Proulz Chandler

8/8/2019 Proulz Chandler

1/27

Paper

Human Development 2009;52:261286DOI: 10.1159/000233259

Jekyll and Hyde and Me: Age-GradedDifferences in Conceptions of Self-Unity

Travis Proulxa Michael J. Chandlerb

aUniversity of California, Santa Barbara, Calif., USA; bUniversity of British Columbia,

Vancouver, B.C., Canada

Key Words

Age-graded differences Conceptions of self-unity Hierarchical self

Multiplicitous self Singular self

Abstract

This research details the changing ways in which young people of different ages

differently warrant the conviction that, notwithstanding evidence of good and bad be-

haviours, selves can be understood as unified across the various roles and contexts that

they occupy. Canadian adolescents and young adults were asked to explain the appar-

ent disunity of self implied by good and bad behaviours manifested by the fictionalcharacter Dr. Jekyll and Mr. Hyde, themselves and familiar others. Age-graded differ-

ences were observed, where young people described themselves as increasingly mul-

tiplicitous and context dependent as they grew older. This developmental trajectory is

understood to represent an emerging desire to imagine ones good behaviours as in-

ternally motivated, and ones bad behaviours as externally provoked.

Copyright 2009 S. Karger AG, Basel

man will be ultimately known for amere polity of multifarious, incongruous,and independent denizens. Henry Jekyll

The Problem of Self-Unity

Here is how it (i.e., our standard-issue Euro-American, Judeo-Christian, Post-Enlightenment cultural narrative) ordinarily goes. Nobody likes a two-face. Threefaces and up are straight out. Rather, we are meant to be ofone face, inside and out.

Travis ProulxDepartment of PsychologyUniversity of California, Santa BarbaraSanta Barbara , CA 93106 (USA)E-Mail [email protected]

2009 S. Karger AG, Basel0018716X/09/05250261$26.00/0

Accessible online at:www.karger.com/hde

Fax +41 61 306 12 34E-Mail [email protected]


2/27

Human Development2009;52:261286

262 Proulx /Chandler

That is, on pain of things otherwise unravelling, we are each meant to be a coherentself, not only continuous in time, but also unified across whatever contexts in whichwe find ourselves embedded. If we lend you a fiver while out on the town, we do notwant to hear, at home, about how you are no longer the same person who borrowedthe money, and so do not owe us a dime. Or, if, to our face, you are acting like we arethe best of friends, we do not want you whispering our weaknesses into the ears ofour enemies when our back is turned. And we certainly do not ever want to standbefore Kierkegaards Judge William and have him declare:

Do you not know that there comes a midnight hour when everyone has to throw off hismask? or can you think of anything more frightful than it might end with your naturebeing resolved into a multiplicity, that you really might become many, become, like thoseunhappy demoniacs, a legion, and thus you would have lost the inmost and holiest thingof all in man, the unifying power of self. [Kierkegaard, 1843/2000, p. 164]

At least, that is how, for a long time, most of us have imagined that things necessar-

ily work.Of late, however, this old and comfortable consensus appears to have begun un-ravelling. The midnight hour has arrived, the masks have been thrown down, andpinned in the glare of certain harsh contemporary postmodern lights we are all re-

vealed as belonging to just that legion of unhappy demoniacs whose holiest thing isheld up as being the mere product of a group delusion meant to obscure our truemultiplicitious nature [e.g., Deleuze, 1994; Derrida, 1978]. On this post-everythingaccount, we, along with Kierkegaard [1843/2000] and other assorted holdouts forcoherence, are simply standing ankle-deep in the same Enlightenment trickle-down,and we should all be looking to lose the Locke and lighten up. Maybe, we are invitedto imagine, those existentialists were a bit too strident. Maybe, all of their talk ofau-thenticity and choosing selves and creating selves and beingand time was just a bunchof hokum generated by a group of surly navel-gazers, men who wrote of freedom

and possibility from sickrooms and shabby offices. Fortunately, for the purposes ofthis research, there is no obligation to wade too deeply into the perpetually murkywaters that surround the existential question of whether people are (really and truly)bounded subjects, unified across roles. Nor is it necessary to attempt to resolve themetaphysical uncertainties that surround the matter of whether or not we can everbe properly considered bounded subjects, unif ied across roles. All of that, we, as psy-chologists, are generally happy to forfeit in favour of the more tractable empiricalquestions of whether, left to their own devices, realpeople (i.e., philosophic ama-teurs), of differing ages, actually do maintain beliefs about their own and otherssynchronic unity, and, if they do, how they go about attempting to warrant such be-liefs in the face of all of the evidence seemingly to the contrary.

Before coming to the details of how we went about addressing these empiricalquestions, and all in search of some alternative to flying completely blind, we rea-

soned that we might get some purchase on the various solution strategies that youngpersons actually do employ in thinking about the problem of self-unity by examin-ing what their more philosophically informed elders have traditionally said, and arecurrently saying, about the same subject.


3/27

Age-Graded Differences in Conceptions ofSelf-Unity

263Human Development2009;52:261286

The Need for Self-Unity

Notwithstanding the heavy counterweight of enduring modernism, there has

arisen, primarily in the course of the 20th century, a contrapuntal chorus of detonatevoices, all of which have been quick to intone a seemingly endless chorus of person-al disunities, big and small, all of which are offered up as proof that truly unifiedselves selves of the sort envisioned by Enlightenment theorists [e.g., Hobbes,1660/1973; Locke, 1632/1996] have never and will never exist [Bakhtin, 1986].Many such more-postmodern-than-thou cultural critics have made the still moreprovocative claim that, not only are we not unified, but we should not have ever re-ally wanted to be unified to begin with [Foucault, 1986]. By design, such anti-mod-ernist claims are provocatively aimed to fly directly into the face of millennia ofWestern thought that proclaims just the opposite the necessity of obeying whatFlanagan [1996] called the one self per customer rule (p. 52).

All such postmodern prospects notwithstanding, it is important to underscorethe fact that a great many of Western societys standard institutions and cultura l prac-

tices (and so its folk psychologies) are stil l quintessentially modern. Consequently, itis not difficult to imagine why adherence to Flanagans [1996] one-self-to-a-customerrule has regularly been so consistently judged to be a good thing. It is hard to imagine,for example, how any society could function if it did not operate from the assump-tion true or false that each of us is fundamentally the same person, not only acrosstime, but across roles and contexts. In short, Enlightenment philosophers were notpreoccupied with the subject for no reason; they had Leviathans and Second Treatiseof Governments to write, and you cannot have laws without stable, bounded subjectsto apply them to. To be a citizen within a nation, an economic subject, a legal andmoral subject who can sensibly be held accountable for her or his actions, we seem torequire some kind of stable individual, some kind of selfhood that does not undergoa complete transformation from one context to the next. If, on this standard account,they are to have any sort of followable meaning, then selves must persist, not only

across changing time, but across shifting contexts; their essence is that, on pain of be-ing something constitutively different entirely, they simply must persist.In addition to all such pragmatic rationale for presumptive self-unity, there is

also a long-standing Western belief that being synchronical ly disunified simplyfeelsbad. This general contention is shared by existentialists, with Kierkegaard [1843/2000]making what is an especially strong case for the affective necessity of a unified self.You could try, of course, to get by without one, but, Kierkegaard argued, the anxietycould kill you. Later, Camus [1955] popularized the related notion of a nostalgia forunity (p. 4) as a general psychological need, one that, if left unsatisfied, necessarilyproduces a great deal of suffering. Psychologists, too, beginning with James[1892/1963] and Freud [1930/1991], have also seemed eager to describe both the needsand consequences thought to surround the quest for both horizontal self-unity (i.e.,self-coherence across external, situational contexts) and vertical self-unity (i.e., co-

herence within the individual psyche, between the conscious and unconsciousminds). Later, Erikson [1968] gave us the expression identity crisis, and Epstein [1981]went a long way in militating for an innately human unity principle that, in additionto not being entirely dissimilar from Camuss nostalgia for unity, is said, not only todrive us to establish a unity within our psyche, but to punish us with misery if wesense that this unity has not been achieved.


4/27



So How Do Young People (Philosophic Novices) Actually Think about

Self-Unity?

To date, there has been only limited psychological research into the develop-mental course by means of which we come to whatever understanding we have aboutourselves as synchronically coherent persons. In attempting to impose some orderon this scarce and scattered literature, it is a matter of special importance to keepdistinct and well apart work that focuses on matters of diachronic continuity, on theone hand, and, on the other hand, those separate efforts that feature, as is done here,more time-limited, cross-sectional matters of synchronic self-unity.

