Introduction
(I) Specified chromatic inflections in vocal sources
(II) The theorists' statements
(III) Instrumental tablatures

> (IV) Doubling the subtonic

(V) General conclusion
 

 
     first part
> second part
 
 
 

The problem of the false relation

After this overall description of the history of the doubled subtonic, we shall now take into consideration the specific problem of the simultaneous false relation (imperfect octave). To ascertain its practicability in the musical langage of that time, we can rely on three different methods: a) the study of vocal sources, b) the theorists' statements, and c) the evidence drawn from instrumental tablatures.
 

a) Vocal sources

As already explained, explicit leading note accidentals are not used in the Franco-Flemish polyphony of the first half of the sixteenth century. It follows that specified imperfect octaves cannot appear in the cadences of this repertoire.
   Things are somewhat different for Italian music, in which sharps tend to reappear from 1530-40. Nevertheless, it has not been possible to find any instance of doubled subtonic with explicit imperfect octave in this repertoire either. Now, if simultaneous false relations had been a characteristic feature of this music, copyists or publishers most probably would have marked a few of them, from time to time.
   The same applies to Franco-Flemish music at the time it passed to the explicit conception of chromatic inflections (around 1560-70). Admittedly, a few exceptions to this rule have been observed by Gustave Reese, namely in Jacob Vaet's works. But these are not conclusive: in this case, the false relations do not appear in the vocal parts printed during the composer's lifetime (1564), but only in an instrumental arrangement published fifteen years after Vaet's death (by Rühling, 1583).

[See REESE, Gustave, Music in the Renaissance, New York, Norton, 2/1959, p. 297; and STEINHARDT, Milton, "Vaet, Jacobus", in New Grove (1980), vol. 19, p. 486a. The critical notes of Vaet's Complete Works clearly show that the sharps entailing a false octave do not appear in any of the vocal sources; see STEINHARDT, M. (ed.), Jacobus Vaet, Sämtliche Werke, Graz-Wien, Akademische Druck- u. Verlagsanstalt, 1961-68 (=Denkmäler der Tonkunst in Österreich), vol. 98, p. 132.]
 

On the other hand, explicit simultaneous imperfect octaves can be found in the works of some famous English composers of the second half of the sixteenth century (Byrd, Tallis...). But the study of this repertoire is fraught with difficult problems, which will be dealt with in a later section of this article.

b) The theorists' statements

Most theoretical treatises of the late Middle Ages and Renaissance contain the famous rule prohibiting mi against fa in a perfect interval. This affects diminished fifths as well as imperfect octaves. But it is well known that diminished fifths are common practice in Renaissance music, notably between the upper voices of a II-I cadence (this question has already been mentioned in our section on instrumental tablatures).
   So it would be possible to understand the prohibition of mi against fa as being only a general rule allowing for exceptions, which could explain the case of the diminished fifth, and maybe of the imperfect octave as well. But some theorists give more precise statements concerning both kinds of dissonances, and tend to make a clear difference between them. Texts of particular interest on this topic may include:

a) Treatises prohibiting the imperfect octave in the specific context of a cadence with doubled subtonic. Theorists of this category include:

    (?) Ramos de Pareia (1482)
    Francisco Tovar (1510)
    Nicola Vicentino (1555)
    Vincentio Lusitano (1558)

[See -WOLF, Johannes (ed.), Musica Practica Bartolomei Rami de Pareia Bononiae (1482), Leipzig, Breitkopf & Härtel, 1901 (=Publikationen der Internationalen Musikgesellschaft, Beihefte, Heft II), p. 66.
-TOVAR, Francisco, Libro de musica practica, Barcelona, Johan Rosebach, 1510; facsimile: Libro de musica practica, por Francisco Tovar, Joyas bibliográficas, Madrid, 1976 (=Viejos libros de musica 6), lib. III, cap. 7.
-VICENTINO, Nicola, L'antica musica ridotta alla moderna prattica, Rome, 1555; facsimile (ed. E. E. Lowinsky), Kassel, Bärenreiter, 1959 (=Documenta musicologica, Erste Reihe, XVII); libro terzo, cap. XXXV, fol. 58r.
-LUSITANO, Vincentio, Introdutione facilissima, et novissima, di canto fermo, figurato, contraponto semplice, et inconcerto..., Venetia, Francesco Mareolini, 1558, f. 14v.]
 

