Song and singing are topics of immense philosophical and aesthetic interest which have received comparatively little attention from philosophers. This chapter surveys some of the main areas of philosophical interest related to song, including the definition and ontology of song, meaning in songs, and some questions surrounding vocal performance and the ubiquity of singing. For a discussion and diagnosis of the entrenched philosophical “prejudice” against songs, see Ridley (2004: ch. 3).
Singing is both a performing art and a cultural practice, and it plays a part in the artistic culture and domestic life of every culture of which we know. Two related features distinguish songs from other musical forms. First, the presence of a text. Although “songs” may be composed for musical instruments (such as Felix Mendelssohn’s “Songs without Words” for the piano), these are not songs in the strict sense, but works that share some characteristics with them, notably a clear melodic line and accompaniment. “Vocalise” is song without a text – a borderline case in which the voice is treated as a musical instrument and the melody is sung on a vowel sound. Second, singing is by definition a vocal activity, although the voice may be accompanied by musical instruments.
While speech and singing may be understood as contrary to one another, the distinction between them is best made not physiologically but on cultural and pragmatic grounds. Using a spectrograph, George List (1963) recorded the pitch contours of various vocal activities and found many gradations between everyday speech and singing. Speech intonation may level out and approach a monotone or be heightened and exaggerated. Examples of vocal communication falling between singing and speech, yet arguably belonging clearly to neither, include rap, children’s skipping and clapping rhymes, auctioneers’ chants, street sellers’ calls and cries, field and street hollers, the chants used in meditation and religious practices, and calls to prayer. Whether any of these are considered examples of “song” (or indeed examples of music) depends on cultural expectations and related attitudes regarding these very categories. For example, in cultures where “music” is understood as a secular activity or has little prestige, participants may be reluctant to describe their vocal activity as singing and prefer to define it as part of a larger spiritual practice. Islamic calls to prayer may sometimes sound like songs but are not usually considered to be singing performances, and the similar role of the cantor or chazzan in Judaism is seen to be primarily moral or spiritual rather than musical.
Today we may find it natural to think of song as a hybrid form combining words and music which are likely to have been composed separately before being brought together. (See, for example, Levinson 1990.) However such an assumption is historically uninformed. In earliest times music and poetry were one; their separation – what has been called music’s “emancipation” from language (Neubauer 1986) – came later. A number of theorists have argued that music originated before language or that the two have a common origin. These include Rousseau (1998) and Darwin (1981), as well as contemporary researchers (Brown 2001; Mithen 2006). The evidence we have suggests that the beginnings of poetry lay in the fitting of words to pre-existing melodies and rhythms (Winn 1981: 1). C. M. Bowra, in his study of the songs of pre-literate cultures, argues for five evolutionary stages in the development of songs: the meaningless line, the repeated intelligible line, the single stanza, the collection of stanzas into longer songs, and the collection of these songs into cycles (1962: 86). In the West, the division of poetic from musical technique can be traced to the development of writing. With another method available for preserving poetic texts, instrumental musicians could pursue melody and rhythm for their own sakes, rather than for the purpose of aiding memorization (Winn 1981: 17–18). Ancient Greek mousiké which had previously been a unified whole, eventually broke down into four distinct pursuits: musical composition and performance, literary composition and performance, musical theory and philosophy, and rhetorical theory (Winn 1981: 30).
Given the ubiquity of songs the world over, there are many different types of songs, sung in many different contexts and serving diverse purposes. A philosophical approach appropriate to one type of song may not be fruitful when applied to another song. I propose a three-part classification of songs. First, while any song can be performed in front of an audience, some songs are “works for performance,” specifically intended to be performed, often in a formal setting (Thom 1993: ch. 1; Davies 2001: 20–5). These include art songs, songs in opera and music drama, jazz standards, and the songs recorded by professional singers for a mass media audience. Second are songs intended for “participation-performance” or communal singing. The “audience” and the performers are one in this case. Such songs include national anthems, hymns, campfire songs, and many folk songs. Even when only one person performs such a song, he or she does so less for an audience than on behalf of an audience (Zuckerkandl 1973: 27). Finally, some songs are best understood as “functional” songs because they serve specific practical or cultural purposes. Examples include lullabies, mnemonic songs, work songs, and laments. Cone (1974) proposes a different classification of songs, based partially on musical and partially on functional considerations. “Simple songs” have no accompaniment or only simple accompaniment (Cone 1974: 58). An “art song” is a poem set to a composed vocal line and united with a fully developed instrumental accompaniment. In contrast to these are “natural songs” such as ballads, in which the roles of the poet and composer are “hardly relevant” (Cone: 1974: 59). Finally, in “functional songs” the vocal persona of the song is an aspect of the actual singer, expressing himself or herself as a member of a specific community taking part in a ritual or assisting at a social event (Cone: 1974: 49–52). “Happy Birthday” is the classic example.
