Michael Burgoon, Vickie Pauls, Denning Laura Roberts. The Persuasion Handbook: Developments in Theory and Practice. Editor: James Price Dillard & Michael Pfau. Sage Publications. 2002.
It appears that at least once every decade or so, there appears a renewed interest, as evidenced by public discussion and publication activity, in two recurring issues in the discipline of communication. The first focuses on the general health, or even the viability, of the study of Persuasion. Nearly a quarter of a century ago, Miller and M. Burgoon (1978) asked the question of whether or not a case could be made for Persuasion research. They claimed that while it would be hyperbolic to state that the guns are silent on the Persuasive battleground, the roar of these guns has been sporadic and muted. They further added that traditional Persuasion research had been swimming against the ideological and scholarly currents of the past decade. They were writing at a time when Persuasion research, not only in the discipline of communication but also in allied social science disciplines, no longer was front and center stage but rather had been displaced by other concerns such as interpersonal communication and social cognition.
A dozen years later, M. Burgoon and Miller (1990) followed up on their earlier essay commenting on whether the former article’s suggestions, injunctions, and conclusions about the directions of Persuasion research had much, if any, influence on the extant state of research in the area of social influence. They speculated that there would be a “rebirth” of interest in Persuasion research during the final decade of the final century of the then current millennium. Specifically, Burgoon and Miller predicted more concern for macrosocial concerns with an interest in the role of massive social change. They also suggested a greater emphasis on natural setting research that would allow researchers to scrutinize cultural and sociological patterns of communication so that Persuasion researchers could participate in socially important, tangible goals such as reducing the incidence of cancer by 50% by the year 2000.
Perhaps most interesting—and not all of the current chapter authors are unbiased observers of their musings—was the following commentary by M. Burgoon and Miller (1990):
If the course of communication inquiry follows the course of normal science (and we suspect it will), then attention to macro-level and micro-level issues in natural settings will trigger a pendulum-like action. A concern for greater ecological validity (our reading of the present intellectual atmosphere) is almost inevitably followed by demands for more control and precision in explanation and prediction. Thus, it would not be surprising if a future call for vigilant experimental control rejuvenates present-day “mechanistic empiricist” defectors and recruits new hands to laboratory research efforts reminiscent of the work being done several decades ago. (p. 158)
Most of the observations turned out to be relatively accurate harbingers of what the decade would bring in terms of the study of Persuasion. Moreover, that the study of social influence is relatively healthy and concerned with a number of areas of social import is not at issue, as it has been in the past, and is attested to by the existence of the current handbook, replete with examples of lively theoretical debate and interesting empirical data.
A second recurring issue is the claim that the discipline of communication is bereft of rigorous formalized theories. Berger (1991) relegated theories to the status of “curios” in one such reprise of the issue. M. Burgoon (1994) joined the chorus by claiming that there was some consensus that the discipline of communication had been somewhat bradytelic in developing rigorous content-specific theories. He further added that theoretical precision was going to be difficult to obtain. Such lamentations about the state of theory development across the full spectrum of communication studies might be warranted. It might be less warranted for Burgoon to cast the critical net so widely as to encompass the study of social influence processes. Again, the current handbook highlights the extent to which theoretical development, debate, and Perhaps debunking are part of the universe of discourse of Persuasion scholars.
While the question of “whether” theory development might not be as relevant in the social influence area as in the rest of the discipline, there are certainly issues about “which” of the many extant theories merit attention. The composition of this handbook probably speaks as well as anything I might write in this prolegomenon to the somewhat ephemeral nature of our Peripatetic fascination with certain theories and, Perhaps, premature dismissal of others. If theories were to be judged scientifically useful by referendum, then the dual-processing models (Chaiken, 1980, 1987; Petty & Cacioppo, 1981, 1986) would win in a walk. However, there is little reason to accept current popularity as any kind of indicator of scientific value for any theoretical position. What is not currently in vogue in communication, at least as evidenced by this handbook, is much of a recognition of the lively interest (evident in the allied social science disciplines) in the reconstruction of the theory of cognitive dissonance (Festinger, 1957) or the refinement and advancement of the theory of reasoned action (Fishbein & Ajzen, 1975) in applied areas such as public health and mass media effects. One could also argue that Persuasion scholars, in general, tend to be less attuned to theoretical formulations emanating from the discipline of communication than is warranted. However, the argument that is advanced in this chapter is that there is at least one theory that merits serious consideration and research attention.
Language expectancy theory (LET) has been refined, refurbished, and remodeled over the past quarter of a century. First, it advances a relatively formalized propositional framework focusing directly on how message features positively or negatively violate (or conform to) macro-level expectations about what constitutes appropriate suasory communication attempts. Those propositions have been refined and extended at various times throughout the development of LET.
Second, the boundary conditions not only are specified but also are, in reality, quite broad compared to extant theories of social influence. This particular theoretical formulation includes propositions about (a) the traditional passive message reception situation, in which a Persuader presents a message to a target or target audience with a desire to change attitudes and/or behaviors; (b) the active participation paradigm, in which individuals are “self-Persuaded” by actually producing messages, usually at odds with their own privately held attitudes, resulting in changing their private attitudes to more closely conform to their public communication behavior; and (c) how language and expectancy violations operate in tandem in the resistance to Persuasion paradigm. Obviously, attempting to cross three distinct Persuasive situations, each with its own voluminous wealth of research findings and collection of sometimes confounded and contradictory results, is an ambitious undertaking.
