ѕоиск по сайту

ѕишите нам: info@ethology.ru

Follow etholog on Twitter

—истема Orphus

 росс-культурный метод
—тарые форумы
Ќе в тему

список статей

Maintenance of genetic variation in personality through control of mental mechanisms: a test of trust, extraversion, and agreeableness

Kai Hiraishi, Shinji Yamagata, Chizuru Shikishima, Juko Ando

Х 1. Introduction

Х 1.1. The enigma regarding heritable personality variation

Х 1.2. Hypothesis: personality as an internal environment

Х 1.3. Examination of our internal environmental hypothesis of personality using twin data

Х 2. Methods

Х 2.1. Participants

Х 2.2. Questionnaires

Х 2.3. Procedure

Х 3. Results

Х 3.1. Basic statistical analysis of general trust scores

Х 3.2. Correlation between general trust and personality factors

Х 3.3. Univariate genetic analysis of general trust

Х 3.4. Multivariate genetic analysis of general trust and personality factors

Х 4. Discussion

Х Acknowledgment

Х References

Х Copyright

1. Introduction 

1.1. The enigma regarding heritable personality variation 

From a Darwinian perspective, the fact that personality traits are heritable poses an enigma. Behavioral genetics studies have repeatedly shown that personalities are typically 30Ц50% heritable (Bouchard & Loehlin, 2001), despite the fact that natural selection produces genetically homogeneous populations with regard to adaptively important traits. In other words, if a trait is adaptively important, it should have zero or very small heritability. Therefore, the relatively high heritability of personality traits appears to suggest that individual differences in personality are, from an evolutionary perspective, meaningless, contradicting common sense.

Several hypotheses have been proposed to address this issue (Buss & Greiling, 1999): (a) genetic personality variance is a result of mutation, which is meaningless with regard to individual fitness (Tooby & Cosmides, 1990); (b) personalities are alternative strategies that are maintained by frequency-dependent selection (Dall et al., 2004, Mealey, 1995, Wilson et al., 1996); (c) personality is reactively heritable through physical traits (Tooby & Cosmides, 1990); and (d) personality is a target of sexual selection (Miller, 2000). In this article, we propose another explanation that could be considered a mixture of these hypotheses.

To begin, we considered the reactive heritability hypothesis of Tooby and Cosmides (1990). They argued that behaving aggressively may be adaptive for those who are superior, but not inferior, in physique and that the superiority of physique is relative. Therefore, all humans should have the mental mechanism to behave aggressively and control its activation in accordance with their environment. In this sense, the ability to behave aggressively is a universal human trait and has zero heritability. However, activation of the aggressiveness mechanism can also be affected by one's own physique, which is highly heritable (Carmichael & McGue, 1995). Therefore, those who have a genetically superior physique are more likely to activate the aggressiveness mechanism, thus appearing aggressive. Accordingly, aggressiveness may appear to be heritable because of the genetic influence on physique.

The above hypothesis has two implications; one is that the heritability associated with mental traits such as personality is spurious and there is no genetic variance in personality. This is an interesting idea; however, in this article, we concentrate on the other implication, that control of mental mechanisms reduces fitness differences among individuals in different conditions.

In the hypothetical example given above, it was inferred that those who have superior physique will employ the aggressiveness mechanism more often. However, it is also inferable that those who are inferior in physique will employ different kinds of mechanisms. For instance, it has been reported within several species that smaller males often employ less aggressive sneaking strategies in mating while larger males employ aggressive and competitive strategies (Alcock et al., 1977, Arak, 1988, Gross, 1985), thereby reducing the fitness differences between larger males and smaller males. For instance, in the salmon species Oncorhynchus kisutch, large hooknose and small jack males differ genetically and employ different mating strategies, yet their reproductive success is almost identical (Gross, 1985). As a result, genetic variation associated with physical traits will be maintained because of the reduced selection pressure.

