The history of the Nights is extremely complex and modern scholars have made many attempts to untangle the story of how the collection as it currently exists came about. Robert Irwin summarises their findings: "In the 1880s and 1890s a lot of work was done on the Nights by [the scholar] Zotenberg and others, in the course of which a consensus view of the history of the text emerged. Most scholars agreed that the Nights was a composite work and that the earliest tales in it came from India and Persia. At some time, probably in the early 8th century, these tales were translated into Arabic under the title Alf Layla, or 'The Thousand Nights'. This collection then formed the basis of The Thousand and One Nights. The original core of stories was quite small. Then, in Iraq in the 9th or 10th century, this original core had Arab stories added to it – among them some tales about the Caliph Harun al-Rashid. Also, perhaps from the 10th century onwards, previously independent sagas and story cycles were added to the compilation [...] Then, from the 13th century onwards, a further layer of stories was added in Syria and Egypt, many of these showing a preoccupation with sex, magic or low life. In the early modern period yet more stories were added to the Egyptian collections so as to swell the bulk of the text sufficiently to bring its length up to the full 1,001 nights of storytelling promised by the book’s title."
Possible Indian origins
Devices found in Sanskrit literature such as frame stories and animal fables are seen by some scholars as lying at the root of the conception of the Nights. Indian folklore is represented in the Nights by certain animal stories, which reflect influence from ancient Sanskrit fables. The influence of the Panchatantra and Baital Pachisi is particularly notable. The Jataka Tales are a collection of 547 Buddhist stories, which are for the most part moral stories with an ethical purpose. The Tale of the Bull and the Ass and the linked Tale of the Merchant and his Wife are found in the frame stories of both the Jataka and the Nights.
It is possible that the influence of the Panchatantra is via a Sanskrit adaptation called the Tantropakhyana. Only fragments of the original Sanskrit form of this work exist, but translations or adaptations exist in Tamil, Lao, Thai and Old Javanese. The frame story is particularly interesting, as it follows the broad outline of a concubine telling stories in order to maintain the interest and favour of a king — although the basis of the collection of stories is from the Panchatantra — with its original Indian setting.
Persian prototype: Hazār Afsān
The earliest mentions of the Nights refer to it as an Arabic translation from a Persian book, Hazār Afsān (or Afsaneh or Afsana), meaning "The Thousand Stories". In the 10th century Ibn al-Nadim compiled a catalogue of books (the "Fihrist") in Baghdad. He noted that the Sassanid kings of Iran enjoyed "evening tales and fables". Al-Nadim then writes about the Persian Hazār Afsān, explaining the frame story it employs: a bloodthirsty king kills off a succession of wives after their wedding night; finally one concubine had the intelligence to save herself by telling him a story every evening, leaving each tale unfinished until the next night so that the king would delay her execution. In the same century Al-Masudi also refers to the Hazār Afsān, saying the Arabic translation is called Alf Khurafa ("A Thousand Entertaining Tales") but is generally known as Alf Layla ("A Thousand Nights"). He mentions the characters Shirazd (Scheherazade) and Dinazad. No physical evidence of the Hazār Afsān has survived so its exact relationship with the existing later Arabic versions remains a mystery. Apart from the Scheherazade frame story, several other tales have Persian origins, although it is unclear how they entered the collection. These stories include the cycle of "King Jali'ad and his Wazir Shimas" and "The Ten Wazirs or the History of King Azadbakht and his Son" (derived from the 7th-century Persian Bakhtiyarnama).
In the 1950s, the Iraqi scholar Safa Khulusi suggested (on internal rather than historical evidence) that the Persian writer Ibn al-Muqaffa' may have been responsible for the first Arabic translation of the frame story and some of Persian stories later incorporated into the Nights. This would place genesis of the collection in the eighth century C.E.
