"Arts, Faculty of"@en . "English, Department of"@en . "DSpace"@en . "UBCV"@en . "Mason, John Neilson"@en . "2010-02-09T22:41:15Z"@en . "1976"@en . "Master of Arts - MA"@en . "University of British Columbia"@en . "Nineteenth-century scholars generally felt that the Exeter Book riddle collection was a unified whole under the authorship of Cynewulf, or that it was made up of two major parts, Riddles \"1\" (now known as \"Wul'f and Eadwacer\") to 59 and 61-95. Most scholars since the first decade of this century, however, have viewed the riddles as a miscellany, with a few individual riddles perhaps sharing common authorship, but with no overall unity or organization in the collection as a whole.\r\nIf the riddles are examined in terms of their point of view (I am..., I saw..., There is...), a distinct pattern emerges which demonstrates Riddles 61-95 to be separate from the rest, and which also divides Riddles 1-59 into two more or less equal groups. The distribution of point of view does not indicate the exact point of division between the first two groups, but if the groups originally comprised 60 riddles (like the collection of Eusebius), and if the two groups are assumed to have been equal collections of 30, then deduction based on the amount of missing material due to the loss of folios between fols. 105 and 106, and between 111 and 112, would locate the break between Riddles 29 and 30. Riddle 30b, then, could have been simply a mis-start of the second group at a point later in the MS. Examination of the distribution of opening and closing formulas arid of the adverbs hwilum, oft and nu over the collection supports the three-part theory.\r\nStylistic diversity in the third group, from crude riddles like Nos. 75 and 76 to the fine 'horn' and 'water' riddles suggests that\r\n\r\nthis group is a miscellany containing the work of a number of authors. Connections between riddles of this group and the two earlier ones appear to indicate some sort of dependence of these on riddles of the first two groups. The relationship in several of the cases can be explained as imitation or modelling of the later riddles on earlier ones. Such a suggestion is not inconsistent with practice at the time, as other riddles of the period appear to have been used as exercises in grammar."@en . "https://circle.library.ubc.ca/rest/handle/2429/19945?expand=metadata"@en . "AUTHORSHIP AND UNITY IN THE EXETER BOOK RIDDLES by John Neilson Mason B.A., University of British Columbia, 1972 A THESIS SUBMITTED IN THE REQUIREMENTS MASTER PARTIAL FULFILLMENT OF FOR THE DEGREE OF OF ARTS in THE FACULTY OF GRADUATE STUDIES Department of English We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA May, 1976 \u00C2\u00A9 John Neilson Mason, 1976 In presenting th i s thes is in pa r t i a l fu l f i lment of the requirements for an advanced degree at the Un ivers i ty of B r i t i s h Columbia, I agree that the L ibrary sha l l make i t f ree l y ava i l ab le for reference and study. I fur ther agree that permission for extensive copying of th i s thes i s for scho lar ly purposes may be granted by the Head of my Department or by his representat ives. It is understood that copying or pub l i ca t ion of th i s thes i s fo r f i nanc ia l gain sha l l not be allowed without my wr i t ten permission. Department of ENGLISH The Univers i ty of B r i t i s h Columbia 2075 Wesbrook Place Vancouver, Canada V6T 1W5 Date MAY 7, 1976 - i i -ABSTRACT Nineteenth-century scholars generally felt that the Exeter Book riddle collection was a unified whole under the authorship of Cynewulf, or that i t was made up of two major parts, Riddles \"1\" (now known as \"Wul'f and Eadwacer\") to 59 and 61-95. Most scholars since the f i r s t decade of this century, however, have viewed the riddles as a miscellany, with a few individual riddles perhaps sharing common authorship, but with no overall unity or organization in the collection as a whole. If the riddles are examined in terms of their point of view (I am..., I saw..., There i s . . . ) , a distinct pattern emerges which demonstrates Riddles 61-95 to be separate from the rest, and which also divides Riddles 1-59 into two more or less equal groups. The distribution of point of view does not indicate the exact point of division between the f i r s t two groups, but i f the groups originally comprised 60 riddles (like the collection of Eusebius), and i f the two groups are assumed to have been equal collections of 30, then deduction based on the amount of missing material due to the loss of folios between fols. 105 and 106, and between 111 and 112, would locate the break between Riddles 29 and 30. Riddle 30b, then, could have been simply a mis-start of the second group at a point later in the MS. Examination of the distribution of opening and closing formulas arid of the adverbs hwilum, oft and nu over the collection supports the three-part theory. Stylistic diversity in the third group, from crude riddles like Nos. 75 and 76 to the fine 'horn' and 'water' riddles suggests that \u00E2\u0080\u00A2oh:' -this group is a miscellany containing the work of a number of authors. Connections between riddles of this group and the two earlier ones appear to indicate some sort of dependence of these on riddles of the fir s t two groups. The relationship in several of the cases can be explained as imitation or modelling of the later riddles on earlier ones. Such a suggestion is not inconsistent with practice at the time, as other riddles of the period appear to have been used as exercises in grammar. - iv -CONTENTS Page CHAPTER I THE CYNEWULFIAN CONTROVERSY 1 Footnotes to Chapter I 18 CHAPTER II IS THERE UNITY IN THE COLLECTION? 21 Treatment of Material 21 Unity and Sources 27 Arguments Based on Date 29 Footnotes to Chapter II 35 CHAPTER III SECTIONS OF THE TEXT AS SEPARATE UNITS 37 Evidence for Distinct Groups in the Collection 37 The Original Groupings Ul Opening and Closing Formulas, and Three Adverbs i|8 The Relationship Between the Groups 57 Footnotes to Chapter III 66 CONCLUSION 69 LIST OF WORKS CONSULTED 71 APPENDIX A 77. APPENDIX B 78 APPENDIX C 80 _ v -LIST OF TABLES Table I Distribution of Points of View p. 38 Table II Distribution of Opening and Closing Formulas p. 50 Table III Frequency of Opening and Closing Formulas p. 51 Table IV Occurrence of hwilum, oft and nu i . Adverb Begins Half-Line p. 53 i i . Scribal Point Precedes Adverb p. 51i - vi -LIST OF FIGURES Figure 1 Forster's Collation of Gatherings XIV, XVI and XVII p. hh Figure 2 Possible Loss from Gathering XVII p. kh - v i i -ABBREVIATIONS REB Frederick Tupper, ed., The Riddles of the Exeter Book (Boston: Ginn and Co., 1910). K-D G.P. Krapp and E.vK. Dobbie, eds., The Exeter Book (New York: Columbia University Press, 1936). EBOEP R.W. Chambers, Max Forster and Robin Flower, eds., The Exeter Book of Old English Poetry (London: Lund, Humphries and Co., 1933). - y i i i -Note on Numbering The four, most widely-used editions of the Exeter Book riddles today are those of Tupper, Wyatt, Mackie and Krapp and Dobbie; earlier scholars usually used Grein's edition. 1 Unfortunately, a l l these editions vary in their systems of numbering the riddles. Tupper includes \"Wulf and Eadwacer\" as his Riddle 1, but presents K-D 68 and 69 as one riddle. Wyatt and Mackie also group 68 and 69, but Mackie omits K-D 90, the Latin Riddle. Grein includes \"Wulf and Eadwacer,\" but omits the fragments K-D 78, 82, 89, 92 and 9U. A l l numbering in this study will follow the Krapp and Dobbie system, but the following key can be used for adjusting the other systems to K-D numbering. THEE er K' -D 1-68 = T 2-69 (K-D = T-l) K-D 69-95 = T 69-95 Wyatt K-D 1-68 = W 1-682 K-D 69-89 = W 68 3-88 (K-D - W+l) K-D 91-95 = W 89-93 (K-D = W+2) Mackie ,2 K-D 1-68 = M 1-68' K-D 69-95 = M 683-94 (K-D = M+l) Grein K-D 1-66 \u00C2\u00BB G 2-67 K-D 68 = G 68 1\" 2 K-D 69-77 = G 68J-76 K-D 79-81 = G 77-79 K-D 83-88 = G 80-85 K-D 90-91 = G 86-87 K-D 93 = G 88 K-D 95 = G 89 1 C W M Grein, Bibliothek der Angelsachsischen Poesie (Goetingen: Georg H. Wigands Verlag, ltJbo), n , Joy-u07-CHAPTER I THE CYNEWULFIAN CONTROVERSY A study of authorship and unity i n the Exeter Book riddles i s inextricably involved with the Cynewulfian controversy. This i s the scholarly dispute which raged i n the nineteenth century, and a l i t t l e into the twentieth, i n which opinions were voiced concerning possible authorship by Cynewulf of almost every piece of Anglo-Saxon poetry extant. The riddles were at the heart of the Cynewulfian dispute, due to charades on the name Cynewulf which were supposedly found i n two of the riddles, and on alleged biographical details i n another. These charades were thought to be similar to the runic signatures found i n the four 'signed' Cynewulfian poems. The idea of Cynewulfian authorship of the riddles came for a while to be regarded almost as an established fact, but began losing favor i n the 1880*s. Many scholars, s t i l l held to the theory, though, and i t continued to be revived sporadically u n t i l 1910. The basis of the theory which assigned authorship to Cynewulf, indeed v i r t u a l l y the only piece of evidence upon which a l l the later opinions were grounded, was Heinrich Leo's solution of the so-called 1 v r 'First Riddle.' This poem, which appears on Folios 100 and 101 of the Exeter Book, and which immediately precedes i n the MS the f i r s t group of what are now considered by most scholars the true riddles, i s now generally known as \"Wulf and Eadwacer.\" The inspiration for Leo's solution took place i n 18U0, with the almost simultaneous, though completely independent discoveries, by 2 3 Jacob Grimm and John M. Kemble, that the runes at the end of the - 2 -Old English poems \"Elene,\" \"Juliana,\" and \"Christ II,\" when properly arranged, formed the name Cynewulf in the fi r s t two cases, and Cynwulf in the third. Grimm and Kemble suggested that the person in question might be Cenwulf, Abbot of Peterborough, who flourished in the early tenth century; Grimm assigned no historical significance to the name, but supposed Cynewulf to have been an eighth-century Northumbrian, per-haps a pupil of Aldhelra. The discovery of an author's name connected with Anglo-Saxon verse, in contrast to the usual complete anonymity of the poetry, was an exciting find which prompted a minute search over the next few decades for more hidden signatures in Old English poetry. As well, specifically Cynewulfian features were claimed for many unsigned poems, and scholars tentatively assigned more and more works to Cynewulf. An early product of this research, and one which added considerable impetus to further investigation, was Leo's 'discovery' in 18\u00C2\u00A37^ of a charade in the First Riddle, which, he claimed, revealed the name of Cynewulf. By a process of emendation and twisting of meaning to suit his purpose, , Leo produced an ingenious interpretation which finds the whole poem to be a play on the syllables of Cynewulf's name. I reproduce here the original and Leo's interpretation. For Morley's English trans-lation of Leo see Appendix A. - 3 -. Leodum i s minura swylce him mon lac gifej willa3 hy hlne abecgan, g i f he on breat cyme5. Ungelic i s us. Wulf i s on iege, i c on oberre. Faest i s baet eglond, fenne biworpen. Sindon waelreowe weras baer on ige; willa8 hy hine apecgan, g i f he on preat cyme5. Ungelice i s us. Wulfes i c mines widlastum wenum dogode; bonne h i t waes renig weder ond i c reotugu saet, bonne mec se beaducafa bogum bilegde, waes me wyn to pon, waes me hwaebre eac lad. Wulf, min Wulf, wena me pine seoce gedydon, pine seldcymas, murnende mod, nales meteliste. Gehyrest pu, Eadwacer? Uncerne earne hwelp bireo wulf to wuda. ?aet mon eabe tosliteS j>aette naefre gesomnad waes, uncer giedd geador. Meine Glieder verhalten sich wie man ihnen Bedeutung zutheilt; Sie werden dieselbe offenbaren, wenn die Bedeutung sich zusammenschaart. (2) Ungleich verhSlt sich's mit uns. Ein Wolf i s t auf einer Insel, ich auf der andern: Vollkommen i s t die Insel mit Sumpfland umgeben. Wilde Manner sind hier auf dem Eilande. Sie werden dieselbe offenbaren, wenn er mit (ihrer) Schaar zusammenkommt. (3) Ungleich verh< sich's mit uns. Ich gebe mich den weitgehenden Sehnsuchten nach meinem Wolf hin. Wenn es regniges Wetter war und ich weinend sass, Dann umfasste mich der Kampfschnelle mit seinen Armen. Das ward mir Wonne, ward mir doch auch Leid. Wolf! mein Wolf! die Sehnsuchten nach dir Haben mich krank gemacht, deine seltenen Besuche; Das trauernde Gemuth (that's), nicht durch Nahrungsmangel. H6rst du? Eadwaccer, unserer beider Jungen, tr'agt ein Wolf zum Holze. (U) Das sondert man leicht aus einander, was nie Zusammenhang hatte, Unserer beider gemeinschaftliches Lied.^ - h -Leo's Olieder, 'limbs,' are the f i r s t and last syllables of Cynewulf's name: the reader is to guess them by recognizing their representations through homonyms and parallel meanings, which Leo supposes to be worked into various parts of the riddle. After the opening section (Leo's division l ) , the first syllable speaks. The riddler metaphorically demonstrates the separation of the syllables by placing them on separate islands. The waelreowe weras, 'cruel or bloodthirsty men,' on the f i r s t syllable's island are to be interpreted as cene, 'keen or fierce men,' which the solver i s then to interpret as cyne, identifying the fi r s t part of the name. In the third section cyne is again identified, this time in coen, or cwen, which, Leo says, is the woman who is in love with a man named Wulf. Wuda, 'wood,' is assumed to stand for cen, 'pine torch.' Leo claims that i f cen i s though of as 'split wood,' the association is simple. Cen, which is the C-rune in Cynewulf\u00C2\u00BBs runic signature, is to represent cyn. The word Eadwacer Leo interprets as merely a personification of the letter e, which joins cyn and wulf; Wulf carries Eadwacer to cyn, the wood. According to Leo the closing section of the poem means that though the charades cyne, coen and cwen are joined together (in the substance of the charade), they are easily taken apart by the solver, as they are after a l l not identical in sound. On the strength of the evidence of a single poem, Leo is cautious about crediting the entire Exeter Book riddle collection to Cynewulf, but i f not a l l the riddles are Cynewulfs, he says, those whose answers are the names of runic letters probably are. ! Dietrich, in his f i r s t article on the riddles,\" fully supports Leo's solution of the First Riddle as a charade on Cynewulfs name, and finds further support for Cynewulfian authorship in two more of the enigmas. In the Latin Riddle he finds the repetition of the word lupus strongly reminiscent of the \"wulf min wulf\" phrase and of the other repetitions of the word wulf in the First Riddle, and feels these suggest some sort of charade. The last riddle Dietrich solves as 'wandering singer.' This riddle, he concludes, again refers to Cynewulf, this time autobiographically as a young itinerant scop. The whole collection cannot possibly be by Cynewulf, though, Dietrich says, because in Riddles 61-95 several of the earlier riddle objects are handled a second time, and in a way similar to their treat-ment in the fi r s t section. A good poet, he reasons, would not repeat himself. Also, as the riddles are separated in the manuscript into two distinct groups, with one additional riddle and the repeat of another inserted along with other material between, the likelihood of unity is further reduced. Yet,' he says, at least two poems of Cynewulfs appear in the final portion. To account for this, Dietrich proposes that the collection comprises the work of two or more authors, the f i r s t group of riddles being Cynewulfs alone, and the last group a collection including single riddles of Cynewulfs, but with the majority by another poet or poets. 