In the second article recently published on Tengwestië, Carl Hostetter gives reasons for seeing two distinct values assigned to the grapheme LL in theMessage 1 of 5 , Dec 9 4:33 AMView SourceIn the second article recently published on Tengwestië, Carl Hostetter gives
reasons for seeing two distinct values assigned to the grapheme LL in the
Sindarin of _The Lord of the Rings_: geminate l, phonetically [l:], and long
voiceless l, phonetically [L:]. The latter would be heard in _mallorn_ (cf.
I am however rather unconvinced, and still think that externally speaking,
the value of [L:] for <ll> is a late development. I doubt very much that Tolkien
already saw things likewise when he wrote LR and its Appendices - without
absolutely convincing arguments, but with some serious presumptions.
[I don't think it takes much presumption to accept Tolkien's own account
of the orthographic choice he made at face value, unless there are very
strong reasons not to. CFH]
Firstly, it is striking that no reference to a double pronunciation of ll is
given in LR's Appendix E, or to a special pronunciation of the (important)
word _mallorn_ in particular. True, absence of evidence is not evidence of
absence, so this does not allow certitude, yet it is troubling. Although the data in
Appendix E are quite compressed, they contain minute details (for instance
the slight palatalization of L in some contexts, or the closer or more open
nature of long E and O's in Quenya and Sindarin respectively) and mention
exceptional cases (such as the stress of _Annûn_ and _Amrûn_ in Sindarin).
[I don't find this striking or troubling at all. Tolkien likewise did not note in
App. E that /d/ in certain phonetic environments in Sindarin words represents
_dh_; yet we know that it does, and that Tolkien chose to use /d/ nonetheless
because he found /dh/ uncouth for English readers. This precisely parallels the
decision to use /ll/ instead of /lh/ or /lth/, and could be explained by
precisely the same reasoning. Yet I don't recall anyone finding the use of /d/
for /dh/ to be troubling, or the absence of an account of this in App. E
Secondly, it is true that other graphemes have double values in Elvish
languages, such as H in Quenya or F in Sindarin, but those cases are explained in
Appendix E. Why would he not have done the same for LL if needed? I also doubt
that it was a way to avoid an "uncouth" spelling, since Tolkien used LH from
the very beginning, which is alien to English as well. Using the ambiguous
spelling LL while LH was at hand, and already seen elsewhere, would be quite
unfriendly to the reader, and slightly odd anyway. The analogy with the use of D for
DH is interesting... but notice that this was revised later.
[Tolkien's explanation of finding /dh/ uncouth was specifically applied to it
use in _The Lord of the Rings_, and for the lay, English audience that would
form its readership. But Tolkien used both /dh/ and edh in his _linguistic_
writing from the very beginning. So his use of /lh/ and /lth/ in his linguistic
writing is no evidence against what Tolkien says about /ll/ and its use in
Thirdly, I do not think that the comparison with English spelling is
relevant. It is well known how complex and sometimes inconsistent English orthography
is, whereas Sindarin's romanized orthography, if admittedly not an entirely
phonological representation, comes quite close to it. The example mentioned in
the article ("It is rather as though it were regarded as contradictory to say
that English S is pronounced /s/, but then to note that the plural marker -es
is pronounced /?z/.") is plainly an instance where English favours
morphological unity at the expense of phonetic accuracy. Moreover, the orthographic issue
is not exactly the same in a living and a constructed language. For a living
language, it is possible to depart from phonetic accuracy for other purposes
(notably morphological unity or etymological preoccupations) without too much
trouble: the competence of users will supply. For a constructed language
presented to others, it cannot be counted on, and one first needs to give a clear
picture of pronunciation, as Tolkien claimed to try doing in LR's Appendix E: "In
transcribing the ancient scripts I have tried to represent the original
sounds (so far as they can be determined) with fair accuracy, and as the same to
produce words and names that do not look uncouth in modern letters".
