Fenja orthography: An abugida with different inherent vowels

Substantial postings about constructed languages and constructed worlds in general. Good place to mention your own or evaluate someone else's. Put quick questions in C&C Quickies instead.
Post Reply
User avatar
Chengjiang
Avisaru
Avisaru
Posts: 437
Joined: Sun Feb 26, 2006 4:41 am
Location: Davis, CA

Fenja orthography: An abugida with different inherent vowels

Post by Chengjiang »

I've had trouble coming up with glyphs I like, but I really wanted to talk about the design principle of the orthography of my Ansai language group, whose most prominent representative is Undreve. This orthography is named for Fenja, the prestige dialect of a dialect continuum from which the modern Ansai languages descend. It is distinguished by being more or less an abugida but having different inherent vowels for different groups of consonants based on their historical secondary articulation.

Fenja phonology

To start off, we need to understand the phonology of Fenja. In the period where the writing system's usage became (more or less) standardized, most consonants contrasted some subset of labialized (probably technically labiovelarized), palatalized, and velarized or neutral articulations. There is some uncertainty as to whether various "velarized" consonants actually had velar secondary articulation or simply lacked a secondary articulation (and obviously "velarized velars" aren't literally velarized), but for the purposes of the orthography and historical phonology of the descendant languages they will be marked as velarized.

Fenja's consonants are as follows:

/mˠ mʲ nˠ nʲ nʷ/
/pˠ bˠ pʲ bʲ tˠ dˠ tʲ dʲ tʷ dʷ kˠ ɡˠ kʲ ɡʲ kʷ ɡʷ/
/fˠ vˠ fʲ vʲ sˠ zˠ sʲ zʲ sʷ zʷ/
/lˠ lʲ rˠ/
/j w/

As you can see, this is an overlay of secondary articulations on this base inventory:

/m n/
/p b t d k ɡ/
/f v s z/
/l r/
/j w/

(This is more or less the inventory of Proto-Ansai. Fenja's inventory came about after numerous monophthongizations, glide deletions, and other vowel shifts made the secondary articulations phonemic.)

Labialization is only distinguished from velarization on non-labial obstruents and nasals.

Fenja's vowel system distinguishes far more qualities in long vowels than short:

/i a u/
/iː aː uː eː oː ʉː/

Fenja allows syllables of the form (C)V(C)(C). Long vowels can occur in open syllables or syllables ending in a single consonant, but not in syllables ending in two consonants. Coda clusters are restricted to a glide or liquid followed by an obstruent, a homorganic nasal followed by an obstruent (with the coronal nasals assimilated to velars), or a glide followed by a liquid or nasal.

On top of this, consonants in a cluster assimilate in secondary articulation to a following consonant, with /j/ treated as palatalized and /w/ treated as labialized when preceded by another consonant. (They aren't affected by following consonants.) Labialization is not distinguished from velarization before rounded vowels.

Glyphs

Every one of the consonants distinguished in Fenja has a glyph. These consonant glyphs have an inherent short vowel whose quality depends on their secondary articulation: Velarized consonants have an inherent /a/, palatalized consonants have /i/, and labialized consonants have /u/. There is also a "dummy" or "carrier" glyph with inherent vowel /a/, used for word-initial syllables with null onsets. (Provided the vowel isn't a close vowel; initial close vowels are written with the "logical" close vowel glyph, with initial /ʉː/ using the glyph for /j/. This doesn't cause ambiguity because Fenja did not allow word-initial sequences of glide plus close vowel.)

These glyphs will be represented as follows:

/mˠ mʲ nˠ nʲ nʷ/ <MA MI NA NI NU>
/pˠ bˠ pʲ bʲ tˠ dˠ tʲ dʲ tʷ dʷ kˠ ɡˠ kʲ ɡʲ kʷ ɡʷ/ <PA BA PI BI TA DA TI DI TU DU KA GA KI GI KU GU>
/fˠ vˠ fʲ vʲ sˠ zˠ sʲ zʲ sʷ zʷ/ <FA VA FI VI SA ZA SI ZI SU ZU>
/lˠ lʲ rˠ/ <LA LI RA>
/j w 0/ <YI WU HA>

The vowels, meanwhile, have the following set of six glyphs:

/iː aː uː eː oː ʉː/ <I A U E O Ü>

These vowel glyphs cannot be written independently and must have a consonant glyph to carry them.

As you can see, only long vowels are typically written. Since the majority of the time a short vowel following a consonant corresponds to that consonant's inherent vowel, and since the range of legal consonant clusters is not too large, there is not much information loss from having occasional other short vowels or the lack of a vowel not written.

