Dante's Conception of the Form of the Canzone

Neoplatonic imitations of the motions of the cosmos such as the one just examined occur in texts that were indisputably known and studied

by Dante. As we shall see, as early as the Vita nuova Dante is striving to give the form of his canzoni a metaphysical (if not yet really cosmological) foundation. Our understanding of his conceptions is considerably enhanced by the fact that we have Dante's own discussion of the nature of the canzone in the De vulgari eloquentia,^[49] which he seems to have written—and abandoned—late in the first decade of the fourteenth century, the same time span that saw the writing of the ambitious Convivio, also abandoned, it seems, in favor of the Commedia. In the De vulgari eloquentia we will find strong confirmation of our claims for the nature of the new microcosmic poetics of the petrose.

The first book of the De vulgari eloquentia moves from general topics—the superiority of the vernacular to "grammar" (i.e., Latin and Greek), the origin of language, the confusion of tongues, the interrelation of the Romance languages—to the nature of the noblest Italian vernacular, the vulgare illustre latium (which Dante urges should be based on the language spoken at the most illustrious courts of Italy rather than on the language of any particular region or city). Dante apparently planned to treat the entire subject of eloquence in the vernacular, and he refers in several passages (1.xix.3, 2.iv.6, 2.viii.6) to a fourth book on the middle and low styles; but the work breaks off in the second book, in the midst of the section on the canzone, which follows a first section establishing that the vulgare illustre latium should be used only by the best poets writing on the noblest subjects in the noblest poetic forms.^[50]

One of Dante's aims is to claim for his own lyric poems the standing in the Italian tradition he felt they should have, and if in the course of his discussion he cites many passages from Provençal and Italian poets as representing an authoritative canon (the Italians feature Guido Guinizelli, Guido Cavalcanti, Cino da Pistoia, and a few others), it is in order to set beside them the poems of his own for which he claims authority. The eleven references to his own poems are interestingly distributed, in fact, among the main categories of his serious lyric: there are three references to the Vita nuova, three to the petrose, one to the Convivio, three to uncollected doctrinal poems, and one to a lost poem. Three canzoni are singled out by being mentioned twice: "Amor, che movi tua virtù dal cielo" (Dante 1979a 174, 216), "Donne ch'avete intelletto d'amore" (pp. 204, 218), and "Al poco giorno e al gran cerchio d'ombra" (pp. 212, 226). Significantly, the rime petrose are cited on an equal footing with "Donne ch'avete," the poem that Dante had identi-

fied as his first major achievement;^[51] they are proudly put forward as instances of the high canzone and are clearly regarded by Dante as embodying the thematic seriousness, as well as the technical mastery, required of the tragic style. They are in no sense being repudiated or apologized for; rather, one purpose of the treatise is to assert normative rather than eccentric status for them. We believe that Dante's references to them, when read in the full context of Book 2, support our readings. For this reason, Book 2 is included, with translation, in Appendix 4.

The section on the canzone, as it comes down to us, argues that the canzone is superior to the sonnet and the ballata, that it is the only form, in fact, worthy to treat the highest subjects (arms, love, virtue) in the noblest (the tragic ) style. It discusses the relation of the canzone to music, the appropriateness of the various verses (Dante excludes verses with even numbers of syllables and those with more than eleven or less than five, effectively limiting the high canzone to the hendecasyllable and the settenario ), the form of the canzone stanza, and the nature of the "tragic" style, covering both vocabulary and syntax. We will focus here on Dante's conception of the form of the canzone stanza; the important sections on the poetic qualities of "smooth" and "hairy" words will come under discussion especially in Chapter 5 (on "Così nel mio parlar voglio esser aspro").

