CHAPTER FOUR

 

TONAL STRUCTURE IN THE TROUBADOUR SONGS

   

            One of the difficulties faced by the student of troubadour and trouvère songs is the lack of a method of recognized validity for analyzing the music. The problem is especially acute when one tries to confront the issue of tonal structure in the songs, by which is meant the ordering or ranking of pitches according to function. A tonal structure can be simply a more or less consistent differentiation between pitches receiving greater prominence (structural tones) and those with less emphasis or importance (ornamental tones). Even at this lowest level, however, we are hampered by the uncertainties resulting from the transmission process. Without any indication of rhythmic values in the notation, one of the strongest potential clues to emphasis becomes a matter for conjecture. Even if one does distinguish structural from non‑structural tones in a given passage, on the strength of other criteria, one must confront the possibility that another version of the song from a different manuscript may display a rather different tonal structure, owing to alterations in the melodic line, in the distribution of pitches over syllables; inconsistencies in the notation of accidentals may also affect the tonal character of the song. Both performers and scribes may be responsible for variants, and one must question the respective roles of the two groups in determining the tonal characteristics of the songs as we now have them. For example, if the music scribes received their training entirely within the context of chant notation and theory, did they alter what may have been a different musical idiom with different tonal features to make it conform more closely to those of Gregorian chant?

            In spite of these uncertainties, two methods have been proposed for understanding the tonal structure of secular monophony, and these are discussed in the first part of the present chapter. However, neither method has been applied in any thoroughgoing way. The second part of the chapter therefore presents the results of a study of approximately half of all the surviving troubadour songs, in which both methods are taken into account. The findings of this investigation will provide a clearer picture of tonal structure in the repertoire, and will also show which aspects of the two approaches are valid or useful, and which are not.

            According to one view, the music of the troubadours and other secular composers shows the influence of the music of the church. Consequently, the modal system expounded by medieval theorists is the most appropriate tool of analysis, and it is also the only one contemporary with the sources. Scholars like Ian Parker and Theodore Karp have therefore found it appropriate to discuss troubadour and trouvère songs in terms of the Gregorian modes, while allowing for certain features that set the two repertoires apart.^[1] The other approach, which is not necessarily incompatible with Gregorian modality, has been associated with the name of the ethnomusicologist Curt Sachs, and seeks to understand the melodies in terms of prominent intervals, especially chains of thirds, which Sachs considered to be the organizing factors behind many popular and folk traditions. Hendrik van der Werf has argued for its applicability to the music of the troubadours and trouvères.^[2] These two approaches are discussed below, and following this, a significant cross‑section of the troubadour songs is examined with a view to determining what evidence of tonal structures or features can be found in the sources, especially from the perspectives of modality and interval chains.

            The system of eight modes and four finals, with their respective tenors and pitch ranges, was established in the traditional form now known to us by the eleventh century, although its history goes back several centuries, and can be traced to the Byzantine system of echoi.^[3] The western system was originally developed as a tool in the training of singers and as means of classifying the very large body of chants they were responsible for. The psalm tones had an especially important connection with modality, since one of the main purposes of the modal system was to determine the correct psalm tone for a given antiphon or responsory. In the ninth and tenth centuries, either the beginning of the antiphon or the final was used to determine the mode, but by the eleventh century the final was the main indicator.^[4] Theorists also adduced other factors in judging the mode of a piece, including range, the tenor or reciting‑note,^[5] and the pitches for initial and medial cadences, these latter criteria coming into greater use after the eleventh century. According to Harold Powers, these elements can be considered hierarchically according to function: "A four‑tiered system of modal pitch‑functions results: at the first level the final, at the second level the tenor, at the third level the other initial‑medial strong points, and at the lowest level the remaining degrees of the scale."^[6] In twentieth‑century discussions of mode, the term includes the concept of melody type as well as a set of intervals and pitches; it has been defined as "either a 'particularized scale' or a 'generalized tune', or both, depending on the particular musical and cultural context."^[7] To some extent, it can be said that the compilers of tonaries recognized this aspect in their classification of the variable psalm‑tone endings, or differentiae, according to general melodic features of antiphon beginnings.^[8]