A few example words:

/sʲiɡʲirˠdˠ/ <SI.GI.RA.DA>
/kʷoːzˠɡˠeːnˠ/ <KU.O.ZA.GA.E.NA>
/fʲirˠkʲinʷtʷʉːmʲ/ <FI.RA.KI.NU.TU.Ü.MI>
/vʲʉːlˠaː/ <VI.Ü.LA.A>

Non-palatalized consonants before a rounded vowel are customarily written using the glyphs for labialized consonants if applicable, although it is likely that their actual articulation was in between those of prototypical velarized and labialized consonants; descendants varyingly treat these sounds as labialized or velarized. Consonants that do not distinguish labialization, such as labial consonants and liquids, are written as velarized in this environment.

Next post will cover the development of the script in later Ansai languages, in particular Undreve.
[ʈʂʰɤŋtɕjɑŋ], or whatever you can comfortably pronounce that's close to that

Formerly known as Primordial Soup

Supporter of use of [ȶ ȡ ȵ ȴ] in transcription

It is pitch black. You are likely to be eaten by a 青.

User avatar
din
Avisaru
Avisaru
Posts: 779
Joined: Wed Jan 10, 2007 10:02 pm
Location: Brussels

Re: Fenja orthography: An abugida with different inherent vo

Post by din »

Nice! I like it when people come up with orthographies that make 'historical' sense.

So, since you don't talk much about the sound changes from Proto-Ansai, I'd like to know what those example words would have looked like. Are they in any way similar to what the syllabary* spells out? And would there be much potential ambiguity due to the lack of marking of the 'fake' vowels?

___
*I called it a syllabary, but I guess it's kind of an abugida-syllabary hybrid
— o noth sidiritt Tormiott

User avatar
Chengjiang
Avisaru
Avisaru
Posts: 437
Joined: Sun Feb 26, 2006 4:41 am
Location: Davis, CA

Re: Fenja orthography: An abugida with different inherent vo

Post by Chengjiang »

din wrote:So, since you don't talk much about the sound changes from Proto-Ansai, I'd like to know what those example words would have looked like. Are they in any way similar to what the syllabary* spells out? And would there be much potential ambiguity due to the lack of marking of the 'fake' vowels?
The sound changes pre-Fenja are under heavy revision and so I can't supply definite forms for those words right now, sorry. Proto-Ansai wasn't written, anyway. The ancient Ansai adapted this script from a syllabary used by speakers of another language in the region, one with a simple three-vowel system. (Actually a distant relative of the Ansai languages, but that's neither here nor there.)

Since this language is being revised anyway, I'm actually considering adding a few glyphs. Although labialization isn't phonemic on labial consonants and liquids (and the rhotic was always velarized), and thus in the current plan there aren't <MU PU BU FU VU LU RU RI> characters, I'm considering adding them since presumably the older syllabary this system derives from would have them. It would cut down on the currently rather frequent writing of round vowels after these consonants.
*I called it a syllabary, but I guess it's kind of an abugida-syllabary hybrid
Well, yes. Honestly this manages to combine elements of an abugida, a syllabary, and an abjad.

Future development and use in descendant languages

Base glyphs

(Although the family containing Fenja has many modern members, we will be looking primarily at Undreve.)

As the family developed starting from late Fenja and related dialects, the palatalized, velarized, and labialized phonemes began to diverge further in pronunciation, and so in the modern descendants they often have dramatically different values from each other. Undreve, for example, had the following as the primary outcomes of the consonants, the outcomes they had before a vowel:

[mˠ] > [m]
[mʲ] > [n]
[nˠ] > [n]
[nʲ] > [ȵ]
[nʷ] > [n] before vowels that stayed rounded, [nʷ]* otherwise
[pˠ] > [p]
[bˠ] >
[pʲ] > [t]
[bʲ] > [d]
[tˠ] > [t]
[dˠ] > [d]
[tʲ] > [tʃ]
[dʲ] > [dʒ]
[tʷ] > [t] before vowels that stayed rounded, [tʷ]* otherwise
[dʷ] > [d] before vowels that stayed rounded, [dʷ]* otherwise
[kˠ] > [k]
[ɡˠ] > [ɡ]
[kʲ] > [ts]
[ɡʲ] > [dz]
[kʷ] > [k] before vowels that stayed rounded, [kʷ] otherwise
[ɡʷ] > [ɡ] before vowels that stayed rounded, [ɡʷ] otherwise
[fˠ] > [f]
[vˠ] > [v]
[fʲ] > [s]
[vʲ] > [z]
[sˠ] > [s]
[zˠ] > [z]
[sʲ] > [ʃ]
[zʲ] > [ʒ]
[sʷ] > [s] before vowels that stayed rounded, [sʷ]* otherwise
[zʷ] > [z] before vowels that stayed rounded, [zʷ]* otherwise
[lˠ] > [l]
[lʲ] > [ȴ]
[rˠ] > [ɾ]
[j] > [j]
[w] > [w]

*The consonants labeled as labialized alveolars here in modern Undreve are actually realized as very different sounds by dialect, but their exact realizations aren't important to this discussion.

