THE ENGLISH Downloaded from

HISTORICAL REVIEW http://ehr.oxfordjournals.org/

NO. LXVIIL—OCTOBER 1902

Dr. Guest and the English Conquest of South Britain at University of Iowa Libraries/Serials Acquisitions on March 17, 2015

T7BEEMAN, in a letter to the Spectator : upon the death of Dr. J. Guest, refers to ' the wonderful series of discourses—more than one, happily, but still too few—made before successive meet- ings of the Archaeological Institute, in which the progress of the English conquest in the southern part of Britain was first really set out' by that writer. His ' almost morbid' love of certainty and accuracy is mentioned, and he is said to have 'belonged to that class of revealers of truth who bring order out of chaos and light out of darkness, who do their work at the first blow, so that it needs not be done again,' and that what little he has left behind him ' is all of the purest gold.' Freeman had previously professed himself ' in all essential points an unreserved follower of that illustrious scholar.'2 Green was hardly less enthusiastic in his acceptance of the results of these essays, which largely appear in his maps in the ' Making of England.' The joint editors of the ' Origines Celticae,' in which these essays were republished in 1883, were Bishop Stubbs and Mr. Deedes, and the editorial intro- duction is almost as unbounded in its praise as Freeman. It is no wonder that these essays, thus endorsed by three great histo- rians, should be received without question by local historians and other writers. Moreover their conclusions are accepted by Dr. William Bright in his ' Early English Church History.' As they

1 Reprinted in Origines Celticae, 1. p. xii. 1 History of tlie Norman Conquest, i. 9. n. 1. Compare also Life and Letters, i. 336. VOL. XVII.—NO. LXVIII. 8 S 626 DR. GUEST AND THE Oct. are thus becoming gradually woven into the fabric of our accepted history, it is time that they and the methods by which they were reached were submitted to critical examination. These papers are, briefly put, an attempt to write history with- out documents. Half a dozen words in the Chronicle recording a battle or two is sufficient material for Guest to draught a map showing the relative positions of the English and Britons in the Downloaded from sixth century. In order to do this he founds arguments upon the physical nature of the country, local names, and ancient fortifica- tions, and the arguments are eked out by references to exceedingly late chroniclers. Great stretches of country are filled up with woodlands, and these are assumed to have been so impassable that http://ehr.oxfordjournals.org/ the English invaders were compelled to leave them in the hands of the Britons. But in fact we have no trustworthy evidence as to the extent of land under trees in the fifth and sixth centuries. Guest takes as his basis the forests of feudal times. It is not until the thirteenth century that we obtain perambulations of the boundaries of these forests, and even then it is impossible to say how much was woodland. Guest nowhere shows any appre- at University of Iowa Libraries/Serials Acquisitions on March 17, 2015 ciation of the fact that these feudal forests were not exclusively or even necessarily woodland. A forest was a district of land subject to the harsh forest laws, comprising within its bounds wood, heaths, and other wild or poor lands, towns, villages, and cultivated land. In many parts the woodlands were nothing more than the strips of wood belonging to the manors within the forest, and in such parts there was little or nothing to distinguish land within the forest from that outside. Guest, however, assumes that all the land within these forests was under trees, and he increases the area of the forests occasionally so as to fill in the space between rivers. Even if it were possible to believe that the outlines thus obtained represented the state of the country in the sixth century, it would still be difficult to believe that all these woodlands were such impassable barriers to semi-barbarous invaders as they are assumed to have been. Green, with his love of the picturesque, has greedily seized upon this theory of impassable woodland belts, and he has found a host of imitators. It has become common and indeed irksome to read how the English invader was kept at bay by a belt of woodland, until he finally did what he apparently might have done earlier—that is, he ' dashes through the brake' and surprises the too confiding Briton. Ancient are a still more uncertain foundation for the reconstruction of the lost history of the sixth century. It is impossible to establish the date of the erection of these works in the vast majority of cases, or even to feel any confidence in the ascription of them to Britons, Saxons, Danes, and so on, in which antiquaries have so much delighted. To illustrate the risk of writing 1902 ENGLISH CONQUEST OF SOUTH BRITAIN 627 history from early fortifications it is not necessary to go beyond the • Origines Celticae.' GueBt adopted the view of Stukeley that the Bokerly Dykes, part of the series of entrenchments upon which that credulous writer bestowed the name of Belgic Dykes, were the work of the Belgae before the invasion of the Eomans, and that they marked their boundaries. Guest builds largely upon this conclusion in his paper on the date of Stonehenge and on the Downloaded from West-Saxon conquest. Yet the patient excavations of General Pitt-Rivers have shown by the irrefragable evidence of Roman coins that these fortifications were not erected until the fifth century of the Christian era. Local names are scarcely more satisfactory. It is not until http://ehr.oxfordjournals.org/ centuries after the dates dealt with by Guest that we get any documentary proof of the existence of these name3, and the great majority of them are not recorded until after the Norman Conquest. This long interval of time leaves open the door for many possibilities in connexion with them. A great part is played in Guest's paperB by names apparently compounded of the names of the English and

Welsh. But even if all these names had an ethnic reference it at University of Iowa Libraries/Serials Acquisitions on March 17, 2015 is obvious that they may refer to events of later date than those with which he unreservedly connected them. It may be doubted whether the West-Saxons called themselves English or Angles in the sixth century. Guest endeavours to meet this objection by drawing attention to the use of both names in the laws of King Ine.3 But these are of considerably later date, and, as they are known only through the medium of King Alfred's laws, there is a possibility that these denominations are not due to Ine. Of the names cited by Guest one only is an undoubted compound of the name of the Angles,4 whilst some of them have clearly no reference whatever to them. Guest does not seem to have realised that there was no genitive plural in -8 in Old English, and, as there is no proof that the singular of an ethnic name could be used for the plural, as it is in modern English by a sort of personification, we may erase from the list all names involving a genitive in -8. It is difficult to believe that the invaders distinguished trivial features or settlements on portions only of their temporary frontiers by their national name, and that such names survived after the frontier had advanced by many successive stages out of sight. Such names as English Bicknor and Welsh Bicknor are of later origin, and are not parallel to the uses assumed by Guest. It is possible that some of these names in Engle- are derived from men named Engel, for the word was in use for forming personal names among most of the Germanic peoples, although there is no instance except in local

* Origines Celticae, ii. 190. 1 That is, Englefield, near Beading, which is mentioned in the chronicle under 871 as ' Englafeld,' and is translated by Florence of Worcester as ' Anglorom Campus.' s s 2 628 DR. GUEST AND THE Oct. names of its use in England before the Franco-Norman Ingelrams, Ingchics, and Ingelberts arrived. It is in non-Anglian districts that we should expect to meet with it, just as it is in Anglian districts that names compounded with Sexe (' Saxon') occur. The case is even less clear in regard to Wealh, ' Welshman,' for that was extensively used in Old English personal namea, as in continental

Germanic names, and we have clear proofs of the use of the Downloaded from single-stem hypocoristic form Wealh in Old English names. This word also meant a ' serf,' and hence in some cases local names compounded with it may represent settlements of serfs, just as Charlton and Carlton derive their names from dwellings of ' churls.'

