The Qenya Lexicon Reviewed

[This is a slightly revised version of an article that originally appeared in Tyalië Tyelelliéva #14]

After several delays, the full contents of the fabled Qenya Lexicon was finally published in Parma Eldalamberon #12 in the early autumn of 1998; the QL now forms a counterpart to the Gnomish Lexicon published in Parma #11 in 1995. Prefixed to the Qenya Lexicon proper, Parma also published 28 pages of dense phonological descriptions, the material referred to by Christopher Tolkien in LT1:247: "Some early phonological description does exist for Qenya, but this became through later alterations and substitutions such a baffling muddle (while the material is in any case intrinsically extremely complex) that I have been unable to make use of it." The editors are probably to be commended for managing to present this material in a reasonably readable format. In this review I will however concentrate on the Lexicon itself.

Originally written in 1915, the Qenyaqetsa or Qenya Lexicon documents the beginnings of Tolkien's development of High-elven - a process that would go on for decades before LotR-style Quenya was reached, and in one sense not ending before it inevitably ceased with Tolkien's death. Organized by "roots", somewhat like the Etymologies, the Qenya Lexicon contains well over 3,300 Qenya words. (One estimate by the editors, "over 2,500", is technically correct but far too conservative.) While one of the editors some time ago made the peculiar contention that we learnt the entire contents of the Lexicon from them, fact is of course that a very substantial part of the Lexicon was published already in 1983-84: Taken together, the words quoted by Christopher Tolkien in the appendices to LT1 (p. 248-273) and LT2 (p. 335-349) amount to no less than a fifth of the total. (Indeed it is sometimes the plight of the editors to point out occasional errors in Christopher Tolkien's transcription; see pages 75, 79, 80, 90, 94, 95, 98, 99, 100, 101, 106.)

Based on the very substantial and representative excerpt in the appendices to LT1 and LT2, the language of the Qenya Lexicon could have been satisfactorily reviewed already fifteen years ago, and some observations were indeed done also before the entire QL at long last appeared. It has long been clear that Tolkien in the earliest period imagined a very different proto-language; six years ago, Anthony Appleyard observed that many early Qenya words "became invalidated when he dropped the BoLT [Book of Lost Tales] version of P[rimitive] E[lvish] from his language mythos and started Etymologies with a completely new PE language with different phonology" (Vinyar Tengwar #30 p. 7; the BoLT version and the QL version are for all intents and purposes identical). In accordance with the general policy of his group, Patrick Wynne in a response did his best to downplay the substantial revisions Tolkien's ideas underwent in the period from 1915 and until Etym was written in the mid-thirties. While he was right in pointing out that quite a few stems and morphemes were carried over into Tolkien's later and more mature concepts, this constitutes nothing like "strong evidence against Anthony's assertion that Tolkien rejected the phonology and system of roots in QL and rebuilt them from the ground up in the Etymologies", as Wynne would like to believe: Appleyard hardly meant to say that there was no continuity whatsoever in Tolkien's evolution of Elvish, but only made the quite obvious observation that "Qenya" and Quenya diachronically speaking do not even belong to the same language family.

The proto-language envisioned in the early Lexicons seems strongly inspired by the proposed reconstructions of Proto-Indo-European; in particular one is struck by the great number of syllabic consonants. In the late thirties, when the Etymologies was written, Tolkien had reached a new vision of Primitive Elvish, as is evident both from the bases themselves and from the actual "reconstructed" words quoted as the origin of words in various later Elvish tongues. This new proto-language, certainly more original (and artistically speaking therefore also more attractive) than the earlier experiments, underlies the kind of Quenya and Sindarin we know from Tolkien's world-famous narratives. Since the diachronic aspect was very important to Tolkien - in LR:342 his son points out that he imagined language "as growth, in time" - one can say that the similarity between the "Qenya" of the Lexicon and LotR-style Quenya is in one very real sense superficial only.

Speaking about his father's coining of individual words, Christopher Tolkien notes that "in some cases it seems clear that the word was 'there', so to speak, but its etymology remained to be certainly defined, and not vice versa" (LT1:246). The same principle may be applied to the entire language. The general style of Q(u)enya was "there" from the beginning; Tolkien knew roughly what he wanted the flavor of this language to be like, and he never discarded many of the words that originally appeared in this first Lexicon of 1915. But the diachronic history "remained to be certainly defined". Twenty years of experimenting would go into developing a satisfying system. Only with the writing of the Etymologies did Tolkien apparently feel that he had got it more or less right; the evidence is that for the rest of his life, he largely stuck to the ideas that manifested there.

