On the Kalīla wa-Dimna

2023.11.02 | By Olga M. Davidson

§0. This pre-edited standalone essay, included in a series of pamphlets published online and also in print (“on-demand”), consists of two parts. Part I is a pre-edited rewriting of a shorter essay (Davidson 2015a) originally published in a book edited by Regula Forster and Neguin Yavari (2015), while Part II is a likewise pre-edited but new followup, presented on November 2, 2023 at the Annual Meeting of the Middle East Studies Association, at Session I-11, “Premodern Fables in a Comparative Frame.” A printed version of the presentations for that session, edited by Neguin Yavari, is already being planned, to be published by the Ilex Foundation in a series distributed by Harvard University Press, https://www.hup.harvard.edu/collection.php?cpk=1222.

Part I: Aetiologies of the Kalīla wa-Dimna as a mirror for princes

§1. It is well known that the stories contained in the multiform text known in Arabic as the Kalīla wa-Dimna were used for purposes of education in statecraft. The literary form of such education is commonly known by its medieval Latin designation as speculum principum or ‘mirror for princes’. In the case of the Kalīla wa-Dimna, the genesis of the stories told in this multiform text is explained in statements that are attached as introductions or prefaces or even appendices to the main body of stories. These statements, which can sometimes include stories in their own right, are what I call aetiologies, or, more precisely, aetiological narratives. As for the actual stories of the Kalīla wa-Dimna, they are animal fables. What I argue here is that the aetiological narratives of this multiform text of animal fables, the Kalīla wa-Dimna, are designed to explain not only the genesis of these fables but also their use as a mirror for princes. I will also argue that these aetiological narratives are multiform in their own right, just as the fables themselves are multiform.

§2. Such multiformity is typical of oral traditions, which require ongoing adjustment to the reception of whatever has been composed. I use the word reception here along the lines defined by Gregory Nagy (2009) in a study of the oral traditions that shaped the poetry attributed to Homer and Hesiod. I quote from the formulation he gives in that study (Nagy 2009:282):

I use the term reception … not in the narrow sense that applies in studies of literature, where this term conventionally refers to whatever happens after a given piece of literature is composed for transmission to the public. A broader sense of the term is needed when we are dealing with literary traditions that stem from oral traditions… . How, then, are we to understand the phenomenon of reception in oral traditions? The answer has to do with the transmission of composition by way of performance. In any oral tradition, … the process of composition is linked to the process of performance, and any given composition can be recomposed each time it is performed. The performer who recomposes the composition in performance may be the same performer who composed it earlier, or it may be a new performer, even a succession of new performers. The point is, such recomposition-in-performance is the essence of transmission in oral traditions.

To quote further from Nagy (2009:283):

This kind of transmission is the key to a broader understanding of reception. Unlike what happens in literature, where reception by the public happens only after a piece of literature is transmitted, reception in oral traditions happens during as well as after transmission. That is because the process of composition in oral traditions allows for recomposition on each new occasion of performance for a public that sees and hears the performer. In oral traditions, there is an organic link between reception and performance, since no performance can succeed without a successful reception by the public that sees and hears the performer or performers.

§3. I add here to Nagy’s formulation another dimension of such reception: in the case of fables, which are used for education in statecraft, the reception depends on the patron as well as on the public. After all, the intended patrons of fables are rulers of state, and it is their statecraft that the fables are intended to shape.

A brief introduction to the animal fable as a traditional form of discourse

§4. Animal fables are stories about talking animals, whose words can be quoted directly. The telling of such stories, which do not exclude humans interacting with animals, is a very old traditional form of discourse that is found all over the world in a wide variety of unrelated cultures. Clearly, such fables originate from oral traditions. And, just as clearly, the many different forms of animal fables cannot be traced back to any single original form.

§5. Still, in some cases, the traditions of this kind of storytelling as attested in different cultures can in fact be traced back in time to a common source. I emphasize that I am saying common here, not original. Such is the case with animal fables attested in the earliest surviving phases of ancient Greek and ancient Indic traditions, which can be traced back to the common linguistic heritage of Greek and Indic as Indo-European languages (Nagy 2011 §§121-124). The ancient Greek tradition of telling fables featuring talking animals is attested already in the Hesiodic Works and Days, which contains the embedded fable of The Hawk and the Nightingale (verses 202-212) and in the iambic poetry attributed to Archilochus, which contains the embedded fable of The Fox and the Eagle (fragment 174 ed. West). The eventual textualization of the poetic traditions attributed to Hesiod and Archilochus can be dated around the sixth century BCE or even earlier. Correspondingly, the ancient Indic tradition of telling fables featuring talking animals is prominently attested in the Buddhist texts known as the Jātaka Tales, composed in the Pāli language, and in such “Hindu” texts as the Pañcatantra and the Hitopadeśa, both composed in Sanskrit. The textualization of the Jātaka Tales is dated around the fifth century BCE, while the dating for the Pañcatantra and the Hitopadeśa is later. With due allowances for the fact that the dating of Sanskrit texts is notoriously slippery, I note that the textualization of the Pañcatantra is conventionally estimated at around the third century BCE, while the Hitopadeśa as we know it was produced as a text in the twelfth century CE.

§6. In other cases, the animal fables of one culture are borrowed from corresponding fables in another culture. One such case is the Kalīla wa-Dimna, which is the Arabic name for a collection of fables borrowed from earlier fables composed in Middle Persian, which in turn were borrowed—at least in part—from still earlier fables composed in Sanskrit.

A sketch of the structure of the Kalīla wa-Dimna

§7. Before we consider the textual traditions that culminated in the collection of fables that we know as the Kalīla wa-Dimna, I offer here a sketch of the structure of this collection. Basically, the fables of the Kalīla wa-Dimna are divided into a sequence of ten main stories, which contain sub-stories, which in turn can even contain sub-sub-stories. The ten main stories are the following:

– The first five of the ten main stories are The Lion and the Ox, The Ring-Dove and her Companions, The Owls and the Crows, The Ape and the Tortoise, and The Ascetic and the Weasel. These stories were borrowed from the five main fables of the Pañcatantra, which I have already mentioned earlier.

