The Singer of Tales

  Use the following persistent identifier:

Chapter 10. Some Notes on Medieval Epic [1]

The formulaic character of the Old English Beowulf has been proved beyond any doubt by a series of analyses beginning with my own in 1949 (see Chart VIII), which Professor Magoun improved and elaborated in 1953, [5] and which Professor Creed has carried to its ultimate detailed conclusion. [6] The documentation is complete, thorough, and accurate. This exhaustive analysis is in itself sufficient to prove that Beowulf was composed orally. Thematic study has also been begun for Beowulf, [7] but the concept of theme which Professor Magoun and others have been using differs to some extent from that presented in this book, although I feel that there is no basic conflict but rather a difference of emphasis. I should prefer to designate as motifs what they call themes and to reserve the term theme for a structural unit that has a semantic essence but can never be divorced from its form, even if its form be constantly variable and multiform. It is not difficult to see that even from this point of view there are themes in Beowulf: repeated assemblies with speeches, repetition of journeying from {198|199}

Chart VIII [8]


Beowulf maþelode, ‖ bearn Ecgþeowes:
–––––––––––––––2   –––––––––––––––3


"Geþenc nu, se maera ‖ maga Healfdenes,
–   –   – –  –  ––––––––––––––––––––––––4
–  –   – –   –  ––––––––5 –––––––––––––6


snottra fengel, ‖ nu ic eom siðes fus,7
–––––––––––8   –  –   –  –  –   –  –  –9


goldwine gumena, ‖ hwaet wit geo spraecon,10
––––––––––––––11  –––––––– – – ––––––––12


gif ic aet þearfe ‖ þinre scolde
–  – –  – –   – – – –  –  –   –  –13
––––  –––––––14 ––––  ––––15


aldre linnan, ‖ þaet ðu me a waere16
–––––––––17  –  –  –  –  – ––––––18


forðgewitenum ‖ on faeder staele.19
–  –  –  –   –  –20  ––––––––   –  –21


Wes þu mundbora ‖ minum magoþegnum,22
––––––  –  –  –   –23 –––––  –––––––––––24


hondgesellum, ‖ gif mec hild nime;25
–  –  –   –   – –26 ––––––––––––––27


swylce þu ða madmas, ‖ þe þu me sealdest,28
–  –   –  – ––––––––––29   – –  –  –  –   – –  –30


Hroðgar leofa, ‖ Higelace onsend.31
–  –  –  –  –  –32 –  –  –  –   –  –  –33


Maeg þonne on þaem golde ongitan ‖ Geata dryhten34
                                                       35   –––––––––––36


geseon sunu Hraedles, ‖ þonne he on þaet sinc starað,37
–   –   –   –   –  –    –  –38  –  –  –  –   –   –  –   –   –  –   –39


þaet ic gumcystum  ‖  godne funde
–  –  –  –  –  –  –  – – –  –  –  – –  –40
–––––   –   –   –  –  –41 –––– –––––42


beaga bryttan, ‖ breac þonne moste.43
–––––––––––44 –  –  –  –  –  –  – –  –45

one place to another, and on the larger canvas the repeated multiform scenes of the slaying of monsters.

Thus the arrival of Beowulf in the land of Hrothgar (lines 223b ff.) and the arrival of the hero back in the land of the Geats (lines 1912b ff.) are multiforms of the same theme, both distinguished by the watchman, yet different in their detailed elaboration. One can see the same theme (a) in the assembly in Heorot, when Hrothgar welcomes Beowulf and there is the wrangling between Unferth and the newly arrived stranger, wrangling in which an old story is told, followed by laughter and the entrance of the queen (lines 405ff.) and (b) the feasting after the fight with Grendel (lines 1063ff.) with its ancient tale of Finn and the appearance of the queen following the minstrel’s song.

There are two larger problems in connection with Beowulf which the method of comparative study may help in solving. One is the question of the structure of Beowulf, its “unity”; the other whether Beowulf might be {199|200} a “transitional” text. What has already been said in Chapter Six above concerning the second of these problems can be brought to bear here. Analysis of Beowulf indicates oral composition. The corpus of Anglo-Saxon heroic poetry is so small that it is scarcely possible for us to know its tradition well enough to assert that Beowulf breaks away from that tradition of formula and theme. It seems to be more logical to assume that a text so analyzed belongs to the category of oral dictated texts unless one has sufficient evidence from the tradition itself on which to base a judgment that it is “transitional.” It might be foolhardy to generalize to the extent of including the Anglo-Saxon religious poetry, for example, the Christ, Genesis, Andreas, etc., in the category of heroic poems, or rather as traditional heroic poems. It must be said, however, that the evidence of Caedman’s dream [9] and of the song of the bard in Beowulf

would seem to indicate the possibility that Christian religious narrative had been accepted into the tradition. Perhaps it would be wiser to say that the pagan myths had given place to or had been reinterpreted in terms of the Judaeo-Christian myth.

In regard to the unity of Beowulf, to the question as to whether we are dealing here with one, two, or three poems, it seems to me that in the absence of manuscript evidence to the contrary we must accept the singleness of our poem. More specifically we should recognize that it represents one dictated performance by a single singer, undoubtedly over more than one day of dictating and writing. It might very well be that the story of Beowulf’s fight with the dragon was sometimes separately sung; it is conceivable, though I must confess not probable, that the incident of Beowulf’s undersea adventure with Grendel’s dam was also a subject that was sung separately. But from Chapter Five above on the subject of “what is a song” it should be clear that the fact that these parts might or could be sung separately would not militate against Beowulf as a single song.

A glance at Parry 6580 from Murat Čustović of Gacko in Appendix III will show how easy it is for the singer to continue to narrate about a hero or about action leading out of incidents that would perhaps in normal performance close a song. In other words, in dictating, a singer would very possibly continue narrating, let us say, about Beowulf, about a third encounter with a monster and about the hero’s death, thus perhaps adding one song to another. One does not need a written tradition, as the Yugoslav example shows, in order to produce such a phenomenon. But I believe that when the singer does add material it is because of an association that often {200|201} goes deeper than association of one rescue with another, or of one monster-fight with another. On these grounds, I believe that the dragon episode in Beowulf may have a deeper significance and be more integrally related to the hero and to his previous adventures than can be explained by any biographical or chronological tendency.

The hero, Beowulf, is marked as one of those with whom are associated story patterns of wandering into another world, there to encounter and overcome the king and queen of that world and to return victorious. He is a mythic figure of death and resurrection. Achilles and Odysseus, as we have seen, belong in this same category. Beowulf shows some detailed relationship with the Achilles-Patroclus pattern as well. A case might be made for Aescere as a Patroclus, that is, the close friend who is killed before the encounter of the hero with the enemy. Indeed, unless one interprets his death in this manner, he is at a loss to understand it; for Beowulf is apparently present at the time of Aescere’s slaughter and does nothing about it. We have said in the previous chapter that mythic heroes of this type can die by substitution or symbolically, or by undergoing an “almost-death.” It would probably be accurate to say that Beowulf, like Achilles, has undergone death twice, even three times, in the poem before the dragon episode. First his journey itself into the land of Hrothgar is an expedition into the other world; he dies by substitution in the person of Aescere; and he undergoes an “almost-death,” similar to Achilles’ fight with the river, in his struggle with Grendel’s dam.

La Chanson de Roland

In the case of Beowulf we are dealing with only a single manuscript, but when we turn to the Chanson de Roland and to Digenis Akritas we come upon a richer manuscript tradition. [13] No lengthy analysis has been made of the formula structure in the Chanson, although Rychner’s work [14] leads in this direction. However, an examination of a passage, chosen at random, illustrates the extent to which formulas are used in the Chanson. Only the Oxford manuscript of the song has been employed as material for the analysis of ten lines in Chart IX. [15]

Chart IX


Li quens Rollant par mi le champ chevalchet
–––––––––––––1 –––––––––––––––––––––2


Tient Durendal, ki ben trenchet e taillet,
––––––––––––3                                     4


Des Sarrazins lur fait mult grant damage.


Ki lui veïst l'un geter mort su l'altre,
––––––––7  –  –  –  –  – –  –  – –  –8


Li sanc tuz clers gesir par cele place!
–––––––––––––9                               10


Sanglant en ad e l'osberc e la brace,
–  –  –  –  –  –11–   –  –  –                12


Sun bon cheval le col e les espalles.
––––––––––––13–   –  –                  14


E Oliver de ferir ne se target,
––––––15           –   – –  –  –16


Li .XII. per n'en deivent aveir blasme,


E li Franceis i fierent e si caplent.
–––––––––19 –   –  –  –  –   – –  –20

It seems clear from the chart that the Chanson is formulaic beyond any question. The first part of the line is obviously much more hospitable to formulas than the second part. This is undoubtedly because of the assonance at the end of each line. Nevertheless, at least half of the lines are formulaic in their second part, and there are parts of formulas even in most of the other lines. {202|203}

From Chart IX and its notes we can readily discern formula systems such as the following, which shows how useful tient, trait, or prent in first position in the line is with a three-syllable noun-object extending to the line break:

               ⎧ Durendal
               ⎪ Halteclere
               ⎪ Tencendur
(trait)   ⎫⎪ ses chevels
tient    ⎨⎨
(prent) ⎭⎪ ses crignels
               ⎪ sun espiet
               ⎪ sun escut
               ⎩ l’olifan

Such analyses seem to indicate that the Chanson as we have it in the Oxford manuscript is an oral composition.

When one approaches the problem of the relationship between several manuscripts of the same song (e.g. or the Digenis Akritas or of the Chanson de Roland), the knowledge that we are dealing with oral compositions, coupled with an understanding of how such songs are collected, is helpful. We have seen from the Yugoslav examples that variation, sometimes not great, sometimes quite considerable, is the rule in oral composition. When there is exact line-for-line, formula-for-formula correspondence between manuscripts, we can be sure that we are dealing with a written tradition involving copied manuscripts or with some circumstance of collecting in which a fixed text has been memorized.

If one were to disregard all other elements and parts of the manuscripts and to judge only from this, single passage, one would conclude after perusing Chart X that the Italianized Venice IV is either copied directly from the Oxford manuscript or is a copy of a copy. In line 3 “des Sarrazins” has been interpreted as “de qui de Spagna” by Venice IV, and the tenth line of Oxford has been omitted; line 10 in Venice IV is the eleventh line in the Oxford manuscript. The second half of line 11 of Venice IV does not correspond, nor does line 12. In spite of these differences it would seem that this passage has been either memorized and turned into Italianate French or copied from manuscript.