b) Treatises prohibiting the imperfect octave, but allowing the use of the diminished fifth (under certain conditions):

    Juan Bermudo (1555)
    Tomas de Santa María (1565)

[See -BERMUDO, Juan, Comiença el libro llamado declaracion de instrumentos musicales..., Osuna, 1555; lib. IV, cap. 48-49; see also cap. 32.
-SANTA MARIA, Tomás de, Libro llamado Arte de tañer Fantasia, assi para Tecla como para Vihuela, y todo instrumento..., Valladolid, 1565; facsimile: Geneva, Minkoff Reprint, 1973; Parte I, cap. 11, fol. 28v, 29v.]
 

c) Treatises asserting that the imperfect octave is a much harsher interval than the diminished fifth.

    Johannes Tinctoris (1477)

[Liber de arte contrapuncti, lib. II, cap. xvii, see SEAY, A. (ed.), Johannis Tinctoris opera theoretica, Rome, American Institute of Musicology, 1975 (=CSM 22), t. 2, p. 104]
 

d) Treatises asserting that the imperfect octave is unbearable to musical ears, or even to people without any musical training.

    Pietro Aaron (1529)
    Ghiselin Danckerts (around 1550)

[See -AARON, Pietro, Toscanello in musica, Venise, Bernardino et Matheo de Vitali, 2/1529, lib. II/20 and Aggiunta, p. [6/9]
-LOCKWOOD, Lewis, "A dispute on accidentals in sixteenth-century Rome", in Analecta Musicologica, Studien zur italienisch-deutschen Musikgeschichte II, 1965, p. 31.]
 

The two earliest authors mentioned here (Tinctoris, 1477, and Ramos, 1482) surely deserve a special commentary, for they both forbid mi against fa for all perfect consonances, i.e. the fifth together with the octave (Tinctoris introduces an example including a diminished fifth as well as an imperfect octave, whereas Ramos mentions the unison, and only implies the other "perfect consonances").
   Moreover, Tinctoris tells us that "almost all composers" make use of these types of dissonances, in spite of their being forbidden. Actually, this last statement could be understood as an allusion to (and a justification of) the imperfect octave in the special case of a cadence with doubled subtonic. But by the time Tinctoris was writing (1477), the doubling of the subtonic (in modes with a subtonium) must still have been extremely rare (if it even existed): we have seen that Ockeghem (c1420-1497) was apparently the first to introduce this kind of formula, which did not become common practice previous to the following generation of Franco-Flemish composers (Obrecht, Josquin, Isaac). Even then, one still can find important composers who seem unaware of this contrapuntal technique, for example De la Rue (c1460-1518), Pipelare (c1450-c1515), Compère (1445-1518). So it is hardly plausible that Tinctoris is here alluding specifically to the imperfect octaves which would appear in the case of a cadence with doubled subtonic. Hence this passage seems only to attest the usage of the diminished fifth by the composers of Tinctoris' time (the latter mainly quotes from Caron, Busnois and Faugues, who do not use cadences with doubled subtonium); but its ambiguity would arise from the fact that Tinctoris rejects the false fifth and the false octave alike, without making any distinction between them.

Summing up, the theoretical literature from the late fifteenth and the first half of the sixteenth century does not seem to adopt the imperfect octave as an interval to be used at the cadence. In this respect, things apparently would not change before the late sixteenth (Thomas Morley, 1597, who attests the dissonant cadence without approving of it) and early seventeenth century (Francisco Correa de Arauxo, 1626).

