Studying theory isn’t as important as you think it is

Studying theory isn’t as important as you think it is

Posted by Stephen Corwin on April 08, 2020

Why aspiring musicians, programmers, linguists, and bakers shouldn’t let theory get in the way of learning something new

Sheet music thanks to sheetmusicdirect.com

I took piano lessons when I was six, for what feels like a long time, but was probably only a year or so, and as far as I recall, I hated it. I don’t actually have many memories of the lessons themselves — that is, sitting on the bench with my hands on the keys playing music. What I do have is a set of tangential memories: the fights I’d get in with my mom for not having practiced a single time between weekly lessons; the basket of chachkie toys my teacher used to get me to pay attention; the way she clicked the mechanical pencil — a device I’d never seen before at the time— when she would write in the names of the notes on my sheets. She wrote the notes, of course, because I never learned to read music.

A particularly potent memory: There was a book of exercises she used to keep in the bench called Finger Power. It was full of sheet music, but rather than songs, the notes on the page were just repeating patterns traversing the octaves of the keyboard. They were supposed to teach you how to perform complicated maneuvers with your fingers, and the idea was that if you studied these maneuvers and did them over and over again, you could eventually use them to play beautiful music.

My bitter hatred of finger power is probably the memory that stands out most from that time.

I’d wanted to learn, but the effort required to learn the theory was just too intimidating.

I can’t claim to know why I had wanted to play piano any more than anyone else can explain the behavior of a six-year-old (though I suspect it may have been a random impulse triggered by the baby grand piano my parents had furnished the living room with), but if there was one thing my young self knew for sure, it was that finger exercises were the last thing I wanted to be spending my time on. I knew what music sounded like, and that wasn’t it. I signed up to play the piano, not memorize the E♭ scale.

I buried those damn books as far down into the bench as possible.

When I quit taking lessons, I didn’t have much to show for it. I knew the names of the keys, but I still couldn’t read music, I didn’t know the theory, and with the exception of a cheesy ballad or two, I couldn’t play almost anything. I had wanted to learn, but the effort required to learn the theory was just too intimidating. I moved on.

As I got older, the urge to learn piano regularly returned, surfacing more frequently as I became a teen and started developing my own music taste. At one point I managed, as many do, to learn Beethoven’s Moonlight Sonata, but it was the result of a painstaking process of brute-force “sounding it out” in conjunction with a near-mathematical analysis of the sheet music (every good boy does fine!) in which I calculated the letter value of each note and wrote it in on the sheet, as my teacher had once done. Though I could eventually play the whole thing in its entirety, it was ultimately only impressive as a feat of rote memorization — a fun party trick that left me no closer to being a fluent musician than I’d ever been.

I dreamed of taking it to the next level. Moonlight Sonata having done the job of convincing friends and family that I was a piano player, I suddenly found myself fielding requests to “play something” whenever a piano happened to be available, and without anything else to play, I’d plop myself down and dive right back into that same piece, whether or not it suited the mood of the moment, which it obviously never did. And then inevitably, people would shout, “You played that last time! Play something else!”, and in those moments, I wanted nothing more than to truly know how to play piano, to play a popular song, one I knew people wanted to hear. Or the ultimate dream: to be able to take a request. I knew it was possible — I’d seen people who could quickly figure out the chords to songs, or even make up music on the fly, and I desperately wanted to be able to do that. But, according to everything I’d been told, that required, theory. I needed to know the theory. Without it, I was doomed to disappoint my would-be fans. I would never be able to play something else.

The story goes that theory is the technical stuff. The story goes that it’s the discrete, fundamental building blocks upon which a craft is built. The story goes that in the process of learning a craft, the theory comes at the beginning. One is expected to learn the fundamental skills and techniques, independent of the craft, and then later on, unify them all in some kind of juggling act called art.

I don’t buy it.

Imagine teaching someone to walk by forcing them to learn the theory first. How does one boil walking down into a series of discrete steps?

  1. Shift weight onto right leg
  2. Simultaneously lean forward and raise left leg
  3. Simultaneously “fall” onto left leg while pressing up with the toes on the right leg.
  4. Push off from the toes of the right leg.
  5. You get the idea. This is terrible.

Imagine sitting in a classroom, learning such steps, and then being asked to link them all together in a fluid sequence afterward. If it’s even possible, it would be an immensely complicated juggling act. The idea that we should be able to take factual knowledge of erroneously platonified concepts from the logical parts of our brain and use them to drive a process that has nothing to do with logic, in real time, is insane.

And yet, this is how we teach. If you take Spanish lessons, you get a book of conjugations. If you take programming classes, you get a book on binary search trees. If you take piano lessons, you get a book of scales. And nothing kills motivation to learn faster than a book of scales.