Chandler and his colleagues [e.g., Chandler, Lalonde, Sokol & Hallett, 2003;Chandler & Proulx, 2006; Chandler, Sokol, & Hallett, 2001] have focused their ear-lier work almost exclusively on the temporally vectored problem of diachronic self-continuity, reporting both age-graded and cross-cultural variations in how youngpeople work to resolve discrepancies between conceptions of their past and presentselves. This work also demonstrated that failure to resolve such discontinuities can

be linked to an increased risk for a variety of adaptational failures, including sui-cidal behaviours. While this research points to a felt need to maintain self-continuityover time, and details some of the possible costs of failing to achieve a sense of dia-chronic coherence, it is orthogonal to, and does not address the present matter ofsynchronic self-unity, or the contradictions that sometimes arise when we compareour behaviours from one social context to the next.

Among those who have concerned themselves with sel f-unity, the work of Erik-son [1950, 1968], along with that of Marcia [1966, 1980], is perhaps the most familiar.Erikson, in particular, characterized self-unity as the ability to bracket together allof ones diverse but simultaneously present features, much as did Kelly [1955] andRogers [1959], who similarly stressed the importance of achieving some cross-sec-tional sense of internal consistency, or unity of the self. Within the more contempo-rary clinical and developmental literatures, Grotevant [1993], Grotevant, McRoy,

Elde, & Fravel [1994], Harter [1986], Harter and Monsour [1992], Fischer [1980], andFischer and Aboud [1994], among others [Aboud, 1979; Broughton, 1978; Leahy &Kogan, 1989; Selman, 1980], to name a few, have all devoted attention to the problemof identifying age-graded changes in the ways that young persons attempt to linkconflicting aspects of the self into some increasingly organized understanding.

Within the identity development literature per se, empirical studies focusing onyoung peoples efforts to achieve a measure of synchronic unity are relatively thin onthe ground. A rare and notable example of this work is Harter and Monsours [1992]Developmental analysis of conflict caused by opposing attributes in the adolescentself-portrait. As did Chandler et al. [2003], Harter and Monsour focused on adoles-cence the developmental juncture at which the need for both a synchronically uni-fied and diachronically continuous personal identity is widely imagined to be feltmost keenly and is arguably the most difficult to achieve. On their account, adoles-

cence is a time when we often experience ourselves as numerous potentially contra-dictory selves clamouring for expression (p. 258). Swimming against all but themost postmodern of philosophical currents, Harter and Monsour also presentedevidence said to indicate that the usual solution to this problem is not, as might havebeen expected, to struggle against mounting odds by working to achieve self-unityat whatever cost, but, rather, to adopt a more differentiated sense of self as we grow


5/27



older a way that is said to be associated with the varying social roles that theadolescent must come to adopt (p. 256). Evidently, rather than seeking to achieve awholly consistent sense of self-unity across roles, what adolescents typically do, ac-

cording to Harter and Monsour, is deal with the negative effect associated with suchdisunity by turning a potential vice into a virtue, by reasoning that it is both under-standable or desirable to act differently in different social situations (p. 256).

Although Harter and Monsours [1992] programme of research comes reason-ably close to a historically well-grounded operationalization of self-unity, two cen-tral issues remain unsettled by their approach. First, it is important to distinguishbetween: (a) the potentially oppositional attributes and behaviours manifested by aself across contexts, and, alternatively, (b) the prospect that there are, in some sense,different selves inhabiting the same body. While Harter and Monsours work elab-orated the manner in which we may come to reconcile our often discordant but con-text-dependent ways of behaving, none of this directly addressed what is arguably amore egregious form of disunity the manifestation of desires or intentions thatmanifestlycontradictone another. The often noted fact that adolescents commonly

behave differently at home and at school perhaps deserves a footnote, but who canforget the patient who spends her days knitting up sweaters, but her nights undoingher own knitting [Wilkes, 1993]? Second, the presence ofcontradictory desires andintentions is, we will argue, more egregious still when such desires and intentionsprovide the impetus for both right and wrong conduct. After all, much of the handwringing practiced by centuries of Enlightenment thinkers was owed, not to someneed for behavioural consistency, but rather the need to have stable, unitary heartsand minds, as well as bodies, to which laws can be applied and that can be held ac-countable. Whatever Harter and Monsour might believe about the normative char-acter of becoming a two-face, what parents do not want to hear is that you are notresponsible for joyriding in the neighbours pickup because that was your otherself,the one who likes to raise hell with his friends.

A Methodology in the Making

Simply asking anyone especial ly young persons to comment upon their de-gree of self-unity will hardly do; even thoughtful adults may not easily understandthe question, or may not have declarative access to the knowledge that they need toexplicitly frame a coherent account of how they ordinarily resolve the riddle of eventheir own self-unity. Through a certain amount of trial and error, the procedure fi-nally settled upon here as a way of getting at such matters was to have young peopleconsider and comment upon examples of the coincidence of what they took to betheir own and others otherwise honourable dealings, on the one hand, and behav-iours that run counter to whatever informal, common sense system of personallyaccepted customs of conduct and right living that ordinarily govern their own be-

haviour, on the other.Having recorded whatever instance of right and wrong conduct these respon-

dents personally confessed to, and whatever similarly honourable and dishonourableactions they identified in others, participants were asked two key questions. The firstwas, Can you (or some other named individual) still be considered one in the sameperson, despite the good and bad behaviours you/they have alternatively engaged in?


6/27



The second follow-up question pressed all of those who claimed for personalunityto explain How can you (or the other individual in question) behave so differently,and stil l be considered one and the same person? As a way of beginning with a com-

mon case, participants were first asked about the possible disunity, not of them-selves, but about a hypothetical and purposefully extreme case drawn from familiarliterature (about The Strange Case of Dr. Jekyll and Mr. Hyde), before also being askedabout their own self-unity, and that of close acquaintances.

Jekyll and Hydes Selfhood

Robert Louis Stevensons The strange case of Dr. Jekyll and Mr. Hyde [1886/2003] and, more particularly, a stripped-down Classic Comic Book version of thisclassic story was chosen to provide such a common starting point. This story isdeeply imbedded within our popular culture, presumably because it presents suchan admittedly extreme example of self-disunity, and because (unlike many philo-

sophical hypotheticals) it focuses on those aspects of self-disunity that are the mostcompelling and immediately problematic. That is, as is most often the case in reallife, Jekyll and Hydes chief disunity lies not only within his/their seemingly contra-dictory intentions, but in the degree to which these intentions possess a normativeor ethicalvalence. Sometimes Jekyll wants to dogoodthings, and at other times, andin other contexts, Hyde wants to do badthings. Although The Strange Case of Dr.

Jekyll and Mr. Hyde is certainly an exaggerated example of right and wrong living,and of seeming disunity, it was also thought to be sufficiently dramatic and straight-forward to allow most young people to both recognize and engage the issue of per-sonal unity.

Personal Selfhood

By contrast, our primary interest lay in how young people might react to suchpotential disunities, not simply within Jekyll and Hyde, but within their own lives.1Since one cannot pull out a dossier on each possible participant and confront themwith their own ethical contradictions, participants were instead asked to offer uptheir own examples. Fortunately, and perhaps against all odds, even younger adoles-cents were, as it turned out, quick to provide examples of their own normatively con-flicted behaviours. Although, if actually carried out by one and the same person, therightness of Dr. Jekylls behaviour and the wrongness of that of Mr. Hyde would eas-ily qualify as meta-ethically contrary, we had no expectation that our research par-ticipants would confess to such extreme instances of right or wrong conduct, or eventhat the oppositional behaviours that they volunteered would necessarily count asfully fledged instances of moral and immoral behaviours. Rather, it was enough

1 In particular, while many, if not most, of the right or wrong actions offered up by the respondentsof these studies involved what have been called common sense moral matters [Gert, 2004], only somewould qualify as involving what Turiel [2002] and Nucci [2004] would regard as moral, as opposed toconventional, or personalprescriptive claims. Consequently, although the terms moral and moralityare occasionally used throughout this account, they are intended here only to broadly refer to good oradmirable, as opposed to bad or dishonourable, conduct.


7/27



for our present purposes if these respondents themselves saw a measure of rightnessand wrongness in the incidents they reported.

With these self-generated examples in hand, the structured interview protocol

employed regarding both the Jekyl l and Hyde story, and their own self-reports, wereas follows. That is, despite having described themselves as sometimes acting like agood person and at others like a bad person, participants were asked whether they,nonetheless, consider themselves to be one in the same person. And if so, how? Fol-lowing a similar line of inquiry, participants were also asked about potentially dis-cordant behaviours shown by friends or family members.

Two Orienting Questions

Although the three studies presented here are primarily exploratory in nature,the first two (studies 1a and 1b) were, nevertheless, organized around two theory-driven questions.

