SOURCE: "The Date of Havelok," Anglia, Vol. 95, 1977, pp. 20-33.

[In the following essay, Jack takes issue with Herlint Meyer-Lindenberg's attempt to date Havelok the Dane more exactly, considering and rejecting all six of his arguments in turn.]

Though it would generally be accepted that the Middle English romance Havelok must have been in existence before 1300, there has been little agreement on any very precise date of composition; and indeed an agnostic view of the matter was taken in the edition by Skeat and Sisam, who concluded that it was impossible to determine how much before 1300 the poem may have been composed.¹ Nevertheless, there has been one recent attempt to assign a much more exact date to the poem, for it has been argued by Herlint Meyer-Lindenberg that Havelok must have been composed between the years 1203 and 1216². This has significant implications, as it provides a basis for the further conclusion that the English text was one of the sources of the French Lai d'Haveloc³; and it is bound to influence our conception of the kind of poem that Havelok is, since the arguments used by Meyer-Lindenberg require us to see Havelok as a literary reshaping of early thirteenth-century events, intended to be recognised as such by a contemporary audience. It is therefore a matter of more than passing importance that the arguments for such a date should be sound; yet it seems to me that they are very much open to doubt, for they involve many difficulties.

The case for dating Havelok between 1203 and 1216 rests on three primary arguments and three secondary ones⁴. The primary arguments are (1) that in Havelok's misfortunes at the hands of Godard there is a reflection of the fate of Arthur of Brittany, who died in 1203, (2) that one theme in Havelok is that of the inviolability of strictly hereditary succession, a theme of special appropriateness during the reign of King John, and (3) that the portrait of Athelwold in lines 27-105 of the poem also has particular appropriateness to the reign of John. The first of these arguments is intended to show that the earliest possible date of composition is 1203, the year of Arthur's death; and the second and third arguments are designed to establish that the latest possible date of composition is 1216, the last year of John's reign. In addition to these primary arguments there are three secondary ones, depending on (1) the allusions to Roxburgh as marking the northern boundary of England (lines 139, 265), (2) the reference to the creation of itinerant justices (lines 263-5), and (3) the use of the name Birkabeyn for Havelok's father. None of these three points indicates so specific a time as 1203-16, but they all, in Meyer-Lindenberg's view, suggest a date in either the late twelfth century or the early thirteenth, and to that extent support the assumption of early thirteenth-century composition.

The view that Havelok cannot be earlier than 1203 rests, as we have just noted, on the supposition that there are significant parallels between the fate of Arthur of Brittany and the treatment of Havelok by Godard in the early part of the romance. On the death of Richard I in 1199 the succession was disputed: John was accepted in England and Normandy, Eleanor (the queen-mother) in Aquitaine, and Arthur of Brittany in Anjou, Maine, and Touraine. Then in 1202, at Mirabeau in Poitou, John captured Arthur and his sister and imprisoned them. What then happened to Arthur is not known with certainty, but the probability is that he was killed in 1203 by John himself.⁵ Meyer-Lindenberg argues (pp. 92-6) that these historical events are reflected in the narrative of Havelok, forming the basis of the scene in which Godard imprisons Havelok and his two sisters, murders the sisters, and plans to have Havelok murdered by Grim (lines 408-544).

