“… nessun libro che parla d’un libro dice di più del libro in questione”
Italo Calvino, Perché leggere i classici
“… what can be said about a work of art can never outsay what a work of art says about itself”
Alan Bennett, Untold Stories
“Analysis is an abominable business.”
John Ruskin to Charles Eliot Norton, May 1857
Earlier scholars of Frescobaldi devoted a large part of their attention to speculation on the sources and formation of his keyboard style. Such speculation generally started from a preoccupation with the most daring and experimental aspects of that style, exemplified in the toccatas and in the chromatic works, and undervalued their counterpoise in the more traditional contrapuntal genres of fantasia, ricercar, and canzona. This led to two contrasting distortions of Girolamo’s musical personality. In the first, he was presented as a unique genius with no historical antecedents, a figure of “grandezza e solitudine.” In the second, the most striking features of his style were taken out of context and traced to a single composer, school, or musical device, the more unlikely the better—Cabezón, Sweelinck, the Neapolitans, the inganno.
Frescobaldi himself acknowledged only one teacher and one formative influence on his music, Luzzasco Luzzaschi and the Este court at Ferrara. Our examination of Ferrarese music during Frescobaldi’s youth has shown that this rich musical milieu was capable of providing the nourishment necessary for a budding musician. The sources of Frescobaldi’s keyboard style are therefore to be sought first of all in the works of Luzzaschi and other composers connected with Ferrara. Indeed, in his Ricercari of 1669 Luigi Battiferri (ca. 1610-82), maestro of the Ferrarese Accademia dello Spirito Santo, praised the organists of Ferrara for cultivating the ricercar, “this form [which is] almost extinct, with the death of those persons” (“quasi estinta questa forma, con la morte di quei soggetti”).
However, the list of musical visitors to Ferrara and the contents of Alfonso II’s music library show that the duchy was also open to outside musical influences. It is possible to establish at least a general context for the examination of Frescobaldi’s earliest keyboard publications by summarizing the most important printed collections (the dating of manuscripts is usually so uncertain as to render them problematic for a chronological study) of works in strict style from the mid-sixteenth century up to 1607, when Frescobaldi was already at work on the Fantasie, and toccatas, variations, canzonas, and of dances up to 1614, when Toccate I had been compiled in its first state and the Recercari et canzoni was presumably being prepared for publication.
There is one stumbling-block in examining possible sources of Frescobaldi’s earliest contrapuntal works: the question of intended medium. Collections could be variously designated “to sing or to play.” The “to play” (i. e., instrumental) category poses the question of keyboard vs. instrumental ensemble. In his history of keyboard music (Apel 1972) Willi Apel omitted from his account of the development of contrapuntal keyboard genres such as the ricercar works in part-book format, which he considered as intended for instrumental ensemble rather than keyboard compositions. Such rigid categorization is untenable.
Three notations for keyboard music were available: four-staff keyboard partitura (open score); two-staff keyboard tablature (intavolatura); and partbooks. Of the nearly fifty known manuscripts and prints (some lost) employing partitura, including six volumes by Frescobaldi, the majority are connected with either Ferrara or Naples, and it may be that Ferrarese practice influenced the Neapolitan composers. (Probably the line of influence runs Willaert-Rore-Luzzaschi-Frescobaldi.) Collections specifically designated for organ on their titlepages could nonetheless be published in the form of partbooks. The Intabolatvra d’organo di recercari (Venice: Gardane, 1549) of Jacques Buus (c. 1500-1565) contains four ricercars in keyboard tablature. Its companion volume, Il secondo libro di recercari (Venice: Gardane, 1549), headed “to Sing, & play on the Organ & other [keyboard] Instruments” (“da Cantare, & sonare d’Organo & altri Stromenti”) is presented in four partbooks, but the first partbook ricercar duplicates the first ricercar of the keyboard collection.
Numerous titlepages indicate an overlap between keyboard and instrumental ensemble. Adriano Banchieri’s L’Organo suonarino (Venice: Amadino, 1603) comprises “Fantasies or canzoni alla francese to play on the organ and other musical instruments” (“Fantasie overo canzoni alla francese per suonare nell’organo et altri stromenti musicali”). In his Partitura moderna armonia di canzoni alla francese (Venice; Amadino, 1612) Banchieri stipulated that not only might the canzoni in partitura be played on organ or harpsichord but that they might also be concerted with one or two instruments, high and/or low. Frescobaldi himself published his 1628 Canzoni in partbooks, while his pupil Grassi issued them transcribed them for keyboard in open score.
Perhaps the most obvious example of this permeation of boundaries is the Turin codices, a collection of sixteen keyboard manuscripts in the Biblioteca Nazionale Universitaria of Turin now subdivided into the Foà and Giordano series, named for their donors. The volumes contain 1,770 pieces of music from a variety of printed and other sources, all copied between 1637 and 1640 in New German keyboard notation, a staffless alphabet tablature. Volumes VI, VII, VIII, and IX contain ricercar-like works (including contrapuntal sections ripped from their context in toccatas), whose originals were scored both in partitura and intavolatura. Vol. VIII (copied 1639-40: Felici 2005, 101), a volume of ricercari by Italian composers, is entirely taken from editions published in parts.
The centers of publication for the most notable keyboard collections of the later sixteenth century and earlier seventeenth centuries parallel the principal Italian schools of keyboard composition. Venice was clearly preeminent, attracting not only the organists of St. Mark’s and other local musicians but also composers from Ferrara, Brescia, Bologna, and other north Italian cities. The Neapolitan houses seem to have published only local composers, especially pupils of Giovanni [Jean] de Macque (1548/50-1614), most of whose own keyboard works survive only in manuscript. Milanese publishing was also largely parochial in character, with a preference for partitura in libro aperto format (“open book”—lines of music running across both pages of an opening) in the productions of Frescobaldi’s first keyboard publisher, Tini and Lomazzo (Newcomb 1965, 117-18). Roman keyboard publication was almost non-existent except for collections printed in two-staff intavolatura di cimbalo from engraved copper plates rather than the more usual and less expensive technique of movable type.
