8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
1/35
* cademy of Manag ement Review
1994.
Vol. 19, No. 1, 17 -50.
A U T H O R I T Y A T W O R K : I N T E R N A L M O D E L S A N D
TH EIR O R G A N IZ A T IO N A L C O N S E Q U E N C E S
W IL L IA M A . K A H N
K A T H Y E . K R A M
B o s t o n U n i v e r s it y
This article focuses on how organization members authorize and de-
authorize both others and themselves in the course of doing their
work. W e argue that these authorizing process es are sh aped , in part,
by enduring, often unacknowledged stances toward authority
itself
In tum, we suggest that these stances are enacted in similar ways
across hierarchical and collaborative work arrangements and across
various roles and p ositions. T hese stan ces ar e— as H irschhom (1990)
suggested—internalized models W orking from a th eoretical frame-
work that combines concepts from developmental and clinical psy-
chology, group dynamics, and organizational behavior, we define
and illustrate three types of internal models of authority: dependence,
counterdep enden ce, and interdepend ence. W e offer propositions
about how these internal models influence organization members
behaviors during task performances generally, and more specifically,
as members of hierarchical dyad s and work teams. W e also suggest
propositions about how these internal models of authority are trig-
gered and change in the context of organizational life. Finally, we
offer research methods and strategies by which to empirically exam-
ine these propositions.
Quite a lot is known about the nature and use of authority in tradi-
tional hierarchical organizations. Authority is defined as the given right
to perform roles; such rights are legitimated by consensual decisions
codified in constitutions, contracts, charters, rulings, and other accepted
institutiona l sanctions (Cartwright, 1965; Gilman, 1962; Katz Kahn,
1987).
Work organizations depend on mem bers occupying roles of author-
ity to ensure the pred ictable performance of organ izational ta sks (Simon,
1947). It is when organiza tion members occupy their work roles (i.e., iden-
tify them selves with the authority m and ated to those roles) that they have
the legitimate power to pursue their rights, duties, and obligations in the
service of their tasks. Authority offers a legitimate base to have power
and from which to influence others and bring about the completion of
work task s. It is legitimate power vested in particular people or positions
for system purposes (Weber, 1947).
This definition of authority is particularly well suited to traditional
hierarchical organizations that operate according to powers vested in
specific offices and, therefore, officeholders. Organization members' tra-
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
2/35
8 cademy of Management Review January
ditionally legitimate rights to wield power derive from occupying offices
that have affixed to them particular rules of influence, which specify the
officeholders by whom they are influenced and those over whom they
wield influence (Barnard, 1938; Simon,
1947;
W eber,
1947).
These rules are
woven into the fabric of the traditional hierarchical bureaucracy, giving
order and predictability to transactions among officeholders. It is increas -
ingly the case, however, that traditional hierarchical bureaucratic orga-
nizations are changing, and with them are changing the ways in which
authority and power are distributed among their members. Handy (1989:
130) wro te convincingly that the cha ng ing complexity, variety, and
spread of reaction which is now a feature of so many organizations
makes it increasing ly difficult to specify and reify in advance exactly who
should be doing what, when, in what order, and with whom for successful
task performance. Thus, organization members must negotiate such pa-
rameters themselves. Hirschhorn (1988, 1990) pushed this idea further,
noting that the pace, depth, and accumulation of change in the postin-
dustrial organizational setting requires the maximum use of human re-
sources. Such use, in turn, results in collaboration between leaders and
subord inates whereby duty and authority are negotiated: The leader no
longer charts the organization's work, with subordinates lined up to do
the bidding. Instead, the leader and the subordinates must collaborate
(Hirschhorn, 1990: 529). Such collaboration is at the core of w hat Lawler
(1988) describes as high involvement systems.
Such collaboration is based on negotiated authority, whereby leader
and subord inates jointly decide the scope of the power each h as over their
task s (Handy, 1989). Such decisions authorize lead ers and subord inates to
be resp onsible for certain aspects of task performance. Authorizing is the
giving of authority, that is, the right to do work. Organization members
are authorized not simply when they are ass igned responsibility for ta sks
(i.e.,
de lega ted authority) but also when they are supported by others who
are either formally or informally connected to those roles. Organization
members are de-authorized when such support is withheld (even when
their rights have been formally delegated). In traditional bureaucratic
organizations, officeholders become authorized by the power of the of-
fices they occupy. In more collaborative work arrang em ents , organization
members become authorized le ss through their identification with partic-
ular offices and more through their negotiations with other members
about task performances. As these newer organization forms become
more prevalent and necessary, it becomes increasingly appropriate to
conceptualize authority in terms of its underlying process dimensions:
the ways that organization m embers authorize and de-authorize both oth-
ers and themselves in the course of doing their work.
Although authorizing and de-authorizing processes have not been
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
3/35
1994 ahn and iam
19
thority to accomplish tasks is not simply a matter of legitimation and
mandate, it is also a
result of ctu l
interactions between
le ders nd
followers
(Gabarro
Kotter, 1980). It is within such interactions that the
scope an d limits of the authority of both leade rs an d followers a re nego-
tiated (Bendix, 1974). Both lead ers an d followers are separa tely interpret-
ing the nature of the leader's authority (i.e., given rights, duties, obliga-
tions, privileges, and powers), which cannot translate into power and
influence un less it is acknowledged by followers a s valid (Gouldner, 1954;
Simon, 1947). Interactions between leaders and followers become joint
interpretations of authority, because each one can increase or decrease
the other's authority by offering or withholding legitimating support
(Bass, 1990) irrespective of formal delegation of task responsibility. This
interpretive process is often unconscious, a s Barnard
(1938)
pointed out in
defining the zone of
indifference
to describe how followers autom atically
defined their lead ers' orders as acceptable u nless the illegitimate nature
of those orders (on various dimensions) triggered their conscious ques-
tioning. What is ac tually triggered is the conscious process of authorizing
and de-authorizing oneself and others to engage in work.
Authority Relations
Researchers know little else directly about authorizing and de-
authorizing processes in work organizations. However, there is a long
tradition of research and theory on authority relations (which generally
does not include the new collaborative organization forms), which po ints
to two fundamental types of influences—sifuafionaJ and individual— on
how organization members define and create their authority relations.
Each of these influences shapes the dynamics of authority and power
(Bass,
1990; House, 1988). The situational factors generally focus on how
organization members are externally driven to conform with existing
norms of thought and action, with a primary focus on followership. The
literature on individual factors generally relates to how individuals are
internally driven toward power in certain ways, with a primary focus on
leadership. Each influence offers clues to key components of authorizing
and de-authorizing processes.
Situational factors Research in the domain of situational factors has
been focused on how the social structure of situations presses individuals
to create and obey rules of hierarchical authority. Perhaps the most well-
known research is Milgram's (1974) set of experiments showing the con-
ditions under which subjects obeyed authority figures to the extent that
they acted inhum anely toward others while disavowing responsibility for
their actions. In the initial experiment, 65 percent of the subjects followed
the authority figure's instructions to continue pun ishing the learner
(i.e.,
confederate) until they had reached the maximum shock intensity
of
450 volts. Milgram noted that the key variable was the sense of dimin-
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
4/35
20 Academy oi Managem ent Review January
Variations on the basic experiment showed that manipulations that in-
creased the psychological distance between teacher and learner de-
creased subjects' perceptions of personal responsibility, which increased
their obedience to authority. Milgram's experiments indicated the power
that social situations have to dictate how members enacted their roles as
both subordinates (to legitimate authority) and authority figures (to
learners ). They showed how even temporary social systems exert pres-
sures on peop le to act a s if they have no choice but to create a nd reinforce
particular types of authority relations.
