Ethics of Archives: Improving Historical Social Science Through the Consideration of Research on Violence

Abstract

In recent years, social scientists have “(re)discovered history” by visiting archives, collecting documents, and analyzing their findings to address concerns about the causes and consequences of violence. Nevertheless, social scientists frequently appear at their archives with little to no training on the methods and ethics of archival research as they increasingly rush to examine primary historical records. This has resulted in a dearth of discourse on how the practice of historical research influences the outcomes of our analyses. Our article, as a result, employs findings from research on political violence in sociology and political science, as well as insights from history and archival studies, to introduce three broad ethical concerns related to politics, interpretation, and harms and benefits that, we argue, have methodological implications for historical social science. These methodological implications are too often ignored in historical social science, but we contend they are necessary to consider prior to and during archival research, as well as afterward when analyzing data, in order to ensure that the results of that research are valid, reliable, and ethical despite the constraints involved in working with historical evidence. We also discuss contemporary conflicts and how data collection on violence influences our understanding of the past. The objective of this article is to identify and address the primary challenges that social scientists who work with archives encounter, as well as to advocate for increased transparency in archival research.

Introduction

In recent years, social scientists have “(re)discovered history” by visiting archives, collecting documents, and analyzing their findings to answer questions concerning violence (Finkel et al. 2019: 2). This follows a broader trend in historical social science, whereby scholars who study the past increasingly draw on primary sources (Adams et al. 2005: 27; Mayrl and Wilson 2020: 12). Historians, of course, have long contended with how best to conduct archival research, and issues of ethics in archival research have become a standard element of historical training as well as a popular focus in information studies programs. ¹ Yet, their peers in the social sciences rarely have similar training on the methods and ethics of archives. In their rush to the archives, the social scientific disciplines have largely sidestepped discourse on how the practice of historical research influences the outcomes of their analyses.

This article is motivated by our shared and simultaneous realizations (Luft 2020a; Subotić 2021) that archival research methods and ethics are not consistently addressed in sociology and political science and that methods and ethics in historical social science are intertwined. Given our own expertise, we draw primarily from research on political violence in sociology and political science, as well as insights from historical and archival studies, to raise three ethical concerns related to (1) the politics of archival research, (2) the challenges of interpretation, and (3) the consideration of harms and benefits. We also propose three interconnected recommendations: scholars should describe and contextualize the provenance of their archives; they should specify alternative interpretations of materials along with an explanation of the reasoning behind their chosen interpretations; and they should discuss how considerations of harms and benefits influence their research methods and writing.

Ethical concerns regarding archival research on violence

The methodological implications of ethical complications in archival research are seldom discussed in historical social science. This holds even for research on political violence, where the focus is on acts of physical violence that result in mass death and suffering, as well as on attempts to uncover individual or group motivations for violence, experiences with violence, decision-making about violence, and violence’s enduring social and political impacts (for recent overviews, see Balcells and Stanton 2021; Berry and Lake 2021; Chenoweth 2023; Davenport et al. 2019). This is surprising given that the ethics of research on violence, in and of itself, has received significant attention in recent years (e.g., Campbell 2017; Cronin-Furman and Lake 2018; Hoover Green and Cohen 2021; Krause 2021; Lake and Parkinson 2020).

Certainly, discussions of social scientific research ethics have proliferated over the past decade on topics as far-ranging as social movements in democratic and non-democratic contexts (Blee and Currier 2011; Gillan and Pickerill 2012; Wackenhut 2018); medical and health care (Anspach and Mizrachi 2006; Bosk 2008); migration (Bloemraad and Menjívar 2022); capitalism (Hoang 2022); policing (Stuart 2016); and research in the wake of COVID-19 (Abedi Dunia et al. 2023; Fine and Abramson 2020). Yet historical research on violence (and historical research writ large) remains under-addressed: only a handful of scholars have discussed in their work on violence the ethical dilemmas associated with working with historical documents (Einwohner 2011; 2022; Finkel 2017, Appendix A; Kligman and Verdery 2011, Appendix II; Macías 2016). Notably, historical research, whether about violence or any other topic, does not require Institutional Review Board (IRB) approval, a fact that may dampen enthusiasm for considering its potential harms. ²

To be sure, historians, as practitioners in a discipline that primarily relies on archives, regularly address the ethical dilemmas raised by their work. Many have developed tools and skills to manage the methodological and empirical consequences of these dilemmas (e.g., Fuentes 2016; Hartman 2008; 2019; Johnson 2023; Martinez 2018; Smallwood 2007). The specific ethical challenges related to historical research with digital archives have also increasingly become a topic of interest and scrutiny (Agarwal 2021; Agostinho 2019; Carusi and Jirotka 2009; Crossen-White 2015; Lerner 2022), as has the notion of analyzing archives as subjects in themselves (Burns 2020; Farge 2013; Sharpe 2020; Steedman 2011; Weld 2014; for anthropological approaches, see Stoler 2008 and Verdery 2014). Similarly, scholarship in information studies has made the ethics of archives and their consequences for research a central topic in recent years (e.g., Caswell et al. 2017; Ghaddar and Caswell 2019; Gilliland and Caswell 2016; Jimerson 2009; Lowry 2019a; Schwartz and Cook 2002). By considering archives as sites of power and narrative creation, historians and information studies scholars have challenged the authority of archives and become more reflexive about how the researcher’s own status or identity shapes their access to archives and experiences (King 2016). ³ Some have even suggested that scholars should include their archive “arrival stories” in their scholarship, just as anthropologists routinely report their “arrival” to the field (Dirks 2002 in King 2016: 27). ⁴

