What follows is a pre-print of Chapter II, «The Chronology of the Proto-Accessus», of my study, Tragedy and Comedy from Dante to Pseudo-Dante, which will appear as a volume of the University of California Publications in Modern Philology (Berkeley: University of California Press). My purpose in the monograph as a whole is to analyze what Dante has to say about tragedy and comedy in his authentic writings, and then to study the concepts of his commentators. I have accepted the arguments of those scholars who consider the Epistle to Cangrande to be spurious. It must therefore be treated as one of the commentaries in its proper chronological order. In Chapter II I try to locate the Accessus portion, which deals specifically with tragedy and comedy, in terms of indebtedness and influence; and incidentally I produce new arguments for the letter's inauthenticity.

The Epistle to Cangrande is clearly divided by its final author into two main parts, the opening epistle and the concluding introduction to the Paradiso.¹ The second part in turn is subdivided into a preliminary discussion of the whole Comedia and the discussion proper of Paradiso.² I will call the epistolary section (paragraphs 14) the Dedication, the discussion of the whole work (paragraphs 5-16) the Accessus, and the Paradiso discussion (paragraphs 17-33) the Exposition. I will call the resulting composite the Compilation and the person responsible for it the Compiler ^__ a designation for the most part interchangeable with «Pseudo-Dante». In this chapter I will be particularly concerned with the Accessus, and insofar as it is thought of as antedating Pseudo-Dante's Compilation I will refer to it as the Proto-Accessus and its original author as the Accessor.

      Among the scholars who have questioned the authenticity of the Epistle to Cangrande, one school, led by Augusto Mancini and Bruno Nardi, impugns the Dantean character only of the Accessus and the Exposition, while accepting the Dedication as a letter that Dante actually wrote to Cangrande. Others, including Giorgio Brugnoli, reject the Dedication as well.³ I will take up the matter later, in Chapter VII. But here I am concerned only to agree that the Accessus and the Exposition are not by Dante. I wish to maintain further that the Accessus in its original form was not by the Compiler posing as Dante. The main reason for this last point is that Giuseppe Vandelli in 1901 presented conclusive evidence that Boccaccio, in one sentence of his Italian commentary, was drawing on the corresponding Latin sentence of the Accessus; that is, he demonstrated that the Italian must derive from the Latin, and not vice versa.⁴ Since it is unthinkable that Boccaccio knew the Epistle to Cangrande in its present form, as the words of the venerated Dante himself, he must have seen it in another form.

      It is possible that the entire Epistle to Cangrande was in existence by Boccaccio's time and that Boccaccio saw only an excerpt of it containing all or some of the Accessus. But I consider it very unlikely that a portion of such an astoundingly revelatory letter by Dante could have been circulated without word of the whole letter getting around. Therefore I postulate that an earlier version of the Accessus was written before Boccaccio's time and used by him, and that it was later taken up by the Compiler and incorporated into his Epistle to Cangrande.

      I will assume as a working hypothesis that it was the Compiler rather than the Accessor who was responsible for the Exposition, and that the Proto-Accessus was substantially the same as the Accessus that now appears in Cangrande, except for adjustments made by the Compiler in adapting it to his context of an introduction to Paradiso. In order to set the earlier chronological limit, the terminus post quem, of the Proto-Accessus, we must look at the material held in common by it and the commentaries written before Boccaccio's time. This material has been most thoroughly studied by Luis Jenaro-MacLennan. Jenaro deliberately sets aside the question of whether Cangrande was by Dante, but his analysis seems designed to make such a conclusion very plausible. Jenaro is able to put Dante into the running primarily by his claim that Cangrande antedates Guido da Pisa's long commentary on the Inferno and was drawn upon by Guido. I hold that the influence is in the opposite direction.

