В. Г. Гузев. Избранное

344 В. Г. Гузев. Избранное: К 80-летию some letters”, the synharmonic principle on which this alphabet was based was decisive evidence of this. The Old Turkic consonantal dualism is of great interest from a phono- logical point of view. In the Turkic languages, lingual distinctions are, as a rule, specific to vowels, not to consonants. It is the vowel of the initial syllable (quite often occurring as the first sound of the word) that deter- mines the synharmonic row of all units (phones) of the given word or word form. According to N. S. Trubetzkoy (1939:251), “Da der Konsonant j keine palatalisierten bzw. velarisierten Spielarten aufweist und viele Wörter nur aus Vokalen und j bestehen (aj “Mond”, aju “Bär” usw.), so können die Vokalphoneme auch unabhängig von der konsonantischen Umgebung eine bestimmte Eigentoneigenschaft aufweisen, während die Konsonanten nur in Verbindung mit Vokalen palatalisiert bzw. velarisiert sind. [...] Daher sind die Eigentongegensätze bei den Vokalen phonematisch, wobei die palatalisierten und velarisierten Spielarten der Konsonanten nur kombinatorische Varianten [...] darstellen.” At the same time, Trubetzkoy (1933: 121–122) was the first to sug- gest that a native speaker will seek to reflect just the phonemes in the script he uses, being, as a rule, unconscious of the phoneme variants, i. e. the units of the subphonemic level. This point is in keeping with treating the phoneme as a “sound representation”, “das psychische Äquivalent des Sprachlautes“ (Boduén de Kurtene 1963. I:352), i. e., an abstraction ac- tualized in individual language usage as a concrete class of functionally equivalent minimum sound segments (phones) occurring in speech (see Kasevič 1983: 33–67). It appears from the above that from a phonological point of view, the phe- nomenon of “consonantal dualism” is a contradiction on two obvious counts. 1. If the inventors of the script were Turks, their representation of palatal vowel harmony would have been based on the qualities of the vowels in their own language. The inscriptions, however, have a typical consonantal hypergraphy (the notation of subphonological units, i. e. allophones of the consonants) and, as a consequence, a bulky allography (the use of two or more alternate signs to render one phoneme) (Shorto 1965), although the palatalization/velarization distinction is phonologically irrelevant in the case of Turkic consonants. 2. If we assume that the script was produced by non-Turks (because only a foreigner would have distinguished between the velarized and palatalized allophones of a Turkic consonant), we still have the surprising use of two signs for the phoneme /y/, which (as is clear from the remark of N. S. Tru- betzkoy quoted above) cannot possess the allophones in question.