On the other hand, a comparison of the same laisse in the Chateauroux manuscript with the Oxford (Chart XI) seems to show a relationship which might be that of oral texts of the same song. The first line is quite different; the second agrees in the very common first half-line formula, “tint Durendart“; the third line is the same, but it is a common line; lines four through eight in Chateauroux are quite different; but line 9 of Chateauroux is the same as line 8 of Oxford; and line 10 of Chateauroux corresponds in its first half to line 9 of Oxford, although this is a very common formula (“li XII. per”). {203|204}

Chateauroux line 11 begins with the last half of Oxford line 10, but finishes differently; the first half of line 12 and the first half of line 13 in Chateauroux correspond to the beginnings of lines 11 and 12 respectively in Oxford, but the rest of the laisse is quite different. And we might note that the words of the archbishop in Venice IV correspond in part to his words in Chateauroux. At the end of the laisse, Venice IV (as elsewhere in fact) has {204|205}

Chart XII [18]

Chateauroux   Cambridge
Rollanz fu proz et de mult fier coraje:   Roullant fut preux et de fier courage,
Tint Durendart par mot ruste bataille;   Tint Durandal par son riche barnage,
De Saraçins a fait mot grant doumage;   De Sarrasins y fait moult grant domage.
Cel jor mostra si ben son vasalage.   Le jour y monstre si bien son vasselage,
Qi l’atendit ne fist mie qe saje:   Cil qui l’atent y fait moult grant folage,
La teste i pert, ne demande autre gaje;   La teste prent, il ne quiert aultre gage,
Sanc et cervelle fait voler en l’erbaje,   Sang et cervele fait voler par l’erbage,
Tot a son cors sanglant et son visage.    
Et Oliver de ferir ne se targe;   Et Oliver de ferir ne targe;
Li .XII. per, qi sunt de haut parage,   Li .XII. pairs qui sont de haut parage,
    Ne ly Franceys ja n’y aront hontage!
Ferent et caplent desor la gent sauvage:   Dist l’arcevesque: “Nostre gent est mult sage!
Murent paien a duel et a hontage.   Fierent et chaplent sur celle gent sauvage:
Dist l’arcivesqe: “Nostre gent est mot sage!   Meurent paiens a deul et a hontage.
Bien se defendent a cest estrot pasage.   Bien se desfendent a ceul estroit passage;
Car pleüst Deu, qi fist oisel volage,   Car pleüst Dieu, qui fist oysel sauvage,
Chi fust li rois cui avons fait domage!”   Que fust су ly rois a qui avon fait homage!”

been influenced by the Chateauroux type of manuscript, in spite of its closeness to Oxford. This is still a kind of relationship that is within a written manuscript tradition.

In Chart XII, which places the same passage from the Chateauroux manuscript side by side with the corresponding passage from the Cambridge manuscript, we can see that the relationship between these two, on the basis of this passage, is that between copies. This is apparent although there is some confusion in the order of lines toward the end of the passage, and although one line has been omitted in Cambridge that is in Chateauroux, while there is one line in Cambridge that is not in Chateauroux.

It is easy to divide these four manuscripts then into two groups. We believe that the Oxford-Venice IV group (that is Venice IV insofar as it follows Oxford, namely to line 3865 at least) is oral as shown by our analyses. The relationship between Oxford and Chateauroux looks like that between two oral versions and may be such, but one must note that Oxford is assonantal and Chateauroux is rhymed. And although the “author” or scribe of Chateauroux is not always consistent, he seems to have changed the lines that do not end properly for his rhyme scheme, which is “-age.” So he changes line 1, but neglects to change line 2 (an oversight that Cambridge remedies), keeps line 3, because it already has his rhyme, but changes lines 4–7, because they do not have it, and so forth. {205|206} Strangely enough he also changes line 12, in spite of the fact that it ends with “barnage“! We had best agree with the majority that the relationship between Oxford and Chateauroux is a written one, a conscious literary changing of one manuscript (or manuscript group) characterized by assonance in order to produce a rhymed text. But I should like to suggest that the whole question of these remaniements should be reviewed again in light of oral composition. For the present we begin with an oral Oxford manuscript, from which the others named have been derived.


In our remarks on Beowulf above we have noted that the dragon episode may very well be organic in Beowulf tradition and that the singer, if joining together separate songs, was doing so in accordance with the subconscious forces of attraction that are operative in oral tradition. It is significant that many of our medieval texts are divisible into at least two and sometimes three parts. It is so with Beowulf, Chanson de Roland, the old Spanish Cid, the Nibelungenlied, and Digenis Akritas. Criticism has tried to defend their unity on artistic or logical grounds, but it has seldom if ever attempted to see these poems as traditional units with their parts belonging together by a kind of mythic necessity or by thematic attraction.

One can see, I believe, a traditional force at work (together with other forces) in the addition of the Baligant episode at the end of the Chanson de Roland. Indeed, we can trace parallelisms of mythic meaning between Beowulf and the Chanson. Let us take as a point of departure the scene in which Baligant has nearly overcome Charlemagne in single combat. Charlemagne’s prayer finally gives him the strength at the last desperate moment by miracle to kill Baligant. The parallel is clear. Beowulf in the mere is lying on his back, with Grendel’s dam astride his chest and ready to dispatch him. He is saved by catching sight of a fragment of a giant’s sword lying nearby with which he kills the monster. We have already noted in Chapter Nine parallels in the Iliad and in the Odyssey, of Achilles nearly overcome by the river, Skamander, but saved by Hephaestus, and of Odysseus about to drown when rescued by Ino. The scene is in the other world and the hero is locked in mortal combat with the king of death. For the myth to be effective, he must overcome death and return.

If we pursue the parallel between Charlemagne (and his representative Roland) and Beowulf, we find a few striking details. But first we must admit the possibility that Roland is a substitute for Charlemagne, that he plays a role not unlike that of Patroclus. If we agree that Roland and Charlemagne are the same, then we see Roland, like Beowulf, wounding his adversary (Marsila in the Chanson, Grendel in Beowulf), and wounding him essentially in the same way in both poems. Beowulf tears Grendel’s arm from its socket, and then the monster escapes to his lair. Roland cuts off Marsila’s arm in single combat, after which Marsila escapes and returns to {206|207} Saragossa. They both seek solace from a female, Grendel from his dam, Marsila from his wife, and they both die as a result of the wounds. Their deaths lead directly to ravages by new characters: by Grendel’s dam and by Marsila’s overlord, Baligant. The structure of incident is amazingly similar in both these songs. Whatever the mythic meaning may be, there is a strong force of association which brings a second encounter with an enemy into the story when the pattern of the first encounter is mutilation, escape, and death elsewhere.

The Baligant episode is as necessary to the story of the Chanson as is that of Grendel’s dam to Beowulf. It may be sufficient to explain this in terms of a feud but it would probably be more accurate to say that some mythic significance has yielded to a sociological explanation of similar intensity. Charlemagne, in the details of his final battle with Baligant, has taken over mythic material, and we should not be surprised that the parallel with Achilles is strong. This is true in spite of the great difference of age between the two heroes, Charlemagne, the old man with the long, white, flowing beard, and the twenty-year-old impetuous Achilles. Although the listeners’ attention is occupied mostly with Roland, it is Charlemagne to whom the song belongs.

Digenis Akritas [19]

Digenis Akritas seems to have been a historical person of the eighth century of our era. The epic about his ancestry, birth, marriage, adventures, and death survives in five Greek metrical manuscripts, a Russian prose version dated by Grégoire in the twelfth century, and a late Greek prose version. [20] I shall be concerned primarily with the five Greek metrical versions, all of which are in the vernacular showing more or less archaizing of language and in the political fifteen-syllable meter. Only one of these is dated, O (Oxford), which belongs to 1670 and is the latest manuscript. For the others I follow Grégoire’s dating: [21] the earliest is G (Grottaferrata) in the fourteenth century; the other three, A (Athens, formerly Andros), T (Trebizond), and E (Escorialensis), are sixteenth century. E and T are acephalic, and E is in very poor condition, with some lacunae. A and T are divided into ten books; O and G into eight books. E has no book division. The story as told in these manuscripts is essentially the same in all of them, but there is, nevertheless, considerable divergence in the telling. I shall not attempt to prove that any of these manuscripts is an oral text. They have been through the hands of learned men, or at least educated men who knew how to read and write. But I shall suggest tentatively that one of these manuscripts, E, is very close to being an oral text, and that the others have enough oral characteristics to show that there is an oral text behind them and that some signs of oral technique of composition have survived in them in spite of their literary, written, and learned character.

It is by now a truism that no two performances of an oral epic are ever {207|208} textually exactly alike. Not only is such textual divergence typical and fundamental in oral style, but also, as we have said earlier, if two texts are nearly word-for-word exact, they cannot be oral narrative versions but one must have been either memorized or actually copied from the other or from the same original. Let us look at the same passage in the five Greek metrical manuscripts. I reproduce them here from Krumbacher’s article, written in 1904, at the time he announced in Munich the discovery of the Escorialensis. [22]

ἀνέγνωσαν τὰ γράμματα καὶ οὕτως ἐδηλῶναν·
καἰ ὡς ἤκουσεν τὰ γράμματα, ἐθλίβην ἡ ψυχή του,
ἐκαύθηκαν τὰ σπλάγχνα του, ἐχάθην ἡ καρδιά του,
ἤκουσεν διὰ τὴν μάνναν του…

Ε 298–301

ὡς δὲ εἶδεν ὁ ἀμηρᾶς τὴν γραφὴν τῆς μητρός του,
ἐσπλαγχνίσθη κατὰ πολὺ ὡς υἱὸς τὴν μητέρα,

G 2. 105–106

καὶ ὡς ἤκουσε τὰ γράμματα, ἐθλίβη ἡ ψυχή του,
καὶ ἡ καρδία του τιτρώσκεται, ἠλέησε τὴν μητέρα,

Τ 235–236

τότε του δίνουν τὴν γραφή, ἀνοίγει καὶ διαβάζει
καὶ τότες ἀφ᾽τὴν πίκρα του βαρειὰ ἀναστενάζει,
γιατ᾽ ἐπικράθηκεν πολλὰ τὸ πῶς του καταρᾶται
ἡ μάνα…

Ο 655–658

καὶ ὡς ἤκουσε τὰ γράμματα, ἐθλίβη ἡ ψυχή του,
καὶ τὴν καρδίαν τιτρώσκεται, ἠλέησε τὴν μητέρα,

Α 685–686

Even from Krumbacher’s brief sampling one could see that the Trebizond and Athens manuscripts are almost word-for-word the same. We might justly conclude that these are copies of the same original; and I do not believe that anyone would quarrel with us. We should also say that Oxford is far from the other four, but that Athens, Escorialensis, and Grottaferrata are somewhat alike.

A better idea of their similarity can be obtained from a comparison of somewhat longer passages from the three manuscripts. I have chosen the beginning of the Escorialensis manuscript, which is acephalic. {208|209}

            "Κρότοι καὶ κτύποι καὶ ἀπειλαὶ μὴ σὲ καταπτοήσουν, 
            μὴ φοβηθῇς τὸν θάνατον παρὰ μητρὸς κατάραν· 
            μητρὸς κατάραν φύλαττε καὶ μὴ πληγὰς καὶ πόνους. 
            Μέλη καὶ μέλη ἂν σὲ ποιήσουσιν, βλέπε ἐντροπὴν μὴ ποιήσης 
5          ἂν κατεβοῦμεν.
               Toὺs πέντε ἂς μᾶς φονεύσουσιν καὶ τότε ἂς τὴν ἐπάρουν· 
            μόνον προθύμως ἔξελθε εἰς τοῦ Ἀμηρᾶ τὴν τόλμην· 
            τὰ δυό σου χέρια φύλαττε καὶ ὁ θεὸς νὰ μᾶς βοηθήσῃ."
            Καὶ ὁ Ἀμηρᾶς ἐκαβαλλίκευσεν, εἰς αὐτὸν ὑπαγαίνει·
10        φαρὶν ἐκαβαλλίκευσεν φιτυλὸν καὶ ἀστερᾶτον· 
            ὀμπρὸς εἰς τὸ μετῶπιν του χρυσὸν ἀστέραν εἶχεν, 
            τὰ τέσσαρά του ὀνύχια ἀργυροτζ άπωτα ἦσαν, 
            καλιγοκάρφια ὁλάργυρα ἧτον καλιγωμένον,
            ἡ οὐρά του σμυρνωμένη μὲ τὸ μαργαριτάριν.
15        Πρασινορρόδινος ἀετὸς εἰς τὴν σέλλαν ἐξ ὀπίσω
            καὶ ἡσκιάζει τὰς κουτάλας του ἐκ τοῦ ἡλίου τὰς ἀκτῖνας·
            κοντάριν ἐμαλάκιζε βένετον χρυσωμένον.