For the purposes of analysis we can consider a song’s text separately from its other elements. What is the proper construal of the relationship between the meaning of a song’s words and the melody, harmony, and rhythm to which they are set, as well as any accompanying melodies? Many different answers have been proposed. Words in songs may be “reinforced, accented, blurred, inspired to a new meaning, in a continual interplay” by accompanying music (Booth 1981: 8). It is worth noting that already by the early Christian era three distinct forms of singing were recognized, each positing a different relationship between words and music. First was the chanting of the Psalms on a single note, with rising introduction and falling cadence. The texts were essentially treated as prose. Second were the elaborated and ecstatic melismata sung to the words “alleluia” and “amen.” Finally hymns, designed for congregational singing, displayed a more equal partnership between words and music, with each word set to one or at most two pitches (Winn 1981: 35–6). These different vocal practices point to different underlying attitudes toward the relationship between a song’s text and its other elements.
Rhetorical models of music, which originated in ancient Greek ideas about the character-forming power of mousiké, dominated theorizing about musical meaning until the early modern period and continue to be significant today. Their influence remains evident in what might be called the “propositional” model of song meaning. According to this, the meaning of a song can be reduced to the meaning of its verbal text. Rousseau subscribed to a variety of this doctrine and his views are typical of his time. Language conveys ideas, but to convey feelings as well words must be set to melody (Rousseau 1998: 324). Words are thus the primary carriers of meaning and the most that music can do is to supplement them. The propositional model becomes unsatisfactory when we take a closer look at song texts (Booth 1981). Songs are a form of oral communication and as such are subject to the burdens and limitations of oral communication. Song texts tend to be highly redundant, predictable, often formulaic, convey a low density of information, and trade in familiar simplifications (Booth 1981). Folk and popular songs must be accessible to their intended audiences; this in turn requires some fidelity to familiar forms (Gracyk 2001: 18–26). So a problem arises: if the meaning of a song is nothing more than the meaning of its text, it becomes difficult to explain the endurance of songs and singing across time and space. This is because singing is not a very efficient means for conveying propositional meaning; song is music and text is not. Hence while most songs do convey a text, their meaning cannot be reduced to the propositional content of that text.
One influential account of the relationship between words and music in song is that of Jerrold Levinson (1996). He notes three dimensions that can be analyzed in this relationship: the relation of the song’s text to its vocal line, the relation of a song’s text to its accompaniment, and the relation of the vocal line to the accompaniment. Levinson proposes that the ideal comportment between music and text is one of “mutual suitability” or “holistic working,” rather than internal matching or mirroring. He compares this relationship to that of a (happily) married couple, whose interaction with one another is mutually rewarding. In Peter Kivy’s “moderate indeterminacy” view, the text of a song particularizes the emotional expressivity of the music and therefore contributes to the music’s overall expressiveness (Kivy, 1989: ch. 10). Ridley (2004: ch. 3) criticizes both Kivy and Levinson for, despite outward assurances to the contrary, improperly treating songs as hybrids of music and text. Asking, as Levinson does, how closely the expressiveness of the music “matches” the emotional quality or tone of the text is to treat songs as a hybrid art form. Ridley, in opposition, argues that songs are a distinct musical form containing words. Any talk of “matching” is inappropriate because there is not one thing to match to another. Ridley notes that a text may have musical qualities before it is set to music, and that the same words read as a poem and sung in a song may have a different sonorous quality and emotional resonance. The best setting of a text to music is not one that “matches” the text in some way but rather one that shows that the composer has understood the text in question. Furthermore, music may particularize a text, just as much as text particularizes music. (For another discussion of these points, see Boykan 2000.)
