Études du Corpus des inscriptions du Campā, XIII
Documenting and Describing a Language on the Basis of an Epigraphic Corpus: The Case of Old Cham (Present Vietnam, 500–1500 ce)
p. 21-89
Résumés
The Chamic languages, today represented only by small or very small minority communities of speakers on the Chinese island of Hainan, in Cambodia and in Vietnam, form a low-order branch of the Austronesian language family. One of the Chamic languages is Cham itself, with pockets of speakers (several 100,000 in total) in some provinces of Southern Vietnam as well as Cambodia, and a bifurcation between at least two clearly distinct dialects. It is the only Chamic language with a premodern writing tradition, but one that goes as far back as the 5th century ce. While Modern Cham has been documented based on manuscripts (none earlier than the 17th century) and by consultation with native speakers, older stages in the development of Cham are preserved in an epigraphic corpus, associated with a kingdom called Campā, that covers a millennium (ca. 500–1500) and a North-South stretch over at least 1000 km of what is today Vietnam. These premodern phases of Cham linguistic history, that we may call Old Cham, have drawn virtually no attention among linguists, and have not begun to be described linguistically by epigraphists. The road towards a grammar and dictionary of Old Cham has to pass through the visual documentation of the text-bearing inscriptions (estampages, photographs), and the decipherment of the writing system. Besides the challenges inherent in relying on inscriptional data (damage to artefacts leading to textual uncertainty, constraints on the type of linguistic expression found in inscriptions, among others), the linguistic analysis of Old Cham needs to be sensitive to the likelihood of chronological and geographic variation in a small corpus covering a lot of time and space. To access Old Cham, moreover, one cannot do without a knowledge of Sanskrit (which was used in Campā for epigraphic expression besides Old Cham), closely related non-Chamic languages that have a relatively well documented history (Malay, Acehnese), the unwritten other Chamic languages (Jarai, Raglai, etc.), and an understanding of the areal pressures that have pushed the Chamic languages rather far from the typological mold of the Austronesian languages of maritime Southeast Asia. This article, conceived as the prelude to grammatical and lexicographical research to follow, presents the Old Cham corpus from the artefactual and textual points of view, outlines the history of discovery and documentation of Old Cham, sketches the challenges that face scholarship in this field, and illustrates some preliminary results of typological and historical interest with special attention to the relevance of comparison with Malay and Acehnese.
Les langues chamiques, uniquement représentées aujourd’hui par de petites ou de très petites communautés minoritaires de locuteurs sur l’île chinoise de Hainan, au Cambodge et au Vietnam, constituent une sous-branche de la famille des langues austronésiennes. L’une de ces langues chamiques est le cham lui-même, dont les communautés de locuteurs (certaines comptant près de 100 000 personnes) vivent dans des provinces du sud du Vietnam ainsi qu’au Cambodge, et qui présente une bifurcation entre au moins deux dialectes clairement distincts. C’est la seule langue chamique qui possède une tradition d’écriture prémoderne, remontant au ve siècle de notre ère. Alors que le cham moderne a été documenté sur la base de manuscrits (dont aucun n’est antérieur au xviie siècle) et via la consultation de locuteurs natifs, les étapes plus anciennes du développement du cham sont préservées dans un corpus épigraphique, associé à un royaume appelé Campā, qui couvre un millénaire (vers 500-1500) et une étendue nord-sud sur au moins 1 000 km de ce qui est aujourd’hui le Vietnam. Ces phases prémodernes de l’histoire linguistique cham, que nous pouvons appeler le vieux cham, n’ont pratiquement pas attiré l’attention des linguistes et n’ont pas fait l’objet de descriptions linguistiques de la part d’épigraphistes. L’établissement d’une grammaire et d’un dictionnaire du vieux cham passe par la documentation visuelle des inscriptions textuelles (estampages, photographies) et le déchiffrement du système d’écriture. Outre les défis inhérents à l’utilisation de données épigraphiques (dommages causés aux artefacts entraînant une incertitude textuelle, contraintes sur le type d’expression linguistique trouvée dans les inscriptions, entre autres), l’analyse linguistique du vieux cham doit tenir compte de possibles variations chronologiques et géographiques au sein d’un corpus limité couvrant un large empan temporel et spatial. Pour accéder au vieux cham, on ne peut, en outre, se passer d’une connaissance du sanskrit (qui a été utilisé dans le Campā pour l’expression épigraphique, à côté du vieux cham), des langues non chamiques proches, dont l’histoire est relativement bien documentée (malais, acihnais), des autres langues chamiques non écrites (jaraï, raglaï, etc.), ainsi que d’une compréhension des contraintes aréales qui ont poussé les langues chamiques assez loin du berceau typologique des langues austronésiennes de l’Asie du Sud-Est maritime. Cet article, conçu comme un prélude aux recherches grammaticales et lexicographiques à venir, présente le corpus du vieux cham du point de vue des artefacts et des textes, esquisse l’histoire de la découverte et de la documentation du vieux cham, les défis qui se posent aux chercheurs dans ce domaine, et illustre quelques résultats préliminaires d’intérêt typologique et historique, avec une attention particulière portée à l’intérêt d’une comparaison avec le malais et l’acihnais.
Entrées d’index
Mots-clés : Aymonier (Étienne), proto-chamique, malais, acihnais, histoire intellectuelle, épigraphie, estampages
Keywords : Aymonier (Étienne), proto-chamic, malay, acehnese, intellectual history, epigraphy, estampages
Note de l’auteur
This article is a result of the project DHARMA ‘The Domestication of “Hindu” Asceticism and the Religious Making of South and Southeast Asia’. The project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement no 809994). I gratefully acknowledge feedback received on earlier drafts from various colleagues: Sander Adelaar, Michel Antelme, Marc Brunelle, Chhom Kunthea, Jim Collins, Mark Durie, Tom Hoogervorst, Salomé Pichon and Alice Vittrant. I am especially grateful to Émilie Aussant for having invited me to give the lecture on which this article is based, and for her support during the months it took me to write it up.
Texte intégral
1. The Modern Linguistic Profile of Mainland and Maritime Southeast Asia
1The languages of Southeast Asia belong to five major linguistic families. Among these, the Austronesian languages form a largely insular family, stretching from Madagascar to Easter Island, the demographically most significant languages in Indonesia and the Philippines being among them, notably Malay (the matrix of the national languages Indonesian and Malaysian), Javanese and Tagalog (the matrix of Filipino) (fig. 1).
Fig. 1. The Southeast Asian language phyla (simplified).

Reproduced from Brunelle & Kirby 2016.
2The Chamic languages are a low-order branch of the Austronesian family and the only branch whose members are almost exclusively represented on the Southeast Asian mainland. They are today represented mainly in Vietnam, with a few Chamic languages also spoken in neighboring Cambodia and on the Chinese island of Hainan. But they have a particularly close relationship with Acehnese and Malay, with which they form the higher-order subbranch called Malayo-Chamic that is itself part of the Malayo-Polynesian branch of the Austronesian family. Although there is some scholarly debate about the higher-order subgrouping of the Malayo-Polynesian languages, about the classification of Malayic varieties (Adelaar 2005, 2008) and about the relationship between Acehnese and the Chamic languages, all scholars agree that the Malayic and Chamic languages, as well as Acehnese, go back to a common ancestor (fig. 2).
Fig. 2. The main language families referred to in this study and their genetic connections, after Brunelle 2019, no. 2.

3In the following pages, I will concentrate on the documentation and description of the only Chamic language that has a written tradition of significant time-depth. The language in question is Cham itself, whose two distinct varieties, that linguists today refer to under the labels Western and Eastern Cham, are associated respectively with speech communities in Cambodia and Vietnam (fig. 3). The location of Eastern Cham most probably represents the homeland from which migration has taken place probably no more than a few centuries ago, leading to the presence of a Cham-speaking community in mostly Khmer-speaking Cambodia. I will take the angle of the history and methodology of research in this field and will limit myself to the documentation of the premodern stages of the development of Cham. In other words, I will be focusing on Old Cham.
Fig. 3. Ethnolinguistic map of French Indochina (1949).

The pockets of Cham speakers are indicated in dark green, other Chamic languages with horizontal green stripes.
Source : Bibliothèque nationale de France. https://gallica.bnf.fr/ark:/12148/btv1b53197047g.
4This article is conceived as the first of what I hope will become at least a pair of studies dedicated to the Old Cham language. The present one furnishes the historical and epistemological background to linguistic analysis proper (grammatical description and dictionary), research on which is ongoing. I try to address a variety of potential readerships at once: descriptive and historical linguists, epigraphists, historians of linguistics, historians of Southeast Asia, scholars of Campā. This exercise forces me to make compromises with the conventions of their respective disciplines and occasionally to test the patience of some readers for the benefit of others.
2. Historically Documented Languages of Southeast Asia
5Most of the languages of Southeast Asia have a relatively shallow historical record. Consequently, although quite a bit has been written on the comparison and reconstruction of Southeast Asian languages, the history of individual languages has not drawn systematic attention and no attempt has been made, that I am aware of, to write a historical grammar of any Southeast Asian language, in anything like the manner that some readers may be familiar with, for instance, for languages of Europe or India.1 The languages that do have a written tradition several centuries deep, including epigraphic evidence, of the type that yields data for analyzing their historical development, are (without attempting to be entirely exhaustive):
- Burmese, belonging to the Sino-Tibetan language family;
- Siamese, Thai Lanna, and Lao, of the Tai-Kadai family;
- Vietnamese, of the Austroasiatic family;
- Mon, Austroasiatic;
- Khmer, Austroasiatic;
- Balinese, of the Austronesian family;
- Malay, Austronesian;
- Javanese, Austronesian;
- Sundanese, Austronesian;
- Cham, Austronesian.
6Of these, Mon, Khmer, Malay, Javanese and Cham have a documented history stretching over more than a millennium, while documentation for the other languages is a few or even several centuries less deep. Mon and Khmer are related Austroasiatic languages and benefit can be had in studying their respective histories by comparison with each other and with other, historically less well documented Austroasiatic languages. (Vietnamese is too distant a relative of Mon and Khmer, and has such a different history, that it is less relevant in this connection.) Analogously, Balinese, Malay, Javanese and Cham are all Austronesian languages of the Malayo-Polynesian branch and their histories too can be better understood by comparison among them, as well as with other, historically less well documented Austronesian languages. And among all these historically documented languages, Cham holds the doubly honorable position of being not only the oldest attested Austronesian language, but also the oldest attested of all languages of Southeast Asia. Its oldest attestation takes the form of an inscription on stone — one of the items (number C. 174, see below) in the corpus of inscriptions of the kingdom of Campā that formerly occupied much of what is today Vietnam.
3. The kingdom of Campā and its epigraphic corpus
7The history of the kingdom of Campā is known from external sources, such as Chinese and Vietnamese annals, or mentions of Campā in Cambodian inscriptions, but the most abundant source material is found in the inscriptions of Campā itself. The epigraphic corpus contains texts that are written in the Sanskrit language, whose use in early Southeast Asia as a cosmopolitan language of cultural and political expression has been quite elaborately discussed as a historical phenomenon and need not detain us much here (Pollock 1996, 2006). What is important is that adoption of this language, and the Indian syllabic writing system serving to transcribe it, became the vehicle for a vernacular language of Campā, namely Cham, to gain a written life of its own, and over time to become the exclusive language of written expression in Campā. Being thus composed in Sanskrit or in Cham, the inscriptions of Campā are mostly engraved on stone supports, such as stelae and doorjambs, and more rarely on metal artefacts. They are very often internally dated with precision, or datable on paleographic grounds, and their geographical provenance is in most cases known with precision too — potentially important facts for historically sensitive linguistic analysis. By way of illustration, let us look at a map showing the territories and some of the major archaeological sites of Campā (fig. 4).
Fig. 4. The Territories and Major Archaeological Sites of Campā.

Map by Federico Barocco, reproduced from Griffiths, Hardy & Wade 2019 (map 1).
8I will cite inscriptions by their “C.” inventory number.2 My first example is C. 1, a 15th-century inscription, entirely composed in Cham, engraved on the back-slab of a sculpture found in the 19th century near a Pagoda at Biên Hòa, just north of Saigon (fig. 5). This is the southernmost location at which a Campā inscription has been found. One of the northernmost locations is Hà Trung, in Quảng Trị province, where the inscription C. 113 was engraved in the 10th-century on a large pillar, still in situ today (fig. 6). This text, running over four faces of the pillar, contains parts in Sanskrit and parts in Cham, alas all severely weather-beaten. C. 40, the oldest inscription of Campā, and of Southeast Asia as a whole, was composed entirely in Sanskrit and found in Khánh Hòa province (fig. 7). It does not contain an internal date but on the ground of paleographic comparison with Indian inscriptions is thought to date to the 3rd century ce. Recourse to distant comparison with Indian epigraphic material is necessary because no other Southeast Asian inscription shows script as archaic as this one.3 C. 41, part of a small group of inscriptions that show only slightly less archaic script and are thought to date to the 4th or 5th century, was found in Phú Yên (fig. 8). And it is to this group of 4th/5th-century texts that the oldest Cham inscription belongs. It bears number C. 174, and we shall return to it in §11.4.
Fig. 5. Inscription C. 1, engraved on a stele at Bửu Sơn pagoda in Biên Hòa province. Photographs Arlo Griffiths, 2013.

(a) The stele viewed from the side. (b) Close-up of the inscription on the back-slab.
Fig. 6. Inscription C. 113, on a pillar at Hà Trung in Quảng Trị province. Photographs Arlo Griffiths, 2015.

(a) General view of the site: the pillar bearing C. 113 is the one on the right. (b) Close-up of one of the four faces of C. 113.
Fig. 7. Inscription C. 40 at the National Museum of Vietnamese History, in Hanoi.

Photograph Adeline Levivier, 2023.
Fig. 8. Rockwall with inscription C. 41, below Tháp Nhạn temple in Tuy Hòa, Phú Yên province.

Photograph Federico Barocco, 2011.
9The “C.” inventory of inscriptions was begun over a century ago by French scholars, but new inscriptions are still discovered quite regularly in Vietnam, so that the corpus of Campā inscriptions today covers nearly 300 inventory items, ranging in length from single words to multiple pages in romanized transcription, though by far the majority fits on a single printed page.
10From this very rapid overview, let us retain, first, that the oldest and the latest inscriptions that I have shown are separated by a millennium in time and that the northernmost and southernmost sites are about a thousand kilometers removed from each other as the crow flies. This means that we’re dealing with a textual corpus that covers a very substantial amount of time and space. Let us also retain, second, that a significant percentage of the epigraphic corpus is made up by Sanskrit text. For comparison, the Cambodian epigraphic corpus, which is in many ways closely comparable to the one of Campā, covers over 1500 items, with Sanskrit and Old Khmer being the main languages of expression, while inscriptions from the Indonesian archipelago, in Sanskrit or vernacular languages, number well over 3000 items. Finally, let us retain that the geographic extent of the Campā epigraphic corpus covers much of the present East-West extent of Chamic languages in Vietnam (with a total of around 250,000 speakers by recent estimates), and more than their present North-South extent (if we exclude Tsat on the island of Hainan).
Fig. 9. Diagrams showing hypotheses on the relationship of the Proto-Chamic dialect chain to its modern reflexes.

