Abstract
This article builds on theories about the expressive function of law and uses Structural Topic Modelling to examine how the prioritisation of civil and political rights (CPR) issues by the Extraordinary Chambers in the Courts of Cambodia (ECCC) has affected the agendas of Cambodian human rights NGOs with an international profile. It asks whether these NGOs’ focus on CPR issues can be traced back to the near-exclusive focus on CPR issues by the court, and whether this has implications for the creation of a “thick” kind of human rights accountability. It argues that, considering the nature of the Khmer Rouge's genocidal policy, it would have been within the mandate and capacity of the court to pay more attention to actions that also constituted violations of economic, social, and cultural rights (ESCR). The fact that the court did not do this and instead almost completely obscured ESCR rhetorically has triggered a similar blind spot for ESCR issues on the part of human rights NGOs, which could have otherwise played an important role in creating a culture of accountability around this category of human rights. Does this mean that violators of ESCR are more likely to escape prosecution going forward?
Keywords
Introduction
In her recent work on economic, social, and cultural rights (ESCR) in international criminal law, Evelyne Schmid opens with a reference to leading Nazi lawyer Hans Frank, who wrote in one of his diaries that the annihilation of Jews should not take place by shooting or poisoning them, but rather through more “indirect” means such as starvation, the withholding of medical care, or relocation to ghettos with poor health infrastructure (Schmid 2015). This strategy, Schmid argues, is indicative of an awareness on the side of the Nazi leadership that deliberately starving people would receive only marginal attention as compared to other types of atrocities. Also in more recent cases, genocidaires have often turned to these “indirect” strategies – which could be defined as massive violations of ESCR – in an attempt to keep their genocidal policies below the international community's radar. It is important to ask why this is. International criminal law, after all, also allows for the prosecution of conduct that de facto constitutes a violation of ESCR.
Cambodia is one of those cases where massive violations of victims’ ESCR (in the form of organised famines) were a central component of the Khmer Rouge's genocidal policy, which brought death to nearly a quarter of Cambodia's eight million people between April 1975 and January 1979 (Genocide Studies Program 2015). Nevertheless, as the quantitative analysis in this article shows, references to issues related to ESCR (including access to food) are notoriously absent from the discourse of the Extraordinary Chambers in the Courts of Cambodia (ECCC), a hybrid tribunal established in 2006 to bring to trial “those most responsible” for international and select national crimes committed under the Khmer Rouge regime.
In this article, 1 I explore how the ECCC's systematic lack of attention to ESCR violations (even in those cases where these violations overlapped with conduct that gives rise to criminal responsibility under international criminal law) influences accountability for ESCR violations in the “post-conflict” period. To do so, I focus on the expressive function of law, and more specifically on how the work of the ECCC affects NGOs’ and rights users’ understandings of what their rights are. 2 I argue that the kind of legal accountability that is foreseen in both international human rights law and international criminal law is only meaningful if there are rights users with an adequate understanding of (a) what constitutes a violation of their rights, (b) who can be held accountable, and (c) by which means.
Acknowledgements and funding details: The author wishes to thank Ken MacLean and the organisers and participants of the “Holding Accountability Accountable Conference” for their feedback on a first draft of this paper. The author also wishes to thank Gwyneth McClendon, Jean-Pierre Charriau, Francis Chateaureynaud, Steve Beissinger, and Jing Liao for advice on the data analysis, and Marcos Zunino for advice on the selection of documents. The research was made possible through the support of the Center for Human Rights and Global Justice at NYU's Law School, the Institute of Advanced Study in Berlin, the Marie-Sklodowska-Curie Actions – COFUND Programme – FP7, and the Human Rights Center at Ghent University.
Declaration of interest statement: The author declares no conflict of interest.
It is particularly important, therefore, to point out the importance of victim participation in the ECCC, which meant that a relatively high number of victims participated directly in the ECCC's proceedings and were directly exposed to its discourse on site. Moreover, because of the focus on victim participation, several dedicated structures (such as the victim support section and the public affairs section) were established within the court to communicate its work and organise outreach programmes (over 400,000 people have been reached since 2009) (ECCC 2014), and many so-called victim-support NGOs emerged, which later evolved into a broader type of human rights NGO. As such, the ECCC's focus on victim participation exposed many Cambodians to an institutional discourse on international criminal justice and human rights for the first time. This discourse was heavily shaped by international criminal law practitioners, who often proposed a rather restricted and hierarchical understanding of the kind of conduct that could (or, indeed, should) be prosecuted and prioritised the prosecution of conduct that violated civil and political rights (CPR) over conduct that violated economic, social, and cultural rights (ESCR).
The supposition inspiring this article is that this blind spot on the part of the ECCC for violations of ESCR also created a blind spot for violations of ESCR on the part of the human rights NGOs that operate in its shadow. This increases the chance that perpetrators of ESCR abuses will also escape prosecution in the future because no mobilisation that could contribute to a culture of accountability for ESCR issues is emerging. In the following, I first outline the theoretical and legal framework as well as the research methods. I then discuss the findings of a large-n comparative discourse analysis to argue that more attention to the expressive function of law is required if we are interested in the long-term, indirect, and potentially unforeseen effects of victim participation in post-conflict societies.