9 In his second article, however, Dietrich expands his claims, this time assigning the entire collection to Cynewulf, and going so far as to propose that the f i r s t group is the product of Cynewulfs youth and the second of his maturity. In the f i r s t group he finds - 6 -evidence of the poet's youthfulness i n the mistaken translation of Aldhelm's pernix aquilis, 'swift eagle,' as pernex, an imaginary bird, i n line 65 of Riddle 1+0; i n the youthful cadence of the poetry; and i n the presence of five 'obscene' riddles, 25, U2, UU, ii5, 5U\u00C2\u00AB Dietrich seeks to support his claims by comparing thoughts and expressions i n the Riddles with other Cynewulfian poems. But, as Tupper remarks, \"The larger number of parallels (granting that such parallelism carries any weight) are drawn from a text of such doubtful authorship as the Andreas. The charade solution was generally accepted for the next 20 years, the only dissenting voice being Reiger's i n 1869. Reiger*\"'\" feels that Leo's argument for extracting cene as the f i r s t part of the name from waelreowe i s \"too thin.\" He also rejects Leo's problematical rendering of the f i r s t l i n e , preferring to find the f i r s t - s y l l a b l e charade i n leodum minum (which Leo had emended to leo&um minum), which he interprets to signify cynn. Reiger admits some d i f f i c u l t i e s , however. In the sense of the passage, he says, leodum would be closer i n meaning to dryht than to cynn. To account for the e i n the f i r s t charade, he adapts his found-word to the form cynne. Similarly, he adapts coen and cwen to coene and cwene. In the third charade he follows Leo i n supposing the f i r s t syllable to be found i n wuda = cen, with the missing e f i l l e d i n by Eadwacer. But whereas Leo had read Eadwacer uncerne earne hwelp as object of bire5 (1. 17), with wulf the subject and gehyrest bu a rhetorical question, Reiger argues that the question mark should be placed after Eadwacer, making Eadwacer (= e) the object of gehyrest. The subject of bire5 would now be uncerne earne hwelp. The result of the adjustment i s that i t i s not now the e that i s taken to the wood and thereby connected to cen, but wulf. Reiger translates the f i n a l lines of the poem: \"Do you hear Eadwacer, our angry whelp? He carries the wolf to the wood, (one easily breaks apart what was never put 12 together) our riddleword together.\" The parenthetical l i n e he interprets as an aside comment by the riddler on the art of charade-making. Despite his differences with Leo, Reiger does not dispute the idea that the piece i s a charade, or that the solution i s the name Cynewulf. In 1883 Trautmann made the f i r s t solid attack on Leo's solution 13 to the Fi r s t Riddle. His argument focusses on LU objections, some of which do not wholly exclude the p o s s i b i l i t y of Leo's being right, but none of which can be ignored. Cumulatively they were devastating to Leo's interpretation. Trautmann's objections are: 1. There exist no other syllable riddles of the form \"my f i r s t i s . . . , my second i s . . . , my whole i s . . . \" i n Old English, nor indeed i n the entire literature of the Middle Ages.-^ 2i Leo's f i r s t two lines are a shambles: a. I t i s unjustifiable to emend leodum to leoSum, since i t makes good sense as i t stands. b. swylce does not mean simply 'as,' but 'as i f . 1 c. lac does not mean 'meaning,' but ' g i f t . ' d. abecgan does not mean 'reveal,' but 'take up.' e. nine, he cannot refer to lac, as lac i s neuter. 3. a. Trautmann doubts that the f i r s t two lines indicate a charade at a l l . In a l l three places where the rune signature exists\u00E2\u0080\u0094\"Elene,\" \"Christ,\" and \"Juliana,\"\u00E2\u0080\u0094the second l e t t e r of the name i s given by the rune Yr, so the f i r s t part of the name cannot be assumed to be anything but cyn: not cen, coen or espe-c i a l l y cwen. b. Riddles are by tradition directed more to the ear than to the eye. U. Line 5 faest i s baet eglond fenne beworpen i s ignored i n Leo's translation. - 8 -3>. The supposed extraction of cene from waelhreowe is much too difficult. 6. paer on ige cannot be translated as 'here onfmy island.' 7. Leo's translation of abecgan as 'reveal' is wrong. It is a forced interpretation, dependent on the meeting of the wild men with the wolf, when actually Leo says they are on different islands. 8. The translation of wen as longing is wrong. 9. Trautmann doesn't see why a lamenting woman would call her beloved beaducafa, when more appropriate would be husband, friend, beloved, splendid one, untrue one, cruel one, inconstant one, or some such. 10. waes me hwae&re eac laS i s not understandable by Leo's interpretation: a longing woman would feel joy, not pain, i f her beloved came and embraced her. 11. It is possible that the e of Cynewulf's name can be called the 'bond' between the first and third syllables, but that this letter should denote the offspring would be rare and inappropriate. 12. To think that wuda = cen is unreasonable. 13. In no sense can the syllable wulf (1. 16-17) be said to carry the e to the syllable cen. Iii. Leo's translation of the last two lines is not valid, as geador cannot be taken as other than adverbial. Along with his attack on Leo, Trautmann offers his own solution, 'riddle,' in which the wolf is the solver and the woman the riddle. Eadwacer, the whelp of the solver and the riddle, is the solution. Trautmann finds the identical solution, 'riddle,' for the last enigma of the collection, which Dietrich had solved as 'der fahrende SMnger,' and further declares that the finding of a reference to Cynewulf in the repeated lupus of the Latin riddle is merely a matter of opinion\u00E2\u0080\u0094 one which he does not share. Therefore, he reasons, i f neither the first riddle nor the last refers to Cynewulf, and i f the Latin riddle shows no allusion to him, then there is no evidence at a l l for assigning - 9 -authorship to him. 15 Nuck attacked Trautmann's new solution i n 1888, claiming that i t was even more d i f f i c u l t than Leo's. Hicketier, 1^ i n the same publication, presents an elaborate defence of Leo, and claims that the hwelp i s Cynewulf himself, with Eadwacer his father and the female speaker his mother. Hicketier also dismisses Trautmann's comments on Riddle 90 and his solution of Riddle 95, and favors a retention of Dietrich's opinions. 17 Henry Morley, i n his English Writers, agrees with Trautmann i n dismissing Leo's solution. \"Noiman bdrn of woman,\" he exclaims, \"would, with the natural wit to which riddles appeal, go through such a process of interpretation as i s here suggested*\" But Morley also rejects Trautmann's solution not only of the Fi r s t Riddle, but of the last, as well as Dietrich's comments on the Latin riddle. Morley takes a l l three to be religious pieces, with the wolf of both the Fi r s t Riddle and the Latin riddle being the devil. In the Latin riddle, he explains: The marvel of the Lamb that overcame the wolf and tore i t s bowels out i s the Lamb of God who overcame the devil and destroyed his power. The great glory then seen was of \"the lamb that had been slain,\" the Divine appointment,of the agony of one of the three Persons of the Trinity. The four feet were the four gospels; and the seven eyes refer to the Book of Revelation, where the seven eyes of the Lamb are the seven s p i r i t s of God sent forth into a l l the earth. The Fi r s t Riddle he solves as 'the Christian preacher': He i s welcome to the people as one who brings g i f t s , the promises of the Gospel^ and i s received when he comes amongst them. He i s on one island, of the spiritual l i f e ; upon the other island, of the fleshly l i f e , i s the wolf, the devil. The island i s surrounded by the swamps of sin, and men i n i t are fierce and cruel. But they w i l l receive the preacher of Christ when he comes among them. Their positions are unlike. Upon that island of the flesh there i s - 10 -suffering from the passionate desires that are of the wolf; when tears come with the sense of distress, there are worldly pleasures from embrace of the strong destroyer; but pain comes with them. There is yearning of the flesh towards the devil, grief from the expectation that the does*not satisfy, not for want of the true bread of l i f e . There is the preaching. Hearest thou it? Eadwaccer\u00E2\u0080\u0094 the word means custos bonorum, watcher over our wealth\u00E2\u0080\u0094the child of us both, of Christian teacher and of the flesh. He carries the wolf to the wood:1\" he brings the power of the devil over us to the rood tree, the wood of the Cross. Men who have never been joined in Christian brotherhood, and who are easily parted from each other, our music brings together. \"So vanishes the name of Cynewulf out of the riddle, into which i t had been too ingeniously read,\" Morely declares. The last enigma Morley solves as 'the Word of God.' \"We have, then,\" he concludes, \"no evidence upon which to ground a belief that Cynewulf wrote any of the First-English riddles.\" 19 Henry Bradley, in a review of Morley, agrees with his rejection of Leo. As for Trautmann's 'riddle' solution, Bradley says, \"Let the reader compare this ingenious interpretation with any translation of the \"riddle\" that attempts to be fairly l i t e r a l , and refrain from smiling i f he can.\" Nevertheless, he could not accept Morley's solution. Bradley's own feelings on the matter supplied the beginnings of what is now the generally held view about the ' First Riddle.' It was his opinion that \"the so-called riddle is not a riddle at a l l , but a fragment of a dramatic soliloquy, like 'Deor' and 'The Banished Wife's Complaint,' to the latter of which i t bears, both in motive and in treatment, a strong resemblance.\" He added that the true form of the piece i s s o far from being really riddle-like that the idea of its being a riddle probably \"would never have occurred to anyone but for the accident that the fragment appears at the head of the riddles of the Exeter Book.\" - 11 -20 Sievers, i n 1891, delivered the f i n a l blow to Leo's theory with his presentation of philological arguments against the charade. F i r s t of a l l , he says, the preservation of the suffix vowel e i n the f i r s t part of the compound, Cyne (or Cyni, i n the older form), proves, according to the laws of syncopation, that the stem vowel must be short, and that i t i s related to cynn, not, as Leo would have i t , with cen. Secondly, Leo's supposed form Coenewulf i s impossible, as the length of the oe demands the syncopation of the old i \u00E2\u0080\u0094 i . e . the only possible form would be Coenwulf, with no mid-vowel. Sievers l i s t s a number of examples of such forms. Thirdly, he argues, cyne-coene-cen-cwoen could not, under the laws of the Old Northumbrian dialect, be interchanged due to similar accent, because Old Northumbrian made the distinction between the sounds oe and e particularly sharp. Further philological analysis led Sievers to conclude that the riddles predate Cynewulf, and so could not have been written by him. Despite the doubts of Trautmann, Bradley and Sievers and their supporters that any of the riddles refer to Cynewulf, the idea of Cynewulfian authorship continued to be held, for a few years at least, by many scholars. Indeed, the whole sphere of ,01d English academic inquiry i n the late nineteenth century seemed almost preoccupied with the Cynewulfian question, with the result that a few scholars exceeded the bounds of reasonability i n assigning authorship to Cynewulf. From 21 the original three 'signed' poems the supposed Cynewulfian canon had swelled to include not only the riddles, but nearly every piece of Anglo-Saxon poetry evensremotely acceptable by date to be his. In 1886 22 Sarrazin even went so far as to credit \"Beowulf\" to Cynewulf, basing - 12 -his opinion on s t y l i s t i c comparison. This claim brought the t o t a l of Cynewulf's supposed accomplishments to nearly three-quarters of a l l the 23 Old English verse extant! Kail, however, showed the absurdity of over-reliance on Sarrazin's s t y l i s t i c method by using i t successfully to show Cynewulfian authorship of poems which chronologically could not possibly be his. The fervor for wide attributions of authorship to Cynewulf subse-quently subsided, and by the end of the century there was general scepticism toward Cynewulfian authorship of most unsigned poems. Following a study i n 1900 by Madert,^ which found the riddles to have l i t t l e i n common with the style and word usage of Cynewulf, the idea of Cynewulfian authorship of the Exeter riddles was rejected by the majority of scholars. The search for evidence of Cynewulf i n the riddles was rekindled 25 i n 1903, however, when Edmund Erlemann claimed to have found another charade on Cynewulfs name, this time i n the Latin Riddle. Mirum uidetur mihi, lupus ab agno tenetur; obcubuit agnus rupi et capit uiscera l u p i . Dum starem et misarem, uidi gloriam magnam, dui lupi stantes et tertium tribulantes; 2g i i i i pedes habebant, cum septem oculis uidebant. It seems to me marvellous; a wolf i s held by a lamb; the lamb has l a i n down on a rock and seizes the entrails of the wolf. While I stood and wondered, I saw an equally glorious thing, two wolves were standing and tormenting a third. They had four feet and saw with seven eyes. ' Erlemann's charade i s combined with a sort of anagram, which he finds easier to explain by numbering the letters of the name as follows: C Y N E W U L F 12 3 4 5 6 7 8 - 13 -The lupus of the f i r s t l i n e , he says, i s WULF, letters 5-8, and ab agnus i s EWU, U-6. Thus line lb reads: \"WULF i s held by EWU.\" The two overlap, W and U being covered twice. The uiscera lupi of line 2b he takes to be the duo lupi (Erlemann emends from dui) of lia, l u p i being taken as genitive and duo as neuter. These i n Erlemann's view represent L and F, the two remaining letters of WULF which are not siezed ( i . e . overlapped) by EWU. Erlemann makes no mention of the f i r s t syllable, CYN, but presumably f e l t i t s charade l i e s i n the obscure second portion of the riddle. He points out that the letter groupings of this charade are similar to those i n Cynewulf's runic signature i n 'Juliana\" 703-711, where they appear as CYN, EWU and LF. Indeed, the duo lupi, standing for L and F, he says, may be related to the LF of the \"Juliana\" passage, which has s t i l l not been adequately explained. Because of Sievers' work on the dates of composition of various riddles, Erlemann agrees that the old theory which held Cynewulf as the single author of a l l the riddles must be abandoned, but suggests that the authorship of each should be judged independently. He admits that Riddle 90 has a unique place i n the collection because i t i s i n Latin, and that the discovery of Cynewulfian authorship of this one riddle need not prove anything for the rest, but prefers to favor the pos s i b i l i t y of Cynewulfian authorship of some other riddles, especially those following Latin models. He proposes that Cynewulf i n his youth (c.JkO) was the author of this riddle and of some others, and that he was the collector of the rest: \"einen jungen, ubermQtigen Scholaren, der mancherlei Wissen, wenig Wurde und noch v i e l derbe Sinnlischkeit -14 -hat.\" 2 8 Joseph Gotzen, in a note appended to Erlemann's article, alters the argument somewhat. The duo lupi, he says, mean WU, not LF. The terbium is L, the third letter of WULF. The i i i i pedes signify CINE in a figurative sense, supplying the missing portion of the name. The septem oculi are CYNEWUL, the fi r s t seven letters of the name. The eighth, F, Gotzen explains, has already been established by the mention of lupus (= WULF) in the first line, and the number seven is brought in to create further subtlety. Fritz Erlemann, twoLyears later, further refined his namesake's theory. He followed Gotzen in seeing W and U as the duo lupi stantes. E, the first letter of EWU, he says, is left over, and is the tertium which is driven out by the duo lupi. Thus, i f the E is driven out (i.e. discarded), the form CYNWULF remains. This form contains seven letters, which are the septem oculis of line Ub. The i i i i pedes mean WULF, the last four letters of the name. Tupper, in the notes to Riddle 90 in his edition of the riddles, says that \"while the Erlemann solution doesisnot compel acceptance, i t 30 surely invites close attention.\" In the introduction to his edition, though, he is less open to the possibility of a connection between the riddles and Cynewulf: \"In the absence of one jot of evidence connecting the Riddles with this poet,...the heavy burden of proof rests upon 31 him who seeks to revive the moribund claim of Cynewulfian authorship.