IN THE LATE FOURTEENTH CENTURY, Egidius de Murino wrote an elementary treatise on how to compose a motet. The first step, he told his readers, was to select a tenor appropriate for the theme of the motet:

After that, Egidius says, the tenor is arranged into phrases and given rhythm (a process he calls ‘ordering and coloring’), and the other parts are added – contratenor first, if there is one, then triplum and motetus. The upper-voice texts, according to this account, are added only at the end, by dividing the music and the words and combining them ‘as well as you can’.

It is important to remember that Egidius is writing for beginners, so he takes each stage in turn, while the experienced poet-composer would surely work on several levels simultaneously. The choice of a specific tenor may reflect not only a general sense of subject matter, but also the knowledge of certain key words, if not entire lines, of the upper-voice texts, or specific melodic features that the composer intends to work out in the polyphonic framework. In perhaps an extreme (but not unique) case, Anna Zayaruznaya makes a convincing argument that the composer of Colla / Bona had already written the texts and worked out the periodic structure of the motet before choosing the tenor and giving it a repetition pattern that is at odds with its upper voices. Moreover, a motet that borrows a tenor used in another motet creates a further dialogue that goes beyond general subject matter to include various aspects of structure, text, and music. The ways in which a composer might work out tenor choice and other aspects mentally resist the kind of clear linearity Egidius's description gives us.