More on Tengwar (cont.) ======================= First of all, this part is not meant to be a complete and detailed description of the Tengwar system. There are several books and web sites that cover the topic; I will try to explain here roughly what I have understood while reading Tolkien's documentation, and these explanations will **only** be used as a basis for the `following `__ discussion on how to use the *tengwar* system to write Lojban. This is not strictly necessary to the rest of this document, and indeed much of it is only of cultural interest (references to old-fashioned uses of *tengwar*, for example). However, it remains necessary to **justify** the choices that will be made later. Please bug me about any omission here that could have proven useful for the next part. Notes before reading -------------------- While the Tengwar is mainly a phonemic system, each sign corresponds to a nearly unique sound, and moreover, athough these sounds may have evolved, in the Middle-Earth of the Third Age, *languages only involved "pure" sounds*. Let's introduce a short list of phonetic transcriptions, that will be sufficient to describe nearly all the Tengwar system: .. _stable: =============================== ============================================================== ==================================================== Latin letters used in this part `ASCII IPA `__ notation words where the sound is found *a* ``[a]``, ``[A]`` park, father *e* ``[e]``, ``[E]`` *"a"* in say, met *i* ``[i]`` pizza, machine *o* ``[o]`` pure *"o"* in joke, *"o"* sound of french *"haut"* *u* ``[u]`` moo, zoo *y* (see note below) ``[y]`` *"u"* in french *"lune"* *c, k* (see note below) ``[k]`` kill, token, flak *sh* ``[S]`` ship, dish *ch, kh* (see note below) ``[x]`` not in english, german *"Bach"* or scottish *"loch"* *th* ``[T]`` voiceless english *"th"*, as in "thin" or "cloth" *dh* ``[D]`` voiced english *"th"*, as in "the" or "clothes' back *ng* ``[N]`` nasal g, as in "finger" front *ng* ``[n.]`` "gn" sound, as in french *"cro-magnon"* =============================== ============================================================== ==================================================== The various consonants which have obvious pronunciations for english-speakers are not shown in the previous table, and further details about them will only be given when needed. Note that I am using the same Latin notations as Tolkien, this is to ease the reading for those people who have read Tolkien's work. In particular, do not confuse the above sound *"c"* with lojban *"c"*, the latter being pronounced *"sh"*. More notes on sounds ~~~~~~~~~~~~~~~~~~~~ *"c"* and *"k"*, according to Tolkien, represent the same consonant. In fact, the *"k"* letter was used by the Elves when transcribing foreign names when the *"c"* sound was found (cf. Kloczko). Same applies for differences between *"kh"* and *"ch"*, which are the fricative equivalents of *"k"* and *"c"*, respectively. .. _yv: Additionally, Tengwar-aware people may have noticed that I am using *"y"* in the Sindarin fashion. The original (Quenya) *"y"* sound was used as a consonant, much like its english counterpart, in words like english "you". Sindarin, along "modern" elvish-like languages, introduced the fronted *"u"* sound (as un french *"lune"*), and latin symbol *"y"* was used for it instead. This letter is therefore very different from the Lojban "y" sound, which was very uncommon in Tolkien's languages. When the Lojban-style sound (IPA *schwa*) was needed, it was also written as *"y"* by Tolkien, but a dot-below *tehta* is used for it in various Tengwar modes. In the following, I shall use *"y"* for the phoneme described in the above table, while the *schwa* sound will be discussed in the `next part `__. Referring to the *tengwa* tables ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Make sure you have given a glance at the `Tengwar tables `__ before reading on. Through the remaining of this description, I shall refer to the *tengwar* by indexes given in the tables, or, in the case of vowels, by their description. .. _cons: The base *tengwar* ------------------ The most complete description on the *tengwar* system is about consonants. The first two tables given correspond to the most common form of presenting the *tengwar*, and we shall show that it is also the most logical. Structure of the base table ~~~~~~~~~~~~~~~~~~~~~~~~~~~ In the first table, columns represent "Series" and rows represent "Grades". The series correspond themselves to the main positions of the tongue in the mouth when forming the consonants, while the grades correspond to the manner in which consonants are pronounced. As already said, nearly all Tengwar modes used the base *tengwar* in the same way. This organization