The Written Text as a Metaphor for the Integrity of Oral Composition in Classical Persian Traditions and Beyond

The centerpiece of this essay is a brief narrative found in the classical Persian text of the Shāhnāma or ‘Book of Kings’, a monumental poem composed in the early eleventh century CE by a poet named Ferdowsi, literally, ‘the man of paradise’. In this narrative, the poet tells the story of the genesis of an archetypal ‘Book of Kings’ celebrating the kings and heroes of the Iranian Empire. This book, says the poet, became the ultimate source of his own classical Persian poem:

According to this story, an archetypal book had once been disintegrated—only to become reintegrated many years later at the initiative of a wise nobleman who put together the fragments of the disintegrated book. This nobleman, described as a descendant of elite Zoroastrian landowners (dehqāns), proceeded to gather experts in poetry (mōbads) from all over the Iranian Empire. [2] Each one of these experts brought with him a fragment of the original book. Then, lining up the experts in order, the nobleman invited each one of the experts to recite, in proper order, what was contained in each fragment. This way, the archetypal book was reintegrated, and it was this book that became the primal source of Ferdowsi’s own Shāhnāma or ‘Book of Kings’. As I have argued in earlier work, “Ferdowsi’s description of this genesis amounts to a myth-made stylization of oral poetry.” [3]

This is not to say that Ferdowsi’s Shāhnāma is oral poetry in and of itself. Ferdowsi’s poem acknowledges that it was written down, and, to that extent, it cannot be described as oral poetry in the sense of a composition-in-performance. Nevertheless, this poem also acknowledges its own derivation from poetry as performed by oral poets.

In the present work, I will argue that the picturing of a disintegrated and then reintegrated book is a metaphor for oral poetry, and that we find this metaphor attested in a wide variety of oral traditions that have evolved independently of each other. As we will see, such a metaphor can serve a variety of purposes in a variety of cultures that differ widely with regard to their use of written texts.

To get a better sense of such a metaphor, I will consider

• other classical Persian narratives that feature variations on the theme of a disintegrated and then reintegrated Book of Kings
• other Iranian narratives about other kinds of books
• non-Iranian narratives found in other Indo-European traditions.

I start with a classical Persian prose text featuring two further narratives that are most similar in content to the poetic narrative I quoted from the Shāhnāma of Ferdowsi. These two narratives are embedded within the framing narrative of the preface of a book stemming from a recension of Ferdowsi’s Shāhnāma, commissioned in 1426 CE by the Timurid prince Ghiāth-al-Din Bāysonghor. [4] The book, known as the Bāysonghori Shāhnāma, is now housed at the Golestan Museum in Tehran, and I will be quoting here from the text of the preface to this book. From here on, I will refer to this text as the Bāysonghori Preface, abbreviated BP. [5] Near the very beginning of the text (BP 368), the framing narrative says that there were numerous copies of the Shāhnāma in the imperial library of Prince Bāysonghor. But none of them ‘agreed with his delicate constitution and elegant disposition.’ So, that was the impetus, according to the logic of the framing narrative here, for a new recension. And the circumstances were right: ‘when in the time of his imperial rule—may this period of peace be doubled—works of art dominated and well-measured speech held immeasurable value, dried up shoots of wisdom and knowledge were refreshed day by day with the subsidizing drops of rain from his learning.’ So, when the prince called for a collation of texts, the production of ‘a more complete and accurate copy’ was acted upon, and the text highlights all the editorial work that went into the production of an enhanced text for this book. The framing narrative now proceeds to aetiologize this editorial work in the form of two embedded narratives that tell about the genesis of an archetypal book of kings:

In terms of these two embedded narratives, the archetypal book of kings that each one of the two shahs had commissioned seems to be the same text that Ferdowsi supposedly turned into Persian poetry around 1010 CE. At least, that is how this archetypal book of kings is imagined in the embedded narratives.