Third, LET has been able to advance a number of hypotheses in a series of research efforts to provide empirical evidence of the explanatory and predictive power of the propositional formulation. In study after study, there has been a confirmation of hypotheses that are not easily interpretable without reference to the theory from which they were derived.
While LET appears to have a considerable amount to add to the study of Persuasion, there are future actions that must be taken to provide additional validation of the explanatory calculus in natural settings in which socially important issues are the focus of suasory attempts. Also, even though one can claim that the current boundary conditions are quite broad, attempts to directly test the Persistence of expectancy violations effects in more interpersonal or face-to-face attempts is overdue. Questions about how expectancies are developed and exactly what expectations are held about specific Persuasion tactics must also be the focus of future research efforts. In addition, questions that plague most experimental research in Persuasion remain a problem for this particular theory. For example, research needs to determine the Persistence of the effects of expectancy over time. Sequential messages, not one message delivered in experimental isolation, need to be examined to determine how expectancies change and how violations either enhance or inhibit Persuasion in ongoing Persuasion attempts. These and other issues are addressed after presenting an overview of LET.
An Overview of Language Expectancy Theory
Origins and Assumptions of LET
Brooks (1970) published research that sparked interest in the role that expectations about what a source might or might not say in Persuasive discourse plays in attitude and/or behavior change. Brooks was interested in “reversals of previously held attitudes” and specifically stereotypes. He concluded his brief research report with an insightful comment that provided the spark to begin to formally develop LET:
The possibility of contrast effects should be considered. This principle assumes that we carry stereotypes into such social situations as the public speech. There, the speaker’s behavior may be discrepant with stereotyped expectations. If the discrepant stimuli cannot be assimilated or ignored, they are likely to be exaggerated in a listener’s Perceptions…. One explanation … is this: unfavorable (or favorable) speakers may be Perceived more (or less) favorably not because their behavior is intrinsically Persuasive (or dissuasive) but because it contrasts with stereotyped expectations which audiences hold. (p. 155)
These comments prompted questions about the nature of what Brooks called stereotypes and about what determines what would later be called expectations. First, to what degree could the individual, as a unit of analysis, be jettisoned in favor of a more aggregate look at shared expectations of groups and even societies? Second, would it be fruitful to pursue research to determine whether there are indeed cultural and sociological forces that shape patterns of ordinary language and determine normative and non-normative usage? There was ample evidence to assume that as communicators mature, they learn not only the mechanics of language but also what to say and when to say it. Finally, the question of whether such normative expectations were limited to notorious (or popular) public figures, as used in previous research, or were applied to all communicators of a given type, group, class, or even all societal members was intriguing.
Prior research (McPeek & Edwards, 1975) bolstered the contention that receivers do have shared expectations about the behaviors a communicator should exhibit. When these expectations are violated, receivers overreact to the behaviors actually exhibited. If a communicator is initially Perceived negatively and then demonstrates more positive behaviors than anticipated, receivers overestimate the positiveness of the unanticipated behaviors. The reverse also holds; when an initially positively valenced communicator exhibits unexpectedly negative communication behaviors, receivers tend to exaggerate their negative evaluation of the communicator and/or the message (Brooks, 1970; McPeek & Edwards, 1975).
One of the first tests of what was to become LET began with a fascination and curiosity with the impact of different linguistic strategies on securing Persuasive outcomes (M. Burgoon, Jones, & Stewart, 1975). This three-part seminal article provided evidence that strategic linguistic choices can be significant predictors of Persuasive success. Following the success of this novel approach to influence inquiry, programmatic research began examining passive reactions to Persuasive messages and evolved into inquiry into active participation (or self-Persuasion) (M. Burgoon & Miller, 1971) and resistance to Persuasion (M. Burgoon, Cohen, Miller, & Montgomery, 1978; Miller & M. Burgoon, 1979), placing the three paradigms under the expectancy umbrella.
Language expectancy theory assumes that language is a rule-governed system and that people develop macro-sociological expectations and preferences concerning the language or message strategies employed by others in Persuasive attempts. These expectations are primarily a function of cultural and sociological norms. Preferences, according to this sociological Perspective, are usually a function of cultural values and societal standards or ideals for what is competent communication Performance.
Burgoon and Miller (1985) published a detailed propositional logic outlining the formative explanatory calculus of LET, which is explicated in some detail later in this chapter. M. Burgoon (1989, 1990) later presented amajor refinement of the model and discussed the effects of both positive and negative violations of expectations in Persuasive attempts. In summary form, LET posits that changes in the direction desired by an actor occur when positive violations of expectations occur. Positive violations occur (a) when the enacted behavior is better or more preferred than that which was expected in the situation and (b) when negatively evaluated sources conform more closely than expected to cultural values, societal norms, or situational exigencies. Change occurs in the first case because enacted behavior is outside the normative bandwidth in a positive direction, and such behavior prompts attitude and/or behavioral changes. In the second condition, a Person who is expected to behave incompetently or inappropriately conforms to cultural norms and/or expected social roles, resulting in an overly positive evaluation of the source and subsequently change advocated by that actor. Negative violations of expectations result from language choices or the selection of message strategies that lie outside the bandwidth of socially acceptable behavior in a negative direction. Negative violations result in no attitude and/or behavioral changes or in changes in the opposite direction intended by the actor.
