Word-dividers in the Greek Script of the
Kushan Period of Afghanistan
Abdul Hai Habibi
Alexander of Macedonia arrived in Ariana (Khorasan of the Sassanid and Islamic periods) with his victorious army in 330 B.C. and was confronted with the great opposition by the people of this land for a period of three years. He reached Taxila and the bank of the river Jehlum in 326 B.C. and one year later made for Iran by a route that ran to the south of Afghanistan, through present day Baluchistan.
The arrival of Alexander in Afghanistan was an important event since it had a deep and significant impact on the politics and culture of later day Afghanistan and all the other countries of central Asia. For more than ten centuries, until the conquest of the Arabs and Islam, the effects of the Greek conquest could be evidently felt on the literature, art and culture of this land.
The Greek descendents of Alexander established a mighty empire in Ariana and Transoxiana. Thirty six kings and one queen ruled for a period of 250 years, the last of whom was Hermaeus who was overthrown by the Saka tribes from Kabul in 30 B.C.
Hellenism lasted in Afghanistan up to the 7th century A.D. Its imprints on the culture of this land are evident in art, architecture and script.
Graeco-Bhuddic art, which spread throughout eastern Afghanistan and Gandahara, constitutes a very important chapter in the history of art in Central Asia. At this juncture, remains of Greet architecture are being unearthed in a great Greek city of Ai-Khanum. In the present discourse we shall confine ourselves to a single subject i.e., Greek script in Afghanistan.
Of the coins minted by Greek kings in Bactria, a great number are preserved in museums in Kabul, London, Paris, the USSR, India and Pakistan. The number of such coins in private collections is also considerable. Some of these coins carry legends only in Greek script, while others have legends in both Greek and Kharoshti. Kharoshti script, which was common from India to China, has been called the Arianian alphabet by H.H. Wilson in his book Ariana Antiqua (p. 262).
But the Greek script, which has become popular with the coming into power of Bactrian Greeks in Afghanistan, Transoxiana and western India, was perfected in accordance with and updated to the needs of the people of this area with letters and sounds found in the languages common here having been incorporated in it. In addition to having been used on coins by the Bactrian Greeks, the Sakas, the Kushanids, the Epthalities, the Kidaras, the Lions of Bamian, etc., it has been employed in inscriptions on rocks from the first to the 7th century A.D. From this it could be inferred that besides the Kharoshti, Brahmi, Sarada and Pahlavi scripts, Greek script was also used in court writings.
As far as is know, so far, in addition to coins, the following rock inscriptions, written in this script, have been discovered in Afghanistan. They belong to the Kushanid and Epthalite periods.
1. A twenty five line inscription on a large flat rock. This inscription, which has been unearthed at Surkh Kotal of Baghlan from the great temple of Kanishka, contains 160 words in cursive Greek alphabet (Kabul museum).
2. Two other copies of the same inscription dug from the walls of a well in this temple. These inscriptions are written in coarser script than the first one (Kabul museum).
3. Two inscriptions on large rocks discovered near Qala-e Achakzai in Shali Pass of Rozgan, 150 km. north of Kandahar, in the heart of Zawolistan. Three lines of these containing about 50 words are readable. It seems that they were inscribed at the orders of an Ephtalite prince of Zawol.
4. Two inscriptions at Jaghatu, 20 km. northwest of Ghazni, one of which has three lines with 25 words. It is the Buddhist Triatana Formula or Three Jewel Formula in Greek script resembling that common during the Epthalite period.
5. The second Jaghatu inscription consists of six lines and has been written in the same script. It contains 25 words and could have been written in 6-7 centuries A.D. It’s inscription could be attributed to a Sagasi Shapor (Sagazi Prince) Vim Sha-Ulugh?
It is for a century that research into the relics and decipherment of inscriptions left by the Bactrian Greeks, the Kushanids and Ephtalites has been going on. As most of the researchers were not well-conversant with the indigenous languages of Afghanistan and their dialects, they have made mistakes in reading these writings.
