This essay offers a novel theoretical perspective on the evolution of music. At present, a number of adaptationist theories posit that the human capacity for music is a product of natural selection, reflecting the survival value of musical behaviors in our species’ past (e.g., Wallin et al., 2000). In sharp contrast, a prominent nonadaptationist theory of music argues that music is a human invention and is biologically useless (Pinker, 1997). I argue that research on music and the brain supports neither of these views. Contrary to adaptationist theories, neuroscientific research suggests that the existence of music can be explained without invoking any evolutionary-based brain specialization for musical abilities. And contrary to Pinker’s claim, neuroscience research suggests that music can be biologically powerful. By biologically powerful, I mean that musical behaviors (e.g., playing, listening) can have lasting effects on nonmusical brain functions, such as language and attention, within individual lifetimes. Music is thus theorized to be a biologically powerful human invention, or “transformative technology of the mind.”
The past decade has witnessed a rapid rise in cognitive and neuroscientific research on music. This has led to renewed interest in evolutionary questions about music, which originate with Darwin’s discussion of the topic in The Descent of Man (1871). There are now several adaptationist theories arguing that musical behaviors originated via biological evolution due to their survival value for human ancestors. In contrast, nonadaptationist theories propose that musical behaviors are a human invention. The most prominent such theory, that of Steven Pinker (1997), regards music as a pleasure technology built from pre-existing brain functions (such as language, emotional vocalization, etc.), and posits, “As far as biological cause and effect are concerned, music is useless” (p. 528).
Pinker’s idea that music is an invention built from existing brain functions provides a useful null hypothesis for evolutionary debates over music. His assertion that music is biologically useless, however, is problematic. While Pinker was likely referring to music’s impact on human biology over evolutionary time, as opposed to within the lifetime of individual humans, his writing does not make this distinction. Furthermore, the metaphors he uses to describe music (e.g., “auditory cheesecake,” or “recreational drugs”) imply a view of music as having little biological significance at either evolutionary or individual timescales.[71]
A central point of this essay is that discussions of the biological significance of music should conceptually distinguish music’s effects over evolutionary time from its effects within individual lifetimes. The need for this distinction is driven by evidence from neuroscience. Neuroscientific research suggests that music is an invention that builds on diverse, pre-existing brain functions, rather than a trait that originated via processes of natural selection. This is consistent with Pinker’s thesis. However, growing evidence from neuroscience also suggests that music is biologically powerful, meaning that it can have lasting effects on nonmusical abilities (such as language or attention) during the lifetime of individual humans. Importantly, these effects can be observed not only in trained musicians but also in ordinary individuals who engage regularly with music. Thus, I believe that music should be regarded as a biologically powerful human invention or “transformative technology of the mind.” (For brevity, henceforth I refer to this idea as TTM theory.)
This essay is organized as follows. Section 2 introduces the evolutionary puzzle of music. Section 3 explains why neuroscience research suggests that music is an invention rather than a biological adaptation. Section 4 provides examples of the biological power of music. Section 5 suggests why music can have lasting effects on nonmusical brain functions. Section 6 provides a non-genetic explanation for why music is so pervasive in human culture. The essay concludes with a brief discussion of the relevance of a Darwinian perspective for the modern biological study of human music.
It is worth clarifying some points regarding TTM theory’s claim that music can shape brain function. It is obvious that engaging in any humanly-invented activity (e.g., kite flying) changes the brain within individual lifetimes, because learning and memory are instantiated by changes in neural networks, e.g., in the pattern of synaptic connections between neurons. Thus, TTM theory does not simply claim that musical behaviors change the brain. (This is trivially true: even learning a simple tune involves changing brain networks in some way in order to store the memory of the tune.) Nor does TTM theory simply claim that learning music results in lasting structural changes to the brain. (This claim would hardly be novel, given the growing evidence of experience-dependent changes in brain structure caused by learning a musical instrument, e.g., Hyde et al., 2009.) Rather, TTM theory claims that music is a human invention that can have lasting effects on such nonmusical brain functions as language, attention, and executive function, and is concerned with explaining the biological mechanisms underlying these effects.
The qualification of “lasting” effects is important, because this distinguishes TTM theory from theories concerned with the short-term effects of music on other cognitive abilities (e.g., Thompson et al., 2001). That is, TTM theory is concerned with musically driven neurobiological changes that impact other brain functions over the course of months or years, not over the course of a few minutes. In this regard, TTM theory has some parallels to neurobiological theories of reading, another human invention with salient impact on the brain within individual lifetimes (Dehaene and Cohen, 2007). Indeed, reading can be considered another transformative technology of the mind, because it is a human invention built from existing brain systems (such as those supporting visuospatial cognition and language) that impacts a variety of mental abilities (Mar et al., 2008; Patel, 2008:400; Dehaene, 2009).