Given the comprehensive revisions of the diachronic history, it is not surprising to notice that the precise etymology of individual words could undergo great changes as well, even if the word as such was retained. Two examples will suffice: 1) In Tolkien's later conception, the word Vala basically refers to a "Power". Hints to this effect are found already in the Etymologies, in which the stem BEL "strong" is tentatively compared to BAL (LR:352), and Vala is glossed "Power" as well as "God" (LR:350). In out best late source, Tolkien made explicit references to the "power" etymology (WJ:404). But in the earliest concept as recorded in the Qenya Lexicon, the Valar were "The happy folk", their name being connected to such words as valin "happy" and vald- "happiness" (p. 99). Later, there is no hint that a Vala was etymologically a "happy one". 2) In the Etymologies, the Quenya word for "book" (parma) is derived from a stem PAR "compose, put together", identifying a book as a "thing that is composed", evidently referring to the efforts of the author. The word parma appeared already in the QL, p. 72, but here it meant primarily "skin, bark" and hence "parchment" - "book, writings" being only secondary, poetic meanings. There is not yet any hint of a connection with a word for "compose"; indeed no word with this meaning appears in the QL at all.

To a person interested in writing in LotR-style Quenya, perusing the Qenya Lexicon is somewhat frustrating. There is a feeling of "well, what in the world are we to think of this?" For here we have a very substantial vocabulary, sometimes filling gaps in the known Quenya vocabulary that have long annoyed us, and yet we feel uncertain about using these words in LotR-style Quenya: Do they belong there at all, or would we be diluting Tolkien's more mature forms of Quenya with foreign elements? Elderly Tolkien's disdainful attitude towards the earliest stage of "Qenya" cannot well be ignored: he called it "very primitive" (PM:379). So asserting that the early Lexicons are in all respects great stuff would reflect a rather misunderstood loyalty to Tolkien. Those who want to develop a useable, LotR-compatible form of Quenya will have to approach the Qenya Lexicon with caution. Mindlessly mixing "Qenya" and Quenya together would produce a hybrid language that does not correctly represent Tolkien's intentions at any of the many stages his conception went through over the decades. To protect the integrity of Tolkien's later system we must in each case make sure that a word taken from the Lexicon fits the phonological, morphological and grammatical structure of LotR-style Quenya. If not, we must either ignore the "Qenya" word or subtly alter it to make it fit. We must also make sure that the "Qenya" word in question does not clash with later words. Some examples will be useful.

When nelde in the Qenya Lexicon (p. 65) is the word for "four", but the same word reappears with the meaning "three" in later Quenya (LR:376 s.v. NEL), we must obviously opt for the later meaning and ignore the QL. Other cases are more complex. A word like rakta- "stretch out" (p. 78) cannot be used in LotR-style Quenya in this form, for Quenya as opposed to "Qenya" does not have the cluster kt. At some point after writing the QL, Tolkien decided that kt actually became ht in Quenya; hence a word like ektele "fountain" (p. 35) reappears as ehtele in the Etymologies (LR:363 s.v. KEL). In light of this example, we should have to alter rakta- "stretch out" to *rahta- if we were to use it in LotR-style Quenya. (Other features of "Qenya" phonology that were later rejected is that "Qenya" unlike Quenya permits final k, as in auk "fool", or that "Qenya" possesses what is evidently final syllabic consonants, like l in findl "lock of hair". - Pp. 33, 38.)

