Is Music Theory a Science?

Is music theory a science?

Many arguments across the internet are, at their roots, arguments about whether music theory is an objective science, or if it is just people’s interpretations of music.

As you get to higher levels of music theory (I say, as a 2^nd-year PhD in Music Theory and Cognition), you’ll find that there’s a lot of room for interpretation. For example, not everyone would agree that there is a key change on the same exact chord. Some might argue that it happens earlier and others later. Are either of these wrong? Does that take away from music theory being a science?

To determine if music theory is actually a science to begin with, it’s helpful to first define what we mean by science. There are tons of definitions of “science” out there, but I’ve narrowed down 4 different ways that people define science.

These are:

1. Level of Specialization
2. The Object of Study
3. Data
4. Methodology

1. Specialization

The idea of defining a field based on level of knowledge comes way back from AristotleThese reports about Aristotle come from: Elisabeth Kotzakidou Pace, “The Techne of Music Theory and the Epistemic Domain of the (Neo-) Aristotelian Arts of Logos,” in What Kind of Theory Is Music Theory? Epistemological Exercises in Music Theory and Analysis, ed. Per F Broman and Nora A Engrebretsen (Stockholm: Stockholm University, 2007), 133–86..

Aristotle describes two ways of knowing things:

1. Personal experience, given to you by your senses
2. Through Art & Science, which is more specialized

He further distinguishes between types of knowledge with a chart that has sensory experience at the top and largest as what everyone experiences every day. Everyone listens to music all the time. It’s inescapable. Gas station, grocery store, walking down the street, everywhere.

As we move down in levels of the hierarchy, they get smaller because they become more specialized.

So from sensory experience that everyone has, we get practical experience. These would be musicians, if we’re talking music. Musicians have hands-on experience creating the sounds that people hear. They know about notes, what it takes to make sound, intonation, timbre, at the very least.

Then we move down to Art and Craft, which is where he seems to place composers, teachers, and potentially music theorists.

Kotzakidou Pace, who talked about this at length in her chapter of the book What Kind of a Theory is Music Theory?Elisabeth Kotzakidou Pace, “The Techne of Music Theory and the Epistemic Domain of the (Neo-) Aristotelian Arts of Logos,” in What Kind of Theory Is Music Theory? Epistemological Exercises in Music Theory and Analysis, ed. Per F Broman and Nora A Engrebretsen (Stockholm: Stockholm University, 2007), 133–86., seems to argue that music theorists are at this level, and that shouldn’t make the field any less important. We need this level to exist in order to get to the most specialized level of science.

From this perspective, music theory does not count as a science.

What really struck me about Aristotle’s hierarchy is that he includes medicine and engineering as part of the Art and Craft level. One other important thing to point out is that “knowledge” is not equivalent with science here. Science is just one, highly specialized way of knowing.

Aristotle also notes that this hierarchy is not a judgment of the value of certain domains of knowledge over others, only a description of specific versus general knowledge. Science, is thus really specific, and Arts a little more general (but still pretty specific compared with everyday experience).

He says,

…experience is in no way inferior to art; indeed we see men of experience succeeding more than those who have theory without experience. The reason is that experience is knowledge of particulars, but art [is knowledge] of universals”
—Kotzakidou Pace (tr.), 156Elisabeth Kotzakidou Pace, “The Techne of Music Theory and the Epistemic Domain of the (Neo-) Aristotelian Arts of Logos,” What Kind of Theory is Music Theory? Epistemological Exercises in Music Theory and Analysis, ed. Per F. Broman and Nora A. Engrebretsen (Stockholm: Stockholm University Press, 2007), 156

Aristotle goes on to say that the difference between the experiential knowledge and more specialized knowledge of Art and Science are about causality. Specialists explore predictability based on reason; they uncover “principled explanatory patterns.”

Kotzakidou Pace sums it up as philosophy and science are “knowing-that,” and Arts are “knowing-how,” which is how medicine and engineering fit into art rather than science. Because they’re knowing how to heal someone and knowing how to build a bridge.

The main takeaway here is that Aristotle (and others) define science based on a level of specialization of explanatory pattern, or “knowledge depth” (in regards to the depth of the above hierarchy).

2. Object of Study

The second way to define science is based on the object of study. If art is the object of study, then your field is the arts. If you’re studying natural phenomena (like lightning or gravity), then you’re doing science.

