Myself coming from Tolkienian linguistics I can't help but inspect amman iar
from this angle. Here are some observations:

Interestingly, lenition in amman iar is caused by _r_, especially by the
definite article _ir_. Some time ago, it was believed that _ir_ is the form
the article takes in Sindarin before vowels (or just before another _i_), as
_a/an_ in English, based on an untranslated text. However, lenition in
Sindarin is still caused by preceding vowels. _Ir_, which had been mistaken
for the article, might be the conjunction 'when'.
I also wonder why the gerund _er- tulo -e-l_ becomes _erthulel_. The
development _rt_ > _rth_ is what is found in Sindarin, but in amman iar I
would have expected its r-lenition to yield _*erdulel_.

Curiously, AI. _adhan_ 'man, human' seems to go through a second lenition
(S. _adan_, Q. _atan_ 'human'), although other Sindarin words are adapted
without it, e.g. _adar_ 'parent' (S. _adar_, Q. _atar_ 'father').

The author writes that the proto-language had pl. _-i_ which disappeared
causing i-mutation. Amman iar words do not just sound Tolkienesque, it quite
apparently borrows a lot of words from Sindarin and some from Quenya.
However, Sindarin by that time already had a developed system of vowel
mutations. Does it mean that amman iar borrowed pl. _-i_ from Quenya and
appended it to Sindarin and Quenya loans alike, later developing its own
vowel mutation? One notices that the effects of umlaut are actually somewhat
different from Sindarin.
A bit strange, however, is that words like _dorris_, _faindir_, _amil_ keep
the first vowel unaffected by the _i_ in the second syllable, but are
pluralized to _derris_, _fendir_, _emil_. I guess there must have been some
kind of analogical leveling.

The only grammatical features amman iar seems to borrow from Elvish are the
class plural _-(h)oth_ (identical in Sindarin), the mentioned pluraliztation
as well as maybe dative _-on_ (Quenya _-n_), genitive _-îon_ (Q. gen. pl.
_-ion_ after consonants and _e_).
Everything else, including pronouns, cases and the complex verb morphology
have nothing to do with Elvish. So I guess the idea was that Sindarin and
Quenya would form a superstratum for a different language when the
Numenoreans arrived?