      Jenaro proceeds by isolating passages common to Guido, Jacopo della Lana, and the second and third versions of the Ottimo commento, and he designates these passages collectively as N. The passages common to Cangrande and Guido are labeled E¹. Jenaro says⁵ that Guido must have taken E¹ from Cangrande, since Cangrande could not have taken E¹ from Guido without its being contaminated by N and by Guido's intercalations in E¹. But since Jenaro is the one who is assigning the labels, he has seen to it that there is no contamination; he has simply designated only unconnected (or «uncontaminated») sections as N. Jenaro's real objection would seem to be that he finds it implausible that Cangrande, if it were posterior to Guido, should have made such a selective use of Guido; but he does not seem to find it puzzling that Guido should have made such a selective use of Cangrande. Why, for instance, would Guido say that there are only four kinds of literature when Cangrande says that there are six?⁶ Furthermore, Jenaro has begged the crucial question by calling the extra material in Guido's E¹ sections «intercalations». In my view, it is much more probable on the face of it that the material was original to Guido and was accidentally or deliberately dropped by the Accessor or by the Compiler.

      There seems to be a likelihood of accidental omission by reason of homoeoteleuton in Cangrande's statement of Dante's literal subject. Here is Guido's commentary on the point, with the passage omitted by Cangrande in brackets (I have italicized the long identical phrases that would have given rise to the scribal eyeskip):

Si enim accipiatur litteraliter, dico quod subiectum huius operis est status animarum post mortem simpliciter sumptus: [qui quidem status dividitur in tres partes, prout conditio animarum est triplex. Primus status sive conditio est illarum animarum que eternaliter sunt damnate, et que in penis habitant sine spe aliqua evadendi ex illis; et ista pars appellatur Infernus. Secundus status sive conditio est illarum animarum que voluntarie stant in penis, ut Deo satisfaciant de commissis, et sunt in ipsis penis cum spe ad gloriam ascendendi; et ista pars Purgatorium appellatur. Tertius status sive conditio est illarum animarum que sunt in beata gloria, ipsi summo et eterno bono eternaliter (hoc est, sine fine) coniuncte; et ista pars appellatur Paradisus. Et sic patet quomodo subiectum huius operis est status animarum post mortem simpliciter sumptus]. Nam de illo et circa illum totius huius operis versatur processus.⁷

The text of Cangrande is as follows:

Est ergo subiectum totius operis, litteraliter tantum accepti, status animarum post mortem simpliciter sumptus. Nam de illo et circa illum totius operis versatur processus.⁸

In this case, according to my hypothesis, it must have been the Accessor and not the Compiler who made the omission, since it is reflected in Boccaccio's commentary:

È adunque il suggetto, secondo il senso litterale, lo stato dell'anime dopo la morte de' corpi semplicemente preso; per ciò che di quello e intorno a quello tutto il processo della presente opera intende.⁹

Guido's authoritative dico in his version of the passage is also an indication of priority.¹⁰

      In another passage, where the topic under discussion is the final cause, the likelihood is that the Cangrande text represents a deliberate truncation of Guido's lengthy discourse. I give Guido's text and bracket the material not in Cangrande:

Circa quartam, id est circa causam finalem, nota quod autor istud opus composuit ad hunc finem principaliter, licet et multi alii possint assignari fines. Est autem principalis eius intentio removere viventes a statu miserie, [relinquendo peccata, et sic composuit Infernum; reducere ad virtutes, et sic composuit Purgatorium;] ut sic eos perducat ad gloriam, [et sic composuit Paradisum. Fines vero alii qui possunt assignari in hoc opere sunt tres: primus, ut discant homines polite et ordinate loqui. ... Secundus finis est ut libros poetarum ... renovaret. ... Tertius finis est ut vitam ... malorum ... condemnaret, bonorum autem ... commendaret. Et sic patet que est causa finalis in hoc opere].¹¹

I have omitted a good deal of the last part, but one can see the sort of expansiveness that Guido indulges in. Cangrande, in contrast, has an air of cutting through garrulity. The text reads:

