(zb) cassese ch21

21SUPERIOR RESPONSIBILITY

Kai Ambos

123456789101112131415161718192021222324252627282930313233343536373839404142

I am grateful to Emily Silverman, J.D. (Berkeley), LL.M. (Freiburg) for refining the English versionand for her critical comments; as well as to Christian Steiner and Katrin Schiefer, both of the MaxPlanck Institute for Foreign and International Criminal Law, for technical assistance.

805

I. Introduction 806II. The Status of Superior Responsibility

in Customary Law 806A. International and National

Jurisprudence 8071. The Yamashita Case: The Difficult

Debut of the Superior Responsibility Doctrine 807

2. Other Post-World War II Jurisprudence 810

(a) Nuremberg 810(b) IMTFE-Tokyo 812(c) UNWCC 813(d) Other National Trials for

War Crimes 8143. The Jurisprudence of the Existing

Ad Hoc Tribunals 815B. Efforts of Codification 819

1. Protocol Additional I to the Geneva Conventions 819

2. Other 821C. Comparative Aspects 823

1. Military Law 8232. Related Aspects of Economic

Criminal Law 825D. Preliminary Conclusions 828

III. Article 28 Rome Statute 830A. Description of the Provision and

Comparison with Earlier Codifications 830

B. Critical Analysis and Interpretation 8321. General Considerations 8322. Actus Reus 836

(a) The Status of the Superior 836(b) Effective ‘command and

control’ or ‘authority and control’ 839

(c) Crimes as a ‘result’ of the Superior’s Failure to Exercise Control Properly 841

(d) ‘Necessary and reasonable’ Countermeasures ‘within his or her power’ 843

(e) ‘Prevent’, ‘repress’, or ‘submit to the competent authorities’ 844

3. Mens rea 845(a) Military and Non-Military

Superior: Positive Knowledge 845

(b) Military Superior: ‘should have known’ 846

(c) Non-Military Superior: ‘consciously disregarded information which clearly indicated . . .’ 851

C. Conclusion 852Select Bibliography 853

(Zb) Cassese Ch21 15/2/02 9:00 am Page 805

I. Introduction

The concept of command or better superior responsibility1 makes the superiorliable for a failure to act to prevent criminal misconduct of his or her subordinates.The superior is punished for a lack of control and supervision of his or her subor-dinates who commit crimes. Thus, the superior is punished both for his or herown failure to intervene and for the crimes of others. As a result, the concept seemsto create, on the one hand, direct liability for the lack of supervision, and, on theother, indirect liability for the criminal acts of others. Finally, as will be explainedbelow, superior responsibility has a double character: it is a genuine offence ofomission (echtes Unterlassungsdelikt) and an offence which creates danger(Gefährdungsdelikt).

These rather theoretical considerations quickly take on practical importance ifone looks, for example, at the evidentiary problems encountered in cases of supe-rior responsibility: first, it must be proved that the superior was in a position ofcommand and control which would have enabled him or her to prevent the crimesof his or her subordinates. Second, the superior must have known of these crimesor, at least, have deliberately remained ignorant of them. The latter issue is partic-ularly controversial.

This paper will first show that the doctrine of superior responsibility is recognizedin customary international law (II) as evidenced by the case law (II.A), codificationefforts (II.B), and comparative law (II.C). Analysis of these sources will also show,however, that the elements of the doctrine are far from clear and uniform. Thus, themain part of the paper is devoted to clarification and refinement of these elements,as established by Article 28 of the Rome Statute (III). Finally, this part of the papercontains suggestions for interpretation and, where necessary, improvement.

II. The Status of Superior Responsibility in Customary Law

Customary law within the meaning of Article 38(1)(b) of the ICJ Statute followsfrom a general State practice, which is the expression of an opinio iuris vel necessi-tatis.2 By analysing codifications, international and national jurisprudence and

General Principles of International Criminal Law

1 Whereas the term ‘command’ seems to limit the doctrine to a military context, ‘superior’ is abroader term extending to civilians as well. As such it was adopted by the Rome Statute. See infra,III.B.2(a); see also W. J. Fenrick, ‘Some International Law Problems Related to Prosecutions beforethe International Criminal Tribunal for the Former Yugoslavia’, 6 Duke Journal of Comparative andInternational Law (1995) at 110 fn. 21.

2 Cf. R. Bernhardt, ‘Customary International Law’, EPIL 1 (1992) 898 et seq.; I. Brownlie,Principles of Public International Law (5th edn., 1998) 4 et seq.; A. Verdross and B. Simma,Universelles Völkerrecht (4th edn., 1984) §§ 549 et seq.

806


comparative law, including doctrine, one may identify custom although compar-ative law can also be an expression of general principles within the meaning ofArticle 38(1)(c) of the ICJ Statute.

The origins of the superior responsibility doctrine can be traced back to the fifteenth century.3 However, the following survey only addresses developmentsoccurring after World War II. It was not until the post-war period that the doc-trine was clearly linked to criminal responsibility4 and obtained a more precisemeaning which ultimately served as the basis of the Rome Statute.

A. International and National Jurisprudence

1. The Yamashita Case: The Difficult Debut of the Superior ResponsibilityDoctrine

The trial against General Tomoyuki Yamashita of the Japanese Army producedthe first judgment based on the superior responsibility doctrine.5 Yamashita tookcommand of the Japanese army in the Philippines on 9 October 1944. He trans-ferred his headquarters to the mountains of Baguio, 125 miles north of Manila, inDecember 1944. By 4 February 1945 US troops reached the capital of Manila,and by 3 March all Japanese naval forces who tried to defend the city had beenkilled. During the fighting the Japanese forces tortured and murdered thousandsof the city’s 700,000 civilian inhabitants.6 Other atrocities were committed in theBatangas Province and elsewhere on Luzon.7 During the defence of Manila,General Yamashita was in Baguio and, according to his defence counsel, knewnothing of the fighting since his communications were cut off. His initial order toevacuate Manila encountered resistance from both the Japanese army and thenavy. Only the 1,600 army troops left the city in January 1945 when ImperialHeadquarters put them under Yamashita’s direct command. The 20,000 navaltroops remained in the city.

123456789101112131415161718192021222324252627282930313233343536373839404142

Superior Responsibility

3 Cf. W. H. Parks, ‘Command Responsibility for War Crimes’, 62 Military Law Review (1973)2 et seq.; L. C. Green, ‘Command Responsibility in International Humanitarian Law’, 5Transnational Law & Contemporary Problems (1995) 320 et seq.; M. C. Bassiouni, Crimes againstHumanity in International Criminal Law (2nd edn., 1999) 423 et seq. See also H. S. Levie,‘Command Responsibility’, 8 Journal of Legal Studies (1997/98) at 1, referring to Grotius.

4 Cf. Parks, supra note 3, 19.5 The following factual information is based on A. M. Prévost, ‘Race and War Crimes: The 1945

War Crimes Trial of General Tomoyuki Yamashita’, 14 HRQ (1992) 309 et seq., B. D. Landrum,‘The Yamashita War Crimes Trial: Command Responsibiliity Then and Now’, 149 Military LawReview (1995) 293 et seq.; and C. N. Crowe, ‘Command Responsibility in the Former Yugoslavia:The Chances for Successful Prosecution’, 29 University of Richmond Law Review (1994) 194–195.

6 Parks, supra note 3, 25 and Landrum, supra note 5, 295, refer to over 8,000 killings and 7,000mistreated or wounded; the latter also cites 500 rapes. Prevost, supra note 5, 314, and Crowe, supranote 5, 195, refer to nearly 100,000 killings.

7 Landrum, supra note 5, 294–295.

807


On 25 September 1945 Yamashita was detained by US forces and charged of hav-ing ‘unlawfully disregarded and failed to discharge his duty as commander to con-trol the operations of the members of his command, permitting them to commitbrutal atrocities and other high crimes against the people of the United States andof its allies and dependencies, particularly the Philippines’.8

The trial before a US military commission began on 29 October and ended on 7December 1945 with a death sentence for the defendant. The Commission drewno express conclusion of knowledge but found only that there had been wide-spread atrocities and that Yamashita had ‘failed to provide effective control . . . aswas required by the circumstances’.9 Yamashita appealed first to the SupremeCourt of the Philippines, which declared itself incompetent ‘to interfere with actsof the U.S. Army . . .’.10 The US Supreme Court accepted the petition for writ ofhabeas corpus but restricted its finding—on the basis of its strict theory of judicialrestraint as applied in previous war crimes cases—to mere procedural questions,stating that the ‘detention of petitioner for trial . . ., subject to the prescribedreview by the military authorities, were lawful, and . . . the petition for certiorari,and leave to file in this Court petitions for writs of habeas corpus and prohibitionshould be, and they are Denied ’.11

With regard to Yamashita’s right to a fair trial according to the Fifth Amendmentof the US Constitution, the Court avoided a substantive decision, declaring itselfincompetent to review this question.12 As to the substantive basis of the prosecu-tion’s case, the Supreme Court first recognized that the crimes in question consti-tuted ‘violations of the law of war’.13 These violations were to be attributed to thedefendant by the superior responsibility doctrine because:

. . . the law of war imposes on an army commander a duty to take such appropriatemeasures as are within his power to control the troops under his command for theprevention of the specified acts . . . [and] . . . he may be charged with personal respon-sibility for his failure to take such measures when violations result. . . .

. . . Hence the law of war presupposes that its violation is to be avoided throughthe control of the operations of war by commanders who are to some extent respons-ible for their subordinates.14

Thus, the Court held that ‘the allegations of the charge, tested by any reasonablestandard, adequately allege a violation of the law of war . . .’.15


8 Quoted according to 327 US 13–14 (1945) (emphasis added).9 Cf. 327 US at 51 (1945).

10 Quoted according to Prévost, supra note 5, at 320.11 327 US at 25–26 (1945) (emphasis in the original).12 Ibid., at 18 et seq. (23): ‘we hold that the commission’s ruling on evidence and on the mode of

conducting these proceedings against petitioner are not reviewable by the courts, but only byreviewing military authorities.’

13 Ibid., at 14, see also at 18.14 Ibid., at 15 (see also at 2).15 Ibid., at 17–18 (Justice Murphy dissenting at 31).

808


The trial and the judgment received strong criticism in judicial and academic cir-cles,16 but they were also defended.17 The criticism is summarized in the followingexcerpt from Prévost’s recent study:

It is a basic rule of law that criminal sanctions should not be imposed on anyonewithout some showing of culpability on their part. However, Yamashita was con-victed and hanged for a crime for which there was no showing of culpability, in a trialwhich eliminated basic Constitutional guarantees. Nonetheless, this study hasshown that even under this strict liability theory of command responsibility,Yamashita should not have been convicted. In addition, this study has demonstratedthat the conviction and the Supreme Court decision which followed were influencedin large part by racism.18

Although one should not go so far as to negate any precedential value of theYamashita decision,19 it cannot be denied that its mixture of technical-legal flaws,as particularly criticized in Jude Rutledge’s dissent,20 and ideological-racial preju-dice, as recently demonstrated by Prévost’s study, severely hampers its legal and,above all, moral value. As to the superior responsibility doctrine itself and theprinciple of personal guilt,21 one cannot underestimate the importance of the fol-lowing criticism expressed by Supreme Court Justice Murphy (dissenting):

. . . there was no serious attempt to charge or to prove that he [Yamashita] commit-ted a recognized violation of the laws of war. He was not charged with personally par-ticipating in the acts of atrocity or with ordering or condoning their commission.Not even knowledge of these crimes was attributed to him. It was simply alleged thathe unlawfully disregarded and failed to discharge his duty as commander to controlthe operations of the members of his command, permitting them to commit the acts

123456789101112131415161718192021222324252627282930313233343536373839404142


16 Cf. Prévost, supra note 5, at 317–319, 325–329; Bassiouni, supra note 3, 426 et seq.; Landrum,supra note 5, 297, 300; C. A. Hessler, ‘Command Responsibility for War Crimes’, 82 Yale LawJournal (1973) at 1283; see also the positive book review of R. L. Lael’s, The Yamashita Precedent(1982) by R. W. Branstetter in 98 MLR (1982) 159 et seq.; critically also W. O’Brien, ‘The Law ofWar, Command Responsibility and Vietnam’, 60 Georgetown Law Journal (1972) at 637; M. Lippman, ‘Conundrums of Armed Conflict: Criminal Defenses to Violations of theHumanitarian Law of War’, 15 Dickinson Journal of International Law (1996) 1–111, at 75–76 andW. A. Schabas, ‘The General Principles of the Rome Statute’, 6 Europrean Journal for Crime,Criminal Law, and Criminal Jusstice (1998) 400–428, at 417.

17 See esp. Parks, supra note 3, 22 et seq., 87 et seq., who, inter alia, rejects a strict liability inter-pretation. Similarly Crowe, supra note 5, at 198–99, refers to the ordering of atrocities byYamashita. See also Levie, supra note 3, passim.

18 Prévost, supra note 5, at 337 (emphasis in the original). See also the strong criticism in JusticeMurphy’s dissenting opinion of, 327 US at 27 et seq.; for criticism of this opinion, see Levie, supranote 3, in fn. 13.

19 In this sense Prévost, supra note 5, at 338.20 327 US at 41 et seq., e.g. p. 53 with regard to the admission of ‘untrustworthy, unverified,

unauthenticated evidence’ or p. 61 with regard to denial of fair trial (‘departure from the most ele-mentary principles of fairness’).

21 This was already recognized by the IMT as ‘one of the most important [legal principles]’;therefore, ‘the Tribunal should make such declaration of criminality [of an organization or group]so far as possible in a manner to ensure that innocent persons will not be punished’ (The Trial of theMajor War Criminals (1950) Vol. 22, at 469).

809


of atrocity. The recorded annals of warfare and the established principles of inter-national law afford not the slightest precedent for such a charge. This indictment ineffect permitted the military commission to make the crime whatever it willeddependent upon its biased view as to the petitioner’s duties and his disregard thereof,a practice reminiscent of that pursued in certain less respected nations in recentyears.22

Although actual knowledge was not proven, it cannot be said that Yamashita’sconviction was based on a theory of strict liability.23 Instead, the Commissionseemed, on the one hand, to disbelieve Yamashita’s plea of ignorance, given theextensiveness of the atrocities24 and, on the other, to have doubts as to what stan-dard of knowledge should be applied. Thus, it is impossible to say whether theCommission actually believed that Yamashita knew or should have known of theatrocities.25 According to Hessler’s interpretation, the Commission and theSupreme Court were convinced that Yamashita’s ‘total ignorance . . . raised unac-ceptable general risks of future subordinate criminality, those crimes becomingchargeable to him as soon as they occurred.’26 Thus, although Yamashita did notknow of the atrocities, his ignorance created risks attributable to him. This ‘risk-oriented approach’ is confirmed by the general structure of the superior responsi-bility offence, as will be seen below.27 In any case, the codifications examinedabove and the post-Yamashita case law show that the broad liability standard setby Yamashita has not been followed; in particular the US military jurisprudencedid not apply this standard in the proceedings for the My Lai massacre.28

2. Other Post-World War II Jurisprudence29

(a) NurembergWhile the IMT was only concerned with the superior responsibility doctrine ‘inthe most indirect fashion’,30 the doctrine was invoked in three trials before the USmilitary courts at Nuremberg. In US v. Pohl et al. the tribunal referred explicitly toYamashita and stated: ‘The law of war imposes on a military officer in a positionof command an affirmative duty to take such steps as are within his power and


22 327 US at 28.23 Cf. Parks, supr note 3, at 37, 87–88.24 In this sense Crowe, supra note 5, 206–07; also Levie, supra note 3, 5, arguing that Yamashita’s

knowledge of the atrocities cannot be denied.25 See also the well-taken criticism of Crowe, supra note 5, 207–208.26 Hessler, supra note 16, at 1283.27 See infra, C and III.B.1.28 See infra, 2(d), note 55 and accompanying text. See also Landrum, supra note 5, at 300: ‘Even

in United States courts, Yamashita has lost favor.’29 The Leipzig Trial cannot be adduced in support of the command responsibility doctrine. The

German Reichsgericht did not even know this doctrine and only judged the defendants on the basisof the ordinary rules of participation as laid down in the Strafgesetzbuch (this is overlooked by Levie,supra note 3, 2–4)

30 Green, supra note 3, 333.

810


appropriate to the circumstances to control those under his command for the pre-vention of acts which are violations of the law of war.’31

Apart from the liability for omission which remained confined to military superi-ors, the case law extended the doctrine to civilian superiors. In Pohl et al., forexample, one of the convicted defendants, Mummenthey, was a civilian adminis-trator.32 US v. Brandt et al. also recognized the superior responsibility of civilians.33

With regard to aggressive war, the case law required that the Defendant be a‘leader’ or ‘planner’34 or belonged to the ‘policy level’.35

In the Hostage case,36 the court argued that a command position implies certainduties of control and supervision with regard to the area of competence. Ifoffences are committed within this area of competence, the responsible officer hasto intervene. As to the mental element, the court stated that a commander is‘obliged to know’37 if he receives credible information about atrocities committedby his subordinates. Thus, actual knowledge was not required; instead, a should-have-known standard was applied. It differed from Yamashita in that the objectiveexistence of widespread atrocities was not considered sufficient to impute subjec-tive knowledge unless the superior had concrete information about these atroci-ties.

