The Arabian Nights (New Deluxe Edition) (2 page)

Dissemination and Manuscripts

THE STORIES OF
the
Nights
are of various ethnic origins, Indian, Persian, and Arabic. In the process of telling and retelling, they were modified to conform to the general life and customs of the Arab society that adapted them and to the particular conditions of that society at a particular time. They were also modified, as in my own experience, to suit the role of the storyteller or the demand of the occasion. But different as their ethnic origins may be, these stories reveal a basic homogeneity resulting from the process of dissemination and assimilation under Islamic hegemony, a homogeneity or distinctive synthesis that marks the cultural and artistic history of Islam.

No one knows exactly when a given story originated, but it is evident that some stories circulated orally for centuries before they began to be collected and written down. Arab historians of the tenth century, like al-Mas'udi and ibn al-Nadim speak of the existence of such collections in their time. One was an Arabic work called
The Thousand Tales
or
The Thousand Nights
, a translation from a Persian work entitled
Hazar Afsana
(A thousand legends). Both works are now lost, but although it is not certain whether any of these stories or which of them were retained in subsequent collections, it is certain that the
Hazar Afsana
supplied the popular title as well as the general scheme—the frame story of Shahrazad and Shahrayar and the division into nights—to at least one such collection, namely
The Thousand and One Nights
.

The stories of the
Nights
circulated in different manuscript copies until they were finally written down in a definite form, or what may be referred to as the original version, in the second half of the thirteenth century, within the Mamluk domain, either in Syria or in Egypt. That version, now lost, was copied a generation or two later in what became the archetype for subsequent copies. It too is now lost, but its existence is clearly attested to by the remarkable similarities in substance, form, and style among the various early copies, a fact that points to a common origin. Specifically, all the copies share the same nucleus of stories, which must have formed the original and which appear in the present translation. The only exception is the “Story of Qamar al-Zaman,” of which only the first few pages are extant in any Syrian manuscript, and for this reason I have not included it in the present translation.

From the archetype there evolved two separate branches of manuscripts, the Syrian and the Egyptian. Of the Syrian branch four manuscripts are known to exist. The first is the copy in the Bibliothèque Nationale in Paris, in three volumes (nos. 3609-3611). It is of all existing manuscripts the oldest and the closest to the original, having been written sometime during the fourteenth century. The other three Syrian manuscripts were copied much later, in the sixteenth, eighteenth, and nineteenth centuries, respectively. They are, however, very close to the fourteenth-century manuscript and similarly contain only the nucleus and the very first part of “Qamar al-Zaman.”

If the Syrian branch shows a fortunately stunted growth that helped preserve the original, the Egyptian branch, on the contrary, shows a proliferation that produced an abundance of poisonous fruits that proved almost fatal to the original. First, there exists a plethora of Egyptian copies all of which, except for one written in the seventeenth century, are late, dating between the second part of the eighteenth and the first part of the nineteenth century. Second, these copies delete or modify passages that exist in the Syrian manuscripts, add others, and indiscriminately borrow from each other. Third, the copyists, driven to complete one thousand and one nights, kept adding folk tales, fables, and anecdotes from Indian, Persian, and Turkish, as well as indigenous sources, both from the oral and from the written traditions. One such example is the story of Sindbad, which, though early in date, is a later addition. What emerged, of course, was a large, heterogeneous, indiscriminate collection of stories by different hands and from different sources, representing different layers of culture, literary conventions and styles tinged with the Ottoman cast of the time, a work very different from the fundamentally homogeneous original, which was the clear expression of the life, culture, and literary style of a single historical moment, namely, the Mamluk period. This is the more significant because the Ottoman period is marked by a sharp decline in Arabic culture in general and literature in particular.

