• Send page
  • Print page

Kosovo 1999: The air campaign -- Have the provisions of Additional Protocol I withstood the test?

31-03-2000 Article, International Review of the Red Cross, No. 837, by Peter Rowe

 Peter Rowe   is Professor of Law, Department of Law, University of Lancaster, UK.  

The air campaign began on 24 March 1999. It ended with the conclusion of the Military Technical Agreement between the International Security Force and the Governments of the Federal Republic of Yugoslavia and the Republic of Serbia of 9 June 1999.1 Between these two dates 10,484 strike sorties were flown by NATO aircraft and 23,614 munitions were released [2 ] . No NATO casualties were reported arising out of these strike sorties. The damage or destruction caused in Kosovo alone was in the order of “181 … in the tank category, 317 in the [armoured personnel carrier ] category, 600 in the military vehicles, and 857 in the artillery and mortars [categories ] .”3 In addition, many targets were attacked in other parts of the Federal Republic of Yugoslavia (FRY). This article will attempt to analyse the campaign from the standpoint of the obligations imposed by Additional Protocol I of 19774 and will draw upon the debates in the United Kingdom Parliament to illustrate how, in practical terms, the Protocol was viewed by those responsible for the actions of the armed forces of one of the NATO member States. [5 ]

The political and military objectives of the campaign were stated to be “to stop the killing in Kosovo and the brutal destruction of human lives and properties; to put an end to the appalling humanitarian situation that is now unfolding in Kosovo and create the conditions for the refugees to be able to return; to create the conditions for a political soluti on to the crisis in Kosovo based upon the Rambouillet agreement.” [6 ] These objectives were often stated, in short, to be to “have a major impact on Belgrade's war machine (…) to degrade its ability to carry out the current acts of violence in Kosovo (…) to attack, degrade, disrupt and further diminish the capacity of the Serb war machine to perpetrate these atrocities against its own people”. [7 ]

These objectives set a difficult task for international humanitarian law to control. This was largely because the conflict did not fit within those types of armed conflict envisaged by the drafters of Additional Protocol I or by those States which had become High Contracting Parties to it. There were no ground forces engaged in combat against each other, but only attacks from the sea, and more particularly from the air, against targets on land. The principal aim was, in reality, to compel FRY to remove its forces from Kosovo, a part of its own territory, after which the other political aims could be achieved. The conflict was, therefore, quite unlike any other. Denied the persuasive value of precedent in a situation where their air attacks were largely unopposed by enemy forces, NATO military commanders and politicians were forced to think through the consequences of each military action.

 A legitimate target  

The Protocol is quite clear on what are not legitimate targets, namely, civilians and civilian objects [8 ] . Military objectives, on the other hand, may be attacked. The problem is in defining such an objective. Thus, Article 52(2) limits military objectives to “those objects which by their nature, location, purpose or use make an effective contribution to military action and whose total or partial destruction, capture or neutralization, in the circumstances ruling at the time, offers a definite military advantage.” [9 ]

Arti cle 52(2) makes use of two phrases that may be said to be designed to limit military action. First, the object must make “an effective contribution” to military action and, secondly, its destruction etc. must offer “a definite military advantage”. In an armed conflict involving the armed forces of a State engaging the armed forces of another State these limiting phrases give rise to little difficulty, since it will not be difficult to show that an enemy's military equipment, or civilian objects used for military purposes, make an effective contribution to military action in some way or other and, if destroyed, would offer a definite military advantage. Thus, all military structures or equipment (whether used for attack or defence) or command and control centres would clearly be military objectives. Objects which have a dual use, such as bridges, railway lines, roads and so on, would also be military objectives providing they satisfy the two limitations set out above. [10 ]

There is, however, a danger of taking too academic an approach to what is essentially a practical test of legitimacy. It is, no doubt, for this reason that a number of States indicated on ratification of Protocol I that “the military advantage anticipated from an attack is intended to refer to the advantage anticipated from the attack considered as a whole and not only from isolated or particular parts of the attack.” [11 ]