But with Guest local names compounded with Wealh are made to http://ehr.oxfordjournals.org/ refer to the Welsh, and the date of their imposition is restricted to the period with which he was dealing. Scarcely less important than these names in his scheme are words which he considers to mean ' boundary,' and in like manner all possibility of the boundaries being other than those of the expanding West-Saxon kingdom in the sixth century is ignored. It is not certain that any one of the names used by him does mean ' boundary.' Thus mere is taken by him to represent at University of Iowa Libraries/Serials Acquisitions on March 17, 2015 the O.E. gemtere, which is never used absolutely in the charters. He passes over the great probability that the word may represent the O.E. mere, ' lake.' It is certainly remarkable that in three out of four of the places in Domesday called Mere, Mera a lake still exists, or is known to have existed.* From his maps it is evident that Guest also considered oare to mean ' boundary,' but the O.E. ora really meant ' bank, shore.' He explains by impossible philology other words as meaning ' boundary,' and even the Norman Devises is brought into his service. He assumes that the memory that this district was in the sixth century a borderland between the West Saxons and. the British lingered on locally long enough to confer this name upon the , and Devizes accordingly is marked as an evidence of the sixth-century boundary. But the greatest part of all is played by the numerous Grim's Ditches in the south of England, which he regards as meaning ' boundary .' The evidence adduced by him in support of this explanation is wholly illusory. It is nothing but a comparison with ' what appear to be its correlatives'—namely ' the Icelandic greini [sic], " to separate," with which must be connected the German grcinze [sic], " a boundary;" the Swedish ren and German rain, " the boundary of a field," and perhaps the A.S. rima, " a margin." 'G Now these words are philo- logically unrelated to one another; the German grenze is a loan word from Slavonic; the Icelandic greina is derived from grein,

5 These are Meare in Somerset,.Mere in Rosthorn, co. Chester, and Meer in Meer- town, co. Stafford. The fourth is Mere, co. Wilts. " Origines Celticae, ii. 149, note 1. 1902 ENGLISH CONQUEST OF SOUTH BRITAIN 629 which does not mean 'boundary' but 'branch,' and hence ' division;' and the Swedish ren and German rain mean primarily a grass balk in an open field. The only other proof vouchsafed is the occurrence of Grimsetene gemcere in a Worcester text,7 and the conclusion that it must have meant the ' inhabitants of the boundary.'8 Assuming

that grim was here a natural feature in the perambulation, it does Downloaded from not follow that it was a boundary.9 But there remains the great philological difficulty of the presence of the personal genitive -es in Grimes die.10 The O.E. form of composition of this hypothetical grim, a' boundary,' and die would have been grim-die. The genitive

would be as unusual and unnecessary in such a compound as it http://ehr.oxfordjournals.org/ would be in modern English ' border-land,' ' march-land,' or ' frontier district.' Quite as unsatisfactory as the philology is the use by Guest of passages in the chroniclers. Freeman has described Thierry and Palgrave as never making a statement for which they could hot give chapter and verse in some form, ' but both of them were too apt to catch at any statement which seemed at all to support their several theories, without always stopping to reflect whether such, at University of Iowa Libraries/Serials Acquisitions on March 17, 2015 statements came from contemporary chronicles or charters or from careless and ill-informed compilers three or four centuries later.' u This description applies to Guest's methods, with the difference that the interval between the period he is dealing with and the chroniclers cited by him is much greater, and that many of them are using the fictitious history of Geoffrey of Monmouth.

' Cartularium Saxonicum, iii. 534, .10, A.D. 969, relating to Witley, co. Worcester. 6 Orig. Celt. ii. 150. He also alludes to Graeme's Dyke, the name of the Wall of Antoninus. Graeme or Graham can hardly be the same word as ' Grim.' • As the adjoining parish of Grimley appears as ' Grimanleah' (Cart. Sax. ii. 55, 17, and ' Grimanleh' in Domesday, i. 173 b, col. 2, by a rare copying of the English spelling of the original return), and as a ' Grimman-hill' is also mentioned in or near it (Heming, ed. Hearne, p. 257; Domesday, i. 172b, col. 2, 'Gremanhil'), it is obvious, from the hypocoristic gemination of the m, that ' Grimma' was an early owner of the land, and it is therefore probable that ' Grimsetene' is miscopied from ' Grimes-setena.' ' Grimma ' and ' Grim ' might be used indifferently; hence we get a ' Grimes-hyll' at Hallow, which can hardly be other than the ' Grimman- hyl' cited above. In the boundaries given in Cart. Sax. i. 496, 33, there is probably an omission of a substantive ' hyll' or ' leah ' after ' Griman.' With the Danish in- vaders Grim was a somewhat common name, and there can be little doubt that many local names in Grim—such as Grimsby, Grimston, and Grimsthorpe—are derived from owners bearing this name. As Danish housecarls were settled even in southern counties, the possibility of a local name in Grims- being derived from the Danish name cannot be excluded. 10 ' Grimes die ' must either be derived from a man named Grim, in which case we are met with the difficulty of accounting for its wide diffusion, or from some unrecorded mythical or divine figure, who was perhaps connected with ditches. If there were any evidence that Woden was also known among the pagan English as Grim, as he was among the Norsemen, we might conclude that ' Grimes-dlc ' was merely the equivalent of' Wodnes-dlc' The Devil's Dyke, occasionally met with as the name of ditches, may be compared. The works were considered as superhuman. 11 Norman Conquest, i. p. xv. DR. GUEST AND THE Oct. Although Guest realised that Henry of Huntingdon amplified from his own imagination the barren annals derived from the ' Old English Chronicle,'12 he frequently cites him without any reserve.13 We are told that the dates in Boger of Wendover ' may be of service,' u and they are advanced as trustworthy, and important 15

