A coherence framework for fact-finding before the International Court of Justice

Abstract

It is the task of the International Court of Justice to establish the operative facts from which to draw normative conclusions in all contentious cases that come before it. This task, however, is complicated where those facts necessitate engagement with specialized epistemic fields other than law such as science. Drawing on legal theory and epistemology I propose a coherence framework for the establishment of the facts in such cases. In essence, this means that the Court need not show that its factual determinations have been established beyond all doubt, nor even that they satisfy a certain standard of probability. Rather, what the Court must show is that a factual determination it has made has been arrived at through a rational process and that it is coherent. The coherence framework also has both descriptive and normative value as it both maps neatly on to the current (best) practice of the Court and provides a justification for why it should operate this way in the future.

1. Introduction

It is the task of the International Court of Justice (the ICJ, or the Court) to establish the operative facts from which to draw normative conclusions in all contentious cases that come before it. ¹ This task, however, is complicated where those facts necessitate engagement with specialized epistemic fields other than law, such as science. Significant academic attention has focused on the particular challenges that science poses to international adjudication in recent years. ² This article presents a fresh take on the debate but also has a broader focus. While the difficulties posed by cases that require the ICJ to engage with science are clear, science is just one example of a different epistemic field which presents challenges for the Court. Others are equally challenging. For instance, the Court’s task of establishing the facts is no less demanding in cases that involve contested or complex economic ³ or historical facts ⁴ for example. ⁵ Such fields, by definition, require:

specialized training in order for a person to attain sufficient epistemic competence to understand its aims and methods, and to be able critically to deploy those methods, in service of these aims, to produce the judgments that issue from its distinctive point of view. ⁶

It is for this reason that this article considers ‘factually complex’ cases more broadly, these being those cases which necessarily involve engagement with specialized epistemic fields other than law. ⁷ In such cases the Court’s judges must make sense of methods and subject-matter in which they do not necessarily have the requisite training or expertise. ⁸ The Court’s experience in the Pulp Mills and Whaling in the Antarctic cases are good examples in this regard. ⁹ This article takes the challenges the Court has faced in such cases as its starting point and attempts to move the debate forward by proposing a fresh framework for such cases.

Drawing on legal theory and epistemology, I propose a coherence framework for the establishment of the facts in cases that come before the Court. In short, when faced with the task of establishing the operative facts in cases that come before the Court, I argue that employing such a coherence framework is the best way for the Court to show that its determinations are justified. In essence, this means that the Court need not show that its factual determinations have been established beyond all doubt, nor even that they satisfy a certain standard of probability. Rather, what the Court must show is that a factual determination it has made has been arrived at through a rational process, and that it is coherent (the contested notions of rationality and coherence are discussed further at Section 5.4). In this sense, what the Court strives to achieve is the best (most coherent) factual determination.

One of the main contributions of this coherence framework is that, unlike proposals which only address how the Court deals with specific types of facts such as scientific facts, it can be used to explain a whole range of different factually complex cases. ¹⁰ The coherence framework also has both descriptive and normative value as it both maps neatly on to the current (best) practice of the Court and provides a normative justification for why it should operate this way in the future.

In defending this coherence framework, I address a number of potential objections including (i) the critical challenge to coherence as a concept; (ii) that the coherence framework does not describe or fit with the practice of the Court; (iii) even if it does, the coherence framework does not provide normative justification for the Court’s practice; and (iv) even if it does, it is not the only (or even the best) way to normatively ground the Court’s practice. I show that each of these objections can be met and that a coherence framework for the establishment of the facts before the ICJ is both normatively and practically appealing. While the present article focuses primarily on the ICJ and the challenges it has faced in this regard, these are challenges that arise before all courts, national and international alike. In this regard, while the examples explored in this article are drawn from the practice of the ICJ, I do not see any reason why my argument about how a court should address these issues should not have relevance beyond this particular court and this particular context. For instance, the framework could be particularly useful in the context of investor-state dispute settlement or human rights litigation whose fact-finding processes have not received anywhere near the kind of attention that evidence scholars have paid to those of domestic courts.

Overall, this article makes three distinct contributions to legal scholarship. First, it takes seriously the theoretical aspects of fact-finding that have received less attention compared to other procedural issues in international adjudication. Second, whereas scholarship on evidence has been largely dedicated to national courts to date, this article presents a theoretical framework for fact-finding specifically in international law. And third, it is one of the first attempts to apply a coherentist theory of legal justification to international law.

2. Playing fields: Epistemic competence and differing methods

A prime example of the ICJ having to engage with epistemic fields other than law is the Pulp Mills case. ¹¹ In this case the Court was faced with a complex and contested claim about science relating to Argentina’s arguments that, inter alia, the construction of pulp mills by Uruguay violated a 1975 treaty between the two states and was harmful to the River Uruguay and the surrounding environment. Judges Simma and Al-Khasawneh in their Joint Dissenting Opinion spelled out the challenge facing the Court:

The Court on its own is not in a position adequately to assess and weigh complex scientific evidence of the type presented by the Parties … a court of justice cannot assess, without the assistance of experts … the implications of various substances for the health of various organisms which exist in the River Uruguay. This is surely uncontroversial: the task of a court of justice is not to give a scientific assessment of what has happened, but to evaluate the claims of parties before it and whether such claims are sufficiently well-founded so as to constitute evidence of a breach of a legal obligation. ¹²

This passage frames the issue succinctly. Science and law are distinct epistemic fields, ¹³ each requiring significant training in order to acquire epistemic competence. ¹⁴ While these two fields share a number of features, such as the use of logic as a system for justifying claims, ¹⁵ they have markedly different methods and standards for evaluating evidence. ¹⁶ There are at least three significant divergences, namely the goals that each seeks to achieve, ¹⁷ the degree of certainty required to consider a claim sufficiently well-established, ¹⁸ and the defeasibility of factual claims in these particular contexts. On this latter issue I mean to say that while scientific knowledge is defeasible and always open to revision, once determined as part of the adjudicative process, findings of fact in law become formal legal facts which are not generally open to revision. ¹⁹

These divergences raise the question which is at the heart of this article: given that a legal decision-maker must show that their factual determinations are justified by reasons, ²⁰ how can they properly do so when such determinations involve epistemic fields other than law?

