The language factor: lessons for communication and community engagement from Translators without Borders’ experience

Communicating in the right language and format is critical to the success of any humanitarian response. The stakes could not be higher. People who don’t understand or speak the language used by humanitarians in a given context are disadvantaged and exposed to greater risks. Humanitarian operations are less effective as critical information is lost in translation and limited resources are wasted. Getting the language factor right improves not only communication and community engagement, but also access to services, needs analysis and accountability.

In the aftermath of the 2010 Haiti earthquake, over-reliance on international languages (English and French) and a lack of translation were identified as a ‘perennial hidden issue’. Harvard Humanitarian Initiative, Disaster Relief 2.0: The Future of Information Sharing in Humanitarian Emergencies, 2011 (https://issuu.com/unfoundation/docs/disaster_relief20_report). Since then, humanitarian organisations have explored new initiatives to communicate with affected people in relevant languages. Key frameworks and commitments now recognise the role of languages in communication and community engagement. See, for example, the Core Humanitarian Standard on Quality and Accountability (https://corehumanitarianstandard.org/the-standard) and the Grand Bargain commitment to a ‘participation revolution’ (https://interagencystandingcommittee.org/grand-bargain-hosted-iasc). Yet operationally, language as a key component in communications is still too often an afterthought, and few organisations have dedicated resources for language support. This is a problem especially for speakers of marginalised languages, who usually make up a significant portion of the at-risk or crisis-affected population. As a result, they are not always able to receive information from and communicate with humanitarians.

This article reflects on the experience of Translators without Borders (TWB) in the 2014–15 Ebola outbreak in West Africa and the European refugee response from 2015.

It also draws on our ongoing experience in the Rohingya refugee response and the internal displacement crisis in north-east Nigeria. The central lesson is that addressing language barriers is both essential and possible. It implies mobilising resources for language support and capacity-building, assessing needs and preferences and developing open resources and tools. In turn, those actions can ensure meaningful communication and community engagement.

Replacing unsafe assumptions with language support and data

Two potentially dangerous assumptions about language prevail in many responses. The first is that local colleagues do not face language barriers and can take responsibility for meeting language needs when necessary. But both international and local staff can face language problems. For example, displaced people in north-east Nigeria speak more than 30 languages. However, most humanitarian organisations rely on core national staff who are predominantly English, Hausa and to a lesser extent Kanuri speakers. TWB’s research found that interviews with internally displaced people often entail a four-stage translation between English, Hausa, Kanuri and a local language – and back again. Translators without Borders, Language Barriers in the Humanitarian Response in North-eastern Nigeria, July 2017 (https://translatorswithoutborders.org/wp-content/uploads/2017/04/Language-barriers-in-the-humanitarian-response-in-north-eastern-Nigeria.pdf). The result is slow planning and implementation, and unknown levels of information loss. Language is also a barrier when staff do not understand complex humanitarian terms and concepts in English. During TWB workshops with data collectors in north-east Nigeria, one group understood as few as 35% of the key terms they were being asked to interpret during surveys. This raises concerns about the accuracy of the data on which response plans are built.

The reality across humanitarian responses is that local staff carry the burden of multilingual communication with little or no support, and often outside their agreed job description. Effective alternatives include hiring vetted, trained translators and interpreters, offering language skills training and developing glossaries of key terms in relevant languages. Training affected people in translation and interpreting can also help overcome language barriers, and promote their active participation in response and recovery efforts.

The second assumption is that all affected people will understand the national or regional language or lingua franca, such as Hausa in north-east Nigeria. Information on the languages people speak and understand is largely unavailable at the level of detail needed for humanitarian planning. Aid organisations do not routinely collect this data – and when they do they rarely share it. As a result, responders too often use a language that many affected people don’t understand. Testing comprehension of simple content in four responses, we found that the language in which information is provided is of critical importance. Most respondents prefer to receive information in their mother tongue; English or another lingua franca is not seen as an adequate alternative. It is not safe to assume a person’s linguistic ability based on his or her country of origin. In Italy, for example, TWB found that humanitarian organisations rarely collect data about which languages refugees and migrants actually speak and understand. Instead, they use country of origin as a proxy for mother tongue. Yet most migrants in recent years have come from linguistically diverse countries, including Nigeria, which is home to over 500 languages. Without a reliable indicator of the languages affected people speak, humanitarians are ill-equipped to communicate with those they aim to help.