The serious contrapuntal ricercar (as opposed to the earlier improvisatory type derived from lute music) first appeared in printed keyboard music in Girolamo Cavazzoni’s Intavolatura of 1543. All four of Cavazzoni’s ricercars, which he described as youthful works, are extended, carefully crafted contrapuntal pieces like instrumental motets. They are based on several successive subjects but overt sectionalization is generally avoided by elided cadences and other types of transition between sections. Although the basic movement is a decorated note-against-note texture, there also occur livelier cadential ornaments and even running figurations in treble or bass against slower-moving chords. The last two ricercars contrast the prevailing duple meter with concluding triple sections, rounded off by duple cadences. The clarity of general conception and the flexibility of detail found in Cavazzoni’s ricercars continued to characterize the genre in the collections of his successors. In his Ragionamento (1588), Pietro Ponzio (1532-96) named some leading composers of ricercari in discussing compositional procedure: “[in the ricercare] it is equally permitted to repeat two, three, four & more times the same idea [inventione], in various ways as will please you, as one can see in the Recercari of Jacques Buus, and of Annibale Padovano, Claudio [Merulo] of Correggio, & of Luzzasco” (“Parimente sia lecito replicare due, tre, quattro & più volte la istessa inventione, sì come vi piacerà per variati modi, come si può vedere nelle Recercari di Iaches Bus, e d’Anibale Padovano, di Claudio da Correggio, & del Luzzasco”: quoted in Durante 1998, I, 9, n. 2).
Mode and Cleffing
Many collections of serious contrapuntal keyboard works were ordered modally. The older system, derived from chant, consisted of eight modes in two configurations (authentic, with a range an octave above the finalis, and plagal, ranging a fifth above and a fourth below the finalis) with finales on on d, e, F, and G. The Dodekachordon of Heinricus Glareanus (1488-1563) of 1547 added two further modal pairs with finales on a and C, a system adopted by Gioseffo Zarlino (1517-90) in his Istituzioni armoniche of 1558. These modes could be transposed up or down a fourth or fifth with the addition of B-flat (notably the standard transposition of mode I Dorian on d down a fifth to g).
Various combinations of clefs were employed for both vocal and instrumental works. In Girolamo Diruta’s keyboard treatise Il Transilvano (Venice: Vincenti, vol. 1, 1593; vol. 2, 1609) he arranged the modes in pairs according to Zarlino with finales on d, e, F, and G and employed several cleffings: natural clefs (chiavi naturali): C1, C3, C4, F4 and a variant of chiavi naturali, C1, C2, C3, F3; low or transposed clefs (chiavi basse, C2, C4, F3, F5). For a higher tessitura in modes on fa, sol, la, or ut or transposed re and mi modes, high clefs (“all’alta”) were used: G2, C2, C3, F3 [variant: G2, C1, C2, C4: Felici 2005, 11]. (These were later called chiavette, a term first known from a Roman manuscript of 1718 copied by one Girolamo Chiti.) Vocal compositions in chiavette were usually transposed down a fourth or fifth in performance, but instrumental works in chiavette were sometimes left untransposed since “playing thus at high pitch they make the harmony more lively” (“suonando così all’alta fanno più viva l’armonia”: Banchieri, Cartella, 21-24). In the Turin manuscripts most of the pieces originally notated in chiavette were transposed down.
Frescobaldi’s knowledge of one set of contrapuntal works by a northern composer closely connected with Ferrara is revealed in a letter from Bernardo Bizzoni (see Chapter 2) in Rome to Enzo Bentivoglio in Ferrara, 22 August 1607: “I have had in the post the Compositions [Musiche] that Your Most Illustrious Lordship was pleased to send me, and they are old Ricercari without words by Adriano made only to be played, and they were left by our messer Geronimo Frescobaldi on his departure for Flanders with my Master [Guido Bentivoglio] to Signor Luzzasco for him to send to me”—perhaps Frescobaldi’s last personal encounter with his teacher. “I kiss Your Most Illustrious Lordship’s hands for the care that you have taken to have them reach me.”
Bizzoni thus acknowledges the receipt from Enzo of an old collection of instrumental ricercari by Willaert that Frescobaldi had left with Luzzasco to be sent to Bizzoni on Girolamo’s departure for Flanders with Guido Bentivoglio. This was presumably Willaert’s FANTASIE RECERCARI / CONTRAPVNTI A TRE VOCI DI /M. Adriano & de altri Autori appropriati per Cantare & Sonare d’ogni / sorte di Stromenti … (FANTASIES RECERCARI / COUNTERPOINTS FOR THREE VOICES OF / M. Adriano & other Authors appropriate to Sing & Play by every / kind of Instrument …), Venice: Gardano, 1551, reprinted in 1559 (the only complete surviving example) and in 1593: in three part books, CTB.
The study of Willaert’s three-part counterpoints fits into a pedagogical pattern. Zarlino (Istituzioni III, 59) and following him Diruta (1593, II, p. 14) treats writing for three parts as a necessary transition from the rules of note-against-note two-part counterpoint to full four-part writing. The publisher Giovanni Angelo Mutij, who claimed Frescobaldi as his teacher in 1673, singled out (convincingly) as elemtents of his instruction “playing from score [partitura] and making Contrapunti à 3.”
Willaert’s Fantasie opens with settings of Regina Celi by Willaert and Rore above the same cantus firmus, followed by nine ricercars of Willaert (I-VIII, X), three ricercari on Gregorian cantus firmi tenors in equal notes by Willaert’s friend Antonio Barges (XII-XIV), two ricercari of “incerto autore,” and one by Girolamo Cavazzoni. The finales of the Willaert ricercari display no apparent modal order. The continuing currency of Willaert’s Fantasie is evident in the fact that they were transcribed almost a century later in publication order (with much additional ornamentation) in volume VIII of the Turin keyboard tablatures, which is dated 1639-40.
Willaert’s ricercars do not seem to have provided a close model for Frescobaldi’s Fantasie, although the identical subjects of Willaert’s nos. I and II are rhythmically and melodically close to the subject of Frescobaldi’s 1608 Fantasia Nona. Willaert overlaps sections to provide a continuous polyphonic texture, unlike the clear sectional divisions of Frescobaldi’s fantasies, which are defined by such devices as augmentation, diminution, and variations of motivic density. Willaert frequently employs inversion to generate new motivic material, as do Luzzaschi and Frescobaldi. Triple sections occur in Willaert’s ricercars I and VI. The long opening duet in ricercar IV recalls the opening of Frescobaldi’s Recercar quinto.
