In Etym. we can find two Noldorin etymologically related words
glossed 'light': _calad_ (s.v. KAL-) and _galad_ (s.v. GAL-, in
A&C:13).

_Calad_ does not occur elsewhere in Etym., while _galad_ is apparently
found in two compounds: _gilgalad_ 'star-light' (GIL-) and _Cilgalad_,
the Noldorin cognate of Qenya _Kalakilya_, 'Pass of Light' (KIL-).

(It also occurs in V:362, in the name _Gilgalad_, as a cross-reference
to _calad_. But this must not be accounted, since Hostetter and Wynne
point out in A&C:19 that the cross-reference is actually to _Glingal_.)

And now I wonder: Is really _galad_ in _gilgalad_ and _Cilgalad_, or
is it the lenited form of _calad_?

===

For _gilgalad_ this last theory is usually held, based in the
explanation given by Tolkien to Rhona Beare about the name of the
character _Gil-galad_ (L:279), in which he wrote that it was formed
from _gil_ + lenited _calad_ (with the same meanings that these words
have in Etym.) However, that letter was written in 1958, and I would
like to know whether this is true according to the inner principles of
Etym., composed long before.

A first overview would suggest that in fact _gilgalad_ should not be
the result of any lenition, by comparison with the names _Gilbrennil_
*'Star-lady' and _Gilthoniel_ 'Star-maker' that occur next to it in
GIL-. _Gilbrennil_ shows no kind of mutation (cp. _brennil_ 'lady',
s.v. BARÁN-), and the mutation of _Gilthoniel_ (perhaps < *_tânielle_
or something similar, cp. _tanô_ s.v. TAN-) is doubtless the liquid
assimilation _l_ + _t_ > _lth_, not a lenition that would have yielded
*_Gildoniel_ instead. (cf. "Consonant mutations in conceptual
evolution of Noldorin/Sindarin phonology", by Ryszard Derdzinski, in
http://www.elvish.org/gwaith/pdf/consonant_mutations.pdf).

However the proper names _Gilbrennil_ and _Gilthoniel_ may be a
misleading clue. It is not difficult to see that the Noldorin
consonant mutation pattern in Etym. is quite complex, depending on
the phonology and the grammatical context of each case (and even on
the stage of composition of the manuscript, which was substantially
revised). Therefore, comparisons must be done with care in order to
not mistake the conclusions.

And though those two names are phonologically similar to _gilgalad_,
they have some grammatical differences. In all three, _gil_- 'stars'
is the "modifier" noun, that is in proclitic position (at the
beginning), while _galad_, _brennil_ and *_toniel_ are in each case
the "principal" or "modified" noun. However, the function of _gil_- is
different in each compound:

- In _gilgalad_ 'starlight', _gil_- is the source or subject of the
action expressed by _galad_. It could be defined as a "subjective
genitive".
- In _Gilbrennil_ *'Star-lady', _gil_- characterizes _brennil_. It
would be a "genitive of description".
- In _Gilthoniel_ 'Star-maker', _gil_- is the object, the product of
the *_toniel_. It could be perhaps classified as an "objective
genitive".

(Note: I have chosen these types of genitive from the "Resumen de la
gramática latina" -- trans. "Summary of Latin grammar" of the
"Diccionario Vox latín-español/español-latín" edited by Bibliograf,
used by Spanish school students of Latin, not from any book that
covers this matter in detail. Surely there are other more accurate
descriptions, and I would be grateful if anyone provides them.)

Apart from _gilgalad_, the other cases that I have found in Etym. of
compounds that have a proclitic modifier, with function of subjective
genitive, are _lhasbelin_ *'leaf-fading', 'autumn' (cf. LAS(1)-,
KWEL-) and _gildin_ 'silver-spark' (cf. TIN-). In both the principal
name shows clearly lenition: _lhasbelin_ must certainly come from
*_lass-pelin_ < *_lassekwelênê_ (this latter form attested in Etym.);
and _gildin_ must be _gil_- + _tin(t)_ 'spark'. From the comparison
with these one may infer that _gilgalad_ probably comes from _gil_- +
_calad_, as Tolkien told Rhona Beare twenty years later.