[Tolkien found the comparison with English spelling relevant, so I don't see
how your personal disagreement with his judgment can count as a relevant
argument against the veracity of Tolkien's account. You don't have to agree
with his judgment, but you do have to accept that it was, as he claims, the
basis upon which he made his decision. As for the analogy I drew with
pronunciation of _s_ vs. _-es_, that was intended to show that two linguistic
statements can be seemingly contradictory, _if_ misinterpreted, but yet remain
both true, when understood appropriately and in context. In this specific case,
the contrast is between a general statement of typical valuation of a grapheme,
_s_, and a more specific statement about its value in a particular historical and
phonetic environment. Which is precisely the case with Tolkien's description
of the value of _ll_ in the particular historical and phonetic environment
where /-L-/ < /-lth-/ < *_-lt-_. And finally, I disagree fundamentally with
your claim regarding the difference between living languages and Tolkien's
art-languages, which were _intended_ by their creator to _appear_ to be and
behave just as actual, historical languages do. As for App. E, the very words
you quote undermine your claims: Tolkien writes of representing sounds
with _fair_ accuracy, not complete accuracy; and to _avoid_ what he judged
to be "uncouth" appearance. We know that he avoided _dh_ for this reason;
and there is no reason to think he would not avoid _lh_, which looks extremely
similar to _dh_, for precisely the same reason. CFH]
Fourthly, it is not the only discrepancy between Sindarin in LR and in VT42.
There are at least two other irreconcilable points:
- former mp, nt, ñk are said here to give mf, nþ, ñx and later the long
unvoiced nasals mh, nh, ñh in the southern dialects of Sindarin including what Men
learnt. Yet in LR and the published S they rather seem to give mm, nn, ng. In
the appendix C of his article "Reconstructing the Sindarin verb system",
http://www.uib.no/People/hnohf/sverb-rec.htm, Helge Fauskanger discusses this
question - in a prescriptive approach, but it does not affect the relevance of
examples. He notes that in LR Appendix E, it is said that "Grade 6 (21-24) should
then have represented the voiceless nasals; but since such sounds were of very
rare occurrence in the languages concerned...", which would not have been
true of Sindarin as presented in VT:42.
[Even if this really were a discrepancy with _LotR_, that has no bearing on the
specific claim Tolkien's account of _ll_ must be revisionist. But in fact I see no
_necessary_ discrepancy. First, the examples that Helge gives are mostly from
the Noldorin of the _Etymologies_, which Tolkien was not describing in this
late essay. Second, Helge's discussion takes no account of the possibility of
analogical levelling in these forms, even though we know that analogy played
a major role in Sindarin grammar. Third, those examples Helge gives from
_The Lord of the Rings_ are only _assumed_ to have developed from the
specific combinations Tolkien is discussing here, and only _assumed_ to
belong to the specific dialects Tolkien is describing. These assumptions are
made for the sake of argument, which is fair enough -- Helge himself points
the uncertainties he is navigating with the _LotR_ forms; but you can't treat these
assumptions as fact and then use them to "prove" that Tolkien was making a
contradictory statement. In fact, when contradictions are arrived at, it is the
_assumptions_ that must be discarded. CFH]
- the correspondence Telerin _glania-_ / Sindarin _gleina-_ shows that in
VT42 the development of a medial sequence VCiV in Sindarin is thought to be
DiphthongCV with the diphthong arising from a mutated vowel + epenthetic i. This is
not uncommon: compare Ancient Greed _bainô / Latin _veniô_ "I come" from an
reconstructed prototype *gwmyô (gw == labiovelar), Latin _ratiône(m)_ / French
_raison_ "reason", _gloria(m)_ / _gloire_ "glory" (the oi diphthong
subsequently > oe > we > wa today). It certainly stands beyond Sindarin plural patterns
like _aran_ "king", pl. _erain_. Yet in the Sindarin seen in LR and S it is
seen finally, and not medially; otherwise we would get things like **Gilthoenel
and not _Gilthoniel_, **peiran and not _perian_, **egleiro and not _eglerio_,
all three well_known from LR; **geneidad and not _genediad_ (King's Letter,
IX:126-9); **arnoeidad and not _arnoediad_ (where clearly oe==fronted o, not a
diphthong) in S, etc. I remember only two similar cases, probably: _Einior_
"Elder" XII:358, and _Eirien_ "Daisy" IX:129-31.