When I was applying for college, I remember being asked to make an unexpected decision. My understanding of the general application strategy was that it was pretty straightforward: one simply picked a major, and then submitted an application to each of the schools that offered that major (ranked according to which ones were most likely to impress your parents’ friends). But after selecting electrical engineering as my major, I was presented with an additional dimension: I had to choose between “theory schools” and “hands-on schools”.

I was told that UCLA, where I eventually ended up, was a theory school, whereas Cal Poly in San Louis Obispo, where several of my good high school friends went, was a hands-on school. At the time it didn’t seem like a significant distinction — both yielded a B.S. in engineering. The end result was bound to be the same, wasn’t it? And yet each time I returned home for a break and connected with friends at other schools, it would hit me that much harder just how different those two approaches were. My school ranked higher. Why didn’t I feel as confident in what I was learning as they did?

Where my friends were being asked to solve problems, I was putting an enormous effort into collecting carefully segmented “chunks” of knowledge, waiting for some kind of electric shock I imagined would at some point spontaneously unite all those chunks, bringing them to life and allowing me to confidently declare myself an engineer. You won’t be surprised to learn that moment never came. I graduated four years later with a tenuous grasp on a set of loosely related concepts, and hardly a clue what to do with them.

What’s worse, I was bored. In the beginning, I had imaged learning would be empowering. I’d chosen EE due to an interest in building my own electronic instruments in high school, and I figured the cutting-edge researchers teaching the classes would broaden my imagination for what was possible and make me progressively more immersed in my major. But the opposite happened: Our courses ended up having nothing to do with the professors’ research. They were perfectly compartmentalized, platonic units of subject matter, sterilized of all context and application, and the more I heard, the less I was interested, and the more I was convinced I didn’t want to be an electrical engineer.

That’s not to say my time spent in college was a waste, though. On the contrary, my four years at UCLA were when I picked up the skills that have been most valuable to me over the last decade. It’s just that those skills didn’t actually have anything to do with being in school.

My lack of engagement in the EE curriculum afforded me time to pursue other interests, especially the type of interests that were actually interesting. I followed spontaneous whiffs of motivation in every direction. I watched Banksy’s Exit Through the Gift Shop and briefly became a street artist. I explored photography and film. But the thing that eventually came to consume most of my time outside school was a music blog I launched called Uh Oh Disco.

Uh Oh Disco was the prototypical mid-2000’s era nu-disco blog. Like everyone else at the time, I’d been wooed by artists like Justice and MSTRKRFT and felt the need to share my obsession on the internet, which is exactly what I did, eventually acquiring a fairly substantial readership and becoming a name-brand in what’s since been dubbed the “blog house scene.” But the real value of that project wasn’t the success of the blog itself, but the unexpected skills I picked up while running it, most of them out of necessity. I knew exactly what I wanted to create. But I didn’t have the skills to do it, and I didn’t have the money to pay someone. So I started Googling.

My initial research took me to Wordpress, which in turn spawned new paths — what’s an FTP server? How do HTML and CSS work? What’s a database? All of these concepts first presented themselves as barriers, blocking me from progressing on the task at hand. But as I figured them out and learned to navigate them, they became tools, and the best part about them was that unlike the concepts I learned in my EE classes, these tools were connected by a story. My story. The skills existed not as discrete, isolated chunks of knowledge, but as interlocking concepts, ready to be put to use in any number of applications. I didn’t need to study them to remember how each individual piece worked, because much like the image printed on a jigsaw puzzle prevents you from having to memorize the shape of each piece, my understanding of the project as a whole gave context to my understanding of the individual tools. And that’s what made those tools truly useful.

I just as easily could have taken a class on databases and learned about them in a neatly organized classroom unit, but had I done that, that knowledge would have sat stagnant in my brain, unconnected to intuition or other ideas, failing to be remembered even in cases where it could be useful, deteriorating until the only thing left is the useless bits I never even tried to remember: the professor’s funny accent. The terrifying elevators in the lecture building. The mechanical pencil.

Whose idea was it to learn like this? Why do we believe learning needs to happen this way, systematically disconnected from application? It’s possible it may have something to to do with the way we teach.

In the purest form of apprentice-master arrangement, learning happens by doing. A baker does not rely on pre-constructed lesson plans to pass his trade on to his student. Rather, the student joins the master baker in his practice each day, directly experiencing the effects of flour combinations, proofing techniques, rising times and environments, etc. Furthermore, the choices being made by the master are directly informed by his immediate needs. When one day he’s on a tight schedule and he throws a little sugar in with the yeast to make it rise faster, the student notices. A few weeks later, he’s similarly pressed for time, and the student reaches for the sugar, but the baker stops him. It’s hot today. The sugar will make the bread proof too fast and come out flavorless.