Diagram (a) is reproduced from Thurgood 1999: 42 (fig. 7); (b) and (c) offer two explicit hypotheses on the position of Old Cham while showing only a selection of living Chamic languages.
11Faced with the Old Cham corpus, historical linguists need to determine how the language reflected in the corpus relates to other Chamic languages (fig. 9). Had the ancestors of other Chamic languages already separated from Cham at the time that Cham began to be written? Or can Old Cham in some ways be considered as equivalent to Proto-Chamic, or to an intermediate stage between Proto-Chamic and Proto-Cham? If Old Cham cannot be considered the direct ancestor of all extant Chamic languages, then it must also be determined what was the sociolinguistic status of Old Cham in the various parts of the extensive area where it was used for writing inscriptions: was it used for writing in areas where the spoken language was some other Chamic or even non-Chamic language?4
4. The Post-Campā Textual Corpus and Modern Cham
12Before turning to the history of the discovery of the Campā epigraphic corpus, and of the study of Old Cham, let me briefly point out that the written documentation in the indigenous writing system does not stop with the demise of an independent Campā state in the course of the 15th century, but continues in the modern period over the 16th century down to the 20th. The principal written sources are literary, documentary and ritual texts written in Cham script on paper or palm leaf, though a small number of stone inscriptions from the post-Campā period also exist.5 All the materials from the post-Campā period show forms of Cham script that belong to the varieties of actively used script documented in manuscripts from the late 19th century onward, and are distinct even from the 15th-century script forms found in inscriptions to the extent that those who are literate in Modern Cham are unable to read inscriptions from the 15th century, or earlier, without some training, while those, like me, who can read the script found in inscriptions, cannot simply pick up a Cham manuscript and read it. Besides this difference of script, there also seems to be linguistic difference between the Old Cham found even in the latest Campā inscriptions, and the Modern Cham found in the post-Campā textual corpus. These are no more than impressionistic statements, however, since no attempt has been made by any scholar, that I am aware of, to define the difference between Old Cham and the earliest varieties of Modern Cham from a linguistic point of view, while the historical development of Cham script, from the ancient to the modern period, has likewise yet to be described in detail.6
5. The discovery of Cham epigraphy
13Let us now look at another colonial-period map (fig. 10). The Second French Empire annexed Cochinchina and established a protectorate in Cambodia in 1862 and 1863, respectively. After the French Third Republic had taken over northern Vietnam — i.e., the protectorates of Tonkin and Annam — through the Tonkin campaign, the various protectorates were consolidated into one union in 1887.7 The discovery of the early epigraphic tradition of Cambodia, and of Campā epigraphy in what were then the territories of Cochinchina and Annam, took place in a whirlwind succession of scholarly events during that 25-year period. Thanks to the exploratory fervor of a few gentlemen, scholarship quickly emerged from a state of profound ignorance as to the early political and linguistic history of these countries.
Fig. 10. Map showing French Indochina (1860–1869) with its constituent territories Cambodia, Cochinchina, Annam and Tonkin.

Source : Bibliothèque nationale de France. https://gallica.bnf.fr/ark:/12148/btv1b530814228.
14To illustrate the state of ignorance of primary sources which reigned until the early 1880s, and the role of missionary scholars in the discovery of Cham epigraphy, let us read what the Abbé Lesserteur, who was settled in northern Vietnam, wrote in 1882:8
If we peruse the old historical annals of the Annamite nation, we often find mention of a people who bordered it to the south and inhabited the country that now corresponds to the portion of Annam named Cochinchina by Europeans. This country was then called Qhiem-Thagne by the Annamites, from which the name Cochinchine was later derived when they had taken control of it. Qhiem-Thagne, written in Chinese characters, is actually pronounced Chin-Chine in the Mandarin dialect. — In the earliest reports published in French, this country was known as Ciampa.
Gradually driven out by the Annamite race, very invasive by nature, the Ciampese took refuge in the south, eventually losing their nationhood towards the beginning of the 18th century. But their race and language did not disappear with the collapse of the nation. While in the northern and central provinces of Cochinchina, the defeated merged in large numbers with the invading people, the Ciampese race remained more or less intact in the south.
This is why today, descendants of this people can still be found under the name of Qhiames (Cham), in the province of Bình Thuận, and in the neighboring mountains, among the half-wild tribes who have withdrawn there like the Jaraïs; in some parts of French Cochinchina; and above all in Cambodia, where they enjoy a certain degree of autonomy, in comfort and prosperity.
What we know of the history of the Ciampese people is very incomplete. In the 18th century, the famous Father Gaubil gave a brief outline, based on Chinese historians. This sketch, entitled Notice historique sur la Cochinchine, first appeared in Lettres édifiantes et curieuses, and was later included in volume XII of Father de Mailla’s Histoire générale de la Chine. Recently, Mr. Bouillevaux published a new study on Ciampa in the Annales de l’Extrême-Orient, drawn mainly from the historical annals of Annam. These are the only historical documents we have.
We know even less about the literature and language of the Ciampese people than we do about their history. The first and only work published so far on this subject is of very recent date: it is due to Mr. Aymonier and appeared in the tenth issue of Excursions et Reconnaissances, towards the end of last year.
I thought it would please you, gentlemen, to offer the Société académique Indo-chinoise some very old inscriptions which I believe belong to this language. If they cannot help to shed some light on any point of history, at least they may be of use in the study of Ciampese literature.9
15Writing in 1882 to present some graffiti found in the far northern province of Quảng Bình that he correctly surmised to be connected to Campā,10 although he was unable to decipher them, Lesserteur mentioned the publication, just one year prior, of a Cham inscription found far to the south and west of Quảng Bình in Cambodian territory. This item (C. 2), the first Cham inscription and probably also the first specimen of Cham language ever published, had been brought to light by Étienne Aymonier, who was to become the central figure in the discovery of Campā epigraphy.11 In the relevant article, the scholar opined that this epigraphic document “must date back to ancient times”. He added:
The Chams of Cambodia have long since lost the habit of inscribing their impressions on stone. The characters and the language indicate a distant era. (Aymonier 1881: 173)
16He could not know it then, but in reality, the inscription with which he inaugurated his work in Cham epigraphy was not that ancient, definitely post-Campā, say around the 17th century. Despite this error of judgment, Aymonier’s pioneering publication makes some linguistic observations that were spot on:
A large number of words shared with Malay, Javanese and the other languages of the archipelago, indicate the close kinship of the Cham dialects with all the idioms of Malaya, disregarding, of course, the terms that the Bani dialect may have borrowed from Malay over the last few centuries. Kinship terms, color names, numeration, etc., are almost identical. In addition, we quote almost at random: minhum, to drink; malam, night; apoey, fire; jalan, way, road; hujan, rain; thau, to know; abih, all; basey, iron; bley, to buy, from which Cham has made pabley, to sell (literally to make buy); tangan (pron. tangoeun), hand; tama (tamau), to enter; tubéak, to leave; main, to play; dalam, inside, in; kayau, wood; barau, recent, new, etc., etc. (Aymonier 1881: 182)
17For anyone who knows Malay, these Cham words are indeed all immediately recognizable. This proximity with Malay is an important point to which I will return.
18But let us first learn more about Étienne Aymonier.12 This military man turned scholar had arrived in Saigon in 1869 and, soon after his arrival, had the chance to learn Khmer while posted among speakers of this language south of the city. His command of the language led to teaching appointments at the Collège des administrateurs stagiaires and, by 1879, to an appointment as representative of the French government in Cambodia. Let me quote from the obituary by George Cœdès, which sketches the important period that was now about to begin.
On January 1, 1879, Aymonier was appointed acting representative of the Protectorate in Cambodia, replacing Moura, who returned to France for good: he was confirmed in this position two years later, and held it until May 10, 1881. In the meantime, he had been promoted to the rank of captain. During his stay in Cambodia as head of the local government, Aymonier made a series of tours that revealed the country’s archaeological wealth. In August 1879, he followed the arm of the lake [Tonle Sap, AG] to Pam Čhkot; in March 1880, he travelled up the Mekong to Hàn Čei; finally, in March 1881, he traversed the entire region of central Cambodia between his two previous itineraries. Cambodian epigraphy was then on the agenda: Kern had just deciphered for the first time fragments of Sanskrit inscriptions brought back by Harmand. Using his knowledge of the language, Aymonier undertook the translation of several modern Khmer inscriptions, and read for the first time the numerical dates of some ancient inscriptions. These early epigraphic works, published in Excursions et Reconnaissances and his travel reports, attracted the attention of Indologists, for whom Aymonier, upon his return to France in 1881, brought back a series of estampages of Sanskrit inscriptions. (Cœdès 1929: 543)
19These promising first results then led to a pathbreaking three-year survey undertaken in the years 1881 to 1884 at French government expense. Let us read a few passages from the fascinating account of his survey and field methods that the pioneer himself furnished (Aymonier 1892).
In December 1881, the Ministry of Education entrusted me with a scientific mission to research the inscriptions and monuments left in Indo-China by two peoples of Indian civilization: the Khmers or Cambodians and the Tchames or Ciampese. (pp. 216–217)
[…], I return to November 1884, at Saigon, when […] I was busy preparing my own trip to Annam. I intended first to spend several months in the province of Bình Thuận, adjacent to French Cochinchina and the southernmost province of Annam. […] In order to tackle the decipherment of the inscriptions in the vernacular language of ancient Champa, which more or less occupied present-day Annam, I had to return to the study of the dialect of the Cambodian Chams, which I had completely neglected for several years; I also had to study the dialect of the Annamese Chams who inhabit this province of Bình Thuận, where on the other hand, inscriptions and monuments are plentiful: I knew this vaguely from information I had obtained in advance. (p. 355)
I set about studying the native language and manuscripts, while I had my Chams from Cambodia crisscross the valley in order to collect manuscripts and preliminary information on the ruins and inscriptions, as well as on the curious customs of these brothers separated from them for long generations. (p. 360)
I led a very secluded life, but the Asians around me, and also my neighbor the missionary, repeatedly warned me that tempers were getting hot, a situation that was sufficiently explained by the general situation in Annam. For a long time, the Hué court had been preparing, almost ostentatiously, to make a last effort to extricate itself from French domination. […] In the midst of these incendiary elements, the safety of Father Villaume and myself, the only Europeans in this remote province, at the mercy of a trivial incident, could be dangerous from one moment to the next. […] In order to double our forces by uniting them, the missionary offered me hospitality in his main Christian settlement, two leagues from my village. But I didn’t want to appear to be abandoning the Chams with whom I had settled, who were worried by all these rumors and who secretly implored me not to leave them. (p. 361)
Before leaving the province, I had recourse to [the kindness of the governor of the Annam empire] to send to Saigon, by land, with the necessary precautions, a large box of estampages of the inscriptions found at Phanrang. I also had to cause him an infinitely greater nuisance. I mentioned earlier that on my way to Annam, I had hired five Cambodian Chams for my stay in the province of Bình Thuận, where they were to help me make good contact with their brothers in that province. As a result, I had tried to recruit in Bình Thuận a new staff of Chams who would travel with me to the other provinces of Annam; I was beginning to express myself sufficiently in their dialect, and they all offered the incomparable advantage of being fluent in the Annamite language. (p. 365)
Left almost alone, waiting for the local Chams, I wandered through the Parik valley. I was also visited by the two missionaries I had already seen in Phanrang: Father Villaume, from Bình Thuận, and Father Auger, from Khánh Hòa. The latter offered to accompany me to his province, which I gladly accepted, and it was agreed that I would pick him up on my way to Phanrang, where the gentlemen were returning immediately. (p. 366)
Then, still accompanied by the good Father Auger, I went back down the river in a pirogue to return to my luggage, to take care of making estampages of the rich inscriptions on the Pô Nagar monument, training for this work my new staff, which consisted of ten Chams recruited from Bình Thuận. (p. 368)
I had not foreseen the events of Hué on July 5, which spontaneously set fire to the whole of Annam […] Hard as it was to accept it, I had to admit that almost the whole of Indo-China was in revolt. There was no safe haven in Cambodia, Annam or Tonkin. The days of peaceful exploration were well and truly over […] From then on, all I thought about was going home and getting some rest. Taking the steamer again, I went to Saigon where I entrusted my Cham staff to my friend Mr. Landes, who gradually had them repatriated, and I embarked for France at the beginning of September 1885 […].
I deeply regretted not having been able to complete the exploration of this last country. Preliminary information taken in advance and from afar, following my usual practice, indicated Cham inscriptions in Quảng Nam, near Tourane; in the vicinity of Hué; at the Troc caves on the Sông Gianh, Quảng Bình province. Some of these inscriptions have since been seen by Europeans, but no estampages have been made of them yet. My regret was tempered by the satisfaction of having sent from all the countries I had surveyed, but especially from Cambodia, the estampages of hundreds and hundreds of epigraphic documents, some in Sanskrit, the religious and learned language of the ancient Khmers and Chams, others in the vernacular Khmer or Cham depending on the country. (pp. 372–373)
20We learn here how Aymonier gained an active command of Cham by immersion into the life of the Cham community in Bình Thuận, how he depended for his fieldwork on the involvement of two teams of Cham assistants, and how political conditions imposed a limit on the number of Campā inscriptions that he was able to document.13
6. Documentation of Campā Inscriptions Through Estampages
21The passages quoted from Aymonier’s account also make clear how crucial was, in his methodology, the documentation of the engraved texts by estampage, a French term which designates both a process and resulting paper artefact — it is also used in English though perhaps more commonly recognized English equivalents are rubbing or squeeze.14 The history of estampage-making in Southeast Asian epigraphy still needs to be written. Suffice it to say that the estampages made, as per Aymonier’s account, by his Cham collaborators from Cambodia or Bình Thuận, were despatched to Paris and are still accessible today in the Bibliothèque nationale and the library of the Société Asiatique. Many of the bundles of estampages have manifestly not been opened for a very long time. Since 2020, my collaborator Adeline Levivier is engaged in the painstaking process of photographing the Bibliothèque nationale collection, one example of which will be given below (C. 53, §11.2). These 19th-century documents are in some cases the only physical trace of texts inscribed on artefacts that have been lost since. It also happens that they preserve the texts in a state much superior to that of the stone artefacts still to be found in Vietnam, when the latter have suffered damage since Aymonier’s time.
22I may mention that I have myself continued in Aymonier’s footsteps by undertaking estampage-making campaigns, and that, like Aymonier, I do so in the company of highly skilled Cambodian collaborators. The technique we use, however, with ink on paper, is not the same as Aymonier’s but rather the one originally developed in China that was adopted in this field a few decades after Aymonier’s campaign in the 1880s (fig. 11).
Fig. 11. Ham Seihasarann and Khom Sreymom making an inked estampage (EFEO n. 2089) of inscription C. 80 at Mỹ Sơn.

