I. PHONOLOGY. CEREBRALS. DENTALS. 35 for k-s IE. ks). The reverse transference of k to the 3. sing. has taken place in prá nak beside á-naț (naś- ‘reach') and in á-srāk (from srj- 'discharge'). b. before consonant suffixes: = 1. the phonetic cerebral appears before case-endings beginning with bh in pad-bhís, from pás- 'look' and 'cord'; vid-bhís from vis- 'settlement'; sarád- bhyas ‘for the bees' (probably from *saráh-); sad-bhis. In anaḍúd-bhyas (AV.), from anad-váh- 'bull', d appears for d by dissimilation; while the guttural of the nom. instead of the phonetic d appears in susamdrg-bhis (from drš- 'see') 'fair to see' and in dig-bhyás (AV.) from dís- ‘region'. 2. before the su of the loc. pl. k is phonetic, and appears in vik-sú, from vis, in spite of the unphonetic t of the nom. vit. But owing to the influence of the other cases the unphonetic cerebral (in the form of t dissimilated for t) appears in anaḍút-su. 3. before the dhi of the 2. sing. impv. the cerebral is phonetic in dididdhi, from dis; also in (s), which after cerebralizing the dh is dropped, z leaving a compensating length, in tāḍhi from takṣ- ‘hew' (— IE. tegzdhi); also in şo-dhá 'sixfold' (for sas-dha: as-, like as-, becoming o before a voiced mute) 3. c. The cerebrals in the following words have not been satisfactorily explained: aghāṭí- and a-ghāṭá- (AV.) 'striker', beside -a-ghāta- (VS.); andá- ‘egg'; itánt- (x. 171¹) 'wandering' (?), kúta- 'house' (?); kúta- 'frontal bone'; kipita- 'fuel' (?); mandúka- 'frog'; ita- (AV.) 'reed'; raráța- (VS.), laláta- (AV.) 'forehead'. Some others, mostly containing b, may be suspected of non-Aryan origin: bát, bada, interjections; batūrin- 'broad'(?), bírița- 'troop'(?); bekanáta- 'usurer'; ādámbara- (VS). ‘drum'; khadgá- (VS. MS.) 'rhinoceros'; cāṇḍalá- (VS.) 'outcast'; markáta- (VS.) ‘ape'. 44. The dentals.-The dentals are at the present day pronounced as interdentals in India, but according to the Prātiśākhyas+ they were post-dental, being produced at the root of the teeth (dantamula). They represent IE. dentals, corresponding to similar sounds in the cognate languages. When two IE. dentals met, there seems to have been a tendency to change the first to a sibilant 5. A survival of this appears in some Vedic combinations of d or dh with dh, which point to an earlier zdh, viz. in de-hí, beside dad- dhí ‘give’; dhe-hí (for *dhadh-dhi) ‘put'; kiye-dhá ‘containing much', in all of which examples e is based on IIr. az6. a. Change of s to t. The dental sibilant as the final of roots or nominal stems becomes #7: 1. before the s of verbal suffixes (future, aorist, desiderative) in the three verbs vas- 'dwell', vas- ‘shine', and ghas- 'eat'8: thus avātsīs (AV.) ‘thou hast dwelt'; vát-syati (MS.) 'will shine'; jighat-sati (AV.) ‘desires to eat, and jighat-sú- (AV.) 'hungry'. corresponding forms). This sibilant, Indians, first became the cerebral muted before the bh-suffixes (as dental s became dental d) when it first spread to the nom., and lastly to the loc. pl. 4 See RPr. I. 19; TPr. II. 38. 1 From this phonetic change of s to d before bh is to be explained the stem id- "refreshment', beside iş- (which occurs before vowel endings only): id-bhis etc. would have led to the formation of id-a, etc. (inst. sing.), which then gave rise to iḍ-ā- as an extension of id-; cp. also idáyata (RV. 1. 1916 MM., ilayata, AUFRECHT) : iláyatì (AV.) be quiet'. 5 For example, Gk. Foiola, Av. voistā, beside vét-tha 'thou knowest'. Cp. WACKER- NAGEL I, 152 b. 2 It is not phonetic in aviḍdhi and viviḍdhi (see above, 42 d). Cp. WACKERNAGEL I, 149 c (end). 6 Loc. cit., also note. 7 See discussion of attempted explanations in WACKERNAGEL I, 153, note. 3 The cerebral which in this paragraph represents (except before s) the old pala- 8 All the other roots in s add the suffix tals, is based on an IIr. sh- sound š ž (as with connecting vowel i. shown by the Avesta having in the 3*