Lessons learned in practice aren’t discrete pieces of knowledge. They’re a web of interconnected ideas — what some might call intuition, or wisdom. That web is the stuff of true understanding. But if my experience at UCLA is evidence, this type of education isn’t what happens at scale. In fact, I believe it may be the very process of “scaling” education that has eroded its functionality.

Let’s pretend the baker from above takes on not one, but 3 apprentices. How does that change the experience? It still works, for the most part. He can still teach by example, and they can still learn by doing, but the experimental nature of the relationship is somewhat diminished. Only one of them gets to make the mistake with the sugar. The others may or may not see it happen. If they don’t, the baker might notice the value of the lesson and call the others over to explain to them what just happened. “Never add sugar when it’s too hot out,” he’ll say, to nods. They never made the mistake themselves or connected it to their own thought process and course of action. But they memorize the rule.

Now what happens if the baker scales up to 30 apprentices?

Remembering the important sugar lesson he taught in the past, and knowing he can’t possibly be lucky enough to catch each of them making the mistake on a hot day, he decides to simply make a note to explain the concept to them.

“If the day is hotter than 30 degrees and you’re making baguettes using any of the following types of flour, you should not use sugar to speed up the proofing process.” The idea that was once learned in context, has been platonified, and is being instructed as a conceptual idea, as a fact to be memorized, underutilized, and eventually forgotten. It has become theory.

Now scale up this teaching arrangement to a late-capitalist environment where education is treated as an economy-stimulating commodity, where 4.5 million students pay large sums of money to receive college degrees every year. How can we be sure each of them are receiving the same education in exchange for their money? How can we be sure they’ve been taught everything they need to know for a job in their field? It’s impossible without boiling true knowledge down into a set of discrete lessons to be learned. It’s simply an unfortunate side effect that doing so has confused us into believing this is the best way to learn, and an even worse side affect that attempting to learn this way may actually be preventing us from learning.

Having spent most of my adult life working as a web developer, I’m often asked by friends and acquaintances for advice on learning to code, and in the vast majority of cases, the person asking has already spent considerable time acquainting themselves with the various coding schools available. They want to know: Is General Assembly as good as it seems, or should I pick one of the niche coding schools? My answer tends to come as a surprise: I recommend self-teaching over either of them.

Since launching my startup in 2016, I’ve done my fair share of sifting through job applications from software developers, and though it initially ran counter to what I expected, I’ve learned that the best applications tend not to come from graduates from coding schools, nor those with engineering degrees from traditional universities. Of everyone I interview, these people have proven to have a solid grasp on how to talk about fundamental concepts, flout knowledge of various algorithms, and even to pass a tricky coding test. But when it comes time to put those concepts to use, they often fall flat. Many of them struggle to figure out how the concepts apply to a real, dynamic problem without a predetermined solution. They simply don’t know where to begin. Or worse: they expect to be taught where to begin.

That process of figuring out where to begin — that’s where real learning happens, and when it comes to software in particular, that learning process starts not when you pay a coding school $12,000 in tuition, but when you pick a project and run with it. It doesn’t matter that you don’t know what you’re doing. The process of figuring it out is worth more than anything you’ll learn in bootcamp.

At 27, I bought a piano. While at home for the holidays, Aphex Twin’s Avril 14 — one of my all-time favorite pieces of music — randomly started playing on the living room speakers, and looking for an excuse to avoid working on the massive jigsaw puzzle my brother had brought out, I decided I was going to try to learn to play it. When the end of the visit rolled around and I still hadn’t mastered the whole song, I was so determined to finish it that I went to Craigslist the moment I got home and starting searching for a piano.

The thought of being able to play the whole song was a powerful motivational force; I found a piano within a week or two, and before long, I had Avril 14 down.

Riding that high, I segued my enthusiasm into picking up a few more of my favorite piano tunes. Philip Glass’s Metamorphosis. Ryuichi Sakamoto’s Merry Christmas Mr. Lawrence. The theme song from Amelie. Granted, I was learning all of them just as I’d learned The Moonlight Sonata so many years prior — translating notes to letters, letters to keys, and them memorizing those keys. It was a brutal way to learn, and the process was contributing nothing to my understanding of theory or my ability to improvise, but I was okay with that. It was about me this time. I simply wanted to be able to play these songs, and I was perfectly content doing nothing more than performing them for myself. I did notice, though, that the more songs I picked up, the easier the transcription process was becoming. The names of the notes started to come to me quicker. I would remember having transcribed similar note combinations in previous songs and was able to recreate them more easily. I kept at it.

At no point had I decided that I was going to learn to site read music. It was simply a habit that had formed organically, as a result of me learning to play my favorite songs, my way.

