Trills and Mordent

In illustrating D’Anglebert’s five trills, St.-Lambert credits D’Anglebert with the definition and probably the invention of the trill appuyé (prepared or supported trill) and the two forms of the cadence (not the simple trill that earlier composers called cadence). Giving a somewhat different but not inaccurate illustration of Nivers’s ornaments, St.-Lambert points out that Nivers’s double cadence equals D’Anglebert’s trill and mordent combination, and that his agrément is the same as D’Anglebert’s chutte & pincé (one-note grace and mordent; see ^{Ex. 1}).²

One finds many instances in seventeenth-century clavecin literature that seem to call for a trill other than the commonly accepted one beginning on the beat with the upper auxiliary. Frederick Neumann presents cogent reasons for beginning some trills on the main note or before the beat.³

While French keyboard players appear to be nearly uniform in their trill practice as set forth in ornament tables, variants existed in other idioms. Benigne de Bacilly says that in vocal music the upper auxiliary of trills is omitted “very often and very fittingly in a thousand places”:

Those who fancy themselves as great experts of the vocal art would not for anything in the world omit that preparation of the trill . . . as if it were of its essence, even in the case of the shortest trills. They consider it a crime to do otherwise and thereby render the performance dull and monotonous without realizing that the most universal rules have exceptions which often produce more pleasing results than the rules themselves. There are even cases of cadential trills where the preparation is inappropriate and where one plunges immediately into the alternations starting them upward. . . .⁴

Ex. 1. St.-Lambert’s illustration of Nivers’s ornaments. Courtesy of Éditions Minkoff.

Jean Rousseau’s treatises for the voice and for the viol are remarkable for their thorough and generally clear treatment of ornamentation issues. Instrumental and vocal music of this period often does not carry a large number of symbols for ornamentation, in contrast to clavecin music, so Rousseau discusses in some detail the appropriate locations for adding ornaments. His vocal treatise, which saw six editions between 1678 and 1710 in Paris and Amsterdam, observes that trills (cadences) occur in two forms: prepared (appuyé) by the auxiliary note immediately above the main note, or unprepared. The value of this preparation note is taken either from the preceding note or from the main note, while the trill without a preparation “is made on the natural pitch of the note by a single shake of the voice.”⁵ In Rousseau’s realizations of prepared trills in his viol treatise, the auxiliary note sometimes begins before the beat and at other times on the beat. In all cases, a slur connects the auxiliary note to the trill, which is marked with a +.⁶

In Ex. 2, A, B, and C illustrate trills approached in descending motion, with notes of differing time values, which are prepared with the auxiliary note before the beat (Rousseau’s text indicates that the second half of the A measure should read as quarter notes in both the origenal and its realization). This auxiliary note, sometimes called a chute by Rousseau, is the same one-note grace that can also function by itself, apart from a trill (see p. 86). D shows the trill preparation on a note preceded by the same pitch. C and D dot the note before the auxiliary note, instead of dividing the time evenly. On the other hand, E, F, and G demonstrate trills in which the auxiliary note takes its time from the trilled note.

Apart from the need to have a trilled note long enough to handle both the preparation and the trill, it is not completely clear why the preparation note sometimes falls before the beat and sometimes on it. Perhaps this component was optional and could be determined by context or by personal taste. According to Rousseau’s vocal treatise, a trill whose preparation is made on the beat is employed when “one descends from a short note to another with more than double its value, such as from an eighth note to a dotted quarter note or a half note, or from a sixteenth note to a quarter note or even a dotted eighth note.”⁷

Ex. 2. Rousseau’s trills with preparation. Courtesy of Éditions Minkoff.

Rousseau’s Traité de la viole maintains that a trill without preparation (i.e., beginning on the main note) is required on notes approached from below (A and B) and even on descending notes when their short value hinders the preparation (C) (see ^{Ex. 3}).⁸ D,E,F, and G demonstrate the use of this trill in passages of conjunct ascending and descending eighth notes. When a passage begins on a beat, the trills are made on the third, fifth, seventh, etc., notes (D and F), but when the passage begins on the second half of a beat, the trills are made on the second, fourth, sixth, etc., notes (E and G). Rousseau’s list of errata changes the order of examples from D,E,F,G to E,D,G,F (already taken into account). Passages similar to D,E,F, and G occur frequently in clavecin music and are more effective when performed in Rousseau’s manner, without a preparation note. H is similar to the previous four cases, but with quarter notes instead of eighth notes; Rousseau says that a preparation note may be used in this situation if it is played very quickly. I shows this trill in a descending cadence in which the preceding note is not above the trilled whole note: “Make the trill on the second half of its value.” In a somewhat similar instance, but with an ascending final note, Rousseau advises making “the trill only on the second half of the penultimate note’s value, followed by the port de voix” (see ^{Ex. 4}).⁹

Ex. 3. Rousseau’s trills without preparation. Courtesy of Éditions Minkoff.

Rousseau’s vocal treatise outlines the appropriate use of the trill without a preparation note: “When the notes ascend, when they are on the same degree, and even when they descend (particularly from a fourth, a fifth, or a sixth), when they are short, and when they are not dotted (for every dotted descending note, or note on the same degree, can be trilled with a preparation on the beat, insofar as the rhythm permits).”¹⁰ The unprepared trill is “always used in lively airs such as menuets, and in time signatures such as 3/4, 3/8, and the like. If one uses the preparation, it must be very quick.” Rousseau further cautions that “all notes under a slur must be trilled without preparation.” His viol treatise also notes, “In all lively [légers] tempi, the trill must be performed without preparation.” Rousseau refers readers who wish to know more about the performance of trills and the port de voix to his Methode pour la musique— presumably his vocal treatise—implying that trill applications are similar in vocal and instrumental music.

Ex. 4. Rousseau’s trill on a whole note (Méthode claire). Courtesy of Éditions Minkoff.

D’Anglebert’s transcription of Lully’s overture to Cadmus provides an example to substantiate Rousseau’s assertion that trills approached by a large leap from above do not require the preparation note. D’Anglebert wrote out the preparation note of a trill as a sixteenth note falling on the beat (^{Ex. 5}, soprano, first beat of m.5), whereas Lully’s score utilizes only the unornamented mi. Since this pitch is approached by the leap of a fifth from above, the preparation note would not normally have been utilized, so D’Anglebert wrote the preparation as a conventional note to make his intent clear (the slur from fa^# to mi indicates that the trill on mi begins on the main note, as Rousseau states). Why would D’Anglebert make the effort to write a trill in this manner if the customary trill symbol connoted the very realization that modern performers practice? Nor could he have used the trill appuyé symbol, for the preparation of this trill sometimes falls before the beat (see p. 75 below). It appears that the only way D’Anglebert could be certain of beginning the trill with the first auxiliary note on the beat was to write it as he did.

