Dwelling With the Letters: Methodological and Epistemological Dilemmas in a Qualitative Analysis of the Correspondence of Jewish Internees From the Drancy Camp

Conceptual Framework

QDA is a systematic procedure for analyzing documents as a source of qualitative data (Bowen, 2009). It is often used alongside other qualitative methods as a form of triangulation, but it can also function as a stand-alone methodological approach (Furlong et al., 2024; Moilanen et al., 2022; Wild et al., 2010). Documents may be examined descriptively or interpretively and can be approached through a wide range of analytic orientations: hermeneutic, discursive, narrative, anthropological, historical or ethnographic (Flick & Coffey, 2014; Prior, 2003). In historical research, close analysis of written documents is common.

A particularly rich body of historiographical scholarship has developed around letters written in wartime contexts (Dauphin, 2002; Debruyne & Van Ypersele, 2011; Doetzer, 2002; Gilbert, 2022; Halberstam, 1995; Hurtubise, 1994; Klacsmann, 2017; Lamprecht, 2001; Lorenz, 1998; Sabbagh, 2005; Saltiel, 2017; Vidal-Naquet, 2014), playing a key role in establishing correspondence as a historical source in its own right. More than a simple individual narrative, a letter bears the imprint of the context in which it is written and emerges as a space in which multiple dimensions of existence intersect, at the crossroads of the social and the intimate (Dauphin, 2002; Hurtubise, 1994; Vidal-Naquet, 2014). These personal narratives are historical materials, they do more than record events, they illuminate the ways in which individuals construct meaning around their existence; In the context of the Shoah, they acquire particular significance, insofar as they provide spaces in which subjective identity and institutional frameworks shaping individual trajectories become intertwined (Balint, 2023; Knight et al., 2025).

The methodological literature on documentary research remains relatively limited (Tight, 2019) and only a few QDA studies have used a phenomenological orientation and always through a combination of descriptive and hermeneutic dimensions “aimed at construing the meaning of the document: both its surface and underlying meanings that the document avails” (Armstrong, 2021). Hermeneutic approaches that seek to interpret how participants make sense of their lived experience were not well suited to this corpus and its historically sensitive context. Analysis of documents produced in settings of extreme violence or oppression exposes researchers to intense dilemmas explored in scholarship on “difficult topics” (Silverio et al., 2022) and ethical qualitative research in sensitive contexts (Laimou et al., 2026b; Stevens, 2013).

These letters appeared amenable to our IPSE inspired collective descriptive phenomenological approach that might allow us to account for their content without betraying the subjectivity of their authors. IPSE was designed for long, co-constructed interviews and not for brief, fragmentary documents. We complemented it with Ethnographic Document Analysis (EDA) (Trundle & Phillips, 2025), whose emphasis on “dwelling with documents” (reading them slowly, repeatedly, attending to their silences, absences, tonalities and materialities) and attending to the material and contextual constraints of texts allowed us to adapt the IPSE analytic posture to short, censored letters written under extreme conditions. EDA invites researchers to consider documents as social and affective sites rather than inert texts. It was impossible not to read the letters through a constant double movement: emic (what the letter explicitly gives to read, the letter as subjective material) and etic (what history compels us to know, the letter as testimony to atrocity).

A Qualitative Study Inspired by the IPSE Approach

This study was conducted from November 2022 to April 2024 and was approved by the Ethics Committee of the Université de Picardie Jules Verne (number: 2022-12-1/2023-21). The report of this study complies with the JARS-Qual guidelines (Levitt et al., 2018). Our approach draws on EDA and the collective group analysis and a descriptive phenomenological orientation IPSE method (Sibeoni et al., 2020).

Our research group consisted of four expert qualitative researchers (ARL, EM, JM, JS), a psychologist-psychoanalyst and principal investigator of the LettresCamps project (DL) and a historian (AV). Data analysis was conducted by individuals of different genders (one man, three women), ages and personal backgrounds, each holding a specific and situated relationship to this period of history. To support the reflexive dimension of the analysis, the qualitative researchers invited another researcher from their laboratory (LV), who had no access to the data, to participate in selected meetings.

The historians of LettresCamps team compiled all available historical information regarding the context in which the correspondence was produced. To preserve an inductive stance and to dwell with the letters as subjective material, the qualitative researchers deliberately chose not to access the detailed historical context of the Drancy camp until after the analytic process was completed. This decision reflects a methodological commitment grounded in phenomenological inquiry and in the IPSE ethos: to begin with what the material gives, to bracket pre-existing assumptions, and to allow categories and meanings to emerge from the documents themselves rather than from prior historical knowledge. This intentional postponement of contextual information protected the analysis from being prematurely shaped by established historical narratives and enabled the researchers to attend carefully to the letters’ linguistic, emotional, and relational textures.