Legal Accountability for ESCR Violations
Legal accountability for ESCR violations is foreseen in international human rights law as well as, to some extent, in international criminal law, both of which are suborders of the international legal order and which have distinct but connected social functions and premises. It is in societies where crimes against humanity have taken place that the most significant overlaps between both suborders tend to exist. Below, I explore how accountability for ESCR violations can be made meaningful in the context of the ECCC by discussing how legal accountability operates, overlaps, and interacts in both suborders.
… in International Human Rights Law
International human rights law is a subset of public international law that entails the commitment of nation states. The applicability of international human rights treaties to ratifying states is not controversial, and there is no ambiguity about the general principle that states are accountable for human rights standards. As such, human rights violations exist only where an act or omission by the state is not in conformity with its human rights obligations. 3
The term “violation” should only be used where a legal obligation exists and where a corresponding failure to meet that obligation can be identified, not where there is merely a suboptimal situation with regard to ESCR.
For each of the rights contained in the International Covenant on Economic, Social and Cultural Rights (ICESCR), the Committee on Economic, Social and Cultural Rights (CESCR) has established a number of criteria (e.g. in its Concluding Observations and General Comments) to assess whether a state is complying with its legal obligations under the Covenant. The Maastricht Guidelines on violations of ESCR and several other instruments provide further guidance in determining whether an obligation related to ESCR has been breached. Combined, the treaties and interpretive instruments facilitate a fairly rigorous understanding of states’ obligation to respect, protect, and fulfil ESCR and establish the legal implications of ESCR. As such, concrete legal obligations for duty-bearers arise from treaties such as the ICESCR. 4
For example, state parties are required to take “progressive action,” to abstain from and prevent discrimination, to ensure minimum subsistence rights, to not engage in violations of ESCR themselves and to sanction non-state actors who do, and to offer judicial remedies where appropriate (Leckie and Gallagher 2006).
Nevertheless, the international human rights enforcement machinery is much stronger for CPR than for ESCR, and despite a few exceptions where human rights courts have adjudicated large-scale abuses of ESCR, 5 many legal practitioners consider violations of CPR more serious than even massive and systematic violations of ESCR. Moreover, the debate about the justiciability of ESCR (i.e. the capacity for adjudicating such cases in court) has to some extent become a self-fulfilling prophecy: a widespread assumption about their vagueness (and thus injusticiability) has led to disengagement by academics and practitioners and has largely deterred courts from using the means at their disposal to remedy violations and from developing the kind of jurisprudence that would facilitate further judicial consideration (Schmid 2015; also see Schmid and Nolan 2014).
For example, the cases of the Mapiripán Massacre and the Ituango Massacres in the Inter-American Court of Human Rights, the cases considered by the Human Rights Chamber for Bosnia and Herzegovina (not to be confused with the International Criminal Tribunal for Yugoslavia), the COHRE vs. Sudan case before the African Commission on Human and People's Rights, or cases of internally displaced persons (El Salado) before the Colombian Constitutional Court (United Nations 2014).
Yet, even if the adjudication of ESCR has historically been a contentious area in international human rights law, a range of measures exist to give substance to, and facilitate accountability for, ESCR. These include (quasi-)judicial, political, and administrative mechanisms and independent bodies such as national human rights institutions, as well as the periodic country reports to the CESCR; the hearing of individual communications regarding violations of the ICESCR by the CESCR, CEDAW, or CRC; the monitoring of compliance with ICESCR and other treaties protecting aspects of ESCR by UN Special Rapporteurs; and, at a regional level, the hearing of cases before the European Court of Human Rights and the European Committee of Social Rights, the Inter-American Human Rights System, and the African Commission and Court of Human and People's Rights. At the domestic level, national courts can be instrumental in adjudicating ESCR (von Tigerstrom 2001; United Nations 2010; Ratner, Abrams, and Bischoff 2009).
As such, it is beyond doubt that accountability mechanisms for violations of ESCR exist and acceptance of their justiciability is growing. 6 Moreover, it is also widely accepted that legal protection of ESCR does not cease to exist in times of armed conflict. 7 However, in this context it is relevant to examine whether conduct qualified by human rights lawyers as violations of ESCR can sometimes also be prosecuted under international criminal law.
For example, CESCR General Comment No. 3: “Among the measures which might be considered appropriate […] is the provision of judicial remedies with respect to the rights which may, in accordance with the national legal system, be considered justiciable.”
Armed conflict may decrease the means available to realise ESCR, but no specific derogation clauses are foreseen in the ICESCR (Alston and Quinn 1987).
… in International Criminal Law
International crimes are those for which (individual) criminal liability is directly established in international law (treaty, customary, or peremptory norms) and for which mechanisms are foreseen for international cooperation and enforcement in the repression of the conduct. While not every ESCR violation gives rise to criminal responsibility per se and while not all conduct amounting to an international crime is also a violation of ESCR, the same facts that constitute a violation of international criminal law might also constitute a violation of ESCR. As such, conduct that violates ESCR can, de lege lata, be addressed in processes dealing with international crimes (Schmid 2015; Drumbl 2009; van den Herik 2014).
Nevertheless, an industry of praxis has emerged in which the conventional position has been to assume that violations of ESCR are beyond the scope of international criminal law, despite a considerable number of precedents to the contrary: they are seen as aspirational goals to be realised progressively, 8 as the backdrop against which CPR violations take place, 9 as too vague to lend themselves to (quasi-)judicial action, 10 as issues unrelated to life and integrity, 11 or as issues for which no direct criminalisation is foreseen in international criminal law. 12 These assumptions – which are learned rather than rooted in positive international law – have also been visible in most institutional mechanisms employed to deal with abusive pasts (criminal proceedings, truth commissions, reparations, etc.) as well as in the jurisdiction of the International Criminal Court (ICC), which in practice treats CPR and ESCR as a dichotomy, prioritising the former (Ocheje 2002).
There seems to be a shared (mis)understanding that ESCR violations arise exclusively because of a state's failure to act, rather than its active conduct, while criminal law is concerned precisely with the commission of acts rather than omission. This leads to a “legal impossibility” argument (on this issue, see van den Herik 2014).
Scholars of “transformative justice” argue for more attention to ESCR, but usually do so based on an interest in tackling “root causes” of problems or offering a more comprehensive answer to violent pasts that also considers structural violence (Carranza 2008; Laplante 2008; United Nations 2014; Gready and Robins 2014).
The principles of legality signify that a person may only be convicted and punished on the basis of clear and ascertainable law that was in force at the time of conduct and prescribe specific rules for interpreting legal definitions of international crimes to avoid miscarriages of justice (e.g. that the crime be prescribed with precision). This requires non-retroactivity, which is not to be confused with retrospective criminal prosecution: national tribunals of states with newly adopted legislations may adjudicate conduct related to ESCR without infringing upon the principles of legality as long as the behaviour was previously criminalised as a crime against humanity in customary international law (Schmid 2015).
For example, when the water sources of a village are poisoned by soldiers and people die as a consequence, their right to life has been violated.
Also CPR do not appear as such in the list of crimes in the statutes of international criminal tribunals, since international criminal law, unlike human rights law, does not stipulate any type of crime – ESCR-related, CPR-related, or otherwise. Instruments of international criminal law are crime-based, not rights-based, and identify a crime in terms of factual and specific elements that are pertinent to interpreting the criminal nature of the conduct. Since “the only relevant legal enquiry is whether the elements of the alleged conduct constitute a crime,” criminal courts do not prosecute human rights violations, regardless of whether they relate to CPR or ESCR (Schmid 2015: 39).
Yet, even if human rights law cannot simply be transposed into international criminal law, there is no legal explanation for why relatively little attention has been paid to ESCR in international criminal law. A critical examination of this selectivity suggests that alternative (or additional) selection of facts for criminal purposes is possible and that international criminal lawyers and judges are free to pay more attention to behaviours that de facto constitute a violation of ESCR, something which is not currently happening (also see Marcus 2003). Moreover, the claim that certain abuses fall outside the scope of international law, not because of the factual conduct that is involved, but because of the fact that they affect people's access to their ESCR rather than CPR, seriously limits the scope of current international law. As Sigrun Skogly (2001: 59) posits, “if malicious state leaders know that they may be brought to justice for massacring people, they may choose to starve them […] instead.” Thus, as long as “scholars, judges [… and] advocates believe that considerations of ESCR have no place in international criminal law, international law will not be able to maximise its potential to contribute to the resolution of complex issues that arise from ESCR violations” (Schmid 2015: 13). 13
It is only in the last decade that attention to this issue has been growing, starting with Louise Arbour's 2006 speech in which she criticised the exclusion of ESCR from transitional justice interventions and challenged the belief that ESCR are mere aspirations (Arbour 2007).
Endorsing a broader and deeper account of criminal accountability that includes violations of ESCR is also crucial from a gendered perspective in that it allows for more focused attention on those types of criminal conduct that tend to disproportionally affect women (Ni Aolain, Haynes, and Cahn 2011; Reilly 2007). Gendered discourses have historically operated to exclude female subject positions, and the uncritical incorporation of international human rights hierarchies into transitional justice practices has resulted in gendered exclusions of certain types of violence (Ni Aolain and Turner 2007).
… in Practice
For either type of accountability to be relevant in practice, however, rights users are needed who have a sound understanding of who can be held accountable for what and how. Desmet (2014) uses the notion of “human rights users” to define any individual or composite entity that engages with human rights, ranging from individuals who “use” human rights in their own name to human rights lawyers and judges, to social movements (even if they do not mobilise human rights as law). 14 The type of actions for which perpetrators are held accountable ultimately depends on these rights users’ understanding of what constitutes a violation, who may speak, and who should be held accountable. Therefore, if ESCR violations are not considered to have any overlap with current definitions of what constitutes an international crime, they are unlikely to gain attention in the criminal justice processes that are predominant in post-conflict societies and their invisibility and silencing is likely to continue.
Desmet acknowledges that marginalised groups often have easier access to the non-legal dimension of human rights, and therefore gives equal weight to the legal and non-legal dimensions.
