1. Introduction
As of 4 November 2016, the ICC had ten situations under investigation and has commenced proceedings against 39 defendants (22 ongoing and 17 completed). During the course of these investigations and the adjudication of particular cases, the chambers of the ICC issue thousands of warrants, orders, decisions, and judgments. These records contain footnotes in which the judges cite various authorities: treaties, law journal articles, party submissions, previous rulings, and judgments from other courts, among others. The aim of this article is to describe the citation practices reflected in the ICC records related to the Situation in Darfur, Sudan. The article uses the data from the citations to address three questions: What authorities does the ICC prefer to cite? Why does it prefer those authorities? and What do the patterns that emerge and the preferences say about the Court? To date, no similar studies have been conducted.
Citation practices are a worthwhile target of study because they ‘potentially open a window to better understanding of judicial decisionmaking, the development of the law, use of precedent, intercourt communication, and the structuring of relations between courts’.Footnote 1 Identifying and analyzing judicial citations provides ‘a better understanding of the inner workings of international adjudicatory institutions and their evolution over time’.Footnote 2 Citations can reveal so much because to a certain extent they form the foundation upon which judicial decisions are constructed. On a micro level, citations explain the basis for decisions by indicating the sources upon which a court relies and they legitimize decisions by anchoring them in authority.Footnote 3 On a macro level, they play an important role in the development of law, even at international courts where the doctrine of stare decisis is relaxed.Footnote 4 This second point needs some explanation. Courts are influenced and often persuaded by the judgments of other courts, including by what lies in the footnotes. An authority cited by a court, by virtue of being cited, may impact the decisions of future courts.Footnote 5 Judges are influenced by citations and in turn are aware that their citations influence other judges.Footnote 6 Indeed, secondary sources such as treatises and academic articles, by virtue of their citation, become more likely to be re-cited and thus influence future citations.Footnote 7 In this way, through the accumulated citations to preferred sources over the course of a court's history and the persuasive effect of past decisions (and their citations) on future ones, preferred authorities affect the development of the law of the citing institution and those that cite to it.
This article examines the citations related to the Situation in Darfur, Sudan – only a portion of the ICC's total citations – for three reasons. First and most practically, the Sudan matter consists of a number of records (346) that have produced a meaningful number of citations (9,203), but not so many as to be beyond the capability of the author to collect and analyze the data (the data was collected manually and each citation was individually analyzed). Second, the Sudan cases involve all three divisions of the ICC – the Pre-Trial, Trial, and Appeals Divisions – making a comparative examination among chambers possible. Finally, Sudan involves the controversial case of current Sudanese President, Omar Al Bashir, the first (and to date only) sitting head of state to be indicted by the ICC and the only case involving the charge of genocide. Al Bashir's case is singular because it ‘starkly illustrates the court's profound limitations, and the reason it is such a lightning rod’.Footnote 8 Indeed, as will be discussed, the charge of genocide significantly impacted the citation count.
The data collected for this article indicates that the ICC has clear preferences in its citation authorities. To name a few: it overwhelmingly cites its own internal decisions, particularly from Pre-Trial Chambers; it cites national court decisions (combined) more than any individual international court or tribunal (this surprising conclusion will be explained and significantly qualified in the article's discussion); it cites judgments of the International Criminal Tribunal for the former Yugoslavia and the International Criminal Tribunal for Rwanda between two to three times as frequently as those of the International Court of Justice and the European Court of Human Rights; and in establishing a factual record, it looks primarily to witness statements and then to prosecution submissions. The article also attempts to at least partially explain the practices, querying for instance whether the disproportionately high amounts of citations to Pre-Trial Chamber records are a function of the Court's young age; whether the surprisingly large numbers of citations to national court judgments are largely the result of dissenting and concurring opinions; whether the disparity between citations to factual statements from the prosecution and the defence, though expected, underscores the unequal resources and power between the parties; and whether the citations to judgments of other international courts reflect citations in the filings of the prosecution and defence. Courts, of course, cite and prefer authorities for varied and multiple reasons. This article does not aim to definitively identify those reasons, but raises some possibilities and then suggests what the citation practices may reflect about the ICC.
2. Scope of the study
This study covers all records issued by the ICC chambers related to Darfur, Sudan that were publicly available on the ICC's website, www.icc-cpi.int, as of 1 October 2015 (when data collection ceased). These 346 records contained 9,203 citations.Footnote 9 A ‘record’ is any document issued by the ICC that might contain citations; these include decisions, judgments, orders, directions, elections, notices, warrants, summonses, and concurring and separate opinions. In all of its situations and cases, the ICC issued 4,008 publicly accessible records as of 11 October 2015.Footnote 10
Citations were any authority in a footnote. Multiple citations in a single footnote were counted separately when separated by a semi-colon (unless it was clear from the context that only one citation was intended, such as a long string of page numbers separated by semi-colons), a period, or some other indication that multiple citations were intended. The highest number of citations in one record was 600 contained in 281 footnotes.Footnote 11
Citations from dissenting, concurring, and separate opinions, though perhaps not as important as those from majority opinions, have also been included in the data (except where otherwise noted) because they too reflect selection in authorities made by ICC judges, albeit individually. Moreover, as will be observed when analyzing citations to national judgments and the changing patterns of citations over time, the data appears to reveal a greater freedom for authors of dissenting, concurring, and separate opinions to cite outside traditional sources of international law. These types of opinions also generally appear to trigger greater amounts of citations, as judges are prompted to defend their decisions to their disagreeing colleagues. In this way, including dissenting, concurring, and separate opinions not only provides insight into, for example, the diminished restrictions perceived by the dissenting or concurring judges, but also necessarily into potential limitations on citing perceived by the Court's majority.