– The sixth, seventh, and eighth of the ten main stories are The Mouse and the Cat, The King and the Bird, and The Lion and the Jackal. These stories were borrowed from fables found in Book XII of the Mahābhārata, which is the main epic of Indic civilization.

– The ninth is The King and his Eight Dreams, borrowed from an unattested Sanskrit source.

– The tenth is The King of the Mice and his Ministers, an external source for which cannot be determined.

§8. Besides the main body of fables organized along the lines of the ten main stories as I have just outlined them, the Kalīla wa-Dimna also features a set of statements that introduce this main body of stories. As I noted from the start, these introductory statements can sometimes include stories in their own right. And, as I also noted from the start, I describe these statements as aetiological narratives.

A sketch of the textual traditions of the Kalīla wa-Dimna

§9. Here is an inventory of the attested texts of the Kalīla wa-Dimna. This inventory tracks a corresponding inventory presented in a monograph by François de Blois (1990:5):

(1) At the core of the textual tradition of the Kalīla wa-Dimna is a wide variety of texts written in Arabic. If we follow the aetiological narratives attached to these texts, the fables of the Kalīla wa-Dimna originated only indirectly from Sanskrit prototypes. There was an intermediate stage, in the form of a text written in Middle Persian. This Middle Persian text, which is not attested, was attributed by the aetiological narratives to a physician named Borzōy (Barzūya in the Arabic script), who lived in the era of the Sasanian king Khosrow I (Kisrā in the Arabic script) Anōshirvān, emperor of the Persian Empire from 531 to 579 CE. This Middle Persian text, as posited by the aetiological narratives, was later translated into Arabic by a Persian savant named ‘Abd Allāh ibn al-Muqaffa‘, the date of whose death is traditionally set at 757 CE. The oldest of the texts stemming from the Arabic translation of Ibn al-Muqaffa‘ is dated to 1221 CE, called by de Blois the “’Azzām manuscript”, while the second oldest is dated to 1338 CE, called the “Shaykhū manuscript.” Finally, I must mention two poetic versions of Kalīla wa-Dimna: one by ’Abān al-Lāḥiqī (died around 815 CE), which survives only in a few quotations, and another by Ibn al-Habbārīyah (11th to 12th century).

(2) The oldest attested text of the Kalīla wa-Dimna, however, is a version composed not in Arabic. It is the Qalīlag w Damnag, composed in Syriac. This “older Syriac” version, as it is called, is dated to the sixth century CE. As de Blois argues, this “older Syriac” text was a “translation” from the same Middle Persian text that became the prototype for all the Arabic versions of the Kalīla wa-Dimna.[1] This older Syriac version was composed by a savant named Bōd – according to a 14th century Nestorian Christian named ʿbed Īshōʿ, who notes that Bōd also composed other treatises, including Discourses on the Faith, Against the Manichaeans, and Against the Marcionites. I observe that the religions of both the Manichaeans and the Marcionites are strongly dualistic, and this observation is relevant in what I will argue later about the Persian savant Ibn al-Muqaffa‘, who as we saw is credited with the Arabic translation of the Kalīla wa-Dimna. But let me return for now to the aetiological narrative given by the 14th-century Nestorian Christian named ʿbed Īshōʿ about the genesis of the “older Syriac” version of the Kalīla wa-Dimna. This narrator goes on to report about Bōd: ‘And he [Bōd] interpreted the book of Qalīlag w Damnag from the Indian’.[2] This figure Bōd was active around the year 570 CE.[3] The expression used by ʿbed Īshōʿ, pashsheq men hendwāyā ‘interpreted from the Indian’, is discounted by de Blois, who claims that the Syriac text was not “translated” from the Sanskrit but from Middle Persian.[4] He cites as his authority the work of earlier orientalists who show occurrences of Middle Persian words in the Syriac text. Even the name Bōd is evidently Iranian, morphologically parallel to Avestan baoδah‑, meaning ‘wise’.[5] No doubt there was Middle Persian mediation. But I must point out an empirical fact that stands in the way of assuming, as does de Blois, that Bōd simply “translated” a Middle Persian text. That empirical fact is this: the older Syriac version aetiologizes itself as an interpretation stemming from the Indic tradition of fables, not from a Middle Persian translation. And the claims made in aetiological narratives about a tradition, whether these claims are true or not, can be used as evidence for reconstructing the actual reception of that tradition, as we will see later.

§10. As de Blois notes, there are medieval “translations”—again he uses that word—some of which are roughly contemporaneous with the oldest Arabic manuscripts, and so de Blois considers them fair game in “reconstructing” the text of ibn al-Muqaffa‘.”[6] So, even though de Blois cannot even reconstruct that text on the basis of existing Arabic manuscripts, he still treats stemmatologically these versions in non-Arabic languages—even though these non-Arabic versions all claim various authors of their own. Here is the inventory of de Blois, which can be added to the first and the second sets of texts I have already listed:

(3) There is a Syriac version, dated to the 10th or 11th century, independent of the older Syriac version of the 6th century. This later version, “translated” from the Arabic, is described by de Blois as “rather free.”[7] He adds: “The translator has padded the text considerably and introduced numerous quotations from the Bible.”[8]  I highlight here the use of the word “padding,” to which I will return later.

(4) There is a Greek “translation” (as de Blois calls it), entitled Stephanites kai Ichnelates, by Symeon Seth, dating from the latter part of the 11th century.

(5) There are also New Persian “translations” (again, as de Blois calls them): [9]

(5a) The oldest known example of a New Persian version was commissioned by the Samanid wazīr Balʿami, and the story of that commissioning is told in the Shāhnāma of Ferdowsi. This version, as de Blois argues, “was apparently the basis for the Persian versification by the celebrated poet Rōdakī [born 858 CE and died somewhere around 941], of which scattered verses survived.”[10] I draw attention to the term “versification” used by de Blois, which for him is a mechanical process.[11]

(5b) A Persian prose “translation” (as de Blois calls it) by ’Abū l-Maʿālī Naṣr Allāh, dated to the 12th century.