[Again, whether or not this is a real discrepancy with _LotR_is irrelevant to the
issue of the value of _ll_. But any claim of revisionism on the basis of this
case founders utterly on the fact that the very essay in question also shows
what you seem to think is the development required in terms of Sindarin
as exhibited in _LotR_, as, for example, in S _seidia-_ < _satya-_ (VT42:20).
Since both developments are exemplified within this essay, it cannot be said
that Sindarin as Tolkien conceived of it when _LotR_ was published did not
also have this alternation, but simply by circumstance happened only to
use forms exhibiting the one but not the other. The proper scholarly response
to this situation ought to be to examine what might account for the
apparent variation of development (if indeed it is not only apparent)
within Sindarin, not to simply dismiss the phenomenon as an inconsistency
with what was previously, and wrongly, thought to be an established,
"canonical" fact about Sindarin of _LotR_. CFH]
All this inclines me to the opinion that Tolkien did change his mind about
several points of his Sindarin in writing _The Rivers and Beacon-hills of
Gondor_. It is even possible that he was beginning a throughout revision of the
language - that he never had the time to complete.
[I certainly do not dispute that Tolkien's conception of Sindarin changed
between the time that _The Lord of the Rings_ was published in 1954-55 and
the time that he wrote _The Rivers and Beacon-hills of Gondor_ in 1969. But
a) I see no _necessary_ change in the specific cases you highlight; and b)
that is not the issue: the issue is whether his account of the two values of
_ll_ in Elvish Sindarin in _LotR_ represents a revision, as David Salo claims
and Helge suggests. I see no reason whatsoever to think it is, as Tolkien's
specific statements regarding _ll_ are not in any way contradicted by the
actual evidence in _The Lord of the Rings_. CFH]
In the post n°551, December 8th, Carl Hostetter also observed:
> While on the subject of long voiceless resonants, and more generally on thedialectal variations Tolkien describes in the passage quoted in my
_Tengwestië_ article, I would like to point out that these developments and dialectal
variations are clearly modelled on very similar themes in Welsh and Welsh
Certainly. I notice that by the changes I alluded to above, Sindarin
evolution as seen in VT:42 seems to parallel Welsh's even more closely. Currently I
have with me only notes from _Language and History in Early Britain_ by Kenneth
Jackson, Edinburgh University Press, 1953, especially his chronology of
phonetic changes in the three Brittonic languages (Welsh, Cornish, Breton) till the
twelfth century, pp. 694 and following. Some are interesting with regard to
- the several waves of i-affection have differences between Welsh on the one
hand, Cornish and Breton on the other hand
- in Welsh mp, nt, nc > mh, nh, ngh medially (in Modern Welsh the h has
disappeared in non-initial unstressed syllables), whereas Cornish and Breton
preserve the stops
- in Welsh specifically, lt > ll.
[Non-text portions of this message have been removed]
... A further thought on this point: The correct value of _h_ in Quenya and _f_ in Sindarin can be readily discerned by the reader strictly on the basis ofMessage 1 of 5 , Dec 9 7:50 PMView SourceOn Dec 9, 2003, at 7:33 AM, Tchitrec@... wrote:
> it is true that other graphemes have double values in ElvishA further thought on this point: The correct value of _h_ in Quenya and
> languages, such as H in Quenya or F in Sindarin, but those cases are
> explained in Appendix E. Why would he not have done the same for LL if
_f_ in Sindarin can be readily discerned by the reader strictly on the
basis of phonetic environment in the various forms involved. This is
not the case with the multiple values of _ll_ and _d_, where the
correct value can only be determined through a knowledge of the
etymology of the forms exhibiting them. Since the reader of _The Lord
of the Rings_ would have no ready means of knowing the etymologies of
the forms involved, information regarding the dual values of _ll_ and
_d_, unlike that for discerning the values of _h_ and _f_, would have
been essentially useless for the target audience of App. E, the
interested lay reader.
Carl F. Hostetter Aelfwine@... http://www.elvish.org
ho bios brachys, he de techne makre.