3. Methodology and coding
3.1. Overview
This article is largely empirical and quantitative. Quantitative work ‘allows us to identify patterns that are not readily discernible in unsystematic reading of opinions, and offers us significant explanatory power in certain discrete categories of cases’.Footnote 12 Although on its surface the systematic collection and reading of documents, recording of consistent features, and drawing of inferences about their meaning appears ‘simple, even trivial’, the simplicity of the method belies its usefulness: content analysis is a way of ‘generating objective, falsifiable, and reproducible knowledge about what courts do and how and why they do it’.Footnote 13
The study also has analytical, qualitative components. It labels each citation by function: whether the citation helps the Court to decide a legal issue or establish a fact. Though not as objective or replicable as the quantitative counting, the labeling increases the usefulness of the data because it allows the exclusion of citations that do not directly impact the Court's decisions. The study's assessment of possible reasons for the emergence of particular citation patterns similarly provides greater insight and depth of analysis, and in this way complements the descriptive empirical portions.
3.2. Extracting the data
Quantitative work usually involves collecting a significant amount of data, and thus the first step in this study was to extract data from the ICC records. This meant copying and pasting information from the footnotes into a database program. For this study, FileMaker 13 Pro Advanced was used. If a footnote contains a citation to a decision of the International Court of Justice (ICJ), for instance, the name of the judgment, the name of the court, the year, and the legal citation are copied and pasted into separate fields in the database. For the next citation, a new database record is created. After data are input, the software is able to tabulate and link them. It can indicate, for instance, how many times ICJ decisions were cited in the particular ICC record. Even better, after the data from all the records is entered, the software can show the number of ICJ citations over a five-year period sorted by the ICC chamber that made the citation.
3.3. Coding the function of individual citations
Extracting data was the first step. The second step was coding (or ‘labeling’) each citation's function. Coding for function is a key feature of this study because it goes beyond mechanical collection and makes the data more meaningful by identifying the reasons underlying the citation. Thus, significant effort will be made to explain the coding, describe its value, and disclose its drawbacks.
To begin, mechanically extracting data from footnotes as described above provides information about which sources a court cites. This addresses source selection. Just as important is source function: why a court chooses to cite. Judges may cite for a myriad of reasons.Footnote 14 This article does not attempt to divine the ICC judges’ reasons for each citation; it is not possible to do so with any degree of reliability, particularly from an examination of the records alone. Instead, the article determines function by evaluating a more discernible quality: how the citation is used by the Court. Was the Court using the citation because it was helpful in deciding a legal issue? Or, was it helpful in determining a fact? Or, was the Court merely reciting a source to explain, for instance, what defence counsel argued, with no direct impact on the decision?
Identifying the function of citations is crucial to generating meaningful citation patterns.Footnote 15 To illustrate, a Court's citation of an ICJ decision to help it decide a point of law almost certainly has a greater probability of impacting the development of law than, for instance, a citation to a document issued by the Registry that mechanically notes a party filing. Future chambers may be influenced by the former, but in all likelihood not by the latter. As another example, the Court uses authorities to establish facts (such as how many people died in an attack) in a markedly different manner from an authority that helps it resolve a legal ambiguity. Without this differentiation, all citations would be lumped together, regardless of function or significance.
Function can only be determined by carefully reading the text to which the footnote is appended. A determination must be made for every citation. Only humans (not software) can do this. And one human's determination of a citation's function will not always be the same as another's. This component of the study is a strength, in that it discriminates between the meaningfulness of each citation, but is also a weakness, in that it is more subjective than mechanical collection. Another drawback to categorizing citations by function is that large numbers of citations are grouped together, disregarding their varying gradations of importance. For instance, all citations that are used by the Court in determining a point of law are coded together, yet each of those citations may play a different role in the decision; one citation may be the most crucial one to the Court's ultimate decision in a judgment, while another is used to reach a tangential or less significant legal point. The coding does not account for this hierarchy of citations within each group.
3.4. Coding whether a citation reflects judicial discretion
In addition to a citation's function, another important distinction made in this study is whether the citation reflects a choice by the Court in selecting among a number of possible sources, or whether the particular authority is predetermined. To explain, the ICC does not always have discretion in selecting among sources in its application and interpretation of law. Article 21 of the Rome Statute provides that the Court must first apply the Rome Statute, the Elements of Crimes (a document adopted by the ICC Assembly of States Parties that elaborates on the crimes described in the Rome Statute), and the ICC Rules of Procedure and Evidence. These are internal ICC sources, they have the highest priority under the Statute, and they are the only ones whose proper names are provided in Article 21 (as opposed to broad categories, such as ‘treaties’ or ‘general principles of law’). Hence, the Court has no leeway – no ‘wiggle room’ – in selection; it must apply those sources, it must apply them first, and it must apply them without discretion of selection. To indicate that this category of sources does not reflect a judicial preference, a special label called ‘binding’ was created.
