TheAngkola,Mandailing…ipll.manoa.hawaii.edu/shared/Mandailing 06022011.pdf · Angkola-Mandailing ethnic group, which is by and large ... both script and language may show influences

The  Angkola-­‐Mandailing  Script:  A  Historical  Perspective  

Linguistically as well as culturally the Mandailing people belong without any doubt to the wider group of the Batak people that are commonly grouped into five distinct ethnic groups, namely Angkola-Mandailing, Toba, Karo, Simalungun, and Pakpak-Dairi.

The ethnic groups of northern Sumatra are

predominantly defined by the languages they speak. According to Adelaar (1981) there are two main branches of the Batak languages, the northern branch

comprising Pakpak-Dairi and Karo and the southern branch consisting of Simalungun, Toba and Angkola-Mandailing. Linguistically the Alas also belong to the northern Batak branch of languages. However, as the Alas people have been Muslims for a long time, their culture has undergone considerable changes so that culturally they cannot be considered as Batak. The three northern dialects Alas, Karo, and Pakpak-Dairi are all mutually intelligible, but they are different enough to be clearly discernible as separate languages. The differences between these three languages or dialects, are not much larger than the differences between Danish, Norwegian, and Swedish.

Simalungun is an early offspring of the southern group whereas Toba and Angkola-Mandailing split later. Simalungun has in fact a kind of an intermediate status between the languages of the northern and the southern branch. In the border region of Simalungun and Karo the languages are mutually intelligible, and the same is true in the border regions between Simalungun and Toba. But the other languages of the southern branch are unintelligible to speakers of the languages of the

Figure 1 The Batak languages

northern branch. Due to increased communication between speakers of Mandailing and Toba Batak, and especially due to the role of the national language of Indonesia which has deeply penetrated all Batak languages, Mandailing and Toba are now more intelligible than it may have been the case in the past:

I realized on my journey [to Mandailing, U.K.] a great difference in clothing, architecture, shape of burial monuments, etc. but I was delighted to see that the script was very similar to that of the other places. The language was, especially in Mandailing, so different from Toba Batak, that people were not able to understand neither me nor my teacher who accompanied me on my journey. But it is also evident that the language can be mastered within a few weeks by those who have studied Toba Batak, while in Sipirok the language is less different so that we were universally understood. (Van der Tuuk 5-5-1852, cited in Groeneboer, 2002, p. 133)

Adelaar lists Angkola and Mandailing as separate languages. There are certainly sufficient differences between Angkola and Mandailing, not only linguistically, but also culturally to regard

them as two separate ethnic groups where Angkola has a kind of an intermediate status between Mandailing and Toba. On the other hand one can also emphasize what the two groups have in common to justify the classification of Angkola-Mandailing as one single ethnic group. Whether Angkola and Mandailing comprise one single, or rather two separate ethnic groups is still a matter of discussion. Ethnicity is, of course, a complex issue, especially among the Batak where clear-cut linguistic and cultural boundaries are missing, but generally there is nowadays a tendency in favor of one single Angkola-Mandailing ethnicity. As this paper is mainly concerned with the script of the Angkola-Mandailing ethnic group, which is by and large identical, we will from now on use the term Mandailing instead of Angkola-Mandailing.

Angkola and Mandailing are not the only Batak people who have made the, not undisputed, decision to avoid segregation. The Toba Batak, the largest of the five ethnic groups, live in districts that are geographically often very diverse. The tropical hot Batak Toru valley is considerably different from the cool climate of Toba at the shore of Lake Toba where the relative fertile soil supports a high population density. Habinsaran, on the other hand, is extremely sparsely populated and consists of rugged mountains with often relatively cool climates due to the high altitude. However, all these Toba Batak sub-groups are genealogically, culturally, and linguistically so closely related that they now are widely, though not undisputedly, considered as one single ethnic group.

Generally, the division of the Batak into five cultural-linguistic groups is widely accepted, not only by scholars but also by the people themselves1. Together with the ethnicization of the Batak people into five distinct ethnic groups, the Batak script is also commonly believed to perfectly follow 1 Nowadays the term Batak is only used by the Toba Batak whereas the other ethnic groups are often reluctant to refer to themselves as “Batak”.

Figure 2 Family tree of the Batak languages according to Adelaar (1981)

the agreed ethno-geographical boundaries. Too some degree it indeed does, but certainly not perfectly as there is considerable overlap on one side, and a relatively high degree of local variants on the other side. Recently, the Batak script was codified by the Unicode consortium, and it was agreed that the regional traditions should be seen as ‘graphical dialects’ within one scriptorial tradition rather than as separate scripts.

Written  Literature  Folk tales, fables, legends, and myths were usually orally transmitted and rarely recorded in

writing – except on request of a foreigner. The Dutch linguist H.N. van der Tuuk, the first person ever to study the Batak language, published a four-volume reader of Batak stories of which the third volume contained prose from Mandailing (Tuuk, 1861b). Although the Batak possessed a writing system for several hundred years, they hardly ever used their script for administrative, legal, or economic purposes.

Instead, most of the literature they produced belongs to the realm of magic, the hadatuon – the “science” of the datu, the Batak magicians and healers. The datu, or guru as they were also called, wrote down their knowledge on leporello shaped books made from the bast (the inner bark) of the alim tree (Aquilaria malaccensis and other Aquilaria species), but occasionally they also used bamboo or buffalo bones as writing mediums. There are an estimated one thousand pustaha as these bark books are called, in various collections, mainly in the Netherlands and in Germany, and a considerable larger number of bone and bamboo manuscripts.

The datu were the professional scribes of the Batak, but in the egalitarian society of the Batak every person, man and woman, could learn the art of writing, and many did, especially the young men, because it enabled them to be engaged in the art of writing love laments that were used in courtship rituals. Knowing to write was also important for the traditional leaders, the raja, who occasionally would exchange letters written on bamboo nodes, and, last but not the least, for everyone engaged in an unsolved legal dispute and who choose the ultimate path of getting justice by writing threatening letters known as pulas on small bamboo nodes. In order to reinforce the threat, little miniature weapons made of bamboo were added such as spears, a knife, a booby trap, and flint and tinder.

It is difficult to estimate the degree of literacy in the pre-colonial society (approximate dating needed) of the Batak as the early reports often contradict each other, but it seems to have been relatively widespread. A very rough estimate is that about 30-50% of male Batak were literate. Female literacy rates seem to have been considerably lower, probably not higher than 10%.

In the pre-colonial Batak society people were reluctant to travel unless they really had to, or if they did so professionally such as the many traders who brought horses and forest products to the lowlands and returned with salt, iron, silver, cloth and other products from the coast. These trading routes were relatively safe, and traveling traders could be easily identified as such. Everyone else who traveled arose the suspicion of the villages he had to pass through with the risk of being held in captivity.

The datu, on the other hand, were protected, and allowed to freely roam within the boundaries of the Batak lands. The pupils (sisean) of a datu often traveled long distances in order to learn from a well-known guru. In Manuscript D2 of the National Library of Indonesia, it is said that the writer of the bark book, originally from Pangaribuan in Habinsaran, south of Lake Toba, traveled to the Karo high lands north of Lake Toba in order to study with a famous datu there who was also the raja (sibayak) of Kuta Bangun. The distance between Kuta Bangun and Pangaribuan is more than 250km.

Because of the high mobility of the datu, it is often difficult to ascertain the provenance of a particular manuscript. It is quite common that bark books, which are usually written in a particular

idiolect called the poda language, contain words from the local dialect where it was produced, and both script and language may show influences of one or more than one Batak dialect, depending on the origin of the writer of the pustaha and the place where he learnt his skills. The mobility of the datu and their apprentices (who often originated from other regions) has also contributed to the fact that the borders between the scriptorial traditions are relatively unclear.

The  language  of  instruction  In order to write a pustaha, a datu needed the knowledge of the so-called poda language (language

of instruction). This is an ancient southern Batak dialect with a large number of Malay loanwords. The datu acquired knowledge of this secretive language during his apprenticeship with his teacher. It is very difficult, and in most cases impossible to translate a pustaha not only because of the nature of the poda language but also because these books were not meant to be read. The served the datu more as a mnemonic aid, and as a vehicle to teach his knowledge to his pupils. Texts in pustaha always assume that the reader of the text already has some knowledge of the nature of the text, which can be an instruction to conduct a ceremony, some kind of divination, or a prescription for a medicine.

Mastering the poda language is a prerequisite for composing a pustaha, but the texts also always contain words from the vernacular. Because of the localized nature of the text it is often possible to roughly locate the region where the text was written. The provenance of the text is in most cases also evident from the script that has been used in composing the text.

Table 1 The Language of Instruction

Mandailing Hata Poda Malay molo jaha jika habinsaran purba timur dohot dongan dengan tuat, susur turun turun on inon ini jolo lobe dulu uli bajik bajik ro dumatang datang di- da- (passive) di-

This list is extendable with dozens words from the language of instruction. Example of hata poda

words are given in the works of Sarumpaet (1982) and Hariara (1987), but the best source remains Van der Tuuk’s (1861a) excellent Batak-Dutch dictionary. The list of loanwords is quite extensive: jaha, dongan, turun, bajik, d=um=atang, are all borrowed from Malay while purba is derived from Sanskrit. The heavy reliance on foreign words is particularly due to the fact that foreign concepts, foreign materials, and foreign terminology were perceived as being stronger charged with magical quality compared to local knowledge. Lead, known by the Batak as simbora (a word derived from Sanskrit tivra) was deemed to have strong magical properties. Amulets made from lead, called porsimboraon, where considered to have especially great power. A strip of lead was rolled up, and worn with a string around the neck. Such an amulet was considered even more powerful if a piece of paper was rolled up together with the lead on which some Arabic letters were written.

The common Muslim formula bismi-llāhi r-raḥmāni r-raḥīm (in the name of God, most gracious,

most merciful) was probably not understood by most Batak, but nevertheless it played, in its shortened form bismilah, an important role in pustaha where chapters were often introduced by bismilah—usually rendered as bisurmila.

The Batak scriptorial tradition, and the culture in general, was not only influence by Islam, but to a much stronger degree by concepts imported from India. A kind of protective magic (pagar) is known as Si Aji Baima (Lord Baima) who is nobody else but Bhīma, one of the central characters of the Sanskrit epic Mahabharata, famous for his great physique and brute strength. Auṃ, the sacred syllable in Buddhism and Hinduism, was also used by the Batak, usually rendered as ong, for the same purpose, and often in combination with bisurmila.

