This week there is a very cool conference about practicing taking place at the Norwegian Academy of Music. Entitled, “Teaching of Practicing – new knowledge and change,” the conference features presentations by many of Europe’s leading researchers of musical practice. The keynote speech today was given by Andreas Lehmann of the Hochschule für Musik in Würzburg. He is a leading proponent of the theory that the cause of elite performance is deliberate practice.
In his speech today, entitled “25 Years of research on deliberate practice in music - What do we know?” Professor Lehmann strived to give the audience a definition of deliberate practice. Deliberate practice is :
“Structured activity, often designed by teachers or coaches with the explicit goal of increasing an individual's current level of performance. (···) it requires the generation of specific goals for improvement and the monitoring of various aspects of performance. Furthermore, deliberate practice involves trying to exceed one's previous limit, which requires full concentration and effort.”
Professor Lehmann contrasted deliberate practice with work, play, and gaining experience, none of which are done solely for the sake of improvement. He spoke about how deliberate practice is the most valuable type of practice that can be done, but it is only a part of most people’s practice regimens. Deliberate practice is the cream at the center of the practice éclair.
Deliberate practice requires resources, effort, motivation and concentration. It consists of optimized training activities performed at sustainable optimal intensity. This limits the amount of daily deliberate practice possible to a maximum of four hours a day.
Having defined deliberate practice, Professor Lehmann admitted that we can’t always recognize deliberate practice when we see it. Self-regulation, though, is a critical component. One must be one’s own teacher in the practice room, apply appropriate problem solving strategies to difficult passages, set goals, and strive toward them. He says that deliberate practice is not always enjoyable. In fact, a good practice session feels like kicking yourself in your rear-end. The most relevant and least enjoyable aspect is practice of problem spots. He stressed that practicing alone is of higher relevance toward musical progress than other “more enjoyable” activities such as playing music with others.
Professor Lehmann stressed that exceptional musicians come from exceptional circumstances – a parent who supervised their practice, a dedicated first teacher who gave multiple lessons a week, the social environment of pre-college music programs, etc. This explains the ability of elite musicians to engage in deliberate practice.
Now that I have summarized Lehmann’s useful advice for students, I will discuss the talent vs. practice debate.
One of the foundational studies in musical practice is “The Role of Deliberate Practice in the Acquisition of Expert Skills” (Ericsson et al 1993). This study provided the framework for subsequent research in musical performance. Ericsson finds that the elite performers in his study spent significantly more time practicing than the non-elite performers. He claims that this extra time spent practicing explains the difference in performance ability.
One of the problems with the study is the lack of random sampling. Anyone who studies statistics knows the importance of random samples in order to provide generalizable results. The subjects for Ericsson’s study were all music students at an elite academy of music in Berlin. It is conceivable, therefore, that talent is controlled for in the study. If all of the students at the music school have similar talent levels, then it is a trivial conclusion that more practice makes one a better musician. That’s why we practice, after all. But, we are looking for an answer to the talent vs. practice question. Ericsson rejects the talent hypothesis (that talent is a significant factor in accounting for elite performance) on the basis of the evidence provided in his study.
To be fair, Ericsson acknowledges this methodological problem: “It is hard to imagine better empirical evidence on maximal performance except for one critical flaw. As children, future international-level performers are not randomly assigned to their training condition. Hence once cannot rule out the possibility that there is something different about those individuals who ultimately reach expert-level performance.” Nevertheless, he sticks with his conclusion.
I took this problem of random sampling up with Professor Lehmann, who has done a meta-analysis of all the studies linking practice and musical performance. He admitted that all studies of elite musical performance have this inherent problem. Randomization, he says, is impossible with elite performers because they have the same specific traits. Perhaps, then, the conclusions of studies on elite performers are not generalizable to the general population, as Malcolm Gladwell implies in his book, Outliers. Maybe studies linking practice and performance level should examine a random sampling of musicians (that may or may not include any elite performers) rather than studying elite performers exclusively.
Professor Lehmann indicated that in order to provide evidence for the talent hypothesis I would need to find two groups of people, people who had practiced enormous amounts without succeeding and those that have succeeded without practicing huge amounts. Here is my partial stab at finding these populations. Christian Tetzlaff makes a point of explaining how little he practiced when he was young. He emphasizes how he learned through sight-reading duets and playing in orchestra, an experience which, if true, directly contradicts the primacy placed on practicing alone. It is worth considering that Tetzlaff may be, for some reason, exaggerating his lack of practice. Even in the following article, though, he takes talent for granted. He says, “if you are talented, you will succeed with more thinking and fewer hours of practicing.” His focus on using your head instead of your fingers also implies that intelligence is important for being a great musician, and extensive evidence demonstrates that intelligence is not equally distributed (i.e. the same may be true for talent). There might be other musicians who fit Tetzlaff’s mold, but I would have to search longer to find them.
As for the first question (where are the populations that practice a lot without achieving excellence), I can venture a hypothesis. It is likely that many of these people don’t find a career in music. If you compare the hours spent practicing by many musicians at music conservatories with those of Christian Tetzlaff, nearly every student qualifies to be in the population of “practices a lot, doesn’t perform at a world-class level.”
The fact is that there are a mere 120 or so studies linking musical performance to various factors. There just isn’t enough research to make definitive claims. While Ericsson 1993 demonstrates the importance of deliberate practice, this study cannot be said to definitively reject the talent hypothesis. I have one final observation directed at social science researchers who are eager to reject the talent hypothesis. How to account for the persistent belief among music educators of the existence of talent? Generations of music instructors have taken the existence of differing levels of talent as a given. It is obvious that some students grasp concepts easily and can master technical challenges with little effort. Teachers can see this with their own eyes in the span of a single lesson. If all “talent” is the result of practice (or extraordinary environments), do social science researchers really think that music educators are too dim to pick up on this themselves?
1. Is there an inherent bias in social science to look for social explanations?
2. Is it perceived to be more socially beneficial to argue that talent is not important and does this affect experimental design?
3. Would it be possible for a university to create an advanced institute of medical science and social science in order to try to settle the nature vs. nurture debate?
4. Can we design an optimal experiment in which test subjects must master a task unrelated to anything people do in real life (hence test subjects have no prior knowledge that can apply to this task) and examine how a large sample masters this task in a laboratory? Would there be different strategies used? Different time taken toward mastery?
5. Why not think of musical performance as a mathematical function of many variables? The inputs are deliberate practice hours, quality of teaching, an index for the social environment, talent, motivation, etc. The coefficients for these variables represent their relative weights for the whole population. Consider that the individual coefficients might have their own distribution across this population and that for one individual talent might be the important factor whereas for another individual the quality of teaching is salient. In such a function, how does one account for the fact that the inputs are themselves dependent variables of the other inputs (social environment, motivation and practice hours clearly being linked)?