¹ Arnould, E. J., Le Manuel des Péchés. Étude de littérature religieuse anglo-normande (xiii ^me siècle) (Paris 1940). At this writing the edition is immured in Paris; I am working from one of Doctor Arnould's two personal copies kindly placed at my disposal. I have published a general comment upon the book, Speculum 20 (1945) 99–103.Google Scholar

² Perhaps the chronology of our studies should be established. At the instance of Prof. W. W. Lawrence, I was attempting a literary reëstimate of Robert Mannyng of Brunne's Handlyng Synne, but found that its Vorlage, the Manuel, was so inadequately represented in the printed versions that one could not speak with any confidence concerning Mannyng's contribution. At that time, no Romance scholar since Gaston Paris had evinced interest in the Anglo-Norman version; accordingly, with the aid of a fellowship from Columbia University, I spent the year 1930–1931 examining the manuscripts of the Manuel. Personal difficulties prevented publication, and in 1935 I learned, through the good offices of Miss Hope Emily Allen, that Doctor Arnould had become interested in the poem with a view to an edition. My dissertation describing the Manuel was finished in 1939 and is on deposit in the Stanford University Library, dated March 1940, under the title, ‘The Source of Robert Mannyng of Brunne's Handlyng Synne, A Study of the extant Manuscripts of the Anglo-Norman Manuel des Pechiez.’ Later in the same spring Doctor Arnould's dissertation was published, but the edition, except for the author's personal copies, remained in occupied France; no copy reached this country until October, 1944. Meanwhile, I have published some of my findings but have reserved the discussion of the text until I could coördinate it with Doctor Arnould's conclusions; cf. n. 18, infra .Google Scholar

³ Furnivall, Frederick J., Robert of Brunne's Handlyng Synne with the French Treatise on which it is founded, le Manuel des Pechiez by William of Wadington. The eiitor used MS Harley 273 through line 10,330, at which point the hand changes; thereafter he used Harley 4657. No manuscript of the Manuel has been printed entire.Google Scholar

⁴ Furnivall, Frederick J., Robert of Brunne's ‘Handlyng Synne’, A.D. 1303, with those Parts of the Anglo-French Treatise on which it was founded, William of Wadington's ‘Manuel des Pechiez’ (Early English Text Society 119, 121; 1901–1903). I shall cite this edition for the portions of the Manuel it reprints, since the publications of the Roxburghe Club are but sparsely available.Google Scholar

⁵ Essais historiques sur les bardes et les jongleurs et les trouvères normands et anglo-normands (Caen 1854) III, 225–233.Google Scholar

⁶ For a list of manuscripts of the Manuel and their sigla, cf. n. 21 infra. Google Scholar

⁷ Paris, Gaston, ‘Wilham de Wadington,’ Histoire littéraire de la France, 28 (1881) 179–207.Google Scholar

⁸ Ibid. 192.Google Scholar

⁹ Romania 8 (1879) 332–334; 15 (1886) 312–313, 348–349, 351; 29 (1900) 47–53; 32 (1903) 145–147; 36 (1907) 111.Google Scholar

¹⁰ Romania 29 (1900) 5–21, 83–84.Google Scholar

¹¹ Catalogue of Romances in the Department of Manuscripts in the British Museum (London 1910) III, 272–303.Google Scholar

¹² The form seems to have become established through deference paid by Furnivall and Gaston Paris to Sir Frederic Madden, who had written, ‘We should certainly read Wadington, as confirmed by the reading of many excellent MSS. I have seen.’ Cf. Manuel, Roxburghe Club p. xviii; Paris, , op. cit. 180. Curiously, no manuscript now known, excellent or otherwise, has exactly this reading. Two approach it, B, from which Furnivall printed the Epilogue, and St. John's College, Cambridge MS 167, which is close to B. They have a in the first syllable; all others have i or y. Cf. section IV infra ; Arnould, , Manuel des Péchés 245–249, 433–435.Google Scholar

¹³ ‘An English Prose Version of the Manuel des Pechiez,’ Modern Philology 13 (1916) 167–168.Google Scholar

¹⁴ ‘The Mystical Lyrics of the Manuel des Pechiez,’ Romanic Review 9 (1918) 154–193.Google Scholar

¹⁵ ‘The Manuel des Pechiez and the Scholastic Prologue,’ Romanic Review 8 (1917) 434–462.Google Scholar

¹⁶ Note 1, supra. I here omit two earlier articles by Arnould, : ‘Un manuscrit partiel du Manuel des Péchés,’ Romania 63 (1937) 228–240 describes the Stonyhurst fragment; ‘On two Anglo-Norman Prologues,’ Modern Language Review 34 (1939) 248–251 suggests, among other matters, that the Graces of Shrift are a late addition. My dissertation was legally published, but not printed; cf. n. 2 supra. Google Scholar

