This is the first in the series of Barlow Lectures delivered at University College London on 17-18 March 1993 by Professor Robert Hollander of Princeton University. The lecture texts distributed now are only a part of the detailed presentation of argument and evidence that will make up the printed book. Scholarly users in particular should be aware that only the printed book will incorporate final revisions and corrections and the printed book should in all cases be used as the authoritative citation of the author's work. Anyone who wishes to be notified when the printed book is available should send e-mail to: michael_kehoe@um.cc.umich.edu or regular mail to Michael Kehoe, The University of Michigan Press, 839 Greene Street, P.O. Box 1104, Ann Arbor, Michigan 48106. This text was put into SGML by J. M. Dean (6/26/96) [Editor's Note: Please direct any questions or corrections to Jim Dean (dean@udel.edu)
(1) The Beginnings of a New Debate
As far as Dante studies are concerned, a debated issue that has only a 174-year history is, relatively speaking, barely out of its adolescence. And such is the case with regard to the Epistola a Cangrande, the authenticity of which was first attacked by Scolari in 1819. If the Epistle only became problematic late in its career, it is also true that it had belated public recognition as having been written by Dante. While, as we shall see, it left its mark on Dante's readers within ten years of its probable date of composition, it was not alluded to as being of Dantean origin in a highly visible way until Filippo Villani, one of Dante's most erratic, if nonetheless interesting commentators, announced its Dantean paternity in the early years of the fifteenth century. The pendulum of the question of its authenticity, within the relatively recent period of "scientific" philological debate, has swung between the poles of acceptance and denial, and has done so not slowly, but rapidly, in several notable "moments."
At the turn of the century D'Ovidio vigorously attacked authenticity, but was immediately opposed by Torraca, Vandelli, and Moore. In the late 1930's Porena and Pietrobono launched urgent assaults, the latter answered by Vandelli, his championing of the Epistle's genuineness unwithered by age. In the 1940's and 1950's Mancini's thrust was met by Mazzoni's parry, which convinced many, if not Bruno Nardi. The 1960's began with two stinging assaults upon Mazzoni's position by Nardi (Colin Hardie, too, entered the fray on the side of inauthenticity); Nardi was answered by such as Mazzeo, Padoan, and Billanovich. In 1974 we had the important study of Jenaro-MacLennan, the burden of which, avoiding the essential question in order to sift through its sandy underpinnings, was to establish the priority of the text of the Epistle, by whomever it was written, so as to disarm those who claimed it was a document composed by a later "forger."
Mazzoni's work, welcomed by so magisterial a presence as Contini, essentially carried the day, if pockets of armed resistance dotted the difficult Balkans of Dante studies. It is only very recently that a series of publications, in the wake of Giorgio Brugnoli's edition of 1979, has vigorously sought to overturn what had become, for practical purposes, an official consensus. This latest rebellion has spread quickly, its fires fanned by Dronke's book in 1986, Kelly's in 1989, and by articles by Hall and Sowell and by Baranski in the last three years. I believe that, if one attends congresses of Dantists, especially in my country, the current moment is one in which one finds the inauthenticists coming into the ascendancy. Dronke and Kelly are treated as the Dioscuri who have finally rid us of the burden of belief in the embarrassing document that would have Dante, in his own voice, say such foolish things about his Comedy.
The monograph on the subject that I completed last summer, and which is scheduled for publication later this year by Michigan University Press, is the result, I confess it, of my desire to attend closely to these negating arguments in the expectation of finding them wanting. It is dedicated to Francesco Mazzoni, a man who is, I am honored to say, one of my closest friends. It is also, more importantly, dedicated to a recuperation of his findings. And, while the first gesture is meant to represent my feelings for the man, the second effort leaves friendship to one side. As Mazzoni will attest, he and I disagree on a number of points concerning Dante's work, some of them of considerable importance. My work on the Epistle's status as object of debate has indicated to me that, on most of the crucial points concerning the Epistola a Cangrande, his work was, and remains, indispensable.
In the three lectures that I have so kindly been asked to deliver here at University College, under the name of a major British Dantist, I will assuredly try your patience, will certainly offer small occasion for delight, and may only instruct, if at all, in matters that are of little interest to this public. Nonetheless, I have chosen to assume an interest on your part to which the facts may give the lie. What, one may reasonably inquire (as one of my country's leading students of Dante has indeed recently inquired, as you shall shortly hear) does it profit us to gain the Epistle to Cangrande? Do we really require its authority in order to interpret Dante's poem? If we were to lose the Epistle to the labor and ingenuity of the inauthenticists, would we be losing Dante's soul? The easy answer is that the authentication of any text of Dante's is of considerable interest and importance. Further, to know that we can count on the authenticity of a text that probably should be understood to offer the (not "a" but "the") interpretive key to the Commedia would be and should be a matter of some significance. I shall thus allow myself to hope for two things, not only your good will, but your acquaintance with this document over which there has been such fierce struggle. I have no illusion that the battle will end with this latest item in a growing bibliography. What I hope to show is that the recent arguments opposing authenticity have failed to dislodge the best working hypothesis: the Epistle is the work of Dante himself.
It seems only yesterday that most Dante scholars were certain that the Epistola a Cangrande was written by Dante Alighieri. Recently, this view of the matter has been challenged. The last few years have witnessed the appearance of several studies that have as their purpose the reopening of the question, whether in such a way as to make the attribution to Dante at least highly unlikely (the position taken by Peter Dronke and Zygmunt Baranski) or even impossible (that taken by Henry Kelly). Their three studies have, with Giorgio Brugnoli's edition of 1979 obviously having opened the way and now performing the function of base of operations, served to embolden still others, especially in the United States, hotbed of allegorism and anti-allegorism, to take up the cudgels against the Auerbach-Singleton position, with its view that the essential allegory of the Commedia is theological, in other words, aptly described by the Epistola. (Dronke is not alone in denying the privileged position of the figural interpretation sponsored by Auerbach as opposed to other forms of medieval allegoresis, while also asserting that the Epistle is not, in any case, a very good guide to the poem.) The present survey of the recent debate over the question of the authenticity of the Epistola a Cangrande is intended to demonstrate two things. First, that the cases made against authenticity are not, on their own terms, as convincing as they would have to be in order to turn the argument in their favor; second, that the argument for authenticity had already been made more convincingly than its recent opponents have been willing to acknowledge.
Does it all really matter? That question, answered in the negative, and con brio, by Teodolinda Barolini in 1989, should detain us. Here is her central observation: ". . . I am convinced that the Commedia's imitation of God's way of writing, in defiance of all theological protocol, does not require the Epistle to Cangrande or any other external document for its substantiation; the poem itself furnishes sufficient and incontrovertible evidence of how it wants to be read. By this I do not mean to say that I find the Epistle uninteresting; for the record, my sense of it is that it is Dante's. But I do mean to say that its authorship, were the matter one day to be decided against Dante, need not in any way impinge on our reading of the Commedia." These remarks are thus not directed to Professor Barolini's attention. Yet I must object to her dismissal of the question on these grounds (even as I agree that the poem itself tells us how to heed its claims for telling the truth and reveals its borrowing of the exegetical method of theological allegory). Almost all of those who are interested in removing Dante's name from the Epistle are motivated --one would better say highly motivated-- by the desire to do away with an important and imposing injunction that they would otherwise have to observe as readers and commentators, an injunction that would at least inhibit (or even invalidate) the procedures and definitions to which they would like to remain accustomed. This stumbling- block may be the four senses (and not "levels," as so many incorrectly say, yielding to the spirit of modernity; the medieval term was always sensus) of Scripture or, among yet other impedimenta and as has become increasingly the case, the definitions of tragedy and comedy found in section 10 of the Epistle. (This is the essential problem that concerned Porena in 1933; the modern version of this argument is most challengingly present in Brugnoli, who spends 368 (or 46%) of the 790 lines of comment which he devotes to the second and third parts of the Epistle, the accessus (sections 5-16) and exposition (sections 17-33), to section 10 (which occupies less than 3% of those parts of the text), thus revealing an extremely acute interest in this problem. Kelly and Baranski are also centrally concerned with this section of the Epistle. In other words, no one who has worked on the problem, with the sole possible exception of Mazzoni (to whose contributions there will be continuing reference), can be said to be free of an ulterior interpretive motive. (Both Brugnoli and Barolini point out that Mazzoni's actual reading of Dante does not make use of the "allegory of the theologians" as his defense of the Epistle's genuineness might make one expect he would.) To that charge your speaker must also plead guilty, insofar as a sense that the Epistle is congruent with the poem represents a parti pris. However, my method here will not be to show how I believe the Epistola a Cangrande squares with Dante's poem except in answer to some specific charges that it does not do so. In fact, we will probably never be in a position to know that the Epistle is Dante's, only that it probably is. The same thing may be said of any disputed reading of a textual variant in the poem. Final proof lies only in the lost autographs; and there is not much hope that we shall ever have these. On the other hand, what is at stake is the superiority of one argument over another. Which seems best adapted to the situation it purports to represent? Which accords best with what we know of Dante's habit of mind? And in these respects I think it is evident that the arguments for authenticity are clearly to be preferred. (Here are the words of Moore in 1903 to similar purpose: "Few, if any, of the minor works of Dante are more worth the effort of salvage. For, if genuine, it not only gives us an authoritative exposition by Dante himself of a portion of his own works; but, what is even more important, it throws light on the spirit and method by which he would have us interpret them generally.") And that is the case which I will attempt to make in these lectures. In order to do so, I will spend a good deal of time walking on the via negativa in a review of the recent arguments against authenticity, not because it is my pleasure to quarrel with my colleagues, but because the question, as it lies under their hands, has seemed to many to have been resolved. Since, in disagreement with Barolini, I think the question of the authenticity of the Epistle has significant impact on the continuing debate concerning Dante's stance as poet, I have chosen to deal at some length with a number of the particulars of their arguments.
The Epistola a Cangrande was first definitively referred to as being written by Dante by Filippo Villani in the first years of the fifteenth century but was also cited, apparently as an anonymous accessus and/or prologue to the Commedia, by many of the fourteenth-century commentators: Guido da Pisa, Jacopo della Lana, the Ottimo, Pietro Alighieri, Boccaccio, Francesco da Buti, and Benvenuto da Imola. Its genuineness has been under review since Scolari first questioned its provenance in 1819, 119 years after its first printing. It is fair to say that all who have doubted its genuineness have done so with the sense that they have a just reason for their suspicion, that is, because the Epistle violates, in one respect or another, their sense of Dante. It is also fair to say that many of those who have defended its authenticity have a similar stake in the matter, namely, that the Epistle seconds their sense of how the Commedia "works."