Photograph Arlo Griffiths, 2011.
7. Aymonier’s Decipherment and Interpretation of Cham inscriptions
23Having been forced to abort his campaign, Aymonier returned to France and some years later had sufficiently advanced in his study of the collected materials to contribute a long article, entitled “Première étude sur les inscriptions tchames [First study regarding the Cham inscriptions]”, in the 1891 volume of the Journal asiatique. It is with this study, appearing ten years after the discovery of the epigraphic corpus of Campā, that the philological study of the Old Cham corpus is inaugurated. From Aymonier’s introduction to this study, let me quote only the following paragraph which again illustrates his methodology:
From the philological point of view, my studies on the ancient vernacular languages of the two peoples of Indian civilization of the Trans-Gangetic peninsula, i.e. the present work on epigraphic Cham, and that which I published in 1883 on Old Khmer, these studies, I would say, have been carried out according to the same method, resulting from a more or less thorough knowledge of modern idioms and a very summary comparison, carried out at first glance, as it were, of these idioms with the languages of the inscriptions. This method, while giving us new and precise information on these ancient languages, inevitably leaves many obscure points and unexplained words. But the future will allow us to refine and deepen these studies. To this end, I have in my possession the necessary documents and materials: inscriptions, manuscripts that have been examined or read, annotated, two voluminous Khmer and Cham dictionaries in preparation — in brief, all the elements that will make it possible to search meticulously for traces, affinities and the certain, or at least probable, nature of the many words that Khmer and Cham have lost or disfigured over the centuries. (Aymonier 1891: 8)
24The brief introduction moves on to elucidate various aspects of the way the Indic writing system is used and adapted for transcribing Old Cham (a subject that I will, in turn, attempt to elucidate below, in §13) but does not discuss any issues of grammatical description. It is assumed in Aymonier’s translations that the language of the old inscriptions shared such basic syntactic features as modified-modifier and subject-verb-object word order and the same (limited) number of prefixes and infixes as the modern dialects.15 Within that implicit syntactic and morphological framework, if a lexical word or a function word is recognized as equivalent to a word in Modern Cham, as known to Aymonier, he interprets it accordingly. This approach could only have taken him as far as the continuities between the different historical varieties of Cham allow. Aymonier was aware of the limitations of the method and hoped to offer improved interpretations of the inscription at a later stage — a hope that never materialized in print. It is striking that Aymonier does not allude, in his list of “necessary documents and materials”, to the potential interest of linguistic comparison with languages of maritime Southeast Asia to determine the meanings of words lost or unrecognizably transformed in Modern Cham. And this despite the fact that he had already recognized the close proximity of Cham with Malay in his 1881 study quoted previously.
8. The Aymonier & Cabaton Dictionnaire and the Role of the EFEO
25In the decade after the publication of his first study of Cham inscriptions, Aymonier seems to have focused on other projects, notably the compilation of the Cham dictionary that he mentioned having in preparation. He was joined in this endeavor by a remarkable young scholar, Antoine Cabaton, who started his career with research on Cham before taking up a teaching assignment at Langues O’ that led him to focus on Malay. Some of his early publications (1901, 1904) form a kind of bridge to the new phase of Campā studies which was ushered in by the creation, at the turn of the 20th century, of a new institution, the École française d’Extrême-Orient. Let us read a little about this scholar from a retrospective dedicated to him in 1983:
In 1898, he came to the attention of Louis Finot, who was in the process of setting up the École française d’Extrême-Orient and asked him to go on a field trip to Indochina. He accepted enthusiastically and spent two years “in the field”, devoted in particular to a vast survey of the Chams of Annam and Cambodia. […] Unfortunately, he fell seriously ill and had to rush back to France in early 1901. He then resumed his service at the Bibliothèque Nationale, worked on the casting of various typefaces (Khmer, Cham, Bengali) at the Imprimerie Nationale, and wrote his first publications on Champa: “Rapport sur les littératures cambodgienne et chame” (C. R. de l’Académie des Inscriptions, 1901), Nouvelles recherches sur les Chams (Publ. EFEO, 1901), “L’inscription chame de Bien Hoa” (BEFEO IV, 1904), and above all the famous Dictionnaire čam-français written in collaboration with Aymonier (Publ. EFEO, 1906). (Lombard 1983: 20)
26In the preface to their Dictionnaire (p. V), Aymonier & Cabaton state among the principal aims which the dictionary was to serve that it should “facilitate the reading of the inscriptions”.16 In the list of abbreviations (p. XLVI), the authors introduce the label “I.” for epigraphic data to be cited. But this label is used only rarely.17 Although it is undeniably of some use for interpreting inscriptions, it cannot be said that the Dictionnaire adequately covers the lexicon of Old Cham. The several Old Cham inscriptions published in the first half-decade of the 20th century (Finot 1903, 1904a, 1904b; Cabaton 1904) perhaps appeared too late to be excerpted for the Dictionnaire. But even for the inscriptions that had been published in Aymonier’s foundational study of 1891, based on the numerous estampages from Southern Campā that the pioneer had been able to collect by 1885, no attempt is made to cover their vocabulary in the Dictionnaire by explicit indications that a given word is attested epigraphically. And a large amount of Old Cham lexical material that does figure in the 1891 article is not recorded in the Dictionnaire at all.
27While both authors were back in France and occupied with other projects, after the publication of the Dictionnaire, the exploration of Campā sites in the field and the documentation of new inscriptions through estampages, with double copies sent to Paris, that had been continued in the 1890s by a cohort of explorers that I only mention in passing,18 was now taken up on a systematic scale by archaeologists and epigraphists working under the auspices of the EFEO. And it is especially thanks to the work of epigraphists Louis Finot, Edouard Huber and George Cœdès that the textual corpus swelled considerably during the first decades of the 20th century. It should be noted, however, that while these scholars generally offered integral Romanized transcriptions of the epigraphic texts, and translated Sanskrit (parts of) inscriptions integrally, they often gave only fragmentary translations or no translations at all for the Old Cham textual material. Their understanding of Old Cham grammar and lexicon was fundamentally insufficient to be able to offer coherent translations of whole texts.19
28Meanwhile, it seems that Aymonier continued working with the estampages collected by himself and other scholars, available to him in Paris. With one minor exception (Aymonier 1911), none of this post-Dictionnaire epigraphic work has ever been published, but an interfoliated copy of the Dictionnaire that is preserved at the Société Asiatique includes many hundreds of manual additions in red ink elegantly reproducing citations from inscriptions in transcription and in original script (fig. 12). I must admit that a lot is still unclear to me about the chronology of Aymonier’s work and the reason why so few epigraphic data were included in the Dictionnaire.20
Fig. 12. Interfoliated copy of Aymonier & Cabaton 1906.

Société Asiatique, Cote B 40 (formerly 4° collection 93, 7).
29While the EFEO produced a steady stream of editions of newly discovered Campā inscriptions between 1904 and 1915, progress became much slower after that, and ground to a complete halt after 1939, when George Cœdès published his article entitled “La plus ancienne inscription en langue chame [The oldest inscription in Cham language]”. By this time, the textual corpus had grown to several dozens of pages of transcribed text, but probably less than half of it had been translated, while very little explicit lexicographical and virtually no explicit grammatical commentary had been published to accompany the texts.
30For the next seventy years, there was not much significant progress in Campā epigraphy at all, and what little there was had no linguistic ambitions. The state of affairs today, in 2024, more than 130 years after Aymonier published his “First study regarding the Cham inscriptions”, is that we still have no grammatical and no lexicographical study of the Old Cham language.
9. Post-WWII American Linguists
31This state of affairs is all the more remarkable as the post-WWII decades saw a flurry of interest in Cham and other (non-written) Chamic languages from the part of American linguists, directly or indirectly associated with the Summer Institute of Linguistics (SIL). Thus, Ernest Wilson Lee defended a thesis at Indiana University in 1966 on a historical topic, namely “Proto-Chamic phonologic word and vocabulary”. But this work is only one of the first of a surprisingly abundant series of contributions to Chamic historical linguistics that were produced by Lee himself and some other American scholars over the second half of the 20th century.21 This line of scholarship generally engages only superficially with the results of French predecessors. Let us read Lee (1966):
I do not know the date of the oldest written records in Cham, but according to Georges Maspero the inscriptions of the 11th century are not greatly different from the present Cham language. The value of the Cham language and its sister languages in the overall picture of Austronesian historical linguistics leads me to reconstruct as much as possible of the vocabulary of the parent language before it diverged to form the modern Chamic languages. (p. 2)
E. Aymonier and A. Cabaton have compiled a large 586-page Dictionnaire Cham-Français. The entries are in the Devanagari script (sic!) with a transliteration and French glosses. The words are alphabetized in the traditional order for Devanagari scripts (sic, plural!). The materials in this dictionary have not been used except occasionally for corroborative evidence. There are three reasons for this: (1) the size of the dictionary precluded a careful use of it in the time available, (2) there are many variants for large numbers of the morphemes with no reference as to the date of the materials from which they were taken, and (3) the inadequacy of the script at some points.
In reference to the second point above, Aymonier and Cabaton include all dialects of Cham, but where there is a difference between Cambodian Cham and Vietnamese Cham, it is usually indicated. There is, however, no indication as to whether the written forms represent the way they were written by a contemporary scribe or whether taken from some older material. As Doris Blood has pointed out, the orthography, in general, represents a much older stage than current spoken Cham. (pp. 13–14)
32Lee, who had obviously not looked at any published inscriptions himself, alludes to Georges Maspero, who was a Sinologist and historian rather than a scholar of Old Cham.22 Lee is evidently ignorant of the distinction between Devanagari and Cham scripts, and of George Cœdès’s publication of an Old Cham inscription dating to the 4th or 5th century (1939), which would have helped him situate the earliest written Cham record (namely C. 174) in time. He refers to Doris Blood to make the point that the orthography in the Dictionnaire is very archaic compared to the spoken language of the 1960s.23 Since the purpose of Lee’s research was to reconstruct Proto-Chamic, deciding not to make use of written evidence that is acknowledged to “represent a much older stage than current spoken Cham” is methodologically hard to justify.
33Whatever be the precise reasons for the lack of engagement with the data contained in the Dictionnaire, the facts that it covers Old Cham so poorly, and that epigraphists had advanced so little in interpreting Old Cham texts, will not have helped to encourage Lee and other linguists to make use of the early written data in their publications from the 1960s to the end of the 20th century.24 We see Lee’s work still serving as model even by the turn of the century, in a seminal monograph published by Graham Thurgood (1999). Despite its title, From Ancient Cham to Modern Dialects: Two Thousand Years of Language Contact and Change, this book has hardly anything to say about the ancient language of the inscriptions and makes virtually no use of epigraphic data.25
34But — to his credit — Thurgood was aware of the potential significance of epigraphic data for linguistic reconstruction:
Finally, it is expected that the forms reconstructed in this work for Proto-Chamic (henceforth PC) will be revised in the direction of the forms found reconstructed for Proto-Malayo-Chamic, as better and more thorough use is made of the Written Chamic inscriptional data. (Thurgood 1999: x)
However, there is one potentially invaluable source that remains only partially tapped—the earlier records of inscriptional Chamic. The more that we know about the earlier inscriptions, the further back that we will be able to push the PC reconstructions. The value of these older inscriptions is greatly increased by the fact that early Chamic itself was a dialect continuum—a string of related and interacting dialects along the coast of Vietnam. As a consequence, there is every reason to believe that there were Post-PC changes that affected all the members of the dialect continuum, with the result that the evidence of an earlier stage was lost everywhere. Thus, when we reconstruct PC, in reality, what we are reconstructing is not the earliest stage of Chamic but a somewhat later stage. What evidence we have of the earlier stages is, in part, provided by Acehnese, which dropped out of the dialect chain, when it left the mainland, thus saving evidence of earlier stages, and, in part, by the inscriptional evidence, which should fill in some of the gaps. (Thurgood 1999: 9)
35To give one example, it indeed seems that inscriptional evidence may require reconsideration of the chronology of some sound changes. For instance, Thurgood thought that Austronesian high vowels in final open syllables had become diphthongized before Proto-Cham, but juxtaposition of Old Cham, Malay and Acehnese cognates suggests that diphthongization had not advanced as far as Thurgood believed by the Proto-Chamic stage (fig. 13).26
Fig. 13. Chamic Diphthongization chains in comparison with Malay and Acehnese.

36The reconstructions postulated by Thurgood in his Appendix appear to be in contradiction to his own observation of evidence from written Cham, known to him from the Aymonier & Cabaton Dictionnaire, that historical forms of Cham reflect earlier stages of the diphthongization chain than -ɛj and -ɔw, such as the forms found in the epigraphic corpus shown in my diagram.27 In fairness, the Cham epigraphic corpus shows spellings or spelling variations that reflect later stages in the diphthongization chain, such as tavov ‘sugarcane’ (Malay tebu, Acehnese teubèe, Thurgood *tabɔw, Shorto 2006: 98, no. 125) and huriy/hrey ‘sun, day’ (Malay hari, Acehnese uroe, Thurgood *hurɛy) as well as asov besides asuv. But even if the words with i and u in the nucleus are considered historical spellings, their very existence would still seem to imply a prior (but necessarily post-Proto-chamic) stage where diphthongization had not yet occurred, or was only incipient, depending on the phonological interpretation of the spellings with -y or -v at word-end (see §13). As we will see below, there are also several lexical items missing from Thurgood’s list of reconstructions but attested in inscriptional Old Cham that can be matched with Malay and/or Acehnese and thereby proven to have existed in Proto-Chamic.
10. The Contribution of a British Scholar of Malay
37A promising exploration of the relevance of Malay lexical material to interpreting Old Cham texts was given by the British Indonesianist Geoffrey E. Marrison in his article “The early Cham language, and its relationship to Malay” (1975). In this contribution, he attempted to translate selected inscriptions in the light of his knowledge of Malay, and reached the following conclusion:
To the above samples of mediaeval Cham, further materials of a similar nature could be added from inscriptions up to the middle of the fifteenth century A.D. Though there are some changes in the language, these are less evident than the later changes evidenced in the Cham manuscripts of the nineteenth century, and the currently spoken language. Mediaeval Cham reveals its affiliation to Malay much more clearly, than does the later language, which has been considerably influenced, particularly in phonology and vocabulary by the neighbouring Vietnamese and Cambodian, as well as other non-Indonesian languages of Indo-China.
What is clear from the inscriptions is that the early Cham language has close connexions with Malay in phonology, vocabulary and syntax; but also from the earliest time, certain characteristics of a non-Indonesian type are in evidence. These include the throwing forward of the word accent to the last syllable, and a consequent tendency to monosyllabism, and the use of a system of affixation more akin to Mon-Khmer than to the Indonesian languages, in the use of prefixes and infixes but the rejection of suffixes.
A further feature of early and mediaeval Cham is the use made of Sanskritic vocabulary, covering a range of subject matter similar to that found in Old Malay and Old Javanese, and to that extent characteristic of a general tendency of borrowing in the Indonesian languages.
A special interest which attaches to mediaeval Cham is that in the field of the history of Indonesian languages they span a period for which there is no direct Malay evidence. The Sri Vijaya inscriptions belong to the latter part of the 7th century A.D. Thereafter little has survived till the thirteenth century, but the whole of this period is covered by the Cham material, which is thus coeval with Old and Early Middle Javanese. A detailed study of the Cham evidence should provide some advances in the historical and comparative linguistics of the Indonesian language family. (Marrison 1975: 58–59)
38Although he emphasizes the significance of his contribution to the study of Indonesian languages, Marrison effectively also demonstrates the reverse, namely the relevance of Indonesian linguistic data to the interpretation of Old Cham. Notably, he proposes several more or less convincing etymological connections that had not yet been noted by French scholars, among which I cite:
huvi ‘root, tuber’ | Malay ubi, Acehnese uboe, Mod. Ch. habei, Thurgood *hubɛj |
vuḥ ‘to offer’ | Malay bubuh ‘to place’, Acehnese bôh, not in Mod. Ch. or Thurgood |
vyau ‘banana’ | Balinese biyu ‘banana’, not in Mod. Ch. or Thurgood |
sumrāṅ ‘bank, shore’ | Malay seberang ‘further shore’, not in Mod. Ch. or Thurgood |
39I will briefly discuss the last word below, in §12.4.
11. Interpreting Texts: Determining Meanings of Words and Grammatical Rules
40Given the fact that we cannot apply the usual methods of descriptive linguistics, that rely on interaction with native speakers (elicitation, grammaticality judgments, etc.), the methodology for determining meanings of words and grammatical constructions needs to be philological, whereby I mean the interpretation of textual data through contextual analysis and corpus-internal comparison informed by linguistic comparison with related languages, particularly the other Chamic languages as well as Malay and Acehnese. Building on the work done by previous scholars and benefiting from the ever-increasing range of digitally available resources, it is now possible to take significant steps forward, as the following examples aim to illustrate. I have selected texts presenting relatively few unresolved challenges.28
11.1. Inscription C. 30 A3 — 1183 ce, Temple of Po Nagar at Nha Trang
41Translated without text edition by Aymonier (1891: 44–45), epigraphic edition in Schweyer 2005: 93–94. In the aforementioned interfoliated copy of the Dictionnaire, we find the following printed entry, one of the rare ones that make reference to epigraphic data, and further specifications in red ink on the facing page (fig. 14).
Fig. 14. Aymonier & Cabaton (1906), interfoliated copy at the Société Asiatique, p. 433 and facing page.

42We learn here several facts. In the face of the occurrence of two variant spellings of the word for ‘eel’ in the epigraphic corpus, namely lanuṅ besides landuṅ, the possible etymological connection with lendong/lendung/lindong (meaning ‘eel’) in Malay, as indicated in the Dictionnaire, suggests that the spelling landuṅ is probably the older one.29 It is this spelling that is attested (twice) in a passage represented only in part in the handwritten additions. The larger context is as follows (fig. 15):
Fig. 15. Two lines from inscription C. 30 A3 as reproduced in EFEO estampage n. 227.

