Almost all languages of the world have more consonants than vowels. Are there some languages of the world with more vowel phonemes than consonants?
Asked
Active
Viewed 4,050 times
18
-
Can someone post an answer involving whistling languages? I can't quite do it. – Joshua Apr 24 '23 at 17:10
-
3@Joshua there are no natural whistled languages, only whistled registers of spoken languages (the key fact here is that they encode a language whose default mode is spoken, rather than being languages that are primarily whistled; in the case of the famous silbo gomero, it is a whistled register of Spanish). The consonant/vowel distinction is not really maintained in whistling, so I don't think the question could really be applied to them, any more than it could to sign languages – Tristan Apr 25 '23 at 08:57
5 Answers
21
Probably the best-known and most often-cited example of this is Danish.
Danish is generally said to have around 17 or 18 consonant phonemes, a fairly invariant number. The number of vowel phonemes cited varies a lot more, because there’s some serious interplay going on between morphophonemes, phonemes and phones, but it’s very difficult to reduce the vowel system to less than 20 vowels – generally, around 20 or 21 morphophonemic vowels and around 25–28 vowel phonemes are cited.
Janus Bahs Jacquet
- 4,673
- 1
- 22
- 28
12
The conventional understanding of "phoneme" is that it is a segment. There is vast disagreement over what constitutes a "segment". Given that, one example of a language with many more vowels than consonants is Vietnamese, which has 84 vowels and 19 consonants. To put the matter in more familiar terms, are [ɛ] and [ɛ̃] in French one vowel or two? Akan has a vowel distinction "Advanced Tongue Root" which is contrastive – are [e] and [e̟] two vowels or one? Various languages have two or three degrees of vowel length: are Estonian [a, a:, a::] one vowel or three?
There are various analytic methods for eliminating vowel from the list of phonemes, factoring out orthogonal properties that are available to vowels in general. Without systematic criteria for saying whether a vowel property is to be taken into consideration (and please note that all of my examples are phonemic), it is likely that "more vowels than consonants" is actually a very common feature of languages.
user6726
- 83,066
- 4
- 63
- 181
-
4Your Vietnamese example works only if you count all the diphthongs and triphthongs as separate vowels. The orthography recognises only 9 vowels and 6 tones. – fdb Apr 21 '23 at 18:33
-
3Diphthongs and triphthongs are often treated as separate vowels, just as "prenasalized stops" and affricates are often treated as separate phonemes. Likewise there is a debate as to whether Chinese has [k, kʷ] as phonemes, of [k,w] and some clusters. – user6726 Apr 21 '23 at 23:25
-
2For the sake of this question: The clearer the vowel majority is, the better. When there is a vowel majority without invoking tone, length, and diphthongs the example is just perfect. – Sir Cornflakes Apr 22 '23 at 19:33
-
1@fdb English orthography "recognizes" five or six vowels, but there are far more vowels actually used. The mapping of phonemes to letters is never one to one. – PC Luddite Apr 23 '23 at 20:44
-
1@PCLuddite. In the case of Vietnamese the assumption of 9 vowel phonemes is an adequate and economical phonological analysis, with tone as a super-sequential feature. – fdb Apr 23 '23 at 21:45
-
-
11
I have yet to see anyone bring up the Iau language of West Papua, Indonesia, which has only 6 phonemic consonants (not counting allophony) but 8 vowel qualities even before accounting for diphthongs and tones.
Arcaeca
- 344
- 1
- 5
6
In general, languages with fewer consonants will have more vowels (and vice versa), so as to have as many unique syllables as possible and preserve redundancy. (Danish is an exception.)
- Hawaiian has eight consonant phonemes and twenty-five vowels (including diphthongs & long vowels). If you don't include diphthongs, it still has ten vowels, including the long vowels.
- Danish has around seventeen consonant phonemes and around twenty-six vowel phonemes.
- Apinayé is analyzed as having seventeen consonants and seventeen vowels.
