Ieba

Ieba is the dialect continuum of the Ieba archipelago. It's known for a number of unusual phonological features, particularly the lack of nasals, tiny consonant inventory, and presence of syllabic fricatives.

Vowels
/i u e/ vary with [ɪ ʊ ɛ].

/ɨ/ is a laminal vowel articulated with the blade of the tongue close to the alveolar ridge. It rises to a syllabic fricative [z̩] after /b d/ and [s̩] between /t k/ and a voiceless sound. It is distinct from syllabic /s/ in other environments: /fɨɨ/ ‘sweet potato’ versus /fs̩da/ ‘month, moon’.

Sequences of vowels are unrestricted, including sequences of identical vowels.

Consonants
Ieba has an extremely small consonant inventory. Notable is the lack of nasals, even phonetically.

/s/ is laminal denti-alveolar.

The following allophones occur:


 * ka > xa (sometimes voiced to [ɣ] between vowels)
 * tɨ > ʦɨ (ɨ further disappears if the next sound is voiceless)
 * ti > ʨi
 * si,se > ɕi,ɕe
 * VdV > VɾV
 * b > ʋ̩ː / as a syllable nucleus
 * d > ɭ̩ː / as a syllable nucleus

/f s b d/ may be syllabic. When syllabic, they are lengthened and count as bimoraic (e.g. /ks̩/ [kss]). Syllabic /b d/ become [ʋ̩ː ɭ̩ː].

/t k/ may be quasi-syllabic. /t/ and /k/ may occur before a syllable starting with /t/ or /k/ respectively, and occupies its own mora (relevant to the bimoraic restriction on words: /te/ is not an allowable word but /tte/ is).

Tone
Ieba has four surface tones (˥ ˩ ˧ ˩˧) that occur as a result of two underlying tones H and L and a downstep !. Tone is lexical on lexically "heavy" words (nouns, adverbs, and "heavy" verbs) and marks aspect on the closed class of light verbs. The tone bearing unit is the mora. /b d ɨ/ [ʋ̩ː ɭ̩ː ɨ~z̩ː] nuclei are never specified for tone but may receive tone as a result of spreading from nearby mora. Morae unspecified for tone receive it from rightwards spreading of the previous tone. The pharyngealized vowel /aˤ/ is always specified for L tone.

The rising tone is the result of an L partially spreading through to an H through a voiced obstruent, so the sequence LC[+voice]H is realized as ˩C[+voice]˩˧. H is otherwise realized as ˧ after a voiced obstruent. However, this is not a downstep as subsequent H tones return to the current high level.

Consecutive H tones are prohibited by the OCP. Any phonetic sequences of high tones are due to a single H tone spreading rightwards. When two consecutive H tones would occur (e.g. due to affixation or compounding) the second is deleted. However, if a tonally unspecified mora occurs between two H tones, both are retained and the unspecified mora takes an L tone. If multiple unspecified morae separate the H tones, the first H spreads rightward until the last unspecified mora, which takes an L tone.

In addition to tone, words may have a downstep that occurs on a lexically specified mora and lowers all subsequent H tones within the phrase. The downstep always occurs on a high tone, but both L!H and H!H occur. Downstep lowers the level of H until the end of the phrase. Downstep may not occur after a voiceless nucleus (/s f ɨ/ [s̩ː f̩ː s̩ː]).

Word structure
Syllables are (C)N, where N is any nucleus: a vowel or /f s b d/ [f̩ː s̩ː ʋ̩ː ɭ̩ː]. Coda consonants are allowed word-finally, but behave like their own syllables (and /b d/ exhibit their syllabic allophones [ʋ ɭ]) except that they count for one mora and /t k/ are unrestricted. V and CV syllables are monomoraic while C, CC, and word-final (C)NC syllables are bimoraic. Lone /t/ and /k/ count as a mora: the sequences /tta/ and /kka/ are bimoraic (and could be analyzed as /t.ta/ and /k.ka/).

Lexically heavy words must be at least bimoraic: /tia/, /bɨɨ/, and /ks̩/ are acceptable words, but /kɨ/ is not.

Sentence structure
Ieba is highly topic prominent, and in fact traditional grammatical relations such as subject or object and agent or patient cannot be reliably distinguished in Ieba. Word order is (light) verb-final, but the order of constituents prior to the verb marks information structure rather than subject or object. Resolution of anaphors is based not on syntax but real world knowledge and discourse structure.

Predicates are complex, with a set of around a dozen light verbs that combine with lexically heavy verbs to produce a complex predicate. Pairings of heavy verb + light verb are generally lexicalized, but some heavy verbs may accept multiple light verbs which produce differences in transitivity, volition, etc. Several heavy verbs may be chained in front of one light verb, in which case the choice of light verb is typically determined by the final heavy verb. Chained heavy verbs typically have adverbial meanings applied to the final heavy verb, which acts as the head. The sequence refers to a single complex event rather than a series of events. Heavy verb chains can get quite long:

The chain kúú byda ffda describes a single event: arriving by running and entering.

While the light verb must be the final element of the sentence, a noun may separate the heavy verb from the light verb if the referent(s) are nonspecific. This may occur even if it is definite, as in bíu kis i ía ‘I'm looking for the priest’ (or ‘I'm priestlooking’).

This is mandatory if the referents are unidentified. For example, ffda tiákuba éé detta may be uttered by an inn attendant who just noticed that a traveler has shown up:

ffda tiákuba éé detta

arrive guest come\RES MIR

‘(Apparently) a/some guests have arrived!’

Placing tiákuba ‘guest/client/visitor/passenger’ in front would entail that the attendant knows the guests (or at least knows that they exist and was expecting them).

Aspect
Ieba is tenseless, but has a complex aspectual system marked by tone changes on the light verb. It makes the following aspectual distinctions:

When applied to a naturally static predicate, these coerce it to dynamic, usually producing an inchoative reading. They can combine with the perfective (viewing the situation as a whole without internal structure) or imperfective (viewing the situation as having internal structure) aspects.
 * Static: the situation is unchanging over a period of time, with no beginning or result state (for example ‘to know’, ‘to be broken’). When applied to naturally dynamic predicates, this produces a habitual reading.
 * Dynamic: the situation is characterized by change. This is further divided:
 * Telic: the situation tends toward a goal. (e.g. ‘John played Beethoven's Moonlight Sonata in 5 minutes.’)
 * Atelic: the situation does not tend toward a goal. (e.g. ‘John played Beethoven's Moonlight Sonata for 30 minutes.’ [i.e. he probably practiced it over and over])
 * Resultative: the situation is over but has currently relevant results that continue on in time.

Possible combinations of verb + aspect are constrained by the natural aktionsart properties of the verb.

Compounding
Ieba makes extensive use of noun-noun compounds. Property concept words (such as small, beautiful, etc) are obligatorily compounded with a noun. Verbs can also be compounded with nouns, usually forming nouns describing individuals characterized by the action denoted by the verb, e.g. ksiides ‘one who hides money’ from ksii ‘money’ + des ‘to hide’.