PERSIAN LANGUAGE i. Early New Persian



i. Early New Persian

Early New Persian is the first phase (8th-12th centuries CE) of the Persian language after the Islamic conquest of Iran. This sub-entry is organized into these sections: Introduction, Historical background, Dialectology, Evolution of Early New Persian, Sources, The Grammar of Early New Persian.


The demise of the Sasanian Empire through Arab-Islamic invaders (632-51 CE) and the ensuing Islamization of Iran marks a major break in Iranian history, providing Iran with the basis for a new culture and identity. Out of the remnants of the former “national” Iranian culture, the first Persian documents written in Arabic script emerged during the 9th/10th centuries, continuing the Sasanian Middle Persian language in a new garb.

Definitions. What seems, seen from a historical bird’s eye view, to constitute a decisive linguistic break was actually a slow and steady development of over 200 years, betraying many moments of continuity. The linguistic difference between the 7th-century Middle Persian (MP) and 10th-century Early New Persian (ENP) is certainly smaller, in many ways, than the one between 10th century ENP and 21st-century modern New Persian (NP). The distinction between Middle and New Persian is due as much to convention, to extra-linguistic factors such as the historical break mentioned, and to changes connected to this break such as the shift of script (from Pahlavi to Arabic), as to linguistic differences.

There is no clear linguistic break, either, between Early New Persian and later stages of Persian. The definition of ENP followed here corresponds to the one implicitly given by Lazard (1963, pp. 18, 24), saying that from the 13th century onwards, dialectal features were declining in Persian documents and a dialectally homogeneous standard of the language was more or less reached. Following this definition, ENP would span the stages of New Persian from its beginning, the 8th century, to approximately the early 13th century. A further distinction may be made between a formative phase of ENP which lasted until the early/mid-11th century, when Persian had more or less attained the status of a language of literature, administration, and science, and the ENP of thriving classical literature between the mid-11th and early 13th century.

Even after the 13th century, Persian did not become fixed or standardized. The development of Persian since the 8th century has been continuous and gradual, and instead of speaking of one “classical” standard of the language which never existed, one should rather speak of a series of temporary standards of stylistic nature that were set by classical authors such as Ferdowsi, Saʿdi, or Hafez (see also Paul, 2002). Iranian national philology has defined certain stylistic standards in terms of region, time, or dynasty, e.g., the Khorasani style of early (10th-12th centuries) “realistic” poetry or the Samanid style of simple 10th-11th century prose (Rypka, pp. 112 ff.). Each of these styles could be emulated by later authors (e.g., the 19th-century Qajar poet Qāʾāni wrote in Khorasani style), but they are no “absolute” standards of New Persian. Even modern “Standard High Persian” is only a transitory, flexible idiom that continues developing.

Varieties of Early New Persian. Dialectal and regional factors are of paramount importance for understanding the history of Early New Persian. Likewise important, and partly overlapping with regional factors, is the affiliation of ENP texts to various religious traditions, according to which the ENP varieties should be distinguished in the first place. The “default” variety of ENP, written by Muslim Iranians in Arabic script, is only one variety of the language. Other varieties are documented more weakly by far in terms of quantity, but contribute a lot to our understanding of the evolution of New Persian.

The other main varieties known today include the ENP used by Persian-speaking Jews (called “Early Judaeo-Persian,” EJP), Manicheans, Christians, and Zoroastrians (called “Manichean,” “Christian,” and “Zoroastrian NP,” respectively). Each of these varieties represents one religious tradition, visible both from the script used for writing it (Hebrew, Manichean, Syriac, Pahlavi/Avestan) and from the contents of the scriptures. In terms of age, quantity, and linguistic importance, the EJP variety comes next to ENP written in Arabic script; the Manichean and Christian varieties offer only a small number of short texts. There was also an early Zoroastrian literature written in New Persian, written partly in Middle Persian and partly in Avestan script, of which there are left some substantial texts.

The ENP texts written in Arabic script originate from all over historical Iran, in the earliest phase (late 9th to 10th centuries) with a clear preponderance of northeast Iranian regions (Khorasan). EJP documents were written in, or have been found in, such diverse areas as Egypt, Palestine, southwest Iran (Ahvāz;), Central Iran (Zefra), Afghanistan, western China (Dandan Öilïq), and South India. This distribution does not, however, represent distinct dialects, but rather regional varieties that had been spread widely and recently, due to the mobility of Persian-speaking Jewish merchants, scholars, and travelers of Sasanian and early Islamic times. The oldest dateable EJP documents, and indeed the oldest of all surviving NP documents, are from northern regions: a short inscription from Tang-i Azao (Afghanistan) from 751-52 (ed. Henning, 1957) and a fragmentary letter from Dandan Öilïq probably from 760 (ed. Utas, 1969). Most other EJP documents are later, from the 10th century onwards; a large part of them originate from southwest Iran, but there are also substantial texts from northern regions, including recent finds from Afghanistan.

The few extant Manichean NP texts are from northeast Iran (the region of Samarkand?) and probably date to the 11th century; the even smaller corpus of Christian NP texts are mostly from Central Asia (Bulayıq), from about the same time. Both varieties, typically, come from borderlands of (historical) Iran, where religious minorities could thrive longer than in the Islamicized heartlands of Iran. The extant Zoroastrian NP texts originate from Fārs and Kermān, the regions where Zoroastrians continued to thrive after the Islamic conquest of Iran. There are also late Zoroastrian Middle Persian (ZMP) texts from the 9th and 10th centuries that already show NP grammatical features; it is sometimes difficult to distinguish late ZMP from early Zoroastrian NP.

Each text of each variety of ENP betrays dialectal features from the region where it was found or written, or from where the author originated. The date of many of the Jewish, Manichean, Christian, and Zoroastrian NP texts is not exactly known; some of them may even postdate the 12th century, but due to the remoteness of their speakers from the development of mainstream (Muslim) Persian, they preserved linguistic features that had disappeared from the latter.

Manuscript transmission. The linguistic “value” that can be attributed to the various ENP texts and varieties depends, not only on the time when and the region where these texts were composed, but also on their transmission. Hardly any ENP works written in Arabic script have come down to us in the form of an autograph, i.e., a manuscript written by the author himself. The oldest surviving ENP manuscript in Arabic script, the Ketāb al-Abniya, was copied in 1056, probably 80-90 years after it was written down. For most other ENP manuscripts, the gap between the composition of the work and the oldest extant manuscript is much longer (e.g., 200 years for the Siāsat-nāma, or over 250 for the Šāh-nāma). Nevertheless, the sheer number of available texts and manuscripts, their dialectal variety, and the high standard of philological analysis that has been attained for ENP texts in Arabic script still make this variety the key variant for linguistic analysis of ENP.

The advantage that EJP manuscripts offer over ENP ones, besides being partly older texts (see above), is that the manuscripts in which they are transmitted are also much older than the ENP ones. The EJP private documents, as non-literary texts, are “naturally” autographs preserved in their original form and thus provide a genuine and lively picture of the Persian spoken in southern Iran in the 10th/11th centuries. The extant EJP tafsir (commentary; see JUDEO-PERSIAN COMMUNITIES ix. JUDEO-PERSIAN LITERATURE) manuscripts, by far surpassing the private documents in quantity, were probably copied not very long after their composition; linguistic differences between various copyists, between different styles within certain manuscripts, and the use of two different systems of vocalization (Tiberian and Babylonian) greatly enhance the linguistic value of these texts. The Man. NP and Zor. NP texts suffer from historical writing conventions which they owe to the MMP and ZMP traditions, respectively. Besides the late date of some manuscripts (especially Zor. NP ones), this factor limits the linguistic value of both varieties of ENP.


The historical context in which Persian written in Arabic script, as a literary and official language, evolved from the 9th century onwards has been described variously as one of an “Iranian renaissance,” “national rebirth,” etc. It is true that, unlike other ancient Near Eastern countries such as Syria or Egypt that were Arabicized as a consequence of Islamization, Iran kept a language of its own, partly preserved its pre-Islamic cultural heritage, and thus laid the foundation for a later cultural renaissance that was bound to a language and identity separate from the Arabic one. Talking of a “national rebirth” or the like, however, is not only misleading, because there was no Iranian nation as a political or social unit at that time, but also because it distracts from the real historical developments, which were more complex (for an overview, see also Lazard, 1975).

The quick conquest of most of the Sasanian empire by Arab-Islamic troops between 632 and 651 was possible only after years of exhausting warfare of the Sasanian with the Byzantine empire. The successful and relatively “smooth” integration of Iran under Islamic rule hints at a lack of political and social stability in late Sasanian Iran and was facilitated through the cooperation of a substantial part of the Iranian landed aristocracy (see DEHQĀN) and administrative elite with the new Islamic rulers. Iran became a constitutive part of the Umayyad caliphate (661-750). The new rulers had to rely on Iranian (and Byzantine) administrative and fiscal traditions to govern the vast area, and it was only the fifth caliph of the Umayyad line, ʿAbd-al-Malek b. Marwān (685-705), who introduced an Arabic coinage of his own and fostered Arabic as a language of administration (Gibb, p. 77).

The historical province of Khorasan, including also parts of contemporary Central Asia and northwest Afghanistan, played a key role in the evolution of New Persian. During the Sasanian period, Persian had already been spread over to this region, as an official language, through garrisons and state administration, at the expense of Parthian, the local and former official language. During the Arab-Islamic conquest of Iran, substantial numbers of Sasanian aristocrats fled the Iranian heartland to northern Khorasan (Samarkand), further spreading the Persian language there at the expense of vernacular Sogdian. After 651, the whole of Khorasan was conquered by Arab Muslims within two generations, and large numbers of Arabs were settled in arid Khorasan, finding conditions for living and cattle-breeding there similar to those of the Arabian peninsula. The relatively quick Islamization of Khorasan, far away from the centers of Arab culture, provided a kind of cultural laboratory for the development of something totally new: the transformation of the (late) Sasanian vernacular, called Dari, into a simplified language written with the Arabic script, Muslim (New) Persian.

