Hostname: page-component-848d4c4894-sjtt6 Total loading time: 0 Render date: 2024-07-03T03:29:34.216Z Has data issue: false hasContentIssue false
  • English
  • Français

On the plural and dual in Sogdian

Published online by Cambridge University Press:  24 December 2009

Nicholas Sims-Williams
Get access Rights & Permissions [Opens in a new window]

Extract

As a result of the gradual attrition of final syllables, which caused the Old Iranian nominal plural endings to become insufficiently distinct from those of the singular, the plural was extensively remodelled in almost all Middle and New Iranian languages. Such renovation could come about in two ways. Either the more distinctive pl. inflexions, such as the gen. *-ānām or the instr. and dat. forms containing *b, could be extended or adapted to supply other cases, as happened in Parthian, Šuγnī, etc., or the old pl. could be replaced by a surrogate pl. derived from a sg. collective noun. The latter process is well illustrated by Sogdian, where the usual pi. suffix is -t (oblique -ty) after heavy stems, -t– (obl. -ty') after light stems. As is indicated by the inflexion, which is that of an old ā-stem, and by the fact that f. sg. forms of the articles and demonstratives are used with pl. nouns, this suffix derives from that of the OIr. f. abstract-collective nouns in *-tā-.

Type
Research Article
Information
Bulletin of the School of Oriental and African Studies , Volume 42 , Issue 2 , June 1979 , pp. 337 - 346
Copyright
Copyright © School of Oriental and African Studies 1979

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

Footnotes

1

Bibliography.

Barth.: C. Bartholomae, ‘Mitteliranische Studien, VI’, WZKM, xxx, 1917–18, 1–36.

Benv.: E. Benveniste, Essai de grammmre sogdienne, II, Paris, 1929.

Gauth.: R. Gauthiot, ‘Du pluriel persan en -hā’, MSL, xx, 1918, 71–6.

GMS: I. Gershevitch, A grammar of Manichean Sogdian, Oxford, 1954.

Morg.: G. Morgenstierne, ‘Archaisms and innovations in Pashto morphology’, NTS, XII, 1942, 88–114.

SGS: R. E. Emmerick, Saka grammatical studies, London, 1968.

Ted.: P. Tedesco, ‘Ostiranische Nominalflexion’, ZII, iv, 1926, 94–166.

References

2 The exception is of course Khotanese, which is unique in preserving a system with five distinct cases in the pl., only one less than in the sg. (cf. Emmerick, , BSOAS, XXVIII, 1, 1965, 24, and SGS, 249–50CrossRefGoogle Scholar).

3 Ted., 151; Benv., 79; GMS, § 1069; differently Gauth., 75 (*-θwa- n.), and Barth., 18 (*-tāt-).—Also the Ossetic pl. suffix -tä goes back to *-tā-, cf. Bailey, , TPS, 1945, 24–6CrossRefGoogle Scholar (but note that the Choresmian cognate quoted there has proved to be illusory, cf. Henning, , Z. V. Togan'a Armağan, Istanbul, 1956, 430–1Google Scholar).

4 AL III.23, cf. Henning, , ‘Mitteliranisch’, Handbuch der Orientalistik, Abt. I, iv Google Scholar. Bd., Iranistik, 1, 33, n. 1. The further examples cited in GMS, § 1185, from SCE are of doubtful value in view of the prolific and inconsistent employment of the ending -' in this text (cf. GMS, § 1182). On M δβr' see below, p. 340.

5 GMS, § 1205. Cf. also § 1217: ‘Heavy-stem adjectives showing no ending in the Plural, can be considered as virtually having the old Plural ending -ā’.

6 I take the opportunity to draw attention to one more occurrence of this compound in a fragment recently published by Gropp, G., Archäologische Funde aus Khotan…, Bremen, 1974, 366–7Google Scholar. The text (in a cursive but legible Sogd. script) may be read thus: (1) ]. wyspγw[nc (2) ]w 'sp[t'] (k). [ (3) *βγ'n β]γtm pw(t)[y. For wyspγwnc see Dhy. 153, 295; for βγ'n βγtm pwty (passim in B texts) cf. also M370, 11–12 (apud GMS, §§ 254 and 1179).

