Given the recent civil unrest in the United States, where after three months of rioting, protesters and counter-protesters have begun shooting each other on the streets with at least three dead, I started wondering about what the definition of a civil war was. The Wikipedia page on civil wars lists a few criteria that certain academics have used to classify civil wars, ranging from "100 people dead" to "1000 people dead per year of the conflict", along with the definition used by the International Committee of the Red Cross, which is a bit more in-depth and doesn't mention a number of casualties at all.

This lead to me wondering if there was an official definition of a civil war in either United States law or official policy, or in international law in general. At what point would intrastate violence be considered a civil war?

  • 31
    That whole debate about whether or not that one perpetrator can be considered "Antifa" and the other perpetrators can be considered "right-wing" does not help to improve this question at all and keeps escalating into petty name-calling. I deleted it and rephrased the question using the hopefully less controversial phrasing "protesters and counter-protesters".
    – Philipp
    Aug 31, 2020 at 14:41
  • 1
    Based on the title, I thought this question was asking what events typically happen for a violent civil unrest to escalate into civil war. Instead it seems the question is regarding the definition of civil war. Sep 2, 2020 at 0:36
  • And just to complicate matters, consider the situation where the civil unrest is actively instigated by agents of another country. Regardless of how bad it becomes, will it really be a civil war, or just an "ordinary" war, where the enemy country uses internal conflicts to weaken the other country? Sep 2, 2020 at 18:32

4 Answers 4


This is usually left for the historians to sort out.

The key point at which something is a "war" is when potentially lethal unprovoked violence is not treated as a criminal matter, but as a military matter.

Therefore, in the US civil war, nobody thought of prosecuting the union and confederate soldiers for "murder" even though they purposefully killed enemy soldiers. But before the war, actions like John Brown's raid on the Harper Ferry Armory were certainly treated as criminal.

There is also a rhetorical matter in choosing to call an event a "civil war" or a "revolution" or "rioting" or "terrorism". The IRA may have seen the period of the "Troubles" as a civil war, but the British government considered it to be criminal terrorism.

  • 5
    Good point on the difference between say, "Freedom Fighters" in the IRA and Mujaheddin and "Insurgents" in the Syrian Civil War
    – David M
    Aug 30, 2020 at 23:27
  • 19
    Could of course also ask if the American Civil War was a civil war - at least as we understand the term today (ie. "war" between groups of people inside a single country). The individual states were much more independent than today, the southern states had seceded and created a new "union" and country. Both sides had armies under command of a central, democratically elected, political authority. So it really was more like a war between two (groups of) nations, than the typical civil war. Aug 31, 2020 at 4:28
  • 3
    @BaardKopperud There's also the fact that the South didn't want control of the whole country, they wanted their own country. I'm not 100% sure this is a requirement, but every other civil war I can think of involved two or more factions fighting for control of one polity. The Spanish Civil War, the German Civil War, the English Civil War, in all of these nobody wanted to secede. They wanted all the marbles. Seems to me the American Civil War was more of a rebellion than a civil war.
    – Ryan_L
    Aug 31, 2020 at 16:22
  • 11
    @BaardKopperud That's a big stretch. If we say something is not a Civil war because both sides consider themselves to be their own states, then has there ever been a Civil War?
    – divibisan
    Aug 31, 2020 at 16:24
  • 4
    @BaardKopperud By that logic, e.g. the Russian Civil War wasn't a civil war either.
    – Alice
    Aug 31, 2020 at 16:32

The concept of 'civil war' has three necessary attributes:

  • Intention: There must be an explicit effort to enact regime change, either by overthrowing the current state or by carving out a new, sovereign state from the territory the current state claims
  • Structure and organization: The regime change effort must have some degree of military/political organization, such that it can coordinate action effectively over distance and time
  • Popular support: The regime change effort must have at least tacit support among a broad segment of the population in the current regime, so they can receive supplies, recruit combatants, and find safe-havens for logistical purposes

None of these attributes are sufficient in themselves to constitute civil war. For example, in the US context there are:

  • militia groups that have structure and want regime change, but lack popular support outside their immediate members
  • protest movements that have broad popular support and a loose structure through organizers and social media, but lack the intention of regime change
  • the occasional firebrand who demands regime change and has a significant popular following, but who does not organize or structure a cohesive military or political system

...but none of these rise to the level of civil war. The secession of the Southern states in the 1860s, by contrast, checked all three boxes, and thus validly constituted civil war.

