lO A. H. Edgren, 



Goth. ska-t/iaQ) ; — Skr. ] tan.: Gr. relv-at, Ti-raiv-co, rev-aiv, 

 ra-iG'i, re-ra-KU, e-rddrjv; Lat. ten-do, ten-eo, ten-ax; Goth. 

 tJian-jan ; — Skr. ■\/inan: Gr. [xev-co, /xe-fiov-a, fiev-o<i ; /mav-la; 

 /ji€-/jia-/jiev, fxe-fxd-Tco; Lat. inon-co ; Goth, ninn-uni, ga-inun-a7i ; 

 Skr. -^/van : d-fd-o), d-aa-To<i ; Lat. ven-ia, ven-ns, vcn-ustiis ; 

 Goth. V7in-an, vinn-an ; — Skr. ^/san: Gr. ev-w (Fick, L 226), 

 dvvw, uvco ; e-T09 ; Lat. sim't, sen-ex; Goth, sin-ista, sin-teino ; 

 and so on. This variation, however, nowise favors tlie theory 

 of independent -an and -a-xooX's>. The nasal is rarely lacking 

 except in Greek, and its absence there is frequently explain- 

 able on precisely the same grounds as in Sanskrit, viz., as 

 owing to an original weakening of the root : cf. tatd : ra-To^; ; 

 tail : rdcrt^ ; inatd : -fxaTO'^ ; satd : ero^, etc. The very fact 

 that cognate forms in other languages usually retain the nasal 

 where it is lost in Sanskrit or Greek, tends to show that it is 

 original, having disappeared in Sanskrit and Greek under 

 certain circumstances which in the cognate tongues have 

 produced different results : cf. Skr. inatd or mati : Gr. jxaro^ : 

 Lat. incnti-, Goth, ga-mnndi- ; Skr. menivid (for "^mc-nin-i-via: 

 n preserved by the following i) : Gr. fiefxafjiev : Goth, innnuvi, 

 and so on. Analogy has in all languages, as especially in 

 Greek, wrought many changes, and must account for some 

 of the seemingly irregular non-nasal radicals. Concerning 

 the Greek forms, compare especially Brugman's article already 

 referred to above. 



If, then, contrary to the opinion of Van den Gheyn and 

 others, it seems incontestable that the nasal of the roots ksan, 

 tan, man, van, and san is genuine and original, at least with 

 reference to the time when the inflectional system of the 

 Indo-European language was developed, does it necessarily 

 follow that in a stem like tano-, the suiBx must be -o and not 

 -no ? Not at all. The stem would have precisely the same 

 appearance were we to suppose either, with Bopp, that the n 

 of the original -//^'-suffix has been dropped, or that the root 

 itself had suffered mutilation before such a suffix. Let us 

 then examine whether there are any considerations that will 

 warrant us in accepting either of these theories, instead of 



26 