1.2. Hypothesis: personality as an internal environment 

In the hypothetical example given above, we argued only that individuals may control the activation level of mental mechanisms in accordance with their physical traits. Logically, however, it is possible that individuals control the activation level of certain mental mechanisms in accordance with their psychological traits.

Individual differences in personality have been reported cross-culturally. McCrae and Costa (1997), for example, reported five basic personality dimensions, neuroticism, extraversion, openness, agreeableness, and conscientiousness, in American, German, Chinese, Korean, and Japanese subjects. In addition, a recent study revealed that not only the phenotypic structure but also the genetic and environmental structure of personality is universal in Canadians, Germans, and Japanese (Yamagata et al., 2006). These studies indicate that there are universal patterns of individual differences in personality in humans.

Given these universal patterns of individual differences in personality, it is conceivable that individuals control their mental mechanisms according to their personality traits just as they would do according to their physical traits. If true, the selection pressure on personality differences will be reduced and genetic variation in personality will be maintained.

Put differently, we propose that individuals deal not only with their outer environments, such as ecology and society, but also with their УinternalФ environments by using their mental mechanisms. It has been argued that humans are endowed with a particularly large number of domain-specific mental mechanisms (Cosmides & Tooby, 1992, Hirschfeld & Gelman, 1994, James, 1890, Pinker, 1997, Pinker, 2002). Via these different mental mechanisms, humans can respond to a wide variety of environmental challenges. For instance, it is reported that humans have a set of domain-specific reasoning mechanisms that enable them to cope with a variety of adaptive social problems such as dyadic social exchange (Cosmides & Tooby, 1992), collective action (Hiraishi & Hasegawa, 2000), and alliance formation (Brown & Moore, 2000). Such mental mechanisms allow humans to cope with situations ranging from relatively small and simple societies to huge complex social structures like global market economies and nation states. Considering the large number of mental mechanisms that enable us to cope with a variety of external environmental challenges, it is therefore also plausible that these mechanisms enable us to enjoy wider internal environmental variation as well. In other words, because humans have a large number of mental mechanisms, they should be able to adapt well regardless of the personality inherited from their parents.

This hypothesis states that the genetic variance in personality will be maintained because it is, from an evolutionary point of view, meaningless. This does not mean, however, that personality is meaningless for individuals, but rather suggests that it is important for individuals to cope with their Уpersonality environmentФ and control their mental mechanisms accordingly. It is also important for individuals to assess the personality of others around them because such knowledge is a useful cue for understanding and predicting their social environments. Hence, personality differences are important although they are meaningless from an evolutionary point of view.

1.3. Examination of our internal environmental hypothesis of personality using twin data 

Our hypothesis is based on several assumptions, the most central being that humans adaptively control their mental mechanisms to meet their personality environment. In the following section, we present evidence that our level of trust toward strangers is controlled by extraversion and agreeableness.

Trust is crucial when engaging in social relationships. If individuals cannot trust their partners, it is impossible for them to construct and maintain mutually beneficial social relationship. Researchers agree that it is important to discriminate trust of neighbors, where there is both a history and future prospect of interaction, from trust of strangers, where both are lacking (Knight, 2001, Macy & Sato, 2002, Stolle, 2001, Yamagishi, 1998, Yamagishi, 2001). Trusting and engaging in social relations with strangers make us susceptible to vicious outsiders, but if strangers are never trusted, we will not create any new social relationships and will incur an opportunity cost if a stranger is trustworthy.

Several researchers have attempted to explain the emergence and evolution of trust of strangers (Cook, 2001). Among them, Yamagishi (1998, 2001) labeled trust of strangers as general trust and discriminated it from trust of neighbors, which he referred to as assurance of security. Based on a game-theoretic analysis, he argued that in a social environment where opportunity costs are high, individuals should set their general trust level higher. A number of studies support this theory. For example, Macy and Sato (2002) conducted an agent-based computer simulation and demonstrated that social mobility, which affects opportunity costs, influences the level of general trust in a population. Yamagishi and Yamagishi (1994) also reported that in the United States, which has an open and flexible society, people are generally more trusting than people in Japan, which has a closed collectivistic society. Differences between people who show high general trust and low general trust have been observed with regard to their level of cognition and their behavior in social settings, such as the Prisoner's Dilemma Game (e.g., Hayashi et al., 1999, Hiraishi et al., 2004, Yamagishi et al., 1999).