The judgment in the High Command case 38 applied an even more restrictive stan-dard, requiring a ‘personal dereliction’ and proof of actual knowledge:

Criminality does not attach to every individual in this chain of command from thatfact alone. There must be a personal dereliction. That can occur only where the actis directly traceable to him or where his failure to properly supervise his subordinatesconstitutes criminal negligence on his part. In the latter case, it must be a personalneglect amounting to a wanton, immoral disregard of the action of his subordinatesamounting to acquiescence. Any other interpretation of international law would gofar beyond the basic principles of criminal law as known to civilized nations.39

123456789101112131415161718192021222324252627282930313233343536373839404142


31 US v. Pohl et al. (case 4), V TWC, 958–1163, at 1011 (emphasis added).32 Ibid., at 1054.33 See US v. Brandt et al. (‘Medical case’, case 1), II TWC, supra note 11, at 171–300, 212, 213.34 The IMT, supra note 21, at 469, also stated in the context of the criminality of certain Nazi

organizations ‘that mass punishments should be avoided’.35 Cf. US v. Krauch et al. (‘IG Farben’, case 6), TWC VIII, 1081–1210, at 1124, 1126; US v. von

Leeb et al. (‘High Command case’, case 12), TWC XI, S. 462–697, at 489.36 US v. von List et al. (case 7), TWC XI, supra note 11, 1230 et seq., at 1256–1257. See also the

extensive analysis of Parks, supra note 3, 58 et seq.37 US v. von List et al., supra note 36, 1281 (‘obliged to know’).38 US v. von Leeb et al., supra note 35, in particular at 510, 533–544, 555. See also the thorough

analysis of Parks, supra note 3, 38 et seq.; also J. J. Douglass, ‘High Command Case: A Study in Staffand Command Responsibility’, 6 International Lawyer (1972) passim; G. R. Vetter, ‘CommandResponsibility of Non-military Superiors in the International Criminal Court (ICC)’, 25 YaleJournal of International Law (2000) 106.

39 Ibid., at 544.

811


Further, it was stated that the duty of control or supervision diminishes with dimin-ishing commanding power. Thus, superior responsibility may apply to a com-manding general and even to a commanding officer; it does not, however, applywhen such officers transmit orders without knowledge of their content.40 Nor doesit apply to staff officers since such officers normally have no commanding power.41

However, under certain circumstances even staff officers may incur criminal liabil-ity: ‘[i]f the basic idea is criminal under international law, the staff officer who putsthat idea into the form of a military order . . . or takes personal action to see that itis properly distributed . . . commits a criminal act under international law.’42

In sum, this survey shows that the Nuremberg jurisprudence did not refine theYamashita precedent, as argued by O’Brien,43 but rather that it developed, for thefirst time, a standard at all. In fact, the tribunal in the High Command case clearlyrejected Yamashita, applying instead a standard of positive knowledge, while theTribunal in the Hostage case opted for a should-have-known standard based on theconcrete information received by the superior.44 The latter standard proved to beof utmost importance in subsequent case law and codifications.

(b) IMTFE TokyoThe Tokyo Tribunal confirmed and extended the Nuremberg jurisprudence tocivilian personnel.45 First, it imposed duties of supervision and control on civiland military command personnel with regard to the proper treatment of war pris-oners. Consequently, abuses of prisoners were attributed to both civil and militarydefendants if they failed to prevent them. Superior responsibility was furtherextended to the (civilian) government, arguing for a type of collective responsibil-ity: ‘A cabinet member may resign. If he has knowledge of ill-treatment of prison-ers, is powerless to prevent future ill-treatment, but elects to remain in the cabinetthereby continuing to participate in its collective responsibility . . . [,] he willinglyassumes responsibility for any ill-treatment in the future.’46

Membership in the cabinet was deemed to justify the causal connection—at leastin the case of wilful or negligent ignorance—between the failure to act and thecommitted abuses. Also, duties of oversight and investigation were increased.


40 Ibid., at 510 (‘for the transmittal in any manner whatsoever of a criminal order’). See also USv. von List et al., supra note 36, at 1281 et seq., 1286.

41 US v. von Leeb et al., supra note 35, at 514. On the responsibility of staff officers, see Douglass,698 et seq. (702–703); also E. J. O’Brien, ‘The Nuremberg Principles, Command Responsibility,and the Defense of Captain Rockwood’, 149 Military Law Review (1995) at 90; I. Bantekas, ‘TheContemporary Law of Superior Responsibility’, 93 AJIL (1999) 581.

42 US v. von Leeb et al., supra note 35, at 513.43 See O’Brien, at 286–287.44 Cf. Crowe, supra note 5, 194, 209 et seq. (213, 215, 219–220), 232.45 The Tokyo Judgment: The International Military Tribunal for the Far East (IMTFE), Vol. I, ed.

B. V. A. Röling and C. F. Rüter (1977) at 29–30. Cf. Fenrick, supra note 1, at 118.46 The Tokyo Judgment, supra note 45, Vol. I, p. 30 (emphasis added).

812

(Zb) Cassese Ch21 15/2/02 9:00 am 12

As a matter of fact, the wide attribution of criminal responsibility to civilian supe-riors for military atrocities merits some criticism. Take, for example, the case ofMamoru Shigemitsu. Shigemitsu was Minister of Foreign Affairs between April1943 and April 1945. In this period the Allies protested strongly because of thetreatment of Allied prisoners of war. The evidence showed that Shigemitsu knewof the situation in the prison camps. According to the Tribunal, he had a specialresponsibility as a member of the government for the well-being of the prisoners.Therefore, he should have taken adequate steps to investigate the situation or evenresigned in order to press the matter. As he did none of these things, he showedgrave disrespect to the law of war and was considered guilty.47 Similarly, theTribunal stated with regard to the responsibility of Koki Hirota, Foreign Ministerfrom 1933 to March 1936: ‘His inaction amounted to criminal negligence.’48

Judge Röling, in this author’s view correctly, criticized these and other individualcases in his dissenting opinion.49 As to Hirota, Röling argued that he should nothave been held responsible, since he personally approached the responsible WarMinister. In general, according to Röling, ‘a Tribunal should be very careful inholding civil government officials responsible for the behaviour of the Army in thefield’. Further, responsibility for omissions should only be recognized ‘in a veryrestricted sense’.50

(c) UNWCCThe Tokyo jurisprudence was confirmed by the British, Canadian, Australian andChinese war crimes trials as documented by the UNWCC.51 Concretely speaking,the competent tribunals—a grosso modo—considered as crucial the position orstatus of the defendant when determining criminal accountability based on com-mand responsibility. The Canadian trial against Brigadeführer Kurt Meyer(Abbay Ardenne case) is of particular importance in this respect.52 The UNWCCitself took the position, however, that ‘the principles governing this type of liabil-ity . . . are not yet settled’53 and formulated the following questions:

(1) How far can a commander be held liable for not taking steps before thecommitting of offences, to prevent their possible perpetration?

123456789101112131415161718192021222324252627282930313233343536373839404142


47 Ibid., at 457–458.48 Ibid., at 446 et seq. (447–448). See also Vetter, supra note 38, at 125 et seq.49 The Tokyo Judgment, supra note 45, Vol. II, pp. 1043 et seq. (1126–1128, 1137–1138).50 Ibid., at 1127.51 See e.g. the explicit reference to the Yamashita case in XV Law Reports TWC (UNWCC,

1949), at 66. For other trials, see also Parks, supra note 3, 69 et seq.52 IV Law Reports TWC, supra note 51, at 98–99. See also L. C. Green, ‘Superior Orders and

Command Responsibility’, 27 CYIL (1989), 186 et seq.; id., supra note 3, 337 et seq.; Bassiouni,supra note 3, 431–432.

53 XV Law Reports TWC, supra note 51, at 67. Also G. Brand, ‘The War Crimes and the Lawsof War’, 26 BYIL (1949) 414–427 (424).

813


(2) How far must he be shown to have known of the committing of offences inorder to be made liable for not intervening to stop offences already beingperpetrated?

(3) How far has he a duty to discover whether offences are being committed?54

(d) Other National Trials for War CrimesAs far as other national trials or inquiries are concerned, the judicial treatment ofthe My Lai massacre is of particular interest. As was already mentioned,55 the USmilitary tribunals did not apply the superior responsibility doctrine in the trialsagainst Captain Medina and Lieutenant William Calley, Jr., as it was understoodsince Yamashita. Whereas Calley was convicted by a court martial of directing themassacre, Medina, his immediate superior, was acquitted because of lack of (pos-itive) knowledge of the atrocities.56 This is surprising since Medina—unlikeYamashita who had allegedly been completely cut off from communication withhis troops—was present at the scene of the crimes57 and in close contact withCalley.58 Also, the requirement of positive knowledge is not consistent with USlaw.59 The doctrine also played a role in the military investigation of MajorGeneral Koster, commanding general of the division to which Medina’s companywas attached. In this context, the Secretary-General of the Army found: ‘. . . [A]commander clearly must be held responsible for those matters which he knows tobe of serious import, and with respect to which he assumes personal charge. Anyother conclusion would render essentially meaningless and unenforceable theconcepts of great command responsibility accompanying senior positions ofauthority . . .’60

Finally, the superior responsibility doctrine was confirmed by Israeli practice inthe Kafr Qassem incident and the massacres at the refugee camps Sabra and Shatilain Lebanon. In the former case, the Israeli Military Court of Appeal held a lieu-


54 Ibid.55 See supra note 28 and accompanying text.56 See R. Hammer, The Court-Martial of Lt. Calley (1971), passim; see also J. Goldstein, B.

Marshall, and J. Schwartz, The My Lai Massacre and its Cover-up: Beyond the Reach of Law? (1976),reprinting the Review and Appeals decisions against Calley on pp. 475 et seq. See also Landrum,supra note 5, at 299; Crowe, supra note 5, 223; W. G. Eckhardt, ‘Command CriminalResponsibility: A Plea for a Workable Standard’, 97 Military Law Review (1982) 11 et seq. and32–34 (summary of evidence by Judge Howard); Bassiouni, supra note 3, at 434. A balanced assess-ment of the US role in Vietnam is provided by O’Brien, supra note 16, at 641 et seq., 663.

57 See Colonel Howard’s statement to the jury as recounted by Green, supra note 3, 353.Howard was the defence lawyer of Medina (see his article in the bibliography).

58 See Lieutenant Calley: His Own Story as Told to John Sack (1970). See also Prévost, supra note5, at 329 note 152. Not until early 1998 did the US Army publicly recognize that there was noenemy in My Lai but an ‘unlawful massacre of noncombatants by American forces’ (GuardianWeekly/Washington Post, 15 March 1998).

59 Crowe, supra note 5, 223–224; Bassiouni, supra note 3, at 434. see also infra, II.C.60 Quoted according to Green, supra note 3, 355. See also the prosecution brief concerning

Medina as reproduced in Eckhardt, supra note 56, at 30–32.

814


tenant responsible because, among other things, ‘he was aware of what was takingplace . . . and did not take any measures to stop the killings.’61 In the Sabra andShatila massacres a commission of inquiry under the president of the SupremeCourt found that the Israeli minister of defence and the military authorities wereaware of the massacres committed by the Phalangists in the refugee camps but didnot adequately intervene or punish the perpetrators.62

3. The Jurisprudence of the Existing Ad Hoc Tribunals

In Delalic et al.—also known as the Čelibici case after the prison camp of thatname—the ICTY characterized the requirements of superior responsibility as fol-lows:63

• existence of a superior-subordinate relationship;• superior’s knowledge or superior having reason to know that the act was about

to be or had been committed;• superior’s failure to take the necessary and reasonable measures to prevent the

criminal act or punish the perpetrator thereof.

The more recent judgments in Aleksovski 64 and Blaškic 65 basically confirmed orcomplemented the principles developed in Delalic et al. and will therefore be pre-sented together with this judgment. With regard to the first requirement, theICTY held that ‘individuals in positions of authority . . . may incur criminalresponsibility . . . on the basis of their de facto as well as de jure positions as superi-ors’.66 A position of command cannot be determined by ‘reference to a formal sta-tus alone’ but by ‘the actual possession, or non-possession, of powers of [effective]control over the actions of subordinates.’67 In this sense, superior responsibilityextends to civilian, non-military superiors, but ‘only to the extent that they exer-cise a degree of control over their subordinates similar to that of military com-manders’.68 This implies that various persons may—as superiors—be responsible

123456789101112131415161718192021222324252627282930313233343536373839404142


61 Quoted according to Green, supra note 3, at 359.62 See the extensive quotes from the Commission’s report in Green, supra note 3, 361 et seq.; see

also id., supra note 52, 200–201; Bassiouni, supra note 3, at 439–440; Fenrick, supra note 1, 120 etseq.

63 T.Ch., Judgment of 16 November 1998, Delalic/Mucic/Delic/Landžo(‘Čelibici’), IT-96-21-T,para. 346.

64 T.Ch., Judgment of 25 June 1999, Aleksovski, IT-95-14/1 (original in French), para. 69 et seq.(69); confirmed by App.Ch., Judgment of 24 March 2000, IT-95-14/1-A, para. 69–77.

65 T.Ch., Judgment of 3 March 2000, Blaškic, IT-95-14-T (original in French), para. 289 et seq.(294).

66 Delalic et al., supra note 63, para. 354 (emphasis in the original); confirmed in App. Ch.,Judgment of 20 February 2001, Delalic/Mucic/Delic/Landžo (‘Čelibici’), IT-96-21-A, also Blaškic,supra note 65, para. 300.

67 Delalic et al., supra note 63, para. 370; see also para. 377 (emphasis added); confirmed inDelalic et al. (Appeal), supra note 66, para. 198; also Aleksovski, supra note 64, para. 76; Blaškic,supra note 65, paras. 301, 335; T. Ch., Judgment of 2 August 2001, Krstic IT-98-33-T, para. 648.

68 Delalic et al., supra note 63, paras. 355 et seq., 363, 378; Aleksovski, supra note 64, paras. 75, 78.

815


for a crime committed by a subordinate.69 However, responsibility is excluded ifsuch control was ‘absent or too remote’,70 or if the superior lacked ‘the materialability to prevent and punish the commission of the[se] offences’.71 Blaškicemphasized that the lack to prevent the commission of the crimes ex ante cannotbe compensated by their punishment ex post, i.e. the superior is obliged to doboth.72 On the other hand, the non-punishment of crimes committed can be con-sidered as aiding or encouraging their commission within the meaning of Article7(1) ICTY Statute.73 Finally, as was stated in Aleksovski, a civilian superior doesnot need to have the same power of sanction as a military superior; were this thecase, the doctrine would be virtually inapplicable to civilian superiors.74

On the subjective level the superior must:

• have actual knowledge that his subordinates were committing or about tocommit crimes; or

• possess information of a nature which would put him on notice of the risk ofsuch offences by indicating the need for additional investigation in order toascertain whether they were committed or were about to be committed.75

While these requirements were not disputed in principle, there was controversyconcerning how actual knowledge can be proved. The ICTY rejected a presump-tion of knowledge but permitted the establishment of such knowledge by meansof circumstantial evidence that takes into account certain indicia.76 The positionof the superior as such constitutes a strong presumption that he had knowledge


69 Aleksovski, supra note 64, para. 106; Blaškic, supra note 65, para. 303.70 Delalic et al., supra note 63, para. 377.71 Ibid., para. 378 (emphasis added); Aleksovski, supra note 64, para. 81; Blaškic, supra note 65,

paras. 302, 335: ‘capacité materiélle’.72 Blaškic, supra note 65, para. 336.73 Ibid., paras. 337 et seq. (339); similarly T.Ch., Judgment of 26 February 2001, Kordic/Čerkez,

IT-95-14-2-T, para. 371.74 On this point, it is worth quoting the Chamber’s statement in full: ‘Exiger dune autorité civile

quelle détienne un pouvoir de sanction semblable à celui détenu par un militaire limiterait le champd’application de la doctrine du supérieur hiérarchique au point de la rendre pratiquement inapplic-able aux autorités civiles. La Chambre considère, par conséquent, qu’il n’est pas indispensable quele supérieur ait eu la capacité, de jure ou de facto, de prendre lui-même des sanctions. La simple pos-sibilité de transmettre des rapports aux autorités suffit’ (Aleksovski, supra note 64, para. 78).

75 Delalic et al., supra note 63, paras. 379 et seq., 383, 386, 393; confirmed in Delalic et al.(Appeal), supra note 66, para. 241.

76 Ibid., para. 386; also Blaškic, supra note 65, para. 307: number, type, and scope of illegal acts;time during which the illegal acts occurred; number and type of troops involved; logistics involved,if any; geographical location of the acts; widespread occurrence of the acts; tactical tempo of opera-tions; modus operandi of similar illegal acts; officers and staff involved; location of the commanderat the time. These indicia have been developed by the UN Commission of Experts for the FormerYugoslavia (UN Doc. S/1994/674 of 27 May 1994), quoted in Bassiouni, supra note 3, 437–438.Cf. Aleksovski, supra note 64, paras. 80 and 114: ‘Il ne pouvait donc ignorer les mauvais traitementsrépétés subis par les détenus’. See also W. J. Fenrick, ‘Article 28: Responsibility of Commanders andOther Superiors’, in O. Triffterer (ed.), Commentary on the Rome Statute of the InternationalCriminal Court (1999) mn. 10.

