Annotated Swadesh wordlists for the Fur-Amdang group (Fur family).

Languages included: Fur [fur-fvr]; Amdang [fur-amd].

DATA SOURCES

I. Fur.

Jakobi 1990 = Jakobi, Angelika. A Fur Grammar. Phonology, Morphophonology and Morphology. Hamburg: Helmut Buske Verlag. // Detailed description of Fur, well illustrated by examples and an annotated text, but without an accompanying glossary.

Beaton 1968 = Beaton, A. C. A Grammar of the Fur Language. Khartoum. // An older description of Fur (structured almost as a real manual, with exercises etc.), less phonetically accurate than recent sources but very well illustrated with lexical data.

Kutsch Lojenga & Waag 2004 = Kutsch Lojenga, Constance; Waag, Christine. The sounds and tones of Fur. In: Occasional Papers in the Study of Sudanese Languages, 9, pp. 1-25. // Detailed description of Fur phonology.

Waag 2010 = Waag, Christine. The Fur Verb and its Context. Köln: Rüdiger Köppe Verlag. // A special monographic treatment of the complicated verbal system of Fur. Includes some morphological and lexical data (on verbs) unavailable in other sources.

II. Amdang.

Wolf 2010 = Wolf, Katharina. Une enquête sociolinguistique parmi les Am\-dang (Mimi) du Tchad. Rapport Technique. SIL International. // A sociolinguistic survey of the Amdang language, including 227-item wordlists collected from speakers in four different villages.

Doornbos & Bender 1983 = Doornbos, Paul; Bender, M. Lionel. Languages of Wadai-Darfur. In: Nilo-Saharan Language Studies. Ed. by M. Lionel Ben\-der. Michigan: East Lansing, pp. 42-79. // A condensed report on the authors' fieldwork with various languages of the Wadai-Darfur area, including excerpts from P. Doornbos' lexical data on Biltine (Amdang).

Jungraithmayr 1971 = Jungraithmayr, Hermann. How many Mimi languages are there? In: Africana Marburgensia, 4, 2, pp. 62-69. // A small wordlist collected by the author from speakers of the "Mimi" language, which, upon close survey, turns out to be different from the previously attested "Mimi" languages and is now reliably identified as a dialect of Amdang.

NOTES

1. General

I. Fur.

Despite the large number of Fur speakers, a definitive dictionary of the language, from which a 100% reliable Swadesh wordlist could be derived, is still lacking. Because of this, our main source for Fur data remains [Jakobi 1990], a detailed description of the language's phonology and grammar whose general accuracy may be trusted; however, it is impossible to construct a complete lexicostatistical wordlist out of its data, so we have to rely on additional sources for completeness - such as the newer, more condensed (but well illustrated by examples) description [Kutsch Lojenga & Waag 2004], as well as the monograph on Fur verbal morphology written by Ch. Waag [Waag 2010]. On the whole, Jakobi's and Waag's data are in mutual agreement, so using all of these sources to fill in primary slots seems permissible.

As an additional control source, we use [Beaton 1968] (in 5 cases, we even have to select it as our primary source): in terms of phonetic accuracy, this is a highly questionable resource, but it also contains a large amount of lexical data, unavailable elsewhere.

Note: Fur is known to have a very complex system of verbal morphophonology, including allomorphs that are widely different from each other in shape, yet can be derived from each other through a set of formal rules (e. g. 'to burn': su- / =uy-, etc.). We follow the general outlines of A. Jakobi's internal reconstruction and, when filling the primary slot, place the allegedly more phonetically archaic allomorph first, and its contextually transformed alternant in the second place.

II. Amdang.

The main source on the relatively recently discovered language of Amdang (= Biltine or "Jungraithmayr's Mimi" in earlier sources), and the only one that allows it to be included into meaningful lexicostatistical analysis, is [Wolf 2010], a sociolinguistic survey with a complete 227-item list for four subdialects of the language as spoken in four different villages in the Biltine department of Chad. Since lexical differences between the four subdialects are very few (not exceeding 10 items on the Swadesh list), and since we cannot properly vouch that these represent genuine lexicostatistical discrepancies and not semantic inaccuracy on the part of the researcher, we only select data from one village (Sounta) in the primary slot. Data from the other three villages (Kouchane, Yaouada, Tere) are listed in the comments section.

For the sake of completeness, the comments section also includes data from earlier sources on Amdang: [Jungraithmayr 1971], the first paper to mention the very existence of this language, and [Doornbos & Bender 1983]. Wordlists given in both of these sources are very small, but on the whole, seem to agree fairly well with the more recent and detailed data in [Wolf 2010]. An ongoing problem is the morphological segmentation of forms given in all sources of Amdang, especially verbal ones: it seems that Amdang has a complicated system of verbal morphology, comparable to Fur, but no clear description of it has appeared so far, meaning that segmentation decisions have to be taken based on systematic structural analysis of the lexical evidence.

2. Transcription

I. Fur.

Fur has a relatively simple phonological system, and differences between systems used by specific researchers, UTS, and IPA are minimal. Only a few notes are necessary.

(a) The voiced palatal plosive (affricate?) is written as j in [Jakobi 1990] and in [Kutsch Lojenga & Waag 2004] = UTS ʓ.

(b) Palatal and velar nasals are respectively transcribed as ñ and N in [Jakobi 1990] = UTS ɲ and ŋ (same way also in [Kutsch Lojenga & Waag 2004]).

(c) Long vowels are written as doubled aa, ee, oo, etc. in [Jakobi 1990] = UTS aː, eː, oː, etc.

(d) In [Jakobi 1990], Fur is analyzed as a 5-vowel language; in [Kutsch Lojenga & Waag 2004], the authors introduce three additional phonemes - ı, ʋ, ǝ. We reflect all these differentiations in our transcription.

(e) Fur has two tonal registers, High and Low; in [Jakobi 1990], only Low is marked with a gravis (V̀), whereas High remains unmarked. Conversely, in [Kutsch Lojenga & Waag 2004] Low tone is unmarked and High tone is marked with an acute diacritic. We consistently mark both High as V́ and Low as V̀.

II. Amdang.

[Wolf 2010] consistently uses IPA to transcribe Amdang data, requiring only the usual "cosmetic" amendments for UTS: thus, Wolf's tʃ = UTS č, dʒ = UTS ǯ, j = UTS y. It should be noted that Wolf's transcription is generally phonetic rather than phonological, for instance, consistently marking aspiration (kʰ, tʰ, etc.), even though there does not seem to be any real phonological opposition between aspirated and non-aspirated allophones of simple voiceless plosives in Amdang. Nevertheless, we do not make any efforts to "phonologize" the transcription, since no detailed phonetic/phonological description of Amdang has been published so far that would allow to make definitive conclusions.

Database compiled by: G. Starostin (March 2017).