The  end  of  the  Batak  scriptorial  tradition  When the Dutch linguist Herman Neubronner van der Tuuk visited Sipirok in 1852, he observed

“a lack of pustaha, datu, and pigs” as a consequence of the invasion of Islamic warriors from neighbouring Minangkabau.2 Willer, the civiel gezaghebber (resident) of Mandailing, also writes that no pustaha are left in Pertibi (Angkola), whereas in Mandailing they have become scarce. Here, too, the lack or scarcity of pustaha is attributed to the Padri invasion in the 1830ies.3

In regards to the divine books of the Batak, the fanatic Wahhabis and the Protestant missionaries were brothers-in-spirit. A few years after the Dutch decisively defeated the Padri in 1838, the colonial officer Isaac Esser (1818-1885), became the Malay teacher of Gerrit van Asselt, a young missionary of H.W. Witteveen, one of the founders of the Dutch Free Evangelical Churches. Esser’s intent was that Asselt should become missionary in West Sumatra to “conquer the Minangkabau kingdom”, but upon arrival in Padang, Asselt was unable to get governmental approval as the Minangkabau region was already firmly Islamised. Instead, the governor advised Asselt to settle in Sipirok. In 1861, the small and not very successful mission was taken over by the Rhenish Missionary Society. The Batak bark books reminded the German Lutherans of the magical books of the Jews and the Greeks in Acts 19:17–20:

This became known to all who lived in Ephesus, both Jews and Greeks; fear came over them all, and the name of the Lord Jesus was praised. Many of those who had believed came forward, confessing and making their deeds known. Large numbers of those who had practiced magic collected their books and burned them up in the presence of everyone.

Missionary Meerwald, for instance, writes: “The books, written by the datu with such a great sacrifice in time and effort, served them, as we can read in Acts 19:19, as items for witchcraft, and were hence, as they were there, ripe to be burned.” (Meerwaldt, 1922, p. 295) 4

2 “Een algemeene schaarschte in poestahas, datoes en varkens, welke laatste bestanddeelen van de Battaksche gemeente voor den invloed van het Islamisme de vlugt hebben moeten nemen.” Letter dated 5-5-1852 published in Groeneboer 2002, p. 134. 3 “Wat Pertibie en Mandheling betreft, kan ik verzekeren, dat in eerstgenoemd landschap geen poestaha's hoegenaamd zijn te verkrijgen, en zij in het laatstgenoemde hoogst zeldzaam zijn te vinden. De Padries hebben deze boeken overal met veel godsdienstijver opgespoord en verbrand, de bezitters daarvoor hooge boeten afge-perst; zij gebruiken hier hetzelfde argument als Omar bij de Alexandrijnsche boekerij.” (Willer, 1846, p. 391) 4 “De boeken door de datoe's met zooveel opoffering van tijd en moeite geschreven, zij dienden hen, die, gelijk wij in Hand. 19 : 19 lezen, ijdele kunsten pleegden, en waren dus rijp, om, evenals daar, ten vure gedoemt te

From Sipirok the missionaries proceeded further north, first to Silindung, then to Toba, and finally to Dairi, Samosir and Simalungun. By the 1920ies a large percentage of Batak was Christianized and only a minority still practiced the religion of the ancestors, especially in Karo, Simalungun and Samosir. Especially in Mandailing the missionaries had to compete with Islam. After the forced conversion by the Padri warriors, many of the new converts abandoned Islam and reverted to the old religion, but others decided to stay Muslims. The Christian mission first made great progress and in the years 1867–1871 there were about 700 new converts, mainly in Sipirok and Angkola. Islam, however, progressed even faster and by the mid 1870ies there were only a few people left who still adhered to the old religion. The overwhelming majority of the population had become Muslims with a sizeable minority of Christians.

Given the particular situation of Mandailing—the invasion of the Padri warriors in the 1830ies and the arrival of Christian missionaries in the 1850ies—it is most likely that the majority of Mandailing pustaha that until now are preserved in mainly European collections, predate the Padri invasion, with the rest being produced between 1840 and 1875. It seems fairly unlikely that any pustaha were written after 1875.

The Batak script, however, continued to be used. Influenced by Herder's focus upon language and cultural traditions as the ties that create a nation, the German missionaries were keen to preserve Batak language and its writing system as an integral part of Batak Volkstum (national character).

The first typefaces for the Batak script were commissioned by the Dutch Bible Society and used by the printing company Spin & Zoon in Amsterdam for Van der Tuuk’s translation of the book of genesis, which was published in 1853 by Frederik Muller. Later, the Rhenish Mission also created its own set of typefaces which were used by three different publishing houses, namely the national printing house in Batavia (Landsdrukkerij), the mission’s printing house in Elberfeld, and the smaller local printing house (Zendelingsdrukkerij/Pangarongkoman Mission) in Laguboti. After World War I the mission decided not to publish any books in the Batak script anymore, and the last book published in Batak script was Arsenius Lumbantobing’s school reader Porgolatanta: Buku sidjahaon ni anak sikola published 1916 in Balige. The mission cited financial reasons for its decision. After the war the financial situation of the Rhenish Mission was very tight, and it was considerably cheaper to publish books in Roman script. Yet, the main reason for ceasing to use the Batak script was simply that the Batak themselves preferred the Roman over their own indigenous script.

The  Batak  script    As already mentioned above, the Batak refer to their script as surat. In his “grammar and

dictionary of the Malay language”, John Crawfurd already stated that the word surat is an Arabic loanword (Crawfurd 1852:cxcvi), and the same view is still reiterated in recent publications (e.g. Siti Hawa Hj. Salleh 2010:5) despite the fact that since Dempwolff’s studies (1934-1938) surat has been firmly established as an Austronesian word, reconstructed as Proto-Malayo-Polynesian *surat ‘to scratch, draw’. The word surat or sulat is also used for all other writing systems of Sumatra, Sulawesi, and the Philippines. In central Sumatra the Kerinci refer to their script as surat incung (“cursive script”), the various peoples in the modern province Bengkulu and Sumatra Selatan refer to their scripts as surat ulu (“the script of the highlands”), and the people of Lampung also refer to their script as surat had, or surat Lampung. In the Philippines, the Bisayans and Tagalog call their script sulat while Tagbanwa and Mangyan call it surat. The Bugis and Makassar people of Sulawesi also call their

worden.”

script surat, which is usually rendered locally as sureq. The common nomenclature is of course not coincidental, but convincing evidence of a common origin of all these scripts.

Attempts have been made to pinpoint the origins of the Philippine scripts to the Cham script (Wade, 1993), while Miller (forthcoming) has suggested a Gujarati origin for the scripts of Sumatra, Sulawesi, and the Philippines and thus dismissing the earlier theory proposed by de Casparis (1975) and others that the insular Southeast Asian scripts are ultimately derived from the Pallava script. The proposed Gujarati origin would also place the surat scripts in the Nagari group of Indian scripts, contrary to the long-held assumption that they belong to the Pallava group of which the scripts of the Bali, Java, and Sunda are derived. While this paper is not the place to address in detail the arguments of Wade and Miller, my findings are supportive of the traditional view.

The history of writing in Indonesia goes back to the early fifth century. The oldest inscriptions of the archipelago are the Kutai inscription of Raja Mulavarman in Borneo (322 Saka or 400 CE) and the Raja Purnavarman inscription (450 CE) from Ci Aruten, West Java. Both inscriptions are in the South Indian Pallava script and the language is Sanskrit. Inscriptions from the seventh century CE are still in Pallava script but the language used is now Old Malay. After the seventh century the Pallava script gradually changed into a more distinct Indonesian script, which developed in Sumatra, Java, and in Bali. Initially this so called Kawi script was still relatively similar to its progenitor but after roughly five centuries, during the heydays of the Majapahit kingdom, the script had not only changed considerably from its progenitor but had also developed distinct local varieties. By the sixteenth century Kawi had developed into the modern Javanese and Balinese hanacaraka scripts, but it is not entirely clear whether the other scripts of the archipelago, the scripts of Sumatra, Sulawesi, and the Philippines, also developed at that time, and whether they developed straight out of the Kawi script.

In summary, the pre-Arabic writing systems of the archipelago can be divided into two groups. 1. The Hanacaraka Scripts of Bali, Java, and Sunda These three very closely related scripts are descendants of the Kawi script, which, in turn, is

derived from the Indian Pallava script (Casparis 1975). 2. The Surat Scripts of Sumatra, Sulawesi, and the Philippines As will be demonstrated later, the scripts of Sumatra, Sulawesi, and the Philippines clearly

constitute one single related group that can be divided into four subgroups. a. The Ulu Scripts of Central-South Sumatra The scripts of Besemah, Enim (around Muara Enim), Kayu Agung, Lahat, Lembak, Lintang,

Lebong, Ogan (along the Ogan and Komering river), Pagar Alam, Rambutan (Banyuasin), Rejang, and Serawai are commonly called surat ulu—the script of the upstream people. In fact, these districts are all located in the Bukit Barisan mountain range where the inhabitants lived in relative isolation just like the Batak several hundred kilometers to the North. The script of Lampung is not typically called ulu script, but as it belongs to this group it will in this paper be treated as belonging to the ulu scripts. In the literature the ulu scripts have also been named rencong. This term, which is used in only one small district, became popular through the work of the Dutch administrator Oscar Louis Helfrich (1904). The British anthropologist Mervyn A. Jaspan, who conducted fieldwork in Rejang in the 1960ies, named the scripts of that region, according to the first three letters of their alphabet, the Ka-Ga-Nga script (Jaspan, 1964). Apparently Jaspan was not aware that in the ulu script consonants are, by and large, ordered according to the phonetic distinctions made in Brahmic scripts proceeding from velar, palatal, dental, labial, sonorant, sibilant to guttural consonants. Ka, Ga, and Nga are thus the first three letters not only of the ulu alphabets but of dozen other alphabets as well.

The indigenous term surat ulu (upstream letters) is little known in Western academic discourse, but is commonly used in Indonesia, especially by Sumatrans themselves.