comes together with the way in which the signs are drawn. Each sign consists of a "stem" (*telco*) and one or more "bow" (*lúva*). The bows can be open (Series I and III) or closed (Series II and IV), and the stem can be either on the left (Series I and II) or on the right (Series III and IV). The Series of a particular *tengwa* can thus be determined by observing the bows and the side of the stem. Each base *tengwa* has at least one bow; another is added to indicate voicing of the consonants. For stopped consonants, the stem goes downwards; it is raised when the consonant becomes spirant, that is, when air is breath through the positioning of the mouth. Shortening the stem indicates a nasal or weakened consonant, as in Grades 5 and 6. Correspondance between the table and phonemes ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ .. raw:: html
Grade 1: tinco (1) parma (2) calma (3) quesse (4)
Each Series corresponds to a family of sounds that are pronounced similarly around a single position of the tongue in the mouth. To distinguish between them, one may use Grade 1, which is applied to the base consonants, that is, stopped and unvoiced consonants. The sounds of this row are, in order, *"t"*, *"p"*, *"k"* and *"kw"*, corresponding to *tengwa* nr. 1, 2, 3 and 4, resp. In newer Elvish modes, *tengwa* nr. 4 is used for *"k"*, where *tengwa* 3 is used, as explained above, for foreign *"c"* transcription. .. raw:: html
Grade 2: ando (5) umbar (6) anga (7) ungwe (8)
Obviously, the doubling of the bow in Grade 2 indicates voicing of Grade 1, therefore we get *"d"*, *"b"*, *"g"*, and *"gw"* for *tengwar* nr. 5, 6, 7 and 8, resp. The sound present in Series III of Grade 2 is not clearly defined; although its first pronunciation was *"g"*, it is sometimes (in the newer Westron mode, for example) pronounced as *"j"* (in which case, Series IV is used for sharp consonants, which are not rounded by *"w"*). Also, the original Quenya modes used Grade 2 for voiced **and** nasal stops, using *tengwar* 5 through 8 for *"nd"*, *"mb"*, *"ng"* and *"ngw"*, hence their names. However, this ancient mode is not commonly used any more. It is interesting to note that according to Tolkien, the *"c"* letter (*tengwa* 3) was in some cases used for a backward (palatized) *"t"* sound, which is allophonic with frontal *"t"*, but introduces continuity in the newer uses of Series III, as voicing such a palatal *"t"* gives the *"j"* sound. However, many Quenya modes insist on having Series III and IV for sharp and "rounded" consonants (with *"w"* appended to them). .. raw:: html
Grade 3: thûle (9) formen (10) charma (11) chwesta (12)
Grade 3 applies, as in Grade 1, to unvoiced consonants, but here they are spirant. Thus *"t"* becomes *"th"* (9), *"p"* becomes *"f"* (10), *"c"* becomes *"ch"* (11), and *"kw"* becomes *"chw"*. Recent modes use Series III (*tengwa* 11) for the sound *"sh"*, following the idea of using Series III for front-palatal consonants. In this case, too, *tengwa* 12 is used for *"ch"*, and not *"chw"*. .. raw:: html
Grade 4: anto (13) ampa (14) anca (15) unque (16)
The original Quenya mode used this grade for consonants stops preceded by a nasal, and thus assigned *"nt"*, *"mp"*, *"nc"* and *"nkw"* to *tengwar* 13 to 16 (hence their names again). However, in newer modes, the meaning of the second *lúva* is respected, and Grade 4 is the voicing of Grade 3, which leads to *"dh"*, *"v"*, *"j"* and *"gh"* for *tengwar* 13, 14, 15, 16, resp. .. raw:: html
Grade 5: nûmen (17) malta (18) ñoldo (19) ñwalme (20)
Grade 5 is used for nasal consonants. Keeping the previous meanings for the four Series, we have *"n"*, *"m"*, and **back** *"ng"* and *"ngw"* for *tengwar* nr. 17, 18, 19, 20. Newer modes, again, have front **and** back *"ng"* for *tengwar* 19 and 20 (front consonant for nr. 19, and no *"w"* appended for nr. 20). .. raw:: html
Grade 6: ôre (21) vala (22) anna (23) wilya (24)
Lastly, Grade 6 is used for the "weakiest or semi-vocalic consonants of each Series". *Tengwa* nr. 21 is commonly used for untrilled *"r"* (but of longer duration than the flapped "r" of english) and nr. 22 and 23 for the **consonant** *"w"* and *"y"* (as opposed to the vowel *"y"*). It should be noted that the "Mode of Beleriand" uses nr. 23 for vowel *"o"*, and that *tengwa* nr. 24 was used in ancient Quenya modes for consonant *"w"*. There is an hypothesis according to which *tengwa* nr. 23 (*anna*) was used in ancient modes as a "carrier" for `vowel tehtar `__. .. _add: The additional *tengwar* ------------------------ .. raw:: html
silme (29) silme nuquerna (30) àze (31) àze nuquerna (32)