            It must be remembered, however, that the needs and aims of the medieval clergy were not the same as those of modern scholars, and that the classification by range and final may not tell us very much about the tonal structure of a given piece. Modal ambiguity can arise in cases of a particularly narrow or wide ambitus, for example, and there are cases of pieces that are substantially identical ending on different finals in different manuscripts, and thus receiving a different modal designation.^[9] Even in the case of Gregorian antiphons, therefore (arguably the class of pieces most readily conforming to the terms of the system of modes), some scholars find it more useful, for analytical purposes, to look at other, internal features rather than the final for analytical purposes.^[10]

            Contemporary evidence for the understanding of secular melodies in modal terms is not very promising. The only medieval theorist to mention the question is Johannes de Grocheo, writing around 1300 in Paris. In two passages of his treatise he seems to deny the applicability  of the church modes to secular music:

 

              Describunt autem tonum quidam dicentes

              eum esse regulam, quae de omni cantu in

              fine iudicat. Sed isti videntur multipliciter

              peccare. Cum enim dicunt de omni cantu,

              videntur cantum civilem et mensuratum

              includere. Cantus autem iste per toni

              regulas forte non vadit nec per eas

              mensuratur. Et adhuc, si per eas

              mensuratur, non dicunt modum per quem

              nec de eo faciunt mentionem. Amplius

              autem, cum plures toni in fine conveniant,

              puta primus et secundus in d‑gravi,

              per hoc, quod dicunt in fine, non

              articulatam differentiam apponunt, nisi

              quis per hoc intellexerit principium

              et medium cum hoc esse. Amplius autem,

              cum dicunt iudicat, peccare videntur.

              Non enim regula iudicat, nisi quis

              metaphorice dicat. Sed est illud,

              mediante quo iudicat homo, quemadmodum

              instrumento mediante mechanicus operatur.^[11]

 

By cantus mensuratus, it is clear that Grocheo means vocal polyphony. In the passage immediately following, he mentions, as examples of cantus civilis, the ductia and stantipes; he may be referring to either vocal or instrumental forms of these dance genres, since both are discussed elsewhere in the treatise. The term cantus civilis is however Grocheo's general term for all the forms of secular music he discusses. Shortly after this statement, Grocheo again rejects the notion that the rules of the modal system may be applied to or recognized in secular monody and polyphony:

 

              Temptemus igitur aliter describere et

              dicamus, quod tonus est regula, per

              quam quis potest omnem cantum

              ecclesiasticum cognoscere et de eo

              iudicare inspiciendo ad initium, medium

              vel ad finem. Dico autem hic regula,

              per quam etc., quemadmodum in

              grammatica et in aliis artibus regulae

              inveniuntur generales propter cognitionem

              et facilem apprehensionem illorum, quae

              sub eis continentur. Dico etiam cantum

              ecclesiasticum, ut excludantur cantus

              publicus et praecise mensuratus, qui

              tonis non subiciuntur. Sed dico inspiciendo

              etc., quoniam per hoc toni ad invicem

              distinguuntur.^[12]

 

            At the very least, Grocheo's statements suggest that the tonal structures of secular songs differ sufficiently from those of liturgical chant that the rules of the latter, namely the system of eight modes, are not those of the former. Some scholars even question whether it is legitimate to look for any kind of tonal structure in these songs. John Stevens, for example, in the chapter on secular song in the new edition of The New Oxford History of Music, does not mention the modes at all, and seems skeptical about the existence of any kind of tonal plan in troubadour or trouvère songs.^[13] Those who do find it useful to refer to modal theory in this context are nevertheless forced to acknowledge the significant departures from strict modal theory and from chant practice in the secular realm. The main differences are aptly summarized by Theodore Karp: "Compared with the great melodic treasure of Gregorian chant, a wider variety of accidentals is employed [in troubadour and trouvère song], there is greater contrast between extremes of range, a larger number of ways in which the final may relate to the melodic ambitus, and a larger variety of finals."^[14] According to Karp, one can divide the troubadour and trouvère songs into groups based on the relation of the final to the tonality in the rest of the song. In one group, the final is the tonal centre of the song, well prepared and expected; in another group, the final seems to have no connection with what comes before; in a third group there is an oscillation between two tonal centres; and in the last group there occurs a shift from one centre to another during the course of the song. (No estimates are given regarding the relative sizes of the groups, or differences between troubadour and trouvère tonal practice.)^[15]

