Abstract
We often say that persons need to change themselves. But what is this “self” that needs to change, and why do “selves” seem to be so resistant to change? I have argued elsewhere that who I really am, in an everyday practical sense, is who I really am with respect to the moral evaluation of me. In other words, the everyday self that (often) needs changing is a moral self. In this article, I am offering further indirect evidence for this thesis by exploring an alternative possibility—that who I really am is my personality in the fashionable “Big Five” or “Five-Factor Model” sense—and showing that this suggestion does not bear scrutiny. I argue that although Big-Five theory has identified relatively stable within-person patterns, it has not shown these patterns to be psychologically meaningful, except to the extent that they are morally salient. I argue further that the nature of our diachronic moral selves suggests that moral philosophy needs to take a developmental and educational turn: a turn for which it may, however, not be well equipped.
Philosophical misgivings notwithstanding (most famously Hume, 1739–1740/1978, Book I), everyday language entertains little doubt about the existence of selfhood. Such language bristles with references to people who “are no longer themselves” after undergoing some life-changing experiences, and to others whose selves “need changing.” In Rilke’s poem “Archaischer Torso Apollos,” the Greek sculpture—headless, armless, yet so full of life—beckons us to change: “Du mußt dein Leben ändern” (Rilke, 1996, p. 513). None of us is exempt from its uncompromising eyeless gaze. We all need to change our lives by changing ourselves. And of course we all do change, at least in the platitudinous sense in which we develop from childhood onwards in a gradual, continuous way. Empirical research shows that stability of personal characteristics from childhood to adulthood, although significant in many cases, is poor enough to indicate that considerable self-change does occur prior to adulthood (Cloninger, 2009, p. 12). What typically happens after that is more debatable. In his trilogy His Dark Materials, Philip Pullman (2011) uses the ingenious ploy of making the personalities of his characters external to them in the form of accompanying “daemons.” Children in these stories—owing to the plasticity of their personalities—have shape-changing daemons, but as they come of age, their daemons settle on a single form that reflects their essence. Yet adults also sometimes undergo radical self-changes in real life: witness Saul’s transformation into Paul on the way to Damascus, where he became—in some sense—another person (cf. Gunnarsson, 2010, pp. 7, 73).
In this article, I am not going to explore the whys and hows of what happens when a person changes (I have dealt with it in Kristjánsson, 2010b, chap. 10). Rather, my focus is on that everyday self which Rilke’s Apollo prompts us to change. When we sit down in the pub or at the kitchen table and talk about the need for ourselves or somebody else to change—whether or not such a change is in fact feasible or does actually occur—what is this alleged self that is said to need changing? What is the vehicle of the called-for transformation?
By referring to the “everyday self” of ordinary language, I am not implying that the word “self” has an unequivocal meaning in everyday parlance that somehow distinguishes it from the vagaries or theory-ladenness of philosophical usage. In both everyday conversation and nuanced philosophical discussion, the word “self” can have many meanings, and what we are after each time—that is, from what perspective the person’s “core” is being considered—must be deduced from the context: is it the “metaphysical self,” the “phenomenological self,” or the “everyday self,” for instance (cf. Kristjánsson, 2010b, pp. 6-7)? It is only the last of those that is on my present agenda. With regard to that specific perspective, however, I am not implying that the concept of an everyday self can be defined unproblematically from the convenience of a philosophical armchair, or through a simple atheoretical and ahistorical study of ordinary language. In a recent book, Svend Brinkmann (2011) chronicles the scientific trajectory of the idea of who we really are, as it has moved from realist notions of self as moral character, through realist notions of self as non-moral personality, to anti-realist notions of self as mere identity or self-concept. He also shows convincingly how this trajectory has developed in tandem with more general paradigm-shifts in psychology involving the psychologization of morality and the subjectivization of psychology. He further demonstrates the historical contingencies of these scientific developments as they reflect shifts in our academic world-view towards the instrumentalization of ends and the individualization of meaning. Those academic twists and turns notwithstanding, I think we have good reason to doubt that the notion of an everyday self at issue in ordinary kitchen-table talk about people who have changed or need to change themselves (if they are to follow Apollo’s call) has evolved quite as dramatically from the time of Rilke, or even that of Socrates, as have academic fashions. In any case, all open-textured naturalistic concepts, such as that of a person’s everyday self, are in the end answerable to empirical facts about human beings and their environment. As I have argued elsewhere, all good analyses of such concepts must, therefore, be critical and potentially revisionary with regard to existing linguistic conventions and intellectual fashions (Kristjánsson, 1996, chap. 7).
