The Formation of Knowledge and Petrarch’s Books

With the development of the fields of the history of the book, codicology and material philology in the last 50 years, since the publication of Leon Gilissen’s Prolégomènes à la codicologie: recherches sur la construction des cahiers et la mise en page des manuscrits médiévaux, we have begun to recover the complex cultural and intellectual relationships inherent in the production of the medieval and early humanist book. It was, as we know, a process that both demanded extraordinary planning and allowed unique flexibility in the layering of texts that would constitute the intellectual canvas of often multiple hands and collaborators. Every aspect of the preparation of the book’s page, fascicle, and ultimate unity

among its parts, from its mise en page and script to the composition of its gatherings linked it in the intellectual processes of those who produced the book and those who read it to traditions as diverse as the mercantile culture responsible for vernacular anthologies that would be essential to the early Italian lyric (such as Vatican Latino 3793 and Escorial e.III.23) or the erudition of highly glossed legal and theological texts (such as Vatican Latino 1411, Vatican Urbinate latino 161 and Cesena Biblioteca Malatestiana MS s.IV.2). 3 These and similar systems of preparation signaled cultural ties to location and professions that in themselves defined ways of thinking and the mechanics of knowledge that reflected philosophical orientation. The use of the 'littera bononiensis' identified the mechanisms and influence of Scholastic legal thought defined by the glossators and professors of law at the University of Bologna. The two-column mise en page of tenth-and eleventh-century books, that contained text and gloss in two distinct scripts, ultimately produced a two-column layout in thirteenth-and fourteenth-century Italian manuscripts in gothic and chancery hands that contained texts as diverse as Boccaccio's Decameron, Dante's Commedia, Petrarch's and Guittone d'Arezzo 's letters. 4 Vol. 1: Introduzione. For a study of the relationship between the compiler and the copyists of the sole, datable fourteenth-century copy of the Convivio (MS Biblioteca Nazionale Centrale II.III.47, ca. 1361), see Beatrice Arduini: Assigning the 'Pieces' of Dante's Convivio: the Compiler's Notes in the Earliest Extant Copy. In: Textual Cultures: Texts, Contexts, Interpretation 3, 2 (2008), p. 17-29. 3 On the influence of these formulaic and highly glossed works produced in Bologna on Petrarch, see John Ahern: Good-bye, Bologna: Johannes Andreae and Familiares IV 15 and 16. In Teodolinda Barolini/Wayne Storey (eds.): Petrarch and the Textual Origins of Interpretation. Leiden -Boston: Brill 2007, p. 185-204. 4 As we see in the case of the Decameron copied by the Benedictine monk "Nicolaus" in 1396 (Firenze, Biblioteca Nazionale Centrale MS Banco rari 37, olim II.I.23), the long predominance of gothic script even late into the fourteenth century is witnessed in distinct scribal and cultural contexts. At the same time and in the same general geographical area but in a different cultural environment, we find the progressive semi-gothic script -proposed as a scribal reform of illegible gothic script by Petrarch -used for the transcription of the so-called 'historia Griseldis,' Petrarch's commentary and translation of Boccaccio's Decameron X 10 in Seniles XVII 3-4, together with a copy of Burley's De vita et moribus philosophorum, in MS Bloomington, Lilly Library Poole 26. These two scripts are distinguishable from the cursive chancery hands that were favored for fourteenth-century copies of the Commedia, proposed by Savino as a 'virtual autograph' and even for what is, according to Folena and others, probably a late fourteenth-century copy of Guittone's letters and poetry, Ms. Firenze, Riccardiano 2533. It is noteworthy that each of these scribal styles defines intellectual and moral associations between works and the readers for whom each was used. See Gianfranco Folena: Üeberlieferungsgeschicte der altitalienischen Literatur. In: Geschichte der Textüberlieferung der antiken und mittelalterlichen Literatur 2 (1964), p. 319-538; Giancarlo Savino: L'autografo virtuale della Commedia. In: "Per correr To say that Petrarch was attentive to the meaning of the association among script, mise en page, and intellectual tradition would be an understatement. In both his Familiares and his Seniles, Petrarch complains about the divisions of labor in the production of modern books, a practice that -in Petrarch's viewdamaged the essential unity of the intellectual's command in the preparation of a text. 5 For Petrarch not only was the design of the book a reflection of the edition's intellectual structure, it was also an integral part of its systems of meaning, from the clarity of its script to the unified organization of its knowledge in the text and its apparatus and glosses. It is, as Armando Petrucci has pointed out, Petrarch's preference for the simplicity and clarity of form that drove him to admire copies of the tenth and eleventh centuries in minuscule Caroline hands and to detest the often calligraphic and illegible minuscule forms of Gothic scripts overburdened by abbreviations and compendia, and tied to Scholastic thought. 6 Recalling the gift copy of his De vita solitaria to Philippe de Cabassoles, in Senilis VI 5 (written in Venice 6 June 1366), Petrarch describes the tribulations of finding a copyist whose simple but elegant and practical writing style stood in direct contrast to the then current writing habits of copyists who "pride themselves on small, cramped lettering that baffles the eye; by heaping and cramming everything together, […][their writing] confuses the spacing and piles up the letters, as . 5 See, for example, Familiaris XVIII 5 to his brother Gherardo about a copy of Augustine's Confessions. 