Use of Force in Self-Defence to Recover Occupied Territory: When Is It Permissible?

Written by and

In a recent piece on Just Security, Tom Ruys and Felipe Rodríguez Silvestre argue that a state whose territory is unlawfully occupied by another state does not have the right to use of force in self-defence to recover the occupied territory. The post considered the position in relation to the recent armed conflict in Nagorno-Karabakh between Azerbaijan and Armenia and suggested that even if it is accepted that the region belongs to Azerbaijan, and is unlawfully occupied by Armenia, Azerbaijan will have lost any rights it may have had to act in self-defence because the status quo has lasted for a quarter of a century. Yesterday, in a post on this blog addressing a range of international legal issues regarding the conflict in Nagorno-Karabkh, Bernhard Knoll-Tudor and Daniel Mueller agreed with the position taken by Ruys and Silvestre arguing that ‘continued occupation cannot be equated with “continued attack” permitting the recourse to self-defence in line with Article 51 [of the UN Charter].’

Ruys and Silvestre set themselves the following question: ‘When part of a state’s territory is occupied by another state for a prolonged duration, can the former state still invoke the right of self-defence to justify military operations aimed at recovering its land? Put differently, can unlawful occupation be regarded as a continuing armed attack permitting recourse to self-defence at any given point in time—possibly years after occupation commenced?’

We think that the question is wrongly put—or at least incomplete. The point here is not whether an unlawful occupation constitutes a continuing armed attack, but whether any occupation that is the direct consequence of an armed attack constitutes a continuing armed attack. We argue below that such an occupation resulting from an armed attack on another state is indeed a continuing armed attack and that the attacked state does not lose its right to self-defence simply because of passage of time.

Distinguishing Between Settling Territorial Disputes by Force and the Use of Force in Self-Defence

First of all, one needs to draw the distinction between a territorial dispute and a situation of armed attack and occupation of territory. 

As Ruys and Silvestre observe, and of course we agree, it is trite law that territorial disputes cannot be resolved by recourse to force. This principle, which is outlined in the Friendly Relations Declaration (UNGA Res 2625) follows from the requirement in Article 2(3) of the Charter requiring that international disputes be settled by peaceful means, which itself follows from the prohibition of the use of force in Article 2(4) of the Charter. However, distinct from this is a situation of an armed attack, to which the right to use force in self-defence applies. Indeed, it would not be coherent for a principle which derives from Article 2(4) of the Charter to prevent the application of Article 51 which after all constitutes an exception to the prohibition in Article 2(4). To be fair, the authors, at the very end of their piece, and after having put much weight of the inadmissibility of force for the resolution of territorial disputes, recognise that ‘defining where self-defence stops and the prohibition of settling territorial disputes by force kicks in remains an extremely difficult exercise’ (our emphasis), and we agree. Yet it is precisely this exercise that needs to be undertaken.

The distinction that needs to be drawn then is between an outstanding territorial dispute where no force has yet been used by any of the disputing parties, and a situation where one party creates (or escalates) a territorial dispute by invading and occupying territory held by another state. While no force can be used by either party in the former instance, the latter instance is clearly one where an armed attack has taken place, and the right of self-defence is triggered.

To put it differently, it is one thing to invoke alleged title to territory in order to justify the use of force against another state (impermissible), and quite another to respond to an armed attack of another state that has led to the occupation of territory you previously held (without having occupied it through resort to force). In the former instance we have an attempt to settle a territorial dispute by force (which is impermissible). In the latter, we have an instance of the use of force in self-defence, even though it may still be possible that the title to  territory continues to be in dispute, and such dispute has to be resolved by peaceful means since no use of force can lead to annexation or otherwise lawful title to territory.

Continuing Armed Attacks

Second, we need to now discuss when, if ever, the right of self-defence ceases where one state has used force resulting in the occupation of (part of) the territory previously controlled by another state. Ruys and Silvestre seem to argue that the right of self-defence ceases at some (unclear) point in time when a status quo is established, eg when the territory has been occupied and administered peacefully and there is a prolonged absence of fighting. They do not tell us where that point in time is, but they seem to imply that a quarter of a century is long enough.