Perhaps the earliest keyboard ricercars of possible influence on Frescobaldi were those of Claudio Merulo, first printed in 1567 and reissued in 1605. The eight ricercars of Merulo’s set follow the order of the eight modes, although a transposed C mode is substituted for the F mode without B-flat. Owing both to their length—an average of one hundred breves—and to the introduction of as many as eight successive subjects, Merulo’s ricercars seem deficient in motivic concentration and contrapuntal artifice.
Four authentic keyboard works by Frescobaldi’s teacher Luzzasco Luzzaschi had been known previously: a toccata in the first part of Il Transilvano (1593), two small ricercars in the second volume (1609), and a ricercar in the Turin manuscripts. The Vatican manuscript Chigi Q. VIII. 206 contains a fantasia on La Spagna headed “Lucciasco a 4,” followed by a fantasia on Ave maris stella “D’incerto” with the notation, “It was attached to the preceding one and therefore is believed to be by Luzzasco” (“Era attacato al precedente e pero si crede esser del Lucciasco,” Jeanneret 2009, 350). In the ricercar on the seventh tone with three subjects [fughe] of his 1615 collection Giovan Maria Trabaci inserted a hand pointing to the second subject with the words, “Luzzaschi uses this at the beginning of his tone 7, Ricercari bk. 3” (“Luzas[chi] vsa q[ue]sto in principio del suo 7. tono, Ricercat. lib. 3.”), indicating the existence of a lost third collection of ricercari by Luzzaschi. (An anonymous volume of paired versets on the eight tones has also been attributed to Luzzaschi [Intavolatura d’organo facilissima, Venice: Vincenti, 1598].) With the discovery of a manuscript copy of Luzzaschi’s Il secondo libro de ricercari a quattro voci (Venice: Gardano, 1578, previously known only from the Gardano catalogue of 1591), containing twelve ricercari, this corpus has been quadrupled.
The ricercari are presented in partitura as opposed to intavolatura, the notation employed for the toccatas of the collection.
Luzzaschi’s two ricercars in the second volume of Il Transilvano inaugurate a series on the twelve modes by a variety of composers and follow the pattern of the other ricercar pairs MESS as stipulated by Diruta. They are complementary in length (twenty-four breves each), in the relation of their subjects—the subject of the first is inverted to produce that of the second—and in their common finalis, d. The higher range and clef system of the first ricercar (chiavi naturali) define the authentic Dorian mode, Primo Tuono; the lower ambitus and clef-arrangement (chiavi basse, trasposte) of the second, the plagal Dorian, Secondo Tuono. Following Diruta’s teaching, Luzzachi’s subjects span an octave, a fifth and a fourth above the final for the authentic first mode, a fifth above and a fourth below for the plagal second mode. In part owing to the external restriction of each of the twelve ricercars in the volume to a single page, both of Luzzaschi’s contributions have a strong sense of harmonic architecture and motivic unity. The rhythmic writing, although devoid of ornamentation, avoids monotony through syncopation and the use of a fluid texture.
Luzzaschi’s recovered second book of ricercars of 1578 is ordered in pairs. The finales are C, C; d, g (with B-flat); e, a (B-flat); F, F ; G, G; a, a. Both chiavi naturali (= Diruta’s) and chiavi alte are employed, but they do not alternate regularly (high clefs appear in ricercari II, VII, IX, and XI). In length, the works average one hundred breves or slightly more.
Luzzaschi’s Ricercar I, of 1578, copied exactly in the Foà collection in Turin except for an added cadential trill, displays some of the same characteristics as the miniature pair in Il Transilvano, notably construction on two subjects related by inversion. The closing stretto presents the last motive both upright and in inversion. The overlapping sections of this large work are defined by permutations of this opening material which reveal that for Luzzaschi the “subject” is less an exact intervallic sequence than a succession of events: a step in one direction followed by a series of steps in the opposite direction. The richness and concentration of the miniature ricercars here expand into repetitive melodic lines and somewhat monotonous pacing. Some of the keyboard writing is awkward, but at least one passage foreshadows Frescobaldi’s creation of idiomatic keyboard sound through contrapuntal textures:
EXAMPLE 12.1: Luzzaschi, Ricercar I, mm. 85-89
The following ricercari of Luzzaschi’s second book display other devices taken over by Frescobaldi. Ricercar II has a tripla section, the only one in the entire collection, and transforms the opening leap of a fourth into a diatonic tetrachord motto C D E F or G A B C stated as a kind of cantus firmus in breves. Ricercar III inverts its theme by omitting the first note, and the theme of IV is an inversion of that of III. In Ricercar V Luzzaschi creates a three-beat line against a duple filigree in the other parts (T: mm 46-51; B: 53-58; S: 59-64; A: 66-70). The sonority of Ricercar VI is marked by third-derived materials and motion in parallel tenths. The exposition of Ricercar VII alternates two themes, of which the second is the decorated inversion of the first. The opening motif of VIII repeats the strong fourth that opens VII, and the work presents a particularly decorated figural surface. The exposition of Ricercar IX again presents two subjects as does that of X, which contrasts a repeated-note theme with the leap of a fourth which opened ricercar IX and develops a long-note descending fifth ostinato that moves through all the voices (mm. 77-100). The opening of XI presents a similar pairwise statement. The final ricercar, the longest of the collection, evolves its opening theme into another long-note motif, presented in stretto to close the series.
The surviving work of the other notable Ferrarese keyboard composer, Ercole Pasquini, is of limited value for stylistic comparison because Pasquini left Ferrara for Rome in 1597, and his keyboard music is transmitted only in manuscripts—often unreliable—that appear to date from well after his death.
Trying to determine the sources, or at least the antecedents, of Frescobaldi’s earliest works, in particular the Fantasie, gives the sensation of looking through a kaleidoscope. One scholar presents what seems to be a coherent arrangement of the material. A twist of the lens, and another scholar composes the same tesserae into a different and momentarily equally convincing pattern.
