1. Introduction
During and after recent military operations, especially those conducted by Western states, issues regarding the law of targetingFootnote 1 have attracted a great deal of attention in academic circles, and amongst United Nations (UN) bodies and human rights non-governmental organisations (NGOs).Footnote 2 Such issues include, for example, classification of certain objects as military objectives, the principle of proportionality and the precautions which need to be taken when carrying out an attack. When an operation results in civilian casualties,Footnote 3 which is unfortunately the case in many military operations, calls are often heard not only to hold the state responsible for violations of international humanitarian law (IHL), but also to investigate and impose criminal liability on individuals responsible for those deaths.Footnote 4 In some cases, UN fact-finding missions also investigate and report on these issues with regard to specific operations; these missions often recommend a similar course of action.Footnote 5
Nonetheless, actual criminal proceedings centring on issues of the law of targeting are extremely rare; these matters remain largely unexplored in both national and international courts.Footnote 6 With just over half a dozen relevant cases, the International Criminal Tribunal for the former Yugoslavia (ICTY) is the international court which has adjudicated the most cases on targeting issues.Footnote 7 Moreover, when examining anticipated cases before the ICTY, the International Criminal Court (ICC) and other international tribunals, it appears unlikely that these issues will be the focus of international proceedings in the near future.Footnote 8
Against this background of scarce jurisprudence, the matter of Gotovina and Others Footnote 9 presented the Trial and Appeals Chambers of the ICTY with an opportunity to adjudicate on issues related to targeting.Footnote 10 The Gotovina case revolved around events which occurred in the context of the largest and most intense combat operation to be reviewed by an international criminal tribunal since the trials which followed the Second World War.Footnote 11 Since the Gotovina proceedings dealt with issues of the law of targeting rarely discussed in international judicial fora, and since they did so in the context of a large-scale, intense military operation, a unique opportunity arose to clarify and develop the law of targeting in general, and its applicability in a criminal context in particular. As a group of eminent scholars noted while the case was still pending, ‘the “Gotovina” judgment has the potential to become the “Tadić” of targeting law’.Footnote 12
This contribution explores how the Gotovina judgments dealt with issues related to the law of targeting in a criminal context, and inquires whether the Trial and Appeals judgments met the above mentioned high expectations. It is not intended to address all of the legal issues that arise from the two Gotovina judgments; instead, the article focuses on some of the questions most pertinent to the law of targeting. Section 2 presents a short historical background of ‘Operation Storm’, the military operation at the heart of these proceedings, and a short summary of the relevant findings of the Trial and Appeal Judgments. It focuses on the method of ‘impact analysis’, which was central to both judgments. Section 3 briefly presents background on the law of targeting, and on the interaction between IHL and international criminal law (ICL). Section 4 examines specific incidents and issues related to the law of targeting that arose in the context of the proceedings: the concept of military objectives; the principle of distinction and insufficient information during an attack; issues related to proportionality analysis and analysing the intent behind an attack. The article concludes that both the Trial and Appeals Chambers failed to explain their conclusions regarding those questions. Moreover, they failed to address the relationship between IHL and ICL, and the tensions that arise between these two branches of law in the context of the law of targeting. Thus, it appears that the opportunity to create the ‘Tadić judgment of targeting’ was missed.
2. Background
2.1. Operation Storm
At a very early stage during the break-up of Yugoslavia, in December 1990, the Serb community in Croatia established an autonomous region on approximately one-third of the territory of what is now Croatia. In December 1991 this region declared independence and became known as the Republic of Serbian Krajina.Footnote 13 Following Croatia's declaration of independence in 1991, it attempted to regain control over that territory.Footnote 14
On 4 August 1995, just weeks after the Srebrenica genocide,Footnote 15 Croatia launched a massive military campaign code-named ‘Operation Storm’, with the goal of taking over the territory controlled by the Serbs.Footnote 16 By 7 August, major combat operations ended in a decisive Croatian victory: Croatia took over the Republic of Serbian Krajina and drove back the Serbian army.Footnote 17 The commander of Operation Storm in the southern region of Krajina was General Ante Gotovina.
According to the Trial Chamber (TC), the operation led to ‘526 Serb casualties, including 116 civilians, in addition to 211 casualties among Croatian soldiers and policemen and 42 Croatian civilian casualties’,Footnote 18 and to the deportation of (at least) 20,000 Serbs.Footnote 19 These events, and especially the artillery attacks on several towns in KrajinaFootnote 20 and the deportation of tens of thousands of Serbs, were at the heart of the proceedings against Gotovina before the ICTY.