            Karp finds no necessary relation between characteristic melodic patterns or formulas and modal type, but it is just this kind of relation that is the subject of one of the few studies of troubadour or trouvère tonality of any extent, a 1958 monograph by Hans Zingerle.^[16] Part of the intent of this study was to investigate the historical transition from a modal to a major‑minor system of tonality, and along with this the development of a sense of the tonic as the final goal of a piece. Zingerle looked at patterns of disjunct intervals in melodic phrases to see whether certain ones occurred more often with some finals rather than others. The finals were grouped into those he considered as precursors of major tonality, F, G, and C, and those with a minor character, namely D, E, A, and G with b‑flat.^[17] Most of the phrases cited as examples are the final ones in the song, though earlier ones are sometimes used.^[18] The intervals considered, mainly fourths, fifths, and progressions containing thirds, may occur a few or several syllables before the final pitch of the phrase. For example, an ascending fourth from the subfinal was found to occur most frequently with D, A, and G (with b‑flat) finals and was therefore considered to be associated with minor tonality. Similar findings were reported for a series of intervallic patterns: ascending fourth from the final; ascending fourth to the final; ascending fourth to subfinal or third below the final; ascending fifth from the final; ascending fifth to the second above the final; descending fifth to the subfinal; descending fifth to the final, and so on. Another series of formulaic patterns involve thirds: two or more ascending or descending thirds in various positions relative to the final; single thirds directly approaching the final from above or below; combination of seconds and thirds. Not all these patterns point consistently to one or another group of final. Where a tendency is suggested, it is difficult to know how significant it is; Zingerle gives no figures or tables, nor does he tell us how many songs were examined, whether all phrases were equally weighted, how multiple versions were evaluated, etc.

            Even if we grant the validity of the findings, and the association between specific types of intervallic progressions and specific finals or tonalities, it is difficult to assess what significance this might have for the tonal structure of an entire song without conducting a further study.^[19] One would naturally like to know more about several questions Zingerle raises. Do the formulae occur most often in the final phrase, or in others as well?  If the latter is the case, how far should each phrase with a different final be considered to have its own tonality?  How often is the tonality found in the final phrase a feature of the song as a whole?

            It is suggestive, nonetheless, to find that chains of thirds make up one of the main classes of melodic formulae in this study, and also that the author divides the modes into major and minor, because both of these elements figure prominently in the other approach to tonal structure in the repertoire of secular monody. The main idea behind this approach is that interval chains, especially third chains, constitute the structural skeleton of a melody. Hendrick van der Werf is the main exponent of this idea as an aid to discerning the tonal organization of troubadour and trouvère song,^[20] but he is simply adapting the ideas of Curt Sachs, who seems to have been the originator of the method, discovered through his lifelong studies of musical cultures from every part of the world.^[21]

            Sachs describes very simple melodies, most of them belonging to primitive tribes, in terms of the intervals formed by their structural tones. Simplest of all are the generally solemn songs consisting almost entirely of one‑note recitations in the manner of psalmody. Then there are one‑step melodies that alternate between two notes which may be a second apart, or have larger intervals between their two main pitches. ("Affixes" or "infixes" may ornament the other notes, but, in his examples at least, it is quite clear which are the essential and which the non‑essential notes, by the vastly greater emphasis on the former.)  Two‑step melodies have three structural notes, three‑step melodies have four, and so on. Following is Sachs's example of a quadruple‑third melody ascribed to the Hottentots, but which Sachs considers especially prevalent in Europe:^[22]

 

       Example 43. From Sachs, 150.

   

Another kind of melody, which Sachs calls "centric," is distinguished by the continual return to a pitch in the middle of its range, which easily stands out from the others by the amount of musical time it occupies in the melody and by repetition.

            Because of the simplicity of the melodies discussed by Sachs, his descriptions amount to much more than what we usually mean by tonal structure, for they constitute a nearly complete definition of those melodies—all that is missing is the rhythmic features, and the particular succession of pitches employed. Needless to say, such simple and straightforward types of melodic construction are not found among the troubadour songs, so that if some of Sachs's types are to be sought there, they will not be anywhere near as obvious as they are in his tribal songs. In spite of the obscurity of Sachs's melodic structures in the troubadour and trouvère repertoires, van der Werf nevertheless believes that these structures account for the coherence and "memorizability" of the melodies.^[23] He has opted for Sachs's types in lieu of the modal scales, which he feels are inadequate to explain the tonal organization in secular songs, owing to inconsistencies between the theoretical system of modes and the preserved notations of the songs. These divergences mainly concern the notation of sharps and flats, which may vary in different manuscripts, altering the mode of a song; there are also many songs that van der Werf would classify as Ionian or Aeolian.^[24] Structures such as Sachs's interval chains are present in troubadour and trouvère songs, but it is rash to assume that they are so pronounced as to constitute the structural core or framework as represented by that which is common to multiple versions of the melodies. On the contrary, as van der Werf himself admits, even among melodies "that move freely up and down without any apparent structure and without any limitations other than those dictated by the confines of an average human voice . . . we find remarkable similarities among the preserved versions."^[25] Our ears may not be in a position to judge the coherence of the melodies, and medieval singers may have had no trouble remembering melodic details that appear featureless to some of us today.^[26]

