The Creation and Spread of Scripts in Ancient India

Harry Falk The Creation and Spread of Scripts in Ancient India Abstract: Like his ancestors before him, King Aśoka (ca. 268–232 BC) was linked to the Seleucid dynasty in the West by family ties. Unlike his predecessors, however, he was the first to react to the immense cultural differences separating the two realms. One of the major deficits concerned the use of dressed stone for public monuments, the other was the lack of a script. India had produced a vast literature of a sacerdotal and secular nature, kept alive solely by oral means. Aśoka had a script designed for his country which amalgamated the best features of the two scripts current in his time, Greek and Kharoṣṭhī. The latter had been created only few years previously, 1600 km away in what is now northern Pakistan in an area formerly under Achaemenid rule, administrated with the help of Aramaic clerks. Although less suited to the language, Kharoṣṭhī script was not superseded by the new Brāhmī script from the East. This paper explains the perseverance of the western script on the basis of its function as a means for legal transactions. The need for documents had provided the incentive for the creation of a local script after the Aramaic clerks with their foreign script for a foreign language had started to disappear, rendering older documents illegible and thus worthless. Brāhmī, on the other hand, had little success as a script of adminis- tration, but served to demonstrate to the West the cultural and ethical standards of its homeland, and of its creator, Aśoka. Zusammenfassung: Mit dem König Aśoka (ca. 268–232 v. Chr.) der Maurya-Dynastie begann eine neue Zeit in Indien, denn er war nicht nur familiär mit den Seleuki- den im Nahen Osten verbunden, sondern er war auch der Erste, der die kulturellen Unterschiede begriff und Veränderungen in seinem Land herbeiführen wollte. Zwei Momente sind für uns besonders deutlich fassbar, einmal die Einführung einer Stein- technologie, die zu seinen monumentalen Säulen mit Tierkapitellen führte und zu den künstlichen Höhlen bei Bodh Gaya, zum zweiten ließ er für sein Land eine Schrift erschaffen und beendete damit eine Phase ausschließlicher Oralität, die von den vedi- schen Brahmanen zum Schutz ihrer Ritual-Literatur gepflegt wurde. Diese Schrift, Brāhmī genannt, übernahm innovative Züge der Kharoṣṭhī-Schrift. Diese war kurz zuvor im heutigen Nord-Pakistan entstanden, das bis zu Alexanders Eroberungen zum Reich der Achaemeniden gehörte. Trotz einiger struktureller Defizite ließ sich die Kharoṣṭhī nicht von der Brāhmī verdrängen. Die Zählebigkeit der Kharoṣṭhī wird hier mit ihrer Funktion als Medium der Rechtssicherheit in Verbindung gebracht: Mit dem allmählichen Verschwinden der aramäischen Schreiber-Kaste waren auf Aramäisch geschriebene Urkunden aller Art unlesbar und damit wertlos geworden. Nicht lange nach der Kharoṣṭhī wurde die Brāhmī entwickelt und verbreitet. Für sie ist eine frühe Verwendung in Handel oder privaten Verträgen nicht nachzuweisen. Für die Verwal- Bereitgestellt von | Freie Universität Berlin 44   Harry Falk tung wurde sie nur unter Aśoka und nur in der Ganges-Ebene nördlich der Hauptstadt eingesetzt. Die kultur- und religionsfördernden Aspekte, die den Ursprung der Brāhmī ausmachten, blieben auch weiterhin dominant. Leaving the Harappan culture of the third millennium BC apart, scripts in ancient India did not develop out of simple beginnings in a gradual fashion, as in the literate cultures around the eastern Mediterranean, but pop out into view late, suddenly and almost fully developed in the early third century BC.1 After the fall of the Harappan culture of unknown linguistic affiliation, around 1900 BC, literature was produced in the so-called Vedic language for Vedic rituals, and to our advantage a great part of it was preserved to this very day by mainly oral means. Even after the Iranian Achaemenids occupied lands up to the banks of the Indus, the adjoining Vedic groups never thought of developing a script for their textual treasures. The mechanics of orality in those days are comparable to what Julius Caesar (Caes. Gall. 6,14) reported about the literature in Celtic France: the Druids knew an amazing amount of verses by heart. Their knowledge was the basis of their social standing and also of their livelihood, causing these specialists to have an unfavourable view of writing. Nobody in India east of the Iranian dominions challenged this attitude because India lived at the fringe of the known world, with very little contacts to the West and North, happy in its own sort of splendid isolation. Things changed with Alexander the Great, coming from outside, and with the indigenous Maurya dynasty, willing to intensify contacts with the Hellenistic rulers of the West. Alexander left a network of newly founded towns from Mesopotamia up to the Indus, in part populated with Greek-speaking and -writing veterans. These urban centres were by definition interrelated through commerce and spread knowledge of Western material culture and also of actual political events. They also helped to intro- duce a range of cultural concepts concerning property rights, building in stone and dressing stone for artistic purposes, and also script as a means of administration and for private purposes, freely available for everyone. Even before Alexander, people and rulers in India had certainly heard of the build- ings of Persepolis, Babylon or Memphis, but without the secure roads and seaways from Mesopotamia and Egypt to the Indus needed for bulk trade in both directions, this had had little impact. The transformation effected by Alexander now allowed a greater number of people in India to become aware of the huge cultural differences concerning material culture, administration and other components of what we call civilization. Candragupta (ca. 321–297 BC) and Bimbisāra (ca. 297–273 BC), alias Ami- traghāta, the first two rulers of the Mauryas, were in contact with Seleuceia, but they 1 An exposition of the facts was given by Max Müller as early as 1859; for a summary of all arguments proposed so far cf. Falk 1993. Bereitgestellt von | Freie Universität Berlin  The Creation and Spread of Scripts in Ancient India   45 seem to have been busy securing their enormous realm, without spending too much time on modernising their country. The State of Stone Working before Aśoka It pays to glance at what an archaeologist can find in Indian soil that dates from the time before Alexander opened the roads to the West. There are a number of old cities and commerce stations, and many of them are known from the itineraries of the Buddha in the decades prior to 400 BC, meaning that they definitely existed before the Macedonians reached the Indus. Härtel has shown how little of their remnants can be dated to times earlier than the Mauryas,2 and Erdosy’s neat summary of the problem still remains valid: “To this day not a single house-plan is available for pre-Mauryan levels”.3 Arrian gives us a reason why and also explains why this state of affairs persists even into Mauryan levels: “. . . such cities as are situated on the banks of rivers or on the sea-coast are built of wood, for were they built of brick they would not last long–so destructive are the rains; (. . .) those cities, however, which stand on commanding situations and lofty eminences are built of brick and mud.”4 Brick and mud, but not stone was used. When things change, non-humans seem to be at work, a topic of talk for centuries to come. The Chinese pilgrim Faxian was told shortly after AD 400 that “Pataliputra is where Asoka ruled. The royal palace was wholly built by ghosts who piled up stones for walls and towers, who incised letters, made inlays which cannot be done in this (human) world. It still exists as of old”.5 A visitor soon realizes that many of the oldest cities are nestled inside a narrow ring of surrounding hills. The most famous such site is Rājagṛha, modern Rajgir. Similar sites are the old Shahbazgarhi in Gandhara, or the old Kopbal in Madhya Pradesh, in between the Aśokan Minor Rock Edicts (MRE) of Palkiguṇḍu and Gavīmath. Bairaṭ, near Jaipur, likewise falls into this category, as does Hathial, the oldest part of Taxila. This kind of settlement allows for only a limited number of inhabitants, who are sup- plied with water from the surrounding hills. A second and younger type of city provides shelter inside huge, mostly square earthen wall constructions in places where there are no major hill formations. Famous examples are Śiśupālgarh near Bhubaneswar in Orissa, then the whole site of Jaugarh, also in Orissa, Sannati on the Bhīmā and Kauśāmbī on the Ganges, all sites famous for their vestiges of Aśoka. Without preserved traces of Aśoka are Adam, Mahāsthāngarh, or Candraketugarh in Bengal. In all cases, the huge earthen walls furnish remarkably 2 Härtel 1991. 