6 Armando Petrucci: La scrittura del Petrarca. Città del Vaticano: Biblioteca Apostolica Vaticana 1967, p. 66. One of the clearest expressions of the positive features of Petrarch's ideal script is found in his Familiaris XXIII 19, written to Giovanni Boccaccio. The letter announces the nearing to completion of his Letters on Familiar Matters (Rerum familiarium libri) in the hand of his young assistant, believed by most to be Giovanni Malpaghini, which Petrarch describes as the antithesis of "that pompous and fancy lettering so typical of contemporary scribes […] as though it were destined for something other than reading." Rather the young scribe's hand is "in neat and clear lettering, affecting more than just the eyes and lacking […]  though they were riding on top of one another, so that the scribe himself could scarcely read them, were he to return a little later, while the patron who commissioned the book would really purchase not so much a book, as blindness because of the book". 7 The paragraph's initial rhetorical play on the general 'artlessness' (iners) and 'faithlessness' (perfidia) of copyists of Petrarch's day is instilled in the opening attack of Senilis VI 5, 6: "Accessit ad causas more scriptorum perfida semper inertia inersque perfidia". 8 Combined with youth's slavishness to fashion, the illegible script against which Petrarch rails also conveys copyists' lack of knowledge and understanding of the very texts they have themselves transcribed. His description of the negative aesthetic and practical dimensions of his day's gothic handwriting amounts to a programmatic condemnation of the very style of writing that forms most of the manuscripts and documents of the late Middle Ages. In this condemnation we find especially Petrarch's rejection of medieval formulae of learning instilled in and identified with the very formation of illegible gothic scripts. 9 At the same time, Petrarch's adoration for the earlier and more legible Caroline minuscule of the tenth-and eleventh-century books he sought out leads us to his 'effective aesthetics of knowledge'. This aesthetics of script and book production constitutes the core of Petrarch's notion of a renewed understandability of the page in the transmission of knowledge, a page that invites study and, as we shall see, intervention and collaboration through emendation  II,p. 226 (Sen. VI 5,[6][7], in which he explains that the copy he is sending has been completed by a presumably older religious whose simple handwriting is especially adept to their age: "Litera non tam anxie exculta quam nostre atque omni etati, nisi fallor, ydonea". Petrarch's initial play on the general 'artlessness' (iners) and 'faithlessness' (perfidia) of copyists is instilled in the opening attack of Senilis VI 5, 6: "Accessit ad causas more scriptorum perfida semper inertia inersque perfidia". Combined with the fadishness of youth, the illegible script against which Petrarch rails also conveys the copyists' lack of knowledge and understanding of their own texts: "[…] acervans omnia et coartans atque hinc spatio, hinc literarum super literas velut equitantium aggestione confundens que scriptor ipse, brevi post tempore rediens, vix legat, emptor vero non tam librum quam precio cecitatem emat" (Pétrarque: Lettres de la vieillesse, Rerum senilium IV-VII. Vol. 2, p. 227: Sen. VI 5, § 7). 8 "Another reason for the delay has been the ever deceitful laziness and lazy deceitfulness of the scribes" (Petrarca: Letters of Old Age,p. 198). 9 For an overview of Petrarch's systematic reform of gothic script in its relationship to legibility and knowledge, see H. Wayne Storey: Transcription and Visual Poetics in the Early Italian Lyric. New York -London: Garland Press 1993, p. 201-224; and Armando Petrucci's chapter on Petrarch's aesthetic and graphological ideals in Scrittura del Petrarca, and glosses. Already a significant influence in his reading in the 1330s, the clarity of the Caroline minuscule had by the 1340s became an essential feature of the 'scriptura notularis' in which Petrarch learned and collaborated in the margins of his ancient manuscripts in 'littera antiqua', through which "Petrarch sought a visual harmony in the composition of the manuscript charta among text, commentary and support". 10 It is, however, in his letter to the Florentine Lapo Castiglionchio, now Familaris XVIII 12, addressed 'Ad Iacobum Florentinum', that we find the essence of the relationship among learning, reading and writing that fed Petrarch's philosophy of knowledge. In his description of his own copying of an unknown work of Cicero's from Lapo's library ('opus rarissimum' [4]) Petrarch clarifies the seamless unity of these three activities. His letter explains the delay of four years in the return of Lapo's Cicero precisely because of the shortage of intelligent scribes. 11 Unable to part with Lapo's volume without having a copy of his own, and unable to turn -as was his habit -to a trusted copyist, literally 'per scriptorum ignaviam ' [3] Petrarch is compelled to pick up his own "worn down and battered pen" ('exesum atque attritum calamum' [3]). 12 The process that Petrarch describes as his own "custom" (mos) is not, however, simply the act of copying, but rather a triple act of engagement that fuses the physical and the intellectual: memory and the pen are unified. The possibility that one activity will diminish the other is held in check by the reciprocity of the movement of the eyes that propel the pen and the pen that paces the eyes. Even what should be contrasting participles, 'frenante' and 'urgente', now act in unison: Nichil legi nisi dum scribo.
[…] as I reached particular passages, I experienced such great pleasure and was drawn with such force that, reading and writing at the same time, I became aware [that] my pen did not move as rapidly as I wished it to. I feared that my pen would outstrip my eyes, that my compulsion to write would lessen because of my reading. Thus did I proceed, with my pen checking my eyes and my eyes propelling the pen; and not only did I find a delight in my toil, but I learned a great deal in the act of writing and I committed much to memory. Since writing is slower than reading, it impresses more deeply and clings more tenaciously in the memory.] 14 While seemingly a letter of excuses for a borrowed volume long overdue, the letter to Lapo Castiglionchio crystalizes Petrarch's philosophy of knowledge as the constructive pleasure of intellectual toil ("ut non dicam tedio animi […] sed labore manuum victus" [I would not say that my spirit got tired but I was defeated by the effort of my hand], Fam. XVIII 12, § 6). Physical fatigue is inextricable from learning, reading and copying, just as the author of his exemplar, Cicero, demonstrates in his own copying of the orations of others.
