Three Questions That Music Scientists Need To Ask

18 October, 2006
Which is the biological adaptation: making music or listening to music (or both)? What makes some music better than other music? Would what I am saying about music make sense if I said the same thing about something else?

The Questions

The Mystery of Music

Our ignorance about what music is remains profound. Music, or the desire for music, has a significant effect on human behaviour. Substantial resources are devoted to the composition, performance and enjoyment of music. Yet it serves no apparent purpose.

If we could believe that God just made us out of nothing, then we could believe that He threw music in as something for us to do. But the theories of biology do not allow such extravagance. If music is a big thing in our lives, then it must be explained, and that explanation must have in it somewhere an adaptation: we must explain how the genes which cause music to exist, or at least make it possible for music to exist, also give their owners increased long-term reproductive success.

The mystery is compounded by the fact that we do not have an objective description of what it is that distinguishes musical sounds from sounds which are not musical. So-called "music theory", based mostly on common-sense observation of how music is constructed, can tell us something about the features that seem necessary if not completely sufficient to create music. But the modern music industry remains very dependent on human composers, and that tells us that something significant is missing from the descriptions provided by formal music theory.

If the mystery of music has not been solved, it should not be for lack of trying, as it has become a popular topic of research, and there are many professional scientists busy at work making observations and testing theories. It should also not be for lack of data, because the whole body of modern and traditional music is available (although some may be hard to access for reasons of copyright), and it is available both in notated forms and as digital recordings.

But if we look at what music scientists were talking about in the year 2000, and we look at what they are talking about now, in the year 2006, there doesn't seem to have been a whole lot of progress. There is a larger body of empirical data from a range of sources, be it the responses of babies to music, perception of sound by non-human animals, brain mapping of professional musicians or the disabilities suffered or not suffered by those with various forms of amusia. Yet no one seems any closer to figuring out what music is, and if there is some big breakthrough that needs to happen before this question is answered, then it hasn't happened yet.


What does strike me, as I read some of the recent scientific literature, and particularly the review papers that are written periodically, is that there are some assumptions that are made, perhaps without anyone realising that they are being made, and one reason that these assumptions persist is that certain questions are not asked.

There are three particularly important questions which I think need to be asked, which are:

The corresponding assumptions challenged by the first two questions are:

The third question is more of a meta-question, and can be used to challenge many assumptions about what makes sense when talking about the biological function of music.

In the following sections I consider each of these questions, and then I apply them to a recent review article that appeared in the journal Cognition.

First Question: Which is Adaptive: Making or Listening?

Music has two major aspects: production and consumption. This leads to a basic question about biological function: if there is a biological function associated with music, is it more associated with the production aspect, or with the consumption aspect, or is it equally associated with both?

Music production consists essentially of composition and performance. Musical consumption consists of listening to music and enjoying it. Production and consumption cannot be completely separated, because anyone performing music is necessarily also listening to it, and, one would assume in most cases, enjoying it. It is however quite possible to enjoy listening to music without playing any part at all in the making of it. This suggests that our enjoyment of music is more biologically fundamental, but it doesn't really answer the question about whether the production of music serves any biological function (other than generating the indirect benefits received by a musician in return for making music that other people like, which doesn't really count as a convincing biological explanation because then one has to explain why the listeners give something to the musician in exchange for something that has no intrinsic biological benefit for the listeners).

If we consider other human activities, such as eating and drinking, we can see that the desire to consume is programmed into us as instincts, i.e. hunger and thirst. There is still a biological function served by producing – or, more generally, acquiring – food and water, but we are not born with any detailed instructions on how to produce or acquire them.

Second Question: What Makes Some Music Better?

Talking about music being "better" or "worse" seems very judgmental. Which of course it is. Judging the "quality" of music can seem undesirable for a number of reasons, including:

These objections may be why the scientific and other academic music literature seems keen to avoid any discussion involving the idea that some music is "better" than other music. Yet, if we consider some other human perceptions, for example perception of "sexiness" of a potential sexual partner, or perception of the "deliciousness" of potential food, or even the perception of the moral "rightness" of a an intended action, we can see that all these perceptions serve very real biological functions, and they all involve judgements, even though they are all subject to both cultural and personal variation.

Third Question: What If We Said the Same Thing About Something Else?

The third question is a meta-question, because it can help us answer other questions and discover weaknesses in our reasoning when we consider possible answers to those questions.

My discussion of the first two questions brought up analogies with other human instincts, perceptions and behaviours, i.e. when I considered instincts in relation to food and water. The third question implies that we should adopt this approach as a general and compulsory principle of inquiry: in any discussion of some aspect of music, would the logic of the discussion survive substitution of "music" with something else? In some cases the discussion would cease to make any sense at all, and the substitution tells us nothing useful. But in other cases we might see that the substituted premises are valid, yet the substituted conclusions are obviously wrong, or at the least very dubious.

