mind, self, and personality: dynamic interactions from ... self... · this paper aims to integrate...

P1: FLT

Journal of Adult Development pp810-jade-462980 April 10, 2003 1:7 Style file version May 31, 2002

Journal of Adult Development, Vol. 10, No. 3, July 2003 ( C© 2003)

Mind, Self, and Personality: Dynamic InteractionsFrom Late Childhood to Early Adulthood

Andreas Demetriou1

This paper aims to integrate research on mind, personality, and self-development using ageneral model which hypothesizes that mind and personality are organized in 3 levels. Thefirst level includes environment-oriented, domain-specific systems that specialize in the rep-resentation of and interaction with particular types of relationships in the environment inboth the cognitive and the social realms. The second level comprises self-oriented monitoringand representation processes that build maps of the environment-oriented systems. Self andidentity derive from the interaction between these 2 levels, which is subject to the constraintsand the system’s processing capacity at the successive ages. These constraints may be takenas the third level of the self. Efficiency in overcoming these constraints is determined by theself-oriented processes, generating feelings and self-representations of self-worth. These feel-ings are idiosyncratic, and function as a personal constant, which is applied on self-evaluationand self-representations. This constant is rescaled and differentiated with development. Thus,we have simultaneously both modular and transmodular constructs in the mind and the self,which change with relative independence of each other under the constraints of the generalprocessing capabilities and the personal constant. A series of studies are presented to supportthis model, and its implications for life-span theories of development are discussed.

KEY WORDS: cognitive development; hypercognition; personality; self-awareness; thinking styles.

Intelligence, mind, personality, and self are com-plementary aspects of the same entity: the individ-ual in interaction with the physical, social, and sym-bolic worlds. Intelligence refers to the abilities relatedto knowledge acquisition, understanding, and learn-ing that allow the individual to cope with the chang-ing demands of the world (Demetriou & Kazi, 2001;Jensen, 1998; Sternberg, 1985). Mind refers to ex-actly the same processes, although the emphasis shiftsfrom the individual’s relationships with the environ-ment to those referring to herself or others as think-ing agents; thus mind refers to self-awareness aboutthinking, understanding, and learning (Perner, 1991;Wellman, 1990). Personality refers to dispositions torelate and interact with the world in particular ways,and so provides the framework within which intel-

1To whom correspondence should be addressed at Department ofEducational Sciences, University of Cyprus, P. O. Box 537, 1678Nicosia, Cyprus; e-mail: [email protected].

ligence and mind operate (Costa & McCrae, 1997;Ferrari & Sternberg, 1998). The self refers to all thesedimensions together as they are experienced, sensed,understood, and defined by each individual to pro-duce the individual’s personal identity (Demetriou& Kazi, 2001; Demetriou, Kazi, & Georgiou, 1999;Harter, 1999). Despite their complementarity, theseaspects of the person have been examined separatelyin psychology, resulting in conceptually and epistemo-logically very different theories. As a result, the com-mon ground and the dynamic relationships betweenthe dimensions have not been clearly defined.

This paper summarizes a series of studiesdesigned to contribute to the formation of anoverarching model that would restore in theory theunity of intelligence, mind, self, and personality thatexists in reality. Specifically, the paper attempts to an-swer questions such as the following: What preciselyis involved in self-awareness and self-representation?That is, how are intellectual processes and personality

151 1068-0667/03/0700-0151/0 C© 2003 Plenum Publishing Corporation

P1: FLT


152 Demetriou

dispositions projected onto a person’s self-construct?How do the products of self-awareness interrelate toproduce an integrated self-system, which includes thedomains of cognition and personality? That is, how isit possible that each person has a strong sense of unityand identity while all persons are very differentiatedand diversified in their abilities, characteristics, andtendencies? What changes with development in thesedimensions and their self-representation? How canthe sense of unity be so strongly preserved throughoutdevelopment despite the fact that abilities, processes,and characteristics change extensively as a result ofdevelopment?

The various dominant traditions of psychologyhave focused on a number of these questions, andwhile valuable work has been produced no modelintegrating the findings presently exists. It is imper-ative to move in the direction of constructing such amodel (Demetriou & Kazi, 2001; Ferrari & Sternberg,1998; Jensen, 1998). Thus, in this paper, we will firstvery briefly summarize the current theorizing and re-search, in order to pinpoint the similarities and differ-ences between the various traditions. This exercise isuseful for the construction of an overarching modelthat can be tested empirically.

ARCHITECTURES OF THE MIND,PERSONALITY, AND SELF

Psychometric (Carroll, 1993; Jensen, 1998;Thurstone, 1938) and other (Gardner, 1983;Sternberg, 1985) theories of intelligence specifythe dimensions and abilities underlying humanintelligence and their interrelations. Theories ofintellectual development, such as the classic theoryof Piaget (1970) or the subsequent neo-Piagetiantheories (Case, 1992; Fischer, 1980; Halford, 1993;Pascual-Leone, 1988), describe the nature andinterpatterning of cognitive processes underlyingintelligence from birth to maturity. Although thesetheories have influenced our model of the developingmind, the presentation below will focus only on ourmodel because of space limitations. The reader whois interested in a discussion of the relationships ofour theory with other theories is referred to othersources (Demetriou, 1998a, 2000; Demetriou &Efklides, 1985; Demetriou, Efklides, & Platsidou,1993; Demetriou, Kazi, et al., 1999; Demetriou& Kazi, 2001; Demetriou & Raftopoulos, 1999;Demetriou, Raftopoulos, & Kargopoulos, 1999).

Our model points that the mind includes twolevels of knowing (the term “level” here refers to

architectural layers in the construction of the mindand not to successive stages or tiers in the construc-tion of concepts during development), one orientedto the environment and another oriented to the self.The first level involves representational and under-standing processes and functions that specialize in therepresentation and processing of information fromthe environment. The second level involves func-tions and processes oriented to monitoring, represent-ing, understanding, and regulating the environment-oriented functions and processes. Optimum orintelligent performance at any time depends on theinteraction between the two levels, because efficientproblem-solving or decision-making requires the ap-plication of environment-oriented functions and pro-cesses under the guidance of representations heldabout them at the level of self-oriented processes.

The interaction of the two levels occurs underthe constraints that stem from the biological conditionof the individual at any given point in his life. Theseconstraints may be viewed as an additional content-free level in the architecture of the mind that involvesprocesses operating as the interface between the twoknowing levels mentioned above. In information pro-cessing terms, this levels refers to the processing ca-pacity available at a given age.

Our research has identified and delineatedseven environment-oriented systems (EOS). (1) Thecategorical system deals with similarity–differencerelationships. Forming concepts about class relation-ships is an example of the domain of this system.(2) The quantitative system deals with quantitativevariations and relationships in the environment;mathematical concepts and operations are examplesof this system’s domain. (3) The causal system dealswith cause–effect relationships; operations such astrial-and-error or isolation of variables strategies thatenable a person to decipher the causal relationshipsbetween things or persons and ensuing causal con-cepts and attributions belong to this system. (4) Thespatial system deals with orientation in space and theimaginal representation of the environment; mentalimages and operations on them belong to this system.(5) The propositional system deals with truth/falsityand the validity/invalidity of environmental informa-tion and in systems of representation. Different typesof logical relationships such as implication, com-patibility, and incompatibility belong to this system.(6) The social system deals with the understandingof social relationships and interactions. Mechanismsfor monitoring nonverbal communication and skillsfor manipulating social interactions belong to this

P1: FLT


Mind, Self, and Personality 153

system. (7) The pictographic system underlies therepresentation of the environment or of thoughtsthemselves by means of drawings or other kind ofsigns. Manual skills or skills enabling the translationof a mental image into a picture on paper belong tothis system (Case, Demetriou, Platsidou, & Kazi, inpress; Demetriou et al., 1993; Demetriou & Efklides,1985, 1989; Kargopoulos & Demetriou, 1998; Shayer,Demetriou, & Pervez, 1988).

These systems are computationally, procedurally,and symbolically specific and they follow partially au-tonomous developmental trajectories. Each systemincludes a characteristic set of operations and pro-cesses which are appropriate for the processing of therelationships specific to each domain and is biasedto different symbol systems (Demetriou & Efklides,1988; Demetriou & Raftopoulos, 1999; Kargopoulos& Demetriou, 1998).

The self-oriented level involves processes andfunctions that guide the monitoring, understanding,and regulation of the processes and functions resid-ing at the other two cognitive levels. Thus, the input tothis level is information arising from the functioningof the environment-oriented systems under the cur-rent processing constraints (sensations, feelings, andconceptions caused by mental activity). This input isorganized into the maps or models of mental functionsand the self to be described below. We used the termhypercognitive to refer to this level and denote thehierarchical relationship between this and the othertwo levels of the mind.