7 There is some variation in usage. Thus, the pl. of the loan-word ykš- ‘yakṣa’ is formed with -yšt(y) in Christian and Manichean texts (C2/12R.2 = BST ii, 849.2; Henning, , BBB, pp. 102 on f8 and 139 s.v. Google Scholar; BSOAS, XI, 1, 1943, H16Google Scholar; XII, 2, 1948, 312, 1. 103) but with -t', -ty– in B Sogd. (P2.338; P5.6; P6.160; PII.16, 19; P21 ii.ll, iii.l; O2, 29).

8 Gauth., 76, n. 1; Benv., 79.

9 On the origin of -išn etc. see Benveniste, , Les infinitifs avestiques, Paris, 1935, 105–7Google Scholar (cf. also Henning, , Sogdica, London, 1940, 17–18 Google Scholar, on bāmbušt < bānbišn).

10 Gauth., 76, n. 1.—For the sake of completeness a reference should here be included to the derivation of Sogd. -yšt from the superlative suffix *-išta- (proposed by Mirza, H. K., Proceedings of the XXVIth International Congress of Orientalists, ii, New Delhi, 1968, 235 Google Scholar).

11 Barth., 2 ff., where however several incorrect forms are adduced.

12 GMS, § 350; rather differently Henning, , ZDMG, xc, 1936, 197 Google Scholar, and GMS, § 13991.

13 Examples in GMS, §§ 1259–60 (*aka-stems), 1229 and 1251 (other heavy stems).

14 ST ii/4.41–2.

15 ST ii/6.18; C2/12R.28; C2/48R.27 (= BST ii, 883.25).

16 Cf. the situation in Pato, where the voc. pl. is identical with the obi. pl. (Morg., 101), and in Khot., where the instr.-abl. pl. in -yau is used for the voc. pl. also (SGS, 250, 265).

17 ST i, 23.4([']fryṭyṭy), 25.6 (nfryṭyṭy).

18 The obl. pl. is pointed thus some 50 times in C5, but only twice -tỵ (mrtxmỵtỵ, Sundermann, , Altorienlalische Forschungen, i, 1974, 254, I. 12, and 255, 1. 15Google Scholar; Müller in fact read o ytỵ in both instances, but Sundermann, who kindly re-examined the passages at my request, confirms that the MS has o ỵtỵ). The evidence of the MS C2 is less straightforward. In this MS the pointing ỵ, is altogether extremely rare; where C5 has ỵ, C2 generally uses unpointed y, occasionally ỵ. A fairly typical example of this mode of vocalization (which probably results from a phonetic shift of [i] to [ẹ], contrasting still with /e/ = [ę]) is provided by the nom. sg. bγy, which is spelt thus, without points, about 20 times in C2, but also twice with y1(23R.28 = BST ii, 859.28; 77V.15) as against C5's bγỵ. Similarly the obl. pl. in -ty is usually left uhvocalized in C2 (about 170 examples); -ty2, occurs six times (31R.13, 23; 31V.20 bis; 39V.3; 56V.9), but a feeling that this pointing is incorrect is indicated by the fact that on three further occasions the points have been deleted (51R.8; 57R.14 bis). The overwhelming preference for unpointed -ty seems to rule out an interpretation with /e/ = [ę], for which the pointing y1 is very liberally employed in this MS.

19 VJ 301, 309, 323, etc.; ST ii/1.40; K16 (unpublished), 1.3 ('γšnk' is either uninflected or more likely nom. sg. f.; cf. the instances of group-inflexion in the voc. pl. referred to above and the even more closely comparable m. sg. equivalent pry z't' ‘dear son’, VJ 24a).