Finis totius et partis esse posset et multiplex, scilicet propinquus et remotus. Sed, omissa subtili investigatione, dicendum est breviter quod finis totius et partis est removere viventes in hac vita de statu miserie et perducere ad statum felicitatis.¹²

I take it that the cut was made by the Accessor, but that it was the Compiler, drawing on scholastic tradition for his terminology of proximate and remote ends, who added the totius et partis phrases when he adjusted the Proto-Accessus to his discussion of the Paradiso. Boccaccio's commentary must preserve something of the original form of the Proto-Accessus passage: «La causa finale della presente opera è remuovere quegli che nella presente vita vivono dallo stato della miseria allo stato della felicità».¹³ But Boccaccio dropped the phrase, omissa subtili investigatione, which I presume was in the Accessor's original text.

      However, one reference to the «whole and part» appears in Guido's commentary: «Non ad speculandum sed ad opus inventum et fictum est totum et pars». This statement is repeated in Cangrande, except that «et fictum» is omitted.¹⁴ Defenders of Dante's authorship of Cangrande, or defenders of the priority of Cangrande, have alleged this passage as proof that Guido must be drawing on Cangrande, since Guido would have had no reason to contrast one of the parts of the Comedy with the whole. The same would be true, I must add, if there was indeed a Proto-Accessus of the sort that I have postulated: the Accessor would be speaking of all three parts of the Comedy, and not of one in particular. Nevertheless, I consider it plausible that the expression was in fact original to Guido, for two reasons. First, Guido goes on to contrast whole with parts by referring to occasional speculative passages within the Comedy as possible exceptions to his general statement: «Nam etsi in aliquo loco vel passu pertractatur ad modum speculativi negotii, hoc non est gratia speculativi negotii principaliter, sed operis».¹⁵ The second reason is suggested by Giuseppe Boffito, who points out that totum and pars are often coupled together without evident reason.¹⁶ I would prefer to say, not that there is no evident reason for such couplings, but rather that the reason is obvious but unstated: the expression totum et pars is an elliptical-pleonastic way of emphasizing completeness, short for totum et omnis pars or in toto et in omni parte.¹⁷

      The Guido/Accessus passage goes on to cite Aristotle: «quia, ut ait Philosophus, secundo Methaphysice, "Ad aliquid et nunc speculantur practici," aliquando».¹⁸ One might be inclined to think it too much of a coincidence that the Compiler also quotes from the same paragraph of the Metaphysics when introducing the Accessus, to justify his dealing with the whole Comedy before treating of Paradiso. He says, «Sicut dicit Phylosophus in secundo Metaphysice, "sicut res se habet ad esse, sic se habet ad veritatem''».¹⁹ But if one accepts my hypothesis that the Compiler is making use of a pre-existing Accessus, there is no coincidence at all. The Compiler did not independently think of drawing on the Aristotle passage; rather, he was inspired by the Accessor's use of it. (This part of the Metaphysics is also referred to in the Exposition; I will discuss the implications of this point in Chapter VII.)

      The above explanation is verified by the circumstance that the translation used by the Compiler is different from that used in the Guido/Accessus passage. Guido employs the twelfth-century Anonymous translation, whether in its original form or in the revision made by William of Moerbeke in the 1260s. Moerbeke's was the fifth version of the Metaphysics to be produced in less than a century. The version used by the Compiler was the oldest of the translations, that made by James of Venice (Iacobus Veneticus Graecus). The Anonymous/Moerbeke version of the first passage is, «ad aliquid et nunc speculantur practici», and of the second, «sicut habet esse, ita et veritatem» (Anonymous) or «sicut se habet ut sit, ita et ad veritatem» (Moerbeke).²⁰ James gives the first as, «ad aliquid et nunc considerant practici», and the second as, «sicut se habet ad esse, sic et ad veritatem» (or, as one manuscript has it, «sicut se habet ad esse, ita se habet ad veritatem»).²¹

      We know that Dante himself was familiar with the Metaphysics, to some extent at least. He drew on the first part of Thomas Aquinas's prologue to his Metaphysics commentary in the Convivio and the Monarchia:²² and in the Monarchia he quotes from Moerbeke's translation at the point where Thomas's copy ended.²³ If he was using the whole work and not simply repeating an excerpt cited by some other author, he may have been using the text of the translation accompanying Aquinas's commentary. In any case, it does not seem very likely that Dante as author of the Epistle to Cangrande would have drawn on two different translations of the Metaphysics (or excerpts from two different translations) to cite portions of the same paragraph.

