¹ See W. A. A. van Otterlo, Untersuchungen über Begriff, Anwendung und Entstehung der griechischen Ringkomposition, Mededelingen der Nederlandsche Akademie van Wetenschappen, Afdeling Letterkunde, NS 7, No. 3 (Amsterdam: Noord-Hollandsche Uitgevers Maatschappij, 1944); “Eine merkwurdige Kompositions-form der älteren griechischen Literatur,” Mnemosyne, 3rd ser., 12 (1944), 192–207; and De Ringcompositie als Opbouwprincipe in de epische Gedichten van Homerus, Verhandelingen der Koninklijke Nederlandsche Akademie van Wetenschappen, Afdeling Letterkunde, NS 51, No. 1 (Amsterdam: Noord-Hollandsche Uitgevers Maatschappij. 1948). Cedric Whitman develops and refines Otterlo's methods of structural analysis in his book Homer and the Heroic Tradition (Cambridge: Harvard Univ. Press, 1958), pp. 87–101, 249–84, 287–90, and the chart at the end of the book, as does Julie Haig Gaisser in “A Structural Analysis of the Digressions in the Iliad and the Odyssey,” Harvard Studies in Classical Philology, 73 (1969), 1–44. See also my note “On the Design of the Hymn to Delian Apollo,” forthcoming in Classical Journal, 75 (1979–80). For examples of the extension of similar methods of analysis to literatures other than early Greek, see Michael Fishbane, “Composition and Structure in the Jacob Cycle (Gen. 25:19–35:22),” Journal of Jewish Studies, 36 (1975), 19–32; John D. Niles, “Ring-Composition in La Chanson de Roland and La Chançun de Willaine,” Olifant, 1 (Dec. 1973), 4–12; and David Buchan, The Ballad and the Folk (London: Routledge. 1972), pp. 87–144, 230.

² D. B. Monro, ed., The Iliad, 5th ed. (Oxford: Oxford Univ. Press, 1963). The translation is my own.

³ Bartlett, The Larger Rhetorical Patterns in Anglo-Saxon Poetry, Columbia University Studies in English and Comparative Literature, No. 122 (New York: Columbia Univ. Press. 1935), pp. 9–29. Bartlett cites the following lines from Beowulf as exhibiting envelope patterning of a simple or complex sort: ll. 115–25, 129–93, 237–57, 491a–646a, 767–70, 837–924, 1063–162a, 1323b–29, 1384b–89, 1425–4la, 1441b–72, 1591–622, 2247–66, and (depending on the emendation “næfne” for “næs he” in 1. 3074b) 3051–75. On envelope patterning, see also Constance B. Hieatt. “Envelope Patterns and the Structure of Beowulf,” English Studies in Canada, 1 (1975), 249–65. In developing Bartlett's mode of analysis, Hieatt pays close attention to the echoing of isolated words. She thus disregards Bartlett's warning that “the mere repetition of a phrase or phrases within the space of ten or twenty or a hundred lines does not suffice to establish the presence of what is here called the Envelope pattern. The group of verses must be a real group to the mind of the reader as well as to his ear and eye” (Bartlett, p. 9). And again: “It must not be forgotten that perfect verbal agreement without content unity does not constitute an Envelope pattern. Repetition alone is, as group pattern, almost meaningless” (Bartlett, p. 24). On mere verbal echo in Beowulf. see J. O. Beaty, “The Echo-Word in Beowulf with a Note on the Finnsburg Fragment,” PMLA. 49 (1934), 365–73. Another recent study of structural patterning in Beowulf is David R. Howlett. “Form and Genre in Beowulf” Studia Neophilologica. 46 (1974), 309–25 (esp. pp. 318–25). Howlett attempts to identify a number of examples of thematic envelopment in Beowulf. In his close attention to the line numeration of the text and to the division of the text into numbered fits, he appears to be influenced by the numerological analyses of Thomas E. Hart, “Ellen: Some Tectonic Relationships in Beowulf and Their Formal Resemblance to Anglo-Saxon Art,” Papers on Language and Literature, 6 (1970), 263–90, and “Tectonic Design, Formulaic Craft, and Literary Execution: The Episodes of Finn and Ingeld in Beowulf” Amsterdamer Beit rage zur älteren Germanistik, 2 (1972), 1–61. Space does not permit a review of the dangers attendant on the numerological analyses of medieval texts, but it should be borne in mind that, unlike the fit divisions, the line numeration of Beowulf is a modern invention. In the manuscript, as in all Old English poetic manuscripts, the lines are written not separately but continuously, like prose. In view of the lacunae in the text, furthermore, the modern lineation of the poem seems rather problematical. A more recent and more successful attempt to find structural symmetry in an Old English poem is Earl Anderson, “Cynewulf's Elene: Manuscript Divisions and Structural Symmetry,” Modern Philology, 72 (1974), 111–22. One further article on this subject appeared after the present paper was completed, and I can do little more than call attention to it here: H. Ward Tonsfeldt, “Ring Structure in Beowulf,” Neophilologus, 61 (1977), 443–52. Although his methods of analysis are similar to my own and to those of Hieatt and Howlett, Tonsfeldt does not cite either of their studies or my “Ring Structure in Beowulf and in Oral Poetry,” the chapter of my doctoral dissertation on which this article is based (see “Aspects of the Oral Art of Beowulf: A Comparative Investigation,” Diss. Univ. of California, Berkeley, 1972, Ch. v, pp. 154–90). After analyzing the ring structure of ll. 129b–49a, 237–70, 1017–168, 1885b–924, 2355–72, and 2426–512a, Tonsfeldt concludes that “the poet does not use ring structure consistently enough or determinately enough to justify any very extensive claims” (p. 452).

