Íslenskt mál og almenn málfræði - 01.01.1985, Síða 100

98
Jörundur Hilmarsson
179) that the formations in *-tro-/*-tero-/*-toro- are frequently used
pronominally oradverbially in Germanic.
Brugmann (1905/06:424) points out that Gmc. *sunþra- (he posits
*sunþera-) probably joined the group of direction words only in Ger-
manic times, as there is no etymological evidence for such a formation
in other Indo-European languages. Indeed, the other three words in the
series, *nurþra-, *austra-, *westra-, are arguably of Pre-Germanic
origin.
Thus *-þra- is probably original in *nurþra-. This Germanic word is
usually compared with Gk. (é)vépxepoq ‘deeper, lower’, Umbr. nertru
‘sinistro’, Osc. nertrak ‘a sinistra’, which seems a likely connection in
spite of the difference in root vocalism. Perhaps the zero grade of
*nurþra- and *nurþa- (for expected *nerþra- and *nerþa-) might be due
to the analogical influence of their antonyms, *sunþra- and *sunþa-,
whose formation, vocalism and accentuation will be discussed below.
Gmc. *austra- has an exact match in Lat. auster ‘southwind’ < I.-E.
*austro-, indicating an original *-tro- in this word also.
If Gmc. *westra- is correctly associated with Skt. avás-‘downward’,
as generally assumed, a formation in *-tro- might be original in this
word too, for adverbs of space frequently produce derivatives in *-tro-
/*-tero-, e.g. Goth. wiþra, Olcel. viðr, OHG. widar ‘against’, cf. Skt. v/
‘apart’: vitarám ‘further’ (Krahe/Meid 1967:180).
On the other hand, since there is no etymological reason to assume
that Gmc. *sunþra- contains an original *-tro-, it would seem permissi-
ble, with Brugmann (loc.cit.) and de Vries (1962:559) to consider this
word as analogically formed on the pattern of the other direction words.
None of the words in the other series of direction words, the nominal
*nurþa-, *sunþa-, *austa-, *westa-, has an exact etymological cor-
respondence in other languages. However, word-formational justifica-
tion foran I.-E. *-to- suffix might be found in the case oi*sunþa- which
would then have served as a pattern for the remaining words.
One function of the I.-E. suffix *-to- was to form denominal adjec-
tives with the sense ‘being fumished with or having the properties of
whatever the basic word expressed, cf. Krahe/Meid (1967:141). Thus
Hoffmann (1975:336) has shown that Skt. párvata- ‘mountain, rock’
was still used as an adjective in the Rigveda. He analyses this word as
*perun-to- ‘rocky’, an adjective in *-to- to the heteroclitic
*pér-ur/*pér-un- ‘rock’ continued in Hittite and Indian.