Lawrence Erlbaum Associates (Taylor & Francis Group)

Autonomy and Self-Regulation Author(s): Charles S. Carver and Michael F. Scheier Source: Psychological Inquiry, Vol. 11, No. 4 (2000), pp. 284-291 Published by: Lawrence Erlbaum Associates (Taylor & Francis Group) Stable URL: http://www.jstor.org/stable/1449622 Accessed: 26/10/2008 15:02
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=lebtaylorfrancis. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].

Lawrence Erlbaum Associates (Taylor & Francis Group) is collaborating with JSTOR to digitize, preserve and extend access to Psychological Inquiry.

http://www.jstor.org

COMMENTARIES

Autonomyand Self-Regulation
Charles S. Carver
Departmentof Psychology Universityof Miami

Michael F. Scheier
Departmentof Psychology CarnegieMellon University The targetarticle summarizeswell the currentstate of the literatureof self-determinationtheory and its predecessors. Deci and Ryan (this issue) and their have generateda largebody of evimany collaborators dence, constituting impressive support for many of theirassertions.Whetheror not one agrees with all aspects of theiranalysis, it's hardnot to be impressedby the breadthof their efforts to develop a viewpoint on humanbehaviorthat is humanisticas well as organismic. In the target article they also extended their discussion to a wide range of theories that have focuses different from their own. In so doing, they made a strong claim for self-determinationas a comprehensive statementon humannature. In our commentary,we devote our attentionlargely to issues that we think still lack clarity, assumptions that seem arbitrary, and points on which we disagree a discussion that is wider ranging, see Carver& (for Scheier, 1999b). ics added]within the self); and autonomy,as a human characteristic,is an extension of this deeply evolved tendency [toward integrated functioning]."It seems clear thatautonomyhas very differentconnotationsin self-determination theorythanit does in everydaylanguage. We think it is confusing to use the word this way. We believe the matter of self-determinationis logically distinctfrom the matterof integrationwithin the self, and that the two should be kept separate. Deci and Ryan wrote elsewhere in the targetarticle that the development of an integratedself reflects a deep inner design of the human organism toward self-cohesion and the avoidanceof self-fragmentation. We have no problem accepting this principle. This principlefits very nicely with a hierarchicalorganization of goals and development (Carver & Scheier, 1999b). We just object to seeing it incorporatedinto the termautonomy.

What Is Autonomy? An aspect of this theory we always have trouble with is the concept of autonomy.In fact, we have several problems with it, startingwith how it is defined within the theory.

Is Autonomy Real? Apartfrom the issue of whetherintegrationshould be included in it, what else is bothersome about the concept of autonomy? Another thing that's bothersome is the questionof whethertrueautonomyactually exists. We raised this question earlier (Carver & Scheier, 1999a), asking whethertrue independenceof action ever really exists-in effect, whether there is free will. Clearlypeople experiencea strongersense of independence and self-determinationat some times than at others.Just as clearly, however, the subjective experienceof free will does not make it true in reality (cf. Wegner& Wheatley, 1999). We arehappyto have the sense of self-determination (at times), but we retain some skepticismover whetherthat sense is illusory. On the otherhand, it isn't clear whetherthe importance Deci and Ryan place on autonomy is about whetheror not people trulyare autonomous.The issue instead may be whether people "need to feel [italics added] autonomous,"as Deci and Ryan (this issue) wroteat one point in the targetarticle.Perhapsthe universalneed is actuallythe need to screen away enough of the controllingpressurestofeel the sense of self-direction,even if it happensto be illusory.RyanandDeci (1999) placed great weight on the fact that feeling a

Defining Autonomy What does autonomymean?The nearbydictionary defines autonomy as "the quality or state of being self-governing; ... self-directing freedom and esp. moral independence."A check on the adjective form "autonomous" adds "undertaken or carriedon without outside control; self-contained;responding, reacting, or developing independentlyof the whole" (Webster's Ninth New Collegiate Dictionary, 1984) Autonomy, then, seems to be self-direction, self-determination-plain and simple. In contrastto this, however, the targetarticle (this issue) includes the following statements:"autonomy concerns the experience of integration [italics added] and freedom;it is in people's natureto develop greater autonomy(as representedby greaterintegration[ital284

COMMENTARIES

sense of self-determination promotesbetter outcomes of varioussorts, as do Deci and Ryan in the targetarticle. Yet those findings cannotshed light on whetherit is the perception of autonomyor the existence of autonomy thatmatters.

