^{Note 1 in page 320} Seventeen of Peirol's thirty-two authentic poems have melodies. Sixteen of the songs with melodies are love poems. It is my intention to examine only the love songs in this paper.

The melodies of Peirol are found in three manuscripts, G (Milan, Biblioteca Ambrosiana R 71 sup.), R (Paris, Bibliothèque Nationale, français 22543), and X (Paris, Bibliothèque Nationale, français 20050). Facsimiles of the manuscripts are included in Peirol, Troubadour of Auvergne (Cambridge, England, 1953), an edition of the troubadour's works by S. C. Aston. All of Peirol's melodies have been published by A. Restori, “Per la storia musicale dei trovatori provenzale,” Rivista musicale italiana, in (1896), 407–451, and by Friedrich Gennrich, Der mvsikalische Nachlass der Troubadours (Darmstadt, 1958 and 1960, 2v.). In addition, the fourteen melodies contained in G appear in Ugo Sesini's edition of that manuscript, “Le mélodie trobadoriche nel Canzoniere provenzale della Biblioteca Ambrosiana (R 71 sup.),” SMed, nuova série, xii (1939), 1–101, xiii (1940), 1–107, xiv (1941), 31–105, xv (1942), 189 ff.

I have assumed in this paper that Peirol wrote both music and words of the songs which are attributed to him since we have no evidence of his having borrowed the melodies or of their having been composed by someone else. The stylistic and structural unity of the songs would support the assumption that they were produced by one person. Restori has emphasized the authenticity of Peirol's music, Rivista musicale italiana, in, 408.

^{Note 2 in page 320} Restori does not go into the question, his monograph on Peirol, referred to in the preceding note, being taken up with presenting the melodies and including only a paragraph or so of comment on their character. In his edition (Peirol, Troubadour of Auvergne), S. C. Aston does not discuss the music at all.

^{Note 3 in page 320} Théodore Gérold, La Musique au moyen âge (Paris, 1932), Ch. xii. This chapter constitutes the most extensive discussion of the relationships between verse content and melody in medieval secular song. Gérold reviews the entire repertory, troubadours and trouvères, and all genres of lyric poetry. He does not examine the output of any one poet in entirety. Other scholars have brought up the question more briefly, usually to point out isolated examples of songs in which the melody seems to reflect the sense of the words. See, for example, Barbara Smythe, “Troubadour Songs,” Music and Letters, ii (1921), 267–269 and 272; Jean Beck, La Musique des troubadours (Paris, 1928), pp. 75–85; Sal-vatore Battaglia, La Lirica médiévale (Naples, 1956), pp. 180–184 and 294–297. Of a different persuasion had been Pierre Aubry in Trouvères et troubadours (Paris, 1909). For him, troubadour and trouvère music can reflect the text only in a narrow, material sense: “D'expression musicale, il ne pouvait être question dans une mélodie comme la mélodie des troubadours et des trouvères, où la carrure la plus absolue faisait loi, mais la phrase était l'image du vers auquel elle était jointe, elle était calquée sur lui, elle en était comme un écho agrandi et sonore” (pp. 10–11). Aubry found in particular the melodies of the courtly lyrics lacking in naturalness and spontaneity (p. 101).

^{Note 4 in page 321} Maurice Valency, In Praise of Love (New York, 1958), pp. 116–117. Valency, of course, is not the only one to have discussed this question. The problem of the sincerity of the troubadours, of the fiction or truth of their loves, of the whole relationship of courtly love to genuine emotion has been commented on by all who have written about them: Friedrich Diez, Die Poésie der Troubadours (Zwickau, 1826), pp. 122–123; Eduard Wechssler, Das Kulturproblem des Minnesanges (Halle, 1909), pp. 213–216; Carl Appel, Bernart von Ventadorn, seine Lieder (Halle, 1915), pp. xxiv-xxix; Alfred Jeanroy, La Poésie lyrique des troubadours (Paris, 1934), n, 94–113; Stanislaw Stronski, La Poésie et la réalité aux temps des troubadours (Oxford, 1943); Pierre Belperron, La Joie d'amour (Paris, 1948), pp. 173–190, to mention only a few. While it is obvious that lyric poetry cannot exist totally unnourished by human feeling, most authorities agree that in courtly love there is little genuine individual emotion.

^{Note 5 in page 321} Aston points out the conventionality of the subject matter of Peirol's poems (Peirol, Troubadour of Auvergne, p. 23); but on the other hand he has drawn from these poems (and the vida) a love story of sorts upon which to base his narration of Peirol's life and the order of the poems in his edition. However, it is difficult to think of these poems as being instigated by specific situations, experiences, or feelings; the weakness of Dr. Aston's arguments has been pointed out by Kurt Lewent in his review of the edition, RR, xuv (1954), 274.