So, what should be substituted for "music", in order to test the validity of our reasoning about music?

A short answer is: anything and everything, because we know so little about what kind of thing music is. Music isn't exactly like any other thing that we already know about – if it was, then we would already have solved the mystery of what music is – and any analogy between music and something else is likely to be imperfect. But the more substitutions we try, the more chances we give ourselves to check that our reasoning is not weak or fallacious.


A good example of a word that can be substituted for the word "music" in discussions about the biology of music is the word "cooking".

Why Cooking is a Bit Like Music

For the purpose of this example, I will define "cooking" to mean the preparation of food by heating it. In the so-called evolutionary environment of adaptation (or EEA), the source of heat would almost always have been fire, so from the point of view of evolutionary biology, cooking can be assumed to be cooking with fire (and in particular, no microwave ovens).

Cooking is an activity of production. The corresponding consumption activity is the eating of cooked food, which is of course a subset of the general behaviour of eating food.

There isn't really any "mystery" of cooking. We cook because that's what you have to do if you want to eat cooked food, and people want to eat cooked food.

There doesn't appear to be any "cooking instinct". Cooking is a goal-directed activity, it's a job to be done, just like many other jobs that have to be done. It can be done well, it can be done badly, and it can be done "good enough".

There do appear to be genetically determined instincts in relation to the eating of cooked food. People prefer eating hot food, which in the EEA would almost always mean food that had just been cooked (because use of firewood to reheat food would probably be considered a waste of time and fuel anyway). And people prefer to eat food – especially meat – which is at least slightly burnt on the outside.

One can plausibly argue that these instincts reflect a necessity to cook certain types of food well enough to kill pathogenic bacteria, and to eat the food before more bacteria have a chance to recolonise the food. Although these instincts are not instincts for cooking as such, they could only have evolved in a context where cooking happens and there is a choice between eating food which is cooked and food which is not cooked (or not cooked properly).

An Apparent Difference Between Music and Cooking

If we were to suppose that music is only produced because people want to hear music, then this would be analogous to the assertion that people only cook food because other people want to eat cooked food.

One apparent difference is that with music there is a considerable enjoyment in performing music, perhaps more so than there ever is with cooking. However, we can explain much of this difference by pointing out that the performer of music is also a consumer of music, since necessarily they hear and enjoy what they perform as they perform it. There may be a minor element of this in cookery, as the cook samples a dish being cooked to check if it is finished yet or if more ingredients need to be added, but for the major part of the enjoyment, the cook has to wait just like everyone else until the cooking process is completed.

Thus what seems like a significant difference between music and cooking may be largely a consequence of the different relationships in time between the production activity and the consumption activity in each case, i.e. music making gives immediate satisfaction, whereas cooking demands a certain amount of patience, because we can't eat the food until it has finished cooking.

Review: "The nature of music from a biological perspective"

The nature of music from a biological perspective is a review paper by Isabelle Peretz (a well-known music scientist) published as part of the May 2006 issue of the journal Cognition. Unfortunately the other papers in that issue are not freely available online, although I was able to find an online version of The Capacity for Music: What Is It, and What's Special About It? by Ray Jackendoff and Fred Lerdahl.

In the rest of this article I will analyse Peretz's paper in relation to the three questions.

Question 1: Which is Adaptive: Making or Listening?

Some of the time Peretz does make a clear distinction between these two aspects of music, for example she tells us that:

Throughout human history and across all cultures, individuals have produced and enjoyed music.

However, in many other parts of the paper, perhaps to avoid the tedium of having to list these two aspects over and over again, she uses terms with a deliberately inclusive interpretation, such as:

Peretz does discuss many specific details about music, and many of these details pertain specifically either to production or to consumption. But when the discussion rises to a more general level, the distinction between the two aspects is not emphasised so much.

The overall aim of Peretz's paper is to consider the extent to which music is biologically determined. If I was writing such a review paper, I would be more inclined to ask questions about biological determination separately for each of these two aspects of production and consumption.

For example, the ability to enjoy music appears to be the more basic capacity, and this suggests that our enjoyment of music is biologically determined, and that the production of music is totally determined by our tendency to enjoy music. This would then reduce all biological questions to questions about enjoyment, and the mechanics of how people create music would not be any more interesting than the mechanics of how people create food, or shelter, or anything else that satisfies basic biological desires. Or rather music-making would be interesting only in as much as it constituted a goal-directed creative activity, but it would not be interesting as a specific biological function.

Or, this view might be wrong, and making music might be the more basic function, with music perception serving some supporting role as a system for judging the quality of the music that is produced.