You may have noticed that the voiced and unvoiced palatal-dental fricatives *"s"* and *"z"* are not covered by the 24 first *tengwar*. These two are commonly represented by additional *tengwar* nr. 29 and 31 (double and single curl) resp., of which nr. 30 and 32 are equivalents used when a *tehta* has to be superimposed. In later Elvish languages, where the *"z"* sound didn't exist any more, *tengwar* nr. 31 and 32 were used for a doubled *"s"*. It is especially interesting to note that the double curl *tengwa àze* used for *"z"* follows the standard of adding a second *lúva* (bow) for voicing. .. raw:: html
rómen (25) arda (26) lambe (27) alda (28)
*Tengwa* nr. 27 (*lambe*) is universally used for *"l"*. Nr. 25 (*rómen*) is a modification of nr. 21, designed to represent a trilled *"r"*. Nr. 26 (*arda*) and 28 (*alda*) are unvoiced or composite counterparts of nr. 25 and 27, therefore represent somewhat *"rh"* and *"lh"*, or in ancient Quenya modes *"rd"* and *"ld"*. .. raw:: html
rómen (25) arda (26) lambe (27) alda (28)
Nr. 33 (*hyarmen*) was originally a weaker counterpart of 11, but in later modes commonly stands for the *"h"* consonant separator. Nr. 34 (*hwesta sindarinwa*), when used, is the voiceless counterpart of nr. 22, that is somewhat the *"hw"* sound. Nr. 35 (*yanta*) and 36 (*úre*) are not often used as consonants (when they were, they stood for consonant *"y"* and *"w"*); the mode of Beleriand uses them for vowel *"e"* and *"u"*, resp. .. _vowels: Tengwar *tehtar* and vowels --------------------------- In most uses of Tengwar, vowels are represented by *tehtar* imposed over adjacent consonant *tengwar*. In languages where most words end by a vowel, such as Quenya, the *tehta* is imposed over the **preceding** consonant, while in languages such as Sindarin or the Black Speech, where most words end by consonants, the *tehta* is set over the **following** consonant. .. raw:: html
Short carrier: short carrier
Long carrier: long carrier
Where a vowel is needed but no consonant is present, the *tehta* is placed above the "short carrier", which looks like an undotted "i". For long (lengthened) vowels, the *tehta* should be placed above the "long carrier", which looks like an undotted "j", but cases where it is doubled over the following or preceding consonant are possible (especially with the curls or accents, as shown below). The various vowel *tehtar* ~~~~~~~~~~~~~~~~~~~~~~~~~~ There exist many *tehtar* to express vowels. The most commonly used are: +-----------------------------------+-----------------------------------------------------+ | |3 dots above| | Three dots above for the *"a"* | | | vowel, often handwritten as | | | dashes or a circumflex accent, | +-----------------------------------+-----------------------------------------------------+ | |dot above| | Single dot used for *"i"*, or | | | for *"e"* when the following | | | "acute" accent is used for | | | *"i"*, | +-----------------------------------+-----------------------------------------------------+ | |acute accent| or |double acute | Single or doubled "acute" | | accent| | accent, for the vowel *"e"*, | | | or sometimes *"i"*, | +-----------------------------------+-----------------------------------------------------+ | |forward hook| or |double forward | .. _ou: | | hook| | | | |backward hook| or |double | Curls for *"o"* and *"u"*, | | backward hook| | where the curl to the right | | | ("forward hook") represents | | | the most used sound. For the | | | Black Speech where *"o"* was | | | rare, the curl to the right | | | was used for *"u"*, as shown | | | on the `Ring | | | inscription <_static/images/ring.gif>`__; | | | but for Quenya and Sindarin, | | | the curl to the right is used | | | for *"o"*, therefore the curl | | | open to the left is used to | | | represent *"u"*. | +-----------------------------------+-----------------------------------------------------+ | |one dot below| or |two dots | One or two dots below (and | | below| | sometimes above) for *"y"*. | +-----------------------------------+-----------------------------------------------------+ As explained before, the doubling of the *tehta* indicates lengthening of the vowel. The signs for the vowel *"y"* should be taken carefully, because the "y" latin letter was used through the various works on the *tengwar* system for several different phonemes. See in `this above paragraph `__ for a brief explanation. Please remember that the "two dots below" *tehta* is used for the english-like "y" sound, and is applied to a consonant *tengwa* to tranform it into another, but compound, **consonant** *tengwa* (thus the specially associated *tyelpetéma* Series), above which another vowel *tehta* can be applied. The preceding explanation might seem a little bit confusing. Here is an example: the complete *tengwa* used to write the phoneme *"tye"* in the word *"tyelpetéma"* itself. .. raw:: html
tye
