None of my teachers ever spoke with me about how to practice. They didn’t suggest strategies or give me tips. I guess they assumed, since I memorized so easily, that I didn’t need any help. I had what’s called a “good ear,” and I could hear the piece in my mind. By the time I had developed the motor skills to play a particular piece, all of the melodies, rhythms, and harmonies were in my head, and I counted on that when I performed. I assumed that good auditory memory was all I needed. But after a couple of bad experiences following grad school, I decided that I needed to develop a more secure system for learning and memorizing.
In my first post on memory, I wrote that many neuroscientists consider learning and memory to be two sides of the same coin – that learning is the process by which new information is encoded in the nervous system and memory is the encoding itself of the information. So how we learn determines the quality of our memory recall.
Practice is the primary way that musicians learn a piece, so how we practice is key to learning and memory. I had very good auditory recall, but I didn’t have much in the way of a backup. So when something happened in performance that interrupted the auditory recall, I was in trouble.
So what kind of practice ensures that learning and memory are more effective – and secure?
Many people are convinced that massed practice is the best way to learn new information. Massed practice is the focused, single-minded repetition of something we are trying to learn, whether it’s the exposition of a Beethoven Sonata, the conjugation of French verbs, or a strong backhand in tennis.
Musicians don’t refer to what we do as massed practice, but that’s what it is, and we’re very fond of it. We spend endless hours repeating a section of music until we are playing it quite fluently and then we feel good about how much we have accomplished.
In sports, continuous repetition is called blocked practice, and it refers to practicing a particular skill. If you’re learning tennis, you may do a lot of repetitions or a “block” of your backhand, then move on to a block of repetitions of your serve, then on to your follow through.
But whether it’s called massed or blocked practice, it doesn’t work.
Massed practice has been found in one research study after another to be the least productive way to practice. And yet, even faced with results of these studies, most people will continue to use repetitive practice because it feels good. We always play our best at the end of a practice session of repetition. We feel as though we have made progress when we spend a couple of hours on a particular section of music and then play it nearly flawlessly – even when we find out later that we can’t play it as well as we thought we could. (We have all either had a student or been the student who comes to a lesson, doesn’t play very well, and then wails, “But I played it perfectly in the practice room!”)
When we practice by repetition, we are really just recalling the material from short-term or working memory. It’s no different than repeating a long password over and over until you have to use it. But we don’t really know how well we know the piece (or the password) until hours or days later when we have to retrieve it from long-term memory.
Continuing to use repetitive practice is the musical version of Einstein’s famous comment: “Insanity is doing the same thing over and over again and expecting different results.” Obviously it is necessary to spend tens to hundreds of hours on a piece – that’s how we learn the motor skills necessary to play it. But having the motor skills to play a piece is quite different from actually knowing the piece and being able to play it from memory. And it turns out there are better ways than endless repetition to achieve more solid learning and memory.
Researchers have found that learning tends to be deeper and longer lasting when it requires effort. So in this and the following two posts, I am going to write about ways to practice that require some effort – that are more of a struggle and cause you to work harder, but also lead to stronger learning and more reliable memory.
Roger Chaffin is a cognitive psychologist at the University of Connecticut who has done a number of studies involving the learning process of professional musicians preparing for performance. He discovered that professional musicians, performing expertly from memory, used the same principles that that had been identified by K. Anders Ericsson and Walter Kintsch* in 1995 in their study on expert memory in other disciplines: 1) meaningful encoding of the musical material, 2) use of a well-learned retrieval structure to access the material, and 3) practice to ensure rapid retrieval from long-term memory.
When we practice, we are encoding the musical material, and I wrote about encoding in “From sight-reading to memorized performance.” When we use massed or repetitive practice, we often are paying attention only to the motor skills necessary to play the piece – to get the tough passages “under our fingers.” That means that we are encoding motor and auditory information, but what about all of the other information found in a piece of music?
Chaffin’s “meaningful encoding” refers to encoding and consolidating our conceptual (declarative) memory of a piece as well – to the facts that we know about the piece, such as what the structure is, where the section boundaries are, where the key changes are, what the notes are in a complex chord, where rhythmic, harmonic or melodic patterns occur, where we may have used an unusual fingering, and something Chaffin calls “switches,” which are places where two (or more) repetitions of the same theme begin to diverge.
