Java took Islam from merchants — then stopped building temples (1300)

Java and Sumatra before the sultanates

The state that described itself in 1365

In 1365 a Buddhist cleric of the Majapahit court finished a poem in ninety-eight cantos and called it the Deśawarṇana — "the depiction of the districts." Its author, Mpu Prapañca, held the office of dharmādhyakṣa ring kasogatan, Superintendent of Buddhist Affairs, an appointment of state; six years earlier he had ridden in the baggage of King Hayam Wuruk's royal progress through the eastern districts of Java and taken notes on everything he saw.³ The poem is the closest thing the pre-Islamic archipelago has to a self-portrait, and it describes a society organised around premises that would be unrecognisable within two centuries.

The king was not a Muslim's idea of a king. He was a manifestation of divinity in the world, and when he died he would be installed as a god in a funerary temple of his own, his ashes under a stone image with his features and Śiwa's attributes. Prapañca lists these installations by name and date; the royal dead of Majapahit occupied a calendar of anniversaries as precise as any liturgy. Religion was an office of state, held in two branches — Śaiva and Buddhist — each with its own superintendent, each drawing revenue from tax-immune estates granted by inscribed copper plate.³⁵ A generation earlier Mpu Tantular, writing the Kakawin Sutasoma at the same court, had put the theological settlement into a single line: bhinneka tunggal ika, tan hana dharma mangrwa — they are different, they are one, there is no duality in truth. He meant that Śiwa and the Buddha were one substance. The Republic of Indonesia took the first three words for its coat of arms in 1950.

What the state claimed and what it had

Cantos 13 and 14 of the Deśawarṇana set out the reach of Majapahit as the court wished it understood: the tributary lands run from the Sumatran polities of Jambi, Palembang, Lamuri and Samudra, up the Malay peninsula, across Borneo, into Sulawesi and out to the clove islands of Maluku.³ Most historians read the list as a map of ambition rather than administration — a claim of prestige over ports that traded with Java rather than obeyed it.⁵ But one entry deserves attention. Samudra, on the north coast of Sumatra, appears in Prapañca's roll of dependencies. By 1365 Samudra had been ruled by Muslim sultans for at least two generations. The Śaiva-Buddhist court of Java was listing the first Islamic state of the archipelago among its tributaries, and thought nothing of it.

That indifference is the most important fact about the receiving culture. Majapahit was not a religiously anxious society. It had absorbed Indic religion by sea over the previous thousand years without an invasion, fused two rival soteriologies into a state cult, and continued to treat the arrival of foreign clergy as a commercial and diplomatic matter rather than a threat. Wealth came from wet rice on the Brantas and Solo plains and from a carrying trade in rice, cloves, nutmeg and pepper.⁵ Trowulan, the capital, was a brick city of walled compounds, canals and reservoirs — the Segaran tank alone measures some 375 by 175 metres — with temple foundations and split gates set among them. Writing was a thousand years old there, in the Kawi script descended from South Indian Pallava letters, incised on palm leaf and copper.⁴

What a woman could do, and what a debtor could not

Two features of the pre-transmission society matter for what follows, because both were altered and neither is usually mentioned.

The first is the position of women. Across pre-modern maritime Southeast Asia women held and inherited property in their own names, ran the markets, controlled textile production and rice-pounding, and appear in Javanese inscriptions as parties to contracts and witnesses to grants. Anthony Reid, whose two-volume account of the region between 1450 and 1680 remains the standard synthesis, treats this economic autonomy as one of the defining regional traits, giving women "an almost uniquely strong position in sexual politics" by the standards of the literate agrarian world.¹⁷ Descent was reckoned bilaterally. Post-marital residence was frequently uxorilocal — the husband moved to the wife's family. Divorce was available on either side.

The second is bondage. It would be false to present the archipelago before Islam as a society without unfree labour. Debt bondage was pervasive, war captives were distributed as property, and inscriptions record servile categories attached to temple estates.¹⁶ What did not exist was a religious criterion for who could be held. Bondage in Java before the fourteenth century was a matter of debt, capture and birth. It was not a matter of belief.

Muslims were already there, and nothing had changed

The single oldest Islamic gravestone in the archipelago sits at Leran, a village near Gresik on the north coast of Java, and it marks the burial of a woman: Fāṭimah bint Maimūn bin Hibat Allāh, dead on 7 Rajab 475 AH — 2 December 1082 — under an epitaph cut in Kufic Arabic. The reading of the date is contested and the stone may be an import, but no proposed alternative moves it later than the twelfth century.⁷¹⁰ Muslims were therefore living, dying and being buried on Java roughly three hundred years before any Javanese ruler professed Islam.