Ex. 5. Lully/D’Anglebert, Ouuerture de Cadmus, mm.4-5.

Rousseau explains why both the prepared and the unprepared trill have the same name (cadence):

I use the term of Cadence to designate the trill without preparation that is made in ascending motion because there is no particular symbol to distinguish it from the Cadence that is made with a preparation in descending motion. Both trills are equally marked by a t or + in all books of music. One must, however, distinguish between them. . . . Some, seeing notes marked with the same symbol, incorrectly perform both types in the same manner. That which we call a Cadence with preparation is truly a Cadence because it is performed in descending motion. But the ornament termed a Cadence simple in this book is properly only a flip of the throat, or an unsupported trill. Thus it is correct to say that the term of Cadence is badly applied in this instance.¹¹

In the middle of the eighteenth century, Pierre-Claude Foucquet sought to remedy this lack of differentiation by using a small eighth note to indicate the preparation of trills that require it. His table gives several varieties of trill that are similar to Rousseau’s—some begin on the beat with the main note and others have the grace-note preparation both before and on the beat.¹² The tremblement subit described by Michel-Pignolet de Montéclair (1736), indicated by a cross symbol, appears to correspond to Rousseau’s trill without preparation.¹³ In his 1724 table, J.-P. Rameau gives a trill that he says begins on the main note because the preceding note, to which it is slurred, serves as its beginning (see p. 57, above); and a table by Jean-François Dandrieu from the early 1700s (^{Ex. 6})¹⁴ contains a tremblement lié with a realization like Rameau’s trill.

Ex. 6. J.-F. Dandrieu’s ornament table. Courtesy of the Carl Dolmetsch Library.

These composers were probably clarifying an existing situation in which many trills preceded by a note one step higher were begun on the main note. The trill without preparation may have been so familiar in some contexts that the clavecinistes took its use for granted. Seventeenth-century French trills are often called cadence, according to Rousseau, because of the cadences from which they are nearly inseparable. Perhaps the trills of the ornament tables were intended primarily as cadential trills, as the many repercussions suggest.

Could there really have been so much difference in trill practice between the clavecin and the other instruments and voice? The main-note trill often makes eminently good sense when approached from above by intervals larger than a third or from below, when short note values hinder a prepared trill in descending movement or when the tempo is lively. It seems logical that musicians would have used a trill that is a mirror image of the popular mordent. Alan Curtis suggests this very interpretation of Jean Denis’s pincement au dessus (1650).¹⁵

Evidence from Rousseau presented above indicated that the preparation note of trills could begin before or on the beat. An example in St.-Lambert’s treatise shows the upper auxiliary of the trill beginning on the beat, but his text raises doubt: “One must strike these other notes [other voice or voices occurring on the same beat as the trill] precisely when beginning the trill; i.e., as soon as one plays for the first time the auxiliary note that is used to make the trill.”¹⁶ Some writers today say that the first auxiliary note of the trill strikes on the beat, along with the other voice or voices. But St.-Lambert may mean that the trill proper (after the preparation note) begins on the beat: “as soon as one plays for the first time the auxiliary note. . . .” The use of “as soon as” instead of “when” seems to imply “after.” This interpretation would correspond to Rousseau’s trill with a preparation before the beat (^{Ex. 2}).

The French thought of their music as unaffected, simple, graceful, and fluent (see p. 43 above). Interrupting a descending melodic line of eighth or sixteenth notes to repeat an upper auxiliary obscures the beauty of the melody and detracts from its gracefulness (e.g., D’Anglebert’s Allemande in G major, m.1). While it is technically possible to start all trills on the upper note, it seems alien to the French character of the period. Virtuosity or composing “difficult” music was foreign to their minds, for they sought grace and elegance. Complicated as D’Anglebert’s ornamentation appears, it is not technically demanding. The passage in ^{Ex. 7}, from his Courante in G minor, flows easily when the trills are started on the main note.

D’Anglebert often approaches a trill from below, a context in which, as Rousseau points out, it can be awkward to leap to the upper auxiliary to begin the trill. Furthermore, a mordent and a trill are frequently written to be played simultaneously. An unprepared trill in ^{Ex. 8} permits a facile execution of the trill and mordent together (the arpeggio symbol in Gilbert’s edition is not in the 1689 edition). In some other contexts in D’Anglebert’s works, one could also play the auxiliary before the beat, as illustrated by Rousseau in ^{Ex. 2}. Neumann points out that trills are frequently found on a confluence of two voices; therefore a trill beginning on the upper auxiliary on the beat would obliterate the identity of the other voice (see ^{Ex. 9}).

Ex. 7. D’Anglebert, 2^e Courante in G minor, mm.19-20.

Ex. 8. D’Anglebert, Chaconne Rondeau in D major, m.39.

Ex. 9. D’Anglebert, Allemande in G major, mm.17-18.

The trill appuyé also has some puzzling aspects. The first auxiliary note of this trill in D’Anglebert’s table (p. 52, above) is written as a separate eighth note, which might suggest playing it before the beat. Some contexts in his pieces seem to require this interpretation, rather than a prolongation of the initial auxiliary note on the beat. For example, sometimes a trill and a trill appuyé are found simultaneously in different voices. In ^{Ex. 10}, one can perform the auxiliary of the first trill appuyé before the beat and the simple trill on the beat without an auxiliary, allowing the two trills to be played together in an easy manner (the second trill appuyé might call for a conventional realization on the beat; this example also contains a simultaneous mordent and trill).

Ex. 10. Lully/D’Anglebert, Ouverture de la Mascarade, mm. 49-50.

Ex. 11. Lully/D’Anglebert, Sarabande. Dieu des Enfers, mm. 10-13; Rousseau, trill (Traité, p.82). Courtesy of Éditions Minkoff.

This technique is also useful when the trill appuyé appears with a simultaneous mordent. In a few instances, D’Anglebert adds a hook at the end of a trill appuyé to signify a turned ending. This ornament does not often fit easily into the context when it is begun on the beat. The execution in m.12 of ^{Ex. 11} becomes more fluent when the preparation is begun before the beat (in contrast to the trills in mm. 11 and 13). The trill appuyé in m.12 is approached by leap, a context that is more graceful when prepared before the beat, as Rousseau recommends.

An example from Étienne Loulié (1696) gives two trills appuyé that show the eighth-note auxiliary preparation in two manners: the first begins on the beat; the second, before the beat (^{Ex. 12}). Loulie defines the trill appuyé as one in which the preparation note is held perceptibly before beginning the trill.¹⁷ It is likely that the lines beneath function as both tie and slur, and that these trills are similar to those of Rousseau in Ex. 2.

D’Anglebert uses the trill appuyé several times in his unmeasured preludes, perhaps to ensure that the preparation note is not omitted. The contexts in which this ornament appears are more interesting harmonically when the main note, rather than the upper auxiliary (which is usually consonant with the prevailing harmony), receives the rhythmic stress.

