The etymological dictionary of the Iṟula language should represent my ultimate and closing work on the Iṟulas — a work which began as early as in 1968 with the publication of a short paper entitled ́Iṟula — a South Dravidian Languagé (New Orient Bimonthly, Prague, 7.3.94 - 95). Since then, I have published, apart from a number of papers, three more or less detailed accounts of the Iṟula language and texts with introduction to their social and cultural structure, and a more ́populaŕ account (what is so nicely termed in French vulgarisation) of the Iṟula Community and their milieu.1

1There is something deeply satisfying in the "salvage operation̏ of a culture/language, however small or insignificant it may be seem to be, just like there is something undoubtedly valuable and even moving in the attempts to save some species in the world of plants and fish and birds and animals, endangered by our civilization, a work which is, unfortunately, so very much needed nowadays.

2The Iṟulas are small tribal-demitribal community in the Nilgiri Mountains of southern India, and are — as an independent sociocultural unit — on the verge of extinction. Having survived the acculturation pressures of lowland Hindu Tamils and other Hindu Indians, of Muslim invaders, of British plantation-owners, having lived in the low forests of the Nilgiris for at least two thousand years and perhaps even longer, they will sooner or later, as a separate, individual community disintegrate and their language and culture disappear, due to the unbearable pressure of contemporary civilization. I doubt whether within a few decades one will see or hear any Iṟula at all.

3I was working in the field and at various home-bases on Iṟula language and culture between 1968 and 1988, and sporadically after that, too, and in 2001 was published a large volume referred to in footnote 1. The months spent among the Iṟula in the Blue Mountains repeatedly in 1968, 1976, 1978 and 1981 belong to the happiest days of my life.

4I would like to remember with a feelings of deep gratitude and warm friendship a number of persons whose help, encouragement and assistance made the work on several Nilgiri communities2 possible and even enjoyable. I am much indebted, naturally, to my many informants of these communities, in particular to P.Sivaraj of Kunjapene, born in 1953, son of Puliyan of the Kuppä kula, my chief Mele Nāḍu Iṟula informant. I owe special thanks to Mr. J.D. Rajiah, retired deputy Tahsildar, Ootacamund, who was my interpreter and assistant during two of my fieldworks, particularly with the Todas and other tribals; to Prof. Dr.D.B. Kapp, University of Cologne, for his hospitality and many friendly discussions with a colleague who loves the Nilgiris as much as I do. I think with affection of Dr. Saskia C. Kersenboom, my former student, for her encouragement and care. I am also grateful to Mrs. Kokilam Subbiah and to Evam Devika Beryl Pilijean-Wiedeman; Evam introduced me in a friendly and profitable manner to the Todas and their customs. To Professor Murray B. Emeneau I am indebted in innumerable ways, particularly for his critical and friendly comments. The more I plunge myself in Emeneaús work, in particular his Kota Texts (1944–46), his Toda Songs (1971) and his Toda Grammar and Texts (1984), the more increases my admiration and gratitude for his immense contribution to Indian, particularly Nilgiri studies. Last but not least I wish to thank my wife Nina for her understanding and generosity.

Nilgiri Areal Typology

When I first began gathering the data in the Iṟula settlements in the Nilgiris, I did not have any clear vision of an "areal̏ approach to the region of the Blue Mountains. After my first data on Iṟula had been gathered and to some extent evaluated, it appeared clear to me that there must have been two tendencies in the development of the language: one, of Iṟula diverging from some common Proto-Tamil (?) ancestor and source, another, of Iṟula converging with its neighbours, Kuṟumba, Muḍuga, Toda, Kota, Badaga, and may be with yet other "undiscovered̏ tongues. After having studied Emeneaús path-breaking article of 1956 (Language 32.1.2-16) and the elaboration of that vision by a colleague and friend, the late A.K. Ramanujan and another colleague, C. Masica and, in particular, Emeneaús excellent article ́Diffusion and Evolution in Comparative Linguisticś (1965), I have adopted, from about 1980 on, the structural and areal approach to the linguistic, anthropological and cultural phenomena of the Nilgiri area. Very kindly, M.B. Emeneau wrote, in 1989, ̏Nilgiri areal studies have been begun by Zvelebil (1980), and already show interesting results̏.3 And again, in a paper on ́Linguistics and Bottany in Nilgiriś,4 "This essay in general owes much to Zvelebiĺs proposal of treatment of the Niligiris as a linguistic microarea… The microareal treatment was begun by Zvelebil in 1980 (Emeneau 1989 and Zvelebil 1990).̏