Of course, music is much older and far more widespread than reading and appeals to humans from infancy. Also, unlike reading skills, basic musical abilities develop without any special instruction (Bigand and Poulin-Charronatt, 2006). These facts make the claim that music is a human invention seem odd. Yet other theories view ancient and universal human communication systems as inventions. For example, Tomasello (2008) has proposed that language originated as an invention based on communicative interactions between primates who had a special socio-cognitive ability for sharing actions and goals with others (“shared intentionality”; see also Lee et al., 2009). In common with such “language as invention” theories, TTM theory proposes that a complex and universal human trait can originate as an invention rather than as a biological adaptation. However, to my knowledge, all “language as invention” theories leave open the possibility that language, once invented, led to co-evolutionary changes in the brain that were aimed at supporting the acquisition of language (cf. Deacon, 1997). Indeed, the idea that our brains have been modified over evolutionary time to support the acquisition of language is favored by at least ten converging lines of evidence (Patel, 2008:358-366). TTM theory, in contrast, posits that there has been no evolutionary modification of our brains specifically aimed at facilitating musical abilities. Instead, music is viewed as a technology that is learned anew by each new generation of human minds. This view is congenial to the tremendous diversity of musical practices that have been described by ethnomusicologists (e.g., Titon, 1996; Nettl and Stone, 1998) and to the seemingly endless growth and development of music as a human art form (Ross, 2007). It is important to note, however, that TTM theory does not amount to the claim that humans are musical blank slates. Since music is theorized as building on preexisting brain functions (such as language and auditory scene analysis[72]), processing predispositions relevant to these other functions are likely to be reflected in the structure and processing of human music (cf. Reynolds, 2005; Dehaene and Cohen, 2007).
A final conceptual point about TTM theory concerning the fundamental question of why humans are drawn to musical behaviors merits discussion here. TTM theory claims that music can have lasting effects on nonmusical brain systems, but it does not propose that humans engage in music in order to produce these effects. Rather, as discussed in section 6 below, TTM theory posits that people are drawn to music because of its emotional power and because of its efficacy for ritual and memory. The lasting effects on nonmusical abilities are thus a consequence of how music engages the brain, not a cause of musical behavior. A better understanding of how and why these effects occur is of interest both for basic brain science and for designing musical activities to address problems in nonmusical domains, i.e., in scientifically-based music therapy (Leins et al., 2009).
Like language, music is a human universal that reaches deep into our species’ past (Nettl, 2000). Recent excavations have revealed bone flutes dating to the late Pleistocene era (~40,000 ybp, Conard et al., 2009). Cross-cultural and developmental research indicates that listening to and/or making music has a profound appeal to most members of our species, starting early in life (Blacking, 1973; Trehub, 2003). Thus, one can predict with some confidence that the few remaining uncontacted tribes of humans, when finally described by anthropologists, will have music as part of their behavioral repertoire.
For those interested in the evolutionary foundations of human behavior, such observations are puzzling. Musical activities lack any obvious survival value. Why then is music so pervasive in human life? Are we musical today because music helped our ancestors survive? Has the human mind been shaped by natural selection for music? Darwin (1871) was the first to wrestle with these questions, noting that “as neither the enjoyment nor the capacity of producing musical notes are faculties of the least direct use to man in reference to his ordinary habits of life, they must be ranked among the most mysterious with which he is endowed” (p. 1207).
In The Descent of Man, Darwin offered an adaptationist theory of music’s origins based on principles of sexual selection (see Kivy, 1959, for a discussion of these ideas in a larger historical framework). For the next century, scholarly discussion of music and evolution was relatively sparse but began to stir again with the rise of cognitive studies of music (e.g., Roederer, 1984). Interest in the topic has grown considerably in the past decade, reflecting the explosion of cognitive neuroscience research on music (Peretz, 2006). Indeed, since 2000, two scientific volumes of essays have been devoted to the evolution of music (Wallin et al., 2000; Vitouch and Ladinig, 2009), and the topic has been addressed in many other books and scholarly articles (e.g., Pinker 1997; Hauser and McDermott, 2003; Mithen 2005; Fitch, 2006, 2010; Hagen and Hammerstein, 2009; Kirschner and Tomasello, in press). Several adaptationist and nonadaptationist proposals are now in existence; some of the more prominent ones are reviewed below.