The language in which, during the Epthalite and Kushanid periods, the above mentioned inscriptions have been written was without any doubt an Arian language that we could call the Bactrian or Tukhari or Kushanid language of Afghanistan. According to Al-Buruni in Athar-al-Baqua (p. 222) and Al-Bashari Maqdisi in Ahsan-ul-Thaqasim (p. 335) the language spoken by the people inhabiting the area between Balkh and Badakshan was called Tukhari, which had kinship with the language spoken in Balkh. Tukhari is contemporaneous with Pahlavi. At the time when southern Pahlavi was spoken in the court of Pars and northern Pahlavi existed in Transoxiana, Khutan and Thoorfan, Tukhari was spoken in Afghanistan. It is considered the mother of the Dari of Afghanistan. Tukhari was the court language of the Kushanids and Ephthalites, Etymologically, synthetically and phonetically it also shows similarity to Pashto. In my book The Two Thousand Old Language of Afghanistan I have dealt with this matter in detail.
Greek script was rearranged in accordance with the phonetic needs of the languages of central Asia, insofar as Greek script written from right to left has been found at Taigak of Salqataw mountain in Alma Ata region. Tukhari possessed sounds which were unknown to Greek. As Greek letters could not represent those sounds therefore new letters were invented for them, for example: the sounds ϸش and ϸх ښ both of which are included in present day Pashto alphabet. According to Huen Tsang the alphabet of this land was composed of 25 letters and was written from left to right. This proves that around 644 A.D., at the time of Huen Tsang;s visit Greet script was common in Afghanistan.
In this alphabet the letter omicron, which has the shape of a small circle and corresponds to the English letter ‘o’, was, at first, used for its special sound, but later on it came to be used as a word-divider on the coins of the Bactrian kings. The word-divider made reading easier by separating the words from one another, but, as after the passage of many centuries, it was forgotten, it became a cause of confusion for recent readers of the script, because they thought that the sound ‘o’ was a part of the original word. Thus this sound was added to the end of every word producing aberrant forms.
As far as is known the word diveder ‘o’ can not be seen on the coins of the first Greek kings of Bactria like Theodotus, 256 B.C., Euthydemus, 190-220 B.C., Demetrius, 190 B.C., Eukratides, 181 B.C. and others, and the words on these coins succeed one another without any separating space. But, on the unique square-shaped coin of queen Agathokleia (the servant of God) the word BAZAILIZAS is followed by an ‘o’ on the left margin of the coin behind the queen. The word EOTRON is inscribed above the queen’s head and is likewise followed by an ‘o’. On the left margin of the coin, facing the queen, her name Agathokleia is inscribed. Thus, it is evident that this sign entered Greet script in Afghanistan in the last century B.C. and became more common after Christ, in the Kushanid period. It is seen more in the beginning of he Christian era and after it in the inscriptions if the Kushanid and later periods.
When the reading of legends on Graeco-Bactrian coins was first undertaken many mistakes were made in deciphering sounds of the Greek letters of the Kushanid period. Wilson’s book, Ariana Antiqua, written around 1840 A.D., is devoted to reading the coins just mentioned. Mistakes were especially made in deciphering the sounds of two letters, namely, the omicron, which was a word-divider or word terminator, but which was thought to be the last letter of words, and the sound sh which was mistaken for the Greer rho, despite the fact that in Bactrian Greek alphabet it had the half circle in the middle and not at the top ϸ. Thus the word SHA or SHAH meaning king and inscribed ϸα-ο on the coins was thought to be RAO, a word of Indian origin. Sh was read as r and the word-divider ‘o’ was considered a part of the word itself. The remaining part of the legend KUSHANSHA and SHANANSHA (Shahansha) was incorrectly read as RAO NANARAO KANIRKI KORANO. This legend in Tukhar found on Kushanid coins has been inscribed instead of the Greek BACILEYC BACILEWN. The author of Ariana Antiqua has unconvincingly tried to establish the Indian origin of these words, while the actual words are evidently SHANANSHA KANISHKI KUSHAN.
These mistakes ought not to be repeated as the Surkh Kotal inscriptions have helped us in deciphering the actual sounds and forms of the Greek alphabet of the Kushanid period. A part of this syntax of and transformation of the letters and words in Tukhari have also been determined.
When the Surkh Kotal inscription became the subject of the study of scholars, M. Maricq in Asiatic Journal, M. B. Henning in Bulletin or Oriental Studies and Benveniste in Asiatic Journal, 1961 were the first to publish their studies on it. But, as all these studies were of a preliminary nature, the authors have complained that it is difficult to read this inscription, because the words have not been spaced, thus making the task of word separation a difficult one.