E 1–17

            "καὶ κρότοι, κτύποι, ἀπειλαὶ μὴν σὲ καταπτοήσουν,
325      μὴν φοβηθῇς τὸν θάνατον, παρὰ μητρὸς κατάραν. 
            Μητρὸς κατάραν φύλαγε, κομμάτια κατακόπτου, 
            καὶ ὅταν ἀποθάνετε ἐσεῖς οἱ πέντε ὅλοι
            τότες ἂς τὴν πάρουσιν ἐκεῖνοι πάντες, ὅλοι· 
            μόνον προθύμως ἔξελθε στοῦ Ἀμηρᾶ τὴν τόλμαν,
330      μὲ τὴν βοήθειαν τοῦ θeoῦ, τοῦ μόνου δυναμένου, 
            ἔχω τὸ θάρρος εἰς αὐτὸν τὴν ἀδελφὴν νὰ πάρῃς." 
            Αὐτὸς δ'ἀκούσας τῆς μητρὸς τοὺς λόγους παραυτίκα 
            τὸν μαῦρον ἐπιλάλησε στὸν Ἀμηρᾶν ἐπῆγεν, 
            καὶ μετ᾽ αὐτὸν οἱ ἀδελφοὶ ἐφθάσασι κατόπιν,
335      ἄλογα ἐκαβαλλικεύσασιν ἀρματωμένα οὗτοι. 
            Καὶ ὡς εἶδεν ὁ Ἀμηρᾶς τὸν νέoν Κωνσταντῖνον, 
            τῆς κόρης τὸν αὐτάδελφον, πού ᾽ ρχετον πρὸς ἐκεῖνον, 
            φαρὶν ἐκαβαλλίκευεν φιτυλόν, ἀστερᾶτον, 
            ἔμπροσθεν εἰς τὸ μέτωπον χρυσὸν ἀστέρα εἶχε,
340      τὰ τέσσαρά του νύχια ἀργυροτζ άποτ᾽ ἦσαν,
            καλλιγοκάρφια ἀργυρᾶ ἦταν καλλιγωμένον, 
            πρασινορρόδινος ἀετὸς στὴν σέλλαν ἦτον πίσω, 
            ζωγραφισμένος ἤτονε μὲ καθαρὸν χρυσάφι, 
            τὰ ἄρματά του λάμπασιν ἡλιακὰς ἀκτῖνας
345     καὶ τὸ κοντάρι, ἤστραπτεν, σὰν Βενετιᾶς χρυσάφι.

Α 324–345

            "Μὴ ὅλως, λέγων, ἀδελφέ, φωναὶ καταπτοήσουν, 
135      μικρόν τι δειλιάσωσι, πληγαί σε έκφοβήσουν·
            κἂν γυμνὸν ἴδῃς τὸ σπαθίν, φυγεῖν οὕτω μὴ δώσῃς, 
            κἂν ἄλλο τι δεινότερον εἰς τροπὴν μὴ ἐκφύγῃς· 
            νεότητος μὴ φεῖσαι σὺ παρὰ μητρὸς κατάραν,
            ἧς εὐχαῖς στηριζόμενος τὸν ἐχθρὸν καταβάλεις· {209|210}
140      οὐ γὰρ παρόψεται Θεὸς δούλους ἡμᾶς γενέσθαι·
            ἄπιθι, τέκνον, εὔθυμον, μὴ δειλιάσῃς ὅλως."
            Καί, στάντες πρὸς ἀνατολάς, Θεὸν ἐπεκαλοῦντο·
            "Μὴ συγχωρήσῃς δέσποτα, δούλους ἡμᾶς γενέσθαι."
            Καί, ἀσπασάμενοι αὐτόν, προέπεμψαν εἰπόντες·
145      "'Η τῶν γονέων μας εὐχὴ γένηται βοηθός σου!" 
            ‘Ο δὲ ἐφ᾽ ἵππου ἐπιβὰς μαύρου, γενναιοτάτον, 
            σπαθὶν διαζωσάμενος, λαμβάνει τὸ κοντάριν,
            ἐβάσταξε καὶ τὸ ραβδὶν εἰς τὸ ραβδοβαστάκιν·
            τὸ δὲ σημεῖον τοῦ σταυροῦ φραξάμενος παντόθεν,
150      τὸν ἵππον ἐπελάλησεν, εἰς τὸν κάμπον ἐξῆλθε·
            ἔπαιξε πρῶτον τὸ σπαθίν, εἶθ᾽ οὕτως τὸ κοντάριν.
            Καί τινες τῶν Σαρακηνῶν ὠνείδιζον τὸν νέον·
               "῎Ιδε ποῖον ἐξέβαλον πρὸς τὸ μονομαχῆσαι 
            τὸν τρόπαια ποιήσαντα μεγάλα εἰς Συρίαν!"
155         Εἷς δέ τις τῶν Σαρακηνῶν Ἀκρίτης Διλεβίτης
            γαληνὰ πρὸς τὸν Ἀμηᾶν τοιόνδε λόγον ἔφη·
               "'Ορᾷς τὸ καταπτέρνισμα ἐπιδέξιον ὅπως, 
            σπαθίου τὴν ὑποδοχήν, γύρισμα κονταρίου; 
            Ταῦτα πάντα ἐμφαίνουσι πεῖραν τε καὶ ἀνδρείαν·
160      ὅρα λοιπὸν μὴ ἀμελῶς τὸ παιδίον προσκρούσῃς." [23] 
               Ἐξέβη καὶ ὁ Ἀμηρᾶς εἰς φάραν καβαλλάρης
            θρασύτατος ὑπῆρχε γὰρ καὶ φοβερὸς τῇ θέᾳ, 
            τὰ ἄρματα ἀπέοτιλβον ἡλιακὰς ἀκτῖνας· 
            κοντάριν ἐμαλάκιζε βένετον, χρυσωμένον.

G 1. 134–164


            "Let clanging and crashing and threats not affright you!
            Fear neither death nor anything except your mother's curse!
            Beware your mother's curse, but pay no heed to blows and pain!
            If they tear you to pieces, see that you shame us not, 
5          If we should go down!
            Let them kill the five of us, and then let them take her!
            Only go forth boldly to meet the might of the Emir!
            Guard your two hands, and may God help us all!"
            And the Emir mounted and set out against him. 
10        He mounted his piebald, star-marked steed.
            In the midst of his forehead he had a golden star,
            His four hoofs were silver adorned,
            The nails in his shoes were of silver,
            His tail was stiff with pearls.
15        Green and red was the eagle that perched behind the saddle
            And shaded his shoulders from the rays of the sun.
            The lance he wielded was of blue and gold.

Ε 1–17

            "Let clanging and crashing and threats not affright you!
325      Fear neither death nor anything except your mother's curse! 
            Beware your mother's curse, and do your utmost!
            And when all five of you die,
            Then let them all take her! 
            Only go forth boldly to meet the might of the Emir, {210|211}
330      With the help of God, who alone has power!
            I have faith in Him that you will find your sister."
            When he had heard his mother's words, straightway
            He urged on the black horse and set out against the Emir,
            And after him his brothers brought up the rear, 
335      Mounted on their fully caparisoned steeds.
            And when the Emir saw the young Constantine,
            The twin brother of the girl, proceeding against him,
            He mounted his piebald, star-marked steed.
            In the midst of his forehead he had a golden star, 
340      His four hoofs were silver adorned,
            The nails in his shoes were of silver.
            Green and red was the eagle that perched behind the saddle,
            Painted it was with pure gold.
            His weapons shone like the rays of the sun, 
345      And his lance gleamed like Venetian gold.

A 324–345

            Saying, "No wise, brother, let the shouts affright you
135      Nor ever shrink, nor let the blows appal you;
            If you see the sword naked, give not way,
            Or anything more terrible, never fly;
            Heed not your youth, only your mother's curse,
            Whose prayers supporting you, you shall prevail.
140      God shall not suffer us ever to be slaves.
            Go child, be of good heart, fear not at all."
            And standing towards the east they called on God:
            "O Lord, never allow us to be slaves."
            Having embraced they sent him forth, saying,
145      "So may our parent's prayer become your helper."
            He mounting on a black, a noble horse,
            Having girt on his sword, took up the lance;
            He carried his mace in the mace-holder,
            Fenced himself all sides with the sign of the cross, 
150      Impelled his horse and rode into the plain,
            Played first the sword and then likewise the lance.
            Some of the Saracens reviled the youth:
            "Look what a champion is put out to fight
            Him who great triumphs made in Syria." 
155      But one of them a Dilemite borderer
            Spoke softly to the Emir a word like this:
            "You see him spurring, and how cleverly,
            His sword's parry, the turning of his lance.
            All this exhibits skill as well as courage; 
160      See then you meet the child not carelessly."
            Forth came the Emir riding upon a horse,
            Most bold he was and terrible to view,
            His arms were glittering with sunny rays;
            The lance he wielded was of blue and gold.

G 134–164

(Mavrogordato translation, pp. 11–12)

It is not easy to understand these three as copies from the same original. Could they be closely related oral versions? {211|212}

The textual differences between E, А–T, and G in many passages look like the textual differences between oral versions of an oral poem. Let us apply the formula test to these Digenis Akritas texts.

The passage in Chart XIII is from the Athens manuscript and shows that on the basis of its nearly five thousand lines there is a fair number of formulas in the sampling, although the result is not so impressive as in Beowulf or in Roland.

Chart XIII


A 628–637 [24] 


Ἡ μήτηρ δὲ1 τοῦ Ἀμηρᾶ, / ἡ τοῦ Ἀκρίτου μάμμη,2 
––––––––––   –  –  –  –   –   –  –  – –  –  –  –   –  –  –


ὡς εἰς ἀλήθειαν ἔμαθε3 τὰ κατ᾽ αὐτῆς υἱοῦ τε,4
–   –   –  –  –   –   –  –  –


ἀπὸ Συρίας ἔστειλεν /5 τὰ γράμματα τοιαῦτα,6
–  –  –  –   –   –  –   –  –  –  –  –  –   –  –  –  –  –  –


θπήνων μεστὰ ὑπάρχοντα7 ὀνειδσμῶν καὶ πόνων,8
–––––––––––––––––––––––    –  –  –  –  –  –  –   –  –


τῆς δὲ τοιαύτης τε γραφῆς9 ἦσαν οἱ οτίχοι οὗτοι.10
                                                  –  –   –  –  –  –  –   –  –


" Ὦ τέκνον μον, ποθεινότατον,11 μητρὸς παρηγορία,12

   –––––––––––––––––––––––––    –   –   –  ––––––––––


πῶς ἐχωρίσθης ἀπ᾽ ἐμοῦ13 κ᾽ ἐπῆγες εἰς τὰ ξένα;14

–  –   –   –  –   –  –  –  –   –   –  –  –    –   –   –  – –   –  –


Ἐτύφλωσας τοὺς ὀφθαλμούς,15 ἔσβεσας καὶ τὸ φῶς μου,16
                                                        –   –   –  –  –  –   –   – –   – –


πῶς ἀπαρνήθης συγγενεῖς,17 πίστιν καὶ τὴν πατρίδα,18
–  –   –  –  –    –   –  –  –  –   –    – –  –  –  –  –   – –   –  –  –


