1.1. Khmer in the landscape of Mon-Khmer languages
Ferlus (1979, 1992) distinguishes among Mon-Khmer languages those that are conservative (such as Khmu or Laven) and those that are innovative in terms of the evolution of initial stops. Khmer is among the innovative languages, as it developed a complex vowel system from the transfer of earlier oppositions on initial consonants -- the same historical source from which Vietnamese, another Mon-Khmer language, developed a two-way split of the tone system.
The innovative developments all constitute compensatory answers to the loss of the voicing feature of initial stop consonants. The three attested scenarios are (i) the development of constrastive voice-type registers (registrogenesis, see Jones 1986 and Matisoff 2001), (ii) the two-way split of the vowel system, or (iii) the two-way split of the tone system (Haudricourt 1965).
Khmer played a major role in the discovery of this fundamental phenomenon, which shaped the prosodic systems of most East and Southeast Asian languages. In a groundbreaking study, Henderson (1952) established the sequence of events from voicing oppositions on initials to phonation-type oppositions and finally to vowel contrasts
Initial consonant
|
Vowels (syllable rhyme)
|
Register
|
*unvoiced
|
Head voice / tense phonation
|
Higher pitch
|
Normal
|
Vowel opening
|
First (1) or loud set
Head register : "a normal or head voice quality" (Henderson 1952)
|
*voiced
|
Breathy voice / lax phonation
|
Lower pitch
|
Constriction
|
Vowel closing
|
Second (2) or low set
Voice register : "a deep rather breathy or sepulchral voice,… pronounced with lowering of the larynx" (Henderson 1952)
|
A more recent development, still undetectable in Henderson's time, is the development of contrastive tone from the reduction of initial consonant clusters: *kr and *kh merge to kh, and the earlier lexical oppositions are preserved as a tonal contrast (Kirby 2014).
Khmer bears traces of diverse influences (Higham 2002 : 14). Sanskrit and Pali principally gave cultural, religious and royal lexicon. Among colloquial registers, Thai, Lao, Vietnamese, Cham gave Khmer several words. According to Higham, "It has been shown that Indian religion, architecture, scripts and the Sanskrit language were incorporated into the indigenous cultures. This occurred among Mon speakers in the Chao Phraya Valley, Khmer speakers of the middle and lower Mekong and the speakers of Cham in central coastal Viet Nam." (Higham 2002 : 288; see also Lewitz 1967).
1.2. The special interest of the Cardamom dialect
Khmer is mainly spoken in Cambodia and close border areas (Surin Khmer dialect, Thailand and Krom Khmer dialect, Mekong River Delta in South Vietnam). The variety known as "Cardamom Khmer" differs notably from Central Khmer. According to Ferlus (1992), all Khmer dialects can be traced back to Middle Khmer: "For us, from the point of view of historical phonetics and within the limits of the present study, Middle Khmer begins at the start of the 16th century." / "Pour nous, d'un point de vue de phonetique historique et dans les limites de cette étude, le khmer moyen commence avec le XVIe siècle." (Ferlus 1992 : 58).

Cardamom Khmer dialect is spoken on both sides of the border between Western Cambodia and Eastern Central Thailand (Chanthaburi Province). It has a vowel system which remains highly similar with that of Middle Khmer. According to Martin (1992), following the conquest of the town of Longwek by Siam in 1591, the staff of the Court and inhabitants of Longwek took shelter in Cardamom mountains area. Since that time, Cardamom Khmer and Central Khmer are thought to have diverged gradually.
Cardamom Khmer has a distinction between first register (high register: slightly taut, modal voice, middle-range pitch, vowel opening) and second register (low register: slightly breathy, lower pitch, stable vowel quality). Cardamom Khmer thus preserves clear traces of the wide range of phonetic correlates of the register opposition, which was one of the main characteristics of Middle Khmer, and has been lost in many Khmer dialects, where vowel differences are now paramount (Ferlus, 1979, 1992; Henderson 1952; Martin 1975; Pain 2014). The Mon language as described by Shorto (1962) and Christian Bauer (PC) seems to be more conservative than Standard Khmer in this respect, preserving a register system with a broad range of phonetic correlates (voice quality, vowel quality, and some differences in pitch).