The scripts of Kerinci and Lampung, although are not usually termed surat ulu, belong to the same group.

b. The Batak Script of North Sumatra This script consists of five subgroups, namely Angkola-Mandailing, Toba, Simalungun, Pakpak-

Dairi, and Karo c. The Bugis-Makassar Script of Sulawesi Like the scripts of Sumatra and the Philippines, the Sulawesi script is also called surat, or, more

precisely, sureq as surat is rendered there. Only a few manuscripts written in the Old Makassar script still exist as this script became extinct in the nineteenth century. This script was then replaced by the Bugis-Makassar script, which consists of the two identical scripts of Bugis and Makassar where the only difference is that Bugis has four additional characters. The Bugis-Makassar script has also been used on the islands of Sumbawa and Flores.

d. The Philippine Scripts The four closely related scripts of the Philippines are called Sulat Bisaya, Sulat Tagalog, Surat

Tagbanwa, and Surat Mangyan. The above-mentioned scripts all belong to the wider Brahmic family of scripts. Before the coming

of Islam that introduced the Arabic-Persian script into parts of the archipelago, and before the arrival of Europeans who introduced the Roman script, nearly all Southeast Asian countries wrote in scripts that were derived from the Indian subcontinent. These scripts have been called “alphabet” and “syllabary” but also a “semisyllabary” (Diringer, 1968), an “abugida” (Daniel, 1996), or an “alphasyllabary” (Bright, 1996). The last three terms basically all refer to the same thing. In alphasyllabic systems, this is the term that I prefer, “syllables are recognized as units but are represented by symbols that acknowledge an awareness of underlying consonants and vowels.” (Bringhurst, 2002, p. 19)

Alphasyllabic systems are similar to alphabets in that they distinguish two types of symbols: consonants and vowels, but they differ in that the vowels are represented by a diacritic when in combination with a consonant, and by an independent grapheme when they are not in combination with a consonant.

The features of the scripts of the Brahmic family, which are structurally all very similar, are described below. Special reference is made to the modern Javanese script, but what is said in respect of the modern Javanese script is also valid for its precursor Kawi.

1. The graphic syllable, called akṣara in Sanskrit, but also in Bali, Java, and Sunda. The akṣara represents either an independent vowel, or a consonant. The unmarked consonantal grapheme is

understood to have an automatic or inherent vowel, which in usually a short a. 2. The consonantal akṣara can take a dependent vowel called mātrā which are represented by

a diacritic placed, depending on the script, to the left of, to the right of, above, below, or on

Table 2 Vocalic diacritics in Javanese

both the left and the right sides of the akṣara. The mātrā deprives the akṣara from its inherent vowel. Nasalization is noted by a special diacritic called anusvāra, and aspiration by visarga. In the modern Javanese script there are five vowel markers known as sandhangan swara, each of which has a proper name: wulu [i], pepet [ə], suku [u], taling [e], and taling-tarung [o].

3. Consonants can be combined in ligatures. Special marks are added to denote the combination of ‘r’ with another consonant. In Java, each consonant has two forms: the aksara form is used at the beginning of a syllable, while the pasangan form, which usually appears below the aksara form, is used for the second consonant of a consonant cluster and mutes the inherent vowel of the aksara.

4. The inherent vowel of the aksara can also be suppressed by use of the virāma to form a “dead” consonant at the end of a syllable.

5. The traditional ordering can be summarized as follows: vowels, velar consonants, palatal consonants, retroflex consonants, dental consonants, bilabial consonants, approximants, sibilants, and other consonants. Each consonant grouping had four consonants (with all four possible values of voicing and aspiration), and a nasalized consonant. Modern Javanese has abolished the traditional order. The order of the consonants makes the following saying, “Hana caraka, data sawala padha jayanya, maga bathanga” which means “There were (two) emissaries, they began to fight, their valor was equal, they both fell dead”.

6. Most Indian scripts, and also the Javanese script, come with their own set of symbols expressing numerals. The Hindu-Arabic numeral system with numbers from 0 to 9 is the predecessor of our contemporary “Arabic” numeral system. Compared to the Indian but also to the Javanese script (and its Bali and Sunda cousins), the scripts of Sumatra (the five Batak scripts, the Kerinci script, the Lampung script, and the numerous Ulu scripts), Sulawesi (Bugis and Makassar) and the Philippines (Tagalog, Hanunóo, Buhid, and Tagbanwa) do show a number of similarities, that clearly place them in the Brahmic family of scripts, but they also show a number of differences that clearly set them aside from the Indian, but also the Javanese script, or the similar scripts of mainland Southeast Asia (as the Cham script, for instance), and confirm that the scripts of Sumatra, Sulawesi and the Philippines are not only closely related to each other, but form a subgroup of their own. The similarities as well as the differences will be described below using the same set of six essential features as described above.

1. The word aksara is unknown in the scripts of Sumatra, Sulawesi, and the Philippines. The name aksara has been replaced by the word surat, which in some regions is rendered as sulat or sureq. The Batak, for instance, call their script surat sampulu sia ‘the nineteen letters’, the people of Kerinci call their script ‘surat incung’ (cursive script). Often the epithet Java is added (surat incung Jawa), probably denoting a foreign, although not necessary Javanese origin. The slightly different scripts of southern central Sumatra are called surat ulu or ‘upstream script’ denoting that it is used by the highland people of the Bukit Barisan mountain range. The Lampung script is called surat had. The script of the Hanunóo of Mindoro is called sulat Mangyan. The Bugis call their script, depending on the orthography, sura, surak, or sureq, etc. The word surat is an Austronesian word, and known in many languages. The

Batak dictionary gives: surat, script, letter; manurat, to write; and in an Tagalog dictionary we find sulat, letter; sumulat, to write. Because of the common terminology I will further on refer to the scripts of Sumatra, Sulawesi and the Philippines as surat scripts, and refer to the other Indian and Indian derived scripts as aksara scripts. As the term surat can also refer to the diacritics mentioned under point 2), the aksara are called by the Bugis ina surak or indok surak and by the Batak ina ni surat (literally “mother letters”) whereas the diacritics are called anak surat (literally “the children of the letters”). The same terminology is also used by the Antemoro people of Madagascar, who, however, use an Arabic derived writing system which they call sora-be, derived from soratra (script) and be (large). The main letters are called reni soratri and the diacritics zana ‘tsoratri (Dahl 1983:71 fn.7). The Antemoro hence use exactly the same metaphor that is also used by the Batak and Bugis people. It is therefore evident that the Sorabe script of the Antemoro derived either from the Malay Jawi or the closely related Javanese Pegon script, both adaptations of the Arabic script. As the Sorabe script dates back to at least the fourteenth century, one can conclude that either the Pegon or the Jawi script (or both) by the thirteenth or fourteenth century still used the mother-child metaphor for character and diacritic, and that this metaphor was carried over from a pre-Arabic writing system. (Adelaar 2005:4) The alphasyllabries of the Brahmic script family are strictly phonetic. The more complex a language is phonetically, the more letter symbols it has. The Devanagari script, for instance has 33 consonants in addition to 7 “independent” vowels, i.e. vowels are used to write a syllable without a consonant. When the Pallava script gradually transformed into the Kawi script it could have reduced the number of letter symbols considerably as Old Malay is phonetically relatively simple. But this was not done as Sanskrit continued to be an important language associated with the court, and of course also with religion. The modern Javanese script dramatically reduced the number of aksara, and the surat scripts generally have no more than 20 aksara (consonants and independent vowels), and between 2 and 10 diacritics. The Philippine scripts have 14-16 aksara (consonants and independent vowels) and 2 diacritics. Batak and Bugis-Makassar have 19 aksara (plus a handful of prenasalized aksara), and between 7 and 10 diacritics. The Ulu scripts have between 20-24 aksara and up to 13 diacritics. The small amount of aksara and diacritics among the surat scripts is sufficient as the languages are phonetically relative simple. All these languages have disyllabic roots, simple basic word patterns, and relatively few phonemes. Toba Batak, for instance, has only 21 phonemes, and the basic word patterns are CVC, CVCCVC, or CVCVC. Pre-colonial Tagalog has 20 phonemes (15 consonants, the vowels a, i, u, and the diphthongs ai, and au) and a very simple basic word pattern of V, VC, CV, VCV. An innovation of the surat scripts are nasalized aksara. These are absent, and also not needed in the aksara scripts where consonants are combined in ligatures to form consonant clusters (see point 3). Austronesian languages are phonetically relatively simple and tend to have few, or even no consonant clusters. Malay—the language that is used for writing the Kerinci and all Ulu scripts—only allows word-medial nasal-stop clusters like mp, nt, ngk, and r plus another consonant, such as in pernah (ever). Like Malay, Bugis also only allows four nasal-consonant clusters (Abas & Grimes, 1994, p. 559) and “the lontara [another name for the Bugis script, U.K.] is principled around open syllable structure and is thus inadequate for indicating final

/ʔ/ or /ŋ/, or for mapping germinate consonants.” (op. cit.). Originally Tagalog also did not have initial or final consonant clusters, and the examples given by (Ramos, 1971, p. 8) indicate that the medial clusters are predominantly (if not all) borrowings from Spanish. Due to the high frequency of word-medial nasal-stop clusters in Malay, special prenasalized aksara were developed such as mpa, nta, ngga, ngka, etc., which tend to closely resemble the aksara they are based upon. Rejang ga gfor instance, becomes ngga G, and ja j becomes nja J. Prenasalized aksara are unknown among the Batak scripts, except for Karo, which has a number of variants for mba (f,B,v) which are evidently based on ba (b), and also for nda, for which there are four variants (q,[, {, }) that are all based on da (d). Bugis has mpa (P) based on pa (p), and also ngka, mpa, nra, and nyca. The Philippine scripts do not have prenasalized aksara due to the lack of nasal-consonant clusters in Philippine languages. As it is also the case with the aksara scripts, the surat scripts have among the aksara independent vowels and consonants. Whereas aksara scripts tend to have a full set of independent vowels, usually between seven and more than 15 (when long vowels are included), surat scripts have either a reduced set of independent vowels, or only a (Surat ulu and Bugis/Makassar). Batak has a (a), i (I), and u (U) only. Both I and U are optional, and in fact relatively infrequently used. Surat ulu has only one independent vowel, a. Tagalog has a, i, and u (where i and u are also used for e and o). Tagbanwa and Mangyan also have a, i, and u, but the i is nothing but the letter a with diacritic i, and this is precisely how independent i is formed in the Ulu, and also in the Batak script. Instead of I (i), and U (u), one can use ai and ao as in the words aid (ida), and AL (ulu). Other independent (not consonant-bound) vowels are likewise formed with the letter a plus the vocalic diacritic: aed (eda), aolo (olo).