Consider, for example, the recently fashionable thesis that the everyday self is the same as self-concept: the set of beliefs one entertains with regard to oneself. Self-change is then equal to self-concept-change. At first blush that may seem to make self-change relatively uncomplicated: one simply needs to change one’s beliefs about who one is. Indeed, postmodernists, with their routine rejection of objective entities or anything of depth, relish anti-self-realism of this kind precisely because of its implication of the unencumbered renegotiation of self-identities (see, e.g., Gergen, 1991). Realizing, however, that this implication runs counter to our everyday experiences of the difficulty of self-change, more sophisticated anti-self-realists among psychologists provide us with intricate explanations for why self-concept-change is such an arduous process and why we remain so set in our ways. Witness, for instance, William B. Swann’s (1996) self-verification theory, which posits our need for homeostatic self-confirmations as being stronger than our need for self-enhancements. Such maneuvers notwithstanding, a problem remains for anti-self-realism of whatever ilk: it seems easy to think of cases of self-concept-change without self-change and of self-change without self-concept-change. Even if you think you are a duck, and even walk and talk like one, this does not make you a duck. The discrepancy between self-concept and self is no mystery. It is called self-deception. The essential advantage of self-realism over anti-self-realism lies in its ability to account for self-deception. According to self-realism, when my self-concept gets things right, it has my actual full self as its cognitive content. When it does, I experience self-knowledge. When it does not, I am self-deceived. It is hard to make do without the assumption of an actual full self; even Swann needs a self-trait that is not part of self-concept to account for the intractability of self-concept change—namely our desire for self-equilibrium or self-stability.
I keep saying what the everyday self is not, but what do I then take it to be? In a recent book (Kristjánsson, 2010b), I argue that the everyday self is a moral self: the subject of moral agency and the object of moral evaluation. I draw there on arguments from ordinary language, from Aristotle’s notion of moral character, and from Hume’s pride-produces-the-everyday-self thesis (in Hume, 1739–1740/1978, Book II—as distinct from his more famous no-metaphysical-self thesis in Book I). Nevertheless, my theory is not purely linguistic, purely Aristotelian, or purely Humean. It is also important to distinguish it clearly from the recently fashionable moral-self thesis in moral psychology, according to which persons of moral virtue are characterized by their defining themselves in moral terms (see e.g., Blasi, 2005). That thesis is all about a moral self-concept, the opposite of which is an immoral or amoral self-concept. My theory in the above-mentioned book is about the everyday self that we all possess, the opposite of which would be a non-moral self. I maintain that what we are really after, in everyday discussions of self and self-change, is that moral self. (Notably, the theory does not entail that all morally evaluable character traits are traits of the moral self; some may be too insignificant to count as parts of who I really am.) The “moral self” in my theory incorporates self-concept (qua beliefs relating to the moral self), but is not exhausted by it. By incorporating self-concept, the theory becomes liable to the regress problem of how a mirror mirrors itself. I suggest, however, that this regress can be halted by assuming that the self is essentially constituted by certain emotions that incorporate both a self-concept part (qua self-beliefs) and a part that belongs to self but not to self-concept (qua self-desires). In general, I argue that the moral self is quintessentially an emotional self: that emotions are implicated in the self at all levels of engagement (see esp. Kristjánsson, 2010b, chap. 4). The emotions that I have in mind there, which are self-constituting and involved in both the self’s creation and its sustenance (cf. Hume, 1739–1740/1978, Book II), are specifically certain self-conscious background emotions of pride and shame that incorporate essentially a moral dimension as they form constitutive parts of our fundamental virtues and vices: states of character that determine who we are “deep down.”
This theory—which I refer to as “an alternative” as opposed to “the dominant” (cognitive, constructivist) self-paradigm—is, I submit, quite radical. It holds that who I really am deep down—what gives me diachronic unity in an everyday practical sense—is who I really am with respect to the moral evaluation of me. What the Apollo statue beckons us to change, then, in order to change our lives, is our moral self. This theory gives succor to objective accounts of wellbeing that consider moral virtues to be conducive to or even constitutive of wellbeing, be they virtue-theoretical accounts of Aristotelian provenance or social scientific accounts of empirical lineage.
As direct arguments for the “alternative” self-paradigm are readily available in the book I just mentioned, I have chosen a different tack for this article. In what follows, I concentrate on another possible answer to the question of what constitutes our Pullmanite daemons: that our everyday selves are our personalities in the currently popular personality-psychology sense of the “Big Five” or “Five-Factor Model.” Simply put, my strategy is this: if another initially plausible alternative account of the everyday self fails to pass muster, and its disadvantages happen to be the flip sides of the advantages in mine, then indirect support has been garnered for my theory by eliminating an attractive alternative. There are two reasons why I consider the Five-Factor Model initially plausible: (a) it upholds a realist account of the everyday self as personality and (b) when what I call “the everyday self” is being discussed in ordinary conversation, the discussion is often couched in terms of “personality.” As Gunnarsson (2010) correctly points out, however, “personality” can, just as “self,” mean a number of different things in different contexts (p. 196, footnote 10). My claim is that what we do mean (or at least should mean if we are being reflective) in everyday conversation when discussing who we really are deep down in terms of personality (and self-change in terms of personality change) is personality understood as moral selfhood. The Five-Factor Model suggests otherwise. More precisely, it suggests that the elastic concept of personality from everyday conversation can and should be replaced with a scientifically derived non-moral concept that encapsulates who we really are deep down. That contention is my polemical target in the remaining two sections. To forestall a straw man objection, I wish to make it clear that I have no complaints with psychologists who consider the Five-Factor Model to have only a limited, domain-specific heuristic utility in enabling us to collocate human behavior (see Hastings, 2007); or to offer particularized practical guidance in helping us, say, to find the job we are best suited for. Rather, my animadversions are directed against those theorists, such as Robert McCrae (2009; see subsequent section), who understand the dispositions of everyday selfhood exclusively in terms of personality dispositions and who think that the model in questions captures most of what is important about individual differences between selves.
