4.1 The Other Shāhnāmes

Whereas in modern discourse, the Shāhnāme usually refers to Firdawsī’s work, well into the twelfth century and even later, Shāhnāme was merely a common title for many works concerning Persian national history.¹ Firdawsī’s was not the first among these in any sense: chronologically, there were several others that had been written before him and even in terms of prestige we can see in many works of the eleventh and twelfth centuries that other Shāhnāme narratives were preferred to Firdawsī’s.² We should not let Firdawsī’s later fame lead us to believe that he was above his peers from the very beginning, except, perhaps, in literary value.

The term Shāhnāme invites comparison to the Khwadāynāmag, as it is how a user of early Classical Persian would have translated the latter title. This, however, does not prove that the Shāhnāmes would have been translations of the Khwadāynāmag, as the term is also rather natural for any version of the national history.

4.2 The Prose Shāhnāme

With Daqīqī we have for the first time reached a situation where we have extensive parts of the text at our disposal. Now we must go a bit backwards in time and take a look at yet another book that has been lost.

The Prose Shāhnāme has been lost, but there is a scholarly consensus that one of the prefaces of Firdawsī’s Shāhnāme has actually been lifted from this book and attached, with some modifications, to Firdawsī’s epic. This Older Preface³⁰ tells how the Samanid Abū Manṣūr Muḥammad ibn ʿAbd al-Razzāq, the one-time governor of Ṭūs, commissioned his minister Abū Manṣūr al-Maʿmarī to gather together owners of books. He found several such owners and gave them orders to compile a book that would perpetuate his memory. Four such men are mentioned in the text, all bearing Zoroastrian names. It is often, but possibly erroneously, believed that the whole compilation work was done by these four men, as a kind of committee.

The title of the book itself is unclear. The Preface exists in many manuscripts and there are several variants, often confused. Although I have adopted the conventional title, the Prose Shāhnāme, the work’s original title may well have been Kārnāme-ye Shāhān (see Chapter 7.4, §2).³¹

The Preface falls into seven separate parts:³²

1. pious introductory formulae (§1);

2. the story of how the work came to be compiled (§§2–7a);

3. general description of the book; what a book should be like (§§7b–9);

4. the beginning of the book proper³³ with an exposition of the Sasanian kishwar system (§10);

5. chronological questions (§§11–14);³⁴

6. genealogy of Abū Manṣūr and some deeds of his forefathers (§§15, 17–20);

7. inserted within the previous there is a short mention of Firdawsī’s Shāhnāme, which clearly is a later addition (§16).

The work was completed in 346/957. It has been lost, but we can deduce something of its contents from the Preface. In general, the text looks rather similar to Firdawsī’s Shāhnāme and, as will be shown in Chapters 4.4–6, it was most probably Firdawsī’s main source. There are, though, occasional differences, and, as de Blois has pointed out (1992–97: 120–124), it is slightly disturbing that the few pieces of information we can glean from the Preface show several differences as compared to Firdawsī’s text. If these few pieces contain such differences, are we entitled to claim that Firdawsī used this work as his main source? For reasons that will become clear in Chapters 4.4–6, I think we are.

Some significant details of the Preface have been ignored in earlier research and need to be highlighted here. The first is that the Khwadāynāmag is not specifically mentioned among the sources of the Prose Shāhnāme. This does not mean that it could not have been one of them, though, as the sources are only mentioned in a very general way and no titles are given. Probably it was, but this cannot be proven, and there is no saying whether the Middle Persian original text was used or its Arabic translation. Even the Zoroastrian names, supposing they are real names in the first place,³⁵ are not proof that their sources were written in Middle Persian: e.g., Bahrām ibn Mardānshāh (Chapter 3.2.6) obviously wrote in Arabic, despite his name. It is probable that many of their sources were, in fact, in Pahlavi, but there is no compelling reason to assume that all were necessarily so.³⁶

The second is that the Preface makes it abundantly clear that the work was composed on the basis of several different texts, quite obviously belonging to different genres.³⁷ On the basis of comparative evidence, it seems clear that the Sistanian Cycle was used for this Prose Shāhnāme through sources other than the Khwadāynāmag and other more “official” historical texts (Chapter 5.1). Other texts, possibly in a variety of languages, were also used as sources. As we will see later (Chapter 4.6), the Middle Persian Wizārishn ī chatrang, Ayādgār ī Zarērān, Husraw ud rēdag-ē, and Kārnāmag ī Ardashīr were among the sources, as were perhaps andarz collections and lost works which we do not always even know by name. There is no reason to assume that such texts would have been parts of any Pahlavi text titled Khwadāynāmag.³⁸ The whole confusion derives from the firm belief of early scholars that all information on pre-Islamic Iran must come from various recensions of the Khwadāynāmag, a notion for which we have little evidence (see Chapter 6.2).

The existence of early Classical Persian books, discussed above (Chapter 4.1), makes it possible that some of the sources might have been in Classical Persian, and there is no reason to exclude the possibility of Arabic sources, including previous translations from Pahlavi.³⁹ Some support for the latter may be found in the Arabic name forms used by Firdawsī and probably derived from the Prose Shāhnāme, such as Ḍaḥḥāk and Būzurjmihr (instead of Pahlavi Wuzurgmihr → Classical Persian Buzurgmihr).⁴⁰ The reference to Ibn al-Muqaffaʿ and Ḥamza in §11 of the Preface also shows that the compilers at least used Arabic sources for mining information and there is nothing to exclude the idea that they translated parts of them.

As the story of Alexander seems to have been included (see below), and as there seem to be strong reasons to doubt the existence of a Pahlavi Alexander Romance (Chapter 2.3), this would imply that at least this major piece of text was introduced into the Prose Shāhnāme from Arabic sources.

One might speculate on the possibility that the four Zoroastrian names, given as examples (all are preceded by chūn “such as”), have been selected to sound authoritative in regaining the national past⁴¹ and there might have been others, bearing Islamic names, involved in the process.

Unfortunately, the Preface has confusing variants (see Chapter 7.4). The main variant tells that Abū Manṣūr “collected (men from) every town” ([az] har shāristān), but in one reading we have har chahār-e shān “all four of them”, which would seem to restrict the number of the “Committee” to four.⁴² This would be slightly incongruous with the preceding use (four times!) of chūn “such as”. If the four are given as examples, there should surely have been more people than just them. However, the variant har chahār-e shān cannot be excluded, in which case there, indeed, was a committee of four.

It has also been ignored that the Preface actually includes the beginning of the Prose Shāhnāme (§11), and we can see that, unlike Firdawsī’s Shāhnāme, the book began with a geographical exposé of the Sasanian kishwar system and a definition of Ērānshahr.

As the Prose Shāhnāme is usually, and with good reason, considered the main source for Firdawsī, it should be evident that Firdawsī is rather far removed from the Khwadāynāmag that was, at most, just one among the many sources of the Prose Shāhnāme, which in itself was only one (though probably the most important) among Firdawsī’s sources. Equating Firdawsī’s Shāhnāme with the Khwadāynāmag is unwarranted, and deducing the latter’s contents from the former is absurd.

Finally, there remains the question about the mutual relations of the Prose Shāhnāme and the other books studied in Chapter 4. The dates of the authors discussed in 4.1 are far from clear, but it seems possible that all of them wrote after 346/957. Perhaps the earliest, Masʿūdī-ye Marwazī, wrote sometime before 355/966, as al-Maqdisī was able to quote him, and Balʿami, the translator of al-Ṭabarī, wrote in 352/963–4 (Chapter 4.3). Balʿamī is able to quote Abū l-Mu’ayyad al-Balkhī and Bahrām ibn Mardānshāh, which would seem to date them securely before 352/963–4, too, but here we have to be very careful, as the manuscript tradition of Balʿamī is unusually complicated and we know that there are many interpolations in the manuscripts, this particular passage seemingly being one of them.⁴³ Daqīqī seems to be somewhat later (d. around 365/975–6) and Firdawsī (d. 411/1020) and al-Thaʿālibī (who wrote around 412/1022) are clearly later, writing up to half a century after the Prose Shāhnāme.

What is abundantly clear is that there was a huge surge of interest in national history in a very short period in mid- to late tenth-century Iran. This has been interpreted as a growth of national feeling, which may be an exaggeration,⁴⁴ but it is clear that the Iranian past became a particular object of interest in the tenth century.

In addition to the Shāhnāmes, there was an equal surge of interest in texts that found no place in Firdawsī’s Shāhnāme. We have already mentioned Asadī’s Garshāsbnāme and its predecessor (Chapter 4.1.3) and, as will be seen in Chapter 4.7, this was not the only early version of Sistanian epics. The case of Wāmiq o-ʿAdhrā’⁴⁵ shows that the interest went even further than that, the epic probably ultimately going back to Greek sources.