The manuscripts that transmit the patrimony of early Italian and Old Occitan literature supply us many of the formulae that Petrarch would have inherited. Large anthologies and even smaller volumes were often constructed according to divisions by poets and by literary genres. This construction was facilitated by copyists' use of booklets that contained mostly homogeneous content -by poet or genre or even theme -that could then be combined with other similar booklets. 15 In fact the vernacular traditions of early Italy were founded on a system of assembling books by virtually free-standing quires, or fascicle booklets. The thirteenth-century anthology MS Vaticano Latino 3793 is a collection of mostly single quires that usually conclude a 'content unit' at the end of most fascicles. 16 If it were not for the overarching historical and critical program with which MS Latino 3793 -an anthology of early Italian poetry -is organized, these single quires could easily have been placed anywhere in the manuscript without disrupting the reading of any of the texts that make up the single quires. The same could be said for the early transmission of the Vita Nova, for example, contained in two symptomatic early-and mid-fourteenth-century manuscripts, respectively MSS Laurenziano Martelli 12 and Magliabechiano Classe VI 143, precisely in two quires of five bifolia each (two quinternions). 17 This 'booklet-structure' served as well for numerous anthologies and even manuals of Old Occitan lyric, vidas, and razos produced in Italy. Of particular note is the early fourteenth-century manual with lyric repertory, vidas, and glossaries, MS Laurenziano 41.42, arranged from distinct fascicles to produce a virtual primer of the language, the poets' lives and their poetry. 18 This notion ultimately of a more rapid execution and diffusion of individual works both as independent booklets and within larger anthologies reveals a cultural orientation to the assembly of texts and knowledge itself in direct contrast to many of the models that Petrarch seems to have chosen for his own books. 19 inf 1]) stands as one of his 'modern' books that played a central role as the site where he noted the deaths of Laura and others, and where he studied, conjectured, and emended his Virgil and Virgil's canonical commentator Servius. 20 This very personal and textual site devoted to Virgil and Statius also taught him about the unifying intellectual force of a book constructed not of smaller, independent booklets but of integrated pieces of a whole context, of a cultural view in which the intricacy of material unity bears out the broader perspective of a classical intellect's world view and his influence on and relationships to other sources of knowledge. Petrarch's Virgil exerted a lasting influence on his notions of the unity of mise en page, mise en livre, and the effective transmission of knowledge in the strata of conversations among the text, the commentary and glosses, and future users of the book. Prepared upon a commission from his fatherthe "Petrus Parentis Florentinus" intuited by Giovanni Mercati as Petracco di Parenzo -well before 1326, the manuscript exhibits the tenets of a harmony of the page and the fascicle within the context of a vast editorial project that lasted the church fathers. But for our purposes in this essay, Nolhac's occasionally overlooked study Pétrarque bibliophile in his Pétrarque et l'humanisme, I, p. 13-85, examines the materials, the passions, and early historical development of "sa collection" (p. 36) and of the practical aesthetics of the book, the "habitudes de Pétrarque" (p. 41), that will guide his own book production in later years. 20 For the extensive bibliography on this pivotal book in Petrarch's development as a codicologist and an historian, see Marco Petoletti's introduction to the manuscript: "Petrus parentis florentinus, qui hoc modo volumen hoc instituit": Il codice, in Marco Baglio/ Antonietta Nebuloni Testa/Marco Petoletti ( all of Petrarch's life. 21 It is the work of a single hand and is -in the expanse of its first 232 chartae -devoted primarily to the works of a single author, in a gothic book hand, together with Servius's commentary on the Aeneid. However, certainly part of its original program would have included Statius's unfinished epic Achilleid and four of Horace's Odes. It is, nonetheless, the unified assembly of these works that would have offered Petrarch a unique material model. Composed entirely of quinions, 26 of them complete with a partial 27th, it is noteworthy that no work concludes at the end of a gathering until we reach c. 250, the close of Horace's Odes, suggesting that the entire book was carefully planned out so as not to divide its contents materially by fascicles. These will be the texts to which Petrarch will turn throughout his life to study, annotate, and remember, to integrate and -at times -contrast his observations and knowledge with that of Servius's standard commentary of the Aeneid. There is nothing sterile or Scholastic in Petrarch's reengagement of these works. Rather we often find the maturation of reflections on passages and works, which develop over years and through different texts. Such is the case most certainly with his marginal gloss of Georgics IV 545, "inferias Orphei Lethaea papavera mittes" [you will make funereal offerings of Lethaean poppies to Orpheus], on c. 51v, when Petrarch recalls Ovid's explanation in the Remedia Amoris (vv. 550-553) of "Lethaeus Amor, qui pectora sanat" [Lethaeus Love, which heals the souls]: "Nil melius amanti quam amoris et curarum oblivisci, ideo papaver sompniferum et obliviosum sacrificari precipitur Orpheo; hinc est ille Amor Letheus in templo Erycis, cuius meminit Ovidius in libro de remediis" [Nothing is better to the lover than forgetting about love and worries, and for that reason it is advised to offer the soporific and amnesiac poppy to Orpheus; hence that Lethaean Love in the temple of Eryx, which Ovid mentioned in the book De Remediis]. While the original verse in the Georgics sites the effects of the poppy offered to Orpheus that brings sleep and forgetting, Ovid's verses in the Remedia focus solely on the Lethaean forgetting of love's vows: "Illic et iuvenes votis oblivia poscunt" [There, young men ask for oblivion through their vows] (553). Nowhere in the Remedia is there a single 21 The entire inscription is "Petrus Parente Florentinus qui hoc modo volumen instituit" [Petrus Parente Florentinus who thus prepared the volume]. Billanovich (Il Virgilio, 52) formulated one set of explanations for the dates and provenance of the manuscript's production in which Petrarch himself would have had a hand, while Michele Feo proposed a much earlier date of preparation (end of the thirteenth century) for which Petrarch's father would have been responsible and, as in Billanovich's reconstruction, the project would have been executed by an Italian copyist. However, the most cogent discussion of the manuscript's preparation comes from Petoletti: Il codice, in which he proposes a southern French provenance of copyist and commission after Petracco's exile and the transfer of his family to Avignon. See also Giovanni Mercati: Opere minori. 5 vols. Città del Vaticano: Biblioteca Apostolica Vaticana 1937 (Studi e testi, 80), IV, poppy; instead Petrarch links the two passages on Lethaean oblivion through the "narcotic effect of the poppy" recalled in Virgil's Aeneid IV 486, when the witch Massyla -the custodian of the temple of the Hesperides -saves the sacred branches "spargens umida mella soporiferumque papaver" [sprinkling moist honey and soporific poppy]. 22 Many years later, Petrarch returns to the remedy of the poppy for forgetting suffering and loss in his long Senilis X 4, sent in 1368 to console Donato Albanzani over the death of his son ("Ad Donatum Apenningigenam grammaticum, consolatoria super illius filii suique simul nepotis immaturo obitu"). 23 The episode is significant in its lesson. The toil of Petrarch's marginal observations has now found a new context and a new twist. Petrarch prefaces his new development of this classical crossroad of citations by professing his preference, for his friend and himself, for forgetful happiness over "mournful recollection": "mallem ego michi et tibi, mallem iocundam oblivionem quam memoriam luctuosam" 24 [I would prefer, for me and you, I would prefer pleasant oblivion to mournful recollection]. The letter's very next line incorporates first the utility of oblivio as an aid to lovers (Remedia) and then the poets' consecration of the "soporific and […] amnesiac poppy". The linkage between the poppy and Orpheus is so strong, especially thanks to Virgil's description in the Georgics, that editors have frequently preferred that Petrarch make the same association. But here he surprises us, if we trust Lachmannian stemmatics, by substituting the oblivion of the dream state, Morpheus, in combination with the Lethaean poppy: Et est, hercle, oblivio, ut aiunt, amantibus utilis, unde apud poetas somniferum ac perinde obliviosum papaver sacrificatur Morpheo et Letheo sua sunt sacra Cupidini. 25 [And by Hercules, this oblivion is useful to lovers, as they say, and according to the poet, they sacrifice soporific and amnesiac poppy to Morpheus and his sacrifices are Lethaean to desire.] Aldo Bernardo and Marco Baglio presume that Petrarch has simply made an error by writing "Morpheo" when he must have meant "Orpheo". 26  be a reasonable correction in light of his previous development of the motif. But perhaps Petrarch's reapplication of readings and expansion of knowledge leads us in a different direction: toward a more complex understanding of forgetting that moves, after years of study and reflection, toward the metaphor of Morpheus's dream state.