Ars longa, vita brevis.
The lyf so short, the craft so long to lerne.
"I wish life was not so short," he thought. "Languages take such
a time, and so do all the things one wants to know about."
... Also lh in Portuguese, like ll in Castilian, has the value of l mouillé. Hence it is unclear (as indeed any Latin spelling of any non-Latin sound, exceptMessage 1 of 5 , Dec 10 11:30 AMView SourceCarl wrote:
> [...We know that he avoided _dh_ for this reason; and there is no reason toAlso lh in Portuguese, like ll in Castilian, has the value of l mouillé. Hence it is unclear (as indeed any Latin spelling of any non-Latin sound, except for those familiar with the conventions: sh is a Schin in English but an h in Gaelic, th Thorn or Edh in English but h in Gaelic, final a is Latin a in Castilian but Sch'wa in Catalan, sz is Schin in Polish but s in Hungarian - were Schin is spelled s! - and so on). No doubt Tolkien used Tengwar to circumvent the difficulty, and he specifically states that the Tengwa alda - ld in Quenya - is used for voiceless l (whatever be the Welsh name for that sound) in Sindarin.
> think he would not avoid _lh_, which looks extremely similar to _dh_, for
> precisely the same reason. CFH]
> [...These assumptions are made for the sake of argument, which is fair enoughA good point in _any_ branch of science, whether grammar (Tolkienian or non-Tolkienian) or astronomy or whatever.
> -- Helge himself points the uncertainties he is navigating with the _LotR_ forms;
> but you can't treat these assumptions as fact and then use them to "prove" that
> Tolkien was making a contradictory statement. In fact, when contradictions are
> arrived at, it is the _assumptions_ that must be discarded. CFH]
Hans Georg Lundahl
Höstrusk och grå moln - köp en resa till solen på Yahoo! Resor
[Non-text portions of this message have been removed]
... Useless only as a way of indicating the correct pronunciation of every single word. But isn t it useful to the interested lay reader to know that theyMessage 1 of 5 , Dec 10 11:46 AMView SourceOn 10.12.2003, at 04:50, Carl F. Hostetter wrote:
> Since the reader of _The Lord of the Rings_ would have no readyUseless only as a way of indicating the correct pronunciation of every
> means of knowing the etymologies of the forms involved, information
> regarding the dual values of _ll_ and _d_, unlike that for discerning the
> values of _h_ and _f_, would have been essentially useless for the target
> audience of App. E, the interested lay reader.
single word. But isn't it useful to the 'interested lay reader' to know
that they _d_ and _ll_ etc. could have such value ? Or indeed simply the
statement on etymology as such. Tolkien does furnish the
reader with etymological details although they're not strictly
necessary for pronouncing the words correctly.
[Apparently, in Tolkien's judgement, no, the information was not
useful _enough_ to be included. Quite reasonably so: with no way
to determine the etymology of most Sindarin words in _ll_, so as
to distinguish the correct pronunciation, what would the reader
do with the information? In the event, the number of words in
_LotR_ that seem at all likely to have the pronunciation /L:/ of _ll_
are small; so the errors of pronunciation would likewise be small,
small enough to satisfy Tolkien's stated goals of describing the
pronunciation with "fair" (not scientific) accuracy, while avoiding
what he judged to be spellings uncouth to his English readers. CFH]
In fact, while Tolkien gives the reason for excluding _dh_ originally
(a point on which he later changed his mind), he doesn't say anything
about _ll_. The exclusion of _dh_, in my view, precisely doesn't
parallel the writing _ll_ as _lh_ is indeed used in the Lord of the
[You're entitled to your view, but in my opinion the visual parallel
between uncouth _dh_ and _lh_ could hardly be more striking.
And note that Tolkien did not merely "exclude" _dh_, he altered its
_representation_ to _d_, which thus, like _ll_, represents two
different values depending on etymology. CFH]
Tolkien writes: In the transcription of Elvish Sindarin in _The Lord of
the Rings_ _ll_ is used in the manner of modern Welsh for the medial
voiceless _l_; as in _mallorn_ < _malhorn_ < _mal�orn_ < _malt_
�gold� and _orn_ �tree�. So apparently we have the following scenario:
S. _lh_ == _ll_ when < *_-lC-_.
[No, that is not at all apparent, nor is that at all what Tolkien is saying.