First, it seemed important to address the question of whether the ways in whichadolescents attempt to understand meta-ethically conflicted selves do or do notchange over the course of their teenage years. Harters [1986] and Harter and Mon-sours [1992] work traced a developmental arc according to which older adolescentsallowed contradictorybehaviouralattributes to remain a part of their self-concep-tion more frequently than was the case for younger adolescents a finding inter-preted as suggesting a more differentiated self-conception manifesting with age. Incontrast to this earlier work, the focus of the present research is on the resolution ofapparently contradictory intentions, rather than behaviours or attributes. It re-mained an open question as to whether strategies for resolving non-normatively va-lenced attributes implicate cognitive processes and capacities that are different fromthose that underlie efforts to maintain or restore a sense of self-unity in the face ofdiscordant good and bad behaviours.

Second, it was also judged to be of interest whether adolescents tend to employthe same or different notions of self-unity in considering the honourable and dishon-ourable behaviours of Jekyll and Hyde, themselves, and familiar others.

Study 1a: Adolescents

Method

ParticipantsThe participants consisted of 70 high school students (grades 812) from an urban, North

American, private school. All spoke English as their f irst language. Of these students, 40 werefemale and 30 were male. Ages ranged from 13 to 18 years, with 14 13-year-olds, 19 14-year-olds,11 15-year-olds, 12 16-year-olds, 11 17-year-olds, and 3 18-year-olds (M = 14.94, SD = 1.51

years). Participants were paid CAD 15 to participate in a half-hour interview.

Materials and ProcedureStimulus materials consisted of a classic comic book version of The Strange Case of Dr.

Jekyll and Mr. Hyde [1886/2003], one that was edited so that it would embody the key concep-tual element of the story that participants were meant to consider: namely, the question ofwhether Jekyll and Hyde can be considered one in the same person in the face of the good and


8/27



bad behaviours on display. The focus of subsequent questioning was on how this evident dis-cordance could be understood. This revised version of the comic book depicted no dramaticphysical transformations, contained no magic potions and was as vague as possible regardingJekyll and Hydes motivations, thereby removing any easy answers that could account for the

alternating good and bad behaviours in evidence.Trained interviewers questioned participants using a standardized interview protocol. Par-

ticipants were asked to read the comic book The Strange Case of Dr. Jekyll and Mr. Hyde, and torecount the story as a way of determining whether they were clear about the characters and events.Participants were then asked to separately describe Dr. Jekyll, as he behaved during the day, andMr. Hyde, as he behaved at night. It was then put to participants: Can Jekyll and Hyde be consid-ered one and the same person, given his good and bad behaviours on display? If they answeredin the aff irmative, they were then asked to provide reasons in support of this conclusion.

Participants were then asked if it was ever the case that they themselves sometimes be-haved like a bad person, and other times like a good person. If they consented to this possibil-ity a prospect universally agreed to they were asked to provide examples. As was the casefor Jekyll and Hyde, participants were then asked whether, in spite of these opposing behav-iours, they nonetheless considered themselves to be one and the same person. If they answeredin the affirmative, they were again asked to warrant this assertion in detail.

Finally, participants were asked to give examples of good and bad behaviours displayed by

their friends or family members, and whether or not they thought these individuals could stillbe considered one in the same person, and if so, to warrant this conclusion.

While all interviews began with questions concerning Jekyll and Hydes self-unity, ques-tions concerning personal self-unity and the self-unity of others were counterbalanced acrossinterviews.

Coding and TypologyEach interview transcript was reviewed by two coders, and participant responses corre-

sponding to their own, Jekyll and Hydes and others self-unity were coded into one or anotherof the categories detailed below with a high degree of reliability. Over the course of each half-hour interview, allparticipants provided at least one account of their own or others good andbad behaviours; in fact, many participants presented more than one potentially codable account,often with various false star ts and changed minds as the interviews progressed. While the num-ber and nature of these false starts could not themselves be reliably coded, it was also t he casethat all of the part icipants eventually settled on a preferred, elaborately detailed conception ofselfhood that they spent the balance of the interview discussing, and that could be reliablycoded. Statements such as now that I think about it, or forget what I said before, regularlymade it clear what they had been working themselves toward.

The following typology of self-unity warranting strategies constitutes what are argued tobe the naturalcategories in which young peoples responses came to rest over the course of theirself-reports. While some philosophical precedence is offered up for the labelling of the catego-ries eventually settled on, it should be noted that these categories were not generated a priori inorder to meet some pre-existing framework of understanding. As such, the typology outlinedbelow came to include four distinct response types. Importantly, these are not to be understoodas developmental levels or stages of increasing conceptual sophistication. Rather, they repre-sent what is argued to be a spectrum of self-unity warranting strategies that implicate two in-terdependent dimensions.

The first of these dimensions concerns the number (i.e., one versus more) of impulses orvolitions that were seen as underlying the apparently contradictory behaviours manifested in

their responses. As will be shown, many (although not all) young people imagined that differ-ent impulses or volitions separately underpinned the good and bad behaviours they described,while other participants imagined that a single volition was dif ferently manifested in the con-texts of the different behaviours on display. Protocols were, then, dichotomously scored as in-

volving only one or more than one volitional explanation.The secondof these dimensions concerns the degree to which these impulses or volitions

were described as being generallyactive or reactive in nature. Many young people described


9/27



themselves as generallyreacting to circumstance responding to external or internal forcesbeyond their control that evoked their good and bad behaviours. Conversely, other participantsimagined themselves or others as acting on circumstance volitionally choosing to behave bothwell and badly, all in the service of some common goal. Still other responses fell somewhere inbetween these two extremes, with many young people imagining themselves as harbouring

both volitions that acted on circumstance, and impulses owed to circumstance.As shown in figure 1, these two interrelated dimensions combined to form four self-unity

categories that lie along a continuum. These four categories are described here as expressive ofa multiplicitous self, hierarchical self, and singular self. The remaining category no self-unity comprised the responses of those that denied the presence of any self-unity whatso-ever. A brief account of each of these scoring categories follows.

No Self-Unity. Responses scored in this litt le used category evidenced no attempt to assertthat the individual in question possessed any semblance of self-unity. Most often, a mind/bodydistinction was evoked, whereby the individual in question could be considered to be the sameperson only insofar as they possessed the same physical body, but qualified as two or more dif-ferent people insofar as that body housed entirely different minds. Participants usually madereference to multiple personalities, or claimed that the individua l was schizophrenic, suggest-ing that some form of mental breakdown was responsible for their discordant behaviours.

Multiplicitous Self. The Multiplicitous Self can be characterized as harbouring many im-pulses that arise as a consequence of passivelyreactingto changing circumstance. As distinctfrom the previous mode of responding, participants coded in this category began by f latly as-serting that the individual in question retains a unified self, even if what they went on to de-scribe could initially be construed as belying this claim. Notwithstanding their claims for self-unity, these participants described themselves and others as no more than the common causeof their actions, referring to a self that is inhabited by different desires or emotions that are in-dependently evoked by changing circumstance, giv ing rise to the individuals otherwise discor-dant behaviours. Unlike the remaining warranting strategies, however, for those employingthis brand of self-unity, what brings one or another impulse to the fore is not a conscious voli-tion, but simply some matter of circumstance. That is, more than any other self-unity warrant-ing strategy, responses that fell into this category carefully detailed the interplay between ex-ternal, situational factors and internal, affective factors that bring out, or elicit, good and badbehaviours at dif ferent times. The multiplicitous self is an essential lypassive and reactionaryentity, regarded simply as mechanical conveyors of animating forces [Bandura & Cervone,1986, p. 12] forces that reside in either the external environment or their own inaccessible in-

ternal emotional world where the implication is that both of these sources of influence lie be-yond the individuals conscious, volitional control.Hierarchical Self. The Hierarchical Self employs a self-structure that allows for the possibil-

ity of both one volition consciouslyacting on circumstance, and many impulses passivelyreactingto circumstance. Many folk psychological accounts of self-unity rest on something approximat-ing the Enlightenment distinction [Locke, 1632/1996] betweenfirst-order, conscious, controlledvolitional actions, on the one hand, and, on the other, second-order, largely unconscious, untamed

Fig. 1. Spectrum of self-unity.

No self-unity Multiplicitous self Hierarchical self Singular self

Many impulses/ Many impulses and One volition/ActiveReactive volitions/

Reactive and active

Least unified Most unifiedDegree of self-unity


10/27



impulses [Frankfurt, 1971]. Like the multiplicitous self, responses scored in this third categorydescribed a self that houses more than one desire. However, in this case, one of these desires risesabove the others and speaks louder than its fellow self-inhabitors. This master desire is seen asactively directing good behaviour mostof the time, t hough it is occasionally overcome by build-

ing impulses that are evoked in certain c ircumstances, resulting in behaviours we are not proudof. Therefore, unlike the multiplicitous self, hierarchical self responses suggest that the individ-ual is more than the common cause of all its act ions, insofar as the self is seen to remain underthe influence of a single, conscious, active volition at least mostof the time. Generally, theseresponses evoke a kind of nave-psychoanalytic stance, where conflict a rises only when otherwiseunacceptable impulses and emotions, generally held in check, occasionally burst forth and giverise to actions that are not in keeping with the individuals characteristic ethical stance.