Burgoon (1995) provided a detailed etiology of the formulation of LET in a festschrift for his collaborator, the late Gerry Miller. The review provides a bit of sociology of knowledge as it describes how this theoretical formulation has taken a few steps forward and some steps backward over the years in attempting to develop a sound basis for understanding how expectations and a host of message variables interact to enhance or inhibit Persuasion effects. That historical Perspective is available elsewhere and might be of interest to researchers experiencing difficulty in developing and refining a theory over the course of a programmatic research effort.
The Propositional Framework of LET
Passive Message Reception Paradigm. Language expectancy theory is an axiomatic theory that, as discussed previously, expounds on the effects of linguistic variations on message Persuasiveness. It is a message-centered theory of Persuasion (M. Burgoon, 1995) that explains why certain linguistic formats in Persuasive messages influence Persuasive outcomes. The theory makes assumptions about human nature that, in turn, explain the effects of using unconventional language styles on message Persuasiveness.
Certain basic assumptions have guided the development of LET and have influenced all of the research completed under the umbrella of this theoretical formulation. Language is viewed as a rule-governed system (M. Burgoon & Miller, 1985) in which people develop norms and expectations concerning appropriate language use in given situations. Cultural and sociological forces shape our patterns of ordinary language and determine normative and non-normative use. Based on these assumptions, the following propositions have been developed and refined (M. Burgoon, 1989, 1990, 1995; M. Burgoon & Miller, 1985):
Proposition 1: People develop cultural and sociological expectations about language behaviors that subsequently affect their acceptance or rejection of Persuasive messages.
As corollaries, LET advances two propositions delineating the impact on Persuasive outcomes that result from conformance or nonconformance to expectations:
Proposition 2: Use of language that negatively violates societal expectations about appropriate Persuasive communication behavior inhibits Persuasive behavior and results either in no attitude change or in changes in position opposite to that advocated by the communicator.
Proposition 3: Use of language that positively violates societal expectations about appropriate Persuasive communication behavior facilitates Persuasive effectiveness.
Original research projects examined the effects of varying language intensity in some manner in Persuasive message (M. Burgoon, 1970a, 1970b). In a series of empirical studies investigating the effects of the use of highly intense language, M. Burgoon et al. (1975) provided experimental tests of the combined effects of source characteristics and language intensity. They identified types of individuals (e.g., females, low-credible communicators) who were presumed to be expected to use less aggressive language choices (low-intensity language) in their Persuasive messages. When these individuals used more instrumental verbal aggression (a term coined much later), they were seen as negatively violating expectations, and thus attitude change was inhibited. Furthermore, the studies revealed that males and highly credible sources could use either aggressive or unaggressive verbal strategies and be Persuasive. However, it seemed that more aggressive behavior was the expected and/or preferred mode of argument only for highly credible male advocates. The results of these empirical tests of expectancy-based predictions were accompanied by a discussion of what was then called a message-centered theory of Persuasion.
What the discussion of the empirical studies of language intensity and source characteristics actually provided was Perhaps a skeletal formulation of what would later be developed as LET. Arguments that were advanced in the interpretation of this experimental research were markedly different from prevailing theories of social influence of the day. First, the focus was distinctly macrosocial in orientation. It was argued that entire social categories (e.g., females, members of different ethnic groups) were bound by relatively rigid normative expectations of what was “appropriate” or expected communication behaviors. Such expectations were not unique to specific communicators but were unique to aggregates of like individuals in this society. Second, the previously discussed concept of normative bandwidths varying in size for different social groups for expected language behaviors was empirically demonstrated. People of high credibility and male speakers in general appeared to have linguistic freedom (wide bandwidths) and could select from a number of Persuasive strategies (low- or high-intense language) without violating preset expectations. On the other hand, large numbers of the population had constricted bandwidths of expected communication behaviors and concomitantly very constrained choices in how they could argue if they wished to be successful at Persuasion. Finally, this elementary theoretical formulation provided a plausible explanation for “boomerang” effects—change opposite to the position advocated by the communicator— that had proven enigmatic to Persuasion researchers at the time. People negatively violating expectations by using intense language in a Perceived extremely inappropriate manner produced such a contrast effect that people moved to opposing attitudinal positions to distance themselves from advocacy of such communicators.
Later, attention shifted from a focus solely on one message variable, language intensity, to a host of message variables (discussed in more detail later in this chapter) that were of interest to Persuasion researchers. Detailed reviews of research on fear appeals, opinionatedness, and language intensity are readily available in published social influence literature (e.g., M. Burgoon & Miller, 1985). M. Burgoon (1989) extended that review to include a number of micro-level message variables as special cases of what has been called a type of instrumental verbal aggression (which includes, but is not limited to, fear appeals, opinionated language, language intensity, and aggressive compliance-gaining strategies) and placed them under the theoretical umbrella of LET. M. Burgoon (1990) also explained the results of more macro-level Persuasion strategies (e.g., sequential message strategies such as foot-in-the-door [Dillard, Hunter, & M. Burgoon, 1984] and door-in-the-face [Cann, Sherman, & Elkes, 1975] techniques), as well as the compliance-gaining message strategy research (e.g., Marwell & Schmitt, 1967a, 1967b; Miller, Boster, Roloff, & Seibold, 1977, 1987), from an expectancy theory Perspective.