conclusions are based upon them. Yet there can be no question Downloaded from that these dates are entirely worthless. The St. Albans compila- tion, the work here referred to as 'Wendover,' who, like Matthew of Paris, merely embodied it in his work, adopted the absurd stories of Geoffrey of Monmouth as genuine history, and worked them in with the material derived from English sources. It is from the http://ehr.oxfordjournals.org/ latter that most of the dates, which are absent in Geoffrey's work, are derived either directly or by inference. Some of them are clearly inventions by the compiler in order to piece his materials together in chronological order. Worse even than this is the use of Thomas Eudborne's ' History of Winchester Priory,' a fifteenth - century compilation that contains almost as many inventions as Geoffrey's more famous work.16 Whether Eudborne derived these

from the authorities that he cites or added them himself is of little at University of Iowa Libraries/Serials Acquisitions on March 17, 2015 moment. Besides William of Malmesbury, the ' Polychronicon,' and the ' Flores Historiarum' (that is, another form of the St. Albans compilation) he quotes mysterious writers, such as Girardus Cornubiensis, ' De Gestis Eegum West-Saxonum '17 and ' De Gestis Britonum,'18 Vigilantius,' De Basilica Sancti Petri (Wintoniensis),'19 Moracius, antiquarian Britannomm conscriptor,™ and Volensis, ' De Laude Britonum.'2I The trail of Geoffrey of Monmouth is discernible in most of these.22 Little better is the fourteenth- century ' Eulogium Historiarum,' upon which Guest founds most important conclusions.23 Even the sixteenth-century Italians, Biondo

!2 Orig. Celt. ii. 164. 11 Ibid. pp. 179, 180, 184. Cf. p. 285. " Ibid. ii. 164. 15 Ibid. pp. 174, 194. '• Ibid. pp. 180, 187. " Wharton, Anglia Sacra, i. 189, 193, 201, 204, 206, 227. Girard seems to have been a real person, and not, as suggested by Bishop Tanner, a perverted form of Giraldus Cambrensis. In the Liber de Hyda, p. Ill, he is quoted for the restoration of the university of Cambridge by Edward the Elder, whioh story sufficiently stamps his character. It is not surprising after this to find that he relates as history the duel of Guy of Warwick with the Danish giant Colbrand in the reign of jEthelstan (ibid. p. 118; Hearne, Chronicon de Dunstaple, p. 147). '" Aitglia Sacra, pp. 180, 186. " Ibid. pp. 180, 186, 189, 192, 199, 203, 205, 206, 210, 215, 223, 226. He is cited by the compiler of the Liber de Hyda, pp. 7, 21. " Anglia Sacra, i. 182. ll Ibid. i. 183. a Budborne, however, critically examines aud rejects some of Geoffrey's tale? about Arthur (p. 189). The Florentius cited by him from a Florarium Historiale (pp. 190, 192, 193) is clearly not Florence of Worcester. This unknown Florentius is cited by the Liber de Hyda, p. 11, for the foundation of the University of Cambridge in 394 B.C. Budborne alludes to a Concordantia Historiarum, or Chronicarum, or Chranicantium, which was arranged in lexicographical order. a See below, p. 639. 1902 ENGLISH CONQUEST OF SOUTH BRITAIN 681 (of Forli) and Polydore Vergil, are used as authorities for the history of Britain in the sixth century.24 By their side the ' Flemish chronicler Sigebert'H is a respectable authority, as he wrote only some six hundred years after the events for which he is vouched. An example of Guest's methods of dealing with historical evidence may be found in his paper on the ' Early English Settle- Downloaded from

