A rejoinder to the incidental civilian harm debate, part I: the untenability of understanding incidental as ‘fortuitous’ killing of civilians (no matter how many) under IHL

[Luigi Daniele is Associate Professor of International Law at the University of Molise, Italy, where he recently moved after working as Lecturer and Senior Lecturer in IHL and ICL at Nottingham Law School (NTU)]

Last year, the Journal of Conflict and Security Law hosted an interesting debate on a surprisingly neglected topic, that is how to understand the notion of incidental civilian harm under international humanitarian law (IHL), and the implications of this understanding on how we construe the normative and operational relationship between the rules of distinction and proportionality in attacks, as well as their respective violations.

I have argued that the notion of incidental civilian harm, never as in the last years, has been unduly stretched in many discourses on the laws of war, with the consequence of radically expanding the scope of proportionality in attacks (to the detriment of the basic rule of distinction), up to the point of encompassing attacks launched with the declared objective of eliminating one, or a handful combatants, despite (and obscuring) the certainty of the attackers that the main outcome of their action, overwhelmingly predominant in comparison to the objectively marginal impact on lawful targets, would have inevitably been the mass killing of crowds of hundreds of civilians, children, and/or the mass destruction of protected objects. In my reading, subject to conditions discussed in the article, such attacks qualify rather as indiscriminate attacks violating distinction, at least in all cases in which the masses of protected persons and objects are treated de facto as extensions of the lawful targets they surround, because they are made direct object of the attack, jointly with the lawful targets.

Aurel Sari has written a dense and thorough article objecting to my analysis, and stimulating this rejoinder. He argues that incidental civilian harm under IHL would have no upper limits related to its magnitude/extensiveness, no quantitative relationship with the lawful targets expressly aimed at, and could be found even in attacks consciously and exclusively shelling civilians and civilian objects, with “no impact on the intended target” (Sari, p. 225).

In this post, while thanking Aurel for the thought-provoking engagement, I am going to summarize what appears unconvincing in the author’s counterproposal about how to understand the notion of incidental civilian harm, in the spirit of encouraging the collegial development of this unsettled debate.

The knots of genuine disagreement between the two analyses revolve around an example discussed in the initial article, that is the

bombing a high-rise tower building full of civilians because an enemy unit was located in its basement.

In the context of my discussion, it was abundantly clear this example referred to the then-ongoing and systematic targeting of high-rise residential towers in Gaza ultra-densely populated areas, with highly destructive artillery, without advance warning, knowing the towers had dozens of families inside at the moment of the attacks, and launched with the virtual certainty of killing and injuring all the civilian residents. Even without considering the context, the sentence made clear that, in the example, the object of the bombardment is, factually, the multi-storey tower, including its residents, as a whole, including its basement (not its basement per se).

Nonetheless, Sari contests that such an attack amounts to an indiscriminate attack violating distinction.

Sari’s initial argument is that my analysis:

“sits uneasily with the fact that the relevant rules in Article 51 of AP I are focused on the manner in which attacks are directed, not on their outcomes”.

(p. 219)

This appears disconnected from the substance of the dispute. We all agree that unforeseeable civilian harm or, conversely, unexpected failures of the relevant weapon system do not affect a legal assessment focused on the knowledge, expectation, calculus of the attackers when launching the attack. In my example, however, the outcomes are precisely those expectable. Nothing more, nothing less. There is no discrepancy between the manner in which the attack is directed (bombing a residential tower teeming with civilians) and its obvious, direct, and prevalent outcomes (an aerial civilian massacre). I do not find in the relevant digression, therefore, anything relevant to refute my perspective.

The author, after this detour, begins excluding in relation to the above example, one by one, every type of indiscriminate attack violating distinction codified in article 51(4) of the 1977 Additional Protocol I (AP I).

He states, first, that the example would not qualify as an attack that was “not directed at a specific military objective”, under Article 51(4)(a) AP I,

as long as it was directed at the enemy unit in the basement.

In Sari’s view, the attack must be conceived as directed at a specific military objective even if the impact of the attack on the lawful targets (the combatants in the basement) was foreseeable only as a temporally secondary and a quantitatively peripheral, indeed marginal effect of the attack itself in comparison to the calculated mass killing of protected persons and mass destruction of protected objects.

The author then proceeds to add that such an attack would not even qualify as an attack which employs a method or means of combat which “cannot be directed at a specific military objective”, under Article 51(4)(b) of AP I,

as long as it involved the use of a weapon capable of being directed at the enemy unit with some degree of accuracy.

