Zarlino on Syntax and Cadence
Zarlino wrote in the wake of a tradition that had long assumed a similarity between verbal and musical structures. Among medieval commentators the notion that music reflected grammatical relationships through its cadences was a commonplace. Virtually all commentators related the grammatical hierarchy of comma, colon, and period to musical phrasing. Fifteenth- and early-sixteenth-century theorists of polyphony, however, had gone some way toward severing this connection for the sake of counterpoint (as we saw in Lanfranco's Scintille and in Spataro's retort to Aaron). In his writing about cadences in polyphonic music, Aaron therefore had to make a point of advocating (albeit tentatively) that composers resume the old grammatical alliances, arguing in his Trattato that without the proper cadences a texted composition simply could not be well understood: "The cadence is a sign by which composers make an indirect ending ["mediato fine"] according to the sense of the words."[105]
Extending the recent tradition of rhetorically minded polyphonic theorists (including del Lago), Zarlino insisted repeatedly that the cadence in polyphony needed to correspond to meaningful divisions of words. In the Istitutioni, admonishments appear prominently in Part III, Chapter 53, on cadences, where only after emphasizing several times the need to coordinate musical with verbal cadences Zarlino admits that cadences in polyphony sometimes do "emerge from a certain con-
[103] "Varietà molto piacere porge alli nostri sentimenti. Debbe adunque ogni Compositore imitare un tale, & tanto bello ordine: percioche sarà riputato tanto migliore, quanto le sue operationi si assimigliaranno a quelle della Natura" (Istitutioni harmoniche, p. 177).
[104] "La natura odia le cose senza proportione, & senza misura; & si diletta di quelle, che hanno tra loro convenienza" (ibid., p. 185).
[105] "Cadenza non è altro che un certo segno del quale gli Compositori per alcun senso delle parole fanno un mediato fine" (Trattato, Chap. 8 [unpaginated]).
trapuntal design initiated by the composer."[106] Lanfranco had described this circumstance as the norm; for Zarlino it was something to be conceded "out of necessity."[107] No other theorist of the Renaissance showed as clearly as Zarlino how to reflect textual syntax and (though to a lesser extent) meaning in music. This is not inscribed so much in his rules of text underlay — many of which, as Don Harrán has shown, are derived from Lanfranco[108] — but in the totality of his teachings on cadences, rests, text repetition, and underlay.
In Part III, Chapter 50 ("Delle pause"), Zarlino theorized questions of rests (like so many others) through the principle of variety. Rests enabled composers to vary the numbers of voices singing at any given point in a work in order to make compositions "lovely and delightful." Just as important, rests functioned like cadences in demarcating text. They were "not to be placed except at the ends of clausule or punti in the text . . . and at the end of every periodo. " Composers were to make sure that rests allowed the parts of the text and the meanings of the words to be fully heard and understood, so that they would have a "true function" and not just be random: "In no case should they be placed . . . in midclause, for anyone who would do so would show himself to be a veritable sheep, an oaf, and an ignoramus."[109] In Part IV, Chapter 32, on accommodating harmonies to words, Zarlino added that composers should neither make a cadence nor use a rest larger than a minim in any place where the meaning of the words is incomplete.[110]
The attention Zarlino gave to meaningful divisions of the text included textual repetitions, which he otherwise rarely discussed. The portion of text repeated, he
[106] "[S]i debbe avertire, che le Cadenze nelli Canti fermi si fanno in una parte sola: ma nelli figurati si aggiungono altre parti. Et in quelli si pongono finita la sentenza delle parole; in questi poi non solamente si fanno, quando si ode la Clausula perfetta nella oratione; ma alle volte si usano per necessità, et per seguire un certo ordine nel Contrapunto, principiato da Compositore" (Istitutioni harmoniche, p. 221).
