Published online by Cambridge University Press: 22 April 2026
The functional history of the IE negative prefix, the variants of which we conveniently designate by their central form *
-, is an integral part of the history of IE composition. A classification of IE compounds, based on historical rather than logical considerations, would include (1) old types, generally declining during the historical period of IE—bahuvrīhis and synthetic compounds (verbale Rektionskomposita), and (2) distinctly younger types—pronominal hypostatics, determinatives, and copulatives.
1 Hamilton, The negative compounds in Greek 27 (Baltimore, 1899) ; Frisk, Gebrauch des Privativpräfixes im indogermanischen Adjektiv 4, 45 (Göteborg, 1941).
2 Jacobi, Compositum und Nebensatz 83 ff. (Bonn, 1897) ; Brugmann, IF 18.59 ff. ; Petersen, IF 34.249 ff.
3 Hirt, Indogermanische Grammatik 4.67 ff. (Heidelberg, 1928) ; Frisk, IF 52.288; Petersen, IF 34.273; Risch, IF 59.1 ff., 245 ff., 292 f.; Schwyzer, Griechische Grammatik 1.453 (Munich 1939).
4 Jacobi 1 ff.; Hirt 4.25 ff.
5 The second type concerns us only indirectly, since a priori it contains no compounds with the negative prefix.
6 Chantraine, La formation des noms en grec ancien 4 (Paris, 1933).
7 Jacobi's explanation of synthetic composition.
8 Buck, Cl.Ph. 12.173 ff.
9 Frisk 4 f.
10 Carruthers, Cl.Ph. 26.181 ff.
11 Buck-Petersen, A reverse index of Greek nouns and adjectives 469 (Chicago, n.d.).
12 Cf. Frisk for Greek, Indo-Iranian, Latin, Germanic; Hamilton for Greek.
13 E.g. Hirt 4.49, 5.390, 7.69.
14 Cf. Miklosich, Vgl. Grammatik der slavischen Sprachen 4.173 (Vienna, 1868-74); Delbrück, Vgl. Syntax der indogermanischen Sprachen 2.522 f. (Strassburg, 1897).
15 Brugmann, IF 18.127.
16 Vgl. slavische Grammatik 1.676 (Göttingen, 1906-08).
17 Meillet-Vendryes, Traité de grammaire comparée des langues classiques 545 (Paris' 1927).
18 A later development seems to have confused the original picture, especially in Greek. The
bahuvrïhis date from a time when
etc. were purely adverbial expressions, as still seen in Homer:
where the ablatival genitive is self-sufficient and
merely strengthens it. With the development of prepositional value, many of the bahuvrïhis could be interpreted as hypostatics, like Slavic bez. It is uncertain in cases like
'out of breath' whether we have to do with original bahuvrïhis or with later hypostatics.
19 Fowler, The negatives of the Indo-European languages 32 (Chicago, 1896); Gray, Lg. 1.127.
20 Walde-Pokorny, Vgl. Wb. der indogermanischen Sprachen 1.95 (Berlin und Leipzig, 1930); Pokorny, Indogermanisches etymologisches Wb. 280 (Bern, 1949-).
21 Walde-Pokorny 1.58; Pokorny 39.
22 Cf. in general Sturtevant, Indo-Hittite laryngeals (Baltimore, 1942); Lehmann, Proto-Indo-European phonology (Austin, 1952).
22a In these reconstructions, X = a voiceless a-colored laryngeal (Hitt. ḫ(ḫ) or nil), Y = a voiceless e-colored laryngeal (lost in Hittite), Z = a voiced e-colored laryngeal (Hitt. ḫ or nil).
23 μαλακός had an initial voiceless laryngeal, as is shown by ‘position’ in Homeric meter (
Od. 1.56). Its allophonic history in formulae is as follows: 1. /HṃlS > /ṃlS', 2. possibly also to /mḷlS; 3. NHṃls > NṃlS (after heavy syllable);4 to NmḷlS (after light syllable); 5. SHmḷlS remains in *hμαλακος; 6. in other IE languages it generally gives SmḷlS. No traces survive of 1 and 3. Upon spread to position after short syllabic, 2, 4, and 6 developed the allophone SmlS seen in
etc., which was generalized in the dialects. Similar to these is the allophonic origin of
Il. 12.198),
(στήθɛσσι λασíοισι, Il. 1.189).
24 E.g. Brugmann, KVG 310 (1933); Hirt 4.50.
25 A useful though incomplete list by Gray, Lg. 1,119 f.
26 Between non-syllabics the laryngeal was lost here as elsewhere (vόθosi). Cases where
‘makes position’ originate in initial or postsyllabic position; the other variant is seen in
Il. 17.243.