Ruys and Silvestre do mention Article 3(a) of the UN General Assembly Resolution on the Definition of Aggression (UNGA Res 3314), which goes directly against their view. It provides that ‘military occupation resulting from an invasion or attack’ constitutes aggression (our emphasis, and see further above our point on what is the right question to ask). They get around this rather significant obstacle to their argument by attempting to cast into doubt the relevance of the UN General Assembly Definition of Aggression for the concept of an ‘armed attack’ under Article 51 of the UN Charter. However, as we have written, the International Court of Justice has practically equated the notion of ‘aggression’ with the notion of ‘armed attack’ by relying on the UNGA Res 3314 Definition of Aggression in order to determine whether a particular use of force constitutes an armed attack in numerous instances, including in the Nicaragua case (para 195) the Armed Activities case (para 146).

Indeed, the Definition of Aggression makes it clear that occupation that results from an invasion or attack is aggression and must be one of a continuing character. And given that an invasion or attack that leads to occupation is certainly an armed attack within the meaning of Article 51 UNC, we fail to see how it simply stops being one if one were to just wait long enough. It is significant that state practice accepts, as it seems Ruys and Silvestre do too, that if there is a short of space of time between the initiation of the use of force that results in occupation and a response in self-defence then the right of self-defence persists. One can think of the UK’s response to the invasion and occupation of the Falkland Islands. If the right of self-defence exists in such cases, it can only because there continued to be an armed attack up to the point when the right of self-defence was being exercised. If it were to be accepted that the armed attack consists only of the initial act of invasion, and also accepted that self-defence is lawful in the sort of cases described above, this would open up the possibility that self-defence can be exercised even after an armed attack has ceased. The only justification for accepting self-defence in these cases is because the armed attack continues. But if it continues, what is it that would make it cease?

Ruys and Silvestre also make the point that Article 51 UNC states that there is a right of self-defence ‘if an armed attack occurs’, and thus the concept must be limited to a specific point in time. In the first place, the use of the word ‘occurs’ does not linguistically confine the attack to a specific point in time. It only means that such an attack must have occurred and presumably must be occurring. It does not set out how long it must occur for. The Charter says ‘if’ an armed attack occurs, and not ‘when’ it does. If it occurs, presumably it can occur for a rather long time. So, the ‘textual’ argument does not seem either compelling or dispositive here. In any case, in practice, armed attacks can occur for a lengthy period. One need only think of a case where after state A invades state B, fighting goes on within the territory of state B for several years. After a while, the tide turns and state B gets the upper hand and pushes state A back on to its own territory. At this point state B is now acting in self-defence on the territory of state A. Does the fact that the initial attack by state A began years ago and that it has been under attack for several years mean that state B no longer has the right of self-defence? One can think of the situation between Germany and the Soviet Union in the Second World War where it took a few years before the Soviet Union began to use any significant force on German territory after the latter’s invasion.

The Importance of the Necessity Criterion

We agree of course with Ruys and Silvestre that there must be some proximity between attack and defence, and that the lapse of time between the two cannot be extended indefinitely. However, this argument does not go to whether there is a continuing armed attack, but rather, as they themselves admit, to the issue of whether use of force in self-defence is necessary. So, let us try and disentangle these two issues.

First, there may be a situation where a lapse of time between an armed attack an attempt to respond is merely an indication that the use of force is no longer necessary to repel it. This will be the case where the armed attack has taken place, is complete or over, and there is nothing to repel anymore, absent any occupation. Second, if an armed attack leads to occupation of territory, then it is continuing for as long as the territory is under occupation. And if years pass between the initial attack and the use of force in self-defence, that may actually mean that the use of force in self-defence is necessary precisely because the state resorting to force in self-defence has no other means to bring the armed attack and occupation to an end (a similar point is made by Corten in this video). So, the passage of time may actually show that there is no other reasonable means of bringing the armed attack and occupation to an end, rendering the use of force in self-defence the ultima ratio—which is precisely the point of the necessity requirement.