He is speaking specifically and _only_ of _-ll-_ < _-lh-_ < _-lth-_
< *_-lt-_. Neither the etymological figure nor Tolkien's statement makes
any claims whatsoever about the value of _ll_ developed from other
combinations (e.g., *_-ld-_). This was _precisely_ the point of my article.
Phonetically that is; they both represent a *short* voiceless _l_.
[No. Tolkien states specifically that "Medially ... _lth_ (_l�_) became _long_
voiceless ... _l_, though the old spelling [_lth_] was mostly retained (beside
... _lh_)" (emphasis mine), and that he used _ll_ in _The Lord of the Rings_
to represent the sound arising from this specific source. Now, we don't know
for certain -- because Tolkien doesn't say, nor does his statement prevent
the possibility -- that the long voiceless /L:/ that developed from medial
_-lth-_, and was spelt as both _lth_ and _lh_, didn't eventually shorten to
simple /L/ by the Third Age, in which case _ll_ could indeed represent short
/L/; but we would need other evidence to show or support this development.
Nor, of course, does this conjecture of possible simplifcation beyond the
stage of development that Tolkien describes have any bearing on the point
of my article, or on the point of Tolkien's statement, which concerns the fact
(previously unknown, and thus now disruptive of certain "canonical" notions,
and therefore eagerly desired to be set aside) that _ll_ was used in _The Lord
of the Rings_ to represent two different values, depending on etymology. CFH]
So, the phonetic value (voiceless l) is represented in _The Lord of the
Rings_ by both _lh_ and _ll_. Thus neither orthography can be considered
uncouth (especially as _lh_ doesn't figure in English but didn't give
rise to worry).
Interestingly Tolkien says that he uses _ll_ *medially* as in Welsh.
Briefly, the differentiation _-lh-_ and _-ll-_ could be an
etymological one but would still not be exactly in the manner of Welsh,
which uses _ll_ for a simple voiceless _l_ in all places.
[That is because Tolkien is specifically describing a development that
only occurs medially: _-lh-_ < _-lth-_ < *_-lt-_. He is not addressing
at all the development of initial _lh-_ < initial *_sl-_. CFH]
Or else, we take Tolkien's above cited statement [VT42:27] as meaning
"...medial [long] voiceless _l_". In that case there is a good
phonetic reason for the orthographic differentiation of _lh_ and
_ll_. This would be in tune with Tolkien saying that: "Medially however
_lth_ ... became long voiceless _l_ .... Still, the usage wouldn't
exactly be in the manner of Welsh either.
At any rate, if _ll_ in _The Lord of the Rings_ is to represent voiceless
_l_ in the manner of Welsh (that is short) what about _nn_, _mm_, and
ng ? (Modern Welsh only has one more voiceless resonant written _rh_ at
all times although in Middle Welsh, _rr_ too occurs). Do they also at
times represent voiceless sounds in the manner of... Tolkien (?)
altough he never says so (the Elves, when not retaining the old
orthography, that is _lth_ etc., wrote _lh, _nh_ etc.) ?
Are we to assume that these doubly written consonants were at some
point meant to primarily represent double consonants rather than
Whatever the answer, I don't think the 'revision' theory can be easily
[I don't think the "revision theory" is a theory at all, as it does not explain
all the available evidence. Again, precisely the point of my article. CFH]
[Non-text portions of this message have been removed]
... Know the truth. [There are very many facts about his languages that Tolkien could have included in the Appendices had his purpose in writing them been toMessage 1 of 5 , Dec 11 4:52 AMView SourceOn 10.12.2003, at 20:46, Carl Hostetter wrote:
> with no way to determine the etymology of most Sindarin words in _ll_,Know the truth.
> so as to distinguish the correct pronunciation, what would the reader
> do with the information?
