Thursday, September 8, 2011

Is this definition of terrorism analytically useless?

Under the heading "20 years of terror," The Economist has a graph mapping global "terrorism" deaths since 1991. The one-year high topped out at nearly 13,000; the trough was around 3,000. The most lethal single incident over that period, of course, was 9/11, but the 2,996 dead were nearly half of an annual total near the 20-year mean. Have a look:

What's most interesting to me, though, is the definition used by the National Consortium for the Study of Terrorism and Responses to Terrorism (!), who compiled the data: for them, terrorism is "the use of illegal force and violence by a non-state actor to attain a political, economic, religious, or social goal."

On its face, that's a little goofy; after all, economic, religious, and social goals are fundamentally political, too. (Presumably a "religious goal" means something roughly like an attempt to advance the social or political status of one's sect, not the performance of some divinely-mandated act of violence, right?) But worse than that, doesn't this definition allow for the inclusion of deaths attributable to insurgency and rebellion? Or organized crime? Maybe I'm just picking semantic nits here, but as the lexicographers say, words mean something.

What's the most useful and accurate definition of terrorism you've seen? Maybe you've got one of your own...?

(Hat-tip to Peter J. Munson on Twitter for the link.)


  1. Gulliver,

    Always a provocative question. Personally, I follow Schmid/Jongman's 1988 definition of terrorism in their volume "Political Terrorism". I still believe it is the best single volume on terrorism, its definition, its role, and our political/legal response. He found over 109 definitions... in 1988. It hasn't gotten better. I digress. Here's their definition:

    "Terrorism is an anxiety-inspiring method of repeated violent action, employed by (semi-) clandestine individual, group or state actors, for idiosyncratic, criminal or political reasons, whereby - in contrast to assassination - the direct targets of violence are not the main targets. The immediate human victims of violence are generally chosen randomly (targets of opportunity) or selectively (representative or symbolic targets) from a target population, and serve as message generators. Threat- and violence-based communication processes between terrorist (organization), (imperilled) victims, and main targets are used to manipulate the main target (audience(s)), turning it into a target of terror, a target of demands, or a target of attention, depending on whether intimidation, coercion, or propaganda is primarily sought".

    The length of this definition speaks not only to the difficulty of pinning down the nuanced perceptions and involved in describing our innate judgement of what is 'terrorism', it speaks to the difficulty of navigating the emotion so omnipresent in the terrorism debate.

    Given this, I think we should generally start over. I devoted one of my first posts to the question of whether we need a neologism for terrorism. A professor of mine argued that perhaps academic discussions of definitions are innately about power, but I like to dream that we can parcel, begin anew, or simply reset, the terrorism debate with new rhetoric. Link to my post below. Apologies if its a bit raw, it was one of my first.

  2. For what it's worth, I like what Conor Gearty has to say on the subject. His is a non-exhaustive definition, but it's a good start:

    Violence is unequivocally terrorist when it is politically motivated and carried out by sub-state groups; when its victims are chosen at random; and when the purpose behind the violence is to communicate a message to a wider audience.

  3. Thomas (@WThomasWebb)September 8, 2011 at 11:01 PM

    Definitely concise and well-formed. The challenge that comes up again and again is the murky strings one can draw from almost any terrorism definition: Immunity of state sponsorship? Definition of “wider” audience, or audience in itself? How does one denote randomized victims instead of civilian/military dichotomy? I haven't read Gearty's before, but I agree its one of the better starts I've ever seen.

  4. To placate Gulliver's desire for comments in lieu of tweets:

    "Deliberate use of violence against civilians in pursuit of political goals."

    This clearly includes uses of state terror; for specific analysis of terrorism by non-state actors one need only stipulate so. But within a empirical (ie, not normative) framework, the phenomenon clearly includes firebombing Tokyo and Dresden.

    On the twitters we discussed whether violence against civilians has to be random/indiscriminate in order to be "terror." I think this is a great unexplored question. I think all parties make some use of wide-area and targeted attacks as part of their broader "terror" campaign, but definitely need more thought here.

  5. As you say, the definition bleeds over into insurgency, rebellion, and organized crime. When we look at the numbers on the chart at first glance, I'm sure we're not thinking these things are included. But to take some recent examples, is the suicide bombing of a line of potential police recruits perpetrated in the midst of an insurgency terrorism? The bombing of a shrine that kills people? What about marauding sectarian death squads that kill people in their homes? These sound a lot like terrorism, but could also be seen as insurgent action. Insurgents ambushing a government army patrol does not sound as much like insurgent action. Over to organized crime. Kidnapping someone of the opposite sect in the midst of an insurgency could be terrorism, if it is meant to incite terror and further a political goal, or is it to be considered separately as resulting from an insurgency, or is it wholly criminally motivated? Some kidnappings are done in the insecure environment offered by an insurgency solely for money. Then what if those organized criminals willingly or under threat pay a kickback to the insurgents? The point is that words do matter and the numbers in the chart may be a little misleading, but I don't know how to disentangle terror, rebellion, and crime.

  6. Thanks for the comments, all.

    *Commenting function going a bit nutty on me again so I'm not logged in, but I promise it's me!*

    Peter -- You've outlined my exact thoughts when I first read the definition, and I agree wholeheartedly with your conclusion about disentangling. I think Gearty's statement takes a run at it by including a randomness/indiscriminate component and mandating a broader message. But I have another concern about this one: hasn't he just equated terrorism with propaganda of/by the deed? Like I said, Gearty's list isn't exhaustive: I think he'd also assert that actually terrorizing people is a necessary component of terrorism, but that's sort of a sub-element of the random victims thing, right?

    Charlie -- I'm going to come back to some of the things I mentioned on Twitter last night when I have a few more minutes. But for now, I suppose I ought ask why people think the definition is significant? (Or maybe you don't.) I'd argue that there's a certain level of emotion associated with the label that allows for demonization of political opponents, etc., and I worry that such an expansive definition as yours isn't sensitive to that fact.

    Thomas -- Provocative idea about "retiring" the word. But: never gonna happen. I suppose you could say the same thing about getting popular consensus (or even expert consensus) on one definition, but I think it's an important effort just the same. The Schmid/Jongman definition is pretty comprehensive and has a lot in common with Gearty's, I think (though not everything, as it's a bit more expansive); I think it's interesting that a lot of the stuff done pre-2001 on terrorism (Gearty's book was from '91) is the most helpful to me as it was written outside the context of this allegedly "new" threat.

  7. "Deliberate use of violence against civilians in pursuit of political goals."

    Such an expansive definition would encompass not only Tokyo and Dresden, but atrocities in East Pakistan, Cambodia, Guatemala, Rwanda, the Balkans, and Sudan; the Great Leap Forward, Stalinist purges, the Armenian Genocide, the March to the Sea... In other words, it's seems too broad to be useful since it includes such wildly different phenomena.

    One problem with Geraty's definition is the word 'random'. Assuming we're distinguishing between terrorism and assassinations, the individuals targeted for attack may not be specified, but the population group certainly is. Whether by nationality, ethnicity, patronage of a certain location or business (restaurant in Bali; Marriott in Islamabad), or something else, there's always some logic that bounds the target population. I suppose Aum Shinrikyo's sarin attack might be an exception, but their aims aren't very coherent to begin with.

  8. The original Economist article includes a link to the GTD's Data Collection Methodology, which clearly acknowledges the existence of divergent definitions of terrorism. It's also worth noting that the second phase of the GTD included criteria that "were constructed to allow analysts and scholars flexibility in applying various definitions of terrorism to meet different operational needs."