Considering the Use of Language in Protection Risk Analysis

Humanitarians often recognize the importance of language in planning and implementing responses but tend to confine their considerations to issues of protection mainstreaming—access, participation, accountability, and safe and dignified service provision—rather than understanding how their use of language can further protection outcomes. CLEAR Global has done extensive research into the protection mainstreaming-related challenges humanitarians face in missing the critical importance of language usage, to the detriment of the efficacy of and safety and inclusion in their programs.

This case example was written with input from CLEAR Global

However, the work of InterAction and partners in Somalia and Colombia further demonstrates how a broader conceptualization of language usage could enhance more effective strategies toward the implementation of the Centrality of Protection across all sectors. Language usage plays a profound role in defining and understanding protection risks, helping to reveal perspectives unique to each community.

Accessibility: A Foundation for Language Considerations in Safe and Inclusive Programming

It is a widely acknowledged and decried fact that humanitarians’ over-reliance on global languages of power or perceived local lingua franca can hinder program accessibility and efficacy. CLEAR Global, an NGO helping people get vital information, and be heard, whatever language they speak, conducted research in Cox’s Bazar which found instances of mistranslation with significant implications for service seeking, such as “violence against women” being rendered as “violent women.” In a similar study in northeast Nigeria, “safe space” was understood as “a space protected by guards.” Such errors can obscure service functions and deter engagement, particularly for marginalized language speakers or those with hearing or visual impairments.

Guidance such as the Core Humanitarian Standards (CHS) Commitment 1 emphasizes the importance of communicating in languages and formats that are accessible, understandable, respectful, and contextually appropriate. While these practices are vital, they often remain confined to planning stages—limited to creating Information, Education, and Communication (IEC) materials or recruiting community mobilizers—and focus on reducing barriers to access.

Participation and Accountability

Language also plays a crucial role in enabling participation and
accountability. Often, these considerations bookend a project, gathering data on the needs of affected persons prior to project design and then considering the impact and effect of implementation through complaints and feedback mechanisms. Effective community engagement for participation and accountability depends on affected persons understanding survey questions and knowing how to provide feedback or report issues.

However, CLEAR Global’s study in northeast Nigeria revealed that many respondents understood as little as 35% of survey terms due to reliance on multiple translation stages. In Nigeria and the Democratic Republic of Congo (DRC), language barriers prevented individuals from reporting Sexual Exploitation, Abuse, and Harassment (SEAH) through hotlines operating only in dominant languages. For example, a woman in Gwoza, Nigeria, recalled dropping a hotline call because the operator spoke only English.

Here, again, the humanitarian sector has well-defined, if too often ignored, guidance in the form of the Global Protection Cluster’s Protection Mainstreaming Toolkit. The toolkit calls for facilitators of protection risk assessments to be fluent in the local language and to communicate in a way that the objectives of the assessment or data collection are clearly understood and unrealistic expectations are not created. The toolkit’s guidance squarely places consideration of language as enabling full participation in data collection.

Safety and Dignity

Language usage, particularly overreliance global languages of power, can reinforce or challenge power dynamics. This is an issue, for example, when humanitarians are mostly recruited among speakers of global languages of power, while crisis-affected people speak marginalized languages. The result can be reduced access to lifesaving information and services, aggravating existing vulnerabilities. A CLEAR Global study in eastern DRC found that aid workers and health care providers often spoke French or Lingala, languages associated with government forces perceived as oppressive or hostile by communities that spoke Congolese Swahili or Nande. Thus, even using the lingua franca in a large part of DRC, but not the eastern part of the country —Lingala—undermined trust in aid agencies and fostered suspicion. Not understanding service providers is also a source of shame for minority language speakers in Somalia, where CLEAR Global found speakers of the Maay Somali dialect said they would feel embarrassed discussing their needs and concerns with predominantly Maahatiri-speaking humanitarians and would rather remain silent for fear of discrimination.

Protection case management best practices offer some guidance that can be used outside of that specific response area, such as emphasizing jargon free and nonjudgmental communication. Reflective listening techniques, drawn from case management, should be fully integrated into data collection to ensure that language choices enhance, rather than hinder, disclosure. However, such approaches often focus narrowly on creating trust in service provision rather than leveraging language to advance broader protection goals.