In the beginning of his chapter in What Kind of Theory is Music Theory?Per F Broman, “Music Theory: Art, Science, or What?,” in What Kind of Theory Is Music Theory? Epistemological Exercises in Music Theory and Analysis, ed. Per F Broman and Nora A Engrebretsen (Stockholm: Stockholm University, 2007), 17., Broman begins his chapter saying, “Given that music is an art…the analyst has to become an artist.” (17)

In a way, this approach to music theory is similar to Aristotle’s. Broman argues that to gain knowledge from an object, that scholar must disassemble it to gain explanatory power of how that object works. To do the disassembling requires a different set of tools than you would use for objects that are not art.

Broman doesn’t hold strictly to this approach in his chapter, though.

Later, he introduces the concept of Models, which he explains are built from the building blocks of theories. Morgan and Morrison, in their book Models as Mediators: Perspectives on Natural and Social ScienceMary S. Morgan and Margaret Morrison, Models as Mediators: Perspectives on Natural and Social Science (Cambridge: Cambridge University Press, 1999)., provide a helpful explanation of Models.

Their explanation of Models also creates a type of hierarchy of knowledge, but instead of knowledge of objects, this approach is knowledge of systems.

A system refers to an entity that changes or grows in time, meaning that there are many possibilities for what the same system might look like in the future. We can refer to each of these possibilities as a “possible state.”

“System” is the object that we grasp at as scholars (or scientists), and systems are full of smaller objects, in the more traditional sense of the word. Models are our way of poking at systems.

Models are representations we create to represent all the possible states of a system.

Below Models in this hierarchy are their building blocks: theories. Theories are more observational in nature, but they are tested and often interact with one another.

BromanPer F Broman, “Music Theory: Art, Science, or What?,” in What Kind of Theory Is Music Theory? Epistemological Exercises in Music Theory and Analysis, ed. Per F Broman and Nora A Engrebretsen (Stockholm: Stockholm University, 2007), 29. uses the example of f=ma as a theory, but “a simple equation is not enough to make a model to mirror the real world.” The equation of force isn’t enough to fully predict what a system (like the Solar System) will look like at a given future point in time.

To do that, you’d need context and surrounding conditions. To have a full model of the Solar System, you’d need to take gravity, friction, and many other theories into account and show how they interact with one another.

In music theory, a good example of a model versus a theory is Fuxian counterpoint. Fuxian counterpoint is made up of a set of rules about Renaissance counterpoint. Each underlying rule is simple on its own (like no parallel fifths, ever) and count as a “theory” in this sense. But when you combine all of the rules together, the model becomes complex and robust.

You can use the full model, made up of the combination of all of the rules, to predict what will happen in a given piece of Renaissance contrapuntal music.

It’s hard to think about that when the Renaissance was in the past, but imagine that a secret tunnel somewhere in Italy were uncovered that led to a library that’s been hidden in the ground all this time, and it’s full of old Renaissance musical manuscripts.

Given that none of us have seen any of these imaginary historical manuscripts, we can’t know for sure what characteristics the music will have. But because we’ve built a model, we can make a pretty educated prediction that we won’t find many parallel fifths.

Getting back to the Models hierarchy, though, below the level of theories are those of the Domain of Objects and History.

To keep with the Fuxian counterpoint example, the Domain of Objects refers to every actual instance of a rule. Every time there is no parallel fifth in a piece of Renaissance counterpoint is an example of this.

Then below this is the level of History: where did these objects even come from?

History shows us a journey of each object from where it started to where it is now, as we observe it. That lets us plot a trajectory for that object (a.k.a. identify a possible state for the singular object).

If you’re into music theory at all, I’m sure you’ve heard the phrase, “Theory follows practice.” That’s a helpful way of summarizing the journey up through this hierarchy.

Practice would be the lower 2 tiers, and we move up to the level of theory through observation, questions, and testing. Models also follow practice (just after Theory). We get to move up a level from Theory to Model through the process of discourse (like peer review and conferences) and deduction. The discourse part comes up later!

If we were to put Aristotle’s hierarchy of knowledge on top of the Models hierarchy, we’d have to flip it upside-down. Aristotle’s model originally had science as smaller because it was more focused on who was doing the science; it is a smaller level because fewer people do it, but here the size of each level is based on the amount of information and complexity.

Also different from Aristotle is the difference between each tier. Science, for Aristotle, was delineated by level of specialization, not about the object under study. The Models hierarchy delineates science based on what’s being studied: objects or systems. The bottom two levels are object-focused, and the upper two are more system focused.

To sum up point 2, some people define science based on what the object of study is.

If you’re looking at a piece of art, some people would describe what you’re doing as an art as well. The other way of thinking about the object of study is whether you’re looking at an object or the model of a system. If it’s a single object, maybe it’s more of an art. If it’s a system, then what you’re doing might be more of a science.