Second Question: What Makes Some Music "Better" Than Other Music?

Peretz does not ask this question (or anything like it) anywhere in her review paper, and she does not give any references to the work of anyone else who has asked it.

But we can find a few places where if she had asked the question then she might have said something different.

The very first sentence of the paper starts off with:

Music is generally regarded as an exquisite art form, ...

This is a direct reference to the "goodness" of music, but it only considers music as a whole, i.e. there is no suggestion that some music might be more "exquisite" than other music.

Further on Peretz tells us that:

Everyone knows what music is but cannot delimit its boundaries.

If we realise that some music is better than other music, then we can also realise that other music is worse, and that there is a continuum from good music to bad music to really bad music which is so bad that it ceases to be perceived as music. The boundary between music and "not-music" is not well defined, because "musicality" is a continuous value, and not a binary yes/no value. (Note: whereas I have adopted the term "musical" to refer to the quality of how "musical" a piece of music is, Peretz uses the same word strictly as an adjective describing people, referring to their ability to make or appreciate music.)

The absence of the concept of judging the quality of music is apparent when Peretz talks about amusia (a subject which she herself has done substantial research on). A number of criteria for measuring musical "ability" are given, including:

Missing from this list is the ability to distinguish good music from bad music. If this ability was included in a test of musical ability it would seem judgmental, and indeed it would be. Especially if we are considering passing judgement on someone else's judgement about how good music is, which sounds like passing judgement on other people's musical tastes just because they differ from our own. But having different musical taste from other people is not the same thing as lacking perception of musicality. Although musical taste varies, it does not vary so much that it is normal for people to like music which no one else likes. To put it another way, if you really like some particular item of music, almost certainly there will be other people who also really like that item of music.

Since there appear to be no tests for amusia described in the literature which are based on ability to judge the quality of music, I will take the initiative and describe a simple protocol, which might be called the album test, and which proceeds as follows:

Even though musical tastes vary, we would expect non-musical subjects to fare much worse on this test than most people with "normal" musical ability.

The tricky part is to find suitable albums where the subject has not been previously exposed to any of the songs. It might be necessary to pick genres of music slightly outside a person's normal experience (e.g. from other countries). On the other hand, anyone who is amusic is probably uninterested in music generally, and they won't know that only two songs from so-and-so's last album are worth listening to unless you're an extreme fan of so-and-so, so it won't matter which album you choose. ("Normal" subjects doing this test should be encourage to state which songs they like the most, even if they already think they know which ones were the hits.)

Ironically, Peretz follows the discussion of amusia with a discussion of great musicians, such as Louis Armstrong, and what it takes to be one. The ironic part of it (in relation to my second question) is that the very notion of a "great" musician implies the existence of a sufficient consistency of musical judgement such that we can all agree on which musicians actually are the "great" ones because they know how to give "great" performances (or if we can't all agree, because some of us are uninterested in jazz trumpet, then at least a large number of us can agree).

Peretz does go as far as saying:

For music to be appreciated, performers and listeners alike must share core processes and knowledge.

However she falls short of mentioning or suggesting that the whole purpose of these "core processes and knowledge" might be to help the audience make some judgement about the quality or "musicality" of the music being performed.

She goes on to talk about "functional music" (apparently a technical term for "popular music"), which is "understood by all members of a community" (italics mine), and again Peretz stops short of recognising that the measured popularity of music might be a function of the music's musicality which might be the very thing that music perception is all about.

There is a short section in the paper on universals, and these are characterised by whether or not they occur across a large number of musical cultures. Candidates include singing, meter, uneven scales with 5 to 7 pitch values repeated across octaves, structure and rhythm. Of course if we were considering the musicality of music, we wouldn't just ask which features happened to occur historically in different musical cultures (and I say "historical" because musical culture is becoming universal). We would ask which features contribute sufficiently to musicality such that once they are incorporated into a musical culture, they are never removed. Western music can be seen as a homogenized culture which annoyingly erases all signs of previous musical history, but it can also be seen as proof of the existence of a set of musical features which consistently improve the musicality of music as judged by most people who listen to music. (This is not the same thing as saying that all improvements in popular music come from Western sources, since many features of modern Western popular music are derived originally from African styles of music.)

While still on the subject of universals, Peretz considers the constraints imposed by a diatonic scale containing only eight distinct notes, and observes that:

This finite pitch set enables the generation of an infinite number of musical structures. Thus, factors related to the discriminability and learnability of fixed and discrete pitches must constrain these choices.

This seems to be an ad hoc hypothesis that anything that we don't already know about what determines people's choice of music must be determined by "discriminability" and "learnability".

