دانشگاه گیلان زبان فارسی و گویش‌های ایرانی 2476-6585 10 2 2026 02 20 From Fā'-e 'Ajamī to Arabic 'Ḍād': Complexities of Phonemic Splits in New Persian از فاء اعجمی تا ضاد عربی: پیچیدگی‌های انواع انشقاق واجی در فارسی‌نو 7 31 9520 10.22124/plid.2026.32792.1745 FA مهرداد نغزگوی کهن استاد گروه زبان‌شناسی دانشگاه بوعلی سینا، همدان، ایران. Journal Article 2026 01 09 This study examines various types of phonemic splits within the consonant system of New Persian. Three types of primary splits are identifiable in the New Persian period. The first case involves an allophone of /b/, which traditional grammarians called <em>Fā'-e 'Ajamī</em> (the non-Arabic Fā'). This was the voiced labiodental fricative /β/, which later became obsolete by merging into the pre-existing phonemes /b/, /v/, and /f/. This research indicates that this split likely did not occur uniformly in Northeastern and Southern varieties of Early New Persian. The other two cases of primary consonant splits in New Persian relate to allophones of /d/. In this context, the split of the dental fricative allophone /δ/ / ɾ̪ / is relevant. This allophone, represented by <ذ> in Early New Persian, was later lost, with most of its instances pronounced as /d/, while some merged into the /z/ phoneme. The third primary split involves the separation of the voiced allophone /ḍ/ from the phoneme /d/ and its subsequent merger into the Persian phoneme /t/. In this study, we demonstrate that this latter change occurred through a process we have termed an <em>intermediary split</em>. An intermediary split refers to the change during the early Islamic centuries where the loan phoneme /dˤ/ <ض> from Arabic split into four phonemes /d/, /dz/, /l/, and /t/, thereby disappearing itself. Since there was no pre-existing allophone of /dˤ/ involved, this process cannot be classified as a secondary split, nor does it fit the classical definition of a primary split. This paper shows that the native allophone [δ] /[ ɾ̪ ] was significant in determining the ultimate pathway of change for the loan phoneme /dˤ/ <ض>. The merger of some instances of the native allophone /δ/ / ɾ̪ / into the Persian /z/ phoneme paved the way for the substitution of those instances of /dˤ/ that were pronounced as [δ] in Persian, leading to their merger into /z/ as well. Through analogy, the remaining sounds whose ultimate origin was Arabic /dˤ/ also merged into the Persian /z/ phoneme. Following the loss of δ / ɾ̪ due to the primary split around the 6th century AH (12th century CE), all new Arabic loanwords containing /dˤ/ <ض> (by analogy with earlier loans) were directly accommodated into Persian as /z/. A similar, though less complex, change occurred for the loan phoneme /ðˤ/ <ظ>.<br /> <br /><strong>Extended abstract</strong><br /><strong>1.Introduction</strong><br />Phonological component of a language is susceptible to change, just like its other components; However, this change typically occurs at a slower rate compared to other linguistic changes, such as lexical, semantic, morphological, or syntactic changes. In phonological changes the number or distribution of phonemes is altered. Phonemic split is considered as one of the chief kinds of change in the phonological system of a language. The origin of splits may be internal, emerging from within the linguistic system, or external, influenced by factors such as language contact, or borrowing. Sometimes, both internal and external factors may contribute to phonological changes. This study examines various types of phonemic splits within the consonant system of New Persian by taking into account both these factors. It should also be noted, however, that the primary cause of such changes is not always readily identifiable.