5The Iṟula Dictionary should have comparative and etymological character, with strong emphasis on areal (or ́microareaĺ) features of the Nilgiris. The entries are compared with corresponding lexemes of Ālu Kuṟumba and Badaga (since for these languages, excellents sources are available thanks to D.B. Kapp and P.Hockings–Ch.Pilot-Raichoor), but also to some extent with Bëṭṭu Kuṟumba, Jēnu Kuṟumba (alias Kāḍu Nāyika), Shōlega and Paṇiyan.

6I would like to mention below several of these areal features. I believe (on the ground of some data gathered thus far) that we are entitled to speak not only of Nilgiri areal linguistics but, more generally of Nilgiri areal typology, including features of culture (material culture as well as e.g. mythology). In somewhat more precise terms, I would like to designate the region in question as the Nilgiri multiethnic and multilingual convergence area.

Iṟula Dialects

I argued the independent status of the Iṟula language several times, lastly in my 1982 publication.5 I have no reason to change my opinion expressed then: Iṟula is a "separatȅ preliterate South Dravidian language, a close relative to Tamil, with a few archaic features and a few striking innovations. Ëṟula na: ya or Iṟula language manifests four different dialects, exemplified well by the Ir. words designating "red colouȑ:

• Mele Nāḍu (MN) Ir. kanape, kenape (DEDR 1931, cf. e.g. Kota ken, Ka. kempu)
• Vëṭṭe Kāḍu (VK) Ir. ratta (cf. Ta. irattam < Skt. rakta-)
• Ūrāḷi cenga (DEDR 1931, Ta. etc. cem- < * kem)
• Kasaba kempu (DEDR 1931, e.g. Ka. kempu)

The stemma indicating the dialect division of Iṟula may be written out as follows:

Since the main body of the entries in the etymological dictionary contains items current in the MN Iṟula dialect, these forms are, for the purpose of this dictionary, regarded as "unmarked̏. This procedure is, in point of fact, in agreement with the actual overall importance and very probable numerical predominance of Mele Nāḍu Iṟula (s) over the other three dialect communities of the language . Strikingly diverging lexical entries current in the other three dialects are marked as such in the dictionary: VK = Vëṭṭe Kāḍu, Kas. = Kasaba, Ur. = Ūrāḷi. Whenever it is necessary to indicate a Mele Nāḍu form it is marked MN.

The Iṟula Phonological system

The language in its four dialects known to date contains 23 contoids, 10 vocoids, and 5 additional sounds.

Contoids

The nasal phonemes are m, n, ṇ and ṅ [უ]. Palatal nasal is a predictable variant of /n/. /n/ is thus a single phoneme with three phonetically conditioned positional variants: dental to postdental in the neighbourhood of t, d, and initially; dento-alveolar in the neighbourhood of ṯ, ḏ and intervocally; palatal in the neighbourhood of c, j. /ṅ/ has contrastive function in a limited number of items, e.g. maṅa son: mana mind. Another phoneme with extremely restricted occurrence is the domal glide /.y/, e.g. va.y to wipe, rub, scrape. The contrast of /r/: /ṟ/ is not equally valid in all the dialects. When fully preserved (as with most Ūrāḷi speakers), /r/ is a dento-alveolar tongue-flat trill, whereas /ṟ/ is a back alveolar tongue-retracted trill, e.g. ve: ru root: vë: ṟu another.