This is the calibration the record needs. The presence of Islam in the archipelago is not the transmission. Muslim ships had been calling at Sumatran and Javanese ports since the eighth century; Muslim quarters existed in the harbour towns; Muslim graves accumulated quietly in the sand behind them. For three centuries none of it altered a single Javanese institution. The transmission is the moment the receiving society's own machinery — its law, its script, its idea of legitimate rule, its category of person — began to be rebuilt around the newcomers' premises. That began somewhere in the late thirteenth century, and it began at sea.

The transmission: monsoon, ledger, marriage, grave

Marco Polo, wind-bound at Sumatra, 1292

In 1292 a Venetian in the escort of a Mongol princess bound for Persia was forced by weather to wait out five months on the north coast of Sumatra. Marco Polo used the time to inventory the kingdoms of what he called Java the Less, and of one of them he wrote a sentence that has been quoted in every history of Islamic Southeast Asia since: "This kingdom, you must know, is so much frequented by the Saracen merchants that they have converted the natives to the Law of Mahommet — I mean the townspeople only, for the hill-people live for all the world like beasts."⁶

Strip out the contempt and what remains is an accurate piece of reporting, and a prediction. The kingdom was Ferlec — Perlak, on the eastern tip of Aceh. The agents were merchants. The mechanism was frequency of contact rather than preaching. And the division Polo drew, between converted coast and unconverted interior, would define the religious geography of the archipelago for the next six hundred years and in places still does. The Javanese distinction between the Muslim north-coast pasisir and the Hindu-Buddhist inland kejawèn courts; the Batak, Toraja and Dayak interiors that never converted; the highland peoples of Aceh whom Polo dismissed as beasts and whose descendants were still being fought as unbelievers in the nineteenth century — all of it is visible in a single sentence written in 1292 by a man complaining about the wind.

The grave of al-Malik al-Ṣāliḥ, 1297

Five years after Polo sailed, a ruler died at Samudra, a few days up the coast from Perlak, and was buried under a marker naming him Sultan al-Malik al-Ṣāliḥ, deceased in Ramaḍān 696 AH — between 23 June and 22 July 1297. He is conventionally the first Muslim sovereign of Southeast Asia, and his headstone is conventionally the founding document of Islamic statehood in the region.⁷

Elizabeth Lambourn's re-examination of the stones complicates that neatly. Reading the monument as what she calls an integral cultural product — text, ornament and material weighed together rather than the inscription read alone — she concludes that al-Ṣāliḥ's markers bear no relationship to South Asian tombstone traditions and belong squarely within the north Sumatran batu Aceh school; and that they were most probably carved in the fifteenth or sixteenth century to replace an earlier monument.⁷ The founding document of Islamic Southeast Asia is, on the best current reading, a later commemoration. That does not make the 1297 date wrong. It makes the stone evidence of something else: a sultanate two centuries on, deliberately building its own genealogy in carved granite.

Around al-Ṣāliḥ lie roughly 150 more graves. Claude Guillot and Ludvik Kalus catalogued the Pasai epigraphic corpus in 2008 and sorted it into six typological classes spanning the thirteenth to the sixteenth centuries, one of which is an ogival form imported from Cambay in Gujarat.⁸ Lambourn traced that import stream separately: Gujarati grave memorials, quarried and carved at Cambay, shipped as cargo to Samudra-Pasai and to Gresik on Java in the fifteenth century.²¹ The same vessels that carried Gujarati cloth south and pepper north carried the headstones. Islam arrived in the archipelago the way everything else did — on a manifest.

