The Arabian Nights - A Companion, page 7
The debate about the origins and early form of the Nights was not dependent only on evidence found within the manuscripts themselves. Other medieval works referred to the existence of the Nights or something very like the Nights in the Middle Ages. Al-Mas‘udi (896–956) wrote a delightfully rambling history entitled Muruj al-Dhahab, or ‘Meadows of Gold’. In a digression on stories, he had occasion to remark that there
are collections of stories which have been passed on to us translated from the Persian, Hindu and Greek languages. We have discussed how these were composed, for example the Hazar Afsaneh. The Arabic translation is Alf Khurafa (‘A Thousand Entertaining Tales’) . . . This book is generally referred to as Alf Layla (‘A Thousand Nights’). It is the story of a king, a vizier, the daughter of the vizier and the slave of the latter. These last two are called Shirazad and Dinazad. There are also similar works such as The Book of Ferzeh and Simas which contains anecdotes about the kings of India and their viziers. There is also The Book of Sindibad and other collections of the same type.9
Al-Mas‘udi’s observations are supported in general terms by Ibn al-Nadim (who died around the year 990). Ibn al-Nadim was a bookseller and the compiler of a catalogue of all the books that were known to have been written up to his own time, called the Kitah al-Fihrist. According to Ibn al-Nadim, the writing and collecting of entertaining stories (which it is clear he does not rate very highly) first became fashionable in pre-Islamic Sassanian Persia:
The first book to be written with this content was the book Hazar Afsan which means ‘A Thousand Stories’. The basis for this [name] was that one of their kings used to marry a woman, spend a night with her and kill her the next day. Then he married a concubine of royal blood who had intelligence and wit. She was called Shahrazad and when she came to him she would begin a story, but leave off at the end of the night, which induced the king to ask for it the night following. This happened to her for a thousand nights, during which time he [the king] had intercourse with her until because of him she was granted a son, whom she showed to him, informing him of the trick played upon him. Then, appreciating her intelligence, he was well disposed towards her and kept her alive. The king had a head of the household named Dinar Zad who was in league with her in this matter.
Ibn al-Nadim says that, although Hazar Afsan means ‘A Thousand Nights’, there were only about 200 stories in the collection, and he adds that ‘it is truly a coarse book, without warmth in the telling’.10 Elsewhere in the Fihrist, when he lists the ‘Names of the Books of the Byzantines about Evening Stories, Histories, Fables, and Proverbs’, he includes ‘Shahriyar the King and the Reason for his Marrying Shahrazad the Storyteller’.11
Then there is evidence from the Geniza (a medieval Egyptian Jewish archive to be discussed later). The Geniza contains a fragmentary record of loans made by a twelfth-century Jewish bookseller and notary in Cairo. One of the books lent out was The Thousand and One Nights.12 (Here for the first time we get the title in its final form.) That these stories were circulating in Egypt at about this time is confirmed by al-Maqrizi, an Egyptian historian of the early fifteenth century, who quotes a thirteenth-century Spanish author, Ibn Said, who in turn quotes a certain al-Qurtubi (‘the Cordovan’), to the effect that tales from The Thousand and One Nights were circulating in Fatimid times, that is, in the late eleventh century.13
Finally, in the preface to a late-eighteenth-century Turkish story collection, Phantasms of the Divine Presence, Ali Aziz Efendi the Cretan claims to be translating from, among other sources, Elf Leyle (i.e. ‘The Thousand Nights’) by al-Asma‘i. Ali’s story collection does indeed contain versions of stories that are common to the Arabic Nights, but he provides no supporting evidence that al-Asma‘i, the distinguished ninth-century Basran philologist and companion of the Caliph Harun al-Rashid, did indeed compile such a collection; and in general, scholars have been chary of attributing the Nights to a single author.14
Although such external sources suggested that something like the Nights was circulating in the ninth or tenth century, it took scholars a long time to identify any text or fragment of the text which could have been written earlier than the thirteenth century. However, an important discovery was made after the Second World War. This was a couple of fragmentary sheets of paper, which had been preserved in Egypt’s dry air, dating from the ninth century. The fragment was acquired by the University of Chicago and published by the distinguished papyrologist Nabia Abbott. It is one of the oldest surviving literary manuscripts from the Arab world, and by great good fortune the fragment which was preserved bears the title Kitab Hadith Alf Layla, or ‘The Book of the Tale of the Thousand Nights’, plus some fifteen lines of the opening of the book, in which Dinazad asks Shirazad, if she is not asleep, to tell her a story and give ‘examples of the excellencies and shortcomings, the cunning and stupidity, the generosity and avarice, and the courage and cowardice that are in man, instinctive or acquired, or pertain to his distinctive characteristics or to courtly manners, Syrian or Bedouin’.15 Obviously, the title is different, and there is no reference to the misfortunes of Shahriyar and Shahzaman in this opening fragment of the frame story; yet, equally obviously, here we have a prototype version of The Thousand and One Nights.
