Tengwar - The Classical Mode


Sample language: Quenya

Samples: DTS 12, DTS 20, DTS 26, DTS 40, DTS 55, DTS 62, DTS 66, DTS 70, DTS 73, DTS 74


Quenya was the mother-tongue of Feanor, and one of the first languages to receive written form in Tengwar. We do not know in detail, however, how Quenya was represented with Tengwar in the beginning, since most of our information about this mode of writing derives from the late Third Age of the sun, or the beginning of the Fourth (the time of the Red Book). This document, therefore, is primarily concerned with how Quenya was written in later ages, although references to earlier usage are given whenever this is known or can be deduced.

Even the contemporary material is however somewhat contradictory. Our most important sources are Tolkien’s description in AppE, and two renderings of the poem Namárie in Tengwar (DTS 20 and DTS 55, which below are referred to collectively as the Namárie manuscripts). Another important source is OP. The point of departure for this description is AppE; deviations in other sources will be noted as such.

It should be observed that many of the published texts in Quenya are not written in the Classical Mode, but rather according to the General Use (e.g. DTS 38, DTS 46, DTS 49, etc.). Such texts will not be considered here.

Note: the different tengwar are frequently referred to below by their Quenya names. For a complete list of these, see the article Tengwar Names.


Only consonants are in the Classical Mode represented by tengwar, while modifications of these, as well as vowels, are represented by diacritic marks — tehtar. Figure 1 shows the sound values of the tengwar for Quenya, according to the transcription method used in The Lord of the Rings (with the exception of the sound ng as in king, which is transcribed {ñ} in accordance with Tolkien’s usual custom). For the pronunciation of Quenya, refer to AppE. The symbol > signifies “developed into”, i.e. because of sound-changes within the language; >> signifies “replaced by”, i.e. because of a concious decision by the writers. Notes are given below. In most browsers you will be able to click on each tengwa to jump to the relevant note.

Figure 1: The tengwar

umbar thúle > súle harma > aha ñoldo > noldo ñwalme > nwalme óre anna wilya > vilya rómen silme / silme nuquerna áze > esse áze nuquerna > esse nuquerna hyarmen halla yanta úre


umbar. This tengwa normally represented mb, but after lambe it represented b. Cf. AppE: “For lv, not for lw, many speakers, especially Elves, used lb: this was written with 27 + 6, since lmb could not occur.”

thúle > súle. AppE states: “TH [. . .] this had become s in spoken Quenya, though still written with a different letter”. I.e., the sound /s/ that developed from /th/ was written with a different tengwa than the original /s/. That this “different letter” was thúle cannot be doubted. However, in DTS 20, silme is used for /s/ even where we would expect thúle, as in hísie. (The shorter text DTS 55 contains no certain novelties of this type. The only word that might have silme representing earlier /th/ in that manuscript is súrinen, but the etymology of this word is not firmly established.)

harma > aha. The ach-laut ch (spelt /x/ phonetically, but always written as {h} by Tolkien) could originally be found in all positions, and harma was used to represent this sound. When the pronunciation changed into a breath [h] in word-initial position the name was altered to aha in order to keep the original sound value in the name. In the Third Age, however, the sound had turned into [h] even medially, and only remained [x] before t, as in Telumehtar. According to OP, the tengwa continued to be used for medial h even after the change in pronunciation. In other positions, hyarmen or halla were used for h.

ñoldo > noldo. The sound ñ [ŋ] originally occurred initially, but had in the Third Age changed into n (or g after n, l, r [OP]). No Classical Mode samples of noldo are known; if the tengwa was used at all in the Third Age, it was probably as a conservative spelling of older ñ (cf. súle above).

ñwalme > nwalme. Only used for initial nw, which developed from ñw. Other occurences of nw (originating in n + w) are written númen + vilya.

óre. Quenya originally possessed “a weak (untrilled) r” [AppE] which was represented by óre. The weak r sound had evolved from d, and occurred between vowels [OP]. In the Third Age, “R represents a trilled r in all positions” [ibid.], and this trilled r was usually written with rómen. In the Namárie manuscripts, óre represents r before consonants and word-finally, while rómen is used before vowels (with one exception, probably by mistake, in DTS 55: the word ómaryo uses óre). The reason for this distribution of the r signs was probably æsthetic, rather than actually signifying a difference in pronunciation. According to OP, the Noldor tended to use óre for original trilled r, “but this was considered incorrect by the loremasters”.

anna. According to FQD, anna originally represented ʒ, which denotes a back spirant or voiced velar fricative [ɣ] (a voiced ach-laut). But this sound had vanished from Quenya before Feanor was born, so if used in this manner it must have been primarily for some other language of Aman that still possessed it — possibly Valarin — or in etymological texts where the lost sound was discussed. More usually, it represented any “silent” or “vanished” consonant, which was thought to have preceded a vowel that still remained in the word (as for instance in alda < ʒalda). Neither of these usages occur in any published specimen. It is possible that they vanished early in the history of the writing system, as the distinction of the old [ɣ] sound became redundant.