3 Erdosy 1995, 110. 4 Arr. an. 10 = Majumdar 1981, 223  f. 5  After Deeg 2005, 547. Bereitgestellt von | Freie Universität Berlin 46   Harry Falk few traces of habitation, recalling our Celtic oppida in size and construction, which served to shelter merchants until they were numerous enough to depart in a secure caravan. What did houses look like in the “wooden” days of Pāṭaliputra? We should not project backwards the depictions on the gates of Sanchi. Those houses are 200 years younger and impress with their high bases, so that the inhabited part is elevated to the second floor. But we have the caves which Aśoka had cut into the boulders at Barabar, yet another old site surrounded by hills near Bodh Gaya. His most ingenious architects copied luxurious private dwellings into the rocks for the ascetics of the Ājīvika order. One cave has received an entrance front that copies in stone the wooden beams and rafters and carvings from a standard house. The most telling part is found inside: all caves without exception show slanting long-side walls, and upright end-walls. This slanting (about 10 cm inwards with a height of 180 cm) copies basic constructions of upright pillars along the sides of houses, topped by a close sequence of semi-circular ribs. This stable barrel-shaped roof construction then received vertical end-walls in the form of the entrance at the Lomas Rishi Cave at Barabar. A stunning example is provided by the Gopālī Cave at the Nagarjuni Hills, one km further east from Bara- bar,6 with two circular ends on the narrow sides, under a rounded roof. A comparable design is also used for long-houses built on the ground at nearby Rajgir. Unfortu- nately, the documentation is very superficial, but one photograph (IA-R 1954–55, pl. XXIX) and two ground-plans (IA-R 1954–55, fig. 4; 1958–59, 13) leave no doubt that the Gopālī Cave copies constructions where the foundations, 1 m thick and more than 1 m high, were not made from burned bricks but from rubble and mud-mortar.7 At the Gopālī cave with its two semi-circular ends as well as in most other Aśokan caves there is a sharp and straight line separating the slanting wall-sides from the vaulted ceiling, reminiscent of the line where in ordinary houses the wall met the ribs shaping the roof. Rubble and mud was in use, but bricks can be expected also for foundations, although so far they are mainly found inside stūpas. For this reason I characterize the pre-Aśokan housing culture in India as a “culture of wood and bricks”, materials which seem to have dominated the public world, and I dwelt here on the simplicity of construction techniques at seemingly undue length in order to make clear how sharply the years before ca. 270 BC differ from the phase to follow. 6 Cf. Falk 2006, 257; these caves have been “handed over” (niṣiṭha) to the Ājīvikas by Daśaratha, prob- ably after the demise of Aśoka, who always uses “given” (dinna) for his personally effected donations. 7 According to IA-R 1954–55, 16, the constructions are found immediately above the Northern Black Polished Ware and yielded copper coins. This looks like a Śuṅga time horizon, but without further ev- idence an occupation by monks not using luxury ceramics in Mauryan times should not be excluded. Chakrabarti 1976, 263  f. considers the site unimportant. Bereitgestellt von | Freie Universität Berlin  The Creation and Spread of Scripts in Ancient India   47 Towards the Transformation After Alexander, and after the first two Mauryas who did not leave a single clear trace to the spade,8 the third in line, Aśoka (268–232 BC), modified his country in a revo- lutionary manner. His realm reached from modern Calcutta to Kabul, from Pokhara in Nepal almost to Madras. He changed India from a “wood and bricks” country to a “stone and art” country. His impact cannot be exaggerated. He had people cut mon- olithic pillars and animal figures in stone, with huge dimensions, the pillars up to 14 m high9, weighing up to 300 tons, the animal figures adding another 2 to 2.5 m. The animal capitals are chiselled in a very naturalistic way, the surface of the stone is polished to a glaze. One may imagine the reactions of the locals as they were first con- fronted with these works of art. Judging by what present-day rustics say, the people of old must likewise have believed that these pillars and the animals were made and transported by gods. Besides introducing stone and art, Aśoka had a script designed and spread, carry- ing his messages. It seems that Aśoka’s intention in doing so was not merely to guide his people, but also to communicate with visiting foreigners, as is suggested by the fact that many inscriptions are located on borders. Why? Indians returning from the Western countries must have told tales of all sorts of wonders they had encountered, be it the pillars of Persepolis or the obelisks of Egypt. At the Mauryan court, people must have realized that there was nothing in India that could impress foreigners in a similar way. The most glaring differences concerned building in stone, plastic art and the use of script, all of them absent in India. We can be sure that building in bricks, wood and bamboo was of a high quality and had its own aesthetic appeal, but any- thing only vaguely similar to the temples or palaces in Bactria, Persepolis, Seleuceia or Alexandria in Egypt was absolutely lacking. Every foreign visitor would sooner or later ask for the local temples. But there were no temples, there were no statues of gods. Aśoka was personally related to the Seleucid dynasty, he must have met Greek and Macedonian guests in the household of his family. He must have heard them talk about what they could not find in the capital: stone foundations of houses, stone-lined city walls, stone-paved roads, temples for stationary gods, statues of gods, and many things more. Once Aśoka had installed edifices and artefacts made from stone, the new tech- nology was spread everywhere and the traders and diplomats from the West returning about half a century after Candragupta had reason to speak with awe of the India created by Aśoka. In principle, Mauryan India was rich; it sold elephants, gems and spices to all parties in the West. Aśoka wanted more than riches, he wanted to be king 8 Jacobs 2016. 9 Falk 2006, 152. Bereitgestellt von | Freie Universität Berlin 48   Harry Falk of a country that stood culturally on par with the countries of the famous Macedonian kings following Alexander. A country without any script certainly appeared backward to the people in the West, but Aśoka was about to fill this gap. However, before he could introduce a script of his own something happened at the western border of his own empire. Script Was “in the Air” Around 550 BC the Achaemenid dynasty in Iran made what is now Pakistan its tribu- tary. Although writers came from the West an indigenous Indian writing system was still not in sight, although we can expect a general knowledge of writing as such. The grammar of Pāṇini from roughly the middle of the fourth century BC knows a term for script, to be spelled lipi or libi, but this grammar itself is a composition of such pho- netic subtlety that it can only be preserved unchanged by oral means.10 The mistakes innate in writing would have destroyed the text in a very short time. The admitted spelling variance shows that lipi or libi are foreign terms with no Sanskritic derivation and thus no dogmatic orthophony. In fact, both forms of pronunciation, libi and lipi, reproduce a term which is ultimately Sumerian, used by the Achaemenid administra- tors active in northern Pakistan from the 6th to the 3rd century BC. Pāṇini, who hailed from a site on the Indus in northern Pakistan, thus knew writing, but did not make use of it for his so-called grammar – which is rather a philosophy of sound and meaning. The reason is obvious: the sounds which can be reproduced by the Aramaic script used by the clerks in the service of the Achaemenid dynasty are not in the least suited to represent the range of sounds analysed by the Indian grammarians as phonetically relevant. The Aramaic alphabet knows of 22 letters, plus number signs, and whereas the sounds to be distinguished by a student of Sanskrit amount to about 50, the