20 SGS, 285.

21 Cf., for instance, Trumpp, E., Grammar of the Pḁstō…, London, 1873, 90 Google Scholar.

22 Justification for these and similar general statements will be found in my forthcoming edition of C2.—In the present article references to my text of C2 are wherever appropriate accompanied by references to the edition of Hansen, O., ‘Berliner sogdische Texte, II’, Abh. der Geistes- und Sozialwisenschaftlichen Kl. der Akademie der Wissenschaften und Literatur in Mainz, 1954, 15 Google Scholar [= BST ii].

23 Recognized by Schwartz, M., Studies in the texts of the Sogdian Christians (unpublished Berkeley dissertation), 1967, iiiGoogle Scholar.

24 Only C2/68R.28 (= BST ii, 843.28): 'y1 γnt'q-qr(y)t y1 ‘O evil-doers’.

25 C2/60R.2 (= BST ii, 888.2); C2/87V.3 (= BST ii, 878.3); C2/94R.23; C2/40V.19 (= BST ii, 907.19); C2/68V.12 and 69R.30 (= BST ii, 843.43 and 846.32). These forms were discussed in AM, NS, XVIII, 1, 1973, 100, but a definitive treatment was not possible at that time.

26 Schwartz, , op. cit., 15 and 151Google Scholar.

27 C2/57R.25 (= BST ii, 904.25); C2/57B.29 (= BST ii, 904.29); C2/69R.14 (= BST ii, 846.16).

28 C2/94R.28 (= BST ii, 826.28).

29 C2/56V.31 and 57R.29 (= BST ii, 904.29). The etymology of prtw is due to Schwartz, , op. cit., revisions (1969), 9 Google Scholar.

30 Sic, cf. IIJ, xx, 3–4, 1978, 260 on Dhu. 269Google Scholar.

31 Sogd. δ(y)βnw surely corresponds to the Khot. gen. dvīnu (on which see SGS, 249). Differently GMS, § 1320.

32 M178 apud Henning, , BSOAS, xii, 2, 1948, 312 Google Scholar (cf. also Sogdica, 53 on t8 ](iv) δβr').— Unfortunately the sg. is not attested of pnc fsp' ‘five rugs(?)’, M178.74–5, which may also belong here.

33 Cf. Beaulieux, L., Grammaire de la langue bulgare, Paris, 1953, 53–4Google Scholar; Lunt, H. G., A grammar of the Macedonian literary language, Skopje, 1952, 32 Google Scholar. I am most grateful to Bernard Comrie for drawing my attention to the Slavonic phenomena.

34 Cf. Morg., 94; Lorimer, D. L. E., Pashtu, I, Syntax of colloquial Pashtu…, Oxford, 1915, 48 Google Scholar; Smirnova, I. A., Formy čisla imeni v iranskix jazykax, Leningrad, 1974, 197, n. 28Google Scholar.

35 Apparently sg. forms are, however, sometimes found in B texts, e.g. δry cšmy, VJ 915; cnn δβnw cšm', VJ 277; 'βt' 'yšph, P6.128. In this respect, as in so many others, the B usage is probably merely ‘sub-standard’.

36 Examples passim, cf. for instance GMS, § 1662 ff.

37 C prtw (cf. above); M rtw ‘ten seconds’ (M14, Rl, E4, and passim apud Henning, , JBAS, 1945, 149–53Google Scholar) and jmnw ‘hour’ (M14, RIO; further examples apud GMS, § 1662). In my view (differently Henning apud GMS, § 5121) jmnw ‘time, hour’ is in origin the nom.-acc. sg. of a neuter žamn- < *ǰamana-: to this stem belong also the loc. sg. jmny- and the pl. jmnd-. It has, however, been re-interpreted according to the analogy of light u-stems such as rtw, prtw, mγw/mwγ-, etc., as is shown not only by the nom.-acc. numv. jmnw but also by the variant loc. sg. jmnwy- and by the gen. sg. (with f. ending for an old n.) γiwmny' in ST i, 84.14 (not loc. as. assumed by Henning). Similarly M whrw ‘kind’ (nom. sg., e.g. BBB 533), treated as a u-stem in the nom.-acc. numv. ii whrw (BBB a5; M697 and M778 apud GMS, § 1662), may originally have been a n. α-stem, a possibility which seems to be favoured by the B form whry (passim, e.g. 'yw whry, δw' wkry). In any case whrw ‘kind’ should be derived from *wi-kurV-, cf. B wk'wr ‘kin’ < *wi-kaura- ( Gershevitch, , BSOAS, xiv, 3, 1952, 493–4Google Scholar).