All other sources are referred to in Article 21 by broad category instead of by name. Similarly, authorities used for interpretation (rather than application) may be selected from a wide variety of sources pursuant to Article 31(3)(c) and Article 32 of the Vienna Convention on the Law of Treaties. Citations to these authorities – treaties, principles and rules of international law, internal ICC judgments, and others – can only be made after a selection by the Court among authorities, and thus they reflect a greater degree of the Court's exercise of discretion. These sources occupy the categories of function other than ‘binding’ and are the main focus of the article because they indicate preference and reflect discretion. The selection of one citation alone means little, but this is where quantitative work helps. By coding hundreds, or in some categories even thousands, of citations, patterns emerge about which sources are preferred and which are not. In their study of citations by the Canadian Courts of Appeal, I. Greene et al. note that deciding what to cite ‘is not mechanical’; the judges interviewed agreed that ‘appropriateness and relevance implied a selection from the enormous diversity of available precedents, and therefore some kind of evaluation of judges and courts’.Footnote 16
3.5. The categories of function
A field for ‘function’ was created in the FileMaker database with nine categories: binding, legally persuasive, factually persuasive, informational, mentioned, discussed, distinguished, criticized, and unknown due to redaction. Each category will be briefly explained (except for ‘binding’, which is discussed in the previous section).
‘Legally persuasive’ authorities were those used by the Court in deciding some point of law. The vast majority of cited authorities that helped the ICC decide legal issues came not from the ‘binding’ sources of law but from ‘legally persuasive’ authorities. These authorities were most often court decisions, but included journal articles, treatises, and UN documents, among others.
Another category that assisted in the Court's decisions, and that reflects the Court's exercise of discretion, is ‘factually persuasive’ sources. These sources convinced the Court of some fact, rather than a point of law. They were most often witness statements, although they could also be items such as non-governmental organization reports, news reports, or party submissions.
Authorities in the next category, ‘informational’ sources, were most often used by the Court to recite procedural history or arguments made by parties. There was no evidence from the records that these citations were used in helping to decide any legal or factual issue. This was the category with the highest number of citations, but is arguably the least meaningful in revealing preferences and in its impact on the Court's decisions.
‘Mentioned’ sources fall somewhere between persuasive and informational sources. They were not mere recitations, nor did the records indicate that they assisted the Court in reaching some legal decision or establishing a fact. The coding of ‘mentioned’ sources was the most difficult because in most cases it was unclear precisely how the Court was using the citation and because some lied near the line of informational or, at the other end, persuasive.
‘Discussed’ sources are identical to mentioned sources, but were discussed at greater length (usually for more than one paragraph). ‘Mentioned’ and ‘discussed’ sources, where labeling was most difficult, only comprised 7.2 per cent of the total citations.
The final three categories encompassed very few citations. ‘Distinguished’ sources were usually introduced in a party filing, but the Court found their applicability lacking. ‘Criticized’ sources were similar, but the source material (rather than its application) was treated more harshly than in distinguishing citations. Some sources were cited but, due to redactions in the text of the record, their importance was unclear. These were coded ‘unknown due to redaction’. Most of these were likely informational. Finally, some footnotes were either completely redacted or did not contain citations. These were included in the database but excluded from the analysis.
Figure 1 illustrates the distribution of citations based on function. The top four categories contained 97 per cent of all citations. Figure 1 makes clear that the majority (by a slim margin) of citations were ‘informational’ and thus of little impact on the Court's legal decisions or in determining facts. Nonetheless, there were significant numbers of factually persuasive (2,067) and legally persuasive citations (1669) distributed over a relatively small number of records (346). These two categories comprise the data primarily analyzed in this article.
Figure 1. Sudan, Citations by Function
4. Data and analysis
4.1. Citations by case
The Situation in Darfur, Sudan was opened on 6 June 2005 after being referred a little over two months earlier by the UN Security Council. The matter was assigned to Pre-Trial Chamber I. Five cases against individual defendants have been opened. One case (against Saleh Mohammed Jerbo Jamus) was terminated without prejudice after the ICC received evidence of his death. Pre-Trial Chamber I refused to confirm charges against Bahar Idriss Abu Garda for lack of evidence that he had participated in an alleged common plan to attack African Union peacekeepers. All defendants remain at large. They have been charged with crimes that include murder, rape, torture, deportation or forcible transportation of population, and pillaging. The case against Al Bashir is the first to involve an arrest warrant for a sitting head of state and the first (and so far only) involving the charge of the crime of genocide. The defendants are both members of the government and its allies, as well as opposition leaders.
Figures 2 and 3 illustrate the amount of court activity generated by the Situation and its five cases, divided by court records in Figure 2 and citations in Figure 3.