Vocoids

Length is contrastive, e.g. nila moon: ni: la blue.

Additional sounds

7s – in Indo-aryan loanwords, rarely in other loans

8ṣ – in IA loanwords, rarely in other loans

9ñ – in a few items in dialects when not positionally conditioned and hence unpredictable, as in Ur. päñḏi pig.

10r. – in Ta./Ma. loanwords and when imitating Ta. pronuncation, e.g. Tamilized Ir. väṛ.i path

11w – in a few items in several districts, e.g VK oggwe mother.

Long consonants are treated as sequences of identitical sounds, e.g kallu [kál: ʉ] stone. /u/in most unstressed positions and in almost all final positions has the predictable variant [ʉ], a somewhat centralized, retracted, only partly rounded back high vowel, e.g. irundu [írʉṉḏʉ] having stayed. /c/ is almost predictable either as [s] or [š] in loanwords, e.g. ca: mi [sá: mi] lord, god, kacta [kásṭa] / [kášṭa] trouble.

Length is indicated by /: /; occasional nasalization by /͠/.

Sources of the Dictionary.

As mentioned above, a very considerable majority of the entries in the Dictionary were provided by my chief informant, P. Sivaraj of Kunjapene, son of Puliyan, of the Kuppä Kula, Mele Nāḍu Iṟula tribe, born in 1953. Hence, the majority of all entries are MNIr lexemes, and since they form the bulk of the Dictionary, and since the MN dialect is the Iṟula sociolinguistic "standard̏ (economically, socially and culturally the most important of the four dialects), these entries are as a rule unmarked. Apart from the data provided by P. Sivaraj in 1981 specifically for the Dictionary, the following sources were used:

1. The two glossaries appended to Zvelebil, K.V., The Iṟula language, Wiesbaden, 1973 (indicated if necessary as ILI-1) and The Iṟula (Ëṟla) Language – Part II, Wiesbaden, 1979 (indicated as ILI-2).
2. Lexical items occurring passim in Zvelebil, K.V., The Iṟula (Ëṟla) language – part III, Wiesbaden 1982, and in Zvelebil K.V., The Iṟulas of the Blue Mountains, Syracuse University, 1988.
3. Zvelebil, K.V., "The Body in Nilgiri Tribal Languages̏, JAOS 105.4 (1985) 653–74.
4. Fieldnotes gathered during fieldtrips in 1968, 1976 , 1978 and 1981 with the view of compiling a comparative dictionary of Nilgiri languages.
5. personal communication Prof. M.B. Emeneau and Prof. D.B. Kapp.
6. Burrow, T., and M.B. Emeneau, A Dravidian Etymological Dictionary, 2nd ed., Oxford, 1984.

Various layers of Iṟula lexis and instances of areal features

The "Wortschatz̏ of Iṟula as reflected in my lexical data does not fundamentally contradict the hypothesis that a pre-Dravidian (?) Melanid population which forms the bulk of the Iṟulas anthropologically adopted (or was made to adopt) an ancient pre-Tamil or proto-Tamil dialect which was superimposed almost totally on their native (pre-Dravidian?) speech, and that this speech-form became the core of the Iṟula language as we know it.

12To go yet deeper into the the past, there exists some very striking and perhaps, originally, rather close connection between Nilgiri Iṟula, Ūṟāḷi and Shōlega, and a yet deeper specific relationship between Iṟula, Muḍuga, Shōlega, and Ālu and Pālu Kuṟumba (probably some other Kuṟumba languages included).

13The Various layers of Iṟula lexicon may be symbolized (in a simplified manner) by the following sketch:

In this sketch,

? represents ́enigmatic itemś, "deep down̏ in the lexicon, without any etymological connection: substratum?