We believe the social sciences, including and especially our own “home” disciplines, have much to contribute to these conversations. Sociology’s engagement with dynamics of social relationships, organizations, and institutions positions it well to investigate archives as sites of inequality and to uncover how these inequalities are naturalized and reproduced through archival research. Political science, as a discipline concerned with power, is similarly suited to examine archives as sites of politics and power. It is thus odd that there is little mention of the ethics of archival research in our disciplines’ methods guides, which also offer only general guidance for research in archives (McNabb 2010: 309). One prominent guide on archival research in political science makes no mention of ethics (Frisch et al. 2012), while prominent discussions of historical methods in sociology not only overlook ethics in archival research, but archival research itself (Lara-Millán et al. 2020: 346). ⁵ Investigating further, we found that, while it is no representative sample, a collection of the 22 syllabi currently on the American Sociological Association’s Comparative-Historical Sociology Section website, compiled as part of the ASA-CHS teaching initiative started in 2020, is similarly avoidant. Only half of these syllabi discuss practical and analytical approaches to historical research methods and theorizing, and only a quarter (23 percent) discuss ethics. Likewise, a recent special issue on “conflict archives” in the Journal of Peace Research provides excellent practical advice, but nothing on ethics, while a special issue in Qualitative Sociology discusses archival work while largely, but not totally, ignoring ethics (Skarpelis 2020 is an exception).

Of course, it is reasonable to object that methods training in other forms of social science already addresses some of the issues related to the ethics of politics, interpretation, and harms and benefits that we discuss here. Certainly, the ethics of ethnography (Bosk and De Vries 2004; Edwards 2021; Murphy et al. 2021; Reyes 2018) and the ethics of experiments (Ferguson et al. 2023) have become important and popular topics in their own right. Both can help guide social scientists who engage with archives. Additionally, recent debates regarding how and even whether academics should use documents unlawfully leaked by organizations such as WikiLeaks have brought up related issues. Along with the legal challenges surrounding the handling of stolen documents, there are a number of immediate ethical concerns related to the use of clandestinely acquired archives: compromising the protections of human subjects, becoming complicit with organizations that operate outside the law, and producing more severe threats to national security (Darnton 2022). The digitization of archives has also led to more explicit discussions of data collection, methods, and ethics, as mentioned above (see also Burdick et al. 2016; Hodder and Beckingham 2022; Kim 2022; Presner 2024). We anticipate that these discussions will increase with the growing use of artificial intelligence to retrieve and organize online information (Colavizza et al. 2021; Makhortykh 2023; Rochford et al. 2023).

Simultaneously, the scant related discussions raised by social scientists who work with archives have engaged in recent years with reflections on the practice of archival research and have proposed ways to improve archival research in turn. These, however, mainly concern issues of credibility, such as how to compensate for incomplete records (Childress et al. 2020; George and Bennett 2005; Tansey 2007), how to avoid archives’ “potentially motivated and informational biases” (Bennett and Elman 2007: 183), how to use archives to make reliable inferences (Darnton 2018; Larson 2017), how to craft generalizable claims (Lara-Millán et al. 2020) and theory (Ermakoff 2019), and how to showcase the richness of archives as invaluable sources for research on violence (Balcells and Sullivan 2018). Still, seldom does this work address the ethical dilemmas raised by archival research and how they bear on our methodological practices, our analyses, and our understandings of the past.

The remainder of this article argues and demonstrates with a range of examples how the collection and analysis of archives requires scholars to make ethical decisions throughout their work. Further, these decisions impact empirical outcomes. We identify three broad ethical minefields: the politics of archives, the role of researchers in interpretation, and the evaluation of harms and benefits to research subjects. We also discuss contemporary conflicts and highlight ways that data collection on violence shapes what can be known about the past. Finally, we offer recommendations for moving forward.

Politics

Archives are not neutral. They do not transmit the unfiltered past. Rather, the political and social contexts in which documents have been collected and stored determine what exists (that is, what has been collected and kept) and how it has been categorized, classified, and made available to researchers (Derrida 1996; Foucault 2012). Archives, for example, can reflect the politics of the state in which they were created, illuminate state-society relationships, and reflect archivists’ politics, as it is they who appraise records, decide on their value for present and future uses, decide what to keep and reject, what to catalog, and the order of priority for organizing materials into collections and determining to whom and how they will be made available (Bastian 2021; Caswell 2016; Weld 2014). Taken together, this means that the social scientist conducting historical research ought to consider the archive as not only a source of data but as a subject itself worthy of interrogation. This is particularly true for archives concerning histories of violence, because battles over the construction, placement, organization, purpose, and meaning of archives can be fought amid (as well as following) conflicts by different groups invested in their possession or destruction, with consequences for research and analysis.

During the 1991 and 2003 breakdowns of state authority in Iraq, for example, diverse Kurdish and Iraqi factions seized state records. Simultaneously, US government and military personnel removed records from Iraq with the support and encouragement of some Iraqi expatriates, such as Kanan Makiya, who founded the Iraqi Research Documentation Project (later the Iraq Memory Foundation) (Alshaibi 2019; Degerald 2021; Montgomery 2019). Still others in Iraq purposefully destroyed records likely to conceal unflattering or incriminating evidence. With the 2003 US invasion of Iraq, large quantities of archival documents were burned and plundered with “permission” from the Department of Defense (DoD) (warned in advance of the looting, the DoD did little to prevent it; Eskander 2004; Montgomery 2019). Others confiscated documents for personal uses, including for blackmail and to make money on the Black Market (Mufti 2004). As Iraqi documents ceased to be records of a repressive regime, they became sites of political conflict (Whiting 2021).

Complicating matters further, Iraqi state records’ “de-territorialization” due to ongoing conflicts, the loss of information about the organization and structure of their original filings, and the rise of new social and political situations made them more open to change by those who took control of them. Degerald (2021: 170) laments, “Writing the history of Iraq under Ba’athist rule now resembles a card game in which some of the deck is missing and no one actor knows exactly which cards were shuffled into the fraction of the deck being played with. Other cards (in this case, Iraqi Ba’athist archival documents) removed from the full deck may appear at any time, and their place in the deck will not be immediately clear or verifiable.” A scholar who fails to consider these politics will fail to adequately examine this past.

Yet violence researchers also know that conflicts over archives extend well beyond those in power at any given time. For example, during and after genocide, victimized populations can disagree regarding the placement and purpose of their documents (Luft 2020a). They can also disagree on whether records of violence perpetrated against them should exist as accessible collections at all. Indigenous survivors of residential schools in Canada, for example, disagree with the Western human rights framework underlying the 2008 Truth and Reconciliation Commission (TRC) and its emphasis on truth telling, truth determination, exposing historical wrongs, and establishing redress so that “this [violence] must never happen again” (Ghaddar 2016: 9). In line with this emphasis, the TRC created an archive of over 5 million historical records about the Indian Reservation Schools system, which is now housed at the University of Manitoba. Several thousand of these records can be viewed online. Some Indigenous survivors of Canadian residential school violence have argued that they do not want their past experiences with violence or their personal information (as well as the information of witnesses and perpetrators of this violence) made accessible for research. To them, these are intimate personal stories of horrific assault and abuse as well as the consequences, for many of them and their families, of addiction, domestic violence, psychological harm, and suicide – all sensitive and traumatic information originally provided to the Independent Assessment Process (a component of the Indian Residential Schools Settlement Agreement) for the purpose of financial compensation (Luker 2020, building on Johnson 2017). Rather than sharing these records for a wider audience, some Indigenous survivors of this violence would prefer to have their records destroyed.

In fact, despite the ostensibly noble aims of researchers and the Canadian government in gathering evidence of past harms for the TRC and making such evidence available to the public, some survivors charge that this project may serve to supplant the structural changes they truly desire. Moreover, many Indigenous people simply lack faith in the Canadian government, in state institutions, and in non-Indigenous researchers, and so they do not want them to create nor make public archives about their past. There is a fear that research and information-sharing will allow a disingenuous Canadian tale of national triumph over shame to dominate their histories of violence. Tuck and Ree (2013: 640), both Indigenous, explain: “I care about your understanding, but I care more about concealing parts of myself from you. I don’t trust you very much.” Their essay, “A Glossary of Haunting,” argues that the haunting that settlers and their descendants experience as a result of the horrors they perpetrated against Indigenous communities or benefited from is the point – “I am a future ghost” – Tuck and Ree write (ibid.: 648), and in response to the harms settlers enacted, Indigenous survivors will now use their “arm[s] to determine the length of the gaze” (ibid.: 640). The destruction of archives by survivors of this violence is viewed from their perspective as a form of resistance to Western and colonial domination, including its methods of knowing, preserving, recalling, and sharing the past, as well as its justifications for why (Ghaddar 2016; Luker 2020). Though some survivors fear that the destruction of these archives would be an additional injustice, the desire of others to control how their information is used and interpreted, as well as their belief that eradication may be the only solution to preserve their own memories and understandings of the past, demonstrates how archives can be contested sites not only between powerholders and those they’ve harmed but also among survivors.

These examples show how archival materials must be contextualized within their proper social and political dynamics with explicit evaluation of what influenced their production and collection as well as their release. Likewise, the resulting research should contain a discussion of the factors that influenced the placement of documents in various archives and their categorization and classification by archivists. Not simply during episodes of violence but also after, select documents are collected and preserved and others hidden or destroyed, whether due to the complex politics involved in collecting information or to suit contemporary political agendas. Ethical issues tied to the politics of relationships between records and various communities impact the methods employed in archival research since they shape the archives researchers rely on, the records that exist, and the information they contain. Therefore, when conducting research and writing publications, we recommend that social scientists who work with archives consider, describe, and contextualize the provenance of their archives, including how they were formed and why they exist as they do currently.

Interpretation

Another challenge concerning the ethical use of archives is how to assess and represent the truth when documents don’t simply represent reality as it occurred, but reality as filtered through the perspectives and interpretations of documents’ “producers.” This is especially important when examining personal documents, such as letters and diaries. Rather than merely reflecting unfolding events, these documents perform the function of interpretation for writers as well as those who will one day access their writing. Narratives of ourselves are as complex as people, never to be understood as unfiltered depictions of contemporary life. Each of us interprets events based on our own positionalities, and both the meanings of those events and our relationships to our own positions can change over time. People may also have reasons to lie or not be wholly truthful when crafting documents, including personal and private documents, especially but not exclusively about violence (Freije and Nolan 2021). Scholars should be mindful of these issues and be clear when working with archives, as well as when selecting records to examine or feature in their research, about their decisions and why they made them.

Consider the example of Hilda Dajč, a 19-year-old Jewish nurse from Belgrade, Serbia, who was detained at the Semlin Nazi concentration and death camp in December 1941. She wrote four letters to her friends back home over the course of several months, from when she was initially deported until she was murdered in a mobile gas van. Although her letters were written over a short period, there are significant differences in their tone, language, and perspective.

Hilda entered Semlin with a positive outlook. She described herself as a volunteer on the transport, eager to help others, and wrote, “Everything is going to turn out all right, perhaps even better than my optimistic expectations.” Her last letter, however, is full of anger and spite: “We’re all just a bunch of animals that I despise,” “I hate every single one of us,” “nothing is so repulsive as the crow of people who deserve to be pitied, but who you are unable to help and can do nothing else than put yourself above them and despise them” (cited in Subotić 2021: 346). So, which is Hilda: a warm and benevolent woman, or a bitter, spiteful one? Which letter is the truth? Which best represents her experiences during the Holocaust? Which would she want us to share? Each letter offers only a snapshot, and we can easily imagine a scholar selecting the latter to depict the horrors Hilda suffered. However, the former letter tells an important story, too, one of optimism and generosity in the face of an unknown future. Both together tell a third story of Hilda’s quickly transformed experiences with and perspectives on Nazi violence and its victims (including herself) (ibid.). Combined, these three possible interpretations show how, when describing individuals’ experiences with violence, scholars’ archival choices matter greatly for their conclusions – even when the documents in question provide varying perspectives on the same or similar events written by the same person.

Similarly, consider the story of Khaim Sygal, a Jewish Yiddish teacher from L’viv, Ukraine (near the Polish border), who served as a Soviet policeman after annexation under a Nazi-Soviet agreement in 1939. Following the Nazi invasion in 1941, Sygal changed his name to Kirill Sigolenko and, identifying as a Ukrainian, joined the Ukrainian Nationalist militia. He then left these groups to serve in the Nazi occupation auxiliary forces and likely participated in the mass murder of thousands of Jews during the Holocaust (several witnesses identified him at the scenes of these crimes). Toward the end of the war, Sigolenko served as a translator for the German SD. When the war ended, he switched his identity a third time to become a Polish Jew by the name of Karl Kowalski. As Kowalski, he sought restitution as a Nazi victim and Holocaust survivor (McBride 2021).

The specifics of Sygal/Sigolenko/Kowalski’s story are unique but not rare: archives are rife with conflicts over credibility and truth, and while they present interpretive difficulties for scholars studying the past, such conflicts are “an important constitutive part of the historical record” that must be considered in their own right (Freije and Nolan 2021: 1). Concerning Sygal’s case specifically, people frequently shift how they identify in times of war and peace, and they frequently shift their behavioral stances, too (Fujii 2017; Luft 2015; Morris-Suzuki 2015; Williams 2022). If not writing about behavioral variation in genocide, however, or about kindred topics such as decision-making in war and violence over time or how ongoing experiences with violence shapes individuals’ self-understandings and social relationships – that is, topics where change is central – how ought scholars write about and examine the pasts to which these phenomena are related? From our perspective, it is vital that social scientists be at least explicit about and attentive to the messiness of lived experience as it shows up in their archives and, when imposing neat categories for purposes of analysis or neat narratives for purposes of description, be clear about the choices they made and why they made them. One need not study why Sygal changed his identity so often during the Holocaust, for example, but if using his archives to study Ukrainian Nationalist collaboration with Nazi Germany or Jewish efforts at seeking restitution, the study of both subjects becomes more complex once his broader trajectory is considered. Both ethically and methodologically, it matters which documents researchers select as representative of the past, and transparency in archival data selection and interpretation is necessary.

Relatedly, scholars must be aware of and explicit about archival silences and the need for interpretation to both uncover where and why they exist as well as how silence as absence influences their analyses and results (Blouin and Rosenberg 2011; Goldthorpe 1994; Stoler 2002; Trouillot 2015). As noted above, social and power dynamics impact not only how people experience the world but also how their experiences are recorded and stored – and by whom. Consider the transatlantic slave trade. Although the Trans-Atlantic Slave Trade Database contains meticulous records of this cruel past, including how many people were on each slave ship and how many died, we know little about who enslaved people were and their experiences from their perspectives. The only records we have are the logbooks of European officers who treated enslaved people as commodities, along with some manuscripts produced by the privileged few who knew how to write (colonists, bureaucrats, and a small number of non-white people not subject to slavery). Even then, this assessment is based on interpretation, such as the implications drawn from the similarities between how European officers described the deaths of men who jumped off a ship and how they described the weather (Rothman 2020). In seeking to describe and explain the dynamics of the transatlantic slave trade, a scholar may therefore analyze all the available logbooks and identify important details about violent deaths on the voyage from West Africa to North America, but without additional interpretation, it is difficult to tell what this journey was like either for the officers or for the millions of men, women, and children subjected to the terrible violence of slavery.

In response, scholars of slavery have developed a number of interpretive methods to read archives for lost perspectives. Hartman’s (2008) “critical fabulation,” which combines speculation with storytelling is one well-known example, but others include Smallwood (2007), who looks at the counterfactuals hidden in dominant narratives to reveal hidden perspectives and possibilities, and Fuentes (2016), who adds to Stoler’s focus on silence and reading against the grain by examining archival fragments. Fuentes, in her words, dwells “on the scars” in the archives to reveal the violence of the archive itself (Stoler 2002: 15). Johnson (2023) proposes in a more recent approach a tactic she terms “informed speculation” that utilizes, in her case, the archives of Moreau de Saint-Méry, a public official and intellectual who opposed slavery, to create a “communal biography” of those in his vicinity. She probes his archives to consider whom he did and did not write about, including enslaved people whose own voices are mostly absent except as filtered through his perspective, and in so doing combines specific and precise details from de Saint-Méry’s archive with conjecture concerning those around him. This approach, explicit about its attention to interpretation, offers a powerful ethical and methodological corrective. ⁶

The importance of considering archival absences and attending, in turn, to interpretation has also been central to historical scholarship on colonialism (Bastian et al. 2018; Mbembe 2015; Trouillot 2015; Spivak 2023; Stoler 2002), yet only occasionally makes its way into historical social science (e.g., Gordon 2008; Hammer and Itzigsohn 2021: 18–22; Sabbagh-Khoury 2024). The case of research into the Fernandeño Tataviam Band of Mission Indians, however, provides another example of how archival silences have ethical and methodological implications.

In recent years, the Fernandeño Tataviam Tribe has learned that the evidentiary materials they need to gain formal recognition from the Office of Federal Acknowledgement are held in over 20 different repositories across the United States. This dispersion is a direct result of the multiple removals and dispossessions of California Indians from their native lands and parallel settler (Spanish, Mexican, and United States) practices of collecting and confiscating indigenous materials (Champagne and Goldberg 2021). ⁷ Yet those aren’t the only difficulties. Numerous documents housed in these dispersed archives – evidence of ceremonial life and rituals, for instance – were initially created by government-commissioned anthropologists sent to document the Tribe’s purportedly uncivilized practices. In a similar fashion, census data were compiled to substantiate the Tribe’s alleged barbarism to rationalize their forcible removal and the sending of their children to residential schools (Montenegro 2019). By contrast, much of what exists in material form about the Fernandeño Tataviam and their history as produced by the Tribe themselves has been obliterated. It has disappeared due to the physical and cultural elimination of California Indians, and it has been filtered through the lens of settlers who created records to further their own violent and administrative agendas (ibid.). What remains better illustrates colonial power and practices oriented toward dispossession, relocation, and elimination rather than truths about the Tribe’s past. ⁸ In still another example, that of Israel/Palestine, sociologist Sabbagh-Khoury (2024: 27) notes how the kibbutz records produced in the early 20th century by socialist-leftist Zionists in the Hashomer Hatzair (Young Guard) settler movement about Palestinians in neighboring villages preserve not simply a history of indigenous presence but also a history of settler violence, even if unintentionally. She designates these as “archives of apprehension” due to how they expose the “reconnaissance practices” of the settlers as well as the “affective state[s] of the [Palestinian] archival subjects.”

Finally, sometimes the consequences of violence are such that there are no records that can speak to the past from either the perspectives of victims or perpetrators. This creates an additional ethical and practical challenge for social science history. The Armenian genocide is a case in point: as Gilliland and Hovhannisyan (2022) explain, the destruction of Armenians’ homelands as a result of the genocide inflicted by Young Turk radicals in Ottoman Turkey resulted in the permanent loss of nearly all documents and artifacts attesting to Armenians’ Ottoman existence. Those documents and artifacts that did survive were carried into the diaspora by displaced survivors, who were predominantly women (men, mainly, were killed). Most of these items were photographs, including portraits of Armenians from before the genocide as well as images of lynching marches and makeshift camps captured by missionaries observing the events. Presently, all possible genocide-related archives in Turkey are censored and sealed; the Turkish government refuses to acknowledge that genocide occurred (Gocek 2014). Turkey’s denial of the violent event has therefore coupled with the violence of the event itself to disappear material traces of it, leaving survivors and researchers unable to know what is lost versus what is irretrievable (Gilliland and Hovhannisyan 2022: 236). In turn, certain readings, knowledge, and interpretations of the Armenian genocide are nearly impossible. One cannot, for example, determine the original meanings of photographs without interpretation – that is, without imposing transgenerational memories and imaginaries influenced by and about the genocide.

Though we all always impose on the past our understandings in the present, this fact is something that scholars ought to be more aware of and explicit about when undertaking comparative-historical social science. Presentism is its own form of social positioning, and it is far too easy, without critically reflecting on the ethics of archival interpretation, to evaluate documents produced in the past in light of what we know, or believe we know, in the present. Sabbagh-Khoury (2024: 40) explains: There is always a “danger of functionalist or teleological reading that assimilates all social action under a singular ‘logic’ or structure, [or] an overarching ideology or practice.” This can especially be the case if scholars read archives without questioning what alternative interpretations for an event or experience might exist but does not. To counter bias in examining the past, and to address possible ethical harms as a result of these biases, we must recognize not simply that silence exists in archives and has consequences for our work, but also that “negative evidence” bears witness in its own way to events “just as does presence” (Gilliland and Hovhannisyan 2022: 250). Social scientists may develop distorted perceptions of the past if they fail to consider the lessons that silence and attention to interpretation can teach.

Harms and benefits

Scholars should also be aware of the many complexities concerning harms and benefits in research on the past with materials produced by or about people who are no longer living. For example, publishing unflattering archival documents can cause significant reputational and even physical harm to authors or their descendants, particularly to “ordinary” individuals (as opposed to public figures) who may not have anticipated that their documents or evidence about them would have been made available for study in certain ways (e.g., online and accessible by anyone anywhere, as opposed to on-site in a physical archive), and published. To be sure, deceased persons do not have absolute rights as creators of documents, nor should the expectation of privacy be absolute. Given, however, the ethical complexities of research with living people, it stands to reason that scholars should at least be mindful of the potential for reputational harm to the deceased and their descendants and to think seriously about the potential negative consequences, in addition to the benefits, of their research.

These issues are not merely theoretical: social and political contexts shift and change throughout history, and unearthed information may become more or less dangerous or damaging to subjects in different environments (Subotić 2021: 347). The case of the creation of Romani testimonies of Nazi atrocities is illustrative. As Joskowicz (2020) explains, European governments and law enforcement agencies continued to view Romani with suspicion after the Second World War. ⁹ For example, in West Germany, authorities regularly monitored, profiled, and detained Romani while enlisting (or trying to enlist) their aid in prosecuting Nazi war criminals. To hold Nazis accountable, the West German government wanted Romani to testify about their traumatic pasts, including what had happened to them and their families during the Romani Holocaust, by whom, how, where, and when. But the pressure the state exerted on Romani to disclose this material served, for many, as a disturbing reminder of how Nazi authorities had also interrogated them to collect genealogical information about them and their families, which the state then used to classify Romani as belonging to a subordinate race and to monitor, detain, deport, and kill them. Moreover, those who interrogated Romani families and demanded their cooperation in prosecuting Nazis often represented the same institution – the criminal police – that had sent so many to their deaths. Lastly, testifying required providing the state with the same information that state and corporate entities used and continue to use to surveil populations they deem suspicious. Romani have therefore had to balance their desires to be heard and for Nazi perpetrators to be prosecuted with their fears regarding potential re-targeting.