After the Abbasid revolt in Khorasan (747-50), led by the Iranian Abu Moslem, Iran became part of the Abbasid caliphate (750-1257), but from the mid-9th century onwards, the caliphate gradually lost its military and political power and disintegrated into more or less independent states. One of these, led by the Samanid dynasty (819-1005), ruled northeast Iran for a long time and was especially important for the renaissance of Persian. At the Samanid court in Bukhara, Persian flourished as a language of poetry, ancient Iranian epic traditions were collected, and historical and scientific texts were translated from Arabic to Persian. The oldest known Persian poets are recorded in the second half of the 9th century at the Samanid court and at that of the Saffarids of Sistān, who would later (911) become the Samanids’ vassals. At the close of Samanid rule, Persian was well established as a language of poetry and epics in northeast Iran, but only partly in other regions of Iran, and not yet in all genres of scientific writing and of administration.

This situation changed in the course of the 11th century. The Ghaznavids (977-1186), who overthrew the Samanids and conquered most of their territories in the 990s, were the first of many dynasties of Central Asian provenance and Turkic origin to rule over considerable parts of historical Iran. They fully continued the Samanids’ patronage of Persian literature and culture, even after they lost Iran proper to the Saljuqs and were restricted to their territories in Afghanistan and north India. This marks an important step towards the internationalization of Persian beyond Iran proper, as the Islamic language of the “Eastern Caliphate.”

When the Saljuq Turks conquered Iran from the Ghaznavids and Buyids in the mid-11th century, they found Persian as a fully developed language, to which their ruling classes soon assimilated. Under their dynasty (1040-1194), a great part of historical Iran was united under non-Arab rule, for the first time since 651. But the Saljuqs, like all other Islamic dynasties ruling Iran before them, did not yet use the name “Iran” for their land or state. Only under the Mongol Il-khanids (1256-1353) was Iran to appear again as a political unit bearing its traditional name (see Krawulsky).

Towards the late 11th century, the status of Persian written in Arabic script may be described as follows. (1) NP had become fully developed as a language of literature, administration, and scholarship. (2) NP had become an “Islamic language,” as is seen from the contents of its texts and is measurable through the high influx of Arabic loanwords into Persian. (3) NP had started expanding over the boundaries of the historical Persian-speaking regions proper, into other areas such as Central Asia and North India.

These three factors, in combination, laid the basis for a remarkable development that would influence the cultural history of the Islamic world over a long period. Persian, as one main carrier of Iranian and Islamic civilization, would play a leading role for the expansion and thriving of Islamic civilization from the 11th up until the 19th century, in a vast area spanning Anatolia in the west up to the fringes of China in the east. During this time, Persian language, literature, and culture exerted a decisive influence as a civilizing tool on the Turkic, Mongol, and Indian dynasties and their cultures that dominated the eastern half of the Islamic hemisphere politically. For the cultural dominance of Persian over this vast area and long period, the term of “Persophonie” has been introduced by B. Fragner (1999), modeled after the French term “francophonie” and being a scientific equivalent of Persian qalamrow-e zabān-e fārsi. While this is correct at the descriptive level, it remains yet to be investigated in what sense the Persian language itself could have been a moving factor in this historical process.


For a proper understanding of the evolution of ENP, a closer look at its dialectal situation in the 7th century is necessary. Most valuable information about this is provided by the Iranian-born Arab polymath Ebn al-Moqaffaʿ (721-57), as transmitted in the 10th-century Ebn al-Nadim’s Fehrest. According to Ebn al-Moqaffaʿ, five languages were spoken in late Sasanian Iran, three of which are Iranian: Pahlavi used in central and northwest Iran, the ancient region of Media (Rey, Isfahan, Hamadān, Azerbaijan, Māh Nehāvand), Dari spoken at the court of al-Madāʾen (Ctesiphon-Seleucia) and in Khorasan, and Pārsi spoken in the province of Fārs and used by the Zoroastrian priests (the two probably non-Iranian languages mentioned by Ebn al-Moqaffaʿ, Ḵuzi and Soryāni, are beyond the scope of the present article).

Lazard (1971) has analyzed this passage carefully, showing that Pārsi and Dari are two functional and regional varieties of Persian, Pārsi representing the literary language, especially that of religious (Zoroastrian) literature, and the regional variety spoken in southern Iran, and Dari representing the language spoken at court and in the northeast of Iran (including large regions of present-day Afghanistan and Central Asia). “Pahlavi” stands for Parthian, in which ancient epic traditions were preserved in various regions of Iran, even after Parthian had ceased to be used as an official language. It is appropriate to discuss separately the main dialect regions distinguished by Ebn al-Moqaffaʿ, the northwest, northeast, and south.

Northwest. According to Ebn al-Moqaffaʿ (Ebn al-Nadim, ed. Tajaddod, p. 15; tr. Dodge, I, p. 24), Pahlavi (= Parthian) was spoken in central and northwest Iran, the ancient region of Media (Rey, Isfahan, etc.), which is called Fahla or Bahla in Arabic sources. Parthian, however, was not indigenous to that region. The homeland of the Parthians in Arsacid times had been Khorasan; at the end of the Sasanian period, Parthian must have been extinct or nearly extinct there. After the 6th century, Parthian continued to be used only as a language of the Manichean liturgy in Central Asia (along the Silk Road; see Sundermann, 1986, p. 315). The idiom used in central and northwest Iran and called “Pahlavi” in the 7th century is not likely to be Parthian proper, but more probably a bundle of “Parthoid” dialects, forerunners of the northwest Iranian languages and dialects spoken there until today, such as Ṭāleši, Southern Tāti, or variants of Āḏari (see Azerbaijan vii. The Iranian Language of Azerbaijan).

The semantic development of the term Pahlavī in Early New Persian seems to corroborate this. In Sasanian times, when Parthian gradually disappeared as a living language, ancient Parthian epics were integrated and assimilated to Middle Persian literature (see, e.g., the MP Ayādgār ī Zarērān whose Parthian loanwords show that the text was based on a Parthian original). The term Pahlavī was now identified with “heroic, old, ancient” (Lazard, 1972, p. 35), and later the name of the bygone Parthian language was transferred to the Persian language written by Zoroastrians that seemed likewise to be ancient or heroic to the Iranian Muslims (note that Pahlavī meaning “Zoroastrian Middle Persian” developed as an exonym used by Iranian Muslims, while the endonym of the Zoroastrian Persians for their own language continued to be Pārsī(g) for some centuries). Another meaning of Pahlavī in Early New Persian, attested later in the Arabicized plural form fahlaviyāt and designating poetry in the dialects from the Fahla/Bahla region from the 10th century onwards, is a direct continuation of that of the 7th-century “Parthoid” dialects.

Northeast. Khorasan, the homeland of the Parthians (called abaršahr “the upper lands” in MP), had been partly Persianized already in late Sasanian times. Following Ebn al-Moqaffaʿ, the variant of Persian spoken there was called Darī and was based upon the one used in the Sasanian capital Seleucia-Ctesiphon (Ar. al-Madāʾen). The regions adjacent to Khorasan, namely, Sogdiana (capital: Bukhara), Bactria (Balḵ) and Transoxiana/Chorasmia, were Persianized later, after the Arabic-Islamic conquest of Iran—a process that coincided with their Islamization. (Centuries later, the Persianization of Sogdiana and Transoxiana should be followed by their Turkicization.)

Under the specific historical conditions that have been sketched above, the Dari (Middle) Persian of the 7th century was developed, within two centuries, to the Dari (New) Persian that is attested in the earliest specimens of NP poetry in the late 9th century. Compared to Middle Persian, Dari New Persian has undergone substantial grammatical changes, e.g., it has given up the MP eżāfa ī(g) as a relative clause marker, the indirect (and sometimes direct) object marker ō, and the ergative transitive constructions of the past. Compared to Dari New Persian, which was the result of a historical break, the Pārsi New Persian of southern Iran remained closer to Zoroastrian Middle Persian.

South. The variety of Persian called Pārsī by Ebn al-Moqaffaʿ, spoken in southern Iran and used as a literary language by the Zoroastrians of Fārs, continued well into Islamic times. Zoroastrian literature in Pārsi (later called “Pahlavi” by Iranian Muslims, see above) thrived during the 8th-9th centuries, perhaps also as a reaction to increasing social and cultural pressure from the Islamic caliphate and its local agents such as the Taherids or Saffarids. The language of the southern EJP documents can be considered more or less a continuation of ZMP.

Up to the 11th or early 12th century, these EJP documents preserved MP grammatical features and lexemes that had already vanished from contemporaneous Dari New Persian in the northeast, e.g., the eżāfa i introducing relative clauses, the preposition o “to” (< MP ō), or traces of ergativity (Paul, 2013, §156). There is no evidence to prove that the Jews and Zoroastrians of southern Iran considered the form of Persian they used during the 8th-11th centuries as a “new” form of Persian detached from (late) Middle Persian. They continued to call their language Pārsī (< MP Pārsīg) well into the 11th century and beyond (e.g., EJP Pārsī “Persian”; Gindin, p. 83, Ez12.11; or MP Pārsīg in a text from ca. 900; Safa-Isfehani, pp. 80-81). The “change of paradigm” for the southern varieties of Persian came with the Saljuq dynasty, under whose rule the Islamicized Persian from the northeast was spread to all other parts of Iran as a language of administration, culture, and communication and increasingly influenced all other varieties, including the EJP ones.

While many of the EJP texts written in Hebrew script are from southwestern Iran (e.g., Ahvāz), a dialect translation of the Koran found in the 1970s was produced in the southeast (Sistān), probably in the early 11th century (see Ravāqi). The language of this Koran translation has certain features in common with that of contemporaneous southwestern EJP texts. This shows that the whole of southern Iran, from Ahvāz in the west to Sistān in the east, was home to a homogeneous group of dialects at that time, which were more conservative than the varieties of (Dari) Persian spoken in northeastern Iran (Lazard, 1990). The affiliation of Sistān to the “southern belt” of Persian dialects does not rule out the fact that, at the Saffarid court in Sistān, Dari Persian “high” literature could be patronized by Saffarid rulers such as Yaʿqub b. Layṯ.

The dialectal basis of Darī New Persian. Modern New Persian is based on the Dari Persian of Khorasan, and (following Ebn al-Moqaffaʿ) Dari shares a common dialectal basis with the variety of Middle Persian that was spoken at the former Sasanian court in the capital Ctesiphon-Seleucia (al-Madāʾen); this is likely to have been the basis for literary Manichean Middle Persian (MMP). All three historical stages of Persian should therefore be dialectally homogeneous; this is, however, not the case. MMP had more purely preserved its Persian (southwest Iranian) linguistic type as compared to Zoroastrian Middle Persian, the official language of the Sasanian empire, which had absorbed more lexical influences from Arsacid Parthian, the language of their dynastic predecessors (e.g., MMP damistān “winter,” bišehk “doctor,” bāw “garden,” as against ZMP zamistān, bizišk, bāḡ, all from Parthian). New Persian, although based on Dari NP, shares the ZMP features of these words (NP zemestān, pezešk, bāḡ), showing that Dari NP is not based solely on MMP.

To explain this, one has to keep in mind that the Dari of the 7th century was not a homogeneous language or dialect, but basically the spoken variety of Sasanian Persian, as against Pārsi, the language of administration and (predominantly Zoroastrian) literature. Dari must have absorbed many influences from ZMP and from Parthian during the Sasanian period, and there must have existed various local varieties or styles of Dari by the 7th century. There are no extant Khorasani Dari documents from the 7th-8th centuries, but it is possible to get an impression of its dialectal non-homogeneity from the later Khorasani or northeastern prose texts of the 10th century (including southwestern forms like damestān “winter” from the Ḥodud al-ʿālam; Lazard, 1963, p. 54, fn. 8; or the lexical doublet (h)ēzom/hēma “firewood”; Lazard, 1963, p. 166; cf. ZMP/MMP ēzm/ēmag “id.,” with ēzm being from Parthian; in later NP, hizom < (h)ēzom< ēzm prevailed).


Starting from Ebn al-Moqaffaʿ’s account of the linguistic situation in the 7th century, accounts and statements of other Arabic and Iranian authors about the linguistic situation in 8th-11th-century Iran serve to complement the information given by Ebn al-Moqaffaʿ and to follow up on the evolution of Persian during that time. These complementary accounts help to draw a more lively and complete picture of the usage of Persian and of the views of the people on the language(s) they used. It must be borne in mind that language names are not always stable, but may be flexible and subject to semantic change in the course of time. Ebn al-Moqaffaʿ himself gave a good example of this flexibility. He used the term “Pārsi” in its specific meaning of “southern Persian” but also as a cover term meaning “Iranian,” covering all languages (even non-Iranian ones) that were spoken in Iran (Lazard, 1971, p. 363); from this meaning, the term Pārsi-(ye-)Dari “the Pārsi (variant, namely:) Dari” would later be derived. Most of the accounts have already been dealt with by Lazard, 1971, Ṣādeqi, and Perry, but will be studied here partly in a new perspective.

Geographical spread. Ebn al-Moqaffaʿ’s opaque remark that Dari was spoken not only in Khorasan, but that the “dominant” variety of Dari was that of Balḵ, becomes clearer through statements by other authors such as Ṭabari that Pārsi was spoken in Balḵ (referring to the situation of 726; Perry, p. 51), or Nażr b. Šamil that Pārsi-Dari is the language used in Balḵ (before 822; Ṣādeqi, p. 44). Balḵ, thus, must have been an established center of Persian language and culture in early Islamic times. Later, Jāḥeẓ writes that the people of Marv spoke in the sweetest and most skillful manner among Iranians (before 868; Lazard, 1971, p. 381); Naršaḵi, that the Koran was read in Pārsi in Bukhara (953; Lazard, 1971, p. 372); and Ebn Ḥawqal, that the inhabitants of Bukhara spoke Dari besides Sogdian (973; Ṣādeqi, p. 49). Although it cannot be ruled out that Pārsi was used for “Sogdian” by Naršaḵi (as Lazard, 1971 assumes) or even for “Bactrian” by Ṭabari and Nażr b. Šamil, these statements more probably refer to (Dari) Persian, and attest to the wide spread of Persian in the whole of eastern Iran in early Islamic times, which had already developed during the Sasanian period.

Functional load: administration. In the former Sasanian parts of the Umayyad empire, Arabic replaced Persian only gradually as a language of administration, from the late 7th to the mid-8th century (see above). This is reflected in statements that the administration of Iraq had always been in Pārsi and was, by now, about to be replaced by Arabic (Balāḏori, Jahšiāri; Lazard, 1971, p. 369; both authors wrote in the late 9th and early 10th centuries, but their statements refer to the 8th century). From later times (early 10th century), Eṣṭaḵri reports that the administration is in Arabic, while the Zoroastrian priests still write in Pahlavi (= now Middle Persian) and speak in Pārsi (Lazard, 1971, pp. 365, 379).

Functional load: literature. The literary reawakening of Persian in eastern and northeast Iran, at the courts of the Saffarids and Samanids during the latter half of the 9th and the 10th centuries, is also well reflected in quotations. The Saffarid ruler Yaʿqub b. Layṯ, who did not understand Arabic, ordered one of his viziers to translate an Arabic poem into Persian (ca. 870; Tāriḵ-e Sistān, pp. 209-10). Similar motifs are attested from various authors at the Samanid court, where Persian literature was actively promoted. In his preface to the Šāh-nāma, Abu Manṣur explains that the ruler ordered old epics that had been translated from Middle Persian into Arabic to be translated into (New) Persian, so that the text would “reach the people“ (957; Qazvini, p. 33). For the same reasons, Balʿami translated parts of the Tāriḵ-e Ṭabari into Persian (963; Lazard, 1971, p. 374), and Meysari decided to write a medical poem in Persian rather than Arabic, because it would be thus understood by a larger audience (978; Lazard, 1971, p. 371).

Defining Persian against “the other”: Arabic. While for the 7th century, the linguistic situation in Iran was described in terms of the interrelation of the predominant Iranian idioms Dari, Pārsi, and Pahlavi, during the 8th century this changed, Arabic becoming an important point of reference, or “the other,” to describe the status and usage(s) of Persian. This is reflected, not only in statements about the usage of Persian and Arabic (see above), but also, once Persian had started its “march through the literary genres” towards its reestablishment, in negative, “reactionary” statements against Persian as a language in its own right. Abu Ḥātem Rāzi said that for many Arabic terms the Persians have no equivalent in their language; more specifically, they do not have a metrical system, or a word for “poet,” and therefore do not really know what poetry is (before 934; Lazard, 1971, p. 371). Maqdesi (Moqaddasi) stated that the most eloquent variety of Pārsi was that of Khuzestan, which was mixed with Arabic (before 985; Lazard, 1971, p. 372). Biruni noticed a loss of clarity if Arabic texts were translated into Persian, and Ṯaʿālebi confirmed the verdict of Abu Hātem Rāzi that for many Arabic words there is no (suitable) Persian equivalent (both before 1038; Ṣādeqi, p. 39). These comments on the deficits of Persian as against Arabic may, or must, be understood as negative reactions to the ongoing process of the establishment of Persian in functions hitherto reserved for Arabic, and thus as indirect proofs for this process.

Defining Persian against “the other”: Darī/Pārsī. After the mid-11th century, when the Persian of Khorasan (called Dari or Pārsi-ye Dari) had become established as a literary and scholarly language, and Arabic was already on retreat in Iranian lands, the relation of Pārsi-ye Dari to other Iranian idioms was again commented upon by various authors. Nāṣer Ḵosrow mentions that the poet Qaṭrān (from Tabriz) did not speak Persian well (called Fārsi here; ca. 1046; Lazard, 1971, p. 380). Asadi Ṭusi compiled his Dari Persian dictionary for the benefit of the people of Azerbaijan, whom he intended to make familiar with this language (ca. 1060; Asadi, p. 4). Both quotes reflect a need for Dari Persian to be explained beyond the limits of Khorasan. This shows that Dari Persian had indeed started spreading beyond the borders of Khorasan, to the west, where the people spoke various “Parthoid” (e.g., Āḏari) dialects, and where the knowledge of Persian was still very limited.

Further information is provided by Šāhmardān b. Abi’l-Ḵayr, saying that the Persians are using a “(too) pure” form of Dari (Darī-yi wēža-yi muṭlaq; 1073; Perry, p. 59), and by Kay Kāvus b. Eskandar in his Qābus-nāma, that it is not good to use “pure“ Persian, especially if it is Dari Persian (1082; Lazard, 1971, p. 384). In both quotes, a discomfort is expressed against the cultural renaissance of Dari Persian, whose “Persianness,” the authors feel, should at least be tamed by the usage of a larger number of Arabic words. The reproachful word of wēža “pure” may target the many Parthian loanwords that had been taken over by Dari Persian recently and were still unintelligible to non-Dari Persian speakers in the late 11th century.


The majority of ENP sources are literary and extant in the form of manuscripts. They may conveniently be grouped according to their religious affiliation(s).

Early New Persian in Arabic script. Poetry. Islamic Persian literature starts with poetry (Lazard, 1975, p. 595). The earliest works of Persian poetry known today go back to the mid-9th century (e.g., the poets Ḥanẓala Bādḡisi or Moḥammad b. Waṣif; Lazard, 1964, Pt. I, pp. 17-18). No old and independent manuscripts of these poets’ works are extant, probably in part because at that time the oral transmission of poetry still predominated; they are known only from short quotations in later collective works such as the dictionary Loḡat al-Fors (mid-11th century) or the anthology Lobāb al-albāb (early 13th century). Despite its old age, ENP poetry is less important linguistically for the study of the evolution of New Persian than other forms of ENP (especially prose literature).