[107] Gaspar Stoquerus expressed the relationship of words and cadences similarly, at least by implication, in De musica verbali of ca. 1570, also based on Willaert's practice. Stoquerus made the traditional equation of music and speech and exalted those composers who were actually capable of realizing it: "therefore for the more learned and careful musicians, a phrase is the same as for the grammarians" (emphasis mine) (Quare doctis quidem et diligentioribus musicis sententia eadem est cum grammaticis; Chap. 20, fol. 30'; See De musica verbali, libri duo: Two Books on Verbal Music, ed. and trans. Albert C. Rotola, S.J., Greek and Latin Music Theory [Lincoln, Neb., and London, 1988], pp. 204-5).
[108] See "New Light on the Question of Text Underlay prior to Zarlino," Acta musicologica 45 (1973): 24-56.
[109] "[N]on porre tali Pause, se non nel fine delle Clausule, o punti della Oratione, sopra la quale è composta la cantilena, & simigliantemente nel fine di ogni Periodo. Il che fa dibisogna, che li Compositori etiandio avertiscano; accioche li Membri della oratione siano divisi, & la sentenza delle parole si oda, & intenda interamente: percioche facendo in cotal modo, allora si potrà dire, che le Pause siano state poste nelle parti della cantilena con qualche proposito, & non a caso. Ne si debbeno porre per alcun modo, avanti che sia finita la sentenza, cioè nel mezo della Clausula: conciosia che colui, che le ponesse a cotal modo, dimostrarebbe veramente essere una pecora, un goffo, & uno ignorante" (Istitutioni harmoniche, p. 212).
[110] "Si debbe . . . avertire, di non separare alcuna parte della Oratione l'una dall'altra parte con Pause, come fanno alcuni poco intelligenti, fino a tanto, che non sia finita la sua Clausula, overo alcuna sua parte; di maniera che 'l sentimento delle parole sia perfetto; & di non far la Cadenza; massimamente l'una delle principali; o di non porre le Pause maggiori di quelle della minima, se non è finito il Periodo, o la sentenza perfetta della Oratione" (ibid., p. 340).
insisted, should make sense as a unit. One should therefore avoid single-word and (needless to say) single-syllable repetition, as explained in the eighth of his rules of text underlay, Part IV, Chapter 33: "[I]n mensural music such repetitions are tolerated; I am certainly not talking about one syllable or one word, but of some part of the text whose idea is complete."[111] Textual repetitions, when they did occur, were to be kept to a minimum and reserved for words bearing "serious thoughts worthy of consideration."[112]
The same view of symbiotic syntactic and semantic functions that governed Zarlino's recommendations on text repetition and cadence led him to insist on faithful musical accentuations of words. Following theorists from Guido to Gaffurius and beyond, he held that long syllables should be matched to long notes and short ones to short notes,[113] placing this at the head of his famous decalogue, or ten rules of text setting. Yet in doing so he was not just nodding obediently toward old grammatically based traditions of music theory.[114] Neither, as Don Harràn has argued, did he mean for the rule to apply only to Latin quantitative verse, as the preponderance of Italian madrigals among his examples attests.[115] Instead, his attitude corresponded to that of Vicentino's L'antica musica ridotta alla moderna prattica, published three years earlier, which lamented that too many composers were inclined to go their own compositional way without considering the nature of words, their accents, and syllable length, whether in the vernacular or in Latin.[116]
More than Vicentino, though, Zarlino conceived the coordination of syllable and musical length in the specific context of Venetian repertory, as suggested in the wording of rule I: "The first rule then will be always to place on a long or short syllable a corresponding note, in such a way that no barbarism is heard. For in mensural music every note that is separate and not tied (excepting the semiminim and all notes smaller than the semiminim) carries its own syllable."[117] Here, first of all,
[111] "[N]el [canto] figurato tali repliche si comportano; non dico gia di una sillaba, ne di una parola: ma di alcuna parte della oratione, quando il sentimento è perfetto" (ibid., p. 341). This conforms to what Stoquerus later said, allowing for repeats of sentences, but not individual words. See Harràn, Word-Tone Relations in Musical Thought, p. 460, nos. 380-81.
[112] "[I]l replicare tante fiate una cosa (secondo 'l mio giuditio) non stia troppo bene; se non fusse fatto, per isprimere maggiormente le parole, che hanno in se qualche grave sentenza, & fusse degna di consideratione" (Istitutioni harmoniche, p. 341).
[113] See Harràn, In Search of Harmony: Hebrew and Humanistic Elements in Sixteenth-Century Musical Thought, Musicological Studies and Documents, no. 42 (Neuhausen-Stuttgart, 1988), pp. 92ff.; and idem, Word-Tone Relations in Musical Thought, pp. 10-15 and 375-87.
[114] See Franchinus Gaffurius, Practica musicae (Milan, 1496), 2.1, fol. 21', who cites the Latin grammarians on the point, ed. and trans. Clement A. Miller, Musicological Studies and Documents, no. 20, AIM ([Rome], 1968), pp. 69-72; facs. ed. BMB, ser. 2, no. 6 (Bologna, 1972).
[115] In Search of Harmony, pp. 92-97.
[116] "Molti Compositori che nelle loro compositioni attendono à far un certo procedere di compositione à suo modo, senza considerare la natura delle parole, ne i loro accenti, ne quali sillabe siano lunghe ne brevi, cosi nella lingua volgare come nella latina" (L'antica musica, Book 4, Chapter 29, fol. P i'); facs. ed. Edward E. Lowinsky, Documenta musicologica, ser. I, Druckschriften-Faksimiles, no. 17 (Kassel, 1959).
[117] "La Prima Regola adunque sarà, di porre sempre sotto la sillaba longa, o breve una figura conveniente, di maniera, che non si odi alcuno barbarismo: percioche nel Canto figurato ogni figura quadrata si accommoda la sua sillaba" (Istitutioni harmoniche, p. 341).
Zarlino explicitly exhorts composers to syllabic text setting, which was normative in midcentury Venetian polyphony. As rule 4 specifies, however, only in rare circumstances — justified expressively — could the declamation occur on semiminims and smaller notes; the actual rhythmic structures of the style are such that most of the syllabic declamation occurs on semibreves and minims.[118] In adding in rule I that each note should carry one syllable as long as the note does not fall within a ligature and is not smaller than a minim, Zarlino presupposed the composite minim rhythm that pulsates at the declamatory surface of Venetian repertory when the voices are configured in their typical staggered relationships. The longs and shorts that he, along with his predecessors, relied on as a traditional means of description hardly pointed to strict quantitative poetics, with two values corresponding to a spoken syllable length. Rather, a system of two surface durations of a 2:1 ratio now reflected local practice, which confined itself almost solely to the declamatory values of semibreve and minim, organizing the syllables in a free arrangement of fluctuating tonic and agogic stresses.[119]
All of this is considerably simpler than the questions surrounding cadence. Along with rests, cadences formed one of the two main devices articulating verbal syntax. Zarlino surpassed any previous polyphonic theorist in trying to account for the multiplicity, variety, and subtlety of cadential events that were possible in an irregular contrapuntal texture. In such a context the tenor-superius framework on which cadences had traditionally hinged was all but gone or, at minimum, was seriously threatened. The main melodic materials — the soggetti — could not necessarily be found in any particular voice or imitated in any particular form, nor would they necessarily even be imitated at all. Even the way text was divided did not warrant a separate category of articulation, as Zarlino made clear in his definition of cadence at the outset of Part III, Chapter 53, "On the Cadence, what it is, its species, and its use." "The cadence," he explained, "is a certain act that the voices of a composition make in singing together, which denotes either a general repose in the harmony or the perfection of the sense of the words on which the composition is based. Or we could say that it is a termination of a part of the whole harmony at the middle or end, or an articulation of the main portions of the text."[120] A text-determined environment of
[118] Rule 4 reads: "La Quarta, che rare volte si costuma di porre la sillaba sopra alcuna Semiminima; ne sopra quelle figure, che sono minori di lei; ne alla figura, che la segue immediatamente" (ibid.). In rule 6, Zarlino adds that when one syllable is placed under a semiminim another may be placed under the following note: "La Sesta, quando si porrà la sillaba sopra la Semiminima, si potrà anco porre un'altra sillaba sopra la figura seguente" (ibid.).
