^{page 355 note 2} It goes back at least as far as Rossbach, O., De Senecae Philosophi librorum recensione et emendatione (Breslau, 1888), p. 2.Google Scholar It is repeated, for instance, by Faider, P., Études sur Sénèque (Gand, 1921), p. 114;Google ScholarNothdurft, Klaus-Dieter, Studien zum Einfluss Senecas auf die Philosophic und Theologie des zwölften Jahrhunderts (Leiden-Cologne, 1963), p. 11;Google ScholarViansino, I., L. Annaei Senecae Dialogorum libri III–IV–V (Corpus Paravianum, Turin, 1963), p. xxii.Google Scholar

^{page 355 note 3} M. G. H., , script. 7, pp. 746–7:Google Scholar ‘non solum autem in aedificiis, verum etiam in libris describendis operam Desiderius dare permaximam studuit. Codices namque nonnullos in hoc loco describi praecepit, quorum nomina haec sunt: Augustinum contra Faustum … Ovidium Fastorum, Senecam.’

^{page 355 note 4} Lowe, E. A. (Loew), The Benemntan Script (Oxford, 1914), p. 71.Google Scholar

^{page 356 note 1} Op. cit., p. 341. There is a facsimile of fol. 52^v–53^r in Chatelain, É., Paléographie des classiques latins (Paris, 1884–1900), plate 167.Google Scholar

^{page 356 note 2} e.g. Lowe, , op. cit., p. 50 n. 5.Google Scholar

^{page 356 note 3} PL 147, 1306 D = const. 10. 2; cf. Manitius, M., Geschichte der lateinischen Literatur des Mittelalters, vol. ii (Munich, 1923), p. 487 n. 1.Google Scholar

^{page 356 note 4} Manitius, , op. cit., p. 489.Google Scholar

^{page 356 note 5} Nothdurft, , op. cit., p. 13 n. 6.Google Scholar

^{page 356 note 6} Quotations from the Dialogues occur, though more thinly, in other parts of the Vita, and also in his Vita Secundini.

^{page 357 note 1} For a more detailed discussion of the missing part of the de ira, see pp. 368–9.

^{page 357 note 2} 1306 C.

^{page 358 note 1} Becker, G., Catalogi bibliothecarum antiqui (Bonn, 1885), p. 246;Google Scholar 119, 12 Seneca de beneficiis I; 119, 18 Senecam magistri amici quern abemus pro alio Seneca monasterii in pignore I.

^{page 358 note 2} f. 2^r ‘Iste liber… … Cf. Lowe, p. 71.

^{page 358 note 3} f. 2^r ‘Antonii Franc. Neapolitae Caraccioli Siculi et amicorum Anno D. MDLXXXIII Kal. Novembr. Messanae.’

^{page 358 note 4} f. 2^r ‘Card. Federici Borrhomaei anno 1603’.

^{page 358 note 5} A note by the first librarian Olgiatus (f. i^v) tells us that it was already in the Ambrosiana by 1603.

^{page 358 note 6} This identification was disputed by Gertz, M. C. (Studia critica in L. Annaei Senecae Dialogos [Copenhagen, 1874], pp. 9–10), who found discrepancies between the readings of Muretus' codex Siculus and those of A. But inaccurate collations were not unknown in the sixteenth century, and one would be hard put to it to find another ‘Sicilian’ manuscript of the Dialogues. His liber Siculus of the Letters is almost certainly Ambrosianus C 85 inf., which also belonged to Caracciolo, though again the readings do not entirely tally.Google Scholar

^{page 358 note 7} ‘Testante philosopho, crudelitatem fati aequalitas consolatur’ (PL 207, 470 C) = ‘ut crudelitatem fati consolaretur aequalitas’ (Polyb. i. 4).

^{page 358 note 8} By Cohn, E. S., English Historical Review 41 (1926), p. 46.Google Scholar

^{page 359 note 1} PL 207, 469 D. It begins ‘vagientibus in cunis adhuc artis grammaticae, natisque recens discipulis, quilibet homo, qui sequitur veritatem, vitam bonam et exitum meliorem’.