There is a story that Alice Hammerstein, daughter of lyricist Oscar Hammerstein, was once at a social gathering where someone referred to the song “Ol’ Man River,” saying that it was “written by Jerome Kern.” She is said to have protested that no, her father wrote “Ol’ Man River,” and what Kern wrote was “Dum dum da dum.” This anecdote, whether apocryphal or not, nicely illustrates some of the ontological complexities inherent in song. Is a song to be identified with its text, its tune, or something else? What are a song’s identity conditions – do the words or the tune make a song the particular work that it is? Like many songs, the text of “Ol’ Man River” is often altered in performance. To what degree can the words be changed before we have a different song? Composers have notoriously “recycled” their work, sometimes using the same melodic line to fit very different texts. Different composers have set the same texts to different melodic lines. In such cases, how are we to determine what counts as the same song? What sort of entity is a song?
The answers to such questions will depend upon one’s prior ontological commitments and one’s understanding of the nature of musical works and artworks more generally. Those committed to the view that musical works are types or universals will have a different take on the ontology of songs than, say, those holding the view that works are abstract particulars. What seems uncontroversial, however, is that many songs are ontologically “thin” – that is, they have few determinative properties, and many qualities of a performance will be aspects of a performer’s interpretation, not of the work as such (Davies 2001: 20). This would seem to be especially true of folk songs, popular songs, and rock songs. Songs originated in non-literate societies and were originally part of oral traditions. This background influences the attitudes taken by audiences and performers alike. For example, many folk songs exist in numerous versions, and verses may be dropped or sung in a different sequence. Even songs that are works for performance may, depending on the musical tradition and audience expectations, be altered in performance and retain their identities. We are better off determining identity conditions for songs on a pragmatic, case-by-case basis.
Any act of singing – with friends at the pub, to a sleepy child, or in the course of a ritual – may be attended to as if it were a performance, but some acts of singing are intended to be listened to for their own sake. Because singers make music with their bodies, instead of or in addition to musical instruments, vocal performances have an element of subjectivity beyond that of solely instrumental performances. “The voice is the person” is both an overused metaphor for individual style and a matter of copyright law in many countries. Public performances of songs thus involve complex issues of gender, social ontology, and personal identity that are not so pronounced in instrumental performances. One consequence is that while some songs are aesthetically appropriate for any singer, in other cases incongruities may arise between what the song communicates and the singer’s public persona (Bicknell 2005).
Performances of songs imply a three-way relationship between a singer, a song, and an audience. As with any human interactions, these may carry moral obligations in addition to the aesthetic obligations inherent in performance (Bicknell 2009). Different conceptions of song meaning contribute to different performance practices among singers. The degree to which a singer (and accompanying musicians) will attempt to make the words of a song comprehensible to an audience will depend on the musical genre in question and on audience expectations. In some genres, the communication of the song’s melody and rhythm, and with it a dominant emotional mood, will be just as important as, or more important than, the communication of the song’s lyrics. In other musical traditions, such as contemporary country music, performers must strive to make each word fully audible and comprehensible.
Vocal performances manifest similarities and differences with other kinds of musical performances, and an adequate account of musical performance must take into account the complications raised by singing. Paul Thom makes a useful distinction between a “reading” and a “rendition” of a work for performance (1993: 76–8). A reading of a work is an understanding of its content, and a rendition is the execution of a work. Renditions are based on readings of works, and may or may not be executed as planned. (For example, a singer may plan a note-perfect performance yet be unable to execute it.) Inaccuracies can occur in the reading of a work, in the plan for the execution of a reading, or in the rendition itself. Performance without interpretation cannot be an artistic ideal because of the teleology of works for performance. They are understood, both by creators and by audiences, as calling for a certain kind of “playful” attention (Thom 1993: 30–2).
Similarly, Stan Godlovitch understands musical performance as “a complex activity which co-ordinates and focuses actions, skills, traditions, and works in order to define and create musical experience for the receptive listener” (Godlovitch 1998: 50). He places more emphasis on the performer–listener axis than on the performer–composer relationship. Performers have certain categorical obligations to listeners, and performances can fail by disaffecting listeners. Performers do not have unconditional obligations to composers, although a performance might certainly fail if it misrepresents a work.
