The table of 29 ornaments (agréments) included with D’Anglebert’s Pièces de clavecin (1689) greatly expanded the number of ornament symbols used to add grace and elegance to French keyboard music. D’Anglebert played an important role in the development of shorthand symbols for keyboard ornamentation. These symbols were devised primarily in France, during the latter part of the seventeenth century, and were widely copied or varied in other parts of Europe.
Early in the century and before, diminutions affecting the entire melodic line often served as embellishment. Ex. 1, containing three written-out trills with turned endings, is the last of several possibilities given in Marin Mersenne’s Harmonie universelle (1636) to demonstrate diminutions by le sieur de la Barre on the first two measures of “Tu crois o beau soleil”, a chanson composed by Louis XIII.1
Concurrent with the use of diminutions was the development of agréments, which affect only single notes or chords. Mersenne describes the performance of various types of lute tremblemens (trills), gives some symbols for them (but no realization), and mentions that these ornaments can also be performed on the harpsichord.2 The vocal portion of his treatise, however, illustrates a trill preceded by a note of the same pitch and the port de voix in its classic seventeenth-century form as a tone that is slurred to the main note and takes its value and its pitch from the preceding note (see p. 85).3 The Traité de l’accord de l’espinette (1650), by Jean Denis, discusses ornaments briefly, but does not include an ornament table.4
Early references indicate that performers could add ornamentation; e.g., Jean Titelouze (1624) instructs his readers to use the “common cadences [trills] that everyone knows”5; and Mersenne notes the extemporaneous ornamentation practiced by Pierre (III) Chabançeau de La Barre (1592-1656): “But it would be necessary to have several special symbols to mark the places of the martelemens, the tremble mens, the battemens, and the other graces with which this excellent organist enriches his playing at the keyboard.”6 In 1655 Constantijn Huygens wrote to Henri Du Mont asking to hear him play Huygens’s allemande with the “ornaments that can scarcely be expressed in a musical score.”7 Perhaps these writers are referring to ornaments that have the same general shape as the turned trills of the diminution in Ex. 1 but are performed much more freely.
Most French keyboard composers from G.-G. Nivers (1665) to those writing in the first part of the eighteenth century included an ornament table in their editions. Their music is carefully marked, in contrast to vocal and instrumental music, where the performer may be expected to add some of the ornamentation. Some ornaments, such as the port de voix and the turn, can be written out in conventional notation; but others, such as the various forms of the trill, can be only approximately realized via notation. The speed and the number of repercussions must have varied with the context. The commonest French ornaments that can be readily translated into English usage include:
|Tremblement or Cadence||Trill|
|Cheute or Port de voix||One-note grace (appoggiatura)|
|Coulé||Slide (later, a one-note grace)|
|Arpegé or Harpegement||Arpeggio|
Since appoggiatura is an Italian term closely identified with an on-beat ornament, let us adopt Frederick Neumann’s “one-note grace” as a generic term to refer to all French ornaments that add a single tone.8
The agréments were considered an integral part of the melody and not merely decoration, according to Benigne de Bacilly (1668):
As in all cases where one makes a distinction between beauty and ornament, it is the same in song, where a piece of music can without a doubt be beautiful, but not please the listener because it has been executed without the necessary ornaments.9
Keyboard Ornament Tables in Seventeenth-Century France
Guillaume-Gabriel Nivers, the first keyboard composer to include a table of symbols for his ornaments (1665), calls the mordent approached from below an agrément ( ), the trill a cadence (), and the trill with a turned ending a double cadence ().10 The last, a particularly common ornament (its written-out form appears in Ex. 1), was termed a cadence parfaite by Denis.11 There was much confusion of terminology in the seventeenth century, particularly with the cadence and the various forms of the one-note grace. Nivers’s form of the mordent was followed by the organist Gilles Jullien (1690), while the harpsichordists often included a lower one-note grace with the mordent, making the execution identical. Nivers advises slurring the port de voix (one-note grace) to the following note and suggests consulting a singing manual since the organ should imitate the voice. Port de voix means literally “carrying the voice,” so the notes, indicated by the small lines above, should be connected smoothly (Nivers writes out the ports de voix in his music). He also describes the slide, consisting of a third with an oblique line between the notes, which is performed with only the outer notes held (Ex. 2).
