Phonology Vowels Semivowels Diphthongs Triphthongs Consonants Nasals Plosives Fricatives & Affricates

 

Phonology

 

In constructing a worldlang, one should be careful to make the phonology as neutral as possible. It should not be especially difficult for native speakers of any particular language. Whenever possible, the difficulties should be distributed as evenly as possible. (JW) (LS)

 

A minimum phoneme set for a worldlang would be five vowels (/a e i o u/), two semi-vowels (/j w/) and seven consonants (/p t k m n s l/). (JW) (LS)

 

Vowels

 

It is generally recognized that a world language should have five vowels: /a e i o u/. They can be known as the "favored five". (JW) (RK) (LS) (FG)

 

Vowel chart

i		u
e		o
	a

 

Semivowels

 

The semivowels /w/ and /j/ should be used. (LS) (JW) However, these semivowels should be allphones of /u/ and /i/. (DN) (JW) (LS)

 

One important issue is the range of /w/. It may be helpful to expand it to cover /v~V~w/. Or /v/ can be an allophone of /w/. (JW) (LS) (FG) (RK)

 

I'd favor a broader phoneme /v~V~w/ (/j/ is OK). (FG)

 

/v/ could be an allophone of /w/. (LS)

 

/ʋ/ is halfway between /v/ and /w/, so I included it, but keep it simple: /w~v/ is OK. (FG)

 

Diphthongs

 

Limited diphthongs should be allowed. /aw/ and /aj/ are definitely acceptable. /oj/ also seems recognizable. (LS)

 

/ew/ is recognisable to the same extent that /oj/ is. (LS)

 

/ew/ and /oj/ don't seem like a very good idea. Even I have trouble with /ew/. (DN)

Okay by me. (LS)

 

The following diphthongs should be allowed: /ai au ia ie io iu ua ue ui uo/ (DN) (LS) (JW)

 

I'm suspicious about /ie io iu ue ui uo/. All of them are rising diphthongs. Is there some reason to think that they are easier than their falling counterparts /ei oi eu ou iu ui/? (RK)

 

I think you are seeing some English-language bias here. For us North-American English speakers, we tend to glide off those vowels, so the falling diphthongs are largely allophonic with their pure leading vowels. To our ears, it's easy to distinguish /uo/ from /o/, but hard to distinguish /ou/ from /o/. But maybe others would find other pairs hard to distinguish...I'd like to see some rationale for including some and excluding others. Perhaps we need to scale back the diphthong set. (LS)

 

I'm on the fence regarding /ie/ and /uo/ because the components are too close together. I have them in Sasxsek, but tend to use them sparingly, especially /uo/.

 

I'm not sure myself. Should sounds like "ia" be considered diphthongs, or are they a semivowel followed by a vowel, i.e. "ja". I think it's OK to consider them together orthographically, but maybe "ia" is not usually pronounced as a diphthongs. I'm confused, so I'll leave it to more expert people among you to answer this. (JW)

 

I suggest considering /ja/ and /ia/, etc., allophones. Then it's simply a matter of orthography. I personally find spellings like "kui" and "uanai" visually prettier than "kwi" and "wanay", but that's a teeny-tiny point. (LS)

 

Well, that's why they are called "semi-vowels" because they are in that grey area between vowel and consonant. It's that position of uncertainty that leads me to just one phoneme for the whole range. (DN)

 

With the preceding six comments in mind, let me suggest the following as permissible diphthongs (and considering /i/ ~ /j/ and /u/ ~ /w/): /ai au ia io iu ua ue ui/. The design point is that the the two components of each diphthong are always at least two steps apart on the vowel tree if one of them is /a/, and three steps otherwise. The omission of /eu/ and /oi/ is not defensible by this rationale, admittedly. (LS)

 

Triphthongs

 

The following triphthongs should be allowed: /iai iau uai uau/ (DN) (LS) (JW)

 

Consonants

 

At the very least, a world language should have the following "sacred seven" consonants: /p t k m n s l/. (JW) (LS) (FG) (DN) (RK)

 

Consonant chart

m	n
p	t	k
	s
	l

 

Larry >>>

If maximum aural comprehension is paramount, the consonant inventory should be limited to these. This makes adaptation of world vocabulary more difficult, but still conceivable because words that fit this limited phonology can usually be cherry-picked (though it may end up with quite a lot of borrowing from Tongan). This also makes for long words, but brevity is not necessarily the virtue that we often think it is.

