Justifying Self-defense against Assisting States: Conceptualizing Legal Consequences of Inter-State Assistance

Written by and

Cause for thought: Israel’s airstrikes directed against Iran and Syria

Israel has acknowledged to have repeatedly struck Iranian military targets in Syria. While confrontations occur frequently, the incident of January 21, 2019 has received  particular attention. Israeli guided missiles, apparently fired over Lebanese territory (UN Doc. S/PV.8449, p. 31f), hit Iranian military targets in Syria, also leading to personal and material damage of Syria. Israel invoked its right to self-defense, apparently reacting to Iran firing a surface-to-surface missile towards the Golan Heights on Sunday, January 20 from Syrian territory. Syria’s precise role in the Iranian action beyond this territorial link remains murky.

The problem: self-defense affecting assisting states

The Israeli claim to self-defense faces various legal questions (e.g. whether the attack meets the necessary threshold or whether annexed territories can be defended). This contribution does not aim to assess the Israeli claim, but shall use this example to shed light on one problem only: May the victim of an armed attack defend itself not only against the attacker state, but also against an “assisting” state?

Even if the use of force by the defending state (here Israel) against the attacking state (here Iran) is assumed to be justified by self-defense, it also forcefully infringes upon the territorial integrity of the assisting State (here Syria), as protected under Article 2(4) UNC, and warrants justification, too. The claim that strikes directed against an actor within the territory of another state are not a prohibited use of force against the territorial state has been repeatedly rebutted.

In fact, in the Security Council debate on the January incident, Syria labelled the Israeli strikes as “acts of aggression targeting the Syrian Arab Republic”, a “gross violation of international law” (S/PV.8449, p. 31f). Iran like Russia condemned the Israeli action, emphasizing the infringement of Syria’s sovereignty. Interestingly, Israel (unlike the USA or Germany) showed awareness of the problem by holding “the Syrian regime responsible for the missile that was launched against Israel from Syrian territory” (S/PV.8449, p. 8). The IDF added that “Syria paid the price for allowing Iran to conduct attacks from its soil.”

While the permissibility of self-defense against states supporting non-state actor violence is being extensively discussed, self-defense against states assisting another state has received little attention. In the non-state actor context, two legally distinct problems arise: first, whether Article 51 UNC requires an armed attack by a state (i.e. whether there is a right to self-defense at all) and, second, whether any and if so what involvement justifies the infringement of the assisting state’s sovereignty. The second question is often answered in light of the first. In the state context, however, there is no doubt that prima facie a right to self-defense against a state attack exists. Only the second question arises, entirely decoupled from the first. The reason for this duty to tolerate the use of force can only be sought in the “third” state’s “provision of assistance”.

“Assistance” is understood here as any action or inaction, short of armed force, that is de facto capable of facilitating another state’s use of force in international relations. “Assistance” can take many forms ranging from territorial support, be it deliberate, tolerating, unwilling or even involuntarily (e.g. through inability to prevent), to other economic or logistical support such as the provision of weapons.

At first sight, there seems little reason to treat the state-actor and the non-state-actor scenario distinctly. On the ground the nature of the attacker does not make a difference. However, states are states – a seal that is precisely denied to even the most powerful non-state actors. What may appear trivial increases – in light of greater structural capacities – the risk of escalation, and is also reflected in the normative stigma of an Article 2(4) UNC violation. Moreover, non-state actors necessarily sit within an “assisting” state. States have their own territory. Yet the theoretical possibility to direct the defense against this other territory (instead of affecting the “assisting” state) may in many cases be ruled out as ineffective and/or unproportionate. Furthermore, states can breach international law, opening doors to complicity and hence new normative assessments. Also, the ius contra bellum frequently distinguishes between (assistance to) states and non-state actors.

All this make it seem appropriate to free oneself from the established path known from the debate on non-state actors and to map out the relevant question for the inter-state context with an unbiased mind. So how can self-defense against assisting states be justified?