A strong argument that Ruys and Silvestre make is based on the following passage from GA Res 2625: ‘Every State likewise has the duty to refrain from the threat or use of force to violate international lines of demarcation, such as armistice lines, established by or pursuant to an international agreement to which it is a party or which it is otherwise bound to respect’. The next sentence of the same paragraph of UNGA Res 2625 says, however: ‘Nothing in the foregoing shall be construed as prejudicing the positions of the parties concerned with regard to the status and effects of such lines under their special regimes or as affecting their temporary character’ (emphasis added).

These last words, ‘temporary character’ are, we think, significant. The aforementioned passages of UNGA Res 2625 should not be construed as referring to the principle of the peaceful settlement of (territorial) disputes, but rather as referring to the principle of necessity. In particular, the latter passage underlines that, contrary to what Ruys and Silvestre argue (time runs in favour of the aggressor, to simplify), time runs ‘against’ the aggressor, as there must be a temporal limit to lines of demarcation, if these are of a ‘temporary’ character. Surely when something lasts for a quarter of a century, then its ‘temporary’ character can be called into question. So, here is the necessity point which we think UNGA Res 2625 makes: when an armistice or demarcation line is agreed, it is no longer necessary to use force: the temporary armistice line provides time to seek other means to deal with the armed attack that has taken place. But when this armistice line is no longer ‘temporary’, rather it turns into status quo, then at some point it becomes necessary again to use force in self-defence, all other means to repel the armed attack having failed.

In sum, a use of force in self-defence will only be lawful where it responds to an armed attack which is ongoing, is necessary, and uses force that is proportionate. In cases of occupation that result from a use of force, the armed attack has not only occurred but has not ceased. Necessity is often taken to refer both to the immediacy of the response in self-defence and to the notion that the use of force is a last resort. Since prolonged occupation arising out of force is a continuing attack, any requirement relating to immediacy is met and the question then turns to the last resort element of necessity. While passage of time may, and hopefully will, open up other means for bringing the occupation to an end, the very passage of years may itself be an indication that the last resort requirement is met.

Print Friendly, PDF & Email

Leave a Comment

Comments for this post are closed


Marko Milanovic says

November 18, 2020

Hi both - a truly excellent piece, as is the one by Tom and Felipe. This seems to be one of those under-determined issues on which the legal argument will revolve greatly around policy preferences. So can I push you on that point directly? Are you comfortable with making an argument under which Ukraine would be entitled to try to recover Crimea from Russia by force, because there is no prospect of Russia returning Crimea to it voluntarily (not that it's likely to happen, but still)? Or that Syria, with the help of Iran, would be entitled to recover the control over the Golan Heights from Israel by force (somewhat more likely, e.g. from today: Because that does seem like the inescapable outcome of your reasoning .

Dapo Akande says

November 18, 2020

Hi Marko, Thanks for the question. Our response is that we wouldn't put the issue in terms of whether we are comfortable with this use of force or that but in terms of policy preferences we do not think that a rule that says if you stay long enough the other side loses their right to self-defence would be a good one. Such rule simply favours the aggressor and favours aggression.

It important to note the point we make about necessity. What we are saying that is that there is a continuing armed attack and the right to self defence is not lost simply because of the passage of time. The attacked state still needs to demonstrate that it is necessary to use force. However, it may be that passage of time actually points towards that necessity rather than against it. If that is actually correct on the facts then as a matter of legal right the action to retake should not be unlawful. Whether it is wise in any given situation is another matter entirely.