2.2. The Trial Judgment and the Impact Analysis
In 2005, after being at large for several years, Gotovina was captured and transferred to the custody of the ICTY.Footnote 21 He was accused, along with two others,Footnote 22 of bearing responsibility for crimes against humanity and war crimes committed between July and September 1995 against the Serb population in the Krajina region. The prosecution alleged that before, during and after Operation Storm, the three defendants took part in a joint criminal enterprise (JCE) which was designed to permanently remove the Serb population ‘from the Krajina region by force, fear or threat of force, persecution, forced displacement, transfer and deportation, as well as appropriation and destruction of property’.Footnote 23
While Gotovina was never formally charged with any war crime to which the law of targeting is relevant,Footnote 24 the artillery attacks carried out over the course of the operation were one of the key issues discussed in the judgment. In order to prove that the elements of the crime of deportation had been fulfilled,Footnote 25 the prosecution argued that the shelling of civilian objects was the coercing factor that drove the Serb population out of Krajina. Thus, indiscriminate artillery fireFootnote 26 constituted the alleged actus reus of the crime against humanity of deportation. As stated earlier, Gotovina was accused of taking part in a JCE, the purpose of which was the deportation of the Serb population from Krajina. The prosecution's perception of the JCE resulted in the crime of deportation being presented as central to establishing the mode of liability upon which all of the crimes in the case were based. Since the artillery attacks were presented as the actus reus of the crime of deportation, and since the deportation was a key element of the alleged JCE, they became a crucial element in establishing the main mode of Gotovina's liability.Footnote 27
In examining the legality of the artillery attacks, the TC reviewed experts' reports, witness statements, transcripts of meetings in which Gotovina had participated, artillery logs and official orders given by Gotovina before and during Operation Storm. Based on all of these sources, the TC could not conclude beyond reasonable doubt that the artillery attacks were intended to target military or civilian objects.Footnote 28 While the TC regarded some evidence as indicating an intent to attack civilian targets, it also found evidence which indicated that there was no such intent.Footnote 29
The TC then conducted an ‘impact analysis’, examining the exact place where the artillery shells landed in order to establish whether they were directed at civilian or at military targets.Footnote 30 Of more than 1,200 artillery shells found by the TC to have been fired at Knin, Benkovac, Obrovac and Gračac (the Four Towns),Footnote 31 the prosecution was able to prove the impact location of only 154 (13 per cent).Footnote 32 As is evident from the Trial Judgment, the TC was fully aware that it was missing a great deal of information both in terms of where the shells had landed, and of the number of shells that had hit each target.Footnote 33
At this stage, the only factor which the TC took into account in determining the intended target of a shell was the location of its actual impact. In making its determination, the TC construed an evidentiary standard according to which any shell landing within a 200-metre radius of a military objective was to be regarded as having been aimed at a military objective (the 200-metre standard).Footnote 34 Consequently, any shell landing more than 200 metres from a military objective was regarded by the TC as having been aimed at a civilian object, and was therefore unlawful. Certain aspects of this methodology are discussed further in Section 4.4.
After careful analysis, the TC found that a total of 74 shellsFootnote 35 had landed more than 200 metres away from any military objective.Footnote 36 As a result of these findings, the TC concluded that forces under Gotovina's command had carried out unlawful attacks during Operation Storm.Footnote 37 The TC then accepted the prosecution argument and determined that these unlawful attacks were the actus reus (the coercive element) of the crime against humanity of deportation.Footnote 38 In turn, this crime of deportation was found to constitute the common purpose of the JCE and the artillery attacks were the manner in which Gotovina had contributed to its execution.Footnote 39 Moreover, the TC found that Gotovina's participation in and contribution to the JCE was the basis of his responsibility for other crimes that took place during and after Operation Storm. After reaching these conclusions, the TC went on to unanimously convict Gotovina of crimes against humanity and war crimes,Footnote 40 and sentenced him to 24 years’ imprisonment.Footnote 41
2.3. The Appeals Judgment
The Appeals Chamber (AC) was fiercely divided over Gotovina's appeal. The majority judges – President Meron, and Judges Robinson and Güney – acquitted Gotovina on all counts, despite a very strong dissent by Judges Agius and Pocar. First, the AC decided unanimously that the 200-metre standard used by the TC was not based in law or on the evidence presented during trial. Thus, the judges agreed that the TC had erred in using the standard as part of its impact analysis.Footnote 42 However, the judges disagreed over the significance of this error. The majority judges declined to analyse the artillery attacks without the 200-metre standard, or to set any other standard in its place. They pointed to the fact that the TC had not considered the evidence as a whole, absent the impact analysis, as proving beyond reasonable doubt that the artillery attacks were aimed at civilian objects.Footnote 43
The majority then noted that the unlawful attacks constituted the coercive element that was considered by the TC to be the actus reus of the crime of deportation. Since the majority judges could not find that any unlawful attack had been carried out, they acquitted Gotovina of the count of deportation. Furthermore, since the TC had found the deportation to be the underlying purpose and an element of the JCE – the mode of liability under which Gotovina had been convicted – the majority judges determined that following Gotovina's acquittal on the count of deportation, the existence of the JCE had not been proved.Footnote 44 Without this mode of liability, and after briefly ruling out alternative modes of liability,Footnote 45 the majority acquitted Gotovina of all counts.Footnote 46
The dissenting judges strongly rejected the reasoning of the majority. Although these judges agreed that the 200-metre standard was based neither on the law nor on the evidence, they could not accept the majority's dismissal of the entire conviction simply because this faulty evidentiary test had been rejected.Footnote 47 The dissenting judges concluded that the totality of evidence, even without the impact analysis, established beyond reasonable doubt that Gotovina had deliberately directed attacks against civilian objects.Footnote 48
3. General Targeting Issues and the Application of IHL in a Criminal Context
The Gotovina case has raised several questions with regard to the law of targeting, in addition to the 200-metre evidentiary standard described above. Before turning to several of the targeting issues that were discussed in Gotovina, primarily in the Trial Judgment,Footnote 49 an overview of the general principles and rules of IHL and the manner in which they are applied in a criminal context is in order.
IHL imposes restrictions on the types of person and object that a party to a conflict may lawfully attack, and also on the means and methods that may be used against those lawful targets.Footnote 50 These restrictions date back as far as the first IHL treaties, and may be traced back to the 1868 St Petersburg Declaration.Footnote 51 The most fundamental targeting principles are distinction, precautions and proportionality.Footnote 52
The principle of distinction is enshrined in Articles 48 (basic rule), 51 (persons) and 52 (objects) of Additional Protocol I, and is considered to reflect customary international law.Footnote 53 Certain violations of this principle may amount to international crimes entailing individual criminal responsibility.Footnote 54 These articles address the parties’ obligation to direct their attacks only at military objectives, combatants and persons taking direct part in hostilities;Footnote 55 parties to a conflict must never intentionally target civilians or civilian objects. The principle of distinction prohibits only the intentional targeting of civilian objects; it does not completely prohibit all damage to such objects as the result of an attack on military objectives. Thus, the principle does not impose strict liability on commanders; nor does it require perfect results.Footnote 56 It ‘only’ requires that commanders should never direct their attacks towards civilians and civilian objects. This fact makes the intent of the attacker relevant not only for the analysis of the attack under criminal law, but also in terms of its analysis under IHL.
