Anthologica Universe Atlas / Universes / The Allosphere / Kaam-Yerte / Ubghuu / Ubghuu Grammar / Phonology and Morphophonology / Phonology

Phonemic Inventory

Ubghuu has the following inventory of sounds:


LabialPalatalized LabialDentalPostalveolarPalatalVelarUvularGlottal
Aspirate Stoppʲʰtsʰʈʰ
Tenuis Stopptsʈkqʔ
Voiced Stopbdzɖgɢ


Ubghuu distinguishes between the oral vowels /a ɛ ɔ e o i u/ and the nasal vowels /ã ɛ̃ ɔ̃ ĩ ũ/. All of these vowels exhibit contrastive length.


There are three contrastive tones: high, mid, and low. Long vowels may exhibit a tone contour in some context; the allowable contours are high-mid and mid-low.

Syllable Structure

Ubghuu allows a maximal syllable of CCVC, where the initial consonant clusters are limited to a) a glottal stop plus any consonant; b) a stop plus any of /w ɳ r ɭ ʐ/ c) a homorganic nasal plus any consonant or d) /zɖ/, /sʈ/ or /sʈʰ/. Thus /ŋg, ʔq, pɳ/ are valid initial clusters, whereas /ŋb, bʔ, pʈ/ are not.

Word-finally, -CC clusters can sometimes occur morphologically, These are always broken up by epenthetic vowels (on which see below).


Ubghuu observes several allophonic rules.

The voiced stops are somewhat devoiced at the beginning of words and are pronounced with a strong breathy element, as in dzaàk [d̤zʱâːk], gáab [g̤ʱâːb]. An exception applies to /ɢ/, which is realized (and written) as [ʁ] everywhere except word-finally and next to the glottal stop.

All pulmonic, non-aspirated stops undergo weakening between vowels word-internally, with tenuis stops voicing and voiced stops fricating. /q/ and /ɢ/ both fall together as [ʁ], and the voiced stops generally merge with the corresponding fricative. An exception is /ɖ/, whose weakened equivalent can be realized as any of [v ð z ɽ ɻ ɭ ʐ] depending on the dialect.

The fricatives are devoiced word-finally or when adjacent to a voiceless stop, including the glottal stop: ʐɛɛ̀w, but miidzį̀f and ʔsáap.

The velar nasal /ŋ/ weakens to a glottal approximant between vowels, nasalizing the following vowel; |ya-ŋúb-aa| -> <yahų́ų́váa>.

Sequences of vowels are broken up by epenthetic consonants; if the vowels on both sides of the sequence are oral, the consonants will be /j/ for following front vowels, /w/ for following back vowels and /h/ if the second vowel is /a/; if any of the vowels are nasal, the epenthetic consonant will also be /h/. If it is the first vowel in the hiatus that is nasal, the second will additionally assimilate in nasality: |u-gɛ̀ɛ̀n-dzaŋ-ii-aa| -> <ughɛ̀ɛ̀ndzàhį̀į̀hąą̀>.

When the glottal stop occurs before a consonant, it drops unless in pausa, when it usually provokes vowel epenthesis in non-careful speech.

Similarly, final consonant sequences as well as globally impermissible clusters produce epenthesis as well. In this case, the cluster is broken up with a short echo vowel.


The mid tone is unmarked; either of the marked tone spreads rightward until another marked tone is encountered or there is a glottal stop (for this, see here).

The tone contour of a word must always be falling. If the last two morae of a word distribute over two syllables, the first will keep its tone and the second experience downstep (high to mid or mid to low) unless they would already both be low. In this case, the first syllable will have its tone lifted to mid.

Note that this applies even if the first syllable has two morae itself — in this case, the entire long vowel will receive mid tone.

If the word ends in a long vowel, then it takes either high falling (ordinary high) or mid falling tone (ordinary mid or low).

Note that for these purposes, a word may contain several tonal clitics (such as true adjectives, the topic marker ŋa, the demonstrative ma etc.)