Personality and the Five-Factor Model
In contemporary personality psychology, personality is viewed as a construct that underlies individual differences in customary thoughts, feelings, and behaviors (Smith & Shoda, 2009, p. 473). The aim of this construct is to pinpoint the aspects that make us who we really are in the sense of identifying us and setting us apart from other persons. Although the word “personality” has its origin in the Latin term “persona,” which refers to theatrical masks and mask-wearers, personality is not meant to capture merely people’s façades or external guises; rather it is meant to capture the internal factors that make them unique.
The search for a person’s “core traits” can be traced all the way back to the stock characters of ancient Greek tragedy, replicated in various guises through medieval times. There may not have been much science behind these sketches, except the grandmotherly and anecdotal one. That was all to change, however, in the 1930s with Gordon Allport, the father of today’s personality research. Allport’s goal was to find an academically sound way of distinguishing the traits of one human being from that of another, and he set out to achieve that goal by carefully choosing all trait-terms from the 1925 unabridged 400,000-word edition of Webster’s New International Dictionary. What he produced was a list of no fewer than 17,953 single-word descriptor terms. He then reduced that list to 4,504 trait terms by excluding items that did not focus on generalized stable modes of behavior and, more interestingly perhaps for present purposes, terms that involved moral evaluations (Allport, 1937; cf. Block, 1995, pp. 191-192).
Allport’s self-limiting ordinances set the tone for future personality research. He defined the method that should serve as the starting point of such research: the lexical or ordinary-language method, based on the hypothesis that the words which have evolved in our language to describe individual differences somehow naturally carve out and stand in proportion to the importance of the domains described; the fundamental characteristics to be identified: stable and robust traits; and the main exclusion criterion: value-ladenness. (To be sure, the “adjustment value” of traits could and should be gauged, but only through an evolutionary lens, not a moral one.) The demand for non-value-ladenness drew on the logical positivism that was spreading through American social science at the time, and also on social trends that called for more individualistic and consumer-oriented approaches to the human psyche. Social science, like all other science, had to be purged of metaphysics and morals. Only personality (“character devaluated”) had a place in this agenda—the notion of character (“personality evaluated”) had no place; rather, it belonged to social ethics (Allport, 1937, p. 52). “Personality,” Allport once remarked, “is a blessed word; it induces in both the writer and the reader a sweet sense of stability, security, and modernity” (as cited in Nicholson, 1998, p. 59). The fear of the normative thus becomes—not only in Allport’s writings, but in those of many of his contemporaries—a salient and lasting background concern of psychological research.
The search for the rigorously individuating traits of personality marched steadily on post-Allport, and continued to be driven by similar considerations. I shall assume that readers have at least a working knowledge of those developments, which—after Allport’s original traits had finally been reduced in number, through factor analysis, to five orthogonal ones—resulted post-1960 in the construction of the Five-Factor Model, later to be known popularly as the “Big Five.” To summarize, prior to the 1960s the theory of stock characters, later evolving into that of personality traits, was much like a cushion bearing the impression of the last theorist who happened to sit on it. The Five-Factor Model has changed all of that. Although not uncontested, as I describe later, it retains the support of the majority of personality psychologists, many of whom remain passionately devoted to it. One of its most vocal adherents is Robert McCrae, who considers this model to represent the most scientifically rigorous taxonomy in behavioral science—inexorably universal—and believes in it as an empirical fact, like the fact that there are seven continents. Rejecting the anti-realist suggestions that personality factors are mere cognitive constructions, he claims that they represent “real psychological structures” (McCrae, 2009; see also citations in Block, 1995, p. 187; Deary, 2009, p. 103). The Five-Factor Model has become a true scientific paradigm in the Kuhnian sense, defended by its “normal-science” practitioners with the usual vigor that tends to be associated with such paradigms. The reason for their steadfast support is simple, say its proponents: because it is true.
