Metaphor and Persuasive Communication: A Multifunctional Approach

Metaphors are pervasive in both mass communication and interpersonal exchanges and can play an important role in persuasion. Metaphor serves multiple functions in persuasive communication, and the effect of metaphor on persuasion is potentially mediated by multiple psychological process mechanisms. Nevertheless, we propose that past and future research in this area can be organized or grouped into three simple categories. First, metaphorical statements can activate information that is directly applied to the communication topic and thereby influence attitudes toward the communication topic. Second, metaphorical language may influence impressions of the communication source and thereby impact attitudes toward the communication topic. Third, metaphors may affect attitudes toward the communication topic by influencing the direction or amount of elaboration that takes place when recipients process literal statements contained in the communication. A review of past research is organized into these three categories, and proposals for future research in each category are introduced. It is concluded that future research within each of these domains should focus on two related questions: under what conditions does metaphor elicit a given psychological process in the receiver (e.g., attribute mapping, valence transfer), and under what conditions will a given process result in an increase versus decrease in persuasion? Metaphor is pervasive in both mass communication and personal linguistic exchanges (Gibbs, 1994; Lakoff, 1987; Lakoff & Johnson, 1980). Poets tell us that love is a ‘light,’ a ‘fire,’ or form of ‘energy’. Politicians have sought to ‘build a bridge to the 21st century,’ ‘ignite a thousand points of light,’ or provide ‘tax relief’. Metaphors can furnish vivid images, convey multiple meanings in a concise fashion, or express that which cannot be stated in literal terms (Gibbs & Bogdonovich, 1999; Graesser, Mio, & Millis, 1989; Ortony, 1975; Paivio, 1979). Some theorists have even argued that all higher-order cognitive functioning is metaphorical in nature (e.g., Lakoff, 1987; Lakoff & Johnson, 1980). Metaphorical statements often pertain to an attitude object (e.g., ‘Pollution is a cancer,’ ‘Elvis is king’), and as such, may play an important role in persuasive communications. A metaphor contains a topic and a vehicle (Ortony, 1975; Richards, 1936). The topic is the object or phenomenon being described, whereas the vehicle is some other object or phenomenon that conveys a certain meaning about the topic. For example, ‘Juliet is the sun’ contains ‘Juliet’ as the topic and ‘sun’ as the vehicle. Models of metaphor comprehension often assume the comprehension process entails mapping semantic or evaluative implications of the vehicle onto the topic (e.g., Gentner, 1983; Gibbs, 1992; see Ortony, Reynolds, & Arter, 1978 for a review). Thus, ‘Juliet is the sun’ might convey that Juliet is a source of energy, warmth, and happiness, or that Juliet simply possesses a positive valence. Alternatively, ‘television is an addiction’ might convey that watching television is unhealthy, habit forming, and difficult to stop, or simply that watching television is undesirable (cf. Bowdle & Gentner, 2005; Chiappe & Kennedy, 2001; Gentner Social and Personality Psychology Compass 4/9 (2010): 783–794, 10.1111/j.1751-9004.2010.00292.x a 2010 The Authors Social and Personality Psychology Compass a 2010 Blackwell Publishing Ltd & Clements, 1988; Gibbs & Gerrig, 1989; Glucksberg & Keysar, 1990; Kintsch, 2001; Ortony, 1979; Ortony, Vondruska, Foss, & Jones, 1985). Work regarding metaphor and persuasive communication includes purely theoretical writings, nonexperimental studies, and experimental research (Mio, 1997). Empirical studies regarding the psychological mediators of metaphor effects on persuasion is growing into a substantial body of literature (e.g., Bowers & Osborn, 1966; Frey & Eagly, 1993; Graesser et al., 1989; Hitchon, 1997; Johnson & Taylor, 1981; Landau, Sullivan, & Greenberg, 2009; Mio, 1996; Ottati, Rhoads, & Graesser, 1999; Read, Cesa, Jones, & Collins, 1990; Sopory & Dillard, 2002). Much of this research has been performed by scholars outside of social psychology (e.g., communication, cognitive psychology). As a consequence, this literature has yet to fully incorporate the implications of mainstream social psychological models of persuasion (e.g., the Elaboration Likelihood and HeuristicSystematic Models). One objective of this study is to more fully incorporate the implications of these models when considering research on metaphor and persuasion. An additional objective of this study is to provide a conceptualization of metaphor effects on persuasion that is both comprehensive and simple. Past research has often focused on very specific effects of metaphor on persuasion, and reviews of the literature have considered a large list of