Constantine Antonopoulos says

November 18, 2020

This is a very interesting piece. However, self-defence is not a continuing situation. There must be immediate reaction to an armed attack. Otherwise it would render the prohibition of the use of force devoid of substance. In a situation as the Falklands, for instance, what matters is the immediate invcation of the right and the taking of measures to implement it. Once defensive force is used then an armed conflict ensues and recovery of occupied territory becomes a legitimate war objective for so long as hostilities are in progress. If they permanently cease and if this is accompanied by a peace process agreed by the parties in the context of international organization then the necessity of self-defence no longer exists. There has to be a "new" occurence of an armed attack in order to invoke the right of self-defence. The Yom Kippur War of 1973 is an example of the inadmissibility of invoking self-defence to recover occupied territory qua a continuation of the initial armed attack (in this case the Six Day War).

Marko Milanovic says

November 18, 2020

Hi Dapo,

Yes, I do get it that you don't want to necessarily get into the specifics of any given case. And it seems to me that the policy issue - which appears decisive here - is precisely with some kind of utilitarian calculus that both Antonios and you on the one hand and Tom and Felipe on the other make, explicitly or implicitly, as to what kind of rule would best maximize peace/minimize loss of human life. You are right to say that their version of the rule rewards (successful) aggression; but I think that they are also right in saying that your version of the rule incentivizes resort to force, and thus loss of human life, even in situations in which there is a relatively stable status quo, all to recover control over territory, as with thousands that have now died in the Nagorno Karabakh conflict. The problem I have with both positions is that it's hard to make that utilitarian calculus since it's impossible to say prospectively what version of the rule will save more lives/spare most suffering in the long run.

In that regard, I don't think necessity has much constraining power at all here, contrary to your comment. You are basically saying that resort to force becomes necessary if the status quo has not produced some kind of process during which the lawful sovereign can recover control over the territory peacefully. But that seems like SUCH a difficult judgment to make. Can one really say to Azerbaijan that they exhausted non-peaceful options, or that they failed to do so? Bearing in mind that these frozen conflicts can last decades, with bursts of negotiations, sometimes with glimmers of hope that they might produce something (e.g. there was a point in which Israel seriously considered returning the Golan), but most often with much disappointment. How exactly could one reliably say, aha, at this point the peaceful options were exhausted and self-defence became necessary? Couldn't one always object that the lawful sovereign should wait a bit more, hoping say for a change of government in their adversary? Couldn't one conversely always say that the lawful sovereign has waited long enough?

So what I'm trying to say is that the imponderability of that assessment is a good reason to favour the other option, protective of the status quo. But I'm happy to sit this one out on the fence!

Adil Haque says

November 18, 2020

Hello all,

Thanks for the post, and the conversation. I want to pick up on 2625.

As I read it, the “temporary character” of an international line of demarcation simply refers to the international agreement that establishes it. If the parties agree to a six-month armistice, then they must refrain from the use or threat of force across the armistice line for six months. If they agree to six years, or six decades, then the prohibition lasts that long instead.

However, I don’t think 2625 permits a State to agree to a six-month armistice, then decide after five months that negotiations are leading nowhere, and resume hostilities. They have to let the agreement run its course, unless the other side materially breaches first.

Put another way, the first sentence prohibits the use or threat of force across international lines of demarcation established by or pursuant to an international agreement. The second sentence clarifies that the prohibition maps onto the terms of the agreement, respecting any special regimes agreed upon, and enduring no longer than the agreement.

In some cases, there may be no agreement that the partially-occupied State is bound to respect. Then 2625 does not apply. But where there is such an agreement, that State cannot be the first to use or threaten force.

At any rate, that is my reading. Do let me know what I’m missing.