      The same unlikelihood does not apply to a later Compiler, for the Anonymous/Moerbeke text would simply have been swept in with the block of material taken from Guido da Pisa ^__ or rather taken over by the Accessor from Guido. The Compiler, then, would be drawing directly only on the James of Venice translation.

      If the Compiler himself was the author of the Accessus, we would have to postulate that the Accessus had somehow become separated from the Dedication (and, probably, from the Exposition as well) before it was consulted by Boccaccio. If the Compiler was not the author of the Accessus (the hypothesis I favor), we must date the Accessus in its original form (before the Compiler adapted it) to a time after Guido's commentary and before Boccaccio's. We can say nothing more about the Accessor's sources until we deal further with Guido and examine what some of the other commentators have to say. But since it is certain that Boccaccio used the Accessus at the time of his Dante lectures in the 1370s, we must ask whether there are any signs of his earlier acquaintance with the work. In the first version of his Trattatello in laude di Dante, which is now dated to the years 1350-1355,²⁴ Boccaccio's statement of Dante's intention in the Comedy, «cioè a volere secondo i meriti e mordere e premiare, secondo la sua diversità, la vita degli uomini»,²⁵ may have been influenced by the Accessor's characterization of the subject of the poem as «homo prout merendo et demerendo per arbitrii libertatem iustitie premiandi et puniendi obnoxius est».²⁶ Furthermore, the Accessor's account of the Comedy's divisions («quelibet cantica dividitur in cantus ... quilibet cantus dividitur in rithimos»)²⁷ is suggestive of Boccaccio's formulation: «De' quali tre libri egli ciascuno distinse per canti e i canti per rittimi».²⁸ Boccaccio speaks of Dante's title as referring to the poem's humble style: «L'andar quieto significa l'umilità dello stilo, il quale nelle comedie di necessità si richiede, come color sanno che intendono che vuole dire "comedia"».²⁹ In so doing he accepts a reason that he will reject in his Esposizioni, where he is clearly drawing on the Accessus. The conclusion that he was drawing on the Accessus in the earlier work as well is strengthened by the fact that he uses the word quieto. As we will see in Chapter VI, quieta was the Accessor's substitution for Guido's grata in his description of the first part of tragedy.

HENRY ANSGAR KELLY

University of California Los Angeles

NOTES

¹ This division is set forth in Cangrande 4.13: «Itaque, formula consumata epistoli, ad introductionem oblati operis aliquid sub lectoris officio compendiose aggrediar». For the text of Cangrande I generally follow Giorgio Brugnoli's edition in Opere minori, vol. 2, ed. P. V. Mengaldo et al. (Florence 1979), pp. 512-521 and 598-643. The original manuscripts are given in Friedrich Schneider's facsimile edition, Die Handschriften des Briefes Dantes an Can Grande della Scala (Zwickau 1933), and the text of an additional MS of the Dedication, B. is given by Augusto Mancini, «Un nuovo codice dell'Epistola a Can Grande», SD 24 (1939), pp. 111-122. I have also consulted the editions of Giuseppe Boffito, L'Epistola di Dante Alighieri a Cangrande della Scala (Turin 1907), and Paget Toynbee, Dantis Alagherii Epistolae (Oxford 1920; reprinted with preface and additional bibliography by Colin Hardie, 1966), pp. 166-211. Toynbee's variants are not based on the manuscripts but on Boffito's edition, which is not reliable.

