¹ The bibliography is as in Lg. 21.184 fn. 1, with additions as given further on in this note.

The sigilla for the languages are as follows: Dr. = Dravidian, PDr. = Proto-Dravidian, Ta. = Tamil, Ma. = Malaylạ̄am (H. Gundert, A Malayalam and English dictionary, Mangalore, 1872), Ka. = Old Kannaḍa (F. Kittel, A Kannaḍa-English dictionary, Mangalore, 1894), Koḍ. = Koḍagu (Coorg),Tu. = Tuḍu, Te. = Telugu (P. Sankaranarayana, A Telugu-English dictionary, Madras, 1927), Kol. = Kolami (usually my field material; SR = P. Setumadhava Rao, A grammar of the Kolami language, Hyderabad, 1950; Kin. = dialect of Kinwaḍt, Adilabad, Nk. = Naiki;Pa. = Parji;Oll. = Ollari;Go. = Gondi (without further sigillum, C. G. Chenevix Trench, Grammar of Gondi, Madras, 1919, 1921; M. = A. N. Mitchell, A grammar of Maria Gondi as spoken by the Bison Horn or Dandami Marias of Bastar State, Jagdalpur, 1942; L. = A. A. Lind, A manual of Madia, Kedgaon, 1913 [not available to me, but quoted sometimes from Burrow]; W. = H. D. Williamson, Gondi grammar and vocabulary, London, n.d. [1890]); Kur. = Kurukh (Oraon); Malt. = Malto; Br. = Brahui. Kota, Toda, Kui, and Kuwi are not abbreviated; Kuwi (F) = A. G. Fitzgerald, Kuvinga Bassa, The Khond language as spoken by the Paḍjas ...of the Madras Presidency, Calcutta, 1913; Kuwi (S) = F. V. P. Schulze, Vocabulary of Kuvi-Kond language, Madras, 1913. IA = Indo-Aryan. Turner = (Sir) Ralph Lilley Turner, A comparative and etymological dictionary of the Nepali language, London, 1931.

1 am indebted to Professor T. Burrow of Oxford University and Sri Sudhibhushan Bhattacharya of the Indian Museum, Calcutta, for allowing me to use items collected by them during their field trips in Bastar and in the Adilabad district of Hyderabad; the languages quoted on their authority are the Kinwaṭ dialect of Kolami, Naiki, Parji, and Ollari. My work has benefited much from discussion with Burrow.

A summary account of the phonetic symbols used will be found in BSOAS 14.98 fn. 1.

² Especially Robert Caldwell, A comparative grammar of the Dravidian languages ³ 397-400 (1913); Julien Vinson, Manuel de la langue tamoule 69-70 (Paris, 1903). Caldwell undoubtedly had examined the material in his 2d edition of 1875 and perhaps even in the 1st edition of 1856; I have no easy access to either of these editions. Most recently P. Meile has discussed some at least of these forms in a lecture which is reported in outline in BSL 46.1.xiii (1950).

³ The Tamil lexicon glosses empi as en tampi, eṇkai as en taṇkai, etc.

⁴ Caldwell (399) suggested that -pi was derived from pin ‘back, rear part, that which is subsequent in time’, i.e. ‘he who is subsequent to me’, etc. He saw the difficulty, viz. the lack of final -n in the kinship words. The presence of pin ‘younger brother’ in the old glossary Tivākaram (Cēntan-tivākaram) is probably nothing but an attempt at an etymology.

⁵ ‘Final short l, n, ḷ y (and m) of original monosyllabic words with short vowels become lengthened when immediately followed by vowels in sandhi contexte.‘ So L. V. Ramaswami Aiyar, Quarterly journal of the Mythic Society 26.173 (1936). He notes that em-, num-, tam-have such doubling.

⁶ Note that tan, singular reflexive, occurs in similar constructions (Vinson, 83) : en tan karpu ‘my chastity’, anuman tan uḪal ‘Hanuman's body’.

⁷ L. V. Ramaswami Aiyar, QJMS 26.94 f.

⁸ Structure grammaticale des langues dravidiennes 24 (Paris, 1946).

⁹ Quoted from Puranānüru 69 by P. S. Subrahmanya Sastri, Tolkāppiyam-Collatikāram (Annamalai University Tamil series, No. 9; 1945) in his commentary on sūtra 415.