Epics. The Iranian national epic Šāh-nāma (composed in the late 10th and early 11th centuries by Ferdowsi) is of paramount importance for the study of ENP. Its historic and mythic contents had been collected from older compilations such as the MP Xwadāy-nāmag (see HISTORIOGRAPHY ii). Ferdowsi based his monumental work on a prose version (now lost) collected under the dehqān and temporary Samanid governor of Ṭus, Abu Manṣur b. ʿAbd-al-Razzāq, integrating also parts from an epic of the earlier poet Daqiqi (de Blois, 1997, p. 105-08). The language of the Šāh-nāma is different from that of contemporaneous poetry or prose: incorporating also “Parthian” epic material, it contains a lot of Parthian loanwords, not all of which found their way later into “standard” Persian (e.g., borz “high,” pur “son,” later NP bālā, pəsar). It shares certain grammatical features with ENP texts of northeast Iranian provenance (e.g., the perfect tense of the type kardastam), but retains its own characteristics. Being a compilation of older material, the percentage of Arabic loanwords in the Šāh-nāma is lower than in contemporaneous poetical or prose works (this explanation seems more probable than to assume that Ferdowsi avoided Arabic words intentionally). The Šāh-nāma stands out as the Iranian national epic that reconciles ancient Iranian myths, legends, and traditions with the predominant Islamic culture of the time.

Prose. For ENP prose texts, Lazard’s seminal study of 1963 remains the reference work that lists all extant ENP prose works of linguistic interest, namely 47 texts of the earlier period (until ca. 1090) and 24 texts from the late 11th and 12th centuries. They include the well-known “classical” texts that made Persian develop into a language of scholarship and prose literature during the 10th-11th centuries, in such fields as historiography, medicine/pharmacy, astrology, philosophy, and “wisdom literature.” Lazard provides ample information about each text, the manuscripts in which it has been transmitted, and its linguistic value.

A selection of works serves to illustrate how Persian was gradually coming of age as an Islamic literary language in the course of the 10th and 11th centuries. The Arabic original of the Tāriḵ-e Ṭabari was compiled until ca. 915 by Abu Jaʿfar Moḥammad b. Jarir al-Ṭabari, at a time when no one yet conceived of Persian as a language of Islamic learning or scholarship. Fifty years later, a need was felt to translate into Persian those parts of the Tāriḵ-e Ṭabari that pertained to the Iranian past, a task entrusted by the Samanid Manṣur b. Nuḥ to his vizier Balʿami. It seems that historiography, important for the rulers to refer to a heroic past of their dynasty, paved the way for other scholarly disciplines, such as medicine or astrology, to be written down.

Persian prose literature of the second half of the 10th century mainly consisted of translations from Arabic works or of works written on the basis of Arabic sources (e.g., the pharmacological Ketāb al-Abniya, probably composed around 970). This changed slowly during the first half of the 11th century. The Chorasmian polymath Biruni still expressed a strong preference for Arabic over Persian as a language of scholarship, e.g., in his famous saying that he preferred to be abused in Arabic, rather than being praised in Persian. Around 1030, he nevertheless composed one of his works, the astrological Tafhim, in Arabic and Persian. The celebrated Iranian-Islamic scholar and philosopher Avicenna also wrote most of his works in Arabic, but around 1030 he likewise chose to write at least one of his major philosophical works, the Dāneš-nāma, in Persian, and even made an effort to create Persian equivalents to philosophical terms that had hitherto existed only in Arabic.

In the second half of the 11th century, when the Saljuq dynasty had taken power in the caliphate, Persian had already been established as a language of literature and science, and increasingly of administration. The high prestige of Persian as an Islamic language on equal right with Arabic is illustrated by the Siāsat-nāma written by the Saljuq vizier Neẓām-al-Molk (q.v.) in 1091, which is the most important example of “wisdom literature,” or a “mirror of princes,” in Islamic literatures. Comprising ancient Iranian (Sasanian) governmental traditions, this genre goes back to Arabic works written by the Iranian genius Ebn al-Moqaffaʿ in the 8th century.

Epigraphy. ENP inscriptions are very scarce and short and provide little evidence for the linguistic development of Persian. The earliest extant inscription is that on a silver bowl from the 9th century; the earliest monumental inscriptions are from the mid-11th century, on buildings in the Qarakhanid territory (i.e., in the periphery of the Islamic-Iranian world). The Samanids, who actively promoted the development of Persian literature, do not yet seem to have questioned the priority of Arabic over Persian in their monuments and artifacts (O’Kane, pp. 11-12, 17, 25).

Early Judaeo-Persian. Most EJP manuscripts originate from the Cairo Geniza; they had been stored in the Geniza and thus had been preserved up to the late 19th century, due to Jewish religious concern about the desecration of the name of the Lord written on them. Most extant EJP works are of religious nature, e.g., tafsirs, translations of and commentaries on books or parts of the Hebrew Bible in EJP. The most important single tafsir, that of the book of Ezekiel, is also the most comprehensive of all EJP manuscripts, probably from the 11th century, comprising 226 manuscript pages (ed. Gindin) and thus more than one-third of the entire EJP corpus. Other religious texts include a halakhic (biblical law)  treatise, a sefer miṣwot (ed. MacKenzie, 1968). Grammatical treatises (like the one edited by Khan) also belong to the religious sphere, because they analyze the language of religious texts. The text edited by Khan and some of the tafsirs originate from the Jewish sect of the Karaites, whose anti-Rabbinic theology was highly influential during the 9th-11th centuries in many parts of the Middle East.

Another group of EJP texts, in quantity much inferior to the tafsirs, but of great linguistic importance, are the private documents. These include two legal documents (contracts), dated 951 and 1020 and published by Shaked (1972) and Asmussen (1965), and about 20 private/commercial letters, probably from the 11th century; some of them are fragmentary or very short, of which so far only four have been published by Utas (1969), Zhan/Guang (2008), and Shaked (2010, two letters). The two letters published by Utas and Zhan/Guang are not from the Cairo Geniza but were found in Central Asia (Dandan Öilïq). The contracts are written partly in official/legal style, but contain also passages in lively or “everyday” Persian of the time, as do the private letters throughout. This makes these documents very valuable linguistically, especially in relation to the tafsirs, whose Persian renders Hebrew syntax, and sometimes morphology, slavishly. The rest of the EJP texts include various genres such as  poetry and  medicine, mostly short texts that are not as early as the private documents and therefore do not have much to contribute to the study of ENP. Recently, new EJP texts have been found in Afghanistan that are at present being prepared for publication under the supervision of Shaul Shaked. A grammar of EJP (not yet including the Afghan documents) is being published by L. Paul (forthcoming 2013).

The most important sources of the other three varieties of ENP (Christian, Manichean and Zoroastrian NP) include the fragments of a Syriac-Persian bilingual psalm found in Bulayıq (Sims-Williams, 2011), possibly from the 11th or 12th century; a Manichean didactical text, probably also from the 11th century (ed. Sundermann, 2003); and Zoroastrian texts such as the Zarātošt-nāma, the ʿOlamā-ye Eslām [q.v.] (de Blois, 2007), and a short but interesting report about the reform of the Zoroastrian calendar, from the early 11th century, which was appended to a Pahlavi Rivāyat (de Blois, 2003).


Investigating the grammar of Early New Persian, one must bear in mind that there cannot be one grammar of ENP, because ENP was not a unified language during the 8th-12th centuries. Due to the disintegration of Iran and of Persian culture and language after the Islamic conquest of Iran, there was a great variety of Persian dialects during this period. They are usually grouped as those of the northeast, where Persian evolved as an Islamic literary language called Dari, and the south, where it continued the variety of Middle Persian called Pārsi (see above).

The following grammatical sketch of ENP will basically compare its two main extant varieties, the one written in Arabic script mainly in northeast Iran, and the EJP written in Hebrew script predominantly in southern, but also in northern Iran. Since “ENP” is a cover term denoting both varieties, the acronym “ENPA” will be used where only the northeastern variety written in Arabic script is meant.

Phonology. The phonemic system of ENP is remarkably similar to the systems of MP and modern NP. ENP consonants are shown in Table 1.

The most important deviation from the MP system includes the ENP additional stop q, and the glottal stop (ʾ), both taken over from Arabic loanwords, and the voiced fricatives β and δ, probably allophones of b and d respectively.

The bilabial fricative β results from a lenition of b after a vowel and before a voiced consonant. Its distinct pronunciation appears most clearly from an f written with three dots in certain ENPA manuscripts (e.g., aβzūdan “increase”; Lazard, 1963, §1). In intervocalic position, b may further become lenited to the bilabial semivowel w (e.g., biyāwān  “desert”; Lazard, 1963, §5). The phonemic distinction between b and w is thus weakened, or partly neutralized, in certain ENPA dialects, β showing an intermediate stage between b and w and being an allophone of b. B may alternate with bilabial glide w also word-initially, in words starting with MP w-, e.g., ENPA bahā/wahā “price,” ENPA barzeš/warzeš “training“ (Lazard, 1963, §8), EJP bāyenda/wāyenda “bird.” A pronunciation of w as [β] in this position is also possible, especially in EJP, where there are spelling doublets like warda/ḃarda “captive” (Paul, 2013, §18). The influx of Arabic loanwords into Persian, “importing” a clear distinction between w and b into the language, may have contributed to the stabilization of this distinction in ENP (Pisowicz, pp. 119-20).

Postvocalic d was written with letter δāl and pronounced [δ] in words of Iranian origin in many ENPA manuscripts up to the 13th century. It may be considered an allophone of d in this position, and it was shifted back to [d] from the 13th century onwards, probably under the influence of certain northeastern Persian dialects that had always preserved [d] (with a few exceptions like guδaštan “pass”; Pisowicz, pp. 107 ff.; Meier, pp. 103-13). In words of Arabic origin, however, d and δ were always kept distinguished in all positions; cf. badr “full moon” vs. baδr “seed.” This led F. de Blois (2006, p. 94) to assume that d and δ were separate phonemes in ENPA, with their distinction “imported“ into ENP through Arabic loanwords. If Persians, however, pronounced postvocalic *d as [δ] in words of Persian origin, they probably pronounced postvocalic d in Arabic words likewise (e.g., bdr [baδr] “full moon”), the written dāl here being due to the conservatism of the script. A postvocalic pronunciation of d [d] may have existed in Arabic loanwords, if at all, only in “learned” pronunciation.