Singular Self. The Singular Self can be characterized as guided byone, and only one, voli-tion consciouslyacting on all circumstances. On this account, there is no mention of active,conscious volitions mingling, or competing with base impulses within a single self. Selves, asunderstood by those whose responses were scored as singular, are seen as seamless, fully activeagents autonomously formulating their own actions in every circumstance. Rather, by this reck-oning, the appearance of seemingly contradictory intentions constitutes nothing more thanthat: mere appearance. Like the Mafia don who both cuts throats and cuddles his grandchil-dren, all to bring about whats best for the family, participants who scored in this category of

self-understanding claimed that all of their actions were owed to one and the same self-chosenvolition that, while differently manifesting itself in different contexts, does not itselfchangefrom situation to situation. That is, while the structure of the self, on this account, may well beunderstood to have its beginnings in society or various intrapsychic turmoils, the singular vo-lition that underpins all of their seemingly diverse actions is understood to operate in ways thatare independent of social circumstance or other passing fancies. In this sense, the singular selfbecomes what Lakoff and Johnson [1999] have called an embodied agent an individual agentwho, while rooted, or situated, in a body and a society of people, employs a more autonomousframework of self-understanding (or what Bratman [2007] and Malle, Moses, & Baldwin [2001]all call responsible agency). Such young persons abandon all passive or reactive conceptions ofmental life, and count themselves instead as authors who plan and deliberate and are the own-ers of, and responsible agents for, all of their actions.

Interrater ReliabilityAll participant responses were independently scored by two coders. Participant responses

could be reliably scored into one category or another with 86% interrater agreement amongcoders (Cohens = 0.82). All disagreements were resolved through discussion.

Results

The following analyses were conducted to determine whether the self-unity cat-egories these adolescent participants ascribed to Jekyll and Hyde, themselves, or fa-miliar others varied in relation to their age. Analyses were also conducted to deter-mine whether some self-unity categories were employed more frequently than othersto the various target persons under consideration. There were no main effects forgender or gender-related interactions, so all results were collapsed across gender.

Age and Self-Unity CategoryThe mean age of participants who ascribed each category of self-unity to Jekyll

and Hyde, themselves, and others was examined by means of a one-way ANOVA.Jekyll and Hyde. There was no significant relationship between the age of the

participant and the category of self-unity ascribed to Jekyll and Hyde [F(3, 66) = 0.44,P = 0.73; table 1].


11/27



Self and Other.Unlike self-unity judgments made for Jekyll and Hyde, there wasa significant relationship between the age of the participants and the category of self-

unity they attributed to themselves [F(2, 67) = 15.02, p!

0.01, h2

= 0.31]. The Tukeypost hoc analysis revealed a significant difference between the mean ages of adoles-cents who characterized themselves as a multiplicitous self (M= 15.87) as opposedto a hierarchical self (M= 14.43) (p ! 0.01, d = 1.09), or as a multiplicitous self (M=15.87) rather than a singular self (M= 13.37) (p ! 0.01, d = 1.9; table 1). There was nosuch significant relationship between the age of the adolescent and the category ofself-unity that they ascribed to familiar others [F(2, 67) = 2.27, p = 0.11; table 1].

Target Persons and Self-Unity CategoriesA 2 goodness-of-fit test was used to determine whether there were differences

in the frequencies of self-unity categories participants ascribed to Jekyll and Hyde,themselves, or familiar others.2 For descriptive purposes, the effect sizes of salientand theoretically relevant frequency differences were explored by means of the

coefficient.Jekyll and Hyde. There was no significant difference within the frequencies ofself-unity categories adolescents ascribed to Jekyll and Hyde [2(3, n = 70) = 3.94,p = 0.27; table 1].

Self. There was a significant difference within the frequencies of self-unity cat-egories participants attributed to themselves [2(3, n = 71) = 36.1, p ! 0.01]. No par-ticipants imagined themselves as possessing no self-unity ( = 0.97), and they morefrequently construed themselves as a multiplicitous self or hierarchical self, ratherthan as an undifferentiated singular self ( = 0.71; table 1).

Other. There was also a significant difference within the frequencies of self-unity categories adolescents ascribed to familiar others [2(3, n = 71) = 24.49, p !0.01]. As was the case with judgments offered about themselves, no participantsimagined familiar others as possessing no self-unity ( = 0.97), and they more fre-

Table 1. Study 1a Mean age and self-unity category ascribed to Jekyll and Hyde (J&H), them-selves, and others by Canadian adolescents

Self-unity

category

Mean Standard deviation Sample size

J&H self other J&H self other J&H self other

No self-unity 14.67 NA NA 1.23 NA NA 15 0 0Multiplicitous 15.22 15.87 15.37 1.63 1.45 1.49 18 30 29Hierarchical 14.83 14.43 14.73 1.55 1.22 1.57 24 30 23Singular 15.08 13.7 14.5 1.65 0.67 1.34 13 10 18

2 In instances where no part icipants chose a given category for a target person, (e.g., no participantchose no self-unity for themselves), a corresponding data point was entered to avoid a structura l zerofor that category and allow for an accurate 2 analysis. This is reflected in the degrees of freedom forthe given 2 analysis, and is present in studies 1a and 1b.


12/27



quently ascribed to them a generally differentiated self-construal (multiplicitous selfor hierarchical self), rather than an undifferentiated singular self ( = 0.48; ta-ble 1).

Discussion

When comparing the self-unity conceptualizations produced for Jekyll and Hydeand for themselves (table 1), it is clear that the manner in which these adolescents con-strue their own self hood is importantly different from their approach to the disunityof Jekyll and Hyde. It seems that when considering the fict ional, massively disjointedJekyll and Hyde, a small but significant number of adolescents (roughly 1 in 5) werehappy to see him as a little more than a singular body housing multiple selves (noself-unity). By contrast, none of the participants understood themselves in this ut-terly disunified fashion. Rather, most of these adolescents (6 in 7) viewed themselvesas a single self, albeit in different manners (i.e., multiplicitous self or hierarchical self),

with relatively few (1 in 7) explaining their own contradictory behaviours as beingowed to a volitional, context-independent singular self. Finally, when it came to un-derstanding the self-unity of familiar others, these same adolescents did so in a man-ner that was closely related to the way in which they understood themselves. That is,they were generally unwilling to view others as utterly psychologically disunif ied, andmost often described them using the same multiplicitous self or hierarchical self ac-counts that they favoured in response to their own contradictory ways of behaving.

Age-graded differences, as they manifest themselves in the application of one self-unity warranting strategy rather than another, were only evident when participantswarranted their own self-unity, in contrast to their accounts of Jekyll and Hydes self-unity, or even that of familiar others (table 1). Apparently, for a given age, just aboutany account will suffice when attempting to reconcile the discordant behaviours offictional characters or close acquaintances. When, by contrast, it comes to warranting

self-unity in the face of what they take to be their own contrastive good and bad be-haviours, these same young people seem less inclined to be quite so arbitrary, oftenacting instead in ways that suggest that there is something uniquely at stake when try-ing to understand such contradictions in their own lives. As was the case with Harterand Monsours [1992] work involving conflicting attributes, the older participantsmost often resolved contradictions by appealing to imagined contextual influences.While all of these participants f latly asserted that they were, in fact, the same psycho-logical as well as physical individual, older adolescents increasingly presented them-selves as passive or primarily reactive to local and changing circumstances, going onto describe a tangled array of desires and emotions that are merely triggered or pro-voked by external events (multiplicitous self). Younger adolescents, by contrast, moreoften imagined themselves as acting from active volitions most of the time (hierarchi-cal self ), and those that were younger still more frequently saw themselves as driven

by a singular, teleologically unified sense of personal agency (singular self) a set ofage-graded relations that many might have intuited would run in exactly the oppositedirection.

Following from these initial findings, we counted as still open the question ofwhat would be found if the same questions were put to young adults. With three termpapers due this week, a boyfriend theyre happy with, another guy theyreally like, a


13/27



boss they hate, or even a child on the way, it may not be the case that they can so eas-ily continue to imagine themselves as context-dependent multiplicities. Or, alterna-tively, it may be the case that they are more motivated than ever to imagine them-

selves in just this way.