¹⁷ Since this article was written, Robertson, D. W. Jr. has revised some of Arnould's discussion of the ancestry of the Manuel in ‘The Manuel des Péchés and an Episcopal Decree,’ Modern Language Notes 60 (1945) 439–447.Google Scholar

¹⁸ My description of the manuscripts contains less paleographical matter, somewhat more textual material; cf. ‘A Fourteenth Century Scribe,’ Modern Language Notes 55 (1940) 601–603; ‘Manuscripts of the Manuel des Pechiez,’ Stanford Studies in Language and Literature (1941) 99–123; ‘Palatinus Latinus 1970, A Composite Manuscript,’ Modern Language Review 38 (1943) 117–121.Google Scholar

¹⁹ Stanford University Bulletin: Abstracts of Dissertations. 1939–1940 (1940) 66–71.Google Scholar

²⁰ The Middle English prose version is included, since it is so slavish a translation that it can be used for some purposes as a manuscript of the original.Google Scholar

²¹ For descriptions of manuscripts, cf. Arnould, , Manuel des Péchés 359–398 and my article in Stanford Studies , 1 (1941) 99–123. These studies will be referred to hereafter as Arnould and Manuscripts respectively. The sigils employed are those in Manuscripts, since my descriptions contain more of the information with which we are concerned in this article. The following table provides a concordance of sigils: Google Scholar

²² Arnould 64; Roy, which lacks Books VI and IX, omits lines 31–34, adds a reference to Book VIII, and thus becomes the only copy of the Manuel with a table of contents which corresponds to the contents of the manuscript; cf. Manuscripts 102.Google Scholar

²³ p. 79.Google Scholar

²⁴ Scribes were zealous in providing cross references; cf. n. 31 infra. Google Scholar

²⁵ Arnould noticed, p. 63, that the passage is absent from HVY; it is lacking, also, from JP, and appears out of context in ArPhZ. Google Scholar

²⁸ This difference occurs also at the end of Book I; cf. n. 35 infra. Google Scholar

²⁷ Romania 15 (1886) 283.Google Scholar

²⁸ In M and Hm. For extracts, see Aitken, Marion Y. H., Etude sur Le Miroir ou Les Évangiles des Domnées de Robert de Gretham (Paris 1922). Suggestive, also, is the fact that the Manuel occurs frequently with the Château d'Amour, which must have appealed much more strongly to clerics than to the ‘lewed’. For another manuscript, cf. Laird, Charlton, ‘Five New Gretham Sermons and the Middle English Mirrur,’ Publications of the Modern Language Association 57 (1942) 628–637.Google Scholar

²⁹ On this basis it is possible to resolve the differences of previous writers on the Prologue. Miss Allen thought it influenced by the scholastic prologue, but Arnould has shown that Miss Allen's case rested too largely on spurious readings; cf. respectively Romanic Review 8 (1917) 434–462; Modern Language Review 34 (1939) 248–251. We may assume, without contravening the evidence of either critic, that the original author wrote a chatty, familiar prologue, which was influenced but not definitely shaped by his scholastic training. This prologue was not sufficiently formal to satisfy pedantically minded revisers, who provided the table of contents appropriate in a formal prologue, and useful in a reference work.Google Scholar

³⁰ I here differ from Arnould, who considers Books I–V of a piece; pp. 65–85, 192–205, and especially p. 104. As to my reasons for excepting Book I, of. Section IV infra. Google Scholar

³¹ Cross references in the Manuel are suspect, per se; more than half show textual disturbance to such a degree that their authenticity should be questioned, and many are unique. The scribes seem to have been librarians rather than creative artists, and they loved cross references. The following lines involving references have some currency, but show suspicious disturbance: 9303–08, 3117–18, 5283–84, 905, 9949–9950, 9441–42, 7957–7962, 7739, couplet in A after 7092, 7765–66, couplet in B after 9601, couplets in Roy after 7995 and 11,635. The order is roughly from lines rather well attested to unique lines. In linking Books II–V, I am relying upon lines so regular in their form and occurrence that they must be part of the original work, or very early additions. Books II and III are linked by 3111–16, 4791–92, 6465–67; Books III and V by 7912–16; Books IV and V by 7483–87, 7878–79; Books V and VII by 7763–64. The Prologue is linked to Book I by 125, to Book II by 2541–2, to Book III by 6562–64; Book I is otherwise referred to in 1107.Google Scholar

³² That Handlyng Synne, and by inference the Manvel, rests upon a strong tradition of penance manuals will be demonstrated historically by Robertson, D. W. Jr. in a forthcoming article in Speculum , ‘The Cultural Tradition of the Handlyng Synne;’ cf. also Arnould, 1–59.Google Scholar