            Let us then consider those of Sachs' melodic types that  van der Werf found most typical of troubadour and trouvère songs.^[27] Simplest is a one‑note melody in the style of recitation, with intonation and termination. Of course, no troubadour or trouvère melody conforms directly with such a basic pattern; at most, one or more lines in a song may feature the repetition of a pitch for a few syllables—from three to five on average.^[28] Van der Werf lists some typical examples of intonation‑type formulas leading to recitations on various pitches. To cite a few examples, recitations on d may be approached through formulas such as G b d; G b‑c d; b c d. Recitations on a have intonations such as D a; D F a; D‑E F‑G a; F‑G a. Intonations for F and c are slightly different—examples are C D F; C‑D‑C F; D C D F; and G a c or G G‑a c. In troubadour songs, one also finds repeated notes on other pitches, such as G, with the intonation E F G, and other formulas, as well as recitations beginning directly on the reciting note. There is no necessary connection between the reciting note and the final of the piece; it may be a fifth above the final or on some other pitch. Although the presence of such reciting notes may help to delineate a tonal structure through emphasis on one or more pitches, the phenomenon is more properly dealt with in terms of melodic formulae, which deserve a separate study of their own.

            Next in order of complexity would be one‑step structures, which are discussed by van der Werf, although he allows they are not found in troubadour or trouvère songs except perhaps in a few individual phrases. It is two‑ and three‑step structures (and larger) that are most often found in these songs, according to van der Werf, especially chains of thirds which he says "occur in abundance and range from chains of only two thirds to chains of four and sometimes even five thirds."^[29] Another common structure consists of two thirds combined with a fourth to outline an octave, although there is sometimes doubt as to whether the top structural pitch in such patterns is a fourth or a third above the middle one. A melody may have two contrasting chains of thirds, one of which tends to predominate over the other, and there are songs in which a third chain may be established temporarily and then become obscured.

            In melodies that do not appear to have any clear structural tones, van der Werf nevertheless finds that one or two notes often function as basis tones. If this tone is in the middle range of the melody, it can be considered similar to Sachs' centric melodic types; a "standing" melody will have its basis‑tone near the bottom of its range, and a "hanging" melody will have it near the top. The basis tone could be compared with the modal final, since van der Werf says that the melody usually ends on it. He also says that compared with melodies with step‑wise structure, "there are many more instances in which it is difficult or impossible to determine the exact place of the basis or center tone."^[30]

            In theory, these categories of melodic types, recitation tones, interval chains, and centric types seem clear enough, and easily distinguishable. In practice, however, their operation, at least in the troubadour songs, can be very ambiguous for the analyst. In The Chansons, van der Werf provides an edition of four complete troubadour and eleven trouvère songs, with commentary. First of these is the well‑known song of distant love by Jaufre Rudel, "Lanquan li jorn son lonc en may" (P‑C 262,2), shown below in its three Provençal versions:

   

            Example 44. Jaufre Rudel, P‑C 262,2.

   

           In his commentary, van der Werf notes the recitation on F in lines 1 and 3 and on c in line 5. He then writes: "In its entirety the melody of Jaufre's chanson has a rather ambiguous structure: most lines have F and one line has the high C as the most important structural tone; only one line, the sixth, encompasses the entire range of the melody; and although the low C is not very pronounced as structural tone, it serves in all versions as ending tone of both pedes and of the entire chanson. Thus, in my estimate, it is difficult to determine whether this melody is a centric one, moving around F, or a standing one with C or perhaps even D as basis tone."^[31] By contrast, Leo Treitler, in a more recent discussion of the same song, points to chains of thirds as the most important structural factors in the song.^[32] The principal third D F (a c in W) is established at the beginning, and later expanded to include the two thirds above it; a secondary third chain, C E G (G b d in W) is contrasted with the first in the second and fourth lines, from the third or fourth syllable on. As Treitler summarizes, "this alternating relationship of phrases elaborating the two third‑chains and polarized to the secondary one is the commanding syntactical idea of the melody."^[33] The usefulness and pertinence of Sachs's categories of melodic structure would appear to be cast in doubt, at least as far as troubadour songs are concerned, since the same song can be understood, with some justification, as exemplifying all three of the main categories.