specific hypotheses, predictions, and findings (e.g., Sopory & Dillard, 2002). What is needed, however, is a coherent overarching conceptual framework for organizing this expanding number of hypotheses, predictions, and findings. Social psychological approaches to persuasion have traditionally emphasized relatively simple taxonomies that subsume a vast array of effects (e.g., ‘Who says what to whom?’). This study takes an analogous approach when considering past and future research regarding the effects of metaphor on persuasion. We assume that metaphor serves multiple functions in persuasive communication, and that the effect of metaphor on attitudes toward the communication topic is complex and mediated by multiple psychological process mechanisms. Nevertheless, we believe it is possible to organize or group this work into three general categories (see Figure 1). First, metaphorical statements might influence the recipient’s impression of the communication topic in a relatively direct manner. When this is the case, the effect of metaphor on attitude toward the communication topic is independent of other cues (e.g., source credibility, argument strength). Second, the effect of metaphor on attitude toward the communication topic might be mediated by the message recipient’s impression of the communication source. Lastly, the effect of metaphor on attitude toward the communicaFigure 1 Three pathways by which metaphor can influence attitudes toward a communication topic. 784 Metaphor and Persuasive Communication a 2010 The Authors Social and Personality Psychology Compass 4/9 (2010): 783–794, 10.1111/j.1751-9004.2010.00292.x Social and Personality Psychology Compass a 2010 Blackwell Publishing Ltd tion topic might be mediated by the manner in which the message recipient processes literal statements contained in the communication. We consider several psychological process mechanisms within each of these three categories. In doing so, we do not exclusively focus on effects of metaphor that are distinct or unique from the effects of literal language. To the contrary, we believe that a comprehensive conceptualization of metaphor and persuasion should not only consider effects that are unique to metaphor (e.g., comprehension of a novel metaphor elicits an ‘ah ha!’ experience) but also effects that are similar to those obtained using more literal forms of communication (e.g., metaphor facilitates organization). Metaphor and Persuasion A meta-analysis has revealed that metaphor generally increases persuasion (Sopory & Dillard, 2002). However, the magnitude and direction of this effect vary considerably across moderating conditions. Metaphor is most likely to increase persuasion when it identifies a novel comparison, appears at the onset of related arguments, or is semantically congruent with other metaphorical statements contained in the communication (Read et al., 1990; Sopory, 2006; Sopory & Dillard, 2002). Some research suggests that metaphor is most effective when the message recipient is familiar with the communication topic or knowledgeable in the topic domain (Johnson & Taylor, 1981; Sopory & Dillard, 2002; Roehm & Sternthal, 2001; but see Bosman & Hagendoorn, 1991). On the other hand, other research indicates metaphor can reduce persuasion (Bosman & Hagendoorn, 1991), especially when the metaphor fails to afford clear semantic linkages with literal arguments contained in the communication (Krumdick, Ottati, & Deiger, 2004). To better understand this conflicting pattern of findings, it is useful to organize our discussion in terms of the three aforementioned categories. These are (1) direct effects of metaphor on attitudes toward the communication topic, (2) metaphor effects that are mediated by impressions of the communication source, and (3) metaphor effects that are mediated by manner in which the receiver processes literal statements contained in the communication. Relatively direct effects of metaphor on attitude toward the communication topic As noted previously, many researchers have suggested that metaphor comprehension entails mapping semantic or evaluative implications of the vehicle onto the topic. Thus, for example, ‘Professor Williams is a Nazi’ might lead the message recipient to infer Professor Williams is domineering, cruel, racist, or simply unpleasant. As a consequence, the message recipient might be persuaded to adopt a negative attitude toward Professor Williams. Consistent with this conceptualization, metaphors can elicit an assimilation effect wherein evaluation of the topic is assimilated toward evaluation of the vehicle (Johnson & Taylor, 1981; Mio, 1993; Mio, Thompson, & Givens, 1993). This effect is relatively direct because it is not mediated by the recipient’s impression of the source or the way in which the recipient processes other literal arguments contained in the communication. Interestingly, this assimilation effect can emerge even when a metaphor is indirectly