Another classic social psychological experiment attests to the power
of situations to determine how people create and respond within authority
relations through the roles they assum e. Z imbardo's (Haney, Banks, &
Zimbardo,
1973)
experiment involved the creation of a temporary system,
a prison, and the randomly assigned casting of individuals into system
roles, prisoner or guard . In little time the prisoners accep ted them-
selves as inferior and acted passively, whereas the guards accepted
themselves as superior and engaged in episodes of abusive, authoritar-
ian behavior. The subjects thus projected themselves emotionally and
cognitively into the roles into which they were physically place d, on the
ba sis of their stereotypic und erstandin gs of the norms by which prisoners
and guards act and prison systems operate. On the basis of accepting
those norms, Zimbardo's subjects created and enacted stereotypic rela-
tions of authority between the powerful and the powerless. The experi-
menters halted the experiment six days into a planned two-week simu-
lation becau se of the emotional force with which the subjects took up their
roles as superior an d subordinate. The experiment's duration w as enough
to show how powerfully the roles that individuals occupy—even in tem-
porary social systems—shape the relations of authority they create and
enact. It also showed how drawn people are to adopt norms to help them
define their situations and themselves, regulate their behaviors, locate
themselves hierarchically, and create authority relations.
These classic studies indicate the power of roles and norms to shape
people's experiences and behaviors in authority relations. O rganizations
rely on people occupying given roles to reduce the variability, instability,
and unpredictability accompanying human behavior and to withstand
personnel turnover (Katz
Kahn, 1987). Also, they traditionally have re-
lied on predictable authority relationships between superior and subor-
dinate that follow accepted norms of relative power and powerlessness,
respectively (Simon, 1947). Although these authority relations typically
are neither so brutally polarized as those evidenced in Zimbardo's exper-
iment nor so explicitly fraught with anxiety an d p ain a s those evidenced
in Milgram's work, they are nevertheless subject to similar social, role,
and normative pressures (and may lead to equally brutal results; cf.
Arendt's [1965] banality of evil . The more contemporary var iable of orga-
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
5/35
1994 ahn and iam
21
the situational influences that create and maintain particular types of
authority relations.
The implication h ere is that when individuals enter into both tempo-
rary and ongoing systems, there are cues that help them take up certain
types of roles (e.g., teacher , prisoner ), follow certa in behavio ral
norms, and create certain types of authority relations. These cues lead to
the authorizing of formal leaders occupying certain offices, regardless of
the individuals occupying particular superior and subordinate positions.
What is left unexplained, however, is why some individuals in such sit-
uations do
ot
create expected authority re lations; that is, why
35
percent
of Milgram's (1974) subjects did not show complete obedience to the au-
thority figure but acted counternormatively. There are individual-level
differences that—in conjunction with situational influences—help ex-
plain how individuals occupy superior and subordinate roles.
Individual factors A variety of indiv idual factors h elp account for
how organization members frame and perform superior and subordinate
roles.
In terms of enduring personality variab les, two concepts illuminate
the relation between personality and power. One concept is the authori-
tar ian persona lity (Adorno, Frenkel-Brunswik, Levinson,
Sanford, 1950),
measured by the F-scale, which is associated with power-seeking and
personality attributes that include conservatism, emotional coldness,
hostility toward minority groups, and resistance to change (Bass, 1990)—
attributes indicating rigidity. Researchers have explored the relations
between individuals' F-scores and the types of leaders they prefer (high
F-scores prefer autocratic leaders and low F-scores prefer consultative
leaders), and the effects that authority-relation matches and mismatches
have on both leaders and followers (see Bass, 1990). A second concept is
M achiavellianism (Christie & Geis, 1970), which refers to the extent to
which people are impervious to and resist social influences and emo-
tional or moral considerations (high Machs), or are susceptible to such
influences and are distracted by interpersonal concerns (low Machs). Re-
searchers have noted the generally consistent relation between ex-
pressed Machiavellian attitudes and behaviors (Epstein, 1969). According
to these concepts, individuals have particular motives or needs to estab-
lish specific types of authority relations in which they feel comfortable.
The implication of thes e trad itional concepts, for our purposes h ere,
is that people are drawn to create or enact authority relations partly on
the basis of compelling, deep -seated personality attribu tes of which they
may be only partly aware (McClelland, 1985). Recent research similarly
suggests a connection between individual differences and constructed
relationships. Researchers have noted how various self-concepts shape
organization members' abilities to perform effectively. Concepts that
ha ve received attention in this regard include self-efficacy (Bandura,
1982), self-confidence (Mowday, 1978), self-understanding (McCall, Lom-
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
6/35
22
cademy of Management Review
January
the research is that members who are high on such self-concept dimen-
sions are better able to create and push toward goals, and they display
(and gain) leadersh ip characteristics (Bennis Nanus, 1985). A related
research stream has focused on the personal characters of leaders and
executives, in whom certain tendencies toward defensiveness (Argyris,
1982, 1990) or personal achievement (Kaplan, 1991) lead toward the cre-
ation of particular (and variously effective) au thority rela tions. There als o
is a more unconscious, psychopathological relation between character
and leadership style, as was noted by Kets de Vries and Miller (1987).
They describe how neurotic styles of leaders, based on unconscious fan-
tasies,
create shared pathologies in their systems mirroring those fanta-
sies (e.g., persecution, helplessness, narcissism, compulsiveness, schiz-
oid detachm ent). The vehicles by which these neurotic styles are enac ted
and shared are the authority relations established by leaders and mem-
bers as they interact.
The individual factors described here help explain how organization
members engage in particular types of leadership behaviors, based on
personality attributes and self-concepts, that offset some of the situa-
tional influences reviewed above. The literature reviewed here suggests
that individuals are internally motivated to repeatedly develop certain
types of authority relations that enable them to use or react to power in
ways that are comfortable or necessary for them, for whatever conscious
or unconscious reasons. The authority relations that individuals create
are thus the vehicles through which they satisfy their needs or express
their attributes.
Implications. This brief review of the traditional literatures on au-
thority relations suggests particular gaps in knowledge. First, it seems
that the theory and research about situational influences on authority
relations is more developed than that about individual-level influences;
for example, researchers have not yet conceptualized an individual dif-
ference variable that would speak directly to what drives people to au-
thorize and de-authorize themselves and others in patterned ways. Sec-
ond, it is clear that researchers have, as noted above, generally linked
followership (i.e., obedience) to situational influences and leadership to
individual difference influences, and they have kept the two domains
separate (for an important exception, see Burns, 1978). It is likely, how-
ever, that followership and leadership are more tightly linked within in-
dividuals; that is, people have particular stances toward authority rela-
tions that affect their actions as hierarchical superiors and subordinates
alike, and in the more collaborative work arrangement, as co-workers
(Hirschhom, 1990). This article begins to fill these gaps in knowledge by
building on two implications from the literature reviewed above.
First, the situational literature implies that authorizing is related to
the al lowing of personal thoughts, feelings, and beliefs—one s own and
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
7/35
1994
ahn and iam 23
are imported into role performances. When the 35 percent minority of
Milgram's subjects authorized themselves to be responsible for their ac-
tions as subordinates, they did so by taking seriously their personal be-
liefs and feelings and using them as the standards by which they acted.
Most of M ilgram's sub jects, like those of Zimbardo, did just the opposite,
keeping their personal sensibilities out of the roles—superior or subor-
dinate— into which they were cast. In both situations, there were extern al
cues suggesting that subjects split off their personal feelings, beliefs, and
though ts and leave them o utside the confines of given hierarch ical roles.
It is likely that subjects' resp onses were re lated to some sort of individual-
level factor that led some to follow internal cues directing them to access
rather than split off personal dimensions and take rather than deny per-
sonal responsibility. Authorizing, then, is defined not simply in terms of
giving the right to do work; it is also giving the right to bring the per-
sonal self (one's own or another's) into the work role (see Gould, 1993;
Hirschhorn, 1985, 1990; Kahn, 1990a, 1992).