… pom̃ parameśvarī vuḥ taliy tulāṅ· landuṅ· māḥ 2 thim̃ | pom̃ rayā vuḥ talim̃ tulāṅ· landuṅ· māḥ 2 thim̃ 8 draṁ | pu-nai bhāgyavatī rājaputtrī vuḥ 1 talim̃ kakaṁ canvak· māḥ | …
43In normalized spelling, this would correspond to:
44pov parama-īśvarī vuh taliy tulāṅ landuṅ māh 2 thiy. pov rayā vuh taliy tulāṅ landuṅ māh 2 thiy 8 dram. pu-nai bhāgyavatī rāja-putrī vuh 1 taliy kakam c-an-vak māh.
45Building on the partial translation “offrir un collier [d’]os d’anguille en or de 2 onces 5 drams [offer a gold eel bone necklace, 2 ounces 5 drams]” on the interfoliated page, I propose to gloss the whole quoted segment word-for-word and to translate it as follows:
lady chief-queen offer cord bone eel gold 2 tael. lady great offer cord bone eel gold 2 tael 8 dram. princess b. king-daughter offer 1 cord encrust (color/instr-wrap) gold.
The Lady Chief Queen (pov parameśvarī) offered a golden eel-bone necklace of 2 tael. The Great Lady (pov rayā) offered a golden eel-bone necklace of 2 tael, 8 dram. The princess (pu-nai) Bhāgyavatī, royal daughter, offered 1 golden color-encrusted necklace.
46The context is here that of a gift of jewelry in precious metal to be applied to a stone statue of a deity. Every subject is marked for status by composite status-marking expressions involving the elements pov and pu, about which I will have more to say in §12.5. In the first two sentences, the repeated sequence taliy tulāṅ landuṅ māh (word-for-word gloss ‘cord bone eel gold’) nicely illustrates how Old Cham forms complex noun phrases in the order modified-modifier and how sentence constituents are arranged in the order subject-verb-object. The interpretation is plausible in the light of extant examples of mobile jewelry as well as its representations in stone.30 By contrast, the interpretation of the third sentence is less certain, but kakam survives in Modern Cham in the meaning ‘encrust’ and could be a regular Cham cognate of kekam ‘scum, froth’ in Malay, while canvak is presumably the Old Cham antecedent of canvak or canuk in Modern Cham. We are dealing with -an- infixed forms connected with the base which is cuk/cvak in Modern Cham. The meaning ‘color’ recorded for canvak (canök in Aymonier & Cabaton’s transcription) in Modern Cham seems to be most fitting, although I do not understand how this meaning would have come to be connected with a noun derived from cvak ‘to tie’.31
11.2. Inscription C. 53 — 11th or 12th century ce, Bình Định province
47Translated without text edition by Aymonier (1891: 45). My epigraphic edition based on the BnF estampage (fig. 16):
Fig. 16. C. 53. Bibliothèque nationale de France, Inscriptions du Cambodge et du Campā, 422 (46).

Photograph Adeline Levivier, inverted horizontally.
48(1) [quatrefoil] yām̃ po ku śrī jaya Indravarmmadeva pu-cim̃ Anāk·
(2) Urām̃ _ _ mapura32 vuḥ ja hamuy· ṅan· sarvva(d)ravya ja hamu-
(3) (y)· di yām̃ pu nagara he(tu) ja hamuy· Adhama
49Normalized representation, word-for-word gloss and translation:
yāṅ po ku śrī jaya indravarmadeva pu-ciy anāk urāṅ paramapura vuh ja hamuy ṅan sarva-dravya ja hamuy di yāṅ pu nagara, hetu ja hamuy adhama…
god lord 1s ś. j. i. prince child person p. offer status h. and all-property status h. to god lord state cause status h. disobedient
My God Lord (yāṅ po ku) Śrī Jaya Indravarmadeva, the prince (pu-ciy) [who was] a child of the man of Paramapura offered ja Hamuy and all possessions of ja Hamuy to the Divine Lady (yāṅ pu) of the State (nagara) because ja Hamuy was disobedient.
50In this short passage — only the beginning of a text the rest of which is lost — we find not only further composite expressions for high status (yāṅ po ku designating the king,33 yāṅ pu designating a goddess) but also encounter an instance of how Old Cham can use personal articles to indicate the low social status of an individual, here with ja.34 In this case, the person named Hamuy is stripped of all his possessions (a concept expressed with the composite expression sarva-dravya, borrowed from Sanskrit) and given — as slave — to the temple as represented by the residing deity. The reason, as stated in a dependent clause opening with the Sanskrit noun hetu, here functioning as conjunction,35 is that said individual was adhama, a Sanskrit loanword used in Old Cham in the meaning ‘disobedient’.36 We encounter the use of ṅan (cf. Malay dengan) to coordinate the object noun phrases ja hamuy and sarvadravya ja hamuy, the latter noun phrase also illustrating the standard possessive construction by juxtaposition (possessed possessor). Finally, we see how the indirect object noun phrase (di yāṅ pu nagara) is marked with preposition di. The constituent order of the main clause is subject, verb, direct object, indirect object.
11.3. Inscription C. 208 — 1130 ce, Provenance Unknown
51Epigraphic edition, French translation and commentary in Griffiths et al. 2008–2009: 477. The text can be represented as follows in normalized spelling, word-for-word gloss and translation:
yāṅ po ku di yāṅ ciy pəṅ pu dhiḥ urāṅ rupaṅ anāk pu-lyaṅ śrī devarāja vuh asaṅ nī pak yāṅ parama-īśvarī-rūpa nī kāla śāka-rāja 1052
god lord 1s among god prince p. status d. person r. child minister ś. d. give bearer this to god chief-queen-statue this time śaka-king 1052
My God Lord among Gods, the Prince Pəṅ, Sire Dhih, man of Rupaṅ, son of the honorable minister Śrī Devarāja, offered this (incense) bearer to this statue of the chief-queen deity. The time was [the year of the] Śaka king 1052.
52This example of an inscription engraved on a metal artefact which must have been a ritual incense bearer illustrates that pak can be used instead of di to mark the indirect object, and that deictics are postposed in attributive position (here asaṅ nī and parama-īśvarī-rūpa nī). It also illustrates the absence — unsurprising for a Southeast Asian language — of any marking for number (yāṅ po ku di yāṅ).
Fig. 17. Inscription C. 174 as reproduced in EFEO estampage n. 1072.

11.4. Inscription C. 174 — 4th or 5th Century ce, Quảng Nam province.
53This is the oldest Cham text, first published by Cœdès (1939). A slightly revised interpretation was published by Marrison (1975). The original stone being lost, the inscription can fortunately still be read from an estampage (fig. 17). I represent it here in epigraphic edition, normalized spelling, word-for-word gloss, and translate it as follows:
54(1) siddhaM [spiral] nī yāṄ nāga puÑ putauV ya UrāṄ spūY di ko
(2) kuruN ko jmāY labuḥ nari svarggaḥ ya UrāṄ paribhū d(i) ko
(3) kuruN saribū thūN ko davaM di naraka dṅaN tijūḥ kulo ko
siddham. nī yāṅ nāga puñ pu-tauv. ya urāṅ spūy di ko, kurun ko jmāy labuh nari svargah. ya urāṅ paribhū di ko, kurun sa-ribū thūn ko davam di naraka dṅan tijūḥ kulo ko.
success this god serpent puñ king. rel person gentle to 3p kurun 3p neg.imp fall from heaven which person insult to 3p kurun one-thousand year 3p fever in hell with seven generations 3p.
Success! This is the divine serpent puñ the king. People who are gentle to Him (ko), kurun they (ko) shall not fall from heaven. People who insult Him, kurun a thousand years they suffer from fever in hell with their seven generations of descendants.
55The first sentence illustrates substantive use of deictic nī as well as the formation of equational sentences without any copula (nī yāṅ nāga puñ putauv). The word puñ, found in no other inscription than C. 174, is with kalañ in §11.5 below one of the very few words with word-final ñ, i.e. /ɲ/, in Old Cam.37 No really convincing interpretation has so far been found.38 Another word never figuring in other inscriptions is ko. Cœdès deemed the context to prove that it must be 3rd person pronominal element, comparable to iko in Minangkabau; Marrison added comparison with iku in Javanese. Although it does not seem entirely impossible to me to interpret the text under the alternative hypothesis of 2nd person reference (in which case ko could be cognate with Malay engkay/kau), I tentatively adopt the 3rd person interpretation, with the specification that ko seems to function both as honorific 3rd-person pronoun with singular reference and as unmarked 3rd plural pronoun.39
56Neither Cœdès nor Marrison recognized jmāy as the antecedent of the Modern Cham negative imperative word jvai. The /m/ in the Old Cham form probably requires reconsidering Thurgood’s reconstruction *juəy (1999: 321).40 Cœdès intuited that davam might be cognate with duam in the Chamic language Jarai and mean ‘fever’. It seems to have escaped his attention that Aymonier & Cabaton have recorded daum for Modern Cham itself (1906: 228) and that both of these Chamic words can plausibly be related to demam in Malay and deumam in Acehnese, although the correspondence of the v in the Old Cham word to the /m/ in Malay and Acehnese still needs to be explained.41 Attention may be drawn to the use of ya to form what seem to be relative clauses. Cœdès was unable to interpret kurun which Marrison connected with Modern Cham krön, ‘to look at, to know’, which he suspected to function as preposition ‘regarding’. This seems unconvincing to me. The structural parallelism between the phrases concerning gentle as opposed to hostile people suggests to me that kurun should either have modal value (‘may’), function as a conjunction expressing consequence (‘so that’) or be a deictic element opening the main clause. However, I have not yet found plausible cognates that would support any one of these hypotheses.42
57The fact that this text contains a few lexical items not found in the rest of the epigraphic corpus (puñ, ko, kurun) must be connected with the fact that it is a chronological outlier, several centuries older than the next oldest Old Cham inscriptions.
11.5. C. 122 (1064 ce), Bình Thuận or Ninh Thuận province
58Epigraphic edition and French translation in Cœdès 1912. My new edition based on an EFEO estampage (fig. 18), followed by normalized text, word-for-word gloss and translation:
Fig. 18. Inscription C. 122 as reproduced in EFEO estampage 781.