The circuit and the people on it

The carriers were plural, and flattening them into "Arab traders" loses the mechanism. The network that reached the archipelago was assembled from at least six distinct communities, each with its own base, its own commodity and its own reason to be there:

• Ḥaḍramī sayyids out of Tarim in southern Arabia, trading partly on documented descent from the Prophet — a form of capital that converted into marriage alliances with local ruling houses wherever they settled. Engseng Ho's study of the diaspora traces how genealogical writing let these families become local in three regions at once without ceasing to be connected across the ocean.²⁰
• Persian Kāzarūnīs, from the Shiraz hinterland, devotees of the Sufi shaykh Abū Isḥāq al-Kāzarūnī (d. 1033). Their order ran hospices in ports around the Indian Ocean; merchants crossing toward China vowed sums to the shrine against storm and piracy and kept written records of the vows, which agents collected on arrival.²⁵ It was, functionally, marine insurance underwritten by a dead saint. A Kāzarūnī tomb complex stood at Pasai.⁷
• Gujarati merchants of Cambay, whose cotton cloth was the working currency of the eastern spice trade and whose stone-carvers supplied the gravestones.²¹
• Tamil-speaking Māppila and Marakkayar traders of the Malabar and Coromandel coasts, Shāfiʿī like the Ḥaḍramīs, with the shortest crossing to the Straits of anyone in the network.
• Bengali shippers out of Chittagong and the Ganges delta.
• Chinese Muslim shipmasters out of Quanzhou and Guangzhou, whose role Geoff Wade has argued was decisive in the second half of the fourteenth century and who arrive in force with the Ming treasure fleets after 1405.¹³¹⁴

Wade's summary of the geography is the cleanest available: "Islam came to the polities and societies of Southeast Asia by sea, along the girdle of trade which extended from the Arab and Persian worlds through the ports of South Asia, to Southeast Asia and onwards to the southern extensions of the Chinese world in the East China Sea."¹⁴ And of the agents: "The earliest introduction of Islam to Southeast Asia was by Muslim merchants who travelled along the maritime routes which, for at least a millennium earlier, had connected the two ends of the Eurasian continent."¹⁴

Why the monsoon did the work

The mechanism that turned visiting merchants into a resident Muslim gentry was meteorological. The monsoon reverses twice a year. A ship from Gujarat or the Coromandel riding the north-east monsoon into the Straits arrived between December and March and could not sail home until the south-west monsoon returned. That is four to six months ashore, every voyage, for a man carrying capital.

What such men did, everywhere in the Indian Ocean, was marry locally. The children were Muslim, raised in the mother's language, connected by kinship to the harbour-master's family and by religion to a commercial network stretching to Aden. Within two or three generations the wealthiest lineages of a port were Muslim, related to the ruler, and in possession of the credit lines that made the port work. Conversion of the ruling house then followed as a commercial and dynastic adjustment rather than a spiritual crisis — which is why the Malay chronicles narrate it as a dream rather than an argument. In the Hikayat Raja-Raja Pasai, the ruler Merah Silu meets the Prophet in his sleep, wakes circumcised and able to recite the whole Qurʾān, takes the name al-Malik al-Ṣāliḥ, and has the conversion witnessed by a shaykh arriving from Mecca.²³ The story is not history. It is a fifteenth-century Malay court explaining to itself, in the only register available, why its legitimacy now ran through Arabia.

Ibn Baṭṭūṭa at Samudra, 1345

Fifty years after Polo, the Moroccan jurist Ibn Baṭṭūṭa spent a fortnight at Samudra as the guest of Sultan al-Malik al-Ẓāhir, and left the first insider's account of a Southeast Asian Muslim court. He found a wooden-walled town, a Shāfiʿī establishment whose observances matched what he had seen among the Māppilas of coastal India, and a sultan he describes as a humble-hearted man who walks on foot to the Friday prayer.⁹ He also recorded, in the same breath, that the sultan's subjects took pleasure in warring for the faith and had the upper hand over all the infidels in their vicinity.⁹ He called Samudra the furthest point of the Muslim world, and left for China in one of the sultan's own junks.

Both halves of that portrait belong in the record. The piety is real; so is the raiding. Within fifty years of the first Muslim sultan's death, the frontier between Samudra and its unconverted neighbours had become a religious frontier, and a religious frontier in Islamic law is not a neutral line. It is the boundary of the dār al-Islām, and it governs — among much else — who may lawfully be taken.

Muslims inside Majapahit: the Troloyo stones

Java's evidence is stranger and more important. A little over a hectare of ground south of Trowulan village, in the Mojokerto regency of East Java, holds a grave complex known as Troloyo or Tralaya. It sits within the archaeological footprint of the Majapahit capital. The graves are Muslim. The dates on them are not.