³³ pp. 192–205. Arnould's discussion implies that Peraldus is the source for Books I–V, but it is notable that none of his quotations is taken from Book I. His whole discussion will be called somewhat into question by Robertson; cf. n. 32 supra. The argument does not, however, concern the discussion here, since both writers seemingly agree that the tradition in the Manuel goes back to Peraldus' Summae or to some penance manual.Google Scholar

³⁴ pp. 66–67.Google Scholar

³⁵ p. 68. He cites a variant common to ArHJYZ; to these should be added VPh; that is, the variant is common to the manuscripts which I call Group II, significantly the oldest manuscripts. This is a variant for lines 913–916; lines 903–912 also show considerable disturbance. It is perhaps significant, as in the Prologue, that many of the better manuscripts, ArEHJPhV, have something like ‘Ore metteray cum ay apris’ (copied from H); thus the author says he has learned about the Ten Commandments, not that he has promts them, presumably in the Prologue; cf. n. 26 supra. Google Scholar

³⁶ Lines 7957–7962 are a late addition which came in with Book VI.Google Scholar

³⁷ In a later study I shall suggest that Mannyng followed his manuscript more closely than has been previously supposed.Google Scholar

³⁸ Article X, however, contains Tale 2, against postponing confession.Google Scholar

³⁹ I should attach but little importance, for this discussion, to Arnould's suggestion that a treatise like Book I is to be expected in the Manuel. Even though ‘le Symbole [that is, the Creed] constitue le point de départ de la plupart des compendia catéchétiques des xiii^e et xiv^e siècles,’ I should doubt that ‘le chapitre sur le Symbole a sa place indiquée en tête d'un ouvrage de ce genre,’ assuming that the genre is that to which the Manuel belongs: a check of bibliographies like Wells and Vising seems to suggest that the Commandments, Sins, and Sacraments were much more popular subjects than the Creed; in any event, the works closest to the Manuel, the Summae of Peraldus and Frêre Lorens, subordinate the Creed. The episcopal decree cited by Arnould, , ‘qui pourrait presque servir de table ou d'introduction à notre Manuel’ clearly separates the Symbolum from the mandata, criminalia, and sacramenta; cf. Arnould 20, 67. But in any event the discussion is not very fruitful. Arnould's elucidation is revealing for the character of popularized doctrine, but not conclusive for the content of the Manuel. That there is precedent for a treatment of Faith is beyond question, but the validation of Book I as part of the original Manuel must rest upon direct evidence, not upon analogy.Google Scholar

⁴⁰ Two have been numbered. If one wishes to call the eighteen lines, 484–501, describing the Phoenix, a tale, the percentage of narrative lines to total lines in the book would be changed from 05 to 07 in the table. The author calls the passage an ensample; the usual designation of a tale in the Manuel is cunte. In the various manuscripts hundreds of illustrative passages are labelled ensample, or the equivalent.Google Scholar

⁴¹ I have estimated the tales in round numbers only; there might be disagreement as to where the narratives begin and end, but I have tried to be consistent. I have not counted the unique tales in A and Roy, the tale in BF, nor Tale 56 in ALR; I did count Tale 11, lacking from Group II.Google Scholar

⁴² Matrimony, for instance, is dispatched in thirty-six lines, and the author points out, lines 7913–16, that the subject is treated under lechery.Google Scholar

⁴³ This difference is observable in all manuscripts I have seen.Google Scholar

⁴⁴ Arnould, , who knows the poem intimately, would dissent; cf. p. 67.Google Scholar

⁴⁵ The contents of the book in the Roxburghe Club edition are confusing because of the manuscript Furnivall printed. Lines 447–535, printed as part of Article V, constitute Article VI. The next article, printed as VI is actually VII. Lines 608–627, printed as Article VII are unique in A and are omitted from this table. Beginning with Article VIII, Furnivall's numbering is correct.Google Scholar

⁴⁶ Article X, containing Tale 2, is an exception.Google Scholar

⁴⁷ Article II is an exception.Google Scholar

⁴⁸ There are only six deadly sins here; one doubts that the methodical author of Book^s II–V would have omitted Gluttony.Google Scholar

⁴⁹ Miss Allen noticed that lines 397–400 are identical with lines 12,526–29 of Book IX; cf. ‘The Mystical Lyrics of the Manuel des Pechiez,’ Romanic Review 9 (1918), 158, n. 12. Lines 397–402 are lacking from Group II, and are presumably an addition. We shall see later that Book IX is certainly composite, and that probably it was not compiled by the author of the main body of the Manuel; cf. section VIII infra. Google Scholar

⁵⁰ One tirade runs to twelve lines, but the rhymes do not follow the thought groups very sharply. I have not found the original of this poem, though there are a number of similar pieces in Anglo-Norman printed collections. The Deu le omnipotent is similar in ecstatic quality, in progression of ideas, and in the use of alas!—cf. Bibliotheca Normannica (Halle 1879) I, 103–105. For poems similar in form see Archiv für das Studium der neueren Sprachen und Literaturen 63 (1880) 59–61, and Tanquerey, F. J., Plaintes de la vierge en anglo-francais, xiii ^e et xiv ^e siècles (Paris 1921) 125–135.Google Scholar

