Sunday, November 16, 2025

Theory of Vietnamese Tonal Development

On Haudricourt's Tonogenesis 

by dchph





Vietnamese tones are often regarded as a defining feature of the language. Haudricourt’s groundbreaking theory (1954) argued that tones developed historically from the loss of final consonants and glottal features, rather than from any innate Austroasiatic inheritance. His insight reshaped comparative linguistics and clarified Vietnamese’s position in relation to Chinese and Mon–Khmer.

I) Haudricourt’s breakthrough (1954)

Haudricourt’s proposal that Vietnamese tones arose late, conditioned by the disappearance of final consonants, remains influential but requires close re‑examination. Evidence from Old Chinese, Middle Chinese, and early Sino‑Vietnamese loans suggests that tonal categories were already present in Vietnamese centuries before the twelfth century, in sharp contrast to his proposition. Tonality in Vietnamese developed in parallel with Chinese, reinforced by a millennium of close contact, and both inherited and regenerated itself according to the same phonological laws that governed tonal evolution in the first place.

Haudricourt’s hypothesis of tonogenesis in Vietnamese held that tonal distinctions arose from pitch changes conditioned by the nature of initial and final consonants. At the time, his proposal was strikingly innovative and has since become the foundational theory for explaining tonogenesis more broadly. Its impact on the study of Vietnamese tones has been profound and enduring, to the point that his view remains widely accepted among scholars.

This article, however, challenges his assertion that Vietnamese tones were not fully established until the twelfth century. Vietnamese tonality is better understood as both historical inheritance and innovation. Haudricourt’s postulation on the formation of tones in Vietnamese can only be sustained if the process is seen as unfolding simultaneously with parallel developments in Old Chinese, interactively and concurrently, as suggested by evidence of cognate correspondences.

II) Middle Chinese and Buddhist sources

Building on related ground, Mei Tsu‑lin, in Tones and Prosody in Middle Chinese and the Origin of the Rising Tone (1970), argued that the rising tone (上聲) of Middle Chinese emerged through the loss of a final glottal stop (‑ʔ). This development corresponds to the hỏi and ngã tones in Vietnamese and to comparable tonal categories in several modern Chinese dialects. By the seventh and eighth centuries, tonal distinctions were even employed to simulate the length contrast of Sanskrit, integrating seamlessly into the four‑tone system of Middle Chinese in terms of pitch, contour, and duration, as recorded in a ninth‑century Buddhist text. Furthermore, rhyming evidence from the Book of Odes indicates that Old Chinese words tended to cluster into three or four tonal categories, which later evolved directly into the four tones of Middle Chinese.

"Argument from analogy is that best suggestive, and without testimony from more direct sources, the theory will remain as one of the many possibilities. Fortunately, three kinds of evidence can now be presented: modern dialects, Buddhist sources bearing upon Middle Chinese, and old Sino- Vietnamese loans."

"Several dialects of the southeastern coastal area preserve a glottal stop in the rising tone, and the Buddhist sources indicate that the rising tone of Middle Chinese is high, short, and level. Our thesis, then, is that the final glottal stop of Old Chinese is retained intact in the coastal dialects and developed into a high and short syllable in Middle Chinese. We know from acoustic studies that a syllable is high and short if it ends in a voiceless stop, low and long if it ends in a voiced stop, and medium in pitch and duration if it is open [-Ø]. It is also reasonable to assume that when a final stop is lost, the tonal features are retained as reflexes. Therefore, if the final glottal stop (which is voiceless) indeed existed in Old Chinese, its descendant should have precisely the features we said the rising tone did have in the Middle Chinese."

(Mei Tsu-lin's Tones and Prosody in Middle Chinese and the Origin of the Rising Tone. 1970. )

III) Old Chinese evidence and early loans

Haudricourt emerged as the principal challenger to Maspero's 1916 theory of the non-inheritance of tones, which had been supported by Mon-Khmer non-tonal cognates. Prior to Haudricourt, the prevailing assumption was that tonal contrast could not be derived from non-tonal contrasts, and that tonality in Chinese was an intrinsic, original feature of the language. Even into the late twentieth century, Tung T'ung-ho (董同龢, "中國語音史", Zhōngguó Yǔyīn Shǐ. p. 183), as quoted by Mei Tsu-lin (1977) (1) , had also stated, "Ever since the beginning of the Chinese language, we not only distinguished tones, but possessed a tonal system not much different from the four tones of Middle Chinese." 

Haudricourt's 1954 study of Vietnamese tonegenesis overturned this view. He argued that the Chinese tonal system developed historically through the loss of certain final consonants. Specifically, the rising tone (上聲) of Middle Chinese, corresponding to the hỏi and ngã tones of Vietnamese, reflected an earlier final /-h/, itself a reflex of an original /-s/. Evidence for this comes from Chinese words borrowed into Vietnamese as early as the Han Dynasty, when the hỏi and ngã tones were still represented by /-s/. Examples include 義 ŋrals > Viet. nghĩa (ngã tone) and 墓 *ma:gs > Viet. mả (hỏi tone).

From this evidence and by analogy, Haudricourt further proposed that morphological derivation in Old Chinese involved alternation between a final /-s/ and its absence /-Ø/, which later gave rise to the departing tone (去聲) of Middle Chinese. For instance, he reconstructed /dâk/ 度 for the verbal form “to measure” (cf. SV độ) and /dâks/ for the nominal form "a measure" (cf. SV đạc). Similarly, /âk/ 惡 for the adjectival form "bad" (cf. SV ác) and /âks/ for the transitive verbal form "to dislike" (cf. SV ố). Mei noted that the second member of these pairs falls into the departing tone category, and indeed in Sino-Vietnamese they are consistently departing tones. However, in Sinitic-Vietnamese vernacular forms such as đo and dò, they appear instead in the level tone (平聲).

As for Haudricourt's hypothesis on the Vietnamese rising tones (hỏi and ngã), further discussion is required. Beyond the examples of 義 and 墓, additional cases may or may not conform to the paradigm he outlined, and these variations will be examined in subsequent sections.

Haudricourt's proposal was further developed by Forrest (R.A.D. Forrest, 1960), who equated the reconstructed -s of Old Chinese with the -s suffix of Classical Tibetan. Pulleyblank (E.G. Pulleyblank, "The Consonantal System of Old Chinese, Part II," Asia Major 9, 1962, pp. 206–265) added further support by pointing to foreign words ending in -s whose Chinese transcriptions, dated to the third century A.D., show -ts > -s, which in his theory gave rise to the departing tone. In the same study, Pulleyblank also proposed antecedents for two other tones: -ɗ and -ɓ for the later level tone (平聲), and -ʔ for the later rising tone (上聲). He argued that Old Chinese contained no open syllables [-Ø]. Having already reconstructed ɗ- and ɓ- as initials, he reasoned by symmetry that they could also occur in final position, i.e. -ɗ and -ɓ. Thus, what appears as an open syllable in Middle Chinese may in fact reflect Old Chinese finals -ɗ or -ɓ, depending on whether the syllable shows contact with a velar or dental coda. Pulleyblank's connection of -ʔ to the later rising tone was based largely on analogy with Vietnamese, given the striking parallels between the tonal systems of Vietnamese and Chinese. The steady accumulation of evidence for the -s hypothesis suggests that such analogies may indeed be valid. Since the sắc (') and nặng (.) tones of Vietnamese developed through the loss of an earlier -ʔ, it is plausible that the Chinese rising tone was derived in the same way.

In practical terms, the reverse logic has also been invoked: toneless words in Mon-Khmer languages that have tonal cognates in Vietnamese may in fact be Vietnamese loanwords. Haudricourt's theory of tonegenesis rested on the observation that certain Mon-Khmer finals correspond to specific tones in Vietnamese, for example, a final glottal stop [ʔ] correlating with the sixth tone. Yet his claim that Vietnamese only became fully tonal in the twelfth century is untenable. By the end of the tenth century, Tang Middle Chinese had already developed into an eight-tone system. For nearly a millennium, Annamese scholars used Mandarin as their medium of learning, just as scholars in other prefectures of the Middle Kingdom did. The pervasive presence of Tang Middle Chinese disyllabic vocabulary in Vietnamese further demonstrates that tonal development in Annam must have been complete by the time the country achieved full independence, otherwise there exists no Tang stanzas which have been widely appreciated and composed by Vietnamese scholars as recently as the early 1970s.

Whether or not Haudricourt's chronology is accepted, the evidence shows that Annamese words were already fully tonal by the thirteenth century. This is confirmed by the Annan Yiyu (安南 譯語, Annam Dịchngữ, "Translation of Annamese"), a wordbook compiled by a Chinese envoy to Annam during the Yuan Dynasty, later translated into modern language by Wang Li (王力, 1997).

Not only do contemporary Chinese loanwords in Vietnamese carry tones that correspond precisely with their Chinese counterparts, but older Sinitic‑Vietnamese forms also reveal tonal alignments traceable to Old Chinese, predating the full development of the eight tonal categories in Middle Chinese, for example:

  • OC *ɦljeds → VS thề (2nd tone) → MC dʑiaj → SV thệ (6th tone) → 誓 M shì ‘vow'

  • OC *ŋʷans → VS nguyền (2nd tone) → MC ŋuan → SV nguyện (6th tone) → 願 M yuàn ‘wish'

These correspondences extend to contrasts between voiced and voiceless initials, as seen in pairs such as:

  • buồng vs. phòng 房 fáng ‘room'

  • buồm vs. phàm 帆 fán ‘sail'

  • bữa vs. phạn 飯 fàn ‘meal'

When tonal factors are taken into account, the close alignment between Vietnamese and Chinese pronunciations in Sino‑Vietnamese and portions of Sinitic‑Vietnamese etyma suggests that many of these forms may have originated from shared sources in antiquity.

The model {C(+tone) : V(+tone) : MK(-tone)} illustrates how the three language groups – Chinese, Vietnamese, and Mon-Khmer – developed concurrently, in contrast to the hypothesis advanced by Austroasiatic Mon-Khmer theorists. Such a scenario could have unfolded more than 2,200 years ago, when early Annamese already possessed two or more tones comparable to those of Old Chinese. This was long before Annamese took shape as a distinct language, layering Sinitic elements atop an indigenous substratum, after its ancestral Vietic branch had split from proto-Viet-Muong. By the time local Muong tribesmen in northeastern Vietnam retreated into the mountains under pressure from the Han invasion of 111 B.C., the speech of those Kinh who remained had already become highly Sinicized. In contrast, the Muong, who mingled with Mon-Khmer speakers in the southwestern highlands, maintained closer contact with Austroasiatic traditions.

From an etymological perspective, when all elements are considered in historical context, the plausibility of Vietnamese and Chinese etyma being cognates expands to encompass a wide range of basic vocabulary. This includes items also found in Mon-Khmer languages, many of which have been classified as Austroasiatic lexicons–for example, chết 死 sǐ 'die', máu 衁 huáng 'blood', ngà 牙 yá 'tusk', and others. (see http://tlmei.com/tm17web/1976a_austroasiatics.pdf). 

On the Sinitic side, words sharing the same phonetic element eventually produced a profusion of homonyms in modern Mandarin, which today preserves only four tones after centuries of northern influence from Altaic languages such as Tartar and Manchu during their rule over parts or all of China. By contrast, the Annamese language mirrored the fuller tonal development of Southern Chinese lects such as Cantonese and Fukienese, both of which retained complete tonal systems.