Thanks to the data collection capacity of the humanitarian sector, that problem is relatively easy to solve. The fastest way to identify language needs is for humanitarian organisations to include a few key questions in their standard assessments or household surveys. REACH and the Inter-Agency Standing Committee (IASC) Task Team on Accountability to Affected Populations and Protection from Sexual Exploitation and Abuse offer useful guidance on such questions. See https://interagencystandingcommittee.org/system/files/reach_iasc_aap_psea_task_team_menu_of_aap_questions_for_needs_assessments_june_2018.pdf. This data should be collected and published, with the necessary safeguards to avoid exposing affected people to risk.

In July 2017, regular International Organisation for Migration (IOM) Displacement Tracking Matrix surveys in north-east Nigeria began including key information on language and communication at site level. Now service providers in the area can consult that data on TWB’s Communications Dashboard. See https://translatorswithoutborders.org/communications-dashboard-internally-displaced-people-in-north-east-nigeria/. This identifies key factors such as the primary language of site residents and their preferred format of communication. This is a big step forward for planning communication with internally displaced Nigerians at these sites.

Communicating with individuals, not a community

No community is homogenous, meaning that there can be no one-size-fits-all solution for communication and community engagement. Language can compound communication challenges and increase people’s vulnerability to the impact of disasters and other crises. As a result, the people who most need to make their voices heard in an emergency are often the hardest to reach. For example, research on the 2014–16 Ebola outbreak in Liberia and Sierra Leone indicates that women initially died in greater numbers than men. ACAPS, Ebola Outbreak: Liberia – Communication: Challenges and Good Practices, April 2016 (https://www.acaps.org/special-report/ebola-outbreak-liberia-communication-challenges-and-good-practices); ACAPS, Ebola Outbreak: Sierra Leone – Communication: Challenges and Good Practices, April 2016 (https://www.acaps.org/special-report/ebola-outbreak-sierra-leone-communication-challenges-and-good-practices). This was in part due to their inability to access critical information. An early shortage of information material for non-literate audiences and in local languages left them in deadly ignorance. Similarly, TWB’s research in north-east Nigeria shows that current humanitarian communication practices favouring Hausa and Kanuri disadvantage minority language speakers, particularly less-educated women. TWB, The power of speech, July 2017 (https://arcg.is/01qeHC).

Communicating in simple, jargon-free language is an important first step to reaching these vulnerable groups. Plain-language content also reduces the risks of mistranslation and inconsistency between languages. This is important for humanitarian staff in our multilingual sector too: a TWB/IASC plain English version of the core principles on preventing sexual exploitation and abuse is helping to ensure that aid workers understand the rules. See https://drive.google.com/file/d/1A72OXTkLwJm9bt-aSKRbv71fOCGHpzE6/view.

The channel of communication used (bulletin board, radio, SMS) also affects who can access information and communicate with service providers. In many societies, women, older people and people with disabilities have less access than others to digital technology, radio and television broadcasts or public meetings. By understanding those constraints, humanitarians can identify the right channels for a given target group. The right mix of language and format then promotes comprehension. In north-east Nigeria, TWB found that, while audio messaging in either Hausa or Kanuri was most effective, comprehension of written information increased significantly when text was accompanied by a picture. TWB, Girl Effect and Oxfam, Language Profile of Five IDP Sites in Maiduguri, North-east Nigeria, July 2017 (https://translatorswithoutborders.org/wp-content/uploads/2017/04/Language-profile-of-five-IDP-sites-in-Maiduguri.pdf). Furthermore, research indicates that even people with limited literacy skills might want written information, such as illustrated leaflets. TWB and Save the Children, Language and Comprehension Barriers in Greece’s Migration Crisis, June 2017 (https://translatorswithoutborders.org/wp-content/uploads/2017/07/Language-Comprehension-barriers.pdf). Audio content is most widely understood, but simple text or graphics offer a more permanent record for information retention and later reference.