(5c) There is another Persian version, composed “in a straightforward prose style” by Muḥammad ibn ʿAbd Allāh al-Bukharī, and also dated to the 12th century. This “translation,” as de Blois calls it, “is evidently quite independent of ’Abū l-Maʿālī’s.”[12]

(5d and 5e) There are also two other Persian versions, both based on the version of ’Abū l-Maʿālī, and dated to the 16th century.

(6) There is an older version of a Hebrew “translation,” as de Blois calls it, which is dated to the first part of the 12th century. A Latin version by John of Capua, which is “translated” from the Hebrew,” is dated to the 13th century.

(7) There is a series of Renaissance-era “translations,” as de Blois calls them, of John’s Latin version into various European languages.

(8) There is also a newer version of a Hebrew “translation,” as de Blois calls it, dated to the 12th century, and “made directly from the Arabic.”

(9) Finally, there is a Spanish “translation,” as de Blois calls it, “made directly from the Arabic” and dated to the 13th century. In view of a the frequent word-for-word correspondences in this case, I agree that this version may be described as a “translation” in the modern sense of the word.

A rethinking of the concepts of “borrowing” and “translation” as applied to the fables of the Kalīla wa-Dimna

§11. In my brief introduction to the animal fable as a traditional form of discourse, I had said that fables of the Kalīla wa-Dimna were borrowed from earlier fables composed in Middle Persian, which in turn were borrowed—at least in part—from still earlier fables composed in Sanskrit. But now I need to rethink the concept of “borrowing.” Then, in my inventory of the attested texts of the Kalīla wa-Dimna, I noted that de Blois keeps describing these texts as “translations” from earlier fables composed in Middle Persian, which in turn were “translated” —at least in part—from still earlier fables composed in Sanskrit. Now I need to rethink these concepts of “borrowing” and “translation.”

§12. We cannot really say that a fable such as The Sage and the Mouse Maiden, which is a sub-story of the fable of The Owls and the Crows as we see it attested in the Pañcatantra, was simply “borrowed” and “translated” in the Middle Persian traditions of animal fables and then further “borrowed” and “translated” in, say, the multiple Arabic traditions of the Kalīla wa-Dimna. That is because the Sanskrit Pañcatantra and such later collections as the Arabic Kalīla wa-Dimna are significantly different from each other in form.[13] And I draw attention to the fact that I just spoke of multiple Arabic traditions of the Kalīla wa-Dimna, not of any single Arabic tradition. We have seen the multiplicity in the inventory I just presented. Further, we cannot be sure whether these multiple Arabic traditions were ultimately derived from the Pañcatantra as we know it. As we are about to see, there must have existed other Indic versions of the Pañcatantra.

§13. The Pañcatantra was composed in a form of discourse known as prosimetrum, where some parts of the discourse are in poetry while other parts are in prose. And the parts that are composed in poetry are not necessarily insertions of poetic “quotations” from other poetry. Rather, the discourse is flexible in its capacity to switch from prose to poetry. As Gregory Nagy points out, “Indic narratives of fables regularly show a mixture of poetry and prose, with the parts composed as poetry tending toward older and higher levels of discourse while the parts composed as prose tend toward newer and lower levels.”[14] Further, “in a typical Indic fable, the part composed in poetry tends to be the moral of the story, while the part composed in prose is the story itself.”[15]

§14. When the register of poetry in the Pañcatantra indicates a more elevated form of discourse, it can refer to a custom—as in an example we find in the fable of the Mouse Maiden, where two passages of poetry refer to norms concerning marriage between members of different castes.[16] Not a single one of the corresponding versions of this fable in the Kalīla wa-Dimna transmits the content of the second verse, but some of the versions do transmit the content of the first, which says that parents who do not marry off their daughter before she menstruates may be considered śūdra-s. But the transmission of this content in the Kalīla wa-Dimna is in prose. And any reference to the workings of a caste system is omitted in all versions of the Kalīla wa-Dimna. That omission shows that we are not dealing with some literal-minded “translation.” And, more interestingly and importantly, the content of the poetry can be transmitted, even if the form is changed to prose.

§15. As de Blois concedes, “all descendants of the Middle Persian version” contain a passage that is missing in the Pañcatantra: it is a quoted dialogue between the Sage and the Mouse Maiden who has by now been transformed into a young girl in his care and whom he now needs to marry off.[17] Only in this dialogue, as we find it attested in the Kalīla wa-Dimna, do we find out that the Sage gives the girl the option of choosing a husband. In the version of the Pañcatantra as we have it, by contrast, the Sage simply decides by himself to marry off the girl to the Sun. So now de Blois is forced to invoke the opinion of earlier orientalists who conceded that “the Near Eastern versions [of the Kalīla wa-Dimna] reflect here a different and perhaps older recension of the Sanskrit text than that represented by the existing texts of the Pañcatantra.”[18]

§16. But there are easier ways to account for a detail that is missing in the texts of the Pañcatantra as we have it. The story of the Mouse Maiden is attested in a wide variety of Indic oral traditions, and it is known to folklorists as the Aarne-Thompson Tale-Type 2031C, “The man who seeks the greatest being as a husband for his daughter,” narrated in oral traditions all over India as also in the textual traditions of the earliest collections, namely, the Jātaka Tales, the Pañcatantra, and the Kathāsaritsāgara.[19]

§17. But de Blois ignores the evidence of oral traditions. For him, any variations he finds in the texts of the Kalīla wa-Dimna are merely the result of vicissitudes in the textual transmission. For example, if he finds a longer version of a detail in a fable contained in the Kalīla wa-Dimna, he resorts to treating such a longer version as “a good illustration of the translator’s fondness for padding the text.”[20]

§18. The existence of many different versions in the texts of the Arabic Kalīla wa-Dimna does not mean that the differences we find are arbitrary. They are organic. For example, an analysis of the various different versions of The Sage and the Mouse Maiden as we find this fable attested in the textual traditions of the Arabic Kalīla wa-Dimna shows that the variations we see in these different versions are not just textual: they are also performative. That is, the multiformity of the textual traditions of the Arabic Kalīla wa-Dimna is the result of multiformity in the actual performance of such fables as The Sage and the Mouse Maiden for real audiences. And such acts of performance need to be explained in terms of the oral traditions at work in the telling of fables.