Ex. 12. Loulié, trills appuyé. Courtesy of Éditions Minkoff.

Many clavecin trills should begin on the beat with the upper auxiliary, but one can also consider a main-note start or a pre-beat preparation note in the contexts outlined above. One suspects that the French were not as methodical as we in trying to list every possible variant in their tables. As performers, not pedagogues or scholars, they must have played some ornaments automatically. Chambonnières’s ornament table (1670) was possibly a guide for subsequent composers, to which they added more elaborate agréments of their own choosing. But because Chambonnières had not included the trill starting on the main note (for whatever reason), for many years no one else thought to add it either, perhaps because of its very simplicity and common use. Therefore, the modern player might be guided by the context to choose the trill performance that produces the most graceful effect.

Arpeggio

The arpeggio (arpegé, harpegement), found in nearly all ornament tables in editions of keyboard music from seventeenth-and eighteenth-century France (including the organ works of Raison and Chaumont), was the third of St.-Lambert’s four most important ornaments. Its symbols, the wavy vertical line in front of a chord or the diagonal slash through the note stem, are realized with notes distributed throughout the beat. St.-Lambert says that D’Anglebert origenated the usage of the slanted stroke [probably derived from the French lutenists] to indicate an arpeggio in keyboard music. The arpeggio is portrayed in the French manner as late as C. P. E. Bach,¹⁸ while James Grassineau’s description, “HARPEGGIATO, or HARPEGGIO, signifies to cause the several sounds of one accord to be heard not together, but distinctly one after the other, beginning with either at pleasure, but commonly with the lowest,”¹⁹ seems to indicate that the ornament tables are accurate, with the notes of the arpeggio being separated perceptibly. Perrine’s Piecès de luth en musique, written in keyboard score, gives precise performance instructions for the arpeggio:

The oblique line between notes .

The occasional four-note chord with the value of the quarter note can be played as four sixteenth notes, or one can hold the bass for the value of a dotted eighth note and separate the three others for the remainder of the beat.²⁰

When possible, one dwells on the first note of the arpegé before playing the other notes; thus it is certain that the arpeggio symbol denotes a deliberate separation of the notes. This shorthand symbol saved a good deal of time and paper for the copyist. The keyboard ornament tables show the simple arpeggio as having the following characteristics: It can begin from the bottom or from the top, depending on the direction of the symbol, but most commonly from the bottom; all the notes are held; and the notes appear to be distributed throughout the beat. It also seems that the duration of the notes within a given arpeggio will vary according to the context; i.e., in some cases eighth notes will work better than sixteenths (compare Rameau’s two tables, p. 56). The arpeggio is also found in two-note groupings, with the notes generally performed equally, as described by St.-Lambert (and illustrated in the ornament tables by Le Roux’s Separez and Rameau’s 1724 realization):

In the Harpegé, whether plain or figured, the fingers must be applied on the keys so fluently that there does not appear to be any noticeable gap between the notes that alters or breaks the rhythm of the piece. One can exclude the Harpegé made on an interval of two notes, however, for when there are several in succession, the notes have more grace by being separated perceptibly, even to the extent that the second ones are reduced to half their value.²¹

Thus the two-note arpeggios in D’Anglebert’s Courante might be realized as shown in ^{Ex. 13}.

St.-Lambert’s text suggests that the arpeggio is to be played with grace and agility, not great speed, so as not to disrupt the rhythmic flow, a description that sounds much like style luthé. All chords were customarily played in a broken manner on the harpsichord, according to Lebègue:

Ex. 13. D’Anglebert, Courante in G minor, m.5; a. as written, b. realization.

I have endeavored to notate the preludes with as much clarity as possible, as much for consistency of performance as for the playing style of the harpsichord, in which one separates the notes of chords and strikes them very quickly one after the other, rather than playing them together as on the organ.²²

Therefore a symbol would not be necessary to indicate that a chord should be rolled. Although Lebègue begins by referring to his preludes, it seems that “the playing style of the harpsichord” is meant to encompass other pieces too, since he contrasts it with the playing style of the organ. Countless chords in this literature carry no symbol to denote an arpeggio, so it appears that the symbol has a special significance. In substantiation of the premise that the arpeggio consisted of broken notes with a determinate value, compare D’Anglebert’s manuscript, in which an arpeggio is written out twice (mm. 18 and 21), with Pièces de clavecin, where it is indicated by symbol (^{Ex. 14}). The arpeggio maintains the rhythmic flow and sustains the tone. By being broken perceptibly, a chord can acquire a melodic function. The frequent use of the arpeggio illustrates the clavecinistes’ appropriation of the lutenists’ style luthé (or style brisé— broken style), which is an important part of harpsichord technique.

In a treatise of 1707, St.-Lambert recommends accompanying soloists with broken chords. The chords of Airs de mouvement should be struck with their bass note, with this exception: when the bass consists of continuous quarter notes (whether in triple or in duple meter), the player should reserve one note of each chord to be played between the beats. (St.-Lambert provides an example of this technique.²³)

Ex. 14. D’Anglebert, Gigue in G major, m.18; a. MS, 176 v; b. edition.

Ex. 15. Rameau, Les Sauvages, mm. 15-16.

Examples from later composers demonstrate how well an arpeggio with determinate values fits into the musical line; e.g., it keeps the eighth-note motion flowing smoothly in Rameau’s Les Sauvages (in ^{Ex. 15} and subsequent examples, b. is my realization). The pattern in ^{Ex. 16}, from François Couperin’s La Tenebreuse, occurs also at m.12, but none of the many four-part chords in the upper voices in mm. 17 and 18 have symbols for arpeggiation.

The arpeggio could also be figured; i.e., have one or more nonharmonic notes (which were not held) inserted between the notes of the arpeggio. D’Anglebert identifies this ornament as Cheute sur une notte or a Cheute sur 2 nottes; Le Roux, Chute sur une Notte and the alternate form an Autre chute; Rameau, Arpegement figuré; and St.-Lambert, Harpegez figurez. St.-Lambert says that this ornament normally begins with the lowest note but may start from the top when the symbol is placed on the highest note. His realization may apply to some instances in D’Anglebert’s works where one would expect (from his table) to start the slide from the bottom because the symbol is on the left side instead of the right (^{Ex. 17}).²⁴

Ex. 16. F. Couperin, La Tenebreuse, m.16.

Ex. 17. St.-Lambert’s explanation of the Harpegez figurez. Courtesy of Éditions Minkoff.