2. Vowels other than the inherent one are indicated by the addition of an extra marker (diacritic): fi (ti) fe (te) F (tu) fo (to). The two other diacritics add a ng or h sound to the syllable: t^ (tang), th (tah). The surat scripts, but also the scripts of Bali, Java, and Sunda, do not make any distinction between vocalic mātrā and the nasal- and aspiration markers anusvāra, and visarga. In Indian languages, the exact pronunciation of the anusvāra can vary greatly depending on the location of the anusvāra in the word and the language within which it is used. In all Indonesian and Philippine scripts the anusvāra as undergone a transformation from a general nasal-marker to a syllable-final /ŋ/-marker. It can be added to any aksara, and also to any aksara + vocalic diacritic. For Batak an example would be b (ba) b^ (bang), and bi (bi) bi^ (bing). For Rejang it is k (ka) k^ (kang), bi (ki) ki^ (king). The Philippine and Sulawesi scripts do not have a diacritic for syllable-final ng despite the fact that Philippine languages have syllable-final /ŋ/. Other syllable-final nasals are created by a combination of two aksara plus virama, e.g. Batak bn\ (ban), but some Ulu Scripts actually do have a special diacritic to represent syllable-final n: which looks exactly like a double ng: k^. As there was no distinction anymore between vocalic diacritic on the one, and nasal- and aspiration markers on the other hand, they were subsumed under one single category of anak surat (“the children of the letter”)—a term that is testified for the Batak and the Bugis/Makassar script, and incidentally also for the Ulu, but not for the Philippine scripts. An innovation of the surat scripts—but also the scripts of Bali, Java, and Sunda—is the introduction of proper names for all diacritics. These proper names will be discussed further

down. 3. While consonant clusters are relative rare or absent in the languages of the surat scripts, they

are very common in Javanese and Sundanese (Hardjadibrata, 1978, p. 165). This may be one of the reasons why the scripts of Bali, Java, and Sunda have maintained ligatures, called pasangan. The surat scripts have taken a fairly radical, but consequent step in an effort to make the script more efficient, and easy to learn, by eradicating the complex consonant clusters involving a second set of subscript characters. Instead, the surat scripts either used additional diacritics for syllable-final r, k, and n (these were employed in the ulu scripts only), or the virama.

4. Consonant cluster can, even in the Indian scripts, be resolved by using the virama to mute the inherent a vowel of the aksara. Rather than using ligatures one can simply write a consonant cluster such as tr by writing the aksara ta and ra and muting each using the virama. In Indian scripts this would be perfectly understood, but not usually done as the convention demands the use of ligatures. In the languages written with the surat -scripts, such a case would not even occur as consonant clusters are limited to mid-word clusters as in the Malay words pancing, kambing, bantah, pernah, pingsan. Sanskrit loanwords with triple consonants such as sastra and istri are usually written (and also often pronounced) by adding an additional Schwa to break up the consonant clusters: sastera, isteri.

5. The Sanskrit alphabetical order has been largely preserved in the ulu scripts and also in the Bugis and Makassar scripts (Matthes, 1858, p. 2) while it underwent profound changes in the Philippine and the Batak scripts. The Batak have at least four or five competing alphabetical orders. In Mandailing alone there are two. The one that is used in the south has the order a ha na ma ta ra ja ga la pa sa da nya ba wa nga ya i u ka ca, whereas the one in Angkola is the same that is also used in Toba: a ha ma na ra ta sa pa la ga ja da nga ba wa ya nya i u.5

6. Numerals are unknown in any of the surat scripts. This is understandable as the scripts were never used for administrative, legal, or economic purposes.

Another feature that distinguishes the surat scripts from the aksara scripts lays in the relative

simplicity of their shapes. The ulu scripts are occasionally also called rencong or incung, i.e. cursive script, which points to the fact that virtually all characters are shaped around a slash.

Most characters are extremely simple and consist between one to four straight strokes, and there is not a single curved line:

Table 3 Selection of simple Ulu characters

ba pa ga ha la wa ta ra na ma b p g h l w t r n m

But even the most complex ones involve no more than five or six straight strokes:

Table 4 More complex Ulu characters

nga nda nja

5 In a paper manuscript from the Van der Tuuk collection in the Leiden University Library (Cod. Or. 3421, p. 83) the position of the last three letters is i, u, nya.

< D J In the Bugis and Makassar scripts the basic pattern is again a stroke in form of a slash (ka), but

more common is the wedge, shaped similar to the Roman letter v as in the letter ma. It usually occurs inverted as in the letter ta, it can be reduplicated, either vertically (ra), horizontally (wa), or both (sa), quadrupled (ha), extended by adding an upstroke parallel to the downstroke (pa)—ca and nga can be considered variants of this shape, it can be distinguished by a dot (na, da, a, ya, ga), and la may be considered a variant of wa with an interrupted line.

Table 5 Progressing complexity of the Bugis-Makassar script

ka ma ta ra wa sa ha pa ca nga na da a ya ga la ba k m t r w s h p c G n d a y g l b

Already in 1858 Matthes observed: “I cannot discern any similarity with the Devanagari or

Sanskrit alphabet except that a) all letters have an inherent a-sound, b) the alphabetic order is almost the same, and c) that the order follows the same linguistic principle: first the gutturals ka, ga, nga, ngka, then the labials pa, ba, ma, mpa, then the dentals ta, da, na, nra, then the palatals ca, ja, nya, nyca […].” (Matthes, 1858, p. 2)

The strikingly uniform appearance of the Ulu, and especially the Bugis-Makassar script, suggests that at some point in time they underwent an intentional process of simplification and standardization which changed their shapes to such an extend that it is now extremely difficult, especially in respect to the Bugis-Makassar script, to detect any similarity to other scripts of the region. The Philippine and the Batak scripts are much less uniform, but here, too, we can discern a number of basic patterns. One can assume that the proto-script—is seems likely that all surat scripts evolved out of one single proto-script—was simplified once it had left the narrow court environment with which is previously was associated, and became a commonly used tool, especially in courtship rituals (Kozok, 2000). It then became a necessity for the script to be a tool as simple as possible to facilitate its use not only by a handful of specialists, but by commoners.

Among the related surat scripts of the archipelago, Kerinci surat incung and Bengkulu surat ulu are closest akin to the Batak script. The following letters demonstrate this convincingly:

Table 6 Affinities between Batak, Ulu and Incung scripts

Batak Ulu Incung A/Ha a h a

Ka k h k k

Ga g g g

Da d d D

On the other hand, there are also clear similarities between Batak, Incung, Ulu, and Philippine

scripts:

Table 7 Similarities between Batak, Incung, Ulu, Tagalog, and Tagbanwa

Batak Ulu Incung Tagalog Tagb Ba b b b

Pa p p p Ha a h h

The structural similarities, especially if contrasted with the Indian scripts, clearly show that the

surat scripts are all related to each other and form a single subgroup of the Pallava-derived scripts of the archipelago.

Although it is difficult to trace the exact relationship between the three subgroups of surat scripts, it is possible to do that within the Batak subgroup. We will start our analysis with the aksara (ina ni surat).

⌘

Aksara  (Ina  ni  Surat)  Palaeography is the study of ancient writing including the practice of deciphering, reading, and

dating historical manuscripts. By tracking the changes that a particular script has undergone it is often possible to determine the origin of the script. According to the studies conducted by Herman N. van der Tuuk, the Batak script developed from south to north (Tuuk, 1971, p. 77). Parkin (1978, p. 100) also concurs that the Batak script originated in Mandailing. These two scholars give the following reasons:

The aksara [ (nya), w (wa) and y (ya) represent three phonemes that do exist in Mandailing but not in Toba Batak. Among all Batak languages Toba Batak is phonologically the least complex language, which, at some point in time, has undergone phonetic changes that sets it apart from its neighbouring dialects. The difference between the Mandailing and the Toba language, lays to a great extent, apart from some different lexical items, in their different sound systems.

Mandailing /ny/ becomes /n/ in Toba Batak, while /w/ and /y/ become Ø: M manyurat T manurat; M sayur, T saur; M tawar T taoar. Although in Toba these three aksara are clearly redundant, they are included in the alphabet. The aksara nya is never used in Toba, while wa and ya are often used to link vocalic clusters as, for instance, in dua (two) which may be written either Dw or Da, or in sia which may be written either siy or sia. The inclusion of nya, and to a lesser extend also of wa and ya, in the Toba alphabet, proves that the Batak script developed in Mandailing and was later brought to Toba. Had the Batak alphabet been created by the Toba Batak, and then later brought to the Mandailing people, there would have been no nya in the Toba Batak alphabet.

This example alone would not be sufficient to fully proof the theory of a south-north development of the Batak script, and there is more evidence that will be provided further down.

In the districts north of Toba, namely Pakpak and Karo there is also no phoneme to be represented

by the aksara nya, but in the Karo script the aksara [ has been preserved, but the value has changed from nya to ca. It still occupies the same position in the alphabetic order (between la and i), which reinforces the theory of a south-north development.

Another factor that points to Mandailing as the point of origin of the Batak script, is the large diversity of characters in Mandailing. The second highest diversity is in Karo, where we have variants for the aksara sa, da, and ca, and especially for the relative new (and always optional) aksara mba (B, v, f) and nda ({, }).

In Toba and Mandailing the aksara ma can have three slightly different shapes: 7, 8, and m. The latter is hardly ever found in Toba manuscripts and is usually associated with Mandailing manuscripts. The other regions, Pakpak, Karo and Simalungun, only have one single representation of the aksara ma. The greater diversity in Mandailing is not a proof, but another piece of evidence that points to the greater age of the Mandailing as compared to the Toba script. Stronger pieces of evidence for our theory of a south north development of the Batak script are provided by the aksara na and ja.

Of the four Batak shapes for the letter na, shows the greatest affinity to the Pallava form as it was written in India during the first half of the first millennium, and also to fourteenth century Kawi (the examples are taken from Col. 2. stone inscriptions from East Java, Col. 3. inscription in Amoghapasa statue from Dharmasraya, West Sumatra; and Col. 4 the Tanjung Tanah manuscript also from Dharmasraya). The na variant seems to form an intermediate stage to the more modern versions n and n.

Table 8: Historical Development of the letter na

Pallava ±6-12th c

Majapahit 14th c

Amoghapasa 14th c

Tj. Tanah 14th c

Bali ±17-19th

Java ±17-19th c

Batak ±17-19th c

n n

The letter na has shown very little development in about one thousand years, but not all letters are

equally conservative and can be traced back with so much ease. Tracing back the origins of the Batak script is difficult because the earliest known Batak manuscript dates back to only the second half of the 17th century. Stone inscriptions in any of the surat scripts are rare, and the few stone inscriptions from the Batak land all use Batak letters that resemble those on manuscripts. The case is different with the Javanese script where there has been a continuous documented tradition for over one thousand years with manuscripts written on daluang bark paper, on lontar palm leaves, as well as inscriptions on copper plates and stone.