Does Everyone Want Autonomy? People in Westernculturedo seem to like to feel autonomous. However, we've also wondered how universal this desire really is (Carver& Scheier, 1999b). Of considerableinterest,in this regard,arethe findings (discussed in the targetarticle)reported by Iyengarand That research found that reflections of Lepper(1999). intrinsic interest among Asian-American children were greater when a close and trusted other (their mother,their classmates)made a choice for them than when they made the choice themselves. Deci and Ryan interpretthese results as indicating that the means throughwhich autonomy is expressed can differ across cultures, an interpretation we find a little strained.Deci and Ryan say that "in some East Asian cultures,people may feel both highly volitional and autonomouswhen endorsingand enacting values of those with whom they identify"(this issue). We will not dwell on the discrepancybetween this and the dictionarydefinition of autonomy.But we are compelled to ask what evidence sustains the conclusion that the children whose mothers chose for them were feeling autonomous,as opposed to the conclusion thatthe desire for autonomyis a Westernphenomenon.We must also ask why the Asian-Americanchildrenwho chose for themselvesapparentlydid not feel autonomous.If the essence of autonomy is self-determination,how could their actions possibly be more self-determined thanby making their own choices?1

therdefinitionalproblemhere. To Deci and Ryan, the "self' in the term self-governing or self-determining means more thanjust an "internal" perceived locus of their occasional causality (despite emphasis on I-PLOC).An internallocus of the originationof the act (i.e., the impetus starts from within the person) does not equal self-determination in this theory.In this thean action is self-determined ory only if it reflects a value of the "true"self.2 Deci and Ryan acknowledge thattheirconcept of self is very differentfrom the self in most otherviews. Fromtheirperspective,not everythingthat'sinside the person'smind is a partof the true self. This point is easy to lose trackof. In an earlydraftof our commentary,one of us wrote that self-determination can be exercisedby steppingonto a busy highway withoutlooking,butthat'snot right.Internal perceived locus of causality could (we assume) be reflected in such an act, if the impetus to act originatesinside the person's mind. So could self-governance,in the sense that the decision to act is made on one's own with no outside interference(the dictionary definition of autonomy).The act could be freely chosen. But upon further review, such an act probably would not be autonomousin the Deci andRyanview, because it fails to advancea value of the trueself, andindeedmay conflict with an important value of the trueself (desire for self-preservation). This difference in assumptionsabout what constitutes the self can make debate difficult. It's not that Deci and Ryan are not explicit abouttakinga position that's different from those taken by others. They are explicit that they do so. But it's hardfor at least some of us to keep the differencein mind, and difficulties in communicationdo arise on thataccount(for a broader critique of the true self as a construct,see Carver& Scheier, 1999b).

The Self In Self-Determination A final point about the use of the terms autonomy and self-determination: Again we put aside the issue of and deal integration only with the dictionarydefinition of autonomousas self-governing.Thereremainsa fur-

The Core Needs Are Not Structurally Equivalent Anotherset of issues is raised by the natureof the basic needs thatDeci and Ryan postulate.Throughout theirtargetarticle, they refer to the three fundamental human needs of autonomy, relatedness, and competence. Relatednessis a relative latecomerto the Deci and Ryan model, which formerly incorporatedonly self-determinationand competence. Adding relatedness broadensthe theory. However, relatedness also differsin a fundamental way fromthe othertwo needs.
The notionof a "true" self is on theone handappealing,andon the otherhandmaddeningly difficultto be preciseabout.One ofthe problems manyobservershave had with self-actualizationmodels is that it is very hardto specify a prioriwhat anyone's true self consists of and thus what kind of behavioris self-actualizingfor that person.

As an aside, we offer the speculationthat Asian-Americanchildrenmay hold the belief thattheirmothersand peers typically make wiser or betterchoices thanthey would makethemselves. If so, they would prefer to relinquishcontrol over the choice, because relinquishingcontrolwould promotea betteroutcome.This would render the patternconsistent with the view that obtaininga desirableoutcome is more important than exertingpersonalcontrolover the outcome, a view thatwe have and othershave promoted(see, e.g., Bur& Scheier, 1998;Skowronski& ger, 1989;Carveret al., 2000; Carver Carlston, 1982). Althoughwe are mindfulof Deci and Ryan's point thatautonomyis not quitethe same as personalcontrol,this extrapolation seems not unreasonable.

