HANDBOOKOF

MORAL BEHAVIOR
AND
DEVELOPMENT

Volume 1: Theory

Edited by
WILLIAM
JACOB

M. KURTINES
L. GEWIRTZ

Florida lnternational University

m
1991

LAWRENCE ERLBAUM ASSOCIATES, PlJBLISIIERS

llillsdale, New Jersey
l love and London

Social Cognitive Theory of

Moral Thought and Action

Albert Bandura

ABSTRACT
A comprehensive theory of morality must explain how moral reasoning, in
conjunction with other psychosocial factors, govems moral conduct. Social cognitive theory adopts a cognitive interactionist perspective to moral phenomena.
Within this conceptual framework, personal factors in the forro of moral thought
and affective self-reactions, moral conduct, and environmental factors ali operate
as interacting deterroinants that inluence each other bidirectionally. Moral thinking is a process in which multidimensional rules or standards are used to judge
conduct. Situations with moral implications contain many decisional ingredients
that may be given lesser or greater weight depending upon the standards by
which they are cognitively processed and the particular constellations of events
in given moral predicaments. There are sorne culturally universal features to the
developmental changes of standards of conduct and the locus of moral agency.
These commonalities arise from basic uniforroities in the types of biopsychosocial changes that occur with increasing age in all cultures. A theory of morality
requires a broader conception than is provided by rationalistic approaches casi in
terros of skill in abstract reasoning. Moral conduct is motivated and regulated
mainly by the ongoing exercise of self-reactive inluence. Self-regulatory mechanisms, therefore, forro an integral part in the conception of moral agency in
social cognitive theory. Development of self-regulatory capabilities does not
create an invariant control mechanism within a person. Self-reactive inluences
do not operate unless they are activated, and there are many psychosocial processes by which self-sanctions can be selectively activated and disengaged from
transgressive conduct. Mechanisms of moral disengagement also play a central
role in the social cognitive theory of morality.

45

46

BANDUAA

Human morality is an issue of considerable import hoth individually and

collectively. lnternalization of a set of standards is integral to the achievement of
self-directedness anda sense of continuity and purpose in one's everyday life. In
the absence of personal standards and the exercise of self-regulatory influence,
people would behave like weathervanes, constantly shifting direction to conform
with whatever is expedient al a given moment. A shared morality, of course, is
vital to the humane functioning of any society. Many forms of behavior are
personally advantageous but are detrimental to others or infringe on their rights.
Without sorne consensual moral codes people would disregard each others' rights
and welfare whenever their desires come into social conflict. Societal codes and
sanctions articulate collective moral imperatives as well as influence social conduct. However, externa! sanctions are relatively weak deterrents because most
transgressive acts cango undetected. But people continuously preside over their
own conduct in countless situations presenting little orno externa! threat. So the
exercise of self-sanction must play a central role in the regulation of moral
conduct. Self-regulatory mechanisms forman integral par! of the conception of
moral agency presented in this chapter.
Most of the recent psychological interest in the domain of morality has centered on analyses of moral thought. The conspicuous neglect of moral conduct
reflects both the rationalistic bias of many theories of morality and the convenience of investigatory method. 11 is considerably easier to examine how people
reason about hypothetical moral dilemmas than to study their actual moral conduct. People suffer from the wrongs done to them however perpetrators might
justify their inhumane actions. The mechanisms goveming the self-regulation of
moral conduct involve much more than moral thought. Even the moral thought is
not solely an intrapsychic affair. The way in which moral principies are applied
in coping with diverse moral dilemmas varies, depending on situational imperatives, activity domains and constellations of social influence. lt is not uncommon for sophisticated moral justifications to subserve inhumane endeavors.
A comprehensive theory of morality must explain how moral reasoning, in
conjunction with other psychosocial factors, governs moral conduct. Social cognitive theory adopts an interactionist perspective to moral phenomena. Within
this conceptual framework, personal factors in the form of moral thought and
affective self-reactions, moral conduct, and environmental factors all operate as
interacting determinants that influence each other bidirectionally. Before presenting the social cognitive theory of morality, the cognitive structural conception will be analyzed briefly.
STAGE THEORIES OF MORAL REASONING
Stage theorists assume that different types of moral thinking appear in an invariant stage sequence from one uniform way of thinking to another. Piagetian
theory (1948) favors a developmental sequence progressing from moral realism,

1.

SOCIAL COGNITIVE THEORY OF MORALITY

47

in which rules are seen as unchangeable and conduct is judged in tenns of

damage done, to relativistic morality in which conduct is judged primarily by the
perfonner's intentions. In the latter stage, well-intentioned acts that produce
much hann are viewed as less reprehensible than ill-intentioned acts that cause
little hann. Moral absolutism stems from unquestioning acceptance of adult
prescripts and the egocentric outlook of young children; moral relativism develops from increasing personal experiences and reciproca! relationships with peers.
Following the lead of Piaget, Kohlberg developed an expanded cognitive
structural theory of morality that revitalized and altered the direction of the field.
Kohlberg ( 1969, 1976) postulates a six-stage sequential typology of moral rules,
beginning with punishment-based obedience, evolving through opportunistic
self-interest, approval-seeking confonnity, respect for authority, contractual legalistic observance. and culminating in principled morality based on standards of
justice. Changes in the standards of moral reasoning are produced by cognitive
conflict arising from exposure to higher levels of moral reasoning. Because the
stages constitute a fixed developmental sequence, individuals cannot acquire a
given fonn of moral reasoning without first acquiring each of the preceding
modes of reasoning in order. The presumption is that exposures to moral reasoning that are too discrepant from one's dominan! stage have little impact because
they are insufficiently understood to activate any changes. Judgmental standards
of lesser complexity are similarly rejected because they have already been displaced in attaining more advanced fonns of thinking. Views that diverge moderately above one's stage presumably create the necessary cognitive perturbations
which are reduced by adopting the higher stage of moral reasoning.

HIERARCHICAL MORAL SUPERIORITY

A universal, though not inbom, latent preference for higher modes of moral
thinking is posited to explain why people do not preserve their cognitive equilibrium simply by adhering to their own opinions and rejecting conflicting ones
(Rest, Turiel, & Kohlberg, 1969). What makes higher-stage reasoning morally
superior is not entirely clear. In thoughtful reviews of the stage theory of moralify. Locke (1979, 1980) identifies and refutes altemative bases of hierarchical
superiority. lt is not that higher stages of reasoning are cognitively superior
because, in most of their judgments, people do not use the highest mode of
thinking they understand. Such findings suggest that in many instances tests of
maturity in moral reasoning may be measuring personal preferences more than
level of competence in moral reasoning (Mischel & Mischel, 1976). On the
matter of stage progression, if people are actuated by an inherent drive for higher
ways of moral thinking it is puzzling why they rarely adopt the uppennost level
as their dominant mode even though they comprehend it (Rest, 1973). lt is
similarly arguable that higher stage reasons are morally superior. By what logical

48

BANDURA

reasoning is a morality rooted in law and order (stage 4) morally superior to one
relying on social regard and concem for others (stage 3)? Minorities oppressed
by a social order that benefits the majority and those subjected to the rule of
apartheid would not think so. Nor would writers who argue that social responsibility and concem for others should be the guiding rule of morality (Gilligan,
1982).
Higher-stage reasoning cannot be functionally superior because stages provide the rationale for supporting either side of a moral issue but they do not
prescribe particular solutions. Developmental stages determine the reasons given
for actions, not what actions should be talcen. Different types of moral thinking
can justify stealing, cheating on income laxes, and military bombing of foes.
lmmorality can thus be served as well, or better, by sophisticated reasoning as by
simpler reasoning. lndeed, the destructive social policies advocated by en-
lightened graduales of renowned academies is better explained by the social
dynamics of group thinking than by the collective level of moral maturity (Janis,
1972). When people reason about moral conflicts they commonly face in their
environment, Kohlberg and his associates find that moral reasoning is more a
function of the social influences operating in the situation than of persons' stages
of moral competence (Higgins, Power, & Kohlberg, 1984).
Kohlberg ( 1971 a) underscores the point that his hierarchical stages of reasoning are behaviorally nonprescriptive because they are concemed with the form of
reasoning not its content. However, the end point of moral reasoning, which
construes morality as justice, carries a fixed behavioral mandate. Unlike the
preceding stages, where it is acknowledged that a given type of moral thinking
can support either the transgressive or the conforming side of a moral issue, at
the end-point stage, thought is said to prescribe what courses of action are
morally right. Because movement through the stages is said to be achieved
naturally by the force of reasoning, empirical "is" thus becomes philosophical
"ought." Rationality dictates morality. The ordering of moral priorities is presumably revealed by switching perspectives in impartial cognitive role taking of
the position of each party in a moral conflict. However, as Bloom (1986) notes,
simple perspective shifting in no way guarantees consensus on what aspects of a
situation are morally relevant, the moral principies considered inherent in those
aspects, and which principie should be granted priority unless there is already
prior agreement on which principie should take precedence. lt should also be
noted that impartial role reversibility is imaginable in the abstract, but social
experiences creiile too many human biases for impartiality of view and universalization of interests to be achievable in reality. For example, no amount of
perspective shifting is likely to produce consensus among those who hold proand antiabortion views. The principie of freedom-for women and personalized
fetuses-provides justification for both moral stances. The consensus most likely to he achieved is agreeing to disagree.
Thc evidencc for the cultural university of the "is" has not gone uncontested

1.

SOCIAL COGNITIVE THEORY OF MORALITY

49

(Locke, 1979; Simpson, 1974). Other theorists argue that the moral idealization
in Kohlberg's theory reflects preference for Western views of moral adequacy
rather than objective standards or the dictates of reason (Bloom, 1986; Shweder,
1982). Societies that are less inclined toward ethical abstractions and idealization
of autonomy come out looking morally underdeveloped even though in their
moral conduct they may exhibit fewer inhumanities than Western societies that
are ranked as morally superior. Kohlberg' s ( 1971 b) prescriptive stance that moral education in the classroom should consist of moving children through the
stages of moral reasoning, even regardless of parental wishes, draws understandable heavy fire (Aron, 1977; Wonderly & Kupfersmid, 1980) and belies the
egalitarian characterization of the theory. The view of moral superiority as an
autonomous self operating above communal nonns and concems does not sit
well with many moral theorists.
Sorne moral philosophers, who hardly lack competence for principled reasoning, regard the principie of justice as only one among other moral principies that
either compete for the role of chief yardstick of morality or share a pluralistic
system of judgment (Carter, 1980; Codd, 1977). lf, however, principled reasoning is defined as using justice as the supreme judgmental rule it becomes a
conceptual truth incapable of empirical disproof (Peters, 1971 ). The common
finding is that adults comprehend different moral principies but use them selectively or in a complementary way, depending on the interplay of circumstances
and the domain of functioning. Moral development produces multifonn moral
thinking rather than follows a single developmental track.
Empirical analyses of Kohlberg's theory generally rely on a test that includes
only a few moral dilemmas sampling a narrow range of moral conlicts. They are
stripped of factors that can greatly complicate efforts to find moral solutions. To
contend that a few sketchy items verify moral truths is to invest a simple assessment tool with extraordinary revelatory power. A test that can offer only a
limited glimpse of moral predicamenls lacking systematic variation of ingredients may provide a shaky empirical basis on which to found a theory of
morality orto classify people into moral types. A person's propensity for principled moral reasoning will vary, depending on the infonnation included in the
depicted moral conflicts. For example, the moral dilemmas devised by Kohlberg
are ambiguous about the likely consequences of transgressive behavior. In the
transactional realities of everyday life, people nol only have to live with lhe
consequences of their moral choices, which they weigh anticipatorily in their
moral reasoning, hui experience of consequences is likely to affect their subsequent moral reasoning. Possible consequences are not taken lightly when moral
decisions can alter the course of one's life. lndeed, when infonnation about
different types of consequences are added even to hypothelical moral dilemmas
used to verify the stage theory, as the severity of personal consequences increases, people favor self-interesl over principled reasoning (Sobesky, 1983).
How often people offer principled solutions for moral conflicts may partly relecl

50

BANDURA

the gravity of rhe consequences lhey happen lo imagine for the sketchy portrayals
ralher than their compelence for principled reasoning.
The way in which hypolhetical moral dilemmas are structured can exert
considerable inluence on the priority given to different moral principies and the
amount of agreement obtained in moral judgment. To pit petty theft against
human life, as in the oft-quoted connict of the husband faced with stealing an
overpriced drug to cure his wife's cancer, will draw consensual judgments from
principled thinkers. Adding more substance to the moral dilemmas in the fonn of
complicating elements will elicit disagreement among principled thinkers over
which moral claims should take precedence (Bloom. 1986; Reed, 1987). Toe
moral dilemmas over which people agonize and feud often involve abhorrent
alternatives that do not lend themselves easily to moral solutions. We shall have
occasion lo review sorne of these later.

PRESCRIPTIVE AMBIGUITY OF ABSTRACT PRINCIPLES

Skeletonized abstrae! principies do not provide much guidance for judgmenl or

aclion until they are leshed out with relevan! details of concrete situations that
are inevitably laden wilh evaluative biases. For purposes of illustration, consider
the example given by Peters ( 1971) on judging what is jusi paymenl for service
rendered under a given set of circumstances. The abstract principie of justness
does not yield a unifonn answer. For inslance, what is a just fee for a surgeon?
Differenl people can arrive at different judgments from the same principie of
justness, depending on what factors they consider relevant and how they weight
them: such as the amount and expense of past training required, operating cosls,
lhe price of malpractice insurance, the effort and risks involved, the surgeon's
financia! needs, the benefits to patients, the patients' financia! status, and the
like. The judgmental thicket becomes even more ensnarled if social comparative
infonnation of remuneration for other occupations, such as poorly paid teachers
and exorbitantly paid superstar singers, is considered.
Given the prescriptive ambiguity of abstract principies, it is not surprising that
. cognitively facile people can find ways to serve lheir self-interests under the
cloak of justice or social contract. The advantaged members of a society have
considerable say in how juslice is defined at the operalional level. Social systems
lhat contain institutionalized inequities provide a set of social justifications that
make inequitable practices appear just ( Bandura, 1986). For example, people can
be persuaded that inequitably high compensation is deserved for activities that
carry substantial responsibility and risks, incur high personal cosls, require specialized skills thal are acquirable only through long arduous effort, and lhat
produce widespread social benefits. An abstrae! principie of justness does not say
much about where to set the houndary between just and unjust disparity in
compensation. Clearly, theories of moralily framed in tenns of moral abstrae-

1.

SOCIAL COGNITIVE THEORY OF MORALITY

51

tions cannot remain divorced from the social realities of how people go about
judging the moral dilemmas they confront in real life.

SEOUENTIAL TYPOLOGIES ANO MULTIFACETED

MORAL JUDGMENT
Stage theories assume that, over the course of development, moral judgments
change into a series of uniform types representing discontinuous stages. A major
problem with typologies is thal people hardly ever fil them. Because differing
circumstances call for differenl judgments and actions, unvarying human judgment is a rarity. A person's moral judgments typically rely on reasoning from
severa! different moral standards ralher than being based on only one type of
moral standard. So stage theorists have lo creale transitional calegories and
substages. Stage theories classify people into types according to their modal form
of reasoning, although any given individual usually displays coexisting mixtures
of reasoning thal span severa! "stages." Most people get categorized as being in
varying degrees of transition between stages.
People not only display substantial variability in their moral reasoning at any
given period, but many years elapse from the time they first adopta new standard
of morality and when they come to use it as a preferred one (Colby, Kohlberg.
Gibbs, & Lieberman, 1983). Fischer (1983) comments that such evidence is at
variance with stage theory, which depicts changes in thinking as occurring by
pervasive transformations of preceding modes of thought. Clearly, moral thought
is not hamstrung by a single cognitive structure that undergoes disjunctive developmental changes, nor does adoption of one standard pre-empt ali others. Rather
than exhibiting wholistic reorganization oftheir moral thinking, people gradually
adopt new moral standards, eventually discard simpler ones, and draw from
among a coexisting set of standards in judging different moral predicaments. The
mature mode of thinking is characterized by sensitivity to the di verse factors that
are morally relevant in any given situation. Choice of judgmental standards
depends partly on which factors are most germane to a particular moral problem.
One might question the practice of treating reasoning that draws on more than
one moral standard as evidence of moral immaturity evolving toward justness as
the ultimate standard of morality. Different moral standards are not necessarily
contradictory. Hence, adoption of a certain standard need not require jettisoning
another. To judge the morality of conduct by a system of complementary standards, such as justness and compassion, reflects a high level of moral reasoning
rather than transitional immaturity in thinking. lndeed, Peters ( 1966) argues that
justice is necessary but not sufficient for a moral system. He points out that
people can be brutal, but enlirely impartial or just in their brutality. A society that
subscribes to a morality that integrales standards of justness and compassion will
be more humane than a society concemed solely with justness.

DEVELOPMENTAL CHANGES IN MORAL JUDGMENT

There are sorne culturally universal features of the developmental changes of
standards of conduct and the locus of moral agency. These commonalities arise
from basic uniformities in the types of biopsychosocial changes that occur with
increasing age in all cultures. Growth of personal competencies and increasing
autonomy alter the types of morally relevant situations with which the growing
child must contend and the social structures within which these transactions take
place. A broadening social reality changes the nature of the moral concerns as
well as the social sanctions for transgressive conduct. Expanding moral choices
require more generalized and complex moral standards. Change in reasoning
from concrete to more abstract form with maturation and experience is also a
natural order of development that all theories acknowledge. No one would contend that young children begin as sophisticated reasoners and become progressively more simple minded as they mature. Nor do young children recognize the
prescripts of the social system before they recognize the prescripts of their
immediate caretalcers or companions. Another obvious natural order of development involves a broadening of perspective from individual to institutional prescripts for promoting human well-being. Change from externa) regulation to
increasing autonomy and self-regulation is still another natural order of development.
The major theoretical disputes center not on whether there are sorne universalities in the order of development, but on the validity of casting developmental
changes in discrete lock-step stages. Preparation for adult roles in society require
adoption of standards appropriate to the new social realities and set of roles. The
standards must serve as guides for conduct over an expanding range of moral
domains in a variety of settings involving multiple sources of intluence. Therefore, developmental change in moral standards is not simply a cumulative process. With increasing age, new standards are adopted rather than merely being
appended to earlier ones. People vary in the standards they teach, model, and
sanction with children of different ages.
The development and exercise of moral self-sanctions are rooted in human
relations and the way in which they are structured by the larger society. At first,
guidance of behavior is necessarily externa! and physically oriented. To discourage hazardous conduct in children who do not understand speech, parents must
rely on physical guidance. They structure situations physically to reduce the
likelihood of problem behavior, such as injurious aggression and, should it arise,
they try to check it by introducing competing activities or by disciplinary action.
Sometimes they pair simple verbal prohibitions with physical intervention, so
that eventually a "no" alone will suffice as a restrainer. At the earliest period of
development, there is little that is asked of young children and there is little they
can do that is transgressive. Their behavior is regulated and channeled mainly by
physical sanctions and verbal proxies for them.

52

1.