² This subdivision is explained in Cangrande 6.17: «Volentes igitur aliqualem introductionem tradere de parte operis alicuius, oportet aliquam notitiam tradere de toto cuius est pars. Quapropter et ego, volens de parte supra nominata totius Comedie aliquid tradere per modum introductionis, aliquid de toto opere premittendum existimavi, ut facilior et perfectior sit ad partem introitus». The transition to the Paradiso discussion comes at 17.42: «His itaque premissis, ad expositionem littere secundum quandam prelibationem accedendum est».

³ See Brugnoli, Cangrande, pp. 518-519.

⁴ Giuseppe Vandelli, review article in the Bullettino della Società Dantesca Italiana, n.s. 8 (1900-1), pp. 137-164, esp. p. 156, n. 2. He cites a line from Cangrande 7.22, which reads in Brugnoli's edition: «Et quanquam isti sensus mistici variis appellentur nominibus, generaliter omnes dici possunt allegorici». Boccaccio read the generaliter as going with the first part of the sentence: «Et quanquam isti sensus mistici variis appellentur nominibus generaliter», etc., with the following result: «E così come questi sensi mistici sono generalmente per vari nomi appellati, tutti nondimeno si possono appellare allegorici». Cited from his Esposizioni sopra la Comedia di Dante (1.2.21), ed. Giorgio Padoan, Tutte le opere, VI (Milan 1965), p. 58. Vandelli's argument is accepted by Luis Jenaro-MacLennan, The Trecento Commentaries on the Divina Commedia and the Epistle to Cangrande (Oxford 1974), p. 105 (he mistakenly cites the argument as appearing on p. 165 rather than 156).

⁵ Jenaro, Trecento (see previous note), pp. 60-61, 67, 75.

⁶ Guido names lyric, satiric, tragic, and comic poets. Cangrande speaks of poetic narrations and names comedy, tragedy, bucolic verse, elegy, satire, and votive sentence.

⁷ Guido da Pisa, Expositiones et Glose super Comediam Dantis; or, Commentary on Dante's Inferno, ed. Vincenzo Cioffari (Albany N.Y. 1974), Prologue, pp. 2-3.

⁸ Cangrande 8.24.

⁹ Boccaccio, Esposizioni, Accessus 8. Boccaccio's «tutto il processo della presente opera» does not correspond to Guido's «totius huius operis versatur processus» or Cangrande's «totius operis versatur processus»; but rather it agrees with the Laurentian version of Guido's prologue given by Vandelli, p 151: «totus huius operis versatur processus». I can only explain the similarity as a coincidence. Cf. n. 15 below.

¹⁰ This point is disputed by Jenaro, pp. 82-83.

¹¹ Guido, p. 4.

¹² Cangrande, 15.39.

¹³ Boccaccio, Accessus 12.

¹⁴ Guido, pp. 4-5; Cangrande 16.40. Boccaccio, Accessus 42, omits the whole explanatory clause.

¹⁵ Guido, p. 5; Cangrande 16.41 omits principaliter and puts a second gratia before operis. Me and M² have pertractatatur, R has passim pertractatur, M² has pertractat, V has pertractamus, and M³ has tractamus. The Laurentian Guido has tractatur, and puts si for etsi (for Cangrande on etsi, see below). Cf. Boccaccio, Accessus 42: «per ciò che, quantunque in alcun passo si tratti per modo speculativo, non è perciò per cagione di speculazione ciò posto, ma per cagione dell'opera, la quale quivi ha quel modo richesto di trattare». Here Boccaccio's quantunque agrees with the etsi of the cited text rather than with the si of the Laurentian Guido (see n. 9 above). As for Cangrande, V, M¹ and M² have si et, Me and R have si, and M³ has et si.

¹⁶ Boffito (n. 1 above), p. 28.

¹⁷ Cf. the somewhat similar English expressions, «whole and some», «all and some», «all and sundry», «one and all», and so on. See the Oxford English Dictionary s.v. «all» A.12; «one» B.29; «whole» A.7b. Legal documents are particularly rich in such tautological expressions; a good example is «all and every», translating the Latin «universi et singuli»; see OED s.v. «every» 7b.