exact number varies depending on whether theoretically possible sounds are included or not. A simple adoption of a Semitic script without additions is thus impossible. Even if more letters were added to make good for the remaining Indian sounds, the system would still lack the ability to represent doubled consonants and vowels of different length. Realizing that there are incompatible languages (Aramaic and Sanskrit), dif- ferent sound systems, and different purposes (administrative and sacerdotal) must have led to the logical conclusion that Aramaic writing, which appeared as writing as such, was of absolutely no use to Sanskrit speaking priests and philosophers. A short time later, after the fall of the Achaemenids in the wake of the conquests of Alexander, the general situation changed completely. Although the Aramaic clerks continued to work for or under the Macedonian rulers, genuine Greek writing became 10 Deshpande 2011. Bereitgestellt von | Freie Universität Berlin  The Creation and Spread of Scripts in Ancient India   49 known to the public, a writing not shrouded in the garb of a Semitic language, but open to everyone interested and used for a language with a number of structural sim- ilarities to the local Indic idioms. Once the attitude towards writing had changed to the positive it must have been tempting to try writing also on the local vernaculars. For us, these vernaculars had clearly branched off from Sanskrit long ago, but the pandits maintaining pure Sanskrit were not part of the process of introducing literacy in the third century BC. On the other hand, those who finally tried to adapt the Aramaic writing system lacked even the basic knowledge of the Sanskrit grammar schools. There were at least two such attempts at adaptation, taking place almost simultane- ously, one in northern Pakistan and another in the capital of the Mauryan empire, at Pāṭaliputra, modern Patna in Bihar. The Invention of the Kharoṣṭhī Script The first such attempt apparently took place in northern Pakistan and drew substan- tially on what people were used to seeing, and that was the Aramaic script. The Ara- means, responsible for the term, had once installed some short-lived kingdoms in Mesopotamia, which were destroyed by the Assyrians, who distributed this people all over their Assyrian realm in the eighth century BC. The Aramaic clerks spoke a lan- guage closely related to Hebrew and served as a sort of postal system for their Meso- potamian oppressors, and continued to do so once the Iranian Achaemenids took over the Neo-Babylonian empire in 539 BCE. As the Achaemenid realm expanded, Aramaic clerks were dispatched “on duty” further east as well, up to sites on the Indus. After the fall of the Achaemenids brought about by Alexander the Great, Aramaic clerks continued to work for the local courts. We have two texts documenting the outcome of court cases, each one mentioning “Vasu the judge” (wʼšw dynʼ / wʼšw ŠHM dynʼbr), written in Aramaic and dated to the regnal years 16 and 17 of the third king of the Mauryan dynasty, Aśoka, showing us that the script of the Achaemenid administra- tion survived the downfall of the empire by at least eighty years. By their very nature, these documents in the old script help us to understand why a new script, named kharoṣṭhī, had to be invented. Both documents (“Laghman I and II”) were found on vertical rock sides in the lower Laghman valley near Jalalabad.11 Such documents are of high value when it comes to defending ownership positions once sanctioned by a local court. The preparation and interpretation of these docu- ments naturally required the presence of Aramaic clerks. Without them the written verdict was useless. How difficult it is to understand the few pieces that have come down to us becomes obvious when comparing the various translations offered so far 11 Falk 2006, 247–250. Bereitgestellt von | Freie Universität Berlin 50   Harry Falk Abb. 1: Aramaic writing from the Laghman valley, Eastern Afghanistan, describing a plot of land and mentioning the year 16 of Aśoka; eye-copy Djelani Davary, cf. Falk 2006, 250. by modern scholars otherwise well versed in the Aramaic language as such. The texts seem to contain a number of juridical terms or their abbreviations, which to some extent still defy philological analysis. A number of Aramaic documents on goat skins, mostly economic contracts and administrative accounts, written in Afghanistan under Greek rule soon after the con- quest of Alexander have been published recently.12 Amongst Macedonian families Greek idioms and script were certainly in use, but for local landlords and their juridical needs documents held in Greek would have been much under the same risk as those requiring readers of Aramaic. It must have been in this period when the Aramaic clerks were on their way out that some agencies dependent on court decisions realized the need for a new script, this time for reproducing the local language. Land was re-allotted in the wake of the Macedonian conquest, the new land-owners needed authoritative documents, legible, held in the local language, usable for generations to come, inde- pendent of clerks or rulers preferring foreign languages. Many of the new land-owners were Greek or Macedonian, but a number of collaborators to the foreign forces received rights as well, as we learn from the works describing Alexanderʼs campaign in Paki- stan. The last and most recent documents using the Aramaic language and script in Gandhara are the court decrees mentioned above, and the need for dependable and comprehensible court decrees and other business documents is proposed here as the reason behind the creation of a new script. Land-owners or traders in bulk, when they come in some numbers are a fraction of society economically strong enough and by nature politically close enough to a local sovereign to be able to promote the spread 12 Naveh & Shaked 2012. Bereitgestellt von | Freie Universität Berlin  The Creation and Spread of Scripts in Ancient India   51 and use of a medium necessary for all sorts of business. The rulers may not even have needed much convincing – it only needed someone bold enough to start and others influential enough to implement the general use of the new form of writing. As was explained above, it was impossible to adopt Aramaic directly for use with the local language, and so the foreign script was modified to suit its new purpose. First, a number of additional letters were designed, and then, most important, the vowels, which are not or only insufficiently represented in Aramaic writing, were given a definite shape for mandatory use in initial and in medial position alike, by adding short strokes to the basic character. Thus, the local language, nowadays generally termed Gāndhārī after the region of Gandhara in and around the Pesha- war valley, could be reasonably well expressed in writing, and understood by all its speakers when read aloud.13 The script looks like Aramaic at first glance, like Aramaic it runs from right to left, and it even has identical ciphers and a number of characters which express the same or closely similar sounds in both languages (yod/ya, waw,va, nun/na, reś/ra, bet/ba, dalet/da). So there can be no doubt that this new script, later called Kharoṣṭhī, had Aramaic script as an antetype with good reason: by retaining the script’s traditional appearance, the documents looked as trustworthy as the purely Aramaic documents had done before. Surprisingly, this new script has also a number of letters that look alike in Aramaic script and Kharoṣṭhī, but express completely different sounds in the Aramaic language and in Gāndhārī (peh/a, taw/pa). To add to the confusion, there are also sounds, used in both languages, which are represented by very differently shaped characters in the two scripts (gimel/ga, he/ha, kap/ka, lamed/la, mem/ma, peh/pa, taw/ta). In the 19th century it was Georg Bühler, an influential Indologist from Vienna, who tried to explain the genesis of Kharoṣṭhī. To explain the cases of “identical letters but different sound values”, and “identical sounds but different shapes”, he twisted and mirrored the Aramaic letters until they resembled some Kharoṣṭhī counterparts. The twisting and mirroring was regarded as a reconstruction of what he thought happened to the original letters through the centuries. Such gradual deformations are known in all scripts that have come of age, but then we have all or most stages of the changes documented. In the case of Kharoṣṭhī, not a single intermediate form of this kind was