===

The case of _Cilgalad_ 'Pass of Light' is somewhat easier. Opposite to
the previous cases, in _Cilgalad_ the modifier noun is _galad_ in
enclitic position (at the end). This is something normal in Noldorin,
but less frequent, at least in the examples of Etym. Of all the cases
that I have found there (in V as well as in A&C), if we only take
those in which the enclitic modifier is a noun (not an adjective) and
begins with a consonant that might be mutated (stops or _h_), we are
left with, besides _Cilgalad_:

- _Dagor Vregedúr_ 'Battle of Sudden Fire' (BERÉK-, MERÉK-, UR-).
- _Din-Garadras_ >> _Din-Caradras_ *'Pass of Caradras' (DEN-).
- _Amon Uilos_ 'Mount Everlasting-snow' (EY-, GEY-, OY-).
- _Eredwethion_ *'Mountains of Shade' (ÓROT-, WATH-).
- (Fingolfin) _Aran Chithlum_ 'King of Hithlum' (TÂ-).

We can see that all of them are mutated according to the Noldorin
lenition pattern: _bregedúr_ > _vregedúr_; _guilos_ >_uilos_ [Note 1];
_gwethion_ > _wethion_, and _hithlum_ > _chithlum_ [Note 2].

[Note 1:-- This case depends on which interpretation of Etym. is chosen:
As explained under OY-, this base replaced GEY-, and this itself replaced
EY-. Only in GEY- the unmutated form is _guilos_; in the other it is
originally _uilos_ and hence it does not undergo any lenition.]

[Note 2:-- I am not really sure that this is a case of lenition. Among
Tolkien's published explanations about the consonant mutation in the
Celtic-like Elvish language, the nearest in time to Etym. is in the
"Early Noldorin Grammar". There, "_hîr_ 'lord', (a) _ihir_" shows no
soft mutation of _h_ (PE13:121).]

I suggest that _(Aran) Chithlum_ might nevertheless be an instance of
lenition, because in the earlier "Gnomish Grammar" _h_ > _ch_ does
occur in the list of the "grammatical mutations" (PE11:7), and it may
also be compared with the later famous title of the _Narn i Chîn
Húrin_, where _hîn_ 'children' > _chîn_. Christopher Tolkien wrote in
LR:322 that both _i Chîn Húrin_ and _Aran Chithlum_ are, among others,
cases of the "Initial Variation of Consonants" of Exilic Noldorin,
what from the explanations in V:298 and V:301 we might infer that is
specifically the lenition or soft mutation.

Another interpretation is given by Helge Fauskanger in "Sindarin - The
Noble Tongue",
http://www.uib.no/People/hnohf/sindarin.htm#nasal,
where he asserts that this _h_ > _ch_ is rather nasal mutation. This
is not what Christopher Tolkien seemed to mean in the notes about the
"Initial Variation of Consonants" (if I don't misinterpret his words);
and the case of _Aran Chithlum_ does not fit that theory either,
since there is no article _in_, or preposition _an_ or _dan_ in it. A
nasal mutation could be, however, justified by the final _n_ in
_Aran_.

We see in the change from _Din-Garadras_ to _Din-Caradras_ that the
change in the rules of lenition pointed out by Christopher Tolkien in
V:298 also affected Etym. But yet in the last pencil layer of Etym. Tolkien
wrote _Din-garadras_ (A&C:9), and the "D"-section should be the last part
that Tolkien revised of Etym. (cf. V:353). So, we may infer with
confidence that when he wrote _Cilgalad_ under the base KIL-, the rule
should be yet that lenition _did_ occur, hence such a name would be
formed from _calad_. The fact that the Qenya cognate is _Kalakilya_
supports this idea.

===

So, that is why I think that both _gilgalad_ and _Cilgalad_ in Etym.
do not actually show _galad_, but rather lenited _calad_. However, I
do not know whether the basis of these ideas is correct, especially
that of the grammatical distinction of genitives and its possible
effect on consonant mutation. Moreover, there are still some obscure
points to me, as why _Certhan_ 'ship-builder' (s.v. TAN-), which shows
the same pattern as _Gilthoniel_, _Gilthonieth_ next to it, was
changed to lenited _C(e)irdan_ (but _Gilthoniel_, _Gilthonieth_ were
not). I would like to read your interpretations.

Regards,
Helios