Abbildungen der Seite

The release of the syllabic becomes the consonantal that is articulated before the vowel e. The group ú+rr really becomes ir.


[merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][ocr errors][ocr errors][ocr errors][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][ocr errors][merged small][merged small][merged small][merged small]




The question of the mutual relation of the various privative nominal prefixes characterised by n in the Indo-European languages, as well as of these prefixes to the n-negative particles, has repeatedly been discussed from more than one point of view. Without recapitulating the theories already advanced, which may readily be ascertained from the appended bibliography, the problem may be examined anew in the light of all linguistic evidence now available. The fullest data are found in Greek, then in Italic, Indian, and Iranian; doublets of the type *n: *ne are rare in Celtic, Teutonic, and Balto-Slavic; they are not found in Armenian, which has only an= *n;1 and their existence is questionable in Albanian.2


Here five types are observed: ἀνάκεστος: ἀνήκεστος : νήκεστος; ἀνάγνωστος : ἄγνωστος; and ἀμφασίη. The material for the first three types, in the alphabetical order of the second component, is as follows:

[merged small][merged small][merged small][merged small][merged small][merged small][merged small][ocr errors][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][ocr errors][merged small][merged small]

1 Hübschmann, Armenische Grammatik, 1. 419, Leipzig, 1895.

'Pedersen (1. 45) would cite here Albanian ésɛlɛ 'fasting': sil 'breakfast,' and égel 'fasting': gele 'food,' but this is doubted by Barić, Albanorumänische Studien, 1. 59-60, Sarajevo, 1919.

'For the distribution of the n-compounds see Hamilton 61-62, and cf. ib. 11-13.

4 For the type αθάνατος : ἀθάνατος, where the first of three short syllables is metrically lengthened in words of four or more syllables, see Hirt, Handbuch,2 51, and for the Sanskrit ádeva 'godless,' ásat 'non-being' (árūpita is too uncertain to admit of satisfactory discussion), see Wackernagel, 2. 131. The Modern Greek colloquial forms ȧvýuжopos ‘powerless’: ἐμπορῶ ‘to be able,” ἀνήξευρος ‘ignorant’: ἐξευρίσκω ‘to discover” are derived by Hatzidakis (Einleitung in die neugriechische Grammatik 431, Leipzig, 1892) from the colloquial forms ἠμπορ(έ)ω, ἠξεύρω. The colloquial ἀνηπρόκοπος ‘unprogressive’beside ¿πρóкожоs is apparently due to analogy with some compound where &vŋ- is justified.

[blocks in formation]

The rather scanty material of the type ἀνάγνωστος : ἄγνωντος is, in

Ancient Greek:

ἀνάγνωστος ‘absolutely secret': ἄγνωστος

ἀνάεδνος ‘without bridal gifts’6

ἀνάελπτος ‘unhoped for’7

ἀνάπνευστος ‘breathless': ἄπνευστος

ἀνάπταιστον 'τὸ μὴ πταῖον' (Suidas): ἄπταιστος

5 Callimachus, frag. 422: μηδὲν ἐθέλω καλὸν ἔχειν ἀνάγνωστον; see Burlingame 301. This ἀνάγνωστος is, of course, quite other than the common ἀνάγνωστος ‘readable.’

6 ἔδνον < *εhεδνον > *εεδνον (Boisacq 215).

7 ἔλπος < * ελπ-, ib. p. 246. For ἀάατος “inviolable' see ib. 1-2, and Ehrlich, Untersuchungen über die Natur der griechischen Betonung 227, Berlin, 1912, the latter deriving it from *a-rafa-to-s ‘disallowed'; ἀάσχετος ‘irresistible' is derived by Schulze, Quaestiones Epicae 495. n. 1, Gütersloh, 1892, from *n-sn-sgh-eto-s 'quasi non continendus.'

In Modern Greek this use of ȧva- has received some extension, as is shown by the following examples, the colloquial forms being marked by C:8

C ἀνάβαθος ‘senza profunditd': ἄβαθος

ἀνάβατος ‘inaccessibile': ἄβατος

ἀνάβολος ‘incomodo’: C άβολος

ἀναβράκωτος ‘sans-culotte': C ἀβράκωτος

C ἀναβρεξά, ἀναβροχή ‘siccità': C ἀβρεξία, ἀβροχία

C ἀνάγνωρα ‘unknown’: C ἀγνώριμος

C ἀνάκακος ‘senza malizia': ἄκακος

ἀνάκαρδα ‘timid’: ἄκαρδα

ἀνακέφαλος ‘headless”: ἀκέφαλος C ἀνάμελος ‘negligente': ἀμελής

C ἀναπάντεχος ‘inatteso': C ἀπάντεχος

ἀναπόλιφτος ‘not urbane, rustic’: ἀπολίτευτος

C ἀνάφαγος ‘senza appetito': C ἀναφάγανος ‘insaziabile'

Of the fifth type the sole examples in Greek, apart from the Homeric ἀμφασίη ‘speechlessness' beside ἀφασία,9 are, possibly, ἀννέφαλος ‘cloudless” beside ἀνέφελος (the latter metrically lengthened to ἀνέφαλος, ζ 45)

and aλoços 'crestless' (only K 258, where the variant reading is aλopos) beside ἄλοφος.10

