Regarding Helios De Rosario Martínez's article, "Early Ilkorin Phonology"
(<http://www.elvish.org/Tengwestie/articles/DeRosarioMartinez/earlyilkorin.phtml\
>):

Being involved in the topic myself I have read the article with great
interest and have some comments to make. The special characters are
changed in the following post (macron to circumflex etc.) so that it
will hopefully be readable.

>Long vowels in prehistoric forms are normally marked in ND with a
>macron (¯), but in the NW instances Tolkien also uses an acute
>accent (´) or a colon (:) for long vowels, except in t'lê'pe,
>where the acute accent means that long ê is stressed (cf. PE13:135).
[...]
>mîye

There is no mention that the acute accent often represents stress (as
is also written in PE13:135), even if not grouped with the macron or
colon. And stress is very important in understanding the derived forms
in the case of _y_. The data is explained remarkably well if one
assumes different developments depending on whether _y_ comes before
or after the stress. It can be summarized as following:

_y_ before accent | _y_ after accent
Quenya i|y | y
Telerin i | r
Noldorin i|j | dh
Ilkorin ? | gg

I cite relevant examples in parts III.3.2. and IV.1.2. of my article.
One can compare the Noldorin changes with an almost identical
stress-dependent development in Welsh as mentioned in Morris-Jones'
_Welsh Grammar_. Therefore I believe that _míye_ has the stress, not
the length marked by the accent with the development indicated in the
second column: Q. _mie_ < _*miye_, T. _mire_, N. _midh_, Ilk. _migg_.

>Under this model, the development of preh. pisye would be explained
>as *pisje > *pisðe, which by assimilation > *pisze or *pisse (or one
>after the other), and eventually > fiss through regular loss of
>final vowel and Ilkorin Sound Shift.

Following the previous idea: There is no stress marked in _pisye_, but
looking at the resulting forms, Q. _pihye_, T. _pirie_, N. _hí_ it
seems that it should be _*pisyé_, otherwise one would expect _dh_ in
Noldorin (I cannot say what _*sdh_ would have evolved into).

Therefore I believe that _pisye_ shows a development different from
_míye_. Looking at the first column of the table above it appears that
_y_ is kept in all the languages (but may become vocalic depending on
its environment). Hence, I would expect that _sy_ was inherited by
Ilkorin and later changed to _ss_ by a different mechanism.

With the little amount of examples availible there is of course a
chance that the change of _y_ is not dependent on stress in Ilkorin,
but the bottom line is: Assuming _migg_ < _*miðe_ does not require the
assumption of _*pisðe_.

>the change from dentals to velars is not unknown in Elvish languages
> — one example occurs in Gn. feigien worse < faiðn (PE13:114)

I'm not convinced here. Both forms are written beside each other, but
can one really be sure that _feigien_ is derived from _faiðn_ and not
vice versa? The adjectival ending is _-(k)ka_ (_fêka_, _fekka_), so
maybe _*feiX'n/*feigh'n_ > _faiðn_?

Also, one has to assume a comparative ending involving _d_ which may
be similar to N. _-iad_ < _*jat-_ (PE13:125), but looking at the
counterpart _mawr_ 'good', adj. _maien_, _mairien_ 'better' one rather
sees _-ien_ once appended to a form combined with the adjectival
ending _-r_ and once not. So maybe we are even dealing with
_*feik-ien_ > _faigien_ (with suffix _-k(a)_) and _*fei-j'n_ > _faiðn_
(without suffix and _y_ > _dh_)?

I find it difficult to believe that the interdental spirant _ð_,
becoming intervocalic, suddenly changes to a velar stop, although
intervocalic _ð_ is perfectly acceptable in Goldogrin/Noldorin phonology.

>And in Old Norse (which is not a West Germanic language, however),
>sometimes /j/ was lengthened to /jj/ according to Holtzmann's Law
>(Prokosch, 1939: §33c), and eventually hardened to /gg/, as here
>in mîye > migg.

This is very interesting and seems to be a likely development for
Ilkorin, given that some fairly complex sound changes of Welsh are
closely imitated in Noldorin of that time.

Being self-educated in linguistics I assumed that /gg/ from a single
/j/ would be unlikely and therefore tentatively suggested /ž/ or /dZ/
for Ilk. _gg_ by pure speculation (cf. the pronounciation of initial
_j-_ in English and French).

>Roman Rausch interprets in his article that Ilk. þerr comes from
>tésare, the primitive form of the N cognate teiar, instead of
>from terar (whence Q. telar). In fact all these words are given
>in the same entry of NWL, and it is likely that terar < tésare.
>But preh. terar seems to be closer to the Ilkorin form, both by
>phonology and by their relative position in the entry of NWL.

As I interpet the entry, _tésare_ is the primitive form of all the
three words mentioned. Noldorin vocalizes _s_ > _i_ (or > _j_ as part
of a diphthong), the Qenya form _telar_ looks strange at the first
glance, so Tolkien explains that it is from _terar_, i.e. with
rhotacism and vowel loss followed by dissimilation. That would make
_terar_ an old Qenya form.

If rhotacism occurs in Ilkorin as well (it's Germanic in style after
all), Ilk. _þerr_ might have evolved from a parallel _terar_ as well;
otherwise I would assume _*þesr_ > _þerr_. In any case I believe that
_tésare_ is the older form.

>Other vowel mutations [...]
>mîgg < smeigê

typo: _smíg_

Of course, I will link 'Early Ilkorin Phonology' in my work and add
some references to it in the text.

Roman Rausch