Judges Simma and Al-Khasawneh’s position that it is the task of the judge to evaluate the claims of the parties and to make an assessment as to whether these claims are ‘sufficiently well-founded so as to constitute evidence of a breach of a legal obligation’ may hold some instinctive appeal. However, when one considers this statement a little more deeply it becomes more challenging. What exactly does it mean that a claim is sufficiently well-founded so as to constitute evidence of a breach of a legal obligation? And when these claims relate to factually complex issues, how can the average ICJ judge assess when such a thing is sufficiently well-founded?

In the following sections I defend two main claims: firstly, that in such situations ICJ judges need not become scientists, mathematicians or economists on the bench. Rather, what the ICJ as a legal decision-making body must do is show that its factual determinations can be rationally justified. And second, the coherence framework set out in the following section is the best path to such rational justification. Each of these claims will be considered in turn.

3. The ICJ’s duty to show that its factual determinations are justified by reasons

My first claim, that the ICJ must show that its factual determinations are justified by reasons, follows from the nature of legal reasoning. I take legal reasoning to be a ‘special case’ (Sonderfall) of practical reasoning. ²¹ This is due to the fact that (i) legal reasoning is concerned with practical questions, i.e., what should or should not be done; (ii) these questions involve a claim to correctness; ²² and (iii) it is a ‘special case’ because it is subject to the constraints of the legal context. ²³ As such, my enquiry presupposes certain criteria of rationality carried over from practical reason. I take the establishment of the facts to be an integral part of the adjudicative process and of legal reasoning, and this is the basis for my normative claim that the Court must show that its factual determinations are justified by reasons. In addition, the law itself requires that factual determinations are justified by reasons. ²⁴ The duty to give reasons is one which applies to most legal decisions with very few exceptions, such as jury decisions in certain domestic proceedings.

4. A coherence framework as a route to rational justification

As mentioned above, I argue that employing a coherence framework is the best way for the Court to show that its determinations are justified. In essence, this means that what the Court must show is that a factual determination it has made has been arrived at through a rational process and that it is coherent. In this sense, what the Court strives to achieve is the best (most coherent) factual determination. To be clear, what I am advancing here is a coherentist theory of justification, not a coherentist theory of truth. As such, it is entirely possible for one to advance a coherentist theory of justification while still advocating the need or the desirability to make factual determinations which are true in the sense that they correspond with reality. And indeed, the notions of epistemic responsibility and the contextualized nature of this framework ensure that truth has a role to pay, as will be shown in Sections 4.2 and 4.3. It is time to set out the framework, which has been adapted for the international legal context, in some detail. In accordance with this framework, a factual determination made by the Court is justified if it is (i) coherent; (ii) the result of an epistemically responsible process; and (iii) in the context of adjudication before the ICJ.

4.1 Coherence

The first element of the framework for justification of factual determinations made by the Court is that the finding of fact must be coherent. Before proceeding any further, it is necessary to say a word about coherence as a central concept. Coherence theories can be found in a range of epistemic fields, including morality, ²⁵ epistemic justification, ²⁶ truth, ²⁷ and even the interpretation of art. ²⁸ What unites coherence theories is the aim of turning complex issues into simpler ones favouring the most coherent alternative (as opposed to, say, the most probable). ²⁹ As such, the idea is not to attain absolute certainty or objective truth, but rather to arrive at the ‘best explanation’ through a rational process. ³⁰ A coherence approach to reasoning involves restructuring the ‘diverse and conflicting considerations that provide equivocal support for different decision alternatives … until they fit into a “coherent” representation’. ³¹

In the legal context coherence has been advocated as a response to dissatisfaction with the traditional positivist approach to law which locates the justification of legal claims in the sources of law itself. ³² Much has been written on, for instance, moral facts in the traditional positivistic concept of law, including by proponents of coherence theories of law as a means of providing ‘a non-sceptical account of legal reasoning and rationality broader than the one defended by classical versions of legal positivism’. ³³ Prominent coherence theories of law have been advocated by scholars such as Dworkin and MacCormick, with the latter, for example, relying on the notion of (normative) coherence to show how the ethical, moral, and value choices that arise in the course of the selection of premises in law, despite not being rational in the strict sense, can still be rationally justified. ³⁴

Those coherence theories of law that have been proposed have focused almost exclusively on normative coherence, coherence in terms of legal norms, and have not considered coherence related to the establishment of facts in the legal context to any meaningful extent. ³⁵ While certain scholars have proposed ‘holistic’ approaches to the evaluation of evidence and the matter of proof, ³⁶ these cannot be said to amount to a ‘full-blown coherence theory of the justification of evidentiary judgments in law’. ³⁷ The first to attempt to draw together scholarship on coherence relating to legal norms, interpretation, epistemology and beyond is Amalia Amaya. ³⁸ Her work is the starting point for my own inquiry which seeks to consider whether a coherence framework can assist the International Court of Justice specifically in addressing some of the difficulties it has faced when dealing with factually complex cases.

All of this is relatively esoteric without an example, so let us consider what such a coherence approach to the establishment of the facts would look like before the ICJ. In practical terms, this would require the ICJ judge, as legal fact-finder, to first of all, (i) specify the parties’ competing factual claims, before (ii) evaluating the coherence of the competing claims against relevant coherence principles, before (iii) selecting as justified the most coherent of the claims. ³⁹

The first of these steps is relatively straightforward in the context of the ICJ. While it may be otherwise before domestic courts with different legal and procedural traditions, the Court’s fact-finding process is one which is primarily driven by the parties. Settled practice shows that the Court has traditionally been happy for the parties as sovereign states to have the initiative in terms of putting whatever facts they believe to be relevant before the Court. ⁴⁰ In accordance with this practice, the Court then typically assesses the probative weight of the evidence put before it by the parties, ⁴¹ evaluating the strength of their claims. Given the relative positions of the parties in terms of resources and proximity to available evidence, this is a sensible approach for the Court in the majority of cases. ⁴²

Of course, in cases of non or partial-appearance where parties refuse to participate in or withdraw from proceedings, the Court struggles to establish the operative facts from which to draw normative conclusions and may be forced to engage in its own fact-finding in order to satisfy itself in accordance with its Statute that its factual determinations are ‘well-founded’ in fact. Nevertheless, the coherence framework proposed can account for such cases in which the facts put before the Court by the parties are relatively scarce through the notion of epistemic responsibility. What is responsible for the Court varies from case to case, and in cases of non-appearance, for example, it may be that in order for the Court to fulfil its epistemic duties it is required to undertake its own fact-finding. These are all issues to which we will return below in Section 4.2.