A pianist I once knew told me about a performance in which he had played Luciano Berio’s Sequenza IV, written in the mid sixties. Sequenza IV is a very thorny, virtuosic piece that alternates between complicated chords and very fast, single-voice chromatic figures. There is a place about half-way into the piece where the musical material is just like that found shortly before the end. In the first statement, you go one direction and the piece continues, toward the end you go another direction, and the piece finishes.
The pianist spoke with the audience about the piece, said it would last 10 minutes and sat down to play it from memory. Unfortunately, he took the wrong turn (Chaffin’s “switch”) at the crucial place and ended up at the end – finishing the piece in 6 minutes, which no one would have thought about, had he not said it would last ten. You’d better believe that when I later memorized the piece, I made sure to pay particular attention to learning the “switch,” so that I knew exactly how the first statement of the material differed from the second.
Of course we need to encode in our brains the motor movements necessary to play the piece and the auditory information about the piece, but we also need to encode all of our conceptual knowledge about the piece. And we need to organize that conceptual knowledge as retrieval cues (Chaffin’s second point), or as a retrieval structure that we can access readily during performance.
Chaffin and pianist Gabriela Imreh, with whom he did one of his first studies, have written Practicing Perfection: Memory and Piano Performance (along with third author, psychologist Mary Crawford), in which they identify stages of the learning process, characteristics of expert practice, and practice strategies.
In speaking about her practice of Bach’s Italian Concerto, the third movement of which was the basis of the study, Imreh comments: “My fingers were playing the notes just fine. The practice I needed was in my head. I had to learn to keep track of where I was. It was a matter of learning exactly what I needed to be thinking of as I played.” (p. 224)
If the only place in the piece we can access when we have a memory slip is the beginning and we have to start over, then we haven’t encoded enough information – we don’t have a retrieval scheme. The structure of a piece is itself a ready-made retrieval scheme. We can encode into long-term memory where the exposition, development, and recapitulation are in a sonata and exactly what is happening at the beginning of each. We can encode into long-term memory how the ABA sections begin in a piece, or how the Coda begins. With complex new music, we can encode the particular structural sections of the work. Knowing the structure of a piece provides access to chunks of information in long-term memory.
While the structure of a piece remains the same for everyone, other retrieval cues may vary. One person may use texture changes as retrieval cues, another difficult fingering patterns, or chord progressions, technical difficulties, melodic or rhythmic patterns, phrasing, pages. We need to use what makes the most sense to us, or perhaps what has been the most difficult to learn.
The retrieval cues may vary, but what is absolutely crucial is Chaffin’s third point: practice to ensure rapid retrieval from long-term memory.
What often happens when we have a memory slip is that we can’t access our conceptual memory of the piece (our retrieval cues) fast enough to tell us where we are and get back on track. Our motor memory (fingers) is moving faster than our conceptual memory. We flounder until we can finally think of some landmark in the piece.
We need to be able to access our conceptual memory (retrieval cues, landmarks) as fast as we are playing the piece in order to feel secure about memory. Imreh says that she needed to be “learning exactly what I needed to be thinking of as I played.” That means identifying a set of cues or markers in the piece that you can access at the same speed as you are playing, so that if (when) the unexpected happens, you always know where you are.
This doesn’t diminish what some people refer to as the “flow state,” that state of being totally immersed in the music during performance. In fact, having ready access to your conceptual memory of the piece makes the flow state more possible, because encoding conceptual (factual) memory simultaneously with auditory and motor memory gives us a great deal more freedom in performance.
No matter how experienced a musician you are, something inevitably goes wrong because we’re all human. But if we have instant access to our conceptual memory, we can restart the motor sequence – usually without the audience being aware that anything has happened.
In the two posts that follow, I’ll write about strategies for practice that are far more work than massed or repetitive practice, but that have been shown to be far more effective for learning and memory.
* Ericsson, K.A., & Kintsch, W. (1995). Long-term working memory. Psychological Review, 102: 211-245.