Data Collection

Access to the letters written by Jews interned in the Drancy camp was made possible through a partnership with the Mémorial de la Shoah in Paris, whose mission includes the preservation and transmission of archival materials (Mémorial de la Shoah, 2025). The letters form part of the CMLXXXVI collection, which brings together correspondence entrusted to the Mémorial over several decades by the families and relatives of internees and deportees. The team of historians of the project classified the letters according to an initial typology (clandestine, official or uncertain), dated them as precisely as possible, digitized them, and transcribed them into a documentary database (Heurist) developed for the project (Laimou et al., 2025, 2026a, 2026b). The entire collection is estimated to comprise several thousand items. For the purposes of the research, the digitized and transcribed letters were compiled into an Excel file organized chronologically and made accessible to the qualitative research team (Laimou et al., 2025, 2026a, 2026b). A preliminary corpus of 300 letters was then selected according to a principle of temporal and situational variability to capture different communicative conditions.

None of the letters were written under conditions of free expression. Each text is shaped by constraint, censure, fear, and strategies of protection toward loved ones. As a result, the subjectivity expressed in these documents is not a native or spontaneous subjectivity but one profoundly transformed and often wounded, by the conditions under which it was produced.

Our qualitative research group received the 300 letters in stages. In November 2022, we first obtained the letters from 1941 (all written by men) and provided an intermediate analysis of this subcorpus at the request of the wider team. The letters from 1942, 1943 and 1944 were shared with us in January 2023.

The letter extracts reproduced in this article are presented in their original French in the appendices and were transcribed by the team of historians of the project. To maintain authenticity, they have been reproduced exactly as written with no attempt to improve readability. The English translations adhere closely to the vocabulary, grammar, syntax and punctuation of the originals, although spelling mistakes have not been replicated. These translations were produced specifically for this article and are not intended as official translations of the corpus (Laimou et al., 2025, 2026a, 2026b).

Data Analysis

The analytic process unfolded in two major phases: an individual descriptive phase and a collective group-based phase, both deeply shaped by the sensitive nature of the letters.

Reflexivity

Working with Shoah-related documents, and particularly with letters written from a space of internment, required a deliberate and sustained reflexive practice at both individual and collective levels. Reflexivity was not simply an ethical expectation but became a methodological requirement to “secure credibility, trustworthiness, and non-exploitative research through self-scrutinization of the lens through which the researcher views the phenomenon studied” (Berger, 2013). It was also essential to ensure the psychological safety of the researchers and to prevent the emotional strain inherent in research on highly sensitive topics.

The Findings

We present here a reduced synthesis of the structure of experience produced through our IPSE-inspired analysis. The full version of the results, including all categories and the complete set of quotations, is available in the Supplemental Material 1.

The Trajectory

The IPSE-inspired qualitative analysis did not yield stable, robust, and consensual categories in the way it typically does when applied to interview data. Instead, it produced a trajectory marked by tensions, dead ends, and successive readjustments.

The Unbearable Nature of Description

The earliest letters (1941) intensified the rupture. They contain no explicit trace of the violence, deprivation, or internment conditions that we, as analysts, know would later unfold, realities that, in 1941, the authors themselves could not yet fully anticipate. Instead, they speak of food, clothing, parcels, health, and minor concerns, an unbearable normality within a death-bound environment. As a researcher wrote: “I read a request for socks and I heard something else. I cannot stay inside the text without invoking History.” (Research diary). This was our first methodological limit: a literal, descriptive phenomenological reading risks unintentionally banalizing extreme historical violence.

IPSE is anchored in a constructivist epistemology, assuming that reality is co-constructed between material and researcher; that data emerges from dialogical, situated exchanges; that meaning is grounded in first-person explicit accounts (Denzin & Lincoln, 2011; Sibeoni et al., 2020). But in this corpus, the problem of what is not written becomes epistemologically central. This directly resonates with Morgan’s (2022) discussion of QDA as an approach constrained by the partial and incomplete nature of documentary data, which requires researchers to adapt their analytic questions to what documents can provide. The letters’ inexplicable omissions (lack of explicit references to violence and fear, and only few about hunger, or Jewish identity) are not simply absences: they are structural silences, imposed by the camp regime and the conditions of terror.

As the coding began, this epistemic divide hardened. The descriptive categories we generated felt both too literal and too blind to the historical context. We faced a contradiction: if we coded literally, we risked naïve description and the ethical danger of banalizing extreme suffering; if we contextualized first, we left the descriptive phenomenological method and entered hermeneutics or history.