<<< Larry

 

Nasals

 

The sounds /m/ and /n/ should definitely be included. The sound /N/ is more complicated. Many languages do not make a distinction between /N/ and /n/, and /n/ often mutates to /N/ before /k/ and /g/, even across word boundaries. (LS) (JW) (FG) It is best to include /N/ merely as an allophone of /n/. (LS) (JW)

 

Pandunia has /m n N/

 

Larry >>>

I would omit /N/; many languages do not make a distinction between /N/ and /n/, and /n/ often mutates to /N/ before /k/ and /g/, even across word boundaries.

<<< Larry

 

I also agree that /N/ should be dealt with carefully. I think it's unnecessary.

 

It certainly need to be treated carefully, but without it you will limit borrowing from a lot of Eastern languages like Chinese. (DN)

 

If /N/ is really useful, why not allowing /ng/ cluster as an allophone? (FG)

 

Yes, let /N/ be an allophone of /n/. (LS)

 

I'm about 50-50 regarding /N/. I like to leave it out of the phonology, but there are too many limitations in borrowing Sinitic and Austronesian words, which is why it was left in Sasxsek. For some of the more Creolistic languages like NP or Pandunia, it may be better to go without it, but either way works fine. Let's just not go for anything really crazy like /J/.

 

Let's look for opportunities for Sinitic borrowing for words without /N/, then. (LS)

 

That limits things quite a bit when you consider Mandarin only has two final consonants /n/ and /N/. Leaving it out may also be a good idea if you are going to do something like maybe have a general "final nasal" which has no specific point of articulation defined, but assimlates with the following consonant to become /m/, /n/, or /N/ as needed. This is something I used in my loglang although the idea comes from the Japanese mora. (DN)

 

Plosives

 

I propose distinguishing fortis and lenis plosives. The fortis plosives "p t k" should be voiceless and possibly aspirated. The lenis plosives "b d g" should be fully or partially voiced and unaspirated. The reason is that there are two major accents:

1. Languages with voicing distinction: Spanish, Portuguese, French, Russian, Indonesian, Japanese, Swahili, Arabic
2. Languages with aspiration distinction: Chinese, English, German

This way the both major accents would fit in.

-- Risto

 

English and German do not have an aspiration distinction. It's not a phonemic feature. /p_h/ and /p/ are both allophones of the phoneme /p/. Phonemic aspiration is something that's not very common so I wouldn't recommend it for any WAL, any more than I'd recommend palatization, labialization, ejectives or implosives. The voiced distinction isn't universal but it's common enough to be worthy of at least considering. (DN)

 

I was under the impression that phonemic aspiration was just about as common, in terms of number of speakers, as phonemic voicing. Mandarin and Japanese, for example. But further, what I've noted is that these distinctions are not even the same from language to language. That is, though both language X and language Y have phonemic voicing, speakers of language X still can't hear the difference when a language Y speaker speaks, and vice versa. It's a continuum and it's astonishing how much variation there can be. The only reason I support even having two series (canonically placed as far apart as possible) is to ease adoption of international vocabulary, and because having two series is more common than three, four, or one. (LS)

 

The following plosives should be included in a worldlang: /p_h b/, /t_h d/, and /k_h g/. (RK) (FG) (LS) (JW)

 

The fortis plosives be voiceless and strongly aspirated; while the lenis be not just voiceless and unaspirated, but "softened", somewhat in the manner that Spanish /b/ is softened toward /v/. (LS) (JW)

 

I'd avoid the softened consonants too. Yes, it's likely to happen with certain speakers, but it's not a good standard. The softening in Spanish is where the voiced stops /b d g/ are pronounced as fricatives [B D G] in *medial* and *final* positions. (DN?)

 

I also would like to add to this that /d/ and /t/ are articulated differently in English than other European languages where it's a dental sounds. So the "soft" /d/ would become z to an anglophone (DN)

 

If we avoid the proposed wide distinction between lenis and fortis, then I have to recommend only a single series, the fortis. The distinction otherwise is so small that it's almost inaudible even between a speaker/listener pair who share a native accent or language; and if they don't share native accent/language, the two series are quite indistinguishable. (LS)

 

Okay, Dana, you make good points; those softened sounds may be just too rare. Hopefully /p_h b/, /t_h d/, and /k_h g/ (which four of us agree to above) are distinct enough pairs that when a, say, Finnish speaker pronounces them more like /p_h p/, /t_h t/, and /k_h g/, while an English speaker pronounces them /p b/, /t d/, and /k g/, they will understand each other's accent and neither accent will be considered right or wrong. (LS)