11 Orig. Celt. ii. 184. " Ibid. Sigebert of Gembloux died in 1112. In piecing together the materials for his universal chronicle he displayed considerable carelessness in regard to dates. Many of those given by him are, as his editor, L. C. Bethmann, has remarked (apud Pertz, Monum. Germ. Hist., Scriptores, vi. 276-7), purely arbitrary, and his dates prior to the eleventh century are of scarcely any authority. Yet he is cited by Guest http://ehr.oxfordjournals.org/ to Bupport his identification of Ambrosius Aurelianus with the Natanleod, whose death is recorded in the Chronicle under 508. Guest takes the date assigned by Wendover for the accession of Ambrosius as 464 (whereas it is really 465), alters it to 463, adduces Sigebert to prove that Ambrosius reigned forty-five years, deducts this period from 508, and so arrives again at 463. He then tells us that ' the coincidence of dates seems almost to demonstrate the identity of Natanleod and Ambrosius ' (p. 184). All this is so carefully arranged as to have the air of demonstration, but upon exami- nation the whole case crumbles to pieces. . The passage in Wendover is taken word for word from Geoffrey of Monmouth, who as usual gives no dates (lib. viii. caps. 1^3). The date is therefore an invention of the twelfth-century St. Albans compiler. The at University of Iowa Libraries/Serials Acquisitions on March 17, 2015 passage from Sigebert occurs under 446, a date that he evidently obtained by de- ducting three years from 449, the year given by Beda, Hist. Eccl. i. c. 15. This deduction was made on the strength of Beda's ' ante paucos sane adventus eorum [sdl. Saxonum] annos,' o. 17, which is intended to give the date of the mission of Gennanus and Lupus to Britain. Sigebert begins under 446 with this mission, gives an account of the Alleluia victory, from Beda's c. 20, and then continues: ' Britanni, qui per tot annos expositi fuerant praedae et hostium ludibrio, Anglos invitant a Germania. Angli Britanniam veniunt, eamque ab hostibus tutam faciunt esse; non multo post earn sibi vindicant. Ambrosium Aurelianum sibi Britanni regem statuunt, eiusque ductu per annos xlv. vario belli eventu contra Anglos con- fiigunt' (M.G.H. vi. 309, 9). This is certainly derived from Beda's fifteenth and six- teenth chapters, and contains echoes of his phraseology. The statement that the Britons made Ambrosius king is a not unnatural deduction from Beda's reference to his 'occisis parentibus regium nomen et insigne ferentibus' (c. 16), while the ' vario belli eventu' is obviously the ' ex eo tempore nunc cives, nunc hostes vincebant' of Beda, who copied it from Gildas. Sigebert has clearly taken his forty-five years of Ambrosius's reign from the sentence immediately following: ' usque ad annum obsessionis Badonici montis, quando non minimas eisdem hostibus strages dabant, quadragesimo circiter et quarto anno adventus eorum in Britanniam.' Sigebert records under 491, on the basis of this forty-five years, the complete conquest of Britain by the Angles. It is now plain that his authority for assigning a reign of this length to Ambrosius is nothing but a misunderstanding of Beda. Guest's other arguments for the identity of Ambrosius and Natanleod are even weaker than this. They consist of the suggestion that' Natan-leod' is compounded of O.E.' leod,' which in poetry means ' prince,' and must have been a word familiar to the authors of the chronicles and to the transcribers, and the Old Welsh ' nawt' (' sanctuary'), which he says, merely to suit his argument, would have been taken over into O.E. as ' Nat-e,' gen. ' Nat-an ;' that the sanctuary was a monastery at Amesbury, for whose existence there is no better authority than the Welsh triads, wherein it is called ' the choir of Ambrosias;' and that the ' lord of the sanctuary ' was therefore Ambrosius. It is satisfactory to note that Green rejected this tissue of absurdities. According to Gildas, cc. 25, 26, the date of Ambrosius was some time before the battle of Mons Badonicus, which, as we shall see, he places within a year or two of 500. Ambrosius can, therefore, hardly be the Natanleod of 508. 682 DR. GUEST AND THE Oct. ments in South Britain,' which is accompanied with a map pur- porting to show the land occupied by the Britons and the Saxons in the year 520. This is the date that he fixes for the battle of Mons Badonicus. Our sole record of this battle is the obscure passage in Gildas, c. 26, usque ad annum ohsessionis Badonici Montis . . . quique quadragesimus quartus ut novi orditur annas mense iam emenso, qui et vieae nativitatis est. This Beda read as Downloaded from meaning forty-four years from the arrival of the Saxons, which would fix the date about 498. If it refers to the time that elapsed between Gildas's birth and the date when he wrote this work, as Mommsen suggests,26 there is some uncertainty as to the date; but, as King Maglocunus, who is upbraided by Gildas, died in 547, http://ehr.oxfordjournals.org/ according to the usual interpretation of the dates in the ' Annales Cambriae,'!7 it would seem that Gildas was born about the year 500. Consequently such evidence as we have points to a date considerably earlier than 520 for the of Mons Badonicus. Guest, after alluding to a treaty between Arthur and Cerdic, which he accepts on the authority of so worthless a writer as Thomas 28

Rudborne, quotes Higden's ' Polychronicon ' for its terms. This at University of Iowa Libraries/Serials Acquisitions on March 17, 2015 fourteenth-century writer tells us that Arthur, about twenty-six years after the arrival of Cerdic, gave to him out of weariness Hampshire and Somerset, and bestowed upon them the name of Wessex.29 Guest says nothing of the doubtful way in which Higden, who displays really remarkable critical powers for his time in dealing with the fables of Geoffrey of Monmouth,:o introduces this passage. It is obvious that the unnamed chroniclers mentioned by Higden were engaged upon the hopeless task of reconciling the stories of Geoffrey and the Arthurian accretions with the English authorities; that the story of Arthur's resigning these two counties out of ennui is merely a clumsy attempt to harmonise the uncon- querable power ascribed to him with the statement that Cerdic became king of Wessex in his time; and that the date is founded upon the entry in the Chronicle recording the commencement of the realm of Wessex in 519, which is twenty-four years after the date ascribed for their arrival in Britain. Guest also quotes Eoger of Wendover (i.e. the St. Albans compilation) for the year 520 for Mons Badonicus,31 but the evidence of the compilation is discounted by the fact that it is hereabouts relying almost exclu-

58 Chroiiica Minora, iii. 8 (' Monumenta Germaniae Historica '). 37 These annals commence with an ' annus primus,' which is not defined, but is usually taken to have been 444. M Orig. Celt. ii. 187. At p. 193 the Isle of Wight is held to be ' one of those wasted districts which were yielded up to Cerdic after the battle of Mount Badon.' As this was a British victory, according to Gildas, the sole original authority for it, it is difficult to realise the magnanimity of the conquerors here imagined. 50 Higden, v. 330 (Rolls series). K Ibid. p. 334. " Orig. Celt. ii. 187. 1902 ENGLISH CONQUEST OF SOUTH BRITAIN 683 sively upon Geoffrey of Monmouth, who practically gives no dates. One of his few dates is that of the death of Arthur in 542.32 As the very late Welsh chronicle, the ' Brut y Tywysogion,' says that Arthur died twenty years after Mons Badonicus, Guest does not scruple to cite its testimony for fixing the date of that battle in 3:i

52O, and upon this evidence he thinks that ' we have ground for Downloaded from believing that this celebrated battle did really take place in the year following that in which Cerdic became king.' These ' authori- ties,' with Nennius and the ' Annales Cambriae,' a tenth-century work largely founded upon the work ascribed to him,34 are held to

be sufficient to prove that Arthur was the leader of the Britons at http://ehr.oxfordjournals.org/ Mons Badonicus.36 On the authority of Wendover his accession is accepted as taking place in 516.36 Not only does the date of Mons Badonicus obtained in this unscientific manner form the basis of his map and paper, but both map and paper very largely turn upon the identification of Mons Badonicus with Badbury Eings, in Dorset. We have no indication of the site of this battle; so little indeed that Skene made out 37 a plausible case for locating it in the lowlands of Scotland. at University of Iowa Libraries/Serials Acquisitions on March 17, 2015 Guest decides that Mons Badonicus was ' so called from the baths (Welsh badon) in its neighbourhood,' and he identifies it with Badbury, although ' no Eoman baths have yet been discovered in the neighbourhood.'38 He was aware that Badbury occurs in the Chronicle under 901 as cet Baddan-byiig, but he attempts to derive this from O.E. bceft, ' bath,' by remarking that 'in Hampshire and its neighbourhood the final and medial th was often pronounced d; hence, if there were Eoman baths near Badbury, the locality would probably be called Badde.' The explanation looks suspiciously like one invented for the occasion. It is quite immaterial what the Hampshire dialect did in later times; what is wanted is proof that