After a year or so of picking up songs this way, I noticed something had changed. Whereas previously I would decode a note, find that it was an ‘A’, and then search for ‘A’ on the keyboard, I was now often skipping the last step. At least for the more common keys, I was associating notes on the page directly with specific keys on the keyboard, and this shortcut, I realized, was saving me a lot of time. The more music I read, the more this became the case, until eventually I could play some simple passages fluidly, without having to analyze the music first. I was sight-reading.

At no point had I decided that I was going to learn to site read music. It was simply a habit that had formed organically, as a result of me learning to play my favorite songs, my way.

Then something even more interesting happened: I tried to learn a Cole Porter song.

Every now and again I re-watch Woody Allen’s Midnight in Paris, and every time I do, I end up spending a few days lost in a Cole Porter rabbit hole, digging up old recordings and listening to everything Spotify has to offer on the guy. On one of those binges, I was possessed to add I Get a Kick Out of You to my queue of songs to learn. It was only when I went to download and print off the sheet music that it occurred to me how naive I was being: I Get a Kick Out of You is a jazz song.

Jazz isn’t played the same way that classical piano music is. Jazz songs typically have a “melody,” which is an extremely simple series of single notes, and a set of chords that back the melody. What happens within those chords is left to the musician. In other words, if I wanted to play this song, I would have to:

  1. Learn what all the chords on the above piece of paper meant
  2. Improvise my own version of the song

As you’ll recall, I never learned how to do either of these things. I still had no working knowledge of scales, chords and progressions, or any understanding as to how those two related to each other.

I wasn’t about to give up on learning the song, but I also knew I wasn’t willing to commit to learning a whole bunch of theory just to figure out how to play one song I liked. And so I decided I’d treat this song just like all the other songs I’d picked up: I’d convert the music on the page to something concrete that I could memorize.

I sat down on the bench with my laptop and picked out the first thing on the page: F-7. I had no idea what that meant, so I Googled it. F-A-C-E♭. I wrote it down and continued. F-6: F-A-C-D. B♭-7: B♭-D-F-A♭

Just as I was starting to get discouraged by the monotony of the task, I started noticing patterns. The -7 chords all had the same spacing between the notes, and they all started with the note they’re named after. The -6 chords were like the -7’s, except the top not was one lower. Of course, I had no idea why these patterns were what they were, but it didn’t matter — they made it easier to figure out the notes.

Eventually I’d finished the whole thing, and a few days after that, I could play a halfway decent version of it, which was exciting, because I’d never played this type of music before. It opened up a whole new realm of possibility. What other jazz songs could I play?

As it turns out, this basic knowledge of how to play various chords was enough for a lot of them. And just like before, with each new piece I learned, they came faster the more I played until I had yet another profound realization: I was recognizing intuitively which chords were going to come next. I’d never studied chord wheels. I’d never practiced scales. I still didn’t know why things fit together the way they did. But a sixth sense had emerged. I could feel how the different sounds interlocked.

And then it hit me.

This is how the theory is created. This is where it all came from. We didn’t make up a bunch of rules and then try to follow them. The learning came first. The exploring and the connection forming and the intuition building led people to develop an innate understanding, and theory was simply an attempt to describe that understanding in words when someone was asked, “How do you do that?” But as we saw in the step-based attempt to explain how walking works, the explanation is imperfect. It’s not that it’s wrong. It just can’t capture what it feels like to truly understand.

I’m now 31 years old. It’s been 25 years since I first sat down in front of the piano and dreamed of being able to play, not as an imposter performing a party trick, but as a musician who feels comfortable calling himself a pianist. Alas, I think I’m there, and I did it without ever having to commit the countless hours of “study” I thought were requisite to success. Instead, I followed a genuine interest, learning only what I needed for what was directly relevant to the task at hand, and eventually, those pieces became something bigger than the sum of their parts. Yes, there are certainly still important things I don’t know. But I will learn them when I need to know them. And if that time never comes, then maybe they aren’t as important as we think.

Of course, there’s a caveat to all this, which is that just as it’s a mistake to over platonify musical concepts, it’s similarly a mistake to take this advice as universal fact. There are always edge cases. In certain fields, particularly scientific ones, simply knowing that some theory exists can change one’s approach to a problem or open doors that may not otherwise have been open, and I’m not suggesting it’s a good idea to intentionally blind oneself to all theory. But it is regardless true that the process of learning — not memorizing, but truly learning — is far more likely to end in success when connected to an application that is genuinely interesting to the learner, and that the excitement that comes from doing something genuinely interesting is oftentimes the motivating force required to uncover such theory first-hand.

Once you’ve done that, it stays with you forever.

Continue reading