I find this implausible. It seems to me initially rather unlikely that the hero of a popular romance such as Havelok would be based on Arthur of Brittany, for Arthur was not a popular hero in England⁶. There are also more particular difficulties, for the narrative of the romance is in various ways unlike the historical events concerning Arthur of Brittany, and such similarities as there are chiefly involve features that are recognised motifs of popular literature and are at least as likely to owe their presence in Havelok to that source as to be reflections of historical events. To begin with the first of these points, we may notice that there are considerable divergences between the historical events surrounding Arthur of Brittany and the narrative of Havelok. Havelok is a young child when imprisoned by Godard (according to line 417 he is less than three years old); but Arthur of Brittany was at the time of his capture by John between fifteen and sixteen years of age—in Powicke's words, nearly a man in those days⁷—and he was engaged in besieging his grandmother at Mirabeau castle, hardly the activity of a young child and certainly without any parallel in Havelok. More generally, the circumstances of the historical succession were unlike those in the romance. In the romance Godard is simply a usurper and Havelok unquestionably the true heir, since he is the king's son; but Arthur was not the king's son and was by no means unambiguously the rightful heir, for John too had a claim to the throne and was in fact named by Richard as his successor. The dispute between John and Arthur was chiefly concerned with the succession to the French provinces and followed a course bearing little resemblance to the romance narrative. Following Richard's death in 1199 there was a brief period of conflict in France, ended by the Treaty of Le Goulet in 1200, whereby John was recognised by Philip of France as Richard's lawful heir, though Arthur was to hold Brittany as John's vassal. Then in 1202 war broke out between John and Philip, who hoped to take Normandy and replace John by Arthur in the other French provinces, and it was in the course of this conflict that John captured Arthur. The dispute between John and Arthur was thus spread out over several years, was complicated by the intervention of Philip of France, and was mainly concerned with the succession to only part of Richard's realm. But in the romance events are different and simpler. Godard, though ostensibly Havelok's guardian, plans to murder Havelok and usurp the throne; the events take place swiftly, involve no complicating intervention from any figure corresponding to Philip of France, and are concerned with the succession to the whole kingdom, rather than only part of it. So the historical events concerning Arthur of Brittany are in many ways unlike the narrative of Havelok; the romance does not seem to match those events in a sufficiently detailed and accurate way to make it reasonable to assume that it is based on them.

It might be maintained, however, that no great weight should be attached to, these differences between the romance and history. A romance, it might be said, is not a chronicle, and detailed historical fidelity is not to be expected; even an approximate resemblance in some essentials of the narrative may be enough to indicate a source in historical events and to alert an audience to an intended parallel to such events. This point may in principle be a fair one, but in the particular case of Havelok it is of doubtful importance, because the aspects of the poem that do resemble events concerning Arthur of Brittany mainly involve motifs of popular literature, and may with as much plausibility be derived from that source as from actual events. Thus imprisonment in a tower, the starvation of the captives, the intention to murder Havelok by drowning, and his rescue by a fisherman, all points cited by Meyer-Linden-berg (p. 96) as common to Havelok and to accounts of Arthur's fate that are found in thirteenth-century chronicles, are likewise all recognised motifs of folk-tale⁸. Moreover, there are other Middle English romances containing episodes that show resemblances to Godard's treatment of Havelok and his sisters. In William of Palerne a plot is hatched by which the king's brother, hoping to gain the throne, bribes two ladies to murder the true heir, the king's young son; but the child is carried off by a wolf and brought up in safety by a cowherd⁹. In King Horn the child Horn is set adrift on the open sea by Saracen pirates who have killed his father, the king, and seized the realm to which Horn is heir; and the pirates intend Horn to die¹⁰. In view of these parallels it seems likely that the comparable episode in Havelok, in which the usurper Godard attempts to kill the young heir to the throne, has its roots in a common story-pattern, rather than being a reflection of particular historical events. And in general the similarities that there are between Havelok and the events concerning Arthur of Brittany appear to me to be of little significance; they can so readily be explained as motifs of popular literature that it seems dangerous to rest any weight upon them as an indication of historicity in the romance.

I turn next to Meyer-Lindenberg's two arguments for 1216 as the latest possible date of composition of Havelok. One argument depends on the fact that Havelok asserts the principle of hereditary succession: Havelok and Goldborough, the hereditary rulers, recover their kingdoms, overthrowing the usurpers Godard and Godrich, and the poem gives some prominence to the theme of the rightful heir. Meyer-Lindenberg maintains (pp. 99-101) that this was intended by the poet to be a comment on contemporary events, and in particular that it was an assertion of the claims of strict heredity at a time—the reign of John—when the succession had been disputed. John's reign ended in 1216, and since the succession was not then disputed the theme of strictly hereditary succession would after 1216 no longer have immediate contemporary relevance. Therefore, Meyer-Lindenberg argues, 1216 is the latest possible date of composition for Havelok. This I find unconvincing, in part because the restoration of the rightful heir after the overthrow of a usurper is a narrative pattern that is found also in other Middle English romances: in King Horn Horn returns to his own kingdom, defeats the usurping Saracens, and regains his throne; in Gamelyn the inheritance properly due to Gamelyn is first appropriated by his eldest brother, then recovered by Gamelyn after the discomfiture of the brother; in Generydes the hero Generydes, heir to the kingdom of India, defeats the usurper Amelok and recovers his realm. The presence of a similar pattern in Havelok may therefore be explained as simply an example of a familiar narrative type; no special historical relevance need be assumed. Nor can it be accepted that the emphasis given in Havelok to the theme of the rightful heir indicates composition at a time when the succession had recently been disputed. This argument would be valid only if we were to assume that the poet's themes were a direct reflection of the time in which he lived, and that his work was shaped primarily by recent history. But if we allow, as surely we must, that literature need not be like this then the argument disappears, for an emphasis on the theme of the rightful heir can as easily be supposed to have its source in the poet's imagination as in his actual experience. It is surely possible to compose a poem asserting the claims of hereditary succession at a time when the succession does not happen to be in dispute.

