CHARISMATIC LEADERSHIP: STRATEGIES FOR EFFECTING SOCIAL CHANGE

C. Marlene Fiol University of Colorado - Denver Drew Harris Fairleigh Dickinson University Robert House University of Pennsylvania

Second revision February 1999 Please address all correspondence to: C. Marlene Fiol University of Colorado at Denver College of Business CB 165 PO Box 173364 Denver CO 80217-3364 303-556-5812 [email protected]

CHARISMATIC LEADERSHIP: STRATEGIES FOR EFFECTING SOCIAL CHANGE ABSTRACT

Due to their unique relationship with followers, charismatic leaders can be powerful agents of social change. Current theories of charismatic leadership have emphasized primarily the personality and behavior of leaders and their effects on followers, organizations, and society. This emphasis fails to uncover why and how the charismatic leader-follower interaction can generate social change. Our study draws on theories of social meaning to develop a process model of charismatic leadership. Empirical exploration of our model suggests that charismatic leaders employ a set of consistent communication strategies for effecting social change.

INTRODUCTION We have substantial evidence that charismatic leaders behave differently than noncharismatic leaders. Further, we know that charismatic leaders can generate radical social changes, and that the performance of charismatic leaders and their followers tends to exceed that of their non-charismatic counterparts. To date, however, we know very little about the processes by which leaders and followers interact to effect social changes (Meindl, 1992). There is a need to address the following unanswered questions: Why do charismatic leaders adopt certain behaviors? Why do their followers respond in predictable ways to those behaviors? How does the leader-follower interaction generate social change? In this paper, we address these questions by drawing on theories of social change and construction and destruction of social meaning. We present a model that begins to explain why and how the charismatic leader-follower interaction can generate social change. The model suggests theoretical propositions that we test empirically by content analyzing speeches of all twentieth century U.S. presidents through Ronald Reagan. The empirical results show that charismatic leaders employ consistent communication strategies for breaking down, moving, and re-aligning the norms of their followers. A REVIEW OF CHARISMATIC LEADERSHIP RESEARCH Sociologists, political historians, and political scientists have widely accepted the theory of charismatic leadership originally advanced by Weber (1947). To our knowledge, no one has subjected Webers theory to quantitative empirical test. However, several scholars have advanced additional theories that invoke the concept of charismatic leadership (Bass, 1985; Conger & Kanungo, 1987; House, 1977; Trice & Beyer, 1986), visionary leadership (Bennis & Nanus, 1985; Sashkin, 1988), or transformational leadership (Burns, 1978). These related theories have been subjected to substantial empirical investigation. We refer to this general class of theory as the neo-charismatic leadership paradigm. 3

While there are some differences among these theories, and while some may argue that there are some substantive differences among these theories (Bass, 1990; Yukl, 1994) which indeed there are, we argue that they all fit well within a more general paradigm. We take this position and use the label neo-charismatic leadership paradigm for a number of reasons. First, this new genre of theory has much in common with the Weberian conceptualization of charisma. As Weber asserted, all of these theories also assert that exceptionally effective leaders articulate visions that are based on normative ideological values, offer innovative solutions to major social problems, stand for non-conservative if not radical change, and generally emerge and are more effective under conditions of social stress and crisis. Second, charismatic behavior (visionary, change oriented, non-conservative) is either implicitly or explicitly a central concept in all of the theories of this paradigm. Third, all of the theories of this paradigm emphasize independent variables that appeal strongly to followers: symbolic leader behavior, visionary and inspirational ability, nonverbal communication, appeal to ideological values, and leader expectations for follower self-sacrifice and for performance beyond the call of duty. Fourth, while all leadership theories imply an underlying theme of performance improvement, the theories of the neo-charismatic leadership paradigm focus primarily on affective rather than cognitive dependent variables: follower emotional attachment to the mission and values espoused by the leader, emotional and motivational arousal, enhancement of valences with respect to the mission articulated by the leader, heightened self esteem, trust and confidence in the leader, and heightened intrinsic motivation. Fifth, all of these theories assert that leaders described as charismatic, visionary, or transformational generally have positive effects on followers and organizations that exceed those of leaders described in theories of non-charismatic leadership. Sixth, the term charisma has had an enduring and honorable tradition in the sociological literature, and the above independent and dependent variables of the neo-charismatic paradigm are consistent with the traditional 4

charismatic literature. Seventh, the similarities among these theories are, in our opinion, far greater than their differences. Eighth, by grouping these theories all within a common paradigm we call attention to their common essential elements. Ninth, providing a common paradigm label for these theories sets them apart from the earlier and more traditional task - person oriented and cognitively oriented leadership theories. Tenth, and finally, we believe that grouping these theories within a common paradigm with a label that is descriptive of their essential commonalties brings coherence to this literature in a meaningful and theoretically parsimonious way. The theories of the neo-charismatic paradigm have been subjected to more than one hundred empirical tests. Collectively, the empirical findings demonstrate with surprising consistency that leaders described as charismatic, transformational, or visionary cause followers to become highly committed to the leaders mission, to make significant personal sacrifices in the interest of the mission, and to perform above and beyond the call of duty. The findings also demonstrate that such leaders have positive effects on their organizations and followers, with effect sizes ranging from .35 to .50 for organizational performance effects, and from .40 to .80 for effects on follower satisfaction and organizational identification and commitment. A recent meta-analysis by Lowe, Kroeck and Sivasubramaniam (1996) of 32 correlations between leader charisma as measured by the Bass (1985) Multifaceted Leadership Questionnaire (MLQ) and independent ratings of leader effectiveness demonstrated a mean corrected correlation of .35. A second meta-analysis by these authors, based on 15 correlations between charisma and subordinates' ratings of their superiors' effectiveness, demonstrated a corrected correlation of .81. Corrected correlations between criterion variables and charisma were higher than corrected correlations between criterion variables and measures of intellectual stimulation, individualized consideration, contingent reward, and management by exception. The effect sizes are usually at the lower ends of these ranges in studies that did not control for 5

