page 122 note 1 Caldwell ³, pp. 150–1.

page 123 note 1 Originally *kiltu and derived from kil>cil above. The change -It > r is well known in Tamil external sandhi (Vinson, Manuel de la langue Tamoule, p. 33), and the same change also takes place in the middle of words as here. The final -tu is to be explained as the 3rd sing. neut. pronominal affix, which is appended to the Drav. adjective in certain syntactical relationships (when used predicatively and to form adverbs). In this case, as a result of the sandhi change it has been incorporated with the original adjective to form a new word. Nevertheless its origin is betrayed by a peculiarity of its grammatical treatment in Tamil. Whereas other adjectives have a nasal added to them when used attributively before words beginning with an occlusive, e.g. perum pulia big tiger, kuṟun tokai a small collection, etc., such is never the case with ciḷu which is always construed without a nasal, e.g. ciḷu kuti a little house, ciḷu paḷai a little drum, etc. The account of its origin given here accounts very well for this anomaly.The commonest word for “small” in Kanarese is cikka. Kittel derives this from ciṟu (= kiṟu), comparing for the sufix Ma. ceṟuklcan boy, lad. This cannot be correct because k- does not become c in Ka. except in rare dialectal forms. Thus ciṟu occurs rarely, probably due to the influence of Te. or Ta., but the regular and usual form is kiṟu. In the case of cikka, on the other hand, no trace of k appears, so that c- here must be regarded as original, and the word can have no connection with ciṟu<kiṟu. The word cikka has been borrowed from Ka. into IA. See J. Bloch, Formation de la Langue Marathe, Index s.v. cike a little, cikkar small. Earlier it appears Skt. cikka small, in the compund cikkagaja a young elephant, and Pkt. (Deśī. iii, 21) cikkā a small thing, but it does not seem to have had any currency outside the Deccan.For the same reason Ka. cillu smallness, cillaṟa small trifles, odds and ends, can only be connected with kil-, kiṟu, etc., if we assume them to be borrowed from Telugu or some palatalizing dialect. Te. has cillara in the same sense as Ka. cillaṟ and may be the origin of it. Ta. cillaṟai is not old, and obviously borrowed form Ka. or Te.J. Bloch is inclined to derive Mar. celā pupil from De.: Ka. cillaṟa, cillu, ciṟu, smallness, Ta. cila a few (Formation ⃛, Index, s.v.). This attempt is to be rejected in view of the fact that Dravg. cil represents earlier kil in these cases.Another word to be separated form ciṟu<kiṟu is Te. cinna small. This is rather to be compared with Brah. cunak small, cunā a child. In this case we have primitive Dravidian c-, not k-. The words are further to be compared with Beng. cunā, cunī small, Sh. cũDṇŭ which Turner (N.D. s.v. cun) connects with worids derived form Skt. cūrṇa, Pkt. cuṇṇa powder. They may, however, be lws. <Drav. Santali cuni small appears to be derived from Bengali (Bodding, A Santal Dictionary s.v.).Ka. saṇṇa small appears to be derived from Pkt. saṇha. Master, A. (JBBRAS, V (1929), p. 110) is inclined to regard the borrowing as having been the other way. Against this it is to be observed that the word is more widespread in Indo-aryan (N.D. s.v. sānu) than it is in Dravidian, where it is confined to Ka. and Te. (Kur. sunnī small cannot be native Dravidian because sdoes not occur in that language in native Dravidian words). Further the Indo-aryan etymology of the word-Skt. ślakṣṇa > Pkt. saṇha (> Ka. lw. saṇṇa) > Nep. sānu, etc., is unobjectionable.+Pkt.+saṇha+(>+Ka.+lw.+saṇṇa)+>+Nep.+sānu,+etc.,+is+unobjectionable.>Google Scholar

page 123 note 2 This is to be distinguished from the following forms, onomatopoeic in origin, which had not original k: Ka. siṭṭu anger, which is probably equivalent to Ta. cīṟṟam, though somewhat irregular in form; cf. cīṟu to hiss. Telugu has ciṟṟu, ciṟṟa angeṙ, and with it can be compared Brah. ciṛing to be annoyed, Kur, ciṛārnā to lose one's temper, fly into a passion. With these latter some IA. words are to be compared. Sir Denys Bray compares Jaṭkī, sindhi chiṛwith the Brah. word; cf. also Hi. chirnā, etc. In Dravidian again we can compare Ka. seṇe to be angry, seṇasu wrath.