The millennium of Chinese colonial rule, from the Han through the Tang dynasties until the 10th century, provided ample time for the eight-tone system of Middle Chinese to become firmly embedded in Vietnamese, where it remains an inseparable feature. The crucial point is that Annamese did not wait until the 12th century to complete its tonal system, as Haudricourt suggested. Rather, the tonal shifts that unfolded from Old Chinese through Early Middle Chinese into Middle Chinese were adopted in Annam contemporaneously, just as they were in other prefectures of the Middle Kingdom.

As a result, the eight-tone system in Vietnamese words of shared ancient roots allowed for precise differentiation of meaning among characters built on the same phonetic stem. For example, with the sound...

1.  口 kǒu (SV khẩu, VS cửa, 'opening')

  • ca 哥 (gē, 'brother')
  • ca 歌 (gē, 'sing')
  • cá 個 (gè, 'each', VS cái)
  • các 各 (gé, 'every')
  • cáo 告 (gào, 'announce')
  • cao 高 (gāo, 'high')
  • cảo 稿 (gǎo, 'manuscript')
  • cẩu 狗 (gǒu, 'dog', VS cầy)
  • cổ 古 (gǔ, 'ancient', VS cũ)
  • cô 姑 (gū, 'aunt')
  • cố 估 (gù, 'estimate')
  • cố 固 (gù, 'cause', VS cớ)
  • cú 句 (jū, 'sentence', VS câu)
  • cục 局 (jù, 'bending', VS cong)
  • cư 居 (jū, 'reside')
  • hà 何 (hé, 'how, which')
  • hà 河 (hé, 'river')
  • hồ 胡 (hú, 'neck', VS cổ)
  • hồ 湖 (hú, 'lake')
  • khả 可 (kě, 'able')
  • kha 珂 (kē, 'jade')
  • khắc 克 (kè, 'overcome')
  • khách 客 (kè, 'guest')
  • khấu 摳 (kòu, 'stingy', VS kẹo)
  • khấu 扣 (kòu, 'knock', VS gõ)
  • khô 枯 (kū, 'dried')
  • khổ 苦 (kǔ, 'bitter', VS cay, 'spicy hot')
  • khốc 哭 (kù, 'weep', VS khóc)
  • khốc 酷 (kù, 'brutal')
2. 方 fāng (SV phương, VS vuông, vửa, mới, 'square', 'recently')
  • bàng 旁 (páng, 'side')
  • báng 謗 (páng, 'slander')
  • biên 邊 (biān, 'border')
  • phòng 房 (fáng, 'room', VS buồng)
  • phóng 放 (fàng, 'release', VS buông)
  • phòng 防 (fáng, 'safeguard')
  • phỏng 仿 (fáng, 'imitate')
  • phỏng 訪 (fǎng, 'visit', VS thăm)
  • phương 芳 (fāng, 'fragrant', VS thơm)
  • phường 坊 (fáng, 'quarter', VS hàng) 
3. 工 gōng (SV công, 'work')cang 扛 (káng, 'carry', VS khiêng, gánh, gồng, cõng)
  • cang 缸 (gāng, 'vat', VS ảng)
  • công 功 (gōng, 'force')
  • công 攻 (gōng, 'assault')
  • cống 貢 (gòng, 'tribute')
  • củng 鞏 (gǒng, 'consolidate')
  • giang 江 (jiāng, 'river', VS sông)
  • hạng 項 (xiàng, 'nape, item', VS càng)
  • hồng 紅 (hóng, 'pink', VS hường)
  • hồng 虹 (hóng, 'rainbow', VS mống)
  • hồng 鴻 (hóng, 'swan', 'grand')
  • không 空 (kōng, 'empty', VS trống, rỗng)
  • khống 控 (kòng, 'control')
  • khủng 恐 (kǒng, 'terribly')
  • xoang 腔 (qiāng, 'hollow', 'accent', VS giọng)

4. 共 gòng (SV cộng, VS cùng, cung, cũng, vòng, 'add', 'common', 'together')

  • cảng 港 (gǎng, 'seaport')
  • cung 供 (gōng, 'supply, offerings', VS cúng)
  • cung 拱 (gōng, 'cup hands before the chest', VS vòng)
  • cung 恭 (gōng, 'respect')
  • hang 巷 (xiāng, 'alley', VS hẻm)
  • hồng 烘 (hōng, 'heat by fire', VS hong, hâm, hầm, hơ)
  • hồng 洪 (hóng, 'flood')
  • hồng 哄 (hòng, 'clamor', 'coax', VS hống)

The essential point is that all of these fundamental words, and hundreds of others, must have been pronounced with tonal distinctions long before the 12th century. Without tonal differentiation, they would have collapsed into homonymy. This demonstrates that Annamese had already developed a full tonal system well before Haudricourt's proposed timeline, regardless of whether one considers it a Sinicized language. That is to say, it inherited mainly at first from the Old Chinese source and continued on creating new ones, that is, new lexicon.

In addition, since Vietnamese is the only language in contrast to the Mon-Khmer group that developed a full tonal system, the argument can be reversed: loanwords from a tonal language such as Vietnamese, when transferred into toneless Mon-Khmer languages, would necessarily undergo morphemic innovation to compensate for the absence of tonal contrast. In such cases, pitch or intonational features may have been recruited to preserve distinctions, much as occurred with Chinese loanwords in other non-tonal languages like Japanese and Korean.

Koichi Honda, in his study, Tone Correspondences And Tonegenesis In the Vietic Family (Austroasiatic), approaches the entire issue from a Mon-Khmer perspective (emphasis by dchph):

"Vietic is known as the only sub-group in the Austroasiatic (Mon-Khmer) language family for having tones. Due to the existence of the tones, Vietnamese (or Viet), one of the members of the Vietic family, has long been discussed in terms of its position to which it belongs. In 1912, Maspero grouped Vietnamese as a member of Tai (Thai) languages, mainly because of its tones. Haudricourt, on the other hand in 1954, claimed it belongs to Mon-Khmer family, due to the correspondence of basic words, and posited a hypothesis which is called "tonegenesis". It seems Haudricourt's hypothesis is widely accepted by linguists. However, his hypothesis has not been well attested due to the scarcely obtainable materials for the comparative method."

And Honda summarizes Haudricourt's hypothesis as follows:

"The Vietic language did not have tones in the first stage around the year A.D. O. The birth of tones dates back to the 6th century, when a 3-tone system was established, depending on the syllable ending types: (1) open and sonorant-ending syllables became level tone; (2) fricative-ending syllables created falling tone; and (3) stop and glottal stop-ending syllables created rising tone (phonemicising of the rhymes to tones). The third shift took place in the 12th century where 3 tones split into 6 tones depending on the initial consonants; voiced ones became lower series of tones accompanied by the devoicing of initial consonants (phonemicising of the voiced initial consonants to tones). The last stage has been continuing to now where the devoiced initial consonants became voiced without changing their tones (voicing)." (Honda, p. 3)

Whether or not Old Chinese, under the same hypothesis outlined above, can be theorized to have developed into a four-tone system through such a phonemicizing process, it is crucial to note that prior to 111 B.C., before the Han Empire's annexation of the Nam Việt Kingdom, ancient Chinese loanwords, already complete with tonal distinctions, may have entered the earlier form of proto-Vietic. By the time Annam gained independence in the early 10th century, the region, having been a Chinese protectorate for nearly a millennium, had become a flourishing center of Tang-style rhymed poetry, second only to the Middle Kingdom itself. Without the already established eight-tone system inherited from Middle Chinese, it would have been impossible for Annamese poets of that era to compose Tang-inspired masterpieces without producing crippled imitations, lacking the tonal prerequisites essential to the tradition. (See Drake, F.S. ed. 1967. Symposium on Historical Archaeological and Linguistic Studies on Southern China, South-East Asia and the Hong Kong Region). In other words, Middle Chinese loanwords with their full eight-tone system must have entered Sino-Vietnamese well before the Middle Ages, already integrated into its two-register framework, so that eight tones were perceptually distinguished, even though only six are visibly marked in the later Vietnamese orthography, and "the register system is well reflected in the present day Vietnamese." (Honda, ibid, p 13)

For what Honda calls "specific irregular words in Viet" as he came along with his comparative work on data at hand, his postulation on some other factors seem to have influenced in the voicing in Vietnamese where they are devoicing. 

			Arem	    Ruc		       Muong	Viet
#7 "chicken" 		lakæ:	təlka:1, rəlka:1	ka	gà
#35 "rice (husked)" 	ŋkɔ:	təlkɔ:3, rəlkɔ:3	kaw.	gạo
"Our expected tone for #7 and #35 are ngang and hỏi tones respectively, both of which belong to high register. However, contrary to our expectations, both of them have low register tones with voiced onsets. Since these two words are so closely related to their daily life, it is hard to believe that only two of them developed in a different course. There must be some other factors for this irregularity. Another factor in common to the above two words is the initial consonant cluster, [tk] or [rk]. For reference, Ferlus' reconstructed forms for the above words are #7 *r-ka: and #35 *r-ko:ʔ This is a supporting evidence where initial consonant (or consonant cluster) has something to do with the voicing.

However, not all the reconstructed forms of Ferlus are reliable. Please look at the following example. 

		Arem	Ruc		Muong	Viet	Ferlus
#85 "near" 	-	təkiɲ1 or 2 	xəɲ`	gə`n	*t-kiɲ
#35 "sand" 	təka:c	təka:c3		kac´	kát	*t-ka:c

When I found word #85, I expected the initial consonant cluster *t-k is working in the same way as *r-k is doing. The expectation, however, was betrayed because of word #35. Word #85 is an old form of a loan word from Chinese called quasi-Sino-Vietnamese. Formal form of Sino-Vietnamese for this word is cận [kə.n]."

(Honda, p 13)

Could the so‑called “irregularities” and “faulty items” identified by Ferlus in fact be explained as Vietnamese loanwords into Mon‑Khmer languages? Terms such as gạo (from wet‑rice cultivation), gà (domesticated fowl), and cát (sand, as found in riverbeds or coastal shores) would not have been essential to the vocabulary of upland Mon‑Khmer montagnards. Perhaps the question itself would never have arisen if one accepts the straightforward anatomy of Vietnamese–Chinese cognates for these etyma, as follows:

  • gà 雞 (jī, 'chicken', SV kê) [See elaboration on the etymology in the previous section.]
  • gạo 稻 (dào, 'paddy, rice', SV đạo) [ M 稻 dào < MC dɑw < OC *lhu:ʔ ~ ɫhu:ʔ (Schuessler: MC dâu < OC *gləwʔ or *mləwʔ). See elaboration in the previous section.]
  • gần 近 (jìn, 'near', SV cận, cấn, ký) [ M 近 jìn (cận, cấn, ký) < MC gɨn < OC *ɡɯnʔ, *ɡɯns |  According to Starostin, also *gərʔ‑s; MC gyn; Mand. jìn 'to come near, keep close to'. In Vietnamese cf. gần 'near, close; adjacent, beside'. For etymology cf. 幾 *kəj 'near' (an old ‑r/‑l variation?). § 雞 jī (SV kê) 'gà'; 記 jì (SV ký) 'ghi'; 寄 jì (SV ký) 'gởi'; 急 jí (VS cấp) 'gấp'. ¶ j‑ ~ c‑(k‑). Note: 近 jìn ~ SV ký ~ VS kề 'close by'.]
  • cát 沙 (shā, 'sand', SV sa) [ M 沙 shā, shà, suō (sa, sá) < MC ʂaɨ < OC *sraːl, *sraːls | ¶ /sh‑, j‑, q‑ ~ k‑/. Examples: 尚 shāng ~ VS còn 'still'; 插 chā ~ VS cài 'stick in'; 擦 cā ~ VS cà 'rub'; 笑 xiào ~ VS cười 'smile'; 吉 jí ~ VS cát 'luck'; 旗 qí ~ VS cờ 'flag'; 棋 qí ~ VS cờ 'checker'. Alternatively, it may be cognate with sạn 'pebble' or 砂 shā.]