⁴ All references to the text of Beowulf are to Frederick Klaeber, ed., Beowulf and the Fight at Finnsburg, 3rd ed., with 1st and 2nd supps. (Lexington, Mass.: Heath, 1950). Diacritical marks are omitted, and all translations are my own.

⁵ For a discussion of the symmetry of the two sea voyages see my “Aspects of the Oral Art of Beowulf.” pp. 59–60 and 167–70; Lee Carter Ramsey, “The Sea-Voyages in Beowulf,” Neuphilologische Mitteilungen, 72 (1971), 51–59; and Tonsfeldt, p. 445.

⁶ Bartlett, pp. 19–22; Hieatt, pp. 250–51, 254–56, 258–59, and the diagram on p. 260; and Howlett, diagram on p. 324.

⁷ It may be useful to relate my concerns in this essay to those of John Nist, “The Structure of Beowulf,” Papers of the Michigan Academy of Science, Arts, and Letters, 43 (1958), 307–14; Thomas E. Carrigan, “Structure and Thematic Development in Beowulf.” Proceedings of the Royal Irish Academy. 66 (1967), Sec. C, pp. 1–51; and John Leyerle. “The Interlace Structure of Beowulf,” University of Toronto Quarterly, 37 (1967). 1–17. All three authors find evidence of detailed structural parallelism and balance in the poem. Nist shows how certain themes (such as “Grendel motifs” or “details of Beowulf's life”) recur in a complex and apparently meaningful sequence, which he calls “spiral” design. Carrigan analyzes the poem into a discrete number of fit groupings and shows how they balance one another thematically. In addition, he shows that one such grouping (covering the events of the dragon fight) has a similar balanced structure on a smaller scale. Leyerle points out how certain individual themes (e.g., the Frisian raid, monsters, or women as the bond of kinship) interweave with one another in a manner that he finds analogous to the interlace design characteristic of much Anglo-Saxon art. Of the three studies Leyerle's seems to me the most rewarding, and yet both his and Nist's methods of analysis are open to the objection that the recurrence of certain themes in Beowulf is in itself of little significance. In such a long and complex work, some themes are bound to recur. What one wants to know is, Is there a meaningful pattern to this recurrence? Leyerle would say yes, but he does not identify this pattern except in general terms. My essay might be regarded as an attempt to supplement Leyerle's case for the interlaced structure of Beowulf by reference to the specific structure of the hero's three great fights. From this point of view, ring composition (like parallel composition) might be looked on as one special type of interlace design.

⁸ On the interplay between Æschere's head and Grendel's, see Hieatt, p. 257, writing in response to Stanley B. Greenfield, The Interpretation of Old English Poems (London: Routledge, 1972), pp. 58–59.