Yamagishi's theory states that our general trust level should be adjusted according to our social environment. This suggests that every individual possesses a mental mechanism for general trust, which is activated by facultative (environmental) rather than genetic factors. However, it is possible that even if general trust itself is not heritable, it is reactively heritable through other psychological traits. This is because the fitness of individuals cannot be improved by simply adjusting their general trust level according to their social environment (i.e., opportunity cost); it also requires adjustment according to their internal environment such as personality.

The advantages of high general trust flow from newly created beneficial social relationships. If an individual has an introverted personality and likes to be alone, that person will have difficulty constructing new social relationships and will fail to benefit from higher general trust. Similarly, individuals who are not agreeable and make the people around them feel uncomfortable will fail to construct new relationships. Therefore, even in a social environment with an abundance of social opportunities, only those who are extraverted and agreeable should set their general trust level higher: In short, personality should affect general trust level.

Behavioral genetics has shown that individual differences in personality, including differences in extraversion and agreeableness, are somewhat heritable (Bouchard, 2004, Bouchard & Loehlin, 2001). If general trust is adjusted according to an individual's level of extraversion and agreeableness, we can expect the individual's general trust to appear heritable because it is affected by genetic influences on extraversion and agreeableness. We can therefore expect there to be an apparent genetic influence on individual differences in general trust level. However, this genetic influence is УreactiveФ and can be explained by genetic influences on extraversion and agreeableness. We therefore propose the following predictions:

1.General trust and extraversion and/or agreeableness are positively correlated.

2.General trust is heritable.

3.General trust is adjusted according to our degree of extraversion and agreeableness; thus, the genetic influence on general trust can be fully or partly explained by genetic factors contributing to extraversion and/or agreeableness.

In order to test the above predictions, we conducted a twin study in which we decomposed the covariance of twin traits into variances of three independent parameters: additive genetic factors (A), environmental factors shared by the twin siblings (C), and environmental factors not shared by the twin siblings (E) (Neale & Maes, 2002, Plomin et al., 2000). According to the second prediction, additive genetic factors (A) are expected to contribute to the variance in general trust. In addition, the theory of general trust states that our familial environment (C) will not affect general trust while our environment outside our family (E) will.

Multivariate genetic analysis of twin data can be used to test whether there are causal relationships between two or more phenotypic traits (Heaths et al., 1993). According to the third prediction, our phenotypic level of extraversion and agreeableness is expected to have an effect on our phenotypic level of general trust, but not the reverse. In addition, it is suggested that all or part of the genetic influence on general trust can be explained by that on extraversion and agreeableness.

2. Methods 


2.1. Participants 

Participants were recruited from the Keio Twin Registry, which was constructed by collecting twin data in the Tokyo metropolitan area from the basic resident register provided by local government offices. A booklet composed of multiple questionnaires was sent to 2073 registered twins, of whom 1501 responded by completing the general trust scale. Twin pairs in which both siblings completed the general trust scale included 150 identical male pairs, 341 identical female pairs, 47 fraternal male pairs, and 91 fraternal female pairs. A total of 1060 participants completed both the general trust scale and the personality scale (356 identical twin pairs, 104 same-sex fraternal twin pairs, and 58 opposite-sex fraternal twin pairs). Data from opposite-sex twins were not included in the genetic analysis. Participants ranged in age from 14 to 31 years (mean=20.37 years, S.D.=4.13 years). Zygosity was initially determined using a questionnaire consisting of three questions about the twins' physical resemblance (e.g., As a toddler, were you and your twin Уas alike as two peas in a podФ?), which has been demonstrated to have 93.2% accuracy (Ooki et al., 1990). For twin pairs from whom DNA information was available through blood or buccal smears (181 pairs), zygosity was diagnosed by examining gene polymorphisms (for details, see Shikishima, Ando, Toda, & Yoshimura, 2006).