The first evolutionary theory for music was offered by Darwin in The Descent of Man (1871). Darwin drew an analogy with birdsong and theorized that music arose in our ancestors via mechanisms of sexual selection. He wrote: “Musical tones and rhythm were used by the half-human progenitors of man, during the season of courtship, when animals of all kinds are excited by the strongest passions” (p. 1209). Darwin speculated that wordless courtship songs predated our linguistic abilities and that such singing provided the scaffolding upon which language itself evolved. This idea of a musical protolanguage has proved of enduring interest to scholars researching the evolution of language and music (e.g., Brown, 2000(a); Mithen, 2005; see Fitch, 2010, for an overview and a recent version of the musical protolanguage theory). Indeed, the idea of a shared origin for language and music is pre-Darwinian, dating at least as far back as French enlightenment writings in the 1700s (Thomas, 1995). Commencing with Darwin, however, scholars have explored the idea within an evolutionary framework, proposing theories for how such a form of communication could have evolved and seeking to explain how it could further evolve into articulate language and fully developed music. Such theories view music as having a biological rather than a purely cultural origin and posit that musical behaviors had survival value for our ancestors.
This section focuses on the three most prominent adaptationist theories of music, based on sexual selection, parental care, and group cohesion. These theories have been proposed and explored independently but are not mutually exclusive. Indeed, musical protolanguage theories often invoke all three such theories to account for the biological origin of musical behavior.
As noted above, the sexual selection theory of music originated with Darwin. Sexual selection has the appeal of being able to explain the evolution of elaborate traits that seem nonadapative, or even maladaptive, in the daily struggle for existence, yet that are beneficial in the competition for mates (the peacock’s tail is a classic example). The sexual selection theory of human music has been explored by Miller (2000) and others and continues to attract interest.
A second set of adaptationist proposals concerns parental care rather than sexual selection. As often noted by biologists, human infants are born remarkably early in their biological development compared to other primates, possibly due to constraints on the size of the birth canal imposed by bipedalism (Mithen, 2005). Dissanayake (2008) and Falk (2004) have pointed to the cross-cultural importance of vocal communication in human infant care, whereby adults use melodious and rhythmic affect-laden utterances (“motherese”) to soothe or arouse prelinguistic infants. Positing that such vocalizations had adaptive value for infant survival, these authors propose that music has its origins in vocalizations aimed at caring for infant offspring.
A third set of adaptationist proposals concerns possible benefits of music to group cohesion. Humans, like most other primates, live in groups where individual competition is balanced with cooperation. Humans are unusual, however, in having relatively low degrees of in-group genetic relatedness (due to high gene flow between groups), yet depending to a large degree on in-group cooperation in order to survive and outcompete other groups (Richerson and Boyd, 2005). There has been much recent interest in the idea that music may have served as a mechanism to promote social cohesion within groups (e.g., Brown, 2000(b)). This idea was first clearly articulated by Roederer (1984), who pointed to “the value of music as a means of transmitting information on emotional states and its effect in congregating and behaviorally equalizing masses of people.” Dunbar (in press) has argued that group singing and dancing replaced physical grooming in ancestral human groups, when increasing group size made physical grooming of allies impractical. According to this view, song and dance led to endorphin release (mimicking the neural effects of physical grooming). This in turn promoted bonding, because endorphins, “as a byproduct of their role in pain control…have the property of making us feel warm and well disposed towards others who share…the experience that stimulates their production” (cf. Cohen et al., 2009, Kosfeld et al., 2005).
One appeal of the social cohesion idea is that music is often a social activity among humans, especially in small-scale cultures, and experimental work suggests that musical group activities promote cooperation between group members on subsequent nonmusical tasks (e.g., Wiltermuth and Heath, 2009; Kirschner and Tomasello, in press). Furthermore, music has certain design features that distinguish it from language, such as discrete pitches (allowing voices to blend together in song) and a distinct beat (enabling synchronized movement through time), which facilitate coordination between individuals and can promote a shared sense of identity and purpose (McNeill, 1995; Bispham, 2006).
Social cohesion hypotheses are currently a focus of much interest within music cognition, mirroring a growing interest within biology in natural selection at the level of social groups (e.g., Wilson and Wilson, 2007; Wilson et al., 2008). Several variant hypotheses have developed. Cross (2009), for example, draws on ethnomusicological literature and emphasizes music’s efficacy in managing situations of social uncertainty, i.e., situations in which linguistic interaction might give rise to conflict. He also emphasizes the role of music as a training ground for social cognition (cf. Boyd, 2009). Merker (2000), in contrast, draws on observations of chimpanzee group vocal displays and theorizes that music may have originated in our ancestors from synchronous calls aimed at mate attraction (see also Merker et al., 2009). Hagen and Hammerstein (2009) draw on comparative data from nonhuman primates and carnivorous mammals thought to be ecologically similar to human ancestors, and suggest that music may have arisen from group vocal territorial advertisements (for antecedents of this idea, see Geissmann, 2000).