The famous five factors are extraversion, agreeableness, conscientiousness, neuroticism, and openness. Extraverted people are friendly, gregarious, energetic, excitement-seeking, cheerful, and optimistic. People low on this scale (namely introverts) are reserved, formal, serious-minded, and laid back. Agreeable people are soft-hearted, modest, generous, considerate, trusting, and cooperative. Disagreeable people, by contrast, are suspicious, uncooperative, arrogant, and hard-hearted. Conscientious people are well-organized and disciplined, self-efficacious, and orderly, whereas their opposites are disorganized, careless, lackadaisical, and weak-willed. Neurotic people are tense, anxious, easily depressed and angry, impulsive, and vulnerable. Those who score low on neuroticism are, in contrast, calm, relaxed, easy-going, and secure. Open people are imaginative, adventurous, and art-loving; people low on this scale are conforming, down-to-earth, prosaic, and dogmatic. But do not at least two of those factors, namely agreeableness and conscientiousness, as well as many of the listed sub-factors, belong to moral character rather than “character devaluated” (contra Allport, 1937)? The official response to that question would be that these traits are to be understood non-morally as mere conative and behavioral tendencies. The conscientious person according to the Big-Five understanding can therefore be well organized and disciplined with regard to criminal activities, and the agreeable person can be cooperative in dealings with fellow gang members. Whether or not these personality traits are—in the relevant context—morally praiseworthy (instantiations of moral character, if you like) is allegedly another and altogether separate question. Nevertheless, I revert to this suggestion at several junctures in what follows.
Big-Five theorists explain the resistance of the human self to change not with a theory about the homeostatic nature of self-beliefs, as anti-realists often do, but with an empirically based theory about the intractability of the five personality traits. McCrae (2009) highlights how “remarkably stable” those traits are over time—in adulthood, where temporal correlations range from 0.5 to 0.8, compared to the correlation of 0.3 in childhood, which is itself statistically significant (cf. Deary, 2009). Although longitudinal changes reveal some general changes over time in adulthood—with agreeableness and conscientiousness gradually increasing as one gets older but neuroticism, extraversion, and openness declining—those changes are dwarfed by the stability of these traits. The obvious explanation for both the universality and the stability of the five traits is said to be their biological foundation. Judging from studies of twins and other relatives, each of the five factors seems heritable to a certain extent: guesses range from 25 to 60%, with extraversion most genetically based and agreeableness least so. It must be admitted, however, that twin studies have tended to produce disturbingly conflicting evidence. Moreover, the most extensive longitudinal study with which I am familiar—the study of 163 men from the Harvard classes of 1939–1944 who were followed for over 45 years—reveals that only three of the five traits, namely extraversion, neuroticism, and openness, exhibited significant correlations across the 45-year interval (Soldz & Vaillant, 1999). In general, little evidence has been produced to support the belief of early Big-Five theorists that people’s personalities are set in plaster by approximately age 30. For even if it were true that temporal correlations after 30 exceed 0.5 and that 50% of personality traits are inherited, neither fact excludes the possibility of significant self-change in a given individual during adulthood (cf. Quackenbush, 2001).
But why should a person with a certain Big-Five personality profile—say, low in agreeableness and high in neuroticism—want to change? If taken at face value, those terms may seem to incorporate an evaluative dimension. Thus, in ordinary conversation, it would be considered odd to hold that it is generally better to be neurotic and disagreeable than non-neurotic and agreeable. We must recall, however, that as used in Big-Five theory, those terms are supposed to have been stripped of their evaluative dimension and are simply used to describe certain predictable behavioral traits. It remains an open question, therefore, if a person should want to become less disagreeable or less neurotic. Again, “should” must not be understood normatively here, but empirically. The measure of whether or not people should want to change is if they do want to change. People generally want to be happier; the psychologists in question take that to be an empirical rather than a conceptual truth. Therefore, considerable research has been undertaken to explore the correlations between the Big-Five traits and happiness. Indeed personality tends to be regarded as one of the strongest predictors of happiness (Vittersø, 2001). At least before the recent virtue-ethical turn of positive psychology (Peterson & Seligman, 2004), few psychologists shared philosophers’ enthusiasm for objective accounts of happiness. The most common psychological instrument to measure happiness thus continues to be a subjective one, SWB, which measures “subjective wellbeing,” understood as the combination of reported life satisfaction and pleasure (the presence of positive mood plus the absence of a negative one), resulting in a total happiness score.
Repeated empirical findings show significant correlations between extraversion and non-neuroticism, on the one hand, and SWB on the other (see, e.g., Diener, Lucas, & Oishi, 2005). Extraversion used to be considered the more salient factor here. More recent studies reveal, however, that after the neuroticism factor has been partialled out, the association between extraversion and SWB diminishes or disappears. Vittersø (2001) hypothesizes that earlier results may have been biased by the fact that they were conducted primarily in the USA, arguably the world’s most extraverted country. The small minority of psychologists who want to incorporate objective (eudaimonic) criteria into their wellbeing measures have found a similar correlation between non-neuroticism, on the one hand, and eudaimonic wellbeing, on the other, but also between openness, agreeableness, conscientiousness, and extraversion and such wellbeing, depending on which objective sub-factor of wellbeing one gauges (Schmutte & Ryff, 1997). Various explanations have been suggested for these findings: that neurotic individuals tend to be hypersensitive to negative events, and that this hypersensitivity may lead to poorer peer relations and increased victimization; that extraverted, agreeable, and conscientious individuals may find it easier to make good friends; that openness may help us to navigate the best path towards a fulfilling life; and so forth (see, e.g., Jensen-Campbell, Knack, & Rex-Lear, 2009; Moberg, 1999).