environmental effects (Lowe et al., 1996), and at the upper ends of these ranges under conditions of environmental threat, crisis, or uncertainty (House et al., 1991; Waldman, Ramirez & House, 1998). Such findings have been demonstrated at several levels of analysis, including dyads (Howell & Frost, 1989), small informal groups (Howell & Higgins, 1990; Pillai & Meindl, 1991), formal work units (Curphy, 1992; Hater & Bass, 1988), major sub-units of large complex organizations (Howell & Avolio, 1993; Koene, Pennings & Schreuder, 1993), overall performance of complex organizations (Koh, Terborg & Steers, 1991; Roberts, 1985; Trice & Beyer, 1986; Waldman, Ramirez & House, 1998), and U.S. presidential administrations (House et al., 1991; Simonton, 1987). The evidence supporting this genre of theory is also derived from a wide variety of samples, including informal leaders of task groups (Howell & Higgins, 1990), military officers (Bass, 1985), educational administrators (Koh, Terborg & Steers, 1991), supervisors (Hater & Bass, 1988), middle managers (Howell & Avolio, 1993), subjects in laboratory experiments (Howell & Frost, 1989), U.S. presidents (House et al., 1991), chief executive officers of Fortune 500 firms (Waldman, Ramirez & House, 1998), high-level executives of large Canadian firms (Javidan & Carl, 1997), Canadian government agencies (Javidan & Carl, 1997), and CEOs of Egyptian firms (Messallam & House, 1997). The evidence shows that the effects of charismatic leader behaviors are rather widely generalizable in the United States and that they may well generalize across cultures. For instance, studies based on the MLQ charisma scale (Bass & Avolio, 1989) have demonstrated similar findings in India (Pereria, 1987), Singapore (Koh, Terborg & Steers, 1991), The Netherlands (Koene et al., 1993), China, Japan (Bass, 1997), Germany (Geyer & Steyrer, 1994), and Canada (Javidan & Carl, 1997). Finally, a recent cross-cultural study has shown that the leader behaviors of the neo-charismatic leadership paradigm are universally included as 6

prototypical behaviors of highly effective organizational leaders, having ratings consistently above six on a seven-point scale of attributed effectiveness for all 60 countries studied (House et al., 1998). The studies cited above have dealt with leader behaviors and their effects. To date, we know very little about the processes by which leaders produce such results. We need to better understand the underlying motivations and psychological forces that result in the extraordinary effects of charismatic leaders. In this paper, we address this gap by drawing on theories of change and construction and destruction of social meaning. We begin by discussing the differing motivations of followers and leaders, using Lewin's (1951) field theory to integrate them into a single theoretical framework. We present a model that depicts charisma as a social process. We then introduce semiotics as a means for operationalizing and testing the theory's predictions. The empirical results show that charismatic leaders employ consistent communication strategies for effecting social change. THE MOTIVATIONAL UNDERPINNINGS OF CHARISMATIC LEADERSHIP Follower Motives - Frame Alignment Shamir et al. (1993) recently advanced a theoretical explanation of an interpretive process, frame alignment (Snow et al., 1986), by which charismatic leaders motivate followers to embrace social change. Frame alignment (Snow et al., 1986) refers to the linkage of individual and leader interpretive orientations, such that some set of followers' interests, values, and beliefs and the leader's activities, goals, and ideology become congruent and complementary. The term "frame" denotes an interpretive scheme (Boal & Bryson, 1988; Goffman, 1974) that enables individuals to locate, perceive, and label occurrences within their life and the world at large. By rendering events or occurrences meaningful, frames function to organize experience and guide action, whether individual or collective. To achieve frame alignment, charismatic leaders engage the self-concepts of followers in 7

the mission articulated by the leader. Strong engagement of the self-concept of followers makes it cognitively dissonant for them not to behave in ways that further mission accomplishment. Charismatic leaders increase the intrinsic value of follower efforts in pursuit of mission accomplishment by linking effort and goals to valued aspects of the follower's self-concept, thus harnessing the motivational forces of self-expression, self-consistency, self-esteem, and self-worth. Shamir et al. (1993) further argued that charismatic leaders change the salience hierarchy of values and identities within the follower's self-concept, thus increasing the probability that these values and identities will be implicated in action. Finally, they argued that charismatic leaders increase self-efficacy and collective efficacy by positive evaluations, communicating higher performance expectations of followers, showing confidence in followers' ability to meet such expectations, and emphasizing followers' ties to the collective. Shamir et al. (1993) specified communicative techniques that charismatic leaders employ to effect frame alignment and mobilize followers to action. They link present behaviors to past events by citing historical examples (Willner, 1984). They articulate an ideology clearly, often using labels and slogans. They provide a vivid and positive image of the future. Further, they amplify certain values and identities and suggest linkages between expected behaviors, amplified values and identities, and their vision of the future. By articulating an ideological vision and recruiting a number of followers who share the values of the vision, charismatic leaders provide for followers a sense of identity with the collective and a sense of efficacy resulting from membership in the collective. Articulation of high performance expectations, together with display of confidence in followers, results in enhancing both follower self-esteem and self-worth. Since such a shift in values and identities is socially based, followers resulting behavior should represent a shift from the instrumental to the moral, and from a concern with individual gains to a concern with contributions to a collective. We need a theory that explains how and why charismatic leaders engage followers in such transitions, and how and why followers become 8

engaged. Leader Motives - Frame Breaking Sociologists (Eisenstadt, 1968), political scientists (Dow, 1969; Willner, 1984) and organizational behavior theorists (Bass, 1985; Conger & Kanungo, 1987; House, 1977; Nadler & Tushman, 1990) have described or defined charismatic leaders as breaking with traditional institutional authority and persuading followers to embrace innovative or revolutionary ideas. These definitions imply a motivation to change the status quo. Charismatic leaders are thus motivated to alter or break the "frame" or interpretive scheme by which individuals locate, perceive, and label occurrences in their life consistent with the status quo. In contrast to the lack of constraint implied by a frame breaking motive, McClelland and his colleagues proposed a theory of leader constraint and activity inhibition. In a seminal work on leader motivation, they (McClelland et al., 1972) argued that high power motivation, in combination with low affiliative motivation and high activity inhibition, predisposes individuals to be effective leaders through satisfying their need for power by making socially desirable contributions to the larger collective rather than by pursuing self interests. McClelland and his colleagues supported their theory with a series of studies (1972, 1975, 1985) linking men's expression of power motive in the Thematic Apperception Test (TAT) with a presumed measure of activity inhibition - the number of times the word "not" appeared in the stories written by subjects in response to TAT stimulus material. These studies presumed that the word "not" expressed moral constraint as reflected in Christian-Judaic caveats such as "Thou shalt not..." Subjects low in this presumed measure of activity inhibition expressed thoughts about the exercise of power that were focused on personal dominance or winning at someone else's expense. In contrast, subjects who scored high on activity inhibition expressed power imagery more often in terms of doing good for others, for humanity, or for some worthy and presumably moral cause. According to McClelland, individuals who have a high need for power and who also 9

