Again, it very greatly limits the probability of Malory's having elsewhere worked with a free hand, inventing and rearranging, when we find, as we have done, that numerous small details, hitherto ascribed to him, are faithful reproductions of his source. We are justified in cherishing very serious doubts as to the originality of any marked deviation from the traditional version of an adventure which we may find in his compilation.
These arguments, of course, apply most strongly to his version of the Charrette adventure, the problem of source of which, so far as Malory is concerned, is absolutely unaffected by the evidence we have collected. This alone is certain, there is no proof whatever that he knew anything of the first part of the Lancelot romance, his treatment of the Lady of the Lake seems to show that he was absolutely ignorant of it. He was not in the habit of departing unnecessarily from his source, his variations as a rule are slight, and their motive can generally be detected; when, therefore, we find him giving an entirely different account of the abduction of Guinevere from that given elsewhere, the probabilities are all in favour of his reproducing a separate source, and all against his original invention. So far as the matter stands in the light of the latest evidence, the question remains unsolved, with a decided balance in favour of the theory advanced by M. Gaston Paris, and against that advocated by Professor Foerster.193
Leaving the question of Malory, what may we hold to be the result of this examination on the problem of the Queste itself? Is the form in which we possess it practically the original form, or are we to postulate a series of successive redactions? I think that every one who has carefully studied the variants given above must have been struck by the fact that in no case is the question one involving a variety of incident or even an alteration in sequence. It is the same story in every case, told in the same order, in the same words, only certain copies give a fuller and more coherent version than others. In fact, as I said above, the variations are the variations of the copyist, not of the compiler. The one point in which we may postulate either omission or addition, i.e. the greater or less fulness, the presence or absence, of the 'improving' sections, is precisely a point in which we might expect a copyist of a more or less didactic turn of mind to assert himself; it was so easy to expand or to contract such passages. And it is a curious feature that precisely in those versions in which the story, as a whole, is the best told (D. L., 1533, and in a minor degree M.), we find the edifying passages in their shortest forms; while Q., the text of which as compared with the others is decidedly poor, gives them at the greatest length.
Of any previous redaction of Galahad's adventures there is no trace; there are no lengthy interpolations as in the Conte del Graal MSS.; there is no conflict, such as we find in other romances, between an earlier and later form; in sundry passages we have allusions to unrelated adventures: we are told that the heroes ride so many days, weeks, or years, and meet with many and strange adventures, but in no copy do we find any hint of what these adventures may have been; yet had there existed an earlier and fuller form, some fragments of it must surely have been preserved.
And this argument becomes more convincing the more closely we look into it. Above we have compared four versions of the Queste (five if we include W.), but one of these, Dr. Furnivall's edition, does not represent one ms. only, but is founded on a critical collation of two, and contains a specimen of the opening columns of twelve mss. of the Bibliothèque Nationale; while Dr. Sommer states that he has examined four other versions and found that, saving details of style, all agree in incident and sequence. We may therefore take it as certain that one of the four variants represents at least five MSS., while scholars of standing assure us of the practical identity of sixteen more!
Now, side by side with these Queste versions, we have compared four versions of the prose Lancelot, and of these four no two agree perfectly throughout, and all differ from the summary given by M. Paulin Paris.
D. L. and 1533, which on the whole correspond well with each other, yet have their distinctive differences, i.e. D. L. contains adventures not related by 1533; M., while on one side condensing arbitrarily, on the other gives two adventures known to neither of the first; and S. omits an important section altogether. The summary in Romans de la Table Ronde, while agreeing on the whole with the two first, deviates from both in the later sections.194
The practical identity of all the versions of a romance transmitted in so large a number of MSS. as the Queste is, I believe, unique in the Arthurian cycle. Such a phenomenon, for it is nothing less, can, I think, only be explained in one way: there was but one version of the story, and that version took shape, not at a period when oral transmission was the rule, but at a later date, when the story could at once find expression in literary form. I do not believe that any story, the earlier stages of which were developed orally, is ever, when committed to writing, found so entirely free from variants.195
Can we decide what special form of the Perceval Queste the Galahad variant was intended to supersede? I think not: it is noticeable that the writer never gives any adventure which finds an exact parallel in the older romances, yet he not only knew the Perceval story, but knew it in various forms. The allusions in Book XIV., though slight, are remarkably instructive: he knew that Perceval was the son of a widow, and that his mother died of grief at his departure (Chrétien, Wolfram, Didot Perceval); that in his wanderings in search of the Grail he came to the dwelling of a female recluse, who proved to be a near relative (only related by Wolfram); that he has a sister (Didot Perceval, and Perceval li Gallois). Thus in these few allusions he is in touch with the whole cycle of Perceval romance! When, therefore, we find that he never elsewhere assigns to Perceval any of the adventures traditionally connected with him, but gives him a new series which are duplicated elsewhere, one can only conclude that it is done of set purpose.
