assembly’, Old English þing, lēah, is not in doubt for Thingley, and Old English or Scandinavian þing with Old English lēah seems plausible as an etymology for Dingley.

A related piece of supporting evidence is that Thing meeting-places tended to be on land near boundaries (Pantos, 2003). Dingley lies at the boundary of four administrative districts: it is on the boundary with Leicestershire and has lands in three different Northamptonshire districts, Stoke Hundred, Rothwell Hundred and Witchley Wapentake, in Domesday Book 1086 (Ryland et al., 1902: 322b, 335a, 386a, land in Stoke Hundred; 334a, land in Rothwelll Hundred; 350a, land in Witchley Wapentake). The boundary of English districts and the Scandinavian wapentake here may be particularly significant. Dingley is a very plausible location for a Thing meeting-place.

Breeze proposes that emendation of the Old English dinges mere to *dingles mere will solve the problem and at a stroke make redundant the argument constructed for the interpretation ‘wetland of the Thing’. But it would mean inserting a Middle English word into the Old English poem and ignoring that word’s dialectal distribution. The supporting evidence for the proposed *dingles mere from place-names, and in particular the etymology of Dingley, is much less secure than Breeze implies. The discussion has shown that the ‘difficulties of phonology and sense’ that Breeze mentions in relation to the Wirral interpretation of dinges mere, and which he appears only to have partially understood, are much more significant in relation to the proposed *dingles mere.

Conclusion

The only way properly to evaluate a hypothesis is to test it against the evidence. No detailed evidence has been presented here for the localizing of Brunanburh in a place other than Lanchester; other sites, such as Bromborough on the Wirral, have been suggested, but there is still vigorous debate on this issue. The main purpose here has been to examine the case Breeze makes. In this regard, Breeze’s Lanchester hypothesis cannot stand for three reasons. First, the Welsh evidence does not support the view that the references to Brune, Brun and -brun- must be to the river name Browney; the Browney was in fact called Brun and would not have taken an Old English weak oblique inflection, so could not be the Brunan- in Brunanburh. Second, the linguistic process by which Breeze posits Weondun to derive from *wenn-dūn is not plausibly shown to be possible in this orthographical or phonological environment. Finally, Breeze’s proposed emendation of the dinges mere expression in the Old English poem on the battle imports a Middle English word while necessitating change in four manuscripts. His speculations depend on superficial similarities of words, and they lack detailed investigation of the sources and cogent argumentation.