Stay well,


Evelyne Schmid says

November 19, 2020

Dear all, this is an interesting debate, thanks for your contributions on both blogs and in the commments. I share many of the concerns expressed in the previous comments by Marko, Constantine and Adil. Here is my question: don’t we need to include the collective security dimension in the analysis? Why? Because the temporal evaluation of a self-defence claim is directly related to the obligation of the attacked state to inform the Security Council and the idea that it is then upon the UNSC to restore international peace and security. Of course, we all know what the problems are with the collective security architecture but isn’t that a missing piece in the analysis?
Dapo worries that “[s]uch rule [the idea that self-defence after years of occupation is not an option] simply favours the aggressor and favours aggression”. This statement may be valid when we look at the empirical difficulties with collective security. But from a legal point of view - maybe an old-fashioned one - the assumption in the Charter is that the Security Council will deal with the breach of 2(4) by the aggressor from that point in time. In this view, time really runs ‘against’ the aggressor.
Tom Ruys and Felipe Rodríguez Silvestre write on Just Security that the following is a minority position but maybe it would indeed be preferable to view the problem ‘simply’ as a previous breach of international law leading to an occupation, an ongoing international armed conflict in which the territorial state tries to regain control over the occupied territory within its own territorial borders (potentially by breaching a ceasefire agreement), hence outside the framework of self-defence. But I admit that this view may – at least in practice - equally be a slippery slope because it reduces the justification requirements re jus ad bellum.

Paula Silfverstolpe says

November 19, 2020

Dear all,

I find the necessity and policy arguments interesting. There appears to different opinions among scholars whether occupation constitutes a continuing armed attack or not. At the same time, I would like to highlight that the demarcation line (“Line of Contact”) flows straight from the Bishkek Ceasefire Protocol of 1994 which effectively ended active hostilities back then, and which, from what I understand, is still in force (it has no end date). The OSCE is since 1994 tasked with monitoring the implementation of the ceasefire and reporting on any violation along the “Line of Contact” (demarcation lines).
On the issue of armistices, ceasefires and the use of force I refer to:
Christopher Greenwood in Fleck (ed.), The Handbook of International Humanitarian Law (2nd ed., 2008) (Chapter 2, p. 68):
“The changes in the law regarding resort to force brought about by the adoption of the UN Charter have had a particular effect on the right of the parties to resume hostilities after the conclusion of an armistice or ceasefire of indefinite duration. Whereas the law once admitted there was a general right to resume hostilities (Article 36 Hague Reg), today it would be a violation of Article 2(4) for a state to resume hostilities unless the behavior of the other party to the armistice or ceasefire amounted to an armed attack or the threat of an armed attack. Similarly, although under the traditional law the conclusion of an armistice did not prejudice the right of a party to exercise belligerent rights against shipping, such action would now be lawful only if it constituted a necessary and proportionate measure of self-defence.”

I guess the reference to the "behavior of the other party" is key here. Could it refer to the continuous occupation as a possible continued armed attack, or is it more forward looking necessitating a new armed attack for the resumption of hostilities? I tend to view it from the latter perspective.

Another aspect is the reaction by the international community following the eruption of the recent spate of violence. The international community including UNSCR, UNSG, EP, NATO, CoE and the Contact Group of the OSCE, have all strongly condemned the use of force by the parties (not restricted to IHL violations). The UNSC has called upon the parties to immediately end hostilities and resume peaceful negotiation’s “without preconditions”. Apart from Turkey and Pakistan I have not seen any other states expressing support for the interpretation that Azerbaijan can "revive" its right to use force in self-defence pursuant to article 51.

Adil Haque says

November 19, 2020

Hello again,

I tend to agree with Paula, that continuous occupation as such can't justify the use of force violating a demarcation line.

For one thing, such lines fall within one State's borders. The invading State will *very* often exercise effective control over the slice of territory on one side of the line. So it's hard to believe that the prohibition wasn't meant to persist in these cases.

For another thing, in these cases, the agreement presumably tolerates the presence of the invading forces. So it's hard to believe that the territorial State can break the agreement to end the occupation.

(I generally avoid incentive-based arguments, but it's worth asking whether Dapo and Antonios's view would make it harder for warring States to agree to an armistice in the first place.)

So I think Paula is correct that the prohibition persists unless there is a new armed attack (or imminent threat thereof).

Perhaps unsurprisingly, I also agree with Evelyne that it would be nice to bring the rest of Chapter VII into this discussion at some point :)

Be well,


Adil Haque says

November 19, 2020

Hello yet again,

After thinking about this a bit more, I think Dapo and Antonios could argue for an even narrower interpretation of 2625.

