STOP Early Journal Content on JSTOR, Free to Anyone in the World This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in the world by JSTOR. Known as the Early Journal Content, this set of works include research articles, news, letters, and other writings published in more than 200 of the oldest leading academic journals. The works date from the mid-seventeenth to the early twentieth centuries. We encourage people to read and share the Early Journal Content openly and to tell others that this resource exists. People may post this content online or redistribute in any way for non-commercial purposes. Read more about Early Journal Content at http://about.jstor.org/participate-jstor/individuals/early- journal-content . JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people discover, use, and build upon a wide range of content through a powerful research and teaching platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please contact support@jstor.org. STUDIES IN GREEK NOUN-FORMATION Based in part upon material collected by the late A. W. Stratton, and prepared under the supervision of Carl D. Buck 1 LABIAL TERMINATIONS Words in -ri or -o (also -tj>d) — in -0jjj or -(pas, gen. -os and -ov, gen. -ov. By E. H. Sturtevant The existence of an Indo-European suffix -bho has long been recognized, and in recent years it has received rather frequent treat- ment. There is an excellent summary of our present knowledge of the subject in the new edition of the second volume of Brugmann's Grundriss (Part I, pp. 386 ff . ; see his references to the more impor- tant literature). Upon this we must base our introductory remarks on the use of the suffix in Indo-European in general (pp. 197 to 199), although we shall depart from Brugmann's treatment at several points, and our third paragraph is entirely new. Only a few words containing the suffix can with any certainty be ascribed to the parent speech. In the first rank stand epios (from *eribhos), Ir. earb "goat," Swed. jarf "gulo borealis" (from *erbhos) : Lat. aries, Lith. eras "lamb"; a\6?) "young deer," Cymr. elain "hind," OB. jelen "deer" (see Osthoff Etymologische Parerga 305); a-6v, OPr. subs "selbst, eigen" (from *s e bho-), Goth. sibja, NHG. Sippe, OB. seb-ru "free peasant," Lat. Sab-ini : *se- reflexive; and Slavic svobo-da, svobo-dd "freedom," Suebi, OHG. Swaba, NHG. Schwaben : *sue- reflexive. 2 1 See Introductory Note, CP. V. 323 ff. 2 Cf . Xiiden Studien z. altind. u. vergl. Sprachgeschichte 54 and Solmsen Ifnter- suchungem z. griech. Laut- u. Verslehre 197 f. Solmsen derives Skt. sabhd "assembly" also from the reflexive stem, but it seems preferable to connect it with the demonstra- tive *se- (see p. 199). Brugmann, Cfrundr. 2. I 2 . 388 (cf . also Demonstrativpronomina [Classical Philology VI, April, 1911] 197 198 E. H. Sturtevant It is impossible to draw a sharp distinction between words with an Indo-European suffix and those with a "root-determinative." For example, icvdo<; "a carding comb," Kvd-rnm "card wool," Lith. knibti "pluck," knebti "pinch" (Sutterlin IF. 25. 66), clearly belong with Kvfjv "scrape, scratch"; accordingly Kvdfyos may originally have been an abstract derivative in -bho, while dyvafos, N.T., may possibly represent an Indo-European adjective. Simi- larly we have (o-)icepao$, cr^epa^os "mockery, abuse," Hesych., Lett, schk'erbs "rude," skarbs "sharp, etc." : Keipco, Lat. careo, etc. (Brugmann IF. 15. 97). Tpvrj "luxury," properly "destruction," rpv<})0<; to "fragment" (perhaps originally masculine, cf. e'Sa^os), pinna, Lett, drubascha "fragment," drubasas "splinter" stand beside 0pava> "break," Opavcnos "frangible," Lett, drupi "frag- ments," drupu "fall to pieces" (cf. p. 209), etc. Scholars have been more successful in tracing certain uses of the suffix back to pro-ethnic times. Thus, all branches of the Indo- European stock employ it in animal names. Adjectives, though rather less numerous except in Baltic-Slavic, are quite as wide- spread. It is hardly possible to classify their meanings except for the several languages separately. 