It is not necessary that the Prose Shāhnāme should have been the first literary work in the process – it may well have been preceded by some of the texts studied in Chapter 4.1, or others of which we are not aware – and it is also clear that some almost contemporary works may have been written independently of it, but it seems obvious that such a royal project would have been noted and the compilation may well have become a central source already for the authors in the final years of the 950s and the 960s. Thus, the Prose Shāhnāme was presumably the main source for several, if not all, of the texts discussed in Chapter 4.1.

It also seems to have been the main source for both Firdawsī’s Shāhnāme and al-Thaʿālibī’s Ghurar, the two extant works that we have at our disposal. The Prose Shāhnāme is, to a certain extent, reconstructable through a comparison of the similarities between these two books (see Chapters 4.4–6).

A few features that we could highlight on the basis of such a comparison are that the Prose Shāhnāme evidently told the national history from the Creation to the last Sasanid, Yazdagird iii, as also indicated in the Preface (§6). In contrast to Ibn al-Muqaffaʿ’s translation of the Khwadāynāmag it included stories from the Sistanian Cycle, as well as material that was of interest to the Arabs (Bahrām Gūr’s early history). It also included a version of the Alexander Romance and many good narratives, but probably lacked Firdawsī’s orphan stories.

Firdawsī clearly added to the material when versifying the book, but al-Thaʿālibī may well have abbreviated it, although he did make additions by bringing in quotations from Arabic historians and some Persian texts. In general, the Prose Shāhnāme may well have been about the same size as al-Thaʿālibī’s Ghurar, which in Zotenberg’s edition has 748 pages, containing both the text and the translation, so the text itself covers some 374 pages in a rather large font. As we will later speculate on the size of the Khwadāynāmag, this number should be kept in mind. If a conglomerate of various sources covers no more than this number of pages, the Khwadāynāmag must have been considerably shorter than this, as there is no reason why passages from the Khwadāynāmag (if it was among the Prose Shāhnāme’s sources in the first place) should have been considerably abbreviated or dropped away.

The Prose Shāhnāme is referred to by few authors. An important testimony is given by al-Bīrūnī, who mentions it in his Āthār, p. 133/116//119: “We have found the chronologies (tawārīkh) of this second part (of the Ashkānians) in Kitāb Shāhnāme, made (al-maʿmūl) for Abū Manṣūr ibn ʿAbd al-Razzāq.”⁴⁶

4.3 Balʿamī

Abū ʿAlī Balʿamī⁴⁷ came from an influential family of state officials. His father, Abū l-Faḍl al-Balʿamī (d. 329/940), may have been involved in translating, or having translated, Ibn al-Muqaffaʿ’s Arabic version of Kalīla wa-Dimna into Classical Persian, to be further versified by the poet Rūdakī.⁴⁸ Both father and son were interested in Arabic literature and were patrons to many poets mentioned in al-Thaʿālibī’s Yatīmat al-dahr.

The younger Balʿamī was still alive in 382/992.⁴⁹ His main contribution to Persian letters is his translation, or Persian redaction, of al-Ṭabarī’s historical work, the Ta’rīkh, which is not a separate Shāhnāme, but deserves some discussion here. Far from being a simple translation, Balʿamī’s Tārīkhnāme modified the original and, what is important in the present context, added information on pre-Islamic Iran from other sources, which the author sometimes quotes explicitly. Balʿamī himself openly says (Tārīkh, p. 2) that when something was missing from the original, he added useful pieces of information.

The work was commissioned in 352/963 by Abū l-Ḥasan Fāʾiq, and at about the same time, the great Qurʾānic commentary of the same author, al-Ṭabarī, was also translated into Persian in the same court. The transmission history of Balʿamī’s Tārīkhnāme is extremely tangled, as Peacock (2007) has shown, and the various manuscripts have major differences between each other.⁵⁰

Among the additional sources of Balʿamī, the most important for our purposes is a certain great Shāhnāme that he quotes in Tārīkh, p. 3 (Tārīkhnāme i: 5). Unfortunately, manuscripts give here various readings. In the Tārīkh, the “great Shāhnāme” would seem to refer to Ibn al-Muqaffaʿ’s work (andar Shāhnāme-ye buzurg īdūn gūyad pisar-e Muqaffaʿ ke). However, Tārīkhnāme adds here the name of Ḥamza as the author of the great Shāhnāme (dar Shāhnāme-ye buzurg Ḥamza-ye Iṣfahānī īdūn gūyad ke pisar-e Muqaffaʿ ) and the comment of Ibn al-Muqaffaʿ thus becomes a quotation through Ḥamza’s book.

Both readings are extremely problematic. No source implies that Ḥamza would have written a great book of kings – his Ta⁠ʾrīkh is a rather slim volume. Moreover, the following piece of information (from the expulsion of Adam from Paradise until the “time of our Prophet” there are 6,013 years) is not to be found in Ḥamza’s book, where the period from the Creation until the end of the Persian kings’ rule is given as 4,071 (p. 15) or 4,409 years (p. 25). Immediately after this passage, though, Balʿamī gives Ḥamza’s list (Tārīkh, pp. 4–5; Tārīkhnāme i: 5).

The attribution of this great Shāhnāme to Ibn al-Muqaffaʿ is equally problematic, and his book is nowhere else referred to as “the Great Shāhnāme” nor, as far as I can see, is it ever referred to with the Persian title Shāhnāme. Moreover, the sentence is garbled in the Tārīkhnāme, where the verb that would have pisar-e Muqaffaʿ as its subject never appears.

It seems probable that the great Shāhnāme actually refers to the work of Abū l-Muʾayyad Balkhī, whose Shāhnāme is referred to in Balʿamī, Tārīkhnāme i: 93 (Tārīkh, p. 90) with the very same title Shāhnāme-ye buzurg. As Ḥamza’s list comes immediately after the problematic quotation it would seem probable that the original form is preserved in the Tārīkh, and in Tārīkhnāme’s version the name of Ḥamza has slipped in erroneously. The text should probably be understood so that the great Shāhnāme is here given anonymously and Ibn al-Muqaffaʿ quoted through it (“In the Great Shāhnāme Ibn al-Muqaffaʿ says,” i.e., is quoted as saying).

Unfortunately, Balʿamī rarely gives explicit references to his sources. He does elaborate on al-Ṭabarī’s history in the part concerned with pre-Islamic Iran, and it is quite possible that much of this additional information comes from this great Shāhnāme. The illusion that Balʿamī excerpted a variety of sources seems to be based only on a misunderstanding of the passage Tārīkh, pp. 4–5, where Balʿamī refers to an impressive number of authorities, although he is, in fact, merely lifting this list of names from Ḥamza or some intermediate source.

In Tārīkh, p. 85, Balʿamī quotes “Khudāynāme-ye Bahrām al-Muʾayyad” (Tārīkhnāme i: 87, only has nāme-ye Bahrām al-Muʾayyad), but this is probably a mere corruption of Bahrām al-Mōbad, which probably refers to Bahrām ibn Mardānshāh (see Chapter 3.2.6), rather than to Abū l-Muʾayyad, whose first name we do not know.

4.4 Al-Thaʿālibī

Although slightly later than Firdawsī, it may be advantageous to discuss al-Thaʿālibī’s Ghurar first.

The author has tentatively been identified with the famous Abū Manṣūr al-Thaʿālibī (d. 429/1038), author of, e.g., Yatīmat al-dahr, but the identification is not certain.⁵¹ The question is, however, not of pivotal importance for the present discussion, as we know that the book was written in 412/1022 or a few years earlier in the circles of Ghazna, and the identity of its author is of secondary importance for us.⁵²

The first part of a history of the world written in Arabic by al-Thaʿālibī is usually known by the title Ghurar akhbār mulūk al-Furs wa-siyarihim, but the whole work consisted of four volumes, of which only the first bears this title and is concerned with Persian history. The second covers the life of the Prophet Muḥammad and early Islamic history and has also been preserved, though it still remains unpublished, while the last two volumes have been lost.⁵³

The Ghurar uses two kinds of sources. The main source is a Persian national history, in all probability the Prose Shāhnāme, but the author also had at his disposal some other Persian sources, such as al-Masʿūdī al-Marwazī’s⁵⁴ muzdawija in Persian (p. 10) and a Kitāb Shāhnāme (pp. 263, 457, cf. p. xxiii). The second group of sources are the Arabic historians, who are occasionally used and quoted: al-Ṭabarī, Ibn Khurradādhbih,⁵⁵ Ḥamza al-Iṣfahānī (see p. xix), and al-Maqdisī (p. xxi) are among the Arabic authors mentioned by name.