The mania for collecting more stories and “completing” the work led some copyists to resort even to forgery. Such is the case of none other than “The Story of Aladdin and the Magic Lamp.” This story is not among the eleven basic stories of the original work, nor does it appear in any known Arabic manuscript or edition, save in two, both written in Paris, long after it had appeared in Galland's translation. Galland himself, as his diaries indicate, first heard the story in 1709 from Hanna Diab, a Maromite Christian of Aleppo, who may have subsequently written it down and given it to Galland for his translation. The first time the story appeared in Arabic was in 1787, in a manuscript written by a Syrian Christian priest living in Paris, named Dionysius Shawish, alias Dom Denis Chavis, a manuscript designed to complete the missing portions of the fourteenth-century Syrian manuscript. The story appeared again in a manuscript written between 1805 and 1808, in Paris, by Mikhail Sabbagh, a Syrian collaborator of Silvestre de Sacy. Sabbagh claimed to have copied it in turn from a Baghdad manuscript written in 1703. Such good fortune, in retrieving not one but two versions of a lost wonderful tale, might be cause for rejoicing, as it indeed was among the scholars. However, a careful examination of the two versions, both in the light of the general style of the
Nights
and in the light of Galland's translation, leads to a less joyful conclusion. Chavis fabricated the text by translating Galland back into Arabic, as is manifest from his French syntax and turns of phrase, and Sabbagh perpetuated the hoax by improving Chavis's translation and claiming it to be a Baghdad version. And this forgery was the source used by Payne and Burton for their own translations of the story.

The Printed Editions

IF THE HISTORY
of the manuscripts is a confusing tale, that of the printed editions of
The Thousand and One Nights
is a sad comedy of errors. The first edition was published by Fort William College in Calcutta, in two volumes comprising the first two hundred nights (vol. 1 in 1814; vol. 2 in 1818). The editor was one Shaikh Ahmad ibn-Mahmud Shirawani, an instructor of Arabic at the college. He pieced this edition together from a late Syrian manuscript and a work containing classical anecdotes, choosing the texts at random. He deleted, added, and modified numerous passages and tried to change, whenever he could, the colloquial to literary expres-sions. He edited as he pleased. Then came the Breslau edition, the first eight volumes of which were published by Maximilian Habicht, between 1824 and 1838, and the last four by Heinrich Fleischer, between 1842 and 1843. For reasons known only to himself, Professor Habicht claimed to have based his edition not on a Syrian or an Egyptian but on a Tunisian manuscript, thus confusing the scholars until they finally disproved the claim by discovering that he had patched the text together from copies of the fourteenth-century Syrian manuscript and late Egyptian ones.

It was on such a late Egyptian manuscript that the first Bulaq edition of 1835 was exclusively based. It is a manuscript whose copyist, by culling, collecting, and interpolating numerous tales of recent vintage and written in a late style, swelled the old text, and by subdividing the material, obtained one thousand and one nights, thus producing a “complete” version of the
Nights
, a version very different from the Mamluk original in substance, form, and style. The Bulaq editor, Abd al-Rahman al-Safti Al-Sharqawi, not content to edit and print an accurate text of the manuscript, took it upon himself to correct, emend, and improve the language, producing a work that was in his judgment superior in literary quality to the original. Then came the second Calcutta edition, published in four volumes by William Macnaghten, between 1839 and 1842. Edited by several hands, it was based on a late Egyptian manuscript copied in 1829, with interpolations and with “corrections” in the substance and the style, according to the first Calcutta and the Breslau editions. Thus “thoroughly edited” and “completed,” as its editors claimed, it has ever since vied with the Bulaq edition in the estimation of scholars and general readers, not to mention all the major translators. Thus “authentic” came to mean complete and, ironically, spurious. (For a full history of the manuscripts and printed editions, see Muhsin Mahdi's Arabic introduction to his edition of the text of the
Nights, Alf Layla wa Layla
, Leiden, 1984, and his English introduction in the forthcoming third volume.)

The Mahdi Edition

IT IS ONE
of the curiosities of literary history that a work that has been circulating since the ninth century, that has been heard and read for centuries by young and old everywhere, and that has become a world classic should wait until very recently for a proper edition. This is curious yet understandable as one of the anomalies of comparative cultural studies. While the history of textual scholarship in the West has been, since the Renaissance, increasingly one of keen accuracy and authenticity, its counterpart in the East, especially in the case of the
Nights
, has been one of error and corruption, at the hands of Eastern and Western scholars alike, the result of ignorance and contempt. It is all the more gratifying, therefore, that the most recent edition of the Arabic text of the
Nights
should be by far the best. After years of sifting, analyzing, and collating virtually all available texts, Muhsin Mahdi has published the definitive edition of the fourteenth-century Syrian manuscript in the Bibliothèque Nationale (
Alf Layla wa Layla
, Leiden, 1984). Mahdi fills lacunae, emends corruptions, and elucidates obscurities; however, he refrains from providing punctuation and diacritical marks or corrected spellings. What emerges is a coherent and precise work of art that, unlike other versions, is like a restored icon or musical score, without the added layers of paint or distortions, hence, as close to the original as possible. Thus a long-standing grievance has been finally redressed, and redressed with a sense of poetic justice, not only because this edition redeems all others from a general curse, but also because it is the work of a man who is at once the product of East and West. And it is particularly gratifying to me personally, because it has provided me with the text for my translation.

Past Translations

NOT SO FORTUNATE
were the major translators of the work into English, Edward Lane (1839–41), John Payne (1882–84), and Richard Burton (1885–86). Lane based his translation on the Bulaq, the first Calcutta, and the Breslau; Payne on the second Calcutta and the Breslau; and Burton on the Bulaq, the second Calcutta, and the Breslau editions. These translators did not, as one might expect, compare the various editions to establish an accurate text for their translations (assuming that, given what they had, such a task was possible); instead they deleted and added at random, or at will, from the various sources to piece together a text that suited their individual purposes: in the case of Lane, a detailed but expurgated version; in the case of Payne and Burton, versions that are as full and complete as possible. In effect, they followed the same Arabic editorial tradition, except that whereas the editors of Bulaq and Calcutta produced a corrupt text in Arabic, the translators of London produced an even more corrupt text in English. Even the two less significant twentieth-century translations followed this pattern. Edward Poweys Mather based his English version (Routledge, 1937) on a French translation by J. C. Mardrus, which was based on the Bulaq and second Calcutta editions. Since he knew no Arabic, he altered the French text, ignorant of what he was doing to the Arabic or how far he had strayed from it. And N. J. Dawood, who translated a selection of tales (Penguin, 1956), which includes less than three of the eleven basic stories of the
Nights
, followed the second Calcutta, “editing” and “correcting” here and there in the light of the Bulaq edition.

Interestingly, the only exception to this pattern is Galland himself, the very first to translate the work in Europe (1704–17). His French translation of the basic stories was based on none other than the fourteenth-century Syrian text, as well as other sources. But instead of following the text faithfully, Galland deleted, added, and altered drastically to produce not a translation, but a French adaptation, or rather a work of his own creation. He did succeed, however, in establishing the work as a classic, for no sooner had his translation begun to appear than a Grub Street English version followed (1706–8), went into many editions, and was itself followed by other translations, pseudotranslations, and imitations, so numerous that by 1800 there were more than eighty such collections. It was such hack versions that inflamed the imagination of Europe, of general readers and poets alike, from Pope to Wordsworth. The
Nights
could shine in the dark.

These translators did not deviate from the letter of the original because they did not know sufficient Arabic. On the contrary, a careful comparison between any given Arabic passage and their own respective translation of it reveals an admirable command of Arabic diction, grammar, and syntax, except where the text itself poses severe problems, as it often does. Although the tales are generally written in the conversational style of the storyteller, they modulate between the colloquial and the literary, and even ornate, within a given passage, from passage to passage, and from story to story, and both types pose problems in regards to diction, grammar, and syntax. A great many words are thirteenth-century Syrian and Egyptian colloquial idioms, which have long since disappeared from usage or whose meanings have been altered; and many others are of Persian origin, either used without alteration or Arabized. The sentences are often ungrammatical, hence capable of several different and often contradictory readings. The typical structure is that of an interminable running sentence, consisting of brief coordinated clauses, often without apparent regard for place, time, or causality. The translator is therefore forced to interpret and reorder the clauses in a subordinated and logical sequence, in order to suit the European habits of reading and thinking, if his reader is to understand the passage at all. To make matters worse, the text, including Mahdi's, normally bears neither diacritical nor punctuation marks. In Arabic, the diacritical marks distinguish one letter from another, thus differentiating between words that share the same letters but have different roots and therefore different meanings. Thus a word may offer two very different readings in a given sentence. This is not a problem when one of the meanings is unlikely in the context, but when both are possible, the translator must choose a single interpretation. The diacritical marks also indicate the forms of conjugation and declension. Their absence, therefore, coupled with the faulty grammar of some sentences, makes every sentence an encounter, assuming, that is, that one always knows where a sentence or a unit of expression begins or ends, for the Arabic text uses no punctuation, not even question marks.