Few thinkers have focused on the intricacies and complications of singing as opposed to other kinds of musical performances. Adam Smith (1723–90) offers an insightful analysis of singing performance that deserves to be better known. In keeping with the attitudes to music then current in Britain, Smith understands music as a form of artistic representation, such that individual musical works may represent the sentiments of a particular person in a particular situation. In the case of a vocal performance, an additional layer of representation is possible. The singer can, “by his countenance, by his attitudes, by his gestures, and by his motions,” convey the sentiments of the person whose situation is depicted in the song (Smith 1982: 194). The singer’s acting enhances the performance and is indeed necessary for a good performance. As Smith writes, “there is no comparison between the effect of what is sung coldly from a music-book at the end of a harpsichord, and of what is not only sung, but acted with proper freedom, animation, and boldness” (Smith 1982: 194).
The composer Edward Cone raises an intriguing question about singing performance and song meaning. In listening to, say, a performance of Schubert’s setting of a poem by Goethe, whose voice do we hear? He suggests four different answers. We hear the actual physical voice of the singer in question, the protagonist of the song, the poet whose words and images characterize the protagonist and the dramatic situation, and in the song’s musical accompaniment we hear the voice of the composer. Yet the composer is not simply one voice among the four. Rather, the voice of the composer constitutes the “complete musical persona,” and the vocal persona – the protagonist of the song – is properly understood as a character quoted by the complete persona. The complete musical persona of a song is not to be strictly identified with the composer but is “a projection of [the composer’s] musical intelligence, constituting the mind, so to speak, of the composition in question” (Cone 1974: 57).
Cone’s analysis of song meaning leads him to a specific account of song performance. Like Smith, he stresses the importance of dramatic impersonation. For example, when Marian Anderson sings “Swing Low, Sweet Chariot,” she recreates, as a dramatic persona, the slave who originally sang the song as an authentic appeal. An ideal or “faithful” performance is one in which the physical presence and vitality of the singer turns the persona of the musical text into an immediate living being. By contrast, in an illegitimate interpretation, it is the singer (rather than the vocal persona) who is seen as embodying and “composing” the song as he or she sings. The singer fails to let us hear the persona – and hence the composer’s voice behind the persona – speak for itself. Yet this is not quite the whole story for Cone. As human beings, singers must produce their own interpretations of songs and insist on their own freedom of action. But as dramatic characters, singers must be faithful to the text and to the dramatic situation. There is thus a tension between these two aspects of a singer’s role, and audiences are (or should be) aware of these tensions, as they are analogues for the tension between freedom and determinism in their own lives. Cone hypothesizes that the presence of such tensions can help explain the peculiar appeal of vocal performance, and of the performing arts more generally.
Finally, it should be noted that not all singing performances are renditions of works for performance or even of songs. Singers may also improvise new lyrics for an existing song, improvise lyrics for an instrumental melodic line (“voca-lese”), and singers in the jazz tradition may improvise nonsense syllables (“scat singing”) in the course of a song performance.
The phenomenon of singing continues to be both widespread and taken for granted. The practice has not atrophied, despite instrumental music’s “emancipation” from words, or the fact that ordinary non-musical verbal communication is more efficient. Composers continue to set words to music, and listeners persist in seeking out vocal music in all the genres in which it has a place. While Plato already in Book Three of his Republic raised the question of the power of song and argued for subjecting it to generally agreed moral imperatives, similar concerns continue to be raised about transgressive popular singing, as evidenced by the imposition of warning labels on some CDs. Why should song and singing have retained their importance? I suggest two possibilities, both related to the social nature of music and song.
First, thinkers in different traditions have described the ways in which communal singing serves to connect the subjective with the social. Victor Zuckerkandl (1973) writes that different interrelations between people are created by speaking and by singing. The spoken word presumes an “other” – the person spoken to, as opposed to the person speaking, who face each other as separate individuals. When tones are added to words and individual speech becomes communal singing, individuals who had previously faced one another are transformed into one group. Tones do not refer yet they are intended to be heard, both by the singers themselves and by others. When we sing as part of a group, we perceive the feeling of our own vocal activity within our bodies, and we hear the tones we make combining with those made by others around us. As Zuckerkandl writes, “the dividing line between myself and others loses its sharpness” (1973: 28). It seems reasonable that traces of these effects can be perceived even when we listen to others sing, and we join in with the singing imaginatively, if not audibly.