SOCIAL COGNITIVE THEORY OF MORAlllY

53

As children mature, they begin to pursue activities sorne of which inevitably

come into connict with others and with social norms. Such occasions elicit social
reactions designed to promote culturally valued behavior. Social sanctions increasingly replace physical ones as inluential guides for how to behave in
different situations. Parents and other adults explain standards of conduct and the
reasons for them. Social sanctions that disapprove transgressive acts and commend valued conduct add substance to the standards. lt is not long before children leam to discriminate between approved and disapproved forms of conduct
and to regulate their actions on the basis of anticipated social consequences
(Bandura & Walters, 1959; Sears, Maccoby, & Levin, 1957; Walters & Grusec,
1977).
Studies of socialization practices show that social sanctions combined with
reasoning foster self-restraints better than do sanctions alone (Parke, 1974;
Sears, Maccoby, & Levin, 1957). The reasoning that is especially conducive to
development of self-regulatory capabilities appeals to behavioral standards and
to empathetic concem for the adverse effects that detrimental conduct inlicts on
others (Bandura & Walters, 1959; Hoffman, 1977; Perry & Bussey, 1984).
Discipline that is used as an occasion for explaining rules of conduct is more
effective in instilling a generalized self-regulatory capability than if a specific act
is simply punished (LaYoie, 1974). Coercive threat may extrae! situational compliance, but cognitive guides provide a basis for regulating future conduct under
changing circumstances.
The social consequences that transgressors might bring on themselves through
their actions do not materialize if they avoid detection. But the injury and
suffering such actions cause others occur regardless of whether or not the wrongdoer is discovered. Thoughts of punishing consequences gain force through selfinterest. However, if the punishment is seen as avoidable or easily tolerable, it
may be less restraining than concems over possible injuries to others. There is
sorne evidence that negative sanctions accompanied by reasons arousing empathy for the victims tend to promote stronger self-restraints than those that try to
impress on wrongdoers that their conduct is likely to bring negative consequences to themselves (Walters & Grusec, 1977). The effectiveness of appeals to
empathy increases with age (LaYoie, 1974). Qualitative differences in the use of
reasoning are evident when comparing families of aggressively antisocial and
prosocial adolescents (Bandura & Walters, 1959). The former families emphasize the punishments misconduct can bring one, the latter families stress the
injury and suffering misconduct inlicts on others.
The extent to which the inluence of social sanctions is enhanced by reasoning
depends on its content and on a person's cognitive capabilities. Appealing to
abstractions is likely to be lost on young children who lack the experience to
comprehend them. They are swayed more by reasons centered on the tangible
consequences of misdeeds than on abstract rules (Parke, 1974). As children gain
social expenence and knowledge about what is right, they become more respon-

54

BANDURA

sive to ahstract appeals to rules and moral directives (Cheyne & Walters, 1970:
LaVoie, 1974).
Parcnls cannot always he presenl to guide lheir childrcn's hehavior. Successful socializalion requires gradual subslilution of symbolic and infernal conlrols for exlemal sanctions and demands. As moral standards are gradually
intemalized, they hegin to serve as guides and deterrents to conduct by the selfapproving and self-reprimanding consequences children produce for themselves.
Not only do the sanctions change from a social to a personal locus, but with
advancing age the range of moral considerations expands. As the nature and
seriousness of possible transgressions change with age, parents and other significan! adults in 1he child's life add new aspecls to the moral persuasion. For
example, lhey do nol appeal to legal arguments when handling preschoolers'
misconduct, but they do explain legal codes and penalties to preadolescents to
influence future behavior thal can have serious legal consequences. 11 is hardly
surprising that adolescenls are more likely than young children to consider legalilies in their reasoning about lransgressive acts.
People develop moral standards from a variety of influences. ley form
slandards for judging their own behavior partly on the basis of how significant
persons in their lives react to il. Parents and others are generally pleased when
children meet or exceed valued standards and disappointed when their peormances fall short of them. As a result of such differential evaluative reactions,
children evenlually come lo respond lo their own hehavior in self-approving and
self-critical ways, depending on how it compares wilh the evaluative standards
set by others.
Standards can be acquired through direct instruction in the precepts of conduct
as well as through the evaluative reactions of others toward one s actions (Liebert
& Ora. 1968; Rosenhan, Frederick, & Burrowes, 1968). In this form of transmission, moral slandards are drawn from lhe tutelage of persons in one's social
environment or !hose prescribed in the wrilings of influential figures. The moral
standards to which adults subscribe guide the type of morality they teach lo
children (Olejnik, 1980). As in other forms of influence, direct tuition is most
effective in fostering developmenl of standards when it is based on shared values
and is supported by social feedback to conduct.
People not only prescribe self-evaluative standards for others, they also exemplify them in responding to lheir own behavior. The power of modeling in
influencing standards of conducl is well documented (Bandura, 1986). Modeling
is a dynamic constructive process. People do not passively absorb standards of
conducl from whatever influences happen to impinge upon them. Rather, they
construct generic standards from the numerous evaluative rules that are prescribed, modeled, and taught. This process is complicated because those who
serve as socialization influencers, whether designedly or uninlentionally, often
display inconsistencies between what they practice and whal they preach. When
these two sources of social influence conflicl, example often oulweighs the

1.

SOCIAL COGNITIVE THEORY OF MORALITY

55

power of precept (Hildebrandt, Feldman, & Ditrichs, 1973; McMains & Liehert.
1968; Rosenhan, Frederick, & Burrowes, 1968). Moreover. people usually differ in the standards they model, and even the same person may model different
standards in different social settings and domains of conduct (Allcn & Liehert,
1969). Such discrepancies reduce the impact of modeling on the development of
personal standards. Exemplified standards also carry mre force when models
possess social power and status (Akamatsu & Farudi, 1978; Grusec. 1971; Mischel & Liebert, 1967).
Parents' level of moral reasoning predicts the level of their children's moral
reasoning (Holstein, 1973). Fine-grained analyses further reveal that children
model the form of the rules their parents use to integrate information in judging
the morality of transgressive conduct (Leon, 1984). Thus, if parents use simple
moral rules so do their children. whereas if parents rely on more complex
relativistic rules, their children do likewise. Parents, of course, are not oblivious
to their children's cognitive capabilities to grasp the moral implications of their
conduct. Parents react differently to their children's misconduct at different ages
(Denny & Duffy, 1974). They increase the complexity of their moral reasoning
as their children get older. The more complex the parent's moral reasons in
dealing with misconduct, the more elaborate is their children's moral reasoning.
Variation in social inluences contributes to developmental changes in what
factors are considered to be morally relevant and the relative weight they are
given.
Parents, of course, are not the exclusive source of childrens' standards of
moral judgments and conduct. Other adults, peers. and symbolic models, who
are by no meaos uniform in their moral perspectives, play inluential roles as
well. Children exposed to adult and peer models who exemplify connicting
standards adopt different standards of conduct than if adults alone set the standard, or if adults and peer models subscribe to the same standards (Bandura,
Grusec, & Menlove, 1967; Brody & Henderson, 1977). As we have already
seen, the power of modeling is attenuated by variation in modeled standards.
Peers can also exert strong inluence on the application of pre-existing moral
standards by evaluative justifications that make transgressive behavior morally
permissible. Even when the evaluative reactions of parents carry more weight
than those of peers, peers can win out because they are the ones who are present
in the behavioral situations to exert inluence on moral choices (Dombusch,
1987). The rnechanisms goveming the conditional application of moral standards
will be analyzed in a later section of this chapter.
To the developing child televised modeling, which dramatizes a vast range of
moral connicts that transcend viewers' immediate social realities, constitutes
another integral part of social leaming. The values modeled in print can similarly
impart moral standards for judging conduct (Walker & Richards, 1976). Symbolic modeling inluences the development of moral judgments by what it portrays as acceptable or reprehensible conduct, and by the sanctions and justifica-

56

BANDURA

tions applied to it. Clearly, a varied array of interacting societal inluences

contribute to the development of moral perspectives.

FAMIUAL ANO SOCIAL TRANSMISSION MODELS

Psychological theories have traditionally assumed that values, standards and
behavioral pattems are transmitted via parent-child relationships. In a provocative paper, Reiss (1965) contrasts theories based on the familia! transmission
model to those emphasizing transmission by broader social systems. He offers
severa! reasons why the familia! transmission model cannot adequately explain
socialization processes and outcomes. Assuming, at least, a 20-year procreation
difference between generations, a long time intervenes between parents' imparting values and standards to their children and when they can, in tum, pass on
!hose values to their own offspring. The long time lag between succeeding
descendants would produce a very slow rate of social change, whereas, in fact,
extensive society-wide shifts in standards and norrnative behavior often occur
within a single generation. The marked changes in sexual standards and practices
and cohabitation pattems within a relatively short time span are but one example.
Reiss, therefore, argues that the parent-child relationship cannot be the major
agency of cultural transmission. Rather, standards of behavior are primarily
disseminated by institutionally organized systems (e.g., educational, mass media, religious, political, and legal agencies) and regulated by collectively enforced sanctions. In Reiss's view, psychosocial changes originate primarily at
the social systems level, whereas changes emerging within the family are of
lesser social impact. Thus, forexample, racial segregation in public accommodations and infringements of voting rights were changed more rapidly by collective
protest and Supreme Court decisions than by waiting for prejudiced parents to
inculcate in their children more acceptant attitudes and values which they would
display toward minority groups when they became restaurateurs and motel operators 30 or 40 years later.
In accord with Reiss's main thesis, social cognitive theory assumes that
values and standards of conduct arise from diverse sources of inluence and are
promoted by institutional backing. Because social agencies possess considerable
rewarding and coercive power, collectively enforced sanctions can produce rapid
and widespread societal changes. However, social systems theory alone is insufficient to explain why there is often substantial variation in values and standards,
even within the same subcultures. Differences arise partly because institutional
prescriptions for the youth of a society must be implemented by parents, teachers, and community members. Those who, for whatever reason, do not subscribe
to the institutional codes, will undermine the broader social transmission effort.
Barring strong sanctions, parents often find new values discordant and resist

1.

SOCIAL COGNITIVE THEORY OF MORALITY

57

adopting them for sorne time. Families who are estranged from the mainstream
social systems also pay linle or no heed to institulional values.
A comprehensive theory of social transmission must also explain what produces and sustains the values, standards and behavioral nonns promulgated by
the cultural institutions. They are products of inluences wielded by members of
the society. Changes in social systems are often initiated by detennined dissenters acting on values modeled largely from individuals who have opposed prevailing social practices (Bandura, 1973; Keniston, 1968; Rosenhan, 1970). Dissenters create their own subsystems lo suppon lheir effons to refonn social
systems (King, 1958).
In discussing the limitations of personality theories of socialization, Reiss
states that, in such approaches, social change can arise only when there is a
breakdown in transmission between generations. This type of criticism is applicable to theories assuming that parental values are introjected by children in
toto and then are later passed on unmodified to their progeny. In social cognitive
theory, the adoption of values, standards and attributes is governed by a much
broader and more dynamic social reality. Social leaming is a continuous process
in which acquired standards are elaborated and modified, and new ones are
adopted. As previously mentioned, intemalization involves construction of standards from di verse sources of inluences rather than mindless mimicry. Children
repeatedly observe the standards and behavior paneros not only of parents, but
also of siblings, peers, and other adults (Bandura, Grusec, & Menlove, 1967;
Davidson & Smith, 1982). Moreover, the extensive symbolic modeling provided
in the mass media serves as another prominent extrafamilial source of inluence
(Liebert, Sprafkin, & Davidson, 1982). Hence, children's values and attributes
are likely to relect amalgams of these diverse sources, rather than simply the
unaltered familiai heritage. Even if psychosocial pattems arose solely from familial sources, significant changes could emerge across generations through
familial transmission. This is because the attributes and standards of the two
parents are rarely identical and siblings add further variety to what is modeled in
the familial environment. The attributes children develop are composites of
different features of parental and sibling values at each generation. Thus, children within the same family can develop somewhat different composite systems
of attributes and values that are neither solely those of the parents nor of the
siblings (Bandura, Ross, & Ross, 1963).
Sorne of the criticisms levied by Reiss against the familia) transmission model
are debatable, but his contention that social institutions often play a heavier role
in perpetuating and changing standards and psychosocial pattems than do familial inluences is well taken. However, an interactional theory that treats human
development as a product of both familia! and social system inluences holds
greater promise of furthering our understanding of the process than does a dichotomized view that pits one system against the other. This broader transmission model provides the vehicle for cultural evolution and the transmission of

58

BANDURA

cultural pattems holh within generalions and from one generation lo the next
( Boyd & R icherson. 1985).

MULTIFACETED NATURE OF MORAL JUDGMENT

AND ACTION
Adoption of intemal standards does nol necessarily encompass every domain of
activity or completely supplanl other fonns of control. Even the most principled
individuals may, in sorne domains of activity and under sorne circumstances,
regulate their behavior mainly by anticipated social or legal consequences. Moreover, during the course of development, children learn how to get around moral
consequences of culpable behavior that can gain them personal benefits. They
discover that they can reduce the likelihood of reprimands by invoking extenuating circumstances for their misdeeds (Bandura & Walters, 1959). As a result,
different types of vindications becorne salient factors in moral judgrnents. Even
very young children are quite skilled in using mitigating factors to excuse wrongdoing (Darley, Klosson, & Zanna, 1978). Later they leam to weaken, if not
completely avoid, self-censure for reprehensible conduct by invoking self-exonerating justifications. A theory of moral reasoning must, therefore, be concemed
as well with how exonerative moral reasoning can make the immoral inconsequential or even moral. We shall return later to the fonns that these mechanisms
of moral disengagernent take.
Stage theories attribute changes in moral judgrnent chiely to interna! reorganization of thought by stage-regulated mental perturbations for modifications channeled by latent preferences for higher moral stages. Such views make
light of the prominent role social inluences play in cultivating moral standards
and commitrnents. lt is not that stage theories take no notice of social factors.
They do, but they grant social inluences a narrow function-the views of others
serve mainly as externa! perturbators for autoregulated change. In fact, they do
much more. People impart moral standards and provide a great deal of social
support for moral commitments.
Developrnental trends obviously exist in moral reasoning and judgment, as
they do in everything else. But the conditions of social leaming are much too
varied to produce unifonn moral types. Even at the more advanced levels, sorne
behaviors come under the rule of law, others under social sanctions, and still
others under personal sanctions (Bandura, 1986). When statistical controls for
other causal factors are not applied, developrnental changes, which have been
attributed to stagelike unfolding of moral modes of thought, may relect changes
in general intelligence, infonnation-processing skills, educational level, and socialization practices with which moral reasoning correlates (Kay, 1982). Evidence of age trends, which every theory predicts, is often accepted as validating
stage theories of morality. Toe validity of stage propositions, however, demands

1.

SOCIAL COGNITIVE THEORY OF MORALITY

59

much more than age trends: They assume ( 1) That there is unifonnity of judgment when a person is at any given stage; (2) That a person cannot evaluate
conduct in tenns of a given moral principie without first adopting a series of
preceding principies; and (3) That attainment of a given judgmental principie
replaces preceding modes of thought by transfonning them. These presumptions
do not fare well when compared to emprica! findings.

SOCIAL CHANGE OF THE MORAL STANDARDS

OF STAGE THEORIES
Moral reasoning involves interpreting available information in moral predicaments against personal standards and situational circumstances for evaluating the
rightness or wrongness of conduct. The standards for moral reasoning, are much
more amenable to social inluence than stage theories would lead one to expect.
Numerous studies have been conducted in which children with differing moral
standards are exposed to opposing views of models who use either malevolent
intentions or severity of hann as the standard for judging the reprehensibility of
conduct. Such modeling influences alter how heavily children weigh intentions
and harm when they judge transgressive acts: Children who had previously
judged wrongdoing mainly by intentions judge conduct by the hann caused, and
those who previously evaluated wrongdoing by the amount of harm caused adopt
intentions as the principal indicant of reprehensibility (Bandura & McDonald,
1963; Cowan, Langer, Heavenrich, & Nathanson, 1969; Le Furgy & Woloshin,
1969). These altered moral perspectives are reflected in moral reasoning as well
as in the judgments made, they generalize across transgressive situations and
different pattems of intentions and damages, and they endure over time (Dorr &
Fey, 1974; Schleifer & Douglas, 1973). Although the modeled perspectives of
both adults and pecrs are persuasive, the moral reasoning of adults is usually the
more influential (Brody & Henderson, 1977; Dorr & Fey, 1974).
Evidence that children apply their altered moral perspective to new moral
predicaments and adhere to it over time attests to the significance of the achieved
effects. Changes promoted by structured social influence are sometimes called
into question by tautological arguments that cognitive change is a slow process,
so if changes are achieved in a short time they must not be "genuine." One can,
of course, point to instances where superficial influences produce circumscribed
change. But it is studies that effect generalized, enduring changes by influences
of sorne substance that speak most persuasively to the issue of whether moral
reasoning skills can be socially cultivated.
Efforts aimed at altering moral reasoning have relied heavily on the influence
of example. Exposure to others modeling an opposing view can alter moral
judgments in severa! ways. Moral judgment involves two separable processes.
Firstly. elements that are viewed as having moral relevance are selected from the

60

BANDURA

configurations of information available in given predicaments. Secondly, the

selected elements are weighted and integrated on the basis of moral rules for
judging conduct. By singling out certain elements in their moral reasoning,
models call attention to the factors lhe moral standards embody. The views
models express also provide supporting justifications for reweighing various
factors in making decisions about the wrongness of certain acts. Things that were
regarded as minor may become important, and visa versa. Evidence will be
presented later that models convey the moral rules as well as invest particular
elements with moral salience. In arcas of morality, for which society places a
premium on socially acceptable altitudes, public opinions may differ substantially from those that are privately held. Expression of moral convictions by
models provides the social sanctions for others to voice similar opinions. Modeling of opposing viewpoints can lhus effect changes in moral judgments through
attentional, cognitive, and disinhibitory mechanisms.
As in other areas of functioning, modeling intluences do not invariably alter
moral reasoning. When lack of effects do occur, they can result from either
comprehension deficits or performance preferences. People cannot be intluenced
much by modeled opinions if they do not understand them. Pre-existing knowledge and cognitive skills place limits on what can be leamed from brief exposure
to opposing opinions. There is substantial difference, however, between making
social influence dependent on knowledge of cognitive-processing skills than on
concatenated unitary thought. In social cognitive theory (Bandura, 1986), cognitive development is analyzed in terms of the sets of cognitive competencies
goveming given domains of functioning rather than discrete uniform ways of
lhinking.
When models voice opinions they transmit their ideas and preferences. But
modeling does not, itself, guarantee that the views so leamed will be articulated
by the learner. Where apparent unintluenceability retlects performance preferences, modeled standards have been learned but are simply not expressed because they are personally or socially disfavored. The case with which judgmental
standards can be shifted in one direction or another depends on the conceptual
skills lhey require and the social effects lhey have. In addition, judgmental
standards vary in how easily they can be discemed, which affects the facility
with which they can be leamed. lt is much easier to recognize damage than to
infer lhe historical antecedents or intentions of actions. When information about
intentions is provided in ways that aid its recall, young children use the intentions
of wrongdoers to judge culpability (Austen, Ruble, & Trabasso, 1977). The
claim, sometimes attributed to social leaming theory, that different moral standards are equally modifiable has no foundation. Sorne judgmental changes are
obviously more difficult to achieve than others. lt might also be noted in passing
that, contrary to what is sometimes alleged (Murray, 1983), social leaming
theory has never proposed lhe implausible assumption that erroneous reasoning
in matters of fact is just as producible by social influence as is accurate reason-

1.

SOCIAL COGNITIVE THEORY OF MORALITY

61

ing. Once children have leamed to reason in accord with evident fact (e.g.,
changing the shape of a clay hall does not change its mass), they will not revert to
fallacious reasoning by exposure to arguments they know to be untrue.