2.2. Questionnaires 

The booklet sent to registered twins contained questionnaires on topics such as social cognition, eating behavior, upbringing environment, and psychological health. In addition, questionnaire items from the general trust scale (Yamagishi, 1998) and the Revised NEO Personality Inventory (NEO-PI-R; Costa & McCrae, 1992), Japanese version (Yoshimura et al., 1998), were also included. The general trust scale consists of five items scored on a seven-point Likert scale (responses range from completely agree to completely disagree). Examples of the items are УMost people are basically honestФ and УMost people are trustful of others.Ф The NEO-PI-R is composed of 240 items scored on a five-point Likert scale and identifies the following five personality factors: neuroticism, extraversion, openness to experience, agreeableness, and conscientiousness.

2.3. Procedure 

Questionnaire booklets were sent to each twin pair in the Keio Twin Project who responded to our invitation letter by agreeing to participate. With the questionnaires, we included instructions stressing that all questionnaire items should be answered individually without discussing or exchanging information with others, specifically twin siblings. Subsequently, participants returned the questionnaires by mail using a separate envelope for each participant.

3. Results 


3.1. Basic statistical analysis of general trust scores 

Data from individuals who completed the general trust scale were used in the analysis (n=1501). No significant difference in mean general trust scores was observed between identical twins and fraternal twins. In addition, no significant difference was observed for mean general trust scores according to sex or birth order.

3.2. Correlation between general trust and personality factors 

Data from individuals who completed both the general trust scale and the NEO-PI-R (n=1060) were analyzed. As predicted, extraversion and agreeableness were strongly correlated with general trust (Table 1). Since the original agreeableness dimension includes a УtrustФ subscale, this subscale was omitted from the analysis and the correlation coefficient was recalculated. Following the recalculation, agreeableness remained significantly correlated with general trust.

Table 1.

Correlation coefficients between general trust and personality factors

Personality factor General trust
Neuroticism −.25
Extraversion .36
Openness to experience .19
Agreeableness .46
A without trusta .30
Conscientiousness .19

All correlations were significant (p<.001).


Agreeableness factor without the trust subscale.

3.3. Univariate genetic analysis of general trust 

We compared four models (ACE, AE, CE, and E) using structural equation modeling. Data from twin pairs in which both siblings completed the general trust scale were used in the analysis (491 identical twin pairs, 138 fraternal twin pairs). Based on the Akaike Information Criteria (AIC; Akaike, 1987), the best-fit model was the AE model (Table 2). Heritability (h2) was estimated to be 31%, while the remaining 69% of the variance was explained by nonshared environmental factors (and error).

Table 2.

Univariate genetic analysis of general trust

Model χ2 (df) p AIC h2 c2 e2
ACE 2.13 (3) .55 −3.869 0.31 0.00 0.69
AE 2.13 (4) .71 −5.869 0.31 Ц 0.69
CE 4.83 (4) .31 −3.175 Ц 0.28 0.72
E 55.56 (5) .00 45.561 Ц Ц 1.00

A smaller AIC indicates a better-fitting model. h2 indicates an additive genetic influence. c2 indicates a shared environmental influence. e2 indicates a nonshared environmental influence and error.

3.4. Multivariate genetic analysis of general trust and personality factors 

The trust subscale of the agreeableness factor was omitted, and multivariate genetic analysis was conducted with regard to general trust, extraversion, and agreeableness. Data from twin pairs in which both siblings completed the general trust scale and the NEO-PI-R (356 identical twin pairs, 104 same-sex fraternal twin pairs) were included in the analysis.