816


about the crimes of the subordinates.77 In contrast, the specific contents of the‘had reason to know’ requirement remained unresolved. While it is clear that asuperior must not remain wilfully blind to the acts of his or her subordinates, it isless clear how far the superior’s duty to know reaches and how specific theinformation in the superior’s possession must be in order to oblige him or her toinvestigate further. Thus, in Delalic et al. the ICTY was careful in its formulationof an abstract standard and made no finding as to the present content of custom-ary law on this point.78 However, in Blaskic the Chamber tried to develop a cus-tomary law rule on the basis of the case law and an interpretation of Articles 86and 87 AP I.79 With regard to the former it follows the conclusion drawn by Parks,according to which the superior is responsible if ‘he failed to exercise the meansavailable to him to learn of the offense and, under the circumstances, he shouldhave known and such failure to know constitutes criminal dereliction.’80 Withregard to the latter, the Chamber interprets the expression ‘possédaient des infor-mations’ of Article 86(2) AP I broadly. This very strict position has not beenaccepted in the Delalic et al. appeal decision, which held that there is ‘no “duty” toknow’ for military or civil superiors.81 However, the appeals decision considered avery low standard with regard to the kind of—available—information which issufficient to trigger command responsibility. According to the Chamber it is, forexample, sufficient that the superior had informations at her disposal ‘that someof the soldiers under his command have a violent or unstable character, or havebeen drinking prior to being sent on a mission’. 82 Thus, a superior must analysethe information at her disposal very thoroughly and take the necessary measuresto prevent crimes from being committed.83 In conclusion, if the superior hasproperly fullfilled her duties but still does not know about the crimes committedby the subordinates, such ignorance cannot be held against her.84

Apart from Delalic et al., Aleksovski, and Blaskic, the ICTY has confirmed, primafacie, the individual responsibility of civil leaders and military superiors in variousRule 61 decisions that have been described elsewhere.85 Further, in the indictment

123456789101112131415161718192021222324252627282930313233343536373839404142


77 Aleksovski, supra note 64, para. 80; Blaškic, supra note 65, para. 308.78 Delalic et al., supra note 63, para. 393.79 Blaškic, supra note 65, para. 309 et seq. (314 et seq.).80 Parks, supra note 3, at 90, referred to in Blaškic, supra note 65, para. 322 (translated into

French).81 Delalic et al. (Appeal), supra note 66, para. 240; confirmed in Kordic/Čerkez, supra note 73,

paras. 435, 437.82 Delalic et al. (Appeal), supra note 66, para. 328.83 Ibid., para. 325 et seq. (328–329), relying on the ICRC Commentary on the Protocols.84 Ibid, para. 332.85 Cf. K. Ambos, ‘Individual Criminal Responsibility in International Criminal Law’, in G. K.

McDonald and O. Swaak Goldman (eds.), Substantive and Procedural Aspects of InternationalCriminal Law, Vol. I (2000) at 26–28.

817


against Miloševic et al.,86 confirmed by the review decision of Judge Hunt,87 theaccused Miloševic, Milutinovic, Ojdanic, and Stojilkovic were also considered,prima facie, responsible for crimes against humanity on the basis of superiorresponsibility.88

The ICTR dealt for the first time with superior responsibility in Akayesu.89

Although characterizing the application of the doctrine to non-military (civilian)superiors—before the ICTY’s Delalic et al. decision—as ‘contentious’,90 theChamber did not reject this possibility but held that ‘it is appropriate to assess ona case by case basis the power of authority actually devolved upon the Accused inorder to determine whether or not he had the power to take all necessary and rea-sonable measures . . .’.91 Thus, for example, one of the charges against Akayesubased on superior responsibility was rejected by the Chamber since the paramili-tary Interahamwe could not be considered his subordinates92 and he thereforecould not be expected to have controlled them effectively. The existence of thematerial possibility to control or intervene is, however, irrelevant if the defendantdid not even try to prevent the crimes and instead collaborated with the direct perpetrators.93

As to the mental element, the chamber in Akayesu rejected the view—based on arule of strict liability—that the responsibility of the superior is independent of hisor her criminal intent.94 Rather, the superior must have acted with ‘maliciousintent, or, at least, . . . negligence . . . so serious as to be tantamount to acquies-cence or even malicious intent’.95

In Kayishema and Ruzindana the ICTR essentially confirmed the Čelibici judg-ment and refered to Article 28 of the Rome Statute.96 It stated that individualcriminal responsibility according to Article 6(1) of the ICTR Statute and superior responsibility according to Article 6(3) of the ICTR Statute are notmutually exclusive.97 The doctrine also applies to the responsibility of non-


86 Prosecutor against Milosevic/Milutinovic/Sainovic/Ojdanic/Stojiljkovic, Indictment, submittedon 22 May 1999. See Vetter, at 138 et seq.

87 Prosecutor v. Miloševic/Milutinovic/Šainovic/Ojdanic/Stojiljkovic. Decision on Review ofIndictment and Application for Consequential Orders, 24 May 1999.

88 Indictment, supra note 86, para. 55 et seq., 84–88.89 T.Ch., Judgment of 2 September 1998, Akayesu, ICTR-96-4-T, para. 486 et seq.90 Ibid., para. 491. See the critical view of Vetter, supra note 38, 132 et seq.91 Prosecutor v. Akayesu, supra note 89, para. 491.92 Ibid. para. 691.93 Ibid., para. 193.94 Ibid., para. 488.95 Ibid., para. 489. See also T.Ch. I, Judgment and Sentence, 27 January 2000, Alfred Musema,

ICTR-96-13-T, para. 131.96 T.Ch. II, Judgment, 21 May 1999, Clément Kayishema and Obed Ruzindana, ICTR-95-1-T,

para. 208–231.97 Ibid., para. 210.

818


military superiors.98 In any case, responsibility depends on the ‘degree of author-ity’ over the subordinates which in turn is based on the superior’s de jure or/and defacto control, the latter being decisive.99 The material ability to control the actionsof the subordinates is ‘the touchstone’ of the superior responsibility doctrine; it is‘inherently linked with the factual situation’ in the concrete case.100 As to the mensrea required, the Chamber considers the Rome Statute’s distinction between mil-itary and civilian superiors ‘an instructive one’.101 Accordingly, the accused, i.e. acivilian superior, would be liable if he ‘either knew or consciously disregardedinformation which clearly indicated or put him on notice that his subordinateshad committed, or were about to commit’ criminal acts.102 This distinction wasnot drawn in Delalic et al.

B. Efforts of Codification

1. Protocol Additional I to the Geneva Conventions

The most important provisions codifying the concept of superior responsibilitythat serve as a basis for all further codifications are Articles 86 and 87 ProtocolAdditional I (PA I) to the Geneva Conventions of 8 June 1977.103 Article 86(1) PAI establishes the general obligation of the State Parties to suppress grave breaches‘which result from a failure to act when under a duty to do so’. Thus, a superiorcan only be held liable if he or she failed to take action against grave breaches com-mitted by his subordinates and if the superior had a duty to act. Accordingly,Article 86(2) PA I describes the failure-to-act-situation as follows:

The fact that a breach of the Conventions or of this Protocol was committed by asubordinate does not absolve his superiors from penal or disciplinary responsibility,as the case may be, if they knew, or had information which should have enabled themto conclude in the circumstances at the time, that he was committing or was going tocommit such a breach and if they did not take all feasible measures within their powerto prevent or repress the breach.104

123456789101112131415161718192021222324252627282930313233343536373839404142


98 Ibid., para. 213 et seq.; confirmed in T.Ch., Judgment and Sentence of 27 January 2000,Musema, ICTR-96-13-T, para. 132 et seq.

99 Ibid., para. 216, 217 et seq. (218, 230).100 Ibid., para. 229 et seq. (229, 231).101 Ibid., para. 227.102 Ibid., para. 228.103 Official Records of the Diplomatic Conference on the Reaffirmation and Development of

International Humanitarian Law Applicable in Armed Conflicts. Geneva (1974–1977), Vol. I (1978)115. For the earlier humanitarian law which can be considered the basis of the doctrine, see in par-ticular Art. 1 of the 1907 Hague Regulations and Art. 43 of the Protocol (Armed forces): J. DePreux, ‘Commentary on Articles 86 and 87 of Protocol Additional I’, in Y. Sandoz, Ch. Swinarski,and B. Zimmermann (eds.), Commentary on the Additional Protocols of 8 June 1988 to the GenevaConventions of 12 August 1949 (1987) mn. 3531, 3540; Crowe, supra note 5, 196–7; Green, supranote 3, 325; Fenrick, supra note 76, mn. 1.

104 Emphasis added.

819


Thus, three conditions must be fulfilled to hold the superior responsible:105

(1) the breach was committed by one of the superior’s subordinates;(2) the superior knew, or had information which should have enabled him or

her to conclude106 that a breach was being committed or was going to becommitted;

(3) the superior did not take the measures within his or her power to preventor repress the breach.

These requirements will be analysed in detail below. At this stage it is sufficient topoint out that Article 86(2) PA I establishes not only the superior’s responsibility(including civilian superiors)107 but also the parallel responsibility of the subordi-nate. This must be stressed since, taken together, Articles 85 et seq. PA I seem toimply an exclusive responsibility of the superior. However, by the phrase ‘wascommitted by a subordinate’, Article 86(2) explicitly refers to a breach of theConventions or the PA I by a subordinate.108 Thus, the subordinate is directlyresponsible as an immediate perpetrator.

Article 86(2) PA I must be read in conjunction with Article 87, which describesthe duty of a commander to prevent the commission of crimes by the subordinatesand thereby establishes the superior’s position and duty as a ‘guarantor’(Garantenstellung und -pflicht). It reads as follows:

1. The High Contracting Parties and the Parties to the conflict shall require militarycommanders, with respect to members of the armed forces under their command andother persons under their control, to prevent and, where necessary, to suppress and toreport to competent authorities breaches of the Conventions and of this Protocol.2. In order to prevent and suppress breaches, High Contracting Parties and Partiesto the conflict shall require that, commensurate with their level of responsibility,commanders ensure that members of the armed forces under their command areaware of their obligations under the Conventions and this Protocol.3. The High Contracting Parties and Parties to the conflict shall require any com-mander who is aware that subordinates or other persons under his control are goingto commit or have committed a breach of the Conventions or of this Protocol, to


105 See also De Preux, supra note 103, mn. 3543 et seq.106 ICRC Draft Art. 76 para. 2 proposed a ‘should have known’ standard but was rejected as too

broad (see De Preux, supra note 103, mn. 3526; also Levie, supra note 3, 8; Landrum, supra note 5,294; Crowe, supra note 5, 225).

107 Cf. Hessler, supra note 16, 1277; T. Wu and Y. Kang, ‘Criminal Liability for the Actions ofSubordinates: The Doctrine of Command Responsibility and its Analogues in United States Law’,38 Harvard International Law Journal (1997) 291–292.

108 Cf. R. Wolfrum, ‘Durchsetzung des humanitären Völkerrechts’, in D. Fleck (ed.), Handbuchdes humanitären Völkerrechts in bewaffneten Konflikten (1994) 413–440 (423); see also W. A. Solfand E. R. Cummings, ‘A Survey of Penal Sanctions under Protocol I to the Geneva Conventions ofAugust 12, 1949’, 9 Case Western Reserve Journal of International Law (1977) 242–243. This viewalso resolves the criticism of K. J. Partsch, ‘Commentary on Articles 86 and 87 of ProtocolAdditional I’, in M. Bothe, K. J. Partsch, adn W. A. Solf (eds.), New Rules for Victims of ArmedConflicts (1982) 524, who claims that Art. 86 para. 2 is incomplete since it does not consider theresponsibility of the subordinate.

820


initiate such steps as are necessary to prevent such violations of the Conventions orthis Protocol, and, where appropriate, to initiate disciplinary or penal action againstviolators thereof.109

2. Other

Subsequent codifications were based exclusively on Article 86(2) PA I and variedonly with regard to the legal consequence and the mental element. Article 12 ofthe ILC Draft Code 1991 and Article 6 of the ILC Draft Code 1996—‘responsi-bility of the superior’—read as follows:

The fact that a crime against the peace and security of mankind was committed bya subordinate does not relieve his superiors of criminal responsibility, if they knew orhad information enabling them to conclude, in the circumstances at the time, that thesubordinate was committing or was going to commit such a crime and if they did nottake all feasible measures within their power to prevent or repress the crime.110

The fact that a crime against the peace and security of mankind was committed bya subordinate does not relieve his superiors of criminal responsibility, if they knew orhad reason to know, in the circumstances at the time, that the subordinate was com-mitting or was going to commit such a crime and if they did not take all necessarymeasures within their power to prevent or repress the crime.111

Thus, the responsibility of the superior has, again, three requirements: the com-mission of crimes by the subordinates, the superior’s failure to take all possiblecountermeasures against these crimes, and his or her positive or possible know-ledge of them.112 The differences between the 1991 and 1996 Draft clearly showwhere the problems lie: on the one hand, the questions arise whether the superiorhas to take ‘feasible’ or ‘necessary’ measures and whether there is a differencebetween the two; on the other hand, one wonders whether the different formula-tions with regard to possible knowledge are supposed to create different standards.According to the ILC, the changes did not imply substantive differences.113 Themajority of its members accepted—despite criticism114—the proposal of theSpecial Rapporteur to retain Article 12 of the Draft Code 1991.115

123456789101112131415161718192021222324252627282930313233343536373839404142


109 Emphasis added.110 Art. 12 of the ILC Draft Code 1991, in YILC (1991) Vol. II, Part 2, pp. 94–97.111 Art. 6 of the ILC Draft Code 1996, in Report of the ILC on the work of its forty-eighth ses-

sion, 6 May to 26 July 1996, UN GAOR, 51st Sess., Supp. No. 10 (A/51/10) 14 et seq. (paras. 50 etseq.) = YILC (1996), Vol. II, Part 2, at 15 et seq. (emphasis added).

112 See also YILC (1988) Vol. I, pp. 288–289 (Mr Tomuschat, chairman of the DraftingCommittee); Vol. II, Part 2, at 70–71. See also the Fourth and Fifth Report of the SpecialRapporteur, Doudou Thiam, YILC (1986), Vol. II, Part 1, pp. 53–86 (65–6) and YILC (1987),Vol. II, Part 1, pp. 1–13 (9), respectively, who mistakenly considered this provision a ‘specific caseof the theory of complicity’.

113 Cf. Report ILC 1996, supra note 111, at 38, para. 5 and 6 = YILC (1996), Vol. II, Part 2, at26, para 5 and 6.

114 Prof. Tomuschat criticized the extensive liability created by Art. 12 (YILC (1994), Vol. I, at 112).115 Thiam, Twelfth Report (1994), para. 127; YILC (1994), Vol. II, Part 2, S. 83; (1994), Vol.

I, p. 111 (Mr Al-Baharna).

821


The unofficial (updated) Siracusa Draft116 deviates only slightly from the ILCDraft Code 1996: ‘The fact that a crime under this Statute was committed by asubordinate does not relieve his superiors of criminal responsibility, if they knewor had reason to know, under the circumstances of the time, that the subordinate wascommitting or was going to commit such a crime and if they did not take all nec-essary measures within their power to prevent or repress the crime.’117

A comparison of Article 86(2) PA I with the provisions of the ILC Draft Codesand the Siracusa Draft shows first, that the draft provisions refer only to ‘criminal’consequences for the superior; no mention is made of ‘disciplinary’ responsibility.This does not, however, preclude disciplinary sanctions based on national mili-tary law; rather, the assignment of a commensurate criminal sanction to thealleged criminal conduct indicates that disciplinary sanctions alone do not con-stitute an adequate sanction. Secondly, the superior is obliged—according to theILC Draft Code 1996 and the Siracusa Draft—to take all ‘necessary’ rather than‘feasible’ measures. The difference, if there is any, will be analysed below. Finally,while the ILC Draft Code 1991 introduces only a minor change with regard to themental element, the ILC Draft Code 1996 and the Siracusa Draft introduce amajor change by replacing the original formula ‘information which should haveenabled them to conclude’ by ‘had reason to know’. The correct formulation washotly debated in Rome. This is understandable since this aspect of the mental ele-ment is arguably at the heart of the superior responsibility doctrine. Although thispoint will be addressed later in more detail, prima facie it seems to be clear that the‘had reason to know’ requirement considerably lowers the mens rea requirementbecause it is a much wider and less precise concept.

Be that as it may, this language was already adopted before by the Statutes of theUN Ad Hoc Tribunals. Article 7(3) of the ICTY Statute118 as well as Article 6(3)of the ICTR Statute119 read: ‘The fact that any of the acts . . . of the present Statutewas committed by a subordinate does not relieve his superior of criminal respon-sibility if he knew or had reason to know that the subordinate was about to commitsuch acts or had done so and the superior failed to take the necessary and reasonablemeasures to prevent such acts or to punish the perpetrators thereof.’120


116 Cf. Association Internationale de Droit Pénal (‘AIDP’)/International Institute of HigherStudies in Criminal Sciences (‘ISISC’)/Max Planck Institute for Foreign and InternationalCriminal Law (‘MPI’) et al., 1994 ILC Draft Statute for an International Criminal Court with sug-gested modifications, prepared by a committee of experts, Siracusa/Freiburg/Chicago, 15 Mar. 1996(‘Updated Siracusa Draft’). For the ‘General Part Draft’ and the earlier Siracusa Draft, see<http://www.iuscrim.mpg.de>.

117 Art. 33-2 para. 3 (emphasis added).118 SC Res. 827 (25 May 1993), reprinted in: 32 ILM (1993) 1159. See also

<http/www.un.org/icty/basic.htm>.119 SC Res. 955 (8 November 1994), see <http://www.ictr.org>.120 Emphasis added.

