Homer does not elsewhere use the word ẞpaxús. But he too did not take βράσσων as the comparative of βραδύς, though this word is very common in Homer, but as the participle of βράσσειν ' seethe,” βρασσόμενος, ταρασσόμενος διὰ τὸ δέος, in which certainly no one will follow him. Hence the derivation from ẞpadús did not once occur to him either, even as a possible one. So weak is the authority for this explanation. Now as we cannot hold with the same rigour to the critical principle of Aristarchus to refer the Homeric usage to a rule as strict as possible - nor, above all, to his disinclination to admit words, which in later times were commonly current, in isolated passages in Homer, inasmuch as the investigations of Friedländer (Zwei homerische Wortverzeichnisse' Leipzig 1860) show that the number of anaέ eipnμéva in Homer is unusually large, and so the decision of Aristarchus cannot carry conviction for us: we shall rather have the less hesitation in accepting the older tradition, according to which Вpáσoov means 'shorter,' in that Bpaxús is common enough in later Greek in metaphorical applications. For instance, μετὰ βραχύτητος γνώμης (Thuc. iii. 42) corresponds to our 'short-sighted' as applied to the mind. We have more than enough confirmation from another dialect, conjecturally Aeolic (Ahrens Dor. 504), in βρίσσονος (ep. βροχέως) βραχυτέρου (Hesych.) Thus Bpárowv has beyond all doubt come from βραχιων, like ἐλάσ-σων from ἐλαχιων ; and this anomaly is removed. All other instances of this change quoted from the Attic dialect are altogether dubious. But there are certainly in other dialects more traces of an interchange of oo with. And indeed or would come from 8, if the gloss of Hesychius πέσσον χωρίον Κύπριοι, πεδίον· Αἰολείς τινὲς ouaλés were quite sound, as Ahrens (66) and M. Schmidt (Hesych.) suppose it to be. This origin for the group of sounds is less certain in yáoσav ýdovýv (Hesych.), for though in the case of this word we naturally think first of rt. 673 1 Cobet Mnemos.' N. S. iv. 236 comes to the same conclusion, as so often without any regard to German science. Gradually Bpánowv 'shorter' is finding its way even into our school books (Franke [and Monro] on II. K 226). Fad (cFad No. 252), with y = F (cp. p. 597), yet it is also possible that it is akin to yné-w (No. 122), and that yáooa is from yao-ja (cp. gaud-iu-m). But the Tarentines (Ahrens 'Dor.' 98) said not only caλníσow for σaλnícw, where a y, and originally, as we shall see immediately, a κ occurred, but also φράσσω for φράζω. In this very word φράζω, however, the δ which appears in ἀρι-φραδής, πέ-φραδ-ο-ν and other words may possibly have originated in 7. In 'Ztschr.' iv. 237 I have attempted to trace the rt. ppad back tо прат 1, and to establish its identity with the Lith, prat 'understand' (pra-n-t-ù ‘I mark,' próta-s 'insight'), and Goth. frath-s understanding,' frath-jan 'understand.' Similarly Fick i3. 679. I did not, however, venture to place this case among the certain instances of aspiration and the extremely rare softening of 7 to 8. In the Tarentine ppáσow and the Boeotian pάTT (Corinna apud Eustath. ad 'Od.' p. 1654, 25) the harder sound may have been retained. The case is probably the same with Xíoσov, which is explained in Hesychius inter alia by eλaooov, and in this meaning is to be regarded as equivalent to oλigov. We saw on p. 534 that the y of oλíyo-s is softened from κ. But for the majority of the dialectic forms, which we have every reason to separate from those of ordinary Greek, another explanation is also possible, which I regard as on the whole more probable. The Romans as a rule expressed a in the middle of a word by 88: atticisso, massa. The ss must undoubtedly have properly expressed the doubled soft sibilant, for which there was no special character. How if it was much the same with those Tarentine, Cyprian, and Aeolic forms? In that case these dialects, earlier than the others, allowed the dz (), at any rate in certain forms, to pass into zz, that is, into the very same soft sibilant, which in Modern Greek is denoted by ¿, only pronounced more thickly, and therefore felt to be doubled. In other instances too the character σ must have occasionally carried with it the soft sibilant, e. g. in Σμúpvŋ (as compared with Zμúpvn). But such softenings cannot guide us at all with regard to the Attic dialect, in which oo 1 Lat. inter-pre-s I now prefer to omit. Cp. Bréal 'Mémoires' iii. 163 ff. and were strictly distinct sounds. The explanation here offered also suits well the statements that we have with regard to a reverse change, which sometimes occurs, and that in the same dialects: Tarent. ȧvágw= áváσow, etc. (Ahr. 'Dor.' 101), and also the fact that the later Greek language shows a similar fluctuation: ovpíoow being found after 674 Christ for the older σvpígo, etc., and conversely in modern Greek -afw, -¿w (i. e. azo, izo) very commonly for the old -ασσω, σow (Ascoli Krit. Stud.' 354). φράσσω In 13 primitive verbs and 7 verbs derived from noun-stems σo (TT) seems to come from y. In the case of three of the primitive verbs, Greek itself seems to show a fluctuation between the harder and the softer sound, i. e. in noσw, the later by-form of пýуvvμ (No. 343), and áσoaλos (p. 523), σάττω (σάκος by the side of σάγη) and ὀρύσσω (p. 529): in three, viz. in párow = farc-io (No. 413), in μáoow (No. 455) and in σow (No. 367), the other languages give us forms with k which are undoubtedly akin. For a fourth verb too, páσow, the Lith. perk-ù 'I sell' is of itself enough to make it probable that we must regard рāк as the stem, derived from the shorter st. pa, which is preserved in π-ρá-σк-ш and in ἔ-πρα-σεν· ἐπραγματεύετο, recorded by Hesych. (ep. No. 358). But on Greek soil too at least one relic of the st. πρακ has been preserved in an unaltered form: πρᾶκο-s, or as it would probably be more correctly accentuated, after the analogy of πηγό-ς, λοιπός, στιλβός, πρακός ‘C. Ι.' 1702 1. 4. The word is equivalent, as Boeckh has seen, to páктiμо-s oг πрактíμо-s (also a Delphic expression), and therefore means 'liable to the exaction of a penalty,' 'punishable' (cp. eioπpáσσειν). σew). The form typń-σow, which occurs even in Homer [1 550 and (in the same verse) P 660], must have come from the shorter stem preserved in type-70, by means of an expanding κ, just as πрāк comes from πpa. The case is much the same with poo, the late by-form of рýy-vvμι (ẹp. p. 542), because of pák-os, only that here the priority of the κ cannot be proved so decisively. Hence for 7 or 8 out of 12 verbs the conjecture is justified, that the reason for the hard sibilant lay in a K originally present. As for the four or five remaining verbs, σpáτT∞, which from Plato downwards becomes more common = than the older opácw, seems to have taken the place of opago in order that too many sibilants might be avoided: and we must remember in this case, that 88 was unknown to the Attic dialect. Hence the 77 is here somewhat in the same position as in the Cretan Trva (p. 621). Of the remaining words there is only ráσow, the origin of which is still doubtful, and which is not indeed used frequently until after the time of Homer. ἄσσω (κατάσσω) ἄγνυμι cannot be proved to exist until after the time of Augustus, púσow = Opúyw is not by any means certain in Theocritus: [Ahrens and Meineke read opúуovтos in xii. 9. From Fritzsch's note it seems that pрÚTTоντоs here rests only on the authority of early printed texts, the MSS. here as in vi. 16, and xi. 12 having opúуovTOS]. In these cases we need not hesitate to admit real deviation from the ancient paths on the ground of 675 the apparent analogies just discussed.-As to the derivative verbs, we have for σαλάσσω the forms σάλαξ (gen. σάλακος) and σαλάκων preserved by the side of σαλαγή. ἀλλάσσω must be referred to a noun-stem ἀλλακo, which is to άλλο precisely as Skt. anja-ká to the equivalent anja (No. 524). A similar origin for μαράσσω, πλατάσσω, πτερύσσω is in no way improbable, as κ is in secondary word-formation an extremely common element, while y is an element which is in hardly any case original. For papuapúσow we learnt on p. 535 to recognize a noun-stem in -ūka as its source. It can hardly have been otherwise with the New-Attic apμÓTT∞ by the side of the older apμów. Thus the general result, recognized also by Schleicher Comp.3' 226, can hardly be contested, that with the exception of a small number of verbs, which are not yet fully elucidated, and of which some are very late, σo (77) came not from y but from κ, which κ however-after the establishment of this group of sounds-in the remaining forms, and especially between two vowels (ἐφράγην, πέπραγα), as was ἁρμόζω. T 1 Ascoli p. 326 conjectures that in some cases in this list oo came from κrj, e. g. opáσow from opay-r-j-w by a combination of the of the present with the syllable jo, je of the i-class. I will not deny the possibility. But there is hardly a certain instance of σo (TT) from κτλ. ἄνασσα and ἀνάσσω have perhaps come not from ἀνακτ, but from the shorter stem Favak (avaк-es, avaк-0-1 Angermann 'Stud.' iii. 119). explained above pp. 533 ff., was softened to y. We must notice here again, as an important fact for the chronology of the history of the language, that the latter softening is of later date than the first establishment of this group of sounds. 8. x0 from ghj. T In one indubitable instance, and that at the beginning of a word, the group xe corresponds to an Indo-Germanic ghj (Skt. hj), i. e. in xoés (No. 193) Skt. hjas, Indo-G. ghjas. The most natural explanation of the here is that a dental sound was developed before j, as in the numerous instances thoroughly discussed above, and that after the change of the gh to X, this necessarily became by the assimilating influence of the X (Ebel 'Ztschr.' xiv. 39). Ascoli Krit. Stud.' 323, 377 will not allow that the here is a true explosive,' but this seems to me dangerous, for the 7 in Tów (rt. spju), which Ascoli himself compares, was certainly an explosive.-The origin of the χθ in χθαμαλό-s compared with χαμαί, and the connexion of these words with Skt. ksham 'earth,' have been discussed under No. 183. — τρίχα must have become τριχθά by the addition of the suffix -ja, the j of which afterwards passed 676 into θ. τριχθά therefore (ep. τετραχθά) is to τρίχα ες υστάτιος to ὕστατος, λοίσθιος to λοισθος. Further combinations with regard to these words are offered by Joh. Schmidt Ztschr.' xvi. 436, and Möller Palatalreihe' p. 25, who refers i-x0ú-s with Lith. zu-v-i-s, with Fick i3. 585, to *xjv, and regards the j in this stem (cp. xa, xáσкw Van. 237) as parasitic after the X. 9. TT supposed to come from pj, bhj, bj. Theoretically it would be a probable conjecture that the labial consonants likewise united with j to form particular groups of sounds. It is true that above for good reasons we rejected the transition of a labial into the sibilant groups. But here we have to deal with a different conjecture. far as I know, Ahrens ( Formenl.' 185) was the first to maintain that the of the labial present-stems like Túπ-T-∞, ВλáπT-ш, KρÚπ-T-ш came from j. Others have since followed him, especially Christ (Lautl.' 159), Grassmann Ztschr.' xi. 40, |