Since we possess a document bearing Dante's name and treated explicitly as authentic by one early commentator (Filippo Villani; but others will also receive our attention), the burden of proof rather falls on those who would deny genuineness. That there have been so many of these, all using different arguments (if sometimes repeating, apparently without knowing so, previous tactical maneuvers), is probably a sign of weakness in the basic position. And where some have argued for the inauthenticity of the whole (in the earlier stages of the debate only D'Ovidio and Pietrobono; in the current one only Kelly), many, following Mancini and Nardi have accepted as genuine the first four paragraphs of the work. These are so deeply Dantean in their style and expression that it is nearly impossible to attribute them to anyone else. However, once the authenticity of the beginning is granted, as it almost certainly must be (and even the insistently negative Kelly hedges his position on this issue), the argument of those opposed to authenticity is in serious difficulty, since we can find evidence of at least one of the first commentators who cites elements of the epistolary opening along with parts of the rest of the Epistle as early as 1328, as will be pointed out. It is perhaps for that reason that the most recent editor of the Epistle, Giorgio Brugnoli, spends a certain amount of ingenuity in attempting to call into question the authenticity of the opening, without risking his argument on a total denial.
The work of Mazzoni and Brugnoli has subsumed the salient points of most of the antecedent scholarship. However, the arguments put forward by Mazzoni and Jenaro-MacLennan have been seriously underattended, even by Brugnoli (who finished work on his edition in 1973, a year before the second's work appeared, but who might have been expected to have seen his preceding article, published in 1968, on the relation of Guido da Pisa's commentary to the Epistola), and nearly totally by Dronke and Kelly, as well as others (only Baranski has done a competent job of indicating the presence of counter-arguments to his thesis, even if he does not deal in detail with them, a task, we may assume, to which he will commit himself in his promised volume on the problems raised by the Epistola). These arguments need restating. Thus my method will be to review the present state of the question as it reflects, or fails to reflect, precedent discussion.
Giorgio Brugnoli's edition of the Epistola a Cangrande was finished in 1973 and published in 1979. Its importance does not lie in its editorial work on the text, which is nearly entirely dependent on the previous edition of Pistelli, published in 1921, but in the interpretive apparatus which sits beneath that text. In fact, contemporary dubiety concerning the Epistola's authenticity stems most directly from Brugnoli's arguments, themselves frequently reflecting judgments offered by Francesco D'Ovidio, Augusto Mancini, and Bruno Nardi. Thus anyone who wants to study the current phase of the dispute must turn to this work, which is the first and perhaps the most challenging of the recent attacks upon the authenticity of the Epistola. As we shall see, in his treatment of the two major antagonists of mid-century, Mazzoni and Nardi, Brugnoli's evident loyalty to the latter's ideas frequently leads him to overvalue Nardi's arguments and allows him to underrepresent Mazzoni's.
The pages of his introduction follow the state of the problem up to 1969. We are told of the two groups of manuscripts, as described by Mazzoni, the elder group (alpha), comprised of three (from the fifteenth century) containing only the epistolary portion (sections 1-4), the younger one (beta), containing six (all of these dating from the sixteenth century). Against the later hypotheses of Henry Kelly, who supposes, repeating Mancini's argument without apparently realizing that he does so, that the dedicator's concluding sentence, in which he turns from the epistolary format (and style) in order to offer his commentary (section 4.13: "Itaque, formula consumata epistole, ad introductionem oblati operis aliquid sub lectoris officio compendiose aggrediar") must have been added by a different hand (he assigns it to the "Compiler" of the Pseudo- Dantean text that he also hypothesizes), we have some fairly hard evidence that, in its original state, the epistolary dedication was as we know it today. Brugnoli points out that the manuscripts of the alpha group would all seem to derive from a larger version of the Epistle, that composed of the familiar thirty-three sections. As I suggested in 1969, it is probably significant that the Epistle, like the Paradiso, is divided into thirty-three parts. Moore's remark (which dates from 1903) is perhaps more interesting: "I do not know what may be the authority for the division of the Epistle into sections. If the work is genuine this is probably due to Dante himself, like the numbering of the chapters in the Convito. It is perhaps significant that this letter introductory to the Divina Commedia, and the Paradiso in particular, should contain just thirty-three sections, the number of Cantos in each Cantica (Canto i of the Inferno being Introductory)." Were the dedication alone to have been written by Dante and the rest by a falsario, the existence of texts which do not contain this "bridge" between the two parts would have provided strong support to those who argue against authenticity. They are denied this comfort and are forced to argue for the presence of the hand of a forger even in the surviving manuscripts of the alpha group. The lost archetype of this family of codices and that of the beta group is thus almost certainly a document very much like the one we know today as the Epistola. Nonetheless, Mancini argued that the last sentence of the dedication was spurious. He had no textual ground for doing so. However, like Kelly (who only mentions Mancini once in his book and makes no reference to his rather similar arguments), he did not want the Epistola to be Dantean. And since, for Mancini and Kelly, if that last sentence is allowed to be part of the dedication, the rest of the Epistola is that much more difficult to classify as spurious, out it goes. The world of the expositors of the Epistola a Cangrande often resembles the world of literary fraud so happily rendered in Umberto Eco's Il nome della rosa.
The dedicatory opening portion of the Epistola a Cangrande is, without doubt, the part of the work that is most easily seen to be "like Dante." It is of some importance that even those most rigidly opposed to the idea that the doctrinal stance of the accessus (sections 5-16, or, for that matter, the commentary procedure of the exposition [sections 17-31]) be considered that of Dante himself, like Mancini and Nardi, had little doubt but that the dedication is genuine. We can be sure that such as they would have been delighted to have had any excuse whatsoever to be rid of the whole Epistola, if only because the attribution of even a part of the document made their arguments more difficult. Their critical integrity, however, stood in the way of convenience. Coming to this question for the first time, any objective reader of the first four paragraphs who is well acquainted with the works of Dante would see why. It is difficult to imagine how anyone would not consent to the notion that this is indeed Dante's work. The likeness is so great that, were we to have the Epistle without its salutation, and thus without Dante's name affixed to it, I feel confident that at least some skilled dantisti would have seen that it is their poet's work. The dedication does exactly the kinds of literary things that Dante alone has made familiar to us. Even one of the most vocal deniers of the authenticity of the rest of the Epistle, Bruno Nardi, has described well the reasons for which we should think that the dedication is of Dantean provenance: "le tipiche cadenze ritmiche dello stile epistolare, le reminiscenze di altre opere dantesche,. . . l'ardita franchezza nel rivolgersi al principe amico e il non meno ardito ragionare col quale questa franchezza si giustifica, la stessa fiera convinzione d'aver trovato finalmente tra le sue coserelle, quella davvero degna dell'ospite liberale e magnifico, da ricambiare in dono" ["the typical rhythmic cadences of the epistolary style, the reminiscences of other works of Dante, the bold frankness with which he turns to his princely friend and the no less bold speech with which this frankness is giustified, the proud assurance that he has finally found, among his possessions of little worth, one truly worthy of his generous and munificent host that he may offer up as a gift."] There are some pieces of evidence to add to Nardi's description, but it serves well as a general understanding of the way in which the writing and thought of the dedicator and of Dante seem identical. Indeed, Nardi had already referred to the Dantean nature of the first four paragraphs in still more insistent ways: they "recano cosi` visibili e inconfondibili i caratteri dello stile dantesco, quasi direi le impronte digitali del loro autore" ["bear, obviously and unmistakably, the characteristics of Dante's style, as I might say, the fingerprints of their author."] Moore's characterization, which is concerned with the Epistle as a whole, is also worth having: "the minute divisions and subdivisions [Moore refers to the Convivio for Dante's description of his habit of so dividing his commentary to his canzoni]; the elaborate analysis and planning out of the whole subject; the same cold-blooded dissection, not to say vivisection, of his own poetry, which is found on almost every page of the Vita Nuova and of the Convito; the tendency to diverge at a tangent into metaphysical or scholastic disquisitions and fine-drawn distinctions, often suggested by a single word, and, often too, seeming to our modern notions irrelevant or pedantic; the enforcement of conclusions by appeals to authority, after they have been established by arguments; and the intermingling in such appeals of quotations from sacred and profane authors. All these are general characteristics." While Kelly is wrong to say that Brugnoli "rejects" the authenticity of the dedication, he would have been correct had he claimed that Brugnoli would like to be able to reject it, and indeed suggests, in the passage to which Kelly refers, that Nardi's arguments may not be convincing. Indeed, Brugnoli spends a good deal of effort in attempts to undermine previous judgments of the genuineness of this part of the text.
Dante's naming of himself in a notably self-conscious way, his use of an extended comparison developed from figural allegory (Sheba sought Jerusalem [and thus Solomon], Apollo sought Helicon [and thus the Muses], so now does Dante seek Verona [and thus Cangrande]), his combining a Hebrew/Christian and an Ovidian source in that typological comparison, the citation of Virgil ("fama vigil volitando disseminat" [Aeneid IV, 173]) and the many citations of his own work, notably the Convivio (first noted by Moore and now evident in Brugnoli's notes) mark the Dantean atmosphere of the first four paragraphs. Also present in the first of these (as previously noted by Moore, Giuliani, and others) are several passages that coincide with Paradiso XVII. What are we to make of these? If the Epistle is genuine, is it quoting Dante or is he quoting it? We must turn for a moment to the question of the date of the Epistle. Mazzoni has drawn the most defensible conclusions. Cangrande, indicated as "victorioso," and "in urbe Verona et civitate Vicentie Vicario generali" in the salutation, had conquered Vicenza in April of 1311 and had been made imperial vicar in the two cities by Henry VII himself on 12 February 1312. Thus, concludes Mazzoni, the date of the Epistle may with comfort be moved back from Mancini's 1319 to sometime between 1315 and the end of 1317, as Dante arrived in Verona with the Paradiso probably only just begun. This dating still seems to me the best hypothesis; if Dante did indeed write the Epistle, it is the simplest and accords well with the facts that we know about Dante's life and the composition of the poem, as well as the context of the Epistle. (Against Nardi's assurance that the [authentic] first four paragraphs are to be assigned to 1319, or at least before August of 1320, after Dante had finished Paradiso, it must be objected that we have no sure sense that the poem was finished then [it might have been], but, more tellingly, that it would have been very strange for Dante to have dedicated the poem and the Epistle to Cangrande while he was so happily lodged in Ravenna by Guido Novello.) One question that results involves the temporal relationship between the parallel passages in the letter and the poem. It seems to me a justifiable working hypothesis that Dante, writing the middle canto of Paradiso in praise of Cangrande, would have wanted to guarantee the earnestness of his praise in the earlier Epistle by citing lines from it in the poem, by certifying, as it were, the Epistle as "genuine." This far from unlikely hypothesis is greeted with derision by Brugnoli, who doubts either that Dante would have cited from his earlier works in the Epistle or, even worse, cited the Epistle in Paradiso XVII. Only if one looks upon the Epistle as inept, unDantean, or indeed as a forgery, is there any logic in this position. If the work is by Dante, there is no reason in the world for him not to have cited it, every reason that he should have. As for Brugnoli's first objection, had we not seen all those other self-citations, observed first by Moore, would we not be hearing from him and from others that the absence of these proved that Dante did not write the dedication? His is a curious argument: there is reference to a lot of earlier Dantean text in the Epistle, therefore it is probably not by Dante, but by a "forger." And how many likely candidates can we think of, who knew intimately not only the Commedia, not only the Convivio (which, as far as we know, had little or no early divulgation), but the Monarchia as well? Is a later "forger" combining in such subtle ways texts from three works of Dante? Or is the master citing himself? The latter seems a far more likely view of the matter, especially since many of the self-citations are subtle enough to have gone unobserved until Moore saw them and, in any case, are not put forward with the kind of clumsy notice that a forger would nearly certainly use, lest the "self"-citation go unnoticed, e.g., "ut in Convivio dixi," etc. In a slightly different context, E. K. Rand, writing in 1910, has pointed out that, if the forger had chosen to take Dante's prose style in De vulgari Eloquentia as his model, he would not have used ego (the first-person form that marks all such utterance in the Epistola), but nos, used to indicate himself thirty-eight times in De vulgari Eloquentia, while ego is used a single time, and then to denote not Dante himself but an exemplary "me." Moore's words are still worthy of consideration: "Would any forger be likely to have not only general culture and knowledge so extensive as these quotations indicate, but also so nearly co-extensive with that of Dante himself, especially in reference to Aristotle? Would any be likely to go behind the quotations offered by the writings of the author he was imitating, and study the original works that the latter was acquainted with in order to find fresh quotations? Again, would he be so crafty, or so venturesome, in the rare cases when he did reproduce any reference or quotation from that author's other works, as to vary the treatment of it in a manner that would give colour to the charge of inconsistency with those works? It seems to me that throughout we find just that amount of similarity in method and thought combined with independence in language and treatment as we should expect in the same author writing at different times and on different subjects."