Civil unrest doesn't become civil war. Civil war occurs when civil unrest is given focus and purpose by an organized alternative. Riots may occur during civil unrest, leading to destruction and even death, but riots do not spontaneously organize into military campaigns. Moreover, people who organized military campaigns do not want ordinary citizens wasting their violent tendencies in unstructured riots.

  • 1
    I think you might add a fourth condition, that the unrest needs to be geographically broad based. E.g. the current civil unrest in the US in happening only in few urban areas, which aren't of real interest to most of the population. Or the unrest in Hong Kong not affecting the rest of China, the Troubles in Northern Ireland not much affecting either the Irish Republic or the rest of Britain...
    – jamesqf
    Aug 30, 2020 at 18:17
  • 8
    @jamesqf: I'd implicitly wrapped that up the 'Popular Support' item. Physical geography isn't as important as political geography. To use your example, it doesn't really matter if the Hong King unrest reaches the other provinces; it has enough popular support within the HK region that it could (theoretically) support a regional civil war (a secessionist thing, rather than an attempt to unseat the current Chinese regime entirely) Aug 30, 2020 at 19:44
  • 1
    Comment to point 1. Often civil war is the result of a power vacuum - what happens when there is no (effective) government, so it may be a simple matter of which fraction should lead the country (rather than overthrowing a regime or creating a new country). And how would things like an independence war or a revolution or coup differs from a civil war with your definition? Aug 31, 2020 at 4:33
  • 1
    There is missing a condition here that civil war is by definition violent (or it's not a war). A fully peaceful movement aiming to overthrow a regime (see Belarus) does not constitute a civil war, yet meets your three criteria.
    – gerrit
    Aug 31, 2020 at 6:21
  • @BaardKopperud: #1 isn't generally called a civil war; depending on context it might be referred to as internecine conflict or simply a power struggle. Wars of independence and revolutions are different names (or perhaps different types) of civil war; the first is usually used for colonies or other loosely-held, semi-autonomous regions, while the second almost always seems to be used for lower-class uprisings against aristocracies. And coup d'état is usually reserved for a sudden, dramatic seizure of power. But I'm not sure how picky people are about using these words in practice... Aug 31, 2020 at 7:01

It was historically relatively easy to define an international war as there were two countries with clear command chains fighting out combat. While modern asymmetric conflicts convolute the issue, the general idea is still rather easy to see and point out: a foreign military acting as a military in combat.

By contrast, it is much harder to apply similarly clear definitions to potential civil wars – again, especially due to the asymmetric nature of many modern conflicts. The First Geneva convention does not use the term civil war, instead speaking of (in article 3):

armed conflict not of an international character occurring in the territory of one of the High Contracting Parties

One may argue that as soon as there are two parties of which one is raising weapons at the other this constitutes a case for article 3 to apply – but this very broad definition might also include the police apprehending a criminal group which would usually not be considered civil war by anyone. Accordingly, the International Committee of the Red Cross (ICRC) comments:

What is meant by “armed conflict not of an international character”? The expression is so general, so vague, that many of the delegations feared that it might be taken to cover any act committed by force of arms – any form of anarchy, rebellion, or even plain banditry. For example, if a handful of individuals were to rise in rebellion against the State and attack a police station, would that suffice to bring into being an armed conflict within the meaning of the Article? In order to reply to questions of this sort, it was suggested that the term “conflict” should be defined or – and this would come to the same thing – that a list should be given of a certain number of conditions on which the application of the Convention would depend. The idea was finally abandoned, and wisely so. Nevertheless, these different conditions, although in no way obligatory, constitute convenient criteria, and we therefore think it well to give a list drawn from the various amendments discussed; they are as follows:

  1. That the Party in revolt against the de jure Government possesses an organized military force, an authority responsible for its acts, acting within a determinate territory and having the means of respecting and ensuring respect for the Convention.