In effect, these two embedded narratives about two earlier shahs serve the purpose of prefiguring the recension undertaken by the would-be shah Bāysonghor (this prince died before he ever became shah), which is the main topic of the framing narrative. In the second of the two embedded narratives I have just quoted, it is made explicit that the textual recension involves not only earlier texts but also performances stemming from oral traditions. By implication, the variations found in the texts of the Shāhnāma that were housed in the library of Bāysonghor were understood to be derived from not only textual but also performative variations in the transmission of lore about kings and heroes.

But this gesture of prefiguration glosses over a major complication concerning the transmission of this lore. The texts commissioned by the two shahs who are figured here as prototypes of Prince Bāysonghor would have been composed not in Persian but in Pahlavi, the imperial language of the Sasanian Dynasty. The shah Khosrow I Anōshirvān was ruler of the Sasanian Empire from 531 to 579 CE, while the shah Yazdgerd III, the very last king of the Sasanian Empire, ruled from 632 to 651 CE. And there is a further complication concerning the transmission: as we read from the words of Ferdowsi himself, the prototype for his poetic version of the Book of Kings is envisaged as a text written in Pahlavi prose, and it is up to the poet to turn the prose of the Pahlavi book into Persian poetry. Here is how the poet speaks about this Pahlavi prototype:

A still further complication concerning the transmission of lore about Iranian kings presents itself in the so-called Older Preface to the Shāhnāma, hereafter abbreviated OP, attested in the London manuscript of the Shāhnāma (= L, British Museum), dated 1276 CE, as also in other manuscripts. [8] The content of the OP is dated all the way back to the middle of the tenth century. [9] According to the narrative of the OP, a Book of Kings composed in Persian prose was commissioned by a certain Ibn ʿAbd al-Razzāq, a local potentate of Tus (which was the native city of Ferdowsi), in concert with his administrator, Abū Manṣūr al-Maʿmarī (OP §6). The resulting work, finished in 957 CE (OP §7), was based on a ‘compilation’ of older books (OP §6). The narrative relates that Ibn ʿAbd al-Razzāq had been inspired to commission this book after hearing the story of the genesis of the book Kalīla and Dimna, which is a celebrated collection of fables stemming ultimately from India (OP §6).

As we are about to see, the events that are adduced in the Older Preface to explain how the book Kalīla and Dimna attained its supposedly perfect form are presented as a prefiguration of the events that explain how the Shāhnāma supposedly reached its own perfection—once it was turned into poetry by Ferdowsi.

According to the Older Preface, the Sasanian shah Khosrow I Anōshirvān commissioned a translation of the Kalīla and Dimna from the ‘Indian’ language into Pahlavi (OP §4). It was this same shah, as we already noted in the Bāysonghori Preface, who commissioned the reintegration of a supposedly disintegrated Book of Kings, likewise written in Pahlavi (BP 369). Having noted this detail once again, I return to the narrative of the Older Preface: so, many years after the translation of the Kalīla and Dimna into Pahlavi, the Caliph al-Maʾmūn (who ruled from 813 to 833 CE) ordered a translation of this book from Pahlavi into Arabic, through the agency of his administrator Ibn al-Muqaffaʿ (OP §4). [10] Then the Samanid Amir Naṣr II b. Aḥmad (who ruled from 914 to 943), through the agency of his secretary Balʿamī, commissioned a translation from Arabic into Persian prose (OP §5). And, finally, Balʿamī’s prose version was transposed into Persian verse by the poet Rudaki, who flourished in the late ninth and early tenth century CE (OP §5). It is said of the Persian prose version that ‘the book fell into men’s hands and every man turned its pages,’ and of the poetic version that from then on, Kalīla and Dimna ‘was on the tongues of the great and the lowly’ (OP §5). [11] Ferdowsi himself draws a parallel between his own Shāhnāma and Rudaki’s Kalīla and Dimna, stressing that the uniqueness of both compositions depends on what is described as the turning of prose into poetry (Shāhnāma VIII 655.3460–3464). [12]