καὶ ἐγενόμην ὄνειδος19 εἰς ὅλην τὴν Συρίαν;"20
–  –  –   –   –  –    –  –  –   –––––––––––––––––––

An analysis of parts of Grottaferrata shows some tendencies toward formulization but the vocabulary differs. From the Grottaferrata manuscript we find the following formulas, for example:

(1) ὦ τέκνον ποθεινότατον  
ὦ τέκνον ποθεινότατον, πῶς μητρὸς ἐπελάθους 2.53
῏Ω τέκνον ποθεινότατον, ὦ ψυχὴ καὶ καρδία, 4.81
Ὦ τέκνα ποθεινότατα, оἰκτείρατε μητέρα 1.70
(2) ὦ τέκνον μου γλυκύτατον  
Ὦ τέκνον μου уλυκύτατον, οἰκτείρησον μητέρα· 2.89
Ὦ τέκνον μου γλυκύτατον, φῶς τῶν ἐμῶν ὀμμάτων, 3.132
Ναί, τέκνον μου γλυκύτατον, ὁ πατὴρ ἀπεκρίθη, 4.291
ὦ παιδίον γλυκύτατον, καβαλλάρην ἐμπρός μου· 2.290
Ὦ ἄνερ μου γλυκύτατε, αὐθέντα καὶ προστάτα, 2.120 {212|213}
(3) ὁ Ἀμηρᾶς εἰς φάραν καβαλλάρης  
Ἐξέβη καὶ ὁ Ἀμηρᾶς εἰς φάραν καβαλλάρης 1.161
καὶ ὕστερον ὁ Ἀμηρᾶς εἰς φάραν καβαλλάρης. 2.294
(4) εἰς τὸ καβαλλικεύειν  
τὸ δὲ παιδίον εὔθιον εἰς τὸ καβαλλικεύειν· 4.242
διελάλησαν ἅπασιν εἰς τὸ καβαλλικεῦσαι. 4.597
(5) ἐν τῷ ἰδίῳ ἵππῳ  
Ταῦτα εἰπὼν εἰσπήδησεν ἐν τῷ ἰδίῳ ἵππῳ 4.663
καὶ ὑποδέχεται αὐτὴν ἐν τῷ ἰδίῳ ἵππῳ, 4.782
Ὡς γὰρ ταύτην ἀνήγαγον ἐν τῷ ἰδίῳ ἵππῳ, 5.237
ὡς πρὸς ἐμὲ κατήρχετο σὺν τῷ ἰδίῳ ἵππῳ. 6.258
Ἡ δὲ εὐθὺς ἐπέβηκεν ἐφ᾽ ἵππῳ τῷ ἰδίῳ, 6.530

But generally speaking one does not find formulas so readily in Grottaferrata as in Athens. If one proceeds by the method of taking a common word and listing all lines in which it is used in the manuscript, one seems to find fewer repetitions of phrase and consequently greater variety of phrasing in Grottaferrata than in Athens.

But the really significant test is not whether one can find formulas or repeated phrases in a manuscript, but rather how frequent they are in any

The Escorialensis, in spite of its roughness and brevity, presents us with a number of formulas. Here are some typical formulas from Escorialensis: {213|214}

(1) τοιοῦτον λόγον λέγει  
καὶ μετὰ τοῦ δακτύλου του τοιοῦτον λόγον λέγει· 54
Καί τὸτε ὁ Φιλοπαπποῦς τοιοῦτον λόγον λέγει. 677
Καὶ τότε καὶ ὁ στρατηγὸς τοιοῦτον λόγον λέγει· 987
ἀντάμα oἱ πέντε ἐστενάξασιν, τοιοῦτον λόγον εἶπαν· 188
Καί τότε πάλιν ὁ Ἀμηρᾶς τοῦτον τὸν λόγον λέγει· 18
γλυκέα τὸν ἐφίλει καὶ τέτοιον λόγον λέγει· 534
(2) τὸ φῶς τῶν ἀπελάτων  
Ὁ θαυμαστὸς βασίλειος, τὸ φῶς τῶν ἀπελάτων, 622
Πότε νὰ δοῦν τὰ μάτια μου τὸ φῶς τῶν ἀπελάτων 636
νὰ γομωστοῦν τὰ ὀμμάτιά μου τὸ φῶν τῶν ἀπελάτων; 637
(3) φαρὶν ἐκαβαλλίκευσεν  
φαρὶν ἐκαβαλλίκευσεν φιτυλὸν καὶ ἀστερᾶτον· 10
φαρὶν ἐκαβαλλίκευε, πολλὰ ἦτον ὡραῖον, 1486
Γοργὸν ἐκαβαλλίκευσαν ἄλλοι τριακόσιοι ἀγοῦροι, 944
γοργὸν ἐκαβαλλίκευσαν θεῖος καὶ ὁ πατήρ του 1031
(4) εὐθὺς ἐκαβαλλίκευσαν  
Εὐθὺς ἐκαβαλλίκευσαν εἰς τὸν κάμπον κατεβαίνουν· 32
Καὶ εὐθὺς ἐκαβαλλίκευσεν εἰς τὸν οἶκον ἀποβγαίνουν, 482
εὐθὺς ἐκαβαλλίκευσα καὶ ἐπῆγα εἰς τὸ κοράσιον· 1578
χθὲς ἐκαβαλλικεύσαμεν ὁμάδι καὶ οὶ πέντε 425
(5) πηδᾷ, καβαλλικεύει  
Καὶ τότε ὁ νεώτερος πηδᾷ, καβαλλικεύει, 1009
Καὶ ὡς εἶδεν τοῦτο ὁ Κίνναμος πηδᾷ, καβαλλικεύει, 1274
Γοργὸν πάλιν σηκώνεται, πηδᾷ, καβαλλικεύει, 1281
Καὶ τότε καὶ ὁ Φιλοπαπποῦς πηδᾷ, καβαλλικεύει, 1357
Καὶ πάραυτα ὁ Ἀμηρᾶς πηδᾷ καβαλλικεύγει, 566
καὶ μὲ τὸν λαόν του καὶ μὲ τοὺς ἀγούρους του πηδᾷ καβαλλικεύγει 567
καὶ ὅσοι οὐδὲν τὸν ἐγνώριζαν, πηδοῦν καβαλλικεύουν. 927
(6) τηδᾷ καὶ ἐκαβαλλίκευσεν  
πηδᾷ καὶ ἐκαβαλλίκευσεν καὶ ἐπῆρεν τὸ σπαθίν του 831
πηδᾷ κ᾽ ἐκαβαλλίκευσε καὶ παίρνει καὶ κοντάριν 1555
Πηδῶ καὶ ἐκαβαλλίκευσα τὴν θαυμαστὴν τὴν φάραν, 1440
(7) Καὶ αὐτὸς ἐκαβαλλίκευσεν  
Καὶ αὐτὸς ἐκαβαλλίκευσεν ὁ Διγενὴς Ἀκρίτης, 752
Καὶ αὐτὸς ἐκαβαλλίκευσεν, εἰς αὐτὸν καταβαίνει 938
Καὶ ὁ λαὸς ἐκαβαλλίκευσεν μετὰ πολυχρονίων· 1060
Καὶ ὁ Ἀμηρᾶς ἐκαβαλλίκευσεν, εἰς αὐτὸν ὑπαγαίνει· 9
Καὶ οἱ πέντε ἐκαβαλλίκευσαν, εἰς τὸν Ἀμηρᾶν ὑπάγουν· 58
Οἱ πέντε ἐκαβαλλίκευσαν καὶ ὑπὰν εἰς τὸ Χαλκοπέτριν, 332
(8) Γοργὸν ἐπῆρα τὸ ραβδίν  
Γοργὸν ἐπῆρα τὸ ραβδὶν καὶ προσυπήντησά τους, 1173
γοργὸν ἐπῆρεν τὸ ραβδὶν καὶ προσυπήντησέν τους 974
γοργὸν ἐπῆρα τὸ σπαθὶν καὶ προσυπήντησά του· {214|215} 1133

One can note from these formulas the following systems built around the verb for mounting or riding a horse; with the kind of horse as the direct object:

φάριν   ⎫
             ⎬ ἐκαβαλλίκευσεν (-σαν)
γοργόν ⎭

with an adverb preceding the verb:

εὐθύς ⎫
           ⎬ ἐκαβαλλίκευσαν (-σαμεν)
χθές   ⎭

or with a preceding subject:

καὶ ὁ Ἀμηρᾶς  ⎫
καὶ οἱ πέντε     ⎪
οἱ πέντε           ⎬ ἐκαβαλλίκευσεν (-σαν)
καὶ αὐτός        ⎪
καὶ ὁ λαός       ⎭

For comparison with the passages analyzed for formulas from the Athens and Grottaferrata manuscripts, the passage from Escorialensis (lines 1274–1280) similarly analyzed will be useful (Chart XV). It is worth stressing that Athens has 4778 lines, Grottaferrata 3709, and Escorialensis only 1867 lines. The analyses in each case are based only on material from the manuscript from which the passage is taken. Because of the frequency of formulas in the evidence presented here, in spite of the limited amount of material for analysis, and because of the irregularity of the lines in the manuscript itself, I might tentatively suggest for consideration by the specialists that Escorialensis may be an oral manuscript unskillfully written down from dictation. It is instructive to compare the irregular lines in Escorialensis with the recited texts in Parry and Lord, Volume II (for example, No. 11).

Turning next to the test of enjambement, one sees necessary enjambement frequently in the Oxford manuscript of the Digenis Akritas, which dates {215|216} from 1670 and is in rhymed couplets. Both the rhyme and the enjambement point here to a “literary” text:

            Ὡς εἶδαν οἱ Σαρακηνοὶ ὅτι ‘πῶς ἐνικήθη
270      ὁ Ἀμηρᾶς καὶ εἰς τῆς γῆς τὸ χῶμα ἐτυλίχθη, 
            τρέχουν καί τον ἁρπάζουσιν νὰ μήν τον θανατώσῃ
            ὁ Κωνσταντῖνος κ᾽ εἰς τῆς γῆς τὸ χῶμα τόνε χώσῃ,
            καὶ καθώς τον ἁρπάξασιν τοῦ λέσιν· " μὴ θελήσῃς
            πλει᾽, ἀφέντη, μὲ τὸν Κωνσταντῆ νὰ ᾽βγῇς νὰ πολεμήοῃς,
275      μόνο ἀγάπην ἄν ᾽μπορῇς κάμε μ᾽ αὐτὸν γιὰ νά ᾽χῃς
            ἀνάπαψιν καὶ ἀφοβιὰ εἰς ὅποιον τόπο λάχῃς. "
            Ὅμως ‘σὰν ἐσυνέφερεν ὁ Ἀμηρᾶς ᾽φοβήθη
            μήπως καὶ ἀφ’ τὸν Κωνσταντῆ ‘πάγῃ μέσα ᾽ς τὰ βύθη
            τοῦ ᾍδου καὶ ἐτρόμαξεν καὶ γιὰ τοῦτο καθίζει
280      εἰς τἄλογο καὶ ᾽γλήγορα ᾽ς τὸ στράτευμα γυρίζει·
            καὶ φεύγοντας ἐγύρισεν πάλι καὶ φοβερίζει
            τὸν Κωνσταντῖνον, κ᾽ ἤρχισεν μὲ λόγια νὰ ὑβρίζῃ.