In this context, expressive theories of international criminal law, such as that of Feinberg (1965), are particularly relevant, as these explore the extent to which the work of courts is relevant beyond the cases that these courts adjudicate. Judicial attention can be read as the expression of collective beliefs, values, and attitudes about what constitutes a crime, and may lead to behavioural changes, also in cases that are beyond the court's purview (Samson 2012; also see Snow 2004). As Fionnuala Ni Aolain (2009) argues, the under-enforcement of certain legal norms in a post-conflict context also calls into question the extent to which women will benefit from interventions, as the existing under-enforcement pattern marginalises those crimes that tend to disproportionally affect women.
This is consistent with Gready and Robins’ claim that transitional justice frames enable activists to “create” an issue and insert it into the existing transitional agenda (2014: 355). Transitional justice, as defined by the United Nations, is the full range of processes and mechanisms (including criminal prosecutions, truth telling, reparations to victims, and institutional reform) associated with a society's attempt to come to terms with a legacy of large-scale past abuses of human rights in order to ensure accountability, serve justice, and achieve reconciliation (United Nations 2014). Teitel (2014: 39), who first coined the term “transitional justice,” argues that the virtue of the international legal scheme is that it contributes to the development of a normative vocabulary, and that the normative impact of international criminal law is vast, as it often becomes the dominant language of the successor regime. This means that the framing and prioritisation that is used during transitional justice interventions can act as a gatekeeper in keeping certain issues that fall outside of this imaginary horizon off the radar. Transitional justice, in this sense, creates what Habermas describes as a disciplined framework for communicative action, where power relationships are negotiated and expressed (McEvoy 2007).
This acknowledgement of the great symbolic and educational function of transitional justice interventions (including criminal justice) renders the question about which issues criminal justice procedures should prioritise even more pertinent, in the sense that these procedures can expose wrongdoings and label these as criminal, foster a culture supportive of socio-economic rights, and offer a more complete picture of justice to victim communities (Drumbl 2009). This expressive function of law is all the more important in a context of victim-oriented transitional justice processes, where many rights users are for the first time systematically exposed to an institutional discourse on human rights and legal accountability for violations thereof. Miller's work on the effect of invisibility should be invoked to examine what exposure to a discourse that invisibilises ESCR does to the agendas of Cambodian human rights NGOs (Miller 2012).
Seeking Accountability for ESCR Violations in Cambodia
To study this question, this article analyses the discourse and priorities of the Extraordinary Chambers of the Court of Cambodia (ECCC) and its influence on the discourses and priorities of major human rights and transitional justice NGOs.
Of all the victims who died at the hands of the Khmer Rouge, approximately 50 to 70 per cent are believed to have been killed through intentional starvation, disease, or overwork. 15 These direct violations of victims’ ESCRs were part of a deliberate genocidal policy that violated the laws of war seriously enough to qualify as a war crime (DeFalco 2014). Because of this, Schmid argues that it would have been possible for the prosecutors to (also) press charges for starvation, which affected many more than those executed or detained (Schmid 2015; also see Schmid 2009). Yet, the ECCC (like most national, international, and hybrid courts that deal with criminal responsibility in the context of transitional justice) referred to “principles of feasibility” to justify a nearly exclusive focus on CPR over ESCR violations.
As witness 2-TCW-971 testified before the ECCC in December 2016, when he was sent to the north of the country in 1978, he saw storage facilities full of rice and sugar and “He did not understand why this was not given to the people. […] the people did not have sufficient food to eat” (Kijewski 2017; also see Kiernan 1996; Sliwinski 1995: 82).
While the ECCC is not unique in its focus on CPR, it is exceptional with regard to victim participation, which has been of a scope unprecedented in other international criminal proceedings (Jasini and Phan 2011). Over time a complex system for consultation and submitting requests for moral and collective reparations has been established, which has brought a large number of people into close contact with the court (ECCC 2015). Furthermore, community screenings of proceedings, radio and television broadcasts, and live attendance of proceedings have become an integral part of the ECCC's outreach work (McGonigle 2009). In addition, a Victim Support Section (VSS, formerly Victims Unit) was established to act as an intermediary between the court and those giving testimony or acting as partie civile.
Despite the existence of the VSS, a significant number of NGOs have emerged in the last decade to support victims who have participated in the ECCC. The Cambodian Human Rights and Development Association (ADHOC), for example, was established with the goal of organising monthly meetings to inform Case 001 civil party applicants about the process, as well as to give them opportunities to meet their lawyers, and to gather information about their needs and their expectations of reparations. The Cambodian Defenders Project (CDP) and Legal Aid of Cambodia (LAC), together with international pro bono civil party lawyers, have provided legal assistance to ECCC participants. Other NGOs, such as the Transcultural Psychosocial Organization (TPO), have offered psychological support before, during, and after trial to victims participating in the proceedings. In parallel with the court's own outreach, these organisations have developed and implemented an array of outreach and information programmes that have used the ECCC's legal proceedings as a stepping stone for further societal discussions and mobilisation about past human rights violations and how these interact with contemporary societal dynamics (Pham et al. 2011). While some organisations (like the Cambodia Tribunal Monitor, the Documentation Center of Cambodia (DC-CAM), the Khmer Rouge Trial Monitor, or Sleuk Rith) have so far managed to maintain their narrow focus on the ECCC, others (like CDP or ADHOC) have evolved into more general human rights NGOs. The ECCC has thus boosted an entire strand of civil society activism whose discourse implicitly or explicitly takes the work of the court itself as a point of reference.