The term hypercognitive is used rather than theterm metacognitive because it is a broader and moreaccurate description of the level of mind that isresponsible for self-monitoring, self-representation,and self-control. In Greek, the adverb “hyper” means“higher than” or “on top of” or “going beyond”and when added to the word “cognitive” indicatesthe supervising and coordinating functions of the hy-percognitive system (Demetriou, 1998a, 1998b, 2000;Demetriou et al., 1993). The adverb “meta” means“coming after”; thus, our proposed term is more ac-curate than the term “metacognitive” because it refersto more of the functions associated with this level ofmind. That is, it can denote an autonomous level ofthe architecture of the mind which exists from thebeginning and which operates on the other levelsrather than as a by-product of their functioning. Infact, metacognition is only one of the functions of thehypercognitive system.

We have shown that the hypercognitive levelinvolves active self-knowing processes and self-

descriptions and concepts. We used the terms work-ing and long-term hypercognition to refer to these twotypes of processes, respectively. Working hypercogni-tion includes on-line self-monitoring, self-recording,self-evaluation, and self-regulation processes andskills that enable the individual to efficiently and ac-curately activate her various cognitive, emotional, orother systems, according to the requirements of themoment. Thus, working hypercognition controls theuse of current processing capacity. The functioning ofworking hypercognition produces knowledge aboutthe mind and the self that is organized into three in-terrelated systems of concepts or models. These threesystems constitute long-term hypercognition. The so-called theory of mind (Flavell, Green, & Flavell, 1995;Perner, 1991; Wellman, 1990), self-regulation rules(Sternberg, 1985), and one’s cognitive self-conceptbelong to long-term hypercognition (Demetriou et al.,1993; Demetriou & Efklides, 1989; Demetriou & Kazi,2001).

Processing capacity is defined in terms of threeparameters: speed of processing, control of process-ing, and storage. Speed of processing refers to themaximum speed at which a given mental act may beefficiently executed. Control of processing refers tothe processes that identify and register goal-relevantinformation and block out dominant or appealingbut actually irrelevant information. Storage refersto the processes which enable a person to holdinformation in an active state while integrating itwith other information until the current problem issolved. Thus, this level is close to the psychomet-ric conception of general intelligence and to otherneo-Piagetian conceptions of processing capacity ormental power (Demetriou et al., 1993; Demetriou,Spanoudis, Christou, & Platsidou, in press).

The Organization of Personality

Personality appears to be organized in a way thatis formally equivalent to the hierarchical organizationof the mind presented above. It can also be describedas a hierarchy of three levels: one including differentmodes of interacting with environment, one referringto constraints operating on these modes, and one in-volving self-representations about these modes.

Specifically, the so-called Big Five factors of per-sonality (that is, extraversion, agreeableness, con-scientiousness, neuroticism, and openness to experi-ence) may be taken as the different modes of relatingwith the environment. Extroverts are sociable, active,

P1: FLT


154 Demetriou

and uninhibited as contrasted to introverts who arewithdrawn, shy, and inhibited. Agreeable individu-als are soft-hearted, generous, forgiving, and warm ascontrasted to individuals low in agreeableness whoare suspicious, headstrong, argumentative, and ag-gressive. Conscientious individuals are organized, en-ergetic, and reliable as contrasted to individuals lowin conscientiousness who are distractible, lazy, care-less, and immature. Neurotic individuals are nervous,anxious, tense, and self-centered as contrasted to in-dividuals low in neuroticism who are emotionally sta-ble, confident, alert, and content. Finally, individu-als who are open to experience are curious and withwide interests, original, imaginative, and nontradi-tional whereas individuals who are not open to experi-ence are conservative and cautious (Costa & McCrae,1997; Kohnstamm & Mervielde, 1998).

Graziano and colleagues (Graziano, Jensen-Campbell, & Finch, 1997; Graziano, Jensen-Campbell, & Sullivan, in press), hypothesize thatthese five dimensions underlie normal everydayfunctioning, but are expressed through a numberof higher-order constructs. These constructs trans-late each of the Big Five into self-definitions andself-representations, which lead to particular waysof acting. Moreover, there are also evaluative con-structs, such as global or domain-specific self-esteem(e.g., social and academic self-esteem). These higher-order constructs in the structure of personality seemequivalent to the self-representations that reside atthe hypercognitive level of the mind.

The constructs included in the two levels of per-sonality analyzed above operate under the constraintsof temperament which refers to more or less stabledispositions to receive and react to stimulation in par-ticular ways (Kagan, 1994). Examples of tempera-mental dispositions include ease as well as intensityof arousal in reaction to stimulation, regulation ofarousal states, approach to or avoidance of novelty, ir-ritability, mood, motor activity, and adaptability. To aconsiderable degree these dispositions are consideredas inherited and function as the substrate upon whichpersonality and the self are formed. Thus, from child-hood these early tendencies and dispositions gradu-ally develop into the more stable dimensions of per-sonality and style that characterize the individual.

The Organization of the Self

Interestingly, the analysis of the self is formallyvery similar to the analysis of the mind and person-ality proposed above. According to the classic theory

of James (1892), the self involves two hierarchical lev-els, the “I-self” and the “Me-self.” The I-self is theknower, and includes all self-observation and self-recording processes that generate the knowledge thatpersons have about their bodily, social, and mentalaspects.This knowledge is the Me-self.

James’s distinction between a knowing (theI-self) and a known self (the Me-self) is present inmodern theories of the self. For example, in Markus’smodel (Markus & Wurf, 1987) the individual’s work-ing self-concept is differentiated from his collectionof self-representations. The working self-concept in-volves any presently accessible self-representationsand it is directly involved in the formation and con-trol of behavior at both the intra- and the interper-sonal level. Therefore, in this model, the working self-concept assumes the functions of the Jamesian I-self(our working hypercognition), which generates theself-descriptions that belong to the Jamesian Me-self(our cognitive self-image).

Empirical research suggests that the self-construct is indeed hierarchically organized. This re-search suggests that there is a general self-conceptat the apex, a number of major domains at a middlelevel (such as self-concepts of the academic, social,emotional, and physical domains) and certain morespecific domains within each of these major domains(such as maths and science in the academic domain orphysical ability and physical appearance in the physi-cal domain). Moreover, each domain includes a de-scriptive part that specifies what an individual canand cannot do and an evaluative part, which specifiesthe importance of possibilities and impossibilities foreach of the domains. Thus, there is global and domain-specific self-worth or self-esteem (see Bracken, 1996;Harter, 1999; Hattie, 1992).

TOWARD AN OVERARCHING MODEL OFTHE DEVELOPING, THINKING PERSON

Draft for an Overall Architecture

We proposed that the three hierarchical architec-tures summarized above are interrelated (Demetriou,2000; Demetriou & Kazi, 2001). For instance, we as-sumed that temperament is to personality what pro-cessing potentials are to intelligence and thinking.That is, processing potentials constrain the complex-ity and type of information that can be understood ata given age, and temperament constrains how infor-mation will be initially received and reacted to. More-over, the Big Five factors of personality may be taken

P1: FLT



to correspond to the domain-specific systems that re-side at the environment-oriented levels of knowing.In other words, the domain-specific systems of un-derstanding channel the functioning of the mind andthe Big Five factors of personality channel patternsof action and relationships with the social and cul-tural environment. The functioning of these systemsis controlled by the active processes of self-knowingimplicated in James’ I-self or Markus’ working self-concept or by the monitoring and control processesinvolved in our working hypercognition. In turn, thedaily functioning of all these systems generates feel-ings and representations of general cognitive effi-ciency or global self-worth and general self-esteem,as well as more specialized self-representations aboutthe domain-specific systems of the mind or the par-ticular dimensions of personality. In turn, the interac-tions between these systems generate different stylesof thinking or action, which, in turn, describe differ-ent types of actual adaptations and behavior. Figure 1illustrates this overarching model.

It seems, therefore, that it is possible to con-struct an overarching model of the thinking andacting individual that simultaneously captures boththe dynamic (motivational and emotional) and themeaning-making (representational) components ofunderstanding, experience, and action. Obviously,testing this model in its totality is an enormous un-dertaking. However, several studies conducted in ourlaboratory, have provided encouraging evidence thatthe model justifies further theoretical and empiricalscrutiny. Some of these studies will be summarizedbelow.