Meyer-Lindenberg's other argument for 1216 as the latest possible date of composition is of a similar form; it depends on claiming (pp. 98-9) that the portrait of Athelwold in lines 27-105 must have been intended by the author to have special relevance to the reign of John, thus indicating that the poem must have been composed during John's reign, and therefore by 1216 at the latest. The argument involves three steps. (a) The portrait of Athelwold is disproportionately long, in view of his relatively small part in the narrative, and the only likely explanation of this is that the author was attempting to fit into his story a comment on his own times. (b) More particularly, in presenting a portrait of Athelwold as an ideal ruler the poet intended to give a measure of the deficiencies of the king reigning in his own time; therefore the poem was composed during troubled times. (c) This would suggest that the poem was composed either during the reign of John or at some time after 1234 in the reign of Henry III; and since for other reasons the reign of Henry III is unlikely as the time of composition, this implies a date at some point in the reign of John. There are several objections to this argument. Even if we accept that the portrait of Athelwold is disproportionately long, this may be explained in ways other than the one favoured by Meyer-Lindenberg: the poet may simply have been an erratic craftsman, who genuinely but mistakenly believed that the passage was appropriate in length; or he may just have liked composing passages in this vein. Simply from the premiss that the passage is disproportionately long we cannot reasonably deduce that it was intended as a tract for the times. Moreover, even if we did assume that the passage was intended to have specific contemporary relevance it would not follow that it must have relevance to a troubled reign; it seems at least as plausible to assume that it might be intended as a celebration of a just and peaceful reign¹¹. And finally, it is by no means clear that the passage is indeed disproportionately long; a good case can be made for the view that the passage is a fitting expression of the values that will be overthrown by the usurpers Godard and Godrich and re-established by Havelok, and that it thus provides a moral framework for the subsequent action of the poem¹². No sound argument for the dating of Havelok can be based on the portrait of Athelwold.

We may now consider the three secondary arguments presented by Meyer-Lindenberg (p. 99), arguments intended to show that the date of composition, though not necessarily in the precise period 1203-16, must fall at some point in the late twelfth century or the early thirteenth. The first of these arguments depends on the references at lines 139 and 265 to Roxburgh as marking the northern boundary of England. These were used in the past as a basis for dating the poem, but in view of the fact that Roxburgh castle first came into English hands in 1174 they have generally come to be seen as rather unhelpful, indicating only that the poem can hardly have been composed before that date¹³. Meyer-Lindenberg's argument, however, is based on the fact that Roxburgh castle, after passing to the English in 1174, was recovered by the Scots in 1189 and held by them from then until 1296. If we assume that a reference to Roxburgh as marking the northern boundary of England would be likely only at a time when the castle either was or had recently been in English hands, then this implies a date of composition either around the late twelfth century or at the very end of the thirteenth. Meyer-Lindenberg argues on other grounds (which I shall consider below) that Havelok certainly cannot be later than 1250; this therefore leaves a time around the late twelfth century as the most likely period for the composition of the poem, given the references to Roxburgh¹⁴.