implied, but not explicitly stated in the communication (see Landau et al., 2009). Assimilation effects of this nature are potentially mediated by two psychological process mechanisms. First, the metaphor might activate specific semantic attributes that are mapped on to (or highlighted in) the topic (Edelman, 1971; Hitchon, 1997). In this case, the metaphor serves as a concise expression that conveys a set of arguments (e.g., ‘Professor Metaphor and Persuasive Communication 785 a 2010 The Authors Social and Personality Psychology Compass 4/9 (2010): 783–794, 10.1111/j.1751-9004.2010.00292.x Social and Personality Psychology Compass a 2010 Blackwell Publishing Ltd Williams is domineering,’ ‘Professor Williams is cruel,’ ‘Professor Williams is racist’) that elicit persuasion in the direction advocated. Of course, only attributes potentially shared by the vehicle and topic are mapped on to (or highlighted in) the topic (Hitchon, 1997; Kintsch, 2001; Newsome & Glucksberg, 2002). For example, the metaphor ‘Professor Williams is pig’ does not lead the receiver to infer ‘Professor Williams has four legs’. Thus, evaluation of the topic is specifically assimilated toward evaluation of the shared features. A second possibility is that the overall valence of the vehicle is mapped on to the topic. In this case, the metaphor simply functions as an evaluative judgment (e.g., ‘Professor Williams is undesirable’) that produces persuasion in the direction advocated. If this occurs, evaluation of the topic should be assimilated toward global evaluation of the vehicle. In fact, research indicates that evaluation of the topic is assimilated toward evaluation of features potentially shared by the vehicle and topic, even when this evaluation is the opposite of the global vehicle evaluation (Hitchon, 1997). Thus, the attribute-mapping model has garnered more support than the global valence-transfer model. Future research should explore conditions that potentially moderate effects of this nature. For example, it is possible that the attribute-mapping model (i.e., metaphor functions as a set of arguments) is supported under high elaboration likelihood conditions (e.g., high personal relevance) whereas the valence-transfer model is supported under low elaboration likelihood conditions (e.g., low personal relevance). From this perspective, attribute-mapping constitutes a form of central processing whereas valence-transfer constitutes a more peripheral form of persuasion (see Petty & Cacioppo, 1986, for a multiple role analysis of persuasion cues). Indirect effects mediated by the recipient’s impression of the communication source Many researchers have suggested that metaphor elicits positive impressions of the communication source which increase persuasion in the direction advocated (Bowers & Osborn, 1966; McCroskey & Combs, 1969; McGuire, 2000; Reinsch, 1971, 1974). This connects metaphor and persuasion research to a vast literature that documents the role of source credibility and source attractiveness in persuasion. Message recipients may simply regard metaphor as a sign of genius or brilliance (Sopory & Dillard, 2002). Alternatively, metaphor might elicit a favorable impression of the source when it enlightens the listener by identifying similarities between the vehicle and topic that have not been previously considered. Research that demonstrates novel metaphors are more persuasive than conventional or ‘dead’ metaphors is compatible with this later possibility (Sopory & Dillard, 2002). In the former case, metaphorical language functions as a simple, heuristic cue that signals the source is highly intelligent and credible. Presumably, this is most likely to occur when elaboration likelihood is low. In the latter case, the message recipient must analyze and appreciate the ingenuity of the metaphorical statement before deriving a positive impression of the source. This should occur when elaboration likelihood is high (for a related discussion see Petty & Cacioppo, 1986). A review of the literature, however, does not support the blanket assumption that metaphor increases persuasion because it enhances source credibility. In some cases metaphor increases perceptions of source credibility (Bowers & Osborn, 1966). However, in other cases metaphor fails to influence or even decreases perceptions of source credibility (Bowers & Osborn, 1966; McCroskey & Combs, 1969; Reinsch, 1971; see Sopory & Dillard, 2002 for null effects in a meta-analysis). Metaphor is associated with an increase in source attractiveness (Read et al., 1990; Mio, Riggio, Levin, & Reese, 2005; see Sopory & Dillard, 2002 for a meta-analysis). Nevertheless, research has failed to document that the 786 Metaphor and Persuasive Communication a 2010 The Authors Social and Personality Psychology Compass 4/9 (2010): 783–794, 10.1111/j.1751-9004.2010.00292.x Social and Personality Psychology Compass a 2010 Blackwell Publishing Ltd effect of metaphor on persuasion