These consonants are mostly written with the same character as the corresponding Fenja consonant. /k/ deriving from /kʷ/ and the like are still written with <KU> et al. when they precede rounded vowels.

After some time a need began to be felt to distinguish between, say, /sˠa/ and /sˠi/, or /nʷutʷu/ and /nʷtʷu/, and so the long vowel glyphs began to be pressed into service for representing the short vowels when they differed from the inherent vowel. This began to happen around the time the length distinction in vowels was being lost in many dialects anyway. Vowelless versions of certain consonant glyphs were also developed, based on what were originally word-final variants of their characters. (Most descendants of Fenja have a more restricted set of final consonants than it does.) Those final consonant characters used in Undreve are <F S N M L R Y W H>, where <F S N M L R Y W> represent final /f s N N l ɾ j w/* and <H>, derived from an older final velar consonant glyph, represents a lengthening of the preceding vowel; Undreve is one of the languages that developed a vowel length distinction once again after losing the original vowel length contrast found in Fenja.

As the distinction between the original short and long vowels leveled and as a few new vowel qualities developed, the long vowel glyphs came to be just, well, vowel glyphs, used whenever the vowel following a consonant was not the "expected" inherent vowel. Which vowel is the expected one is somewhat more complex in Undreve and its relatives than it was in Fenja, because most of the modern languages have developed vowel harmony.

Undreve has the following base vowel inventory, with length additionally distinguished on all vowels:

/i ʉ u/
/e ɵ o/
/a ɑ/ (/a/ being actually cardinal open front, unlike the probably open central /a/ of Fenja)

Aside from /i u/ that continue Fenja-era /iː uː/, these have pretty complex correspondences to Fenja's vowels, but when they are written they are written as follows:

<I Ü U>
<E Ö O>
<Ä A>

<Ö> was created by modifying <Ü>, and <Ä> by modifying <A>. All of these letters represent the short vowel by themselves, requiring <H> after them to mark length. All of these letters are also omissible depending on vowel-harmonic context.

Usage

Undreve and its closer relatives only distinguish three archiphonemic vowel qualities in syllables outside the root. They will be represented here by |I A U|. They participate in vowel harmony along two dimensions: closeness and frontness. The vowels /i e/ are neutral for frontness and the vowels /a ɑ/ are neutral for closeness, yielding four vowel sets:

Front-close: /i ʉ a/
Back-close: /i u ɑ/
Front-open: /e ɵ a/
Back-open: /e o ɑ/

In each of these sets, the first vowel, |I|, is the inherent vowel of historically palatalized glyphs like <TI KI>. The second vowel, |U|, is the inherent vowel of historically labialized glyphs like <TU KU>. The third vowel, |A|, is the inherent vowel of historically velarized glyphs like <TA KA>. In a stem, where the possible vowels aren't narrowed by vowel harmony, the inherent vowels of these three groups of consonant glyph are /i u ɑ/ respectively, the close and back vowels being somewhat more common than the open and front vowels.

Undreve has contrastive tone contours on stressed syllables, deriving from simplification of old codas, but this is not indicated by the writing system.

That all said, here are some examples of this orthographic system as used in Undreve:

/tiȵaɡʉN/ <PI.NI.Ä.GU.N>
/zoːvezɑ/ <VI.O.H.VA.E.ZA>
/lɑɾfɑNtsis/ <LA.R.FA.N.KI.S>
/jɵlsaȵef/ <YI.Ö.L.SA.NI.F>

Note that, as per the sound correspondences above, several consonants have more than one possible spelling in Undreve, e.g. /z/ can be <VI> or <ZA>, as well as <ZU> if it isn't followed by one of <I E Ä A>. These unpredictable spellings are supposed to be etymological, and most of them are, but a fairly large minority are not.

*/f/ and /s/ assimilate in voicing to a following consonant, and |N| assimilates in POA to a following consonant, being spelled <M> before labial consonants and <N> everywhere else.
[ʈʂʰɤŋtɕjɑŋ], or whatever you can comfortably pronounce that's close to that

Formerly known as Primordial Soup

Supporter of use of [ȶ ȡ ȵ ȴ] in transcription

It is pitch black. You are likely to be eaten by a 青.

Richard W
Avisaru
Avisaru
Posts: 363
Joined: Sat Oct 16, 2010 8:28 pm

Re: Fenja orthography: An abugida with different inherent vo

Post by Richard W »

Chengjiang wrote:It is distinguished by being more or less an abugida but having different inherent vowels for different groups of consonants based on their historical secondary articulation.
That doesn't disqualify it from being an abugida. The Khmer writing system has different two different implicit vowels, but the difference depends on whether the consonant was originally voiced. Implicit vowels can also depend on other elements of words, such as open v. closed syllable and special effects for some final consonants.