822


In addition, the Statutes do not incorporate AP I, Article 82(2)’s temporal refer-ence to ‘in the circumstances at the time’ and introduce the following modifica-tion with regard to the superior’s duty to intervene: Whereas, according to Article86(2) PA I, intervention had to take place when the subordinate ‘was committingor was going to commit’ a crime, according to Article 7(3) ICTY Statute andArticle 6(3) ICTR Statute, intervention must take place when the subordinate‘was about to commit’ a crime or ‘had done so’. It must be recalled, however, thatArticle 87(3) PA I contained similar language referring to crimes that the subor-dinates ‘have committed’. In any case, this change affects the actus reus of theoffence since the way the superior is supposed to react to a crime depends on thetime of its commission; once the crime is committed the issue becomes one of fail-ure to punish. This is a different problem that requires a more detailed analysisbelow. Furthermore, the Statutes require the superior to take ‘necessary’ and ‘rea-sonable’ measures, thereby adding still another qualification to the manner inwhich he or she must act. On the other hand, the Statutes omit the reference tomeasures ‘within their power’ and thus seem to extend the superior’s liability.

C. Comparative Aspects

Norms delineating the principle of superior responsibility can generally be foundin national military laws. They are, as Röling has stated, ‘the logical consequenceof the duty of every State to ensure respect for the laws of war.’121 Given that thePA I has been ratified by more than 150 States,122 it is to be expected that theseStates, at least, in compliance with their treaty obligations, will incorporate therequisite norms on superior responsibility in their national law. Thus, the follow-ing overview can be limited to a few representative legal systems. Further, it willconsider parallel problems of employer’s responsibility that arise in the context ofeconomic criminal law.

1. Military Law

The most cited provision is the Field Manual 27-10 of the US Army which pro-vides:

Such a responsibility [of a commander] arises directly when the acts in question havebeen committed in pursuance of an order . . . The commander is also responsible ifhe has actual knowledge, or should have knowledge, through reports received by him orthrough other means, that troops or other persons subject to his control are about tocommit or have committed a war crime and he fails to take the necessary and rea-sonable steps to insure compliance with the law of war or punish violators thereof.123

123456789101112131415161718192021222324252627282930313233343536373839404142


121 B. V. A. Röling, ‘Criminal Responsibility for Violations of the Laws of War’, 12 Revue Belgede Droit International (1976) 8–26, at 16.

122 In September 1999, the number of ratifications was 155 <www.icrc.org>.123 US Department of the Army, Field Manual 27–10: The Law of Land Warfare, para. 501

(1956) [hereinafter FM 27–10] (emphasis added). Cf. Röling, supra note 119, 16; Crowe, supra

823


It must not be overlooked, however, that the Field Manual has no legally bindingeffect, it only provides ‘authoritative guidance’.124 Further, the Department of theArmy itself considered the ‘should have known’ standard as being too broad.125 Asto the rank of the commander, the Army training materials emphasize duties foreveryone in the chain of command notwithstanding the different levels andresponsibilities within the chain.126

The British Manual of Military Law employs the same wording, with a slight vari-ation, adding the words: ‘and if he fails to use the means at his disposal to ensurecompliance with the law of war’.127

In France, the Ordonnance of 28 August 1944 concerning the suppression of warcrimes does not provide for a concrete rule but does, at least, recognize superiorresponsibility. Article 4 reads: ‘Lorsqu’un subordonné est poursuivi commeauteur principal d’un crime de guerre et que ses supérieurs hiérarchiques ne peuvent être recherchés comme coauteurs, ils sont considérés comme complicesdans la mesure où ils ont organisé ou toléré les agissements criminels de leurs subordonnés’.128

In Germany, § 357 (1) alt. 3 StGB punishes the superior who intentionally ‘per-mits . . . an unlawful act by his subordinates to happen’. Article 41 of the militarycriminal code (Wehrstrafgesetz-WStG) punishes the superior who fails to superviseor to properly delegate supervision of his or her subordinates, and grave conse-quences are thereby caused, e.g. the commission of a crime by the subordinates.The norm is a genuine offence of omission (echtes Unterlassungsdelikt)129 that con-sists of two acts: the failure to properly supervise and the causation of the graveconsequence.130 The latter implies that the failure of supervision must have causedthe occurrence of the grave consequence; i.e. the grave consequence would not


note 5, 223–224; O’Brien, supra note 41, 288. For criticism of the US law, see Hessler, supra note16, 1287 et seq. On the US law shortly after World War II, see III Law Reports TWC 103 et seq.

124 Field Manual, supra note 123, 3.125 US Dept. of Army, Pamphlet No. 27-1-1, Protocols to the Geneva Conventions of 12 August

1949, p. 65. See also Eckhardt, supra note 56, 18.126 See O’Brien, supra note 16, at 640.127 Great Britain War Office, The Law of War on Land: Being Part 3 of the Manual of Military

Law, para. 631 (1958); see also para. 507 b. Cf. Röling, supra note 119, 16; Green, supra note 3,343.

128 H. Pelletier and J. Perfetti (eds.), Code Pénal 1998–1999 (11th edn., 1998), at 1118–1119.In English in III Law Reports TWC, at 94: ‘Where the subordinate is prosecuted as the actual per-petrator of a war crime, and his superiors cannot be indicted as being equally responsible, they shallbe considered as accomplices in so far as they have organised or tolerated the criminal acts of the sub-ordinates.’

129 Cf. J. Schölz and E. Lingens, Wehrstrafgesetz (3rd edn., 1988), § 41 mn. 2, 17. The corre-sponding active conduct is covered by §§ 33 and 34 WStG which punishes the (attempted) induce-ment of the subordinate by the superior to commit a crime.

130 Ibid., mn. 2.

824


have occurred if the superior had complied with his duty of supervision.131 Theduty of supervision corresponds to the competence to give orders.132 If the supe-rior delegates this duty to another superior (normally his or her subordinate), hehas to guarantee that this superior is adequately instructed and able to perform theduty of supervision.133 Thus, arguably, the duty of supervision is converted into aduty of correct selection, instruction, and follow-up control. The duty of super-vision itself is based on § 10 (2) Soldatengesetz (SG).134 It is generally recognizedthat this duty creates the obligation for the superior to prevent the commission ofcrimes by the subordinates.135 According to § 1 (4) SG a superior is any personwho may give orders to a soldier. As to the mens rea standard, § 41 WStG requiresat least gross negligence (Leichtfertigkeit) with regard to the violation of the dutyof supervision and negligence (Fahrlässigkeit) with regard to the causation of thegrave consequence.136

Finally, §§ 28 et seq. of the Wehrdisziplinarordnung contain rules on the exercise ofthe disciplinary competence of the superior. According to § 28(1), the superiormust initiate an investigation and clarify the factual situation if there exists a sus-picion of a disciplinary offence. If necessary the superior must impose disciplinarysanctions or submit the matter to the competent authorities of criminal prosecu-tion (§ 29).137

2. Related Aspects of Economic Criminal Law

From a comparative viewpoint it is worthwhile mentioning the efforts to draft aEuropean Model Code of Criminal Law. These efforts are still limited to eco-nomic criminal law, more concretely speaking, to the protection of the financialinterests of the European Union. The corresponding proposal, the Corpus Juris, inits most recent version states in Article 12:

(1) If one of the offences under Art. 1 to 8 is committed for the benefit of a busi-ness by someone acting under the authority of another person who is the head of thebusiness, or who controls it or exercises the power to make decisions within it, thatother person is also criminally liable if he knowingly allowed the offence to be com-mitted.

(2) The same applies to any public officer who knowingly allows an offences underArt. 1 to 8 to be committed by a person under him.

(3) If one of the offences under Articles 1 to 8 is committed by someone actingunder the authority of another person who is the head of a business, or who controls

123456789101112131415161718192021222324252627282930313233343536373839404142


131 Ibid., mn. 13.132 Ibid., mn. 7.133 Ibid., mn. 9, 11.134 Cf. W. Scherer and R. Alff, Soldatengesetz. Kommentar (6th edn., 1988), § 10 mn. 7;

W. Stauf, Soldatengesetz: Taschenkommentar (2nd edn., 1998), § 10 mn. 3.135 Explicitly Stauf, supra note 134, mn. 3; also Scherer and Alff, supra note 134, mn. 8.136 Schölz and Lingens, supra note 129, § 41 mn. 2, 16.137 See also Wolfrum, in Fleck (ed.), supra note 108, para. 1213, at 434.

825


it or exercises the power to make decisions within it, that other person is also crimi-nally liable if he failed to exercise the necessary supervision, and his failure facilitated thecommission of the offence.138

Also, an alternative proposal by Tiedemann139 is of particular interest, since it isbased on a comparative approach to the European criminal doctrine on omissionand takes into account aspects of the Anglo-American theory of vicarious liabil-ity,140 the French responsabilité du fait d’autrui and the German Garantendog-matik.141 In fact, this proposal is based on § 130 of the German law of pettyoffences (Ordnungswidrigkeitengesetz-OWiG) which sanctions the companyowner if he or she violates the duty of supervision within an organization possess-ing legal personality. This provision is structurally similar to the provisions onsuperior responsibility in that it makes the superior liable for an omission; there-fore, the doctrine under § 130 OWiG and employer’s responsibility in generalmay be analogously applied to the superior responsibility concept. § 130 OWiG,like § 41 WStG, is considered a genuine offence of omission (echtesUnterlassungsdelikt).142 The duty of company owners (and other responsible per-sons within the company)143 derives from their position which is similar to a ‘guar-antor’; this position, in turn, follows from the various duties of care that they arelegally obliged to fulfil.144 The object of the offence is the company owner’s failureto effectuate the necessary measures of supervision. This failure allows violationsof the company’s general duties towards the public to occur. These violations are


138 M. Delmas-Marty and J. Vervaele (eds.), The Implementation of the Corpus Juris in theMember States, Vol. I (2000), 193 (emphasis added). For the earlier version, see Delmas-Marty(ed.), Corpus Juris (1997) 68.

139 Prof. Tiedemann was one of the members of the expert group that drafted the Corpus Juris.140 See on the especially important ‘Responsible Corporate Officer’ doctrine US v. Dotterweich,

320 US 277 (1943); US v. Park, 421 US 658, 674 (1975); see also B. S. Hustis and J. Y. Gotanda,‘The Responsible Corporate Officer: Designated Felon or Legal Fiction?’ 25 Loy. U. Chi. L. J.(1994) 169, 176.

141 This proposal provides:

(1) Whoever must legally guarantee that another person does not commit a criminal act will bepunished as a perpetrator if he knew of the commission and could have prevented or impededit.

(2) The following persons, in particular, are responsible according to paragraph (1):(a) Members of a government, public officials and soldiers for acts that have been commit-

ted by subordinates (Weisungsabhängige, Untergebene oder Befehlsempfänger);

See Tiedemann, at 505–506, 511 (§ e para. 1 and 2 a of the proposal; translation by the author).See also O. Tsolka, Der allgemeine Teil des europäischen supranationalen Strafrechts i.w.S. (1995) 268et seq. (279, 280–281).

142 E. Göhler, Kommentar zum Ordnungswidrigkeitengesetz (12th edn., 1998) § 130 mn. 9; K. Rogall, ‘Dogmatische und Kriminalpolitische Probleme der Aufsichtspflichtverletzung inBetrieben und Unternehmen (§ 130 OwiG)’ 98 Zeitschrift für de gesamte Strafrechtswissenschaft(ZStW) (1986) 588.

143 See Göhler, supra note 142, mn. 4 et seq.144 Ibid., mn. 2. Rogall, supra note 142, 613 et seq., correctly argues that the company owner has

the position of a full garant.

826


committed by the subordinates who are not properly supervised. Thus, theowner’s omission causes the realization of the concrete dangers inherent to theworking and functioning of the company (which implies that the company oper-ates in a type of business that is inherently dangerous). The realization of thesedangers is not an element of the offence but only an objective condition of theowner’s liability (objektive Strafbarkeitsbedingung).145 In this sense, we can speak ofan offence that creates concrete danger (konkretes Gefährdungsdelikt).146

If the owner delegates his duty to other company staff, he is responsible for properselection and supervision of these persons. This duty diminishes with the increas-ing level of qualification of the delegates.147 The causal relationship between thefailure of supervision and the violation exists if adequate supervision would mostprobably have prevented or at least diminished the likelihood of the violationoccurring. Thus, the generally recognized conditio sine qua non formula—in theAnglo-American doctrine also known as the ‘but for’ test148—is applied in anexpanded version since any probable increase of risk (Risikoerhöhung) caused bythe lack of proper supervision is considered sufficient.149

As far as mens rea is concerned the owner must act intentionally or negligentlywith regard to the omission element of the offence, i.e. with regard to his or herfailure of proper supervision and the possible consequences thereof. However, theowner need not foresee that a certain violation will be committed as a conse-quence of the failure of supervision.150 Still, this low standard is highly suspect inview of the principles of guilt (Schuldgrundsatz) and equality (Gleichheitsgrund-satz).151

123456789101112131415161718192021222324252627282930313233343536373839404142


145 H. E. Rotberg, Gesetz über Ordnungswidrigkeiten. Kommentar (5th edn., 1975) § 130 mn. 8.Critical B. Schünemann, ‘Strafrechtsdogmatische und Kriminalpolitische Grundfragen derUnternehmenskriminalität’, 1 Zeitschrift für Wirtschaft, Steuer, Strafrecht (Wistra) (1982) 48.

146 Göhler, supra note 140, mn. 9; Rogall, supra note 142, 599–600. The latter, however, limitsthe owner’s liability further by considering the violations by the subordinates not only as a condi-tion of the owner’s liability, but as the exclusive point of reference of his omission (ibid., at 591).Further, he requires a specific relationship between the omission and the violation (ibid., at 595).

147 Göhler, supra note 140, mn. 11; Rogall, supra note 142, at 603.148 Cf. G. Fletcher, Basic Concepts of Criminal Law (1998) at 62–63.149 Göhler, supra note 142, mn. 17, 22; Rotberg, supra note 145, mn. 12. See also Schünemann,

supra note 145, at 45; critically Rogall, supra note 142, 608, but in the end, at 610 et seq., in favour.The latter (at 613) argues for a stricter nexus between the omission and the realization of the dan-ger, i.e., the violation must appear as realization of the danger which the owner should have pre-vented by his duty of supervision (Gefahrverwirklichungszusammenhang). See also B. Schünemann,Unternehmenskriminalität und Strafrecht (1979), at 120: ‘schutzzweckkongruente Kausalität’(causality limited by and based on the protective purpose of the norm).

150 Göhler, supra note 142, mn. 9, 16 a; Rotberg, supra note 145, mn. 16; Rogall, supra note 142,588–89, 600–601.

151 See Schünemann, supra note 145, 49; see also Rogall, supra note 142, 591, 597, 607, whoseeks an interpretation of § 130 OWiG compatible with the principle of guilt.

827


Furthermore, the construction of the norm as a (negligent) violation of a duty ofsupervision entails its subsidiarity with regard to any conduct of the owner thatconstitutes intentional participation in the crimes committed. In other words, ifthe owner him- or herself participates in a violation, he or she is liable for this participation but not for the less blameworthy failure of supervision, which hisparticipation also might imply.152

Finally, it is worthwhile mentioning Resolution II. 1 of section I of the XVI. AIDPCongress 1999 in Budapest, which once again stresses both the importance anddifficulty of establishing criminal liability of superiors or ‘men in the background’(Hintermänner).

Because it is often difficult to prove that leaders and members of organized crim-inal groups have actually participated in the perpetration of particular offences,traditional forms of perpetration and accessorial liability can be insufficient tomake these individuals accountable. To the extent the traditional law of perpetra-torship and complicity is deemed insufficient, one should consider a cautio[u]smodernization based on the principle of organizational responsibility. In hierarchi-cally structured organizations, persons with decision and control power can bemade responsible for acts of other members under their control if they haveordered these acts to be committed or have consciously omitted to prevent theircommission.153

D. Preliminary Conclusions

In the light of this survey, there can be no doubt that the doctrine of superiorresponsibility, which was first recognized by Yamashita,154 is a well-establishedprinciple of customary international law.155 However, Yamashita also makes clearthat superiors must not be held responsible in contravention of the requirementof personal guilt. From the perspective of an international criminal law compati-ble with general principles of human rights and criminal law, respect for this prin-ciple is of utmost importance and cannot be underestimated. In addition,excessive obligations imposed on superiors can also lead to undesirable overreac-tions, as was demonstrated in the case of Captain Rockwood.156


152 Göhler, supra note 142, mn. 25 et seq.; also Rotberg, supra note 145, mn. 16; W. Ferner,Ordnungswidrigkeitengesetz. Kommentar, Vol. II (loose-leaf, 47th update, February 1999) § 130mn. 2; Rogall, supra note 142, 620.

153 Emphasis added.154 Cf. Parks, supra note 3, 37; Levie, supra note 3, 6; This is even conceded by the critics; see

Landrum, supra note 5, 298.155 See also Green, supra note 3, 350, 371; Levie, supra note 3, 12.156 Cf. O’Brien, supra note 41, passim: Rockwood, misunderstanding his command responsi-

bility, tried to prevent alleged abuses by Haitian soldiers in the National Penitentiary in Port-au-Prince instead of remaining at his place of duty and investigating the killing of Haitians by fellowHaitians.

828


These reflections may in part explain why the subsequent efforts to refine thesuperior responsibility doctrine focused on the mens rea requirement. From theseefforts it is clear that mens rea cannot be established on the basis of a theory of strictor vicarious liability157—as seen by some in Yamashita—nor can positive know-ledge—as in the High Command and My Lai cases—be required. Rather, the new‘should have known’ standard as developed in the Hostages case was taken up in thenegotiations of the PA I158 and confirmed by the ‘reason to know’ standard of theILC and ICTY/ICTR statutes as well as by the ICTY decision in Delalic et al. Inthis sense, one can identify a consensus in the post-war efforts, although the con-cept needs further clarification and refinement.159

In sum, the UNWCC’s criticisms in 1949 that ‘the principles governing this typeof liability . . . are not yet settled’160 has not completely lost its validity, in particu-lar with regard to the criminal law problems inherent to the doctrine.161 Despitethe increasing application of the doctrine since World War II its elements have notbeen defined precisely enough to be indubitably in accordance with the nullumcrimen principle as laid down in the Rome Statute (Articles 22, 24),162 especiallywith its requirement of legal exactness and strictness.163 As a result, the deterrenteffect of the doctrine is weakened.