have high activity inhibition should be more effective leaders because they manifest their need for power in socially appropriate ways, while meeting the role demands of positions of influence such as those found in large complex organizations. House et al. (1991) integrated McClelland's leader motive theory with House's (1977) theory of charismatic leadership in a model that includes the most important variables of both theories. House et al. (1991) tested their model using archival data relevant to all elected U.S. presidents from George Washington through Ronald Reagan. They studied the presidents' needs for power, achievement, affiliation, and activity inhibition. They measured these motives by applying the TAT coding scheme to the inaugural addresses of all U.S. presidents. The researchers assumed that inaugural addresses represented presidents' fantasies, hopes, and desires for their terms in office and therefore projections of their non-conscious motives. Three sets of dependent variables measured presidential success: 1) their effectiveness in implementing international, economic, and social/domestic policies; 2) presidential greatness as measured by opinion polls of present day political scientists; and 3) a measure of successful direct actions such as victory in war, great decisions, and near war avoidances such as the Cuban missile crisis. They interpreted their results as consistent with McClelland: Presidential need for power as measured by the use of the word "not" in presidential writings significantly predicted presidential charismatic behavior and effectiveness. However, Spangler and House (1991) noted that presidents who used the word "not" most frequently were highly unconstrained in their behavior. They found that the use of the word "not" was associated with the manner in which presidents exercised power, rather than the ends for which they exercised power. Specifically, presidents who used the word "not" most frequently were more impatient, forceful, radical, demanding and active, and they frequently by-passed the chain of command. Spangler and House (1991) concluded that with respect to presidential speeches and writings, the count of the word "not" was not a measure of activity inhibition as 10

defined by McClelland. Charismatic leaders who frequently invoked the word "not" appeared less disciplined and less psychologically constrained in the way they exercise power than other leaders. The consistently strong ability of "nots" to predict charismatic and effective leadership suggests substantial practical importance in studying the meaning of "nots." However, Spangler and House's (1991) observation that presidents who used the word "not" most frequently were highly unconstrained in their behavior raises questions about McClellands (1975, 1985) interpretation that the frequent use of the word "not" reflects expressions of respect for institutionalized authority, self-discipline, and belief in a just world. Since interest in charismatic leaders stems, in part, from their ability to bring about radical change, how does this fit with McClellands notion of restraint on action and respect for current institutions? Perhaps "not" does not represent unconscious motives. Perhaps, instead, it is a conscious rhetorical device in the repertoire of communicative tools consistently employed by charismatic leaders to bring about innovation and gain acceptance for revolutionary ideas. The use of not may thus reflect charismatic leaders motivations to break current frames through negation. Here again, we need a theory that better explains how and why charismatic leaders engage followers in radical change. Integrating Leader-Follower Motivations Lewin's (1951) field theory provides a useful starting point for integrating theories of the neo-charismatic leadership paradigm with follower motives in generating change. Lewin's theory attempted to describe and explain stability and change in social norms and conduct. He began by defining a social "field," consisting of the collective and its setting. The distribution of social forces within the field determines what happens throughout the field. For example, conflicting social forces act for greater and less discrimination against selected ethnic groups. If the forces for each are equally strong (fgreater + fless = 0), the field maintains a quasi-stationary social 11

state. The equation says nothing about the absolute strength of either of the opposing forces; only that they are equal. Lewin argued that social change can be achieved most effectively if one first decreases the tension between the opposing forces by reducing the strength of both. According to this view, instead of attempting to bring about social change by defining and promoting the objective of the desired change, more effective change efforts begin by reducing tensions. This tension reduction "unfreezes" the average state of collective norms around which opposing forces have stabilized. The next step "moves" the collective norms to a new state. The final step "re-freezes" collective norms in the new state. Lewin's theory suggests that both frame breaking (unfreezing) and frame alignment (moving and re-freezing) are critical processes for bringing about social change. Beyond this, Lewin did not address how unfreezing, moving, and re-freezing of social norms occurs - only that it revolves around individual perceptions of the value of those norms. He suggested that the major cause of resistance to social change lies in individuals' beliefs in the value of existing social norms. To bring about social change, then, one cannot focus exclusively on the level of the collective, nor at the individual level, but rather on the interface between the two: The value that individuals place on the norms of the collective. A SEMIOTIC APPROACH TO STUDY CHARISMATIC EFFECTIVENESS Though Lewin's theory provides a useful framework for integrating the motivational forces underlying charismatic leadership, it offers little guidance about how one might operationalize and test the theory's predictions. Since Lewin's theory addresses socially-constructed meaning, an appropriate tool for empirical investigation is semiotics - the science of signs. Semiotics explicitly deals with the interface between individual and social values, the critical issue in effecting social change. As a formal mode of analysis, it identifies the rules that govern the construction and destruction of meaning in a particular social system (Greimas & 12