285

COMMENTARIES

Relatedness Is Different in Form Relatednessconcernsa "content" domainof behavioralexperience(albeita broadone). The othertwo basic needs are not content needs. The need for competence applies in principleto any domain of behavior,as does the need for self-determination. Indeed, competence and self-determinationare qualities that could readily be applied to relatedness.That is, some people are good at maintainingrelatedness(are competent at it), others are not. Some people feel thattheir efforts to be connected to others are self-determined and authentic,othersfeel they "ought"to be connected to othersandare tryingto be so in orderto satisfy those conditions of worth. This difference between relatednessand the other supposed core needs raises a number of questions. Why is this particularcontent need special? Are there no otherpsychological contentneeds thatareuniversal besides this one? The other two needs that Deci and Ryan postulatehave to do with the "why"of behavior (actually, the more we think about it, the more they seem to be partof the "how"of behavior-that is, behaviorbeing done well and done freely). Can it be that the need for relatednessis the only universal"what"of behavior?

Self-Determination Moderates Competence The two noncontentneeds also turnout to have another importantrelationshipto each other, which renders suspect the statusof at least one of themas a basic need in and of itself: Deci andRyan say thattheirview is very differentfrom othersthatemphasizeconstructs such as personal efficacy. Their position is more narrow and specific. In theirview, efficacy or competence is not valuable unless it is efficacy at the right activities, being done under the right circumstances.More specifically, competenceis desirableonly if it pertains to an activity that authenticallyreflects some value of the trueself, and is being engaged in freely ratherthan being controlled. This conditionalqualitydoes indeedmaketheirtheory differentfrom others that emphasize competence. At the same time, however, this conditionalqualityis a double-edged sword. It also places a boundaryon the relevancefor humanwell-being of the need for competence. It means thatthe need for competencecannotbe fundamental in the sense of applyingto all domainsof behavior. Thus, competence does not standon its own in this theory. Competence matters only in interactionwith self-determination.Being highly competent at breaking into houses is not good, because housebreaking does not reflect the true self, even if one is choosing 286

freely to engage in it (question:does this remaintrue even if the housebreakingstems from such motives as the desire for fun, or curiosity,ratherthanthe desireto steal?).Being competentat the piano is not good if the reason for it is your mother standingover you with a switch for 10 years forcing 2 hr of practiceevery day. Competenceis good only if it furtherssome value of the self, and does so freely and without coercion. What's beneficial is behavior that simultaneouslyreflects competenceand self-determination. This moderationof the positive value of competence by self-determinationraises a methodological question. Moderationor synergistic effects should be tested via interactions between the predictors(Carver, 1989). Do researcherstesting the role of competence and self-determinationapproachthe question in designs that permit testing for interactions?If not, this would seem to be an importantavenue for furtherexploration. Althoughwe've focused on the issue of moderation with respect to competence, the question can also be raisedfor autonomy.Does self-determination standon its own? Is behaviorbeneficial if it is self-determined but not competent?If it is not, the same two problems as applyto competence.It pertainto self-determination cannot be fundamentalin the sense of applying to all activities, and the propertest of its effect is its interaction with competence.

Why Is Competence Necessary for an Activity to Relate to the Authentic Self? A final question about these needs (also raised in Carver& Scheier, 1999b) concerns the role that the theory assumes for competence in the authenticself. We've always found it odd that competence per se should be a hallmarkof the authenticself. What if a person wants to do something for perfectly authentic and intrinsic reasons, but is horrible at it? We can readilysee how this personwould have troublehaving a "flow"experiencewhile engaged in the activity. But why shouldthis desire(this goal) not be partof the person's authenticself? We are not arguing here that Deci and Ryan's position reduces to an efficacy model. As noted earlier, in their view it is not beneficial to be efficacious at an activity that's imposed on you, or is irrelevantto your true self. Efficacy is desirable only with regard to values that are authentic. But if the value is authentic, efficacy seems very important indeed. In fact, it seems from what Deci and Ryan have written in various places that efficacy is one determinantof whether the value is authentic. We find that hard to understand.