Additional considerations - Haudricourt's hypothesis is questionable for further reasons:

  1. As Honda (ibid.) has noted, the hypothesis is poorly attested due to the scarcity of reliable comparative materials.

  2. From the outset, the tonal table constructed by Haudricourt for comparison was flawed, diverging from the scheme traditionally employed in Chinese historical linguistics and widely adopted by subsequent philologists.

 It is not like this:

1. ­­ 3. ´ 5. ʔ 7. ´ -p, -t, -c, -ch
2. ` 4. . 6. ~ 8. . -p, -t, -c, -ch

(Sources: Norman. 1988, p. 55)

but it should be in correct alignment like this:

1. ­­ 3. ʔ 5. ´ 7. ´ -p, -t, -c, -ch
2. ` 4. ~ 6. . 8. . -p, -t, -c, -ch

Hence, according to Honda's observation that "(3) stop and glottal stop-ending syllables created rising tone (phonemicising of the rhymes to tones)." (Honda. Ibid.) , such a process should have given rise instead to the two departing tones (去聲) in Vietnamese, namely, the high [´] and low [.] registers, as classified in the tonal categories of the second table. These categories, faithfully preserved in Chinese historical linguistics and in the classic rhyme books, illustrate how ancient tonal schemes were devised and interpreted, and how the two‑tiered register system evolved out of the earlier three‑ and four‑tone system of Old Chinese.

Although the discussion concerns the emergence of three tones in Old Chinese, the distributional order of the pitch scheme in Middle Chinese is crucial, for it reflects how Chinese philologists of antiquity classified tones according to their contours as they appeared in Middle Chinese. Strikingly, this framework does not appear in Haudricourt's tonal schema, revealing a surprising gap in his awareness of the already well‑established Middle Chinese tonal system, even though he was otherwise one of the foremost Sinologists of the early twentieth century.

Moreover, regarding Honda's observation that "the last stage has been continuing to now where the devoiced initial consonants became voiced without changing their tones (voicing)" (Honda, p. 3), it should be noted that tonal change is not as rigid as Haudricourt's hypothesis suggests. Haudricourt attempted to correlate Vietnamese tonal registers directly with Old Chinese categories, for example, assigning sắc and nặng to shàngshēng (上聲 'rising') tones, and hỏi and ngã to qùshēng (去聲, 'departing') tones, and then reversing these correspondences for Middle Chinese. Such a mapping, however, oversimplifies the situation.

In reality, both initial and final consonants in Vietnamese words are distributed across all tones in both registers, low and high. For clarity, rather than using the traditional notation of tones 1-4 in two registers, the system here numbers them sequentially from 1 through 8 in paired sets. This reflects how the eight Vietnamese tones are classified (see the second table above) and provides a more precise framework for the discussion that follows.

As several philologists have noted, Haudricourt tended to see only one‑to‑one correspondences between certain initial or final consonants in Mon‑Khmer words and specific tones in Vietnamese. This limitation arose, first, because many of the Vietnamese words of Old Chinese origin in his list were drawn from a relatively small set of attested lexical items; second, because the comparative materials available at the time were scarce and earlier surveys incomplete; and third, because he was likely unaware of the much larger body of Vietnamese words of Chinese origin that exhibit multiple tonal layers, effectively allomorphs. These forms fall into his so‑called "last stage" of tonal development, but in fact they may have undergone repeated cycles of change in the distant past, distancing themselves considerably from their original shapes.

Moreover, one must also consider the possibility of loanwords moving in the opposite direction: Vietnamese items may have been borrowed into Mon‑Khmer languages, where speakers adapted them to fit a toneless phonological system, just as Chinese lexicons were adapted into Japanese and Korean, rather than the other way around.

Before proceeding further, it is useful to recall Haudricourt's hypothesis on the development of the six tones of modern Vietnamese, summarized in the following table (Honda, p. 2): 


			Haudricourt's hypothesis (1954)

	AD 0		6th Century	12th Century	Today
	toneless	3 tones		6 tones		6 tones

	pa		pa		pa		ba
	ba		ba		pà		bà
	pas, pah	pà		pà		bả

	bas, bah	bà		pã		bã

	paX, paʔ	pá		pá		bá
	baX, baʔ	bá		pạ		bạ

and here is the premise for that supposition:

"It was presupposed that the number of tones indicates chronological development of the Vietic family, i.e. from Arem (0 tone), Ruc (4 tones), Muong (5 tones) and Viet (6 tones). In the light of Haudricourt's hypothesis, Arem shows the first stage (AD.O), then Ruc just after the second stage (6th century), Muong around the third stage (12th century), and Viet the last stage (today)." (Honda, p. 4)

IV) Sino‑Vietnamese reflexes and tonal diversification

Let us examine the oft‑cited example of mả (rising tone) [ mis‑cited as mã ] from Haudricourt, which he presented as a cognate of Chinese 墓 mù = SV mộ. The latter is embedded in the nặng register, the 6th tone (陽去 yángqù, 'low departing tone'). In fact, this etymon has evolved into several distinct Sinitic‑Vietnamese and vernacular Vietnamese forms, each pronounced with different tones that are explicitly cross‑tonal and cross‑staged.

  • 墓 mù 'grave' (SV mộ) [M 墓 mù < MC muo < OC *ma:gs. | Chinese dialects: Cantonese mou6, Hakka mu5, Amoy boŋ6, Chaozhou mo4, Fukienese muo5, muoŋ5. According to Starostin, Standard Sino‑Vietnamese is mộ. Cf. also probable borrowings: ma 'funeral', mồ 'tomb'. For mh‑ cf. Amoy boŋ6, Chaozhou mo4, Fuzhou muo5, muoŋ5. GSR: 0802 f. Cf. môđất (土墓 tǔmù, 'mound'), maquỷ (魔鬼 móguǐ, 'ghost'), and machay (墓祭 mùjì, 'funeral ceremony').]

 From this root we find the following Vietnamese reflexes:

  1. mô (thanh ngang, 1st tone, 陰平 yīnpíng 'high level'), e.g., môđất 土墓 tǔmù 'earth mound'
  2. ma (thanh ngang, 1st tone, 陰平 yīnpíng 'high level'), e.g., thama #墓地 mùdì 'graveyard', machay 墓祭 mùjì 'funeral ceremony'
  3. mồ (huyền, 2nd tone, 陽平 yángpíng 'low level')
  4. mả (hỏi, 3rd tone, 陰上 yīnshàng 'high rising')
  5. mã (ngã, 4th tone, 陽上 yángshàng 'low rising'), Haudricourt posited VS mã < OC mâg instead of #4 mả < OC *ma:gs and #7 SV mộ < MC muo
  6. mố (sắc, 5th tone, 陰去 yīnqù 'high departing'), meaning 'bumper' (Fr. butée)
  7. mộ (nặng, 6th tone, 陽去 yángqù 'low departing')
  8. mốc (sắc nhập, 7th tone, 陰入 yīnrù 'high entering'), cf. biamốc (墓碑 mùbēi, 'gravestone')
  9. mộc (nặng nhập, 8th tone, 陽入 yángrù 'low entering'), cf. biamộc (墓碑 mùbēi, 'gravestone')

The form #(4) mả [mả], or even #(5) mã [ma4] (as cited by Haudricourt), is in fact a Sinitic‑Vietnamese reflex of the academic Sino‑Vietnamese word mộ [mo6], which is itself a clear cognate of Middle Chinese 墓 mù. In other words, the Vietnamese forms mồ, mô, and related variants all derive from the same root, or else represent Chinese loanwords that entered at later stages, possibly during Early Middle Chinese. These items may have been borrowed into Vietnamese at different historical periods, beginning as early as Old Chinese, with semantic development from mô 'mound' to mồ 'tomb'. Alternatively, they may share a common Yue source, perhaps reflected in a form like mã.

As Bo Yang (1983, vols. 1-2) notes, in the early Han dynasty mô denoted an 'earth lump', and graves of that period were flat and level. From this evidence, two scenarios emerge:

  1. Tonal correspondence: A relationship between voiced and devoiced initials or finals could have given rise to the 4th tone. In this specific case, {ʔ ⇒ ~}, e.g., OC mâg > mã, as Haudricourt postulated. This fits within his system of consonantal alternation producing tonal categories.
  2. Later voicing stage: As Honda (ibid., p. 3) observed, "the last stage has been continuing to now where the devoiced initial consonants became voiced without changing their tones (voicing)." This accounts for other cases where tonal distribution reflects subsequent phonological developments rather than Haudricourt's rigid correspondences.

In short, the aim here is to demonstrate that a single Chinese word could yield multiple reflexes in Vietnamese, each with different sounds and tones, regardless of whether the original Chinese form involved voiced or voiceless initials or finals (cf. Mei Tsu-lin's Tones And Prosody In Middle Chinese And the Origin of the Rising Tone. 1970.)

Much like the case of mã discussed above, the following examples further illustrate how multiple tonal shifts could develop from a single Chinese etymon, giving rise to several distinct Vietnamese reflexes – patterns that Haudricourt may not have fully recognized. The extended list serves two key purposes: first, to demonstrate that divergent tones had already taken shape well before the 12th century, functioning to mark subtle semantic distinctions; and second, to affirm their overwhelmingly Sinitic origins, in contrast to the comparatively scant Mon‑Khmer etymons, by showing how all derivatives remain systematically interconnected closely with the Chinese historical roots phonetically, tonally, and semantically, which the other side seriously lacks.