is mediated by source impressions. In some studies, metaphor influences attitudes toward the communication topic, but fails to influence source ratings (McCroskey & Combs, 1969; Reinsch, 1971). In other studies, metaphor influences source ratings but fails to influence attitudes toward the communication topic (Reinsch, 1974). No study has presented a formal mediational analysis that demonstrates source impressions mediate the effect of metaphor on attitudes toward the communication topic. Future research on this question should explicitly acknowledge that metaphor can increase or decrease the favorability of source ratings (McGuire, 2000). As a consequence, the effect of metaphor on source impressions should produce increased persuasion in some circumstances and decreased persuasion in other circumstances. For example, metaphors that afford clear semantic linkages with other information contained in a communication might increase the favorability of source ratings and thereby increase persuasion in the direction advocated. In contrast, metaphors that are irrelevant to other information contained in a communication might reduce the favorability of source ratings and thereby decrease persuasion in the direction advocated. Approaches that assume metaphor must elicit a blanket positive influence on source credibility will fail to uncover more nuanced effects of this nature. Indirect effects mediated by the processing of literal arguments in the communication: metaphor influences the direction of elaboration Metaphorical language might influence the degree to which the recipient generates positive or negative thoughts about the topic when processing literal arguments contained in a communication. Or, metaphors might influence the degree to which the recipient generates proor counter-arguments when processing the literal arguments. In these cases, metaphor influences the direction of elaboration upon the literal arguments. Of course, metaphor can only influence the direction of elaboration if message recipients are, in fact, elaborating on the message. Therefore, these effects should primarily emerge when elaboration likelihood is high (for a related discussion, see Petty & Cacioppo, 1986). This might result in a corresponding directional effect on attitudes toward the communication topic. Effects of this nature have yet to be fully explored, and therefore should be considered in future research. At least three psychological mechanisms might produce this sort of effect. Specifically, a metaphor might activate an expectancy, influence interest in the communication, or influence the coherence of a persuasive communication. In each of these instances, metaphors might produce an effect on attitude toward the communication topic that is mediated by the direction of elaboration. Metaphor, expectancy, and the direction of elaboration. When a metaphor precedes literal communication arguments, it can activate a positive or negative expectancy regarding the topic. This might produce cognitive responses to the literal arguments that are congruent with the expectancy, resulting in attitude change in the direction of the expectancy (see Renstrom, Krumdick, & Ottati, 2008, for related evidence). Assume, for example, that ‘Professor Williams is an encyclopedia’ precedes ‘Professor Williams assigns a lot of homework’. In this case, the metaphor vehicle (i.e., ‘encyclopedia’) implies a positive trait (e.g., ‘knowledgeable’). As a consequence, cognitive responses to the literal statement may be relatively positive (e.g., ‘The homework challenges us to expand our range of knowledge’). Alternatively, assume ‘Professor Williams is a slave driver’ precedes the same literal statement. In this case, the metaphor vehicle (i.e., ‘slave driver’) implies a negative Metaphor and Persuasive Communication 787 a 2010 The Authors Social and Personality Psychology Compass 4/9 (2010): 783–794, 10.1111/j.1751-9004.2010.00292.x Social and Personality Psychology Compass a 2010 Blackwell Publishing Ltd trait (e.g., ‘sadistic’). Thus, cognitive responses to the literal statement may be negative (e.g., ‘He likes to make us suffer’), and the recipient may form a negative attitude. In most cases, the source of a communication will select a metaphor that possesses a valence that is congruent with the desired direction of influence. Thus, in most cases, this process should increase persuasion in the direction advocated. To test this hypothesis, the valence of a metaphor that precedes related literal arguments should be manipulated (negative versus positive valence). After receiving the communication, participants should be asked to list their cognitive responses and rate their attitude toward the communication topic. If the ‘expectancy hypothesis’ is correct, valence of metaphor should produce an assimilation effect on attitudes toward the topic, and this effect should be mediated by the direction of elaboration (number of pro-arguments