1 Abstracts occur with some fre- quency in Baltic-Slavic and Greek, and occasionally elsewhere. The abstracts and adjectives seem to be very closely related. The suffix-form -tbha- common in abstracts in Baltic-Slavic, stands beside -ibho in Lithuanian and Greek adjectives. 2 Such pairs as Lith. dalybas "sharing" : dalybos (plural) "division," especially of an inheritance, and Goth, halhs " half" : halba fern. "Seite, Teil" are significant. The animal names probably represent, in part at least, a specialization of the adjectives. For example, several animal names 31), on the other hand, connects only Slavic svobo-da with the reflexive stem. Appar- ently he loses sight of the fact that these etymologies are quite independent of Solm- sen's theory that I-E. su- became s- under certain unknown conditions, since the semantic equivalence of *smp- and *se- is indubitable. •Brugmann, toe. cit., lays stress upon the color words; but, of the half-dozen he names, Lat. galbus is very likely a Celtic loan word with suffix -no (Walde), while &\wos may be specifically Greek. 2 The constant length of the -i- in Baltic is not original, as is shown by Slavic -«-. Lithuanian -yba, -ybS in abstracts may have been influenced by the equivalent suffix ■ysta, -ysU. The adjectives in -ybas clearly stand in some secondary relation to those in -yvas, Slavic -ivii (seeLeskien Bildung d. Nomina im Lit. 353). But Greek a4plo<; "black- bird" is perhaps onomatopoetic in its first member. Prellwitz sug- gests a derivation from *kukiubhos (with dissimilation of the first vowel?), and compares Skt. kukkubhas "phasianus gallus." Brug- mann, Grundr. 2. I 2 . 390, sees onomatopoeia in the bird names Ken-o<;, Kov/covo<;) seems to be an Egyptian word for eirotjr (Thompson op. cit.; cf. Horapollo 1. 55). aira6<;- throyjr, to Spveov, Hesych., is very likely a variant of its synonym, under the influence of the animal names in -$o?. The Hesychian Kopafyos- Troths opvis is derived from the stem seen in Kopavi], icopal;, Lat. cornix. oi' . . . . oi aypioi, Terriyes, Hesych., is to be connected with aiyrfkos "mute," Used of animals in Arist. H.A. 1. 1. 29, and with \ol, while L. Dindorf (Thesaurus 7. 216) prefers triyaKioi. Doubtless one or the other is right as far as the first definition (ol &o<;, Hesych., belongs with «e/ua? -d8o<;, Hom.,-J— The suffix is -d-, although it is possible that the influence of the primi- tive carried the word over to the -a? -d8o<; declension. xpvo<;, Marcell. Sid., and x/>vffo<£o9, CGL., a fish, Lat. "aurata," show the influence of the suffix, but they are probably modifications of %pv- trofypw, the name which is usually applied to the fish in question, rather than new derivatives from xpva6<;. icipa7n?| is a derivative of the word which precedes it in Hesychius, idpa' akdnrni;, Adjemve; ; unless we should rather suspect a corruption of the text. An article in Antonius Monachus' lexicon runs: fiovfov opveov. /3oi)$o? e7rip6va)v eiprjrai. Whether the bird in question was merely a creature of the imagina- tion or a real bird of sluggish habits, its name is derived from /Sou?. The late Kovrd\aa<: as the " insect on the twigs." 6p$o<;, 1 a kind of fish, is Attic 6pd>s. 2 Gf . Fiok KZ. 42. 288. This view now seems to me more probable than the one which I expressed in CP. V. 331. Studies in Gheek Noun-Formation 201 serpents) ical r&v ivrofioov ixSvvei to y-fjpas oaa e«8wet, olov eri\r). .... The change from yfr- to «r- would have to be explained as an instance of dissimilation (from *TT(ri\irhr]). Both meaning and ety- mology of Ti^>7), Ar. Ach. 920, 925, are uncertain. It is indeed said by Phrynichus, 359 Ruth., and the scholiast on the passage to be equal to a-i\rj ; and probably we should read Titfrq instead of o-fyt) in Herodian 1. 108. 14 L.: iyivero 8e eWeityis, as irapci rols 'Atti- koIs tj <^Ty{ri trap' rjfiiv Xejofievrj But the two words can hardly have been equivalent in Aristophanes' day. In the Acharnians, Nicharchus propounds a scheme to burn the ships by putting a wick is ri^-qv, and sending it to the dockyard through a sewer. Dicaeopolis replies : o-ekayolvr' av virb ricprjs re kol OpvaX- \lhos; Clearly a book worm is out of the question; a glow worm would fit the context admirably. 