As we have shown above, the Mandailing variant q of the aksara na is so similar to the na of the Kawi script that Voorhoeve has called variant q the “ancient na”. This variant is not entirely limited to Mandailing, but its occurrence in Toba is rare, and in Mandailing it is common. North of Toba it does not occur at all. The similarity between the two aksara shall not be seen as evidence that the Batak script is descendent from the Kawi script, but rather that either both scripts share the same historical root, or that, at some point in the development of the Batak script, it may have been influenced by the Kawi script. Voorhoeve is in favor of the latter interpretation. Another strong piece of evidence supporting our theory is provided by the aksara ja.

The shape of the letter ja (j) is precisely identical to the letter da (d) except that the former has an additional horizontal stroke. Looking at Holle’s “Tabel van oud- en nieuw-indische alphabetten” (Holle, 1882) it seems to be obvious that the shapes of the letters da and ja have been inspired by the letters da and ja as it appears in the stone inscriptions and copper plates found in Java and Sunda from the ninth to the sixteenth century. Here, too, the essential only difference between the two letters is the addition of a horizontal line. The same feature is also discernible on the 762S (840 CE) copper plate from Java (Holle, 1882 p.4), the 900 CE Laguna Copperplate Inscription (LCI) from the Philippines (Postma, 1992), the 1347 CE inscription from Padang Candi (PC), Dharmasraya (Sumatra) (Kern, 1907), and on the fourteenth century Tanjung Tanah manuscript (TT), also from Dharmasraya (Kozok, 2006).

Table 8 Development of the Letters da and ja

Java 840 CE

Philippines (900 CE)

Sumatra (TT) 14th c.

Sumatra (PC) 14th c.

Batak ±17-19th c

da

d

ja

j

The aksara j was later simplified to j. Both forms are common in Mandailing and Toba, the

latter form is the only form attested for Karo Batak. This, again, is a piece of evidence testifying a south-north development of the Batak script.

The two southern scripts Toba and Mandailing, are very similar. Mandailing Batak has variants of the aksara sa and ha/ka which are unknown in Toba whereas in Toba new variants for wa (v) and ta (f) were developed—a word of caution seems to be necessary here: one cannot exclude the possibility that the variants f and v are the older forms that in Mandailing and in parts of Toba became t and w.

The variant v is the dominant form of wa in Pakpak-Dairi, and a common variant in Toba whereas in Karo only w is known. f is the only ta form in Karo and Pakpak-Dairi. Simalungun uses t and w.

In his Toba Batak grammar, Van der Tuuk mentions that the ta variant t and the wa variant w are used in “East Toba” and the variants f and v in “West Toba”. In a letter dated 5 May 1852 he is more specific and explains that f is used “in the interior of Barus and Siboga” and also “in the vicinity of Sipirok.”

I was able to ascertain that the Toba dialect is also spoken in the vicinity of Sipirok. In a letter addressed to Mr. Hamers I found the alphabet of that district [Toba, UK]; among others I found that the letter t was represented by t instead of t (the t is found in the script as it is used in the interior of Barus and Siboga, and the t is used in Batang Toru, Angkola, and Mandailing). (in Groeneboer 2002:135)

Table 9 Comparison of Mandailing and Toba characters

Ta Wa Sa Ha/Ka

Mandailing t w s h f

Toba f t w v s h The shape of the Mandailing letter ha, which has also the value /ka/, is very similar to Toba ha, but

it appears to be the older form of the letter ka, which can be traced back to early Kawi and remained very stable until later Kawi, and this is also the form of the letter ka used in all Ulu scripts including Lampung but excluding Kerinci.

Figure 3 Pustaha from Mandailing (Museum für Völkerkunde. Berlin 36841)

Table 10 Historical Development of the Letter Ka

Pallava Java 840 CE

Philippines 900 CE

Sumatra (TT) 14th c.

Sumatra (PC) 14th c.

Java 16th c.

Ulu ±17-19th c

Batak ±17-19th c

k h k f

We believe that the above evidence is strong enough to convincingly identify Mandailing as the

region of origin from where the Batak script developed. More evidence will be provided soon when we discuss the diacritics where we will also demonstrate that from Mandailing the Batak script most likely spread first to the border region of Toba and Simalungun. On the first sight, the Simalungun script may seem quite different from the other Batak scripts because of its unique feature of disconnected lines. When we disregard this peculiarity, we can discern a number of similarities between the Simalungun and the Mandailing script. The aksara sa has two variants in Simalungun, namely s and 0, the latter is identical to Mandailing s, and the former can probably simply be seen as a variant with a disconnected line so characteristic for Simalungun. The Simalungun aksara ha k resembles closely Mandailing h and w. The fact that Mandailing and Simalungun share these similarities among each other, but not with Toba, points to the possibility that the Batak script entered Simalungun at a very early stage. The curved horizontal line of the Simalungun aksara ya (y) may also be contributed to Minangkabau influence.

The  Diacritics  (Anak  ni  Surat)  Each diacritic, or anak ni surat (minor letter), has its own name. This is a feature the Batak scripts

share with all other Indonesian scripts of Indian origin (see Table 12). In India, there are certain diacritics that have proper names such as the anusvāra, used to mark a type of nasalization, and the virāma that is used to suppress the inherent vowel that otherwise occurs with every consonant letters, but the diacritics used to express dependent vowels do not have any special name.

The names of the diacritics in the Batak script differ from region to region. A comparison can provide us clues about the historical development of the Batak script. As we will see, the diacritics too confirm our theory of a south to north evolution of the Batak script, and they also strengthen the theory that the Batak script is closely linked to the Ulu scripts, and that is shares the same ancestor with the modern Hanacaraka scripts of Bali, Java, and Sunda.

Table 11 Proper names of the diacritics in Indonesian scripts

e ě i o u ng h virama

Karo kětelengěn

kěběrětěn

kělawan

kětolongěn sikurun kěbincarěn kějěringěn pěněngěn

Pakpak kětadingin kaběrětěn podi

kaloan sikora kaběrětěn kěbincarěn sikorjan pangolat?

Simalungun hatalingan haluan sihorlu haboritan haminsaran hajoringan panongonan

Toba hatadingan haluain

hauluan

haulian

siulu

uluwa

siala

sihora

hasialaan

haborotan

haboruan

haminsaran

hamisaraan

paminggil

pangolat

Mandailing talinga uluwa siala ulu boruta

buruta

amisara pangolat

Rejang kamica - kaluan - kamitan katulang

Lampung ulan bicek

bicak

ulan bitan bitan tekelubang

ketelobang

keleniah

kaliniah

nengen

tanda mati

Jawa taling pěpět wulu taling-tarung

suku cěcak wignyan pangkon

pangku

Bali taling pěpět ulu taling-tedung

suku cěcěk bisah adeg-adeg

Sunda paneleng paměpět panghulu panolong panyuku panyecek pangwisad paten pamaeh

Sanskrit anusvara visarga virama, halanta

e

BE (Java taling, Lampung keteliling, Rejang katiling) The Mandailing name of this diacritic is talinga – showing strong resemblance to the name of this

diacritic in Java, Rejang, and Lampung. Simalungun adds the suffix -an to the name. It is interesting to note that the names of the diacritics

in Karo and Pakpak almost all bear affixes whereas affixation is absent in Mandailing. Apparently this feature was added once the script had crossed the northern border of Mandailing. Two types of affixation are used: the prefix ha- and the circumflex ha-…-an, which in the northern regions regularly becomes kě-…-ěn.

The Mandailing talinga hence became hatalingan in Simalungun. It will remain uncertain whether the base word of hatalingan is taling, or, by means of contraction, talinga. Karo kětělengěn clearly has těleng and not tělenga as its root because otherwise the form should have been kětělengan with a long a, but this is of minor importance as Karo kětělengěn is clearly not directly derived from Mandailing talinga but via Simalungun hatalingan.

In Toba hatalingan becomes hatadingan. What has happened here is not difficult to guess. The word taling does not mean anything, and has hence been reinterpreted by means of tading ‘to stay behind’. The circumflex ha-…-an can form the adversative and hatadingan hence has the meaning ‘left behind’, which most likely refers to the position of the diacritic in relation to the aksara it modifies. The hatadingan is placed to the upper left, and is hence ‘left behind’ when one progresses in

the direction of writing, which is from left to right. Although this may seem somewhat speculative, this is not the only case where a diacritic was renamed in order to give the name a meaning. The diacritic ng, which will be discussed further down, has undergone a similar transformation. It seems most likely that the term hatadingan is not based directly on Mandailing talinga, but entered Toba via Simalungun hatalingan. This may seem illogical, as Mandailing and Simalungun do not share a common border but are separated by eastern Toba. This is the part of Toba in which the southern ta is used. Unfortunately we do not know for certain the extent of this “eastern Toba” but it certainly includes Uluan, the district south of the Simalungun border along the shore of Lake Toba. South of Uluan at the location of the town Porsea is the outflow of Lake Toba. Further to the south is the fertile district Toba (i.e. Toba proper, the densely populated area around Balige and Laguboti), and south of Toba are the large mountainous and very sparsely populated districts of Habinsaran and Pangaribuan. Fortunately we know from local sources (Harahap, 1960, p. 148; Tampubolon, 2002, p. 296) what we already expected, namely that besides hatadingan the term hatalingan also existed in Toba. That both van der Tuuk as well as the German missionaries were unaware of the term hatalingan is not surprising as Van der Tuuk never visited eastern Toba, and the missionaries too had their base of operation in southwestern Toba, namely in Silindung. We can hence assume that hatalingan was the name of the diacritic in eastern Toba where the southern ta (t) and the southern wa (w) were used, whereas the term hatadingan was used in western Toba where the northern ta (t) and the northern wa (w) were predominant.

The name of the diacritic e hence developed as follows (M=Mandailing, ET=Eastern Toba, S=Simalungun, WT=Western Toba, K=Karo, P=Pakpak-Dairi):

M Talinga | ET Hatalingan | S Hatalingan WT Hatadingan | | K Kětelengan P Ketadingin

o, u

Bo This x-shaped diacritic has the value o, except in Karo where is becomes u. In Mandailing its

name is siala ulu. Siala does not have any apparent meaning, but ulu is ‘head’, which may related to the position of the diacritic, which ‘heads’ the aksara it modifies. This may seem far-fetched, but the other diacritic that occupies the same position in relation to the aksara, namely the diacritic i, has a similar name, ulua.