How does all this apply to the Kosovo air campaign? Given the non-involvement of ground forces the air campaign was predicated on the political and operational objectives described above. There were substantial military assets and personnel within Kosovo during the conflict, but it is unlikely that the majority of these were involved in attacks upon the Kosovar Albanian population who were not taking a part in the hostilities [12 ] . NATO's objectives included the protection of this category of civilian from “the brutal destruction of human lives and propertie s” and, it must be assumed, not the protection of the KLA fighters. Nevertheless, since it was clear that the Kosovar Albanian population could not be protected effectively unless the FRY military presence was removed from the Kosovo region, this became the primary political goal. The situation can be likened (in fact but not in law) to the unlawful occupation of the territory of another State as in the Falklands/Malvinas conflict (1982) or the invasion of Kuwait by Iraq in 1990, where military campaigns were directed at removing the respective unlawful occupiers. There can be little doubt that FRY military assets were, therefore, legitimate targets since their destruction etc. would offer a definite military advantage if their presence in Kosovo was “removed”. Especially would this be the case where the attack is considered as a whole.

Members of the armed forces of FRY would be legitimate combatants who might also lawfully be attacked [13 ] . Whether police officers can be attacked depends upon whether they are members of a force that has been incorporated into the armed forces of a State [14 ] , or civilians who have taken a direct part in hostilities [15 ] . It is not clear whether the police forces who were active in Kosovo and who took part in actions against Kosovar Albanian civilians were members of a force that had been incorporated into the FRY armed forces. If they were not, the issue arises as to whether they took a “direct part in hostilities”. By using the term “hostilities” the Protocol clearly envisaged armed conflict between the armed forces of States. The FRY police were not attacking NATO forces but a particular group of their own fellow-citizens. They were therefore doing something which NATO forces wished to prevent. The difficulty, however, is that even if it could be argued that they were taking a “direct part in hostilities” by attacking members of the Kosovar population, they would only be legitimate targets while they were so acting. This is in marked contras t to members of the armed forces who may be attacked at any time simply because they have that particular status. It is possible that such a distinction might be made by individual members of ground forces had they been involved, but unrealistic to have expected NATO aircrew to be able to make this distinction from their operational height.

The difficulty with dual purpose objects such as bridges, roads, railway lines, power stations, broadcasting facilities and the like is to determine whether the potential targets of this type make an “effective [16 ] contribution to military action” and whether their destruction etc. would offer “a definite [17 ] military advantage”. The limiting effect of these words is particularly significant here. The said objects will obviously make an effective contribution to the ordinary commerce of life in general (including military activity), but this is not the same as saying that they make an effective contribution to military action. Nor will their destruction (taken as a whole) necessarily lead to a “definite military” advantage. The dangerous line of thinking is that merely because the military may make use of an object (and they are unlikely to do so unless it is necessary) that object becomes a military objective and, as such, a legitimate target. Were this argument to be taken to its logical conclusion, every civilian object that could possibly be used by the military would become a military objective. [18 ]

The limiting phrases referred to above try to prevent this forbidden line of reasoning. It is, however, still possible to argue that only the destruction of all forms of certain types of dual purpose object would make an effective contribution to military action and offer a definite military advantage. If there are, for instance, two bridges across a strategically significant river, the destruction of one only may give no military advantage; only the destruction of both would achieve this objective. Simil ar arguments can be presented for the destruction of all telecommunications centres and all power generating facilities,19 but they run into problems of direct or indirect collateral damage, discussed below.

 Collateral damage  

Attacks on military objectives may be indiscriminate if the injury or damage to civilians, civilian objects, or a combination of these, is expected to be excessive in relation to the concrete and direct military advantage anticipated [20 ] . Any other incidental injury or damage is considered to be collateral and one of the unavoidable consequences of armed conflict in modern times.

Incidental injury to civilians or to civilian objects may be direct or indirect. Direct injury or damage needs no further explanation, but indirect damage is often overlooked. To take an example, suppose all the electricity generating stations are destroyed by air attack to deprive the enemy military forces of all sources of power in bulk. Should this be achieved it may hamper the effectiveness, to some extent [21 ] , of the enemy military forces, but it is likely that it will also lead to the deaths of many civilians (particularly those most vulnerable on health grounds) for a variety of reasons, from the inability of hospitals to carry out their normal functions to sewage appearing on the streets and resultant disease [22 ] . Again, attacks that lead to pollution of the environment [23 ] may have a similar effect, as will the presence of unexploded weapons in civilian areas.