page 124 note 1 Attempts to derive Skt. Śiva n. of the god from the Drav. word for “red” (Konow, , LSI., IV, p. 279Google Scholar; Chatterji, S. K., Origin and Development of the Bengali Language, I, p. 41, etc.) are to be rejected in view of the fact that the cis only a late development of kin three of the southern languages. The same objection applies to attempts to derive Skt. Śambhu, a name of Śiva from Ta. campu *redness, copper (Chatterji, loc. citGoogle Scholar, Collins, M., Dravidic Studies (University of Madras), III (1923), p. 61, etc.)Google Scholar

page 124 note 2 Lws. from this in IA. would seem to be:Hi. kharpā sandal (<*kharappa: Ka. kerahu <*kerappu), and from a palatalizing dial. Mar. sappaḷa sandal.

page 125 note 1 A problem is presented by the Kui and Kuvi forms in which -r appears. Ramaswamy Aiyar in drawing attention to this (educ. Rev., 37 (1931), p. 552) compares to the Kui forms Tu. kirmbilụ ear. With Tu. kirmbilụ we can compare also Ta. kuṟumpi ear-wax, Ka. kuṟugaṇi, kogaṇi, gugge, Te. gubili, gulibi, gulimi, Go. koranged, Kuvi kirpejja. There is a certain doubt about the connection between these words with the Kui forms of the word for ear because of the vowel of the first syllable which form the majority of the forms seems to have been -u, and also because the Tamil from can be quite naturally explained as the Tamil Lexicon explains it, namely as kuṟu small + pī excrement. The latter element is in any case pī which has this secondary sense generally in Dravidian (cf. Ma. cerippī ear-wax, Kur. pīk in both. senses, etc.). If Ramaswamy Aiyar's explanation is to be retained, the first element in a form like Ta. kuṟumpi must be regarded as having been altered by popular etymology, and on the whole I am inclined to regard this as being the case. The same element appears also to be found in Ta. kuruṭu tragus.Tu. ceraḁu ear is isolated in Dravidian. Initial k is occasionally palatalized in Tulu, which is somewhat of a mixed language, and anyway ceraḁu is to be regarded as a dialect form since the standard word for “ear” in Tu. is kebi. Tu. ceraḁu is therefore to be regarded as representing primitive Dravidian *keraṭu and though so far it has not appeared in any other Dravidian language, we have in Sanskrit a word karaṭa an elephant's temple; the spot between the forehead and ear of a bird, which is obviously the same word and is to be regarded as a lw. fromDravidian. This root ker with the rather vague sense of “cheek, temples, ear” (cf. Ma. ceppi “ear” and “cheek”, koviṭu cheek which is a variant form of *keviṭu which appears in Ma. as ceviṭu ear) appears in a number of derivative words in Drav.: Ma. kenni, cenni temples, Ka. kenna the upper cheek <*kerni, *kerne, Te. carãpa cempa cheek, Te. cekku, kekkili, Ka. kekke <ceppi) <*kervi or kerpi*. Skt. karṇaear, which has no Indo-European etymology, is not doubt borrowed form an early Dravidian *kerne.

page 125 note 2 Cf. Ma. cuva = Ta. civa to be red. It is to be observed that Ta. cuvar is pronounced cevar in the colliquial (Pillai, R. P. Sethu, Tamil literary and Colloquial, p. 10: Annals of Oriental Research, University of Madras, II (1937–1938)Google Scholar).

page 126 note 1 The first mention is in Mahābhāsya on 4.1.175. The Vārttika here mentions Coḁa with an -ādi that includes the Keralas whom Pat. takes the trouble to mention. According to the usual view Patanjali is a century or two later than Aśoka, Kātyāyana roughly contemporary. The references in the Mahābhārata which are not numerous, belong to the later stage of the compilation of that work, and are not likely to be earlier than Patanjali.

page 126 note 2 An attempt has been made to see in the term Cerapādāḥ, which occurs in an obscure passage of the Aitareya Āraṇyaka (ii, 1, 1. See Keith's note in his edition: the theory has been adopted by Iyengar, P. T. Srinivasa, History of the Tamils, p. 29Google Scholar, and others). There is no foundation for this suggestion, which is based on a misconstruction of the words in question. The passage runs: uā vai tā imāḥ prajās tisro atyāyam āyaṃs, tānîmāni vayāṃsi vaṅgā vagadhāś cerapādāḥ “The three races that transgressed were these birds-vaṅgas, vagadhas, and cerapādas”. The words vaṅga-, vagadha, and cerapāda are the names of three species of birds not otherwise mentioned. Since the actual number of bird-names current in the Brāhmaṇa period must have been far larger than the comparatively small number of which record is preserved, the fact that these words turn up nowhere else need not be a cause of surprise. Nor need there be any doubt about the words being bird-names, since the text if construed as above definitely states that they are such. The error of the commentators, followed by the modern interpreters, is to regard vayāṃsi birds as one of the three prajāḥ instead of a general them describing the three which are mentioned after it; and, as a result of this, to turn vaṅgā vagadhāś which is actually two nominative plurals, into a compound expression (vaṅga- + avagadha). From this have sprung other errors, including the attempt to see in the passage the names of peoples or tribes.