In the Namárie manuscripts, anna is used to carry a palatal tehta, where no other consonant is present in the required position (see Consonant Modifications below). This usage probably developed because the tengwa, usually representing “no consonant”, could bear the diacritic signifying y without adding any sound value of its own. A short carrier could theoretically have been used in the same way, but the Elves perhaps thought the carrier tended to look “over-loaded” with both a superscripted vowel-mark and a subscripted palatal tehta, more so than anna would.

wilya > vilya. Quenya w turned into v initially, but often remained w medially. In the sole Quenya text where vilya occurs (DTS 20), it is used for medial w, in the word vanwa. According to OP, the change of w to v was “usually also reflected in spelling”, which presumably means that vala was used to write the new v sound.

rómen. The additional letter rómen probably represented a trilled r through its entire history. In the Namárie manuscripts it is used for prevocalic r (even when palatalized or long), while óre is used in all other positions.

silme / silme nuquerna. Both silme and silme nuquerna are used for s, but the latter was “much used when accompanied by superimposed tehtar” [AppE]. Both the Namárie manuscripts attest this.

áze >> esse. The sound z that occurred originally in Quenya developed into r, which was represented by rómen. Unlike in the case of thúle, the Eldar felt no need to distinguish the new r sound from the older in writing. Therefore, the tengwa áze, then temporarily known as áre, “was in Quenya used for the very frequent ss of that language, and the name esse was given to it.” [AppE]. The usage of esse for ss is well attested in our sources.

áze nuquerna >> esse nuquerna. The variant form of esse had the same significance as its counterpart, but was “much used when accompanied by superimposed tehtar” (AppE; cf. silme nuquerna above). It was not required to, though: in DTS 20 there are several samples of esse rather than esse nuquerna being used with superimposed tehtar.

hyarmen. This tengwa first represented hy, a palatalized h. But when a tengwa was required for the newly developed breath h (cf. aha above), hyarmen was assigned to this sound (its older value was represented by adding the palatal tehta; see Consonant Modifications below). The tengwa was never given a new name, despite the fact that the pronunciation of hy /ç/ is quite distinct from h /h/.

halla. Before the development of breath h from initial ach-laut (see aha above), Quenya possessed another breath h represented by halla. Later the two h sounds merged, and halla apparently ceased to be used for breath h altogether.

Another function of halla was to “be placed before a consonant to indicate that it was unvoiced and breathed; voiceless r and l were usually so expressed and are transcribed hr, hl” [AppE]. This custom seems to have survived in later ages (in words like hríve, hlóce), although no samples are known. The sound hl “was in the Third Age usually pronounced as l” [AppE], but since Tolkien claims to have transcribed something from his source texts as hl, something that was usually expressed with the aid of halla, the sound must still have been represented in writing.

According to OP, the Noldor before the Exile pronounced hr, hl as r, l, but the spelling with halla was usually maintained. Later, after contact with Sindarin, many of them reverted back to the older pronunciation. Also according to OP, halla could in older Quenya be placed before medial hy, hw to indicate that the sound was long (in contrast to initial hy, hw which were short).

yanta. According to AppE, yanta was “mostly applied to y”. In the Namárie manuscripts, the tengwa is used for i, when it stands as the second vowel in the diphthongs ai, oi, ui (where it is pronounced similarly to Quenya’s palatal consonant y). See Diphthongs below.

úre. AppE states that úre usually represented w. In the Namárie manuscripts it signifies u standing as the second vowel in the diphthong au (where it is pronounced similarly to w). It would certainly also be used in the diphthongs eu and iu (at least when these were still falling diphthongs; “But iu in Quenya was in the Third Age usually pronounced as a rising diphthong” [AppE]). See Diphthongs below.

Consonant Modifications

Tehtar were used in the Classical Mode to indicate long consonants, palatalization, and following s.

Long or double consonants were indicated by putting a horizontal line beneath the baseline of the tengwa.

Palatalized consonants (i.e. consonants followed by a consonantal y) were indicated by the palatalization tehta, which most usually had the form of two subscripted dots. When occurring solely, and not following a consonant, the palatal sound y was represented by the silent tengwa anna together with the palatalization tehta.