COGNITIVE CONFLICT ASAN AUTOMOTIVATOR

A theory of morality must explain both the motivators for cognitive change in
moral principies and the motivators for acting morally. Stage theorists address
the motivation for cognitive change but largely ignore the motivation for pursuing moral courses of action, sorne of which are self-denying while others may
bring adverse reactions from certain quarters. Standards alone do nol drive
action. Cognitive conflicl is posited as the major motivator of cognitive change
in stage theories. According to this equilibration mechanism (Piaget, 1960),
discrepancies between the cognitive schemas that children already possess and
perceived events create intemal conflict that motivates exploration of the source
of discrepancy until the intemal schemas are altered to accommodate to the
contradictory experiences. Events that differ markedly from what one knows or
expects are too bewildering and those that differ minimally are too familiar to
arouse interest and exploration. lt is moderately discrepant experiences that
presumably arouse cognitive conflict that prompts cognitive reorganization.
Piagetian theory thus proposes cognitive perturbations by moderately discrepant
experiences as the basic autornotivator for cognitive change.
Empirical tests of this type of autornotivator reveal that discrepancy of experience alone does not guarantee cognitive change (Kupfersmid & Wonderly, 1982;
Wachs, 1977). lndeed, if disparities between perceived events and mental structure were, in fact, automatically motivating, everyone should be highly knowledgeable about the world around them and continually progressing toward ever
higher levels of reasoning. The evidence does not seem to bear this out. Although motivation presumably springs from cognitive conflict between beliefs
held and the information conveyed by situations encountered, surprisingly little
effort has been made to verify the causal links between discrepanl influences,
indicants of intemal conflict, and the quest for new understanding. What little
evidence there is on this point shows that discrepant influences foster cognitive
changes but they seem unrelated to level of cognitive conflict (Zimrnerman &
Blom, 1983). This finding receives support from a study by Haan (1985) comparing the power of induced social and cognitive disequilibrium to change moral
reasoning. Cognitive disequilibrium has little effect on moral reasoning. However, the experiences of coping with social discord around issues o morality
produced changes in moral reasoning. The impact o divergent views seems to
stem from how persuasive they are than from how cognitively conflictful they
are. Role-playing higher levels o moral reasoning is no more effective in alter-

62

BANDURA

ing moral judgments than simply observing the same moral arguments being
modeled (Matefy & Acksen, 1976).
Simply demonstrating that children are unmoved either by what they already
know or by whal they do not comprehend because it exceeds their cognitive
capabilities is a mundane finding that can be explained without requiring an
elaborale automotivating mismatch rnechanism. Until objective criteria are specified for whal level of disparity constitutes moderate discrepancy, the equilibration model of self-motivation does not lend itself readily to empirical test. Langer
( 1969) maintains that it is the cognitive perturbations children spontaneously
produce by themselves rather than those extemally activated by discrepant events
that are the effective instigators of cognitive change. Moreover, the cognitive
conlict is said to be often unconscious, which makes it even less accessible to
study. Unless independenl measures of unconscious self-perturbation are provided, the posited incongruity motivator is incapable of verification.
As a rule, people do not pursue most activities that differ moderately from
what they know or can do. lndeed, if they were driven by every moderately
discrepant event encountered in their daily les they would be rapidly overwhelmed by innumerable imperatives for cognitive change. Effective functioning requires selective deployment of attention and inquiry. Self-motivation
through cognitive comparison requires distinguishing between standards of what
one knows and standards of what one desires to know. lt is the latter standards
that exert selective inluence over which of many activities that create discrepant
experiences will be actively pursued. A moderately discrepant experience, even
in areas of high personal involvernent, does not guarantee cognitive change.
When faced with views that are discordant from their own conceptions, people
often resolve the conlict by discounting or reinterpreting the discrepant infonnation rather than by changing their way of thinking. lt has been shown in other
domains of cognitive functioning that the degree of cognitive change generated
by exposure to discrepant information is better predicted from the credibility of
those voicing discrepant views than from the degree of disparity per se. Sources
of high credibility produce increasing cognitive change the more their views
differ from those held by the person being inluenced whereas, for sources of low
credibility, the more discrepant their view, the more they are rejected (Bergin,
1962; McGuire, 1985). Social factors exert a powerful inluence on how discrepant conceptions are cognitively processed and received.
Sorne efforts have been made to test the equilibration mechanism of developmental change within Kohlberg's framework by exposing children to moral
arguments that increasingly diverge from the views children already hold. In the
initial investigations of stage constraints on moral change, children were presented with a few hypothetical moral dilemmas and they were given confcting
moral advice by persons using reasons from different stages (Rest, Turiel. &
Kohlberg, 1969; Turiel, 1966). The investigators report that children rejecl modeled opinions below their dominan! mode of thinking, are unaffccted by those

1.

SOCIAL COGNITIVE THEORY OF MORALITY

63

that are too advanced, but are likely to adopt modeled views one stage above
their own.
Subsequent research indicates that the restricted changeability o moral reasoning may lie more in how the modeling influence was used than in constraints
o children's stages. lt is unreasonable to expect entrenched moral perspectives
to be altered markedly by a transitory influence, especially i presented in a weak
form. Theories predicting null results should apply social influences in their most
powerful form because one can easily fail to produce cognitive changes by using
weak influences. Children do not remember the essential details o moral situations presented to them briefly. but they show good recall with greater exposure
(Austen, Ruble, & Trabasso. 1977). Fleeting information that goes by unrecognized or unrecalled cannot affect moral thinking. In the studies conducted by
Rest and Kohlberg. not only is the modeling influence unusually brief, but the
models disagree with their views by advocating opposing solutions. Although
results are not entirely uniform (Walker, 1983), models who are consistent in
how they judge different moral predicarnents generally have greater impact on
children's moral reasoning than do models who disagree with each other (Brody
& Henderson, 1977; Keasey, 1973). When the modeled views are consistent,
children's moral perspectives are changed more by exposure to moral reasoning
two stages above their own than by reasoning one stage more advanced (Arbuthnot, 1975; Matey & Acksen, 1976). These findings are in accordance with
substantial evidence in social psychology cited earlier that the more discrepant
persuasive reasoning is from one's own views, the more one's attitudes change.
lmmaturity, of course, places sorne limits on the power of discrepant influences.
Young children cannot be influenced by reasoning so advanced that it is completely incomprehensible to them.
Children also adopt modeled modes of reasoning labeled as more primitive in
the stage hierarchy, but the findings are mixed on how well they adhere to them
over time. Here, too, the variable adherence may refiect more how persuasively
modeling is used than stage constraints. The views of a lone model, or one who
disagrees, can be easily discounted as atypical. lt is consensual multiple modeling that carries the strong persuasive impact necessary to override pre-existing
orientations. lndeed, the propensity of children to pattem their preferences after
models increases as the leve! of consensus among models increases (Perry &
Bussey, 1979). Viewers are likely to conclude that i everyone firmly believes
something. it must have rnerit.
lt could be argued that judging by the intentionality of actions does not
necessarily represen! a higher level of reasoning than judging by the consequences that fiow from the acts. In judging the morality of nuclear strategies, for
example, the awesome destructiveness o a nuclear attack should be the overriding consideration, rather than the intentions o the launchers o such attacks. To
give utmost priority to the devastating consequences o a nuclear strike would
hardly be considered "regressive" or "primitive" thinking. Rather, to judge as

64

BANDURA

morally well intended, nuclear strikes that can take a massive toll on human life
and render much of the planet uninhabitable would reflecl an unthinking reverence for intention and personal principie.
Results showing thal there are sorne age trends in moral judgment, that
children fail to adopt standards they do not fully comprehend or about which
there is disagreernent, and that they are disinclined to stick to views considered
immature for their age can be adequately explained without requiring stage
propositions. Evidence that moral reasoning can be changed by exposure to
modes of thinking that invert or skip stages is al variance with the contention of
stage theory that, to alter how one thinks about moral issues, one has to pass
through an invariant sequence of stages, each displacing lower ones along the
way from which there can be no retum. Acknowledging the intraindividual
diversity of moral reasoning, sorne stage theorists (Rest, 1975) have redefined
stage progression as a shifting distribution of mixed modes of thinking that are
affected by many environrnental factors. Such a view reduces the mismatch
between the theoretical conception and the actuality. But it raises the issue of
what purpose is served by adhering to a stage doctrine stripped of its major
defining properties of change by structural displacernent, steplike discontinuity,
unifonnity of cognitive structure, and judgrnent unarbitrated by either the situational factors or the domain of activity? lf stage progression is recast as a
multifonn gradualistic process cultivated by environrnental influences, such a
model differs linle from developrnental theories that do not invoke stages.
Apparent deficiencies in moral reasoning, often anributed to cognitive limitations or insensitivity to certain moral issues, have also been shown to depend
partly on how moral thought is assessed (Chandler, Greenspan, & Barenboim,
1973; Gutkin, 1972; Hatano, 1970; Leming, 1978). The same individuals express different types of moral judgments, depending on how morally relevant
factors are presented, whether children judge verbal accounts or behavioral portrayals of transgressions, whether they judge common or outlandish moral conflicts, whether they reveal their moral orientations in abstract opinions or in the
severity of the sanctions they apply to different acts, and whether they judge the
transgressive acts of others or give moral reasons for how they would behave if
faced with similar moral dilemmas. The view that stages constrain people to
think in a unifonn way receives little support in the notable variability of moral
thinking even with small changes in how moral conflicts are presented and how
judgments are rendered.
MORAL JUDGMENT AS APPLICATION
OF MULTIDIMENSIONAL RULES
In the social cognitive view, moral thinking is a process in which multidimensional rules or standards are used to judge conduct. Situations with moral implications contain many decisional ingredients that not only vary in importance

1.

SOCIAL COGNITIVE THEORY OF MORALITY

65

but may be given lesser or greater weight, depending on the particular constellation of events in a given moral predicament. Among the many factors that enter
into judging the reprehensibility of conduct are the nature of the transgression; its
base rate of occurrence and degree of norm variation; the contexts in which it is
perfonned and the perceived situational and personal motivators for it; the immediate and long-range consequences of the actions; whether it produces personal
injury or property damage; whether it is directed al faceless agencies and corporations or at individuals; the characteristics of the wrongdoers, such as their age,
sex, ethnic, and social status; and the characteristics of the victims and their
perceived blameworthiness. In dealing with moral dilemmas, people must extract, weigh, and integrate the morally relevant information in the situations
confronting them.
We saw earlier that moral rules or standards of conduct are fashioned from
varied social sources including precepts, evaluative social reactions, and models
of moral commitments. From such diverse experiences people learn which factors are morally relevant and how much weight to attach to them. With increasing experience and cognitive competence, moral judgments change from singledimensional rules to multidimensional rules of conduct. The more complex rules
involve configura! or relativistic weighting of morally relevant information. That
is, factors that are weighed heavily under sorne combinations of circumstances
may be disregarded or considered of lesser import under a different set of
conditions.
Researchers who approach moral thinking as a process of information integration havr, studied the rules by which children weigh and combine information
about different factors in making moral judgments (Kaplan, 1989; Lane & Anderson, 1976; Surber, 1985). Much of this research has examined how children
combine information about intentions and consequences in judging transgressive
actions. When presented with situations varying in degree of maliciousness and
hann, children do not reason dichotomously. that is, using hann when young and
intention when older, as proposed by Piagetian theory. Rather, they apply varied
integration rules in which the different factors are combined additively with the
same absolute weight regardless of other information, or configurally in which
the amount of weight given to a factor depends on the nature of another factor.
However, additive rules seem to predominate (Leon, 1980, 1982). The form of
the integration rule used vares more across individuals than ages. Parental
modeling accounts for a large part of the individual differences in complexity of
moral decision making (Leon, 1984). Parents differ in how they integrate information into moral judgments, ranging from a simple rule based solely on damage
done, to a composite linear rule combining intent and damage, to a more complicated configura! rule that weighs damage differentially, depending on intent. In
their own cognitive processing of information regarding the morality of conduct,
children model their parents' rules in form and complexity.
Children at ali ages use both intention and harm in forming their judgments,

66

BANDURA

with developmental changes in the weight given these factors being gradual
rather than stagelike (Grueneich, 1982; Surber, 1977). Analyses that separate
what judgmental factors are selected from constellations of events, what weight
is given to the factors that are singled out, and the decision rule by which they are
combined are especially well suited to darify developmental changes in moral
reasoning. Multifaced analyses of judgments of factorial combinations of different types of information are more informative than coding verbal protocols or
selecting global attributions of whether outcomes are attributed to personal
causation or to externa! circumstances.
Kaplan (1989) has examined the integrative rules of moral decision making
with scenarios that include different combinations of factors characterizing the
various stages of Kohlberg's theory. For example, a transgressive act may be
portrayed as both fulfilling a social obligation and serving a moral principie, or
as inlicting punishment but fulfilling a social obligation. People's judgments
revea! how much weight they give to the different factors and the type of
integration rule they use. The findings show that students combine factors from
different stages in their moral decision making rather than reason in terms of a
particular stage-constrained moral rule. Efforts to develop morality based on
Kohlberg's framework rely on guided moral argumentation that provides exposure to more mature levels o reasoning. This form of moral training presumahly improves cognitive skill in making decisions about moral problems rather
than inculcates particular values. Kaplan found that such training is more likely
to inculcate values than to increase the complexity of moral reasoning. However,
students can leam to combine information in configura! or relativistic moral rules
through discussions of nonmoral problems in which they come to understand that
particular factors may be given more or less weight, depending on the configuration of other elements.
More work remains to be done on how people deal with large sets of morally
relevant factors, how social inluences alter the weight they give to different
factors, what types of combinatoria! rules they use, and how these different
aspects of moral judgment change with development. Humans are not ali that
adept at integrating diverse information (Kahneman, Slovic, & Tversky, 1982).
As in other judgmental domains, when faced with complexities most people
probably fall back on judgmental heuristics that give too much weight to a few
moral factors while ignoring other relevant ones. Consistent social feedback can
produce lasting changes in the rules used to judge the morality of action
(Schleifer & Douglas, 1973). However, in everyday life social consensus on
morality is difficult to come by. thus creating ambiguity about the correctness of
judgment. In the absence of consistent feedback, reliance on convenient
heuristics may become routinized to the point where moral judgments are rendered without giving much thought to individuating features of moral situations.
The susceptibility of moral judgment to change depends in part on the effects of

1.

SOCIAL COGNITIVE THEORY OF MORALITY

67

the actions it fosters. Over time, people alter what they think by experiencing the
social effects of their actions.

RELATION BETWEEN MORAL REASONING AND

CONDUCT
An issue that has received surprisingly little attention is the relationship between
moral reasoning and moral conduct. The relationship between thought and conduct is mediated through the exercise of moral agency (Bandura, 1986; Rottschaefer, 1986). The nature of moral agency will be examined shortly. The study
of moral reasoning would be of limited interest if people's moral codes and
thoughts had no effect on how they behaved. In the stage theory of moral
maturity the form of moral thought is not linked to particular conduct. This is
because each level of moral reasoning can be used to support or to disavow
transgressive conduct. People may act prosocially or transgressively out of mutual obligation, for social approval, for duty to the social order, or for reasons of
principie. A person's level of moral developrnent may indicate the types of
reasons likely to be most persuasive to that person, but it does not ensure any
particular kind of conduct.
The implications for human conduct of the stage theory of moral maturity are
difficult to test empirically because conflicting claims are made about how moral
reasoning is linked to behavior. On the one hand, it is contended that the level of
moral reasoning does not sponsor a particular kind of behavior (Kohlberg,
1971a). 1be theory is concemed with the form of the reasoning not with the
moralness of the conduct. Hence, in studies designed to alter moral perspectives
through exposure to moral argument, the same leve! of reasoning is used, for
example, for and against stealing (Rest, Turiel, & Kohlberg, 1969). 0n the other
hand, a positive relationship is clairned between leve! of moral reasoning and
moral conduct-the higher the moral reasoning, the more likely is moral conduct, and the greater is the consistency between moral judgment and conduct
(Kohlberg & Candee, 1984).
Studies on whether stages of moral reasoning are linked to characteristic types
of conduct are inconsistent in their findings (Blasi, 1980; Kurtines & Greif,
1974). Sorne researchers report that moral conduct is related to the level of moral
reasoning, but others have failed to find strong evidence of such a relationship.
Sorne of the studies routinely cited as corroborating such a link have not withstood replication. Others are seen under close scrutiny as contradicting it or as
uninterpretable because of methodological deficiencies (Kupfersmid & Wonderly, 1980). Moreover, relationships may disappear when controls are applied
for other differences between persons at varying levels of moral reasoning, such
as general intelligence (Rushton, 1975).

68

BANDURA

Efforts to verify the link between moral thought and action have raised disputes about the designation of moral conduct. Kohlberg and Candee ( 1984) argue
that it is perfonners' intentions that define their actions as moral or immoral. lf
the morality of conduct is defined by the intentions voiced by transgressors, then
most behavior that violates the moral codes of society will come out laundered as
righteous. People can easily find moral reasons to redefine their misdeeds as
really well-intentioned acts. They become more adept at self-serving justifications as they gain cognitive facility. Presumed intent always enters in as one
factor in the social labeling of behavior (Bandura, 1973), but intention is never
used as the decisive definer of conduct. A robber who had a good intent would
not thereby transfonn robbery into nonrobbery. A theory of morality must explain the detenninants and the mechanisms goveming transgressive conduct, not
only how perpetrators justify it. This requires a broader conception of morality
than is provided by a rationalistic approach cast in terms of skill in abstract
reasoning. Affective factors play a vital regulative role in moral conduct.

CONCEPTION OF MORAL AGENCY IN TERMS

OF SELF-REGULATORY MECHANISMS

Moral self-regulation is not achieved by disembodied moral thought or by a feat

of willpower. Explanation of the relation between moral reasoning and conduct
must specify the psychological mechanisms by which moral standards get translated into actions. In social cognitive theory (Bandura, 1986), transgressive
conduct is regulated by two major sources of sanctions-social sanctions and
intemalized self-sanctions. Both control mechanisms operate anticipatorily. In
control arising from social sanctions, people refrain from transgressing because
they anticipate that such conduct will bring them social censure and other adverse
consequences. In self-reactive control, they behave prosocially because it produces self-satisfaction and self-respect and they refrain from transgressing because such conduct will give rise to self-reproof.
For reasons given earlier, moral conduct is motivated and regulated mainly by
the ongoing exercise of self-reactive influence. The major self-regulatory mechanism, which is developed and mobilized in concert with situational factors,
operates through three major subfunctions. These include self-monitoring of
conduct; judgment of conduct in relation to personal standards and environmental circumstances; and affective self-reaction. To exercise self-influence, people
have to monitor their behavior and the situational circumstances in which they
find themselves enmeshed. The process of self-monitoring is not simply a mechanical audit of one's perfonnances and social instigators. Pre-existing conceptions and affective states can bias how one's actions and the instigators for it are
perceived and cognitively processed.
Self-monitoring is the first step toward exercising influence over one's con-

1.

SOCIAL COGNITIVE THEORY OF MORAUTY

69

duct but, in itself, such infonnation provides little basis for self-directed reactions. Actions give use to self-reactions through a judgmental function in which
conduct is evaluated against moral standards and environmental circumstances.
We saw earlier that situations with moral implications contain many judgmental
ingredients that not only vary in importance but may be given lesser or greater
weight, depending on the particular constellation of events in a given moral
predicament. In dealing with moral dilemmas, people must, therefore, extract,
weight, and integrate the morally relevant infonnation in the situations confronting them. Factors that are weighted heavily under sorne combinations of circumstances may be disregarded or considered of lesser import under a different set of
conditions. This process of moral reasoning is guided by multidimensional rules
for judging conduct.
Self-regulation of moral conduct involves more than moral thought. Moral
judgment sets the occasion for self-reactive influence. Affective self-reactions
provide the mechanism by which standards regulate conduct. The anticipatory
self-respect and self-censure for actions that correspond with, or violate personal
standards serve as the regulatory influences. People do things that give them selfsatisfaction anda sense of self-worth. They ordinarily refrain from behaving in
ways that violate their moral standards because it will bring self-condemnation.
There is no greater punishment than self-contempt. Anticipatory self-sanctions
thus keep conduct in line with interna! standards.
There is a difference between possessing self-regulatory capabilities and
being able to apply them effectively and consistently under the pressure of
contravening influences. Effective self-regulation of conduct requires not only
self-regulatory skills but also strong self-belief in one's capabilities to achieve
personal control. Therefore, people's belief in their efficacy to exercise control
over their own motivation, thought pattems and actions also plays an important
role in die exercise of human agency (Bandura, 1986). The stronger the perceived self-regulatory efficacy, the more perseverant people are in their selfcontrolling efforts and the greater is their success in resisting social pressures to
behave in ways that violate their standards. A low sense of self-regulatory
efficacy heightens vulnerability to social pressures for transgressive conduct.
lf people encounter essentially similar constellations of events time and again,
ttiey do not have to go through the same moral judgmental process of weighting
and integrating moral factors each time before they act. Nor do they have to
conjure up self-sanctions anticipatorily on each repeated occasion. They routinize their judgment and action to the point when they execute their behavior
with little accompanying thought. However, significant changes in morally relevant factors reactivate evaluative processes for how to behave under the altered
circumstances.
In social cognitive theory, the self is not disembodied from social reality.
People make causal contribution to their actions and the nature of their environment by exercising self-influence. However, in accord with the model of re-

70

BANDURA

cprocal causation, social influences affect the operation of the self system in al
least three major ways. They contribute importantly to the development of selfregulatory competence. Analyses of regulation of moral action through affective
self-reaction distinguish between two sources of incentive motivation operating
in the process. There are the conditional self-generaled incentives thal provide
guides and proximal motivators for moral courses of action. Then there are the
more distal social incentives for holding lo a moral syslem. Thus, the second way
in which social influences contribute to morality is by providing collective support for adherence to moral standards. The third way in which social realities
affect moral functioning is by facilitating selective activation and disengagement
of moral self-regulation. The forms that the various psychosocial mechanisms of
moral disengagemenl take are analyzed in the sections thal follow. 11 might be
noted in passing that the wealth of particularized knowledge on how self-regulatory competence is acquired and exercised (Bandura, 1986) stands in stark
contrast to the ill-defined intemali1,ation process commonly invoked in theories
of morality. A complete theory of rnorality, whatever its theoretical allegiance,
musl include these verified mechanisms of self-regulation.