Petrarch often supplies glosses as well that bring the history of his sources into the present day in a kind of dialogue with the future, unknown reader of his marginalia. While he engages Virgil directly, Petrarch seems at times to establish a trajectory between Servius's commentary and the present as a means of documenting the continuity of history. When at the beginning of Aeneid XI, when Servius comments on the triumphal arch as a reflection on Aeneas's desire to create a tribute to Mars out of Mezentius's arms ("ingentem quercum decisis undique ramis / constituit tumulo fulgentiaque induit arma, / Mezenti ducis exuvias, tibi magne tropaeum / bellipotens"  [he cut off all around the branches of a huge oak / he set it on a tumulus and hung from it the shining armor, / spoils of war from their leader Mezentius, as a trophy to you / great and powerful in war]), Petrarch draws the tradition into his own day by commenting on his own experience in Rome: "Mos arcuum triumphalium, quibus Roma nunc etiam plena est". 27 In some places, Petrarch's glosses mix sources and personal experience in complex combinations and over long stretches of time. In Book I of the Aeneid, Neptune calms the stormy waters created by Juno's wrath for Aeneas (v. 131). Petrarch's annotation points to Apuleius's Metamorphoses V 9, 84 and corrects it by reminding his reader of the true dominance of a Christian God in controlling the seas, heavens and the land: "'Ventis ipsis imperat' ut ait Apuleius […]; quod verius non de puella ut ibi vel de maris de aut hic, sed de celi terreque et maris domino dicitur" ['The winds obeyed her commands,' said Apuleius (…); but truthfully I am not talking about the girl there or that god of the sea, but of the Lord of heaven and earth and sea]. 28 As Caterina Tristano has noted, the passage "Ventis imperat" is repeated in the margin by Petrarch in his own copy of Apuleius's Metamorphoses, MS Vaticano Latino 2193, c. 56v. 29 Many years later, in 1362, Petrarch wishes his friend Paolo de Bernardo di Venezia a safe journey by sea in Senilis X 3, drawing upon his previous engagement of Virgil and Apuleius and reiterating that it is not Neptune but Christ who is the sole ruler of the sea, the land and the heavens: "maris et terre celique regnator Cristus omnipotens" [Christ almighty ruling over the sea, the earth and heaven]. 30 It is noteworthy that the letter concludes with a passage from Statius's Achilleid that Petrarch would have recalled from the very same manuscript (Ambrosiano A 79 inf. Sala Prefetto 10/27): "what Achilles' wife said in Statius: 'Go safely and come back true to us'" (Achilleid IV [I] 942). 31 It is possible that we are at a personal commonplace for Petrarch, who dreaded the dangers of sea voyages. 32 But even stronger -I believeis the nexus of text, faith, and knowledge to which Petrarch often turned. For the passage in Statius bids Achilles not only a safe return but also recalls that Thetis -remembered not only as Achilles' mother but also one of the deities of the sea -did not fear in vain. Petrarch returns obliquely to the question of who is the true god of the seas: i cautus, nec vana thetim timuisse memento, i felix nosterque redi! nimis improba posco 33 [And go about carefully, Thetis was not afraid for nothing, Go safely and come back true to us! I am asking for too many things] For Petrarch the gloss or, perhaps more accurately, the space of the gloss represents a place of exchange and of learning in a dialogue between the past and the future. This transmission of knowledge is instilled in the larger site of the text and gloss of the Aeneid. Petrarch's annotations are themselves a building site that is intimate, reserved for himself and for the intimates who will inherit his library, his Virgil, his Apuleius, for those who will consult the margins. 34 30 See Pétrarque: Lettres de la vieillesse: Rerum senilium VIII-XI, vol. 3, p. 305, § 48. For the passage from Senilis X, 3, see Ibid., p. 275, § 2. 31 "i felix nosterque redi! nimis improba posco" (v. 942). We should note that Petrarch's Achilleid in his precious Virgil manuscript was divided into five books, according to the medieval tradition. After the incipit on c. 234r ("Magnanimum eacidem formidatamque tonanti"), Book II begins on c. 236v ("At tetis undisonis per noctem irrupibus astans" [v. 198] This nexus between the intimacy of textual space and learning, especially its unique transmission in Petrarch's manuscripts, certainly is most concentrated in his Virgil. But, as we remember, the manuscript was -by Petrarch's own account -out of his possession from 1326 until 1338. These are, as Armando Petrucci reminds us, essential years -especially 1337-1343 -in the development of Petrarch's studies as well as of his consistent gloss script. 35 In even earlier years, 1325 to 1328, the young scholar will acquire a copy of Augustine's De civitate Dei (today MS Padova, Biblioteca universitaria 1490). They are as well the years in which he will prepare what Petrucci calls the first critical edition of Livy's Ab Urbe condita: MS London, British Museum Harley 2493 (see Plate 1). 36 Already in both annotations and texts of these two manuscripts, especially in his gloss on c. 219v (Plate 1), we see the elegance of Petrarch's lighter gothic hand as well as his control of the text's presentation in the coordination of three other contemporary copyists charged with executing Petrarch's edition. 37 Only in 1962 was the young Petrarch's ownership of a late tenth-century Italian manuscript in a Caroline hand documented in the possession of the Swiss bibliophile Martin Bodmer (today Cologny, Bodmer Library MS 146 [see Plate 2]). 38 It would seem that the codex held a particular fascination for Petrarch both at the time of its original acquisition, sometime around 1330, and again later in his life.