⁵¹ The reliance which homilists reposed, during the early thirteenth century, on the Holy Ghost, seems to have been replaced by faith in rhetoric and the Virgin Mary as the century progressed; cf. Charland, Th. M., Artes praedicandi, contribution à l'histoire de la rhétorique au moyen âge (Publications de l'Institut d'Études Médiévales d'Ottawa 8, 1936) 111–126.Google Scholar

⁵² This tale resembles in content those which appear in Books II–V, coming, as it does, from Gregory. To me the manner of narration seems somewhat different; one misses, for instance, the conventional couplet which says, in effect, ‘this I will confirm with a story.’ But too much importance should not be attached to an isolated example.Google Scholar

⁵³ Romania 29 (1900) 5–21, 83–84.Google Scholar

⁵⁴ Arnould, 85–87; Manuscripts 116–118, 121; Modern Language Review 38 (1943) 117–121; Romania 29 (1900) 7–8.CrossRefGoogle Scholar

⁵⁵ Variants are moderately numerous in the independent poem. A check of them suggests that the version in the Manuel is related to Group I as described by Meyer, , Romania 29 (1900) 5–7; that is, it is like the version found in MS Rawlinscn Poetry 241 (R), not like that in MS Arundel 288 (Ar).Google Scholar

⁵⁶ This numbering is from the Early English Text Society edition; the other references in this column are from the text printed only for the Roxburghe Club.Google Scholar

⁵⁷ pp. 215–218.Google Scholar

⁵⁸ Lines 7957–7962 were certainly written for the Manuel. Google Scholar

⁵⁹ Arnould calls Ar ‘probablement le plus anciende nos mss.’ Cf.p.373. He would date Ar ‘du troisième quart … du xiii^e siècle;’ this is also the approximate date of the primitive Manuel, which had no Book VI.Google Scholar

⁶⁰ Meyer, , Romania 29 (1900) 5–6; Vising, Johan, Anglo-Norman Language & Literature (London 1923) No. 154; Tanquerey, F. J., L'évolution du verbe en anglo-français (Paris 1916) p. xvii; Arnould seems to accept the dating; cf. p. 218.Google Scholar

⁶¹ Noticing that lines 9951–58, 9963–68 of the Points are almost identical with lines 7934–47 of Book VI—called in the rubrics of several manuscripts a Sermun—and noticing also (p. 104) that line 9949 runs ‘Car, ausi cum en le sermun disei,’ Arnould suggested the possibility of the author of the Points being also the author of the Sermun, and, as we shall see, the Points seems to be part of the original Manuel. Were this cross-reference genuine, it would be important, but it is certainly spurious, absent as it is from ArBEFJPPhUY, that is, from all manuscripts except those in the late and revised Group I. The important manuscripts HZ are fragmentary at this point; V has changed its family to Group I.Google Scholar

⁶² I have used the following abbreviations: Lum, Lumere Indeficient Prologue; Gr, Graces and Vices of Shrift; P, Points of Shrift; P′, the Points of Shrift begun a second time; VIII, Book VIII; IX, Book IX; IX′, abbreviated form of Book IX; IX″, lyrics from Book IX; Epi, Epilogue; Epi′, an abbreviated form of the Epilogue. For the content of IX′ and IX″ see the discussion of Book IX, infra. Google Scholar

⁶³ No extant manuscript of the Manuel agrees with Mannyng in the order Lum-P-Gr. Mannyng certainly used a manuscript of Group III, which traditionally would have had the Points first and the Prologue illogically in the middle. Whether he or a scribe made the sensible alteration is a matter for conjecture.Google Scholar

⁶⁴ They do not appear in H, but were involved, presumably, in the loss of a gathering or so; we know only that in H Book VII ended with the Graces and Vices.Google Scholar

⁶⁵ Arnould, who presented evidence that part of the doctrinal material in the Points is derived from Sermon CX of Peraldus, suggests some parallels also between the Graces and the same sermon, and scrupulously notes that this seeming similarity of source damages to a degree his belief in a different origin for these two parts of Book VII; cf. pp. 226–235. Arnould's identification of material in the Points as having come from Sermon CX will be called into question shortly; cf. n. 32 supra; but even if this identification can be accepted, one might remark that whereas the similarities between Sermon CX and the Points is sharp and detailed, the likeness of the sermon to the Graces is general, except for material that both might have borrowed from the Vulgate.Google Scholar

⁶⁶ Tale 55. Tale 56, occurring only in ALR, is certainly late.Google Scholar

⁶⁷ They are so called in Handlyng Synne, but are not frequently given a general name, either in the rubrics or in the text of the Manuel. They are called choses against which checun se garde. Google Scholar