As far as I have studied the Surkh Kotal, Rozgan and Jaghatu inscriptions, in all of them, written in the Greek script, a word-divider in the form of ‘o’ exists, and if the word-dividers are struck off from the ends of words, the words obtained are very meaningful and original. These words are found in ancient Dari and Pashto literature.
The word SHA equivalent of shah was read SHAO, its word-divider having been included in it and KIRD, kard of Dari, was read as KIRDO mutilating these two very original words. Similarly, hundreds of other words have met with the same fate in this spree of incorrect reading.
In all the discovered remains of he Greek script of Afghanistan a small circle is observable at the end of every word. This circle serves as a word-divider which should not be considered the last letter of the word it follows. Sometimes it has the form of a small square.
Word-dividers having the form of < \ are seen in the cuneiform characters of Achaemenian inscriptions.. In the Surkh Kotal inscription of Baghlan, the small circle can, in no way whatever, be considered a part of the word. The very original words, BAGLANG, MANDAR, BAGPHOR, AB, NARD, NOBIXT, FROMAN, BORZHIHR and NEICAN should not be read in their mutilated form of BAGOLANGO, MANDARO, BAGOPOHRO, ABO, KANDO, NOBIXTO, FROMANO, BORZOMIHRO and NEICANO. It is possible for all the words of a language to end in ‘o’ and for no word not to end in it in several inscriptions containing two hundred words in all?
Now we shall focus our attention on a number of coins of that period which legends in the Greek script. The symbol ‘o’ was usually used for dividing words and it could be seen on the coins of emperors Gondophares, who ruled in the first century A.D. (19-48 A.D.). On a coin belonging to Hermaeus, HERMAIOZ-O is inscribed under the figure of the king and is separated from BAZILEUZZUTER-O above by the symbol ‘o’ as shown.
From a general study of the Graeco-Bactrian and Kushanid coins it could be established that this symbol was seldom used with Greek words proper, but when instead of the Greek BACILEYC BACILEWN, the Dari title SHANANSHA was used, the word-divider was put beside it. Use of the word-divider was one of the characteristics of the Greek script of the language of the Kushanid period. For instance:
1. One of the coins carries this legend. KANISHKIKUSHAN SHA-NAN-SHA. It begins from eleven o’clock and the word-divider was used in the following manner: ΚΑΝΗϸΚΙ ΚΟϸΑΝϸΑ-Ο-ΝΑΝ-Ο-ϸΑ-Ο. From this it is evident that the title of emperor was composed of three parts, namely, SHA+NAN+SHA.
2. On another coin at three o’clock is inscribed ΚΑΝΗϸ-Ο. The word-divider as can be seen is at the end of the word.
3. An Okshki coin (around 160 A.D.) carries the following legend beginning from eleven o’clock: ΟΟΚϸΙ ΚΟϸΑΝϸΑ-Ο-ΝΑΝ-Ο-ϸΑ-Ο.
4. One of the very strong proofs of the symbol ‘o’ being a word-divider or word terminator is that on most of the Kushanid coins the names of gods have been inscribed from which, with the deletion of the omicron, the original Pashto or Dari words are obtained, for instance:
ΟΡΛΑΓΗ-Ο Orlagi on a coin of KanishkiKushan Sha Nan Sha. The Wardagi of Pashto has been derived from this word.
On another coin of KanishkiKushan Sha we have ΑΘϸ-Ο Atsh (atash of Dari). ΜΑ-Ο MA on a coin of Kanishki (mah of Dari).
ΑΡΔΟΧϸ-Ο Ardosh on a coin of Kanishki (Ardoxsh was the name of a god). ΜΙΟΡ-Ο. Mir on a coin of Kaniskhi (mir of Pashto and mehr of Dari). Also written as ΜΙΡ-Ο or ΜΙΙΡ-Ο.
ΟΑΔ-Ο Wad on a coin of Kanishki (baad of Dari)
ΦΑΡ-Ο Far on a coin of Ohshki (far of Dari)
ΟΝϸ-Ο Wish on several coins (wesh of Pashto).
All these words prove that the terminal ‘o’ is not a part of the word. These words have preserved their original form even to this day. On the coins of later Sassanid kings, discovered at Hadda and Bagram, sometimes Greek letters have been used. On two of these coins, small in size, on the obverse the figure of a king with a Sassanid crown is seen, while on their reverse there is the picture of a temple. The king faced to the right and in front of him, starting from two o’clock ϸΑΒΟΡ. It is claimed that it was minted in Balkh around 356 A.D. From the above it could be inferred that the Sassanids of Persia at times did and at other did not use this symbol.