COMMENTARIES

Avoidance and Autonomy In consideringthe distinctionbetween self-determination and control, we have argued that many controlled actions appear to have their origins in the attemptto avoid some undesiredstate, eithersanctions from outside or self-sanctions (Carver & Scheier, 1999a, 1999b). In this respect,they appearto resemble "ought" self-guides (Higgins, 1987, 1996)-values thatpeople approach,but which seem to have a partial basis in the attemptto avoid other self-guides such as self (Carver,Lawrence,& Scheier, 1999). the "feared" and Deci (1999) rejectedthis argument,holding Ryan thatthe distinctionbetween approachand avoidanceis unrelated to the distinction between self-determined and controlledbehavior. They reiteratedthat position in the targetarticle. Some of the data suggest otherwise, however. As noted in the target article, Ryan and Connell (1989) found that introjected(controlled)and identified (autonomous) regulation related to motivated effort in schoolchildren. However, introjectedregulation also related to anxiety, whereas identified regulation related to enjoyment.Why were the childrenwho were engaged in introjectedregulationanxious?Anxiety relates to avoidanceprocesses (Carver& Scheier, 1998; Davidson, 1992; Gray, 1987; Higgins, 1987, 1996; Roseman, 1984). This patternof findings thus seems consistent with the idea that introjectedregulationin these children is rooted in an avoidance impulse-avoidance of a sense of guilt or shame. Eliot and Sheldon(1998) conductedanotherproject bearing on this issue. They had participants report 10 were and characterize goals they actively pursuing each as being primarilyan approachgoal or an avoidance goal. Participantsalso rated the extent to which their pursuit of each goal was based on reasons that were extrinsic, introjected, identified, and intrinsic. These ratingswere used to create indices of autonomy (from intrinsic and identified reasons) and controlledness (from extrinsic and introjected reasons). Elliot and Sheldon found thatpeople who were pursuing a higher proportionof avoidance goals reportedless autonomyand more controllednessin their goal pursuit. This patternis also consistent with the idea thatmuch of controlledbehaviorhas avoidanceas its core basis. Let us be clearaboutwhatwe arenot saying here,as well as what we are saying. We are not makingthe assertion, which Deci and Ryan incorrectlyascribe to Carverand Scheier (1999b) that approachis autonomous and avoidance is controlled. Our assertion is asymmetrical.WhatCarverand Scheier(1999b) wrote was "we can think of no case in which a value of the trueself as [RyanandDeci] discuss it has an avoidance tendency as its core motivationalbasis" (p. 291). Can an action thathas approachas its core motivationalba-

sis be controlled?Yes. Can an action that has avoidance as its core motivationalbasis be autonomous? Because the process leading to avoidance is always coercive (some danger forces the avoidance), we believe the answeris no.

Internalization and Compensation We also have some comments about how values that are not at first intrinsicbecome part of the self. Deci and Ryan describe the process of interalization into the as moving along a continuumof incorporation moves the controlinside the person's self: introjection mind but not into the self; identificationbegins an assimilationinto the self; integrationmakes that assimilation more complete. We have no doubt that this is a useful and accuratedescriptionof how people change over time. However, it leaves tantalizing questions hanging. In particular,what is the process by which this happens? Deci and Ryan discuss evidence that certain environmentalconditions foster the occurrenceof this inand othersdo not. But the mechanismby ternalization which it happens was less clearly specified. They wrote that "for integrationto occur there must be an for the individualto freely process and enopportunity dorsetransmitted values... ," (this issue) andthatpeople must grasp the importance of the values "and synthesize their meaning with respect to other values and motivations"(this issue). It appearsthat they are saying that internalizationtypically involves giving the child (a) a rationalefor why a value is important, (b) a chance to think about the rationalefor a while, and (c) a chance to see thatthereare links from this to other values that the child alreadyholds. It would appear that the key element is the discovery (through whatever means) of links to other values that are already in place within the self. Let us reframethis slightlyin termsof a hierarchical modelof goals,usingthe corevaluesthatDeci andRyan emphasize.We startwith the situationof externalcontrol. Externalregulationof children'sbehaviorentails pressurefrom the parent,eitherin the form of cajoling and rewardsor in the form of threatsand sanctions.As an example,we use the child who doesn'twant to practice the piano.Typicaltacticsto elicit practicing areemphasizing how much the parent wants the child to practice,and emphasizingthat the child will be able to go outsideandplay (or whatever)only afterthe practicing is done (Figure 1, SectionA). These two tacticsare with autonomy.Nonecontrolling,in thatthey interfere theless, they rely ultimatelyon core needs. Satisfying a sense yourmother'sdesiresis one way of maintaining of relatedness in the family.Gettingto the goal of being able to go outside and play is a way of pursuingan intrinsicmotivation. 287