minus number of counter-arguments listed). Metaphor, interest, and the direction of elaboration. Following the lead of Aristotle, McGuire (2000) has proposed that metaphor can increase interest in the communication, and thereby increase persuasion in the direction advocated. Other scholars have contended that metaphorical language can elicit vivid and evocative images (Billow, 1977; Graesser et al., 1989; Ortony et al., 1978; Paivio, 1979), excite the imagination (Mio, 1993; Reinsch, 1971), enliven discourse (Bowers & Osborn, 1966; Zashin & Chapman, 1974), produce a pleasurable ‘ah ha’ experience (Kreuz & Roberts, 1993; Mio, 1997), and engage active intellectual attention (Bowers & Osborn, 1966; McGuire, 2000). Moreover, research confirms that metaphor can indeed increase interest (Deiger & Ottati, 2002; Kreuz & Roberts, 1993; see Read et al., 1990; for related evidence). According to McGuire (2000), this should attract attention to the message, induce a sympathetic mood, and thereby elicit favorable cognitive responses to the message. This model predicts that metaphor will increase persuasion and that this effect will be mediated by interest, attention, sympathetic mood, and the generation of favorable cognitive responses. A formal empirical test of the mediational pattern predicted by McGuire does not exist in the literature. Moreover, McGuire (2000) has acknowledged that metaphors can reduce interest under some circumstances, and research confirms that this is indeed the case (Ottati et al., 1999). Clearly, McGuire’s proposed model needs to be adjusted to accommodate findings of this nature. Ottati et al. (1999) have demonstrated that a metaphor (e.g., a sports metaphor) increases interest when it is compatible with the recipient’s personal preferences and interests (e.g., sports metaphor presented to a sports enthusiast). Does McGuire’s (2000) model apply under conditions of this nature? It probably does not. As will soon become apparent, effects of this nature occur when the literal arguments are strong, but the reverse effect emerges when the literal arguments are weak (Ottati et al., 1999). Thus, even when metaphor increases interest, it does not necessarily increase the generation of favorable cognitive responses. Although this finding is compatible with an alternative conceptualization that assumes metaphor influences the amount (not direction!) of elaboration, it is not consistent with the implications of McGuire’s (2000) proposed model. Thus, extant research has failed to document a pattern of findings that conforms to McGuire’s proposed model, and related work suggests interest influences the amount (not simply the direction) of elaboration. Metaphor, coherence, and the direction of elaboration. Independent of effects involving interest, metaphors may also influence the ‘coherence’ of a communication. Because metaphors can evoke and highlight a number of semantic pathways and relational connections among concepts and ideas (Gentner, 1982, 1983), the accompanying arguments in a com788 Metaphor and Persuasive Communication a 2010 The Authors Social and Personality Psychology Compass 4/9 (2010): 783–794, 10.1111/j.1751-9004.2010.00292.x Social and Personality Psychology Compass a 2010 Blackwell Publishing Ltd munication may be integrated together more easily and coherently when a metaphor is present (Allbritton, McKoon, & Gerrig, 1995). Thus, metaphorical statements may provide an overarching schematic framework that guides interpretation, facilitates organization, and facilitates systematic processing of literal statements contained in a communication (Billow, 1977; Mio, 1993, 1996, 1997; Read et al., 1990; Whaley, 1991). For example, ‘Work is an addiction’ might facilitate processing of literal statements that indicate ‘Corporate executives are unable to relax’ or ‘Vacations make a business person anxious’. If a metaphor increases coherence in this fashion, it might increase the generation of pro-arguments (or decrease counter-arguing), and thereby increase persuasion in the direction advocated. This ‘cognitive coherence’ model is supported by research that indicates metaphor is more persuasive if it appears at the onset as opposed to conclusion of a communication (Read et al., 1990; Sopory & Dillard, 2002; for related evidence see Renstrom et al., 2008). To guide encoding, interpretation, and organization of a message, a metaphor should presumably be activated before (not after) the specific message arguments are processed. This model is also supported by research that suggests metaphor is most effective at producing persuasion when the message recipient is knowledgeable with regard to the topic domain (Johnson & Taylor, 1981; Roehm & Sternthal, 2001; Sopory & Dillard, 2002). Presumably, this is because knowledgeable individuals are better able to apply the metaphor when encoding, interpreting, and organizing the specific message arguments. It seems unlikely, however, that all