2.4 : Introduction
Part 1 of this chapter referred to two methods of classifying evidence, that is, as internal or external evidence (evidence which is, or is not, contained in the text), and as testimonial or circumstantial evidence respectively, statements about authorship and facts from which the likelihood of authorship may be inferred). Tables 2.5 and 2.6 present all the types of evidence used in the studies discussed, arranged in accordance with those two classifications. This Part identifies the factors affecting the value of each type of evidence and the extent to which its use in the studies reviewed has led to secure attributions of authorship.
Table 2.5 Non-stylistic criteria used in the studies reviewed in Part 1
Type Internal External
Criterion Used in Criterion Used in
Testimonial Statements of authorship Self-ascription within text (including epigraphic works) Self-ascription in manuscript522 Direct 3.1, 3.2 Direct Indirect 3.1, 3.2 Indirect 4.1 Other manuscript ascriptions 2.1, 3.1.3.2, 5.2 Contemporary or later attributions 3.1, 3.2, 4.1 Incorporation in other works securely attributable to the author 3.2 Circumstantial
Evidence derived from the manuscript
Manuscript history (date, style, place of origin) 4.1, 5.1a 5.2 Transmission history Associations 2.1, 5.1a History-based evidence General historical context 5.1a,b, 5.2 Biographical material 4.1, 5.1a, 5.2 Evidence of commissioning or patronage 4.1 Text indicating date,
place, circumstances of composition, and other cogent content
3.1, 4.1 Association between author and subject- matter of work
3.1, 3.2, 4.1
Table 2.6: Metrical, lexical and other stylistic criteria used in the studies reviewed in Part 1 Metrical criteria
Metrical form 2.1, 5.1a,b
Prosody 1.2, 3.1, 3.2 Metrical patterns 1.1, 1.2 2.1 Stress patterns 1.1, 5.1b 5.2 Caesura patterns 1.2 Elision and hiatus 1.1, 1.2
3.1, 3.2 Lexical localisation 1.2 Lexical and other stylistic criteria
Rhyme523 4.2
Formulae 1.1,1.2
Borrowings from and parallels with other sources 1.2, 3.1, 3.2, 4.1, 5.2 Alliteration 1.1, 1.2, 3.2 Vocabulary 3.1, 3.2 4.1 5.1a Grammar and syntax 1.2, 3.1
3.2
Spelling 4.1
2.5 : Statements of authorship
A statement that A is the author of a work may be made by A himself or by some other person (X) and may be direct (where A is named as the author) or indirect (where A is referred to in a way intended to identify him as the author). There are several ways in which such statements, whether within the text or external to it, may be unreliable or inconclusive.
Reduction of a work to written form is often a collaborative process, so much so that Love has warned that ‘collaborative authorship is so common, and so often disguised, as to constitute a central concern of attribution studies.’524 Collaborative authorship may take many forms and the appearance of the author’s name on the title-page or in an inscription may conceal the extent to which the style and content are due to the producer of the written work. However, that is 523 Orchard, Aldhelm, does contain some discussion of the use of rhyme in the chapter concerned with Aldhelm’s octosyllabic verses, at 39-42. The discussion of Orchard’s study in this work is confined to the characteristics of Aldhelm’s hexameter verse.
not an issue in any of the studies reviewed in Part I of this chapter and it is not considered further.
Statements of authorship apparently made by A himself may generally be regarded as more reliable than any other statements of authorship, but even in those cases, external evidence may raise the possibility that the statement was not in fact made by A but by some unknown person, for the purpose of giving more lasting currency or greater authority to the work than it would have had if published under his own name. The controversy over the authorship of the Pauline epistles provides a good illustration of this. The canon of the New Testament, as it appears in the King James Bible, contains fourteen epistles believed then, and by many scholars
subsequently,525 to be the work of S. Paul526. All except Hebrews begin with an assertion of his authorship.527 Several German scholars, the earliest of whom was F.C. Baur, doubted Paul’s authorship; Baur accepted only four,528 though later followers of Baur529 accepted three more as authentically Pauline. Morton530 describes the stylometric studies of sentence length in the Greek texts of the epistles carried out by Wake531 and adds some observations of his own on the frequency of occurrence of the words και and ει as first words, and γαρ and δε as second words, in sentences in the epistles. These results supported Baur’s conclusions and the controversy remains unresolved.