There is one other particular in the epistolary section to which I would like to turn. In section 3.11 the author offers to dedicate the Paradiso to Cangrande. It is referred to as sublimem canticam. Brugnoli's gloss takes considerable trouble over the first of these words, citing several earlier glossators who opt for the reading ultimam canticam, and arguing for a certain ambiguity in the term, which he thus translates as "suprema," with the sense of "last and highest," following Pe'zard's translation. However, the parallelism between the qualities of the recipient and those of the gift, first noted by Vandelli in 1901, the *preheminentia* of Cangrande and the sublimita; of Paradiso, erodes the base of support from under this possible reading.
The word on which I want to focus attention is not so much sublimem as it is its neighbor, canticam. And here we have some new evidentiary material in the form of two recent articles by Lino Pertile (published in 1991 and 1992). It is at the very least possible that this is only the second use of the word (rare enough in early Italian; Pertile has found only one use prior to or coeval with Dante's in Purgatorio XXXIII, 140) to describe a part of the Commedia. What Pertile argues is that the commentary tradition, with the single near exception of Guido da Pisa, is a great deal less secure in dealing with this term than is the Epistola. His findings may be summarized as follows: Among the earliest commentators only Guido da Pisa uses the term "in senso proprio," while several (Jacopo Alighieri, Graziolo Bambaglioli, the Anonimo Selmiano) do not use it at all. Others (Jacopo della Lana) use it sparsely, while still others use it in some but not in all versions of their commenti (the Ottimo, Pietro di Dante). With the exception of Guido, according to Pertile, those who do use the term do so uncertainly. From a consultation of the Dartmouth Dante Project (and it must be remembered that Francesco da Buti is only the most illustrious early commentator missing from that database) I have been able to check Pertile's results, and find that, while his main point stands, they are in need of slight adjustment. For instance, Jacopo della Lana uses the word "correctly" some ten times; Guido da Pisa (27 occurrences in his commentary, limited as it is to Inferno) does in fact offer the most dense use of the term in the early commentators; on the other hand, the Ottimo, in his first redaction, displays 42 uses (if only ten in his commentary on Inferno). In all others, it is used only sparsely, until we come to Benvenuto da Imola. He never uses the term until it appears in his comment to the word's first appearance in the poem (Purgatorio XXXIII, 140); once it is "authorized" by Dante's own use, it appears seventeen times in the commentary to Paradiso. Zygmunt Baranski has now pointed out to me that Benvenuto in fact preceded his pupil in the observation, which occurs near the close of his accessus, and thus in a position similar to that in which we find John's. On the basis of this evidence, it would seem that Benvenuto only composed his accessus after he had finished his commentary, for there is no occurrence of the word cantica between its occurrence in the accessus and in the commentary to Purgatorio XXXIII. Pertile goes on to show that it was only with John of Serravalle (in fact, Benvenuto da Imola) that a somewhat better awareness of what the term implies (and thus why Dante might have chosen it) begins to take hold, indicating the relatedness of the term to the Canticle of Canticles: "Sicut liber Canticorum Salomonis dicitur cantica, per excellentiam, propter modum loquendi . . . ." (Benvenuto's words are the nearly certain source, given his authority for the commentator who on five occasions refers to him as his "magister": "Hic liber [the Commedia] merito appellatur Cantica; sicut enim in sacra Scriptura quidam liber Salomonis appellatur Cantica Canticorum per excellentiam, ita iste liber in poetria." Pertile's argument concludes with his doubts that Guido da Pisa, the only early commentator, in his opinion, who uses the word more or less confidently, is the source of the passage in the Epistola; Pertile is instead of the opinion that the Epistola is Dante's. His case, which can be countered only if Guido is the source of Henry Kelly's "Pseudo-Dante," is further strengthened when we consider that the word is used in the Epistle not only in the accessus, whence Pertile cites it (section 9.26; it appears there twice), but here in the dedication, once again in the accessus (section 13.37), and finally in the exposition (section 17.43). Where some early commentators, in the face of this strange new term for a part of the poem, blench, the author of the Epistola a Cangrande is uniquely sure in his use of it, and is so throughout his opusculum, still another sign that the author of the dedication and of the rest of the work is one and the same.
Turning to the accessus of the Epistle (sections 5 to 16), I shall be brief, not because I am not interested in the problems it presents, but for three reasons: first, because there has been so much written about it; second, because my own views on the central issue we find in it, the question of the nature of the allegorical procedures of the poem, have been expressed, and at some length; third, because my disagreements with Brugnoli can be put forward fairly briefly.
Section 6.18: Concerning the infamous sex . . . inquirenda Brugnoli cites Nardi for notice that Curtius had already shown that these terms dated back to antiquity, to Boethius and others, and that there was therefore nothing original in this schema. Brugnoli's tactic here is doubly culpable. First, it was Mazzoni who first discussed the significance of Curtius's treatment of the sex . . . inquirenda; second, Mazzoni did not claim they were original, only that Dante's use of them was entirely his own, since he did not link them to the Aristotelian "four causes," as every other medieval accessus to the Commedia which he had studied did. Mazzoni's claims for the boldness of the Epistola a Cangrande are always based on his sense of the understanding which Dante brought to familiar materials, the way in which he gave them a radically new point, not on the originality of the materials themselves. (It is exactly this sense of the cultural sophistication of the author of the Epistola a Cangrande which comes through the pages of the recent studies of the text by Martinelli, Paolazzi, and Placella.)
Section 8.23: The passage begins "manifestum est quod duplex oportet esse subiectum, circa quod currant alterni sensus." This is translated by Brugnoli as follows: "Š chiaro che il soggetto di un'opera, sottoposto a due diversi significati, sar… duplice." The middle clause in that translation may be misleading. A better translation would run as follows: "Š chiaro che il soggetto [dell'opera], circa il quale corrano alterni sensi [= i sensi letterale e allegorici], deve essere duplice." ["it is clear that the subject, around which run the one sense and the other, must be twofold."] The discussion of the four senses in the preceding paragraph of the Epistle makes the meaning of this sentence clearer than some contrive to find it. The first sense is the literal; the second set of "alterni" senses comprises the allegorical, moral, and anagogical, and it of these that the writer speaks when he subsumes them under the term allegorice in the following sentence. That term has caused much confusion, beginning with Nardi in 1961, in recent discussions. Brugnoli here cites Nardi to the effect that the subject of the poem is Dante's voyage and not the state of the souls after their deaths. Nardi's objections reveal that he really did not fathom how the Epistle meant its reader to understand its deployment of theological allegory; his complaint, strange as it is, is that the Epistola reduces the literal sense of the poem to a "semplice finzione poetica." That, in my opinion, is not a competent reading of the text. The scandal offered by the Epistle is precisely in making this claim; and that is the very mark of its Dantean origin (or, should we one day discover that Dante did not in fact write it, of its consonance with Dante's thought). Maria Simonelli has proposed the odd argument that, while Dante did not use the allegory of theologians for fear of the Inquisition, he did in fact write the Epistle (in which he claims that he did), a far more dangerous text for him to have set his name to. There is no overt claim for a theological allegoresis in the poem, while the claim made in the Epistle is manifest, and hence potentially extremely dangerous.
Section 9.27: Of the ten terms of the "forma sive modus tractandi" Brugnoli says: "The schema adopted is not identifiable." Yet the same Curtius, so recently lauded in the words of Nardi cited by Brugnoli, and who has discussed the significance and general provenance of the list, is now not mentioned. Of all the terms found in the list, the first two, "poeticus" and "fictivus," have caused the most unrest. These have served both opposers of authenticity to show either that Dante couldn't have written that part of the Epistle because the poem is veracious vision and not a fiction (for instance, Nardi) and those who claim that even if Dante wrote the Epistle it does not claim theological truth for the Commedia, which is only another poetic fiction (for instance, John Scott). Against Nardi, one argues that the poet knows perfectly well that there are fictive elements in his work, whether in the larger sense that he made it, was its fabbro, or in the more limited one that some parts of it are fabula (as contrasted with the central narrative, e.g., the passages marked off as being "sotto il velame"); against Scott that the ten terms are none of them all-inclusive, but identify various techniques used by the poet from time to time. How else, for instance, would Dante have responded if someone asked him whether or not he "believed in" the characters found in classical fiction that he includes in his text (e.g., the Minotaur, Geryon)?
Section 10.28-30: We come to the heart of Brugnoli's commentary. As we have seen, Brugnoli devotes the lion's share of his discussion to this paragraph. As are others, he is sorely troubled by the theory of the genres that is put forward in the Epistle. Since I disagree with nearly every part of his exposition, it would require a lengthy treatment to deal with it in toto. Rather, I will deal with a single detail. Brugnoli puts an important piece of his pleading in the following terms: "It is doubtful that Dante, if he was the author of the Epistola, could have claimed that he had chosen as the title of his work Comedia . . .; along these lines he would have had to call comedia the Aeneid as well, which, on the contrary, is said to be 'alta tragedia' at Inferno, XX, 112-3." This claim is only true if Dante did in fact think of the Aeneid as having a happy ending. Since I have had my opportunity to make the case for the obverse point of view, I do not here choose to argue against Brugnoli's interpretation, only against his critical procedure. He has assumed as true what is in fact the heart of a dispute, one perhaps even more important than the question of the authenticity of the Epistola, the complex problems that are found in the study of the components of Dante's understanding of Virgil. On this particular point I will merely note that Francesco da Buti, for one, believed that the Aeneid was a tragedy, and not for reasons of style alone.
