Outstanding Performers: Created, Not Born?


from Science Spectra, 1999, Number 18.

Outstanding levels of performance in areas such as memory, chess, sports or music are commonly ascribed to innate talent. Dr. David Shanks of University College, London, England, describes evidence for the role of deliberate practice in achieving these levels of performance and questions the need for the notion of talent. This "anti-nativist" hypothesis is also evaluated in the area of intelligence, where, again, it appears that outstanding levels of achievement may not be due to innate ability.

As a young man, Rajan Mahadevan regularly amazed his friends in India with his exceptional numerical memory. Eager to prove the superiority of his memory, and on the advice of the editors of the Guinness Book of Records, Rajan set himself the task of committing to memory as many of the digits of pi as he could manage. The culmination of this effort came on July 4th, 1981, when Rajan correctly recalled the first 31,811 digits of pi. It took nearly three hours at a rate of over three digits per second, and was obviously thirsty work: according to newspaper reports he had to pause after the first 10,000 digits for a can of Pepsi. "Why did you stop at that particular point?" Rajan was asked. "The 31,812th digit, I don't know why, I am always stumbling over that one," he replied.

Figure 1.
Figure 1: The first 1024 digits of pi. This represents a mere three percent of the 32,000 or so digits that Rajan Mahadevan was able to recall from memory.

To those of us who struggle to remember people's names or telephone numbers, where we put our keys, and the things we have to do today, Rajan's memory ability seems quite exceptional, and it is almost impossible to avoid the inference that Rajan was born with a gift for memory, some innate talent that later in life endowed him with a memory well beyond the norm, in much the same way that Mozart was born with a gift for music and Garry Kasparov with a gift for chess. Yet recent research has begun to challenge the idea of giftedness or talent as the basis for excellence in performance. Instead of attributing Rajan's extraordinary memory abilities to nature, evidence is now emerging that exceptional performance in memory, chess, music, sports and other arenas can be fully accounted for on the basis of an age-old adage: practice makes perfect.

To appreciate the extent to which memory capacity can be improved by practice, consider some data collected by K. Anders Ericsson, then at Carnegie-Mellon University in Pittsburgh, and his colleagues. Ericsson was interested in a memory capacity called digit span. This refers to the number of digits that can be held in memory when presented at a rate of one per second and which can then immediately be recalled in the correct order. If read a six- or seven-digit number, such as a telephone number, most people can correctly recall the number from short-term memory. However, if the number is eight or nine digits or more in length, correct recall becomes extremely difficult. On the face of it, memory span would appear to be a relatively fixed property of human memory, yet Ericsson and his colleagues showed that span can be increased by practice up to in excess of 100 digits. Ericsson trained two people, called SF and DD, to improve their memory spans over many months of practice during which they were given numerous digit sequences to memorize. At the outset, both SF and DD had normal digit spans, but their spans improved at a rate of about one digit for every two hours of practice.

How was this improvement achieved? SF and DD were given no particular guidance in how to improve their spans. However, SF was a very good long-distance runner and had extensive knowledge of expected times for different race distances. Thus when a digit sequence included 3492, he encoded it as 3 min 49.2 sec, near world-record time for the mile. It is obvious that with an encyclopedic knowledge of running times, a long digit sequence can be more easily remembered if it is translated into a smaller number of distinctly memorable running times. DD was also a long-distance runner and used a method similar to SF's, whereas a former winner of the World Memory Championships, Dominic O'Brien, used a different method.

O'Brien's memory span, also in excess of 100 items, was based on the classical "method of loci" originally attributed to the poet Simonides of Ceos. Legend has it that Simonides was in the middle of reciting a lengthy poem at a banquet when he was urgently called outside. Just then, the roof of the banquet hall fell down, killing everyone inside and mangling the bodies so badly that they couldn't be recognized. However, Simonides was able to remember who had been sitting at each place and so the bodies were identified. As the legend suggests, the method of loci involves creating a series of mental images in which each of a set of to-be-remembered objects is visualized in a different place on a route. To recall the set, you simply imagine yourself visiting all of the places on the route.

Although Rajan's extraordinary feat in memorizing pi is not based on memory span but on storage in long-term memory, it is clear that like SF and other mnemonists, Rajan relies very heavily on mnemonic strategies that he has practiced extensively. Thus he recognizes familiar patterns such as 312 as the area code for Chicago. Moreover, after training specifically to improve his memory span, SF became much faster at committing lengthy sequences of digits to long-term memory, exactly the ability required by Rajan in memorizing pi. Thus memory feats such as those demonstrated by Rajan are enormously dependent on practice. Finally, and most importantly, when tested on other forms of material such as prose passages or complex geometrical shapes, Rajan's memory turns out to be distinctly average.