In light of these conclusions, the central part of this paper attempts to refine theelements of the doctrine and to reduce its ambiguity.

123456789101112131415161718192021222324252627282930313233343536373839404142


157 For further arguments against strict liability in comparison with US environmental law, Wuand Kang, supra note 107, 279–282, in particular because the underlying crimes are no strict liabil-ity offences. See also Parks, supra note 3, passim and 103. Against vicarious liability (pour le faitd’autrui) also the Defence in Aleksovski, supra note 64, para. 72; see also Delalic et al., supra note 63,para. 647.

158 Cf. Crowe, supra note 5, 225 and supra note 104.159 Thus, Hessler’s assessment (at 1281) that there was no consensus of the post-war jurispru-

dence regarding the mental element is not completely correct if one takes into account the codifica-tion efforts.

160 See supra note 53.161 See already De Preux, supra note 103, mn. 3527.162 On this principle, see K. Ambos, ‘General Principles of Criminal Law in the Rome Statute’,

10 Criminal Law Forum (1999) 4–6 with further references.163 Cf. Hessler, supra note 16, 1288: ‘this body of law is too inexact and inconsistent . . . raises

problems of unconstitutional vagueness.’ Similarly critical regarding the principle of legalstrictness in the case of § 130 OWiG also Rogall, supra note 142, 591; Schünemann, supra note145, 48.

829


III. Article 28 Rome Statute

A. Description of the Provision and Comparison with Earlier Codifications

The provision164—‘subject of extensive negotiations and . . . delicate compro-mises’165—distinguishes between the responsibility of (de facto) military com-manders (paragraph (a) ) and other, i.e. non-military or civilian, superiors(paragraph (b) ). As we have seen, the previous codifications or drafts, in particu-lar Article 86(2) of PA I, did not make this distinction, at least not explicitly, butonly referred to the ‘superior’, leaving open the question of civilian or military sta-tus. The distinction in the Rome Statute was inspired by a proposal from the USdelegation, whose fundamental objective was to introduce distinct mens rearequirements for military and civilian responsibility.166 According to the proposalmilitary commanders would be held liable—in accordance with the recognizedstandard—for knowledge or negligence (‘should have known’) but the civiliansuperior only for knowledge. The high threshold for the latter, however, was notaccepted by most delegations taking part in the informal discussions. An informalcounterproposal by Argentina, Canada and Germany argued for a negligence


164 Art. 28 of the Rome Statute (‘Responsibility of commanders and other superiors’) reads(emphasis added):

In addition to other grounds of criminal responsibility under this Statute for crimes within thejurisdiction of the Court:

(a) A military commander or person effectively acting as a military commander shall be criminallyresponsible for crimes within the jurisdiction of the Court committed by forces under his or hereffective command and control, or effective authority and control as the case may be, as a result of his orher failure to exercise control properly over such forces, where:

(i) That military commander or person either knew or, owing to the circumstances at the time,should have known that the forces were committing or about to commit such crimes; and

(ii) That military commander or person failed to take all necessary and reasonable measureswithin his or her power to prevent or repress their commission or to submit the matter to the compe-tent authorities for investigation and prosecution.

(b) With respect to superior and subordinate relationships not described in paragraph (a), a supe-rior shall be criminally responsible for crimes within the jurisdiction of the Court committed bysubordinates under his or her effective authority and control, as a result of his or her failure to exercisecontrol properly over such subordinates, where:

(i) The superior either knew, or consciously disregarded information which clearly indicated, thatthe subordinates were committing or about to commit such crimes;

(ii) The crimes concerned activities that were within the effective responsibility and control ofthe superior; and

(iii) The superior failed to take all necessary and reasonable measures within his or her power toprevent or repress their commission or to submit the matter to the competent authorities for investi-gation and prosecution.

165 UN Doc. A/CONF.183/C.1/WGGP/L.4/Add.1, p. 2. See also P. Saland, ‘InternationalCriminal Law Principles’, in R. S. Lee (ed.), The International Criminal Court: The Making of theRome Statute (1999) 202 et seq.; Schabas, supra note 16, at 417.

166 UN Doc. A/CONF.183/C.1/L.2 (1998); also Saland, supra note 165, 204.

830


standard in both cases.167 The opposing positions were finally merged in informalconsultations chaired by Canada into a compromise formula, which replaced thenegligence standard for the civilian superior with the wording ‘consciously disre-garded information which should have (clearly) enabled him or her to concludein the circumstances of the time that the subordinates were committing or aboutto commit a crime . . .’.168 Subsequently, in a later version, the wording was sim-plified to ‘consciously disregarded information which clearly indicated that sub-ordinates were committing or about to commit such crimes’.169 Thus, a newformula for the mens rea requirement of civilian superiors was invented whichemphasizes the main difference between their responsibility and the responsibil-ity of military superiors.170 As far as the remaining aspects of the mental elementare concerned, the Rome Statute accepts the traditional temporal restriction ‘inthe circumstances at the time’ (Article 86(2) PA I) with regard to military superi-ors using the slightly different phrase ‘owing to the circumstances at the time’.With regard to non-military superiors it omits this reference entirely, therebyemulating the ICTY and ICTR Statutes. However, it is self-evident that mens reacan be proven only with regard to the time of the commission of the crimes by thesubordinates; therefore, the circumstances at that time always have to be takeninto account. In this sense, the explicit reference is redundant. Concerning theobject of the superior’s (possible) knowledge, the formulation ‘about to commitsuch crimes’ prevails over ‘going to commit’. Again, this formulation reflects theICTY and ICTR Statutes, replacing the wording of the PA I and that proposed bythe ILC Draft Codes.

As far as the actus reus of Article 28 is concerned, and compared to the issues iden-tified in the earlier codifications, five objective elements of Article 28 are apparent:

(1) The perpetrator or agent of the offence is a (de facto) military or non-military (civilian) superior who has ‘forces’ or ‘subordinates’ under his orher command; there is no more precise description or delimitation of hisor her status within the military hierarchy; any kind of ‘superior and subor-dinate relationship’ seems to be sufficient.

(2) The ‘command and control’, in the case of the military superior, or the‘authority and control’, in the case of both types of superiors, over the sub-ordinates must be ‘effective’; this restrictive requirement of the superior’sliability is reaffirmed with regard to the civilian superior, who must, inaddition, have ‘effective responsibility and control’ over the activities thatled to the crimes concerned.

123456789101112131415161718192021222324252627282930313233343536373839404142


167 Proposal by Argentina, Canada, and Germany, 17 June 1998 (on file with author).168 UN Doc. A/CONF.183/C.1/WGGP/L.7 (1998). The word ‘clearly’ was only inserted later

in the official working paper.169 UN Doc. A/CONF.183/C.1/WGGP/L.7/Rev. 1 (1998).170 Bassiouni, supra note 3, 422, explains the higher military standard with two policy reasons:

higher standard of discipline in a military structure and the effectiveness of deterrence.

831


(3) The crimes committed by the subordinates are a ‘result’ of the superior’sfailure to exercise proper control over them; this element can be called thecausal requirement.

(4) The superior fails to take the ‘necessary and reasonable measures within hisor her power’ against the crimes committed; the power to take these coun-termeasures obviously derives from the ‘effective’ control.

(5) The countermeasures are supposed to ‘prevent’ or ‘repress’ the commissionof the crimes or the superior has ‘to submit the matter to the competentauthorities for investigation and prosecution’; the latter option was notcontained in the earlier codifications.

If these requirements are fulfilled and the superior had the necessary mens rea, hisor her criminal responsibility is established. Thus, international criminal law nolonger seems to be concerned with possible disciplinary sanctions, as originallyprovided for in Article 86(2) PA I.

In the following section the objective and subjective elements of the offence haveto be analysed in more detail, starting with some general considerations.

B. Critical Analysis and Interpretation

1. General Considerations

Superior responsibility establishes liability for omission. The superior is punishedbecause of the failure to supervise the subordinates and to prevent or repress theircommission of atrocities. This kind of liability—for omission—is unique ininternational criminal law. A general provision defining act and/or omission pro-posed by the Draft Statute171 was deleted in Rome.172 It was argued, inter alia, thatArticle 28 only creates and should create liability for omission.173 The delegatesfollowed the French approach, which rejects general liability for improper omis-sion (commission par omission) because of a strict interpretation of the legalityprinciple.174 Indeed, if the duty to act (Garantenpflicht), which justifies the moralequivalence between the failure to prevent harm and the active causation of harm,is not regulated by statute but only supported by case law and scholarly writings,its compatibility with a strict legality requirement is more than uncertain.175


171 Draft Statute for the International Criminal Court, UN Doc. A/CONF.183/2/Add.1 (14 April 1998), Art. 28.

172 UN Doc. A/CONF.183/C.1/WGGP/L.4/Add.1 (1998).173 Edward Wise, in L. Sadat Wexler (ed.), Model Draft Statute for the International Criminal

Court based on the Preparatory Committee’s Text to the Diplomatic Conference, Rome, June 15–July 171998 (1998), at 48–50, argues for a general norm on omission.

174 See J. Pradel, Droit pénal comparé (1995) 233 et seq. (236). The French position can beexplained with the historical importance of individual freedom and the only slow recognition of theprinciple of solidarity which increasingly allowed for specific (genuine) offences of omission, e.g.Art. 223-6 (2) CP (non-assistance à personne en péril ).

175 See the critical discussion of Fletcher, supra note 146, 47–48; critical also Pradel, supra note174, 236–237.

832


Article 28 can be characterized as a genuine offence or separate crime of omission(echtes Unterlassungsdelikt),176 since it makes the superior liable only for a failure ofproper supervision and control of his or her subordinates but not, at least not‘directly’, for crimes they commit. These crimes are ‘directly’ imputed to the sub-ordinates, the superior is only responsible for the failure to prevent them fromoccuring; that means that the superior is not liable for an improper omission or—to use a better term—a commission by omission (unechtes Unterlassungsdelikt) aclassical example of which is the father who lets his child drown and, therefore,would be guilty of manslaughter by omission. This characterization of Article 28is also supported by the fact that the structurally similar provisions of Article 41WStG and § 130 OWiG are also considered separate offences of omission.177

Finally, Article 28 is much more precise than the general provisions that form thebasis of offences of commission by omission, e.g. § 13 Criminal Code(Strafgesetzbuch-StGB).178

Thus, Article 28 is a separate crime of omission179 that consists, on an objectivelevel, of the superior’s failure properly to supervise subordinates. The underlyingcrimes of the subordinates are neither an element of the offence nor a purelyobjective condition of the superior’s punishability. Rather, they constitute thepoint of reference of the superior’s failure of supervision; therefore, a specificcausal relationship between the failure and the occurrence of the crimes mustexist.180 Unlike in the case of the company owner, the superior’s mens rea need notextend only to the failure to supervise, which creates the risk or danger that thesubordinates commit crimes,181 but also to the underlying crimes themselves.This clearly follows from the wording of Article 28.182 Indeed, Article 28 has apeculiar structure in that it extends the superior’s mens rea, beyond his or her own

123456789101112131415161718192021222324252627282930313233343536373839404142


176 See for more information on the terminology and difference between ‘echt’ and ‘unecht’:H.-H. Jescheck and T. Weigend, Lehrbuch des Strafrechts, Allgemeiner Teil (5th edn., 1996) at605–608; S. Mir Puig, Derecho Penal, Parte General (5th edn., 1998) at 299–301; also Pradel, supranote 174, 233–234. C. Kreß, ‘Die Kristallisation eines Allgemeinen Teil des Völkerstrafrechts: DieAllgemeinen Prinzipien des Strafrechts im Statut des Internationalen Strafgerichtshofs’, 12Humanitäres Völkerrecht (1999) 4–10 (9) takes the view that Art. 28 is an ‘unechtesSonderunterlassungsdelikt’, i.e. a special offence of improper omission or—better—a crime of com-mission by omission. As will be shown in the text, this view is not correct.

177 See supra note 129.178 § 13(1) reads: ‘Whoever fails to prevent a harm which is part of the constitutent elements of

a crime may be punished under this Code only if he was under a legal duty to prevent the harm, andif his failure to act was equivalent to an affirmative act for purposes of establishing the statutorilydefined constituent elements of the crime’ (translation by Darby, The American Series of ForeignPenal Codes, Vol. 28, 1987).

179 Cf. Wu and Kang, supra note 107, 289.180 See already the similar interpretation of § 130 OWiG, supra, II.C.2. as well as infra, 2(c).181 See supra, II.C, notes 136, 150 and accompanying text. Similarly Wu and Kang, supra note

107, 289.182 Art. 28(a)(i) and (b)(i): The superior ‘knew or, owing to the circumstances at the time, should

have known’ or ‘consciously disregarded information which clearly indicated’ that the subordinateswere committing or about to commit such crimes. See supra note 162 and infra, 3.

833


failure to supervise, to the concrete acts of the subordinates. This structure is over-looked by those who consider superior responsibility a form of complicity.183 Inaddition, it is not sufficiently reflected in the case law which, as has been shownelsewhere,184 normally does not clearly distinguish between accomplice liabilityand liability for failure to supervise. However, simultaneous liability can concep-tually only arise if the mental object of accomplice liability and liability for omis-sion are the same, i.e. if both arise in respect of the crimes committed by thesubordinates. Given the structure of the offence, this is normally not the case: inthe case of superior responsibility, the main object of the offence is the superior’sfailure properly to supervise and, consequently, his or her mens rea needs to extendto this failure; in contrast, in the case of complicity, the superior takes part in thecrime of the subordinates and must share the subordinates’ mens rea, i.e. he or shemust have mens rea with regard to his or her own contribution and with regard tothe main offence. If this is the case, then and only then could the superior be liableas an accomplice; in such a case liability for superior responsibility would be sub-sidiary.185

Further, the peculiar structure of superior responsibility leads to a stunning con-tradiction between the negligent conduct of the superior and the underlying intentcrimes committed by the subordinates. In such a case we are dealing, as Schabasrightly points out,186 with a commission of ‘a crime of intent by negligence’:187 thesuperior fails negligently to properly supervise the subordinates, who commit theircrimes intentionally within the meaning of Article 30 of the Rome Statute. Thus,the question arises how the intent crimes of the subordinates can be imputeddirectly to the superior. In such a case, at least mitigation of punishment must beconsidered, since the superior’s conduct is less blameworthy than in the case ofintentional conduct, i.e. if he or she is aware of the commission of offences by thesubordinates. Indeed, it would be more logical not to find the superior criminallyliable for the intent crimes of the subordinates at all but this solution is not com-patible with—certainly contradictory—wording of article 28. As to the classifica-tion of the superior’s conduct in such a case, the superior cannot be considered amere accomplice (aider and abettor) by omission to the crimes of the subordinates


183 See most recently Fenrick, supra note 76, mn. 2; Bantekas, supra note 41, 577. ContradictoryWu and Kang, supra note 107, who argue, on the one hand, for accomplice liability (278 et seq.) andfollow, on the other, the view expressed in this paper (at 289).

184 See Ambos, supra note 162, at 20. The case law holds the superior either responsible as anaccomplice (aiding and abetting) for, at least psychologically, encouraging and supporting thetroops, or for failure to prevent the crimes on the basis of superior responsibility.

185 See also the identical view with regard to § 130 OWiG, supra, II.C, notes 150–152 andaccompanying text.

186 Schabas, supra note 16, 417.187 See also the example given by K. Howard, ‘Command Responsibility for War Crimes’, 21

Journal of Public Law (1972) at 20: ‘the punishable act in the case of the subordinate is the unlawfuldeath of the . . . person whereas in the case of the commander . . . the negligent omission . . . torestrain the act of the subordinate’.

834


according to Article 25(3)(c), since complicity requires intent (Article 30) and heor she only acts negligently. Thus, in fact the superior is liable because of a merenegligent failure of supervision. This is certainly a very broad liability.

The superior’s liability for his or her omission stands and falls—on an objectivelevel—with his or her effective authority and control: the possibility of controlforms the legal and legitimate basis of the superior’s responsibility; it justifies hisor her duty of intervention. In fact, the Garantenpflicht as described above can bederived directly from Article 87 PA I, since it establishes the duty of commandersto prevent crimes committed by their subordinates. In a way, therefore, one canspeak of a legal or positive duty to act, since the Garantenpflicht is based on a pos-itive norm of treaty law which, in addition, is regarded as customary law. Thisgeneral duty to act is complemented by the various specific rules of positive con-duct as laid down in the PA I.188 Although these rules were initially addressed onlyto State Parties, they can now be considered the basis of rules of responsibility foran individual’s failure to act, since the doctrine of superior responsibility and themajor part of the offences established by the Geneva law (including PA I) havebeen ‘individualized’ by the Rome Statute and by national implementation laws.As a result, broad liability has been established that can only be restricted, as willbe examined in more detail below, by the elements of control and the requiredmens rea.