Point of departure: No absolute toleration-duty for all infringements necessary for the act of self-defense

Things would be relatively easy if self-defense covered any necessary corollary of any lawful act of self-defense against an attacking state. “Assisting” states would be obliged to tolerate any such force, irrespective of the form and extent of their assistance. Accordingly, for instance, even Lebanon, although it did no more than unsuccessfully preventing the passage through its airspace would have to tolerate Israeli means of self-defense. And in fact, Article 51 UNC refers to self-defense without specifying against whom, whenever an “armed attack occurs”. Arguably, otherwise an effective exercise of self-defense might be impossible.

However, such a “situational” interpretation is not easy to square with practice, and hence is to be rejected. States like here Syria, Russia and Iran usually protest against any kind of territorial intrusion, regardless of whether the action may be potentially justified with regard to the actually targeted state (so does e.g. also Lebanon (p. 30) protesting against Israeli aerial intrusion or here for earlier incidents). Also, states acting in self-defense normally ask the “assisting state” for permission, or provide additional justifications (see e.g. US on Cambodia discussed under 3) rather than just assume that the ability to infringe that state’s sovereignty is necessarily implicated in their right to self-defense. More importantly, it may be considered “a question of (legal) logic” that at least at the outset self-defense is a riposte not against any actor but only against states somehow responsible for the armed attack.

To justify self-defense against assisting states, three alternative conceptual legal avenues seem worth considering.

(1) The attacker state’s attack is considered an attack of the assisting state itself

First, through providing assistance, the attack might be attributed to the assisting state. The attack is then considered its own conduct and hence a “full” right of self-defense against the assisting state emerges. This avenue is frequently used for self-defense with non-state actors, but may also apply to the state-context.

The attribution of the attack is relatively uncontroversial when the requirements of Chapter II ARSIWA are fulfilled. For inter-state assistance, attribution may be achieved primarily through control or acknowledgement of the attack as its own (Articles 8, 11 ARSIWA) (see here on a related discussion on UK’s responsibility for involvement in US-action in Iraq 2003). Yet, many decisive forms of assistance, such as provision of territory, would arguably not be captured. And in light of Article 16 ARSIWA, which treats complicity explicitly distinct from attribution, it remains doubtful whether the relevant criteria can be expanded beyond the confinements of Chapter II ARSIWA to include general complicity, as suggested for non-state actors.

(2) The assisting state’s contribution constitutes an armed attack itself

Second, the assistance to an armed attack could be (artificially) understood as an armed attack itself, justifying force against the assisting state under Article 51 UNC. In practice, States adopt this approach, yet not generally, but for specific forms of assistance only.

For the inter-state context, the sole universally (at least explicitly) accepted principle is Article 3(f) Aggression Definition. Accordingly, the provision and permission to use territory for an act of aggression is classified as act of aggression itself.

But this raises questions. Depending on the circumstances, cannot any act of assistance (e.g. providing targeting intelligence) be as crucial as the permission to use one’s territory? In other words, is the rule articulated in Article 3(f) arbitrary? To illustrate: why should Iraq be only allowed to use force in self-defense against Jordanian military bases, but not against Italian forces providing targeting intelligence, both decisively assisting the Anglo-American intervention in Iraq 2003? A feasible, though not conclusive distinguishing feature may be the fact that in case of territorial support infringements of the territorial integrity are intrinsically linked to an effective exercise of self-defense against the attack. But why limit this then to the inter-state context, and chose different standards for non-state actors (Article 3(g))? A comprehensive and consistent approach classifying assistance as independent aggression/armed attack is hence still to develop.

In this light, it is interesting to see that Israel did neither adopt this approach nor invoked Article 3(f), and accused only Iran, but not Syria, of “a grave act of aggression”(S/PV.8449, p. 8).