  • 母 mǔ (SV mẫu, mô) [ M 母  mǔ, mú, wǔ, wú < MC məw < OC *mɯʔ || Starostin: MC mʌw < OC *mǝ̄ʔ. For initial *m- cf. Min forms: Xiamen bo3, Chaozhou bo3, Fuzhou, Jianou mu3. | GSR: 0947 a-e || According to Thiều Chửu's Hán-Việt Dictionary: also SV 'mô': VS 'men', 'mẻ'. 母 mǔ ~ VS 'mái', 'cái' | § 海 hǎi (SV hải) | Example: 酵母 (jiàomǔ, VS menrượu), 母雞 (mǔjī, VS gàmái), 父母大王 Fùmǔ Dàwáng (VS Bốcái Ðạivương), 母系 (mǔxì, SV mẫuhệ), 繼母 (jìmǔ, VS mẹghẻ) ]
    1. VS men 'yeast' (thanhngang the 1st tone, 陰平 yīnpíng, 'high level tone'),
    2. VS me 'mother' (the 1st tone),
    3. SV mô 'mold' (the 1st tone),
    4. SV mỳ 'venter' (the 2nd tone),
    5. VS mẻ 'female elder' (hỏi, 3rd tone, 陰上 yīnshàng 'high rising')
    6. SV mẫu 'mother' (ngã or the 4th, 陽上 yángshàng, 'low rising tone')
    7. VS mái 'female of animal' (sắc or the 5th, 陰去 yīnqù, 'high departing tone'),
    8. VS nái 'female of animal' (the 5th tone) [ ex. 'heonái' (母彘 mǔzhí, 'female pig') ],
    9. VS cái 'female' (the 5th tone) [ ex. 'con dại cái mang', literally translated, '子呆母忙 zǐ dài mǔ máng'. cf. 海 hăi (SV 'hải' /ha̰ːj/) VS 'khơi' ]
    10. VS mạ 'mother' (nặng or the 6th tone, 陽去 yángqù, 'low departing tone'),
    11. VS mệ 'mother' (the 6th tone),
    12. VS mẹ 'mother' (the 6th tone),
    13. VS mợ 'aunty' (the 6th tone)
  • 梅 méi (SV mai) [ M 梅 méi < MC moj < OC *mɯː| According to Starostin, Japanese apricot (Prunus mume), plum. Viet. 'me' has a narrowed meaning 'tamarind' (cf. Chin. 酸梅 'tamarind', lit. 'sour plum'). An older loanword is probably Viet. mơ 'apricot'. The regular Sino-Viet. reading is mai. For *m- cf. Min forms: Xiamen m2, Chaozhou bue2, Fuzhou muoi2, Jianou mo2. ] we have:
    1. VS mai 'plum' (thanhngang or the 1st tone, 陰平 yīnpíng 'high level tone')
    2. VS me 'tamarind' (thanhngang or the 1st tone, 陰平 yīnpíng 'high level tone')
    3. VS mơ 'apricot' (the 1st tone)
    4. VS muội 'salted dried plum' (nặng or the 6th tone 'low departing tone') [ ex. 'xímuội' 鹹梅 (xiánméi, 'preserved salty plum') vs. 'ômai' (烏梅 wūméi, 'black preserved salty plum') ]
  • 海 hǎi (SV hải) [M 海 hǎi < MC həj < OC hmlɯːʔ |  Etymological: § 母 mǔ > mái > mệ > mể ~ QT 海 hǎi ~ bể > biển. ¶ Sound correspondences: /h‑ ~ m‑/, /m‑ ~ b‑/. | Dialectal reflexes : Cant. hoi2, Hakka hoi3, Teochew hai2. Related forms: For khơi (open sea), note the alternation ¶ /k‑/ ~ /h‑/. Phonological note: 開 kāi (SV khai) in Hainanese is /k'uj1/, while in Cantonese it is /hoj1/.  Cf. 悔 huǐ (SV hối), 況 kuàng (SV huống). Examples: 出海 (chūhǎi, VS rakhơi), 海外 (hǎiwài, VS ngoàikhơi), Interchange "bể" ~ "biển": Compound words illustrate the sound‑change patterns:
    • 大海 (dàhǎi, VS bểcả, biểncả)
    • 苦海 (kǔhǎi, VS bểkhổ, khổải)
    • 海浪 (hǎilàng, VS sóngbể)
    • 海口 (hǎikǒu, VS cửabể)
    • 海寇 (hǎikòu, VS cướpbể)
    • 海賊 (hǎizéi, VS giặcbể)
    • 海域 (hǎiyù, VS vùngbiển) ]
  1. VS khơi 'sea' (the 1st tone, 陰平 yīnpíng 'high level tone')
  2. SV hải 'sea' (hỏi, 3rd tone, 陰上 yīnshàng 'high rising')
  3. VS bể 'sea' (the 3rd tone),
  4. VS biển 'sea' (the 3rd tone),
    • Toponymic usage: 北海道 Běihǎidào (Hokkaidō) → SV Bắchảiđạo.
    • Comparative notes: Cf. 繁 (緐) fán, pó (SV phồn, bàn), 敏 mǐn (SV mẫn), 每 měi (SV mỗi), 梅 méi (SV mai).
    • 溟 míng (SV minh) [ M 溟 míng, mǐng, mì (SV minh, mình, mịch) < MC mieŋ < OC *meːŋ, *meːŋʔ | Etymologically, it is thought by classical commentators to be the same word as 冥 (OC *meːŋ, 'dark', 'black(of water)') (likely in light of the parallelism with the unrelated 海 (OC *hmlɯːʔ, 'sea', 'ocean') < 晦 (OC *hmɯːs, 'dark')). Schuessler (2007) proposes that there's an outside chance this can be instead connected with Proto-Tibeto-Burman *mlik, whence Old Burmese မ္လစ် ('river'), Burmese မြစ် (mrac), Rakhine (mreik, 'sea'), Daai Chin [Term?] (mlik (tui), 'big water, river, sea')]
    1. SV minh 'dark' (thanhngang, the 1st tone)
    2. VS mênh 'expanse' (thanhngang, the 1st tone)
    3. VS mưa 'crizzle' (thanhngang, the 1st tone)
    4. SV mình 'muddle' (huyền, 1st tone)
    5. VS mùng 'vast' (huyền, 1st tone)
    6. VS mờ 'indistinct' (huyền, 1st tone)
    7. VS bể 'sea' (hỏi, the 3rd tone)
    8. VS biển 'ocean' (hỏi, the 3rd tone)
    9. SV mịch 'dull' (nặngnhập, the 8th tone)
    10. VS mịt 'obscure' (nặngnhập, the 8th tone)
    • 放 fàng (SV phóng) [ M 放 fàng < MC pwoŋ < OC *paŋʔ, *paŋs | According to Starostin, to put away, put aside; neglect; banish. In Viet. cf. also a colloquial word: phỗng 'to take away, to carry away'. | ¶ /f- ~ b-/ : Ex. 房 (fáng, SV phòng, VS buồng ('room') ], we have:
      1. VS bỏ 'discard' (hỏi the 3rd tone, 陰上 yīnshàng 'high rising tone')
      2. VS phỗng 'take away' (ngã or the 4th tone')
      3. SV phóng 'release' (sắc or the 5th, 陰去 yīnqù 'high departing tone')
      4. VS buông 'let go' (thanhngang or the 1st, 陰平 yīnpíng 'high level tone')
      5. VS bắn 'shoot' (sắc or the 5th tone')
          • 會 huì (SV hội, cối) [ M 會 huì, kuài, guā, guài, huǐ, guì (hội, cối) < MC kwaj, ɦuɑi < OC *ko:bs, *go:bs ], we have:
          1. VS hay, 'aware' (thanhngang, the 1st tone)
          2. SV hồi 'festival' (huyền or the 2nd tone)
          3. VS hiểu 'understand' (hỏi or the 3rd tone); (cognate to or an alternation of the modern Mand. 曉 xiáo 'know', 'understand', SV hiểu)
          4. VS đỗi, 'moment' (ngã or the 4th tone)
          5. VS sẽ, 'will' (ngã or the 4th tone)
          6. SV hội 'festival' (nặng or the 6th tone 'low departing tone')
          7. VS hụi 'loan' (nặng or the 6th tone, from Fukienese or Amoy)
          8. VS họp 'meeting' (nặngnhập or the 8th, 陽入 yángrù 'low entering tone')
          9. VS hẹn 'dating, appointment' (nặng or the 6th tone)
              • 賊 zéi (SV tặc) [ M 賊 zéi < MC dzək < OC *zɯːɡ | Etymology: Possibly Sino-Tibetan; compare Tibetan ཇག (jag, 'robbery') (Coblin, 1986). Schuessler (2007) points out that a palatalized consonant in Tibetan does not usually correspond to an unpalatalized one in Chinese; instead, he compares it to Khmer ឆក់ (chɑk, 'to snatch; to steal');  Based on evidence from early loans from Chinese, e.g. Lakkia kjak⁸ ('bandit') and Rục kəcʌ́ːk ('bandit'), Baxter and Sagart (2014) reconstructs the Old Chinese with a *k preinitial.]
                1. VS chích 'burglar' (sắcnhập or the 7th tone, 陰入 yīngrù 'high entering tone'); as in đạochích: 盗賊 dàozéi 
                2. VS cắp 'steal' (the 7th tone);  as in đánhcắp: 盗賊 dàozéi 'steal'
                3. SV tặc 'enemy' (nặngnhập or the 8th, 陽入 yángrù 'low entering tone'),
                4. VS giặc 'enemy' (the 8th tone),
                • 粉 fén (SV phấn) [ M QT 粉 fěn, fèn < MC pun < OC *pɯnʔ | Dialects: Minnan, including Hainanese hun2, Amoy hun2, Chaozhou huŋ21, Fuzhou xuŋ2 | According to Starostin, the later (and usual) meaning is 'flour'. The word is also used in compounds meaning 'noodles', thus it seems possible that Viet. bún 'vermicelli' is an independent loan from the same source. | ¶ /f- ~ ph-, b-, v-/ ]
                1. VS phở 'noodle soup' (the 3rd tone)
                2. SV phấn 'powder' (nặng or the 5th tone)
                3. SV phớn 'powder' (the 5th tone)
                4. VS bún 'rice vermicelli' (the 5th tone)
                5. VS bột 'flour' (nặngnhập or the 8th, 陽入 yángrù 'low entering tone')
                6. VS bụi 'dust' (the 8th tone) [ Probably associated with 灰 huì (SV muội) ]
                7. VS vụn 'shard' (the 8th tone) 
                  • 照 zhào (SV chiếu) [ M 照 zhào < MC tɕiɛu < OC *tjews ]
                  1. VS soi 'look at the mirror' (thanhngang the 1st tone)
                  2. VS noi 'follow' (the 1st tone)
                  3. VS theo 'according to' (the 1st tone)
                  4. SV chiếu 'reflect' (sắc or the 5th, 陰去 yīnqù 'high departing tone')
                  5. VS chói 'reflect' (the 5th tone)
                  6. VS chụp as in 'chụphình' 'take picture' (nặngnhập or the 8th tone')
                  7. VS rọi 'shine' (nặng or the 6th tone)
                  8. VS dọi 'shine' (nặng or the 6th tone)
                    • 染 rǎn (SV nhiễm) [ M  染 rǎn, ràn < MC ȵiam < OC *njomʔ, *njoms | According to Starostin, 'be soft'. Somewhat later (since late Zhou) the character was also used for a homonymous *nam (~-emʔ) 'to dye, smear; ('dye' <) infect' (with a variant *namʔ-s, MC ɲe\m). Viet. nhiễm is a standard reading; there also exists a colloquial loan nhuộm 'to dye'. Coblin (1986) compares this to Tibetan ཉམས་པ (nyams pa, “be stained, tarnished, spoiled”); Pan (1987) also notes Proto-Tai *ɲuɔmᴬ as well as Vietnamese nhuộm, both meaning "to dye". Schuessler (2007) cites Downer (1986)'s opinion that form with 上 (shàng) tone is the verb, while form with 去 (qù) tone is the noun meaning "kind of cloth" (Lǐjì). ]
                    1. VS lây 'contagious' (thanhngang or the 1st tone)
                    2. VS sang 'spread a virus' (the 1st tone)
                    3. VS lem 'stain' (the 1st tone)
                    4.  SV nhiễm 'extract a disease, habit' (the 4th tone)
                    5. VS vẩn 'smear' (the 4th tone)
                    6. VS nhuốm 'extract a disease' (sắc or the 5th, 陰去 yīnqù 'high departing tone')
                    7. VS nhuộm 'dye' (the 5th tone)
                    8. VS ruộm 'dye' (the 5th tone)
                    9. VS nhẹm 'tarnish' (the 5th tone)
                    10. VS mắc 'get sick' (sắcnhập or the 7th tone)
                    • 深 shēn (SV thâm) [ M 深 shēn (thâm, thẩm) < MC ɕim < OC *hljum, *hljums | Etymology: Most of dialects read /sjəm1/ | ¶ /sh- ~ đ-/ : Ex. 燒 shāo (SV thiêu, VS đốt, 'burn'). Unger (1995) suggests that 深 (shēn) may have had an Old Chinese initial n‑, based on the phonetic evidence of 淰 (shěn), whose component 念 (niàn*) points in that direction. Schuessler (2007) reconstructs 深 as OC nhəm and proposes connections to several Tibeto‑Burman forms: Mizo hniam 'to be low, to sink into (land)', Burmese နိမ့် (nim.) 'low', Tangkhul Naga kʰənim 'to be humble', and Tibetan ནེམས་ (nems) 'sink a little, give way'. He traces these to Proto‑Tibeto‑Burman nem 'low', which STEDT further derives from Proto‑Sino‑Tibetan *s‑n(i/u)(ː)p/m ~ r/s‑nyap/m 'pinch, squeeze; press, oppress; submerge, sink, west, low, soft'. If so, 深 (OC nhəm) ultimately reflects a Proto‑Sino‑Tibetan root. Schuessler also notes the similarity between 深 (OC nhəm) and 沉 (OC d‑ləm), which he interprets as an areal etymon. In the Rites of Zhou, 深 (OC nhəms) 'depth' appears as a nominal derivation with the suffix ‑s, which yielded the departing tone (MC ɕiɪmH). By contrast, 淰 (OC nhəmʔ) 'to be startled and flee (of fish); to sink into the deep' (in the Liji) represents an endoactive derivative with ‑ʔ, which produced the rising tone (Mandarin shěn). ]
                    1. SV thâm 'profound' (thanhngang or he 1st tone)
                    2. VS sâm 'deep' (the 1st tone)
                    3. VS sâu 'deep' (the 1st tone),
                    4. VS sẫm 'dark' (ngã or the 4th tone)
                    5. VS thẩm (hỏi or the 5rd tone)
                    6. VS thắm 'profound' (sắc or 5th tone)
                    7. VS đậm 'dark' (nặng or the 6th tone, 'low departing tone')
                    8. VS sậm, 'dark' (the 6th tone)
                    • 扛 káng (SV cang) [ M 扛 káng, gāng < MC kaɨwŋ < OC *kroːŋ | cf. cõng, gánh, gồng, chống: M 抗 kàng (SV kháng) 抗 kàng, káng < MC  kʰaŋ < OC *ɡaːŋ, *kʰaːŋs  ]
                    1. SV cang 'carry' (thanhngang or the 1st tone, 'high level tone'),
                    2. VS khiêng 'carry on one's shoulder' (the 1st tone)
                    3. VS gồng 'to shoulder' (huyền or the 2nd tone 'low level tone')
                    4. VS cõng 'carry on one's back' (ngã or the 4th tone 'low rising tone')
                    5. VS gánh 'carry on one's shoulder' (sắc or the 5th, 陰去 yīnqù 'high departing tone'),