Rastier, 1968; Eco, 1979). Like Lewin's (1951) field theory, it rests on the assumption that all meaning is contextualized and resides in a system of underlying oppositions. Semiotic analysis offers a systematic means of linking multiple surface-level expressions of opposition to the system of meaning that underlies them. The Structure of Meaning The units of signification in any communicative act express meaning through their differences (Eco, 1979). The social codes that regulate meaning arise from underlying oppositional structures. Two kinds of opposition interact to give meaning to a sign, or a unit of signification. For a given sign, the first opposition, called contradiction, expresses the total absence of the sign. For example, non-love contradicts love; non-conventional contradicts conventional. A sign can express either end of the contradiction, but not a combination (e.g., one cannot express love and non-love at the same time). For a given sign, the second kind of opposition, called contrariety, includes another sign. The contrary sign, by social construction, also opposes the meaning of the original sign. For example, hate is contrary to love; innovation is contrary to convention. A given sign may have multiple contraries. Hate, loathing, and disgust each opposes love. Innovation, deviance, and spontaneity oppose convention. In contrast to contradictions, a sign may include combinations of contrary meaning or values. For example, one can express love and hate at the same time, though this combination expresses a complex and unstable condition. Based on our theoretical framework, a charismatic leader performs the task of translating innovative ideas into socially conventional ideas, that is, translating a value into its contrary. Semioticians typically employ a visual representation of a value's contradictions and contraries. Figure 1 depicts such a representation for the values of convention and innovation. Convention is contrary to innovation, and non-convention is contrary to non-innovation. Convention contradicts non-convention, and innovation contradicts non-innovation. This framework sets the stage for our 13

subsequent discussion of leader-follower interactions. Figure 1 about here By practice, the dominant positive social value, in this case social convention, occupies the upper left corner of the semiotic square. Conventions represent socialized, institutionalized, and endorsed ideas or values, the frame through which most people experience their world. The contrary, innovation, occupies the upper right corner. The contradictions occupy positions diagonal to the values they oppose. The values depicted in a semiotic square relate hierarchically; the assertion of the dominant value presumes the negation of its contrary (convention presumes non-innovation). However, negation of the contrary only allows the possibility of the dominant value (noninnovation makes convention possible). Understanding the process by which one can effect change in social norms depends on this hierarchy. Replacing one dominant value with another directly (e.g., replacing convention with innovation), or shifting the balance of one in favor of the other, will lead to increasing tension that is likely to undermine change efforts (Lewin, 1951). According to Lewin, an effective change strategy begins by "unfreezing" the dominant value. Here, this means advocating its contradiction (e.g., replacing convention with non-convention before advocating innovation, as indicated by the change trajectory of least resistance in Figure 1 (Greimas & Rastier, 1968)). The terms and structural relations defined by a semiotic square can provide a theoretical starting point for identifying the components of meaning of any set of values within a social system. In this study, they depict values concerning leadership, values believed to both motivate leaders and to serve as standards by which society judges leader effectiveness. The next section examines current conceptions of leadership in relation to the semiotic structure outlined above.

14

The Social Structure of Values Leaders operate and are judged within a social system whose values define what is "effective." Values associated with charismatic leadership imply rejection of the status quo and reliance on non-conventional solutions to existing social problems. These values oppose the socially-endorsed dominant cultural values represented by conventional leadership. Neither the conventional nor the innovative values contain "objective" content. Conventional leadership may mean one thing in the U.S. and something quite different in Cuba. Yet in a given social context, innovation/ non-convention always opposes convention/non-innovation. Within such a meaning system, charismatic leadership, by definition, attempts to persuade society to embrace a contrary of a current social convention (S2 - the dominant negative value in Figure 1), that is, innovation. The logical relations depicted in the semiotic square imply that charismatic leadership is, at its essence, the contrary of conventions. As one adopts more institutionalized conventions, one appears less charismatic. According to this model, negation of existing beliefs is a fundamental characteristic of charismatic leadership. The social structure of leaders' values identifies value components at a societal level. However, it does not reflect the personal values that motivate or predispose charismatic leaders to promote non-endorsed social values. Nor does it describe the personal value changes required when members of society follow the charismatic leader. Therefore, we need to more closely examine the personal value structure of leaders and of society's members. The Personal Sub-Structure of Values Following Greimas and Rastier (1968), two sets of motivators define individual human behavior in relation to social values: desire and fear. Following their premise that desire is a prime human motivator and fear its corollary inhibitor, Figure 2 shows the semiotic structure of personal values. The relations among the values in this square reflect the same structure and hierarchy as those of the social model in Figure 1. 15

Figure 2 about here The terms of the personal value system interact with the social value system to generate what one observes in human behavior. Different value combinations lead to either conflicting or compatible relations (Greimas & Rastier, 1968). For example, desire - that is, aspirations, intentions or a conscious impulse toward something positive - interacts with endorsed social values to produce a compatible or balanced relation of personal and social values. Similarly, fear - that is, the anticipation of danger or the impulse to avoid or overcome a negative - combined with unendorsed social values (innovation), produces compatibility in personal-social values. Both relationships are stable over time, are deemed "effective" and represent socialized members of society. On the other hand, people experiencing fear of society's conventions are unlikely to remain stable over time or to be "effective." Similarly, desire combined with socially unendorsed values (innovation) produces an unstable tension. Charismatic leaders represent this combination of tension desire for innovation. Their leadership aims at replacing current social convention with their personal values in order to achieve a state of balance for themselves. Shamir et al. (1993) described charismatic leaders as having the power to modify the beliefs and preferences of individuals in order to create a new compatibility between personal and social values. Previous theory suggests that charismatic leaders effect such a change through frame re-alignment (Conger & Kanungo, 1987; Weber, 1947). Our analysis suggests that before followers will align around a new interpretive frame, they must reverse their previous relationship with the existing socially dominant values. As described below, this occurs through introducing conflict in a compatible convention-based structure and compatibility in a conflictive innovation-based structure. FRAME BREAKING, MOVING, AND RE-ALIGNING Operationalizing Lewins (1951) three-phase process for changing social values with 16