(3) Assistance to another state’s armed attack as basis for self-defense? – Limited self-defense against a “derivative attacker”

Finally, self-defense against an assisting state could be justified by its illegal involvement in another state’s armed attack. The assistance’s connection with an armed attack might justify a duty to tolerate acts of self-defense. The duty would originate in the assisting state’s specific contribution to a conduct that reaches the threshold of an armed attack. Hence, the threshold for self-defense would not be lowered, as the act of assistance itself does not suffice. And as such, self-defense would be limited in scope, purpose, time and space: to acts that are necessary in the course of self-defense against the primary aggressor, and confined to the assistance itself.

This avenue appears to best reflect in legal terms the nuanced situation on the ground. It acknowledges that there is only one attack by one actor, receiving maybe decisive, but also not more than support by another actor. Unlike the first and second option, it does not artificially, and arguably in contravention to a “principle of fair labeling”, blur the line between perpetration and participation or double an armed attack.

What involvement is necessary to justify a duty to tolerate is however controversial. It is beyond this piece to answer this comprehensively. Yet, it is argued that in determining what assistance is sufficient, one may draw on established normative judgments comprised in rules governing assistance that when violated establish responsibility for the assisting state’s involvement in the armed attack.

To be clear: it is not argued here that any such determination always suffices to justify self-defense, nor that there may not remain a gap between the threshold of outlawed assistance and the threshold for triggering self-defense, as prominently advocated by the ICJ in the context of support to rebels.

The proposal is that such rules that protect the interests of the attacked state and that derive the assistance’s wrongfulness from the armed attack may indicate who is to bear the burden of the assisted armed attack – the assisting or the attacked state (which otherwise would be denied to effectively defend itself). Thereby, in the interest of assuring a coherent normative framework, they may constitute the basis for and inform subsequent state practice concerned with the dilemma. A divergence may require special justification.

For example, the regulation of permission to use territory is relatively unambiguous: Article 3(f) Aggression Definition strongly indicates towards a duty to tolerate acts of self-defense. It not only outlaws such involvement in an aggression, but makes clear that this (derivative) contribution to the attack suffices to equate the legal consequences with those of a direct aggression.

Indication from other norms regulating other forms of assistance may not be as clear. For instance, Article 16 ARSIWA prohibits the deliberate support to an aggression. Deriving the assisting State’s responsibility for its own conduct from the connection with the (wrongful) armed attack, (while remaining silent on self-defense itself), Article 16 ARSIWA may be taken as normative indicator that it is rather the assisting than the attacked state who shall bear the burden of an assisted armed attack. Yet, it is not conclusive.

Relevant rules could also point against a right of self-defense. For example, the hurdle to establish a duty to tolerate force in self-defense is arguably higher for a state being merely unable to prevent the attack. Not only does one face similar difficulties in the inter-state context as in the non-state actor support to establish that mere inability constitutes a violation of international law (possibly of the customary no-harm-rule or the law of neutrality). But the scenario in the inter-state context adds a layer of complexity: The attack could also violate the rights of the unable (and hence involuntarily assisting) state. To illustrate: if Lebanon’s airspace was used without permission and (constructive) knowledge about an attack, Lebanon arguably discharged any potential positive (due diligence) obligations. The factual contribution to the attack is in fact a violation of its rights by the very same attack. It is not argued that this normative evaluation excludes that Lebanon ultimately needs to tolerate acts of self-defense; but it sets the justification threshold higher.

Importantly, all these considerations may be no more than indicative theoretical underpinnings. In the end, it is for States to prove these indications right or wrong, and to make the step from viewing support as wrongful to requiring the assisting state to tolerate acts of self-defense. In pursuing this proposed avenue states however would not enter uncharted waters, as the US justification of its invasion of Cambodia in the context of the Vietnam War along those lines indicates.


In this light, the Israel-Syria-Iran-situation illustrates that the (specifics of) inter-state support requires more discussion than just the side notes which scholars and states alike dedicate the topic when dealing with non-state actor support. It is interesting to see Israel invoking “Syria’s responsibility” for what happens on its territory. Ambiguous as Israel remains, it at least indicates that the responsibility and the extent of involvement may – on a conceptual level – be the key to unravel the legal dilemma states face when another state’s protection may operate as a shield against the effective exercise of self-defense.