On the other hand, a word like vára "other" (p. 100) must be rejected despite the fact that it fits the phonology of later Quenya: It would have filled a gap in our vocabulary, but it unfortunately clashes with later vára "soiled, dirty" (LR:397). Other words we can easily afford to ignore, since we have later words for the same thing: as the conjunction "and" the Lexicon (p. 104) has ya, but ar is well attested in later Quenya (and ya later became the relative pronoun "which" - apparently so already in the "Arctic" sentence in The Father Christmas Letters). Inevitably rejecting some words, we are however left with the question of whether we can then accept related words. We have no late word for "or" (in desperation, people have been trying to reconstruct one from Sindarin egor, leading to such suggestions as *ecar; a more likely reconstruction may be *ercë). Now the Qenya Lexicon provides var for "or" (p. 100). Writers have also missed a word for "also"; the QL has yando (p. 104). The words var and yando are often used on the Quenya mailing list, where we try (and largely succeed!) to keep conversation going in Quenya. I use them myself. Yet attentive readers will have recognized the page references: Var "or" belongs to the same group of words as vára "other", and yando "also" is listed right below ya "and". Rejecting, as we have to do, ya and vára as valid words in LotR-style Quenya, can we still feel free to accept the related words yando and var just because they conveniently fill gaps in our vocabulary and because these word-forms as such do not conflict with later words? (In all likelihood, quite different words for "also" and "or" are found in later, still unpublished material - perhaps a word for "also" related to ar as the word for "and", and a word for "or" that is indeed cognate with Sindarin egor.) There are no easy answers to these questions; we can only hope that material closer to the LotR-period will eventually be published.

Luckily, the QL also provides useful words that do confirm with later Quenya while not clashing with later words, e.g. net- pa.t. nente as the word for "get" (p. 66). It should also be noted that the Lexicon contains some words that are not found in the Etymologies of the mid-thirties, but turn up in later writings, indicating that they never really left Tolkien's mind. One example is koirea "alive, lively" (p. 48); our next attestation of this word is in a document that was written more than half a century later: In one of his essays, elderly Tolkien used the phrase coirëa quenya, "living speech" (PM:399). This is somewhat encouraging to scholars who are wondering whether or not words from the Lexicon - found there but not in such later sources as the Etymologies - can still be accepted as valid in LotR-style Quenya.

Having read this far, the reader will rightly have inferred that this student is most interested in Quenya as an actual, potentially useable language, and that I give much thought to the practical aspects of establishing a LotR-compatible standard. But if we stop trying to adapt the language of the Lexicon to later Quenya and look at the QL from a more academic angle instead, what do we learn about young Tolkien and his inner and outer world - not to mention his invented world?

Several words are rather surprising and can hardly be fitted into the world of Middle-earth (though there can be no doubt that Tolkien had already decided that this was the setting where Qenya primarily "belonged"). Qenya names for various primary-world countries and regions are provided: Warwickshire, England, Ireland, Germany, Norway/Scandinavia, Oxford (pp. 29, 42, 43, 44, 74, 89). Quite a few words reflect Tolkien's religious beliefs. The word Atar "father", also known from LotR-style Quenya, was already in place - but in the QL, a note specifies: "Usually to 1st Person of the Blessed Trinity" (p. 33). The word Ion (compare yondo "son") is here defined as a "mystic name of God", the "2nd Person of Blessed Trinity" (p. 43). As far as the Trinity is concerned, this leaves only the Holy Ghost, of which the word "Fire" is said to be a "mystic name" (p. 81). Other words with specifically Christian reference can also be found: On page 36 we have evandl "Christian missionary" and evandilyon "gospel" (Qenya adaptation of New Testament Greek euangelion?), and there even turns up a word for "crucifixion" or "crucifix" (anatarwesta, pp. 31, 89). Even a person ignorant of Tolkien's religion could probably have narrowed it further down to Catholicism when the author carefully includes words for "saint" (masc. aimo, fem. aire), "monk" (anustar or anuon), "nun" (qinde or qinne), "monastery" (anusta) and "convent" (qindesta) (pp. 34, 31, 77).

In the background, we also hear the guns of World War I. The Great War had already been raging for a year when Tolkien, himself a soldier at the time, wrote the Lexicon. On p. 94, we find the onomatopoetic tompo-tompo, ominously glossed "noise of drums (or guns)". Patakatapaka (or pataktatapakta) is simply defined "rat-a-tat" (p. 72), but we need not doubt that it is the sound of machine guns we are hearing. Young Tolkien inevitably did not think very highly about the enemy. The following word-group (p. 44) is a testimony to that:

kalimbo (o) a savage, uncivilized man, barbarian. - giant, monster, troll.
kalimban (n-) "Barbary", Germany.
kalimbardi the Germans.
kalimbarie barbarity.