The so-called Bourdeney ricercars provide a classic illustration of this. Four fantasias in fascicle 41 of BAV, Chigi Q. VIII. 206 are headed “Fantasie di Giaches.” They have four anonymous concordances in the Bourdeney manuscript, Paris Bibliothèque Nationale 851, formerly the property of the musicologist Clarisse Bourdeney; this source also contains ten other ricercars attributable to “Giaches,” plus two additional ones. Oscar Mischiati (19751) inventoried the Bourdeney manuscript (cataloging it by M numbers) and placed its origin in the Po valley. James Ladewig (1978) accepted Carol McClintock’s attribution of the Chigi fantasias to Giaches de Wert (1535-96), thus also identifying their four anonymous concordances in Bourdeney. Newcomb 1987 reassigned the pieces to Jacques Brumel (d. 1594) and dated them around 1560, while Fabris 2013 suggested that they might even be the work of “Fabritio Denticj” (ca. 1539-ca. 1581), to whom the ricercar M302 is attributed in Bourdeney.
Newcomb describes the Bourdeney ricercars as deriving their entire texture from a small set of materials with a greater degree of economy, in which new subjects are generated by evolving variation, than any other collection of the century. These materials, which retain a consistent hexachord shape, are constantly manipulated by devices such as diminution, augmentation, inversion (hexachordal and non-hexachordal), including the creation of countersubjects by inversion and inganno (see below). Tripla sections occur in six of the fourteen ricercari. The ricercars are ordered by modes: I: finales d, d; II: g, g; III: e, e; IV: a; V: F; IX: d, a; XII: C, F with Bb. Newcomb sees the line of development as running from the Bourdeney (Ferrara- Mantua) ricercars through the manuscript ricercars of Giovanni de Macque, thence to Macque’s Neapolitan pupils Trabaci, Mayone, Rinaldo, Filomarino, and Stella, ending with Frescobaldi.
The ricercari of Giovanni [Jean] de Macque (c. 1548-1614), an attractive and important set of twelve, are preserved in the manuscript Florence, Biblioteca Nazionale Magl. XIX. 106 bis, dated late 16th-early 17th century. Macque moved to Naples from Rome in 1585. There he worked for the Gesualdo household c. 1585-90, became organist of the Annunciata (1590) and to the Spanish viceroy (1594), and maestro of the Neapolitan royal chapel (1599). In 1597 he dedicated his third book of madrigals à 5 (Ferrara: Baldini) to Alfonso II d’Este. Macque’s ricercari are laid out in four-staff partitura (C1, C3, C4, F4 or G2, C2, C3, F3), the last four pieces in campo aperto. They are identified only by modal number (“Primo [tono],” etc.) and their finales follow the system of Zarlino, whose work Macque knew, with transpositions by the use of B-flat for modes II (g), IV (a), XI (F), and XII (F). Each Macque ricercar employs three subjects (four in VIII), some of them quoted from Zarlino 1558, introduced in the first seven bars.
Andrea Gabrieli (1533-85), in his two posthumous volumes of keyboard ricercars (Venice: Gardano, 1595, 1596), is credited with raising still higher the level of an already serious genre through the systematic employment of learned devices such as inversion, augmentation, and diminution. Both collections are arranged modally. Andrea’s 1595 collection is a set of thirteen works (see Chapter 11, Organ Registration), filled out with two pieces by his nephew Giovanni; Andrea’s 1596 volume contains a group of six ricercars on tones I, II, V, and IX. His ricercars are written on single subject, double subjects derived by inversion, other subject-countersubject pairs, and from two to five subjects treated successively. Except for two concluding triple sections in the first and fifth ricercars of the 1595 book, cadences are generally elided, and rhythmic motion is continuous. Augmentation and diminution appear within the prevailing texture rather than defining discrete sections. The degree of surface ornamentation varies widely: some ricercars are completely unadorned, others decorate cadences by ornamental flourishes, and some are dominated by decorative figurations. Occasionally the composer is seduced by the open-score format into voice-crossings that are more apparent to the eye than to the ear. Gabrieli 1605 VI: 7 named chansons, 1 madrigal, ricercar homophony + decoration; à 4, 2-staff keyboard intavolatura.
The ricercars of the Neapolitans Ascanio Mayone and Giovanni Maria Trabaci show the influence of their teacher, Giovanni de Macque. They have in common with Frescobaldi 1) the use of three three to four subjects 2) a fondness for the inversion ricercar 3) strict contrapuntal textures 4) free rhythmic and melodic variation of subjects (also in canzoni) 5) use of the inganno (Ladewig 1978, 92: see below).
The four ricercars in Mayone’s 1603 Capricci are based on multiple (double, in one case triple) subjects, of which the first tends to predominate. Virtually all of the material in the working out of each ricercar is derived from the initial motives. The texture is without passaggi (perhaps with the impromptu addition of modest cadential ornaments). Each ricercar has a kind of concluding stretto in all four voices.
The twelve ricercars of Giovanni Maria Trabaci’s 1603 Ricercate follow those of Macque in their arrangement by modal finales (d, e, F, G, a, C, with some transpositions). Trabaci, who was not averse to advertising his own ingenuity, indicated in the title of each ricercar the number of subjects (fughe), from one to four. The subjects are introduced together at the opening of each ricercar; they are developed throughout and are usually presented in stretto at the end. In construction the subjects frequently show some reciprocal relation (a rising figure may be balanced by a descending one outlining the same interval, for example); literal inversions (riversi) also appear. Ten of the ricercars are continuous, with no full internal cadences. The two ricercars that show clear structural divisions (VIII, on the Ruggiero, and X) add a contrasting triple section with a duple cadence; both are longer than the single-section ricercars. Ornamental figuration occurs only in these two ricercars and in Ricercar V.
Trabaci calls the reader’s attention to two devices that were to become hallmarks of Frescobaldi’s style. The fourth ricercar is headed “con tre fughe, & inganni” (“with three subjects, and deceptions”). According to Artusi, an inganno occurs “every time, that one part beginning a subject the answer follows it not by the same intervals; but by the same [solmisation] syllable-names.” In other words, new motivic materials can be derived by substituting for a note represented by a hexachord syllable the same syllable in another hexachord. Ut Re Mi = C D E in the natural hexachord, F G A in the soft, and G A B in the hard. Thus Ut could be realized as three possible pitches: C, F, or G. Trabaci’s other musical device is the obligo, a “certain [external] set condition” or subject imposed upon a contrapuntal composition. This is evident in the title of the fifth ricercar, with four subjects and “notes that pass for false” (“note che passano per false”—perhaps owing to the ambiguity B-natural/B-flat in the F mode).