The Hebrew letter dālet may represent both postvocalic stop d and fricative δ; the occasional writing of postvocalic d as d̄ in a few EJP documents, however, further corroborates its pronunciation as [δ] (Paul, 2013, §16). A comprehensive discussion of the issue of d/δ is given in Filippone, pp. 184-89.

The lenition of b and d may already have started in MP, but it was certainly not phonemic there. The velar fricative γ (< *g) should, however, be considered a phoneme already in MP, even if it is attested only in ZMP, and mostly in loanwords from Avestan and Parthian (e.g., ZMP bāγ ~ MMP bāw “garden”; see MacKenzie, 1967, p. 22).

The most important phonetic and phonemic changes from ENP to later NP include the following: (1) The pronunciation of wāw shifted from a bilabial glide [w] to a labiodental fricative [v], possibly due to Turkic influence (Pisowicz, p. 120). (2) ḵʷ, which should be considered a phoneme in MP and ENP, lost its labial component (e.g., ḵʷār “mean” → [ḵār], except for ḵʷa which turned into ḵu, later ḵo; e.g., ḵʷad “self” → [ḵod]). (3) The ENP distinct phonemes ġ and q turn into positional allophones with post-velar articulation in modern spoken NP, especially in the prestigious variety spoken in the capital Tehran (word-initially and before consonants [q], between two vowels fricative [ġ]). In “learned“ pronunciation and in many areas such as Yazd, Kerman, and Kabul, the two sounds may still be distinguished.

The ENP system of vowels, which also differs remarkably little from the MP one, is:

ā ē ī ō ū

a (e) i (o) u

Diphthongs: ai, au

In MP, short e and o were probably marginal phonemes (MacKenzie, 1967, p. 23). In ENPA, they do not seem to exist, but their existence is likely in EJP (e.g., hyst [hest] “he/she is,” Paul, 2013, §26). In modern NP, ENP i/u have been shifted to e/o, and the majhūl vowels ē/ō to i/u. More importantly, vowel length is no longer the main distinctive factor of the vowel system in modern NP, but place of articulation, the NP system being ā (=[å]) i u / a e o. The distinction of a/ā may be neutralized in ENP, especially in such phonetic environments as before h (Paul 2013, §28).

Other phonetic features. The ZMP consonant clusters with s or š as a first element like ZMP stadan “take” show a prothetic i- already in MMP (istadan), and both prothetic and anaptyctic i side by side in ENP (istadan, sitadan, Lazard, 1963, §105; Paul, 2013, §47). In later forms of Persian, there is a preponderance of anaptyctic forms (modern NP setadan), with very few prothetic exceptions (e.g., estaḵr “pool”). Another feature of ENP is the partial preservation of initial a- of many MP words that would regularly be dropped in later NP, where both varieties of ENP represent an intermediate stage of the dropping process (e.g., MP abar “upon,” ENP (a)bar, NP bar; Lazard, 1963, §108; Paul, 2013, §46).

Table 2 lists the phonemic and phonetic changes from MP to NP discussed above.

Morphology. While Old Persian was a highly inflecting language of the ancient Indo-European type, MP had given up most categories of nominal, and was about to give up most categories of verbal, inflexion. In MP, most case relations were expressed by prepositions, and a totally new, synthetic system of verbal tenses was about to be established. The ENP and later NP tense systems can be seen as continuations of that of MP, even if not all details can as yet be fully explained.

Nouns. In the earliest MP texts from the 3rd-4th centuries, a case distinction for relationship nouns in the singular still existed, e.g., Rct. brād, Obl. brādar “brother” (Sundermann, 1989, p. 154, with literature). In later ZMP, both brād and brādar still occur side by side, but they are used indiscriminately, without case distinction. Barād “brother” is attested (besides barādar) also in the EJP private letters, likewise without case distinction (Paul, 2013, §66).

The plural of ENP nouns is marked by the suffixes -(g)ān (< MP) and -(i)hā; the latter goes back to a MP adverbial ending -īhā. The form -ihā (with -i-) occurs regularly in many EJP texts but only in a few ENPA texts (Paul, 2013, §78; Lazard, 1963, §152), and the -i- was dropped in later stages of Persian. The distribution of -(g)ān and -(i)hā is the same in all varieties of ENP and modern NP; -(g)ān is used for animate and a small number of other nouns such as  plants and body parts (e.g., deraḵtān “trees”; Lazard, 1963, §149), and -(i)hā is used for all other nouns. In modern NP, - shows a tendency to be generalized at the expense of -(g)ān, especially in colloquial registers.

The particle -ē (< MP -ēw) gives the sense of indefiniteness and unit when attached to an ENP noun. It is called yā-ye waḥdat/nakara in Persian grammar and seems to be used in ENP less often than in later stages of the language; for example, it is not used in EJP with nouns that are qualified by čē “what (kind of)” or attached to kas “(some)one” (Paul, 2013, §75.b). Its specifying function, i.e., indicating restrictive relative clauses (yā-ye ešārat), seems to be also largely absent from ENP and to have evolved only in later NP (see Lazard, 1966).

Arabic words were borrowed into Persian in great numbers especially during the ENP period. Among the borrowings, nouns ending in -a(t) (with tāʾ marbūṭa) constitute a large and important section. The borrowings were mostly literary and words in -a(t) were borrowed more often in the construct state (-at) than in pausa form (-a, later -e). Perry (1991) has shown in a comprehensive study how in the history of Persian there is a tendency of Arabic borrowings in -at to cross over to the -a/-e-group (e.g., ENP istifādat “use,” modern NP estefāde “id.”); this is not a purely phonetic development, but a complex process involving also semantic, syntactic, and other factors.

Adjectives. While many ENP adjectives cannot be recognized as such formally (e.g., bozorg “big”), there are some suffixes that derive adjectives from nouns, such as -ōmand, -ēn, -ī (ziyānōmand “harmful,” āhanēn “(made) of iron,” arzānī “worthy”). The most widespread suffix denoting general relationship in ENP (and later NP) is -ī (MP -īg), also as a nesba suffixed to places of origin (e.g., EJP rōmī “Byzantine”). Material adjectives like ENP āhanēn “(made) of iron” show an interesting development in the history of New Persian: in modern NP they are used only metaphorically (erāde-ye āhanin “iron will”), while the concrete material meaning has been shifted to the -ī-suffix (dar-e āhani “iron door”; see Paul, 2009).

ENP adjectives usually follow the noun they qualify with an eżāfa, but there are exceptions. such as quantifying, indefinite, or interrogative adjectives, or superlatives, that precede the noun (e.g., bas “many,” čand “several; how many”; bisyār “many,” however, follows the noun; see Paul, 2013, §101). Certain frequent qualifying adjectives such as “good,” “bad,” “big,” and “small” may also precede the noun in ENP (Lazard, 1963, §165 ff.). This shows as an intermediate stage between MP, where the position of adjectives was quite free (Boyce, 1964, p. 44), and modern NP, where the pre-noun position of adjectives is restricted to quantifying and interrogative adjectives, and superlatives, among others.

The ENP comparative adjectives are normally built by the suffix -tar (< MP -tar, or -dar after voiced consonant). In all ENP sources, however, old MP comparative forms without -tar still exist (e.g., beh “better,” bēš “more”), sometimes side-by-side with forms with secondary -tar (EJP meh = mehtar “greater”; Lazard, 1963, §177; Paul, 2013, §94). In modern NP, there are only comparatives with -tar.

Some MP quantifying adjectives have been preserved in ENP that no longer exist in modern NP, e.g., abārī “other” (MP abārīg; Lazard, 1963, §266; Paul, 2013, §100); others like ENP bas “many” have preserved their MP meaning as against modern NP bas “enough.” In EJP but apparently not in ENPA, the demonstrative ēn “this” may take the plural suffix -ān also when it is used as an adjective (EJP pa ēnān rōzigārān “in these times,” Paul, 2013, §102.b; Lazard, 1963, §245).

The ENP numbers show some peculiar forms, some of which occur also in later forms of NP, e.g., čār “four” (= modern coll. NP), duyum/dudum “second,” sayum “third.” An interesting ancient dating formula occurs in EJP Du2.14 pa dah sagd “on the tenth (day),” with sagd probably taken over from Sogdian sγt “passed”and corresponding to MP saḵt “id.” (used in MP dating formula); it is not attested in any other form of ENP (Paul, 2013, §107.b).

Pronouns. The EJP pronouns show a southern dialectal form absent from MP and later NP in 1 pl. ēmā (perhaps ēma) “we” that also occurs in the Tāriḵ-e Sistān (dubbed as “forme douteuse” by Lazard, 1963, §222). ENP (and later NP) ēšān is still used for 3 pl., differently from modern NP, where it has become the polite form of the animate 3 sg. pronoun and where the demonstrative ānhā has replaced it for the 3 pl. pronoun. The EJP variants 1 pl. ēmān and 2 pl. šumān seem to be analogies to ēšān. The ENPA variants 1 pl. amā (or imā?) and 2 pl. šumāh preserve an old anlaut or auslaut (MP amā(h), šumāh). The ENPA forms of 3 sg. way, 1 pl. , and 2 pl. išmā occur also in the northern EJP letters Du1/2, which shows the fundamental disctinction of the language of Du1/2 as against most other EJP texts, and corroborates that ENPA way is a northern form (Lazard, 1963, §224; Paul, 2013, §110). Differently from modern NP, ENP ō(y) was used also for inanimates. Table 3 shows further phonetic variants of the pronouns.

The ENP enclitic pronouns may be attached to any other part of speech (noun, adjective, preposition, verb, conjunction) and serve any possible syntactic function (including, although seldom, that of subject). As shown in Table 4, the ENPA and EJP suffixes differ from each other (for details, see Lazard, 1963, §281 ff., and Paul, 2013, §112 ff.).

