Аннотация:Speech research presumes: speech production and perception; speech analysis and synthesis; speech recognition; speech pathology; speech processing; laboratory phonetics and phonolo-gy. This kind of knowledge can be applied to every language. Relatively good level of speech science in Russia caused investigations in Iranian languages. In last decades this kind of study became very popular in Iran too. 95% of works are dedicated to Persian, but there are some remarkable achievements in other languages as well.
The instrumental study of Pashto speech has begun in 1969. At that time most of the Soviet speech scientists were under the influence of the Leningrader Phonologic School. They de-veloped a theory that all Indo-European languages have dynamic stress, i.e. the stressed syl-lable is marked by higher intensity. In the West phoneticians were influenced by the works of D. Bolinger who in early 50s found the pitch accent in English. In other words, the stressed syllable in English is marked by higher tone (pitch). After that Western phoneticians found pitch accent in other languages, while in the Soviet Union the linguists were discover-ing dynamic stress. Pashto was no exception in that respect. Mrs. Sharifa Asmati studied Pashto stress at our Institute and after some experiments came to the conclusion that the word stress was dynamic.
In 2000–2001 I have led a set of experiments and saw no proof of dynamic stress in Pashto. My study showed that the word stress in Pashto is quantative, i.e. the stressed syllable is marked by duration. In 2009-2012 Ms. Tarbeeva N. made her experimental studies of Pashto stress and rhythmics. She also found the importance of duration for Pashto stress, but intensi-ty was important too. So in her opinion Pashto stress was quantative-dynamic. She also mod-elled the rhythmics of Pashto speech by synthesis by rule.
In 2006 18 researchers from USC Viterbi School of Engineering (LA) made a multilingual Speech to Speech Translation System. Persian and Pashto Automatic speech recognition en-gine was a part of that project. Another Text-to-Speech project that included Dari and Pash-to, was fulfilled in 2014 by Michael H. Lee.
The first linguist to study Ossetic experimentally in 1948–1950 was Dr. Sokolova V. Her book was published in 1953. She came to the conclusion that in the more archaic Digor vo-calism there is an opposition of long ↔ short vowels, while in younger Iron system strong vowels are opposed to the weak ones. She also showed a collection of kymograms of Ossetic consonants.
Another problem is characteristic for the Soviet-Russian Iranian studies and is not observed in Western phonetic research. It is the question of so called double-focused fricatives. P. Ladefoged and I. Maddieson called them “multiply-articulated fricatives” and consider them to be unlikely to exist. But V. Sokolova was one the first iranianists to introduce the double-focused fricatives into the speech research. Every Iranian language, she describes, contains at least 2 double-focused fricatives – [š], [ž]. She uses her description of Iron double-focused fricatives to show the difference in the articulation regarding their previous state described by A. Sjögren and Vs. Miller 70 years before. In our opinion there are no double-focused fricatives in Iranian languages at all (in Ossetic as well). That was proved by radiography at least for Persian and Tajik. Speech researcn in Ossetia continued and in 2009-2016 Dzakhova V., Parsieva L., Gatsalova L. and Andieva M. published a series of articles on Os-setic vowels, consonants and intonation.
Tradition of describing Pamirian languages has started about a century ago. Many books have been written by Pamirian native speakers like M. Fayzov and Sh. Yusufbekov. Still most phonetical references are made to (Sokolova V. Essays on phonetics of Iranian lan-guages, v. II, Ossetic, Yaghnobi, Pamirian languages, 1953), where she began to study the vowels’ duration. She found that word stress influenced the vowels’ length.
Our studies of Sarikoli word stress showed that it is quantative (p=0,033). It caused the lengthening of the vowel by 20%.
First instrumental analysis was done on Ishkashimi vowels and consonants by T. Pakhalina in 1959. She published some palatograms of the Ishkashimi consonants. But the most important experimental study in this field was the dissertation of S. Sheshenin “Ishkashimi consonants” (2011). He made a very thorough investigation of spectral properties of the consonants and described their acoustic features.
In 1953 Sokolova V. studied vowels’ duration in Shughni. She proved that long vowels op-posed the short ones. Short vowels were approximately 2 times shorter than the long ones in identical phonetic positions.
In the same book V. Sokolova shows us the duration of Vakhi vowels in various phonetic po-sitions. She also demonstrates the palatograms of Vakhi consonants. In 1975 T. Pakhalina continued her study on vowels’ duration. She used another classification of vowels and their phonetical positions. But the main result was very similar: long vowels are opposed by the short ones. Our studies of Vakhi (including Pamirian, Pakistani and Chinese variants) in 2012 showed that its word stress is multicomponent: duration, pitch and intensity turned out to be very significant to mark the stressed syllable.
In 2017 3 Iraqi authors A. Al-Talabani, Z. Abdul, A. Ameen have developed an automatic Kurdish dialect recognition system using one-dimensional Local Binary Patterns (LBP) fea-ture. The acquired data involved in this study were 3 Kurdish dialects (Sorani, Badini and Hawrami) with 3 neighboring languages (Arabic, Persian, Turkish). They proposed a new method to interpret the closeness of the Kurdish dialects and their neighboring languages us-ing confusion matrix and a non-metric multi-dimensional visualization technique. Now they can cluster the Kurdish dialects and separate them from the neighboring languages.
In 2008 a Russian forensic system that can identify Talysh accent in Russian speech has been built. In 1997 V. Yefimov published Parachi vowels’ formants.
In 1974 Pierre Lecoq stated that the word stress in Ābyānei language was dynamic (marked by intensity; accent d’intensité). In 2009 V. Ivanov and L. Dodykhudoeva came to the con-clusion that it is quantative and intensity seems to play no significant role in marking the stressed syllable. Our research on Gavruni allowed to correct the vowel system chart. In Kerman vernacular word stress is quantative (marked by duration). In Yazd word stress is multi-componential (quantative-dynamic-tonal). In Mazandarani word stress was described as dynamic, while we found that it is tonal. In Gilaki previously word stress was described as purely dynamic. My analysis showed that it is 2-componential: tonal-dynamic. In Tajik and Persian word stress is tonal, in Dari – quantative.
Speech research on different Iranian languages is quite uneven: most of the studies were done in the field of Persian (ca 100 authors). Some of the works include other 19 languages: Abyanei, Baluchi, Dari, Gavruni, Gilaki, Ishkashimi, Kurdish, Mazandarani, Ossetic, Parachi, Pashto, Rushani, Sarikoli, Shughni, Tajik, Talysh, Tat, Tati, Vakhi. But most of the Iranian languages and dialects are still to be studied. Primarily we must verify their vocal-ism, consonantism, word stress and main types of intonation (narrative, interrogative, imperative).