1 We may at least, therefore, be sure that ai\r} and tm£»? have no etymological connection, and that in the classical period they differed in meaning. The semantic con- fusion which we have noticed in Herodian and elsewhere was respon- sible for Lucian's ri\r) and Ki8ar), KivSdtpiov, Hesych., and a/c(v8aos, Oudendorp's conjecture for aicivSaicos in Aelian. The forms with a nasal represent a contamination with the synonymous Kiva8os, but the etymology of Ki8aij>os itself is unknown. Other obscure words are (aa-)KaXaof opveis rives, Hesych., /3aia)/j,(f)af al alyes eV iepariKols, Hesych., os' 0r/pi8iov fwcpov, oirolov ifxirCs, Hesych., icvtyrj, a kind of shrimp, Epich.,+, KaQovfy^v 3 aXcoireKa, Hesych. Mlpaid)rr}s, Lesb. 'TLppafadrrrjs, an epithet of Dionysus, h. Horn., +, suggests *eipaos "goat" (Froehde Z?J5. 21. 199), or "fox" (Ehrlich KZ. 39. 567), or the like. a.Ka\rjri l and (a)aKa\dr] "sea-nettle," as well as "nettle," and y6fios- i%#«5?, Hesych. : y6fuf>os "nail" are animal names only in 1 Herwerden, Lex. Or. Suppl. 2 , identifies Aristophanes' rlo<;, the name of an insect, is a variant of the adjective o$ (see below). There are also several compound animal names in the following word-lists, as rpayeXao<;, cr^ivorpo^o';, ki^Xokoo-ctv^o^. 2. Adjectives. — Several of the Greek adjectives with suffix -$o? have a derogatory force. But there is hardly sufficient evidence that this is anything more than accidental and that such a force had become consciously associated with the suffix. Xi'o-o? "a7ri>7o?," Moer., Tzetz., as substantive, Xiaipa rd lax" 1 ••••■, E.M., is cited as Attic for XtWo?. A derogatory connotation attaches to the latter form in Ar. Ran. 826, Xio-jtt) jXaxraa (of Euripides), to the compounds, Xio-iroirvyos and XunroTrvl;, and to viroXia-iros in Ar. Eq. 1368, 7roXX.ot9 7' vTToXurirois irvyiBwicnv i^apiaw. Elsewhere Xwm-o? has no disagreeable connotation, and vttoXio-^os in Philostr. De Gymn. 35, ra 8' vtto t to"%ta> y-r\ff viroXicrfya ecnw /Arjr' av irepiTrd, is inno- cent enough. Xto-o? is a derivative of Xipd<: "dry, parched, lean," Hipp., implies a base *o7ceXt<£o?, from axeXXco "make dry, parch." The form aKeXep6<:, Erot., shows assimilation of 1 to the preceding e, and. so does the primitive adjective in its two recorded substantival uses: KeXeo'?, Hegemo- nius,+, is one of the numerous late Greek words for "leper"; ovee- Xe(j>op (Schmidt; MS a/ceXecf>ep) • ySoXou 8vop,a, Hesych., was perhaps a throw in which some or all of the tali rested upon the narrow side, hence the "narrow throw." According to Antonius Monachus and Suidas, the phrase os, a small island (cf. Fick Vorgriech. Ortsnamen 57), in ov? a kind of wormwood, and in the Hesychian gloss: Septc^o?- ' A.puno$dvr}<; ttjv AaKeSai/xova 2e/3«£oi\ ecrri Se ical 7ro'a X L. and S. cite os tj from Diosc. 3. 27 ( = 3. 24 Spengel). That passage, how- ever, contains the accusative singular without the article. The gender is shown to be neuter by the article in Galen 13. 126 K. Studies in Greek Noun-Formation 203 ov ttjv AatceSatfiova, Sea to o-kXtjp&s %rjv Apparently aept^>oo<; "a dried-up old woman"; while o-epiov "wormwood" and 2e/3t^o?, the island name, as well as Septc^o? = AaKeSai/Awv, show a derived sense, "unpleasant, harsh" (cf. f»7/w opposed to fjhrk in Eur. Andr. 784). o$, an insect, differs only in the form of the suffix. For the root, Prellwitz suggests a comparison with fe/ao? "dry." os "barren," of women, animals, and plants, Ar.,-f, is a derivative of the synonymous o-reipos; likewise o<; "firm" from o-repeos. dypoov opeivov, Galen Lex. Hipp., is derived from dypfc. l ievo ' i i Hesych., prob- ably represents f. There remain four color words, of which one, a\o'?, is certainly inherited. dXco6^ "white," Hesych., is made on the same root as a\o'?, Lith. alvas, etc., and is probably an inherited form (see Brug- mann Grundr. 2. I 2 . 388). dpyvfos, 1 a Homeric epithet of sheep, contains the stem of apyvptx; and Skt. drjunas "white." Kivupw 6 7row«\o?, Anton. Mon., Suid., seems to go with Kivem, icivvfiai "go" and Kivvo-o-ofiai "waver, sway." 3. Abstract substantives. — Most of the Greek abstracts in our suffix differ from those in Baltic-Slavic in having the o-stem and masculine gender. All the masculines have recessive accent. The inherited forms (o-)/eepa09, cr%e/)a^>09, icvdo<;, yvdo$, rpiufxy;, and rpvfyrj may contain the suffix (see p. 198). It is certainly present in the following (Brugmann Grundr. 