[See REESE, Gustave, Music in the Renaissance, New York, Norton, 2/1959, p. 824, note 34; HUDSON, Barton, "Correa de Arauxo, Francisco", in New Grove (1980), vol. 4, p. 799b-800a.]
 

c) Tablatures

Instrumental tablatures also provide us with valuable information on the topic. They can be classified into three different categories:

a) Those which almost completely avoid imperfect octaves (but include diminished fifths at the cadence).
b) Those which use imperfect octaves only for shorter durations (possibly with very rare exceptions), and without special connection with cadential contexts.
c) Those which make use of longer imperfect octaves, sometimes (but not necessarily) in the context of a cadence.

The following table displays all three categories:
 

Without imperfect octaves	With shorter imperfect octaves	With longer imperfect octaves
Dalza	Spinacino	Valentin Bakfark
Vincenzo Capirola	Attaingnant/Blondeau	Diego Pisador
Francesco da Milano	Le Roy	Fridolin Sicher
Luys de Milán	Morlaye	Leonhard Kleber
Luys de Narváez	Albert de Rippe
Alonso Mudarra	Jean Belin
Valderrábano	Judenkünig
Juan Bermudo	Gintzler
Cabezón	H. Newsidler
Kotter	Ochsenkun
Jean de Lublin	Hans Buchner (2 exceptions)
Clemens Hör

The third and last column is obviously the most interesting one for our enquiry. One will observe that it is quantitatively the least significant of the three categories. Moreover, Kleber's tablature has only four cases of sustained imperfect octaves (for examples by this composer, as well as by the others mentioned in this paragraph, see the previous section of the present article, "The instrumental tablatures"). Bakfark's has nine, but all of them appear in one and the same piece, which is an arrangement of a Franco-Flemish work with six voices (by Loyset Pieton); besides, in this special case, the notation technique of the Italian lute tablature does not enable us to ascertain the exact duration of the notes, so that it is not absolutely certain that the false relation is a simultaneous one here, rather than a successive cross relation.
   Thus, among the composers who most clearly make use of the imperfect octave, Fridolin Sicher (about twenty cases, not necessarily connected with cadences) and Diego Pisador need to be mentioned above all. One should remember that the latter's artistic status is not altogether easy to assess; he was probably an amateur musician, maybe the one Bermudo castigates (without mentioning his name) for the awkwardness of his instrumental arrangements, which are being mocked by professional singers. As for Sicher, his tablature may also show signs of incompleteness, as already seen (in section 3, "The instrumental tablatures"). At any rate, it cannot be considered as representative of the entire instrumental literature of its time.

As one can see, the large majority of tablature composers of the Renaissance tend to be very cautious in their handling of the simultaneous false octave. Now if the latter were a common-usage interval in Franco-Flemish vocal polyphony (more especially in the case of cadences with doubled subtonic), one could expect to find more freedom in this respect within contemporary instrumental sources as well.
 

English music

Vertical false relations do not seem to be attested earlier than in the English music of the second half of the sixteenth century. The earliest examples which can be dated with some accuracy appear in the first printed volume of some importance published in London, i.e. Byrd's and Tallis' Cantiones quae ab argumento sacrae vocantur of 1575.
 
 

ex.: T. Tallis, When Jesus went into Simon the Pharisee's house

 
[See ELLINWOOD, L. (ed.), Thomas Tallis : English sacred music : I, Anthems, London, Stainer and Bell, 1973 (=Early English Church Music 12), p. 72/41. Other similar examples: ibid., p. 71/29, 71/36, 72/40.]
 

But by as late a date as 1575, doubled subtonics have already completely disappeared from continental music; hence examples similar to the Tallis quotation above can only be considered as a specific characteristic of English music alone.
   It seems more interesting to try to find earlier examples. But here one has to face the difficult problem of the sources, which are in a poor state for the whole period previous to 1575, mainly on account of the religious conflicts of mid-century.