available for Bühler and so he tried to find help through the Aramaic papyrus docu- ments from Elephantine, written in Achaemenid times in southern Egypt in the 8th century. Certainly, there are shapes differing from the prototypical Phoenician script used in the Levant, but all changes follow the standard principles of formal modifi- cations. And no such change resulted in any of the letter-forms Bühler would have needed to prove that Kharoṣṭhī had locally and gradually emerged out of the Aramaic 13 An intermediate state is in evidence in a text of Aśoka found at Kandahar, which presents its con- tent twice: first in Aramaic script and language, then in Aśokaʼs own language rendered in Aramaic letters (cf. Falk 2006, 246). Bereitgestellt von | Freie Universität Berlin 52   Harry Falk Abb. 2: Kharoṣṭhī writing, rubbing from Rock Edict 12 of Aśoka at Shahbazgarhi, Northern Pakistan. After Hultzsch 1925, 64. script. Because of this flaw, other, and better models for the origin of Kharoṣṭhī were proposed already in his time, but for various reasons they were not well received.14 One of Bühler’s basic assumptions was that all intermediate forms necessary for his hypothesis are undocumented because they were written on perishable material. The Indian termites, certainly fond of birch-bark and palm-leaves, are an indispensa- ble element in all attempts at making Indian scripts older than they can demonstrably be. Fortunately, the Indian termites donʼt travel to Bactria and they may also shrink back from gnawing at animal skins, and so a number of documents were preserved, not from Elephantine but from northern Afghanistan, not from the time of Kyros, but from Alexander at the end of the fourth century BC.15 These documents reveal what was to be expected, that the Aramaic script in Bactria had remained the same regard- ing shape and sounds, with only some minor modifications; and not a single one of those indicated a twisting or bending in the sense of Bühler. Even the letters which developed out of the Aramaic script in the 2nd century AD for the Parthian, Middle Persian and Sogdian scripts still continue the shapes of standard Aramaic,16 as it is 14 Falk 1996, 127  ff. 15 Falk 1996, 127  ff. 16 Frye 2006, 60. Bereitgestellt von | Freie Universität Berlin  The Creation and Spread of Scripts in Ancient India   53 found i.a. in Gandhara, in the Aśokan texts in Laghman, in the 3rd century BC. To expect evidence of “twisting” in Elephantine Aramaic centuries earlier seems rather anachronistic. Although to my mind the “twisting technique” has nothing to its credit it produced a long-term model which still lives on and assumes an “adaptation of the Aramaic script” “developed sometime in or around the fifth century BCE”,17 i.  e. in Achaemenid times, when all other Achaemenid provinces retained their Aramaic script as it was.18 With the twisting hypothesis eliminated, we have to reconsider how identically shaped letters can receive completely different sound values, and how identical sound values can have taken on a completely different graphical representation. The only solution, to my mind, is taking leave from a gradual change over centuries resulting in a script in use for centuries alongside Aramaic. Instead, I consider a spontaneous process more likely, involving an Aramaic clerk, who shows an interested person of definitely non-Aramaic stock how the letters work as a system, writing down parts or the whole alphabet on a medium – to take away. This supportive Aramaic clerk was not the one who invented Kharoṣṭhī, as then he would have had to allot some com- pletely new sound-values to letters he had used for years for very different sounds. For such a radical break with traditional sound values to occur the creator must have been free from Aramaic cleric learning. Our unknown local Gandharan inventor to-be kept in mind correctly a number of sound-attributions from the instruction he received. However, he had forgotten a number of sounds attributions and thus combined the remaining sounds at random partly with characters he still had unused on his exem- plar or he invented other forms freely for those sounds of Gāndhārī still remaining unrepresented, occasionally enlarging signs for related sounds.19 This model can explain the similarities as well as all deviations, and it does away with the almost schizophrenic situation that a correct Aramaic script could be used in the time of Aśoka side by side with its own derivation wherein sounds and letter-shapes had received different and thus confusing realisations. The assumed spontaneous creation must be considered against the political back- ground. The Aramaic clerks lived in their own little world, with colleagues all over the Achaemenid realm. They, and only they, knew to listen to words in one language, e.  g. Gāndhārī, write them down in Aramaic letters and language, and retranslate it for the recipient into Greek, or Babylonian or Egyptian, whatever needed. Imagine a clerk in Gandhara using any of the twisted or mirrored shapes for a document. The 17 Salomon 2012, 175. 18 According to Baums (2014, 214), O. von Hinüber and myself are said to agree that Kharoṣṭhī “is a derivative of the Aramaic script used in the administration of the Achaemenid empire”. I cannot find such a statement in the work of O. von Hinüber and my own view is diametrically opposed and has been repeated in print since 1993. A spontaneous idiosyncratic creation using Aramaic as a base model (“Vorbild”) is far removed from a “derivative” that would have required centuries to take shape. 19 Strauch 2012, 147. Bereitgestellt von | Freie Universität Berlin 54   Harry Falk recipient in Babylon would have been at a loss. A popular version of writing could only be launched once the postal lingua franca had already started to become supplanted by Greek. The inventory of letters of most writing societies is conventionally taught in a fixed sequence to make sure that the full range of letters is kept in memory. The Greek and Roman sequences are inherited from an age-old20 North Semitic inventory and are thus free from any systematic ordering.21 If Kharoṣṭhī had been developed out of Aramaic writing its sequence should also follow a North Semitic prototype. But it does so only in part, in that it has no logical sequence. With the exception of the starting letter /a/, its other half, the sequence of sounds, is completely different. The conver- gence and variance seen when comparing the North Semitic alphabet with the ara- pacana order are once more better explained by my spontaneous creation model than by Bühler’s gradual change hypothesis and its variants. The sequence of Kharoṣṭhī has been found in literature and on writing pads alike,22 all identical with only trifling deviations, starting with a, ra, pa, ca, na, la, da, ba, ḍa, ṣa, va, ta and 30 more letters,23 showing that already at this point there is no internal phonetic system apart from a separation of the one initial vowel and all following consonants. From no. 19 onwards there is a mixture of simple consonants and biphonematic sounds, as in tha, ja, śpa, dha, śa, kha, kṣa, sta, ña, proving, to my mind, that no brahmin trained in phonetics can have had anything to do with this hotchpotch. How then did the arapacana originate? The outcome reveals the mechanism behind its origin, happenstance. When did it originate? In the Gāndhārī used for the Aśokan texts, a considerable number of signs later in use are still missing. Bipho- nematic sounds are scarce, as we find only kṣa, sta and spa and their signs are not formed as ligatures of ka and ṣa, of sa and ta or pa, but single graphemes independent of the shapes of their constituents. This shows that kṣa, sta and spa24 were not reduced to their constituents, probably out of ignorance, and not because original clusters had changed to monophonematics. On the other hand, a correct analysis of sound clusters had taken place before Aśoka in the case of pre-consonantal r- and post-consonantal -ra, -va and ‒ya. The classical arapacana range of 42 letters represents a much larger number of sounds compared to those used for the edicts of Aśoka, indicating that this full list was not in existence at the beginning but is rather the product of a gradual improvement 20 Salomon 2013, 9. 21 Salomon 2013, 12. 22 Salomon 2016. 23 Strauch 2012. 