2. I 2 . 390): /c6Xao<; "buffet" : koXos, KoXd^m, etc.; <£\??vao? "ineptiae" : Xr)vea>, (pXrjvdw, Xn- vv6a>v "whirling" implies an abstract substantive *feXwo? and its derivatives we have a bewildering array of meanings. a-KipcKfros itself means "trickery, cheating" in Hipponax, "gambler" and "licentious" according to Herodian 1. 225. 13 f., 2. 581. 22 ff. L., and "dice box" according to the E.M. etv means "play at dice" (Hesych.); evTi]<;, "dice player" (Amphis); and (ncipaeiov was situated. The various recorded meanings can most easily be arranged if we start with o'; in an abstract sense "rascality, gambling." This developed into the concrete "dice box" and "rascal, gambler." The connecting link between liclpov and (TKip-a-^>v "frequenter of the liclpov." Several Greek abstracts show the feminine which appears con- stantly in Baltic-Slavic, jpvfyrj has already been mentioned, kvtj^t] "the itch," LXX, Suid., is a specifically Greek derivative from Kvrjv "scratch, tickle," and has no direct connection with the inherited Kvdrj, we find a masculine Kopvov, the name of a mountain near Epidaurus, Paus. 2. 28. 2, suggests that the word is Epidaurian rather than Megarian. A wholly different meaning of the o-stem appears in the Hesychian gloss, ic6pvos- icopv/xfto*; jvvaiKelo 1 ;. The force of icopvfyr), as of Kopvfos, is usually concrete, "summit, top"; but a trace of the abstract meaning "height" is preserved in Pindar's Kopvcfra \6>ycov, Xoywv icopvfyal "the climax," icopval TroXemv "the best of cities," etc.; although here too the concrete idea was present to consciousness, as is shown by Plato's use of the noun with the verb "to go" (Crat. 415 A) — 1 Solmsen Untersuehungen z. griech. Laut- u. Verslehre 325. His alternative sug- gestion of an original adjective is less probable, since it would not correspond in meaning with the Greek adjectives in -r)v &v eipvica, .... iirl ttjv ieopvr)v Sel to<: "gambler," Xrjvaoi] "summit." Other concrete substantives, as we have seen above (pp. 202 ff.), sprang from the substantival use of adjectives in -$o?, as ice\e6s "leper" and aiceXefyop "a throw at dice" from *o-Ke\t0o? "dry, thin"; 2e/«<£o?, an island, and aepifyov "wormwood" from o-epio<; "dry, unpleasant"; Xiov "(small) buttock" from Xt6s "a dull-white leprosy" from aX6<; "white." After the model of these and similar words, the suffix -<£o?, -r) was used to form concrete substantives. 7a\e$o?, Interpol. Diosc, is a variant of the plant names jdXiov, yaXeo/38oXov, and 7aXeo'i/rts. Walde, s.v. galea, derives these three words from yaXev "weasel," in the secondary sense of "helmet" which is seen in the borrowed Latin galea. ydXeov ovpd, Hesych., is a derivative of eyiciXXov ovpdv, Hesych. aypfyv 1 "harrow, rake," with its derivative ayplva, Anth. P. 6. 297, and aypio<; "grapple," Const. Porph., may be from aypevw, aype'co. aypio<; "wild olive," Hesych., whence comes 'Aypia- 'Advvd, Hesych., seems to be connected with dypios "wild." ovXacfrw ve/epos, Hesych., suggests oXXv/m, ovXo/j.vv, etc. rap^cjiv' rdpirv, Hesych., goes with rdpirrj, and rapo-os, rapp6s "crate, basket." "AXdv^ov or " A.\6r)o$, the name of a district in Troezenia, belongs with dXdero, cnraXOi]aea-0ov "heal," Horn., ak0r)aic6o$ may be regarded as the short forms of full names the second member of which begins with , e.g., "Aw^o? beside 'Avndvr)s, 'Ai'Tt'^tXo?, 'Avrufr&v, etc., or STapi-o^o? beside lTpa,TOdvr)<; (so Fick-Bechtel 255). But there are several for which no such full names are quotable, and there is no reason why we should not recognize the adjective suffix -<£b? as one of those which, like -Xos, etc. (cf. Fick-Bechtel 24 ff.), may be employed in short names derived from full names or in those of the class treated of by Bechtel op. cit. Thus TeKTao<;, which Fick-Bechtel say implies a full name TeKTa-cfrwv or the like, is more probably formed directly from the stem of re/crav ; similarly "Acrra<£o? from "Ao-twv, while "Aaro^o? is a short form of 'Ao-to-^Xo? or the like; Kepica.o$, a mythical ances- tor of the Bhodians, from icepicv&i- iepa%, Hesych. (Fick BB. 21. 