In DTS 55 a variant form of the palatalization tehta is sometimes used, which consists of three dots instead of two. The primary reason for the alternate form is probably that long i is in that text represented by two superscripted dots (see Long Vowels below), and the scribe wanted distinct symbols for í and y.

Following s could be indicated by attaching a downward hook to the bow of the tengwa, “especially in the combinations ts, ps, ks (x), that were favoured in Quenya” [AppE]. No known Quenya specimen attests an s-hook, but luckily they are abundant in documents in other languages.


Figure 2: The ómatehtar

Vowels are in the Classical Mode represented by tehtar (or, more specifically, ómatehtar, vowel-marks [QE]). Figure 2 displays the tehtar for the vowels of Quenya.

The vowel tehtar were placed above the preceding consonant tengwa. When there was no preceding consonant, the tehtar were placed above a short carrier (shown in grey in figure 2). According to AppE, a “common form” of the short carrier was “like an undotted i”; all the attested short carriers have this appearance.

Long Vowels

Figure 3: Long vowels

The Quenya long vowels (in Roman transcription marked with an accute accent) were “usually represented by placing the tehta on the ‘long carrier’, of which a common form was like an undotted j. But for the same purpose the tehtar could be doubled. This was, however, only frequently done with the the curls, and sometimes with the ‘accent’” [AppE]. In DTS 55, all long vowels except á are consistently represented by doubled tehtar whenever there is a preceding tengwa available. That is, even í is indicated by a doubled dot, despite the indirect remark that this was not very common. Doubled tehta for long i is also attested in DTS 62. Known methods of indicating long vowels are shown in figure 3.

The Vowel A

Figure 4: Known ways to write ‘calma’

The most frequent vowel in Quenya was a, and it is notable that the tehta representing a is also the one most complicated to write. Tolkien notes that “The three dots, most usual in forming writing for a, were variously written in quicker styles, a form like the circumflex being often employed”. The circumflex a-tehta is attested in DTS 20 and DTS 62. Furthermore,

In Quenya in which a was very frequent, its vowel sign was often omitted altogether. Thus for calma ‘lamp’ clm could be written. This would naturally be read as calma, since cl was not in Quenya a possible initial combination, and m never occurred finally. A possible reading was calama, but no such word existed. [AppE]

Nevertheless, there was an obvious risk of misinterpretation. In some documents, this risk is minimized by marking consonants not followed by any vowel with a subscripted dot. Samples include DTS 42, DTS 51, and, perhaps most notably, DTS 64, where one of the words seems to be calma. In this case the text could not be read calama, since the l is marked as having no following vowel. It should be noted, though, that the only Classical Mode sample using the subscript dot is DTS 42, and that sample is dubious, since it contains one other feature not consistent with our knowledge of this mode. See SDT for a more detailed study of the subscript dot. There is also a thorough discussion of a omission in QS.

Another practice was to write out the vowel a only when stressed and / or initial, and to omit it in all other positions. This usage is evident in DTS 55. Although being just one word (namárie), DTS 12 apparently follows the same principle. (Also cf. the Quenya treatment of a in the Sarati.)

A summary of known or probable ways to write calma is given in figure 4.


Figure 5: The diphthongs

The Quenya diphthongs are ai, oi, ui, iu, eu, au. All were originally falling diphthongs, i.e. stressed on the first vowel. In the Namárie manuscripts, the first vowel in each diphthong is represented by a vowel tehta, the second by a tengwa. If the second vowel is i, it is represented by yanta; if u, it is represented by úre. The diphthongs are listed in figure 5. Unattested diphthongs are marked with an asterisk (*).

Normally, the tehta constituting the first part of the diphthong is superimposed on the tengwa representing the second part, which means that the symbols are read in the opposite direction than normally: superscripted tehta first, then tengwa, instead of the other way round. On one occasion in DTS 20, the tehta is however placed on the preceding tengwa (in the word caita), i.e. above the calma instead of above the yanta. This need not be a mistake, since the intended meaning should be equally clear.

The diphthong iu “was in the Third Age usually pronounced as a rising diphthong as yu in English yule” [AppE], and since this diphthong is not attested in Tengwar we do not know if it should also be written using úre. It might be that it should rather be represented by a palatalized anna (cf. Consonant Modifications above) + an u-tehta.

Sample text

Using the Classical Mode of writing, one could for instance produce the following text:

Íre hanyuvalye i tixi tehtaron ar i lúvar tengwaron, san istuvalye tece tengwainen.

“When you understand the dots of the tehtar and the bows of the letters, then you can write with the Tengwar.”

More Quenya texts can be found at the page Tengwar Illuminations.