            The analytical prospects appear even more disheartening when one considers that this song of Jaufre Rudel must surely stand as one of the most solidly structured in the entire repertoire. Symmetry, simplicity and balance ensure the song's immediate appeal today as they may have at the time of its creation. All these features are present from the first line, as the ascending third of the beginning is answered by the same third in descending form, filled, at the cadence. The second line begins with the same minor third as the first and then expands the range a tone higher, with a melodic figure recalling that of syllables 6 to 8 in the previous line; it ends with a cadence like that in line one, but shifted down a tone—the "secondary" major third does indeed produce a sense of contrast and lessening of tension. Repetition of the first two lines in lines 3 and 4 (the song being in ABAB form) further enhances the effect of stability and balance before the contrasting shift to a higher tonal register in lines 5 and 6. The manuscript versions diverge the most in these lines, but all agree in maintaining elements of earlier lines while varying them. There is the minor third of the opening, shifted up a fourth, similar melodic turns as in lines 1 and 2, and finally the return to the cadences of line 1 and 2 in line 6, ms. X following the former, R and W the latter. To complete the formal and tonal balance of the song, line 2 is repeated as the final line.

            If one may distinguish tonal structure, an emphasis or centring around one or more tones, from melodic or formal structure, then both modality and third‑chains seem relevant to the song. The versions of R and X could be considered authentic D‑mode melodies except that their finals are C. The version of W could be considered transposed D or A mode; the b‑flat of X makes it closer to W than to R in pitch structure, allowing for transposition. The ambitus is an octave for the versions of R and W, and a ninth for X, with the final as the lowest pitch in all three. Minor thirds are prominently featured in six of the song's seven lines, at the beginning and end of the line; the secondary or contrasting major third a tone below is also not difficult to perceive in lines 2 and 4, along with its upper third.

             Many other songs are considerably more ambiguous, but as a relatively clear case, Jaufre's song could provide a hypothetical model for examining a larger portion of the repertoire, taking into account both modal features and third intervals or chains as possibly compatible elements in the tonal structure. Accordingly, an adequate sample (of about half the surviving troubadour songs) was examined for the presence or absence of certain general features relating to both kinds of approach. In terms of modality, data on finals, ambitus, and phrase finals was compiled and then other features of the songs were examined to see whether and to what degree the final could be considered a basis tone or structurally significant as a tonal centre in the song. (I think it can be safely assumed that this kind of question is more meaningful than whether or not it is possible to fit each song into one or another modal designation.^[34]) Factors that contribute to the perception of some pitches as having more structural weight than others include position in the phrase and in the melodic ambitus, with initial and final positions, highest and lowest notes tending to achieve more importance, along with peaks in the individual phrase; intervallic progression, since pitches approached or left by disjunct intervals such as a third or greater are thereby given greater prominence; repetition, both immediate and intermittent, if the latter is frequent enough; first notes in ligatures, except in cadences, where first and last notes have more weight. It is also assumed that in syllabic passages the single pitch for a syllable will carry greater weight than each individual pitch in a ligature in neumatic passages, although the syllable itself receives more emphasis in the latter. A certain amount of subjectivity seems unavoidable in the procedure, although I have tried to be as consistent as possible in evaluating each song.

            A total of 173 song versions were examined for general aspects of tonal structure, representing 117 distinct songs with attributions. All composers with four or more songs are represented in the sample, which thus includes songs by 17 troubadours and roughly half the total number of extant songs with music. To further reduce and proportionally balance the sample, only a percentage of the songs by composers with a greater number of surviving songs were included. For most composers this is roughly half of their total, and the selection was made arbitrarily, except that a preference was given to songs in multiple versions. Guiraut Riquier, however, has 48 songs (all in a single manuscript) more than double the number of the troubadour with the next highest number, so only 16 of his total were selected.^[35]

            The majority of the song versions (43 percent) have a range of a ninth; 25 percent fall within an octave, and 13 percent  extend to a tenth. A small number of songs have narrower or wider ranges. Ten songs (6 percent) have the range of a seventh, and seven (4 percent) the range of only a sixth.^[36] The widest range found in the sample, that of a fourteenth, is found in two of the three versions of Peire Vidal's "Be.m pac d'ivern e d'estiu" (P‑C 364,11) only; in the other version it has the range of a ninth. Two songs (1 percent) reach a thirteenth, four (2 percent) reach a twelfth, and six (4 percent) have an eleventh as their range.