⁶⁸ See especially Table V, infra, and the subsequent discussion.Google Scholar

⁶⁹ Manuscripts 114–115.Google Scholar

⁷⁰ Arnould, 97; Manuscripts 106–107; pp. 301–302 infra. Google Scholar

⁷¹ Arnould has printed the significant passages, pp. 90–98. My analysis was essentially the same; cf. my unpublished dissertation (n. 2 supra) pp. 215–216, 229–230, 248–257.Google Scholar

⁷² Supra n. 22.Google Scholar

⁷³ The variants were recorded on the basis of the printed text. I endeavored to make the selection of variants as objective as possible, and for this tabulation defined a variant as follows: absence of one line or more, found in the printed text; presence of one line or more not found in the printed text; difference in order affecting one line or more; difference within a line or lines, sufficient to affect the sense materially. The tabulation is based on variants for E, which I chose because I happen to have a microfilm of the manuscript; most other manuscripts would have served, except, of course, that Book VI does not occur in Group II. In every manuscript there are unique variants; 1 have arbitrarily removed these, since they are significant principally for a study of the individual manuscript.Google Scholar

In addition to abbreviations conventionally recognized, I have used the following: ad. —additional line or lines; tr.—transposed; v.—variant reading; w.—wanting. In dealing with lines not in the printed text, and hence unnumbered, I have cited the last previous numbered line. I have also used an irregular form like the following: ‘4967–68 (1 for 2) ArHJZ.’ This entry is intended to mean that at this point Manuscripts Ar, H, J, and Z have only one line, not like either of those in the printed text.Google Scholar

The numbering of lines in Book VI is that of the Roxburghe edition. I have omitted the variant lines for 9597–9600, since one may be uncertain where this wandering quatrain belongs.Google Scholar

⁷⁴ H is fragmentary; the fragment Har offers variants here, and beginning with Book VI, V agrees with manuscripts of Group I.Google Scholar

⁷⁵ Romania 15 (1886) 348–349. For dissent from this judgment, cf. Manuscripts 105–106, Arnould 377–378.Google Scholar

⁷⁶ Concretely, I have endeavored to show that V was copied through the first five books from a manuscript of Group II, and thereafter from a manuscript of Group I. Cf. Manuscripts 116–118 and ‘Palatinus Latinus 1970, A Composite Manuscript,’ Modern Language Review 38 (1943), 117–121. EF, likewise, were finished from manuscripts of Group I; cf. pp. 301–303 infra. Google Scholar

⁷⁷ It also closes Har, but this fact probably is not significant, since Har is composed only of Books VII and VIII. Curiously, it appears in R, otherwise composed only of tales.Google Scholar

⁷⁸ At least one other person noticed this lack. In B, Book VIII closes with lines calculated to relate the book to sin and confession, as the lines apparently inserted into Book VI relate that work to sin. The lines occur also in PPhY, but in those manuscripts become part of the Epilogue, occurring after line 12,753; they were printed by Furnivall, , p. 404 Rox.Google Scholar

⁷⁹ Herbert, , Catalogue of Romances III, 278–284; Arnould, 113–192. The only other source in which more than three tales occur is Jacques de Vitry, from whom, directly or indirectly, five tales seem to have been drawn.Google Scholar

⁸⁰ For the basis on which variants were taken, and for abbreviations employed, cf. n. 72 supra. Google Scholar

⁸¹ Furnivall's shift from A to B as the basis of his text would make some difference in the tabulation, but not enough to be significant.Google Scholar

⁸² Arnould clearly considers the book late, but offers little evidence against its authenticity; cf. pp.99, 105.Google Scholar

⁸³ Romanic Review 9 (1918) 154–193.Google Scholar

⁸⁴ It should be recalled that Hf and Z are fragmentary, and that V has become a manuscript of Group I, so that these manuscripts offer no testimony for Group II in the later books; cf. n. 76 supra .Google Scholar

⁸⁵ For Book IX the A-variants are generally to be preferred; they would add a few lines to the total.Google Scholar

⁸⁶ Arnould noticed the correlation, but pressed his investigation no further; cf. p. 100.Google Scholar

⁸⁷ If one assumes that each tirade must be introduced by a phrase of direct address, there are a few more irregularities. In this connection, the most aberrant passage in the poem consists of lines 12,515–12,550. Of three tirades which constitute a unit on Mary's passion at the Cross, the first is introduced with Chaitiff, the other two not introduced.—Some matter may have been inserted, of course; occasional manuscripts lack occasional tirades. Details will be found in my unpublished dissertation, available from Stanford University, especially, pp. 277–282. Personally, I doubt that these seeming irregularities should be so considered. Lesser irregularities in the lyric disappear, in most instances, if one accepts the A-variant.Google Scholar