COMMENTARIES

Is" I

mother

Satisfy

Getto do funactivity

Develop pianoskill

Practice the piano

Trueself

Maintain autonomy

Maintain relatedrnss

Do intrinsic Acquire interest competence

Satisfy mother

Get todo fun activity

Develop pianoskill

Practice the piano

Figure 1. Internalizationof a goal into the self, as changes in hierarchical links among self-goals over time and experience. Importanceaccrues to a concrete action goal in two ways. Theaction goal maycontributein a large way to attainmentof a higher ordergoal (indicated by a thickerline here), or it may contributeto attaining several higher order goals at once (indicatedby a larger numberof upwardprojections).Interferencewith attaining a higherorder goal (what one mightthinkof as an inhibitorylinkupward)is indicatedhere by barred lines. (A) This child is being induced to practice the piano by the mother,who emphasizes how happy it will make her, and that the child can go do somethingfun only after practicing. Both of these tactics rest in part on presumedcore needs (maintainingrelatedness,and intrinsicinterestin the otheractivity), thoughboth of these controlling tactics also interferewith the desire to maintainautonomy.(B) Later,these controlling reasons become less important;this child has discovered thatdeveloping skill at thepiano is a path to an overall sense of competence, to having connections with otherpeople, and also autonomy.Evena sense of intrinsic interesthas emerged, with respect both to the developmentof skill and to some of the activities ofpractice itself. Thegoal of practicing the piano has been internalized.

Now fastforward a few years.Afterhaving thought

about why (or whether) learning the piano is valuable-most importantly, after having experienced other connections among values in the hierarchyof goals within the self, the child is now organizeda bit differentlyregardingrelationsamonghigherordervalues that supportthe practicing(Figure 1, Section B). 288

Now the child realizes that acquiringa musical skill enhancespersonaloverall competence,and providesa new pathto relatedness(i.e., playing along with others andfor the enjoymentof others),and thatboth of those pathspermitthe activity to be autonomous.This child has also begun to discover an intrinsic interest in the music makingitself, and even in some of the activities

COMMENTARIES

of practicing.3The relevance of the controllingpaths (the links to satisfying mom and getting free to go do somethingelse) has faded. This activity is now identified, possibly even integrated. As suggested previously, the key to interalization seems to be the discoverythatthe activityor value connects to some othervalue thatis alreadyin place within or because it the self (eitherbecause it is fundamental, has alreadybeen acquired).This discoverycan be quite serendipitous(see the longer discussion on how new goals are acquired,in Carver& Scheier, 1999b). This line of thoughtwould also seem consonantwith Deci and Ryan's assertionsat severalpoints thatpeople can be filling core needs even when they are not tryingto do so (although they may later discover that they are doing so), and with their statementthat finding an acis influencedby priorextivity interestingor important of need satisfaction. periences It seems important,however, to be explicit about the underlyingassumption:thatfor goals andvalues to be internalized, their attainmentmust be enhancing congruencewithin the self. Withoutthis principle,we of the see no obvious way to accountfor the absorption goals into the self. This example thus illustrateshow a hierarchicalconception of self can render more explicit one of the processes that is less explicit in self-determination theory.

Compensatory Activity Deci and Ryan (this issue) contrastthe process of acwith what they called compensatory internalization The fulfillments. substitute or tivities, developmentof is compensatoryactivity hypothesized to occur when fulfillment of basic needs is repeatedlythwarted.The compensatory motives do not satisfy the thwarted need, but provide some "collateralsatisfaction."Presumably thatmeans that they satisfy needs or motives thatare not as importantas the core ones (e.g., people will work for money to buy food, even if they don't like the work). The notion of compensatoryactivities is a part of the theorythatseems less well explicatedconceptually thanothers.For example, it is not obvious why the failure to experiencerelatednessshouldlead to intensified attemptsto accumulatemoney, a "wrong"goal. Nor is it obvious why accumulatingmoney will enhance the person's focus on the pursuitof this goal.

This example is framedin termsof a child who actuallyhas some intrinsicinterestin the piano. If the interestisn't thereat all, thatlink won't solidify. Further, if the child hasno inbornskill at the piano,the sense of competencewon't be fed, not will the pianobecome a means to connectwith otherpeople. Forsuch children,practicingmaynever be absorbedinto the true self.

