198 Walter Petersen, 



104. Feminines in -ktxtj are sometimes formed as a counterpart 

 to a previously existing masculine in -ktv.o?, without being directly 

 traceable to a feminine name as primitive. So ^^//rrxt^ ^ (Plant. 

 Cist.) is probably patterned after the masculine 'A7icrxo? : "kXtoc, ; 

 MeridYJ] (Cos CB. 3706. i. 18, 3. 63, 7. 36) after AJEvtirxo? : JMsvTj? 

 or l\Ievo)V ; Eixj ()oriGy.t( (Lipara CB. 3552) after E-j^povicr/.o? : E5opo)v; 

 possibly also Aafxia/.Tq after Aa[jicy.oc instead of : % Aaij.ia, Bola/M 

 after Boit/vO? instead of : r, (jou?, etc. 



105. As a successor to I. E. -ko- -it/.o- could probably take upon 

 itself the function of forming " Kosenamen," i. e. shortened forms 

 of compound proper names of which only one member, usually 

 the first, was used in full, whereas the other was suppressed en- 

 tirely or nearly so. Even if this function was not due to syn- 

 cretism with prehistoric -ko- or -I0-, it could easily develop in the 

 Greek language itself by a shift of grouping of etymologically 

 related names. Thus Aaij.t(r/.o?, which was a diminutive of Aaij.t? 

 or Aa[xia?, could be referred to the compound Aap.o^^paTTjC or Aa- 

 \}.6-KpiToq or Aa[xoxX^c, particularly since Aajj.ic and Aajjiac were them- 

 selves " Kosenamen " of these same compounds. Similarly yVaij.7:pi(7y.oc 

 might be "Kosename" of Aaij.7ip6[j-a/o? or 7Va[j.xpox>.Tj? instead of be- 

 ing a diminutive of Aa[x>:po$, JMevioy.o? e. g. of Msvstevo? -'Xao? instead 

 of : MsvTjC Msvfov, IIoBi'Txoc of IloBavsij.o? IIoBapxiqc instead of : IIo'SYic, 

 $aviaxo$ of <I>avi7C7ro? instead of : tI>avo? -la?. Also a number of words 

 listed above as coming from appellatives mighl have been formed or 

 secondarily felt as" Kosenamen," e. g. 'AvSpCoxoc ^ of 'AvSpoxpainr]? etc. 

 instead of being 'little man,' Apo[xi(7xo? of Apo[j.oxXrj? -xIsiSy]? in- 

 stead of 'little runner,' Auxioxo? of AuxoEipYoc etc. instead of 

 ' Wolfy,' M7]vtoxO(; of M7]v6Boto? -Bcopo? instead of : [j.rjVY] ' moon,' 

 IlaTpioxo? of llaTpoxArj? etc. instead of being 'little father,' $1X1- 

 axoQ of ^ilnzTzoi; <I>i};oxpa-rj5 instead of ' little friena.' Since a 

 number of names from similar or identical stems usually existed 

 alongside of each other we can not, however, often tell whether 

 an individual word was a " Kosename " or not. But that -icrxo- was 

 at least occasionally used in their formation is shown by several 

 words in which the addition of the suffix is accompanied by such 

 mutilation of the stem of the primitives that it can not possibly 

 be explained by the rules of sandhi (cf. § 12). These are Ilav- 

 •zcckicixoi; : IlavxaTveojv and rTappovvOc : llap[jiv(ov (§ 103 D). Di- 

 rectly formed from a compound stem is also the Boeotian ^Iqavi- 



1 For the inorganic h of the Latin form Halisca cf. Stolz, Lat. Gram.* 128; 

 Sommer, Handb. d. lat. Laut- u. Formenl. 217. 



^ This word e. g. is considered a " Kosename " by Fick-Bechtel. 