It seems probable from all the above that the Persian Hazar Afsaneh was translated into Arabic in the eighth or early ninth century and was given the title Alf Khurafa before being subsequently retitled Alf Layla. However, it remains far from clear what the connection is between this fragment of the early text and the Nights stories as they have survived in later and fuller manuscripts, nor how the Syrian manuscripts related to later Egyptian versions. In the absence of a critical text of the Nights, all opinions were at best speculative. Duncan Black Macdonald first formulated the project of producing a critical edition of the Nights in the earliest form which could be deduced from the surviving manuscripts.16 Macdonald was born in Glasgow in 1863. He studied Semitic languages at Glasgow and Berlin and subsequently taught in the United States at the Hartford Theological College. His main academic interest was in Muslim theology and spirituality. His interest in the Nights really developed as a subsidiary to his main interest, for he believed that the stories could be used to illustrate the concerns and imagery of Muslim popular piety. Macdonald himself was particularly interested in the invisible world and in parapsychology.
Macdonald followed Zotenberg’s trail, examining the manuscripts of the Nights, and he began to publish studies on the subject in 1908. Among his achievements in the field of Nights scholarship was the discovery in the Bodleian Library, Oxford, of a unique Arabic manuscript of the story of Ali Baba. (‘Ali Baba’ does not feature in the Bibliothèque nationale manuscript, and some had therefore speculated that Galland himself had made up the story.) Macdonald also demonstrated in devastating detail that Habicht’s ‘Tunisian manuscript’, the basis for the Breslau edition, had never existed, the printed text being based on a variety of manuscript sources. Macdonald also planned to publish an edition of the Galland manuscript, collating it with another early manuscript of Syrian provenance which was preserved in the Vatican. However, this project does not seem to have got very far by the time he died in 1943.
The techniques involved in editing medieval texts, whether western or eastern, are based on those pioneered for the production of editions of the literature of classical antiquity.17 These techniques depend heavily on what may be called the psychopathology of scribal error. The textual critic, working from a number of late and variant copies of a no longer extant original manuscript source, seeks to reconstruct as accurately as possible the appearance of the original manuscript. He does this by establishing a stemma, or hypothetical pedigree, in which some manuscripts are shown to descend from one or several earlier manuscripts. Common errors are crucial in helping to establish the mutual dependence of manuscripts upon a common source; for, while scribal mistakes which are shared between an early and a late manuscript of the same work may be the result of coincidence, this is not likely. It is much more likely that the later manuscript was directly copied from the earlier one (or from a copy of the earlier one), or that the two manuscripts had a common parentage and that they derive their shared errors from a manuscript from which both were copied. Common errors in scribal psychopathology include haplography (writing once what should have been written twice), dittography (writing twice what should have been written once) and saut du même au même (moving to the same word or phrase further down the page). Through the detection of shared errors, a family tree of manuscripts is established (curiously similar to the system devised by the eighteenth-century naturalist Linnaeus for classifying biological organisms).