To the average philosopher, however, these may seem to be somewhat platitudinous explanations of connections that are wrongly understood as factual when they are, in fact, conceptual. I say this because extraversion is defined as involving cheerfulness; non-neuroticism is defined as the absence of painful emotions. Similarly, one of the factors gauged in SWB measures is pleasure and lack of negative emotions. There is not an empirical connection but a logical one between cheerfulness and pleasure or between painful emotion and negative emotion. Hence, it must count as a diplomatic understatement to say that the negative correlation between neuroticism and SWB is “somewhat tautological” (Ryan & Deci, 2001, p. 149). It is tautological. In any case, if one really wants to focus on variables with a proven record of standing in a clear causal relationship to SWB, there are more obvious candidates to choose from than the Big-Five factors—regular exercise or charity work, for instance—variables that are considerably more specific and easier to manipulate than are wholesale personality traits.
Attempts at establishing empirical links between the Big-Five traits and another positively valenced variable, namely educational achievement, have not fared much better. Agreeableness and extraversion seem to have negligible or inconsistent effects on student learning. Non-neuroticism and openness are slightly more relevant, but only high conscientiousness is consistently linked to good results at school (Zeidner, 2009). Nevertheless, the effects of any personality traits here are small in comparison with IQ score, which remains by far the best single predictor of academic success (Kline, 2000, pp. 126-127). Furthermore, as conscientiousness in the Big Five is defined inter alia in terms of achievement striving and self-discipline, it teeters on the brink of the tautological to claim that having those traits “predicts” academic success. At least in my vocabulary, “doing well academically” includes—rather than is empirically caused by—self-discipline and achievement striving. This is not to say that the Big Five is useless in predicting how well one does in one’s chosen field of study. Different combinations of Big-Five personality traits provide better or worse preparation for different study areas, just as they do for different occupations in the job market. Even for the latter, however, the one psychometric measure that stands above all others in the prediction of occupational success is not a Big-Five trait—not even conscientiousness, although that surely makes one a priori better qualified for any job. The best predictor is, in fact, the notorious IQ score once again (Kline, 2000, pp. 134–135).
The most common complaint lodged against the Five-Factor Model is not about its lack of clear empirical correlates, but rather its reliance on self-reports. Self-reports and peer-reports played a part in the construction of the Five-Factor Model, during which the thousands of original lexical factors were reduced, via factor analysis of their reported distribution, to a mere five. Moreover, self-reports continue to be the fundamental method by which individual personality profiles are generated. It is well known, however, that self-reports often involve hefty dollops of self-serving spin. Willful deceptions of others as well as unwillful self-deceptions pose a constant threat to outcome validity. When I myself took the Big-Five test, it struck me how transparently the specific questions tracked the specific personality traits, so that anyone who came to the test with a clear image of his or her profile would have no trouble securing—whether consciously or unconsciously—the expected result. The creators of such tests seem blissfully unaware of decades of research findings in social psychology which show how quick people are in rationalizing or confabulating their self-attributions in response to subtle environmental primes—so much so that we easily accept bogus (randomly administered) personality profiles as containing accurate and revealing truths about ourselves (Nisbett & Ross, 1980; cf. Jopling, 2008; Kristjánsson, 2010b, chap. 2). To be sure, substantial agreements between self-reports and peer-reports do mend matters somewhat (McCrae, 2009, p. 150), but the fact that not only I, but also my friends and family, consider me to be a duck is not sufficient to make me a duck.
The critique by Block (1995) is still the most trenchant one leveled at the philosophy behind the Five-Factor Model. I say “philosophy” because it is so far from being the case that this model is an “atheoretical construct,” standing above the philosophical fray (as its proponents proudly claim), that the opposite seems true. The model is anchored in a chillingly grand philosophy with its clear epistemology: a radical form of ordinary-language philosophy concerning the foundations of knowledge; its ontology: a rejection of moral properties as belonging to the fabric of scientific reality but the upholding of an adaptable and morally unencumbered realist self as part of that reality; and its methodology: a belief in factor analysis as an appropriate and sufficient basis for the creation of its main theoretical construct. But (a) the ordinary-language or lexical hypothesis is subject to all the same objections that were lodged against Austin and Ryle in the 1960s; (b) the model’s moral anti-realism suffers from the same problems that afflict such anti-realism in general; and (c) a mere statistical procedure, pulled out by sleight of hand from some black box during the process of data distillation, can hardly be treated as an automatic truth generator (cf. Block, 1995; Nicholson, 1998).