INTERPLAY OF PERSONAL ANO SOCIAL SANCTIONS

The self-regulation of conduct is not entirely an intrapsychic affair as the more

radical forms of cognitivism might lead one to believe. Nor do people operate as
autonomous moral agents impervious to the social realities in which they are
enmeshed. Social cognitive theory favors a causal model involving triadic reciproca! causation (Bandura 1986). The three constituent sources of influencehehavior, coxnition and other personalfactors, and environmental influencesall operale as interacting determinanls of each other. From this interactionist
perspective, moral conducl is similarly regulated by a reciprocity of influence
between thought and self-sanctions, conduct, and a network of social influences.
After standards and self-reactive functions are developed. behavior usually produces lwo sets of consequences; self-evaluative reactions and social effects.
These two sources of consequences may operate as complemenlary or opposing
influences on behavior.
Conduct is most congruent with moral standards when transgressive behavior
is not easily self-excusable and the evaluative reactions of significant others are
compatible with personal standards. Under conditions of shared moral standards,
socially approvable acts are a source of self-pride and socially punishable ones
are self-censured. To enhance the compatibility between personal and social
influences, people generally select associates who share similar standards of
conduct and thus ensure social support for their own system of self-evaluation
(Bandura & Walters, 1959; Elkin & Westley, 1955; Emmons & Diener, 1986).

1.

SOCIAL COGNITIVE THEORY OF MORALITY

71

Diversily of slandards in a sociely. lherefore, does nol necessarily creale personal conlicl. Selective association can forge consistency out of diversily.
Behavior is especially susceptible to externa! influences in the absence of
slrong countervailing interna! standards. People who are not much commiued lo
personal slandards adopl a pragmatic orientation, tailoring their behavior lo fil
whatever the situation seems to call for (Snyder & Campbell, 1982). They become
adept al reading social situations and guiding lheir actions by expediency.
One type of conlict between social and self-produced consequences arises
when individuals are socially punished for behavior they highly value. Principled
dissenters and nonconformists often find themselves in this predicamenl. Here,
the relative strength of self-approval and social censure determine whether the
behavior will be restrained or expressed. Should the threatened social conscquences be severe, people hold in check self-praiseworthy acts in risky situations
but perform them readily in relatively safe settings. There are individuals, however, whose sense of self-worth is so strongly invested in certain convictions that
they will submit to prolonged maltreatment, rather than accede to what they
regard as unjust or immoral.
People commonly experience conlicts in which they are socially pressured to
engage in behavior that violates their moral standards. When self-devaluative
consequences outweigh the benefil for socially accommodating behavior, the
social influences do not have much sway. However, the self-regulation of conduct operates through conditional application of moral standards. Self-sanctions
can be weakened or nullified by exonerative moral reasoning and social circumstances. People display different levels of detrimental behavior and offer different types of moral reasons for it, depending on whether they find themselves in
social situations that are conducive to humane orto hurtful conduct (Bandura,
Underwood, & Fromson, 1975). Because almost any conduct can be morally
justified, the same moral principies can support different actions, and the same
actions can be championed on the basis of different moral principies. However,
moral justification is only one of many mechanisms that affect the operation of
moral standards in the regulation of conduct.

SELECTIVE ACTIVATION ANO DISENGAGEMENT

OF MORAL CONTROL
Development of self-regulatory capabilities does not create an invariant control
mechanism within a person, as implied by theories of internalization that incorporate entities such as conscience or superego as continuous internal overseers of
conduct. Self-reactive influences do not operate unless they are activated, and
there are many processes by which self-sanctions can be disengaged from inhumane conduct (Bandura, 1986, 1990). Selective activation and disengagement
of interna! control permits different types of conduct with the same moral stan-

72

BANDURA
MUR~

.. ,,..,w,r,..c..

JUST1fl(AT1()H

fli&lllAIIV(

ot Hl.MAtrilll AIION
Al HUBIJTt()N

Of Bt.Mf:

tUlltSlOUtfrtl.fS

lABlllfil(,

REPREHENSIBLE
CONOUCT

1CH>A1frfu

(IA M1',(_0f,l':,fHt)1f\ll(, IH(

(l)MP,t.Ho':,()frt

(uPHfWISl1(

'

DETRIMENTAL

--VICTIM

EfF~CTS
J

'
l' 01SPLACIME~l

'
'

el A(SPQtr18lll11' :

OlffUS10N ()f R(Sfl'ONSIBM.llY

'
FIG. 1.1. Mechanisms through which interna! control is selectively
activated or disengaged from reprehensible conduct at different points
in the regulatory process (Bandura, 1986).

dards. Figure 1.1 shows the four major points in the self-regulatory process at
which interna! moral control can be disengaged from detrimental conduct. Selfsanctions can be disengaged by reconstruing conduct, obscuring causal agency,
disregarding or misrepresenting injurious consequences, and blaming and devaluating the victims.
These mechanisms of moral disengagement have been examined most extensively in the expression of violent conduct. But selective disengagement of moral
self-sanctions is by no meaos confined to extraordinary inducements to violence.
People often experience contlicts in which behavior they themselves devalue can
serve as the means for attaining valued benefits. As long as self-sanctions override the force of externa! inducements, behavior is kept in line with personal
standards. However, in the face of strong externa! inducements such contlicts are
often resolved by selective disengagement of self-sanctions. This enahles otherwise considerate people to perform self-serving activities that have detrimental
social effects. The processes by which self-regulatory capabilities are acquired
have been examined in sorne detail. However, the selective activation and disengagemenl of interna! control, which have considerable theoretical and social
import, have only recently received systematic study.

Moral Justification
One set of disengagement practices operates on the construal of the behavior
itself. People do not ordinarily engage in reprehensible conduct until they have
justified to themselves the morality of their actions. What is culpable can be
made righteous through cognitive reconstrual. In this process, detrimental con-

1.

SOCIAL COGNITIVE THEORY Of MORAUTY

73

duct is made personally and socially acceptable by portraying it in the service of

moral purposes. People then act on a moral imperative.
Radical shifts in destructive behavior through moral justification is most
strikingly revealcd in military conduct. People who have been socialized to
deplore killing as morally condemnable can be transformed rapidly into skilled
combatants, who may feel little compunction and even a sense of pride in taking
human life. Moral reconstrual of killing is dramatically illustrated in the case of
Sergeant York, one of the phenomenal fighters in the history of modero warfare
(Skeyhill, 1928). Because of his deep religious convictions, he registered as a
conscientious objector, but his numerous appeals were denied. At camp, his
battalion commander quoted chapter and verse from the Bible to persuade him
that under appropriate conditions it was Christian to fight and kili. A marathon
mountainside prayer finally convinced him that he could serve both God and
country by becoming a dedicated fighter.
The conversion of socialized people into dedicated combatants is achieved not
by altering their personality structures, aggressive drives, or moral standards.
Rather, it is accomplished by cognitively restructuring the moral value of killing,
so that it can be done free from self-censuring restraints (Kelman, 1973; Sanford
& Comstock, 1971). Through moral sanction of violen! meaos, people see themselves as fighting ruthless oppressors who have an unquenchable appetite for
conquest, protecting their cherished values and way of life, preserving world
peace, saving humanity from subjugation toan evil ideology, and honoring their
country's intemational commitments. The task of making violence morally defensible is facilitated when nonviolent options are judged to have been ineffective, and utilitarian justifications portray the suffering caused by violent counterattacks as greatly outweighed by the human suffering inflicted by the foe.
Over the years, much reprehensible and destructive conduct has been perpetrated by ordinary, decent people in the name of religious principies, righteous
ideologies, and nationalistic imperatives. lndividuals espousing high moral principies are inclined to resist arbitrary social demands to behave punitively, but
they will aggress against people who violate their personal principies (Keniston,
1970). Throughout history, countless people have suffered at the hands of selfrighteous crusaders bent on stamping out what they consider evil. Rapoport and
Alexander ( 1982) document the lengthy, blood-stained history of holy terror
wrought by religious justifications. Acting on moral or ideological imperatives
relects a conscious offense mechanism, notan unconscious defense mechanism.
Although moral cognitive restructuring can be easily used to support selfserving and destructive purposes, it can also serve militant action aimed at
changing inhumane social conditions. By appealing to morality, social reformers
are able to use coercive, and even violent, tactics to force social change. Vigorous disputes arise over the morality of aggressive action directed against institutional practices. Powerholders often resist, by forcible meaos if necessary,
making needed social changes thal jeopardize their own self-interests. Resistance

74

BANOURA

to warranted changes invites social activism. Challengers define their militan!

actions as morally justifiahle means to eradicate harmful social practices.
Powerholders, in tum. condemn such activism as representing impatient resort to
violent solutions or efforts to coerce changes that lac:k popular support.
There are those who argue for a high moral threshold as a criterion of coercive
activism (Bickel. 1974). In this view, unlawful conduct is justified only if
traditional rneans have failed and those who break the law do so publicly and
then willingly accept the consequences of their transgressive behavior. In this
way, specific unjust practic:es can he challenged while maintaining respect for
the judicial process itself. lt is presumably the suffering endured by the aggrieved
pmtesters that shakes the moral complacency of compassionate citizens and,
thereby, mobilizes the widespread support required to force warranted reforms.
lf challengers demand amnesty for unlawful conduct, it not only defeats the
purpose of conscientious disohedience, but it is morally wrong. lf individuals do
not have to accept responsibility for their actions, violent tactics and t.hreats of
force will be quickly used whenever grievances arise. lt is further argued that
illegal defiance of the rules in a representative society fosters contempt for the
principie of democratic authority. Anarchy would flourish in aclimate in which
individuals acted on private moral principies and considered coercive tactics
acceptable whenever they disliked particular social practices or policies representing majority decisions.
Challengers refute such moral arguments by appeal to what they regard as a
higher leve! of morality, derived from communal concems. Their constituencies
are expanded to include ali people, hoth al home and ahroad, victimized either
directly or indirectly by injurious social practices. Challengers argue that when
many people benefit from a system that is deleterious to disfavored segments of
the society the harmful social practices secure widespread public support. From
the challengers' perspective, they are acting under a moral imperative to stop the
maltreatment of people who have no way of modifying injurious social policies
because they are either outside the victimizing system, or they lack the social
power to effect changes from within by peaceable means. Sorne are disenfranchised, most feel they have no voice in decision making, and legal efforts to
remedy their grievances are repeatedly thwarted. Even if the judicial procedures
were impartially administered. few could afford the heavy expenses and the
protracted time required to exhaust legal remedies. Not only is one not obligated
to obey authorities who preside over inequitable systems that protect them with
layers of bureaucratic barriers and coercive power, so the reasoning goes, but
one is morally right to disohey them. When leaders gain widespread support
from a populace that henefits from exploitive policies, the social activism of an
aggrieved minority is more likely to arouse demands for coercive social control,
rather than sympathy for them. lndeed, people in advantaged positions excuse
high levels of violence for social control, hui they are quick to condemn dissent
and protest for social change as acts of violence (Blumenthal, Kahn, Andrews, &

1.

SOCIAL COGNITIVE THEORY OF MORALITY

75

Head, 1972). Submitting lo !he punitive consequenccs of lheir disruptive proles!,

challengers argue, places inslitutional procedures above the welfare of human
beings and simply allows the system to perpeluate ils exploitation of the
disadvantaged.
As !he preceding discussion shows, adversaries can easily marshal moral
reasons for aggressive actions for social control or for social change. When
viewed from divergen! perspectives, violent acts are different things to different
people. In conflicts of power, one person's violence is another person's selless
benevolence. lt is often proclaimed that one group's criminal terroristic activity
is another group's liberation movement fought by heroic freedom fighters. This
is why moral appeals against violence usually fall on deaf ears. Adversaries
sanctify their own militan! actions but condemn those of their antagonists as
barbarity masquerading under a mask of outrageous moral reasoning.
Terrorists invoke moral principies to justify human atrocities. Moral justification is also brought into play in selecting counterterrorist measures. This poses
more troublesome problems for democratic societies than for totalitarian ones.
Totalitarian regimes have fewer constraints against using institutional power to
control media coverage of terrorist events, to restrict individual rights, to sacrifice individuals for the benefit of the state rather than make concessions to
terrorists, and to combat threats with lethal means. Terrorists can wield greater
power over nations that place high value on human life and are thereby constrained in the ways they can act. Hostage taking has become a common terroristic strategy for wielding control over govemments. lf nations make the
release of hostages a dominan! national concem they place themselves in a highly
manipulatable position. Tightly concealed captivity thwarts rescue action.
Heightened national attention along with an inability to free hostages independently conveys a sense of weakness and invests terrorists with considerable
importance and coercive power to extract concessions. Overreactions in which
nations render themselves hostage to a small band of terrorists inspires and
invites further terrorist acts. Hostage taking is stripped of functional value if it is
treated as a criminal act that gains terrorists neither any coercive concessionary
power nor much media attention.
Democratic societies face the dilemma of how to justify morally countermeasures to stop terrorists' atrocities without violating the societies' own fundamental principies and standards of civilized conduct (Carmichael, 1982). lt is
hard to find any inherent moral rightness in violen! acts designed to kili assailants
orto deter them from future assaults but that sacrifice the lives of sorne innocent
people in the process as well. Because of many uncertain factors, the toll that
counterterrorist assaults will take on innocent life is neither easily controllable
nor accurately calculable in advance. Therefore, the use of violent counterrneasures is typically justified on utilitarian grounds in tenns of the benefits to
humanity and the social order that curbing terrorist atlacks will bring. lt is
generally considered legitimate to resort to violen! defense in response to grave

76

BANDURA

threats that inlict extensive human suffering or that endanger the very survival of
the society. The gravity criterion is fine in principie but slippery in specific
application. Like most human judgments, gauging the gravity of threats involves
sorne subjectivity. Moreover, violence is often used as a weapon againsl lhreals
of lesser magnitude on lhe grounds that, if left unchecked, they will escalale in
severity to the point where they will eventually extract a high toll on human
liberties and suffering. Gauging potential gravity involves even greater subjectivity and fallibility of judgment than does assessment of present danger. Construal of gravity prescribes choice of options, but it is also often true that choice
of violent options shapes construal of gravity. Thus, projected grave dangers to
the society are commonly invoked morally to justify violent means to squelch
limited present threats.
The mass media, especially television, provide the best access to the public
through their strong drawing power. For this reason, television is increasingly
used as the principal vehicle of justification. Struggles to legitimize and gain
support for one' s causes and to discredit those of one' s opponents are now waged
more and more through the electronic media (Ball-Rokeach, 1972; Bassiouni,
1981).
The nuclear age has ushered in new magnitudes of risk that create major moral
imperatives and paradoxes. Major disputes revolve around the morality of the
development of nuclear weaponry and on nuclear retaliatory policies (Churchill,
1983; Johnson, 1984; Lackey, 1984). Proponents of the deterrence doctrine justify
threat of nuclear retaliation as a necessary meaos to protect against a nuclear attack
by rival powers. The moral justifications take the following fonn: Self-defense
against grave dangers is morally obligatory. Threats to strike back in kind if
attacked with nuclear weapons not only safeguards the populous from adversaries
with nuclear arsenals, but deters them from nonnuclear assaults against vulnerable
allies as well. Unilateral disarmament is untenable because it leaves a nation open
to coercive control by adversaries through extortive nuclear threats.
Threats of nuclear retaliation have no deterrent effect unless the feuding nations
believe that their adversary has every intention to use such weapons in the event of
a nuclear attack. But virtually everyone concedes that it would be suicida! to use
them. A nuclear deterrence doctrine paradoxically seeks to achieve a deterrent
effect with threats that no one in their right mind could conceive of ever using.
Hence, in efforts to add credibility to deterrence policies, nuclear weapons are
menacingly deployed and nuclear systems are said to be preprogrammed so that a
launch of offensive missiles will trigger a massive nuclear counterstrike semiautomatically. The intent is to strengthen the deterrent threat by creating the mindset that retaliatory reactions cannot be checked by a loss of retaliatory nerve. In the
justificatory arguments of proponents, national security is ensured by maintaining
a balance of nuclear destructiveness that will be mutually deterring. They remain
suspecl of treaties aimed at limiting or reducing ballislic arsenals on the grounds

1.

SOCIAL COGNITIVE THEORY OF MORALITY

77

that agreements are unlikely to be honored and verification procedures are inadequate to safeguard against cheating.
Opponents of nuclear deterrence policies consider the development of nuclear
weaponry and threats to use it, even in retaliation, as morally wrong. They
regard a retaliatory strike that would inevitably produce vast human and ecological devastation as a ghastly act of vengeance that is irrational as well as immoral.
A counterstrike to a failed deterrence would most likely achieve only massive
mutual destruction through a series of nuclear exchanges with surviving missiles
dispersed on land, in aircraft, and in submarines. In the afterrnath, survivors
would find themselves in a largely uninhabitable environment. Drifting radioactive fallout would spread the devastating human and ecological toll both within
and across nations. These are unique indiscriminate consequences of nuclear
weapons that limit the value of deterrence models developed for conventional
arrned conflicts. In short, the moral logic of counterstrike threat is underrnined
by its self-destructive consequences. Nuclear deterrence thus rests on a retaliatory threat that paradoxically is too self-destructively irrational and too immoral in
innocent human toll to carry out. What is immoral to do is immoral to threaten
(Kavka, 1988). Deterrent credibility must depend on perception of one's adversary as sufficiently irrational and immoral to be able to launch nuclear missiles
that can destroy each other's societies. Opponents call into question other aspects
of deterrent effect. The heavy military involvements of superpowers with nuclear
stockpiles (e.g., in Korea, Eastem Europe, Vietnam, Afghanistan) dispute the
argument that the existence of a nuclear threat deters nonnuclear military venturesomeness (Lackey, 1984). Nations are understandably unwilling to risk selfdestruction to repulse invaders abroad.
Development and deployment of nuclear weapons consume huge financial,
technical and creative resources. To justify and gain public support for continua!
investment of a large share of national resources in nuclear bombs, proponents
usually portray their adversaries as possessing nuclear superiority. Mutual disadvantageous comparison has fostered spiraling escalation of ever deadlier arsenals. Any advancement in either offensive or defensive missile technology is
likely to create a destabilizing effect that sparks a new escalation of destructive
potential. Proponents contend that a powerful deterrent threat must be maintained until such time as a space-based impenetrable defense system is developed
against ballistic missiles. Critics argue that any defensive shield will be porous
and orbiting battle stations would be highly vulnerable to counterrneasures, thus
requiring continued reliance on deterrence by retaliatory nuclear threat to bolster
the partial defense (Long, Hafner, & Boutwell, 1986). Rather than shifting the
effort from retaliatory deterrence to defensive self-protection, adding a new
defensive system to offensive retaliatory forces will only create more sophisticated nuclear systems poised for mutual destruction. Erecting new defensive
systems undercuts efforts to reduce offensive nuclear forces.