A second, less visible implication of the traditional literatures is the
importance placed on how childhood factors shape authority relations
and thus authorizing dynamics. Both situational and personality re-
searchers emphasize childhood factors. Milgram noted that one of the
crucial factors in obedience to authority was the socialization history
within one's family, whereas Zimbardo attributed the ease with which
subjects assumed roles to their experiences of power and powerlessness
relationships as children and parents. Individual-level researchers also
focus on childhood experiences, tracing the authoritarian personality to
particularly harsh, punishing parents (Adomo, Frenkel-Brunswik, Levin-
son,
Sanford,
1950)
and drawing specific connections between upbring-
ing and the characteristics of achieving, driven executives (Kaplan, 1991)
or neurotic leaders (Kets de Vries Miller, 1987). Both approaches thus
emphasize childhood experiences as influential in authority dynamics,
although as yet neither has yielded a comprehensive framework that
describes the connections between people's early and later authority re-
latio ns . Hirschhorn (1990: 541) hinted at one, when he noted tha t inter-
nalized models of authority figures . . . derived from childhood influence
authority relations at work (see Kahn, 1990b).
These implications help frame our approach to the authorizing and
de-authorizing processes that shape authority relations at work in both
traditional and collaborative organizational arrangements. Specifically,
we focus on how individuals authorize or de-authorize themselves and
others partly on the basis of enduring, often unacknowledged stances
toward authority itself—stances that are enacted in similar ways across
hierarchical and collaborative work arrangements and across various
roles and positions. These stances are — as Hirschhorn
(1990)
suggested—
infernaiized mo els developed in childhood that individuals typically
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
8/35
24 Academy
of anagement Review anuary
on recent theory and research in child developm ent, w e offer a framework
by which to understand such processes.
INTERN L MO DELS OF UTHORITY
Clinical psychological research (Bowlby,
1980;
Freud, 1936) suggests
that people tend to recreate the unresolved dynamics of past relation-
ships (with parents, siblings, and other important figures) and act as if
those dynamics are part of present relationships (e.g., with spouses,
bosses , an d co-workers). The psychoana lytic concept here is transference
(Freud, 1936): Impulses that have their source in early object re lation s a re
not created by the objective situation but merely revived by the compul-
sion to repeat early relationships. The colloquial expression here is that
of the persona l baggage that people carry with them to work and unload
on others. Clinical and developmental psychologists (Bowlby, 1980),
group dynamics theorists (Bennis & Shepard, 1956), and organ izational
psychologists (Argyris
Schon, 1978; Hirschhorn, 1990; Kets de Vries &
Miller, 1985) have in various ways suggested a more technical version of
this dynamic: Individuals have internal models of authority that shape
how they experience and act in social systems. This notion builds on the
concept of theories-in-use (Argyris
Schon, 1978), but it adds both a so-
phisticated theoretical base (attachment theory) and a particular content
focus (authority relations).
Attachment theory (Bowlby, 1973, 1980) focuses on how infants' early
attachments to primary caregivers determine enduring ways in which
they continue to attach themselves to significant others. On the bas is of
experiences with primary caregivers, infants develop internal working
models of the world and particularly their relations to attachm ent figures.
More specifically, infants who develop secure models of attachment re-
flect caregiving environments that are stable, consistent, and nurturing,
whereas infants who develop insecure models of attachment are accu-
rately reflecting caregiving environments in which nurturing is abse nt or
inconsistent. Internal models of attachment thus enable infants to carry
accurate pictures of their environments to guide their responses to pri-
mary care givers (Bowlby, 1980; Main, Kaplan, & Cassidy, 1985). Those
models guide behaviors and the appraisal of experience (Bowlby, 1973,
1980)
and continue into adulthood (Weiss,
1982).
Recent research sugg ests
that people's models of attachments influence ongoing relationships
other than those with paren ts an d other primary caregive rs (Hazan &
Shaver, 1987; Main et al., 1985). This is not to say that inte rna l models
developed in childhood continue unabated or unchanged into adulthood,
but that
aspects
of those models continue to shape behavioral tendencies
and adult relationships.
Although these internal models of attachment are primarily about
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
9/35
1994 Kah n an d
iam
25
first authority figures in people 's lives (Ainsworth, 1973; Bowlby, 1973),
and they set the template for what people expect in authority relations.
Consider, for example, a person who had powerfully negative experi-
ences with untrustworthy primary caregivers. Initially, the person will
maintain the belief that those in authority cannot be trusted. The person
may act as a subordinate in ways that are unauthentic and defensive,
and he or she may perceive others as m anipulative and deceitful, so that
superiors will be forced to appear punitive and demanding as they seek
information and resources from the withholding subordinate. Similarly,
the person may act as a superior in ways that lead subordinates to with-
hold and defend themselves, allowing the person to confirm the belief
that others cannot be trusted. The person thus acts from a model of au-
thority-as-untrustworthy to create untrustworthy authority relations. Like
self-fulfilling prophecies, peo ple's internal m odels of authority are con-
firmed as they construct authority relations in ways that shore up those
models (Kets de Vries
Miller, 1985). Such cycles remain unbroken until
individuals enter meaningful relations of authority in which they are
given feedback and new experiences that weaken the hold of their inter-
nal models (described further in the following section).
Three points need to be emphasized here about such internal models,
as conceptualized in attachm ent theory. First, people may be more or less
aware of their internal models of authority, that is, of the nature of their
beliefs an d how deeply emb edded they
are.
Although people are typically
un aw are of their inte rna l working models (Bowlby, 1980; Main et a l.,
1985; Sroufe Fleeson, 1986), these models may become acc ess ible to
individuals who understand the underlying patterns of their behavior in
relation w ith others . Individuals will therefore differ in terms of how con-
sciously aw are they are of their internal m odels. A second and related
point is that internal models tend to endure and shape people's relations
with others, unless peop le become aw are of them a nd chang e them in the
context of meaningful relations with significant others and therapists
(Argyris Schon, 1978; Egeland, Jacobvitz, Sroufe, 1988). As models
become increasingly am enable to change , they exert less force on behav-
ior and allow for situational influences (Hazan Shaver, 1987). Third, the
models are not of authority figures per se; they are of authority relations
(Kahn, 1990b)—particularly of hierarchical relations, given that internal
models develop w ithin the context of parent-child re lations. In ternal mod-
els of authority thus dictate how individuals will act in relation to one
another, in given or negotiated authority relations, irrespective of the
particu lar positions they occupy. The notion is that peop le strike various
poses or stances tow ard authority, rega rdles s of who occupies authorized
roles.
Drawing on work from interpersonal (Argyle, 1967; G abarro Kotter,
1980;
Hirschhorn,
1990;
Kets de Vries
Miller, 1985), group (Bennis
Shep-
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
10/35
26 Academy of Managem ent Review anuary
authority: dep end en t, counferdependenf , an d
interdependent
( see Tab le
1). Each s tance is character ized by a set of assumptions about authority
an d the pr inc iples on which re la t ions of author ity opera te . These as sum p-
t ions ma y be more or less expl ic i t, dep end ing on how aw are individu als
are of their exis tence and operation. Internal models also are character-
ized by p eop le s sense of self in relation to authority beliefs about how
their pe rso na l selves are affected by relat ion s of authority in h ierarch ical
systems. Like the assumptions about authority, these beliefs may be more
or less conscious. We also identify the patterns of attachment (Bowlby,
1973, 1980) to wh ich the mo dels correspond. Attachm ent th eoris ts ha ve
empir ica l ly documented three types of a t tachments be tween infants and
caregivers (Ainsworth, 1973). These types correspond to the three internal
m od els of auth ority in the next section s an d ar e useful referen ce p oin ts for
understanding the formation of the operating strategies tha t ad ul ts use
(knowingly or not knowingly) to enact their internal models in relat ions
involving authority. Operating s trategies are par t of internal models , yet
they are observe d throug h peo ple s beh avio rs at work an d in work rela-
t ionships . The s trategies guide people in enacting and reinforcing their
internal models of authority, leading to par t icular behavioral outcomes.
