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

14 LÁSZLÓ DEZSŐ In Indo-European there was only one sibilant s with palatal, labiovelar coloring: s, s° and neutral s, a characteristic shared by velar stops and laryngeals (cf. Gamkrelidze - Ivanov 1984: 116-134). In Proto-Uralic the sibilants were represented by two phonemes s and s with or without palatal feature. They are preserved in Finno-Ugric; in Samoyedic s is usually replaced by s, s by t. The two approaches to PU and PFU differ in the reconstruction of affricates: Rédei and Hajdú postulate a neutral с and a palatal с for Proto-Uralic; Sammalahti admits them for Finno-Ugric but for Uralic reconstructs only a cacuminal с (UEW: IX, Hajdú 1993: 164-166, Sammalahti 1988: 482-94). The phoneme s figures in both Uralic and Finno-Ugric according to the first approach, and only in Finno-Ugric according to the second. In the Altaic proto-languages there was an affricate с with its voiced phonemic pair dz, and a sibilant s without voiced z in both initial and medial positions with possible modifications (Poppe 1960: 25-31,62-65). There were laryngeals in Indo-European, Gamkrelidze and Ivanov (1984: 195-214) reconstructs three of them: one with palatal, another with labio-velar coloring and a neutral one. For Uralic Sammalahti (1988: 482) postulates a sound, denoted by the symbol x, whose phonetic nature is not clear: on the basis of Ob-Ugric evidence „it is possible that /x/ in fact represents two different sounds, a laryngeal (?/h/, ?/?/) yielding a long vowel in Proto-Ob-Ugric, and a velar (/g/ or /g/) giving the spirant /g/. It is possible that /x/ is merely syllable boundary reconstructed in cases which originally had two successive heterosyllabic vowels: the phonotactic peculiarities of /x/ call for thorough examination". Thus, the existence of a laryngeal in Uralic remains an open question. Greenberg (1995 section 2) notes: „In Chapter 2 I suggested, as a working hypothesis, that the Indo-European *H ultimately derives from one or more velar stops in many, though perhaps not in all, instances" . The Uralic /x/ or /Y/ is marginal in Uralic, its probable origin is к from which they can derive through fricativization and voicing. The change of £ to a laryngeal in Uralic was marginal and к has preserved its very high frequency (cf below and Dezső forthcoming). 2.3. Obstruents and Résonants in Root Structures In the two-consonant roots which are typical for both PIE and PU and the only possible or dominant ones in PrelE and PreU, the first consonant position is dominated by non-resonants, the second by résonants. Gamkrelidze and Ivanov (1984: 216) observe the statistical limitation on the use of résonants in initial position which implies higher frequency of obstruents and laryngeals. The obstruents dominate the Ci position and the résonants the C2 also in Kartvelian (Gamkrelidze - Macavariani 1982: 122-3, Boeder's Nachwort). In Uralic and Finno-Ugric we find a similar situation; Fancsaly's data of the two protolanguages (1988: 42^14) are summed up in order to have statistically a more