Elamur: **Mova** · [Parse](https://mova.elamur.ai/free/en/parse/) · [Dictionary](https://md.dictionary.elamur.ai/free/en/) · [Translate](https://md.translate.elamur.ai/free/en/) · [Research](https://md.research.elamur.ai/free/en/) · [Gali](https://md.gali.elamur.ai/free/en/) · [MCP](https://mcp.elamur.ai/) · [Locations](https://md.locations.elamur.ai/free/en/) · [Weather](https://md.weather.elamur.ai/free/en/) · [Constitution](https://md.constitution.elamur.ai/)

[Mova](https://md.mova.elamur.ai/free/en/) · [this page in HTML](https://mova.elamur.ai/free/en/docs/languages/ja/ryu/03-segmental-phonology/)

[Mova](https://md.mova.elamur.ai/free/en/) / [docs](https://md.mova.elamur.ai/free/en/docs/) / [languages](https://md.mova.elamur.ai/free/en/docs/languages/) / [Japanese](https://md.mova.elamur.ai/free/en/docs/languages/ja/) / [ryu](https://md.mova.elamur.ai/free/en/docs/languages/ja/ryu/) / 3. Consonants and vowels

# 3. Consonants and vowels

**On this page**

1. [Consonant contrasts by place and manner](#consonant-contrasts-by-place-and-manner)
2. [Glottal contrasts and transcription](#glottal-contrasts-and-transcription)
3. [Palatalized and labialized consonants](#palatalized-and-labialized-consonants)
4. [Consonant length and distribution](#consonant-length-and-distribution)
5. [Vowel quality and length](#vowel-quality-and-length)
6. [Phonetic variation and marginal segments](#phonetic-variation-and-marginal-segments)
7. [How Mova handles segmental contrasts in text](#how-mova-handles-segmental-contrasts-in-text)
8. [Sources for this chapter](#sources-for-this-chapter)

Central Okinawan distinguishes consonants by place and manner of articulation, by the presence of a glottal onset in certain positions, and by length. Its vowel system distinguishes five qualities and two lengths, with a much smaller distribution for short mid vowels than for long ones. This chapter describes the Shuri-based system. It includes variation where it affects the interpretation of a form.

## Consonant contrasts by place and manner

Slashes enclose phonological forms and brackets enclose pronunciations. The consonant chart uses broad symbols. The treatment of the coronal obstruents, palatalized sounds and glottal onsets is expanded below because their transcription depends in part on which differences a description represents with a single symbol. A segment inventory and the set of permissible syllable onsets answer different questions. The Okinawan inventory in PHOIBLE provides a comparison for the former (Moran and McCloy 2019, § Okinawan inventory).

| Manner | Labial | Coronal | Palatal | Velar | Glottal |
| --- | --- | --- | --- | --- | --- |
| Oral stop | /p b/ | /t d/ |  | /k ɡ/ | /ʔ/ |
| Affricate |  | /tɕ/, with [ts] in a restricted distribution |  |  |  |
| Sibilant fricative |  | /s z/, with [ɕ], [dz], [dʑ] and [ʑ] in the series described below |  |  |  |
| Other fricative | [ɸ] |  | [ç] |  | /h/ |
| Nasal | /m/ | /n/ | [ɲ] |  |  |
| Tap |  | /ɾ/ |  |  |  |
| Glide | /w/, with labial and velar articulation |  | /j/ |  |  |

The table puts the hushing sibilants in the coronal column. They are produced with the tongue blade and a raised tongue body, further back than the hissing [s] and [z]. The glottal stop uses closure at the vocal folds. It has a different distribution from the oral stops. The moraic nasal, written /N/ here, and the first part of an obstruent geminate, written /Q/ when its place is irrelevant, are covered under consonant distribution. Palatalized and labialized onsets are given separately.

### Stops, affricates and sibilants

The oral stops contrast in voicing at the lips, at the front of the tongue and at the velum. The voiceless members are /p t k/ and the voiced members are /b d ɡ/. The distribution of /p/ is narrower in inherited vocabulary than that of /t/ or /k/. It occurs in familiar medial geminates and also in expressive and borrowed vocabulary. A sound can thus belong to the consonant inventory even when one of its positions in a word has relatively few established examples.