⁸⁸ Phillipps 8336. I have seen no notice of the sale of this manuscript. It may still be impounded with others at Cheltenham, . Romanic Review 9 (1919) 157; Romania 13 (1884), 513; other references seem to rely upon this last by Meyer: Naetebus, Gotthold, Die nichtlyrischen Strophenformen des Altfranzösischen. Ein Verzeichnis (Leipzig 1891) III, 1; Långfors, Arthur, Les incipit des poèmes français antérieurs au xv ^e siècle (Paris 1917) I, 101; Vising, , Anglo-Norman, No. 160.Google Scholar

⁸⁹ I have not been able to determine the length of the Phillipps lyric with much confidence. Meyer supposed that the poem was unique in Phillipps 8336; Naetebus, Långfors, and Vising mention no other manuscripts. Meyer did not print the end of the poem, and since there are several hands in the manuscript one cannot confidently estimate the lines to a page. So far as I may guess on the analogy of adjacent pages, however, the poem should contain something more than 300 lines, that is, about as many as the lyric to Mary in the Manuel without the introduction.Google Scholar

⁹⁰ Romanic Review 9 (1918) 154–157.Google Scholar

⁹¹ For other translations of the Latin hymn into Anglo-Norman, see Meyer, P., ‘Mélanges de poésie anglo-normande,’ Romania 4 (1875) 371–372; Harrison Thomson, S., ‘The Dulcis Jesu Memoria in Anglo-Norman and Middle French,’ Medium Aevum 11 (1942) 68–72. For the Latin version, cf. Etienne Gilson,‘Sur le Iesu Dulcis Memoria,’ Speculum 3 (1928)322–334.CrossRefGoogle Scholar

⁹² Vising, op. cit. No. 191; Patterson, F. A., The Middle English Penitential Lyric (New York 1911) 173; Meyer, P., ‘Notice du MS. Bcdley 57,’ Romania 35 (1906) 575; Arnould 238.Google Scholar

⁹³ Beginning with the address to the soul at line 12,135, no manuscript follows B; some are close to A. The principal divergences are represented in the EFHJPh group. In F the sequence is as follows: two lines not printed (fol. 141a, col. 2),Google Scholar

Google Scholar

lines 12,147–48 as they appear in the A-variant; lines 12, 148–52 B-variant; followed by ten lines I have not identified, as follows (fol. 141a, col. 2):Google Scholar

Google Scholar

Lines 12,135–46, B-variant; line 12,153, A-variant; lines 12,153–57. EHJPh have been in essential agreement. At this point, HJPh add the closing prayer, lines 12,321–32. E and F have one unique line each (E—La ioye de cel auerunt talent; F—Yl les receit mout dulcement) and go on with Book VII, that is, with the Lumere Prologue. Both of them, however, begin too early, in Book VI; F contains lines 8437–78; E has 8437–40, 8467–78. When I described E and F, I had not yet identified these lines from Book VI; cf. Manuscripts 106, 108.Google Scholar

⁹⁴ There are couplets and monorhyme passages of varying length. Three of the virtues are introduced in tirades, and by analogy with the remainder of the prayer one would expect tirades developed with couplets. This may be the situation, obscured by confusion in the text, which here has rather numerous variants. For instance, for the confusion after line 12,234, the manuscripts make clear that the next two lines should be transposed, and the last four lines of the A-variant inserted; but not always is the evidence clear.—Arnould, p. 238, calls attention to analogous matter in the Speculum Ecclesiae. Google Scholar

⁹⁵ Romanic Review 9 (1918) 158; cf. lines 12,305ff., 12,410ff., 12,606ff.Google Scholar

⁹⁶ Allen, Miss, Romanic Review 9 (1918) 151–193, argued for a single author, and with the evidence at her disposal her conclusions were plausible. It is now apparent, however, that the similarities upon which she relied characterized only the body of the lyric to Mary and tirades which comprise a portion of the lyric to Christ.Google Scholar

⁹⁷ It is at least curious that the compiler of Book I seems to have had a fondness for couplets, also. Cf. 265–267 supra. Google Scholar

⁹⁸ Romanic Review 8 (1917) 437.Google Scholar

⁹⁹ pp. 433–435. For discussion, see pp. 102–103, 245–253. In Ph, which Arnould did not see, the closing lines are as follows: Google Scholar

Google Scholar

¹⁰⁰ Modern Language Review 38 (1943) 117–121; cf. note 76 supra. CrossRefGoogle Scholar

¹⁰¹ E and Y lack the lines, but we should probably assume that Y once had them, since the manuscript has lost its final leaf, and P, to which Y is very close, has the lines.Google Scholar