                    • 蟲 chóng (SV trùng) [ M 蟲 chóng < MC ɖuwŋ < OC *l'uŋ, *l'uŋs | From Proto-Sino-Tibetan *djuŋ ('insect'; 'bug') (STEDT). According to Starostin: 'insect', 'small bird' Used also for a homonymous *ɬhuŋ (nóng) 'be hot (of weather)'. Standard Sino-Viet. is trùng. ] 
                    1. VS giun 'earthworm' (thanhngang or the 1st tone, 'high level tone')
                    2. VS sâu 'insect' (the 1st tone)
                    3. SV trùng 'insect' (huyền or the 2nd tone, 'low level tone'),
                    4. VS trùn 'earthworm' (the 2nd tone),
                    5. VS sán 'worm' (sắc or the 5th, 陰去 yīnqù 'high departing tone')
                    6. VS nóng 'hot' (the 5th tone)
                    • 種 zhǒng (SV chủng) [ M 種 zhǒng, zhòng, chóng (chủng, chúng, chùng) < MC tɕiowŋ < OC *tjoŋʔ, *tjoŋs | Etymology: Sino-Tibetan 'Chepang' (tuŋʔ-, 'to plant'), दुङ् (duŋ, 'shoot; sprout'), दुङ्‌सा (duŋ-, 'to sprout; to grow'). Compare 腫 (OC *tjoŋʔ, 'to swell') and 踵 (OC *tjoŋʔ, 'heel'). Related to Proto-Vietic *k-coːŋʔ ('seed') (Vietnamese giống ('seed')), which is likely a loanword from Chinese (Wang, 1948). Pronunciation 2 (trồng, 'to sow'; 'to plant') is the exoactive derivation of pronunciation 1 (giống, 'seed'). According to Starostin: seeds; cereals. Also read *toŋʔ-s, MC couŋ (FQ 之用), Mand. zhòng 'to sow'. The word also means 'kind, sort, race' ( > 'seed'), which is reflected in a colloquial Viet. loanword (from another dialectal source) giống 'kind', 'sort'; 'race', 'breed', strain'. For this word, An Chi (Ibid. 2016. Volume 2) boldly posited it as 'trứng' (egg) that ought to be 蛋 dàn (SV đản), which is the case that Vietnamese scholar still in the mentality of trying to match word-by-word phonologically. ]
                    1. VS trồng 'to plant' (huyền or the 2nd tone, 'low level tone', read 'zhòng' in Mandarin)
                    2. VS dòng 'breed' (the 2nd tone)
                    3. SV chủng 'type' (hỏi or the 3rd tone)
                    4. VS giống 'strain' (sắc or the 5th tone, 陰去 yīnqù 'high departing tone')
                      • 臭 chòu (SV xú, khứu, khứu) [ M 臭 chòu, xìu (xú, khứu) < MC tɕʰiəu < OC *kʰljus | Etymology: Schuessler (2007) considers it to be cognate with 犨 (OC *kʰju, 'sound of an ox breathing') and connects it to Burmese ဟိုက် (huik, 'to pant'). Also compare 朽 (OC *qʰluʔ, 'to rot; to decay') (Baxter and Sagart, 2014). According to Starostin, in MC there also exists a reading xjəw (Mand. xiu) (Jiyun); it is interesting to note that standard Sino-Viet. renders it as khưu. These are most probably dialectal variants of the original *khiw-s which gave the standard MC reflex chjəw (note that Viet. thiu 'stale' is a colloquial reflex of the latter; the standard Sino-Viet. form is xú.]
                        1. VS hôi 'smelly' (thanhngang or the 1st tone)
                        2. VS ôi 'rotten' (the 1st tone)
                        3. SV khưu 'smell' (the 1st tone) 
                        4. VS thiu 'stale' (the 1st tone) [ doublet 餿 sòu ]
                        5.  VS ngửi 'to smell' (hỏi or the 3rd tone) [ doublet of M 嗅 xìu < MC xǝ̀w < OC *xus ]
                        6. VS hửi 'to smell' (the 3rd tone)
                        7. VS hủi 'rotten' (the 3rd tone)
                        8. SV xú 'bad smell' (sắc or the 5th tone)
                        9. SV khứu 'smelling sense' (the 5th tone)
                        10. VS thối 'foul' (the 5th tone)
                        11. VS thúi 'foul' (the 5th tone)
                        • 按 àn (SV án) 'case' [ M 按 àn < MC ʔɒn < OC *ʔa:ns || Comments: cf. 案 àn (SV án, VS bàn, 'table'), '安心 ānxīn (VS yêntâm, 'not worry')' ~ '放心 fàngxīn (vữnglòng, 'feel reassured') | ¶ /Ø- ~ y-, nh-, b-,  f-, v-/ ]
                        VS ịn 'press' (nặng or the 6th tone),
                          1. VS nhồi 'to stuff' (huyền or the 2nd tone)
                          2. SV án 'press' (sắc or the 5th tone)
                          3. VS ấn 'press' (the 5th tone)
                          4. VS nhấn 'press' (the 5th tone)
                          5. VS bấm 'press' (the 5th tone)
                          6. VS nhận 'stuff' (nặng or the 6th tone)
                          • 利 lì (SV lợi) [ M 利 lì (lợi, lị) < MC li < OC *rids | Dialects: Amoy li32 (lit.); lai32, Hai. lai32, Cant.: lei32 | Etymology: From Proto-Sino-Tibetan *ri:t ('to reap, scrape, shave, cut, sever') (STEDT, Schuessler, 2007); cognate to Mizo rîit ('to scrap with a hoe'), Western Gurung (wriqba, 'to scratch'), Burmese ရိတ် (rit, 'to cut', 'reap', 'mow', 'shave').]
                            1. SV lị 'benefit' (nặng or the 6th tone, 'low departing tone')
                            2. SV lợi 'advantage' (the 6th tone)
                            3. VS lãi 'profit' (ngã or the 4th tone, 'low rising tone')
                            4. VS lời 'profit', 'interest' (huyền or the 2nd tone, 'low level tone')
                            5. VS lẽm 'sharp' (ngã or the 4th tone)
                            6. VS lém 'witty' (sắc or the 5th tone)
                          and the list goes on and on. There are too many of them to list!

                          It should be noted that the preceding list contains only monosyllabic items. If each of these were placed into formations yielding disyllabic words, such as those etyma built with the morphemic syllable 海 hǎi exemplified above, the roster would likely expand to colossal proportions.

                          From the deliberately extensive set of examples, three points stand out. First, the phenomenon of multiple sound changes radiating from a single root is not unique to Vietnamese but is equally attested in Chinese, where a single character may bear several tonal values within one dialect. Second, the correspondences between Old Chinese tones and those of Sino‑Vietnamese are varied and complex, extending far beyond the narrow set of Vietnamese words with the 3rd and 4th tones (hỏi and ngã) that Haudricourt hypothesized to derive from final ‑ʔ in certain Mon‑Khmer languages on a one‑to‑one basis. Third, and most importantly, if ancient Annamese tones had only reached full development by the 12th century, as Haudricourt proposed, it would be impossible to account for the one‑to‑many Chinese‑Vietnamese monosyllabic reflexes that had already emerged in much earlier periods.

                          In general, tonal changes in the Sino‑Vietnamese lexicon have occurred both diachronically and synchronically. A single Chinese character or word could generate multiple Vietnamese reflexes, each accented with distinct tonal contours. The case of 墓 mù (grave) illustrates this: OC mâg > mã versus ma:gs > mả. Such variation is abundant throughout the examples already cited and elsewhere in this study. Moreover, tonal differentiation often carries subtle semantic distinctions. For instance, Vietnamese Phật for 佛 Fó (Buddha) coexists with Bụt, which also denotes 'Buddha' but shades toward the semantic field of Sino‑Vietnamese thánh (聖 shèng, 'saint') and thần (神 shén, 'god').

