Extreme Ideas and Extreme Tactics?

By Joseph Young

What makes an extremist extreme? It is a conventional wisdom in studies of political violence to assume that extreme preferences on a policy outcome equal a desire for extreme measures. Is this true? If a person is strongly against abortion, are they more likely to commit acts of violence to stop it than someone who is moderately against abortion?

Max Abrahms of Johns Hopkins challenges this common knowledge. He agrees that extreme tactics by dissident groups leads to an audience perception that the group has extreme beliefs. In his estimation, however, this can be quite counterproductive. For Abrahms, if elites wish to persuade their domestic audiences that a group cannot be negotiated with, extreme dissident tactics can actually harden the resolve of the targeted government. For example, Hamas is perceived by the Israelis as a group that is extreme in their preferences, because of their use of suicide attacks. Are they? Since the Israeli-Palestinian conflict is maybe the most emotive and divisive conflict on Earth, the answer to this question is hotly debated. Instead, let’s consider a group like the Army of God in the US. While some people question their groupness, their goal is clear: ending legalized abortion. A recent Gallup poll suggests that about 1 out of 5 Americans agrees with this position, a rough percentage has remained stable since 2004. In a separate question, about half the respondents considered themselves pro-life. We can all probably agree that Eric Rudolph bombing the Atlanta Olympics or Paul Hill gunning down an abortion provider (and his bodyguard) in Pensacola, Florida are extreme tactics. But are their positions extreme?

As I mentioned in a previous post, political scientists are warming to the use of experimental methods to unpack these kinds of questions. Abrahms uses a survey experiment in a forthcoming piece in International Studies Quarterly to take on this assumption central to many recent bargaining models of conflict (including ones that I borrow from). Begun by Schelling around 1960 and refined by Fearon and others in the mid 1990s, bargaining theory is arguably one of the most productive research programs in International Relations explaining everything from internal ethnic conflict to interstate war. One of the key points of this approach is that violence lends credibility to threats in an anarchic world filled with uncertainty and asymmetric information.

While Abrahms agrees that violence can increase the credibility of a threat, he claims it can also undermine the credibility of a group’s/individual’s claim that they will refrain from violence in the future, what he calls the Credibility Paradox. The reason (causal mechanism) that violence can be counterproductive for groups is what Abrahms calls the Correspondence of Means and Ends bias. In short, as the extremeness of a challenger’s means increase, so too do the perceived extremeness of his/her ends particularly in the minds of citizens of the target country.

To test this, Abrahms presents subjects with a simple story about a group. The group has a relatively moderate demand: release some of their confederates from prison. In exchange, the group promises to lay down their weapons. Each subject then was assigned to one of two conditions. The first condition (the control), the group captures some American hostages but does not hurt them. The second condition (the treatment) is nearly identical except the group kills the hostages. In sum, each scenario is the same as a group with moderate demands takes hostages. They differ based on the extremity of their tactics.

Comparing respondent’s scores between the treatment and control group, Abrahms finds respondents exposed to the more extreme tactics condition were:

  • 20 percent more likely to rate the group’s preferences as the most extreme on a standard 7-point ordinal scale.
  • 23 percent more likely to believe the group would not demobilize upon achieving its demand to free the imprisoned leaders.

Since the end goal was the same, Abrahms attributes this disconnect to the Correspondence of Means and Ends bias.

Abrahms recent research is ambitious as it challenges the hegemonic bargaining models of conflict that make these assumptions. It brings psychology back into the center of conflict studies a la Ted Gurr. It is also part of his larger attempt to show that terrorism is generally associated with failure.

As an end user of bargaining theory, I, of course, have a stake in the argument. The easiest criticism is that is a single experiment, and the results don’t generalize to other populations. Since this was a US sample, the experiment could be tapping into US post 9/11 perceptions of terrorism rather than larger empirical patterns across countries and time periods. Additionally, a casual look at many groups that use terrorism suggests that they have fairly extreme positions on the issues that drive their violence. Al-Qaeda would like to establish a global Caliphate where Sharia is the law of the land. This is far from moderate. Hamas would like all of the land from river (the Jordan) to sea (the Mediterranean), not exactly a compromising position. The LTTE wanted their own state, so did the PKK, ditto ETA. The Army of God is an exception, but on average it seems like many of these groups have objectively extreme demands. These goals shift, but it seems there is a positive relationship between tactics and demands. With that said, Max Abrahms is pushing the boundaries of the debate and inspires me to want to do work that casts doubt on his claims, something I can’t say about most recent research.

  1. Since any extremist group would actually have to convince some real people to commit (and support the commission of) extremist acts, perhaps a more psychological framing would be more useful? Seems to me that the degree to which people will be prone to committing extremist acts is a function of how acutely they feel a particular pain and the amount they stand to lose by taking extreme actions (both directly and as a result of forsaking less extremist means). Put otherwise, one will only be willing to commit extremist acts if one faces an irredeemable status quo and sees no reason to believe that a satisfactory solution can be reached quickly enough through less extreme methods.