Language expectancy theory, as suggested previously, has been used as a theoretical framework to explain the effects of several source, message, and receiver variables on message Persuasiveness. Each of those research foci merits some detailed attention in that different propositions are advanced from these three somewhat distinct domains of published literature. The research that focused on just source variables has included gender (M. Burgoon, Dillard, & Doran, 1983; M. Burgoon, Dillard, Koper, & Doran, 1984), physician gender (M. Burgoon, Birk, & Hall, 1991; Klingle, 1993), credibility (M. Burgoon et al., 1975), and trustworthiness (Miller & Baseheart, 1969). All of this source-oriented research indicates that highly credible communicators are privileged, with a greater degree of freedom to use a range of language appeals and remain Persuasive. However, individuals who are Perceived as low-credible sources have a more limited bandwidth of messages they may use to be effective (Hamilton, Hunter, & M. Burgoon, 1990).
Derived from the fundamental propositions and supported by copious empirical findings, the following propositions are proffered:
Proposition 4: Highly credible communicators have the freedom (wide bandwidth) to select varied language strategies and compliance-gaining techniques in developing Persuasive messages, while low-credible communicators must conform to more limited language options if they wish to be effective.
Proposition 5: Because of normative impact of source credibility, highly credible sources can be more successful using low-intensity appeals and more aggressive compliance-gaining messages than can low-credible communicators using either strong or mild language or more prosocial compliance-gaining strategies.
Proposition 6: Communicators Perceived as low credible or those unsure of their Perceived credibility will usually be more Persuasive if they employ appeals low in instrumental verbal aggression or elect to use more prosocial compliance-gaining message strategies.
In addition to general source credibility differences, there are also gender-specific expectations about appropriate communication behavior. LET posits that males and females have differing bandwidths in terms of Persuasive message choices. Fairly rigid norms develop about what is and what is not acceptable use of language by males and females. Subsequently, enactment of compliant behaviors reinforces these norms. M. Burgoon et al. (1975) found support for the relationship between gender and variations in language intensity such that males could use more intense language and maintain their Persuasiveness, whereas females were not Persuasive using intense language.
Gender-specific expectations also exist about the appropriate use of more general compliance-gaining strategy use. The early research examining Marwell and Schmitt’s (1967a) compliance strategy checklist did not deal with the various strategies in any conceptually oriented manner. Hunter and Boster (1978, 1979) were the first to suggest that compliance-gaining strategies were a unidimensional continuum ranging from high to low on empathy. M. Burgoon et al. (1983) disagreed with this conceptualization and argued that the continuum was best viewed as variations in instrumental verbal aggression, not empathy.
Burgoon et al. (1983) then tested hypotheses derived from LET about gender and the use of compliance-gaining strategies, conceptualized as a continuum of instrumental verbal aggression. They reasoned that because male sources are more Persuasive using intense language, male sources should also be expected to use more antisocial and verbally aggressive compliance-gaining strategies. As anticipated, they found that males were expected to use more threat and aversive stimulation (verbally aggressive) strategies, while females were expected to use positive moral appeal and altruism (less verbally aggressive) strategies. Supported by research, LET posits the following:
Proposition 7: People in this society have normative expectations about appropriate Persuasive communication behavior that are gender specific such that (a) males are usually more Persuasive using highly intense Persuasive appeals and compliance-gaining message attempts, while (b) females are usually more Persuasive using low-intensity appeals and unaggressive compliance-gaining messages.
In addition to the vast amount of research on source characteristics, message variables have received considerable research attention under the LET framework. Those message features include language intensity (Buller, Borland, & M. Burgoon, 1998), opinion-atedness (Miller & Lobe, 1967), fear appeals (Miller & Hewgill, 1966), verbal aggression (M. Burgoon, 1989, 1990), and sequential message strategies such as foot-in-the-door (Dillard et al., 1984), and door-in-the-face (Cann et al., 1975).
One example from the research on such message variables is the interpretation of the fear appeal literature from this particular Perspective. Language expectancy theory is able to serve as an explanatory mechanism for existing—and sometimes contradictory— findings from various research efforts on fear-arousing messages. Extant research findings on fear appeals have shown that in some cases high fear appeals were more effective, while other studies have produced results suggesting the relative efficacy of appeals low in fear. Yet other studies have found no consistent effects based on intensity. Similar findings have been reported by researchers for opinionated language, profanity, and instrumental verbal aggression. M. Burgoon (1989, 1990) conceptualized fear appeals, opinionated language, profanity, and instrumental verbal aggression as special cases of language intensity that could be incorporated under the general framework of LET. The following proposition was then added to the original LET formulation:
Proposition 8: People in this society have normative expectations about the level of fear-arousing appeals, opinionated language, language intensity, sequential message techniques, and compliance-gaining attempts varying in instrumental verbal aggression appropriate to Persuasive discourse.
In extensive literature reviews, M. Burgoon (1989, 1990) suggested that highly fear-arousing messages, instrumentally verbally aggressive message strategies, profanity, and so on are negative violations of expectations for low-credible sources and other groups such as women and minorities. The use of such strategies should result in the same inhibition or enhancement of Persuasion as outlined in the earlier propositions about the effects of intensive language.
There has been very little research on receiver variables conducted directly from an expectancy Perspective. While the message variable research literature has looked at some individual difference variables, such as open-versus closed-minded individuals (Miller & Lobe, 1967) and receiver’s need for approval (Baseheart, 1971) in the fear appeals arena, such attention to receiver characteristics has not yielded much of interest.