A particularly striking aspect of Dante's discussion of the canzone stanza is his use of apparently musical terminology, which has led to serious misunderstanding.^[52] All canzone stanzas, he asserts, are based on musical principles; each is "constructed to receive a certain melody" ("omnis stantia ad quandam odam recipiendam armonizata est," 2.x.2 [1979a 210]), even though it may never be set to music. In fact, in 2.viii Dante has established the independence, even the superiority, of the words of a canzone relative to the music. After asking whether the term canzone (cantio ) refers to the words or to the music, he decides in favor of the first possibility, never even mentioning that the term might refer to the union of words and music:

Nullus enim tibicen, vel organista, vel citharedus melodiam suam cantionem vocat, nisi in quantum nupta est alicui cantioni; sed armonizantes verba opera sua cantiones vocant, et etiam talia verba in cartulis absque prolatore iacentia cantiones vocamus. Et ideo cantio nichil aliud esse videtur quam actio completa dicentis verba modulationi armonizata.
(2.viii.5–6 [Dante 1979a 202])

For no wind player or organist or string player calls his melody a canzone except insofar as it is wedded to some canzone; but the harmonizers of

words do call their works canzoni, and such words lying on the page in the absence of a performer we still call canzoni. And thus it is clear that a canzone is nothing other than the action, complete in itself, of writing words harmonized with a view to musical setting.

Here Dante clearly is not thinking of the author of the words of a poem as also the composer of an eventual musical setting. Far from it: he thinks of the musical settings as produced by musicians: their melodies (the authorship is clearly assigned) are called canzoni only when "wedded" to texts that carry the name, marriages in which the texts are clearly thought of as the male partners, the melodies as the subordinate, female ones. Likewise, and significantly, there is no mention of the musical aspect in the formal definition of the canzone that sums up this chapter: "a yoking together of equal stanzas, in tragic style, without refrain, treating one thought" ("equalium stantiarum sine responsorio ad unam sententiam tragica coniugatio," 2.viii.8 [1979a 204]).

Dante's discussion of the canzone stanza does agree with what is known today about troubadour musical practice,^[53] but he never even mentions the properly musical factors—intonation, melisma, high points, low points, final tones, cadences—that mark the structure of a melody as such; he never refers even to the concept of musical modes, only to repetition and change of melody (or lack of them). His reference to manuscripts of canzoni implies that it was normal for them to lack music: ("talia verba in cartulis absque prolatore iacentia cantiones vocamus" fails conspicuously to mention musical notation). The one representation Dante gives of the singing of one of his poems (Purgatorio 2.106–114) leaves us in the dark about who its composer may have been, though it makes it perfectly clear that it was not Dante. In fact in Italy there had almost certainly never been a direct connection between the writing of a canzone and the composition of its musical setting, such as there had been—in theory, at any rate—in Provence. No Italian melodies for canzoni survive from the thirteenth and early fourteenth centuries, and the leading student of the subject, though strongly persuaded of the inseparability of canzone and music, concludes that the Italians always borrowed melodies (Monterosso 1970b 808). In any case, nowhere in the De vulgari eloquentia is there any indication that Dante planned to write about the rules of musical composition. He accurately characterizes Italian poetic practice; he tells us nothing about musical practice, and there is no reason to suppose he had any particular technical knowledge of music (his terms derive rather from the rhetorical tradition).

Dante's references to the musical structure of the canzone stanza, then, have quite another function. In the first place, they are to be taken as characterizations of the verbal patterns, the literary-rhetorical forms, de facto and, in his judgment, de jure established as appropriate for musical setting; they retain the literary awareness of a connection between lyric poetry and song. In the second place, they reflect his conviction that the composition of verse is inherently musical in a broader sense, since it is a fashioning of words governed by numerical proportion. For this reason, and also because verse involves the tempering of opposites, such as "combed" and "bristling" words (for discussion of these terms, see Chapter 5) or sweetness and harshness (2.vii), verse is in itself a valid imitation of the musica mundana, the harmony of the universe, and probably for Dante, because of its greater intellectuality, a higher one than music itself.^[54] We shall return to this point after our survey of Dante's discussion.