Study 1b: Young Adults

Method

ParticipantsThe young adult participants were 50 undergraduates, all born and raised in North Amer-

ica. There were 23 males and 27 females; ages ranged from 18 to 25 years, with 8 18-year-olds,9 19-year-olds, 9 20-year-olds, 11 21-year-olds, 6 22-year-olds, 4 23-year-olds, 2 24-year-oldsand 1 25-year-old (M= 20.46, SD = 1.79 years).

Materials and ProcedureThe Jekyll and Hyde stimulus materials were identical to those administered to partici-

pants in study 1a. However, participants in study 1b were not interviewed, but instead filled outa self-administered questionnaire package that posed the same questions that were previouslyemployed in the standardized interview protocol questions put to participants in study 1a. Toensure that answers generated in the study 1b questionnaire were comparable to answers gener-ated in the study 1a interview, 15 participants filled out the questionnaire either before or aftera study 1a interview, and their responses were compared by two coders. Codings from the par-ticipants interview and self-administered worksheet were in agreement 87% of the time acrossall judgements for themselves, others, and Jekyll and Hyde (Cohens = 0.83).

Undergraduate participants were recruited in psychology courses, where they were of-fered CAD 10 to pick up, complete, and return the self-administered questionnaire package.Only those packages from students born and raised in North America were included in thestudy.

Interrater Reliability

All participant responses were independently coded by two coders. The responses of par-ticipants could be reliably coded into one category or another, with 86% interrater agreementamong coders (Cohens = 0.82). All disagreements were resolved through discussion.

Results

There were no main effects for gender or gender-related interactions, so all re-sults were collapsed across gender.

Age and Self-Unity CategoryA one-way ANOVA determined that there was no significant relationship be-

tween the age of these young adults and the self-unity category they ascribed to Jekyll

and Hyde [F(3, 46) = 0.79, p = 0.51], themselves [F(2, 47) = 0.19, p = 0.83] or others[F(2, 47) = 1.01, p = 0.34; table 2].

Person and Self-Unity CategoryA 2 goodness-of-fit test was used to determine whether there were differences

in the frequencies of self-unity categories participants ascribed to Jekyll and Hyde,


14/27



themselves, or familiar others. Effect sizes of salient and theoretically relevant fre-quency differences are reported.

Jekyll and Hyde. There was a significant difference in the frequencies of self-

unity categories participants ascribed to Jekyll and Hyde [2

(3, n = 50) = 8.71, p =0.03]. These undergraduate participants, like their younger counterparts, also imag-ined that Jekyll and Hyde possessed no self-unity more frequently than either amultiplicitous self ( = 0.50), hierarchical self ( = 0.27) or singular self ( = 0.35)(table 2).

Self. There was again a significant difference within the frequencies of self-uni-ty categories these participants ascribed to themselves [2(3, n = 51) = 29.71, p ! 0.01].No participant imagined themselves as possessing no self-unity (= 0.96), and theymore frequently attributed to themselves a largely differentiated multiplicitous self,rather than a less differentiated hierarchical self ( = 0.4), or an undifferentiated sin-gular self ( = 0.47) (table 2).

Other. Again, as had been the case with the younger participants of study 1a,there was a significant difference in the frequencies of self-unity categories ascribed

to familiar others [2

(3, n = 51) = 33.31, p ! 0.01]. No participant imagined familiarothers as possessing no self-unity ( = 0.96), and as was the case with judgmentsmade about themselves, they more frequently ascribed to others a largely differenti-ated multiplicitous self, rather than a less differentiated hierarchical self ( = 0.57),or an undifferentiated singular self ( = 0.38) (table 2).

Discussion

So how do such young adults imagine themselves as they enter university, getjobs, get dates, and have a few more responsibilities heaped on their shoulders? Forthe most part, just as they did as they left high school multiplicities reacting to cir-cumstance, a skein of moods and momentary wants real demoniacs, if you will.

Over half of these young adults again relied upon this mode of understanding fortheir own self-unity, as well as the self-unity of familiar others, with the remainderdivided between either a hierarchical or singular self. While Jekyll and Hyde weregenerally seen to possess no common identity, these young adults continued to avoidimagining themselves or familiar others as a legion, though perhaps for differentreasons (more on this shortly). Overall, the young adult participants appeared to ac-

Table 2. Study 1b Mean age and self-unity category ascribed to Jekyll and Hyde (J&H), them-selves and others by Canadian young adults

Self-unity

category

Mean Standard deviation Sample size

J&H self other J&H self other J&H self other

No self-unity 20.1 NA NA 1.7 NA NA 21 0 0Multiplicitous 21.29 20.32 20.55 2.28 1.98 1.96 7 28 29Hierarchical 20.5 20.66 19.62 1.73 1.92 1.59 12 12 8Singular 20.6 20.6 20.76 1.79 1.07 1.48 10 10 13


15/27



tively avoid imagining themselves as entirely lacking in self-unity (no self-unity),just as they seemed to shy away from counting themselves in singular, agentive terms(singular self).

The cross-sectional data in hand are consistent with the prospect that, while atleast some of the young people in study 1a underwent fundamental changes in howthey viewed themselves over the course of their adolescence, the proportion of re-spondents that employed the most frequently chosen solution strategy (the multi-plicitous self) appears to remain constant across these two study samples. Whetherthis dominant solution strategy remains intact when college-aged young adults be-come real adults with spouses and children and angry bosses is a matter forconjecture and future research.

What, then, is motivating young peoples gradual shift in their understandingof self-unity from singularand active to multiplicitous and reactive self-construals?While one might speculate on the pragmatics of one form of self-construal or an-other, it remains to be worked out what, exact ly, is at stake when we adopt a particu-lar mode of self-understanding in the face of our sometimes irresolute, morally haz-

ardous behaviours.

Study 2: Freedom and Responsibility

As reported earlier, the work by Harter and Monsour [1992] exploring age-grad-ed differences in self-construal found a similar age-graded shift in how adolescents

viewed (in their case) contradictory personality attributes. Similarly, with advancingage, the young people in our studies came to more frequently incorporate opposingattributes into their own self-structure, and attaching greater importance to the rolethat changing circumstances can play in drawing out different aspects of our char-acter. Harter and Monsour offered a compelling argument for why our self-conceptsmay change over the course of adolescent development, an argument that may be

entirely correct when considering ethically neutral personality attributes (e.g., shyvs. outgoing), rather than intentionally driven desires and behaviours with a strongnormative valence. Harter and Monsour reasoned that, as we grow older, experiencesimply teaches us that someone may possess different or even contradictory attri-butes attributes that lead them to behave differently in different situations (i.e., shywith strangers, outgoing with friends) without causing the entire self-system tounravel.

When one attempts to apply such a practice makes perfect explanation to theresults of studies 1a and 1b, however, it becomes apparent that such an account doesnot entirely fit, and for the following reason. That is, when considering contradictionsin their own morally hazardous behaviours, older participants in our studies moreoften favoured a multiplicitous account of self-unity, but did not do so when consid-ering such contradictions in other people. Clearly the same experience-based account

cannot be made to work for such different outcomes. What is obviously needed issome accounting strategy capable of making sense of why progressively older personsincreasingly offer situational explanations for their own contradictory behaviours,but not those of others. That is, it was often the case that older adolescents and youngadults were increasingly likely to provide examples in which their own good actionswere described as wholly voluntary, whereas when describing themselves as some-


16/27



times acting like a bad person, responsibility rapidly shifted to their lousy mood,their rowdy friends or the weather. Could it be, we wondered, that such an external-izing of their bad self underlies young peoples gradual shift towards an increas-

ingly multiplicitous, contextualized conception of their own self hood?Psychologists have long demonstrated that people generally claim that personalefforts are responsible for their successes, but often discount their failure as owed tothe situation [for a review, see Heckhausen & Schulz, 1995]. Children as young as 9seem inclined to causally att ribute their own and others actions to both internal andexternal forces, and to demonstrate a common bias towards imagining personal fail-ures as owed to circumstances beyond their personal control [Skinner, Chapman, &Baltes, 1988]. A meta-analysis conducted by Mezulis, Abramson, Hyde, and Hankin[2004] found that this self-serving attributional bias diminishes somewhat by earlyadolescence (1214 years), and increases again over the course of the late adolescenceyears (1518 years). As yet, however, there is no work that directly explores how suchshifting control strategies in adolescence may play a part in perceptions of self-unity.Insights into such self-serving practices, if they are to be found at all, need to be lo-

cated elsewhere. At least anecdotally, our discussions with young people in the firstof our studies seemed to mirror the moral disengagements observed in adolescent par-ticipants by Bandura and his colleagues [Bandura, Barbaranelli, Caprara, & Pastorel-li, 1996; Bandura, Barbaranelli, Caprara, Pastorelli, & Regalia, 2001]. Since the previ-ous investigations had not been carried out with specific concerns about self-unity inmind, our own interviewers had not separately probed for young peoples internal orexternal attribution of what Bandura called moral agency when pressing respondentsto explain their own good and bad behaviours.