Moreover, the assumption that participants in the ECCC would “take the benefits of that participation back to their communities and families, thus magnifying the advantages of their participation,” (OSF 2016) seems to have partly materialised, in the sense that other types of civil society activism have also increased significantly over the last decade. While no quantitative data is available, all the available qualitative sources – local and international – point out the expansion (and professionalisation) of civil society activism since 2006 (Bottomley 2015; OHCHR 2006-2016). 16 Many of these organisations claim to defend human rights (in the broadest sense) and seek accountability for human rights violations that are taking place today. Examining how the discourses and priorities of these organisations relate to the discourse and priorities of the ECCC is therefore a highly pertinent question.
Downie and Kingsbury (2001) identify another boom in civil society activism around the 1998 multiparty elections.
Methods
To study this question, publicly available documents from the ECCC and its support structures were analysed and then compared with documents produced by the 15 most prominent human rights and transitional justice NGOs that produce materials in English. These organisations were selected on the basis of expert interviews, snowballing, and a survey of publicly available organisation profiles. Of these 15, five organisations defined themselves largely with reference to the transitional justice process and often had their direct roots in mobilisation around transitional justice issues. The other 10 self-identified, instead, as human rights NGOs in a more general sense. However, many issues and topics were cross-cutting and there were many mutual citations of each other's work between these NGOs.
Two corpora of written documents were created: one consisting of the documents produced by the ECCC and its support structures, and one consisting of the documents produced by the NGOs. The two corpora were constructed in such a way as to allow for the analysis of further subsets of actors – for example, to assess differences between the official, publicly available decisions of the ECCC and the communications of its support structures (two subsets of the first corpus), or to compare different years. For reasons of feasibility and comparability, our sample was restricted to English-language documents. This means that NGOs producing only materials in Khmer or other languages were excluded, which might have resulted in the overrepresentation of NGOs with an “international profile.” However, because these are among the most prominent NGOs in Cambodia, and because the aim of this article is not to give a comprehensive overview of all civil society activism, this selection criterion should not invalidate the results of the analysis.
A vast textual dataset emerged from the collection, identification, and organisation of those documents produced by the actors examined here. These corpora were analysed using computational text-mining techniques to discover and trace the evolution of topics emerging from these texts.
For the purpose of this article I applied a probabilistic topic-modelling technique called Latent Dirichlet Allocation (LDA) (Blei 2012). In the basic LDA model used here, any document can be described as a mixture of topics that are more or less prominent in the document, and every topic is identifiable by a number of words (Blei and Lafferty 2009). For example, an LDA analysis of newspaper articles might find one topic that could be defined by the words “Russia,” “foreign,” “relations,” “secretary,” “power” and another topic with the words “health,” “debate,” “senate,” “government,” “affordable.” The analyst can then apply topic labels to indicate that one topic is focused on foreign affairs and the other topic is focused on Obamacare.
Combined, the two corpora initially consisted of a collection of 22,262 documents. After reviewing these documents for technical and substantive validity by removing files that contained no content (i.e. less than one hundred words), files that contained only information pertaining to the structure of the website (i.e. site-taxonomy files), files that had been posted on the English version of the website but were written in another language (mostly Khmer), and documents that were not archived in a usable format (e.g. photos), 12,845 documents remained: 9,645 published on the websites of NGOs and 3,200 published by the ECCC and its support structures (see Table 1). Duplicates (e.g. the same press release published on the websites of several organisations) were not removed from this final dataset because the fact that a document occurs multiple times in various places is in itself relevant to this analysis, as it is an indicator of the prominence and centrality of its content in the debate.
Corpora Overview
The documents in both corpora were then prepared for analysis by removing general and topic-specific stopwords. 17 Metadata for each document included the date of publication and original source, which allowed for the analysis of different “slices” of the corpora. This facilitated, for example, an analysis of the evolution of topics over time or of subsets of the corpora.
Examples of general stopwords in English would be “have,” “we,” etc. Examples of context-specific stopwords are “Cambodia,” “judge,” “witness” – that is to say, those words which appear frequently in these specific corpora but tell us little about the substance of the documents. These technical terms, like the names of accused persons or NGO directors, were removed after a first round of analysis so as to arrive at a more substantial understanding.
Results of the Topic Modelling
A first round of experimental analyses was undertaken to establish the appropriate number of topics and high-probability words along with the topic assignments and the perplexity. 18 The model that was adopted on the basis of these tests was then applied to the combined corpora (not accounting for publication date or source but instead combining the data) to establish which topics appeared across the board and to determine the optimal parameters for the analysis. This showed that those topics with the highest probability were all issues related to CPR and more specifically to the rule of law (T1), access to justice (T2), killings and executions (T3), detention and arbitrary arrest (T4), and fair trial (T5) (see Table 2). It was not until relatively far down the list that the first issues clearly related to ESCR started to emerge, with topics revolving around land and land rights (T13), work and poverty (T14), food and food security (T15), education (T16), and (mental) health (T17). 19
The perplexity of held-out documents describes how well the high-probability topic words in fact describe the topic. The lower the perplexity, the better the model.