Figure 2
Figure 2: Average digit span for four subjects studied by Ericsson and his colleagues. All began the study with normal spans, but after months of practice in which they were given many digit sequences to memorize, their spans increased. This is particularly notable in the cases of SF and DD. It is unknown how much further digit span can be increased.


No one would dispute that practice is an important component of achieving exceptional levels of performance in music, chess, sports and so on. After all, even the chess prodigy Bobby Fisher spent many years immersed in chess strategy and tactics before becoming world champion. But it is commonly assumed that both talent and practice are needed to achieve renown, where talent refers to some innate predisposition to make rapid advances in a particular field.

Yet evidence for the contribution of talent over and above practice has proved extremely elusive. In another recent study, Ericsson and his colleagues studied young pianists and violinists in their early 20s at the Music Academy of West Berlin, Germany. They asked the music professors to nominate the best young musicians, those who they thought had the potential for careers as international soloists, as well as others whose potential they regarded as not quite so great, and a third group who were most likely to become music teachers. Hence, in terms of achievement, the first group comprises the most exceptional musicians, the second group the next most outstanding, and the last group the least exceptional.

If "talent" is the primary factor, we might assume that these three groups differ in their innate giftedness for music and that this explains their different levels of achievement. If a person is innately gifted, then he or she can very rapidly attain an outstanding level of performance once the basic skills and knowledge required have been mastered. Yet Ericsson and his colleagues obtained a surprising finding: the best musicians had simply practiced more across their lives than the next best ones, who in turn had practiced more than the ones likely to become music teachers. Each of the musicians was asked to estimate approximately how many hours a week they had practiced each year since the outset of their musical training, and these estimates yielded cumulative totals of about 10,000 hours for the best musicians, followed by 8,000 for the next best ones and 5,000 for the least accomplished. The musicians also kept diaries for a week, recording their exact amounts of practice, and these yielded comparable differences, suggesting that the retrospective estimates were roughly accurate.

Figure 3
Figure 3: Musical Mozarts, 1763, after Carmontelle. Leopold Mozart (1719 - 1787) playing the violin, accompanied by his seven-year-old son Wolfgang Amadeus (1756 - 1791), while his daughter Marie Anne sings. Mozart is often regarded as the quintessential child prodigy, whose achievements are attributable to natural giftedness, yet his early environment was hardly normal. His father was himself an outstanding musician and exerted great pressure on the young Wolfgang to develop his musical abilities.

The estimate of 10,000 hours from this study is interesting because there is now abundant evidence across a range of abilities that roughly this much practice is needed to achieve international levels of performance. For instance, it has been estimated that about this many hours of practice is usually completed between the time of first learning the rules of chess and becoming an international master. Ericsson has argued that similar amounts of practice are seen in first-rate sportspeople, writers and scientists.

Two other lines of evidence converge on the conclusion that innate talents play a minimal role in the development of exceptional performance. First, if giftedness is important, then it ought to be the case that children identified at an early age as having high ability in a particular domain are exactly those children who go on later to achieve high levels of accomplishment. If a child learning the violin possesses an innate musical talent, this should be fairly evident as soon as the child begins playing the instrument. In contrast, if later ability is simply dependent on amount of accumulated practice, there should be little or no relationship between early signs of ability and later achievement.

Figure 4
Figure 4: Cumulative amounts of practice achieved over their lifetime by different groups of violinists in a study by Ericsson, Krampe and Tesch-Römer (Psychological Review, 100, 363-406, (1993)). By the age of 20, the best musicians at the Music Academy of West Berlin (as judged by the music professors) had practiced for about 10,000 hours, the "good" ones for about 8,000, and those trained to become teachers for about 5,000. Also shown are the data from a group of older professional violinists. The results highlight the close relation between practice and achievement.

To test these predictions, John Sloboda, Michael Howe and their colleagues at the Unit for the Study of Musical Skill and Development at the University of Keele, England, studied a large number of children between 8 and 18, some of whom were sufficiently good musicians to have won places at a selective music school. The remainder were divided into further groups of differing musical ability, with the least musical group comprising children who had been relatively unsuccessful in learning an instrument, most giving up after less than a year.