The definition of ‘military objectives’ under IHL is found in Article 52 of Additional Protocol I,Footnote 57 which is considered to reflect customary lawFootnote 58 even by states not party to it.Footnote 59 It contains a two-pronged test for determining whether an object is a military objective.Footnote 60 The first aspect of this test is whether – because of its nature, location, purpose or use – the object makes an effective contribution to the adversary's military action. The second aspect of the test is whether the total or partial destruction, capture or neutralisation of the object, in the circumstances ruling at the time, offers a definite military advantage to the attacker.Footnote 61 The two tests are cumulative.Footnote 62
A further set of obligations imposed on an attacker is found in Article 57 of Additional Protocol I,Footnote 63 which lists several precautions that should be taken. These precautions include, but are not limited to, giving warnings before an attack unless the circumstances do not permit, choice of the means and methods of attack, and the timing of an attack.
After identifying the target being attacked and after taking relevant and feasible precautions, the principle of proportionality comes into play. This principle is enshrined in Article 57(a)(iii) of Additional Protocol I, and is also considered to reflect customary international law.Footnote 64 It restricts, to a degree, a party's ability to attack military objectives: the attacker must weigh the anticipated military advantage against the expected harm to civilians, and must not go through with the attack if anticipated harm is excessive in relation to the military advantage. Under the ICC Statute, certain violations of this principle amount to war crimes entailing individual criminal responsibility.Footnote 65 Moreover, IHL obliges parties to ‘take all feasible precautions in the choice of means and methods of attack with a view to avoiding, and in any event to minimizing, damage to civilians and civilian objects’.Footnote 66 It should be noted that this obligation is more demanding than merely verifying that an attack is proportional.
In accordance with these basic principles, every attack should be analysed first in light of the principle of distinction; then, and only if it was directed against a military objective, in light of the precautions taken by the attacker; and finally in accordance with the principle of proportionality.
IHL and ICL share certain features. They both apply during armed conflict, and it may be said that they are both intended to advance similar overarching goals, although they do so by different means. However, the two branches of law are not identical. IHL focuses on the state and is meant to guide and control its behaviour during armed conflict, ex ante. It is a regulative branch of law, intended to be applied in the battlefield by non-lawyers. ICL, on the other hand, applies ex post, and focuses on individuals and their criminal responsibility. It is intended to be applied by lawyers and judges in a courtroom. Because of these fundamental differences between the two branches of law, in conjunction with their parallel applicability in times of armed conflict, the interaction between them is complex.Footnote 67
IHL is the normative basis of certain aspects of ICL; war crimes are violations of these norms.Footnote 68 Thus, what is legal under IHL can never be a war crime under ICL.Footnote 69 The opposite is not true: not all violations of IHL are considered war crimes.Footnote 70 Moreover, it appears that the ICTY OTP traditionally applies the same logic in relation to crimes against humanity committed during armed conflict. Since what is punishable under ICL is narrower than what is prohibited under IHL, it is important not to confuse the two; this would result in reducing the scope of protection afforded under IHL.
The opposite, confusing ICL with IHL, is also disturbing; it would unjustifiably narrow the scope of what is permitted under IHL in a way that does not really reflect IHL but rather other branches of law. In this case, IHL might become irrelevant and practically inapplicable during fighting, ultimately leading to a rejection of the law altogether. In turn, rejection of the law would create greater suffering for civilians and combatants alike.Footnote 71 These risks require judges and practitioners to be explicit about the branch of law that they apply in any given determination or argument. This issue was of some relevance in the Gotovina proceedings.
Despite the fact that some aspects of ICL are based on IHL, IHL rules cannot be automatically applied in a criminal context. Because of the different logical basis of each branch of law,Footnote 72 caution and prudence are required when applying IHL in a criminal context. Moreover, various rules of IHL may require some adjustment before they can be used properly in a criminal context.Footnote 73 Unfortunately, the lack of jurisprudence and scholarly work on this matter hinders a better understanding of the proper way in which to apply IHL in a criminal context.
A good illustration of the complex relationship between IHL and ICL, and the need to make certain adjustments in situations where they are mutually applicable, can be found in the Galić case in the ICTY. This case concerns, among other aspects, attacks on persons. In this case, the judges noted the IHL rule that clearly states ‘[i]n case of doubt whether a person is a civilian, that person shall be considered to be a civilian’.Footnote 74 In accordance with this rule, the state is required to conclude positively that a person is not protected from attack. Despite that, the ICTY determined that in a criminal context the burden is reversed and ‘the prosecution must show that in the given circumstances a reasonable person could not have believed that the individual he or she attacked was a combatant’.Footnote 75 The Gotovina judgments triggered, or should have triggered, similar questions regarding the proper parallel application of IHL and ICL. The following section will examine the way in which the TC and AC dealt with several such questions.