 Mistakes in targeting  

During the air campaign there were a number of mistakes made when an unintended civilian object was attacked. These ranged from the air strike on a train on 12 April 1999 and the bombing of a convoy near Prizren on 14 April [24 ] to the attack on t he Chinese Embassy on 5 May. Different reasons were given for each mistaken attack [25 ] . The Protocol is silent as to the effect of a mistake in these circumstances. Since the mistakes were ones of fact alone, the issue becomes that of whether the attack was against a military objective and -- if the facts had been as those who carried out the attack believed them to be -- whether the injury or damage to civilians or civilian objects was caused indiscriminately. [26 ]

 Warning of attack  

It is often thought that the obligation, in the 1907 Hague Regulations [27 ] , of a commander to warn the authorities before commencing a bombardment is a reflection of a past age of warfare [28 ] . It is, however, replicated in modern form in Article 57 of Protocol I. During World War II such warnings were given, “particularly in the case of objectives situated in occupied territory” [29 ] . In the air campaign in 1999 such warnings might have been given to enable the civilians located within a particular military objective, or close to one, to avoid the consequences of the attack. The often-quoted phrase “surprise in attack is the key to victory” does not have a great deal of significance if the attacking State has complete supremacy of the air, is virtually immune from the defensive measures of the attacked State and wishes, for political purposes, to avoid civilian casualties. Such a warning would, of course, be feasible only if the military objective was fixed, otherwise upon receipt of a warning it could be moved.

Moreover, any warning would have to be short-term to avoid essential equipment being removed from a building to be attacked. Whether such a warning would be effective in protecting the civilians working within the building concerned would turn upon whether the attacked State's authorities permitted them to leave. [30 ]

 The air campaign and the role of the UK Parliament  

No vote is required to be taken in the UK Parliament to commit British forces to an armed conflict, this being a decision made by the Government under the royal prerogative [31 ] . Ministers may, however, be questioned in either House and will make frequent statements as to the conduct of any particular conflict. The air campaign over FRY was no exception. There were arguments presented by members of both Houses of Parliament that the whole enterprise was unlawful under international law or that the actual campaign was being conducted badly (or illegally [32 ] ) but such views were expressed by only a very small minority. The reason for this is to be found more in the realms of psychology than in a detailed analysis of the legal position. When the armed forces are committed to a particular conflict it is easy to have independent judgment controlled by a desire to avoid appearing'disloyal'to those undertaking the fighting or to believe that it is the'duty'of Parliament to support them. [33 ]

It was the NATO attack on particular buildings in Belgrade, the mistaken attacks and the loss of civilian life that caused much debate. Insight is given in the various ministerial statements as to the reasons or explanations for such attacks. There is, however, no detailed analysis of the applicability of the 1949 Geneva Conventions for the protection of war victims nor of Protocol I to particular situations, merely a statement that “action by our forces is in strict conformity with international humanitarian law including the 1949 Geneva Conventions and their additional protocols ( sic )” [34 ] . An interesting feature of this particular conflict was that it was carried out by NATO and ministers were therefore required to make statements on the conduct of particular operations even though British forces might n ot have been involved.

The attack on the TV station on 23 April drew much comment. In the House of Lords the Minister stated that the reasons for attacking it were that “the media has been used to incite racial hatred and to mobilise the Serbs. Therefore, where media facilities are relevant to military operations or the capacity of Milosevic to continue his campaign of terror, they will be considered as possible targets… It is the propaganda machine that helps to keep the military machine moving.” [35 ] There was no analysis here of the limiting conditions contained in Article 52(2) of Protocol I, referred to above. Thus, did the TV station make “an effective contribution to military action” and would its destruction offer a “definite military advantage”? If so, how? It might well be argued that, like electricity generating stations and telecommunications centres, an attack on a TV station is likely to have little impact unless this is followed up by denying the enemy all, or at least a substantial amount, of these facilities. This is especially the case where the attack on a structure, such as a TV or radio station, is likely to involve the death of, or injury to, civilians.

