Nyelvtudományi Közlemények 96. kötet (1998-1999)

Tanulmányok - Dezső László: Typological Comparison of Root Structuring in Uralic and Early Indo-European. [Az uráli és korai indoeurópai tőstruktúrák tipológiai összevetése] 3

Typological Comparison of Root Structuring in Uralic and Early Indo-European 29 nology where the etymons are in the foreground. In the pre-language the major variant of a phonological word was a root with a possible, even frequent, formative element and clitics, most probably enclitics, attached to the noun stem and person suffixes added to verbal bases. Both the clitics and the suffixes related the word to the rest of the sentence. The noun base functioning as subject or object also had suffixes denoting active, animate persons or inactive objects, beings. In my view, it is possible to hypothesize that they were of pronominal origin and were clitics at the beginning. The verbal person suffixes could also have been clitics in an early stage like in Uralic. The fact that certain suffixes did not participate in apophony can be explained by their independence from the word: when the apophony was being completed and extended to the whole of the base, they were still clitics. 3.2.4. On the Stage Preceding Pre-Indo-European Greenberg (1990: 119-121) assumes that the Indo-European vowel system emerging from Northern Eurasian could be presented by the opposition of high and low vowels reflecting level harmony: High i e и Low e a<0] 02 where 02 is a stable о not alternating with e. Greenberg's article was devoted to the presentation of the different vowel systems in Northern Eurasian protolanguages and did not deal with the evolution of early Indo-European to Pre-Indo-European examined by me. I consider Greenberg's systems as a typological hypothesis for one of the stages preceding the reconstructed forms of Indo-European and Uralic. Greenberg's article did not deal with accentuation. The other Western Eurasian languages (Uralic, Yukaghir and the three languages of the Altaic group) had fixed accent on the first root syllable. Such accent can be hypothesized only for the earliest period of Pre-Indo-European, if it is acceptable at all. From the point of view of Pre-Indo-European, free accent can be reconstructed with certainty. It could be on the root vowel according to the testimony of some lexemes (like the primary verbs of VC structure), but also beyond the root; the CC-eC apophonic variants shows a clear tendency to „postroot" accentuation. The dynamic accent was later replaced by melodic accentuation in Proto-Indo-European. The early free accent must not lead to the reduction of unaccented syllables, to their „obscuring", but later it certainly did. Similar „obscuring": the loss of quality of unaccented vowels took place in Uralic; it caused changes in the quantity of vowels in the Altaic group: Mongolian and Manchu-Tungus; in Turkic, however, the unaccented vowel was lost. The whole process was ac-