<br /><strong>2.Theoretical framework</strong><br />In historical linguistics, primary and secondary splits are traditionally distinguished. In a secondary split, the allophones of a phoneme become two or more distinct phonemes. This typically occurs as a result of the loss of the conditioning environment of each allophone, leading to the unpredictability of the presence of these segments and, consequently, an increase in the number of consonant phonemes. In contrast, a primary split refers to a process in which some allophones of a single phoneme merge with existing phonemes of the same language. A primary split can also be viewed as a split followed by a merger.<br /><strong>3.Research Methodology</strong><br />This study employs a qualitative, descriptive-analytical methodology. Our data is primarily based on Early New Persian. Texts of Early New Persian survives in two distinct varieties: Northeastern and Southern varieties. However, Southern Persian texts which are very limited in number, had closer relations with Middle Persian. This variety played a marginal role in history of Persian language and it did not continue to be used in the same way as Northeastern Persian did. This latter variety gained the status of the common language of Iranians and developed into the Modern Persian of Today.<br /><strong>4.Discussion and Results</strong><br />The linguistic distinction between Northeastern and Southern Persian is also observable even in the phonemic splits of /b/ within both varieties. In Southern variety, the phoneme /b/ had two allophones: [b] and [w]. Later, [w] evolved into [ß]. During the subsequent developments, the /b/ allophones merged and mostly resulted in [b]. However, in some cases, the allophone [β] split from /b/ and merged with the pre-existing phonemes of /f/ and /v/. However, it should be kept in mind that the Northeastern Persian was not a uniform language throughout Iran. As far as the primary split of /b/ is concerned, we do not have strong evidence that the development of [ß] in Northeastern variety followed the same path as in Southern variety. More precisely, the intermediate stage of the transformation of [w] into [ß] did not occur in this variety. However, similar to Southern Persian, in Northeastern Persian, some cases of the [β] allophone split from /b/ and merged with /f/ and /v/.<br />Another primary split attested in Early New Persian involves the allophones of /d/, namely, [δ]/ /ɾ̪/ and [ḍ]. In some cases, the allophone [δ] split from /d/ and merged with the pre-existing phoneme of /z/. In other cases, this allophone reverted back to /d/ and no trace of [δ] remains in the Modern Persian.  In the same vein, some instances of the devoiced allophone [ḍ] merged with the phoneme /t/. This primary split was a temporary phenomenon and attested only in some specific areas of Iran. However, this latter case had some bearing for the last phonemic split we are concerned with. This involves the accommodation of with Arabic where Arabic /dˤ/ <ض>, and ðˤ<ظ> in Early New Persian. The Arabic /dˤ/ merged with four different phonemes of /d/, /dz/, /l/ and /t/. The instances of /t/ were actually the ones developed from the split of the allophone [ḍ] from the phoneme /d/, mentioned earlier. In a later stage, all instances of the Arabic /dˤ/ merged with the Persian phoneme of /z/. I have termed this path of change as intermediary split, while it does not exactly fit the definition for primary, or secondary split. Furthermore, still in another development, the Arabic phoneme of /ðˤ/ shifted to allophone of [δ] which later merged with the phoneme /z/.