Early Music and Historically Informed Performance Practice

Rhythmic alteration and articulation in 18th-century French flute music: a reappraisal of Jacques Hotteterre le romain

flute player (probably Hotteterre)


The practice of playing the rhythm of notes differently from how they are notated is one of the most controversial areas of Baroque performance practice.

This case study is about the French convention of notes inégales (unequal notes).

After giving some context and background information, I look in particular at what Jacques Hotteterre has to say in his treatises and what we can learn from his published music.

I also look at Hotteterre’s articulation syllables (tu and ru) and how they apply to different degrees of inequality.

Author: Greg Dikmans

Historical Performance Practice

During the 1980s, numerous writings appeared which looked critically at the state of research into what has been termed historical performance practice.[1] A number of authors are critical of the ‘positivist’ or ‘objectivist’ goal of amassing more and more facts. This approach was aimed at providing the ‘scientific’ support for a ‘code of performance’ of universal applicability ‘into which the score need only be plugged to set the music aglow with authenticity’.[2] Laurence Dreyfus believes that, by itself, this empiricist methodology is unhelpful in the study of performance practice, and that a new approach should acknowledge that ‘historical performance…[is] an evolving and necessarily incomplete paradigm rather than a set of documented index cards set atop inferences culled from freshman logic texts’.[3] Richard Taruskin is ‘convinced that ‘historical’ performance today is not really historical; that a thin veneer of historicism clothes a performance style that is completely of our own time, and is in fact the most modem style around.’[4]

Nevertheless Taruskin sees value for the modern musician in the study of historical evidence: ‘Really talented performers are always curious, and curious performers will always find what they need in the sources and theorists—what they need being ways of enriching and enlivening what they do.’[5] ‘It is not the elimination of personal choice from performance that real artists desire, but its improvement and refreshment. And for this purpose original instruments, historical treatises, and all the rest have proven their value.’[6]

Jacques Hotteterre le romain

Jacques Hotteterre le romain (1674–1763) was a composer, theorist, teacher, performer and instrument maker. Happily, much of the relevant source material relating to these activities has survived. Of particular importance to the present study are his three didactic treatises:

  • Principes de la flûte-traversière (1707),[7]
  • L’Art de preluder (1719)[8] and
  • Methode pour la musette (1737).[9]

Principes is the best known and most influential of the three, as the flute treatise (which occupies 34 of the 52 pages of text) was the first set of instructions ever published specifically for the transverse flute.[10] In addition to explanations of the technicalities of flute playing (holding the instrument, posture, embouchure and fingerings), it includes important information about the more general aspects of performance practice, such as articulation, rhythmic alteration and the agréments that Hotteterre considered ‘necessary for playing correctly and with taste’.[11]

The instructions contained within the pedagogical works can be supplemented by an examination of his published music, in particular his first book of Piéces (Op. 2) for flute and basso continuo.[12] This was published the year after Principes and includes an Avertissement that expands on material discussed in Principes.

In 1715, when Hotteterre published his second book of Piéces (Op. 5),[13] he reissued Op. 2 in a newly engraved edition.[14] This nouvelle edition has a table of agréments added to the Avertissement and contains a large number of changes to the score of the 1708 edition, especially additional agréments and other clarifications of the original text.[15]

Hotteterre’s instructions relating to rhythmic alteration are brief and quite simple, and should cause us no problems if we are willing to accept the simple meaning of his French.[16] Unfortunately, he does not answer every question the modern performer needs to ask, and it is beyond the scope of this study to offer definitive solutions (if this is at all possible). I will, however, briefly address some of them by drawing on other relevant primary sources, as well as offering my own suggestions and those found in the secondary source material. I will also suggest how this information may help us come closer to an understanding of the tasteful application of Hotteterre’s principles of articulation.

Rhythmic Alteration

Rhythmic alteration has become one of the most controversial areas of Baroque performance practice. In its broadest sense it includes all the expressive fluctuations of tempo and rhythm which Riemann called agogics.[17] However, the issues which concern us here are more specific.[18] These are:

  1. the application of the French convention of inequality or notes inégales, especially when the composer does not indicate its use either by an instruction, such as croche inégales, or by writing it out in dotted rhythms;
  2. the degree of inequality when its use is appropriate (whether indicated or not);
  3. inconsistency of notation: for example, when dotted rhythms are written, whether this is an indication of inequality or double dotting;
  4. inequality on disjunct intervals;
  5. inequality with slurs.

