The Temi Alphabet

Temi is the community name (endonym) for the Sonjo language (ISO 639-3: soz), a Bantu language spoken by approximately 10,000–15,000 people (the WaTemi community) in the Sonjo Highlands of northern Tanzania, near Lake Natron and the Eastern Rift Valley ^[1].

The people call themselves WaTemi (plural) or MuTemi (singular), and their language is Temi. The name "Sonjo" is an exonym — the name given by outsiders, including neighbouring communities and early colonial administrators. Both names refer to the same language and people ^[2].

The name Sonjo is an exonym — the name given by outsiders to the language and community. The community's own name for their language is Temi, and they call themselves WaTemi. This dual naming is common in linguistics across African languages ^[2].

Linguistic literature may use either name. ISO 639-3 codes the language as soz under Sonjo. Community documentation projects increasingly use Temi as the preferred endonym, reflecting a wider trend in African linguistics of prioritising community self-designations over externally imposed names ^[1].

Temi (Sonjo) is spoken in the Sonjo Highlands of northern Tanzania, in a geographically isolated area east of the Serengeti and near Lake Natron — one of the most alkaline lakes in the world. The community lives in a cluster of villages in this highland area ^[1].

The isolation of the Sonjo Highlands — bounded by the Rift Valley escarpment and the hostile Lake Natron shores — has helped the WaTemi community maintain their distinct Bantu language and culture despite being surrounded by Maasai-speaking communities ^[3].

The WaTemi (Sonjo) are Bantu-speaking farmers who live in a region otherwise dominated by Maasai-speaking pastoralists. The Maasai speak Maa, a Nilotic language completely unrelated to the Bantu family. The two communities have coexisted for centuries ^[3].

This coexistence has resulted in loanwords from Maa entering Temi, particularly vocabulary related to cattle, land, and seasonal practices. Temi/Sonjo is therefore a linguistically interesting case of Bantu-Nilotic contact — a Bantu language that has absorbed Nilotic vocabulary through proximity rather than through Bantu-internal processes ^[2].

Temi (Sonjo) has been documented by linguists at the University of Dar es Salaam (UDSM) in Tanzania and at LLACAN (Laboratoire Langage, Langues et Cultures d'Afrique, CNRS) in France. Field recordings include wordlists, grammatical elicitation, and narrative texts ^[1].

Published research on Sonjo/Temi appears in journals of African linguistics, including SAGE Publications' Journal of African Languages and Cultures. The language is not endangered in terms of intergenerational transmission, but its small community size makes documentation important for comparative Bantu studies ^[2].

Temi shares all core Bantu features: five-vowel system (a, e, i, o, u), prenasalized consonants (mb, nd, ng, nj) as single phonemes, a noun-class prefix system, verb agglutination, and a two-tone (high/low) system ^[2].

The pan-Bantu numeral root *-kumi (ten) appears in Temi as in Swahili (kumi), Shona (gumi), Tumbuka (khumi), and dozens of other Bantu languages — demonstrating that Temi/Sonjo shares a common ancestor with all Bantu languages despite its peripheral location in the Rift Valley ^[3].

The ISO 639-3 code for Temi (Sonjo) is soz, registered under the Sonjo name. The code is the same whether the language is called Temi or Sonjo — they are the same language with two names: Temi (endonym, the community's own name) and Sonjo (exonym, the outside name) ^[1].

The community of approximately 10,000–15,000 WaTemi is located in Rift Valley Region, Tanzania. There is no official writing system or government recognition for the language, but documentation efforts ensure that Temi is recorded for future linguistic comparison and community use ^[2].