1. items occuring only in Iṟula, so far without convincing etymology.
2. items found only in Iṟula plus one more Tamil and/or Malayalam dialects; also items occurring only in Iṟula and Old Tamil (or rather, some early form/s of Tamil) ± sporadically in some other Dravidian language.
3. items found only in Iṟula and or more Kuṟumba languages (Ālu and Pālu Kuṟumba, Muḍuga, Jēnu Kuṟumba, Bëṭṭu Kuṟumba) and Shōlega. This is an important layer of the lexicon since I propose a specific Iṟula-Kuṟumba-Shōlega linguistic plus cultural relationship (this cultural relationship being convicingly demonstrated in mythology).
4. items found in Iṟula plus some other language (s) of the Nilgiri area (other that Kuṟumba and Shōlega only): the proper realm of Nilgiri areal linguistics.
5. items found in Iṟula, one or more Tamil and/or Malayalam dialects, some others South Dravidian non-literary language (s), and some other SDr language.
6. items with plausible SDr or Dravidian etymological connection, with or without typical Iṟula developments: the majority of Iṟula vocabulary.
7. Peripheral layer of obvious borrowings from Badaga, and/or from non-Nilgiri language (mostly Tamil, Malayalam or Kannaḍa, but also from Indo-Aryan and English).

Items ́beloẃ the double line go back to a hypothetical proto-Iṟula-Kuṟumba-Shōlega entity; items ́beloẃ the interrupted line go back to a hypothetical Nilgiri microareal linguistic reservoir.

To quote a few illustrations of each of these seven groups:

? – ḍëkëṇe, ḍökëṇe, dekkaḍa (VK) panther; buṇḍri (VK) grasshopper: rïe-, rïa-, rï: ya- a kind of edible tuber (rïekaṅgu, rïyapu: etc.); a: ḷi small vegetable garden, small plot of cultivated land.

1. vënḍi, vëndi idol of deity (cf. DEDR 5530? 4438? 5529?); anducu beauty (cf. DEDR 2328?); omuda, omeda heap (cf. DEDR 5065?); a: re digging stick (cf. DED 371?); however, this item goes more probably with 4094 Ta. pārai "crowbar, small hoȅ; Ka. pāre short hoe, etc.
2. meguru, megaru (VK), moyiru, möyiru (MN) hair: Yerukala mogurú, Korvī magara, Kaikāḍi magri (DEDR 4704): retention from pre- or proto-Tamil * makVr; algane: ra, algapo: du evening: DEDR 235 old Ta. al night, darkness, evening, alku night, afternoon, alkal night; found also in Malayalam and Kuṛux. Iṟula has preserved the old Ta. form in current use.
3. mattu (etymology?) lip(s) in all Ir. dialects: Pālu Kuṟumba, Muduga mattu; ja: ya (etymology?) trance, possession by deity, demon, divination: Ālu Kuṟ. já: ya [dza: ya] Prophezeiung; muṅä [muუæ], maṅe [maუε] son (DEDR 4616): Ālu Kuṟ. maṅa Sohn; Jēnu Kuṟ. maუən; Bëṭṭu Kuṟ. maṅənu, Shōlega maṅa [maუ^].
4. kunni, künni bee (DEDR 1867): Ālu Kuṟ. kúnni, "Insekt, bes. Bienȅ, Pālu Kuṟ. kunni bee, Kota kuṇy id.; bugari, bugiriya a large bamboo flute (DEDR 4239): Ālu Kuṟ. búguri bamboo flute, Pālu Kuṟ. bugiri id., Kota bugi: r flute, Toda puxury Toda flute, Badaga buguri id.
5. maṇi to talk, speak (DEDR 4671): Ko. mayṇ-, Tuḷu maṇipuni, manipuni; mä: rain (DEDR 4753) Ta. mar.ai, Ma. mar.a, Ko. may, To. maw, Ka. mar.e, Koḍagu maḷe, Tuḷu maḷe: South Dr. and Tulu * mar.ay.
6. mëkku wax (DEDR 5082): Ta mer.uku cow-dung, wax etc. through to Brahui; i: ppi fly (DEDR 533): Ta. ī fly, bee etc.through to Brahui; pammu, pa: mu ripe fruit (DEDR 4004): Ta. par.am etc. through to Malto; mele mountain (DEDR 4742); Ta. malai etc., through to Brahui.
7. la: ba profit (< Ta./Skt.); badilu reply (< Ta./Urdu); sku: lu school (< Engl.).