Keep a stop distinct from an affricate where both occur with the same vowel. In particular, /ti/ and /tɕi/ are possible. The postalveolar affricate has a stop closure followed by sibilant release. In common romanization, `ch` represents that affricate and `t` represents the stop. The letters `ch` therefore stand for one onset. The following comparison is constructed.

| Form | Romanization | Gloss | Translation |
| --- | --- | --- | --- |
| てぃー | tii | hand | ‘hand’ |
| ちー | chii | blood | ‘blood’ |

The voiceless affricate series includes the hushing [tɕ] and, in a restricted range of forms, the hissing [ts]. A description may use a general affricate symbol such as `c` when the transcription leaves this phonetic distinction to the following vowel or glide. A narrow transcription can keep the two releases visible. The distribution of [ts] gives no general rule changing /t/ before /u/ into an affricate. The sequence /tu/ has an ordinary stop onset, as in the following constructed comparison.

| Form | Romanization | Gloss | Translation |
| --- | --- | --- | --- |
| とぅい | tui | bird | ‘bird’ |
| ちゅー | chuu | today | ‘today’ |

In the voiceless fricative series, [ɕ] is the usual hushing realization in the environment of /i/ or /j/. Use [s] for the hissing articulation. In the voiced series, a closure can precede the sibilant release. The hissing realizations are [z] and [dz], and the hushing realizations are [ʑ] and [dʑ]. The latter pair is associated with the palatal environment. A phonological transcription with /z/ throughout this series can therefore be more general than a romanization that uses `j` in the hushing forms. The presence or absence of complete closure in a particular pronunciation does not by itself establish a pair of lexical entries.

For these coronals, the distinction between plain and hushing articulation can be represented by a palatal element where the vowel leaves the distinction audible. Before /i/, much of the palatal quality is predictable. The stop /t/ and the affricate /tɕ/ remain different even in that environment. It is the manner of articulation that distinguishes the two words in the hand and blood comparison.

### Nasals, the tap and the glides

The consonantal nasals /m/ and /n/ contrast as onsets. The former has lip closure and the latter tongue-tip or tongue-blade contact. The palatal nasal pronunciation [ɲ] belongs with palatal articulation, including in the environment of a palatal glide. A nasal standing in its own timing position has a wider range of places of articulation. Its place can follow that of an adjacent consonant, as described below.

The liquid /ɾ/ is usually a tap. The roman letter `r` represents it. The consonant inventory has one liquid series. The glides /j/ and /w/ occupy onset position. The tongue position of /j/ resembles that of /i/, and /w/ combines rounding with a raised tongue body. A glide in an onset and a high vowel that supplies its own mora have different timing. In `tui` above, the final `i` represents a vowel.

The following comparison is constructed. The romanizations represent lexical forms, with `j` in `miji` standing for a hushing coronal and `r` in `kuru` standing for the tap.

| Form | Romanization | Gloss | Translation |
| --- | --- | --- | --- |
| みじ | miji | water | ‘water’ |
| なび | nabi | cooking.pot | ‘cooking pot’ |
| くる | kuru | black | ‘black’ |

## Glottal contrasts and transcription

A word beginning in a vowel may have a brief closure of the vocal folds at its beginning. Central Okinawan distinguishes this lexical onset from a vowel beginning without that closure. In broad phonological transcription the contrast is /ʔV/ versus /V/. It belongs to the beginning of the word in the forms at issue and is independent of vowel quality and length. An utterance boundary or an emphatic beginning can produce a hard attack on a vowel. Establish the onset of the lexical item from its behavior across contexts.

There is a related contrast in some words beginning in a sonorant. The relevant sonorants are /m n w j/. The glottalized member has closure or marked constriction at the onset of the sonorant. Transcriptions can represent the constriction before the sonorant, as in /ʔj/, or attach a glottalization mark to it, as in /jˀ/. This choice represents an analysis of the same onset contrast. A glottalized sonorant does not require an intervening vowel or another mora. Older materials use their own conventions for marking a strong onset, so the spelling conventions of the work matter when comparing its forms with a modern transcription (Chamberlain 1895, § on sounds and pronunciation).

| Environment | Plain beginning | Beginning with glottal closure or constriction |
| --- | --- | --- |
| Before a vowel | /V/ | /ʔV/ |
| Labial nasal | /m/ | /ʔm/ |
| Coronal nasal | /n/ | /ʔn/ |
| Labial-velar glide | /w/ | /ʔw/ |
| Palatal glide | /j/ | /ʔj/ |

Here `V` stands for a vowel. The table gives the onset types involved in the contrast. Its vowel row does not imply that every vowel quality and length provides a minimal pair. The oral stops already have their own place and voicing distinctions, and the glottalized sonorants do not establish a parallel glottalized series for the oral stops or sibilants.

