The Panchatantra Was Put Into Pahlavi for Persia's King Khosrow I in the 6th Century and Into Arabic as Kalila wa Dimna in the 8th, and That Chain of Translations Carried Its Fables to English by 1570

The Panchatantra, a Sanskrit collection of animal fables meant to teach princes how to live and rule, is the Indian source of the Arabic book Kalila wa Dimna. The Sanskrit original is lost. What survives outside India is a chain of translations: a Pahlavi version made in Sasanian Persia in the 6th century (also lost), then a Syriac book of about 570 and an Arabic book of the 8th century, both made from it.

Pahlavi is the written Middle Persian of Sasanian Iran. The Pahlavi book is gone, but Syriac and Arabic books made from it survive and are the evidence for what follows. Keith-Falconer opened his 1885 history of the tales by saying that few books, ancient or modern, have been so widely circulated or rendered into so many languages.

A book of five books for three slow princes

The book introduces itself with a story. In Ryder’s 1925 English translation, a king called Immortal-Power has three sons who will not study. A counselor named Keen tells him that grammar alone takes twelve years and that a short digest would serve better. The Brahmin Vishnusharman, who says he is eighty, promises to make the princes masters of practical life in six months and composes five books for them. Ryder’s titles for the five are The Loss of Friends, The Winning of Friends, Crows and Owls, Loss of Gains and Ill-considered Action.

The Encyclopedia of Religion classes the Panchatantra as niti-shastra (the science of right conduct) and partly as arthashastra (the science of statecraft), the practical knowledge a king needs to run his kingdom. Most of its actors are animals, and stories sit inside stories. A character in one tale recites a verse, someone asks for an explanation, and the explanation becomes the next fable. The Encyclopedia gives the tortoise and the two geese as a typical case: the geese carry the tortoise through the air on a stick, he opens his mouth to answer people below, and he falls.

Book One of Ryder’s translation opens with a lion and a bull who are friends until a jackal parts them. The two jackals of that book are Karataka and Damanaka in Sanskrit. The Encyclopedia of Religion says the Arabic title Kalila wa Dimna contains their Arabicized names.

Nobody knows who wrote it or exactly when

The sources put the Sanskrit original anywhere between 500 BCE and the 6th century CE.

• Encyclopaedia Iranica says the story cycle originated in India between 500 and 100 BCE.
• The Encyclopedia of Religion says it was compiled by an unknown author before the 6th century CE, possibly as early as the 4th.
• Ryder reports that Johannes Hertel, the editor of the Sanskrit text he translated, believed the original was composed in Kashmir about 200 BCE.
• Keith-Falconer says only that the collection passed into Persia not later than 570 CE.

Dagmar Riedel, writing in Iranica, adds that the Syriac and Arabic translations, made from the lost Pahlavi, give the only upper date for the lost Sanskrit original, a point Theodor Benfey was the first to make. She notes that most Sanskrit scholars still assume one original compiled by one author, though nothing is known about him. The book’s own introduction names Vishnusharman, and Ryder’s translation reproduces it.

The Sanskrit texts that do survive are later. Ryder says the one he translated dates from 1199 CE.

The physician Burzoe and the Pahlavi translation

Encyclopaedia Iranica describes Burzoe (Borzuya in Persian) as a physician at the court of Khosrow I, who reigned from 531 to 579, and as responsible for the Pahlavi translation from Sanskrit. Iranica’s article on him follows the introduction in the Arabic book, where Borzuya tells his own story. It says his father was a soldier and his mother came from a prominent religious family, that he went to school at seven and then studied medicine, that he treated patients without charge, and that his doubts about religion sent him to India. When he returned he brought back Kalila and Dimna and several other works.

Two accounts of the trip survive, and the sources treat them differently. The Arabic preface, as Keith-Falconer summarises it, has King Khosrow hear of a book in India that holds every kind of instruction and tell his vizier Buzurjmihr to find a man who knows Indian and Persian. The vizier picks Burzoe, who has trouble in India, makes friends with an Indian who helps him, and translates by day and night in case the Indian king asks for the book back. Keith-Falconer calls the whole story one embellished with much pleasing fiction.

The Shahnameh tells it differently. In Iranica’s summary, Burzoe reads in Indian books of a mountain plant that revives the dead and asks Anushirvan (Khosrow I) for leave to find it. An ascetic tells him that the plant is the word, the mountain is learning and the dead are the ignorant, and that the remedy is a book called Kalila in a treasure chamber. In this version the minister Bozorgmehr translates it into Pahlavi. Khaleghi-Motlagh calls this tale apocryphal.

Scholars disagree on how much can be trusted. Francois de Blois, in Iranica’s account, concluded that Burzoe’s authorship of the Middle Persian version is historically possible because it can neither be proved nor disproved. Khaleghi-Motlagh judges the attribution to the vizier Bozorgmehr a late Sasanian fabrication, since there is no evidence he knew Sanskrit. He thinks Burzoe may have gone to India, learned Sanskrit and brought the book back while handing the Pahlavi translation to someone else.

The date of the Pahlavi version is also an estimate. The Encyclopedia of Religion puts it at about 550. Keith-Falconer dates the Syriac made from it to about 570, so the Pahlavi was in existence by then.

The Syriac copy found at Mardin

The old Syriac translation was a lost book until 1870. Keith-Falconer tells the story from Benfey’s introduction to the 1876 edition. The medieval Syriac writer Ebed-Jesu, in a catalogue Assemani dated to between 1290 and 1318, credited a Bud, called periodeutes, who oversaw Christians in India and Persia and lived about 570, with translating Kalilag and Damnag. Gustav Bickell, remembering this, wrote to Benfey while Catholic bishops gathered at Rome in the summer of 1870. Benfey asked friends there, and Ignazio Guidi heard from a Chaldaean bishop that he had seen a Syriac Kalilah and Dimnah in the episcopal library at Mardin. Professor Socin, who was travelling in the region, found it but could not buy it and had it copied. The transcript reached Basel in April 1871 and was bought for the University Library at Gottingen. Bickell and Benfey published it in 1876.