There is also a small number of items occurring (as far as our knowledge goes) only in Iṟula, Tamil-Malayalam, and one of the Kuṟumba languages: Ir. ku: re house (DEDR 1904): Ta. kūrai sloping roof, small hut, Ma. kūra hut, thatch, Bëttu Kuṟ. ki: ri house.

14Some of the enigmatic items may ultimately be connected with known etymologies; thus a: ḷi small lot of culivated land may be perhaps connected with DEDR 4112 Ta. pār.i temple, town, city, town of an agricultural tract, Ko. va: y < pa: y fields nears village which are ploughed in the sowing ceremony, Ka. ha: i field near village.

15The areal features, i.e. features of phonology, morphology, lexicon, semantics (and even some mythemes) which define the Nilgiris as multiethnic and multilingual convergence area were discussed in some detail in my most recent publication Nilgiri Areal studies, Charles University in Prague, 2001, Introduction, pp.21–33. It has been demonstrated, e.g., that centralized vowel phonemes form part of the phonemic inventories "of Toda, Iṟula, Ālu Kuṟumba, Muḍuga, Jēnu Kuṟumba, Shōlega, Bëṭṭu Kuṟumba and Paṇiyan̏,6 i.e. of eight Nilgiri languages. Another areal feature, occurring in Kota, Toda, Iṟula, Ālu Kuṟumba, Bëṭṭu Kuṟumba and Paṇiyan, is the three-way contrast among dentals, alveolars and retroflexes in addition to labials, palato-alveolars and velars.7 There are yet other unique and peculiar phonological developments found in some Nilgiri languages. A few features could be quoted from the realm of morphology, and we know at present a considerable number of lexical units with only Nilgiri etymologies and no related etyma elsewhere, and/or which may be regarded as Nilgiri areal words.8 Even several semantic features seem to be specific for at least some Nilgiri languages.

Arrangement of entries

The order of entries is that of the Devanāgarī alphabet, with letters and diacritics added in agreement with the Iṟula phonological system in the following sequence: a a: i i: u u: e e: o o: ai au ä ä: ï ï: ü ü: ë ë: ö ö: k g ṅ c j ṭ ḍ ṇ ṯ ḏ t d n p b m y r l v ṟ .y ḷ w s ṣ ṛ.

16Double (or more) phonological shapes of the same entry are due to dialectal differences. The assignment of word-classes is purely functional; the labels designate functional "parts of speech̏ as valid in contemporary Iṟula speech.

Several model entries

I would like to include here just eight "model̏ entries to show that almost every entry is linguistically and culturally interesting, relevant or even important. The entries were chosen at random from the data under preparation.