In the analysis, we compared the Cholesky decomposition model (Neale & Maes, 2002) and the direction of causation (DOC) model (Heaths et al., 1993). First, we constructed a full Cholesky model in which the number of genetic (A) and nonshared environmental (E) factors was equal to the number of traits (n=3 in this analysis). We then removed the paths with nonsignificant loadings. Subsequently, the best-fit model with the lowest AIC value was adopted (Fig. 1A).

View full-size image.@@START_COMMENT@@END_COMMENT

Fig. 1. Multivariate genetic analysis. A1 to A3: additive genetic factors. E1 to E3: nonshared environmental factors and error. The best-fitting model was Model B with unidirectional causal paths from extraversion and agreeableness to general trust.

Next, we constructed three DOC models (Fig. 1BЦD). By using latent etiological factors (i.e., genetic and environmental factors) as instrumental variables, DOC modeling of twin data can differentiate DOC between two or more variables, which is usually impossible with single-wave data. The first model assumed that extraversion and agreeableness have effects on general trust but not the other way around (Fig. 1B). The second DOC model assumed that general trust influences extraversion and agreeableness but not the other way around (Fig. 1C). The third DOC model assumed reciprocal causal relationships between extraversion and general trust and between agreeableness and trust (Fig. 1D). Since measurement errors have a strong effect on model fitting in the DOC model analysis (Heaths et al., 1993), we controlled such errors using alpha coefficients of all three traits in each model, including the Cholesky decomposition model.

The best model with the smallest AIC was the DOC model with unidirectional causal paths from extraversion and agreeableness to general trust (Table 3). Through these causal paths, genetic influences on extraversion and agreeableness had effects on general trust. However, there remained an independent genetic factor specifically affecting general trust. If we omit the independent genetic factor from the model, the model fit was significantly reduced (from AIC=−60.81 to AIC=−40.27). Therefore, not all of the genetic influences on general trust can be explained by the genetic influences on extraversion and agreeableness.

Table 3.

Multivariate genetic analysis: model comparison

Model χ2 (df) p AIC
Cholesky decomposition (A) 6.28 (32) 1.0 −57.72
DOC: personality to trust (B) 7.19 (34) 1.0 −60.81
DOC: trust to personality (C) 37.56 (34) .03 −30.63
DOC: reciprocal causation (D) 24.17 (34) .83 −41.84

A lower AIC score indicates better model fitting. A to D correspond to the models in Fig. 1.

On the best-fitting model (Fig. 1B), the causal influence of extraversion explained 18% of the variance in general trust (square of the path coefficient from extraversion to general trust) while agreeableness explained 14%. Heritabilities of extraversion and agreeableness were estimated to be 51% and 45%, respectively (square of the path coefficient from A1 to extraversion and from A2 to agreeableness, respectively). Therefore, general trust was estimated as being 9% and 6% heritable through extraversion and agreeableness, respectively. We also observed a relatively large nonshared environmental factor with an exclusive influence on general trust, which explained 46% of the variance. In addition, an independent genetic influence on general trust (A3 of Fig. 1B) explained 22% of the variance.

4. Discussion 


All three of our predictions were supported by the data. First, significant correlations were observed between general trust and two of the personality factors, extraversion and agreeableness. Among the five personality factors, extraversion and agreeableness had the first and second strongest correlations with general trust, respectively, even after omitting the trust subscale of agreeableness.

Second, the univariate genetic analysis revealed that the variance in general trust is best explained by additive genetic factors (A) and nonshared environmental factors and error (E), thus supporting our hypothesis. The contributions of A and E factors were calculated as 31% and 69%, respectively, indicating that nonshared environmental factors have a comparatively large influence. This observation supports Yamagishi's (1998) theory, which proposes that the level of general trust is adjusted according to nonshared environmental factors, not familial environmental factors.