24 For spa this original state was later changed and a true ligature designed where a Śa-head came to stand above a pa-hook. Bereitgestellt von | Freie Universität Berlin  The Creation and Spread of Scripts in Ancient India   55 of this script.25 Once enlarged, the arapacana order had a long life, even surviving the change of script from Kharoṣṭhī to Brāhmī.26 The Invention of Brāhmī In principle, the Aramaic monopoly on writing in ancient India collapsed with the Macedonian conquest. With the new creation called Kharoṣṭhī in Gandhara, writing was ready to be taken up by anyone. Naturally, a script needs some time to spread, and we see that it was already sufficiently widespread when around 250 BC Aśoka, from his capital at Pāṭaliputra, 1600 km further east, sent his “edicts” to Gandhara in the West in another new script, which we call “Brāhmī”. At the local level, the newly popularized Kharoṣṭhī was regarded to be firmly enough established to disregard the script in which the texts arrived. Replacing Kharoṣṭhī by yet another new script in such a short time was not attractive to those who just had managed to get Kharoṣṭhī accepted. In my model the invention and proliferation of Kharoṣṭhī occurred following the inroad of the Macedonian army around 330 BC with their gradual invalidation of the Aramaic clerks; the encounter between the first and the second new script hap- pened when the Aśokan edicts were sent west around 250 BC. The refusal in Gandhara to replace Kharoṣṭhī by Brāhmī was somewhat short- sighted, because the script from the East had a number of advantages over the local pioneer. In fact, the system of vowel strokes is basically the same as in Kharoṣṭhī and shows that clerks from the West were part of the developing team in the capital of the Mauryan empire, Pāṭaliputra, today Patna. One of Aśokaʼs clerks came from the West, he disseminated Aśokan edicts in Southern India in Brāhmī script, added an explanatory part which he signed with his name and function in Kharoṣṭhī, “made by Capaḍa the clerk”. Despite all attempts at constructing alternative models, the Aśokan texts are the earliest material evidence of the use of Brāhmī, and Kharoṣṭhī as well. Initially, this king must have planned for an illiterate population. The edicts prescribe at which date the words of Aśoka “have to be made to listen to”, not “have to be read”. The verve perceptible in the distribution of his texts from Orissa to Kandahar, from the Terai of Nepal down to just outside Tamil Nadu shows that he wanted to spread not only his words, but also the knowledge of and the capacity to use writing. Whoever starts using a script in a culture not based on writing will have to tackle the question: How to make people write and read? The answer is simple: You just have to start using it and continue teaching it, begin the impossible and stick to it. Fortunately he succeeded. 25 Strauch 2012, 147. 26 Salomon 2016, 20. Bereitgestellt von | Freie Universität Berlin 56   Harry Falk Abb. 3: Government order in Brāhmī script on stone from Mahasthangarh, the old Puṇḍranagara, north of Calcutta, regulating the use of victuals in state-controlled store-houses. After Sircar 1941, pl. XXI. What induced him to try the impossible? He had grown up in a capital with links to the West. Around 300 BC Megasthenes had come from the Seleucid court in Babylon as an ambassador, and was surprised to see the Indians go to court without any written evidence in hand.27 There were more ambassadors in the decades after Megasthenes,28 there was the transport of figs and grape juice from the Seleucid court in Mesopota- mia29 to Pāṭaliputra in the time of Aśokaʼs father Bimbisāra. Aśoka was well-informed about some rulers in the Hellenistic East. He knew that there were 600 yojanas (Rock Edict 13) to the West, that is equivalent to 6000 km, covering the distance between Pāṭaliputra and Alexandria as the crow flies, or over land to Seleuceia following the land roads. Naturally, he knew the dynasties of Antiochos I (r. 281–261 BC) or Antio- chos II Theos (r. 262–246) at Seleuceia, whom he calls the yonarāja, “the king of the Greeks”, here as in Rock Edict 2. In addition, he gives four simple “kings” (rājā) of 27 von Hinüber 1989, 20. 28 Karttunen 1997, 99–100. 29 The source is Athenaios of Naukratis in his Deipnosophistai (ed. C.D. Yonge, Book 14, 67), preceded by an episode which shows how Xerxes was led to implant figs from Athens into Achaemenid soil, an import the Seleucids would later pass on to India. Bereitgestellt von | Freie Universität Berlin  The Creation and Spread of Scripts in Ancient India   57 descending importance, mentioned by name in Rock Edict 13, summarily called “his adjoining kings” (ye ca tasa samaṃtā lājāne) in Rock Edict 2. In Rock Edict 13(Q) the first “adjoining” king is Ptolemy II (tulamaya; r. 285–246 BC) in Egypt, follows Anti- gonos II (antekina) Gonatas (r. 283–239 BC) of mainland Greece. Of clearly less impor- tance was Magas (maga/maka) with his short-lived (276–250 BC) kingdom in Libya, and the least important was governor of the western Seleucid dominions in Anatolia, Alexander of Sardis (Merkelbach 2000).30 With all the roads and seaways secured by the Seleucid administration, people, goods and news could spread at a pace unheard of before in this part of the world. Even one lady from Seleucosʼ family was married into the Mauryan family and vice versa, so that in principle there was sufficient reason to stay in contact.31 The lady from Seleuceia certainly spoke much about the differ- ences she had to witness. How did Aśoka communicate with his in-laws in the Near East? Did he send messengers who conveyed the texts from memory? Or did he supply them with letters in Aramaic or Greek? With all this correspondence recurring on a regular basis there was no way to go on without a script to express oneʼs own cultural standing. Why did Aśoka not venture to adopt the Greek script? Like Aramaic, the Greek script cannot be used unchanged to represent Aśokaʼs local vernacular, called Māgadhī after the region Magadha, now Bihar. Like the inventor in Gandhara who probably knew “script” only in the shape of the Aramaic letters, Aśoka may have known “script” primarily in the shape of Greek, in those days written in nothing but capital letters, but Kharoṣṭhī was definitely known to him and his advisors. The Brāhmī script result- ing from a comparison of Greek and Kharoṣṭhī shows that all their advantages and shortcomings were discussed in order to perfect their own creation. The Greek script definitely won over Kharoṣṭhī with regard to its visual impression, with its broad, upright and symmetrical forms. Writing from left to right also seems natural for all right-handers who can thus avoid having to move the pen against the medium. From Greek came the inspiration to distinguish between short and long letters, as found in Greek (E-H, O-Ω). On the other hand, Kharoṣṭhī was better than the Greek script in that it features the short -a “inherent” to every consonant, inherited from Aramaic. 30 Merkelbachʼs view has garnered little attention in Indological literature so far, even though his solution is the only one which includes Anatolia in the range of competing diadochs. Unfortunately we know only that Alexander of Sardis as a vicegerent was on the side of Antiochos Hierax and that he signed a decree in 246 BC. This has no bearing on Aśokaʼs personal chronology. – All four named persons after Antiochos played a role in Seleucid politics during the decades in question where three difficulties needed a solution, first the futile attack of Antigonosʼs father on Seleucid possessions in Anatolia, then the pincer attack of Antiochus II Theos along with Magas on Magasʼ stepbrother Ptolemy II Philadelphos, and then the possibly foreseeable difficulties over the succession of Antio- chus II involving his stepbrother Hierax and his uncle Alexander of Sardis. A personal visit by Aśokaʼs emissaries to Egypt, Libya and Anatolia was not necessary to learn about the Who Is Who in ongoing and pending wars. 31 Strab. 