^{page 359 note 1} Huillard-Bréholles, A., Vie et correspondance de Pierre de la Vigne (Paris, 1865), pp. 300–2.Google Scholar

^{page 359 note 3} What is known about Terrisio has been assembled by Torraca, F., ‘Maestro Terrisio di Atina’, Archivio storico per le province napoletane 36 (1911), pp. 231–53,Google Scholar reprinted in his Anedotti di storia letteraria napoletana (Città del Castello, 1925), pp. 33–59.Google Scholar Cf. also Haskins, C. H., Studies in Medieval Culture (Oxford, 1929), PP 129, 135.Google Scholar

^{page 359 note 4} For Bene, see Gaudenzi, A., ‘Sulle opere dei dettatori bolognesi’, Bulletino dell'Istituto storico italiano 14 (1895), 150 ff.;Google ScholarFrati, C., A proposito di maestro Bene, Rome, 1895.Google Scholar

^{page 359 note 5} MS. Bodley 633, edited by Delhaye, P. and Talbot, C. H., Analecta Mediaevalia Namurcensia, 5–6 (Namur-Lille, 1955–1956).Google Scholar

^{page 360 note 1} aequo animo sustinenda sunt imperitorum convitia, etc. (PL 205, 302 D) has nothing to do with the de constantia but, as has been noted by Nothdurft, (op. cit., p. 149),Google Scholar comes from letter 76. 4. Again, quid refert an garciones isti superius an inferius intonent? sicut in posteriori parte, sicfetunt et in ore (ibid.) was not inspired by any passages in the dialogi, as has generally been supposed, but surely by letter 91. 19: Demetrius noster solet dicere eodem loco sibi esse voces imperitorum quo ventre redditos crepitus. ‘Quid enim’ inquit ‘mea, susum isti an deosum sonent?’ Similarly 351 D–352 A is not a free adaptation of ideas from the de brevitate, as has been thought: it is the beginning of letter 101.

^{page 360 note 2} Edited by Holmberg, J., Uppsala, 1929. For the quotation, cf. p. 50. 1–3.Google Scholar

^{page 360 note 3} Edited by Sundby, Thor, Chaucer Society, Ser. ii, pt. 8, London, 1873.Google Scholar

^{page 360 note 4} The only other alleged quotation from the Dialogues in the works of Albertano which have so far been edited will not bear examination. Liber consol., p. 55. 3–4Google Scholarfacilius est vitia excludere quam admissa comprimere is not, as has been assumed, an echo of ira 1.7.2. facilius est excludere perniciosa quam regere, et non admittere quam admissa moderari, close though it is: it is a direct quotation from letter 85. 13 cum facilius sit excludere quam admissa comprimere.

^{page 360 note 5} It is found in the de ira of Marin of Braga: cum superiore contendere furiosum est, cum pari anceps, cum inferiore iam sordidum (Barlow, C. W., Martini Episcopi Bracarensis Opera Omnia [American Academy in Rome, 1950] P. 153. 35) but this is not the source of the two later quotations.Google Scholar

^{page 360 note 6} Opus tertium, edited by Brewer, J. S., Rolls Series (London, 1859), p. 56.Google Scholar

^{page 361 note 1} ‘(Libri) Senece, qui sunt optimi et rarissime inveniuntur’ (Opus tertium, frag. Duhem, p. 164);Google Scholar ‘protraxi hanc partem terciam Moralis philosophic gratis propter pulcritudinem et utilitatem sentenciarum moralium, et propter hoc quod libri raro inveniuntur’ (Opusmaius, p. 187. 1–3 Massa).

^{page 361 note 2} Part vii of the Opus maius, which has now been separately edited by Massa, E., Baconis Operis Maioris Pars Septima seu Moralis Philosophia, Zürich, 1953.Google Scholar

^{page 361 note 3} The three following are (1) the de brevitate vitae + the ad Polybium, (2) the de vita beata + the de otio, and (3) the de tranquillitate.

^{page 361 note 4} Massa, p. 133. 1–7.