<br /><strong>5.Discussion and Results</strong><br />Intermediary split does not fully align with the traditional concept of a primary split, although it shares several key characteristics with it. During this process, the Arabic phoneme /dˤ/ underwent a split, resulting in the merging of four distinct phonemes already present in Persian. Among these, the phoneme /d/ played a particularly significant role. Alongside its allophone (i.e., /δ/ / /ɾ̪/), it created the necessary phonological environment that facilitated the eventual replacement and merger of /dˤ/ with the Persian phoneme /z/.<br /><strong>Select Bibliography</strong><br />Anttila, R. Historical and comparative linguistics. Amsterdam/ Philadelphia: John Benjamins, 2009.<br />Barthélemy, A. <em>Gujastak Abāliš</em>. Paris: Vieweg, 1887.<br />Bodrogligeti, A. <em>The Persian Vocabulary of the Codex Cumanicus</em>. Budapest: Akadémia Kiadó, 1971.<br />Kang, Y. “Loanword phonology”. In: Marc van Oosetendorp, Colin Even, Elizabeth Hume.<br />Keren Rice (eds.). Blackwell companion to phonology. 2258-2282. Malden, MA: Wiley-Blackwell, 2011.<br />Meier, F. “Aussprachefragen des älteren Neupersisch”. <em>Oriens</em>, 1981, Vol. 27-28, 70-132.<br />Naghzguy-Kohan, M. <em>Historical linguistics</em>. Tehran: Agah Publication, 2026. [In Persian].<br />Pisowicz, A. Origins of the New and Middle Persian phonological systems. Krakow: Nakladem Uniwersytetu Jagiellonskiego, 1985.<br />Sadeqi, A. A. <em>Evolution of Persian language</em>. Tehran: Daneshgah-e Azad-e Iran, 1978. [In  Persian].<br />Vajda, G. “Les lettres et les sons de la langue arabe d’aprés Abū Hatim al-Rāzi”. <em>Arabica</em>, 1961,  VIII/2, 113-130.<br />  در انشقاق اولیه، واج‌گونه‌(های) یک واج از آن جدا و در واج‌ دیگری در همان زبان ادغام می‌گردد. در فارسی نو سه نوع انشقاق اولیه قابل تشخیص است. مورد اول مربوط به واج‌گونه‌ای از b است که دستورنویسان سنتی آن را فاء اعجمی نامید‌ه‌اند. فاء اعجمی همخوان دولبی واک‌دار سایشی ß بوده که بعدها با ادغام در واج‌های /b/، /v/ و /f/ (که از قبل در زبان موجود بوده‌اند) از بین رفته‌است. به احتمال زیاد این انشقاق در فارسی شمال شرقی و فارسی جنوبی قرون متقدم هجری به یک صورت به‌وقوع نپیوسته‌است. دو مورد دیگر انشقاق اولیۀ همخوان‌ها در فارسی نو مربوط به واج‌گونه‌های d است. در این راستا، انشقاق واج‌گونة سایشی دندانی [δ/ɾ̪] قابل طرح است. این واج‌گونه که در فارسی نو متقدم با <ذ> نمایش داده می‌شده، بعدها به تلفظ اصلی d بازگشته‌است. مواردی از آن نیز در واج /z/ ادغام شده‌است. انشقاق اولیۀ سوم مربوط به جدا شدن واج‌گونۀ واکرفته [ḍ] از واج /d/ و ادغام آن در واج /t/ فارسی است. که تغییر اخیر جزو مسیرهای تغییر در انشقاقی بوده‌است که آن را انشقاق بینابینی نام نهاده‌ایم. انشقاق بینابینی ناظر بر تغییری است که طی آن در قرون اولیۀ هجری، واج قرضی dˤ <ض> به چهار واج /d/، /dz/، /l/ و /t/ انشقاق یافته و خود از بین رفته‌است. واج‌گونۀ δ/ɾ̪ در تعیین مسیر نهایی تغییر واج قرضی dˤ <ض> اهمیت اساسی داشته‌است. با ادغام بعضی از موارد واج‌گونۀ بومی [δ/ɾ̪] در واج /z/ فارسی، راه برای ابدال مواردی از dˤ (که در فارسی به‌صورت δ تلفظ می‌شده‌اند) هموار شده و این موارد نیز در واج /z/ ادغام شده‌اند. با فرایند قیاس بقیۀ آواهایی نیز که اصل‌شان  dˤعربی بوده‌است، در واج /z/ فارسی ادغام شده‌اند. در حدود قرن ششم هجری، با از بین رفتن δ/ɾ̪ بر اثر انشقاق اولیه، dˤ <ض> در تمامی کلمات قرضی جدید عربی )به قیاس با کلمات متقدم قرضی( مستقیما به‌صورت z درآمده‌اند. همین تغییر آخر با پیچیدگی کمتری برای ðˤ <ظ> قرضی اتفاق افتاده‌است.<br />  https://zaban.guilan.ac.ir/article_9520_87b95bc525e746993abbfd0f0d897c4d.pdf