38 Cf. Henning, , BSOAS, xi, 4, 1946, 728 Google Scholar; GMS, § 807.

39 The passage is poorly preserved, but it is hard to imagine a plausible alternative to the restoration here proposed. Cf. Av. yuxta- ‘pair, team’ (rightly assumed by Bartholomae to be a n. rather than a m. noun, though direct evidence of gender is lacking), N P ǰuft.

40 Tale E apud Henning, , BS0AS, xi, 3, 1945, 473 Google Scholar. That βγn- is neuter is most clearly indicated by the nom. sg. βγnw, which occurs in a verse from the unpublished Sogd. version of the MPers. hymn-cycle ‘Gōwišn ῑg grῑv zῑndag’: zwkyh β't ZKn tw’ βγnptw oo ky ZY βy ZK βγnw pršt't prw- 'z-prtyh ‘Hail be to Thee, lord of the temple, whose temple has been prepared in holiness’ (13399a, R12–14; Sogd. script, unpublished).

41 See Miller, W., Grundriss der Iranischen Philologie, i, Anhang, 47–9Google Scholar; Abaev, V. I., A grammatical sketch of Ossetic, 22 Google Scholar.

42 In Dig., numerals themselves, when used substantivally, also take -e- + pronominal sg. suffixes in the obl. cases. For a further idiosyncratic use of this -e- see Bailey, , TPS, 1956, 125–6CrossRefGoogle Scholar.

43 Abaev, , op. cit., 82, n. 4Google Scholar.

44 C2/51R.27–9 (= BST ii, 913.56–8).

45 Ephraem, , De Nativitate, 1.67 Google Scholar, apud Beck, E., Des heiligen Ephraem des Syrers Hymnen de Nativitate (Epiphania) (Corpus Scriptorum Christianorum Orientalium, Scriptores Syri, LXXXII), Louvain, 1959, 9 Google Scholar.

46 C2/60V.28, 30 (= BST ii, 890.60, 891.62); Б15.10.

47 Cf. Ted., 153; GMS, § 1187. It is possible that a few of the examples quoted may be sg. (with merely graphic -h) but this can hardly be true of all, in particular not of ‘spy’.

48 VJ 47, 86, 164 (-y'), 416, 13c, 618, 10e, 11e, 843, 1166; Vim. 64; P2.980.

49 SCE 226.

50 VJ 618 (MS 'γwt'ryh!). Note that the spelling 'γwštryh cited by Tedesco and Gershevitch does not in fact occur.

51 P2.784; VJ 86.

52 e.g. 'spy and 'γwštry in VJ 497, 560, 977, etc.

53 VJ 416, 13c, 1165.

54 The loc. sg. (in -y y2) of 'γwštr- does not occur.

55 Cf. for instance < y in the inflexion of Sogd. y y2 -stems: AM, NS, XVII, 1, 1973, 98 (on 1. 80), 99–100 (on I. 107)Google Scholar.

56 The pointing indicates [-(y)].

57 See Emmerick, , SOS, 289, 293Google Scholar.

58 A m. i-stem may also be assumed to lie behind Sogd. kyrm- ‘snake’ (cf. OInd. kĭmi- m. ‘worm’), but this word would hardly have been used sufficiently often to provide the sole starting-point for the spread of these pl. endings.

59 Abaev, , Istoriho-ètimologičeskij slovar' oselinslcogo jazyka, I, Moscow-Leningrad, 1958, 605 Google Scholar; Gershevitch, , The Avestan hymn to Mithra, Cambridge, 1959, 173, fnGoogle Scholar.