L.-C. Damais published the corpus in the Bulletin de l'École française d'Extrême-Orient in 1957, and the arrangement he described has never been satisfactorily explained away.¹⁰ The headstones carry, on one face, a date written in Old Javanese numerals in the Śaka era — the calendar of the Hindu-Buddhist state — and on the other, a pious formula in Arabic language and script. M. C. Ricklefs gives the earliest precisely: "The earliest of the dated Tralaya gravestones carries the dates 1298 (AD 1376-77) in Old Javanese numerals on one side and a pious Islamic formula in Arabic language and script on the other."¹ The series runs forward from there into the seventeenth century. Some of the markers carry the Sūrya Majapahit, the eight-rayed sun-rosette that served the dynasty as an emblem, and the kāla mask that guards the gateway of every Javanese temple.

Damais's conclusion, reaffirmed and extended by Hadi Sidomulyo in 2012, was that the dead were not foreign merchants.¹⁰¹¹ The carving is of court quality; the ground is beside the palace; the dating convention is the state's own. These were Javanese Muslims, and not a few of them were members of the Javanese nobility.¹¹ By the 1370s there were Muslims of rank living and dying inside the capital of a Śaiva-Buddhist empire, buried under the emperor's sun.

What Ma Huan saw in 1416

The last pre-conquest snapshot comes from a Chinese Muslim. Ma Huan sailed as interpreter with Zheng He's fleets and wrote up the ports in the Ying-yai Sheng-lan. Describing Java, he divides the population into three classes: Muslims from the western kingdoms who had migrated there as merchants; Chinese from Guangdong, Zhangzhou and Quanzhou, of whom many likewise followed the Muslim religion, fasting and doing penance; and the natives of the country, whom he describes with open contempt.¹² He names the ports where these communities lived — Tuban, Gresik, Surabaya — the same north-coast harbours that would supply the sultanates.

Set against Prapañca's poem of fifty years earlier, the passage is a measurement. In 1365 the court could list a Muslim sultanate among its dependencies without comment. By 1416 a foreign observer counted the Muslim mercantile population of Java as one of the three constituent categories of the island's people, and the Chinese half of it as substantially Muslim too.

The argument scholars are actually having

There is a genuine, unresolved dispute about where the archipelago's Islam came from, and the atlas will not pretend to settle it. The principal positions:

1. Gujarat. Advanced by J. Pijnappel, developed by Christiaan Snouck Hurgronje and given its epigraphic form by J. P. Moquette, who read the Pasai tombstones as Cambay products. The argument rests on the antiquity of Gujarat–archipelago trade and on stone typology.
2. Bengal. S. Q. Fatimi's objection: the Pasai stones resemble Bengali work more closely than Gujarati.
3. Coromandel. Morison's objection, the sharpest of the three: Gujarat was still under Hindu rule when al-Malik al-Ṣāliḥ died in 1297, so it is an unlikely source of missionary Islam at that date.
4. Arabia and Egypt. The doctrinal argument: the archipelago is overwhelmingly Shāfiʿī, as are Ḥaḍramawt and Lower Egypt, whereas Gujarat is largely Ḥanafī. A school of law is harder to acquire by accident than a tombstone.
5. Persia. Hoesein Djajadiningrat's case from the Persian stratum in Malay religious vocabulary, reinforced by the Kāzarūnī evidence.

G. W. J. Drewes reviewed the whole field in 1968 and found none of the theories evidentially secure.²⁶ The reason is probably that the question is badly posed. The network was plural by construction: a Ḥaḍramī sayyid in a Gujarati-owned ship crewed on the Coromandel and financed against a Persian saint's shrine does not have a single point of origin. A. H. Johns's influential 1961 proposal — that the decisive agents were wandering Sufis whose cosmology mapped easily onto Javanese mysticism — captures something real about the content that travelled and rather less about the carriage, and Johns himself later softened it.¹⁵ Anthony Reid's commercial reading, that Islamisation in the archipelago accelerated with and because of the fifteenth- and sixteenth-century boom in the spice trade, explains the timing better than any of the origin theories explain the route.¹⁶

What changed, and what it replaced

A script, a calendar and a legal person

The first durable changes were bureaucratic. Arabic script was adapted for Malay as Jawi, with letters added for sounds Arabic lacked, and Malay so written became the administrative and religious lingua franca of the whole archipelago — the ancestor of the two national languages spoken today by some three hundred million people. The earliest surviving monument of it is a granite stele found in 1887 at Kuala Berang in Terengganu, whose damaged date has been read as either 702 AH (1303) or 789 AH (1387). Its text is not devotional. It is a ruler's proclamation setting out ten provisions of sharīʿa for his subjects.¹³

That is the shape of what arrived: not primarily a set of beliefs but a portable legal and administrative technology. Contract law enforceable in ports two thousand miles apart. A notarial script. A lunar calendar that made the Friday assembly and the fast a shared timetable from Aden to Ternate. And a definition of the legal person — the Muslim — that carried both a protection and its shadow.