Print Friendly, PDF & Email

Leave a Comment

Comments for this post are closed


Federica Paddeu says

May 23, 2019

Dear Benjamin and Paula,

This is a very interesting question you raise and one which, as you note, is not often discussed in these terms. I very much enjoyed reading your thoughts on this and hope this will lead to a more extensive work on the issue!

I have a general observation which relates to the framework in which you assess the question. You speak throughout of a ‘duty to tolerate the act of self-defence’ which, I assume, you focus on because of the existence of a ‘right’ of self-defence. But to speak of a ‘duty to tolerate’ the act of self-defence implies that the right of self-defence is, in Hohfeldian terms, a claim-right. And this, i think, is at the very least doubtful both in theory and in practice. I will explain the issue by reference to the classic self-defence scenario of inter-State force between two States and then modify it to fit the scenario you discuss. For ease, I will use ‘victim’ State to refer to the State who has been the victim of an armed attack and is therefore entitled to act in self-defence.

In the Hohfeldian scheme, correlation exists between two parties and in respect of the same thing. State A has a claim-right that State B not-p, and State B has a duty to State A to not-p. For example, State A has the obligation not to use force and State B has the correlative right that A not use force against it: States A and B are the two parties, and the ‘thing’ is not using force. Correlatives are, really, just two ways of saying the same thing: to say that State A has a claim-right that State B not use force against it is the same thing as saying that State B has a duty not to use force against State A. I’m not sure you can analyse self-defence in this way. In your analysis, State A has a right to use force in self-defence, and State B has a duty to tolerate that force, namely a duty not to interfere with the self-defensive force. We have the two parties here (A and B) but the ‘thing’ is different for the right and for the obligation: in one case (the right) it is to use force in self-defence (so it is about A’s conduct); in the other (the duty) it is about not interfering with the self-defensive force (so it is about B’s conduct). I’m not saying the latter duty cannot exist - but it would not be a correlative of the right of self-defence. It would be the correlative of a right of A that B not engage in acts of interference with self-defence (here the ‘thing’ is B’s acts of interference: A has a right that B not do it, B has an obligation to A not to do it).

In practice, moreover, aggressor States often ‘interfere’ with the (alleged) self-defensive force. Namely, they respond to defensive force with (more or renewed) force. In doing so, they are not infringing the victim’s right of self-defence. This is not to say that they act lawfully; they act unlawfully, but as a result of the breach of other obligations such as the prohibition of force (the response would be either a continued, or a new, violation of the prohibition of force).

The reason why I raise this is that the choice of framework has an impact in the way you go about setting out your argument. If self-defence is a claim-right, then you are looking for a duty to tolerate the self-defensive acts. If, on the other hand, self-defence is a privilege/liberty then what you need to look for is its correlative: a ‘no-right’. If State A has a privilege to act in self-defence, then State B must have a no-right that A not act in self-defence. That is, you would need to find a justification for why the assisting State, by virtue of its assistance, is in a position of no-rights vis-a-vis the victim State.

This framework (of self-defence as a privilege/liberty) is compatible with deontic (of threshold deontic) theories of self-defence: by virtue of its armed attack, the aggressor State forfeits the legal protection afforded by the legal system to its rights such as the right to be free from force, the right to territorial sovereignty, etc, vis-a-vis the victim State (or, to be more precise, some of its rights - Article 21 of ARSIWA and its Commentary clarify that ‘obligations of total restrain’ remain in place).