- Finnish has fourteen native consonants (a few more are found in loanwords) and sixteen vowels (including long vowels). Edit: Including diphthongs, it has thirty-four vowels.
nearsighted
- 395
- 1
- 11
-
4When counting length in vowels, shouldn't length in consonants (gemination) also be counted as distinguishing feature? This would change the balance for Finnish. – Sir Cornflakes Apr 21 '23 at 20:58
-
Rotokas technically fails the OP's requirement, but comes really close, with 5 vowels and 6 consonants. – dan04 Apr 21 '23 at 21:17
-
@SirCornflakes I suppose so; however, the phonology charts I see don't show geminates. On the other hand, geminates are sometimes analyzed as sequences of two of the same consonant (e.g. akka is [ak.ka]), but I guess by that logic, diphthongs shouldn't count, taking Hawaiian off the list. – nearsighted Apr 22 '23 at 13:22
-
@nearsighted But diphthongs are usually vowels that change in quality as they progress. They are often vectors rather than a sequence of two vowels. I'm not sure what the situation is in Hawaiian though. – Araucaria - him Apr 22 '23 at 14:51
-
@Araucaria-him Well, as I think about it I suppose you're right. But they're still written either as two vowels or as a vowel followed by a glide. – nearsighted Apr 22 '23 at 15:26
-
if you count long vowels as distinct from short vowels (in Finnish), should you also count vowel combinations as distinct? i.e. sana != Saana != sauna – ilkkachu Apr 23 '23 at 11:39
-
-
-
@ilkkachu I counted diphthongs for Hawaiian; forgot to include them for Finnish though ─ thanks! – nearsighted Apr 23 '23 at 15:39
0
Other languages likely have more unbalanced ratios than this, but for interest' sake, the Tai-kadai languages (Thai, Lao, Shan, etc.) have a few more vowel phonemes than consonant phonemes. This is not true, however, of their graphemes, as many consonants have multiple grapheme representations.
Taking the Thai as the dominant example, we can represent its sounds as follows.
CONSONANTS
| Approximate Phoneme | Thai Grapheme(s) |
|---|---|
| b | บ |
| bp | ป |
| ch | ช / ฉ / ฌ |
| d | ด / ฎ |
| dt | ต / ฏ |
| f | ฟ / ฝ |
| gk | ก |
| h | ห / ฮ |
| j | จ |
| k | ค / ข / ฃ / ฅ / ฆ |
| l | ล / ฬ / ฦ |
| m | ม |
| n | น / ณ |
| ng | ง |
| p | พ / ผ / ภ |
| r | ร / ฤ |
| s | ซ / ส / ษ / ศ |
| t | ท / ฑ / ธ / ฐ / ถ |
| w | ว |
| y | ย / ญ |
NOTES: The "w" and "y", like in English, can also function as vowels, and the "r" also sometimes functions as a vowel in Thai. Lao lacks the "ch" and officially it lacks the "r" (which it had in the past, but officials have "simplified" the language, converting the "r" to either an "h" or an "l" in pronunciation), so it has even fewer consonants than Thai.
VOWELS
| Approximate Phoneme | Thai Grapheme(s) |
|---|---|
| ah | -า |
| aht | -ะ |
| i | -ี |
| it | -ิ |
| eu | -ื |
| eut | -ึ |
| u | -ู |
| ut | -ุ |
| ae | แ- |
| aet | แ-ะ |
| e | เ- |
| et | เ-ะ |
| o | โ- |
| ot | โ-ะ |
| au | -อ |
| aut | เ-าะ |
| euu | เ-อ |
| euut | เ-อะ |
| ia | เ-ีย |
| iat | เ-ียะ |
| eua | เ-ือ |
| euat | เ-ือะ |
| ua | -ัว |
| uat | -ัวะ |
| ai | ไ- / ใ- |
NOTES: The dash/hyphen in the grapheme column represents a consonant. (Some will still show the hollow circle to indicate a missing grapheme.) No vowel in Thai or Lao can be written without a consonant to which it is attached (Shan is written in Burmese script, with similar rules). The "t" represents a shortened vowel in which the sound cuts off abruptly almost as if culminating in a "t" in English.
Adding them all up, the vowels (inclusive of diphthongs and long/short vowels--which are, to the speakers of the language, entirely different phonemes), vowels outnumber consonants by a small margin (about 25 to 20). In Lao, this margin is increased, as Lao lacks some of the consonants which Thai uses, and might be said (debatably) to have one or two vowels which Thai does not have.