280) ; Ti5Xt<£os, a mythical shepherd, from ti5Xj?, tvTuk. Is €>e\epocf>o<;, name of a poet, formed from a *0e'Xe/jo? "wishing, wilful" ? There are some examples of similar extensions in -a$a?, or 'E7ro'i>»79 in the adopted Attic form, IG. 12. 3. 1187, Studies in Greek Noun-Formation 207 must be of the same type (from an *"E'7roH'?). 'Ow^«, Herodian 1. 57. 15, 2. 655. 4, may be a nickname derived from ow. 6. Regressive formations from diminutives in -acpiov. — Such a form as i\d-iov, Ar.,-|-, diminutive of e\a4>os, gave rise to drjp-dfaov from 6r)p, and to a whole series of diminutives having no relation to animal names, e.g., frpdfaov, ■xjpvcdfyiov, yppdfyiov, /jlvt) p,d(f)iov, etc. In a similar fashion arose the diminutives in -iiov and -vfaov. Cf. now especially Walter Petersen Greek Diminutives in -ION, pp. 276 ff . In late times these were felt as diminutives in -iov, and by regressive derivation there arose many new forms in -$os, -v, -a. So modern Greek \o>pd^>a, £vpda, from ^copd^iov, l;vpdiov, like viica ical via 'EXX^w/ca 2. 139 ff. 0ijpao<;, a late medical term for a small insect, arose in the same way, and deao$, 6eiaiov, rather than an original formation in -<£o? and the source of Oedfaov, as assumed by Petersen loc. cit. The form of the words that we have been discussing calls for some treatment. It will be enough, however, to summarize briefly from that point of view some of the facts already stated in detail. In par- ticular we must notice the two dissyllabic forms of the suffix, -io<; and -afyos. The suffix -ibho- appears in Baltic-Slavic abstracts and Lithuanian adjectives as well as in Greek adjectives (see p. 198). Probably it arose from the incorrect analysis of derivatives of z-stems (cf. e/st-^o? : Lat. aries). We have discussed aep-los (beside the substantive creptyos), crrepi-^o 1 ;, Kivi-(f>o<;, and *o7£eX-t$o?. Tv\-io<; ( : tvX?7, tuXo? "callus") was probably in the first place an adjective meaning " callous." In a-y/ai-^o? the i belongs to the base. The suffix -<£os was frequently appended to the weak form of nasal stems as in e\a-o<; : eXXo? from *eXi>09 (above, p. 197), Kopa-o<; "ras- cality" seems to be from *v "frequenter of the 1/cipov" (see p. 204), (f)Xi]va-o<; "ineptiae" may be from *o$, from *KpoTos which appears in xpva-a(}>tov. Nearly all words containing our suffix have recessive accent. The exceptions are: the inherited a\6<; (with its congener aXax^o'?), the feminines Kvrj, /cadovifrrj, Tpvi], and Kopvr/, of which only the last certainly contains the suffix, the Hesychian glosses criya\6<; and Kep,d<;, which may have been influenced by the primitives o-io? and the obscure word p,a6<;, and finally, o-th^o'?. VARIATION BETWEEN THE THREE LABIAL MUTES So many of the words in -r), -»??, and -o<: show variants with /8 or tt, that we are compelled to inquire into the causes of the varia- tion. For completeness' sake we shall include in the discussion a number of words which do not properly fall under the topic to which these studies are devoted; and we shall find it convenient to treat here the few cases of variation between ir and /S. 1. ir : /3. — The Indo-European variation between tenuis and media, particularly frequent in root-finals (Brugmann Grundr. I 2 . 629 ff.), is not often reflected in the Greek labial stems. It occurs, for example, in aKcnrdvr) "mattock," (o-)«a'7reTo? "trench," etc., Goth. skaban "shear" : Lat. scabo, Lith. skabus "cutting" (Brugmann Grundr. I 2 . 629); and in Lat. stlpo, Lith. stimpxi "stiffen," Eng. stiff, etc. : aTifios "path," trrtyS^ "hoar frost," o-Tet/36), Lith. staibus "strong," etc. (Walde). The y3 of /3\dftr), beside Cretan a/3\o7re'?, afiXoiria, KarafSXairecrOai, with original ir as is shown by Lat. mulco, is due to assimilation to the initial consonant of the root (cf. vulgar Eng. Babdist for Baptist). NtoTrt}, on an Attic vase, gets its ir from the rather numerous mythological names in -dirt) (Meisterh. 3 77. 3). ypvyfr, ypviros, beside 7/sv/3o? with original y3, shows the influence of ypviros "crooked" (see above, p. 200). aarpawr) for currpdfiri "saddle," Anon, in Rhett. Gr. 8. 