O 269–282


            When the Saracens saw how the Emir 
270      Was being overcome and covered in the mound of earth,
            They ran and took him, in order that him might not slay
            Constantine and put him into the mound of earth;
            So straightway they took up his body. "Do not wish
            More, my lord, to go and do battle with Constantine.
275      Only be reconciled with him, if you can, that you may have
            Rest and freedom from fear whatever may befall."
            When the Emir recovered his senses, he feared lest
            He might be despatched by Constantine to the midst of the depths
            Of Hades, and he trembled and for this reason sat
280      Upon his horse and quickly returned to the army.
            And he turned back the fleeing and he affrighted
            Constantine, and began to taunt him with words.

On the other hand the following example from the Athens manuscript, dated by Grégoire in the sixteenth century, shows the kind of unperiodic enjambement we have seen in the Slavic examples earlier:

2030    ‘Ως τὸν εἶδεν ὁ Διγενὴς τὴν κόρην τότε ᾽λάλει·
            "Βλέπεις, καλή μου, Σαρακηνόν, ποὺ μᾶς καταδιώκει,
            ἄρτι, κυρία, πρόσεξε τῶς θέλω νὰ τὸν κάμω. "
            Ἐσήκωσε τὴν λιγυρὴν οτὴν γῆν τὴν ἀποθέτει, 
            αὐτὸς ἐκαβαλλίκευσεν ἐπῆρε τὸ κοντάρι,
2035    καὶ πρὸς ἐκεῖνον ἔτρεξεν καὶ προσυπήντησέν τον,
            καὶ πρῶτον τὸν ἐλάλησε· "Σαρακηνέ μου δέχου."
            Καὶ τὸ κοντάριν ἔσυρε ὀμπρὸς στὴν κεφαλήν του,
            εὐθὺς ἀπεθανάτωσε κεῖνον καὶ τ᾽ ἄλογόν του,
            καὶ πάλιν μεταστράφηκεν ὄπισθεν εἰς τὴν κόρην.
2040    Καῖ ἄλλοι τριακόσιοι ἀγοῦροι τὸν ἐφθάνουν
            καβαλλαραῖοι καὶ πεζοὶ ἤλθασι πρὸς ἐκεῖνον,
            ἔκαραζον δέ, ἐφώναζον καὶ ταραχὰς ἐποίουν. 

A 2030–2042 {216|217}


2030    When Digenis saw him, he spoke to the girl:
            "My dear, you sec the Saracen pursuing us;
            Pay heed now, my lady, to how I shall deal with him."
            He raised the sweet maid and put her upon the ground,
            While he himself mounted and took up his spear, 
2035    And he set out to meet him and confronted him,
            And first he addressed him: "Saracen, receive my blow!"
            And he hurled his spear at his head.
            Straightway he killed him and his horse,
            And he went back again to the girl. 
2040    And another thirty youths carne up,
            Riding and on foot they came toward him,
            And they cried out and shouted and made a great din.

In respect to enjambement, therefore, the Athens manuscript might be oral, but this feature, unlike formulaic structure, is far from being sufficiently decisive for us to call this manuscript oral. All we can say is that it is not the same kind of “literary” style as that of the Oxford manuscript or of Virgil. Indeed, the Oxford manuscript is the only Digenis Akritas manuscript that has a predominance of necessary enjambement. All the other manuscripts exhibit the unperiodic, adding, style of oral poetry. This feature as we have said before, is symptomatic, however, rather than decisive because it persists into written poetry.

If one cannot reconstruct an original text, and if one cannot reconstruct with any degree of exactness the myriad thematic complexes which the poem has shown in the past, one can, I believe, reconstruct a basic form, a more or less stable core of the story. No matter how fluid the song content may be, there is always this stable core of narrative or of meaning that distinguishes one song from another. In the case of Digenis Akritas, it is the age-old tale of the demigod who lives a wonder-working life among men, who champions and saves, but who has within him a mortal element which leads inevitably to his death. The tale of Digenis must begin with the story of the marriage of his mother and father; it must as inevitably end with the hero’s death. Indeed, I would suggest that this epic has tenaciously survived, even when misunderstood, because of the basic grandeur of its myth.

But the epic of Digenis in time went through the many sea-changes inherent in oral composition and recomposition. I do not think that we should conceive of these as Grégoire does as redactions of an original text or as remaniements. How then are we to envisage the composition of the several manuscripts which we possess? I think we may justly hazard the opinion that the Escorialensis manuscript is directly from oral tradition; and at the other end of the spectrum, that the Oxford manuscript is a literary reworking from some previous manuscript of the song, presumably one like the Grottaferrata, which is also divided into eight books. Certainly behind the other two manuscripts, Grottaferrata and Athens, is an oral form of the story, as we have indicated above. It might be that this oral form was written down and formed a canonized text for singers who were like the rhapsodes of ancient Greece (as opposed to its ἀοιδοί) or like the narodni guslari of Yugoslavia. It would not be inconsistent with the facts, I believe, to suppose then that (1) Escorialensis is a rhapsode version of this canonized text, written down either by or from the rhapsode and (2) Grottaferrata and Athens are rhapsode versions that have been retold once more by a man whose repertory of tales included as well the current romances of chivalry, and who has attempted to relate the story of Digenis as a romance. Yet, these romances may also be from oral tradition, and the wedding here of epic and romance is a most natural one. In other words one can see here that epic and romance are not really separate genres, but actually the same genre of oral narrative poetry. In a chivalric and religious age the older heroic epic naturally assumes the coloring of its age, and the oral style {219|220} allows for change, for multiplication of incident, and for general expansion.

Only when these versions exist on paper can we speak of the learned editor who has divided the tale into books, eight or ten as the case may be, and provided these books with introductions as in the Athens manuscript. Our texts have been touched up to look like Homeric epic as it existed in Alexandrian manuscripts and later editions. To this editor would certainly be due the references in the text to Homer and the “shield of Achilles” type of description of Digenis’ palace. Yet even the little introductions and the “learned” references are done in a style almost indistinguishable from the rest, by analogy with the patterns and rhythms of oral poetry; for vestiges of this method of composition survive for a long time into the age of writing.

In Conclusion

Yet after all that has been said about oral composition as a technique of line and song construction, it seems that the term of greater significance is traditional. Oral tells us “how,” but traditional tells us “what,” and even more, “of what kind” and “of what force.” When we know how a song is built, we know that its building blocks must be of great age. For it is of the necessary nature of tradition that it seek and maintain stability, that it preserve itself. And this tenacity springs neither from perverseness, nor from an abstract principle of absolute art, but from a desperately compelling conviction that what the tradition is preserving is the very means of attaining life and happiness. The traditional oral epic singer is not an artist; he is a seer. The patterns of thought that he has inherited came into being to serve not art but religion in its most basic sense. His balances, his antitheses, his similes and metaphors, his repetitions, and his sometimes seemingly willful playing with words, with morphology, and with phonology were not intended to be devices and conventions of Parnassus, but were techniques for emphasis of the potent symbol. Art appropriated the forms {220|221} of oral narrative. But it is from the dynamic, life principle in myth, the wonder-working tale, that art derived its force. Yet it turned its back on the traditional significance to contemplate the forms as if they were pure form, and from that contemplation to create new meanings.

The nontraditional literary artist, sensing the force of the traditional material whence his art was derived, but no longer comprehending it, no longer finding acceptable the methods of the traditional, sought to compensate for this lack by intricacies of construction created for their own sake. The old patterns were not only thus given new meanings, but a kind of complexity, which could be attained only through writing, was also cultivated as an end in itself. When we look at oral poetry and observe in it something that looks like these new forms and complexities, we may be deluded. Enamored of the meretricious virtues of art, we may fail to understand the real meaning of a traditional poem. That meaning cannot be brought to light by elaborate schematization, unless that schematization be based on the elements of oral tradition, on the still dynamic multiform patterns in the depths of primitive myth. {221|}


[ back ] 1. The material for this chapter is drawn in part from a lecture on medieval epic at the English Institute in New York in September 1956, and from a lecture at Dumbarton Oaks in the spring of 1955. For assistance with the Byzantine Greek material I am especially indebted to Dr. George C. Soulis of Dumbarton Oaks.

For the most recent discussion of comparative Slavic epic see V. M. Žirmunskij, “Epičeskoe tvorčestvo slavjanskih narodov i problemy sravnitel’nogo izučenija eposa,” in the IV. meždunarodnyj s’ezd slavistov, doklady (Moscow, 1958); P. G. Bogatyrev, “Nekotorye zadači sravnitel’nogo izučenija eposa slavjanskih narodov,” also in IV meždunarodnyj s’ezd slavistov, doklady (Moscow, 1958); and the various articles in Qsnovnye problemy eposa vostočnyh slavjan (Moscow, 1958), published by the Institut mirovoj literatury of the Soviet Academy of Sciences in Moscow.

[ back ] 2. See Francis P. Magoun, Jr., “Oral-Formulaic Character of Anglo-Saxon Narrative Poetry,” Speculum, 28:446–467 (July 1953), and Adrien Bonjour, “Веowulf and the Beasts of Battle,” PMLA 72:563–573 (September 1957).

[ back ] 3. Ronald A. Waldron, “Oral-Formulaic Technique and Middle-English Alliterative Poetry,” Speculum, 32:792–804 (October 1957).

[ back ] 4. Jean Rychner, La Chanson de Geste, Essai sur l’Art Épique des Jongleurs (Geneva and Lille, 1955).

[ back ] 5. See Magoun, Speculum, 28:446–467 (July 1953).

[ back ] 6. See the unpublished doctoral thesis of Professor Robert Creed of Brown University in the Harvard University archives. For an excellent example of the application of the oral theory to textual criticism, see Professor Creed’s article “Genesis 1316″ in Modern Language Notes, 73:321–325 (May 1958).

[ back ] 7. Stanley Greenfield, “The Formulaic Expression of the Theme of ‘Exile’ in Anglo-Saxon Poetry,” Speculum, 30:200–206 (April 1955), and Francis P. Magoun, Jr., “The Theme of the Beasts of Battle in Anglo-Saxon Poetry,” Bulletin de la société néo-philologique de Helsinki, LVI:81–90 (1955). Professor Creed is planning a full-scale investigation of the theme in Beowulf.

[ back ] 8. Notes to Chart VIII follow:

[ 1 ] 529, 631, 957, 1473, 1651, 1817, 1999 (Biowulf, Ecgðioes), 2425 (Biowulf). Rhythmic pattern is Pope D*2, 45 (p. 312). Syntactic pattern is subject (1), verb (2), appositive (3), patronymic (4). The figures in parentheses indicate the number of the measure in which the syntactic unit is included. Cf. 371 (Hroðgar m. / helm Scyldinga), 1321, 456 (same), and also the variant pattern caused by alliteration 499 (Unferð m. / Ecglafes bearn), 2862 (Wiglaf m. / Weohstanes sunu), 3076 (Wiglaf m. / Wihstanes sunu).

[ 2 ] 405 (on him byrne scan), 2510 (beotwordum spraec), 2724 (Biowulf m. / he ofer benne spraec). For rhythmic pattern and syntactic pattern see note [1]. Cf. also 286 (Weard m. / ðaer on wicge saet), 348 (Wulfgar m. / þaet waes Wendla leod), 360 (Wulfgar m. / to his winedrihtne), 925 (Hroðgar m. / he to healle geong), 1215 (Wealhðeo m. / heo fore þaem werede spraec), 1687 (Hroðgar m. / hylt sceawode), 1840 (Hroðgar m. / him on andsware).