4. Targeting Issues in the Two Gotovina Judgments
4.1. Definition of ‘Military Objectives’: Failure to Articulate Criteria for Classification
The indictment against Gotovina did not formally contain any count of unlawful attack against civilians; nevertheless, the TC addressed the legality of the artillery attacks conducted during the operation. The TC implicitly adopted the ICTY's traditional approach – also evident in the way in which the prosecution argued the caseFootnote 76 – that only unlawful attacks could constitute the actus reus of crimes against humanity, in this case deportation.Footnote 77
As part of analysing the legality of the attack, the TC treated a rather large variety of objects as military objectives. These include dual-use objectsFootnote 78 (such as fuel stations,Footnote 79 post offices and railway stations);Footnote 80 objects that have been considered controversial in other contexts (such as police stations);Footnote 81 and various locations that have been considered to be targets in and of themselves (such as open fieldsFootnote 82 and intersections).Footnote 83
The TC's willingness to regard these objects as military objectives is especially noteworthy considering that it ultimately held that Gotovina did intentionally attack civilian objects. However, the TC did not explicitly refer to the definition found in Article 52 of Additional Protocol I, or to any other definition of ‘military objectives’. Rather, it only briefly mentioned that certain objects are considered to be military objectives or that they provide a ‘definite military advantage’,Footnote 84 without providing any real analysis. The TC's failure to explain the reasoning behind its conclusions is regrettable for several reasons. To begin with, most of the scholarly discussion on the principle of distinction thus far has focused largely on persons; the exact scope of the concept of military objectives requires further development and refinement even in IHL, and far more so in a criminal context.Footnote 85 Even if one were to concede that the definition of ‘military objectives’ is almost undisputed, its application to specific circumstances tends to be highly complex. This disagreement on the application of the law may result, at least partly, from the scarcity of judicial rulings that have actually applied the law to a specific set of facts. While there are ICTY judgments that address similar questions,Footnote 86 none of them dealt with intense combat situations similar to that in Gotovina.
Another area in which questions were left unanswered by the TC is in respect of the burden of proof. It is uncertain which party, in a criminal trial, must prove that the conditions for lawful attack under Article 52 of Additional Protocol I have been fulfilled. For example, it is unclear whether the prosecution is required to prove beyond reasonable doubt that an object could not possibly be used, in the future, in a manner that would make an effective contribution to military action.Footnote 87 As mentioned above, the ICTY's jurisprudence had already addressed this question in the context of targeting persons, but no such determination had been made in relation to targeting objects. While it is reasonable to assume that the same logic – that is, which led the AC in Galić to place on the prosecution the burden of proving beyond reasonable doubt that a person was a civilian – should apply also to objects, the TC should have explicitly stated this. Moreover, the TC should have been explicit as to the branch of law it applied: were its determinations based on its interpretation of IHL or of ICL?
4.2. The Attack on Martić's Residence – Private Homes as Military Objectives and Attacking in Situations where Insufficient Information is Available
The TC treated one attack, involving Milan Martić, differently from all of the other attacks and targets mentioned in its judgment. As is explained below, the TC's analysis of this attack encompassed issues of both distinction and proportionality.Footnote 88 This section deals with the classification of military targets in the context of this attack, and with issues relating to launching an attack in cases where insufficient information is available.
Martić was Commander-in-Chief of the Serb forces and President of the Republic of Serbian Krajina. According to the TC, early on the morning of 4 August 1995 (the first day of Operation Storm), Croatian forces fired 12 shells at Martić's apartment. Later that day, they fired an unknown number of shells at a different location, referred to as ‘area marked R’.Footnote 89 As in the case of other targets, the TC began by analysing whether the target attacked was a military objective. It noted Martić's high rank in the Serb forces and found that the purpose of the attack was not to kill him, but rather to harass him and to disrupt his ability to command and control Serbian forces.Footnote 90 The TC went on to determine that ‘firing at [Martić's] residence could disrupt his ability to move, communicate, and command and so offered a definite military advantage, such that his residence constituted a military target’.Footnote 91
While the TC concluded that the attack was directed at a military objective, it is not clear from the judgment whether the target was Martić himself or his residence. The statement above explicitly and clearly refers to Martić's residence as the military objective.Footnote 92 Moreover, when the TC reviewed artillery logs and target lists prepared by the Croatian forces, it noted that they listed as a military target, among others, ‘Martić's residence’.Footnote 93 However, other parts of the judgment imply that the target of the attack was Martić himself, and not his residence.Footnote 94 This confusion was not resolved in the Appeals Judgment. The majority judges did not address this question; they merely commented on the TC's proportionality analysis. The dissenting judges similarly did not present any coherent position on this issue. For example, one of the dissenting judges stated that Martić was the object of attack but that his location was not known to the Croatians. He referred to the attack as ‘firing at two locations where the HV believed Martić could be found’.Footnote 95 However, the same judge later seemed to regard Martić's residence as the military objective when he stated that ‘[a]t no time did the Trial Chamber doubt the legitimacy of targeting Martić's residence’.Footnote 96
In light of the ambiguity in the TC's analysis,Footnote 97 the following section discusses questions that arise from each possibility: the classification of Martić's residence as the military objective attacked, and the classification of Martić himself as the object of attack.
4.2.1. Martić's Residence as a Military Objective
If the military objective attacked was Martić s residence, then the Trial Judgment failed to properly substantiate this conclusion. The TC did not explain how the residence satisfied the first prong of the test found in Article 52 of Additional Protocol I. Did the apartment, by its nature, make an effective contribution to military action by the Serb forces? Or was it used, or was it intended to be used in the future, to make an effective contribution? As for the second prong of the test, during its brief discussion of the attack the TC explicitly referred to the definite military advantage that the Croatians sought to achieve: to harass Martić and to disrupt his ability to command and control. However, the TC does not explain how the residence's destruction, neutralisation or capture promoted this military advantage.Footnote 98 To this end, the judgment should have provided more information regarding Martić's residence. Was the residence actually used by Martić to command the Serbian forces? If not, was it prepared and equipped for such use in the future? For example, did it contain communications equipment that would potentially enable Martić to exert command and control from his apartment? As it stands, this part of the judgment is not factually substantiated.