             As we have just seen, the majority of songs fall within the standard modal ambitus of an octave plus one or two steps. Most finals are located at or near the lowest pitch in their songs' ambitus, corresponding closely with the authentic maneriae of modal theory. Of the 171 finals from the sample,^[37] 131 or 77 percent lie within one or two, more rarely three pitches of the lowest note in their song's range. Only three songs (2 percent) have a final near the highest pitch in their range, and two of these also have relatively rare pitches for finals, namely e and b.^[38] The remaining 37 song versions in the sample (21 percent) have a final in the middle or lower middle part of their range, and thus could be viewed as plagal.

            In order to consider the relation of the final to the whole song, it may be simplest first to group the songs according to final, and then to see what kind of relationship may be discerned in each group. The largest group comprises song versions with D as final, of which there are 64 in the sample, or 37 percent of the total. Next most common is G as final; it is found in 41 or 23 percent of the songs. C occurs almost as often as G and is found in 34 or 20 percent of the songs.^[39] Other pitches occur somewhat less frequently: F is the final in 16 songs (10 percent); a occurs in nine songs (6 percent);^[40] E and e occur in six songs (4 percent); b in one song only.^[41]

            Of the 117 distinct songs in the sample, there are 39 that are transmitted in more than one version; only 12 of these have the same final in all versions, while 27, or two‑thirds, have a different final in at least one of the other versions. Such a high degree of variability in finals could be taken as negative evidence for a meaningful correlation between finals and the songs' overall tonal structures. In ten of these 27 cases, however, the variability in finals is obviously the result of transposition of the entire song—if they were notated at the same pitch level, the finals would be the same.^[42] Of the remaining 17 songs with different finals, not quite half of this group of 39, there are eight in which the finals from other versions are one tone apart, four in which they are a third apart, and four in which they are a fourth apart; one song has a different final in each of its three extant versions.^[43] Most of these variants occur in songs by Folquet de Marseille and Gaucelm Faidit.

            Each song was examined in light of the criteria outlined above to assess whether and to what degree the final could be considered to function as a tonal centre in the song. Since this is a procedure most liable to subjective variation, an effort was made to include only the clearest cases in the positive and negative categories, the others being assigned to a middle category of ambiguous cases. In the positive examples, the final was structurally prominent in a majority of the song's phrases, appearing as initial and/or cadence pitch several times, or emphasized in other ways. It was found that thirds formed a strong association with the final in these cases, as the pitches in the chain of thirds above the final were also prominent. In the songs considered ambiguous, the final may have some weight, but it is not as obvious and exclusive a focal centre. Other pitches may seem equally likely as prepared finals, or else there are not enough indicators to allow a determination of the tonal orientation. For a number of the ambiguous cases, however, it would be possible to argue that the final is functionally related to the tonality or modality of the song, even though it may only be sounded a few times. This is because of the correlation with upper thirds found in songs assigned to the positive category; since in the majority of examples, the final is at the bottom of the song's range, while the greater part of the melodic unfolding takes place in the middle and upper parts of the ambitus, it is natural that the associated thirds would receive more play than the final itself. In songs given a negative designation, there was often evidence pointing to one or more tones as structurally prominent, but the final was some other tone unrelated to these.

            From the 172 songs in the sample, 91, or slightly more than half, had finals which were considered clearly functional and prepared; 46 (27 percent) were considered ambiguous, and 35 (20 percent) were negative. The greatest number of positive cases occurred in the group of songs with D finals, which is not surprising, since this is the largest group of finals. This group also displayed the highest proportion of positive cases, with 46 or 71 percent of the 64 songs with D finals showing a positive correlation. Songs with F, G, and C finals also had a significant proportion of positive cases, while songs with a, E or e, and b finals were found to be mainly negative or ambiguous in showing any functional relation between their final and the rest of the song.^[44]

            Example 45 below illustrates a typical D‑final song in which the final and associated thirds F, a, and c provide, for our ears at least, a clear tonal orientation throughout. The song is "Molt era.m dolz mei conssir," (P‑C 30,19) by Arnaut de Marueil.