The sultan replaces the god-king

The deepest institutional change was to the theory of rule. A Majapahit king was a manifestation of divinity who would be installed as a deity in his own temple. A sultan is none of those things. He is a Muslim, subject to a law he did not make, whose legitimacy runs through the khuṭba said in his name at Friday prayer and, ideally, through a document from Mecca or a genealogy reaching the Prophet. Ternate is the clearest case: Sultan Zainal Abidin, who took the throne about 1486, was the first ruler of the clove island to abandon the title kolano for sultan, restructured the government on Islamic lines and created a council of clerics, the jolabe, to advise him on religious matters. Tidore and Bacan followed.

The Javanese did the same thing in a more Javanese way. The north-coast lords who converted through the fifteenth century — Tomé Pires met their successors in 1513 and wrote them up as merchant princes of recent and often foreign descent — kept the ceremonial apparatus of the courts they had left and grafted Islamic legitimacy onto it.²² What went out of use was the specific machinery of divine kingship: the funerary temple, the deified ancestor, the dharmādhyakṣa priesthoods that had drawn state revenue. In their place came the mosque, the Friday sermon and a scholar class.

Photograph by Ihsan Ariswanto. Menara Kudus (minaret of the Al-Aqsa Mosque), Kudus, Central Java. CC0 via Wikimedia Commons. · CC0

The mosque that is a temple

The synthesis is visible in brick. At Kudus on the north coast stands a mosque whose minaret is not a minaret in any Middle Eastern sense: it is a red-brick tower built in the split-gate candi idiom of Majapahit, with the mouldings and proportions of a Hindu-Javanese temple gateway, serving as the place from which the call to prayer is given. The Great Mosque of Demak — seat of the sultanate that would finish Majapahit — carries a tiered pyramidal roof descended directly from the Javanese meru, not a dome. The Javanese did not import a mosque. They rebuilt one out of the parts they had.

M. C. Ricklefs has given the resulting formation its standard name and its standard account. "The purpose of this book," he writes at the opening of Mystic Synthesis in Java, "is to present the history of how a synthesis of being Javanese and being Muslim was achieved by the early nineteenth century."¹ The synthesis rested on three commitments held simultaneously: a firm Islamic identity, observance of the five pillars, and acceptance of the reality of a local spiritual universe of Javanese deities and spirits. Sufism supplied the vocabulary in which the third could be reconciled with the first two. This is why the Javanese conversion looks so unlike the conversions in Iran or North Africa, and why the standard Indonesian story of Islamisation is a story of accommodation.

That story has a documentary problem, and Ricklefs states it plainly. The Wali Sanga — the nine saints credited with converting Java by adapting wayang shadow theatre, gamelan and the Javanese ceremonial calendar to Islamic ends — "survive only in manuscripts from the eighteenth and nineteenth centuries and can provide little help in matters as subtle as evolving cultural identities in the sixteenth century."¹ They are not evidence of how Java converted. They are evidence of how Java, three centuries later, chose to remember it: as a negotiation, conducted by saints, in which nothing Javanese had to be given up. The historical record is less generous.

What stopped being made

Set against the things that arrived, here is what went out of production in Java between roughly 1450 and 1600:

• Monumental temple building. Candi Sukuh, on the slope of Mount Lawu, carries a chronogram on its western gate reading gapura buta abara wong, which resolves to 1359 Śaka — 1437 CE. Sukuh and its neighbour Cetho are the last significant Hindu temples raised in Java, and they are already anomalies: crude, terraced, more Polynesian than Indic in feel, built high on a mountainside as the lowland order was going. After them, for four and a half centuries, no more candi.
• The kakawin in Java. The Old Javanese literary corpus — the Rāmāyaṇa, the Arjunawiwāha, the Sutasoma, the Deśawarṇana itself — stopped being composed on the island and largely stopped being copied there. What survives, survives because Balinese scriptoria kept copying it on palm leaf for the next four hundred years. P. J. Zoetmulder's survey of the corpus rests almost entirely on Balinese manuscript traditions; the Deśawarṇana itself was recovered in 1894 from the sack of the palace at Cakranagara on Lombok.⁴³
• The state clergies. The paired Śaiva and Buddhist superintendencies, and the tax-immune estates that funded them, disappear from the record.
• Deified royal burial. The last Javanese king installed as a god in his own temple predates the sultanates. Sultans are buried in the ground, facing Mecca, under a stone.