Jacques Champion de Chambonnières’s Les Pièces de clauessin (1670) was the first harpsichord publication to include a table of ornaments. His realization of the port de voix appears to be on the beat, but its repeated note and rhythmic pattern resembles Nivers’s port de voix. This ornament is the same as Nivers’s and should be notated as two sixteenth notes preceding a quarter note (see p. 92 below). The five-note turn beginning on the main note is especially important in seventeenth-century French literature (see p. 102), while the remaining ornaments and their symbols became widely used in France and elsewhere (see Ex. 3).
In Nicolas-Antoine Lebègue’s table of ornaments (1677) the trill is called a Cadence ou tremblement and the symbol is given for the petite reprise, the short final repeat found in many French pieces.12 Otherwise, his table is the same as Chambonnières’s, but it lacks the port de voix and double cadence (see Ex. 4).
Elisabeth-Claude Jacquet de La Guerre’s recently discovered first book of Pièces de clauessin (c.1687) employs the same ornament symbols as are in Chambonnières’s books. In the only extant copy, a leaf has been torn out after the title page, so that the first part of the dedication is missing; and two leaves, which probably contained an ornament table, have been cut out after the dedication. La Guerre’s ornaments, except for turns, work well when performed in the manner of Chambonnières’s. Her turns, seldom used in a cadential formula before a trill (see p. 102 below), are usually approached stepwise from below, so one can speculate that they are to be performed in the manner of D’Anglebert’s four-note turn instead of Chambonnières’s five-note version.
André Raison’s (1688) table shows the port de voix, indicated by a slur, beginning before the beat.13 Raison inserts a rest before the termination of most trills and begins the last trill with a mordent. His double cadence is yet a different rendition of this term, and it uses two symbols, the first of which is a modified turn symbol. Noteworthy is the inclusion of the arpeggio in an organ table. Raison’s table (Ex. 5) illustrates a wider variety of possibilities, suggesting that ornament performance in France at this time offered many choices.
In D’Anglebert’s ornament table, shown in Ex. 6, the basic trill is now termed tremblement, rather than cadence. Like Jean Denis, D’Anglebert distinguishes between single and double mordents. Several ornaments make their first appearance in this table:
The trill appuyé, in which the upper auxiliary is held at the beginning before starting the trill.
The cadence, an unusual form of the trill (which can start above or below the main note), apparently not encountered elsewhere in the clavecin tables, but included in J.S. Bach’s table in the Clavier-Büchlein for Wilhelm Friedemann and used by later German composers.
The double cadence, consisting of two ornament symbols—a turn and a cadence, with the turn being the same as Chambonnières’s five-note turn. When the turn is not followed by a trill (sans tremblement), it has four notes and begins on the upper auxiliary—the form that became commonly accepted throughout Europe.
The coulé sur 2 notes de suitte (slide on two successive notes), consisting of filler notes between two melodic notes, in contrast to the conventional slide, which adds the inner note between two notes written as the harmonic interval of a third. D’Anglebert’s notation indicates precisely that the melodic slide may begin before or on the beat, probably depending on the context. The first note of this type of slide is not held.
The symbol for cheute or port de voix (one-note grace) placed before a chord, indicating a figured arpeggio with one or two nonharmonic tones inserted. Also, a unique symbol for the cheute can indicate two unusual renditions: double cheute à une tierce (double cheute on a third) and idem à une notte seule (the same on a single note).
The detaché (detaching before a trill or mordent), using the seventeenth-century form of the eighth rest (attached to the note head) to indicate a rest before the ornament.
D’Anglebert also inaugurated the use of different signs for previously known agréments:
D’Anglebert’s Influence on the Ornamentation
of Later Composers
The ornament tables of a number of composers of the next generation bear a clear visual resemblance to D’Anglebert’s.
“Rules for Graces,” printed in both French and English in Six suittes de clavessin of Charles (François?) Dieupart (c.1702), was largely modeled after D’Anglebert’s table.14 (The suites were subsequently published in an incomplete form by J. Walsh of London, apparently without the table.) An almost exact copy of Dieupart’s table appears in the Mollersche Manuscript (dating perhaps from before 1707),15 which contains works by J. S. Bach (including an autograph), Georg Bohm, numerous other German composers, and Lebegue (including his ornament table).16 The table patterned after Dieupart immediately precedes the pages in Bach’s hand. The copyist has enlarged the curved lines in the last and third from last examples (see Ex. 7).
French influence extended to Germany through the court at Celle when Georg-Wilhelm, Duke of Braunschweig-Lüneburg, married Eleonore Desmier d’Olbreuse, who was originally from Poitou. Many French musicians were in their service from before 1685, and Georg Böhm became part of this circle. J. S. Bach too was influenced by the Celle court during his student days in Lüneburg, so the Möllersche Manuscript is possibly an outgrowth of relationships formed in this region.