                          In this light, Haudricourt's proposal that Vietnamese tones developed from an absence of tone to partial formation by the 12th century, and only later reached their present six [eight] tones, appears internally inconsistent for several reasons as some have been briefly summed up previously, to be elaborated later on. At this stage, it suffices to note that his theory rests on a limited set of correspondences between toneless Mon‑Khmer and Vietnamese basic vocabulary. Yet many of these fundamental words have demonstrable cognates in Chinese, a language that is highly tonal. Examples such as 口, 方, 母, and 海 show that Vietnamese pronunciations of Chinese characters, as preserved in Sino‑Vietnamese readings, can be verified against the definitions and fǎnqiè (反切) keys recorded in sources compiled in the Kangxi Zidian (康熙字典). Consider 母, for example:

                          【說文】蜀人 (dchph: ngườiThục) 謂 母 ("mẹ") 曰 姐 ("chế")，齊人 (ngườiTề) 謂 母 曰 嬭 ("nạ")，又 曰 㜷 ("mợ")， 吳人 (ngườiNgô) 曰 媒 ("mẹ")。 【眞臘 (Chânlạp) 風土記】呼 父 為 巴駞，呼 母 為 米 ("mẹ")。 方音 不同，皆 自 母 而 變。 又 乳母 亦 曰 母 ("mụ, vú")。 【越語 Việtngữ 】生 三 人，公 與 之 母 ("mẹ")。 又 禽獸 之 牝 皆 曰 母 ("mái")。

                          English Translation: "Shuowen": The people of Shu (ngườiThục) called 'mother' jie ('chế'); the people of Qi (ngườiTề) called 'mother' nai ('nạ'); they also said mu ('mợ'); the people of Wu (ngườiNgô) said mei ('mẹ').  "Record of Customs of Zhenla" They called 'father' batuo, and 'mother' mi ('mẹ'). The regional pronunciations differ, all deriving from mu [M 母 mǔ, mú, wǔ, wú (mẫu, mô) < MC məw < OC *mɯʔ]. A wet nurse was also called mu ('mụ', 'vú').  "Yueyu" When three children were born, the Duke shared them with their 'mother' ('mẹ'). Among birds and beasts, the female was also called mu ('mái').

                          This evidence underscores the tonal and semantic breadth of Sino‑Vietnamese correspondences, extending far beyond the narrow scope envisioned in Haudricourt's original hypothesis.

                          The author maintains that there is no need to employ electronic instruments to measure tonal intensity when comparing tonal correspondences between Vietnamese and the Chinese dialects. Such similarities can be reliably assessed by the human ear alone. For example, Mandarin 入 rù /ʐu⁴/ is realized as /zuː⁵/, in Sino‑Vietnamese as nhập /ɲəp8/, and in Cantonese as /jap⁶/. These forms clearly demonstrate cognate tonal values, without requiring mechanical validation of decibel precision. Vietnamese speakers, like other Taic peoples, are capable of reproducing 'Chinese' sounds and tones with remarkable accuracy, without perceptible deviation. Thus, Vietnamese /rú/ aligns closely with Mandarin 入 rù /ʐu⁴/, while nhập /ɲəp8/ corresponds to Cantonese dập (the local rendering of /jap⁶/). In each case, the tonal contours and values remain consistent across the cognate forms, something that speakers of most non‑Sinitic languages would find difficult to replicate.

                          One of the principal reasons for rejecting Haudricourt's hypothesis of Vietnamese tonegenesis is that the modern Vietnamese tonal system aligns so closely with the Middle Chinese tonal scheme. This is further reinforced by the nine inherited tones of modern Cantonese (formerly known as the Tang language, 唐話 T'ongwa), whose tonal formation must have been complete long before the 10th century, the very period marking Annam's political separation from China. The strict rhyming matrices of Tang poetry, still observed in Vietnamese Tang‑style verse, testify to this continuity (see Drake, ibid.; Sung Shee, ibid.). In modern Vietnamese, tones are organized into four categories across two registers (traditionally counted as eight, or six in modern orthography, which omits the two entering tones 入聲 Rùshēng). Strikingly, when compared with the nine tones of Cantonese, the degrees of tonal intensity are virtually identical, with the Cantonese ninth tone corresponding to Vietnamese syllables ending in ‑p, ‑t, ‑k, or ‑ch, spoken with level intonation.

                          The question of whether tonality is acquired or inherited can be clarified by comparison. Japanese and Korean, like Vietnamese, borrowed vast numbers of Chinese words. Yet despite their deliberate adoption of Chinese vocabulary, neither language preserved tonal distinctions. Japan, for instance, began sending students to the Tang court in the 9th century, importing Chinese literary culture and loanwords with Kan‑on, Tō‑on, and Go‑on readings, but without tones (Bo Yang 1983). Similarly, Korean absorbed large numbers of Chinese words, especially from the Ming period, again without tones (An Chi, ibid., vol. 3, p. 284). These parallels demonstrate that tonality is not the product of acquisition but of inheritance. No matter how intensively a language borrows, tones cannot be artificially imposed. The situation is analogous to modern Chinese borrowings into English, terms in biotechnology, high‑tech, or trade espionage retain their English phonology without tonal adaptation. Vietnamese words borrowed into English likewise appear toneless (Vietcong, Vietminh, Ho Chi Minh, Nu‑yen [Nguyen], bunmee, banhmi, aodai, pho, dong), further illustrating the point.

                          V) Comparative perspectives and critiques

                          Austroasiatic Mon‑Khmer languages are toneless, while Chinese, Vietnamese, and Mường are tonal, as are related families such as Tai‑Kadai, Thai, Lao, and Hmong. The prevailing Austroasiatic view that many Vietnamese–Mon‑Khmer cognates originated from Mon‑Khmer underpins Haudricourt's theory of tonegenesis. He argued that Vietnamese was originally toneless, deriving from Mon‑Khmer, and only later developed tones. The author proposes the reverse: Vietnamese loanwords in Mon‑Khmer became toneless through the latter's contact with Mường dialects. Unable to reproduce embedded tones, Mon‑Khmer speakers stripped them away, compensating through other phonetic adjustments. Over time, this produced paradigms such as the tonal pattern {~ → ʔ}. This explanation is more consistent than assuming tones were generated ex nihilo in Vietnamese.

                          Maspero, in fact, was correct in asserting that tonality cannot be acquired. Middle Vietnamese must already have had a full tonal system, given the long‑standing adoption of Chinese tonal words, especially from Middle Chinese during the Tang Dynasty, when Annamese officials in Chang'an spoke the same Mandarin as their Chinese counterparts. Sino‑Vietnamese pronunciations are simply variants of Tang phonology (Nguyen Tai Can 2000, ibid.). It is inconceivable that Annamese could have borrowed Middle Chinese words without their tones, since tones were morphemically embedded in every syllable. To accept Haudricourt's chronology would require believing that Middle Chinese loanwords entered Vietnamese stripped of tones, as in Japanese and Korean, an implausible scenario.

                          Moreover, ancient Annamese words were already attested with eight tones. For example: vuông ('square') ~ 方 fāng → SV phương; VS buồng ('room') ~ 房 fáng → SV phòng; VS buông ('let go') ~ 放 fàng → SV phóng; SV phỏng → 仿  fǎng ('imitate'). In earlier stages, labial‑dental initials /v‑/ and /f‑/ did not yet exist, so vuông may have been buông, and phóng related to bắn ('shoot'), just as buồng corresponds to phòng. The crucial point is that sounds and meanings were already differentiated by at least four tonal categories in Old Chinese: (1) level tone, (2) lower level tone, (3) departing tone, and (4) low departing tone. This is consistent with other cases such as 墓 (mô, mồ, mộ) and 母 (me, mẹ, mái).

                          It is untenable to suggest that the phenomenon of tonality in Vietnamese arose as late as the 12th century. Such a claim is both illogical and inconsistent with the evidence. As previously noted, the highly Sinicized vocabulary of ancient Annamese already demonstrates that most Vietnamese words must have acquired tones contemporaneously with Old Chinese. During the Eastern Han (東漢, 25–220 A.D.), Viceroy Sĩ Nhiếp (士攝, 187–226 A.D.) established schools and promoted Han culture in Giao Chỉ Prefecture (交趾部), raising the study of Chinese to unprecedented prominence. This fact underscores that Han language learning had been firmly rooted in Annam for at least 250 years since the annexation of the NamViet Kingdom in 111 B.C.

                          VI) Implications for tonogenesis

                          Vietnamese tones as both inherited innovations and independent creations; methodological lessons.

                          If we were to strip away the tones from Sino‑Vietnamese words of Old Chinese origin, the resulting forms would collapse into homonymy, unlike the Kan‑on borrowings in Japanese, where tones were never adopted. For example, Japanese ichi for 一 /yi¹/ contrasts with Sino‑Vietnamese nhất and Cantonese jat⁵. The Vietnamese evidence instead points to the opposite process: tones were inherited directly from Old and Middle Chinese. Without them, semantic differentiation would have been severely weakened, as seen in words with Mandarin /yáng/ like dê (羊 yáng, SV dương, 'goat'), giông (颺 yáng, SV dương, 'windstorm'), rượng (痒 yáng, SV dương, 'aroused'), nắng (暘 yáng, SV dương, 'sunshine').

                          The Japanese case further illustrates the weakness of Haudricourt's hypothesis. He proposed that tonality developed from morphemic contrasts, but Japanese and Korean, despite massive borrowing from Chinese, never acquired tones. By contrast, Vietnamese did, which suggests inheritance rather than innovation. Consider words such as mặc ('put on'), chồmhỗm ('squat'), and chànghãng ('stand with legs apart'), often claimed to be of Mon‑Khmer origin. Even if their roots lie in Mon‑Khmer forms like /pec/, /chrohom/, or /choho/, the tonal system in which they are embedded is distinctly Vietnamese. Internally, the tonal contrasts are obvious: mặc, mắc, mắt, mặt, etc. Even without diacritics, Vietnamese speakers instinctively supply tonal contours, much as Cantonese speakers do with their entering tones (thanhsắcnhập, thanhnặngnhập), peculiarly in syllables ending in /‑p/, /‑t/, /‑k/, or /‑ch/.

                          Haudricourt also overlooked the many Vietnamese words carrying thanhsắc [5] and thanhnặng [6] that differentiate meaning in subtle ways, as in mô vs. mộ, me vs. mẹ, tan vs. tán. Therefore, his claim that Vietnamese only became fully tonal after the 12th century fails to capture the reality: tonality in Vietnamese developed in tandem with Chinese, from Old Chinese through Middle Chinese, as reflected in the phonological record. This trajectory is consistent with other southern Chinese languages that followed the same evolutionary path.

                          Given that Annamese scholars had been studying official Mandarin continuously from the beginning of the Common Era until at least 939 A.D., it is inconceivable that their language lacked tones during this period. Ancient Annamese must already have possessed the full eight tonal categories of Middle Chinese long before the 10th century.

                          There is nothing remarkable about an Asian linguist writing on topics in English linguistics. Yet when a Western linguist happens to publish something on Vietnamese, it is treated as a sensation. The situation is almost comical, reminiscent of Vietnamese or Chinese variety shows on VTV, CTV, or YouTube, where local audiences erupt with applause, cheers, and exaggerated delight whenever a Westerner, often an exchange student, steps on stage and delivers a few simple jokes in Vietnamese or Chinese.

                          What explains the enthusiasm of many Vietnamese specialists for Haudricourt's theory? Part of the answer lies in a long‑standing cultural tendency: out of a sense of inferiority, Vietnamese intellectuals have often displayed excessive admiration for Westerners who demonstrate even modest fluency in Vietnamese. What is celebrated as a positive openness to Western scientific methods is, in practice, also a reflection of pride that their language has attracted attention from foreign scholars, an attitude shaped as much by a superiority complex as by deference.

                          In reality, most Western linguists possessed only limited fluency in Vietnamese and relied primarily on the analysis of available data. Some achieved a degree of technical mastery through academic training, but they often lacked what might be called "linguistic feeling".  (2). Imagine the scenario: a Western linguist publishes a proposal on Vietnamese linguistics, and immediately a chorus of local admirers responds with "wows" and "hoorays".

                          Haudricourt's own listings, as cited by Shafer (1972) and discussed in What Makes Chinese So Vietnamese? - Chapter 10 on Sino‑Tibetan etymologies, raise doubts about his ability to distinguish Sinitic‑Vietnamese from Sino‑Vietnamese forms. His confusion of VS mã (which should properly be mả) with SV mộ is a telling example. Yet his disciples repeated such errors uncritically, and the flaws went largely unnoticed because the hypothesis appeared to fit neatly into the intellectual climate of the time. Even today, some Western linguists continue to perpetuate the same mistakes as their predecessors (see Ding Bangxin, ibid., 1977, p. 263).