¹⁰ This type of simplification of homorganic nasal and stop by loss of the nasal is chronologically prior to the simplification by loss of the stop that was stated in Kota texts 1.18, §11.9.

The alternation of ni and di is to be classed as partial assimilation of nasal (n-) to the stop-type of production that begins the following words (b, d, g).

¹¹ There is no question here of assimilation of ŋg to m before b; contrast eŋg bāynālas ‘my husband's elder brother’.

¹² The phonemic system of Kolami is simplest if word divisions are so placed that long vowels occur only in the first or in the only syllable of a word. Consequently, avreta–kammaner ‘their parents’ has a word division putting a. in the first or only syllable of a word. Comparison of this utterance with other utterances yields ta·k as the word meaning ‘father’, and any possessive pronominal form preceding it is a separate word. The resulting syntactic statements are also simpler than if there were no word division at this point.

¹³ The references for most of the American Indian languages mentioned are to Linguistic structures of native America (Viking Fund publications in anthropology 6; New York, 1946). Algonquian: Leonard Bloomfield, LSNA 96, §32; Chiricahuan Apache: Harry Hoijer, LSNA 75; Chipewyan: Li Fang-kuei, LSNA 401, §12c; Eskimo : Morris Swadesh, LSNA 35; Tunica: Mary R. Haas, LSNA 358, §5:73; Tunica 64, §4.83 (New York, 1940).

¹⁴ Benjamin Lee Whorf, LSNA 171, §7.5; 381, §3:2.

¹⁵ C. F. Voegelin, Tübatulabal grammar 140 ff. (UCPAAE 34.2; Berkeley, 1935).

¹⁶ Morris Swadesh, LSNA 326, §3:42.

¹⁷ A. M. Halpern, LSNA 264, §23; IJAL 12.209, §27 (1946).

¹⁸ Edward Sapir, The Takelma language of southwestern Oregon 231–4 (Handbook of American Indian languages, Part 2; Bureau of American Ethnology, Bulletin 40; Washington, 1912).

¹⁹ Franz Boas, Handbook of American Indian languages, Part 1,45 ff. (Bureau of American Ethnology, Bulletin 40; Washington, 1911); Lg. 5.1-7 (1929).

²⁰ P. O. Bodding, A Santali grammar for beginners 30-2 (Benagaria, India, 1929), records a set of pronominal possessive suffixes which do not distinguish the number of the possessor; they certainly occur with kinship terms and it is possible that they occur only with kinship terms, but Bodding is not explicit.

²¹ D. L. R. Lorimer, The Burushaski language 1.127-39, esp. §§128, 129 (Oslo, 1935).

²² Georg Morgenstierne, preface to Lorimer's work referred to in fn. 21, p. x. It does not seem that Morgenstierne has reported this elsewhere in more detail.

²³ L. Lévy-Bruhl, La possession dans les langues mélanésiennes, MSL 19.96-104 (1914). Before I found this reference, my colleague C. Douglas Chrétien had called my attention to this Melanesian trait. The chief primary sources and discussions are to be found passim in: R. H. Codrington, The Melanesian languages, esp. 128 ff. (Oxford, 1885); Sidney H. Ray, Reports of the Cambridge anthropological expedition to Torres Straits, Vol. 3 : Linguistics, esp. 437 (Cambridge, 1907) ; id., A comparative study of the Melanesian island languages, esp. 69 (Cambridge, 1926). For Micronesian I am indebted to my colleague Denzel R. Carr for an oral communication and the use of unpublished U. S. Navy textbooks.

²⁴ See e.g. A. Capell in Studies in Australian linguistics 59 (The Oceania Monographs, No. 3; Sydney, 1938), and A. Capell and A. P. Elkin, ibid. 99. I am indebted for this reference to Chrétien. The literature on the Australian languages is hard to control, and I must rely on Capell's seemingly exact statement (loc.cit.), the implication of which is that it is only these North Kimberley languages that have this feature. Certainly, Aranda as described by T. G. H. Strehlow (Aranda phonetics and grammar; The Oceania Monographs, No. 7; Sydney, n.d.; reprinted from Oceania 12–4, 1942-4) does not have it.

²⁵ Wolseley Haig, A comparative vocabulary of the Gōṇḍī and Kōlāmī languages, Journal of the Asiatic Society of Bengal 66.1.185-91 (1897).