This is the second in a series of guest posts by Stuart J. Kaufman of the University of Delaware. Stuart advances a long-running dispute with PTJ about whether “what goes on inside people’s heads” is relevant to social constructionism. PTJ doesn’t think so; Stuart disagrees. The first post can be found here. After the final post, we will make the entire piece available as a PDF — consider it our first true “working paper” publication.

Since each theory begins with a metatheoretical judgment about human nature, I think the place to start looking for insights is in psychology, which focuses on the empirical questions of how people actually think and feel under what circumstances, and what they tend to be inclined to do. For an example of how psychology can inform constructivism, let us return to Krebs and Jackson’s “Twisting Tongues and Twisting Arms,” which suggests a constructivism based on the notion that rhetoric operates as a sort of coercion. In this very creative piece, they lay out a model in which much of the action of politics comes in the form of rhetorical competitions in which competing forces try to frame issues in terms of societal values that favor their argument. One side wins if the other runs out of plausible responses to refute the implications of its opponent’s frame.

One gap in this argument is that the plausibility of arguments depends fundamentally on pre-existing “rhetorical commonplaces” familiar to the public audience, but in their empirical illustration Krebs and Jackson do nothing to show what the relevant rhetorical commonplaces were before the debate they analyze. In principle, constructivists can do this by sampling the discourse prior to any particular debate to get a sense of what those commonplaces are.

What constructivists cannot do, however, is measure how widely believed and strongly influential those commonplaces are with the relevance audience. This audience is always multiple—divided into subgroups by myriad cleavages. How is the constructivist to know which rhetorical commonplaces are the ones that most powerfully influence the relevant audiences, and therefore demonstrate the power of rhetorical jiujitsu? Krebs and Jackson do so by assumption, picking out one particular rhetorical commonplace in Israel—the notion that those who serve in the military have thereby earned equal rights—to explain why Druze Arabs, who do serve in the Israeli military, have been granted rights other Israeli Arabs have not.

The trouble with this argument is that, even if one retains a constructivist methodology, Krebs and Jackson fail to consider other discourses that may better explain the outcome. For example, perhaps the key point in Israeli discourse is not that the Druze have earned citizenship, but that they have proven their loyalty—their military service proves that they are not a security threat. Much Israeli discrimination against Arab citizens is justified on security grounds. How do we know that the more important reason for the outcome was not that the notion of earned citizenship was unanswerable, but that the notion of Druze as a threat was a non-starter?

The deeper problem is that even if Krebs and Jackson had considered both discursive effects, constructivism offers no way to assess which one was more important, if both were present and prominent. The only way to assess these competing hypotheses is to think more systematically about the interaction between discourse at the social level and attitudes and beliefs at the individual level. In other words, one must resort to the methods of sociological framing theory (e.g. Benford and Snow 2000) that Krebs and Jackson reject—examining the pre-existing beliefs, values, attitudes and perceptions of the audiences (including their perceptions of the credibility and other qualities of the leaders proposing alternative narratives) to explain why some rhetorical moves resonate with different audiences while others do not.

The study of pre-existing beliefs, values, norms, attitudes and perceptions, in turn, leads us back to the realm of political psychology. It is political psychologists who have studied these issues most carefully, and have come to some important conclusions about the power of different discourses with different audiences. One of the most important of these findings is the importance of emotional or symbolic predispositions in influencing attitudes. Some stark examples are in the work of Drew Westen (2007, pp. 107-8). For example, when a group of respondents were asked their views about whether President Clinton deserved to be impeached, 85% of the variance in their answers was predicted by their emotional feelings about the political parties, Clinton, infidelity and feminism as measured in those same respondents six to nine months earlier. When cognitive constraints were added to the model, they increased the explanatory power only to 88%. Obviously these respondents had been exposed to some combination of pro- and anti- impeachment discourses, but their responses varied with their symbolic predispositions.