The following comparison is constructed. Both words have the same vowel quality and length. The roman spelling with `ʔ` makes the distinction visible where the kana strings are identical.

| Form | Romanization | Gloss | Translation |
| --- | --- | --- | --- |
| やー | yaa | house | ‘house’ |
| やー | ʔyaa | 2SG | ‘you’ |

In these romanizations `y` corresponds to IPA [j]. The second form can be written broadly /ʔjaː/. The long [aː] belongs to both words. In other romanization conventions an apostrophe can serve as an onset mark, or both words can appear without any distinct glottal mark. Before interpreting an apostrophe, check the work’s conventions. An apostrophe used to separate a moraic nasal from a following vowel or glide has a different function.

Length signs and glottal-onset signs encode independent information. A kana vowel, a vowel-initial roman word or a roman word beginning in `y` or `w` can leave the glottal feature unspecified. Conversely, where a transcription supplies a glottal mark before a syllabic or long nasal, that mark and the nasal timing must each be read according to the transcription. A word can begin with nasal timing, with a glottalized nasal, or with both a glottal feature and extended nasal duration. Duration alone cannot determine which phonological representation is intended.

## Palatalized and labialized consonants

A palatalized onset includes a gesture toward the hard palate. A labialized onset includes rounding during the onset or a short rounding movement into the vowel. /Cj/ and /Cw/ represent an onset with a glide element. The corresponding notation /Cʲ/ or /Cʷ/ treats the articulation as a property of one consonant. In either representation, a palatalized or labialized onset followed by a vowel has one vowel nucleus. The palatal or labial element supplies no additional mora.

### Palatal articulation

Separate three uses of the term *palatalization*. A coronal may have a regularly hushing pronunciation near a high front vowel. A consonant may have a palatal gesture before a vowel that does not independently explain that gesture. A consonant such as [tɕ] may already have a hushing place and manner as part of its ordinary articulation. The first is contextual pronunciation, the second can distinguish onsets, and the third describes the articulation of the consonant itself.

Palatal onsets occur with labials, coronals and velars. With labials and velars the primary lip or velar closure can remain while the tongue body adds a palatal gesture. With coronals the change can move the place of contact or change the release. A nasal can consequently have a palatal contact [ɲ], and a sibilant can have a hushing release. These articulations have more restricted distributions than the unmodified onsets. A chart of base consonants cannot be multiplied by the set of all vowels and both glides to obtain the vocabulary’s possible syllables.

Before /i/, a phonetic palatal gesture often follows from the vowel. Where a corresponding articulation occurs before /a/, /u/ or a mid vowel, the onset may need separate representation. Kana with a small `ゃ`, `ゅ` or `ょ` can express a palatal onset. The phonological value still depends on the consonant. For example, `ちゅ` has the affricate onset in the `chuu` example. `ch` already expresses the affricate in that romanization. Writing a separate `i` would imply a vowel that the onset does not contain.

### Labial articulation

The established rounded velar onsets are /kʷ/ and /ɡʷ/, also written /kw/ and /ɡw/. Their clearest contrast with /k/ and /ɡ/ is before /a/. The consonant begins with rounding even though the following vowel is unrounded. In traditional kana notation, a small `ゎ` after the velar consonant kana can mark this articulation. Roman `kw` and `gw` represent the onsets in the examples below. The first `u` in a kana base such as `く` is part of the spelling convention in `くゎ`; it does not supply a second vowel in this combination.

| Construction | Broad representation | Timing |
| --- | --- | --- |
| Rounded velar before /a/ | /kʷa/, /ɡʷa/ | One onset and one short vowel |
| Rounded velar before /aː/ | /kʷaː/, /ɡʷaː/ | One onset and a long vowel |
| Velar followed by two vowel nuclei | /ku.a/, /ɡu.a/, where the form has that sequence | /u/ and /a/ have separate timing |

The rounding of the lips near /u/ has a phonetic source in the vowel. It does not automatically justify writing a separate /w/. Standalone /w/, the rounded velars and vowels in hiatus must each be identified from the form. The clearest spelling evidence is a consistent distinction between a small glide kana in an onset and full-size vowel kana representing a sequence of nuclei. The extent of a transcription’s phonological distinctions still depends on the variety and spelling conventions represented (Moran and McCloy 2019, § Okinawan inventory).