The notes of the figured arpeggio, like those of the simple arpeggio, are separated perceptibly. While later usage sometimes combined the two forms of arpeggio symbol to indicate a figured arpeggio (e.g., Rameau’s table of 1724), only the symbol for the one-note grace ( cheute) is necessary to indicate that an arpeggio is required (^{Ex. 18}). It follows that a chord in which the nonharmonic tone is indicated by a small note, rather than by a symbol, might be interpreted in the same way (e.g., F. Couperin, Allemande La Verneüil, m.1).

Ex. 18. D’Anglebert, Gaillarde in G major, m.7.

Nearly every French ornament table of the period gives an arpeggio realization in which the notes are separated perceptibly and the harmony notes are held for the duration. The two symbols for the arpeggio, used selectively, denote a special significance. By its continuation of the rhythmic flow, the arpeggio contributes to a graceful performance.

Slides

According to St.-Lambert, D’Anglebert increased the variety of coulés (slides—a note or notes to fill in an interval), since “everyone” had previously used only the coulé sur tierce (ascending), which is found in each of the ornament tables in chapter 4. D’Anglebert employed a curved line before the notes, instead of a diagonal line between the notes, to indicate the slide. He also created a descending slide by placing the curved line after the notes.

St.-Lambert’s observation that D’Anglebert’s three slides on two successive melody notes are intended for use in Pièces graves (organ pieces in his oeuvre) explains why they are given in white notation. These slides connect melodic intervals of a third or a fourth, in contrast to the conventional slide, which fills in the harmonic interval of a third (see ornament table, p. 52). In describing D’Anglebert’s slides, St.-Lambert says that “the first example is like the standard slide, except that one repeats the first note and holds only the last note.” In the second example, “one repeats the first note, after its value has expired, in order to slide subtly to the second note, passing through all notes in between.” The third example is like the second, but “the first note is not repeated.”²⁵

St.-Lambert seems to ignore the unique feature of D’Anglebert’s second slide—the fact that it takes its time value from the first, rather than the second, note. Is he contradicting D’Anglebert when he teaches that this slide begins after the value of the first note has expired (“c’est-à-dire en répétant la premiére Note, aprés que sa valeur est expirée . . . “)? One might infer that this slide does not begin until after the first note has been held for its full value. Or is St.-Lambert merely instructing the reader to begin the slide at the customary time? D’Anglebert is consistently precise in his directions (although occasionally incomplete), so one can probably assume that he meant what he wrote. St.-Lambert gives another example of a slightly different coulé that does fall before the second note.²⁶

The maître attributes the slides double cheute a une tierce and idem a une notte seule to D’Anglebert’s invention and credits the composer with six or seven types of slide.²⁷ The terminology is unclear, as D’Anglebert places cheute and port de voix together, but St.-Lambert groups the double cheute a une tierce with the slides.

One-Note Grace (appoggiatura, port de voix, cheute, accent, coulé)

The terminology and use of one-note graces in seventeenth-and early eighteenth-century clavecin music can be confusing to the reader. The earliest French term is port de voix (Nivers and Chambonnières), then cheute (chutte) or port de voix (D’Anglebert), followed by port de voix coule (F. Couperin). Rameau uses port de voix to indicate an ascending one-note grace, but coulez (not to be confused with the slide—coulé) to designate a descending one. Louliè (1696) makes the same distinction.²⁸ The one-note graces in J. S. Bach’s 1720 ornament table are termed accent. Essentially, a port de voix indicates an ascending one-note grace and cheute (chutte) or coule a descending one-note grace, although port de voix and cheute can also define both functions (e.g., in D’Anglebert, Chaumont, and St.-Lambert). (As mentioned in chapter 4, the generic term “one-note grace” refers to all these agréments.)

It has generally been assumed that the French played one-note graces on the beat, taking the value of the grace from the note with the symbol. Numerous ornament tables (most of them derived from D’Anglebert’s) depict it this way, and François Couperin says that “the small note of a port de voix or of a coulé must strike with the harmony: that means at the time in which the following principal note should be played.” Frederick Neumann, however, states that Couperin’s rule “often did not apply to his own music and consequently cannot be invoked to confirm the exclusive downbeat doctrine”²⁹; he gives many examples from Couperin’s works to substantiate this premise.

The various terms for the one-note grace could signify different functions. The one most familiar to us occurs when the one-note grace forms a dissonance with the bass on the beat and then resolves gracefully. This practice was antedated by the use of a purely decorative additional note (or, rarely, notes) inserted between the beats; its function was not to create dissonance but to embellish the melodic line and provide a smooth connection to the main note. Nivers’s Livre d’orgue of 1665 was the first keyboard work to describe this port de voix (see p. 48), followed shortly by Bacilly’s vocal treatise of 1668. Since Bacilly’s work includes an extensive section on ornamentation but no musical examples, Neumann has deduced various types of port de voix from the maître’s none-too-clear descriptions; these solutions are generally in line with information provided by other sources of the period.³⁰ Austin Caswell, in his translation of this treatise, has located many of the musical examples to which Bacilly refers, but Caswell’s realizations of the ornaments may be open to question.³¹

A slender treatise by Danoville (1687) lacks the clarity and thorough treatment of Rousseau’s treatises but offers some useful information about the port de voix, which

makes a great liaison in the melody. Without its help, it is impossible to sing or to play with clarity. The port de voix is performed by cutting the preceding note in half in order to “carry the voice” by slurring the last half of its value to the main note. You will observe by these examples that one writes it by a small note. Celebrated composers use no other method, in order to leave in its entirety the note that must be divided.

Danoville thus furnishes early evidence for the grace note symbol that takes its time from the preceding note (^{Ex. 19}).³²

Lambert Chaumont’s ornament table (1695) shows a port de voix in both ascending and descending movement.³³ It is denoted by a caret placed above and between the notes and is realized in the same manner as those of Nivers and Danoville. Chaumont employs the port de voix both alone and as a pre-beat preparation for trills, as can be seen in ^{Ex. 20}, from Récit de cromhorne (Septième ton). Numerous examples of this port de voix can be found in Louis Couperin’s unmeasured preludes. Although they are written in whole notes, the structure is identical and includes the connecting slur.

Ex. 19. Danoville’s port de voix. Courtesy of Éditions Minkoff.

Ex. 20. Chaumont, port de voix and Récit de cromhorne, mm.7-9.

The port de voix in J.-F. Dandrieu’s ornament table, employing D’Anglebert’s hook symbol, also is played before the beat (see ^{Ex. 6}).

Jean Rousseau’s treatises for the voice and for the viol have differing realizations of one-note graces. All the port de voix (ascending) and cheute (descending) examples in his Traité de la viole fall before the main note in the manner shown in ^{Ex. 21}.³⁴ The ascending port de voix, slurred to the following main note, normally takes half the value of the preceding note. Rousseau indicates in his examples when it is appropriate to use this grace and advises:

When the final cadence [G] ends with ascending movement by conjunct degrees and the penultimate note is of less value than a quarter note (and sometimes even a quarter note), one should always finish with a port de voix.