Figure 2. Sudan, Number of Court Records by Case
Figure 3. Sudan, Number of Citations by Case
In its Sudan cases, the ICC averaged approximately 27 citations per record (9,203 citations in 346 records), but the citations were not evenly distributed (see Figure 5 below). Prosecutor v. Banda (SUD4), the only Sudan case that has reached the trial stage, generated the most records and citations, for obvious reasons. Noticeably, Prosecutor v. Al Bashir (SUD2) and Prosecutor v. Abu Garda (SUD3) generated identical amounts of records but very different amounts of citations. This difference is at least partially a result of records in Al Bashir's case addressing the ground-breaking and complex charge of genocide. At the pre-trial level, a single record on this issue generated 982 citations (almost the exact difference in the amount of citations between the cases).Footnote 18 Prosecutor v. Hussein (SUD5) did not commence until 2012, which explains the lower records and citations (the other cases were opened in 2009 and Prosecutor v. Ahmad Harun and Ali Abd-al-Rahman in 2007).
4.2. Citations by chamber
The ICC is divided into three divisions comprising Pre-Trial Chambers, Trial Chambers and one Appeals Chamber. Segregating citations by chamber, especially ‘legally persuasive’ citations, may indicate which chambers face more legal issues that require extensive citing. Figures 4 and 5 illustrate the distribution of records, total citations, and average citations per record (total and legally persuasive) by chamber.
Figure 4. Sudan, Number of Records and Citations by Chamber
Figure 5. Sudan, Average Citations per Record by Chamber
The Pre-Trial Chambers produced the highest amounts of records and citations, and the Appeals Chamber the fewest, indicated by Figure 4.Footnote 19 The highest average of all types of citations (informational, legally persuasive, mentioned, etc.) per record, shown by the black bars in Figure 5, was generated by the Trial Chamber. The Pre-Trial Chambers, however, generated the highest rate of legally persuasive citations, indicated by the grey bars in Figure 5. This discrepancy can be explained by the difference in informational citations. In only 72 records, the Trial Chamber generated 1,769 informational citations, while in 239 records (over three times as many as the Trial Chamber), the Pre-Trial Chambers just barely topped the Trial Chamber with 2,301 informational citations. Nearly all of the citations of the Appeals Chamber (with the lowest average persuasive citation rate) were informational: 555 of 697 total citations.
These differences probably mean little more than that certain chambers faced a higher percentage of legal issues that required extensive citing than other chambers. Additionally, recall that the ICC is relatively young (it commenced operations in 2002; its first situation, Uganda, was referred to it in December 2003). Because cases reach the Appeals Chamber after the Pre-Trial Chamber, the Appeals Chamber has not had as much time to generate records. This also means that it has had fewer ICC records than the other chambers from which to cite. To elaborate, the Appeals Chamber produced 67 legally persuasive citations to its own previous decisions, but only two to those of the Pre-Trial Chambers and five to the Trial Chambers. It makes sense that an appellate court rarely looks to pre-trial or trial chambers as a persuasive authority, even if the ICC has no formal hierarchy. In contrast, the Pre-Trial Chambers referenced previous decisions of the Pre-Trial Chambers 556 times for persuasive citations. Hence, the Appeals Chamber had much fewer ICC records from which to cite at the time of data collection.
4.3. The effect of time: ICC sources and non-ICC sources
The ICC's relatively recent establishment makes it a particularly interesting target of time-related study because a still-maturing court's citing behaviour will likely change as it accumulates more internal precedent. Presumably, the ratio of citations to non-ICC sources in comparison with citations to ICC sources will decline over time because the Court will have more internal judgments to which it can cite.Footnote 20 The data collected for this article provides a starting point for that study. An ‘ICC source’ is any source that is issued by the Court or one of its branches (primarily judgments, but items like the Registrar's or President's orders also qualify).
Figure 6 plots the average legally persuasive citation rate to ICC sources and non-ICC sources over time. In other words, the black line in Figure 6 reflects the following formula:
$$\begin{equation*}
\frac{{{\rm{Number\ of\ legally\ persuasive\ citations\ to\ ICC\ sources\ in\ a\ given\ year}}}}{{{\rm{Number\ of\ records\ in\ the\ given\ year\ that\ contain\ legally\ persuasive\ citations}}}}
\end{equation*}$$
Figure 6. Sudan, Legally Persuasive Citation Rates Over Time
The grey line measures the same, but for legally persuasive citations to non-ICC sources. Measuring averages per record (as opposed to averages per year) eliminates the effect of any fluctuations in the number of records. It also eliminates the impact of a varying number of issues per year that require persuasive citations, because in any given record the Court can equally cite either ICC or non-ICC authorities. Records without legally persuasive citations were excluded because they would merely lower both averages equally.
Based on the hypothesis that the Court will increasingly turn to internal ICC authorities as its case law grows, one would expect the average citing rate over time to ICC sources (the black line) to increase in relation to the average citing rate to non-ICC sources (the grey line). This is to say, even though both averages may independently increase or decrease, the distance between them should increase. Figure 6 shows the opposite occurring. From 2006 to 2007, during the Court's earlier years, the inward-looking citation rate increased at a faster pace than the outward-looking rate, but in 2009 and 2012, the rate to outside sources increased more quickly. In 2015, the rates were almost equal.