We have come a long way from Allport’s original question about the essential units of personality. The story of personality research, culminating in the Five-Factor Model, shows us that the answer to that question will always be an irredeemably philosophical one. No scientific approach or method—of whatever level of catchpenny tricks and justificatory acrobatics—absolves us of the responsibility to get that philosophy right. My fear is that the model of the Big Five is saddled with an impoverished and disabling philosophy. This is not the place to pursue all the possible philosophical misgivings about its implicit epistemology, ontology, and methodology. Let me simply note that the exploration of the history and nature of the Big Five leaves one with the eerie suspicion that there is something essentially arbitrary—if not essentially dodgy—about it. Its supposedly salient empirical correlates turn out to be non-existent, negligible, or tautological. It remains likely that knowing your Big-Five profile may aid you in finding a suitable job or even a suitable partner in a dating agency. But so, presumably, would knowledge of many other factors that Big-Five theorists have decided to ignore. Take being religious or non-religious, deceptive or non-deceptive, sexy or non-sexy, thrifty or non-thrifty, masculine or non-masculine, witty or non-witty. In other words, there is considerable variance in human action and reaction that is not accounted for by the Big Five (cf. Paunonen & Jackson, 2000). To what extent such variance can be considered self-constitutive is, however, another question.
The Five-Factor Model, morality, and self
The prototypical examples of self-transformations that one reads about in novels or biographies are moral and/or religious conversions. A person then changes adherence to a thick evaluative framework and develops new character states (Aristotelian hexeis): dispositions to act, react, and comport in previously unfamiliar ways—witness again Saul’s mutation into Paul. Big-Five theory does not consider adherence to or expression of evaluative frameworks to define who one really is in the everyday (or, if you like, psychological), as distinct from the metaphysical, sense. Rather, what makes you who you are and distinguishes you from other individuals is your unique combination of the Big-Five traits. The rigidity of those traits—caused partly by their heritability and partly by their reinforcement or suppression through childhood—allegedly explains the adult’s relative resistance to self-change.
Despite the alleged non-moral nature of those traits, considerable energy has been spent charting their empirical correlations to moral characteristics. In a study of the personality profiles of acknowledged moral exemplars from everyday life, such exemplars scored highest in agreeableness and conscientiousness; and the author concludes that those two are the classic dimensions of moral character (Walker, 1999). The remaining three traits are typically seen as more ambiguous morally. Whereas high openness does predict the moral virtues of sensitivity and tolerance, for instance, low levels of openness may predict other virtues such as temperance, stoicism, and community solidarity (Miller, 2007, p. 106). By contrast, if one understands “moral characteristics” less in a virtue-ethical sense and more in a Kantian/Kohlbergian sense, as a high level of moral reasoning, openness is most closely associated with it (McAdams, 2009, p. 15). Moberg ruminates over the expected profiles corresponding to the four cardinal virtues of temperance, courage, wisdom, and justice. He concludes that all four virtues would register low on neuroticism. Additionally, temperance would most likely register somewhere in the middle with regard to extraversion; courage would be high on extraversion; justice high on agreeableness and conscientiousness; and wisdom high on openness (Moberg, 1999, p. 251).
The very idea that moral characteristics are not parts of the traits that individuate us but are, at best, empirically correlated to them adds ammunition to the suspicion expressed at the end of the previous section that there is something essentially arbitrary about the Big-Five project. By excluding moral properties from their potential list of traits that comprise what I have called our everyday self, personality psychologists have already risked obscuring and diluting what is, in fact, central to us. It may be retorted that this exclusion is now a thing of the past, as, with the arrival of the Five-Factor Model, moral character turned up again in the guise of agreeableness and conscientiousness (see, e.g., Cawley, Martin, & Johnson, 2000, p. 1009). Moreover, attempts are underway in personality-psychology circles to insert a new factor into the model, and a “pure” moral one at that—namely honesty-humility—so that it becomes a six-factor model (Ashton & Lee, 2005). I submit, however, that the problem with the Big Five lies much deeper than that which can be cured with simple facelifts. It was not really the case that moral character returned to personality profiling through agreeableness and conscientiousness. For as those factors are defined, they are supposed to be operationalizable independent of moral evaluations. The orderly obsessive-compulsive individual or the self-disciplined criminal can score as high on this factor as the morally conscientious Mother Teresa does (cf. Prinz, 2009, p. 121). If we think, however, of conscientiousness as a trait that defines us in a way that is relevant to our everyday engagements with other people—that sets us apart in a non-arbitrary fashion as a person among persons—the role of conscientiousness seems to me to be exhausted by its moral import. The same would apply, mutatis mutandis, to agreeableness—notably understood by Aristotle as a comprehensive moral virtue (for a discussion, see Kristjánsson, 2007, chap. 10). Looking for “empirical correlates” of conscientiousness or agreeableness in specifically moral characteristics is, at best, grasping the wrong end of the stick—at worst, falling prey to a tautology. In so far as those factors matter for who we are, they are moral characteristics.
David Funder (2009) hits the nail on the head when he says that there is a crucial question facing research on within-person variance: can it be shown to be interesting or useful? “It is one thing to identify within-person patterns that have some degree of stability,” he observes; it is “quite another to show that these patterns are psychologically meaningful” (p. 122)—which, of course, is an instantiation of the old dictum that an instrument’s reliability does not guarantee its validity. To take a parallel example from the field of education, considerable emphasis used to be placed on the difference between a teacher’s classroom style (which was supposed to reflect non-moral personality traits) and a teacher’s classroom manner (meant to capture what was moral in relation to the teacher’s conduct). A closer look revealed, however, that the two could not be separated for any relevant purposes. In so far as a teacher’s “style” matters in the classroom, it is because of its moral implications: its impact on student wellbeing (see Kristjánsson, 2007, pp. 152–155).