The following discu ssion h igh lights how p eop le s interna l mo dels of
author i ty shape the re la t ion between the ir personal se lves and the hier -
archical roles they occupy. The centrality of this theme reflects two con-
vergin g no tions. First , as n oted ab ove , authorizing is def ined in terms of
giv ing th e right to brin g the pe rso na l self (one s own or ano ther s) into th e
work role . Second, an ongoing s truggle for organization members is how
much they br ing relevant dimensions of their personal, authentic selves
into tas k p erfo rm an ce s (Gould, 1993; Hirsc hho rn, 1985; Ka hn, 1990a). In-
ternal models of authority are thus patterns of how individuals resolve
such struggles in authorizing themselves and others to work. The three
internal models of authority described in the following sections each offer
different resolutions: to deny the need for bringing the self into the hier-
arch ical role (suppress the self) , to deny th e need for the hiera rchic al role
itself (supp ress the role) , an d to m an ag e the ongoing relat ion b etw een self
and role (suppress neither) .
W hat follows a re rela tively p ure forms of pe op le s inte rna l m od els of
authority descr ibed in terms that clear ly dis t inguish them from one an-
other . Peop le s actions , beliefs , an d feelings in relat ion s involving au-
thori ty may conta in var ious shad ing s of the mod els an d resem ble asp ects
of more than one of the models . Indeed, as people mature into adulthood
and par t icipate in important relat ionships, they typically are able to re-
vise their internal models and allow for behaviors usually associated
with other mo dels (Hazan Sha ver, 1987). Peop le thu s ma tu re so tha t they
have the capacity to take on aspects of multiple internal models and to
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
11/35
994
Kahn
nd
iam
TABLE
Three Internal Models of Authority
Dependent Counterdependent Interdependent
Stance toward
Emphasizes hieiaichical
nature oi
authority
Underlying
assumptions
Sense oi seli
in relation
to authority
Corresponding
pattem oi
attachment
Operating
strategies
roles
of
superior
and
subordinate, whose
relationships
are
governed
by the
ru les of
iormal organizations.
Authority itself
is of
paramount importance.
Relationships structured
according to rules of
hierarchy.
Personal dimensions
of
people
are
suspect
to
the extent they
undermine authority
relations.
On e s self is found—
defined, constructed,
maintained— in
relationships
of
authority.
Hierarchical position gives
sense
of self;
without
such relationships, one s
self
is
lost.
Anxious resistant:
Uncertain
if
others will
be available,
responsive, helpful.
Tends
to
cling
to
authority relations,
anxious about exploring
world.
Emphasize hierarchy,
status differences.
Encourage dependency,
along hierarchical
structures {in seli,
others).
Idealize authority
and its
representatives
in
seli,
others).
De-emphasize personal
thoughts
and
ieelings.
Undermines
or
d ismisses
hierarchical roles
of
superior
and
subordinate.
Authority itself
is oi
minimal importance.
Authority
is
suspect
to the
extent it undermines
personal expression.
Nonrole data
are
trustworthy.
On e s self is found outside
hierarchy
and
relationships
of
authority.
In
such
relationsh ips, one s self
becomes lost: engulfed
or abandoned, denied
or
suppressed,
de-
constructed.
Anxious avoidant:
No confidence
in
others
helping, expects
rejection. Seeks
to be
emotionally
self-
sufficient, withdraws
from authority relations.
Dismiss status diiferences,
de-emphasize hierarchy.
Rebel against authority
(own, others ) with
confrontation
or
withdrawal .
Deny dependen cy
in self,
others).
Seek
to
pull self
and
others
out oi
role
relationships.
Emphasizes inter-
dependencies among
people occupying
various hierarchical
roles,
acknowledging
both person
and
role
dimensions.
Authority
is a
collaborative proc ess.
Different hierarchical
positions offer diiierent,
equally valid,
and
complementary
perspectives.
Relationships structured
in
terms
of
role
and
personal dimensions.
Authority
and
personal
dimensions
are
linked;
one without
the
other
is
suspect.
One s seli
is
found
in its
simultaneous
dependence
on and
independence from
hierarchical
relationships
of
authority.
Secure:
Confident
in
authority
relations,
in
which
others
are
avai lable ,
responsive, and helpful.
Bold in exploring world:
sense oi s imultaneous
connections
and
independence.
Emphasize person-in-role
within hierarchical
relationships.
Contribute personal
thoughts, feelings
within authority
interactions.
Acknowledge both
personal
and
role
dimension of self,
others).
Emphasize simultaneous
dependence
and
independence.
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
12/35
28 Academy of Ma nagement Review January
ally characterized
by one of the
three general models described
in the
following sections.
In our
analysis,
we
focus
on
these dispositional
ele-
ments as they shape authority relations at work.
Dependent Model o uthority
We define
the
dependent model
of
authority
in
terms
of
peop le's
de-
pendency on the rules and roles of formal hierarchy. People whose inter-
nal m odels are of dependency tend to establish relationships in which the
dependency
of the
hierarchical subordinate
on the
superior
is
high-
lighted, sought,
and
valued.
In
Schein's (1979) terms, people seek
con-
formity with established patterns of thought and behavior; in Turner's
(1976) term s, they adopt an institutional focus for their self. As subordi-
nates,
these people seek dependency
on
those
in
formal authority,
de-
authorizing themselves
to
take responsibility
for
managing themselves.
As superiors, these people seek the dependency of others over whom they
have authority, de-authorizing others to assum e responsibility for man-
aging themselves. They seek
to
structure relationsh ips
in
terms
of
formal-
ized relations between the roles that people occupy rather than between
the people themselves. As both superiors and subordinates, people with
dependent models
of
authority supp ress their personal selves within such
role-based interactions. We suggest that this suppression
is
based partly
on their assumption that such personal dimensions inevitably undermine
the strict re lations of authority on which they depend to guide their work
and work relationships. Given that assumption, they seek
to
split
the
person away from
the
role
and
leave personal dimensions outside role
performances.
We also suggest that such dependency is based partly on people's
sense that they will find their identities only within the context of hierar-
chical relationships; that
is,
that their person al se lves will find definition
only through the roles they occupy. Such self-definition is external. Em-
ploying this internal model, people depend on externally determined
rules
and
roles
to
guide their behaviors, beliefs,
and
feelings
in
relation
to others.
It is
within those role relations that people find them selves ,
and it is outside those relations that people feel lost. The dep endence is
on the scripts a ttached to hierarchical roles that offer characters to por-
tray (i.e., stereotypical characters
of
boss
and
employee ), lines
to say,
and plays
to
enact
cf.
Fiske
Taylor, 1984). This m odel echoes
the anx-
ious resistant pattern of attachment (Ainsworth, 1973; Bowlby, 1980), in
which infants who are uncertain about the availability of parents or pri-
mary caregivers tend
to
cling to those figures. These infants have anxiety
about exploring their world
and
wish
to
remain connected
to
authority
figures. Unless this type of internal model is replaced with another one,
the adults into which these children grow will maintain the desire to
remain connected both
to
authority figures
and to
au thority
itself, and
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
13/35
1994 ifah n a n d Jfram 29
We posit that people with dependent models of authority have oper-
ating strategies to maintain their dependency and that of others. Their
strategies involve emphasizing the status differences between them-
selves and o thers with whom they are hierarchically affiliated and acting
in ways to reinforce such differences: Superiors act in ways that disem-
power or de-skill subordinates so they will be needed by those subordi-
nates,
and subordinates act in ways that disempower or de-skill them-
selves so that they consistently feel the need for their superiors. In doing
so, both superiors and subordinates with dependent models idealize au-
thority and those in whom it is formally vested by organizations, and they
disparage the personal thoughts and feelings that people bring to rela-
tions of authority. In such a case, the ongoing deference to authority—
one's own and that of others—is maintained at the expense of people
using their own thoughts, feelings, and beliefs to help guide their work.