Omidsalar (2011): 53, takes Kitāb Shāhnāme to refer to the Prose Shāhnāme. This is possible, but it is also possible that the main source is translated without any indication of source and the twice-quoted Shāhnāme is another, secondary source, possibly that of Abū l-Muʾayyad al-Balkhī.⁵⁶ The references to Masʿūdī-ye Marwazī prove beyond doubt that the author did use other works on Persian national history as his sources.

The contents of the book bear close resemblance to Firdawsī’s Shāhnāme, but there are also differences between the two sources (see Chapter 4.6 for some comparisons), not deriving from al-Ṭabarī or other identified historians.⁵⁷ In the preface to his edition of the Ghurar, Zotenberg has convincingly argued that al-Thaʿālibī cannot be dependent on Firdawsī.⁵⁸ These significant differences prove that Firdawsī was not the main source for al-Thaʿālibī. Firdawsī had also completed his work only a few years before the Ghurar was written.⁵⁹

It is also noteworthy that al-Thaʿālibī lacks Firdawsī’s orphan stories, and it is not easy to see why al-Thaʿālibī should have taken just these parts away and would have accidentally returned to an earlier form of Persian national history. A reverse process – al-Thaʿālibī and Firdawsī following the same model and the latter adding originally unrelated stories – presents no problems.

Still, al-Thaʿālibī may well have known Firdawsī’s epic and may occasionally have used it as a secondary source (cf. Chapter 5.2). The later fame of Firdawsī’s Shāhnāme should not lead us to suppose that it must have been an instant success. The voluminous, and hence expensive and hard-to-get work left little mark in the literature of the early eleventh century, so al-Thaʿālibī would be a unique example of a work strongly dependent on Firdawsī merely a few years after its completion.⁶⁰

Firdawsī says that he used “an old book” as his source (see Chapter 4.5). Although his own testimony can by no means be used as binding evidence, it does match the strong evidence provided by the comparison of al-Thaʿālabī’s and Firdawsī’s works with some preserved Pahlavi texts (Chapter 4.6). It seems an obvious solution that both authors used the same book as their main source. The Prose Shāhnāme is usually considered to have been this source, which is supported by the fact that we know in Firdawsī’s case that his main source was in prose.⁶¹

With reference to the Khwadāynāmag, this means that al-Thaʿālibī is in the same situation as Firdawsī: the Ghurar is largely based on a lost book, one of whose many sources was probably the Khwadāynāmag, either in the Middle Persian original or in Arabic translation.

As we have already seen (Chapter 2.4), al-Thaʿālibī takes liberties in quoting from al-Ṭabarī and the same may be supposed to have happened when he translated from his Persian original, the Prose Shāhnāme. See also Chapter 4.6.

With this in mind, we may now proceed to a comparison of the Arabic with the lost original, of which we may take Firdawsī as a representative. As a sample, I will select the episode of Ḍaḥḥāk, which is found in many early Arabic and Persian sources, ranging from a passing quotation to an elaborated narrative. If we consider Firdawsī as representing the original, the most conspicuous change is the abbreviation of the text, but here we have to be very careful as Firdawsī has clearly elaborated his version and invented details which were not in the original source⁶² and has, perhaps, also used other sources, whether oral or written.

The two most conspicuous and clear changes in the text are the use of rhymed prose, not in common use in tenth-century Persian prose, and the slight Islamization of the story. Neither goes deep into the text, but they remain superficial elements. Interestingly, both seem to feature at the beginning and the end of the episode, as if the translator had given more thought to these crucial parts of the story and then translated the rest more quickly. The similar use of rhymed prose may be seen in al-Bundārī’s translation of Firdawsī’s Shāhnāme some two centuries later (Chapter 2.4).

In the episode of Ḍaḥḥāk, Qur’ānic echoes are found at the beginning (p. 16) where the megalomaniac Jamshīd is made to use the words of the Pharaoh in Q 79: 24 (ana rabbukumu l-aʿlā)⁶³ and at the end (pp. 33–34), where the realization of the dream of Ḍaḥḥāk provides an opportunity to allude to the Surah of Joseph (Q 12: 100 twice; also Q 56:1 and Q 18: 64 are alluded to on these pages). These almost exhaust the Qur’ānic allusions in the story, and the middle section (pp. 17–32) contains very little Qur’ānic or religious vocabulary.

The same holds for the use of rhymed prose. The episode begins with a cluster of rhymed prose and a play on the sequence amm(a) (p. 16):

lammā tamma amru Jam wa-jammat ʿindahū amwālu l-dunyā wa-ʿaẓuma shānuh wa-ʿalā mulkuhū wa-sulṭānuh wa-mtadda zamānuh (…) lam yalbath an khabā qabasuh wa-kabā farasuh (…)

After the first page of the episode, rhymed prose more or less disappears, only to return in a few passages towards the end.

Al-Thaʿālibī freely inserted passages from al-Ṭabarī, which he usually marks as such, presumably because of the prestige al-Ṭabarī already enjoyed at his time, although sometimes he quotes him without acknowledgement (e.g., pp. 17–18, and Abū Tammām’s verses on p. 35 derive from al-Ṭabarī, Ta⁠ʾrīkh i: 201//ii: 2). Al-Ṭabarī, however, is only a secondary source for al-Thaʿālibī, as can be seen from the order of the material: al-Thaʿālibī uses al-Ṭabarī without any order, even inserting (p. 24) a quotation from al-Ṭabarī i: 174//i: 344 into this episode, which otherwise relies on al-Ṭabarī i: 201–210//ii: 1–9. The structure of the Ghurar comes from the Prose Shāhnāme.

Minor embellishments aside, the similarity of al-Thaʿālibī’s text with Firdawsī’s epic shows that both sources were in their main lines following their common source.

4.5 Firdawsī

The similarities between al-Thaʿālibī’s Ghurar and Firdawsī’s Shāhnāme show beyond the slightest doubt that the two works are interrelated. As has been shown in the previous chapter, there are strong reasons to assume that al-Thaʿālibī is not translating Firdawsī (nor, of course, the other way round), but the two must go back to a common source. The close resemblance of material and its near identical order exclude the possibility that both were compiling their works from the same selection of texts.⁶⁴

This is supported by Firdawsī’s own testimony. Both in the Preface and later, he refers to an old book as his source.⁶⁵ Although on its own such testimony would be far from conclusive, it gains credence from the fact that all other evidence points in the same direction. Given the dearth of evidence, it is impossible to prove that this old book, the common source of al-Thaʿālibī and Firdawsī, is the Prose Shāhnāme and not, e.g., the Shāhnāme of Abū l-Muʾayyad al-Balkhī (Chapter 4.1.3). However, the royal prestige of the Prose Shāhnāme makes it a good candidate, and the insertion of the Prose Shāhnāme’s preface into some manuscripts of Firdawsī’s Shāhnāme might also be induced in favour of this. If the common source would turn out to be, e.g., Abū l-Muʾayyad’s Shāhnāme, this would not change the picture in any significant way.

The comparison between some preserved Pahlavi texts and their reproductions in the works of Firdawsī and al-Thaʿālibī gives us a possibility to see how these two authors handled their original source (see Chapter 4.6). As the comparisons show that in different cases one or the other author comes closer to the original, it is not possible to assume that the author(s) of the Prose Shāhnāme, their common source, would be the one(s) who had modified the original texts: in that case, neither Firdawsī nor al-Thaʿālibī could come closer to the Pahlavi originals.

As already mentioned, Firdawsī does have stories that are not found in al-Thaʿālibī’s Ghurar. In some cases, it is possible that al-Thaʿālibī has abbreviated the work by dropping stories that are not relevant to the main story line, but as Firdawsī’s orphan stories tend to be missing also in other early narratives of Persian national history (al-Ṭabarī, al-Masʿūdī, etc.) it is rather clear that it is Firdawsī who added these stories to his epic from other sources.

It is not evident whether Firdawsī’s additional sources were oral or written. As we have seen, there was an extensive literature in Arabic and Persian on Persian national history (Chapters 2.2.1, 3.6, and 4.1) and some of these texts may well have been available to Firdawsī. On the other hand, it is also possible that he had heard some of the epic tales in oral performances, whether narrated in prose or sung in verse, as there is no reason to suggest that the oral tradition had died out.⁶⁶

The continued existence of an oral tradition is made probable by the extensive nāme literature (Chapter 4.7). The contents of nāmes are sparsely documented in Arabic sources and we have little evidence to claim that they would ever have been translated into Arabic. It is tenuous to claim that each and every nāme is based on a lost Middle Persian text, even though some of them may, in fact, be.⁶⁷

On the other hand, we have occasional information on historical books having been compiled from oral sources. Such, e.g., is the case of the Bāwandnāme, which, according to the testimony of Tārīkh-e Ṭabaristān, p. 4, was “collected (at the end of the eleventh century) in verse (…) from the lies of the country folks and the mouths of the common people.”⁶⁸ Here, and presumably in other similar cases, we have a literary composition based on oral narratives rather than a transcript of oral poetry as such.

