Back to journal

What deliberate practice actually means, and why most apps get it wrong

Most people who use the phrase “deliberate practice” have never read the paper it comes from.

In 1993, the psychologist K. Anders Ericsson and two colleagues published a study in Psychological Review that would reshape how the world thinks about expertise. The paper — “The Role of Deliberate Practice in the Acquisition of Expert Performance” — studied violinists at the Berlin Academy of Music. Its findings were specific, nuanced, and carefully qualified.

Fifteen years later, Malcolm Gladwell turned those findings into a single number: ten thousand hours. The nuance vanished. The qualifications disappeared. And an entire industry of apps and courses built itself on a misreading.

This matters if you are serious about improving how you speak. Because the difference between what Ericsson actually said and what Gladwell popularised is the difference between training that works and training that merely feels productive.

The violin study

Ericsson’s team divided violinists at the academy into three groups: the best (judged by faculty as having the potential for solo careers), the good (strong but not exceptional), and a third group training to become music teachers. Then they examined their practice histories in forensic detail.

The headline finding: by age twenty, the best violinists had accumulated roughly ten thousand hours of solitary practice. The good violinists had logged around eight thousand. The future teachers, about four thousand.

But the finding that mattered — the one Gladwell skipped — was not about hours. It was about what happened inside those hours.

Three levels of practice

Ericsson drew a sharp distinction between three kinds of practice. Most people never get past the first.

Naive practice is repetition without structure. Playing a song you already know. Running the same route at the same pace. Using a language app for fifteen minutes a day while half-watching television. It feels like practice. It is not. Performance plateaus quickly and stays there.

Purposeful practice adds goals, focus, and feedback. The musician works on a specific passage. The runner tracks split times. The language learner targets a weak area. This is better. But it lacks one critical element.

Deliberate practice is purposeful practice plus an established training methodology, designed by or with a teacher who understands the path from novice to expert. It requires activities that are specifically designed to improve performance, that demand full concentration, that are not inherently enjoyable, and that incorporate clear feedback.

The distinction is not academic. It is the reason most self-improvement efforts fail. People log hours. They feel busy. They confuse effort with progress. And they plateau — sometimes for years — without understanding why.

What Gladwell got wrong

In Outliers (2008), Gladwell reduced Ericsson’s research to a rule: ten thousand hours of practice makes you world-class. Ericsson spent the rest of his career correcting this.

In his 2016 book Peak, Ericsson is explicit: there is no magic number. The ten thousand figure was an average for a specific group of violinists at a specific point in their careers. It was never a universal threshold. Some domains require far fewer hours. Some require far more. And crucially, ten thousand hours of naive practice produces nothing but ten thousand hours of mediocrity.

The type of practice matters more than the quantity. This is the single most important sentence in the entire body of expertise research, and it is the one most consistently ignored.

Why most apps are naive practice in disguise

Consider the typical language-learning app. It offers daily lessons of five to fifteen minutes. It tracks streaks. It awards points and badges. It congratulates you for showing up.

By Ericsson’s taxonomy, this is textbook naive practice — pleasant, bite-sized, low-concentration repetition with no meaningful feedback on the dimension that matters: how you actually sound.

Research confirms the problem. A 2015 study by Hanus and Fox, published in Computers & Education, found that gamified learning environments led to lower motivation, lower satisfaction, and lower exam scores compared to non-gamified ones. The extrinsic rewards (badges, leaderboards, streaks) actively undermined intrinsic motivation — a finding consistent with decades of self-determination theory research by Deci and Ryan.

The apps are not designed for mastery. They are designed for retention. Their metric is daily active users, not skill acquisition. The streak is not a training tool. It is an engagement mechanism. And it works — for the company, not for you.

What deliberate practice looks like for speech

If naive practice is running through flashcards, and purposeful practice is studying grammar with intent, then deliberate practice for speech is something more specific: isolating a unit of speech, listening with full concentration, reproducing it, and comparing the result against the original.