semiotic analysis leads to two distinct strategies for generating innovation (shown in figure 3). Both scenarios assume a starting point of dominant and compatible personal-social values. In the first case, individual members of society desire current conventions; in the second, they fear innovation. Both cases are compatible within the bounds of current conventions. The phases of unfreezing, moving and re-freezing represent the paths of least resistance in the semiotic squares (e.g. convention to non-convention to innovation) and are described in detail below. Figure 3 about here Frame Breaking (Unfreezing) To effect a change in social values, a charismatic leader must first attempt to reduce the strength of the value individuals place on conventional norms (Lewin, 1951). This is labeled the frame-breaking or unfreezing phase. If the current value is a desire for convention (first scenario in Figure 3), the leader must negate this desire to create a more neutral state, non-desire for convention. A leader can do this by convincing society that conventional thinking is not fruitful, but rather dysfunctional. For example, before President Bush could successfully press for intervention into the Persian Gulf in 1991, he had to convince the congress and the public at large that conventional wisdom, which interfered with his vision, was wrong. He had to persuade them that it would not be another Vietnam, that it would not be another military defeat, that it would not be an embarrassment even in military victory, and that the U.S. was not intervening for the sole reason of maintaining access to low price crude oil. This involved discrediting people's ties to convention. If the current value is fear of innovation (second scenario in Figure 3), the leader must negate this fear to create a more neutral state, non-fear of innovation. The second approach involves convincing society that non-innovation is not viable. For example, President Bush tried to portray the U.S. as committed by treaty, precedence, and moral obligation to not sit by and watch an ally fall to an aggressor. That is, non-intervention equated with non-innovation, which 17

according to the President's arguments, was not a viable option. Frame Moving To build a new stable and compatible value structure, leaders must eventually move personal values from a neutral to a more active state, and social values from opposing to conforming with the desired innovation (Lewin's (1951) second phase). An effective way to initiate the shift is to first negate the endorsed social norms that are contrary to the innovation (e.g., convert convention to non-convention as in Figure 1 above). At the same time, followers' values must move from a passive state (non-desire or non-fear) to an active state (desire or fear). Thus frame moving will include a double negation: non-desire for convention must be transformed into desire for non-convention, and non-fear of innovation into fear of non-innovation. Following our previous example, having discredited conventional thinking about U.S. military intervention, President Bush attempted to shift non-fear of innovation to fear of non-innovation by portraying Saddam Hussein as a neo-Hitler who must be stopped before he reached full power. Negating or inverting both personal and social values minimizes potential resistance since the resulting values remain compatible with the prior ones. The resulting values, however, now encompass the personal motivators (desire and fear) needed to move collective values to a new level (Lewin, 1951). Frame Re-aligning (Re-freezing) Finally, the third phase of the change process involves re-freezing new and compatible values (Lewin, 1951). If successful, the second phase results in personal motivators that a leader can now channel in the desired direction. Through substituting a compatible positive image for the negated social norm, leaders mobilize followers to action (Shamir et al., 1993). The first scenario in Figure 3 entails substituting innovation for non-convention, leading to the final desire for innovation. In the second scenario, it entails substituting positive values for both personal and social negative values (from fear/non-innovation to desire/innovation). In the President Bush 18

example, the administration needed more than discrediting conventional views of U.S. military intervention and creating fear of Saddam Hussein. A new vision (innovation), the "new world order," had to replace the discredited convention. This final change phase will likely meet with minimal resistance as it represents filling a void rather than opposing an entrenched position. The success of this final phase of value transformation critically hinges on a leader's ability to provide for followers a sense of positive identity with the change (Shamir et al., 1993). COMMUNICATION STRATEGIES FOR EFFECTING SOCIAL CHANGE The semiotic modeling of the social value transformation process suggests the need for distinct communication strategies in each of the three phases. To summarize, effective change agents bring about a new set of social and personal value combinations by first reducing the strength of a current value through neutralizing follower ties to the value. They then move the value through a process of negating both the social and personal values. Finally, they solidify the links between their innovative vision and the values of their followers by substituting the negated social value with a positive value. The following briefly describes several communication strategies for effecting these changes and summarizes the discussion with a series of propositions. Negation As described above, a change process requires breaking, neutralizing, negating and substituting. The use of the word "not" is an essential rhetorical device for breaking, neutralizing and negating. Since charismatic leaders, by definition, attempt more innovations and are seen as more successful in those change efforts, they should use the word "not" more frequently than non-charismatic leaders. We include this as our first proposition, though it has already received significant empirical support (House et al., 1991). Here, the proposition reflects a process rather than a personality trait, and is fundamental to our overall model of value transformation. Proposition 1: Charismatic leaders will use the word "not" more often than non19

charismatic leaders. During the initial phase of a change process, charismatic leaders must negate the followers' personal values towards convention or innovation. During the second phase, they must move the neutralized values toward a less neutral position by negating both the original convention and the neutralized personal links to that convention. This phase thus calls for a double negation. During the final phase, charismatic leaders must substitute a new and positive social norm for that which they have negated. During this phase, they no longer rely on negation as the critical means of effecting change. Thus, they will use fewer "nots" during this final phase of transformation. From the above, it follows that the use of the word "not" through the change phases is curvilinear. One would expect "nots" to be used frequently in the frame-breaking phase, more frequently in the form of double negation in the frame-moving phase, and less frequently in the final frame re-freezing phase. Because non-charismatic leaders are less likely to attempt radical change, we would expect them to not follow this pattern. Proposition 2A: During a transformation, charismatic leaders will use the word "not" frequently during the initial phase, more frequently during the middle phase, and infrequently during later phases. Proposition 2B: The use of nots by non-charismatic leaders will not follow the curvilinear pattern of charismatic leaders. Inclusion and Consensus Building While negation is a critical rhetorical device for the breaking down or unfreezing that occurs in the early phases of social change, endorsement and affirmation are important in later phases. Given the instability induced in earlier phases, change agents must generate relations of trust with followers in which parties believe that "things will work out" (Gambetta, 1988; Gartner 20