Some scholars (especially Dick Davis 1996 and Olga Davidson 1998 and 2006) have maintained that not only did Firdawsī gather his material from oral sources but Firdawsī’s own work at first lived on in oral tradition, which would explain the wide variation of the preserved manuscripts.

This theory remains supported by some scholars, even though it is improbable for various reasons which have been pointed out, among others, by de Blois (1992–97): 53–58, and Omidsalar (1998) and (2011): 11–31. The proponents of the oral theory have mainly by-passed the very valid arguments of their critics. As the theory cannot be supported by any evidence, it will only be discussed briefly here. Besides the strong evidence for a literary source, discussed above, there are also other arguments against the theory of the centrality of oral sources. Practically all contemporary and slightly later sources show that there were written texts that could have been used by Firdawsī. Had his sources been oral, we should assume that al-Thaʿālibī’s Ghurar is a translation of Firdawsī’s Shāhnāme, which is a problematic claim, as already discussed. Otherwise, we should claim that, by a curious coincidence, al-Thaʿālibī happened to come across the very same oral performances as Firdawsī and, by an even more curious coincidence, happened to organize them in an identical order. There is also negative evidence: no source claims that either Firdawsī’s sources or his work itself were solely or mostly oral. Finally, the comparison of Firdawsī’s Shāhnāme, al-Thaʿālibī’s Ghurar, and some Pahlavi texts (Chapter 4.6) proves beyond the slightest doubt that the existing texts of both al-Thaʿālibī and Firdawsī go back to literary sources, at least in these cases.

In addition to being extremely improbable, the oral source theory is also irrelevant from the point of view of the Khwadāynāmag. Whether and to what extent Firdawsī may have used oral sources, it is certain that the Middle Persian Khwadāynāmag was not transmitted orally, or at least we do not have any evidence for such an improbable theory.

On the other hand, some scholars have recently claimed that Firdawsī faithfully used only one source, the Prose Shāhnāme. The most prominent among these is Mahmoud Omidsalar, who claims in a recent book (2011: 26) that “[t]he notion that Ferdowsi could fake a whole book (…) is at best unrealistic” and “Ferdowsi could not have gotten away with fabricating his source because his contemporaries knew their sources, and he would have been unable to fool them”. He puts this even more clearly on p. 33: “I will argue that (…) it is not possible to believe that Ferdowsi incorporated any stories from other sources – oral or written – into his narrative, and [I will argue] that the prose Shāhnāmeh served as the exclusive source material for his epic.”

This presupposes that Firdawsī and his contemporaries shared our ideas of textual fidelity to the original sources, which is hardly true. It also ignores the fact that later sources (even slightly later ones) seem to be quite content with making similar formulaic claims of using an old book as their source, and these claims are so close to those of Firdawsī that they cannot be taken as anything but topoi. Even though it is not possible to retroject the attitudes of the authors of the late eleventh century and later back to Firdawsī, this at least shows that they had no ideal of absolute fidelity. Firdawsī was not faking or lying, but using a familiar and acceptable literary topos of finding (implicitly all) his stories in an old book, while in fact he used a variety of auxiliary sources as well. In addition, of course, it should be remembered that Firdawsī also refers to old dihqāns from whom he had heard stories: if these are accepted as poetic liberties, why should Firdawsī suddenly be taken literally when he implies (not even says explicitly!) that one old book was the source of all his stories?

Besides references to Firdawsī’s moral character, Omidsalar’s argumentation is to a large extent based on the claim that Firdawsī’s Shāhnāme makes coherent reading and, hence, the stories must stem from one and only one source. In some cases, Omidsalar’s argument for an absolute coherence of the text is forced: proving that a narrative is not completely out of place in its context hardly proves that the elements of the narrative must derive from only one source. It is, moreover, generally admitted that Firdawsī was a good author, and a good author will undoubtedly be able to use several sources without becoming incoherent. In fact, Omidsalar’s attitude would push back the credit of composing such a magnificent epic merely to Firdawsī’s predecessors.⁶⁹

Omidasalar also claims (2011: 27) that Firdawsī “worked in a highly refined literary environment, which considered oral tradition vulgar and uncouth” and goes on to cite Bayhaqī’s negative comment on “impossible lies” (Tārīkh, p. 905).⁷⁰ This, however, does not refer to the orality of such stories, and there are many passages where the khurāfāt are condemned, irrespective of their mode of transmission.⁷¹ The main target of such criticism was the fabulous and supernatural content of the stories, not their mode of transmission, and there is no reason why Firdawsī could not have used historical materials from oral sources besides his main source, the Prose Shāhnāme.

There still remains the question of the possible Arabic sources of Firdawsī. Since Theodor Nöldeke’s groundbreaking study (1879: xxiii), the more or less universal opinion has been that Firdawsī did not use Arabic sources.⁷² Jackson Bonner (2011): 65, has recently argued against this, but his evidence is inconclusive: the fact that Firdawsī uses Arabicized forms of Syriac words does not prove that he was using Arabic sources, as he was using Persian sources which themselves were (at least partly) based on Arabic sources. One should make a clear distinction between two separate things:

1. Firdawsī may, or may not, have used sources in Arabic;

2. Firdawsī certainly used earlier Classical Persian sources, which partly went back to Arabic ones.

There does not seem to be any clear evidence that Firdawsī knew Arabic, and as his Persian source (the Prose Shāhnāme) covers most of his stories and the remaining orphan stories are not well documented in Arabic, it seems advisable to consider his sources as having been predominantly, if not exclusively, in Persian. He may, of course, have known Arabic and could also have received some relevant information from Arabic books, but there is no concrete evidence that he did so. There is also no evidence that Firdawsī would have known Pahlavi, and it seems rather improbable that, as a Muslim of the late tenth century, he would have known the old language and script. At least, again, there is no sign that he did.

As to his possible relation to the Khwadāynāmag, Firdawsī is separated from this Pahlavi book by many steps. There is no evidence that he would have been using the Khwadāynāmag in the original language or in Arabic translation, so the only way the Khwadāynāmag could have influenced Firdawsī’s epic is through the following chain: Firdawsī used as his main, though not sole, source the Prose Shāhnāme, which in its turn used a variety of sources in a variety of languages. Some of these sources were in Pahlavi and one of them may have been, and probably was, the Khwadāynāmag.

In practice, this means that Firdawsī’s Shāhnāme is a poor representative of the Khwadāynāmag and there is no reason to assume that the contents of the Shāhnāme could give us any clear idea of the contents of the Khwadāynāmag. Thus, e.g., we can see that it was the Pahlavi Kārnāmag ī Ardashīr through its Classical Persian translation in the Prose Shāhnāme, not the Khwadāynāmag in any language, that provided Firdawsī with the story of the founding of the Sasanian dynasty (cf. Chapter 4.6).

Although undoubtedly one of the most valuable jewels of Persian literature, for Khwadāynāmag studies Firdawsī’s epic has been a cause of much confusion. In a sense, there may well be a line from the Khwadāynāmag to Firdawsī, but this is buried under several influxes of other materials and there is no direct contact between Firdawsī and the Middle Persian text written almost half a millennium before him.

Although overwhelmingly important in later Persian literature, especially from the twelfth century onwards, Firdawsī’s epic did not take the other Shāhnāmes, or material deriving from them, out of the market. In fact, even late historians, such as Mīrkhwand, still base their narrative on sources which often tell the story in a way contradictory to Firdawsī’s Shāhnāme and quote other authors, including the earlier Shāhnāmes, as authoritative and sometimes implicitly more authoritative than Firdawsī.

Firdawsī added to his main storyline new episodes, which cannot be located in any earlier version of Persian national history. These episodes were probably not invented by Firdawsī but were only integrated by him into an existing storyline. Later epics, such as the Garshāsbnāme and the Farāmarznāmes, continued the process of incorporating more material into national history.

In addition, Firdawsī versified the stories he received from the Prose Shāhnāme and other sources. The versification of earlier prose texts was a common practice in the tenth and eleventh centuries (and later). Kalīla wa-Dimna was first translated into Persian prose and then versified by Rūdakī (cf. Chapter 4.3).⁷³ Azraqī boasted of his ability to improve on the prose Sindbādnāme.⁷⁴ Daqīqī and Masʿūdī-ye Marwazī preceded Firdawsī in this versification, and Asadī Ṭūsī came soon after and versified a prose Garshāsbnāme (see Chapter 4.1.3), to select but a few examples both before and after Firdawsī.