Today a fragment, the MS carries a clear note of its young owner as well as evidence of the young scholar's interest in Cicero's Partitiones oratoriae, which Petrarch marks on c. 35r as comprising 15 chartae (Plate 2). Even in the first four chartae that remain of the work in the Bodmer manuscript, we find Petrarch annotating and "marking up" his copy of the text, to whose title he has added "sub dialogo" ("Marcij tulij ciceronis de partitiones […] re rhetorice sub dialogo incipient" [Begins Marcus Tullius Cicero's dialogue On Oratorical Partitions]) and accordingly added the letters 'C' and 'M' to indicate the changes in voice in the dialogue between Cicero and his son Marcus. But for Petrarch the lesson of the Bodmer codex was far more complex than simply being a copy of one of Cicero's minor works on oratorical skills. It is, in fact, the material regularity of this book that gathers together four works by four authors on the theme of rhetorical skills that presented itself as a unique model for a genre that would eventually have greater significance to Petrarch: the unified miscellany.
Produced by a single, late tenth-century copyist on at least six regular quaternions (up to c. 48) on chartae of 36 ruled lines in two columns, the codex would have provided a unique formula of clarity and consistency in the ordering of texts by four different authors on the same topic. The following table is helpful in visualizing the material construction of the miscellany: Cologny By most accounts, in April 1338, Petrarch began to devote time to his recovered Virgil. The intense work of Petrarch's study and glossing reveals a fully formed hand that imitates the clarity and lightness of the Caroline hands in his miscellany on rhetoric (Bodmer 146) and Augustine's Enarrationes. To understand the significance of this development, we can compare Petrarch's use in the same period, 1337-1339, of a gothic chancery -or semi-cursive -hand to transcribe drafts of lyric poems in the vernacular that will ultimately find their way into the Fragmenta. The regular and posed hand of Petrarch's scriptura notularis in this period reflects the scholar's engagement in a process of learning that Plate 3: Cologny, Fondation Martin Bodmer, Cod. Bodmer 146, c. 18v. Manuscript with rhetorical works, owned by Petrarch; published with the kind permission of the Bodmer Library required a stable notational style. Around the same time, the variation evident in the earliest entries of his draft copies in MS Vaticano Latino 3196 -a manuscript to which we shall return later -reveals the experimental nature of his poetic works in the vernacular. As the margins of Latino 3196 demonstrate, Petrarch's writing is here destined for emendation and for subsequent transcription in a fair hand elsewhere, "in alia papiro". It is the script that Petrarch will use to communicate with other scholars in the margins of his copies of classical and patristic authorities that has undergone the most urgent reform away from the dense gothic minuscule of his Scholastic predecessors and toward what would become an elegant semigothic adopted, as Lilly Poole 26 demonstrates, in the last decades of the fourteenth century. While we possess precious few copies of Petrarch's vernacular hand, the witnesses we do have suggest a slower process in adapting the features of his beloved Caroline script, confirming -I would conjecture -Petrarch's attitude toward what he called his little trifles, nugellae, in the vernacular ("Nugellas meas vulgares") in his letter of 1373 to Pandolfo Malatesta (Sen. XIII 11) as something more than a rhetoric stance. 40 With the acquisition of his prized eleventh-century Horace in Genova in November of 1347 (today MS Laurenziano 34.1 [see Plate 4, a detail of c. 26v]), Petrarch secures for his own study and references in the glosses of numerous other manuscripts -including his prized Virgil -a codex that offered him not only an authoritative Caroline model in quaternions of an edition of Horace's opera omnia but also a site which will collect the widest chronological variety of annotated entries over the years of his ownership, allowing us to chart the often microscopic changes in Petrarch's gloss style. Well beyond a collection of variations in script, Petrarch's possibly oldest manuscript offers us details on the material construction of the transmission of knowledge that influenced Petrarch's own notions of book making.
Central to this manuscript's construction are: 1) an intricate system of strata of glosses that no longer relies on a two-column presentation but on clearly divided spaces of textual function laid out in regularly ruled 29-line chartae for the main text and 52 lines for the glosses in the wide external margins; and 2) the unifying dynamic of the single copyist responsible for the transcription of 140 chartae, recto and verso, in 18 gatherings, or 16 quaternions (or gatherings of eight chartae) plus a single quinion (Fascicle III) and a final bifolium (or 2 chartae) added to complete the second book of Horace's Satires (cc. 139-140 of 104r-140v) at the end of the codex. Examining the structure of the gatherings presented in the table below, we see that Horace's six works never conclude at the material end of a quaternion: the Odes (cc. 1v-56r [end of Fascicle VII is at c. 58]), the Ars Poetica (cc. 56r-64v [end of Fascicle VIII is at c. 66]), the Epodes (cc. 65r-76v), the Carmen Saeculare (cc. 76v-78r [quaternion X finishes with c. 82]), and the Epistles (cc. 78r-104r [Fascicle XIII, a quaternion, ends on c. 106]). Only the Satires, the final work in the manuscript, closes the 18th and final gathering, a bifolium (cc. 139-140) added to the final quaternion so that the copyist can complete the final satire: Like Petrarch's 'Virgilio Ambrosiano', his eleventh-century Horace scrupulously observes the same unifying rule that no work should conclude at the end of a gathering. Considering Gilissen's methodological norm of examining the placement of irregular gatherings, we should note that the third fascicle, cc. 17-26 (a quinion) and the final gathering, a folded sheet or a bifolium (cc. 139-140), break the consistency of the book's construction in quaternions.