^{Note 6 in page 321} One song could constitute an exception. It is not included in this discussion although placed by Aston among the love poems: No. xxvi, “Ren no val hom joves que no-s perjura.” Pillet-Carstens (Bibliographie der Troubadours, Halle, 1933, p. 326) followed by Istvân Frank (Répertoire métrique de la poésie des troubadours, Paris, 1953 and 1957, i, 195) considered this poem a sirventes, to which type it seems more closely related in style and content. Aston does not give any reason for his classification. In fact, he does not discuss the poem at all, which is strange, as Kurt Lewent has noted: RR, XLV (1954), 274–275. For if the poem has certain stylistic features in common with Peirol's crusading song, “Pus flum Jordan ai vist e-1 monimen,” it is startlingly different from all the other songs, especially the cansos. And serious question could be raised about its authenticity. The poem shows a high percentage of nouns, particularly concrete nouns, and a vigorous style. In no other canso did Peirol sing except at the instigation of love, nowhere did he suffer heat or cold or speak of the night being dark, at no other time did he concern himself with such folk as merchants and how they acquire wealth, yet all of these things are in this song. The poem ends, moreover, with a powerful image unlike anything else Peirol invented. In view of the fact that this poem is so different from the others, that its authenticity is doubtful, and that it appears rather to be a sirventes than a canso, I have not taken it into consideration in this paper.

^{Note 7 in page 322} The Roman numerals refer to the order of poems in S. C. Aston's edition, the Arabic numerals to the lines. All subsequent references to Peirol's poems will be to that edition.

^{Note 8 in page 323} The song begins :

Si be-m sui loing et entre gent estraigna eu ai pensier d'amor en que-m conort, e pens d'un vers cossi-1 fass' e l'acort tal que sia bos e valens e fis; et on horn plus mos chantars mi grazis e mieils me dei gardar que no-i mesprenda ni diga ren don savis me reprenda.

In conformity with the poet's intention, the song has a complicated organization of rhymes, more so than is Peirol's wont. It contains six coblas capcaudadas with one tornado,. The rhymes of the even-numbered stanzas reproduce the rhymes of the odd-numbered stanzas backwards; thus they are retrogradatz.

^{Note 9 in page 323} Posing the problem in this manner seems to assume that the music was always written after the words. The statements of some troubadours might lead us to believe that this was not necessarily the case. Arnaut Daniel begins one of his songs with the assertion : “En cest sonet coind' e leri / Fauc motz e capuig e doli.” Bernart de Ventadorn also informs us in the poem “Lone tems a qu'eu no chantei mai” :

Ara no tem ploya ni ven, tan sui entratz en cossire com pogues bos motz assire en est so, c'ai… (ed. Appel, lines 3–6)

To be sure, where melodies were borrowed and new texts fitted to them, the music was obviously composed first; the technique of inventing words for a pre-existent melody was well-known in the Middle Ages. But a canso melody was supposed to be original. With reference to the quotation from Bernart de Ventadorn just given, Carl Appel has denied that the troubadour wrote first his melody and then the words; rather he imagined the poet to have had in mind a melodic schema which determined the metrical structure of the poem, with the actual music coming into existence only when the text was done (Bernart de Ventadorn, p. xcviii). This position seems justified; however, one cannot entirely rule out the possibility that other procedures may have occurred.

^{Note 10 in page 324} This could match the first stanza of the poem or the third, is less suited to some of the other stanzas—the fourth, for example.

^{Note 11 in page 324} In a note, Restori qualifies this melody as “insolita-mente manierata” (Rivista musicale italiana, iii, 429), which is exactly the impression it conveys—extraordinary for Peirol, none of his other melodies being quite like it. Perhaps this is the result of the poet's avowed intention to fabricate an intricate work of art.

^{Note 12 in page 324} Friedrich Gennrich, Grundriss einer Formenlehre des mittelalterlichen Liedes (Halle, 1932), pp. 204–20S, and Hans Spanke, Beziehungen zwischen romanischer und mittellateinischer Lyrik (Berlin, 1936), p. 12, have both left out the middle phrase of this melody, indicating its structure to be ABC ABC, an error corrected by Gennrich in Der musikalische Nachlass der Troubadours, i, 124, and ii, 73. Elsewhere I have discussed the structure of this song more in detail, “Metrical and Musical Structure in the Songs of Peirol,” RR, LI (1960), 241–255.

^{Note 13 in page 325} This was also the opinion of Restori, Rivista musicale italiana, iii, 408.

^{Note 14 in page 325} Besides interval structure and general contour, the scales on which the melodies are based could also be mentioned as determining their character. The melodies lie for the most part within the ecclesiastical modes, lacking therefore the leading tone and a drive toward the tonic. In this they are not different from other troubadour melodies. It is often exceedingly difficult, if not impossible, to assign the melodies to a specific mode; one may say, however, that Peirol prefers the Dorian and Mixolydian modes, and that the ambitus, especially with the Mixolydian mode, tends to be plagal. The Lydian mode does not occur at all. Discussion of the tonality of the songs preserved in manuscript G may be found in Sesini, SMed, XII (1940), 105–107, and xiv (1941), 33–55.