Of the parallels given above, the existence of the sister appears to me to be the most important, judging from the prominence of the rôle here assigned to her. She only appears in the two forms of the Perceval Queste which show traces of having formed part of a cycle; and inasmuch as Perceval li Gallois represents the mother as living to see her son return, and regain his kingdom, the correspondence is closer with the Didot Perceval, but the question can hardly be settled.
As a Grail romance the Queste is extremely poor. The utter confusion of the writer as to the identity of the Fisher King and Maimed King; the inter-relation of Grail Winner, owner of Grail Castle, Fisher King and Maimed King; the neglect of the most obvious conditions of the quest, such as ignorance on the part of the predestined Grail Winner; his giving proof of identity by fulfilment of a test; the inaccessibility of the Grail castle to all but the elect knight—all show a most extraordinary carelessness on his part, were he intending to write a Grail romance pure and simple. Ignorance we cannot postulate. He knows too much about Perceval not to know more about the Grail! It is evident throughout that the main anxiety of the author is to keep himself in touch with the Lancelot rather than with the Grail tradition. He is extremely careful to introduce references to that portion of the Lancelot story with which he is familiar; to explain that the adventures foreshadowed in Grand S. Graal and Lancelot have been really fulfilled, and so long as he can demonstrate his hero to be a worthy upholder of the glories of the race of King Ban, he cares very little if he fails to fulfil the necessary conditions of the original Grail Winner. This latter may know from the first what the Grail is, where it is, his own predestined relation to it, his final winning it may be reduced to an absurdity by the presence of eleven or twelve others all equally worthy of beholding the sacred talisman, but that matters nothing to the author; he has contrived to bring the Grail into more or less harmony with the Lancelot legend; he has crowned the most popular of Arthur's knights with reflected glory as father of the Grail Winner, he has put the last touch to the evolution of the Lancelot legend, and in so doing he has achieved the task which he set himself to perform. The Queste is in all essential features not a Grail but a Lancelot romance, and as such primarily it should be judged.
CHAPTER XI
THE MORT ARTUR
This, the final section of the Lancelot cycle, offers less opportunity for criticism. The versions of D. L. and 1533, though still closely in accord with each other, differ much less from the summary given by Dr. Sommer, and show less affinity with Malory.196
So far as Malory is concerned I differ from Dr. Sommer, who says that 'he cannot have derived his account from the prose Lancelot.'197 On the contrary I think there is little doubt that Malory had the latter portion of the Lancelot before him, but dislocated it by the introduction of the Charrette and Urre of Hungary198 episodes, which he most probably knew in an independent form; though of course, as I have suggested above, it is quite possible that some Lancelot MSS. may have included the latter. But considering the clear proof that the English compiler was following an Agravain MS. for the earlier part of his Lancelot adventures, and that he includes the Astolat and Patryse stories, which are a part of the ordinary Mort Artur199 section, I see no reason to doubt that his Lancelot MS. represented all the latter part of the cycle (as we know he had, and followed, an alternative version of the M. A. proper). I have carefully compared both D. L. and 1533 with the abstract given in the Studies, and give the following as the most important of the variants, but I should like to make it clearly understood, both as regards this and the previous sections, that the instances I quote by no means represent all the points of contact and departure to be noted between the different versions. I have many others in my notes, and a critical edition will certainly very much strengthen the case I have here stated in outline.
As we have before noted, D. L. and 1533 agree against S. in incorporating with the Queste the passage generally given as the opening of M. A. Otherwise all three versions are in practical agreement as regards the events leading up to the tournament at Winchester. D. L. does not mention Hector when Lancelot inquires on which side his kinsmen are fighting, but only Bohort and Lionel. 1533 agrees here with S.
According to D. L. and 1533 Gawain and Gaheriet take no part in the fighting at Arthur's desire: he fears they may fight with Lancelot, and ill-will arise from it. S. does not mention this, so I cannot say if it be in the 1513 edition or not.