668, involves a rather grotesque popular etymology. «?j/3o? = icfj7ro<;, aTvfiri — o-tvitt), and «o'XXa/3o? = KoWoyjr, KoWoiroi show the influence of the suffix -/So?, -fa (see CP. V. 332, 333, 341). Studies in Greek Noun-Formation 209 2. 7r : . — A larger number of Greek words show a variation between ir and , and in some cases this may go back to proethnic times. There was in the parent speech an occasional interchange of tenuis and tenuis aspirata (Brugmann Grundr. I 2 . 632 f.), as in 7r\a- tw : Skt. ppthus, Lat. rota : Skt. rdthas, Av. rapo "wagon." We have an additional example of this kind in Lat. pampinus, papilla, Skt. pippalas "berry," etc. : 7ro/a<£d? "blister," Trop,(p6\vi-, ire/x^il; "bubble" (see Walde s.v. pampinus). The interchange of Indo-European bh and ph appears in rpvo<; "stile," do/iai "scratch, sketch," Schol. Ar., Lat. scribo, OHG. scriban, OS. scrtban, OFris. skriva "write" : Lett, skripat "to scratch," skrlpsts "curved knife," OIsl. (h)rifa "to scratch." If, however, we assume I-E. *sgnph- (Prellwitz), we can trace all the forms to one original, for the Germanic words with b are probably Latin loan words, as is indicated by their meaning and by the / of OIsl. (h)rifa. The meaning of OFris. skriva, OE. serif an "to enjoin punishment upon," OIsl. skript "confession, punishment" may easily be derived from the meaning "write"; cf. ypd^eaOai "to indict" and ypa4>v "indictment." reOijTra "be astonished," Horn., beside eraov, Horn., ra^o? "astonishment," Horn., owes its tr as well as its t to the dissimila- tion of aspirates. More frequently the Greek variation between <£ and ir is due to analogy. As Osthoff, Perfect 301, long ago showed, the

a>, beside \i-ira, Xnro?, Xnrapo*;, etc., arose from the analogy of ipeyfrio, TJpeiJra : ipe, and the like. Siniilarly the presents in -7tto> form the connecting link between forms with original tt and later forms 1 It is possible that drup- represents I-E. *dhrup-, but an Indo-European variation between bh and p is problematic. The relationship of Eng. drop, etc. (I-E. *dhrub-), is not certain. It was suggested above (p. 198) that the forms with bh contain the suffix -bho, -bhd. If so, this group differs from many others that exhibit "root determinatives" only in that we can identify one of the " determinatives " with a formative element that continued to be productive. In any case, the words certainly belong under this head. 210 E. H. Sturtevant with <£. Thus we have Ka-wrf "crib, manger," Horn., but eyKaos "mouthful," Eupolis (eyicd-ma) "gulp down," perfect iy/ceiccMpa) ; aKairdvTj "mattock," (o-)/ca7reT09 "trench," Lat. scapula "shoulder blade," OB. kopati "dig," but o-icacfrrj, cncdcJHx; "a digging," o-/caev<;, icncdcfivv, etc. (cr/cdiTTa) ; Solmsen Beitr. z. griech. Wortforschung 196 ff.); pairfc "needle," Epich. 138 Kaib., Lith. verpu "spin," but pari "a stitching, seam," pafaifc, etc. (pdirrea; cf. Prellwitz); pnrr\, Horn., +, but pirj "a throw," Lye, ava-7r6pio<; "a/ie/u7TTo?," papyri, epptyi)v, Eur., -j-, piriaoiiai, LXX, + (pfcTco); (a)Kvnro<; "niggardly," Lucillius in Anth. P. 11. 172, (o-)/cw7tot)?9, Ps.-Hipp., -}-, but o?, Phryn., +, Tvfymv, Mto-6yvi of o-icvMpos, 1 beside (a)Kvnr6<; "dim- sighted," KeKvtTrto/jievoc . . . . f} kvittoI row b4>8a\p,ov<;, Hesych., is due to the pair a- 'Adrjvd, Hesych., suggests a popular connec- tion with the phrase aypios iWo?. Perhaps Cyprian e\o<;- fiovrv- pov, Hesych., beside eXirov e\aiov, areap, eiidtjvia, Hesych., owes its to the influence of akei(f>a> and the related forms. Doric pd(j>a "a large radish" and patjxivfc : pdirvs "turnip," Lat. rapum, etc., were perhaps connected in the popular mind with pak "needle," an implement that was not unlike a radish in shape. From their first appearance until the present day the two words 7pf7ro? "fishing net," Artemid., Diog. L., -f-, and yptyos "riddle," Ar., +, have been kept distinct, with the exception of two occur- rences of yptyos "net" in Plut. 2. 471 D and Opp. H. 3. 80. 2 The iv of ypliro<; is further attested by ypiirem "fisher," Theocr., -f-, Tpi-irav, the name of a fisherman, Leonidas in Anth. P. 7. 504, ypnrrjk, Antipater in Anth. P., ypnrewo, ypC-marna, Anton. Mon., ypnri£a>, Hesych. It seems more likely, then, that the copyists have blundered than that Plutarch and Oppian really knew a by-form yptyos — 7/317T0S. 1 On the provenience of the word, Bee Herwerden. 