⁹ See, in addition, Christopher Knipp, “Beowulf 2210b–2323: Repetition in the Description of the Dragon's Hoard,” Neuphilologische Mitteilungen, 73 (1972), 783, and Carrigan, p. 44. In Fig. 2 of a plate appended to the end of his essay, Carrigan diagrams his conception of the design of the dragon fight (ll. 2221–3182), with particular reference to speeches before and after the fight, to the roles of Wiglaf and of the last survivor, and to allusions to Beowulf's peaceful reign; see also Carrigan, pp. 30–49. On how the “lay of the last survivor” helps to prepare the way for the dominant mood of the end of the poem, see also Adrien Bonjour, The Digressions in Beowulf, Medium Ævum Monographs, No. 5 (Oxford: Blackwell, 1950), pp. 68–69.

¹⁰ J. R. R. Tolkien, “Beowulf: The Monsters and the Critics,” Proceedings of the British Academy, 22 (1936), 245–95; rpt. in The Beowulf Poet: A Collection of Critical Essays, ed. Donald K. Fry (Englewood Cliffs, N.J.: Prentice-Hall, 1968), pp. 34–35.

¹¹ Lord, The Singer of Tales (Cambridge: Harvard Univ. Press, 1960), p. 201.

¹² Gabriel Germain, Genèse de l'Odyssée (Paris: Presses Universitaires de France, 1954), p. 333; Whitman, p. 288.

¹³ Grendel, for example, is introduced as a “fiend in hell” ‘feond on helle’ (l. 101b) and is elsewhere called by names that seem modeled on Latin terms denoting the Christian devil—“helle hæfta” (captivus inferni), “ealdgewinna” (hostis antiquus), and “feond mancynnes” (hostis humani generis). During his struggle with Beowulf, Grendel wishes to flee “to seek out the throng of devils” ‘secan deofia gedreag‘ (l. 756a), as if his pool were to be identified with the Christian underworld. For a further discussion of the poet's association of Grendel with the Archfiend, see Marijane Osborn, “The Great Feud: Scriptural History and Strife in Beowulf” PMLA, 93 (1978), 975–77.

¹⁴ For a tenth-century Latin version, see H. C. Kim, ed., The Gospel of Nicodemus: Gesta Salvatoris (Toronto: Pontifical Institute of Mediaeval Studies, 1973). There is also an Old English version dating from the tenth or eleventh century: The Gospel of Nicodemus. ed. S. J. Crawford (Edinburgh: Hutchen, 1927). The theme of the descensus, which was probably widely known in Anglo-Saxon England by the early eighth century, recurs frequently in Old English devotional literature, notably in the poem known as “The Descent into Hell,” which is included on fols. 119b–21b of the Exeter Book (see George Philip Krapp and Elliott Van Kirk Dobbie, eds., The Exeter Book [New York: Columbia Univ. Press, 1936], pp. lxi–lxiii, 219–23, 356–59). We need not conclude, however, that Grendel's mere is to be regarded as a symbolic representation of hell or that Beowulf's descent into the waters is to be read as a thinly veiled allusion to Christ's harrowing of hell and to the related liturgy of baptism, as is argued independently by Allen Cabaniss in “Beowulf and the Liturgy,” Journal of English and Germanic Philology, 54 (1955), 195–201, and by Maurice B. McNamee, S.J., in “Beowulf—An Allegory of Salvation?” Journal of English and Germanic Philology, 59 (1960), 190–207. Such parallels as exist between the two stories fall short of the sort of precision that one would expect if Beowulf were a conscious allegory.

¹⁵ See Klaeber, n. to 1. 1357, for the relevant passage, which is based on a Latin Visio Pauli.

¹⁶ Scholars who have supposed Vergilian influence in Beowulf have tended to concentrate their attention on Aeneid vi.237–42 and Beowulf, ll. 1357b–76a (the description of Grendel's mere); again see Klaeber, n. to l. 1357. Alan Renoir compares the artistry of the two passages without making any commitment about whether or not the Beowulf poet knew Vergil (“The Terror of the Dark Waters: A Note on Virgilian and Beowulfian Techniques.” in Larry D. Benson, ed., The Learned and the Lewed: Studies in Chaucer and Medieval Literature, Harvard English Studies, No. 5 [Cambridge: Harvard Univ. Press, 1974], pp. 147–60). For a review of earlier scholarship and a sensitive attempt to find new evidence for Vergilian influence, see Theodore M. Andersson, “The Virgilian Heritage in Beowulf.” Ch. iv of his Early Epic Scenery: Homer, Virgil, and the Medieval Legacy (Ithaca: Cornell Univ. Press, 1976), pp. 145–59.