[There are very many facts about his languages that Tolkien could have
included in the Appendices had his purpose in writing them been to allow
the reader to know every linguistic truth. But that was not his purpose;
instead, his purpose was, as he stated in the Appendices, "to represent
the original sounds (so far as they can be determined) with _fair_ accuracy,
and at the same to produce words and names that do not look uncouth in
modern letters" (emphasis mine). CFH]
> [You're entitled to your view, but in my opinion the visual parallel1) Let me get this straight: Was there no _#lh-_ in the first edition
> between uncouth _dh_ and _lh_ could hardly be more striking.
> And note that Tolkien did not merely "exclude" _dh_, he altered its
> _representation_ to _d_, which thus, like _ll_, represents two
> different values depending on etymology.
of the LotR?
If there was, why wouldn't it be used medially ? That's my point.
[I can think of only one instance: _Amon Lhaw_. This might be an isolated
case where Tolkien forgot his own decision. On the other hand, how else
could he have represented it? _Ll_ would be out of the question, because
_in initial position_ _Ll_ would be even more uncouth and foreign to
English readers than _Lh_ (or so feels this English reader). Tolkien could
instead have used _L_, but for whatever reason did not. In the particular,
rare (if not unique) case, Tolkien may have weighed the balance of accuracy
and uncouthness in favor of _Lh_ because it was in initial position; medially,
_lh_ would tend to be pronounced by English readers as _l_ + _h_ -- as,
indeed, it is _sometimes_ to be pronounced; so again, one would have to
know the underlying etymology in order to decide on the correct
pronunciation; and thus it would be no improvement, in addition to
being uncouth. CFH]
2) Let's not play with words! By using _d_ instead of _dh_ the spelling
_dh_ is excluded from the text. Your point about the double value is
valid but note that I never doubted the fact that _ll_ is meant to
represent two different values in strict accordance with Tolkien's
statement cited in VT42.
> So apparently we have the following scenario:That's a typo. It should of course read <*_lt_.
> S. _lh_ ==== _ll_ when < *_-lC-_
Regarding the issue of whether _ll_ is single or double voiceless _l_,
the big point, which you seem to miss, is that Tolkien does say that
_ll_ represents short voiceless _l_. He implies it when saying (VT42:27):
"In the transcription of Elvish Sindarin in _The Lord of the Rings_ _ll_
is used in the manner of modern Welsh for the medial voiceless _l_;
as in _mallorn_ < _malhorn_ < _malþorn_ < _malt_
‘gold’ and _orn_ ‘tree’."
Clearly, there is no 'conjecture' as Welsh _ll_ represents short
voiceless _l_ (1). Also Tolkien writes "... voiceless _l_" not
**"...voiceless _ll_". So the passage you cite must refer to a
transitional stage, not the one reflected in the LotR and its spelling.
[I certainly recognize this as a possible _implication_ of what Tolkien
writes, as I have already said; but I do not consider this an explicit
statement or proof. To my mind, Tolkien may only be referring to an
orthographic convention in Welsh, not intending thereby to make a
precise claim as to phonetic length. Alternatively, he _may_ have had
such in mind, but I don't see this as a _necessary_ implication of his
Thus, if we take what Tolkien himself says seriously, we have a double
representation of one and the same sound (unless _#lh-_ doesn't figure
in the original version of the LotR, which I don't know).
[I do take Tolkien's words "seriously"; seriously enough to consider the
context in which statements are made, and to distinguish possible
implication from established fact. CFH]
If _lh_ figures the reason for not using _lh_ medially is etymology not
[I maintain that the reason Tolkien _ll_ instead of _lh_ _may_ have
been for visual aesthetic reasons, as explained further above. CFH]
(Or does initial _lh_ look less uncouth to English speaking eyes that
medial _lh_ ? Seriously.).
[Again, see above. And thank you for giving my article serious attention,
and for sharing your thoughts on it, and giving me an opportunity to
further consider, explore, and clarify my own thoughts on the matter. CFH]
(1) In Modern Welsh, that is. In Middle Welsh _ll_ could represent both
long voiced _l_ or short voiceless _l_.