15,2,9 = Majumdar 1981, 98. Bereitgestellt von | Freie Universität Berlin 58   Harry Falk Brāhmī also copies the vowel strokes from Kharoṣṭhī. As a new achievement, it devel- ops a consistent system of linking graphs vertically for two succeeding consonants, a system which was later incorporated into Kharoṣṭhī. The history of the scripts of ancient India shows that both of them have Semitic ancestors. One line of descent runs from a Western Semitic prototype, such as Phoe- nician, to Aramaic, reaching northern Pakistan through Achaemenid administrators where it inspired the creator of Kharoṣṭhī. The second thread likewise starts with Western Semitic and, with Phoenician settlers, leads to Greece and from there, after centuries of standardization, to India through the Macedonian armies. At the court of Aśoka parts of Kharoṣṭhī were united with parts of the Greek script into a system which is capable of rendering Indian idioms almost perfectly. With very few additions it is still in use for most of the Indian languages, be they of Sanskritic or Dravidian origin. In a very short time, two writing systems were designed for an India on the verge of writing. How to employ them? While we do not have the density of documenta- tion required for a full picture, it seems that use of writing remained tied to specific purposes. In the West, writing was used for administration, and after its invention, Kharoṣṭhī may have served just this purpose, to document transactions and court decisions. In Gandhara no court verdicts written in Kharoṣṭhī have been discovered so far, but in the documents on wood from the Taklamakan desert, particularly from Niya, we have mainly court decisions written in Kharoṣṭhī. In addition, we also have one business transaction,32 put to writing and kept in a monastery, probably as a duplicate, which shows that profane business also made use of the means of literacy as present in Buddhist monasteries. In Gandhara and at other sites in Xinjiang, Buddhist texts are reproduced on the oldest manuscripts preserved in this script, proving that this “heretical” religion made use of writing unimpeded by a brahminical preference for oral transmission. With Brāhmī the picture is different. Instantly, we find the edicts of Aśoka spread- ing across mainland India, on stone pillars and on natural rock faces. These texts have a decidedly religious background. Aśoka wanted people to understand his atti- tude towards Buddhist groups, which he personally preferred over other religious movements. He wanted to spread certain ethics for the public, hoping to improve the relations inside families as well as social interactions. He must have been deeply convinced of the importance of his mission, which led him to promote Buddhist com- munities and to meddle with popular, pre-“Hindu” rites, which he felt he could not recommend.33 What inspired him to engage in this new activity? Public laws on dressed stones or rock-faces have a long history in the Near East, and Hammurapi (r. 1792–1750 BC) was not the first to erect stone pillars with laws in public places. The most obvious formal parallels for the Rock Edicts are the rock inscriptions installed by the Achae- 32 Falk & Strauch 2014, 72, § 3.2.4.2. 33 Falk 2006, 56  f. Bereitgestellt von | Freie Universität Berlin  The Creation and Spread of Scripts in Ancient India   59 menid kings at Naghsh-e Rustam and Persepolis, seen by every visitor to the capital, but illegible on account of the distance between the reader on the ground and the texts high up on the rock walls. The Achaemenids had a particular cuneiform script designed for their own language which demonstrated that they were not forced to use the script of subject (and older) cultures, but could command their own means of scriptual expression when necessary. The same urge to express cultural independence can be expected as one of the reasons behind Aśoka’s insistence on using a script of his own. After Alexander, with all nations in close contact, and India important as the main exporter of war elephants, gold and spices, the Mauryan empire needed its own script to avoid being considered a cultural backwater amongst the educated nations. Official documents in Brāhmī from the offices of state administration are extremely rare and seemingly confined to the time of Aśoka. Two such documents mention storing goods for public use for times of distress or famines, both issued by the high officials (mahāmātra) specified as saṃvagiya, probably for Skt. saṃvargya or saṃvargīya. One is a copper plate found in Sohgaura34 the other a flat stone from Mahasthangarh (Fig. 3)35 north of Calcutta, 630 km apart from one another. Both docu- ments are small, the cast copperplate with nail-holes measures 6.5×5 cm, the flat stone with some parts missing about 6×9 cm, and both texts deal with depots of victuals in storehouses (koṭhāgāla) for times of distress (atiyāyika). The script is no different from the one used for Aśokaʼs religious edicts; on the Mahasthangarh stone vertical strokes are used in the way Aśokaʼs scribe at Sassaram36 applied them to indicate rhetoric pauses between words or phrases. The Mahasthangarh stone shows us how official decrees were dispatched: on small flat stones; the Sohgaura plaque shows us that metal was also in use, and the occurrence of ideograms above the letters shows that the time of picture “writing” had not yet passed. These two singular pieces of evi- dence show that in the time of Aśoka official decrees were communicated in writing, and the pause of evidence after Aśoka shows that his efforts were not continued by his successors. There is a manual of state administration, the Arthaśāstra, composed in the decades before Aśoka and enlarged in the four centuries after his time, in which writing plays no role at all, at least in its old parts. Even in the younger passages there is no term for “register”, “index” or “file”. The first text to mention written documents used as evidence in lawsuits is the law book of Yājñavalkya, dating to the second century AD.37 Oral communication and memorization of texts was effective and more in line with common habits. Even the British used servants instead of registers in court offices to remember exactly in which of the hundreds of paper bundles a certain case was filed. 34 Near Gorakhpur; Sircar 1943, 85; Falk 1993, 177. 35 Cf. Sircar 1943, 82; Falk 1993, 180. 36 Hultzsch 1925, 228  ff. 37 Falk 1993, 251. Bereitgestellt von | Freie Universität Berlin 60   Harry Falk The Didactics of Writing Brāhmī, to my mind, owes its shape and logic to a combination of improvements which started with some of those which had already led to Kharoṣṭhī. The didactic means of an alphabet, by which I mean the sequence by which the pupil learns all existing letters of Brāhmī by heart, did not draw upon the a-ra-pa-ca-na hotchpotch of Kharoṣṭhī. Instead, we see on terracotta figurines from the first cent. BC of children holding small writing-pads38 that the sequence of all vowels is followed by a sequence of well-sorted consonants, starting in the classical braminical fashion with the velar group, ka, kha, ga, gha, ṅa. Thus, the sequence of letters to be learned was inde- pendent of older prototypes, be that Kharoṣṭhī (a-ra-pa-ca-na), Greek (α-β-γ-δ-ε-ζ), or Aramaic (’-b-g-d-h-w). This new and systematic sorting made learning the gamut of all defined sounds much easier, and this sorting certainly came about through the influ- ence of brahmins with their age-old classification of vowels and consonants and their similarly mature didactic techniques. Combining the best features from all current systems led to an almost perfect script which needed only a few refinements (gemi- nata, vowel-blocker and two rare vowels) to be usable even for Sanskrit, – and for all of its derivations to this very day. Terms for Writing and Utensils All terms used to describe the technical procedure of writing were introduced from foreign languages into Indian languages. The term for “script/written document” occurs first as libi and lipi in the 4th cent. BC in the sūtras of Pāṇini (3.2,21), as was said above. Aśoka uses lipi in his Rock Edicts, although in the places close to Pāṇiniʼs home on the upper Indus the redactors preferred dipi, a derivation of an Iranian form.39 With some nationalistic undertones, the Indian use of lip(i) for “writing” is often traced back to the notion of “smearing”, innate to a truly Indian root li(m)p. However, the Iranian lipi was also adopted further west in East-Iranian languages40 such as Bactrian, where the term was pronounced δipi, which later led to Bactrian libo for “document”.41 A further case of Iranian influence can be seen at Shahbazgarhi (RE 4,J, K) in the Peshawar valley where the redactor often used the Iranian nipista/ nipesita instead of the Indian term likhita/likhāpita, “written”. Nipista/nipesita can be “directly compared to nipištā, perfect participle from the root paiθ, found, e.  g., on the so-called Daiva-inscription of Xerxes at Persepolis”.42 Already at Mansehra, however, 38 Salomon 2016, 8, fig. 1. 39 Witzel 2006, 460 fn. 7. 40 Henning 1957, 337 fn. 1. 41 Sims-Williams 2002, 227. 