The vowels of the EJP suffixes correspond to the MP ones (except 2 sg.), while some of the ENPA forms are harmonized to -a-. The modern NP forms with -e- seem to be a secondary development from ENPA -a-, rather than a continuation of EJP -i- (which Lazard, 1963, §282, proposes). For the plural, the predominating forms are without -i- in EJP and with -i- in ENPA. In EJP, the suffixes are often (sometimes also in ENPA) written separately from the word they follow, giving rise to the question if they should be called “enclitic” in these cases; 1 pl. -mān is actually not reduced in form as compared to the full pronoun ENPA , as would be expected from an enclitic pronoun.

Reflexivity, i.e., co-reference of the subject with another actant (direct/indirect object, complement) of the same sentence, is expressed by tan-i ḵʷēš, ḵʷēštan, ḵʷēš, and ḵʷad in ENP. In MP, tan-i ḵʷēš/ḵʷēštan was used for objects and prepositional complements only, ḵʷēš for possessive complements; and ḵʷad was still an adverb meaning “indeed.” In the course of the development of ENP and later NP, ḵʷad (later > ḵod) gradually took over the functions from the reflexives tan-i ḵʷēš/ḵʷēštan and ḵʷēš, a process during which the semantics of the reflexives began to overlap (e.g., ḵʷēš becoming used also for objects); for an overview, see Paul, 2002.

The following specific features of the ENP reflexives may be noted (see also Paul, 2013, §§116 ff.): (1) they are generally used without pronominal suffixes in ENP, except in two EJP texts (e.g., Kd.9 ḵʷēštan-iš) that anticipate the later NP development; (2) tan-i ḵʷēš occurs often in ENPA, but in EJP only one (doubtful) case is noted; (3) once in EJP, in L18.15, k̄yštn [ḵēštan] occurs without -w-; (4) in translations from Arabic and Hebrew, possessive reflexivity may be expressed with the plain anaphoric pronoun.

The demonstrative pronouns ēn “this” and ān “that” have the EJP plural forms ēnān and ānān. According to Lazard (1963, §245), the plural forms ēnhā/ānhā occur in ENPA besides ēnān/ānān (but he gives no examples). In both varieties of ENP, the singular forms ēn/ān may also be used in a plural sense (Lazard, 1963, §245; Paul, 2013, §120.g). For ēnān used as an adjective, see above.

MacKenzie (1968, p. 253) has shown that the NP noun čēz (later čiz) “thing” goes back to the MP interrogative pronoun čē “what” with the generalizing suffix -(i)z, originally meaning “what-ever.” In various EJP texts, the original form and function of čēz is still visible, mostly in negative or generalizing contexts, e.g., Du1.30 čēz andōh ma ḵʷar “do not suffer any grief”; without -z in Gen 5.8 wa-ēn čē nēst “and this is nothing”; with hiatus-deleting -h- in Ez1 82.7 čēhiz … na bāšad “nothing will be ….” An interrogative meaning seems to be still visible in Ar.e9 wa-abar čēz āmadē “and what indeed would it have come to?”

Verbs. While there had been a decisive break in the verbal system from Old to Middle Persian, with many tenses and moods (like perfect, aorist, optative) having been given up or reduced to a marginal status, the verbal system of MP seems to be similar to that of modern NP at a first look, with ENP taking an intermediate position. All MP and NP verbal forms are based on two stems, present and past, going back to the OIr. present stem and past participle respectively. With the help of auxiliary verbs and verbal particles, a complex system of tenses and moods develops in New Persian, whose details vary greatly from MP to ENP, within ENP, and from ENP to later NP.

In verbal stem formation, there are many dialect variations in ENP; e.g., past stems in -īd are more widely used in certain ENP texts than they are in later NP (e.g., ENP oftīdan “fall” and estīdan “stand” instead of oftādan and estādan; Lazard, 1963, §333; Paul, 2013, §125.e; ENPA āvarīdan “bring” instead of later NP āvardan; Lazard, 1963, §332).

Present tenses. The EJP present personal endings resemble those of MP more than the ENPA endings do, e.g., they share with ZMP 1 sg. -om, and with MP the palatal vowel of the 3 sg. and pl. (Paul, 2013, §142). See Table 5.

The ENP present indicative continues the MP one, which was built from the plain present stem (e.g., EJP nibēs-om “I write”). The adverb hamē (later ) could still complement the ENP present verb with an almost free position in the clause to give it a durative or progressive sense, and the particle/preverb be could precede the ENP present verb to emphasize its syntactic autonomy (Lazard, 1963, §403; Paul, 2013, §144) or, in the case of verbs of motion, to modify its semantics (ravad “he goes” ~ be ravad “he goes away”; Lazard, 1963, §405). Both (ha)mē and be would be grammaticalized as markers of the present indicative or subjunctive only in the course of the 10th-13th centuries; their usage in ENP is rather stylistic, and often unpredictable.

The MP present subjunctive was preserved in ENP in the 3 sg. only, in the hortative ending -ād, e.g., L18.4 āf(a)rīn pa šumā bād “may blessing be upon you” (Paul, 2013, §146; Lazard, 1963, §474). In modern NP, it continues to exist in fossilized expressions like mabād-ā “may it not be” (the suffix -ā is attested in ENPA, Lazard, 1963, §761, but not in EJP). EJP and ENPA share the inherited imperative endings -ø/-ēd (sg./pl.), and also the  singular “polite imperative” in -ē, e.g., EJP āgah-om kunē tō ay sayyidī “(please) inform us, oh my master”; it occurs also in the MP Psalter, but not in other varieties of MP or in later NP (Lazard, 1963, §477; Paul, 2013, §150.a).

The ENPA irrealis particle -ē(δ) (< MP hēd), attached to present and past verbal forms, corresponds to EJP -ē(h) (the -h occurring in 2 texts only), in which it is only attested with past verbs. In later NP, the -ē (> i) was used mainly in conditional clauses of the type agar raftami “if I had gone,” and it is no longer used in modern NP. There is a “hortative” particle EJP (h)ē going back to MMP hēb / ZMP ē(w) that occurs with present indicative verbs, e.g., Du1.9 yazīd ḵudāh ē yār bāšad “the Lord God may be a helper,” and does not seem to occur in ENPA or later forms of NP (Paul, 2013, §146.b, §151).

Past tenses. Although the MP, ENP and NP systems of past tenses are based mainly on a combination of the past stem, or the past participle, with various suffixes and auxiliary verbs, there are substantial changes of the verbal system from MP to ENP and from ENP to modern NP, and the ENP varieties also differ greatly from one another. The MP simple past (āmad hēm “I came,” etc.) is amalgamated into the ENPA simple past āmadam “id.” For transitive verbs, this formal amalgamation went along with a syntactic change of paradigm from the ergative to the accusative system (MP kušt hēm “(he) killed me” > ENP kuštam “I killed”). From late MP ergative constructions, it would seem that the formal amalgamation of the endings preceded the transition from ergative to accusative, although this process need not have happened simultaneously in all regions of Iran (Paul 2008a).

In southern EJP texts, the former passive understanding of the plain past stem is still visible in passages like L3.45 nibišt ēn nāma “this letter has been written (on the date …)” (Paul, 2013, §156), where no agent or subject is mentioned or understood; here the passive understanding of nibišt seems preferable to that as “(he) wrote.” The active understanding of the plain past stem, however, was already prevalent in EJP.

Besides the simple past (āmadam “I came”), there are various compound tenses in ENP, the evolution of some of which has not yet been adequately explained. The past participle in -a (āmada “(having) come”) is an adjectival derivation from the past stem, and the present perfect built with it (āmada am “I have come”) still betrays the past participle’s stative-adjectival origin in ENP in expressions like EJP gufta hest “it is said,” nibišta hest “it is written,” (Paul, 2013, §165.b); or ENPA kušta am “I am dead” (Lazard, 1963, §487). The active understanding of the present perfect, however, was prevalent in all forms of ENP, e.g., EJP Ez1 ḵabar dāda hest “he has given notice.” The present perfect was probably developed to a verbal tense in the northeast and then shifted to other regions of Iran.

The present perfect of the kardastam type occurs often, and almost exclusively, in ENPA texts (and in very few northern EJP texts). Despite its sometimes being called “Perfect II,” it is not secondary to, or derived from, the present perfect of the kardaʾam type. It is rather derived from the past stem (kard), either with the auxiliary verb est- (past partic. estād), or with the inflected copula (h)ast-. The first explanation, advocated by MacKenzie, 1984, would gain support from the EJP verbal form krdystym [kardestēm] (attested in a northern EJP text; Paul, 2013, §167.b), but there are no 3 sg. forms like *kardested that would corroborate this. For the second explanation, one should compare the EJP present perfect, which is built from the simple past with the invariable copula hest (e.g., kerd-om hest “I have done,” kerd hest “he/she has done”; Paul, 2013, §164.c). It is tempting to see ENPA kardastam as a reversed form of *kard-am ast, corresponding to EJP kerd-om hest.

The ENP Past Participle in -a may be combined with auxiliary verbs to form paradigms like rafta buvam, rafta bāšam, rafta hastam, etc., yielding more or less the same sense as raftaʾam “I have gone” (but rafta bām is subjunctive “I may have gone”; Lazard, 1963, §482-83); these forms occur seldom also with the plain past stem (e.g., āmad bāšad; Lazard, 1963, §485). A past tense peculiar to certain northern ENPA dialects and sometimes occurring also in the EJP text Ez1 is the ending in -agē (e.g., raftagē “he has gone”), which is still found in modern Tajik dialects (Shaked, 1986). In some southern EJP texts, a characteristic past tense is formed by the active participle in -ā and the copula, often (but not always) translating a Hebrew participle (e.g., šināsā būdand “they knew”; Paul, 2013, §160).