This state distrust was well-founded. In one instance, a Romani man named Bruno K., who survived Nazi concentration camps and prisons, was apprehended and imprisoned for two months for driving without a license after he missed a summons to testify about Nazi violence and was tracked down by prosecutors. Joskowicz (2020: 1213) observes that, while the resulting testimony purports to represent the experiences of a consenting witness to and survivor of the Romani Holocaust, only the protocol header for Bruno K.’s testimony makes it evident that he was at the time imprisoned in Germany. Likewise, one of the few Romani children to survive the so-called Auschwitz “Gypsy Camp” (Auschwitz-Birkenau Subcamp B-IIe), Alfred L., testified at his local police headquarters in Hannover after avoiding his scheduled interview only after the state prosecutor sent a patrol car to his address and forced him to speak about his experiences (others departed town prior to their interrogations). Joskowicz suggests Roma may also hesitate to testify because they are seen as “people without history” (see also Kapralski 2014), as “unreliable witnesses” (see also Von dem Knesebeck 2011), and as inherently immoral thieves (see also Kligman 2001). He summarizes, “Legal proceedings demanding their testimonies may have challenged the broader erasure of Romani voices from society, but they also regularly forced Romanies to speak against their will” (Joskowicz 2020: 1210). Whether or not scholars should use these testimonies in their analyses is unclear. Certainly, IRB would not have allowed Romani to have been interviewed in such circumstances, but now that the testimonies exist and are available for researchers to examine, is it ethical for them to do so? Or for them to publish excerpts in their academic work? We would suggest that, at a minimum, scholars who use these testimonies must contextualize in their writing the processes by which they were collected.

Further complicating matters is that, as mentioned with regard to behavioral variation in genocide in the preceding section on interpretation, some Romani were not only victims but also perpetrators of violence during the Holocaust. This, too, has played a role in Romani’s reluctance to provide testimony. Former Romani inmates of Auschwitz’s Gypsy Camp, for example, refrained from sharing their memories of fellow Romani who worked as block leaders and kapos (prisoner functionaries) in Nazi concentration camps out of concern for retaliation from within their own communities. Other Romani, due to their close social and personal ties with other Romani and only limited acquaintance with Nazi Germans, have only shared general details regarding the atrocities committed by SS personnel (and, because the risks associated with providing testimony on other Romani are often considerably higher than providing details on Nazi violence, they have abstained from providing specific information that could lead to reprisals or challenge the idea of Romani victimhood) (Joskowicz 2020: 1215–16). Already, as noted above, Romani are frequently stereotyped in Europe as dubious and criminal. For fear of harming their larger community, some Romani understandably hesitate to testify about others who were both victims of Nazi violence and contributed to it.

Of course, pressures to portray proper victimhood have influenced how many survivors – not just Romanis – frame and have framed their experiences with violence, as well as what questions interviewers have and have not asked of them when recording oral histories and gathering testimonies for archival purposes (Krystalli 2024). In Rwanda, for example, civilians have been required in post-genocide trials, testimonies, and memorial events to recount their experiences according to a strict government narrative that frames Hutu as only perpetrators and Tutsi as only victims. Questions and answers that deviate from this frame place interviewers and witnesses at risk of imprisonment (Chakravarty 2016; Luft and Thomson 2021; Luft 2023, Appendix A; Meierhenrich 2024; Purdeková 2015; Thomson 2009). It is extraordinarily difficult to collect oral history of Rwanda and the 1994 genocide that does not align with the current government narrative of the country’s past.

Relatedly, consider the ongoing Colombian conflict. Civilian women’s strategic submission to unwanted sex has in some instances protected them, but, when testifying about it, has also undercut their access to rights and redress by challenging expectations of how victims ought to behave; this, too, has discouraged transparent testimony with consequences for research on conflict-related sexual violence (CRSV) (Stallone 2023). In general, social scientists’ knowledge of CRSV remains limited, despite recent gains, specifically because of how ongoing stigma – especially of men and boys, sexual and gender minorities, and combatant victims and survivors – shapes data collection processes (Nordås and Cohen 2021: 204). This stigma affects the questions interviewers ask and the narratives survivors present (or omit) in oral testimonies, which, too, begs the question of what to do if those details emerge in other testimonies or other archival sources such as written records or photographs (see below). How do we weigh the importance of material to its originators versus its importance to researchers? (Bastian 2021: 34).

Each of these issues points to the uneasy and uneven dynamic between the rights of individuals and the rights of researchers, and each has practical methodological consequences for historical social science. How testimony is collected and by whom shapes what people are or are not willing to share, particularly, but not exclusively, in contexts of war and violence or contexts in which war and violence have ended, but their legacies continue to manifest in ways that impact archival data production. Contemporary politics may also shape how stories are told and what people choose to share or not share, as well as what questions oral historians ask, or archivists think to conceal. Of course, calling for IRB anonymity standards for private individual data is a potential solution to the issue (and we note that despite oral history’s similarities to qualitative interviewing in the social sciences, most oral history, as with historical research more broadly, is exempt from IRB requirements (Federal Register 2017), but choosing to anonymize introduces new challenges. ¹⁰ As Einwohner (2011) explains in her work with Jewish Holocaust survivors, for many of them, the purpose of sharing details of their experiences with violence was specifically to bear witness. Were a researcher to erase their names as well as the specifics of their testimonies – names of family members and what they know of how they died, information on neighbors who helped or harmed them, details about Nazi perpetrators, collaborators, and their actions, and so on – this could be experienced as an additional act of violence.