Comments on the multiformity of the textual tradition of the Kalīla wa-Dimna

§19. Already in the first paragraph of my essay, I described as “multiform” the received textual tradition of the Kalīla wa-Dimna. Now I am ready to say why I had used that term multiform there: in terms of my argument, it is because multiformity, even in textual traditions, is a clear sign of an earlier multiformity in oral traditions.

§20. Returning to the inventory of the attested texts of the Kalīla wa-Dimna, I note that many of the textual differences that we see there can be explained in terms of such multiformity:

– In the case of the Greek version (number 4 in the inventory), de Blois remarks: “Where the later Syriac version stretches the text, the Greek version reduces it to a bare minimum.”[21] From the standpoint of oral traditions, such variation can be described in terms of expansion and contraction.

– In the case of the Persian prose version attributed to ’Abū l-Maʿālī (number 5b), here is how de Blois describes this text: “The translator has stuffed it with quotations from Arabic and Persian poems, from the Qur’ān and Ḥadīth, and so on, all of which sounds rather quaint in the mouths of animals in the jungles of India.”[22] But if this work by ’Abū l-Maʿālī is merely a “translation” and if ’Abū l-Maʿālī is merely a “translator,” then how do we explain the reception of such a work? Also, de Blois reveals here a misunderstanding of the animal fable as a literary form. From a comparative point of view, this form is not something that suits only “the jungles of India.” As we saw earlier, animal fables can be found in a worldwide range of cultures. And, from a comparative point of view, there is nothing “quaint” about the utterances of high literary forms—even if they come from “the mouths of animals.” I see a further assumption made here by de Blois, and I quote: “In order to fit these quotations into the book, the translator has padded the prose text, too, to a considerable extent.”[23] Once again, de Blois uses the term “padded.” Clearly, he does not think of recomposition in the context of reception. He thinks of a passive “translator” who cannot recompose creatively but can only be capable of “padding.” But the comparative perspective shows that discourse expands to accommodate the “quotations.” And de Blois has actually seen it happen in the Pañcatantra, noting that the work of ’Abū l-Maʿālī actually resembles the Pañcatantra. But he does not see that such expansion in the context of “quotations” is a common feature of animal fables in general. And he does not take into account the discourse of the Pañcatantra as a form of prosimetrum.

– In the case of the newer Hebrew prose version (number 8), de Blois describes it this way: “This version is much freer than the older [Hebrew] one,” and “the work is spiced with innumerable quotations from the Bible.”[24] I note simply that such “spicing” is analogous to the elevations of register in the discourse of prosimetrum.

§21. For de Blois, however, all signs of multiformity in the textual tradition of the Kalīla wa-Dimna have to be explained in terms of whatever it is that scribes do:

A comparison of the various [Arabic] manuscripts reveals at once such a degree of discrepancy that one must often wonder whether they are really copies of one and the same book. It appears that the Arabic Kalīla wa-Dimna has to a large degree become a victim of its own popularity.[25]

§22. The divergences among the manuscripts of the Kalīla wa-Dimna as described throughout the monograph of de Blois reinforce this initial impression as he formulates it here at the beginning of his work. And I note his use of the word “popularity.” You would think that “popularity” is an indication of positive reception by audiences who hear the fables performed, and such positive reception would lead to more and more variations. But de Blois will have none of that. He follows up his formulation of his first impressions by explaining away the variations in terms of scribal initiatives. Here is the way he puts it:

For one thing, the frequent reading of the book insured that all the old copies were rapidly worn out and had to be replaced. For another, editors and copyists felt free to alter the text, to add new stories and rewrite old ones, to combine material from various manuscripts, and so on, in a way which would have been unthinkable in the case of a “serious” work, say on theology.[26]

§23. It is as if “popularity” were a matter of reading only, not of listening as well. So, the primary indication of “popularity” is that there must have been lots and lots of wear and tear at the hands of lots and lots of readers. But what about the “serious” works, as on theology? Should we not think that they too suffered from lots and lots of wear and tear at the hands of lots and lots of readers?

§24. So I find untenable the first explanation offered by de Blois for the instability of the textual tradition of the Arabic Kalīla wa-Dimna. But then de Blois goes on to offer a second explanation, saying that “editors and copyists felt free to alter the text, to add new stories and rewrite old ones, to combine material from various manuscripts, and so on, in a way which would have been unthinkable in the case of a ‘serious’ work.”[27] As we saw, de Blois cites “theology” as an example of something “serious,” which he now goes on to juxtapose with the very idea of a collection of fables:

For Kalīla wa-Dimna was generally (if not always) considered to be a “popular” work, a piece of entertainment, which one did not need to approach so respectfully.[28]

§25. So now de Blois is contrasting “popular” in scare quotes with what he had earlier described as “serious,” also in scare quotes. But the way he uses both terms is confined to writing and reading. What is missing in this narrow usage of both terms “popular” and “serious” is the idea that fables were meant for performance before audiences and that the tellers of the fables had to adjust to their audiences. By contrast, a discourse on, say, theology could be notionally non-adjustable.

§26. Whereas some things, like scripture, are notionally unchanging, other things, like fables, are meant to change, applying in different ways in different situations. In the case of the Greek Fables of Aesop, for example, the earlier forms show no explicit formulation of any moral to the story. Only the later forms spell out the moral: ‘and the moral of the story is’. And we can see a comparable kind of non-explicitness in the Sanskrit Pañcatantra as well. The moral has to be read from the contextualization.

§27. We may compare the reception of fables to the reception of recipes. Your ingredients will have to change from place to place. Similarly in the case of fables, your performance will change from audience to audience. It is a matter of transmission and reception. Moreover, you can have changes even in structures that are notionally unchanging, but there is appreciably more change in structures that are meant to be applied and reapplied in different contingencies. This kind of thing is brought to life in Book XII of the Mahābhārata, where the contextualization of performance is dramatized by the fact that the narratives of animal fables are embedded in a narrative that shows how and why the fables are performed by the character of Bhiṣma to the character of Yudhiṣṭhira.