A study of he three inscriptions of Baghlan will show that the differences among them are trivial. In the first the words Sha and Ma have been inscribed as ϸΑ-Ο and ΜΑ-Ο with the word-divider having been placed after the last letters., But, in the second inscription the words have been written this: ΜΑΥ-Ο and ϸΑΥ-Ο. In the first instance the words should be read as Sha and Ma and not as Shaho and Mao because ‘o’ is the word terminator. On the same basis, the words in the second inscription are Shah and Mah and not Shaho and Maho as ‘o’ is not the last letter of the words. It seems that at that time, too, these words had two written forms each, namely, Sha and Shah, and Ma and Mah. In Dari literature, too, the word shah is sometimes written as sha. In the first and third inscriptions we have ϸΑ-Ο and ΜΑ-Ο, but in the second one it was intended to add an epsilon , Υ=H to Sha and Ma, therefore the epsilon has been attached after Α and the word-divider added after it. This proves that ‘o’ after the words, Sha and Ma was not their original part. When it was intended to add the sound h or epsilon to the end of the words, the word-divider was placed after the epsilon and the words were written as ϸΑΥ-Ο and ΜΑΥ-Ο.
In addition to the three inscriptions just described, the word-divider ‘o’ is also seen in one of the Rozgan inscriptions with the word-dividers present. The inscription reads: the great prince Mir Sang Zmik?
The symbol is also seen in an inscription at Jaghatu of Ghazni stating the Buddhist Tri-Ratna.
The Sanskrit verse of he Tri-Ratna is as follows:
Namo Buddhasya, reverense to the Buddha
Namo Dharmasya, reverence to the Dharma
Namo Sanghasya, reverence to the Sangha
In Sughdi the Tri-Ratna was written in the following manner with no ‘o’ at the end of the words PUT, DARM and SANK: NM’W PWT.NM’W DRM.NM’W SANK, reverence to the Buddha, reverence to piety and reverence to society. It should be noted that in the Greek script there are three word-dividers between the six words, each word-divider dividing the words of a line.
In another Jaghatu inscription words and word-dividers have the following relationship: BAG (O) SAGASISHAPOR (O) VIM (O) SHA (O) ULUGH? Doubt exists as to the correct reading of the second word. But according to the rules of the script four word-dividers have been placed among five words. In Dari it would read as Bag Sagzishapur Vim Sha Ulugh.
In the Islamic period remnants of this word-divider are seen in some copies of the Holy Koran. It has been used for separating verses from one another and has this very same form of ‘o’. Diacritical marks and vowelization were not used in the old Kufic script. It was Abu-al-Aswad Doueli (death in 96 H, 714 A.D), who used diacritical marks for vowelization. It is thought that he may have copied them from the Caldians. According to Ibn-i-Khallikan, Hajjaj, the governor of Iraq, employed punctuation to distinguish Arabic words of similar form. From the preceding statements it becomes evident that the Arabs had not need for vowelization and punctuation, but in order to make the reading of Arabic easier for non-Arabs, they adopted punctuation and vowelization from other nations. For example, Khalil Ibn-i-Ahmad, who died in 170 H (786 A.D.) arranged Arabic letters in Kitab-ul-Ain in accordance with the Sanskrit alphabet.
During the reign of Abd-ul-Malik son of Marwan Umavi a point was put at the beginning and end of verses of the Holy Koran. Later on it was changed to three points.
As far as is known, these symbols have not been used in the old Kufic copies of the Koran, but they became common in Khorasanian copies after the 4th century Hejera. The present system of the separation of Koranic verses is that of Sujawandi. It is probable that Sujawandi adopted the symbol of space ‘o’ from ancient Kushanid works. Afterwards it was the custom of scribes and painters to add a variety of decorative marks to it and paint it in gold and many other colors. But these symbols were not common in Kufic copies of the Koran before the age in which Sujawandi lived.
 Kabul, 1963.
 Archiv Orientalani, 35, 1967.
 Ariana Antiqua, p. 358.
 No. 4, 1958, Paris
 London University, 1960
 N.S. p. 30, 1937.
 History of Islamic Civilization, p. 111, 80.
 Wafiyyat, 1, 135.
 Islamic Civilization p. 111, 115.
 Al-Ithqan, 1, 84.
 Encyclopoeadia of Islam, IV, 54.