[ 3 ] 2177 (swa bealdode). Rhythmic pattern is Pope D1, 1 (p. 358). For syntactic pattern see note [1]. Cf. 1550 (haefde ða forsiðod / sunu E.), 2367 (oferswam ða sioleða bigong / sunu E.), 2587 (þaet se maera / maga E.), 2398 (sliðra geslyhta / sunu Ecgðiowes). Cf. also 620 (Ymbeode þa / ides Helminga), 194 (þaet fram ham gefraegn / Higelaces þegn), etc. In the last instance note the reversal of noun and patronymic because of alliteration.

[ 4 ] Syntactic pattern is imperative-adverb (1), demonstrative-adjective (2), vocative (3), patronymic (4). Cf. 2587 (þaet se maera / maga Ecgðeowes), 2011 (sona me se maera / mago Healfdenes).

[ 5 ] Cf. 489 (site nu to symle / ond onsael meoto), 1782 (ga nu to setle / symbelwynne dreoh), and 2747 (bio nu on ofoste / þaet ic aerwelan) for the first measure; and 762 (mynte se maera / þaer he meahte swa), 675 (gespraec þa goda / gylpworda sum), 2971 (ne meahte se snella / sunu Wonredes), and lines 2587 and 2011 given in note [4] for the second measure. Rhythmic pattern is Pope A3, 89 (p. 270). For syntactic pattern see note [4].

[ 6 ] 189 (swa ðа maelceare), 2143 (maðma menigeo), 1867 (mago H. / maþmas twelfe), and 2011 given in note [4]. Cf. also 1465 (huru ne gemunde / mago Ecglafes), and 2587 given in note [4]. Rhythmic pattern is Pope D2, 11 (p. 361). For syntactic pattern see notes [1] and [4].

[ 7 ] Syntactic pattern is adjective (1), vocative (2), adverb-pronoun subject (3), genitive-adjective (4).

[ 8 ] 2156 (sume worde het). Cf. 1507 (hringa þengel / to hofe sinum), 1400 (wicg wundenfeax. / Wisa fengel), 2345 (oferhogode ða / hringa fengel). Rhythmic pattern is Pope B1, 4 (p. 336). For syntactic pattern see note [7].

[ 9 ] Cf. for the first measure 251 (aenlic ansyn. / Nu ic eower sceal), 946 (bearngebyrdo. / Nu ic, Beowulf, þec), and also 335 (heresceafta heap? / Ic eom Hroðgares), 407 (waes þu, Hroðgar hal! / Ic eom Higelaces), and 2527 (Metod manna gehwaes. / Ic eom on mode from). For the second measure cf. 579 (siþes werig. / Ða mec sae oþbaer), and 1794 (sona him seleþegn / siðes wergum). Rhythmic pattern is Pope A1, 1b (p. 247). For syntactic pattern see note [7].

[ 10 ] Pattern is vocative (1), genitive (2), relative-pronoun subject (3), pronoun indirect object-verb (4).

[ 11 ] 1171 (ond to Geatum spraec). Cf. 2419 (g. Geata. / Him waes geomor sefa), 2584 (g. Geata. / guðbill geswac). Rhythmic pattern is Pope A2a, 34 (p. 258). For syntactic pattern see note [10], and cf. notes [3] and [6]. This is common pattern. Cf. also 120 (wonsceaft wera. / Wiht unhaelo), 467 (hordburh haeleþa; / ða waes Heregar dead), etc.

[ 12 ] Cf. for first measure 1186 (hwaet wit to willan / ond to worðmyndum), and 1707 (freode, swa wit furðum spraecon). For the second measure cf. 2252 (gesawon seledream. / Nah, hwa sweord wege), 3126 (Naes ðа on hlytme, / hwa þaet hord strude), etc. Rhythmic pattern is Pope C1, 2 (p. 348). For syntactic pattern see note [10].

[ 13 ] This is a closely knit line syntactically and rhythmically, with no pause between the second and third measures. Note also the necessary enjambement at the end of the line. Syntactic pattern is conjunction-pronoun subject-preposition (1), noun-object (2), possessive pronoun (3), verb (4). Cf. 293 (swylce ic maguþegnas / mine hate). See also note [40].

[ 14 ] Cf. for the first measure 1822 (gif ic þonne on eorþan / owihte maeg), 2519 (waepen to wyrme, / gif ic wiste hu), and 1185 (uncran eaferan, / gif he þaet eal gemon), 2841 (gif he waeccende / weard onfunde), 1140 (gif he torngemot / þurhteon mihte), 944 (aefter gumcynnum, / gyf heo gyt lyfað). For the second measure cf. 1525 (ðeodne aet pearfe; / ðolode aer fela), 2709 (þegn aet ðearfe! / þaet ðam þeodne waes), and cf. also 1456 (þaet him on ðearfe lah / ðyle Hroðgares), 2694 (Ða ic aet þearfe (gefraegn) / þeodcyninges), 1797 (þegnes þearfe, / swylce þy dogore), and 2801 (leoda þearfe; / ne maeg ic her leng wesan). Rhythmic pattern is Pope A3, 68 (p. 265). For syntactic pattern see note [13].

[ 15 ] For the third measure cf. 2131 (þa se ðeoden mec / ðine life), 2095 (þaer ic, þeoden min, / þine leode), 1823 (þinre modlufan / maran tilian), 1673 (ond þegna gehwylc / þinra leoda), etc. For the fourth measure cf. 230 (se þe holmclifu / healdan scolde), 280 (gyf him edwenden / aefre scolde), 1034 (ongean gramum / gangan scolde), 1067 (aefter medobence / maenan scolde), etc. Out of 19 cases of “sculan” observed, all but 3 are in the fourth measure. For syntactic pattern see note [13]. See also notes [40] and [42].

[ 16 ] The syntactic pattern is object (1), infinitive (2), conjunction-pronoun subject-dative of reference (3) adverb-verb (4). Note that this line is connected to both the preceding and the following line by necessary enjambement.

[ 17 ] 2443 (aeðeling unwrecen / ealdres linnan). Cf. also 680 (aldre beneotan, / þeah ic eal maege), 1524 (aldre sceþðan / ac seo ecg geswac), 2599 (ealdre burgan. / Hiora in anum weoll), 2924 (þaette Ongenðio / ealdre besnyðede), 661 (gif þu þaet ellenweorc / aldre gedigest), 1469 (under yða gewin / aldre geneþan), 1655 (Ic þaet unsofte / ealdre gedigde), etc. Cf. also such formulas as 1002 (aldres orwena. / No þaet uðe byð), 1565 (aldres orwena, / yrringa sloh), 1338 (ealdres scyldig, / ond nu oþer cwom), etc. For syntactic pattern see note [16].

[ 18 ] Cf. 313 (torht getaehte, / þaet hie him to mihton), 1833 (wordum ond weorcum, / þaet ic þe wel herige), 203 (lythwon logon, / þeah he him leof waere), 2161 (hwatum Heorowearde, / þeah he him hold waere). For the fourth measure cf. also 881 (eam his nefan, / swa hie a waeron), 754 (forht on ferhðe; / no þy aer fram meahte), etc. Rhythmic pattern is Pope C1, 5 (p. 349). For syntactic pattern see note [16].

[ 19 ] The syntactic pattern is participle (1 and 2), preposition-genitive (3), object of preposition (4). Note the necessary enjambement with the preceding line.

[ 20 ] Cf. 1937 (handgewriþene; / hraþe seoþðan waes), 59 (Ðaem feower bearn / forðgerimed). Rhythmic pattern is Pope A1, 12b (p. 252). For syntactic pattern see note [19]. It is interesting to note the following from other Anglo-Saxon poems: Gu. 1107 (waeron feowere / forðgewitene), El. 1267 (nu sind geardagas / forðgewitenum), Met. 1052 (waeron gefyru / forðgewitenum), El. 636 (is nu feala siðan / forðgewitenra).

[ 21 ] Cf. 1950 (ofer fealone flod / be faeder lare), 21 (fromum feohgiftum / on faeder bearme), 1114 (Het þa Hildeburh / aet Hnaefes ade). Rhythmic pattern is Pope C1, 34 (p. 356). For syntactic pattern see note [19]. Cf. also Reb. 11 (on bearna staele), and Gen. 1113 (on leofes staele).

[ 22 ] The syntactic pattern is imperative-pronoun subject (1), appositive (2), possessive pronoun (3), dative of reference (4).

[ 23 ] For the first measure cf. 269 (leodgebyrgean; / wes þu us larena god), 407 (Waes þu, Hroðgar, hal! / Ic eom Higelaces), 386 (Beo ðu on ofeste, / hat in gan), 1226 (sincgestreona. / Beo þu suna minum). Cf. also such lines as 2946 (Waes sio swatswaðu / Sweona ond Geata), and especially 2779 (þam þara maðma / mundbora waes), in which the order of the verb has been changed for the sake of the alliteration. For the second measure cf. 349 (waes his modsefa / manegum gecyðed), 373 (waes his ealdfaeder / Ecgðeo haten), 3046 (haefde eorðscrafa / ende genyttod), etc. Rhythmic pattern is Pope C2, 22c (p. 295). Pope lists 118 examples of this rhythm. For instances of it in the second half line see Pope pages 352–53, where he cites 166 examples, and notes the frequency of the compound noun in the second or fourth measures. For the syntactic pattern see note [22].

[ 24 ] Cf. the following examples in note [15] above: 2131, 2095, 1823, and 1673; and in note [13] above: 293, where “minum” occurs in this position in the line, but modifying a following noun. Cf. also 2804 (se scel to gemyndum / minum leodum), 2797 (þaes ðe ic moste / minum leodum), and for “magoþegn” in the second measure 2079 (maerum maguþegne / to muðbonan). For syntactic pattern see note [22].

[ 25 ] The syntactic pattern is dative of reference (1 and 2), conjunction-pronoun object (3), noun subject-verb (4).

[ 26 ] “Hondgesellum” is a hapax legomenon in Beowulf. Cf., however, the many instances in which the first half line is taken up by such a compound: 1495 (hilderince. / Ða waes hwil daeges), 1511 (hildetuxum / heresyrcan braec), 1520 (hildebille, / hond sweng ne ofteah), 1526 (hondgemota / helm oft gescaer), etc.

[ 27 ] Cf. 452 (Onsend Higelace, / gif mec hild nime). Cf. also 447 (dreore fahne, / gif mec deað nimeð), 1491 (dom gewyrce, / oþðe mec deað nimeð). The rhythmic pattern is Pope C2, 22 (p. 353). For the syntactic pattern see note [25].

[ 28 ] Syntactic pattern is adverb-pronoun subject-demonstrative (1), direct object (2), relative-pronoun subject (3) indirect object-verb (4). Note that this line is linked to the following by necessary enjambement.

[ 29 ] Cf. 293 (swylce ic maguþegnas / mine hate), in which the first syllable of the noun takes the place of the demonstrative. For the first measure cf. 757 (swylce he on ealderdagum / aer gemette), 1156 (swylce hie aet Finnes ham / findan meahton), and 2869 (þeoden his þegnum, / swylce he þrydlicost), and 2767 (Swylce he siomian geseah / segn eallgylden). For the second measure cf. 2490 (Ic him þa maðmas, / þe he me sealde), and cf. also 2788 (He ða mid þam maðmum / maerne þioden), 2779 (þam ðara maðma / mundbora waes), etc. Klaeber notes that “swylce” in this sense is used in all but one instance at the beginning of the half line (Glossary page 377 of the 1928 edition of Beowulf). The rhythmic pattern is Pope A3, 70b (p. 266). For the syntactic pattern see note [28].

