¹ Nyberg, H. S., A manual of Pahlavi. I. Texts, Wiesbaden, 1964.Google Scholar

² Handbuch der Orientalistik, Abt. I, IV. Bd., Iranistik, 1, 122.

³ See pp. 201, nn. 12, 13.

⁴ op. cit., 123.

⁵ Tedesco, P., Dialektologie der westiranischen Turfantexte, MO, xv, 1921, 184258.Google Scholar

⁶ ibid., 249 f.

⁷ For Avestan, v. Morgenstierne, G., NTS, xii, 1942, 72 ff.Google Scholar; for OP, R. G. Kent, Old Persian, 59.

⁸ Initial - need not have had the same development as medial --, so it is not necessary to assume the survival of an OP *sata- beside ata-, etc., either by dialect differentiation or chance development, with I. Gershevitch, TPS, 1965, 13, to account for M and NP sad. Nor is his postulate (ibid., n. 2) of an occasional Old Persian realization of dental as s, to explain the s- of M and NP sanfidan, saxt, etc., at all to be expected. (The s of the Pahl. spellings l's, g's, nswb'l-, etc., is no more likely to preserve an OP sound than the s for h in such words as I 'lswmwk^Y, B ywsd'sl.) Sound laws, or trends, may develop slowly, but at least they do so in one direction, in a limited channel, and to a definite end. Even if there were the occasional backwash at a confluence of channels it could not be expected to survive against the main current, still less zigzag into another channel. By the same token, there is no other example of initial w- > h-to support G.'singenious etymology (ibid., 15 f.) of MMP hagj-n- < Av. *wazga-.

⁹ See Hbschmann, Persische Studien, 95100.

¹⁰ See, e.g., G. Lazard, Grammaire du persan contemporain, 5.

¹¹ Though the very existence of a distinct Parthian orthography, plainly clothing a different dialect, should in itself be sufficient to call in question any lingering tendency to assume Parthianisms on a large scale in the non-Parthian Pahlavi.

¹² In either case (z- for z- or -), it is clear that no *jm'n was ever written in Pahl.; indeed no letter j ever really existed, only a historic or pseudo-historic y (e.g. y'n = jn). The spelling with a seeming y- must be ẕm'n, with a shrunken z. In view of this it is more than likely that non-Persian z- was also intended, rather than Pers. d-, both in those words with alternative spellings (zmyk' = ẕmyk') and those without (ẕmst'n'), cf. M zmyg but dmyst'n, NP zamestn.

¹³ Also jam ; see previous note.

¹⁴ The pseudo-historic c in Pahl. also indicates that the post-vocalic c of contemporary MMP spelling was at least archaistic, if not already historic. A similar case of a sound change developing at about this time is presented by postvocalic f > h. While M has -f in lwp, kwp, and similarly I kwl'p^Y, kwpd'l^Y, B has both kwp, kwl'pk (the cap, i.e. calyx of a date, n, 2.122) and 'lwh, and ENP luh, kh, and kulh.

¹⁵ See Nyberg, Dr. J. M. Unvala memorial volume, p. 100, n. 2, where b, d, g > , , applies to Olr. b, d, g, while p, t, k remain unchanged!

¹⁶ e.g. B 'py, npt' = M 'by, nbyt > ENP b, nibit (Lazard, La langue des plus anciens monuments dela prose persane, p. 137, n. 3), B wtl- = M wdr- > ENP guar-, B gwk'dy = M gwg'y > *guy > ENP guv(h).

¹⁷ e.g. apa, B 'pkn- = M 'bgn-; upa, B 'plywm = M 'brywm, B 'p'l-; abi, B 'pzwn = M 'bzwn, BM 'pswn.

¹⁸ e.g. B 'wst't, 'wt'p = M 'wyst'd, 'wyt'b.

¹⁹ Lazard, Langue persane, p. 140, 12 ff.

²⁰ Pahl. ptgl- = M pdyr- padr- is a different case, with early -r- > *-iir- > *patigrbya-.

²¹ Zum altir. Wrterbuch, p. 36, Excurs 14, Aber nicht nur i und u, sondern auch kurzes e und o wird durch Jod und Waw gar nicht selten dargestellt.

²² Despite OP *dam-, baudi- (Av. zam-, baoi-), cf. Arm. burastan.

²³ With the loss of d before m, via *-ehm-, there was compensatory lengthening, e.g. M nym = ENP nim (Pth. ndm), xyym = xm (Pth. xdm).

²⁴ ENP also dax (Lazard, Langue persane, 148), but Arm. dox > duox = M dw(w)x.

²⁵ This contrast of * > e > a : ȧ > o > u may throw light on the distinction between M -yr- and B -rl- (NP -ar-), but M = B -wr- (ENP -ur-), < -r- in different contexts. Taken in conjunction with such Arm, loans as kerp (M kyrb: B kip < krpa-) and kert (M kyrd: B krt < krta-) as a suffix and in aakert (M h'gyrd: B h'kll), etc., it suggests that the simplest development of -ṛ- was to *-er-, thus MP kerb form, kerd made, hagerd disciple, ters- (M tyrs-: B tls-) fear, etc., while other factors influenced the shift to -i-, e.g. in kirm worm < *kṛmi-, kit sowed < *kṛta-, dil heart < *dṛd-, etc. There is no evidence to show whether the labialization -ṛ- > -ur- proceeded via *-or-.

²⁶ The stems karya-, tarya- were evidently exempt from the normal metathesis -arya- > *-ayra- observable in MP rn, er, mrag, nrg, etc.

²⁷ With this word and the two (equally non-Persian) forms B m = Av. maa- ande DKR' = *mu (Kumzr mu, NP mox) date palm o appears to be established by the minimal pairs of ma: mo: mu, but there is no certainty of a distinction between the last two.

²⁸ Hubschmann, Armenische Grammatik, 235, s.v., already im Phi. immer sardr geschrieben, aber zur Sassanidenzeit slr gesprochen.

²⁹ The mention of this word permits a comment on Pth. wx-. The process of Olr. hw-, x ^v- > the simple w- of certain northern dialects (Zaza, Avromn, Grn, and proto-Bal), noted by Christensen and quoted by Tedesco, op. cit., 207, is remarkably similar to that of Germanic xw- (IE q ^u-) > OEng. hw- > Mod. Eng. wh- w or w. The inversion in the Man. Pth. spelling was probably therefore a device, in effect like Eng. wh-, to represent a new sound, viz. devoiced w. It lends itself to the transcription w^x-, thus w^xar, w^xbeh, w^xa, etc. This assumption would the more easily explain the loss of -x-, i.e. of voicelessness, after a voiced consonant in dyjw'r (?), pwnw'r (v. Henning, Annali 1st. Orient. Napoli, Sez. Ling., vi, 1965, p. 33, n. 1).

³⁰ Zoroastrian problems, p. 152, n. 1. AV, i, 7 KLYT' npst = DkM, 405.20 dc Y npt.

³¹ So Bailey, TPS, 1959, 71 f.

³² Bang and von Gabain, Trkische Turfantexte, II, 15.

³³ Or npyd- nibey- < nipa'ya-, v. Geiger, WZKM, XL, 1933, 116.

³⁴ See Morgenstierne, Acta Or., i, 1923, 274.