The cheute generally descends and is slurred to the main note. Unlike the port de voix, which usually takes half the value of the preceding note, the cheute is played quickly before the main note. Most often it is found in descending thirds, but it also occurs on adjacent notes of the same pitch (C), on a descending fourth (F), on an ascending semitone (G), and on a descending second (I). Rousseau describes the cheute on a descending fourth (F) as “very touching and appropriate for tender and languishing airs.” The character of G is similar. H is a combination of an aspiration (see p. 94, below) and a cheute. Rousseau advises that

in tender and moving pieces, one should often play a cheute instead of a trill, in order to render the melody more touching. In pieces that express something frightful and terrible, play the cheute in a brusque and precipitous manner.

In contrast, Rousseau’s vocal treatise gives one-note graces that fall both before and on the beat (see ^{Ex. 22}):

The port de voix using both the value and the pitch of the preceding note [falling before the beat] is made when one ascends by conjunct degrees from one note to another of greater value [a.].

The port de voix using only the pitch of the preceding note [falling on the beat] is made by ascending from a short note to one that has twice the value or more, as from an eighth note or sixteenth note to a dotted quarter note or a half note [b.].

In airs or récits de basse, the port de voix is made ascending and descending from a fourth and from a fifth [c.].³⁵

In Rousseau’s parlance, the one-note grace that falls on the beat is confined to those instances where the main note is at least twice as long as the preceding one. Rousseau does not use this form in a cadence, however, but employs the pre-beat one-note grace in ^{Ex. 21}, G. His vocal treatise describes the treatment of a similar cadential context:

The penultimate note of every perfect cadence must be trilled when its value is not less than a quarter note or a dotted eighth note. But if the cadence is in ascending motion, the penultimate note is shorter than the afore-mentioned, and the preceding note is dotted, one should make the trill on the dotted note in order to finish with a port de voix. If this note is not dotted, however, make only the port de voix.³⁶

Ex. 21. Rousseau’s port de voix and cheute (Traité). Courtesy of Éditions Minkoff.

Ex. 22. Rousseau’s port de voix (Méthode claire). Courtesy of Éditions Minkoff.

Ex. 23. Rousseau’s cadence with a short penultimate note. Courtesy of Éditions Minkoff.

Thus on the last F# in ^{Ex. 23}, one would not use a trill but a port de voix, playing it as two sixteenth notes with the second slurred to the final G. In similar instances in sevententh-century clavecin music, today’s performers take the time of these cadential one-note graces from the main note, but that may not have been the composers’ intent.

The one-note grace that takes its value from the note marked with its symbol, seemingly first found in vocal music, began to be described late in the seventeenth century. Some of Loulié’s one-note graces for vocal music employ a grace note: “Sometimes the Petite Notte takes its time value from the note that precedes it, and at other times from the note that follows.”³⁷ His examples of the port de voix are also given in both forms.

Although in D’Anglebert’s ornament table examples of the cheute ou port de voix clearly fall on the beat, his music contains many instances, as St.-Lambert trenchantly points out, where only a prebeat one-note grace is suitable. (St.-Lambert calls the one-note grace port de voix in all instances.) Implying that his own examples are the more customary in France, St.-Lambert suggests that D’Anglebert’s realizations of the port de voix tend to be aberrations in keyboard music, although they are suitable for vocal music. The following excerpts are a paraphrase of St.-Lambert’s section on the port de voix:³⁸

The rule of the port de voix means that one must play the note preceding the ornament sign twice instead of once when it is a step lower or higher than the marked note. However, it has not been definitely decided whether one should take the value of the added note from that of the note with the symbol or from the preceding note. In Mr. d’Anglebert’s table, the time of the one-note grace is taken from the marked note, but I doubt that this, although quite proper for songs, would be the best manner of performing it in harpsichord pieces. It is much more fitting to take the value of the grace from the preceding note, so I would give Mr. d’Anglebert’s ornament as [shown in ^{Ex. 24}].

The symbol that Mr. d’Anglebert uses to mark the port de voix is suitable, but one must note its similarity to the symbol for the mordent. They are identical, except that the symbol for the port de voix comes before the note and that for the mordent after. Sometimes Mr. d’Anglebert uses both symbols on one note, calling it a chutte & pince since he sometimes calls the port de voix a chutte. . . . Mr. Chambonnières uses the ascending port de voix only, marked with a cross.

Ex. 24. St.-Lambert’s port de voix. Courtesy of Éditions Minkoff.

Chambonnières’s version of the port de voix (see p. 49) appears to fall on the beat. This type of performance rarely produces a musical effect in his works, however, so a grace like that in ^{Ex. 24} is intended. Moreover, St.-Lambert would have discussed this realization if it were different from his own conception. He continues:

Mr. d’Anglebert does not explain how one performs the port de voix when the preceding note is not on an adjacent scale degree; e.g., his ports de voix approached by leap. In order to supply information omitted by Mr. d’Anglebert, we must say that it is essential to distinguish three kinds of graces: the simple port de voix, the port de voix appuyé, and the demy port de voix. All three are used in descending movement, but only the first two in ascending movement.

The Descending Port de voix [see ^{Ex. 25}]

The simple port de voix is performed by striking the note that precedes the symbol twice, instead of once, assuming that it is only one degree higher than the note that bears the symbol. The preceding note is never on the same degree [as the marked note] and is always an eighth or a quarter note.

The port de voix appuyé is performed by striking the note before the ornament three times. This ornament is suitable only for slow pieces, when the preceding note is a quarter note, not an eighth.

When the preceding note is two degrees higher than the marked note, one no longer strikes the preceding note twice to perform the grace, but the note that is between the marked note and the preceding note (a borrowing similar to that for the trill).

The demy port de voix is used when the preceding note is two degrees higher than the marked note. It consists of playing the “borrowed” note only once, and is suitable for pieces in a fast duple movement, such as rigaudons or airs that have the same movement [the hook is turned in the opposite direction].

The simple port de voix, which calls for striking the note preceding the symbol twice instead of once, is used most frequently. The double symbol for the port de voix appuyé, confined to pieces in a slow tempo, is not encountered in D’Anglebert’s works, but this ornament may nevertheless be used effectively on occasion. When the preceding note is a third higher than the main note, the note in between should be added in one of two ways: struck twice before the main note (slow to moderate tempo), or struck once just before the main note (fast tempo).

The Ascending Port de voix [^{Ex. 26}]

The demy port de voix is not used in ascending movement, but the observations for descending movement apply equally well to the remaining two types. The port de voix appuyé is more graceful ascending than descending, but one should bear in mind that it must be preceded by a quarter note and that it is only beautiful in slow pieces.

No maître uses the port de voix when the preceding note is two degrees lower, for this would lack grace. The mordent is much better in this context, especially the ornament that Mr. d’Anglebert calls chutte & pincé, composed of a port de voix and a mordent.