[ 30 ] The rhythmic pattern is Pope C1, 2 (p. 348). Out of 74 instances he cites 22 with noun or pronoun plus verb in the fourth measure, including line 1482. Of these the following involve either the verb “sellan” or a pronoun before the verb: 72 (geongum ond ealdum, / swylc him God sealde), 1271 (gimfaeste gife, / þe him God sealde), 2182 (ginfaesten gife, / þe him God sealde), 2490 (Ic him þa maðmas, / þe he me sealde), 417 (þeoden Hroðgar, / þaet ic þe sohte), and 563 (manfordaedlan, / þaet hie me þegon). Cf. also 1751 (forgyteð ond forgymeð, / þaes þe him aer God sealde), etc. For syntactic pattern see note [28].

[ 31 ] The syntactic pattern is noun (vocative) (1), adjective (2), indirect object (3), verb (4). This line is linked with the preceding by necessary enjambement. “Onsend” has both its subject and its object in line 1482.

[ 32 ] Cf. 2745 (Wiglaf leofa, / nu se wyrm ligeð), 1216 (Bruc ðisses beages, / Beowulf leofa), 1758 (Bebeorh þe ðone bealonið, / Beowulf leofa), and, with the reversing of noun and adjective for the sake of the alliteration, 1854 (licað leng swa wel, / leofa Beowulf), 1987 (Hu lomp eow on lade, / leofa Beowulf), 2663 (Leofa Biowulf, / laest eall tela). Cf. also 618 (leodum leofne; / he on lust geþeah), and its opposite, 3079 (Ne meahton we gelaeran / leofne þeoden), etc. The rhythmic pattern is Pope A2a, 28c (p. 256). For the syntactic pattern see note [31].

[ 33 ] Cf. 5 (monegum maegþum / meodosetla ofteah), 690 (snellic saerinc / selereste gebeah), 884 (sweordum gesaeged. / Sigemunde gesprong), etc. Cf. also 452 (Onsend Higelace / gif mec hild nime), in which the shift in position is due to alliteration. The rhythmic pattern is Pope E, 7 (p. 370). For the syntactic pattern see note [31].

[ 34 ] Since the rhythmic pattern of the first half of this line is found here only, the line as a whole could not be considered as either formula or formulaic, although the second half of the line is a very common formula.

[ 35 ] The rhythmic pattern is Pope B2, 48 (p. 285). It is found only here in the first half of the line and once in the second half line, 1585 (reþe cempa, / to ðaes þe he on raeste geseah). See Pope, p. 345. This half line is nonformulaic.

[ 36 ] 1831 (Geata dryhten, / þeah ðe he geong sy), 2483 (Geata dryhtne / guð onsaege), 2560 (wið ðam gryregieste, / Geata dryhten), 2576 (Geata dryhten, / gryrefahne sloh), 2991 (geald þone guðraes / Geata dryhten), and the reverse 2402 (dryhten Geata / dracan sceawian), 2901 (dryhten Geata / deaðbedde faest). Cf. also 2419 (goldwine Geata. / Him waes geomor sefa), and 2584 (goldwine Geata; / guðbill geswac), etc.

[ 37 ] Since the rhythmic patterns in both halves of the line are rare (see notes [38] and [39]), the line as a whole must be considered nonformulaic.

[ 38 ] The rhythmic pattern is Pope D1, 8 (p. 302). The syntactic pattern in both instances cited by Pope is the same, verb (1), substantive complex (2), although in 501 (onband beadurune — / waes him Beowulfes sið) the compound noun takes the place of the two nouns in 1485. For the second measure cf. 1847 (hild heorugrimme / Hreþles eaferan), 2191 (heaðorof cyning / Hreðles lafe), 2358 (Hreðles eafora / hiorodryncum swealt), 2992 (Hreðles eafora, / þa he to ham becom) (it is interesting to note that in this instance the phrase is also preceded by “Geata dryhten” in the line before it), and 454 (hraegla selest; / þaet is Hraedlan laf), which are all cases of the reverse order for the sake of alliteration. Cf. also 2025 (geong goldhroden, / gladum suna Frodan), and many other formulas with “sunu” as a base: 524 (sunu Beanstanes / soðe gelaeste), 645 (sunu Healfdenes / secean wolde), 980 (Ða waes swigra secg, / sunu Ecglafes), 1009 (þaet to healle gang / Healfdenes sunu), etc. Cf. also other related rhythmic patterns (D1), such as 758 (Gemunde þa se goda, / maeg Higelaces), etc. (See Pope, p. 359)

[ 39 ] The rhythmic pattern is Pope C2, 30 (p. 355). This is the only case of this pattern in Beowulf, but cf. also the related patterns, 996 (secga gehwylcum / þara þe on swylc starað), 2864 (þaet, la, maeg secgan / se ðe wyle soð specan). Cf. also 2796 (ecum Dryhtne, / þe ic her on starie), and 1603 (modes seoce / ond on mere staredon).

[ 40 ] The syntactic pattern is conjunction-pronoun subject (1), dative (2), adjective (accusative) (3), verb (4). Note the necessary enjambement with the following line.

[ 41 ] Cf. 260 (We synt gumcynnes / Geata leode), 378 (þa ðe gifsceattas / Geata fyredon), 556 (þaet ic aglaecan / orde geraehte), 571 (þaet ic saenaesses / geseon mihte), and 894 (þaet he beahhordes / brucan moste). The rhythmic pattern is Pope C1, 2c (p. 289). Pope notes the frequency with which compounds occur in the second measure, and cites 58 such instances out of the 118 of this pattern. For the syntactic pattern see note [40].

[ 42 ] Cf. 2789 (dryhten sinne / driorigne fand), 1810 (cwaeð, he þone guðwine / godne tealde), 1969 (geongne guðсуning / godne gefrunon), 199 (godne gegyrwan; / cwaeð, he guðcyning), and for another instance of “funde” in the second measure of the half line cf. 1415 (ofer harne stan / hleonian funde). This rhythmic pattern (Pope A1, 1, page 325) is extremely common. Pope notes 460 instances of it in the second half line, and 371 in the first (see note [8]). For the syntactic pattern see note [40].

[ 43 ] The syntactic pattern is objective genitive (1), direct object (2), verb-conjunction (3),verb (4).

[ 44 ] 35 (on bearm scipes), 352 (swa þu bena eart). The rhythmic pattern is Pope A1, За (р. 248). For the syntactic pattern see note [43].

[ 45 ] Cf. 1177 (beahsele beorhta; / bruc þenden þu mote), 894 (þaet he beahhordes / brucan moste), 3100 (þenden he burhwelan / brucan moste), 2241 (brucan moste. / Beorh eallgearo), and for the position of “breac” cf. also 1216 (Bruc ðisses beages, / Beowulf leofa), and 2162 (breostgewaedu. / Bruc ealles well). For the syntactic pattern see note [43].

[ back ] 9. Francis P. Magoun, Jr., “Bede’s Story of Caedman: The Case History of an Anglo-Saxon Oral Singer,” Speculum, 30:49–63 (January 1955).

[ back ] 10. Beowulf, lines 90–98, R. K. Gordon translation.

[ back ] 11. In the Aethiopis. See H. G. Evelyn-White, Hesiod, the Homeric Hymns, and Homerica, pp. 507–509.

[ back ] 12. See the most recent article on the symbolic interpretation of Beowulf: Peter F. Fisher, “The Trials of the Epic Hero in Beowulf,” PMLA, 73:171–183 (June 1958).

[ back ] 13. For an edition of the various manuscripts see Raoul Mortier, Les textes de la Chanson de Roland, 10 vols. (Paris, 1940–44).

[ back ] 14. Rychner, op. cit.

[ back ] 15. Oxford manuscript, lines 1338–1347. Notes to Chart IX follow:

[ 1 ] Lines 194, 355, 663, 707, 751, 777, 792, 803, 1145, 1321, 1545, 1580, 1629, 1671, 1691, 1761, 1785, 1869, 1897, 2066, 2099, 2124, 2134, 2152, 2163, 2166, 2215, 2233, 2246, 2375, 2701.

[ 2 ] 3421. Cf. also 1098 par grant irur chevalchent; 2851 par grant vertut chevalchent; and 3463 li amiralz chevalchet par le camp.

[ 3 ] 1583, 1870. Cf. 1550, 1953 tient Halteclere; 1324 trait Durendal; 3622 prent Tencendur; 2287 tient l’olifan; 2992, 3152 tient sun espiet; 3114 prent sun escut; 2596 trait ces chevels; 2906 trait ces crignels.

[ 4 ] Cf. 925 veez m’espee ki est e bone e lunge; 1276 ki est a flurs e ad or; 1354 ki est ad or e a flur.

[ 5 ] 1340. Cf. 1007, 1929 de Sarrazins; 1030, 1186 E Sarrazins; 202 de ses paiens; 588 de vos paiens; 177 des Francs de France.

[ 6 ] 3422. Cf. 3479 i ad mult gran damage; 1224 sin ad mult grant irur; 1987 en avrat grant damage; 2660 m’at fait guere mult grant.

[ 7 ] 1970. Cf. also 1680 ki puis veïst; 3483 ki dunc veïst; 1181 ki dune oïst.

[ 8 ] Cf. 1971 un mort sur altre geter; 3878 vait ferir l’uns li altre.

[ 9 ] 1980, 3925.

[ 10 ] The only evidence I can find here is line 1694 . . . veez gesir par tere.

[ 11 ] Cf. 1056 sanglant en ert.

[ 12 ] Cf. 994 des osbercs sarazineis; 1227, 1575 e l’osberc li derumpt; 1647 e l’osberc jazerenc; 1721 jamais entre sa brace; 3939 Tierri entre sa brace; 3250 de elme ne d’osberc.

[ 13 ] 1610. Cf. 1266 sun bon espiet; 2032 sur sun cheval.

[ 14 ] Cf. 647 Guenelun par l’espalle; 1826 el col un caeignun; and 1109 e li colps e li caples; 2206 le doel e la pitet; 2276 sun cors e sun visage; 2902 ma force e ma baldur.

[ 15 ] 176, 576, 586, 903, 1351, 1990, 2216, 2963, 3186, 3690, 3755, 3776.

[ 16 ] Cf. 2805 puis escriet: “Baruns, ne vos targez!“; 1681 de lur especs e ferir e capler; 1415 li .XII. per ne s’en targent nient; 338 quant aler dei, n’i ai plus que targer; 1366 kar de ferir; 1198, 1226, 1584, 3424 vait le ferir; 1092 par ben ferir.

[ 17 ] 262, 325, 547, 560, 826, 937, 948, 965, 1415, 2792, 3187; des 1308; les 1513, 3756.

[ 18 ] 1718. Cf. 681 nel devez pas blasmer; 1063 pur mei seient blasmet; 1174 ne funt mie a blasmer.

[ 19 ] 1416, 1835, 3476.

[ 20 ] Cf. 3475 ben i fierent e caplent; 1416 i fierent cumunement; 1681 e ferir e capler.

[ back ] 16. Oxford, laisse 105, lines 1338–1350; Venice IV, laisse 101, lines 1256–1267.

[ back ] 17. Oxford, laisse 105, lines 1338–1350; Chateauroux, laisse 144, lines 2277–2292.

[ back ] 18. Chateauroux, laisse 144, lines 2277–2292; Cambridge, laisse 39, lines 577–592.

[ back ] 19. John Mavrogordato, Digenes Akrites (Oxford, 1956), is the only English translation.