But to leave the influences from our own world behind, much of the invented world can already be seen to have come into place. Many well-known names, never changed throughout Tolkien's long life, are already present: Manwë and Varda, Aulë and Yavanna, Ulmo, Túrin Turambar, Valinor, Tol Eressëa and many others. God's title Ilúvatar is also in place (p. 42), but here it is interpreted "Heavenly Father", not as later "All-Father". In the QL, the divine "name" is given as Enu, in sound clearly the predecessor of later Eru, but the meaning seems to be *"Creator" rather than "The One": It is hinted that Enu is related to a verb enye "devise"; see p. 35. Interestingly, on the next page we find the stem ERE "remain alone", so the ingredients for the later name Eru "The One" are already in place. It seems that Tolkien first intended Noldo to mean "goblin", but the classical meaning of the word swiftly arose (p. 67). Such strange shifts in Tolkien's conception are no surprise to serious students of his languages; compare the very first entry in the Etymologies, which shows that Tolkien at first intended Avari as the name of the Elves who did go to Valinor, then changed his mind 180 degrees and instead made it the name of those who refused the summons of the Valar (LR:347-348).

There are, however, also traces of more primitive concepts, like the idea of Elves or "fairies" as small, pretty creatures dwelling in flowers; Ailinóne is the name of "a fairy who dwelt in a lily on a pool", while Tetille is "a fairy who lived in a poppy" (pp. 29, 92). This seems to reflect precisely such a popular notion of "Elves" as the childish ideas Tolkien later scoffed at (in LotR Appendix F he laments the fact that the word Elves "may now suggest fancies either pretty or silly, as unlike to the Quendi of old as are butterflies to the swift falcon - not that any of the Quendi ever possessed wings"!)

Again turning our attention to the language as such, we may ask whether the Lexicon gives a reasonably complete account of this language. After all, the well over three thousand words here provided exceed the number of words known from later Quenya (about 2,200). So can we now express ourselves relatively freely in "Qenya", or at least as well as we can in LotR-style Quenya? The answer to that has to be no. What we have is a vast mass of isolated words, but little information on how to inflect them or combine them into actual text. Just about the only real sentence quoted is found on page 73, perilme metto aimaktur perperienta, "we indeed endure things but the martyrs endured and to the end". For a substantial "Qenya" text more or less contemporaneous with the Lexicon we must rather turn to the poem Narqelion; with the help of the QL, Christopher Gilson was able to plausibly interpret most this very early poem (the substantial content of Vinyar Tengwar #40).

As for what little grammar can be inferred from the Lexicon itself, the information is in some respects contradictory. What is already the classical example involves the pronominal endings -r and -n. The group enin (or emin), emil, emir on page 35 is glossed "I (etc.) am called". The "etc." following the pronoun "I" would seem to indicate that the other forms would be translated with other pronouns, so that if emin or assimilated enin = "I am called", emil is probably 2nd person *"you are called", while emir would be the 3rd person *"he is called" (possibly covering *"she is called" as well). So the pronominal endings would seem to be -n = "I" (as is still the case in LotR-style Quenya), -l "you" (cf. later Quenya -lyë and the short ending -l mentioned in WJ:364) and -r "he" (compare -ro in antaváro "he will give" in LR:63). There are many examples besides enin to indicate that -n is the 1st person ending, such as aqin "I seize", elin "I drive", fengin "I cut", hatin "I fling", hotin "I sneeze", iltin "I trust home", iqin "I pray", kakin "I laugh", kapin "I jump" (pp. 31, 35, 38, 39, 41, 42, 43, 44, 45). Yet in one section of the Lexicon the glosses provided would seem to indicate that it is not -n, but -r that means "I"; we have for instance mokir "I hate" or wastar "I dwell" (pp. 62, 102). The ending -n is suddenly reassigned to the third person and is translated "he" or "it" (e.g. avin "he departs", hilkin "it freezes", usin "he escapes", vildin "it matters", vilkin "it cuts" - pp. 33, 39, 98, 102, 101). This strange reversal of the first and third person endings occurs almost exclusively in only one end of the Lexicon, so there is little reason to doubt that this is a genuine revision undertaken while the QL was written, leaving the document internally contradictory because Tolkien only sporadically went back to change what he had written earlier. While Christopher Gilson on the TolkLang list tried to argue that the endings -n "I / he" and -r "he / I" could indeed coexist within the same variant of "Qenya", the consensus seems to be that maintaining such a position requires theories so elaborate that they are transparently overwrought. In later Quenya, -n is in any case the 1st person ending once again; notice that while both the Lexicon and the Etymologies list the negative verb umin (QL:98, LR:396), the meaning was changed from "it does not" in the QL to "I do not" in Etym.