Third, the multivariate genetic analysis of general trust, extraversion, and agreeableness showed that phenotypic levels of extraversion and agreeableness have effects on the phenotypic level of general trust. This causal relationship supports our hypothesis that individuals control their mental mechanisms in accordance with their personality disposition. General trust was estimated as being 9% and 6% reactively heritable through extraversion and agreeableness, respectively. We also observed a relatively large independent nonshared environmental factor on general trust (46%). Given Yamagishi's (1998) theory on general trust, this nonshared environmental factor might reflect the level of opportunity cost in the social environment.

We observed a significant level of independent genetic influence on general trust (22%); however, this does not necessarily contradict our hypothesis. There are several possible explanations. One is that it is the result of reactive heritability of other psychological and/or physical traits. Yamagishi (1998) argued that because people with higher general trust risk outing their trust in vicious people, they should use more cognitive resources to assess the trustworthiness of their counterparts. Given this argument, it is plausible that general trust is reactively heritable through cognitive abilities concerning social information, if the latter is heritable at all. It is also possible that the level of general trust is controlled in accordance with physical traits such as physical attractiveness. Solnik and Schweitzer (1999) reported that in the ultimatum game, attractive people were offered more, but at the same time, more was demanded of them. This kind of Уbeauty premiumФ might have an effect on individuals' perception of others' trustworthiness.

Another possibility is that the remaining heritability of general trust is not reactive but real. If true, the genetic variation might reflect the behavioral strategical differences that are a result of frequency-dependent selection (Dall et al., 2004, Mealey, 1995, Wilson et al., 1996). This might reflect genetic mutation, which is neutral to individual fitness (Tooby & Cosmides, 1990). These points should be examined in future studies.

In this article, we proposed a hypothesis about the origin of individual personality differences and empirically tested one of the key assumptions, that is, that individuals adaptively control mental mechanisms with regard to their personality. We believe that reexamination of existing behavioral genetics studies would be a promising way to further test the hypothesis. In other words, the framework presented in this article could be used to bridge the two fields of evolutionary psychology and behavioral genetics, which, until now, have experienced only limited interaction despite shared interests with regard to understanding the human mind from a biological point of view (Segal & MacDonald, 1998).



This study was supported by a Keio University special grant-in-aid for innovative collaborative research projects, and the Human Frontier Science Program. It was also supported by a grant-in-aid for scientific research (A) No. 13309014 (project leader: Juko Ando) by the Japanese Ministry of Education. We thank all the participants and the staff of the Keio Twin Project.



Akaike, 1987. 1.Akaike H. Factor analysis and AIC. Psychometrika. 1987;52:317Ц332. CrossRef

Alcock et al., 1977. 2.Alcock J, Jones CE, Buchmann SL. Male mating strategies in the bee Centris pallida Fox (Anthophoridae: Hymenoptera). American Naturalists. 1977;111:145Ц155.

Arak, 1988. 3.Arak A. Callers and satellites in the natterjack toad: Evolutionarily stable decision rules. Animal Behaviour. 1988;36:416Ц432. CrossRef

Bouchard, 2004. 4.Bouchard TJ. Genetic influence on human psychological traits: A survey. Current Directions in Psychological Science. 2004;13:148Ц151.

Bouchard & Loehlin, 2001. 5.Bouchard TJ, Loehlin JC. Genes, evolution, and personality. Behavior Genetics. 2001;31(3):243Ц273. MEDLINE | CrossRef

Brown & Moore, 2000. 6.Brown WM, Moore C. Is prospective altruist-detection an evolved solution to the adaptive problem of subtle cheating in cooperative ventures? Supportive evidence using the Wason selection task. Evolution and Human Behavior. 2000;21:25Ц37.

Buss & Greiling, 1999. 7.Buss DM, Greiling H. Adaptive individual differences. Journal of Personality. 1999;67:209Ц243.