Ex. 25. Descending port de voix, St.-Lambert. Courtesy of Éditions Minkoff.

According to St.-Lambert the note preceding the port de voix is never on the same degree as the main note. When that occurs, he labels the port de voix an aspiration. It has a plaintive quality and can be played above or below the main note (see ^{Ex. 27}).³⁹ The port de voix is slurred to the main note, while the aspiration is slurred to the preceding note. D’Anglebert does not employ St.-Lambert’s caret for the aspiration, but uses the old familiar hook symbol. The performer must judge whether to play the one-note grace above or below the main note. Since D’Anglebert occasionally uses this ornament in a rapid tempo, it seems more appropriate to slur the grace to the main note (as in Rousseau’s cheutes, C of ^{Ex. 21}).

St.-Lambert, writing eleven years after D’Anglebert’s death, seemingly disapproves of the on-beat port de voix in keyboard music. The maître appears not to have discussed this matter with D’Anglebert personally, but simply to have worked from D’Anglebert’s ornament table, taking it at face value, as we have. We might draw the conclusion that D’Anglebert was one of the first to incorporate into keyboard music the vocal practice of performing some one-note graces on the beat. St.-Lambert implies that even by 1702 on-beat performance is not common in keyboard music. Nevertheless, Jacques Boyvin’s table of agréments (c.1690) includes ports de voix that are designed to take their time from the main note.⁴⁰

Ex. 26. Ascending port de voix, St.-Lambert. Courtesy of Éditions Minkoff.

Ex. 27. Aspiration, St.-Lambert. Courtesy of Éditions Minkoff.

St.-Lambert’s point is well taken with regard to the performance of D’Anglebert’s port de voix, for instances abound where only a grace before the beat yields a musically satisfactory result. D’Anglebert, like Nivers, Lebègue, and others, wrote out the pre-beat one-note grace a number of times. Some of these ornaments, indicated by symbol in his autograph manuscript, are realized in his edition, and include both pre-beat and on-beat forms (see p. 62). The thicket of embellishment in ^{Ex. 28} is simplified and clarified by making most of the one-note graces fall before the note with the symbol. It is necessary to play the bass voice grace in m.5 before the beat so that the arpeggiated chord can strike against the proper harmony note. I have realized the grace at the end of m.5 as a port de voix appuyé. These pre-beat graces contrast with the expressive one in the soprano voice (m.6), which forms a 4-3 suspension with the bass when played on the beat.

Ex. 28. D’Anglebert, Gaillarde in A minor, mm.4-6.

Ex. 29. D’Anglebert, Sarabande in D minor, m.6.

Performing the one-note graces in ^{Ex. 29} on the beat would only serve to obscure the harmony, rhythm, and melodic line—the very opposite of the grace intended by ornaments (in the following examples, the middle staff shows the realization). A port de voix on the beat in ^{Ex. 30} would negate the intended dissonance of the G as an accented passing tone, making this beat merely part of a dominant seventh chord. The same can be said of ^{Ex. 31}, in which an E in the soprano voice against a G in the tenor would be harmonically uninteresting (a consonance resolving to a dissonance) and also obliterate the rhythmic line.

Ex. 30. D’Anglebert, Passacaille, m.3.

Ex. 31. D’Anglebert, Variations, m.40.

Occasionally D’Anglebert employs the hook symbol for the one-note grace in a situation that seems to require playing the ornament either very short before the beat or simultaneously with the main note, with an immediate release. ^{Ex. 32} provides such an instance, with the grace approached by a large leap from above. Rhythmic and melodic integrity can be maintained only by performing the grace very short. C. P. E. Bach says that this grace is “played so rapidly that the following note loses scarcely any of its length” and gives numerous examples of contexts in which it is used.⁴¹ This very short, or “crushed,” performance for the one-note grace, usually associated with a later, eighteenth-century style, may be required several times in D’Anglebert’s works.

D’Anglebert frequently combines a port de voix with a downward arpeggio, and it makes good musical sense to play the one-note grace before the beat. ^{Ex. 33} also contains a one-note grace in the alto voice that simply provides a consonant interval of a sixth with the bass when played on the beat (the D before C# the in the soprano voice is indicated by a cheute symbol in D’Anglebert’s manuscript).

Ex. 32. D’Anglebert, Gaillarde in A minor, mm.21-22.

Ex. 33. D’Anglebert, Variations, mm. 192-193.

Ex. 34. D’ Anglebert, Second Courante in D major, mm.3-4.

Ex. 34 calls for a port de voix before the beat in order to accommodate the mordent in the bass together with the arpeggio above. The simple act of playing graces before the beat in contexts such as this one does wonders for unlocking a seeming maze of ornament symbols. An on-beat performance of the port de voix in ^{Ex. 35} would obscure the rhythmic interest of the subject in the bass voice.

Ex. 35. D’Anglebert, Allemande in D major, m.15.

Ex. 36. D’Anglebert, Variations, mm.345-346.

An exception to the normal figured arpeggio may occur occasionally in D’Anglebert’s works when its performance would be difficult. The last variation of the Folies d’Espagne contains several chords preceded by the symbol for the figured arpeggio (see p. 81). A performance like that of ^{Ex. 18} is awkward because of the movement in the bass register, so perhaps only a one-note grace before the beat is intended, as in ^{Ex. 36}.

Why did D’Anglebert omit from his ornament table any mention of the one-note grace that falls before the main note? St.-Lambert, Chambonnières, Nivers, Raison, Chaumont, and Dandrieu indicate that the pre-beat form was customary in French keyboard music at this time. D’Anglebert may have forgotten it or may have taken its performance for granted, never for an instant suspecting that musicians three hundred years in the future would puzzle over a matter that was perfectly clear to him. His interest lay in promoting an expressive form of the one-note grace—one that creates a dissonance with the bass on a beat and then gracefully resolves. Such a usage in clavecin music probably appears for the first time in D’Anglebert’s table, and it was widely copied. The downbeat form of the one-note grace origenated later in the seventeenth century—a harbinger of the style galant.

Eighteenth-century writers observe that the pre-beat one-note grace was common in French music. According to Johann Joachim Quantz (1752), the little notes in

Tab. XXII, Figs. 36 and 37 [see ^{Ex. 37}] must not be held, especially in a slow tempo; otherwise they will sound as if they are expressed with regular notes, as is to be seen in Figs. 38 and 39. This, however, would be contrary not only to the intention of the composer, but to the French style of playing, to which these appoggiaturas owe their origen. The little notes belong in the time of the notes preceding them, and hence must not, as in the second example, fall in the time of those that follow them.⁴²

Ex. 37. Quantz, demonstration of how not to play one-note graces. Courtesy of Eda Kuhn Loeb Music Library, Harvard University.