PresumablyDeci and Ryan are not saying thatsuccess in pursuingmoney "reinforces"the tendency to pursue money, as reinforcementis not a key part in their theory. Is it that accumulatingmoney, possessions, and fame producesat least a semblanceof relatedness thatfeels betterthanno relatednessat all? Thus people hold tightly to the ersatzrelatedness?But people presumablyfeel the differencebetween actualand ersatzrelatedness.Why should people who have only ersatz relatedness not simply experience more and more acutely the absence of satisfactionof their core needs? Why shouldn't involvement in pursuing the "wrong"goals lead people to be more ready to abandon those goals? Why (and how) do they become enmeshed in compensatoryactivities? How (and under could they ever get out of thatenwhat circumstances) meshment? We have suggested a different way of thinking about this kind of situation (see Carver & Scheier, 1998), one that rests partly on the ideas of dynamic-systems theory. We suggest that people often find themselves in less than ideal circumstancesand become adapted to those circumstances. This idea, which is hardlyrevolutionary,views the person's behavior less as compensatorythan as being "adaptive" in the currentlife situation(thoughpossibly not at all admirableto an outsider).As long as the behaviorremains adaptive(helps the person fit into personal life space-including fitting with the person's other values), thereis no pressureto change. Only if the person "error" (which may come from experiencessubstantial conflict with othervalues, fromdemandsfromoutside, from changes in otheraspects of the situation)is there pressureto shift, reorganize,move from one life patternto a differentone. This view would be consonant with the ideas discussed by Deci and Ryan, if one were to accept their core values as being alreadyembeddedin the person's hierarchyof self (as Deci & Ryan do, of course). This view adds some interestingpredictions,however, including the notion that a shift (if it ever does occur) may be expected to be fairly abrupt (cf. Hayes & Strauss,1998). On the otherhand,the fit with Deci andRyanwould be less good if one were not to assume that the needs for relatedness,competence,andautonomyarealready there and actively pulling at the person. One of the problems many observers have had with self-actualization models is that it's hardto specify a prioriwhat anyone's true self consists of. If an individual's true self incorporatesan intrinsic interestin accumulating wealth, for example,who can say thatthatinterestisn't part of that person's true self, and that the accumulation of wealth isn't self-actualizing for that person (Carver& Baird, 1998)? Although it is dishearteningto consider the possibility, it is not clear thatthe "trueself' of every person 289

COMMENTARIES

is really rooted in values that affirm human connectedness and excellence (cf. Baumeister & Campbell, 1999). Might it not be the case thatthe true self of the sociopath is exactly what it seems-exploitive, unconnected, and entitled; inimical to society, but supportiveand protectiveof its own autonomouswell being? Concluding Comment Despite theircriticismsof cyberetic theories,Deci and Ryan suggest that there is the potential for a successful integrationacross conceptualboundaries.We agree. For example, we think the notion of hierarchicality adds considerably to discussions of how needs are interwoven, and we think Deci and Ryan's model benefits from explicit considerationof this idea. As anotherexample,Powers (1973, Chap. 14 & 17) discussed in control-theoryterms some of the same issues as Deci and Ryan discuss in the context of autonomy needs, including the idea that giving a reward to induce a behavior can impede a natural process (p. 193). self-correctingreorganization Deci and Ryan criticized the cyberneticmodel primarilybecause it doesn't specify a basis for determining what the higher ordergoals are, thatit "seem[s] to suggest that what lies at the top of goal hierarchiesis not organismicallydetermined"(this issue). We suspect (once again) that the differencebetween views in this respect is not nearly as sharp as Deci and Ryan think it is (see also Powers, 1973, Chap. 17). We offer three bases for this opinion. First, the organismic and humanisticprinciple of integration,coherence, or congruence within the self is entirely commensuratewith self-regulatorycontrolprinciples. That'swhat discrepancy-reducing loops do: createand maintaincongruency. Second, a numberof people have begun to invoke the principle of self-organizationas a basis for emergent properties in dynamic systems (Nowak & Vallacher,1998; Prigogine& Stengers, 1984; for basic introductionsee Carver & Scheier, 1998, Chap. 17). Interestinglyenough, MacKay (1956) anticipatedthis notion and described a system of feedback processes that could evolve its own goals (see also Beer, 1995; Maes & Brooks, 1990). In such an arrangementthe goals at the top of the hierarchy would truly be "organismicallydetermined"-that is, determinedby the characteristicsof the entity as a whole in interaction with its environment-not just postulated,as Deci and Ryan did with respect to competence, autonomy, and relatedness. This is not to say that competence, autonomy,and relatedness are not perhaps key emergent properties withinthe framework of humanfunctioning.They may well be. Our thirdreason for believing that the cybernetic and organismic viewpoints are not as different 290