In practice these issues are closely related, or even linked.[19]

Inequality is usually defined as the uneven performance of evenly written values.[20] The following definition by David Fuller provides a useful background to the issues discussed in this study:

As it existed in France from the mid-16th century to the late 18th the convention of notes inégales was first of all a way of gracing or enlivening passage-work, especially improvised diminutions. As styles changed and the figurations born of diminution entered the essential melodic vocabulary, inequality permeated the musical language. Its application was regulated by metre and note values; it always operated within the beat, never distorting the beat itself.

The degree of inequality (i.e. the ratio between the lengths of the long and short notes of each pair) could vary from the barely perceptible to the equivalent of double dotting, according to the character of the piece and the taste of the performer.

Inequality was considered one of the chief resources of expression, and it varied according to expressive needs within the same piece or even within the same passage; where it was felt to be inappropriate it could be abandoned altogether unless explicitly demanded.

Although the practical problem is certainly that of deciding when to alter what appears on the page, the rhythmic convention itself is independent of questions of notation. French composers frequently wrote out inequality with dotted figures, sometimes to resolve doubt, sometimes to ensure a sharply dotted effect, and sometimes for no apparent reason.[21]

Hotteterre on notes inégales

Hotteterre first mentions inequality during his discussion of the articulation of quavers in Chapter 8 of Principes. He writes:

You would do well to observe that you must not always play the quavers equally, [but] that, in certain time signatures, you make one long and one short, which is also regulated by their number.

When [the number of quavers] is even, you make the first [quaver] long, the second short, and so on for the others.

When it is odd,[22] you do exactly the opposite: this [practice] is called dotting (pointer).

The time signatures in which this is most usually done are 2 time, simple 3 time and 6/4 time.[23]

In other words, the metrically strong quaver is long, and the metrically weak quaver is short. Hotteterre calls this practice pointer. Literally this means ‘to dot’, that is, to add a dot to a note.[24] So pointer can be translated as ‘dotting’ (noun) or ‘to dot’ (verb). He also uses the adjectival form pointée, as in croche pointée (dotted quaver).

L’Art de preluder includes a chapter which explains the various time signatures, including instructions about which notes are equal and which are dotted in each. For example, in discussing slow common time (Mesure à 4 temps lents), Hotteterre says: ‘the quavers here are equal and…the semiquavers here are dotted, that is to say, one long and one short’.[25] In cut common time (Mesure du C barré) he says: ‘the quavers here must be equal…unless the composer puts dots’.[26] Again this implies that inequality is the result of adding dots.

Lastly, in Methode pour la musette Hotteterre includes a section on note values. After explaining the relative values of the notes, he goes on to describe two meanings of the dot found after a note:

[1] You sometimes put dots after the notes, which augment them by half of their value: thus a dotted minim is worth three crotchets; a dotted crotchet is worth three quavers, etc.

[2] In time signatures where the quavers are unequal, the dot which is after the crotchet makes the equivalent of a dotted quaver; the sort of quaver which follows [such] a dotted crotchet is always short.[27]

In the first case, the dots (whether written or added by the performer) behave as we would expect. Note that Hotteterre says the dots ‘you put’ or ‘one puts’ (on met). It is not clear whether he means the dots the performer adds when playing pointer, or the dots added by the composer. He probably means both.

In the second case, the written dot added to a crotchet does not behave as we would expect. This is in fact a description of double dotting which can be notated as in Example 1.[28] Note that Hotteterre is specific about the effect the dot has on the length of a note (in the first case) and about the length of the dot itself (in the second case), but he is not specific about the quaver following the dotted crotchet (second case). He does not say it becomes a semiquaver, only that it is ‘always short’. This raises further questions.

Perhaps he did not need to be specific about the quaver becoming a semiquaver, because he thought it was obvious that this was the case. Or perhaps he left it open, because the short quaver is not always exactly a semiquaver.

Example 1: Hotteterre’s description of double dotting.
musical example 1

An examination of Hotteterre’s music

The above is all Hotteterre has to say about rhythmic alteration in his didactic treatises. More information can be deduced by examining his published music. A comparison of the two editions of his Pièces (Op. 2) reveals a number of differences. Hotteterre obviously took the opportunity, in preparing the engraved 1715 edition, to clarify certain points.