The verbal particle be and the adverb (ha)mē could be combined freely with all past tenses in ENP, e.g., beraftam, berafta buvam, hamē raftastam, etc., and also with -ē(h), e.g., hamē raftand-ē “they would have gone” (Paul, 2013, §162.c). For their usage and meaning with past verbs, the same is true as has been explained for the present tenses (see above); an example of verbs of “low syntactic autonomy” that would be used rather without than with be, are guftan “say” and dīdan “see” followed by object clauses (Paul, 2013, §159.c). Before the plain verb (e.g., raftam) became the simple past tense in later NP, the prefixed forms (beraftam) were widely used in the same sense for a long time in ENP and later stages of NP.

The MP pluperfect (āmad būd hēm, etc.) is continued identically in a few EJP and ENPA forms that are, if transitive, always used accusatively, e.g., EJP L4.18 nibišt būd “he had written,” Du2.32 nibišt būdom “I had written” (Paul, 2013, §168.a; for its forms see Lazard, 1963, §485). Unlike the northern form from Du2 just quoted, the EJP non-3 sg. forms usually follow the peculiar pattern of the EJP present perfect (see above), replacing invariable hest by būd (L7.5 suʾāl-tān kerd-om būd “I had asked you”; Paul, 2013, §168.b). ENPA had also a past perfect corresponding to the kardastam-type of the present perfect, namely karda būdastam (Lazard, 1963, §484). In both EJP and ENPA, however, the pluperfect paradigm of the type karda būdam that would later hold sway in NP, occurred already, but with the adjectival-stative origin of the past participle in -a being still visible, e.g., nibišta būd “(that) were written …” (Paul, 2013, §169). Table 6 shows the ENP tenses in overview, omitting some seldom or marginal forms.

The arrangement suggests possible, but partly conjectural, historical developments, e.g., EJP raftom hest and ENPA raftastam are not necessarily derived from MP raft ēstēm. Since in principle all tenses may be combined with (ha)mē or be, and these combinations are not yet grammaticalized in ENP, (ha)mē and be are not represented here.

Special verbs and constructions. The MP morphological passive in -īh/-īhist (present/past) is preserved in EJP (as -ih/-ihist) but has vanished from ENPA (e.g., MP tabāhīhistan “be ruined,” EJP tabāhihistan “id.”; Paul, 2013, §171). A periphrastic passive may be formed with āmadan “come” as an auxiliary in EJP and ENPA (e.g., āfarīda āmadand “they were created”; Paul, 2013, §172; Lazard, 1963, §490 ff.). The MP causative in -ēn is preserved as -en in southern, but as -ān in northern EJP, in ENPA, and later NP (e.g., MP āgāhēn- “inform,” EJP (south) āgahen- “id.,” ENPA, NP āgahān- “id.”; Paul, 2013, §173).

There is a great variety of forms of the substantive verb (“to be”) in ENP. The MP copula 1/2/3 sg. hēm/hī/ast (etc.) has been preserved as the regular copula in EJP, with a velarisation of the vowel in EJP 1 sg. hom, and with EJP 3 sg. hest/hast having taken the h- in analogy to the other forms; in ENPA, the copula has been replaced by the enclitic -am/-ī/ast (etc.). Hom/hī (etc.) is also found in the Qorʾān-e Qods from Sistān (Honar), and occurs very seldom also in ENPA (Lazard, 1963, §496), besides other forms of the verb with h-, like 2 sg. hayē (Lazard, 1963, §495). Other forms are built with the stems buw- and bāš-, occurring as variants of the main forms mentioned above, both as full verbs and auxiliaries (bāš- is derived from the MP sg. imperative bāš). Plain b- is also noted, especially in subjunctive forms like bād (see above, and Paul, 2013, §175.a). An interesting form of the substantive verb is attested in EJP est-, corroborating MacKenzie’s (1984) thesis of a MP substantive verb ist- (Ez182.31 saḵʷan pa miṣrim ested “there is a speech about Egypt”; Paul, 2013, §175.c).

The ENP modal verbs ((a)bāyed “it is necessary,” šāyed “it is possible, one can,” dānistan “be able,” ḵʷāstan “want,” etc.) correspond largely to their MP predecessors, with the exception of ENP tawānistan “can, be able,” which is a new formation from the MP (?) adjective tawān “(it is) possible.” Invariable (a)bāyēd and šāyēd are used in impersonal constructions and dānistan, ḵʷāstan and tawānistan are used personally; as in MP, all these verbs are usually complemented by the infinitive (e.g., EJP L3.32 nadānam nibištan-it “I cannot write you”). In ENPA but not in EJP, apocopated infinitives without -an (which are identical to plain past stems) occur besides the full infinitives, possibly representing a dialectal feature of the region of Herat (Lazard, 1963, §500 ff.). The construction of modal verbs with dependent inflected verbs that would replace the infinitive construction in later NP occurs already in EJP (mostly in Ez, e.g. 181.28 bāyad ke be-dānad “he must know”), but it does not seem to occur in the earliest ENPA texts (at least, Lazard, 1963 gives no examples); note, however, that examples do occur in ENPA texts from the mid-11th century onwards (Paul, 2002, p. 26). An account of the development of modal constructions from the 10th-16th centuries is given by A. Lenepveu-Holtz.

The most important ENP participles include the passive one (past participle in -a), primarily used in compound past tenses (see above), and the three active participles that are built from the present stem with -anda, -ā, and -ān. The one with -anda is mostly used adjectivally in MP and ENP (e.g., EJP wāyenda “bird” < “flying”). The participles in -ā (< MP -āg) and -ān may be used adjectivally and in verbal constructions in MP (Skjærvø, pp. 244 ff.); while the adjectival usage of -ā is continued into EJP and ENPA (e.g., rawā “possible”), its verbal function survives only in EJP (often translating a Hebrew participle; see Paul, 2013, §145) but not in ENPA (Lazard, 1963, §508). The participle in -ān, in turn, has been preserved only in ENPA, both adjectivally and in verbal function (Lazard, 1963, §505-06), but does not seem to occur in the EJP texts.

A specific feature of the language of the Qorʾān-e Qods is the frequent usage of old agent nouns in -dār/-tār suffixed to the past stem (and merged with a stem-final -d), translating Arabic active participles, e.g., šīr-dādār “milk-giving” (Ravāqi, p. 211). Only the verb “do” uses the present stem here with the suffix -ār (e.g., ḥesāb-konārān “(those) who count”; Ravāqi, p. 208), showing that forms like dādār were probably understood as dād-ār. These agent nouns are sometimes combined with forms of the verb “to be” to translate full verbs, e.g., budim āzemudārān “we have examined” (Ravāqi, p. 220; see also Filippone, pp. 195-96).

Particles. Prepositions play an important grammatical role in MP and ENP, because they have taken over most case functions after the loss of the nominal inflectional system of Old Persian. Certain ENP prepositions show a transitory stage formally from MP to (later) NP, e.g., ENP (a)bar “upon,” (an)dar “in,” furō(δ) “down to”; cf. MP abar, andar, furōd, as against SNP bar, dar, furō (later foru). The function of certain ENP prepositions is straightforward and quite stable from MP to ENP and beyond, e.g., az “from” or (an)dar “in.” That of others is more complex and difficult to describe, e.g., the ENP preposition abā(z) “back to,” which is derived from the two separate MP prepositions abāz “id.” and abāγ “with” and has retained the corresponding two separate meanings of “back to” and “with,” which cannot always be clearly separated (Lazard, 1963, §680).

The most complex preposition semantically is EJP pa(d) / ENPA ba(d) “by, to, towards, with.” It is derived from MP pad “to, at, in, on,” the lenition of p- > b- in ENPA being possibly due to the influence of the ZMP/MMP preverb be/ba that took on the meaning of “to, towards” in EJP and was amalgamated with pa(d); an influence from the Arabic preposition bi “in, at, on, with” is also feasible. The -d of pa(d)/ba(d) was preserved in certain combinations with pronouns or suffixes in ENP, but not in the same way in both varieties: EJP pad-iš “to/with him, it,” but ENPA bad-ān “to that,” bad-ō “to him” (there is neither EJP pad-ān, nor ENPA bad-iš; later NP bed-ān, bed-u).

Another important ENP preposition is o “to, towards” (< MP ō), occurring exclusively in EJP texts and being written mostly with plain ʾaleph (and therefore perhaps pronounced as a in later EJP?). It may occur “infixed” between a verbal form and a personal suffix, e.g., Lr.6 gwptyndʾš [guftend-a-š] “they said to him” (Paul, 2013, §181). It is possible that the EJP preposition be/ba (< MP preverb be/ba) developed from the combination MP be/ba ō, where the ō was eventually reduced to zero, but there is no certain example of a combination be u occurring in the extant EJP texts to substantiate this claim.

The ENP prepositions pa(d)/ba(d), be/ba, and o (/a) overlap semantically with each other and also with other prepositions like abā(z) “with,” e.g., in marking indirect objects of certain verbs (Lazard, 1963, §§586, 665, 667). There is also an overlap, in the same function, with the ENP suffix -; the semantic proximity of o and - has actually brought about the EJP circumposition o…-rā “for.” For the complex interplay of these markers of direction, indirect object, etc., whose usage differs substantially in the various EJP texts, see Paul, 2003 and Paul, 2013, §180, but there remain several unanswered questions. For -, the function of direct object marker is already attested in ENP, besides the (earlier) one of indirect object marker. Differently from later NP, where - is used as a marker of definite direct objects only, in ENP it is still used as a marker of animate direct objects (Paul 2008b).

The MP subordinating conjunction “that” and the relative particle “who, which” have merged into the ENPA multifunctional particle ki but are preserved in most EJP texts (probably with a short vowel, ku resp. ki/ke). Some tafsirs, however, have merged and into ki already (e.g., Ez1 and Ez2). In a small number of EJP texts, ku may be used as a relative particle, perhaps representing another variety of the confusion that eventually led into the merger of MP and . In the EJP Argument text alone, MP ka is preserved in its MP meaning of “if, when” (Paul, 2013, §185).


Asadi Ṭusi, Loḡat-e Fors, ed. F. Mojtabāʾi and ʿA.-A. Ṣādeqi, Tehran, 1365.