After having ranked the various verses in hierarchical order, led by the hendecasyllable, Dante turns to the canzone stanza. Just as the canzone includes the entire art of poetry, so also the canzone stanza, he says, includes the entire art of the canzone, the term stanza itself indicating the fact:

. . . hoc vocabulum per solius artis respectum inventum est, videlicet in quo tota cantionis ars esset contenta, illud diceretur stantia, hoc est mansio capax sive receptaculum totius artis. Nam quemadmodum cantio est gremium totius sententie, sic stantia totam artem ingremiat; nec licet aliquid artis sequentibus arrogare, sed solam artem antecedentis induere.
(2.ix.2 [1979a 206])

. . . this term stanza has been chosen for technical reasons exclusively, so that what contains the entire art of the canzone should be called stanza [room], that is, a capacious dwelling or receptacle for the entire craft. For just as the canzone is the container [literally, lap or womb] of the entire thought, so the stanza enfolds its entire technique; and successive stanzas are not permitted to introduce any new technical devices, but must clothe themselves in the devices set by the preceding.

In addition to speaking of the canzone and the stanza as "rooms" or "containers," Dante uses the revealing metaphor of a bundle of sticks, a fascis, in which the sticks correspond to the verses, the bindings to the factors—especially the grammatical constructions and the rhymes—that tie the lines together into a whole (2.v.8, 2.viii.1; cf. 2.iii.1, 2).^[55] The basic unit of construction is not so much the individual line (though

that must receive its due attention) as the group of lines. In this light Dante's description of the stanza form of the canzone is more comprehensible, particularly his assertion that the principal division among stanza types is between those with only one "melody" and those involving a transition from one "melody" to another, the latter always involving the repetition of musical phrases. In poetic terms, then, what Dante is referring to by "melody" is in fact a fixed bundle, a group of lines (of fixed length): the stanza is conceived as built out of a small number of such units (from one to five), depending on whether or not any are repeated within the stanza:

Quia quedam sunt sub una oda continua usque ad ultimum progressive, hoc est sine iteratione modulationis cuiusquam et sine diesi—et diesim dicimus deductionem vergentem de una oda in aliam (hanc voltam vocamus, cum vulgus alloquimur)—: et huiusmodi stantia usus est fere in omnibus cantionibus suis Arnaldus Danielis, et nos eum secuti sumus cum diximus "Al poco giorno e al gran cerchio d'ombra."
(2.x.2 [1979a 210])

For some are governed by one melody progressively from beginning to end, and this without repetition of any musical phrase and without diesis—diesis is the name we give to the passage from one melody to another (which we call volta when speaking to the unlearned).—Arnaut Daniel used this type of stanza in almost all of his canzoni, and we followed him when we wrote "To the short day and the great circle of shade."

In other words, the scheme of "Al poco giorno" would be ABCDEF (the lines are all hendecasyllables, hence the capital letters; the italics indicate rhyme-words rather than rhymes); the stanza has no division and no repetition (it presents the further peculiarity, not mentioned by Dante, of rhyming only between rather than within stanzas).

Other stanzas, however—most, in fact—have diesis and therefore at least one further subdivision. Dante's terminology for the subdivisions in the parts of the stanza is used today:

Quedam vero sunt diesim patientes: et diesis esse non potest, secundum quod eam appellamus, nisi reiteratio unius ode fiat, vel ante diesim, vel post, vel undique. Si ante diesim repetitio fiat, stantiam dicimus habere pedes; et duos habere decet, licet quandoque tres fiant, rarissime tamen. Si repetitio fiat post diesim, tunc dicimus stantiam habere versus. Si ante non fiat repetitio, stantiam dicimus habere frontem. Si post non fiat, dicimus habere sirma, sive caudam.
(2.x.3–4 [1979a 212])

Others, however, involve diesis: and there cannot be diesis, as we use the term, without the repetition of a melody, whether before the diesis, or after it, or both. If there is repetition before the diesis, we say that the stanza has pedes [feet]; and it is fitting for it to have two pedes, though occasionally it is given three, very rarely however. If repetition occurs after the diesis, then we say that the stanza has versus [turnings]. If there is none before, we say the stanza has afrons. If there is none after, we say it has a sirma, or tail.