Biblasia
- 163
- 6
-
2“Officials” didn’t “simplify the language” to convert /r/ to /h/ or /l/; that’s not how sound change works. The change happened naturally and gradually in regular speech, not because any officials decreed that it should happen: /r/ became /h/ initially in monosyllabic words and /l/ elsewhere. It’s not known when exactly the latter change happened (though it had definitely happened by the early 20th century), but the change to /h/ seems to have happened around 500 years ago. – Janus Bahs Jacquet Apr 24 '23 at 12:38
-
@JanusBahsJacquet I could be more specific, actually. It was a recent king of Laos, just prior to Communism taking over in the country, who reduced the alphabet and removed certain consonants from it, including the "r". I have spoken with native Lao people, within Laos, who explained this--note that Wikipedia does not appear to have the full details, but a brief mention of the fact that the "r" "was dropped as part of a language reform" can be seen HERE. – Biblasia Apr 24 '23 at 13:20
-
1@Biblasia: that sounds more like a spelling reform than a language reform. It didn't change the language, just the way it was written. – TonyK Apr 24 '23 at 13:22
-
1While this answer contains a good insight, it "unfairly" favors vowels over consonants. Namely, it promotes vowel clusters (like
iaandua) into full-featured vowel sounds while discourages consonant clusters (likekr,khr,kl,khl,kw) at all. It promotes vowel duration (thus promotingaanda:into two full-featured ones) but discourages the final values of many consonants (e.g. written/initialrbecoming finaln) etc. Thankfully, it did not consider lexical tones; that would multiply the whole multitude of vowels by 5. – Be Brave Be Like Ukraine Apr 24 '23 at 13:24 -
@TonyK I should also note that in addition to simplifying the Lao script, one of the purposes of the king was political in nature. He deliberately wanted the Lao language to be different from Thai, hence the removal of certain consonants, and pressure placed on the population to adopt the revised pronunciations. – Biblasia Apr 24 '23 at 13:24
-
@BeBraveBeLikeUkraine From an outsider's perspective, you make a valid point. From within the language, however, those consonant clusters are treated as combinations of separate phonemes, not new phonemes. For example, the "ng" and "ch" are distinct and indivisible phonemes, but something like "kr" is divisible into its "k" and "r" components--which is why I do not count "kr" as its own consonant. – Biblasia Apr 24 '23 at 13:31
-
@Biblasia As Tony mentioned, this was a change in writing, not in speaking. The /r/ sound (note the /phoneme slashes/, rather than ⟨grapheme brackets⟩) had disappeared much earlier, and reforming the alphabet was just a way to catch up with the existing change in pronunciation. Spelling reforms like that are common, but they don’t affect the inventory of actual sounds used in the language. – Janus Bahs Jacquet Apr 24 '23 at 14:05
-
@JanusBahsJacquet I speak, read, and write Lao. I can assure you that the spellings with the "r" very much exist. However, it is not pronounced as an "r". If one says "โรงเรียน/rongrien" (school) in Thai, the "r" is like the rolled "r" of Spanish. But if one speaks this word in Lao, it (ໂຮງຮຽນ) is pronounced as "honghien." However, in "borisut" (pure), the "r" becomes an "l", i.e. "bolisut." This is reflected in the spelling (Lao: ບໍລິສຸດ vs. Thai: บริสุทธิ์). Lao actually has two "r" characters (ຮ/ຣ), in addition to characters for "l" (ລ) and for "h" (ຫ). They just aren't pronounced. – Biblasia Apr 24 '23 at 14:16
-
@Biblasia I know they exist, and that they’re not pronounced with a [r] sound. But they were not pronounced with [r] before the alphabet reduction you mention either. Officials didn’t suddenly tell everyone in Laos, “You cannot pronounce the sound [r] anymore – from now on, you have to pronounce it as either [h] or [l]”. That did not happen. What happened was that, about 500 years ago, in word-initial position, the original /r/ phoneme (written ຣ) naturally started to be pronounced first as unvoiced [r̥], then later on as [h]; i.e., it merged with the previously distinct phoneme /h/. → – Janus Bahs Jacquet Apr 24 '23 at 15:00
-
→ The letter ຮ (a variation of ຣ) was probably invented while that change was happening, perhaps to denote the difference between [r] and [r̥]. Some time later (no one knows exactly when, but definitely no later than the late 19th or early 20th century), the remaining instances of /r/, such as in borisut, started being pronounced as [l], merging with the previously distinct phoneme /l/. When that happened, Lao no longer had any [r] sounds at all, but the words were still written with the original letters, so you could see whether a word had originally contained an /r/ or not. → – Janus Bahs Jacquet Apr 24 '23 at 15:00
