In 2023, our ongoing journey with Loop unveiled key insights into the importance of language for accessibility and trust building. Through practical prototyping and testing, we identified language and literacy as substantial barriers affecting feedback, reporting, information access, and agency. This exploration highlighted the critical necessity of addressing these challenges in communication. Join us as we delve into the pragmatic lessons learned, where language proved pivotal in overcoming barriers and enhancing community engagement.
Data analytics and research in 2023 illuminated the divergence between standard written languages and oral languages in various communities. For instance, in Somalia, Maxatiri, the primary language for education, dominates both written and spoken communication. In contrast, Maay, the second most prominent language, lacks a standardised written form, leading to multiple dialects and comprehension challenges.
For example, in Somalia, Maxatiri - the primary language for education, as a result, is the main language that is written and spoken across Somalia. So, if someone can read and write in any Somali dialect they will have first learnt to read and write in Maxatiri. This scenario appears to be representative of other contexts such as Bemba or Tonga in Zambia for example. In addition, we observe that some populations may speak 3 or 4 languages or dialects but not feel confident writing in any.
We learnt that efficiency and accuracy in moderation is essential to be able to scale the platform. Language is also a key element of inclusion and a valuable offer we can provide to marginalised communities. Based on this learning we have developed two tiers of languages: written and oral languages or dialects which are treated slightly differently on the platform.
By written language we mean languages such as Maxatiri or Nyanja, these are available across all platforms and feedback and data analytics can be accessed in these languages. Instructions and communication materials are also available in these languages, just like English, Spanish and French for example. This inclusion of national languages targets people who are literate and do not speak a UN language (English, French, Spanish etc).
On Loop an oral language refers to languages which tend not to be taught in schools but which are widely spoken as a first language by communities. On Loop these oral languages are available for people to listen to instructions, to feedback in, to read small, primarily visual communications materials and to receive voice mail replies in. Loop tags this feedback in the Oral language (Maay, Tonga etc) but the written text on the platform is only available in the written language of the country and the other Loop written languages. This targets people who speak an underserved language and are not literate in this language. Having the predominantly oral language in a written format on the Loop platform does not necessarily increase accessibility to information.
To implement this distinction for Somalia (Maay, Benadiri, Benadiri af Murke, Banjuli) and Zambia (Tonga, Lozi) in 2024, we will be seeking additional funding.
Loop is available in 17 languages and three scripts, covering a wide linguistic spectrum:
1) English, Arabic, French and Spanish
2) Nyanja, Bemba, Lozi and Tonga for Zambia
3) Tagalog and Cebuano for the Philippines
4) Bahasa Indonesia for Indonesians
5) Somali Maay and Somali Maxatiri for Somalia
6) Ukrainian, Polish and Russian
Throughout 2022 and 2023, both sensitive reports and open stories were submitted in all 15 of the languages. Russian and three Somali dialects will be added in the first quarter of 2024. Our ongoing learning emphasises the importance of inclusivity, especially concerning oral dialects and languages in Somalia. Collaborating with Minority Rights Group (MRG), we developed a language strategy to engage communities effectively. The strategy recognises the significance of mother tongues, often the most comfortable means of communication around sensitive issues.
In 2023 Loop received the majority of feedback in Somali Maxatiri, for both Open and Sensitive feedback. It was also the majority of sensitive feedback. We received the second largest amount of feedback in Somali Maay.
Amount of Feedback received on Loop in 2023 by Language
If we remove Maay and Maxatiri, the other languages used show that English, followed by Tagalog and then Bemba are the most used languages on Loop in 2023.
Feedback on Loop, Open and Sensitive feedback (tagged by author or moderator) by language, with Somali Maxatiri and Maay removed.
Nyanja and English are most prominent for Sensitive feedback. We see in Zambia feedback coming in a range of the 5 available languages and a larger percentage of sensitive reporting, specifically Gender Based Violence compared to other countries. This is received in both English and Nyanja.
The below table shows the amount of feedback received by input channels (SMS, WhatsApp etc) by each different country.
IVRR / Voice is only available in Somalia (all year) and Zambia (since August) due to delays with Mobile Network contracts. People from Ethiopia, affected by the Somali floods have been able to call the Loop number. We hope to also set up IVRR in the Philippines but await funds to meet the Mobile charges for the year. IVRR is also available on a paid international number which is expressed here as Poland.
Facebook messenger was requested by the Philippines and is the primary channel of feedback received. There are small amounts of feedback received via facebook messenger in other countries as well.
Telegram, which was requested by Ukrainians, was only used in Poland and Ukraine as would be expected. We have not done any promotion or added on all of the languages to Telegram, based on community feedback.
We can see that WhatsApp is most dominant in Indonesia, where it was requested to be added.
Web feedback is used across all countries including Somalia where the only other feedback option during 2023 was through IVRR/ Voice.
Feedback on Talk to Loop in 2023 by language and feedback channel, with Somalia removed
Together, this data shows us that different countries prefer different input channels but there is always a subsection of people who will prefer to use a different channel and method. Similarly we can see different languages being used across different input channels. Providing choice adds agency and increases inclusion.
Oral language emphasis is needed: We must prioritise oral language accessibility, recognizing its significance for non-literate communities and seek funding for the inclusion of more languages and dialects in 2024
The well functioning of Loop's platform also depends on efficient Moderation:
How do we enhance moderation efficiency to ensure scalability? We are exploring tech enabled solutions coupled with human moderation, that can also address language-specific moderation challenges for accuracy.
Communication Clarity: How can we effectively convey the availability of languages on our channels to users and communicate our ongoing efforts to incorporate additional languages. We also need to foster awareness of language-specific features to encourage organisational use.
