(Introductory question: Are these reasonably plausible?)

Okay.  This shouldn't be too nightmarish.  I figgered I should set up some
sound change rules that I can derive Nrit vocab from.  Joy!  None of that
obvious combinatoric compounding-and-slavishly-following-rules stuff

i y /i-/ u
e o

/e/, /o/, and /a/ have phonemic length, which we'll mark the Finnish way.
We'll pretend that vowels follow the same laxing rule as Nrit: Vowels other
than /o/ and /a/ are laxed when short, unstressed, and non-final, even
though that's unimportant, so I won't bother marking it anywhere.

Generally, stops distinguish by aspiration and fricatives by voice.

Labials: p, v, m
Dentals: d /t[/, dh /t[^h/ þ, ð, nh
Alveolars: t, th, s, z, n
Palatals: c, ch /C/, j, nj
Velars: k, kh, h, ng
Labiovelars: kw, khw, hw, ngw /Ng^w, N^w/ (trigraphs! *hug*)
Liquids: l, lh /l^o/ , ln /l^n/, lj /L/, r, rh /r./

Some sort of simple syllable structure:
final: (X)V(C)
I'm tending to a (S)(T)V(N)(L) pattern.

Soo, we'll make up some words, and try to devise soundchange rules that make
them sound like Nrit:

aalhkyr    vija    khweng    sthoon    yrkhe    zukhunh    ljanj    caalid
earte    lnach

(Key to symbols:    C: any consonant    T: a stop    S:a fricative    N:a
nasal    L:a liquid    V:a vowel    X:any consonants    U:a back vowel
I:a front vowel    #:word boundary    {}: Nrit orthography)

Phase 1, vowels shift:

yXU > uXU
yXI > iXI
y(X)# > i(X)#

Va > V:
iu > i:
io > ja
ia > ja
ie > i:
eu > u:
eo > o:
ea > ja
ei > e:

Phase 2, palatals are lost:
c > /Z/ when intervocalic
c > /Z/ when adjacent to a voiced consonant or aspirate
c > /S/ elsewhere
ch > h except where intervocalic, where ch > <hh> (<hh> is the mysterious
Nrit voiced fricative thingy.  It might be hooktop-h, or gamma or
nj > {ni} (This varies between /ni/ and /nj/ depending on the environs.)

Phase 3, dentals go rotten:

d > {tth}
dh > {ddh} (Affricates)
þ > {th}
ð > {dh}
nh: Nasalizes the adjacent vowel and is lost, except finally, where it

Phase 4, velars go strange:
ng > n.
kh > /tS/ except intervocalically, where it is unchanged.
kV: > khV
intervocalically, h voices.  high vowels drop it completely.
w > {u} (which like i can be a vowel or glide)

Phase 5, liquids become Nritlike:
lj is no longer regarded as a phoneme.  rh and r conflate.
lh >h
ln nasalizes the adjacent vowel and loses its nasality.

Phase 6: Ephenthesize or reorder the clusters Nrit doesn't like.

The >> denotes a respelling into Nrit orthography.
aalhkyr > aalhkhir > aakhir >> âqir
vija >> víia
khweng > /tSwen.g/ >> cueng ([n.g] and [N] are allophonic, phonemically the
sthoon > stoon > tsoon >> tsôn
yrkhe > irkhe > irhe > ire >> irre
zukhunh > zukhu^nn >> zuqùn
ljanj > ljani >> liani
caalid > /Sa:lid/ > /Sa:lit[T/ > /Sa:lit[Ta/ >> xâlittha
earte > jarte >> iarte
lnach > lnah > la^nh >> l h

Hmm.. more Nrit-looking than I had guessed, but rather too systematic and
regular.   Hmm... some consonants that care about adjacent consonants would
be interesting.