^{page 361 note 5} ‘Roger Bacon and the ‘Dialogues’ of Seneca’, Manly Anniversary Studies in Language and Literature (Chicago, 1923), pp. 243–53.Google Scholar

^{page 361 note 6} No one has looked at Guaiferius in this connection, and the importance of the epitome of the de ira by Martin of Braga was not pointed out until 1937 (Barlow, C. W., TAPA xlviii [1937], pp. 26–42).Google Scholar

^{page 361 note 7} Quoted from a MS. in the Bodleian Library (Digby 65, f. 102) by Paetow, L.J. in The Morale Scolarium of John of Garland (Memoirs of the University of California, iv, Berkeley, 1927), p. 101 n. 87.Google Scholar

^{page 361 note 8} Paetow, p. 113.

^{page 362 note 1} Compendium studii philosophie (ed. Brewer, ), P. 453.Google Scholar

^{page 362 note 2} Balduinus ab Amsterdam (Collectanea Franciscana 32 [1962], 234 ff.)Google Scholar has drawn attention to a similar reworking of passages from the de ira in three works by Guibert, the de septem verbis, the de pace, and the de nomine lesu. In the de pace I have noticed an adaptation of the beginning of the de vita beata (De pace, ed. Longpré, E., Bibliotheca Franciscana Ascetica Medii Aevi VI [Quaracchi, 1925], p. 162), so that his borrowings are not confined to the de ira.Google Scholar

^{page 362 note 3} For the chronology I have mainly relied on Bonifacio, E., Gilberto de Tournai; De modo addiscendi (Turin, 1953), pp. 7 ff.Google Scholar

^{page 362 note 4} For his use of the Dialogues see Little, A. G., Studies in English Franciscan History, Publications of the University of Manchester, Historical Series, xxix (Manchester, 1917), p. 188,Google Scholar and in particular Pratt, Robert A. in Speculum, xli (1966), pp. 627 ff.Google Scholar For recent studies of John of Wales, see Smalley, Beryl, English Friars and Antiquity in the Early Fourteenth Century (Oxford, 1960), pp. 51–5;Google ScholarPantin, W. A., ‘John of Wales and Medieval Humanism’, in Medieval Studies presented to Aubrey Gwynn, S.J. (Dublin, 1961), pp. 297–319.Google Scholar

^{page 363 note 1} Op. cit., pp. 248 ff.

^{page 363 note 2} In particular those of Gertz (1886), Castiglioni (1946), and Viansino (1963).

^{page 363 note 3} Castiglioni, L., ‘De quibusdam deterioribus codicibus Senecae opuscula De Ira continentibus disputatio’, Athenaum i (1913), 98–111;Google ScholarMarouzeau, J., ‘Ce que valent les manuscrits des Dialogi de Sénèque’, RPh, xxxvii (1913), 47–52;Google ScholarWagenvoort, H., ‘De codice Senecae Angelico (MS. Lat. 1356)’, Mnem. 1913, 153–63;Google ScholarBourgery, A., ‘A propos des manuscrits du “De Ira”’, REL xi (1933), 369–78,Google ScholarFontan, A., Algunos códices de Séneca en bibliotecas españolas y su lugar en tradición de los diálogos', Emerita xvii (1949). 9–41 and 22 (1954) 35–65;Google Scholar B. L. Hijmans B. L. F. and Forder, M. P., ‘De xxxii codicibus recentioribus L. A. Senecae libellum De providentia continentibus’, Mnem. 1960, 39–62.Google Scholar The last marks an advance on the others; although I cannot accept their stemma or their conclusions, Hijmans and Forder give some useful information about a large number of manuscripts and their affiliations.

^{page 364 note 1} Laurentianus 76. 32, known usually as L, and containing the de ira, was at one time dated to the twelfth century, more recently to the thirteenth. It should be placed firmly in the second half of the fourteenth century. On the strength of its spurious seniority it was taken up by Gertz and eventually won a place in the Teubner text. It is a poor piece of work. Laurentianus 76. 38 (dated to the thirteenth century by Marouzeau and Viansino) is a perfectly decent fourteenth-century manuscript. Perugia 57, catalogued as thirteenth, belongs to the fifteenth century.