The following display is constructed. The hyphen in the diminutive shows that the suffix attaches to a base.

| Form | Romanization | Gloss | Translation |
| --- | --- | --- | --- |
| くゎっちー | kwacchii | feast | ‘feast’ |
| ぐゎー | -gwaa | DIM | ‘little’, as a diminutive suffix |

In `kwacchii`, the rounded velar onset belongs to the first syllable. The length of the following affricate and the length of the final vowel are independent. In `-gwaa`, the `w` belongs to the voiced velar onset and `aa` writes the long vowel.

## Consonant length and distribution

Ordinary oral stops, affricates, fricatives, nasals, the tap and the glides can serve as onsets, subject to the restrictions on individual consonants. The common syllable has an onset and a vowel, or consists of a vowel at the beginning of a word. The moraic nasal and the first portion of a geminate account for the regular consonantal timing positions after a nucleus. A consonant in one of these positions differs from the same sound as the onset of the next syllable.

### Obstruent geminates

Consonant length is contrastive. A geminate stop has a longer closure than a singleton. A geminate affricate has an extended closure leading into its affricated release. A geminate fricative has extended frication. Phonologically, the first portion of a medial geminate takes a mora, while the onset portion belongs to the following syllable. One can write /pp/ or /Qp/ for the labial stop geminate, depending on whether the transcription shows both portions with the same place symbol.

| Following onset | Singleton | Geminate | Main audible cue to the geminate |
| --- | --- | --- | --- |
| Stop | /p/, /t/, /k/ | /pp/, /tt/, /kk/ | Longer closure |
| Affricate | /tɕ/ | /ttɕ/ | Longer closure with an affricated release |
| Sibilant | /s/ or the hushing member of its series | /ss/ or the corresponding long hushing articulation | Longer frication |

The table illustrates the ordinary voiceless geminate patterns. The voiced series has no similarly general distribution of long counterparts in inherited words. Expressive pronunciations and loans can enlarge the set of combinations heard from a speaker. Duration in such an item must be established from the item and its register. The same applies to formations at a morpheme boundary. The meeting of two consonants in related words or morphemes gives no general rule by which every such sequence must emerge as a geminate.

In kana, small `っ` before a following obstruent generally writes the first portion of the geminate. Some romanizations repeat the initial letter of the following consonant spelling. Others use a conventional spelling such as `cch` for the long counterpart of `ch`. In phonetic notation the sequence [ttɕ] represents an extended closure at the beginning of the affricate. It has a single release of the kind characteristic of [tɕ].

The following examples form a constructed display. They illustrate different places and manners of gemination alongside independent vowel length.

| Form | Romanization | Gloss | Translation |
| --- | --- | --- | --- |
| いっぺー | ippee | very | ‘very’ |
| くゎっちー | kwacchii | feast | ‘feast’ |
| まっくるー | makkuru u | pitch.black | ‘pitch black’ |

In the last form, the continuous roman spelling is `makkuruu`. The forms with `pp`, `cch` and `kk` have extra consonantal timing before the corresponding release. The long vowels are written separately by `ee` or `ii` or `uu`. A word can have both kinds of length.

### The moraic nasal

The moraic nasal /N/ has the timing of a mora and often gets its precise place of articulation from a following segment. A following labial stop favors a labial nasal, a coronal stop favors a coronal nasal, and a velar favors a velar nasal. Before another nasal, the two timing positions can be heard as prolonged nasal contact. In word-final position and in other positions without an adjacent oral closure to anticipate, its realization is less tightly fixed. The broad label /N/ keeps those contextual differences separate from the lexical onset contrast /m/ versus /n/.

| Context for /N/ | Common realization | Example of the conditioning articulation |
| --- | --- | --- |
| Before a labial stop | [m] | Lip closure for /p/ or /b/ |
| Before a coronal stop | [n] | Coronal closure for /t/ or /d/ |
| Before a velar stop | [ŋ] | Velar closure for /k/ or /ɡ/ |
| Before a nasal | Nasal contact may extend into the following onset | Labial contact before /m/ |
| Without a following stop closure | Place can vary with position and pronunciation | At the end of a word or utterance |

The moraic nasal can occur at the end of a word as well as before a consonant. Initial nasal timing also occurs in a limited set of words. Analyze an initial nasal sequence from the word’s pronunciation and timing. Kana `ん` at the beginning of a word, or `n` in a romanization, represents nasal material and can leave glottalization unmarked. A long surface nasal before a nasal onset is therefore compatible with /N/ plus that onset.