Sloboda, Howe and their colleagues then interviewed the parents of these young musicians and tried to find evidence for early signs of musical talent in those who later went to music school. Despite the range of musical ability in the three groups, almost no differences between them were found. Thus the most musically gifted young people were reported to have first shown a liking for musical sounds at around 1.9 years of age, but this was no younger than in any of the other groups: for example, the children who went on to be unsuccessful in learning an instrument first showed signs of liking musical sounds at around 1.7 years of age. Similarly, there were no differences in age of first making rhythmic or dance movements to music or of first requesting involvement in musical activity. Only one characteristic, the age at which the children first sang, appeared significantly earlier in the most able group. However, even this is probably not evidence of innate musical ability, since these children experienced a greater degree of early musical input from their parents.

The second line of evidence suggesting that innate talents play a minimal role in the development of exceptional performance concerns the rate of improvement with practice. If talent or giftedness plays a significant role over and above practice, then we would predict that a talented individual would make more progress from a given amount of practice than a less talented one. Surely the young Garry Kasparov learned the complexities of a new chess opening in far less time than it would take an average chess player? On the other hand, if practice is essentially the only ingredient in the development of outstanding ability, then talented and less talented individuals will require the same amount of practice to progress by equal amounts.

The available evidence suggests that the latter of these predictions is closer to the truth. Thus in a further study of their young musicians, Sloboda, Howe and their colleagues asked them to estimate how many hours of practice per day they had engaged in each year since taking up their instrument, just as Ericsson and his colleagues had done in their study. Since the musicians were regularly taking musical grade exams, Sloboda, Howe and their colleagues were able to use this as a measure of musical progress and could therefore calculate the amount of practice that took place between successive grades. The surprising result was that the most gifted children required just as much practice as the less gifted ones: in fact, if anything, there was a tendency for them to require more. For instance, the most gifted group required on average 971 hours of cumulated practice to reach Grade 4, while a less talented group took 656 hours. The high figure in the former group is boosted by a small number of young people who practiced for exceptionally long periods of time, but even when these individuals are excluded, there is still no evidence that more gifted people can get by on less practice.

Overall, then, there is little evidence that talent contributes to the achievement of exceptional levels of performance over and above practice. Maybe we should dispense with the notion of talent altogether?


In the debate over nature versus nurture, it is clear that the results obtained by Ericsson, Sloboda, Howe, and others strongly emphasize the importance of nurture. The environment is dramatically different for a child who practices a great deal and one who does not. Yet it is not only in specific abilities such as memory, music, chess, or sports that the effects of the environment can be seen. It is now clear that even intelligence, that traditional favorite of psychologists, can be dramatically influenced by changes in the environment.

The most radical way to change the environment is to adopt a child at birth from a household of low social and economic status (SES) to one of high status, or vice versa. Christiane Capron and Michel Duyme from the Université Paris V, France, searched adoption agency records for children given up for adoption at birth and who were from either high or low SES families and who were adopted by either high or low SES families. They found overall that adoption into a high-SES family compared to a low-SES one increased average intelligence, as measured by intelligence quotient (IQ), by over 11 IQ points. Thus a relatively extreme change of environment can have a significant impact on "intelligence."

However, Capron and Duyme also found that regardless of the SES of the adoptive family, children whose biological parents were of high SES had IQs some 16 points higher than ones whose biological parents were of low SES. Since the children were separated at birth from their biological parents, it seems that the only way the biological parents could influence the intelligence of their children is through genes. Surely this indicates that intelligence is partly inherited? And if that is the case, then might there not also be genetic factors in memory, musical ability, and so on?

Some of the most compelling evidence for this suggestion comes from studies of identical and non-identical twins. Identical twins (who are called monozygotic, MZ) share 100 percent of their genes whereas non-identical twins (called dizygotic, DZ) share only 50 percent of their genes. If we measure memory, intelligence, or other abilities in pairs of twins and if genes contribute to these abilities, then we would expect the correlation between identical twins to be greater than the correlation between non-identical twins. Since identical twins are genetically identical, they should appear more similar in abilities that have genetic components than non-identical twins. There is now abundant evidence that this is true: in the case of memory span, for instance, correlations are greater than 0.60 for monozygotic twins compared to only about 0.20 for dizygotic twins.

Figure 5a
Figure 5b
Figure 5: Panel (a) illustrates the separate placentas and fetal circulation systems of dichorionic twins, while panel (b) shows the single placenta and shared vascular communication of monochorionic twins. Dizygotic (non-identical) twins are always dichorionic, whereas monozygotic (identical) twins can be either mono- or dichorionic. Thus the prenatal environment tends to be more similar for identical than for non-identical twins and this in turn might explain the higher correlation of cognitive abilities in the former.


