В. Г. Гузев. Избранное

345 Some Puzzling Aspects of the Turkic Runiform Script It might help to look for a solution to the puzzle in two separate direc- tions: V. Thomsen’s suggestion that, in view of the extraordinarily large num- ber of signs (thirty-eight), we are, in fact, having to do not with an ordinary alphabetical script (where about thirty letters would be normal) but with a syllabary; and everything that we know about the vocalic signs in OTRS. We might recall that V. Thomsen first discovered the following three vo- calic letters: O (/u/, /o/), i (/ï/, /i/), o (/ö/, /ü/); later, the letter a , (/a/, /ä/) was established (Thomsen 1984: 332–334). As regards the Yenisei-Tuba type of OTRS (which, it is generally conceded today, is no older than the Orkhon type), it had an additional vocalic sign: A for /ä/ and /é/. Furthermore, it was only vocalic signs that could be “both incised or omitted” (V. Thomsen), with the last one ( a ) occurring in the absolute final position. My first point is that the very presence of the “vocalic” signs, as well as their number (four + one), bears out the suggestion of the syllabic nature of OTRS; (it is vowels that form syllables). Secondly, the signs themselves, or their phonematic contents (o/u, ö/ü, ï/i, a/ä, /ä/, /e/), are evidence that they reflect, to some extent, the lingual features of vowels or syllables. Developing the hypothesis of the syllabic nature of OTRS further brings us to the conclusion that not only the vocalic signs, but also the twenty-two consonantic signs might have been syllabograms, at least for some time in the past. “Consonantic” signs, in that case, should perhaps be interpreted as having, at one time, represented twenty-two syllables of some Turkic type. The fact that the “vocalic” sign a/ä is absent in the absolute initial position and always occurs in the absolute final position brought O. Pritsak (1980:85– 86) to the correct conclusion that it must have been of the VC (vowel plus consonant) type 1 . In view of the great morphonological value of palatal vowel harmony, every syllabogram had, no doubt, to reflect the lingual features of vowels by means of the given syllable as a whole (presumably, originally there were no signs for vowels at all). We may, thus, regard the twenty-two “paired letters” as syllabograms with the following values: ab ( B ), äb ( b ), ad ( D ), äd ( d ), a y ( G ), äg ( g ), ay ( J ), äy ( j ), aq ( q ), äk ( k ), oq/uq ( [ ), ök/ük ( V ), al ( L ), äl ( l ), an ( N ), än ( n ), ar ( R ), är ( r ), as ( S ), äs ( s ), at ( T ) and ät ( t ). We must also add ïq ( Q ) and ič ( w ), two syllabograms which are not “paired”. 1 According to Pritsak (1980: 85–86), one of the fragments found in Toyoq (Turfan, 1905) and published by A. von Le Coq (1909) serves to support the conclusion that the values of nineteen of the Turkic runic can be arrived at by means of the Manichaean alphabet. Talât Tek’n reached the same conclusion quite independently; his views have given rise to objections on the part of Osman F. Sertkaya (1990: 13–15).