Let's consider just borrowing between unrelated, national/standardized tonal languages, just in case borrowing between related languages might be a special case and borrowing between non standardized languages might be more variable:

  • Cantonese ↔ Lao
  • Cantonese ↔ Thai
  • Cantonese ↔ Vietnamese
  • Lao ↔ Mandarin
  • Lao ↔ Vietnamese
  • Mandarin ↔Thai
  • Mandarin ↔ Vietnamese
  • Thai ↔ Vietnamese

Given that each language has a different number and selection of tones just like each has a different number and selection of phonemes, is it generally true that the "closest" tones will be borrowed as well as the "closest" phonemes?

Of course this would be within reason since there are going to be phonological constraints such as possible phoneme clusters.

I've put "closest" in scare quotes since this may not be scientifically measurable, so there may be two more or less acceptable mappings possible.

Basically are tones as stable or less stable than phonemes when borrowing between languages?

(Yes we already have questions about borrowing from non-tonal into tonal languages. This question is for when both the source and destination languages are tonal.)

  • 3
    Would the downvoter be interested in offering some constructive criticism? Sep 29, 2013 at 15:36
  • 1
    I would be careful about calling languages unrelated. They have millennia of contact. Pronunciation rules, alphabets and so on - including tones - can be borrowed and influenced. Essentially each language establishes rules for borrowing from other languages, a lot depends on the context of the original borrowing. (Was it via a third language? From which dialect? Written or spoken only? Do the borrowers know the origin language well?) Jan 8, 2016 at 11:31
  • If you borrow your neighbours hammer you don't become relatives. It's the same with languages. Jan 8, 2016 at 11:34
  • Yes, but if you're neighbours, you surely borrow more than his hammer. Languages can be related in more ways that sharing a common ancestor. en.wikipedia.org/wiki/Sprachbund (For example, it's hardly coincidence that languages in that region have tone.) It happens constantly, everywhere. Jan 8, 2016 at 11:36
  • Just think of the borrowing rules for the languages that you know. It's only partly determined by the pronunciation abilities of the native speakers. eg if I now borrow a word from Latin into English, I will change -tio to -tion, by convention. (And that is really the effect of English originally borrowing from French and Italian.) Jan 8, 2016 at 11:38

1 Answer 1


Let me speak only about Thai language and what rules govern the loanwords. Other languages may well "behave" the other way.

is it generally true that the "closest" tones will be borrowed as well as the "closest" phonemes?

Short answer: Yes, but not always for phonemes; Usually No for tones.
I would also say that it's useful to know how words borrowed from non-tonal languages (English) change. It would highlight the things better.

Many words are borrowed along with their pronunciation

  • บาร์บีคิว [baː biː kʰiw] barbecue
  • 孔子 [kʰʊŋ3 tsɨ3] -- ขงจื๊อ [kʰǒŋ tɕɯ́ː] Confucius, a philosopher

...but sometimes not accurately. There are several observations:

  • Consonants are preserved in their written form, but pronounced according to the target language rules. This often happens to final consonants:

    • disco ดิสโก [dìt koː].
      Note the low tone appeared "from nowhere", because the do-dek consonant + dead syllable implies the Low tone (unless an explicit tone mark presents).
      Also, s in final position is pronounced t.
    • apple แอปเปิ้ล [ʔɛ̀ːp pɤ̂ːn]
      Here, the 1st syllable has implied low tone, like the above, and the 2nd syllable receives an explicit tone mark to denote Falling tone; obviously, the original English word has no tones at all, and moreover, the 1st syllable is stressed, not the 2nd (stress somewhat resembles the Falling tone);
      Also, l in final position is pronounced n.

    • football ฟุตบอล [fút bɔːn]
      Again, the implied high tone for the 1st syllable (fo-fan consonant + short vowel) and l in final position;

  • Tones may be totally missing:
    • 叩头 [kʰɤʊ̯4 tʰɤʊ̯2] -- เข่าเท้า [kʰàw tʰáːw] to bow one's head. Note different vowel length and different tones: [kʰɤʊ̯4 tʰɤʊ̯2] is Falling+Rising, while [kʰàw tʰáːw] is Low+High. Different diphthongs must be fine since [aːw] seems to be closest to [ɤʊ̯].
    • 老子 [lɑʊ̯3 tsɨ3] -- เล่าจื๊อ [lâu tɕɯ́ː] Lao-Tze, a philosopher. Again, the original FallingRaising+FallingRaising tones become Falling+High due to explicit tone marks.

