Malin Petzell, language Gothenburg University

2009-2014

This project describes and thoroughly analyses the language referred to as Kami, spoken in the Morogoro region in Tanzania, approximately 150 km from Dar es Salaam. Kami (G36) is both under-described and endangered - there is one scanty (and out-of-date) description of the language from 1900 by Velten. It is said to be spoken by approximately 5,000 people, but our field work in the area reveals that the number of fluent speakers left is much less. The purpose of the project is to describe (document and analyse) this endangered variety for posterity, but also to analyse the structure of the language and relate the new findings to current linguistic assumptions, especially those relating to comparative Bantu. The study comprises a grammar of Kami, i.e. the phonological, morphological and syntactic structures, as well delving into specific linguistic phenomena that are particularly interesting and/or distinctive for Kami.

The three most important findings
There is variation in the Kami phonetic inventory. When interviewing Kami speakers and eliciting data, there is a striking variety of allophones representing the same phoneme. Some of these sounds can be traced to Swahili influence and some to local languages such as Luguru, while others are probably Bantu cognates. There is variation between speakers but also within the idiolect of individual speakers. We are not opposing the claim that language obsolescence can lead to simplification, but in our case the opposite seems to be true. A factor that most likely plays a role in the abundance of forms is the fact that more than two languages are influencing the endangered language. The Kami language has ceased to exist as a uniform language since it is not the main language for any individual speaker, nor is it the main language in any domain. We believe that the Kami speakers have expanded their sound inventory by incorporating the phonetic variants from surrounding languages. This unstable situation - an obsolescent language which is heavily influenced by more than two other languages - has led to the wide variety of phones within Kami.

In many Bantu languages, there is a pre-prefix (PP) that appears ahead of the obligatory noun class prefix. In the neighbouring language Luguru the PP is very frequent, but in Kami, it is virtually non-existent. In Luguru in the 1970s, the PP was used mainly by older speakers, but today young people are using the PP more and more, and using it in contexts where it used to be disallowed. These opposing standpoints ? hardly any PP-usage in Kami vs. increasing usage in Luguru ? are remarkable for such closely related languages. We propose that the wide-spread use of the PP in contemporary Luguru may have started out as some form of hypercorrection, but that the PP has now become grammaticalized as a topic marker. The very low level of usage in Kami can partly be explained by historical factors, but also by the fact that there is no PP in the dominant language Swahili. In the multilingual contact situation that characterizes the area, a vulnerable language like Kami is more prone to losing language-specific features than Luguru.

Finally, we propose a reanalysis of the passive in Bantu. In Bantu languages, tense, aspect and mood [TAM] markers (both pre- and post-root) are by and large termed inflection and the verbal extensions termed derivation. Inflection is here taken to mean a change in a word that expresses a grammatical phenomenon, such as noun class membership, person, tense, aspect or mood. Verbal derivation stands for the extensions that derive verbs from verb roots. These verbal extensions can be valence-increasing, decreasing or maintaining. In the Bantu tradition, the passive is often regarded as a verbal extension as a result of behaving the same way as other extensions morphosyntactically and appearing in the same slot. If the passive is truly derivational, it is highly unlikely to be preceded by inflectional markers. Nonetheless, this is often the case, and when the passive occurs in a slot where it 'should not be', it is every now and then explained by imbrication or similar. What is more, there is an ongoing debate on the status of the passive and whether it really is valence-changing (i.e. derivational) or not. We make a case for a gradual continuum between inflection and derivation in Kami, as well as in other Bantu languages.

A new research question that has arisen is how we can account for the variation found within a language. As mentioned, Kami varies not only between speakers but the same speaker may use a variant word forms or morphemes within the same sentence. Our paper 'Phonetic variation in an obsolescent language (Kami, G36)' has been accepted for the 8th World Congress of African Linguistics (WOCAL) in Kyoto in August 2015.
Another new research question, which also generated an article, is what level of 'editing' is accepted in documentation. When the speakers of a language themselves regret their choice of words in a recording and want to replace it with a 'better', more pure Kami word, should the researcher documenting the language accept such editions? What version of a language should be documented? In the case of Kami, as well as many other Bantu languages in Tanzania, Swahili is the official language, the language used in school and basically everywhere but in the homes. This means that code switching is ubiquitous, and a 'pure' version of a language is almost never heard. We show that all versions of the language - both the original and the edited (purer) one are valid, 'authentic' representations of their own linguistic reality, and that both of them are part of a comprehensive documentation of a language and its linguistic ecology.

International connections: The researchers in this project and the co-authors of articles come from universities in four different countries. The field research is carried out in a fifth country, namely Tanzania. The findings have been presented at international conferences such as The 5th International Conference on Bantu Languages, and it will be presented the 6th International Conference on Bantu Languages in 2016 and the above-mentioned WOCAL in 2015. All publications are in English.

Carrying out field work is in itself a way of disseminating the research. Every person interviewed, as well as passers-by, are briefed on the project. In terms of dissemination to the broader public, Petzell participated in the so called 'Academic quarter' (a short lecture open to the public) in Gothenburg.
https://www.youtube.com/watch?v=p3XMrjh4eZc&index=25&list=PL1629A9886730CCE5

The most important publication is a chapter in an anthology on Bantu languages. We have been asked to write a description of Kami. Although invited to write this, the work will be peer reviewed. The grammatical description will be coupled with a discussion of phonetic variation in Kami, as well as the structure (and status) of the passive. The second most important publication is an article entitled 'The impact of sociolinguistic factors on the usage of the pre-prefix in some G-languages'. Petzell has been invited to the University of Kiel in February 2015 to give a talk on this paper.

The publication strategy is to reach as many researchers as possible, which is why we have opted for open access publications, although this is not mandatory for our project.

Publications

Petzell, Malin and Aunio, Lotta. (in prep.) ‘A comparative description of Kami.’ (For inclusion in the second edition of The Bantu languages.)

Petzell, Malin. (in prep.) ‘The impact of sociolinguistic factors on the usage of the pre-prefix in some G-languages.’ (To be submitted to Studies in African Linguistics).

Marten, Lutz and Petzell, Malin. (under review) ‘Linguistic variation and the dynamics of language documentation: Editing-in “pure” Kagulu.’ (Submitted to Language Documentation and Conservation).

Petzell, Malin and Hammarström, Harald. 2013. ‘Grammatical and lexical comparison of the Greater Ruvu Bantu languages.’ Nordic Journal of African Studies 22 (3) 2013. pp. 129–157.

Petzell, Malin. 2013 ‘Språklig stigmatisering i Tanzanias skolor.’ ASLA:s skriftserie 24 Språk i undervisning. pp. 83–92.

Petzell, Malin. 2012. ‘The linguistic situation in Tanzania.’ Moderna språk 106 Uppsala: Riksförbundet för lärarna i moderna språk. pp. 136–144.

Petzell, Malin. 2012. ‘The under-described languages of Morogoro - a sociolinguistic survey.’ The South African Journal of African Languages 32. African Language Association of Southern Africa and the Foundation for Education. pp. 17–26.

Website (Petzell): http://sprak.gu.se/kontakta-oss/larare/petzell-malin

Website (Aunio): http://tuhat.halvi.helsinki.fi/portal/en/person/lharjula