That said, in those cases in which I am interested for the purposes of the present article, namely cases involving the complex facts in which the Court must necessarily engage in epistemic fields other than law, the Court will have the various factual claims of the parties before it. It is then the task of the Court to establish what it believes to be the most convincing account of those facts. This leads us to the next stage of the coherence framework, namely the Court’s evaluation of the competing factual claims against relevant coherence principles.

Given the Court’s traditional preference to allow the parties as sovereign states to place before it whatever evidence they see fit, the assessment of this evidence has always been a highly significant part of the establishment of the facts before the Court. ⁴³ In its jurisprudence over the years the Court has indicated on a number of occasions the types of evidence that it considers to be of ‘prima facie superior credibility’ ⁴⁴ but this approach has been criticized. ⁴⁵

Returning to the coherence framework, central to the role of the ICJ judge in assessing the factual claims made by parties in cases before it is the concept of inference to the best explanation (IBE); ⁴⁶ ‘the pattern of reasoning whereby explanatory hypotheses are generated and evaluated’. ⁴⁷ This is the move from the factual claims made by the parties to the selection by the Court of the hypothesis which explains the facts better than any other explanation. ⁴⁸ It is this move, the selection of the best (most coherent) explanation, which provides the normative justification for the Court’s factual determination. ⁴⁹

In considering which of the factual claims best explains the contested facts of the case it is for the Court in accordance with the coherence framework to have resort to those coherence principles that are relevant in relation to explanatory coherence (the pertinent form for the purposes of reasoning about legal facts). ⁵⁰ Prominent epistemologist Thagard has identified seven coherence principles that are central to the evaluation of the explanatory coherence of claims. ⁵¹ These are symmetry, ⁵² explanation, ⁵³ analogy, ⁵⁴ data priority, ⁵⁵ contradiction, ⁵⁶ competition, ⁵⁷ and acceptance. ⁵⁸ Rather than explain each here in list form, which would be dry and abstract, I will use a real example to illustrate their meaning, namely the aforementioned dispute between Argentina and Uruguay regarding alleged breaches of the 1975 Statute of the River Uruguay related to the construction of pulp mills used for turning wood chips into fiber board used to make, among other things, paper. ⁵⁹

In contentious cases that come before the Court which require it to deal with contested or complex facts such as those requiring engagement with epistemic fields other than law, such coherence principles guide the legal fact-finder when evaluating the coherence of factual claims. Once the ICJ judge has considered, for instance, whether the factual claim in question is coherent as it explains what is claimed, and that there are no competing claims which offer a better explanation, the judge can select the most coherent of the claims, and crucially that finding of fact can be justified as the result of a rational process.

Take our example, and one particular issue contested by the parties. Argentina claimed that Uruguay had breached Article 36 of the 1975 Statute which obligated the parties to co-ordinate through the Administrative Commission of the River Uruguay (CARU) to avoid changing the ecological balance of the river. ⁶⁰ Argentina argued that discharges from the Orion pulp mill into the river had changed its ecological balance, citing as an example in this regard the algal bloom (or eutrophication, caused by the enrichment of water with minerals and nutrients) of 4 February 2009. This algal bloom was, in Argentina’s opinion, a clear example of modified ecological balance of the river as a result of the operation of the Orion mill and the discharge of toxins into the river. ⁶¹ Uruguay, for its part, denied that the algal bloom had been caused by the Orion mill, and instead suggested that it had originated upstream, most likely caused by an annual carnival held in Gualeguaychú, and the resulting increase in sewage caused by visitors. Ultimately, Uruguay argued that the data (including that provided by Argentina) showed the Orion mill had not added to the concentration of phosphorus in the river since its operations began. ⁶²

The Court’s treatment of the issue is an example of the kind of factual determination for which this the coherence framework would be helpful in guiding the Court to elaborate just exactly why it was not satisfied that Argentina had proved its claim. Both parties led expert evidence during the oral proceedings before the Court. In order to determine whether the algal bloom had been caused by effluence from the Orion mill as Argentina alleged the Court had to address questions regarding the level of phosphorous in the river and the link between this and the operation of the pulp mill. In order to do so, the Court necessarily had to engage with epistemic fields other than law, in this case marine biology.

The Court, in an example of the type of cursory treatment of complicated scientific issues that was the subject of much criticism of the Pulp Mills judgment, (including by Judges Al-Khasawneh and Simma in their joint dissenting opinion ⁶³ ) addressed the contested issue of the algal bloom in one paragraph:

250. The Parties are in agreement on several points regarding the algal bloom of 4 February 2009, including the fact that the concentrations of nutrients in the River Uruguay have been at high levels both before and after the bloom episode, and the fact that the bloom disappeared shortly after it had begun. The Parties also appear to agree on the interdependence between algae growth, higher temperatures, low and reverse flows, and presence of high levels of nutrients such as nitrogen and phosphorus in the river. It has not, however, been established to the satisfaction of the Court that the algal bloom episode of 4 February 2009 was caused by the nutrient discharges from the Orion (Botnia) mill. ⁶⁴

How could application of the coherence framework have improved matters? Applying this framework, in accordance with the practice of the Court, it fell to the parties first of all to make their respective factual claims and to produce evidence in an attempt to convince the Court that each has met the standard of proof. It was then for the Court to consider the factual claims of the parties and through inference to the best explanation, select as justified the factual claim that is most coherent. In doing so, the coherence principles mentioned above provide guidance for the Court.