In fact, the insertion of the quinion, or the 10-charta gathering (=III), avoids the aesthetic problem of three of the book's six works concluding at the end of a fascicle: the Odes at Fascicle VII, the Ars poetica at the end of Fascicle VIII, and the Epistles at the close of Fascicle XIII. The copyist's simple alteration in the size of the third gathering in the middle of the Odes can have little other function besides that of creating a more unified book. Long before the less precise method of constructing manuscripts by confining whole works to individual gatherings and then suturing those gatherings to make miscellanies, we find a model of the integrated edition of a single author carefully constructed across many gatherings, a large edition that would have been complete only through the ordering and sewing together of its dependently component parts. Some might argue that Petrarch would not have counted chartae and stitchings in the middle of gatherings (signaled in our list by "|") as Gilissen taught us. However, as we see in the examples of cc. 26v (Plate 5), 34v (Plate 6) and 114v (Plate 7), the original copyist of Laurenziano 34.1 has accommodated future readers, and most certainly the observant Petrarch, by inserting a Roman numeral at the bottom of the final verso of each gathering to guarantee the book's sequence of quires when they are sewn together.
The gathering number is always in the same ink and Caroline hand as Horace's central text. While the copyist would have meant it as a tool for guaranteeing the The careful planning of the quires and the insertion of irregular fascicles will be techniques that Petrarch himself will utilize in the construction of his own copy of Rerum vulgarium fragmenta. To that end, in fact, the final, irregular bifolium added by the copyist of Laurenziano 34.1 to complete the second satire would also have been a particularly useful lesson to Petrarch. At the close of Fascicle XVII, the copyist must add at least two chartae -or a single sheet of parchment, folded in two to make a bifolium -to conclude the second satire. From the last charta (138) of Fascicle XVII onto the recto of the first charta (139) of Quire XVIII, he continues the same mise en page of 29 ruled transcriptional lines for the main text (the gloss lines have expanded to 62).
On the verso of c. 139 he increases the writing canvas to 31 lines. But on the final charta he has at his disposal (c. 140; see Plate 8), the copyist of Laurenziano 34.1 has to fill the recto and verso with 35 and then 36 transcriptional lines to keep from adding another bifolium, which would have left at least three full sides of the two leaves, or three modern pages, completely blank at the close of the codex. The original copyist elects to break the system that he has maintained, the 29-line canvas, for over 138 chartae rather than conclude his work with blank parchment. This is a noteworthy decision. It safeguards the integrity of the book's unity as an integral edition of Horace's opera. But it also provides Petrarch with an additional, curious model. For blank space in this manuscript and in medieval manuscripts in general announce divisions between works, or incompletion and poor planning by the copyist or compiler. 41 Only once before has blank space been used in Laurenziano 34.1 to indicate the conclusion of one of Horace's works (see Plate 9). The empty space of c. 64v in MS Laurenziano 34.1 visually announces the pause between the end of Horace's canonical Ars poetica and the incipit of the first of his Epodes on c. 65r (note the added rubric "De arte poetica explicit Incipit Epodon ad mecenatem […]").
This medieval tradition of spacing between works within the construction of a manuscript would have been a commonly accepted practice by Petrarch's day. And, in fact, those manuscripts of the Fragmenta that Petrarch either 41 For example, numerous manuscripts of Dante's Commedia contain a blank recto or verso of a charta to signal the division between canticles.
Plate 9: Firenze, Biblioteca Laurenziana, Pluteo 34.1, c. 64v-65r; published with the kind permission of the Biblioteca Laurenziana prepared or oversaw, or that can reasonably be assumed to derive from an early generation of manuscripts that he approved, all demonstrate a division between what we might call a Part I and a Part II that Petrarch utilized in his own final service copy of the work (Vaticano Latino 3195). 42 By tradition, the blank space between Parts I and II in Vaticano Latino 3195 should have been confined to the verso of c. 52. But as we know, the blank space between Rvf 263 ("Arbor victoriosa triumphale") on c. 49r and Rvf 264 ("I' vo pensando") on c. 53r extends over six sides of ruled parchment (cc. 49v-52r) and one unruled side (c. 52v). Petrarch's extensive experience with the preparation and structure of books like the Laurentian Horace (34.1) as well as the construction of early copies of the Rerum vulgarium fragmenta suggest that the extensive blank space at the close of Part I in Vaticano Latino 3195 is an anomaly and potentially a space intended for additional poems that would have filled the six ruled sides of cc. 49v-52r. While the outer folded sheet of this binion, cc. 49|52, accommodated Rvf 260-263 and the requisite blank side (c. 52v) before the beginning of Part II on c. 53r, the internal folded sheet that is ruled but blankthe bifolium cc. 50|51 -is completely without function in the bound manuscript. 43 The anomaly was certainly noted by two subsequent manuscripts (New Haven, Beinecke M 706 and Vaticano Reginense Latino 1110) as a way of authenticating their own copies and identifying their antegraphs as copies authorized by Petrarch. 44 The unique state of those unfinished chartae hardly reveals a Petrarch who ignores the structural standards of his time, especially in light of other decisions regarding the spatial organization of what would become one of the most influential manuscripts of early Italian literature (Vaticano Latino 3195). As I have demonstrated elsewhere, the technique of adding bifolia and binions to expand the spatial potential of a work in-progress is nowhere more evident than in the second sections of both Part I and of Part II. 45 Petrarch's knowledge of book assembly is integral to the process of what we could call the "formation of the work's poetic knowledge". This intricate process was founded on the iron-clad repetition of visual-poetic structures that define the 'poetic page', and the component parts of the structures in which that 'page' builds its lyric and intellectual case: the folded sheet (bifolium), the gathering (from the binion [two folded sheets] to the quaternion [four folded sheets]), and the unified book. We know the original project of the Fragmenta at the time that Petrarch began what would become Vaticano Latino 3195 contained materially six quaternions (today's cc. , 53-60 [VI, but today's VII]) and a binion (today's cc. 61-62 and 71-72), all carefully calculated by Petrarch and his copyist to accommodate the contents of the work as it was conceived in 1367. In that plan, Petrarch applied the visual aesthetic of the integrated gatherings, making sure that the first three quaternions were bridged or linked by a canzone. 