Finally, if one wants to compare the superior responsibility concept with a posi-tive act, it relates systematically to the liability for ordering a crime. In both casesthe criminal responsibility of a superior is at stake: in the case of ‘ordering’ for apositive act—the order to commit a certain crime—and in the case of superiorresponsibility for an omission (non-act)—the failure to prevent the crime. In thissense, ordering crimes and failing to prevent them, although conceptually dis-tinct, seem to be different sides of the same coin.189 The practical similarity of thetwo forms of criminal conduct is also reflected in the case law analysed in this arti-cle.190 Normally, superior responsibility cases are accompanied by orders to com-mit certain acts; if such orders cannot be proven or attributed to the defendant,superior responsibility serves as a kind of subsidiary form of criminal liability.191

Thus, the Secretary-General stated in his report on the establishment of the ICTYthat a person in the position of superior authority should either be held respons-ible for giving the unlawful order to commit a crime or for failure to prevent a

123456789101112131415161718192021222324252627282930313233343536373839404142


188 De Preux, supra note 103, mn. 3536; for further sources of command duties, see Bantekas,supra note 41, at 576–577. See for the similar interpretation with regard to the responsibility of acompany owner (§ 130 OWiG ), supra note 144.

189 Cf. Green, supra note 52, 167. See also Eckhardt, supra note 56, at 4–5; W. H. Parks, ‘A FewTools in the Prosecution of War Crimes’, 149 Military Law Review (1995) 73, at 75.

190 See e.g. US v. List et al., supra note 36, at 1257: ‘Those responsible for such crimes by order-ing or authorizing their commission, or by a failure to take effective steps’.

191 See also Wu and Kang, supra note 107, 274.

835


crime.192 The ILC distinguishes between a direct and indirect contribution to thecommission of a crime.193 Similarly, a resolution of the most recent AIDPCongress in Budapest recognized the parallel importance of ordering and failingto prevent crimes within criminal organizations.194 This resolution also remindsus of the importance of the so-called doctrine of domination by virtue of organ-izational control (Organisationsherrschaftslehre) with regard to the criminal attri-bution of positive acts within a criminal organization.195 However, the differencebetween ordering and failing to prevent, in legal terms, lies in the fact that respon-sibility in the case of the former is accepted in international criminal law,196 whilethe latter still needs further refinement.

Although superior responsibility is ‘the legal and logical concomitant to thedefence of “obedience to superior orders” ’,197 there is no logical paradox betweenthe parallel responsibility of the superior for giving an illegal order and of the sub-ordinate for executing this order. This view, defended by D’Amato and correctlyrejected by Levie,198 overlooks the different forms of participation developed incriminal law doctrine, especially in the so-called differentiated model of theEuropean continent, which is also recognized by Article 25 of the RomeStatute.199 Such a differentiated view of criminal participation makes it perfectlypossible to consider the superior an accomplice (instigator) or indirect perpetra-tor (perpetration by means) and the subordinate a direct perpetrator. The fact thatthe superior’s order is illegal and, therefore, should not be executed by the subor-dinate is only a normative expectation and cannot absolve the superior fromresponsibility. Otherwise, any crime committed within the framework of a hier-archical relationship—e.g. the killing by a hitman on the order of a mob leader—would entail impunity for the ‘man in the background’ whose conduct is normallyconsidered even more blameworthy than that of the direct perpetrator.

2. Actus Reus

(a) The Status of the SuperiorThe Nuremberg jurisprudence limited the criminal responsibility for acts ofothers to commanding generals and officers, excluding military personnel with-


192 S/RES 827 (1993), 25 May 1993, reprinted at 32 ILM 1203 (1993), para. 56.193 Report ILC 1996, supra note 111, at 35, para. 1 = YILC (1996) Vol. II, Part 2, at 25, para. 1.

Similarly Wu and Kang, supra note 107, 289: ordering as direct command responsibility; alsoBassiouni, supra note 3, at 419; Vetter, supra note 38, 97, 99.

194 See Resolution II.1, supra note 153 and accompanying text.195 See C. Roxin, Täterschaft und Tatherrschaft (6th edn., 1994) at 242–252, 653–654. I have

already referred to the importance of this doctrine for international criminal law in the article‘Individual Criminal Responsibility’, supra note 85, at 18–19.

196 See Bassiouni, supra note 3, at 441; concurring Wu and Kang, supra note 107, 272, 289.197 Bassiouni, supra note 3, 420; also Vetter, supra note 38, 102.198 See Select Bibliography.199 See K. Ambos, ‘Article 25: Individual Criminal Responsibility’, in O. Triffterer (ed.),

Commentary on the Rome Statute of the International Criminal Court (1999) mn. 2, 7 et seq.

836


out command power.200 With regard to aggressive war, the defendant had to be a‘leader’ or ‘planner’ or had to belong to the ‘policy level’.201 The IMTFE limitedthe responsibility to cabinet members;202 the ICTY jurisprudence requires ‘a posi-tion of authority’203 or ‘hierarchical power’.204 Thus, the international case lawdoes not really address the problem of whether the superior has to belong to a certain hierarchical level and, if so, to which level. Although the post-war juris-prudence seems to limit the responsibility to the leadership or the policy level, itdoes so implicitly and does not explicitly exclude commanders below this level. InPohl et al. staff officers are excluded, since they normally have no commandingpower.

As to the codification efforts, the ILC clearly stated that Article 6 of its 1996 DraftCode refers to any superior in a chain of command, not only the immediate supe-rior within a certain hierarchy.205 Similarly, Article 87(1) PA I has been interpreted‘to refer to all those persons who had command responsibility, from commandersat the highest level to leaders with only a few men under their command’.206 Thisimplies that the superior–subordinate relationship requires a chain of com-mand,207 but it does not mean that all persons in the chain are equally responsible,regardless of the concrete circumstances and the level of hierarchy involved,208 butonly makes clear that the concept of superior responsibility is not per se limited toimmediate superior on the higher echelons of command. This view is supportedby the wording of the corresponding provisions in other codifications. Article 6 ofthe ILC Draft Code 1996, for example, refers to the responsibility of all superiors(in plural) of the subordinate who committed a crime.209 Similarly, Article 28(b)of the Rome Statute refers to ‘superior and subordinate relationships’; the mostimportant feature of this relationship is the superior’s commanding power (con-trol/authority) over the subordinates, not the particular rank of the superior.210

123456789101112131415161718192021222324252627282930313233343536373839404142


200 See US v. Pohl & Others, II.B.2. supra note 31 et seq.201 See supra, II.B.2.(a).202 See ibid., supra note 46.203 See Delalic et al. as well as Blaškic, both supra note 66 and text.204 Aleksovski, supra note 64, para. 78.205 Report ILC 1996, supra nota 111, at 37, para. 4 = YILC (1996) Vol. II, Part 2, at 25–26, para

4.206 See De Preux, supra note 103, mn. 3553 quoting from the Official Records; see also mn.

3561: ‘Every commander at every level has a duty to react’. Similar Partsch, supra note 108, 528.207 Bantekas, supra note 41, 580–581, 584.208 De Preux, supra note 103, mn. 3554; see also A. P. V. Rogers, Law on the Battlefield (1996) at

142; Bassiouni, supra note 3, at 421. Similarly, Röling, supra note 121, 15, argues for a ‘specialresponsibility for the field in question’.

209 See supra, II.A.2.210 Cf. Fenrick, supra note 76, mn. 4, 15. Similarly, it was stated in US v. von Leeb et al., supra

note 35, at 489: ‘It is not a person’s rank or status, but his power to shape or influence the policy ofhis state, which is the relevant issue for determining his criminality under the charge of crimesagainst peace.’

837


In other words, the minimum requirement of command responsibility is that thesuperior concerned have command.211 Although this seems to be evident giventhat the term ‘command’ is part of the concept itself, this does not necessarilymean that command entails responsibility. This implication only follows from thelink between the two. Thus, as stated in the US Field Manual: ‘The military com-mander has complete and overall responsibility for all activities within his unit.’212

On the other hand, in the case of non-military superiors the Rome Statute speaksof ‘authority and control’ and consequently the title of Article 28 is not ‘commandresponsibility’ but ‘responsibility of commanders and other superiors’. If onewants to use a more concise title that covers both types of superiors, one may speakof ‘responsibility by authority and control ’, since this criterion is used for both typesof superiors; a handier formula which has been used throughout this text wouldbe, though, superior responsibility. The requirement of command also gives thesuperior–subordinate relationship a new quality. The question is no longer whatkind of command the superior must exercise but whether he or she possesses‘authority and control’ over subordinates; this does not exclude the possibility thatthe superior concerned is him- or herself a subordinate.213

One may argue that it clearly follows from the fact that a person has subordinatesthat he or she also has authority or control over them. However, this is not neces-sarily true. The argument is too formal: Article 28—like the ICTY and ICTRjurisprudence—does not refer to the formal status of the superior alone but to thede facto status (‘effectively acting as . . .’).214 Thus, the provision does not refer to asituation which ought to be (the normative ideal world) but rather to the situationas it really is (the real world). In conclusion, the subject or agent of the concept ofsuperior responsibility as understood by Article 28 is the superior who has effec-tive ‘authority and control’ over his or her subordinates, independent of his or herformal status;215 the concept of a superior should be seen ‘in terms of a hierarchyencompassing the concept of control’.216 Thus, the superior’s status is intimatelylinked to the element to be analysed in the next section.


211 On the sources of de iure command, see Bantekas, supra note 41, 578–579 distinguishingbetween policy, strategic, operational, and tactical command.

212 § 22 FM 22-100 quoted according to Howard, supra note 187, 14 (emphasis added).213 Fenrick, supra note 76, mn. 17214 Cf. ibid., mn. 5, 18; see also Wu and Kang, supra note 107, 292–293 and Bantekas, supra

note 41, 579–80.215 See the convincing definition of Bantekas, supra note 41, 582: ‘Being feared by others and

enforcing one’s might over others renders an individual superior to those with lesser power orgreater fear. . . . actual and effective subordination as a result of the exercise of one’s influence overpersons is necessary.’

216 De Preux, supra note 103, mn. 3544. Rogers, supra note 208, 140, wants to distinguishbetween Art. 86 para. 2 and Art. 87 para 3. PA I in this respect. The two provisions, however, haveto be read together.

838


(b) Effective ‘command and control’ or ‘authority and control’The first question is if there is a difference between the concepts of ‘command andcontrol’ (paragraph (a) of Article 28) and ‘authority and control’ (paragraph (b) ).This question is only relevant with regard to military superiors, since only in thiscase are both concepts applicable. As both concepts contain the term ‘control’, itis clear that control is linked to or dependent on command or authority. In a way,‘control’ is a kind of umbrella term encompassing command and authority. Whilethe command refers in a rather material sense to ‘an order, a directive’, possiblybacked by threats;217 ‘authority’ seems to have a rather formal meaning in the senseof the ‘the right or permission to act legally’.218 Both terms imply control: com-mand explicitly as power to control, authority implicitly as a right to command.Thus, a superior with command and authority normally controls his or her‘forces’219 or subordinates220 and has the capacity to issue orders.221

This is less clear if military functions are split, e.g. if a commander only exercisesoperational but no administrative control. In this case it is expected that the com-mander will take all measures which are within his or her physical power to pre-vent the commission of crimes by subordinates. A lack of administrative controldoes not foreclose or preclude the use of all other measures at his or her disposal.222

This applies vice versa if the commander has administrative but lacks operationalcontrol. In this case, he or she is expected at least to use the available administra-tive means or sanctions to prevent the commission of crimes. A duty to act mayonly be rejected if there is no control at all. This may be the case if the subordinatesare totally out of control and no longer obey the orders of the superior, commit-ting widespread or isolated excesses.223 It is also the case for a military adviser whohas neither operational nor administrative control. Such an advisor can only beexpected to report the commission of crimes to the commanding authority or toresign his or her position.224 On the other hand, the commander’s duty to actrefers to the entire territory over which he or she has executive authority and con-trol, e.g. in case of an occupation, even if this territory reaches beyond his or her

123456789101112131415161718192021222324252627282930313233343536373839404142


217 Black’s Law Dictionary (7th edn., 1999) at 262. Similar Fenrick, supra note 76, Art. 28 Rn.124.

218 Black’s Law Dictionary (7th edn., 1999) at 127.219 The term ‘forces’ used in relation to military commanders include armed forces, groups, or

units under a command (Art. 43 PA I) as well as police and paramilitary units (Fenrick, supra note76, mn. 6).

220 Cf. Parks, supra note 3, 83.221 See Bantekas, supra note 41, 582–583.222 Parks, supra note 3, 84–85; Bantekas, supra note 41, 585–586.223 Cf. Schünemann, supra note 145, 45, who refers in the economic context to excesses due to

the subordinate’s ‘emancipation’. According to Rogall, supra note 142, 619, in this situation theunlawful result cannot be attributed to the superior.

224 Cf. Parks, supra note 3, 86. See also De Preux, supra note 103, mn. 3557, who correctlyemphasizes that advisers are to advise, not ‘to replace’ the commander; also Partsch, supra note 108,529.

839


own (formal) command.225 This view is confirmed by the reference to ‘authorityand control’ in para. (a) and (b) of Article 28.226 Further, the superior must activelysupervise certain trouble spots.227

The control (command, authority) has to be ‘effective’.228 According to Fenrickthis word ‘is intended to encompass both de [i]ure and de facto command and toensure that, when multiple chains of command appear to exist, responsibility isassigned to the chain of command wherein resides the power to give orders . . .’.229

As Article 28(b)(ii) shows, the requirement of effective control is particularlyimportant with regard to non-military superiors.230 The repetition of this require-ment makes clear that the drafters wanted to limit the liability of a civilian as com-pared to a military superior already on the objective level. There is no effectivecontrol, for example, with regard to non-work related activities of the subordi-nates.231 In this sense, Wu and Kang argue that the scope of the relationshipbetween superior and subordinate should be considered as a limitation of thesuperior’s duty of control ‘to that which is part of their relationship’.232 This seemsto be applicable to a civilian, but not to a military relationship since the latterencompasses almost all of the subordinate’s conduct. In general, control in civil-ian hierarchies is less strict than in military ones.233 A civilian superior does notnormally have the power to sanction subordinates in the same way as does a mili-tary superior; therefore, as stated by the ICTY in Aleksovski, the same power ofsanction cannot be a requirement of superior responsibility for civilians.234

A delegation of the duty of supervision does not absolve the superior of responsi-bility;235 rather this duty is converted into a duty of proper selection, instruction,and follow-up control. This view is supported by the Tokyo judgment, whichdeveloped ‘systemic’ duties,236 and by the German doctrine concerning Article 41WStG and § 130 OWiG.237 Similarly, Article 12(4) of the recent Corpus Juris draft(2000) states:


225 Parks, supra note 3, 85–86; also De Preux, supra note 103, mn. 3555; Rogers, supra note 208,at 141; Bantekas, supra note 41, 586.

226 Fenrick, supra note 76, mn. 8.227 See Eckhardt, supra note 56, 23–24 with various examples.228 See also Bantekas, supra note 41, 580.229 Fenrick, supra note 76, mn. 7.230 See also Vetter, supra note 38, at 115 et seq.231 Fenrick, supra note 76, mn. 19.232 Wu and Kang, supra note 107, 295.233 See Fenrick, supra note 1, 117: ‘Most bureaucratic leaders do not wield the same type of life

and death authority.’234 See supra note 74.235 In particular if the delegation only serves the purpose to evade criminal responsibility, see

Bantekas, supra note 41, 585.236 Wu and Kang, supra note 107, 293–94; see also Green, supra note 3, 355 and supra, II.B.2.237 See II.C, supra notes 133, 147 and accompanying text.

840


. . . the fact that he [the head of business] delegated his powers shall only be a defencewhere the delegation was partial, precise, specific, and necessary for the running ofthe business, and the delegates were really in a position to fulfil the functions allotedto them. Notwithstanding such a delegation, a person may incur liability under thisarticle on the basis that he took insufficient care in the selection, supervision or con-trol of his staff, or in the general organisation of the business, or in any other matterwith which the head of business is properly concerned.

The delegation of powers and criminal liability is only valid as a defence if it is par-tial, precise, and specific . . . and if the delegates are really in a position to be ableto fulfil the functions delegated to them. Such delegation does not exclude thegeneral responsibility of monitoring, supervision, and selection of personnel, anddoes not include matters proper to the head of business such as general organiza-tion of work within the business.238

Still another question is how far the, once established, duty of a superior goes. It isbeyond doubt that the superior must act to prevent crimes committed by the sub-ordinates by positive acts. But what about crimes committed by omission? Takethe case of a soldier responsible for the death of a prisoner of war by starvation. Orthe case of a soldier who does not intervene in a crime committed by another sol-dier without sharing the mens rea of the perpetrator, i.e. without being an accom-plice. Admittedly, the problem is not of great practical importance since the RomeStatute does not contain a general norm on omission, it only recognizes genuineoffences of omission as established in Articles 5 to 8. Still, there are some crimesthat do not necessarily require a positive act but for which an omission suffices;239

the superior’s duty to intervene encompasses these crimes of omission as well sinceArticle 28 refers in a general way to ‘crimes within the jurisdiction of the Court’.One can take these examples even further by imagining a situation in which a sol-dier refuses to help a civilian or a comrade in arms. Would this soldier be liable fora failure to fulfil the general duty to render assistance as recognized in criminal law(see e.g. § 323 c StGB)? Would the superior consequently be liable for not havingprevented the soldier’s failure to render assistance? Arguably, such cases will rarelybecome practical given the evidentiary problems involved. Besides, the superiormay have the necessary mens rea only with regard to the really heinous and wide-spread crimes. Nevertheless, the examples show that clarification is needed con-cerning the types of crimes that fall within the superior’s duty to intervene.

(c) Crimes as a ‘result’ of the Superior’s Failure to Exercise Control ProperlyThe requirement that the crimes of the subordinates be ‘a result’ of the superior’s‘failure to exercise control properly’ implies a causal relationship between the

123456789101112131415161718192021222324252627282930313233343536373839404142


238 See Delmas-Marty and Vervaele (eds.), supra note 138, at 193. See also § e(3) of theTiedemann proposal, supra note 141, at 511.