Table 2.5 shows that the studies by Schaller on Theodulf532 and Burghardt on Alcuin533 both rely on self-ascription in the text. In so far as these are statements embedded in the text and not formulaic statements at the beginning (as in the Pauline epistles) or end of the text, which would be relatively easy to tack on, they can be accepted as reliable. In that context it is interesting that Burghardt does not base his attribution to Alcuin of the large number of Schlussverse appended to Alcuin’s letters on self-ascription in the text but on the fact that they 525 See A.Q.Morton, ‘The Authorship of the Pauline Epistles’ in Literary Detection: how to prove authorship
and fraud in literature and documents, (New York: Charles Scribner’s, Sons, 1978), 165-83.
526 Romans, I and II Corinthians, Ephesians, Philippians, Colossians, I and II Thessalonians, I and II Timothy, Titus, Philemon and Hebrews.
527 In various forms, e.g., ‘Paul, a servant of God’; ‘Paul, an apostle of Jesus Christ’; ‘Paul, a prisoner of Jesus Christ’. In Philippians, Colossians, I and II Thessalonians and Philemon, others (Timotheus, Silvanus, Timothy) are named with him.
528 F.C. Baur, Paulus, der Apostel Jesu Christi, sein Leben und Wirken, seine Briefe und seine Lehre (Stuttgart, Becher and Müller, 1845), founder of the Tübingen school of theology. He accepted Romans, I and II, Corinthians and Galatians as Pauline.
529 E.g., A.B.C. Hilgenfeld, Histor-kritisches Einleitung in der Neue Testament (Leipzig: Fues’s Verlag, 1875). The further three are Philippians, I Thessalonians and Philemon.
530 A.Q. Morton, Literary Detection, 167-81.
531
W. C. Wake ‘The Authenticity of the Pauline Epistles: A Contribution from Statistical Analysis’,Hibbert Journal, 47 (1948-9), 50.
532 See Table 2.4, above, listing six instances of direct, and one of indirect self-ascription in the text. 533 Burghardt, Gedichten Alcuins, 13-15, and the Table at 31-33 therein. Burghardt finds ten instances of direct, and twenty-nine of indirect self-ascription; many of them are in the form of references to Flaccus, Alcuin’s nom de plume in Charlemagne’s court circle of poets.
are preserved with other works of Alcuin, namely, the letters themselves. Schaller534 and Burghardt535 have also both made firm attributions based on the author naming himself as the originator of the subject-matter of epigraphic poems, though Schaller qualifies the criterion by the requirement that it must be difficult (to believe) that anyone else can have been the author.