However, more drastic changes happen when loanwords omit tone marks, so they have default tone governed by the initial consonant (1).

  • カラオケ [ka ra o ke] คาราโอเกะ [kʰaː raː oː kè], karaoke.
    This word is a shining example: the original word has all syllables of equal length (has no lexical vowel length), while being borrowed, it has long vowel for syllables 1-3, but short vowel for syllable 4 (plus, short vowel produces Low tone according to Thai grammar rules).
  • อินเทอร์เน็ต [in tʰɤː nét] Internet.
    Note that due to vowel length, it sounds rather like "inTERnet", plus the last syllable has High tone.
  • 公司 [kʊŋ1 sɨ1] -- กงสี [koŋ sǐː] -- company.
    Different vowels [ɨ] vs. [iː], different length, and different tones: High+High tones vs. Mid+Rising

All above applies per syllable, of course. If the word is polysyllabic, the same rules apply that govern polysyllabic words of Thai language (see Enepenthetic ‘Leading Consonant’ Clusters).

(1) In Thai language, syllable's tone is defined by its initial consonant, unless overridden by an explicit tone mark. Each consonant belongs to one of three consonant classes. Each consonant class is associated with a default tone, governed by live/dead syllable or long/short vowel.

So, the summary:

  • Consonants are often preserved in their written form;
    • And become governed by the target language rules;
  • Vowels are more vulnerable for changes (than the consonants);
  • Some loanwords attempt to keep original tone;
    • Often, it happens to be not very accurate;
    • A large number of words do not retain the original tones;

(A) Fundamental frequency contours of the four Mandarin Chinese tones (spoken by a female speaker of Mandarin). Each tone on the syllable /ma/ represents a different lexical item. “ma1” is the level tone, “ma2” the rising tone, “ma3” the dipping tone, and “ma4” the falling tone.
(B) Fundamental frequency contours of the five Thai tones (spoken by a female speaker of Thai). Each tone on the syllable /ma/ represents a different lexical item. “ma0” is the mid level tone, “ma1” the low level tone, “ma2” the falling tone, “ma3” the high level tone, and “ma4” the rising tone.

Tone contours for Chinese and Thai Image from here


  • 1
    Very interesting to see that some vowels became long in Thai as well. I guess it would be quite similar with maybe slightly relaxed results into Lao due to its phonetic spelling compared to Thai's etymological spelling. And now it makes me wonder about borrowing between the other tonal language families. Still it's confirmed my hunch that tones might not survive borrowing just because phones tend to. I probably should've emphasized modern borrowings to because obviously the older a borrowing is the more the languages diverge which can obscure the processes. Sep 28, 2013 at 18:34
  • 1
    @hippietrail This is exactly what I observe. Admittedly, it's not a scientific approach, but probably high and raising tones do not clash with low and falling. Other than that, the target ("borrower") language rules apply more aggressively than the original phones. Sep 28, 2013 at 21:14
  • 1
    "IPA for Chinese". What you use for Mandarin looks more like just pinyin to me. "k" in pinyin is aspirated, so your point about "differences, including aspirated kʰ" isn't quite clear.
    – dainichi
    Sep 30, 2013 at 2:40
  • 2
    You give a few examples where falling+rising becomes low+high and high+high becomes mid+rising. From my limited knowledge of autosegmental phonology I see a certain similarlity between the original and the adapted tones. Is it possible that there are phonological rules in Thai that account for the changes? For example, a hypothetical *H initial rule could easily account for the H>Rising and H>M transformations (I'm not saying such a rule exists, just giving an example of how such rules could explain the discrepancies)
    – acattle
    Sep 30, 2013 at 3:19
  • 2
    Thai [koŋ sǐː] is possibly a loan from Cantonese, whence gung1 si1, although the tones correspond to neither Mandarin nor Cantonese.
    – jogloran
    Sep 30, 2013 at 8:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.