a Hist. Britonum, xi. c. 2. M Orig. Celt. ii. 187. 31 A. de la Borderie, in Revue Celtique, vi. 2, points out that the compiler of the Annales has transferred to Mons Badonicus, which in Nennius is the twelfth of Arthur's battles, some of the details in Nennius that relate to the eighth battle. As Mr. Egerton Phillimore remarks, this ' creates a doubt as to which of these two battles the annalist found placed (or reason to place) under 516 ' (Owen's Description of Pem- brokeshire, i. 229, n. 3). It also deprives his testimony as to the leadership of Arthur at Mons Badonicus of an; value it might otherwise have possessed. *> Orig. Celt. ii. 186. " Ibid. p. 194. " Four Ancient Books of Wales (Edinburgh, 1868), i. 57. He identifies it with Bouden Hill, near Linlithgow, which, however, occurs in 1710 in a quotation given by him as ' Buden,' so that it would appear to have had ' a,' not ' a,' in the root syllable. ra Orig. Celt. ii. 189. The alleged modern Welsh ' baddon ' seems to be derived from Bath, O.E. ' at BaiSon,' possibly conditioned by its identification with Mons Badonicus, and is not a current word for ' bath.' In form it is an irregular plural of ' badd,' which is a loan-word from the English ' bath.' For this information I am indebted to my friend Professor Rhys. Neither word, therefore, could have existed, early enough in Welsh to have given name to Mons Badonicus, which only super- ficially resembles the late O.E. dat> pi. ' Ba'San,'' BaSon,' earlier ' BaSura.' 634 DR. GUEST AND THE Oct. such a change had occurred in the tenth century. No such proof is attainable, and it is impossible to believe, in face of the evidence of the very great number of West-Saxon texts, that such a change occurred. Guest does not seem to have had any perception that the double d meant anything more in Old English than it does in modern English. It alone is a very serious obstacle to the deriva- tion of Baddan-byrig from Mons Badoniais. The word is, in fact, Downloaded from derived from Badda, which was a man's name. Precisely the same form cet Baddan-byrig occurs in a contemporary charter of 944 as the name of Badby, co. Northampton.39 Two other important factors in constructing this map of 520 http://ehr.oxfordjournals.org/ are the capture of Winchester by the English in the year of Cerdic's landing, which has no better evidence than an absurd calculation in Eudborne,40 and the identification of Cerdicesleah with Chearsley, co. Bucks, which involves the transference of the name to Bernwood Forest. No valid reason exists for rejecting the usual location of Cerdices-leah somewhere near Cerdices-ford, Charford, in Hants, and the name of Chearsley was evidently not 41

Cerdices-leah. Guest leaves Eynsham, Aylesbury, and Bensington at University of Iowa Libraries/Serials Acquisitions on March 17, 2015 in the hands of the Britons because their capture is entered in the Chronicle under 571, and for similar reasons Salisbury andBarbury are also in British possession. The evidence by which he per- suades himself that he has identified the British chief assumed to be resident at Amesbury is quite inadequate.42 But he makes

™ Cart. Sax. ii. 540, 8, 31; 542, 13. A Baddan- in Warwickshire occurs in a charter of 704-9 ; ibid. i. 179, 5, 27. 40 Orig. Celt. ii. 180, 185. 41 From the Domesday form ' Cerdes-lai' we may conclude that the O.E. form was 'Ceardes-leah,' and this is supported by innumerable later instances. In none of them is there anything corresponding to the ' ic ' of ' Ceardices-leah.' 41 Orig. Celt. ii. 183, 195. The ' Caradoc' of the lists of British cities is identified with Amesbury (which was not a Roman city, the first requirement for the identification) on the authority of Caradoc of Llancarvan, who wrote in the twelfth century, and of so credulous a person as Owen Pughe, the compiler of the Welsh dictionary that has caused so much mischief through its imaginary words and meanings (such as ' andred,' which duly appears in Guest in several places as meaning ' uninhabited,' whereas it is merely a figment to explain ' Anderida; ' ' gwent,'' a champaign,' which was also accepted by Guest, and has been still further popularised by Green; and ' caint,' which is also said to mean ' open country,' whereas it is merely a guess as to the meaning of ' Cantium,' although it has been adopted by Guest and Green). Then Caradoc, whom Guest identifies with Caradawg Vreichvas, a Welsh chieftain of uncertain date (if not entirely mythical), is held to be the ' prince who gave the name of Caer Caradoc to the stronghold which adjoins to Amesbury.' From this the unsuspecting reader might conclude that there was a camp bearing this name near Amesbury. The camp referred to is now known as Vespasian's Camp, a figment of Stukeley's. Its older name, Fripsbury, is the only one known to Camden and Aubrey. But the existence of this Caer Caradog at Amesbury, adduced in this misleading way, depends entirely upon the views of 'the best informed Welsh scholars,' who considered the name to apply to the earthworks immediately adjoining Amesbury (p. 265). This is not an unfair specimen of many other arguments in these extraordinary essays. 1902 ENGLISH CONQUEST OF SOUTH BRITAIN 635 Amesbury British on the strength of this, although he is compelled to restrict the brave British chieftain to a small angle of land that it would be hardly possible to defend. The only evidence discernible beyond these points and the lines of Grim's Ditches, Aves Ditch, and Grimsbury Castle is the occur- rence of a Wealas-huth (= O.E. *Weales-hy$) and Weala geat in Downloaded from a very late and corrupt copy of the boundaries of Chertsey.43 The former embodies the genitive singular, and can therefore at once be excised. The latter is explained as 'Welshmen's road,' although geat had never this meaning in O.E. Englefield by Egham and Englemoor or Englemere, the name of a lake near Ascot, are further evidences of the boundary. In default of early forms of http://ehr.oxfordjournals.org/ the names their derivation from the Angles is not free from doubt. Another point is obtained by the impossible explanation of Theale as a form of ' dole,' and the explanation of the latter as ' boundary.'44 A Weala brucg at Brimpton affords another indica- tion of the boundary,45 and another is found in the very doubtful assumption that Mare Bidge, in Upton Nervet, is derived from

mere, ' boundary.' Wallingford is also held, although with some at University of Iowa Libraries/Serials Acquisitions on March 17, 2015 hesitation, to derive its name from the Britons; but it is quite certain that it does not.40 Inglesham, in Lechlade, is treated as a record of the English, but, as its O.E. form was Incgenes-ham," derived from a man's name, we may safely cross this off the map. Then ' Ingleford or Hungerford' and Inglewood, near Hungerford, are adduced.48 From this region we get no indications in local names