§28. The need to adjust the content of fables to ever new situations is illustrated in the quotations from Kalīla wa-Dimna in the ʿUyūn al-’Akhbār by Ibn Qutaybah (died in CE 889) and the Kitāb alʿIqd by Ibn ʿAbd Rabbihī (died in CE 940). As de Blois notes, following the opinions of previous orientalists, “the passages in Ibn Qutaybah do not seem always to be literal quotations: often one finds instead a free paraphrase of ibn al-Muqaffa‘’s text.”[29]

Multiform aetiologies for the multiform text of the Kalīla wa-Dimna

§29. Returning one last time to the first paragraph of Part I here, I reiterate that I described as “multiform” not only the received text of the Kalīla wa-Dimna, which is attested in many different versions, but also the aetiological narratives that were attached to this text, which are also attested in different versions. In Part II, I will concentrate on two such different versions of aetiologizing the Kalīla wa-Dimna, and I will argue that such aetiological narratives were designed to explain not only the genesis of such a book but also its multiform reception in the many different societies that were affected by this collection of fables—and that in turn affected the content of these fables.

Part II:  A West European parallel to the mythological version of the aetiology of the Kalīla wa-Dimna 

§1. Part II here follows up on Part I, where I focused on the “aetiology” of the Kalīla wa-Dimna. In terms of that aetiology, a quest for a plant that confers immortality results instead in the finding of an archetypal book, which is a corpus of beast fables known today as the Kalīla wa-Dimna. Here in Part II, I follow up with an extended analysis of a parallel found in West European traditions, where the central narrative concerns a quest for what is known today as the Holy Grail.

§1a. This narrative is attested in countless versions, the most extensive and comprehensive of which is a massive text written in Old French and dating to the early thirteenth century CE. This text is commonly designated by English-speaking experts as the Vulgate Cycle or the Arthurian Vulgate—and by French-speaking experts as la vulgate arthurienne en prose. The narrative of this Vulgate Cycle is contained in a text extending to roughly four thousand quarto pages (the three attested manuscripts are catalogued this way: London, British Library, “Add.” MS 10292, 10293, and 10294). This text is remarkably cohesive in its central purpose, which is, to represent the Grail as essential for life not because the overall content of this narrative about the Grail has any single meaning but because the narrative, as narrative, highlights one simple and obvious fact, which is, that we see here a monumental book containing a vast variety of stories—stylized as fables—about King Arthur, his knights, and beyond. I highlight the idea that this book “contains” the stories—in the sense that the Vulgate Cycle, as a book, is a “container” of these fables.

§1b. But I also highlight another fact about this monumental book, which is, that it avoids fidelity to any single ideology, political view, or religious outlook. Rather, what makes this book one single book is simply the fact that all the fables contained inside the book are contained simply by the container itself, which is the master narrative, imagined as the totality of the book that contains all the fables. Thus the integrity of the “book” is to be found not in the content of the book that contains it but in the book itself as a glorified container. For a thoroughgoing demonstration, I recommend a study by Alexandre Leupin (1982), the title of which is Le Graal et la littérature (hereafter abbreviated simply as “L”).

§1c. Let me restate my argument by borrowing a playful aphorism once formulated in a book by Marshall McLuhan (1964), where he asserts, quite simply, that the medium is the message. Applying this playfully universalizing assertion to the specific subject of my study here, I will now assert, also somewhat playfully, that the quest for the Holy Grail is the same thing, so to speak, as the mental process of reading the book of the Holy Grail, in that the medium of the book is of and by itself the message about the meaning of the book. Similarly, as I argue, the quest for a Plant of Immortality in Near Eastern narratives known under the general heading of Kalīla wa-Dimna is the same thing, so to speak, as the mental process of reading the book known as Kalīla wa-Dimna, since this book, viewed as a prototypical book, is the message of the book itself.

§2. I highlight in particular two points of comparison that can be made in juxtaposing the West European narrative about the Quest for the Holy Grail and the Near Eastern narrative about the Quest for the prototype of the Kalīla wa-Dimna, especially as attested in Arabic sources.

§2a. First, here is my Point A. I note that the primary narrators of the Book of the Grail were supposedly composing the text in a “classical” language, that is, in Latin, while the secondary narrators were composing the text in French—or maybe we should say “recomposing,” not just “composing.” Similarly, as I noted in Part I, the primary narrators of the Kalīla wa-Dimna were supposedly composing the text in the “classical” language of India—that is, in Sanskrit—while a secondary narrator was recomposing the text in Middle Persian—or maybe we should say that this secondary narrator was “translating” the text into Middle Persian. And, in this case, we must keep in mind not only the secondary narrator but also tertiary narrators, who are viewed today as “translators” of the Middle Persian text—now lost—into Arabic and into other languages, including New Persian—which is the language known retrospectively as “classical” Persian.

§2b. Next, here is my Point B. I note that the secondary narrators of the Book of the Holy Grail, who are speaking in French, not in Latin, can talk about themselves and about their world views in the first person, speaking in French.

§3. Similarly, the secondary narrator of the Kalīla wa-Dimna, can talk about himself and about his world views in the first person, speaking in the Middle Persian language, which is then translated by tertiary narrators into Arabic and into other languages, including New Persian. This secondary narrator of the Kalīla wa-Dimna, I should add by way of review, identifies himself as a physician named Borzōy—or Barzūya, to add here the Arabic version of his name—who lived in the era of the Sasanian king Khosrow I Anōshirvān (Kisrā in Arabic), emperor of the Persian Empire from 531 to 579 CE. The Middle Persian text of this secondary narrator, as we learn from the aetiological narratives that accompany the transmitted text of the Kalīla wa-Dimna, was later translated into Arabic by a Persian savant named ‘Abd Allāh ibn al-Muqaffa‘, the date of whose death is traditionally set at 757 CE. And the point is, this secondary narrator named Borzōy or Barzūya can talk about himself and about his world views in the first person, speaking in the Middle Persian language, which is then supposedly translated by tertiary narrators into Arabic and into other languages, including New Persian.