S. says, 'The people think the two knights' (Lancelot and his comrade) 'cannot be the sons of the lord of the castle of Escalot.' This does not agree with the other versions: the people think they are the lord's sons at first; then Gawain says, one of them cannot be. D. L.:
'Ende man waende daer wel dat Lanceloet
Ware een vanden broderen van Scaerloet.'—ll. 851-2.
and Gawain proceeds to say, 'This knight with the red sleeve is not he whom I thought, no one ever saw such valour by one of the "Kinder van Scaerloet."' Arthur asks what knight he may be. D. L. makes Gawain say simply he does not know, 'but he is certainly a good knight'; while 1533 goes on to add 'if Lancelot had not been left at Kamalot he would have said that this was he.' This does not at all agree with S.
Both 1533 and D. L. agree against S. in saying Lancelot's wounds will take six (not seven) weeks to heal.
When Gawain and Gaheriet follow Lancelot, S. says they meet a wounded knight; in 1533 the knight is dead. D. L. omits the incident.
When Gawain returns to Arthur, S. represents the king as saying 'it was not the first time he took trouble without results, nor will it be the last.' 1533 and D. L. here add 'through that knight,' which is evidently correct.
S. simply says the second tourney is fixed at Tanebor, 'du lundi dapres en ung moys'; whereas the other versions carefully specify the wherabouts of this place, 'dat een casteel es, Staende in den inganc van Nortgales.' D. L. spells it 'Caneborch.'
Again, according to S., Lancelot, unable to go to the tourney, sends greeting to the queen and Gawain, 'from the knight who wore the red sleeve'; whereas D. L. and 1533 say 'the knight who won the tournament at Winchester' and make no mention of the sleeve, which, considering the relations between Lancelot and Guinevere, seems to me the better version. Neither do these mention that Guinevere tries to persuade Bohort to return to Camelot.
When Gawain comes to Escalot S. represents him as admiring the maiden's beauty and envying the knight 'with the red sleeve.' 1533 says, more correctly, 'the knight who wins her love'; he has not yet learned to whom the sleeve belonged.
In the account of what happens after Gawain's return to court, and Guinevere's learning the truth, all the versions agree on the whole, and it is noticeable that M., though making Bohort a more energetic defender of his cousin's good faith, yet correctly reproduces all the main features of incident and speech. I think any one comparing his version closely with two or three others can hardly fail to come to the conclusion that it is the prose Lancelot and no other account he is reproducing.
According to S. Lancelot's kinsmen only remain for a week at court; according to 1533 and D. L. it is 'that week and the next.'
When they leave the court on the second occasion after the tournament of Tanebor, neither D. L. nor 1533 say (as S. does) that the queen tries to persuade Bohort to remain, though they agree in making her regret his departure.
After Lancelot's return to court when Bohort lectures Guinevere on the mischief done by women, with reference to David, Solomon, etc., D. L. omits the reference to Tristan, while 1533 amplifies it by saying 'it is not five years since Tristan died for love of Iseult.'
D. L. omits all reference to Lancelot's being wounded in the wood, condensing considerably at this point, and gives no account of the arrival of the dead body of the maiden of Escalot.
In the account of how Lancelot learns of the queen's danger from Madoc de la porte, all three versions differ. According to S. he meets a knight from Kamalot who tells him of the queen's plight, and at once resolves to rescue her. The next day he meets Hector and reveals his intentions; and a few days after both meet Bohort, who asks if they know the news. 1533 says that as the first knight rides off, Hector appears from a cross-road; he is on his way to defend Guinevere.
In D. L. it is not said how Lancelot first learns the news, but he meets Hector and Bohort together, and on their asking him if he has heard, replies in the affirmative.
I suspect M. had a version akin to this last before him as he makes Bohort Lancelot's informant.
In the account of the final detection of Lancelot and Guinevere, S., as I have before pointed out, goes wrong, by substituting Guerreshes for Gariët. All the texts I have consulted agree in stating that it is this latter who sides with Gawain, and refuses to be a party to the betrayal.200
D. L. omits the fact that Arthur hears of Lancelot's victory at the tournament of Cahere; and also the remark of Bohort that only Morgain or Agravain can have betrayed him.
In the details of the detection all three versions agree closely.