The unanticipated pattern appears to be the result of certain legal issues, particularly in 2009 and 2012, prompting higher rates of citation to external authorities. During these years, two records in particular dramatically raised the average non-ICC citation rate: the 2012 ‘Decision on the Defence Request for a Temporary Stay of Proceedings’ (‘Decision on Stay of Proceedings’)Footnote 21 and the 2009 ‘Decision on the Prosecution's Application for a Warrant of Arrest against Omar Hassan Ahmad Al Bashir’ (‘Decision on Application for Warrant of Arrest’).Footnote 22 The majority opinion in the Decision on Stay of Proceedings cited legally persuasive non-ICC sources 31 times, while Judge Chile Eboe-Osuji in his concurring opinion cited them 162 times. Meanwhile, the majority in the Decision on Application for Warrant of Arrest cited judgments of international courts other than the ICC 80 times and treatises 24 times, while Judge Anita Ušacka's Separate and Partly Dissenting Opinion generated 92 persuasive citations to international courts other than the ICC. The large amounts of citations in these two records to non-ICC sources clearly account for the spikes in the grey line in Figure 6.
This data indicates that the presence of concurring and dissenting opinions appears to be a strong factor in the generation of citations, probably because it signals particularly contentious legal issues and prompts greater justification by the judges as they respond to their colleagues. Ušacka's dissent illustrates this point (Judge Eboe-Osuji's citations will be analyzed in detail in a subsequent section). The primary issue that prompted the citations was the interpretation and application of the elements of the crime of genocide. The Pre-Trial Chamber declined to include the charge of genocide in the arrest warrant for Al Bashir because:
. . . the existence of reasonable grounds to believe that the [Government of Sudan] acted with genocidal intent [was] not the only reasonable conclusion of the alleged commission by [Government of Sudan] forces, in a widespread and systematic manner, of the particularly serious war crimes and crimes against humanity.Footnote 23
In her partial dissent, Ušacka took issue with the majority's conclusion. It was sufficient for the purposes of an arrest warrant, in her view, to show that the inference of genocidal intent was reasonable, rather than requiring it to be the only reasonable conclusion. She also questioned the majority's use of the jurisprudence of other ‘legal and quasi-legal bodies’ with mandates different from the Pre-Trial Chamber.Footnote 24 A little less than one year later, the Appeals Chamber agreed with her, reversing the majority decision in part and remanding it for reconsideration.Footnote 25
If dissenting and concurring opinions are omitted from the data, the gap between legally persuasive citations to ICC and non-ICC sources changes measurably. There were only eight dissenting opinions and two concurring opinions in all the records examined, yet they accounted for 736 of the 9,203 citations studied, 330 of which were legally persuasive citations. Figure 7 shows average legally persuasive citation rates with all dissenting and concurring opinions excluded.
Figure 7. Sudan, Legally Persuasive Citation Rates Over Time Without Dissenting and Concurring Opinions
With these individual opinions excluded, the average citation rates to ICC authorities remained almost unchanged, while the rate to non-ICC sources significantly decreased, particularly in 2009 and 2012. The separation between the rates, however, did not see any noticeable increase over time. Additional time and data will help clarify the effect, if any, of time on the citation rates. Although the data did not fully support the hypothesis that the gap between the rates would increase over time, it served useful to highlight the significance of dissenting and concurring opinions to citation rates.
4.4. Legally persuasive citations to court judgments
The vast majority of the ICC's legally persuasive citations – 85 per cent (1,422 citations) – were to court judgments (nearly three-quarters of these to ICC records). The Court made only 247 legally persuasive citations to sources other than court judgments (these are analyzed in the next section). This article, therefore, devotes significant attention to these citations to court rulings. Figure 8 shows their distribution.
Figure 8. Sudan, Legally Persuasive Citations to Courts
The points of interest in Figure 8 that will be analyzed in greater detail are the noticeable difference among the amounts of citations to the three ICC chambers; the surprisingly high number of citations to national courts; and the potential impact of Prosecution and Defence filings on the citations to judgments of the ICTY, ICTR, ICJ and ECHR.
4.4.1. Differences in citations to the three ICC divisions
Article 21(2) of the Rome Statute provides that the ICC may (but need not) apply principles and rules of law as interpreted in its previous decisions. Of the 1,017 legally persuasive citations to previous ICC records, 59 per cent are to Pre-Trial Chamber records, 27 per cent to Appeals Chamber records, and 14 per cent to Trial Chamber records. The high number of citations to Pre-Trial Chamber records is probably best explained by the fact (alluded to previously in relation to Figures 4 and 5) that in any given case the Pre-Trial Chamber begins its work before the Appeals and Trial Chamber, and thus the number of records available at the time that the Court made a citation was disproportionately in favour of Pre-Trial Chambers records. Figure 9 helps explain this phenomenon by showing the rise and fall of records issued in absolute amounts over the 13 years measured (all records, not only those related to Sudan).
Figure 9. All ICC Records Over Time
Figure 9 clearly indicates the earlier accumulation of Pre-Trial Chamber records, and the late and sudden production of records by the Trial Chambers. (Area graphs like Figure 9 are best used to understand proportionality among the amounts. Generally, the y-axis (here, Amount of Records) is only useful to measure the lowest data (here, the records of the Pre-Trial Chambers). The other amounts are stacked on top of, rather than overlapping, the bottom amount.)