Since the political objectives of the campaign included action to “stop the killing in Kosovo [and ] to put an end to the appalling humanitarian situation (…) unfolding”, the targets chosen for attack had to be selected to achieve this. One MP noted that a number of Kosovars were being picked out from groups of other civilians by FRY officials and he argued that he would extend targets to include “motor spare parts depots and major data recording departments (…) Yugoslavia's identity registration system (…) inland revenue and national insurance departments, the military pensions departments and the national records departments.” [36 ] There is a certain logic to this, since it would have made it much more difficult for FRY officials to isolate particular ethnic groups, bu t it would also be difficult to determine whether back-up copies of such documents were not stored in one or more computer systems and whether such action would offer a definite military advantage.

The principal concern of members of both Houses of Parliament was the killing of civilians as a result of the bombing raids. They wanted to be assured that as much care as possible was being taken to avoid such collateral damage and mistaken attacks. The Foreign Secretary told the House of Commons Defence Select Committee, before the air campaign began, that “a tiny number [of the 5,000 combat aircraft available to NATO ] have a capability for night-time bombing or for the kind of precision bombing which increasingly the legal authorities expect us to engage in.” [37 ] Events showed, however, that even with aircraft possessing these qualities, bombing from the air does result in unplanned damage and loss of civilian life. Indeed, the Gulf War 1990-91 had illustrated this. [38 ]

One decision that some argued increased the risk of such unwanted damage was the height at which NATO aircraft flew. It was asserted that by setting the height at 15,000 feet the politicians were guarding against the possibility of aircrew casualties and, in so doing, were transferring the risks to the civilian population. The Foreign Secretary assured the House of Commons that the decision to fly at this height was “entirely an operational matter for the military. Politicians would be extremely foolish to override the military on the height at which our pilots fly.” [39 ]

This leads to another issue, namely, who approved the targets to attack? Was it military commanders or the relevant government ministers? Given that this was a NATO action and the air forces of a number of nations were involved, the issue is likely to be less clear-cut than when a State is engaged alone in an armed conflict with another. The Secretary of State for Defence informed the Hous e of Commons that “I personally approve some of the targets, but for most, I have now delegated the decisions to the operational commanders. This allows them to make decisions quickly and to respond to changing requirements. However, I retain the ultimate authority and responsibility for those decisions.” [40 ] The corresponding Minister in the House of Lords gave a much more bland statement. She told the House that “all targets are selected and approved at the highest levels in NATO and that rigorous criteria are used to assess the suitability of any target, including its military utility and the risk of environmental damage or civilian casualties.” [41 ] The conclusion must be that government ministers were involved in the approval of some targets where the consequences of the bombing might give rise to political implications, otherwise the military commanders were left to make the decisions by themselves. The point is a practical one, since in an armed conflict involving the use of air power alone the decision as to whether to attack a particular target is likely to be taken by a high-ranking military commander or a senior government minister. It is, therefore, quite unlike a land campaign where decisions to use weapons are most often taken by the lowest, or lower, ranks. In the circumstances of an air campaign alone it is much more likely that legal advice will be available to a high-ranking commander and that the precautions required prior to an attack can, and should, be taken.

A further concern expressed in Parliament was the effect of using cluster bombs and depleted uranium weapons by NATO forces. The Foreign Secretary assured the House of Commons Select Committee on Foreign Affairs that neither anti-personnel cluster bombs nor depleted uranium weapons were being used. [42 ]

Who is liable if breaches of international humanitarian law occur as a result of the bombing campaign in circumstances where the government minister has devolved decisions on targeting to military commanders? Clearly the latter will be, but what of the former? He or she has not ordered the attack to take place but is unlikely to be held personally responsible unless he or she was actually, or should have been, aware of the circumstances. [43 ]

Finally, a member of the House of Lords suggested that some of the NATO actions amounted to the “murder” of innocent civilians44 and thus, by inference, national law also applied to govern the actions of those who carried out the bombing raids. This is not, however, the position under English law where, at the very least, liability for murder or manslaughter is coextensive with a killing amounting to a breach of international humanitarian law.

 Did Protocol I add anything?  

It may seem an extreme position to conclude that Protocol I had little impact or influence upon the conduct of this particular air campaign, yet it might be argued that all the detailed rules so carefully drafted in 1977 were of little consequence. If the Protocol had never come into existence, would the air campaign have been conducted any differently? The answer, I regret to say, is likely to be in the negative. The customary international law rules of proportionality and those relating to the distinction between civilian objects and military objectives [45 ] would have been perfectly adequate. For the reality of the situation is that those objects which military commanders wished to attack, for whatever reasons, were attacked. There is no evidence that the rigours of the limitations imposed on the definition of a military objective were applied to attacks against dual purpose objects. In any event, the Protocol is, when it comes to the test, very weak in determining what may and what may not be attacked. There is no argument over attacking military personnel and their equi pment. It is when civilians are most likely to be placed in danger that Protocol I, designed to protect them, shows its faults.