J. P. Asmussen, “Judaeo-Persica II. The Jewish-Persian Law Report from Ahwaz, A. D. 1020,” Acta Orientalia 29, 1965, pp. 49-60.

F. de Blois, Persian Literature. A Bio-Bibliographical Survey, begun by the late C. A. Storey V. Poetry of the Pre-Mongol Period, London, 1997.

Idem, “The Reform of the Zoroastrian Calendar in the Year 375 of Yazdgird,” in ĀTAŠ-E DORUN. The Fire Within. Jamshid Soroush Soroushian Memorial Volume II, Bloomington, 2003, pp. 139-45.

Idem, “Glossary to the New Persian Texts in Manichaean Script,” Dictionary of Manichaean Texts II. Texts from Iraq and Iran (Texts in Syriac, Arabic, Persian and Zoroastrian Middle Persian), ed. F. de Blois and N. Sims-Williams, Turnhout, 2006, pp. 89-114.

Idem, “The Two Zoroastrian Treatises called ʿULAMĀ I ISLĀM,” in Religious Texts in Iranian Languages. Symposium held in Copenhagen May 2002, ed. F. Vahman and C. V. Pedersen, Copenhagen, 2007, pp. 199-210.

M. Boyce, “The Use of Relative Particles in Western Middle Iranian,” in Indo-Iranica. Mélanges présentés à Georg Morgenstierne à l’occasion de son soixante-dixième anniversaire, ed. G. Redard, Wiesbaden, 1964, pp. 28-47.

E. Filippone, “The Language of the Qorʾān-e Qods and its Sistanic Dialectal Background,” in M. Maggi and P. Orsatti, eds., The Persian Language in History, Wiesbaden, 2011, pp. 179-235.

B. Fragner, Die “Persophonie”: Regionalität, Identität und Sprachkontakt in der Geschichte Asiens, Berlin, 1999.

H. A. R. Gibb, “ʿAbd al-Malik b. Marwān,” in EI2 I, 1960, pp. 76-77.

T. E. Gindin, The Early Judaeo-Persian Tafsirs of Ezekiel: Text, Translation, Commentary I. Text, Vienna, 2007.

W. B. Henning, “The Inscriptions of Tang-i Azao,” BSOAS 20, 1957, pp. 335-42.

ʿA.-M. Honar, “Ṣurathā-ye qadimi-e feʿl-e ‘būdan’: ham, hī…,” Majalla-ye zabānšenāsi 9/2, 1992, pp. 21-24.

G. Khan, Early Karaite Grammatical Texts, Atlanta, 2000.

D. Krawulsky, “Zur Wiederbelebung des Begriffes ‘Iran’ zur Ilkhanzeit,” in D. Krawulsky, Mongolen und Ilkhane. Ideologie und Geschichte, Beirut, 1989, pp. 113-30.

G. Lazard, La langue des plus anciens monuments de la prose persane, Paris, 1963.

Idem, Les premiers poètes persans (IXe-Xe siècles). Fragments rassemblés, édités et traduits. I: Introduction et traduction française II. Textes persans, Tehran and Paris, 1964.

Idem, “L’enclitique nominal -i en persan: un ou deux morphèmes?” BSL 61/1, 1966, pp. 249-64.

Idem, “Pahlavi, Parsi, Dari: Les langues de l’Iran d’après Ibn-al-Muqaffaʿ,” in C. E. Bosworth, ed., Iran and Islam, in memory of the late Vladimir Minorsky, Edinburgh, 1971, pp. 361-91.

Idem, “The Rise of the New Persian Language,” in R. N. Frye, ed., The Cambridge History of Iran IV. The Period from the Arab Invasion to the Saljuqs, Cambridge, 1975, pp. 595-632.

Idem, “La source en ‘fârsi’ de Vis-o Râmin,” Trudy Tbilisskogo ordena trudovogo krasnogo znameni gosudarstvennogo universiteta 241 [Mélanges D. Kobidzé], Tbilisi, 1983, pp. 34-39.

Idem, “Lumières nouvelles sur la formation de la langue persan: Une traduction du Coran en persan dialectal et ses affinités avec le judéo-persan,” in S. Shaked and A. Netzer, eds., Irano-Judaica II, Jerusalem, 1990, pp. 184-98.

A. Lenepveu-Hotz, “Des verbs modaux persans et de l’évolution de leur construction entre les Xe et XVIe siècles,” Studia Iranica 39/1, 2010, pp. 79-107.

N. MacKenzie, “Notes on the Transcription of Pahlavi,” BSOAS 30, 1967, pp. 17-29.

Idem, “An Early Jewish-Persian Argument,” BSOAS 31, 1968, pp. 249-69.

Idem, “ḤWYTN: Stance and Existence,” in W. Skalmowski and A. van Tongerloo, eds., Middle Iranian Studies. Proceedings of the International Symposium organized by the Katholieke Universiteit Leiden from the 17th to the 20th of May 1982, Leuven, 1984, pp. 45-56.

F. Meier, “Aussprachefragen des älteren Neupersisch,” Oriens 27-28, 1981, pp. 70-176.

B. O’Kane, The Appearance of Persian on Islamic Art, New York, 2009.

L. Paul, “A Linguist’s Fresh View on ‘Classical Persian’,” in M. Szuppe, ed., Iran. Questions et connaissances. Actes du IVe Congrès Européen des Études Iraniennes, organisé par la Societas Iranologica Europaea, Paris, 6-10 Septembre 1999 II. Périodes médiévale et moderne, Paris 2002, pp. 21-34.

Idem, “Early Judaeo-Persian in a Historical Perspective: The Case of the Prepositions bē, o, pa(d), and the Suffix -rā” in L. Paul, ed., Persian Origins – Early Judaeo-Persian and the Emergence of New Persian .Collected Papers of the Symposium, Göttingen 1999, Wiesbaden 2003, pp. 177-94.

Idem, “Remarks on the Evolution of the New Persian System of Past Tenses,” in S. Shaked and A. Netzer, eds., Irano-Judaica. Studies Relating to Jewish Contacts with Persian Culture throughout the Ages VI, Jerusalem, 2008a, pp. 185-95.

Idem, “Some Remarks on the Persian Suffix -rā as a General and Historical Linguistic Issue,” in S. Karimi, V. Samiian, and D. Stilo, eds., Aspects of Iranian Linguistics, Newcastle 2008b, pp. 329-37.

Idem, “Remarks on the Evolution and Distribution of the New Persian Adjectival Suffixes -īn and ,” in C. Allison, A. Joisten-Pruschke, and A. Wendtland, eds., From Daēnā to Dîn. Religion, Kultur und Sprache in der iranischen Welt. Festschrift für Philip Kreyenbroek zum 60. Geburtstag, Wiesbaden, 2009, pp. 105-10.

Idem, A Grammar of Early Judaeo-Persian, Wiesbaden, forthcoming 2013.

J. R. Perry, Form and Meaning in Persian Vocabulary. The Arabic Feminine Ending. Costa Mesa, Cal., 1991.

Idem, “The Origin and Development of Literary Persian,” in J. T. P. de Bruijn, ed., A History of Persian Literature I. General Introduction to Persian Literature, London, 2009, pp. 43-70.

A. Pisowicz, Origins of the New and Middle Persian Phonological Systems, Krakow, 1985.

M. Qazvini, “Moqaddama-ye qadim-e Šāhnāma,” in idem, Bist maqāla II, Tehran, 1953, pp. 1-64.

ʿA. Ravāqi, ed., Qorʾān-e Qods. Kohantarin bargardān-e Qorʾān ba fārsi, Tehran, 1984.

J. Rypka et al., History of Iranian Literature, Dordrecht, 1968.

ʿA.-A. Ṣādeqi, Takwin-e zabān-e fārsi, Tehran, 1978.

N. Safa-Isfehani, Rivāyat-i Hēmīt-i Ašawahistān. A Study in Zoroastrian Law. Cambridge, Mass., 1980.

S. Shaked, “Teʿuda qaraʾit qeduma be-parsit yehudit (An early Karaite document in Judaeo-Persian),” Tarbiz 41, 1972, pp. 49-58.

Idem, “An Unusual Verbal Form in Early Judaeo-Persian,” in R. Schmitt and P. O. Skjærvø, eds., Studia Grammatica Iranica. Festschrift für Helmut Humbach, Munich, 1986, pp. 393-406.

Idem, “Persian-Arabic Bilingualism in the Cairo Genizah Documents,” in B. Outhwaite and S. Bhayro, eds., From a Sacred Source: Genizah Studies in Honour of Professor Stefan C. Reif, Leiden, 2011, pp. 319-30.

N. Sims-Williams, “Early New Persian in Syriac Script: Two Texts from Turfan,” BSOAS 74/3, 2011, pp. 353-74.

P. O. Skjærvø, “Middle West Iranian,” in G. Windfuhr, ed., The Iranian Languages, London and New York, 2009, pp. 196-278.

W. Sundermann, “Studien zur kirchengeschichtlichen Literatur der iranischen Manichäer II,” AoF 13, 1986, pp. 241-319.

Idem, “Mittelpersisch,” in ed. R. Schmitt, ed., Compendium Linguarum Iranicarum, Wiesbaden, 1989, pp. 138-64.

Idem, “Ein manichäischer Lehrtext in neupersischer Sprache,” in L. Paul, ed., Persian Origins – Early Judaeo-Persian and the Emergence of New Persian. Collected Papers of the Symposium, Göttingen 1999, Wiesbaden, 2003, pp. 243-74.

Tāriḵ-e Sistān, ed. M. Bahār, Tehran, 1935; repr., 2010.

B. Utas, “The Jewish-Persian fragment from Dandān-Uiliq,” Orientalia Suecana 17, 1968 (1969), pp. 123-36.

Zhang Zhan and Shi Guang, “A Newly-discovered Judaeo-Persian Letter,” Journal of the Dunhuang and Turfan Studies 11, 2008, pp. 71-99.

(Ludwig Paul)

Last Updated: November 19, 2013