§4. Having made these two points of comparison, I will now elaborate on the parallelisms that we find between the narrative strategies in the stories about the making of the Vulgate Cycle and in the stories about the making of the Kalīla wa-Dimna. Such parallelisms show clearly that the first-person narratives that we see in stories about the making of the Kalīla wa-Dimna are hardly unusual from the standpoint of comparative folklore studies. And there is no justification for simply assuming that the first-person references of the secondary narrator Borzōy or Barzūya are necessarily “historical.” Such an assumption, as I have noted in Part I, pervades the argumentation of François de Blois (1990) in his monograph about the making of the Kalīla wa-Dimna.

§5. The multiform textual transmission of the Kalīla wa-Dimna includes multiple introductions to the actual fables that are narrated, and these introductions are narratives in their own right, aetiologizing as they do the origins of the Kalīla wa-Dimna as a book. In the opinion of de Blois (1990), there were basically two different ways of narrating the story of the making of the Kalīla wa-Dimna. In a “mythological” version, as viewed by de Blois, the physician named Borzōy is sent to India by King Khosrow on a quest to find a wondrous plant that heals mortality and confers immortality. He finds instead the book of fables that becomes ultimately the Kalīla wa-Dimna. In the other version, the mythological aspects of the quest are shaded over: Borzōy, physician that he is, goes to India to find medicinal herbs but brings back the book of fables instead. In both versions, Borzōy supposedly “translates” the book from the language of the Indians—what is Sanskrit to us—into his own language, which is of course Middle Persian.

§6. The “mythological” version of the aetiology of the Kalīla wa-Dimna is deemed to be “spurious” by de Blois (1998:423, summarizing the argumentation of his 1990 monograph), while the demythologized version is by contrast “historical.” In the supposedly demythologized version, we find a first-person statement, as if spoken by Borzōy, where the physician talks about himself and about his world view. This statement, which is missing in an older Syriac version, is deemed by de Blois to be the centerpiece of what he calls the “autobiography” of Borzōy. In his monograph, de Blois consistently uses the term “autobiography” in arguing that this first-person account is the historical core of the genesis of the Kalīla wa-Dimna as we know it. Here is how de Blois translates the first-person statement, and I quote (de Blois 1990:26):

Then, I thought about medicine and realised that a physician cannot give his patient a remedy which would heal his illness to such a degree that he would never again suffer from it, or from any other illness. Seeing that there is no guarantee against the same disease, or an even more serious one, recurring, I came to the conclusion that knowledge of the hereafter is the only thing which brings permanent salvation from all diseases. Thus, I came to hold medicine in contempt and to long for religious knowledge. But when I had come to this conclusion, I felt uncertainty concerning religion. I found nothing in my books of medical learning which could show me which religion was the true one. I found that there are many religions and creeds and that the followers of these creeds differ one from the other. Some inherited their religion from their ancestors, others, adopted it on account of fear and coercion, yet others hoped by means of it to acquire worldly goods, pleasures, and prestige. But every one of them claims that his religion is the true and correct one and that whoever contradicts him lives in error and deception. Concerning the creator and what he created, the beginning and the end of the world, and other questions they have violently different opinions, but every one of them despises, opposes, and attacks the others, I decided to frequent the scholars and leaders in every religious faction and to examine what they teach and stipulate in the hope that perhaps I could learn to distinguish truth from falsehood and attach myself confidently to the truth without having to accept on the authority of others something that I could not know or understand by myself. I pursued this plan, investigated and studied. But I discovered that all of these people merely repeat what was handed down to them. Each one praises his own religion and curses the religion of those who disagree with him. It became clear to me that their conclusions are based on illusions and that their speech is not motivated by a sense of fairness. In not one of them did I find that degree of honesty and rightmindedness which would induce rational persons to accept their words and be satisfied with them.

§7. As I noted, what de Blois sees in this first-person statement is the historical core of what he calls the “autobiography” of Borzōy. What I see instead, however, is a dramatization of Borzōy as a speaker who levels the playing field before telling the fables of the book of Kalīla wa-Dimna. For de Blois, the words of this passage are “rejecting all established religions.” For me, however, these words are not so extreme: rather, the speaker is simply resisting the zero-sum attitude of dualist religions such as Manichaeanism. We may compare the Syriac figure Bōd, who was the reported “translator,” as it were, of an earlier Syriac version of the Kalīla wa-Dimna. It is relevant that Bōd is known to have composed relevant treatises arguing against dualist religious. We know of treatises of his that bear such titles as Discourses on the Faith, Against the Manichaeans, and Against the Marcionites.

§8. As I have argued in Part I, the narratives about the finding of the Kalīla wa-Dimna were designed to explain not only the genesis of such a book but also its multiform reception in the many different societies that were affected by this collection of fables—and that in turn affected the content of these fables.

§9. We see a comparable situation at the beginning of the medieval French book known as the Vulgate Cycle, taking up fourteen pages in the quarto format of Sommer’s edition, which presents a massive prologue to the Estoire del saint Graal. The narrator of this prologue introduces himself in the third person singular, without naming himself at this point, and claiming that he is inspired by the mystery of the Holy Trinity. Then he switches to the first person as he says that he will tell the story of the Holy Grail. At this point the second person of the Trinity, God the Son, appears to our narrator in a dream and speaks to him with a Voice, taking him by the hand and giving him a small book, the size of the palm of a hand. Next, God the Son claims that he wrote this small book himself with his own hand. And now the narrator becomes the reader. The narrator starts reading and finds his own genealogy in what he has read. Then he falls asleep and dreams that he continues reading. In this dream, he is transported by an angel to heaven, where he sees heavenly visions. Returning to earth, he puts the book written by God the Son, Jesus, into a ciborium, right next to consecrated hosts. This way, by a contagion métonymique (L p. 27), the small book becomes part of the corpus Domini. The narrator leaves the ciborium for a moment and, by the time he returns to it, the little book is gone. And he hears a voice declaring that the little book has gone away from this world just as Jesus goes away after he is resurrected. So now we see the ciborium speaking with its own voice, and that voice is the voice of the little book. The narrator is now given a blank book into which to copy the contents of the narrative. He wakes up to find all the ingredients for making a book: pen, ink, parchment, and a pen-knife. So now the narrator becomes a scribe. After this is done, what follows is not an act of transcription but an act of transformation—which must become an act of translation (L p. 33). The narrative must be transformed from Latin to French, and the transformer is identified in the third person singular as Robert de Boron (I 95; L p. 33). So, up to now, everything has been composed in Latin before being written down in French by Robert.

