The languages of Southeast Asia are known for their tones. Sure, there are notable exceptions such as the Austronesian languages spoken in the Malayan Peninsula, but for the majority of languages, including the Kra-Dai languages spoken in Thailand, Laos, and northern Vietnam, Hmong-Mien languages spoken in northern Vietnam and Laos, and the Sino-Tibetan languages spoken in Myanmar, these languages tend to have their own system of phonemic tones.

However, there is particular glaring exception to the arguable norm here. The Austroasiatic languages stand out amongst this linguistic landscape by mostly lacking a system of phonemic tones. And even more bizarrely, Vietnamese and some other Vietic cousins, which belong to that language family, have a system of phonemic tones. This leaves Khmer as the major Austroasiatic language without a tone system. With over 17 million native speakers, Khmer is the official language of Cambodia.

For the past several decades, linguists have been studying an interesting linguistic phenomenon in Khmer, along with other Khmeric languages spoken in Cambodia. The most recent publications have noted the rise of phonemic tone, a tonogenesis of some sort, in the variant of Khmer spoken in its capital, Phnom Penh.

Before we examine how this tone is characterised in Phnom Penh Khmer, we have to define what makes a minimal pair. Minimal pairs are pairs of words that only differ by a single sound. Compare the words “such” and “touch”. Phonologically speaking, these are minimal pairs as these words only differ in the initial consonant sound. Similarly, “broom” and “bloom” are minimal pairs as they only differ in the second consonant in the initial consonant cluster.

In Khmer, there may be minimal triplets present. In words containing a single syllable, minimal triplets may be distinguished by aspiration and a consonant cluster consisting of two consonants. In most cases affected by tonogenesis, the second element of the consonant cluster is the /r/ sound. Thus, you can have words like /ku:/ (pair), /kʰu:/ (old), and /kru:/ (teacher) in Standard Khmer.

But in Phnom Penh, the /r/ sound has sort of ‘transformed’ into something else. In 1966 and 1967, Noss and Huffman respectively have noticed that this sound has become a low-rising tone in Phnom Penh, while in 1994 and 1999, Pisitpanporn reported that the pitch contour was a falling-rising one, accompanied by aspiration. And in 2005, Wayland and Guion reported seeing both pitch contours occurring in Phnom Penh Khmer, with aspiration. Sometimes, this loss of /r/ sound also carries some vowel changes and vowel quality changes along with the pitch contours.

This lead to the 2014 study by Kirby, who aimed to characterise what exactly is happening with this loss of the /r/ sound, and to provide some insights on what could potentially be a case of tonogenesis happening in real time.

To approach this, he split his study into two major parts — one investigating how the sound change is produced, and one investigating how the sound change is perceived. All of his participants were from Phnom Penh, or Cambodians who have migrated to Phnom Penh early in their lives. I think that it would have been interesting to see how Phnom Penh Khmer and the Khmer from other regions of Cambodia would communicate, examining how the sound change is perceived by Khmer speakers from outside of Phnom Penh. Maybe that is a potential research avenue, although managing the logistics of that study would be more complicated than the one that was conducted.

In the production study, Kirby used a reading and repetition task approach. The first reading task involved reading a certain item in a frame sentence, which translates to “I say [word] again”, at a normal speaking rate. There were 41 items in total, which were presented in a randomised order. The repetition task involved an oral prompt in Standard Khmer, while the participant would respond with colloquial Khmer. Instead of the 41 items in the reading task, there were only 15 items used for the repetition task.

As we are particularly interested in the development of tone in Khmer, we would pay particularly close attention to the fundamental frequency trends yielded from these tasks. This frequency largely reflects the pitch used in speech, from which we could evaluate the presence of pitch contours. The general pattern Kirby observed was, the pitch contour seemed to be mostly low or low-falling, although in prosody, that is, speech rhythm patterns, this could develop from a falling-rising tone. By analysing other frequencies, Kirby was able to identify that this tone is accompanied by aspiration or breathiness in vowel quality, and some vowels tended to become diphthongs as well. However, this realisation of the /r/ sound seemed to vary from speaker to speaker.