Developmental Dynamics

All functions, processes, skills, dispositions, andcharacteristics included in the various levels and sys-tems of the mind as specified in the model aboveundergo extensive changes from birth to maturity.Different scholars and traditions have focused on dif-ferent components of this model and have attemptedto describe and explain their development. Thus weknow, for instance, that the various functions of theprocessing system, such as speed of processing, inhi-bition and control of processing, and short-term stor-age, improve systematically from birth until levellingoff at the end of the second or the beginning of thethird decade of life. After this age, processing slowsdown, inhibition and control are effected with moredifficulty and less efficiency, and short-term memorydeclines (see Baltes, 1991; Shaie & Willis, 1991).

The environment-oriented systems, includingcategorical, analogical, verbal, mathematical, andspatial thought, seem to follow a similar course. Thatis, they grow and expand until early adulthood andthey then also begin to decline, probably a few yearslater (i.e., in the fifth decade) than information pro-cessing, not always at the same time or rate. Forinstance, spatial reasoning appears to decline ear-lier than verbal reasoning (Demetriou, 1998a, 1998b;Shaie & Willis, 1991).

Models of personality development are less de-tailed than those of cognitive development, proba-bly because personality is considered stable from anearly age. Although there is strong evidence that thebasic Five Factor structure of personality stabilizesin early childhood, if not earlier, we now know thatsome important personality changes occur. For exam-ple, there is evidence that from adolescence to adult-hood individuals tend to become more conscientious,more emotionally stable, and more agreeable, but lessextroverted and less open to experience (Costa &McCrae, 1997).

It is generally agreed that there are devel-opmental changes in at least three dimensions ofthe self-system. First, with age self-representationsinvolve more dimensions that are better inte-grated into increasingly more complex structures.Thus, second, emphasis gradually moves fromconcrete self-descriptions to more abstract trait-and characteristic-like self-definitions. Third, self-representations become more accurate with regard tothe actual characteristics and abilities to which theyrefer (see Harter, 1999).

Finally, and probably as a result of these changes,self-regulation becomes increasingly more focused,refined, efficient, and strategic. That is, the individ-ual gradually acquires command of his or her in-formation processing capabilities and temperamen-tal dispositions. Practically, this implies that cognitivecapabilities and personality dispositions come un-der increasing a priori control of the person’s long-term hypercognitive maps and self-definitions. In fact,capitalizing on these maps by training individualsto use them efficiently enhances and stabilizes theeffects of learning experiences and development ofthe environment-oriented abilities (Demetriou et al.,1993; Kallio, 1998). In Freudian terms, mind and per-sonality come under ego control; in Jamesian terms,they come under the control of the I-self.

Various models that capture the development ofmind, personality, or self have been proposed. How-ever, there have been very few and only sketchy

P1: FLT


156 Demetriou

Fig. 1. The general model of the mind and personality. (Note. SCS stands for specialized structural system. P stand for personalityfactor.)

P1: FLT



attempts to study and model changes in the relation-ships among these three realms. Some of our studiesto be summarized below are directly addressed to thisquestion. These studies aim to highlight how mind,personality, and self interact from late childhood toearly adulthood, and thus contribute to the forma-tion of tendencies and abilities underlying the normaladult’s everyday functioning.

STUDY 1: DYNAMIC RELATIONSHIPSAMONG THE LEVELS OF MIND,PERSONALITY, AND SELF

The first study, fully presented in Demetriouand Kazi (2001), aimed to answer the followinghypotheses:

1. How are self-representations of cognitive abil-ities organized? The theory is based on theassumption that the hypercognitive systeminvolves maps of mental abilities that repre-sent the actual organization of abilities; thissuggests that self-representations of cognitiveabilities must be organized in the same struc-tures as the corresponding actual abilities.

2. How are factors of personality and think-ing style related to cognitive factors? We hy-pothezise that these factors will depend onthe actual cognitive performance and self-representation factors, because they reflect ad-justments that individuals make to their owncognitive profile so as to maximize efficiencyand satisfaction in their personal interactionwith the world.

3. During development what changes occur inthe relationships between the two main lev-els of the mind, namely the environment-oriented and the self-oriented level? Our as-sumption is that large developmental changes,such as those associated with major develop-mental transitions, necessitate reorganizationsin the persons’ self-representations about theirown abilities aiming to restore, at the level ofself-representation, the state of actual cogni-tive abilities residing at environment-orientedlevel of the mind.

Participants

This study involved 840 individuals, aged 10.6–15.6 years. Socioeconomic status and gender werefairly equally represented in each age group.

Tasks and Inventories

Participants were examined with the followingfive types of tasks or inventories.

Cognitive Tasks

A set of tasks addressed seven domains ofthought, six of which correspond to the environment-oriented systems specified by our theory. Specifi-cally, a set of mathematical analogies and algebraicreasoning tasks addressed the quantitative system(Cronbach’s alpha: .62). A set of hypothesis-testingand experiment-design tasks examined the causal sys-tem (Cronbach’s alpha: .61). A set of tasks focused onunderstanding social intentions and dialectical think-ing (that is, grasp of the complementarity of differentperspectives in regard to issues of social concerns)represented the social thought system (Cronbach’s al-pha: .58). A set of mental rotation and image manip-ulation tasks tested the spatial system (Cronbach’salpha: .50). A set of drawing tasks, which requireddrawing both persons and landscapes, addressed thepictographic system (Cronbach’s alpha: .70). Finally,to assess creativity, we used a set of tasks requir-ing the production of new ideas and symbols aboutthe situation described (Cronbach’s alpha: .70). Ascan be seen, the reliability of these tasks, which arefully described in Demetriou and Kazi (2001), wassatisfactory.

Self-Evaluation

Here, participants were asked to evaluate cer-tain aspects of a number of these tasks. Specifically,they were asked to judge the difficulty and successof solution to an easy and a difficult task among thetasks addressed to each environment-oriented sys-tem. These evaluations allowed the organisation andinterrelationships of cognitive processes at the levelof performance to be compared with the organiza-tion and interrelationships of these processes at thelevel of self-awareness; thus the interrelationships be-tween the environment and the self-oriented level ofthe mind could be more effectively studied.

Self-Representation About Cognitive Abilities

To examine self-representations of the abilitiesexamined by the tasks above, we constructed an in-ventory asking participants how they represent them-selves in regard to various processes involved in

P1: FLT


158 Demetriou

each of the cognitive domains represented in thethese tasks. Statements about quantitative thought re-ferred to the subject’s facility in solving mathemat-ical problems or applying mathematical knowledgeto everyday problems (e.g., “I easily solve every-day problems involving numbers” Cronbach’s alpha:.83). Statements about causal thought referred to theability to formulate hypothesis and design experi-ments (Cronbach’s alpha: .59), or to construct mod-els that would accommodate experimental findings(e.g., “To find out which of my guesses is correct, Imethodically consider only the things my guess pro-poses”; Cronbach’s alpha: .63). Statements about spa-tial thought referred to visual memory (e.g., “I retaina very clear picture of things”) and facility in thinkingin images (“When I have to arrange things in a certainspace, I first visualize what it will be like if I place themin certain way and then I physically arrange them”;Cronbach’s alpha: .66). The statements on drawingreferred to ability to draw a man, a landscape, anda map (e.g., “I can draw a person very accurately”;“ I can paint a building as if it were a photograph”;Cronbach’s alpha: .84). Finally, the statements aboutsocial thought referred to the facility in understandingother’s thoughts and feelings (e.g., “I understand eas-ily the intentions of others before they express them”;Cronbach’s alpha: .77). As can be seen, this inven-tory was very reliable. Also, this inventory addressedgeneral cognitive characteristics, such as learning (“Ilearn easily,” “I retain a lot of what I hear”; Cronbach’salpha: .76) and reasoning ability (“I like drawing log-ical conclusions, which can be justified by the data Ihave”; Cronbach’s alpha: .70).

Personality Characteristics and Thinking Styles

To examine how personality and thinking stylescompared to various aspects of the mind addressedby the tasks and inventories above, an inventory as-sessing various social and personality characteristicswas designed. Regarding personality, this inventoryaddressed ambition and ideal self (“I am an ambi-tious person,” “I want to be the center of attention”;Cronbach’s alpha: .76), impulsivity (“I do the firstthing that comes to my mind; Cronbach’s alpha: .69),and systematicity (“I will deal with various thingsin a certain order, depending on their importance”;Cronbach’s alpha: .59).

Regarding styles of activity and thought, this in-ventory addressed the following dimensions: original-ity (“I like working on problems for which there are nopre-prepared solutions”; Cronbach’s alpha: .51), an

executive style, which requires following rules (“Whensolving a problem, I prefer to follow existing rules”)(Cronbach’s alpha: .56), or a judgmental or evalua-tive style, which requires evaluating other people (“Ilike to judge others’ choices”; Cronbach’s alpha: .59),and professionalism (“I like choosing the best of vari-ous solutions”; Cronbach’s alpha: .62). This inventorywas based on Sternberg’s theory of self-government(Sternberg, 1988).