In his preface to the Partito delle Canzoni (Milan: Erede di Simon Tini e Lomazzo: Sartori 1605a, published in score (libro aperto) as well as CATB partbooks), Gio. Dominico Rognoni Taegio sounded a note later echoed by Frescobaldi: “where they [the students of this profession] become practiced in scoring, and draw from it much profit … since I know that in every case the score is better than the Basso continuo.”
The Ricercari (Sartori 1606a) of Giovanni Paolo Cima (ca. 1570-1630: organist of S. Maria presso S. Celso in Milan 1595, acting maestro, 1607-11) were published two years before Frescobaldi’s Fantasie by the same Milanese firm, l’herede di Simon Tini, & Filippo Lomazzo, and in the same libro aperto format. There is some evidence that Frescobaldi knew Cima’s work since the final phrase of Cima’s Enigma musicale, “one cannot reach here without effort” (“qui non si può arrivare senza fatica”), recurs at the end of Frescobaldi’s Toccata IX in the 1627 set as “non senza fatiga si giunge al fine.” Camillo Angleria’s Regola del contraponto (Milan: Rolla, 1622), dedicated to Cima, contains a canon à 4 by the latter with the Petrarchan superscription, “Understand me who can, for I understand myself” (“Intendami chi puo, che m’intend’io”), later quoted by Frescobaldi in the Fiori musicali of 1635.
Cima’s seven ricercars do not display any obvious modal arrangement. The subject that opens his sixth and seventh ricercars appears also in Frescobaldi’s Fantasia XI, and Cima’s treatment of it as a tenor ostinato in a wide variety of note-values (Ricercar VII) anticipates Frescobaldi’s use of ostinato oblighi in the 1615 Recercari et canzoni.
The ricercars of Costanzo Antegnati’s (1549-1624) L’Antegnata (Venice: Gardano, 1608) are arranged in the traditional twelve modal finales. Printed from movable type in keyboard intavolatura rather than partitura, they display a more generally figural texture, with decorative passaggi at cadences and penultimate tripla sections in nos. VIII and XI.
Therefore, at the point where Frescobaldi entered its evolution, the ricercar presented a wide spectrum of possibilities as one of the normative genres of serious contrapuntal keyboard composition. Subjects could vary in number from one to four introduced simultaneously and up to as many as eight in succession. They could be either distinct in character or related in a variety of ways, most obviously by inversion. In character they might range from the classically profiled motto to the ornamental figure. The possible structural organizations of ricercari ranged from monosectional works with no full internal cadences to constructions whose multiple sections were further set off by cadential ornamentation and shifts from duple to triple meter. Length could vary from the twenty-odd-breve miniatures of the Transilvano pieces to the monstruous ricercars of Jacques Buus, over eleven times the length of the Transilvano pieces. Despite the basically contrapuntal conception of the genre ornamental figuration was not excluded, and the Turin redactions suggest that more may have been added in performance. All of the modes were employed, and modal organization by the systems of eight or twelve modes was frequent. Some composers eschewed the more arcane contrapuntal devices such as augmentation, diminution, inversion, inganno; others employed them as integral elements in motivic development; and a few set them up as self-imposed hurdles for the greater display of compositional virtuosity.
Frescobaldi’s Fantasie of 1608
The designation “fantasia” occurs sporadically in Italian keyboard sources, attached to everything from cantus firmus settings to canzoni alla francese. On the evidence of the twelve fantasias that Frescobaldi published in 1608, he regarded the genre as a contrapuntal study in the tradition of the ricercar. Within this tradition Frescobaldi’s fantasias conform most closely to Luzzaschi’s ricercars. Like the ricercars of Luzzaschi, Il Transilvano, Cima, and the Neapolitans, the Fantasie were printed in keyboard partitura and libro aperto format. For the possible influence of Frescobaldi’s Milanese printers on this arrangement see Newcomb 1965, 117-18.) Like the ricercars of Merulo, Macque, Andrea Gabrieli, Trabaci, and Antegnati, the Fantasie are arranged by modes and within modes by plagal/authentic pairs: g (d transposed with Bb), g; e, a (e transposed with Bb); F, F; G, G; a, a; F (C transposed with Bb), F. High clefs are employed in Fantasias I, V, VII, IX, XI. Triple sections appear on Fantasias I, III, V, and XII. Fantasies II, IV, VI, VIII, and XI have a concluding section in longer note-values, as in the concluding “drag” of an English viol fantasia.
Frescobaldi’s fantasias resemble Trabaci’s ricercars in being based on varying numbers of simultaneous subjects indicated in their titles, but here the arrangement is systematic: a set of three fantasias each, on one, two, three, and four subjects. In seven of the twelve fantasias Frescobaldi includes at least one triple section, like Luzzaschi, Andrea Gabrieli, Trabaci, and Antegnati, creating a clear three- or five-part structure. The consistent employment of augmentation and diminution suggests the influence of Andrea Gabrieli’s ricercars, but unlike Gabrieli, Frescobaldi sometimes defines entire subsections by the employment of a single one of these devices rather than by a combination of them.
Frescobaldi’s Fantasie are clearly sectional, defined by cadences (sometimes lightly overlapped) and different levels of rhythmic activity in contrast to Luzzaschi’s ricercars, where cadences are elided even at significant points such as changes between duple and triple meter. All of the Fantasie open in relatively long note-values (whole, half, quarter); a succeeding section presents the thematic material in somewhat frantic diminution as eighth-note figures. Augmentation of the subject as a kind of cantus firmus appearing in three or four voices (cf. Bourdeney quinto tono, mm. 72-88, SAB) occurs in Fantasies III (ABST), V (ABTS, where it lasts for 67 measures, mm. 30-97), V (ATTBS, mm. 30-92), VI (ABS, mm. 77-102), X (canon partly by inversion, SB, ST, mm. 28ff.), and XI (ABST, mm. 52-99). (These statements in augmentation suggest the sustained sound of the organ as the appropriate performing medium.) An occasional stretch in the left hand of a tenth to a low E implies a C/E short octave; in Fantasia VII the chromatic repertory comprises C, F, G, and D sharps.