\" By November of the same year, however, Tupper had changed his mind. It was he himself who sought to revive the claim. Tupper developed a new interpretation of the 'First Riddle' which again found i t to be a -15 -riddle, incorporating a new charade on the name Cynewulf, and a runic signature as well. \"The guise of the lyrical monologue i t certainly has,\" he says, \"but i t seems also to bear the stamp of Cynewulf's cypher....The poem, whether by coincidence or no (and the chances are enormous against a merely accidental concourse of so many elements), may easily be read as a cryptogram..., combining acrostic and charade. 32 Both were popular at this time.\" Acrostics were used by Aldhelm, Tatwine and Boniface, he points out, and a charade was used in AEthel-wald's priscus cassis for Aldhelm. Tupper also cites one tenth-century example in which both acrostic and charade are used at once; folio 78b of Bodleian MS C697. The key to the solution lies in the second line, Tupper claims. He translates: \"They (leod or Cyn) will oppress him (Wulf), i f he comes to want or need (fareat or Nyd)\u00E2\u0080\u0094that i s , Cyn will oppress or press upon Wulf i f the syllable comes to N, the last letter of Cyn.\"^ The line is so important, Tupper argues, that the poet repeats i t a few lines later. According to Tupper hine and he refer to Wulf, who has probably been referred to in a line now missing. As in the previous solutions of this type, a l l of Tupper's found-words and rune names must be inferred by deduction through synonyms. - 16 -Lcod = Cyr. \u00E2\u0080\u00A2 Leodum is mlnurn swylce him mon lac gife : hy=Cyn hint, he =z Wulf villaS hy hint iipecgan, gi( he on J\u00C2\u00BBrea\u00C2\u00A3 cymeiS. Ungelic is us. \"Wulf is on Tege, io on 5>erre ; 5 fiest is )*ct ejlond fenne biworpen, sindon tccclrcowe weras J5ir on Ige: willa'3 hjj hint ij>ecgan, gif he on //real cyme's. Un gel ice is us. \u00E2\u0080\u00A2\Vulfcs ic mines widlastum venum hogode; 10 fonne liit was rcnig ueder ond ic reotugu stet, ponne mcc 6C beaducafa bogum bilegdc : woes me wyn to pon, was me liwa'^rc Oac liiS. [Mln] wulf, ruin wulf, wena me pine seoce gedydou, J>Inc seldcymas, 15 murnende mod, nsllcs metellste. Goliyrest )>u, Eadwacer? Uncerne eame htcelp bire\"S wulf to wuda. Wulf i c - C y n ( ? ) Cyn Wulf ren \u00E2\u0080\u0094 Lagii (?) beaducafa = Ccne (C) h-xelp = C y n ( ? ) lac = Feoh {F) /treat = '2iyd ( J V ) eg, i. c ea \u00E2\u0080\u0094 Lagu (L) ualreowe = Ctne (C) ( ?) Nyd (iV) 6oe loslltefi Jwette niefre gesomnad uncer ciedd gcador. ^ * For the acrostic: lac = Feoh = F = N preat = Nyd . . . . . eglond may originally have read ealond = ea = Lagu = L waelreowe, beaducafa = Cyn . . . . . = C bogum is to suggest boga = Yr . . . . = Y wyn, wen . . . . . . . . . = W uncer = Ur . . . . . . . . ^_ U FNLCYWU\" = CYNWULF Given this new evidence, Tupper proclaimed the burden of proof had now shifted to those who claim the poems are not, with a few exceptions, Cynewulf's: \"The undoubted variations in meter, language and style from the usage in the generally accepted poems of Cynewulf are after all'too slight to avail against the explicit evidence of the 35\" First Riddle and the substantiating testimony of Riddle 90.\" 16 Trautmann, in \"Das Sogenannte Erste RStsel\" in 1912, rebutted Tupper*s argument. If the syllable wulf often seems so obvious, he - 1 7 -argues, why is the other, cyn, only vaguely hinted at? The extraction of cyn out of leodum (1), hjr (2, 7), ic (1+) and hwelp (16) seems unlikely, he says, and the relation of waelreowe and beaducafa (mascu-line) to cyn (neuter), and ultimately to cen, is dubious. Equally unlikely Trautmann finds the uncerne = ur association. And Tapper's explanation for yr, bogum = boga = [ft, Trautmann points out with a smile, starts with the same word Leo used to produce cen. It would be just as easy, he exclaims, to produce the runes T H K C - M ^ (TUPPER) from the poem as h f h } frh T Y (CYNWULF). \"Sicher aber 1st, todsicher, dass das kleine stuck night ein ratsel ist das den namen Cynewulf 37 aufgibt.\" But, he admits, nothing conclusive has really been brought to light to prove i t anything else. Scholars since Tupper have generally treated the question of Cyne-wulfian authorship of the riddles with great caution. Wulf and Eadwacer has been since solved as a riddle, but not with reference to Cynewulf. Most modern scholars prefer to treat the piece as a sort of enigmatic 39 lament. As for the other two supposedly Cynewulfian pieces, Riddles 90 and 95, there is l i t t l e agreement as to solution, and few scholars would risk supporting earlier beliefs in their Cynewulfian traits. Cynewulfian authorship of some individual riddles is s t i l l considered a possibility, but belief in the unity of the collection under one pen is now unanimously rejected. - 18 -FOOTNOTES TO CHAPTER I 1. Heinrich Leo, \"(juae de se ipso Cynevulfus (sive Cinevulfus sive Coenevulfus) poeta Anglosaxonicus tradiderit\" (Halle: Programm, 1857). 2. Jacob Grimm, ed., Andreas und Elene (Kassel, 18J+0). 3. John M. Kemble, \"On Anglo-Saxon Runes,\" Archaelogia 28(l8i;0), pp. 360-363. i | . Leo, loc. c i t . 5. G.P. Krapp and E.vK. Dobbie, eds., The Exeter Book (New York: Columbia University Press, 1936), pp. 179-180. 6. leodum, 'people,' emended to leofrum, 'limbs.' 7. Cited i n Albert S. Cook, ed., Christ of Cynewulf (Boston: Ginn and . Co., 1900), p. l i v . 8. Franz Dietrich, \" Die RSthsel des Exeterbuches: Wtirdung, Losung und Herstellung,\" Zeitschrift fur Deutsches Alterthum 11(1859), hh8-h90. 9. Dietrich, \"Die RSthsel des Exeterbuches: Verfasser, Weitere Losung und Herstellung,\" Zeitschrift filr Deutsches Alterthum 12(1865), 232-252. 10. Frederick Tupper, Jr., ed., The Riddles of the Exeter Book (Boston: Ginn and Co., 1910), p. Iv. ' 11. Max Reiger, \"fiber Cynewulf,\"'Zeitschrift fur Deutsche Philologie 1(1869), 215-226. 12. i b i d . , p. 218. 13. Moritz Trautmann, \"Cynewulf und die Ratsel,\" Anglia 6(1883), anzeiger 158-169. l U . Though there were no other riddles i n Old English i n a form similar to Leo's suggestion for the 'First Riddle^' such charades did exist i n contemporary Medieval literature i n other languages. 15. R. Nuck, \"Zu Trautmann's Deutung des Ersten und Neunundachtzigsten Ratsels,\" Anglia 10(1888), 391-39U-16. F. Hicketier, \"Funf Ratsel des Exeterbuches,\" Anglia 10(1888), 564-600. 17. Henry Morley, English Writers (London: Cassel and Co., 1888), vol. II, pp. 219-226. 1 8 . Morley seems to mistranslate here, or at least to follow Leo in twisting the grammar of the original. Uncerne earne hwelp seems clearly accusative, so i t should be the wolf which carries the child to the wood. See also Leo. 1 9 . Henry Bradley, review of Morley's English Writers, vol I I , i n Acadamy 3 3 ( 1 8 8 8 ) , 1 9 7 - 1 9 8 . 20. Eduard Sievers, \"Zu Cynewulf,\" Anglia 1 3 ( 1 8 9 1 ) , 11-21. 2 1 . A fourth signed poem was added when Napier found another runic signature CYNWULF in the \"Fates of the Apostles\" in 1 8 8 8 . 2 2 . Gregor Sarrazin, \"Beowulf und Kynewulf,\" Anglia 9 ( 1 8 8 6 ) , 515-550. 23. J. Kail, \"Ueber die parallelstellen in der angelsachsischen poesie,\" Anglia 1 2 ( 1 8 8 9 ) , 21-UO. 2U. August Madert, Die Sprache der altenglischen Ratsel des Exeterbuches und die Cynewulffrage (Marburg: H. Bauer, 1 9 0 0 ) . ' 25- Edmund Erlemann, \"Zu den altenglischen Ratseln,\" Archiv 1 1 1 ( 1 9 0 3 ) , 4 9 - 6 3 . 26. Krapp and Dobbie, p. 2U0. 2 7 . Mackie's translation, The Exeter Book Part I I : Poems I X - X X X I I , EETS No. 194 (London: Oxford University Press, 1 9 3 1 0 , p. 2 3 1 . 28. E. Erlemann, p. 62. 2 9 . Fritz Erlemann, \"Zum 90. altenglischen Ratsel,\" Archiv 115(1905), 3 9 1 - 3 9 2 . 3 0 . REB, p. 2 3 2 . 3 1 . REB, p. l x i i . 3 2 . Tupper, \"The Cynewulfian Runes of the First Riddle,\" MLN 2 5 ( 1 9 1 0 ) , 2 3 7 . 3 3 . ibid., pp. 2 3 8 - 2 3 9 . 34. ibid., p. 2 3 8 . 35. ibid., p. 241. 3 6 . Anglia 3 6 ( 1 9 1 2 ) , 1 3 3 - 1 3 8 . 3 7 . ibid., p. 1 3 8 . 3 8 . H. Patzig, \"Zum ersten Ratsel des Exeterbuches,\" Archiv 1U5(1923), 204-207, solves as 'millstone.' - 20 -Notable exceptions are Norman E. Eliason, \"On Wulf and Eadwacer\" in Old English Studies in Honour of John C. Pope, ed. R.B. Burlin and E.B. Irving (Toronto: University of Toronto Press, 1974), 225-234, who thinks i t is a mock lament by the author of a mis-transcribed poem, and Donald K. Fry, \"Wulf and Eadwacer: A Wen Charm,\" Chaucer Review 5(1970), 247-263, who thinks i t is a wen charm. - 21 -CHAPTER II IS THERE UNITY IN THE COLLECTION? The question of unity or disunity in the Exeter riddles can be divided into several parts. Most basic, of course, is the question of whether the collection is a unified whole or some type of miscellany. If i t is a whole, is i t a l l by one author (and i f so who was he) or is i t unified in some other way, such as by the organizing hand of an anthologist? If on the other hand i t is not a unit, is i t simply a jumble of enigmas put together more or less at random, or is there some detectable trace of organization in smaller segments of the collection? In this chapter I wish to examine arguments relating to unity from the approaches of treatment of material, use of sources and dating. Treatment of Material Nearly a l l of the advocates for unity in the riddles believed that i f any one poet was the author of a l l of them, that poet was Cynewulf. Believing that they saw the name Cynewulf in the \"First Riddle\" and perhaps in one or two more (as discussed in Chapter I), and seeing a handful of other riddles with styles somewhat reminiscent of Cynewulf, they concluded by a line of fallacious reasoning that a l l of the riddles must be by him, and that therefore the collection must be a unified whole. More cautious scholars were often guilty of less serious forms of the same logic, though they were perhaps a l i t t l e more justified in their claims. The central arguments for unity arise mainly from Dietrich's work, though Dietrich himself did not firmly believe in the total unity of - 22 -the collection. In his first article Dietrich felt that the fir s t section, Riddles \"l\"-5>9, was a unit by one author, and that the last, 61-95, was a miscellany containing single riddles by the author (Cynewulf) of the f i r s t . Evidence for unity in the first section was to be found, he believed, in the interconnection between the subjects, in the pervasive employment of Latin sources, and in the similarity of treatment of subjects and use of language and opening and closing formulas. Dietrich categorizes a great many of the riddles under headings relating them to natural phenomena, religion, books and writing, weapons, implements for work and so on, and argues that though the riddles he lis t s are generally separated in the collection, this has been done by the poet only to avoid making solution too easy. It must be noted, however, that a l l of the classifications are based almost.'entirely on solution, and that this type of organization i s possible in the most miscellaneous of groups of things, and can prove nothing in itself about any genuine connection between individual riddles. (See also Ebert's similar categorizations of the enigmas of Symphosius.\"'\") The riddles of the final group, he says, exhibit similarities with many in the f i r s t section. Some of the subjects appear again, often with similar handling. Moreover, he continues, the last riddle and the Latin riddle both refer to Cynewulf. However, especially because of the presence of Riddle 66, which is just a short, poor-quality version of kO, Dietrich feels that the final section must, to some extent at least, be different from the rest. Latin sources are less pervasively and more freely used there than in the earlier section, he says, and the opening and closing formulas are far less - 23 -elaborate. He suggests that the latter part is a miscellany containing individual riddles by Cynewulf, but in his second article proposes that i t might after a l l be entirely by Cynewulf, but taken from a different manuscript\u00E2\u0080\u0094one from Cynewulf\u00E2\u0080\u00A2s maturity. The idea is presented as a possibility only, leaving considerable doubt (Tupper's comments to the contrary) as to whether he placed much weight on the argument himself. The riddles of the minor section, 30b and 60, Dietrich justifiably declines to assign definitely to either the preceding section or to the following one, but says that he is inclined to the opinion that they should go with the f i r s t ~ 3 0 b because i t is a repeat of a riddle in the f i r s t section, and 60 largely because of i t s use of Symphosius. He suggests, though, that these two riddles may have been drawn by the scribe from yet another MS. Herzfeld somewhat altered Dietrich's Cynewulfian theory,3 claiming that a l l the riddles are by a youthful Cynewulf. He points out their fresh interest in a l l manner of objects, their joi de vivre. which seems compatible with the frequent inclusion of salacious wit, and attributes the large number of hapax-legomena in the collection to \"a young poet...fond of choosing rare words which may seem to his audience new and surprising.\"^ However, because of the subsequent arguments of later scholars which cast grave doubts on the question of Cynewulfian authorship of the riddles (see Chapter I), Herzfeld later abandoned these opinions. Of the scholars who doubted the unity theory, Trautmann was the most vocal. Though there are, he said, a few very limited areas in the collection which seem to show the influence of some ordering hand\u00E2\u0080\u0094 - 2k -7, 8, 9, 10 are a l l bird riddles, and hh, U5 and 61, 62, 63 are a l l 'obscene'\u00E2\u0080\u0094that is about a l l that permits any thought of the plan and connection which might be expected i f the riddles were a l l from one pen. When one weighs the differences between individual riddles in style, tone, ability and language, then i t must be concluded that the Exeter riddles are not only no unified and similar collection, but that they must originate with different poets. And this conclusion is not surprising, as in the years 700-750 there were in England a whole line of men who composed Latin riddles, and so i t seems only natural that there should have been also several who wrote in English. Trautmann further remarks that the fair l y wide agreement of words and phrases found in the riddles is as easily explained by assuming the presence of several authors as one, particularly as i t is possible that the poets knew each other personally or through their writings. Trautmann feels that Cynewulf, because of the dates of composition of the riddles, could not have been the possible author of more than hO, and that i t is quite likely he was author of none. If any named poet is to be suggested as possible author of these riddles, he says, Eusebius (or Hwaetbert, by his Old English name) has the best claim, for \"remarkably often these remind one of him and his riddles remind 7 one of the Exeter riddles.\" The most extensive study of the question of unity in the riddles, though, has been Tupper's, which systematically examines the entire collection, sifting out and comparing f i r s t the opening and closing formulas (REB lxv-lxvi) and then specific motives (lxvi-lxxv). This \"cursory survey,\" which goes on for nine tightly-packed pages, seems very impressive, and obviously impressed Tupper: \"Such likenesses as I have pointed out between the various riddles are sufficiently striking to establish homogeneity, and indeed they often compel belief in the g presence of a single hand in many of the problems.\" He concludes: The Riddles, then, are homogeneous in their artistry. One of the finest proofs of this lies in the striking circumstance that almost every dark saying or obscure periphrase in our poems finds illuminating explanation elsewhere in the collection.... Now i f certain art-riddles are found grouped in what is really a single collection; i f , moreover, these riddles, after close analysis, are found to be homogeneous in their diction; i f , too, large collections from single hands were common at that period,\u00E2\u0080\u0094 the burden of proof rests not upon him who argues for unity of authorship, since every precedent and presumption are in his favor, but upon him who champions diversity of origin. Even so, he says, there is a small group of riddles which could not have been written by the same author as the rest: The servilely imitative temper of Aldhelm's translator in the enigmas of the 'Mail-coat.' and 'Creation' (Rid. 36, hi CK-D 3$, 40j ) differs so utterly from the prevailing tone in the collection, which is at its highest in the unchecked range of imitr-ation of the 'Storm' riddles (2-U [K-D 1-3J), that this inferiority cannot be explained with Dietrich by the changing inclination of one poet.