[119] See the respective Latin- and Italian-texted musical examples from the Musica nova in Harrán, In Search of Harmony, pp. 96-97.
[120] "La Cadenza adunque è un certo atto, che fanno le parti della cantilena cantando insieme, la qual dinota, o quiete generale dell'harmonia, o la perfettione del senso delle parole, sopra le quali la cantilena è composta. Overamente potemo dire, che ella sia una certa terminatione di una parte di tutto 'l concento, & quasi mezana, o vogliamo dire finale terminatione, o distintione del contesto della Oratione" (Istitutioni harmoniche, p. 221).
the kind Zarlino describes here made it difficult to distinguish the purely musical components of cadences. Zarlino opted to define cadential events as comparable syntactically (at least in theory), whether they brought to an end harmony or text (or both). Thus, certain passages might be deemed cadences that were not articulative in contrapuntal terms. In principle, cadences could even include events involving neither the suspended dissonances nor the imperfect-to-perfect intervallic movements that we normally associate with classic Renaissance polyphony.
Nevertheless, cadences that consist of contrapuntally discernible structures do have clearly defined musical features.[121] According to Zarlino's description, they routinely consist of two voices moving through three successive dyads, with at least one of the simultaneities progressing in contrary motion. Zarlino first says that cadences arrive on unisons or octaves, then admits that they may also arrive on other intervals, such as fifths or thirds (and even additional ones). Cadences on octaves and unisons are perfect (i.e., with a major sixth to octave progression for the last two of the three dyads, or a minor third to a unison), while those on other intervals are imperfect since they are not strictly speaking cadences ("non assolutamente cadenze"). Later on Zarlino adds that an imperfect cadence is a "cadenza impropiamente" (improper cadence).[122] Either type, perfect or imperfect, may be simple — that is, unornamented, homophonic, and fully consonant; or diminished — that is, varied rhythmically and including dissonant suspensions.
Once Zarlino begins to discuss specific cases he implicitly calibrates cadences from strong to weak, though decidedly not in any rigid, categorical, or (often) explicit way. For Zarlino perfect cadences carried the most weight. They were to be used when ending all the parts in a work or when punctuating complete sentences.[123] Zarlino's stance on them was unequivocal, but in his detailed contrapuntal descriptions he made other (sometimes ambiguous) refinements regarding weight. Imperfect cadences, by contrast with perfect ones, were to be used for "intermediate divisions of the harmony or for places where the words have not completely reached the end of their thought." Whether ending on thirds, fifths, or sixths, imperfect cadences always had to move into their last sonority from a third and with the upper voice ascending stepwise.[124] As with his descriptions of the soggetto and imitatione, Zarlino saw that many instances in current practice transcended his guidelines, making it impossible to illustrate every case: cadence types
[121] For what follows I am indebted to Meier, The Modes of Classical Vocal Polyphony, Part I, Chap. 4; Siegfried Hermelink, "Über Zarlinos Kadenzbegriff," in Scritti in onore di Luigi Ronga (Milan, 1973), pp. 253-73; Karol Berger, Musica ficta: Theories of Accidental Inflection in Vocal Polyphony from Marchetto da Padova to Gioseffo Zarlino (Cambridge, 1987), pp. 122-39; and Michele Fromson, "The Cadence in Zarlino's Le istitutioni harmoniche," Chap. 1 in "Imitation and Innovation in the North Italian Motet, 1560-1605" (Ph.D. diss., University of Pennsylvania, 1988). Fromson provides a close reading of Zarlino's Part III, Chapters 53 and 54 and parts of 61, with astute analyses of his musical examples.
[122] Istitutioni harmoniche, p. 224.