However, one interesting line of inquiry suggests that the amount of anxiety experienced by the receiver will influence evaluations of a Persuasive message. M. Burgoon et al. (1975) tested whether people experiencing irrelevant fear (arousal) would respond differently from individuals in a low- or no-fear condition. They found that messages that were intense or verbally aggressive were generally ineffective with people already in state of high arousal such as that produced by the induction of irrelevant fear (anxiety unrelated to the Persuasive message itself). This is consistent with previous research that individuals who are highly aroused will resist experiencing additional arousing stimuli (Carmichael & Cronkhite, 1965) and suggests that bombarding aroused people with highly intense language or verbally aggressive message strategies would negatively violate expectations and inhibit Persuasion. More recently, Hamilton et al. (1990) examined the relationship between receiver anxiety and language intensity. Results indicated that when the receivers of the message were experiencing high anxiety, male sources were more Persuasive using low-intensity language. On the other hand, receivers not experiencing irrelevant fear were more persuaded by male sources using highly intense language. Overall, the state of mind that the receiver is experiencing at the time of receiving a Persuasive message will significantly affect the evaluation and outcome of the Persuasive attempt:
Proposition 9: Fear arousal that is irrelevant to the content of the message of the harmful consequences of failure to comply with the advocated position mediates receptivity to different levels of language intensity and compliance-gaining strategies varying in instrumental verbal aggression. Receivers aroused by the induction of irrelevant fear or suffering from specific anxiety are most receptive to Persuasive messages using low-intensity and verbally unaggressive compliance-gaining attempts but are unreceptive to intense appeals or verbally aggressive suasory strategies.
Active participation paradigm. This Persuasive situation involves an exchange in roles taken by the Persuader and the Persuade (target). Within the active participation paradigm, the target prepares a counterattitudinal message for public display. By developing and presenting a belief-incongruent message, the persuade is considered actively involved in the Persuasion process. It is argued that the act of preparing counterattitudinal messages increases arousal in the persuadee in the form of cognitive stress. The level of cognitive stress, in turn, affects the production of the counterattitudinal message. People experiencing high levels of cognitive stress do not want to increase their arousal level and, thereby, will produce less intense and more ambivalent messages and will elect to use less verbally aggressive messages (M. Burgoon & Miller, 1971). Under circumstances of increased cognitive stress, self-Persuasion is induced as people change their private attitudes to conform to public communication behavior. M. Burgoon and Miller (1985) translated these findings into an expectancy framework claiming that the act of producing messages that are not privately accepted induces a violation of expectations about one’s own norms and expectations about appropriate communication behavior. They offered the following propositions:
Proposition 10: Communicators experiencing cognitive stress produce less intense, more ambivalent messages and elect to use less aggressive compliance-gaining message strategies.
Proposition 11: When forced to violate their own norms about appropriate communication behavior by encoding highly intense messages or using aggressive compliance-gaining message strategies, communicators experience increased cognitive stress, which facilitates attitude change toward the belief-discrepant position being advocated.
Proposition 12: There is a direct linear relationship between level of language intensity used in counterattitudinal advocacy and concomitant attitude change.
Resistance to Persuasion Paradigm. Language expectancy theory was extended into the context of resistance to Persuasion as an extension and a refinement of McGuire’s (1964) inoculation theory. The concern was with how to make receivers more resistant to future Persuasive messages. A program of research looking at sequential message strategies in which pretreatment messages were viewed as developing expectancies and attack messages as either positively or negatively violating those induced expectations yielded some interesting and often counterintuitive findings (M. Burgoon & Chase, 1973; M. Burgoon et al., 1978; M. Burgoon & King, 1974; Miller & M. Burgoon, 1979). It was determined that inducing resistance to Persuasion is a two-step process: forewarning the target about the forthcoming message (pretreatment) followed by delivering the actual Persuasive message. Language expectancy theory posits that the pretreatment message provides the receiver with expectations about the subsequent Persuasive message. Whether the Persuasive message confirms or disconfirms the expectations afforded by the pretreatment message affects the Persuasive outcome. M. Burgoon and Miller (1985) translated this empirical research from the inducing resistance paradigm into a propositional framework consistent with their theorizing about the role of expectancies in the attitude change process:
Proposition 13: When supportive pretreatment strategies are used, attitude change following a subsequent Persuasive attack varies inversely with the linguistic intensity of the supportive pretreatment strategy.
Proposition 14: Refutational pretreatments create expectations by forewarning receivers about the nature of the forthcoming attacks such that (a) when Persuasive attack messages do not violate the linguistic expectations created by refutational pretreatments, maximum resistance to Persuasion is conferred, but (b) when the linguistic properties of attack messages violate the expectations created by refutational pretreatments either positively or negatively, receivers are less resistant to Persuasion.
According to LET, refutational messages create expectations about future attack messages simply by virtue of the language used in the pretreatment message. When highly intense pretreatment messages are delivered, expectations are induced for highly intense subsequent attack messages. If those expectations are violated, the receiver will be less resistant to the attack messages and Persuasion may occur. Thus, a receiver expecting a high-intensity attack message but receiving a more moderate one (a positive violation) may be Persuaded by the reasonableness of the moderate message (M. Burgoon & Miller, 1985). On the other hand, a receiver expecting a low-intensity attack message but receiving an intense one might not have felt threatened enough to prepare for the unexpectedly Persuasive attack (M. Burgoon & Chase, 1973). In this resistance to Persuasion model, the objective is to minimize violating expectations induced by pretreatment messages. However, in general, high-intensity encoding tends to be less Persuasive, especially when used in an attack message following any type of pretreatment effect (M. Burgoon & King, 1974).
Proposition 15: Given passive message reception, low-intensity attack messages are generally more effective than highly intense messages in overcoming resistance to Persuasion conferred by supportive, refutational, or combination pretreatments.
