Annotated Swadesh wordlists for the Gondi-Kui group (Dravidian family).

Languages included: Konda [gku-kon]; Pengo [gku-pen]; Manda [gku-mnd]; Kui [gku-kui]; Kuwi [gku-kuw].


I. General

Burrow & Emeneau 1984 = Burrow, Thomas; Emeneau, M. B. A Dravidian Etymological Dictionary. 2nd edition. Oxford: Clarendon Press. // The major source of reference on comparative Dravidian etymology, with more than 5000 comparanda. Although the dictionary mainly relies on previously published sources for its materials, it also contains some previously unpublished fieldwork data, most notably on the Gadba and Manda languages.

II. Konda

Krishnamurti 1969 = Krishnamurti, Bhadriraju. Koṇḍa or Kūbi: A Dravidian Language (Texts, Grammar, and Vocabulary). Hyderabad: Tribal Cultural Research & Training Institute. // An extensive description of Konda Dora, focusing on the Araku valley dialect, and accompanied by a small, but representative glossary.

III. Pengo

Burrow & Bhattacharya 1970 = Burrow, Thomas; Bhattacharya, S. The Pengo Language. Oxford: Clarendon Press. // The only extensive monograph/data source on the Pengo language, containing a detailed phonological and grammatical sketch, a collection of texts, and a mid-size vocabulary of the language.

IV. Manda

Reddy 2009 = Reddy, B. Ramakrishna. Manda-English Dictionary. Mysore: Central Institute of Indian Languages. // A mid-size vocabulary of the Manda language; this is currently the most detailed source on this little-studied language.

V. Kui

Winfield 1929 = Winfield, W. W. A Vocabulary of the Kui Language (Kui-English). Calcutta: Baptist Mission Press. // Comprehensive and fairly accurate (for its epoch) dictionary of Kui. Based primarily on the Gumsur-Udayagiri dialect, but with notes on deviating forms in several other dialects (Barma, Phulbani, etc.).

Maheswaran 2008 = Maheswaran, C. A descriptive grammar of the Kui language. Bangalore: Dravidian University. // Detailed grammar of Kui, accompanied by illustrative texts and a representative vocabulary. Based primarily on the Balliguda (China Kimedi) dialect.

Burrow & Bhattacharya 1961 = Burrow, Th.; Bhattacharya, S. Some notes on the Kui dialect as spoken by Kuṭṭia Kandhs of North-East Koraput. In: Indo-Iranian Journal, 5, pp. 118-135. // A short paper describing the authors' field research on the Kuṭṭia dialect of Kui. Does not include a comprehensive grammatical description or any wordlists/vocabularies, but contains a significant amount of lexical material that may be used in etymological comparison.

VI. Kuwi

Israel 1979 = Israel, M. A grammar of the Kuvi language. Trivandrum: Sangam Printers. // Detailed grammatical description of the Kuwi language.

Burrow & Bhattacharya 1963 = Burrow, T.; Bhattacharya, S. Notes on Kuvi with a short vocabulary. In: Indo-Iranian Journal, v. 6/3-4, pp. 231-289. // A description of Kuwi grammar, based on the authors' own fieldwork with several subdialects. Includes a small, but representative vocabulary for Sunkarametta and Parja subdialects of Kuwi.


1. General.

I. Konda

The main source of lexical and grammatical information on Konda is Bh. Krishnamurti's monograph on the language [Krishnamurti 1969]. Judging by some elements of the description, as well as a number of lexical discrepancies in the glossary, dialectal variety within Konda is substantial; unfortunately, available data does not allow for the construction of comprehensive lists for more than one dialect, represented by Krishnamurti's informant Bōyi Sombra from Gorṛa Guṛa, Araku valley.

Where they are available, we quote concurrent forms from the Sova dialect, also from [Krishnamurti 1969]; and forms from an unnamed dialect represented in the field notes of Th. Burrow & S. Bhattacharya that they collected in 1956-57 and partially published in the Dravidian Etymological Dictionary [Burrow & Emeneau 1984]. In quite a few cases, Krishnamurti's Sova and Burrow & Bhattacharya's unnamed Konda dialect actually preserve more archaic equivalents for Swadesh meanings (e. g. we count at least 9 lexical discrepancies between Konda proper and Sova Konda, 4 of which are archaisms in Sova), but it is impossible to establish a complete statistical picture because the dialectal data are not systematic.

II. Pengo

The only significant source on data that is available on the Pengo language is the monograph [Burrow & Bhattacharya 1970], which contains a detailed grammar sketch, some examples of transcribed and translated texts, and a comprehensive, though, most likely, far from exhaustive vocabulary, along with scattered observations on dialectal varieties (some of which are mentioned in the Notes section of the wordlist).

It should be noted that Pengo is under heavy areal pressure from the "prestigious" Oriya language: quite a few Swadesh meanings are represented in Burrow & Bhattacharya's vocabulary by several equivalents, including an inherited term and one or more Oriya equivalents. In most such cases, analysis of texts shows that the inherited form is usually the preferred equivalent, while the Oriya term is used in more specialized (e. g. bound idiomatic) contexts. However, this is not a 100%-accurate rule, and in some cases, the Oriya borrowing even turns out to have completely replaced the inherited Dravidian word (e.g. 'rain', 'tooth').