3. Data

Related to the object of study is the type of data that researchers get from their work. Data about Art and history is generally different from data from experiments. But of course, it’s a blurry line that music theory walks.

David Huron, who founded the Music Cognition Lab at Ohio State University, explains that empirical knowledge is just knowledge gained through observation. Science is only one type of empirical knowledge.(cite)

This means that empiricism ≠ science.

Even when a field uses empirical methods, that field is not automatically a science.

HuronDavid Huron, “The New Empiricism: Systematic Musicology in a Postmodern Age,” The 1999 Ernest Bloch Lectures Lecture 3, no. Methodology (1999): 1–32. also explains how the type of data we collect as researchers can be misconstrued to represent if a field is a science or not. He defines two data types: prospective and retrospective.

Retrospective data is evidence that you already possess, whereas prospective data is evidence you still have to collect.

Prospective data can be historical, as long as the researcher doesn’t have it in their hands yet. And this is how historians can actually test theories by predicting an unopened bunch of documents about a specific historical event. The event may be in the past, but their possession of the documents about it are in the researchers’ future.

I will note that Huron does not take the stance that empirical activity is a defining feature of science, but many others do.

His take is just that you work with what you’ve got, and if you call it science, great. If you call it something else, also great! Just use the tools that make the most sense for the job.

4. Methods

If it looks like science, acts like science, and talks like science, it’s science! Right?

Silly reasoning, but some people consider the method to define the field.

Karl Popper, one of the most influential epistemologists, was one of the strictest definers of science when it comes to method.

If you didn’t design your experiment to test for falsifiability, then you’re just not doing science. That’s all there is to it.

If you want to better understand why scientists test for null hypotheses, then you should read Popper’s The Logic of Scientific DiscoveryKarl Raimund Popper, The Logic of Scientific Discovery (Psychology Press, 2002)..

Continuing from Huron’s types of data from the previous point, the ability to control prospective data is usually what is associated with science. Huron explainsDavid Huron, “The New Empiricism: Systematic Musicology in a Postmodern Age,” The 1999 Ernest Bloch Lectures Lecture 3, no. Methodology (1999): 16, “Disciplines that can or cannot influence the phenomena under study are methodologically distinct.”

In experimental methods, researchers are able to control conditions to uncover causal relationships. In historical fields, however, researchers can’t do anything to control their experiments. They can’t, at least as far as I know, go back in time and change one variable, which means their methods are correlational.

Again, Huron doesn’t specifically delineate between science and non-science, but he does note that historical disciplines rely on correlational methods.

Distinguishing Features of Science

So there you have it, four distinguishing features of science that some or none of these authors would agree with.

1. Specialization
2. Object of Study
3. Data
4. Methods

If we take all 4 to be true, then science may be characterized as a highly specialized type of knowledge about natural phenomena (a.k.a. “knowing-that”) that creates models not just theories, which also means that it is more future-oriented in its knowledge. It focuses on prospective data, specifically controlled prospective data so that it can determine causality and have more explanatory power.

Science may defined by:

• Highly specialized knowledge
• About natural phenomena
• “Knowing-that”
• Creates models, not just theories
• Future-oriented
• Controlled prospective data
• Determines causality
• Has explanatory power

But we don’t necessarily have to take all of these to be defining features of science, and I don’t know that any of the authors take all four to be completely true.

Is Music Theory a Science?

A quick pause here for some discussion questions!

• Do you agree with these 4 defining features of science? Would you throw any to the wolves?

• Does music theory fit as a science based on these? Would we have to toss out any of these features for music theory to count of science?

Just Because You Can Doesn’t Mean You Should

Now that we have all the definitions out of the way, I want to bring in some voices who were fighting to say that music theory doesn’t need to be a science.

There has been a push for music theory to be a science, so it’s more pure and gets more respect. Some argue that music theory is more scientistic than scientific because it tries to look like a science but isn’t actually a science. I’ll leave that up to you for now.

But many say that music theory is valuable as an art or anything else. It’s valuable in and of itself.

In fact, there are those who argue that science isn’t the most esteemed type of research. Discourse is.

Sayrs and Proctor, in their chapter, “Playing the ‘Science Card:’ Science as Metaphor in the Practice of Music Theory,”Elizabeth Sayrs and Gregory Proctor, “Playing the ‘Science Card:’ Science as Metaphor in the Practice of Music Theory,” in What Kind of Theory Is Music Theory? Epistemological Exercises in Music Theory and Analysis, ed. Per F Broman and Nora A Engrebretsen (Stockholm: Stockholm University, 2007), 36. explain, “Discourse is credited with the highest degree of rigor, objectivity, utlity, and truth-value”

In other words, it’s not just the act of “doing science” that places a person in the most specialized level of Aristotle’s hierarchy of knowledge. It’s the rigorous discourse.