Proposition 16: When the Persuasive attack relies on an active participation strategy, a direct relationship exists between language intensity in the actively created attack and overcoming resistance conferred by any kind of pretreatment message strategy (supportive, refutational, or combination) such that matching intensity in the pretreatment and attack messages confers the most resistance to Persuasion.
Proposition 17: When receivers are exposed to more than one message arguing the same position, the confirmation or disconfirmation of linguistic expectancies in the first message systematically affects the acceptance of the second message such that (a) when linguistic expectations are positively violated in an initial message, the initial message is Persuasive, but a reversal of attitudes to the original position occurs after exposure to a subsequent message advocating the same counterattitudinal position, and (b) when linguistic expectations are negatively violated in an initial message, the initial message is not Persuasive, but receivers are more vulnerable to the arguments of a subsequent message advocating the same counterattitudinal position.
Future Directions of Language Expectancy Theory
Attention to the Theoretical Formulation and Generalizability of LET
Research emanating from LET has been claimed by its own primary formulator to being subject to criticism for possibly being teleological in nature (M. Burgoon, 1995). Because normative sociological or expected communication behaviors have not routinely been specified on an a priori basis in many empirical tests of this theory, it can be difficult—if not impossible—in some situations to determine when a positive or negative violation of expectations has actually occurred. Some investigators have concluded that when attitude/behavior change occurs, a positive violation of expectations must have occurred. Similarly, when no attitude/behavior change occurs or there is a boomerang effect, the conclusions drawn are that a negative violation must have occurred. Such interpretations of the empirical data make the theoretical model unfalsifiable. While such a criticism should be directed more at the way experimental social science is sometimes conducted than at the theoretical model Per se, a priori specification of expectations in experimental situations and, certainly, naturally occurring research settings makes for much stronger scientific claims.
While attention to specification of exactly what expectations people have about appropriate communication behavior requires additional scrutiny, there is also an attendant need to specify even more clearly exactly what is meant by “expectations.” A pivotal definitional distinction that needs to be made regarding expectancies is whether they are regarded as predictive or prescriptive. Predictive expectancies Pertain only to the typicality of behavior—to its central tendency or regularity of occurrence (as might be measured by the mean, median, or mode). Prescriptive expectancies Pertain to idealized standards for conduct. They capture evaluative connotations of behavior such as appropriateness and desirability; in other words, they carry an associated valence ranging from negative to positive (J. Burgoon & M. Burgoon, 2001). Recent work (e.g., J. Burgoon & White, 1997; Floyd & J. Burgoon, 1999) has attempted to create a crisPer distinction by reserving the term expected for the predictive variety and using the term desired for that which is valenced. Language expectancy theory has avoided dealing almost entirely with such definitional issues, and by default it has accepted predictive expectancies and not really idealized behavior as earlier formulations might have implied. The working definition of appropriate Persuasion behavior was clearly about predicted future behavior. It is not entirely clear what might be gained by refining the predictions to incorporate distinctions between violations of preferences and predictions, but it is of sufficient theoretical import to warrant serious consideration.
Burgoon and M. Burgoon (2001) pointed out that the ubiquity of expectancies in guiding human conduct is so widely conceded that it is perhaps unsurprising that expectancy-related concepts populate, under one guise or another, so many theories of human communication and psychology. They dealt in detail only with LET, Berger and colleagues’ (Berger, Fisek, Norman, & Zelditch, 1977; Ridgeway & Berger, 1986; Ridgeway & Walker, 1995; Shelly, 1998) expectation states theory, and J. Burgoon’s expectancy violations theory (J. Burgoon 1992, 1993, 1995; J. Burgoon & Hale, 1988). In addition to having generated a substantial body of research, these theories were highlighted because they represent a good sampling of Perspectives on expectancies and their violations emanating from the fields and foci of social psychology, sociology, and communication. Even though expectancies are central to many research efforts, there have been very limited attempts to incorporate theorizing and compare empirical findings across what are now quite separate lines of inquiry. That remains a challenge for people interested in better explaining the role of expectations in social influence processes.
Obviously, the empirical evidence supporting LET is the product of research conducted in but one culture, namely North America. However, it should be noted that if the theory is robust, the effects of violations of expectations, whether positive or negative, should be culturally invariant. That is certainly not to say that expectations about what is appropriate or even desired Persuasive communication behavior will be the same in different cultural contexts. Obviously, no one would expect this to be the case. Not only would it be valuable to know the extent to which expectations are different or similar across cultures, but it would also be enlightening to determine whether the seeming magnitude of effects of expectancy violations on attitude and/or behavior change are as dramatic in other cultures as the research evidence suggests they are in this culture.
Testing LET in Natural Settings
The bulk of the research supporting LET has been conducted in laboratory-like research settings. Data in support of the theoretical formulation provide considerable encouragement for proponents of this Perspective. Clearly, testing in natural settings with a variety of methods and media being used to change socially important attitudes and behaviors is the long-term goal. That testing is now in its first stage. Because of funding opportunities, a step has been taken in testing a small part of LET in a longitudinal field study concerning skin cancer prevention.