I am not a weapons expert (see here what a renowned one had to say on the effects of the one-ton bombs used to flatten residential towers and tower blocks in Gaza), but I am unaware of the existence of weapons capable of accurately piercing through, say, 15 or 20 floors of a residential tower, leaving behind just a hole in the perforated flats’ ceilings and floors, killing and/or injuring a limited number of civilian residents, and then exploding when impacting the basement with a small-radius blast, expected to kill the few combatants located therein, without collapsing the tower itself on the heads of all its civilian residents.

Even if such weapons exist, turning to war realities surrounding us (in homage to the pragmatism evoked by Sari), we can verify that the flattening of the residential towers of Gaza has been carried out with wide-area effects projectiles directed at the towers themselves, often tearing them down from the roofs, or striking them with the stated objective (see this Amnesty’s report) of levelling them as such. As an example, among many, of this military policy and its intolerable costs for civilian lives, see this Airwars investigation on footage of aerial attacks on residential towers and densely populated neighborhoods shared by the IDF themselves as examples of military efficiency.

Not once in his response does the author make any reference to this reality, in which proportionality in attacks on residential towers has been mobilized to justify mass civilian harm unseen in aerial campaigns of the 21^st century, the decimation of an entire national group, and the complete desertification of its living space, with cities without a single stone left standing by bombs.

The author then admits (at least) that the attack in the initial example could violate article 51(4)(c) of AP I, as an attack

the effects of which cannot be limited as required by the Protocol.

But here, too, Sari argues that the limitation of the effects required by the Protocol, evoked in this part of article 51 AP I, circles back directly to proportionality and, therefore, invalidates my argument about the violation of distinction.

I find this reading unconvincing: it ignores the fact that the limitations required by the Protocol evoked by the prohibition under article 51(4)(c) AP I, before proportionality, include first and foremost the limitations imposed by distinction.

The subsequent sentence in article 51(4), above all, affirms, referring the three typologies of attacks prohibited in the paragraph

and consequently, in each such case, are of a nature to strike military objectives and civilians or civilian objects without distinction.

The message is clear: “without distinction”. It denotes the absence of distinction, not a merely insufficient degree of it, and cannot be interpreted as referring solely to lack of proportionality, i.e. excessiveness of incidental civilian harm, or other limiting rules (such as the protection of the environment, or the rules surrounding installations containing dangerous forces). This reading conflicts with the plain text of the article.

This dismissal is key to understanding Sari’s underlying effort, which appears aimed at circumventing this normatively explicit violation of distinction and ultimately at collapsing the IHL prohibition of indiscriminate attacks—or at least most of it—into the prohibition of disproportionate attacks.

On the contrary, the recurrence of indiscriminate attacks of the types discussed has been duly noted by many competent bodies, including the UN Office of the High Commissioner of Human Rights, which has dedicated to indiscriminate attacks in Gaza ample parts of a specific thematic report, noting inter alia

GBU-31, 32 and 39 […] are mostly large and heavy munitions used to penetrate through several floors of concrete and can completely collapse tall structures. Given how densely populated the areas targeted were, the use of such a wide area effect weapon would have in all likelihood resulted in an indiscriminate attack. Explosive weapons with such wide-area effects cannot be directed at a specific military object in densely populated areas of Gaza and the effects cannot be limited, resulting in military objects, civilians and civilian objects being struck without distinction

The author’s whole article, in contrast, comes down to one essential argument: incidental civilian harm, triggering proportionality assessments and implying that distinction was respected, under IHL should be construed as including all harm calculated and inflicted on civilians and/or objects, but not accompanied by a desire of the attackers of killing or injuring the relevant crowds of civilians, and/or damaging or destroying the relevant compounds of civilian objects, as such.

A key provision mobilized in support of this view by Sari is article 2 of the 1907 Hague Convention IX on Naval Bombardment in Times of War, actually its French translation, referring to ‘dommages involontaires’, which would include:

“succession of ranging shots adjusting their direction by correcting range and deflection until the target is found, eventually allowing the gun or battery to open fire for effect to destroy the target”.

In Sari’s opinion, this concept of involuntary harm, to be associated to the contemporary notion of incidental civilian harm, would cover:

“not only harm that an exploding shell may cause to nearby civilian persons and objects when it scores a direct hit on its intended target”,

but also:

“damage caused by projectiles that fall short, overshoot or otherwise fail to strike their desired point of impact, including projectiles that cause civilian harm directly and exclusively”.