Another version of the social cohesion hypothesis is notable for the relatively small degree of biological specialization for music that it proposes (Kirschner and Tomasello, in press). According to this view, music originated as an invention in ancestral human groups. Because music promoted group cohesion and survival, it acted as a cultural (vs. biological) adaptation, so that musically-oriented groups outsurvived other groups. Subsequently, due to feedback between cultural group selection and biological natural selection, there was selection for individuals who were biologically predisposed toward musical behavior.[73] Thus, according to Kirschner and Tomasello, modern humans are hypothesized to have “an innate proclivity for musical sounds and actions” without necessarily having any other brain specializations for music processing (cf. Trehub and Hannon, 2006).[74]
For the sake of brevity, a critique of the above theories is not provided here (the interested reader is referred to Patel, 2008: 368-371). For the current purposes, the relevant point is that all despite their different points of emphasis, all adaptationist proposals view the human mind as having been specifically shaped by evolution to support musical behavior.
In sharp contrast to adaptationist theories, nonadaptationist theories of music posit that there has been no natural selection for musical abilities in our species. Herbert Spencer implicitly took this position (even prior to the publication of Darwin’s Origin of Species) in his essay, “On the origin and function of music” (Spencer, 1857). Spencer argued that music grew out of the rhythms and cadences of impassioned speech and launched a debate that engaged Darwin and many other scholars (for a fascinating discussion, see Kivy, 1960, 1964, and Rehding, 2000).
Some thirty years later, William James voiced a nonadaptationist view of music in The Principles of Psychology (1890). James regarded the human love of music as “a mere incidental peculiarity of the nervous system” (Vol. 2, p. 419) and asserted: “It has no zoological utility…it is a pure incident of having a hearing organ…it has entered the mind by the back stairs, as it were, or rather [has] not entered the mind at all, but got surreptitiously born in the house” (Vol. 2, p. 627).
A modern descendant of James’ view is that of Pinker (1997: 528-538), which has become the most prominent nonadaptationist theory of music. Pinker’s proposal starts with the theory that many human mental faculties have been direct targets of natural selection. Music is chosen as a counterexample and is argued to be a human invention that is universal because of its link to pleasure: “Music appears to be a pure pleasure technology, a cocktail of recreational drugs that we ingest through the ear to stimulate a mass of pleasure circuits at once” (p. 528). In a later essay, Pinker (2007) elaborates this point to propose that music and many other human arts are “by-products of two other traits: motivational systems that give us pleasure when we experience signals that correlate with adaptive outcomes…and the technological know-how to create purified doses of these signals…” (p. 171).
Pinker’s proposal is notable for its specificity in suggesting the nonmusical foundations upon which music builds. These are: 1) the prosodic component of language, 2) auditory scene analysis, 3) emotional calls, 4) habitat selection, and 5) motor control.[75] According to Pinker, music brings us pleasure because it “tickles the sensitive spots” of these faculties. Specifically, 1) music has prosody-like properties, and the brain rewards the analysis of prosodic signals (patterns of linguistic rhythm and intonation) because prosody is an important component of language; 2) music is rich in harmonic sounds (sounds in which frequency components are integer multiples of some fundamental frequency), and the brain rewards the analysis of such sounds because harmonicity is an acoustic cue used to identify sound sources, an important part of auditory scene analysis; 3) music can evoke strong emotions because it contains pitch and rhythm patterns that resemble our species’ emotional calls, and 4) because it contains sound patterns reminiscent of evocative environmental sounds (e.g. “safe” or “unsafe” sounds, such as thunder, wind, or growls); 5) musical rhythm engenders rhythmic movement (e.g., in dance), and such movement is rewarded by the brain because rhythmic motor patterns are associated with biologically meaningful behaviors, such as walking, running, or digging.
