International Criminal Tribunals as Epistemic Engines Or Why Legal Truth Is Not Sui Generis

This post is part of a symposium, Doing Justice to Truth in International Criminal Courts and Tribunals. All currently available contributions to the symposium can be found here.

A PDF of this post can be downloaded here.


International criminal tribunals (ICTs) are epistemic engines. That is, they are institutions that systematically produce knowledge or find truths. And they do so not only in the way usually recognised in doctrinal scholarly works on international criminal law, i.e. in the sense emphasized, e.g., by the ICTR of “producing a substantial body of jurisprudence on genocide, crimes against humanity, war crimes, as well as forms of individual and superior responsibility.”[2] ICTs are epistemic engines also in the sense that they find (or at least claim to find) factual truths about certain past events. More specifically, they (claim to) find truths about such past events that qualify as genocide, crimes against humanity and war crimes.

This epistemic output is very tangible, covering a wide spectrum of historic facts ranging from the specific to the more general.[3] All factual knowledge that seems indisputably valuable even regardless of its significance as a means for achieving just international criminal convictions. The facts unearthed by tribunals are of value to victims and relatives who naturally have an interest in knowing as much as possible about the circumstances surrounding their suffering. These facts are valuable also to post-conflict societies as such – if only in virtue of their ability, as one commentator has it, to “reduce the number of lies that can be circulated unchallenged in public discourse.”[4] Finally, the factual truths produced by ICTs have value well beyond the societies immediately involved. They are of value to humankind as such in so far that the occurrence of mass atrocities forces all of us to face questions about the human capacity for evil. It is no exaggeration that the Holocaust led to an intellectual and civilisational crisis, and since then understanding this phenomenon has been a primary concern of the arts and social sciences. Truths of the kind provided by ICTs have repeatedly proven indispensable to this end.[5]

This leads to a puzzle. Once, at a conference in Oslo on the legitimacy of ICTs, Professor Andreas Føllesdal stated that the goal was to “try to establish a taxonomy of challenges to the legitimacy of ICTs that can plausibly be raised against them.” Considering the manifest value of accurate historical records of mass atrocities, one would think that the apparent truth finding competence of ICTs would be a natural candidate for such a taxonomy. If, on the one hand, ICTs could be demonstrated to be generally reliable epistemic engines, i.e. to be good at producing factual truths about mass atrocities, this would seem to provide a strong argument in favour of their legitimacy. If, on the other hand, they could be demonstrated to thwart or distort truth more often than not, this would seem to provide a strong argument against them.

In actual fact, however, this particular epistemic aspect of the work of ICTs seems only to have attracted limited attention in legitimacy debates – especially compared to questions about their alleged capacity to bring an end to impunity or to deter potential perpetrators of mass atrocities etc. At a more fundamental level, this dismissive treatment seems to reflect a deep disagreement among scholars, politicians and practitioners alike about the proper level of epistemic ambition. Or more precisely, about whether it is at all proper for ICTs to have epistemic ambitions as such.

On the one hand, there is a widespread tendency to play down the role of truth finding in relation to ICTs. This view finds a prominent precursor in Hannah Arendt’s notorious remark on the Eichmann trial in Jerusalem:

The purpose of a trial is to render justice and nothing else; even the noblest of ulterior purposes – “the making of a record of the Hitler regime which would withstand the test of history” as Robert G. Storey, executive trial counsel at Nuremberg formulated the supposed higher aims of the Nuremberg Trials – can only detract from the law’s main business: to weigh the charges brought against the accused, to render judgment, and to mete out due punishment.[6]

Correspondingly, the Preamble to the Rome Statute does not mention finding the truth or making a historic record, whereas it explicitly lists a number of other goals closely related to the traditional topics in relation to the legitimacy of tribunals. In the same vein, it is not uncommon to hear, in court proceedings or in judgments, express rejections of the goal of truth as an independent objective of the tribunals. This perception on the part of at least some proponents of ICTs is at least indirectly supported by those who favor the use of so-called truth commissions over ICTs. Here, the name alone brings the message across: Unlike ICTs, these institutions provide truth.

