24/08/2026

The Circular Definition of "Music"

TL;DR

I propose the following history of the evolution of music and word-based language:

The Problem of Definition

“Music” is a hard word to define. Wikipedia even has a whole page https://en.wikipedia.org/wiki/Definition_of_music devoted to the difficulty of defining “music”, separate from its main page on Music.

When we define a word, we usually hope to define that word in terms of other things.

If we define “X” as being “Y”, then we might ask what the meaning of “Y” is, and if the answer is “X”, then we haven’t really given any useful definition of “X”.

In the case of music, we know that it has relationships with various other things, including emotion, and dance, and normal prose language (AKA song lyrics). But the exact nature of those relationships is not certain, and in practice our understanding of the meaning of the word “music” is almost entirely dependent on our subjective experience of actual musical items and the fact that other people use the word “music” to describe those items.

Indeed, if our knowledge of what “music” means comes entirely from our subjective experience of actual examples, then perhaps the correct dictionary definition is:

(And somehow your dictionary includes the ability to play audio files – perhaps there’s a CD included with the book, or maybe it’s an online dictionary that can link to MP3 files.)

But I would like to suggest an alternative approach.

Maybe our first mistake is the assumption that music should be defined in terms of something else.

Maybe the correct definition of “music” is actually circular.

Maybe music is something that exists for its own sake, and therefore cannot be defined in terms of anything else.

On the one hand this might solve the problem. On the other hand it does feel like a cheat. Also, from the viewpoint of theoretical evolutionary biology (a viewpoint that the current author definitely subscribes to), defining a thing entirely in terms of itself remains problematic, because that definition fails to solve the question of biological functionality, ie what is “music” actually for?

So, here goes.

A Proposed Circular Definition of “Music”

In summary:

I believe that this definition corresponds fairly well to our personal and subjective experiences of making music and listening to music.

However it does not obviously tell us anything about what the biological functionality of music might be.

In particular, “pleasure for pleasure’s sake” does not count as a plausible biological function.

Musical Identity

Music does not consist of a homogenous substance, even though we might talk about it in the abstract sense of “music”, in the same way we might talk about “food” or “water”.

Actual music consists of musical items.

And it is important to observe that creating new musical items with a level of musicality similar to the items we already know is a non-trivial thing to do.

In the world of music that we know of, both the composition and performance of music are quite competitive. A musical item is only “good enough” if it is as good as or better than the musical items that we already know of.

We can observe that musical items have a fairly strong identity in the mind of the music listener.

That is, when we hear a musical item that we have heard before, we easily recognise it as being that musical item.

We can assert a stronger proposition, which is that musical items with a high level of musicality have a definite qualia, in the sense that listening to a strong item of music “X” has a feeling which is unique to the feeling of listening to that particular item of music.

We can consider the possibility that the identity of musical items is their biological function.

That is, each musical item serves a biological function which consists entirely of being a sequence of sounds that can be uttered, and when listeners hear that utterance, they know which musical item they are listening to.

Because the experience of this identity is pleasure-driven, the result is that people are motivated to create musical items with an identity, and to perform them so that other people hear the same items with the same identity. The existence of the musicality function acts socially, motivating the development of a shared repertoire of musical items within a society.

But how does that provide any kind of biological functionality?

Identity suggests Meaning

One possible purpose for the composition of audio items with a strong unique identity is for those items to act as symbols in a language.

The only problem with this hypothesis is that, as far as we know, distinct musical items do not carry any strong meaning.

It is true that some musical items are always used in specific circumstances, like “Happy Birthday”, or that wedding march by Mendelssohn.

And many popular musical items consist of a tune with lyrics, and the lyrics give the tune an associated meaning.

But for the most part these associations are not fixed or strong in the same way that the meanings of words are.

Pachelbel’s Canon in D is popular at weddings, but it can easily be used in other situations where there is no wedding and no intention to reference or suggest a wedding.

And lyrics in songs can be replaced with different lyrics – as long as the new lyrics are sufficiently consistent with the expressed emotional feel of the music.

It might seem that we have to abandon the idea that musical identity determines meaning, but in an evolutionary context there is an alternative hypothesis, which is:

In other words, music was a system of communication, and I will use the term communicative music to describe music when it had that function.

Relating this to the current state of things, where music is not communicative, the implication is:

To find a possible reason as to why musical identity may have ceased to determine meaning, we need only look at how modern humans communicate symbolically, ie via word-based language. (In this article and elsewhere I use the phrase “word-based language” in lieu of “language”, so that I can still use the abstract concept of “language” to refer to any system that involves communicating using symbols that have agreed shared meanings.)

We can consider the differences between a hypothetical musical language and modern word-based language in terms of symbols, units of meaning and utterances.