Likewise, Einwohner writes, anonymizing data by assigning numbers to Holocaust testimonies for coding purposes has “eerie similarities” to the Nazis’ dehumanizing tactics of tattooing Jews with serial numbers at Auschwitz, and it could be experienced as traumatic for survivors and their descendants (ibid.: 422). This differs from the oral history of Romani survivors of the same genocide, many of whom had their testimonies gathered under duress and who would, in some cases, prefer anonymization, either of themselves or of the information provided in their interviews. We note, too, that within both victim groups, there is variation in people’s preferences for what should be done with their testimonies, and this variation increases once we consider the kind of material a scholar is using.

A productive consequence is that we can compare the difference between the oral histories of Romani and Jewish survivors of the Holocaust taken with or without their consent and written traces in former Nazi archives of their experiences with violence. The International Tracing Service (ITS), for example, which contains over 30 million concentration and labor camp documents as well as displaced persons files and post-war testimony, has made its materials publicly searchable on the internet since 2019. ¹¹ Alfred L.’s testimony can now be found online, with his full name as well as his story and the names and details he shares in it – something he surely did not envision that day he was brought by patrol car to the Hannover police station (and which, in turn, has shaped our decision to follow Joskowicz’s lead in providing only the first letter of his surname, p. 18).

Meanwhile, the first author of this article, whose grandmother and great aunt provided oral testimony in 1997 for the USC Shoah Foundation Institute for Visual History and Education, has also found ITS records of the transport lists they were on, their concentration camp registration cards and numbers, displaced persons documents that included their full names, birthdays, marital status, parents’ names, and health information, photographs taken of them during this time period for recording purposes, and so on. While she can imagine them wanting their testimonies shared with a broad audience (thus shaping her decision elsewhere to quote them and their names in full; c.f. Luft 2020b; Subotić 2019 is another such example), she can also imagine that, were they alive, it would be traumatic for them to see some of the documents compiled on them by Nazis published in an academic article or book. The difference, of course, comes down to consent: they provided the oral testimonies willingly; they did not consent to being numbered and listed in Nazi transport and concentration camp records. There is no simple solution to the dilemma of considering harms and benefits in historical social science, but this should not stop academics from pausing, reflecting on, and justifying their choices when working with individuals’ private documents in their research.

Considering contemporary conflicts

The current Russian invasion of Ukraine has brought issues of archives and research ethics into sharp relief. In June 2023, for example, the Russian government announced the release of a trove of newly declassified documents pertaining to the Holocaust, purporting to demonstrate that the local populations of Estonia, Lithuania, and Latvia actively and independently carried out anti-Jewish pogroms in 1941 without participation or encouragement from occupying Nazi forces (Jerusalem Post Staff 2023). Local participation in the Holocaust in the Baltics has been a major source of political as well as scholarly contention (e.g., Kopstein and Wittenberg 2018; Mishkin 2023; Subotić 2019). Russia has always insisted on the Baltic populations’ complicity in the Holocaust (and, more broadly, sympathies with the Nazi occupation), while Baltic scholars and politicians have insisted that any local participation was coerced and orchestrated by the Germans. These archival documents (most of which, in fact, are not new as the Russian government claimed, but have simply been re-released with renewed media attention) are very important as they do, in fact, provide further evidence that local participation in pogroms against Jews was often independent from Nazi orders. At the same time, the documents’ release in the midst of Russia’s invasion of Ukraine serves a direct political purpose for the Russian military project: to discredit contemporary Baltic states as NATO members and Ukraine’s allies and to further advance the Russian narrative about the invasion of Ukraine being a project of “denazification” (Kuzio 2023).

Meanwhile, archivists and archive scholars in the region have established digital archive spaces and networks to facilitate the collection and preservation of war-endangered Ukrainian materials. In addition, they have made digital spaces available for Ukrainian refugees to submit their artifacts for preservation and classification. These projects have enabled Ukrainians to document their plight, particularly as it pertains to the bureaucratic nightmares of crossing borders and seeking asylum while under invasion. But they have not done so equally because of unequal border crossing policies and procedures, which already impact what documents are available and will be accessible to future researchers. For example, the massive refugee influx in the wake of Russia’s invasion led authorities to construct simplified border-crossing procedures en route to Poland and Hungary that prioritized white Ukrainian women and children while informally discriminating against Black, Asian, and Romani refugees, Ukrainian and foreign alike (Wiśniewska-Drewniak et al. 2023: 263; see also Busari et al. 2022; Lowry 2019b). People with disabilities similarly confronted outsized difficulties leaving Ukraine, and most Ukrainian men between the ages of 18 and 60 have not been allowed to leave at all due to martial law imposed in February 2022. Despite numerous memory initiatives seeking to broaden collection efforts and record oral histories from a cross-section of the Ukrainian population remaining in Ukraine, the overwhelming majority of documentation that exists is documentation from white adult women.

Added to this, we can already see how the trauma of war and violence is shaping the materials Ukrainians who are and have been able to contribute to archives have chosen to submit. Ukrainian refugees in Poland who responded to a survey collected by archival studies scholars (Wiśniewska-Drewniak et al. 2023, for example), emphasized that upon fleeing their homes, they chose to prioritize taking with them legal identity documents, proof of professional qualifications, evidence of property ownership, and COVID certificates, not personal or familial artifacts. Simultaneously, they often felt as if recording (and sometimes sharing on the internet) their wartime experiences with cellular telephones was far too emotionally difficult for them, felt exploitative – especially once they had already seen traumatic photos, e.g., of dead children, circulating on social media – and led them to avoid online communities where others were sharing images and videos of the war. This, Wiśniewska-Drewniak et al. note, further raises the question of “whose experiences are being preserved” when technology access, proficiency, and comfort with some kinds of evidence rather than others are shaped by gender, age, and economic status (ibid.: 260–61). Ukrainian respondents to their survey recognized the importance of having and sharing information to raise awareness of the horrors of war in general, but they themselves sought to avoid producing and consuming this information. The authors in turn emphasize that research centered on social media, mobile phones, and visual ethnographies should be especially careful to employ trauma-informed methods in their work (262).