During a second wave of data collection (study 2), utilizing parts of the samplethat comprised study 1a, these same adolescents were subsequently asked to furtherprovide separate examples of behaviours of which they were proud, and behavioursof which they were not proud. These participants were then asked to separately ex-plain why they believed that they demonstrated these seemingly inconsistent behav-

iours, and their answers were coded as expressing an internalor externalattribution.It was hypothesized that younger participants would primarily express an internalattribution for both their good and bad behaviours, while older adolescents wouldexpress an internal attribution only for their good behaviours, while expressing anexternal attribution for their bad behaviours.

Such a pattern of results, if obtained, would suggest the possibility that the age-graded shift in self-understanding observed in study 1a cannot be entirely laid at thedoor of some general increase in cognitive sophistication. After all, if older adoles-cents primarily understand themselves as multivoiced selves reacting to circum-stance, and if they do so simply because they have gained (as Harter and Monsour[1992] suggest) the intellectual capacity necessary to appreciate the role of circum-stance in their contradictory behaviours, then it would be expected that these sameadolescents would offer an external attribution for both their good and bad behav-

iours. Evidence that only bad behaviours are being accounted for with referenceto the impact of circumstance would, by contrast, suggest that other factors areat play.

Furthermore, if it is imagined that an increasing desire on the part of older ad-olescents to externalize responsibility for negative behaviours actually underlies thechoice of a multiplicitous, contextual ly bound conception of self-unity, then it would


17/27



be expected that there would be an independent relationship between the categoryemployed to account for their own contradictory behaviours, and their own directexpressions of either an internal or external attribution for their bad behaviours.

Specifically, it would be expected that the more unif ied and autonomous adolescentsconception of their own self-unity proved to be, the less likely they would be to ex-ternalize responsibility for their negative behaviours.

Method

ParticipantsThe sample of consisted of 45 high school students drawn from a second wave of data col-

lection in study 1a. There were again 17 males and 28 females; ages ranged from 13 to 17 years,with 11 13-year-olds, 15 14-year-olds, 6 15-year-olds, 4 16-year-olds, and 9 17-year-olds (M=14.67, SD = 1.46 years).

Materials and ProcedureThe study materials consisted of the same Jekyll and Hyde comic book stimulus materials,

followed by a standardized interview protocol probing the degree to which the same partici-pants employed an internal or external attribution style when explaining their own and othersgood and bad behaviours.

As was the case in study 1a, one of two interviewers asked participants to separately de-scribe themselves behaving in ways of which they were proud, and ways of which they were notproud (these quest ions were counterbalanced across conditions). For each good and bad behav-ioural anecdote that was offered up, participants in study 2 were also asked to separately explainwhy they thought this event had taken place. Participant responses were subsequently coded ashaving employed a generally internal or external attribution for their good and bad behav-iours.

Coding and TypologyParticipant behaviours assigned an internalattribution style put the weight of responsibil-

ity on their own desires, beliefs, and choices, without laying emphasis on the pressures of the

situation. Alternatively, participants who explained their behaviours in a manner consistentwith an externalattribution of agency were those that made no explicit reference to their owndesires, beliefs, or choices, and instead placed a heavy emphasis on the role of the situation.

Interrater ReliabilityAll participant responses were independently coded by two coders. The responses of par-

ticipants could be reliably coded into one category or another, with 91% interrater agreementamong coders (Cohens = 0.80). All disagreements were resolved through discussion.

Results

The following analyses were conducted to determine whether participants morefrequently provided internal or external attributions for their good and bad behav-

iours, respectively. Additional analyses were conducted to determine whether olderand younger adolescents would more frequently externalize their bad behaviours.Finally, an analysis was conducted to determine whether participants who external-ized their bad behaviours were those same persons who also characterized them-selves as having a multiplicitous self. There were no main effects for gender or gen-der-related interactions, so all results were collapsed across gender.


18/27



Moral Valence and AttributionA 2 goodness-of-fit test determined that there was a significant difference be-

tween the frequencies of internal and external attributions for good behaviours [2(1,n = 45) = 27.22, p ! 0.01, = 0.78], with the overwhelming majority of participantsdescribing themselves as the primary cause of behaviours of which they were proud.In contrast to their responses to questions about good behaviours, there was no sig-nificant difference between the frequencies of internal and external attributions for

their bad behaviours [2

(1, n = 45) = 0.02, p = 0.88], where internal and external at-tributions were roughly evenly split (table 3).

Age and Moral AttributionIndependent-samples t tests were conducted to determine any differences be-

tween the mean ages of adolescents who offered up an internalas opposed to an ex-ternal attribution for their good and bad behaviours. For descriptive purposes, amedian split of participant ages in months was produced and 2 tests of differencewere conducted to determine any differences between younger and older adolescentsin their frequencies of internal and external attributions offered up for good and badbehaviours.

Good Behaviours. There was no significant difference between the mean ages ofadolescents who offered up internal (M= 14.75), or external (M= 14.0), attributions

for their good behaviours [t(43) = 1.08, p = 0.28], nor was there a significant differ-ence between the frequencies of internal and external attributions offered by ado-lescents coded as either younger or older for their good behaviours [2(1, n = 45) =0.18, p = 0.68; table 4].

Bad Behaviours. There was, however, a significant difference between the meanages of adolescents who offered up internal (M= 13.74), as opposed to external (M=

Table 3. Study 2 Number of adolescents who ascribed inter-nal and external attributions to their good and bad behav-iours

Internal External

Good 40 5Bad 23 22

Table 4. Study 2 Attribution ascribed to good and bad behav-iours by older and younger adolescents

Good behaviour Bad behaviour

internal external internal external

1315 years 20 3 18 51618 years 20 2 5 17


19/27



15.64), attributions for their bad behaviours [t(43) = 5.72, p ! 0.01, d = 1.69; table 5],such that younger adolescents more frequently internalized their moral failings[2(1, n = 23) = 7.35, p ! 0.01, = 0.57], and older adolescents more frequently exter-nalized their moral failings [2(1, n = 22) = 6.55, p ! 0.01, = 0.55; table 4].

Self-Unity Category and Attribution of Moral AgencyA 2 test of difference indicated that while no relationship existed between at-

tribution and self-unity category when considering good behaviours [2(2, n = 45) =1.97, p = 0.374], a relationship did exist when considering bad behaviours [2(2, n =45) = 7.38, p = 0.025, = 0.40]. As anticipated, participants who were coded as evi-dencing a multiplicitous self predominantly externalized their bad behaviours, andparticipants who were coded as evidencing a singular self predominantly retainedan internal attribution for these bad behaviours (table 5).

Discussion

When the results of study 2 are considered as a whole, three findings becomeespecially salient. First, both younger and older adolescents displayed a strong ten-dency to take full volitional credit for the good things that they do. Even older ado-lescents chose to describe themselves as volitional agents who provide the primaryimpetus for the behaviours in which they take pride; this, despite the fact that theysubsequently demonstrate a clear capacity to generate external, contextualized be-havioural accounts when explaining away the behaviours in which they take no pridewhatsoever.

Second, older and younger adolescents differ significantly in how they makemoral attributions about their bad behaviours. Younger adolescents more often re-tained an internal attributional style when explaining their bad, as well as their goodbehaviours, whereas older adolescents regularly externalized responsibility only fortheir own negative behaviours. In general, the apparent desire of those on the cusp

of adulthood to externalize their self-perceived failings is in keeping with similarfindings in the social psychological literature, where adults imagine that desired out-comes result from their own conscious effort, while undesired outcomes are sloughedonto elements of the situation that lie well outside the purview of volitional control[e.g., Covington & Omelich, 1979]. Furthermore, the age-graded trajectory evidentin the present data conforms with previous research demonstrating an increase in

Table 5. Study 2 Self-unity category and attribution ascribedto good and bad behaviours by adolescents

Good behaviour Bad behaviour

internal external internal external

Multiplicitous 19 1 6 14Hierarchical 12 3 9 6Singular 9 1 8 2


20/27



self-serving attributional bias over the course of adolescence [Mezulis et al., 2004].If it is the case that older adolescents seek to internalize their good behaviours andexternalize their wrong conduct, this impetus is not, however, uniformly present at

all earlier points in the course of development. Young adolescents seemingly placeless of a premium on slipping the leash of responsibility a tendency that may growin importance as they make their way into young adulthood.