But the notion of "musicality" provides a different explanation for the constraints on music: the constraints that determine music are that people want to listen to the best music, i.e. the most musical. If we consider the set of all possible musical pieces, and a mapping from this set onto some measure of musicality, then it is plausible that the higher demand we make of musicality, the lesser number of items we will have to choose from. Of course current musical knowledge and technology may not be creating the best possible musics – there may be something much better "out there", if only we knew where to look. Music is "constrained" in the same sense that a gold mine is "constrained" to be where the gold is. But if you had one gold mine, and then you discovered a new location where the gold was ten times more concentrated, then you would likely abandon your first mine, and it would cease to actually be a "mine", in the sense that no one would be going into it any more and extracting gold ore from it.

In a section on "domain-specificity", Peretz mentions Steven Pinker's theory that music processing may be a system for processing "something else". Of course if you assume this, and you assume that music perception is the perception of musicality, then you end up with my theory, which of course isn't mentioned in the paper at all, so I'll say no more about that.

The sections on musical modules and innateness include more discussions of amusia, again with references to various tests on musical ability, which include tests on various aspects of music perception, but do not include any tests of the ability to tell which music is good music. And there is some discussion of the musical abilities of infants, and even discussion of their preferences, but no mention of their ability to determine musicality. (From the observation that the musical preferences of young children are somewhat distinct from those of adults, we could deduce that the ability of young children to measure musicality is less than that of adults – if we regarded adult taste as the benchmark. Of course it could be that infant musical preferences serve the purpose of infants in the same way that infant food preferences serve the purposes of infants, rather than just being an undeveloped version of adult preferences, in which case measuring the musical abilities of infants against adult benchmarks is an irrelevant comparison.)

Towards the end of the paper there is a section on the emotional effects of music, and Peretz specifically tells us that "music can have a profound impact on listeners" and "music is a powerful tool for emotion and mood modulation" (the latter quote referring to a separate paper by Bharucha, Curtis and Baroo), and again we might expect a realisation that how profound and how powerful the effect of music is has something to do with how musical it is. But the discussion remains on the level that music can invoke emotions (and that it can invoke various different emotions), not on how much emotion different items of music invoke or why some items invoke emotion more strongly than other items.

The Third Question: What If We Swap Something Else for Music?

There is relatively little discussion in Peretz's paper of analogies between music and other aspects of human behaviour and perception. In the section on domain specificity Peretz considers the example of reading and literacy, particularly that reading is localised in certain brain structures, even though those same brain structures cannot possibly evolved for the purpose (because reading is a historically recent invention, and widespread reading is even more recent). She also mentions Morse code in a discussion of culture versus genes.

There is also some discussion of language, but this is more because language has specific similarities to music, and it doesn't really count as a test of substitution.

Using the previously mentioned example of "cooking", the substitution test can be applied to Peretz's paper by deleting the word "music" wherever it appears appears and replacing it with "cooking" if the reference is production-oriented, "eating" if the reference is consumption-oriented or "cooking/eating" if the reference is inclusive of both production and consumption. The substitution breaks down when the discussion refers to too many specific details, but works well when the discussion dwells mostly on generalities. For example, here is an extract from section 6 "Why are humans musical?":

Two main evolutionary explanations have been offered. The initial account was provided by Darwin (Darwin, 1871) himself who proposed that music served to attract sexual partners. This view was revived by Miller (2000), who argues that the impulse to make music (and create art in general) is a way to impress prospective sexual partners with the quality of one's brain and thus, indirectly, one's genes. Musical virtuosity is unevenly distributed, demanding, hard to fake, and widely prized.

and here is the same extract with the word "cooking" substituted for "music" as shown in bold (and with recent references sanitized to avoid any possible confusion):

Two main evolutionary explanations have been offered. The initial account was provided by Darwin (Darwin, 1871) himself who propose that cooking served to attract sexual partners. This view was revived by M***** (2000), who argues that the impulse to cook (and create art in general) is a way to impress prospective sexual partners with the quality of one' brain, and thus, indirectly, one's genes. Cooking virtuosity is unevenly distributed, demanding, hard to fake, and widely prized.

And for an alternative substitution, emphasising the possibly drug-like nature of music, substitute "alcohol" into a later portion of the same paragraph, here shown before substitution:

However, the dominant view about the adaptive value of music lies at the group level rather than at the individual level, with music helping to promote group cohesion (Wallin et al., 2000). Music is present in all kinds of gatherings – dancing, religious rituals, ceremonies – strengthening interpersonal bonds and identification with one's group.

and after substitution:

However, the dominant view about the adaptive value of alcohol lies at the group level rather than at the individual level, with alcohol helping to promote group cohesion (W***** et al., 2000). Alcohol is present in all kinds of gatherings – dancing, religious rituals, ceremonies – strengthening interpersonal bonds and identification with one's group.