[ back ] 20. For an edition of all the manuscripts see Petros P. Kalonaros, ed., Basileios Digenis Akritas, vols. I and II. Important discussions of the Russian version are found in Russian Epic Studies, Memoirs of the American Folklore Society, vol. 42, edited by Roman Jakobson and E. J. Simmons (Philadelphia, 1949), and A. I. Stender-Реtersen, “O tak nazyvaemom Devgenievom Dejanii,” in Scando-Slavica, I (Copenhagen, 1954).

[ back ] 21. Henri Grégoire, Digenis Akritas (New York, 1942). See particularly the stemma on page 301.

[ back ] 22. K. Krumbacher, “Eine neue Handschrift des Digenis Akritas,” Sitzungsberichte der Koenigliche Bayerische Akademie der Wissenshaften, 2:309–355 (1904).

[ back ] 23. Grottaferrata 1.155–160 are equal to Athens 356–361 and E 21–29.

[ back ] 24. Notes to Chart XIII follow:

[ 1 ] 294, 604; cf. 1274: καὶ ὁ λαὸς τοῦ Ἀμηρᾶ; 2483: πατρὸς αὐτῆς τοῦ Ἀμηρᾶ; 4303: Μητέρα δὲ τοῦ Ἀμηρᾶ.

[ 2 ] Cf. 2280: ἡ μάννα τοῦ Ἀκρίτου; 4302: τὴν μάμμην τοῦ Ἀκρίτου; 2245: ἡ μήτηρ τοῦ Ἀκρίτου; 4304: πατὴρ ὁ τοῦ Ἀκρίτου; 3234: τοῦ Διγενοῦς τὸν λόyov.

[ 3 ]Cf. 1002: Ἡ μήτηρ τοῦ δὲ ὡς ἔμαθεν… and 1885: ὁ πατήρ μου τὸ ἔμαθε.

[ 4 ] There seems to be nothing very close.

[ 5 ] Cf. 3970: ἀπὸ Συρίαν ἅπασαν; 2496: εἰς τὴν Συρίαν ἠθέλησα.

[ 6 ] Cf. 2352: ἐπιστολὴν τοιαύτην.

[ 7 ]4367.

[ 8 ] Cf. 550: θλίψεών τε καὶ πόνων.

[ 9 ] There seems to be nothing very close.

[ 10 ] Cf. 2075: ἦσαν οἱ ἀδελφοί της; 3078: ἦτον ὁ λόγος οὗτος.

[ 11 ] 1.

[ 12 ] Cf. 466: μικρὰν παρηγορίαν; 471: κόσμου παρηγορίαν; 2009: οἶμοι τέκνον γλυκύτατον, φῶς καὶ παρηγορία; 2925: αὐτοῦ παρηγορίας; 3281: ψυχῶν παρηγορίαι.

[ 13 ] Cf. 718: οὐ χωρισθῆναι θέλω σοῦ; 465: πῶς νὰ σὲ ξεχωρίσωμεν ἐκ τῶν λοιπῶν σωμάτων, and 242: … ἀπ᾽ ἐμοῦ … and many other prepositions plus pronoun in the same position.

[ 14 ] Cf. 2797: ἐπῆγα’ εἰς τοὺς γονέους; 2805: ἐπῆγα εἰς ἄλλον τόπον; 1672: ἐπῆγεν εἰς κυνῆγιν; 979: κρυβέντες εἰς τὰ δάση; 1767: ἐξαίφνης εἰς τὰ νέφη; 1217, 1274; ἔφθασεν εἰς τὸ κάστρον; 2197; δοκοῦσα εἶναι ξένα.

[ 15 ] There seems to be nothing very close, but cf. 4111: καλύψαι σου τοὺς ὀφθαλμούς.

[ 16 ] Cf. 458: καὶ ἔσβεσας τὸ φῶς μας; 4260: ἠμαύρωσας τὸ φῶς μου; 4399: κ᾽ἐθάμπωσεν τὸ φῶς του.

[ 17 and 18 ] Cf. 578: δι᾽ οὗ πίστιν ἠρνήσατο καὶ συγγενεῖς καὶ φίλους; 848: ὁς δι᾽ ἐμὲ ἠρνήσατο γένος τε καὶ πατρίδα; 1067: πίστιν πατρίδα ἠρνήσατο καὶ συγγενεῖς καὶ φίλους; 1352: τὰ πάντα γὰρ ἠρνήσατο πίστιν τε καὶ πατρίδα; 1006: ὁμοίος δὲ οἱ συγγενεῖς; 1012: εἰς μέρος μὲν οἱ συγγενεῖς; 1881: μετέπειτα οἱ συγγενεῖς; 731: οἰκείος καὶ πατρίδα; 2777: γονεῖς τε καὶ πατρίδα.

[ 19 ] Cf. 3898: καὶ γέγονε περίφημος εἰς ἅπαντα τὸν κόσμον; 4286: καὶ φοβερὸς ἐγένετο εἰς ἅπαντα τὸν κόσμον; 4351: καὶ γέγονεν περίφημος εἰς ἅπαντα τὸν κόσμον; 3757: ἀνδρῶν γὰρ εἶναι ὄνειδος.

[ 20 ] 522. Cf. also 302: κ᾽ἐτράφη εἰς τὴν Συρίαν; 309: μέσα εἰς τὴν Συρίαν.

[ back ] 25. Notes to Chart XIV follow:

[ 1 ] Cf. 1.147: σπαθὶν διαζ ωσάμενος; 4.1071: ποδὸς αὺτοῦ δραζ άμενος; 6.216: σπαθὶν ἀράμενος αὐτοῦ; 6.252: καὶ τὸ σπαθὶν ἐπὶ τὴν γῆν.

[ 2 ] Cf. 4.132: ἔφθασε τὸ θηρίον; 4.136: βαλὼν τὸ θηρίον.

[ 3 ] 111, 4.816 (ὅτε).

[ 4 ] There seems to be nothing very close.

[ 5 ] There seems to be nothing very close.

[ 6 ] There seems to be nothing very close.

[ 7 ] There seems to be nothing very close.

[ 8 ] Cf. 1.150, 1.172: εἰς τὸν κάμπον ἐξῆλθε; 4.872: μετὰ λαοῦ ἐξῆλθε; 6.385: πρὸς Μαξιμοῦν ἀπῆλθε.

[ 9 ] Cf. 4.135: τὸ δὲ παιδίον σύντομα; 4.242: τὸ δὲ παιδίον εὔθιον; 4.913: καὶ τοῦ Χοσρόου τὸ σπαθίν; 6.61: καὶ ἐξελκύσας τὸ σπαθίν; 1.193: πόρρωθεν ρίπτει τὸ σπαθίν.

[ 10 ] Cf. 1.193, 1.200, 2.250, 4.684: χεῖρας εἰς ὕψος ἄρας; 6.74: εἰς ὕψος ὅλω τῷ θυμῷ τὸ σπαθὶν ἀνατείνας.

[ 11 ] Cf. 6.224: τῇ ράβδῳ κατὰ κεφαλῆς.

[ 12 ] 3.99.

[ 13 ] But cf. 6.653: καὶ μέχρι γῆς τὴν κεφαλήν.

[ 14 ] But cf. 6.259: μέσον τῶν δύο ὤμων.

[ back ] 26. Notes to Chart XV follow:

[ 1 ] 1539, 132: καὶ ὡς εἶδεν τούτους; 421: καὶ ὡς εἶδεν ὁ νεώτερος; 634: καὶ ὡς εἶδεν τὸν λέοντα; 789: καὶ ὡς τὸν εἶδεν; 1126: καὶ ὡς εἶδεν ὁ Φιλοπαπποῦς; 1359: καὶ ὡς τὸν εἶδεν ἡ Μαξιμοῦ.

[ 2 ] 566, 567, 1009, 1281, 1357; cf. 466: ἵνα καβαλλικεύσῃ.

[ 3 ] Cf. 1684: καὶ ἀπάνω κεῖται πιλωτόν; 430: καὶ οἱ πέντε ἐπιλαλήσαμεν; 1283, 1528: καὶ σύντομα ἐπιλάλησεν; 1262: καὶ τὴν καλήν μου ἐλάλησα.

[ 4 ] Cf. 768: ἵνα ραβδέα τοῦ δώσῃ; 1246: ἵνα ρaβδέαν μὲ δώσῃ; 1270: ἵνα σπαθεὰν μὲ δώσῃ; 1283: καὶ κονταρέαν μὲ δώσῃ; 1540, 1557: τὴν κονταρέαν μὲ δώσῃ.

[ 5 and 6 ] Cf. 941: καὶ κονταρέαν τὸν ἔδωκεν ὀμπρὸς εἰς τὸ μπροστοκούρβιν; 1452: καὶ κονταρέαν μ᾽ἔδωκεν τὴν φάραν εἰς τὰ μηρία; 1558: Σπαθέαν τῆς φάρας ἔδωκα, ἀπάνω εἰς τὰ κεφάλιν; 1727: καὶ ὡς ἔδωκα τὴν λέαιναν εὶς τὸ κεφάλιν; 975: … καὶ ἔδωκέν του ραβδέαν; 166: οὔτε φίλημαν μ᾽ἔδωκε.

[ 7 ] There seems to be nothing very close.

[ 8 ]1251. Cf. 1690: κaí στ’εκουν ίμπροσθίν του; 1265: ‘έμπροσθεν μου.

[ 9 and 10 ] 1286: καὶ ἐγὼ ταῦτα τὸν ἔλεγα, ἂς σηκωθῇ μὴ κεῖται; 1750: τοιοῦτον πάλιν λέγω σας; 1329: … πάλιν τὰ τοῖα λέγει.

[ 11 ] 1287. Cf. 1298: Ἑγείρου ἀπ᾽ αὔτου; 189: Ἐγείρου ἡ Βεργόλικος.

[ 12 ] 1287. Cf. 417: κοιτόμενον εἰς τὴν κλίνην.

[ 13 ] There seems to be nothing very close.

[ 14 ] Cf. 1101: . . . καὶ ὅπου κελεύεις ἔλα; 1379: καὶ ἂν θέλης, κυρά.

[ back ] 27. Digenis Akritas, lines 259–261. Mavrogordato translation.

[ back ] 28. Digenis Akritas, lines 387ff.

[ back ] 29. See William J. Entwhistle, European Balladry (Oxford, 1939).

[ back ] 30. See Roman Jakobson and Gojko Ruzičić, “The Serbian Zmaj Ognjeni Vuk and the Russian Vseslav Epos,” Annuaire de l’Institut de Philologie et d’Histoire Orientales et Slaves, X (Brussels, 1951), 343–355; and Roman Jakobson, “The Vseslav Epos,” Russian Epic Studies, Memoirs of the American Folklore Society, vol. 42 (1949), pp. 56ff.

[ back ] 31. J. B. Pritchard, ed., Ancient Near Eastern Texts Relating to the Old Testament, pp. 75–76.

[ back ] 32. Francis P. Magoun, Jr., The Gests of King Alexander of Macedon (Cambridge, Mass., 1929), p. 150, lines 778ff.

[ back ] 33. Henri Grégoire, “Le Digénis Russe,” Russian Epic Studies, Memoirs of the American Folklore Society, vol. 42 (1949), especially pp. 152ff.

[ back ] 34. William J. Entwhistle, “Bride-snatching and the ‘Deeds of Digenis’,” in Oxford Slavonic Papers (Oxford, 1953), IV, 1–12.

[ back ] 35. See Chapter Seven, note 7.