Figure 9 indicates that there were more Pre-Trial Chamber records than Appeals or Trial Chamber records until 2009, at which point the Trial Chamber records gained the lead. To illustrate the effect of different amounts of records available for citing at any particular date, a chamber issuing a record in 2008, for instance, had very few Trial or Appeals Chamber records from which to cite, but a multitude of Pre-Trial Chamber records. Recall, moreover, that the Pre-Trial Chambers in the Sudan cases issued over three times as many records as the Trial Chamber and over six times as many as the Appeals Chamber. As mentioned earlier, Pre-Trial Chambers will likely often, though of course not always, address issues that are common to Pre-Trial Chambers – pre-trial issues – and thus will likely look more frequently to either their own previous decisions or to decisions of other Pre-Trial Chambers than to Trial or Appeals Chamber records.
4.4.2. The high number of citations to national judgments
An unanticipated statistic was the high proportion (10.1%, in Figure 8) of citations to national judgments. One would expect a prominent international court like the ICC to try to avoid significant amounts of citations to national courts in a case – here, involving grave crimes in Sudan – that has nothing to do with those nations or their law (no Sudanese or any other African nation's judgments was cited). Article 21(1)(c) of the Rome Statute provides that, as a last resort when it cannot find other law to apply, the Court must apply general principles of law derived from national laws of legal systems of the world. Were the citations to national judgments reflecting a broad-based application of general principles?
They were not. Instead, the citations to national courts (and indeed many of those to international courts) in this instance were a demonstration of the impact of a contentious issue: all 143 citations to national judgments are contained in one record. Like the citations to non-ICC authorities generally, this notable concentration appears to reflect the infrequent, but nonetheless important, presence of concurring and dissenting opinions.Footnote 26 Authoring one's own opinion ostensibly unshackles judges from the citation constraints that marshalling a majority view might present.Footnote 27 Of the 143 national judgments, 19 came from the majority opinion and 124 from Eboe-Osuji's concurring opinion in the Decision on Stay of Proceedings.
Were so many citations to national judgments inappropriate? Probably not. Defence counsel had requested a temporary stay of proceedings, claiming that the Government of Sudan's restrictions on access to witnesses and other evidence in Sudan made an effective defence impossible, and thus deprived the defendants of a fair trial. The primary issue, and what prompted the unusually large number of citations to domestic court judgments, was whether a stay of proceedings was appropriate. Noting that the concept of a stay originated from common law, and in particular the jurisprudence of England and Wales, the Court looked to ICTR, US, and English law (and internal ICC decisions) to help it determine the standards for granting stays. It concluded nonetheless that such a ‘drastic remedy’ was inappropriate.Footnote 28
In his concurring opinion, Judge Eboe-Osuji found the majority's analysis inadequate. Delving ‘deep in the forest of common law jurisprudence on stay’,Footnote 29 Eboe-Osuji cited Australian judgments six times (two separate judgments), Canadian judgments 12 times (eight judgments), English judgments 79 times (33 judgments), one Irish judgment once, Scottish judgments three times (two judgments), and US judgments 23 times (16 judgments). Like the majority, he justified citations to national courts by pointing to the origins of the stay and the other courts’ ‘experience working with the concept’.Footnote 30
A close examination of Eboe-Osuji's citations indicates that his choice of authorities may to some extent reflect his personal history. Personal attributes and experiences like a judge's religion, political party, region of origin (such as from the southern and northern parts of the US), and prior political or judicial experience have been shown to affect judicial decision-making.Footnote 31 No study has been conducted on the impact of a judge's background on citation selection, yet it appears almost certain that some of Eboe-Osuji's would not have been cited but for his background. Citations were made to, for example, eight Canadian judgments, the 1867 Constitution Act of Canada, the website of the Supreme Court of Canada, the website of the Superior Court of Justice of Ontario, the website of the Courts of British Columbia, and the website of Canada's Department of Justice. Eboe-Osuji practiced and lectured law in Canada, and earned a Master of Laws from McGill University, Canada.Footnote 32 This connection should not be overstated, and the suggestion here is not that an ICC judge's background determines the authorities that judge will cite (for instance, there is no evidence that Korean judge Sang-Hyun Song has ever cited a Korean judgment or website); yet the citations indicate that the backgrounds of the ICC's judges, which are diverse, probably influence the Court's citation patterns, particularly in individual opinions. If correct, it follows that these backgrounds also have the potential to impact (perhaps only slightly but any influence is noteworthy) the development of the Court's law. Indeed, Eboe-Osuji implicitly acknowledges this when he explains that his discussion of the stay is particularly important because the Court's jurisprudence is ‘not set in stone on any reasonable view; but will continue to evolve’.Footnote 33 It would behove an advocate before the Court to consider these preferences.
4.4.3. The impact of Prosecution and Defence filings on citations
In choosing sources to cite, judges do not act in isolation. Perhaps foremost at their attention are the pleadings of the parties before the court, in particular those of the Prosecution and Defence. Legal scholars have pointed to a ‘history of judicial opinions “plagiarizing” attorney briefs’Footnote 34 and have called the adoption of the prevailing party's submission, at least in US courts, a ‘widespread practice’.Footnote 35 In assessing citations, I. Greene et al. advise: ‘The role of counsel in providing judges with grist for their decision-writing mills must never be overlooked. It is counsel who provide appellate judges with most of their ideas for sources of citations.’Footnote 36
This study tests this idea by analyzing the impact of prosecution and defence filings on ICC citations to judgments of the ICTY, ICTR, ECHR, and ICJ (these courts and tribunals were selected because, after citations to ICC records, they received the highest number of legally persuasive citations). Reportedly, parties before the ICC have ‘constantly’ referred to decisions of the ICTY and ICTR.Footnote 37 A finding that the ICC citations mirror those of the parties would be a strong indication that judicial preferences are highly influenced by party filings rather than a product of independent court research.