Carmichael & McGue, 1995. 8.Carmichael CM, McGue M. A cross-sectional examination of height, weight, and body mass index in adult twins. Journal of Gerontology Biological Sciences. 1995;50:B237ЦB244.

Cook, 2001. 9.Cook KS. Trust in society. New York: Russell Sage Foundations; 2001;.

Cosmides & Tooby, 1992. 10.Cosmides L, Tooby J. Cognitive adaptations for social exchange. In:  Barkow J,  Cosmides L,  Tooby J editor. The adapted mind. Oxford University Press; 1992;p. 163Ц228.

Costa & McCrae, 1992. 11.Costa PT, McCrae RR. Revised NEO personality inventory and NEO five-factor inventory: Professional manual. Florida: Psychological Assessment Resources; 1992;.

Dall et al., 2004. 12.Dall SRX, Houston AI, McNamara JM. The behavioural ecology of personality: Consistent individual differences from an adaptive perspective. Ecology Letters. 2004;7:734Ц739. CrossRef

Gross, 1985. 13.Gross MR. Disruptive selection for alternative life histories in salmon. Nature. 1985;313:47Ц48. CrossRef

Hayashi et al., 1999. 14.Hayashi N, Ostrom E, Walker J, Yamagishi T. Reciprocity, trust, and the sense of control: A cross-societal study. Rationality and Society. 1999;11:27Ц46.

Heaths et al., 1993. 15.Heaths AC, Kessler RC, Neale MC, Hewitt JK, Eaves LJ, Kendler KS. Testing hypothesis about direction of causation using cross-sectional family data. Behavior Genetics. 1993;23:29Ц50. MEDLINE | CrossRef

Hiraishi et al., 2004. 16.Hiraishi K, Ando J, Ono Y, Hasegawa T. General trust and individual differences on the Wason selection task with sharing-rule. Psychologia. 2004;47:226Ц237.

Hiraishi & Hasegawa, 2000. 17.Hiraishi K, Hasegawa T. Sharing-rule and detection of free-riders in cooperative groups: Evolutionarily important deontic reasoning in the Wason selection task. Thinking and Reasoning. 2000;7:255Ц294.

Hirschfeld & Gelman, 1994. 18.Hirschfeld LA, Gelman SA. Mapping the mind: Domain specificity in cognition and culture. Cambridge: Cambridge University Press; 1994;.

James, 1890. 19.James W. Principles of psychology. New York: Henry Holt; 1890;.

Knight, 2001. 20.Knight J. Social norms and the rule of law: Fostering trust in a socially diverse society. In:  Cook K editors. Trust in society. New York: Russell Sage; 2001;p. 354Ц373.

Macy & Sato, 2002. 21.Macy MW, Sato Y. Trust, cooperation, and market formation in the U.S. and Japan. Proceedings of the National Academy of Sciences. 2002;99(Suppl 3):7214Ц7220.

McCrae and Costa, 1997. 22.McCrae RR, Costa PT. Personality trait structure as a human universal. Am Psychol. 1997;52:509Ц516. CrossRef

Mealey, 1995. 23.Mealey L. The sociobiology of sociopathy: An integrated evolutionary model. Behavioral and Brain Sciences. 1995;18:523Ц599.

Miller, 2000. 24.Miller GF. Mental traits as fitness indicators: Expanding evolutionary psychology's adaptationism. Annals of New York Academy of Sciences. 2000;907:62Ц74.

Neale & Maes, 2002. 25.Neale MC, Maes HHM. Methodology for genetic studies of twins and families. Dordrecht: Kluwer Academic Publishers; 2002;.

Ooki et al., 1990. 26.Ooki S, Yamada K, Asaka A, Hayakawa K. Zygosity diagnosis of twins by questionnaire. Acta Geneticae Medicae et Gemellologiae. 1990;39:109Ц115. MEDLINE

Pinker, 1997. 27.Pinker S. How the mind works. New York: WW Norton; 1997;.