On the other hand, however, others seem more reluctant to denounce truth as a relevant target for international criminal tribunals. Thus, quite a considerable number of judgments and other official documents from ICTs reveal an apparent need among at least some representatives to express stronger epistemic ambitions on behalf of the tribunals, for instance in Halilovic:

It is one of the purposes of the International Tribunal to establish a historical record of the event in the former Yugoslavia. It is also a purpose of the International Tribunal, and a duty owed by the prosecution to the International Tribunal, to look for the “truth” of those events.[7]

Thus, opinions are divided among proponents and representatives as to whether veracity should be a desideratum for ICTs or not – and hence whether pointing to epistemic limitations is relevant in legitimacy debates as a challenge that could plausibly be raised against them. This contribution looks into this division of opinions, and argues that it is relevant because legal truth is not sui generis.

Legal truth as a “false friend” of truth? The objection against truth restated

At a first glance, it appears downright inconsistent to claim that truth is not a desideratum of ICTs. As already remarked, it is plainly obvious that even if truth is absent from the preamble, ICTs do produce an overwhelming number of factual assertions about past atrocious events. Furthermore, ICTs instruct witnesses to “speak the truth, the whole truth and nothing but the truth,”[8] the Prosecutor “to establish the truth,”[9] and the Court “to request the submission of all evidence that it considers necessary for the determination of the truth.”[10]

This serves to remind us that it would be flatly inconsistent for ICTs to denounce completely the ambition of uncovering the truth. If ICTs really had no interest in the truth it would simply be impossible for them to pursue the twin goals of every criminal justice system: to avoid judging the innocent while simultaneously making sure that crime does not go unpunished. True findings of fact are a necessary precondition for just findings of guilt.

As we have seen, however, an influential strand nevertheless seems to maintain persistently that truth is not a desideratum for ICTs (or for domestic criminal trials for that matter). The question is why anyone would claim not merely that the epistemic competence of ICTs is limited in certain ways, but that truth simply is not a desideratum for them?

At first, we should observe a certain ambivalence also in this apparent denial of truth. On one radical reading, it is simply a blanket denial of truth as such. On another, however, the call of Arendt and others is simply for moderation of the scope of the epistemic ambition of ICTs. Here, the point is merely that we cannot reasonably expect of ICTs to provide a “historical record” understood as a full, comprehensive, wall-to-wall narrative of the conflicts they adjudicate. Like in all trials, ICTs deal only with a well-defined factual problem. To provide a “historical record” is simply too broad an undertaking.

Whereas the latter, moderate reading is in my view the more plausible, I shall focus here on the former. I do so, because the radical reading actually holds traction among both scholars and practitioners of ICL, even if only as a dim notion and not always spelled out in detail. I am therefore discussing not a straw man, but a rational reconstruction of an ideal type in the Weberian sense, and in order to expose clearly the implications of a view that is perhaps not always thought through to its logical conclusion.

Looking for arguments for this more radical position, we observe first of all that those who categorically deny that ICTs are in the truth-finding business tend to emphasise the sui generis character of assertions produced in tribunals, consistently calling not merely of “truth” but of “legal truth” or “trial truth.”[11] Furthermore, this emphasis is often coupled with a reference to the distinction between common law and civil law countries with their adversarial and inquisitorial systems respectively, and an emphasis that ICTs are primarily molded on the adversarial model. According to this line of argument then, the overall goal of adversarial legal procedures is simply not truth as it is in inquisitorial systems. It is a fair trial[12], and this difference, it is maintained, has far-reaching implications for the way we should understand legal truth in ICTs.

One way of explaining this is in terms of the distinction between concept and conceptions first invoked by William Gallie.[13] Thus, on this interpretation, truth is not a genuine desideratum for ICTs because the kind of legal truth they find differs from ordinary truth not merely in the sense of constituting a different conception of the same underlying truth-concept. Truth in law is in fact, and in spite of the homonymy, an altogether different and unique concept. When explicitly mandated as a goal for prosecutor and court in statutes or rules of evidence and procedure, truth simply means something categorically different from truth in ordinary usage; it is sui generis. In this particular context, “truth” has nothing to do with being “in accordance with fact or reality” as e.g. The Concise Oxford Dictionary would have us believe.

More specifically, being “legally true” refers to a strictly processual property attributed to particular assertions, i.e. to the property of being justifiably assertable at the end of a trial properly conducted in accordance with certain legal procedures. Traditionally, in common law cultures, this process is defined in terms of fairness; equality of arms, etc. but it can in principle be (and has historically been) defined completely differently. The crucial point is that the process is strictly internal to the trial and that it has finality. The legal truth or falsity of a proposition is determined exclusively on the basis of procedural criteria and without recourse to the outside world.

