Author: Anton Friedrich Fischer
Why Definitions Matter
In previous essays, I have used the terms radicalization, extremism, and terrorism — and, implicitly, also the concept of ideology that underlies all three — without pausing to clarify what precisely they stand for. This I want to do in this essay. The task is more than housekeeping: as Peels et al. (2025) argue, pervasive conceptual confusion surrounds these notions, with much research relying on tacit understandings and inconsistent usage. And, as I will show, the problem is not merely definitional. Vague or self-serving definitions can also invite abuse, particularly once they enter policy.
Before taking the four concepts in turn, two broader points should be made clear from the outset. The first is that the four terms do not rest on the same conceptual level: ideology names a substrate on which the other terms sit, extremism a type of belief, terrorism a method, and radicalization a process. Much confusion arises from treating them as if they were the same kind of concept. The second is that the four concepts are best taken in the following order: ideology, extremism, terrorism, radicalization. This order runs from the most basic to the most dependent: extremism is a specific kind of ideology, terrorism (on one account) is defined partly by the extremist beliefs behind it, and radicalization cannot be defined at all without borrowing its endpoint from extremism or terrorism.
Ideology: The Substrate Everyone Stands On
Ideology is the widest of the four concepts. In its broadest scholarly usage, an ideology is “a distinctive system of normative, semantic, and/or reputedly factual ideas, typically shared by members of groups or societies, which underpins their understandings of their political world and shapes their political behavior” (Leader Maynard 2014, p. 824). Two elements of this definition matter for what follows. First, ideologies are epistemically organizing, supplying “both an interpretation of the environment and a prescription as to how that environment should be structured” (Denzau and North 1994, p. 24, cited in Zmigrod 2022). Second, they are socially organizing: an ideology is shared within a group and marks its boundaries in some way or another (Leader Maynard 2014). Simply put, an ideology tells its adherents both what the world is like and (sometimes implicitly) who “we” are within it.
Defined this way, ideology reaches further than one might expect, both in how it works and in whom it covers. The first concerns how ideologies work. Holbrook and Horgan (2019) warn against an overly doctrinal and static conception: ideology is not a fixed body of tenets that individuals either “possess” or don’t. Ideology is more fluid, often transmitted through emotionally resonant cultural products — a song, a video, a meme — rather than through manifestos. One “does not need to ‘qualify’ with any level of ‘expertise’” to be affected by it (Holbrook & Horgan 2019, p.8). Ideology is thus better understood as a collective and emotional map than as a mastered textbook. The second concerns who holds an ideology. On a definition this wide, everyone operates with such a map. This deserves emphasis, because the term is often used pejoratively. One standing temptation is to treat ideology as what other people have, while one’s own commitments are just common sense.
Hall (1986) sums it up by stating that ideology spans interpretive frameworks at every level, from the “high, systematic, philosophical level [to] the level of casual, everyday, contradictory, common-sense explanation” (cited in Holbrook and Horgan 2019, p.9). What varies between people is therefore not whether they hold an ideology but how: namely, with what rigidity and with what attitude toward those holding a different map. This leads to the definition of extremism.
Extremism: A Belief Structure, Not the Distance from the Political Center
In common parlance, extremism is understood as a matter of distance from the political center: extremists are those whose beliefs sit far from what most of us think. This everyday intuition also finds its way into scholarly usage. But Peels et al. (2025) consider the definition defective and raise objections that are hard to escape. A measure of distance needs a fixed point, and here the fixed point is whatever the surrounding society currently happens to believe. Extremism then becomes a fact about the audience rather than about the movement or group itself.
The consequences of this common-sense definition are quite uncomfortable. The German Nazi movement of the 1920s to 1940s would have to count as extremist while it drew only a few percent of the vote and progressively less so as that share climbed, until, in power, it defined the standard against which everything else was measured. Thus, a criterion built on distance from the political center cannot reach a movement that has become popular, and it catches small unpopular groups automatically, whatever they believe or do. Applied by a government, this supplies a scholarly-sounding vocabulary for suppressing dissent from the norm, and this is exactly what some legal instruments have done.
Peels et al.’s (2025) alternative, grounded in Tajfel’s social identity theory (Tajfel 2010), defines extremism structurally rather than relatively: extremism is an ideological belief that one’s in-group can never be healthy or successful unless it is engaged in hostile action against an out-group. Four features of this definition deserve unpacking.
First, extremism is here a belief structure rather than a set of violent actions: the hostility the belief calls for can be segregation, discrimination, or harassment. Decoupling extremism from direct violence in this way allows extremist movements to be studied across their whole life cycle, as they move through violent and nonviolent phases.
Second, the definition is not norm relative. The reference point is not society’s center but the constructed out-group. Extremism can therefore become a society’s dominant, norm-setting force, as arguably in Nazi Germany, without ceasing to count as extremism.
Third, the definition is content-neutral: the extremist in-group can be religious, ethnic, racial, national, or class-based. On this account, extremism is what happens when an ideology hardens into the conviction that hostility toward the out-group is necessary. In the vocabulary of political theory this is an anti-pluralist belief structure: it denies, in principle, the legitimacy of enduring coexistence with an out-group.
Fourth, the definition sets a high threshold, which keeps it from covering ordinary group antipathy. It requires more than dislike: the claim is that the in-group cannot be healthy unless it acts against the out-group, with an end state in mind in which the in-group dominates and the out-group is permanently controlled or removed (Peels et al. 2025).
Terrorism: A Method Rather Than a Cause
The term terrorism is, if anything, even more contested than the term extremism. Peels et al. (2025) cite Louise Richardson’s remark that the only universally accepted attribute of the term is that it is pejorative, and Ganor (2002) reports a survey yielding 109 different definitions of terrorism. But the apparent chaos is overstated: most are variations on a theme, differing mainly in which feature an author emphasizes (Peels et al. 2025). These features correspond to three dimensions any definition of terrorism must address: the target, the actor, and the motive. I take them in turn, drawing on two well-developed definitions of terrorism, those of Boaz Ganor and John Horgan.
As said, on the target of terrorism, the two agree. But Ganor (2002) offers the fuller explanation. He extrapolates from the Geneva and Hague Conventions, which distinguish soldiers, who target combatants, from war criminals, who deliberately target civilians.
On this basis he defines terrorism as the deliberate use, or threat, of violence against civilians to attain political aims. Guerrilla warfare he defines as the deliberate use of violence against military and security targets for the same kinds of aims. The contrast is the whole point of the maneuver, which is aimed at defeating the relativist slogan that “one man’s terrorist is another man’s freedom fighter.” That slogan defines terrorism by its goals: if the goal is legitimate, the actor cannot be a terrorist. Ganor’s move is to define terrorism by its method instead. The aims of the terrorist and the guerrilla can be identical, what distinguishes them is whom they deliberately attack.
Since combatant status does not depend on the observer’s political standpoint, the definition achieves a form of objectivity: an organization that deliberately targets civilians is a terrorist organization “according to objective measures, and not according to the subjective perception of the definer” (Ganor 2002, p. 298). One should note, however, that the criterion doing that objective work is itself a normative commitment: the immunity of noncombatants is one of the oldest principles of the liberal law of war.
On the actor, the two concur as well. Both reserve the term for nonstate actors: deliberate state violence against civilians is already defined and proscribed under international law, as war crimes in wartime and crimes against humanity in peacetime, so the term terrorism adds nothing there. The gap it addresses is politically motivated violence by organizations and individuals (Ganor 2002).
The motive is where they finally part ways. Both require some political or ideological gate: neither counts every deliberate killing of civilians as terrorism. Both, for example, exclude grievance-driven mass shootings. The disagreement is over where that gate is exactly.
Ganor (2002) places it in the aim: an act is terrorism if it pursues a political objective, and he treats the ideological motive behind that objective as irrelevant. For Ganor, “political” is broad enough to absorb nationalist, revolutionary, or religious goals.
Horgan places it in the belief itself. He defines terrorism as the use or threat of violence by nonstate actors against noncombatants in pursuit of some social, political, religious, or other ideological goal, and for him that ideological component is constitutive: it is part of what terrorism is (Peels et al. 2025). On this view, ideology is the shared set of beliefs that gives the actor legitimacy, direction, and a way to justify the violence, and it is this component that ties the terroristic method directly to extremist belief. Horgan’s definition also mirrors how terrorism cases are actually investigated and prosecuted, where the ideological element is often exactly what has to be proven (Holbrook and Horgan 2019).
The price Horgan pays is that this definition requires interpreting an actor’s beliefs. Ganor’s motive-blind definition avoids that interpretive step and stays externally checkable, but at the cost of making ideology definitionally irrelevant. On Ganor’s account, the ideology is a label one may attach afterward, doing no work in deciding whether the act was terrorism in the first place. Each chooses a horn of a definitional dilemma. The radicalization section that follows shows that the two definitions do different jobs.
Radicalization: A Process That Borrows Its Endpoint
Radicalization is the only one of the four terms that names not something a person or group has or does, but a direction in which they move. Peels (2024) defines it as a temporal process, short or long, in which an individual, a group, or a wider social entity shifts from the moderate toward the extreme. Because it is defined by its endpoint, radicalization cannot be specified without a prior account of the extreme. This is where our definitions of extremism and terrorism come into play, since each supplies a candidate endpoint.
It is instructive to unpack this distinction on the level of policy first, because this is where the choice of endpoint has its most tangible consequences. Neumann (2013) shows that each candidate endpoint underwrites a different approach to counter-radicalization, each with a characteristic cost: the European and the Anglo-Saxon. The European approach reflects a continent whose democracies were destroyed from within by extremist movements in the interwar years, and so treats non-(yet-)violent extremism as a genuine threat. The Anglo-Saxon approach reflects the confidence of the United States and Britain in the robustness of their institutions and their strong free-speech traditions, and so intervenes when extremism turns into deviant behavior.
If the endpoint is extremist belief (extremism), we have the European approach, which treats extremist ideas as a threat even when they are non-violent. But because the target is now ideas rather than conduct, the line between subversion and legitimate dissent becomes movable, and a government can redraw it at will. The approach thus lends itself to overreach, in the wrong hands even to the suppression of opposition as such.
If the endpoint is violent behaviour (terrorism), we have the Anglo-Saxon approach, which intervenes only at the intention to break the law. That danger of overreach ebbs, but two costs replace it. The concept now surrenders everything before that threshold, including the non-violent phases through which extremist movements pass. And it pushes enforcement toward eliciting the very conduct it waits for, which is how Neumann (2013) reads the American sting operations.
The policy debate, however, only concerns which endpoint to choose. It leaves untouched a deeper, conceptual question: how the two endpoints, extremist belief (extremism) and violence against civilians (terrorism), relate to each other in the first place. And here the trouble is that they are only loosely coupled: whoever is radicalized in belief is not thereby radicalized in action. This is the problem of the few. Only a small share of those exposed to extremist ideology come to hold extremist beliefs, and only a small share of those go on to act (Peels 2024). Belief is thus neither sufficient nor necessary for violent involvement.
The key to this looseness is that radicalization does not operate at a single level. The process is not confined to individuals: groups and whole societies can radicalize. The West German left-wing terrorist group Rote Armee Fraktion moved from non-violent protest to terrorism within a few years, and Iran after 1979 radicalized as a state (Peels 2024).
This multi-level character is what dissolves the coupling problem. Cognitive radicalization does concern individuals, since each member holds a extremist belief personally. But because that belief derives from a shared group ideology, the process runs much the same across the members of an extremist group. And this is precisely why belief cannot separate the few who act (terrorism) from the many who don’t: within a movement whose members already share the ideology, the merely cognitively radicalized and the behaviourally radicalized hold the same convictions.
Holbrook and Horgan (2019) confirm this empirically. Individuals who carry out terroristic acts often have only rudimentary knowledge of the ideology ascribed to them. Many routes into violence run through friendship, status, identity, a need for significance. All of this fits the collective, emotional character of ideology described earlier: it is the map of a whole milieu, not a doctrine held by one person. Its explanatory power therefore lies at the level of movements rather than persons. This is Neumann’s (2013) point: the successive waves of modern terrorism (like Islamist or right-wing terrorism) take their names from belief systems, and the pattern explains why radicalization surges in particular settings and eras rather than others. Ideology answers why here, why now, not why this person.
The two levels just distinguished are what finally make the concept tractable. Radicalization looked messy because its two forms were run together, kept apart and anchored to the definitions of extremism and terrorism given above, each becomes clearer.
Cognitive radicalization, tied to Neumann’s European approach, is movement toward the belief that the in-group cannot be secure without hostile action against an out-group. Because that endpoint is extremism in the structural sense, rather than a position relative to the center, cognitive radicalization no longer means mere drift from whatever a society happens to believe at a given time. This reduces the objection that the concept can be used to persecute legitimate dissent, since the line can no longer be redrawn at will: what one moves toward is a specific belief, not an arbitrary distance from the mainstream.
Behavioural radicalization, tied to the Anglo-Saxon approach, attaches instead to terrorism as defined above, specifically to Ganor’s definition. Terrorism there is a method, violence against civilians, with motive deliberately set aside. Behavioural radicalization supplies the motive that the method leaves out: it is the individual process, within a group already cognitively radicalized, that carries a particular person from sharing the extremist conviction to employing the violent method. That the method brackets motive is not a defect here but the very feature that makes room for radicalization to work. Ganor’s thin definition of terrorism and behavioural radicalization are complementary, each supplying what the other omits.
This division of labour also dissolves the apparent rivalry between the two definitions of terrorism. Cognitive radicalization is common to the members of a group, behavioural radicalization is what distinguishes the individuals within it. The two rival definitions of terrorism map onto exactly this contrast. Horgan makes ideology constitutive, which ties terrorism upward, to the shared belief of the movement, and so to cognitive radicalization and the extremist substrate. Ganor sets motive aside, which leaves it open below, to be supplied at the individual level by behavioural radicalization. What looked like a contest between two definitions of terrorism turns out to be a description of the same structure at two different levels.
References
Ganor, B. (2002). Defining terrorism: Is one man’s terrorist another man’s freedom fighter? Police Practice and Research, 3(4), 287–304. https://doi.org/10.1080/1561426022000032060
Holbrook, D., & Horgan, J. (2019). Terrorism and ideology: Cracking the nut. Perspectives on Terrorism, 13(6), 2–15.
Leader Maynard, J. (2014). Rethinking the role of ideology in mass atrocities. Terrorism and Political Violence, 26(5), 821–841. https://doi.org/10.1080/09546553.2013.796934
Neumann, P. R. (2013). The trouble with radicalization. International Affairs, 89(4), 873–893. https://doi.org/10.1111/1468-2346.12049
Peels, R. (2024). Towards a fruitful concept of radicalisation: A synthesis. Journal of Contemporary European Studies, 32(3), 610–624. https://doi.org/10.1080/14782804.2023.2185594
Peels, R., Tietjen, R. R., Horgan, J., & Berger, J. M. (2025). Extremism, fanaticism, fundamentalism, terrorism: A conceptual map. In R. Peels & J. Horgan (Eds.), Conceptualizing extreme beliefs and behaviors: Definitions and relations (pp. 241–271). Oxford University Press. https://doi.org/10.1093/9780197760222.003.0014
Tajfel, H. (2010). Social identity and intergroup behaviour. In T. Postmes & N. R. Branscombe (Eds.), Rediscovering social identity (pp. 77–96). Psychology Press.
Zmigrod, L. (2022). A psychology of ideology: Unpacking the psychological structure of ideological thinking. Perspectives on Psychological Science, 17(4), 1072–1092. https://doi.org/10.1177/17456916211044140