  2. That the legal Government is obliged to have recourse to the regular military forces against insurgents organized as military and in possession of a part of the national territory.

  3. (a) That the de jure Government has recognized the insurgents as belligerents; or
    (b) That it has claimed for itself the rights of a belligerent; or
    (c) That it has accorded the insurgents recognition as belligerents for the purposes only of the present Convention; or
    (d) That the dispute has been admitted to the agenda of the Security Council or the General Assembly of the United Nations as being a threat to international peace, a breach of the peace, or an act of aggression.

  4. (a) That the insurgents have an organization purporting to have the characteristics of a State.
    (b) That the insurgent civil authority exercises de facto authority over the population within a determinate portion of the national territory.
    (c) That the armed forces act under the direction of an organized authority and are prepared to observe the ordinary laws of war.
    (d) That the insurgent civil authority agrees to be bound by the provisions of the Convention.

(Slightly modified for typographic reasons and to allow for list formatting)

It is not fully clear to me whether the ICRC intended this list as a set of necessary conditions (i.e. all four must be fulfilled) or as a set of indicators (i.e. any one may be fulfilled). The document’s following sentence provides a clue by saying:

Does this mean that Article 3 is not applicable in cases where armed strife breaks out in a country, but does not fulfil any of the above conditions? [emphasis mine]

Thus, it seems valid to check whether any one of the four conditions applies to the current situation in the United States; if any one is met, the current conflict would be a civil war.

  1. ‘the Party in revolt against the de jure Government’ – in my opinion, this can be dismissed as neither side is attempting to wage war against the government and state as a whole. At best, one side intends to bring change within the constitutional provisions at the next election while the other intends to support the current leader of government.

  2. ‘against insurgents […] in possession of a part of the national territory’ – one might very broadly stretch this definition to include cases like Seattle’s Capitol Hill Autonomous Zone but note that in the part that I have left out it is required that these insurgents organise as a military. To the best of my knowledge, no group has taken control over part of US territory while being organised as a military so the second possible condition fails. In addition, CHAZ itself has been disbanded in early July.

  3. This set of conditions considers the view of the de jure government. (3c) and (3d) can be definitely excluded, I believe so can (3b). For (3a), President Trump has tweeted out the following on 15th June:

    Interesting how ANTIFA and other Far Left militant groups can take over a city without barely a wimpier from soft Do Nothing Democrat leadership, yet these same weak leaders become RADICAL when it comes to shutting down a state or city and its hard working, tax paying citizens!

    At the very least, this suggests that the current administration has considered at least part of the participants miliant (i.e. belligerent) as far back as June. This can be considered as an admission that a civil war is occurring. Arguably, however, the resulting actions of the US government do not fully support this interpretation.

  4. These four sub-points seem to apply to neither side (‘antifa’ and ‘right-wing militia groups’ in OP’s words) in the current situation as neither side – again, to the best of my knowledge – is exercising state-like control over parts of US territory or authority over its population and neither side has declared themselves as a conventional war force or bound to the Convention.

Therefore, applying the ICRC conditions while assuming that any single positive suffices, one should arrive at the conclusion that the US has been in a civil war situation for ten weeks as President Trump’s Tweet suffices to fulfill condition (3a).

However, this conclusion has severe limitations. The entire ICRC commentary is built on the idea that either there is an insurgence against the central government or alternatively that the central government is actively participating. On the other hand, the current events in the United States seem to suggest that these are conflicts fought out between various civilian or at best paramilitaristic groups, neither of which has the intent to overthrow the government by non-constitutional means or secede to form their own state from part of the central government’s claimed territory.