Up to the early years of the sixteenth century, the main sources of English polyphonic music (Eton Choirbook, Caius Choirbook, Lambeth Choirbook...) comprise works in a style quite distinct from continental music. The typical cadence formulas of the contemporary Flemish or Italian schools (i.e. the progression from the sixth to the octave, with suspension of the upper voice) are generally absent from this repertoire. Even when they do appear in it, the attraction principle is not explicitly applied to them, and is even found to be contradicted, albeit in rare instances:
 
 

ex.: Richard Davy, Stabat mater dolorosa (Eton Choirbook)

 
[See HARRISON, F. Ll. (ed.), The Eton Choirbook, London, Stainer and Bell, 3 vol., 1956-61 (=Musica Britannica X-XII), vol. 2, p. 89/140.]
 

Such an example could remind one of passages drawn from certain contemporary German organ tablatures (for example Buchner's), some of which have been quoted and discussed in the previous section of the present article.
   So it could be that attraction is not a component part of the English repertoire of the late fifteenth and early sixteenth century. From the 1530s, however, some of the English composers start introducing in their works cadence formulas which are quite similar to those already in use by continental musicians. One should name here such figures as Taverner, Tye or Shepphard. But it is very difficult to know whether the attraction principle already applies to the music of that period. Indeed, works belonging to this mid-century repertoire have been transmitted to us mainly through manuscripts which are of a much later date, and whose version of the music could reflect a more modern conception of it than that of the composers themselves:
 

"But whilst acknowledging our indebtedness to this and similar sources, we should note that they convey to us music from around the mid-century seen (with regard to such matters as accidentals) through the eyes of a later generation [...]."

[See CHADD, D. (ed.), John Sheppard: I, Responsorial music, London, Stainer and Bell, 1977 (=Early English Church Music 17), p. x.]

There remain, however, a few contemporary manuscript sources, which are of particular importance for our research:

    Forrest-Heyther partbooks: early 1530s.
    Henrican partbooks: around 1539-41.
    British Library, Ms. Roy.Add. 74-7: around 1547.
    Gyffard partbooks: between around 1540 and 1580, but mainly 1553-58.

In this case as well, chromatic inflections can be shown to be additions by later hands, and it is very difficult (if at all possible) to uncover the original version.
   In these mid-century sources, indeed, simultaneous false relations will be found in not insignificant numbers. Some of them correspond to the case under consideration:
 
 

ex.: Thomas Tallis, Benedictus (Roy.Add. 74-7)

 

[See Early English Church Music 13, p. 112/80.]

But a still greater part bear no resemblance to the subject of our inquiry. Many of them even seem to be quite absurd; at any rate, it is hardly to be expected that they could match the composer's intentions:
 
 

ex.: Christopher Tye, Alleluya. Per te Dei genetrix (Gyffard Partbooks)

[See Early English Church Music 33, p. 5/13.]

Problems of this kind have already been emphasized by several specialists of this repertoire, who even give up any attempt at reconstructing what could be called an original version.

[See BENHAM, H. (ed.), John Taverner : I, Six-part masses, London, Stainer and Bell, 1978 (=Early English Church Music 20), p. xvi; BENHAM, H. (ed.), John Taverner : V, Five-part masses, London, Stainer and Bell, 1990 (=Early English Church Music 36), p. 134; CHADD, D. (ed.), John Sheppard : I, Responsorial music, London, Stainer and Bell, 1977 (=Early English Church Music 17), p. x, xv; SANDON, N. (ed.), John Sheppard : II, Masses, London, Stainer and Bell, 1976 (=Early English Church Music 18), p. xi; MOREHEN, J. (ed.), Christopher Tye : I, English sacred music, London, Stainer and Bell, 1977 (=Early English Church Music 19), p. xi-xii.]
 

In short, it does not seem possible to define more precisely the date of appearance of the dissonant cadence typical of Byrd, Tallis and Morley. It certainly does not antedate 1530; and even if one admits that the false relation appears in British music from the mid-sixteenth century (which strictly speaking does not seem to be demonstrable), this kind of dissonance applies to a variety of situations, where it certainly would not have been allowed by Flemish or Italian musicians. One only needs to reread the declarations of Aaron or (better yet) Danckerts, who distinctly states that the false octave is an interval which hurts the ears even of uneducated people (for references see above).