It is, for instance, common for those who make the targeting decisions to exaggerate (albeit unwittingly) the contribution to military action and the military advantage to be gained from a particular bombing mission and, in addition, to expect the “smart” weapons to produce precision bombing (and thus leave civilians untouched) [46 ] . The reason for this is that Protocol I sets the dividing line between an indiscriminate attack (illegal) and one that is not (legal) on the basis of expectation before the attack is commenced. At this stage of military operations those planning the attack are at their most optimistic and civilians are at most risk. The position has been reached where the well-established principle of distinction between civilian objects and military objectives has been rendered of little practical significance when some military advantage in the destruction of the object is discernible and smart weapons are used. Add to this a political objective of attempting to persuade a foreign government to take a particular action within its own territory, and it becomes much easier to argue that anything that affects that government's actions becomes a military objective, wherever it is situated.

 Is reform of international humanitarian law necessary?  

Those who formulate international humanitarian law, particularly in treaty form, must always guard against the dangers of creating unrealistic expectations in times of armed conflict. In practical terms the aerial (or artillery) bombing of objects in heavily populated cities carries with it much greater responsibility to avoid civilian casualties than does a similar discharge of high explosives in a battle area, given that in such an area civilians are unlikely to be present. Although Part IV of the Protocol refers to the protection of the civilian population and civilian objects, protection of the former is clearly more important than of the latter [47 ] . The protection of civilians during an armed conflict must therefore be the primary restraint upon military action.

How is this restraint to be achieved? Articles 51 and 57 of Protocol I are fine so far as they control “crude” methods of bombing, but they do not deal effectively with attacks using smart weapons, targeted at an alleged military objective.

The “do-nothing” option does not seem appropriate since the increased use of “smart” and “not so smart” weapons is clearly foreseeable [48 ] , if only because it appears to be more acceptable to bomb targets within a city by using such weapons than by using free-fall bombs. For the reasons stated above, in particular the relative ease with which a State can convince itself not only that a selected target is a military objective, but also that mistakes in targeting are rare when compared with the number of bombing raids or that the enemy's actions are much worse, it seems not unreasonable for the law to take into account the development of these smart weapons. Moreover, these “smart” weapons can go “dumb”, as failures in their guidance systems are not uncommon. If the political and military leaders of all States were asked the question “do you wish to prevent as many civilian casualties as possible during an international armed conflict,” a substantial majority would be likely to answer in the affirmative. As argued above, it seems clear that the law, in Part IV of Protocol I, does not achieve this.

Proposals to amend the definition of a military objective in Article 52(2) or the definition of an indiscriminate attack in Article 51 are unlikely to succeed, since any reformulation would create as many definitional problems as exist in the current law. These articles may be left to deal wit h what may be styled as the normal run of military operations. This suggests that an additional norm is needed to deal with the problem addressed in this paper, namely, the protection of the civilian population from air-delivered “smart” weapons. The fact that relatively few civilians will be killed or injured in these circumstances (compared with an indiscriminate attack using free-fall bombs or poorly guided missiles) should not affect the legal regime involved. One possibility, it is suggested, may be to add an additional protocol to the 1981 Conventional Weapons Convention and draw upon the analogy of Article 2(3) of Protocol III to that Convention. Thus, it could be provided that:

It is prohibited to make any military objective located within a concentration of civilians the object of attack, except when such military objective is clearly separated from the concentration of civilians, and all feasible precautions are taken with a view to limiting the effects of the attack to the military objective and to avoiding, and in any event to minimizing, incidental loss of civilian life, injury to civilians and damage to civilian objects. [49 ]