This discourse is where the entire idea of peer review comes from, which is probably the first thing you learned about when writing research papers in high school: always use peer-reviewed sources. I’m writing a post on what peer review actually means, so stay tuned for that. 😊

But probably the most important part of science (or scholarship in general) isn’t just in rigor or lack of bias. These come from being precise in using your tools.

Huron would add to this that you also want to make sure you’re using the right tools.

For Huron, regardless of if what you’re doing can be considered science or not, your research should still be led by the questions you ask. Instead of picking a method of study because you’re in a scientific field or a history field, you should pick a method based on how well it’s able to answer your research questions. He citesDavid Huron, “The New Empiricism: Systematic Musicology in a Postmodern Age,” The 1999 Ernest Bloch Lectures Lecture 3, no. Methodology (1999): 19. Maslow’s hammer analogy.

When you’re holding a hammer, everything around you starts to look like nails.

So when working on a research project, start by putting the hammer down. Have your full collection of tools around you, better yet, look at all the tools that exist, and ask which one best solves the problem you have.

If you don’t have the tool, or don’t know how to use the tool, ask to borrow it or to collaborate with someone who does.

This is also part of discourse!

All of this to say, even if we can squeeze music theory into the definition of science, it doesn’t mean we should.

What do you think? Are there benefits to calling music theory a science? What about to calling it an art? Or are the real benefits in being in this in-between art and science place that music theory’s been in for a long time?

The Importance of Discourse in Any Field But Specifically Music Theory

In the book, What Kind of Theory is Music Theory, Kotzakidou PaceElisabeth Kotzakidou Pace, “The Techne of Music Theory and the Epistemic Domain of the (Neo-) Aristotelian Arts of Logos,” in What Kind of Theory Is Music Theory? Epistemological Exercises in Music Theory and Analysis, ed. Per F Broman and Nora A Engrebretsen (Stockholm: Stockholm University, 2007), 181. focuses her chapter on Rhetoric. She compares what music theorists do to Rhetoric, as Rhetoric is “a method for doing systematically what others may have come upon by accident.”

In other words, music theorists put into words or diagrams what musicians have been doing by accident. We name things like chords, progressions, rhythms, and other musical phenomena, so they can be organized and repeated. (Music theorists do a lot more than this too)

But it’s important to note that music theorists don’t always agree.

This is where a chapter in Korsyn’s Decentering Music: A Critique of Contemporary Musical ResearchKevin Korsyn, Decentering Music: A Critique of Contemporary Musical Research (Oxford: Oxford University Press, 2003). comes in.

Part of this book focuses on what the object of musical study actually is (this should sound familiar from the discussion about what science is). Korsyn explains that most people cannot agree on what the objects are we actually study.

For an example, he looks at many analyses of Chopin’s Preludes, Op. 28.

We don’t have a letter or anything from Chopin telling us if this set of preludes is meant to be a cycle, played together, or if they’re just solo pieces he put together in one spot. Theorists have made arguments about this, and Korsyn grouped these analyses into four different ways of thinking about these preludes.

1. Monads – standalone pieces that just got grouped together arbitrarily
2. Nomads – pieces that actively try to keep anyone from thinking of them as being a cycle (they are all very different in vibe)
3. Cryptocycle – a cycle but one you have to dig really deep into to see how all the pieces actually connect together; you have to uncover the secret
4. Ironic or Paradoxical Cycle – the argument here is that Chopin did compose them to be a cycle but one that calls its own unity into question, similarly to a satire

The reason that picking one of these stances is so important is because, as Korsyn says, by naming the preludes one of these options, we simultaneously “unname” them by rejecting the other options. The Preludes, Op. 28 cannot be all of these.

Korsyn says,

…unnaming changes the relationship between text and context, inside and outside, content and frame, establishing a different hierarchy among the contexts in which the Preludes might be heard. (105)Kevin Korsyn, Decentering Music: A Critique of Contemporary Musical Research (Oxford: Oxford University Press, 2003), 105.

In other words, by choosing one of these four analyses, we change the object of study.

We change the musical work.

For some, the silences between the Preludes are an important part of the music (aka the object of study). But for others, those silences aren’t “real.” They’re just something that happens in between moments of music.

Korsyn continues,

Thus, “the music itself” changes radically in each construction of the Preludes: the very identity of the music is at stake. (106)Kevin Korsyn, Decentering Music: A Critique of Contemporary Musical Research (Oxford: Oxford University Press, 2003), 106.

And that’s pretty important given how the object of study has the potential for defining the field.