In the mid-20th century, the scholar Nabia Abbott found a document with a few lines of an Arabic work with the title The Book of the Tale of a Thousand Nights, dating from the 9th century. This is the earliest known surviving fragment of the Nights. The first reference to the Arabic version under its full title The One Thousand and One Nights appears in Cairo in the 12th century. Professor Dwight Reynolds describes the subsequent transformations of the Arabic version:
Some of the earlier Persian tales may have survived within the Arabic tradition altered such that Arabic Muslim names and new locations were substituted for pre-Islamic Persian ones, but it is also clear that whole cycles of Arabic tales were eventually added to the collection and apparently replaced most of the Persian materials. One such cycle of Arabic tales centres around a small group of historical figures from 9th-century Baghdad, including the caliph Harun al-Rashid (died 809), his vizier Jafar al-Barmaki (d.803) and the licentious poet Abu Nuwas (d. c. 813). Another cluster is a body of stories from late medieval Cairo in which are mentioned persons and places that date to as late as the thirteenth and fourteenth centuries.
Two main Arabic manuscript traditions of the Nights are known: the Syrian and the Egyptian. The Syrian tradition includes the oldest manuscripts; these versions are also much shorter and include fewer tales. It is represented in print by the so-called Calcutta I (1814–1818) and most notably by the Leiden edition (1984), which is based above all on the Galland manuscript. It is believed to be the purest expression of the style of the mediaeval Arabian Nights.
Texts of the Egyptian tradition emerge later and contain many more tales of much more varied content; a much larger number of originally independent tales have been incorporated into the collection over the centuries, most of them after the Galland manuscript was written, and were being included as late as in the 18th and 19th centuries, perhaps in order to attain the eponymous number of 1001 nights. The final product of this tradition, the so-called Zotenberg Egyptian Recension, does contain 1001 nights and is reflected in print, with slight variations, by the editions known as the Bulaq (1835) and the Macnaghten or Calcutta II (1839–1842).
All extant substantial versions of both recensions share a small common core of tales, namely:
- The Merchant and the Demon.
- The Fisherman and the Jinni.
- The Story of the Porter and the Three Ladies.
- The Hunchback cycle.
- The Story of the Three Apples, enframing the Story of Nur al-Din and Shams al-Din
- The Story of Nur al-Din Ali and Anis al-Jalis
- The Story of Ali Ibn Baqqar and Shams al-Nahar, and
- The Story of Qamar al-Zaman.
The texts of the Syrian recension do not contain much beside that core. It is debated which of the Arabic recensions is more "authentic" and closer to the original: the Egyptian ones have been modified more extensively and more recently, and scholars such as Muhsin Mahdi have suspected that this may have been caused in part by European demand for a "complete version"; but it appears that this type of modification has been common throughout the history of the collection, and independent tales have always been added to it.
The first European version (1704–1717) was translated into French by Antoine Galland from an Arabic text of the Syrian recension and other sources. This 12-volume work, Les Mille et une nuits, contes arabes traduits en français ("Thousand and one nights, Arab stories translated into French"), included stories that were not in the original Arabic manuscript. "Aladdin's Lamp" and "Ali Baba and the Forty Thieves" (as well as several other, lesser known tales) appeared first in Galland's translation and cannot be found in any of the original manuscripts. He wrote that he heard them from a Syrian Christian storyteller from Aleppo, a Maronite scholar whom he called "Hanna Diab." Galland's version of the Nights was immensely popular throughout Europe, and later versions were issued by Galland's publisher using Galland's name without his consent.