In these terms the stanza of "Io son venuto al punto de la rota" has two pedes: ABC.ABC and a sirma: CDEeDFF (the last two letters are italicized to indicate repetition of the entire word). "Così nel mio parlar voglio esser aspro" also has two pedes and a sirma: ABbC.ABbC: CDdEE. "Donne ch'avete intelletto d'amore," however, has both pedes and versus: ABBC.ABBC:CDD.CEE; and "Amor, tu vedi ben che questa donna" is similar except that it has rhyme-words rather than rhymes: ABA.ACA:ADD.AEE (this stanza, however, would seem to depart from Dante's rule that any verse unrhymed in the first pes or versus must be answered in the second; he perhaps thought of the B and C rhyme-words, unanswered within the stanza, as instances of claves ["keys"] like those practiced by Gotto of Mantova, to whom he attributes the term in 2.xiii.5).^[56] Dante apparently wrote only one canzone with (undivided) frons, a lost poem he mentions in 2.xi.5. Dante also says (z.xiii.11) that the versus may be treated as separate from the concatenatio and concluding couplet;^[57] he wrote only one canzone of this type, "Io sento sì d'Amor la gran possanza." (See Figure 1 for these various stanzas; we have specified their rhyme schemes as well as line lengths for convenient reference, but Dante's discussion of the melodic divisions and subdivisions of the stanzas does not specify rhyme.)

Dante's statement that diesis requires the subdivision of one of the two parts of the stanza is not logically founded,^[58] though it does reflect Italian practice, which departs here from Provençal practice. We believe that the petrose reflect Dante's characteristic tendency to establish a metaphysical—cosmological—basis for this traditional practice.^[59]

As Dante points out, the rules of the canzone stanza offer the greatest possible freedom to the poet:

Vide ergo, lector, quanta licentia data sit cantiones poetantibus, et considera cuius rei causa tam largum arbitrium usus sibi asciverit; et si recto calle ratio te duxerit, videbis autoritatis dignitate sola quod dicimus esse concessum.
(2.X.5 [1979a 212])

See, therefore, reader, how much freedom is allowed to those who write canzoni, and consider for what reason custom has assigned to itself such large choice; and if reason leads along the right path, you will see that it is only on account of the dignity of authorship that this freedom has been granted.^[60]

It is in the disposition of the parts of the stanza, the establishing of the complex harmonies of proportion resulting from the interaction of the numbers of syllables and the number of lines from part to part, that the poet's mastery is especially shown; an entire chapter is devoted to this topic (2.xi), and another (2.xii) to the question of the proportion of hendecasyllables to settenarii. It is especially striking to the modern reader, who is used to thinking of stanzas in terms of rhyme schemes, that so far rhyme has not been referred to as in any way constitutive of the form of the stanza. Dante does think of rhyme as a unifying or binding factor, but he speaks of it as if, in the planning of the stanza, it were added to the lines after their number and order (and therefore the syntactic units they can hold) had been determined. He does not speak of it as a musical or melodic factor; there is no association between line endings and musical phrase endings (cadences or semicadences), and therefore the choice between alternate, equally permissible rhyme schemes does not affect the melodic structure of the stanza. Rather, Dante thinks of rhyme in rhetorical terms, as the rhetorical ornament of similiter desinens. In assigning the rhymes, the poet again has great freedom—virtually unlimited in frons and sirma, subject to a few rules in pedes and versus. He clearly thinks of a given "melodic" structure as capable of receiving a number of different rhyme schemes; given a first pes with the rhymes ABBC, for instance, the second pes could evidently be rhymed ABBC, ADDC, CDDA, BAAC, CAAB, and so forth; the first of these alternatives (i.e., ABBC.ABBC) is the scheme of the pedes in "Donne ch'avete intelletto d'amore." Dante's practice was mainly to have pedes that were duplicates in every respect, but the other alternatives are fully allowed for by his theory—as well as practiced by other poets.^[61]

There is no doubt some inconsistency in Dante's analysis, for if rhyme had no constitutive relation to stanza form, there would be no way of determining the division, or lack of it, of any canzone stanza whatever. In practice, and historically, rhyme is constitutive of stanza form in the vernacular, and it is precisely for that reason that Dante's exclusion of it from the numerical or spatial phase of stanza design is so significant. He is establishing ontological priorities among activities

[Full Size]

Fig. 1. 
Terminology of the Canzone (through page 29 )

that could hardly be separated in practice, and his priorities privilege the abstract, intellectualistic aspects of poetic creation, those which most clearly support the analogy with God's creation of the world.

