Many animals communicate using sound. The sounds involved are often short and simple calls, like the croaking of frogs or roaring of lions. But they may be much longer and more elaborate, and are then usually referred to as ‘song’ by analogy with those produced by humans: examples are the trilling of crickets, the marvellously evocative songs of whales and birdsong with its great diversity. Many composers have been inspired by such sounds, notably those of birds (see Birdsong).
Whether or not the sounds produced by animals should be classed as music is a more complex issue. It may nevertheless be useful to consider such questions as why animals sing; whether animal songs can in any sense be regarded as examples of music; and whether they might shed light on the origins of the music of the animal species Homo sapiens. Analogies that have been drawn between structural factors in human music and, especially, birdsong (for example in transposition, inversion, rhythmic variation and the use of motifs) invite further enquiry (see Scholes, 1938; Mâche, 1983).
The use of sound as a means of communication for animals has several advantages. As with human language, much detailed information can be packed into a short sequence of sound signals, and these travel rapidly through air or water in all directions, by day or night, little affected by obstacles. Other modes of communication are usually less efficient. Smells diffuse slowly and largely downwind; they cannot be rapidly changed. Visual signals move at the speed of light and have the advantage that they can be quickly altered, but they are disrupted by obstacles in their path and usually depend on light, so are of little use in a dense environment or at night. Given these considerations, it is not surprising that sound is the channel of communication most often employed for complex animal signals, particularly where these involve advertising, where it is important that the signal covers the widest possible area. A particularly common use of sound by animals is in advertising for mates: the greater the area reached by the signal, the more likely it is to be received by a member of the opposite sex.
Simple animal sounds, such as call notes, convey a variety of messages. Examples are the calls used by members of a flock to maintain contact, or the alarm signals produced by certain animals when they spot a predator, which lead others to ‘freeze’ or seek cover. Some such calls are ‘referential’, like words in a language. Thus a vervet monkey (Cercopithecus aethiops) has different calls for eagle, snake and leopard, and other vervets hearing these calls behave appropriately (Seyfarth, Cheney and Marler, 1980). To the snake call they look down and approach with caution, to the eagle one they rush from the trees and into the thickets, while to the leopard alarm they run up into trees. The three calls might as well be words representing these three different kinds of animal.
Such instances, however, are rare. Much animal communication is affective rather than referential, representing emotional states rather than particular objects in the outside world, and in this respect it is more akin to music than to language. Further, the elaborate songs of some birds, in which each individual may have several hundred or even thousand different phrases, do not convey many different messages. It seems that this variety has evolved for its own sake to convey the same message in many different and perhaps more persuasive ways rather than because there are different messages to transmit.
Why do animals sing? Two main reasons have been proposed and there is good evidence for both of them (Catchpole and Slater, 1995). A clue is provided by the fact that song in birds, at least in temperate regions, is usually produced only by males and is restricted to the breeding season. At this time males, for example many songbirds, fight for territories. They often have duels in song across boundaries and, if the birds involved have repertories of different phrases, each tends to match the song of the other as they sing. A male redwinged blackbird (Agelaius pheoniceus) that cannot sing suffers more intrusions on to his territory (Smith, 1979). If a male great tit (Parus major) is removed from his territory, it will be less rapidly invaded if recordings of the song of his species are played from loudspeakers within it (Krebs, 1977). Such evidence points to song having a role in rival repulsion.
Repelling rivals with song may explain why some birds have a small number of distinct song phrases, for this enables them to match different intruders. Many such birds sing the same song several times in succession before moving on to the next and, unless countersinging with a rival, they will often cycle through their whole repertory before returning to the first song (e.g. the chaffinch, Fringilla coelebs: Slater, 1983).