Brugnoli's commentary to the Epistola a Cangrande has left its mark in recent work on the Epistola in England and America perhaps more than it has in Italy. I would like to suggest a reason for this, namely, that he failed to deal with the details of Mazzoni's argument so as to show that it was the weaker. For, while many Italian dantisti do not choose to commit themselves on the question of the authenticity of the Epistola, their silence tends to show that they have been impressed by Mazzoni's evidence. It is in the details that Mazzoni remains effectively unchallenged by Brugnoli. Dante, like God, is in the details.
(2) The Evidence Offered by Dante's Use of the Cursus
In this second lecture, which will run its course more quickly than the first, I will offer a response to arguments advanced by Peter Dronke, Henry Kelly, and in the joint effort by Ralph Hall and Madison Sowell. All three of these contributions have a common aim, to demonstrate that the accessus and exposition found in the Epistola a Cangrande do not bear the tell-tale sign found in all of Dante's other letters, including the epistolary portion of the thirteenth Epistle itself. That sign is the presence of the cursus, the cadenced conclusion of period and/or clause, the stylistic mannerism particularly dear to the epistolists of the high middle ages.
We will not descend to the level of micro-analysis of Dante's use of the cursus, but will concentrate on the arguments themselves, analyzing the adequacy of their logical procedures and the applicability of their findings. If the commentary portion of the Epistola a Cangrande is found to lack the nearly total recourse to the cursus (and this is not a matter in dispute) that characterizes all of Dante's other letters and the epistolary portions of this epistle itself, is that fact reason to believe that Dante did not write it? Or is it, we should add (as Dronke, Kelly, Hall and Sowell have not done), reason to believe precisely that it is?
Among those who currently attempt to find grounds upon which to base an assault upon the authenticity of the Epistola a Cangrande there has been a renaissance of concern about the cursus, more particularly, the relative lack of rhythmic periodic (and/or clausal) closure in the accessus and exposition of the Epistle. I say "renaissance" because some of those who have shaped this argument are apparently unaware of the extent to which it is not a new subject in Dante studies. It is Peter Dronke, in the monograph which he published in 1986, who has centered current interest on the argument that, on stylistic grounds, the commentary portion of the Epistola could not have been written by Dante. Whereas Dronke's first chapter attacks authenticity in the more usual ways, his Excursus is devoted to the question of the cursus and its relative absence from the non-epistolary parts of the Epistola. He is followed by Kelly, who, in his book of 1989, also finds that the relative failure of the author of sections 5-31 of the Epistle to use the cursus in either the accessus or the exposition supports his belief that this portion of the Epistola is not by Dante. The same basic understanding is found in the article by Hall and Sowell published in 1989. What is striking in all three of these studies is their lack of balance and of dispassionate attitudes toward the phenomena which they are examining. Hall and Sowell may particularly be faulted for the polemical tone of their study, which opens with a fairly vitriolic attack on the motives of the sustainers of the Epistle's authenticity, especially those Americans who are "tenacious" in maintaining their conviction that the Epistle is genuine, "in spite of," in their words, "compelling, if not overwhelming, clausular evidence to the contrary." It seems fair to put the question this way: if there is a major disparity between the way in which Dante uses the cursus here and in his other "similar" writings, would that fact in itself indicate that he is not the author of the treatise? And there is a second question, of even greater interest. Is the practice of the author of the final two sections of the Epistola a Cangrande in fact so very greatly different from Dante's in his other non-epistolary Latin writings?
What has not occurred to any of these recent discussants is that the facts that the first four paragraphs of the Epistle are almost entirely rhythmically cadenced and that the following ones are considerably less so may be the result of the author's turning from the epistolary style to that of commentary. They are all aware of the argument that the change from epistolary to expository prose may be significant, and deal with that problem by comparing the prose of other expository works by Dante (De vulgari Eloquentia, Monarchia) with the longer part of the Epistle (with less than convincing results, as we shall see shortly). On the other hand, this turning from the style of epistolary prose to that of commentary (not merely that of other forms of expository writing) seems to me to frame the crucial question, and it is not even raised. (On the other hand, we have Mengaldo's article on the cursus in the Enciclopedia dantesca, describing the clear distinction between "il grosso dell'Epistola, cioe` il vero e proprio commentario... e l'inizio," with "[il] suo taglio e linguaggio piu` propriamente epistolari....") ["the body of the Epistola, that is, the commentary proper,... and the beginning" with "its style and language that are more fitting to epistles..."]. It is instructive to consider a point made by Hall and Sowell; they admit that Dante did in fact resort to the cursus to a lesser extent than he did in his Epistolae in De vulgari Eloquentia and Monarchia; nonetheless, "the forger chose to include his methodology for analyzing the Commedia in a corpus of letters, and it must be judged by the standards of that corpus, wherein it is found to be wanting." This dire judgment not only makes Dante a conscious compiler of a "corpus" of letters more than is warranted, it, still more significantly, completely neglects any consideration of the Epistola a Cangrande's clear signal of a change in genre at the end of section 4, with the phrase sub lectoris officio, not to speak of the previous precision, between the poem that is to follow the Epistle and its opening epistolary gesture, that is found at section 3.11: "et illam [sublimem canticam] sub presenti epistola, tanquam sub epigrammate proprio dedicatam." For a far different view we may resort to the firm comment of Vandelli (1901): "Non <&grave e;> dunque vera e propria lettera tutta la parte che segue fino al 31o, dove, infatti, si parla in terza persona" ["All the part that follows, up to the thirty-first section, in which , in fact, the text uses the third person, is therefore not truly or properly an epistle"]; and that of Parodi (1912): "soltanto in essi [the first four paragraphs] consiste la vera epistola" ["the true epistle consists only of these" {the first four paragraphs}]. However, even the authenticity-opposing D'Ovidio had seen the point in 1899: "Qui finisce, dice, la parte veramente epistolare; adesso salgo in cattedra... e faccio sull'opera mia quel che farebbe un qualunque lettore, ossia professore" ["Here finishes, says he, the truly epistolary part: Now I leap to the lectern... and I make of my work that which would no matter what lecturer or professor"]. His sarcasm is responded to by Torraca in the same year: "Il D'Ovidio vede bene la differenza; ma quasi se ne maraviglia. A me il cambiamento, prova di sincerit…, porge buon indizio dell'autenticit…" ["D'Ovidio sees the distinction clearly; yet he almost seems surprised by it. For me the change, a proof of sincere intention, offers a clear indication of {the letter's} authenticity"]. The phrase sub presenti epistola also indicates that a line is drawn between the "superscription" (epigrammata) that constitutes the opening four paragraphs and the rest of the text (and, eventually, the text of Paradiso). (In 1989 Paolazzi followed Vandelli and Moore for Uguccione da Pisa's definition of epigramma as "superscriptio" and its implications. Paolazzi paraphrases the passage as follows: "sub brevi annotatione eorum que diffusius dicuntur in sequenti opere" ["beneath the brief formulation of those things which are said more copiously in the work that follows"]. One is reminded of Castelvetro's reference to the "soprascritto della pistola mandata a messer Cane della scala.") Since Dante's only other auto- commentaries are in Italian (contained in Vita nuova and Convivio), we have no means of checking the cursus in the Epistle against Dante's Latin stylus commentarius. Have any of the discussants taken the trouble to study the cursus in the Latin commentaries to Dante? That might have been an indicative exercise, if not one that could offer binding results. None has done so (except for Kelly, briefly and inconclusively). The relentless counting and assembling of tables in Dronke, Kelly, and Hall and Sowell, in light of the surprisingly diverse and hence inconclusive results they have drawn from these tabulations, have doomed their individual and combined efforts to use the cursus to undermine the authenticity of the Epistle. Anyone will admit, even insist, that the use of the cursus in epistles is always very high in Dante (as has long been acknowledged) and in his time (Table 6 in Kelly's book offers some confirming data), while I would consider it likely that, in Latin commentaries on poems, it is considerably less present. Kelly has a single tabulation of such as these (Tables 3 and 4), analyzing the prologue of Guido da Pisa and the introduction of Pietro di Dante. These samples are not as valuable as Kelly might hope for two reasons: the material is not very extensive and the nature of exordia is such that one expects to find more stylistic flourish in them than in their continuations as line-by-line commentaries (such as we find in sections 17-31 of the Epistola). (Dronke reports that he has left "the more rhetorical opening chapter" of Monarchia out of his accountancy for exactly this reason. Much earlier in this debate, we find Di Capua's discussion, in 1919, of Dante's use of the cursus in his Latin works; this begins by pointing out the relative lack of rhythmic closure in Monarchia. For a similar point we may turn to the sensible discussions in Toynbee (1920) and Rajna (1932), both of whom point out that in non-epistolary prose the cursus tends to be employed more fitfully, a conclusion borne out by the tabulations of Dronke, Kelly, as well as those of Hall and Sowell, but not admitted by them as having any effect on the question of the Epistle's authenticity.) It is especially interesting to note that the concluding period of the Epistle (section 33.90) is examined by Kelly and found to contain eight clauses, all of which, according to him, end rhythmically. It is at least worth considering that the author of the Epistola a Cangrande wanted to remind us that he knew perfectly well how to employ the cursus uniformly, but had chosen not to in the commentary portion of his work (I shall return to this point shortly). (In exactly such a vein, writing in 1912, E. G. Parodi had this to say: "E concludiamo che neppure in questa tanto discussa e maltrattata e ingiuriata Lettera, nulla v'e`, rispetto al 'cursus,' come non v'e` in alcuna di tutte le Lettere esaminate, che sembri menomamente scostarsi dall'uso e dal gusto dantesco" ["And let us agree that not even in this so debated, misunderstood, and maligned Epistle is there found anything, with regard to the cursus, as there is not in any of all of {Dante's} Epistles that have been studied, that seems even slightly foreign to Dante's taste and practice"].)
Let us return to Kelly's discussion of the parts of the two Latin commentaries that he has chosen to analyze. The eighty-seven periods in the prologue of Guido da Pisa show a maximum of 75% of rhythmic cadence; the 29 of Pietro's introduction a maximum of 72%; when all the clauses of these samples are construed in this way, the figures are as follows: of Guido's 259, a maximum of 66%; for Pietro's 101, a maximum of 62%. What of the Epistola?: Here are Kelly's figures: Accessus (sections 5-16), 102 periods, maximum of 56%; 229 clauses, maximum of 58%; Exposition (sections 17-32), 53 periods, maximum of 59%; 226 clauses, 53%. In sum, the shortfall is not of so cataclysmic a nature as our counters of cursus might like us to believe. Nor, because the samples are atypical (occurring in prologues) and too brief to be conclusive, do we have any reason to believe that such comparison is meaningful in the first place.