Lapidge’s study of Bede provides some instances of unreliable ascription by later
commentators and editors. He observes that one of the problems in dealing with mediaeval texts preserved only in modern printed editions is that it is impossible to control the
information which they transmit.536 As he says, ‘the distinction of Bede’s name attracted to it many hymns which could not-on stylistic grounds-be his’, and he notes a number of instances in which hymns attributed to Bede either in mediaeval texts or in modern editions exhibit hiatus (which Bede never used in any hymn securely attributable to him) or contain metrical and grammatical faults which are so uncharacteristic as to eliminate him from consideration as their author.537
Another reason for discounting an apparent self-ascription in the text is that it may be inconclusive or misleading. The studies by Lapidge538 and Herren539 of the quantitative poems attributed to Columbanus exemplify both these possibilities. They accept that, since
Columbanus and its variants is a common name for clerics of Irish origin, the occurrence of the acrostic COLUMBANUS HUNALDO in Ad Hunaldum and the line ‘dicta Columbani fida te voce monentis’ in Ad Sethum cannot shed any light on the authorship of those poems. However, a factor in Lapidge’s attribution of those poems and of Ad Fidolium to Columbanus of Saint-Trond is his supposed authorship of the planctus de obitu Karoli which commemorates the death of Charlemagne in 814 and contains the words ‘O Columbane, stringe tuas lacrimas’. Herren argues that the occurrence of that line may have caused Muratori to interpret the subscription ‘hymnus Columbani ad Andream episcopum de obitu Caroli’ of the poem in a seventeenth-century collection of the poems of Hrabanus Maurus540 as meaning that Columbanus of Saint-Trond was its author.541 In rejecting that attribution he adopts the argument of Löwe542 that the author of the poem addresses himself throughout in the first person in the refrain heu mihi misero, which is 534 Schaller, Theodulf,’ criterion Ic; 20: ‘ferner, für epigraphische Gedichte, in denen Theodulf als Urheber des
jeweils zugrundeliegenden Gegenstandes gennant ist und die schwerlich ein anderer als er selbst verfasst haben kann’. He makes ten such attributions.
535 Burghardt, Gedichten Alcuins, criterion III, 22; expressed in the unqualified words ‘A[lcuin] nennt sich
als Urheber der jeweils gemeinten Sache’. He makes four such attributions.
536 Lapidge, ‘Bede the poet’, Anglo-Latin Literature, 326. 537 Ibid., 330 and nn. 71, 72 thereto.
538 Lapidge, ‘Ad Fidolium’ and ‘Epilogue’.
539 Herren, ‘Quantitative poems attributed to Columbanus’, 99-112.
540 See the proemium to the poem at PLAC I, 434, in which Dümmler rejects its attribution to Columbanus; the poem is printed without any attribution.
541 Herren, ‘Quantitative poems attributed to Columbanus’, 101. 542 Löwe, ‘Columbanus und Fidolius,’ 3.
inconsistent with the address in the second person in the words quoted, and regards it as placing a ‘practically insuperable grammatical difficulty’ in the way of Lapidge’s attribution.
2.6 : Contextual and biographical evidence
Both types of evidence are potentially valuable in determining whether an author is a credible candidate for authorship of the work in question. Such evidence may locate the author in time or place, indicate associations between him and the subject-matter of the work, identify the influences on his writing and the genres in which he wrote, and throw light on his beliefs and attitudes. More specifically, there may be evidence of commissioning or patronage, or the text may contain relevant information bearing on the time, place or circumstances of composition.
The only studies reviewed in which general historical and biographical evidence has played a significant part are those of Lapidge and Herren, already discussed with reference to the value of self-ascription, on the quantitative poems Ad Fidolium, Ad Hunaldum and Ad Sethum. As discussed in section 2.3.8, the evidence has failed to distinguish between the candidates for authorship, though I observe one possible clue to their dating which does not appear to have had any significance for either Lapidge or Herren. The fourth of the six hexameter verses which form the conclusion to Ad Fidolium reads ‘nunc ad olympiades ter senae venimus annis’. That period of ‘thrice six Olympiads’ amounts to seventy-two years, and Raby543 reads the line as a statement that the author was seventy-two years old when he wrote the poem. If that is correct, and if Columbanus of Bobbio (ca 543-21 November 615) was the author, the poem would have been composed in the last year of his life. However, Ad Fidolium is transmitted in the
manuscript Berlin Staatsbibliothek Diez B. Sant. 66, which dates from 796 or shortly
afterwards.544 That militates against the attribution to Columbanus of Saint-Trond. If he was seventy-two years old in or around 796, and still alive in 814, as he must have been for the planctus de obitu Karoli to have been attributed to him, he would have lived well past the age of ninety. It seems unlikely that such longevity of a senior cleric would have gone unnoticed and unrecorded in the annals of the abbey or elsewhere.