In details of structure, motivic treatment, and sectional relationships each fantasia is a unique solution. What they most clearly share is the rigor with which Frescobaldi derives each work from its initial subject at a level of concentration that even he never again attempted. The subjects that he invented are particularly suited to such treatment. Like those of many of his predecessors, they usually begin by outlining some strong interval—third, fourth, fifth—in long notes which then dissolve rhythmically into scalar figures. (In the sixth fantasia the same effect is achieved by adding a syncopated ostinato countersubject to the initial motto, a treatment prefigured in Luzzaschi, Ricercar VIII). Since Frescobaldi follows Trabaci in presenting the subjects of a multiple fantasia simultaneously, complex interrelations increase with the number of subjects. In the fantasias on two subjects the most common relation is inversion, either literal, as in the fifth fantasia, or allusive, as in the fourth. (For detailed analyses of each fantasia including motivic quotations between pieces see Bellasich 1995, viii-xvi.)
In the fantasias on three subjects one or two of the motives are strongly intervallic; the remainder are less striking owing either to scalewise construction or to a less emphatic presentation. Frequently the subjects are related by inversion or by a stress on the same interval, as in Fantasia VIII, where all three subjects outline tetrachords, subject 3 is the inversion of 1, and 2 is also built on the intervals of the fourth and whole tone. In Fantasia X the four subjects are in fact two pairs in literal inversion.
Like Luzzaschi, Frescobaldi regarded his subjects as a series of musical gestures rather than as a fixed interval series, and in some cases a subject may develop throughout a piece to emerge transformed at the end (see Fantasia VIII and the analysis of Fantasia II below).
The most obvious agents of such thematic manipulation are augmentation, diminution, and the inganno. The effect of regular augmentation is most apparent where repeated statements of an augmented subject are employed to produce a cantus firmus or ostinato (Fantasia V, mm. 72-97). Diminution is commonly employed in the manner of Luzzaschi, to generate figural material from the original subjects (Fantasia IV, mm. 25-26). Although either technique can set the general rhythmic pace of a section, both often occur together as a means of creating a complex texture saturated with motivic material.
The inganno assumes an underlying identity of hexachord syllables despite a changed intervallic sequence. In some cases it not only creates new figural patterns but even challenges the authority of the original motif (Fantasia X, mm. 43-48: subject 3 as an inganno in augmentation).
Chromatic inflection of basic diatonic materials also plays an important role in Frescobaldi’s motivic development, sometimes with far-reaching results. In Fantasia VII an F is sharped at m. 26 in a statement of the first subject, answered by an inflection of C in subject 3 in bass and soprano (mm. 29, 31); from then on, F and C are regularly sharped. The inflection moves up a fifth in m. 35, where G is sharped in an inversion of 1, and by m. 40 the same subject (now on B) includes G-, C-, F-, and D-sharps. In the Fantasie tonal shapes are generally less important than linear counterpoint (some fantasias have no cadences on anything but the tonic), but here the harmony suggests D, A, and E major (to employ a useful if anachronistic terminology), and cadences on D.
Most of the fantasias, even those with no decisive internal cadences, have sections defined by different levels of rhythmic subdivision and consequent varying density of texture. (The exception is Fantasia IX, which in its poorly differentiated subjects, monotonous pacing, and occasional miscalculations of sonority is the weakest member of the set.) Frequently some detail in the preceding cadence prepares a new rate of movement, and sometimes it is reached by a rhythmic modulation (Fantasia XII, mm. 80-82).
Frescobaldi’s techniques of thematic transformation and rhythmic variation are strikingly represented in the second fantasia, “sopra un soggetto solo.” Structurally, the work is laid out in six sections, two of them in triple meter. Section 1 is essentially an exposition in which the subject preserves its original rhythmic outline, although even here diminished versions of the subject appear as countersubjects, and the second half of the subject is stated in augmentation (Fantasia II, mm. 20-22). After the cadence in m. 23 the subject is stated in jagged syncopations, and from then on its characteristic descending fourths and fifths permeate the texture at every rhythmic level from eighths to whole-notes. The short triple section (in a rhythmically ambiguous pattern, like many triplas in the Fantasie) employs both the descending fourth of the original subject and a variant derived by an inganno:
Ex. 12.1: Fantasia II, m. 1: natural hexachord: C: sol fa mi re mi fa mi. mm. 33-34: natural hexachord: C: sol fa soft hexachord: F: mi re mi fa
The texture of the next section is derived from the diminution of the subject, which also appears in its original values and in an augmentation that continues to clarify the basic descending tetrachord. The tetrachord also underlies the following section (mm. 62-64), which is reached by a rhythmic modulation. The tetrachord becomes almost an ostinato in section five (mm. 70-76), whose shimmering overlapping figuration prepares the last section: here the subject is presented in a magisterial rhythm and in its most compelling melodic form, a descending chromatic tetrachord.
It is easy to give the misleading impression that the fantasias are nothing but intellectual exercises, and certainly they are works of formidable concentration. Further, each piece is conceived essentially within the relatively limited range of three octaves around its finalis, which means that there are passages in which the clarity of the partitura format betrays the composer into voice-crossings intelligible to the eye but not to the ear (e.g., Fantasia V, mm. 43-45). Nonetheless, there are also moments in which what appears on the page to be independent part-writing dissolves in performance into an almost impressionistic keyboard sonority:
Ex. 12.2: Fantasia II, mm. 70-73
When dealing with a group of superficially similar works in an unfamiliar style, it is easier to give the impression of uniformity than to communicate their individuality. An examination of the Fantasie shows that within the boundaries of the genre Frescobaldi chose for his first keyboard publication he achieved a considerable variety of formal structure and proportion, motivic construction and exploitation, rhythmic manipulation and texture. Despite their occasional technical flaws (cf. the parallel octaves in Fantasia XII, m. 80), the fantasias display many of their young composer’s gifts: a remarkable grasp of the contrapuntal potentialities of a subject, and both logic and imagination in their development; an emerging sense of form and structural relationships, climax and de-emphasis; and an extraordinary command of keyboard sonority within a theoretically abstract polyphonic medium.
Unlike all of Frescobaldi’s other collections, the Fantasie were not reprinted in his lifetime except for excerpts in a theoretical work. The Architectonice musices universalis of Wolfgang Schönsleder (Ingolstadt, 1631) contains fifteen musical examples ascribed to Frescobaldi, which have been identified as excerpts from fantasies I, II, III, VI, VIII, IX, and X and are thus early disseminations of Frescobaldi’s music outside Italy. There is a strong stream of manuscript transmission, beginning in the Kraków manuscript before 1625 and including a complete scribal copy made ca. 1690-1710 with a later inscription that attributes it (falsely) to Bernardo Pasquini, as well as other copies in a Paris manuscript (see Catalogue I.1). This manuscript tradition is not so much a “second best” Samizdat version as an alternate method of circulation practiced in the late Renaissance, sometimes for particularly hermetic texts (see Richardson 2009).