As scholars were looking for the presumed "complete" and "original" form of the Nights, they naturally turned to the more voluminous texts of the Egyptian recension, which soon came to be viewed as the "standard version". The first translations of this kind, such as that of Edward Lane (1840, 1859), were bowdlerized. Unabridged and unexpurgated translations were made, first by John Payne, under the title The Book of the Thousand Nights and One Night (1882, nine volumes), and then by Sir Richard Francis Burton, entitled The Book of the Thousand Nights and a Night (1885, ten volumes) – the latter was, according to some assessments, partially based on the former, leading to charges of plagiarism. In view of the sexual imagery in the source texts (which Burton even emphasized further, especially by adding extensive footnotes and appendices on Oriental sexual mores) and the strict Victorian laws on obscene material, both of these translations were printed as private editions for subscribers only, rather than published in the usual manner. Burton's original 10 volumes were followed by a further six entitled The Supplemental Nights to the Thousand Nights and a Night, which were printed between 1886 and 1888. It has, however, been criticized for its "archaic language and extravagant idiom" and "obsessive focus on sexuality" (and has even been called an "eccentric ego-trip" and a "highly personal reworking of the text").
Later versions of the Nights include that of the French doctor J. C. Mardrus, issued from 1898 to 1904. It was translated into English by Powys Mathers, and issued in 1923. Like Payne's and Burton's texts, it is based on the Egyptian recension and retains the erotic material, indeed expanding on it, but it has been criticized for inaccuracy.
A notable recent version, which reverts to the Syrian recension, is a critical edition based on the 14th or 15th-century Syrian manuscript in the Bibliothèque Nationale, originally used by Galland. This version, known as the Leiden text, was compiled in Arabic by Muhsin Mahdi (1984) and rendered into English by Husain Haddawy (1990). Mahdi argued that this version is the earliest extant one (a view that is largely accepted today) and that it reflects most closely a "definitive" coherent text ancestral to all others that he believed to have existed during the Mamluk period (a view that remains contentious). Still, even scholars who deny this version the exclusive status of "the only real Arabian Nights" recognize it as being the best source on the original style and linguistic form of the mediaeval work and praise the Haddawy translation as "very readable" and "strongly recommended for anyone who wishes to taste the authentic flavour of those tales". An additional second volume of Arabian nights translated by Haddawy, composed of popular tales not present in the Leiden edition, was published in 1995.
In 2008 a new English translation was published by Penguin Classics in three volumes. It is translated by Malcolm C. Lyons and Ursula Lyons with introduction and annotations by Robert Irwin. This is the first complete translation of the Macnaghten or Calcutta II edition (Egyptian recension) since Burton's. It contains, in addition to the standard text of 1001 Nights, the so-called "orphan stories" of Aladdin and Ali Baba as well as an alternative ending to The seventh journey of Sindbad from Antoine Galland's original French. As the translator himself notes in his preface to the three volumes, "[N]o attempt has been made to superimpose on the translation changes that would be needed to 'rectify' ... accretions, ... repetitions, non sequiturs and confusions that mark the present text," and the work is a "representation of what is primarily oral literature, appealing to the ear rather than the eye". The Lyons translation includes all the poetry (in plain prose paraphrase) but does not attempt to reproduce in English the internal rhyming of some prose sections of the original Arabic. Moreover, it streamlines somewhat and has cuts. In this sense it is not, as claimed, a complete translation.
Scholars have assembled a timeline concerning the publication history of The Nights:
- One of the oldest Arabic manuscript fragments from Syria (a few handwritten pages) dating to the early 9th century. Discovered by scholar Nabia Abbott in 1948, it bears the title Kitab Hadith Alf Layla ("The Book of the Tale of the Thousand Nights") and the first few lines of the book in which Dinazad asks Shirazad (Scheherazade) to tell her stories.
- 10th century: Mention of Hazār Afsān in Ibn al-Nadim's "Fihrist" (Catalogue of books) in Baghdad. He attributes a pre-Islamic Sassanian Persian origin to the collection and refers to the frame story of Scheherazade telling stories over a thousand nights to save her life. However, according to al-Nadim, the book contains only 200 stories. Curiously, al-Nadim also writes disparagingly of the collection's literary quality, observing that "it is truly a coarse book, without warmth in the telling".