 

Note that /p_h/ will disadvantage speakers of Classical Arabic. But will it disadvantage speakers of the various Arabic spoken languages? (Almost no Arabic speaker speaks only Classical Arabic.) ---Larry

Further "research" suggests that only a few Arab sub-languages have /p/. (LS)

 

Standard Arabic has no /p/. Borrowings usually use /b/ or /f/, hence words like "Farsi" which is "Parsi" in its native form. (DN)

 

If /f/ is not used, it could be an allophone of /p_h/. (FG) (LS)

 

Keep in mind the connections /f/ or /P/ have to /h/ in some languages. The Japanese mora is spoken /PM/. Etymologically Spanish words with initial /0/ come from Latin words in /f/. (DN)

 

I think it's a false problem. From my experience with Arabic speakers, I noticed some problem with vowels, like saying /i/ for /e/ and /u/ for /o/, but I never heard a /f/ instead of a /p/ or anything similar. Maybe it's due to French influence in Maghreb? (FG)

 

Florent, you likely have heard a /b/ for a /p/, but given how close they are, you would likely not notice in normal speech. (LS)

 

You are right. Within context, it's possible I didn't notice a /b/ for a /p/. However, bilabials are the earliest phonemes mastered by babies. They are incomparably easier than, for example, /v/, rothics, or the dreaded /T D/. (FG)

 

Larry >>>

(Actually, English has voicing distinction, not aspiration distinction. Yes, English speakers do naturally make both aspirated and unaspirated plosives, but the differences are not phonemic.)

 

Notwithstanding my views on aural comprehensibility, one must still consider the written word, which is how a world language would begin to get off the ground. And it must be admitted that most languages do have at least two series of plosives (although it can be astonishingly difficult to distinguish them in a language/accent not one's own).

 

Therefore I concur that two series of plosives is not a bad idea. I further suggest that the fortis plosives be voiceless and strongly aspirated; while the lenis be not just voiceless and unaspirated, but "softened", somewhat in the manner that Spanish /b/ is softened toward /v/. This will give a distinctive sound to the language and set a target pronunciation that does not nestle so preferentially within some speakers' natural inclinations.

 

An alternative to "softening" the lenis plosives would be forcing strong aspiration of fortis plosives via palatilisation. For example, a minimal pair might be /kwan/:/gan/, rather than /kan/:/gan/ or /kwan/:/gwan/. If this were a rule, it would obviate the need to monitor and control borrowings for vocabulary. But this definitely smacks of artificiality.

 

Finally, if voiced+unaspirated versus unvoiced+aspirated distinctions are to be made, then /z/ should be considered as the lenis fricative for its fortis fricative /s/.

 

Other common consonants — /f/, /v/, /S/, /Z/, /h/ are more troublesome. If /S/ and /Z/ are adopted, then /t_S/ and /d_Z/ should probably be permitted as allophones.

<<< Larry

 

I think that the idea of palatalization (actually your example looks more like labialization) is not suitable for worldlang. Voicing and aspiration are as strong cues as palatalization and labialization if they are done consistently and they are far more common than the other two.

Risto

 

Fricatives & Affricates

 

Phoneme proposal: the "throat back" phoneme /x~R~h/. (FG)

 

There should be a fortis fricative /s/ and a voiced /z/. (LS)

 

I think everybody agree on the following phonemes: /s z S Z t_s d_z t_S d_Z/, but not on the manner to regroup them. (/s S t_s t_S/ would be fortis, and /z Z d_z d_Z/ lenis.) (FG) (RK)

 

Proposals

 

RK:	s~S	z~Z	t_S~t_s	d_Z~d_z
LS:	s	z	S~t_S	Z~d_Z
DN:	s	z	t_S	d_Z	/S Z/, /t_s d_z/ or /t_s\ d_z\/ can be allophones of either depending on etymological considerations
JW:	s		S~t_S		I know this is quite limited, but I prefer not having voiced versions
FG:	s	S	t_S_h	d_Z	Missing phonemes would be re-allocated case by case

 

Some languages (like French) doesn't have true affricates, but allow near-equivalent consonants clusters, for example: /tS/ instead of /t_S/. I think this fact should be taken in account. (FG)

 

That's okay, the two sound quite close to one another. If we have an affricate, we will not also have a consonant cluster that could be confused with the affricate. (LS) (DN)

 

Phonology

 Create your own educational wiki!HelpPublic, not yet premium Privacy Policy RSS