41 Cart. Sax. i. 57, 27, 30. 14 Orig. Celt. ii. 191, where we are told that' in the old Hampshire dialect th was often substituted for initial d, as thar, there, thoth, &c, for dare, dear, den, doth, &c. Theale seems to*be the provincial pronunciation of dole, a boundary.' The philo- logist will not need any comments to see the absurdity of this. It is one of many similar instances of perversion of meanings to fit theory, for ' dole' never meant ' boundary,' but ' part,' ' division.' 41 Cart. Sax. ii. 559, 29. " It occurs as ' Wealingaford' in the Chronicle and elsewhere. The suffix ' -inga' is out of place in an ethnic name, and there can be little doubt that' Wealinga- ford ' derives its name from a man named ' Wealh.' A ' Wealingaham,' a similar formation, occurs in the boundaries of Knoyle, co. Wilts, in Cart. Sax. iii. 16, 33. This is to the east of the Welsh boundary assigned hereabouts by Guest. 47 It occurs in 965-71 as ' Incgenaeshain' (Cart. Sax. iii. 432, 9) and in 995 as ' Inggenesham' (Codex Diplomaticus, vi. 131, 15, 25). Cf. also Liber de Hyda, p. 255, where it is copied ' Igeneshamme' from the will of Ealdorman /Ethelmeer, who died in 982 (cf. p. 218 and Chronicle C). This is instructive, for the name occurs in Domesday, the Pipe Bolls, and later records as ' Inglesham.' It is possible that a similar corruption has occurred in some of the other names in ' Ingles.' 4S Whatever may be the source of Camden's statement that Hungerford was anciently called Ingleford, it is certainly wrong, for ' Hungreford' appears in a fine of 1199 (Fines, sive Pedes Finium, p. 101), in the Close Bolls for 1205 (p. 29 a), in the Charter Bolls for 1215 (p. 218 a), &c. An Everard de Hungreford, who derived his name from this place, is mentioned in the Pipe Bolls for 11 Henry II and succeeding years. There is probably some confusion with the Domesday ' Ingeflot,' ' Inglefol,' and ' Ingheflot,' which occurs as ' Hingelflod' in 1199 (Rot. Chart, p. 14 a), and is now 636 DR. GUEST AND THE Oct. beyond Oare and Fisherton-de-la-Mere (where the surname is that of a post-Norman family who owned it),49 and Britford and Sher- field English. The boundary is placed halfway between these two latter places, evidently on the assumption that the English goes back to the sixth century. But it is only another instance of Guest being misled by the addition of a much later lord's surname to the village name.60 Downloaded from We may now turn to Guest's paper on the ' Welsh and Eng- lish Boundaries after the Battle of Bath, A.D. 577.' The map appended to it represents the Britons as in possession of Somerset south of the Axe,61 and of a long strip of land ' some fifty miles long by fourteen broad,' projecting from Wookey to Braden Forest. It http://ehr.oxfordjournals.org/ is, in fact, the valley of the Frome and the Avon, between the Wilt- shire Downs and Bath. This tongue of land is supposed to be impassable woodland, and the English conquest consequently pro- ceeds along two sides of a triangle formed by Koman roads, the base of which lies between Bath and Marlborough. Guest re- marks that 'it may be thought strange that the Welsh should

retain' this tongue of land ' in the midst of a country that had at University of Iowa Libraries/Serials Acquisitions on March 17, 2015

Inglewood in Kintbury. It is probable that ' * Engel-flod' was the name of the brook by Inglewood, and that ' Engel' has nothing whatever to do with the Angles, for ' Angil,' ' Engil' occur as continental German river names (see Forstemann, Altdeutsclies Nainenbuch, ii. 84). To this also may be referred the Wiltshire Ingleburn, which Guest unhesitatingly connects with the Angles, and the Devonshire Englebourne. "> Cf. Patent Rolls, 1390, p. 335, gTant by Philip son and heir of John de la Mare of the advowson of Fisherton. The family name seems to be derived from ' mere ' (lake), either through the Norman cognate ' mare ' or the English ' mere.' In the Nomina Vilkirum in 1316 the village appears as Fisherton Babington. M In the Nomina Vilkirum the lady of Shirefelde (which has no addition) is returned as Alina la Engleys. 51 This is because he believes the stories that Glastonbury was a British monas- tery. He states that it was known ' by the Welsh, as early as the sixth century, as Ynis Wytrin— that is, the glass island '—on the authority of William of Malmesbury's Gesta Rcgum, i. cc. 27-8, and he accepts the charter of 601 given in this author's De Antiquitate Qlastoniemis Ecclesiae, ed. Hearne, p. 48, and endeavours to identify the nameless king of Damnonia, who is alleged to have granted by it the ' five hides ' (!) called ' Yniswitrin' (Orig. Celt. ii. 65, 270). At the former page, by aid of the non- existent 'Anglo-Saxon thing, " a council, a place of meeting," ' he manages to explain the ' Glffistings ' of ' Glfestingabyrig' as ' the buildings used for public purposes, whose walls were vitrified.' The ' Yniswitrin' is an obviously late and erroneous translation of ' Glsestinga,' which has been associated with O.E. ' glses ' (glass), and arises from the attempts to connect Glastonbury with the Arthurian cycle (cf. Ehys, The Arthurian Legend, p. 328 sqq.) William of Malmesbury, in his Oesta Pontificum, § 91, ascribes the first foundation of the monastery to King Ine, thus agreeing with the Chronicle. Unfortunately William allowed himself to become an active agent in the dissemination of the lying history of this pretended British monastery, whose founder was Joseph of Arimathea. For the growth of this mythical history see Stubbs, Memorials of St. Dunstan, p. lxxviii, n. 3, and p. lxxx sqq. This ridiculous ' Yniswitrin' appears in the map of Britain in 597 given in the first volume of Freeman's Norman Conquest, which faithfully reproduces the results of Guest's paper on the boundaries in 577. 1902 ENGLISH CONQUEST OF SOUTH BRITAIN 637 become English territory.'M In justification he alleges ' that everything tends to show us that these anomalies were of frequent occurrence in the territorial arrangements of the period;' but no proof of this is given. Not only Bath but the Mendips are assigned to the West-Saxons, and this is ascribed to Ceawlin's insisting upon retaining the lead mines in some unrecorded peace negotiations. So enamoured is Guest of his imaginary history Downloaded from that he says, 'If it were conceded to him [Ceawlin], no line of demarcation could be drawn which would more neatly or more effectually secure his object than the one we have been describing.'M Yet the map shows a Roman road proceeding along the base of the triangle, a feature that lessens the impassability of the land assigned http://ehr.oxfordjournals.org/ to the Britons. Guest does not explain why the heathen monarch should attach so much importance to the lead mines. He refers to the great value in which lead was held ' in times immediately preceding those we are treating of,'64 which is not to the point. After referring to the battle of Fethanleah, which he persuaded himself was so far north as Cheshire,55 he comes to the conclusion