metaphorical utterances increase cognitive coherence. Under some conditions, metaphors might confuse or distract the listener, and thereby interfere with the listener’s ability to encode, organize, or systematically process literal statements (Billow, 1977; Frey & Eagly, 1993; Krumdick, Renstrom, Aalai, & Ottati, 2007; Mio, 1996; Mio & Lovrich, 1998). This might occur if the metaphor fails to afford clear semantic linkages with other statements contained in the communication. Interestingly, some have suggested that this might distract listeners who would otherwise counterargue the message (Sopory & Dillard, 2002). This implies that distracting metaphors will actually increase persuasion in the direction advocated. Of course, it is also possible that this kind of metaphor will reduce the perceived coherence and strength of the communication and thereby reduce favorable cognitive responses to the communication. From this perspective, distracting metaphors should reduce persuasion. To date, research findings are more compatible with this later possibility (Ottati et al., 1999; Sopory & Dillard, 2002). Future research needs to more explicitly contrast conditions in which metaphor will function as a distracter from conditions in which metaphor will increase cognitive coherence. Past research has taken an indirect approach to this question by comparing metaphor effects on receivers possessing high versus low levels of expertise, presumably because metaphors will increase coherence more among experts but perhaps serve to confuse or distract novices. As we have noted, a more direct approach might involve manipulating the degree to which the presented metaphor actually affords semantic linkages with the other literal statements contained in the communication. A metaphor that semantically ‘fits’ the literal statements should produce a coherence effect, whereas a metaphor that is irrelevant to the literal arguments might be expected to elicit a distraction effect. Indirect effects mediated by the processing of literal arguments in the communication: metaphor influences process style Just as metaphors may influence the direction of elaboration, metaphorical language may also influence the amount of elaboration. In doing so, metaphors can elicit a process style Metaphor and Persuasive Communication 789 a 2010 The Authors Social and Personality Psychology Compass 4/9 (2010): 783–794, 10.1111/j.1751-9004.2010.00292.x Social and Personality Psychology Compass a 2010 Blackwell Publishing Ltd effect (Ottati et al., 1999). Inspired by dual-process models of persuasion (Chaiken, 1980; Petty & Cacioppo, 1986), Ottati et al. (1999) posit that metaphors can influence the degree to which message recipients engage in central or systematic processing. Central (i.e., systematic) processing occurs when the message recipient carefully attends to, scrutinizes, and elaborates on communication arguments. It is marked by increased generation of message-relevant cognitive responses, decreased generation of irrelevant cognitive responses, as well as a more complex and organized representation of the attitude object (Chaiken, 1980; Gernsbacher, 1991; Graesser, Singer, & Trabasso, 1994; Petty & Cacioppo, 1986; Trabasso & Magliano, 1996; Woike, 1994). Importantly, central processing increases sensitivity to argument strength when listeners’ evaluate the communication topic (Chaiken, 1980; Eagly & Chaiken, 1993; Petty & Cacioppo, 1986). The tendency for strong arguments to produce more persuasion than weak arguments is accentuated when individuals engage in central processing, and reduced when central processing is absent. A metaphor influences ‘process style’ when it influences the message recipient’s motivation or ability to engage in central processing (Ottati et al., 1999). When this occurs, inclusion (as opposed to exclusion) of the metaphor in the communication moderates the effect of argument strength on attitudes toward the communication topic. Importantly, an effect of metaphor on ‘process style’ is not equivalent to the previously described effects of metaphor on ‘direction of elaboration’. The ‘process style’ models presume that the direction of elaboration is predetermined by the strength of the literal communication arguments. That is, strong literal arguments are assumed to elicit proarguments whereas weak literal arguments are assumed to elicit counter-arguments. A metaphor that promotes central processing should therefore increase the generation of pro-arguments if the literal message arguments are strong but increase counter-arguments if they are weak. A metaphor that reduces central processing should decrease pro-arguments if the literal message arguments are strong but decrease counter-arguments if they are weak. Of course, a metaphor cannot increase central processing if the message recipient is already engaging in extremely high elaboration. Likewise, a metaphor cannot decrease central processing if the recipient