Writing vowels in full ~~~~~~~~~~~~~~~~~~~~~~ The mode of Beleriand, of which a sample is given on `the inscription of the Western gate of Moria `__, is a mode where vowels are written in full. The sample includes a representation for each vowel, here is the correspondance map: .. raw:: html
"a" "e" "i" "o" "u" "y"
bow open to the right tengwa yanta short carrier tengwa nr. 23 tengwa úre tengwa nr. 30
You may wonder if the use of the short carrier and *tengwa* nr. 35, 23 and 30 do not clash with their possible use as consonants. In fact, *tengwa* nr. 30 is the form of *tengwa* 29 used when a *tehta* is imposed; it is thus of no use in a mode where there are no *tehtar*. Similarly, *tengwa* nr. 23, in ancient Quenya modes, is a *tehta* carrier (so is the short carrier in any modes involving *tehtar*), and *tengwa$ nr. 35 is used for particular* "y"\* sounds not found in the language of Beleriand. It is important to note that the *"y"* sound in the language of Beleriand is the one described in the `table `__ at the beginning of this part, which sounds like *"u"* in the french word "*lune*". The full-letter form of this vowel, which has only been found in the inscription of the Gate of Moria, should thus be taken as a representation for this sound, and supposedly **not** for other elvish *"y"* sounds. Finally, in this mode, vowel lengthening is indicated with an "acute" accent set over the vowel *tengwa*. .. _punct: Punctuation and other *tehtar* ------------------------------ .. raw:: html
Consonant underbar: underbar
Consonant doubling is almost always represented with a horizontal bar placed under the *tengwa*. .. raw:: html
"s" curls: bottom curl bottom curl bottom curl or bottom curl
When a *"s"* sound is clustered with a consonant, it is very common (especially in *"ps"*, *"ts"* and *"ks"*) to write it as a curl *tehta* attached to the final bow, stem or bow closing. Please notice the special curl used for *tengwa lambe*. .. raw:: html
tengwa halla: simple upward stem
The smooth breathing introduced between sounds in words which require one is represented in the "modern" Elvish tongues by *tengwa* nr. 33 (*tengwa hyarmen*). However, ancient Quenya modes, like this of Elvenhome, used a simple raised stem without bow, called *halla*. .. raw:: html
pusta: medium dot
double pusta: colon
quadruple pusta: double colon
Separation between parts of a single utterance is made with the *pusta* symbol, or medial-dot. Although Tolkien doesn't describe it, the West-gate inscription, the Ring inscription and the Fëanorian poem `Namárie `__ also feature a colon-like shape, called "double *pusta*", to divide utterances. A "quadruple *pusta*", as shown in the poem *Namárie*, can be put at the end of a speech block, such as a written paragraph. .. raw:: html
End-of-text curl: horiz. curl
End-of-speech curl: vert. curl
The horizontal curl is found on many inscriptions after a double *pusta* to indicate the end of the written text. A vertical curl can be used to indicate the end of speech; however its similarities of meaning with the horizontal curl kept it from being often used (you can find it in *Namárie*, though). .. _digits: Digit *tengwar* --------------- .. raw:: html
0 123 456 789 1011
The above digit *tengwar* were used in two ways: for decimal (base 10) and duodecimal (base 12) numbering. When the *decimal* numbering is used, it is common to specify it by **overlining** the number, or, in the case of a single digit, by imposing a single dot *tehtar* **over** it. When the *duodecimal* numbering is used, it is common to specify it by **underlining** the number, or, in the case of a single digit, by imposing a single dot *tehtar* **under** it. Nearly all writing modes for the *tengwar* system have numbers written in little-endianness, that is, with the digit of least significant written **first** (which is the opposite of the big-endian Western world manner). However, it is possible to find numbers written the other way around. To distinguish the two modes, a small circle is drawn below the digit of least significance. If the small circle is not present, little-endianness is assumed. .. |3 dots above| image:: _static/images/t/vau3.gif :height: 50px .. |dot above| image:: _static/images/t/vau2.gif :height: 50px .. |acute accent| image:: _static/images/t/vau1.gif :height: 50px .. |double acute accent| image:: _static/images/t/vau1b.gif :height: 50px .. |forward hook| image:: _static/images/t/vau4.gif :height: 50px .. |double forward hook| image:: _static/images/t/vau4b.gif :height: 50px .. |backward hook| image:: _static/images/t/vau5.gif :height: 50px .. |double backward hook| image:: _static/images/t/vau5b.gif :height: 50px .. |one dot below| image:: _static/images/t/vau6.gif :height: 50px .. |two dots below| image:: _static/images/t/vau6b.gif :height: 50px