 

            Example 45. Arnaut de Mareuil, P‑C 30,19.

     

Although there are no actual recitations in the song, it has the typical C D F initial formula that is characteristic of many in this group, without being a necessary feature. In some D‑final songs the tonal centre may be clear, but with a secondary or alternate orientation towards another set of thirds, such as C E G or G B D; after a shift in the middle of the song, there is a return to the main centre.^[45] Where the secondary tonal centre is not secondary, but tends to have equal or more weight than that indicated by the final, the song was considered ambiguous. Some evidence for this kind of uncertainty or interplay between two centres can perhaps be adduced from those songs preserved in multiple versions where the finals differ by a tone.^[46] Some of the songs with D finals which were assessed as negative regarding the relation between final and the rest of the song also showed strong implications of a tonal orientation around C with thirds E and G.^[47]

            The group of songs with F finals is perhaps too small to allow us to discern any pronounced association with opening formulas; some begin with recitations on a or c, a few have a step‑wise ascent from the F, and one also finds the intonation‑like patterns associated with the D mode, as well as other figures. In several of the ambiguous cases, the source of the ambiguity is the similarity between songs with F finals and those with D finals. (Some songs with D finals are ambiguous in a like manner.) This again may be viewed as evidence of the affinity between tones, and tonal centres, that are a third apart. Only a few of the songs with D finals are notated with a b‑flat; for those with F finals, the proportion is somewhat higher at four out of the nine with a positive functional relation to the song, which gives these songs a major rather than strictly Lydian character.^[48] The song "Us gays conorts me fay gayamen far," (P‑C 375,27) by Pons de Capdoill, is preserved in two manuscripts, R and X; in X, with an F final, there is a b‑flat at the head of every staff, while in R, the song is notated a fourth lower with a C final. Both versions are given in Example 46 below.

     

            Example 46. Pons de Capdoill, P‑C 375,27.

     

            None of the C‑final songs has an e‑flat, but a few (like P‑C 421,2, version of X, and P‑C 392,9) do have the b‑flat, which aligns them with the interval structure of G‑mode or Mixolydian melodies. It may not be significant, but one may note that in almost all the C‑final songs that are also found in other versions at a different pitch level, the other version has a G final. In several of the songs where the role of the final was deemed ambiguous or negative, one finds an emphasis on the D F A C set of pitches related to typical D‑final songs, as well as some oriented around G B D. In most of these the final with thirds E and G will become prominent only in the last line or the last few lines of the song. Two of the negative C‑final songs are interesting for the role of form in making the final seem justified, even though it is hardly heard throughout the song. "Conortz, era sai eu ben," (P‑C 70,16) by Bernart de Ventadorn, consists of four phrases which are repeated; the musical form may be represented as ABCD/ABCD'. All phrases except D are clearly focussed on the pitches D and F, with the standard formulas of positive D‑final songs (in ms. G; the version in R is notated a fifth higher with G final). The fourth and eighth phrases, which serve to articulate the form, are built around C. In "Molt m'entremis de chantar volunters," (P‑C 366,21) by Peirol, most of the phrases are framed within the fifth above G, with some emphasis on b and d. The basic musical form is ABC/D/ABC* (the asterisk represents a new or altered cadence, as explained in Chapter II). The high c is heard as a brief recitation in the A phrases, and the low C is part of an ascending fifth in the middle of the C phrases, the first of which cadences on D, the second on C. The two may be said to work together in a manner analogous to the ouvert/clos types of cadence, thus revealing some logic in use of C as final, even though by our criteria the final must be considered unrelated to the tonal centre of the song. Example 47 is a transcription of the song by Peirol.

   

            Example 47. Peirol, P‑C 366,21.

     

            For songs with G finals in the positive category, there seems to be little consistency regarding opening formulas, although a good number of the songs begin in the upper part of their range. Where alternate or secondary centres appear, they tend to lie within the thirds F A C, or to a lesser extent, C E G. The latter also tend to appear as the actual centres in those of the G‑final songs that were judged ambiguous or negative. A song by Guiraut Riquier, "Mentaugutz auch que Dieus es," (P‑C 248,55, Example 48) may serve to illustrate one of the clearer cases where the G final is functional and prepared.