Finally, while scholarly investigations of the Russian invasion of Ukraine illustrate how war influences how and which victims contribute to archival data and collection, conflict and violence also influence the documentation practices and possibilities for researchers and journalists operating concurrently in the field. As the brief summaries below from several cases – Syria, Iraq, South Sudan, and Gaza – demonstrate, documenting the human consequences of violence presents immense difficulties.

For example, it is often easier to gather evidence from urban rather than rural contexts of a conflict. This is due, in part, to the fact that urban violence is frequently more visible and accessible to outsiders (Dawkins 2021). This limitation is further compounded by the fact that media organizations increasingly depend on “reporters who ‘parachute’ in,” which increases the likelihood that they will only visit conveniently located areas when they travel to document war (Parkinson 2023: 2, building on Arjomand 2022). Additionally, it is often safer to visit urban than rural areas, although this, of course, depends on the circumstances of the conflict at the time; for instance, it was difficult for outsiders to enter Homs, Syria, during the 2011–2014 siege, and for witnesses to recount their accounts with violence in turn (Price et al. 2015: 18).

It is also typically easier to document killings by and of soldiers and men than the elderly, women, and children. Documenting civilian killings, whether unintentional or deliberate, is also more challenging than documenting conventional war and its consequences. This is due in part to the relationships that researchers and journalists form with state agents to gain access to the sites of a conflict: outsiders frequently face the risk of having their special access revoked, or even threats to their lives and safety, if they expose serious human rights violations (Parkinson 2023: 3). It is also, for many reasons, difficult to distinguish between combatant and civilian deaths, which affects casualty counts, as well (Kinsella 2015). In situations involving potential under- or over-reporting of combatant or civilian casualties “for tactical and reputational reasons” (Lynch and Parkinson 2023: 3), demographic information, such as the victim’s gender, may be used by researchers and journalists as a proxy. This, we note, alongside Lynch and Parkinson (2023), is an imprecise and troubling assumption with life-or-death consequences: When men, even when they are civilians, are presumed to be combatants, they become “acceptable” targets of violence; we should avoid incorporating such assumptions and the mischaracterization of men as combatants because they are men into our data.

Documentary practices are also influenced by the type of violence that is occurring. Massacres are frequently easier to document than gassings or mass burnings. While censorship may be implemented (or attempted) in all cases, severe disfigurement from the latter complicates the determination of victims’ demographic characteristics and casualty counts, particularly in the immediate aftermath of such horrors.

Violent events with larger numbers of victims are more likely to be reported by more sources; however, in situations where there are few or no remaining witnesses, researchers and journalists must depend on estimates (Price et al. 2015: 19).

Finally, accurate evidence of violence is more difficult to collect during periods of contested control, as violence is likely to escalate during such times (Kalyvas 2006). This makes information more difficult for outsiders to obtain. As territorial control varies throughout a conflict, the types of access that various parties are prepared to grant journalists and researchers also changes, impacting recording patterns, as well.

These are merely a few examples of the biases that can affect violence documentation dynamics. The site of violence, victim and perpetrator characteristics, kinds of violence, and features of a violent event such as victim counts or local territorial dynamics, shape what scholars can know about a conflict as it is occurring and years and decades later, when they probe these data in the archives.

Conclusion

This article builds on the work of scholarly peers to raise three ethical issues in historical social science and their implications for research methods and analyses. Drawing on a range of cases, we show how archival research is never neutral. Rather, it is always infused with the politics of the past, which shaped how documents were constructed, compiled, stored, and shared; the politics of the present, which shapes how documents are interpreted; and the politics of the future, which shapes how documents may be used or misused for different purposes, causing harm or providing benefits to different populations. In research on violence and conflict, because the practice of collecting, preserving, publishing, or displaying archival collections is directly and inextricably linked with histories of violence, it is urgent that scholars carefully weigh and consider appropriate ethical protocols.

Our own work and experience inspired us to focus on archival research on violence. However, we believe our broader argument is relevant to all archives. There are issues of past politics, interpretation, and harms and benefits embedded in archival research that relies on, for example, university archives (e.g., issues of power, inclusion, and exclusion), newspaper archives (e.g., representation, censorship, and ownership), or corporate archives (e.g., bias, access, secrecy, and inequality). Our observations are also relevant to state and religious institution archives and historical societies, as well as to special collections such as personal papers, small and local organizations’ collections of archives, and so on. We mention oral history projects and archives built on testimonies throughout this essay, too, though future work on similarities and differences is needed.

Finally, we do not expect this article to cover all the ethical issues involved in archival research or to provide all the answers for how to conduct ethical work on past violence. Instead, it is our intention that this article will serve as an invitation for social scientists to critically consider how they conduct historical research. By attending to issues of politics, interpretation, and harms and benefits, we suggest scholars should strive to account for the origins of the materials in their archives and their data selection and collection processes; they should offer rationales for their interpretive choices where other, possible interpretations might exist; and they should explain how considerations of harms and benefits impacted their research and writing. Addressing these issues will bolster historical research methods’ validity, reliability, and transparency, and, of course, the ethical consideration of the subjects of our research.

Therefore, ethical consideration should be more central to archival research than it is currently. At every stage of a project, from identifying archives, determining their provenance, and historicizing their contemporary locations, to collecting data, examining documents, writing findings, and ultimately publication and dissemination, scholars must be able to consider, make, and defend their decisions. Research ethics in archival methods should not be an afterthought but rather a central component of scholarly training and practice.