Indonesian is the clean case of something the rest of Aksarium only glimpses: a language whose writing system was changed by decree, twice, within living memory, and whose spelling was settled by treaty with a neighbouring country. It has been written in three unrelated scripts and, in the last of them, in four different orthographies since 1901. The second half of the page is about the neighbour: Indonesian and Malaysian are one language administered by two states, and the distance between them was made in offices rather than by geography. The keyboard sits in the gallery; the wider politics of script reform belong to Codex Lexpressia, and this page is the ground both stand on.
Three scripts, and four spellings of the third
Malay has been written down for about thirteen hundred years and has changed alphabet twice in that time. Neither change was internal. An Indic script arrived with Hinduism and Buddhism, the Arabic alphabet arrived with Islam, and the Latin one arrived with European administration; the language received every writing system it has ever used. What makes Indonesian unusual is not the borrowing, which is ordinary, but what happened after the last borrowing: four separate Latin orthographies inside a single century, each of them designed for a different reader, and the last two imposed by government decree.
The oldest surviving specimen of Malay is a stone found in November 1920 at Kedukan Bukit, on the banks of the Tatang River in South Sumatra, and now held in the National Museum in Jakarta. It dates itself to 1 May 683 CE. The language is Old Malay and the script is Pallava, an Indic system carried into maritime Southeast Asia along the same trade routes as the religions; the text records a Srivijayan expedition, and its concerns are military and administrative rather than literary.
Pallava and its descendant Kawi seeded a whole family of archipelago scripts, Javanese and Balinese and Batak and Lontara among them, most of which are now ceremonial where they survive at all. For Malay itself the Indic layer is the earliest of three, and the only one nobody proposes to bring back.
Islam brought the Arabic alphabet, adapted for Malay as Jawi. The adaptation is the familiar one: existing Arabic skeletons re-dotted to carry sounds Arabic does not have, so the inventory grows by a handful of letters without inventing a single new shape. The earliest widely cited Malay text in Arabic letters is the Terengganu Inscription Stone, from the east coast of the peninsula, usually placed in the early fourteenth century; the reading of its date has been contested since the stone was first published, and several different years have been proposed from the damaged text.
Jawi was not marginal. It was the ordinary written form of Malay for centuries, the script of the manuscript tradition and of correspondence between courts, and in Java the closely related Pegon convention carried Javanese and Sundanese in the same letters. It remains official in Brunei and in use in Kelantan and Terengganu. In Indonesia it now belongs mostly to religious education, manuscript scholarship and signage that wants to look old.
Latin script, Rumi, came with the Portuguese and the Dutch, and the Dutch turned it into policy. Charles Adriaan van Ophuijsen had been a school inspector in Bukittinggi in West Sumatra before taking the chair of Malay at Leiden, and in 1901 he published Kitab Logat Malajoe, followed by a grammar, Maleische Spraakkunst, in 1910. He worked with two Malay assistants, Engku Nawawi and Moehammad Taib Soetan Ibrahim, whose names appear on the work and rarely anywhere else.
The system spelled Malay through Dutch habits. It used ⟨oe⟩ for /u/, ⟨dj⟩ and ⟨tj⟩ for the affricates, ⟨j⟩ for the glide English writes as ⟨y⟩, ⟨nj⟩ and ⟨sj⟩ for the palatals, ⟨ch⟩ for /x/, an apostrophe for the glottal stop and a diaeresis to keep adjacent vowels apart. Each choice makes the language easier to pronounce if you already read Dutch, and marginally harder if you do not. That was the design brief: the orthography was built for colonial officials, and it served them well for forty-six years.
The Republican spelling was introduced on 19 March 1947 and is named for Raden Soewandi, then Minister of Learning. It cut the most conspicuously Dutch feature first: ⟨oe⟩ became ⟨u⟩, so goeroe became guru. Apostrophes for hamzah went, replaced by ⟨k⟩ in final position, so ta' became tak and ma'lum became maklum. Reduplication could be abbreviated with a superscript numeral, ubur2 for ubur-ubur. The distinction between ⟨e⟩ and ⟨é⟩ was abolished, which removed a diacritic and lost a vowel contrast in the process.
But ⟨dj⟩, ⟨tj⟩ and ⟨ch⟩ survived, and so the language still looked Dutch on the page. That is the sentence the module turns on. A country that declared independence in 1945 went on spelling its national language in the orthographic conventions of the colonial administration for another twenty-five years, and the man who eventually signed the reform away had himself been elected under the old spelling.