In Toba siala ulu is either shortened to siala or affixation occurs resulting in hasialaan. Besides, there is also a third name, sihora, which does not have any particular meaning, but the o in sihora reflects the value of the diacritic. In Pakpak it has precisely the same name (if written), but the pronunciation is sikora as the aksara h which in Toba is ha, becomes ka in Pakpak-Dairi. In Karo this diacritic does not have the value o, but u instead, which is also reflected in the name: sikurun which still resembles sihora, as does the Simalungun name sihorlu, but why Karo and Simalungun

underwent these changes may remain a mystery. If we have two names for the same diacritic in Toba Batak, then one can assume that these two

names were used in two different places. One is hence tempted to assume that siala is the term used in eastern Toba, and sihora is the term used in western Toba, but this is a totally unproven assumption:

M Siala Ulu | ET Siala WT Sihora | P Sikora (S Sihorlu, K Sikurun)

u, ě

Bu (M, T, S, P) Be (K, P) This diacritic appears twice in Pakpak, once with the phonetic value [u] and once with the

phonetic value [ə] (Schwa). The former is called kaběrětěn, and the latter kaběrětěn podi. The word kaběrětěn can clearly be traced back to Mandailing boruta (which is also spelled buruta) of which, by means of affixation, Toba haboruan and haborotan are derived. Both names are an interpretation of the words boruta and buruta. Boruta unquestionably consists of the base boru ‘daughter’ and the 1st person plural pronoun -ta.

We have already seen above that ha-...-an affixation regularly occurs in the regions north of Mandailing, where talinga becomes hatalingan, and amisara becomes hamisaran. This also happened with the root boru which became ha=boru=an. Once affixation had occurred, the word lost its meaning. The root boru is still recognizable but in combination with the circumflex this newly created word has become meaningless. When hatalingan became hatadingan, this was done in order to make the word meaningful by linking the meaning to the position of the diacritic relative to the aksara. The same happened also with haboruan, which was redefined as haborotan (ha=borot=an) because it is tied (the meaning of borot is ‘to tether’) to its aksara.

M Boruta (Buruta) | T Haboruan T Haborotan | | P Kaběrětěn [u] P Kaběrětěn Podi [ə] S Haboritan [u] | K Kěběrětěn [ə]

ng

B^ (India anusvāra) The diacritic with the phonetic value [ŋ] is one of the most interesting and revealing diacritics. Its

name, amisara in Mandailing, can be traced back directly to Sanskrit anusvāra, the diacritic used in

the Brahmic family of scripts to mark nasalization, and which usually takes the form of a dot which is placed on top of the aksara it modifies. While, depending on the location of the anusvāra in the word and the language within which it is used, its exact pronunciation can vary greatly, its Batak cousin, the amisara, has the value of the nasal ng only.

As usually, affixation occurred as in Toba, where it became hamisaraan. The s then became prenasalized while at the same time the final vowel was dropped resulting in haminsaran, which was used not only in Toba but also in Simalungun. The prenasalization did not only change the pronunciation considerably as ns is pronounced ts, but is also lead the path for a reinterpretation of the word. The root of the meaningless ha=misara=an changed to minsar, which then was apparently perceived as the prenasalized form of binsar –pronounced bitsar–‘to raise (of the sun)’. But it was only in Pakpak and in Karo where the reinterpretation was fully completed: in both districts the diacritic is known as kěbincarěn ‘the place where the sun rises, the east’ (T binsar, P/K bincar ‘to rise’). Here again, we can assume that the name of the diacritic is related to its position, which is to the top right of the aksara it modifies. One may as well imagine it as the sun that has just started rise above the horizon, which is the upper line of the aksara. This character also has a second name in Toba – paminggil ‘high-pitched sound’.

We have shown that amisara is quite clearly the Batak rendering of the Sanskrit anusvāra. If that is the case, then one should expect that the names of the other diacritics would also show at least some kind of Indian linkage. The fact that they do not is hardly surprising considering that only two of the many diacritics in Indian scripts, the anusvāra and the visarga have proper names.

M Amisara | T Hamisaraan T S Haminsaran T Paminggil | K P Kěbincarěn

i

Bi (J Ulu, L Olan, R Kaluan) The diacritic representing the i sound is called ulu ‘head’ in Java and in Bali, and this is due to the

fact that the ulu is a more or less round circle which is placed on top of the aksara it modifies. Calling this diacritic ‘the head’ is indeed very descriptive. In Lampung this diacritic was called ulan, or, according to Van der Tuuk (1855, p. 3), olang. Its position is still on the top of the aksara. It does not form a circle but looks like the Roman letters v or u. In most ulu scripts it forms a little dot that is placed above the character it modifies. In Mandailing as well as in Toba this diacritic is known as ulua. As usual, affixation occurs in Toba, where is became hauluan but there is also a second name, haluain. Hauluan has ulu ‘head’ as its root whereas haluain is a rather odd, and probably corrupted form, but the Javanese ulu, the Rejang kaluan, and the Lampung ulan, are still recognizable. Pakpak-Dairi kaloan and Karo kělawan are based on Toba haluain or Simalungun haluan.

M Uluwa | T Uluwa T Haluain (Hauluan), S Haluan

| P Kaloan K

Kělawan

o (ou)

BO L Kětulung (au), R Katulung (au)

In the Lampung (Funke, 1961, p. 342) and Rejang scripts there is a special diacritic for the diphthong /au/ which is known as kětulung or katulung, clearly recognizable in Simalungun hatulungan. The Simalungun language is the only Batak dialect with regularly occurring diphthong [ou] represented by the hatulungan. Some districts in Karo also have diphthong /ou/ which however is written with the same diacritic that is also used for /o/. Despite this, Karo actually has two diacritics, namely BO and Bo that both have the same name, kětolongěn, and which both can be used to

represent /o/ as well as /ou/. It is possible that one of the two diacritics once was used to specifically for /ou/ as it still is the case in Simalungun.

The existence of the diacritic for the diphthong au in Lampung (kětulung) and Rejang (katulung) proves that the diacritic BO is not a recent introduction. It also confirms the hypothesis that the Batak script of Simalungun, or more precisely East-Toba/Simalungun, is older than that of West Toba, Pakpak, or Karo.

M T (?) | S Hatulungan /ou/ | K Kětolongěn /o/

Ancient  Batak  Script?  The Batak script was almost exclusively used on perishable writing media such as bark, bone, and

bamboo. The number of stone inscription is very limited and the few existing inscriptions are also very difficult to decipher. This is especially true for the Sumbu inscription. This stone inscription which was found in the vicinity of Sidikalang (see Figure 4) is written in the typical northern variant of the

Figure 4 The Spread of the Batak Script

Batak script but also contains a number of unknown characters. In the first half of the 19th century a few rock inscriptions were discovered in the Toba and Simalungun area, but these are even more difficult to decipher. Two very short stone inscriptions are well legible, and both are written in a script, which most closely resembles the Mandailing script.

Candi  Manggis  Inscription  

The Candi Manggis inscription (Provincial Museum of North Sumatra No. 3495) was recently discovered in the Sosa district of the Padang Lawas regency of North Sumatra. The Padang Lawas archeological site which dates roughly into the 13-14th century comprises the river basin of the Barumun and its tributaries Pane, Sihapas, Sirumambe, and Sangkilon. Manggis is located southeast of Sangkilon, and hosts the remains of a temple (candi). The Candi Manggis inscription was found in the vicinity of the temple and hence it is possible that the Candi Manggis inscription also dates to the same period as the Padang Lawas archealogical site. The script, however, seems to be modern. The inscription is incised into the surface of an oval stone measuring roughly 80x30cm. The edges of the stone appear to be polished. The text of the first line could be reconstructed as I ma Tano Mandi “This is the land of Mandi”. However, as the Batak script does not use any spacing between words, we can also read this line as I ma tanoman di “This is the grave of”.

ai 4t qo m q\di

The first letter is easily identifiable as the letter a. Following the letter a there is a short horizontal line and possibly a second shorter horizontal underneath. There are only two options: If it is a character

(ina ni surat), the it can only be the letter la (l) which, however, is unlikely, not only because it only very vaguely resembles the letter la, but also because the space between the a and the following character ma seems to narrow to host a character. It is hence much more likely that this is not a character but a diacritic. In this position, only the diacritics o and i are permissible. The diacritic o (o) can be ruled out and hence it must be the diacritic i. Usually the i has the shape of a small circle, but occasionally, especially in the northern Batak scripts Simalungun and Karo, it can consist of two short horizontal lines shown here in combination with a: ai. In the Batak script initial /i/ can be expressed either by using the character i (I) or, more commonly, by a combination of the character a and the diacritic i as it here is the case.

The second letter of this line is without any doubt the character ma, which in the Batak script can have various shapes, among others m and M. The third letter is the southern ta (t), common in Mandailing, Angkola, Simalungun, and in parts of Toba. The fourth letter is unmistakably the letter

na, or more precisely the “ancient” variant which is very similar in shape to the na of Old Sumatran and Old Javanese inscriptions and manuscripts. This “ancient” na is only found in Mandailing and Angkola, and occasionally also in Toba, but it does not necessarily indicate that the manuscript or the inscription is ancient as the “ancient” na has been used until well into the nineteenth century. The more “modern” and common shape of the letter na is n. The letter na is followed by the diacritic o which has the shape of a cross (o). The next letter is not clearly identifiable. On a first look it looks like a da (d), but I am more inclined to see a ma. The following letter is unmistakably another “ancient” na that is followed by the virama known locally as pangolet, which has the shape of a backslash and serves to eliminate the inherent a-sound of the preceding letter rendering it n. The following letter is virtually unrecognizable. It may be da (d), but this is not certain at all. This letter is followed by the diacritic i which has the form of a small dot (i).

Line 2 can be tentatively reconstructed as amo\Pqisib Ompu ni si Ba… “Grandfather of Ba…”:

a 4 o \ P q i si b Calling an adult by name is considered an offence by the Batak. Adults are called by the names of their first child or grandchild. They are thus referred to as either ama ni "father of" or ompu ni "grandfather of" which is then followed by the name of the oldest child or grandchild.

The first letter is clearly the letter ‘a’ followed by the letter ma with a slightly rounded upper line which is very typical for Angkola and Mandailing: 4.The ma is followed by the diacritic o which has the shape of a cross (o). There is a slight possibility that it could be the letter ta (t), which would render the sentence amatpu ni diba which would not make sense at all, whereas the combination a plus ma plus o plus virama results in the syllable om. This is followed by a character with attached diacritic u which is always written under the character it modifies and often forms a ligature with it. It is hard to decide whether the combination of consonantal character and vocalic diacritic has to be read as du (D), lu (L) or pu (P) but I clearly favor the latter as it makes perfect sense in the combination with om, resulting in ompu “grandfather”.