2 yipios in Pap. Tebt. 2. 486 ia very dubious, and the perfect participle yeypiipiis in Hesych. is in line with irtwofujxi, etc. Studies in Greek Noun-Formation 211 The Latin cupa "cask, vat" (kovtto,, IG. 14. 1342, rfpi-Kofot), inscr. in Comptes Rendus de VAcad. des Inscr. 1866, 384) appears as Kovov in ostraka and papyri (95 a.d., +) and in the Apophthegmata Patrum. 1 The word has evidently been brought into semantic rela- tionship with Kov(f>o(Sa^)oi = anfyndirrfi, Pap. Ox. 2. 298. 9, 10 (first century a.d.), is for *afi(f>iSa'7ro<;. Cf. a,n4>iT(nro<;,~L~KK,-\-, and Sams = Tairrj 1 ;. The second is due to assimilation. Cf., with assimilation in the opposite direction, 'AvOiXoxo? — 'Ait/Xo^o?. We have noticed the effect of the suffix -o?, -r) in airao<; : eVoi/r (p. 199) and Tap^rj : tapiri) (p. 205). icdfos, E.M., Eust., beside kcLtto<; "breath," is of doubtful authenticity. Apparently the word was coined to explain the Homeric KeKafyrjora Ovfiov. Eust., 1280. 34 f., says: ' ' A/iroKatrvaai, Se yfrv^Tjv to airotrvevcrai. Kaxpo<; yap (pao~i to Trvev/ia, ifj oS to iceicaT]o'Ta 6vp*6v. to 8' ax/TO ical 6 «a7ro?, ef o5 Kal 6 /c^7ros, o>? aWa^ov eipijTcu. 3 I doubt the connection of otjhto? "stick" with o-tv<$>ok6tto6<; beside p,ao-Tpoir6<;, see p. 206. 3. ft : . — One of the best-known features of the Macedonian dialect is the development of the Indo-European mediae aspiratae into mediae. In the present connection we need mention only those words in which Macedonian /? stands as stem final beside general Greek (f>. The most certain instances are /ea/j«/3o? : ^Kapobrj, ko/j,- /3o? : 7o'/io?, 7o/i<£to?, and vCfia : vUpa. (See Hoffmann Die Make- donen 28, 37, 52.) For the personal name K6ppa@o<;, Hoffmann, op. tit. 146, thinks of the suffix -<£o?, but the suffix -yQo? is more common in personal names (see CP. V. 337 ff.). We may add to Hoffmann's material /a//3o? = /euro's, Orion 38. 6 (Herwerden), and the geographi- cal name Ottolobos = 'Oktooao^o?, which Livy (31. 36, 40, 44. 3) uses 1 The papyri and ostraka show that Sophocles, Lexicon s.v., was wrong in assum- ing a nominative Kov, vIttt(o, vitrTpov, ^epvi-^r, ^epvt^ov, Skt. nenekti, nejayati "wash," Av. naenizaiti "wash something off," etc. In that case there would be nothing in the form of the word to suggest Macedonian origin. The Indo-European interchange between media and media aspi- rata appears, for example, in 6'XySo? "wealth," Skt. drjati "earn" : a\(f>ri "gain," Skt. drhati "deserve," arghd "worth"; kv/3o<; "a hollow above the hips of cattle," Eng. hip, Lat. cubo : kv<$>6<; "stooping, hump-backed" (Walde, s.v. cubitum); <7T/)a/3o? "squinting," arpo^o<; " a whirling," o-T/ae/SXo? " distorted," Lat. (originally Umbrian) strebula "meat from the hips of oxen" : a-rperi "covering," papyrus of the first century a.d., in Studies in Gbeek Noun-Formation 213 the phrase wo ica\vr)<; alyiaXov "from high water mark" (or pos- sibly "from low water mark"), beside ko.\v/3t) "hut, cell," Hdt., -)-, was formed from Ka\xnrTrj : Kpinrra, Tpvcj>i] : dpviTTO), /3ao<;, N.T., -+-, : pdrrTco, ayvcMpos, N.T., +, : yvdirrw, ara^o? : Qd-inco, aT/3v<£o?, Alcm., : Opvirra). ico\vpov, Hesych., beside icoXvfifida) "dive," ko\i/^/3o? "diver," Ar., etc., Goth, hlaupan, NHG. laufen (Zupitza German. Gutturals 118) shows the influence of e\a/30?. If Tpfyos, which Du Cange cites from a gloss as equivalent to Tyu'jSo?, is a genuine form, it was made from Tpiyfra and erpi-^a on the model of o-Kdo<; : epetya) and tfpeyjra, etc. If ctt(o? "throng, mass" and o-tk£/)o? "firm, solid" are akin to areiftca, their is due to some similar analogical influence (cf. the ambiguous o-rarTo? "trodden down" and "tough, sturdy"). Uhlenbeck's (Etym. Worterb. d. ai. Sprache) connection of Skt. stibhis "tuft, bunch" with these words is not probable enough to warrant the assumption of a third root form beside *steip- and *$teib-. For ypv§6<; beside ypvftos, see p. 200. The reverse analogical change of original to ft also occurs. In icpvftrjcroticu, Eur., eicpvftov, iicpvftrjv, Apollod., +, icpvftrj, airoicpvfti), cnroicpvfir)o-op,ai, LXX, +, iyicpvftoo, Diod., +, Kpvj3r)\o<; Hesych., Kpvftoi, a Byzantine geographical name, etc. (cf. also Lobeck Phryn. 