Attacking the personal residences of high-profile military or political leaders is not a new or unique phenomenon. The most well-known attacks of this kind include the NATO bombardment of Slobodan Milošević's house,Footnote 99 the attack by NATO on Muamar Al Qaddafi's presidential palace,Footnote 100 and an attack on the residence of Ivory Coast president, Laurent Gbagbo.Footnote 101 Indeed, some of these attacks, such as that on Milošević's house, have been justified in light of the use or potential use of the residence as a command and control centre. However, it is not entirely clear exactly which features of Milošević's residence led to this conclusion or what was the reasoning behind the other attacks mentioned.Footnote 102
If the TC did classify Martić's residence as a military target, it should have explained exactly the basis on which this branch of law was adopted. Such an explanation could have shed light on some aspects of the law, especially on the interpretation of some of the components of the definition of ‘military objectives’. Moreover, an explanation could have clarified whether the TC made its determination based on IHL and established that this was indeed a military target, or whether the conclusion was based on ICL and was the result of the prosecution's failure to prove the opposite beyond reasonable doubt.
4.2.2. Targeting Martić – Attacks in Situations where Insufficient Information is Available
If the target was in fact Martić, rather than his apartment, there is no dispute that he was a lawful target because of his role in the Serb armed forces. However, it is significant that his location was not known to the attacking forces either when his apartment was shelled or when, later the same day, the ‘area marked R’ was attacked.Footnote 103 Consequently, the question arises as to the level of certainty required before ordering an attack.Footnote 104 In an age in which targeted killings are increasingly common, allowing attacks on specific people without clarifying the level of information and certainty required could have severe ramifications.
IHL treaty law contains several articles dealing with situations of doubt during attack.Footnote 105 While they are not directly relevant to the question at hand – what to do in cases of lack of certainty regarding the location of a target – they serve as an indication that under IHL the attacker is expected to have a high level of certainty before attacking a target.
The Trial Judgment does not establish, as a factual matter, to what extent the attacking forces knew of Martić's whereabouts at the time of the attack. This silence is especially troubling considering that the TC explicitly found that the Croatians were unable to produce real-time intelligence on military targets within the Four Towns.Footnote 106 It is possible that before the first attack, which took place very early in the morning at the beginning of the operation, the attackers could reasonably have assumed that Martić would be in his residence.Footnote 107 However, it is unclear from the Trial Judgment how, why, and on the basis of what information the ‘area marked R’ was chosen as a target.Footnote 108
Moreover, as in other parts of its judgment, the TC failed to clarify upon which branch of law its analysis was based. This has great significance for the issue of the burden of proof: was the attacking force required to show that it had fulfilled the requirements of IHL, and to prove that Martić's location was known, or, as accepted in a criminal context, was the prosecution required to prove the opposite beyond reasonable doubt? It appears that here, as in other contexts, there is tension between IHL, which places the burden on the attacker, and ICL, which places the burden on the prosecution. Thus, it is extremely important for the judges to be explicit and clear about what they did and why they did it. If they decided this issue based on considerations of ICL, they should have said so explicitly in order to avoid their determination affecting in any way the obligations imposed by IHL.Footnote 109
To conclude the discussion of the attack involving Martić, each of the possible interpretations as to what the TC regarded as the military target, and even a combination of both options, poses serious legal questions that should have been answered or at least addressed by the TC. These questions include (i) the categorisation, in certain circumstances, of private residences of senior commanders as military objectives; and (ii) situations of doubt regarding the exact location of the target. These issues are extremely relevant and important to modern-day combat operations, and both military and civilians could have benefited from a more methodological, consistent and rigorous analysis. Such analysis might have resulted in greater clarity and legal predictability in this field for the sake of guiding future behaviour. Perhaps more importantly, the TC should have clarified the way in which IHL is to be used when analysing individual criminal responsibility. Since the AC also failed to address any of these questions, both Chambers missed an opportunity to clarify the law.
4.3. The Non-Existent Proportionality Analysis
This section briefly presents the decisions of both Chambers regarding the proportionality of the attack on Martić's residence. It then focuses on the proportionality analysis that should have taken place, but which was not in fact conducted by either Chamber. As mentioned earlier, the TC only analysed the proportionality of the attack on Martić's residence and concluded that this attack was disproportionate.Footnote 110 According to the TC, the attack created a risk to civilians that was excessive in relation to the anticipated military advantage.Footnote 111 The majority of the AC criticised the TC's finding, stating that it ‘was not based on a concrete assessment of comparative military advantage, and did not make any findings on resulting damages or casualties’.Footnote 112 Based on this alone, it is not clear whether the AC majority judges actually overturned the TC's decision on this point.Footnote 113
Various commentators have thoroughly discussed the proportionality analysis of the Martić attack conducted by the TC;Footnote 114 it is not necessary, therefore, for this article to further address that issue. Rather, it will discuss the proportionality analysis that was not conducted by the Chambers.
The analysis of whether an attack was lawful under IHL must include, inter alia, the consideration of two cumulative requirements: (i) it must be directed against military objectives, and (ii) the expected damage to civilians must not be excessive in relation to the concrete and direct military advantage anticipated from the attack.Footnote 115 While the TC discussed the principle of distinction – and meticulously examined whether the attacks were aimed at military objectives (albeit with all the shortcomings described in this article) – it never even mentioned the issue of proportionality regarding the vast majority of the attacks other than the attack on Martić's residence.Footnote 116 The TC stated explicitly that it would not pronounce on the proportionality of the attacks on targets other than that specific incident.Footnote 117 The AC, similarly, did not examine the proportionality of any other incident or target. From an IHL perspective, this led to perhaps the greatest legal mistake made by the both Chambers – the complete lack of proportionality analysis of the attack on the Four Towns.