The data analyzed to test the impact of Prosecutor or Defence filings was restricted to legally persuasive citations in majority opinions (dissenting and concurring opinions were omitted due to their skewing effect on citation patterns, described above). First, each of the ICC records with legally persuasive citations to the ICTY, ICTR, ECHR, and ICJ was identified. Conveniently, at the beginning of each of these records, the ICC chamber describes the relevant related filings of the Prosecution and Defence. The second step, therefore, was to check each Prosecution or Defence filing for matching citations to the ICTY, ICTR, ECHR, or ICJ when the judgment cited in the ICC record was the exact same one as in the party filing.
The method is not foolproof. Some of the ICC's citations may have originated, for instance, from an amicus curiae brief, although these and other non-party filings were relatively uncommon. Additionally, a number of confidential records, including annexes, filed by the Prosecution and Defence were unavailable, but they were scarce. Importantly as well, the data does not show citations by the Prosecution or Defence to international courts and tribunals that were not subsequently used by the ICC (throughout the examination a number were detected). Finally, even when there was a match, it is not possible to know whether the Court actually relied on the citations from the parties’ submissions or whether the Court found those citations independently.
Despite these limitations, the data reveals probable influence by the Prosecution and Defence. Figure 10 shows the ICC's citations to the decisions of the ICTY, ICTR, ECHR, and ICJ that were previously cited by the Prosecution or Defence in their related filings.Footnote 38
Figure 10. Sudan, Legally Persuasive Citations Cited Previously by Prosecution or Defence
The data indicates that the Court may look about half of the time to Prosecution and Defence filings for citations to the ICTR and about a quarter of the time for ICTY citations, but that it never (or perhaps rarely, given that the data in this study is limited) adopts its citations to the ECHR and ICJ from the parties’ pleadings. A few preliminary ideas may be advanced regarding these patterns. First, the ICC does not adopt wholesale the prevailing briefs or even authorities of the parties, as reportedly occurs in some other jurisdictions. Second, the parties may be over-citing the ICTY and the ICTR (there were many ICTY and ICTR citations in party filings that were not cited in the Court's records) and are almost certainly under-citing the ECHR and ICJ, given the different proportions of matching citations. They may want to reconsider how the decisions of the latter two courts substantiate their legal arguments. Third, the Court appears to look disproportionately to the Prosecution's rather than the Defence's citations. This may be partially the result of the imbalance in related Prosecution and Defence filings in the Sudan matters during the studied period (161 and 92 respectively).Footnote 39
4.5. Citations to factually persuasive sources
In addition to deciding legal issues, the judges of the ICC must at times weigh the evidence to determine whether, for instance, the facts indicate that an arrest warrant would be appropriate. To do so, the Court looks to factually persuasive sources. Trial Chamber I has declared that it ‘enjoys a significant degree of discretion in considering all types of evidence’ and that the drafters of the Rome Statute deliberately avoided limiting the Chamber's ability to assess evidence freely.Footnote 40 Rule 63 of the Court's Rules of Procedure and Evidence supports this view, providing that the Court has the authority to ‘assess freely all evidence submitted in order to determine its relevance or admissibility’. These rules apply to the Trial Chamber; there is no indication that the other chambers do not have equal or even greater liberty to accept factual evidence. This freedom makes a study of these citations particularly compelling because, perhaps even more than with legally persuasive citations, the Court's factually persuasive citations reflect an exercise of relatively unfettered discretion. They similarly reveal preferences, indicating to future benches (and advocates before the Court) the types of sources that the Court finds most reliable (and thus which sources should be raised to the Court's attention). Figure 11 shows the Court's reliance on sources that it found factually persuasive.
Figure 11. Sudan, Factually Persuasive Citations
Prosecution submissions and witness statements, indicated by the black portions of the top two bars, comprise the bulk of factually persuasive sources. As the prosecution is responsible for investigations, its submissions naturally contain more facts than the defence's. The data reflects this incongruity: of the factually persuasive citations to prosecution and defence records, the prosecution garnered 94 per cent (535) of the citations but only submitted 61 per cent (161 records) of the total prosecution and defence records combined (253), while the defence submitted 35 per cent (92) of the records but only garnered 3 per cent (16) of the citations.Footnote 41
This discrepancy arguably reflects the greater power of the prosecution. In the investigation and production of evidence in criminal cases, the state (or at the ICC, the prosecution) has a clear advantage over defence counsel due to its vast resources and power.Footnote 42 This difference is magnified in international cases where a number of disadvantages arise for defence counsel.Footnote 43 For instance, state and international agencies and non-governmental organizations sometimes refuse to co-operate with defence counsel; the prosecution's clout gives it an advantage in access to evidence; states are sometimes reluctant to provide exculpatory evidence to the defence because that evidence may be potentially embarrassing; witnesses are ‘disinclined to stick their necks out for an accused, whose fate is of no consequence to them’ (while refusal to testify for the prosecutor could lead to a potential indictment); and the prosecution's experts are often salaried-employees who have been involved in the strategizing and planning of the prosecution's case.Footnote 44 The data collected in this study, which indicates a disproportionately high reliance by the ICC on the prosecution's filings for facts, supports the argument that this disparity disadvantages the defence.
