The verb 'to be' in Tolkien's Elvish languages

Abbreviations used:
Q: Quenya
S: Sindarin
N: Noldorin
CE: Common Eldarin

PE: Parma Eldalamberon
VT: Vinyar Tengwar
LR: The Lost Road and other writings
WJ: The War of the Jewels
UT: Unfinished Tales
LOTR: The Lord of the Rings

The verb 'to be' is an integral and often used part of languages, yet in the grammatical writings by Tolkien about his languages it doesn't seem to have featured very prominently. Almost nothing definite is known about it in Sindarin, some forms are known in Quenya, and only in early Qenya we have a complete table of inflected forms. And for all we know, the verb has undergone a series of changes from Tolkien's early Goldogrin and Qenya to later Sindarin and Quenya. This essay aims to collect what we know about the verb from the various stages of language development.

1. Logical considerations

The verb 'to be' has a number of functions in English, and some grammatical features change when the function of the verb changes. For the sake of the argument, assume that A is a noun or pronoun and B may be another noun, an adjective or any phrase which can be joined with A using 'to be'.

First of all, it can be used to express a (mathematical) identity A=B. If used in this sense, the sentence 'A is B' can be altered into 'B is A' without a significant alteration in meaning, 'Aragorn is Elessar' contains the same information as 'Elessar is Aragorn'.

Second, it can be used to denote that A belongs to class B. In this sense, it cannot easily be rearranged into 'B is A' - 'A mouse is an animal.' makes prefect sense whereas 'An animal is a mouse.' doesn't. (However, in spoken language and with strong emphasis, a sentence like 'An ANIMAL is a mouse, not a plant.' is probably understood.)

Furthermore, it can be used to join predicative adjectives to a noun, e.g. 'Arwen is beautiful.' The word order can be altered here, however then emphasis is places on the adjective 'Beautiful is Arwen (not ugly).'

Prepositional phrases can also be joined - as in the case of adjectives, we observe an emphasis if the word order is inverted: 'Elrond is in Imladris.' vs. 'In Imladris is Elrond (not in Gondor).'

Finally, in English it can be used with verbal participles to form various auxiliary tenses, for example with the present participle the present progressive tense 'Aragorn is fighting.'. Again, an inversion of the word order is possible but strongly shifts emphasis: 'Fighting is Aragorn (not resting).'

At this point, one should keep in mind that the verb does not have to be in the same position as in English, or that it may be left out at all, cf. the Latin 'Caius asinus est' 'Caius is an a**' but 'In vino veritas' 'In wine is truth'.

In the following, we aim to investigate to what degree we observe the same constructions in Elvish and where we see differences from the English use.

2. Evidence from the different languages


We find a few sentences involving 'to be' in the corpus of Goldogrin sentences.

i·vrog na cuid arog 'the horse is a swift animal' (PE11:9)
i·weg na an fofrin 'man is a foolish creature' (PE11:9)
i·weg fof '*man is a fool' (PE11:9)
i·winin na gwandron 'women are beautiful' (PE11:9)

In addition, the verb 'to be' is glossed as na in the GL, its plural is nain, the past tense is given as thi (presumably the plural is *thin rather than ?thith) and the participle is ol·. A remark tells us that the verb is 'quite irregular', so it may not be permissible to generalize to other verb forms.

We see the verb used to describe one of a class and to join a predicative adjective. Although a verb exists, as i·weg fof indicates it is permissible to leave it out and still get a valid sentence. In all examples both the order of A and B and of the verb 'to be' is just as it is in English.

In addition, the following examples

im len 'I have come, I am come.' (PE11:53)
um lenin '*We have come.' (PE11:53)

apparently indicate that 'to be' (here left out) can be used with verbal adjectives like len 'come, arrived' to form auxiliary tenses (this assumes that forms like len or lenthol (approaching) are indeed participles in Goldogrin and not just adjectives derived from the same root - the kinship of the ending ol with the participle ol· of 'to be' makes this seem likely, though).

In addition we find a negative verb û 'not to be, not to do' with plural uin and past tense ûthi, ûthin.

Early Qenya

A full paradigm of the conjugation of the verb 'to be' is found in PE14:57. We have in present tense: sg. e pl. il, in past tense sg. ie/ye pl. iel and future sg. va pl. val. In addition, early Qenya has an impersonal form present ir past tense ier and future var, a reflexive form present ikto past tense *iekto and future vakto, a verbal noun present ista past tense iesta, yesta and a future vasta and a participle present tense in(d-), inda past tense ien, yen, yenda and future tense van(da).

They can be used together with the participles of other verbs to form compound tenses, e.g. using tulien(d-) 'having come' as e tulien 'is having come > has come' or tuliende 'has come' where the compound is actually finished. We do not see the verb in any other use.