¹⁸ This is the reading of Guido's text. As for Cangrande 16.41, Brugnoli puts Metaphysicorum, but V has metaphysice, M³ has Methaphysice, and the other MSS abbreviate the title. Five of the MSS cite the text as in Guido, including the aliquando at the end, but V substitutes «aliquando etiam speculantur practici» for the whole quotation.

¹⁹ Cangrande 5.14. Brugnoli again has Metaphysicorum, but all of the MSS except V abbreviate it, and V has metaphysice. The same is true of 20.56. As for the text of the citation from Aristotle, only M¹ gives it as I cite it above (where I follow Brugnoli's edition). V, Me, and M³ invert the second se habet and read habet se, and for the first se habet M³ has sese habet. For the second se habet R has habet with et se careted in after it.

²⁰ Aristotle, Metaphysica 2.1 (993b22-23): Translatio Anonyma sive «Media», ed. Gudrun Vuillemin-Diem, Aristoteles latinus 25.2 (Leiden 1976), p. 37; Translatio Guillelmi de Moerbeka, as given in Thomas Aquinas, In duodecim libros Metaphysicorum Aristotelis expositio, ed. Raimondo M. Spiazzi (Turin 1950), p. 84. Dr. Vuillemin is preparing an edition of Moerbeke, to appear as Aristoteles latinus 25.3. See her «Untersuchungen zu Wilhelm von Moerbekes Metaphysikübersetzung», in Studien zur mittelalterlichen Geistesgeschichte und ihren Quellen, ed. Albert Zimmermann and Gudrun Vuillemin-Diem, Miscellanea mediaevalia 15 (Berlin 1982), pp. 102-208.

²¹ Aristotle, Metaphysica 2.1, Translatio Iacobi, ed. G. Vuillemin-Diem, Aristoteles latinus 25.1 (Brussels 1970), p. 37. For details of these and the other translations, see the Praefatio, pp. xi-xii. In a personal communication, Dr. Vuillemin tells me that she knows of no manuscript (or florilegium) that mixes translations in this passage.

²² Dante, Convivio 4.4.4-5; Monarchia 1.3.10. See Lorenzo MinioPaulello, «Dante's Reading of Aristotle», in The World of Dante, ed. Cecil Grayson (Oxford 1980), pp. 61-80, esp. 74-76.

²³ Dante, Monarchia 1.10.6, ed. Pier Giorgio Ricci (Milan 1965), p. 153: «Et hanc rationem videbat Phylosophus cum dicebat: "Entia nolunt male disponi; malum autem pluralitas principatuum. Unus ergo princeps"». Moerbeke's translation reads: «Entia vero nolunt disponi male, nec bonum pluralitas principatuum. Unus ergo princeps» (Thomas Aquinas, In Meta., p. 611; Aristotle, Meta. 12.10, Bekker 1076a3-4). The James of Venice and Composite translations of this portion of the Metaphysics do not survive (or did not extend this far); the Anonymous version reads as follows (p. 224): «Entia vero nolunt tractari male. Nec bonum plures dominatus. Unus ergo dominatus» (some texts read princeps for the last dominatus).

²⁴ See Pier Giorgio Ricci's introduction to his edition of Boccaccio's Trattatello, Tutte le opere, III (Milan 1974), pp. 426-427.

²⁵ Boccaccio, 1 Trattatello 177 (Ricci, p. 481).

²⁶ Cangrande 8.25. This and the other similarities noted in the text above between Cangrande and the Trattatello were pointed out to me by Professor Giuseppe Velli.

²⁷ Cangrande 9.26, ed. Brugnoli. I will note in Chapter VII that the mss have rhythmos or other disyllabic forms.

²⁸ Boccaccio, 1 Trattatello 177 (p. 482).

²⁹ Ibid. 226 (p. 494).