As the TC found that some attacks were directed against civilian objects, no proportionality analysis was required in order to decide upon their illegality. However, it also regarded a considerable number of the attacks to have been directed at military objectives. As mentioned above, this determination does not suffice to conclude that these attacks are legal.Footnote 118 Since Gotovina was not formally charged with unlawful attacks, the TC perhaps believed that its finding that civilian targets were intentionally attacked in some instances was sufficient to establish the actus reus necessary to establish the charge of deportation. From a criminal law point of view, this may excuse the TC's avoidance of a proportionality analysis. The AC's avoidance cannot be justified in this manner, as it found that no attack had been proved to be directed at civilian objects. Its lack of proportionality analysis is especially striking since it went on to acquit Gotovina without examining the mandatory factor of the proportionality of the attacks.
The AC's reluctance to decide the case based on proportionality may, nonetheless, be understandable.Footnote 119 First, as is often mentioned, there has never been any conviction, international or domestic, for the crime of disproportionate attack.Footnote 120 Furthermore, the questionable customary status of the crime of disproportionate attack, at least at the time of the Yugoslavian conflict, contributes to the difficulties of relying on lack of proportionality to substantiate a conviction.Footnote 121
Both Chambers would have had to decide difficult legal questions if they had discussed proportionality. Were they required to analyse every single military target attacked separately – as the TC did concerning the dozen shells directed at Martić's residence – or should they have treated the entire use of artillery in the context of Operation Storm (over 1,200 shells in total) as an overall attack, the proportionality of which required assessment?Footnote 122 Should the focus have been somewhere in between, perhaps analysing each of the Four Towns separately?Footnote 123 How should have the military advantage sought been described? May an attacker legally take into account the fact that the civilians have fled the conflict zone when analysing anticipated civilian damage? Would the answer be any different if, rather than civilians fleeing voluntarily, the attacker had deported them?
In addition to these questions, the Chambers would have been faced with the question of the weight, under the proportionality test, that is to be ascribed to the actual damage to civilians and civilian objects during an attack.Footnote 124 The AC majority judges criticised the TC for failing to address this factor as part of its proportionality analysis.Footnote 125 As mentioned earlier, Operation Storm – especially when taking into account its wide scope – caused relatively few civilian casualties and little damage to civilian objects.Footnote 126 Yet, if the anticipated military advantage is extremely low – which cannot be said to be the case with Operation Storm – even very little civilian damage may be excessive in comparison. Conversely, it is possible that the attacker launched the attack anticipating much greater damage than that which actually occurred, in a way that could potentially make the decision to launch the attack illegal even though the actual damage was relatively minor. Finally, it is important to remember that the law refers to anticipated damage, and not the damage that was actually caused. The actual damage may perhaps serve as an indication of what was anticipated, but it has no legal meaning in and of itself.Footnote 127
Perhaps the relatively small amount of actual damage makes Gotovina a bad case for a groundbreaking judgment on issues of proportionality.Footnote 128 This is especially true considering the uncertain customary status of the criminal prohibition on disproportionate attacks, the precedential nature of any potential determination that a crime of disproportionate attack was committed, and the difficulty of conducting a criminal trial on issues of proportionality. Thus, it is conceivable that both Chambers completely avoided the issue for policy reasons. They did not want to state explicitly that firing 1,200 shells into urban areas, even if those shells are directed at military targets, is proportionate. On the other hand, they felt unable to determine that Gotovina had conducted unlawful attacks solely on the basis of a proportionality analysis.
The law of targeting could have benefited from more clarity regarding such important and complex legal questions, and the Chambers' failure to address them, while somewhat understandable, is legally incorrect. More importantly, the Chambers failed to state explicitly which branch of law formed the basis of their decisions. Even after reading both judgments, it is not clear whether Gotovina violated the principles of proportionality under IHL but was nonetheless acquitted based on considerations of ICL, or whether his actions were not even a violation of IHL. Thus, both judgments have failed to guide future actions of military commanders and have created ambiguity that may potentially endanger civilians in future conflicts.
4.4. How to Establish the Intent Behind an Attack?
4.4.1. The Impact Analysis
Though the issues discussed above raise very interesting and important questions relating to the law of targeting, they attracted very little, if any, attention during the appeal. The main area of contention was the impact analysis conducted by the TC. As described earlier, after reviewing the evidence the TC could not establish Gotovina's intent beyond reasonable doubt. At that point, it chose to analyse the locations of the impact of the shells in order to assess the intended targets of those shells. It did so using the 200-metre standard, regarding every shell that landed less than 200 metres from a military objective as having been directed at that target, and every shell landing more than 200 metres from a recognised military objective as having been directed at civilian objects.
During the appeal, the AC discussed the specific 200-metre standard established by the TC. As mentioned above, all five judges agreed that the standard was not based on either evidence presented at the trial or on the law, and that it should therefore be rejected.Footnote 129 The majority judges declined to set a different standard and to re-analyse the evidence accordingly, a position heavily criticised by the minority judgesFootnote 130 and by various commentators.Footnote 131 One reason given for the majority's determination was that a single standard is not sensitive enough to operational reality, and that it fails to take into consideration factors that were deemed by the TC itself as affecting the accuracy of the weapons used.Footnote 132
The AC majority judges were obviously correct on this point. As convincingly argued by prominent IHL scholars in their amicus curiae brief submitted during the appeal,Footnote 133 in establishing the mens rea for acts committed during intense fighting, any court should take into account the specific operational considerations and the technical aspects associated with the weapons used by the attacker.Footnote 134 This is all the more true when attempting to establish the intent of an attacker by looking only at the results of an attack.