The basis for my hypothesis about the role of security fears in Krebs and Jackson’s Israeli case comes from another strand of political psychology, the unfortunately named “terror management theory” (see., e.g., Greenberg et al. 1990; Cuillier et al. 2010). In a series of experiments, these scholars have shown that subconscious concerns about death systematically drive political opinions to the right, making respondents more respectful of their own national and religious values and symbols, more favorable to those who praise such values and symbols, more unfavorable toward those with different values of any sort, more punitive toward moral transgressors, more physically aggressive toward those who differ politically, and less concerned with incidental harm to innocents. In a particularly striking study, Cohen et al. (2005) found that respondents who were asked to think about death preferred George Bush over John Kerry by 45% to 20%, while respondents in the control condition preferred Kerry to Bush by 57% to 13%. If this pattern holds up in Israel, then it seems plausible that security arguments against Arab rights are more important than failure-to-serve arguments regarding Muslim and Christian Arabs. Therefore the lack of credibility of such arguments regarding Druze Arabs should similarly be more important than rights-for-service arguments.

The reason that systematic attention to audiences’ actual beliefs and values (as measured in survey research) is so important is that failure to do so makes it too easy for the analyst implicitly to impute his or her own values to the audience. For example, in a generally persuasive and well-executed study, Lobasz and Krebs (2007) show how Democrats were “rhetorically coerced” by the “war on terror” discourse into acquiescing in the Iraq invasion that many of them were uncomfortable with and later opposed. While this positive argument is persuasive as far as it goes, the counterfactual argument is not: the suggestion that the most promising alternative discourse would have been a “jeremiad” arguing that the 9/11 attacks were a reaction to American behavior, and that the U.S. should reform itself rather than launching a crusade in the Middle East.

Lobasz and Krebs, not inattentive to findings in political psychology, note that there are some psychological obstacles to acceptance of the “jeremiad” discursive mode, mentioning in particular the fundamental attribution error. However, they vastly underestimate those obstacles, in particular by overlooking the values widely embraced by conservatives and moderates but not liberals or leftists (Haidt 2012). Most important of these is the value of loyalty. The trouble with the jeremiad narrative is that it leaves the would-be Jeremiah vulnerable to the question: “Whose side are you on, ours or the terrorists’?”

The power of the “war on terror” narrative is further boosted by other psychological effects Lobasz and Krebs overlook. First, any “us against them” narrative draws its power from the ingroup-outgroup effect demonstrated by decades of experiments in the social identity theory tradition. Just making the ingroup-outgroup distinction salient leads to increased stereotyping of the outgroup and increased pressure for ingroup cohesion (adding to the power of the “whose side are you on” question). Second, the credibility of the “war on terror” justification for the Iraq war was enhanced by prejudice—both cognitive stereotypes of Arabs and emotional dislike for them—that was prominent among an important subset of the American population. Third, the terror management effect from the lingering fear of terrorism was simultaneously driving attitudes toward the right on issues of nationalism.

Finally, the jeremiad narrative lacked credibility on the issue of 9/11 itself: even if I believe that most Arabs dislike the U.S. for what it does, not what it is, that does not invalidate the logic of a war on terror. If I make that distinction, I must also make another: most Arabs were not involved in the 9/11 attacks, either. Those that were—the militants of al-Qaeda—were violent extremists who did need to be fought. The only plausible alternative to Bush’s War on Terror, then, was Obama’s later war on al-Qaeda. Many plausible discursive traditions were available to purse this argument against the Iraq war, most importantly the security discourse itself, perhaps stated frontier-style: “You’ve got the wrong man (Saddam) there, Sheriff. We can’t let the real culprit (bin Laden) get away with this”.

My argument, then, is that responsible theory-building requires that we build not only on the findings of those within our narrow academic niche, but much more widely beyond it. For the relationship between psychology and constructivism, there is a whole host of psychological mechanisms—in social identity theory, terror management theory, prejudice and ethnocentrism theory, cognitive dissonance theory, cognitive network theory, etc.—that provide important insights into which rhetorics are most likely to resonate with which audiences and in which conditions. Sociological framing theory additional insights regarding the importance of the credibility of the leader offering a particular frame or narrative, among other factors. All of these considerations widen the scope for agency in constructivist analysis, not only by identifying the psychological tools available for leaders to manipulate, but also by identifying the psychological resources available to audience members in deciding how to respond.