But are such findings really evidence for innate talents? It must be borne in mind that the MZ-DZ difference is evidence for a genetic contribution to ability only if it can be convincingly argued that the environment is as similar for a pair of DZ twins as it is for a pair of MZ twins. Certainly, this will be the case with respect to important factors such as the amount of intellectual stimulation the parents provide: regardless of whether twins are MZ or DZ, the mental stimulation they receive from their parents will be highly similar. Moreover, it is known that the IQ correlation of same-sex and different-sex DZ twins is virtually identical. The environment should differ more for different-sex twins than for same-sex twins, but this seems to make little difference in terms of the observed IQ correlation.

Thus it is unlikely that differences in the post-natal environment of MZ and DZ twins can explain the fact that MZ twins tend to be more similar in ability. But what about the pre-natal environment? One important question is whether or not the twins shared a chorion in the womb. The chorion is part of the placenta and constitutes an important ingredient of the pre-natal environment. A shared chorion implies a shared blood supply and a higher probability of infections passing between the twins.

Dizygotic twins are always dichorionic: each twin has his or her own chorion. However, about 60 percent of MZ twins are monochorionic, sharing a single chorion. This means that on average the pre-natal environment is more similar for MZ than for DZ twins. Is it possible that the greater similarity of MZ compared to DZ twins in terms of mental ability is due to this? The critical test comes from comparisons of dichorionic MZ and DZ twins. Such twins have roughly comparable pre-natal environments, so if the MZ twins still show a higher IQ correlation than the DZ twins, that would bolster the case for a genetic contribution to ability. On the other hand, if such twins are equally similar intellectually, that would suggest that the standard MZ-DZ difference is largely due to the fact that some of the MZ twins share a chorion, which makes them especially similar.

Unfortunately, chorionicity is rarely measured in twin studies, and researchers tend to lump together mono- and dichorionic MZ twins. But in the one study to date which has compared IQ similarity in twins as a function of their chorionicity, conducted in 1978 by Michael Melnick and his colleagues at the National Institutes of Health in Maryland, there was no difference in the correlation between white dichorionic MZ and DZ twins. For these twins, environmental factors should be highly comparable, and the only remaining difference is that the MZ twins are more genetically similar than the DZ ones. The results of this study suggest that genetic factors play only a minimal role in mental ability, and that the inheritance of intelligence is very much in doubt.

One problem with this study is that chorionicity is in fact very hard to determine post-natally. However, James Davis and his colleagues from the Southwest Missouri State University have recently developed a technique for determining chorionicity in adult twins which holds enormous promise for future research on the mental abilities of twins. Davis' technique relies on the fact that fingerprints are more similar in mono- than in dichorionic twins. Although fingerprints are under fairly strong genetic control, they are also influenced to a considerable extent by environmental factors such as nutrition, infection, and the mother's health. When twins share a chorion in the womb, they are about equally affected by these environmental factors and thus their fingerprints look similar. Dichorionic twins, on the other hand, experience more distinct pre-natal environments and hence their fingerprints tend to differ more.

The use of this technique in future twin studies should allow a clearer picture to emerge of whether the similarity of identical twins in terms of memory, intelligence, and so on is due to their shared genetic endowment or instead to their shared pre- and post-natal environment. It will be fascinating to see whether such research confirms the emerging view that there is no such thing as "talent" or "giftedness." To become an outstanding performer, one does not need an innate endowment of the right sorts of genes; instead, one simply needs to engage in deliberate practice for around 10,000 hours.


Ericsson, K. A. and Lehmann, A. C. "Expert and exceptional performance: Evidence of maximal adaptation to task constraints" Annual Review of Psychology, 47 , 273-305 (1996).

Howe, M. J. A., Davidson, J. W., and Sloboda, J. A. "Innate talents: Reality or myth?" Behavioral and Brain Sciences, 21, 399-442 (1998).

Phelps, J. A., Davis, J. O., and Schartz, K. M. "Nature, nurture, and twin research strategies" Current Directions in Psychological Science, 6, 117-121 (1997).

Science Spectra is published by G+B Magazines Unlimited Ltd. in association with
Gordon and Breach Publishers in the Netherlands under license from
OPA (Overseas Publishers Association) N.V.
Copyright © 1996 -- 1999 OPA (Overseas Publishers Association) N.V. All rights reserved.
Online site maintained by
OBS. Send comments to webmaster@gbhap.com.