-1-0 These reservations are expanded in Tupper's notes to the creation riddles, where he assumes two hands at work in Riddle K-D 40, the first of which is that of the translator of 35, and the second of which is also responsible for 66. But is Tupper's conclusion of unity justified? Careful inspection reveals that many of Tupper's so-called parallels are dubious to say the least. Some are simply parallels in opening and closing formulas. Some arise in a l l probability only out of similarity of subject treated, U 5 as in the two bird riddles 8 and 2h: 8 hleobre ne mibe, 2h guofugles hleobor. Both birds mimic and make noise (the subject of both may be 'jay'). Others no doubt are parallel only because of common use of the formulaic language of riddle tradition. The phrases streamas stapu - 26 -6 8 b e a t a S , 2 , and mec stondende streamas beata5, 81 , seem parallel enough until i t is seen that the f i r s t refers to waves beating on the shore and the second to rain beating on a weathercock. The image of water beating is one which is s t i l l in common use today, so the employ-ment o f this not unusual image in such unrelated contexts seems to me to be indicative of nothing at a l l . Similar metaphoric technique due not to common authorship but to riddle tradition probably underlies supposed parallels such as 10\"'\" Neb waes min on nearwe, which refers t o the barnacle form of the barnacle goose, 21^ \" Neb is min neberweard, which describes a plowshare, 31^ niberweard...waes neb hyre, the lower pipe of a bagpipe, and 34^ nebb bib hyre aet nytte niberweard gongefl, describing a rake. The fi r s t is probably unrelated to the other three, and the similarity between these last no doubt arises from the common technique of animation, employed through the popular riddle metaphor which turns a pointed protruberance into a nose. Most important, the possibility must always be considered that specific parallels may arise only from common formulaic or traditional patterns of wording. Michael Onwuemene, in a 1970 doctoral dissertation for the University o f Kansas, has calculated that as many as h9-S% of the 2,476 half-lines of the riddles are repeated wholely or in part when compared with the whole Old English poetic corpus. Certainly there may be valid parallels among the riddles pointing to common authorship of individual poems. But, as they are mixed in with a huge mass of what must be mainly coincidental parallels, i t seems to me unlikely that any truly substantiable results could be obtained from this type of study. Unity and Sources A number of scholars have attempted to use the handling of source material in the riddles as evidence for showing unity or disunity of parts or the whole of the Exeter riddle collection. Thorpe1'1' was the firs t to point out the indebtedness of some of the riddles to Latin authors. He believed the riddles to be a l l by one poet, and that any traceable sources merely provided inspiration for essentially original composition. Dietrich agreed with Thorpe about the seeming originality of most of the riddles, but was not as ready (in his first article, at least) to attribute the entire collection to one author. Dietrich felt he could distinguish a difference between the two major sections of the collection in the extent of their use of Latin sources. The fi r s t section, he says, is characterized by considerable use of Symphosius and Aldhelm; 12 indeed, 16, U7 and 60 contain word-for-word borrowings from Symphosius, 13 and 35, 38 and hO contain sentence-for-sentence renditions of Aldhelm. In addition Dietrich points our six instances of more limited borrowing inothe fi r s t section. In the second major part of the collection a freer connection is seen between the Anglo-Saxon riddles and the Latin, he claims. Examples of actual translations, two of which are found in the f i r s t section, do not reappear in the last, and there are only a few uses of Symphosius (65, 8U, 85, 86, 91),\"^ and even fewer of Aldhelm (63, 71, 84). Subsequent scholars1'' generally did not attempt to show a difference in use of sources in different parts of the collection, but concentrated mainly on tracing the sources of riddles, on assessing the extent of - 2 8 -borrowing, and on examining the related question of the o r i g i n a l i t y of the Old English riddles. Tupper f e l t that the collection was mainly by one author, and pointed out that the Latin sources are employed throughout the collection. But though he c a l l s Dietrich's evidence \"very doubtful premises,\"^ a 17 quick tabulation of Tapper's own results shows that of the 35 riddles which he feels demonstrate debt to Latin sources, 27 are i n the section 1-59 and only seven are i n the f i n a l group. The remaining riddle i s 60, \u00E2\u0080\u00A2reed pipe,' which i s found i n the miscellaneous material between the two major riddle sections and cannot be assigned d e f i n i t e l y to either. Actually the data seems to bear out Dietrich's opinion more than Tupper's, though a certain amount of caution must be used i n assessing these figures due to dispute about whether specific riddles depend on Latin sources or not (see Appendix B). No attempt should be made to iise this method to compare the sections on the basis of the presence of folk elements, however. 18 Tupper*s and Wyatt's studies vary widely on this matter, largely due to the quite different principles of selection employed by the two scholars. Both believed that a large number of the riddles are folk productions, yet of the 18 riddles i n Tupper's l i s t and the 29 i n Wyatt's (more than 1-| times as many), only five (13, 29, 33, 45, 46) enjoy the agreement of both scholars. Trautmann, on the other hand, 19 f e l t there were v i r t u a l l y no folk elements at a l l i n the riddles. - 29 -Arguments Based on Date To the early scholars, dating of the Exeter riddles went hand in hand with authorship. By general consensus the riddles were written by Cynewulf and therefore belonged to Cynewulf's period, though just what that period was was a matter of debate. As understanding of the philological principles underlying Old English sound changes increased, however, more supportable opinions about the date of the riddles sometimes resulted in arguments relating to the question of unity. 20 Sievers, in 1891, used datable features in four of the riddles in an attack on the theory of Cynewulfian authorship (see also Chapter I). In the Leiden Riddle, which is a Northumbrian version of Exeter Riddle 35, there are a number of examples of unstressed i . Comparing several manuscripts whose date is more or less reliable due to other evidence, Sievers concluded that the unstressed i > e change occurred in the South and Midlands at about 750, with few exceptions, and that, judging from the few datable examples available from Northumbria, the change probably occurred there at about the same time. The few late examples of i forms, he says, are probably just mechanical repetitions of older spellings. The Leiden Riddle, then, and therefore Riddle 35, must predate this change. The cryptic word agof of Riddle 23 was no doubt originally agob, which is simply boga, 'bow,' in reverse. The original terminal soundSievers says, must have been represented earlier by b, but later by f; Riddle 23 must predate the transition, which took place a l i t t l e earlier than the i> e change. The spelling of HANA and HAEN in Riddle k2 with a before the nasal instead of o in the fir s t word and ae instead of a in the second also indicates a date - 30 -at the beginning of the eighth century, he says. Finally, the word formed by the runes in Riddle 19 is H A O F O C . The form ao, Sievers argues, is not a likely form for the umlaut of a, as i t appears nowhere else. He therefore emends to HAFOC, which, with unumlauted a, he says, would again indicate the first half of the eighth century.; For the second part of his argument Sievers points out that the name Cynewulf is spelled two ways in the runic signatures: CYNEWULF and CYNWULF. These, he reasons, must be alternates. Both forms appear several times in Bede, but there the Cynewulf form, with its unstressed medial vowel, is always spelled Cyniwulf. Thus Sievers concludes that the riddles, which date from the period of unstressed i (on the evidence of Riddle 35, at least), must predate Cynewulf, who part of the time spelled his name with an unstressed e. Though Sievers was s t i l l assuming a unified collection as an almost unquestioned assumption, the groundwork for further development was laid. 21 Barnouw, in his linguistic study of Old English poetry in 1902, claims to have confirmed Sievers' results. By computing the frequency of occurrence of the definite article and other grammatical forms in datable Old English verse, Barnouw points out a general tendency toward a heavier use of the article in later poems and a more sparse occurrence in earlier ones. Comparing the Riddles with the accepted Cynewulfian works, Barnouw concludes that because the frequency of the article is greater in the accepted works, the majority of the riddles must predate Cynewulf. Further, comparing the frequency in the riddles individually, he asserts that the riddles are scattered over a wide period of time, with l i t t l e apparent organization in the collection except for one small group, 37, 38, 39, which could a l l be by the same - 31 -poet. Riddles 23 and h2, however, which Sievers had used as examples of early date, Barnouw declares to be late, due to their frequent employment of the article\u00E2\u0080\u0094four in the 16 lines of Riddle 23 (\"the 22 article appears wherever possible\" ) and seven in the 17 lines of Lj.2 23 (\"we have here without doubt a very young riddle\" )\u00E2\u0080\u0094and to the use of articles instead of demonstratives in U2. 2li Tupper rejects both of these claims for dating the riddles. He feels that Sievers' arguments for dating the i> e change were hasty and based on insufficient data, and provides several examples of the use of unstressed i in the ninth and tenth centuries. In addition, Tupper declares, no generalization should be made from Riddle 35, as i t has l i t t l e in common with the other enigmas. He completely rejects Sievers' argument concerning the datability of the b>f change in Riddle 23.. \"Are we to believe that a riddler in the latter part of the eighth or even in the ninth and tenth centuries was prevented by phonetic laws from inverting any word with an i n i t i a l b and thus 26 forming a nonsense-word with an uncouth ending?\" he asks, and presents three examples of terminal b from the ninth century. The presence of a instead of o and of ae instead of a in H A N A and HAEN of Riddle U2 are both \"well established West Saxon forms,\" he says. Similarly, the runic word HAFOC of Riddle 19 (if the unemended form is rejected) 3 67 is also an accepted West Saxon variant, found in Riddle 2h and I4.O , and in other places outside the riddles. But though he rejects Sievers' evidence, Tupper favors his conclusion of a date for the riddles somewhere in the f i r s t half of the eighth century, but cautions that 27 this is \"only a surmise, which is perhaps incapable of proof.\" - 32 -Barnouw's method is also unreliable, Tupper says, as i t depends in part on article use in opening and closing formulas which may not have anything to do with the particular habits of the poet who used them\u00E2\u0080\u0094for example ic pa wiht(e) geseah 37, 38, 68; ic wiht geseah 29, 3ii, 36\u00E2\u0080\u0094and makes no allowance for incidental inclusion of such archaic 12 forms as 12 hygegalan hond, where a weak adjective appears without an article. Tupper notes two tenth-century examples of weak adjective without an article in the \"Battle of Brunanburh.\" Barnouw's arguments, Tupper declares, \"seem to me to carry l i t t l e weight. The normalizing of later scribes, and the tendency to archaize, to use traditional formulas and expressions, so strong in Anglo-Saxon poetry, render this test almost valueless.\"^ 2\u00C2\u00B0 Hacikyan, in 1966, ' attempted to apply Barnouw's techniques to the riddles in order to prove a division between the sections of the collection. Though Barnouw's methods have been considered dubious by the majority of scholars even when applied to fair l y large bodies of literature, Hacikyan embraces them without much reservation. By examining the frequency of the definite article and other grammatical 30 forms in the fi r s t section (1-59) and comparing this to the frequency of the same forms in the last (61-95), he concludes that the two sections probably share no overall unity of authorship (as they are stylistically different in their employment of demonstratives), and that the last section is probably older. As well, he l i s t s more features in the riddles which he feels to be datable: 1. Comparison using the dative instead of a bonne phrase in kO^' e ^ \u00C2\u00B0 * (Madert^\"1\" points out 10 instances i n this riddle) and 84 ' Hacikyan takes to be indicative of an early date. 2. The form salwonge i n 32 (instead of saelwonge) perhaps indicates a date previous to the unstressed a> ae change, but Hacikyan admits that the same feature could as easily be explained by a late date as an early one, and that at any rate the spelling could be merely a variant, as i n Riddle 42 (nana). 3. \"Another similar example i s the occurrence of hyra instead 1 5> of hyre i n Rid. 70 or grenne instead of grene i n Rid. 66 .\" Hacikyan does not make i t clear what these are similar to, or upon which philological changes his claim i s based. 4. The confusion between 5es and 5aes i n 66^ Hacikyan feels results from an inadvertent preservation of an early eighth-century form. 5>. The spelling wambe 62 for wombe Hacikyan takes to indicate an early date (a> o before nasal), but he admits i t may be only due to scribal inconsistency. . 6. The preterite borcade 87^ as a form of beorcan, he says, 32 may indicate a date previous to the o> e change. Tupper and 33 Bosworth and Toller, however, see this as a regular preterite form of an alternate verb, W2, borcian. Hacikyan's s t a t i s t i c a l approach to dating the riddles, apart from i t s reliance on a method which does not have the approval of many scholars, i s suspect by his own admission. \"The absence of the definite a r t i c l e does not always indicate an early date as certain analyses based on late poems, e.g., the Battle of Maldon, or even - 3h -the Battle of Brunanburh, show the opposite....It seems that the frequency of the article, generally in late works, is mainly a stylistic problem, whereas its lack in early poems is a matter of grammatical and syntactical convention.\" The difference in his frequency figures between the first section and the last, then, is as easily explained by a difference in style as by a difference in date. Some of the riddles may indeed be older than those of the first group, but there is nothing conclusive in this study to prove them so, and certainly nothing to prove that they a l l are older. - 35 -FOOTNOTES TO CHAPTER II 1. Adolf Ebert, \"Die RStselpoesie der Angelsachsen,\" Berichte uber die Verhandlungen der koniglich sachsischen Gesellschaft der Wissenschaften zu Leipzig (Philol.-Hist. Classe, April, 1877), XXIX, 20-56. 2. REB, p. l x i i i (cf. p. l i i i ) . 3. Georg Herzfeld, \"Die RStsel des Exeterbuches und ihr Verfasser.\" Acta Germanica, Bd. II, Heft I, Berlin, 1890. l i . Tupper's translation, REB, p. I v i . 5. Archiv 106(1901), 390. 6. \"Zeit, Heimat und Verfasser der Altenglischen Ratsel,\" Anglia 38(19110, 372. 7. i b i d . , p. 373. 8. REB, p. lxxv. 9. REB, p. lxxvi. 10. REB, p. l x x v i i i . 11. Benjamin Thorpe, ed., Codex Exoniensis: A Collection of Anglo-Saxon Poetry (London: William Pickering, 181+2), p. x. 12. For the purpose of this argument Dietrich includes 30b and 60 with the f i r s t section. 13. Tupper (REB p. x l i ) l i s t s Riddle 38 not among those with'.direct debt, but with those which only \"suggest direct borrowing.\" Iii. Tupper finds only points of similarity between Riddle 8U and Aldhelm. 