Example 2: Hotteterre, Allemande La Royalle (Gracieusement) from Pièces (Op. 2) 1708, bars 1–2.
musical example 2
Example 3: Hotteterre, Allemande La Royalle (Gracieusement) from Pièces (Op.2) 1715, bars 1–2.
musical example 3

Examples 2 and 3 bear directly on the present discussion. The time signature is 2, in which, according to Hotteterre, the quavers are normally played pointer. In the 1708 edition, he has added dots (and shortened the following quaver to a semiquaver) in the bass line in the first bar, and in the 1715 edition he has done the same in both bass and flute lines. So the quavers, written equally in the flute line of the first edition, are played dotted (pointer) as shown in the 1715 edition. Hotteterre does not continue this in the following bars—the first bar is sufficient reminder of what he wants.

As we have seen, Hotteterre uses pointer to describe both a note value and the convention of notes inégales. He does not make a distinction, so it is probable that there is none: a dot is a dot. If there were an important distinction (for example, that the dot means something different in each case), we could reasonably expect that he would mention it in his books intended for beginners.

Hotteterre does in fact attempt to clarify two particular cases. When using the term to describe a note value, such as a dotted crotchet, all such notes will have a dot; however, when using the term to describe the practice of inequality, only every second note will have a dot. Hotteterre clarifies this latter situation by adding that ‘one is long and one is short’. Note that this is a clarification of application and does not imply that the dot is different in each context. The second clarification is the particular case, found in the passage describing double dotting, of the dot after a crotchet in time signatures where the quavers are unequal. In this case the dot does behave differently, so Hotteterre explains the difference (see Example 1).

From the above evidence we can conclude that for Hotteterre a dot after a note, whether written or added by the performer, creates a degree of inequality between the lengths of the long and short notes of each pair that can be expressed as a ratio of 3:1 (3 semiquavers to 1 semiquaver in a pair of quavers—see Examples 2 and 3). In the case of double dotting, the ratio is 7:1. This also has the effect of making the short quavers (the semiquavers) synchronise.[29] So a passage of music which had one line moving in equally notated quavers with another line moving in dotted crotchet-quaver rhythm can be written, following Hotteterre’s explanations, as in Example 4.[30] Example 5 shows how Hotteterre could have continued the passage quoted in Example 3 to indicate the inequality and synchronisation of semiquavers.

Example 4: Synchronisation of notes inégales following Hotteterre’s explanations.
musical example 4
Example 5: Hotteterre, Allemande La Royalle from Pièces (Op.2) 1715, bars 1–4.
musical example 5

Varying the degree of inequality

Hotteterre does not discuss the possibility of varying the degree of inequality. The only passage which may allow for the possibility of a different ratio between the long and short notes is the one describing double dotting, in which Hotteterre does not specify the note value of the quaver following a dotted crotchet, saying only that it is ‘always short’. However, Fuller’s aforementioned definition of inequality states that ‘the degree of inequality…could vary from the barely perceptible to the equivalent of double dotting, according to the character of the piece and the taste of the performer’. Therefore we need to examine the relevance of this statement to Hotteterre.

Étienne Loulié (1654–1702) includes an often quoted discussion of even and uneven notes in his Elements ou principes de musique (1696).[31] He describes four manners of performance. The first manner concerns disjunct intervals and will be discussed below. The fourth manner, a short-long performance notated as a 1:3 ratio in his example, is not relevant to Hotteterre. The second and third manners, however, bear directly on the present discussion:

Sometimes the first half-beat [quaver] is made a little longer than the second. This manner is called lourer; it is used in melodies in which the sounds follow each other in conjunct motion… There is yet a third manner, in which the first half-beat is made much longer than the second, but then the first half-beat should have a dot. This third manner is called piquer or pointer.[32]

In his unpublished Supplément des Principes ou Elémens de Musique,[33] Loulié includes Example 6 with the text just quoted. His description of the third manner and the example (Piqué in Ex. 6) are in accord with Hotteterre and the 3:1 ratio. Loulié’s description of the second manner (Louré in Ex. 6) indicates a milder ratio of 3:2.

Example 6: Loulié, Supplément des Principes ou Elémens de Musique
musical example 6

In arguing against the 3:1 ratio for notes inégales, Frederick Neumann has used this passage to suggest that ‘clearly only the lourer with its mild unevenness represents notes inégales, since the concept of inégalité refers solely to evenly written pairs of notes’.[34] John O’Donnell answers this argument conclusively: ‘[b]ut this is impossible, since the literal meaning of lourer is ‘to slur’, and no contemporary writer limits inequality to slurred notes. It was unnecessary for Loulié to define lourer, for he would simply have been saying ‘to slur’ means ‘to slur’.[35]

So Loulié supports the more vigorous 3:1 ratio for pointer, which is tongued, and allows a milder ratio of 3:2 for slurred pairs. In light of Hotteterre’s vagueness about the length of the quaver following a double dotted crotchet, it is possible that he might also have used the milder inequality when tonguing, if the character of the piece and taste so dictated.