^{page 364 note 2} The sigla are a headache. As a sample of the confusion, Fickert's two Ambrosiani (B 2 sup. and C 293 inf.), which he called E and D respectively, became D and E in Hermes' Teubner text, and finally B and C in the Pravia editions; Laur. 76. 38, 1 to Marouzeau, is F3 to Viansino in his edition of the consolationes, while in his edition of the de ira F3 is Laur. 76. 32, known to Gertz and Hermes as L; P3 and P4 are usually—and also in Viansino's edition of the consolationes—Palatine manuscripts, but in his edition of the de ira they become Parisini (6379 and 6380), known to Hijmans and Forder as Z and β β being Fickert's designation for B, known to Viansino as Ber. As unused sigla are scarce (both the Greek and Latin alphabet have been exhausted), I see no alternative but to start afresh and to allot unsophisticated sigla to those few manuscripts which deserve them.

^{page 364 note 3} It remains a very important witness in the ad Polybium, most of which is missing in A.

^{page 364 note 4} The only other certain thirteenth-century manuscript I have noticed is in Rome, Biblioteca Angelica 505. It belongs to the latter part of the century, and contains the first two books of the de ira and part of the third. It was used by Barriera for his edition of the de ira (Paravia, 1919) and is a scrappy piece of work with a mixed text.Google Scholar

^{page 365 note 1} P has nothing after dial. 9. 15; Q_ has only a fragment of the de ira and omits the ad Helviam matrem completely. Of the three manuscripts used only C(B) is available for the ad Helviam, so that β readings for this dialogue can be reconstructed only by calling in other and later manuscripts.Google Scholar

^{page 366 note 1} Line omission on this scale and to the nearest letter of the line is as good evidence as we are likely to get. But there is other evidence for the dependence of β, not valid in itself, but corroborative. Glosses in A appear in the text of β, and are particularly informative where an intelligible gloss in A is garbled in β, as, e.g., 4. 23. 1: ecquis] pro an quis A^c above the line, pro antiquis et quis β. Wrong division in A causes trouble, e.g. 9. 1. 1 neutrum] ne utrum A, utrum ne β. Less conclusive are the errors in β arising from the confusion of letters written in Beneventan, as at 10. 12. 6 natare] noctare, and 10. 9. 2 ituro] lauro, a classic miscopying in β which has led to an orgy of emendation in its progeny: laturo, labituro, fluituro, duraturo, casuro.

^{page 366 note 2} And even whole stretches of γ text.

^{page 366 note 3} An important step forward was the emergence of these four as a group in the stemma constructed by Hijmans and Forder.

^{page 366 note 4} But the Laurentian manuscripts have been used. Laur. 76. 41 was collated by Gertz for the ad Polybium and the ad Marciam, and his collations were taken over by Hermes in the Teubner text; both have been drawn upon by Viansino.

^{page 367 note 1} Cf. T.L.L. i. 317. 58 ff.Google Scholar

^{page 367 note 2} As in Balliol College MS. 129.

^{page 367 note 3} e.g. Castiglioni, , AlhenŒteum 1913, p. 100: tamquam fundamentum diiudicandarum adfinitatum ponimus.Google Scholar

^{page 367 note 4} 4. 1. 141.

^{page 367 note 5} Hist. 2. 6.Google Scholar

^{page 367 note 6} There are two extant Beneventan manuscripts of Orosius, Monte Cassino 303 and Vaticanus lat. 3340, both of the eleventh century.

^{page 368 note 1} This manuscript is known as D, E, or C, and has found a place in most editions. A number of corrections attributed in the Teubner text to s can be traced back to marginal readings in this manuscript and are indeed not but γ variants.

^{page 368 note 2} e.g. Beeson, , op. cit., p. 246.Google Scholar

^{page 368 note 3} a has errors from which β (and γ) are free 1.4 depravantium se] se om. a; 1. 5 in abdito] in om. a; quietumque]–que om. a; 2. 3 viritim] virium a. β and γ have conjunctive errors, e.g. 2. 1. reorum a, eorum βγ; 2. 3 si tibi a, tibi si βγ.

^{page 369 note 1} L. Annaei Senecae dialogorum libri xii (Copenhagen, 1886), pp. ix–xx.Google Scholar

^{page 369 note 2} Many of the later corrections, particularly those of A⁵, a vicious meddler, are simply taken from one or more of the recentiores. It may be worth investigating whether some of them, such as those assigned to A⁶, are in the hand of Zanobi da Strada, now well known for his habit of annotating Cassinese manuscripts: cf. Gius. Billanovich, , I primi umanisti e le tradizioni del classici latini (Fribourg, 1953), pp. 29 ff.Google Scholar

^{page 369 note 3} There are two glosses in the margin of one page of A: 2. 2. 2 abstractus] in marg. vel arreptus, and 2. 3. 1 tectum] in marg. vel vestitum. There is no sign of these in γ, but the β manuscripts read abstractus vel arreptus and tectum vel vestitum (or something similar).