46 Like all author-copyists, Petrarch knew the material tools of his trade and how to use them. As the collection began to grow beyond the material limits of its original plan, Petrarch drew upon the experience of manuscripts such as Laurenziano 34.1 to expand the material support eventually by another quaternion (today's cc. 41-48) and three additional binions (today's cc. 49-52, 63-66, and 67-70). Just as the copyist of Laurenziano 34.1 added the bifolium to complete Horace's second satire, Petrarch himself inserted first two binions (cc. 63-66 45 One of the first things that must strike us about the material construction of Vaticano Latino 3195 is Petrarch's decision to use the form he found in his tenth-and eleventh-century MSS: quaternions, rather than the far more popular quinion, in which his early fourteenth-century Virgil had been constructed.  (VI, that Petrarch had begun in 1368 in order to expand his original project for Part I eventually with an additional 60 poems (2 canzoni, 3 sestine and 55 sonnets). 47 The expansion offered by the binion's four chartae (or eight pages) in Part I would have allowed Petrarch to include numerous compositions between the allegorical "arbor […] triumphale" of Rvf 263 and the contrastive, even warring, "pensieri" of Rvf 264, the beginning of Part II. The metaphor of Laura's "bel tesoro" (Rvf 263, 13) could have opened additional reflections on beauty, just as we see in the definitive turn of sonnets 260-263 on c. 49r in tone and topic toward ancient virtues of glory (Rvf 260, 12), fame (Rvf 261, 1), and infinite beauty instilled in moral refinement (Rvf 261,12;Rvf 262,2;Rvf 263,12). The final entry in Fascicle VI of "Cercato ò sempre solitaria vita" (Rvf 259) and Petrarch's recall of Laura in the moral mud of Avignon (vv. 9-11) linked to his own adversarial fortune prepares us -as Petrarch will so often do in the Fragmenta -materially and thematically for the shift that he is undertaking in the added and ruled binion, ready for a new set of lyrics that might well have filled the next three chartae recto and verso (or six pages).  [208][209][210][211][212][213]238, When he prepared the second binion of cc. 67-70 in a service copy, he first had to fit the poems he intended to add within the material limits of the binion's four chartae. With few erasures present and the clarity of the hand and ink noticeable between the previous binion (X, cc. 63-66), it would seem that Petrarch had started by preparing a clean copy of the 23 poems he would eventually add (21 sonnets and 2 canzoni). As he had done in previous addenda in a service-copy form, Petrarch still maintains the intricate transcriptions of his visual poetics for sonnets in Fascicle XI. 48 However he must alter his transcriptional style at first slightly, adding a thirty-second line to his normally 31-line canvas to fit the 71 verses of "Quando il soave mio fido conforto" (Rvf 355 [revised as Rvf 359]) onto a single side, the recto, of c. 69. But it is the much longer "Quel' antiquo mio dolce empio signore" (Rvf 356 [revised 360]) that will force Petrarch to alter his transcriptional style in order to avoid adding yet additional material support to his book. As I have demonstrated, the transcription of the 157 verses of "Quel' antiquo mio dolce empio signore" on cc. 69v-70r should have occupied three sides of two chartae rather than a little over a side and a half of parchment as it now does. 49 To fit the canzone into the binion Petrarch had to forego the strict visual poetics of the canzone's prosodic sister, "Una donna più bella assai che 'l sole" (Rvf 119, cc. 24v-25v]) to keep from having to insert an additional bifolium (a folded sheet or two chartae), that would have meant that the book would have had a blank charta (recto and verso) among its final pages. Here in a service copy Petrarch is forced to abandon his signature layout, virtually restored in the edited form of the canzone in the Petrarchive, for a 'service version' that crammed three verses -often hendecasyllables -onto the same transcriptional line. 50 As an intellectual and an impassioned and prolific writer-scribe, a grafomane, Petrarch was for most of his life a master of the gloss. But when it came to producing 48 Petrarch's process of organizing materially the Fragmenta was founded on three essential principles. The first is that Petrarch conceived of an ideal 'canvas' of 31 transcriptional lines for every charta. An essential dynamic that reinforces and is defined by that 31-line canvas is the repeated structure of two fundamental designs for the charta as a material unit as well as for the second and third principles: the placement of four sonnets per charta using seven transcriptional lines of two verses per line with a dividing line between each sonnet; and the coupling of the sestina, the only genre to be read vertically in Petrarch's system of visual poetics, with a single sonnet on a single charta. These two organizing principles account for just over 40 % of the manuscript's material space of 172 sides of chartae, or pages. See H. Wayne Storey/ John Walsh/ Isabella Magni: Glossary: Visual Poetics; Sonnet; and Transcriptional Canvas. In: Petrarchive: http://petrarchive.org for a description of Petrarch's visual poetics applied to the sonnet. For the principles governing the other genres of the Fragmenta, see Storey: All'interno della poetica grafico-visiva di Petrarca, 49 See Storey: All'interno della poetica grafico-visiva di Petrarca, For the form that Petrarch's "Quel' antiquo mio dolce empio signore" would have taken in a fair copy, see the Petrarchive (http://petrarchive.org), c. 69v. the 'published' fair copy of one of his own works, the assignment usually went to one of his trusted copyists who could be counted on for consistency throughout the copy. For Petrarch -as well as for most medieval readers -the consistency of a single copyist across an entire manuscript lent an additional sense of unity to the book that was difficult to attain with two or more scribes, even when subsequent copyists attempted to imitate the hand of their predecessor. Once Petrarch's primary copyist abandons the project of the Fragmenta, clearly visible at cc. 38v and 62r in Vaticano Latino 3195, Petrarch himself will attempt to complete sections of Fascicles V and IX in a fair hand before the manuscript was sent to Milano in 1368 for rubrication. 