Repeating a song several times may ensure that its particular message gets across, and that could be important where males match each other’s songs precisely. On the other hand, the need to repel rivals does not account for the extreme elaboration of song found in some other species. Here the most likely explanation lies in the second proposed reason why animals sing: that it attracts females. Males of many birds stop singing once they are mated (e.g. the sedge warbler, Acrocephalus schoenobaenus: Catchpole, 1973), and song increases enormously if a male loses his partner (e.g. the great tit: Krebs and others, 1981). In several species song has been found both to attract females (Eriksson and Wallin, 1986) and to stimulate them to build nests and lay eggs (Kroodsma, 1976). One might suppose that a simple song labelling the male as belonging to his species, so that only the right females are attracted, would be sufficient. This may be so in some cases, where a simple little song fulfils that function; but in others females are known to be most attracted by males with large repertories of different songs (Catchpole, 1980; Eens, Pinxten and Verheyen, ‘Male Song’, 1991). If, for whatever reason, females prefer males with larger song repertories, the prize will go to the male with the most elaborate song. This process of runaway inflation, known as sexual selection and originally described by Charles Darwin (1871), is thought to be a prime reason why animals have large repertories of different sounds. The message of each sound is the same: ‘I am a male sedge warbler in breeding condition’; but the male that can say it in the most varied way is more attractive to females and thus most likely to be successful in leaving his genes to the next generation. In such birds, unlike those with small repertories, it is much less common for the same song to be repeated several times consecutively, for the main message is variety itself (Slater, 1981).
The sedge warbler is a good example of a bird with a large song repertory (Catchpole, 1976). Each male’s song is composed of relatively few elements but these are combined to make a complex whole. At the start, two elements alternate; then numerous others are introduced in quick succession in the middle of the song. Two of these are then selected to alternate in the closing section. These two then introduce the next song, which starts after a brief pause. This way of combining a small repertory of elements leads to an almost infinite number of possible songs, but just how varied they are depends on the number of elements that a male has: the effectiveness of a male in attracting and stimulating females depends on his element repertory. Males with a large repertory both attract females earlier in the season and stimulate them to mate more effectively (Catchpole and others, 1984).
Unlike the sedge warbler, nightingales have fixed repertories of song types, each of which is near identical every time it is produced (Todt and Hultsch, 1996). But the repertory itself is very large, usually consisting of over 200 different phrases. A bird tends to cycle through its repertory, though it misses out many of the phrases each time it goes through the sequence; thus the same phrase rather seldom occurs twice close together, but will often recur after some 70 or 80 others. What the nightingale seems to be doing, as is the sedge warbler but in a different way, is maximizing the variety of its output. That is exactly what sexual selection would lead us to expect.
The song of the humpback whale (Megaptera novaeangliae) is not unlike that of the nightingale in its patterning, with a series of themes through which each animal cycles, although each theme tends to be repeated several times (Payne and McVay, 1971). Remarkably, all the animals in a population share these themes, yet the songs change during the singing season, some dropping out of the population’s repertory and new ones being introduced (Payne and Payne, 1985). Cultural change is a notable feature of learnt vocalizations in songbirds (and of course in humans) as well as in whales. However, the whale example is unusual in that changes take place within a single season, each animal modifying its song in synchrony with the rest of the population. (In birds, it is usually only young males in their first year that copy from others.) Changes may take place at this stage: a note may be miscopied, or two songs may be blended, so that new songs are created. However, once learnt, the structure of the songs tends to be fixed, and this is true even for those birds that learn new songs each year throughout their lives. What is striking – and this contrasts markedly with human music – is that there is little evidence for innovation or improvisation. While birdsong may sound to the casual listener to be endlessly varied, that impression is created from a fixed, albeit sometimes very large, repertory of sounds.
One dramatic aspect of song learning in many birds is mimicry, the copying by one species of another. A good example is the European starling (Sturnus vulgaris) in which each male, in addition to the distinctive whistles and rattles that make up its species-specific song, will incorporate the songs of several other species (Eens, Pinxten and Verheyen, ‘Organisation of Song’, 1991). However, the most remarkable case must be that of the marsh warbler (Acrocephalus palustris), a small European bird which migrates to East Africa in the autumn. Adults cease to sing before their chicks hatch, but the young birds learn the sounds of many other species from Europe and Africa during their first winter and incorporate them into a song which, while of distinctively marsh warbler patterning, is largely or entirely based on mimicry (Dowsett-Lemaire, 1979). On average, a male marsh warbler mimics some 77 other species. It is not known why birds mimic, but it seems to be a way of building up a varied repertory. It appears to be easier for birds to copy the sounds that they hear than to generate variety by improvisation.