- 10th century: Reference to The Thousand Nights, an Arabic translation of the Persian Hazār Afsān ("Thousand Stories"), in Muruj Al-Dhahab (The Meadows of Gold) by Al-Masudi.
- 11th century: Mention of The Nights by Qatran Tabrizi in the following couplet in Persian:
- هزار ره صفت هفت خوان و رويين دژ
- فرو شنيدم و خواندم من از هزار افسان
- A thousand times, accounts of Rouyin Dezh and Haft Khān
- I heard and read from Hazār Afsān (literally Thousand Fables)
- 12th century: A document from Cairo refers to a Jewish bookseller lending a copy of The Thousand and One Nights (this is the first appearance of the final form of the title).
- 14th century: Existing Syrian manuscript in the Bibliothèque Nationale in Paris (contains about 300 tales).
- 1704: Antoine Galland's French translation is the first European version of The Nights. Later volumes were introduced using Galland's name though the stories were written by unknown persons at the behest of the publisher wanting to capitalize on the popularity of the collection.
- 1706: An anonymously translated version in English appears in Europe dubbed the "Grub Street" version. This is entitled The Arabian Nights' Entertainment — the first known use of the common English title of the work.
- 1768: first Polish translation, 12 volumes. Based, as many European on the French translation.
- 1775: Egyptian version of The Nights called "ZER" (Hermann Zotenberg's Egyptian Recension) with 200 tales (no surviving edition exists).
- 1814: Calcutta I, the earliest existing Arabic printed version, is published by the British East India Company. A second volume was released in 1818. Both had 100 tales each.
- Early 19th century: Modern Persian translations of the text are made, variously under the title Alf leile va leile, Hazār-o yek šab (هزار و یک شب), or, in distorted Arabic, Alf al-leil. One early extant version is that illustrated by Sani al-Molk (1814–1866) for Mohammad Shah Qajar.
- 1825–1838: The Breslau/Habicht edition is published in Arabic in 8 volumes. Christian Maxmilian Habicht (born in Breslau, Kingdom of Prussia, 1775) collaborated with the Tunisian Murad Al-Najjar and created this edition containing 1001 stories. Using versions of The Nights, tales from Al-Najjar, and other stories from unknown origins Habicht published his version in Arabic and German.
- 1842–1843: Four additional volumes by Habicht.
- 1835: Bulaq version: These two volumes, printed by the Egyptian government, are the oldest printed (by a publishing house) version of The Nights in Arabic by a non-European. It is primarily a reprinting of the ZER text.
- 1839–1842: Calcutta II (4 volumes) is published. It claims to be based on an older Egyptian manuscript (which was never found). This version contains many elements and stories from the Habicht edition.
- 1838: Torrens version in English.
- 1838–1840: Edward William Lane publishes an English translation. Notable for its exclusion of content Lane found "immoral" and for its anthropological notes on Arab customs by Lane.
- 1882–1884: John Payne publishes an English version translated entirely from Calcutta II, adding some tales from Calcutta I and Breslau.
- 1885–1888: Sir Richard Francis Burton publishes an English translation from several sources (largely the same as Payne). His version accentuated the sexuality of the stories vis-à-vis Lane's bowdlerized translation.
- 1889–1904: J. C. Mardrus publishes a French version using Bulaq and Calcutta II editions.
- 1973: First Polish translation based on the original language edition, but compressed 12 volumes to 9, by PIW.
- 1984: Muhsin Mahdi publishes an Arabic edition which he claims is faithful to the oldest Arabic versions surviving (primarily based on the Syrian manuscript in the Bibliothèque Nationale in combination with other early manuscripts of the Syrian branch).
- 1986–1987: French translation by Arabist René R. Khawam
- 1990: Husain Haddawy publishes an English translation of Mahdi.
- 2008: New Penguin Classics translation (in three volumes) by Malcolm C. Lyons and Ursula Lyons of the Calcutta II edition