§10. The chronological facts of history about this Robert, who is otherwise known as the composer of two surviving poems (in octosyllabic format) about Arthurian themes, show that he cannot be traced beyond the early thirteenth century. So, the attribution in this narrative to Robert as the translator is at least chronologically fitting.

§11. At the end of the Estoire del saint Graal, however, there is an announcement (I 296, L p. 35) of what will now be added to this Estoire, and the addition will be une Estoire de Merlin. But the person who is doing the adding is “Monsieur Robert”—so all along, it has been “Monsieur Robert” in charge. After all, the language of the narrative as written down has been French, not Latin.

§12. As we start the Estoire de Merlin, there is no prologue. In the middle of the section attributed to Merlin, the narrator who has been narrating from the start finally situates himself as a scribe named Nascien who supposedly wrote down his text in 717 CE. And there is at least an implication of an ancestor named Nascien who lived in the days of King Arthur, supposedly in the 5th century CE (V 318; L p. 36n10). As for the story by Merlin, we see at the beginning of his story (II 18) that he asks a cleric named Blaise to take ink and parchment, and he dictates his story to this cleric, even though he is aware that the title of dictator should be reserved for divinity (L p. 37). And Merlin now tells not only his own story but the whole story of the Grail, starting with Joseph of Arimathia, dating back to Biblical times. By now we have seen three kinds of narrator: a notarius who copies the text, a scriptor who writes a text that can be copied, and a dictator who ‘dictates’ only God’s truth (L p. 37).

§13. Then (II 222) Nascien says in the third person that he added his part of the writing to the writing that Blaise wrote, and so the previous book written by Nascien was supposedly a supplement to the book written by Blaise (L p. 37). But it was still Nascien who wrote things down at the beginning, inspired by the Holy Spirit (II 221-222; we may compare also I 3). Further, as Merlin says to Blaise, what he writes down above and beyond the book of Merlin will still be the book of the Grail (II 28). Blaise puts everything into his book (II 35; L p. 39). Merlin as dictator is replaced by the knights themselves, and Blaise is replaced by royal scribes (L pp. 40-41). In one version of the book of Merlin (II 86 “n.1”), Robert de Boron recurs, saying that ‘I’ will now switch the topic from Arthur to Alan.

§14. In the Book of Lancelot, which is Part III of the Cycle, the narrative gets started without an introduction. The story tells how Lancelot was raised by Nascien (L p. 44), and that is why Lancelot himself is the best narrator of narrations about the Grail. So Arthur commissions Lancelot to narrate to the royal scribes a book that is all his narration (V 190; L p. 44).

§15. In La Queste du Saint Graal, which is Part IV of the Cycle, a new translator comes along: he is Walter (or Gautier) Map. In the transition at the beginning of Part IV, some versions of the multiple text even claim that Part III was already the work of Walter Map (L p. 48). Here again, as we saw before, we are dealing with aetiology, not history: Walter Map died somewhere around 1208-1210. Elsewhere (in Q 279), it is said that Walter Map translated from the Latin out of love for his King Henry (L p. 49).

§16. In La Mort Artu, which is Part V, the figure of Walter Map emerges again: he was already the translator of the Queste, but now he becomes again a translator—this time translating La Mort Artu (L. p. 50). It is said (M 1) that he did it for his King Henry. But Henry, notionally Henri II Plantagenêt, wants something more to express the totality. And Walter does it in the third person (M 1). He makes it all final by finishing the story, though not in the present text. He will finish the whole Lancelot narrative (M 263). So we see here a fiction of exhaustiveness (L p. 52).

§17. The whole story of the Grail is a form of arborescence (L 57). The Grail stems from a mystical Plant of Salvation, a Tree of Life and its branches. We see here one of many metaphors for the Grail—typologically comparable, I would add, to the Plant of Immortality in Near Eastern narratives about the finding of the Kalīla wa-Dimna.

§18. A related metaphor is the genealogical family tree of the heroes. The story begins with a branch from the tree of life, with the fruit still on it (L. 57), that Eve takes with her when she is expelled from paradise. This branch will become the Redemption for the Fall. This branch will be planted in non-paradisiacal land and will flourish (L. 60). The tree now becomes a ship (L. 60-61); or a bed with the Sword of David on it. The carpenters, when they cut the tree, make it bleed.

§19. I bring to an end, for now, my own narration of the aetiological narrations about the Quest for the Grail—but not before making one last observation, again only for now. I end by saying here that the Grail and its prototype, the box, are metaphors of the entire book about the Grail (I follow here the thinking of L. p. 71). The medium, then, of the book is the message that is the book. And what I say here about the book that tells of the Grail is typologically comparable to what I have already said about the book that tells of the Plant of Immortality, which is the same thing, so to speak, as the book known as the Kalīla wa-Dimna.


Bartholomae, C. 1904. Altiranisches Wörterbuch. Strassburg.

Bertels, Y. E. et al., eds. 1960-1971. Ferdowsi: Shāhnāma I-IX. Moscow.

Blois, F. de. 1990. Burzōy’s Voyage to India and the Origin of the Book of Kalīlah wa Dimnah. London.

Blois, F. de. 1998. “Kalīla wa-Dimna.” Encyclopedia of Arabic Literature, ed. J. S. Meisami and P. Starkey, II 423-425.

Chehabi, H. E., and G. Neville, eds. 2015. Erin and Iran: Cultural Encounters between the Irish and the Iranians, ed. H.E. Chehabi and G. Neville, 36-46. Ilex Series 16. Cambridge MA.

Davidson, O. M. 2001.“Some Iranian Poetic Tropes as Reflected in the ‘Life of Ferdowsi’ Traditions.” Philologica et Linguistica: Festschrift für Helmut Humbach, ed. M. G. Schmidt and W. Bisang, Supplement pp. 1–12. Trier.

Davidson, O. M. 2002. “Haft K̲vān.” Encyclopaedia Iranica 11:516–519.

Davidson, O. M. 2005. “Persian/Iranian Epic.” A Companion to Ancient Epic, ed. J. M. Foley, 264–276. Malden.