    Most reasonable people are willing to bargain for reasonable requests provided they can get something in return. So extreme actions only make sense if (i) the request is extreme or (ii) one side has nothing to offer the other (e.g. an already colonised population reasonably seeking freedom from their colonisers). Of course, if one has little to offer, one also has little to lose, giving a second incentive. On the flip side, once one has already forsaken goodwill through extremist actions, why not demand as much as one can. As a result, extreme tactics are always the best choice for those who *both* have extreme ideas *and* feel extremely aggrieved, but they are also sometimes the only tactics available to dispossessed groups with (initially) non-extreme demands.

  2. The first part, that to understand individual participation in extremism, a psychological approach can be useful, I totally agree with. There are a lot of reasons people use political violence (extreme tactics)–revenge, money, boredom, beliefs, etc. A simple rational approach assumes that violence is used by groups either to reach some goal or it is the easiest/cheapest/last option open to them. This can also be true. A group like Al-Qaeda may use violence for much different reasons, for example, then the person who is the human bomb or bombplanter.

    The aggrieved actor stories are limited though as aggrieved people are everywhere. It may be a necessary condition, but it is a trivial necessary condition. Some other factor(s) has/have to be involved to explain why aggrieved people in one location just complain while others do horribly violent things. The recent Sikh temple attack illustrates this. The shooter was frustrated with other folks with similar beliefs that wouldn’t act on them. As far as we know right now, his act was putting some of his belief into practice. Why though did his friends/compatriots/bandmates not use violence? I would argue we need a blending of psychology, sociology (group dynamics), and strategic theories.

  3. I’ll hold off until I read the whole paper before making more extensive comments, but my first impression is that Abrahams is introducing new language and suggesting a slightly different mechanism to explain a phenomenon we’ve known for quite some time and that applies not only to extremist groups but all strategic interaction (and that is neglected by rat choice models). I’m referring of course to the notion that actors come to attribute the behavior of the other to dispositional and not situational factors, and thus fail to distinguish between actors with different sets of preferences (greedy/offensive/revisionist/revolutionary/extremist/etc. vs. status quo/conservative/satisfied/moderate/etc.), and that these resulting enemy images can make the type of signalling and bargaining that rat choice models posit very difficult (the key works here being Jervis’ Logic of Images and Perception and Misperception).

    In any case, good to know about the paper and to see this discussion going on. Looking forward to reading the whole thing.

  4. Here is a link to the paper:


    Thank you again, Joe, for your analysis.

    Dani, I do not fully disagree with you. I challenge mainstream bargaining theory by showing that enhancing the credibility of a threat with violence can in reality be counterproductive for gaining strategic concessions due to what I call the “Credibility Paradox.” Under specified conditions, increasing the credibility of a threat with violence risks discrediting the promise, undermining the logic of bargaining. I label the micro-mechanism as “The Correpondence of Means and Ends Bias,” which posits that extreme tactics risk signaling extreme intentions, thereby closing off the apparent bargaining space. This mechanism does indeed have roots in the attribution literature within social psychology, particularly the work of Edward Jones.

    Max Abrahms

  5. None of the following seeks to discount Abrahms’ demonstration that promises of restraint may be considered less credible after increases in violence. However, couldn’t some of the following provide additional explanations for why an increase in terrorism reduces the likelihood of receiving a concession? I’m sure that most (if not all) of these suggestions have been thoroughly addressed elsewhere, but my familiarity with bargaining literature is quite limited, so my apologies in advance.

    1) The costs associated with concessions may escalate in the wake of an attack, even if the challenger’s offer of restraint remains credible. For example, granting concessions may erode a state’s ability to credibly utilize deterrence by denial against other challengers. By acquiescing to a terrorist’s request in the aftermath of an attack, the defender increases the probability of success associated with aggressive action and may encourage the use of similar tactics from future opponents. If so, that could result in higher cumulative costs associated with the strategy even if the original challenger exercised restraint as promised. Similarly, given the response of citizens to violent action, it may be more damaging politically for a state to offer concessions in the wake of an attack than it would be in the absence of terrorist activity. Finally, if an attack is a demonstration of one’s capability, then challengers may adopt more aggressive demands after an event. If so, that may decrease the likelihood of concessions, if the demands are relatively more expensive to the defender.

    2) The payoffs associated with alternative strategies may improve following an attack. For example, in the process of conducting the attack the terrorist organization may provide the state with information that either lowers the costs or improves the probability of success for retributive action, thereby improving its payoffs relative to conciliation. Likewise, it may be politically beneficial to deny concessions to terrorists.

    3) Psychologically, people may simply harbor an aversion to rewarding those who have engaged in violence. In the scenario offered by the paper a hostage-taking has already occurred, but suppose a terrorist organization merely demonstrated their capability to take hostages. Would demonstrating the capacity for terrorism but restraining from actually engaging in an attack (as in the 1995 Ismailovsky Park RDD threat, for example) be more or less effective in encouraging government concessions than actually acting on that capability? Such a tactic would not be as effective a demonstration of a challenger’s resolve, but could it convince the state of the challenger’s capabilities without engendering in the defender an attitude of hostility or resentment, thereby improving the net likelihood of receiving concessions?

    4) Policymakers may seek to deprive their political opponents of recognition or reward. Terrorists are more likely to be perceived as politically significant in the wake of an attack, which could decrease the likelihood that they receive concessions.

Leave a Reply

Your email address will not be published. Required fields are marked *

You May Also Like