Betty Bang Mather quotes Marie Dominique Joseph Engramelle’s La tonotechnie, ou l’art de noter les cylindres (1775) as allowing

a possible ratio of 3:1 in his text, but his examples showed only the less unequal ratios of 2:1, 3:2 (the most frequently used), 5:3, 7:5 and 9:7. The latter three relationships may be perceived simply as an agogic accent to our ears, and that is their function and effect.[36]

But agogic accent is not inequality as defined by Hotteterre and Loulié. The confusion of the two has no doubt led to the ‘generalized rubato’ and ‘audible analysis’ which Fuller says is ‘practiced by modern players of Baroque music [but] which has nothing to do with notes inégales and cannot be justified as a rendition of it or a substitute for it’.[37]

The notation (or not) of inequality

In his definition of notes inégales Fuller also states that this ‘rhythmic convention…is independent of questions of notation’, and elsewhere that ‘there was no clear boundary between dotted rhythms, whether exaggerated or not, and notes inégales’.[38] Hotteterre sometimes writes the dots and sometimes does not, often within the same piece, and often on notes which would normally be played pointer anyway. There seems to be no consistent logic in his notation.

The problem, then, is to decide whether the notated dotting

  • is a reminder to make the rest of the short note values unequal (as I have suggested is the case in the Allemande La Royalle);
  • indicates a contrast of mild and vigorous inequality;
  • or whether, following Graham Pont’s ‘paradigm of inconsistency’, the undotted notes are to be played as written.

In the absence of any instructions from Hotteterre we will have to judge each case on its context and let taste decide. I do believe, however, that the last possibility is least likely because Hotteterre is so inconsistent that it would be impossible to establish any coherent character in a piece if the notated rhythms were followed exactly.

Unequal (or not) disjunct intervals

The unequal performance of disjunct intervals is possible, as shown in the Allemande La Royalle. Hotteterre sometimes indicates the equal performance of disjunct melodies, or parts of melodies, by instructions in the score such as croches égales (equal crotchets) or piqué (pricked), or by adding dots above the notes. The problem is knowing his intention when he does not give an indication.

Some theorists did associate inequality with conjunct motion and equality with disjunct motion in those situations where inequality would normally be the case. For example, Loulié explains his first manner of performing the half-beats (mentioned above) as follows:

They are sometimes made equal. This manner is called “to detach the notes” (détacher les Nottes); it is used in melodies in which the sounds follow each other in disjunct motion, and in all sorts of foreign music, where one never alters them rhythmically except where indicated.[39] (Example 7)

Example 7: Loulié, Elements ou principes de musique (1696).
musical example 7

Hotteterre makes no specific reference to this practice in Principes, but there is one reference in L’Art de preluder. In his discussion of simple 3 time, where the quavers would normally be unequal, he includes an example from the beginning of a couplet in a passacaille by Lully together with a brief explanation:

[Here is an] example in this same time signature with equal quavers… The quavers are made equal for this reason: it is firstly that they leap by intervals, and over and above this, that they are mixed with semiquavers.[40] (Example 8)

Example 8: Hotteterre, Couplet de la Passacaille d’Armide in L’Art de preluder, p. 58
musical example 8

This is too brief a reference to allow us to take it as a rule with general applicability. Also the language used implies that the more compelling indication of equality in this example is that the quavers are mixed with semiquavers.

Hotteterre often shifts between conjunct and disjunct motion within the same piece, where there are no smaller note values mixed in with the normally unequal notes. This may or may not indicate that the equality and inequality should also shift. For example, a melody which moves in disjunct quavers against a bass moving in conjunct quavers probably implies inequality in both. Again each case needs to be judged on its context with taste being the arbiter.

Unequal slurred pairs

The practice of performing slurred pairs of notes unequally, which Loulié and others called lourer, has already been mentioned.[41] Confirmation of this practice can be found in Hotteterre’s music by comparing the two versions of the Sarabande La Fidelle from Op. 2 (Examples 9 and 10). In the 1708 edition the instruction is to play Coulé & Pointé (slurred and dotted) and in the 1715 edition it is Croches inégales et coulées (quavers unequal and slurred).[42] Hotteterre has also added the slurs above the quavers in the first bar of the later edition, thus pairing them. As mentioned earlier, Hotteterre does not discuss a degree of inequality other than 3:1; either this or possibly Loulié’s 3:2 would be appropriate in performing this pièce.