Not being composed of 15 or 17 career scientists, it was not for the Court to determine whether or not a link existed between the algal bloom and the operation of the Orion mill beyond all doubt, nor even to a certain standard of probability. As stated above, there are significant differences between the scientific and legal methods of proof, and the Court operates under a number of time and resource constraints which preclude open-ended scientific inquiry. Rather, in accordance with the coherence framework, it only had to select as justified the most coherent claim using recognized coherence principles.

For instance, in accordance with the coherence principle of explanation, Argentina’s claim that the algal bloom had been caused by effluence from the Orion mill was supported by the evidence it provided of the operation of the mill combined with ‘… effluent products in the blue-green algal bloom and … various satellite images showing the concentration of chlorophyll in the water’ it claimed ‘are produced during the warm season by the explosive growth of algae, particularly cyanobacteria, responding to nutrient enrichment, mainly phosphate, among other compounds present in detergents and fertilizers’. ⁶⁵ In such a situation Argentina’s claim would cohere with what it explains, namely that the effluence from the Orion mill caused the algal bloom. The principle of explanation is the one that the Court will most often have to apply in practice since, according to Thagard, it is ‘by far the most important for assessing explanatory coherence, since it establishes most of the coherence relations’. ⁶⁶

It is worth noting two implications of this principle. The greater the number of factual claims that it takes to explain a certain hypothesis then the lower the degree of coherence (in other words the degree of coherence is inversely proportional to the number of propositions that it takes to explain a hypothesis). ⁶⁷ Further, the more a hypothesis explains the more coherent it is, and that hypotheses made up of compatible hypotheses are preferable to those that require unconnected or ad hoc theories. ⁶⁸ In this sense, coherence provides a theoretical basis for the adage that the simplest explanation is often the best explanation. ⁶⁹

Further, according to the principle of competition, while Argentina’s claim that the algal bloom had been caused by effluence from the Orion mill, and Uruguay’s factual claim the algal bloom was most likely caused by an increase in people (and sewage) attending a carnival upstream, both independently explain why Uruguay had (not) breached its obligations under the relevant treaty, these explanations are not connected in an explanatory manner, and as such compete with each other, and cannot be considered to be coherent by the Court. Rather, the Court had to choose as justified the claim which best explains the disputed fact, in accordance with recognized coherence principles.

In contrast, in accordance with the coherence principle of contradiction, the emergence of evidence that showed that the level of phosphorous in the river had not been significantly altered by the operation of the Orion mill does not cohere with Argentina’s claim that the algal bloom had been caused by effluence from it. It is also worth saying a word about the principle of data priority, in accordance with which a claim which describes the result of an observation has a degree of acceptability on its own. Coming back to our example, Argentina’s claim that the algal bloom had been caused by effluence from the Orion mill coheres with our observation that increased levels of phosphorous in water can lead to eutrophication. Of course, this is not to say that such claims are beyond challenge, but they do carry a certain weight, provided that they can be explanatorily connected to other claims, such as those that the result of industrial practices have in the past had harmful effects on the environment when not properly regulated or mitigated.

Finally, the principle of acceptance provides that the acceptability of a claim that forms part of a system depends on its coherence with other claims in that system. So, for example, the acceptability of Argentina’s claim that the algal bloom had been caused by effluence from the Orion mill depends on its coherence with other claims relating to what science tells us about the harmfulness of similar mills for the marine environment, acceptable levels of phosphate in water, the causal link between the operation of the pulp mill and algal bloom, and so on. The acceptability of any claim which explains only some evidence is reduced, such as the claim that pulp mills do not lead to dangerous increases in phosphorous in water (a claim which does not explain the effect of the such chemicals in analogous situations or shown in scientific studies).

All of this sounds rather neat in the abstract. But it is not difficult to anticipate a number of obvious objections. It is all well and good arguing that the ICJ judge must evaluate the coherence of factual claims using the coherence principles set out above, but how is the average ICJ judge meant to understand what the ‘best explanation’ is when dealing with epistemic fields other than law? Are we not back where we started, asking the Court’s judges to assess evidence in relation to which they lack epistemic competence, except just with the addition of some abstract theoretical framework?

4.2 Epistemic responsibility

While the coherence framework may have both descriptive and normative appeal, in situations involving complex factual issues the problem of evaluating the coherence of parties’ claims remains. The question posed at the beginning of this article was how ICJ judges could determine whether or not parties’ factual claims were sufficiently well-founded. But even if we say that ICJ judges need not become scientists on the bench, but rather need only show that their factual determinations are rationally justified, and that rational justification can be achieved through the selection of the most coherent factual claim put forward, is the ICJ really any better placed to assess coherence?

The coherence principles set out in the previous section can scaffold the kind of thinking that judges undoubtedly already do when considering competing factual claims. But more than that, these principles provide a vocabulary for engaging with and expressing the ‘intimate conviction of the judge’ in a way which fleshes out and demystifies the fact-finding process, allowing the Court to avoid the unsatisfactory kind of abbreviated reasoning such as its conclusion on the algal bloom issue in the Pulp Mills case.

It is here that the notion of epistemic responsibility becomes central. The coherence framework that I propose here demands an epistemically responsible process. ⁷⁰ As Amaya has stated, since the coherence framework requires the specification, consideration and eventual selection of a factual hypothesis, ‘how well legal fact finders behave’, i.e., how ‘responsible’ they are, throughout this process is relevant to the extent to which they are justified in selecting a particular hypothesis. ⁷¹ Applying the notion of epistemic responsibility to the legal context:

… how well legal fact finders behave when generating and evaluating alternatives is relevant to determining how much reason they have for accepting as justified a theory of the case on the grounds that it is the most coherent one available. ⁷²

Accordingly, part of assessing the coherence of the relevant alternatives in the ‘proper way’ is adhering to a number of epistemic duties. ⁷³

Essentially, then, epistemic responsibility equates to fulfilling a number of epistemic duties. ⁷⁴ It is widely accepted that people have epistemic duties. ⁷⁵ I will define these duties as those duties that individuals have to fulfil in order to be considered rational believers. ⁷⁶ To be clear, these are not legal duties. Nor are they even (or necessarily) moral duties. Rather, I argue that ICJ judges have a role-based duty to be rational believers, i.e., to believe evidence that has been arrived at through a rational process. ⁷⁷ In other words, the source of the epistemological duty to be a rational believer is the judicial function. In defining this epistemic duty in this way, I follow Feldman in arguing that ‘certain kinds, simply in virtue of being that kind, have certain duties’. ⁷⁸ Thus, ICJ judges, simply by virtue of the judicial function, have certain epistemic duties. That is not to say that ICJ judges are not subject to other duties, such as moral duties (they certainly are, in fact) but the epistemic duties in which I am interested for present purposes are ones which have their basis in the occupation the judge has chosen for themselves.