Columbanus of Bobbio was a prolific writer of prose works, but the body of verse securely attributable to him is very small, and his ability to write adonics or any other form of
quantitative verse is not established. Columbanus of Saint-Trond, unless he was the author of the planctus de obitu Karoli, is not known to have written anything in either prose or verse. The debate regarding the sources on which Ad Fidolium draws and the availability of those sources 543 Raby, Christian Latin Poetry, 139.
544 L. Boyle, Mediaeval Latin Palaeography, where the MS is noted as ‘Caroline, shortly after AD 796, possibly from the court of Charlemagne at Aachen’.
to the candidate authors remains unresolved, but its resolution may well provide the best opportunity to achieve a secure attribution.
General historical and biographical information also has very little part to play in Freeman’s study of the authorship of Libri Carolini. 545 The work is, ostensibly, Charlemagne’s response to the Second Nicene Council of 787 and its restoration of images, but the only contenders for authorship are Alcuin and Theodulf. It is dated a ‘triennium’ after the Council, by which time Alcuin had returned to England from Charlemagne’s court; that, however, has never been regarded as evidence against his authorship. The chronology of Theodulf’s early life is not well established and it is not known when he arrived at Charlemagne’s court, though Andersson surmises that Charlemagne recruited him not so many years after he had recruited Alcuin.546 The earliest of Theodulf’s poems securely datable to his time at Charlemagne’s court is Ad
Carolum regem547 written in 796 to commemorate his victory over the Avars.
Historical information has proved more useful in the study of Theodulf’s poems. The three which are attributed to Theodulf on the basis of Schaller’s criterion IV (cogent content-related material) are all datable to a period when Theodulf is known to have been at Charlemagne’s court and relate to historical events.548 Quid faciunt cycni549 is a satire, whose composition Schaller dates to 798,550 in which the author mocks the court circle and, in particular, the Irishman Cadac-Andreas (the Corvinus addressed in the last line). The names of many of the persons satirised also appeared in Theodulf’s Ad Carolum regem,551 in a much more laudatory context. Dümmler dates Rex benedicte, vale552 to the year 800. Its subject is the attack on Pope Leo III in April 799 and Charlemagne’s support for him. Sumito quae misi laetus 553is a short poem to Fardulf, who was abbot of Saint-Denis 793-806, accompanying gifts to him which are not identified; Andersson554 suggests that they may be the other two parts of carm. xxxiii, qui iuvat ad tempus and grande habet initium.
545 A. Freeman, ‘Theodulf of Orleans and the Libri Carolini,’ 663-709. 546 T.M. Andersson, Theodulf of Orléans: The Verse, 3.
547 Ibid., 3, 65-66; Dümmler, PLAC I, carm. xxv, 483.
548 This classification is due to Andersson, Theodulf of Orléans: The Verse. The arrangement of poems in his edition follows that of Dümmler, PLAC I, Theodulfi carmina, 433-581.
549 PLAC I, carm. xxvii, 491. Dümmler identifies the ‘Getulian’ in line 64 as ‘Goth’ (i.e., Theodulf himself), for which, see n. 7, 492. Note the Visigothic spelling ‘cycni’ for ‘cygni’ to which Freeman drew attention.
550 D. Schaller, ‘Der junge “Rabe” am Hof Karls des Grossen’, in Festschrift Bernhard Bischoff am 65.
Geburtstag, ed. J. Autenreith and. F. Brunhölzl, (Stuutgart: Anton Hiersemann, 1971), 123-41.
Schaller doubts Dümmler’s identification and makes a case that ‘Getulian’ means ‘Moor’ that is, Maurus (Hrabanus Maurus, Alcuin’s pupil).