From here, Korsyn explains how important it becomes to ask a higher question: what makes these different constructions of the music possible?

Instead of looking just at the music, we should analyze the analyses.

He goes on to use White’s fourfold tropology that of figures of speech: metaphor, synechdoche, metonymy, and irony. He explains that these tropes concern all the possible relationships between parts and wholes. Because the problem we’re facing with the disagreement between whether the Preludes are a cycle or not, or something else, then it makes sense to use White’s tropology to frame the discussion.

You don’t need to know or understand the figures of speech for this discussion because we could still have discourse that seeks to understand what makes these interpretations possible without this specific framework. It’s just a handy tool that fits the problem really well.

I think Huron would approve.

I won’t go into detail about the rest of Korsyn’s discussion, but the important thing here is that he’s moving from a focus on the object of study—the Preludes themselves as the object of study—to a higher–level issue: how all of these interpretations can exist and how they relate.

Moving up to a higher level should sound familiar given the focus on hierarchies earlier in this post.

We can think of each of these constructions as a theory of the Preludes, and it’s only in their discourse that we could possibly get a higher-level Model for Chopin’s work.

They provide more context to the system. They all bring up important points about what the object looks like and how it behaves, so we need discourse to cope.

For what Korsyn does, we could think of his tropological framing as moving these individual theories up to the level of model, where the model is of a system of how humans think of musical collections. Each theory provides context for the others, filling in the other figures of speech. If we didn’t have any representatives for synechdoche, for instance, this framework wouldn’t work, and we’d need a different kind of model based on what context we do get from placing these theories next to each other.

In other words, we need the juxtaposition of theories to create models, and this juxtaposition occurs in discourse.

Additionally, taking what Sayrs and Proctor said about the importance of discourse for defining specialization, you need discourse to get to that top level thinking about Aristotle’s hierarchy of knowledge that I put next to the models hierarchy.

Sure we could try and test these theories some way, maybe using audience perception, but we can never really know what Chopin intended with this set of Preludes without asking him or finding a note where he wrote the answer.

And in a way, what Chopin intended for the Preludes also doesn’t matter because these pieces have the potential to be taken these many ways.

Instead of this being a problem of “what is the object,” we get more explanatory power for music more generally by asking “why can we think of this object as different things?” Therein lies the rigor for any field.

Music Theory as a Practice of Empathy

One of the things I say most in my social media presence is that music theory can be a practice of empathy.

I have been defining “music theory” as any individual’s perception of music, so in a way we all have our own theories of music.

To put that onto the models hierarchy, these “theories” would cover the two lowest levels, so they may not be full-fledged, formal theories, but they definitely feed into the theory level.

Some people do move their perspectives up to the level of theory: composers are a great example. They interact with music in specific ways and codify it as their process of making new music.

Music theory as a practice of empathy, then, occurs when people entertain other people’s theories. When we listen to the way that other people experience music differently than ourselves.

Even though I may not conceive of a piece of music in the same way they do, I have the capacity to understand where they’re coming from when they express it through an actual theory.

When we create a discourse (have a conversation) about the same object, we create a shared space where we entertain all of these perspectives and don’t dismiss them immediately as being “wrong.”

This is true, even if the status of the object is in question, based on differing perspectives like in the Chopin Op. 28 Preludes example that Korsyn describes. All perspectives go on the meeting table, and instead of dismissing any of them, we try to find the answer about where all the perspectives even came from.

All of these theories get to be valid for the time being, and this is what strikes me as a practice of empathy.

Without going down the rabbit hole of what empathy actually is, I’ll just take “empathy” to mean a shared space where experiences are beheld and understood by those present.

In this way, music theory (a.k.a. creating a model from multiple theories) becomes a practice of empathy.

I agree with the authors I cited here that music theory is valuable on its own without it having to be a science or anything else. But practicing empathy as a result of the discourse required to create a model is an important human thing.

Instead of being robots as researchers, we have even more opportunity to be humanistic and put into practice what it means to be human.

Of course, this isn’t a feature of only music theory or the arts, but the arts do have much more opportunity for it, given the personal and subjective nature of our objects of study.

Discussion Questions

I am genuinely curious about your thoughts on any and all of this. Being in academia tends to isolate people from the outside world, so any insight you want to share in the comments or via email ([email protected]) is super welcome here!

If you need help centering your thoughts, here are some discussion questions:

• Do you agree with the four defining features of science I outlined in this post? Would you toss any out?
• Does music theory fit as a science based on all 4 features of a science? Which would we have to toss out for it to count as a science?
• What are the benefits of calling music theory a science? What about the benefits of calling it an art?
• Do you think music theory can be a practice of empathy?