To justify our descriptive results, we drifted into counting occurrences, an approach explicitly incompatible with IPSE. We counted: mentions of “God”, uses of the word “Jew”, occurrences of “Yiddish” references to food and parcels. This drift was defensive. It emerged during the interdisciplinary discussions that the first letters we analyzed contained no explicit mention of violence, of the harsh conditions of arrest and internment documented elsewhere, nor of Jewish identity. This prompted us to clarify at the beginning of our results how “being Jewish” appeared in the corpus. These exchanges highlighted the gap between the historical reality of persecution and what the letters themselves were able -or willing- to express under conditions of censorship, fear, and surveillance.

The arrival of the 1942–1943 letters (women, children, deportations) intensified the discomfort. These letters contained clearer signs of hunger, fear, and separation, and even then if explicit references to the horror were present, they remained limited, indirect, or markedly restrained. The phenomenon to describe is not in the letters alone, but in the abyss between the letters and History. Only describing the letters is unbearable because the letters themselves make description impossible. They resist interpretation yet demand context.

Toward a Dialogical and Reception-Oriented Reading

If the letters were marked by constraint, fear, and the looming threat of deportation, they were equally traversed by another recurrent gesture: an unwavering desire to protect the addressee, to reassure, to maintain a fragile thread of intimacy across separation. This led us to a decisive epistemological shift: from phenomenological description and abductive testimony work to a dialogical, reception-based approach, positioning ourselves (consciously and reflexively) as Model Readers.

In contrast with interviews, the letters were never addressed to researchers. Their meaning is inseparable from the presence-in-absence of those to whom they were written. This insight aligns with Jauss’ aesthetics of reception (1982), which posits that the meaning of a text emerges through the anticipated horizon of its reader. It also echoes Iser’s “implicit reader” (1993), Genette’s narratological work (1980), Derrida’s Post Card (1980/1987) and Bakhtin’s dialogism (1975/1982), where every utterance is fundamentally relational, anticipating response and shaped by the presence of another consciousness. Under this lens, the letters cease to be mere descriptions or testimonies. They become dialogical gestures, forms of reaching out, of holding on, of sustaining relational life within a system designed to annihilate it.

Umberto Eco’s Lector in Fabula (1985) offered a decisive turning point. Eco argues that every text constructs a Model Reader whose encyclopedia (the sum of their cultural knowledge, memories, prior texts) is mobilized to complete the text’s gaps. The letters, laden with censorship and survival constraints, are quintessential open texts full of silences, ellipses, and implicit meanings. Our reflexive work therefore led us to adopt the position of Model Readers, not to pretend that we were the intended recipients, but to recognize that these letters were meant for someone, to acknowledge that their meaning is inseparable from that fact, to accept that our own encyclopedia is inevitably activated in reading them, and to reintroduce a form of co-construction central to IPSE but disrupted by the archival nature of the material.

Positioning ourselves as Model Readers also activated an ethic of reading. Levinas’ philosophy of the encounter (Dougan, 2016) with the other suggests that any face-to-face (even a deferred or textual one) imposes an ethical responsibility: to recognize the otherness of the other without absorbing it into one’s own categories. Reading the letters dialogically meant resisting the temptation to appropriate or overinterpret but also acknowledging their address as a gesture of care. In this sense, the dialogical axis is not only methodological but ethical: an attempt to honor the relational intention of the letter-writers.

This shift required a practical change: transforming the material from chronological Excel tables into correspondence-based groupings. Grouping letters by sender and recipient reintroduced narrative continuity, personal identity, relational history, and a sense of voice. This reorganization also helped counteract the disembodied feel of the initial dataset and supported the reconstruction of the addressee’s presence, essential for dialogical analysis.

The dialogical stance offered an unexpected emotional buffer. The researchers found a place where compassion and analysis could coexist and yet this proximity also intensified emotional reactions. This oscillation between protection and vulnerability is consistent with literature on analyzing sensitive or traumatic archives (Silverio et al., 2022).

By shifting focus from “what happened” to “what meaning was addressed,” we honor both subjectivity and historical truth. In line with Levitt et al. (2018), documenting reflexive tensions, positionality, and epistemic shifts becomes itself a methodological contribution. Dialogism offers a way to analyze traumatic archives without collapsing into reductionism or denial. Two limitations remain: first we lack the letters sent to the internees, making the dialogical reconstruction partial and second, the Model Reader stance risks reintroducing researcher affect and memory. Triangulation with historians and systematic reflexive journaling were needed to limit this.