a Orig. Celt. ii. 255. Green, Making of England, p. 129, describes ' this wedge of at University of Iowa Libraries/Serials Acquisitions on March 17, 2015 unconquered ground' as running up into West-Saxon territory ' for the next hundred years.' Tet we find Cenwealh fighting at Bradford-on-Avon in 652, in the heart of this imaginary wedge of unconquered land. a Orig. Celt. ii. 249. " Ibid. " Ibid. pp. 288,314. His identification of Fethanleah with Faddiley has been accepted by Freeman and Green and others, but there is no case for it. We have no indication whatever of the site of this battle. The basis of Guest's identification is the Welsh elegy on the death of a Kyndilan at or near Pengwern (Shrewsbury). No indica- tion of date exists beyond its ascription to Llywarch Hen. who is assumed to have lived in the sixth century. The text and ascription depend upon thirteenth-century MSS. (see Skene, Four Ancient Books of Wales, p. 2). Sharon Turner had already connected this elegy with the king Condidan, whose death with two other Welsh ' kings' is recorded in the Chronicle at the battle of Dyrham in 577. But Guest reiects this identifica- tion (Orig. Celt. ii. 283) on the ground that Condidan and Kyndilan are said to have been slain at different places. This contention is surely fatal to his whole paper on the English conquest of the Severn valley in the sixth century, for with the severance of Kyndilan from the Condidan in the Chronicle disappears all reason for ascribing the sack of Pengwern to Ceawlin. Yet it is regarded by Guest as the work of Ceawlin solely on the basis of the entry in the Chronicle, which, how- ever, does mention the sack of many towns by the West-Saxon king. Guest having, as he thinks, proved that Ceawlin was at Shrewsbury, secures a very necessary point for bridging over the distance between Dyrham and Cheshire. This, method of founding one guess upon another, and then usiDg the second to prove the truth of the first, is a common device in these papers. Guest concludes that Ceawlin met with a reverse at Fethanleah, which at p. 289 is supposed to have been the result of his officers pushing on into the Vale Royal without ordeis. To support his view he '. would willingly receive' the details of this battle given by Henry of Huntingdon. Now it is obvious thai, if Fethanleah was the site of a defeat inflicted upon the West-Saxons that caused their king to withdraw in anger to his own kingdom, the name must have been one confeired by the raiders upon the site of their defeat, and therefore one that would not be taken op and preserved by the victorious Britons. Moreover the Britons must have handed over this English name to the later Anglian conquerors of Cheshire. Is this at all probable? The name has every appearance of being a temporary one, for it is, to all appearances, a compound 638 DR. GUEST AND THE Oct. that the site of the battle of Wodnesbeorh, recorded in the Chronicle under 591, was ' beyond all question Wanborough, near Swindon.'56 This forms one of the two ' fixed points' " upon which the whole paper' hinges. He contrives to identify Wodnesburh [sic] with Wanborough by pointing out that the s might disappear, and that then • Wanborough would be the modern corruption of Wodenburgh [sic], just as Wansdyke is of Wodnes-dic,'58 &c. But he had only Downloaded from to turn to the Domesday to see that Wanborough is not a modern form. It there occurs as Wemberge, and that or Wanberge is the form it usually appears under in medieval records. A corruption from Wodnes to Wan so early as the eleventh century is impossible. Any doubt that Wanborough might perchance be an O.E. Wodnes- http://ehr.oxfordjournals.org/ burh must disappear before the O.E. form of the name, which has been overlooked owing to Mr. Birch's erroneous reference of the charter in which it occurs to Hampshire. The charter is a grant of land at Wenbeorgan to Winchester priory,59 which appears in Domesday as the owner of Wanborough.60 The identification is made certain by the occurrence in the boundaries of the Icknield Way, a continuation of the welJ-known Berkshire road, in which county it is recorded at an early date. When to this may be added at University of Iowa Libraries/Serials Acquisitions on March 17, 2015 that a Wiltshire Wodnesbeorh is also recorded in two charters that have in like manner been erroneously assigned to Hampshire by Eemble and Birch,61 we may say that' beyond all question ' Wodnes-

of ' feSa,' gen.' feftan' (' a band or troop of foot soldiers'), and ' leah,'. which has pro- bably its older meaning of ' wood.' Moreover, although Guest is so confident about the agreement between Faddiley and Fethanleah, they do not correspond accurately. Nothing need be said of the blundering manner in which he tries to account for the change of "S' to ' d' (p. 288), as the real difficulty is with the root vowel. Faddiley occurs in 1271 as ' Fadile' (Catalogue of Ancient Deeds in the Public Record Office, B 3624) and as ' Fadilegh,' ' Fadelegh,' within the next few years (Twenty-sixth Report of the Deputy-Keeper, appendix, pp. 49, 53, 95, 96, 100,

the castellum at Caer Bladon (Malmesbury). It is the old story—that http://ehr.oxfordjournals.org/ contempt of enemies which has ever been characteristic of our country- men, and which, if it often leads tnem to victory, has sometimes entailed upon them very humiliating reverses.65 This breached fortress is as unsubstantial as Guest's English and Welsh frontiers. The fourteenth-century writer knew only of borough in the sense of town; hence the imaginary city.