And there are some other troublesome problems that confront us when we look at the results gathered by the three cursisti, one of which is that their results are not uniform. Dronke, whose tabulations are, even according to those sustainers of his basic position, Hall and Sowell, based on faulty methods of comparison (although it should noted that Dronke, unlike either Kelly or Hall and Sowell, at least studied all thirteen of the Epistolae currently attributed to Dante, and not only Toynbee's ten), begins by opposing Lindholm's opinion that the discursive parts of the Epistle are not marked by the regular use of cursus found in sections 1-4 because "technical expressions occur that cannot be fitted into a rhythmic system" (Lindholm's formulation as cited by Dronke). (On the other hand, Toynbee, Rajna and Mengaldo all sustain the point, general to all of Dante's non-epistolary Latin writing, that his use of the cursus in all but the most rhetorical passages (particularly exempted are the exordia) is sporadic. For Mengaldo the clausulae of even De vulgari Eloquentia (Dante's most rhythmic non-epistolary Latin work) "appaiono nel complesso prive di ogni carattere di sistematicit…" ["in their totality appear to be without any systematic character"].) Dronke goes on to assert that the presence of rhythmically cadenced endings of periods in the De vulgari Eloquentia and Monarchia shows that Lindholm is incorrect in drawing that conclusion. Taking 96 sentences from the Epistle and the same number from the treatise on the vernacular, Dronke discovers that De vulgari Eloquentia contains 55 periods marked by one of the three kinds of cursus (velox, tardus, planus), the Epistola only 26. Proceeding in a similar manner, he takes 96 periods from Monarchia and gets a less stirring result; in fact, it is a catastrophic one for his argument. (Perhaps for that reason he has not created a Table for this part of his research, where he had done so for more supportive evidence.) In Monarchia he discovers 15 velox and 7 tardus, but no examples of planus. Thus this sample shows 22 rhythmic period-endings, four fewer than he found in the Epistle. Not one to be defeated by the evidence, Dronke twists it to his purpose. He has not even the grace to mention the discomforting nature of the results. Here, in toto, is his evaluation of his own evidence concerning Monarchia: "While the number of velox cadences here is smaller than in the De Vulgari Eloquentia sample, it is still more than twice that in the Cangrande exposition, and the proportion of velox to tardus closely resembles that in De Vulgari Eloquentia" (as though it had been his purpose to show that Dante had written both De vulgari Eloquentia and Monarchia). Nowhere in view is the fact that there are fewer examples of cursus in his sample from the Monarchia than in that from the Epistle. And whatever happened to the planus? Why is it totally absent in the periods examined in Dronke's sample? We are told not a word. This example is not offered to indicate the insufficiency of Dronke's statistical method, which is fairly evident; even had his results worked out better, we would, along with Hall and Sowell, have to reject them because the samples on which they are based are not significant. And of course I do not suggest that Dronke has shown that the Epistle is likely to be Dantean because of these figures. I wish only to underline the tenacity with which supporters of the inauthenticity of the Epistola a Cangrande hold to their cause, even when it is undercut by their very own evidence. Dronke does not discuss Monarchia in the remaining pages of his argument (we can readily understand why not), but goes on to consider other relations between cursus in De vulgari Eloquentia and in the Epistola. However, he never adduces a reason for his having deserted his search through Monarchia for similar textual evidence. This is all the more surprising when we consider that, no matter when we date that text, it is without significant doubt (and this is an understatement) closer in time to the composition of the Epistola, if we grant its authenticity, even only for purpose of argument, as Dronke does, than is De vulgari Eloquentia. As I have suggested in the penultimate chapter of my monograph, my sense is that Dante's prose in the political work is not in fact dissimilar from the prose of the more extensive passages of the Epistle. And that opinion is given considerable support by the neglected essay of E. K. Rand (published in the Annual Report of the Dante Society in 1910). Rand, working through Dante's Latin works, as he and Wilkins prepared their concordance, came up with a series of dates for these works that is convincingly argued and in keeping with the most impressive contemporary scholarship on the issue. He dates De vulgari Eloquentia to 1304, the Epistola a Cangrande to 1316-1319, the Questio to 1320, and Monarchia to 1318-1321 and in any case posterior to Paradiso V. His study remains the closest stylistic consideration of the question of the authenticity of the Epistola that we have, full of carefully handled evidence that justifies his conclusion: "This evidence is enough to refute once and for all the hypothesis that [the Epistle] and [the Questio] are forgeries; coincidences so numerous and minute could have been vouchsafed a forger only by plenary inspiration, proceeding in this case from the Father of Lies." That claim, failing supporting evidence, will convince no one. I have found no discussion, however, of the considerable evidence offered by Rand in any who argue against authenticity.
A still more disturbing absence from the contributions of all the recent discussants of the use of the cursus in the Epistola a Cangrande is the Questio de aqua et terra. (On the other hand, this is a subject that has occupied other less partisan students of the problem of the Epistle's authenticity, among them Toynbee and Di Capua, the latter examining the similarity in the sporadic uses of cursus in the Questio and the Monarchia, and Mengaldo.) This is a work that most now accept as genuine, a neighbor in time (ca. 1319) to the Epistle (I realize that there may still be some who choose to doubt its ascription, but they do so at some peril, given the nature of the evidence advanced by many and most recently by Mazzoni in 1979, who demonstrates that Pietro's commentary cites the Questio literally and as his father's work). We find in it a pattern of the use of cursus that is very like what we find in the Epistle. It is surprising to find that none of those who have recently taken up the problem has attempted to study the Questio for what it might reveal. (Father Paolazzi, on the other hand, has recently insisted on the importance of Mazzoni's evidence; and we have the words of Giuseppe Billanovich, which date back to 1966, to similar effect: "Veramente l'epistola a Cangrande, la Questio e il Paradiso furono scritti da una sola penna e negli stessi anni!" ["Indeed, the Epistle to Cangrande, the Questio and the Paradiso were written by a single hand in the same years"].) In fact, the clausular procedures found in the Questio support the case for the authenticity of the Epistola.
To conclude his unconvincing treatment of the relationship between cursus and Epistola, Dronke puts forward a last sally in order to question the authenticity of the rest of the Epistle (we must remember that Dronke says he believes, as did Nardi, that the dedicatory portion is genuine). After noting that Brugnoli has pointed out that both families of manuscripts derive from an exemplar that contains that jarring (to all those who oppose authenticity) linking sentence, in which the author prepares to assume the role of commentator on his own text, he responds as follows: "If the dedication is genuinely by Dante, was the writing of this link sentence [with, we remember, its glaring sub lectoris officio] an act of deliberate deception, aimed at passing off the introduction as a work of Dante's own? Yet it seems to me at least possible that the writer of that sentence was making what he thought to be a necessary 'editorial bridge': if he believed that the dedication and the fragment of commentary --which he may well have come across copied side by side-- actually belonged together, he might have honorably supposed that he was merely filling in a lacuna, rather than doing anything more fateful --that he was clarifying, rather than causing confusion." If this wildly problematic thesis were to be accepted --and there are no grounds on which it can be-- we are left with a still more puzzling problem in place of the one with which Dronke begins. Why do the three manuscripts containing only the dedication all include the supposedly bogus ending? Why would any copyist cease his labor at so extraordinarily awkward a point, leaving thereby a document, purporting to be by Dante, with so much hanging in the resulting void? The least embarrassing answer is that he was copying from a complete version of the whole text, but was only interested, either for himself or because of the directions of the person for whom he copied it, in the epistolary relations between those imposing figures, Dante and Cangrande, and not in all that literary critical rot. The best hypothesis available is that now held by Francesco Mazzoni. I am grateful to him for sharing it with me. One of the three truncated manuscripts is found in a collection of letters belonging to Bernardo de Vallibus, who evidently assembled it (or had it assembled) as exemplary of epistolary excellence. To such a collector, the rest of the Epistle was simply not "epistolary," and thus of little or no interest. Bernardo, or one like-minded before him, presented with the Epistle in its entirety, would have copied (or had copied) only the first four paragraphs. The evidence of the manuscripts, if it suggests anything at all, would rather make us believe that the sentence was found in the original text, which even Dronke (if not Kelly) takes to be Dantean. Therefore, the linking sentence must be excised from the "authentic" brief epistle (indeed, this was Mancini's earlier tactic, of which Dronke is apparently unaware) and the question is moved to new ground: was the falsario a hoodwinker, taking on the identity of Dante, or was he an earnest expositor, doing only what seemed the intellectually honest thing to do? The manuscript tradition that we have (whatever its weaknesses, it is all that we possess) would suggest an entirely different scenario.
Hall and Sowell present what is certainly the most developed and challenging of the three recent treatments of cursus in the Epistola. They, too, are committed to Nardi's thesis that the first four paragraphs (and only the first four paragraphs) are genuine. Their analysis of the cursus in Dante's Epistles (like Kelly, they include only the ten published as probably genuine by Toynbee, thus neglecting three epistles considered authentic today; this makes no difference to their result, if it does raise other questions) shows the following result: I-IX & X.1-4: 99%; X.5-33: 70%. It is interesting to compare these figures with Kelly's for the accessus and exposition of the Epistola a Cangrande: he counts 102 periods, 36-40% of which are rhythmical, while Hall and Sowell count 101, 70% of which are rhythmical. (Dronke found 96 periods which do not end with citations in the same parts of the document, of which only 36, or 37%, are rhythmical.) These disparities among the three cursisti are a bit unnerving. (It is in part for this reason that one looks forward to Mazzoni's new edition of the Epistola a Cangrande, which will contain a more definitive examination of its use of cursus. It is the burden of Orlandi's review of Janson's book, published in 1975, that accurate accountancy of the use of cursus is a difficult matter. As Rand has said: "Stylistics and statistics... must be handled with the greatest caution.") And that is not all, as we have seen, that is unnerving. Hall and Sowell move to the culminating phase of their argument. With those tenacious supporters of the authenticity of the Epistola a Cangrande evidently in mind, they gently hector us: "It would certainly seem to be more difficult to maintain that Dante was the author of both I-X.1-4 and X.5-end, for it strains the credibility to accept the proposition that Dante, after a series of near-perfect rhythmical sentence-closings in the bulk of the corpus of his Letters, would suddenly shift to a quite sloppy or even indifferent use of the cursus." They then spend several pages showing us how the Epistolist might with ease have changed his non-rhythmic clausulae to rhythmic ones. (We can imagine the Pseudo-Dante, his head hanging low in hell or heaven, sadly counting his numbers, realizing how close he came to fooling all those sharp cursus experts in the twentieth century.) As I suggested a few moments ago, the author of the Epistle knew perfectly well how to write employing clausular rhythm, but chose not to do so in a thoroughgoing way in the commentary part of the text. This is precisely the point made by Di Capua in 1919 (he is speaking of Monarchia, but his words apply equally well to the Epistola a Cangrande: "...spesso, con una semplice e facile trasposizione, avrebbe [Dante] potuto ottenere delle clausole regolari; il non averlo fatto mostra chiaramente che non l'ha voluto" ["...often, by means of an uncomplicated and easy transposition, {Dante} could have obtained regular clausulae; his not having done so clearly demonstrates that he did not wish to do so"]. We should remember that Hall and Sowell cite Di Capua's study. If they have cited it, have they read it? If they have read it, have they understood it? If they have understood it, why have they ignored its main point? We remember their opening assertion, referring to the "compelling, if not overwhelming, clausular evidence" of the inauthenticity of the Epistle. Yet exactly that evidence serves better to support authenticity than deny it.