-
→ It wasn’t until a long time after that that the spelling reform you mentioned above took place, seeking to get rid of this complication by just writing words as they are pronounced. The reform you mention changed the letters, but the sounds had already changed much earlier. – Janus Bahs Jacquet Apr 24 '23 at 15:00
-
@JanusBahsJacquet Where are you getting your information? Did you catch that fact I mentioned earlier that in the case of "borisut" there was a spelling change? You seem not to have followed that one. I assume Lao is not one of your languages. Until you have an intimate knowledge of Laos, it is likely you will place undue confidence in English-language sources for information on this subject. The only spelling reform on the "r" was the removal of one of them--the other remained. The removed "r" has been reintroduced. Many Lao words have dual spellings (with/without "r" and/or combined "h"). – Biblasia Apr 24 '23 at 15:21
-
If you’d read and understood my comments, you would know that I did catch that, yes. But it’s clear you’re not understanding anything that I’m saying, despite my attempts to make it as clear as possible, so I’ll not try any further. I’ll just say that this statement from your answer is incorrect: “officials have "simplified" the language, converting the "r" to either an "h" or an "l" in pronunciation”. – Janus Bahs Jacquet Apr 24 '23 at 17:03
-
@Biblasia I can't believe that anyone would believe that government officials would have the ability to consciously change the language's pronunciation. Spelling, yes, but not pronunciation. – nearsighted Apr 30 '23 at 16:25
-
@nearsighted Not just pronunciation, but vocabulary/language itself has been changed by governments. Consider Franco's regime in Spain, per the scant details offered via Wikipedia. There is much more to that story, as the older generation in Spain might still remember. – Biblasia May 01 '23 at 05:49
-
@Biblasia All Franco did was ban the usage of languages other than Spanish, including in names. There is no record, among the scant details offered (which, as Wikipedia helpfully notes right at the top of the article, have multiple issues), of any change to the vocabulary or grammar. And not everyone would obey, anyway. – nearsighted May 01 '23 at 18:28
-
@nearsighted I studied some of this history, in Spanish, while in Spain, but that was decades ago. I remember clearly that the people chafed at the policies, and when Franco was gone, the local dialects again emerged. But those policies did affect the language in lasting manner. – Biblasia May 01 '23 at 18:50
-
@Biblasia Please give an example of a linguistic policy that Franco enacted other than making Spanish the only official language. – nearsighted May 01 '23 at 19:00
-
@nearsighted Franco prohibited the local dialects from being used. Effectively, this meant that, for example, Valenciano (the dialect of the region of Valencia) could not be used in legal documents or public venues. The same would be true for each of Spain's other regions, e.g. Madrileño for Madrid, Andaluciano for Andalucia, etc. In Andalucia, the "ceta" (C) is pronounced as "S" whereas in other parts of Spain it is pronounced as "TH"--how this was affected during Franco's time I do not know, but it's almost like a lisp--the Andalucians seem unable to pronounce the "TH". – Biblasia May 01 '23 at 22:44
-
@Biblasia So all he did was try to force some people to pronounce /s/ as /θ/? (By the way, there's no need to have the English approximations for the Thai vowels in your post. We've all studied linguistics here; the IPA would work much better.) – nearsighted May 02 '23 at 18:06
-
@nearsighted I have studied IPA at a master's degree level, and I do not like the system. It poorly represents, or even lacks, many of the phonemes that I deal with on a day-to-day basis in the Asian languages, and seems much better suited to the more traditional European language family. In any case, I avoid using IPA to avoid these weaknesses. If only there were a better system....but alas, one cannot even well hear a sound, much less pronounce or represent with a symbol, that he or she did not learn in early childhood. – Biblasia May 02 '23 at 21:50
-
@JanusBahsJacquet (fyi, Janus, this argument is still going) Biblasia: the IPA represents quite nicely all of the complicated tones of the Tai-Kadai, Sino-Tibetan, and other tonal Asian families. But getting back to the point of our discussion, it seems all Franco did was try to force people to pronounce /s/ as /θ/. That's not a significant change, and amounts only to forcing people to speak in a certain dialect ─ not as extreme as your Thai example above. – nearsighted May 02 '23 at 22:18
-