Although young adolescents no doubt suffer undesirable consequences as a re-sult of the transgressions they own up to, it may be the case that they lack the cogni-tive horsepower required to selectively contextualize their good and bad behaviours.It may also be the case that neither the transgressions quickly volunteered in thisresearch context, nor the resulting punishments they were likely to incur for themare suff iciently severe that such young persons are moved to build independent com-partments within their self-construals to house them. Sisters get locked out of bath-rooms, friends get teased, and favourite TV shows are viewed well past bedtime, allwithout serious concerns about some impending judgment day. As we grow older,however, our mistakes or the price we are asked to pay for them often become

greater, the resulting setbacks more off-putting, the punishments more severe, andthe guilt more palpable. Friends are abandoned and betrayed, and tests are cheatedon along with the boyfriends they were partying with when they should have beenstudying. As our failings exact a greater pragmatic if not emotional toll, it wouldseem reasonable to imagine that we might increasingly seek to offset these costs byconstruing such failings as somehow not our fault even if in our heart of heartswe might not entirely believe it.

Finally, and most central to the work at hand, there is a strong linear relation-ship between young peoples conception of personal self-unity and their attributionsconcerning bad behaviours. Participants who view themselves as unified and au-tonomous singular selves generally emphasize an internal attribution, even for theirbad behaviours, whereas participants who typically view themselves as multiplici-tous and contextualized selves most commonly adopt an external attributional style

for these same behaviours. Given that adolescence appears marked by a shift fromsingular self to multiplicitous self conceptions, and given that this same period ofour lives is marked by an increasing desire to externalize our failings, it was thoughtreasonable to imagine that the two were somehow related, particularly when thesefindings fall in line with numerous theoretical frameworks that emphasize such arelationship [e.g., Fromm, 1941/1994]. Nevertheless, if it turned out that there was norelationship between these two sets of findings, then future efforts to spin a pre-sumptive causal account would become moot. The fact that such a relationship doesexist not only offers strong support for some escape from freedom account at leastpartially underlying adolescents changing conceptions of their own self-unity, butoffers more evidence than do more exclusively cognitive accounts, which generallyhave little to say regarding adolescent self-understanding.

General Discussion

In studies 1a and 1b, the story unearthed here is one of increased dif ferentiationand context-dependent self-construals as one grows older. In the young adult sampleof study 1b, the dominant mode of self-construal in the face of seemingly conflict-


21/27



ing, normatively regulated behaviour involves splitting off at least one portion ofthe self and shackling it to circumstance. It is speculated that this repeated self-split-ting, and the attendant flight from agency that it implies, can be understood as mo-

tivated by something like an escape from freedom a desire to generally downplayones personal responsibility as a way of slipping the leash on ones responsibility forquestionable behaviours. Study 2 offers support for this hypothesis, though, as hasalready been noted, much of the research on offer exploring similar elements of iden-tity development have generated findings that ultimately diverge from those report-ed here, at least insofar as the multiplicitous self-construals strongly evident in thepresent data never seem to emerge. The reason for this divergence, it will be sug-gested, is twofold.

First, the methods historically employed by most identity development re-searchers have been informed primarily by identity theorists who were themselvesinformed by an intellectual tradition whose clearest spokesperson was WilliamJames [1892/1963, 1900/1950, 1910/1984] the same William James who envisioneda mature self-system as involving a dualistic hierarchy comprised of a singular, voli-

tional I, governing over a socially variable and contextually controlled, objectifiedme. Not only are the great bulk of such available studies informed by James per-spective, it is suggested that they were perhaps inadvertently designed such that theirfindings are more likely to conform to the Jamesian belief that one such split (in thiscase the I/me split) is enough for any mature self-conception.

Second, few prior studies confronted young people with behaviours that signalthe sort of disjunct between right and wrong conduct that could easily prompt alter-native forms of self-splitting, or the volitionally impoverished self-understandingpresented by most of our older participants.

How Many Splits Are Enough?

Blasi and Milton [1991], to take a prime example, explored several interrelatedfacets of identity development, including young persons capacity for self-reflection,their relative awareness of interpersonal contradiction, their awareness of the impactof circumstance, and their understanding of both public and private selves. Under-standably, Blasi and Milton did not stumble blindly into matters of human self-un-derstanding without having some idea of what they were looking for and what theyexpected to find. In the history of Western thought, explicit definitions of person-hood, and the cultural expectations that follow from these definitions, go back mil-lennia, where the threads of these various conceptions were eventually collected upby William James and woven into a template that has been followed scrupulously too scrupulously, we will argue by social scientists addressing matters of personalidentity on into the present.

When comparing the self-conceptions of students in grades 6 and 12, Blasi and

Milton [1991] found that young persons: (a) attained a clear conception of their ownpersonal agency in early adolescence; (b) increasingly reflected upon their own agen-cy and how it is influenced by circumstance; (c) increasingly noted contradictionswithin their own agentic aims and behaviours across circumstances; and (d) resolvedthese contradictions by means of a dualisticpublic self versus private selfsplit, wherethe private realself houses a singular agency, and apparent contradictions are ban-


22/27



ished to the public self, or to unconscious depths that are beneath the reach of thereal selfs volitional control a form of self-construal that, in our own typology, islabelled as hierarchical.

Perhaps reassuringly, Blasi and Miltons [1991] findings closely mirror parts ofthe findings presented here. Much like Blasi and Miltons young respondents, theyoungest adolescent participants in study 1a generated self-understandings that weresaturated with centralized agency. When explicitly required to address instances oftheir own good and bad behaviours, the 11- to 13-year-old participants reported onhere were most commonly scored as a singular self, at least insofar as they made ex-clusive reference to a unified will operating across contexts. Similarly, among olderparticipants there was a greater dependence upon a more hierarchical self a dual-istic split almost identical to the public/private conceptualization observed by Blasiand Milton. Again, as was the case with Blasi and Miltons participants, the presentsample of young adolescents also demonstrated an increasing awareness of the gen-uine tensions between their own contradictory desires and behaviours across cir-cumstances and dealt with this conflict by reorganizing themselves into a stable

dualistic hierarchy where their real self is preserved as synchronically coherent.It is at this midpoint in these accounts of adolescent development, however, thatthe similarities between the adolescent participants in the present study sequenceand Blasi and Miltons young respondents apparently end. For Blasi and Milton, onesplit is enough to account for their data. In contrast, the young people who com-prised our admittedly cross-sectional studies 1a and 1b kept splitting. Although theolder adolescent participants sometimes maintained a hierarchical self when ac-counting for their good and bad behaviours, the developmental momentum evidentin the age-graded samples presented here clearly lay with the emergence of increas-ingly multiplicitous, and pointedly non-hierarchical self-construals of the sort fa-

voured by the university-aged young adults. Given that this further splitting wasclearly in evidence in the research presented here, why do Blasi and Milton presentwhat is, by comparison, a relatively truncated developmental arc where a dualistic

real self-solution triumphs? A hint at the answer may be found in the interview pro-tocol that Blasi and Milton employed:

Sometimes people say This is the real me. This is my true self. What do they mean? Arethere certain parts of yourself that are more true and real? Can you give me some exam-ples? Why would these parts be especially real? Who decides what is real and what is not?How does one go about deciding?

Given such a line of questioning more or less common, we assume, to thosesimilarly rowing in the wake of William James it is perhaps not so surprising thatBlasis participants eventually constructed (consented to?) a real self-solution to theproblem of interpersonal contradiction, and that such participants appeared not tobe motivated to move on to invoking the multiple splits evident in the responses of

the older adolescent and young adult samples of studies 1a and 1b. The suggestion be-ing militated for here is that Blasi and Miltons interview procedure, like that of manyof their contemporaries, essentially begs the question, in that the solution to the prob-lem of synchronic unity is explicitly stated in the presentation of the problem.

It is not likely the case that Blasi and Milton were deliberately angling for a spe-cific developmental trajectory in their participants self-understanding. After all, the


23/27



present research presents findings derived from adolescent responses categorizedand coded from an open-ended interview protocol; even this hands-off method gen-erated some very similar results findings that suggest a developmental course often

moving from the undivided agency of some singular self to the real self-solutionimplied by an end-state hierarchical self. To the extent that Blasi and Miltons resultsdiffer from those presented in study 1a, this difference is arguably owed, we suggest,to prior theoretical commitments that have tacitly shaped Blasi and Miltons meth-odology commitments that are rooted in the classic work of William James, andthat are shared by the great bulk of contemporary researchers who have exploredadolescent identity development.

As demonstrated by the work of Chandler et al. [2003] exploring construals ofdiachronic self-continuity, adolescent participants, when faced with the paradox ofsameness and change, most often express an essential ist solution. That is, our West-ern European cultural legacy prompted many of their participants to locate some-thing that doesnt change, and mark it as real. Without delving into the long intel-lectual history of Western essentialism already summarized by Chandler et al . [2001],

it would not be out of order to suggest that an essentialist bias may not only informlay solutions to problems of self-coherence, but also the social science theorists whoseown tacit essentialist commitments allow prescriptive expectations of real self-solu-tions to bleed into their accounts of how ordinary people understand their own self-coherence, either across time or across contexts.