1. ákkara, ákkarä n. ear-wax (VK); kà: d’ákkara (VK) ear-wax (specif.). ILI-2, 1. Cf. Ālu Kuṟ. akáṟa "Kopfschuppen̏; Bad. akkure dandruff, scurf; Vēṭṭuvar (S. Bhattacharya) kaaram ear-wax. Cf. DED S 55 Ta. acaṟu dandruff, scurf; scab in sheep and goats; ayaṟu excrescence resulting from a sore; akir a kind of scurf. Ka. agaru, hagaru, adaru, aduru dandruff. Another possible connection: with DEDR 3811 Ka. hakku crusted or dry mucus or rheum, scab, Ta. pakku scab of a sore, dry mucus of the nose, etc. It is possible that ultimately all these items are related (? ** pak-V- > * hak- > ak-) cf. Te. pakku scab, Ga. (S³) pakku dried portion of any bodily secretion, scab.
2. a: yiraka: la künni n.cpd. is interesting for several reasons: first, it is a compound of three members; the meaning is "centipede", lit. "insect of one thousand legs̏. Second, more importantly, it shows that its last component, künni (Cf. DEDR 1867 kunni, Bhattacharya, 1958; ex. DED{S} 1403) has a broader and more general meaning than the one given in DEDR, i.e. "beȅ. It seems that, in fact, the proper designattion of ’ (honey) bee’ is te: nukunni / të: nukünni (which is attested in my data) while kunni / künni designates "insect̏ in general or, in fast speech, may be used for the more precise te: nukunni. Finally, it is also interesting that we have here a compd. of a: yira derived ultimately from IA (DEDR App.11 /Skt. sahasra-) "thousand̏ with Dr. ka: lu (DEDR 1479) ’leg’ and Ir. kunni / künni, so far without etymological connection, strictly a Nilgiri word, found in three Ir. dialects, in Pālu Kuṟ. kunni and Kota kuṇy.9
3. a:ṟe, a:re n. (simple) digging stick with sharp end. This is obviously a very important tool used in gathering tubers and roots, and in the swidden fields. Its picture occurs in Zvelebil, 1982, 97.10 Sometimes it is called gaḍepare cpd. etc., however, this tool is rather a pick-axe (see below). According to a Mele Nāḍu myth (ibid. pp.228–9), it originated from a veḷḷi taḍi silver stick, given to the earliest Iṟulas by Paramaśiva. According to a comment on the Central Ūrāḻi Myth (ibid. 230 ff.), the Shōlegas told the Ūrāḷis how to dig tubers with digging sticks (pa: re). This comment offers its etymology: DEDR 4093 Ta. pārai crowbar, small hoe for cutting grass, Ma. pāra iron crowbar, lever, bar used for digging, Ko. pa: r, To. pa: r "crowbaȑ etc. DED (S) 3367. It occurs also in Bad. cf. a: re / ha: re n. crowbar (made by Kotas), and in Mala Adiyar, Mala Ulladan tribal speeches as pāra wooden digging stick. The Iṟula word may be a loan from Badaga. gaḍepare (transcribed by me previously incorrectly as gaḍepaṟe and etymologized incorrectly as DEDR 984 + 3317) is a compound to be compared with Ālu Kuṟ. gāḍḍa-pā: re "Eisenstab mit zugespitztem vorderen Ende (wird zum graben verwendt) ̏ (D.B.Kapp). The first part of the cpd. remains a problem. I would prefer a connection with 984 Ka. gaḍḍe, geḍḍe "any bulbous root̏, but 1148 Ka. gaḍḍe "a mass, lump, concretion̏ is also a possibility, though more remote. The second component is certainly 4093 as in a:re, a:ṟe. The short vowel [ă] occurs because it is in the second member of a compound; the - ṟ - in Irula form of 4093 may be a mistake on my part. Thus the Ir. term for this sort of pick-axe/digging stick could possibly be glossed as "tool used for digging bulbous roots̏.
4. a: ḷi n.small (private?) vegetable garden; small cultivated field. This word designates a small area of cultivated land gained by the slash burn process in the jungle; a small forest-field or forest-garden (cf. Zvelebil, 1982, 101-2; W.A.Noble 1978).11 This very frequent, current term presents an etymological problem. There may be possibly a connection via * ha: ḷi < * pa: ḷi < ** pa: r.i with DEDR 4112 Ta. pār.i town of an agricultural tract, Ko. e: r iṭ va:y (< pa:y) fields in village which are ploughed in the sowing ceremony, Bad. ha: i / ha: yi "farmland near a villagȅ (BED 581; Emeneau in Lg. 15.45, 1939, ha: l "field near villagȅ). DEDR 4112.