& Low, 1990). Again, specific communication strategies help generate and sustain trusting relationships. Charismatic leaders will include non-believers within the innovative frame (Goffman, 1974) they wish to generate by managing the boundaries around the subjects of their discourse. They can do this in two ways. First, they employ inclusive rather than exclusive referents (Fiol, 1989). One would expect a charismatic leader to use more associative referent terms such as "we," "us," "our group," or "our organization" rather than terms that imply disassociation or noninclusion such as "I," "you," or "me." Proposition 3: Charismatic leaders will use more inclusive language than non-charismatic leaders. Second, effective change agents enlarge the boundaries of their discourse by employing high levels of abstraction. Eisenberg (1984) has argued that the ambiguity associated with values at a high level of abstraction allows consensus building around those values without necessarily achieving consensus around their meaning. To effectively engage their followers in a movement toward innovation, charismatic leaders will likely employ high levels of abstraction in their discourse during the frame-moving phase. Following the example of the Persian Gulf intervention of 1991, President Bush attempted to rally society around innovation by calling for the U.S. to create a "new world order." This phrase has many meanings. Its ambiguity aided Bush in building a consensus around his views. Proposition 4: Charismatic leaders will communicate at higher levels of abstraction than non-charismatic leaders. Inclusive language and abstract representation are rhetorical techniques that serve a similar purpose: to include and engage followers in a change process that defies conventions. Inclusion explicitly invites followers to engage and embrace the leaders values, while higher levels of abstraction open the space for followers to align their personal values with those of the 21

leader. Since engagement and inclusion are necessary for maintaining or changing social values (regardless of the content of the values) one would expect both charismatic leaders and noncharismatic leaders to combine the two techniques for maximum effectiveness. Proposition 5: All effective leaders will use more inclusive language with higher levels of abstraction. However, like negation, inclusion and abstraction carry out more specific roles in each phase of a transformation process. According to our model, the early unfreezing phase is a period of breaking personal ties to convention. The focus is on individuals, rather than society. One would thus expect relatively low levels of abstraction and less use of inclusion in the first phase. In contrast, one would expect greater use of abstract and inclusive language during the frame-moving phase, when charismatic leaders must actively engage their followers in a process of visualizing a change at the level of the collective. In the final phase, one would again expect inclusion and abstraction to be less critical since the move toward social-level change has already taken place. As in Proposition 2, we would not expect to see this pattern for noncharismatic leaders. Proposition 6A: During a transformation, charismatic leaders will use higher levels of inclusion and abstraction during the middle phase than in earlier or later phases. Proposition 6B: The use of inclusion and abstraction by non-charismatic leaders will not follow the curvilinear pattern of charismatic leaders. In sum, our model suggests that charismatic leaders employ specific communication strategies to move the change process through the three phases of frame breaking, moving, and re-alignment. Table 1 summarizes the communication patterns that characterize each of the three phases. Table 1 about here

22

METHODOLOGY Sample We applied semiotic analysis to forty-two speeches from all 20th century United States presidents through Ronald Reagan (a sample of fourteen presidents)1. A prior study of "nots" (House et al., 1988) provided some of the speeches; others were drawn from various archives of presidential speeches. We selected speeches that addressed a wide, national audience either in topic matter or in physical audience. Most of the speeches were inaugural addresses or addresses to congress. In some cases those were not available, and we chose substitutes from a set that was available to reflect subject matter similar in scope and audience to inaugural or congressional addresses. Appendix A lists the presidents and the speeches. Though one might worry that professional speechwriters create most presidential addresses (at least in more contemporary speeches), substantial evidence demonstrates that presidents greatly influence the language and motive imagery in their speeches. Winter and Stewart (1977) demonstrated the construct validity of inaugural motive imagery for 20th century presidents. House et al. (1991) found that motive scores derived from the motive imagery in the inaugural addresses of all elected presidents predicted both presidential leader style and presidential effectiveness with respect to the implementation of their economic, international, and social domestic policies. For each president, we chose a speech from his first year in office, a middle year in office, and his last year in office. The sequence of these speeches approximated three general

1 Why only 20th century presidents?

Presidential researchers contend that the 20th century presidency varied in three important ways from the pre-20th century presidency. First, the style of language changed in the 20th century. Pre-20th century presidents used more flowery language, as well as more subordinated and conditional language (differences based on Flesch index significant at the .0001 level). Second, the U.S. ended a period of isolationism, joining in international affairs. Finally, mass media changed the speed, means, and reach of presidential communications, allowing them to reach national audiences rapidly and simultaneously. Why only through Ronald Reagan? We wished to replicate the 20th century sample of presidents used in the House et al. (1991) and the Spangler and House (1991) studies.

23

phases of social transformation. While value transformations may involve a specific issue, e.g., income tax, one can argue that presidential leaders have broad agendas that take many years to accomplish. We may especially view the charismatic presidents as attending to a general change in national direction or general societal values (e.g., F.D. Roosevelt and The New Deal, R.R. Reagan and The New Dawn in America). Coding Communication Acts The unit of analysis in each speech was a logical sentence. Logical sentences were defined as complete grammatical phrases (i.e., "subject...verb...object") or sub-phrases which were separated by a hyphen from the remainder of the sentence. Compound sentences were treated as two (or more) logical sentences. Each speech provided 11 to 18 sentences according to the following rules: (1) Code a minimum of 10 sentences, (2) complete coding to the end of the paragraph containing the 10th sentence. This procedure yielded 645 coded sentences. Samples of speeches and their coding are provided in Appendix B. Each sentence provided coding for three independent variables relating to the propositions above. First, as mentioned above, coders recorded the temporal sequence of a speech (1,2,3) and treated these as phases of transformation (unfreezing, moving, and refreezing). Second, coders identified sentences as inclusive/non-inclusive (I/NI) according to the use of inclusive versus non-inclusive pronouns. Presidents set an agenda for the nation whether they address the nation directly or through groups like congress. Examples of inclusive language include uses of "we," "our," and "us," when applied to a large collective, usually the nation, but occasionally the world if the nation was implied as part of it. If pronouns referred to a small, specifically identified group (e.g., "Tom and I...We..."), then coders treated the pronoun as non-inclusive. Coders treated all uses of proper nouns, non-inclusive pronouns, and passive constructions as non-inclusive. If they found two different types of usage in a sentence (e.g., "us 24

