Hello,

In an exhaustive study published in _Tengwestie_, the online journal of the
ELF (http://www.elvish.org/Tengwestie/articles/Hostetter/noldpat.phtml),
Carl F. Hostetter has set forth a carefully constructed classification
of Noldorin verbs in _The Etymologies_ with regard to how they form
their past tenses. This classification is in theory orthogonal to the
criterion proposed by Helge Fauskanger in his article about Sindarin of
the 'Ardalambion' website (http://www.uib.no/People/hnohf/sindarin.htm),
since Carl's classification focuses on whether the stem of a verb is
modified in forming the past tense (thus 'strong' vs. 'weak'), and Helge
concentrates on the form of the stem (thus 'basic' vs. 'derived'), and
on how they form the present tense (hence the equivalence of the terms
'basic stems' and 'i-stems' in Helge's article).

In constructing a full classification of the Noldorin/Sindarin verbs,
both of these criteria will have to be taken into account, since the
divisions they produce are clearly not isomorphic (cf. _degant_, a
'weak' pa.t. of _degi_ (NDAK), apparently a 'basic' verb; all
references are to stems in _The Etymologies_). Nevertheless, this post
is not intended to construct such a classification. I would instead
like to draw attention to some parallels between the Noldorin system
and the verb system of the Germanic languages, including Old English.

The principal distinction in the Germanic verb is that of strong vs.
weak - whether the different verbal stems (usually four) are formed by
affixes ('weak' verbs) or modifying the stem itself ('strong verbs').
The change of root vowels, or Ablaut, is the principal characteristic
of these verbs (the distinction is not limited to Germanic, of course,
cf. the five verbal stems of Old Irish and its strong verbs). Now
Noldorin does not seem to exhibit the kind of highly developed Ablaut
patterns seen in older Indo-European languages, yet Carl's Strong Past
I class can be said to indeed show Ablaut. What is commonly called
'a-infixion' (clearly seen in pa. t. _daul_ from DUL-, and perhaps in
_haul_ (KHAL-) and _trenor_ (NAR-)) can be seen as _a_-nil ablaut,
commonly seen in Germanic verbs. In fact, most Germanic strong verbs
(of classes I to V) derive from classes with the Indo-European
_e_-_o_-nil ablaut, which yields Proto-Germanic _i_-_a_-nil in various
guises. The _a_-nil alternation can be seen in such forms as Old Norse
_kaus_ (pret. sg.) - _kurum_ (pret. 1pl.) from _kjósa_ 'to choose'
(class II) (the _s/r_ alternation is due to Verner's Law), or Gothic
_rais_ - _risum_ from _reisan_ 'to rise' (class I). The difference
between Germanic and Noldorin (aside from the obvious ones, such as
that Germanic strong verbs have four stems, and the Noldorin ones two)
is that while the null grade is exhibited in the preterite plural in
Germanic, it is manifested in the present in Noldorin.

That was qualitative Ablaut. Quantitative Ablaut (well, its relics) is
exhibited in Germanic by class VI strong verbs, such as Old English
_faran_ - _fór_ - _fórum_ - _faren_. Since Germanic *_a_ partly goes
back to IE short _o_, the quantitative pattern here is easily
discernible, even though Germanic has transformed the purely
quantitative alternation into a quantitative *and* qualitative one. A
clear (well, almost) case of quantitative Ablaut can be seen in *_nîdê_
pa. t. of 'hurt, bruised' from NID-. Another solution here can be
postulating 'i-infixion' (thus transforming this into qualitative
_i/nil_ Ablaut, also with a Germanic parallel - vide the _reisan_
example above), but it seems to be _ad hoc_ (yet another one, Carl
notes, is supposing it is an adjective and not a verb at all). If
_haul_ and _trenor_ are also to be derived by lengthening of the root
vowel rather than a-infixion, then they are also cases of this kind of
Ablaut (and we can as well see the transformed pattern where both
quality and quantity play a role).

Yet this all seems to be a rather peripheral issue. I would like to
consider in more detail the behaviour of those verbs that have the
vowel _i_ following the root syllable, which thus causes i-affection,
or i-umlaut, in the root-syllable. In Noldorin, there are two classes
of such verbs - what Helge calls 'basic' verbs, i. e. those which
represent the bare stem without any suffixes, and the 'derived' verbs
in *_-jâ_.