Might our understanding of these complex animal sounds shed light on the origins of human music? Any similarity is more likely to be by analogy than because of any shared musical ancestry with other singing animals. The closest living relatives of humans, the great apes, communicate more by gesture and facial expression than by sound. They have loud vocal displays, such as the ‘pant hoot’ of chimpanzees (Pan troglodytes), but these are far from elaborate, stereotyped or musical. Further, there is little evidence that any monkey or non-human ape learns the sounds that it produces from other individuals. Humans do so; and whales and songbirds, the most notable singers among animals, also copy their sounds from other individuals. Indeed, learning seems essential to the building-up of large song repertories. For some reason, therefore, elaborate singing behaviour has arisen separately in different parts of the animal kingdom; in the case of humans this was in the relatively recent past, since the time of the common ancestry with chimpanzees about two million years ago.
Straight comparison may not be justified, but analogy with birds or whales may help to suggest why singing and other musical activities may have arisen in humans. With any complex or varied display, sexual selection is a prime suspect and the fact that in many cultures singing (and in our own, composition) is predominantly an activity of young males confirms that suspicion. However, why singing behaviour should have been favoured in early man in particular rather than in other species remains a matter of speculation. The singing of humans also has features, such as the simultaneous chanting of the same tune by groups of individuals, which have not been described among animals.
Do animals produce ‘music’? This partly depends on how ‘music’ is defined (see Music). Many animal sounds are rhythmic, such as the trill of a stridulating grasshopper; others are pure and tonal, such as the whistles common in birdsongs. Energy efficiency alone might predict these features: a regular rhythm is shown by a mechanism operating at its resonant frequency, and this is where the energy cost is least. Concentrating all the energy in a narrow frequency band to produce fairly pure sounds is also economical, as such sounds carry further. But there are other good reasons why rhythmical and tonal sounds may have arisen in the animal kingdom. For the great majority of animal signals, and especially those concerned with mate attraction and rival repulsion, the signal must incorporate species identity. In some areas of the world, notably tropical rainforest, there may be hundreds of bird species in a small area; to stand out against both this cacophony of sound and other environmental noises, and to be distinctive, may impose features such as tone and rhythm as each species homes in on a unique broadcasting bandwidth. The complex patterns of songs, and species differences in the rules that underlie them, may also have their origins in the need for distinctiveness.
While animal signals need to be distinctive and clear, there are further features that those carrying similar messages tend to have in common. For example, many small bird species have a thin, high-pitched ‘seeep’ alarm call which they produce when they spot a hawk. This is not because different species are communicating with each other but simply because characteristics of this particular sound make it very hard to locate (Marler, 1955). It functions well in indicating to others of the caller’s own species that there is a hawk about, but it does not encode the species of the caller (nor does it need to) and it minimizes danger to the caller by being very hard for the hawk to locate. For all these reasons the alarm calls of different species have evolved in the same form.
At a more general level, it has been pointed out that similar sounds in different animals may very often convey similar messages (Morton, 1977). For example, deep and gruff sounds tend to be aggressive and hostile while pure and high ones are more affiliative and friendly. Probably there are two factors involved. First, only large animals can make deep sounds, so the deeper the sound the more intimidating it will be to smaller individuals, who would be well advised to retreat rather than risk a fight. Secondly, to be easily understood by other individuals, signals should be as distinctive as possible (the so-called ‘principle of antithesis’: Darwin, 1872). Probably, friendly signals that are pure and high have come to contrast maximally with hostile ones simply to preclude confusion. Given these arguments, it may be more than just by analogy that deep, loud and ponderous sounds tend to be aggressive and threatening in both animal communication and human music. The parallels in emotional expression between animal sounds and music may indeed go further: it has been argued that there are similarities between the rules underlying sequences of calls in Arabian babblers (Turdoides squamiceps) when they are excited and when they are calm and those both in equivalent speech situations and in musical counterpoint (Cohen, 1983).