page 126 note 3 This rule is formulated by Aiyar, L. V.Ramaswamy, Educ. Rev., vol. 37 (1931), p. 553Google Scholar, and Thyagaraju, A. F., IA., vol. 61 (1932), p. 142Google Scholar.

page 127 note 1 E.g. KuT. 145 tuṟai keḷu ciṟ kuṭi the little house near to the ford, 170 maḷai keḷu nāṭaṉ lord of a land possessed of mountains. The Tamil Lexicon's treatment of this word is inadequate. It calls it “an euphonic increment” without defining the meaning. Actually it is the stem of the verb which appears in fuller form as keḷumu is naturally merely a phonetic variation of keḷuvu.

page 128 note 1 Appendix to Arden's Tamil Grammar (4th ed.), p. XXII.

page 130 note 1 Early Tamil has also tīrvai mongoose which appears isolated. Another set of words in Dravidian is: Ta. mūṇkā a species of mongoose (Tolk.˚), Ka. muṅgisi, muṅguri, Te. muṅgisa, cf. Pkt. mangusa, muggasa, Sgh. mugaṭ.

page 130 note 2 The ultimate origin of this word is no doubt to be sought in South-East Asia (Thomas, F. W., JRAS., 1905, p. 169Google Scholar). On the other hand the IA. forms seem to be borrowed directly form a Dravidian source, on account of the second element ˚vera which undoubtedly represents Dr vēr root.

page 131 note 1 for the omission of the liquid before the consonant group -ṅk in Skt., cf. Ka. neggil as opposed to Ta. neruñcil, and for final -ola corresponding to Ta. -il, cf. Ma. ñeṅๅōl plough, as opposed to Ta. ħāñcil.

page 131 note 2 The Drav. forms probably represent original *nerhi or *lerki. The irregular I in Ka. which is not allowed initially in proper Drav. words, suggests that the word may have been borrowed originally from some non-Dravidian source.

page 132 note 1 Cf. Tuttle, E. H., “Dravidian Researches” (AJPh., 50, 152Google Scholar) and “Dravidian Developments” (Language Monographs, v), $ 44; Aiyar, L. V. Ramaswamy, Ed. Rev., 37 (1931). pp. 545–8Google Scholar; Sir Bray, Denys, The Brahui Language, II (1934), pp. 17–18Google Scholar.

page 133 note 1 A number of Skt. words are derived from this Dravidian family: Skt. kalaṅka stain, impurity, kaluṣa turbid (=Ta. kaluḷ), and kalaha quarrel.

page 133 note 2 Skt. kaṣāya astringent, Pa. kasāya, kasāva, and Pa. kasaṭa, bitter, acrid, nasty, are derived from this Drav. root. There seems, however, to have been a confusionof two Drav. bases in Skt. kaṣāya (kāṣāya) because the meaning “reddish-brown” as applied to Buddhist robes cannot be connected with the meaning “astringent”, and its origin is probably to be sought in Dr. kes-, kēs red, the various forms of which are listed above.

page 133 note 3 Cf. Bloch, J., BSOS., V, 741Google Scholar.

page 135 note 1 The same idea is behind the etymology of Skt. aravinda. Cf. Te. araviri a flower half opened, Ka. areviri to half open. Skt. ˚vinda is a contracted form of what appears more fully in the Ta. part. virinta. the idea is “opening half the time”.

page 135 note 2 The IA. words for “mast” are derived from this: Skt. Pa. kūpa-, kūpaka, Pkt. kūva, kūvaya, kūvaga, Sgh. kumba.

page 135 note 3 All derived from a root meaning “young, fresh, tender, new” which appears also in the following words: Ta. kuḻa young, tender, kuḻaku youthfulness, beauty, infant, kuḻakaṉ a youth, a beautiful person; Skanda, kuḻantai infant, babe, childhood, kuḻavi infant, babe, young of certain animals, young of the vegetable kingdom, kuḻavu youth, tender age, Ka. koda tenderness, tender age, youth, koṇasu a young one of wild beasts, Te. koḁuku a son, krotta new, fresh, krov-viri a fresh flower or blossom. IA. words derived from this are Skt. kuṇaka a young animal justborn, Khar. Doc. kuaǵa small boy, kui girl, Pj. kuṛī; girl, etc. Cf. also Sgh. kolu a lad, koli a girl (-1 ?). Sant. Muṇḁ kuṛi girl may have been derived from Dravidian through Indo-aryan.