What of the substance of these duties? First of all, and least controversially, ICJ judges have a duty to believe all (and only) claims that are supported by evidence. ⁷⁹ A logical corollary of this duty is that ICJ judges are also obligated not to believe any claims which are not supported by evidence. Secondly, ICJ judges are subject to a further role-based duty, namely the obligation to seek further evidence when it is not clear which competing claim is better supported by evidence, (in order to fulfil their epistemic duty of only believing claims which are supported by evidence). ⁸⁰ Of course, this duty is context-dependent, ⁸¹ as discussed in Section 4.3, in the sense that the judicial fact-finding process is constrained in a number of ways by resource and procedural limitations. Nevertheless, in order for the Court’s fact-finding process to be considered responsible, it is my submission that its judges are subject to this second (and more burdensome) duty.

Thirdly, ICJ judges are subject to a group of further duties which are a sub-set of the first two, and which more often are discussed under the heading of ‘epistemic virtues’. Extended engagement with the body of virtue epistemology scholarship which has been built up in preceding decades is orthogonal to our purposes, but a word does need to be said as to why I group these ‘virtues’ together as a sub-set of the epistemological duty to believe all and only claims that are supported by evidence and to seek further evidence where it is unclear which competing claim is best supported by evidence. If we accept that ICJ judges are subject to these duties, then it may well be the case that they also are subject to a sub-set of ‘other epistemological duties that will help them achieve this end’. ⁸² This sub-set of duties I frame as follows, that ICJ judges have a duty to behave in ways that will maximize the number of justified beliefs and minimize their unjustified beliefs. I believe that those attributes usually described as ‘epistemic virtues’ neatly cover the sub-set of duties that, if displayed, would maximize the ICJ judge’s justified beliefs in this way. ⁸³

It is necessary to elaborate. A virtue in epistemology can be defined as ‘characteristics that promote intellectual flourishing, or which make for an excellent cognizer’. ⁸⁴ Those epistemologists who take a responsibilist approach to virtues conceive these as ‘trait-virtues’ in that they are cultivated character traits as opposed to innate or faculty virtues. ⁸⁵ These virtues include rigour, perspicacity, open-mindedness, impartiality and intellectual autonomy. ⁸⁶ In cases that come before the ICJ, where the parties maintain the initiative in putting the facts before the Court, such epistemic virtues would include impartiality in assessing the different factual claims made by the parties, a willingness to change their existing views when presented with new evidence and courage to think freely and not simply go along with the views of their fellow judges. ⁸⁷ It is through demonstrating such epistemic virtues that the legal fact-finder has the greatest chance of fulfilling their duties of believing all (and only) those claims that are supported by evidence, as well as gathering additional evidence when it is not clear which claims are best supported by evidence.

In this respect the notion of epistemic responsibility also provides an important link to the value of truth, a value that the ICJ as part of the adjudicative process, strives to reflect. In other words, those epistemic duties that make up epistemic responsibility are truth-conducive, as they increase the chances of the ICJ judge arriving at a true belief. ⁸⁸ In that sense, epistemic responsibility acts as a check against the creation of factual determinations that are coherent in accordance with the relevant coherence principles, but which are a ‘mere abstraction’ which ‘would address facts that are either moot or disconnected from a legal dispute …’. ⁸⁹ Any factual determinations which do not bear any relation to reality are of limited value (if any value at all) to the Court and may in fact be counterproductive to the achievements of the goals of the Court such as the settlement of the dispute before it, or the attainment of contextual justice. An example of this from the Court’s practice is the Cameroon v. Nigeria case concerning the delimitation of the land and maritime boundary between the two states. ⁹⁰ Due to a number of technical errors the Court in its judgment indicated a boundary line the co-ordinates of which were so unclear as to prove impossible to implement by the parties. Consequently, the parties were forced to work with the United Nations through a joint commission to try to understand and implement the Court’s proposed boundary. ⁹¹

Of course, saying that epistemic responsibility offers a path to justified belief is not the same as saying that it will always guarantee that the Court makes factual determinations which are true. As Amaya has stated:

it cannot be shown that endorsing a coherence theory for the justification of factual and normative propositions in law will never misguide us, for it may certainly lead us to accept beliefs about the facts and the law as justified which are, nevertheless, false or incorrect, [nevertheless] there are still good reasons for supporting the desirability of such methods from the perspective of advancing the goal of truth in law.

In other words, a factual determination which has been arrived at as the result of an epistemically responsible process is one which, at the very least, can be rationally justified, and at the most has a very good chance of being true due to the judge’s epistemically responsible behaviour.