The argument developed here is that William Jamess solution to the paradox ofsameness and change entrenched what amounts to an essentialist dualistic hierarchyconcerning the possible meaning of self-unity within the psychological literaturefrom the outset. According to James, there is an I self a unitary, consciously ex-perienced sense of agency that binds the self over time and across contexts. But, thereis also a me self the self as object of contemplation made up of different contex-tually driven desires and attributes acquired and manifested over ones life andacross social interactions. On Jamess account, some part (the me part) of self is al-

lowed to change, while another (the I part) stays the same. More to the point, thatpart that stays the same is that part of the self which is in control, and the part thatchanges represents the social selves (p. 282) whose active-feeling states (p. 181)are, at the best of times, undercontrol.

Later in the century, the Russian literary theorist Mikhail Bakhtin [1986] writ-ing in opposition to the unitary I position posited by James assigned a controllingI to every social self, arguing that the self represented a polyphony of agencies risingand falling in different circumstances and occasionally working against one anther.Prior to James, Kierkegaard [1843/2000] had also imagined the self as such a nest ofmultiple social selves reacting to circumstance, though with no agentic I self at all.If James were to have considered either of these positions, he would, no doubt, haverejected them, be it Bakhtins multiple Is, or Kierkegaards multiple mes.

Jamess reasons for construing the self in classic essentialist terms appears to

have been both prescriptive and descriptive, and there is no doubt that he was in-formed by every prominent theorist to have discussed self-concepts, going back asfar as Plato. There can be even less doubt that almost every prominent 20th centurytheorist and researcher to have addressed the notion of self-concept (the paradox ofsameness and change, in particular) has been deeply influenced by William Jamessarguments for the I and me selves. Not only has this conception been reified by


24/27



psychologists for a century, many developmentalists have understood somethinglike this division to be the end of the history of identity development, whether it isthe young persons active construction of this division, or their explicit realization

that this division has been in place all along [e.g., Damon & Hart, 1982; Selman,1980].In the present studies, those older adolescents and young adults whose answers

were coded as a multiplicitous self, appear to jettison the I while the number of mesare multiplied. In this sense, multiplicitous self responses address the problem ofmorally conflicted behaviours across contexts in ways that are similar to the topo-logical accounts offered up by the young research participants in Chandler et al.[2003] when attempting to resolve their behavioural and psychological changes overtime. Chandler and his colleagues described the topological self as a shifting behav-ioural phenotype without a genotype to account for phenotypic irregularities in be-haviour. Young people who offer up such topological accounts of personal changetend to outgrow them by early adolescence presumably because they acquire theinsights that allow them to fill in the hollow self-structure with a genotypic account,

much in the same way that Loevingers [1979] pre-adolescents eventually organizedthe faade of situational mes around a central volitional I.Given all that has been said by so many, it may seem remarkable, then, that when

the young people examined in the studies reported here attempted to understandtheir conflicted behaviours across contexts, something like a topological account ap-pears to arrive at the endof this developmental story instead of the beginning rath-er than filling in their hollow self-structure with a volitional I, these adolescentsappear to have worked to empty out this superstructure of any agentic infrastruc-ture. While this mode of self-understanding may become available in late adoles-cence, it may not be called upon unless there are circumstances that make a JamesianI/me infrastructure/superstructure construal less applicable or desirable. Oneof these prospective circumstances may involve rationalizing our good and bad be-haviours and the avoidance of the personal responsibilities that an I infrastructure

necessarily implies.

Limitations and Future Directions

The time-consuming nature of our open-ended interview procedure, combinedwith the limited access we had to the time of our high school samples, meant that itwas not feasible to also ask participants to fill out a bevy of questionnaires and mea-sures that would have allowed for a better understanding of how evolving concep-tions of self-unity would bear on a broad array of matters, including moral develop-ment, epistemic development, and personal agency. For example, how do Marcias[1966, 1980] stages of role commitment map on to the categories of self-unity under-standings presented here? Do hierarchical selves and multiplicitous selves corre-

spond to Marcias identity achievedand moratorium stages, respectively? It mightalso be imagined that the present research would relate to existing stage theories in-

volving moral development, in general, and Kohlbergs [1984] work in particular. Isa multiplicitous self more likely to f ind itself at Kohlbergs moral relativistic way-sta-tion, the once envisioned stage 4 1/2? And how do such multiplicitous selves other-wise make sense of the world? Does their multiplicitous self-understanding imply a


25/27



relativistic understanding of competing knowledge claims, or have they adopted arationalist[Chandler et al., 2001] approach to mediating their contradictory behav-iours in addition to others conflicting beliefs?

Despite these limitations, we believe that this program of research constitutesan important first step in understanding how it is that young people resolve mattersof personal self-unity in the face of morally valenced behavioural evidence to thecontrary.

Acknowledgements

We thank Jessica P. Flores, Erica Gehrke, Mitsumi Kawai and Patrice Kong for their helpin collecting data.

References

Aboud, F. (1979). Self: An identity, a concept, or a sense? In L. Strickland (Ed.), Soviet and Western per-spectives in social psychology (pp. 335352). New York: Pergamon Press.

Bakhtin, M.M. (1986). Speech genres and other late essays. Austin: University of Texas Press.Bandura, A., & Cervone, D. (1986). Differential engagement of self-reactive inf luences in cognitive mo-

tivation. Organizational Behaviour and Human Decision Processes, 38, 92113.Bandura, A., Barbaranell i, C., Caprara, G., & Pastorelli, C. (1996). Mechanisms of moral disengagement

in the exercise of moral agency.Journal of Personality and Social Psychology, 71, 364374.Bandura, A., Barbara nelli, C., Caprara, G., Pastorelli, C., & Regalia, C. (2001). Sociocognitive self-regu-

latory mechanisms governing tra nsgressive behaviour.Journal of Personality and Social Psychol-ogy, 80, 125135.

Blasi, A., & Milton, K. (1991). The development of the sense of self in adolescence.Journal of Personal-ity, 59, 217242.

Bratman, M. (2007). Structures of agency: Essays. New York: Oxford University Press.Broughton, J.M. (1978). Development of concepts of self, mind, reality and knowledge. In W. Damon

(Ed.), Social cognition (pp. 75100). San Francisco: Jossey-Bass.

Camus, A. (1955). An absurd reasoning. In A. Camus, The myth of Sisyphus and other essays (pp. 349).New York: Vintage Books.

Chandler, M.J., Lalonde, C.E., Sokol, B.W., & Hallett, D. (2003). Personal persistence, identity develop-ment, and suicide. In W. Overton (Ed.), Monographs of the society for research in child develop-ment. Vol. 68 (2, Seria l No. 231).

Chandler, M.J., & Proulx, T. (2006). Changing selves in changing worlds: Youth suicide on the fault-linesof colliding cultures.Archives of Suicide Research, 10, 125140.

Chandler, M.J., Sokol, B., & Hallett, D. (2001). Moral responsibility and t he interpretive turn: Childrenschanging conceptions of truth and rightness. In B. Malle, L. Moses, & D. Baldwin (Eds.), Inten-tions and intentionality: Foundations of social cognition (pp. 345365). Cambridge: MIT Press.

Covington, M.V., & Omelich, C.L. (1979). Effort: The double-edged sword in school achievement.Jour-nal of Educational Psychology, 71, 169182.

Damon, W., & Hart, D. (1982). The development of self-understanding from infancy through adoles-cence. Child Development, 53, 841864.

Deleuze, G. (1994). What is philosophy?New York: Columbia University Press.Derrida, J. (1978). Writing and difference. London: Oxford University Press.Epstein, S. (1981). The unity principle versus the reality and pleasure principles, or the tale of the scor-

pion and the frog. In M.D. Lynch, A.A. Norem-Hebeisen, & K. Gergen (Eds.), Self-concept:Advances in theory and research (pp. 2738). Cambridge: Ballinger.

Erikson, E.J. (1950). Childhood and society. New York: W.W. Norton & Co.Erikson, E.J. (1968). Identity: Youth and crisis. New York: W.W. Norton & Co.Fischer, K.W. (1980). A theory of cognitive development: The control and construction of hierarchies of

skill. Psychological Review, 87, 477531.


26/27



Fischer, K.W., & Aboud, C. (1994). Affective splitting and dissociation in normal and maltreated chil-dren: Developmental pathways for self in relationships. In D. Cicchetti, & S.L. Toth (Eds.), Roch-ester symposium on development and psychopathology: D

Proulz Chandler

Documents

Transcript of Proulz Chandler