This is not a good argument, for the historical assumptions on which it is based seem to be faulty. There is no real justification for the view that the holding of Roxburgh castle by the English between 1174 and 1189 underlies the references to it in Havelok as marking the Anglo-Scottish frontier. By the Treaty of Falaise in 1174 William of Scotland became the feudal vassal of the English king and ceded five castles to England; Roxburgh was one, and the others were Berwick, Edinburgh, Jedburgh, and Stirling. The effect of the treaty was not to alter the boundary of the two countries so that Roxburgh became part of England instead of Scotland; what the treaty meant was that the whole of Scotland was brought into subordination to England. Edinburgh castle was restored to Scotland in 1186; then in 1189 by a charter of quit-claim Richard I restored the castles of Berwick and Roxburgh to Scotland (Jedburgh and Stirling were not mentioned, and may in fact never have been garrisoned by the English)¹⁵. The importance of this is that it shows that when Roxburgh castle came into English hands it would not thereby have been seen as marking the northern boundary of England, for the whole Scottish kingdom was then in subjection to England, and Edinburgh, a castle much further north than Roxburgh, was also garrisoned by the English. Therefore there is no reason to suppose that the references to Roxburgh as marking the border of the kingdoms depend on the actual holding of the place by the English between 1174 and 1189. The correct explanation must be that the author of Havelok referred to Roxburgh as marking the border, not because it was (or had recently been) held by the English, but simply because it was a known town in the border area. It would be unrealistic to expect the author of Havelok, belonging (as the linguistic evidence shows) to the East Midlands of England, to have exact knowledge of the Anglo-Scottish border; there would certainly be no map available with the border conveniently inked in, and it is likely enough that he was relying on hearsay rather than first-hand knowledge. Moreover, Roxburgh is in fact very close to the border, for it is less than ten miles from the border that was effectively recognised in 1237 by the Treaty of York and had apparently been tacitly accepted for many years before then¹⁶; and so it would not be unreasonable to use Roxburgh to indicate more or less where the border lay. From this, however, it follows that no conclusion about the dating of Havelok can be based on the references to Roxburgh, for these would have been possible at any time in the thirteenth century or in the twelfth.

Before going on to the next of the secondary arguments we may consider Meyer-Lindenberg's reasons for thinking that, quite independently of any arguments for a date between 1203 and 1216, Havelok certainly cannot have been composed later than 1250. The basis for this assumption lies in the fact that there is a version of the Anglo-Norman Brut incorporating an an account of Havelok that displays certain features otherwise found only in the English Havelok.¹⁷ As Meyer-Lindenberg suggests (p. 97), it is reasonable to assume that these features were in fact derived from the English Havelok. Meyer-Lindenberg also assumes, following Brie, that the features in question were incorporated first into a version of Wace's Brut and then taken from there into the version of the Brut in which they actually appear; and since Brie suggested that this presumed version of Wace cannot have been much later than 1250¹⁸, this means that the English Havelok must have been in existence by that date. This appears to be too early.

Brie gave no reason for the date he ascribed to the version of Wace through which he assumed that the Havelok story reached the Brut with which we are concerned. All that can reasonably be concluded is that the English Havelok, since it was apparently the source of features in the Brut (either directly or through a version of Wace), was in existence by the time at which the text of the Brut in which these features actually appear was composed. The earliest manuscript of this version of the Brut (Paris, Bibliotheque Nationale, fonds francais 14640) is circa 1300¹⁹. The chronicle contained in this manuscript continues to the year 1272²⁰ and was presumably composed around that date. On the assumption that this chronicle derived features from the English Havelok, this implies that Havelok was in existence by about 1272; but it does not support a terminus ante quem of 1250.

The next of Meyer-Lindenberg's secondary arguments arises from the mention in lines 263-5 of itinerant justices. In these lines Godrich is described as creating justices who will travel the length and breadth of England, and Meyer-Lindenberg sees this as reflecting the historical fact of the establishment of itinerant justices by Henry II. I take it that Meyer-Lindenberg is here thinking particularly of the Assize of Clarendon (1166). This is hardly a satisfactory argument. First, even if we agree that the lines in question are a reflection of the Assize of Clarendon (and this may be doubted) it is far from obvious that this provides support for the assumption that Havelok was composed between the years 1203 and 1216. An allusion to an event of 1166 would show only that the text could not have been composed before that date; it would not indicate how much after 1166 the poem might have been composed, and would in principle be consistent with a date either before or after 1203-16. Moreover, it is not in fact clear that lines 263-5 do allude to the Assize of Clarendon. The assumption that they do depends on interpreting Iustises dede he maken newe (line 263) as meaning 'he had justices of a new kind established'. But line 263 may mean rather 'he had new justices appointed'—that is, that the incumbents were new, not the office itself.²¹ If this is the true meaning of line 263, then it removes the basis for claiming that Havelok shows knowledge of the establishment of itinerant justices. Lines 263-5 would then show only that the author knew of the existence of itinerant justices, and this would be compatible with a date of composition at almost any time from about 1130 onwards; for although the Assize of Clarendon gave instructions on which itinerant justices were to act, the practice of sending royal justices into the shires had existed before then, at least as early as 1130, and it continued during the thirteenth century²². For these reasons, then, lines 263-5 cannot be regarded as lending any firm support to the assumption of a date of composition between 1203 and 1216.