Denys Lombard's three-volume reading of Java as a carrefour — a crossroads where Indic, Islamic, Chinese and European strata lie on top of one another rather than replacing one another — is the best corrective to reading this list as simple erasure.²⁴ Much of the Indic stratum went underground rather than away: into the pesantren, the residential religious school whose form owes a good deal to the Hindu-Buddhist forest hermitage; into the calendar; into kejawèn mysticism; into the shadow-puppet repertoire, where the Mahābhārata is still performed in a country of 240 million Muslims. But going underground is not the same as continuing. The people who had held the offices lost them.

Women, and what the settlement cost them

The change here was slow, uneven and real. Islamic law brought a coherent body of provision on marriage, divorce, inheritance and testimony, and it did not map onto a society in which women held property bilaterally, ran the markets and divorced at will. Barbara Watson Andaya's study of the early modern centuries places Islam alongside Christianity, Theravada Buddhism and Confucianism as one of four scriptural traditions arriving in the same period, each carrying models of female modesty and submission that the region had not previously had in written, authoritative form.¹⁸ The written form is the point. Custom can be argued with; a text cannot.

The most exact illustration is also the latest. Between 1641 and 1699 the sultanate of Aceh — the most powerful Muslim state in the archipelago — was ruled by four women in succession, beginning with Taj ul-ʿAlam Ṣafiyyat al-Dīn Shāh, who reigned for thirty-four years. Sher Banu Khan's study of the period shows female rule legitimated both by adat, indigenous custom, and by Islamic argument, and shows the sultanahs governing effectively with the support of the ʿulamāʾ and the merchants.¹⁹ From 1683, Acehnese factions began putting the question of female rule to the scholars of the Ḥaramayn. In 1699 a fatwa, said to come from Mecca, declared that a woman's rule was contrary to the sharīʿa. The fourth sultanah, Zainat al-Dīn Kamalat Shāh, was deposed and replaced by an Arab sayyid, Badr ul-ʿAlam Sharīf Hāshim. No woman has ruled Aceh since.

What the cost was

1478 and 1527: the wars that finished it

The conversion of the archipelago was not accomplished by conquest. Its consolidation partly was.

Demak, a north-coast port with a Muslim ruling house, repudiated its tributary obligation to the Majapahit court around 1478 — the year Javanese tradition marks with the chronogram sirna ilang kertaning bhumi, "vanished and gone is the prosperity of the earth," yielding 1400 Śaka. Through the 1520s, under Sultan Trenggana, Demak conducted a series of campaigns against the surviving inland Javanese courts. In 1527 the last of them, at Daha in the Kediri region, fell. In the same year Demak's forces took the Sundanese port of Sunda Kelapa on the western end of Java, renaming it Jayakarta.

The record must be precise about what this does and does not prove. Majapahit was already in an advanced state of collapse before Demak lifted a finger: the Paregreg succession war of 1404–06 had split the dynasty and burned out a generation of the elite; Malacca had captured the Straits trade after 1400; the north-coast harbours had been draining revenue from the interior for a century.²⁵ Islam did not destroy Majapahit. What a religious frontier supplied was language and finality — a way for a coastal war of secession to be understood by its participants as something other than treason, and a reason not to reinstate what had been overthrown. The temples were not rebuilt.

Photograph by Hastosuprayogo. Masjid Agung Demak, Demak, Central Java. CC BY-SA 4.0 via Wikimedia Commons. · CC BY-SA 4.0

Conversion at the point of a sword

The accommodationist account of Javanese Islamisation is broadly right about the fourteenth and fifteenth centuries and misleading about the sixteenth and seventeenth. Ricklefs, whose whole book is an argument for synthesis, does not hide the other mechanism. Quoting the Dutch envoy Rijklof van Goens on the founder of the Mataram dynasty, he records that Senapati, having taken a district, "immediately filled with great cruelty all the royal family and their retainers... He thereupon introduced his newly acquired religion."¹

That is one sentence about one campaign, and it should not be inflated into a general picture. But it is the general picture's necessary correction. Java's Islamisation ran on two tracks. On the coast, over two centuries, it was commerce, marriage and gradual accommodation. Inland, in the sixteenth and seventeenth centuries, it was frequently annexation by Muslim states that arrived with their religion in the baggage train. Both are the transmission. Only one of them is usually told.