The next letter is unmistakably another instance of the “ancient” Mandailing na followed by the diacritic i and hence resulting in the possessive marker ni “of”. The next letter is difficult to decipher: it may be la (l), but it looks more like da (d), or sa (s). The latter has the variant s which is most common in Simalungun. In combination with the following diacritic i the reading si (a kind of an article preceding proper names) makes perfectly sense. The last letter of this line is clearly the character ba (b). The indenture at the bottom is not visible, which, however, is quiet a common occurrence. In the northern Batak script Karo the indenture was even totally lost resulting in b.

Hence we can read this line as either “Grandfather of Diba” or “Grandfather of Si Ba”. The latter is correct as we will show in the discussion of line 3.

q\ d^ i P ia

The first letter of line 3 is another “ancient” na followed by the virama, and then by the letter da to which two diacritics are attached: the diacritic i in the form of a small circle, and the diacritic ng (the Sanskrit anusvāra rendered haminsaran by the Batak) which is placed on the top right of the letter it modifies and has the shape of a short horizontal line. This is the continuation of the name in Line 2 which results in the reading Si Banding. Si is a kind of article for proper names. As Banding is not a common name it is possible, albeit unlikely, that a letter was omitted, and that the writer intended to write baganding. Si baganding is not only the name of a snake, but also the name of a mythological being in form of a snake called si baganding tua “beneficial Baganding” that is said to bring prosperity to the owner of a house. But this is a remote possibility only and we have to accept that the person’s name is indeed Ompu ni Si Banding.

Unfortunately it is impossible to make sense of the remainder of this line. The next letter seems to be a pa, which usually is straight but sometimes also curved (p). In Simalungun we have the form p which is closest to the shape of our letter pa, which here comes with the diacritic u resulting in pu.

The following letters can not be deciphered except for the very end of the line where we can recognize the diacritic i and the letter a.

Although it is impossible to decipher the inscription in its totality, we can make the following observations:

1. The inscription consists of three parallel lines. Line two is written underneath line 1 and has exactly the same starting and the same end point, and the same is true for line 3. From this we can conclude that the three lines are complete.

2. The text can tentatively be reconstructed as I ma

tanoman di Ompu ni Si Banding … “This is the grave of Ompu ni Si Banding…” followed by a few undecipherable letters. We can hence assume that this is a tombstone.

3. The fact that the inscription was discovered in the vicinity of Candi Manggis, a temple from the thirteenth or fourteenth century, leads us to assume a similar age for the inscription. The script does not show any traces of antiquity, but since the earliest known Batak text dates back to the second half of the eighteenth century, we know that the Batak script has not undergone any changes within the last two and a half centuries. It is hence premature to rule out that this text is indeed seven hundred years old, but on the other hand, we are also unable to confirm an antiquity of the text based on the script.

Figure 5 Padang Bujur Inscription (Schnitger 1939, Plate VII)

Padang  Bujur  Inscription  

The administrative unit (desa) Padang Bujur (Desa Padang Bujur, Kecamatan Padang Bolak Julu, Kabupaten Padang Lawas Utara, Sumatra Utara 22753) is located within the Padang Lawas

archaeological complex.

The so called megalithic remains of Padang Bujur consist of 15 stone planks, one

of the planks has an inscription in Batak script. It has an u-shape with both legs in the west. The structure measures about 4.5 x 6.3 meters. The inscription is flanked by two monitor lizards, one above the text facing to the right (on Figure 5 this is the left lizard facing up as the text runs from bottom to top), and the other beneath the text facing to the left. The inscription on the stone plank is reported to be “13 cm in height, 107 cm in length, and 60 cm in width” (Sukawati Susetyo 2006:319)

Schnitger describes the inscription rather briefly: “Inscription at Padangboedjoer, announcing the name of Radja Soritaon. The two lizards are supposed to entice all the lizards in the neighbourhood and urge them to weep for the dead.” (Schnitger 1939: Plate VII).

The two lizards are known as boraspati ni tano, representatives of the underworld (banua toru), and the deity of the fertility of the earth. Warneck describes the boraspati6 as “a goddess dwelling under the earth that manifests itself in shape of a lizard and is as such present in the houses. Offerings are presented at the time when the fields are being worked on to assure fertility. The offerings are in the form of rice cakes, rice flour, betel, and flowers. […] When the post holes for a new house are dug and the supporting stones are set (by which both the soil and the boraspati ni tano are disturbed), the magician (datu) has to utter a prayer, and present an offer to the boraspati. Offerings to the diety are also due when a grave is dug so that it will not get angry towards the human beings disturbing the earth.” (Warneck 1909:38)

Raja sari rjsri taon ma tano\ M n puna tano n\ Pn tno

i ma halak na mora I M hlh\ n Mor di padang bujur di pd^ BjR\

6 Derived from Sanskrit Bṛhaspati, the lord of prayer.

Although the remainder of the inscription is illegible, the first lines give us sufficient clues about the nature of the text, which should, of course, also be analyzed within the context of this megalithic structure. Unfortunately I do not have sufficient information about this megalithic structure, but it seems to be likely that it is part of the Mandailing burial tradition as Schnitger suggests. Our inscription could hence be interpreted as the grave stone, or a memorial, but this is not supported by the text. The inscription has more the nature of a piagam or a royal edict. In our case it is issued by Raja Soritaon and indicates that the land where the inscription is erected, is his land. This prasasti can most likely be considered as a border marker. The text can be reconstructed as Raja Soritaon nampuna tano i ma halak na mora di Padang Bujur. “Raja Soritaon, the noble from Padang Bujur, owns this land”. On the other hand, the fact that the two boraspati are engraved in the stone may be indicative of its function as a grave marker.

Schnitger interpreted the first two words as Raja Soritaon. From the photograph all I can read is Saritaon, but I agree with his interpretation. Soritaon is a classical Batak name still in use today, especially in the southern Batak districts from where this inscription originates. Manpuna strikes me as odd as one should expect the artisan to be more careful in inscribing the text onto the stone. Yet, it appears to be a mistake for nampuna (nM\Pn), derived from na ompuna, “the owner” which has to be read together with tano: na ompuna tano “the owner of the land”. The following character is hard to read, and could be the character i (I) but I am not certain at all. The next character is clearly ma (M)—a particle. This is followed by halak na mora – all characters are clearly legible. Line 5 is also easy to read: di Padang Bujur. The last character of this line is the isolated character a, which must probably be read in combination with the following line, which may be legible–but not from the photographs that are at my disposal. A more literal translation of the text is “Raja Soritaon owns this land, a rich man in Padang Bujur”. Nobles in Sumatra are often literally termed ‘the rich one’ (cf. Karo Batak sibayak, and Malay urang kaya). From Mandailing it is reported: “The leaders called Na Mora (the Nobles) are the ruling elite from the landed clan (marga tanah). The landed clan or nobility are the rulers (rajas). In Lower Mandailing, the Nasution are the noble house while in Upper Mandailing, the Lubis are the noble house. By virtue of being victorious in war or being founders of new settlements in the past, they are rajas until today.7”

7 Cited 2011-03-11 http://www.mandailing.org/Eng/namorana.html

The  Muara  Takus  inscription  

This inscription was found in Candi Muara Takus, Kabupaten Kampar, Riau Province. The Muara Takus archaeological complex consisting of a Buddhist temple complex surrounded by a one metre tall stone perimeter wall that measures 74 x 74 metres. The site is not dated with precision but is very probably contemporary with Padang Lawas. As it appears from the photograph, the stone seems to have belonged to a larger mural, and all that is left is the present fragment. The surface on which the inscription is engraved appears to have been flattened, and probably even polished to create an even surface. The surface appears much darker than the material of the stone itself. From the photograph it is impossible to judge whether the dark surface is has been artificially created or whether it is the result of natural patination. As the scratches, and the other elements, of which we believe that they have purposely been created, are much lighter than the surface of the stone, one can conclude that either the stone was naturally flat, and that the inscription was chiselled onto the patinated surface, or that the stone was flattened and polished and subsequently darkened so that the text and the other elements are clearly visible. Because of the numerous scratches and the general wear of the surface it is difficult to decide which elements were purposely chiselled onto the surface. There is a relatively large oval of which six rays emerge as if the illustrator intended to draw a stylized picture of a fire, or a light. As it sites perpendicular over the text box it seems to be somehow related to the text. The text is, as we just mentioned, written in a perfectly drawn rectangular box with a height of approximately one third of the width. There are at least three aksara. The fourth letter is vary vague, and also much smaller than the preceding letters so that we don’t believe that this an aksara, but it could be the diacritic o.

The inscription could be read pata ro if the text is correctly reconstructed as ptro, of which I am not convinced as a lot of fantasy is needed to reconstruct the diacritic o (o). Besides, as pata has no evident meaning, it would have to be interpreted as a personal name: Pata comes. I am somewhat more inclined to reconstruct patar ‘clear, evident’ (ptr\).

The first letter, which we believe is the curved variant of the letter p (pa) appears to be written slightly too high in relation to the following letter which without no doubt is the southern ta (t). Moreover, there is a dot below, and slightly to the right of the letter. This opens the vague possibility that this is not the letter pa, but the letter ha (h) instead. It could also be the variant g of the letter ga (g). The third letter seems to be the aksara ra (r) but here, again, it is, in relation to the preceding letter ta, positioned slightly too high and also a little bit undersized. Yet, we firmly believe that this is the letter ra, especially as there is no plausible alternative to this interpretation. It seems to be evident that there is something following the letter ra, but what it is, is virtually impossible to decide. The size of the textbox clearly indicates that the scribbled lines where intended to represent either an aksara or a diacritic. If it were an aksara, then the two most plausible letters would have been da (d) or ma (m), but both would not make any sense. The interpretation of this undecipherable scribble as the diactitic o, would at least make sense, as it would form the word ro ‘to come’. There is also the possibility that the text should not read ptro but ptr\ (patar) which means ‘clear, evident’ as in patar tarida ‘clearly visible’, and could hence explain the graphical element that we interpreted as a stylized light. I will not attempt to interpret this text any further as this can only result in more or less wild speculation. Some speculation, however, is necessary in the dating of the three inscriptions discussed in this chapter. Both the Muara Takus as well as the Candi Manggis inscriptions where both unearthed in the vicinity of archaeological complexes that are at least 700 years old. The third text, the Padang Bujur inscription, originates from an undated megalithic site in Padang Lawas. Miksic, who has worked on megalithic sites in Malaysia and West Sumatra, comes to the conclusion that “the tradition of erecting batu tagak [megaliths, UK] in West Sumatra actually began or greatly increased in the fourteenth century, as a by-product of social change which arose in the Minang heartland after it became the centre of a late classical Indonesian kingdom with contacts reaching Java China, and India.” (Miksic 1987) It is, of course, problematic to conclude a parallel development for the Padang Lawas area, which in many respects (especially demographically and ecologically) is profoundly different from the Minangkabau highlands. Yet, in both regions we can observe the same phenomenon: the import of a highly stratified classical culture of mainly East Javanese (Majapahit) provenance into a largely tribal society. By the mid-fourteenth century the Minangkabau-Dharmasraya area (Malayu) was ruled by king Adityawarman, who is said to be the offspring of a marriage between a Malay princess and a high court official from Majapahit, but even before Adityawarman’s time, the Minangkabau-Malayu area had already become part of the cultural sphere of East Java as it is testified by the Dharmasraya temple complex. The local culture was of course not radically transformed, but many elements of a more tribal culture were first retained, and later restored. When the first Europeans visited the Padang Lawas area in the nineteenth century, the local population had very little, if any recollection of their “classical” historical past while the Minangkabau still had some vague recollection of their past, but the name of the illustrious king Adityawarman had become forgotten and all that was remembered was his title maharajadhiraja (king of all great kings). It is hence perceivable that the megaliths, manifestations of a more ‘tribal’ culture, were erected at the same time when temples of East Javanese