Based on LET, Buller et al. (2000) predicted that members of credible institutions, such as school administrators, health care professionals, and physicians, would have a wider bandwidth of language choices and would be able to use high-intensity, graphic messages aimed at parents and children, encouraging family sun safety. Furthermore, the researchers proposed that pediatricians and other medical doctors would be most successful at gaining compliance when they used more intense, verbally aggressive messages as opposed to low-intensity messages. It was further hypothesized that, overall, highly intense (instrumentally verbally aggressive) prevention messages would be more effective than messages lacking in instrumental verbal aggression (low-intensity messages). Moreover, as derived directly from LET, an interaction between instrumental verbal aggression and organizational structure of Persuasive messages was hypothesized such that (a) highly verbally aggressive messages organized in a deductive format would be the most effective strategy to produce change, and (b) low-intensity (low instrumental verbal aggression) messages organized in an inductive pattern of reasoning would be superior to high-intensity, inductive messages. The results of this 4-year longitudinal study, in which each family received multiple messages across time, were supportive of the LET predictions.
Buller et al. (1998) examined data from the same field study to test the prediction that the effects of message intensity may vary as a function of the stage of behavior change of individual participants. The concept of stage changes claims that behavior change is a process whereby people progress through phases from some form of contemplation to changes in beliefs, thoughts, and receptivity until at some point the desired behavior change is enacted. They found that highly graphic and intense messages were effective with people who were already receptive to taking sun safety measures. Such messages from highly credible sources were both expected and effective. However, for the subset of the population who had already decided not to comply with the desired changes, highly intense messages were seen as inappropriate and were not Persuasive.
Data from the family sun safety project also provided valuable information about the Persistence of change when highly intense, graphic messages initially induced attitude and behavioral changes. While most sun safety campaigns concentrate their Persuasive activities during peak sun Periods, this study looked at on-peak and off-peak seasonal differences in message effects. It was encouraging to find that when people were exposed to highly intense, deductively constructed messages, their sun safety behavior remained at very high levels 6 months later when the risk of sun damage was at its lowest point.
This campaign is an exemplar of the kind of research that needs to be done in further testing the LET framework. It tested but a very small part of the entire formulation. However, it did provide support in a natural setting for some very fundamental elements of the theory. It also allowed the use of multiple messages delivered by different sources in various media over time. This demonstration of the predictive power of the theory (or at least a small number of hypotheses derived from the more complete formulation), coupled with data on the Persistence of behavior changes, is at least a small step toward establishing claims of ecological validity for a theory mostly borne of laboratory research with too many convenience samples as the basis for inference to some larger population. This is the kind of natural setting research that needs to be conducted in arenas such as anti-tobacco campaigns, drug abuse reduction efforts, elimination of risky behavior, and a variety of other arenas ripe for research attention by Persuasion scholars.
Extending LET to Interpersonal Influence Situations
As with most extant theories of Persuasion, there is more evidence on suasory communication in one-to-many contexts than in face-to-face influence attempts. It has already been suggested that much would be gained by incorporating the thinking of other theories with a more interpersonal orientation. It is also the case that future research needs to extend LET to better focus on interpersonal influence attempts.
While any number of interpersonal contexts are replete with social influence attempts, the clinical practice of medicine is one worthy of study, if only because it is an exemplar of face-to-face Persuasion with outcome measures of extreme import. One relatively recent research project built directly on the earlier LET propositions (a) to provide more precision in specifying what communication behaviors are expected of physicians by their patients, (b) to incorporate additional variables to directly test the effects of verbal aggression on satisfaction and compliance, and (c) to determine whether LET could be used to generate hypotheses and provide results that would be parsimonious explanations and potent predictors of the outcome variables of concern in the medical context. This research effort, in line with earlier comments about needed theoretical precision, eliminates the possibility of a teleological interpretation of results in that differences in enacted and expected behaviors are clearly specified in advance. Moreover, given such precision in specifying expected behaviors in advance and measuring the valence of such communication styles, clear distinctions can be made among the expected, enacted, and preferred communication behaviors of physicians. Such a procedure allows an unequivocal specification of what sets of behaviors should be considered as positive or negative violations of expectations and what behaviors merely conform to societal norms about communication in the medical context.
Burgoon et al. (1991) applied the general expectancy model to the medical context. As indicated previously, early research (M. Burgoon & Miller, 1985) clearly demonstrated that the bandwidth of normatively expected behaviors is not invariant for individuals or groups of people. For example, it has been clearly demonstrated that females have a much narrower bandwidth of socially acceptable behavior than do males in this society. In other words, females are limited in their choice of strategies if they wish to avoid negatively violating expectations and being ineffective in compliance-gaining attempts. It is very difficult for a female to positively violate expectations. Yet it is quite likely that any deviations—even relatively trivial changes—from the expected roles of females will result in negative violations of expectations andincrease the probability of noncompliance. A significant problem in studying physicians’ enacted and expected behavior is that any resulting conclusions may generalize only to male physicians. Little research has addressed the influence of gender on the physician-patient interaction and subsequent outcomes. The lack of research in this area is due, in part, to the fact that, until recently, few women have entered those medical specialties most responsible for adult health care, namely family practice and internal medicine. Thus, generalizations advanced about expectations in clinical medicine (M. Burgoon & J. Burgoon, 1990) may be applicable only to male physicians.
While there has been little research on the effects of gender differences in the medical context, there is a wealth of research to suggest that females are more nurturing, more likely to express caring and concern, more empathic in their enacted communication behaviors, and less verbally aggressive (e.g., M. Burgoon et al., 1983; Eakins & Eakins, 1978; Infante & Wigley, 1986). To the extent that such behaviors are products of early sex role socialization, they are most resistant to change (Scanzoni, 1975). It is therefore unlikely that professional socialization would completely counteract sex role differences among physicians. This is supported by the finding that female medical students tend to elicit more psychosocial issues from patients, while males tend to see themselves in control of communication with patients and give directions and make demands (Rosenberg, 1979). If female physicians have indeed been socialized to the traditional female sex role, they will be more nurturing and expressive and will have stronger interpersonal communication orientations than male physicians.
