Language abstraction as a strategic tool used by political leaders to convince different types of audience

Extensive research has shown that politicians are particularly skilled at using rhetorical devices to persuade an audience of whatever they wished it to believe. To this aim, politicians recognize the importance of appearing similar to their audience in order to be liked by its member. However, modern media coverage permits the actual audience of a political speech to be much broader than the immediate one. As a result, politicians must often reveal their views to audiences with attitudes and opinions that may be at odds with one another. In the present study, we argue that a specific structural property of language, namely the level of abstraction of interpersonal terms, is a subtle tool used by politicians to compose effective message for convincing different type of audiences. Specifically, when politicians’ opinions match the audience’s they may be concerned with reinforcing this common identity and motivated to fix shared believes by using abstract language. Conversely, when the audience has heterogeneous political positions, the most effective linguistic tactic may be using concrete language, one that furnishes details that will render the speaker accountable and that reduces the perceived generalizability of the opinion expressed. In order to demonstrate that language abstraction is a tool used by politicians to obtain the agreement of the audience, we examined the natural language of political leaders. The analysis of natural language allowed us to capture real-life conditions in which political leaders are aware of talking to their supporters or to audiences with attitudes and opinions that may be at odds with one another. For example, events of this kind are respectively represented by party conventions and by politicians’ participation in press conferences or radio/television programs. Specifically, we expected that political leaders preferred abstract language when speaking to an audience of similar ideological orientation and concrete language when speaking to an audience of heterogeneous ideological orientations. To test this general hypothesis, we used the linguistic category model (LCM; Semin & Fiedler, 1988; 1989), which operationalized the distinction between concrete and abstract language. This distinction can be easily explained by the following example. Imagine that person A is moving fast his/her foot till touching violently person B's leg. According to the LCM, this event may be described using four different linguistic categories that correspond to four levels of abstraction but refer to same semantic domain. For instance, at the most concrete level, the event can be objectively described as “A kicks B”, namely using a descriptive action verb (DAV); at a slightly more abstract level the same event can be described with an interpretative action verb (IAV) by interpreting the action as “A hurts B”; alternatively, one may refer to an enduring psychological state of person A and use a state verb (SV) to utter that “A does not like B”; finally, at the highest level of abstraction, one can employ an adjective (ADJ) as “A is aggressive” to refer to the enduring personal characteristics of person A. According to the LCM, aabstract statements, as opposed to concrete ones, are perceived as revealing more about the person and less about the situation, imply greater temporal stability and are more likely to produce expectations of being repeated in the future. Overall, these properties imply that abstract language has the general function of consolidating and maintaining knowledge and beliefs. In contrast, concrete language is better suited to communicate descriptions of persons and events that are situationally and temporally bounded. Moreover, messages composed in concrete language are less likely to cause disagreement since observable features of events readily validate the references of concrete messages. The LCM has been applied to code for language abstraction 67 speeches by well-known Italian political leaders (36 left-wing, 31 right- wing). These speeches were transcribed from video and audio recordings accessible on the Internet and chosen in equal distributions from politicians’ personal and political party websites, television channel, and radio stations. Speeches made during party conventions or party-organized events, with an audience of party activists, were assigned to the similar audience condition. Speeches made during meetings organized by groups out with the political arena, press conferences or radio/television programs, where the audience is politically heterogeneous, were assigned to the mixed audience condition. Results supported the hypothesis, showing that politicians used more abstract language when they faced a similar audience than when they faced a mixed one. Moreover, variations of language abstraction are more pronounced for left-wing than right-wing politicians. These findings provide evidence on how, not only the content, but also a structural property of language, namely abstraction of interpersonal terms, is really used by politicians to persuade different types of audiences. Implications for political communication and a broader perspective of language as socially motivated action will be discussed