551 Dümmler, PLAC I, carm. xxv, 483, where the poem is dated to 796. 552 Ibid., carm. xxxii, 523.
553 Ibid., carm. xxxiii, 33, 524.
2.7 : Metrical, lexical and other stylistic evidence
At a general level, such evidence will provide an image of the authorial fingerprint, at any rate when the author is writing in a particular genre; that qualification is made because (for example) an author might choose to employ different metrical patterns, or different vocabulary, or draw on different sources, depending on the purpose and subject-matter of the work in question. However, it may be less valuable in making an attribution to a particular author, due to the difficulty of distinguishing in that way between authors who share a common cultural background and are influenced by the same exemplars. In that context, it is noteworthy that the studies by Schaller on Theodulf555 and Burghardt on Alcuin556 rely on similar features of
prosody and that the classical authors from whom the vocabulary of Theodore’s spiritual- didactic poems557 is drawn (namely Virgil, Ovid and Prudentius) were well known to Alcuin, as is apparent from the versus de sancto euboricensis ecclesiae, 558 which contains a catalogue of the authors whose works were in the library at York. The footnotes to Dümmler’s edition of
Alcuin’s poems show how often he drew upon classical authors, particularly Virgil and, to a far lesser extent, Ovid.
Orchard’s detailed study of the metrical, lexical and stylistic features of Aldhelm’s poetry provides a valuable illustration of the power and the limitations of the use of such evidence in attribution studies. In his analysis of the epitaphs to Bugga559 and to Archbishop Theodore,560 he is able to eliminate Aldhelm as author of the first by the occurrence of inaccurate scansion and grammatical solecisms entirely foreign to his style, and to make a strong case for his authorship of the second by the exactitude of the scansion, the metrical placing of words, the absence of elision, the licence, characteristic of Aldhelm, of lengthening the vowel a- before sp, the occurrence of phrases found in the Aldhelmian corpus (though six such phrases also occur in Bugga’s epitaph) and his fondness for alliteration (which also occurs five times in the fifteen lines of Bugga’s epitaph). Understandably, he did not analyse the occurrence of metrical patterns in those poems; that would have been impracticable, since Bugga’s epitaph consists of only fifteen verses and Theodore’s, thirty-four.
555 Schaller, ‘Theodulf,’ 38-40.
556 Burghardt, Gedichten Alcuins, 39-42, and, in the investigation of the poems not securely attributed, (ungesicherten Gedichte), 47-268, passim.
557 Dümmler, PLAC I, Theodulfi carmina, carm. i (contriti et cordis) 445, and carm. ii (quarto libelle, entitled
ad episcopos), 452; see Schaller, 33-38.
558 Ibid., Alcuini carmina, carm. i. (Christe deus, summi virtus), 169. The catalogue of authors begins at 203, v. 1540, with Jerome. Prudentius (Aurelius Clemens Prudentius) appears as Clemens in v.1551, and Virgil in v.1553. Ovid is absent from the catalogue but appears in the footnotes to the poem at n.4, 203 where illic invenies, v.1535 (just before the start of the catalogue) is identified as a borrowing from Ars
Amatoria I, 91. He appears again in n.4, 223, to carm. iv (cartula, perge cito), in which v.3 borrows the
phrase volvitur undis from Metamorphoses I, 570. 559 Orchard, Aldhelm, 243-48.
Finally, Freeman’s study of Libri Carolini merits discussion in this context as well as in that of general historical evidence. Although she states that ‘the customary procedures of determining authorship on the basis of style and content appear ineffectual here561’, that statement must be understood against the background that almost all of Theodulf’s extant work is in verse, not prose. Indeed, her attribution to Theodulf rests heavily on content in the sense that Libri
Carolini contains scriptural citations influenced by the Mozarabic liturgy562 and displays many
instances of Hispanic orthography563, including one spelling (cerubin for cherubim) which also occurs in an inscription in the church at Germigny-des-Pres, built and dedicated by Theodulf in 806564. These features of Libri Carolini make a very strong case for Theodulf’s authorship.
2.8 : Conclusion
All of the types of evidence employed in the studies reviewed in Part 1 of this chapter are available, and capable of contributing usefully towards an investigation of Paul’s possible authorship of the sixty-eight poems associated with him. Some of the poems include statements of authorship in the text, and there are also manuscript ascriptions. There is some historical and