In modern word-based language we have:

In communicative music, all of these things had the same granularity, ie:

(There is also the issue of context which I can touch on – with word-based language each utterance has meaning within the context provided by previous utterances, and each utterance contributes to the evolution of that context. Whereas with music, each musical item defines its own context. And if one musical item stops, and a new item starts, the effect of starting the new musical item is generally the complete elimination of any context from the previous item.)

Communicative music would have been both much less efficient and less powerful than modern word-based language in terms of what it could express. Each melody had a length longer than any individual syllable or word in word-based language, the whole melody expressed just one meaning, and there wasn’t any way to join different melodies together to express more complex meanings.

This gives us a straightforward explanation for why communicative music ceased to operate as a language: it was replaced by the vastly superior word-based language.

But there is an additional possible twist to this story.

A major component of music as we know it consists of word-based language embedded in the music, ie song lyrics.

On the one hand the lyrics do not directly affect the musical quality of the music. On the other hand the lyrics are generally required to have a meaning that is consistent with the emotion feel of the music.

This suggests the possibility that words originally evolved embedded in communicative music.

There are a few reasons why this possibility makes sense:

There is a second question that needs to be answered. We can understand that communicative music ceased to operate as a language because it was replaced by the vastly superior word-based language. But why then did music not just disappear altogether? Why did a partially but not fully disabled communicative music continue to operate as something that continued to consume the time and effort involved in composing, performing and listening to music?

To answer this question, we have to suppose that the partially disabled communicative music evolved to serve some other secondary purpose.

Furthermore, given that we still don’t know of any major biological function of music in the current day, we have to suppose that this secondary purpose, whatever it was, itself has disappeared. Also, given that music still exists, consuming resources of time and effort with no obvious purpose, we have to assume that this final disappearance is something that has occurred quite recently in a prehistoric sense, or even that it is still in the process of fully disappearing.

So, what significant changes have occurred “recently” in the environment that humans live in which might affect the relevance of some particular biological functionality?

In the human case, one “recent” change in circumstances is the rise of “civilisation”, or to be more precise, the growth of complex societies (where the leaders or members of those complex societies might not always be “civilised” in the moral sense of the word), which is something that started to occur about 12,000 years ago (cf 200,000-300,000 years for the amount of time that “anatomically modern” humans are believed to have existed).

With all that in mind, I would like to suggest a plausible candidate for how a secondary function of communicative music evolved when it ceased to operate as a symbolic communicative language:

As a result of this, the communicative music evolved from being part of a communication system to being a motivator of imagination.

In effect communicative music evolved into what we might call evocative music, which no longer communicated, but which evoked emotions and temporary imagined meanings consistent with those emotions.

On the one hand we know that imagination plays a very important role in the development of human culture and technology. So it is plausible that a partially disabled communication system that accidentally evolved into a system for motivating the imagination of imaginary things might have played a significant role in the cultural evolution of modern humans.

At the same time, in the modern world, we do not observe ourselves imagining useful things as a result of listening to music.

Indeed there are some people who have imaginations strongly driven by listening to music, ie so-called maladaptive daydreamers, but for the most part the imaginative activities of those people do not provide any significant benefit to themselves or others.

In the modern world, most of the craziest things that we can imagine build in some fashion on the known crazy imaginings of other people. In the modern world, with modern technology, we have access to any imagined idea or situation that has been publicly expressed or recorded by any other person in the world, whether they be living now or in the recent historical past.

Because of this ready access to the output of other people’s imaginations, there is less necessity to be extremely motivated to imagine crazy things from scratch.

So this is my final hypothesis about how communicative music evolved from a communication system, to a system of motivating imagination, and finally to a system of not much of anything at all.

That is, music evolved into a system of motivating imagination, but with the development of complex societies, the benefits of imaginative thought to the individual person doing the imagining became less than what it used to be, so even that level of motivation has evolved away (or is in the process of evolving away and that process hasn’t finished yet).

The Expression of Emotion in Music

I haven’t actually said much about the musical expression of emotion in this article, other than to observe that the meanings that can be associated with an item of music must be consistent with the emotional feel of that musical item.

The hypothesized musicality function is a function that has a 1-dimensional output, ie answering the question “how musical is the music?”

Whereas emotion is multi-dimensional – although the exact number of dimensions might be difficult to pin down (see https://en.wikipedia.org/wiki/Emotion_classification for some discussion on that question).

We can fit the expression of emotion into this theory if we assume that the emotional aspect of communicative music actually existed before the appearance of a musicality function.

So initially where was protomusic, and this was a fixed language where vocalisations expressed a set of possible emotions. Protomusic did not have a discrete repertoire of “items” in the sense that communicative music had (or that modern evocative music has). The set of possible protomusical utterances existed in a multi-dimensional continuum that mapped continuously to the multi-dimensional continuum of possible emotions that it could express.