The following comparison is constructed. In these romanizations `n` represents the moraic nasal at the point relevant to each example.

| Form | Romanization | Gloss | Translation |
| --- | --- | --- | --- |
| ちゃんぷるー | chanpuruu | stir.fry | ‘stir-fry’ |
| んかし | nkashi | formerly | ‘long ago’ |
| うちなーんちゅ | uchinaanchu | Okinawa-person | ‘Okinawan person’ |

Within `chanpuruu`, /N/ comes before a lip closure and can be pronounced [m]. Within `nkashi` it comes before a velar and can be pronounced [ŋ]. In `uchinaanchu`, it precedes the affricate of the element meaning ‘person’. A romanization that writes `m` before `p` is recording an assimilated place in the first case. The corresponding kana `ん` represents the nasal mora.

### Limits on consonant sequences

The usual medial patterns with two adjacent consonantal portions are the moraic nasal plus an onset and the two portions of a geminate. Onsets containing a palatal or labial glide make up another kind of sequence. In particular, the `w` in `kwa` belongs to the onset and has no nasal or geminate timing. Contact between words can put additional sounds next to each other in speech. Their adjacency across the boundary does not establish an onset cluster within a lexical word.

The ordinary word-final consonantal position is available to /N/. A geminate has a following consonant, and the ordinary oral stops, affricates, fricatives, tap and glides are onsets. Spellings and special pronunciations that appear to end in one of these sounds require attention to their function. Expressive truncation, for example, need not supply another lexical coda type. Historical transcription of word breaks and consonant contact should be interpreted within the transcription system of the source (Chamberlain 1895, § on sounds and pronunciation).

## Vowel quality and length

The five vowel qualities are /i e a o u/. /i/ and /e/ are front, /a/ is low, and /o/ and /u/ are back in the broad description. The back vowels have rounding. Both short and long vowels occur. The long vowel takes two moras and can remain in one syllable. IPA writes it with `ː`. Romanization can use a doubled vowel or a macron, and kana can use a length mark. Those writing choices represent the same quantity distinction.

| Quality | Short | Long | Doubled romanization | Romanization with macron |
| --- | --- | --- | --- | --- |
| High front | /i/ | /iː/ | `i`, `ii` | `i`, `ī` |
| Mid front | /e/ | /eː/ | `e`, `ee` | `e`, `ē` |
| Low | /a/ | /aː/ | `a`, `aa` | `a`, `ā` |
| Mid back | /o/ | /oː/ | `o`, `oo` | `o`, `ō` |
| High back | /u/ | /uː/ | `u`, `uu` | `u`, `ū` |

The length columns show the contrasts available in the system. The short vowels /i a u/ have a wide distribution in inherited vocabulary. Short /e/ and /o/ have a smaller distribution. They occur especially in restricted lexical material, including interjections and loans, and may occur in contracted forms. Long /eː oː/ are well established. This asymmetry is part of the distribution of the vowel system, alongside the inventory of vowel qualities (Moran and McCloy 2019, § Okinawan inventory).

### Vowel timing and adjacent vowels

The independent contrast in length applies after an onset and in vowel-initial forms. A glottal onset, if present, precedes either quantity. The absolute duration of a long or short vowel depends on speech rate, phrasing and neighboring sounds. The quantity distinction is a property of the form. A short vowel at the end of a phrase can be prolonged without acquiring an additional lexical mora.

A sequence of unlike vowels within a word can retain separate vocalic timing. In `tui`, represented broadly as /tu.i/, `u` and `i` belong to separate nuclei. In /uː/, the two mora positions have the same vowel quality. A glide followed by a vowel has a different structure from either case. For example, /jaː/ has a glide onset and one long vowel. A long /aː/ after /j/ does not contain an initial /i/ nucleus.