Photograph Adeline Levivier, inverted horizontally.
59(1) | di śakarāja 977 nan· kāla Im̃śvaramurtti si daḥ yām̃ po ku śrī parameśvaravarmmadeva santā-
(2) na Uroja ya cakravarttirāja di nagara campa nim̃ ra pratiṣṭhā yām̃ vihāra rājakula niy· mulaṅ·
(3) tra ra paralap· kalañ· ya ruṅ· nan· jeṅ· Avista bharuv· ra vuḥ Urām̃ maddan· lumvauv· kravāv· hu-
(4) mā ṅan· samasta Upakaraṇa tuy· devārccaṇa panūjā devatā nim̃ sadākāla nau |
di śaka-rāja 977 nan kāla īśvara-mūrti si dah yāṅ po ku śrī parameśvaravarmadeva santāna uroja ya cakravarti-rāja di nagara campa niy. ra pratiṣṭhā yāṅ vihāra rāja-kula niy mulaṅ. tra ra pa-ralap kalañ ya ruṅ nan jeṅ avista bharuv. ra vuh urāṅ maddan lumvauv kravāv humā ṅan samasta upakaraṇa tuy deva-arcana p-an-ūjā devatā niy sadā-kāla nau.
in śaka-king 977 that time īśvara-embodiment rel look! god lord 1s among god ś. p. lineage u. rel universal-king in state c. this. 3p found god monastery royal-family this again also 3p caus-shine tower rel empty that become all new 3p offer person with cow buffalo field that and all utensil follow god-cult instr-worship deity this all-time go
In [the year of the] Śaka-king 977: that was the time of the manifestation of Īśvara (i.e., Śiva), namely My God Lord Śrī Parameśvaradeva, of the lineage of Uroja, the universal king in this land of Campā, [that] he re-established the (statue of the) deity of this monastery of the royal family. Also, he rendered glittering that tower which was abandoned, in order that (jeṅ) everything be new. He gave personnel with cows, buffaloes, fields, and all utensils according to the cult of the gods, the worship of this deity, for all time to come.
60In this text, we see two affixed forms, pa-ralap and p-an-ūjā. The causative pa- prefix and instrumental -an- infix are well known throughout the Chamic family.43 The base ralap is unknown in Modern Cham and also has not survived, to my knowledge, in any other Chamic language. I attribute to it the meaning ‘to flash, to glitter’ based on the hypothesis that it is cognate with the words relap and gemerlap in Malay, which have such meanings. The noun p-an-ūjā is derived from the base pūjā, a Sanskrit loanword meaning ‘worship’. The compound expression panūjā devatā seems to convey the same meaning as devārcana (a Sanskrit endocentric or tatpuruṣa compound deva-arcaṇa meaning ‘worship of the god/gods’). Such redundant pairs of synonyms using Sanskrit and vernacular-language elements side-by-side are a stylistic device that is widespread in various ancient Austronesian languages from “Indianized” cultures, including Malay, and are also found elsewhere in Old Cham.44 Finally, we see here how two verbs, jeṅ ‘become’ and nau ‘go’, seem to be used in a grammaticalized manner that resembles closely what has been described for these verbs in Modern Cham. According to Moussay (2006: 238), jeṅ can be used in a multitude of constructions, one of them being a “coordinating conjunction introducing a logical sequence: consequently, that’s why”, and this seems to be fitting in the phrase jeṅ avista bharuv. Regarding postposed nau in Modern Cham, Moussay (2006: 258) states that it marks future or outward movement.
12. Further Discussion of Some Grammatical Topics
61In the preceding section, I have presented several short inscriptions and one segment of an inscription to illustrate some of the grammatical and lexicographic observations that can be made at this stage of research on Old Cham. Many other texts, especially longer ones, are rife with problems, although they always also contain sentences that can be satisfactorily interpreted. In what follows, I will make use of such shorter excerpts to highlight some further topics that are of particular interest for the synchronic or diachronic/comparative description of Old Cham grammar and lexicon, again highlighting the relevance of both Malay and Acehnese data for interpreting Old Cham.
12.1. Negation
62This is a topic in Chamic languages that has been relatively better explored than others in a comparative/typological framework, though attention has been focused mainly on bipartite negative constructions.45 Modern Chamic languages seem to show diachronic continuity in the usage of the negative imperative (prohibitive) word which is jmāy in Old Cham.46 The inscriptional data otherwise seem to represent a very different system of negation than what has been described for modern Chamic languages, the bipartite constructions combining oh with kā or buh being notable by their absence.47
63First, it is interesting to observe that Old Cham expresses ‘not yet’ with the single word vlom which, though absent from Thurgood’s reconstruction (1999), is easily interpretable as cognate with Malay belum. Two occurrences have been identified so far:
C. 89, A2–3 (Finot 1904b: 946–951)
hetu yāṅ po ku śrī jaya indravarmadeva nan vāla-svabhāva vlom kən pyan thuv hita-ahita-vastu
cause god lord 1s ś. j. i. that child-nature not-yet really accustomed know good-bad-thing
because My God Lord Śrī Jaya-Indravarmadeva, there, had the nature of a child, really not yet used to discerning the good and the bad things
C. 123 (Griffiths et al. 2008–2009: 470–472)
kāla drəṅ nāma pov pu-lyaṅ śrī yuva-rāja vlom rai mak panrāṅ tmuv
time reign name lord minister sir junior-king not-yet come take p. obtain
when he bore the name Lord of Ministers Sir Heir Apparent, had not yet (vlom) gone to take Panrāṅ for himself (? tmuv)
64It must be admitted that cognate words are not widespread at all, and it is hence possible that Old Cham had borrowed the word from Malay, although its origin in Malay would then still remain unexplained. If vlom and belum do both descend independently from Malayo-Chamic, then the word has apparently been lost in Chamic after Old Cham.
65Other negations are normally built with the word oh, as in the following two examples:
C. 167, D4 (Griffiths, Pichon & Southworth 2021–2022, app. 2)
oh jeṅ si top hulun dravya vihāra niy
neg become rel steal slave property monastery this
May no-one steal the slaves owned by this monastery.
C. 167, E5–6 (Griffiths, Pichon & Southworth 2021–2022, app. 2)
siy ya oh pa-matah proṅ bhoga-upabhoga si ma-tmuv va driy tra inā amā driy di svarga
rel rel neg caus-break big pleasure-enjoyment verb-meet uncle self also mother father self in heaven
He who does not break the extent (proṅ) of the means of subsistence,48 he meets his uncles (and aunts) and his mother and father in heaven.
66But there is an interesting exception:
C. 142, D12–13 (Huber 1911b: 285–298)
ya siy urāṅ nāma tan pa-matah ma-tandāh niy urāṅ nan nau dauk di svarga yathā-iṣṭa driy
rel rel person however not caus-break verb-annul this man that go stay in heaven as-wish self
The person who, however, does not break [or] annul this, that person will reside in heaven according to his own desires.
67The strict parallelism between the phrases oh pamatah and tan pamatah in the last two examples proves that tan could serve as equivalent to oh. No trace of a negation tan seems to survive in modern Chamic languages, and the word also does not exist in Malay, but it does in Acehnese as well as in Javanese. The combined evidence suggests that it may have been part of the Proto-Chamic negation repertoire, although it is also possible that it was borrowed from another source into both Acehnese and Old Cham.
12.2. Pronouns
68In the text passages presented so far, we have encountered the pronominal morphemes ku, ra and ko. While ku is solidly identifiable as 1st person singular pronoun, based on context and sound correspondence with 1st person singular morphemes in the other Chamic languages, its apparent correspondence of sound with freestanding aku besides clitic ku in Malay, whereas Acehnese has stressed kèe beside unstressed ku, may require revision of Thurgood’s (1999: 248) Proto-Chamic reconstruction of *kəu. The other two pronominal elements, ra and ko, do not find evident matches in the Chamic data assembled by Thurgood. While the explanation (offered above) of the latter in relation to deictics in some non-Chamic Austronesian languages remains somewhat unsatisfactory, the former is reminiscent of (si)da (3rd person plural) in Old Malay and seems to share with the latter an ability to serve as honorific with singular reference.49 The reminiscence is even stronger with sira (3p and honorific 3s) in Old Javanese. But in neither comparison is the sound correspondence regular. In Old Cham texts not reviewed here, we find the further pronouns kamiy (1st person plural exclusive) and ñū (3rd person singular), both of which are unsurprising in the light of other Chamic languages. An example of ñū may suffice:
C. 167, E4–5 (Griffiths, Pichon & Southworth 2021–2022, app. 2)
siy ya pa-matah, asov lūh, asov hitam pāt matā ñū ndoy inā ñū inā amā gəp gotra ñū lac dauk di avīcī annan naraka təl yuga-antah pralaya
rel rel caus-break dog wasted dog black four eye 3s mount mother 3s father group clan 3s fall stay in a. namely hell reach age-end conflagration
He who breaks [this]: a wasted (? lūh) dog, a four-eyed black dog mounts [his] mother and father [while] his group, [i.e.] family, falls to reside in the Avīci, that is (? annan) hell, until the end of the ages.50
69The attributive expression pāt matā ñū meaning ‘its eyes four’ is noteworthy.
70So far, a single instance has been identified of the interrogative pronoun haliy, that still exists in modern Chamic languages.51 It is in the short question humā haliy ‘which fields?’, preceding a list of field names (again in inscription C. 167, line C3). It seems that the somewhat more frequently occurring word vrəliy is related to this haliy, although I cannot explain the correspondence between the opening syllables.52 This vrəliy is so far attested only as a marker of indefiniteness, as in this example:
C. 106, A13–15 (Finot 1904a: 99–105)
siy vrəliy urāṅ yəṅ ma-rakṣā sarva-dravya yāṅ pu-pov ku śrī rudramadhyeśvara tra yāṅ pov ku śrī śivaliṅgeśvara tra urāṅ nan nau svarga maddan inā amā avis-kāla53
rel indef person rel verb-protect all-property god lord 1s ś. r. also god lord 1s ś. ś. also person that go heaven with mother father all time
Whichever person who protects all properties, both of My God Lord Śrī Rudramadhyeśvara and of My God Lord Śrī Śivaliṅgeśvara, that person goes to heaven together with his mother [and] father for all times.
71We have already seen above, in §12.1, several occurrences of the morpheme driy (which can also serve other grammatical functions, notably that of numeral classifier for humans) in reflexive use.
12.3. Relative-correlative clauses
72The examples quoted so far have shown varying versions of a relative-correlative construction:
73siy vrəliy urāṅ yəṅ … urāṅ nan …
siy ya … si …
ya siy urāṅ … urāṅ nan …
ya urāṅ … or ya urāṅ … kurun … (depending on the interpretation of kurun in C. 174)
74As noted above (§12.2), vrəliy seems to be an interrogative pronoun used to generate indefinite expressions. All the other elements (si, ya/yəṅ, nan and urāṅ) are familiar from related languages, but their precise functional space in Old Cham still needs to be determined. The element si, confusingly, does not match with the Modern Cham word si, but rather with sĕi.54 Although in Modern Cham, its syntactic function appears to be slightly different, it clearly serves to form relative clauses in Old Cham.
C. 8, C40–42 (Griffiths & Lepoutre 2016)
nī limān si yāṅ pov ku śrī jayasiṅhavarmadeva pu-ciy śrī harijit vuh dī yāṅ pu-pov ku
this elephant rel god lord 1s ś. j. prince ś. h. offer to god lord 1s
this is the elephant that My God Lord Śrī Jayasiṅhavarmadeva, prince Śri Harijit, offered to My God Lord (the deity of the temple)
12.4. Derivational affixes
75The distribution and function of five derivational affixes has been discussed for the Chamic languages as a group by Thurgood (1999: 238–246). The morphemes in question are tabulated in fig. 19.
Fig. 19. Old Cham derivational affixes, productive and fossilized.
pa- | causative | ||
-an- | instrumental | ma- | ? |
ta- | inadvertent | -um- | ? |
76Although pa-, ma- and -an- are all attested quite frequently,55 only pa- and -an- seem to have had a significant functional load in Old Cham, while ta- is too rarely attested so far to be meaningfully discussed here. According to Thurgood (1999: 241),
Aymonier and Cabaton (1906:364) describe the prefix mə- in Cham as a causative prefix, indicating a state, possession, reciprocity, or mutuality, further noting that it also occurs in Bahnar as mə-. They specifically equate it with Malay mě-/mĕng. An examination of the mə- section of their dictionary makes it clear that the prefix is essentially the same prefix as in Malay and that its use, although much diminished in modern Cham, was once far more productive.
77Use of the prefix ma- is frequent in the inscriptional corpus, but it seems to have been optional already in Old Cham. There are plenty of occurrences that could be used to support the idea that it is related to one of the verbal prefixes meN- or ber- in Malay. It must be noted, however, that Thurgood mentions only the former, while Blust (2013) distinguishes no less than three Austronesian prefixes some or all of which may have coalesced in Old Cham ma-.56 There are some double-affix forms like ma-pa-hudip ‘bring to life’ and ma-pa-knā ‘implement’ (see below) that can be compared with causative mem-per- forms in Malay. And there are some occurrences of pra- instead of pa- that also seem to compare nicely with the Malay per- prefix. Nevertheless, it is impossible to label Old Cham ma- as marker of dynamic or transitive verbs like meN- in Malay, because ma- is also used on stative/intransitive verbs. For this reason, I use the neutral gloss verb. The following passage illustrates several of the affixes mentioned so far, with two rather clearly stative/intransitive ma- forms:
C. 139, 5–8 (Huber 1911a: 10–13)
siy pu-pov tana rayā yaṅ ma-putau andap niy siy urāṅ rayā yaṅ ma-rayā andap niy pu-pov tana rayā ṅan urāṅ rayā ma-pa-knā tuy t-an-atap kuv niy pu-pov tana rayā nan ṅan urāṅ rayā nan tmuv puṇya di dalam rājya kuv tra di dalam rājya driy tra
rel lord earth great rel verb-king andap this rel person great rel verb-great andap this lord earth great and person great verb-caus-touch follow instr-fix 1s this lord earth great that and person great that obtain merit in inside kingdom 1s and in inside kingdom self and
Whatever sovereigns shall reign here (? andap niy),57 whatever dignitaries shall have dignity here — sovereigns and dignitaries implementing [the tax exemption] in conformity with this rule of mine — those kings and those dignitaries obtain merit (puṇya) in my kingdom and in their own kingdom.
78Finally, Thurgood (1999: 241–242) also discussed what he calls “the -əm- nominalizing infix” which, as he rightly observes, is inherited from the Austronesian verbal infix *-um-. The Old Cham shape of this morpheme, -um-, agrees precisely with the proto-form and it therefore seems necessary to reconstruct *-um- instead of *-əm- for Proto-Chamic. It also seems necessary to insist more on the verbal function of the infix, as opposed to Thurgood’s idea of nominalization.58 The forms so far identified in Old Cham are:
- s-um-rāṅ (variant s-m-rāṅ), ‘direction’? ‘bank, shore’? Marrison (1975: 56) recognized this as possible cognate with Malay seberang ‘opposite side’, and this seems attractive even though it must then be explained how Malay seberang with /b/ would have developed from a putative antecendent *sumaraŋ.59 Whether or not the Malay word is related, it is certain that sumrāṅ corresponds to Acehnese seumeureuëng /sɯmɯrɯǝŋ/.60
- k-um-vrāp (perhaps related to tumvrāk), whence k-an-umvrap — the meaning of these words is uncertain, and it therefore remains uncertain whether we are dealing with a cognate of one of the several similar-sounding Malay words (kerap, terap, or kedap).
12.5. The Morpheme pu
79As most Old Cham texts were issued by the king and concerned with his entourage, the lexical data for this semantic domain are relatively rich. Most of the text selections above have contained one or more instances of terms involving the morphemes pu or pov (sometimes spelled po).
pu-tau | king — as observed by Marrison (1975: 53), the morpheme tau is probably related to similar words in other Malayo-Polynesian languages, such as Tagalog tao ‘man’,61 but it seems never to occur by itself in any Chamic language |
pu-vyā | queen — vyā also often occurs by itself62 |
pu-ciy | prince — ciy also often occurs by itself63 |
pu-lki | prince — lki is a reduced form of Old Cham lakiy ‘man’ (Malay laki ‘man’, Acehnese lakoe) |
pu-nai | princess — the nai element is probably a reduced form of Old Cham vinai ‘woman’ (Malay bini) |
pu-lyaṅ | minister/heir apparent — the lyaṅ element is unexplained64 |
pu-pov | lord/lady — chiefly in the expressions pu-pov tanā rayā ‘lord of the great earth’ and pu-pov ku ‘my lord’, both commonly used royal titles |
80Thurgood (1999: 301) reconstructs only one morpheme *pɔ ‘master, lord’ for Proto-Chamic and it is not clear at first sight to which of the two Old Cham morphemes, pu or pov, it corresponds. While pu looks identical to the high-status personal article pu found in Old Javanese, its usage in Old Cham, where it does not introduce proper names, is rather different. I suspect that the pu element is in origin the same morpheme as pov, principally confined to unstressed syllables.65 If so, the expression pu-pov might originally have meant ‘lord of lords’, just like pu-tau may have meant ‘lord of men’ (cf. Sanskrit nr̥-pati). Both pu and pov can also occur separately. We have seen, for instance, the common expressions yāṅ pov ku (besides which yāṅ pu-pov ku also exists) and yāṅ pu nagara. I am still unsure how to analyze the latter. In my epigraphic publications so far, I have treated putau as a single word and the other expressions as two words. Based on the new analysis presented here, it now appears to me that all the terms listed above must be treated as lexicalized compounds, i.e. as single words.
12.6. The Word kā
81This is a word of particularly common occurrence in the Old Cham corpus. It would seem to be related to the word ka that still exists in Modern Cham, but the meanings recorded for the latter do not seem to match closely with the usage of kā in Old Cham and even the syntactic position of kā in Old Cham seems to be different from modern usage of ka.66 In Old Cham, kā is most commonly placed before a predicate, in most cases right before it, though there are exceptions to this rule, as in the example selected in the interfoliated copy of the Dictionnaire (fig. 20).
Fig. 20. Aymonier & Cabaton (1906), interfoliated copy at the Société Asiatique, page facing p. 41.

82I believe that the glosses “pouvoir, être, avoir” and “tous ces champs pouvoir gens donner au Dieu” proposed for the headword and the phrase, respectively, can be improved. Having experimented unconvincingly with other interpretations in some previous publications,67 I now believe that the function of kā in Old Cham is best understood in the light of its apparent cognate ka in Acehnese, which expresses perfective aspect. Acehnese ka can often (roughly) be translated as ‘already’, as in ka-gopnyan-jak ‘in-3s-go, i.e. he went; he has gone already’ and ka uroe ‘it is already day’.68 Accordingly, the example in fig. 20 can be restated as follows.
C. 142, D lines 23–24 (Huber 1911b: 285–298)69
avih humā nan kā ra vuh di yāṅ
all field that kā 3p give to god
they have given all those fields to the god
83Such interpretations seem to work in all contexts I have analyzed so far. Two further examples will suffice.
C. 167, D2–3 (Griffiths, Pichon & Southworth 2021–2022, app. 2)
kā ājñā pov ku mahīndrādhipatih grāc vuh di vihāra śrī mahīndralokeśvara | yāṅ pu-pov kuv kā vriy vihāra niy ma-tandāh sarvatah
already status lord 1s m. worship offer to monastery ś. m. god lord 1s already give monastery this verb-annul completely
My lord Mahīndrādhipati reverently (grāc) offered them to the monastery of Śrī Mahīndralokeśvara. My God Lord allowed this monastery to annul (payment of taxes) completely.
C. 108, B7 (Huber 1911b: 269–277)
kā ñū matai nau dauk di yop naraka tauṅ yuga-antah pralaya
already 3s die go stay in darkness hell until aeon-end deluge
after he dies, he resides in the darkness of hell until the end of the aeon, the deluge
13. Script, orthography and phonology
84The last set of issues to be addressed here is the phonological interpretation of the written data. In this study, epigraphic data have been presented in Romanized form. Most texts, especially those that had not yet been adequately published elsewhere, have been represented first in what I have called “epigraphic editions”, which consist in a strictly one-to-one interpretation of the original script in Roman alphabet enriched with some diacritical marks.70 With or without preceding “epigraphic edition”, all data have been represented in what I have called a “normalized orthography”, which is in essence a re-interpretation of the strict transliteration, some graphic distinctions being eliminated as phonologically irrelevant and some ambiguities of the writing system being rendered transparent on the basis of script-internal and linguistic considerations, but otherwise staying close to the transliteration system from which it is derived.
85Simplifying the graphic facts as much as possible, it can be stated that Cham script has a repertoire of ten vowels (monophthongs and diphthongs) and of thirty-three consonants, arranged in fig. 21 in their traditional alphabetical order, from top to bottom and left to right.
Fig. 21. The repertoire of vowels and basic consonants in the Old Cham writing system.
vowels | a ā i ī u ū (r̥) e ai o au | ||||
basic consonants | k | kh | g | gh | ṅ |
c | ch | j | jh | ñ | |
ṭ | ṭh | ḍ | ḍh | ṇ | |
t | th | d | dh | n | |
p | ph | b | bh | m | |
y | r | l | v | ||
ś | ṣ | s | |||
h |
86In the absence of extra notation, a basic consonant character is read with default vowel a. For all other vowels distinguished in the system, there are dedicated characters for use in conjunction with a consonant (or group of consonants), and these are transliterated as a lower-case letter. When the vowel occurs as a stand-alone syllable, the corresponding character is transliterated with an upper-case letter. When a consonant is to be read without any vowel, in syllable-final position, the normal pattern is for a “vowel-killer” sign to be marked on it, and this is transliterated as median dot (·).71 However, there is another pattern that is used to express a limited set of consonant phonemes in final position and consists in using one of the special markers transliterated as ḥ, ṁ, m̃ and m̐. Although it is in the Indian tradition defined as a vowel marker, the character transliterated as r̥ is connected in the Cham implementation of Indic script to this special set, to which I return below.
87In the writing system thus (all too briefly) described some of the signs called “basic consonants” (the ṭ series as well as ś and ṣ) are typical for Sanskrit and relevant for transcribing Old Cham mainly when they occur there in loanwords. For most other basic consonant signs, the transliteration system seems to form a rather tight match to the consonant phonemes reconstructed for Proto-Chamic (fig. 22).
Fig. 22. The reconstructed Proto-Chamic consonant repertoire.
voiceless obstruents | *p, *t, *k, *c, *ʔ, *s, *h |
voiced obstruents | *b, *d, *g, *j |
aspirated consonants | *ph, *th, *kh, *bh, *dh, *gh |
nasals | *m, *n, *ŋ, *ñ |
liquids | *r, *l, *y, *w |
Table compiled from chapter 5 of Thurgood 1999, with exclusion of the glottalized voiced obstruents.
88The only differences of Romanization convention that the linguist familiar with Thurgood’s work needs to be aware of in interpreting the normalized texts are the epigraphist’s use of ṅ for /ŋ/ and the fact that spelled v can represent both /b/ and /w/, the dedicated character b being used only very rarely.
89With respect to the previously mentioned special set of markers, ḥ always maps to /h/ and ṁ mostly maps to /m/. This means that the writing system has a redundancy, with two signs (h and ḥ) being used for transcribing the single sound /h/ and two signs (m and ṁ) for transcribing /m/. The choice is determined by the way the writing system breaks strings of sound down into syllables that need not be detailed here any further.
90It is important to note that the marker transliterated as m̃ has nothing to do with /m/, but rather appears in the following four functions:
- as equivalent to syllable-final ṅ (e.g., urām̃ = urāṅ)
- as equivalent to syllable-final y (lakim̃ = lakiy)
- as equivalent to syllable-final v (vatum̃ = vatuv)
- to express a schwa vowel in the syllable nucleus (gam̃p = gəp)72
91A clear illustration of the fact that the distribution between im̃ and iy (function 2) is purely based on graphic convenience is found in C. 122 above (fig. 18), where the word niy figures in line 2 first in the spelling nim̃ and then as niy· for the reason that a descender from line 1 made the notation of m̃ difficult in the second case.73 Analogous patterns determine the spelling distribution in functions 1 and 3. The fourth function has introduced the vowel sound /ə/ which there is no inherited Indic vowel sign to represent, except that one of the values of the aforementioned character r̥ in Old Cham appears to be to transcribe the consonant-vowel combination /rə/.74
92Although there is a strong tendency for reading as ṅ (1) to be required after ā, as y (2) after i, as v (3) after u, and as ə (4) in other contexts, there is ambiguity in the system, as variant spellings prove this tendency not to be a hard rule. Notably, m̃ is proven to be capable of serving as equivalent to y even after other vowels other than i by heterographs like supum̃ besides spūy and supuy, chām̃ besides chāy, or tralām̃ besides tralāy.75
93The marker m̃ is a Cham innovation upon the Indic system that gains firm ground in Cham spelling only some centuries after the beginning of the Old Cham writing tradition. In periods and areas where the m̃ is not yet in use, or is used only sporadically, the functional scope of the marker ṁ (the Indic anusvāra) covers that of the m̃ described above. In texts that make use of the m̃, the function of ṁ is limited to transcribing /m/. Much more rarely than ṁ and m̃, we encounter another sign inherited from India, namely the m̐ (anunāsika), apparently sharing its functional scope with m̃.
94These aspects of the sign-sound nexus having been established, a near-phonological normalized orthography can be derived from the transliterated texts by applying several operations, partly mechanical and partly dependent on recognition of the underlying word. The process involves the following steps:
- adjusting the spelling of words borrowed from Sanskrit, as much as possible without disturbing what are likely to have been local pronunciations, to their dictionary spelling
- simplifying geminated consonants after r
- removing all vowel-killer signs ·
- eliminating the difference between stand-alone vowels (represented as capital letters) and dependent vowel markers
- interpreting the members of the group of special characters as required by contexts
95The normalized spelling can at this stage be no more than near-phonological insofar as the precise phonological value of every sign in the writing system can only be determined with reference to the phonemic system to be postulated for Old Cham, an issue which in turn depends on the phonemic distinctions reconstructed for Proto-Chamic. As repeatedly emphasized above, and as Thurgood himself admitted, to the extent that his reconstruction makes no use of Old Cham data, there is a good chance that a revised reconstruction taking due account of the inscriptional data would map more closely to spelled Old Cham. If the normalized spelling of syllable nuclei proposed here is linguistically tenable, it will bring Old Cham considerably closer to the Proto-Malayo-Polynesian ancestor than Thurgood’s reconstruction of Proto-Chamic implies. See the discussion of the diphthongization chain at the end of §9 above and compare the diagram in fig. 23.
Fig. 23. Vowel repertoire of second vs. main syllables in PMP and PC.