fromeach otheras Deci andRyanthinkthey arecomes from Powers (1973). In closing his discussion of the controlhierarchyhe had proposed,Powers speculated on the developmentof further layers of control.Partof his description of this possibility was not unlike Maslow's view of transcendent self-actualizers: whatsome see as a universal Perhaps urgetoward Oneness theglimmerings of a modeof perrepresents inwhich allsystem areseenasexamception concepts
ples of higher versions of reality, so that ... what we

call 'realities' will somedaybe manipulated as casuin service of sysallyaswe nowmanipulate principles tems.(Powers, 1973,p. 174) We suggest that the cybernetic(which Powers has also pointedout was an analogyfromliving to artificial systems, ratherthanthe otherway around)and the humanistic-organismicsharea greatdeal. Further explorations of their intersection can only enrich both of them. Notes of this commentwas facilitatedby NCI Preparation grantsCA64710, CA64711, and CA78995. We thank Sheri Johnsonfor her commentson an earlierdraft. of Psychology, UniCharlesS. Carver,Department versity of Miami, Coral Gables, FL 33124-2070. E-mail:[email protected] References
Baumeister,R. F., & Campbell,W. K. (1999). The intrinsicappealof evil: Sadismsensationthrillsandthreatened egotism. Personality and Social Psychology Review,3, 210-221. Beer, R. D. (1995). A dynamicalsystems perspectiveon agent-environmentinteraction. ArtificialIntelligence, 72, 173-215. Burger,J. M. (1989). Negative reactions to increases in perceived personalcontrol.JournalofPersonality and Social Psychology, 56, 246-256. Carver,C. S. (1989). How shouldmultifaceted personalityconstructs be tested? Issues illustratedby self-monitoring, attributional Journalof Personalityand Social Psycholstyle, andhardiness. ogy, 56, 577-585. Carver,C. S., & Baird,E.(1998). The Americandreamrevisited:Is it what you want or why you want it that matters?Psychological Science, 9, 289-292. Carver,C. S., Harris,S. D., Lehman,J. M., Durel, L. A., Antoni, M. C. (2000). How imporH., Spencer,S. M., & Pozo-Kaderman, tantis the perceptionof personalcontrol?Studiesof earlystage breastcancerpatients.Personalityand Social Psychology Bulletin, 26, 139-150. J. W., & Scheier,M. F. (1999). Self-discrepCarver,C. S., Lawrence, ancies and affect: Incorporating the role of feared selves. Personality and Social Psychology Bulletin,25, 783-792. Carver,C. S., & Scheier, M. F. (1998). On the self-regulationofbehavior. New York:CambridgeUniversity Press. Carver,C. S., & Scheier, M. F. (1999a). Themes and issues in the of behavior. In R. S. Wyer,Jr.(Ed.), Advancesin self-regulation social cognition (Vol. 12, pp. 1-105). Mahwah,NJ: Lawrence Erlbaum Associates, Inc.