In the account of Guinevere's trial S. again diverges from the others. We read201 'Arthur decides to punish Guinevere with death. He will have her tried at once. P. L. introduces here, and a little later, a certain "roy Yon" who counsels moderation. The trial takes place; Arthur, with Gaheriet, Mordred, and Agravain, doom the queen to the stake.' I do not know if this accurately represents the text 1513, it certainly differs widely from the reading of D. L. and 1533.
D. L. does not mention Yon; 1533 simply introduces him as telling Arthur that the trial cannot take place that evening, while both agree in saying that Gawain (whom S. does not mention at all) warns the king not to proceed to extremities, threatening to give up all his lands if the queen be burnt. Mordred and Agravain doom the queen to death, Arthur alone specifies the nature of that death.202
In the account of the fatal fight at the stake, D. L. represents Lancelot as slaying both Gawain's brothers, while 1533 agrees with S. in saying that Bohort kills Guerresches and Lancelot Gaheriet. M., it will be remembered, agrees in this with D. L. It may be noted that all three, 1533, D. L., and M., while making no remark about Guerresches, especially lament Gaheriet: the two first say that Lancelot knows Gawain will never forgive him for this, and M. speaks of him as 'the noble knyghte,' making the identity with Gareth quite clear.
The castle at which the queen and Lancelot stay en route for Joyous Garde, called Scalee by S., Scalle in 1533, and Calet in D. L. does not, I think, belong to Keux the seneschal, as S. supposes; D. L. does not mention him, and 1533 speaks of 'ung Keux,' a friend of Lancelot's, which cannot be Kay. Both here, and in the 'Keux du Parc' of the Turquine adventure, I suspect that we have not a proper name at all, but a misreading of 'Queus'=count. In the latter instance D. L. renders Keux by Grave.
On p. 255, S. must surely have misread his source, as he says that Lancelot sends messengers to King Ban of Benoyc, asking his aid. King Ban was of course dead long before; D. L. and 1533 say to the barons of Benoyc, which must be the right reading
Again, the summary of the battle, S., p. 256, differs very materially from D. L. and 1533. S. says Gawain fights like a madman and kills thirty of Lancelot's men with his own hand, wounding others, Lionel among them. The next day there is another battle, in which occurs the incident of Arthur being unhorsed by Bohort, and remounted by Lancelot. Now in the other two versions Bohort and Gawain wound each other so desperately at the first onslaught that they are carried off the field half dead, and it is Hector who overthrows Arthur.
Later on, after the return of Guinevere, when S. represents Hector as challenging Gawain, the other two versions give Bohort.
After the kinsmen return to Benoyc we find D. L. in apparent contradiction with the other versions. S. says that he makes Bohort king of Benoic and Lionel of Gannes, while he himself keeps the crown of Gaule, because Arthur gave it to him. 1533 seems to agree with this latter phrase, as it says, 'et pour ce que le roy Artus me donna le royaulme de Gaule ie le tiendray.203 D. L. on the contrary says:
'Ende vanden conincrike, dat secgic u,
Van Gaule sone doe ic niet nu,
Ende ne houder gene tale af,
Om dat mi die coninc Artur gaf;
Want al haddi mi gegeven vor nu
Al die werelt, dat secgic u,
Ic gavese hem al weder te hant,
Bedie ic ne soude en geen lant
Nu ter tijt van hem willen houden.'—ll. 7407-15.
Now in the earlier portion of D. L., after the war with Claudas, we are told that Lancelot has made Bohort king of Gannes, Hector of Benoyc, and Lionel of Gaul, an arrangement which exactly agrees with that which M. takes from the English M. A. In this earlier passage Lancelot gives as reason for not taking the crown that he prefers to remain a simple knight, and 1533 represents Bohort and Hector as following his example and declining the offered kingdoms.
I think the lesson of this discrepancy is that the Lancelot and the M. A. were fundamentally independent of each other, and each contained an account of the crowning of the race of Ban. When brought into close contact this caused a contradiction of statement which D. L. and 1533 evaded each in their own way. S. gives no clue to what happened on the earlier occasion.
The number of knights Arthur takes with him on his last expedition agrees in D. L. and M., sixty thousand, against forty thousand in the two French versions.204
In both 1533 and D. L., Guinevere does not, as in S., ask for a week's respite, but for a day, and Mordred himself suggests she shall have the week. Labor, whom S. calls simply 'a faithful knight,' is in both these versions a near kinsman—neve, cousin. D. L. gives as a reason for Guinevere's rejection of Mordred's offer that she suspects his true relation to Arthur. This is not in 1533.