Note that (1) highest-probability words are stemmed; (2) topics are ranked according to probability in this table; (3) the number of topics is fixed and written into the script here, as opposed to a Bayesian non-parametric topic model that would let the data determine the number of topics during posterior inference; (4) labels have been determined by the researcher; (5) this article does not perform a position analysis (i.e. an assessment of which argumentation is raised) but only considers whether a topic is present or not.
Tracing the Evolution of Topics
Table 2 shows that, for the combined corpora, those topics related to CPR are indeed significantly more prominent than those related to ESCR. To establish whether this is the case for both the ECCC corpus (n=3200) and the NGO corpus (n=9645), these documents can now be disaggregated again into two corpora. Moreover, the metadata can be used to “slice” the two corpora into annual segments so as to evaluate the prevalence of topics in each of the two corpora over time. This analysis shows strong parallels in how much attention each topic receives from the ECCC and the NGOs, respectively. Figure 1 shows this evolution since 2006. For analytical clarity only the first three CPR-related topics (T1: rule of law, T2: access to justice, T3: killings and executions) and the first three ESCR-related topics (T13: land, land rights and displacement; T14: work and poverty; T15: food and food security) are plotted. The intermediary topics, which follow similar patterns, are not plotted in this figure for reasons of readability.
Most Common Topics
The first thing Figure 1 shows is the greater prevalence of topics related to CPR in both the ECCC corpus and the NGO corpus. 20 The three most prominent CPR-related topics are present in 36 per cent (T1, blue line), 40 per cent (T2, gray line), and 25 per cent (T3, yellow line) of the documents produced by the ECCC and its support structures, whereas the three most prominent ESCR-related topics are present in only 6 per cent (T13, purple line), 3 per cent (T14, green line), and 9 per cent (T15, orange line) of the documents produced by the ECCC and its support structures. The topic prevalence in the NGO corpus follows suit, with topic proportions for the same topics reaching 36 per cent (T1, blue dotted line), 39 per cent (T2, grey dotted line), 22 per cent (T3, yellow dotted line), 14 per cent (T13, purple dotted line), 10 per cent (T14, green dotted line), and 4 per cent (T15, orange dotted line), respectively.
Please see online version for colour lines and colour dotted lines in all figures presented in this article: www.CurrentSoutheastAsianAffairs.org.
Topic proportions are highly similar across the board, with similarities between the ECCC and NGO corpora being most striking in the case of the CPR topics, for which the prevalence is virtually the same between the two. In the case of the ESCR topics there is a slightly greater difference in topic prevalence between the two corpora. With regard to land and land rights (T13), for example, it can be observed that since 2006 this topic has been more prominent in the NGO corpus than in the ECCC corpus and that the attention granted to this topic has steadily grown in both corpora. The topic of food (T15), on the contrary, has been more present in the discourse of the ECCC, which can probably be understood in light of the Khmer Rouge's use of organised famine as an integral part of its genocidal policy.
Evolution of Topic Proportions, ECCC and NGOs
Comparing the Evolution of Topics for Different Sets of Actors
The above discussion of how topics have evolved over time refers to a fairly high level of aggregation and says little about potential variations in the prevalence of topics between the various types of actors in both corpora. In the case of the ECCC, for example, the distinction between core decisions and publications by the court itself (such as the introductory and supplementary submissions, the closing orders by the investigating judges, the trials and appeal judgements, and the communications and publications related to these) and the documents produced by the ECCC's support structures (such as the Victim Support Section, the Lead Co-Lawyers, and the Public Affairs Section) is likely to be a significant one, since the former focus purely on the cases before the court and are therefore more likely to deal more narrowly with the exact elements pertaining to the cases, whereas the latter are produced by bodies whose primary role is to communicate with witnesses, civil parties, and a broader audience, which gives them more room for interpretation and contextualisation.
However, when the metadata is used to further disaggregate the ECCC corpus and break it down into these two subcorpora, the data still show strong similarities between the discourse of the ECCC proper and that of its support structures. Figure 2 plots the same six topics presented in Figure 1 and shows consistent similarities between the prevalence of these topics in court documents on the one hand and the documents produced by the court's support structures on the other. The three dominant CPR topics are dominant for both subcorpora and receive similar amounts of attention: rule of law (T1) appears in 38 per cent of the court documents and 35 per cent of the court's support structures’ documents, while access to justice (T2) appears in 43 and 40 per cent of the documents, respectively. What is remarkable is that the topic of killing and executions (T3) receives significantly less attention in the documents produced by the support structures’ than in those produced by the court. In both corpora, however, ESCR topics are equally uncommon, with land rights (T13), work and poverty (T14) and food security (T15) appearing in 4 per cent, 1 per cent, and 6 per cent, respectively, of the court documents proper, and in 6 per cent, 3 per cent, and 9 per cent, respectively, of the documents from the court's support structures. In each of the six cases plotted in Figure 2, the differences in topic proportions between the two subcorpora are limited, meaning that the distinctions between what the court itself prioritises and what its support structures prioritise is modest.