239 For example, ‘causing great suffering’ (Art. 82(a)(iii) ) can be committed by the failure to givefood and water to the victims.

841


superior’s failure and the subordinate’s commission of crimes. It correspondsexactly to the kind of causation required by the German Article 41 WStG and § 130 OWiG.240 In concrete terms, the prosecution—in accordance with the gen-erally recognized conditio formula or ‘but for’ test241—must prove that the crimeswould not have been committed if the superior had properly supervised the sub-ordinates. Thus, the conditio formula must be inverted. While normally a positiveact causes a certain consequence, i.e. the consequence would not have occurredwithout this act, in the case of omission the argument goes the other way around:the omission ‘causes’ the consequence, since the omitted act would have pre-vented it from occurring. Admittedly, this formula conflicts with the understand-ing, based in philosophical-naturalistic considerations, that omissions cannotdisplay ‘causal energy’;242 however, it must be recognized on the basis of a legal-normative concept of causation.243 Indeed, the inverted conditio formula is alsorecognized, as was mentioned at the beginning of this section, in Article 41 WStGand § 130 OWiG. As a result, it is sufficient that the superior’s failure of supervi-sion increases the risk that the subordinates commit certain crimes. Any higherstandard would overstretch the causation requirement, since we deal with a hypo-thetical causation of events ‘in an imaginary world’:244 it is empirically impossibleto say what would have happened if the superior had complied with the duty ofsupervision. In other words, the existence of an exact causal relationship betweenthe failure of supervision and the commission of the crimes can hardly be provenex post.245 It can only be said for sure that the necessary causal relationship does notexist if the subordinates commit excesses and do no longer obey the orders of thesuperior.246

This does not mean, however, that the traditional criticism of the conditio formuladoes not apply here. One can imagine, for example, a scenario in occupied terri-tory in which crimes have been committed, either by improperly supervised sub-ordinates or, in their absence, by hostile insurgent forces operating in that sameterritory. In such a case of alternative causation (alternative causes for the sameresult) the responsible superior could claim that the crimes would have been com-mitted anyway, even if the subordinates had been properly supervised. Thisdefence, however, would only apply if the two groups acted simultaneously in thecommission of the crimes since only then could one speak of a true alternative


240 See supra, II.C, notes 129 and 142 et seq. and accompanying text.241 Cf. Fletcher, supra note 148, 62–63.242 Cf. ibid., at 67.243 See for the predominant German doctrine, Jescheck and Weigend, supra note 176, at

618–619.244 Fletcher, supra note 148, at 63.245 Cf. Schünemann, supra note 145, 48; Rogall, supra note 142, 611; Göhler, supra note 142,

mn. 22.246 See already supra note 223 and accompanying text.

842


causation.247 In any other situation, i.e. the group that reached the place of com-mission first committed the crimes, criminal responsibility would be attributed,accordingly, to that group. A hypothetical alternative causation would be of noimportance. The example shows, however, that there are cases where the pure(inverted) conditio formula could lead to unsatisfying results. In such cases, nor-mative theories of (objective) attribution248 or the proximate cause doctrine249

could be helpful.

(d) ‘Necessary and reasonable’ Countermeasures ‘within his or her power’The superior must take countermeasures that are ‘feasible’ (article 86(2) PA I; arti-cle 12 Draft Code 1991), ‘necessary’ (article 87(3) PA I; article 6 Draft Code1996, article 7(3) ICTY Statute and article 6(3) ICTR Statute) or ‘reasonable’(ICTY and ICTR Statutes). There is no substantive difference between thesequalifiers. This has been confirmed by the ILC with regard to ‘feasible’ or ‘neces-sary’ measures.250 In any case, the superior must have both the ‘legal competence’and the ‘material possibility’ to prevent or repress the crimes.251 It is obvious thatthese two requirements enable the superior to raise an important defence.252

According to Fenrick253 the commander must do in particular the following:

• ensure that the forces are adequately trained in international humanitarianlaw

• ensure that due regard is paid to international humanitarian law in operationaldecision-making

• ensure that an effective reporting system is established so that he or she isinformed of incidents when violations of international humanitarian lawmight have occurred

• monitor the reporting system to ensure it is effective and• take corrective action when he or she becomes aware that violations are about

to occur or have occurred.

123456789101112131415161718192021222324252627282930313233343536373839404142


247 See the example given by Fletcher, supra note 148, at 63.248 These theories try to limit attribution by normative criteria (cf. Roxin, Strafrecht, Allgemeiner

Teil I (3rd edn., 1997), sec. 11 mn. 39–136; see also Fletcher, Rethinking Criminal Law (1978) at495–496). For our context in particular, see Rogall and Schünemann, as quoted supra note 149 andaccompanying text.

249 Fletcher, supra note 148, 64–67.250 See supra, II.A.2, note 113 and accompanying text.251 Report ILC 1996, supra note 111, at 38–39, para. 6 = YILC (1996) Vol. II, Part 2, at 26, para.

6 (emphasis added); similar Partsch, supra note 108, 525. See also the ICTY’ position, supra note71.

252 Cf. L. C. Green, ‘Article 12: Responsibility of the Superior’, in M. C. Bassiouni (ed.),Commentaries on the International Law Commission’s 1991 Draft Code of Crimes against the Peace andSecurity of Mankind (1993) at 195–196 (196); see also Hessler, supra note 16, 1285 et seq.: lack ofphysical control or legal authority as an excuse.

253 Fenrick, supra note 76, mn. 9.

843


In any case, only those measures can be expected that are within the superior’spower (Article 86(2) PA I) and covered by the superior’s command and control(Article 87(1), (3) PA I) as defined above. Thus, the concrete measures depend onthe superior’s position in the chain of command.254 If the superior is him- or her-self responsible for the situation that left him or her powerless to prevent a crime,he or she must be held responsible ‘for allowing such a situation to develop’.255 Thesuperior must display good faith in meeting his or her responsibilities.256

The time of commission of the offences is also important as will be seen in the nextsection.

(e) ‘Prevent’, ‘repress’, or ‘submit to the competent authorities’Besides ‘prevent’ and ‘repress’ (Articles 86(2) PA I and 87(3) PA I), the terms ‘sup-press’ and ‘report’ have been used in earlier codifications (Article 87(1) PA I). Incontrast, the formulation ‘submit to the competent authorities’ is new; however,it corresponds in substance to the earlier ‘report’ requirement. It fills a gap in thatit formulates a specific duty for those superiors who have themselves no discipli-nary powers to ‘repress’ a crime. Since this is not limited to non-military superi-ors, the formulation has been included in both paragraphs (a) (ii) and (b) (iii) ofArticle 28.257

Whether the superior has to ‘prevent’ or to ‘repress’ depends on the concrete cir-cumstances of each case. In general, the following distinction can be made:258 if acrime has not yet been committed (is ‘about’ to be committed), the superior isobliged to intervene, e.g. by issuing the appropriate orders; if the crime has alreadybeen committed, the superior can only react with repressive measures, i.e. orderan investigation and punish the perpetrators or submit the matter to the compe-tent authorities. If the subordinates are ‘committing’ crimes, as formulated inArticle 28(a)(i) and (b)(i), the superior may still be able to ‘prevent’ the continua-tion of the commission and ‘repress’ the crimes already committed.

Article 28 does not adequately reflect the difference between preventive andrepressive measures in its requirement that the superior ‘prevent or repress theircommission’. ‘Repress’ in this context seems to imply only that the superior muststop the (ongoing) commission of the crimes but not punish the perpetrators,


254 Fenrick, supra note 76, mn. 12.255 See Wu and Kang, supra note 107, at 296.256 Fenrick, supra note 76, mn. 14, 25.257 This seems to overlook Fenrick who comments on this formulation—at least explicitly—

only in relation to non-military superiors, supra note 76, mn. 25.258 See also Fenrick, supra note 76, mn. 12–14; Bantekas, supra note 41, 591–2. On the tempo-

ral factor, see already supra, II.A.2.

844


since the crimes have not been committed yet.259 Similarly, the ILC does not dis-tinguish between ‘prevent’ and ‘repress’ stating that both entail the initiation ofdisciplinary or penal sanctions.260 However, although Article 28—like the earliercodifications—clearly requires both preventive and repressive action,261 this mustnot blur the distinctive line between the different kinds of measures required. Inany case, it would go too far to impose on superiors the duty to either discover orpredict the conduct of their troops unless crimes are likely to occur.262

3. Mens rea

The ‘unless otherwise provided’ clause in the general provision on mens rea(Article 30 of the Rome Statute) allows for other subjective standards: Article 28makes use of this option.263 As has been pointed out, it establishes different mensrea standards for military and civilian superiors, introducing a totally new stan-dard with regard to the latter. While the mens rea requirement would hardly beproblematic if the superior had positive knowledge, any reliable standard belowthe positive knowledge requirement is highly controversial and difficult to estab-lish. Taking into account earlier codifications, recent jurisprudence and doctrine,this section attempts to examine and clarify the different standards.

(a) Military and Non-Military Superior: Positive KnowledgeAccording to Article 30(3) of the Rome Statute, knowledge means ‘awareness thata circumstance exists or a consequence will occur in the ordinary course of events’.As to the evidentiary method which is used to prove actual knowledge, the ICTYhas correctly stated that knowledge may be established by circumstantial evidence(indicia) but it may not be presumed.264 If knowledge is not real, but only pos-sible,265 proof that it exists can only be based on facts not on mere presumptions.Otherwise, one would opt for a presumption of the knowledge of the superior andthus violate the principle of guilt.266

123456789101112131415161718192021222324252627282930313233343536373839404142


259 See Van den Wyngaert, in Bassiouni (ed.), Commentaries (1993), at 59, with regard to theDraft Code 1991. In contrast, according to Partsch, supra note 108, 524, this is the meaning of ‘sup-press’ while repress refers exclusively to penal measures.

260 Report ILC 1996, supra note 109, at 36–37, para. 3, 4 = YILC (1996) Vol. II, Part 2, at 25–6,para. 3, 4.

261 De Preux, supra note 103, mn. 3548.262 Cf. Bantekas, supra note 41, 593–4.263 See D. Piragoff, in Triffterer (ed.), supra note 199, Art. 30 mn. 14–15.264 See supra, II.B.3. notes 75–79. Cf. Fenrick, supra note 76, mn. 10; Bantekas, supra note 41,

587–489.265 Critically Green, supra note 252, at 195.266 De Preux, supra note 103, mn. 3546, apparently, supports such an interpretation when, in

referring to the war crimes jurisprudence, he states: ‘taking into account the circumstances, a know-ledge of breaches committed by subordinates could be presumed.’ See also L. C. Green, TheContemporary Law of Armed Conflict (1993) at 271. Even further goes Bantekas, supra note 41, 590,594 identifying an ‘emerging rule of customary law’ in this respect.

845


Actual knowledge proven by circumstantial evidence has to be distinguishedstrictly from so-called constructive knowledge. As will be seen below, constructiveknowledge has nothing to do with actual knowledge but belongs to the lower‘should have known’ standard.

(b) Military Superior: ‘should have known’The ‘should have known’ standard reintroduces the standard unsuccessfully pro-posed by the ICRC during the negotiations of the PA I267 and contained in theAmerican and British military manuals.268 As mentioned above, this standard canalso be traced back to the Hostages case, Article 86(2) PA I (‘information whichshould have enabled them to conclude’) and the ‘reason to know’ standards of theILC and ICTY.

The wording of the PA I gave rise to some confusion since the French text—’desinformations leur permettant de conclure’ -, covers apparently only the objectiveelement whereas the English—’should have enabled him’—embraces both objec-tive and subjective elements. According to the commentaries the French versionshould, in light of the object and purpose of the treaty,269 prevail.270 However, thetwo versions do not really have a different meaning in the sense of Article 33(4) ofthe Vienna Convention on the Law of Treaties. This is confirmed by both the ILCand the ICTY. While the former interpreted both versions equally,271 the ICTYindicated in Delali_ et al. that the textual difference was not considered as ‘one ofsubstance’ by the delegates of the Protocols conference.272 In fact, the French ver-sion is not purely objective, since the superior can only draw conclusions from cer-tain informations if he or she has some pre-existing knowledge.273 Consequently,both versions make equally clear that at least conscious ignorance in the sense ofwilful blindness does not exclude criminal responsibility.274 As a result we are notdealing with any negligence (‘toute négligence’) but with a kind of negligence that,from a normative point of view, may be put on an equal footing with intent, i.e.‘negligence so serious that it is tantamount to malicious intent’.275 It must not beoverlooked, however, that intent and negligence are very different states of mind


267 See supra note 104.268 See supra, II.C, notes 123, 127 and text. Critically in general: Eckhardt, supra note 56, 19–20.269 Cf. Art. 33(4) of the Vienna Convention on the Law of Treaties.270 De Preux, supra note 101, mn. 3545; Partsch, supra note 106, 525; Fenrick, supra note 76,

119; also Blaškic, supra note 65, para. 326.271 YILC (1988) Vol. I, pp. 288–289 (Mr Tomuschat, chairman of the Drafting Committee);

also Rogers, supra note 208, at 139.272 See also Delalic et al., supra note 63, para. 392.273 Partsch, supra note 108, 525–526, however, takes the view that the objective standard of the

French text differs from the English one in that it does not permit subjective considerations in deter-mining whether the superior should have drawn the right conclusion from the information avail-able.

274 De Preux, supra note 103, mn. 3545–46.275 Ibid., mn. 3541.

846


which only in the exceptional case of ‘weak’ intent (dolus eventualis) and ‘strong’(gross, conscious) negligence approach each other.

The concept of so-called constructive knowledge must also be understood as requir-ing reliable and concrete information enabling the superior to know of the com-mission of crimes. Hessler correctly states that constructive knowledge consists ofa duty to make inferences from actually known facts and to carry out reasonableinvestigation of actually known ‘suspicious’ facts.276 On this basis he formulateshis Rule Three, according to which the superior has the duty to know of specificcrimes and policies on the basis of such facts. Constructive knowledge is definedwith regard to the mental objects used in Rules One and Two, i.e. known crimesor policies which must be prevented.277 Additionally, the superior has the duty toeliminate non-negligible risks of future crimes, which Hessler considers to be thethird mental object.278

Constructive knowledge does not, however, encompass the imputation of know-ledge on the basis of purely objective facts,279 which would imply that the superiormust have known of the atrocities given their large scale and widespread commis-sion. Such a standard is a fiction since knowledge is presumed even though it didnot exist or, at least, cannot be proven. In fact, if knowledge does not exist or can-not be proven the superior can only be punished for negligently not having knownof the crimes, i.e. because he or she should have known. Only this interpretationdoes not conflict with the principle of guilt.

As to the ‘had reason to know’ standard, it does not differ substantially from theearlier formula of the PA I (‘information which should have enabled him to con-clude’). Although the ILC argued that the ‘had reason to know standard’ permitsa more objective assessment than does the PA I standard,280 it explained the formerwith the words of the latter, explicitly referring to the commentary on Article86(1) PA I.

In this situation, a superior does not have actual knowledge of the unlawful con-duct being planned or perpetrated by his subordinates, but he has sufficient, rele-vant information of a general nature that would enable him to conclude that this isthe case. A superior who simply ignores information which clearly indicates thelikelihood of criminal conduct on the part of his subordinates is seriously negli-gent in failing to perform his duty to prevent or suppress such conduct by failing

123456789101112131415161718192021222324252627282930313233343536373839404142


276 Hessler, at 1278–1279, 1298–1299. See also the similar definition of the UN Commissionof Experts, supra note 76, quoted in Bassiouni, supra note 3, at 437; see also Fenrick, supra note 76,mn. 124.

277 Hessler, supra note 16, 1295 et seq.278 Ibid., 1299 et seq. Concerning this mental element Hessler considers the law as unsettled

(1282); see also his general critical assessment of these mental objects at 1281 et seq.279 See Parks, supra note 3, 90; similar Rogers, supra note 208, at 139.280 Report ILC 1996, supra note 111, at 38, para. 5 = YILC (1996) Vol. II, Part 2, at 26, para. 5.

847


to make a reasonable effort to obtain the necessary information that will enablehim to take appropriate action.

The similarity of the two standards is supported by the fact that the ‘had reason toknow’ formula of the ICTY Statute is based on a U.S. proposal that referred topossible knowledge ‘through reports to the accused person or through othermeans’,281 thereby relying on the Hostage case and the PA I.282 Given that the‘should have known’ formula in the Rome Statute also goes back to a US proposal,it is clear that the two formulas are not substantively different. This view is sharedby other writers.283

If these different formulas used for the mens rea requirement do not imply or arenot intended to imply a difference in substance, it seems to be most logical tointerpret the ‘should have known’ criterion in the light of the PA I—as the origi-nal source of superior responsibility—and general criminal law. The problemwith this approach is that the PA I formula is far from clear. One can even say thatthe PA’s vague and ambiguous formula is the cause of all subsequent problems ofinterpretation. Wu and Kang, in a very thoughtful study, offer four interpreta-tions of Article 86(2) PA I:284

(1) A superior has an obligation to monitor the actions of his subordinates andwill be held responsible for the knowledge that a reasonable agent in hisposition would have possessed;

(2) The superior must be guilty of ‘wilful blindness’ in order for knowledge tobe assumed, because deliberate ignorance and positive knowledge areequally culpable;

(3) Knowledge must be constructively imputed based on his position if thereis no way that he could not have known;

(4) Actual knowledge must be proved, but may be inferred from circumstan-tial evidence such as the defendant’s position.