¹⁷ For a recent revival of the analytical heresy concerning the structure of Beowulf, see Francis P. Magoun. Jr.. “Béowulf Á: A Folk-Variant,” Arv, 14 (1958), 95–101, and “Béowulf B: A Folk-Poem on Béowulf's Death,” in Early English and Norse Studies Presented to Hugh Smith in Honour of His Sixtieth Birthday, ed. A. Brown and P. Foote (London: Methuen, 1963), pp. 127–40. That the action of Beowulf is based on two kinds of tale (the “bear's son” type and the dragon-slayer type) is beyond dispute: see Gwyn Jones, Kings, Beasts and Heroes (London: Oxford Univ. Press, 1972), pp. 6–19. Few modern scholars follow Magoun, however, in seeing the extant text of Beowulf as the work of two or three authors.

¹⁸ Certain questions must remain beyond the scope of the present inquiry. The question of the relation of the extant text of Beowulf to oral tradition, for example, is too complex and problematical to be entered into here. Clearly, patterning of any sort would be mnemonically useful to an oral poet or to a performer of oral poems, just as it is useful to any stage performer (whether singer, storyteller, actor, musician, or nightclub entertainer) who does not rely on a fixed text as the basis for his performance. Ring composition could serve as one elementary type of mnemonic patterning: as Whitman observes, “the oral poet, having mentioned A, B, and C, picks them up later on in the order C, B, and A, since it is natural to reconstruct a train of thought backwards” (p. 98). Whitman's conclusion that the complex ring structure of the Iliad marks the poem as a work of oral literature, however, does not necessarily follow. Such an argument is based on a priori reasoning rather than on a careful examination of indubitably oral texts, for oral texts recorded in the field have not been shown to exhibit complex ring structures of the same kind. Perhaps these structures exist, but the necessary fieldwork and analysis have not been done. Of the works referred to in n. 1 of this paper, for example, none (not even Buchan's ballad texts) could be proved to be exact records of oral tradition unmediated by literary hands, although all appear to have a close relation to oral tradition. Until such fieldwork and analysis are done, the presence of various sorts of ring composition in Beowulf may plausibly be taken as evidence that, like the Iliad and the Song of Roland, the poem shows traits that would be useful to an oral poet or performer, but further conclusions seem premature. Moreover, even polished literary works (such as Tom Jones and Paradise Lost) may show complex ring patterning, as R. G. Peterson reminds us with appropriate bibliographical citations in his outstanding essay “Critical Calculations: Measure and Symmetry in Literature.” PMLA, 91 (1976), 367–75. A person who wished to make a case for the specifically oral character of the ring patterning of Beowulf would have to distinguish between the sort of patterning in this poem and the sort in, for example, Tom Jones.

¹⁹ That the poem was intended to be performed aloud (as opposed to being created aloud) is beyond doubt. See Dorothy Whitelock. The Audience of Beowulf (Oxford: Oxford Univ. Press. 1951), passim, and Kenneth Sisam. The Structure of Beowulf (Oxford: Oxford Univ. Press, 1965), pp. 1, 8–10, et passim; and note further Ruth Crosby, “Oral Delivery in the Middle Ages,” Speculum. 11 (1936). 88–110.

²⁰ In fact one need not assume that such expectations were unconscious, although I do not wish to press the point here. An audience that had heard the same story not once but often, with variations, might have become sufficiently discerning to appreciate subtle instances of thematic echo. The Beowulf poet, moveover, took pains to point out how the parts of his story interrelate, as in ll. 2349b–54a, an allusion to the Grendel fight, and in l. 2521b, a direct reminiscence of the Grendel fight at the moment before Beowulf sets out to challenge the dragon.

²¹ See Herbert G. Wright, “Good and Evil; Light and Darkness; Joy and Sorrow in Beowulf”Review of English Studies, NS 8 (1957), 1–11, and Jerome Man-del, “Contrast in Old English Poetry,” Chaucer Review, 6 (1971), 1–13.

²² Joan Blomfield, “The Style and Structure of Beowulf,” Review of English Studies, 14 (1938), 397.

²³ I should like to thank John Leyerle, Dean of the School of Graduate Studies at the University of Toronto, for having read this paper sympathetically and for having made a number of suggestions for its improvement.