In other words, it seems that the case of English music is not sufficient to justify the introduction of false octaves in the cadences of continental music: it is as if its testimony comes too late for that.
 

Conclusion

So the imperfect octave apparently does not belong in the Franco-Flemish music of that time, and the doubling of the seventh degree forces to admit of modal cadences without leading note.
   But it remains to be seen whether this is valid only for the very cadences with a doubled subtonic, or for all cadences of this repertoire. Two hypotheses are possible:

1. The doubling of the subtonic is a technique which is consciously used by the composer to force a modal cadence; in consequence, all other cadences should comply with the rules of attraction (this is Lowinsky's hypothesis).
2. All cadences, not only those with a doubled subtonic, should be left modal; thus the doubled subtonic would only be the visible part of a more general problem.

In the former case, one could expect that the modal form of the cadence, with its austere, almost ascetic, but grandiose colouring, would be reserved for sacred music as a priority. On the other hand, there should be no palpable difference between pieces with three, four, five or more voices. But as we have seen, statistical studies clearly tend to refute both suggestions.

The second supposition is thus to be considered in detail. As is well known, the vast majority of cadences in Renaissance music are based on the following pattern:

Additional voices may come in, but will remain secondary, and cannot change the nature of the cadence.
   This basic pattern is liable to be interpreted essentially in two ways: the former strictly modal, without any leading note; the latter with a leading note, and thus more similar to classical tonality:

If the composer decides on the second interpretation, he will have to be very cautious about the way he handles the leading note: should the latter be doubled, false relations might occur. On the other hand, in the case of the second interpretation, he may double the subtonic, and even present the doubling note in a melodic leap, an arpeggio, a fast run, etc. This is exactly the approach taken by Franco-Flemish composers from the age of Ockeghem and Josquin up to around 1560. If the doubled subtonic only appears with four, and more especially with five and six voices, the reason lies in the fact that it is not common practice to double the third of a triad with only three contrapuntal parts (or even with four).

The advantages of this new hypothesis are as follows:

1. It accounts for the dependence of the doubled subtonic on the number of voices.
2. It accounts for its non-dependence on musical genre (sacred or secular).
3. It avoids most harmonic problems arising from the application of the attraction principle to the Franco-Flemish repertoire (especially for pieces with many voices).
4. It accounts for the differences found to exist between northern and southern theoreticians; more specifically, it explains Danckerts' defence of modal purity, and the surprising silence of German theorists.
5. It explains the appearance of modal features in a group of German tablatures of the first half of the sixteenth century (Buchner, Gerle*, Hör, maybe Sicher and Kleber as well).

[*The case of Gerle does not appear as such in our study, but it has been examined by Robert Toft. See TOFT, Robert, Aural Images of lost Traditions. Sharps and Flats in the Sixteenth Century, Toronto-Buffalo-London, University of Toronto Press, 1992, 199 p.; p. 95-102.]

6. It fits perfectly into a more general picture of the period, which experienced the renewal of church modes, as illustrated for example by the theoretical works of Tinctoris and Glarean, or by the contemporary appearance of titles such as "primi toni", "quarti toni" or even "cuiusvis toni" (Ockeghem). It also seems to correspond to more general changes in the very conception of musical sound and musical taste, as Willi Apel suggests:
 

"The contrast between the Burgundian School (Dufay) and the first Flemish masters (Ockeghem, Obrecht) can be briefly characterized as the change from three-part writing to four-part writing; from a relatively high range to a considerably lower range (first appearance of the bass); from a medieval timbre (Sound-ideal) of "instrumental transparence" to a full vocal sonority, probably a-cappella; from fauxbourdon to the full triad; from a (decorated) chordal style, frequently of the melody-accompaniment type, to a truly polyphonic style with highly embroidered lines in all the parts; from aristocratic subtleness and refinement to pious devotion and mystic expression."

[See APEL, Willi, "Flemish School", in Harvard Dictionary of Music, Cambridge, Mass., Harvard University Press, 1966, p. 268-270; p. 269.]
 
 

Continued: (V) General conclusion >>>