78

BANOURA

No technical system is ever foolproof. As long as nuclear weapons exist there

is always a risk that sorne day they may he fired accidentally through malfunction
o missile-monitoring systems or human error, or launched intentionally in an
extreme crisis by an enraged. panic-stricken. or suicida! leadership. On four
occasions the United States went into a state o nuclear war alert and only lastminute efforts revealed malfunclions or errors in the computer warning system
(Falk, 1983). Nuclear proliferation and swifter missile systems that cut short the
time for decisions raise !he level of risk. To seek security in a fallible system that
can produce massive nuclear annihilation is to invite human calamity of appalling proportions. Because of the vast scope and magnitude of indiscriminale
nuclear devastation, the traditional just-war tenels that sanction self-defense to
avert grave hann affords little guidance in the use of nuclear weapons. For
opponents of nuclear systems, their indiscriminate destructiveness challenges the
moral permissibility of nuclear powers inlicting the catastrophic risks of nuclear
deterrence on the people of innocent nations who are granted no say in the matter
(Lackey, 1985). Whal is immoral lo do is also immoral to risk.
11 is generally acknowledged lhat human security is advanced by multilateral
nuclear disannament. However, control of behavior by mutual threal has become
deeply entrenched in the political rhetoric and military doctrines and practices of
nations. They seek lo gain nuclear advantage as a bargaining chip and, in so
doing, spur reciproca! escalation of destructive power. Human survival in the
nuclear age requires nalions lo develop and leam de-escalative modes of thinking
and hehaving in regard to nuclear weapons. Sorne models of graduated deescalative reciprocalion have been proposed (Osgood, 1980) and occasionally
tried successfully (Scoville, 1985 ). In this approach to reversing the nuclear arms
race, a nation initiates a calculated de-escalation designed to prompl a reciprocating action by an opponent. For example. President Kennedy announced that the
United States would cease nuclear tests in the atmosphere as long as the Soviet
Union exercised similar restraint. This publicized iniliative quickly produced an
inlemational agreemenl barring atmospheric nuclear tests. De-escalative
initiatives are lhus gradually introduced within security limits to prod reciprocation. lf concertedly applied, such iniliatives might achieve drastic multilateral
reductions in nuclear arsenals. Proliferation of nuclear weapons among nalions
thal distrusl and fear each olher makes lhis lask more difficult. As long as sorne
nations refuse lo part with their bombs, others insisl on remaining nuclearly
armed lo deter nuclear threats. Graduated de-escalative stralegies thal are exercisable when nuclear weapons are in the hands of only a few nalions encounter
greater obstacles when many nations wilh chronic animosities are nuclearly
armed. Therefore. to add social and moral force lo de-escalative modes of
change, social mechanisms need to be created whereby societies collectively
applaud reciprocation to initiatives and reprove failures to reciprocate.

Euphemistic Labeling

Language shapes people's thought pattems on which they hase many of their
actions. Activities can take on a very different appearance, depending on what
they are called. Euphemistic language thus provides a convenient device for
masking reprehensible activities or even conferring a respectable status upon
them. Through convoluted verbiage, destructive conduct is made benign and
those who engage in it are relieved of a sense of personal agency. Laboratory
studies reveal the disinhibitory power of euphemistic language (Diener, Dineen,
Endresen, Beaman, & Fraser, 1975). Adults behave much more aggressively
when assaulting a person is given a sanitized label than when it is called
aggression.
In an insightful analysis of the language of nonresponsibility, Gambino
( 1973) identifies the different varieties of euphemisms. One form, palliative
ex.pressions, is widely used to make the reprehensible respectable. Through the
power of hygienic words, even killing a human being loses much of its repugnancy. Soldiers waste people rather than kili them, intelligence operatives terminote
(them) with exJreme prejudice (Safire, 1979). When mercenaries speak of
fulfilling a contract, murder is transforrned by admirable words into the honorable discharge of duty. Terrorists label themselves asfreedomfighters. Bombing
attacks become clean, surgical .ttrikes, invoking imagery of the restorative handicrafts of the operating room, and the civilians they kili are linguistically converted to collateral damage (Hilgartner, Bell, & O'Connor, 1982).
Sanitizing euphemisms, of course, perform heavy duty in less loathsome but
unpleasant activities that people are called upon to do from time to time. In the
language of sorne govemment agencies, people are not fired, they are selected
out. as though they were receiving preferential treatment. A corporate memo
speaks not of laying people off work, but of resizing our operations to the leve/
of profitable market opportunities. In teaching business students how to le in
cpmpetitive transactions, the instructor speaks euphemistically of strategic misrepresentation (Safire, 1979). The television industry produces and markets
sorne of the most brutal forms of human cruelty under the sanitized labels of
action and adventure programming (Baldwin & Lewis, 1972). The acid rain that
is killing our lakes and forests loses much of its acidity in its euphemistic form as
atmospheric deposition of anthropogenically derived acidic substances (Hechinger, 1985). The nuclear power industry has created its own specialized set of
euphemisms for the injurious effects of nuclear mishaps; an ex.plosion becomes
an energetic disassembly, a reactor accident is a normal aberration, and plutonium contamination is merely infi/tration (San Francisco Chronicle, 1979a).
The agentless passive form serves as a linguistic device for creating the
appearance that culpable acts are the work of nameless forces, rather than people

79

80

BANDURA

(Bolinger, 1982). lt is as though people are moved mechanically but are not
really the agents of their own acts. Even inanimate objects are sometimes invested with agentive properties: The telephone pole was approaching. I was auempting to swerve out o/ its way when it struck. my front end (San Francisco Chronicle, 1979b). Gambino further documents how the specialized jargon o a
legitimate enterprise can be misused to lend an aura o respectability to an
illegitimate one. In the Watergate vocabulary criminal conspiracy became a
game plan, and the conspirators were team players calling for the qualities and
behavior befitting the best sportsmen. The disinhibitory power o language can
be boosted further by colorful metaphors that change the nature o culpable
activities.

Advantageous Comparison
Whenever events occur or are presented contiguously. the first one colors how
the second one is perceived and judged. By exploiting the contrast principie,
moral judgments o conduct can be intluenced by expedient structuring o what it
is compared against. Thus, self-deplored acts can be made righteous by contrasting them to tlagranl inhumanities. The more outrageous the comparison practices, the more likely il is thal one's own destructive conduct will appear tritling
or even benevolent. Promoters o the Vietnamese war and their supporters, for
example, minimized the slaying o countless people as a way o checking massive Communist enslavemenl. Given the tritling comparison, perpetrators of
warfare remained unperturbed by the facl that the intended beneficiaries were
being killed atan alanning rate. Domestic protesters, on the other hand, characlerized their own violence against educational and poltica! institutions as tritling, or even laudable, by comparing il with the camage perpetrated by their
country's military forces in foreign lands. Terrorists minimize their slayings as
the only deense weapon they have to curb the widespread cruelties intlicted on
their people. In the eyes o their supporters, risky attacks directed at the apparatus o oppression are acts o seltlessness and martyrdom. Those who are the
objects of terrorist attacks, in tum, characterize their retaliatory violence as
tritling, or even laudable, by comparing them wilh camage and terror perpetrated
by terrorists. In social contlicts, injurious behavior usually escalates with each
side lauding its own behavior but morally condemning that of their adversaries as
heinous.
Historical advantageous comparisons are also invoked as justifications of
violence. Advocates of terrorist tactics are quick to note that the democracies o
England, France, and the United States were bom o violence against oppressive
rule. A fonner director o the CIA effectively detlected, by favorable comparison, embarrassing questions about the morality and legality of CIA-directed

1.

SOCIAL COGNITIVE THEORY OF MORALITY

81

covert operations designed to overthrow of an authoritarian regime. He explained

that French covert operations and military supplies greatly aided the overthrow of
oppressive British rule during the War of lndependence, thereby creating the
modern model of democracy for other subjugated people to emulate.
Social comparison is similarly used to show that the social labeling of acts
may depend more on the ideological allegiances of the labelers than on the acts
themselves. Airline hijackings were applauded as heroic deeds when East Europeans and Cubans initiated this practice, hut condemned as terrorist acts when
the airlines of Western nations and friendly counlries were commandeered. The
degree of psychopalhology ascribed to hijackers varied, depending on the direction of the rerouted lights. Moral condemnations of politically motivated terrorism are easily blunted by social comparison because, in international contests
for political power, il is hard to find nations that categorically condemn terrorism. Rather, they usually back sorne terrorists and oppose others.
Cognitive restructuring of behavior through moral justifications and palliative
characterizations is the most effective psychological mechanism for disengagement of moral self-sanctions. This is because moral restructuring not only eliminates self-deterrents but engages self-approval in the service of destructive exploits. What was once morally condemnable becomes a source of self-valuation.
After destructive meaos become invested with high moral purpose, functionaries
work hard to become proficient at them and take pride in their destructive
accomplishments.

Displacement of Responsibility
Self-sanctions are activated most strongly when personal agency for detrimental
effects is unambiguous. Another set of dissociative practices operates by obscuring or distorting the relationship between actions and the effects they cause.
People will behave in ways they normally repudiate if a legitimate authority
accepts responsibility for the consequences of the conduct (Diener et al., 1975;
Milgram, 1974). Under conditions of displaced responsibility, people view their
actions as springing from the dictates of authorities rather than their being personally responsible for them. Since they are not the actual agent of their actions,
they are spared self-prohibiting reactions. Displacement of responsibility not
only weakens restraints over one's own detrimental actions but diminishes social
concern over the well-being of those mistreated by others (Tilker, 1970).
Most of the research on attributional analysis of moral judgment is concerned
with whether people view their behavior as determined by externa! circumstances
or hold themselves responsible for it (Ross & DiTecco, 1975; Rule & Nesdale,
1976). Perceptions of causal responsibility are reduced if the harmful consequences of actions are viewed as unintended, unforeseeable, or the actions arose

82

BANDURA

from thc dictatcs of thc situation. Within thc attrihutional framcwork. thcsc
factors are usually studicd as mitigators of moral judgmcnt rathcr !han as discngagcrs of moral sclf-sanctions.
faemption from sclf-devaluation for hcinous dceds by displacement of responsibility has heen most grucsomely rcvcaled in socially sanctioned mass
ellecutions. Nazi prison commandants and their staffs divested themselves of
personal responsihility for their unprccedented inhumanities (Andrus, 1969).
They were simply carrying out orders. Impersonal obedience to horrific orders
was similarly evident in military atrocities, such as the My Lai massacre (Kelman. 1973). In an effort to deter institutionally sanctioned atrocities, the Nuremherg Accords were estahlished. declaring that obedience to inhumane orders,
even from the highest authorities. does not relieve subordinates of the responsihility of their actions. Howcver. since victors are disinclined to try themselves as
criminals, such decrees have limited deterrence without an intemational judiciary
system empowered to impose penalties on victors and losers alike.
In fonnal studies of disengagement of self-sanctions through displacemcnt of
responsibility. authorities explicitly authorize injurious actions and hold themselves fully accountable for the hann caused by the activity. However, in the
sanctioning practices of everyday life responsibility for detrimental conduct is
rarely assumed so explicitly. because only obtuse authorities would leave thcmselves accusable of authorizing heinous acts. They are concemed not only with
adverse social consequences to themselves should advocated courses of action
miscarry. hut with the loss of self-rcgard for sanctioning human atrocities in
ways that leave blood on their hands. Thcrefore. authorities usually invite and
support detrimcntal conduct in insidious ways that minimize personal responsihility for what is happening. Moreovcr. thc intcndcd purpose of sanctioned
destructiveness is usually disguiscd so that neithcr issucrs nor perpetrators regard
their actions as censurable. When rcproachful practices are publicized, they are
officially dismissed as only isolated incidcnts arising through misunderstanding
of what, in fact, had becn authorizcd.
Kramcr (1990) describes the great lengths to which Shi'ite clerics go to
provide moral justifications for violen! acts that seem to breach lslamic law, such
as suicida! bombings and hostage taking. These efforts are designed not only to
persuade themselves of the morality of their actions but to preserve their integrity
in the eyes of other nations. The religious codc pennits neither suicide nor
terrorizing innocent people. On the one hand, the clerics justify such acts by
invoking situational imperativcs and utilitarian rcasons, namely that tyrannical
circumstances drive oppressed people to unconventional means to rout aggressors who wield massive dcstructive power. On the other hand, they reconstrue terrorist acts as conventional mcans in which dying in a suicida! bombing
for a moral cause is no different than dying al the hands of an enemy soldier.
Hostages simply gel relaheled as spies. When the linguistic solution defies credihility. personal moral rcsponsihility is discngagcd by construing terroristic acts

1.

SOCIAL COGNITIVE THEORY OF MORALITY

83

as dictated by their foe's tyranny. Because of the shaky moral logic and disputable reconstruals. clerics sanction terrorism by indirection. they vindicate successful ventures retrospcctively, and they disclaim endorsing terroristic opcrations beforehand.
States sponsor terrorist opcrations through disguised roundabout routes that
make it difficult to pin the blame on them. Moreover, the intended purpose of
sanctioned destructiveness is usually linguistically disguised so that neither issuers nor pcrpctrators regard the activity as censurable. When culpable practices
gain public attention. they are officially dismissed as only isolated incidents
arising through misunderstanding of what, in fact, had been authori1.ed. Efforts
are made to limit the blame to subordinates who are portrayed as misguided or
overzealous.
Displacemcnt of responsibility also opcrates in situations in which hostages
are taken. Terrorists wam officials of targeted regimes that if they take retaliatory
action they will be held accountable for the lives of the hostages. At different
steps in negotiations for their release. terrorists continue to displace the responsibility for the safety of hostages on the reactions of the regime. lf the captivity
drags on, terrorists blame the suffering and injuries they inflict on the hostages
on the regime for failing to make what they regard as warranted concessions to
right social wrongs.
A number of social factors affect the ease with which responsibility for one's
actions can be surrendered to others. High justification and social consensus
about the morality of an enterprise aid in the relinquishment of personal control.
The legitimacy of the authorizers is another important detenninant. As can be
seen in Figure 1.2, the greater the legitimation and closeness of the authority
issuing injurious commands, the higher is the level of obedience. The higher the
authorities, the more legitimacy, respect, and coercive power they command,
and the more amenable are people to defer to them. Modeled disobedience,
which challenges the legitimacy of the activists, if not the authorizers themselves, reduces the willingness of observers to carry out the actions called for by
the orders of a superior (Meeus & Raaijmakers, 1986; Milgram, 1974; Powers &
Geen, 1972). lt is difficult to continue to disown personal agency in the face of
evident hann following directly from one's actions. People are, therefore, less
willing to obey authoritarian orders for injurious behavior when they see firsthand how they are hurting others (Milgram, 1974; Tilker, 1970).
Obedient functionaries do not cast off ali responsibility for their behavior as
though they were mindless extensions of others. lf this were the case, they would
act like automatons, only when told to. In fact, they are much more conscientious and self-directed in the performance o their duties. lt requires a strong
sense of responsibility to be a good functionary. In situations involving obedience to authority, people carry out orders partly to honor the obligations they
have undertaken (Mantel! & Panzarella, 1976). One must, therefore, distinguish
between two levels of responsibility; duty to one's superiors and accountability

BANDURA

84
100

90

80
c/J

a:

oc/J

70

c/J

a:

CJ

~
zw

60

1-

ow

50

40

1-

zw

a:

30

Q.

20

'

AUTHORIIY
COMMANOS

AUTHORITY
COMMANDS
DIRECTLY

COMPLIANT
PEE AS

AUTHORITY
COMMANOS

RE.MOTELY

NON AUTHORITY
COMMANOS

AUTHORITY
COMMANDS
+
DHIANT
PEERS

CONFLICTING
AUTHORITIES
COMMAND

FIG. 1.2. Percent of people fully obedient to injurious commands as a

function of the legitimation and closeness of the authority issuing the
commands {plotted from data by Milgram, 19741.

for lhe effecls of one's aclions. The self system operales most efficiently in lhe
service of aulhority when followers assume personal responsibility for being
duliful executors while relinquishing personal responsibility for the hann caused
by their behavior. Followers who disowned responsibilily without being bound
by a sense of duty would be quite unreliable.

Diffusion of Responsibility

1e deterrent power of self-sanctions is weakened when the link belween conduct and ils consequences is obscured by diffusing responsibility for culpable
behavior. This is achieved in several ways. Responsibility can be diffused by
division of labor. Most enterprises require the services of many people, each
perfonning fragmentary jobs thal seem hannless in themselves. The fractional
contribulion is easily isolated from lhe eventual function, especially when participants exercise liltle personal judgment in carrying oul a subfunction thal is
relaled by remole, complex links lo lhe end resull. Afler aclivilies become

1.

SOCIAL COGNITIVE THEORY OF MORALITY

85

routinized into programmed subfunctions, attention shifts from the import of

what one is doing to the details of one's fractional job (Kelman, 1973).
Group decision making is another common bureaucratic practice that enables
otherwise considerate people to behave inhumanely, because no single individual
feels responsible for policies arrived at collectively. Where everyone is responsible no one is really responsible. Social organizations go to great lengths to devise
sophisticated mechanisms for obscuring responsibility for decisions that will
affect others adversely. Collective action is still another diffusion expedient for
weakening self-restraints. Any harm done by a group can always be ascribed, in
large part, to the behavior of other members. People, therefore, act more harshly
when responsibility is obfuscated by a collective instrumentality than when they
hold themselves personally accountable for what they do (Bandura, Underwood,
& Fromson, 1975; Diener, 1977; Zimbardo, 1969). Figure 1.3 shows the level of
punitiveness of individuals given punitive power over others under conditions in
which the severity of their sanctions was determined personally or jointly by a
group (Bandura, lJnderwood, & Fromson, 1975).

p...._
/
(fJ
(fJ

w
z
w

>
:::

_.o- - _.o- - -0----o-

z
::>

Q.

IL

o_.

w
>
w
_.

.......... 'O

r ....

_d

o- -

I
I

/
- - GROUP RESPONSIBIUTY
0---0

__ L_

_ _ _ l ~_

INDIVIDUAL

__ _ _ _t___~----- __ L____

__.._ ____J

10

OCCASIONS

FIG. 1.3. Level of punitiveness by individuals under conditions in

which severity of their punitiveness was determined personally or
jointly by a group. Occasions represent successive times at which
punitive sanctions could be applied (Bandura, Underwood, & Fromson, 1975)

86

BANDURA

This is not to say that shared responsihility has no legitimate purpose. In

efforts to serve diverse constitucncies. actions heneficial to one group may he
detrimental to another. Because differences are not always reconcilahle, someone will inevitahly he hurt, whatever is done. Those who must makc tough
decisions and perform society's nasty duties are at least spared sorne personal
distress by sharing the accountahility. They could not function for long if they
had to hear the full load alone.
People oftcn hehave in hannful ways, not hecause responsibility is diffused
by formal organizational arrangemenls, but hecause they all routinely engage in
activities that conlribule lo negalive effecls. They pollule lhe air they brealhe
with their automobiles and degrade lheir environmenl lo produce the vast
amounls of energy and producls lhey consume. As a resull of collective action,
good environmentalists can also he good pollulers by blaming others for degrading the environment. The more detrimental the collectively produce effects, the
less people feel personally responsible for them (Shippee & Christian, 1978).

Disregard or Distortion of Consequences

Additional ways of weakening self-deterring reactions operate through disregard
or misrepresentation of lhe consequences of action. When people choose to
pursue activities harmful lo others for personal gain. or hecause of social inducemenls, lhey avoid facing the harm lhey cause or lhey minimize it. They readily
recall prior information given them ahout the potential henefits of lhe behavior
bul are less able to rememher its harmful effects (Brock & Buss, 1962, 1964).
People are especially prone to minimize injurious effects when they act alone
and, thus, cannot easily escape responsibilily (Mynatt & Herman, 1975). In
addition to selective inattention and cognitive distortion of effects, the misrepresentation may involve active efforts to discredit evidence of the harm they cause.
As long as the detrimental results of one's conduct are ignored, minimized,
distorted. or dishelieved, there is little reason for self-censure to he activated.
lt is relatively easy to hurt others when their suffering is not visible and when
causal actions are physically and temporally remote from their effects. Our death
technologies have hecome highly lethal and depersonalized. Mechanized weapon
systems and explosive devices in which many people can he pul to death by
destructive forces unleashed remotely. illustrates such dcpersonalized action.
Even high personal responsibility is a weak restrainer when aggressors do not
know the harm they inlict on their victims (Tilker, 1970). In contras!, when
people can see and hear the suffering they cause, vicariously aroused distress and
self-censure serve as self-restraining influences. For example, in his studies of
commanded aggression, Milgram ( 1974) obtained diminishing ohedience as the
victims' pain hecame more evident and personalized (Figure 1.4).
Most organizations involve hierarchical chains of command in which superiors formulate plans and intermediaries transmit them to executors, who then

1.