Evolution of Topic Proportions, ECCC
The same exercise can be undertaken for the NGO corpus. This corpus can be divided into a subcorpus of NGOs focusing primarily on transitional justice and one of NGOs focusing on human rights issues more broadly (see Table 1). The former subset could be expected to stick more closely to the rhetoric and priorities of the ECCC (i.e. to prioritise the same topics as the ECCC), whereas the latter could be expected to mobilise around a broader set of human rights issues, including those that are not necessarily prominent in the court's proceedings. Here too, however, it can be observed that the differences in topic prevalence between the two subcorpora is limited. Rule of law (T1) and access to justice (T2) are still the most dominant topics, followed, albeit somewhat more distantly than in the previous cases, by the topic of killings and executions (T3). Additionally, the ESCR topics of land (T13), poverty and work (T14), and food (T15) are less prominent for both the transitional justice and human rights NGOs. Here, too, the topic of land and land rights (T13) is something of an outlier in the sense that it was the only ESCR-related topic to receive similar amounts of attention as one of the dominant CPR topics in the corpus of human rights NGOs in 2016, and more attention than some of the intermediary CPR topics (T4-12, not plotted here) throughout.
Evolution of Topic Proportions, NGOs
Overall it can be observed that the CPR topics are somewhat more prevalent in the case of the transitional justice NGOs: T1, T2, and T3 appear in 38 per cent, 41 per cent, and 24 per cent of the documents produced by transitional justice NGOs, as opposed to 35 per cent, 38 per cent, and 21 per cent of the documents produced by human rights NGOs. ESCR topics, like T13, T14, and T15, on the contrary, are somewhat more important on the side of the human rights NGOs (16 per cent, 11 per cent, and 5 per cent, respectively) than on the side of transitional justice NGOs (12 per cent, 8 per cent, and 3 per cent, respectively).
The prevalence of topics such as rule of law (T1), access to justice (T2), executions (T3) and the like in both subcorpora is striking in a context where a majority of the population identifies a range of issues that fall entirely within the realm of ESCR as its human rights priorities for the justice process (see Figure 4). However, this can likely be understood at least partly in light of the government's violent crackdown on civil society, which might make the former issues more vital for NGOs themselves than for their constituencies.
Popular Priorities for the Justice Process
A 2008 survey conducted among adult Cambodians suggested that their priorities for the justice process were all related to ESCR rather than CPR (Pham et al. 2011). This had not changed significantly in 2010, when respondents continued to prioritise issues such as poverty and jobs (98 per cent), health (35 per cent), and land and infrastructure (both 13 per cent) over issues of justice or rule of law (3 per cent combined). To compare, in that same year, the topic prevalence for these issues on the part of human rights NGOs was 12 per cent (work and poverty), 7 per cent (health), 15 per cent (land), and 3 per cent (infrastructure), respectively – versus a 37 per cent and 32 per cent topic prevalence for “justice” and “rule of law.” While this is hardly an either-or matter, reading the findings of this topic modelling in light of the population survey suggests that NGOs are not, or not primarily, responding to people's needs and priorities when selecting their priorities, but rather to another incentive.
The ECCC's Agenda-Setting Power and the Priorities of Cambodian NGOs
The discourse analysis reveals that both the ECCC and its support structures on the one hand and Cambodian transitional justice and human rights NGOs on the other hand strongly prioritise issues related to CPR.
In the case of the ECCC, this holds true for the court itself, as well as for its support structures, which have a broader mandate. With regard to the court itself, the lack of discursive attention to ESCR is remarkable if one considers the Khmer Rouge's policy of intentional starvation and relocation (both issues related to ESCR), which affected a majority of the regime's victims. Despite the possibility under international criminal law of trying criminal conduct that de facto also constitutes a blatant violation of ESCR, this path has not been explicitly taken by the ECCC, whose discourse has revolved strongly around CPR issues. With regard to the court's support structures, the almost exclusive focus on CPR means that there have been few attempts to frame the work of the court in ways that might be more relevant to the population, which has been shown to be more concerned with ESCR-related issues. This is remarkable, considering that the role of these support structures is precisely to engage in outreach and information activities for a broader audience.
In the case of the NGOs, the strong dominance of CPR issues can be witnessed both on the part of NGOs with an explicit focus on the transitional justice process and on the part of NGOs that mobilise around human rights more generally (and that are sometimes rather critical of the work of the ECCC, as is the case for Cambodian League for the Promotion and Defence of Human Rights (LICADHO)). As such, the strong focus on CPR issues in the entire NGO corpus suggests that the professionalisation and expansion of NGO activism in the last 10 years has not been accompanied by a diversification of priorities. With regard to the transitional justice NGOs, the focus on CPR can be understood by considering that these NGOs explicitly base their work on the work of the court, since their main goal is to document, comment on, and inform people about the court's proceedings. On the side of the human rights NGOs, however, the dominance of CPR is more striking, in the sense that this does not at all reflect popular human rights priorities for the justice process. This suggests that it is not responsiveness to popular concerns but rather other influences that shape the agenda of these human rights NGOs. 21
For a cautionary note on setting agendas in service of popular demands, see Mohan and Stokke (2000). On the potential for a more dialectical agenda-setting process, see Lundy and McGovern (2008). On the potential influence of funders, see, for example, Oomen (2005).