Results and Discussion

The data generated by this study was first exam-ined to test the first of our hypotheses (the natureand organization of self-awareness) in regard to var-ious processes. Specifically, a series of confirmatoryfactor analysis and structural equation models wererun to test if the organization of self-representationsmatches or reflects the organization of performanceon the various cognitive tasks. Bentler’s EQS pro-gramme (Bentler, 1989) was used to run these models.Readers interested in studying all models tested arereferred to Demetriou and Kazi (2001). In this paperwe present two models that integrate findings on allbatteries and inventories described above ( see Figs. 2and 3).

Figure 2 presents a model that includes threesets of factors; that is actual cognitive abilities, self-representations of cognitive abilities, and personal-ity and thinking styles. Moreover, there are severalhigher-order factors, which represent higher level or-ganizations in each of these three kinds of factors.At the highest level, there are two general factors: afourth-order factor for general cognitive ability (theGcog factor, which, in psychometric terms representsg) and a fifth-order factor, the Ghyper factor, for theself-oriented level of the mind. Attention is drawnto the similarity in the organization of the cognitiveperformance factors and the organization of the cor-responding hypercognitive or subjective factors. Thus,this model confirms the first assumption put forward:the subjective architecture of the mind accurately re-flects its objective architecture.

This model was also tested separately on eachage group in order to test our third hypothesis andfound to have a good fit (the parameters of the modelfound for each age group are also shown in Fig. 2). Theconsistent good fit strongly suggests that the mentalarchitecture described above is stable from late child-hood to middle adolescence despite the fact that dur-ing this period the kind of problems that can be solvedchanges radically. Therefore, it can be concluded that

P1: FLT



Fig

.2.

The

gene

ralh

iera

rchi

calm

odel

appl

ied

onal

ltas

ksan

din

vent

orie

s.(N

ote.

Num

bers

inbo

ldin

each

colu

mn

refe

rto

the

mod

elap

plie

don

the

who

lesa

mpl

e.T

heot

her

five

num

bers

refe

rto

the

mod

elap

plie

don

each

ofth

efiv

egr

oups

.For

mto

pto

bott

om,t

henu

mbe

rsre

fer

toag

egr

oups

11–1

5,re

spec

tive

ly.)

P1: FLT


Fig. 3. The final model of the dynamic relationships between modules across the different levels of the mental architecture.

160

P1: FLT



the major units and levels of the mind as described byour theory do not depend on development.

Our research revealed an especially interestingfact: the correlation between the general cognitivefactor representing the environment-oriented levelof the mental architecture and the general hyper-cognitive factor representing the self-oriented levelwas practically null at the age of 11 years, but in-creased systematically and became strong at the ageof 15 years. That is, the correlation between these twohigher order factors was .05, .11, .20, .45, and .60 atages 11, 12, 13, 14, and 15 years, respectively. Thisfinding suggest that although the constructs and lev-els of the mind are the same, their relationships maychange over the years. Therefore, many changes tradi-tionally associated with the transition from childhoodto adolescence seem to primarily affect the commu-nication between the two knowing levels of the men-tal architecture rather than the modules and actualprocesses at each level. At the hypercognitive level,therefore, self-monitoring and self-representation ofthe environment-oriented level become increasinglyaccurate and detailed (Demetriou & Kazi, 2001).

Interrelationships among cognitive, hypercogni-tive, personality, and thought styles constructs are de-lineated in the model shown in Fig. 3. As in our firstmodel (Fig. 2), the cognitive factors here are organ-ised in two levels, one including the performance fac-tors and the other including the self-representationsfactors that reflect these performance factors. More-over, there was a second-order general factor re-lated to all the domain-specific performance factors.At the level of self-representation, there were twofactors representing general cognitive abilities, logi-cal reasoning and learning ability. One of these fac-tors, logical reasoning, was directly and significantlyrelated to the second-order general cognitive fac-tor and both of them, logicality in particular, ap-peared to play a very important role in the forma-tion of domain-specific self-representations. That is,the domain-specific self-representation factors ap-pear to depend directly upon the self-representationof logical reasoning and, to a lesser extent, learning.Moreover, the self-representation factors were morerelated to the second-order cognitive performancefactor than to the corresponding domain-specific per-formance factor. An exception to this pattern is quan-titative thought and drawing, which seem to be moretransparent to awareness than the other domainsstudied here (negative relationships between cog-nitive performance and self-representation indicatethat, with cognitive development, self-attribution of

abilities becomes more conservative; Demetriou &Kazi, 2001). These relationships suggest that domain-specific self-representations derive directly from self-representations of more general or dynamic cognitivefunctions rather than from domain-specific factors. Itseems that domain-specific experiences of problemsolving, especially when they are not easily amenableto self-monitoring and objective evaluation, such ascausal or social understanding, first contribute to theformation of a sense of general cognitive efficiency.This is then translated back into domain-specific rep-resentations, if needed.

We note here that another study (Demetriou &Kazi, 2001), which cannot be presented because ofspace limitations, provided further support to the as-sumptions about the relationships between the vari-ous levels of the mind. This study showed that skillin processing and analogical reasoning tasks directlycorresponded to individual’s self-representations ofreasoning and learning abilities. This finding suggestsclearly that the hypercognitive system directly regis-ters and represents the condition of the processingsystem and general inferential processes. Therefore,one might say that the famous g construct of psy-chometric theories of intelligence (Jensen, 1998), andthe recently revitalised constructs of global self-worth(Harter, 1999) and self efficacy (Bandura, 1989) ofpersonality psychology may be two sides of the samecoin.

Regarding the organization of personality andthe self, it can be seen that ambition and systematic-ity were highly related to the two factors represent-ing general cognitive efficiency in self-representation,that is, learning and logical reasoning. The factorfor impulsivity was closely and negatively related tothe second-order general cognitive ability factor, butwas not related to either of the two cognitive self-image factors. Thus, it seems clear that both the self-representation of logical reasoning and impulsivityplay pivotal but rather independent roles in the orga-nization of the self-image at both the cognitive and thepersonality level. For instance, they both contributestrongly to the formation of the ideal self. However,systematicity is formed in relation to logicality but in-dependently of impulsivity.

The relationships between personality dimen-sions and thinking styles are equally interesting, andsuggest the following conclusions. First, the factorstanding for the ideal self exhibited the strongest ef-fect on three of the four thinking styles factors, sug-gesting that this dimension is as important in theformation of a person’s choices and style as logical

P1: FLT


162 Demetriou

reasoning is in the organization of the more cognitiveaspects of one’s self-image. Second, the pattern of re-lationships between personality and thinking stylesoffers insight into how personality characteristics af-fect thinking styles. Specifically, the factor represent-ing the executive style was more closely related tosystematicity than to ideal self or impulsivity. Orig-inality was positively and highly correlated to idealself, negatively to implulsivity and it was not related tosystematicity. The evaluative style was related highlyto ambition, much lower but positively to impulsiv-ity, and low to systematicity. It seems, therefore, thatnonimpulsive individuals with high ambitions are ori-ented to activities requiring originality; impulsive andambitious individuals tend toward evaluative activi-ties; impulsive and systematic individuals orient them-selves to the rule-abiding activities.

Development and Individual Differences

It is well known that individual attitudes to self-evaluation and self-representation differ; for exam-ple some people are systematically lenient in theevaluation of their own performance while othersare systematically strict. How can this obvious factbe reconciled with the findings indicating that self-representations actually reflect the organization ofcognitive abilities?

To investigate this question, we extended thestudy described above, further testing subjects whovaried systematically in their accuracy and atti-tude to self-evaluation. By combining cognitiveperformance and self-evaluation of this perfor-mance we formulated the following four groupsof subjects: (1) low performance–low success eval-uation, (2) low performance–high success evalua-tion, (3) high performance–low success evaluation,and (4) high performance–high success evaluation.Obviously, subjects in groups 1 (low-accurates) and 4(high-accurates) are accurate in their self-evaluation,because their evaluations are consistent with theirperformance; while those in groups 2 and 3 are notaccurate. In other words, group 2 subjects (law inac-curates) are lenient in their self-assessment whereasgroup 3 subjects (high inaccurates) are strict withthemselves. These subjects were tested on all tasksand batteries described above once a year for 2 moreyears, (3 consecutive years in total). The effects ofinterest here are concerned with the condition of self-evaluation in these four person categories along withtime. That is, are the differences between the fourgroups preserved across the three testing waves?