Ejaan yang Disempurnakan, the perfected spelling, was announced on 16 August 1972 and took effect from the beginning of 1973. The digraphs went in a single stroke, the letters ⟨f⟩, ⟨v⟩ and ⟨z⟩ were formally admitted so that loanwords no longer had to be disguised, the prefixes di- and ke- were separated in writing from the identical prepositions, and reduplication went back to being written out in full.
Djakarta became Jakarta. Soeharto, who signed the decree, stayed Soeharto, and Soekarno stayed Soekarno. Personal names were left alone, which is why Indonesian surnames still carry an orthography the language abandoned half a century ago, and why a Dutch digraph survives most visibly on the people who ended Dutch rule.
The standard has been revised quietly since: ministerial updates in 1987 and 2009, a 2015 edition that renamed it PUEBI and added the ⟨ei⟩ diphthong, and a fifth edition on 16 August 2022 that restored the name EYD, added ⟨eu⟩ as a monophthong and reaffirmed the optional ⟨ê⟩. None of these is a reform in the 1972 sense. They are the ordinary maintenance of a codified standard, which is itself the point: Indonesian now has an institution that does this, and the institution is the durable outcome of the upheaval.
The same language, two standards
Indonesian and Malaysian are one language in the sense that matters to a linguist and two languages in the sense that matters to a publisher. They did not drift apart across distance; the speech communities are separated by a strait people have crossed daily for a thousand years. They were pushed apart by administration. Two colonial powers supplied two technical vocabularies, two independent states built two language agencies, and the agencies then spent fifty years trying to close a gap their own predecessors had opened.
The clearest split is in the borrowed layer, and it maps exactly onto the colonial boundary. Indonesian took its modern administrative and technical vocabulary through Dutch; Malaysian took its through English. Television is the standard demonstration: Indonesian televisi, from Dutch televisie, against Malaysian televisyen, from English. Office is kantor in Indonesian, from Dutch kantoor, against Malaysian pejabat, which is native Malay put to a new job.
The pattern matters more than any individual pair. Both standards were building vocabulary for the same new things at the same time, and each reached for the nearest available donor. Where Indonesian calqued or borrowed from Dutch, Malaysian borrowed from English or coined from Malay roots, and the result is two lexicons that agree almost completely on the old words and diverge on the ones invented after about 1850.
The awkward cases are not the words one standard lacks. They are the words both have and use differently, because those fail silently: nothing signals to the reader that anything has gone wrong.
The spelling reform in the first half of this page was not an Indonesian decision. It was announced on the same day in two capitals, 16 August 1972, by Suharto in Jakarta and Tun Abdul Razak in Kuala Lumpur, and each side gave something up to get it. Indonesia surrendered ⟨tj⟩ for ⟨c⟩ and ⟨ch⟩ for ⟨kh⟩, and took over Malaysia's ⟨j⟩, ⟨y⟩ and ⟨ny⟩. Malaysia surrendered its own ⟨ch⟩, which had stood for a different sound entirely, and accepted ⟨kh⟩ in its place. Dewan Bahasa dan Pustaka ran classes on the Malaysian side and allowed a five-year grace period.
Read against the decade before it, this is a striking thing to have happened. The two countries had been in armed confrontation as recently as 1966. Indonesian sources record an earlier joint spelling, Ejaan Melindo, drawn up in 1959 and never brought into force; within four years the two countries were in armed confrontation. What survived 1972 was not only the orthography but the machinery: MBIM, the Majlis Bahasa Indonesia-Malaysia, was founded on 29 December 1972, Brunei joined on 4 November 1985 and the body became MABBIM, and Singapore attends as an observer. Its work is coordinating new terminology across three national standards, which is the unglamorous half of language planning and the half that decides whether the standards keep converging.
Far enough to be a working problem, not far enough to be a translation problem. An intelligibility test in 1998 had Malaysian linguistics students read Indonesian newspapers, and found that odd, unintelligible or unusual items came to about 30 per cent of the whole. That is a high number for a shared language and a low one for two languages.
Pronunciation adds distance the spelling hides. Malaysian speech on the Johor-Riau standard reduces final vowels, so a final ⟨a⟩ surfaces as schwa; Indonesian keeps a full /a/ and is spoken close to the way it is written, in a clipped and even rhythm that learners find unusually tractable. Two people can read the same sentence off the same page and sound as though they are doing different things, which is roughly the position, and the 1972 reform did nothing about it because spelling reform never can.