15. See Trautmann, \"Quellen der Altenglischen Ratsel,\" Anglia 38(19110, 3U9-354 for the best summary. 16. REB, p. xxxviii. 17. REB, pp. x l - x l i v . 18. Tupper REB, pp. l i - l i i i j A.J. Wyatt, ed., Old English Riddles (Boston: D.C. Heath, 1912), pp. xxx-xxxi. 19. \"Quellen der Altenglischen Ratsel,\" p. 353. - 36 -20. Anglia 13, pp. 1-25. 21. A.J. Barnouw, Textkritische ITntersuchungen nach dem Gebrauch des Bestimmten Artikels und des Schwachen Adjectivs i n der Altenglischen Poesie (Leiden: E.S. B r i l l . 1902). pp. 211-223. 22. Barnouw, p. 2lU. 23. Barnouw, p. 215-24. REB, pp. l v i - l v i i i , l x x v i i - l x x v i i i . 25. Kenneth Sisam, Studies i n the History of Old English Literature (Oxford: Clarendon, 1953), pp. 4-6, points out that the so-called Northumbrian genealogies, upon which Sievers had largely based this opinion, are not Northumbrian at a l l , but Mercian. \"It i s not a witness for e i n Northumbrian usage; and there i s nothing else to discount the testimonly of the Liber Vitae that the Northumbrian spelling i n the early years of the ninth century i s Cyni-\" (p. 6.). 26. REB, p. l v i i . 27. REB, p. l v i i i ; 28. REB, p. l x x v i i i . 29. Agop Hacikyan, A Linguistic and Literary Analysis of Old English Riddles (Montreal: Mario Casalini, 1966), pp. 16-25, 45-62. 30. \"The frequency of total demonstratives per line i s .15 i n section one and ;039 (.0U7 [not counting mutilated lines]) i n section three, and the frequency of the definite article i n the f i r s t section os .12 and .032 (.042) i n the third\" (p. 19). 31. Madert, pp. 69, 128. 32. REB, p. 246. 33. p. 116. 34. Hacikyan, p. 19. - 37 -CHAPTER III SECTIONS OF THE TEXT AS SEPARATE UNITS Despite the arguments of Tupper and the others for the complete, or almost complete, unity of the Exeter riddle collection, i t seems to me that there i s such a diversity of subjects, of approaches to themes and of quality of workmanship i n these riddles that no argument for overall unity can be made convincing. C r i t i c a l opinion since Tupper has indeed tended more and more toward the idea of wide diversity of authorship and an abandonment of any notion of extensive organization i n the collection. I believe, however, that though this diversity i s certainly present, there i s s t i l l some pattern of organization detectable i n the riddles. In this f i n a l chapter I wish to explore the p o s s i b i l i t y that individual sections of the collection may be separate units. Evidence for Distinct Groups i n the Collection Let us examine the point of view of the riddles. Three different types appear: A. The object speaks for i t s e l f i n the f i r s t person. B. The riddler speaks i n the f i r s t person, talking about the object. C. Neither the object nor the riddler speaks, and the whole matter i s presented i n the third person. The riddles themselves may be classified accordingly: Type A Type B Type C Fragments 1 2 3 a 5 6 7 8 9 10 11 12 lit 15 16 17 18 20 21 23 2k 25 26 27 30 35 ho (30bT 60 61 62 63 65 66 71 72 73 71+ 77 78 79 80 81 83 85 88 91 92 93 95 13 19 _29 31 32 3k 36 37 38 1*2 1*3 U5 1+7 1+8 1*9 51 52 53 55 56 58 61+ 67 68 75 76 87 [90 (Latin)] 22 28 33 39 1+1 hk kS 50 51+ 57 69 70 (82) 81+ 86 (89) (91+) - 39 -The l i s t 1 may be readily organized into three parts (omitting for now 30b and 60). In the f i r s t part, from the beginning to about Riddle 30, nearly a l l the riddles are Type A. In the second part, following the f i r s t group up to Riddle 59, most are Type B, with several C, and almost no A. The third part shows again a preponderance of Type A, but with a few more B and C than i n the f i r s t group. The most striking thing about the l i s t i s the abrupt break i n pattern after Riddle 59, corresponding with the end of the f i r s t section i n the MS. This would seem to lend considerable support to the arguments of Dietrich and Hacikyan (see Chapter II, who have claimed that more than just a scribal separation i n the text occurs here. Further, the division i n the section 1-59 seems to cut i t almost exactly i n half, breaking at about Riddle 30. The result i s a f a i r l y symmetrical three-part grouping. What could account for this pattern? Evidence for a division between Riddles 61-95 and the earlier ones comes as no real surprise, because i f there i s to be a break i n the collection at a l l , the arrangement i n the MS would lead one to expect i t here. In l i g h t of this new evidence, however, i t would appear that not only are the sections 1-59 and 61-95 distinct groups, but the f i r s t group i s i t s e l f r e a lly the sum of two parts. Many things can become distorted i n transmission, but something as basic as the point of view i n the presentation i s certainly one of the least l i k e l y to be changed, regardless of the embellishments, revisions or miscopyings of generations of scribes. If i t i s assumed that writers, or even anthologists, d i f f e r from each other i n their subconscious preferences between points - Uo -of view, and that even miscellanies are quite l i k e l y as a result to show characteristic and discernable differences i n the proportion of the different, points of view used, then one collection which more or less consistently prefers the first-person point of view would be distinguishable from, say, a second collection predominantly i n the third person, even following garbled transmissions and wholesale revisions, providing the revisers did not change the point of view of a large number of pieces. - hi -The Original Groupings Partly because of the precedent of the 60-riddle collection of Eusebius and partly because of desire for a neatly-fitting number, i t i s tempting to speculate that the f i r s t section of the riddles originally comprised an even 60 pieces, and that this was i t s e l f made up of two smaller collections of 30 riddles each. If the makeup of this supposed original form could be reconstructed, then i t should be possible to locate, speculatively, the position of the break between the two smaller groups. There appear to be 59 separate riddles i n the f i r s t section (thus Krapp and Dobbie), but because of textual problems i n the MS there i s some debate about the divisions between the f i r s t three, and thus about 2 3 the t o t a l . Trautmann and Erlemann take Riddles 1-3 as a single piece; li 5 Baum and Erika von Erhardt-Siebold take 1 as a separate riddle, and 2-3 as one together. The scribe of this MS at least seems to have f e l t Riddle 1 to be a separate piece. An end-mark follows i t , and 2 begins with a capital l e t t e r . At the end of 2, however, there i s no end-mark, and no capital to begin Riddle 3. The break occurs, incidentally, at the end of a page\u00E2\u0080\u0094Riddle 2 ends the bottom of the recto of f o l i o 101, and Riddle 3 begins the verso. The inadvertent omission of a clear break at such a point seems highly unlikely (cf. Riddles U2-U3, where the break i s omitted, but i n mid-page). To produce a total of 60, then, the number of riddles i n the f i r s t section must be increased by at least one ( i f Krapp and Dobbie are followed), but more l i k e l y two or even three (following Trautmann and - U2 -Erlemann). Where are these riddles to come from? There i s only one obvious hiatus i n the MS i n the section 1-59. Riddle i+0 breaks off i n mid-sentence at the end of f o l i o 111 , and Riddle Ul, which begins the new f o l i o , has no opening. Clearly at least one f o l i o has dropped from the text at this point. Now the average number of verse lines per f o l i o side i n the MS i n this section of the riddles i s 33 1/3. The total number of verse lines missing i f one f o l i o (two sides) has dropped out, then, would be about 67. Fortunately, the break occurs i n one of the only two riddles of the entire collection which are direct translations from a known source. It should be possible, then, to calculate the approximate number of lines of verse which have been lost from Riddle UO, assuming a complete translation was made. Tupper^ sees the piece as having been the product, i n i t s present state, of two translators. The f i r s t , A, who was probably also the translator of the other Aldhelm riddle, 35, i n nearly every case carefully expanded each line of Latin into two of Anglo-Saxon. The result i s 79 Old English lines from U3 of Latin. Following line 79, though, Tupper finds the method, and i n his opinion the translator, to change. The translation i n the later part of the riddle does not attempt to follow this two-for-one pattern, and assessment i s not at a l l helped by the fact that the B translator evidently had as his model a text with a different ordering of ideas than that i n the extant version. B renders the six lines of Aldhelm 61-66 as 16 of Anglo-Saxon ( l i O 8 2 - 9 7 ) , and the seven of 'De Creatura* iili-50 as 10 ( U O 9 8 - 1 0 7 ) . The t o t a l , however, i s 26 of the vernacular for 13 of the Latin, s t i l l - U3 -about two for one. I f lines $1-60 and 67-83 of Aldhelm s t i l l remain to be dealt with, a total of 27 lines, then about 5U might be expected 7 to be missing from the Exeter text. Comparing this to the estimate of 67 lines missing by the page calculation, I conclude that the 13 or so excess lines are a l i k e l y number to be missing from the beginning of Riddle Ijl, which, i n i t s rather leisurely pace, has not r e a l l y supplied much information by the end of i t s nine surviving lines. We have accounted for the missing f o l i o , then, but s t i l l have not accounted for any more riddles. The only other place where material could have dropped out of the present MS i s between folios 105 and 106, where Riddle 20 ends the verso of 105, and a new piece, 21, begins with a capital letter at the top of the new f o l i o . Though Riddle 20 could quite possibly be complete\u00E2\u0080\u0094\"There i s nothing i n the MS to indicate any lacuna here,\" Wyatt declares \u00E2\u0080\u0094there i s no end-mark at the end v 9 of 105 \u00E2\u0080\u00A2 Most scholars f e e l that there i s something missing. Forster, i n his collation of the MS i n The Exeter Book of Old English Poetry, 1^ f e l t that a leaf had dropped out at this point. As the missing page would comprise the other two faces of the f o l i o missing between 111 and 112 (Riddle kO), the p o s s i b l i l i t y i s an attractive one. The gathering containing folios 106-111, gathering XIV i n Fflrster's numbering, i s composed of only six f o l i o s , two l e s s \u00E2\u0080\u0094 o r one folded sheet less\u00E2\u0080\u0094than the usual eight i n most of the other gatherings (see Figure 1). This missing f o l i o , then, would provide a further 67 or so verse lines. Judging from the general length of riddles i n the collection, i t i s unlikely that a single riddle would be this long. - hh -Rid. 20 Rid. U0 I , I X \u00E2\u0080\u00A2 106 \u00E2\u0080\u00A2 107 \u00E2\u0080\u00A2 108 109 \u00E2\u0080\u00A2 110 \u00E2\u0080\u00A2 111 \u00E2\u0080\u00A2 X I i \u00E2\u0080\u0094 ' i i Gathering XIV Resignation' Rid. 73 Gathering XVI (F = fold) Rid. 71^ -71* Rid. 75 Rid. 95 ' 1 , 1 126 + 127 * 128 129 * 130 l . ,' 1 ' t Gathering XVII Figure 1 FSrster's Collation of Gatherings XIV, XVI and XVII (EBOEP, p. 59) a) 126 + X + 127 \u00E2\u0080\u00A2 128 I 129 * 130 b) 126 \u00E2\u0080\u00A2 127 ' 128 I 129 \u00E2\u0080\u00A2 130 * X : 1 \u00E2\u0080\u00A2 ' \u00E2\u0080\u00A2 1 : Figure 2 Possible Loss from Gathering XVII Apart from the two long pieces i n the f i r s t section\u00E2\u0080\u0094UO (107 lines) and 3 (7k lines) or 2-3 (89 l i n e s ) \u00E2\u0080\u0094 t h e longest are Riddles 15, 26 and 39, each with 29 l i n e s . (In the second section only one riddle approaches this length, No. 8U, which has 56 lines,) As two riddles of about 33 lines each the pieces would s t i l l be longish, but more reasonable i n length. But i f the 67 lines were divided into three poems averaging 22 lines or so each, though s t i l l greater than average i n length for the section,^\" the numbers would come quite within the range of p o s s i b i l i t y . There are half a dozen riddles of about this length or a l i t t l e more i n the f i r s t section. Thus, i f the 'storm* riddles are counted as two and not three pieces (1, 2-3), or especially i f they are considered a l l as a single piece, i t i s quite possible that there might have been o r i g i n a l l y an even 60 riddles i n this section of the MS. Scribal alterations by accident or design i n earlier manuscriptal versions may, of course, have had a hand i n changing the number of pieces, but there i s no way to t e l l . Thus i f the division between the f i r s t group (henceforth to be called Group I) and the second (Group II) i s to be made at the mid-point of the original number, dividing the f i r s t section into an 12 o r i g i n a l l y complete pair of 30-riddle groups, i t should be placed between Riddles 29 and 30. If the f i r s t section could orig i n a l l y have comprised 60 riddles, l i k e Eusebius's collection, could the compiler(s) of the l a s t part, 61-95, have been trying, l i k e Tatwine, to add another UO to bring the t o t a l to 100? FSrster f e l t that there i s nothing to indicate that - h6 -gatherings XVI and XVII, which contain these riddles, ever contained more folios than they now have, i.e. seven and five respectively. There i s , however, one place i n gathering XVII where internal loss could possibly have occurred on the evidence of the text. This i s between folios 126 (ends with Riddle 74) and 127 (begins with 75), though the p o s s i b i l i t y here seems slight. It would after a l l be rather unlikely that the missing riddles would begin exactly at the beginning of the f o l i o and end exactly at the end of i t . Forster reconstructs gathering XVII (the backs of the sheets are now worn away, so the gathering i n i t s 1933 condition was composed only of five single leaves glued together) as having originally comprised two folded sheets, f o l s . 127-130 and 128-129, and a single leaf, f o l . 126. This possible break does come, then, at a place where the loss of a single leaf i s a possibility, as two single s h e e t s \u00E2\u0080\u0094 f o l . 126 and the missing one (see Figure 2a)\u00E2\u0080\u0094would make the beginning of the gathering similar to the beginning of gathering XVI. If there i s any loss, though, i t i s much more reasonable to suppose that i t might have occurred at the end of the collection. Riddle 95 ends at the bottom of the last f o l i o , with a l l the usual number of f o l i o lines completed (22 i n this part of the MS, so there i s nothing to indicate conclusively, despite Forster's contrary opinion, that this was actually the original end of the book. Some leaves, or even a gathering or more, might have been lost from 13 the end, as certainly was the case at the beginning, and Fol. 126 may not i n fact have been a single leaf, but only the f i r s t half of a folded sheet, with a missing leaf possibly providing the second half - 1+7 -(Figure 2b). Judging from the harsh treatment the f i r s t few pages of the Exeter Book have received over the years, the likelihood seems actually rather small that a l l the pages should have been retained at the end. The po s s i b i l i t y does exist, then, though certainly i t i s much more hypothetical than for the f i r s t section, that some material has been lost, and that the missing parts could have brought the f i n a l group to an even 1+0. But what of the minor group, 30b and 60? Do these two riddles belong to the f i r s t section, to the last, or are they independent of both groups? As 30b i s merely another copy of 30, with only minor variations, i t has undoubtedly been taken from the same original as the earlier one. Because of the repetition i t does not seem l i k e l y that 30b i s merely a misplaced member of the f i r s t section. However, i f Riddle 30 was originally the f i r s t riddle of Group II, then 30b could have resulted from an accidental mis-start of this group later i n the MS. Such an explanation does not solve the mystery of the presence of Riddle 60, though. The best guess seems to be that ( i f i t i s a riddle at a l l , and not just the beginning of the \"Husband;'s Message\"^) i t i s independent of the grouped riddles, and that i t found i t s way into the MS at this point only because i t was among miscellaneous short material, and somewhat resembled the \"Husband's Message.