Eric P. Hamp	CDU 801.52 : 809.198.3-3
University of Chicago
ALBANIAN THIKE KNIFE'
In the volume of essays Languages and Areas: Studies presented to George V. Bobrinskoy (Chicago 1967) 66—9,1 struggled to find cognates and an Indo-European background for the obviously inherited Albanian thike, which is the ordinary word for a knife. I was exploring, without finding anything anything conclusive, the possibilities of Indo-Iranian. It seemed to me then, and it still does, that Indo-Iranian assures us of an etymon *kika 'sand, gravel'; I preferred then, in consideration of sita- = Lat. catus etc., to trace this to *ka-ka, a zero-grade of *ko- (= *k'eH0-), rather than to *ki-ka. I further discussed the problems and limitations of the Albanian vocalism and of a suffix in -k-.
Beside Skt. sana, Mod.Pers. san, Lat. cos 'whetstone' and Skt. sila 'crag', it seems best to regard *kika 'gravel as *kd-ka. But a much more direct solution for Alb. thike now seems possible.
It is understandable that Gustav Meyer and Norbert Jokl never considered comparing Vedic (RV) sikvon-, sikvas-, (AV) sikva since Grassmann glosses the first two 'stark, tiichtig'. However, Wackernagel (AiG II, 2, 903 §720b) glosses them 'schnitzend, beilfuhrend'. We may then extract the base sik- with the approximate meaning 'sharp blade'.
In discussing this question with me may late colleague J.A.B. van Buitenen further called my attention to sikha 'tuft of hair, crest, point, tip, border, knife-edge, arro-w'1.1 suggest that we have here a direct comparison with the Albanian. The aspirate of sikha may be analyzed like that of panthas, thisthati, ratha- (: Lat. rota); in fact, as in panthas, we may have here a duplicate representation of the Indo-European laryn-gel both in the aspirate consonant and in the long vowel. We may therefore reconstruct sikha as sik-Ha-. This then would be a laryngeal-suffixed form with a base matching that of sik-van- sik-va-.
In point of fact, this laryngeal-suffixed form may well have an exact correspondent in the feminine Alb. thike. The normal reconstruction for the latter would be *kika or *keika2. If we choose the latter it may be re-expressed as *keikeHa schematically. Thus sikha and thike differ only by ablaut. The indie word would be an old ablauting noun *kik-(e)H£Z; the Albanian would represent a guna stem (perhaps by the same rule as was productive in Baltic and Slavic) *keikeHa, corresponding to Skt.
1	Note too Pall sikhara 'peak, point or edge of a sword, crest', sikha 'point, edge, crest'.
2	For *k we may equally have *kw.
67
pantha, Av. panta. If we reason further, following the paradigm that jani/gna- once had, we arrive at a possible nominative *keik-Ha and an oblique (e.g. genitive) *kik-eHa-os. Such paradigms would immediately account for the ablaut in the base seen in *kika and keika, as well as the aspirate consonantism and the stem-class.
This leads us to a root *keik- or *keikw-, hitherto unrecognized for Indo-Europe-
an3.
Povzetek ALBANSKO THIKE "NOŽ"
Albansko thike "nož" se v prispevku primerja s sti. šikha "čop las, greben, konica, rob, noževo rezilo, puščica" in oba leksema izvaja iz ievr. sklanjatvenega vzorca leksema z laringalnim sufiksom z verjetnim imeno-valnikom ednine k'eik-Ha in s stranskosklonsko, npr. rodilniško obliko k'ik-eHa-os. Albanski leksem z možnima izhodiščima k'ika in k'eika lahko predstavlja polno stopnjo k'eik-eHo, staroindijski pa ničto *k'ik-(e)Ha- in z aspiriranim soglasnikom in dolgim soglasnikom tako kot sti. panthas kaže na dvojno zastopanost ievr. laringala. Primerjava odkriva novo ievr. osnovo k'eik- ali k'eik~\ ki se ohranja tudi v ved. šOcvan- šikvas-"stark, tiichtig" (Grassman), "schnitzend, beilfuhrend" (Wackernagel).
3
68
See my discussion of the descendants of this etymon, Lingua 34, 1974, 229 ff.