Reproduced from Thurgood 1999: 113, figure 8.
96It must be emphasized that the Old Cham spelling system presumably reflects the time when Indic script was adapted for Cham language and comprises several peculiarities (such as the use of v to transcribe both /b/ and /w/ or the doubling of consonants after r) that reflect Indian spelling habits. One must assume that a phenomenon of historical spellings became more relevant as the orthography came to correspond progressively less well with the spoken language during the development from its most archaic attested stage through various stages of Old Cham to the modern idioms. But spellings that look familiar from the Malay point of view while they look foreign when compared to modern Chamic languages can hardly be anything other than vestiges of how the language was pronounced at an early stage.76
97Besides the vowel system, already touched upon above, Thurgood’s reconstruction of Proto-Chamic vocabulary includes no instance of medial /nd/ while it does contain some reconstructed medial /nt/ and /mp/ clusters, apparently based on comparison with Acehnese and/or Malay because the extant Chamic languages show no prenasalization in these cognate sets (294 *kəntut ‘fart’, 296 *lumpɛy ‘dream’, 298 *məntah ‘raw, unripe’, 304 *tampɛy ‘winnow’). And yet the inscriptional data include a number of words with medial nd that can be matched safely with Malay words that contain the same. This implies that medial /nd/ needs to be admitted at the Proto-Chamic stage. Cf. Thurgood’s 297 *manɛy ‘bathe’ (Old Cham mandiy, Malay mandi, Acehnese manoe). Thurgood’s reconstructions do not include the ancestor of Old Cham raṇḍam ‘drench’ that can be matched with Malay rendam and Acehnese reundam. There are several other Old Cham words containing medial /nd/, one of them being the word landuṅ ‘eel’ discussed in §11.1. The cluster seems systematically to have lost its nasal in the pronunciation of Chamic languages, including Modern Cham, but there is a trace of an ancient consonant cluster in the Modern Cham writing system.77
14. Conclusion: Challenges and Perspectives
98The study of the Old Cham language offers a nice counterpoint to much of the other material presented at the conference Documenter et décrire les langues d’Asie : histoire et épistémologie. The materiality of the linguistic documentation in the forms of estampages, placing one extra documentary layer between the physical original and the abstract linguistic content, is one particularity. Another is Étienne Aymonier’s clear statement of his pathway to interpreting the inscriptions philologically, passing through the necessary preliminary of mastering the Modern Cham dialects. Despite the availability of this, and other heuristic approaches, the road forward is fraught with difficulties. We are dealing with a small corpus of texts, many of which have come down in damaged or incomplete state and hence present challenges already on the level of what the words of the texts are. This challenge is compounded by our limited knowledge of the lexicon, making it harder to offer educated guesses for which words stood in now damaged passages. Even perfectly legible words are often attested in only a single inscription (i.e., they are what philologists call hapax legomena), leaving us without the possibility of testing any hypotheses we might have on their meaning. And even if they do occur more than once, the number of occurrences is often insufficient to triangulate their meaning. This situation lends extra importance to the heuristic value of diachronic and interlinguistic comparison: plausible phonological correspondence with words in Modern Cham or related languages often yields crucial insight into the meaning or grammatical function of Old Cham words. Alas, the fact that Old Cham has so far been ignored as much by comparative linguistics as it has by scholars of Modern Cham after Aymonier, means that making use of historical/comparative evidence requires more spadework than would otherwise have been necessary. The nature of the texts presents another challenge. Whole swaths of the Old Cham lexicon that once existed will forever elude us, because many lexical items would never have had occasion to be used in the epigraphic context. Meanwhile, significant linguistic changes seem to have occurred between the stages of the language encountered in inscriptions and those which have been documented based on later written sources, or on linguistic fieldwork among speakers. The fact that 20th-century linguists have not engaged with historically documented stages of the language, while epigraphic research was in a lull for more than half of the 20th century, has stood in the way of cross-fertilization between descriptive/historical linguistics and epigraphy. On the bright side, I believe that there is a real scope for progress given the fact that nobody has even tried to apply a corpus approach, to systematically describe the grammatical constructions encountered in the corpus, and to tease meanings out of the texts using such angles as comparison with vocabulary of other Chamic languages, Acehnese, or related non-Chamic languages, and especially with Malay.
List of labels for morpheme glossing
caus | morpheme to derive causative verbs |
indef | indefinite marker |
instr | morpheme to derive instrument nouns |
neg | negative marker |
neg.imp | negative imperative marker |
rel | relativizer or topic marker |
status | status-marking personal article |
verb | verbal prefix of undetermined function |
Bibliographie
Adelaar, K. Alexander. 1992. The Relevance of Salako for Proto-Malayic and for Old Malay Epigraphy. Bijdragen tot de Taal-, Land- en Volkenkunde 148(3). 381–408.
Adelaar, K. Alexander. 2005. Structural Diversity in the Malayic Subgroup. The Austronesian languages of Asia and Madagascar, ed. by K. Alexander Adelaar and Nikolaus P. Himmelmann. London: New York: Routledge (Routledge Language Family Series). 202–226.
Adelaar, K. Alexander. 2008. On the Classifiability of Malayic. Language and Text in the Austronesian World: Studies in Honour of Ülo Sirk, ed. by Yuri A. Lander and Alexander K. Ogloblin. Munich: Lincom Europa (LINCOM Studies in Austronesian Linguistics, 6). 11–22.
Auwera, Johan van der & Frens Vossen. 2015. Negatives between Chamic and Bahnaric. Journal of the Southeast Asian Linguistics Society 8. 24–38.
Aymonier, Étienne. 1881. Recherches et mélanges sur les Chams et les Khmers, II : L’inscription chame de Dambang Dek. Notions sur les écritures et les dialectes chams. Excursions et Reconnaissances 4(10). 167–186.
Aymonier, Étienne. 1889. Grammaire de la langue chame. Excursions et Reconnaissances 14. 5–92.
Aymonier, Étienne. 1891. Première étude sur les inscriptions tchames. Journal asiatique. 5–186.
Aymonier, Étienne. 1892. Une mission en Indo-Chine : Relation sommaire. Bulletin de la Société de Géographie de Paris 13(7). 216, 339.
Aymonier, Étienne. 1896a. Découvertes archéologiques de M. C. Paris dans la province de Quang-nam ; Inscriptions découvertes par M. C. Paris au Tchampa. Bulletin de géographie historique et descriptive. 92–95.
Aymonier, Étienne. 1896b. Rapport sommaire sur les inscriptions du Tchampa, découvertes et estampées par les soins de M. Camille Paris. Journal asiatique 9(7). 146–51.
Aymonier, Étienne. 1911. L’inscription čame de Po Saḥ. Bulletin de la commission archéologique de l’Indochine. 13–19.
Aymonier, Étienne & Antoine Cabaton. 1906. Dictionnaire čam-français. Paris: Ernest Leroux (Publications de l’École française d’Extrême-Orient, 7).
Balogh, Dániel & Arlo Griffiths. 2020. DHARMA Transliteration Guide. Project documentation. Paris; Berlin: Centre d’Études de l’Inde et de l’Asie du Sud; École française d’Extrême-Orient; Humboldt-Universität. https://hal.science/halshs-02272407v3.
Bhattacharya, Kamaleswar. 1961. Précisions sur la paléographie de l’inscription dite de Võ-cạnh. Artibus Asiae 24(3-4). 219–224.
Bloch, Jules. 1919. La formation de la langue marathe. Paris: Champion (Bibliothèque de l’École des hautes études, Sciences historiques et philologiques, 215).
Blood, David L. 1962. A Problem in Cham Sonorants. STUF – Language Typology and Universals 15(1–4). 111–114.
Blood, Doris E. 1961. Women’s Speech Characteristics in Cham. Asian Culture 3. 139–143.
Blust, Robert. 2000. Review of From Ancient Cham to Modern Dialects: Two Thousand Years of Language Contact and Change, by Graham Thurgood. Oceanic Linguistics 39(2). 435–445.
Blust, Robert. 2013. The Austronesian Languages: Revised Edition. Canberra: Australian National University Press.
Brocheux, Pierre & Daniel Hémery. 2009. Indochina: An Ambiguous Colonization, 1858-1954. Berkeley, CA: University of California Press (From Indochina to Vietnam: Revolution and War in a Global Perspective, 2).
Brunelle, Marc. 2019. Revisiting the Expansion of the Chamic Language Family: Acehnese and Tsat. Champa: Territories and Networks of a Southeast Asian Kingdom, ed. by Arlo Griffiths, Andrew Hardy, and Geoff Wade. Paris: École française d’Extrême-Orient (Études Thématiques 31). 287–302.
Brunelle, Marc & James Kirby. 2016. Tone and Phonation in Southeast Asian Languages. Language and Linguistics Compass 10(4). 191–207.
Bukhari Daud & Mark Durie. 1999. Kamus Basa Acèh = Kamus Bahasa Aceh = Acehnese-Indonesian-English Thesaurus. Canberra: Pacific Linguistics, Research School of Pacific and Asian Studies, The Australian National University (Pacific Linguistics, Series C 151).
Burnell, A. C. 1870. A Few Suggestions as to the Best Way of Making and Utilizing Copies of Indian Inscriptions. Madras: Higginbotham.
Cabaton, Antoine. 1901. Nouvelles recherches sur les Chams. Paris: Leroux (Publications de l’École française d’Extrême-Orient, 2).
Cabaton, Antoine. 1904. L’inscription chame de Bien-Hoa. Bulletin de l’École française d’Extrême-Orient 4. 687–690.
Chatterji, Suniti Kumar. 1926. The Origin and Development of the Bengali Language. 2 vols. Calcutta: Calcutta University Press.
Clavé, Elsa & Arlo Griffiths. 2022. The Laguna Copperplate Inscription: Tenth-Century Luzon, Java, and the Malay World. Philippine Studies: Historical and Ethnographic Viewpoints 70(2). 167–242.
Clémentin-Ojha, Catherine & Pierre-Yves Manguin. 2001. Un siècle pour l’Asie : l’École française d’Extrême-Orient, 1898-2000. Paris: les Éditions du Pacifique ; École française d’Extrême-Orient.
Cœdès, George. 1912. Note sur deux inscriptions du Champa. Bulletin de l’École française d’Extrême-Orient 12. 15–17.
Cœdès, George. 1923. Liste générale des inscriptions du Champa et du Cambodge. Listes générales des inscriptions et des monuments du Champa et du Cambodge, by George Cœdès & Henri Parmentier. Hanoi: Imprimerie d’Extrême-Orient. 1–180.
Cœdès, George. 1929. Etienne-François Aymonier (1844 † 1929). Bulletin de l’École française d’Extrême-Orient 29. 542–548.
Cœdès, George. 1939. La plus ancienne inscription en langue chame. Eastern and Indian Studies in Honour of F.W. Thomas, C.I.E., New Indian Antiquary Extra Series I, 48. 39–52.
Cumming, Susanna. 1991. Functional Change: The Case of Malay Constituent Order. Berlin & New York: Mouton de Gruyter (Discourse Perspectives on Grammar 2).
Durand, Eugène-Marie. 1903. Le temple de Po Romé à Phanrang. Bulletin de l’École française d’Extrême-Orient 3. 597–603.
Durie, Mark. 1985. A Grammar of Acehnese on the Basis of a Dialect of North Aceh. Dordrecht: Foris (Verhandeling van het Koninklijk Instituut voor Taal-, Land- en Volkenkunde, 112).
Finot, Louis. 1903. Notes d’épigraphie, V : Pāṇḍuraṅga. Bulletin de l’École française d’Extrême-Orient 3. 630–648.
Finot, Louis. 1904a. Notes d’épigraphie, VI : Inscriptions du Quang Nam. Bulletin de l’École française d’Extrême-Orient 4. 83–115.
Finot, Louis. 1904b. Notes d’épigraphie, XI : Les inscriptions de Mi-Sơn. Bulletin de l’École française d’Extrême-Orient 4. 897–977.
Finot, Louis. 1915. Notes d’épigraphie, XX : L’épigraphie indochinoise. Bulletin de l’École française d’Extrême-Orient 15(2). 113–135.
Gonda, Jan. 1973. Sanskrit in Indonesia. Second edition. Śata-Piṭaka Series 99. New Delhi: International Academy of Indian Culture.
Griffiths, Arlo, Andrew Hardy & Geoff Wade, eds. 2019. Champa: Territories and Networks of a Southeast Asian Kingdom. Paris: École française d’Extrême-Orient (Études Thématiques, 31).
Griffiths, Arlo, Bob Hudson, Marc Miyake & Julian K. Wheatley. 2017. Studies in Pyu Epigraphy, I: State of the Field, Edition and Analysis of the Kan Wet Khaung Mound Inscription, and Inventory of the Corpus. Bulletin de l’École Française d’Extrême-Orient 103. 43–205.
Griffiths, Arlo & Amandine Lepoutre. 2016. Études du corpus des inscriptions du Campā, VIII: Les inscriptions des piédroits des temples de Po Klaong Girai (C. 8-11), de Linh Thái (C. 109-110) et de Yang Prong (C. 116). Bulletin de l’École française d’Extrême-Orient 102. 195–296.
Griffiths, Arlo, Amandine Lepoutre, William Aelred Southworth, & Thành Phân. 2008–2009. Études du corpus des inscriptions du Campā, III : Épigraphie du Campa 2009-2010 : prospection sur le terrain, production d’estampages, supplément à l’inventaire. Bulletin de l’École française d’Extrême-Orient 95. 435–497.
Griffiths, Arlo, Amandine Lepoutre, William Aelred Southworth, & Thành Phân. 2012. Văn khắc Chămpa tại Bảo tàng Điêu khắc Chăm – Đà Nẵng / The inscriptions of Campā at the museum of Cham sculpture in Đà Nẵng. Ho Chi Minh City; Hanoi: VNUHCM Publishing House and Center for Vietnamese and Southeast Asian Studies University of Social Sciences and Humanities Vietnam National University Hồ Chí Minh City; École française d’Extrême-Orient.
Griffiths, Arlo, Salomé Pichon & William Aelred Southworth. (2023) 2021–2022. Études du corpus des inscriptions du Campā, XII: Buddhist Vihāras in Campā, from the 7th to the 14th Centuries. Buddhism, Law & Society 7 (Special volume on Epigraphic Evidence on Patronage and Social Contexts of Buddhist Monasteries in Medieval South and Southeast Asia), ed. by Ryosuke Furui, Arlo Griffiths, & Annette Schmiedchen. 275–326.
Groupe de recherches cam. 1977. Études cam, I : Essai de translittération raisonnée du cam. Bulletin de l’École française d’Extrême-Orient 64. 243–255.
Guérin, Mathieu. 2017. Tup, un lettré aventurier au service de l’orientaliste Etienne Aymonier. Journal asiatique 305(1). 111–118.
Guillon, Emmanuel. 2011. Art et archéologie du Champa : Une ancienne civilisation du Viet Nam. Paris: Picard.