Does animal musicality go any further than this? It is not difficult to find examples in animal song of complex features that can also be attributed to human music. Some birds sing in near-perfect scales (e.g. the musician wren, Cyphorhinus aradus); some pairs show antiphonal duets (Levin, 1996); and some groups sing in chorus (Brown and others, 1988). Humpback whales with large song repertories often start or end different phrases with the same subsection, a feature that has been likened to rhyming (Guinee and Payne, 1988); indeed it has been suggested, somewhat speculatively, that the common features of successive themes may help the whales to memorize the long sequence of sounds that they sing, as rhyming is known to do in humans. Caution is required here, as it is easy to jump to conclusions from chance similarities. Considering only the songbirds (Oscine passerines), there exist close to 4000 species; all are thought to learn their songs. The variety in the form and patterning of these songs is impressive and there can be few possibilities that remain unexplored given this huge array of species. So it would not be surprising if almost any characteristic found in human music were discovered in one or more of them. Such similarities are likely to be coincidental, attributable to convergence rather than to musical features arising in a common ancestor. Nevertheless, while animals may not share music in the strict sense with human beings, there is no doubt that some of them have complex and beautiful vocal displays. Understanding why these have evolved may help to shed light on why human beings, uniquely among the primates, have taken a similar pathway.
It has been suggested from time to time that the songs of some birds, which seem to us especially beautiful, may be more so than is strictly necessary for their biological function (Thorpe, 1961; Boswall, 1983): could this indicate some primitive aesthetic sense, and that the bird takes pleasure in song for its own sake? Candidate songs here would be that of the song thrush (Turdus philomelos) in Europe, the superb lyrebird (Menura novaehollandiae) in Australia and the mockingbird (Mimus polyglottos) in North America; all these have large, varied and beautiful repertories. It is difficult to test such ideas. Sexual selection is an open-ended process, which will lead to larger and larger song repertories until other constraints, such as storage space in the brain, set limits. Where it is responsible, it is unlikely that song could be demonstrated to be more elaborate than sexual selection demanded. On the other hand, there is nothing incompatible between this and either aesthetics or the enjoyment of song; indeed, sexual selection is likely to have been the basis for their evolution in humans. But that is where the problem of testing comes in. We know that humans feel enjoyment in hearing or performing music; we can ask them about it and discuss their feelings with them. When it comes to animals, however, we have no access to their inner feelings, so that the question can only be a matter of speculation.
P.J.B. SLATER
C. Darwin: The Descent of Man and Selection in Relation to Sex (London, 1871)
C. Darwin: The Expression of the Emotions in Man and Animals (London, 1872)
P.A. Scholes: ‘Bird Music’, Oxford Companion to Music (London, 1938, 10/1970)
P. Marler: ‘Characteristics of some Animal Calls’, Nature, clxxvi (1955), 6–8
W.H. Thorpe: Bird Song (Cambridge, 1961)
R.S. Payne and S. McVay: ‘Songs of Humpback Whales’, Science, clxxiii (1971), 585–97
C.K. Catchpole: ‘The Functions of Advertising Song in the Sedge Warbler (Acrocephalus schoenobaenus) and Reed Warbler (A. scirpaceus)’, Behaviour, xlvi (1973), 300–20