The radical character of this position can be illustrated with Heinrich von Kleist’s “Der Zweikampf.”[14] This novella recounts how in a 14th century murder trial the “legal truth” of a disputed fact is determined not by any investigation into facts or reality, but by the outcome of a duel fought between the two noblemen, Count Jacob Rotbart and Sir Friedrich von Trota. After having first inflicted a surface cut on his opponent, Friedrich loses the duel after being stabbed deeply three times. In accordance with procedure, this decides the case in Rotbart’s favor: his alibi is accepted, and he is acquitted of murder, whereas Friedrich is sentenced to death.

Ex hypothesis, this outcome constitutes a paradigmatic example of a legal truth conceived as categorically distinct from ordinary truth: a factual assertion (“Rotbart was not on the scene of the crime on the night of the murder.”) is held legally true solely in virtue of being the outcome of a particular procedure strictly internal to the legal system (a duel) and without recourse to extraneous investigations of the external world. Of course, those who hold today that legal truth is sui generis do not subscribe to this particular legal procedure. But, again ex hypothesis, this has nothing to do with the obvious epistemic limitations of the particular process (that the outcome of a duel is a poor measure of past events). It has to do, rather, with the perceived arbitrary and unfair character of that process. Not any process will do from the point of view of modern day criminal justice. The legal truth has to be determined through a due process of law as defined through centuries of refinement of the rule of law.

On this account, then, it is simply a popular misunderstanding when people point to the vast number of factual assertions about atrocious past events produced by ICTs in combination with a few procedural rules containing the word “truth,” and jump to the conclusion that ICTs are in the truth-finding business. More specifically, it is the misunderstanding, which in Pierre Bourdieu’s apt words “occur[s] when words from ordinary usage have been made to deviate from their usual meaning by learned usage and thus function for the layperson as ‘false friends.’”[15]

Why legal truth is not sui generis

While this whole line of argument and analysis of legal truth as a false friend of, and hence a different concept from, ordinary truth may be appealing, it is nevertheless incorrect. This can be illustrated by taking a closer look precisely at Bourdieu’s broader analysis of the role of such “false friends” in the legal field.

First of all, Bourdieu’s field theory simultaneously provides us with a powerful argument why, even if they are wrong, international criminal lawyers may have a strategic interest in maintaining that legal truth is strictly sui generis. As Bourdieu remarks, the drive towards creating monopolies constitutes a fundamental logic in the legal field: “The constitution of the juridical field is inseparable from the institution of a professional monopoly over the production and sale of the particular category of products’ legal services.”[16] Maintaining that legal truth is in fact a false friend of ordinary truth undoubtedly constitutes one very powerful way of thus excluding competition. If legal truth is indeed the result of an esoteric guild process that has finality and is inaccessible to the outsider, this creates a strong bulwark against competition from other professionals like journalists or historians who are also in the market of offering true statements about past atrocious events.

In this way, the widespread insistence among legal scholars and practitioners on the incommensurability of legal truth illustrates how:

legal qualifications comprise a specific power that allows control of entry into the juridical field by deciding which conflicts deserve entry, and determining the specific form in which they must be clothed to be constituted as properly legal arguments.[17]

However, this Bourdieusian analysis simultaneously serves to explain why the interpretation of legal truth as a genuine “false friend” cannot consistently be upheld.

Even the strongest monopolies necessarily have to exercise some kind of openness towards the surrounding world. Otherwise there would be no external demand for their services from outside actors. In the same way, ICTs cannot assert absolute sovereignty with regard to the legal outcome of the conflicts they handle and hence neither with the procedures through which they deal with them. Considering the momentous importance of an atrocious past to both victims, war-torn societies and the international community, there simply has to be some incentive for these parties to hand this past over to ICTs.

To establish and maintain a demand for their services, ICTs therefore have to deal in currencies that are recognised by the outside world. One such currency is truth – however philosophically challenged as a concept. The outside world simply wants to know what happened. This is a substantial part of the intuitive idea of justice held widely by non-legal actors, and ICTs cannot afford to ignore this if they are to be considered legitimate.