\" - 48 -Opening and Closing Formulas, and Three Adverbs Some support for this pattern of riddle groupings can be found by examining the opening and closing formulas. Dietrich, i n his f i r s t a r t i c l e , attempted to use the similarity and frequency of opening and closing formulas i n the section \"l\"-i>9 as evidence for unity of author-ship. In contrast, he said, the f i n a l section has few of these formulas, with much less variety than i n the f i r s t section, and what formulas do appear are more clumsily handled. Dietrich's study was, however, not re a l l y comprehensive enough to be considered valid, and though i t did point out a difference between the last section and the f i r s t i n the frequency and i n the type of formulas used, i t was rightly rejected by Tupper. A more systematic examination, though, not only substantiates Dietrich's claims, but provides positive evidence for the three-group theory. Firs t l e t us examine the formulas themselves. A complete l i s t i s presented i n Appendix C. Certainly the greatest impediment to! making any sort of grouping of the riddles on the basis of these expressions i s their diversity. Some, though they are without doubt formulaic, appear only a few times i n the collection. Others\u00E2\u0080\u0094for example raed hwaet i c maene i n 61 and saga hwaet i c hatte i n 62\u00E2\u0080\u0094appear repeated i n variations dissimilar enough to make i t doubtful whether they are 12-15-the same formulas or different ones. Yet others, such as 2 , saga boncol mon hwa mec bregde of brimes faebmum bonne streamas eft s t i l l e weorbaS yba gebwaere be mec aer wrugon though clearly drawing their substance from traditional formulas, - h9 -are so altered for originality by the poet that they can no longer be rightly called formulaic. And, for any individual case of common formula use, the rebuttal that the similarity i s due merely to common tradition i s irrefutable. Yet, because the three groups of riddles are roughly equivalent i n size\u00E2\u0080\u00942 7 (considering Riddles 1-3 as one) riddles extant i n Group I, 30 i n Group II and 35? i n the third group (Group III)\u00E2\u0080\u0094some evidence may be gained from an examination of the distribution of the opening and closing formulas over the collection. Much of the data i n Tables II and III appears to support the claim for three major divisions i n the text. The three most fundamental tests, A, B and C, should be examined f i r s t . A shows a clear division between the three groups: the frequency of opening formulas i s more than twice as great i n Group II as i n either of the other groups. Test B, the distribution of closing formulas, i s not as conclusive, but does seem to indicate a heavier incidence i n Group I than i n the others, with the frequency i n II and III being more or less the same. It i s possibly significant to note here with Tupper the long succession of riddles ( a l l i n Group I, incidentally) with no opening formula, 1-17, though I think the i c seah of 13 i s a formula. Another 18 successive riddles, h.2-59 ( a l l i n Group II), have no conventionalized c l o s i n g . ^ The figures from Test C, absence of both opening and closing formulas, again show a difference between the three groups, with the frequency i n Group II only about half as great as 'in>Gfoup III. Individual formulas are not distributed evenly over the collection, but usually appear concentrated i n a single group. Of the 12 instances of D, i c seah, eight range across Group II, whereas only two each are Group-I Test 1 j 2 1 3 lli !5 6 7 i 8 19 10111 i 12113 H i , l y l 6 l 7 18J19 20 21 22 23 2lr2$i26127!28 29 3P 31 33 - t - - t -Group II -1 \u00E2\u0080\u00A2 3L35 56 |37j38 39 1*0 frl 1*2, 1*3 1*1* U5 jl*6 jl*7 [1*8 ;U9 i5o 51,52 J53 jft & \u00C2\u00A36 57 58 j59 |60 pi 62 63 61* ;65 66 ^7 \u00E2\u0080\u0094 f . . .Group,111 ,' , ; 72 j 73 j 7li j 75 j 76 j 77 j 78' 79 80 8X 8218^181, i R< I ftA ' fl? j 8 Q Q 9 ! 9 Q 9 1 9 p Indicates the test is not applicable because of the fragmentary condition of the MS (also the Latin riddle, 90). As 1-2-3 are likely a l l one riddle, 1 and 3 should only be counted as one using; a closing formula. See also Riddlle 36. They are both counted, Ihowever, for the later tests. Counted as one for this test, though Riddle 36 actually contains two closing formulas. raed hwaet ic maene may be of Type L or N, but more likely is a single example of a further type. Though damaged, the opening of 82 was probably of the wiht is wraetlic type. \u00E2\u0080\u0094 \u00E2\u0080\u0094 Table II Distribution of Opening and Closing Formulas - 51 -Frequency of Opening and Closing Formulas Test A. Opening Formulas Group I II III possible 27 29 33 occurred 8 19 9 frequency .30 .66 .27 Test B. Closing Formulas possible occurred 26 11 29 6 28 8 frequency .U2 .21 .29 Test C. No Formulas possible 26 28 27 occurred 10 7 13 frequency .38 .25 \u00C2\u00ABU8 Opening Formulas: D. occurred 2 8 2 frequency .07 .28 .06 E. occurred 1 5 3 frequency .OU .17 .09 F. occurred 5 - -frequency .18 -G. occurred - 3 frequency - .10 H. occurred - 2 1 frequency - .07 .03 Closing Formulas: L. occurred 7 2 frequency .27 .07 M. occurred 1 5 frequency .Ok .17 N. occurred h frequency .15 Sample calculation: Test D, Group I: possible to occur: 27 actually occurred: 2 frequency =2y = -07 Table III - 52 - . found i n Groups I and III. The related formula E, i c wiht geseah, shows a similar distribution, with five of the nine occurrences i n the middle group, of which four are clustered. Of the f i v e cases of F, i c eom wunderlicu wiht, however, a l l occur i n a clump i n Group I. Formulas G and H do not really occur often enough to provide r e l i a b l e information, but a l l three cases of G, i c wat, appear i n Group I I . The closing formula saga hwaet i c hatte and i t s variants, L, are frequently seen i n the f i r s t and la s t groups (seven times i n Group I and six times i n Group III), but only twice i n Group I I . Conversely, the elaborate formula M, hwaet seo wiht sy, which incorporates a l l the micel i s to hycganne... type, i s found five times i n the middle group, but only once each i n the others. Though N, fr i g e hwaet i c hatte, occurs i n only four riddles i n the collection, a l l four are i n Group I. Of the remaining individual opening and closing formulas, none appears often enough for any v a l i d conclusion to be made. An examination of the distribution of the adverbs hwilum, oft and nu also demonstrates the three-part division, as shown i n Table IV. Remarkably often these words appear at the beginning of a h a l f - l i n e . 17 Indeed, of the 100 times hwilum begins a half - l i n e i n a l l Old English poetry, f u l l y 5\ of them are i n the Exeter riddles I Hwilum i n fact appears only seven times i n the riddles when i t does not begin a half-l i n e . Hwilum, oft and nu frequently begin a sentence or a syntactical unit, and so attain often an almost rhetorical function, directing the movement of the poem at a c r i t i c a l point, or sometimes, especially i n the case of hwilum, setting up a parallelism within the riddle which may even serve as i t s dominant logic. Occasionally i n Group I - 5 3 -Occurrence of Hwilum, Oft and Nu Adverb Begins Half-Line: Total by oft: nu: Group I Group II Group III riddles 2 i i 9 61 I 10 3 ( x 8 ) 57 62 ( x 2 ) 6 ( x 2 ) 63 II 2 7 73 12 ( x 5 ) 80 III 10 m (xio) 83 22 17 85 20 89 2h ( x 7 ) 91 27 93 ( x 5 ) ii 30 ( f u l 61 I 5 5 oft, and 63 -6 3 0 b ) 72 ( x 2 ) II k 16 iili 77 20 ( x 3 ) U9 78 III 10 5 3 80 19 8i i ( x 2 ) 88 (ful oft) 91 93 Hi UO ( x 2 ) 71 I 3 16 l i2# 73 27 53 77 II H 5 5 * 83 88 ( x 2 ) III 8 92 Total by-occurrence I 37 II 2 III 15. ' 7 ii 12 23 3 5 10 IE 93 ( x 2 ) 95 * introduces closing Table IV, i - 5U -Scribal Point Precedes Adverb Total by Total by hwilum: Group I Group II Group III riddles occurrence 2 73 I Mi I 1U 3 (x3) 85 LU (x3) 89 II - II -2k (x7) 91 (25) 93 (x3) III 5 III _7 9 21 oft: 20 93 I II III 1 2 1 2 nu: 26 27 73 88 93 I II III 3 5 3 5 Total Occurrence of Adverb, A l l Positions I II III Total hwilum: 39 2 19 60 oft: 8 7 15 30 nu: 3 6 11 20 Table IV, i i hwilum i s found i n profusion i n such a role: five times i n the 15 lines of Riddle 12, 10 times i n the 19 lines of Riddle LU and seven times i n only 10 lines i n Riddle 2U. Though l i t t l e i s known about the Old English scribal practice of pointing, which appears to have served as a sort of punctuation, this kind of rhetorical importance, especially of hwilum, i s perhaps emphasized by the fact that i n two-fifths (21 of 5U) of the cases i n which hwilum begins a half-line, and even once (25^, and here hwilum may have been placed at the end of the b-line only because of syntactical twisting to accommodate the alliteration) when i t does not, i t i s preceded by a scribal point. In Riddle 2U a l l seven cases of hwilum are set off this way. Though this phenomenon does not seem to apply as much to oft (only i n two of 23 cases where oft begins the half-line), oft has been chosen as the opening word i n four of the riddles, 16, 61, 63 and 78. A scribal point precedes nu i n five of 18 cases i n which i t begins the half-line, and remarkably i n three of these cases, 27, 88 and 93, nu i s also capitalized. This occurs at the beginning of the a^lihe i n 88^ and 93^^, but at the beginning of the b-line i n 27^. In 88^ the scribe even goes so far as to precede the capitalized Nu with two points. A difference i n the frequency of employment of these adverbs between the three groups i s easily seen i n Table IV. Of the 5U places i n 22 riddles where hwilum begins the half-line, the great majority, 37, are found i n Group I. Fifteen appear i n Group III, but hwilum i s seen only twice i n any position i n a l l of Group II. Though the usage i s heavier i n Group I, hwilum occurs i n 10 riddles in both the f i r s t and last groups. The figures for oft and nu do not indicate as much difference between the f i r s t two groups as those for hwilum, but definitely show a far heavier usage, twice that for any other group, i n Group III. Though the difference i n the frequency of occurrence of scribal points preceding hwilum, oft and nu i s perhaps not great enough for a strong indication, i t i s notable that though there are 28 examples of i t i n Groups I and III, this arrangement i s never seen i n Group I I . - 57 -The Relationship Between the Groups I f the riddle collection i s made up of three distinct parts, can anything be determined about the relationship between the groups? Actually, there seems to be far less difference between the riddles of Groups I and II than between those of the third group and the others. Of course Group III stands alone at the end of the MS, whereas Groups I and II are presented together, but there are other reasons for making such a judgement. S t y l i s t i c a l l y , though there i s nothing i n Group II to match the magic of the magnificent 'storm' series which opens the f i r s t group, and though there i s nothing i n Group I to equal the sustained dullness of Riddle UO, the fluctuations i n quality between the riddles of the f i r s t two groups are by no means the wide variations found i n Group III. Despite whatever differences there may be between the riddles of the two earlier groups, there i s s t i l l i n every case an attempt made to produce a l i t e r a r y creation offering the listener or reader a certain minimum standard of l i t e r a r y adornment. In several riddles of the last group, however, this i s not the case. Riddles 68 and 69 (Tupper and Mackie put them together as a single poem, but the point holds nonetheless), 75, 76 and 79 (though i t may be just a mis-start of the following riddle) are bald enigmas even by folk-riddle standards, 18 and i f they are i n fact, as Trautmann would have us believe, not 19 folk but l i t e r a r y pieces (and even Tupper's c r i t e r i a would seem to admit Riddle 75 to be l i t e r a r y because of i t s runes), then they are very crude pieces indeed. Yet, on the other end of the scale, a few of the riddles, such as 88, 93 'horn' and Qk 'water,' come close at - 58 -times to the best i n the earlier groups. Such a wide divergence i n riddle quality i n the third group can only lead me to conclude that this group, at least, i s a very miscellaneous collection containing the work of a number of poets of a broad range of a b i l i t i e s . As the overall differences between Groups I and II are not great\u00E2\u0080\u0094indeed no one has before found reason to suggest a firm division of any kind i n Riddles 1-59\u00E2\u0080\u0094it i s most profitable to study the relationship of the riddles of the last group to those of the earlier parts. Group III opens with a t r i o of 'obscene' riddles, but though these do have a number of expressions i n common with earlier 'obscene' pieces, as well as with each other, i t would perhaps be incautious to conclude that any definite connections exist, because of the great p o s s i b i l i t y of independent use of the same traditional material i n this type of riddle. However, the following s i m i l a r i t i e s should be 8 9 * 8 noted: 61 on nearo fegde, 25 fegeo mec on faesten, 62 on nearo 9 5 5 nathwaer; 61 ruwes nathwaet, 25 neoban run nathwaer, 5U stibes 5 8 7 nathwaet; 63 t i l l i c esne, Sh t i l l i c esne; 63 wyrceS his w i l l a [ n j , 5U^ worhte his willan. Immediately following these i s another series, f i v e riddles this time, a l l with demonstrable connections with earl i e r pieces. Riddle 6h i s a runic 'hunt1 riddle, about which Krapp and Dobbie remark: \"Regardless how we interpret the runes i n d e t a i l , i t i s clear that 20 this riddle i s a companion to Riddle 19.\" Next i s the 'onion' riddle, 65, which i s merely a non-salacious version of i t s 'obscene' 21 counterpart. In both, as Tupper points out, the biter of the onion i s seen as the destroyer. As well, both enigmas share the 'bite only - 59 -t the biter' idea which originates with Symphosius 1+1+: 65 monnan i c ne 2-3 bite nymppe he me bite, 25 naengum scebbe / burgsittendra nymbe 2-3 bonum anum; the attack on the head: 65 aeghwa mec reafa&...ond min heafod scireb, 25 reafa5 min heafod; and the confinement of the subject: 65^ hafaS mec on headre, 25^ fege5 mec on faesten. Riddle 66, as mentioned earlier, i s nothing but a re-working of Riddle 1+0. \"It owes nothing to Aldhelm's De Creatura directly,\" Tupper remarks, \"but i s a very free reshaping of some of the material furnished by the second hand i n 1+1,\u00E2\u0080\u009E[K-D 1+0') 82f.\" 2 2 The damaged riddle 67, 'Bible,' i s only a new treatment of the subject of 26, and shares with i t at least one p a r a l l e l : 67 golde gegierwed, 26 gierede mec mid golde. In addition the wisdom and teaching of the book appear i n both, though handled i n different ways: 67^ [...} wearS / leoda lareow, 26\"^ g i f min bearn wera bruean willa3 hy beo8 by gesundran ond by sigefaestran heortum by hwaetran ond by hygeblibran ferbe by frodan...etc. The last riddle i n the series i s 68, which i s v i r t u a l l y identical to the f i r s t two lines of Riddle 36. In 36, however, the lines are only an opening formula, so i t i s not at a l l clear what to make of their reoccurrence here as a complete poem. Many scholars have taken 68 and the next riddle, which is only one line long, as one poem together, but the objections of Krapp and Dobbie that \"there i s nothing i n the 23 texts of the two riddles which would j u s t i f y us i n associationg them\" seems a sound judgement. The elevation of the last word of 68 i n the MS ( f o l . 