Second, singing is linked with the human desire for recognition and the obligation to recognize others. There is something ineluctably human about the voice. Hence the observation of Mladen Dolar that the impersonal or mechanically produced voice always has a touch of the uncanny (2006: 22). As Cone has argued, one cannot help but interpret a vocalist as a protagonist, rather than as the player of an instrument, even when the singer produces nonsense syllables: “For when the human voice sings, it demands to be heard, and when it is heard it demands recognition” (Cone 1974: 79).
See also Authentic performance practice (Chapter 9), Opera (Chapter 41), Popular music (Chapter 37), and Rock (Chapter 38).
Bicknell, J. (2005) “Just a Song? Exploring the Aesthetics of Popular Song Performance,” Journal of Aesthetics and Art Criticism 63: 261–70.
—— (2009) “Reflections on John Henry: Ethical Issues in Singing Performance,” Journal of Aesthetics and Art Criticism 67: 173–80.
Booth, M.E. (1981) The Experience of Songs, New Haven: Yale University Press.
Bowra, C.M. (1962) Primitive Song, Cleveland: World Publishing Co.
Boykan, M. (2000) “Reflections on Words and Music,” Musical Quarterly 84: 123–36.
Brown, S. (2001) “The ‘Musilanguage’ Model of Music Evolution,” in N.L. Wallin, B. Merker and S. Brown (eds) The Origins of Music, Cambridge: MIT Press, pp. 271–300.
Cone, E.T. (1974) The Composer’s Voice, Berkeley: University of California Press.
Darwin, C. (1981 [1871]) The Descent of Man and Selection in Relation to Sex, Princeton: Princeton University Press.
Davies, S. (2001) Musical Works and Performances: A Philosophical Exploration, Oxford: Clarendon Press.
Dolar, M. (2006) A Voice and Nothing More, Cambridge: MIT Press.
Godlovitch, S. (1998) Musical Performance: A Philosophical Study, London: Routledge.
Gracyk, T. (2001) I Wanna Be Me: Rock Music and the Politics of Identity, Philadelphia: Temple University Press.
Kivy, P. (1989) Sound Sentiment: An Essay on the Musical Emotions, Philadelphia: Temple University Press.
Levinson, J. (1990) “Hybrid Art Forms,” in Music, Art, and Metaphysics: Essays in Philosophical Aesthetics, Ithaca: Cornell University Press, pp. 26–36.
—— (1996) “Song and Music Drama,” in The Pleasures of Aesthetics: Philosophical Essays, Ithaca: Cornell University Press, pp. 42–59.
List, G. (1963) “The Boundaries of Speech and Song,” Ethnomusicology 7: 4–13.
Mithen, S. (2006) The Singing Neanderthals: The Origins of Music, Language, Mind, and Body, Cambridge: Harvard University Press.
Neubauer, J. (1986) The Emancipation of Music from Language: Departure from Mimesis in Eighteenth-Century Aesthetics, New Haven: Yale University Press.
Ridley, A. (2004) The Philosophy of Music: Theme and Variations, Edinburgh: Edinburgh University Press.
Rousseau, J.J. (1998 [1753]) “Essai sur l’origine des langues,” in E. Lippman (ed.) Musical Aesthetics: A Historical Reader, vol. 1, New York: Pendragon Press, pp. 323–37.
Smith, A. (1982 [1795]) “Of the Nature of that Imitation which takes place in what are called the Imitative Arts” in W.P.D. Wightman and J.C. Bryce (eds) Essays on Philosophical Subjects, Indianapolis: Liberty Fund, pp. 176–213.
Thom, P. (1993) For an Audience: A Philosophy of the Performing Arts, Philadelphia: Temple University Press.
Winn, J.A. (1981) Unsuspected Eloquence: A History of the Relations between Poetry and Music, New Haven: Yale University Press.
Zuckerkandl, V. (1973) Man the Musician, Princeton: Princeton University Press.
Clayton, M. (2008) Music, Words and Voice: A Reader, Manchester: Manchester University Press. (An interdisciplinary anthology about song and singing.)