The Recercari et Canzoni of 1615
The full title of Frescobaldi’s next keyboard publication in strict style is Recercari, et canzoni franzese fatte sopra diversi oblighi in partitura (Rome: Zannetti, 1615), dedicated to his new patron, Cardinal Pietro Aldobrandini, Chamberlain of the Church. The collection contains ten ricercars dominated by pre-existing compositional requirements—the “diversi oblighi”—and five canzonas. Like the Fantasie, the Recercari, et canzoni is arranged modally in pairs by finales: d=g Dorian, d=g Dorian [plagal]; e, a with Bb; F, F; G, G; a, a. It is printed in keyboard partitura with chiavi alte for the higher ambitus and chiavi naturali for the lower range.
The oblighi which govern all but two of the ricercars are of varying types and show the influence of a wide range of models. Recercars I and IX are closest to the southern Italian tradition, most recently expressed in Ascanio Mayone’s second book of Diversi Capricci (1609). They present a cluster of subjects (three and four respectively) at the beginning, develop them jointly in the course of the work, and combine them in stretto at the end. Recercars IV, VI, VII, and X are built around solmization-ostinati, perhaps in emulation of the first known ostinato ricercar, the seventh ricercar of Cima’s 1606 volume. Recercars II and VIII have oblighi of external conditions: II follows a preexisting model, Luzzaschi’s Ricercar I of 1578; VIII has the obligo of avoiding stepwise motion. Like Recercar II, Recercari IV, V, IX, and X follow the northern Italian tradition of sectional organization.
These varying oblighi are realized in a variety of formal and musical solutions. Each of the sectional ricercars, for example, creates and employs its material in different patterns. Recercar II follows its Luzzaschian model in defining its three sections by new material, each subject accompanied by a countersubject derived by strict inversion and resolved by cadences of increasing weight. Recercar III is additive in its treatment of its three successive subjects (A, BA, CBA). Recercar V presents its three subjects in a long opening duet, treats each of them separately in four voices, and combines them in the last section.
As one might expect from the appearance of four solmization subjects, ostinato techniques play a large part in the Recercari. The subject of the fourth ricercar, Mi Re Fa Mi, appears in all four voices on A (soft hexachord) or E (natural). (Frescobaldi never employs more than two of the three traditional hexachords in treating a single ostinato.) Its first set of appearances presents the subject in semibreves, its second in breves, and its third in double breves, each in all four voices sequentially. Each section has a different countersubject, the second and third being freely treated in inversion as well. In the seventh ricercar, on Sol Mi Fa La Sol, the subject appears as an ostinato in the tenor in the hard hexachord on D and the natural hexachord on G in semibreves and breves (or double semibreves tied across the bar) and double breves. Unlike Recercar IV the texture is continuous and is dominated throughout by a figure of three repeated notes (cf. Fantasia VI). The ostinato La Fa Sol La Re of the tenth ricercar appears in the other voices at the beginning of the works and constitutes the entire contents of the soprano part in a variety of rhythmic patterns. (In the original partitura format the ostinato treatments leap to the eye.)
The sixth ricercar is a delightful study on the ostinato Fa Fa Sol La Fa. The subject has been traced back to the Dodekachordon of Glareanus (1547) and recurs in the Bourdeney ricercars and elsewhere, but it is also identical with the tune “Fra Jacopino went off to Rome, a staff on his shoulder and a pilgrim’s gown on his back” later employed by Frescobaldi as an ostinato in a set of variations on the Ruggiero. The tune is presented throughout in the alto on F (natural hexachord) or C (hard). It is first announced in the bass in its original rhythmic form, but its succeeding entries in the alto are in notes of equal length. The first four statements of the ostinato occur at regular intervals (mm. 4, 8, 12, 16) in even whole notes. At mm. 20 and 26 it appears in breves. In the first statement each breve corresponds to a measure, but the second foreshadows the approaching rhythmic delirium by splitting the breves across the bar. In m. 39 the subject appears in dotted halves against the undotted halves and quarters of the surrounding material, as in Luzzaschi’s Ricercar V, while in m. 45 it is stated in syncopation against the other parts. In m. 49 there begins a regular augmentation, first in dotted whole notes (always against a one-breve bar), then in dotted breves (m. 54) and finally in double breves (m. 65). As elsewhere in the Recercari, Frescobaldi not only develops the ostinato material in the other voices but also invents new countersubjects.
The obligo of the eighth ricercar, “Obligo of never resolving by step” (“Obligo di non uscir mai di grado”) represents the first appearance in Frescobaldi’s published work of the external limitations that so stimulated his imagination. The subject matter of this ricercar—thirds, fourths, fifths, and octaves—does not encourage the definition of sections by new material. Instead, Frescobaldi employs the principles of rhythmic variation from the ostinato ricercars to produce a study in changing rhythmic textures whose increasing offbeats and syncopations provide the rhythmic equivalent of a contrapuntal stretto. (This is the only ricercar of the set that was copied into the Turin manuscripts without any added cadential trills.)
By comparison with the fantasias, the ricercars show a less obsessive concentration on deriving all the motivic material of a work from a single nexus of ideas. Although most of them are organized into sections, these are now more subtly defined than in the fantasias. Unlike the wide range of rhythmic levels characteristic of the fantasias, the ricercars move in a steady minim pulse, and such obvious devices for differentiating sections as triple passages and florid decoration are totally absent. Variety is achieved more unobtrusively through manipulations of texture and the placing and relative emphasis of cadences. This is reflected in an enhanced freedom of keyboard writing; although the range of the ricercars does not exceed that of the fantasias, their part-writing generally avoids the occasional thickness and obscure voice-crossings of the earlier collection. In its imposition of oblighi and the variety of their treatment, the Recercari shows the first overt signs of two characteristics that animate much of Frescobaldi’s subsequent work: his predilection for self-imposed technical challenges and his musical wit in their solution.