The last of Meyer-Lindenberg's secondary arguments depends on the name Birkabeyn, used of Havelok's father in the English text. This name is of Norse origin, evidently deriving from one applied to King Sverrir of Norway (1184-1202) and coming originally from the nickname Birkibeinar used of Sverrir's followers²³. Meyer-Lindenberg takes the use of the name in Havelok to be an indication that the date of composition is likely to be during or not long after the reign of Sverrir. But this argument has no force, since we do not know the process by which the author of Havelok learned the name Birkabeyn and therefore cannot assume that his knowledge of the name indicates proximity in date to the reign of Sverrir. If the name came to England through the fame of Sverrir, it could well have remained in currency for some time before being used in Havelok. And in any case we do not know whether the name came to England in that way, for it was used of other Scandinavians in the thirteenth century²⁴ and may not have reached England until considerably after Sverrir's time. So the use of the name Birkabeyn indicates nothing more precise than that the date of composition cannot have been before the time of Sverrir, the earliest date at which the name could have reached England from Scandinavia.

I would therefore suggest that there are insuperable difficulties in the arguments for dating Havelok between 1203 and 1216. More generally, the evidence does not seem to be sufficient for any dating of the poem aiming at that degree of precision, and a less definite conclusion is all that can reasonably be expected. Clearly the poem must have been composed by the early fourteenth century, the date of the Laud MS.; and since the textual and linguistic character of Havelok indicates that the Laud version was not the first copy that had been made of the poem, it is at least plausible that a gap of some years separated composition from the production of the Laud version. As we have already noted, moreover, the presence in a thirteenth-century version of the Brut of features apparently deriving from the English Havelok implies a probable date of composition no later than about 1272; and the use of the name Birkabeyn suggests composition no earlier than the late twelfth century. Taking these points together, therefore, we arrive at a date of composition at some time between the late twelfth century and about 1272. It might also be argued that the references at lines 1006 and 1179 to a parliament held at Lincoln imply that the poem does not antedate 1226, when, it has been suggested, parliament first met there²⁵. But this seems an unjustified conclusion. The poet could well have made parliament meet at Lincoln in his romance even though it had never done so in fact, for as Meyer-Lindenberg points out (p. 92) the narrative of Havelok requires Godrich to be at Lincoln and the holding of a parliament provides a convenient explanation of his presence there. We should not attach great weight to the particular term parlement used in Havelok, for it may simply mean 'council, assembly' rather than having any more technical sense. Certainly this more general meaning of the word 'parliament' is found in Middle English, as in Latin and French in the twelfth and thirteenth centuries²⁶; and the context in Havelok indicates that by parlement is meant an assembly of barons and bishops summoned to meet in the king's presence, a council of a kind already in existence before the thirteenth century²⁷. Such a council could meet in a variety of places²⁸, and it would be quite reasonable for the author of Havelok to depict one as meeting at Lincoln, whether or not he knew that it had done so in fact. From the evidence of datable references in the poem, therefore, the composition of Havelok cannot be more precisely fixed than within a time between the late twelfth century and circa 1272.

Notes

¹ W. W. Skeat and K. Sisam, The Lay of Havelok the Dane, 2nd ed., revised impression (Oxford, 1967), p. xxv. This also provides (pp. xxiii-xxv) a convenient summary of discussion of the date of Havelok. But it should be noted that Skeat and Sisam are in error when they state (p. xxiii) that Deutschbein "inclines to a date in the second half of the thirteenth century", for in fact Deutschbein suggests that the first version of Havelok was composed in the last quarter of the twelfth century (M. Deutschbein, Studien zur Sagengeschichte Englands [Cöthen, 1906], p. 165).