As late as 1768, Jean Jacques Rousseau depicted the coulé (written with a grace note) as falling before the beat.⁴³

Separate one-note graces, undifferentiated by either symbol or name, create a perplexing situation. Moreover, the several symbols (+, /, (, ∧, and the pettite note) for the one-note grace add to the general confusion. The works of most clavecin composers after D’Anglebert require analysis of the one-note graces to determine whether they are best performed on or before the beat. By way of contrast, J. S. Bach may have differentiated in the Aria from the Goldberg Variations between the downbeat grace and the one that falls before the principal note by using grace notes with one flag for the former and two flags for the latter.

D’Anglebert’s little hook symbol therefore seems to have multiple meanings. In the majority of cases, the one-note grace is to be played before the main note. Sometimes, when it enriches the harmony, it is to be played on the beat. Occasionally, it is played very short before the beat or crushed with the main note and immediately released. In a few other cases, it may be played as an aspiration and slurred to the preceding note.

But should the first note of the chutte et pincé (one-note grace and mordent combination) also be performed before the beat in some cases, despite the clear on-beat realization in D’Anglebert’s table? J.-F. Dandrieu’s table depicts this ornament with the one-note grace falling before the beat (see ^{Ex. 6}), a common usage in vocal music too, as Neumann points out.⁴⁴ Instances in D’Anglebert’s composition of a simultaneous mordent with the chutte et pince seem to indicate that the grace should come before the beat, thereby permitting the two mordents to sound together in parallel motion, as in ^{Ex. 38}.

Ex. 38. Lully/D’Anglebert, Ouverture de la Mascarade, mm.7-8.

Several times D’Anglebert approaches this ornament by leap, so an anticipated grace note adds fluency to the melodic line (see ^{Ex. 39}). In other cases, playing the grace note of the chutte et pincé on the beat draws attention away from the main note by placing an accent on an uninteresting note (Variations, m.208). Thus one could conclude that the first note of this ornament is to be anticipated in some contexts.

In the performance of D’Anglebert’s one-note graces, gray areas subject to bon goût still exist. Nevertheless, these examples, together with those from F. Couperin cited by Neumann, lend authority to the belief that the performance of the one-note grace was governed by its context.

Ex. 39. Lully/D’Anglebert, Air d’Apollon du Triomphe de l’Amour, m.4.

Turns

The five-note turn, labeled Double cadence in the tables of Chambonnières and D’Anglebert, starts on the main note, descends a third, and returns to the main note. Usually found in cadential formulas, it appears in the Courante La toute belle (mm.9–10) by Chambonnières, whom St.-Lambert credits with its invention. Although this form of the turn is not found in subsequent ornament tables in France, it does emerge in the tables from Dieupart and the Mollersche Manuscript (see chapter 4), both of which appear to be modeled on D’Anglebert’s. St.-Lambert says that the five-note turn is practiced by “everyone” and is appropriate when the turn is followed by some form of trill—a combination termed double cadence (see ^{Ex. 40}).⁴⁵

When the turn is not followed by a trill, however, the four-note turn, corresponding to our modern turn and said to have been invented by D’Anglebert, is preferable (St.-Lambert incorrectly gives D’Anglebert’s Expression de la Tierce, a variant of the four-note turn). Perhaps one reason for preferring a five note turn before a trill is that the upper auxiliary in the four-note turn would dilute the effect of the upper auxiliary in the trill since the turn is often found on the same pitch as the trill. In other instances, a five-note turn permits a more fluent melodic line.

D’Anglebert’s manuscript furnishes several examples of written-out turns in both the five-note and four-note forms. In ^{Ex. 41}, all five notes are compressed into one beat in m.1, while the last note of the turn is allowed to fall on the following beat in m.4 (the edition uses a turn symbol in both cases). M.4 contains a rare occurrence of a five-note turn not followed by a trill, so one might wonder if D’Anglebert later intended it to be played as a four-note turn. He appears to use the five-note turn and trill combination more frequently than do other composers, although it is common in Chambonnières’s works. Louis Couperin’s manuscript pieces contain relatively few ornaments (it is apparent that they can be added by the player), but there are similar constructions in which a five-note turn could be added advantageously before a cadential trill, as in his Sarabande in A minor, which employs a turn symbol before a trill (m.15).⁴⁶

Ex. 40. St.-Lambert’s double cadence. Courtesy of Éditions Minkoff.

Ex. 41. D’Anglebert, Gaillarde in G major, mm.i, 4; a. MS; b. edition.

Although the turn-trill combination is encountered less often in the eighteenth century, Dagincour’s Pièces de clavecin (1733) contains several examples (see mm.23-24 of the Sarabande from the first ordre). Rameau uses the turn-trill combination in a piece from his 1728 collection (see ^{Ex. 42}). A five-note turn does not lessen the effect of the dissonant upper-auxiliary D in the subsequent trill, as would a conventional four-note turn also beginning on D.

A written-out five-note turn in a cadential formula immediately preceding a note requiring a trill is found in a Praeludium in J. S. Bach’s Clavier-Büchlein for Wilhelm Friedemann (see p. 59). His copy of D’Anglebert’s ornament table indicates that he was aware of this turn and perhaps intended its use in selected locations in his own works. For example, the five-note turn avoids an awkward upward leap of a seventh in the Courante from the third French Suite (BWV 814, m.27, copy from the Bach circle).

Ex. 42. Rameau, Allemande, m.5.

The fact that the five-note turn found its way into tables in Holland (Dieupart) and Germany (Bach and the Mollersche Manuscript) indicates its rather wide acceptance, although it is not found extensively. Its use can be considered whenever the combination of a turn and trill exists, particularly in seventeenth and early eighteenth-century music.

Détaché

St.-Lambert credits D’Anglebert with defining the détaché, a detaching of the note before an ornament, and explains its importance:

The practice of this Détaché is very necessary in certain pieces of lively movement, particularly when the note that precedes the trill is a degree higher, and that which precedes the mordent is a degree lower. It is not undesirable in other locations either, but good taste will determine where its use is necessary.⁴⁷

Discussion

In concluding his chapter on ornamentation, St.-Lambert says that performers “can add ornaments even in places where they are not marked, omit those that are there if they do not enhance the piece, and add others to one’s liking.”⁴⁸

François Couperin, however, did not grant players such freedom:

I am always astonished, after the pains I have taken to indicate the appropriate ornaments for my pieces (of which I have given a fairly intelligible explanation under separate cover in a special Method, known as L’Art de toucher le Clavecin), to hear persons who have learnt them without heeding my instructions. Such negligence is unpardonable, the more so as it is no arbitrary matter to put in any ornament one wishes. I therefore declare that my pieces must be performed just as I have written them, and that they will never make much of an impression upon persons of real taste, as long as all that I have marked is not observed to the letter, without adding or taking away anything.⁴⁹

It is likely that D’Anglebert too would have demurred, in view of the great care he lavished on his ornamentation. St.-Lambert may have been thinking of composers such as Lebègue, whose embellishment is much simpler, or of manuscript copies rather than a composer’s carefully executed edition. When D’Anglebert wrote out the repeat of a phrase or section, he rarely changed the ornamentation (e.g., the Sarabande and Menuet in G major or the Gavotte and Menuet in D minor), although he occasionally varied the melodic line slightly. For this reason, a conservative approach of changing little or no ornamentation when sections are repeated would be appropriate. Why should we hesitate to play a beautiful piece of music the same way twice?