against them"), they coded the sentence as inclusive if the "us" referred to the country. Finally, the domain level of a discourse indicates its level of abstraction. Coders assigned each sentence one of the following domain levels (from least to most abstract): (1) individuals (e.g., I, F.D.R., George), (2) particular things or events (e.g., the income tax, this battle, the meeting), (3) the country or nation, including the people of the nation, and (4) the world, foreign countries, and universal beliefs (e.g., mankind, truth, justice). Thus, the greater abstractions at the higher domain levels allow a larger collective to locate itself in the presidents message. Coding Charisma Charisma was treated as a binary variable, with charismatic presidents coded as 1 and non-charismatic presidents coded as 0. We followed the classification procedures used by House et al. (1988). In their study, eight reputable political historians identified American presidents as charismatic, non-charismatic, neither charismatic nor non-charismatic, or uncertain, using the following guidelines: Charisma is the ability to exercise diffuse and intensive influence over the normative or ideological orientations of others (Etzioni, 1961). As a result we can identify charismatic leaders by their effects on their followers such that followers of charismatic leaders: a) have a high degree of loyalty, commitment, and devotion to the leader; b) identify with the leader and the mission of the leader; c) emulate his values, goals, and behavior; d) see the leader as a source of inspiration; e) derive a sense of high self-esteem from their relationship with the leader and his mission; and f) have an exceptionally high degree of trust in the leader and the correctness of his beliefs. The historians classified the leaders according to their relationship with their cabinet 25

members rather than to the public since that study was interested in organizational leadership rather than mass or political leadership. Their classification procedure resulted in two groups charismatic and non-charismatic. In the current study, four of the fourteen presidents qualified as charismatic, and ten as non-charismatic. Appendix A lists the charismatic and non-charismatic presidents in our sample. Reliability Inter-rater agreement among the political historians was .88. House et al. (1991) confirmed the validity of the political historians' classification of presidents. They demonstrated that presidents classified as charismatic had stronger and more positive affective ties with their cabinet members than did non-charismatic presidents. House et al. also confirmed the classifications by use of data collected from presidential biographical writings collected in an independent study by Simonton (1987). Thus the ratings were triangulated with three independent sources. Doctoral candidates in management coded the speeches used in the current study. Coder training involved a review of coding rules, detailed analyses of a pre-coded speech selected for its potentially ambiguous phrasing, and trial coding against a key of three previously coded speeches. Agreement with the key was 95-100 percent indicating adequate reliability. Further, an inter-coder reliability test was conducted based on coding by three coders of onefourth of the total set of speeches. Inter-rater reliability was over 90 percent for all coded dimensions in the sub-sample of texts. ANALYSIS AND RESULTS The propositions fall into two groups - simple comparisons between groups and comparisons over time. Propositions 3 and 4 suggest simple comparisons of means. Using a least squares means test to adjust for the unbalanced sample sizes (Searle, Speed & Milliken, 1980), Table 2 shows support for Propositions 3 and 4. The charismatic leaders used more 26

inclusive language and higher levels of abstraction than did the non-charismatic leaders. Table 2 about here Proposition 5 also involves comparing group means. The Table 3 analysis of variance shows that for all leaders the level of inclusion increased as the level of abstraction increased. The means tests between levels in Table 3 shows that the two upper levels of abstraction varied significantly in their amount of inclusion from the two lower levels, but within the two upper levels and within the two lower levels the average amount of inclusion did not vary significantly. This suggests that the distinction between individuals (level 1) and non-personal, specific topics (level 2) may not be important, at least in terms of the boundaries of discourse. Nor, apparently, did the leaders create different boundaries around country, world, or universal values (levels 3 and 4). However, they did, as proposed, create different boundaries around specific topics (levels 1 and 2) and broad, abstract topics (levels 3 and 4). Table 3 about here The remaining propositions address timing issues. Here, graphical views of the data clearly show the trends suggested in Propositions 2 and 6; also, the graphs help illuminate the statistical analyses. Figures 4 and 5 show that the charismatic leaders followed the pattern of moderate use of "nots" in phase 1, higher use of "nots" in phase 2, and low use of "nots" in phase 3. These figures also show that the pattern for the non-charismatic presidents was quite different, with the first phase having higher use of "nots" than the subsequent phases. The statistics in Table 4 indicate that for all leaders, as a group, the third phase differed significantly from the first two phases. Additionally, the leader-year means tests show that this third-year difference arises because of the charismatic leaders. The difference between Phase 1 and Phase 2 does not achieve statistical significance; in this regard Proposition 2 is not fully supported. However, our general argument holds: Charismatic leaders used "nots" during the unfreezing and moving phases at a significantly higher level than during the re-freezing phase, and their pattern of 27

usage was different from that of non-charismatic leaders. Figures 4 and 5 about here -----------------------Table 4 about here Figures 6 and 7 show the pattern of leaders' use of inclusive language in relation to Proposition 6A and 6B. All leaders used inclusive language more during the middle, moving phase, than the other two phases. The statistics in Table 5 support the impression given by the figures. So the data strongly support Proposition 6A, that charismatic leaders used more inclusive language in their middle phases than in other phases. The data do not support Proposition 6B. However, charismatic leaders in the sample did exhibit a more pronounced pattern in their use of inclusive language than their non-charismatic counterparts. Figures 6 and 7 about here ----------------------Table 5 about here

Figures 8 and 9 show patterns of usage for high levels of abstraction (combined levels 3 and 4)2. Here, the pattern for charismatic leaders is distinctly different than for non-charismatic leaders as predicted in Proposition 6B. Table 6 shows statistical support for the level of abstraction arguments in Propositions 6A and 6B. The analysis of variance shows significant relationships between abstraction, phase, charisma and the interaction between phase and charisma. The combined means tests show higher use of abstractions in the middle phase for all leaders. The separated means tests show charismatic leaders followed the proposed pattern, with significantly higher levels of abstraction than non-charismatic leaders. Figures 8 and 9 about here ----------------------Table 6 about here Overall, the propositions receive consistent and strong support from the data. This
2 The lower levels of abstraction (1 and 2) create an inverse pattern of the higher levels.