SOCIAL COGNITIVE THEORY OF MORALITY

87

-----

60
ffl

a:

ffl
ffl

w
a:

50

(!)

~
z

1-

w
w

-~

40

i5
al

1-

w
u
a:
w

30

n.

20

REMOJE
INFUCTION
OF SUFFERING

HEARS
SUFFERING
OF REMOJE
INFUCTION

SEES ANO HEARS

OIRECT
SUFFERING
INFUCTION
OF REMOTE
OF SUFFERtNG
INFUCTION

FIG. 1.4. Percentage of people fully obedient to injurious commands

issued by an authority as the victim's suffering becomes more evident
and personalized (plotted from data by Milgram, 1974).

carry them out. The farther removed individuals are form the end results, the
weaker is the restraining power of the foreseeable destructive effects. Kilham
and Mann ( 1974) set forth the view that the disengagement of personal control is
easiest for the intennediaries in a hierarchical system-they neither bear responsibility for major decisions nor are they a party to their execution. In performing
the transmitter role, they model dutiful behavior and further legitimize their
superiors and their social policies and practices. Consistent with these speculations, intennediaries are much more obedient to destructive commands than are
those who have to carry them out and face the results (Kilham & Mann, 1974).

Dehumanization
The final set of disengagement practices operates on the recipients of detrimental
acts. The strength of self-evaluative reactions to injurious conduct partly depends
on how the perpetrators view the people toward whom the behavior is directed.
To perceive another as human activates empathetic or vicarious emotional reactions through perceived similarity (Bandura. 1991 ). The joys and sufferings o
similar persons are more vicariously arousing than are those of strangers or
individuals who have been divested of human qualities. Personalizing the injurious effects experienced by others also makes their suffering much more salient.

88

BANDURA

lt is difficult to mislreal humanized persons without risking personal distress and

self-censure. Vicarious emotional activation is cognitively mediated rather than
automatically elicited by the experiences of others. Ascriptions of insensateness
to victims weakens vicarious self-arousal of distress to their suffering. People are
unmoved by "unfeeling" recipients of maltreatment. Subhumans are not only
regarded as lacking sensitivities, but as being inluenceable only by severe
methods.
Self-sanctions against cruel conduct can be disengaged or blunted by divesting people of human qualities. Once dehumanized, they are no longer viewed as
persons with feelings, hopes, and concems but as subhuman objects. They are
ponrayed as mindless savages, goob, satanic fiends, and the other despicable
wretches. lf dispossessing antagonists of humanness does not blunt self-reproof,
it can be eliminated by anributing bestial qualities to them. They become
degenerates, pigs, and other bestial creatures. lt is easier to brutalize victims
when they are referred to as worms (Gibson & Haritos-Fatouros, 1986). The
process of dehumanization is an essential ingredient in the perpetration of inhumanities. A Nazi camp commandant chillingly explained that the extreme
lengths to which they went to degrade victims they were going to kili anyway
was nota matter of purposeless cruelty (Levi, 1987). Rather, the victims had to
be degraded to the point of subhuman objects so that those who operated the gas
chambers would be less burdened by distress. Over the years, slaves, women,
manual laborers, and religious and racial minorities have been treated as chattel
oras subhuman objects (Ball-Rokeach, 1972).
When persons are given punitive power, they treat dehumanized individuals
much more punitively than those who have been invested with human qualities
(Figure 1.5). Dehumanization fosters different self-exonerative pattems of
thought (Bandura, Underwood, & Fromson, 1975). People seldom condemn
punitive conduct and they create justifications for it when they direct it toward
individuals who have been deprived of their humanness. However, people
strongly disapprove of punitive actions and rarely excuse their use toward individuals depicted in humanized terms.
When severa! disengagement factors are combined, they potentiate each other
rather than simply produce additive effects. Thus, combining diffused responsibility with dehumanization greatly escalates the level of punitiveness, whereas
personalization of responsibility, along with humanization, have a powerful selfdeterring effect (Figure 1.6).
Many conditions of contemporary life are conducive to impersonalization and
dehumanization (Bemard, Onenberg, & Redl, 1965). Bureaucratization, automation, urbanization, and high geographical mobility lead people to relate to
each other in anonymous, impersonal ways. In addition, social practices thal
divide people inlo ingroup and outgroup members produce human estrangements
that fosters dehumanization. Strangers can be more easily casi as insensate than
can impersonal acquaintances.

1.

SOCIAL COGNITIVE THEORY OF MORALITY

89

e
7

6
U)
U)

..

~ 5
>

::

z::,
Q.

...........

11.

o
l
>
w

.. .... ...

..... ..

_.o,

_,,o- -

..J

//

-0-- -

..... ....

p...._

.....

,,-0
''-cf/

/
'O.. -

-el

....-.d
O"/

- - OEHUMANIZED
- NEUTRAL
o---o HUMANIZED

.___.___~---~-~-~--~L--~--~-~-~
2
3
4
5
6
1
e
9
10
OCCASIONS

FIG. 1.5. Level of punitiveness on repeated occasions toward people

characterized in humanized terms, not personalized with any characterization (neutrall. or portrayed in dehumanized terms (Bandura, Underwood, & Fromson, 19751.

Under certain conditions, the exercise of institutional power changes the users
in ways that are conducive to dehumanization. This happens most oten when
persons in positions o authority have coercive power over others and adequate
safeguards for constraining the behavior o powerholders are lacking. Powerholders come to devalue those over whom they wield control (Kipnis, 1974). In a
simulated prison experiment (Haney, Banks, & Zimbardo, 1973), even college
students, who had been randomly chosen to serve as either inmates or guards
given unilateral power, began to treat their charges in degrading, tyrannical ways
as guards. Thus, role assignment that authorizes use o coercive power overrode
personal characteristics in promoting punitive conduct. Systematic tests o relative influences similarly show that social influences conducive to punitiveness
exert considerably greater sway over aggressive conduct than do people's personal characteristics ( Larsen, Cole man, Forges, & Johnson, 1971 ) .
The overall findings from research on the different mechanisms o moral
disengagemenl corroborate the historical chronicle of human atrocities: ll requires conducive social conditions rather than monstrous people to produce

90

BANDURA
7
- - GROUP RESPONS181LITY
0----0 INDIVIDUAL RESPONS181LITY

f/)

zw

>

::

::>

ll.

4
/

u.

/Y

-- --

--- --- --- ---

---

/
/

...J

~
3
...J

/
/

<f

HUMANIZED

NEUTRAL

DEHUMANIZED

FIG. 1.6. Level of punitiveness as a function of diffusion of responsibility and dehumanization of the recipients (Bandura, Underwood, &
Fromson, 19751.

heinous deeds. Given appropriate social conditions, decent, ordinary people can
be led to do extraordinarily cruel things.
Power of Humanization. Psychological research tends to focus extensively
on how easily it is to bring out the worst in people through dehumanization and
other self-exonerative means. The sensational negative findings receive the
greatest attention. Thus, for example, the aspect of Milgram's research on ohedient aggression that is widely cited is the evidence that good people can he
talked into perfonning cruel deeds. However, to get people to carry out punitive
acts, the overseer had to be physically present, repeatedly ordering them to act
cruelly as they voiced their concems and objections. Orders to escalate punitiveness to more intense levels are largely ignored or subverted when remotely issued
by verbal command. As Helm and Morelli ( 1979) note, this is hardly an example
of blind obedience triggered by an authoritative mandate. Moreover, what is
rarely noted, is the equally striking evidence that most people steadfastly refuse
to behave punitively. even in response to strong authoritarian commands, if the

1.

91

SOCIAL COGNITIVE THEORY OF MORALITY

situalion is personalized by having lhem see lhe victim or requiring lhem lo

inlicl pain direclly ralher lhan remolely.
The emphasis on obedienl aggression is underslandable considering lhe prevalence and hannfulness of people's inhumanities to one another. However, of
considerable lheorelical and social significance is the power of humanization to
counteract cruel conduct. Studies examining this process revea! that it is difficult
for individuals to behave cruelly toward others when they are humanized or even
personalized a bit (Bandura. Underwood. & Fromson, 1975). Even under conditions in which punitive sanctions are the only means available and they are highly
functional in producing desired results, those exercising that power cannot get
themselves to behave punitively toward humanized individuals (Figure I. 7). The
affinnation of common humanity can bring out the best in others. In contrast,
when punitive sanctions are dysfunctional because they usually fail to produce
results, punitiveness is precipitously escalated toward dehumanized individuals.
The failure of degraded individuals to change in response to punitive treatment is
taken as further evidence of their culpability that justifies intensified punitiveness
toward them.

OEHUMANIZED
NEUTRAL
o--o HUMANIZED

(/)
(/)

w
z 6
w
>
:::

::>

Q.

11.

.8.
,--,

..J

~I

>
w

..J

FUNCTIONAL

'

b---el'

A
,,

I
'

I \

'\J

4
6
DYSFUNCT IONAL

\
\

~/

OCCASIONS
FIG. 1.7. level of punitiveness on repeated occasions as a result of
dehumanization and the effectiveness of punitive actions. Under the
functional condition, punishment consistently produced good results;
under the dysfunctional condition, punishment usually failed to
achieve desired results IBandura, Underwood, & Fromson, 1975).

10

92

BANDURA

The moderating inluence of humanization is strikingly revealed in situations

involving great threat of violence. Mosl abduclors find it difficult to harm their
hostages after lhey have gotten to know lhem personally. Wilh growing acquaintance, it becomes increasingly difficull to 1ake a human life cold-bloodedly.
Humanizalion, of course, is a two-way process. Caplives may also develop sorne
sympathy for lheir captors as they get to know them. This phenomenon is
graphically illustraled in a Stockholm incident in which people who were held
hoslage for 6 days by bank robbers began to sympathize with their criminal
captors and sided with them againsl the police (Lang, 1974). This hoslage
incident included several features that are especially conducive to developmenl
o human affinity (Bandura, 1990). Most people suppon the death penalty in lhe
abstract, but the more they know about panicular cases, the less they favor
eJtecuting them (Ellswonh, 1978). As Ellswonh eltplains it, in the absence o
personal information people conjure upan image of the most heinous criminal,
an image that disposes them to favor punishment by death.

Attribution of Blame
lmputing blame to one's antagonists orlo environmental circumstances is still
anolher eJtpedient that can serve sel-eJtonerative purposes. In this process people
view themselves as faultless victims and their detrimental conductas compelled
by forcible provocation. Detrimenlal interactions usually involve a series of
reciprocally escalative actions, in which the antagonists are rarely faultless. One
can always select from the chain of events an instance of the adversary's defensive behavior and consider it as the original instigation. lnjurious conduct thus
becomes a justifiable defensive reaction to belligerent provocations. Those who
are victimized are not entirely faultless because, by their behavior, they usually
contribute at least panly to their own plight. Victims can, therefore, be blamed
for bringing suffering on themselves. Self-eJtoneration is similarly achievable by
viewing one's destructive conduct as forced by circumstances rather than as a
personal decision. By blaming others or circumstances, not only are one's own
aclions eJtcusable hui one can even feel self-righteous in the process.
Observers of victimization can be disinhibiled in much the same way as
perpetrators are by the tendency to infer culpability from misfonune. Seeing
victims suffer maltreatment for which they are held panly responsible leads
observers to derogate them (Lerner & Miller, 1978). The devaluation and indignation aroused by ascribed culpability, in tum, provides moral justification for
even greater maltreatment. That attribution of blame can give rise to devalualion
and moral justification illustrates how the various disengagement mechanisms
are often interrelated and work logether in weakening interna! control.
lnputting blame operates as a prominent disengagement mechanism in sexually assaultive behavior lowards women. Rapists and males who acknowledge a
proclivily to rape subscribe to mylhs about rape embodying the various mecha-

1.

SOCIAL COGNITIVE THEORY OF MORALITY

93

nisms by which moral self-censure can be disengaged (Feild, 1978; Malamuth,

1981). These beliefs hold rape victims responsible for their own victimization
because they havc supposedly invited rape by sexually provocative appearance
and behavior and by resisting sexual assault only weakly. Men blame rape
victims more than women do. Trivialization and distortion of consequences to
rape victims is another disengagement mechanism that comes into play. Men
who are inclined to assault sexually believe that women secretly enjoy being
raped. Anticipatory self-censure is eliminated when the traumatic effects of
sexual assault are twisted into pleasurable ones for the victim. Such self-disinhibiting pattems of thinking predict proclivity to rape, whereas sexual altitudes,
frustration, and quality of sex life do not (Briere & Malamuth, 1983).
Cross-cultural studies reveal that aggressive sexuality is an expression of the
cultural ideology of male dominancy (Sanday, 1981). Rape is prevalen! in societies where violence is a way of life, male supremacy reigns, aggressive
sexuality is valued as a sign of manliness, and women are treated as property.
Rape is rare in societies that repudiate interpersonal aggression, endorse sexual
equality, and treal women respectfully. Cultural ideologies that attach prestige to
male dominance and aggressive sexuality weaken self-censure for sexual abuse
of women. Cultural practices that belittle the role of women and a tlourishing
industry of pomography that dehumanizes them contribute further to the selfdisinhibition of aggression toward women (Malamuth & Donnerstein, 1984;
Zillman & Bryant, 1984).
Justified abuse can have more devastating human consequences than acknowledged cruehy. Maltreatment that is not clothed in righteousness makes the
perpetrator rather than the victim blameworthy. But when blame is convincingly
ascribed to victims they may eventually come to believe the degrading characterizations of themselves (Hallie, 1971). Moreover, ascriptions of blame are
usually accompanied by discriminatory social practices that create the very failings that serve as excuses for maltreatment. Vindicated inhumanity is, thus,
more likely to install self-contempt in victims than inhumanity that does not
attempt to justify itself.

Gradualistic Moral Disengagement

The aforementioned disengagement devices will not instantaneously transform a
considerate person into an unprincipled, callous one. Rather, the change is
usually achieved through gradual diminution of self-sanctions in which people
may not fully recognize the changes they are undergoing. lnitially. individuals
are prompted to perform questionable acts that they can tolerate with little selfcensure. After their discomfort and self-reproof have been diminished through
repeated performances, the level of reprehensibility progressively increases until
eventually acts originally regarded as abhorrent can be performed without much
distress. Escalative self-disinhibition is accelerated if inhumane behavior is con-

94

BANDURA

strued as serving moral purposes and the people being subjected to maltreatment
are divested of human qualities (Bandura, Underwood, & Fromson. 1975; Gibson & Haritos-Fatouros. 1986).
Analyses of moral disengagement mechanisms usually draw heavily on examples from military and political violence. This tends to convey the impression
that selective disengagement of self-sanctions occurs only under extraordinary
circumstances. Quite the contrary; such mechanisms operate in everyday situations in which decent people routinely perform activities having injurious human
effects to further their own interests or for profit. Self-exonerations are needed to
neutralize self-sanctions and to preserve self-esteem. For example, institutionalized discrimination, a practice which takes a heavy toll on its victims,
requires social justification, attributions of blame, dehumanization, impersonalized agencies to carry it out, and inattention to the injurious effects it causes.
Different industries. each with its public-spirited vindications. may cause
harmful effects on a large scale, either by the nature of their products or the
environmental contaminants they produce.

Disengagement of Self-Sanctions and Self-Deception

The issue arises as to whether disengagement of self-sanctions involves selfdeception. Because of the incompatibility of being simultaneously a deceiver and
the one deceived, literal self-deception cannot exist (Bok, 1980; Champlin,
1977; Haight, 1980). lt is logically impossible to deceive oneself into believing
something, while simultaneously knowing it to be false. Efforts to resolve the
paradox of how one can be the agent and the object at the same time have met
with little success (Bandura, 1986). These attempts usually involve creating split
sel ves and rendering one of them unconscious. However, the self-splitting solution annihilates such a phenomenon rather than explains it. The split-self conceptions fail to specify how a conscious self can lie toan unconscious self without
sorne awareness of what the other self believes. The deceiving self has to be
aware of what the deceived self believes in order to know how to concoct the
deceptions. Different levels of awareness are sometimes proposed as another
possible solution to the paradox. lt is said that "deep down" people really know
what they believe. This attempl to reacquaint the split selves only reinstales the
paradox of how one can be the deceiver and the one deceived al the same time.
People, of course. often misconstrue evenls, they lead themselves astray by their
biases and misbeliefs, and they act uniformedly. However, to be misdirected by
one's beliefs or ignorance does not mean that one is lying to oneself.
People's values and beliefs affect what information they seek and how they
interpret what they see and hear. Most strive to maintain or enhance their positive
self-regard. Therefore, they do not go looking for evidence of their culpahility or
adverse effects of their actions. Selective self-exposure and distorted interpretations of events, which confirm and strengthen preexisting beliefs, relect biased

1.

SOCIAL COGNITIVE THEORY OF MORALITY

95

self-persuasion. nota case of self-deception. To be misdirected hy one's preconceptions does not mean that one is lying to oneself.
Self-deception is often invoked when people choose to ignore possihly countervailing evidence. lt could be argued that they must believe its validity in order
to avoid it, otherwise they would not know what to shun. This is not necessarily
so. Staunch believers often choose not to waste their time scrutinizing opposing
arguments or evidence because they are already convinced of their fallacy. When
confronted with evidence that disputes their heliefs, they question its credibility.
dismiss its relevance, or twist it to fil their views. However, if the evidence is
compellingly persuasive, they alter their original beliefs to accommodate the
discrepan! evidence.
People may harbor sorne doubts conceming their beliefs but avoid seeking
certain evidence because they have an inkling the evidence might disconfinn
what they wish to believe. lndeed, they may engage in ali kinds of maneuvers,
both in thought and in action, to avoid finding out the actual state of affairs.
Suspecting something is not the same as knowing it to be true. lnklings can
always be discounted as possibly being ill-founded. As long as one does not find
out the truth, what one believes is not personally known to be false. Both Haight
(1980) and Fingarette (1969) give considerable attention to processes whereby
people avoid painful or incriminating truth by either not taking actions that would
revea! it or not spelling out fully what they are doing or undergoing that would
make it known. They act in ways that keep themselves intentionally uninformed.
They do not go looking for evidence of their culpability or the harmful effects of
their actions. Obvious questions that would revea! unwelcome information remain unasked so they do not find out what they do not want to know. lmplicit
agreements and social arrangements are created that leave the foreseeable unforeseen and the knowable unknown.
In addition to contending with their own self-censure, people are concemed
about how they appear in the eyes of others when they engage in conduct that is
rnorally suspect. This adds a social evaluative factor to the process. Haight
( 1980) argues that, in much of what is called self-deception, persons are aware of
the reality they are trying to deny, but they create the public appearance that they
are deceiving themselves. Others are thus left uncertain about how to judge and
treat persons who seem to be sincerely deluding themselves in efforts to avoid an
unpleasant truth. The public pretense is designed to head off social reproof.
When people are caught up in the same painful predicament, the result may be a
lot of collective public pretense.
The mechanisms of moral disengagement involve cognitive and social machinations but not literal self-deception. In moral justification, for example, people
may be misled by those they trust into believing that destructive means are
morally right because the means will check the human suffering of tyranny. The
persuasive depictions of the perils and benefits may be accurate, exaggerated, or
just pious rhetoric masking less honorable purposes. The same persuasory pro-

96

BANOURA

cess applies 10 weakening o sel-censure by dehumanizing and hlaming adversaries. In lhe rheloric o connicl, opinion shapers ascrihe lo lheir foes irralionalilies, barbarilies, and culpabililies 1ha1 color puhlic belies (lvie, 1980). In
lhese differenl inslances, lhose who have been persuaded are nol lying 10 1hemselves. le misleaders and lhe misled are differenl persons. When 1he misleaders
are lhemselves operaling under erroneous helies, lhe views lhey voice are nol
inlenlional deceptions. ley seek 10 persuade Olhers inlo believing whal they
lhemselves believe. In social deception, public declaralions by Olhers may belie
lheir privale belies, which are concealed from lhose being deceived.
In reduclion o sel-censure by ignoring, minimizing, or misconslruing 1he
dcleterious effecls o their actions, peoplc lack thc evidence to disbelieve whal
lhey already bclieve. le issuc o sel-dishonesty does not arise as long as one
remains uninfonned or misinfonned about thc outcomcs o one's actions. When
disengagement o sel-ccnsure is promoted by diffused and displaced responsibili1y, unctionaries carry oul lhc ordcrs o superiors and oten peorm only a small
subunction, al thal. Such arrangements enablc people to lhink o thcmselves
merely as subordinate instruments, rathcr than as agcnls, o thc entire cnterprise.
l 1hcy regard lhcmselves as cogs in thc intricate social machinery, they have
little reason to bclieve othcrwise conceming thcir iniliatory power. This is not lo
say that discngagemenl o sel-censure operates flawlessly. l serious disbclies
arise, cspecially al thc poinl o moral justificalion, people cannot gel 1hemselves
to bchave inhumanely, and i lhcy do, lhey pay lhe price o sel-conlempt.