All these show how several epics and other books were versified based on earlier prose texts, most of which have later disappeared, just like the Prose Shāhnāme. It seems that it is even typical that once there were more modern versified versions, the older ones were reduced in a few centuries to insignificance and later disappeared, at the latest in the Mongol disturbances.⁷⁵

4.6 Firdawsī, al-Thaʿālibī, and Pahlavi Texts

There are some Pahlavi texts that have been preserved and are duplicated in Firdawsī’s Shāhnāme and al-Thaʿālibī’s Ghurar, thus providing us with a possibility to study how each of the authors has changed the story which they took from the Prose Shāhnāme. One of these is the story of chess and backgammon, Wizārishn ī chatrang ud nihishn ī nēw-Ardashīr.⁷⁶

The story is told in the Shāhnāme vii: 314–319 in a much expanded version – as the manuscript tradition of the Wizārishn is unusually good for a Pahlavi text, there is no reason to speculate on the existence of a lost longer version in Pahlavi. The text of the Wizārishn is also very coherent and shows no signs of omissions.

The two texts have next to no identical passages, and several significant details, including the name of the Indian King, Dēbshalm, and his Vizier, Takhtarītos (Tātarītos),⁷⁷ are missing from Firdawsī’s text. Likewise, there are significant changes, such as the time taken by King Khusraw’s Vizier, Buzurgmihr, to solve the riddle sent by the Indian King. In the Pahlavi version, Buzurgmihr (Wuzurgmihr) waits for the expiration of the three-day deadline to show that no one else is able to solve the riddle (§§4–6). After that he seems to withdraw for one night, as the text continues by telling how the next day he returned to solve it (§9). In Firdawsī, Shāhnāme vii: 307, v. 2699, the deadline falls in seven days, after which Buzurgmihr (Būzurjmihr) takes a day and a night to solve the riddle (vii: 308, vv. 2712–2714).

The story is rarely found in Arabic and Persian literature. Mujmal, p. 60/75, seems to be the only place where the name of the Indian King, here Dābshalīm, is given, though his emissary remains anonymous.⁷⁸ Otherwise, the story is told there with minimal details, covering only a few lines. The paucity of references makes it improbable that the story would have been told in Ibn al-Muqaffaʿ’s translation of the Khwadāynāmag, and there is no reason to assume that this text, which we know in an independent version, would at any stage have been made part of the Khwadāynāmag or its translation.

The story is also briefly told in al-Thaʿālibī, Ghurar, pp. 622–624,⁷⁹ again with no Indian names. In this version, Khusraw immediately understands that only Buzurgmihr will be able to solve the riddle, which he does, with no deadline indicated, as if this happened straight away. Al-Thaʿālibī may have abbreviated the story, whereas Firdawsī certainly expanded it. Their basic agreement implies that the text was found in their source, the Prose Shāhnāme.

The second case where we are able to compare Firdawsī’s Shāhnāme with Pahlavi texts is Ayādgār ī Zarērān (Chapter 1.2.1), which falls within the section taken from Daqīqī.

The story of Zarēr is found in al-Thaʿālibī, Ghurar, pp. 262–276, prefaced by an account (taken from al-Ṭabarī and other Arabic sources, pp. 256–262) which explains the origin of Zarathustra and his religion and gives a brief summary of the Zoroastrian religion. It is also found in Firdawsī’s Shāhnāme v: 85–149 (written by Daqīqī; again prefaced by a narration of how Zarathustra’s religion began, v: 76–85), yet again a sign that both authors are using the same source, where this addition had already been made. Firdawsī’s version is substantially longer and more detailed and even al-Thaʿālibī’s version is slightly longer than the Pahlavi original, which covers only 17 pages in the modern edition. The story is also briefly told in al-Ṭabari, Ta⁠ʾrīkh i: 676–677//iv: 71–73.

Al-Thaʿālibī also explicitly refers in this passage to ṣāḥib Kitāb Shāhnāme (p. 263), which does not refer to Firdawsī’s work but either to the Prose Shāhnāme or one of the other early Shāhnāmes (see Chapter 4.1). There are significant differences between al-Thaʿālibī and the Ayādgār. E.g., in the latter the events start with Arjāsp having heard about the conversion of Wishtāsp (§2 ud pas Arjāsp ī Khyōnān-khwadāy azd mad kū), whereas in al-Thaʿalibī, Ghurar, p. 263, they start with Bishtāsf writing a letter to Arjāsf and calling him to Zoroastrianism. One has to note, though, that this first letter is only referred to and not quoted, which may have to be interpreted as only al-Thaʿālibī’s elaboration of what the Ayādgār says. In Shāhnāme v: 85, it is a demon who informs Arjāsp about what is happening, suggesting that the latter should refuse to pay tribute to the Iranians now that they have converted.

The gist of the first letter that is quoted (Ayādgār §§10–12; Ghurar, pp. 263–264) is similar in all three sources, although there are few identical passages. One of these, though, is highly significant: in the original, Arjāsp threatens Wishtāsp that: “we will come upon you, eating (i.e., having our horses eat) the fresh (grass) and burning the dry and taking as captives from your land (all) four-legged and two-legged (beings) and put you in heavy chains and misfortune.”

In the Ghurar, p. 264, the same passage reads: “They (my armies) will eat up the fresh and burn the dry and kill (your) men and take the women as prisoners.” This, more or less, reads like a direct translation of the original. The Shāhnāme has no clear parallel to this, although the letter ends with threats to burn the palace and the land of Gushtāsp and to kill the old men not suitable for slavery while enslaving women and children (vv. 165–169), which reads like a free poetic version of the original.

The Ghurar and the Shāhnāme share some details which the original lacks. Hence, Ghurar, p. 263, lets Arjāsf call Zarathustra “a liar who claims that he came from Heaven” in his letter, which coincides with Firdawsī’s (Daqīqī’s) “he claims to have come from Heaven” (v: 86, v. 104). This, though, is not given in this letter, but occurs in Arjāsp’s words to his army before the conflict commences at the beginning of the story. Thus, it would seem that the Prose Shāhnāme had already added this detail, and the two sources dependent on it reflect the addition.

Whereas in the Ghurar, p. 264, the letter is delivered by one anonymous messenger, the Shāhnāme names two messengers (Bīdrafsh and Nāmkhwāst, v: 88, vv. 124–125), as does the Ayādgār (Wīdrafsh and Nāmkhwāst, §6). Likewise, the letter in the Ghurar is full of insults, whereas both the Ayādgār (§§10–11) and the Shāhnāme start in extremely polite terms (v: 88–89). As to the length of the letter, it takes only a few lines in the Ayādgār and the Ghurar, whereas the Shāhnāme’s version is very long (v: 88–92, vv. 133–171).⁸⁰

Similar results arise from other parts of the texts. It seems hard to avoid the following conclusions:

Both the Ghurar and the Shāhnāme go back to a source (the Prose Shāhnāme) which resembles the Ayādgār and contained a translation of the Ayādgār set in the frame of Persian national history. Al-Thaʿālibī has abbreviated this source by, e.g., dropping the names of the messengers, whereas Firdawsī/Daqīqī has freely rewritten the story, elaborating it with details. In passages where al-Thaʿālibī rather closely follows the Ayādgār, we may with good reason assume that his source, the Prose Shāhnāme, was also close to the original. This, furthermore, means that it is Firdawsī/Daqīqī who elaborated the text – had the Prose Shāhnāme contained a long and elaborate narrative, it is hard to see how al-Thaʿālibī could have come by a version which is so close to the Pahlavi original.

The battle scenes with all their details in the Ayādgār resemble those of the Shāhnāme in general, which implies that the Shāhnāme’s ways of describing a battle are basically taken from the earlier tradition, where they already had consolidated in a rather fixed form: kings following the battle from the side, single combats, the promise of a daughter of the king and a high position to the hero who takes it upon himself to become involved in these combats, and the heroic exaggeration of the scene, where single heroes kill myriads of enemies (§§55–61, 70, etc.). There are significant details that are echoed in the Shāhnāme’s narrative, such as arrows that are specifically blessed in order to kill an enemy (§§74, 101; 92, 106).⁸¹

As in the case of the Wizārishn, it seems hard to find exact parallels between the Ayādgār and the Shāhnāme, which would make it very tenuous to claim that Daqīqī, or Firdawsī, used the original Pahlavi text. This would also leave unexplained those cases where al-Thaʿālibī and Firdawsī agree with each other but differ from the Pahlavi text.

It also seems that the Ghurar is not an abbreviation of the Prose Shāhnāme to any great extent, and the differences between the Ghurar and the Shāhnāme should, prima facie, be taken as elaborations by Firdawsī (or Daqīqī in this case).

Finally, al-Ṭabarī’s version contains differences vis-à-vis all the other three texts, which implies that the source he had at his disposal may have slightly differed from the Ayādgār as we now have it. Some of the differences may be mere errors or radical abbreviations. Thus, e.g., Nastūr’s (Bastwar’s) part in the battle is glossed over⁸² and it is Isfandiyār who kills Bīdrafsh,⁸³ while in the Ayādgār §§99–106 and the Ghurar, pp. 274–275, it is the young son of Zarēr, Bastwar, who does this deed.

