شواهد آکوستیکی و بررسی میزان تقطیع‌پذیری واژه‌های مشتق پیشوندی زبان فارسی در پردازش گفتار

اطلاعات آکوستیکی می‏تواند به‌منزلۀ شاهدی برای بسیاری از مسائل مطرح در حوزه‌های دیگر زبان‌شناسی به‌خصوص صرف گفتاری و پردازش گفتاری باشد. مقالۀ حاضر با استفاده از واژه‏های مشتق پیشوندی زبان فارسی به بررسی شواهد آکوستیکی دال‏بر میزان تقطیع‏پذیری آن‌ها در پردازش، درک و دریافت معنی پرداخته است. برای این منظور واژه‏های مشتق از پیشوندهای «نا-»، «هم-»، «فرو-» و پایه‏های آن‌ها همراه با بسامد رخداد نمونۀ آن‌ها از پیکرۀ پژوهشگاه علوم انسانی استخراج شد. از هر پیشوند دو واژۀ تقطیع‏پذیر و تقطیع‏ناپذیر انتخاب و با قرار دادن آن‌ها در جملات حامل هم‌بسته‏های دیرش، زیروبمی، سازه‏ها، شدت، دیرش خیز و شفافیت منطقۀ گذر بررسی شد. یافته‏ها نشان داد دیرش نسبی واژه‏های تقطیع‏پذیر بیشتر از واژه‏های تقطیع‏ناپذیر است. در منطقۀ گذر وند و پایه و در حالت ایستایی وند واژه‏های تقطیع‏ناپذیر به‌علت پیوستگی میزان هم‏تولیدی در بسامد زیروبمی، سازۀ اول، سازۀ دوم و سازۀ سوم آن‌ها بیشتر از واژه‏های تقطیع‏پذیر است. در واژه‏های تقطیع‏ناپذیر، دیرش نسبی شدت خیز بیشتر و روند هموارتری را نشان می‏دهد. همچنین در واژه‏های تقطیع‏پذیر، شفافیت تولید سازه‏ها در منطقۀ گذر بیشتر از واژه‏های تقطیع‏ناپذیر است. 

The Survey of Acoustic Evidences and the Degree of Parsability in Persian Prefix’s Derived Words in Speech Processing

The acoustic information can be evidence for many considered matters in other linguistics areas, especially in spoken morphology and speech processing. The present study investigates the Persian derivational words of prefixes and the degree of parsability in processing, understanding, and receiving the meaning of derived words by some acoustic evidence. For this reason, the derived words of the prefixes «nɑ», «ham», and «farɑ» and their bases were extracted from the Institute of Humanities.  We chose two parsable and non-parsable derived words of each derivational prefix and put them in carrier sentences. Then, We investigate the acoustic features such as duration, pitch, formants, intensity, duration of maximum intensity, and the clarity of transition point The results show that the relative duration of parsable words is more than non-parsable words. The degree of co-articulation in pitch frequency, first, second, and third formant frequency of non-parsable words in transition points of affix and base and statically position of affix is more than parsable words because of incorporation and without potential pauses. In non-parsable words, the relative maximum duration of intensity is more with the smooth trend. So, the clarity of formants in parsable words is more than in non-parsable words at the transition points.1. IntroductionAccording to the presence of the structure of derived words in spoken language, the mental lexical processing can be assessed by its phonetic expression and the variations of morpho-phonetic structures can be surveyed in derived words. According to Plag (2014), we can use morpho-phonetic variations for morphological theories and lexical processing. Hay and Bayeen (2003) through the investigation of phonotactics in parsable and non-parsable derived words in the Complexity-Based Ordering approach considered that in non-parsable words the boundary of affix and base is not discerned very easily due to the affix and base are in phonological Entanglement and during the morphological processing, it’s considered as a whole word because the frequency of the derived word is more than the frequency of its base. This approach by using spoken morphology criteria provides some proofs based on the parsability of the base from the affix. In this regard Hay (2007), Plag and et. (2017) by using the acoustic feature of duration investigated the co-articulation of parsable words. They believed that the co-articulation in non-parsable words is more than the parsable words because in non-parsable words the effect of the adjacent phoneme in the transition point is more.2. Research HypothesisThis study tries to answer the question of whether in Persian derived words the acoustic features of duration, pitch, intensity, and clarity of the transition point of affix and base can explain the degree of parsability. The results of this research can be used in the theories of the mental lexicon, speech producing, understanding, and perception of the lexical meaning. 3. Research MethodIn the present study, we chose three prefixes of «nɑ-», «ham-», and «farɑ-» from Hemasian (2019). We selected two derived words of each prefix in the way that the non-parsable derived word with high token frequency is placed below the trend line of its scatter plot and the parsable derived word with high token frequency is placed above the trend line of its scatter plot. The chosen words are Hamdasti and Hamdardi, Namardi and Namaree, Foruraftan, and foruravande. For creating the same situation of speech we put the derived words in the same carrier sentences and the same position. 4. ResultIn this study, we investigate some acoustic evidence in the degree of parsability for Persian-derived words. For determining whether the duration feature can have any effect on the perception of the degree of parsability of affix from the base, it was found that affix and base duration and relative duration in non-parsable derived words are less than parsable words. Also, the duration differentiation of the last phoneme of the prefix and the first phoneme of the base in non-parsable words are more than in parsable words. These results are in accordance with Hay (2007). Then our hypothesis based on the duration of Persian derived words can be used as a piece of evidence for the degree of parsability is confirmed. In responding to this question that the pitch and formants can be described the morphological structure of derived words according to the transition point in the attachment slot of the affix to the base and the static position of the affix, it showed that the degree of co-articulation in non-parsable words based on the continuity and no potential pause is more. For answering the effect of intensity on the morphological structure of derived words according to findings it determined that in non-parsable words raising duration is more and intensity is very smooth to reach maximize. Then our hypothesis based on the relative raising intensity in non-parsable words is confirmed. In responding to the degree of derived words parsability and the relation with clarity of transition point and with concerning the formants by mapping spectrum it found that the clarity in a transition point in parsable words is more than in non-parsable words. Also, in the transition point, we can see the beginning of the next phoneme very easily. Ultimately, it was found that we can determine the degree of parsability in prefixed derived words by acoustic evidence.