Witness statements were the most-cited sources of factually persuasive material. The most-cited statements were from Witness 446 (45 factually persuasive citations), Witness 419 (43), and Witness 420 (40), all members of the African Union Mission in Sudan allegedly attacked by government forces or their militias on 29 September 2007 at the Military Group Site in Haskanita, South Sudan.Footnote 45 Twelve peacekeepers and civilian police officers were killed and at least eight seriously injured.Footnote 46 The cases against defendants Abu Garda, Abakaer, and Jamus stemmed largely from this attack.Footnote 47 All of the factually persuasive citations to witness statements in the Sudan cases (822) were contained in only eight court records, and all of those were made by Pre-Trial Chamber I. The vast majority (64.7%) of factually persuasive citations to witness statements in the Sudan cases came from confirmations of charges, and these citations were primarily by witnesses who were the targets of the alleged attacks. One would of course expect citations to factually persuasive materials in confirmations of charges to be dominated by prosecution witnesses, and when related to an attack, for those witnesses to be members of the attacked group. Combined with the disadvantages experienced by defence counsel in international criminal cases, however, the data paints a picture in which, at least at pre-trial proceedings, the prosecution enjoys distinct advantages.
5. Conclusion
The ability to scrutinize the practices of courts has never been greater. Records are increasingly available online, and in the case of the ICC, at no cost. Computing software, with a few clicks, instantaneously categorizes and sorts almost unlimited data. Citation analysis has gradually made its way into the legal field. Analyses of citations to court opinions have been used to measure judicial influence,Footnote 48 the quality of judicial decision-making,Footnote 49 citation patterns of non-legal information,Footnote 50 citation use by specific judges,Footnote 51 citation of foreign judgments,Footnote 52 and the accuracy of citations.Footnote 53 Legal scholars have also used network analysis (the collection and analysis of citations within and across courts to build a web-like network) to measure the importance of cases at the US Supreme CourtFootnote 54 and to determine whether the ECHR chooses case citations to satisfy domestic legal audiences.Footnote 55 Citation analysis has, in short, revealed practices that were previously obscured due to the impracticability of measuring them.
Ultimately though, the value of citation analysis depends on its usefulness. It is the hope of this author that the data and analysis in this article will be of use to ICC judges, the ICC Prosecutor, defence counsel, amicus curiae, and legal scholars. As an example, the fact that citations to non-ICC sources dramatically increase in judgments with dissenting or separate opinions may prompt judges to reflect on how controversial issues lead them to look to unconventional sources, and whether this is proper given the Rome Statute's and the Vienna Convention on the Law of Treaty's guidelines. The fact that Court citations to ICTY and ICTR judgments, but not ECHR and ICJ decisions, match citations in the prosecution and defence pleadings may help counsel recalibrate the authorities used. For legal scholars who study the balance of power between the prosecution and defence, the disparate amount of factually persuasive citations to prosecution pleadings may be useful in gauging the impact of different levels of involvement between the parties at different stages of the proceedings, and the dissimilar resources and obstacles faced by the parties in collecting evidence.
This article merely provides partial insight, however, into the ICC's citation practices. The data has other points of interest and the potential to reveal much more. For instance, the majority (66%) of factually persuasive citations to NGO reports are to those authored by Human Rights Watch. Legally persuasive citations to treatises outpaced citations to treaties, UN reports and journal articles, with nearly half of them made to three tomes: Commentary on the Rome Statute of the International Criminal Court: Observer's Notes, Article by Article, edited by Otto Triffterer and Kai Ambos (15 citations), Abuse of Process in Criminal Proceedings by David Corker, David Young, and Mark Summers (ten citations), and An Introduction to International Criminal Law and Procedure by Robert Cryer, Håkan Friman, Darryl Robinson, and Elizabeth Wilmshurst (seven citations). The impact of decisions of the ad hoc tribunals could be evaluated in greater detail. Greater analysis of the citations to UN sources would be another topic of potential interest. Future studies could expand beyond the Sudan cases to include a comparative analysis of citations between situations and cases, between chambers, or even between courts.
The ICC's objective to help end impunity for the perpetrators of the most serious crimes of concern to the international community is far-reaching and historic. As a result, the Court has been the subject of intense scrutiny, both in the popular press and in academia. Its citation practices, however, have remained outside the limelight. That these practices are revealed now is particularly important considering the Court's young age. These first decades will determine to a large extent the directions in which it develops.Footnote 56 The judges of the ICC have acknowledged their role in this formational phase.Footnote 57 It is now, during its early years, that the Court, its proponents and detractors, and the advocates who appear before it have a greater chance to consciously mould citation practices. It is this author's hope that this study will better enable such moulding to occur in an informed manner.