[123] "[S]e le Cadenze furono ritrovate, si per la perfettione delle parti di tutto il concento; come anco, accioche per il suo mezo si havesse a finire la sentenza perfetta delle parole" (ibid., p. 225).
[124] Ibid., p. 224.
Ex. 7.
Zarlino, Le institutioni harmoniche (Venice, 1558), p. 225.
are "almost infinite," he claimed, "for the contrapuntist . . . seeks new ones, constantly searching for new procedures."[125]
In order to suggest some of the possible exceptions, Zarlino embedded in a passage of continuous polyphony a group of examples that he later, in the 1573 edition of the Istitutioni, dubbed "extravagant cadences." In this passage (Ex. 7), we find a welter of melodic syncopations that descend and return by step, but no orthodox cadential progressions. Six of these syncopations occur in the lower voice (mm. 1-2, 2-3, 4-5, 6-7, 9-10, 11-12) and just one in the upper voice (the last one, mm. 13-14). The nonsyncopated voice in some cases makes an unconventional descent by leap to the cadential pitch (m. 3, g to c; mm. 4-5, g to c; m. 10, e to c; and m. 12, c to a); in other instances it stays put from the penultimate to the final sonority, instead of changing pitch at the point of cadence, as usually happens (e.g., the c over mm. 1-2; the c in m. 7). Furthermore, a number of the syncopations do not form dissonances with the counterpoint (as in mm. 1, 7, 10, and 13). What Zarlino seems to be drawing attention to here is the cadential impulse inherent in these syncopated melodic figures; he seems to be grappling with the Willaertian blur between such syncopated figures and the expectations of cadence that they set up in the listener.
Since the contrapuntal events in Ex. 7 lack orthodox cadences, their implied hierarchic strength will be low. In principle, they should carry less weight on our deduced scale than those described in Chapter 54 under "fuggir la cadenza," or "evading the cadence." Zarlino describes the latter as involving a "certain event in which the parts, pointing toward the desire to make a perfect ending according to one of the procedures shown above, turn elsewhere."[126] In other words, unlike the variety of figures grouped under extravagant cadences, evaded cadences should theoretically have all the necessary conditions of true cadences, up until the final sonority.
[125] "Sarebbe cosa molta tediosa, se io volesse dare uno essempio particolare di ogni Cadenza propia, & non propia; conciosia che sono quasi infinite; onde è dibisogna, che'l Contrapuntista s'ingegni di ritrovare sempre di nuove, investigando di continuovo nuove maniere" (ibid.).
[126] "'[L]Fuggir la Cadenza sia . . . un certo atto, il qual fanno le parti, accennando di voler fare una terminatione perfetta, secondo l'uno de i modi mostrati di sopra, & si rivolgono altrove" (ibid., p. 226).
Ex. 8.
Zarlino, Le istitutioni harmoniche, mm. 1-19 (Venice, 1558), p. 226.
But in real use the designation "fuggir la cadenza" turns out to be a catchall for any kind of cadence that appears to start out normally and ends up going astray. This includes imperfect cadences ending on thirds, fifths, or sixths in connection with which Zarlino first introduces the term at the end of Chapter 53. Evaded cadences are useful, Zarlino tells us, "when the composer has a lovely passage on his hands which would suit a cadence perfectly, but the words have not come to an end." Then it would "not be honest" to insert one, so the composer instead evades it.[127] To show techniques for evading the cadence, apart from the imperfect cadences already demonstrated, Zarlino again adduces a lengthy continuous passage, thirty-five breves in all (about half of which is given in Ex. 8).
