152 M^a/ler Petersen, 



ex})lain. The existence of all of these forms in -fr/.ov should make 

 us dubious about invariably counting as masculines those forms 

 from neuter primitives of which the gender can not be determined. 

 Thus the ace. sing. Xsxiotcov (: to >ixo?) in the Ionian Hippocrates 

 and the gen. pi. (7aij.[ja>.t(7>:ojv (: crajj-piaXov) in Hcrondas are certainly 

 neuter rather than masculine, probably also the dual (^uyi'Txc.) in an 

 inscription from Eleusis,^ though the fact that there are no certain 

 examples of the neuter in Attic would make any conclusion doubt- 

 ful, as are also the following forms in literature : YajXfj.aTtTxotc (: yajj.- 

 [xa) Jo. Laur. De Magistr. Reip. Rom. 2. 4, XTspiT/.otr (: zTspov) 

 Babr. 118. 5, (jxsXictxoiv (: to oy.s}^oc) Ar. Eccl. 1167. 



19. The residue of certain or probable masculine derivatives from 

 feminine or neuter primitives is exceedingly small, and no larger 

 proportionately than aberrations of gender in other directions. All 

 of these changes, moreover, can be explained as caused by the 

 same forces and will be treated to-gether. 



20. When it was stated that the gender of words in -itxo- followed 

 that of the primitive several classes of exceptions were ignored. 

 In the first place it is evident that names of persons must follow 

 the natural gender regardless of the gender of their primitives. 

 While TO [xsipaxiov could itself be neuter because that was the 

 regular gender for substantivized adjectives of all kinds, and there 

 was no corresponding masculine form to be had, yet from this 

 word could come only [xstpaxioTto?, not *[;.£ipaxi!Gxov, since a form in 

 -icrxo- grammatically characterized as masculine was available. This 

 applies particularly to proper nanies,- the gender of which depends 

 altogether on the sex of the person to whom they are applied. 

 Thus from 6 >.'Jxoi; comes indeed Auxtcnto? when a man is desig- 

 nated, but Auxioxa for a woman, from 6 aypoi; comes 'Aypicyvdc, an 

 epithet of the goddess Athena, from the neuter to [jiXoc Mslioy^oc, 

 from TO {xupov Mupi'Txo?. 



21. Moreover, when the derivative is not a living 'diminutive,' 

 being cither faded or with a suffix expressing hkeness or perhaps 

 appurtenance from the beginning, the derivative would not be in 

 as close touch with its primitive as with its congeneric words, and 

 would consequently be disposed to follow the gender of the latter. 

 Thus xomoTio? ' a pastille ' followed the gender of the congeneric 

 TpoyicTAOC and xuxAicrxo? rather than of its primitive yj v-OTic ' chopper,' 

 and xavva(3t(7xov ' a shoe made of hemp ' followed (Ta[j.pa}i<7xov (: to 

 <7a[j.[3aXov) rather than its ])rimitive y] xavva[3ic. Congeneric attrac- 



1 Cf. the editor's note Arch. Eph. 1899 p. 184. 

 - Cf. Janson, op. cit. 68. 