Can something similar be said about the assisting State? Has the assisting State done something which warrants its forfeiture of the right to be free from force vis-a-vis the aggressor State? Arguments along these lines have been made in the literature about self-defence against non-state actors (eg: the breach of the due diligence obligation to respect the rights of other States in its territory would warrant the forfeiture of its right to be free from force and others vis-a-vis the victim State). I’m not sure these arguments are entirely successful in that context (as I have argued elsewhere: https://www.cambridge.org/core/journals/leiden-journal-of-international-law/article/use-of-force-against-nonstate-actors-and-the-circumstance-precluding-wrongfulness-of-selfdefence/3B607A56CCE90FEEA4CFC973FDC65991). For one, I think a good argument can be made that the only wrong that justifies the forfeiture of the right to be free from force is a violation of the prohibition of force. Indeed, no other violation of international law can be responded to by force since forcible countermeasures are now prohibited. Here, however, I would say that an argument could be made that ‘indirect’ use of force (à la Nicaragua) could be sufficient to warrant that forfeiture, thus paving the way for defensive force.

There’s no need for a reply - especially if you haven’t thought about it in these terms. I just thought I would raise it in case it is helpful in your investigation!

All the best,


Paula Fischer says

May 24, 2019

Dear Federica,

Thank you for sharing your very interesting thoughts. They really get to the heart to what is at issue here.

We believe that you are absolutely right about the nature of the attacker state's (what we call) "duty to tolerate". It is not a duty in the sense of a "conduct obligation" (mainly because it exists irrespective of whether the attacker state is in a position to assess the relevant circumstances). It is, as you say, simply a "no-right", that is, a forfeiture of the (claim-)right that the defending state not use force against it. To justify, why and under which circumstances the assisting state should have this no-right vis-à-vis the defending state is precisely the question that is so difficult to answer.

In our variant (3) where the armed attack by the defending state is neither attributable to the assisting state (variant 1) nor the contribution can be qualified an attack itself (variant 2), one is constantly on the verge of the reproach of allowing force in a situation below the Article 51 UNC requirement of an armed attack. (And as you say, a similar problem arises in the non-state-actor context). From a normative point of view, we agree with you that an argument could be made that the only wrong justifying such a no-right of the above-mentioned nature is a violation of Article 2 (4). However, even this warrants conceptual justification because Article 51 does not content itself with a "prohibited use of force" but requires an "armed attack". Our idea to connect existing normative values that outlaw the assisting state's contribution with the existing armed attack by the attacker state is simply a proposal to close this conceptual gap. (And that this attack exists beyond doubt is a relevant difference to the non-state actor context). It then becomes only a normative question (that in the end will be decided by state practice) which norm violations by the assisting state suffice to justify the forfeiture of its right to be free from force.

This issue is definitely not solved yet and it is worth to keep thinking about it. Thank you once again for your very helpful comment!

All the best,

Ben and Paula

Ayman Salama says

May 27, 2019

Dear Ben and Paula ,
really I would like to congratulate you on your boldness to address a unique and innovative aspect of self-defense practiced by UN member states . Still , I prefer to primarily focus on the conventional and restrictive approach of the whole theory that had been provided in the UN charter and eventually supported by International court of Justice in the Hague in its advisory opinion .
The wording of Article 51 of UN charter is simple and clear when it necessitates an armed force aggression practiced by a state against a UN member state .In the same vein the International Court of Justice in its advisory opinion in 2004 concerning the Israeli fence affirmed the above mentioned legal prerequisite of an armed aggression attributable to a state denying the Israeli claim that she was in a states of genuine and natural right provided in UN charter .Besides I can not ignore the controversy raised by many jurists towards the jurisprudence of the court judges by virtue of escalating development and surge of non-state actors in recent decades such armed militias and terrorist groups , on the contrary article 51 of UN charter reflects customary international law .The validity and referentiality of the non-binding opinion is still undeniable .

Backing to the cardinal article , I do not agree with many lawyers who claim that the obligation on the victim state to notify UN security about its already adopted military measures is a procedural duty on the ground claiming that victim states rarely adhere to this duty . It is inevitable to mention that the same article authorizes security council to to take at any time to respond to the aggressor state as it deems necessary to restore international peace and security .

I hope my comment would be appropriate to the context presented by the two excellent authors

Dr Ayman Salama .