Interestingly enough, the Shāhnāme in a way combines the two. First, Bastūr is sent by Gushtāsb to combat Bīdrafsh, and he joins the battle (v: 141, vv. 702–712). Soon after, Arjāsp notices the new hero and sends Bīdrafsh to fight him, and they meet in single combat (v: 141–142, vv. 714–723). Thus far, the Shāhnāme seems to be following the version of the Ayādgār, but al-Ṭabarī’s version is also there. Inserted between vv. 702–712, 714–723, which highlight Bastūr, there is a single verse, v. 713, which reintroduces Isfandiyār (“on the other side, the hero Isfandiyār was killing countless enemies”) and he suddenly returns on the scene in v. 724, where he is told about the on-going battle between Bastūr and Bīdrafsh. Isfandiyār intrudes in the combat and kills Bīdrafsh, dispossessing him not only of his head but also of the loot he had taken from Zarēr (v: 142–143, vv. 725–733). Bastūr is suddenly dropped from the narration and the reader does not hear of him until after the battle, v. 740.

The first impression would be that Firdawsī/Daqīqī has slightly rewritten the original story in order to make Isfandiyār its main hero. This may well be so: al-Thaʿālibī’s evidence would seem to show that in the Prose Shāhnāme, it was Bastūr who killed Bīdrafsh, as in the original Ayādgār. However, al-Ṭabarī’s extremely concise narrative causes a problem, as it, too, seems to imply that it was Isfandiyār who killed Bīdrafsh. The passage, however, is ambiguous and deserves to be quoted in full in Arabic (Ta⁠ʾrīkh i: 677):

wa-shtadda dhālika ʿalā Bishtāsb fa-aḥsana l-ghināʾ ʿanhu ibnuhu Isfandiyār wa-qtl Bīdrafsh mubārazatan.

The most unforced translation of this would be:

This grieved Bishtāsb. His son Isfandiyār lamented him (Zarēr) in a beautiful song and (then he) killed (qatala) Bīdrafsh in single combat.

Changing the verb into the passive voice (qutila) would make the end congruous with the Ayādgār and the Ghurar, “and Bīdrafsh was killed in single combat (by someone)”. Though not perhaps the most obvious choice, there is nothing to prevent this reading, as the text of al-Ṭabarī is very cursory, and he only lists the main events of the battle, telling the whole story of the Ayādgār in a mere 16 lines.

Leaving al-Ṭabarī aside for a while, the slaying of Bīdrafsh would clearly seem to indicate that Firdawsī/Daqīqī added Isfandiyār as the main hero of the battle, as he also otherwise has a remarkably important role in the events at Gushtāsp’s time. The original narrative, where Bastwar and Spandiyād are both represented as heroes,⁸⁴ is first interrupted by a single verse, v. 713, to reintroduce Isfandiyār and then the end of the narrative is cut just before Bastūr would slay Bīdrafsh and the deed is left for Isfandiyār to accomplish. The Ghurar would seem to confirm that the Prose Shāhnāme did not as yet have this crucial change, which shows how Firdawsī/Daqīqī worked with this episode, making a major change in the story to tie this episode up with the general story line, where Isfandiyār is the central figure until he meets Rustam.

Coming back to al-Ṭabarī, it seems that instead of selecting the more natural reading (Isfandiyār killed Bīdrafsh) we should opt for the other, equally possible one (Bīdrafsh was killed [by someone]). It is slightly difficult to see how al-Ṭabarī and Firdawsī/Daqīqī could have separately made the same change – al-Ṭabarī does not make Isfandiyār a central hero in his narrative – and speculating on an early version which differed from the present Ayādgār is perhaps too complicated.

Al-Ṭabarī mentions Isfandiyār’s dirge for Zarēr (Zarīn), though he does not quote it. Although al-Ṭabarī is extremely concise here, the dirge is mentioned just before the passage where Bīdhrafsh is killed in single combat, so the lamentation seems to take place in the middle of the battle.

This probably echoes Ayādgār §§84–86, where, also in the middle of the battle, Bastwar laments the loss of his father. In al-Ṭabarī, the crucial sentence (wa-shtadda dhālika ʿalā Bishtāsb fa-aḥsana l-ghināʾ ʿanhu ibnuhu Isfandiyār) is open to two readings: “This grieved Bishtāsb. His [Bishtāsb’s or Zarēr’s] son Isfandiyār lamented him (Zarēr) in a beautiful song”. In his translation, Perlmann has opted for the latter, but as Isfandiyār was Bishtāsb’s son, not Zarēr’s, this is clearly wrong.

In Firdawsī’s Shāhnāme, when he finds the body of his father, Bastūr bursts into a short speech, where there are elements of lamentation (v: 139, vv. 681–687). One does, however, find a passage where it is Isfandiyār who, after the battle, finds the body of Zarēr and laments his death (v: 148–149, lament in vv. 782–784).⁸⁵ The lament of Isfandiyār, with its numerous vocatives in -ā, is, in fact, closer in tenor to the original.

In the Ghurar, no dirge is mentioned, although the description of the finding of the body of Zarēr and other nobles after the battle, p. 276, would perfectly serve as the locus where it could have been inserted. Al-Thaʿālibī often abbreviates the scenes, so this may be explained as his abbreviation. This might again indicate a certain duplication, by which Firdawsī/Daqīqī has attributed to Isfandiyār things Bastūr had done, in order to put Isfandiyār more into the focus.

Like the Ayādgār, Kārnāmag ī Ardashīr finds parallels in both Firdawsī’s Shāhnāme (vi: 138–214) and al-Thaʿālibī’s Ghurar (pp. 473–480). After setting the scene in general, all three works begin with dreams. In the Kārnāmag, Pābag has three dreams over three nights. First, he dreams of brilliant Sun shining from the head of Sāsān (i: 8). Second, he sees Sāsān on a white elephant (i: 9), and on the third night he dreams how three Fires shine out of his house (i: 10). Firdawsi, Shāhnāme vi: 140, gives the last two dreams, whereas al-Thaʿālibī, Ghurar, p. 474, mentions the first dream and then attributes another dream to Sāsān, who has seen a ray of light (shuʿāʿ) coming out of him and filling the horizons with light, which either is a duplicate of the first dream or a version of the third. Whichever it is, the dream scene implies that the common source of Firdawsī and al-Thaʿālibī had all the three dreams, and both authors excerpted the passage in slightly different ways and also modified it freely. The dreams are very concise in al-Thaʿālibī’s version, but slightly longer in Firdawsī than in the Kārnāmag.

Al-Thaʿālibī also drops the scene of dream interpretation (Kārnāmag i:12–13, Shāhnāme vi: 141). Throughout the story, Firdawsī adds verses that are descriptive or show the emotions of the characters: thus, he, e.g., describes the coming of the chief shepherd Sāsān to Bābak’s court in v. 117 (vi: 141): “The shepherd came to him with a gilīm, his woollen garment full of snow, his heart full of fear.” This has no parallel in al-Thaʿālibī and only a very general one in the Kārnāmag, where, i: 18, Sāsān is given a princely garment to wear, but nothing is said about his earlier garments.

Occasionally, al-Thaʿālibī, too, adds details that are not found in the other two sources. Thus, Ghurar, p. 474, relates that Sāsān died soon after the birth of Ardashīr, who was (obviously for this reason) linked in genealogy to Bābak. The following letter from Ardawān to Bābak is only mentioned by al-Thaʿālibī (Ghurar, p. 475) whereas the version in the Kārnāmag ii: 6–7, and the Shāhnāme, vv. 146–150 (vi: 143) are reasonably similar to each other, showing that it was most probably al-Thaʿālibī who decided to summarize the letter’s contents in a few words.

Al-Thaʿālibī abbreviates the story considerably by cutting off episodes that are not relevant to the main story line. Thus, he narrates the escape of Ardashīr from Ardawān in a very concise form (Ghurar, pp. 476–478; cf. Kārnāmag iii–iv) and ends the story with the death of Ardawān and Ardashīr’s ascent to the throne (Ghurar, pp. 480–481; Kārnāmag v), before giving some scattered sayings by Ardashīr (Ghurar, pp. 482–484), which have no counterpart in the Kārnāmag. Ardashīr’s other deeds and battles, told in the Kārnāmag, are not brought into the Ghurar.