² "Zur Datierung des Havelok", Anglia, 86 (1968), 89-112.

³ See Meyer-Lindenberg, op. cit., pp. 102-12.

⁴ This division is my own.

⁵ See A. L. Poole, From Domesday Book to Magna Carta, 2nd ed. (Oxford, 1955), pp. 378, 381-3.

⁶ See F. M. Powicke, The Loss of Normandy, 2nd ed. (Manchester, 1961), p. 314.

⁷ Powicke, op. cit., p. 310.

⁸ See S. Thompson, Motif-Index of Folk-Literature, 2nd ed. (Bloomington and Copenhagen, 1955-58), motifs K958, R41.2, R51.1, R131.4.

⁹William of Palerne, ed. W. W. Skeat, EETS, ES 1 (London, 1867), pp. 1-12 (this includes an opening section taken from the French text, for the beginning of the English text is missing).

¹⁰King Horn, ed. J. Hall (Oxford, 1901), lines 1-116 (MS. C).

¹¹ This was in fact the assumption made by Deutschbein, who suggested (op. cit., p. 165) that if a historical basis were to be sought for the portrait of Athelwold it might reasonably be found in Henry II.

¹² For discussion see D. Mehl, The Middle English Romances of the Thirteenth and Fourteenth Centuries (London, 1968), p. 170, and J. Weiss, Speculum, 44 (1969), 249f.

¹³ See Skeat and Sisam, op. cit., p. xxiii.

¹⁴ In Meyer-Lindenberg's article the argument is not explicitly set out in this form; but this must, I think, be the reasoning lying behind the use of the references to Roxburgh.

¹⁵ See W. Croft Dickinson, Scotldnd from the Earliest Times to 1603, 2nd ed. (London, 1965), p. 78; D. W. Hunter Marshall, Scottish Historical Review, 25 (1927-28), 20-23.

¹⁶ See G. W. S. Barrow, The Kingdom of the Scots (London, 1973), ch. 4; this was earlier published in Northern History, 1 (1966), 21-42.

¹⁷ See F. Brie, "Zum Fortleben der Haveloksage", Englische Studien, 35 (1905), 359-71.

¹⁸ Brie, op. cit., p. 364.

¹⁹ Brie, op. cit., pp. 361-2; J. Vising, Anglo-Norman Language and Literature (London, 1923), p. 98.

²⁰ Brie, op. cit., p. 360.

²¹ The use of the verb make in the sense 'appoint' is recorded from the twelfth century; see the Middle English Dictionary, s. v. maken, v. 1, sense 6.

²² See Poole, op. cit., pp. 399f. But see also W. T. Reedy, "The Origins of the General Eyre in the Reign of Henry I", Speculum, 41 (1966), 688-724, for discussion of differences between the practice during the reign of Henry I and that following the Assize of Clarendon.

²³ See L. A. Hibbard, Mediaeval Romance in England, 2nd ed. (New York, 1960), p. 104; E. Björkman, Nordische Personennamen in England, Studien zur englischen Philologie, 37 (Halle, 1910), s. v. Birkabein.

²⁴ See E. H. Lind, Norsk-isländska personbinamn fran medeltiden (Uppsala, 1920-21), s. v. Birkibein.

²⁵ See W. van der Gaaf, "Parliaments Held at Lincoln", Englische Studien, 32 (1903), 319-20. This assembly was not in fact officially referred to as a parliament; the earliest known official use of the term is in 1236, referring to an assembly that was to take place in 1237 (see H. G. Richardson and G. 0. Sayles, "The Earliest Known Official Use of the Term 'Parliament'", English Historical Review, 82 [1967], 747-50).

²⁶ See OED, s. v. parliament, sb. 1, sense 2, and H. G. Richardson, "The Origins of Parliament", Transactions of the Royal Historical Society, 4th Series, 11 (1928), 137-83.

²⁷ See G. O. Sayles, The King's Parliament of England (London, 1975), ch. 2, esp. p. 32.

²⁸ In the reign of Henry II there were, for instance, meetings of council at Northampton, Winchester, and London (Gesta Regis Henrici Secundi Benedicti Abbatis, ed. W. W. Stubbs, Rolls Series, 49 (London, 1867), I, 107, 118-9, 138-9).