¹⁰² They are lacking from the important manuscripts ArRoy. Google Scholar

¹⁰³ pp. 301–303 infra. Google Scholar

¹⁰⁴ pp. 245–249. One might add to his list of forms, Widdindune from Ph. There was also a William de Wygeton (cf. the Wygetoun of Hm) cf. Sixth Report of Historical Manuscripts Commission, Appendix, 565a-566a.Google Scholar

¹⁰⁵ Note 11 supra. Google Scholar

¹⁰⁶ Arnould, , Appendix I, and especially pp. 105–106; Manuscripts, especially pp. 120–122.Google Scholar

Perhaps one should examine Arnould's, graphique. It is intended to show the development of the Manuel, and as such it is suggestive of at least one way in which the mingling of manuscript traditions may have taken place. As Arnould points cut, any set of manuscripts so complicated as these cannot be represented in a simple stemma without deceptive simplification and he has wisely refrained. There are, however, some details which warrant individual comment, and the whole is subject to misinterpretation if it is accepted hastily. I shall use my sigils, followed by Arnould's when his differ from mine.Google Scholar

The derivatives of Arnould's or are very close to my Group I, including as they do AC (H)Hm(N)L(Z)O(G)R. Presumably they would include G(S), except that this fragment has been omitted from the graphique. One should remember, however, that Roy (D) has been revised so that it resembles manuscripts like A in many of its readings, that V(O) follows Group I after Bock V, that Hm(N) stands in some relationship to manuscripts in Groups II and III, that O(G) seems to preserve an older and better text than its bracketing with C(H) and Hm(N) would imply, and that AL(Z)R stand in some close relationship; they share Tale 56, for instance. In general, however, we agree as to the constitution of Group I; the only inclusion surprising to me is Z(T), to which I shall return. Meanwhile, one should note an important typographical error; in the fifth line under the diagram we should surely read ‘la version de G (not C) dont le scribe’ etc.Google Scholar

Arnould derives ‘la majorité de ncs manuscrits’ from a contamination ‘entre E et L (alors complet) ou A.’ Presumably he here refers to manuscripts anterior to Ar(E) and A, since each of these manuscripts has so many unique readings that it can scarcely have a direct descendant. The most surprising inclusion is that of Z(T): its line readings relate it closely to manuscripts of Group II, as Arnould recognizes in part in his description of the manuscript; that it could have combined with Ar(E) to provide the late books of the Manuel must be pure conjecture, since the manuscript breaks off with line 6700. Since the line readings of Z(T) relate it to Ar(E)H(C)J(W)PhV(O), this seems an improbable conjecture. The text of Z(T) is not close to L(Z), nor close enough to A so that one would expect the manuscript to have been close to the hypothetical A ¹ .Google Scholar

With Arnould's supplementary statement of the relationships of manuscripts within what I have called Groups II and III, I am in general agreement, as will be apparent from my descriptions of these manuscripts, but the diagram leads readily to misconceptions; we should be entitled from the diagram to suppose that Y(M) comes from F(U)O(V), whereas in all the manuscripts of the Manuel there are probably no others so close as Y(M) and P(L); they agree, even to the rubrics, and not merely ‘à partir de VII,’ but throughout. For the rest, I should add that Ph, the early part of Hf(Q), V(O), and Z(T) are also close to H(C)J(W). U(F) and P(L)Y(M) agree more generally with B and with E(I)F(K) than is here indicated in their earlier portions, and E(I)F(K) with H(C)J(W)Ph in their later books. All this confusion, coupled with direct evidence of contamination which we have noticed repeatedly, especially in V(O) and in E(I)F(K), seems to me to require more than ‘une contamination’ by way of explanation. An orderly hypothesis of this sort has its value in promoting understanding, provided one remembers that it does not represent very clearly the ‘correcting’ and cross-copying that must actually have taken place.Google Scholar

¹⁰⁷ S was identified by Arnould after I was in England; M could not be located when I was at Birdsall House; cf. Arnould, 393–398; Manuscripts 105, 120.Google Scholar

¹⁰⁸ One should remember, also, that Z breaks off at line 6700 at the end of a quire; if this early manuscript approached the primitive Manuel in its content, not much was lost. H has a lacuna, cf. n. 63 supra; 1 have suggested elsewhere that H probably did not contain Book VI; cf. Manuscripts 103 n. 20.Google Scholar

¹⁰⁹ EFV offer the most conclusive examples of this phenomenon; cf. 301–303 infra; Modern Language Review 38 (1943) 117–121. What with revision and addition, most manuscripts of Groups II and III, and some in Group I, have a various ancestry.Google Scholar

¹¹⁰ For the bases on which variants were taken, cf. n. 72 supra. Google Scholar

¹¹¹ A tabulation of variants from a manuscript of Group III, for the later books, will be found in Tables III and IV supra. Google Scholar