                          Historically, by the 9th century Vietnam had already become a distinguished center outside China that produced prodigious poets skilled in composing Tang‑style stanzas. It is inconceivable that speakers of a supposedly toneless ancient vernacular could have fully appreciated or mastered such poetry (see Drake, F.S., ed. 1967, ibid.). Even today, many Vietnamese poets continue to compose in Tang‑style forms, and the practice of chanting these poems remains popular among the public, an appreciation that, ironically, is now far less common among the general populace in China itself.


                          Was Vietnamese originally toneless, or did it possess two or three tonal distinctions–low and high–like certain Mon‑Khmer languages, as Haudricourt proposed for the 12th‑century timeframe? For those seeking a quick answer, the reality is that many Vietnamese speakers find it difficult to accept the notion that their tonally rich folksongs, fixed expressions, proverbs, and idioms, believed to date back to antiquity, were once toneless. Each Vietnamese word is intimately tied to a corresponding Sinitic‑Vietnamese etymon, making the idea of a toneless origin implausible (see the author's Hán‑Nôm etymology dictionary at http://vny2k.com/hannom/ for detailed entries and meanings). For example:

                          • Bốcái Ðạivương;
                          • Con dại cái mang;
                          • Gậtđầu lắccái;
                          • Chácđược củarẻ;
                          • Chồng chúa vợ tôi;
                          • Cõng rắn cắn gà nhà;
                          • Giặc đến nhà đànbà phải đánh;
                          • Ăn coi nồi, ngồi coi hướng;
                          • Một miếng khi đói bằng một gói khi no;
                          • Nghèo cho sạch, rách cho thơm;
                          • Giấy rách phải giữ lấy lề;
                          • Nghêdại chẳng hay cóc;
                          • Tiên học lễ, hậu học văn;
                          • Đi một đàng, học một sàng khôn;
                          • Bỏ thì thương, vương thì tội;
                          • Không mợ thì chợ vẫn đông;
                          • Có mống tựnhiên lại có cây;
                          • Rán đàngđông vừa trông vừa chạy;
                          • Nựccười châuchấu đáxe;
                          • Bàcon xa khôngbằng lánggiềng gần;
                          • Thương cho roi cho vọt, ghét cho ngọt cho ngào;
                          • Cảnh cũ non quê nhặt chốc mòng;
                          • Quântử hãy lăm bền chí cũ;
                          • Cổ tới nhẫnkim, sinh thời có hoá;
                          • Thà làmquỷ nướcNam cònhơn làmvua nướcBắc;
                          • Bầu ơi thươnglấy bí cùng, người trong một nước phải thươngnhau hoài;
                          • Nhiễuđiều phủlấy giágương, người trong một nước phải thươngnhau cùng,

                            • etc.

                          and so forth with hundreds of similar folk‑styled expressions, whose rhyming stanzas and lyrical structures clearly indicate that they must have existed long before the 12th century, even if some examples were later recorded in 14th‑century works. In the centuries preceding that period, the Vietnamese not only absorbed the Chinese language of the mandarins from the Han through the Tang dynasties – the basis of Sino‑Vietnamese (Hán‑Việt) – but also composed rhymed prose and poetry of notable literary elegance (Drake, F.S., ed. 1967, ibid.). Could such achievements have been possible if their speech were still toneless, like Khmer? If so, how would they have addressed their beloved Princess Huyền Trân (Huyềntrân Côngchúa)? How were historical names pronounced at that time?

                          The author therefore challenges Austroasiatic Mon‑Khmer theorists to reconstruct even a fraction of these folklores, idioms, and proverbs using a hypothetical toneless lexicon, of which many in Vietnamese are proved cognate to those of the Chinese, say, 'lárụngvềcội' (落葉歸根 luòyèguīgēn, SV lạcdiệpquycăn, 'a falling leaf returns to the roots') or 'bỏlàngquêxưa' (背井離鄉 bèijǐnglíxiǎng, SV bộitỉnhlyhương, 'tear oneself away from one's native place'). In short, it is inconceivable that ancient Vietnamese remained partially toneless until the 12th century.

                          Lest any contemporary Vietnamese philologist forget, or fail to appreciate, the simple expressions cited above, they must first deepen their knowledge of Vietnamese before presuming to write anything serious about the language. More importantly, lest newcomers be led astray, we must guard against their retracing the worn paths of early pioneers in Vietnamese linguistics, that is, paths that obscure access to Sino‑Tibetan etymologies and the overwhelmingly Sinitic strata of modern Vietnamese.

                          Turning specifically to Haudricourt's claim that the eight Vietnamese tones were not fully formed until the 12th century: such a view is highly improbable. It is difficult to imagine that the tonal system would have reached completion only two centuries after Vietnam had already secured independence. If his theory had instead placed tonal formation as far back as the 2nd century B.C. – when the 3rd and 4th tones in Old Chinese were beginning to evolve – it might have been more plausible. 

                          In fact, the Vietnamese huyền [`] and sắc [´] tones predate their Sino‑Vietnamese first‑tone doublets, which means that by then the full set of tonal registers (1st through 4th, each in upper and lower registers) was already in place. Examples include:

                          Early tone contrasts (pre‑Tang)

                          Vietnamese pair Tone value Chin.     Pinyin     Meaning
                          mồi (lower 1st = modern 2nd) vs.
                          môi (upper 1st = modern 1st)                           		huyền vs. ngang     煤     méi     coal
                          mối (upper 3rd = modern 5th) vs.
                          môi (upper 1st)                          		 hỏi vs. ngang     媒     méi     go‑between
                          dì (2nd) vs. di (1st) huyền vs. ngang     姨     yí     aunt
                          lấm (5th) vs. lâm (1st) hỏi vs. ngang     淋         lín     soak


                          Established 8‑tone system (by end of Tang)


                          Vietnamese pair         Tone value Chin.     Pinyin     Meaning
                          dâm (1st) vs. dầm (2nd)         ngang vs. huyền         淫 yín     wet
                          mả (3rd) vs. mã (4th)         hỏi vs. ngã     墓 mù     tomb
                          bố (5th) vs. phụ (6th)         sắc vs. nặng     父 fù     father
                          mắt (7th) vs. mục (8th)         sắcnhập vs. nặngnhập     目 mù     eye


                          Register splits

                          Vietnamese pair         Register contrast          Chin.      Pinyin     Meaning
                          buồng vs. phòng     voiced vs. voiceless initials             房     fáng     room
                          đục vs. trọc     voiced vs. voiceless initials             濁     zhuó     murky

                          On the broader question of whether languages are 'tonal' or 'non‑tonal', it is reasonable to assume that most human languages began with simple utterances and only later developed either tonal systems or complex consonant clusters (kl‑, kr‑, bl‑, etc.). This is evident in Daic languages of southern China or in tonal Tibetan of Lhasa (see Ding Bangxin, ibid., 1977, p. 263), as opposed to the monosyllabicity of Old Chinese or proto‑Vietic (l‑, s‑, tr‑, etc.), which later differentiated tonally in Sino‑Vietnamese and Middle Chinese. If Vietnamese had truly evolved from a purely toneless Mon‑Khmer root, then many inherited Mon‑Khmer words should have remained toneless. Yet in Vietnamese they are tonal, raising the question: why would tones have been added if they were unnecessary? The existence of tonal correspondences in scores of Mon‑Khmer–Vietnamese cognates suggests otherwise. It is inconceivable that Vietnamese folk songs, proverbs, and lyrical traditions that are so deeply tonal were ever performed in a flat, toneless manner.

                          As argued earlier, Haudricourt's theory seems to invert the actual process: rather than Vietnamese developing tones late, it is Mon‑Khmer languages that lost them. His postulation excludes the multiple tonal contours of Vietnamese words that are clearly cognate with Chinese etyma. The sheer volume of Chinese‑origin vocabulary in Vietnamese at every historical stage is sufficient to show that his theory requires revision. Like other Yue‑descendant languages (Zhuang, Daic, etc.), Vietnamese tonality evolved in parallel with Chinese, beginning before the Han Dynasty and continuing through a millennium of colonization after 111 B.C. During this period, Vietnamese tones developed within the Chinese linguistic sphere, as reflected in etymological cognates across the region.

                          From 111 B.C. to 939 A.D., Vietnam experienced long stretches of colonization punctuated by brief intervals of independence, mirroring the rise and fall of Chinese dynasties. This history makes clear that Vietnamese tonality was never an isolated phenomenon, but part of the same continuum as Old and Middle Chinese. If Haudricourt's hypothesis were correct, then all Mon‑Khmer languages should have become tonal as well. Yet only a few developed two or three tones, much like Old Chinese during the Han, with its basic level, high, and low intonations. Vietnamese, by contrast, had already inherited and elaborated a full tonal system long before the 12th century.


                          In comparison, within less than three centuries of Spanish colonial rule in Latin America, the Indigenous peoples became minorities while racially mixed Spanish‑speaking populations emerged as the dominant majority. This historical reality is recalled here as a reminder intended to stir those idle, nationalistic minds into reconsidering their assumptions.

                          In practice, most Mon‑Khmer speakers tend to neutralize or omit the tones embedded in Vietnamese loanwords when speaking, as can be heard in their Mon‑Khmer‑accented Vietnamese across the southernmost and western highland provinces of Vietnam. By contrast, the Kinh Vietnamese consistently apply tonalization to all foreign borrowings. This is most evident in loanwords from French and English, such as xìcăngđan ('scandal'), xìtăngđa ('standard''), or quánhhtùtì ('one, two, three'). The same process naturally accounts for the tones assigned to Mon‑Khmer borrowings for local objects such as bòhóc ('prahok' fish paste), sàrông ('sarong') as well as to placenames like Buônmêthuột (Buon Ama Thuot), Đắklắk (Daklak), and Đàlạt (Dalat).

                          If Mon‑Khmer and Vietnamese cognates truly shared the same linguistic root and phonological system, native speakers would have used them without modification, just as Thai, Lao, or Hmong speakers employ cognates within the Daic family without adding extraneous phonemic features. Put differently, if Vietnamese had originally been a toneless Mon‑Khmer‑type language, its speakers would not have needed to 'retrofit' tones onto loanwords in order to compensate for a lack of tonality. For this reason, Haudricourt's argument that Vietnamese tonal development was independent of Chinese influence must be rejected. If that were the case, Vietnamese speakers could just as easily have stripped tones from Chinese borrowings, as Japanese and Korean speakers did.

                          At the same time, Vietnamese preserves Old Chinese loanwords that already exhibit the four tonal categories Haudricourt associated with a 12th‑century 'tonegenesis'. These may in fact represent ancient tonal strata preserved in Central Vietnamese dialects, particularly in the Thanhhoá-Thuậnhoá region and especially Nghệ an. For example, chimchóc ('birds') contains chóc, cognate with ancient 雀 què [cf. modern SV tước, VS sóc 'sparrow']. When the Trần Dynasty acquired southern Chamic territory through the marriage of Princess Huyền Trân to the King of Champa Chế Bồngnga (Po Binasuor), Annamese settlers carried with them these fossilized four‑tone systems, which still survive in northern Central dialects. This helps explain why, like Chế Bồngnga (Cei Bunga), originally toneless Chamic placenames were fully tonalized in Vietnamese, e.g., Đànẵng, Quynhơn, Nhatrang.