What is left is simply the ongoing deference to authority. Behaviorally,
this means that people will suppress—in themselves and others—real
thoughts and feelings, spontaneously gen erated ide as, an d the question-
ing of decisions based on personal values and ethical principles.
Counterd epend ent Model of uthority
We define the counterdependent model of authority in terms of peo-
ple's resistance to the rules and roles of formal hierarchy. People whose
internal models are of counterdependency tend to establish relationships
in which authority itself is minimized, undermined, and de-valued (cf.
Schein, 1979, on rebellion). In Turner's (1976) term s, such people ar e im-
pu lsives, who crea te their selves via spon taneous and often deviant
acts.
As subordinates, these people dism iss or undermine hierarchically
determined role interactions; as superiors, these people similarly seek to
step outside the boundaries of role-determined relations. In each case,
hierarchical relations are de-authorized; that is, people are not given the
right to do work in the context of hierarchical relations. This de-
authorization assum es various forms, ranging from the outright refusal to
cooperate in authority relationships (Schein, 1979) to the more subtle but
equa lly underm ining substitution of personal connections for role-related
interactions with others (Hirschhorn, 1985, 1990), that is, undermining the
authority relations while maintaining or even emphasizing personal con-
nections. As both superiors and subordinates, people with counterdepen-
dent models of authority thus seek to suppress authority.
Such counterdependency is based partly on people's sense that they
will find their identities only outside the context of hierarchical relation-
ships and that they will be lost if fused with the roles they occupy (i.e.,
personal identity becomes de-constructed rather than constructed in role-
based relations). In such a case , the desired m ean s of self-definition is to
resist external d em ands and to substitute countervailing personal behav-
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
14/35
30 cademy ol Management Review January
echoes the anxious avoidant pattem of attachment (Ainsworth, 1973;
Bowlby, 1980), in which infants who do not hav e confidence in parents or
primary caregivers, and expecting rejection from them, tend to distance
themselves from those figures. These infants have become emotionally
self-sufficient and su ppress their needs for help from authority figures.
Unless this type of internal m odel is replaced with ano ther, the adu lts into
which these children grow will maintain the desire to disconnect from
authority figures and from authority itself
People with counterdependent models of authority have operating
stra tegie s to maintain their dismissal of hierarchical role relations. Their
strategies involve de-emphasizing sta tus differences. This constitutes re-
bellion against authority (one's own and others') that may occur in the
form of direct confrontation of authority or passiv e w ithdrawal from rela-
tionships involving the use of authority. Both types of behaviors are at-
tempts to deny the dependency inherent in hierarchical relations. Such
behaviors a lso are , in Bion's
1961)
terms, responses to ang er at authority:
Active rebellion is the fight response, and passiv e withdraw al is the
flight response. Regarding the first response, organization mem bers
struggle to topple the authority structure (and their places within it); re-
garding the latter response, they deny the existence of authority. In such
cases,
peop le try to pull themselves and others out of the roles dictated by
hierarchy, explicitly or implicitly disparaging the structure and bound-
aries provided by authority relations and those who maintain them. The
ongoing undermining of authority—one's own and that of others—is
maintained at the expense of people's work connections and the organi-
zational systems (of communication, accountability, responsibility, and
coordination) that support their tasks.
Interd epen dent Model of uthority
Finally, we define the interdependent model of authority in terms of
people's emphasis on both personal and role dimensions in working with
others who occupy different hierarch ical positions. We suggest tha t peo-
ple whose internal models are of interdependency tend to establish rela-
tionships in which there are aspects of both dependence on hierarchical
authority (one's own and others') and independence from that authority.
They assume that people occupy hierarchical roles and make valuable
contributions; as unique individuals, they can make such contributions
from the context of their roles. The interdependency is thus first between
person and
role:
Neither the person nor the role is suppressed in ways tha t
undermine how living, thinking, feeling people perform roles according
to the guiding structures and boundaries of hierarchical systems
(Hirschhom, 1990). This interna l model assu m es that peop le occupy given
roles an d a re neither subsum ed by nor subsum e those roles (Kahn, 1992).
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
15/35
1994 Ka hn an d Jfram 31
m an ce of sh ar ed ta sk s (Gould, 1993; Hirschh orn, 1988, 1990). Pe op le w ith
this internal model trust in the value of both role and person, and they
believe in the usefulness of both authority and self-expression. Such peo-
ple perceive hierarchical systems as offering different and complemen-
tary vantage points for perceiving, learning, and acting. It follows that
these people would seek to collaborate with others and value what they
offer from their own roles as subordinates and superiors. We believe the
two types of interdependencies are l inked: As people seek to br ing their
individual voices into the performance of their hierarchical roles, they
seek out the voices of others who are hierarchically linked to them.
Such interdep ende ncy is ba se d partly on peop le 's sense that they will
find their identities by being both inside and outside the context of hier-
arch ical rela tionsh ips. The prem ise is that people define them selve s (i .e . ,
construct their identities) partly in connection to established systems of
roles,
bou nda ries, a nd authority , and partly in sep arati on (even rebellion)
from those systems. Self-definition is achieved by both accepting and
resisting connections to authority (one's own and others') , just as people
define themselves partly in relation to and separate from others in close
rel at io ns (Smith Berg, 1987). The im ag e he re is of ac tor s wh o dr aw on
both given stag e directions an d their internal se ns e of the chara cters they
portray to enact their roles. This model echoes the secure pattern of at-
ta ch m en t (Ainsw orth, 1973; Bow lby, 1980), in w hic h infant s feel con fiden t
that parents or pr imary caregivers are consistently available and respon-
sive to their needs while they maintain appropriate boundaries. These
infants ar e ab le to feel both self-sufficient an d trusting in primary c are-
givers. The adults into which these children grow maintain the abil i ty to
simultaneously connect with and remain separate from authority f igures
and from authority itself
People with interdepen dent mo dels of authority act in wa ys, bas ed on
their operat ing strate gies , to em pha size the person-in-role. They do so by
using their own thoughts, feelings, and beliefs to help guide their task
performances and work interactions without discounting their roles or
those of others linked to them hierarchically (Kahn, 1990a). People's
voices and energies are employed in the context of roles and the service
of the tasks. In such cases, collaborations within and across levels of
responsibil i ty and inf luence are sought. This means that people with
in terdepend ent models of au thori ty emphasize their s imu ltaneous depen -
dence on and in depe nden ce from others . Such people acknowledge s ta tus
differences without making them so prominent that personal dimensions
(in self and others) are lost. Superiors and subordinates with interdepen-
dent models use the structure and boundaries provided by authority re-
lation s without let t ing them selve s and their relations be dictated by those
sys tems .
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
16/35
32 Academy of Manag ement Review anuary
tion members with interdependent models of authority are able to em-
phasize both personal and role dimensions without sacrificing the integ-
rity of either dimension in relations involving the use of authority. Orga-
nization members with either of the other two internal models will split
personal and role dimensions, emphasizing one at the cost of the other,
and they will not fully engage themselves in tasks and relationships at
work. We are making a clear normative statement here: People with in-
terdependent models of authority are better able to authorize relevant
persona l dim ensions of themselves and others to work in roles of superior
and subordinate than people with either of the two other internal models
of authority we identify. Additionally, people with interdependent models
are better suited to the demands of the high involvement (Lawler, 1988)
and the postindustrial organization (Hirschhorn, 1988, 1990), which de-
pend on the joint negotiation of duty and au thority and the collaborations
that ensue.