For the perception part of the study, Kirby wanted to study how Phnom Penh speakers could distinguish differences in spoken forms like the tonal /kru:/ from /ku:/. There were two experiments conducted, one focusing on tonal distinction, and the other focusing on the role of breathiness in vowel and aspiration on this distinction. This made use of stimuli from a Phnom Penh Khmer speaker, and the participants were prompted to select with of the two lexical items the stimulus sounded like the most. Instead of using the Khmer script to represent the lexical items, despite all participants being literate in Khmer, Kirby used pictures for the options. A similar procedure was used for the second experiment, although the options for lexical terms were different.

So, what did Kirby find from these experiments? The main conclusion drawn was, Phnom Penh Khmer speakers were able to use the pitch contour as a cue to identify the /kru:/ forms from /ku:/. However, aspiration and breathiness did not seem to play much of a role in this distinction. Thus, this could be interpreted as sort of a rise of a phonemic tone, where tone is used to distinguish lexical (or grammatical) meaning.

A similar pattern of tonogenesis has also been found in the dialect of Khmer spoken in Kiên Giang province in southwest Vietnam, noted by Kirby in his abstract for a 2012 conference. Here, he attributes some part of tonogenesis to linguistic influence from Vietnamese, which has a much longer history as a tonal language compared to Khmer. Unlike Phnom Penh Khmer, /kru:/ is realised as a tonal /ku:/, with no aspiration nor breathiness attached to it. However, the tonal contour formed is largely similar to Phnom Penh Khmer.

Hyslop classified this sort of tonogenesis as a sonorant-conditioned one, and a similar phenomenon has been seen in a Tibeto-Burman language called Kurtöp, where tones developed following sonorant sounds. However, this did not incur a change in a sound like /r/ to reflect that tone. So, this /r/ sound change seems to be unique to Phnom Penh and Kiên Giang.

Whether or not this development of tone has been observed in other parts of Cambodia and Khmer-speaking regions in neighbouring countries remains to be studied. There has already been a case made for Kiên Giang Khmer, and has been remarked to be ‘more developed’ than in Phnom Penh. However, we do not know if this sound change could spread to other variants of Khmer as well. Additionally, we are unsure why this sound change occurred. Perhaps there have been building linguistic pressures by other tonal languages in the region, but questions would remain on why it has taken a considerably longer time to influence this sound change than say, a neighbouring Austroasiatic language like Vietnamese. Notwithstanding, this observation of a sound change is rather interesting, especially since Khmer seems to be traditionally an atonal language in a sea of otherwise tonal ones.

Further reading

Huffman, F. E. (1967). An outline of Cambodian grammar (Ph.D. dissertation). Cornell University.

Hyslop, G. (2022). Toward a typology of tonogenesis: Revising the model. Australian Journal of Linguistics, 42:3-4, 275-299.

Kirby, J. P. (2014). Incipient tonogenesis in Phnom Penh Khmer: Acoustic and perceptual studies. Journal of Phonetics, 43, 69-85.

Noss, R.B. (1966). The treatment of */r/ in two modern Khmer dialects. In: N. H. Zide (Ed.), Studies in comparative Austroasiatic linguistics (pp. 89-95). London: Mouton Co.

Pisitpanporn, N. (1994). On the r >h shift in Phnom Penh Khmer. Mon-Khmer Studies, 24, 105–113.

Pisitpanporn, N. (1999). A note on colloquial Phnom Penh Khmer. In: G. Thurgood (Ed.), Papers from the ninth annual meeting of the Southeast Asian Linguistics Society (pp. 243–248).
Arizona State University: Program for Southeast Asian Studies.

Wayland, R. P., & Guion, S. G. (2005). Sound changes following the loss of /r/ in Khmer: A new tonogenetic mechanism?. Mon-Khmer Studies, 35, 55–82.