Then I noticed a similarity in the Proto-O and Pre-Tim Ar plurals. In Proto-O, the plural prefix was *r (to be more specific, it was */ʁ/), which had a syllabic allophone [a] in certain circumstances. In Pre-Tim Ar, there were two plurals: A prevocalic vowel, which was kind of indeterminate but I usually conceived of as *ə, that caused metathesis of *əCV- to *əVC- and then dropped, and the plural particle ar which follows vowel-initial words.
I further noticed something else. In Tim Ar, gerunds and verbs appear the same. In O, agent nouns in the accusative are identical with the verb stems they are derived from.
The above leads me to think there's hope for unifying these languages in the same macrofamily. There's a few hurdles to overcome, however.
Phonologically speaking, here's Proto-O:
*p *t *k (with fricative allophones)
*n (with syllabic and assimilatory allophones)
*w *r (= */ʁ/) *j *h (with vocalic allophones)
For comparison, here's what O developed into:
/p t k kʷ q qʷ/
/f β~w s z x xʷ χ χʷ/
/a aː ɛ ɛː e eː i iː ə əː ɨ ɨː o oː u uː/
/ae̯ ei̯ ao̯ ou̯/
Now, here's Ngade n Tim Ar:
/m n ŋ/
/s S ɬ x h/
/ɹ l ʕ/
/à á è é ø̀ ǿ ì í ỳ ý ɤ̀ ɤ́ ò ó ɯ̀ ɯ́ ù ú/
/m̩ n̩ ŋ̍/
As regards the Tim Ar phonology:
- * The front-rounded and back-unrounded vowels are the result of vowel mergers, keeping the tone of the first vowel but the POA and roundedness of the second. These vowels "decompose" when a noun is pluralized; e.g., Cô- becomes úCe-.
* /S/ is highly dialect-dependent; in the Imperial dialect it merged with /s/ but it comes from an original *θ, which spazzed out in the various other dialects.
* /h/ is from an earlier *p, which lenited to *f and then debuccalized.
* As for the tone…I have an idea of having the proto-protolang having either an original voicing distinction in the consonants or phonation in the vowels, and having that be reflected in Tim Ar and dropped in O as the result of a merger.
Tim Ar also is heavily marked for deixis, but I might just handwave that as extensive contact with a substrate language.
I'd also have to deal with the Proto-O *e-infix, which would have been absent in Pre-Tim Ar.