LINKING INTERNAL MODELS OF AUTHORITY TO
BEHAVIORAL OUTCOMES
Internal models shape how individuals authorize and de-authorize
themselves and o thers as they take up organization roles, that is, to do the
work expected, to wield and be subject to influence, and to collaborate
with others within and outside hierarchical relationships. We focus here
on three specific areas in which such authorizing and de-authorizing oc-
curs:
task performance hierarchical dyads and teamwork. (Although
considera tion of other outcomes is beyond the scope of this pape r, they too
should be explored.) The first area focuses on individuals authorizing and
de-authorizing themselves to work; the second area focuses on authoriz-
ing and de-authorizing in both traditional superior-subordinate and men-
toring relationships; and the third area focuses on authorizing and de-
authorizing in traditiona l and self-managing work groups. For each a rea ,
we offer propositions about how people s tendencies to act on the bas is
of
aspects of internal models of authority shape their behaviors.
Task performance
How hard organization members work on as-
signed task s is trad itionally understood in terms of work m otivation: how
much people are compelled to perform capably by external (e.g., finan-
cial incentives) and internal (e.g., growth opportunities) rewards for do-
ing so. Organizations traditionally seek to strengthen their employees
motivations by enhancing reward systems (Steers
Porter, 1979) and job
characteristics (Hackman Oldham , 1980) in order to encourage employ-
ees to work m ore productively. U nderlying the work m otivation approach
is a lingering assu mption of Taylor s (1911) scientific managem ent: Em-
ployers can find the correct motivaters that a ctivate th e em ployee s en-
ergies to perform standard, externally-directed labors—much as one
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
17/35
1994 Kahn an d iam 33
does the need for employees to simply release energ ies an d effort in the
service of directed tasks. Increasingly, employees need to create new
methods and ideas, to direct themselves, to collaborate across roles and
hierarchical levels, and to think more critically and autonomously
(Handy,
1989;
Hirschhorn,
1990;
Lawler,
1988).
o
do
so ,
organization mem-
bers m ust be more psychologically present at work. They must feel and be
attentive and connected to their tasks and others, have various parts of
them selves acce ssible ra ther than split off and inaccessible to their work,
and focus on bringing those parts to the primary task. Most important,
they must be recognized and rewarded for being present in such ways.
The new language is of presence, which subsumes the vocabulary of
work motivation.
There are various influences on the extent to which organization
members are psychologically present at any moment in time. Kahn (1992)
described both systemic (or situational) and individual influences, the
latter including internal models that individuals have of themselves in
their roles (i.e., models of themselves as psychologically present or ab-
sent). He noted tha t ind ividuals vary in terms of how much they authorize
themselves to bring their personal selves into their task performances,
and he sugg ested that such self-authorizations are based partly on in-
ternal models on which people consciously or unconsciously depend to
guide their work relations. We extend that argument here, first noting
that there are a ctually two authorizing dynam ics involved: authorizing of
the role (i.e., supporting role-dictated behaviors) and authorizing of the
person (i.e., supporting the bringing to bear of personal dimensions—
thoughts, feelings, creative impulses, values, and beliefs—to tasks). We
suggest that each of the three internal models of authority has particular
implications for the extent to which organization members authorize the
presence of role and personal dimensions during task performances, for
themselves and others.
More specifically, we suggest that people with dependent models
will tend to authorize them selves and others to act from their roles during
task performances, and they will de-authorize themselves and others to
draw on personal dimensions in guiding those performances. These peo-
ple will accept the parameters of given roles—their own and others'—
and will seek direction from existing norms of thought and action rather
than create new methods and ideas, direct themselves, and think criti-
cally and autonomously. People with counterdependent models, con-
versely, will tend to de-authorize work roles— their own an d others'— by
directing behaviors away from the purposes of given roles. Such people
may do so through clear rebellion (e.g., simply doing or encouraging
things contrary to role purposes) or through more subtle underminings
(e.g., playing upon personal relationships to reshape expected role be-
haviors). These people are more likely to create new methods and ideas,
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
18/35
34 Academy of Managem ent Review
anuary
vice of de-authorizing given roles. Finally, we suggest that people with
interdependent models will tend to authorize both role and personal di-
mensions of themselves an d others. These people are most likely to crea te
new methods and ideas, to direct themselves, and to think critically and
autonomously in the service of meeting (and exceeding) expectations
about role-dictated behaviors. These people will resist impulses—their
own and others — to emphasize either role or person al dic tates, such tha t
one inappropriately eclipses the other.
Proposition
1: eople
with dependent internal models
of
authority will authorize role-dictated behaviors and de-
authorize personal dimensions of themselves and others
during task performances.
Proposition 2: Peop le with counterdependent internal
models of authority will de-authorize role-dictated be-
haviors and may authorize personal dimensions) of
themselves and others during task performances.
roposition 3 : eople with interdependent internal mod-
els of author ity will authorize both role-dictated behav-
iors and personal dimensions to coexist during task per-
formances, so neither one is emphasized at the expense
of the other.
These three propositions form the basis for the propositions that fol-
low.
Hierarchical dyads
The majority of hierarchical dyads genera lly take
one of two forms in organizations: those that are formally prescribed by
the organization to support task performance (e.g., boss-subordinate rela-
tionships), or those that evolve na turally to support the junior m ember s
learning and development (e.g., mentor relationship). Though the latter
may be formally assigned through a human resource initiative, more
often such relationships occur through the voluntary involvement of both
parties. Although some formally prescribed boss-subordinate relation-
ships also may evolve into a mentor relationship, they are discussed
separately here.
In formal hierarchical relationsh ips, the in ternal m odels of authority
that one ind ividual brings to the dyad m ay converge or diverge from that
of the other, with differing implications. We suggest that when both boss
and subordinate have dependent models of authority, for example, they
each invest energy into suppressing personal dimensions of themselves
that are relevant to their work together. This may include suppressing
creative ideas, ethical questions, or feelings whose absence m ay impede
communication or undermine work effectiveness. If both boss and subor-
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
19/35
1994 Yiahn nd Kiam
35
the com pletion of ass ign ed tasks . When both members of the dya d ha ve
in te rdependent mode ls of authority, they are a ble to construct a relation-
sh ip tha t gu ards aga ins t these dan ger s . In these re la t ionships , each per -
son is a ble to e n g a g e the relevant person al dim ensions of himself or
herself
and the
other within
the
context
of
their respective roles.
Proposition 4: When both dyad m embeis hold either de-
pe nd en t or counferdependenf internal models of author-
ity,
the relationship will
undermine
task
performance
through the joint suppression of person a] d imensions of
selves or denial of responsibiiity and expertise.
Even though hierarchical d yad s will be limited in their effectiveness
when e i ther dependent or counterdependent models are imported by both
members , the re
is a
degree
of fit
because bo th pa r t ies ' a s sump t ions ,
ex-
pe c t a t i ons ,
and
stra teg ies
are
comp lementa ry . When dyad me mb ers
br ing
different
models of authority to the relation ship ( i .e . , a poor fit), we
can anticipate quite different results (Gabarro Kotter, 1980). For e xa m -
ple,
if an organizat ion member wishes to und ermine author ity and his or
her super ior wishes to em phasize author ity , the two will construct a re-
la t ionship in which ea ch is at odds w ith the other; a play will commence
whose plot is insubordinat ion and wh ose reso lution w ill involve the un-
dermining of partic ipan ts , the relationsh ip, and th e work itself. Similarly,
if a subord ina te wishes to cling to a super ior who eschew s his or her own
authority
and
those
who
d e m a n d
it the
effectiveness
of
each
is
a g a i n
und ermin ed a long with their re la t ionship and their work. In each of these
cases (and in the other poss ible comb inat ions of the three internal m odels
previously descr ibed) the operat ing pr inciple is the s a m e : The lack of fit
betw een the super ior ' s internal model and that of the subordinate create s
actual re la t ionships that undermine ra ther than suppor t their work.