Gaspard Le Roux’s ornament table of 1705 (Ex. 8) is patterned after D’Anglebert’s, with the exception of the Autre chute and the Separez.17 The inclusion of the Separez indicates that the diagonal line between two notes a third apart could mean either a slide or a separation of the two notes (see p. 79).
The two tables from Jean-Philippe Rameau’s books of clavecin pieces (1706 and 1724) also use many of D’Anglebert’s forms for the symbols (see Exx. 9 and 10).18 In both tables, the mordent is shown only in multiple repercussions. There are several interesting features in the 1724 table:
1. The Double cadence equals a trill with a turned ending.
2. The four-note turn is called a doublé.
3. The one-note grace is shown as being tied against the resolution momentarily.
4. The Coulez is not a slide but a descending one-note grace.
5. The Suspension (a delaying of the melody note—also given by F. Couperin in 1713) is added.
6. The arpeggio illustration at the bottom depicts two figures—one with four notes played as sixteenth notes and one with two notes played as equal eighth notes.
7. “A note slurred to the following trilled note serves as its preparation,” so the trill proper begins on the main note (right column).
8. The Liaison (slur), a useful bit of shorthand, is used over a group of notes to indicate that they are to be held. (St.-Lambert also discusses this use of the slur.19)
François Couperin’s table in his first book of Pièces de clavecin (1713) differs markedly from those of his predecessors. Often unclear rhythmically, it casts doubt on the custom of beginning ornaments on the beat. This table is readily obtained in either the facsimile or the performing editions.20
D’Anglebert’s Cadence appears in J. S. Bach’s Clavier-Büchlein vor Wilhelm Friedemann Bach of 1720 (in its original form and also with a mordent at the close) for the first time since D’Anglebert’s 1689 edition. Bach’s table is clearly modeled on D’Anglebert’s and on other French sources (see Ex. 11).21
A Praeludium of disputed origin from this Clavier-Büchlein (No.29, BWV 931) uses D’Anglebert’s symbols for the arpeggio, trill appuyé, one-note grace, and trill with a turned ending. The wavy lines in front of the first beat in m.i and the third beat in mm.3 and 4 have been interpreted as arpeggio symbols.22 It seems unlikely, however, that both symbols for the arpeggio would be used in the same piece. A more probable explanation is that the symbol in front of the first chord denotes a figured arpeggio, and the symbol in mm.3 and 4 a slide (see Exx. 7 and 12).
The handling of D’Anglebert’s ornamentation symbols is skillful enough to be his own, but the style of the piece is not like any of his known works. Nor is it likely that the Praeludium was written by another French composer, for the French considered the prelude an improvisatory form. The brevity of No.29 implies that it was a didactic example and suggests that the prelude was composed by Bach himself, as a teaching piece in imitation of the French style. The fact that Bach copied de Grigny’s Livre d’orgue, some of Dieupart’s clavecin pieces, and D’Anglebert’s ornament table into a manuscript is further evidence of his interest in the French style.23
D’Anglebert’s Autograph Manuscript
and 1689 Edition Compared
Some clarification of details occurred between D’Anglebert’s autograph manuscript (1660S–1670S) and his Pièces de clavecin (1689), but there is little change in the overall shape of the pieces. In contrast to the 29 ornaments found in the table of the edition, some basic symbols (primarily the trill, mordent, one-note grace, slide, and arpeggio) are used in the manuscript. The manuscript shows evidence of a transitional stage in the development of some symbols, particularly that of the arpeggio (see p. 64). Some repeated or conjunct notes are written equally in the manuscript but dotted in the print; e.g., all the eighth-note groups in the Gavotte in G major. The ornamental figure in the soprano voice in the Sarabande in D minor is written as an eighth and two sixteenth notes in the manuscript, but as a dotted eighth and two 32nd notes in the edition. The Gaillarde in G major (Ex. 13) contains an example of rhythmic alteration (in this and subsequent examples, a. indicates D’Anglebert’s manuscript, b. the edition).
D’Anglebert was careful to indicate the duration of notes in the edition in order to avoid misunderstanding. For example, the broken-chord figures (mm. 178–191, 218) of the Variations sur les folies d’Espagne are written as simple eighth notes in the manuscript, but their full duration is given in the edition. This is simply good harpsichord technique written out.
The only instance of a double dot occurs in the manuscript of his double of a Sarabande by Chambonnières (Jeunes zéphirs), an example that also contains a written-out five-note turn (Ex. 14). Perhaps the 32d note is intended to strike with the one-note grace in the upper voice (see p.85).