                          Further south, earlier Annamese resettlers also transmitted distinctive consonantal features that preserve keys to the Early Middle Chinese phonological system, including initials such as 知 zhī (SV trí), 徹 chè (SV triệt), 澄 chéng (SV trừng), 于 yú (SV vu), and 匣 xiá (SV hiệp) (see Ding Bangxin, ibid., 1977, pp. 266–269).

                          Bernhard Karlgren in his research entitled Tones in Archaic Chinese (1960) reached the conclusion that Archaic Chinese of pre-Han periods after analyzing rimes in Shijing Odes that

                           "Archaic Chinese, like Ancient and Modern Chinese, had distinct tone classes. One of them corresponded to the p'ing-sheng of the Ancient Chinese, another to the shang-sheng, another to the k'ü-sheng and the last (words ending in p-, t-, k-) to the ju-sheng (..) [His conclusion that] Archaic Chinese had tone classes roughly corresponding to those in Ancient Chinese [..] that words figuring in the 'pure one-tone sets of rimes' in all probability belonged to the Arch. tone class corresponding to the Anc. class concerned (p'ing, shang, k'ü, ju). (p. 133)"

                          We may therefore state with confidence that when the Han conquered ancient Annam in 111 B.C., the tonal system of Ancient Chinese profoundly shaped the local speech, leaving an imprint that endured well beyond 939 A.D. The Vietnamese did not need to wait until the 12th century to begin accentuating their language with tones; tonal influence was already deeply embedded. Indeed, Vietnamese tonality corresponds in every respect to the tonal categories of Chinese dialects.

                          In conclusion, the sheer volume and vitality of Sino‑Vietnamese vocabulary – thousands of items actively used in everyday speech, not merely the etyma catalogued in dictionaries – provides sufficient evidence for this claim. Moreover, the vernaculars spoken by the common people in the Southern Han period, across the mountainous regions south of Lingnan from Jiaozhou (Giaochâu) to Guangzhou, must have been broadly similar (see Bo Yang, ibid., vols. 71-72, 1993). And one must ask: would any historical linguist seriously argue that the ancestors of today's nine‑toned Cantonese were speaking with only four or five incomplete tones as late as the 10th century?

                          Conclusion

                          This article's emphasis on Haudricourt's theory of Vietnamese tonal development demonstrates that Vietnamese tones are both primordial inheritances and historical innovations, born from the transformation of final consonants and voicing distinctions into suprasegmental contours traceable to Old Chinese. In this process, Vietnamese did not merely echo lost consonants and inherit the resulting tones; it simultaneously created a new tonal system, as in the case of 墓 mù (mộ, mô, mồ, mố, mả, ma), which stands as an independent development within Southeast Asia. This dual lesson – that similarity across families is plausible, and that phonological creativity can reshape language history – makes Vietnamese tonogenesis a model case for comparative linguistics and for the broader understanding of how languages both inherit and reinvent themselves.

                          Concluding point - Dual lesson: similarity across families is plausible, but phonological creativity reshapes language history.


                          References

                          Foundational Works

                          • Maspero, Henri. Études sur la phonétique historique de la langue annamite. Les initiales. Paris: Imprimerie Nationale, 1912.

                          • Maspero, Henri. Grammaire de la langue chinoise. Paris: Imprimerie Nationale, 1921.

                          • Haudricourt, André‑Georges. L’origine des tons en vietnamien. Journal Asiatique 242, 1954, pp. 69–82.

                          • Haudricourt, André‑Georges. Problèmes de phonétique diachronique: la nasalisation vocalique en vietnamien. Bulletin de la Société de Linguistique de Paris 49, 1954.

                          Tonogenesis and Comparative Studies

                          • Mei, Tsu‑lin. Tones and Prosody in Middle Chinese and the Origin of the Rising Tone. Harvard Journal of Asiatic Studies 30(1), 1970, pp. 86–110.

                          • Pulleyblank, Edwin G. Middle Chinese: A Study in Historical Phonology. Vancouver: University of British Columbia Press, 1984.

                          • Chen, Shu‑Fen. Vowel Length in Middle Chinese Based on Buddhist Sanskrit Transliterations. Language and Linguistics 4(1), 2003, pp. 29–45.

                          Vietnamese and Austroasiatic Context

                          • Thomas, David D. Basic Vocabulary in Some Mon–Khmer Languages. Mon–Khmer Studies, 1960.

                          • Sidwell, Paul. Austroasiatic Dataset for Phylogenetic Analysis: 2015 Version. Mon–Khmer Studies 44, Mahidol University / SIL International.

                          • Alves, Mark J. An Updated Overview of the Austroasiatic Components of Vietnamese. Languages 9(12), 2024.

                          • Alves, Mark J. Linguistic Research on the Origins of the Vietnamese Language: An Overview. Journal of Vietnamese Studies 1(1–2), 2006.

                          Sino–Vietnamese layering

                          • Sa, Quoc Hoang. Study on the Understanding and Use of Sino‑Vietnamese Words: Perspectives from Secondary School Students in Ho Chi Minh City. Sprin Journal of Arts, Humanities and Social Sciences 4(5), 2025.

                          Comparative methodology

                          • Campbell, Lyle. Historical Linguistics: An Introduction. Edinburgh University Press, 2013.

                          • Haspelmath, Martin. Comparative Linguistics and the Problem of Spurious Similarities. Linguistic Typology 9(1), 2005.


                          FOOTNOTES

                          (1)^ Also, see the Chinese translation by Huang Xuan-Fan 黃宣範. "中古 漢語 聲調 與 上聲 的 起源" (in <中國 語言 學論集>. 1977. pp. 175-197

                          (2)^ Without the mastery level of "linguistic feelings" that a specialist needs with near native level of the target language due to lack of first-hand experience in modern Chinese, both standard and colloquial, they would never know the roots of many Vietnamese words such as:

                          - 'đầunậu' (ring leader) 頭腦 tóunăo (SV đầunão),
                          - 'dàydạn' (experienced) 經驗 jīngyàn (SV kinhnghiệm),
                          - 'láibuôn' (merchant) 大販 dăipán (Cant. /tai2pan3/),
                          - 'lẻtẻ' (trivial) 零星 língxīng (SV linhtinh, 'miscellaneous'),
                          - 'ănnhậu' (social engagement) 應酬 yìngchóu (i.e., 'eat and drink'),
                          - 'cụngly' (raise glasses and cheers) 碰盃 bèngbèi,
                          - 'đừnghòng' (don't you ever) 甭想 péngxiăng,
                          - 'luônluôn' (always) 牢牢 láoláo,
                          - 'lạcloài' (solitude) 落落 luòuò [ ~ '失落 shìluò (SV thấtlạc) ],
                          - 'đượclắm' (pretty good) 得來 délái,
                          - 'đượclòng' (pretty good) 心得 xīndé,
                          - 'giờgiấc' (time) 時間 shíjiān [ while 'thuở (thủa)' (a period of time), a contraction of phonetic sandhi of 時候 shíhòu (SV thờihậu) ],

                          all that match exactly the same usage and meanings of the Chinese counterparts, not to mention in-depth knowledge required for the Chinese phonological historical linguistics to appreciate the roots of basic lexicons such as

                          - 'chỉ' 線 xiàn (thread) and 'chỉ' 錢' (ancient monetary unit weighed approximately a 10th of a Chinese unit of 兩 tael) [ cf. 錢 qián (SV tiền) 'money' ],
                          - 'đường' 唐 táng (road, as apposed to 途 tú (SV đồ), to 道 dào (SV đạo),
                          - 'lá' 葉 yè (leaf) [ the pattern /j-/ ~ */l-/ is very common in Chinese. ],
                          - 'lúa' 來 lái (paddy, as opposed to 稻 dào 'gạo' rice) [ cf. 麥 mài (SV mạch) ],
                          - 'cá' 魚 yú (fish) [ /ke-/ and /ca-/ in English 'ketchup' and 'catsup' is cognate to V 'cá' ],
                          - 'sông' 江 jiāng (river) as opposed to 川 chuān (SV xuyên) [ cf. 水 shuǐ (SV thuỷ) 'water', another word for 'river' ],
                          - 'mây' 霧 wù (cloud), as opposed to 雲 yún (SV vân),
                          - 'mưa' 雨 yǔ (rain) [ the pattern /y-/ ~ /m-/ is very common in Chinese ~ Vietnamese. ],
                          - 'nắng' 陽 yáng (sunshine) [ Who says there is no Chinese word for 'sunshine'? ],
                          - 'cóng' 寒 hán (chilly) [ Hai. /kwɔ5/ ],
                          - 'biển' 海 hăi (sea) as opposed to both VS 'bể' and 'khơi' [ SV 'hải', for 'khơi', cf. Cant. /hoj3/; it is not hard to associate the 2 related sounds. Ex. 海外 hăiwài V 'hảingoại' (overseas) vs. VS 'ngoàikhơi' (out in the seas) ],
                          - 'bữa' 飯 fàn (Hainanese /buj2/ 'meal' as opposed to SV 'buổi' (period of the day),
                          - 'ăn' 唵 ăn (eat) [ cf. 吃 chī (cf. 乙 yǐ (SV ất) as opposed 'xơi 食 shí (SV thực) ],
                          - 'uống' 飲 yǐn (drink) as opposed to 'hớp' 喝 hè (SV hát) 'sip',
                          - 'đi' 去 qù (go) as apposed to 走 zǒu (SV tẩu) 'run' for 'chạy',
                          - 'đứng' 站 zhàn (stand),
                          - 'ỉa' 屙 é (to poo), 'đái' 尿 niào (to pee, same as VS 'tiểu' conotatively as 'urinate', cf. 尿尿 niàoniào 'điđái'),
                          - 'ngủ' 臥 wò (lie down to rest, hence 'sleep', as opposed to 睡 shuì, connotatiively 'somnus'),
                          - 'đụ' 嫖 piáo (fuck, a derivative of VS 'đéo', coloquially 他媽 Tāma ('Your mother's fucker'),
                          - 'đẻ' 生 shēng (Hainanese /te1/) 'give birth to', in addition to 'tái' (Hai. /ta5/) 'uncooked',
                          - 'việc' 活 huó (work) as apposed to 務 wù (SV vụ),役 yì (SV dịch),
                          and of a great number of other words cited in this paper. For the same reason, due to lack of first-hand experience in modern Vietnamese the same authors will never know dissyllabic words such as
                          - 'đốivới' (with respect to) 至於 zhìyú giving rise to 'đếnnổi' (to such a degree that) as apposed to 對於 duìyú,
                          - 'vòmtrời' 重圓 chóngyuán (SV trùngviên 'sky vault') instead of 宇宙 yúzhōu (SV vũtrụ 'universe'),
                          - 'gỏi' 膾 (鱠) kuài (SV khoái 'mince meat (fish) salad') instread of 'chopped meat or fish',
                          - 'quà' 饋 kuì (SV quỹ 'gift') instead of 禮物 lǐwù,
                          - 'cảirỗ' 菜蘭 càilán (Chinese brocolli) instead of 'cảilàn' or 'cảilan',
                          - 'dưahấu' 塊瓜 kuàiguā (SV khốiqua 'watermelon') in stead of 西瓜 xīguā,
                          - 'ănmày' 要飯 yàofàn (beggar) in stead of 乞丐 qǐgài,
                          - 'thầymô' 巫師 wùshì as opposed to 'phùthuỷ' (shaman), etc.,

                          all are cognates.


                          By at