To answer this question, a 5 (the five agegroups) × 4 (the four person categories) × 3 (thethree testing waves) × 5 (the five SCSs, i.e., quantita-tive, causal, imaginal, social, and drawing) ANOVA,with repeated measures on the last two factors, wasapplied on postperformance success evaluation pro-vided at each of the three testing waves for eachof the five SCSs. Results are summarized in Fig. 4,which shows performance averaged across the fiveSCSs as a function of person category, testing wave,and age. The person category exerted a highly sig-nificant effect, F(3, 210) = 49.27, p = .0001, indicat-ing clearly that the expected differences between thefour groups remained stable across the three test-ing waves. Age exerted a marginally significant ef-fect, F(4, 210) = 3.16, p = .02, indicating a tendencyfor self-evaluation scores to increase with age. How-ever, the interaction between age and person categorywas nonsignificant, F(12, 210) = 0.86, p = .592, indi-cating that differences among the four person cate-gories remained stable across the five age groups.

The effect of testing wave was also highly sig-nificant, F(2, 420) = 65.18, p = .0001, indicating thatself-evaluation scores increased from one wave to thenext, particularly from the first to the second wave.This increase seems to reflect the performance im-provements, which occur during this phase of life.The significant Person category ×Wave, F(6, 420) =25.72, p = .0001, and Person category × Age inter-action, F(8, 420) = 2.76, p = .006, indicated that thedifferences among the four person categories tendedto diminish from one wave to the next or with increas-ing age, respectively.

These findings suggest that self-evaluationand self-representation remain generally consistentthrough time. Individuals tend to be consistently ac-curate, lenient, or strict in their self-evaluation of cog-nitive functioning and other dispositions. This sug-gests a kind of internal transformer, which adjustsevaluations and representations, related to particu-lar behaviours or events to remain consistent withone’s self-definitions regardless of environmental ordevelopmental changes. This transformer seems to be“driven” by a factor that may be termed personal con-stant, meaning that the individual adopts a particularattitude to his or her performance and ability and con-sistently applies it across different domains. This con-stant will adjust any signals regarding the individual’sfunctioning to a level that is personally characteristic.

This personal constant undergoes major redef-initions or rescalings at crucial points in develop-ment when cognitive potentials are transformed. For

P1: FLT



Fig. 4. Postperformance success evaluation averaged across SCSs as a function of age, person category, and testing wave.

instance, during the transition from childhood to ado-lescence, self-evaluation scores first decrease but arethen restored to previous levels. After this change,self-evaluations become much more accurate (Harter,1999).

These interpretations imply, therefore, that hy-percognitive maps may accurately represent the ar-chitecture of an individual’s mind and self and yetmay be accurate or inaccurate in the evaluation ofpersonal performance and behavior. Thus, the hyper-cognitive maps accurately reflect relative differencesbetween cognitive and personality functions and abili-ties within the individual. The personal constant, how-ever, is consistently applied across all cognitive andpersonality functions and abilities, thereby explainingdifferences in self-evaluation and self-representationamong individuals. The personal constant can there-fore explain why the two differences are generally sta-ble, both the relative differences between abilities andtendencies within persons and the relative differencesin self-evaluation and self-representation across per-sons vis-a-vis each of the abilities and tendencies. Inother words, it explains why hypercognition can simul-taneously preserve an accurate structure of cognitive

processes in the mental maps it constructs and also in-volve a strong personal element in terms of how theseprocesses are effected (or are thought to be effected)in the real world.

STUDY 2: COGNITIVE DEVELOPMENT,SELF-REPRESENTATION, AND THE BIGFIVE FACTORS OF PERSONALITY

The studies thus far presented have focussedon cognitive performance, self-representations aboutcognitive performance, thinking styles, and person-ality. However, our tests did not directly addressthose dimensions of personality considered dominantby current personality theory (see Costa & McCrae,1997; Graziano et al., 1997), nor did our inventoryof cognitive self-image examine some important di-mensions of the mind as specified by our theory. Toremedy this limitation, a series of new studies weredesigned to examine in detail the various dimensionsof cognitive self-representation and thinking styles inrelation to the Big Five factors of personality (that is,extraversion, agreeableness, conscientiousness, neu-roticism, and openness to experience). The study’s

P1: FLT


164 Demetriou

aim was to specify how the Big Five factors interrelatewith both cognitive self-representation and thinkingstyles.

Participants

Three hundred and twenty two students(265 female and 57 male) from the faculties of hu-manities and social sciences of the University ofCyprus participated in this study. The mean age of thisgroup was 19.3 years. These students came from allover Cyprus and from varying socioeconomic groups.However, because admission to the University ofCyprus is highly selective, these students representthe top 10–15% of the population, in relation to aca-demic performance is concerned.

Inventories

Three inventories were used in the study. Per-sonality was examined with the Greek version ofan inventory addressed to the “Big Five” factors(Besevegis, Pavlopoulos, & Mourousaki, 1996), andproved highly reliable, as Cronbach’s alpha rangedfrom .86 to .90 for the set of items addressed to eachof the five factors and .92 for the whole test.

The second inventory focussed on self-representation and addressed the same dimensionstested in the first study as well as a series of newadditional items. The new items examined self-imagein regard to various aspects of (i) speed of process-ing, (ii) working memory, (iii) self-monitoring andself-regulation, and (iv) deductive and inductivereasoning. Thus, in this study, the cognitive self-imageinventory considered most dimensions of each ofthe three levels of the mind (i.e., the processingsystem, the environment-oriented systems, andthe hypercognitive system). The reliability of thisinventory was again very high: it ranged between .72and .87 for the items addressed to each of the variousdimensions of cognitive self-representation and .85for the whole test.

Thinking styles were evaluated with a shortenedversion of the inventory used in the first study (de-scribed earlier), and reliability of this inventory wasalso satisfactory (.60).


Exploratory factor analysis was first applied toall three inventories together to verify that all dimen-

sions of cognitive self-representation, personality, andthinking styles were actually present. The analysis re-vealed that all dimensions appeared as distinct fac-tors, thus justifying their use in the structural equationmodel to be described below (see Demetriou & Kazi,2001).

To specify the relationships among the variouskinds of factors, a complex structural equation modelwas built (see Fig. 5). The model assumed that thefactors are organised in several tiers: the cognitivefactors are located at the lower layer of this architec-ture, personality factors are placed in the middle, andthinking styles at the top. Locating personality factorsin the middle does not imply that they are consideredless basic than the cognitive factors in determining be-haviour; rather, this architecture simply indicates thatcognitive factors influence derivative factors (such asthinking styles), and actual behaviour through the in-terface of the personality factors.

The pattern of relationships is highly interesting.Specifically, it can be seen that four of the five di-mensions of personality are closely related to those ofcognitive self-representation, for instance, individualswho have a strong sense of cognitive efficiency are alsoopen to experience. In fact, the relationship betweenthe factors of perceived general cognitive efficiencyand openness to experience is so high (.95) that theymay be considered as two sides of the same coin. Alsonotably high were the relationships between extraver-sion (.82) , agreeableness (.63), and conscientiousness(.52) and perceived cognitive efficiency. Only neuroti-cism was found to be completely unrelated to cogni-tive efficiency.

Different combinations of personality character-istics appear to result in different thinking styles. Themodel suggests that individuals who are open to ex-perience and emotionally stable tend to score highon the ideal self factor; in other words, they thinkhighly of themselves and want to be the centre of so-cial attention. This is suggested by the rather strongrelationships between these two personality dimen-sions with the factor representing ideal self. In turn,thinking highly of oneself together with being open toexperience orients the individual to activities requir-ing originality. Those who think highly of themselvesbut are not open to experience tend to have an evalu-ative style of thought, which orients them to activitiesrequiring evaluation. Interestingly, a combination ofemotional stability with extraversion seems to lead toimpulsivity; impulsivity itself is negatively related toconscientiousness, which is, however, related to sys-tematicity. In turn, systematic personalities tend to

P1: FLT



Fig. 5. The model of the relations between cognitive self-image factors, personality factors, and thinking styles factors.

have an executive style, which leads them to activitieswith predefined rules. We believe that these findingsare a first step in the direction of integrating the psy-chology of the mind with the psychology of person-ality. The study presented below provides more evi-dence in this direction.

STUDY 3: COGNITIVE DEVELOPMENT,SELF-REPRESENTATION, AND THE BIGFIVE FACTORS OF PERSONALITY

All measures generated by the second study sum-marized above stand for self-representations or self-characterizations. Thus, while this study addresses therelationships between the various dimensions of theself-oriented level of the mind and the self, it cannotdirectly clarify correlations among abilities residingat the environment-oriented level, the various dimen-sions of self-representation, and personality. A seriesof studies currently under way in our laboratory aim toremedy this weakness by including both measures of

actual performance and also various measures of self-representation and personality. One of these studiesis summarized below.