317), beside earlier Kpvr)B6v, Horn., icpv, pdftco, Kkefta, etc. Cf. Hatzi- dakis KZ. 27. 76. If we assume an original aspirate for ftofiftos on account of Skt. bambharas "bee" (found only in lexicons), bambhara- vas "lowing" of cows, bambharalis "fly," a persistent feeling for the onomatopoetic character of the word would help explain the change from voiceless to ft. The same consideration applies to o-ro/i/3o? • fiapvrixo<;, ftapv6oyyo<;, Galen Lex. Hipp., beside oro>£o? "bom- bast," <7To/i(/>o? "bombastic," (TTopfai;, <7TOfjuj>d^co, Ar. 1 •These words have no connection with or fi and fifi, on the basis of which a number of scholars have inferred a change in Greek itself of aspirate to media after nasals. 1 The group nv, /jlo/jlcJ)')]'), and no satisfactory limitation of the operation of the sup- posed phonetic law has yet been proposed. In some cases the variation is clearly Indo-European (see Brug- mann Grundr. I 2 . 633), as ao-refMprjs "firm," <7Tefx,(f>v\a "olives or grapes pressed dry," Skt. stambhate "become firm or rigid": "shake," OHG. stampfon, Eng. stamp; apos, i>e<£o?, etc.: opfipos, Skt. dmbu "water," Arm. amp "cloud;" Odpfios "astonishment," ON. dapr " sad " : rac^o? "astonishment;" icdpco "wither," Kap^ "hay," /ca/a<£o?, Lith. skrebiu "become dry": icpdp,fios "dry, loud," Ar., /cp6fi/3o<; "dry, roasted," Hesych., /cpopfioco "roast," Diphil., NHG. riimpfen. In fiefjuf>op,ai : Ir. mebul, Goth, bimampjan (see above, p. 212) Greek has preserved only the aspirated form of the root. We have a similar interchange of I-E. labio-velars in 6'<£i?, Skt. dhis "snake": OHG. unc "snake," 'lfifii]pi<;- eyxe\v<;. Mvdvp^ valot,, Hesych. Perhaps, however, we should omit the last word on account of its vocalism (Brugmann Grundr. I 2 . 634). <7T/)o'/t/8o5 "top" : <7Tpe(o, crrpofos, etc., was probably in- fluenced by pd/it/3o? "top." The tendency established by such pairs as aarefufrfc : o-refifico, cKfrpos : S(i/3pco, sometimes became effective even where it was not assisted by similarity of meaning with any of these. 2 Thus we find /u/9 instead of (f> in tcopvp,f3o<; "summit, hairpin," icopvfifir) "hairpin": Kopvtyr), Kopvo<; "summit, peak, hairpin" (see above, p. 204); fipepfioi = yS/>e'<^)09, Hesych.; «v/i/3o?, Kvp,/3n "cup" : Skt. kumbhds "cup," Av. xumba- "pot"; arep.0a) "cheat" : Skt. dabhnoti, ddbhati "injure, deceive." Perhaps we should set down here 1 Most recently Otto Hoffmann, Die Makedonen 240 i ., who gives references to earlier discussions. All the examples contain Greek labials except irtvdai; irv9n^v, for which Brugmann, Grundr. I 2 . 633, suggests Indo-European variation between d and dh. 2 For such imitative substitution, which is not to be denied a wide scope in language, and which many regard as a fundamental factor in the regular phonetic changes, compare Wheeler " Causes of Uniformity in Phonetic Change," Trans. Am. Phil. Assn. 32, 5 ft., especially p. 14, and Thurneysen Etymologie 17 ff., KZ. 44. 111. Studies in Greek Noun-Formation 215 Kv(fi)07] "head," E.M., Georg. Sanguinatius : Cretan Kvd "head," Hesych. \t,p,6<; • avKOeo-0ai "curdle," rap^w, etc., and Open/So? "fat," CGL. : rpe^to (cf. CP. V. 334), we have to account for the initial aspirate by the analogy of such forms as Opetyw, e6pey]ra, and riOpafipLai. The occasional interchange of bh and b in the parent speech and of and nfi in Greek furnish us two possible points of contact between the suffixes in and /3. Perhaps, then, it is not altogether accidental that both the suffixes are employed to form animal names and derogatory adjectives. See CP. V. 331, 333 ff., and Niedermann, IF. Anz 19. 32 f. The present writer, however, does not know of any word with Indo-European suffix -bo from -bho. We may, perhaps, suspect some such relationship between do-/ca\o<^o?, icaXa seems to have affected a few words in which was a formative element. The preceding discussion covers the words in which $ appears to be a formative element. The great majority of words in -i??, and $o? are forms with radical , which, with their numerous com- pounds, e.g., -ypao<;, -o