On one point only does the Qenya Lexicon provide extensive information about a grammatical feature: the formation of the past tense of verbs. In well over 300 cases the past tense is listed alongside the more basic form of the verb. We find all the types we are familiar with from later Quenya, such as past tenses formed with the ending -ne (e.g. sesta- "compare", pa.t. sestane), with nasal infixion (e.g. kap- "jump", pa.t. kampe) or by lengthening the stem-vowel and adding -e (e.g. mel- "to love", pa.t. méle; see pp. 82, 45, 60). But there is also a great number of highly exotic formations, such as the ending -ya turning into past tense -sine or -tine (e.g. mauya- "to cry", pa.t. mausine; panya- "arrange", pa.t. pantine, pp. 60, 72), or even past tenses involving internal-vowel shifts (like milk- "have, keep, possess", pa.t. malke, or tump- "build", pa.t. tampe - pp. 62, 93). The "Qenya" past tenses should be subjected to a thorough study. True, some of the information provided clearly applies to "Qenya" only, presupposing its own peculiar phonological history; for instance, vowel-shifts in the past tense occur where we have stems involving syllabic consonants - the forms milk- > malke and tump- > tampe come from stems MLKL and TMPM. Such stems are no longer possible in Tolkien's later vision of Primitive Elvish, so in later Quenya, shifts like milk- > pa.t. malke would not be possible either (there is nothing to parallel this in the Etymologies). Yet in some cases the Qenya Lexicon may provide clues to mysteries in Tolkien's later Elvish. For instance, Gilraen's Sindarin linnod in LotR Appendix A has onen for "I gave". This must be assumed to be the 1st person past tense of the "Noldorin"/Sindarin verb anno "to give", listed in LR:348 s.v. ANA1 as the cognate of Quenya anta-. But how do we get from anno "to give" to onen "I gave"? The QL (p. 31) likewise lists the verb anta- "give", one of many words destined to survive into later Quenya, but the QL also provides the past tense áne, a form not listed in the Etymologies but not necessarily conflicting with this later stage of Quenya - notice that in Etym itself, the verb onta- "beget, create" has the past tense óne besides more regular ontane (LR:379 s.v. ONO). If, in the common ancestral language of Quenya and Sindarin, the verb anta- "give" had the past tense áne "gave", whence *ánen- "I gave", we only have to apply the normal sound-shifts to get straight from archaic *ánen- to Sindarin onen - precisely the form found in LotR. In light of this, writers may consider áne as a perfectly valid alternative to *antane as the past tense of the verb anta- "give" also in LotR-style Quenya.

In summary, what are we to say? Is the Qenya Lexicon worth acquiring? It is of course of the utmost interest if you want to study Tolkien's language-making as such, the Qenya and Gnomish Lexicons of 1915-17 standing out as landmarks near the beginning of a long road. However, I find it regrettable that this very early material appears at a time when our knowledge of the languages in their more mature incarnations is still so full of holes. If we had already known Tolkien's intentions at the LotR-stage more fully, we could have enjoyed the Qenya Lexicon for what it is: the first draft for a system of great subtlety and beauty, to be further refined and beautified in the decades to come. But at present, when researchers very justifiably divert most of their energies into understanding the more mature stages of the languages based on the scanty evidence available, what could have been interesting studies of "Qenya" in its own right almost inevitably drown in one overriding concern: "Is this information valid for LotR-style Quenya?" We must hope that the Qenya and Gnomish Lexicons have not been completely forgotten when Tolkien's later writings on Elvish are finally made available to scholarship; one fears that the Lexicons will spend years gathering dust on many a shelve before comprehensive comparative studies of Qenya vs. Quenya at last become possible.

Index to the Qenya Lexicon by English glosses
Index to the Qenya Lexicon by Elvish words

Ardalambion Index