125V) for economy of space would seem to indicate strongly that the scribe of this MS, at least, saw no connection at a l l between - 6 0 -the two. Beyond this series there are s t i l l further connections between Group III riddles and earlier ones. The 'ox' riddle 7 2 has i n common with i t s two predecessors 1 2 and 3 8 the association of i t s subject with the swarthy slave: 72^ swearturn hyrde, 7 2 1 2 mearcpapas walas, 1 2 ^ Q swearte wealas, 1 2 wonfeax wale, and i t s source of nourishment i n youth: 7 2 oft i c feower teah / swaese broper, 3 8 feower wellan. Related to the drinking of the young ox i s the provision of milk by the adult i n 1 2 ^ ^ , hwilum i c deorum drincan selle / beorn[e] of 8 5 bosme. Note the verbal similarity between 7 2 drincan sealde and 1 2 drincan selle. Even closer i s the similarity between the 'bellows' riddles 8 7 and 3 7 , which at the beginning are almost doublets: Ic seah wundorlice wiht wombe haefde micle prypum geprungne pegn folgade ^ _ maegenstrong ond mundrof ( 8 7 ) Ic pa wihte geseah wombe waes on hindan bripum aprunten pegn folgade , maegenrofa man ( 3 7 1 \" 3 ) The fragment 8 9 also appears to have been very closely related to these two. Though l i t t l e i s l e f t of i t , the parallels i n two elements 2 li are obvious: wombe haefde f...] 8 9 , [...] on hindan 8 9 . Riddle 9 1 ('key'), lik e 65, i s a non-salacious treatment of the subject of an earlier 'obscene' piece, though i t i s not nearly as closely related to kh as 65 i s to 25. One point i n common should be noted, however: 5 2 - 3 9 1 hearde wi5 heardum hindan byrel, hh ~ foran i s byrel / bi& stiS ond heard. And the fragmentary Riddle 9k i s , with l i t t l e doubt, another 'creation' riddle, since i t employs the same comparative \-technique (smaller than..., higher than..., etc.) as kO and 6 7 , and - 61 -shares some of their motives: 9k hyrre bonne heofon, kO hyrre i c 6 3 3 eom heofone, 67 heofonas oferstige; 9k glaedre bonne sunne, 67 6 76 swiftre bonne sunne; and possibly 9k leohtre bonne w \u00C2\u00A3... ] , kO leohtre i c eom micle bonne bes l y t l a wyrm, though the damaged word i n 9k^ cannot have been wyrm, as there are no descending strokes on the next four letters (the tops are obliterated by a hole i n the MS) after w. Finally, Paddle 95, despite i t s confusion arid lack of an adequate solution, has much i n common with the 'body and soul' riddle, 1+3. Indeed Trautmann solves^it, though unconvincingly, i n his e d i t i o n ^ as 'der Geist.' Most obvious of course are the opening half-lines, which d i f f e r only i n point of view: 95^~ Ic eom indryhten, k3^ Ic wat indryhtne. Also, the subject's wide acquaintance arid frequent sojourns 2 2 with the rich and lowly i n 95 are paralleled i n k3 , where the soul i s a guest i n the courts. The journey of the soul, k3^, i s matched by the wide travels of the subject i n 95 . The conditional wiste ond i 8 blisse which the soul may win under favorable circumstances i n k3 5-6 may even shed some light on the passage 95 gif i c habban sceal / blaed i n burgum obbe beorhtne god, which has caused many a headache to riddle scholars. A l l of the above parallels have been between riddles related by subject (or, i n the case of the 'obscene' riddles, by genre) to earlier ones. Even more striking, however, are three parallels between riddles which seem otherwise completely unrelated. Most peculiar i s the lin e haebbe me on bosme baet on bearwe geweox, 80^. The expression paet on bearwe geweox does not, to my ear at least, have the well-worn and 2 familiar ring of the traditional formula. Only by reference to 27 -62 -brungen of bearwum does i t become plain that what the horn of 80 holds i n i t s bosom i s mead, which i s made from honey from the woods, h i l l s , valleys and downs of 27. Much the same relationship i s found between 81^ ^ aglac dreoge / baer mec wegeS se be wudu hrere9 and 1^~^\u00C2\u00B0 bonne i c wudu hrere / bearwas bledhwate beamas f y l l e , and between 92 1 Ic waes brunra beot beam on holte and UO1^^ amaested swin / bearg bellende Cbe] on bocwuda / won wrotende wynnum l i f d e . Here too the extremely cryptic circumlocutions of the later riddles are explained by passages from earlier riddles. In some of the related riddles above, the parallels are limited to only a half-line or two; i n others they are more extensive. But i n those where the similarity i s greatest\u00E2\u0080\u009465-25, 66-UO, 68-36, 87-37, 9U-U0-66\u00E2\u0080\u0094the parallels seem strong enough to eliminate any likelihood of them being caused merly by chance or by the use of traditional expression. Tupper and other scholars have seen such connections as indications of unity of authorship (Tupper uses the word 'homogeneity,' as he finds a few of the riddles to be different), but the overall differences between Group III and the other two groups, discussed earlier i n this chapter, would seem to discourage this explanation. The three peculiar parallels between unrelated riddles\u00E2\u0080\u0094 8 0 - 2 7 , 81-1, 92\u00E2\u0080\u0094Lj.0\u00E2\u0080\u0094however, appear to suggest some sort of dependence of Group III on the earlier riddles. In these three examples, at least, i t does not seem l i k e l y that the later passages would have been written without the prior existence and influence of the ear l i e r ones. P a r a l l e l passages with this kind of relationship could easily have occurred i n a situation i n which a l i t e r a r y work had become well-loved and thoroughly - 63 -known to a number of people. The l i f t i n g of passages from mutually familiar pieces of literature for use i n argument and witty conversation has always been common practice not only among scholars but among liter a t e people generally. Could not such a practice also have been employed for enigmatic purposes? If such a dependence exists between these three pairs of riddles, then we might reasonably conclude that the writers of the third group had an intimate knowledge of the f i r s t two groups. If this p o s s i b i l i t y i s considered, then some of the other links between Group III riddles and earlier ones can reasonably be explained. Riddles 66 and 9h may be new riddles modelled on Riddle 1*0. Riddle 87 could similarly have been begun i n imitation of 37 (the f i r s t f i v e half-lines are very close), with a subsequent d r i f t into independence as the riddle progressed. Riddle 95, though i t probably does not share the same solution as h3, may have been modelled on i t as a beginning, much the same as 87-37, but with the motives used for a purpose now obscured by garbling. Riddles 6U, 65, 67, 68, 72 and 91 f a l l into place as less closely related, but similar cases. An assumption that at least some of the riddles i n Group III are the products of imitation of Group I and II riddles, done by a number of authors, though not a conventional one, i s nonetheless i n keeping 25 with known 'practice among'.the Anglo-Saxons. F.H. Whitman has pointed out that Latin enigmas appear to have been regularly used for the instruction of students i n grammar during this period, and that Aldhelm himself seems to have considered his riddles \"an elementary exercise before the undertaking of a more important task....Coming where they - 6k -do, i n his Epistle to Aldfrid, which includes a dialogue on metrics, i t seems l i k e l y that they were intended to serve as illustrations 26 of the theoretical principles which he had exhibited i n the dialogue.\" Whitman also notes that Latin riddles are often found gathered with material pertaining to grammatical studies, and that interlinear glosses i n the MS containing the riddles of Tatwine, Aldhelm and Symphosius \"seem to indicate that the manuscript was once used by a master for 27 the teaching of rudimentary Latin.\" If riddles were used as creative exercises i n Latin and as models i n the teaching of Latin composition, then i t seems l i k e l y that they would have been employed for grammatical purposes i n the vernacular as well. It i s conceivable, therefore, that Group III i s actually a collection of the efforts of a number of student writers, some riddles of which are imitations of Group I and II enigmas, and others perhaps original compositions or imitations of other riddles now lost, a l l anthologized with a few more accomplished pieces (81*, 88, 93), which may themselves be imitations. Some older riddles may be included, but the dearth of demonstratives and definite articles which Hacikyan points out i n a l l of Group III, i f taken as a' s t y l i s t i c feature rather than as evidence for dating (see Chapter II), would indicate a certain uniformity of usage easily explainable as the product of a common overriding style\u00E2\u0080\u0094perhaps one imposed by a grammar master. Such imitation by a number of authors would also provide a credible explanation for the presence of one, two and even three treatments i n Group III of single subjects dealt with i n earlier parts of the collection: two 'creation' riddles, 66, 9U(iiO), one \"Bible' riddle, - 65 -67(26), one \u00C2\u00BBox\" riddle, 72(12, 38), three 'horn' riddles, 80, 88 arid 93(Hi), probably two 'bellows' riddles, 87, 89(37), and the two non-salacious'treatments, 65(25) arid 91(U0, of earlier 'obscene pieces. - 66 -FOOTNOTES TO CHAPTER III 1. I have placed Riddle 61 i n Type A and 86 i n Type C, though i n each case c l a s s i f i c a t i o n i s d i f f i c u l t . In 61, though the whole riddle i s obviously of the A type, the formulaic tag raed hwaet i c maene might seem to be more appropriate to a B-type riddle i f i c i s taken as the riddler (as i n Baum's translation). More l i k e l y the o r i g i n a l wording was similar to the tag on the following piece, saga hwaet i c hatte, but with raed replacing saga because an r - a l l i t e r a t o r was needed. Of course, i f maene i s taken (as i n Mackie's translation) to mean 'signify,' the problem disappears. In Riddle 86 an error i n transmission has resulted i n an A-type tag being appended to a C-type riddle. Though Riddle 1*1 i s incomplete due to a break i n the MS between f o l i o s 111b and 112a i t has been placed i n Type C, as there i s nothing to suggest any first-person forms were included i n the missing part. Riddle 67 i s placed i n Type B, as both the beginning and the ending seem to demand i t . The section though, seems to have the object speaking. Mackie translates the passage as a quoted speech of the object, told by the riddler. This interpretation would seem to solve the problem, and the riddle would remain Type B. Fragment 78, though l i t t l e sense can be made of what i s l e f t of i t , i s obviously Type A. Three fragments were deemed too doubtful for c l a s s i f i c a t i o n . No. 82 i s probably C, though i t could possibly be B. No. 88 i s l i k e l y of the B type, but missing the i c seah opener. Of No. 9k there i s no evidence for cl a s s i f i c a t i o n l e f t , but Mackie's reconstruction would make i t an A. 2. \"Die AuflBsungen der altenglischen RStsel,\" Anglia Beiblatt 5(1916), 1*6, and i n his edition, pp. 1-1*-3. E. Erlemann, \"Zu den altenglischen RStseln,\" Archiv 111(1903), 1*9-63. 1*. Paull F. Baum, trans., Anglo-Saxon Riddles of the Exeter Book (Durham, N.C.: Duke University Press, 1963), pp. 3-7. 5. \"The Old English Storm Riddles,\" PMLA 61*(19l*9), 181*-188. 6. REB, pp. 162-161*. 7. The obvious fact that this would produce a very long riddle indeed (161 lines or so, whereas the next longest possible would be the combined 'storm' riddles 1-2-3 of 109 lines) need not be t e r r i b l y alarming, as Aldhelm's original i t s e l f i s five times as long as any of his others. 8. p. 16. 9. Tupper, REB, pp. xcvi, 16; Krapp and Dobbie, p. 191; Baum, p. 1*2. 10. p. 59. - 67 -11. The average number of lines per riddle i n the f i r s t section (1-59) i s Hi, taken over 53 riddles. Those not counted were 1-3, as they are under dispute here; 18, as there i s clearly a hiatus i n i t ; 20, which may be a fragment; and Ul, which has no beginning. If UO i s omitted because of i t s unusual length, the adjusted average i s I2f- lines per riddle. 12. FOrster's suggestion that a further single leaf may have been l o s t between f o l s . 111-112 should not endanger this placement of the division, as he advances i t only as a p o s s i b i l i t y . If just two leaves are indeed missing, and I think my arguments show that such an assumption i s reasonable (as far as can be judged from this single version of the riddles, at least), then the likelihood i s much stronger, due to the apparent break i n the text, that the second leaf has f a l l e n out between f o l s . 105-106 rather than between 111-112. 13. cf. Cook, p. xiv. lU- See F.A. Blackburn, \"The Husband's Message and the Accompanying Riddles of the Exeter Book,\" JEGP 3(1900), 1-3; R.W.V. E l l i o t t , \"Runes i n the Husband's Message,\" JEGP 5U(1955), 1-8. 15. REB, p. lxv. 16. Tupper l i s t s U2, 55, 57 and 59 as closing with a formula, but I disagree. See Appendix C, closing formula differences. 17. I am indebted to F.H. Whitman for leading me to this information. 18. \"Quellen der Altenglischen Ratsel,\" p. 353. 19. REB, p. lxxxv. 20. p. 368. 21. REB, p. 123. 22. REB, p. 207. 23. p. 369. 2U. Die Altenglischen Ratsel (New York: G.E. Stechert, 1915), p. 1U0. 25. F.H. Whitman, \"Medieval Riddling: Factors Underlying i t s Development,\" Neuphilologische Mitteilungen 71(1970), 177-185. 26. Whitman, p. 181. 27. i b i d . - 68 -The reappearance i n Riddle 93 of the unusually heavy use of hwilum clauses of Riddle lij. (hwilum also appears twice i n 80, though only once at the beginning of a half-line) should be pointed out as further evidence for a connection between the 'horn' riddles of Group III and the earlier one. - 69 -CONCLUSION Opinion about the unity of the Exeter Book riddles has been divided. Earlier scholars generally f e l t the collection was a unified whole, but since the decline of the Cynewulfian authorship theory, such a view has f a l l e n into disfavor. Most modern c r i t i c s have seen the collection as a miscellany with no overall plan or unity. Scholars who have proposed block divisions i n the text have made the lo g i c a l suggestion that only one break occurs (besides the separation of 30b and 60 from the rest), and that that i s at the textual break following Riddle 59. The distribution of points of view in the riddles, however, establishes a three-part structure (excluding 30b and 60) i n the text of the riddles. One break predictably occurs after Riddle 59, but an earlier break also appears, one which cuts the section 1-59 more or less i n half. If i t can be assumed that this section of the riddles once totalled 60 (like the collection of Eusebius) and that the break i n i t was between two originally equal groups of 30, then the mid-break may be tentatively located between Riddles 29 and 30. This calculation assumes the loss of one sheet which would have provided missing folios between Fols. 105-106 and between 111-112, though i t i s necessarily confined to assumptions of loss only from this MS, and not from preceding ones. Examination of the distribution of opening and closing formulas over the collection, and of the distribution of the adverbs hwilum, oft and nu, seem to indicate support for this three-part grouping. Riddle 30b could be seen as merely a mis-start of the second group of - 70 -riddles (30-59) at a point later i n the MS. Riddle 60 probably follows only because of i t s similarity to the \"Husband's Message.\" Parallels between the f i r s t two groups of riddles do not afford any good opportunities for conclusions about their relationship, but some striking points i n common between Group III and the f i r s t two groups appear to suggest that Group III i s to some extent based on Groups I and II, and that a number of the later riddles may have been modelled on earlier ones. The diversity i n quality between the riddles of Group III also suggest a number of authors rather than one. It i s possible that Group III may be an anthology containing i n part the work of grammar students, as the writing of riddles as composition exercises and the use of them as examples for study and imitation appears to have been common practice during the Anglo-Saxon period. - 71 -LIST OF WORKS CONSULTED Alexander, Michael. The Literature of the Anglo-Saxons. Princeton: Princeton University Press, 19U9. Barnouw, Adrian Jacob. Textkritische Untersuchungen nach dem Gebrauch des Bestimmten Artikels und des Schwachen Adjectivs i n der Altenglischen Poesie. Leiden: E.S. B r i l l , 1902. Baum, Paull F., trans. Anglo-Saxon Riddles of the Exeter Book. Durham, N.C.: Duke University Press, 1963. Benson, Larry D. \"The Literary Character of Anglo-Saxon Formulaic Poetry.