This of course does not exclude that once established ICTs will be granted considerable professional discretion about designing the processing of conflicts in accordance with guild training and expertise. For the whole arrangement to make sense, there has to be an added value in handing over past atrocities to ICTs. But as Dworkin reminds us, “[d]iscretion, like the hole in a doughnut, does not exist except as an area left open by a surrounding belt of restriction.”[18] For the reasons given above, the experts of international criminal law cannot be granted complete license to define their own process at will. Any autonomy they exercise in terms of process and outcome must be conditional. In other words, legal truth cannot be sui generis. It cannot be a complete “false friend” but has to be a conception of our ordinary concept of truth.

We should generally expect this interdependence to be particularly pertinent when, as in ICL, the legal actors are trying to move into new markets. But it is found even when the monopoly is long-established. On closer inspection, we find an apt illustration of this in “Der Zweikampf” too. Thus, as events further unfold, surprisingly the outcome of the duel is turned on its head: Friedrich recovers miraculously from his deep wounds, whereas Rotbart’s superficial cut becomes badly infected. His general condition rapidly deteriorates and becomes untreatable. On his deathbed, Rotbart testifies to Friedrich’s innocence and his own guilt in the murder. Subsequently, it is decided not only to reverse the previous judgment, but also to have “the statutes governing divine trial by duel amended so that everywhere where they presumed that duels would immediately reveal guilt, the words were added: ‘if it be God’s will.’”[19]

This points to an important truth about truth in law: Under normal conditions, the outcome of the legal process has to be respected as final. Otherwise, the legal system cannot perform properly. But the legal system is entrusted with conflicts only because we generally trust its judgments and presuppose the veracity of its evidentiary procedures. This is a precondition of its autonomy as witnessed by the way we respond, when it is violated too blatantly, and the outcome is too much at odds with our ordinary concept of truth.

In real life, this deference to ordinary truth can in fact be found even in the adversarial American legal system. Here recent years have seen a large number of post-conviction exonerations based on DNA-testing.[20] In the great majority of these cases, convicts were exonerated not because of breaches of the right to a fair trial but solely because DNA-evidence proved that the original procedure had reached false conclusions by ordinary extra-legal notions of truth and falsity. In other words, questions regarding legal procedure played no role. The key issues were general unreliability, i.e. failure to be “in accordance with fact or reality.”

This illustrates how the rationale implicitly adopted in the DNA exonerations actually resembles the proviso adopted in Kleist’s novella: even in the American system the deference owed to the outcome of the legal process is only conditional; i.e. it is owed only in so far as the veracity of the evidentiary procedures remains fundamentally uncompromised. This proves that the law aims i) to adopt procedures that are generally truth conducive; ii) to replace them with new and more reliable procedures when available; and iii) even to apply such new procedures retrospectively in particularly grave cases of past legal falsity. Truth in law, in other words, remains thoroughly connected to ordinary truth.[21]


The upshot of these considerations is that legal truth cannot be sui generis. Truth as ordinarily understood simply remains a central goal of the legal process—even in adversarial legal systems like the American.[22] And this brings us back to the ambivalence noted regarding the skepticism expressed by Arendt and others. It seems that the only viable version of this skepticism is not as a blanket rejection, but simply a call for moderation in our ambitions with regard to truth. It is a call for moderation with regard to the scope. ICTs should not look for the full story (whatever that is) of past atrocities, but only to those facts relevant for the case at hand. And it is a reminder that truth can only be one, possibly even subordinate, desideratum among many others, including e.g. fair trial, justice, prevention, reconciliation and peace.

But this is sufficient to establish that truth-finding does indeed constitute a desideratum for ICTs. It belongs in our taxonomy of “challenges that can plausibly be raised against them.” This in turn implies, as the next natural step in developing such a taxonomy, that we should turn to the epistemological question proper as described in the introduction, and try to map various critiques that have been launched against the actual truth conduciveness of ICTs.


[1]                 This post briefly summarizes my contribution “’One of the Challenges that can plausibly be Raised Against Them’? On the Role of Truth in Debataes about the Legitimacy of International Criminal Tribunals,” in The Legitmacy of International Criminal Tribunals, ed. Hayashi and Bailliet (Cambridge University Press, 2017), 206-227.