However, it remains doubtful whether any single standard, even one sensitive to operational considerations, could be used to establish beyond reasonable doubt the intent of the person physically carrying out an actual attack, let alone the intent of the person ordering or commanding the attack.Footnote 135 Artillery experts have submitted that ‘outliers’ (shells that, for some reason, fall very far from their intended and probable area of impact) are technically inevitable.Footnote 136 These outliers may be the result of technical problems in the shell itself or in other components of the firing process (such as the explosives used for firing);Footnote 137 or they may be the result of human error made by someone involved in the long, complex process of planning and conducting the attack.Footnote 138 Since none of these factors reflect the intent of the attacker, it is doubtful whether any single standard, as sensitive to operational concerns as it may be, could properly reflect the mens rea of the attacker.Footnote 139
If one nevertheless chooses to use this sort of standard to analyse an attack, despite its inherent incapability of establishing an attacker's intent at the time of the attack, such use raises another set of problems and difficulties. What is the exact percentage of missed shells that justifies a conclusion beyond reasonable doubt regarding the attacker's mens rea? Intuitively, if one shell out of a thousand were to fall outside the determined standard, it would surely not be convincing beyond reasonable doubt that the attack as a whole was directed at civilian objects. However, if 999 out of the thousand shells fired were to fall outside the properly determined standard, it would be easy to conclude that the attack was directed at civilian objects. It appears that the percentage of shells impacting outside the determined standard is a relevant factor in establishing the intent behind an attack.Footnote 140
The Trial and Appeals Chambers did not specifically address this question. In some instances, the TC hinted that the large number of shells falling outside the determined standard was a factor in its conclusion regarding the unlawfulness of the attacks.Footnote 141 For its part, the amicus curiae brief submitted by IHL scholarsFootnote 142 stressed the relatively low percentage of shells (4 per cent)Footnote 143 that landed outside the impact zone as a key factor in establishing the legality of the attacks. How is it possible to settle the difference between the TC's reference to ‘too many’ shellsFootnote 144 and the experts' reliance on the low percentage of shells falling outside the impact standard?
One way is to understand the TC as suggesting that 4 per cent of the shells falling outside the determined standard is high enough to justify an affirmative conclusion regarding the attacker's mens rea. Intuitively, this seems unreasonable. There is, perhaps, a better way to understand the difference between the TC and the experts. The Trial Judgment could be understood as relying on the number of shells landing outside the determined standard: not compared with the total number of the shells fired, but calculated out of the total number of shells, the impact location of which had been established.Footnote 145 As mentioned earlier, the TC could only establish the impact location of a relatively small percentage of shells.Footnote 146 Of those 154 shells, 74 (almost half) fell outside the impact zone.Footnote 147 This reasoning seems to be the most compelling means of explaining the TC's view that too many shells landed very far from military objectives.Footnote 148
However, this approach merits criticism, because it contradicts one of the most basic tenets of ICL: the presumption of innocence.Footnote 149 According to this presumption, the burden of proof lies with the prosecution to prove its case beyond reasonable doubt.Footnote 150 The TC was not entitled to disregard the shells in the many instances where their impact location was not established, and instead to rely only on the impact locations that the prosecution was able to prove.Footnote 151 The impact location of more than 1,000 shells was not unknown; it was simply not proved by the prosecution.Footnote 152 In considering only the shells in respect of which the impact location had been established, the TC infringed Gotovina's presumption of innocence.
The degree of deviation required in order to reach any conclusion regarding the intent of an attacker should also be influenced by the shortcomings inherent in any ex post impact analysis. Such an analysis is usually conducted months or years after the occurrence of the attacks, thus hindering the collection of ‘real time’ evidence and increasing the reliance on secondary evidence such as NGO and news reports.Footnote 153 This often results in judges being exposed to only a partial picture of the events examined.Footnote 154 Furthermore, the passage of time arguably makes the collection of evidence biased towards shells that landed on civilian targets, since shells landing on military targets are presumably less recorded and draw less attention.Footnote 155 Moreover, there are the inherent difficulties involved in relying on witness testimonies of events that occurred under the fog of battle, including issues such as the number of shells impacting upon a certain location.Footnote 156
All of these practical and legal difficulties should discourage reliance on any means of impact analysis in order to establish the mens rea of an attacker. At the very least, they require that a very high percentage of deviation from a sensitive standard be demanded, in order to balance against the inherent conditions favouring presentation of evidence regarding hits on civilian targets.
In light of the difficulties that any impact analysis raises – and, most importantly, the inability to accurately reflect the intent of the attacker at the time of an attack – if all of the surrounding evidence fails to prove beyond a reasonable doubt the intent of the accused, he or she should be acquitted except in the most exceptional cases where the vast majority of the attacks are found to deviate from any sensitive impact standard (such as in cases previously decided by the ICTY).Footnote 157 Judges should not lose sight of the fact that what they must establish is what the attacker knew and intended before the attack. This is what the law mandates. Conducting an impact analysis entails relying on what happened after the attack. It should be used only as a means to help determine the intent and knowledge of the attacker; it should not become an end or a legal test in its own right. Although merely conducting an impact analysis cannot be ruled out completely as contrary to the law, a prudent judicial policy suggests that it should be used only on rare occasions when it can truly expose, beyond reasonable doubt, the intent and knowledge of the attacker before the attack.Footnote 158
4.4.2. Mistaken Identification
The article now turns to examine an incident in which the TC adopted an approach that was more sensitive to the actual intent and knowledge of the attacker. This approach is also more in line with fundamental legal concepts of IHL and ICL. The TC analysed an incident that involved a factual mistake by an attacker, a mistake which led to an attack on a civilian object. The TC, over the course of its impact analysis, discovered that one of the targets attacked by the Croatian forces was erroneously marked on the map used by the attackers as being located 150 metres away from its actual position. For the purpose of the impact analysis, the TC took as its point of reference the location marked on the map, which in reality did not represent any military objective.Footnote 159 Thus, when conducting the impact analysis, it reached the conclusion that the impact was less than 200 metres from the place marked on the map; it was therefore not an unlawful attack, despite the fact that in reality the shell impacted more than 200 metres from any military target.