Participants

This study included a total of 891 participants, dis-tributed by age as follows: 80 (mean age 10.5 years)and 83 (mean age 11.5 years) primary school childrendrawn from grades 5 and 6, respectively; 67 (mean age12.5 years), 73 (mean age 13.5 years), and 74 (meanage 14.5 years) young adolescents drawn respectivelyfrom grades 1, 2, and 3 of junior high school; 147 (meanage 15.5 years), 131 (mean age 16.5 years), and 137(mean age 17.5 years), older adolescents drawn fromgrades 1, 2, 3, of senior high school, respectively; fi-nally, 99 (mean age 21 years) university students. Thegenders and three socioeconomic classes (rural, work-ing class, and middle class) were about equally repre-sented in each age group. Moreover, the participantsfrom senior high school were about equally drawn

P1: FLT


166 Demetriou

from the two main directions in Cypriot secondaryschools, namely humanities and sciences.

Task Batteries and Inventories

Participants were examined with a cognitive de-velopment test, which included batteries addressedto five of the SCSs specified by our theory, (quanti-tative, causal, spatial, qualitative, and propositional).The quantitative thought battery addressed skills innumerical operations, numerical analogies, and alge-braic reasoning (Cronbach’s alpha .86). The causalthought battery focused on the ability to formu-late and test hypotheses by designing experimentsand the ability to understand various types of ca-sual relationships, such as necessary and sufficient,necessary and insufficient, unnecessary and insuffi-cient, etc. (Cronbach’s alpha .55). The spatial thoughtbattery included items addressed to formation ofmental images, mental rotation, and coordination ofperspectives and frames of reference (Cronbach’s al-pha .85). The qualitative thought battery tested classinclusion, Raven-like reasoning, and verbal analo-gies (Cronbach’s alpha .77). Finally, the propositionalthought battery examined class reasoning, proposi-tional reasoning, and reasoning in a pragmatic con-text (Cronbach’s alpha .62). As a whole, this test washighly reliable (Cronbach’s alpha for the entire testis .90).

Moreover, the participants were examined by thecognitive self-representation inventory used in Study2 described above (Cronbach’s alpha .91). The BigFive Factors of personality were addressed by 50 ofitems drawn from the personality inventory used inStudy 2. Specifically, the 10 items with the highestloading for each of the five factors were selected to ad-dress personality in this study (Cronbach’s alpha .83).Thus, this study provides the data required to examinethe precise relationships among the various SCSs, aswell as those among cognitive self-representation andpersonality from late childhood to early adulthood.


A complex model was fit to the data for boththe entire sample of 891 participants and for thefour important phases of development, namely latechildhood, early adolescence, late adolescence, andearly adulthood. Primary school children, junior highschool participants, senior high school participants,and university students were pulled together to form

each of these four groups. This manipulation allowedthe various factors and relationships over the wholesample to be specified, as well as how these factors andrelationships change, if at all, across these four crucialphases of development. Fitting the model to each agegroup was not possible because of the limited numberof subjects.

The model included both first- and second-orderfactors (see Fig. 6). Each of the five SCSs, which repre-sent performance on the corresponding task battery,was assigned a first-order factor that was specified inreference to three mean scores representing perfor-mance on the items addressed to a domain. Thesefactors were then regressed on a second-order fac-tor. In terms of our theory, this factor represents theenvironment-oriented level of the mind and it corre-sponds to the general cognitive factor specified by thefirst of the studies presented here.

There was also one first-order factor for self-representation of each of the five SCSs based on theself-attribution of ability scores given to the respec-tive items on the self-representation inventory. Again,each of these factors was specified in reference tothree scores representing self-attribution of ability inregard to a domain. All five self-representation factorswere regressed on a second-order factor, representingthe self-oriented or, in our terms, the hypercognitivelevel of the mind. This factor corresponds to the gen-eral hypercognitive factor of the first study. The fiveSCS-specific self-representation factors were also re-gressed on the residual of the corresponding perfor-mance factor. This manipulation enables us to specifyhow (if at all) the domain-specific self-representationfactors depend on their corresponding ability-specificfactor, as well as their dependence on forces regardingself-representation.

The Big Five factors of personality were also as-signed a first-order factor, specified in reference tothree scores representing self-attribution of person-ality characteristics related to a personality dimen-sion. Each factor was regressed on both the generalcognitive performance and the general cognitive self-representation factor. This manipulation allows therelations among the personality factors and both thecognitive and the hypercognitive levels of the mind tobe specified.

Finally, the two second-order factors were al-lowed to correlate in the fashion of the model shownin Fig. 2.

The fit of the model, applied on the whole sam-ple was good, χ2(830) = 1627.812, χ2/df = 1.961,p = .001, CFI = .919, GFI = .920, AGFI = .909,

P1: FLT


Fig. 6. The model captures the organization of cognitive abilites, self-representation about cognitive abilites, and the Big Five Factors ofpersonality. VP, CE, QA, QR, and SI stand for the verbal–proportional, the causal–experimental, the qualitative–analytic, the quantitative–relational, and the spatial–imaginal system, respectively. A, O, N, C, and E stand for agreeableness, openness to experience, neuroticism,conscientiousness, and extroversion, respectively. Asterisks indicate significance of parameters. Asterisks indicate significant loadings.

167

P1: FLT


168 Demetriou

RMSR = .053. The model was further tested with amultiple group analysis, involving the four age groups;no equality constraint was imposed in this model.This analysis allowed the model parameters to bespecified separately for each age group so that theywould be compared for possible differences. The fitof this model was also satisfactory, especially consid-ering the complexity of the model and the presence ofmultiple groups χ2(3279)= 4658.809, χ2/df = 1.420,p = .001, CFI = .844.

We note first the parameters of the model testedon the entire sample, because this sample’s size(891 subjects) and variability (in terms of age andorigin of subjects) lend considerable robustness tothis model. The parameters of foremost importance inthis model are those concerned with the relationshipsamong both the personality factors and the cogni-tive factors. Inspection of the model reveals first, thatthe effects of the general self-representation factor onthe personality factors were very similar to the corre-sponding effects revealed by Study 2. Thus, there wasa very strong effect of self-representation on open-ness to experience (.81), lower but still significant andsubstantial effects on agreeableness (.57), conscien-tiousness (.54), and extraversion (.46), and no effecton neuroticism (.04).

Second, the effects of the cognitive performancefactor on the personality factors were much weakerin all cases. Moreover, it is noticeable that, when-ever significant, these effects were negative, (agree-ableness:−.10; conscientiousness:−.25; extraversion:−.17). We interpret these results to indicate some-thing broader than a direct relationship betweencognitive ability and personality or between intel-lectual development and personality development.That is, we suggest that the size and the directionof these relationships does not justify assuming adirect relationship between cognition and person-ality. Rather, this pattern of relationships reflectsa broader phenomenon regarding the relationshipsbetween development of cognitive abilities and thedevelopment of self-representation. That is, the cog-nitive changes, which occur at some phases of devel-opment, make the persons conservative in their self-attribution of ability or characteristics. As a result,increases in cognitive performance scores are as-sociated with a general decrease in self-attributionscores (Demetriou & Kazi, 2001). In terms of thefindings of our Study 1, this pattern of relation-ships may reflect the rescaling of the personal con-stant that occurs from early adolescence to earlyadulthood.

Third, the intercorrelation of the two general fac-tors was significant, indicating that the two levels ofthe mind do in fact interact. This confirms the findingsof Study 1.

Fourth, the direct effects of the domain-specific factors on their corresponding self-representation factors were generally very low.In fact, this effect was significant only in the caseof quantitative thought (.20); this supports theassumption that self-representations about specificdomains are primarily derived from a general self-representation of cognitive efficiency. Neverthelesssome domains (e.g., mathematics), are more trans-parent to awareness than others, and contributedirectly to both the formation of self-representationand general cognitive efficiency.

The parameters of the model tested on the fourage groups were generally the same as those ofthe model tested on the whole sample. However, iftaken in relation to each other, the four age-specificmodels do reveal some interesting developmentaltrends. First, the effect of the self-representation fac-tor on three personality factors (agreeableness, con-scientiousness, extraversion) diminished systemati-cally from the younger to the older age group. Thisdevelopmental trend was not observed for openness,where the effect of self-representation remained al-ways very strong, nor for neuroticism, where the ef-fect was always negligible. Second, the effect of thegeneral cognitive performance on the general self-representation factor was nonsignificant at primaryschool but became significant thereafter. This find-ing, which concurs with the model shown in Fig. 2,further supports our assumption that a relatively ac-curate communication between the two levels of themind (in regard to the processes studied here) is notestablished until early adolescence.