Hardy, Andrew & Tiến Đông Nguyễn. 2019. The Peoples of Champa: Evidence for a New Hypothesis from the Landscape History of Quảng Ngãi. Champa: Territories and Networks of a Southeast Asian Kingdom, ed. by Arlo Griffiths, Andrew Hardy, & Geoff Wade. Paris: École française d’Extrême-Orient (Études Thématiques, 31). 121–144.
Huber, Edouard. 1911a. Études indochinoises VI-VII. Bulletin de l’École française d’Extrême-Orient 11(1–2). 1–24.
Huber, Edouard. 1911b. Études indochinoises VIII-XII. Bulletin de l’École française d’Extrême-Orient 11(3-4). 259–311.
Huffman, Franklin E. 1986. Bibliography and Index of Mainland Southeast Asian Languages and Linguistics. New Haven, London: Yale University Press.
Jenner, Philip N. 2009. A Dictionary of Pre-Angkorian Khmer. Canberra: Pacific Linguistics, Research School of Pacific and Asian Studies, The Australian National University (Pacific Linguistics, 597).
Jenner, Philip N. & Paul Sidwell. 2010. Old Khmer Grammar. Canberra: Pacific Linguistics.
Lee, Ernest W. 1974. Southeast Asian Areal Features in Austronesian Strata of the Chamic Languages. Oceanic Linguistics 13(1-2). 643–668.
Lee, Ernest Wilson. 1966. Proto-Chamic Phonologic Word and Vocabulary. Ph.D. Thesis, [Bloomington]: Indiana University.
Lee, Ernest Wilson. 1996. Bipartite Negative in Chamic. Mon-Khmer Studies 26. 291–317.
Lepoutre, Amandine. 2015. Études du corpus des inscriptions du Campā, VII : L’inscription de Jaya Parameśvaravarman à Tháp Đôi (C. 213) avec en annexe deux nouvelles inscriptions du même souverain (C. 218.2 et C. 219). Bulletin de l’École française d’Extrême-Orient 101. 117–157.
Lesserteur, E.-C. 1882–1883. Inscriptions qhiames de l’ancien Ciampa. Bulletin de la Societé Académique indo-chinoise de France 2. 7–9.
Lesserteur, E.-C. 1885. Inscriptions qhiames (cham). Revue française de l’étranger et des colonies 11. 476–479.
Lombard, Denys. 1983. Un grand précurseur : Antoine Cabaton (1863-1942). Archipel 26. 17–24.
Mahdi, Waruno. 2005. Old Malay. The Austronesian Languages of Asia and Madagascar, ed. by Alexander Adelaar and Nikolaus P. Himmelmann. London, New York: Routledge (Routledge Language Family Series). 182–201.
Marrison, Geoffrey E. 1975. The Early Cham Language, and Its Relationship to Malay. Journal of the Malaysian Branch of the Royal Asiatic Society 48(2 [228]). 52–59.
Maspero, Georges. 1928. Le royaume de Champa. Paris-Bruxelles: Vanoest.
Moussay, Gérard. 2006. Grammaire de la langue cam. Paris: Les Indes savantes.
Pacquement, Jean. 2000. Linguistique historique et histoire de la langue : 80 ans après La formation de la langue marathe de Jules Bloch. Bulletin de l’École française d’Extrême-Orient 87(2). 741–763.
Pollock, Sheldon. 1996. The Sanskrit Cosmopolis, 300-1300: Transculturation, Vernacularization, and the Question of Ideology. Ideology and Status of Sanskrit: Contributions to the History of the Sanskrit Language, ed. by Jan E. M. Houben. Leiden: Brill (Brill’s Indological Library 13). 197–247.
Pollock, Sheldon. 2006. The Language of the Gods in the World of Men: Sanskrit, Culture, and Power in Premodern India. Berkeley: University of California Press.
Schweyer, Anne-Valérie. 2005. Po Nagar de Nha Trang, seconde partie : Le dossier épigraphique. Aséanie, Sciences humaines en Asie du Sud-Est 15. 87–119.
Shorto, H. L. 2006. A Mon-Khmer Comparative Dictionary, ed. by Paul Sidwell, Doug Cooper, & Christian Bauer. Canberra: Pacific Linguistics, Research School of Pacific and Asian Studies, The Australian National University (Pacific Linguistics, 579).
Singaravélou, Pierre. 2005. De la découverte du Champa à la fondation de l’École coloniale, itinéraire d’Étienne Aymonier d’après ses mémoires inédits. D’un Orient l’autre : Actes des troisièmes journées de l’Orient, Bordeaux, 2-4 octobre 2002, ed. by Jean-Louis Bacqué-Grammont, Angel Pino, & Samaha Khoury. Louvain: Peeters (Cahiers de la Société asiatique, 4). 237–248.
Thomas, David D., Ernest W. Lee & Nguyễn Đăng Liêm, eds. 1977. Chamic Studies. Canberra: Dept. of Linguistics, Research School of Pacific Studies, Australian National University (Papers in South East Asian Linguistics, 4).
Thurgood, Graham. 1999. From Ancient Cham to Modern Dialects: Two Thousand Years of Language Contact and Change: With an Appendix of Chamic Reconstructions and Loanwords. Honolulu: University of Hawai’i Press (Oceanic Linguistics Special Publications, 28).
Vickery, Michael. 1998. Society, Economics and Politics in Pre-Angkor Cambodia: The 7th-8th Centuries. Tokyo: The Centre for East Asian Cultural Studies for Unesco, The Toyo Bunko.
Weber, Nicolas. 2014. Note sur les inscriptions en cam du pāḷāt’ Tut, estampeur d’Étienne Aymonier. Bulletin de l’École française d’Extrême-Orient 100. 253–258.
Zapata-Aubé, Nicole. 2017. Victorien Pierre Lottin de Laval et La Lottinoplastie. Le Caire Dessiné et Photographié Au XIXe Siècle, ed. by Mercedes Volait. Paris: Publications de l’Institut national d’histoire de l’art. 139–156.
Notes de bas de page
1 Since some linguists who read a draft of this article seemed not to be familiar with the scholarly genre I am alluding to here, I explain that I mean a work that deals with the historical development of a given language from its reconstructed prehistory, through any attested premodern stages, to its modern form. For Indian examples, see Bloch 1919 (with Pacquement 2000) and Chatterji 1926. In the Southeast Asian domain, Cumming 1991 partially exemplifies the genre.
2 See Cœdès 1923 and Griffiths et al. 2008–2009.
3 Bhattacharya 1961; see also Griffiths et al. 2017: 72 and 86 n. 95.
4 These sociolinguistic questions have begun to be posed in recent scholarship (Hardy & Nguyễn Tiến Đông 2019), though from a historical rather than a linguistic angle, and will not occupy us further here.
5 The relevant inventory numbers are C. 2, C. 15, C. 16, and C. 27. See Aymonier 1881, Durand 1903, Weber 2014.
6 The publication Groupe de recherche cam (1977) is rather unsatisfactory for anyone wishing to understand details of how modern Cham script works from a historical point of view.
7 See the discussion in Brocheux & Hémery 2009.
8 For the comfort of readers unable to read French, this and all subsequent quotations from French secondary literature are presented in translation. The original passages have not been included for reasons of space-efficiency.
9 Translated from the original French of Lesserteur (1882: 7–8, reproduced in Lesserteur 1885: 477–478).
10 The inscriptions alluded to at the end of the quote are the graffiti of Phong Nha cave in Quảng Bình province (C. 114), that remain undeciphered to this day. See also the Chronique in Bulletin de l’École française d’Extrême-Orient 38 (1938), pp. 412–413. An EFEO team has visited the cave in 2015 and confirmed the presence of Old Cham graffiti.
11 Finot (1915: 118–119): “From 1882 to 1885, he travelled through Cambodia, Siam, Laos and southern Annam, as far as Bình Định, where he was stopped by the insurrection that broke out at the time. This trip to Annam marked a milestone in the history of Indochinese philology: it unquestionably fixed the geographical location of Champa (!); it revealed the existence of an abundant epigraphic literature in both Sanskrit and the vernacular, originating from this kingdom; and, finally, it brought to light the oldest inscription in Indochina, that of Võ Cạnh, which moved the history of Indian colonization back several centuries”.
12 On Aymonier’s life and work, I have consulted Cœdès 1929; Clémentin-Ojha & Manguin 2001: 125; Singaravélou 2005.
13 Aspects of Aymonier’s work with native personnel in Cambodia and Laos are studied in Guérin 2017. As Guérin notes (2017: 112), “In his publications, Etienne Aymonier acknowledges the importance of the help he received from his assistants, a dozen Khmers and at least two Chams, without, however, telling us much about these men who accompanied him on his journeys, helped him to make the estampages and decipher the manuscripts he worked on.” While Aymonier mentions the names of several of his assistants recruited in Cambodia, some of whom accompanied him also in the initial phase of his Campā fieldwork, and praises their work (e.g., 1892: 353), he seems to have been less satisfied by his assistants recruited in Bình Thuận: “In this new workforce, I no longer found the moral and intellectual resources I’d been accustomed to with my Cambodians, who could be sent out on their own in perfect confidence” (1892: 360). Unless I have overlooked anything, the names of the Chams who made up the second team are never mentioned.
14 For more on the technical aspects of such estampages, see Burnell 1870: 4; Finot 1915: 118 (n. 2); Zapata-Aubé 2017.
15 Like other languages of mainland Southeast Asia, modern Cham is poor in inflectional and derivational morphology, but it does (or at least used to) have some affixes which Aymonier had duly described in his 1889 Grammaire de la langue chame. In the 1891 article, Aymonier mentions infixation with -an- on p. 26 and the causative prefix pa- on pp. 34, 43 and 64 (n. 2).
16 On p. IX, they also state that “the spelling of the old inscriptions is more stable” compared to the spelling of post-Campā texts. This repeats an observation already made by Aymonier in 1891.
17 The entries marked with “I.” are inơ̄ (p. 30), udaya (33), kakai (43), čauṅ (132), jadrơṅ (144, citing epigraphic driṅ « régner »), tavuḥ/patavuḥ (187/259), tuy (194), thil (206), liṅi̯ak/paliṅi̯ak (282/442), pasli̯aṅ (285), pu (287), mơn (375), vuḥ (456), vukān (456), vunuḥ (456). In addition, there are some entries referring to inscriptions in other ways, though not always for their Old Cham lexical contents: inrā (30), ĕi (32), ău (35), galaṅ (102), tablaḥ (180), pali̯ơṅ/pōli̯ơṅ (283, 310), batău (320), baber (325), barak (326), raḍai (408), lanuṅ (433), lĕi (455), çadaḥ (460), harơḥ (516), humā (526).
18 Two of the explorers involved were Étienne Navelle and Camille Paris. See Finot 1915 and Aymonier 1896.
19 See the entries for Cœdès, Finot and Huber in my bibliography.
20 A handwritten note, presumably by Aymonier, at the first page of the Société Asiatique copy of the dictionary that was itself published in 1906, seems to state: “Travail de report commencé le 8 Décembre 1906. d’après le Répertoire préparé pour l’Introduction et prêté à Mr C. de Janvier à Octobre 1906. Fini le 18 Xbr. 1906”. But I am not sure whether this means that all of the material handwritten in red ink predates the 10th of October 1906.
21 See Huffman 1986: 530–531.
22 Maspero (1928: 7) speaks of “the Cham language, which today does not differ significantly from that used in the oldest inscriptions in the vernacular language.” In a footnote, Maspero attributes the oldest inscriptions to the 11th century. Both on the dating of the oldest Old Cham inscriptions and on the proximity of Old to Modern Cham, the French scholar was unfortunately quite unreliable.
23 Since only one of Doris Blood’s publications on Cham, that I am aware of, predates 1966, it seems that Lee was relying on a personal communication. The point is valid.
24 Some publications that may be cited are (Doris) Blood 1961, (David) Blood 1962, the papers by the Bloods in Thomas, Lee & Nguyễn 1977, and Lee 1974.
25 A useful review of this work was published by Bob Blust (2000).
26 Malay and Acehnese data are cited in their contemporary Roman orthographies. It is perhaps not superfluous to recall that in Malay orthography, spelt e most often expresses the phoneme /ə/. On the Romanized representation of Old Cham data, see §13.
27 See Thurgood 1999: 116 and Appendix II, as well as Brunelle 2019: 290–291.
28 As some of the examples are a bit long, in order to save space, I do not present word-for-word glosses in the interlinear way most linguists are accustomed to nowadays.
29 See below (§13) on nd ~ n variation. In Acehnese, we find lindông in the same meaning. A similar-sounding proto form has been reconstructed for the Mon-Khmer languages: *[l]nduŋ ~ *dnduuŋ ~ *dnduŋ ~ *nduŋ ‘eel’ (Shorto 2006: 194–195, no. 579).
30 See Guillon 2011: 198–199 with further references in n. 2.
31 See n. 15 above on Cham derivational morphology. See also Thurgood 1999, chapter 9 on affixes reconstructible to Proto-Chamic. See Aymonier & Cabaton 1906: 43 for kakaṁ “Incruster; faire adhérer; s’incruster, se couvrir d’une croûte”. See also p. 123 canuk “Modèle, forme; accord, ensemble, règles (?), usages (?); rangée, brochette” and canök “Condition, état; couleur, teinte; lignée, race, famille. Cf. pour le sens le skt. varṇa”. See further p. 134 cvak “Nouer, entrelacer, enlacer; ceindre, enrouler autour de la tête; se retrousser, relever son pagne, ses manches, etc., en les serrant, retroussis, bord, nœud” and p. 130 čuk “Enfiler, passer, mettre, embrocher, se vêtir. — čuk karaḥ, passer, porter une bague. čuk dṷa sauṅ kauṅ mơḥ, [elle] portait deux paires de bracelets d’or”. Michel Antelme draws my attention to mod. Khmer ច្រវាក់ cravāk’ ‘chain,’ and ឆ្វាក់ chvāk’ ‘to tie (around, e.g., with a cord), make a knot’ — and this connection was already made by Aymonier & Cabaton (p. 134).
32 Space had been left blank for two syllables. Restore ⟨para⟩mapura, as in C. 116 S, 3: pu (nai) gaurendralakṣm(ī) Urām̃ paramapura (Griffiths & Lepoutre 2016: 211).
33 Cham kings typically bear the title yāṅ po(v) ku which is strictly analogous to the title vraḥ kamratāṅ ’añ borne by Khmer kings. My translation is very literal. An expression like ‘His Majesty’ would be more idiomatic in English. The word yāṅ is not cognate with Malay yang but rather with Old Javanese hyaṅ and Old Malay hiyaṅ.
34 See Aymonier & Cabaton (1906: 140–141): “ja, jā. Appellatif des gens de médiocre condition, familier ou de mépris : toi, lui, individu ; ce (C.). —jā nī, ce gars, cet individu-ci. jā ban nam, ce garçon-là. jā daṃ, le gars, le garçon, l’individu. jā rauṅ yṷon kaliḥ, les petits Annamites ennemis, les espions, les mouchards (insulte). jā bol, la plèbe. jā baiḥ, nom d’un gardien de buffles.”
35 See Gonda 1973: 583–587 on the adoption of Sanskrit nouns to serve as conjunction.
36 Lepoutre 2015: 126 n. 58.
37 According to Thurgood (1999: 103), Proto-Chamic had no final /ɲ/. This claim may have to be reconsidered.