51 Even after the manuscript's return from the rubricator, Petrarch struggles to maintain his principle of consistency of hands in the preparation of the fair copy. His most famous attempt is the recycling the rounded section of the rubricated gothic D of "Donna mi vene spesso ne la mente" on c. 26r to produce, over the erasure of the rest of the ballata, the madrigal "Or vedi, Amor". 52 In spite of Petrarch's efforts, a glance at any of the chartae of Fascicle VI (cc. 41-48) or IX (cc. 61-62/71-72) tells us that the character of the fascicles has changed from 'fair copy' to 'service copy'. Even on a charta that has undergone what is normally the final phase of manuscript production (the addition of rubrication and colored initials) such as 39r, we see the multiple strata of erasures and experimentation that mark many of the clearly service-copy sections of the manuscript. From the entire palimpsest of "L'aura gentil, che rasserena i poggi" (Rvf 194) to the multiple layers of erasures and rewritings in "L'aura serena che fra verdi fronde" (Rvf 196) and "L'aura celeste che 'n quel verde lauro" (Rvf 197) we see Petrarch returning to erase, emend and revise in a hand designed to communicate more to a copyist who will then prepare a clean, fair copy of the work than to a reader who would have to study carefully Petrarch's text (see Plate 10). 51 See especially cc. 38v-39v and c. 62r-v in Vaticano Latino 3195. On c. 38v, only "Una candida cerva sopra l'erba" (Rvf 190) is in the hand of the primary copyist. The sonnets from "Sì come eterna vita è veder Dio" (Rvf 191;c. 38v) to "O bella man, che mi destringi 'l core" (Rvf 199;c. 39v) are in Petrarch's hand. On c. 62r, "Tranquillo porto avea mostrato amore" and "Al cader d'una pianta che si svelse" (Rvf 317 and 318) are in the primary copyist's hand, while "I dì miei più leggier' che nesun cervo", "Sento l'aura mia anticha e i dolci colli" (Rvf 319 and 320) are in Petrarch's hand, as is "È questo 'l nido in che la mia fenice" (Rvf 321) on c. 69v. All are rubricated. However, the variation in ductus, ink and even the pen used for transcription, even in the short space from "Di dì in dì vo cangiando il viso e 'l pelo" (Rvf 195) on c. 39r to "O bella man" (Rvf 199) on c. 39v, demonstrates the inconsistency of Petrarch's more formal, or "fair", hand, that did not match the primary copyist's regular ductus. 52 For an analysis of the scribal process and its implications, see H. Wayne Storey: Mobile Texts and Local Options: Geography and Editing. In: Textual Cultures: Texts, Contexts, Interpretation 8, 1 (2013), p. 10-20.
On c. 39r we are virtually at the same level of multiple interventions by the poet as we see in heavily reworked versions of poems in Petrarch's earlier autograph manuscript of draft copies of poems with numerous cross-outs and emendations (Vaticano Latino 3196). Most revealing perhaps is the case of Rvf 197, Plate 10: Vatican City, Biblioteca Apostolica, Latino 3195, c. 39r; published with the kind permission of the Biblioteca Apostolica Vaticana "L'aura celeste che 'n quel verde lauro", originally drafted as "L'aura amorosa in quel bel verde lauro" at the bottom of c. 2r of Vaticano Latino 3196 (see Plate 11). A careful examination of the poem's extensive experimentation and revision in Latino 3196 and the accompanying notation "transcriptus per me" (transcribed by me) suggest that Petrarch might have decoded his complex marginalia and final wishes for the poem and transferred it to c. 39r (Plate 10). 53 But, in fact, there are still at least two phases of revision that will take place between what we could call a final form suggested by Latino 3196 and the last version that Petrarch leaves us on c. 39r of Latino 3195. Even in the drafts of Latino 3196 Petrarch considers two relatively different versions of the second quartina, especially vv. 5 and 7.
The first version of Latino 3195, what should be a clean copy for rubrication, shows additional alterations, such as in the opening of vv. 7 and 9 ("Pò quello in me […]" and "Ne posso [dal] bel nodo […]"), not found in the final draft of Latino 3196, but also at least a second round of revisions that includes two major and two smaller but pivotal erasures and emendations respectively in vv. 10 and 11 and in vv. 7 and 12 (see Plate 12). In v. 10 Petrarch adapts part of a later emendation, "lega et stringe", but alters significantly the syntax of vv. 9 and 10, eliminating the "spirto gentil" and intensifying the power of v. 9's "crespo laccio, / Che sì soavemente  […]"; the erasure of the entire v. 11 clarifies the object of the two verbs in a powerful enjambment. Notably, the erasure in v. 7 after "Ne posso" destabilizes the parchment and the reading "dal", and eliminates the ascender of the 'b' of "bel". 54 These microscopic details bear witness to a state of Petrarch's Latino 3195 that we can only call, even in this section that has been rubricated, a service copy that abandons the consistent aesthetics -Petrarch's and his contemporaries'of the fair copy. While many medieval fair copies demonstrate scribal interventions to correct errors, seldom do we find a manuscript such as Vaticano Latino 3195 that has become a site of authorial experimentation on so many levels: from the reordering of compositions and extensive erasures and revisions to the material expansion of the macrotext itself through the addition of quaternions and binions. It is a unique site that draws upon Petrarch's extensive and attentive experience with earlier Latin codices, such as Bodmer 146 and his Laurentian Horace (34.1), and that model book that supplied him his knowledge of Virgil and Statius, Ambrosiano A 79 inf. Sala Prefetto 10/27, and became one of the precious sites of his own memory, both intellectual and personal, in which he recorded not only his observations on key passages of both Virgil and his commentator Servius, but also where he recorded the dates of the deaths of his most beloved friends and, as we know, of Laura.
Plate 12: Vatican City, Biblioteca Apostolica, Latino 3195, c. 39r (detail); published with the kind permission of the Biblioteca Apostolica Vaticana 54 While Modigliani reads "dal" over an erasure, but my own examination of the erasure suggests that "dal" is, like the 'b' of "bel", partially -and mistakenly -erased. See Ettore Modigliani