Of course, talk of epistemic responsibility, comprising various duties, raises the question of the lengths that the ICJ judge must go to in order to satisfy themselves that they have acted in an epistemically responsible manner. For instance, to what lengths is the ICJ judge expected to go to in conducting their own fact-finding in those situations in which they are unable to assess a factual claim against coherence principles? Is it enough that the judge asks a clarificatory question from the bench during oral proceedings, or is the onus on the judge to go much further, perhaps pushing for a visit to the site, or undertaking private study in order to gain a better understanding of the complex factual issues at the heart of the case? The answer to these questions, namely what counts as epistemically responsible behaviour, is shaped by the context in which the legal fact-finder is operating, namely adjudication in the context of law as a highly institutionalized specialized epistemic field. ⁹²

4.3 Coherence in context

The coherence of a factual claim made by a party cannot be assessed in a vacuum, hermetically sealed from the context in which it is made. ⁹³ Context is crucially important, in that it is what provides the outer limits of the coherence framework, i.e., what can realistically be expected of any legal fact-finder, including ICJ judges. As Amaya has stated, a ‘contextualized version of coherentism would take it that whether a belief is justified is a matter of coherence but that how severe this coherentist standard should be is a contextually variable matter’. ⁹⁴

The ICJ and its judges do not have unlimited time and resources to conduct fact-finding in fulfilment of their epistemic duties and in demonstration of their epistemic virtues. What’s more, and perhaps more importantly, it is not the sole goal of the adjudicative process to attain knowledge. Rather, adjudication pursues a number of goals including the settlement of the dispute between the parties, the achievement of contextual justice, procedural fairness and perhaps even the progressive development of international law. The establishment of the facts of the case, in other words, is not an end in itself but rather a means to establishing the operative facts of the case from which to draw normative conclusions.

Accordingly, what can be said to be coherent, epistemically responsible, and ultimately rationally justified in terms of facts before the Court depends on, and is shaped by, the context of adjudication before the ICJ. The contextual nature of coherence is significant in a number of ways, perhaps most prominently due to the fact that this ensures an element of realism, or of practicality, in terms of what we can and do expect from the Court. No one expects the Court’s judges to become scientists on the bench nor the Court to spend millions upon millions of the United Nations’ budget to make site visits or to appoint expensive teams of experts to assist the Court in every case that comes before it. But it is nevertheless important that we are able to say why we do not expect this of the Court. If we wish the Court to establish the operative facts then why should it not move heaven and earth to do so?

The answer lies in the highly institutionalized nature of adjudication before the Court. For instance, the standard of proof applied by the Court from one case to another shapes not only when the Court can be said to have arrived at facts that are sufficiently well-founded but also influence the Court’s assessment of coherence and justification.

Similarly, in cases where the Court feels that it is not in a position to assess the factual claims made by the parties as, for instance, each party’s claims appear equally convincing without further expert assistance, the Court may decide that in order to conduct an epistemically responsible process it must utilize its own fact-finding powers. ⁹⁵ The time limits that the Court operates under (themselves a reflection of one of the other goals of adjudication before the Court, namely the timely settlement of the dispute between the parties) constrain the options open to the Court in utilizing those fact-finding powers. For instance, perhaps the Court will only have time for one site visit instead of two. Similarly, the fact that the Court operates in accordance with a budget that is not inexhaustible but rather set by the UN General Assembly, means that it may simply not be able to afford to appoint its own experts and travel to the site. The Court operates under these earthly constraints and any normative justification of the Court’s fact-finding process worth its salt must recognize this and take it into account. This is exactly the role that contextualizing coherence in the relation to adjudication before the ICJ plays.

4.4 The coherence framework summarized

The coherence framework set out in the preceding sections is offered as a path to rational justification of the Court’s factual determinations. In accordance with the framework, the Court may consider that a factual claim is ‘sufficiently well-founded’ when it has considered the factual claims in cases before it, assessed them against relevant coherence principles, and chosen as justified the claim that explains the contested factual scenario better than any other. In doing so the Court aims for coherence through inference to the best explanation rather than any objective truth or probabilistic assessment. The coherence framework proposed, crucially, is also a responsibilist and contextualist one, in accordance with which the Court’s judges must fulfil certain epistemic duties. Ensuring epistemic responsibility also helps to provide a crucial link to the value of truth. Finally, the lengths that the Court must go to in demonstrating coherence and responsibility are limited by the context of adjudication before the ICJ, a context which provides the outer limits of the coherence framework.

5. Potential objections to the coherence framework

Now that the coherence framework has been elaborated it is necessary to defend it against the most plausible objections. For the purposes of this article, I have grouped the main objections under four main headings which I believe represent the most significant criticisms that could be made of this framework. That said, I do not pretend that these four groups of criticisms exhaust those that may be made of the framework, but I hope that by anticipating them I may go some way to convincing the reader that this is a path worth pursuing when thinking about the normative justification of the Court’s fact-finding process. These objections, which I will explore in the following sections, are: (i) the critical challenge to coherence as a concept; (ii) that the coherence framework does not describe or fit with the practice of the Court; (iii) even if it does, the coherence framework does not provide normative justification for the Court’s practice; and (iv) even if it does, it is not the only (or even the best) way to normatively ground the Court’s practice. Each objection can be met, and a coherence framework shown to be both a descriptively and normatively desirable way to conceptualize fact-finding before the ICJ.

5.1 Objection 1: Criticisms of coherence as a notion

Before dealing with several specific objections to coherence, I would like to acknowledge important and long-running critical approaches to coherence generally. Here I focus on realist and critical scholarship which, over the course of the last century, has been spectacularly effective in, for instance, exposing the ‘transcendental’ ⁹⁶ nature of certain legal concepts, and laying bare the power dynamics that are woven throughout and perpetuated by law. Such scholarship has fundamentally altered the way that both domestic and international lawyers think about the law and its role in society. By that I mean to say that, surely, any scholar today making the argument that the law and its structures operate in a neutral, dispassionate way, equally applicable to powerful and powerless alike, would be laughed out the room.

Realist scholars such as Cohen ⁹⁷ and Llewellyn ⁹⁸ clearly showed the indeterminacy of legal norms; how they can often support completely opposite conclusions. Not only is the law often internally indeterminate in that it fails to produce determinate outcomes, it is also externally indeterminate in that one can never point to an objective justification for why one interpretation should be chosen over any other. ⁹⁹

Taking up this challenge to traditional legal doctrine, critical legal scholars such as Singer, ¹⁰⁰ Kennedy, ¹⁰¹ Dalton, ¹⁰² Freeman, ¹⁰³ and others ¹⁰⁴ later extended the critique of law’s indeterminacy. Perhaps most famously in relation to international law, Koskenniemi critiqued international law’s objectivity and normativity, or rather highlighted the supposed incompatibility between the two. ¹⁰⁵ While critical legal scholars do not deny that reasons are advanced to justify decisions, they, like the realists before them, have shown that there is often a disconnect between the stated and the actual reason why it was taken (such as power, ideology, personal preference and so on). Consequently, the critique of indeterminacy of legal concepts and norms can certainly be added to critical approaches taken to knowledge and meaning in other contexts such as linguistic theory and post-structural philosophy.

