7.1. Introduction

Every human society uses music in some or other form, and music must therefore be assumed to be an ancestral trait in humans, like other art forms. It is natural then, to speculate about the origin of music in the context of evolutionary history. Compared to other art forms, however, the problem of music is that it does not fossilize well or leave other permanent traces. The traveller was able to form an impression of Ozymandias from the shattered visage and inscription in the desert, but the rocks were silent. When standing in the ruins of a monastery like Cluny, so important in the development of Western polyphony, the traveller could have no impression of the music once resounding there, were it not for the development of a notational system that has preserved the rudiments of musical practice. Therefore, (too) much weight is often laid on the archaeological remnants of musical instruments, e.g. old wind instruments like the much debated Divje Babe flute, found in a Neanderthal settlement in present Slovenia. Music could have originated in human vocalization (e.g. song) long before any instruments were invented, however, and the early history of music will probably always remain obscure. The best chance we have to understand the history is to integrate information from many different fields – ethnomusicology, archaeology, cognitive science and neurophysiology. A valid attempt at such a synthesis is offered in the book The Singing Neanderthals by the archaeologist Steven Mithen, and while the synthesis offered is necessarily speculative, it is well-argued. His evolutionary narrative in my opinion is one of the most interesting presented so far.

Mithen reviews the current theories of the origin of music in an evolutionary context. Here, one speculates about the evolutionary meaning of music, i.e. in which ways music could increase the chances of survival of its protagonists and their offspring. Such selection benefits could accrue in many different ways, both as increased reproductive success of the individual, as featured in sexual selection hypotheses, through improved communication between mother and infant or through benefits to a group of (related) individuals, for example by furthering social structure and group cohesion. Central in all discussions of human evolution is the role of language. There is a complex interdependence between language and social structure: A stable society is beneficial for language acquisition, but language also promotes social structure. If music has had a role in stabilizing early human society, it could have influenced both society and language.

7.2. Music and Language.

It is clear that music, especially song, is closely related to human language. Language is such a complicated phenomenon that this observation in itself is not really helpful, however, and some researchers, notably Steven Pinker (1997), have seen music as a spin-off from the language processing functions of the brain. This idea is dismissed by Mithen, partly because music is such a universal human phenomenon, and partly because there are plausible adaptive hypotheses.

The first chapters describe the relationship between music and language and the underlying neural processing. Much of the information here is relatively well-known, mostly coming from lesions either affecting ‘language’ or music processing. The results are not very clear-cut, but indicate a certain dissociation between music and language, and according to the findings of Peretz (2002) a certain modularity of music processing, although centers dedicated to music processing have not been convincingly demonstrated. Also, there are many different aspects of language – semantics, syntax, prosody to name a few, and it is likely that a defect in processing with regard to one or the other of these aspects will not affect music processing to the same extent. Mithen surprisingly does not mention a very interesting finding by Robert Zatorre and co-workers (Zatorre et al. 2002) suggesting a specialization of the left and right temporal lobe; in most subjects the left predominantly responds to high-frequency modulation of tones and the right to low-frequency modulations. High-frequency modulation characterizes the fast envelope fluctuations in speech that carry most of the semantic content, whereas low-frequency modulation characterizes not only prosody, but also melodic contours. I think that the connection between music and prosody (in the sense of speech melody) that Mithen endorses is very convincing and backed by these studies (see also Christensen-Dalsgaard 2004). Another strong example of the connection between music and prosody is mother-infant communication, excellently reviewed in ch. 6. As shown by Colwyn Trevarthen (1979), Dissanayake (2000) and others, the infant-directed speech (IDS) of parents is characterized by “exaggerated prosody”: it is slower and exhibits steeper pitch contours than normal speech. Furthermore, studies reviewed in the chapter show that infants will respond appropriately to IDS even when the semantic content is filtered out, i.e. to prosody only.

7.3. Music and Emotion

It is often stated that music is the language of emotions. Unfortunately, what is meant by “emotion” is never very clearly defined. In ch. 7, it is stated that all emotions are “variants of happiness, anger, sadness, fear and disgust.” That leaves out basic urges as hunger, thirst and sexual desires, but also a very important feeling – that something is just and right. In other words, I think this definition of emotions may be too restrictive. A much more inclusive definition would be that emotion was all human thinking that could not be characterized as rational. In that case, it is clear that music communicates emotion. I am strongly sceptical of the claims of specific tone intervals being able to precisely and universally communicate emotion, as proposed by Cooke (1959, discussed in Christensen-Dalsgaard 2004). Firstly, music does not work that way for me (or for the people I have asked in an unofficial survey) – rather than communicating specific emotions like happiness, anger etc, what I sense is being communicated is a feeling of the music being “well made” or “just and right.” Secondly, the examples cited by Cooke to show that specific intervals have specific emotional meaning are not universal and can be contradicted by examples from medieval and renaissance music. Thirdly, if there were a well-defined emotional code, is it not surprising, that it is never mentioned in the composers’ usually very straightforward instructions with regard to compositional technique? The Affektenlehre in baroque music, as featured e.g. in Matthesohn (1739), is often held in support of well-defined emotional codes; the aim here, however, was rather to depict emotion (the ravings of an angry man, for example) than to communicate emotion (i.e., in this case, to make the listeners angry), in my opinion.