^{page 370 note 1} In addition to the Dialogues A has a text of the spurious correspondence between Seneca and St. Paul, edited by Barlow, C. W. (Papers and Monographs of the American Academy in Rome, x, 1938).Google Scholar A glance at his stemma shows that this text is closely connected textually with four other manuscripts, all of German provenance and associated respectively with St. Emmeram, Saint-Arnoul (Metz), St. Gall, and Cologne. This seems to me to be a clear result of the strong ties existing in the eleventh century between Germany and Monte Cassino, which had recently had two German abbots. But the Dialogues need not have come from Germany; there is no evidence that they had ever left southern Europe.

^{page 370 note 2} The alien origin of the second part of P is corroborated by the fact that other manuscripts which are very close to P (e.g. Hunterian MS. U. 1. 9 in the University Library in Glasgow, of the fourteenth century) contain only dialogues 1–4 and 9.

^{page 371 note 1} All three of the fourteenth century and all now preserved in the libraries of Oxford colleges: Balliol College 129, Merton College 297, and University College 6. A manuscript which appears to have been related to them—to judge from such readings as have been preserved—is the lost Coloniensis of Gruter, lent to him by the Fratres minores of Cologne (Animadversiones in L. Annaei Senecae Opera [Heidelberg, 1594], p. iv.)Google Scholar It is perhaps worth mentioning that the Balliol manuscript, while in the possession of William Gray, later Bishop of Ely, travelled with him to Cologne in 1442: cf. Mynors, R. A. B., Catalogue of the Manuscripts of Balliol College, Oxford (Oxford, 1963), pp. xxix, 108.Google Scholar A connection between a Cologne manuscript and a purely English group would be more explicable than appears at first sight; but this text may well have travelled directly from Paris to Cologne, and the Friars seem once again to provide the link.

^{page 371 note 2} e.g. 2. 5. 5 movetur iactura; 5. 4. 1 defixis et haerentibus] deftxo inhaerentibus; 7. 1. 1 post lapsus est add. vir; 8. 6. 4 maiora egisse … gessissent errores … maiora gessisse … egissent errores. If such common readings prove to be more widespread than they appear to be, then the affinity between Bacon and the English manuscripts will be more tenuous.

^{page 371 note 3} For example at 1. 4. 9 the manuscripts read velut perpetua ebrietate sopiti, but the structure of the period demands a finite verb. Beeson has pointed out that Bacon (p. 71. 8 in Massa's edition) supports Feldmann's sopiuntur. But if we read further, we find Bacon quoting the same passage again (p. 106. 30), and this time he has sopiti. Bacon does not support anyone; he has just had the same idea as Feldmann did a long time after him, and both are wrong: a finite verb has to be inserted somewhere, but the rhythm shows that the end of the period should be left undisturbed.

^{page 372 note 1} It is clearly right at least twice against the other two: 2. i reorum (eorum β γ) and 2. 3. si tibi (tibi si β γ). An interesting case is 1. 4. Here most editors read (with a) flagrant ac micant oculi. For ac micant β offers emicant, V et micant, R micant. These readings open up possibilities, but the same phrase appears in Martin of Braga's epitome o:the de ira, made in the sixth century, and he reads ac micant. One cannot build an empire on a conjunction, but a is strikingly supported, and this support should not be undermined by the fact that the only authoritative edition of Martin's works (by Barlow, C. W., American Academy in Rome, 1950)Google Scholar reads et micant: the one medieval manuscript, on which Martin's work (and Barlow's edition) mainly rests, is Escorial M. III. 3, of the tenth century, and that has ac micant, as pointed out by Fontàn, A. (Emerita xviii [1950], p. 378) and checked by myself (in fact it needs hac micant, with the false aspiration common in Visigothic manuscripts).Google Scholar