2. The archived documents: Marie Alexandrine Martin's audio recordings
Marie Alexandrine Martin (1932 - 2013) was a botanist and an ethnologist. She worked for the CNRS Research Center on East Asia at Paris Sorbonne University (CREOPS). Between 1965 and 1972, she conducted ethnobotanical and ethnolinguistic fieldwork in the Cardamom mountains (Khmer and Pear villages) and also with Khmer and Pear people from Thailand.
A selection of Marie Alexandrine Martin's bibliographic records can be found on the website of the AEFEK (Association d’Echanges et de Formation pour les Etudes Khmères). In addition to her scholarly publications, she wrote books for the general public based her observations about Cambodia as she experienced it from 1965 to the late 90's (Martin, 1989 and 1997).
The recordings made available here consist of a collection of audiotape records, collected by herself from the late 60’s and the middle 70’s. She entrusted a copy of her recordings to Michel Ferlus, who suggested that they should be digitized along with his own collections. Marie Alexandrine Martin’s heirs enthusiastically agreed to the archiving and online diffusion of her collection. Digitization was conducted by Alexis Michaud, and cataloguing by Michel Ferlus and Julien Heurdier. The set consists of several recordings of Khmer dialects and Pear languages (Pear, Chong, and Samre) spoken in or around the Cardamom mountains (West Cambodia and East Thailand). It also includes a short recording of Brao and Tampuan (Bahnaric languages of Ratanakiri province, North-East Cambodia).
The Khmer corpus of the Marie Alexandrine Martin's collection mainly contains recordings collected in the Cardamom area in two major spots:
1) in South-Western Cambodia:
- Koh Kong province, Thma Bang district (11°49’N – 103°25’E ; close to Pursat province), communes of Ta Tey Leu and Ruessei Chrum, which are in the middle of the Cardamom mountains. In 1998, Thma Bang district had fewer that 3000 inhabitants, which are largely rural-based and mostly rely on agriculture and forestry.
- Koh Kong province, Srae Ambel district (11°30’N – 103°10’E ; close to Kampot province), coastal commune of Chhkê Prus at the foot of the Damrei/Elephant mountains, a southern offshoot of the Cardamom mountains.

Map of Koh Kong province (from Mappery.com)
2) in Eastern Thailand, Chanthaburi province, Pong Nam Ron district (12°54’21’’N – 102°15’46’’E), Thap Sai subdistrict, commune of Thung Krang.
These recordings include narratives about agrarian works, local festivities or toponymy, and lexical fieldworks.
The set also contains one recording of Khmer Krom (a short linguistic enquiry), presumably collected in Vietnam (Mekong River Delta?), where the majority of Vietnam's ethnic Khmer population lives. Krom, in Khmer, means "lower" or "below" and formerly refers to the inhabitants from Kampuchea Krom, an area in southern Vietnam that was once part of of lower the Khmer Empire.
Marie Alexandrine Martin (who could speak Khmer) worked with a female Khmer-speaking collaborator. During lexical elicitation, she gives a Khmer version of an item as a prompt, and the Cardamom Khmer consultant says the word twice in Cardamom Khmer. Marie Martin's collaborator also acted as respondent during the collection of tales, to help rekindle the storyteller's eloquence.
A technical note: in cases where we were unable to identify the place of recording, the values indicated in the document's metadata are, by convention, those of a place in Cambodia (arbitrarily set at latitude: 11, longitude: 105 for Cambodia, and at latitude: 13, longitude: 102 for Thailand). We would be most grateful if visitors of this site could help us determine the present-day coordinates of these locations. Likewise, we hope to be able to add more information about participants: not all participants in the recordings have been identified so far.
The Khmer materials were digitized by the International Research Institute MICA (HUST – CNRS/UMI-2954 – Grenoble INP) in partnership with CNRS-LACITO, as part of the DO-RE-MI-FA project (Sept. 2014-Feb. 2016), funded by the Digital Scientific Library, a programme of the French Ministry for Higher Education and Research.
Last updated 2016.
This page was written by Julien Heurdier. Contact:Alexis Michaud