Firdawsī tells all this more extensively and continues following the story where al-Thaʿālibī cuts off. Thus, he tells of Ardashīr’s battles against the Kurds (Shāhnāme vi: 166–169; Kārnāmag vi) and Kirm-e Haftuwād and Mihrak-e Nūshzād (Shāhnāme vi: 170–189; Kārnāmag vii–ix). The incident with the daughter of Ardawān and the birth of Shāpūr comes in Shāhnāme vi: 194–204 (Kārnamag x–xi), followed by the enquiry of the Indian sages (Shāhnāme vi: 204–207; Kārnāmag xii), and ending in the story of Shāpūr and the daughter of Mihrak and the birth of Hormizd (Shāhnāme vi: 207–214; Kārnāmag xiii–xiv).

Throughout the text, Firdawsī freely embellishes the narrative and invents details, but the main story line clearly follows the Kārnāmag, which is, without doubt, the ultimate source for much of the story of Ardashīr, through the Prose Shāhnāme.

Although Firdawsī did not include the short Husraw ud rēdag-ē in his Shāhnāme, al-Thaʿālibī did take it into his Ghurar, which strongly suggests that the story was found in the Prose Shāhnāme and was excluded by Firdawsī, presumably because it does not tie up with any action and he may have considered it superfluous. Although the text is not found in Firdawsī’s Shāhnāme, a comparison of the Pahlavi original with the Ghurar may throw more light on the way in which Pahlavi originals changed in the hands of the authors of the Prose Shāhnāme and/or al-Thaʿālibī, which further helps us to understand what Firdawsī may have done with the same material.

Al-Thaʿālibī has situated the story in the reign of Khusraw Abarwīz, not Khusraw Anōshagruwān, but the mistake is understandable as the king is better identified in the Pahlavi text only at the end, §125 (Husraw ī shāhān shāh ī Kawādān “the Great King Khusraw, son of Kawād”), which al-Thaʿālibī or his source did not include in his version.

The Pahlavi text begins with identifying the page as the main character (§1 “There was a page named Wāspuhr from Ērān-winnārd-Kawād …”). Al-Thaʿālibī presents the text as part of the wonders of King Khusraw Abarwīz, switching the focus from the page to the King and, thus, tying it up with the general flow of Persian history, making an independent text part of a greater, unified narrative. The Pahlavi text begins, after the shortest of introductions (§§1–2), with a lengthy speech by the page (§§3–18), in which he tells of his highly educated background and the subsequent death and destruction of his family. Al-Thaʿālibī resumes the contents of this speech in a mere two lines of third-person narrative introducing the page in very general terms only (Ghurar, pp. 705–706) and then lets the King start the action by asking the page about which dishes are the best, as well as most suitable and enjoyable.

The two texts go on with the King asking which of various luxurious things is the best and the page answering each question to the King’s satisfaction. The general similarity of the texts is obvious but in details they have a lot of variation. This may primarily be due to the problems involved in the translation of this difficult Pahlavi text, full of names of luxury items and rare vocabulary which may not have been too well understood in the tenth century (and which still defy the attempts of contemporary scholars). In addition, al-Thaʿālabī or his source has also abbreviated the text by, e.g., dropping the standard polite formula anōshag bawēd “may you be immortal” used by the page in his answers. This may well be a stylistic solution, as the repetitive style of the Pahlavi story may have been felt to be too archaic for contemporary taste.

Al-Thaʿālibī ends the story with the page answering the King’s question about the most beautiful and desirable woman, after which the King rewards him with 12,000 mithqāls of silver, the exact amount that is given to him in the Pahlavi text, too (§105). The Pahlavi texts continues with another episode (§§105–124) where the advice given by the page to the King is tested and the page is ordered to catch and later kill two lions, which he promptly does, although tempted by a woman along the road. At the end, the page is created a marzbān of a great province. Like the beginning, the end focuses on the page, not the King, which may again explain why al-Thaʿālibī or his source has dropped it.

All four Pahlavi texts that we have discussed, Wizārishn ī chatrang, Ayādgār ī Zarērān, Kārnāmag ī Ardashīr, and Husraw ud rēdag-ē, have been variously modified before they found their way into Firdawsī’s Shāhnāme and al-Thaʿālibī’s Ghurar. The latter two differ from the Pahlavi texts in various ways, each being sometimes closer to the original than the other, which strongly implies that whatever changes the Prose Shāhnāme had made to the texts, both Firdawsī and al-Thaʿālibī took further liberties with it. This is by no means surprising but in fact tallies well with what we know about contemporary strategies of translation and transmission (Chapter 2.4). In other cases, Firdawsī and al-Thaʿālibī agree with each other, but differ from the preserved Pahlavi original. It is always possible that the Pahlavi texts we have may have undergone changes after they were used for the Prose Shāhnāme, but it seems more probable that the compilers of the Prose Shāhnāme also worked in a similar fashion as Firdawsī and al-Thaʿālibī, changing the text to their liking to create a coherent narrative covering the whole of Persian history, which explains why al-Thaʿālibī and Firdawsī sometimes agree with each other but disagree with the Pahlavi originals.

4.7 Nāme Literature

Nāme literature is extensive and the Sistanian part of it has been well described by van Zutphen in a recent book (2014).⁸⁶ Usually the later epics, nāmes, are seen as epigonal literature composed after, and inspired by, Firdawsī’s magisterial epic.⁸⁷ The first to have done so seems to be the anonymous author of the Mujmal, who calls, p. 2/2, Firdawsī’s work aṣlī “root; origin” and the other nāmes shuʿbahā “branches”.

There is little doubt that Firdawsī did impress many of the authors of these epics; Asadī Ṭūsī (de Blois 1992–97: 83–90) mentions him in his Garshāsbnāme (written 458/1068) with admiration, and was, in fact, one of the first to do so. Likewise, it seems clear that Firdawsī was known even to those later authors who did not mention him by name. Thus, one finds echoes of his famous verse, quoted already by Niẓāmī-ye ʿArūḍī in his celebrated Chahār maqāle, p. 82, man o-gurz-o maydān-o Afrāsiyāb, in many nāmes, including Asadī’s Garshāsbnāme, p. 72 (v. 57): man o-azhdahā o-kuh o-gurz o-tīr.⁸⁸

Likewise, Firdawsī must have consolidated the use of mutaqārib for epics, although he was not the first to use the metre which may well have dominated the scene before him (cf. Chapters 4.1.1 and 4.1.4).

However, Firdawsī did not single-handedly create the genre of epic narratives. We have already seen that Shāhnāmes, both in verse and prose, were in vogue since the mid-tenth century, and we also know that some epics, such as Asadī’s Garshāsbnāme, versified after Firdawsī, go back to prose versions before him. Others, though, may well have been based on minor characters of Firdawsī’s Shāhnāme and be without predecessors, i.e., being the fiction of the author with perhaps some (folk?) narratives to inspire them. Thus, the late epic known by the name of Bānū-Gushaspnāme is clearly a mix of Firdawsian elements, made popular also by the other nāmes, and there need not be any independent story behind this epic.⁸⁹

The early existence of prose epic tales of the Sistanians should also make us wary of seeing all the other epics as epigones of Firdawsī. In time, Firdawsī’s influence became enormous, but the fact remains that many prose tales either preceded him or were written at about the same time as he wrote his Shāhnāme, much before it had its enormous influence on Persian literature. Without underestimating Firdawsī’s influence, it seems safe to say that his overwhelming influence on the early Sistanian epics has been exaggerated and we should see many of these epics, not as epigones of Firdawsī, but as deriving from the same interest in Persian national history that led to the surge of various Shāhnāmes before Firdawsī.

Asadī’s Garshāsbnāme shows that the Sistanian and royal histories had by Asadī’s time been linked together, but unfortunately we cannot know whether this was the case already in the book of Garshāsb he worked on, or whether this was an element added by Asadī, possibly under Firdawsī’s influence. The kings of this epic, Ḍaḥḥāk and Ferīdūn, have a minor role to play and they may well have been added by Asadī as a framework for independent episodes and as a way to tie the whole story up with Persian national history in general.

Asadī’s Garshāsbnāme also shows that the author had other sources of inspiration besides Firdawsī and the epic tradition of Iran. The frequent narratives of wonders in India, China, Maghrib, and elsewhere find their closest parallels in the Alexander Romance, travellers’ stories, and the stories that later found their way into the Arabian Nights.⁹⁰

Most of the later epics centre on the Sistanians. As we have seen and will see in Chapter 5.1, the Khwadāynāmag clearly included little, if any, material on them, which shows that the nāmes do not derive from the Khwadāynāmag, but from a separate, epic tradition. It is also conspicuous that we do not have information on any Arabic translations of these texts, other than those related to Rustam. The reason is not that they contained supernatural elements, although many serious historians may have avoided khurāfāt. The case of many other translated texts, such as Kalīla wa-Dimna and Hazār afsāne, clearly show that such stories could be and were translated into Arabic. This might imply that they did not yet exist, at least not in a written form, at the heyday of the translation movement from the mid-eighth century onward. One has to remember, though, that the huge majority of what was translated into Arabic between 750 and 1000 were scientific or philosophical texts. Historiography and entertainment literature formed a tiny minority.