Davidson, O. M. 2008. Review of Yamamoto 2003. Orientalistische Literaturzeitung 103.3:305–316. Recast in Davidson 2013b:99-112.

Davidson, O. M. 2013a. Poet and Hero in the Persian Book of Kings. 3rd ed. Cambridge MA (2nd ed. Costa Mesa CA 2006; 1st ed. Ithaca NY 1994).

Davidson, O. M. 2013b. Comparative Literature and Classical Persian Poetics. 2nd ed. Cambridge MA (1st ed. Costa Mesa CA 2000).

Davidson, O. M. 2015a. “Aetiologies of the Kalīla wa Dimna as a Mirror for Princes.” Global Medieval: Mirrors for Princes Reconsidered, ed. R. Forster and N. Yavari, 42–57. Ilex Series 15. Cambridge MA.

Davidson, O. M. 2015b. “Parallel Heroic Themes in the Medieval Irish Cattle Raid of Cooley and the Medieval Persian Book of Kings,” Erin and Iran:Cultural Encounters between the Irish and the Iranians, ed. H. E. Chehabi and G. Neville, 36–46. Ilex Series 16. Cambridge MA.

Edgerton, F. 1924. The Pañcatantra Reconstructed. I-II. American Oriental Series 2-3. New Haven.

Forster, R., and N. Yavari, eds. 2015. Global Medieval: Mirrors for Princes Reconsidered. Ilex Series 15. Cambridge MA.

Hanson, K., and P. Kiparsky 1997. “The Nature of Verse and its Consequences for the Mixed Form.” Prosimetrum: Crosscultural Perspectives on Narrative in Prose and Verse,  ed. J. Harris and K. Reichl, 17–44. Cambridge and Rochester NY.

Harris, J., and K. Reichl, eds. 1997. Prosimetrum: Crosscultural Perspectives on Narrative in Prose and Verse. Cambridge and Rochester NY.

Leupin, A. 1982. Le Graal et la littérature: Étude sur la vulgate arthurienne en prose. Paris.

McLuhan, M. 1964. Understanding Media: The Extensions of Man. New York. Reprint edition 1994, with an introduction by L. Lapham. Cambridge MA.

Nagy, G. 1996. Poetry as Performance: Homer and Beyond. Cambridge. http://nrs.harvard.edu/urn-3:hul.ebook:CHS_Nagy.Poetry_as_Performance.1996.

Nagy, G. 2009. “Hesiod and the Ancient Biographical Traditions.” Brill’s Companion to Hesiod, ed. F. Montanari, A. Rengakos, and C. Tsagalis, 271–311. Leiden. http://nrs.harvard.edu/urn-3:hlnc.essay:Nagy.Hesiod_and_the_Ancient_Biographical_Traditions.2009.

Nagy, G. 2011. “Diachrony and case of Aesop. Classics @ Issue 9 (Defense Mechanisms in Interdisciplinary Approaches to Classical Studies and Beyond). http://nrs.harvard.edu/urn-3:hlnc.essay:Nagy.Diachrony_and_the_Case_of_Aesop.2011.

Ramanujan, A. K. 1997. A Flowering Tree and Other Oral Tales from India (ed. and with preface by S. Blackburn and A. Dundes). Berkeley.

Sommer, H.-O. 1908-1916. The Vulgate Versian of the Arthurian Romances. I-VII. Washington DC.

Witzel, M. 1997. “Saramā and the Paṇis: Origins of Prosimetric Exchange in Archaic India.” Prosimetrum: Crosscultural Perspectives on Narrative in Prose and Verse, ed. J. Harris and K. Reichl, 387-409. Cambridge and Rochester NY.

Yamamoto, K. 2003. The Oral Background of Persian Epics: Storytelling and Poetry. Brill Studies in Middle Eastern Literatures 26. Leiden.

Ziolkowski, J. 1993. Talking animals: medieval Latin beast poetry, 750-1150. Philadelphia.



[1] de Blois 1990:1. Only later, at p. 12n1, does de Blois inform the reader that this Syriac Qalīlag w Damnag was composed in verse.

[2] de Blois 1990:2 cites the edition of Giuseppe Simone Assemani (Rome 1725) 219-220.

[3] Assemani p. 219 n2, via de Blois 1990:2n9.

[4] de Blois 1990:2.

[5] For the Avestan form, see Bartholomae 1904 column 919. I agree with de Blois (1990:3) when he writes: “c.f. [sic] budhaḥ, ‘wise’.” (Actually, de Blois writes ““c.f.” instead of “cf.” everywhere (e.g. at p. 2 n5, etc.).

[6] de Blois 1990:5.

[7] de Blois 1990:5.

[8] de Blois 1990:5.

[9] de Blois 1990:5. All the remaining quotations in this list, are from the same page of de Blois.

[10] de Blois 1990:5.

[11] Davidson 2013b [2000]:46. See also Davidson 2013:a [2006]:38, 45.

[12] de Blois 1990:6. All the remaining quotations in this list come from this same page of de Blois and from the next page.

[13] de Blois 1990:8 notes that in the “older Syriac” version, the Sage is called a dēnīg, which is a Middle Persian word; then, in the Arabic versions, the Sage is nāsik. But in fragments 299 and 300 of Rōdaki, which show poeticized versions of the fable of the Ring-Dove, the word is dīnī (p. 8n2).

[14] Nagy 2011 §122, following Witzel 1997, especially p. 388.

[15] Nagy 2011 §122, following Hanson and Kiparsky 1997:37.

[16] Pañcatantra (I pp. 341-342; II p. 381, ed. Edgerton 1924); via de Blois 1990:7n3.

[17] de Blois 1990:9.

[18] de Blois 1990:9.

[19] Ramanujan 1997:245.

[20] de Blois 1990:10

[21] de Blois 1990:5.

[22] de Blois 1990:5.

[23] de Blois 1990:5.

[24] de Blois 1990:6.

[25] de Blois 1990:3.

[26] de Blois 1990:3.

[27] de Blois 1990:3.

[28] de Blois 1990:3.

[29] de Blois 1990:4.

Leave a Reply