This sceptical challenge, that concepts such as coherence have no rational basis and can only express the subjective preferences of the individual and never be objectively established, ¹⁰⁶ has been the subject of much academic debate. ¹⁰⁷ It is not for me to contribute to such debates here. Rather, more modestly, my aim is to point towards what would need to be demonstrated to rebut the sceptical claim that concepts such as coherence can never have any truth-value, that they are simply subjective in nature. ¹⁰⁸

First, I do not pretend to say anything about whether the Court only ever makes factual determinations that are in fact ‘rationally justified’ – that would require empirical investigation that is beyond the scope of this article. ¹⁰⁹ My argument is a normative one, specifically that legal fact-finders, such as the Court, should only make factual determinations that can be rationally justified.

Second, it is important to say that the response to the sceptical challenge does not require me to show that coherence is a concept that exists ‘somewhere in the “external” world and that we possess a special … faculty for sensing [its] presence in the way we sense the existence of celestial bodies, coins, or vacuum cleaners’. ¹¹⁰ Rather, I need only show that evaluative propositions can have truth-value, and are not dependent on the identity of the speaker in each case. ¹¹¹

This challenge has been addressed by prominent philosophers such as Davidson, perhaps the most influential English-language philosopher of the second half of the twentieth century, who adhere to a holistic approach to meaning and knowledge. ¹¹² Such philosophers have shown the interdependence of knowledge, whether that be knowledge of the content of our own minds or the minds of others, or knowledge of the world. ¹¹³ It is through such interdependence of knowledge that we can claim that some truth-value can attach to concepts such as coherence, without claiming that these concepts exist ‘out there’ in some pure form. ¹¹⁴

All of this is to say that I believe that the undeniable contribution of critical scholarship in showing us the, often indeterminate, nature of legal concepts and the power dynamics at play does not mean that we can never attribute any meaning to concepts such as coherence. Ultimately, the coherence framework set out above assumes that the law exerts a normative force on legal actors, that such actors are at least to some extent able to separate their own personal preferences in an attempt to do justice according to the law, and that concepts are not so malleable so as to be meaningless, but rather that through intersubjectivity we can to a greater or lesser extent construct a core meaning. ¹¹⁵

5.2 Objection 2: The coherence framework does not describe the practice of the Court

The second objection that may be anticipated is that the coherence framework is not descriptively appealing or is somehow unsuited to be applied in the context of the ICJ. On the contrary, the coherence framework in fact describes to a large extent the Court’s best practice in terms of fact-finding. By this I mean to say that the coherence framework I have set out in preceding sections fits comfortably with how the main parties to the adjudicative process go about making claims about, and ultimately establishing facts in cases before the Court. This holds true both in general and also in the specific context of the ICJ.

In short, the constitutive elements of the coherence framework, namely (i) specification of the parties competing factual claims; (ii) evaluation of those claims; and (iii) the selection of the claims that are ‘sufficiently well-founded’ to use Judges Simma and Al-Khasawneh’s words is what the Court has typically done in contentious cases that have come before it. ¹¹⁶ The settled practice of the Court in encouraging the parties to submit almost any evidence they so wish before focusing on assessing the probative weight of this evidence and utilizing a variable standard of proof in making factual determinations is well-documented in this regard. ¹¹⁷ As such, the coherence framework does not require a drastic alteration of the Court’s practice, but rather represents a justification of the Court’s practice, as well as an argument in favour of reading ‘well-founded’ as coherent (with all that that entails) rather than ‘true’ or even ‘most probable’.

Readers will have noted that I took care to speak of the Court’s best practice and it is necessary to note that the Court’s practice has been criticized by commentators and its own judges alike. ¹¹⁸ There is merit to many of the criticisms made of the Court such as its use of so-called experts fantômes or its previous practice of hitherto allowing experts to evade cross-examination by appearing as counsel, and it does seem that it is now taking steps towards remedying some of the most significant issues. ¹¹⁹ That said, issues remain, and are always likely to remain given the rudimentary nature of the fact-finding provisions in the Court’s constitutive instruments, ¹²⁰ and the likelihood that factually complex cases will continue to come before it in the future. The ICJ is well-placed, however, to deal with factually complex cases. It has extensive fact-finding powers; its reputation remains high and it has control over many of the means to address the weaknesses in its fact-finding process. The coherence framework proposed in this article has both descriptive and normative value in this regard, and points the way forward for the Court. It provides the answer to the question of what we can expect of ICJ judges and, most importantly, why.

More generally, experimental empirical research suggests that a coherence approach to the establishment of the facts in the legal process has descriptive appeal. ¹²¹ The constituent elements of the coherence framework, such as inference to the best explanation, are comparable to the way that individuals typically assess the establishment of the facts. ¹²² As Amaya has stated, empirical studies show coherence ‘as a key criterion for assessing the relative plausibility of the representations that determine legal fact-finders’ decisions about disputed questions of fact’. ¹²³ Not only this, studies have shown that legal decision-makers are more likely to evaluate facts in a holistic manner with reference to empirical evidence, in line with the holistic nature of the coherence framework. ¹²⁴ In this regard it has more descriptive appeal than the mainstream probabilistic approach to the facts (that we will explore in the following section).

Of course, the fact that a coherence approach to evidence is closer to what we do as individuals or the Court does in practice does not necessarily mean that this is what we should do. In other words, the fact that the coherence framework provides support for holistic approaches to evidence such as coherence is not the same as saying that such an approach is normatively justified. A normative justification for the coherence framework is required and will be considered next.