After this "demonstration," like Dronke (but without his caution), they make a final and fatal mistake: they turn to the other Latin works to test their results. What do they discover? De vulgari Eloquentia contains 90% rhythmic clausulae, Monarchia 77%. They do not discuss the calculation of Di Capua that the percentage is, in fact, 67%, nor do they refer to Mengaldo's citation of Di Capua's statistic. (Dronke, who is sternly criticised for his incorrect procedures, came up with quite different results, based on his collection of 96 randomly chosen periods, 57% and 23% for De vulgari Eloquentia and Monarchia, respectively. I would agree that his procedure is invalid; on the other hand he was at least clever enough to distance himself from the Monarchia.) Hall and Sowell proceed by informing us that these percentages are well below those found in the Epistolae, but rightly point out that epistles are traditionally more the field of operation of the rhythmically intent than other forms of Latin prose, so that there is no major problem with this result. Or is there? If we happen to remember that our cursus-sloppy slouch, the author of the exegetical portion of the Epistola, at least according to Hall and Sowell, comes in at a Monarchia-challenging 70% (or, if Di Capua is correct with regard to Monarchia, the clauses of which he calculates to be 67% cadenced, that figure is slightly higher than what is found in the political treatise). In other words, Hall and Sowell have come closer to "proving" that, because of its relatively similar count of rhythmic periods when it is compared with Monarchia, the Epistola a Cangrande really is by Dante than they have proven the opposite. And with that final example of their use of questionable evidence, we may depart these precincts, unless someone would like, on the heels of their discussion, to open the question of the authenticity of the Monarchia.
As one who has never before evinced particular interest in Dante's use of the cursus and as one whose training in medieval Latin stylistics is minimal, I must say that I find the rather sloppy and surely illogical presentations of this subject by Dronke, Kelly, and Hall and Sowell difficult to account for. They have variously not consulted texts central to their arguments (not one of them apparently even took the trouble to open his copy of the Enciclopedia dantesca to the entry "cursus," which was written by one of the most expert living students of Dante's prose, Pier Vincenzo Mengaldo), distorted evidence or misinterpreted it, revealed a motive for finding the "correct" result whatever the evidence might reveal, and, in general, conducted their inquiries in such a way as to give aid and comfort to their opposers, for the reason that their procedures are so evidently and deeply flawed that whatever their conclusion is, it would seem likely to be false. When I first and hastily consulted Dronke's book I said to a colleague that Dronke had just proven that the Epistola a Cangrande was indeed written by Dante. That was an initial, waggish response. Having now studied the quality of the evidence put forward in these three recent studies, I think there is reason to conclude that that first response was not far from the truth.
(3) Theoretical Considerations: "allegory" and "comedy"
While more limited and discreet questions regarding style ("how could Dante have compared Cangrande to the Queen of Sheba?"), presentation of a grovelling self ("how could Dante have sought patronage in so servile a way?"), language ("how could Dante have translated his own vernacular verses into Latin?"), and interpretive adequacy ("how could Dante have written so lamely about the opening verses of Paradiso?") have occupied any number of the opposers of the authenticity of the Epistola a Cangrande, the two most provoking questions raised by the text have been and remain those concerning its generic distinctions ("how could Dante have so confused the nature of tragedy and comedy?") and its insistence on a theologically based allegoresis ("how could Dante have made so outrageous a claim?"). Like others (most notably Brugnoli and Baranski) who have recently opposed the case for the authenticity of the Epistle, Henry Kelly, in his recent book, Tragedy and Comedy from Dante to Pseudo-Dante (1989), claims to be most bothered by what seem to him the strange things it says about tragedy and comedy. (Hall and Sowell evince the preoccupation of the previous generation of opposers of genuineness: the Epistle's claim for Dante's use of the allegory of the theologians. Pietrobono [in 1937] bridged the gap between the two schools of objectors, spending two pages in fairly high dudgeon against Moore's argument that the things said in section 7 are acceptable in the context of Dante's beliefs about the theory of allegory, before becoming greatly disturbed by the notions of comedy advanced in section 10. Where supporters of its authenticity find that the remarks made in section 10 either support or else are not in opposition to the views of Dante as these are revealed in his poem, some opposers are driven into a state approaching paroxysms of dyspepsia, in this half of the century most notably Bruno Nardi. The argument is a weak one, and always has been. Its logic is simple and not critically mature: since Dante didn't believe such things about the nature of tragedy and comedy, he could not possibly have written this section of the Epistle. In my view, these detractors simply fail to understand the nature of the claims made in section 10, which are not at significant odds with Dante's views in these matters and even, in my opinion, are precisely congruent with them (a position well stated by Pio Rajna over seventy years ago).
Kelly summarizes his thesis as follows: "According to the hypothesis I have developed in this study, an unknown student of Dante's Comedy set to work sometime in the last quarter of the fourteenth century to create an introduction to Paradiso that he attributed to Dante himself. He made use of a preexisting Accessus to the whole Comedy, prefaced it with a Dedication to Cangrande, and followed it with an Exposition of the beginning of Paradiso. The resulting Compilation we now know as the Epistle to Cangrande." (An anonymous marginator in the Princeton library has added the following question and remark in Kelly's text at this point: "Why only Paradiso? only Dante would do that." Kelly never offers an hypothesis that would account for the putative forger's decision so to limit himself.) Kelly's dating of his "Pseudo-Dante's" effort ("sometime in the last quarter of the fourteenth century") is immediately in difficulty when one considers that Boccaccio's translation of a number of passages in the accessus was completed before 1375, the year in which he died. Since the accessus offered by Boccaccio is unmistakably identical in many of its statements with the Epistle, Kelly (who downplays the extent of the identical elements) must either date his forgery before 1375 or create another pseudo-document. It is not surprising that he chooses the latter course: "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." Now this is not impossible. Yet, as we shall see, several early commentators do attribute the Epistle to Dante (a fact Kelly either does not know or simply chooses to overlook). Would it not have been simpler, more "economical," for Kelly to have argued that the "forger" had written before Boccaccio? In any case, Kelly never makes the case for the necessity of a later forgery, since he denies Boccaccio the role of source for the document. And how does Kelly explain the fact that only Boccaccio's commentary gives the same examples for the illustration of the four senses of Scripture that we find in the Epistola a Cangrande, based on the 113th Psalm? That would indicate that the "Proto-Accessor" had them in his version of the accessus, which reasonably might be expected to have influenced the other chiosatori, but somehow did not, a fact that is, at the very least, perplexing. We are left, as we are over and over again in Kelly's book, with his postulations. I should note that my monograph on this subject spends considerable time examining the weaknesses in many of Kelly's major arguments; these will not be addressed today. Instead, I want to turn to the work of my friend Zygmunt Baranski. I hope I need not add that my strenuous disagreement with much of what he has had to say in no way diminishes my admiration for his work or my affection for him. I am not attacking him, but disagreeing with his position.
Baranski has recently devoted a lengthy article to our question (in Lectura Dantis [virginiana] spring, 1991). While it is mainly concerned with what he perceives to be the vast difference between the treatise and the poem with respect to their diverse reflections of a theory of the genres (in one of his more enthusiastic moments he characterizes this difference as a "fathomless abyss"), he concludes with considerations of the supposedly diverse modes of allegory he claims to find in sections 7 and 8. His work is well researched (only one major omission will detain us in a moment), thoughtful, lively, and provocative. Precisely because it is as attractive a critical performance as it is, it requires close attention.
As he warms to his argument, Baranski becomes less cautious in his judgment that the Epistle may not be by Dante, concluding with the following assertion: "To argue against the Epistle's Dantean origins is --I believe-- to perform a humble service on the poet's behalf." Let us examine his arguments in order to test the merit of his project. What follow are the causes of my principal disagreements with him.
(1) Baranski begins by asserting that Nardi "had convincingly refuted" Mazzoni's principal claims. Not even Brugnoli goes that far. It is already clear that I strongly disagree with this evaluation. Nardi had indeed made some telling points against several of Mazzoni's arguments. What he failed to do was to dislodge the evidence Mazzoni had given that the early commentators were better acquainted with the document than Nardi wanted to admit. It is also worth noting that Baranski pays little attention to Nardi's strong support for the authenticity of the first four paragraphs of the Epistle.
(2) While indicating that Brugnoli both studied with Nardi and continues to share many of his thoughts on the question, Baranski inexplicably fails to take any notice of Giorgio Padoan's important contribution, published in 1965, which opposes Nardi and supports the authenticity of the Epistle, even though Padoan also considers himself Nardi's student. What Padoan did was to counter Nardi's surprising miscomprehension (one shared by Brugnoli) that the Epistola's schema for the adaptation of the four theological senses of allegory made the poem a mere, or traditional, fiction, rather than the prophetic vision composed by Dante so passionately believed in by Nardi.
(3) Unsurprisingly, given his basic agreement with the thrust of Kelly's work, Baranski holds that "it is impossible to assert with any degree of certainty that, say, Guido da Pisa depends on the letter rather than vice versa." Only if one supposes that the Epistle is not genuine may one so argue, a tactic which makes of the matter in dispute the first tenet in one's argument. Those who oppose the Epistle's authenticity are the ones under the burden of proof in this matter. We are not dealing with a manuscript that turned up without ascription that a modern scholar has tried to assign to Dante, but one that surfaced in the hands of Filippo Villani, through what paths we know not, bearing Dante's name. If we return to Mazzoni's dating of the Epistola a Cangrande (between 1315 and 1317), then clearly Guido depends on it. And even Nardi, who believes the first four paragraphs genuine, only moves their date forward to 1319. Thus, as I have argued, if the clearly Dantean opening is accepted as authentic, and if we find Guido da Pisa and Jacopo della Lana referring to it and later parts of the Epistle at least by 1328, it is probably impossible to argue for Guido's (or anyone else's) prior composition of elements found in the Epistola. Baranski says that recent discussions have questioned the date of the Epistle. In recent years only Kelly's, however, has postponed it so drastically. And there is the question of the validity of these discussions, which is not addressed by Baranski.