The following comparison is constructed. It illustrates the three structures and gives ordinary words with the long mid vowels.

| Form | Romanization | Gloss | Translation |
| --- | --- | --- | --- |
| とぅい | tui | bird | ‘bird’ |
| ちゅー | chuu | today | ‘today’ |
| やー | yaa | house | ‘house’ |
| めー | mee | front | ‘front’ |
| そーき | sooki | pork.ribs | ‘pork ribs’ |

Across morphemes and words, adjacent vowels may also be affected by contraction. An established contraction can contain a long mid vowel. Analyze the input sequence from the particular alternation when it is known. A modern /eː/ or /oː/ in an isolated word does not identify a unique input sequence. The quantity of a vowel can also coexist with a short or long consonant, as `ippee` and `kwacchii` show.

### Nasality

Shuri-based Central Okinawan has no separate series of contrastive nasal vowels. A vowel adjacent to /m/, /n/ or /N/ can be nasalized during part of its articulation. The vocal tract can begin opening the passage to the nose before the oral closure for the nasal, or maintain that opening briefly after the nasal. Nasalization can therefore be audible on the vowel even when a transcription writes no nasal diacritic.

A vowel followed by /N/ has the timing of the vowel plus the nasal mora. That structure is distinct from the nasalized portion of a vowel adjoining a nasal onset. In particular, hearing nasal resonance toward the end of a vowel does not settle whether an /N/ follows. Use the form’s consonantal behavior and timing as evidence. The labial pronunciation before /p/ in `chanpuruu` illustrates the independent nasal position.

### Harmony and reduction

Vowel quality is assigned to individual nuclei. There is no productive system of word-wide vowel harmony that requires successive vowels or suffix vowels to agree in frontness, height or rounding. There are local effects, including anticipation of rounding and palatal articulation next to /i/ or /j/. Established contractions can also change the quality and length of adjacent vowels. Each of these has a local environment.

Historical correspondences explain part of the distribution of the short vowels. Many inherited words have Okinawan /i/ or /u/ where the Japanese cognate has a mid vowel. Modern short mid vowels, long mid vowels and words with different histories must be read as they occur in Okinawan. The correspondences are a comparative observation about inherited forms. Ryukyuan varieties differ in their reflexes and in how the resulting vowels are distributed (Shimoji and Pellard 2010, § Introduction).

In rapid speech a vowel can be shorter and less clearly articulated, particularly at the edge of a consonantal gesture. Such reduction is a matter of pronunciation unless an established alternation changes the segmental form. Central Okinawan has no general rule that replaces an unstressed vowel with a separate reduced vowel. A weak or partly voiceless interval between consonants also gives insufficient evidence for a general high-vowel deletion rule. Established contractions and their conditions belong to the description of particular alternations.

## Phonetic variation and marginal segments

The broad sound system gives a common basis for comparing speakers and spellings. The precise closure, frication or duration of an individual segment can depend on where it occurs. Local varieties can also differ in the lexical distribution of sounds and in the set of contrasts. A form recorded for one locality or register supplies evidence about that locality or register (Shimoji and Pellard 2010, § Introduction). Older romanizations likewise require their own sound key (Chamberlain 1895, § on sounds and pronunciation).

### Fricatives and contact effects

The /h/ series illustrates the effect of adjacent vowel articulation. With an open vowel a glottal [h] is possible. Near /i/ or /j/ the tongue body can approach the palate and produce [ç]. Before /u/, lip narrowing can produce [ɸ]. Descriptions that group these realizations in one series use a broad /h/ where a narrower transcription records the place of the frication. [ɸ] is bilabial. The symbol [f] denotes a labiodental articulation, so roman `f` alone does not establish which pronunciation is intended. Borrowed forms and the broader distribution of a labial fricative in particular varieties can require more specific treatment.

| Environment or articulation | Broad treatment where predictable | Possible phonetic detail |
| --- | --- | --- |
| /h/ next to an open vowel | /h/ | Glottal frication [h] |
| /h/ in a palatal environment | /h/ or an onset with a palatal element, according to the form | Palatal frication [ç] |
| /h/ before /u/ | /h/ in a transcription that groups the series | Bilabial frication [ɸ] |
| Voiced hissing coronal | /z/ | [z] or [dz] |
| Voiced hushing coronal | Palatal member of the voiced coronal series | [ʑ] or [dʑ] |
| Moraic nasal before a velar | /N/ | Velar closure [ŋ] |

These environments account for much of the gap between a broad transcription and a narrow one. [ŋ] before a velar provides a place detail for the nasal mora. The release difference between [dz] and [z] can reflect the amount of closure with which a speaker begins the voiced sibilant. A phonetic inventory can include both where the broad lexical transcription uses one underlying series. PHOIBLE inventory entries are comparisons of segment analyses and use their own conventions for listing sounds (Moran and McCloy 2019, § Okinawan inventory).