¹¹² Some omissions are not significant. We have noted above that Har, Hf, and Z are fragmentary, and that I have no variants for M and S. My time with Ph was very restricted, so that the absence of that manuscript from a list of variants is not necessarily significant. Had Roy not been extensively revised, it certainly would show greater adherence to the characteristics of the group; I believe that the same statement would be true of Hm, although here the revision has further obscured what I believe to have been the original character of the manuscript.Google Scholar

¹¹³ The best evidence that material of this sort was added to the manuscripts, not omitted from them, is its occurrence as unique readings and the infrequency with which such material is involved in unique omissions. By its nature, this evidence is extensive, too extensive to be reproduced here. It will be found in the appendix of my dissertation; cf. n. 2 supra. Google Scholar

¹¹⁴ Matter lacking from Groups II and III, presumably added to Group I, will be found to fall into the same categories; cf. Table II supra.Google Scholar

¹¹⁵ The rhyme is correct, since il appears at the end of line 3888.Google Scholar

¹¹⁶ The rhyme is not disturbed, since line 4340 appears with a reading like the following from Ar: En Ivr peigne part avera. Google Scholar

¹¹⁷ On the other hand, tales occur at the end of Ph, as though they might once have been omitted.Google Scholar

¹¹⁸ There are odd numbers of lines where the text seems not to be corrupt; Furnivall supplied lines by translating the Handlyng Synne into Old French but the manuscripts do not support his conjectures.Google Scholar

¹¹⁹ That there are three manuscripts from Group II which contain the obvious blunder of mari for femme in line 1607 may be some evidence against the antiquity of the group. The error, however, does not follow group lines, but is found in scattered manuscripts in all groups. Presumably we should recognize here the familiar textual principle that the most conspicuous and obvious errors are the least valuable for purposes of filiation; cf. Manly, John M. and Rickert, Edith, The Text of the Canterbury Tales studied on the basis of all known manuscripts (Chicago 1940) II, 14–18, and the references cited there. In any event, argument from blunder is not entirely reliable, for the author of the blunder may not be obvious. It has been argued, for instance, that the version of Book VII in A is original because of supposed blunders in B, but we now know that the apparent correctness of A is due to late revision. Similarly, the slip of the pen in line 1607 looks like one which any author could commit and any scribe correct. The scattering occurrence of the blunder suggests an early slip, corrected perhaps several times. For a method of dealing with contamination, see especially Burke Severs, J., ‘Quentin's Theory of Textual Criticism,’ English Institute Annual, 1941 (1942) 83–93.Google Scholar

¹²⁰ Ph is the only clear exception; H may have contained the book. Book VI appears in V, but only after V has changed its family; cf. the discussion of Book VI (sec. V supra).Google Scholar

¹²¹ Hm is especially interesting. It is a late manuscript, copied from an exemplar of Group I; several hundreds of lines have been erased entirely or in large part, and now read more like Group III; on some leaves one line in every three or four has been tampered with. This revision has taken place in Hm itself; a tabulation of variants suggests that Hm stems from a manuscript which had been even more extensively revised, and which preserved archaic readings, apparently from a manuscript of Group II. Thus the manuscript, which must now be classified in Group I, shows characteristics of both the other groups.Google Scholar

¹²² pp. 245–256.Google Scholar

¹²³ Ascription of the Manuel to Grosseteste probably need not be taken seriously; there would be chronological difficulties, even if one could believe that the author of Le château d'amour would write such verse as that in Books II–V of the Manuel. Professor S. Harrison Thomson, in a masterly study of the Grosseteste canon, rightly rejects the ascription; cf. The Writings of Robert Grosseteste Bishop of Lincoln 1235–1253 (Cambridge 1940) 233. Arnould concurs, pp. 249–251.Google Scholar

¹²⁴ Hist. litt. 28, 192.Google Scholar

¹²⁵ Unfortunately, Arnould has not distinguished in his discussion those portions of the Manuel which presumably were later additions. For instance, lines 5411–14, including the famous mention of wasseil and dringhail is lacking from ArHJVZ, that is, from Group II, and is presumably a late addition. See Table V supra and the subsequent discussion; Arnould 269.Google Scholar

¹²⁶ He sets the limits 1250–1270; cf. p. 253.Google Scholar

¹²⁷ p. 373.Google Scholar

¹²⁸ Romanic Review 9 (1918) 161–193.Google Scholar

¹²⁹ Previously, the manuscript had been dated 1307 on the basis of an included scrap, but this piece is in a different hand; cf. Kjellman, Hilding, La deuxième collection anglo-normande des miracles de la Sainte Vierge et son original latin avec les miracles correspondants des mss.fr. 375 et 818 de la Bibliothèque nationale (Paris 1922) pp. xvi–xvii. Matzke thought the end of the thirteenth century or the beginning of the fourteenth would be right; cf. Matzke, John Ernst, Les oeuvres de Simund de Freine publiées d'après tovs les manuscrits connus … Société des anciens textes français (Paris 1909) p. xii.Google Scholar