(4) "It is not at all clear why, given their cult of Dante and their culture's obsession with authority" the commentators do not cite Dante as its author. To this observation is added the following challenge: "I'll be especially interested to see how supporters of the Epistle's authenticity would explain away a version of the letter which did not bear Dante's name." It is possible that Baranski's challenge reflects a similar moment in Brugnoli's introduction, in which Brugnoli reflects with bemused impatience that neither Jacopo della Lana nor the Ottimo, if they do cite the document, feels obligated to refer to Dante as its author, an argument confronted by Mazzoni, but treated by Brugnoli as though it had not been and were a major point on his side. (In fact, the argument had already been put forward by D'Ovidio in 1899, arguing that supporters of genuineness could not supply "alcun trecentista che citasse l'Epistola.") Neither Brugnoli nor Baranski cites or discusses Jenaro-MacLennan's substantial discussion of the way in which attributions and other indications of person tend to slip away in later versions of commentary material. Nonetheless, both of these questions raised by Baranski are important, and Mazzoni has dealt with them in what I consider intelligent ways, admitting, even insisting, that the Epistola circulated without its first four paragraphs, after having been transcribed in that form, perhaps by someone close to Cangrande. (Mazzoni is now of the opinion that there was likely to have been a copy of the exegetical portion made somewhere in the area delimited by Bologna, Verona, and Ravenna sometime shortly after Dante's death; this is a working hypothesis, not a fact; it is nonetheless worth recording.) We can, I believe, make the case, on negative evidence, that most of the fourteenth-century commentators who saw the text did not see it in full. Given the fact that so many deal with it in their proemi, we may speculate that it could have been copied into at least one manuscript as the accessus to the work, and thus neither clearly Dante's own nor clearly not his own. But that is mere speculation. All we can say is that it seems more than likely that the truncated form of the Epistle (no more than sections 5-33, probably considerably less) made its way into the hands of a number of commentators, whether or not of all who quote from it (since some may be citing from the texts of other commentators). Thus the lack of any citation of Dante as its author results from the absence of his name as author of the Epistle, which is only found in the dedication. There is a second point. As Padoan has shown, the early commentators (and particularly Pietro Alighieri) were extremely uncomfortable with the truth-claims made in the poem itself, which they frequently attempted to undermine by making the text a more "ordinary" kind of fiction ("here the author feigns that he saw . . ." is a typical locution in this mode) in response to Dante's continual representation of the state of the souls after death and his experience of them as being veracious. Thus the Epistola, with its similar claims, is a difficult document for most of the commentators to accept. At the same time, the fact that it had been cited by such as Jacopo della Lana, Guido da Pisa, and Dante's son Pietro gave it a kind of tentative authority. Thus it is used, abused at times, and held at a certain distance. It is almost always there somewhere in the early commentaries, but its authority is uncertain, its doctrines at times disconcerting or even dangerous. It has, in short, exactly the sort of liminal and simultaneously controversial status that its sporadic treatment would indicate. And, finally, we should observe that in fact Guido da Pisa does refer to the Epistle as being by Dante (in his commentary to Inferno XV, 69: "Et hoc semper in suis licteris ostendebat dicens: Dantes florentinus natione, non moribus"), that Jacopo della Lana does so as well (whether or not he was following Guido, as Jenaro-MacLennan believes, glossing the same text, when he wrote "si si scrivea Dante da Firenze per nazione, e non per costume," and that Paolazzi has now shown that Benvenuto is also in this company. He cites from the lectures given by Benvenuto da Imola at Bologna in 1375 (the version of his commentary the authorship of which was attributed incorrectly to Stefano Talice da Ricaldone in the nineteenth century): The passage addresses Dante's reasons for composing in the vernacular. "Ratio prima est ista, que habetur in sua epistola, ut faceret fructum et delectationem pluribus gentibus, tam literatis quam illiteratis" ["The first reason is this, as we read in his epistle, in order to offer fruition and pleasure to many, whether lettered or unlettered"]. Paolazzi argues for the closeness of the thought of the passage to that found in sections 10.31 and 15.39 of the Epistle, and to the phrase "utilitatem et delectationem" in section 33.89. To his argument may now be added Francesco Mazzoni's notice, in a recent conversation, of the same phrase, fructum et delectationem, reflected in the opening verse of the capitolo ternario of the Commedia by Busone da Gubbio: "Per• che sia pi£ fructo e pi£ dilecto." In addition, the phrase, in any declined form, "utilitate et delectatione" appears only once (strangely enough) in the Latin commentaries currently found in the Dartmouth Dante Project; it is used by Benvenuto in his comment on Paradiso II, 7-9. This fact may lend Paolazzi's thesis some support.
Baranski's supporting argument is more provocative and interesting. If Benvenuto was so interested in having the authority of Dante for his discussion, why did he omit the reference in the final version of his commentary? Henry Kelly had already, as Baranski notes, offered a complex and interesting discussion of Paolazzi's ascription of the reference to Epistola XIII to Benvenuto. In Kelly's view, the words sua epistola in this second version of the commentary may either refer to the letter from Petrarca about Dante referred to by Benvenuto in the final version of his commentary, where this passage is reelaborated in a discussion of Inferno II, or to Boccaccio's letter of "Frate Ilaro" about Dante. However, the passage in question is wholly unambiguous: "Ratio prima est ista que habetur in sua epistola." And the following text, while perhaps not faithfully rendering Dante's argument, is nonetheless to be taken as Benvenuto's version of it, which turns the Epistle's description of the poem's status as vernacular work into the commentator's understanding of the purpose that lies behind that choice. The simplest and most logical inferences that we may draw are, first, that in sua epistola refers to a letter by Dante (and not one about him) and, second, that the only epistle thus indicated is the Epistola a Cangrande. If these things are true, we are still left with Baranski's good objection. Why did Benvenuto withdraw the citation in his final version of his commentary?
Paolazzi, in the understandable pleasure of discovery, has not, it is fair to say, considered the strange practice of Benvenuto if he is both acquainted with the Epistola a Cangrande and knows it is by Dante. If these two things are true (and I think they may be), how do we explain the fact that his references to the Epistle are so very few and far between? And so faint when they do seem to be present? (Mazzoni, in 1955, spoke of Benvenuto as being "at the margins of the problem.") When we compare his accessus to the Commedia to Boccaccio's, perhaps the most faithful recapturing of the key elements in the Epistle of any fourteenth-century commentator, and a text which Benvenuto knew more closely than perhaps any other commentary, how can we account for his nearly total distancing of himself from the document? I think there is a reason: Benvenuto did not approve of what the Epistle was saying. Boccaccio treats it as authoritative (if not in nomine Dantis); Benvenuto seems to think it was by Dante, and desires to be free from its yoke (like so many contemporary students of the problem). This hypothesis would account for Benvenuto's only occasional, nearly involuntary, reference to it, and his suppression of his single earlier reference to it as Dantean in the final version of his exegesis.
Last, but hardly least, Filippo Villani still more explicitly attributes the Epistle to Dante. Does Baranski really believe that all this is not significant evidence? One assumes so, given his total silence about the two glosses to Inferno XV, 69, despite the fact that he has read the texts that put forward these pieces of evidence. His only response is to list the bibliographical sources of the various opinons that we have examined and to offer a blanket denial: the claim may only safely be made for Filippo Villani.
(5) The understated and "conventional" nature of many of the claims made by the author of the Epistle strikes me rather as arguing for Dante's authorship than against it. See, for example, his brief explication of the invocation at Paradiso I, 10-12 (section 31.87). All he says that is of any interpretive significance (and it is of considerable significance) is contained in the laconic remark that the author of Paradiso "petit divinum auxilium." We remember the related text at section 18.47, the first discussion of the invocation, where we are told that what must be invoked by poets is "quasi divinum quoddam munus" ["almost a certain divine gift"], a passage which allows a "pagan" understanding to those so minded, but which is now, it is clear, revealed to have referred exactly to the "divinum auxilium" of the true God. That the author of the Epistola a Cangrande here speaks briefly does not indicate that he speaks without portent. Continuing, at section 32.88, he refuses to develop the sententia of the invocation, precisely at the moment he would need to reveal to us that "Apollo" is the Triune God, that the Holy Spirit is his inspiration, etc. It is typical of Dante that here, as in the poem, he forces us to mouth his deep truths for him. It is a technique he may have learned from the Christ of the Gospels, as in his use of the locution "tu dicis" in order to render Pilate's unbelieving formulation of Jesus' kingship no less than truthful (Matthew 27:11; Mark 15:2; Luke 23:3; John 18:37).
(6) Brugnoli had previously offered the opinion that Dante would have never called the Commedia "sublime" because it is written in the low style, an opinion which fails to take into account so many things, whether in section 10.30 or in the Commedia itself (e.g., "poema sacro" [Paradiso XXV, 1], "t‰odia" [Paradiso XXV, 73]), that reveal to us Dante's desire to transcend the usual generic limitations of stylistic range and register. Further, if the accessus (section 10.30) holds that tragedy speaks "elate et sublime," the exposition only confirms what is said of comedy in the dedication: the opening verses of Paradiso promise to tell things "tam ardua tam sublimia" (section 19.51). It seems clear enough that the author of the Epistle, like the author of the Commedia, wants to have things both ways. The apparent contradiction is not one, but a brilliant paradox: the poem is a comedy that can attain the stylistic height and seriousness of tragedy without being "tragic." Precisely such an understanding is found in the Declaratio super profundissimam et altissimam Comediam Dantis, the vernacular poem in terzine composed by Guido da Pisa (around 1327). In its title and then in the text (vv. 7 and 23) Dante's poem is referred to as "l'alta Comed¡a." What seems to have been easily grasped by Guido causes considerably more difficulty in our time.
Baranski is another who is displeased by the Epistle's formulation of a theory of genre. Its equation of, and now I quote from Baranski, "the vernacular and the 'comic' ('ad modum loquendi, remissus est et humilis, quia locutio vulgaris in qua et muliercule comunicant') is most bewildering, when Dante's glorification of his native language in De vulgari eloquentia is remembered; to say nothing of the uses to which he puts the vulgaris in the Commedia itself." I think that Baranski here is caught on the horns of a dilemma of his own making. Since he wants the Commedia to be simultaneously "comic" (in the widest possible terms) and "plurilingual" (his term in a number of his recent studies), he tends to want to deny (for reasons I do not understand) the "tragic" as one of these styles. (To return to Guido da Pisa's formulation, alluded to a moment ago, Baranski somehow fails to understand that what Dante has composed may be described as his "alta Comed¡a.") At the same time he wants Dante's definition of the lofty vernacular from De vulgari Eloquentia, which is pronouncedly "tragic," to apply here. This objection contradicts Baranski's purpose as it is elsewhere expressed.
In the same vein, the phrase at section 3.11, "sublimem canticam," also seems to Baranski the work of a forger. Why? Because the term (and he here disagrees with a part of Brugnoli's argument, as do I), means not "last cantica" but "sublime cantica," that is, one of loftiness. I would say that the Paradiso is meant to be seen as "high" in style (if not always, essentially, and more so than the preceding cantiche) and "happy" in content. Baranski takes "sublime" to refer only to the "tragic" style. And that bothers him considerably. He cannot fathom how, in light of his other judgments, the author of the Epistle would say that "a comedy could actually be 'sublime.'" We have seen him caught in this bind before. If, as he himself insists, the Commedia is "plurilingual," will not that plurilinguismo include the high style? Reading the poem is enough to convince me that it does. I agree with Baranski that the poem contains many styles. Surely the tragic is one of them, as he himself elsewhere acknowledges. What is perhaps most exciting in Dante's "high style" is that it rests so easily alongside the sermo humilis that is his trademark. Let me, aware of how many might serve, offer only two examples to demonstrate the point. Dante says in the Convivio (2.2.7; 2.4.2) that angeli are only so denominated by the volgar gente, that, in fact, they really should be referred to as intelligenze. (His practice in the Commedia, unsurprisingly, is precisely to follow the volgar gente: the word angelo is used a total of thirty-one times, intelligenza, four, and only once [Paradiso XXVIII, 78] with a possible undertone of the angelic nature of the motion of each celestial sphere.) Such "vulgarization" is in keeping with what is perhaps, and rightly, our central perception of Dante's linguistic direction in the Commedia, from high to low. And I think it is right to conclude that that is what he himself wanted to be most notable about his new creation, its ability to use the vernacular for subjects that had hitherto been treated in "lofty" Latin. However, if we find him "lowering" scholastic Latin to its vernacular equivalent, we also find him "raising" the vernacular to its level, as we may conclude from such evidence as that commented upon by Torquato Tasso, cited by Raffaele Andreoli (1856), discussing Dante's use of a Latin negative where an Italian would have served in Paradiso XXXII, 145): "quasi giudicasse le parole latine esser piu atte ad esprimere la maesta e l'altezza de' concetti del Paradiso" ["as though he considered Latin words more fitting to express the majesty and loftiness of the ideas of Paradise"].