While (1) and (2) are straightforward standards of negligence or recklessnessrespectively, in (3) and (4) knowledge is ‘constructed’ on the basis of objectiveevents or circumstantial evidence. Unfortunately, Wu and Kang do not furtherexamine these standards because they consider that ‘the specification of a single,rigorously defined, unambiguous mens rea requirement . . . would be a fruitlessexercise as it is almost impossible to discern the precise holdings of derivative liability cases with respect to mens rea in practice.’285 However, this surrender to


281 Quoted according to Crowe, supra note 5, 229–230.282 Ibid., 230.283 See Levie, supra note 3, 10: ‘reason to know’ and ‘should have known’ ‘strikingly similar’.

Landrum, supra note 5, at 300: PA I and ICTY standard ‘quite similar’. Dissenting apparentlyVetter, supra note 38, 122–123.

284 Wu and Kang, supra note 107, 284–285.285 Ibid., at 286.

848


practical considerations—without even attempting to develop a theoretical solu-tion less ambiguous than the existing one—is too summary. Although it is correctthat each case ‘is decided largely on its particulars’,286 this should not precludemore profound efforts to develop a theoretical model that would allow more fore-seeable solutions. If one takes a closer look at the standards developed by Wu andKang, one can quickly exclude standards (3) and (4) in the present context. Whilestandard (3) would, as just mentioned,287 violate the guilt principle, by relyingexclusively on objective facts and presuming knowledge on this basis, standard(4), in reality, does not deal with the mens rea requirement at all but only with theadmission of circumstantial evidence for the proof of actual knowledge.288 As faras standards (1) and (2) are concerned the ‘should have known’ formula consti-tutes rather a negligence than a recklessness standard. This view is shared by vari-ous official and private statements, according to which the superior responsibilitydoctrine creates liability for criminal negligence.289 This was already discussedwith regard to the PA I, where it was made clear that the formulation of the PA Iwas written with negligence in mind.290 Further, the UN Secretary-General’sReport about the establishment of the ICTY describes the ‘had reason to know’standard as ‘imputed responsibility or criminal negligence’.291

Still more important than these formal arguments is the material observation thatthe ‘should have known’ standard corresponds to negligence as understood in gen-eral criminal law. According to § 2.02(2)(d) of the US Model Penal Code(MPC)292 a person acts negligently ‘when he should be aware of a substantial andunjustifiable risk that the material element exists or will result from its conduct’.Negligence is distinguished from the other forms of mens rea (purpose, know-ledge, or recklessness) in that it does not involve a state of awareness. A person actsnegligently if the person creates a risk of which he or she is not, but ought to be,aware.293 The person is liable if the failure to perceive the risk ‘involves a gross devi-ation from the standard of care that a reasonable person would observe in theactor’s situation’ (§ 2.02(2)(d) MPC). Thus, there is a clear-cut distinctionbetween negligence and recklessness in terms of the actor’s awareness of the riskinvolved:294 a negligent actor fails to perceive the risk, i.e. is not aware of it; a reck-less actor ‘consciously’ disregards the risk (§ 2.02(2)(c) ), i.e. perceives but ignores

123456789101112131415161718192021222324252627282930313233343536373839404142


286 Ibid., at 287.287 Supra note 279 and accompanying text.288 See supra (a), notes 264–266 and accompanying text.289 See Hessler, supra note 16, 1284: ‘type of criminal negligence’; see also his Rule Three: ‘duty

to eliminate non-negligible risks’ (emphasis added; see supra note 276 et seq. and accompanyingtext); Schabas, supra note 16, at 417: ‘liability for negligence’; also Bantekas, supra note 41, 590.

290 See supra note 275 and accompanying text; also De Preux, supra note 103, mn. 3541.291 Supra note 192, para. 56.292 American Law Institute (1985), Vol. I, at 226.293 Ibid., at 240.294 Ibid., at 242.

849


it.295 Thus, recklessness and negligence can be equated to the German concepts ofconscious and unconscious negligence (bewußte und unbewußte Fahrlässigkeit).296

However, this difference between conscious recklessness and unconscious negli-gence is ignored when such contradictory notions as ‘wilful’, ‘wanton’, or even‘conscious’ negligence are used.297 This accounts for much of the confusion withregard to the mental element involved in superior responsibility.298 It should beclear now, however, that the ‘should have known’ standard must be understood asnegligence and that it, therefore, requires neither awareness299 nor considers suffi-cient the imputation of knowledge on the basis of purely objective facts.

The general criticism of negligence as a form of mens rea 300 has not impeded itsuniversal recognition.301 Although a more profound analysis is not possible here,it may be pointed out that the vague ‘reasonable man standard’302 can be comple-mented by certain subjective criteria, i.e., it can be ‘individualized in order toachieve a fair standard of judging individual behaviour’.303

The common denominator of all these attempts to define a standard less demand-ing than the positive knowledge requirement is that the superior must possessinformation that enables him or her to conclude that the subordinates are com-mitting crimes.304 Accordingly, the ICTY refers to information of a nature thatwould put the superior on notice of the risk of such offences and indicate the needfor additional investigation in order to ascertain whether crimes have been com-mitted.305 Fenrick sees the ‘should have known’ standard as satisfied if the superior


295 Cf. Fletcher, supra note 148, 115.296 See Fletcher, supra note 148, 115, pointing out, however, that the English law uses the term

‘reckless’ not to denote risk-consciousness, but to refer to egregious cases of negligence. See alsoPradel, supra note 174, 257 et seq. who distinguishes between insouciance and négligence, includingin the former recklessness and in the latter conscious and unconscious negligence.

297 American Law Institute, supra note 292, at 242.298 A good example is the often quoted study of Parks who opts for a ‘wanton negligence’ involv-

ing the ‘doing of an inherently dangerous act or omission with a heedless disregard of the probableconsequences ‘ (at 97) and a negligence ‘so great as to be tantamount to the possession of the neces-sary mens rea to so become such an active party to the offense’ (99).

299 As wrongly implied by Art. 87(2)(3) PA I. According to Rogers, supra note 208, 142, thisstandard covers actual and constructive knowledge. However, this is only true if constructive know-ledge is understood as construction of knowledge on the basis of facts which enable the superior toknow of the commission of crimes.

300 See American Law Institute, supra note 292, at 243–244; Fletcher, supra note 148, 117 et seq.;Pradel, supra note 174, 263 et seq.

301 Fletcher, supra note 148, at 146; Pradel, supra note 174, at 261.302 Critical in this context Parks, supra note 3, 90; Hessler, supra note 16, 1285. For a different

reasonableness standard in case of military defendants, see Green, supra note 52, at 169.303 Fletcher, supra note 144, 119; see also the subjective criteria with regard to command respon-

sibility proposed by Parks, supra note 3, 90 et seq.; see also Bassiouni, supra note 3, 423.304 See e.g. Landrum, supra note 5, 301: ‘ “had reason to know” appears to mean “had the

information from which to conclude” ’; Levie, supra note 3, 12: information about violations of thelaw of war available to a commander; also Fenrick, supra note 76, mn. 11.

305 See supra, II.B.3, note 75 and text.

850


‘fails to obtain or wantonly disregards information of a general nature within hisor her reasonable access indicating the likelihood of actual or prospective criminalconduct on the part of subordinates . . .’306 Crowe distinguishes between ‘reportsmade to the commander’ and ‘widely published press accounts of the atrocities’.307

However, this distinction is not relevant since the opportunity to learn of atroci-ties depends on the quality of the information, not on the source. A thoroughlyresearched press report can certainly make the superior aware of irregularities andoblige him to order further investigation. Similarly, the argument brought for-ward by Rogers that ‘[t]he fact that a report is addressed to a commander does notmean that he sees it or is even aware of its existence’308 does not necessarily absolvethe superior of liability since he is responsible for an effective reporting systemwithin his command.309 Further, if the superior possesses sufficient informationbut does not analyse it sufficiently and therefore reaches a wrong conclusion, he orshe is liable because of the blameworthy lack of thorough analysis. It is, however,less clear what legal consequence an erroneous evaluation of the existing informa-tion entails. Take the case of a superior who has analysed the information thor-oughly but draws an erroneous conclusion with regard to the imminentcommission of crimes by subordinates. In this case, one has to determine, first, thebasis of the superior’s error. If he or she incurred in a mistake of fact—although anunlikely assumption if there is sufficient (factual) information—he would act orrather not act (omit) without mens rea within the meaning of Article 30 of theRome Statute. Consequently, he must be exempted from criminal responsibility.In the light of the ‘should have known standard’, however, the superior would becriminally liable since because of the information available he or she should haveknown. If he or she incurred in a mistake of law, misunderstanding his or her legalobligations, the superior would be criminally responsible, since internationalcriminal law has opted for the error iuris doctrine, according to which a mistake oflaw does not affect the criminal responsibility except if it negates the mental ele-ment.310

(c) Non-Military Superior: ‘consciously disregarded information which clearly indicated . . .’This new standard essentially repeats the ‘wilfully blind’ criterion known fromcommon law and war crimes trials.311 The ‘wilful blindness’ standard presents an

123456789101112131415161718192021222324252627282930313233343536373839404142


306 Fenrick, supra note 76, mn. 11.307 Crowe, supra note 5, 226.308 Rogers, supra note 208, 139.309 See supra (d), note 253 and accompanying text.310 See Art. 32 of the Rome Statute, supra note 164; see also my critical view in supra note 162,

at 29–30.311 See e.g. R. v. Finta (1990), 98 ILR (1994) 520 (Ontario CA) at 595. See also Eckhardt, supra

note 56, albeit concerning military commanders, at 14: ‘no room . . . for a “stick your head in thesand” approach.’; also Vetter, supra note 38, 124.

851


exception to the positive knowledge requirement in that it is considered satis-fied—regarding the existence of a particular fact—‘if a person is aware of a highprobability of its existence, unless he actually believes that it does not exist’ (§ 2.07MPC).312 Knowledge is not presumed, however, when what is involved is theresult of the defendant’s conduct.313 Wilful blindness, and thereby the new stan-dard of Article 28(a)(i), stands between knowledge and recklessness. The provi-sion clearly establishes a higher threshold than the negligent ‘should have known’standard for military superiors.314 Given the travaux préparatoires and the fact thatactual knowledge is already covered by the first mens rea requirement of Article 28,it establishes a recklessness rather than a knowledge standard. This also followsfrom the similar formulation of § 2.02(2)(c) MPC, which defines recklessness as‘consciously’ disregarding a risk. The second standard developed by Wu and Kangalso views wilful blindness as constituting recklessness.315

As a result, it will be more difficult to prosecute non-military superiors than mili-tary superiors for a failure of supervision.316 According to Fenrick the indicia usedto prove actual knowledge317 can also be taken into account with regard to non-military superiors; besides, it is necessary to establish:318

• that information clearly indicating a significant risk that subordinates werecommitting or were about to commit offences existed

• that this information was available to the superior and• that the superior, while aware that such a category of information existed,

declined to refer to the category of information.

C. Conclusion

Article 28 establishes a broad liability of the military and civilian superior whichcan certainly be based on customary international law but nevertheless gives riseto serious concerns. There must be a clear distinction between the responsibilityof the superior for acts of his or her subordinates of which he or she was aware orfor acts which were unknown to him or her. In the latter case—leaving aside the


312 See also American Law Institute, supra note 292, at 248; W. R. LaFave and A. W. Scott, Jr.,Criminal Law I (1986) 307–308.

313 American Law Institute, supra note 292, at 248.314 Cf. Fenrick, supra note 76, mn. 21. See also Schabas, supra note 16, at 419, who sees the new

standard as establishing a ‘full knowledge requirement.’315 See supra note 284 and accompanying text316 See the recent criticism developed by Vetter, supra note 38, 94, 96, 103, 116, 141, who takes

the view that the civilian mens rea standard of the Rome Statute reduces the efficacy of the ICC (butsee also his counter-hypothesis in note 171). Thus, in his view, for example, the Japanese diplomatHirota, convicted by the IMTFE (supra note 48), could not be held responsible on the basis of thisnew standard (ibid., 126–127; with a further analysis of the Roechling, Akayesu, and Miloseviccases on pp. 128 et seq.).

317 See supra, II.B.3, notes 75–79 and (a), note 264 and accompanying text.318 Fenrick, supra note 76, mn. 21.

852


slight difference between the mens rea standard for the military and non-militarysuperior—the literal application of the Rome Statute would entail negligence lia-bility for intentional acts, a construction which is not only logically imposible but,more importantly, hardly compatible with the principle of guilt. De lege lata theonly way out of this logical impasse consists in the creation of a separate offence of‘failure of proper supervision’ which certainly deserves a much more lenient pun-ishment than the superior’s intentional omission to prevent his or her subordi-nates from committing international crimes. In any case, the future task of theinternational jurisprudence is not to extend the superior responsibility even moreconverting it into a form of strict liability but to develop criteria which limit thetoo broad liability established by Article 28 of the Rome Statute.

Select Bibliography

CUSTOMARY INTERNATIONAL LAW

J. J. Douglass, ‘High Command Case: A Study in Staff and CommandResponsibility’, 6 International Lawyer (1972) 686–705; K. Howard, ‘CommandResponsibility for War Crimes’, 21 Journal of Public Law (1972) 7–22; W. V. O’Brien, ‘The Law of War, Command Responsibility and Vietnam’, 60Georgetown Law Journal (1972) 605–664; C. A. Hessler, ‘CommandResponsibility for War Crimes’, 82 Yale Law Journal (1973) 1274–1304; W. H. Parks, ‘Command Responsibility for War Crimes’, 62 Military Law Review(1973) 1–104; W. G. Eckhardt, ‘Command Criminal Responsibility: A Plea for aWorkable Standard’, 97 Military Law Review (1982) 1–34; K. J. Partsch,‘Commentary on Articles 86 and 87 of Protocol Additional I’, in M. Bothe, K. J. Partsch, and W. A. Solf (eds.), New Rules for Victims of Armed Conflicts (1982)523–529; B. Schünemann, ‘Strafrechtsdogmatische und kriminalpolitischeGrundfragen der Unternehmenskriminalität’, 1 Zeitschrift für Wirtschaft, Steuer,Strafrecht (Wistra) (1982) 41–50; A. D’Amato, ‘Superior Orders vs. CommandResponsibility’, 80 AJIL (1986) 604–608; H. S. Levie, ‘Some Comments onProfessor D’Amato’s “paradox” ’, 80 AJIL (1986) 608–611; K. Rogall,‘Dogmatische und kriminalpolitische Probleme der Aufsichtspflichtverletzung inBetrieben und Unternehmen (§ 130 OwiG)’, 98 Zeitschrift für die gesamteStrafrechtswissenschaft (ZStW) (1986) 573–623; J. De Preux, ‘Commentary onArticles 86 and 87 of Protocol Additional I’, in Y. Sandoz, Ch. Swinarski, and B. Zimmermann (eds.), Commentary on the Additional Protocols of 8 June 1988 tothe Geneva Conventions of 12 August 1949 (1987), marginal notes 3524–3563;L. C. Green, ‘Superior Orders and Command Responsibility’, 27 CYIL (1989)167–202; A. M. Prévost, ‘Race and War Crimes: The 1945 War Crimes Trial ofGeneral Tomoyuki Yamashita’, 14 HRQ (1992) 302–338; L. C. Green,

123456789101112131415161718192021222324252627282930313233343536373839404142


853


‘Command Responsibility in International Humanitarian Law’, 5 TransnationalLaw & Contemporary Problems (1995) 319–371; B. D. Landrum, ‘The YamashitaWar Crimes Trial: Command Responsibility Then and Now’, 149 Military LawReview (1995), 293–301; E. J. O’Brien, ‘The Nuremberg Principles, CommandResponsibility, and the Defense of Captain Rockwood’, 149 Military Law Review(1995) 275–291; T. Wu, and Y.-S. Kang, ‘Criminal Liability for the Actions ofSubordinates: The Doctrine of Command Responsibility and its Analogues inUnited States Law’, 38 Harvard International Law Journal (1997) 272–297; H. S. Levie, ‘Command Responsibility’, 8 Journal of Legal Studies (1997/98)1–18; K. Tiedemann, ‘Die Regelung von Täterschaft und Teilnahme imeuropäischen Strafrecht’, in A. Eser (ed.), Festschrift für Haruo Nishihara zum 70.Geburtstag (1998) 496–512; M. C. Bassiouni, Crimes against Humanity inInternational Criminal Law (2nd edn., 1999) 419–446; I. Bantekas, ‘TheContemporary Law of Superior Responsibility’, 93 AJIL (1999), 573–595.

ICTY

C. N. Christopher ‘Command Responsibility in the Former Yugoslavia: TheChances for Successful Prosecution’, 29 University of Richmond Law Review(1994) 191–232; W. J. Fenrick, ‘Some International Law Problems Related toProsecutions before the International Criminal Tribunal for the FormerYugoslavia’, 6 Duke Journal of Comparative and International Law (1995)103–125.

ICC STATUTE

P. Saland, ‘International Criminal Law Principles’, in R. S. Lee (ed.), TheInternational Criminal Court: The Making of the Rome Statute. Issues, Negotiations,Results (1999) 189–216; W. J. Fenrick, ‘Article 28: Responsibility ofCommanders and Other Superiors’, in O. Triffterer (ed.), Commentary on theRome Statute of the International Criminal Court (1999); G. R. Vetter, ‘CommandResponsibility of Non-military Superiors in the International Criminal Court(ICC)’, 25 Yale Journal of International Law (2000) 89–143.


854


(zb) cassese ch21

Documents