The Mardin manuscript lacks its first and last leaves and omits several chapters found in Arabic, including the Burzoe biography. Keith-Falconer, following Benfey, concluded that this Syriac book was not made from the Arabic. Iranica’s article on Ibn al-Muqaffa calls the old Syriac version pre-Islamic and crucial to any attempt to form an idea of the Middle Persian original.

A note in the Gottingen transcript, whose author Keith-Falconer says is uncertain, dates the manuscript to the year 1837 of the Greeks, which he converts to 1526 CE. That fits Riedel’s statement that the oldest copies of the older Syriac translation date to the early 16th century.

Ibn al-Muqaffa and the Arabic Kalila wa Dimna

The Arabic version is the one that spread. Iranica names its translator as Abd-Allah Ibn al-Muqaffa, a chancery secretary and Arabic prose writer born in Fars about 721 who died in Basra about 757. He added a prologue of his own and may have added four stories. Keith-Falconer dates the translation to about 750. From his Arabic text, Iranica says, come all later Arabic versions and one of two Syriac versions, along with the medieval Greek, Persian, Hebrew, Latin and Castilian ones.

Latham’s Iranica article adds a warning about the manuscripts. Ibn al-Muqaffa’s own text is not among them, and the oldest dated Arabic copy was written almost five centuries after his death. Riedel dates it to July 1221 and names it as a manuscript in the Suleymaniye Library in Istanbul. The surviving Arabic copies differ so widely, Latham says, quoting de Blois, that what we have is not strictly his translation but a variety of Arabic texts derived from it.

The Arabic book also drew a medieval complaint. Al-Biruni, in his book on India as translated by Sachau, wrote that he wished he could translate “the book Pañcatantra, known among us as the book of Kalila and Dimna”. He accused Ibn al-Muqaffa of adding the Barzoya chapter to plant doubts in people of weak belief and prepare them for Manichaeism. Nöldeke, in 1912, thought the passage on the uncertainty of religions suited Ibn al-Muqaffa, who had given up his ancestral religion for Islam late in life, better than Burzoe. Khaleghi-Motlagh disagrees and attributes both the doubt and the changes to Burzoe, adding that Burzoe and Ibn al-Muqaffa both reworked the book. Nöldeke likewise calls Ibn al-Muqaffa an adapter, not a mere translator.

Sachau’s preface adds that the fables were translated again for the Buyid princes who ruled western Persia and Iraq between 932 and 1055. He groups the work with the Indian learning that reached Baghdad, such as Brahmagupta’s astronomy, covered in Brahmagupta and the rules for zero.

Persian, Greek, Hebrew, Latin and English

Iranica’s Riedel traces the next steps. The vizier Bal’ami wrote a Persian prose version for the Samanid ruler Nasr ibn Ahmad (r. 914 to 943), and the poet Rudaki turned it into verse. Both are lost except for fragments of Rudaki. Two 12th-century Persian versions went back to Ibn al-Muqaffa’s Arabic. One, by Muhammad al-Bukhari, was written for a Zangid ruler of Mosul, and the other, by Abu’l-Ma’ali Nasr-Allah Munshi, for the Ghaznavid sultan Bahramshah. Their oldest dated copies are from 1149 and 1156, which are older than the oldest dated Arabic copy.

Around the Mediterranean, Riedel writes, the Arabic book was translated between the 11th and 13th centuries into Greek, Hebrew and Old Spanish. Keith-Falconer gives the Greek version as made by Symeon Seth about 1080 and the Spanish one as written in 1251. The Hebrew version, usually credited to Rabbi Joel, became through John of Capua’s Latin the source of every European vernacular translation until the 1690s, and Keith-Falconer places that Latin text between 1263 and 1278.

The first English version appeared in 1570. Thomas North translated an Italian version by Anton Francesco Doni as The Moral Philosophy of Doni. The Encyclopedia of Religion counts that English text as seventh in line from the original, more than a thousand years after the Panchatantra was composed.

The fables return to India

The book also went back. Iranica’s O’Kane writes that Husayn Va’iz Kashefi’s Anwar-e Soheyli, written at the end of the 15th century, became the most popular Persian version. Abu’l-Fazl Allami revised it for the Mughal emperor Akbar as the Iyar-e Danesh, made in 1587 to 1588, which simplified the style and added more stories of Indian origin. Keith-Falconer dates its completion to 1591, so the two sources differ by a few years. The court that commissioned it is described in Akbar and the machinery of the Mughal state. O’Kane names the Chester Beatty Library’s Iyar-e Danesh, about 1595, as one of the finest illustrated copies made for Akbar and Jahangir.

The Metropolitan Museum of Art holds folios from a Kalila wa Dimna of the second quarter of the 16th century, attributed to Gujarat and probably based on an Egyptian original (accession 1981.373.23 for one folio).

Riedel names the literary sources of the Panchatantra that Sanskrit scholars agree on as the technical literature of ethics and statecraft, such as the Arthashastra attributed to Kautilya, and the popular story-telling tradition. The site’s article on Chanakya and the Arthashastra covers the first of these. In 1924 Franklin Edgerton published The Panchatantra Reconstructed, an attempt to rebuild the lost Sanskrit text from the principal surviving versions, including the offshoots of the Pahlavi translation.