In sum, it appears that the relationships betweenpersonality and actual cognitive abilities are indirectrather than direct; more specifically, are mediated bythe hypercognitive level of the mind. Our findings re-veal that the dimensions of personality are not closelyrelated to cognitive performance itself but rather areclosely related to cognitive self-representation. Theserelationships are so close that it may be accurate toassume that personality as examined by self-reportinventories is part of the general self-representationsystem. Alternatively, it may be concluded that all butone (neuroticism) of the Big Five dimensions of per-sonality have a strong reflected cognitive component.We, however, note that, apart from openness to expe-rience, the dependence of personality dimensions on

P1: FLT



the dimensions of cognitive self-representation tendto weaken with age. This process of differentiationevolves throughout adolescence, becoming clear inearly adulthood.

CONCLUSIONS

The studies summarized here suggest that themind is simultaneously hierarchical, modular, andtransmodular. It is hierarchical because it involvessystems that are organised at different levels, accord-ing to orientation and functioning. By definition, self-oriented systems stand on a level higher than theenvironment-oriented systems because they are ap-plied on them in order to monitor and regulate theirfunctioning. Processing potentials stand in-betweenas a dynamic field that interlocks the two know-ing levels, thus constraining their functioning and in-teractions. Moreover, the mind is modular at boththe level of environment-oriented cognitive processesand personality dispositions and the level of their self-representation. In fact, modularity at both levels isvery strong for good reasons: it enables specializedand therefore efficient interaction with the domainconcerned.

However, the mind is also unified or transmodu-lar, insofar as the very ability to oversee, record, anddifferentiate modules is by definition a transmodularfunction. Moreover, the personal constant identifiedhere is a transmodular mechanism because it operateson all modules to allow systematic adjustments to bothperformance and subjectivity. This constant integratesthe effects of general processing efficiency and tem-perament. Thus, this constant represents processesthat directly interconnect mind and personality. Allthree studies presented here showed clearly that self-representations about cognitive functions and person-ality dimensions are considerably cohesive, althoughdirect relationships between cognitive abilities andpersonality were very weak and negative.

Therefore, it may be said that humans have mindbecause there are environment-oriented systems andinternal dispositions of which to become mindful. Theself is the personalized aspect of this system and itmay refer to cognitive abilities and processes (e.g., Iam good in reasoning, learning, mathematics, draw-ing, social interaction, etc.) or to personality traits ordispositions (e.g., I enjoy being with people; I am sta-ble; I am irritable, etc.). These traits and dispositionsdetermine individual idiosyncratic functioning. At thesame time, everything is embedded in a social contextwhere minds, selves, and personalities interpret, in-

teract, and interrelate (Demetriou & Kazi, 2001). Infact, observing and interacting with others may bea good platform for testing assumptions about one’sown self and mind. These assumptions are consistentwith modern evolutionary theory, which argues thatmindfulness emerged as a result of self-observation,self-evaluation, and self-mapping as well as the humancapacity for observation and manipulation (Bogdan,2000; Mithen, 1996).

Development appears to exert two seeminglycontradictory kinds of influences. First, it seems tostrengthen the tuning between the two basic know-ing levels of the mind. The reader is reminded thatthere is a structural change in the relationship be-tween the hypercognitive and the other levels ofthe mental architecture, which begins at the age 11–12 years. It is widely accepted that cognitive abili-ties are transformed extensively at this age, and ourstudies have shown that at this age the personal con-stant is rescaled, obviously to tune the person’s self-representations with his or her new cognitive abili-ties and potentials. This interlevel tuning improvesuntil the end of adolescence. In turn, this results inincreasing communication between the environment-oriented and the self-oriented level of the mind. Asa side effect of this increasing communication, self-attribution of abilities and characteristics tend to be-come stricter and more accurate with either age orimprovement in cognitive performance.

Second, the findings of Study 3 suggest that thisincreasing communication between the levels of mindcoevolves with an increasing differentiation betweenself-attributions related to personality and those re-lated to cognition. The reader is reminded that dur-ing the transition from adolescence to adulthood cor-relations between the cognitive and the personalityaspects of the self seem to decline. Taken together,these findings suggest that from late childhood toearly adulthood humans become more accurate self-observers and, probably as a result of this, more able toclarify and articulate various abilities and attributes.

This conclusion implies that the transition toadulthood is associated with a kind of multifurgationof the personal constant. That is, the two principalrealms of the self (cognitive and the personality), andpossibly certain subrealms acquire their own constantwhich is then used to calibrate domain-specific per-formance. Extensive study is needed to map the pro-cesses that cause this multifurgation, and to specifythe new domain-specific constants, their possible in-terrelationships, and of course, their relationships tothe original and a general personal constant. It may

P1: FLT


170 Demetriou

be noted that these findings agree with other researchwhich indicates that young adults begin to operate assystematic self-theorists who are concerned with boththeir own and others’ profile of abilities and charac-teristics (Fischer & Ayoub, 1994; Harter, 1999).

It would be interesting to extend this model toaccommodate development in the years of maturityand late adulthood. It has been mentioned in the intro-duction that the various aspects of the environment-oriented systems show positive changes until aboutthe age of 50 years when they begin to show negativechanges. For instance, speed and control of process-ing decrease, working memory declines, and reason-ing and problem solving loose some of their efficiencyand flexibility. There are also personality changes,particularly in those dimensions most closely relatedto cognition, as individuals become less open to ex-perience and less extroverted with age. Two of ourstudies (Studies 2 and 3) showed clearly that self-representation of cognitive efficiency is closely relatedto both actual cognitive performance, and these twodimensions of personality. These links justify studyingwhether these personality changes in the latter yearsare caused by coexistent cognitive changes. Furtherresearch is needed to provide a direct answer to thisquestion.

A number of scholars have argued that losses atthe more fundamental levels of the mind are compen-sated by gains in the many functions and skills associ-ated with self-understanding, self-management, andmanagement of complex social and interpersonal is-sues, at least until the end of middle age. These authorshave argued that the mind becomes dialectical aftera certain phase (Labouvie-Vief, 1990; Pascual-Leone,1983; Riegel, 1973). Dialectical thought enables in-dividuals to envisage and accept alternative concep-tions of the same aspect of reality to attain a higherorder synthesis. In this synthesis alternative concep-tions are recognized for whatever unique or valuablecontribution they make to the process of understand-ing. All these theorists believe that dialectical thoughtrequires explicit reflection and self-awareness.

Dialectical thought is also implicated in the ac-quisition of wisdom, which is supposed to come withage. According to Baltes and Smith (1990), wisdomis “exceptional insight into human development andlife matters, exceptionally good judgement, advice,and commentary about difficult life problems” (p. 95).Wise judgement integrates factual knowledge of theevents and pragmatics of life, strategies for dealingwith them, knowledge which considers uncertainties,and a relativist attitude toward life values and goals. Infact, wisdom is possible because minds interact with

each other. The attainment of wisdom is highly use-ful for both the individual and the social group, sinceit can integrate the best interests of both into bal-anced decisions and policies. Moreover, wisdom en-ables the individual to accept the negative changesthat aging brings, such as the cognitive changes notedabove and the decline in general physical fitness andhealth (Baltes, 1991; Baltes & Staudinger, 1996).

However, there is no evidence on how thesechanges are related to possible changes in the relation-ships between personality and the mind. It could be as-sumed, for instance, that the weakened relationshipsbetween personality and cognition discussed abovemake the person more sensitive to the various combi-nations of personality, cognitive talents, and abilities.In turn, this may make the person more tolerant ofalternative ways of interpreting the world and actingin it. By implication, increasing tolerance may leadto a dialectical approach to problems and to wisdomin judgement. Thus, the changes discussed above inthe relationships between the levels of the mind andthe personal constant may lie at the background of theacquisition of dialectical thought and wisdom. Obvi-ously, there is no evidence directly related to these as-sumptions. It is hoped that the model and the studiespresented here indicate the directions we would haveto follow, in order to highlight these phenomena.

REFERENCES

Baltes, P. B. (1991). The many faces of human aging: Toward apsychological culture of old age. Psychological Medicine, 21,837–854.

Baltes, P. B., & Smith, J. (1990). Toward a psychology of wisdom andits ontogenesis. In R. J. Sternberg (Ed.), Wisdom: Its nature,origins, and development (pp. 87–120). New York: CambridgeUniversity Press.

Baltes, P. B., & Staudinger, U. M. (1996). Interactive minds in alife-span perspective: Prologue. In P. M. Baltes, P. B. Baltes,& U. M. Staudinger (Eds.), Interactive minds: Life-span per-spectives on the social foundation of cognition (pp. 1–32).Cambridge: Cambridge University Press.