The rhetoric of the narrative I just analyzed in the Older Preface operates on a principle that I describe as an ascending scale of prestige. At the climax of the narrative, the Persian prose version of the Kalīla and Dimna is capped by the Persian poetic version. And now we are about to see that the same kind of rhetoric is at work in the “happy ending” built into the narrative of the Older Preface about the genesis of the Shāhnāma. In this instance, just before the climax of the narrative is reached, the Book of Kings has been translated from Pahlavi prose into Persian prose (OP §15–§16). Then, in a final and perfect culmination of the process of composition, the Persian prose text is superseded by the Persian poetic version. This time, it is Ferdowsi who converts the prose into poetry, and his patron is specified as Maḥmud of Ghazna, who ruled from 998–1030 CE (OP §16). In this narrative, there is no explicit time-gap between the compilation of the Persian prose version and its subsequent conversion into Persian poetry: ‘And after they had compiled it in prose, Sulṭān Maḥmud, son of Sebüg-tegin, ordered the sage Abu ’l-Qāsim Manṣur Ferdowsi to turn it into the dari [‘court’] language in verse, and the circumstances of it will be mentioned at their proper place’ (OP §16). [13]

According to one influential theory, this crucial part of the narrative, the climax, is an “interpolation.” [14] But the advocates of this theory overlook the rhetorical integrity of the Older Preface when they proceed to excise the supposedly interpolated part of the text. With the excision in place, the truncated text is then “archaeologized” as having had a completely different purpose from that which the OP itself announces as its raison d’être, namely to introduce the poetry of the Shāhnāma. Instead, in terms of the theory, the text was intended to introduce a hypothetical “prose Shāhnāma,” which, although no longer extant, supposedly corresponded to the original text commissioned by Ibn ʿAbd al-Razzāq.

Opposing this theory, I argue that the narrative of the Older Preface requires the text that preceded the poetic Shāhnāma to serve as a foil, marking the penultimate stage in the process that led to the perfection of the final product. Narrative precedents provide a clear model for the construction of such a foil: at least in terms of the rhetoric of the narrative, it is evident that the predecessor-text must have been written in prose, just as Balʿamī’s prose Persian Kalīla and Dimna constituted the predecessor-text for the poetic Persian Kalīla and Dimna of Rudaki. Similarly, in terms of the logic that takes shape in the narrative of the Older Preface, the Shāhnāma must have existed in a prose version that functioned as a predecessor-text to the poetic Shāhnāma of Ferdowsi. Such a rhetorical crescendo, in which a prose Shāhnāma commissioned by Ibn ʿAbd al-Razzāq becomes a stepping-stone to a grander and more perfect work, articulates a point of view marked by a poetic Shāhnāma commissioned by Maḥmud of Ghazna.

The purpose of the preceding discussion is not to suggest that a prose Shāhnāma commissioned by Ibn ʿAbd al-Razzāq, the potentate of Tus, never existed. Rather, it is to demonstrate that the narrative in the Older Preface about the commissioning of such a prose Shāhnāma is inextricably tied to its narrative concerning the conversion of this prose into poetry by Ferdowsi, the poet of Tus. In other words, the narrative of the Older Preface rhetorically motivates the Shāhnāma of Ferdowsi instead of the earlier “prose Shāhnāma” that serves as its foil.

There is a comparable narrative in the Bāysonghori Preface concerning the same Sultan Maḥmud, would-be patron of Ferdowsi in the eleventh century CE: this king showed a keen interest in locating texts of the Shāhnāma stemming from western Iran (in this case, Fars, at BP 371) and from eastern Iran (in this case, Sistān, at BP 373). In this case, the activity of gathering regional variants of the Shāhnāma in the era of Maḥmud is presented as a prototype for the interest of Prince Bāysonghor himself in attempting to enhance the text of the Shāhnāma by way of augmenting it with regional variants.

