Being Betrayed by Your Own Side Causes Far More Trauma Than Being Attacked by the Enemy

There is something the standard model of wartime trauma gets badly wrong. For decades, researchers studying the psychological aftermath of armed conflict have focused almost entirely on violence between opposing sides: ethnic groups, armies, factions with clear, legible lines of enmity. This made a sort of intuitive sense. You are attacked by the enemy; you suffer; you heal, perhaps, or you don’t. But it turns out the source of the violence matters enormously, in ways that the field has largely ignored.

A new study from postwar Sri Lanka suggests that being harmed by members of your own community produces trauma that is, by measurable clinical standards, roughly three to five times worse than the harm inflicted by an ethnic outgroup. Worse, in one of the more unsettling asymmetries in psychological research, only one of these two types of violence opens a pathway to psychological recovery and growth.

The setting is Jaffna, a district in northern Sri Lanka and the historic centre of the country’s Tamil community. Sri Lanka’s civil war lasted 26 years, ending in 2009, and it was brutal in ways that don’t fit neatly into a simple narrative of ethnic conflict. Yes, there was interethnic violence: Sinhalese government forces fought Tamil separatists in a war that killed tens of thousands of civilians. But within the Tamil community, the Liberation Tigers of Tamil Eelam, the dominant rebel group, also turned violence inward, against Tamil civilians they deemed insufficiently loyal, rival factions, forcibly conscripted fighters, and families who resisted.

Joan Barcelo and Keshana Ratnasingham, researchers at New York University Abu Dhabi, surveyed 628 Tamil adults in Jaffna in 2022, thirteen years after the war’s end. Respondents were asked about 35 types of war-related experiences, rating both the intensity of each event and, crucially, who was responsible: Sinhalese forces (intergroup violence) or Tamil actors, including the LTTE (intragroup violence). Using validated clinical instruments, the researchers then measured two distinct psychological outcomes. One was post-traumatic stress disorder (PTSD), the familiar constellation of intrusive memories, hypervigilance, and emotional numbing. The other was something less well known: posttraumatic growth, the positive psychological changes that can, somewhat counterintuitively, emerge from surviving serious trauma, things like a stronger sense of personal resilience or a deepened appreciation for life.

The PTSD findings were stark. Nearly 43% of respondents in contemporary Jaffna met the threshold for probable PTSD, more than a decade after the fighting stopped. That is, on its own, a sobering number.

What distinguished the two types of violence was not just their scale but their psychological signature. Each additional traumatic event linked to Sinhalese forces increased PTSD scores by roughly 0.1 units on the clinical scale. Each event linked to Tamil perpetrators increased those scores by between 0.36 and 0.43 units. Across ten events, that is about one additional PTSD unit for intergroup violence versus three to four for intragroup. The gap held across different statistical models and after controlling for age, gender, wealth, and education. Thirteen years of peacetime made no real difference to the disparity.

Barcelo and Ratnasingham propose a mechanism that is, perhaps, almost obvious once stated. Intergroup violence, though genuinely traumatic, allows something psychologists call external attribution: the blame can be located squarely outside yourself and your community. The enemy did this. That cognitive clarity, grim as it is, provides scaffolding for recovery. It leaves certain psychological resources intact even as others are destroyed.

Intragroup violence offers no such scaffolding. When the harm comes from your own community, from people whose safety was inseparable from your own, the psychological consequences are of a qualitatively different order. Betrayal is involved, and what the researchers describe as identity dissonance: a fracturing of the categories through which you understood your world. The social networks that would otherwise buffer trauma are precisely the ones that have been weaponized. Meaning-making, the slow cognitive work by which survivors build a coherent story about what happened, becomes very difficult when the perpetrators share your ethnicity, your language, in some cases your neighbourhood.

The posttraumatic growth data add another dimension. About 38% of respondents showed moderate-to-high posttraumatic growth, which sounds encouraging until you look at where it came from. Intergroup violence was positively associated with growth; intragroup violence showed no such association. And in cases where respondents had been exposed to both types (roughly 11% of the sample), the presence of intergroup violence seemed to buffer the worst effects of intragroup trauma, perhaps because it gave survivors an external locus of blame even when the closer wound was the deeper one. PTSD did not worsen when both types of violence were present together.

There are real limits to what this study can claim. Its cross-sectional design means it captures a snapshot rather than a trajectory; there is no way to watch PTSD and growth evolve over time in the same individuals, or to determine cause and effect with certainty. Self-reported trauma exposure, thirteen years on, carries the usual risks of recall bias. And Jaffna is a specific place with a specific history, one that may not translate cleanly to other postconflict settings, even those with similarly complex internal dynamics.

Still, the broader pattern the study points to is probably not unique to Sri Lanka. Intragroup violence is, in a sense, more common in civil conflict than is generally appreciated: researchers who study civil wars estimate that something like 45% of conflicts since the end of the Cold War have involved significant infighting within rebel movements or affected communities. That is an enormous share of the global burden of political violence, and one that existing mental health frameworks were not really designed to address.

For clinicians and policymakers designing post-conflict mental health interventions, the distinction matters practically. Current screening and treatment protocols focus on overall trauma exposure, and do not routinely ask who inflicted the harm. Barcelo and Ratnasingham argue that this is an oversight with real consequences. Survivors of intragroup violence may need interventions focused less on individual symptom reduction and more on community reintegration, on rebuilding exactly the social trust that was destroyed. Standard trauma therapy models, developed largely in the context of intergroup conflict, may be poorly suited to what is, at its core, a problem of shattered solidarity.

None of this is easy to fix. Trust, once weaponized, does not simply reconstitute itself because a war has ended. But understanding that the identity of the perpetrator shapes the nature of the wound, not merely its severity, is probably a necessary starting point for thinking about how recovery might actually work.

DOI / Source: https://doi.org/10.1093/pnasnexus/pgag058