In the account of the fight between Lancelot and Gawain, all three versions apparently differ at the outset. Gawain will send the challenge by a squire. S., squire refuses, fearing Lancelot's wrath; 1533, refuses, fearing to bring about Gawain's death; D. L., goes at once.
The issue of the fight too is different in D. L. and 1533. In S., Gawain receives a mortal wound in the head and retreats. In 1533, Lancelot appeals to the king: it is vesper-tide, and a fight for treason must be concluded by nightfall. Arthur, seeing Gawain is getting the worst of the battle, stops it at once. D. L. apparently condenses a similar version, but makes Arthur appeal to Lancelot, who says that he will be dishonoured if he leave his foe in possession of the field, but Arthur entreats him to do so for his sake, and Lancelot retires. Both agree in saying that Gawain is over twenty years Lancelot's senior, and is now eighty-two years old!205
D. L. represents the war with Rome as lasting twenty years, which would make both Arthur and Gawain well over a hundred at the time of their death!
After the news of Mordred's treachery D. L. makes no mention of Gawain being carried in a litter on the return journey, or of his desire for Lancelot's forgiveness; nor does he warn Arthur against fighting with Mordred. This is, I suspect, due to the compiler's desire to condense, as 1533 agrees in the main with S. The warning against Mordred appears, however, to be fuller in the former, e.g. Studies, p. 265. Gawain is represented as saying briefly, 'Avoid, if possible, fighting with Mordred, for it will cause your death,' which is in 1533, 'Car ie vous dy vrayement que se vous mourez par une homme [=q] vive vous mourrez par lui et madame la royne,' p. 154, which certainly seems to point to an earlier redaction of the M. A., where Guinevere was a partner in Mordred's treason.206
In the description of Arthur's death there are some interesting variants. Both 1533 and D. L. account for Lucan's death by the weight of Arthur's armour; it is that, and not the vehemence of the king's embrace, which really kills the sorely wounded knight.
They again differ in the details of the final scene. S. says 'a boat full of ladies arrives; they land, go ashore, put Arthur, his horse, and armour into the boat, and row off.' D. L. says they call Arthur, who rises, takes his horse and armour, and goes into the ship. 1533 says the mistress of the party is Morgain; she calls Arthur, who rises at sight of her, she takes him by the hand (which would seem to imply her landing), and bids him bring horse and arms and enter the boat, which he does. Dr. Sommer evidently regards the entire account as absurd, but I not only accept it, but regard the versions of D. L. and 1533, which would merit his strictures more fully than that in which he finds such difficulty, as representing the earlier and more primitive form of the story. There is no doubt that Arthur was conceived of as living and ruling in Avalon. This account of his practically voluntary departure for the mysterious island is much more in accord with that idea than the version which represents him in the extremity of mortal weakness, and subsequently dead and buried. Arthur's tomb is not compatible with Arthur in Avalon, and I strongly suspect that the earlier redaction of the M. A. made no mention of it; it is certainly omitted in the corresponding section of the Didot Perceval, which only says he departed to have his wounds healed in Avalon, and has not since been seen; but Bretons claim to have heard his horn, and seen his armour, and believe he will return.
There is a curious discrepancy in the accounts of Lancelot's death, which seems to point to two distinct versions of that event. S. says he died August 5th, but does not say how long he was ill. D. L. says he fell ill on May 15th, and died after four days. 1533 says he fell ill May 15th, was ill four days, and died August 5th! Evidently a combination by some unintelligent compiler of the two previous accounts, but it is unusual to find such an obvious bévue in so otherwise admirable a version as that of 1533.
All three agree that Lancelot is buried in Galehault's tomb, and that Bohort becomes a hermit in his stead.
From the above comparison it seems clear that though offering less striking and interesting variants, the Dutch version and that of 1533 yet maintain, on the whole, their previous agreement as against S.; while M., which here possesses an alternative source the English M. A., yet occasionally betrays the same curious agreement with D. L. which we have noted before. The result appears to confirm the conclusion previously arrived at, that D. L. and 1533 represent a common French original, and that M.'s source, whether complete or incomplete, was a MS. belonging to the same family.