28

sample suggests that leaders, in general, follow discernible, purposeful patterns of discourse. While attempting social change, charismatic leaders appear to apply different and more pronounced patterns of discourse than non-charismatic leaders. These differences in rhetorical technique provide insights into how the charismatic process works, and support our arguments about why followers attach themselves to charismatic leaders and their causes. DISCUSSION AND CONCLUSIONS To date, the qualities and motives that define charismatic leadership have been as elusive as those defining entrepreneurship. For decades, researchers have tried to identify the personality and characteristics that predispose individuals to be effective agents of social change. The emphasis has been on leader motives and personality characteristics as well as leader behaviors and their effects. The results have not disclosed the psychological processes that explain why leaders of the neo-charismatic paradigm have such extraordinary effects on followers and organizations. This study has attempted to redirect attention to such psychological processes. The theoretical framework emphasizes frame breaking for leaders and frame alignment for followers. It combines these into a process model of how the leader/follower interactions can result in social change. The empirical results of the study suggest that charismatic leaders employ a predictable, consistent set of linguistic techniques to break down, move, and re-align certain beliefs of their followers. Specifically, the presidents of our sample employed techniques of negation, inclusion, and abstraction more frequently during the middle phase of their tenure as leaders than in the earlier and later phases. To explain why these techniques are effective and how they operate, it is necessary to discuss them within the larger context of social interaction. It is not possible to separate the role of language from its social context. The power of language resides in its potential to both reflect and shape social norms and attitudes. To discuss the empirical results of this study, then, we must locate them within the broader context of leaders strategic 29

communications and follower responses to such communications. Lewins (1951) field theory provided a framework for describing the phases of social change. Though the theory suggests the general need for strategies of negation for unfreezing, and strategies of affirmation for re-freezing, it offers little guidance about how to operationalize and test the theorys predictions. A semiotic framing of the interactions of personal and social values allowed us to systematically trace the dynamics of such a change process. By identifying both personal and social components of a change process, semiotics provided a basis for describing and explaining the interactions among the negating and affirming aspects of a charismatic leaders change strategy. Moreover, by highlighting the changing interactions of personal positive and negative motivators and social values across the phases of social transformation, a semiotic perspective suggested specific change strategies appropriate for different phases of the process. For example, during the initial frame-breaking phase, a semiotic perspective argues that the aim must be to neutralize individual ties of desire (fear) to a (non-)convention, rather than to break down the convention itself. Following this perspective, one would expect change strategies to include negation (frequent nots) that focuses on particulars rather than universals (low levels of abstraction) and that does not emphasize the inclusion of individuals in the collective (low levels of inclusion). The relative infrequency of abstract and inclusive language, combined with a moderate degree of negation in the first years speeches of our charismatic presidents (see Figures 4-9), may thus reflect interrelated components of a more general strategy aimed at loosening individual ties to a collective norm within the bounds of compatibility (see Table 1). Similarly, the reduced level of negation, abstraction, and inclusion in the language of charismatic leaders in their final year of presidency, again may reflect a coherent approach to the requirements of refreezing attitudes and norms: channeling personal motivators (already developed) in the desired direction. 30

According to our model and empirical results, the middle phase - frame moving represents the most challenging and critical period of a social change process. Our empirical results show that negation, inclusion, and abstraction all peaked during this period. The results are consistent with semiotic theory, which calls for a negation or inversion of both personal (nondesire to desire) and social (convention to non-convention) values. The theory and data suggest that movement towards a new compatible link between personal motivators and social norms requires the simultaneous construction and destruction of what people know and believe. It calls for a high level of negation combined with equally high levels of affirming forms of identity and consensus building. The interdependence of negative and positive aspects of change strategies follows Gambettas (1988) theory of trust building: The greater the break from tradition, the greater the need for trust-building activities. LIMITATIONS AND IMPLICATIONS FOR FUTURE RESEARCH This study opens the door to a research agenda that to date has remained largely unexplored. If charismatic leadership is about social transformation, whether transformation of a nations or a corporations agenda, then it is time that we concentrate on the processes underlying the effects of leader behaviors. The word charisma derives from the Greek word for gift. We have for too long focused on trying to identify and define the gifted, without recognizing that the gift of change agents resides in their social interactions over time. A redirection of research efforts toward a focus on the social processes underlying charismatic leadership will require that we include characterizations of followers, as well as leaders, into our models. The present study begins by testing a model of leader communication strategies employed in effective change efforts. The communication model behaves consistently with our semiotic modeling of leader communication strategies and follower frame realignment. An important next step is to incorporate empirical data concerning if, how, and when follower frames actually shift during a change process. 31

This points to limitations of our empirical testing that need to be addressed in future research. Following the lead of prior studies of charisma using U.S. presidents (House et al., 1988, House et al. 1991; Winter & Stewart, 1977), we have assumed presidents have general agendas (innovation or stability) that they convey through their public speeches. Testing our model in a wider variety of settings, in particular in settings that might allow for more specific identification and tracking of values being addressed, will add to our understanding of the charismatic process. Also, other settings would allow for larger sample sizes, greater variation in charisma (one might argue that to some extent all U.S. presidents have been more charismatic than the average leader) and a shift from dichotomous to continuous measures of charisma. Focusing on charisma as a social process will also require that we develop additional dimensions of leader/follower interactions. In this study, we focused on the role of three communication techniques - negation, abstraction, and inclusion - because they correspond closely to the change strategies of negation and consensus building suggested by our model. Other communication techniques (e.g., feedback processes, active versus passive forms of communication, use of communication media) may be critical as well. Moreover, the research agenda needs to be broadened to include nonverbal forms of leader/follower interchange. Our model offers an alternative explanation for the well-documented phenomenon of charismatic leaders frequent use of the word not. In its primary grammatical role, not negates a word or a group of words. The results of this study suggest that negation may serve the rhetorical functions of unfreezing and moving attitudes and values, rather than indicating personal restraint (McClelland, 1975, 1985). An important contribution of this study, however, lies in its portrayal of nots as only one component of a systematic strategy for generating social change. It suggests that a comprehensive view of the processes that define charisma is more enlightening than is research based on single components of the process (such as the use of nots) in isolation. It also suggests that future studies should expand beyond the much studied 32

not and look at other forms of negation (neither/nor, no, none, however, but,). Finally, the theoretical framework and results of this study provide a forceful argument that charismatic leadership is a dynamic process that is impossible to capture in a single snapshot. The effectiveness of change strategies at one point in a leaders tenure depends importantly on preceding leader/follower interactions. Further progress in our understanding of this important phenomenon will be substantially enhanced if we recognize the interactive elements of the charismatic process over time.

33