The Impossible We

IPSE rests on the assumption that the analytic group forms a stable we capable of jointly interpreting, reducing, and structuring experience. But under the weight of the Drancy letters, that we fractured. The group dissolved not because of dysfunction but because the material itself made co-analysis impossible. The letters did not simply demand reading; they demanded bearing. And bearing, we discovered, is unevenly distributed.

Some researchers withdrew for a while. Others read but could not analyze. Others analyzed but only in solitude. The analytic collective progressively lost its coherence, giving way to a plurality of destabilized subject positions. This dissolution resonates with findings in qualitative work with traumatic or violent data, where emotional overload fragments analytic communities (Silverio et al., 2022). It also echoes Ellis and Bochner’s (2000) work on emotional autoethnography, which shows that the researcher’s emotional life becomes part of the phenomenon. Yet our experience reveals something more radical: in extreme contexts, the collective is not always the optimal epistemic unit. It can become the obstacle.

IPSE’s intersubjective strength, that is its requirement for shared reading, affect and construction, became untenable. The project unfolded instead as a collective solitude, a paradoxical configuration where each researcher worked alone under the same storm. The impossibility of the we is not a limitation but a methodological result.

Them and Us: The Ethical Asymmetry

A second dilemma emerged from the ontological distance between them -the authors of the letters, persecuted and largely exterminated- and us, researchers reading their last words decades later. This asymmetry shaped every interpretive decision.

Qualitative methods often assume the possibility of encounter, even when mediated by text. But these letters refuse symmetry. They are acts of communication produced under surveillance (Morariu, 2011; Prior, 2003), emotionally protective performances (Goffman, 1959), textual fragments shaped by censorship and self-censorship (Sankofa, 2023), artefacts circulating within a system of domination (Trundle & Phillips, 2025) and testimonies written under conditions where full expression was structurally impossible (Felman & Laub, 1992). Thus, the very ontological assumption of descriptive phenomenology collapses.

The fear of producing a naïve descriptive account, one that could be misused or inadvertently feed negationist readings, was not paranoia; it aligned with warnings from Todres (1999), Gill et al. (2018) and Silverio et al. (2022) regarding the dangers of insufficient contextualization in traumatic histories. While phenomenological bracketing is a core methodological principle of descriptive phenomenological psychology (Giorgi & Giorgi, 2003), our analysis suggests that, in the context of extreme historical documents, its application may become ethically problematic. Suspending historical knowledge is impossible and dangerous. Neutrality becomes a moral impossibility (LaCapra, 2014; Levinas, 1982). Thus, unlike typical qualitative inquiry, our task was not only to remain faithful to the subjectivity of the letters; it was also to remain faithful to the dead. This asymmetry is not soluble. It is a result.

The Fractured I: Mirrored Subjectivities

The letters are written in the first person, an I that appears to offer direct access to experience. Yet this I is not the sovereign, self-present phenomenological subject assumed in qualitative inquiry. It is a wounded I, censored, constrained, fragmented, speaking under terror and often speaking to protect others rather than reveal itself. This raises fundamental questions: Where and who is the I that writes? Does it have access to the horror it is living? Is it protecting itself or its loved ones? Is it a subject of its experience, or so stunned and displaced that subjectivity becomes unstable? The I is both present and self-effacing, a fractured voice whose very fragmentation is part of the phenomenon. What is given is inseparable from what must remain unsaid. We expected to analyze subjectivity. Instead, we encountered the limit of subjectivity.

At the same time, the I of the researchers was destabilized. Shame, guilt, breathlessness and what we came to call methodological unworthiness surfaced: the sense that no analysis was adequate, no interpretation legitimate.

This produced a mirroring effect: the internees’ I is incomplete, forced into silence; the researchers’ I is destabilized, forced into reflexivity. Two subjectivities confront each other: one fractured by violence, the other fractured by responsibility. In this context, Blanchot’s (1955) reflections on subjectivity offer a way of thinking about the limits we encountered: the impossibility of fully restoring the internees’ subjectivity, and the impossibility for researchers to step outside their own.

Toward a New Epistemology for Traumatic Archives

These three contributions converge to suggest that analyzing extreme historical documents requires a multilayered epistemology, one that departs from inductive phenomenology and moves toward: critical realism, reception theory, ethical philosophy, trauma studies, and document theory, as discussed earlier in the manuscript (e.g., Felman & Laub, 1992; Gill et al., 2018; Prior, 2003).

Phenomenology relies on analyzing what the phenomenon gives. Here, what is given is structurally incomplete. These letters cannot be read as full narratives; they are strategic communications, acts of survival, testimonies through silence. Thus, a new epistemic model is required: a three-level analytic stance combining descriptive, testimonial and dialogical/reception layers, grounded in reflexivity and historical responsibility. This is not a generalizable model. It is a model for materials that wound.