\" PMLA 81(1966), 33U-3U1. Bosworth, Joseph. An Anglo-Saxon Dictionary. Edited and enlarged by T. Northcote Toller. London: Oxford University Press, 195U. Bradley, Henry. Review of Morley's English Writers, vol. II. Acadamy 33(1888), 197-198. Brooke, Stopford A. The History of Early English Literature. London: Macmillan, 1892, vol. I. Blackburn, F.A. \"The Husband's Message and the Accompanying Riddles of the Exeter Book.\" JEGP 3(1900), 1-13. Brown, Carleton F. \"The Autobiographical Element i n the Cynewulfian Rune Passages.\" Englische Studien 38(1907), 196-233. Cassidy, Frederic G. \"How Free was the Anglo-Saxon Scop?\" Franciplegius: Medieval and Linguistic Studies i n Honor of Francis Peabody Magoun, Jr. ed. Jess B. Bessinger, Jr. and Robert P. Creed. New York: New York University Press, 1965, 75-85. Chambers, R.W., Max Ftfrster and Robin Flower, eds. The Exeter Book of Old English Poetry. London: Lund, Humphries and Co., 1933. Conybeare, John J., ed. Illustrations of Anglo-Saxon Poetry. London: Harding and Lepard, l82o~! Cook, Albert S., ed. Christ of Cynewulf. Boston: Ginn and Co., 1900. Courthope, W.J. A History of English Poetry. London: Macmillan, 1895, vol. I. Das, Satyendra Kumar. Cynewulf and the Cynewulf Canon. Calcutta: University of Calcutta, 19U2. ' - 72 -Dubois, Marguerite-Marie. Les Elements Latins dans l a Poesie Religieuse ,, de Cynewulf. Paris: Librairie E. Droz, 19^3. ,, Dietrich, Franz. .''Die Rathsel des Exeterbuchs: Wurdung, LcJsung und f Herstellung.\" Zeitschrift fur deutsches Alterthum 11(1857) uu8-u90. . \"Die RSthsel des Exeterbuchs: Verfasser, Weitere Losungen.\" Zeitschrift fur deutsches Alterthum 12(1865), 232-252. Eliason, Norman E. \"On Wulf and Eadwacer.\" Old English Studies i n Honor of John C. Pope, ed. Robert B Burlin and Edward B. Irving Jr. Toronto: University of Toronto Press, 1971*, 225-23U-. \"Riddle 68 of the Exeter Book.\" Philologica: The Malone Anniversary Studies, ed. T.A. Kirby and H.B. Woolf. Baltimore: Johns Hopkins Press, 19U9, 18-19. E l l i o t t , R.W.V. \"Runes i n the Husband's Message.\" JEGP 5U(1955), 1-8. Erlemann, Edmund. \"Zu den altenglischen Ratseln.\" Archiv 111(1903), 1x9-63. Erlemann, F r i t z . \"Zum 90. angelsachsischen Ratsel.\" Archiv 115(1905), 391-392. Fry, Donald K. \"Wulf and Eadwacer: A Wen Charm.\" Chaucer Review 5(1970), 2U7-263. Gradon, P.O.E. Cynewulf's Elene. London: Methuen and Co., 1958. Greenfield, Stanley B. A C r i t i c a l History of Old English Literature. New York: New York University Press, 1965. \u00E2\u0080\u00A2 . Interpretation of Old English Poems. London: Routledge and Kegan Paul, 1972. Hacikyan, Agop. A Linguistic and Literary Analysis of Old English Riddles. Montreal: Mario Casalini, 1966. Henry, P.L. The Early English and Celtic Lyric. London: Allen and Unwin, 1966. Herzfeld, Georg. Review of August Madert, Die Sprache der altenglischen Ratsel des Exeterbuches und die Cynewulffrage. Archiv 106(1901), 389-390. Hicketier, F. \"Funf Ratsel des Exeterbuches.\" Anglia 19 (1888), 56U-600... Holthaus, E. Review of August Prehn, Komposition und Quellen des Exeterbuches. Anglia 7(l881i), anzeiger 120-125. - 73 -HuppI, Bernard F. The Web of Words. Albany: State University of New-York Press, 1 9 7 0 . Isaacs, Neil D. Structural Principles i n Old English Poetry. Knoxville, Tenn.: University of Tennessee Press, 1968. Kemble, John Mitchell. \"On Anglo-Saxon Runes.\" Archaelogia 28(181*0), 237-272. Kennedy, Charles W. The Earliest English Poetry. London: Oxford University Press, 191*3. . The Poems of Cynewulf. London: George Routledge and Son, 1 9 1 0 . Krapp, George Philip, and E l l i o t t van Kirk Dobbie, eds. The Exeter Book. New York: Columbia University Press, 1 9 3 6 . Lawrence, W.W. \"The F i r s t Riddle of Cynewulf.\" PMLA 1 7 ( 1 9 0 2 ) , 21*7-261. Lichtenheld, A. \"Das schwache Adjectiv im Angelsachsisch.\" Zeitschrift fur Deutsches Alterthum 1 6 ( 1 8 7 3 ) , 3 2 5 - 3 9 3 . Mackie, W.S., ed. The Exeter Book Part I I : Poems IX XXXII. EETS os. No. 19l*. London: Oxford University Press, 1931*. Mather, F.J. \"The Cynewulfian Question from a Metrical Point of View.\" MLN 7 ( 1 8 9 2 ) , 193-213. Morley, Henry. English Writers. Vol. II. London: Cassel and Co., 1 8 8 8 . Nuck, R. \"Zu Trautmann's Deutung des ersten und neunundachtzigsten Ratsels.\" Anglia 1 0 ( 1 8 8 8 ) , 391-391*. Patzig, H. \"Zum ersten Ratsel des Exeterbuches.\" Archiv 11*5(1923), 201*-207. Quirk, Randolph. \"Poetic Language and Old English Metre.\" Early English and Norse Studies Presented to Hugh Smith, ed. Arthur Brown and Peter Foote. London: Methuen and Co., 1 9 6 3 , 150-171. Raffel, Burton, trans. Poems from the Old English. Lincoln, Neb.: University of Nebraska Press, I960. Reiger, Max. \"Uber Cynewulf.\" Zeitschrift fur Deutsche Philologie 1 ( 1 8 6 9 ) , 2 1 5 - 2 2 6 . Renoir, Alain. \"Wulf and Eadwacer, a Noninterpretation.\" Franciplegius: Medieval and Linguistic Studies i n Honor of Francis Peabody Magoun, Jr. ed. Jess B. Bessinger, Jr. and Robert P. Creed. New York: New York University Press, 1 9 6 5 , 11*7-163. - 7U -Renwick, W.L. and Harold Orton. The Beginnings of English Literature to Skelton. London: Cresset Press, 1939. Ricci, Aldo. \"The Chronology of Anglo-Saxon Poetry.\" The Review of English Studies 5(1929), 257-266. Sarrazin, Gregor. \"Beowulf und Kynewulf.\" Anglia 9(1886), 515-550. Schlauch, Margaret. English Medieval Literature and i t s Social Foundations. Warszawa: Panstwowe Wydawnictwo Navkowe, 1956. Schofield, William H. \"Signy's Lament.\" PMLA 17(1902), 262-295. Shippey, T.A. Old English Verse. London: Hutchinson and Co., 1972. Shook, Laurence K. \"Old English Riddle No. 20: Heoruswealwe.\" Franciplegius: Medieval and Linguistic Studies in Honor of Francis Peabody Magoun, Jr. ed. Jess B. Bessinger, Jr. and Robert P. Creed. New York: New York University Press, 1965, 19U-20U. Sievers, E. \"Zu Cynewulf.\" Anglia 13(1891), 1-25. Sisam, Kenneth. \"Cynewulf and his Poetry.\" Proceedings of the British Acadamy, 1932. London: Oxford University Press. . Studies i n the History of Old English Literature. Oxford: Clarendon Press, 1953. Smith, A.H. Three Northumbrian Poems. London: Methuen, 1933-Smith, Lucy Toulmin. \"Kynewulf\" i n The Dictionary of National Biography, ed. Sir Leslie Stephen and Sir Sidney Lee. London: Oxford University Press, 1892, repr. 1938, vol. II, 358-360. Stevens, Martin and Jerome Mandel. Old English Literature. Lincoln, Neb.: University of Nebraska Press, 1968. Strunk, William Jr. The Juliana of Cynewulf. Boston: D.C. Heath, 190U. Sweet, Henry. \"Sketch of the History of Anglo-Saxon Poetry.\" i n Thomas Warton's History of English Poetry, ed. W. Carew Hazlitt. London: Reeves and Turney, 1871, vol. 2, 16-18. Taylor, Archer. \"The Varieties of R i d d l e s . P h i l o l o g i c a : The Mal one Anniversary Studies, ed. T.A. Kirby and H.B. Woolf. Baltimore: Johns Hopkins Press, 19U9, 1-8. Ten Brink, Bernhard. The History of English Literature, trans. Horace M. Kennedy. London: Bell and Sons, 19LU, vol. I. - 75 -Thomas, P.G. English Literatue Before Chaucer. London: Edward Arnold and Co., 192IH Thorpe, Benjamin, ed. Codex Exoniensis: A Collection of Anglo-Saxon Poetry. London: William Pickering, 181*2. Trautmann, Moritz. \"Cynewulf und die Ratsel.\" Anglia 6 ( 1 8 8 3 ) , anzeiger 1 5 8 - 1 6 9 . . \"Das Sogenannte Erste Ratsel.\" Anglia 3 6 ( 1 9 1 2 ) , 1 3 3 - 1 3 8 . . \"Quellen der Altenglischen Ratsel.\" Anglia 3 8 ( 1 9 1 1 * ) , 31*9-351*. . \"Sprache und Versbau der Altenglischen Ratsel.\" Anglia 38(1911*), \" 1 5 5 - 3 6 U . . \"Zeit, Heimat und Verfasser der Altenglischen Ratsel.\" Anglia ~3BTl911t), 3 6 5 - 3 7 3 . . Die Altenglischen Ratsel. New York: GtE'., Stechert, 1 9 1 5 . Tupper, Frederick, Jr. \"Originals and Analogues of the Exeter Book . Riddles.\" MLN 1 8 ( 1 9 0 3 ) , 9 7 - 1 0 6 . . \"Solutions of the Exeter Book Riddles.\" MLN 2 1 ( 1 9 0 6 ) , 97 -105. . \"The Cynewulfian Runes of the First Riddle.\" MLN 2 5 ( 1 9 1 0 ) , ~ 2 3 T - 2 l * l . , ed. The Riddles of the Exeter Book. Boston: Ginn and Co., 1 9 1 0 . Von Erhardt-Siebold, Erika. \"The Old English Loom Riddles.\" Philologica: The Malone Anniversary Studies, ed. T.A. Kirby and H.B. Woolf. Baltimore: Johns Hopkins Press, 191*9, 9-17. . \"The Old English Storm Riddles.\" PMLA 61*(19l*9), 881*-888. Whitman, F.H. \"Medieval Riddling: Factors Underlying i t s Development.\" Neuphilologische Mitteilungen 71 ( 1 9 7 0 ) , 177-185. Williams, Blanche C. Gnomic Poetry i n Anglo-Saxon. New York: Columbia University Press, 1911*. Wilson, R.M. The Lost Literature of Medieval England. London: Methuen and Co., revised ed., 1 9 7 0 . Wrenn, C.L. A Study of Old English Literature. New York: W.W. Norton, 1967. - 7 6 -Wulker, Richard. Grundriss zur Geschichte der Angelsachsischen Literatur. Leipzig: Veit, I t ibS. Wyatt, A.J., ed. Old English Riddles. Boston: DvC. Heath, 1912. - 77 -APPENDIX A (1) My limbs are related, as one gives them meaning They w i l l disclose i t , when the meanings join the throng [i.e., come togetherJ. (2) It i s unlike with us. A wolf i s on one island, I on the other. The island i s wholly surrounded by fen; Fierce men ,[that i s CeneJ are here on this island, They w i l l make i t \u00C2\u00A3the sense of the riddle] clear when the meanings join i n one. (3). It i s unlike with us. I give myself far-wandering longings toward my Wolf. When i t was wet weather and I sat weeping, Then the brisk warrior embraced me with his arms; That was bl i s s to me, but also i t was pain. Wolf, my Wolf, my longings toward thee Have brought me sickness, thy seldom coming The mourning mood, not want of meat. Hearest thou? Eadwaccer, the whelp of us both, Carries a wolf to the wood. (li) It i s easy to separate what never was joined, Our song together. 1 1 Morley, p. 218-219. - 78 -APPENDIX B Tupper divides the riddles showing borrowings from Latin originals into three groups. 1 The six riddles of the f i r s t group show strong debt to the Latin, he says. Of these, four\u00E2\u0080\u0094K-D 1*7, 60, 85 and 86\u00E2\u0080\u0094 show close use of Symphosius (though 85 and 86 employ only one line of Symphosius's three), and two\u00E2\u0080\u009435 and 1*0\u00E2\u0080\u0094are v i r t u a l l y line-by-line translations of Aldhelm (two lines of Old English for one of Latin). The eleven riddles of Tupper*s second group \"employ motives of Symphosius and Aldhelm i n such fashion as to suggest direct borrowing from the Latin enigmas.\" These are K-D 9, 16, 37 and 65, which show some debt to Symphosius, and 12, 26, 38, 1*9, 51, 63 arid 81*(?) , which bear some 3 resemblance to Aldhelm. The f i n a l group comprises riddles which demonstrate slight links with the Latin, \"so slight indeed that the likeness may often be accidental, or else produced by identity of topic.\" These are K-D 1-3, 5, 8, 10, 11, 20, 27, 28, 3l*, 1*8, 53, 56, 57, 59, 71, 73, 83, 91. Wyatt^, agrees, but for a few riddles, with Tupper's study of debt to Latin sources, but l i s t s a further four, K-D 30, 58, 66 and 87, which he says contain \"clear borro*ri.ng\" from the Latin, and he sees more \"clear borrowing\" i n 1*8, which Tupper had seen as only doubtfully 1 REB, pp. x l - x l i v . 2 \"But that long poem during i t s longer part declares i t s independence of Latin sources,\" REB, p. x l i i . 1* 3 1*9 shows debt only i f i t i s solved as 'bookcase.' pp. xxxx-xxx. -79 -connected. Trautmann, i n his study of the sources i n 191U , saw definite use of Latin originals i n 18 of the riddles (8, 12, 13, Hi, 16, 26, 35, 37, 38, UO, U7, 55, 60, 65, 83, 85, 86, 9 3 ) 6 , and traces of the Latin \ detectable i n five more (5, 31, Ul, U9, 50) . Unlike Tupper and Wyatt, however, Trautmann did fe e l Eusebius, and to a lesser extent Tatwine, to have had some influence on the Exeter riddles. Thus LU and 93 of his f i r s t l i s t (both influenced by Eusebius 30) do not appear i n either Tupper's or Wyatt's l i s t s . Riddle 13, i n which Trautmann finds the influence of Eusebius 38, Tupper l i s t s with the riddles from the popular tradition. Three of Trautmann's second l i s t d i f f e r from Tupper's and Wyatt*s classifications. In two of these, 31 and Ul, Trautmann follows Dietrich i n noting a pa r a l l e l to Aldhelm. 12. Two things are certain, Trautmann concludes: \"1, for the majority of the Exeter riddles Latin models can not be proven, and 2, even where the Anglo-Saxon poet borrows, he does i t , except i n the cases of 33 [K-D 35] Mailcoat and 38 [K-D U0] Creation, with the greatest individuality: he paints over, reconstructs, reasons differently and 7 wins from the objects new significance.\" * \"Quellen der Altenglischen Ratsel,\" Anglia 38(19lU), 3U9-35U. 6 86 \" i f i t really means one-eyed garlic seller.\" ^ loc. c i t . , p. 352. - 80 -APPENDIX C OPENING AND CLOSING FORMULAS Opening Formulas (see Table II) D. ic seah: 13 19 31 ...ic seah sellic ping 32 ...sipum sellic ic seah U2 51 52 53 55 59 6h 87 E. ic wiht geseah: 29 (and variants) 3k 36 37 Ic pa wihte geseah 38 Ic pa wiht geseah 56 Ic waes baer inne baer ic ane geseah winnende wiht 68 Ic pa wiht geseah 75 Ic swiftne geseah 76 Ic ane geseah F. ic eom wunderlicu wiht: 18 ~ 20 23 Ic eom wraetlic wiht 2k 25 G. icwat: k3 U9 58 H. ic gefraegn: U5 Ic on wincle gefraegn U8 67 Ic on binge gefraegn (damaged, but Wyatt, Tupper, Mackie and Krapp and Dobbie a l l agree on the restoration) - 81 -I. wiga i s on eorban wundrum acenned: 5o 8U An wiht i s on eorban wundrum acenned J. wiht i s wraetlic: 70 . 82 wiht i s [...] (frag.) K. Is bes middangeard missenlicum wisum gewlitegad wraettum gefraetwad: 31 32 This formula i s prefixed only to these riddles, which both also contain another opening formula of the i c seah variety. The repetition i n these consecutive riddles may be only the result of a mistake. Closing Formulas li. saga hwaet i c hatte (and variants): 1 saga hwa mec pecce oppe hu i c hatte 3 8 10 12 19 23 36 saga hwaet hio waere 39 g i f bu maege reselan recene gesecgan sobum wordum saga hwaet hio hatte 62 66 73 80 83 86 M. hwaet seo wiht sy: 28\" \" micel i s to hycganne wisfaestum menn hwaet seo wiht sy 31 micel i s to hycgenne wisum woSboran hwaet sio wiht sie 32 rece g i f bu cunne wis worda gleaw hwaet sio wiht sie 35 saga so&cwidum searoboncum gleaw wordum wisfaest hwaet bis gewaede sy - 82 -36 bu wast g i f bu const to gesecganne baet we so6 witan hu baere wihte wise gonge 1*1 baet i s to gebencanne beode gehwylcum wisfaestum werura hwaet seo wiht sy 67 secge se be cunne wisfaestra hwylc hwaet seo wiht sy N. frige hwaet i c hatte: Ik 16 26 27 0. raed hwaet i c maene: 61 (This may be a variant of L or N, but more l i k e l y i t i s a single example here of a further type. cf. \"Solomon and Saturn,\" 237, saga hwaet i c maene. DIFFERENCES BETWEEN TUPPER*S LISTS (REB, p. lxv) AND THE PRESENT STUDY Opening Formulas 3 13 I think Ic seah i n this riddle i s an opening formula (cf. 31 , 323). 33 Wiht cwom aefter wege wraetlicu lipan: Though this has something i n common with the wiht i s wraetlic openings, the theme of the riddle i s worked i n , so I f e e l i t should not be called properly formulaic. U7 me paet puhte: By the economy of words this may be formulaic, but i t appears nowhere else i n the riddles as an opening device. Compare, however, 873 micel me puhte, where i t i s simply part of the text. 82 Though fragmentary, this was probably of the wiht i s wraetlic type. 89 Tupper l i s t s this riddle as having no opening formula, but the test i s not applicable, as l i t t l e remains of the opening. Closing Formulas 20, UO Tupper mistakenly l i s t s these i n his 'no closing' l i s t . - 83 -i s obviously incomplete, and he elsewhere (REB, pp. xcvi, 16) feels 20 i s also incomplete. U3, 55, 59 Though many of the traditional terms are used, the handling of these closings i s far too s k i l l f u l for them to be properly called formulaic. Note the common use i n i;2, 55 of Nu to lead into the closing, and that a reference back to the subject of the riddle i s worked i n i n a l l four cases. nemnaS hy sylfe: Though i t certainly has a formula-like economy of expression, this ending, i f i t i s translated 'They name themselves' (Baum, p. 21), could not have been a very portable traditional phrase. "@en . "Thesis/Dissertation"@en . "10.14288/1.0093791"@en . "eng"@en . "English"@en . "Vancouver : University of British Columbia Library"@en . "University of British Columbia"@en . "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en . "Graduate"@en . "Authorship and unity in the Exeter Book riddles"@en . "Text"@en . "http://hdl.handle.net/2429/19945"@en .