The passage begins with a canon at the fifth, spaced at the semibreve. The guide in m. 2 sets up a typical syncopated cadential figure but fails to return by ascent to the pitch of preparation. Owing to the close spacing of the canon, the consequent initiates an identical syncopated figure in the middle of m. 2, so that the two end up overlapping in mm. 2-3. Had the upper voice returned to the f at the beginning of m. 3, the cadence would still have been unorthodox, since the lower voice would
[127] "[Q]uando si vorrà fare alcuna distintione mezana dell'harmonia, & delle parole insieme, le quali non habbiano finita perfettamente la loro sentenza; potremo usar quelle Cadenze, che finiscono per Terza, per Quinta, per Sesta, o per altre simili consonanze: perche il finire a cotesto modo, non è fine di Cadenza perfetta: ma si chiama fuggir la Cadenza; si come hora la chiamano i Musici. Et fu buono il ritrovare, che le Cadenze finissero anco in tal maniera; conciosia che alle volte accasca al Compositore, che venendoli alle mani un bel passaggio, nel quale si accommodarebbe ottimamente la Cadenza, & non havendo fatto fine al Periodo nelle parole; non essendo honesto, che habbiano a finire in essa; cerca di fuggirla, non solamente al modo mostrato: ma nella maniera ch'io mostrerò nel seguente capitolo" (ibid., p. 225).
then have anticipated its own cadential pitch. By moving instead to d, it forms a suspended 2-3 dissonance with the consequent. Zarlino fails to sustain the d long enough for it to sound against the lower voice's resolution to b, leaping instead to g in m. 3 on the second part of the implied cadence's second sonority. The lowervoice consequent then undergoes the same lack of resolution as had the guide, b moving to a in m. 3. To double the injury done to the expected cadence at the third beat of that measure, the upper voice leaps again, now down to f and finally (in m. 4) to c.
More obvious deflections away from cadences affect the syncopated voice later in the passage. In m. 6, it leaps up from e to g. In m. 11 what had seemed to be a perfectly prepared Phrygian cadence to aa/a finds the diminished voice dropping suddenly a fifth to c to form an evaded cadence on a minor third. And in m. 19, the Phrygian cadence set up so conventionally is undercut by the sudden disappearance of the upper voice at the moment of final cadence.
To summarize, although Zarlino's discussion does not fix hierarchical relationships of cadence and text in an absolute or exhaustive way, certain hierarchical relationships emerge implicitly from his remarks. Perfect cadences, which mark off complete sentences or thoughts, rank highest.[128] They sit above imperfect cadences, which mark off incomplete sections of text. Evaded cadences seem to carry syntactic weight in proportion to the extent to which they set up cadential expectations. Thus if all the conditions of cadence are present but one (such as the actual cadential pitch in one of the voices), then the evaded cadence may rank fairly high. Those that go astray in more ways, and hence imply cadences less strongly, are weaker syntactically. The so-called extravagant cadences, having little resemblance to conventional cadences apart from the syncopated melodic figure, would seem to be weakest of all.[129]
Taken together, Zarlino's remarks on cadences, rests, textual repetition, and text underlay constitute a novel effort to conceptualize the new relation between text and music in the highly irregular and variegated polyphony of Willaert's circle. Zarlino fused his broad learning and acculturation in literary circles with his exceptional musical perspicacity and practical experience to codify Willaert's vocal polyphony within the terms of current rhetorical issues. These combined gave him a unique place in mediating between the new preoccupations of Venetian literary culture and the long-standing technical traditions of polyphonists bred in the church.
[128] It might seem to us that perfect cadences would be strongest when their lowest voice leaps up a fourth or down a fifth, yet Zarlino asks that such leaps be confined mainly to the bass in multivoice polyphony. This is because he views cadences essentially as two-voice contrapuntal progressions.
[129] We might have hoped Zarlino's ranking of appropriate cadential degrees would complement the foregoing, but he was not very effective in this respect. His chapters on individual modes (Part IV, 18-29) subsume each under a rationalistic arithmetic system that unrealistically favors root, third, and fifth. This directly contradicts contemporary practice (as pointed out by Palisca, On the Modes, pp. xiii-xvi; Meier, The Modes of Classical Vocal Polyphony, pp. 105-7; and Carl Dahlhaus, Untersuchungen über die Entstehung der harmonischen Tonalität [Kassel, 1968], p. 198). Aaron gives a better idea of how modal counterpoint was actually being written (Trattato, Chaps. 9-12), even though his discussion is largely based on chant.