ACKNOWLEDGMENTS
Preparalion o this chapter was acilitaled by Public Heallh Research Granl MH5162-25 from thc National lnslilute o Mental Health. Sorne sections o 1his
article includc revised and expanded material rom thc book, Social Foundations
o/ Thought and Action: A Soal Cognitive Theory, 1986, Prenlice-Hall.

REFERENCES
Akamatsu, T. J., & Farudi, P. A. (1978). Effccls o modcl s1a1us and juvcnilc offcndcr1ypc on lhc
imilalion o scl-rcward crilcria. )011,nal ofCons11/1ing anJ Clinical Psyclwlogy. 46. 187-188.
Allcn, M K , & Licbcn, R. M. (19691. Effccls o livc and symbolic dcvian1-modcling cues on
adoplion o a prcviously lcamcd slllndard. Jour,wl of P,rsonali1y and Soda/ Ps}'llwlogy, 11,
253-260.
Andrus, B C. (19691. TM infamo,u ofN,,mnrg. Loodon: Fnvin.
Arbulhnol, J. ( 1975). Modificalioo o moral judgmcnl duuugh role playing. D,v,lopm,n1a/ Psycltology, 11. 319-324.
Aron, 1. E (1977). Monl philosophy and moral cducalion: A criliquc of Kohlbcrg's lhcory. S,hool
R,vi,w. 85, 197-217
Auslcn, V. D., Rublc, D. N., & TrabasS(), T. (1977) Rccall and ordcr cccls as faclors in
childrcn's moral Judgmcnls Child D,v,lopm,m. 4H. 470-474.

1.

SOCIAL COGNITIVE THEOAY OF MOAALITY

97

Baldwin, T. F, & Lewis, C. ( 19721 V1oleoce m 1elev1,ion. The mdustry lool<s al uscl In (i A
Com!>lock & E. A. Rubmstein Cl'.ds l. Te/e,imm un,l sonul l>,,h,nwr Vol I M ..J,u nml<'nl und
mn,r.,/ (pp 290- 37.l). Washmglon, IX': U S Govemmenl l'nntmg Of:e
Ball Rokeach, S J ( 19721. lc leg1111na1111n ni v1oleoce In J I' Shon, Ir & M E Wolgang
(l'..ds. ), c.,11..,11v<' ,.,.,/.,n,<' (pp IOO 1111 Chi<:ago: Aldme Athenon
Bandura, A. (19731. A1111r<'nion A so, iul leurnm11 unu/ysu Englew,Kid ( 'hlls, NI. l'renticc- llall
Bandura, A. ( 19116) Soul founda1wns o/ 1lwu11h1 und u,1wn. A ,.,....,, n11n111<' th<'ory
Englcwood Clifs, NJ: l'renlke Hall
Bandura, A. ( 1990). Mechanisms of nK>ral disengagemenl In W Reu:h Cl'.d J. Orr11ms. psy,hol
"II<'s. id<'11l,111<'s, 1h""'"/1t'S o/ 1uwrrsm Stutes o/ minJ 1,p 161 191) New York. Cambridge
Universily Press.
Bandura, A. (1991). Social cogmlive 1heory and social reerenung In S hmnum (bl l. Soou/
u/a<'nctn11 und s,,.,ul ,ons/Tu<"lwn ,1,..ulif1 New York Plenum
Bandura, A., Grusec, J. E , & Menlove, F L. ( 19671. Sume ><1<:1al determmanls o self-RK11111 .. nng
reinfo,cemenl systems. Jm.rnul " Pusonulitv und Sonul Psv,holo11.v. 5. lb-B.
Bandura, A , & McDonald, F. J. (19631. lnHueoce o !>t11:1al remfonemenl and 1he behav1or o
models in shapmg chddren's RKJral judgmenls. Journal 1 A.bnormal und Soul Psy,hology, 7,
274-281.
Bandura, A., Ross, D , & Ross, S A. (1%.H. A cornparative test o the status cnvy, social powcr,
and secondary reinf01Ccmenl theories of 1den1ificalory leanung. Journal 1 A.bn.,nnal und Sonul
Psychology, 67, 521-534.
Bandura, A., Undcrwood, 8., & Frornson, M. E. (1975) l>1sinh1b11ion of aggression through
d1ffusion of respons1b11ily and dehunlllm.talion of vicllms J.,u,nul o/ R<'uunh m P<'rs"""lity. \1,
253-269.
Bandura, A , & Waltcrs, R. H. (1959). Adolnant ug11r<'uion New York: Ronald Ptcss.
Busiouni, M. C. (19111 ). Tcrrorisrn, law cnforccmenl, and the mus media: Pcrspec1ivcs, problems, proposals Journal of Criminal lw dr Crimm.,lo11v, 71. 1- 51.
Bergin, A. E. (1%21. lbc cfccl of dissonanl persuas1ve communications upon changes in a sclfrefcrring altiludc. Journal o/ P<'rsonality, 30. 423-4311.
Bemard, V., llcnberg, P., & Rcdl. F. (1965). Dehumanizallon: A composilc psychological dcfense in rclation lo modem war. In M. Schwcbel (F.d. ), 8 ..hu,,ru/ sci<'na anti human surv,vul
(pp. 64-82). Palo Alto, CA: Scieoce and Behavior Buoks
Bickel, A. (1974). Watergatc and 1hc legal ordcr. Comm..ntury, 511, 19 25.
Blasi, A. ( 1980). Bridging moral cognilion and nKJral action. A cri1iu1I rcv1ew of lhc lilcralure
PsycholotJicaJ Bull<'ti11, 88, 1- 45.
Bloom, A. H. (1986). Psychological ingredicnls o high level RKJral lhinking: A critique of lhe
rh<' TJi...,,y o/ s... ,ul B~huvwur, l. 119- IOJ.
Kohlberg-Gilligan paradigm. Journul
Blumenthal, M., Kalm, R. l.., Andrews, f. M., & Head, K. B (1972) Justifyin[l vwl<'na. Th<'
attitudrs of A.muicOII ffl<'n. Ann Arbor, MI: lnslilule for Social Research.
Bok, S. (1980). lbe self dcceivcd. Social Sci<'na lnforrnarwn, 1\1, 923 -936.
Bolingcr, D. (19112) wngua[l<'. Th<' lood<'d weupon London Longman
Boyd, R., & Richcrwn, P. J. (1985) 1,1.,,hunisms o/ ,ulturul nulu1wn. Ch1cago: Umversily uf
Chicago Ptcss.
Briere, J., & Malamulh, N. M. (191131. Sel-reponed likclih<KMJ uf seaually aggressive behav1or.
Alliludinal versus seaual caplanalions. JourfllJI 1 R<'s<'affh in P<'rsonul1ty, 17, 315-323
Brock, T. C., & Buss, A. H. (1%2). Dissonance, aggression, and evaluati<Hl of pain Journal uf
A.bnormal und S"'ial Psycholuu, 5, 197 - 202.
Bruck, T. C., & Buss, A. H ( 1964). Eecls o us1ifica1ion for aggression and commumcaliun w11h
thc viclim on poslaggress1on dissonance Journul 1 A.bn<>rmal und Sonul P1y,hulu11y. ll, 403
412.
Brody, G. H , & llendcrson, R. W. ( 1977). Efls o mulllple RK>del vamtllons 1U1d rallonak

.,,

98

BANDURA

provision on the moral 1ud11rnents and eplanation, of yrnm11 children Child n,,,lopmrnl. 4R.
1117 1120
C"armicharl. 1) J C f 19R2l Of hcasts. ?ods. and civili,ed rnen 11ie justificati,m of terrorism and
of counterterrorist rneasures T,rrori.,m. 6, 1 26
C"arter, R f (19Rllt Whal is Lawrente Knhlhcr! doin11? Journal nf Moral l:ducalinn, 9. RRI02.
<limplin. T S 11977) Sdf-deception A reflexive dilemma. Philo.,nphv. 52. 2RI - IQQ
C'hudler. M J . Green,pan. S . & Barenhoim. C (197.1) Judp:rnents of intentionality in response
to videotaped and vemally presenled moral dilemma: 11ie rnedium is the rnessage. 01ild D,vrlnpmrnl, 44, J"-320
(11-yne. J A , & Walters. R H (1970) Punishrnenl and prohihiti,m: Sorne orip:ins of self-control.
In T. M. Newcomh ((:d.). NrM"dirr,1ion., in p.,whofoRV (Vol. 4, flP. 2Rl-H7). New Yorlt: Holt,
Rinehart. & Winslon
(liurchill, R P ( 19RJI Nuclear arms as a philosophical and moral issue. Annab nf 1hr Am,rican
Arnd,mv nf Pn/i1ical and Social Srn,,. 469, 46-S7.
f'odd, J A. (1977) Sorne conceptual prohlems in the cognitive develorment arproach lo morality.
}ourna/ nf Moral F.ducalion. 6, 147- 1or;7_
Colhy. A Kohlhcrg. L . (iihm. J . & Liehcrman, M (19RJI A longiludinI study of moral
judp:rnent MonoRm('lt., nf 11t, Sof'irlV for Rr.,,.arclt in Child Drv,lnpmrnl, JR ( 1-2. Serial No.
200)

f'owu. P A., Lan11er, J Heavenrich, J & Nathanson. M. (1969). Social leaming and Piaget's
c~nilive theory of moral developrnenl }ournal nf P,r.mnalitv and Social P.,vcltolnRV, I I, 261274.
Darley. J. M , Klosson. E (' . & Zanna. M. P (197R). lntentions and their conluts in the morI
judp:rnenls of children and adulK Child n,vrlnpm,nt, 49. 6(, .. 74.
llllvidson, E. S . & Smith. W P .. ( IQRZ, lmilati,m, social comparisoo and self-rewanl. Child
/Jr,rlopmrnt. H. 92R-9J2.
Oenny. N W, & Duffy. 1) M. (19741 Possihle environrnenlal causes of slages in moral reasoning Journa/ nf <i,nrtic P.,v.-1,oloRV. 125. 277 - 2R4.
Diener, E (1977) l>eindividuati,m: Causes and consequences. Social Brlunior and P,r.mnalitv. 5.
141- 1"i6.
lliener. E. Dinttn, J .. fndresen. K. Reaman, A L. & Fraser. S. e (197"i). Effecls of altered
responsibility. cognilive sel. and modelin,: on physical aggression and deindividualion. Jnurnal
nf P,nonali,v and Social P.,vcholnRV, JI, J2R-B7.
Oomhusch. S M. (19R71 Individual moral choice, and social evaluations: A research odyssey. In
f. J Lawler & B. Markovsky fEd, l. Ad,ana., in Rrnur proa.,.,,.., Throrv and r,uarclt (Vol.
4. rr 271. :107) Greenwich. CT JAI Press.
llorr. 1) . & f'ey, S (1974) Relative power of ,ymbolic adull and pttr models in the modification
ol children , moral choice hchavior Jo11rnal nf Pn,onalirv and Social PncholORV, 29, J.15-341 .
Hkin, J'. &. Weslley, W A. (19~51 11ie mylh of adolcscenl culture. Amnican Sociol11Rical
Rrii,"' 20. 6RO 6R4.
1'11,worth. P ( 197R) Allitud,., tn1<11rd rnpital puni.thm,nl: Frnm applicalinn lo th,orv. llnpublished manuscript. Vale Universily. New Haven, CT
Emmcm,. R. A . & l>iener. E f l <.IK6 t Situation selecti,m a, 1 moderator of response consislency
and stabilily. }ourna/ of Pninn11/i1V and !i,Hia/ P.,v..hnloR.V, 5/. IOn-1019.
hlk, J f 19RJ) Tal:inR Au.,tralia nJJ tht' map New York: Penp:uin Rooks.
l'eild. H. S (197RI Allilude, loward rape: A comparalive analysis of pnlice, rapists. crisi~ counselors, nd cilizt,ns '""""'' "' Pnmnalitl' and Soda/ PncholnR_V . .16. 156-179.
l'ingarelle. H (1969) S,lf dnrption New York. Humanilit,s Press
l'is,-her, K W. fl9R.1). lll11minatin11 lhe pmcesses of moral developrnenl MonoRraph., oftht' Sodt'IV fnr R, ..,ar,-h in Child
.IRI 1 2. Serial No. 2110)

t>,,.,.,,.,.,,,,.,,,,

1.

SOCIAL COGNITIVE THEORY Of MORAlllY

99

Gamhino, R. (197.l, Novr,mhcr-Ottr,mhcrl Watr,rgatr, longo: A language o non-re<ronsihihty

Frudnm 01 l.uur. (No. 22). 7-9, l.'i 17.
Gihffln, J. T., & Haritos-Fatourm. M ( 19R6, Novemhcr). Thc cducation of a torturcr. Pw.-1,oloRY
Todoy, 1'1' ~-.'ill
Gilligan, C. (19112). In o diffrrrnr vnia: P.,vcholoRico/ rh,nrv ond wmnrn., drvt'l"f'wwnr.
Cambridge. MA: Harvard Univer!iity Prcss.
Grucncich, R. (191121 Thc dcvelopmcnl o childrcn's inle(lit11lion rules for makin(li monl jud(l:rncnts. Cliild Drvt'l"P"'t'nl, 53, 11117-ll<U.
Gruscc, J. E. ( 1971). Power and thc intcrnali1.ation of !IClf-dcnial Cliild Drwl,,,,,,.,n,. 41. 93- IM.
Gutkin, D. C. (1972). Thc effcct o systematic slory chan(l:CS on intcntionality in childrcn's monl
jud(lirncnts. Cliild Drvt'l"f'wwnt, 43, 1117-195.
Haan, N. ( 1911'.I). Processcs of moral dcvelopmcnt: Cognitive or MKial discquilihrium? Drvrlnpwwnrol Pnchnlt11tY, 2 I, 996- 1006.
Hai(l:hl. M. R. (1911C.I). A stuJv nf .,rlf drupfi,,,.. Atlantic Hi(l:hlands, Nl Humanilics Prcss.
Hallic, P. P. ( 1971). Justiftcalion and rchcllion. In N. Sanford & C. Comstock (Ecfs.), Sanctionsfnr
rvil (J'I'. 247- 263). San Fnnci~o: Jt~!ICy-Oa!<.s.
Hancy, C., Danks. C., & Zimhardo, P. (1973). lnterpcr.;onal dynamics in a simulatcd pri!IOfl.
lnlt'rnoti"""I Jnurnol nfCriminnln1tv and Prnmnu. l. 69-97.
Halano, G. ( 1970). Subjcctive and objcctive cucs in monl jud(limcnt. Jopon,11r PsycholoRic'ol
Rrsf'orch, 11. 96-106.
Hcchin(l:Cr, F. M. (19115, February 24). Down with doublcspcak. Son Francisco Chroniclr (This
Wnrld scction), p. 19.
Helm, C.. & Morclli, M. (1979). Stanlcy Mil(l:nim and !he obcdicncc ellpcrirncnt: Authority, le(l:ilimacy, and human IIClion. Political Thf'ory, 7, 321-346.
Higgins, A., Powcr. C., & Kohlhcrg, L. (1984). Student judgmcnts of rcspomihility and thc monl
atmosphcrc of high schools: A compmlfivc study. In W. Kurtincs & J. l. Gcwirtz (Eds.).
Morality, moral b,ltavinr and mnral df'wlopm,nt: Bo.,ic i/l/lUf'/1 in thf'ory and r,uarch (J'I'. 74106). New York: Wilcy lntencicnce.
Hildcbnndl, D. E . Feldlllllft, S. E . & Ditrichs. R. A. (1973). Rules, modcls. and sclf-rcinforttmcnl in childmt. }Ot1rnol nf P"-lity ,md Social Psyclenlnffy, 15. 1-5.
Hilgartner, S., Dell, R. C., & O'Connor, R. (1982). Nu.rsp,oA: Nuclf'Or la1tffUOR', 11isi01111. and
mind.,,.,. S1111 F11111Cilco: Sicrn Club Doots.
Hofflllllfl, M. l. ( 1977). Moral incemiilil.alion: Cumnl rhcory and rcscarcll In L. Bcrlcowitz (Ect. ),
Advanus in f'xpt'rimmtal !IOCial PIIYClenlORY (Vol. 10, pp. 86-135). New YOl't: Acadcmic ~111.
Holstein, C. E. (1973). Thc rei.tion of children's moral judgrncnl level to thal of thcir parenls and
to communication pdems in !he family. In M. S. Smart & R. C. Smart (Eds.), Adnlrscmu:
Drwl"f'mt"III ,md rf'lati01111hips (pp. 238-248). New York: Macmillan.
lvic, R. l. (1980). lma(l:CS of savagery in American justiftcalions for war. Cmnmunication Mnnnffroplu, 47, 270-294.
Janis, l. L. (1972). Victilfu nf 1rm,pdiinA:: A psycholnffical studv nffnrf'iRlf-poljcv drcisimu ond
fia.,rnn Roston: Houghton Mifflin.
Johnson, J. T. (19114). Can nwdrrn -r b, ftut' New Haven, CT: Vale Univer.city Prcss.
Kahncman. E., Slovic, P., & Tver!iky, A. (Eds.). (19112) Jud1mrnt undrr unurtainty: H,11ri.flic.,
and hio.<r.,. New Vorlt: Camhridge University Prcss.
Karlan, M. F. (19119, in prcssl. lnformation intetration in moral rcaMming: Conceptual and rncthodological imrlications. In J. Reykowski. N. F..i!ICnhcrt, & E. Staub (F.ds.). S,,ciol and mnral
valu,.,. Individual and .mcirrol p,r.<pt'Cti,rs. Hillsdalc, NJ: lawrcnce Erlbaum A!<.sociates.
Kavka, G. S ( 191111). Mnral porado:u.< nf nuclrar d,rrrrrncr. New Yort: Cambridge Univer!iily
Prcss.
Kay. S R 119112) Kohlhcrg's thcory of monil dcvelormcnt Critica! analysis o validation studics
with thc dclining issucs test. lnrrrnational Journal o/ P.,,clinlnR.Y, I 7. 27 -42.