A result of such attempts at enhancement by augmenting the actual text of the poetic composition by way of regional variants is the enormous size of the Bāysonghori Shāhnāma. This fifteenth-century book contains over 58,000 verses. It has 346 folios (with six columns and 31 lines per page) and only 20 illustrations, by comparison with most books of the Shāhnāma in the fourteenth century, which typically have around 300 folios and many more illustrations. [15]

At the beginning of a thirteenth-century narrative poem about the early life of Zoroaster and the beginning of his public life as a prophet, the poet speaks of his source in the following words:

Later on in the poem, the poet says that a mōbad urged him to translate the book into Persian verse. Here is what the mōbad tells the poet:

Such medieval Persian narratives about the rescuing of books written in Pahlavi can be traced back to still earlier Iranian narratives. Already in the ancient Avestan tradition, a book could be seen as the foundation of performance. A case in point is a narrative that we find in the Dēnkard, which is a tenth-century Pahlavi compendium of Zoroastrian beliefs and customs (in Pahlavi, the word Dēnkard or Dēnkart means ‘Acts of Religion’). In the Dēnkard, we find a story about the transmission of archetypal Avestan texts—texts that date back ultimately to the second millennium BCE. According to this story, three pivotal moments in the transmission happened in three successive eras of Iranian imperial kingship. The three eras are represented by three kings who stem from three successive Iranian dynasties, which are [1] the Achaemenid, [2] the Parthian, and [3] the Sasanian. Here is the relevant text, as translated by the noted Iranist Mary Boyce:

So we see a spanning of three consecutive dynasties in this narrative:

On the basis of such comparative evidence, I argue that the medieval Persian narrative of Ferdowsi’s Shāhnāma about a disintegrated and then reintegrated book can be traced back to earlier Iranian traditions. I especially note, in phase two of the narrative in the Dēnkard, the reference to the scattering of the texts of the Avesta: to quote again from the translation of Boyce, the texts were ‘scattered through the land of Iran’. In brief, then, the ancient Avestan tradition is seen in terms of a book that serves as the foundation of performance. So there is a deep prehistory to be found in the linkage of book and performance.

Over a century and a half ago, Jules Mohl had already interpreted the passage I quoted from Ferdowsi’s Shāhnāma at the beginning of this presentation as an example of a well-known type of myth that serves to explain the genesis of a national epic poetic tradition. [20] Following my analysis of this passage from the Shāhnāma, Gregory Nagy interprets it as evidence for the kind of culture “where written text and oral tradition coexist.” [21] In such cultures, as Nagy argues, “the idea of a written text can even become a primary metaphor for the authority of recomposition-in-performance.” [22] The consequences are enormous:

For Nagy, the most striking comparative parallel with the Persian model is a set of ancient Greek myths that tell of the disintegration and subsequent reintegration of the Homeric corpus itself, culminating in the historicized narrative of the “Peisistratean Recension.” [24] Other comparable parallels include the medieval Irish aetiology of the “lost” book of the Táin Bó Cúailnge ‘Cattle Raid of Cúailgne‘. [25] Another parallel adduced by Nagy is the medieval French narrative of Guiron le courtois, which “lays the foundation for its authority by telling of the many French books that were produced from what is pictured as an archetypal translation of a mythical Latin book of the Holy Grail.” [26] Just as the whole tradition of medieval French romance presents itself as an overall corpus that derives its individual parts from translations of a mythical Ur-text written in Latin, I argue that the whole of Ferdowsi’s Shāhnāma derives itself from a comparable Ur-text written in Pahlavi. [27] Yet another parallel comes from ethnographic fieldwork on the oral epic tradition of an illiterate society, the untouchable Malas of India: “The epic, it is claimed [by the performers], was first written by a Brahmin poet, torn into shreds, discarded, and then picked up by the present performers.” [28]

I started with Iranian traditions as I examined the theme of the disintegrated and then reintegrated book. And I concluded by surveying parallel themes in non-Iranian traditions, both cognate (as in the case of Greek, Irish, and other Indo-European traditions) and non-cognate (as in the case of substrate traditions in India). On the basis of these parallel themes, I argue in general that oral traditions can compare themselves metaphorically to written traditions.