A vertical line is occasionally used in the preludes to mark a concurrence of the upper and lower voices. Ex. 15 shows one that occurs in a different location in the manuscript from that of the edition. In the edition, the D in the bass of the first concurrence comes before the Bߕ in the upper voice. D’Anglebert may have preferred the latter version because the ornament (a cadence) actually begins on the Bߕ, whereas in the manuscript the bass note occurs in the middle of this ornament. Might his cadence begin after the beat in some appropriate measured contexts?
D’Anglebert’s edition substitutes a trill and mordent symbol for a five-note turn (m.2) in the manuscript version of the Gavotte in G major. Conversely, his transcription of Lully’s Courante gives a turn (m.2) in the edition, but a trill and mordent in the manuscript. This Courante contains a passage with more difference than usual between the edition and the manuscript (see Ex. 16). Three-note slides appear in the manuscript (mm. 13, 15), and the first F in m.16 is a natural in both sources. This courante (m.4) and its double (m.13) furnish two more examples of five-note turns written out in the manuscript but indicated by symbol in the edition. Two five-note turns in the double are written out in both sources (mm.8 and 16). D’Anglebert’s transcription of Lully’s Les Songes agreables illustrates his differing conceptions between the time of the manuscript and that of the edition (Ex. 17).
An interesting melodic variation occurs in the Variations sur les folies d’Espagne. D’Anglebert omits a bar line in the manuscript (probably because of an ornamented note that would tie over the bar), but writes out a five-note turn in the edition (see Ex. 18). A melodic variant occurs in variation 21; each measure opens with alternating forms of the cadence in the edition, but with a sixteenth-note pattern beginning on the upper auxiliary (the same execution as the cadence beginning from above) in the manuscript. The rhythm of the closing figure of each measure is also different. The fact that D’Anglebert wrote 32nd notes in the manuscript and sixteenth notes in the edition implies that he changed his mind and meant the sixteenth notes to be interpreted as written (see Ex. 19).
In several instances a one-note grace, indicated by symbol in the manuscript, is written out in the edition (e.g., in Ex. 20). Examples in the edition of written-out one-note graces before the beat occur in the Gavotte in G major (m.5), the Passacaille (m. 2), and the Gaillarde in G major (Ex. 21).
D’Anglebert’s rendition of a one-note grace in the edition, with the nonharmonic tone sounding alone at the beginning, is at variance with that given in his ornament table (see Ex. 22, in which c. indicates my realization). The version in the edition is probably an exceptional realization that needed to be written out because it could not be accurately conveyed by a symbol. The performance of a double cheute (figured arpeggio) from the same piece is also at variance with the ornament table. The manuscript depicts a figured arpeggio with two one-note grace symbols, but the edition shows the graces written out first, followed by a simple arpeggio in a downward direction. This interpretation is stronger, for it introduces contrary motion between the upper voices and the bass. The use of eighth rather than sixteenth notes can probably be attributed to the longer duration of the chord as well as the movement in the bass voice (Ex. 23). The one-note grace in Ex. 24 is replaced with a different rhythm in the edition, perhaps to avoid crowding the arpeggio on the first beat.
D’Anglebert probably borrowed from the lutenists the oblique line that denotes the arpeggio. In the manuscript, the diagonal line is usually placed between the staves instead of on the note stem; e.g., the Double of the Gaillarde in C major (mm.5 and 6). Sometimes it is placed between note heads a third apart (Allemande du Vieux Gautier, m.2), giving it the same appearance as Chambonnières’s slide. D’Anglebert often uses his own conventional curved-line symbol for the slide in these same pieces, so perhaps he means arpeggiation and not a nonharmonic tone. An oblique line is placed between two notes a sixth apart in the Courante du Vieux Gautier (La petite bergere, m.15), two are found in one chord between note heads in the Ouuerture d’lsis (m.8), and one is through the note stem in the double of Couperin’s Allemande (m.17). The oblique line of the Gaillarde in G major (m.1) is placed between the staves in the manuscript, but on the right-hand note stem in the edition. The diagonal slash in all these locations, reflecting transitional stages of the arpeggio symbol, indicates the arpeggiation of chords or intervals.
A comparison of D’Anglebert’s manuscript and edition clarifies the performance of arpeggios, two types of turn, and one-note graces before and on the beat. His meticulous care with rhythmic values suggests that there is little basis for altering note values except for the application of notes inégales in appropriate locations.