In his three motets with tenors taken from secular songs, Guillaume de Machaut experiments with upper voice structures that borrow from the talea principle used in chant-based motets, creating hybrid forms that reflect aspects of the motet's overall subject. In two cases, Machaut sets up upper-voice taleae that do not coincide with their song-based tenor but interact with it in interesting ways. Trop plus est bele / Biauté paree de valour / Je ne sui mie certeins (M20) balances these formal principles to reflect a perfect love balanced between the dedicatory and the sacralised, while Lasse! comment oublieray / Se j'aim mon loyal ami / Pour quoy me bat mes maris? (M16) creates three opposing forms that reflect a Lady looking in two different directions, towards a beloved and a husband who abuses her. Dame je sui cilz qui weil endurer / Fins cuers doulz, on me deffent / T. Fins cuers doulz (M11) does not define regular upper-voice taleae, but rather uses the tools by which taleae are defined in the upper voices – long rests, hocket sections, and melodic repetition – to merge disparate formal principles in the service of a motet that discusses a woman who merges a soft appearance with a hard reality. Here Machaut also uses hexachordal punning, combining sharps and flats to express the Lady's contradictory qualities.

The ninth century witnessed a fundamental change in the way Western musicians thought about music. Before the Carolingians assimilated ancient music theory, they had no functional concept of how the intervals between pitches of the scale differed from one another and how those differences affected melodic structure. The transition to interval-based thinking may be traced in writings about music. The first half of Aurelian of Réôme's mid-century Musica disciplina quotes from Boethius, Cassiodorus and other ancient authors, but fails to make sense of what they say about intervals. The second half describes the rise and fall of chant melodies without reference to intervals. Treatises of the later ninth century (the Enchiriadis treatises, Hucbald's Musica) are the first to treat music in terms of individual pitches and explain how patterns of whole tones and semitones define modes and scales. However, an early draft of Musica enchiriadis, the Inchiriadon, still displays no awareness of the role that semitones played. A parallel evolution occurred in notation. Neumes, which outline melodic direction but not precise intervals, can be documented from the second quarter of the ninth century and are likely older. They lack pitch content because musicians who invented them lacked a conceptual framework for understanding pitch. Pitched notations do not appear until late in the century and their use is confined to examples in theory treatises.

Word setting in Machaut's refrain songs poses a problem, for whilst it is clearly indicated in the manuscripts, it often does not comply with recognised principles or values. To understand the situation, a dualistic relationship of words and music is proposed. It is founded in the coordinated but independent operation of principles of musical mimesis and musico-poetic dislocation. The music is constructed at a primary level as an imitation of the poetic form; but it is fundamentally independent of this model and may thus be detached from it and displaced against it. Devices such as ‘cross-cadencing’, ‘quasi-declamation’, ‘complementary-cadence inversion’ and ‘dissonance’ between implied and actual word setting are manifestations of this technique. The proposal accounts on the same basis for both the close relationship of words and music observable in the virelais and for the more abstract connection apparent in the rondeaux. There is a technical unity at work across the genres in Machaut's song composition.

Most standard musicological references attribute to Bishop Stephen of Liège (†920) the composition of three Offices: for the Holy Trinity, the feast of Saint Lambert (bishop of Liège in the early eighth century to whom the city's cathedral is dedicated) and the Invention of Saint Stephen the protomartyr. From statements by Richarius, Stephen's successor at Liège, and Folcuin of Lobbes (both from the tenth century), and the eleventh-century account of Anselm of Liège, along with the evidence in Brussels, Bibliothèque royale, MS 14650-59 (a tenth-century manuscript produced at Liège during Stephen's episcopate), I conclude that Stephen of Liège did have a hand in the Offices for Saint Lambert and the Holy Trinity. Although he may have composed chants for Saint Lambert, he more likely revised existing ones for the Trinity. No tenth- or eleventh-century testimony attests the attribution of the Office for the Invention of Saint Stephen to Bishop Stephen.

The dualistic relationship of words and music in Machaut's refrain songs (proposed in Part 1 of this article) enables abstraction of the music from the poetic model that inspires it. This situation is at its most extreme in certain rondeaux, representatives of a genre for which a special function in Machaut's output is argued. Studies of specific groups of songs (B9–R4 and B35–R13–R21) illustrate the gradual development and detachment of material in related compositions. Through these accounts, it can be seen that the relationship of words and music proceeds on the same fundamental basis in the seemingly melismatic rondeaux as in the syllabic virelais, despite the apparently closer connection found in these latter. The technical unity at work in Machaut's song composition is constitutive of an aesthetic position. The abstracted musical text is an aspect of the multifaceted voice of Machaut's lyric œuvre and tends to subsume the poetry it sets creating an aesthetic of pure music.

The extensive melodic revision of plainchant in editions of the Graduale Romanum published in Italy from the late sixteenth century onward resulted in musically diverse repertoires that could depart widely from earlier chant traditions. The scale of the changes in these sources, both in type and in number, has obscured certain aspects of their editors’ work: their familiarity with the corpus, their aims and techniques, and their approach to the task. Previous analyses concluded that the editors worked on a chant-by-chant basis, and were either unaware of or ignored any shared melodic relationships between pieces of plainchant. An examination of the revisions to the recurrent melody used by the eight Ostende alleluias in three influential Italian printed graduals – Gardano 1591, Giunta 1596 and Medici 1614/15 – provides a different perspective. Analyses of the reshaped chants reveal that the editors possessed knowledge of the repertoire guiding aims, and favoured revision techniques. The combination of these factors, whether intentionally or not, resulted in the chants’ continued structural connection in the midst of increased melodic diversity. The individuation evident the chants did not necessarily signal the editors’ unfamiliarity with the repertoire, but could have been indicative of their intentional rejection of shared elements. Further, the revisions to the Ostende alleluias reveal that the editorial process could be flexible, with the chants approached both as individual entities and as groups. These findings demonstrate the complexity of the editorial process in early modern Italian printed graduals, and deepen our understanding of this multifaceted repertoire.