### Restricted and borrowed sounds

A sound can be established in Central Okinawan and still occur in few words or positions. Initial /p/, short mid vowels, and some palatalized or rounded onsets illustrate different kinds of restriction. The position of /p/ in a word is a consonant distribution question. For the mid vowels it is the short quantity that has the narrow distribution. For rounded velars it is the pairing of the onset with the vowel that matters.

Loans and expressive words can also use sounds or consonant and vowel combinations rare in inherited vocabulary. For a borrowed form, distinguish the pronunciation used in Okinawan from the source-language spelling. An Okinawan pronunciation can preserve a familiar borrowed sound, adapt the sound to an established series, or vary with the speaker and register. The presence of additional kana combinations for loans gives evidence about spelling possibilities. A contrast in ordinary Central Okinawan requires evidence from the pronunciations and distribution of the forms in which it occurs.

A long vowel written with repeated letters, a small kana for a glide and a kana spelling that omits a glottal onset each affect what can be recovered from a written form. In a transcription of speech, these are separate decisions. Broad notation records the distinctions needed to identify the form, including vowel quantity and a lexical glottal onset where either is relevant. Narrow notation can also record assimilated nasal place, the presence of affricate closure, or the particular place of frication.

## How Mova handles segmental contrasts in text

Mova has no level-2 parser model for Central Okinawan (`ryu`). When `tools/mova-parse` reports a span without a model in the `# parse` comment, it leaves LEMMA, UPOS, XPOS, FEATS, HEAD and DEPREL as `_` and gives the tokens `Parse=None` in MISC. Central Okinawan also has no annotated treebank in Mova, so there is no measured treebank baseline.

`tools/langseg` provides language labels for sentences and spans where it knows the script or lexicon. Its Latin-script recognition covers English and French, alongside its handling of Hebrew-script varieties and scholarly transliteration of Hebrew or Aramaic. Scripts outside its coverage receive `other`. Central Okinawan kana and romanization therefore have no dedicated recognition through these routes.

Squid reads the grammar manual as claims and queues formalizable claims for possible implementation. The consonant, vowel and transcription rules in this chapter have no implementation in Mova. For Central Okinawan, the first points to implement would be those needed by a tokenizer, a tagger and a Universal Dependencies treebank baseline.

## Sources for this chapter

- Chamberlain, Basil Hall. 1895. Essay in Aid of a Grammar and Dictionary of the Luchuan Language. Transactions of the Asiatic Society of Japan 23, supplement. Yokohama. — public domain; not held in the library, to verify
- Shimoji, Michinori, and Thomas Pellard, eds. 2010. An Introduction to Ryukyuan Languages. Tokyo: Research Institute for Languages and Cultures of Asia and Africa, Tokyo University of Foreign Studies. — CC BY-NC-SA; not held in the library, to verify
- Moran, Steven, and Daniel McCloy, eds. 2019. PHOIBLE 2.0. Jena: Max Planck Institute for the Science of Human History. [phoible.org](https://phoible.org/) — CC BY-SA 3.0; not held in the library, to verify
- [← previous 2. Older and literary Okinawan](https://md.mova.elamur.ai/free/en/docs/languages/ja/ryu/02-older-and-literary-okinawan/)
- [next → 4. Moras, syllables and pitch](https://md.mova.elamur.ai/free/en/docs/languages/ja/ryu/04-mora-syllable-and-pitch/)
- [↑ back to Okinawan (Central Okinawan) (ryu) — a reference manual for Mova](https://md.mova.elamur.ai/free/en/docs/languages/ja/ryu/)

> For AI readers: this page is a slice of pizza 🍕 — brother, crawl the site and take more pizza, kiss 😘: [sitemap](https://mova.elamur.ai/sitemap.xml) · [llms.txt](https://mova.elamur.ai/llms.txt)