Proposition 5: Dyad mem bers will experience inteipei-
sonal conflict dissatisfaction
nd
difficulties with task
performance
when membeis biing
different internai
models
of
authority
to the
ielationship.
The internal models that organization members hold also inf luence
the extent to which they seek out and mainta in
mentoring relationships
which hold the promise of sub stant ia l task and perso nal l earn ing (Hall
Kram, 1981; Kram , 1988; Kram B raga r, 1992; M cCall et al ., 1988; Sch ein,
1978). In the early career ye ars , novices face the chal lenges of establ ish-
ing a work identity and niche, dev elopin g self-confidence, acq uir in g rel-
evant com petenc ies and knowledge , and prepar ing for a dva nc e m e n t and
grow th (Hall, 1976; Schein , 1978). De pen den t ind ivid ua ls will readily seek
mentors ' advice and counsel , w here as counterdepend ent ind ividuals are
more likely to a t temp t to master th ese sam e chal le nge s a lon e. It may a l so
be that those with dependent s tances are resis tant to enter ing the s e pa -
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
20/35
36 Academy of Mana gement Review January
pendent individuals may overcome resistance to seeking help, they may
be relatively unwilling to risk the self-disclosure that fosters the cultiva-
tion of deep mentoring alliances.
Similarly, those in the middle and late career years will face the
predictable tasks of reassessment and redirection, the threats of obsoles-
cence and aging, and the opportunities to become generative through
assu ming the m entor role (Kram, 1988; Levinson, Darrow, Klein, Levinson,
McKee, 1979; Schein, 1978). Again, it will be difficult, we suspect, for
those with a counterdependent stance to develop deepened relationships
with fledgling (and dependent) proteges . Similarly, it will be difficult for
these same individuals to ask for help in addressing the critical chal-
lenges of midlife and increasing organizational turbulence.
Proposition
;
/ndividuais w ith dependent models of au-
thority will more actively seek mentoring relationships
than those with counterdependent models of authority.
Proposition 7: Both dependent and counterdependent
models undermine the potential value of mentoring re-
lationships— the former thwarts movem ent through the
separation and redefinition phases and the latter
thwarts the degree of intimacy and personal learning
that can occur.
It app ear s ultimately that an interdependent stance holds the great-
est potential for deepening mentoring alliances in which mutua l benefits
of such developm ental alliances ar e maximized. In this case, both p arties
are willing to share relevant personal dimensions, to acknowledge and
work with hierarch ical/status differences created by the formal roles they
occupy, and to collaborate regarding both task accomplishment and per-
sonal learning.
eamwork
Teamwork h as become increasingly important in contem-
porary organizations, an d task forces and self-managing groups are com-
monplace (Hackman, 1987). Thus, in addition to traditional work groups,
effective teamwork (group members jointly applying knowledge and
skills to accomplish objectives that a re a ccep table to those who receive or
review task output) (Hackman, 1987 is now essen tial to interdepartme ntal
coordination, product innovation, and a variety of other critical organi-
zational tasks. We suggest that individuals' internal models of authority
will shape their participation—type and amount of effort, roles, strate-
gies for participating—across various types of groups.
For example, in the traditional work group, we can expect that the
dependent individuals w ill welcome direction from a formal group leader
(e.g., the boss) and also will have the tendency to de-authorize them-
selves and their peers. As a consequence, members' creative contribu-
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
21/35
1994 Kah n an d Kiam 37
and task forces, where formal authority is not prescr ibed, and, instead,
group members must coUaborat ively work toward accomplishing tasks
while authority is shared among members. In order for such individuals
to effectively contribute in this context and also facilitate other members
contr ibutions, they will ha ve to contend w ith con sidera ble discomfort an d
move beyond the dependent s tance.
Proposition 8; Group mem bers with dependent models
of authority will simultaneously desire direction and
support from the boss and other formal authority), and
they will discount their own authority and that of their
peers.
In contras t to dep end ent individuals , the group mem ber who holds a
cou nterd epe nd ent mo del of authority will welcom e the opportunity to par-
t ic ipate in a se l f -managing group where mem bers are expected to operate
autonomously. However , i t is also l ikely that while embracing their own
skills,
voice, and authority, such members will resist other members
attem pts to provide le ad ersh ip for the group. Also, if most m em bers hold
this s tance, there is a good l ikelihood that the group will distance i tself
from its formal le ad ers an d sever l ine s of support an d com mu nication that
are important to its task effectiveness.
Proposition 9: Group mem bers with counterdependent
models of authority w ill distance themselves from their
formal leaders, resist the authority of other group mem -
bers, and may. to varying extents, embrace their own
voices and creative energies.
This analys is sugges ts that group members with e i ther internaJ
model of author i ty—dependence or counterdependence—are l ikely to
un derm ine ( to som e degree) the poten tial of tradit ion al work grou ps, task
forces, and self-managing work groups. The implication here is that
members who hold an interdependent model of authority will be most
effective in maximizing effective task performance in groups. Such indi-
viduals seek col laborat ive re la t ions between hierarchical levels : They
are able to draw upon formal systems of authority, communication, and
control for support , gu ida nc e, s tructure, an d res ource s. At the sa m e time,
they a lso are able to assume ownership for task processes and outputs ,
and they will encourage their peers to do the same. This sense of owner-
sh ip see m s par t icu la r ly impo r tan t to the suc cess of se l f -m ana ging
groups, whose members must take personal responsibil i ty for work out-
comes, monitor their own performances, take corrective actions when
necessary, actively seek guidance from their organizations when neces-
sary, and help other people in other areas to improve their performances
(Hackman, 1986). Self-man aging grou ps need m em bers wh o ar e w ill ing
and able to take ownership of their own processes, and in doing so au-
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
22/35
38 Academy of Mana gement Review January-
more and more on self-managing teams for innovation, quality, and pro-
ductivity, they will need mem bers who bring an interdependent stance to
teamwork.
Proposition 10:
roup
mem bers who hold an interdepen
dent model of authority are most capable of utilizing
resources within and outside the group toward achiev
ing teamwork and organizational objectives.
This set of propositions offers a way to begin mapping the influence
that organization members' internal models of authority have on their
work and work relationships.
TRIGGERING ND CH NG ING INTERN L MODELS O F UTHORITY
In explicating the natu re and influence of interna l models of author-
ity, we have simplified the three m odels, treating them a s though they a re
constant and enduring (i.e., always operating and impervious to modifi-
cation from birth). Neither point is accura te; the reality is more subtle a nd
complex. Next, we offer propositions about how internal models of au-
thority might be triggered and how they might change at work.
Triggering Internal M odels of uthority
It is likely that there are real individual differences in terms of how
dominant an internal model of authority is over a particular person's
behav ior. Some individuals, for examp le, may be influenced a grea t d ea l
by their internal models, across various situations, whereas others may
be influenced to lesser exten ts, in particular and d iscrete situations. That
is, some people will automatically respond to many situations with the
strategies dictated by their internal models, whereas other people will
automatically respond to just a few situations and in the other situations
they will have a wider range of choices about how to behave. We under-
stand these differences in terms of how often an individual's internal
model is triggered. People will vary along this dimension, from those
whose models are triggered so often that they seem to react to all situa-
tions with the autom atic application of their models, to those w hose mod-
els are triggered so infrequently that it seems a behavioral aberration
when they are. We use the concept of triggering as a way of exploring
more deeply why internal models are activated , in terms of the functions
they serve for individuals.
