575Pronunciation and writing of _r_ in Quenya
- Jan 4, 2004Thank you all (specially to Ales) for your comments on the matter of
rhotacism. I find them very instructive and helpful.
Some time ago I started to write an article for the Spanish linguistic site
Lambenor (http://lambenor.free.fr/) on the matter of the _r_ in Quenya,
with the ideas provided by the members of the Lambenor group (those
who know Spanish may read the messages 384-395, 3287-3290,
4999-5003, 5005, 5009, 5010, 5015, 5026, 5031, 5036, 5038, 5040,
5052, 5123-5124, 5127-5128 and 5765-5768 in
The issue of rhotacism is one of the most complex items of the article,
that was mainly discussed by me and Lambenor member Javier Lorenzo.
But, as may be expected, the discussion dealt mainly with the issue of
pronunciation and spelling in _tengwar_.
I am aware that in other lists (I suppose TolkLang and Elfling, the
most known) this must have been already discussed; but I have searched
in their archives (through the mirror in TolkLang), and never found a
comprehensive and coherent statement on the matter. We are trying to
do it in Lambenor, but since there it shall be only available for
Spanish-speaking people, I translate the conclusions and post them
here, for you to evaluate them.
I skip the discussion on the variants _rd_, _ry_, _hr_ and _rr_, and
go directly to the controversial item: the duality of the pronunciation and
spelling of the single _r_.
We found the main information in the Appendix E:
"R represents a trilled _r_ in all positions; the sound was not lost
before consonants (as in English _part_)." (LR:1088)
"[Óre] was often used for a weak (untrilled) _r_, originally occurring
in Quenya and regarded in the system of that language as the weakest
consonant of the _tincotéma_." (LR:1094)
"[Rómen] (in origin a modification of [óre]) was used for 'full'
trilled _r_" (LR:1095)
From those citations we have seemingly contradictory information,
since first one says that _r_ is always trilled and second that there
was an untrilled _r_. But it is clear that the key is on the words
"originally", "in origin". The second and third citations are from the
section about writing, that deals with the history of the _tengwar_, the
phonemes they _originally_ represented, and their changes, while the first
citation deals with how they must be pronounced Quenya and other
languages in the context of LR, that is, in the late Third Age.
So, we may infer that when Fëanor invented the _tengwar_ there were
a weak untrilled and a full trilled _r_, clearly different, but in
later stages there was only one trilled phoneme.
The phonemes we are dealing with must be the trilled, tap and
approximant alveolars. If we take the literal meaning of the
citations, we learn that the phoneme the of _r_ in Third Age was the
alveolar trilled, as well as the original sound of _rómen_; and that
the original sound of _óre_ should be tap or approximant. We found the
approximant a likelier option, since the other _tengwar_ of the sixth
grade also represented Quenya approximants (_anna_ == /j/, and _wilya_
But others stated that the citations must be interpreted otherwise,
since they are oriented to a general (though interested) English
public, not to phoneticians, and that the terminology used in the
Appendixes does not accurately match the phonetic terminology we
are used to, say, the one of the IPA.
According to this idea the weak untrilled _r_ must undoubtedly be the
weakest for English, the approximant. And we may also suppose that the
trilled _r_ of _rómen_ and in the Third Age is rather the tap, more
usual for English-speakers (in opposition to the approximant, stated
to be _un-trilled_); however we cannot absolutely discard that it was
the phonetically trilled _r_, although this is more difficult for
English-speakers to identify.
A more elaborate version of this interpretation is that the
emphasized adjective "'full' trilled" means something different from
just "trilled", similar to the distinction of the Spanish "vibrante".
We Spanish do not have any approximant _r_, but the tap and the
trilled, which are called "vibrante simple" and "vibrante múltiple".
This idea means that:
- "'full' trilled" is like "vibrante múltiple" ("trilled" for IPA).
- "trilled" is like "vibrante" ("trilled" or "tap" for IPA).
So original _rómen_ would be the trilled (according to IPA) and _óre_
approximant, but the _r_ in the Third Age could be either trilled or
The weak point of this interpretation is that (according to a native
English speaker in Lambenor), the parallelism between the English term
"trilled" and Spanish "vibrante" is quite odd for most languages, except
for Spanish, and that such an interpretation may be regarded as a
typical interference of Spanish speakers, having nothing to do with
So, the first interpretation (original _óre_ approximant, original
_rómen_ and Third Age _r_ trilled) seems to be likelier. And I think
that it can be supported by the nature of Finnish _r_. As far as I
know (a Finnish member will please tell if I am mistaken) in that language
there is only "a trilled _r_ in all positions" (paraphrasing Tolkien in LR:1088),
the trilled alveolar according to IPA. Since Quenya is inspired by Finnish,
it is a good hint.
However, this does not mean that there was only one _sound_ for _r_.
Although the only Finnish phoneme for _r_ is the trilled, I believe
that there are distinct allophones for it. I may be wrong, since I
have no knowledge of Finnish, but listening to the samples of Finnish
names in http://www.saunalahti.fi/~kajun/finns/alpha.htm I hear that
before vowels the _r_ is really trilled (like Spanish _perro_), but
before consonants and in final position it is tap (like Spanish
_pero_). Am I right?
Moreover, this duality of allophones may be compared with the duality
in the spelling. We have few samples of tengwar in Quenya: following
the notation of Mellonath Daeron
(http://www.forodrim.org/daeron/mdtci.html) they are only DTS
12, 19, 20, 38, 40, 42, 46, 54, 55 and 59. The longest and most important
is the tengwar version of the _Namárië_ in _The Road Goes Ever On_, where
we see that _rómen_ goes before vowels and _óre_ before consonant and
final, the same pattern of the Finnish allophones previously commented.
Although the other samples of _tengwar_ are too short for any conclusion to
be found, that pattern seems to be kept, with the only exception the
illustration 182 of _Artist and Illustrator_, where the word _moruvan_ is
written with _óre_ (althoug the _r_ is before a vowel).
The dual writing of _r_ does not seem to be of etymological origin
(in contrast with the other dual-form writings, as the one of _thúle_ and
_silme_ deduced from the statement in LR:1088, "[_th_] had become _s_ in
spoken Quenya, though still written with a different letter"). In the
text of _Namárie_ we have the words _oromardi_ ("lofty halls") and
_ortane_ ("[she] raised"), both with the evident particle _or(o)-_
related to "raise", but in one case spelt with _rómen_ and in the
other with _óre_; as well as the plural suffix _-r_ is always spelt
with _óre_ but the genitive form _-ron_ (as in _aldaron_) with
So it is possible that the pattern is of phonetic origin, and each
_tengwa_ matches one of the allophones of _r_. However others have
discussed other possible reasons for that spelling. Some said that it
is just an stylistic issue (as the duality between _ss_ and _ß_ in
German). Some highlighted David Salo's hypothesis in
that such a pattern helps to indicate whether there is or not an _a_ after
_r_ in the style of writing which omits that vowel. However, there is
little evidence for any such hypothesis.
And that it all the information (in summary) that we gathered in
Lambenor about the issue. What do you think?
- Next post in topic >>