I suggested earlier that the Big-Five Model could at least help us in finding partners in a dating agency. Even that may not be true, however. Research shows that people regard moral traits such as honesty and trustworthiness higher than Big-Five traits when looking for a potential spouse (Baumeister & Exline, 1999; cf. Miller, 2007). If the Big-Five Model suffers from arbitrariness, as I have suggested, it is because it overlooks the fact that for the everyday sense of selfhood, moral characteristics cannot be ignored with impunity. I have not argued directly in this article for the radical alternative: that who I am in the practical everyday sense is exclusively who I am with regard to the moral evaluation of me (Kristjánsson, 2010b). But the weaknesses in the Big-Five Model do, arguably, point in the direction of an alternative model such as mine. To be sure, revealing those weaknesses does not suffice as an argument by elimination, for there are various theorists—for instance within social psychology—who oppose the Big-Five Model while upholding non-moral accounts of selfhood. I have, however, taken precisely such accounts to task in a previous work (Kristjánsson, 2010b, esp. chap. 3).
An objector might complain at this point that I have been offering readers a gambit that they can easily reject. Why assume that the everyday self must necessarily be specified in either moral or non-moral terms, and snub the Big-Five Model because it (mistakenly) only offers the latter option. The objector could point out that even the hard-liner McCrae has in a recent article modified his stance to accommodate the notion that the Big-Five Model does not exhaust the range of relevant individual differences, as “comprehensive does not mean exhaustive” (McCrae, 2010, p. 58). My quick response to that would be to say that although McCrae now claims that the model is not, and was never meant to be, an exhaustive catalog of individual differences, he still insists that it purports to offer a comprehensive taxonomy of individual dispositions. That is, however, precisely a claim that my arguments above have aimed at confuting. For example, when I have spoken of certain morally imbued emotions as constitutive of the everyday self, I have been referring to emotional dispositions rather than to individual episodic emotions.
The objector might continue, however, by noting that a more moderate advocate of the Big-Five Model could easily make room for moral dispositions in an integrative concept of everyday selfhood, where the Big-Five traits only form one aspect of it. On such a complementary account, “who I truly am” in an everyday sense would in part be my personality in terms of the Big-Five traits, in part who I am with respect to the moral evaluation of me, and in part who I am in terms of other non-moral considerations such as my interests and allegiances: say, my deep love of Chinese food and my impassioned support for football team A. In response, let it be said that I did argue earlier that the role of the Big-Five traits in the everyday self is exhausted by their moral import. As far as deep commitments are concerned, I suggest that the same applies. Would we consider persons to have undergone self-change if they had started to love Thai instead of Chinese cuisine or changed their support from football team A to team B? Hardly. We would, however, pause to think what had happened to a lover of Chinese food who lost interest in food altogether, or to a football fan who completely abandoned the game. Why is that? Might it not be because we expect people to have strong commitments, and that we value such commitments morally, irrespective of their specific content? Is the type (“a committed person”) not what matters here for selfhood rather than the token (“a person committed to football team A”)? Would we not primarily be worried that the persons in the latter set of examples had lost their zest for life, thus putting their selfhood in jeopardy? Or could we consider commitments to be constitutive of the everyday self, simply in terms of their non-morally evaluable (and even non-reason-responsive) content? Those are, I believe, empirical questions that call for empirical answers.
To help us answer such questions empirically—and to return to the allegedly non-moral nature of the Big-Five Model—it is instructive to consider briefly recent positive psychology’s take on personality and the virtues (for a critical review, see Kristjánsson, 2010a). The chief proponents of positive psychology, Christopher Peterson and Martin Seligman (2004), share many of the philosophical assumptions of Big-Five theory: they describe themselves, for example, as circumspect moralists, initially worried that their much-touted virtue project—the “social science equivalent of virtue ethics” (p. 89)—was doomed from the start because of its value-ladenness. After doing their empirical research on the strengths that set us apart and define who we are, they were left “somewhat reluctantly,” however, with the conclusion that such strengths differ from mere talents and abilities precisely because of their moral nature (p. 20). Yet, they explain that conclusion as being “descriptive of what is ubiquitous” rather than “normative” (p. 51). Positive psychologists acknowledge the scientific import of Big-Five research, but they think that focusing on the different competences inherent in the Big-Five traits is not enough if we are to grasp fully the way in which those competences become human strengths. In order to understand that process, we need to focus on moral competence, instantiated through virtues and strengths of character, which guides all the other competences in constructive ways. So when comparing their list of allegedly universal virtues (wisdom/knowledge, courage, humanity, justice, temperance, and transcendence) with the Big-Five traits, positive psychologists conclude that after controlling for the Big-Five indices, distinctive features remain, suggesting that the virtue scale reflects “something more than the Big Five measures—specifically, the moral flavor of the character strengths” (Park & Peterson, 2006, p. 903). This fact also suggests, in their view, the possibility of a deep theory about human moral nature, couched in evolutionary terms (Peterson & Seligman, 2004, p. 51). The conclusion of our “reluctant moralists” that human character strengths or virtues contribute to fulfillments that define the good life for the individual in ways that Big-Five traits cannot capture—and that the Big-Five tradition is misguided not only morally, therefore, but also psychologically (cf. Peterson & Seligman, 2004, p. 68)—adds empirical backbone to my skepticism about any sort of a non-moral take on our everyday self.