The term “concentration of civilians” could be given the meaning found in that Protocol, as “any concentration of civilians, be it permanent or temporary, such as in inhabited parts of cities, or inhabited towns or villages, or as in camps or columns of refugees or evacuees, or groups of nomads”. Should an attack be carried out within a concentration of civilians the key concept would then be to determine whether all “feasible precautions” had been taken to avoid the prohibited damage. Given the Oxford English Dictionary definition of “feasible” as “practicable, possible, easily or conveniently done” [50 ] , an attacking State which did not take such steps would be in breach of the provision. The steps required to be taken would clearly depend upon the circumstances existing at the time, but a planner would have to consider various possibilities, such as the guidance system malfunctioning, a mistake as to the proposed target itself and whether the attack should be carried out only when the objective can be clearly identified. The possibility of anti-aircraft missiles, or other defensive measures, will also need to be considered, as well as the height from which an attack should be launched. There is no balancing act here, as there is in Article 51(5)(b) of Additional Protocol I, between destroying the military objective and the expectation of civilian deaths or injury or the destruction of civilian objects. The obligations upon the attacking State are therefore much greater under this proposed addition to the law. It would also be necessary to provide that a breach of this provision would be classed as a war crime under the Rome Statute of the International Criminal Court (1998) [51 ] for which an individual might be indicted.

An alternative approach is to tackle the problem from the standpoint of the State's responsibility to pay compensation to civilians (or to those claiming on their behalf) for injury and damage to them caused by a failure to take all feasible precautions to prevent such loss or damage by those who, acting on behalf of the State, have attacked targets within an area where there is a concentration of civilians. The liability of a State to pay compensation is, of course, well settled although the mechanisms for doing so are not [52 ] . Like criminal responsibility, any obligation to pay compensation must be incumbent upon all States involved in an international armed conflict and not merely the defeated State. It should also be awarded by an independent body whose deliberations are open to the public, so that the accountability of military or political leaders for actions or omissions during an armed conflict may be transparent, as it is in crimi nal cases. To achieve this objective, international criminal courts would need to have jurisdiction to prosecute States, the only punishment available in this type of case being the payment of compensation to the victims of the military action.

The two proposals made above require an alteration to existing treaty arrangements, which may be considered, at the present time, to be unrealistic of achievement. At the very least States should be encouraged to set out in a public document the grounds by which they assert that their actions are in conformity with international humanitarian law. In particular, where an air campaign is mounted, an expectation that a State will assert why a particular object is a military objective will heighten the awareness of those who make decisions about individual targets upon the fact that their actions will be judged by the law. This is no new phenomenon, since in any democratic State a government's actions are normally limited by legal boundaries. These boundaries are to be found mainly in Additional Protocol I or in customary international law.


1. See also Security Council Resolution 1244 of 10 June 1999.

2. Lord Robertson, Kosovo: An Account of the Crisis, Ministry of Defence, London, 1999.

3. Brigadier-General Corley at NATO press conference on the Kosovo Strike Assessment, 16 September 1999.

4. Protocol additional to the Geneva Conventions of 12 August 1949, and relating to the Protection of Victims of International Armed Conflicts (Protocol I), 8 June 1977.

5. The UK carried out 15.4% of all NATO strike sorties but it released only 4.2% of all air-delivered munitions (author's analysis of figures to be found in op. cit. (note 2)).

6. Secretary-Genera l of NATO at NATO press conference, 1 April 1999. See also the Foreign Secretary, H.C.Debs. Vol. 331, col. 887, 18 May 1999; Vol. 329, col. 665, 19 April 1999; Joint Statement on the Kosovo After Action Review, presented by the Secretary of Defense William S. Cohen and General Henry H. Shelton, Chairman of the Joint Chiefs of Staff, before the United States Senate Armed Forces Committee, 14 October 1999; Shea, Briefing on Kosovo, NATO, 30 April 1999 -- none of these statements are identical.

7. General Wesley K. Clarke, Supreme Allied Commander Europe (SACEUR), at NATO press conference, 1 April 1999.

8. Protocol I, Art. 48, combined with the other provisions of Part IV.

9. For a discussion of the history of this concept see Y. Sandoz, C. Swinarski and B. Zimmermann (eds.), Commentary on the Additional Protocols of 8 June 1977, ICRC/ Martinus Nijhoff, Geneva, 1987, and for an attempt to draw up indicative categories of military objectives see the ICRC Draft Rules for the Limitation of Dangers Incurred by the Civilian Population in Time of War, 1956,
Art. 7, reprinted ibid., p. 632. See also Art. 8(1)(a) of the 1954 Hague Convention for the Protection of Cultural Property in the Event of Armed Conflict.