The category that made people property

The doctrinal architecture that came with the merchants included a proposition with commercial consequences: a Muslim may not lawfully be enslaved by another Muslim, and an unbeliever outside the dār al-Islām may be. In a society that already practised bondage extensively — and the archipelago did, on a large scale, before Islam arrived — this did not create slavery. It re-categorised it. Debt and capture had been the criteria. Now belief was one too, and belief mapped onto geography: converted coast, unconverted interior and eastern islands.

Ibn Baṭṭūṭa's Samudra was already operating this way in 1345, when he recorded the sultan's subjects taking pleasure in warring for the faith against the infidels of their vicinity.⁹ Over the following centuries the pattern held across the archipelago: Muslim coastal states raiding non-Muslim uplands and outer islands for captives, with the religious status of the victims functioning as the licence. Reid's collaborative survey of bondage in the region documents the resulting structures, and also warns against importing plantation-slavery assumptions: Southeast Asian bondage tended toward household incorporation and debt servitude rather than gang labour, and the largest and most violent raiding complexes — the Sulu and Iranun fleets of the eighteenth and nineteenth centuries, which carried off tens of thousands — were driven by European demand for marine and forest produce, not by anything in the fourteenth-century transmission.¹⁶

The honest accounting is therefore narrow and still damning. The transmission did not invent enslavement in the archipelago. It supplied a durable religious criterion for who was eligible, and that criterion outlived every polity that first applied it.

Cloves, and the states the cloves built

The eastern end of the story is commercial. Cloves grew, in the fifteenth century, on five small islands off Halmahera and nowhere else on earth; nutmeg and mace grew in the Banda group and nowhere else. Islam reached Maluku roughly eighty to ninety years before Ternate's ruler took the title of sultan in the 1480s, carried by Malay and Javanese merchants, and it reached it because the spice was there.¹⁶ The conversion of Ternate, Tidore and Bacan turned scattered clove-growing chiefdoms into sultanates capable of taxing, arming and warring — and, within a generation of Zainal Abidin's death, of being fought over by the Portuguese and the Spanish, and then destroyed by the Dutch. The Banda islanders, who had converted in the same period, were largely exterminated by Jan Pieterszoon Coen's forces in 1621. That atrocity belongs to the Dutch East India Company and not to this transmission; it is noted here only because the political units the Company annihilated were the ones this transmission had built.

What the transmission does not owe

Three costs commonly attached to this record belong elsewhere, and the atlas will not charge them here.

• The fall of Majapahit was over-determined by dynastic war and the loss of the Straits trade. Demak finished a corpse.
• The Sulu and Iranun raiding economy of the eighteenth and nineteenth centuries was a response to European commercial demand and Spanish pressure, three to four hundred years downstream.
• The destruction of the Bandanese and the Moluccan spice wars were European acts.

What the transmission does owe is what happened at the moment of exchange and in its direct institutional wake: the end of a millennium-old temple civilisation in Java, the loss of two state priesthoods and the literary tradition they curated, the slow legal demotion of women in the most gender-egalitarian pre-modern society in the literate world, the conversion of the coast-interior distinction into a doctrinal frontier, and the establishment of belief as a criterion of enslavability in a region that had not previously used one.

The bill

Set against that: an alphabet in which two national languages are written, a legal order that let strangers contract across an ocean, a literate scholarly tradition that produced a body of Malay and Javanese religious writing of real originality, a shared calendar and ritual life running from Aceh to Mindanao, and the religion now professed by roughly 240 million people in Indonesia, 20 million in Malaysia and several million more across the southern Philippines and Thailand — the largest Muslim population on earth, assembled without a single caliphal army ever reaching it.

The temptation with this record is to write it as the good transmission: the one that proves religion can travel without conquest. It mostly did. Merchants, marriage and monsoon really are the mechanism, and the Javanese really did keep the shadow puppets, the gamelan, the calendar and a large part of the metaphysics. But the reason a peaceful mechanism is worth studying is not that it is free. It is that its costs are deferred, distributed and hard to attribute — paid three generations later by people who never met a Gujarati merchant, in the form of temples that were never rebuilt, a literature that had to emigrate to survive, a queen deposed by a letter from Mecca, and a line on a map that told one village it could sell the next one.