The textual critic seeks, among other things, to establish which is the earliest manuscript. However, it is important to remember that the earliest manuscript is not necessarily the best, for a late manuscript might easily turn out to be a good copy of a now lost manuscript of an even earlier date. More importantly, the textual critic seeks to establish and date the archetype of the stemma, that is, the hypothetical manuscript from which all the surviving manuscripts with their different chains of transmission of error descend. According to the textual scholar Paul Maas, the archetype is defined as ‘the exemplar from which the first split originated’. The archetype, however, is not necessarily and not often the same as the very first version of the manuscript, and the textual critic may attempt to go beyond the archetype, to divine what the source looked like when it was first written down, with all errors and deliberate interpolations removed (that is, to present a hypothetical constitutio textus).
A stemma gets a bit complicated if it can be demonstrated that cross-contamination has taken place – that is, if more than one manuscript has been consulted when making a copy of the work in question; with the result that one finds a confluence of readings (and scribal errors) deriving from two or more branches of the stemma. Textual criticism is, by its nature, a conjectural science, and the results it produces are often controversial. Several textual critics have noted the suspiciously frequent production by textual scholars of bipartite stemmata. E.J. Kenny has argued that what often happens in such cases is that one group of manuscripts has been treated as the chief chain of descent, while all the rest have been (unjustifiably) lumped together in a single divergent group, even though members of the divergent set are a miscellaneous residue rather than a genuine group with significant common characteristics.
It is worth noting finally that the great pioneers of textual criticism, among them Bentley, Pasquali and Lachman, worked on texts which had, or were presumed to have had, a single author and of which there was once in truth a single original manuscript (perhaps even an autograph written in the writer’s own hand), from which all surviving manuscripts ultimately derived. Moreover, those scholars worked on authors like Homer, Callimachus and Lucretius, whose works were treated with reverence by later generations, with the consequence that copyists were often at considerable pains to transcribe them accurately. When faced with a problematic reading, reverentially careful scribes might even go so far as to compare the manuscript they were copying with another manuscript of the same work (and hence there was a possibility of cross-contamination).
Muhsin Mahdi, a professor of Arabic at Chicago and later at Harvard and at one time a colleague of Nabia Abbott, inherited the project – first conceived of by Macdonald – of editing the Galland manuscript and reconstructing the archetypal manuscript of the Nights, from which the Galland manuscript and all other surviving Syrian and Egyptian manuscripts derived. It involved him in years of back-breaking work, making a word-for-word, diacritical-point-by-diacritical-point comparison of widely dispersed and sometimes hard-to-read Arabic manuscripts. Mahdi began work in 1959, and the impressive outcome of his labours was published in 1984.18
Mahdi took as his base text the three-volume Syrian manuscript which had been used by Galland. He compared it with the other surviving Syrian manuscripts, paying attention to variants and errors. He compared the Syrian family of manuscripts with a parallel, though on the whole later, family of Egyptian manuscripts. Some of the Syrian manuscripts showed signs of contamination from the Egyptian branch; but, by relying mainly on the early versions, Mahdi was able to reconstruct the common ancestor of all the Syrian manuscripts (their archetype or, in Arabic, al-dustur), purged of all later additions and corruptions. He then carried out a similar though more cursory operation examining the mainly later Egyptian manuscripts and posited the existence of a common ancestor for this group too. Finally, he compared the two reconstructed manuscripts – the Syrian and the Egyptian – which were the sources of all surviving manuscripts of the Nights and deduced the form of their common ancestor, the ultimate archetype. It was not possible to go beyond the text of the archetype, but Mahdi was able to offer some plausible speculations about the circumstances of composition of the original or ‘mother’ source (in Arabic, al-nuskha al-umm) from which the archetype derived. According to Mahdi, the ‘mother’ source was produced in Syria, some time in the thirteenth or early fourteenth century, probably not many years before the archetype was in turn copied from it (and the Galland version followed on close behind the archetype). The ‘mother’ source manuscript was in turn based on an earlier version of the Nights, composed in Iraq, but Mahdi did not think it profitable to speculate on what form this Iraqi version would have taken.
For the first time, an authentic medieval text of the Nights has been made readily accessible to Arabists; and, for the first time, it is possible to make detailed criticism of the language, style and narrative technique of the Nights. The style of the dustur archetype (and of the Galland manuscript) displays a remarkable variety, ranging from dialect and common colloquial to a high-flown and very correct classical Arabic. Mahdi argues that this range is the product of design and that the ‘author’ adjusted his language according to the social context of the stories and the rank of his speakers. It is also possible that this variation in language merely reflects a haphazard and unintelligent compilation from diverse sources. Mahdi, however, does not believe that the compiler of the Syrian Nights was unintelligent or acted haphazardly. The stories in the Syrian recension were not thrown together, but have been linked to one another to fit an underlying design. In particular, Mahdi has argued that the exemplary tales that come framed within the main narrative are carefully placed to give a covert message about the fatuity of exemplary tales. In other words, the ‘author’ has used this genre of stories to undermine itself.19
It might also be added in support of Mahdi’s thesis that, if one takes the first 270 to 280 nights as a unit (that is to say, the stories found in the Galland manuscript), it is possible to detect the recurrence of certain common devices and images in them. The crucial device is, of course, the framing one of a person talking to save a life. Sheherazade does it. So do the sheikhs who tell stories to a jinn (or genie) in order to save the merchant’s life, and so do the guests in the house of the three ladies of Baghdad, and so do those who are arrested after the death of the hunchback. Then there is a preoccupation with mutilation (the three one-eyed dervishes; the young men who lost their hands in ‘The Christian Broker’s Tale’ and ‘The Jewish Physician’s Tale’; and the various deformed brothers of the barber). The breaking of an interdiction often sets the story in motion or keeps it moving along. Underground spaces are associated with sex and danger. The ‘author’ has a playful, Shandyesque tendency to promise stories which are in fact never told (for example, the allusions to the Sindibad cycle and to ‘the story of the crocodile’ in ‘King Yunan and Sage Duban’). Then there are the careful enumerations of things to be shopped for, something which features at the beginning of the story of ‘The Porter and the Three Ladies of Baghdad’ and at the beginning of the Hunchback cycle.
However, while it is possible that even more common themes and images can be found in this small core of stories, the existence of a common author or compiler is impossible to prove. Be that as it may, Mahdi contrasts the coherence of this early group of Syrian stories with the ragbag of tales which make up ZER and of which he has written that ‘copyists who missed what [the Syrian compiler] was after and thought the book was like a hole in the ground in which one could dump one story after another regardless of their styles, structures or contradictory aims, disfigured the book’.20 In Mahdi’s opinion, the old idea that the Nights had no original designer and that its structure had always been fluid is based on nothing more than insufficient research.
Mahdi believes that the Galland manuscript incorporates all or very nearly all of what was available in the early Mamluke period, some forty stories related over 282 nights and written down in Syria. Subsequently in Egypt more and more stories were added to meet consumer demand and perhaps to match the number of nights to the title. While Mahdi’s main achievement lies in his critical edition of the Galland manuscript, he has also made some interesting discoveries along the way. Mahdi shows that the Baghdad manuscript, which would have been the sole survivor of an Iraqi tradition of transmission and was allegedly copied by the Lebanese copyist Michael Sabbagh in the early nineteenth century, was a fake. What Sabbagh actually did was put his manuscript together by copying from various Egyptian and Syrian manuscripts of the Nights that he found lying around in Paris. Since he was a good writer, Sabbagh added little touches of his own, and, not having an Arabic original for ‘Aladdin’, he translated a French version of the tale back into Arabic. His ‘discovery’ of this alleged Baghdad manuscript earned him money as well as some academic fame. Similarly, Mahdi has shown that another manuscript, which purported to be the old continuation and completion of the Galland manuscript, was actually a fake perpetrated by a Syrian priest called Chavis, who in fact copied an eighteenth-century Egyptian manuscript. Chavis, like Sabbagh, also seems to have translated Galland’s French version of ‘Aladdin’ into Arabic. A similar case arises with the manuscript of ‘Ali Baba’, which Macdonald had discovered in the Bodleian and which was in the hand of Jean Warsy, a French pupil of de Sacy’s who had settled in Egypt as a merchant in the late eighteenth century. This too has been shown to be a retranslation back into Arabic of Galland’s French version of the story.