42 Falk 2010, 209. Bereitgestellt von | Freie Universität Berlin  The Creation and Spread of Scripts in Ancient India   61 on the eastern side of the Indus, the plain Indian verb likhati was used in place of the loan word nipista. Iran and Bactria also provided other loan words, such as mudrā, “seal”, pustaka, “book”, divīra “writer”,43 and from Greek-speaking groups India borrowed melā, μέλαν “ink”, kalama, κάλαμος “pen”, and melanduka, meraṇḍu etc. “ink-pot” μελανδόχιον.44 Apart from these initial loans, India was very productive in transferring meaning and coining new terms, foremost likh, “to scratch”, for the incising of letter lines in wax tablets or leaves of plants. That being said, our original impression remains valid: modern writing and its terminology came from outside with (late) Achaemenid terms and Greek literacy. Picture Symbols for Pots and Coins Despite Aśoka’s efforts to disseminate writing, one of the most common uses for script outside India was disregarded by his own officials: coins were issued without any written reference to the issuing agency. Instead, the officials continued to use various groups of five symbols for recognition, a pictorial program common from the fourth cent. BC onwards, with each symbol being punched separately into the silver flan. Although the meaning of these symbols is disputed, it seems clear that they carried some sort of significance. It could be that the technique used to create the coins pre- vented the use of letters as maintaining it would have meant using a much larger number of punches, one for each letter, to express all that can be read, for instance, on the coins of the Indo-Greeks. The idea of a single large die imprinting a greater number of letters was obviously not considered attractive. The dynasties succeeding the Mauryas, the Śuṅgas and a number of local rulers, did use script occasionally for coins. Their coinage is mainly made of copper, lead or potin and these softer metals facilitated using larger dies. Brāhmī for Seals There is a certain number of seals for personal uses found all over northern Pakistan and India which show one strange “mis-used” letter: instead of the common sa, in use since Aśoka for the only sibilant necessary for his local language Māgadhī, the die-cutters use a letter which was introduced for a second sibilant śa as required for languages outside Magadha. It seems that this wrong use is restricted to die-cutters, as 43 Witzel 2006, 461. 44 Falk 2010, 211. It seems that the Aramaic clerks left no trace in the vocabulary, despite their cen- turies of clerical dominance. Bereitgestellt von | Freie Universität Berlin 62   Harry Falk Abb. 4: Seals, mirror-inverted as they appear in print: a) Brāhmī śidhathaśa, Kharoṣṭhī sidhathasa (Taxila; photo Aman ur Rahman); b) Br. aśaḍhaśa (ḍha unmirrored for seal) Skt. āṣāḍhasya (unknown find-place; Collection Derek J. Content; c) Br. śaghiyaśa, Skt. saṃghikasya (Khao Sek, Thailand, 3rd/2nd cent. BC; private collection Chumphon province; CNRS – Silpakorn). if this trade had some kind of clerical training of its own. The most famous example is a bilingually inscribed cubical die from Taxila, reading sidhathasa (Skt siddhārthasya) in Kharoṣṭhī and śidhathaśa in Brāhmī, which is not an East Indian dialect form, but a peculiar new interpretation of a letter,45 – just one more piece of evidence that writing in Mauryan times had no century-old tradition but produced some ad hoc decisions, erratic forms, and short-lived misunderstandings. I add a few more seal and token legends all of which contain more than one sibilant, and all reduced to the one written śa, spelled /sa/: aśāḍhaśa, Skt āṣāḍhasya (collection Content), śoradvaraśa, Skt saura- dvārasya (Verma 1983, 26: Mathura Museum acc.no. 72.9), haśikaśa, Skt haṃsikasya (excavation Kumrahar), śaghiyaśa, Skt saṃghikasya (Khao Sek, Thailand, Mauryan times), śunakhataśa, Skt sunakṣatrasya, śuviśakhaśa, Skt. suviśākhasya (Ashmolean; traderʼs token probably from Erich). Much larger is the collection of seals showing no second sibilant but only the final genitive -sa written as -śa. Apart from seals, coins with such a legend are extremely rare and comparatively late; a curious case of a “double genitive” is found on coppers of Rāmadatta from Sonkh, reading ramadata- saśa (Baldwin 2007, 670). Weighing the large number of seals with śa-genitive against other types of epigrams we are led to the suspicion that seal cutters had developed their own orthography which cannot be explained through differences in pronunci- ation.46 45 Strauch 2012, 156. 46 Skt. madhyama, “the middle one”, was a common first name in ancient Sri Lanka. It is tempting to imagine a seal bought in northern India by a Ceylonese, reading majhimasa, “Of Madhyama”, in the popular northern language, pronounced /majimasa/ on the island. This seal would convey the most glaring mistakes of jha for /ja/ and of śa for /sa/. The peculiar form of the ma could also have been included, twice. Bereitgestellt von | Freie Universität Berlin  The Creation and Spread of Scripts in Ancient India   63 Brāhmī in Sri Lanka Sri Lanka shows inscriptions in Brāhmī from an early time, in fact from the time of Aśoka. This is not surprising as Aśokaʼs own son, Milinda (Skt. Mahendra), went south from Pāṭaliputra and propagated Buddhism on the island. We can assume that he or some of his immediate successors had an alphabet on board. Inscriptions from Mihin- tale, where Mahendra had his headquarters, have a rather Aśokan appeal, but at other places other letterforms prevail which seem again to be based on misunderstandings, including the seal-cutters’ blunder, in that they use the pointed śa for spoken /sa/. A similar misconception affected the signs for the soft palatals, ja and jha, in Sri Lanka. Since old Sinhalese – like its modern successor – did not use aspiration due to a close phonetic relationship to Dravidian idioms, the sign for jha had no local phonetic counterpart and could be used for ja without any loss in spoken realisation. As a result, rājā, the king, would be written rājhā, adding aspiration where none was required, and the Buddhist order, saṃgha, came in a graphical form which would be read as śaga at other places, dropping the aspiration where it was expected. The earliest evidence comes from dedicatory legends incised inside or above the entrance of caves, cut to divert the rains and financed by laypeople donors. Many legends give the impression that the stonemasons responsible for these dedicatory legends drew upon a standard set of phrases and used them mechanically, with lots of mistakes, as though to them this “literary” language was a foreign idiom, and writing an as of yet ill-known craft. Monks kept on coming to Ceylon from all quarters of India, with the clerical skills acquired at home, and so this idiosyncratic local script system was supplemented by and corrected to standard Brāhmī again and again. By the first century AD the old and irritating misuses of letters had completely disappeared. Apart from cave dedi- cations, we have a great number of ownership inscriptions on dining and drinking vessels, certainly incised by the owners themselves. They again display a wide range of letter-shapes and also demonstrate that female helpers at the Buddhist monasteries knew how to write at least to the same degree as their male colleagues.47 Once an illiterate is confronted with writing for the first time he may be surprised, stupefied or enthusiastic. He or she may think of uses the inventors at home had never thought of. In Rahman & Falk48 I proposed that the so-called nandivardhana sign may in fact express nothing but the term namo, “veneration”, being composed of the Brāhmī letters na and mo. A certain dissatisfaction arose because in India, even in the oldest evidence at Sanchi and Bharhut, the two arms of the mo only rarely show the required form with arms extending sideways,49 while in almost all cases shallow crescents sit symmetrically on the raised arms of the ma. However, when we turn to 47 Falk 2014. 48 Rahman/Falk 2011, 26. 