They could say that 2625 applies where there is no existing international boundary, but only a line of demarcation.

In contrast, they are concerned with cases in which an existing international boundary has been violated, and a resulting occupation remains ongoing. An armistice line within those existing international boundaries would not be protected under 2625.

I'm not sure that this narrow interpretation is correct. I've dipped into the negotiations (which are very interesting) but haven't found anything pointing definitively one way or the other. But I thought it worth putting to the group.

Take good care,


Liron A. Libman says

November 20, 2020

Thank you all for a thought-provoking post responding to another excellent post in just security and triggering very good comments, too.
I must admit, before reading it all, I would think the question is a "no-brainer" and use of force is explicitly prohibited by UNGA res 2625 obliging states to refrain from the threat or use of force “to violate international lines of demarcation, such as armistice lines.”
Well, I realize now it is more complicated, yet still find the piece by Tom and Felipe more convincing. Here are my three cents:
1. One cannot assume that the occupier is the aggressor. A state may find part of its territory occupied by the adversary after it has initiated an aggressive war. Therefore, the ex iniuria jus non oritur argument may not always support the state using force to retrieve territory.
2. Dapo and Antonios writes: "The distinction that needs to be drawn then is between an outstanding territorial dispute where no force has yet been used by any of the disputing parties, and a situation where one party creates (or escalates) a territorial dispute by invading and occupying territory held by another state". However, is it usually a clear and workable distinction? I think not. A territorial dispute usually exists because at some point in the past, state A took what state B think is its own territory. If we move further backwards in time we will probably find state B taking that territory from state A. Note that common article 2 of the Geneva conventions stipulate that an armed conflict exist even if an occupation meets no armed resistance. Are there many territorial disputes where none of the parties used force against the other in the past? Under the same logic I agree with Tom and Felipe that a distinction between "genuine" territorial dispute and other situations is futile and counterproductive.
3. An analogy to domestic law on self-help in cases of trespass may be useful: usually the owner's right to use reasonable force to evacuate the intruder is limited in time (in Israel it is 30 days). Of course, a counterargument will emphasize the difference between the domestic legal system and the international one: in domestic law there is an assumption that if you have the legal title, you may adjudicate it and finally, get the police to exercise a court order to evacuate the intruder. While the black-letter law of the charter seems to envision a similar role for the Security Council, therefore limiting self defense to the time period "until the Security Council has taken measures necessary to maintain international peace and security", in practice, there may be a widely acknowledged legal title that the SC does not enforce.
Thanks again!

Marcelo Kohen says

November 20, 2020

Maybe you can have a look at what I wrote more than 20 years ago:
in chapter 5. It is in French, but I suppose readers of EJIL:Talk!, as good international lawyers, also read languages other than English...
Best wishes,

Bertie Harrington-Wall says

November 29, 2020

If a state has a right to consent to the use of force on its own territory without that offending art 2(4) why can’t a state whose territory is occupied unlawfully still use or consent to force on that unlawfully occupied territory? Is it even an international use of force falling for justification in accordance with art 51 if the ousted state is not directing force beyond merely securing its own territory (eg targeting its own military facilities that have been taken over by the occupier in that territory?)
Alternatively, even if long term occupation is no longer considered an extant armed attack surely the ousted lawful sovereign govt still has lawful rights over the territory? If so, it must be lawful for it to try to re-enter that territory - doing so shouldn’t be regarded a violation of 2(4) and any resistance should be enough to formally trigger 51 (or anticipatory rights thereunder).

Bertie H-W

Fabián says

November 30, 2020

The agression over Malvinas was in 1833 and Argentina doesnt lost their rights to self defense and recovery of their territory . The negotiations with UK were fruitless . The resolution 502 was a mockery to the international system of law . No one can be judge and part in a conflict . This is a universal law landmark . The UK´s vote is a shame . Moreover they have the right to vetoe any resolution promulgated against them . That is not justice .