38 Cœdès (1939: 47–48) rightly called puñ a “challenging word”. He formulated the following hypotheses: (1) It could be a vernacularized spelling of Sanskrit puṇya ‘pious work’, attested elsewhere in mainland Southeast Asian languages, in which case the sentence would mean: ‘This divine serpent is the work of the king’. (2) It could be a possessive marker identical to Malay punya: ‘This is the divine serpent of the king.’ (3) It could be a personal pronoun of the first person, attested in Old Javanese as (*pun >) pinun: ‘This is the divine serpent of me, the king’. (4) It could be a honorific title pu-ñ, built with a pronominal element, like Old Cham po-ku or Old Malay pu-nta, but this presupposes the use of a Mon-Khmer pronominal element such Old Khmer añ: ‘This is the divine serpent of his majesty the king’. Cœdès only marks hypothesis 2 explicitly as implausible. I agree with this judgment, as Old Cham has 3p pronoun ñū corresponding to Malay -nya and possessive constructions are always unmarked elsewhere in Old Cham. (It is this interpretation that Marrison favored, fudging the spelling difference between puñ and punya.) Alas, I do not find any of the other hypotheses significantly more plausible. Perhaps we must consider a suggestion made to me by Michel Antelme, namely, to connect puñ with the pre-Angkorian Old Khmer word poñ, elaborately discussed by Vickery (1998: 190–205) and glossed by Jenner (2009: 320) as “Courtesy title for males (and females?), borne by commoners of honorable status”.
39 See below §12.2 on Pronouns.
40 See below §12.1 on Negation.
41 Sander Adelaar (pers. comm.) suggests that one might speculatively assume *dawam as proto-form becoming *dumawam with *-um- infix, after which the antepenultimate vowel (as often happens) gets reduced to schwa while *awa is reduced to a under pressure of bisyllabic tendencies of the Indonesian languages.
42 Note that Malay kurun ‘century’ is a loanword from Arabic and hence irrelevant in this context.
43 See nn. 15 and 31 above.
44 See the discussion and examples cited in Clavé & Griffiths 2022: 192, with nn. 63 and 64.
45 Lee 1996, van der Auwera & Vossen 2015.
46 See Lee 1996: 304–305, where the Modern Cham cognate is transcribed jôi.
47 Besides the linguistic studies referred to just above, see the quotation in n. 66 below.
48 The fact that the normally attributive word proṅ (‘big’) stands before and not after the Sanskrit compound noun bhoga-upabhoga (‘enjoyment of pleasures’ or ‘resources and minor resources’, a common expression for the means of economic subsistence of religious institutions) suggests to me that the first word is here nominalized (‘bigness’) and stands in a possessive construction with the second. This manner of nominalizing predicates is familiar in languages of Indonesia, though normally with linker between possessed and possessor (ex. Malay miskin-nya orang tua ‘the poverty of parents’). I do not know if they are also known in Mainland Southeast Asian languages.
49 Mahdi 2005, table 6.5.
50 The interpretation I propose here is slightly different from the one recently published in Griffiths, Pichon & Southworth 2021–2022, as I have only now seen that gəp gotra and yuga-antah pralaya are both likely to be twin-forms. On the potential equivalence of the Cham word gəp with the Sanskrit word gotra, see Lepoutre 2015: 126 n. 57.
51 See Thurgood (1999: 338): “halɛy ‘who; question word’, Rade hlɛi, Chru hələi, N. Roglai aləi, Tsat ʔa³³ zaːŋ³³ -i v?, Haroi həlăi -v (grammatical particle), W. Cham hlay ‘whatever’, PR Cham halɛ̆y; lɛ̆y, Wr. Cham halɛ̆i”. See also Moussay 2006: 234–235.
52 See Huber 1911b: 311. This scholar’s explanation of vr̥- with reference to the bər- prefix in Malay and mra-/pra- prefixes in Khmer and Mon does not seem convincing. Is there a connection with the Malay indefinite morpheme barang? To make the issue even more complex, several inscriptions spell vriy (which normally means ‘to give’) in contexts analogous to those where others use vrəliy.
53 This word is a half-Cham calque on the fully Sanskrit sadā-kāla found in the text in §11.5. The word avis is more often spelled avih.
54 See Aymonier & Cabaton 1906: 482–483. And then p. 484: “Voir thĕi. Pronom interrogatif et indéterminé: qui, lequel, quiconque, quelqu’un ; personne, aucun, ensemble. sauṅ sĕi, avec quelqu’un; avec qui ? kā sĕi, à qui ? sĕi nau, qui va ? halā sĕi čiḥ yău nī, le bétel, qui l’a préparé ainsi ? oḥ sĕi khin glaṅ, personne ne veut voir ou ne peut voir.” Moussay (2006: 270) described the same usages of sĕi. See also Thurgood 1999: 303.
55 Sometimes, apparently for euphonic reasons, the -an- infix surfaces as -and- (e.g. gandrāc from grāc).
56 Sections 6.3.1.3 *ma- ‘stative’, 6.3.1.6 *maŋ- ‘actor voice’/*paŋ- ‘instrumental noun’, and 6.3.1.7 *maR- ‘actor voice’/*paR- ‘instrumental noun’. In Malay, *maŋ- has become meN- and *maR- has become ber- but *ma- does not survive except in fossilized forms like merah ‘red’ and makan ‘to eat’.
57 I suspect a connection with Malay endap and Modern Cham kandap (Aymonier & Cabaton transcribe kaḍap), words which imply meanings like ‘hide, conceal, stay’ — hence perhaps ‘place’.
58 See discussion in Blust 2013, § 6.3.2.1 *-um- ‘actor voice’.
59 I suspect an analogical process involving sebelah ‘side’ may have played a role. Adelaar (1992: 392) proposes a completely difference hypothesis on the origin of seberang in Malay.
60 See Aymonier & Cabaton (1906: 495): “sraṅ, srơṅ (bahn. çơlō̃ṅ). Passer, traverser ; être porté sur, flotter”. In a personal communication, Mark Durie suggest that “Cham srāṅ is cognate with Acehnese sreueng, which means ‘to cast up onto a bank or beach’, e.g. by a powerful flood water. It can, metaphorically, also mean ‘to move rapidly’. seumeureueng is sreueng + the -m- infix. In Acehnese the -m- infix detransitivizes the verb. So seumeureueng means to move forward forcefully (like a floodwater rushing forward).” He rejects the hypothesis of a connection with Malay seberang.
61 The term has survived into Modern Cham. See Aymonier & Cabaton (1906: 260): “patau (atj. pauté; jar. patao). Voir putau. Roi.” Thurgood’s reconstruction *pataw (1999: 342) may need to be revised in the light of the Old Cham evidence for the vowel in the presyllable. Incidentally, Mark Durie (pers. comm.) has informed me that pauté cited by Aymonier & Cabaton for Acehnese is a mistaken transcription of Jawi (Arabic script) p-w-t, as this should be transcribed po-teu ‘our lord’, a title for God or a sultan, and is not cognate with Cham patau.
62 The term vyā has survived into Modern Cham. See Aymonier & Cabaton (1906: 345): “bi̯ā, bi̯a, bi̯eā (bahn. bia; jav. biyuṅ ; kawi biyada ; mal. biyāda). Reine, dame; princesse, déesse, appellatif des filles de sang royal : Madame. — patau bi̯ā, pō bi̯ā, reine”.
63 The term ciy has survived into Modern Cham. See Aymonier & Cabaton (1906: 129): “čĕi (atj. čĕhik). Voir kačĕi. Prince, maître, appellatif des divinités inférieures, des membres de la famille royale ; monsieur ; une constellation. — čĕi nai, monsieur et madame”. See also p. 45 “kačĕi (atj. kečehik, čĕhik). Voir cĕi. Appellation moins honorifique que čĕi : maître ; prince, seigneur”. Mark Durie informs me that the Acehnese words referred to by Aymonier & Cabaton with “atj.” should have been transcribed chik and keuchik (where final -k is a glottal stop). Despite what Aymonier & Cabaton suggest, it seems unlikely that Acehnese (keu)chik is related Modern Cham (ka)čĕi and hence to Old Cham ciy, as the Acehnese /ch/ cannot correspond to Cham /c/.
64 The whole term has survived into Modern Cham. See Aymonier & Cabaton (1906: 283 and 310): “pali̯ơṅ. Voir pōli̯ơṅ. Titre de deux premiers ministres (inscr. pu lyaṅ).”; “A. pōli̯ơṅ. Titre de certains seigneurs (inscr. pulyaṅ).” In the second entry, the A. means that the term was recorded specifically for ‘Annam’, i.e. Eastern Cham.
65 On the distinction between unstressed initial syllables (“presyllables”) and stressed main syllables, and the more restricted range of vowels that are found in the former in Chamic languages, see Thurgood (1999: 104–106).
66 See Aymonier & Cabaton (1906: 41): “ka, kā (ka, prép. d’usage analogue dans les langues indonésiennes), particule, souvent explétive, très usitée en čam. — a. Signe du datif, signe du régime indirect : à, vers. —b. D’abord, de prime abord, dès le premier instant, au préalable, avant tout, provisoirement, un peu ; indique une chose à accomplir avant une autre. — c. Avec la négative o, oḥ ou dī, kā donne l’expression « pas encore ». — d. Que, pour que, afin que ; toutes les acceptions de la conjonction « que » ; sert à relier les phrases. — e. De, sur, que...ne, ce qui, pour que, à cause de, car, également, aussi, au lieu de, en place, dans, etc.” Summarizing Moussay’s section on ka (2006: 239–240), he distinguishes the following functions: (A) subordinating conjunction; (B) preposition to, for, etc.; (C) verbal auxiliary, postposed, indicating that the action has been accomplished previously. Moussay has separate entries for ka mang and ka...o.
67 E.g., Griffiths et al. 2012: 212–213 (on C. 43 A, lines 10 and 15, with kā twice translated as “then”) and 223 (speculating on a connection of kā with the word pakā and of both with the Malay conjunction maka). The translation “then” was also used in the publication of C. 167 for which a revised proposal will be made just below.
68 The examples are 4-215 and 4-5 in Durie’s grammar (1985: 47, 101). On p. 248 of that work, the author explains: “ka is used for an event that has happened, or a state that has begun. It is glossed in for inchoative”. See also the thesaurus by Bukhari Daud & Durie (1999: 112), where ka is included among “Preverbal Auxiliaries”, glossed with Indonesian “sudah” and English “already, now”.
69 Some lines having been left unread in Huber’s edition, the passage is there found in lines numbered 18–19 (p. 292).
70 The precisely transliterated representation of the original texts used here follows a set of rules adapted from the ISO standard 15919 for transliteration of Indic scripts. The adaptations have been made in order to render this standard capable of representing graphic phenomena found in South and Southeast Asian inscriptions. See Balogh & Griffiths 2020.
71 In archaic writing, as in inscription C. 174 illustrated above (fig. 17), the vowel-killer method was not yet in use, and vowel-less consonants are, instead, represented below the line (a method represented in transliteration by upper-case letters).
72 See Thurgood (1999: 248, 318).
73 Cf. also the numerous cases in the inscriptions edited in Griffiths & Lepoutre (2016), and the discussion on p. 208 of that article.
74 Epigraphic haR̥ḥ corresponds to Modern Cham haram̃ḥ, i.e. /harəh/. See n. 17 above. Use of r̥ to transcribe /rə/ is also typical of the Javanese implementation of Indic script, which however developed a rather different notation than Cham did for transcribing vowel /ə/ in other contexts.
75 Another heterograph, namely jmāy besides jmai, shows that graphic diphthongs were indeed perceived to transcribe vowel-consonant combinations.
76 For comparison, it is useful to consider some informed statements about the closely related Khmer script and its relation to Khmer phonology. See Jenner 2009: xi–xiv, Jenner & Sidwell 2010: 3–5. Almost all of what is said by these authors about the Old Khmer language and its writing system applies equally well to Old Cham language and script.
77 Groupe de recherche cam (1977: 250–252): “As E. Aymonier understood without daring to assert it, the three so-called ‘added’ consonants ꨒ ꨙ ꨡ are, in fact, nothing more than consonantal groups, existing in Old Cham, but whose memory present-day Chams have totally lost. These groups are each formed by the nasal followed by the corresponding voiced stop, which explains why today’s Chams classify these letters, without explaining it, immediately after the nasals.”
Auteur
École française d’Extrême-Orient
Arlo Griffiths received his PhD in Sanskrit philology from Leiden University, the Netherlands, in 2004. After holding a position as lecturer in Indian Religions at the University of Groningen, the Netherlands (2004–05), and holding the chair of Sanskrit at Leiden University (2005–08), he joined the École française d’Extrême-Orient (French School of Asian Studies, EFEO) in 2008 as Professor of Southeast Asian History, the position he still holds today. He is affiliated with the Centre Asie du sud-est (UMR 8170 EHESS/CNRS). He was posted at the EFEO’s Jakarta Center from 2009 through 2014, then based in Lyon, France, from 2015 through 2022, and since recently posted again at the EFEO’s Jakarta Center, starting in September 2022. In the framework of an institutional partnership, he teaches and supervises graduate students at the École Pratique des Hautes Études (EPHE) in Paris. From 2019 to 2025, he is serving as one of three principal investigators of a major project funded in the European Research Council’s Synergy scheme : DHARMA (The Domestication of “Hindu” Asceticism and the Religious Making of South and Southeast Asia). His current research focuses on textual materials in early languages of South and Southeast Asia (Sanskrit, Old Khmer, Old Cham, Old Javanese, Old Malay), both in inscriptions and in manuscripts, as source materials for exploring socio-religious and linguistic history.

Le texte seul est utilisable sous licence Creative Commons - Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International - CC BY-NC-ND 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.
La linguistique et ses formes historiques d’organisation et de production
Actes du colloque de la Société d’histoire et d’épistémologie des sciences du langage et du laboratoire Histoire des théories linguistiques, Paris, 24-26 janvier 2019
Didier Samain et Pierre-Yves Testenoire (dir.)
2022
Eugen Wüster et la terminologie de l’école de Vienne
Actes du colloque de la Société d’histoire et d’épistémologie des sciences du langage et du laboratoire Histoire des théories linguistiques Paris, 3-4 février 2006
Danielle Candel, Didier Samain et Dan Savatovsky (dir.)
2022
Simplicité et complexité des langues dans l’histoire des théories linguistiques
Actes du colloque de la Société d’histoire et d’épistémologie des sciences du langage et du laboratoire Histoire des théories linguistique, Paris, 23-25 janvier 2020
Chloé Laplantine, John. E. Joseph et Émilie Aussant (dir.)
2023
Documenter et décrire les langues d’Asie : histoire et épistémologie
Actes du colloque de la Société d’histoire et d’épistémologie des sciences du langage Paris, 26-28 janvier 2022
Émilie Aussant et Fabien Simon (dir.)
2025