We have little information of their existence as Pahlavi texts and one is well advised not to speculate on non-existant Pahlavi texts, as the case of the Alexander Romance shows (Chapter 2.3). It is also curious that none of those Sistanian stories that we know to have been translated into Arabic (Chapter 2.2.1) survives in a nāme form outside Firdawsī’s Shāhnāme. This may imply that the written tradition of the nāmes does not go back very far, but that the genre developed only later and the early texts, such as Sīrat Isfandiyār, were not nāmes and did not live on within the tradition. In fact, we have evidence of the existence of nāmes only from the late tenth century onward.

We also know that some (prose) epics existed in the tenth/early eleventh century in written form and, obviously, in Classical Persian, and these formed the immediate source for the writers of the nāmes, at least in some cases. In others, too, one might speculate that most of the nāmes were versifications of existing Classical Persian prose stories, while some, such as the Bānū-Gushaspnāme, may have been completely fictional, with no source other than some hints in Firdawsī’s epics or the other nāmes.

In some cases, the Islamic origin of the nāmes is clear. Thus, in the Rustamnāme, it is ʿAlī who, in fact, is the main character of the story and Rustam is only a dummy to show ʿAlī’s superiority in comparison to the Sistanian hero.⁹¹ The Mujmal also shows that at least the Pīrūznāme contained already before 520/1126, the year of the composition of the Mujmal, materials that must be of Islamic origin and must have either been composed or at least substantially modified in Islamic times (p. 54/67: Shāpūr had heard a prophecy about a new prophet who would end Zoroastrianism and marched against Mecca). Likewise, we find elements of synchronization in, e.g., Asadī, Garshāsbnāme, p. 58, v. 283 (chunān dān ke Hūd andar ān rūzgār / payambar bud az Dāvar-e kirdigār), a feature which we know quite well from the historical tradition since, at least, Ibn al-Muqaffaʿ.

There is no evidence that the earliest epic stories would have been sung by storytellers. As far as we know, prose narratives in a written form, in circulation at the latest in the eleventh century, preceded the versified epics. The prose narratives most probably go further back in time but there is no evidence to show that the earlier texts would have been in verse or would have been sung.⁹² An informed guess would be that they, or at least parts of them, had lived as orally narrated prose stories.

In a few cases, we know the name of the author of a nāme, but mostly the nāmes are anonymous pieces, and some, especially the shorter ones, may first have been inserted into a manuscript of Firdawsī’s Shāhnāme before starting a life of their own.⁹³ In the Mujmal, p. 45/54, there is a report that it was Zāl himself who, when taken prisoner by Bahman, wrote a series of books on the members of his family, i.e., the Sistanian nāmes. This might be taken as an indication that the genre was already mainly anonymous at the time.

Some information on the early history of the nāmes comes from a rather surprising source. Shahmardān ibn abī l-Khayr wrote an encyclopaedia of popular science, Nuz’hatnāme, around 1100.⁹⁴ In this book, the author inserts a rather incongruous chapter (pp. 319–344) on historical matters, discussed under the title Zamān, “Time”. The chapter concentrates on the Sistanian heroes Rustam and Farāmarz, whereas Persian kings have next to no role and the story is not taken to historical times. Two legendary kings (Manūchihr, Gushtāsb) are synchronized with prophets.

The chapter contains five different episodes in the life of Rustam (killing a mad elephant as a child; taking revenge for the death of Narīmān; bringing Kay Qubād from the Alburz to Iṣṭakhr and setting him on the throne; the first attempt to capture Afrāsiyāb; bringing Kay Khusraw from Turkistān and taking revenge for Siyāwush, pp. 319–329). The chapter then goes on to narrate some of the adventures of Rustam and his son Farāmarz, leading to the capture and execution of Afrāsiyāb and including the episode of Āghush-e Wahādān (pp. 329–340). Finally, some partly negative evaluations are given of Rustam’s manners, including a corrective to the well-known story of how Isfandiyār was killed. Isfandiyār was not killed by Rustam but either by a mortar dropping on him or a snake biting him,⁹⁵ after which there is an additional story, a very brief version of Rustam and Suhrāb (pp. 341–342). Later, the author also mentions the episode of the White Demon (p. 343), which either shows that he knew Firdawsī’s Shāhnāme or that these orphan stories do go back to earlier written sources.

Shahmardān, Nuz’hatnāme, p. 342, mentions three authors who may well have been (among) his sources, although he does not explicitly say so. Abū l-Muʾayyad al-Balkhī (Chapter 4.1.3) is just briefly mentioned, but the other two authors receive more attention. Shahmardān tells us that Rustam-e Lārijānī had composed a book which was to stretch from Gayōmard until the reign of the Būyid Shams al-Dawla Abū Ṭāhir (r. 387–412/997–1021). Shahmardān had seen some volumes of this book and, based on them, supposed the whole to be around 500 kurrāses. Otherwise, this book is not described in more detail.

The third author, Pīrūzān, was the teacher of the Kākūyid Shams al-Malik Farāmarz ibn ʿAlāʾ al-Dawla (r. 433–443/1041–1051) and is said to have known both Pahlavi and Persian. Farāmarz had ordered him to translate text(s) (not further identified) from Pahlavi into Persian. The resulting volumes Shahmardān managed to get, and they made a total of between 1,500 and 2,000 pages (waraq). These Shahmardān abbreviated, excluding stories that were fantastic but including those that could be given an allegorical or symbolical meaning, such as the story of Ḍaḥḥāk and the snakes. Shahmardān goes on to list some such legendary stories (Nuz’hatnāme, pp. 342–344) and mentions a Shāhnāme in passing, possibly meaning Firdawsī’s Shāhnāme. There is no indication that Pīrūzān’s work would have continued to cover the historical periods of Persian national history, although this cannot be excluded.

Shahmardān claims that he used this source, but if so, he either distilled a mere 25 pages out of this huge collection or wrote another, unknown, and later lost book, where this material was more extensively used. In either case, it is noteworthy that Shahmardān has a lot to tell about the Sistanians but nothing about the kings, except where they tangentially meet the Sistanians. Rustam-e Lārijānī’s book should have continued the history up to contemporary times, but we know nothing about his sources and the book may well have been a mere continuation of some earlier historical book, such as the Prose Shāhnāme. Pīrūzān’s patrons, as van Zutphen points out (2014: 257–258), bore names such as Rustam, Garshāsb, and Farāmarz, which implies that they were particularly interested in the Sistanian heroes. Based on both this and the contents of Shahmardān’s book, it seems very probable that Pīrūzān’s work heavily centred on the Sistanian heroes and may have completely ignored the kings and their history. If so, his work may have been a central piece in the development of nāme literature.

An intriguing question is whether Pīrūzān really derived all of the 2,000 or so pages of texts from Pahlavi sources or whether, in fact, he compiled his book largely from oral sources and/or the early nāme literature, and merely pretended that he found all this in prestigious Pahlavi books. Likewise, Firdawsī’s references to an ancient “Pahlavi” (i.e., heroic; of hoary antiquity) book could easily be misunderstood as referring to books written in Pahlavi. Such Pahlavi sources on the Sistanians are otherwise completely unknown and, as has been pointed out in Chapter 2.2, all the existing evidence points to the secular Pahlavi texts having been of a rather limited size. Sources that would total up to 2,000 pages in translation would be anomalous. The evidence we have does not allow us to resolve this question. It seems somewhat hasty to hypothesize on the existence of a veritable library of Pahlavi Sistanian texts, against other, admittedly circumstantial evidence, merely on the basis of a short mention in Shahmardān’s book, describing a lost book of Pīrūzān, completely unknown from any other sources. On the other hand, this piece of evidence cannot be brushed aside, either, so the question must remain open. What is noteworthy in this context, though, is that there is no indication that Pīrūzān’s book would have contained anything on the Persian kings and Persian national history, so this lost book has little bearing on the question of the Khwadāynāmag.

The ṭūmār texts, covering the story of the Shāhnāme and some other nāmes, are usually seen as popular storytellers’ aide-mémoires. This may in many cases be so, but the collection and harmonization of a huge repertoire of nāmes into a single, continuous narrative is also a tour-de-force which should not be looked down on. Although the question cannot be studied in the present context, I would yet like to raise the question whether at least some of the ṭūmārs could actually go back to a rather early period and might even retain vestiges of early versions of the stories, whether by Pīrūzān or by others. At least the text edited as Ṭūmār-e naqqālī-ye Shāhnāme by Āydinlū is a valuable summary of a number of epics and would merit a close study of its own, of both its narrative structure and its use of sources, besides Firdawsī’s Shāhnāme and Asadī’s Garshāsbnāme, both openly referred to at the beginning of the story as its sources (p. 155).