is already engaging in extremely low elaboration. Therefore, the effect of metaphor on process style should primarily emerge when elaboration likelihood is initially at a moderate level (see Petty & Cacioppo, 1986 for a related discussion). Metaphor, interest, and motivation to engage in central processing. Metaphorical language has the potential to impact the motivation to centrally process. As noted previously, many researchers have proposed that metaphorical language increases interest in a communication (Billow, 1977; Graesser et al., 1989; McGuire, 2000; Mio, 1993; Ortony et al., 1978; Paivio, 1979; Zashin & Chapman, 1974), and this claim has been empirically confirmed (Deiger & Ottati, 2002; Kreuz & Roberts, 1993; see Read et al., 1990, for related evidence). If metaphor increases interest in a communication, it might motivate central processing of literal arguments contained in the communication. That is, inclusion (as opposed to exclusion) of metaphorical content in a communication might increase the generation of relevant cognitive responses and increase the effect of literal argument strength on the recipient’s attitude toward the communication topic. As noted previously, however, it is possible that metaphorical content can increase or decrease interest in a communication, depending on the extent to which it resonates with the recipient’s personal preferences and interests. On the basis of these considerations, Ottati et al. (1999) developed the motivational resonance model. According to this model, metaphor increases central processing among 790 Metaphor and Persuasive Communication a 2010 The Authors Social and Personality Psychology Compass 4/9 (2010): 783–794, 10.1111/j.1751-9004.2010.00292.x Social and Personality Psychology Compass a 2010 Blackwell Publishing Ltd message recipients only if it contains content that is of personal interest to the message recipient. If the metaphor contains content that alienates or disinterests the message recipient, it will decrease central processing. In accordance with this conceptualization, Ottati et al. (1999) found that a sports metaphor increased central processing among sports enthusiasts, resulting in an increase in cognitive elaboration and a magnified effect of argument strength on attitude toward the communication topic. However, the same sports metaphor decreased central processing among listeners who dislike sports, reducing elaboration and reducing the magnitude of the argument strength effect on attitudes (see McGuire, 2000; McQuarrie & Mick, 1999; for related conceptualizations). Metaphor, coherence, and ability to engage in central processing. A possibility that has yet to be fully explored is that metaphor influences the message recipient’s ability to engage in central processing. As noted previously, some theory and research suggests that a metaphor can increase the coherence of a communication, providing an overarching schematic framework that guides interpretation, facilitates organization, and facilitates central processing of literal statements (Billow, 1977; Mio, 1993, 1996; Read et al., 1990). On the other hand, some theory and research suggest that metaphors can also confuse or distract the listener, and thereby interfere with the listener’s ability to process literal communication arguments (Billow, 1977; Frey & Eagly, 1993; Mio, 1996; Mio & Lovrich, 1998). Research has yet to clearly document when metaphor will elicit a ‘coherence effect’ and when metaphor will elicit a ‘distraction effect’. As previously noted, however, we suspect that a critical moderating variable involves the degree to which the metaphorical utterance shares semantic linkages with the literal statements contained in the communication. If the metaphor activates a schematic structure that affords semantic linkages with the literal arguments, it should increase the cognitive coherence of the communication and thereby facilitate central processing of the literal arguments. This should result in a more organized representation of the communication topic, increased cognitive elaboration, and an amplified effect of argument strength on attitudes toward the communication topic. On the contrary, when the metaphor is semantically irrelevant to the literal arguments, it should function as a distracter that fails to elicit (or even reduces) central processing. This ‘process style’ hypothesis predicts that coherent metaphors will magnify the effect of literal argument strength, whereas distracting metaphors will reduce the effect of literal argument strength. Thus, coherent metaphors should increase the persuasiveness of strong arguments but decrease the persuasiveness of weak arguments. Conversely, distracting metaphors should decrease the persuasiveness of strong arguments but increase the persuasiveness of weak arguments. Clearly, these ‘process style’ effects differ from effects predicted by the ‘direction of elaboration’ models described earlier.