This is a rare occasion on which a judicial decision analyses, albeit without particular thoroughness, what appears to be an honest mistake by an attacker.Footnote 160 The TC's analysis accurately reflects the state of mind of the attacker – who honestly believed that he was attacking a military objective because of the manner in which it was marked on the map – at the time of the attack.
The approach of the TC in this incident is a proper application of IHL and is consistent with the view of most scholars, emphasising the intent of the attacker and what he knew and believed at the time of the attack, and not the results of the attack.Footnote 161 This is a well-known legal requirement, but is one that is often disregarded by academia, NGOs and even UN bodies that pronounce on targeting issues despite having little or no information about the attacker's knowledge, basing their analysis solely on the consequences of the attack.Footnote 162 Ironically, this legal requirement was also, to a degree, disregarded by the TC in other parts of its impact analysis.
This part of the judgment sends a clear message that without information about an attacker's knowledge and intent at the time of the attack, no conclusion may be drawn regarding the legality of such an attack.Footnote 163 This important message went largely unnoticed because the TC failed to explicitly explain its determination. More importantly, it was lost because the reliance on the impact analysis, as explained above, was based on diametrically opposed logic and sent a contradictory message. A clearer legal position by the TC, and a more consistent approach in applying the logic of this analysis to all other targets, would have resulted in a more legally correct and positively influential Trial Judgment.
5. Conclusion
The Gotovina proceedings presented the Trial and Appeals Chambers of the ICTY with an extremely rare opportunity to adjudicate on core issues related to the law of targeting in the context of intense fighting. Though such issues are extensively discussed and analysed by academics and NGOs, the law of targeting is sadly deficient in terms of authoritative judicial clarification regarding some of its most basic concepts. Unfortunately, as demonstrated above, both Chambers missed this opportunity, which is not likely to repeat itself in the near future. They have failed to influence the law of targeting, and especially the way in which it is manifested in the context of a criminal trial.
The Chambers failed to address some of the difficult IHL questions that arose in this case. Moreover, they failed to identify and address the tensions in the application of IHL in a criminal context. Both Chambers failed to appreciate, or at least they failed to state explicitly, that the situation is governed by two branches of law: IHL and ICL. They failed to explain which branch of law controlled each determination that was made, whether the decision involved classifying an object as a military objective or conducting an impact analysis. Furthermore, despite their parallel applicability, these two branches of law are based on different foundations and attempt to achieve different, yet related, goals.Footnote 164 They often provide different answers to the same questions. This makes it all the more important for any court to be extremely clear and explicit about which body of law forms the basis of its analysis.
One of the main issues of divergence between these two branches of law is the burden of proof. The Chambers did not develop clear jurisprudence on this issue. It appears that basic notions of defendant's rights and the presumption of innocence demand that the burden in criminal trials should rest on the prosecution, even if IHL places the burden on the accused (usually the attacker). The burden of proof required by ICL must prevail over that required by IHL. This notion has also been reflected, in other contexts, in previous ICTY judgments. However, before Gotovina it was not clear exactly which aspects of IHL are affected by this burden shift, nor how they are affected. Unfortunately, the issue is no clearer following Gotovina.
There are a few practical lessons which future Chambers may learn from the Gotovina trials. These proceedings, and mainly the TC's impact analysis, are a good example of the fundamental importance, in analysing targeting issues, of knowing what information was available to the attacker at the time of the attack. The impact analysis was an attempt by the TC to bypass this legal requirement and to establish the intent of the attacker by other means. This is a crucial shift from relying on evidence about what happened before the attack, most likely to be found only in the attacker's possession, to evidence about what happened after the attack, which is usually held by the victims of the attack. This shift was rejected unanimously by the AC, although the rejection was never set out explicitly. This rejection, as well as the analysis of the honest mistake incident, re-emphasised the importance of proving intent and analysing targeting decisions based on the information that the attacker had at the time of the decision. This is true for all analyses of targeting decisions, and it is especially true in the context of criminal proceedings.
This is an important lesson to anyone who pronounces on issues of the law of targeting without the information that is essential for any legal analysis. This information is essential not only because of notions of fairness and natural justice, but also because it is legally required by the two branches of law that govern targeting decisions: IHL and ICL. While taking such information into account is important for the professionalism and integrity of NGOs or academics, it is absolutely crucial with respect to judicial and other official review mechanisms. The Gotovina proceedings remind scholars and practitioners that it is not legally possible for them to conduct a serious analysis by relying solely on what is, from a legal perspective, the less important half of the picture.
Finally, the Gotovina proceedings exemplify the great difficulties inherent in conducting criminal trials revolving around events that occurred during intense fighting, including issues relating to the law of targeting. Cases involving targeting issues in actual combat situations entail, in addition to various practical difficulties, extremely difficult legal and factual questions, such as those discussed in this article. Such cases therefore pose an exceedingly difficult dilemma for international prosecutors: whether to proceed with a targeting case knowing that the prospect of success is not high, or whether instead to avoid taking the case to begin with. It is not surprising, therefore, that most prudent international prosecutors do not initiate cases on the law of targeting. The ultimate result is the very low number of cases, nationally and internationally, dealing with those issues. This fact makes Gotovina's missed opportunities even more regrettable.