Another line of reasoning, based on notions of credibility, suggests that female physicians have a high probability of negatively violating expectations if they do not use verbally unaggressive message strategies in compliance-gaining attempts. While physicians, both male and female, have relatively high normative status in this society, there are differences in the Perceptions of credibility of males and females in the health care professions. For example, Engleman (1974) found that a majority of both men (84%) and women (75%) preferred a male for their regular physician. A limited amount of research attests to the fact that while female physicians are held in somewhat high esteem by most people in this society, they are still seen as less credible than their male counterparts. Therefore, according to the tenets of LET, female physicians should have less freedom to use aggressive strategies. This credibility differential, coupled with socialization processes, results in the enactment of unaggressive strategies because part of the female role suggests that any deviation from such verbally unaggressive compliance-gaining strategies will increase noncompliance.
It is clear that male physicians have a great deal of freedom to select compliance-gaining strategies. First, they have high normative status, and the socialization process provides them with an extremely wide bandwidth of acceptable behaviors. Second, much research demonstrates that socialization makes aggressive behavior not only acceptable but also preferred for highly credible male communicators in this society. Yet the expected communication behaviors for male physicians, as evidenced by the generalizations put forth by M. Burgoon and J. Burgoon (1990), are to be affectively neutral, give directions, and use negative expertise. All of these strategies fall near the center of the instrumental verbal aggression continuum. However, there is considerable evidence to suggest that high levels of verbal aggression on the part of male physicians are often perceived by patients as expressions of Personal concern and considered a positive violation of expectations, in creasing levels of compliance. Moreover, because male physicians are expected to be affectively neutral in treatment and prevention situations, affiliative strategies such as the expression of caring and concern can also be a positive violation of expectations for male physicians. Such Personalization of the clinical visit is preferred by most people but rarely experienced in visits to male physicians. Thus, the only strategies that seem to be ineffective for male physicians are the ones currently used most often: a combination of simple direction giving and expertise. M. Burgoon et al. (1991) found convincing support for these expectancy-based hypotheses in this applied context.
Obviously, this kind of research needs to be replicated and extended to other face-to-face influence attempts to determine whether LET has the same predictive power in other situations. Again, this one study is offered only as an example of how LET should be tested in different contexts.
Reinforcement Expectancy Theory: Future Developments
What is needed is not only to follow the above injunctions to extend the boundary conditions of LET by dealing with multiple messages and face-to-face interactions but also to resurrect some important theoretical legacies that have been prematurely dismissed by Persuasion researchers. Klingle (1993) developed a theoretical Perspective that extends LET in these important ways. She was interested in methods for improving both initial and long-term medical adherence and for examining physician-patient communication over the course of several different clinical interventions. Again, while the medical context was selected for this particular study, the important point is to urge the extension of this theory to explain situations in which sequential messages violate or conform to established expectations.
There is a lack of research assessing how physicians should communicate to assist people in dealing with problems that occur over time. Using the recently developed reinforcement expectancy theory (Klingle, 1993), a two-part study was undertaken (Klingle & M. Burgoon, 1995) to assess the effectiveness of communication strategies designed to improve both initial and long-term medical adherence. The first study analyzed patients’ evaluations of communication regard strategies and the effectiveness of these strategies in initial encounters. As predicted, negative regard strategies used by male physicians were perceived as more appropriate than negative regard strategies used by female physicians. In addition, physician gender interacted with strategy effectiveness such that male physicians were Persuasive if they used either positive or negative regard strategies, whereas female physicians were limited to using only positive regard strategies. The results also indicated that the use of negative regard strategies by male physicians did not hinder patient satisfaction or physician Perceptions, whereas the use of negative regard strategies by female physicians was negatively related to these outcome measures. This provided a relatively complete replicate of the earlier research conducted from a LET Perspective that dealt with one-time medical encounters (M. Burgoon et al., 1991).
The reinforcement expectancy framework tested in the second study argued that occasional use of nonrewarding communication would facilitate communication effectiveness for both male and female physicians in ongoing physician-patient relationships. A very complex design was used to test the effects of neutral, negative, and positive reward strategies over five different clinical visits. As predicted, mixed types of strategies over time were generally superior to either all negative or all positive strategies over time with the same patient.
The investigation proved successful on several fronts. The fact that previous research on strategy effectiveness in initial encounters could be replicated using a different influence categorization scheme with a different population of individuals over a variety of consultation sessions suggests that the findings are probably quite robust. This investigation extends previous research by offering tentative support for the claim that health care providers who have repeated exposure to a patient can strategically use nonaffiliative communication to improve adherence rates while maintaining satisfaction and Person Perception. Specifically, it was argued that strategies involving a combination of reinforcing and non-reinforcing communication (i.e., reinforcement violations) are more Persuasive than continual reinforcement or continual nonreinforcement.
While many psychologists and communication scholars have recently been less concerned with reinforcement/learning models, Perhaps O’Keefe (1990) was premature in his unsubstantiated claim that for “good conceptual and empirical reasons,” such reinforcement-based inquiries “are poor prospects” for people interested in the efficacy of Persuasive messages (p. 7). Steps are being taken to determine whether out-of-favor positions, such as learning models and the theory of psychological reactance, can indeed inform and be informed by LET. This is at least one journey worth taking as we ponder future efforts to extend the predictive and explanatory power of what has proved to be an intriguing explanation of social influence processes.