He was also familiar with castles in towns, and transported them at University of Iowa Libraries/Serials Acquisitions on March 17, 2015 backwards to the times of Meildulph. Guest regards ' borough ' as meaning fortress (O.E. burh). But as a matter of fact Brokenborough does not contain the word burh at all, but beorge, the dative singular of beorh, a 'hill,'66 which developed through the Middle English berewe to borough in compound names in this part of England. On the basis of this passage and the mention of Dunwallo Molmutius Guest proceeds to identify the Welsh king who fought against Ceawlin.67 After exposing the unsubstantial nature of the foundations it is not necessary to examine the super- structure. The evidence of the local names upon which these boundaries are founded is as weak as in the other paper.68 The 05 Orig. Celt. ii. 252, n. 3. " Cf. Cart. Sax. iii. 91, 22, ' Brokene-beregge,' and 94, 9, Codex Diplomatics, v. 166, 33, ' Brokeneberge,1 which are all from the Malmesbury chartulary. In Domesday the name appears as ' Brocheneberge' (i. 67 col. 1) and represents an O.E. 'set Brocenan beorge,' 'at the broken hill.' Examples of this name may be found in Cart. Sax. i. 262, 18; ii. 106, 8; 245. 31, 32 ; iii. 48, 13 ; 143, 13; 297, 28. The meaning is explained in a contemporary charter of 944, where the boundaries proceed ' to the broken hill (to brocennn beorge), which is there cloven asunder (toclofen)' (Cart. Sax. ii. 542, 11). 07 Orig. Celt. ii. 257 SJJ. 63 They are Englishcombe and Englishbateh in Englishcombe, near Bath, which he connects with the English, 'for Englishcombe is mentioned in Domesday, which was compiled before surnames were known in England' (ii. 249). But the Domesday form is ' Engliscome' (i. 88 b, col. 2), and it occurs in the Exon Domesday, p. 136, as ' Ingeliscuma,' and this form is represented by the modern collateral' Ingles Combe.1 The name therefore seems to be an O.E.' *Engles-cumb,' and to be derived from an owner's name, ' Engel.' Such a compound as English-comb, meaning a combe in the possession of the English (West-Saxons) in the sixth century or even much later, is 1902 ENGLISH CONQUEST OF SOUTH BRITAIN 641 Wodnes-beorh cited above was near to Wodnes-dic and Wodnes-denu. We have in this strong proof of the influence of the Woden cultus in this district.69 We can hardly be wrong in referring the imposi- tion of these names to pre-Christian days. That is, they were probably conferred before the conversion of the first West-Saxon king in 684, within forty-three years of Ceawlin's fight at Wodnesbeorh. May we not therefore conclude that this Wiltshire Downloaded from Wodnesbeorh was the site of this fight ? By so doing we sweep away whatever may remain of Guest's edifice of guesses, for we find Ceawlin fighting, and apparently against West-Saxons,70 some distance to- the south-west of the position assigned for the battle by Guest, and on the south side of the Wansdyk and close to the http://ehr.oxfordjournals.org/ Roman road from Cunetio (near Marlborough) to Bath—in other words, at the base of the triangle. Guest's lucubrations belong to a class that must be painfully familiar to those whose lot brings them into close connexion with local histories and the proceedings of antiquarian societies, local or otherwise. The methods are old. A theory is formed;

anything that can be considered evidence in its support is eagerly at University of Iowa Libraries/Serials Acquisitions on March 17, 2015 seized, while the eyes are closed to anything that conflicts with it; the evidence is loosely interpreted or perverted ; a guess is furtively slipped in, and is used as a basis for another, which is then held to prove it. It may be true that we must be content with a low standard of proof in an age so dark as the one with which Guest was dealing, and that the student may say to himself, in the words of Livy, in rebus tarn antiquis si quae similia veris sint pro veris accipiantur, satis habeam, and that opinions may reasonably differ as to the degrees of verisimilitude. But there can be no justification for the thoroughly uncritical manner in which Guest uses medieval chronicles, and his philology is equally indefensible.

without any parallel in Old English. Englishcombe and Englishbatch seem to be com- paratively recent corruptions of Inglescombe and Inglesbatch. Guest then takes the village of ' Merkbury,' which he explains as ' the borough or fortress of the march.' But the village is named Marksbury, and, as it occurs with this gen. sing, in Domesday as ' Mereesberie ' (i. 90 b, col. 2), it is certain that it cannot have this meaning—first because the gen. would not appear in such a compound of ' mearc,' and secondly because the gen. sing, of that word is ' mearce,' the gen. pi. ' mearca' or ' mearcena.' Marksbury can only be derived from a masc. personal name. ' Mere' is taken as undoubtedly meaning a boundary (ii. 255), and in the map Leigh Delamere, which derives its surnames from its lords, is evidently regarded as recording the sixth- century boundary. The use of Devizes has already been dealt with (p. 628). *> After all there is something to be said for Lappenberg's view, which Guest rejects, that Wodnes-beorh was a temple of Woden. His cultus seems to have been specially connected with hills both in England and on the continent. See Grimm, Deutsche Mythologie, p. 139; E. Mogk in Paul's Grundriss der germanischen Philologie, i. 1073. A Norse example occurs in England in ' Othenesberg,' the older name of Bosebery Topping, co. York, which occurs early in the twelfth century in the Guisborough Chartulary (Surtees Society), i. 4, &c. 70 Guest adopts Malmesbury's statement that Welsh and English conspired against Ceawlin (p. 243). VOL. XVII.—NO. LXVIII. T T 642 DR. GUEST AND THE ENGLISH CONQUEST Oct. Yet there still exists a tendency among antiquaries and local historians to treat these results as unimpeachable. It is too much to expect that this caveat will suflBce to prevent the production of imitations of Guest's papers, but there can be little doubt that the ultimate verdict of history upon such efforts will be that of perhaps the greatest critic of the nineteenth century: Possumus aliorum incertis opinionibus novas addere, sed nemo unquavi r&v ahwardv Downloaded from ipau et quaerens id quod reperiri non potest, fecit ant factuns est operae pretium.71 W. H. STEVENSON. " C. G. Cobet, Observationes Critieae et Palaeographicae ad Dionysii SaUcar- nassensis Antiquitates Romanas (, 1877), p. z. http://ehr.oxfordjournals.org/ at University of Iowa Libraries/Serials Acquisitions on March 17, 2015