page 136 note 1 DS., i. It thus provides additional evidence for the conclusions reached in that paper. Similar cases in various Drav. languages are Kuvi tīma an island <Pkt. diva <Skt. dvīpa, toda tīv life <*cīv <jīva, Go. kōḁā horse <Pkt. ghoḁaa, Malt. tanyare to become rich < Skt. dhana, etc., tātru a scythe or sickle < Skt. dātra, Ka. purbu brow <Skt. bhru, Ka., Te., tu. titti leather bag, Ka. also tidi bellows < Skt. dṛti. All these words must have been borrowed form IA. at a time when the languages in question were still devoid of initial sonants and consequently substituted the corresponding surd.

page 137 note 1 E.g. by Ramaswamy Aiyar, loc. cit., p. 547.

page 137 note 2 Cf. Sir Denys Bray's remarks (Brah. Lang., ii, p. 28) “⃛On the contrary Indian and Iranian philologists are now again brought up against the whole question of Dravidian words, and this time Brāhūi loanwords in particular in Balōchī, Sindhī, Jaṭkī, and even Pashto”. The words under discussion suggest that often it may not be a question of Brah. words in particular.

page 137 note 3 For the confusion of a and u here, compare Ta. kaṭai to churn with a churning rod, Ma. kaṭayuka, Ka., Tu. kaḁe to churn, to stir up: Ta. also kuṭai to churn; cf. Skt. (lw.) kuṭhara (or kuṭara) the post round which the string of a churning stick winds. Instances can be multiplied.

page 137 note 4 Professor Turner classes Nep. kuri hole, mouse-hole, with these words. It can be more easily explained, as far as meaning is concerned, out of another Dravidian word: Ta., Ka. kuḷi hole, pit, etc. Tu. kuḁe a rat's hole (treated below).

page 138 note 1 Another Skt. word belonging to the same Dravidian group is kuṭa pot. The word is not at all common, or attested very early. BR. quote only from lexica, a few references to literature being added in Schmidt's Nachirāge. Likewise it is absent from the earlier Pali literature, though it turns up in the later texts (Jātaka prose, Comm., etc.). It appears to be a late loanword, borrowed in the south, in all probability direct from Tamil, and never to have had a very wide currency. The common word for “pot” in Skt. and the derived languages is ghaṭa (Pkt. ghaḁa, Hi. ghaṛā, etc.), which, though it is not obvious at first sight, is another adaptation of the same Drav. word. It represents a Drav. *goḁa < koḁa (Ka.) with secondary voicing. As evident from the list given above such secondary voicing is common in Ea. and Te. in words of this group, and though the form in question is not actually attested, it must have been current at some time in the Ka.-Te. area, parallel with the other related words in g. It would not be surprising if it should turn up in an inscription or a dialect. Given the form *goḁa the derivation of Skt. ghaṭa presents no difficulties. Skt. gh represents Drav. g (<k with secondary voicing) also in ghūka owl: Ta. kūkai, Ka. gūge, Te. gūbe. Drav. short o is foreign to the phonetic system of Sanskrit, and is normally replaced by short a: e.g. Skt. pala, palala flesh: Ta. pulavu, pulāl, Te. pola, polasu Skt. mallikā jasmine: Ta. mullai, Ka. molle, Te. molla, etc. The intervocalic -t is naturally a mere Sanskritization of Pkt. -d. There is also in Skt. a rare word kulija which is quoted from the Kauśika-sūtra in the sense of “pot, vessel”, and from Pāṇini in the sense of a certain measure. This is another lw. from Drav. and can be compared with Ta. kuḷici pot. It has been further suggested that Skt. kuḁava (also written kuṭapa, kuḁapa) a measure of grain is connected here. The word, however, cannot very well be separated from Ta. kuḷakam dry or liquid measure, Ka. koḷaga, koḷagu, kuḷa, koḷa a measure of capacity: 4 bḷḷas, Tu. koḷaga a measure of grain. These words are naturally to be connected with the root koḷ (what a thing takes or holds): of. Te. kolapātramu a measure, a vessel of measurement, kolata measurement, kolādi measure), and in that case are naturally to be separated from the words under discussion.

page 239 note 1 Further Skt. words derived from this Drav. base are kulāya nest, Pa. kulāvaka; with the -m suffix of Ta. kuḷumu, kulmi herd (TS. Gavāṃ kulmi; it appears to be a hapax legomenon), and with secondary voicing of the initial k-, gulma group, cluster, etc.; Ka. gummu, gummi mass, crowd (<*kuḷm); probably also grāma multitude, troop; village; Ta. kuḷām.