My main source of inspiration for this language was North Sámi, particularly its phonology and morphophonology. That's where I cribbed the preaspirated consonants from. (Before you ask: yes, /ʰʰp ʰʰt ʰʰk/ are a real thing.) The consonant gradation is actually fairly straightforward compared to North Sámi. In Sámi, there are three consonant grades, but any particular word only ever takes two of them; which combo (e.g. 1+3, 2+3) a word takes is lexically determined. In Ayuma, there are three consonant grades and each lexeme takes the same grade in each particular inflection/declension.
Today's topic of interest: Dependent Clauses
Ayuma is a topic-prominent language. The general gist of topicalization is that
1. The topic is always definite.
2. Only subjects and direct objects may be topicalized. Indirect objects may be raised to the status of direct object and subsequently topicalized.
3. The topic tends to be a constant which figures in several consecutive sentences; to topicalize a newly introduced referent (given that it does not immediately appear in a subsequent sentence) is a marked construction which confers focus. Otherwise, topicalization is not a marked construction.
4. The "topic form" of a noun sometimes functions in a case-like manner, with select verbs.
I'm currently thinking about how dependent clauses work. I want to do away with nonfinite verbs as much as possible and replace such constructions with nominalized phrases, but I'm afraid sometimes the results get a bit too verbose for my liking.
So, relative clauses are simply nominalized verb phrases whose referent is identical to its topic. They are formed with the clitic =tehht/dehht. Examples:
Code: Select all
(1)
pidin tibahi tahpas-wa dihas-ya=tehht
woman.TOP eat.IMPF red-GEN.SG meat-ACC.SG=NMNZ
"a/the woman who eats red meat"
(2)
a do-ht utumo pihhti-n tibahi-mme=tehht-en?
3INTR.TOP be.IMPF-Q place.TOP woman-NOM.SG eat.IMPF-LOC=NMNZ-NOM.SG
"where does the woman eat?" more lit. "the place where the woman eats, where is it?"
Subordinate clauses are also serviced by nominalized VPs, using the clitic =teht/deht. I'm a little unsure of how this should work. I know that some languages do use nominalized clauses where SAE langs tend to use infinitive constructions, but it's very foreign to me, and I'm not sure if I'm getting it right. Things like "Aya likes eating red meat" become the mouthful "qide tibahi tahpaswa dihasyateht Ayas munun", which translates literally to "Aya likes that she eats red meat", but then in order to distinguish between "likes to eat red meat (she does so habitually)" from "likes to eat red meat (a cherished activity, but she doesn't actually get to do it)" I have to add in the irrealis mood marker, which makes the whole thing even longer...
With the aim of shortening some use cases, I'm considering adding in a few words which expect a nominalized VP as a direct object, where case marking of the nominalized VP is absent. For instance, the verb noana "believe":
Code: Select all
(3a)
nis noanna Amelika dos sihti=teht
1DAT.SG believe.PF America be.PF great=NMNZ
(3b)
nis noanna Amelika dos sihti=teht-ya
1DAT.SG believe.PF America be.PF great=NMNZ-ACC
"I believed that America was great"
Code: Select all
(4a)
nis bizea deanas bizea nis bizea=teht
1DAT.SG know.IMPF 2DAT.SG know.IMPF 1DAT.SG know.IMPF=NMNZ
(4b)
nis bizea [deanas bizea [nis bizea]=teht]=deht
1DAT.SG know.IMPF 2DAT.SG know.IMPF 1DAT.SG know.IMPF=NMNZ=NMNZ
"I know (that) you know (that) I know"