Every presidential campaign season, candidates seem to vary their language as they try to persuade audiences to perceive them favorably. Many politicians do not seem to know when to express emotional language and when to restrain themselves. This may be because we, as their message recipients, do not seem to know what we prefer either. For example, during presidential primaries, the more impassioned language usually draws more attention and more forceful applause at debates; however, in the primary voting booths, voters tend to whittle down the candidates to a broadly palatable, less rhetorically extreme pick for the nomination (Ritter 1980).
Whether we prefer low- or high-intensity language during presidential campaigns probably depends on the situation. Presidents govern within the confines of situations, therefore their strategies for persuading must operate with the situation in mind (Edwards 2009). We can rhetorically analyze presidential elections post hoc, deriving anecdotal evidence from situations in which the loser presumably should have employed different language choices. However, our understanding of when a presidential candidate should use different language choices has largely eluded empirical study.
Such lack of clarity concerning the expectations of a politician's language intensity can receive insight via language expectancy theory (LET; Burgoon, Denning, and Roberts 2002) and Edwards' (2009) theory of presidential influence based on contexts. In this experiment, we vary the language intensity and circumstances of a hypothetical presidential candidate campaigning for the White House. We explore whether low- or high-intensity language serves politicians better during stable or exigent economic conditions. Because we extend the theories specifically to a presidential context, language choices should indicate whether the candidate is perceived as being more or less "presidential." Also, because the above theories both concern persuasion, the candidate's language choices should have a measurable impact on particular outcomes relevant to the candidate's trustworthiness (Hamilton 1997).
In line with theories of persuasion applied to presidential campaign contexts, we find that the effects of language intensity and circumstances each depend on the other. Significant interactions arise between language intensity and economic conditions, and they moderate people's perceptions of presidential candidates. This study helps elucidate our understanding of perceptions of political messages, specifically the synergistic effects of presidential candidates varying their language in different contexts.
Language intensity is most commonly defined as "the quality of language which indicates the degree to which the speaker's attitude toward a concept deviates from neutrality" (Bowers 1963, 345). For example, when Herbert Hoover called the New Deal "Fascism," "despotism," and "the poisoning of Americanism" (Sundquist 1983, 301), he was using high-intensity language on the presidential campaign trail. Language intensity deviates from neutrality in two linguistic ways: (1) directness toward the audience and (2) emotionalism of word choice (Bradac, Bowers, and Courtright 1979). According to Hamilton (1998, 12), "Most researchers have manipulated language intensity as a combination of language specificity and nonobscene emotional intensity" (see also Hamilton and Hunter 1998). High-intensity, direct language includes personalized, specific, assertive messages explicitly directed at the audience. Low-intensity, indirect messages are more ambiguous, unclear, and imprecise. Nonobscene, emotional intensity involves a speaker using extremity in word choice, through exaggerations or inflated adjectives. For example, in a high-intensity message of both emotionalism and directness, a politician might say: "This election is the most important election of your life." And as an example of a low-intensity message of indirect, unemotional language, a politician might say: "This election presents a choice between two contrasting visions for our country."
Communication experiments from the past 50 years have demonstrated a persuasive effect when a source's message varies in language intensity (e.g., Bowers 1963; Buller et al. 2000; Burgoon, Jones, and Stewart 1975; Hamilton, Hunter, and Burgoon 1990; Miller et al. 2013). Although, as with most communication phenomena, the effects of language intensity seem to depend on accompanying variables. Early experiments with language intensity indicated conflicting results. On a basic level, according to the Yale School, the more a speaker advocated change, the more the audience complied (Hovland and Pritzker 1957). However, in a study of high school students listening to lectures about dental hygiene, the low-intensity message triggered more conformity because--as Hovland, Janis, and Kelley (1953) conjectured--the high-intensity message stirred too much emotional tension. The Yale School opined that the persuasiveness of language intensity probably depended on other moderating factors, such as personality predispositions (Janis 1954) or the memorability of the message (Hovland and Weiss 1951).
Bowers (1964) took up the charge to discern the moderators of language intensity. His studies varied language intensity and source characteristics as independent variables. He expected to find that the more a hypothetical politician used high-intensity language, the more the politician's image would be enhanced. But he found the opposite. "The best post-mortem explanation," Bowers (1963) wrote, "is that the extremities in word choice produced a boomerang effect" (351). He concluded that language intensity "must be regarded as a complex variable which is subject to interactions with many other variables" (352, italics added for emphasis).
For decades the most evident interaction appeared to be source characteristics moderating language intensity's persuasiveness. Most of those treatments focused on the source's sex (Burgoon, Denning, and Roberts 2002). For example, research indicated that males and highly credible people were granted more latitude to use high-intensity language (Burgoon et al. 2002).
Language Expectancy Theory
Language intensity and other research exploring the suasory effects of linguistic message features inspired a formal theory of persuasion: LET (Burgoon, Jones, and Stewart 1975; Burgoon and Miller 1985). LET provides an explanatory and predictive framework for how language intensity and expectations "interact to enhance or inhibit persuasion effects" (Burgoon et al. 2002, 121, italics added for emphasis). Similar to other persuasion theories with latitudes of acceptance (e.g., expectancy violations theory, social judgment theory), LET posits that message receivers have a normative bandwidth or continuum, in which persuasion may occur depending on the intensity of the message interacting with other factors. Propositions of LET are based on societal norms that audience members bring into situations. If a speaker negatively violates an audience member's expectations (e.g., a female uses aggressive language deemed inappropriate for the situation, or a low-credibility source uses an aggressive fear appeal) attitude change is inhibited (Burgoon et al. 2002). Positive violations can occur as well. For example, if the speaker is a member of a minority group and his audience expects him to give a high-intensity attack message but he uses moderate language instead, a positive violation will occur and audience members may be persuaded by the perceived reasonableness of the argument (Burgoon and Miller 1985).
LET has most prominently been extended to the field of health communication. Language intensity enhanced the effects of argument styles in sun protection messages (Buller, Borland, and Burgoon 1998). Buller et al. (2000) found that physicians should use high-intensity messages to more persuasively encourage parents to put sunscreen on their children. LET has also been extended to online settings in e-commerce. When an online reviewer provided two-sided arguments highlighting positive and negative features of a product, the positive expectancy violations enhanced perceptions of the review's credibility (Jensen, Averbeck, Zhang, and Wright 2013).
Despite LET originally being proposed as a launching pad for interactions of suasory effects (Burgoon et al. 2002), few researchers have explored interaction effects via LET. Hamilton, Hunter, and Burgoon (1990) described three-way interactions involving language intensity, source gender, and credibility, as well as language intensity, source gender, and receiver anxiety (247). Additionally, Averbeck (2010) found that ironic messages mitigated criticism relative to literal messages delivering compliments. Most recently, Averbeck and Miller (2014) demonstrated LET's versatility by examining interactions between cognitive complexity and lexically complex or simple messages, and Averbeck (2015) reported interactions between expectancy violations of a physician's controlling language and patients' anger and behavior change.
Various parts of LET have received peripheral support in political communication research, without the researchers testing for interactions between language intensity and expectancies. For example, voters' expectations of a political message, contrasted with a politician's actual message, have been tested in the political sphere (Pfau 1987; Pfau and Burgoon 1988; Yawn, Ellsworth, Beatty, and Kahn 1998). These studies indicated that a message sender perceived highly in source credibility would be expected to use high-intensity language, while a source not perceived as authoritative and credible should use low-intensity language to achieve attitude change. In a study applying the theory to political campaign ads, Pfau, Parrott, and Lindquist (1992) stated that messages will be "more or less persuasive,...