Pinker’s proposal is much more detailed than that of James (1890), informed as it is by the century of cognitive science research that separates the two books. (For example, Pinker discusses in detail the influential music-cognition theories of Lerdahl and Jackendoff, 1983.) Nevertheless, James and Pinker arrive at a similar view of the biological significance of music. James wrote that music has “no zoological utility,” and Pinker asserts, “As far as biological cause and effect are concerned, music is useless.” Perhaps James and Pinker were referring to evolutionary utility as opposed to utility during the lifetime of individual humans, but their writings do not specify this. Furthermore, Pinker’s metaphor of music as a recreational drug implies a view of music as having a rather superficial relationship to human biology.
There are now several nonadaptationist theories of music, each offering distinct hypotheses about the brain systems upon which music builds. Livingstone and Thompson (2009), for example, argue that music builds on a recently evolved human theory of mind ability to serve the primary purpose of affective engagement. Panksepp (2009), in contrast, emphasizes music’s connection to evolutionarily ancient socio-emotional brain circuitry. There are other nonadaptationist proposals (e.g., Sperber, 1996), but none systematically considers music’s power to shape human brain function. It is on this point that TTM theory differs from existing nonadaptationist theories of music.
Given the debates over the evolutionary status of music, it is parsimonious to adopt the null hypothesis that there has been no natural selection for musical abilities in our species and then ask if there is enough evidence to seriously challenge this null hypothesis. When this strategy is applied to language, there appears to be enough evidence to refute the null hypothesis, as reviewed in Patel (2008: 358-366).
What of music? To some, the universal and ancient nature of human music may imply that it originated as a biological adaptation. The danger of such an assumption is illustrated by another remarkable human trait, namely the control of fire. This trait extends deep into our species’ past and is found in every human culture, yet few would dispute that it arose as an invention rather than a biological adaptation. The universality of the trait can be explained by the fact that it provides things that are universally valued by humans, including the ability to cook food, keep warm, and see in dark places. The example of fire-making teaches us that when we see a universal and ancient human trait, we cannot simply assume that it has been a direct target of natural selection (Patel, 2008: 356).[76]
It is tempting to think that brain specialization for certain aspects of music cognition (Peretz, 2006) and the existence of genetically-based deficits of music perception (Drayna et al. 2001; Peretz et al., 2007) point to natural selection for music. Yet upon closer examination, these facts provide no compelling support for adaptationist theories. Here, reading and writing provide useful analogies. These are indisputably human inventions, probably no more than about six thousand years old, making them too young to be associated with any evolutionary brain specialization for these abilities. Yet brain imaging studies of literate individuals have shown that certain aspects of reading, e.g., recognizing written characters, are associated with functional specializations in specific brain regions (Dehaene and Cohen, 2007; cf. Stewart et al., 2003). This specialization is clearly a product of experience-dependent neural plasticity, i.e., long-lasting changes in neurons and brain networks driven by experiences within an individual lifetime (Dehaene, 2009). Furthermore, certain reading disorders have a genetic component (Fisher and Franks, 2006), even though one can be confident that humans have not undergone natural selection for reading abilities. That is, specific genes can influence brain circuits that happen to be important for a complex human ability without any implication of natural selection for that ability.
The examples of fire-making and reading show that the evolutionary null hypothesis for music is not challenged by music’s universality, age, association with some degree of brain specialization, or its influence by specific genes. Challenges to the null hypothesis thus must come from other sources. Patel (2008: 367-400) reviewed a wide range of evidence in this regard, including data from neuroscience, infant studies, and animal studies, and argued that at present the null hypothesis for music could not be rejected.
Rather than rehearse those arguments here, sections 3.1 and 3.2 below take a different approach and illustrate two lines of research that support the idea of music as an invention. These studies illustrate a comparative approach to the evolutionary biology of music (McDermott and Hauser, 2003; Justus and Hutsler, 2005). The basic logic of this approach is as follows: If one can show that an aspect of music cognition is rooted in other, nonmusical human brain functions or is shared with other species, then it is parsimonious to assume that this aspect has not been shaped by natural selection for music. This approach is particularly powerful when applied to aspects of music which seem domain-specific, i.e., not related to other types of cognition, such as tonality processing and synchronization of movement to a musical beat (Peretz and Coltheart, 2003; Bispham, 2006).
Most of the world’s musical systems use discrete pitches and intervals to create melodies, with the pitches drawn from musical scales of five to seven tones per octave (Reck, 1997). A widespread feature of music is the differential use of scale pitches such that some are perceived as more stable or structurally significant than others (Krumhansl, 1990). This differentiation of scale pitches in terms of stability or prominence has been termed a “tonal hierarchy,” and implicit knowledge of such hierarchies develops without any special musical training (Tillmann et al., 2000). This knowledge contributes to our subjective impressions that tones in a musical context have abstract perceptual properties, such as