COMMENTARIES Carver, C. S., & Scheier, M. F. (1999b). Several more themes, a lot more issues: Commentary on the commentaries. In R. S. Wyer, Jr. (Ed.), Advances in social cognition (Vol. 12, pp. 261-302). Mahwah, NJ: Lawrence Erlbaum Associates, Inc. Davidson, R. J. (1992). Anteriorcerebralasymmetryand the nature of emotion. Brain and Cognition,20, 125-151. Elliot, A. J., & Sheldon,K. M. (1998). Avoidancepersonalgoals and the personality-illness relationship. Journal of Personalityand Social Psychology, 75, 1282-1299. Gray, J. A. (1987). Thepsychology of fear and stress. Cambridge: CambridgeUniversityPress. Hayes, A. M., & Strauss,J. L. (1998). Dynamicsystems theoryas a An applicaparadigmforthe studyof change in psychotherapy: tion to cognitive therapyfor depression.Journal of Consulting and Clinical Psychology, 66, 939-947. A theoryrelatingself andafHiggins, E.T. (1987). Self-discrepancy: fect. Psychological Review, 94, 319-340. Higgins, E. T. (1996). Ideals, oughts, and regulatoryfocus: Affect and motivation from distinct pains and pleasures. In P. M. Gollwitzer & J. A. Bargh (Eds.), Thepsychology of action: Linking cognition and motivation to behavior (pp. 91-114). New York:Guilford. lyengar, S. S., & Lepper, M. R. (1999). Rethinkingthe value of choice: A culturalperspectiveon intrinsicmotivation.Journal of Personality and Social Psychology, 76, 349-366. MacKay, D. M. (1956). Toward an information-flow model of human behaviour. British Journal of Psychology. 47, 30-43. Maes, P., & Brooks,R. A. (1990). Learningto coordinatebehaviors. Proceedings of the AmericanAssociation of Artificial Intelligence (pp. 796-802). Los Alto, CA: MorganKaufmann. Nowak, A., & Vallacher,R. R. (1998). Dynamicalsocialpsychology. New York:Guilford. Powers,W. T. (1973). Behavior:Thecontrolofperception.Chicago: Aldine. Prigogine,I., & Stengers,I. (1984). Orderout of chaos: Man's new dialogue with nature.New York: RandomHouse. of emotions:A strucRoseman,I. J. (1984). Cognitive determinants turaltheory.In P. Shaver(Ed.),Reviewofpersonalityand social psychology (Vol. 5, pp. 11-36). Beverly Hills, CA: Sage. Ryan, R. M., & Deci, E. L. (1999). Approaching and avoiding self-determination: Comparing cyberneticandorganismicparadigms of motivation.In R. S. Wyer,Jr.(Ed.),Advancesin social cognition (Vol. 12, pp. 193-215). Mahwah, NJ: Lawrence Erlbaum Associates, Inc. Ryan,R. M., & Connell,J. P. (1989). Perceivedlocus ofcausality and interalization: Examiningreasons for acting in two domains. Journal of Personalityand Social Psychology, 57, 749-761. Skowronski,J. J., & Carlston,D. E. (1982). Effectsof previouslyexperiencedoutcomeson the desirefor choice. JournalofPersonality and Social Psychology, 43, 689-701. Webster's Ninth New Collegiate Dictionary. (1984). Autonomy. Springfeild,MA: Merriam-Webster. Wegner,D. M., & Wheatley,T. (1999). Apparentmentalcausation: Sources of the experience of will. AmericanPsychologist, 54, 480-492.

Aging and the Satisfaction of Psychological Needs

Peter G. Coleman
Departmentsof GeriatricMedicineand Psychology Universityof Southampton Deci and Ryan (this issue) provide new impetusto research on human motivation by revisiting the concept of psychological need, and specifying competence, relatedness and autonomy as three needs essential to goal-related activity. Their fundamental postulate is that "humansare active, growth-oriented organismswho are naturallyinclined towardintegration of their psychic elements into a unified sense of self and integrationof themselves into larger social structures" (this issue). As mightbe expectedthey provide evidence for this model from researchon human development.However they cite relatively little from the expandingfield of gerontology. All researchareasin personalityandsocial psycholof aging. ogy would benefit fromgreaterconsideration But this is particularlytrue of the study of human needs. The experience of aging can and does place harsh limitations on needs satisfaction. Research on processes of adjustmentand of continued development in adulthoodemphasizes the value of modifying asgoals in late life. Forexample,models of adaptation sociate flexibility of goals with successful aging and avoidance of depression (Brandtstiidter& Greve, 1994). But as Deci and Ryan's theory implies, this flexibility is limited by the persistingcharacterof the underlyingpsychological needs. Motivation theorists should make more effort in testing theirmodels on olderpeople. The reluctanceto give properacknowledgementto the study of aging is relianceon labopartlydue to psychology's traditional and student ratoryexperiment participation.Partly it reflects gerontophobia. Yet thereare some strikingexampleswhere the study of aging has led to new theory building. Research on the importance of subjective control for health and well-being, for example, began in Americannursinghomes (Langer, 1983). Likewise, growing attentionto the epidemiology of depressive illness amongolderpeople, helped the identificationof maintenanceof self-esteem, as well as social support, as centralto the preventionof depression.Subsequent researchhas establishedthat self-esteem generallyremains remarkably stable until late life, but shows progressive loss from the beginning of the ninth decade (Atchley, 1991). Social contacts show the same pat291