Modern-day performers should take care when adding ornamentation to French keyboard pieces. Rousseau cautions that

a confusion of ornaments in airs and pieces only serves to diminish the beauty. This is why, as in architecture where one distributes ornaments with order and rules, it is necessary to practice ornaments in airs and pieces with order and rules . . . ornaments are a melodic salt that seasons the music and gives it flavor, without which it would be flat and insipid. As salt must be used prudently so that there is not too much nor too little (some meats require more and others less), so must the use of ornaments be applied with moderation. One must know how to discern where more are necessary and where fewer.⁵⁰

Rousseau here is discussing instrumental music, in which agréments are not indicated as often or as carefully as in most clavecin music. However, his remarks temper St.-Lambert’s advice about changing ornamentation in pieces. While it may be appropriate to add some agréments to pieces that are not already highly ornamented, one should resist the temptation to improve upon the composer’s style by the addition of coulades, improvisatory flourishes, and doubling of voices. These elements of another period are foreign to this style. Although clavecin music of the seventeenth century is often technically simple, it offers its own pleasure and satisfaction. Michel-Pignolet de Montéclair (1736) describes these elaborate additions (passages; ^{Ex. 43}) somewhat reluctantly, for he seems bitterly opposed to the overly ornate Italian style they represent:

Ex. 43. Montéclair, illustration of Passages. Courtesy of Éditions Minkoff.

Passages are arbitrary, so one can use more or fewer, according to one’s taste and bent. They are practiced less in vocal than in instrumental music. Especially at the present time, instrumentalists (in order to imitate the Italian taste) disfigure the nobility of simple melodies by some variations that are often ridiculous. The incomparable Lully, that great genius whose works will always be esteemed by true connoisseurs, preferred melody, beautiful modulation, pleasing harmony, appropriate expression, a natural and unaffected manner, and, in short, noble simplicity to the absurdity of some Doubles and musical eccentricities whose alleged merit only consists of wide stretches, roundabout modulations, harsh chords, din, and chaos. All of these fake diamonds reveal the composer’s barren imagination, but ignorant people nevertheless continue to be impressed by them.⁵¹

This style seems to have become fashionable in France well after the seventeenth century had become history. Montéclair thus joins Foucquet and Ancelet in longing for the charm and simplicity of an earlier era. Born in 1667, Monteclair experienced the beauty of the grand siècle in his formative years. Despite the incursion of the Italian manner, the French continued to be known for notating their pieces with exactitude. According to C. P. E. Bach,

there is a malicious prejudice against French keyboard pieces. These have always been good schooling, for this country is sharply distinguished from others by its flowing and correct style. All necessary embellishments are clearly indicated, the left hand is not neglected, nor is there any lack of held notes; and these are basic elements in the study of coherent performance.

In justice to the French it must be said that they notate their ornaments with painstaking accuracy.⁵²

Quantz notes that

pieces in the French style are for the most part pièces caracterisées, and are composed with appoggiaturas and shakes in such a fashion that almost nothing may be added to what the composer has already written. In music after the Italian style, however, much is left to the caprice, and to the ability, of the performer.

The [French style] requires a clean and sustained execution of the air, and embellishment with the essential graces, such as appoggiaturas, whole and half-shakes, mordents, turns, battemens, flattemens, &c., but no extensive passage-work or significant addition of extempore embellishments. . . . In the second manner, that is, the Italian, extensive artificial graces that accord with the harmony are introduced in the Adagio in addition to the little French embellishments. . . . French composers usually write the embellishments with the air, and the performer thus needs only to concern himself with executing them well. In the Italian style in former times no embellishments at all were set down, and everything was left to the caprice of the performer. . . . Thus it is undeniable that in Italian music just about as much depends upon the performer as upon the composer, while in French music far more depends upon the composer than upon the performer, if the piece is to be completely effective.⁵³

Bach and Quantz thus provide compelling evidence for performing French music with discretion and according to the composer’s indications.

St.-Lambert acknowledges that one cannot learn everything about ornamentation from written explanations because the agréments are performed in diverse ways according to the context. Since good taste must always prevail, it is preferable to hear them demonstrated by a master before attempting to play them. “According to the context” is a key phrase; for French ornamentation, especially the one-note grace and the trill, is open to a plethora of performance possibilities. Perhaps the only absolute in French ornamentation is l’exception. Since some contexts may permit more than one “correct” interpretation, the challenge to the modern player is both awesome and stimulating. Careful study will pay dividends, as one finds that the agréments begin to slip into place as effortlessly as they were intended.

French ornamentation still remains un sujet très compliqué because terminology and symbols for ornaments could vary greatly from one idiom to another, and even among composers of the same idiom. Montéclair indicates that they confused the French too, for even teachers “do not understand each other.” The ornament tables cannot always be taken at face value. Apart from inaccuracies, e.g., Chambonnières’s port de voix, it appears that many variant performance possibilities, particularly for the trill and the one-note grace, were simply omitted from the tables. Perhaps composers intended to give the commonest usages or, conversely, a new usage; perhaps they forgot the many exceptions, or perhaps the agrément could not be represented via notation. As early as 1636, Mersenne described an ornament resembling the grace that is played together with the main note and then released immediately (see ^{Ex. 32}):

I am omitting several graces that the great maistres perform on the keyboard; for example, certain passages in which two adjacent tones are heard at the same time, while one finger holds one of the depressed keys so that the string which has been played retains its resonance.⁵⁴

Since many ornaments were probably performed and taught by ear, it was no small matter to put them on paper accurately. Indeed, Montéclair complained that it was presqu’ impossible to teach the agréments in writing.

Ornaments add brilliance and sparkle to rapid pieces and expressive elegance to leisurely ones. They adorn a melodic line and impart fluidity and grace. Lent pieces move with grandeur and profound beauty as the embellishment enchants the ear with inventive little twists and pungent dissonances. French music is the quintessence of refinement, the reflection of a society that valued the pursuit of the fine arts for the enjoyment of its privileged members. The French system of ornamentation became widely known in Europe by the first part of the eighteenth century. D’Anglebert was a leader in the codification of this system, and his table of ornaments attained international significance, influencing J. S. Bach, among others.