10. See D. Fleck (ed.), The Handbook of Humanitarian Law in Armed Conflicts, Oxford University Press, 1995, p. 160 (by Stefan Oeter).

11. Reservation (i) entered by the UK upon ratification of Protocol I, 28 January 1998, see IRRC, No. 322, March 1998, p. 186 ff., and www.icrc.org/ihl. -- The German reservation is identical, see Fleck, op. cit. (note 10), p. 162.

12. E.g., the Kosovo Liberation Army (KLA).

13. Protocol I, Art. 43(2).

14. Protocol I, Art. 43(3).

15. Protocol I, Art. 51(3). See C. de Rover, “Police and security f orces”, IRRC, No. 835, September 1999, p. 638.

16. For the history of the inclusion of this adjective, see M. Bothe, K. J. Partsch and W. Solf, New Rules for Victims of Armed Conflicts, Martinus Nijhoff, The Hague/ Boston/London, 1982, pp. 325/6.

17. The Oxford English Dictionary defines “definite” as “clear and distinct, not vague”.

18. Note the comment of the Defence Secretary in the House of Commons: “Milosevic set out to create a military machine in which there was an intimacy between his military and every other aspect of society.” H.C.Debs. Vol.304, col. 364, 26 May 1999.

19. Op. cit. (note 10), p. 161.

20. Protocol I, Art 51(5)(b). This should of course be taken into account in planning attacks, see Art. 57(2)(a)(iii).

21. Military forces operating in the field will have their own electricity generating equipment.

22. See the speech of Mrs Mahon, MP, who stated that “In Novi Sad, which was bombed continuously, the sewage works are damaged beyond repair and the drinking water has gone. That will cause health problems for a long time.” H.C.Debs. Vol. 333, col. 610, 17 June 1999.

23. For concern voiced by MPs see the speech of Mr Dalyell, MP, who expressed the view that “The environmental consequences of bombing know no borders.” H.C. Debs. Vol. 330, col. 893, 5 May 1999. Mrs Mahon, MP, concluded that “the bombing of petrochemical works and nitrate fertiliser plants in Pancevo and other areas has released considerable quantities of toxic chemicals into the Danube and into the atmosphere.” loc. cit. (note 22).

24. Brigadier-General Leaf at NATO press conference, 19 April 1999.

25. In bombing the Chinese Embassy NATO believed that it was attacking the Federal Directorate of Supply and Procurement for
the Yugoslav Army. See H.L.Debs. Vol. 600, col. 989, 10 May 1999. British Ministers explained such mistaken attacks as follows: “I must be blunt with the House; one cannot wage a military campaign of this intensity without mistakes. There have been 6,000 bombing sorties over the past seven weeks, and very few have resulted in the wrong target being attacked.” The Foreign Secretary, H. C. Debs. Vol. 331, col. 26, 10 May 1999. Alternatively, a Government Minister may refer to the “evil actions of the enemy” as a justification. See the Minister's reply, when questioned by a member in the House of Lords: “I have in front of me an appalling and sickening document. It is a chronology of the atrocities committed in Kosovo by Serbian forces … I hope that he [the member ] will study it as carefully as I have studied the protocol that he sent me.” H. L. Debs. Vol. 601, col. 149, 18 May 1999.

26. This is the effect of the word “expect” in Art. 51(5)(b) and 57(2)(a)(iii). In addition, all feasible precautions to avoid prohibited attacks would need to have been taken, in the circumstances ruling at the time (Arts. 52(2) and 57).

27. Art. 26, 1907 Hague Convention IV.

28. See in general T. Meron, War Crimes Law Comes of Age, Clarendon Press, Oxford, 1998, chapters 1-5.

29. Op. cit. (note 9), p. 686. For a modern discussion, see R. K. Goldman, “The legal regime governing the conduct of Operation Desert Storm”, University of Toledo Law Review, Vol. 23, 1992, p. 384.

30. The obligation to remove civilians from the vicinity of a military objective is contained in Art. 58(1), Protocol I. National law may, however, be at variance with this.

31. Mr Corbyn, MP, stated that “60 days into the bombardment, there has been no substantive vote in the British Parliament on the authority of the Government to proceed with the bombardment of Yugosla via.” H.C.Debs. Vol. 331, col. 923, 18 May 1999.