I suggest that the first class corresponds largely to the Germanic
thematic verbs. The thematic conjugation means that the IE thematic
vowel *e/o (Germanic *i/a) is inserted between the stem per se and the
rest of the endings. In Germanic, it was the strong verbs and the weak
verb of class I that belonged to this conjugation (in Gothic also weak
class IV, but since Gothic does not show umlaut, it is irrelevant, and
the rest of the Germanic languages have not kept class IV). Thus, the
strong verbs and the class I weak verbs exhibit i-umlaut.

Strong verbs show thus i-umlaut in the 2 and 3 singular present
(provided the vowel is mutatable, of course), a pattern seen in Old
English _helpan, ic helpe, he hilpþ_, or in Modern German _helfen, ich
helfe, er hilft_. In the Noldorin case, the i-mutation would seem to be
carried almost through the whole paradigm, with the exception of the
endingless 3sg forms (we don't seem to know much about 2 and 3pl at
this point). Now it must be noted that many basic verbs are also strong
verbs. In fact, with the possible exception of Carl's Strong B class
(where, incidentally, the root vowel is the immutatable _i_), all
Noldorin strong verbs belong (at least presumably, with several
exceptions where the class cannot be determined with certainty,
such as _sogo_, pa.t. _sunc_) to the i-stem class and thus show
mutation where applicable (of course, neither do all Germanic strong
verb show the umlaut, cf. Old English _singan, he singþ_). The
important difference is that while Germanic strong verbs show additional
differences between the present and preterite vowels due to Ablaut, the
Noldorin ones do not, yet if we ignore the particulars and concentrate
on the strong/weak distinction per se only, the similarity is there.

Another point where the Germanic verbs show an intervening _i_ is class
I weak verbs. Some languages, such as Old English, usually keep the
umlaut intact all over the paradigm, yet not always - cf. Old English
_þencan_, pa. t. _þóhte_, _cwellan_ 'kill', pa. t. _cwealde_. Old
Norse (and to varying extents modern Scandinavian languages) has
preserved the distinction between the present and preterite stems much
better thanks to the fact -j- was lost everywhere except before _a_,
_u_ (so the -j- is more apparent), cf. Old Norse _þysja_, pa.t.
_þusta_ 'make a noise', Modern Swedish _dölja_, pa.t. _dolde_ 'to
hide' (incidentally, this verb looks tantalizingly similar to N
_d[oe]lio_, _delio_ of similar meaning - and even the presence/absence
of umlaut is distributed identically). ([oe] =oe-ligature, not the
diphthong, vide VT45:11 s. v. DUL-. Noldorin verbs of similar behaviour
(i. e. weak verbs with (at least presumably) i-umlaut in the present
tense) include _hadhant_ (KHAD-), _hafant_ (KHAM-).

So there remain weak verbs without alternation between mutated and
non-mutated forms for various reasons. These include weak ya-stems
(where the -ya- is present throughout the paradigm, and thus so is the
mutation), which can be compared with the Old English weak class I
verbs with umlaut all over the paradigm, and other derived weak verbs
(which never show i-suffixion, thus cf. Germanic weak class II with the
suffix -a-). For some verbs the exact attribution is quite impossible,
and probably irrelevant, since the stem vowel is immutatable. Some
verbs undergo transition from the unproductive strong class into the
productive weak one, characteristically showing the mutated present
tense vowel instead of the historically correct one (_degant_ rather
than **_dagant_).

In this message I have tried to show how Tolkien constructed his
art-language, Noldorin (Sindarin) on principles similar to older
Indo-European languages, in this case Germanic. The use of similar
general points (distinctions of strong/weak, presence/absence of _i_ in
the present tense stem) account for the fact that the Noldorin verb
classes can correspond to the Germanic ones almost point for point.
This short essay (rather preliminary, and open to criticism!) has
shown, I hope, the fact the Germanic (and not only Gothic) influence on
Tolkien's art-languages must never be underestimated, as it often is.
Noldorin is often assumed to be based largely on a Celtic language,
Welsh, yet the Welsh verbal system does not show such a pervading
similarity to the Noldorin one. Even though some aspects indeed can
provide points of similarity (such as the i-affection seen in the Welsh
2pl and 3sg present), there are many more points of departure (in
particular, Welsh has all but lost the strong/weak distinction). Old
Irish presents a more complex system, yet it is well known not to be a
major influence on Tolkien, and anyway this remains a subject for
further study.

In any case, comments are most welcome!

Pavel
--
Pavel Iosad pavel_iosad@...

Nid byd, byd heb wybodaeth
--Welsh saying