style were built. If such as dual culture existed side-by-side, then it is also perceivable that the Kawi script, which, as we can assume, was associated with the Java-influenced court culture, was used contemporaneous with the local (“tribal”) script such as the Batak script. The fourteenth century Tanjung Tanah manuscript is the only known paper manuscript in Kawi script, which, spatially separated from the main text in Kawi letters, also contains a text in a local script. As the Tanjung Tanah manuscript was intended for a Kerinci audience, and presumably dispatched to Kerinci right after it was written, the text in the local script was likely added in Kerinci and not in Dharmasraya. The script is very similar to the Kerinci script known from eighteen and nineteen century sources. The same can also be concluded from the three Batak inscriptions discussed in this chapter. If we assume that these three inscriptions date back to the fourteenth century, one can conclude that there was little, if any, development in the Batak script from the fourteenth to the eighteenth and nineteenth century. The Muara Takus inscription also suggests that in the past the Batak script was used beyond the borders of what presently is known as the Batak settlement area, which comprises most of the North Sumatra province. It is known that in the past the settlement area of the Batak extended beyond its current borders. Place names along the east coast of North Sumatra, and in the northern part of West Sumatra, currently settled by coastal Malays and Minangkabau, testify that Batak settlements used to be located in these areas. In the Minangkabau area the entire region up to and including Rao was reported as being former Batak settlements. Muara Takus, however, cannot be considered a Batak border region, and it is very doubtful whether it was ever settled by Batak. The existence of an inscription in Batak script—so far we have assumed that the text is also in Batak language, but as the text consists of one word only, this cannot be ascertained— in an area located more than 150 kilometres from the borders of the current Batak settlement area, raises more questions than it answers.

Conclusion  It seems to be evident that Mandailing must be considered the home-region of the Batak script.

From here it spread to East-Toba and Simalungun, which most likely at that stage still had the same script of which later the two slightly different scripts developed. Simalungun developed its characteristic unconnected lines, but remained otherwise quite conservative. East Toba also did not develop that much. Together with Simalungun it retained the southern ta and the southern wa: t, and w, while West Toba developed two very different shapes of the ta and the wa: The northern ta f and the Northern wa v. It is doubtful that the northern ta and wa were the exclusive forms in West Toba, as there are quite a few manuscripts where both forms occur, and there are also manuscripts with a likely West Toba provenance where “eastern” shapes dominate, but the northern ta and wa certainly became exclusive once they entered Pakpak. Karo seems to have been influenced by Pakpak as well as by Simalungun. It retained exclusively the northern ta and the southern wa.

Table 12 The 5 Batak scripts

Karo Pakpak Toba Simalungun Mandailing

a a a a A a

ha a a h K h

ka k k k K k

ba b b b B b

pa p p p P p

na n n n N n

wa w w w v W w

ga g g g G g

ja j j j J j

da d d d d d

ra r r r r r

ma m m m m m

ta t t f t t t

sa s s s s s

ya y y y y y

nga < < < < <

la l l l l l

nya [ [ [

ca c C c c

nda q mba B i I I I I I

u U U U U U

References   Abas, H., & Grimes, C. E. (1994). Bugis. In D. T. Tryon & S. Tsuchida (Eds.), Comparative

Austronesian dictionary: an introduction to Austronesian studies. Part 1 : Fascicle 1 (pp. 549-561). Berlin, New York: Mouton de Gruyter.

Adelaar, K. A. (1981). Reconstruction of Proto-Batak phonology. Nusa, 10, 1-20. Bright, W. (1996). The Devanagari script. In P. T. Daniels & W. Bright (Eds.), The world’s

writing systems (pp. 384-390). Oxford, New York: Oxford University Press. Bringhurst, R. (2002). Voices, languages, and scripts around the world. In J. Berry (Ed.),

Language Culture Type : International type design in the age of Unicode (pp. 3-23). Association Typographique Internationale (ATypI)/Graphis.

Casparis, J. G. d. (1975). Indonesian palaeography: a history of writing in Indonesia from the beginnings to c. A.D. 1500. Leiden: E.J. Brill.

Crawfurd, John (1852). A grammar and dictionary of the Malay language with a preliminary dissertation. Vol. 1 Dissertation and Grammar. London: Smith . Elder & Co.

Daniel, P. T. (1996). The first civilizations. In P. T. Daniels & W. Bright (Eds.), The world’s writing systems (pp. 21-32). Oxford, New York: Oxford University Press.

Dempwolff, Otto (1934-1938). Vergleichende Lautlehre des austronesischen Wortschatzes. Volumes 1-3. Berlin: D. Reimer

Diringer, D. (1968). The alphabet: A key to the history of mankind. London, Melbourne, Sydney, Auckland, Bombay, Toronto, Johannesburg, New York: Hutchinson & Co.

Funke, F. W. (1961). Orang Abung : Volkstum Süd-Sumatras im Wandel. Leiden: E. J. Brill. Groeneboer, K. (2002). Een vorst onder de taalgeleerden: Herman Neubronner van der Tuuk,

Taalafgevaardigde voor Indië van het Nederlandsch Bijbelgenootschap 1847-1873. Leiden: KITLV Press.

Harahap, E. S. (1960). Perihal bangsa Batak. Djakarta: Bagian Bahasa, Djawatan Kebudajaan.

Hardjadibrata, R. (1978). Consonant clusters in Indonesian. Paper presented at the Second International Conference on Austronesian Linguistics, Canberra.

Hariara, J. M. (1987). Hata Batak maninggoring. Bagian rangsa ni andung dohot hadatuon. Jakarta: Balai Pustaka.

Helfrich, O. L. (1904). Bijdragen tot de kennis van het Midden Maleisch (Bĕsĕmahsch en Sĕrawajsch dialect). Verhandelingen van het Bataviaasch Genootschap van Kunsten en Wetenschappen, 53, 265-273.

Holle, K. F. (1882). Tabel van Oud- en Nieuw- Indische alphabetten. Batavia/'s Hage: W. Bruining & Co. / M. Nijhoff.

Jaspan, M. A. (1964). Folk literature of South Sumatra: Redjang KA-GA-NGA texts. Canberra: Australian National University.

Kern, H. (1907). De wij-inscriptie op het Amoghapāça-beeld van Padang Candi (Batang Hari-districten); 1269 Çaka. Tijdschrift voor Indische Taal-, Land en Volkenkunde, 49, 159-170.

Kozok, U. (2000) On writing the not-to-be-read: literature and literacy in a pre-colonial 'tribal' society. Bijdragen tot de Taal-, Land en Volkenkunde 156, 33-55.

Kozok, U. (2006). Kitab undang-undang Tanjung Tanah : Naskah Melayu yang tertua. Jakarta: Yayasan Naskah Nusantara, Yayasan Obor Indonesia.

Matthes, B. F. (1858). Makassaarsche spraakkunst. Amsterdam: Nederlandsch Bijbelgenootschap, F. Muller.

Meerwaldt, J. H. (1922). De nieuwe Bataksche Letterkunde. Mededelingen van wege het Nederlandsch Zendelinggenootschap, 66, 295-311.

Miller, C. (forthcoming). A Gujarati origin for scripts of Sumatra, Sulawesi and the Philippines. Paper presented at the 36th Annual Meeting of the Berkeley Linguistics Society. 6-7 February 2010, Berkeley, California.

Parkin, H. (1978). Batak Fruit of Hindu Thought. Madras: Christian Literature Society. Postma, A. (1992). The Laguna copper-plate inscription: text and commentary. Philippine

Studies (Quezon City), 40(2), 183-202. Ramos, T. V. (1971). Tagalog structures. Honolulu: University of Hawaii Press. Sarumpaet, J. P. (1982). Linguistic Varieties in Toba-Batak. In A. Halim, L. Carrington, & S.

A. Wurm (Eds.), Papers from the Third International Conference on Austronesian Linguistics (3: Accent on variety), pp. 27-28). Canberra: Australian National University.

Siti Hawa Hj. Salleh (2010). Malay Literature of the 19th Century . Kuala Lumpur: Institut Terjemahan Negara Malaysia Berhad.

Sukawati Susetyo (2006). Some notes on the megalithic remains in Padang Lawas. In: Truman Simanjuntak, M. Hisam, Bagyo Prasetyo, and Titi Surti Nastiti (eds.). Archaeology: Indonesian perspective : R.P. Soejono's festschrift . Jakarta: Lembaga Ilmu Pengetahuan Indonesia, p. 317-

Tampubolon, R. P. (2002). Pustaha Tumbaga Holing: Adat Batak - Patik Uhum. Buku I dan II (Cetakan ke-2 ed. Vol. 1). Jakarta: Dian Utama.

Tuuk, H. N. v. d. (1855). Over schrift en uitspraak der Tobasche taal als eerste hoofdstuk eener spraakkunst. Amsterdam: C. A. Spin & Zoon.

Tuuk, H. N. v. d. (1861a). Bataksch-Nederduitsch Woordenboek. Amsterdam: F. Muller. Tuuk, H. N. v. d. (1861b). Stukken in het Mandailingsch (2). Amsterdam: F. Muller. Tuuk, H. N. v. d. (1971). A Grammar of Toba Batak (13). The Hague: Martinus Nijhoff. Wade, G. (1993). On the possible Cham origin of the Philippine scripts. Journal of Southeast

Asian Studies, 24(1), 44-87. Willer, T. J. (1846). Verzameling der Battahschen wetten en instellingen in Mandheling en

Pertibie. Gevolgd door een overzigt van land en volk in die streken. Tijdschrift voor Nederlandsch Indië, 8(2), 145-424.