Alongside access and comprehension, language is bound up with another powerful factor in effective communication: trust. Trust may determine whether information is openly shared and properly acted upon. A shortage of female interpreters with the right languages can prevent women and girls from reporting abuse and accessing support. TWB, Putting Language on the Map in the European Refugee Response, September 2017 (https://translatorswithoutborders.org/wp-content/uploads/2017/04/Putting-language-on-the-map.pdf). An interpreter speaking the language of a government accused of repression may not be the right person to help asylum-seekers present their case to a host country. Ibid. More broadly, when a dominant language is spoken by humanitarians and other service providers and not universally understood, this can reinforce existing power dynamics and further marginalise vulnerable groups.

To meet the communication needs of the widest possible audience, humanitarians must first understand how those needs differ between sections of the population. With that knowledge they can then gear their language, formats and channels to those groups that are hardest to reach. A range of local languages and a mix of formats and channels will probably reach furthest and provide the most opportunities for people to triangulate information and respond. At a minimum, the use of plain language and audio or visual formats expands access to non-native and non-literate groups.

The value of a common approach and shared language resources

Poor coordination can leave affected people picking their way through a mass of humanitarian information from multiple sources. This makes it hard to find the specific information they need or know which version of the facts to trust. Uncoordinated communication also makes it less likely that translations into local languages will be consistent, clear and reliable. For example, in the Rohingya response a lack of coordinated emergency warning messaging in relevant languages and formats led to confusion when the 2018 monsoon season began. Strategic Executive Group, Joint Response Plan for Rohingya Humanitarian Crisis March–December 2018 – Mid-Term Review (https://www.humanitarianresponse.info/sites/www.humanitarianresponse.info/files/documents/files/2018_jrp_mid_term_review_v28.pdf). According to a recent evaluation, the Rohingya community in Bangladesh feel they have enough information on health, safety and security, but not about keeping their family safe during heavy rainfall. Nicola Bailey et al., How Effective Is Communication in the Rohingya Refugee Response?, September 2018 (http://downloads.bbc.co.uk/mediaaction/pdf/research/rohingya-research-report.pdf). Low literacy levels and a shortage of skilled humanitarian interpreters are part of the challenge.

To support community engagement efforts, a collaboration between BBC Media Action, Internews and TWB has contributed common service tools and activities to help aid organisations communicate better with the affected population. Ibid. Ensuring relevant content is available in the Rohingya language is a critical component.

The TWB Glossary for Bangladesh and related socio-linguistic guidance are also supporting responders, especially when translating humanitarian terminology. The use of technical terminology that is not readily conveyed in local languages can cause confusion and misunderstanding. For example, TWB’s research in Cox’s Bazar found ‘violence against women’ translated as ‘violent women’. In north-east Nigeria, ‘safe space’ was understood as ‘a space protected by guards’. Glossaries can help humanitarians use consistent, accurate and easily understood words, and produce relevant communication materials in local languages.

When resources are scarce, it makes sense to build a library of materials in the right languages that all service providers can draw on. This can also contribute to the use and usefulness of language technology in humanitarian crises. It can help build the capacity of machine translation in voice and text in those languages, and encourages humanitarian tech developers to integrate language technology into their tools. In time, that capacity will allow speakers of marginalised languages to have conversations with responders and access the information they want directly, and in their own languages.

Challenging the current approach

It is within the grasp of every humanitarian organisation to improve how they address the language factor. As a fundamental component of effective and accountable action, it is also within everyone’s remit. Yet shifting the responsibility onto unsupported national staff remains common practice, as does the use of jargon and written complaints mechanisms in non-literate communities. Are we ready to make effective multilingual communication the default approach in humanitarian action?

Mia Marzotto is a Senior Advocacy Officer at Translators without Borders.