on recognizing this identity. This poet, while firmly anonymous, is by no means what Leo Spitzer, in a seminal essay, calls an “everyman” whose words are completely fungible. Spitzer’s categorization relies on texts where a known author appropriates language that cannot possibly apply to him or her, usually by direct quotation or literal translation. In one example, he cites a passage from Marie de France where she claims personal experience of a story’s teller in an echo of the original but elsewhere acknowledges her use of a written source.11 The final stanza of “When the Nightingale Sings” uses voice differently. While drawing on the French envoy tradition, it neither directly appropriates its language nor relies on the poet’s identity for its effect. Instead, this stanza transforms the French convention tactically, suggesting an expansive but not limitless range of possibilities for its speaker, who would be familiar with the women and terrain “Bituene Lyncolne ant Lyndeseye, Norhamptoun ant Lounde.” In other words, while much of the poem consists of an arrangement of conventional lyric language (though beautifully executed), the final stanza locates this conventional language within a specific horizon of practices. As we will see, this poem’s compilation and layout in MS Harley 2253 amplifies how its voice mediates between such textual and performative practices, which in turn informs the arrangement of the larger miscellany.
To understand how this works in the manuscript as a whole, it is helpful to consider how medieval theories of voice express its multimodal capacities and how these can be put in dialogue with modern discussions of lyric voice. This discussion occupies the first half of this chapter. The second half of the chapter provides an overview of the manuscript’s contents and social context and reads a selection of English and French lyrics from the manuscript, both well known and lesser known, as well as some of the manuscript’s Latin devotional texts.
Voice and the Lyric
Voice is where sound meets language, a grammatically vexed and poetically fertile medium fundamental to human communication. In this section, I show that in medieval theory, voice—particularly literary voice—is inherently tactical, articulating relationships between writing and performance and between a subject’s interiority and his external world. In medieval philosophy, grammar, and music theory, theories of voice rely on and elucidate ontological distinctions. Aristotle says that vox—here meaning voice or even “vocal sound”—consists of “a certain kind of sound belonging to what has a soul” that has signifying power.12 While musical instruments produce a sound similar to voice, only living beings—including animals—have voices. In Aquinas’s commentary on this passage, voice is what turns meaning into sound: “Not every sound belonging to an animal is a vocal sound. For the tongue may make some sounds that are nevertheless not instances of vocal sound—just as those who cough make a sound that is not a vocal sound. For if there is to be vocal sound, what forces the air [against the windpipe] must be something with a soul, along with some imagination intended to signify something. For a vocal sound must be a certain significant sound, either naturally or by convention.”13
For Aristotle and Aquinas, voice is produced by, yet distinct from, a speaking subject. It is the medium through which a subject’s internal image—stored in the imaginative faculty—encounters the external world. Because voice in these definitions is not only a linguistic medium, medieval grammarians further distinguished types of voice by their semantic potential. According to Donatus, “Every sound [vox] is either articulate or confused. Articulate sound can be captured in letters, confused sound cannot be written.”14 Priscian describes four different types of vox based on two binary pairs: articulate or inarticulate (depending on whether the sound has intentional meaning) and literate or illiterate (depending on whether it is writable). He acknowledges that unwritable sounds (such as a human groan) may have meaning, and writable sounds (such as a nonsense word or transcribed animal noise) may lack signification.15 Later grammarians drew on these four types to classify sounds ranging from human spoken language, applause, and whistling to animal noises and the sounds of natural phenomena like crashing waves.16
The difference between Donatus’s and Aristotle’s definitions of vox captures a medieval concern with the relationship between voice and speaker. Whereas for Aristotle, the ontological status of the producer of sound determined whether or not it was a “voice,” for Donatus and the early grammarians, a sound’s resolution into the phonetic alphabet was central. This association between writing and voice not only speaks to grammar’s close relationship to performed oratory in the early period (it was a foundational discipline for the study of rhetoric) but also reveals how writing and voice are co-constitutive in premodern culture. Their interdependence is particularly evident in drama: the multivocal drama of ancient Greece owes its conception to a phonetic alphabet that weds sound to letters, for example, and medieval drama borrows many of its conventions from legal rhetoric.17
But while access to writing was controlled, in the Middle Ages, by institutions of literacy like the schools and the church, voice is a biological attribute shared by humans and animals. As poets have long recognized, human language draws on both aspects of voice, its grammatical and systematic principles and its nonrepresentative sounds. Ezra Pound described these, respectively, as logopoeia (meaning) and melopoeia (sound).18 Later medieval grammarians like John of Garland recognized both aspects of language. In addition to his better-known Poetria Parisiana (discussed in the next chapter), John composed an “equivocal” grammar, a medieval genre of long poem that distinguishes and contextualizes like-sounding words.19 As he puts it in the prologue to this poem, “equivocum celat sub eadem plurima voce, / quorum nomen idem” (an equivocum hides under a voice [word] these many [meanings], which have the same name).20 The treatise (like others in the genre) distinguishes among like-sounding words in a mnemonic verse:
Augustus, -ti, -to Cesar vel mensis habeto,
Augustus, -tus, -ui vult divinacio dici
Mobile si fiat, augustus nobile signat,
Augeo dat primum, dant gustus avisque secundum.
[Augustus, -ti, -to means Caesar or the month (of August); Augustus, -tus, -ui means divination. If it becomes an adjective, augustus means noble. The verb “augeo” (to grow) gives us the first meaning; “gustus” (taste) and “auis” (bird/omen) give us the second.]21
Just as equivocal grammars unite the melodic and grammatical qualities of voice, rhetorical treatises also give primacy of place to voice, an essential component of the rhetorical canon of actio, or delivery. Quintilian acknowledges that a “good voice” is among the natural gifts necessary for success in oratory22 and devotes part of his discussion of delivery to the correction of vocal infelicities: “Again our teacher must not tolerate the affected pronunciation of the s, with which we are so familiar, nor suffer words to be uttered from the depth of the throat or rolled out hollow-mouthed, or permit the natural sound of the voice to be over-laid with a fuller sound, a fault fatal to the purity of speech.”23 The ethics of oratory are implicit in the privileged faculty of speech: “If therefore we have received no fairer gift from heaven than speech, what shall we regard as so worthy of laborious cultivation, or in what should we sooner desire to excel our fellow-men, than that in which mankind excels all other living things?”24
Following the newfound popularity of the pseudo-Ciceronian Rhetorica ad Herennium in the twelfth century, delivery became a subject of serious rhetorical discourse in the later Middle Ages, as well as an influence on the broader culture, permeating scholastic treatises and performative practice.25 The Herennium describes tones of voice appropriate for different kinds of material: “Conversational tone comprises four kinds: the Dignified, the Explicative, the Narrative, and the Facetious. The Dignified, or Serious, Tone of Conversation is marked by some degree of impressiveness and by vocal restraint. The Explicative in a calm voice explains how something could or could not have been brought to pass. The Narrative sets