7.4. Sound Communication in Primates

To get an idea of the sound communication of our ancestors Mithen reviews sound communication in our nearest relatives among the primates (ch 8.). The chapter is a good review of the diverse types of communication in the primates. It is notable that the vocal communication of our nearest relatives – the great apes – is fairly simple with few distinct call types. In other, more distantly related species, however, such as the celebrated vervet monkeys, communication is more sophisticated. In a fascinating study, Struhsaker, Cheney and Seyfarth (Cheney and Seyfarth, 1990) were able to show that vervet monkeys have distinct calls for different predators, corresponding to different behavioral responses. The alarm call for a snake, for example, induces other vervets to raise themselves high on their legs and peek into the ground, whereas the alarm call for a hawk induces the monkeys to seek cover under trees. The calls are distinct and categorical. They are also, as Mithen claims quoting Alison Wray, holistic, however, in the sense that they cannot be combined or constructed from meaningful subunits – the snake call means ‘beware of the snake’ (or maybe more accurate even ‘beware of dangerous animal creeping in grass’). Therefore, this is not a language in the sense that it enables the monkeys to refer in the abstract to snakes, leopards or hawks. In these calls, like in most other animal calls, there is also a reference to the emotional state of the animal – whether it is scared, angry etc. It should be noted, though, that no animal would benefit from communicating its emotional state reliably –rather, what is communicated is a filtered, sometimes deceptive version of its emotional state.

Other primates have very long calls used in social interactions. The most celebrated case, the great call of the gibbons can last between six and eighty minutes; these calls, like the calls of gelada monkeys are tonal and therefore sound ‘musical’ with similar use of rhythm and melody as in human speech and song. It is unclear, however, whether this is just caused by analogous use of similar vocal systems in humans or geladas (there are limited ways to excite the physical structures in the vocal system) or whether the similarity is an ancestral property. It would have been useful to discuss the calls in relation to environmental acoustics. Generally, tonal calls propagate much better in cluttered environments such as rain forests than more bark- or grunt-like calls, and the different call types may be specializations for long vs. close range communication. The question of range is highly relevant in the context of music (song), and it might have been helpful here if Mithen had tried to delineate the acoustical differences between speech and song. It is clear that there are two major differences: duration and the relative importance of vowel sounds. Most speech elements are short (50 ms or less) compared to the basic rhythm of song (0.5 to 1 s), and song can be described roughly as a way to increase the sound energy by lengthening the vowels whereas normal speech essentially is a short-range signal. This observation already suggests one function of song – long distance communication at the expense of speed.

Turning now to the evolution of early man, Mithen reviews the archaeological/fossil evidence in three major areas: evolution of language, cognition and bipedalism. There are many interesting points here; one that I find particularly interesting is the hypothesis that bipedalism itself had great implications for rhythm (by freeing up the upper part of the torso, the arms and especially the hands to perform independent movements) and maybe also for song (changes in the attachment of the spine to the cranium may have caused changes in the position of the larynx). The keystone of the book is Mithen’s hypothesis that early human ancestors and also the Neanderthals did not have a symbolic language, but rather a ‘musical’ form of vocal communication that Mithen calls ‘hmmm’ (for holistic (not composed of meaningful subunits), multi-modal (both sound and gesture), manipulative and musical). This type of communication is believed to be used in mother-infant interactions as in recent humans, but also for ritualistic group activities. This communication type is similar to the musilanguage proposed by Brown (Brown, 2000), and is very likely to have existed in our ancestors, since this is essentially the communication found in other primates (except that the early human ancestors are assumed to have had a much larger vocal repertoire). Direct evidence for the assumptions is, of course, meager. In particular, I am not convinced that the early humans should have had no symbolic language. The argument is based on the absence of symbolic art like cave paintings from the Neanderthals, and the observation that their society seemed to be static. It is tenuous to conclude from the absence of evidence. The Neanderthals might have decorated perishable objects or even had – in line with Mithen’s view of their musicality – song and dance rites that would leave no trace. Secondly, since Hmmm- communication is generally similar to communication in other primates, it is unclear to me what the special musical feature of the communication should be. After all, any animal communication could be called musical. What does the hypothesis really tell us about music? Music as we know it today is neither holistic, nor multimodal, necessarily. Conversely, music is characterized by subunits that can be combined in many different ways, and the existence of musical instruments (that are devices precisely for producing many different combinations of these subunits) suggest that this non-holistic approach is ancient. Furthermore, as a means of communication about anything, music is highly ambiguous. Finally, I think Mithen exaggerates the group synchronization element in music. Certainly, much of the music making in Western culture has been solistic, for example in the form of long, epic chants. Let us now assume, however, that there was such a Hmmm musilanguage among our early ancestors, as stated above; a safe assumption given what we know about primate communication. It is likely that this ancient communication system would have been retained when the “modern” syntactical and combinatorial language evolved, but not unchanged of course; it probably co-evolved into what we know as prosody, the speech melody that carries important information not only about emotion, but also about syntax (delineating the elements of speech). Therefore, we have no access to the original Hmmm language, and if music is related to prosody, it probably also changed beyond recognition. Evolution provides many examples of structures whose function becomes completely changed through evolutionary history, so common ancestry is not necessarily any clue to function (and, conversely, common function is not a clue to common ancestry).

One problem here is that so very little is known about the Neanderthals, and one would be on safer grounds talking about music among our immediate ancestors, where we at least can infer many details about language, symbolism and general cognition from the shared similarities among recent humans. The shift in focus to the Neanderthals seems to me to introduce an unnecessarily wide berth for speculation.

I find the book very stimulating and highly readable nevertheless. Mithen concludes by requesting the reader to listen to music and even to make his/her own music. I deeply sympathize with this; one of the problems we have in understanding music is that it is a language so few of us are using, even though we all have the ability to formulate musical thoughts by ourselves.