Example 9: Hotteterre, Sarabande La Fidelle (Coulé & Pointé) from Pièces (Op. 2) 1708, bars 1–2.
musical example 9
Example 10: Hotteterre, Sarabande La Fidelle (Croches inegales et coulées) from Pièces (Op. 2) 1715, bars 1–2.
musical example 10

Inequality with long slurs

Hotteterre does not discuss the effect of slurs over more than two notes; neither do other French sources of the time. However, the issue has been taken up in the modern literature. Fuller is correctly cautious when he states that ‘the long slur seems at times to be intended to cancel pairing and to suggest to the player that only the first note should be emphasized’,[43] whereas Mather is unequivocal: ‘[s]lurs over more than a pair of notes also signalled equality’.[44]

The evidence provided by the music is also unclear. In the Musette from his second suite for two flutes Op. 6, Hotteterre slurs the quavers in groups of four and gives the instruction to play Doucement, et les croches pointées (sweetly or softly, and the quavers dotted). Unfortunately, it is not clear whether Hotteterre is reminding the player of the normal practice or whether he is asking the player to employ a special practice in this specific case (Example 11).

Example 11: Hotteterre, Musette (Doucement, et les croches pointées) from Deuxième suitte de pièces a deux dessus… Oeuvre VIème (1717).
musical example 11

Inequality and articulation syllables

Various authors have put forward the view that there is a relationship between Hotteterre’s articulation syllables (tu ru) and inequality.[45] Others have gone a step further and argued that the use of tu ru causes inequality.[46] However, Fuller argues against this view, saying:

there is no necessary connection between playing technique and rhythmic unevenness unless the tempo is very fast or the player unskilled… Inequality was essentially a matter of rhythmic style, not performing technique, and it obeyed laws of metre and expression, not of fingers or tongue.[47]

Fuller’s case is supported by the evidence in Principes.

Hotteterre uses tu ru in both its normal and reversed forms for the articulation of unequal notes.[48] What we can say is that the more vigorous type of inequality (including double dotting) is more easily expressed with the normal form of tu ru. This is not so much because of the speed with which one can move the tongue from tu to ru, but because of the articulation silence which precedes tu, the length of which can be varied at will. Indeed, this type of inequality needs the articulation silence. So a figure such as (a) in Example 12 can be played as shown in (b) or even (c), which also creates an agogic accent. This type of phrasing is not possible if the reversed form of tu ru is used.

Example 12: Articulation silence with tu ru (normal form).
musical example 12

Hotteterre explains the use of the reversed form at the end of the section on articulation in Principes. He says that ‘although these rules [governing the normal form of tu ru] apply generally, there are nevertheless some exceptions in certain passages, as you can see here’.[49] (Example 13)

Example 13: The use of tu ru (reversed form) in exceptional cases.
musical example 13

These passages can be played just as easily with the normal form of tu ru; however, Hotteterre says that the reversed form (that is, without the articulation silence between the long and short notes) is used ‘for a greater softening or sweetening [of the articulation], and it is taste which decides it’.[50] The implication is that he equates softness or sweetness with a more legato effect. He then goes on to say:

you must consult this same taste when the tongue strokes seem rough doing them in the way that I have explained in the first examples, and you must decide upon that which seems most agreeable to the ear, without regard to the arrangement of the notes [that is, whether they are odd or even in number, and whether they move by leaps or conjunct intervals], nor the different passages.[51]

It is therefore reasonable to suggest that in pieces with indications such as gayment or majesteusement the more articulated normal form of tu ru is appropriate, while in pieces with indications such as tendrement or douloureusement, the more legato reversed form of tu ru is suitable.

It is also possible that the use of the reversed form of tu ru implies a milder degree of inequality, such a s Loulié’s lourer ratio of 3:2.


In this case study I have confined myself to a very specific problem: how one particular performer approached rhythmic alteration in the performance of his own music.[52] In so doing I hope also to have addressed the need for the specialised and critical study that should support a new approach to historical performance practice, an approach that exploits two modes of interpretation: restoration of meaning and demystification of meaning.

This is an updated, online version of an article published in Context: Journal of Music Research (Uni. of Melb.), No. 9, Winter 1995: pp. 32–42. It is based on research undertaken for my Masters thesis:

  • Greg Dikmans. ‘The performance practice of early 18th-century French flute music: a critical translation of Jacques Hotteterre’s Principes de la fûte traversière, commentary and recording of selected works.’ (MA thesis, La Trobe University, 1991).

© Greg Dikmans | updated November 2019