User avatar
Chengjiang
Avisaru
Avisaru
Posts: 437
Joined: Sun Feb 26, 2006 4:41 am
Location: Davis, CA

Re: Fenja orthography: An abugida with different inherent vo

Post by Chengjiang »

Richard W wrote:
Chengjiang wrote:It is distinguished by being more or less an abugida but having different inherent vowels for different groups of consonants based on their historical secondary articulation.
That doesn't disqualify it from being an abugida. The Khmer writing system has different two different implicit vowels, but the difference depends on whether the consonant was originally voiced. Implicit vowels can also depend on other elements of words, such as open v. closed syllable and special effects for some final consonants.
That's a good point. Sorry, abugidas are the type of writing system I'm least familiar with.

Any thoughts on possibly introducing characters with an inherent /u/ (or |U| in Undreve) for the consonants that don't have phonemic labialization?

For those who might have been curious: Fenja is the name of the ancient language in Undreve; it is spelled <FA.E.N.DI.A.H> and pronounced [feȵdʒɑː]. The equivalent in Fenja itself would have been <FA.E.NI.DI.GA> and pronounced /fˠeːnʲdʲaɡˠ/, with the usual uncertainty about the exact values of consonants. Not one of the most different words between Fenja and Undreve by a long shot (compare Fenja /vʲikʲawbˠ/ > Undreve [zetsof]), but note that the vowel lengths have actually reversed.
[ʈʂʰɤŋtɕjɑŋ], or whatever you can comfortably pronounce that's close to that

Formerly known as Primordial Soup

Supporter of use of [ȶ ȡ ȵ ȴ] in transcription

It is pitch black. You are likely to be eaten by a 青.

Richard W
Avisaru
Avisaru
Posts: 363
Joined: Sat Oct 16, 2010 8:28 pm

Re: Fenja orthography: An abugida with different inherent vo

Post by Richard W »

Chengjiang wrote:Any thoughts on possibly introducing characters with an inherent /u/ (or |U| in Undreve) for the consonants that don't have phonemic labialization?
Well, I suppose you could do it with a system modelled on Persian cuneiform. In world, you could take the precedent of Tamil, and allow <PA.U> to become a ligature <PU> which was then seen as unrelated to <PA>. Indeed, in world, how did the consonant symbols arise? In the real world, they only seem to have arisen as adaptations of Egyptian phonetic complements. By contrast, syllabaries have arisen independently.

User avatar
Chengjiang
Avisaru
Avisaru
Posts: 437
Joined: Sun Feb 26, 2006 4:41 am
Location: Davis, CA

Re: Fenja orthography: An abugida with different inherent vo

Post by Chengjiang »

Richard W wrote:
Chengjiang wrote:Any thoughts on possibly introducing characters with an inherent /u/ (or |U| in Undreve) for the consonants that don't have phonemic labialization?
Well, I suppose you could do it with a system modelled on Persian cuneiform. In world, you could take the precedent of Tamil, and allow <PA.U> to become a ligature <PU> which was then seen as unrelated to <PA>. Indeed, in world, how did the consonant symbols arise? In the real world, they only seem to have arisen as adaptations of Egyptian phonetic complements. By contrast, syllabaries have arisen independently.
They arose as an adaptation of another language's syllabic characters, with each consonant having the inherent vowel of the syllabic character it was adapted from. Thus, in the early syllabary <TA> would have represented /ta/, whereas in its use as a consonant glyph in Fenja it represented /tˠ/ with the default, unwritten following vowel being /a/. I don't think this is an unreasonable transition. It is because of this origin as syllabic glyphs that I was considering having <PU BU> etc. in the first place.
[ʈʂʰɤŋtɕjɑŋ], or whatever you can comfortably pronounce that's close to that

Formerly known as Primordial Soup

Supporter of use of [ȶ ȡ ȵ ȴ] in transcription

It is pitch black. You are likely to be eaten by a 青.

Richard W
Avisaru
Avisaru
Posts: 363
Joined: Sat Oct 16, 2010 8:28 pm

Re: Fenja orthography: An abugida with different inherent vo

Post by Richard W »

Chengjiang wrote:They arose as an adaptation of another language's syllabic characters, with each consonant having the inherent vowel of the syllabic character it was adapted from. Thus, in the early syllabary <TA> would have represented /ta/, whereas in its use as a consonant glyph in Fenja it represented /tˠ/ with the default, unwritten following vowel being /a/. I don't think this is an unreasonable transition. It is because of this origin as syllabic glyphs that I was considering having <PU BU> etc. in the first place.
Then it's reasonable to have the weirdnesses of Persian cuneiform.

Post Reply