49 Cf. Cunningham 1854, pl. XXXII, 20. Bereitgestellt von | Freie Universität Berlin 64   Harry Falk Ceylon, we find exactly what I had in mind, a ma with arms outstretched sideways and not centrally. Evidence comes from saddle querns buried below the circumambulation path of a stūpa at Tissamaharama,50 in the very South of the island, one of the earliest centres of Buddhism on the island. The idealised form is found in Bharhut panels usually added to veneration scenes with people kneeling in front of thrones or relic containers. Sri Lanka, however, preserves this prototypical form. I would interpret this as an act of give and take: Brāhmī was brought to Ceylon, and the combination of letters into a monogram seems to have been formed there and was then re-trans- ferred back to India, where people knew its general meaning, and used the symbol for veneration scenes, but in a shape already removed from its monogrammatic origins. Summary In sum, the culture of writing in India shows several aspects that are unique by com- parison with other ancient cultures: – The introduction of writing for Indic languages was twofold, both arising from the same political conditions which exerted their influence on two places 1600 km apart: – a) After the forced linkage with Achaemenid Iran had ceased to affect Gandhara and when the Seleucid dynasty had set new standards, financial transactions con- tinued to require the juridical stability offered by written contracts. As the old and foreign speaking clerical caste was slowly disappearing, the economically influ- ential groups needed a new script and had Kharoṣṭhī designed to express the local language for all future needs. For the new script the pre-existing Aramaic script was copied to a certain extent, but mis-allocations of signs show the non-Aramaic nature of the design-shaping agency. – b) Until the time of the Seleucids the area of Magadha was devoid of any writing. The introduction of writing in the form of the Brāhmī script was not a process necessitated by legal procedures as in Gandhara, but was an act of cultural adap- tation and was used mainly for cultural purposes. A single person, king Aśoka, appears as the dominating force behind an entire range of innovations, which turned the tide, transforming a self-content but illiterate society cultivating an incredible mass of orally preserved literature into a society with breath-taking new tokens of civilization. Aśoka, and he alone, effected the public proliferation of writing with his rock inscriptions, he had huge pillars and other monoliths cut from stone, and he dotted the whole wide realm with stūpas, visible evidence for the high morals of his subjects and himself. Despite these changes, business, administration, and coinage would go on without making large-scale use of 50 Weisshaar 2014, 123, fig. 4g, 4h. Bereitgestellt von | Freie Universität Berlin  The Creation and Spread of Scripts in Ancient India   65 writing. For the new script, inspiration was taken from familiar scripts, including Kharoṣṭhī, but the result surpassed all previous models in both aesthetics and usability. – Both newly invented scripts were not meant to remain confined to political elites, ethnic groups or a particular sex. However, there were some groups that adopted the new technique faster than others. Buddhist monks were far more inclined to take up the skills for their sacerdotal literature, while brahmin priests withheld their consent for quite a while; Buddhist nuns were at least as interested and dexterous in writing as their male counterparts. – As the art of writing spread fast and wide all over India, some particular groups developed idiosyncratic features: early North-Indian seal cutters used a “wrong” sibilant for the genitive ending, which Sri Lankan stone-masons seem to have imported along with one more misunderstanding. Some particular font styles of Brāhmī are found exclusively on seals or on coinage, while some Aśokan types lived on unaltered for two hundred years and more. That means that there were different strands of literacy in different professional groups for a long time, which seems to speak against a predominant agency intent on standardizing writing for all parts of society. Bibliography Baldwin 2007 = Baldwin’s London auction 50, April 2007. Baums 2012 = S. Baums, Gandhāran Scrolls: Rediscovering an Ancient Manuscript Type, in: J.B.D. Quenzer/D. Bondarev/J.-U. Sobisch (ed.), Manuscript Cultures: Mapping the Field, Berlin 2012, 183–225. Chakrabarti 1976 = D.K. Chakrabarti, Rājagṛiha: an Early Historic Site in East India, World Archaeology 7, 1976, 261–268. Cunningham 1854 = A. Cunningham, The Bilsa Topes; or, Buddhist Monuments of Central India, London 1854. Deeg 2005 = M. Deeg, Das Gaoseng-Faxian-Zhuan als religionsgeschichtliche Quelle, Wiesbaden 2005. Deshpande 2011 = M.M. Deshpande, From Orality to Writing: Transmission and Interpretation of Pāṇiniʼs Aṣṭādhyāyī, in: Travaux de Symposium International Le Livre. La Roumanie. LʼEurope. Troisième édition – 20 à 24 Septembre 2010, Tome III: La troisième section – Études Euro- et Afro-Asiatiques, Bucarest 2011, 57–100. Erdosy 1995 = G. Erdosy, City States of North India and Pakistan at the Time of the Buddha, in: F.R. Allchin (ed.), The Archaeology of Early Historic South-Asia, Cambridge 1995, 99–122. Falk 1993 = H. Falk, Schrift im alten Indien – Ein Forschungsbericht mit Anmerkungen, Tübingen 1993. Falk 1996 = H. Falk, Aramaic Script and the Kharoṣṭhī – A Comparison, Berliner Indologische Studien 9/10, 1996, 151–156. Falk 2006 = H. Falk, Aśokan Sites and Artefacts, Mainz 2006. Falk 2010 = H. Falk, Foreign Terms in Sanskrit Pertaining to Writing, in: A. de Voogt/I. Finkel (ed.), The Idea of Writing: Play and Complexity, Leiden 2010, 207–217. Bereitgestellt von | Freie Universität Berlin 66   Harry Falk Falk 2014 = H. Falk, Owners’ Graffiti on Pottery from Tissamaharama, Zeitschrift für Archäologie Außereuropäischer Kulturen 6, 2014, 45–94. Falk/Strauch 2014 = H. Falk/I. Strauch, The Bajaur and Split Collections of Kharoṣṭhī Manuscripts within the Context of Buddhist Gāndhārī Literature, in: P.M. Harrison/J.-U. Hartmann (Hg.), From Birch Bark to Digital Data: Recent Advances in Buddhist Manuscript Research: Papers Presented at the Conference Indic Buddhist Manuscripts: the State of the Field, Stanford, June 15 – 19, 2009, Wien 2014, 51–78. Frye 2006 = R.N. Frye, The Aramaic Alphabet in the East, Journal of Inner Asian Art and Archaeology 1, 2006, 57–60. Härtel 1991 = H. Härtel, Archaeological Research on Ancient Buddhist Sites, in: H. Bechert (ed.), The Dating of the Historical Buddha / Die Datierung des historischen Buddha. Symposien zur Buddhismusforschung, IV,1, Göttingen 1991, 61–89. Henning 1957 = W.B. Henning, The Inscriptions of Tang-i Azao, BSOAS 20, 1957, 335–342. Hinüber 1989 = O. von Hinüber, Der Beginn der Schrift und frühe Schriftlichkeit in Indien, Wiesbaden 1989. Hultzsch 1925 = E. Hultzsch, The Inscriptions of Asoka, Oxford 1925. IA-R Indian Archaeology – A Review. Jacobs 2016 = B. Jacobs, Beschreibung von Palibothra und die Anfänge der Steinarchitektur unter der Maurya-Dynastie, in: J. Wiesehöfer/H. Brinkhaus/R. Bichler (ed.), Megasthenes und seine Zeit / Megasthenes and His Time, Wiesbaden 2016, 5–26. Karttunen 1989 = K. Karttunen, India in Early Greek Literature, Helsinki 1989. Majumdar 1981 = R.C. Majumdar, The Classical Accounts of India, Calcutta 1981. Merkelbach 2000 = R. Merkelbach, Wer war der Alexandros, zu dem Aśoka eine Gesandtschaft geschickt hat?, Epigraphia Anatolica 32, 2000, 126–128. Müller 1859 = M. Müller, On the Introduction of Writing into India, Journal of the Asiatic Society of Bengal 28, 1859, 136–155. Naveh/Shaked 2012 = J. Naveh/Sh. Shaked, Ancient Aramaic Documents from Bactria (Fourth Century B.C.E.) from the Khalili Collections, London 2012. Rahman/Falk 2011 = A.U. Rahman/H. Falk, Seals, Sealings and Tokens from Gandhāra, Wiesbaden 2011. Salomon 2012 = R. Salomon, Gandharan Reliquary Inscriptions, in: D. Jongeward/E. Errington/ R. Salomon/S. Baums (ed.), Gandharan Buddhist Reliquaries, Seattle/London 2012, 165–199. Salomon 2013 = R. Salomon, On Alphabetic Ordering: Some Principles and Problems, Scripta 5, 2013, 1–20. Salomon 2016 = R. Salomon, Siddham, Across Asia: How the Buddha Learned his ABC (23nd J. Gonda Lecture 2015), Amsterdam 2016. Sims-Williams 2002 = N. Sims-Williams, Ancient Afghanistan and its Invaders: Linguistic Evidence from the Bactrian Documents and Inscriptions, in: id. (ed.), Indo-Iranian Languages and Peoples, Proceedings of the British Academy 116, 2002, 225–242. Sircar 1942 = D.C. Sircar, Select Inscriptions I, Calcutta 1942. Verma 1983 = T.P. Verma, A Catalogue of the Seals and Sealings in Government Museum Mathura, Mathura 1983. Weisshaar 2014 = H.-J. Weisshaar, Legged Saddle Querns of South Asia, Zeitschrift für Archäologie Außereuropäischer Kulturen 6, 2014, 119–144. Witzel 2006 = M. Witzel, Brahmanical Reactions to Foreign Influences and to Social and Religious Change, in: P. Olivelle (ed.), Between the Empires – Society in India 3000 bce. to 400 ce., Oxford 2006, 457–499. Bereitgestellt von | Freie Universität Berlin