C.K. Catchpole: ‘Temporal and Sequential Organisation of Song in the Sedge Warbler (Acrocephalus schoenobaenus)’, Behaviour, lix (1976), 226–46
D.E. Kroodsma: ‘Reproductive Development in a Female Songbird: Differential Stimulation by Quality of Male Song’, Science, cxcii (1976), 574–5
J.R. Krebs: ‘Song and Territory in the Great Tit Parus major’, Evolutionary Ecology, ed. B. Stonehouse and C. Perrins (London, 1977), 47–62
E.S. Morton: ‘On the Occurrence and Significance of Motivation-Structural Rules in Some Bird and Mammal Sounds’, American Naturalist, cxi (1977), 855–69
F. Dowsett-Lemaire: ‘The Imitative Range of the Song of the Marsh Warbler Acrocephalus palustris, with Special Reference to Imitations of African Birds’, Ibis, cxxi (1979), 453–68
D.G. Smith: ‘Male Singing Ability and Territory Integrity in Redwinged Blackbirds (Agelaius phoeniceus)’, Behaviour, lxviii (1979), 193–206
C.K. Catchpole: ‘Sexual Selection and the Evolution of Complex Songs among Warblers of the Genus Acrocephalus’, Behaviour, lxxiv (1980), 149–66
R.M. Seyfarth, D.L. Cheney and P. Marler: ‘Vervet Monkey Alarm Calls: Evidence of Predator Classification and Semantic Communication’, Science, ccx (1980), 801–3
J.R. Krebs, M. Avery and R.J. Cowie: ‘Effect of Removal of Mate on the Singing Behaviour of Great Tits’, Animal Behaviour, xxix (1981), 635–7
P.J.B. Slater: ‘Chaffinch Song Repertoires: Observations, Experiments and a Discussion of their Significance’, Zeitschrift für Tierpsychologie, lvi (1981), 1–24
J. Boswall: ‘The Language of Birds’, Proceedings of the Royal Institution of Great Britain, lv (1983), 249–303
D. Cohen: ‘Birdcalls and the Rules of Palestrina Counterpoint: Towards the Discovery of Universal Qualities in Vocal Expression’, Israel Studies in Musicology, iii (1983), 96–123
F.-B. Mâche: Musique, mythe, nature, ou, les dauphins d’Arion (Paris, 1983; Eng. trans., 1992)
P.J.B. Slater: ‘Sequences of Song in Chaffinches’, Animal Behaviour, xxxi (1983), 272–81
C.K. Catchpole, J. Dittami and B. Leisler: ‘Differential Responses to Male Song Repertoires in Female Songbirds implanted with Oestradiol’, Nature, cccxii (1984), 563–4
J. Hall-Craggs and R.E. Jellis: ‘Birds in Music’, A Dictionary of Birds, ed. B. Campbell and E. Lack (Calton, Staffs., 1985), 369–72
K. and R.S. Payne: ‘Large Scale Changes over 19 Years in Songs of Humpback Whales in Bermuda’, Zeitschrift für Tierpsychologie, lxviii (1985), 89–114
D. Eriksson and L. Wallin: ‘Male Bird Song attracts Females: a Field Experiment’, Behavioural Ecology and Sociobiology, xix (1986), 297–9
E.D. Brown, S.M. Farabaugh and C.J. Veltman: ‘Song Sharing in a Group-Living Songbird, the Australian Magpie, Gymnorhina tibicen: Part I, Vocal Sharing within and among Social Groups’, Behaviour, civ (1988), 1–28
L.N. Guinee and K.B. Payne: ‘Rhyme-Like Repetitions in Songs of Humpback Whales’, Ethology, lxxix (1988), 295–306
M. Eens, R. Pinxten and R.F. Verheyen: ‘Male Song as a Cue for Mate Choice in the European Starling’, Behaviour, cxvi (1991), 210–38
M. Eens, R. Pinxten and R.F. Verheyen: ‘Organisation of Song in the European Starling: Species Specificity and Individual Differences’, Belgian Journal of Zoology, cxxi (1991), 257–78
C.K. Catchpole and P.J.B. Slater: Bird Song: Biological Themes and Variations (Cambridge, 1995)
R.N. Levin: ‘Song Behaviour and Reproductive Strategies in a Duetting Wren, Thryothorus nigricapillus’, Animal Behaviour, lii (1996), 1093–1117
D. Todt and H. Hultsch: ‘Acquisition and Performance of Song Repertoires: Ways of Coping with Diversity and Versatility’, Ecology and Evolution of Acoustic Communication in Birds, ed. D.E. Kroodsma and E.H. Miller (Ithaca, NY, and London, 1996), 79–96
M. Head: ‘Birdsong and the Origins of Music’ JRMA, cxxii (1997), 1–23