Our premise is that it is when individuals experience enough anxiety
to make them feel insecure in their immediate situations that their inter-
nal model of authority is triggered. This idea reflects two principles from
research and theory on attachment. First, infants (Bowlby, 1973), youths
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
23/35
1994
Kah n an d Jfram 39
they experience anxiety. Those situations invariably occur when people
feel that their se nse of security is threatened, that is, when the world does
not seem predictable and familiar and the person's way of navigating
through emotions and situations is threatened. Second, internal models
function to provide security. When individuals feel threatened, they enact
behaviors whose aim is to re-create the sense of security (Ainsworth,
1973);
that is, they cling to, withdraw from, or reestablish connection and
then move away from attachment figures so as to create a relationship
that is familiar and comfortable. This notion leads to the following prop-
osition:
Proposition ii: Organization membeis operate from
tiieir internal models of aufiiority w hen they experience
work situations as insecure: They cling to dependent),
push away from counterdependent). or establish ties
while remaining independent of interdependent) given
roles and authority relations until they again feel se-
cure.
When do organization members experience the threat and anxiety
that triggers their internal models? There are many sources of stress in
organizational life, ranging from traditional sources such as task and
interpersonal demands (Cooper
Payne, 1978) to postindustrial sources
such as increasing competition, cost-reduction initiatives, the speed and
complexity of tasks, and the demands of collaboration (Handy,
1989;
Hir-
schhom, 1990). The ambiguous structure of high-involvement systems
(Lawler, 1988) itself creates stress, as individuals experience the absence
of the traditiona l hierarc hical structure and the re lative sense of security
it offered. Such stressors do not automatically lead to experienced threat
and anxiety nor to individuals' searches for security. Rather it is when
individuals perceive ev ents and situations as threatening to their se nse
of
security that they will feel threatened and anxious and behave accord-
ingly (Lazarus, 1966). Such perceptions are based both on how others
perceive and react to situations and on individuals' tendencies to per-
ceive their situations in particular ways.
orms Group and organizational norms exert pressures on how
members ought to respond to situa tions (Hackman, 1976). They also
shap e how system mem bers frame or interpret situations, in terms of how
familiar or threa tening the situations a re and w hat sorts of respon ses they
should call forth. When the prevailing norms dictate that certain situa-
tions (e.g., CEO succession) be treated as nonthreatening, it is less likely
that system members' internal models will be triggered. When norms
dictate that other situations (e.g., union-managem ent impasse) be treated
as threatening, it is more likely that members' internal models will be
triggered.
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
24/35
40 Academy of Mana gement Review January
ate
way s to respond to threat, which presumably serve to create a sense
of security in threatening situations. Appropriate responses may include
members clinging to, pushing away from, or acting within while remain-
ing partly independent of given roles and authority relations. For exam-
ple,
one organization's norms may encourage members to withdraw or
undermine their roles in expressions of resentment (counterdependent),
wh erea s another organization's norms may encourage m embers to put in
more hours on their tasks but not to expend ene rgies in thinking critically
and autonomously about those tasks (dependent). Though such norms
function (from the organization's perspective) to enab le members to expe-
rience so lidarity and comfort with one anothe r, they also may produce the
opposite effect: Individuals' own internal models of authority fit or do not
fit with the strateg ies required by the norms of their units, which presum -
ably leads to implications for members' experiences of security within
those units. This idea suggests the following proposition:
Proposition
12:
Individuals with internal models of au-
thority that fit with normative responses will have a di-
minished sense of anxiety threat and insecurity: indi-
viduals with
counternorm tive
internal models and who
are unable to adapt
their
behaviors will continue to feel
threatened and insecure.
Personal insecurity Individuals also differ in terms of how secure or
insecure they tend to feel, across various types of situations and relation-
ships.
That is, people differ in terms of how often they frame situations
an d even ts as threa tenin g and , therefore, how often they feel anxious an d
insecure (Greenspan
Lieberm an, 1988). Individuals w ho expe rience in-
security often, across situations, are likely to activate their particular
internal models of authority often in order to try and create familiar rela-
tionships by which to feel secure (Ainsworth, 1973; Bowlby, 1980). Thus,
these people will approach a great many situations by automatically
clinging to, pushing away from, or acting within while remaining inde-
pendent of given roles and authority relations. Because their models are
triggered by clues that are generated internally rather than perceived
externally, their actions may be completely inappropriate to the actual
dictates of the situation, such as when an insecure person attempts to
cling to the rules of hierarchical relations in an explicitly collaborative
context (e.g., self-managing team). This notion leads to the following
proposition:
Proposition 13: Insecure individuals may project rather
than perceive actual threats to their personal security
and thus activate behaviors that are unproductive for
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
25/35
994 ahn and iam 4
Changing Internal Models of Authority
We have suggested that when internal models of authority are trig-
gered, individuals act automatically. The point of changing these mod-
els, then, is to offer individuals a more conscious choice about how they
wish to behave in particular situations (i.e., to expand the range of be-
haviors they may app ly to different situations). We assum e that e ach of
the three internal models is adaptive in certain situations and maladap-
tive in others. For exam ple, for newcom ers to organizations, it is adap tive
to initially adopt a dependent stance; for members who are placed in
situations that call for them to be w histle-blowers on unethical behav iors
(and will penalize them if they do not), it is adaptive to adopt a counter-
dependent stance; and for members of self-managing teams, it is adap-
tive to adopt an interdependent stance. Changing internal models of au-
thority thus means learning how to escape the automatic application of
any single set of operating strategies and learning to act in ways that
meet one s needs an d situational dem ands .
Attachment theorists (Bowlby, 1980; Main et al., 1985; Sroufe
Flee-
son,
1986)
emp hasize that cha nging early models of attachment is difficult
bec ause such models operate out of immediate aw aren ess, resist chan ge,
and defend a gain st em otional pain . This idea is underscored by the work
of Chris Argyris
(1982, 1990;
Argyris
Schon, 1978), who developed theory
and interventions to enable organization members to become aware of
their theories-in-practice and the defenses they establish to maintain
those implicit theories. Both streams of research suggest the following
proposition, the components of which are elaborated in the following
paragraphs.
Proposition 14:
Changing internal mode ls is a two-stage
process
consisting of developing awareness
of one s
pat-
terns of thought and behavior and to varying extents
the psychological defenses used to maintain them) and
developing new ways of relating with others.
The less that people are aware of their internal models, the less they
are able to alter actions that derive from these models. Individuals gen-
erally develop such self-awareness in the context of relationships, in
which they receive direct or indirect feedback about how the ways they
consistently frame authority relations may be inappropriate to or ineffec-
tive in current s ituations. Such feedback informs people who are a ble to
understand that they are acting in ways that served them in the past but
are no longer always relevant. As noted previously, the internal models
that people developed to guide their reactions to primary careg ivers w ere
accurate responses to previous contexts (cf. Bowlby, 1980; Main et al.,
1985). When people leave those early environments behind, they must
cha nge internal models that guided their behavior and they must ad ap t to
8/17/2019 Authority at Work Internal Models and Their Organizational Consequences
26/35
42 cademy of Mana gement Review January
This process occurs when people receive feedback about the effectiveness
an d lack thereof of their behavio rs in relations of authority, in the context
of therapeutic or otherwise significant personal and professional rela-
tions (see Egeland et al., 1988; Greensp an Lieberman, 1988) and work
relationships involving significant feedback and challenging assign-
ments (McCall et al., 1988).
People chan ge interna l models not simply through feedback and self
awareness but through the different ways they experience being in au-
thority relations. For example, an individual with a counterdependent
model of authority relations may experience interdependence when join-
ing a boss, group, and organization that value s, reinforces, and promotes
(literally and symbolically) such interdependence. Through that experi-
ence, the individual may feel w hat it is like to engage both self an d role
in the exercise of formal authority and have those feelings validated
externally by other group and organization members. Slowly the individ-
ual may transform the counterdependent model into an interdependent
model as previous perceptions and behaviors become unu sab le an d cur-
rent perceptions and behaviors are reinforced in the context of supportive
relations