5. ebbukaṭṭe the typical Nilgiri plant of Strobilanthes, conehead, the term found in various shapes in Iṟula, Ālu Kuṟumba, Toda, Kota and Badaga; it is one of the most convincing areal etymologies. The Ir. term is a compound, since Ālu Kuṟ. has kaṭṭe for any Str. species as well as ebbu-kaṭṭe "üppig blühende Strobilanthes Art̏ (D.B. Kapp). For Toda, Emeneau has recorded kaṭ, cf. further Kota kaṭ and Bad. kaṭṭe. Bad. has in addition also hebbukaṭṭe, plus a number of toponyms connected with this plant. Emeneau has provided the original form, DEDR 1154 * kaṭṭay. As for the Bad. hebbu- and Ir. and Ālu Kuṟ. ebbu-, D.B. Kapp has come up with a very plausible etymology for this part of the compound, namely * per-pu: -, i.e. ’big blossom’, so that, finally, the Ir. ebbukaṭṭe would correspond to * per-pu: -kaṭṭay.
6. ga: vu n. blood sacrifice; ritual sacrifice; then, sacrifice in general. This is culturally a most important term, while it is easy to come up with its etymology. Cf. Ālu Kuṟ. gá: vu n. Opfer, (bes.) "Tieropfer (zur Besänftigung böser Geister…) ̏ (D.B. Kapp). This is a loan from Ta. kāvu "sacrifice, oblation to inferior deities̏ (Tamil Lexicon), cf. Pkt. ghāu-kāma- desirous of striking (CDIAL), Skt. ghātuka- killing. In terms of etymology, then, the Nilgiri words go back rather to the Prakritic form. In the Iṟula indigenous cult, the local place of worship contains usually two or three flat stones; the tallest represents the deity, the second, smaller one (if present) the deity’s śakti, and the third stone, ga: vukallu, represents the ga: vutoga, lit. "the deity of the ritual (bloody) sacrificȅ. As a rule, if an animal is slaughtered in sacrifice, its blood is offered only to the ga: vutoga, i.e. the blood is poured over the ga: vukallu. The largest/larger stone is considered to be the cele idol, the image of the deity to which only nonbloody sacrifice is offered.12
7. mattu n.lip; cf. Pālu Kur. mattu, Muḍuga mattu id. There is no DEDR entry for these words for ’lip’. This word of high frequency has so far no etymological connection, is thus a real Nilgiri isolate, and reflects perhaps a Proto-Iṟula-Kuṟumba word (although the Ālu Kuṟ. word for "lip̏ is duḍi DEDR 3296).
8. A: negaṭṭi n. cpd., name of a Mele Nāḍu and Kasaba hamlet below Nilgiri slopes, 16,5 km NNE of Ootacamund, Ootacamund Taluk, revenue village Ebbanad. Tribal residential school. In existence before 1812. Another A: negaṭṭi refers to a former Iṟula hamlet 9 km ENE of Kotagiri (the site has old elephant tethers). DEDR 5161 +1148. Southern dialects of Ir. prefer the form ä: ne, even e: ne for "elephant̏; gaṭṭi connects with Ta. kaṭṭi, Ka. gaṭṭi anything hardened, lump, clod. Hence, the meaning of the local name seems to be "solid lump of elephant dung̏. This may sound bizarre, but elephant dung is of considerable importance for the tribals. The name apparently points to (former?) density of elephant population in the area. Cf. Zvelebil, JDL XII, 1 (Jan. 1983) 48. Engl. versions of the name: Anaikatti, Anekatti, Anna Kattie, Anehatti. Bad. A: nekaṭṭi. The word for elephant (a: ne, ä: ne, e: ne, DEDR 5616) appears in several Iṟula local names, cf. also A: nepaḷḷa Elephant Valley (Coonoor Taluk), A: nekëyi "Elephant Trunk̏ (east of Coonoor) and A: nemele "Elephant Mountain̏ (5161 + 4742), Bad. A: nemale Beṭṭu, name of a mountain c. 2150 m in elevation, 31,5 km SW of Ootacamund.

The proposed Iṟula etymological dictionary would thus be not only a fundamental contribution to Nilgiri and Dravidian linguistics, but would also contain a wealth of cultural material, pertaining to the Nilgiris and to South India.