Published online by Cambridge University Press: 22 April 2026
[Sanskrit has five nasal phonemes: velar, retroflex, dental, and labial nasals, and anusvāra. The palatal nasal is an allophone of the dental.]
1 See e.g. L. Renou, Grammaire Sanscrite 38.
2 Quoted in Mahāsvāmin's
(Kielhorn, in Indische Studien 10.421):
[sic] dīrghā na santi. Whitney on Atharvavedaprātiśākhya 1.39 (JAOS 7.364) gives details of the occurrence or lack of occurrence of
in the schemes of the Prātiśākhyas. For discussion of Āpiśali's denial of the existence of
, see Kielhorn loc.cit. and Wackernagel, Altindische Grammatik l.§30.
reference establishing Āpiśali as a predecessor of his is 6.1.92. Whitney on
in Sanskrit Grammar2 §23a: ‘only for the sake of an artificial symmetry’. Note that according to Faddegon, Studies on
Grammar 69 f. (Verh. d. k. Ak. v. Wet. te Amsterdam, Afd. Lett., N.R. 38.1), even ‘vocalic
has to be omitted [from the operation of 1.1.3].‘
3 I follow essentially the method and definition of Trager and Bloch set forth summarily in Lang. 17.223.
4 The list is taken from Benfey, Vollständige Grammatik der Sanskritsprache 21-6, with some modification. It is to be noted that a few of the clusters given are slightly dubious (e.g. occurring only in forms from
roots), but they could not be rejected without searching examination, and the conclusions are not affected by such occurrences.
5 E.g. vān̄maya-.
6 Whitney, Sanskrit Grammar2 §231.
7
treats nasa phoneme, e.g. in his handling of the nasal-infix verbs in sütras 6.4.111, 8.3.24,4.58, and in his handling of the primary suffix na in sūtra 3.3.90, to which the phonetic rule 8.4.40 applies. He is followed by the Western grammarians.
8 Böhtlingk,
Grammatik 69*.
9 Similar treatment of a word in Vedic Sanskrit by Edgerton in Lang. 10.263, and in the Shawnee language by Voegelin in Lang. 11.27. Hockett has some general remarks on the interpretation of interjections in Lang. 18.4 (§2.22), which might be extended to this case.
The status of the root n̄u is as follows. It occurs in the list
1.1000-3 which has roots ku, ghu, u, n̄u, khu, all with the meaning ‘make a noise’. Of these only the first in the list is attested (Panini 7.4.63). On u, which was at one time thought to occur in two
, Textkritische und exegetische Noten (AGGW 11.5), on RV 5.31.1. The whole list of roots was probably formed on ku, in the same way that on sev ‘honor’ was formed the list sev, gev, glev, pev, mev, mlev, śev, khev, plev, kev
43 [Indische Forschungen 8/9]). Oh the whole question of the methods of formation of these unreal lists, see A. H. Edgren in JAOS 11.1-55.
10 cñ is not discussed in Pischel, Grammatik der Prakrit-Sprachen, and apparently no reflexes of this Sanskrit combination occur.
11 E. Hultzsch, Inscriptions of Asoka, chapters VI, VIII, XI (Corpus Inscriptionum Indicarum, vol. 1).
12 Geiger, Pāli Literatur und Sprache, §§48, 53, 55.
13 Because their history was unclear to the copyists. Pischel, Grammatik der Prakrit-Sprachen, §§274, 276 end, 282, 488 end.
14 Op.cit. §§237, 276 end, 282.
15 For the occurrence of anusvāra in general, 8.3.23,24,4.59.
16 Critical Studies in the Phonetic Observations of Indian Grammarians 148-55 (London, Royal Asiatic Society, 1929).
17 So
, by interpretation.
18 Cf. Roth, Zur Litteratur und Geschichte des Weda 80; and Whitney, Sanskrit Grammar2 79.
19 Anunāsika has not been mentioned, since it is equivalent to anusvāra.
20 For such early Middle Indic influence as is here suggested, see Wackernagal, Altindische Grammatik l.xviii f. and the literature there cited; more recently such works as Maurice Bloomfield and Franklin Edgerton, Vedic Variants, vol. 2, and P. Tedesco in Lang. 20.215, 21.136-41 (this passage most important for the whole question). Though the reference in all these is to linguistic forms, yet similar statements could apply to the chronology of the development of the writing, which seems to be datable between upper and lower limits 800 b.c. and an unknown period anterior to the time of
.
21 It is unscientific to follow the over-refined Hindu method of writing nasalized l by ṅd or any other method meant to insist on the lateral character of the anusvāra;
or whatever other symbol may be agreed upon for anusvāra would be sufficient in this case as in all others; so Whitney's judgment in Sanskrit Grammar2 §206a and on Taittirīyaprātiśākhya 2.30 (JAOS 9.70).