5.3 Objection 3: The coherence framework does not provide normative justification for the Court’s practice (as an alternative to formalistic and sceptical approaches)

Not only does the coherence framework describe what the Court does, it is also capable of providing a convincing argument as to why it should continue to do so. The normative claim at the beginning of this article is that the ICJ, when making a factual determination that it wishes others to accept, must show that such a claim is justified by reasons. The coherence framework laid out in the intervening pages provides such a route to rational justification. There is widespread agreement across disciplines that coherence plays an important role in justifying claims. This includes the justification of epistemic claims, ¹²⁵ as well as moral claims, ¹²⁶ not to mention those claims relating to legal norms. ¹²⁷ While critics have challenged the vagueness or circularity of coherence approaches or their link to truth as a value, it is widely accepted that coherence approaches of the kind laid out in this article, with their attendant coherence principles and so forth, amount to a rational process. As such, any factual determination made by the Court on, for instance, a violation of the 1975 Statute of the River Uruguay which is made on the basis that the factual claims of Argentina and Uruguay are most coherent can be said to be rationally justified. Admittedly, this is not the end of the argument. Just because coherence provides a route to rational justification does not mean that it is the only, or even the most preferable route for the Court. This point still needs to be established, and it is to this that we now turn in the final section.

5.4 Objection 4: Coherence is not the only (or the best) way to normatively ground the Court’s practice

The coherence framework proposed in this article is not the only path to rational justification. Probabilistic methods or even methods based on artificial intelligence could be said to a greater or lesser extent to provide a rational basis for the Court’s decisions. Due to limitations of space, it is not possible to consider each potential route to rational justification so I will limit myself to considering the dominant, probabilistic approach to the justification of factual claims in the context of adjudication.

There is significant debate in the literature regarding the relationship between coherence approaches to the justification of factual claims and probabilistic ones, ¹²⁸ and it is perhaps worth noting that certain prominent philosophers see coherence and probabilistic approaches as complementary rather than competing. ¹²⁹ This is not the place to rehash such debates, but I will instead draw upon them where relevant. In short, I believe there are a number of advantages that the coherence framework has over probabilistic approaches in the context of international adjudication and I will address each in turn.

A common place to start when situating a coherence framework for evidence is with what Richard Lempert termed ‘new evidence scholarship’, ¹³⁰ namely the increased interest in issues of evidence, and in particular in the process of proof, in Anglo-American scholarship. ¹³¹ New evidence scholarship can be characterized by its focus on reasoning and the process of proof rather than in the formulation or interrogation of particular rules of evidence. ¹³² Central to this interest in reasoning and the process of proof has been the role that probability should play in the determination of facts. ¹³³

The dominant theory in the context of legal reasoning about facts is a particular kind of Pascalian theory, ¹³⁴ Bayes’ Theorem. ¹³⁵ It is as a response to this dominant model that coherence approaches such as the one proposed in the present article have gained traction in recent years. ¹³⁶ Bayes’ Theorem has played a prominent role in debates about evidence in more recent years, as it has been used by scholars arguing that a piece of evidence confirms a hypothesis if it increases the probability of the hypothesis. ¹³⁷ Bayes’ Theorem is designed to calculate the probability of a hypothesis relating to an event that has already taken place before and after the introduction of a particular piece of evidence, in order to determine whether such evidence increases the probability of a hypothesis or not. ¹³⁸

While Bayes’ Theorem is an undoubtedly influential and important tool for the calculation of probabilities in certain situations, a number of important concerns have been raised about its appropriateness for the legal context. Some of the main criticisms of this theorem relate to the fact that it does not account for the assignment of a particular probability value (upon which everything else rests). Proponents of the theorem argue that even incorrect or absurd values assigned as the probability of a hypothesis will be corrected over time as more pieces of evidence are added into the calculation. ¹³⁹ But this is problematic in the legal context due to the fact that both time and pieces of evidence are limited. Furthermore, and just as problematically, Bayes’ theorem assumes that agreement will be found on the likelihood of the hypothesis in the end, even though no agreement could be found regarding the assignment of the initial value of the hypothesis in the first place. ¹⁴⁰

In addition, the selection of particular pieces of evidence is a normative question, as is the assignment of a value of probability. Bayes’ Theorem cannot account for such normative questions as it operates on the assumption that such actions are taken purely rationally for the purposes of mathematical calculation. ¹⁴¹ Additional problems, such as those relating to the lack of relevant evidence upon which to make the necessary calculations, or the fact that individuals do not seem to accord probability values in a predictable manner, all cast doubt on the utility of Bayes’ Theorem in the legal context. ¹⁴²

Such are some of the shortcomings of the dominant probabilistic approach. I believe that the coherence framework set out above does not have the same drawbacks. The coherence framework avoids the (at times seemingly arbitrary) assignment of numeric values to certain facts and is generally better suited to the context of adjudication where it is not possible for problematic values to be ‘washed out’ or corrected over time since the requirement to make factual determinations in a particular case does not afford such a luxury. Furthermore, the coherence framework is unashamedly normative. Accordingly, the framework can account for the normative decisions that are made regarding the evaluation of evidence, while at the same time providing a means for the ICJ to show that its factual determinations are justified by reasons. This, in turn, is the basis for the normative value of the framework, the driving force behind the argument that the Court should continue to make factual determinations in this way and, most crucially, the explanation why.

6. Conclusion

When faced with factually complex cases such as determining whether a whaling programme is carried out for ‘scientific’ purposes, whether chemicals released into a river are harmful, or which direction a river naturally flows, it is the responsibility of the ICJ to demonstrate that its factual determinations are rationally justified. The coherence framework set out in this article provides a way of doing just that. In practical terms, this requires the ICJ judge, as legal fact-finder, to first of all: (i) specify the parties’ competing factual claims, before (ii) evaluating the coherence of the competing claims against relevant coherence principles, before (iii) selecting as justified the most coherent of the claims. In doing so the ICJ judge must act in an epistemically responsible manner by fulfilling a number of epistemic duties. This notion of epistemic responsibility also provides a crucial link to the value of truth, preventing the Court’s findings of fact from being mere abstractions from reality. Taken together, I hope the reader will agree, these constituent elements of the coherence framework do what I set out to do at the beginning of this article, namely provide a theoretical account of the Court’s fact-finding process, and a normative argument for what it should do in the future.