One of the most interesting aspects of the metaphors Dante uses for the construction of the canzone stanza is that they envisage the establishing of the stanzaic scheme as a kind of demarcation of space. This is

[Full Size]

Fig 1
(continued)

perhaps more obvious in the metaphor of the room (which is, however, even more abstract than that of the bundle of sticks), since what is to occupy the room is the sententia of the poet and the provisions of his technique. The metaphor of the bundle of sticks is in a real sense closer to the literal, for if the sticks are the lines of verse that are bound together, then the bundle that is formed of these actual lines of verse literally does occupy space. The design of the bundle, then, which is prior to the writing of the lines, is the assignment of empty spaces; that is, the number of syllables in each line, the number of lines in each frons, pes, versus, or sirma, are the limits that demarcate the spaces. The activity of

[Full Size]

Fig. 1.
(continued)

designing the stanza is thus a kind of geometry, an assigning of spatial determinations to what was initially undifferentiated. Dante's terminology, from coartare (2.iii.1) to mansio to fascis to modum quo ligare (2.iii.2) to conditor (2.iii.6) to comprehendere (2.iii.8) and contenere (2.ix.2), reflects this conception, which looks back to the figure of God the geometer delineating with his compasses the space within which the world will exist.^[62] It has not, we believe, been observed previously that Dante's other terms for the stanza, receptaculum and gremium, seem to be

[Full Size]

Fig. 1.
(continued)

echoes of Plato's terms for space in the cosmogony of the Timaeus, as translated by Calcidius:

Quam igitur eius vim quamve esse naturam putandum est? Opinor, omnium quae gignuntur receptaculum est, quasi quaedam nutricula.
(Timaeus 46a [Plato 1962 46]; emphasis added)

What then shall we consider its power or its nature to be? I think that it is the receptacle of all things that come into being, as it were a kind of nurse.

Decet ergo facere comparationem similitudinemque impertiri illi quidem quod suscipit matris, at vero unde obvenit patris, illi autem naturae quae inter haec duo est prolis.
(Timaeus 50d [1962 48])

Let us use a comparison and say that that which receives is comparable to the mother, that which is the source to the father, and the nature that is between them to the child.

Calcidius's commentary introduces further terms:

Quae quidem corpora cum sola et per se ac sine suscipiente [ex] eadem essentia esse non possunt, quam modo matrem, alias nutriculam, inter-

dum totius generationis gremium, non numquam locum appellat quamque iuniores hylen, nos silvam vocamus.
(Plato 1962 277–278; emphasis added)

For since these bodies cannot exist without a recipient or by virtue of their essence alone, he speaks of the womb of all generation, which more recent writers call hylê and we call matter, sometimes as their mother, sometimes as their nurse, and frequently as space.

Although this theme is not made explicit in what survives of the De vulgari eloquentia, it is implied by its entire analysis; and, as we will try to show, the conception is the key to the form of the petrose, in which the principle of imitation of the musica mundana pervades the poems from the most abstract level of stanza design to theme and the shape of sententia (see Chapter 5, note 9).

The idea that human song imitates the cosmos is explicit in the Timaeus, of course, but Dante also knew of Plato's idea of the circling orbits of the planets as song (the idea later known as the music of the spheres) from Cicero's Dream of Scipio and Macrobius's Commentary on it, one of the most widely read books of the Middle Ages, and a principal source of Dante's knowledge of Platonic and Neoplatonic lore. Macrobius explains that human music imitates the music of the cosmos even in its strophaic forms:

Hinc Plato in Re publica sua cum de sphaerarum caelestium volubilitate tractaret, singulas ait Sirenas singulis orbis insidere significans sphaerarum motu cantum numinibus exhiberi. nam Siren dea cantans intellectu valet. theologi quoque novem Musas octo sphaerarum musicos cantus et unam maximam concinentiam quae confit ex omnibus, esse voluerunt . . . ideo canere caelum etiam theologi comprobantes sonos musicos sacrificiis adhibuerunt, qui apud alios lyra vel cithara, apud nonnullos tibiis aliisve musicis instrumentis fieri solebant. in ipsis quoque hymnis deorum . . . per stropham rectus orbis stelliferi motus, per antistropham diversus vagarum regressus praedicaretur ex quibus duobus motibus primus in natura hymnus dicandus deo sumpsit exordium.
(Commentary 2.3.1–6; Macrobius 1970a 103–105)

Therefore Plato, when in his Republic he came to treat of the revolutions of the celestial spheres, said that a Siren was sitting on each orb, signifying that the motion of the spheres was to the gods audible music. For Siren in Greek means a singing goddess. The theologians have explained the nine Muses to be the musical song of the eight spheres plus the one

greater harmony that is made up of all of them . . . and so, the theologians agreeing that the heavens are singing, they included music in their sacrifices, which was customarily made with lyre or harp or in many cases with pipes and other musical instruments. In the very hymns to the gods they demonstrated out of what two motions that first hymn sung to God by nature took its beginning, . . . in the strophe the direct motion of the sphere of fixed stars, in the antistrophe the varying returnings of the planets.

Exactly what Macrobius may have had in mind in comparing the classical strophe to the motion of the Same, the antistrophe to the motion of the Other, is difficult to say, since we cannot determine what kinds of performance of choral odes may have been accessible to him. But for Dante the basis of the analogy lies in the parallel between the recurring form of the stanza and the motion of the heavens as cyclical, as well as in the notion that the stanza is made up of contrasting or opposing motions. For, returning to the De vulgari eloquentia, diesis divides the canzone stanza into two parts governed by different melodies, different formal patterns. Diesis thus involves a transition to difference. In the stanza with diesis as practiced by Dante, there are always two pedes: that is, the stanza proceeds through a subdivision involving repetition of an identical scheme (the pes ); at the diesis the identity is abandoned and the stanza enters into its other motion, usually involving many more rhymes. In the canzone as a whole, then, the entire complex system of cycles and subcycles is repeated in each stanza, the principle of identity (the motion of the Same) thus governing and carrying forward the whole.

The elements of stanza form are not of Dante's invention, but he was tireless in seeking theoretical foundations for his practice, even—or perhaps especially—when it was based on tradition. And the influence of Dante's practice (though perhaps not the influence of the De vulgari eloquentia, which remained virtually unknown until published by Trissino in the sixteenth century) reinforced certain structural possibilities at the expense of others. Dante's preferences were certainly based in part on the ontological and cosmological considerations we have outlined. After him it became normative for canzoni to consist exclusively of hendecasyllables and settenarii, for stanzas with diesis to have two pedes and a sirma, and for them to be joined by what Dante calls a pulcra concatenatio —a last element of stanza form which in the petrose has clear cosmological significance. In the De vulgari eloquentia he explains that

the poet is free to interweave rhymes from the first part into the second (cf., once more, the stanza of "Donne ch'avete intelletto d'amore"):

Et quidam diversos faciunt esse rithimos eorum que post diesim carmina sunt a rithimis eorum que sunt ante; quidam vero non sic, sed desinentias anterioris stantie inter postera carmina referentes intexunt. Sepissime tamen hoc fit in desinentia primi posteriorum, quam plerique rithimantur ei qui est priorum posterioris: quod non aliud esse videtur quam quedam ipsius stantie concatenatio pulcra.
(2.xiii.7 [Dante 1979a 228–230])

And some make the rhymes after the diesis different from those before it; others do not so, but interweave rhymes from the first part of the stanza with those of the second. But this is done most frequently with the first line of the second part, which many rhyme with the last line of the first part: and this seems to be none other than a lovely chaining together [concatenation] of the stanza.

The concatenation is a particular kind of echo effect, which Dante uses skillfully for a variety of purposes in different poems; significantly, in the context of the microcosmic concerns of the petrose Dante thinks of it as a kind of concrete linking of the two parts of the stanza. In the petrose, in a small way, it is like the point at which the Demiurge joins the circle of the Same to the circle of the Other.