[3]           In the specific end of the spectrum, we find for instance this detailed description: “On 7 April 1994 […] Ngirabatware went to the Bruxelles area of Nyamyumba commune with two vehicles transporting weapons. […] Ngirabatware provided ten machetes to Bagango, who in turn gave them to Jean Simpunga for further distribution. Simpunga distributed nine of these machetes to roadblocks in the Bruxelles area, and kept one for himself.” Prosecutor v. Ngirabatware, Case No. ICTR-99-54-T, para. 1335.

At the opposite end, we find broad assertions like this well-known conclusion from the ICTR Trial Chamber’s judgment in Akayesu: “Consequently, the Chamber concludes from all the forgoing that genocide was, indeed, committed in Rwanda in 1994 against the Tutsi as a group. Furthermore, in the opinion of the Chamber, this genocide appears to have been meticulously organized.” Prosecutor v. Akayesu, Case No. ICTR-96-4-T, para. 126.

[4]           M. Ignatieff, Articles of Faith. Index on Censorship, (1996, 25, 110-122) at p. 113.

[5]           Cf. for example “Many important histories of the Holocaust […] could not have been written without the massive archive of documentary evidence assembled through Nuremberg’s act of legal discovery.” L. Douglas, The Memory of Judgment : Making Law and History in the Trials of the Holocaust (New Haven, CT.: Yale University Press, 2001), 2.

[6]           H. Arendt, Eichmann in Jerusalem: A Report on the Banality of Evil (Penguin Books, 2006), 253.

[7]           Motion for Judicial Notice, March 1, 2005, para. 9.

[8]           Rule 66(1), Rules of Procedure and Evidence of the International Criminal Court.

[9]           Art. 54(1), Statute of the International Criminal Court.

[10]          Art. 69(3), ibid.

[11]          Or they sometimes even prefer the less philosophically demanding cognates referred to earlier (facts, findings, narrative, etc.).

[12]          Cf. for example “The purpose and the result of such a trial [in Anglo-Saxon systems] is not the real truth, but the trial-truth. Continental proceedings differ, in that they aim to discover the real truth.” Röling in A. R., Cassese, The Tokyo Trial and Beyond (Oxford: Blackwell Publishers, 1993), 50.

[13]          Gallie 1956.

[14]          H. Kleist, Der Zweikampf in Der Zweikampf / Die heilige Cäcilie / Sämtliche Anekdoten / Über das Marionettentheater / und andere Prosa (Stuttgart: Philipp Reclam, 1998 [1811]), 3-41.

[15]          P. Bourdieu, “The Force of Law: Toward a Sociology of the Juridical Field,” The Hastings Law Journal 38 (July 1987): 829.

[16]         Bourdieu, supra note 17, 834-35.

[17]         Ibid., p. 835, emphasis added.

[18]          R. Dworkin, The Model of Rules I, in Taking rights seriously (Cambridge: Harvard University Press, 1977), 14-45 at p. 31.

[19]         Kleist, supra note 36, p. 41, my translation.

[20]          The so-called Innocence Project recounts 330 such cases:

[21]          This of course does not imply the presumption that outside the law we have access to a procedure-transcendent truth. It merely implies that there is no categorical difference between the concepts of truth involved and the procedures associated with the legal field and outside of that field. The concepts and methods applied inside and outside the field are commensurable.

[22]          This sobering conclusion corresponds well with at least some comparative studies of the common law/civil law-divide. Cf. e.g. “Above all—and contrary to what is sometimes heard—the two groups of systems are united in what the rules of evidence have as their essential aim. In every system this is the manifestation of truth.” Spencer 2002, p. 636f)

Print Friendly

About Jakob v. H. Holtermann

Jakob v. H. Holtermann is an Associate Professor in Legal Philosophy at iCourts, Centre of Excellence for International Courts, University of Copenhagen. He has a Ph.D. in Philosophy from Roskilde University. His research focuses in part on international criminal justice, particularly on the legitimacy of international criminal tribunals; and in part on the philosophy of legal science, particularly on legal realism and foundational issues in relation to interdisciplinary approaches to international law. He has published extensively on these and related issues in Ratio Juris, Leiden Journal of International Law, Santa Clara Law Review, Human Rights Review, Res Publica, Criminal Law and Philosophy, among other journals. He is currently editing a new translation of the main work of Scandinavian Legal Realist Alf Ross forthcoming on Oxford University Press.

Speak Your Mind

Your email address will not be published. Required fields are marked *

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>