(The etymology of KeσTOS 'ungoaded' (?) found thrice (Z 94, 275, 309) in the phrase βοῦς ἥνις ἠκέστας with the variant reading ἀκέστας -is quite uncertain. It is explained by Froehde1 as equivalent to a Sanskrit *āsasta 'hoped for, desired, praised.' One is half-tempted to associate the formation with that of Old High German dwiggi 'wayless, astray' (cf. infra, VI); but the problem is yet too obscure to admit of satisfactory solution.13)

8 Brighenti, Dizionario greco-moderno-italiano, Milan, 1909; G. Meyer 11-12. 'Kretschmer, KZ 31. 407 (1902); Boisacq 57-58. Collitz, in Hamilton 11, suggests the division ȧμ-aoin and connexion with the base *as (d)- 'to parch, wither'; Hatzidakis, 'Akadnμeikà 'Avayvwoμara, 2. 230, Athens, 1904, makes it a metrical formation on the analogy of ἄμβροτος ‘immortal: βροτός.

10 In the Epidaurian poem of Isyllus, line 66, auμopos is to be read instead of &μñopos (Baunack, Aus Epidauros 18, Leipzig, 1890).

[All examples of the type are probably metrical lengthenings, graphically designated

in various ways. Cf. also Walde, Streitberg Festschrift 153 f. G.M.B.]

11 Cf. ἠκέστης ὁ ἀδάμαστος (Suidas); ἠκέστας δὲ ἀντὶ τοῦ ἀδαμάστους, ἀκεντήτους, ἀκενTplotovs, ¿voxeútovs (Etymologicum Magnum, s. v. ñvis; cf. Etymologicum Gudianum, s. v. ἠκέστας).

12 BB 7. 328-9 (1883).

13 Kluge, in Grundriss der germanischen Philologie, 476, connects * with *ě, ** 'there' (cf. Brugmann, Grundriss, 2. 2. 816-9, Strassburg, 1911), but Brugmann (837 f). associates it rather with Greek ǎvev 'without.'

Brugmann1 explains the type of ȧvnpeons as due to a 'secondary initial prolongation,' and the type of vηкepdηs 'unprofitable,' as for *n-kepdns, though he had previously held15 that vηkepôýs is formed by analogy with the type νήνεμος for *νε-ανεμος. As for the ava-type, Froehde16 derived ἀνάπνευστος from *ἀν-αναπνευστος: ἀναπνέω, and ἀνάεδνος from *ἀν-αεδνος, the other three instances being analogical with these. Schmidt, on the other hand, held1 the second a of ȧva- to be a case of svarabhakti; while Burlingame1 has advanced the view that ȧva- is a double negative, *a-na, the starting point being such compounds as Sanskrit an-a-vrata 'not without austerities.' None of the words regarded by Baunack1 as showing a negative prefix *ne (ȧveμwλios 'windy, empty,' veẞpós 'fawn,' véкTap 'nectar,' vérous, a Homeric epithet of seals) is really so to be explained.20

I myself interpret the Greek series νηκερδής, ἀμφασίη, ἀνήγρετος, νήγρετος, ἀνέγερτος, ἀνάγνωστος as follows. The first five represent various grades of a base *ane- 'not.' The second prolonged grade (or possibly the second full grade of a secondary base *anē-) (*nē-) is found in vη-keρons; the full grade I a (*an-) in ȧu-yaoin; the full grade II a (*。né-)21 in ἀνήγρετος for *ἀνε-εγρετος; the full grade II b (*ne-) in νήγρετος for *ve-eypetos; and the null-grade (*-n) in ȧv-éyeptos; while the reduced grade b of a secondary base *aně- (*anǝ-) is represented by ȧvá-YvWOTOS,22 and the second full grade (*nē-) (or possibly the second prolonged grade of *ane-) by vη-κερδής.

"Grammatik, 194.

15 Grundriss,2 2. 1. 23, Strassburg, 1906; cf. Hirt, Akzent 312 f., Handbuch3 458; Boisacq, 667 f.

16 BB 20. 213 (1894); cf. Solmsen, Untersuchungen zur griechischen Laut- und Verslehre 264-6, Strassburg, 1901, Brugmann, Grundriss,2 2. 1. 22, Grammatik,♦ 611.

17 KZ 23. 274 f. (1877).

18 AJP 39. 299-300, and similarly Andersen, Pāli Glossary 2, 8, 9, Copenhagen, 1904-05; cf. also, in general, Hamilton 11-15.

19 Studien auf dem Gebiete des Griechischen und der arischen Sprachen, 1. 271-6, Leipzig, 1888.

20 Cf. Boisacq. 61. 660f., 664f. [On véкTap cf. H.Güntert, Kalypso. R. G. K.]

21 For the representation before n of * by a in Indian, Greek, and Latin, by u in Teutonic, and by i in Balto-Slavic, see Hirt, Der indogermanische Ablaut 18, Strassburg, 1900, Der indogermanische Vokalismus 86, Heidelberg, 1921.

22 For Indo-Greek accentuation of privative syllables see Wheeler, Der griechische Nominalaccent 45-9, Strassburg, 1885. The grade *na- seems also to appear in Old High German una-holda and una-odhi (cf. infra, VI).

« ZurückWeiter »