Indeed, a casual observer might well arrive at the conclusion that the central government is considering this mainly a police question – not unlike gang wars between various criminal groups –, that its main response is increased policing and that any military forces have been but support to the local police force on the ground. Considering this, one must fully reject the idea that the US be in or close to a civil war in the present situation.

To answer your title question: an unrest would be termed a civil war by extension of the conditions above, if the government is targeted by or targets the insurgents and if the conflict between the two features military or militaristic traits.

  • 8
    I don't understand how Therefore, applying the ICRC conditions one should arrive at the conclusion that the US has been in a civil war situation for ten weeks. follows from the list just before, where you conclude most of the conditions aren't (remotely) met.
    – gerrit
    Aug 31, 2020 at 6:42
  • 1
    @gerrit Because I interpret any single one of the conditions to be sufficient based on my understanding of the sentence following them. And further, condition 3a would be met (all four sub-conditions of condition 3 are connected by or), as the President on an official government channel declared some groups militant.
    – Jan
    Aug 31, 2020 at 6:45
  • 1
    @Jan I had the same reaction as gerrit. As written, it sounds as if you consider 3a, and after some discussion ultimately decide that it has not been clearly met. If your intent is to say that the definition has been met because of 3a, then you might want to say that explicitly in the "Therefore, applying..." sentence. Otherwise, great answer.
    – Nobody
    Sep 2, 2020 at 19:12

Organized groups striving for power

I believe that the threshold of a civil war is crossed when there are multiple competing governments (or government-like organized groups) at the same time who all aim to take power in the same country in a violent conflict.

Obviously, if there's no violent conflict, it is not a civil war.

If the scale of the conflict is low, it is not a civil war. For example if a community of 100 people rebels, fight against the army, and all of them get killed deaths, that's horrible but that's not war. The boundary is generally set at 1000 deaths in a year, or, conservatively, at least 100 deaths on each side, so if there's 500 insurgent deaths and 23 dead government soldiers, that's below the line.

If one of the two parties is not local, then it's not a civil war, that's normal war.

If a government is replaced by another government, but there's no stage where both old and new governments have de facto control of at least part of the country at the same time no matter how, that's not civil war, that's something else - a coup or a revolution or voluntary handover, depending on circumstances.

If one of the parties does not make a claim to leadership of the whole country but explicitly claims only part of it, then that would more accurately be named as a war of secession or independence, but that's usually treated as a type of civil war (e.g. American Civil War).

If a group is not well organized, if it does not have leadership, self-governance and some intent to actually govern whatever territories and people they (perhaps temporarily) control or will take control of, then that's not a civil war, but a riot or disturbance.

If the opposition is not well organized, that's also not a civil war - for example, if well organized paramilitaries are performing a mass slaughtering some minority, that's ethnic cleansing or genocide, but not civil war unless that minority is fighting back in an organized manner instead of as scattered individuals. If only one party is applying organized, systematic violence, it's not a civil war, that's generic application (assertion?) of sovereignty and monopoly on violence. War needs two or more opponents like that.

So if we look at the current USA situation, then that seems definitely not a civil war according to these criteria. If we consider "protesters" and "counter-protesters" (as described in the original question) as the hypothetical parties to such a war, then both of them (a) don't meet the criteria for organization and leadership; (b) aren't currently taking over governance of any area or population, but still acknowledge the power of the current country/state/city leadership even if they current leaders are their opponents.

The boundary would be crossed when you either have an alternative government or are trying to build an alternative government that does not acknowledge the competing government as legitimate, and that governance is violently contested. I would say that USA is in a civil war if either "protesters" or "counter-protesters" would assert control and governance over some territory and the other group (as an organized group!) or the current government would violently contest it with significant casualties.

The closest thing yet IMHO was the Capitol Hill Autonomous Zone - if there was a sufficiently large organized group that would try to violently retain control of it, then that would be a direction to civil war. But as long as it's just angry mobs without self-governance, occassional unorganized, unsystematic violence, limited numbers of dead and no attempt to violently enforce your policies and governance on others, it's far from a civil war. And let's hope it stays that way, as real civil wars are very, very horrible both for participants and bystanders.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .