The Writing Revolution

by Amalia E Gnanadesikan

  Officially the language of the following Sasanian period was Middle Persian, written in a script much like Parthian; but other members of the Aramaic family of alphabets flourished as well. The Iranian Sogdian language had its own Aramaic script – or rather, several. The Christian speakers of Sogdian used a descendant of Nestorian Syriac to record their language, and the Manichees, followers of the third-century AD religious teacher Mani, wrote their texts (whether in Persian, Parthian, or Sogdian) in their own Manichaean script derived from Syriac Estrangelo.

  A good thing, too. As the Persian script evolved, its letters became connected, and it became what is known as Pahlavi. Many letters were now indistinguishable from each other. By the end of the Sasanian period, the Pahlavi script was nearly illegible. Words in late Pahlavi papyri have to be read as whole entities – almost as logograms – because the individual letters making them up cannot be distinguished. Scholars use spellings in Manichaean texts, as well as earlier, more legible Persian spellings, to break Pahlavi words up into individual letters. Native speakers of the time, applying their top-down reading skills, would not have had anywhere near as much trouble. As is so often the case, the intelligence of readers made up for the deficiencies of the writing system.

  The Avesta, the sacred book of the Zoroastrian religion, was first written down in the Sasanian period. These scriptures had been passed down orally for centuries, long after Avestan, the Iranian language they were composed in (a sister language of Old Persian), had become extinct. Since no one spoke Avestan natively anymore, it had to be written more accurately than the living languages of the time. Persian and Parthian were written with matres lectionis when vowels were considered necessary, but often vowels were simply left out, infrble frm cntxt. In the dead, sacred language of Avestan, however, inferring vowels from context using one’s native-speaker knowledge was not going to work. The same problem faced the compilers of the Masoretic text of the Hebrew Bible. So the alphabet was considerably beefed up, and Avestan was given a 51-character fully voweled alphabet to accurately portray its pronunciation. For once, precision trumped tradition.

  Having been successfully adapted to the Iranian languages – a linguistic group entirely unrelated to Semitic Aramaic – the Aramaic alphabet was ready to continue eastward, conquering new worlds and new languages. In front of it stretched the wide belt of Inner Asia, from the eastern border of Persia to Mongolia and Manchuria. This was the home of the Altaic language family, comprising, from west to east, the Turkic, Mongolian, and Tungus language groups. The Altaic peoples have historically been nomads. Though they have ruled great empires (those of Genghis and Kublai Khan and of the Ottoman Turks come to mind), they were not themselves progenitors of great civilizations. They did not found great literary traditions, but they did find uses for the technology of writing, especially in periods of empire building (see figure 9.3).

  Figure 9.3 Three Altaic descendants of the Aramaic alphabet. Uighur, which follows Aramaic alphabetical order, is shown next to the original Aramaic prototypes. Initial (left) and final (right) forms only are given here, omitting medial forms, except in a few cases where only medial forms exist. Altaic scripts are read from top to bottom.

  The ancestral Turkic people – the Türks – were the first Altaic people to write. Originally from Mongolia, they began to expand westward in the sixth century ad, founding an empire that was to reach all the way to the Black Sea. Their oldest surviving inscriptions date to the eighth century and are found along the Orkhon River in their Mongolian homeland. Though inspired by the noncursive version of the Sogdian script, their writing took on an idiosyncratic form, vaguely resembling the runes of Europe rather than the Aramaic scripts of Persia. And while it was written from right to left, like other Aramaic-derived scripts, its rows proceeded from bottom to top. The Türks, clearly, owed little allegiance to Middle Eastern tradition.

  In AD 745 control of the Türk Empire passed into the hands of their Turkic kinsmen, the Uighurs, who also learned to write Turkic runes. A century later the Uighurs were themselves displaced from Mongolia by yet another Turkic people, the Kirghiz. They settled in what is now Xinjiang (formerly Chinese Turkestan; see appendix, figure A.4) in northwestern China, where they adopted a sedentary lifestyle and invested in civilization. They cultivated a new script they found in Xinjiang, an adaptation of cursive Sogdian, with which they created significant works of literature.

  This Uighur alphabet (also spelled Uyghur, pronounced [wí gr] in English) contained 22 letters (not all of them originally Aramaic, as certain alterations had been made in the transfer from Sogdian). The Uighur language contained 25 consonant and 8 vowel phonemes, so some letters were pressed into double duty. Adding extra letters to the alphabet seems not to have occurred to the Uighurs, just as it did not to the early Hebrews or Aramaeans who used the 22-letter Phoenician alphabet – or to most peoples on the alphabet’s eastward march.

  At some point, probably under the influence of Chinese, the direction of writing was turned by 90 degrees. Instead of being written in rows from right to left, the page was given a quarter turn counterclockwise, with the result that Uighur was now written from top to bottom in columns that proceeded from left to right across the page (Chinese vertical writing, by contrast, goes from right to left across the page).

  With the introduction of Islam in the thirteenth century, the Uighur script was discontinued in favor of the Arabic script, which was itself replaced by Latin-based pnyn in the twentieth century (and, across the Chinese border in the Soviet Union, by Cyrillic). First, however, it was passed on to others.

  The Mongols were the next Altaic conquerors, overrunning large parts of Asia under Genghis Khan in the early thirteenth century, and even taking China under Genghis’ grandson Kublai in 1279. The Mongols mainly left administration to their conquered peoples; but still, an empire needed a script. They borrowed the Uighur vertical script – quite probably by making use of Uighur scribes – and only after the passage of a few centuries added letters to adapt it to their own language. This new Mongolian script was used for translations of Buddhist works which the Mongols, new converts to Buddhism, read enthusiastically, creating a literate culture that still remembers Mongolian script today (though it was replaced in Outer Mongolia by Cyrillic in 1946).

  An update of the vertical Mongol script, the so-called clear script, was invented in 1648 to improve the fit of Mongolian writing to the Mongolian language. For the first time in its migration eastward, other than Avestan, an Aramaic alphabet acquired letters for all the short vowels, giving it an unambiguous way of writing all the vowels of the language. Clear script came to be used by Oirat Mongols and is still in use, but the majority of Mongols clung to the classical vertical script. It was more familiar, and it was good enough.

  The Mongols were replaced in the arena of empire building by the Manchus, a Tungus people. The Manchus, like the Mongols before them, conquered China, giving it its last imperial dynasty, the Qing (1644–1912). As the Manchu were assembling their empire and preparing to take on China, they realized that they needed writing. Their first solution was to employ literate Mongols as scribes and to do the government’s business in Mongolian. Within a few decades, however, the script was adapted to Manchu, first with the addition of some letters, and then with the further addition of diacritical marks. The result included a full set of vowels, and represented Manchu more fully and unambiguously than vertical Mongolian had represented spoken Mongolian. With the Manchu script, the Aramaic alphabet had finally reached easternmost Asia, 3,400 years or so after the first Semitic alphabet was used in Egypt.

  Manchu was made an official language of China. Government documents were bilingual in Manchu and Chinese, and Manchu was used as the language of diplomatic correspondence. Many Chinese literary and historical works were translated into Manchu, and Western scholars used these translations to study Chinese. Despite this preferential treatment, spoken Manchu languished as the Manchus became assimilated to the Han Chine
se ways and language. Today the language is moribund.

  The last Altaic adaptation of the Aramaic script was designed in 1905 for the Buryat dialect of Mongolian. It was a systematic revision of the vertical Mongol and the clear script, designed to fit the phonology of modern Buryat. Type was cast for it, but the alphabet made little headway against the traditional vertical Mongolian.

  The alphabet that superseded many of the Aramaic scripts of the Middle East and Inner Asia was Arabic, today the most widely used Semitic alphabet. Since the Arabic script itself was descended from Aramaic, the replacement was a kind of alphabetic fratricide. Early Arabic writing had developed out of Nabataean Aramaic in pre-Islamic days, but the rise of Islam spurred the full development, codification, and spread of a specifically Arabic script.

  Even before Muhammad (c. AD 570–632) received the Qur’n in Arabic, the language had spread from its Arabian homeland into many parts of the Fertile Crescent (though some Arabs, such as the Edessans and Palmyrenes, probably gave up speaking Arabic for Aramaic, keeping only their Arabic names). Spread out as it was, the Arabic language consisted of numerous vernacular dialects. There was also a pan-Arabic formal dialect used in oral poetry. This dialect was the basis for the qur’nic language, the Qur’n being the first prose work written in that style. Its formal language and inspired source meant that the style of the Qur’n became the standard for literary Arabic (known as Classical Arabic) in the succeeding centuries. Classical Arabic was no one’s vernacular dialect, though it was more like some people’s speech than others’.

  Tradition tells us that Muhammad himself was illiterate; his words were memorized or jotted down by his followers. After the death of the prophet and a number of his inner circle, it became clear that Muhammad’s revelation was at risk of being lost. The text began to be systematically collected and copied. The third caliph (the caliphs were successors to the prophet in his secular role as leader of the Islamic community) was ‘Uthman ibn ‘Affan (644–56), who ordered the collation of the authoritative version that is still used today.

  As Islam spread along with the Islamic Empire, the Arabic language acquired many new speakers. By the time of ‘Uthman, the fledgling empire covered all of Arabia, the Fertile Crescent, and Egypt. By the end of the seventh century it included Morocco (with Spain to follow soon) in the west and Persia in the east. The Sasanian Persian Empire had been absorbed, and the Byzantine Empire had been drastically reduced.

  As is frequently the way of new empires, administration was at first transferred wholesale from existing systems. The bureaucracy functioned in Greek in the west and in Pahlavi in the former Persian lands. As the switch to Arabic occurred, in the 690s, knowledge of Arabic became an advantage and the language began to gain speakers.

  As the Muslim faith spread within the Islamic Empire, interest in Arabic grew yet further. Arabic held special status, for it was the language in which God had spoken to Muhammad. The Qur’n, as the revealed Word of God, should not be translated; it could not be the same in another language – it would not be God’s word verbatim. Devout Muslims of any language, therefore, had good reason to study Arabic. But unlike native speakers, they needed much more guidance as to how the words should be pronounced. And even native speakers of Arabic might have dialectal disagreements as to the correct pronunciation of words.

  To address these concerns, the Arabic script was revamped. The Arabic language had 28 consonants as compared to the 22 consonants available in the Aramaic script. Some letters had been doing double duty for two consonant sounds; now they were divided into distinct letters by the addition of diacritical dots. Thus (representing the sound in English thing), was distinguished from .

  Additionally, some originally distinct letters of the highly cursive script had come to resemble each other. So were also differentiated by the placement of dots. The tendency for letters to resemble each other was especially true in the middle of a word. Most letters in an Arabic word must be joined to the one that follows. At the end of a word, however, they tend to have expressive “tails.” Thus the letters developed distinct forms depending on whether they occurred independently or initially, medially, or finally in a word (see figure 9.4). The independent form, usually a combination of the initial form and the final form, is the one shown in simple alphabet charts. Some letters which were distinct in independent form were not distinct in initial or medial form, so diacritics were added to these as well: nu:n, are in initial position but for their dots just like initial These diacritical dots are now obligatory, forming an integral part of the letter.

  Figure 9.4 The Arabic alphabet with, from right to left, the independent, initial, medial, and final forms of each letter, plus the letter’s name in Arabic and its pronunciation in IPA. As compared to the original Aramaic prototype, Arabic has reordered its letters, putting similarly shaped letters together.

  Some letters, in contrast to the norm, must not be joined to the following one: the letter ?alif, if joined to a following letter would look just like the connected version of the letter la:m, alkitaab, “the book,” for example, the first (rightmost!) letter is an ?alif, which does not connect forward, while the next is a la:m. ?alif does join to a preceding letter, however, and the sequence la:m–?alif is written with a special ligature, .

  Another part of the writing system that needed to be addressed was the writing of vowels. Arabic has six vowels, three of which are pronounced short – [i], [u], and [a] – and three long – [i:], [u:], [a:]. Arabic inherited the system of matres lectionis from Aramaic and used it thoroughly, so that all long vowels are written. The ?alif, originally representing a glottal stop, now stands for the long vowel [a:] except word-initially, while the glottal stop is represented by an extra symbol that is not considered a true part of the alphabet, the hamza, (This situation arose because the dialect of the Prophet had lost word-medial and -final glottal stops. When later scribes speaking other dialects wanted to add them, they felt they should honor the original spellings and added the hamza as merely a diacritical symbol. Nowadays it may occur on an initial ?alif, as in the name of the letter, or alone as in the name of the letter ba:?.) Long [u:] is represented by the letter wa:w, which may stand for either the consonant [w] or the vowel [u:]. Similarly, the ja:?, stands for either the consonant [j] (normally transliterated as y) or the long vowel [i:].

  To show the exact pronunciation of the Qur’n, and to resolve all ambiguities of meaning, short vowels had to be represented. The doubling of consonants also needed to be shown – like the choice of vowels, consonant doubling is part of the morphological composition of a word and is traditionally omitted in spelling. So an additional layer of diacritical marks was added to indicate short vowels, consonant doubling, the glottal stop (hamza), and even the absence of a vowel. So spells ba, is bi, and is bu, while indicates a bb without intervening vowel, and indicates that no vowel follows the b. A number of other diacritics also exist for accurately showing the proper pronunciation of words.

  The short-vowel marks and other related diacritics are called vocalization. Full vocalization is used only for the Qur’n, for reading textbooks (for elementary school students and foreigners), and sometimes for literary classics – parallel to the Hebrew use of vowel pointing. Most texts just use the “skeleton” of alphabetic letters – the consonants and long vowels, including the dots which distinguish consonants. A fully vocalized text exists on three tiers – the central written line, the obligatory dots above and below the consonants, and the vocalization shown at the very top and bottom.

  To the native speaker and fluent reader, marking of long vowels only is sufficient. With their top-down processing, fluent readers do not particularly care to have every phoneme unambiguously spelled out. In Arabic this is especially true because short vowels can largely be determined by context. Their placement follows certain specific rules: all words begin with a single consonant, and sequences of more than two consonants do not occur. The identity of the vowels is determined by the morphologica
l form of the word being used. The writing of long vowels helps to reduce potential ambiguity, distinguishing kitaab, “book,” from katab, “wrote”, from kaatab, “corresponded,” all derived from the triconsonantal root ktb, which is used for words that involve writing. Since the triconsonantal ktb represents the core meaning of the word, a writing system that locates these letters at the core of the spelling and relegates some of the other features to the peripheries matches the structure of the language.

  Another reason not to use full vocalization is that modern Arabic, no less than in the time of the Prophet and probably more, is a language rich in spoken dialects. A great deal of dialectal variation exists in the vowels, especially in the short vowels. In fact, if it were not for written Classical Arabic and its somewhat updated version, Modern Standard Arabic, no one would consider Arabic a single language: there are many mutually unintelligible spoken forms. One form, Maltese, is written in the Roman alphabet and is now considered a separate language. Just as Chinese characters unite a diverse group of dialects or languages, so does written Arabic.

  In becoming literate, therefore, Arabs must learn a significantly different dialect – almost a new language. To some extent this is true in any written language – one learns to use more formal vocabulary and complex phrasing in writing than in colloquial speech. In Arabic the differences are much greater than in English, to the extent that literate people must become virtually bilingual, a situation known to linguists as diglossia. In Arabic, therefore, one could almost say that no one is really literate in their own language, a situation reminiscent of the Nabataeans and Persians who wrote in Aramaic, or the New Kingdom Egyptians who clung to written Middle Egyptian, but in marked contrast to the inventors of the original proto-Canaanite alphabet who developed the script to write their own mother tongue. The benefit derived from diglossia is that it allows the whole Arabic world to share the same literature – and, of course, the same Qur’n. The omission of short vowels from texts helps make the written language look more familiar to Arabic speakers who have not fully mastered the standard dialect.