 

            Example 48. Guiraut Riquier, P‑C 248,55.

 

            Only a handful of songs in the sample, 15 in all, have either a, E or e as final, and an equally small percentage of these were considered positive regarding the functional role of the final as tonal centre, namely two of the A‑final songs.^[49] One of these, "Del seu tort farai esmenda," (P‑C 366,12) by Peirol, is preserved in two manuscripts; in G, it is notated a fifth lower than in X, which has the a final. Since neither version has any accidentals, the one with a D final has a B‑natural where the other has an F, although the latter pitch occurs only once in the version it is transmitted in. The song is one of the relatively few with a clearly plagal ambitus, ranging a fifth above and below the final. The other is a song by Bernart de Ventadorn, "Pos mi pregatz senhor," (P‑C 70,36) which is also found in two versions, the other having an F final. Both versions showed a functional relation with their finals, but because of the divergence between the melodies, it is not clear whether one should be viewed as a transposition of the other. Most of the ambiguous cases with A or a finals are centred around C E G, and thus display the third relationship found in many songs, but the A itself was sounded very little.

            In four of the six songs with E or e finals there was little evidence of a functional role for the final; three of these are preserved with different finals in other versions. The other two were judged to be ambiguous because both seem to be as equally centred on G as they are on E or e. Perhaps the closest to a genuine "Phrygian" melody is Berenguier de Palazol's "Totz temoros e duptans," (P‑C 47,12), shown below in Example 49; the final is located near the very top of the song's ambitus.

 

           Example 49. Berenguier de Palazol, P‑C 47,12.

     

            Even within the limitations of the present investigation, I think one may conclude that the troubadour repertoire bears strong traces of a functional modality, and that, far from being at odds with the kinds of structures discussed by van der Werf, such as interval chains, it is closely associated with one of these, namely the chains of thirds. There is an important distinction to be made, however, between the role of thirds in defining the tonal characteristics of a troubadour melody as outlined in the preceding pages, and their role as basic melodic structures as discussed by both Sachs and van der Werf. For these authors, the thirds are the skeleton of the melody; other pitches are secondary and clearly ornamental. Such cases are extremely rare among the troubadour songs if there are any uncontentious cases at all (possibly in limited passages of a "lower style" or simple dance-type melody). Where two or three thirds occur in direct succession in a troubadour song, they normally form only part of a single phrase, and the pitches outlined in thirds may have no necessary connection with the main tonal centre of the song. What does occur, is that a series of three or more pitches a third apart tend to acquire more emphasis than others through repetition or placement; they may be considered structural in the sense of contributing to the "modal" or "tonal" character of the song, but they perform this function in a much more fluid and general way than the "structural tones" of van der Werf or Sachs.

            By the most conservative estimate, slightly more than half of the songs examined had a final which was clearly functional as the tonal centre of the song, and in most of these, the two or three thirds above the final were also prominent enough to suggest a deeper affiliation based on this interval. As mentioned above, a large number of the ambiguous songs did show evidence of orientation around pitches a third apart, and located a third, fifth, or seventh above the final, but the final itself did not appear except sporadically or towards the end of the song. It will be recalled that Theodore Karp discerned two groups of songs in which there was either a shift from one tonal centre to another, or an oscillation between the two, besides the category in which the final had no connection with the rest of the song. Another of the consistent patterns found in the present study is that the alternate or secondary tonal centre is to be found a tone below the main one, along with its associated thirds. This interplay between two centres might also be interpreted as a common feature of the tonal style in the troubadour songs. In considering again the ambiguous or negative cases from this point of view, one will find that much of the ambiguity in the former category derives from the almost equal emphasis on both the two centres; conversely, in many of the negative cases, where the final is not a functional centre, the actual tonal centre will be a tone above the final. Though few in number, the multiple versions also tend to support this judgment; where the versions differ as to final, the finals are usually a tone apart.

            One may therefore conclude that in a majority of the troubadour songs, there is evidence for a flexible sense of tonal structure based upon one or more interval‑chains of thirds; there is also a flexible but fairly consistent association between third‑chains and finals. The present chapter has shown the different ways these structures function in a large portion of the repertoire. The detailed examination of 173 songs has also permitted the clarification and refinement of previous opinions on the subject. Appendix III contains a list of the songs examined and the tonal features displayed.

 

 

  «  Chapter IIIe                    Contents                    Chapter V  »