100

BANDURA

Kcascy, C. B. ( 197.H. Expcrimcmally 111du<:cd changes in moral up1m1H1s and reasonmg. Journal 1
Prnonality and Soda/ P.rydw/ogy, l . .lO- .111
Kclman, H C ( 197.l)_ Violcncc wilhoul moral restrainl. RelleclitHlS oo lhe dehumaniTalion o
viclims and vicllm1zcrs. JtJurNJI of StJoal luu,s, l'J. 25 61
Kcnislon, K. (19611) You11g radicals. Ncw York: Harcourt, Bracc, & World.
Kcniston, K. (1970) S1udcn1 aclv1sm, mor.il devclopmcnl, and moraluy. Amrn.-an Journa/ ,1
Onltopsycliiatrv. 40, 577-592.
Kilham, W., & Mann, L. (1974) Levcl o des1ruc1ivc obcdicncc as a functioo o lransmitter and
cxccu111111 roles in the Milgram obcdicnce parad1gm. JourltUI o/ PrrsoNJlity and S,H1a/ Psy,h.,/ugy. 19. &>6- 702.
King. M. L. (1958). Stridr ,o,..ardfrudom Ncw York: Ball1111tine Books.
Kipnis, D. ( 1974). 1c powcrholdcrs. In J. T. Tcdcschi (E.d.), Prrsp,aivrs on social powrr (pp.
82-122) Chicago: Aldine.
Kohlbcrg, L. ( 1969). Stagc mid scqucncc: 1c cognitivc-devclopmcntal approach to socializatioo.
In D. A. Goslin (F.d.), HIJltdbooA of socializatiolt tlirory (pp. 347-480). Chicago: Rand
McNally.
Kohlbcrg, L. ( 197 la). From Is 10 Ought: How lo cornrnit lhe nahlralistic allacy and gel away with
il in thc study o moral dcvclopmcnt. In T. Mischel (F.d.), Co1mtivr tkwlop,MIII and rpistrmoloty (pp. 151-232). Ncw York: Ac.dcmic Prcss.
Kohlbcrg. L ( 19711>). Stagcs o moral dcvck,pmem as a basis for moral educatioo. In C. M. Bcck,
B. S. Cnncridcn, & E. V. Sulliv1111 (F.ds.), Moral rdwati01t: l1ttrrdiscipli1111ry approachrs.
Toron10: Univcrsily of Toron10 Press.
Kohlbcrg, L. ( 1976). Moral stagcs and moralization. In T. Lickona (F.d.), Moral tkvrlopmrnt and
b,liavior (pp. 31-53). Ncw York: Holl, Rineluut, & Win51on.
Kohlbcrg, L., & Candce, O. (1984). 1c rcllllion of moral judgmcnl to moral action. In W.
Kurtiaes & J. L. Gcwirtz (F.ds.), Morality, moral lwliavior and moral drvrlopmrnt: Basi<' iuurs
in tMory and usrarcli (pp. 52- 73). Ncw York: W1lcy lntcrscicncc.
Kramer, M (1990) 1c moral logic o Hizballah. In W. Rcich (F.d.), Origins. psvch<,loirs.
idrologirs. tlirtJlogirs o/ trrrorism. Statrs ., mind (pp. 131-157). Ncw York: Cambridge Univcrsily Prcss.
Kupcrsmid. J. H . & Wondcrly, D.M. (1980). Moral maturity and bchavior: Failure 10 firid a link.
Jountal of YOMlli and Adolrscr11a, 9, 249-262.
Kupcrsmid, J. H., & Wondcrly, D. M. (1982). Discqoilibrium as a hypolhetical construcl in
Kohlbcrgillll moral dcvclopmcnt. Cliild Stwly JOflrltUI. 12. 171 -- 185.
Kurtines, W., & Grcif, E. G. ( 1974). 1c dcvclopmcnl of moral thooght: Rcvicw and cvaloalion of
Kohlbcrg's approach. Psydw/ogical Bullrti11, B, 453-470.
Lackcy, D P. ( 1984). Moral pri11C1plrs and 1111.-lrar wrapons. Totowa, NJ: Rowrnan & Allanheld.
Lackey, D. P. ( 1985). lrnmoral risks: A dcoo1ological criuquc o nuclear dc1crrcnce. StHial Pliilosopliy & Policy. J. 154-175.
Lllllc, J., & Aridcrson, N H. (1976). ln1cgrlllion of intcnlioo and outcomc in moral judgmcnl.
Mrmory Ultd C ogniti01t. 4. 1- 5.
Lllllg, D. A. ( 1974). A reportcr al largc: 1c bank drama (Swedish hostagcs). N~w Yorlcrr. 50(40),
56-126.
Lllllgcr, J. (1969) Discqoilibrium as a sourcc of dcvclopmcnt. In P Musscn, J. Lllllgcr, & M.
Covinglon(F.ds.), Tultdsandisswsindrvrlopmrntalpsycholog_y(pp. 22-37). Ncw York: Holl,
Rinehart, & Winston.
Lancn, K S , Colcman, D., Forgcs, J , & Johnson, R. ( 1971). Is lhe subjccl's pcrsonalily or lhe
upcrimcmal &1lua1ion a bcncr prcdic1or of a subJCCl's willingness 10 adminisacr shock 10 a
viclim'' Journal of Prrsona/ity and Social Psy,:ho/ogy, 22, 287-295.
1..aVoic, J. C ( 1974) Typc of punishmcnl as a dc1crmin11111 of resistance 10 dcviation. Drvrlopmrn,al P,y,h<,/ugy, 10. 1111-1119.

SOCIAL COGNITIVE THEORY OF MORALITY

101

Le furgy, W. G., & Woloshin, G W ( 196'>1 lmmed1a1e ami long lenn dk.:ts ol expennienlally
induccd social influeocc III lhe modificauon uf adolesccnls' mural udgnienl lourna/ oj l'usona/lly anJ Soal Psy,ho/ogy, ll, I04 I IO
Leming, l S (197K> Cheaung hehaviur . .,,ualional 1111luence, and moral developmenl Journal oj
t:ducalwnal Rrsrarch. 71. 214 217
Lcon, M. ( 19801 ln1egral10n o 1111enl and conse<1ucnce informallun 111 <h1ldre11, moral 1uilgmcn1s
In F Wilkening, l Becker, & T Trabasso (Eds. 1, l,formalwn 1nlel(ra1wn by ,h,/Jrrn tpp. 71
97 > ltillsdale, NJ: l ..owrencc Erlbaum Asso<:iates
Lcon, M. ( 19K2> kulcs in children's moral udgnienls. lntegra11un ul 111le111, ilamage. and ral1unale
infonnation. Orvrlopmemal 1'1ydu,logv. IH, KJS K42
Leon, M. ( 1'>84). Rules molhcrs and sons use lo inlegrale 1111e111 and damage mlormallon m lheu
moral judgnienls. Ch,ld /)evrlopmrn1, 55. 21(16 211,
Lcmcr. M J., & M1ller, D T ( 1'>78) Jusi world resear,h and lhe a11ribu11on process. IAtok111g
back and altead. Ps.vdu,logi.-al llullt'lln. H5. HHO 1051
Lev,, P ( 1987). Thr dwwnrd anJ 1hr savrd New York. Su11111111 Books
L1eben, R.M., & Ora, J. P , Jr. ( 1%11). Ch,ldrcn's adopllon ul sell-reward pallems: lnccnuve level
and melhod o 1ransmission. Child Drvrlopmrn1 ..19, 537-544.
Lieben, R. M, Sprafkin, J. N., & l>avidson, E. S (19821 Thr rar/_y w1ndow. l:.ffr,n o/1rlrvisinn
on ,hildrrn and you1h 12nd ed > F.lmsford, NY Pergamon.
Locke, D. (1979). Cognilive s1a1es ol developrnenlal phases-Cri11<1ue o Kohlberg's s1age-s1ruc1ural lhcory of moral reasonmg. Journal uf Moral t:du,a1ion. H. 1611 181
Locke, D. ( 1911(1). lbe illusion o slage s,x. Journal o/ Moral t.Ju,a1ion. 9, 1(1] Hl'I.
Long, F. A , Hafner, D , & 8oulwcll, J (l:Js ) (1'>86) Wrapons ,n ,paa New York. Norton
Malamulh, N. M. 0981). Rape prochvily among males Journal ., S"nal /uurs. _17, 118 157
Malamulh, N. M., & Donnerslcm, E. (1984). Porno11ra1h.v anJ snual aggrrssion. New York:
Acadcmic Prcss.
Manlcll, O. M., & Panzan:lla, R. ( 1976). Obedience and resp<Klsibilily. 8ri1ish Journa/ uf Soal
and Clnica/ Psyclwlogy, 15, 239- 246.
Malcfy, R. E., & Ackscn, 8. A. (1976). lbe effccl o role-playing discrepan! p<mlions on change
in monl judgmcnls and auilUdes. )1111.rnal uf Grnrli< P,v,h11logv. llH. 18'> 200
McGuire, W. J. (1985). A11i1udes and ani1udc changc. In G. Lindzey & E. Ai-onson (hls l. Thr
ltandbouAofso.-ialpsyd1ology(Jrded., Vol 2, pp. 233-3461 Hillsdale, NJ. l.awrencc Erlbaum
Associarcs.
McMains, M. J., & Lieben, R. M. (1968). lnfluencc o discrepanc,es he1ween succcssivcly
modelcd scl-reward cri1eria oo lhc adopllon o a self-imposed slaridard. 11111.rnal "J Pers11na/uy
and Soda/ Psydwlogy, B. 166-171.
Mceus, W H. J., & Raaijmakcrs, Q A. W (1986> Administrallvc obed1ence. Carrying oul orders
10 use pi;ychological-admims1ra1ive violence. f:.umpran J.,urnal oj Sonal /',y,h.,/ogv. /, JI 1
324.
Milgram, S. (1974). Obrdima lo au1hllri1y. An rxprrimrn1al ,.,,,,.. New York. Harper & Row
M1schcl, W., & Licben, R. M. ( 19671 lbe role of p,.1wer 1111hc adoptioo of self rewanl pallems
Child Dr11elopnwn1. 38, 673-6113.
Mischcl, W., & Mischcl, H. N. (1976) A social-cognilivc learning approach lo moralily and selregulalion. In T. Lickona (Ed.), Moral drvrlopmrnt unJ IX'havwr (pp 84 I07> Ncw York:
Holl, Rinehan, & Winstoo.
Murray, f. 8. (19113). Lcarmng and developmcnt thmugh social m1cracllon and coof11d: A challengc 10 social lcarning lhcory. In L S Liben (Ed l. 1'1ugrl anJ 1hr Jound,.,ions o/ luu,wledgr
(pp. 231-245) Hillsdale, NJ: l..owrence Erlbaum Associalcs.
Mynall, C., & Herman, S J. (197') Resp,.1ns1bili1y a11nbu1ioo in groupi; and md1viduals. A d1rec1
1es1 of lhe diftusioo of resp<K1sib11i1y hyp,.t1hesis Journal o/ l'erwnality anJ S,,.,al Psvdu,loK~.
_ll, 1111 1118.

102

BANDURA

Olejnil<. A B (1 <JRO, Adults' moral rnmin11 with childttn Child n,v,l"f'm,nt . .H. l 2R5- l 2RR.
Osgood. C F.. ( l<)ROI The GRIT ,1rale1?y Ru/1,tin ,,( rh, Atmnic .'i,i,nti.m . .lt'i, 5R 60
Parl<e, R ll (1<)741 Rule,. role,. and ttsistance lo deviation: Recent advance, in punishmcnt,
di!ICipline, and scl-control. In J\ D. Piel< ffd. l. Minn,.mra .vmpmia t1n child p.<v.-11<,lo,tv (Vol.
11. l'I' 111-141). Minnearolis: llnivrrsity o Minne>la
f'tony. r, G . & Bu,scy. K f lCJ7<JI The icial leamin11 thcory o se diffettnces: lmitation i alive
and well J..,,rnal nf Pn,,.nnlitv ,md .fodal r.,ycholt>,ti . .17, l(,CJl)-1712.
Peny, D.G. & Busscy, K ll<Jll41 Soal d,v,l"f'm,nl. Englewood Clim. NJ: Prentice-Hall.
Peten. R. S. ( 1%61 F.thi.-.. and ,ducation London: Allen & Unwin.
Peten, R. S ( l<J71) Moral developrncnt: A plea or pluralism. In T. Mi!IChel (Ecl. ). Cn,tnitivt'
d,v,l"f'm,nt and ,pi.u,molo,tv (l'J' 237-2671 New York: Academic Press.
Piaget. J. ( 11)4111. Th, moral jud,tm,nt nf rh, child Glencoe, IL: Free Press.
Pia(!el, J. 11%01 F.quilil>ration and develtlJlmcnt o lo1tical ,tructutts. In J. M. Tanner & B.
lnhelder (F..ds l. Oiuu.uion., on ,hild J,...,,,,,,,,,,n, (Vol 4) New York: lntemalional Univenitie5
Pre,,
Powers. P C. & Geen, R. (i (l<J721 F.ffect, othe hehavior and lhe perceived arousal oa model
on instrumental ag(!tti,m. Journa/ nf r,r.mnalitv and Social Psvchnln,tv. 1.1. 175-110.
Rapoport. D. C.. & Aleunder. Y. IF.ds. l ( l<Jll21 Th, mnralitv """"ri.,m: Rrli,tinu.rnnd .t'Cular
iutifit-ation. F.lmsord. NY Per(!amtm J>re,s.
Reed. T M. (1')1171 (.kvel<lJlrncnlal moral lhcory. l:rhin. 97, 441-4~.
Reiss, A J., Jr. ( 1%5) . .forial or,tanizati"n and .,ocializatinn. Variation."" a th,m,
,trnrratinn.,. Workin(! paper No 1. Center for Rescarch on Social Or1tani7.ation. Univenity o Michi(!Bn, Ann Arhor
Re,1, J. R. (ICJ7.l) Panem o pttfettnce and comprehen,ion in moral jud1tmcn1. Journal nf P,rsonalitv. 41 . 116 Hl'J.
Re,1. J R. ( l<J75) Longitudinal sludy o the defining issues lest o moral judgmcnl: A strategy or
analyzin(! developrncnlal change n,.,,,,.,,,,,,n,a/ P.vchnlo,tv. 11. 7JII- 7411.
Re,1. J. R., Turiel. F. . & KohlherJ. L (1%<JI. l..evel o moral developmcnl as a determinan! o
preettnce and comprehension o moral judJrncnt, made hy others. Journa/ nf P,,.mnalitv . .17,
225- 252.
Roscnhan. D. L. (1970) The natural sociali7ation o altruistic autonomy. In J. Macaulay & L.
~rkowit7 (Eds ), Altrui.m and h,lpin,t bt'havior: Social p.rvchnlo,tirn/ .,tudirs nf '""'" anlt'crd
rnls arul con..,qu,ncr. (pp 251-268). New York: Academic Pre,s.
Roscnhan. D .. f'ttdericl<, F .. & Burmwe,. A. ( 1<)611) PreachinJ and practicinJ: Effects o channel
discttpancy on norm inlemali7ation Child l)rv,l"f'mt'nl, 39, 291-:lOI.
Ros,. M., & DiTecco, D ( 1975) l\n attrihulional analysis o moral judJmcnts. Journal nf Social
/.UUf'.<, .1 /, 91-109.
Rott!IChaeer. W. A. (191161 Leamin(! to he a moral aJent P,r.,nnali.,1 Forum, Z. 122-142.
Rule, R G , &: NeMlale, A. R. (1976> Moral judJmcnts o l!f!"'s,ive hehavior. In R. G. Geen &
E C O'NeI (F..ds ), P,r.,prcri,,., on a,t,tr,.uion (l'f' J7-60) New Yort: Academic Press.
Ru,hton. J. P (l<J75). (ienemsily in childttn lmmediate and ltmJ-term effects of modelin(!,
preachin(!. and moral jud1trncnt /""'"''' ,, P,,.,,.nality and Soda/ P.,vcholo11v. ZI. 4W-466.
Safitt. W (IQ7<J, May DI Thr fine ar1 n euphemism San Franci.,co Chronid,. I' IJ.
Sanday, P R ( l QR 1) The wcio-rnltural conlexl o rape: A cmss-cullural study. Journal ,,( Social
.17. ~ 27.
Sanord. N., & Comstock, (' (l<J71 l Sanction.< for ,../_ San Francisco: Jossey-Bass.
San Frann.,m Chmnid, (l<J7<Ja. Novemher 22). Award-winnin~ noclcar jar1ton. I'- 24.
San f'ran,i.rn Chmni,I, (19791,, April 221 Would you helieve iP I' B~
Schleier, M . & D0u11la,. V 1 (1<1711 F.ffect, of traininJ on the moral jud(!rncnt o youn(!
rhildttn. J,,u,nal o( P,r.wnalil\ and Smial r.,v,holo,t,. 2R. 62 611

Pre,,_

"""'

l.<.,u,., .

1.

SOCIAL COGNITIVE THEORY OF MORALITY

103

Scoville, H , Jr. ( IQR5) Reci,rocal nalional reslrainl: An allemalive palh lo anns rnnlrol Rullrr,n
of ,,., Am,rican AraJ,my of Aru and Sri,nrn. 311. 15-2Q.
Sears.R.R .. Maccoby,E E.&Levin.H (1957). Patt,rn.<ofrhildr,arinf(.Evanslon,IL Row.
Pelerson.
Shi,ptt. G. E & Christian. M. (197R). F.f(n-d,f,n.civ, attril,utinn.< fnr an nff,n.civ, pr,,t,/,m:
Attril,utinn.< of rr.<("'n.cil,ilitv and air pnllutinn .,.,,,;,v_ Pa,er presenled at lhe meeting of the
Western Psychological Association, San Francisco.
Shweder. R. W. (IQR2). Liberalism a desliny. Cnnll',npnrary P.<yrltnlnf(_V. 27. 421-424.
Simpson, E. L. ( 1974). Moral developmenl research. Human D,v,lnpm,nt, 17, 111- 106.
Skeyhill. T.(Ed ). (19211). S,rf(,ant YnrA:. Hi.< nwn lif, .<tnrv and war diary. Ganlen City, NY:
Ooubleday, Doran.
Snyder. M., & Cam,bell. B. H. (19112) Self-monitoring: The self in action. In J. Suls (Ed.),
P.c_vchnlnf(ical p,r.cp,rtiv,.< "" thr .<l'lf (,p. 1115-207). Hillsdale. NJ: Lawrence Erlbaum
Associates.
Sobesky. W. F.. ( 19IIJ) The effects ohituational faclors on moral judgmenls. Child D,v,lnpm,nl,

54. 515 5114.

Surber, C. F. (1971). llcvelo,menlal ,rocesses in social inference: Averaging of intentions and
con~nce in moral jud[!menl. t>,v,/npm,ntal P.<.vrhnlof(_V, 13. 6S4-665.
Surber, C. F. (19R5). Applicalions of informalion inle[!ration lo children's social cottnitions In J.
B. Pryor & J. D. Day (Eds.). Th, d,v,lnpm,nl of .enria/ rnf(nition (pp. 59-94). New York:
Springer- Verlag.
Tilker. H.A. (1970). Socially responsible behavior a. a function of observer responsibility and
victim feedback. Journal of P,r.cnnality and Snrial P.cyrltnlnf(V. 14, 95- 100
Turiel. E. ( 1966). An e~,erimental test of the !ICqllffltiality of developmental stages in lhe child"s
moral judgment5. Jnurnnl of P,r.cnnality and Snrial P.<vchnlnf(y, 3, 611-6111.
Wachs. T. D. (1977). The optima! lllimulation hypothesis and early development: Anyoody got
match7 In l. C. Uzgirill & F. Wei1.mann (Eds.). Th, .<trurturinf( of ,xp,ri,nr, (pp 153- 177).
New York: Plenum.
Walker, L. J. (1983). Sources of cognitive conflict for stage transition in moral development.
Drv,l~ntal P1ycltolof(y, 19, 103-1 IO.
Walker, L. J., & Richards, B. S. (1976). The effecls of a nanative model on children'5 moral
judgmenls. CaNMlian Journal of Brhavinral Scirfl<'I', 11, 169-177.
Walters. G. C., & Gru11e1:, J. E. (1977). Puni.,hm,nt. San Francisco: Freeman.
Wonderly. D. M. & Kupfersmid, J. H. (19IIOJ Kohlberg's moral judgment pmgram in the clas.sl'CXffll: Pnc:lical considerations. Alb,rta Jnurnal of Educatinnal R,1,arch. 16. 128-141.
Zillman. D . & Bryant. J. (19114). Effectsofmassiveexposure lo~raphy. In N. M. Malamuth
& E. Donnerstein (Eds.), Pornof(rt1ph_v and 1,xual Of(f(Tl'J.fi"" (pp. 115-1311). New York: Academic Press.
Zimbardo, P. G. (1969). The human choice: lndividuation. reason, and order versus deindividualioo. impulse. and chaos. In W. J. Amold & D. Levine (Eds.), N,bra.cla .<ympo1i11m nn motivatinn (pp. 237-309) Uncoln: UnivefSily of Nehrash Press.
Zimmerman. B. J., & 81nm. D. E. (1983). Toward an empirical lesl oflhe role of cognitive conflicl
in leaming. 0,.,,1,.,,m,ntal R,vi,w, 3. 18-J8.