Bandura, A. (1989). Regulation of cognitive processes through per-ceived self-efficacy. Developmental Psychology, 25, 729–735.

Bentler, P. M. (1989). EQS: Structural equations program manual.Los Angeles: BMDP Statistical Software.

Besevegis, E., Pavlopoulos, V., & Mourousaki, S. (1996). Children’spersonality characteristics as assessed by parents in naturallanguage. Psychology: The Journal of the Hellenic Psycholog-ical Society, 3, 46–57.

Bogdan, R. J. (2000). Minding minds. Cambridge, MA: MIT Press.Bracken, B. (1996). Clinical applications of a context-dependent

multi-dimensional model of self-concept. In B. A. Bracken(Ed.), Handbook of self-concept. New York: Wiley.

Carroll, J. B. (1993). Human cognitive abilities: A survey of factor-analytic studies. Cambridge, UK: Cambridge University Press.

Case, R. (1992). The mind’s staircase. Hillsdale, NJ: Erlbaum.Case, R., Demetriou, A., Platsidou, M., & Kazi, S. (2001). Integrat-

ing concepts and tests of intelligence from the differential anddevelopmental traditions. Intelligence, 29, 307–336.

P1: FLT



Costa, P. T., Jr., & McCrae, R. R. (1997). Longitudinal stability ofadult personality. In R. Hogan, J. Johnson, & S. Briggs (Eds.),Handbook of personality psychology (pp. 269–290). San Diego,CA: Academic Press.

Demetriou, A. (1998a). Nooplasis: 10+ 1 Postulates about the for-mation of mind. Learning and Instruction: The Journal of theEuropean Association for Research on Learning and Instruc-tion, 8, 271–287.

Demetriou, A. (1998b). Cognitive development. In A. Demetriou,W. Doise, & K. F. M. van Lieshout (Eds.), Life-span develop-mental psychology (pp. 179–269). London: Wiley.

Demetriou, A., (2000). Organization and development of self-understanding and self-regulation. In M. Boekaerts, P. R.Pintrich, & M. Zeidner (Eds.), Handbook of self-regulation(pp. 209–251). New York: Academic Press.

Demetriou, A., & Efklides, A. (1985). Structure and sequence offormal and postformal thought: General patterns and individ-ual differences. Child Development, 56, 1062–1091.

Demetriou, A., & Efklides, A. (1988). Experiential Structuralismand neo-Piagetian theories: Toward an integrated model. In A.Demetriou (Ed.), The neo-Piagetian theories of cognitive de-velopment: Toward an integration (pp. 173–222). Amsterdam:North-Holland.

Demetriou, A., & Efklides, A. (1989). The person’s conceptionof the structures of developing intellect: Early adolescence tomiddle age. Genetic, Social, and General Psychology Mono-graphs, 115, 371–423.

Demetriou, A., Efklides, A., & Platsidou, M. (1993). The architec-ture and dynamics of developing mind: Experiential structural-ism as a frame for unifying cognitive developmental theories.Monographs of the Society for Research in Child Development,58 (5, Serial No. 234).

Demetriou, A., & Kazi, S. (2001). Unity and modularity in the mindand the self: Studies on the relationships between self-awareness,personality, and intellectual development from childhood toadolescence. London: Routledge.

Demetriou, A., Kazi, S., & Georgiou, S. (1999). The emerging self:The convergence of mind, personality, and thinking styles. De-velopmental Science, 2, 387–422.

Demetriou, A., & Raftopoulos, A. (1999). Modeling the developingmind: From structure to change. Developmental Review, 19,319–368.

Demetriou, A., Raftopoulos, A., & Kargopoulos, P. (1999). Interac-tions, computations, and experience: Interleaved Springboardsof Cognitive Emergence. Developmental Review, 19, 389–414.

Demetriou, A., Spanoudis, G., Christou, C., & Platsidou, M. (2002).Decomposing information processing in the Stroop phe-nomenon: A developmental study. Cognitive Development, 16,987–1005.

Ferrari, M., & Sternberg, R. J. (1998). The development of mentalabilities and styles. In W. Damon (Series Ed.), D. Kuhn, &R. Siegler (Vol. Eds.), Handbook of Child Psychology: Vol. 2:Cognition, Perception and Language (5th ed., pp. 899–946).New York: Wiley.

Fischer, K. W. (1980). A theory of cognitive development: The con-trol and construction of hierarchies of skills. Psychological Re-view, 87, 477–531.

Fischer, K. W., & Ayoub, C. (1994). Affective splitting and dissocia-tion in normal and maltreated children: Developmental path-ways for self in relationships. In D. Cicchetti & V. Toth (Eds.),Rochester Symposium on dysfunctions of the self (pp. 149–222).Rochester, NY: Rochester University Press.

Flavell, J. H., Green, F. L., & Flavell, E. R. (1995). Young chil-dren’s knowledge about thinking. Monographs of the Societyfor Research in Child Development, 60(1, Serial No. 243).

Gardner, H. (1983). Frames of mind: The theory of multiple intelli-gences. New York: Basic Books.

Graziano, W. G., Jensen-Campbell, L. A., & Finch, J. F. (1997). Theself as a mediator between personality and adjustment. Journalof Personality and Social Psychology, 73, 392–404.

Graziano, W., Jensen-Campbell, L. A., & Sullivan, G. (in press).Temperament, activity and expectations for later personalitydevelopment. Journal of Personality and Social Psychology.

Halford, G. (1993). Children’s Understanding: The development ofmental models. New-York: Erlbaum.

Harter, S. (1999). The construction of the self. New York: GuilfordPress.

Hattie, J. (1992). Self-concept. Hillsdale, NJ: Erlbaum.James, W. (1892). Psychology: The briefer course. New York: Henry

Holt.Jensen, A. R. (1998). The g factor: The science of mental ability.

New York: Praeger.Kagan, J. (1994). On the nature of emotion. Monographs of the

Society for Research in Child Development, 59, 7–24 (2/3, SerialNo. 240).

Kallio, E. (1998). Training of students’ scientific reasoning skills.Jyvaskyla: University of Jyvaskyla.

Kargopoulos, P., & Demetriou, A. (1998). Logical and psychologi-cal partitioning of mind. Depicting the same map? New Ideasin Psychology, 16, 61–87.

Kohnstamm, G. A., & Mervielde, I. (1998). Personality develop-ment. In A. Demetriou, W. Doise, & K. F. M. van Lieshout(Eds.), Life-span developmental psychology (pp. 399–445).London: Wiley.

Labouvie-Vief, G. (1990). Wisdom as integrated thought: Histori-cal and developmental perspectives. In R. J. Sternberg (Ed.),Wisdom: Its nature, origins, and development (pp. 52–83). NewYork: Cambridge University Press.

Markus, H., & Wurf, E. (1987). The dynamic self-concept: A socialpsychological perspective. Annual Review of Psychology, 38,299–337.

Mithen, S. (1996). The prehistory of the mind: The cognitiveorigins of art, religion, and science. New York: Thames &Hudson.

Pascual-Leone, J. (1983). Growing into human maturity: Towardsa metasubjective theory of adulthood stages. In P. B. Baltesand O. G. Brim (Eds.), Life-span development and behavior(Vol. 5., pp. 117–156). New York: Academic Press.

Pascual-Leone, J. (1988). Organismic processes for neo-Piagetiantheories: A dialectical causal account of cognitive develop-ment. In A. Demetriou (Ed.), The neo-Piagetian theories ofcognitive development: Toward an integration (pp. 25–64).Amsterdam: North-Holland.

Perner, J. (1991). Understanding the representational mind.Cambridge, MA: MIT Press.

Piaget, J. (1970). Piaget’s theory. In P. Mussen (Ed.), Carmichael’smanual of child psychology (Vol. 1). New York: Wiley.

Riegel, K. F. (1973). Dialectical operations: The final period of cog-nitive development. Human Development, 16, 346–370.

Shaie, K. W., & Willis, S. L. (1991). Adult development and aging.New York: Harper Collins.

Shayer, M., Demetriou, A., & Pervez, M. (1988). The structureand scaling of concrete operational thought: Three studies infour countries. Genetic, Social, and General Psychology Mono-graphs, 114, 307–376.

Sternberg, R. J. (1985). Beyond IQ: A triarchic theory of humanintelligence. New York: Cambridge University Press.

Sternberg, R. J. (1988). Mental self-government: A theory of in-tellectual styles and their development. Human Development,31, 197–224.

Thurstone, L. L. (1938). Primary mental abilities. Chicago: ChicagoUniversity Press.

Wellman, H. M. (1990). The child’s theory of mind. Cambridge, MA:MIT Press.

mind, self, and personality: dynamic interactions from ... self... · this paper aims to integrate...

Documents