(7) Baranski argues that the citation of Horace in the service of allowing a mixture of styles in a single work is not really a very good reading of Horace, who allows only a limited mingling of genres. Dante, however, or the author of the Epistle, not having had the benefit of Baranski's counsel, simply decides that he allows as much he wants him to (Horace, he says, "licentiat aliquando comicos ut tragedos loqui, et sic e converso" ["allows that comic poets at times speak as tragic poets do, and vice versa"]). Baranski, differing from this opinion, argues that this use of Horace was a commonplace in the exegetical tradition of the middle ages, and thus is further proof of the passage's "conventionality," an attribute which Baranski assumes renders the Epistle more likely to be spurious, one supposes on the ground that some who have argued for authenticity regard it as "revolutionary." The Epistle, on the other hand, like the Commedia, is a work that combines the old and the new. If Baranski is right about the conventionality of the claim made on the authority of Horace, that is still no reason to believe it is not by Dante; if Vandelli, who believed that Dante's citation is part of a wholly original tactic, is right, there is every reason to believe that it is.
(8) When Baranski turns to the question of the allegorical exposition found in sections 7 and 8 of the Epistle, his main point is that the discussion of theological allegory in section 7 is followed by a description in section 8 of what is essentially "the allegory of the poets": according to him, this description "is rather bland and lacks any Biblical overtones." He then turns to Alastair Minnis's similar argument, which has it that the eighth section of the Epistle reverts to a more usual "allegory of the poets" in its statement of the "ethical" meaning of the text. The problem with the interpretation of Minnis and Baranski is, simply, that it is almost certainly incorrect. First, Baranski's omission of the first part of the eighth paragraph of the Epistola obscures the consecutive nature of the thought of the two connected statements found in sections 7 and 8. Having finished the description of the four senses in section 7, the writer links his next discussion to it: "Hiis visis, manifestum est quod duplex oportet esse subiectum, circa quod currant alterni sensus." As we have earlier seen, Brugnoli's Italian translation of this passage could be more accurate. I do not know whether or not it may help account for Minnis's misinterpretation. It reads: "e` chiaro che il soggetto di un'opera, sottoposto a due diversi significati, sara duplice"; it would better read: "e` chiaro che il soggetto, circa il quale corrano alterni sensi [= i sensi letterale e allegorici], deve essere duplice" ["it is clear that the subject, around which the other senses, one after the other, may run, must be twofold"]. The phrase "alterni [and not 'alii'] sensus" is a far stricter and "theological" precision than most editors and commentators understand. It includes the literal and the allegorical (i.e., the three spiritual senses), the one and the other. The text continues by saying that the subject of the work must be interpreted first literally, and then allegorically; that the subject of the entire work is the state of the souls after death (its literal sense), "nam de illo et circa illum ("statum," not "subiectum," as in Brugnoli's translation; the pronominal form would be "illud" were the latter the antecedent) totius operis versatur processus" ["for from it and around it the progress of the entire work develops"]. The movement forward of the entire work, its plot, or narrative, is concerned with the state of the souls after death, if it is understood literally. And now we join the text where Baranski intercepts it. And what it does is give the allegorical interpretation of the entire work, that of the three "alternating" non-literal senses of Scripture, allegorical, moral, and anagogical, all subsumed, as St. Thomas authorized, in the familiar passage from the Summa (1.1.10), which allows for or, rather, insists upon the identity of the three "allegorical" senses (allegorical, moral, anagogical) and the single threefold (trifariam) "spiritual" sense, where the past lives of the damned are seen as prefiguring their present "status post mortem," also indicating our need to incorporate the lessons offered by their lives and deaths into the choices in ours (for example, Paradiso XXV, 32, where St. James is said to have "figured" hope. The use of the exegetical term figuri helps us understand how we are to "read" James's life "morally," namely as a figure to be "fulfilled" by the hope in each of us), by living their lives in ours, as it were, either by fleeing what was vicious in their thoughts and actions or following what was virtuous in them, and also prefiguring their future damnation or glory under God's justice. Further, Minnis and Baranski are apparently, at least in this instance, not aware of the sort of question a theologian puts to such a text--and the author of the Epistle, as we have seen, is clearly speaking "theologically" in these passages. It is instructive to consult Thomas Aquinas, Quodlibeta, sixth quaestio, reflecting issues raised in his definition of the four senses in Summa Theologica 1.1.10 (the very text which is the closest we have been able to find to Dante's self-exegesis in section 7, as Baranski admits. He answers the following objection, which is much as that of Minnis and Baranski to section 8: "The moral sense is that which pertains to the teaching of morals. But Holy Scripture, in a number of passages, gives moral instruction literally. Therefore, the moral sense cannot be distinguished from the literal." Here is Thomas's resolution of the quaestio: "Let it be said that the moral sense [he is referring to the third of the four theological senses] does not refer to every sense through which morals are taught, but to that through which instruction in morals is understood from a likeness to certain things which have been done; for the moral sense is a part of the spiritual sense, because the moral and literal sense is never the same" ["quod nunquam est idem sensus moralis et litteralis"]. As James E. Shaw reminds us, Thomas defined allegory strictly, where others were content with far looser procedures. In his study of theological allegory, Dante seems to have been formed by Thomas far more than by anyone else. And his own distinctions are a great deal more "strict" than some contemporary theorists recognize. However, that he follows Thomas's formulations does not mean that he accepts his prescriptions (or proscriptions). I am grateful to Francesco Mazzoni for pointing out to me the following. The definition of the "literal subject" of the Commedia in the Epistle, section 8.24, "status animarum post mortem simpliciter sumptus," almost certainly reflects the following proscription of Thomas, who, in his commentary In X libros Ethicorum ad Nichomachum, 3, lectio 14, locating fortitude's place between boldness and fear, discusses our only normal fear of death. Admonishing us to concentrate on this life as we prepare for the next, he concludes as follows: "Ea enim quae pertinent ad statum animarum post mortem, non sunt visibilia nobis." Nardi makes a similar point to a different end. In a previous passage in his commentary to the Ethics, 1, lectio 9, Thomas had already specifically denied such knowledge as Dante would be claiming if he wrote the Epistola: "In this book the Philosopher speaks of the happiness that may be possessed in this life; for the happiness of the other life exceeds any rational investigation." Thus, if the non-epistolary sections of the Epistle were to have been written by Dante, Nardi concludes, their author would have had to deny exactly such purpose as section 8 arrogates unto itself, "per la tomisticissima ragione che lo 'status animarum post mortem' non e di competenza di nessuna parte e di nessun genere della filosofia, semplicemente perche 'omnem investigationem rationis excedit'" ["for the most Thomist of reasons, namely that the state of the souls after death is not in the competence of any part or branch of philosophy, simpy because {quoting Thomas} 'it exceeds the limit of any investigation of which reason is capable'"]. In both these instances (and the text found by Mazzoni constitutes a truly arresting detail, a far more interesting proof that Dante had Thomas in mind here [as he was in Dante's definition of theological allegory]), what we see reflected in the Epistle is Dante's precise knowledge that Thomas would not have warranted what he was writing. And Nardi's argument is a strange one, since it is he who has argued most forcefully for Dante's positioning himself as poet of revealed truth, if anything an even stronger prise de position against St. Thomas than Nardi finds in the Epistle. My view is that both Mazzoni and Nardi are correct in thinking that the author of the Epistola a Cangrande is in polemic with Thomas; the former sees him as citing Thomas, but is less willing than I to argue that the procedures of his poem exactly manifest his insistence on theological allegory; the latter argues that Dante would not have violated Thomas's "rules" in the Epistle, but then did so (and not a little) in the poem. I still find Padoan's resolution of the problem the most convincing one that we have: the Epistle and the poem share the same givens, the same techniques, the same effects.
The question of the adequacy of its description of the "subject" of the poem has long troubled doubters of the Epistle's authenticity; they have been offended that Dante's poem is referred to in what they consider trivializing terms. Remnants of this view remain in some recent work. For instance, Dronke believes that it does not yield an interesting result to say that the subject of the poem is the state of the souls post mortem, since it is in fact "the itinerarium mentis of Dante Alighieri"; thus the Epistle does not seem to him likely to be genuine. But the subject of the poem is the state of the souls after their deaths, from the moment in which we enter Inferno proper until that in which we see the vita of St. Bernard and the culminating vision of humanity as the blessed in the rose. And the fact is not trivial, for here is a poem which takes as its subject the history of the future, as it were. Further, that there is a "second subject" is clear, if it would have been out of order for Dante to say so. He is the second subject. If I may be allowed to quote myself, a quarter of a century ago I tried to put the case as follows: "The twin subject of the poem includes what Dante does and what Dante sees"; again, "The Letter to Can Grande is right as far as it goes, but there is another subject matter and it involves what Dante the Pilgrim does in the poem, while the state of the souls after death involves what he sees."
The above arguments offer, I hope, a sufficient answer to the attempts of Minnis and Baranski to pull the author of the Epistle back to the poet's side of the divide of allegory; that is not where he has elected to be found. One must be careful not to make Dante more of one's kinsman than he wanted to be. There is absolutely no reason by merit of which one can claim with confidence that section 8 of the Epistle is any less "theological" than the one with which it is inextricably bound and that precedes it, only the desire to "detheologize" Dante, as Minnis seeks to do, or to "undertheologize" the "false" Dante of the Epistle, Baranski's aim. (It is amusing to note that Baranski wants the true Dante to be a "theological" poet, and so eventually disagrees with Minnis's evaluation of Dante's "mixed" allegoresis, a position not very distant from Nardi's.)
Baranski's conclusion is that "as things stand today, the most 'economic'--to use Contini's term-- philological conclusion is that the Epistle is forgery." I do not think he has come close to making good his case. And his final gesture toward Contini has the unintended effect of reminding us that Contini accepted at once (in 1956) the genuineness of the Epistola a Cangrande as a result of Mazzoni's research and then underlined his agreement with some force in 1958; I do not believe he ever modified his opinion. (See the rejoinder to Baranski's concluding remark in Pertile's article of 1991, citing the same term ["economical"] in Contini in support of his view that the Epistle is in fact genuine.)