III. Manda

Until recently, the only source of lexical knowledge on Manda was a short lexical set, extracted from the field notes collected by T. Burrow and S. Bhattacharya in 1964-66 and subsequently published in [Burrow & Emeneau 1984]. These extracts, however, were tailored specially for etymological purposes, including only those Manda words that had reliable external parallels in other Dravidian languages, and were not suitable for the construction of a proper Swadesh wordlist.

The recent publication of [Reddy 2009], a proper dictionary of the language, finally allows to construct a proper list (including all the replacements with Indo-Aryan borrowings or words without a transparent etymology) and include Manda in the general lexicostatistical matrix for Dravidian languages. The dictionary is not perfect as far as semantic glossing and differentiation of synonymy is concerned; in particular, several Swadesh items on the list are represented by doublets (often consisting of an inherited term and a recent borrowing) without any indication as to which one is the most basic/frequent. It may be hoped, though, that a few semantic errors on the list will not skew the final results too much.

IV. Kui

The Kui language is well represented lexicographically by W. W. Winfield's "classic" dictionary that seems to have been fairly accurate, phonetics-wise, for 1929 (although the quality of semantic glossing is sometimes dubious). Besides that, there is a number of rather scattered sources on numerous Kui dialects that cannot be used for lexicostatistical purposes, but are of serious value for etymological research (e. g. [Burrow & Bhattacharya 1961] on the highly divergent Kuṭṭia dialect, which could really be said to represent an "intermediate" stage in the continuum between Kui and Kuwi).

The only source at our disposition for which an additional 100-item wordlist could be constructed is the Balliguda dialect, recently described and provided with a vocabulary by C. Maheswaran [Maheswaran 2008]. However, lexicostatistical discrepancies between Winfield's Udayagiri dialect and Maheswaran's Balliguda are few and far in between, and what little there is may just as well be ascribed to semantic inaccuracies in glossing as it could be to genuine lexical divergence. Out of caution, we therefore select [Winfield 1929] as our default source for Kui, and list the Balliguda equivalent(s) from [Maheswaran 2008] in the comments section.

V. Kuwi

There is actually an abundance of data on Kuwi, the major problem being that the majority of sources are old and may not necessarily be up to modern standards of phonetic and semantic accuracy. We decided to limit ourselves to but a few sources, largely ignoring the oldest (but still relevant) lexicographical descriptions by A. G. Fitzgerald (1913) and F. Schulze (1911-1913). Instead, our main source for the compilation of a Kuwi wordlist is [Israel 1979], a large grammatical description accompanied with a good selection of folktales and a vocabulary, based on the author's own fieldwork in Saptamaha, on the Deomali hill (with the language variety said to be "the standard dialect of Kuwi"). As a control source, we have used [Burrow & Bhattacharya 1963], a paper dealing with two more sub-dialects of Kuwi (Sunkarametta and Parja). Some of the dialects show rather transparent lexical discrepancies; however, on the whole they were judged too insignificant, and the absolute reliability of data too questionable, to allow for a rigorous construction of several different wordlists. We have, therefore, limited ourselves to [Israel 1979] as the primary source, and have relied upon Burrow & Bhattacharya's data for questionable cases where Israel's data alone do not allow to make an unambiguous choice between two or more potential quasi-synonyms.

2. Transcription.

I. Konda

For the most part, no special transliterational actions except for the usual conversion of retroflex notation from the traditional Indologist format into UTS ( = ʈ, etc.; long vowels ā, ē = , , etc.). Palatal affricates c, j are transcribed as palatal ɕ, ʓ.

WARNING: Compared to most other Gondi-Kui languages, Konda has the most complex (and archaic) system of coronal taps and trills. Here we differ from the traditional Indological transcription and convert everything to IPA/UTS, namely:

Phoneme Krishnamurti 1969, etc. UTS
Dental tap r ɾ
Retroflex tap ɽ
Alveolar trill (voiced) r_ r
Alveolar trill (voiceless) R

It is important, therefore, to remember that the simple transcription of r actually corresponds to what is usually transcribed as "alveolar r_" in the literature.

II. Pengo

No special transliterational actions except for the usual conversion of retroflex notation from the traditional Indologist format into UTS ( = ʈ, etc.; long vowels ā, ē = , , etc.). Palatal affricates c, j are transcribed as palatal ɕ, ʓ.

III. Manda

Reddy's transcription of Manda data needs very little transliterational effort, apart from the standard recoding of retroflex consonants (ṭ = ʈ, etc.). It should be noted that the phonemes which Reddy transcribes as fricatives s and z in Burrow and Bhattacharya's fieldnotes regularly correspond to palatal affricates c and j (= UTS ɕ, ʓ); this may be a subdialectal differentiation (the entire Manda-speaking community hardly exceeds 4,000 speakers, but this does not exclude the possibility of minor variations between villages).

IV. Kui-Kuwi

Transliteration conventions for both Winfield's and Maheswaran's data, as well as Israel's data for Kuwi, are generally the same as for Pengo and Manda.

Database compiled by: G. Starostin (last update: November 2016).