In recent times, a substantial body of material has been amassed in the field of experimental studies of Iranian languages. This accumulation enables us to draw new generalizations regarding the acoustic attributes of word stress as a part of the speech recognition problem. Numerous acoustic investigations have been conducted, including those carried out by Soviet and Russian scholars. However, the examination of Iranian languages has been very uneven: the majority of the acoustic studies were focused on Persian, and only a few addressed 11 other languages: Abyanei, Dari (Afghanistan), Gavruni (Zoroastrian Dari), Gilaki, Mazandarani, Ossetic, Pashto, Rushani, Sarikoli, Tajik, and Wakhi. Furthermore, most of the studies by Russian scholars were published in Russian and therefore remained unfamiliar to the broader international linguistic community, which led to “rediscoveries” similar to those previously described. The authors of the studies had diverse and sometimes controversial viewpoints concerning the nature of the word stress. Consequently, the primary aim of this study is to provide a comprehensive overview and analysis of the research on the topic in order to present a well-rounded portrayal of the acoustic properties of the stressed syllable in Iranian languages. An extensive examination of all known scientific literature on this subject written in Russian was undertaken. Various points of view, including those of foreign scholars, were compared. Those with weak points in reasoning were found and screened out, and the most cogent ones were adopted as the most probable characteristics of word stress in Iranian languages. Different patterns of stress were identified in Iranian languages. Tonal stress (based on pitch) was found in Mazandarani, Persian, and Tajik. Quantitative stress (based on duration) was discovered in Dari (Afghanistan) and Sarikoli and theoretically in Rushani. Multicomponent stress (with a mixture of different significant acoustic parameters) was found in Abyanei (quantitative-tonal), Gavruni (quantitative-dynamic-tonal), Gilaki (tonal-dynamic), Pashto (quantitative-dynamic-spectral (based on vowel quality)), and Wakhi (quantitative-tonal). The Ossetic is likely to have quantitative word stress, but statistical evidence of it is still lacking, necessitating further search for the exact stress correlate. Apparently, the overall trend emerges that tonal and quantitative stress are typical for numerous Iranian languages. Dynamic stress is found in several languages, albeit only as a part of multicomponent stress, while spectral stress is the rarest phenomenon. The results achieved can find application in automated transcription and speech recognition services.

Confidence Correlates of Word Stress and Prosody in Iranian Languages

工具书词条
发表日期：

2026年

查看全文

Vladimir Ivanov,
Lyubov Silanteva

摘要

In recent times, a substantial body of material has been amassed in the field of experimental studies of Iranian languages. This accumulation enables us to draw new generalizations regarding the acoustic attributes of word stress as a part of the speech recognition problem. Numerous acoustic investigations have been conducted, including those carried out by Soviet and Russian scholars. However, the examination of Iranian languages has been very uneven: the majority of the acoustic studies were focused on Persian, and only a few addressed 11 other languages: Abyanei, Dari (Afghanistan), Gavruni (Zoroastrian Dari), Gilaki, Mazandarani, Ossetic, Pashto, Rushani, Sarikoli, Tajik, and Wakhi. Furthermore, most of the studies by Russian scholars were published in Russian and therefore remained unfamiliar to the broader international linguistic community, which led to “rediscoveries” similar to those previously described. The authors of the studies had diverse and sometimes controversial viewpoints concerning the nature of the word stress. Consequently, the primary aim of this study is to provide a comprehensive overview and analysis of the research on the topic in order to present a well-rounded portrayal of the acoustic properties of the stressed syllable in Iranian languages. An extensive examination of all known scientific literature on this subject written in Russian was undertaken. Various points of view, including those of foreign scholars, were compared. Those with weak points in reasoning were found and screened out, and the most cogent ones were adopted as the most probable characteristics of word stress in Iranian languages. Different patterns of stress were identified in Iranian languages. Tonal stress (based on pitch) was found in Mazandarani, Persian, and Tajik. Quantitative stress (based on duration) was discovered in Dari (Afghanistan) and Sarikoli and theoretically in Rushani. Multicomponent stress (with a mixture of different significant acoustic parameters) was found in Abyanei (quantitative-tonal), Gavruni (quantitative-dynamic-tonal), Gilaki (tonal-dynamic), Pashto (quantitative-dynamic-spectral (based on vowel quality)), and Wakhi (quantitative-tonal). The Ossetic is likely to have quantitative word stress, but statistical evidence of it is still lacking, necessitating further search for the exact stress correlate. Apparently, the overall trend emerges that tonal and quantitative stress are typical for numerous Iranian languages. Dynamic stress is found in several languages, albeit only as a part of multicomponent stress, while spectral stress is the rarest phenomenon. The results achieved can find application in automated transcription and speech recognition services.