32. See, e.g., Lord Jenkins of Putney, who stated: “Is not the whole operation contrary to what is laid down in the convention [Proto-
col I ] , and particularly in Articles 51 and 52?” H.L.Debs. Vol. 601, col. 148, 18 May 1999.

33. “I deplore the criticisms of the chairborne aces who are so critical of the actions of NATO's air and ground crews.” Mr Wilkinson, MP, H.C.Debs. Vol. 923, col. 955, 18 May 1999. Compare, however, Mr. Clark, MP (a critic of the bombing campaign): “I wholly reject any suggestion that to say these things is unpatriotic in some way.” H.C.Debs. Vol. 331,
col. 923, 18 May 1999.

34. The Parliamentary Under-Secretary of State, Foreign and Commonwealth Office, H.L.Debs. Vol. 600, col. 1000, 10 May 1999, and Vol. 601, col. 148, 18 May 1999.

35. H.L.Debs. Vol. 600, col. 38, 41, 26 April 1999.

36. Mr Campbell-Savours, MP, H.C.Debs. Vol. 331, col. 943, 18 May 1999.

37. 17 February 1999, Q. 318.

38. See F. Hampson in P. Rowe (ed.), The Gulf War 1990-91 in International and English Law, Routledge and Sweet & Maxwell, London, chap. 5, pp. 95-100.

39. H.C.Debs. Vol. 331, col. 864, 18 May 1999. Some aircraft “operated down to 6,000 feet when target identification or weapons delivery profile required it.” Op. cit. (note 2).

40. H.C.Debs. Vol. 329, col. 667, 19 April 1999.

41. H.L.Debs. Vol. 600, col. 907, 6 May 1999. Note that the Royal Air Force “refused at least twice to bomb targets given it by American commanders during the Gulf war because the risk of collateral damage (…) was too high.” Tenth Report of the House of Commons Select Committee on Defence, 1991, H. C. 287/1, p.38.

42. Seventh Report, Kosovo: Interim Report, H.C. 188, 20 July 1999, Q. 156. Letter to the Chairman of the Committee from the Secretary of State for Foreign and Commonwealth Affairs, 26 April 1999, ibid., p. 74. Cluster bombs were distinguished by the Minister from anti-personnel mines banned by the 1997 Ottawa Treaty, H.C.Debs. Vol. 304, col. 370, 26 May 1999.

43. Art. 86(2), Protocol I. The term “superiors” is wider than “commanders” and “should be seen in terms of a hierarchy encompassing the concept of control.” Op.cit. (note 9).

44. H.L.Debs. Vol. 600, col. 865, 6 May 1999.

45. See the 1868 St. Petersburg Declaration Renouncing the Use, in Time of War, of Explosive Projectiles under 400 Grammes Weight, preambular para.: “the only legitimate object which States should endeavour to accomplish during war is to weaken the military forces of the enemy.”

46. It is consequently unlikely that a breach of international humanitarian law took place, since Art. 85(3)(a) or (b) of Protocol I would not apply.

47. The protection of civilian objects in Part IV is essentially centred on the protection of the civilian population, as indeed is Protocol III to the 1981 Conventional Weapons Convention. For attempts to protect the environment, and thus the civilian population, see R. Falk in G. Plant (ed.), Environmental Protection and the Law of War, Belhaven Press, London/New York, 1992, pp. 91-95.

48. See The Times, 23 September 1999, interview with Lord Robertson: “the public (…) expect precision attacks. So, in terms of future procurement, nations will have to buy weapons that are relevant.”

49. This is more specific than Art. 57(2)(ii), Protocol I, since it deals with attacks where there is a concentration of civilians.

50. On ratification of the Additional Protocols on 28 January 1998 the UK stated that it understood “feasible” to mean “that which is practicable or practically possible, taking into account all circumstances ruling at the time, including humanitarian and military considerations” loc. cit. (note 11), statement (b).

51. At present it is not. See Art. 8 of the Rome Statute of the International Criminal Court, UN Doc. A/CONF. 183/9.

52. See H. P. Gasser, “International humanitarian law”, in Hans Haug, Humanity for All, Henry Dunant Institute, Geneva, 1993, pp. 87-88.

Abstract in French