60 GMS, § 947; SGS, 250, 289; Morg., 93.

61 Cf. also Sogd. qwš- m./f. ‘side’ ˜ OInd. kukṣi- m. (see BSOAS, XLII, 1, 1979, 134 Google Scholar).

62 Morgenstierne, , Etymological vocabulary of the Shughni group, Wiesbaden, 1974, 40aGoogle Scholar.

63 Turner, , Comparative dictionary of the Indo-Aryan languages, London, 1966, No.3275Google Scholar.

64 Morgenstierne, , Irano-Dardica, Wiesbaden, 1973, 105 Google Scholar.

65 Full details will be given elsewhere.

66 P13.17; cf. IIJ, XVIII, 1–2, 1976, 59 Google Scholar.

67 Morgenstierne, , Irano-Dardica, 104 Google Scholar.—It is tempting to explain OPers. uša-bᾱr - ‘camel-riding’, whose suffix contrasts oddly with that of asa-bᾱra- ‘horse-riding’, as a bahuvrῑhi ‘having a camel as steed’, the second element of the compound being a f. noun *bᾱrῑ- ‘riding-animal’ (cf. Chor. β'rcyk f. ‘id.’, Henning, , Togan'a Armağan, 431 Google Scholar).

68 Cf. also Šuynῑ wirjin ‘she-wolf’ < *wṛčῑ-nῑ- ( Morgenstierne, , Irano-Dardica, 107 Google Scholar) and, for the non-palatalization of k, Pato spǝῑ ‘bitch’ < 8spakῑ- (ibid., 103, cf. 106).

69 In GMS, §§ 140–1, an occasional development of unpalatalized *ṛ to yr is assumed, but all three examples quoted may in fact show palatalization: wyrk- < *wṛkῑ-; kyrm- ‘snake’ < *kṛmi- (or *kṛmῑ-); M 'ngyr(p)[ ‘with united bodies’, Kawᾱn, G14, perhaps to be read 'ngyr (f)[ (a possibilit}’ specifically mentioned by Henning, , BSOAS, XI, 1, 1943, 68, n. 6Google Scholar) and derived from some such form as *han-kṛf-ya-.

70 This transcription is based on the pointing in ST i, 6.8–9: wỵrqỵštỵ, qrmšỵt. No reliance can be placed in the pointing -y1ỵt four times in C2 (6E.15, 12R.3, 12V.6, 24R.8), cf. n. 18 above.

71 To the same fact one may attribute the still wider use of the nom. pl. -išt in Waxi, whether one regards this as a borrowing from Sogd. or as a common inheritance.

72 Bailey, , TPS, 1945, 25 CrossRefGoogle Scholar.

73 GIP, I, 1, 227, 229–30.

74 SGS, 288–9; Morg., 93.

75 Cf. GMS, §§ 114 ff., 163 ff.

76 An explanation sometimes proposed for OPers. B-a-x-t-r-i-š, etc., cf. for instance Kent, R. G., Old Persian…, 2nd ed., New Haven, 1953, 60 Google Scholar.

77 It is noteworthy that a considerable proportion of the Ved. nouns declined like υṛk - consists of f. animal names, mostly corresponding to masculines in -a-.

78 Bǝn ‘co-wife’, etc., cf. Morg., 93.

79 Differently Morg., 99–100, where an extension of the obl. forms (mor < *mᾱθrah, etc.) is assumed.

80 Olr. ῑ-stem nom. sg. forms in -š such as OPers. B-a-x-t-r-i-š or Av. dᾱθriš are inevitably ambiguous, containing either -iš (from the i-declension) or -īš, while no later Ir. language could be expected to show distinct reflexes of final *-iš, *-īš, and -ῑ.

81 For several valuable references used in this article I have to thank B. Comrie, B. R.E. Emmerick, I. Gershevitch, D. N. MacKenzie, M. Schwartz, and W. Sundermann. To Gershevitoh I am further indebted for his fruitful suggestion quoted on p. 343 above.