Pinker, 2002. 28.Pinker S. The blank slate. Viking Press; 2002;.

Plomin et al., 2000. 29.Plomin R, DeFries JC, McClearn GE, McGuffin P. Behavioral genetics. 4th ed. New York: W H Freeman; 2000;.

Segal & MacDonald, 1998. 30.Segal N, MacDonald KW. Behavioral genetics and evolutionary psychology: Unified perspective on personality research. Human Biology. 1998;70:159Ц184.

Shikishima et al., 2006. 31.Shikishima C, Ando J, Ono Y, Toda T, Yoshimura K. Registry of adolescent and young adult twins in the Tokyo area. Twin Research and Human Genetics. 2006;9:811Ц816. MEDLINE | CrossRef

Solnick & Schweitzer, 1999. 32.Solnick SJ, Schweitzer ME. The influence of physical attractiveness and gender on ultimatum game decisions. Organizational Behavior and Human Decision Processes. 1999;79:199Ц215. CrossRef

Stolle, 2001. 33.Stolle D. Clubs and congregations: The benefits of joining an association. In:  Cook K editors. Trust in society. New York: Russell Sage; 2001;p. 204Ц244.

Tooby & Cosmides, 1990. 34.Tooby J, Cosmides L. On the universality of human nature and the uniqueness of the individual: The role of genetics and adaptation. Journal of Personality. 1990;58:17Ц67. MEDLINE | CrossRef

Wilson et al., 1996. 35.Wilson DS, Near D, Miller RR. Machiavellianism: A synthesis of the evolutionary and psychological literatures. Psychological Bulletin. 1996;119:285Ц299. MEDLINE | CrossRef

Yamagata et al., 2006. 36.Yamagata S, Suzuki A, Ando J, Ono Y, Kijima N, Yoshimura K, et al. Is the genetic structure of human personality universal? A cross-cultural twin study from North America, Europe, and Asia. Journal of Personality and Social Psychology. 2006;90:987Ц998. MEDLINE | CrossRef

Yamagishi, 1998. 37.Yamagishi T. The structure of trust. University of Tokyo Press; 1998;[in Japanese; English version is available at the author's website: http://lynx.let.hokudai.ac.jp/members/yamagishi/english.htm].

Yamagishi, 2001. 38.Yamagishi T. Trust as a form of social intelligence. In:  Cook K editors. Trust in society. New York: Russell Sage; 2001;p. 121Ц147.

Yamagishi et al., 1999. 39.Yamagishi T, Kikuchi M, Kosugi M. Trust, gullibility and social intelligence. Asian Journal of Social Psychology. 1999;2:145Ц161.

Yamagishi & Yamagishi, 1994. 40.Yamagishi T, Yamagishi M. Trust and commitment in the United States and Japan. Motivation and Emotion. 1994;18:129Ц166.

Yoshimura et al., 1998. 41.Yoshimura K, Nakamura K, Ono Y, Sakurai A, Saito N, Mitani M, et al. Reliability and validity of the Japanese version of the NEO five-factor inventory (NEO-FFI): A population-based survey in Aomori prefecture. The Japanese Journal of Stress Sciences. 1998;13:45Ц53.

a Graduate School of Arts and Sciences, The University of Tokyo, 153-8902 Tokyo, Japan

b The Graduate School of Human Relations, Keio University, 108-8345 Tokyo, Japan

c Faculty of Letters, Keio University, 108-8345 Tokyo, Japan

@@START_COMMENT@@END_COMMENTCorresponding author. Graduate School of Arts and Sciences, University of Tokyo, 3-8-1 Komaba, Meguro, Tokyo 153-8902, Japan. Tel.: +81 3 5465 8837; fax: +81 3 5465 8837.

 This study was supported by a Grant-in-Aid for Scientific Research (A) from the Japanese Ministry of Education, Culture, Sports, Science and Technology (No. 13309014).

PII: S1090-5138(07)00072-4