Let me conclude by raising two additional points, one psychological and the other moral. In my recent book (Kristjánsson, 2010b), I argue that our everyday self essentially comprises a particular set of morally imbued background emotions. Although I have provided no further evidence for that view here, the reluctance of Big-Five theorists to engage directly with the emotions gives us further reason to view their self-theory with a beady eye. Cognitions about the self are inherently affect-laden (see Smith & Shoda, 2009, p. 478), and if the Big-Five personality traits are to be understood as traits that lay the foundation of our independent selfhoods, they must be traits of, so to speak, our “heart-minds.” Learning how to be a distinct self is learning how to feel things in a distinctive way. As Reisenzein and Weber (2009, p. 60) freely acknowledge, however, in-depth investigation of the emotions from a personality-psychology perspective has begun only recently; it was not on the agenda of the creators of the Five-Factor Model, who operated with a much more restricted notion of the cognitive. Reisenzein and Weber suggest that Big-Five research needs to be, and can be, redefined so as to focus specifically on the emotionality of the relevant factors. After all, neuroticism may best be understood as a broad emotional disposition to experience so-called negative (or painful) emotions; extraversion and agreeableness will presumably comprise dispositions towards positive affect and social emotions; openness to experience seems to be related to a disposition for experiencing aesthetic emotions. These belated suggestions notwithstanding, the fact that the emotionality of the five factors has been largely neglected in the literature must count as a psychological shortcoming.
The moral point I want to make is this: Big-Five theorists pride themselves on their ability to give a more plausible account than anyone else can of the intractability of radical self-change. Their original idea was that people’s personalities are set in plaster from about the age of 30; yet recently, they have been adopting a more embracing view of the dynamic processes by which we continue to constitute ourselves, even in adulthood (McCrae, 2009). In any case, it is clear that at least in childhood and adolescence our Pullmanite daemons are largely shape-changing. Here current Big-Five theory fortunately seems to accept something more like an Aristotelian moral developmental theory, according to which young people are not merely small adults, but rather human beings in the making, with specific traits/virtues of their own (for an elaboration, see Kristjánsson, 2007; cf. positive psychology’s take on the virtues of the young—gratitude, humor, and love—in Park & Peterson, 2006, p. 898; recall also Erikson’s, 1963, psychosocial stages, each with its own corresponding virtues). Moral development and self-development are then seen as going hand in hand. Rather than trying to help the shoots grow faster by pulling them upwards—which, as any gardener can tell you, will simply dislodge their roots and make them wither away—we water the plants, but otherwise allow their developmental course to run naturally. As situations alter and individuals move from one developmental level to the next, their projects change and their respective virtues shift.
Such a view seems to call for an essentially diachronic approach to morality. Endemic to most traditional moral philosophy is, however, a purely synchronic—and hence static—view of morality and the moral virtues: the view of the old cow who thinks she was never a calf. The typical solution to a moral problem is this: reflective (read: adult) agents, reasonably versed in moral theory, try to bring their intuitions about the particularities of the problem into line with the generalities of (their preferred) moral theory, and to reach, through a process of dynamic compromises between the two, some sort of “reflective equilibrium.” As Cottingham (2009) notes, there is nothing wrong with this process as far as it goes. Nevertheless, he correctly complains, there is something missing—namely the way an individual’s moral outlook shifts and develops over time.
The fact that my 14-year-old son approaches a problem differently than I do does not mean that his solution will be inferior to mine had I stepped into his shoes; rather, we could say that if he solved the problem in exactly the same way I would solve it, the solution would be inferior, to the extent that it failed to represent the fulfillment of something central to him: his adolescent moral self. This drama of moral change and renewal does have its philosophical dramaturgs: Kierkegaard, Hegel, Nietzsche, Charles Taylor, and Alasdair MacIntyre, to name but a few. The fact that theirs are not the dominant voices in contemporary moral philosophy, however, may suggest that philosophers would be well advised to seek guidance from their colleagues in psychology (if not necessarily Big-Five theorists!) on how to refertilize a ground that has, for some reason, lost its philosophical lushness.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Kristján Kristjánsson received his Ph.D. at the University of St. Andrews, Scotland. He is currently Professor of Philosophy of Education at the University of Iceland. He has published extensively in the fields of educational philosophy, philosophy of mind, and moral philosophy. His publications include Justifying Emotions (Routledge, 2002) and The Self and Its Emotions (Cambridge University Press, 2010). Address: School of Education, University of Iceland, IS-105 Reykjavík, Iceland. Email: