Definitions of war and conflict typologies

Definitions of war

One can differentiate between quantitative and qualitative approaches to a scientific definition of war. Quantitative definitions of “war” require that the number of direct or indirect deaths caused by violent clashes crosses a certain threshold. The most known and influential approach was developed by David Singer and Melvin Small in the framework of the ‘Correlates of War (COW)’ project at Michigan University. This project attempts to assemble statistical data on wars that have been waged globally since 1816. According to this approach, a war is any violent conflict with at least 1,000 killed combatants per year. To exclude genocides and sporadic massacres from this definition, both parties to the conflict must have organised themselves to commit collective violence, or the party with the least combatants must have inflicted at least five percent of their own losses on the opponent.

The COW definition is quite controversial. Historian Spencer R. Weart for instance uses the criterion of 200 killed soldiers per year to classify a violent conflict as a war. Ted Gurr and Barbara Harff have a threshold of 100 deaths in their data base on failed states and add the criterion of at least 1,000 combatants per conflict party.

Other studies point to the difficulties resulting from the reference to an absolute number as threshold value. One must assume that the number of deaths as a result of violent conflicts is perceived and assessed differently depending on the size of the affected or involved population. Some researchers have therefore suggested not taking the absolute number of deaths as a critical benchmark, but rather their proportion of the total population of the affected country.

Most quantitative approaches, however, have not yet been able to compete with the popularity of the threshold of 1,000 killed soldiers defined by Small and Singer. The conflict data base developed by the University of Uppsala in Sweden offers a further development of the definition by COW. It does not limit itself to the losses of the regular armies but also includes civilian casualties of direct physical violence. It allocates violent conflicts to three different intensity levels. While its definition of “war” still follows the COW approach of more than 1,000 deaths per year, it additionally differentiates between “smaller armed conflicts” (at least 25 deaths per year but less than 1,000 deaths in the entire conflict), and “medium-sized armed conflicts” (more than 1,000 deaths in the entire conflict but less than 1,000 deaths in each single year) (cf. map layer number of victims of war). This quantitative differentiation of different levels of intensity is useful as it takes into consideration violent conflicts beyond the “war” threshold.

A definition of war that is based on empirically verifiable criteria (number of victims) has methodological advantages for a statistical-comparative analysis of violent conflicts—in particular against the background of long historic time frames. But it can also be highly problematic. Precise data on war deaths are extremely difficult to obtain as they are often distorted by warring parties. It is also difficult to understand why the threshold lies at exactly 1,000 deaths, and why a violent conflict with 999 deaths per year is not called a war. This arbitrariness seems all the more problematic when a war or the degree of destruction of a violent conflict is to be recognised solely by the number of deaths resulting from immediate physical violence. Neither social, economic and cultural impacts of military conflicts nor the victims of epidemics or famines, for instance, that directly result from infrastructure having been destroyed by wars are taken into consideration.

Finally, it remains questionable whether a war should indeed be purely defined by its direct effects and not by its characteristics, cause patterns or functional logic. This aspect generally finds more recognition in qualitative definitions of war. One such definition is used by the Working Group for Research on the Causes of War (AKUF) at Hamburg University. Inspired by a definition of the Hungarian peace researcher Istvan Kende (1917–1988), a “war” is a “violent mass conflict” when it has the following characteristics:

  1. "Two or more armed forces are participating in the fighting, at least one of which are regular fighting forces (military, paramilitary units, police) of the government;
  2. There must be a minimum of central control of the fighters on both sides, even if this manifests itself as organised armed defence or planned assaults (guerrilla operations, partisan war, etc.);
  3. Armed operations continue and are not occasional, spontaneous clashes; i.e. both sides operate according to a planned strategy regardless of whether fighting takes place on the territory of one or more societies and of how long they take."

Furthermore, the AKUF differentiates between wars and armed conflicts. The latter are “violent conflicts (…) that do not entirely fulfil the criteria of a war.”

Due to its comparatively degree of openness, the qualitative definition of war as the continuing and systematic use of collective physical violence between at least two organised groups has an advantage over quantitative approaches in that it focusses on the inner logic of violent action rather than on exogenous consequences. Still, the qualitative definition, favoured by AKUF, also has its weaknesses. On the one hand, due to the first criterion, it remains a state-centric perspective. The current manifestation of war as a collective use of violence beyond state actors and indeed borders is not taken into account. On the other hand, war—as in the quantitative definitions—is understood as a determinable status quo. Even though no clear criteria for an analytically valuable differentiation between wars and violent conflicts are given, the definition derived from Kende gives the impression that the phenomenon “war” is a distinctive state of societal interaction.

Typologies of war

How can one differentiate between the various types of war? In the literature, one can find at least two kinds of typologies:

One approach takes the object of the conflict and the goals of the conflicting parties as a criterion for differentiation. The Working Group for Research on the Causes of War (AKUF) at Hamburg University, for instance, differentiates between “anti-regime wars” (“wars about the overthrow of the ruling party or the change or maintenance of the political system”), “wars about autonomy- and secession” (“wars about a larger regional autonomy within the state or the secession from the state”), and “decolonisation wars” (“wars about the liberation from colonial rule”). Other approaches differentiate between “ethnic wars” and politically motivated “revolutionary” wars.

Such a typology is problematic as more than one cause often manifests itself in one individual violent conflict. Furthermore, the goals of the conflicting parties often tend to change in the course of violent action.

This is why a second approach does not differentiate violent conflicts by referring to their causes and goals but rather to the political status or group of the actors involved. Mostly it is about whether parties to the conflict are state or non-state actors. Until the end of the 1990s, only wars with at least one state actor were taken seriously. Consequently, one can differentiate between two different types of war:

  • Symmetric, intra-state wars, i.e. violent conflicts between two states;
  • Asymmetric wars between one state and one non-state party.

The second type, asymmetric war, can be divided into two sub-categories:

  • Intra-state violent conflict, i.e. wars between one non-state actor and a state within the existing state’s borders;
  • Extra-state or extra-systemic violent conflict between one non-state and one state actor outside of the existing state’s borders (as for instance in the war of Western NATO states against the Taliban in Afghanistan).

The Uppsala Conflict Data Program, for instance, follows this approach in that its conflict database not only records extra-state and extra-systemic, inter-state and intra-state violent conflict, but also internationalised intra-state conflicts, where a state, supported by assistance of other states, fights against a non-state actor within existing borders (cf. map layer wars and violent conflicts).

This second approach of a conflict typology is still incomplete and needs to be complemented. It is difficult to understand why at least one side of the conflicting parties needs to be a state actor. It is true that the state still plays an important, if not central, role in the global organisation and enforcement of violence. But the described removal of boundaries as is observable in many of today’s violent conflicts shows itself best in those conflicts in which both parties to the conflict are non-state actors. The German political scientist Sven Chojnacki subsumes these violent conflicts under the group of “sub-state wars”, which complements the other types of war. In total, Chojnacki defines four “core types of armed violence”:

  • Inter-state violent conflicts (between two or more states);
  • Intra-state violent conflicts (between state and non-state actors within existing borders);
  • Extra-state violent conflicts (between state and non-state actors beyond existing borders);
  • Sub-state violent conflicts (between non-state actors independent of existing borders).

Sources and further information:


Buchhandlungen bangen um die Buchpreisbindung

Data tables

For some select map layers, the information portal ‘War and Peace’ provides the user with all used data sets as tables.

More ...
Magnifying Glass in front of a Boston map

Country portraits

In the country reports, data and information are collected by country and put into tables that are used in the modules as a basis for maps and illustrations.

More ...
Compass with Mirror

Navigation and operation

The information and data of each module are primarily made available as selectable map layers and are complemented by texts and graphs. The map layers can be found on the right hand side and are listed according to themes and sub-themes.

More ...