This methodology has a name. It is called shadowing, and it has been the core training technique of conference interpreters for decades. Lambert’s research on interpreter training (1992) documents how simultaneous listen-and-repeat drills build phonological processing speed, prosodic accuracy, and deep listening comprehension — skills that no written exercise can develop.

The neuroscience supports this. Liberman and Mattingly’s motor theory of speech perception, first proposed in 1985 and supported by subsequent neuroimaging research, argues that perceiving speech and producing speech share the same neural substrate. When you hear a sentence, your brain activates the motor commands needed to produce it. Listening and speaking are not two separate skills. They are two expressions of the same system.

This is why passive listening alone does not work. You must produce the sound. You must engage the motor system. And you must do it with a specific reference — a sentence spoken by someone whose delivery you want to internalise.

Baddeley, Gathercole, and Papagno’s work on the phonological loop (1998) adds another dimension. The phonological loop — the part of working memory that rehearses heard speech subvocally — is a primary mechanism for acquiring new phonological patterns. When you listen to a sentence and repeat it aloud, you are engaging this system at full capacity. When you swipe through a flashcard, you are barely activating it at all.

The compound effect of one sentence

Ericsson’s research emphasises that expert performers do not practise everything at once. They isolate the smallest meaningful unit of skill and work on it until it is internalised. The best violinists do not play full concertos in practice. They take four bars and repeat them — slowly, precisely, with a clear model of what correct sounds like — until the gap between their performance and the model closes.

Speech training should work the same way. One sentence. Not a paragraph. Not a dialogue. One sentence, spoken by someone whose clarity, rhythm, and intonation represent the standard you are working toward.

Listen to it. Study the stress pattern. Notice where the speaker breathes, where they pause, which syllables they lengthen. Then reproduce it. Compare. Adjust. Repeat.

This is not glamorous. It is not gamified. It does not award you a badge for showing up. It is the kind of work that most people abandon after one session because it feels too slow, too repetitive, too ungratifying.

It is also the only kind of practice that reliably closes the gap between where you are and where you want to be.

The discipline, not the tool

Ericsson was careful to note that deliberate practice is effortful and not inherently enjoyable. This is not a design flaw. It is the point. The effort is the signal that your brain is building new representations, not reinforcing old ones.

Any tool that makes practice feel effortless is, by definition, not facilitating deliberate practice. It may be facilitating exposure. It may be facilitating recognition. But it is not building the motor patterns, the phonological precision, or the prosodic control that constitute genuine improvement in speech.

The question is not which app has the best streak mechanic or the most languages or the cleverest AI. The question is simpler and harder: are you doing the work that actually changes how you sound?

If the answer involves badges, the answer is no.

References

  • Ericsson, K.A., Krampe, R.Th., & Tesch-Römer, C. (1993). “The Role of Deliberate Practice in the Acquisition of Expert Performance.” Psychological Review, 100(3), 363–406. DOI: 10.1037/0033-295X.100.3.363

  • Ericsson, A., & Pool, R. (2016). Peak: Secrets from the New Science of Expertise. Houghton Mifflin Harcourt.

  • Gladwell, M. (2008). Outliers: The Story of Success. Little, Brown and Company.

  • Lambert, S. (1992). “Shadowing.” Meta: Journal des traducteurs, 37(2), 263–273.

  • Liberman, A.M., & Mattingly, I.G. (1985). “The Motor Theory of Speech Perception Revised.” Cognition, 21(1), 1–36. DOI: 10.1016/0010-0277(85)90021-6

  • Baddeley, A.D., Gathercole, S., & Papagno, C. (1998). “The Phonological Loop as a Language Learning Device.” Psychological Review, 105(1), 158–173. DOI: 10.1037/0033-295X.105.1.158

  • Hanus, M.D., & Fox, J. (2015). “Assessing the effects of gamification in the classroom.” Computers & Education, 80, 152–161. DOI: 10.1016/j.compedu.2014.08.019

  • Deci, E.L., & Ryan, R.M. (2000). “The ‘What’ and ‘Why’ of Goal Pursuits: Human Needs and the Self-Determination of Behavior.” Psychological Inquiry, 11(4), 227–268.