To effectively achieve the Centrality of Protection (CoP), humanitarians should take a nuanced view of the language they use, going beyond safe and inclusive programming principles. Results-Based Protection (RBP) presents a framework for humanitarians to operationalize the Centrality of Protection to undertake risk-based—as opposed to strictly need-based—assistance to achieve a protection outcome. RBP supports integrating the consideration of language into continuous, context-specific protection analyses. This approach prioritizes community-led risk identification, ensuring that linguistic nuances inform a deeper understanding of threats, vulnerabilities, and coping capacities. Language is more than a tool for communication—it shapes how communities define and understand risks. InterAction is working with several operational humanitarian partners in Somalia and Colombia to respond to conflict-induced food insecurity using RBP approaches. In both locations, InterAction and partners undertook qualitative, iterative protection context analyses to inform risk-based programming. The project teams conducted participatory analysis to understand community perceptions of protection risks that either led directly to or resulted from food insecurity and conflict.

2. Data Analysis: Processing What Communities Say Through Humanitarian Minds

Just as communities’ lived experiences impact how they understand the verbiage used in data collection—that “protection” might not include an array of risks that do not result in displacement—so too do humanitarians often fail to accurately process the nuances captured in qualitative data. Communities often communicate risks through euphemisms or culturally specific terms that do not align with rigid analytical frameworks and taxonomies. As InterAction’s partners in Somalia sought to gather data on conflict-induced food insecurity, they found communities in Jowhar were reluctant to even name Al-Shabaab (A.S.) directly, instead using an array of euphemisms that each carried its own significance for how the threat actor was perceived: Arsenal, as in the Premier League football team, mostly used by young people; Ahalinyaradda, meaning “the youth”; Odayaashii, meaning “the elderly” and carrying an association with power and authority; and Gaagaabka, meaning “short men” and carrying the sense that A.S. can infiltrate any location or group. Some men refused to speak of A.S. entirely, with one focus group respondent ending not his own focus group discussion (FGD) prematurely but calling on two others being held with different enumeration teams to stop, as well. This reluctance revealed how fear of A.S. stifled open discussion—a critical finding, though initially unreported by field teams. Only through thorough debriefing was this insight captured.

Similarly, community members in Somalia employed varied translations for sexual violence, colored by different components of risk, most notably the nature of the threat actor. For respondents in Jowhar, “taken by force” referred exclusively to stranger-perpetrated assault, excluding domestic or transactional sexual violence that occurred in the workplace. The term was also never applied to men and boys who had experienced sexual violence. Men also tended to use the term “sexual harassment” or “sexual exploitation” to describe marital disputes, including over money. The fluid use of these terms, especially their different understandings between respondent genders, risked skewing data. Field teams had to disentangle these nuances to avoid erasing critical distinctions.

In another example from Somalia, this time in Abudwaq, respondents differentiated between targeted assassinations, interclan militia attacks, and assaults on civilian infrastructure, all of which were related to clan conflict. While standard UNHCR taxonomies might group these under “attacks on civilians,” communities’ linguistic distinctions highlighted unique vulnerabilities, distinct relationships to community capacities, and implications for food insecurity. Initially translating all incidents into “killings,” enumerators needed training to preserve these distinctions in their notes.

3. Training and Sensitization for Field Teams

Engaging communities in defining protection risks requires humanitarian actors to understand and adopt local terminology. This is not without challenges. Even local staff may sanitize sensitive disclosures, using generic humanitarian language that erases the granularity of community narratives.

While it was previously noted that respondents in Somalia often used “taken by force” as a euphemism for rape (khusi) under specific circumstances, some enumerators preemptively used the more oblique term to create a safer space for discussion. While this might have been effective for initial engagement, the phrasing risked obscuring nuances and implicitly directed respondents to consider only a subset of instances of sexual violence. In one location, Las’anood, field teams admitted to using even more elliptical verbiage, asking only about “violence,” to avoid their own discomfort with the subject matter.

Participatory methodologies, though promising, often resulted in emotionally charged disclosures. In Abudwaq, storytelling elicited vivid accounts of displacement, such as a woman recalling artillery fire and fleeing her home. While these narratives enriched the data, they overwhelmed some staff, necessitating extensive debriefing and data cleaning. When engaging in data collection on protection risks, it is essential to recognize that participants might relive or share traumatic experiences, even if questions are designed to minimize re-traumatization. Non-protection staff might also lack experience in receiving and responding to stories of trauma; this is a skill in and of itself, and psychological first aid (PFA) training, especially if not regularly refreshed, might be insufficient to build this skill. It is critical to ensure that both community members and field teams have access to appropriate referrals to mental health and psychosocial support services. In the cases described, debriefing and emotional support mechanisms were not systematically integrated, leading to challenges in processing difficult narratives. Future efforts should incorporate structured post-interview debriefings, peer support mechanisms, and clear referral pathways for both enumerators and respondents experiencing distress. This ensures that qualitative data collection not only maintains integrity but also safeguards the well-being of all involved. Preparing non-protection staff for such challenges is essential to ensure data integrity and emotional well-being.