The north Italian keyboard canzona dates back to Marc’Antonio Cavazzoni’s intabulations of four French chansons in his publication of 1523. Two decades later, his son Girolamo moved the genre a step further toward independence from vocal models by writing canzonas as paraphrases or parodies rather than simple diminutions of vocal originals. The distinction was continued by Andrea Gabrieli, who titled his intabulations “canzoni” and his paraphrases “ricercari.” The keyboard canzona retained close links with the instrumental canzona, so that four of Merulo’s keyboard canzonas of 1592, shorn of their figuration, were printed as ensemble canzonas in Raverij’s 1608 collection, which also contained three instrumental canzoni by Frescobaldi. In L’Organo suonarino (1603) Adriano Banchieri presented “Fantasies or canzoni alla francese to play on the organ and other musical instruments” (“Fantasie overo canzoni alla francese per suonare nell’organo et altri stromenti musicali”). In his Moderna armonia di canzoni alla francese (Venice; Amadino, 1612) Banchieri envisioned not only solo keyboard but also ensemble performance.
Vincenzo Pellegrini’s Canzoni of 1599 marks the full emergence of the independent keyboard canzona, which in the Venetian tradition is characterized by contrasts of theme and texture between clearly defined sections, often emphasized by the repetition of a section. (In Marc’Antonio Cavazzoni’s canzonas the patterns ABA, AAB, and AABA occur; G. P. Cima usually repeats the closing section, ABB.) North Italian composers such as Pellegrini and Cima reinforced these contrasts by inserting one or occasionally two sections of triple meter into the prevailing duple. Ornamental figuration appears, especially at cadences, but the conventional dactyllic themes and generally rather abstract part-writing still betray the influence of the instrumental canzona. Even in the case of canzonas derived from preexisting chansons, northern composers often added allusive family names as canzona titles.
The keyboard canzonas of the Neapolitans Trabaci and Mayone present a much livelier picture. Their figuration—less ornamental than functional—is exuberant, sometimes bizarre, and occasionally unplayable. Triple sections are a regular feature of these canzonas, and several Neapolitan works even open with chordal passages in triple meter. The subjects of the three or five sections of a typical Neapolitan canzona are generally derived from the opening material, a notable instance of the variation canzona procedure.
James Ladewig has employed the variation canzona, one of the most innovative and consistent of Frescobaldi’s compositional choices, to trace influences on his development (Ladewig 1978). The formal procedure in which the subjects of the various clearly defined sections of a canzona are all derived from the same material appears in works by Ferrarese composers, “Giaches” (whoever he may have been), Ercole Pasquini (six of his nine surviving canzonas, including one preserved in no less than six manuscripts), and in an ensemble canzona by Luzzaschi. The appearance of the variation canzona in the works of the Neapolitans Trabaci and Mayone may derive from the work of their mentor, Giovanni de Macque, or the form could have been encountered by the musicians of Carlo Gesualdo’s entourage during their stays in Ferrara. The subjects themselves are more varied than their northern counterparts. Two of the canzonas from Trabaci’s 1603 collection state their initial subject chordally, and another subject is chromatic in outline. Unlike some northerners, Neapolitans did not give titles to their canzonas.
Frescobaldi’s first published canzonas, the three ensemble works in the Raverij collection, are clearly derived from the north Italian tradition. The first canzona, for four voices, is continuous in texture except for a coda and derives its material from the first subject and its countersubject. The second, in five parts, begins with a standard dactyllic theme and countersubject, subsequently simplified to an ascending and a descending scale, from which the remaining material is derived. (Luzzaschi’s posthumous four-part canzona in the same collection employs similar material.) The continuous pulse and overlapped sections of the work are balanced by a variety of rhythmic surfaces. The third canzona is even more clearly Venetian, since it is scored for double choir and is marked by fanfare motives, alternation of the two choirs, and a sectional construction emphasized by a repeat of the second half of the work.
The five keyboard canzonas with which Frescobaldi concluded the Recercari et canzoni show little trace of Neapolitan influence. According to Ladewig, Frescobaldi’s great innovation in his canzonas of 1615 was to purge non-contrapuntal effects such as pseudo-polyphony, animated homophony, and toccata-like figuration over a chordal accompaniment, thus achieving a purified linear texture throughout (Ladewig 1978, 55). The canzonas are arranged modally: g Dorian (“primo tono ‘); g Dorian (“primo tono”); g Dorian (“secondo tono”); F (“sesto tono”); a (“nono tono”). They all begin with imitative figures marked by the repeated notes in dactyllic patterns typical of the north Italian canzona. Each canzona is clearly divided into sections cadencing on the tonic or some other prominent degree of the mode (B-flat in G Dorian, for example), and their sectional organization is further emphasized by the inclusion of one to three passages in triple meter in each canzona.
The triple sections are continuous, but the duple ones—especially the opening ones—may be subdivided by cadences and the re-presentation of the original material, the introduction of new material, or both. In all except the first canzona, material from the opening section is brought back in the concluding one, as in ensemble canzonas. Frescobaldi’s treatment of thematic relations between these sections varies from new and unrelated material for each section (Canzona I) through the return of the original material or its derivatives in a concluding section (Canzonas IV, V) to works in which all the major sections are based on the same material (the variation-canzonas II and III).
Although not a variation-canzona, Canzona IV is the largest and in some ways the most stylistically advanced of the set. Its opening section has three subdivisions: the first (mm.1-12) presents the first subject against a running countersubject; the second (mm. 13-16) treats the second half of the subject as an independent figure; and the third (mm. 17-25) develops the two motives together. The short chordal triple section that follows concludes with a toccata-like flourish, a feature absent from the rather square rhythmic outline of the other canzonas. A second, more contrapuntal triple section (mm. 30-36) introduces another duple passage presenting new material and cadencing into a third triple section (mm. 45-53) , again concluded by a free passaggio. The final duple section returns (mm. 55-end) to the opening subject, accompanied by a figure derived from the initial subject and developed in the second area of the opening section.
The canzonas of the 1615 collection form a pendant to the ricercars, juxtaposing sectional works in a texture marked by consistent and regular imitation (insofar as those two adjectives can be applied to any production of Frescobaldi) with the highly individual formal and textural solutions of the ricercars. In both sets of works Girolamo carries the tradition of his musical training to a logical conclusion in his mastery of formal and thematic treatment within the context of strict part-writing.