In support of this reading, Sari cites a work by Morton William Royse of 1928, antecedent to the codification of both the Geneva Conventions and the Additional Protocols, tellingly stating:

“The Hague rules, permitting the shelling of military objectives in civilian areas, thus permit, in reality, a certain degree of indiscriminate bombardment.”

By means of this reference, among others, to the legal archaeology of naval warfare (in the era of military reconnaissance capable of identifying a pencil on a desk from a satellite), Sari construes incidental civilian harm as a category comprising all the harm not supported by the attackers’ desire to achieve it, including harm:

• knowingly inflicted on civilians, even if
• directly, even if
• exclusively (while ‘searching with projectiles’ the location of the lawful target), and
• irrespective of its predicted entity, including when the calculated mass killing and destruction of protected persons and objects is so massive as to render the impact of the attacks on lawful targets a consciously marginal outcome of the attack, quantitatively.

This is an exceptionally broad construction of the notion of incidental civilian harm, framed as negative of purposeful killing or destruction, and hardly distinguishable from a self-certification by the attackers claiming they did not ‘desire’ to kill the masses of civilians they have (in Sari’s own words) “knowingly and willingly” killed.

This in turn expands the scope of proportionality assessments to an exceptional degree, radically shrinking the prohibition of indiscriminate attacks violating distinction.

Sari has, in truth, communicated this peculiar orientation in several instances. Take the example of the unprecedented pagers attack in Lebanon, with thousands apparently harmless devices detonated simultaneously, in crowded street markets, shops, cars, family houses, and hospitals, severely injuring and killing more than 3000 people, including children. Sari saw no reason to exclude thousands of simultaneous “robust proportionality assessments”, and therefore did not see (even in this extreme circumstance) why article 51(4) “could have been relevant”. While I disagree, Sari must be credited for his consistency in his broad construction of proportionality, unlike those who do so only with ‘otherized’ civilians and civilian victims of non-allied states. This emerged, for example, in relation to the author’s assessment of the Israeli army flooding with sea water of tunnels in Gaza, in relation to which he made clear that also the killing of Israeli hostages, as well as the widespread, severe and long-term harm to aquifers, amid an already catastrophic deprival of drinkable water, would have been a matter to be evaluated under proportionality and, as such, potentially lawful.

The author uses this expansive reading of proportionality to conclude that incidental, under IHL, must be interpreted merely as the contrary of “deliberate”, as synonym of “accidental”, and understood as “fortuitous and undesired”. This is also the understanding of some military manuals, indeed contradictory on the point (I have discussed this in the initial article).

But here is why this reading, in my opinion, does not work, both logically and legally.

It is indisputable that under IHL foreseeable incidental civilian harm must be calculated as carefully and accurately as possible in proportionality assessments, before being weighed against the concrete and direct military advantage sought. How can something that must be so carefully predicted, and so thoroughly evaluated, be dismissed merely as accidental or fortuitous?

The terms ‘accidental’ and ‘fortuitous’ indicate mainly the opposite, that is the characteristic of an event, or a circumstance, of being unforeseen. English dictionaries define ‘accidental’ as “happening by chance; not planned”, or “happening or liable to happen in fortuitous or subordinate conjunction with something else of primary importance”, or as “happening or existing by chance”. Essentially the same meaning, with an even more marked opposition with anything predicted or planned, is connected to the term ‘fortuitous’.

Revealingly, during the Geneva Diplomatic Conference, the ICRC, opening the 1974 discussion on the relevant rules (see p. 13), used the expression ‘incidentally or accidentally’, pitting the two adverbs against each other as alternatives. In sum, translating ‘incidental’ as ‘accidental’ or ‘fortuitous’ denies and obscures the fact that under IHL incidental civilian harm is the exact opposite of accidental or fortuitous, it is indeed something thoroughly planned and accurately calculated. IHL, following the approach indicated by the author, would be the only legal framework in modern history qualifying killing and destruction of protected persons and objects precisely pondered, thoroughly calculated, and willfully inflicted by a subject, even when constituting the totality or near totality of the foreseen outcomes of an action, as a random and unintentional side-effect of the action itself. This cannot be right.

Furthermore, there is an even stronger legal basis imposing to reject the other element Sari identifies as a criterion for determining when calculated civilian harm (irrespectively of its magnitude) can be said to be incidental, that is ‘undesired’. This legal basis will be discussed in the second part of this rejoinder.