The fact that many of these NGOs are led and staffed by highly mobile professional elites with (inter)national contacts seems relevant in this context. The priorities these elites adopt show more similarities with those established by international transitional justice practitioners and human rights bodies than with the issues considered most pressing by the largely impoverished sections of Cambodian society, such as land rights, housing, education, and access to water and sanitation facilities (Pham et al. 2011). Also, the close alignment of the priority structures in the various (sub)corpora suggests that this is more than a mere correlation, and that the ECCC's focus on CPR is not only shaping the nature of its own support structures’ outreach, but also that of transitional justice NGOs more broadly, and even the work of human rights NGOs that sometimes claim to have no formal interest in the work of the court. As such, the data suggest that the topics that are prioritised in international or hybrid criminal tribunals such as the ECCC have an agenda-setting power far beyond what has commonly been assumed.
This article does not argue that the focus on CPR is inherently problematic. Both on the side of the ECCC and on the side of the NGOs valid reasons for choosing this focus can be cited (including limited organisational capacity or recent violent crackdowns on civil society). Neither is the influence of the court on civil society actors problematic or surprising in itself. In fact, the attention to victim participation in the ECCC, the prominence of its outreach programmes, and its collaboration with NGOs are all part of a systematic attempt to reinforce civil society. Practitioners acknowledged early on that because of the limited resources and time frame of the court itself, supporting the emergence of a viable civil society would be crucial in order to generate not only a legal, but a thicker form of accountability for human rights violations – which entails various actors, processes, and realms – so that this accountability can continue to materialise after international actors leave. However, precisely because of the importance of this concern, it is important to better understand the unforeseen, long-term, or indirect effects of the ECCC's work on civil society.
Especially when considering Miller's work (2008) on the effect of invisibilisation and the extent to which a strong focus on CPR can crowd out other kinds of human rights and social justice activism, the findings of this discourse analysis need to be taken seriously. They raise questions about the nature of the human rights activism that is emerging in the shadow of the Cambodian transitional justice process and whether this is likely to generate a thicker and more multi-layered kind of accountability for all human rights violations going forward.
It should be acknowledged that this analysis has only considered NGOs that produce materials in English. As such, it has not argued that CPR are prioritised by all sectors of civil society. Indeed, it is likely that grass-roots organisations that mobilise closer to the communities they serve reflect popular demands more accurately. However, since the NGOs in this sample are amongst the most prominent and professional civil society organisations, with greater access to media, (international) funders, and policymakers, their influence on the public debate and on other sectors of civil society should not be underestimated.
Conclusion
This article has illustrated the extent to which the omission of ESCR issues from the ECCC's work has coincided with the invisibilisation of such issues in the work of 15 prominent Cambodian human rights NGOs with an international orientation, thus creating a blind spot for ESCR violations – past or present. While more work is needed to make strong claims about causality, the data underline the need for more critical reflection about the extent to which the priorities of powerful international (or hybrid) bodies such as the ECCC also – often inadvertently – shape the agendas of NGOs active in its shadow. Going forward, it would be useful to complement this quantitative analysis with a qualitative analysis of how activists and practitioners on the ground understand these results in order to assess the nature, extent, and potential bidirectionality of the influence in a more textured and fine-grained way. Precisely because of the transitional justice literature's growing attention to the strengthening of civil society, the effects that concrete interventions have on civil society actors must be better understood.
In this case the blind spot for ESCR issues on the part of an important sector of civil society raises questions about (a) the extent to which this agenda is representative of the demands of Cambodians who articulate a broader human rights and justice agenda and (b) whether a thick kind of accountability (beyond a purely legal form of accountability) for all human rights violations is emerging or whether violators of ESCR are just going to continue to escape prosecution going forward. As such, it is crucial to examine whether foregrounding ESCR more explicitly during the transitional justice process, and specifically during international criminal prosecutions, would allow for a more encompassing understanding of human rights violations and for a fuller exploration of how the existing human rights law framework can be used to hold perpetrators accountable going forward.
This raises the question of the generalisability of this case study. On the one hand, it could be argued that the fact that the ECCC had an explicit focus on victim participation makes generalisation more difficult because civil society might be less exposed to a court's discourse in other cases where there is less attention to victim participation. Yet on the other hand, what this case also shows is that even in a case where victims have exceptional access to the justice system, they still do not seem to influence the setting of priorities. This, then, suggests that victims’ agenda-setting potential might be even weaker in other situations where there is less attention to victim participation, and that in these situations the courts’ discourses are even more likely to be out of line with popular priorities.
To return to Schmid's (2015) claim that attention to ESCR is possible within the existing framework of international criminal law and that this would endorse a broader and deeper understanding of criminal accountability, this article suggests that Schmid's claim is not only relevant for legal practitioners, but also has great relevance for human rights users in a broader sense. Acknowledging and utilising the existing provisions for legal accountability for ESCR violations in international criminal law could potentially improve the expressive function thereof and could facilitate a more comprehensive rights understanding amongst rights holders. This could increase their potential to hold perpetrators accountable going forward. As such, exploiting the full scope of international criminal law would offer broader protection for human rights violations both in a narrow legal sense and in a “thicker” sense.