In order for something like a musicality function to evolve, there had to be some pre-existing set of possible utterances that the function could be applied to, and this is exactly what the emotional language of protomusic provided.

The evolution of the musicality function allowed communicative music to express more specific meanings in a symbolic fashion, but the underlying expression of emotion did not go away. So the result was a language consisting of discrete items capable of expressing specific culturally-assigned meanings, but, those meanings still had to be consistent with the emotional feelings expressed by the original protomusical language of emotion.

Hypotheticalness

There is another change that had to happen to the nature of what music expressed.

In the transition from communicative to evocative, music ceased to pragmatically communicate facts about the real world. It followed that the emotions it expressed no longer described the current state of reality. To fit better with the function of motivating imagination, those accompanying emotions were hypotheticalised – that is, music now expressed hypothetical emotions.

Originally, with protomusic and then communicative music, the emotions were real, as in “I feel this way”, or, in some cases “We all should feel this way”. But with modern evocative music the expression has changed to “Consider what it is like to feel this way”.

Musical Identity and “Melodic Identity”

I have talked about “musical items” in this article.

With modern evocative music, a musical item can be one person singing unaccompanied by anything else, or it can be a 4 piece rock band, or it can be a whole orchestra.

But if we assume that communicative music was a system of pragmatic communication, then almost certainly it was only “spoken”, or sung, by one individual at a time – there might sometimes be multiple listeners, but there was only ever one speaker.

And it is also unlikely that the speakers carried musical instruments around with them all the time for the purpose of communicating pragmatically.

So musical items would actually have been just musical melodies, and in that case musical identity would actually have been melodic identity (which is the terminology I have used previously in other articles).

With regard to prehistoric musical instruments, if we assume that musical communication did not involve hand-held instruments, this implies that when musical instruments came into existence (ie at least 42,000 years ago), communicative music had already evolved into modern evocative music, and music was no longer part of a pragmatic system of communication.

The Glial Implementation of the Musicality Function

In order to evolve quickly, the musicality function had to be something simple.

I have developed the hypothesis that the musicality function is at least partly determined by the occurrence of certain physical patterns of activity in regions of the auditory cortex involved in processing sounds, and especially those regions involved in processing the sounds of utterances uttered by other individuals of your own species.

In particular, these patterns consist of constant regions of activity and inactivity within those cortical maps (“maps” in the sense that there is a correspondence between the location of neurons in the map and the perceptual values that the activity of those neurons represent). The occurrence of these patterns corresponds to the perception of some things happening and other things not happening.

The prototypical example of this pattern is musical scales, where pitch values on the scale happen, and pitch values not on the scale don’t happen.

A similar example is nested regular beat, where certain sustained regular beat frequencies occur, and others don’t. For example, with 4/4 time at 100bpm with shortest notes being 1/16 notes, there would be sustained regular beat frequencies of 25bpm, 50bpm, 100bpm, 200bpm and 400bpm, and no sustained beat frequencies at any values in between those values. So if we presume that there exists some cortical map where active neurons represent the frequency of sustained beats, then that cortical map would have constant patterns of activity and inactivity when responding to musical rhythms.

The occurrence of these constant regions of activity and inactivity can also be characterised in terms of exact repetition, where most of the perceptual values occurring with respect to a particular cortical map are exact repetitions of values that have previously occurred.

In order to implement this criterion for musicality, you might think that evolution would have to create a whole new population of “observer” neurons to observe the occurrence of these activity patterns in the cortical maps of the neurons doing the actual perception of those values.

But, as it happens, there already exists a major population of brain cells whose job it is to respond to the activity of nearby neurons, ie according to the location of those neurons. These brain cells are the glial cells, which play a major role in supporting the function of neurons.

So it is plausible that the implementation of the musicality function based on observation of physical patterns of neural activity has occurred not by creating a whole new set of neurons, but rather by adding this functionality to the existing glial cells that inhabit the relevant cortical maps.

The strongest possible confirmation of this hypothesis would of course be to observe some specific response of glial cells to musically activated patterns of neural activity. A major difficulty is that the major premises of the hypothesis imply that we are talking about something that only happens in the human brain (and which happened in the brains of some of our extinct hominid ancestors), and the types of observations that scientists might perform on non-human animal brains are not the types of observations that one can make on human subjects.

One might also expect some type of genetic signature associated with the evolution of this function (which avoids the problem of experimenting on live human brains), but it’s hard to say exactly what such a signature might be.

(A common question about music is: “Which is more essential to music, rhythm or melody?” The theory of constant patterns of activity and inactivity suggests are more fundamental explanation of what defines musicality, because it applies equally to important aspects of both rhythm and melody.)