Late Qenya of 'The lost road'

Most of the evidence can be found in Fíriel's Song (LR:72)

númessier 'they are in the West'
toi aina, mána, meldielto 'They are holy, blessed, and beloved'
talantie 'he is fallen'
toi írimar 'which are beautiful'
Ilu vanya 'the World is fair'
írima ye Númenor 'lovely is Númenor'
sí ye tyelma 'here is ending'
yéva tyel 'there will be an end'
íre ilqa yéva nótina 'when all is counted'
ananta úva táre fárea 'but yet it will not be enough'

In addition we have

Man-ie, atto? 'What is it, father?' (LR:59)

The use of 'to be' in late Qenya agrees partially with what we have attested in Early Qenya - but phonetical elements have changed their role. We see a verb ye, future yéva and a negative verb úva 'will not be'. Several examples indicate that the verb can be left out (at least in present tense).

The form ye can also merge with a preceding word, in the process changing to -ie. So *númesse 'in the west' númesse + yer > númessier 'they are in the west'. As this example, the translation of Man-ie 'what is it' and talantie 'he is fallen' indicate, the same forms are used for 3rd person 'he/she/it is' and personless use 'A is'.

The word order is as in English for the free form ye but (unless in a question) is inverted if the ending -ie is used instead.


Not much is known about the role of 'to be' in Noldorin. One bit of evidence is the entry NA2 in the Etymologies (LR:374) where we find:

NA2 (cf. ANA) be. Stem of the verb 'to be' in Q. Cf. nat thing, N nad

The entry is significant in two ways. First, we learn that in Quenya a transition from the e/ie/ye based forms to the stem na (which has been known from Goldogrin only so far) has taken place. Second, the entry points out specifically that this is the stem of 'to be' in Q(uenya) which suggests that this may not be the relevant stem for 'to be' in Noldorin.

There is one significant sentence found in 'Tolkien - Artist and Illustrator' which seems to be in the Noldorin of the Etymologies:

Lheben teil brann i annon ar neledh neledhi gar godrebh '*five feet high [is] the door and three may walk abreast'

Here the verb 'to be' is ommitted, but it is possible to ommit it from the English translation as well.


There is a plethora of examples to be found for the use of 'to be' in later Quenya texts:

ar i Eru i or ilyë mahalmar eä tennoio 'and of the One who is above all thrones for ever' (UT:305)

Mana i·coimas Eldaron 'What is the 'coimas' of the Eldar?' (PM:396)

Sí vanwa ná, Rómello vanwa, Valimar! 'Now lost, lost to those of the East is Valimar!' (LOTR)
Nai hiruvalyë Valimar! 'Maybe thou shalt find Valimar!' (LOTR)
Nai elyë hiruva! 'Maybe even thou shalt find it!' (LOTR)

Various examples can be found in the 6 different versions of the Ataremma in VT43:

In Ataremma I:

ataremma menelzea na 'our father who is in heaven' (VT43:8)
na aire esselya 'holy be thy name' (VT43:8)
nâ túle turinastalya 'thy kingdom come' (VT43:8)
na carina mendelya 'thy will be done' (VT43:8)

The Ataremma II has a slightly different wording:

A Ataremma i menelzea 'our father who is in heaven' (VT43:9)

Ataremma IIb has yet another rephrasing

Ai Ataremma i meneldea 'our father who is in heaven' (VT43:10)
esselya na aire 'holy be thy name' (VT43:10)

The Ataremma III brings the new line

san cemende ya menelde na 'on earth as it is in heaven' (VT43:11)

Ataremma V has

aranielya na tuluva 'thy kingdom come' (VT43:12)
na carima indómelya 'thy will be done' (VT43:12)

and finally in Ataremma VI we find

Átaremma i ëa han ëa 'our father who is in heaven' (Vt43:12)
na care indómelya 'thy will be done' (VT43:12)

An interesting note in VT43:14 reveals Tolkien's ideas about the difference between imperative and present tense:
na preced. = 'let it be': na aire 'be holy' aire na 'is holy'. The note is found below Ataremma IIb and holds for all following versions.

In the Aia María I we find:

na héru olesse 'the Lord is with you' (VT43:26)
elye na manna mi nínaron 'you are blessed among the women' (VT43:26)
ar manna i yáva carvalyo Yésus 'and blessed is the fruit of thy womb' (VT43:26)

The Aia María II has the lines as:

i Héru olesse 'the Lord is with you' (VT43:27)
manna nalye mi nínaron 'you are blessed among the women' (VT43:27)

Likewise, the Aia María III has

i Héru as elye 'the Lord is with you' (VT43:27)
aistana elye imíca nísi 'you are blessed among the women' (VT43:27)
ar aistana i yáve mónalyo Yésus 'and blessed is the fruit of thy womb Jesus' (VT43:27)

In the Alcar i ataren we have

Alcar i ataren ar i yondon ar i airefean tambe enge i et 'Glory [be] to the Father, and to the Son, and to the Holy Spirit as it was [in] the [beginning]' (VT43:36)

A few more examples can be found in 'Negations in Quenya', VT42:32-34:

lá karita i hamil mára alasaila (ná) 'not to do what you judge good (would be) unwise'
lá karitas, navin, alasaila ná '*not doing it, I deem, is unwise'
lá karitas alasaila ké nauva 'not doing this may be unwise'
alasaila ná lá kare tai mo nave mára 'it is unwise not to do what one judges good'

The bulk of Tolkien's late ideas known so far with regard to 'to be' however can be found in VT49. VT49:13 has the conjugation of a negative verb with Aorist #la (the form lanye '*I am not' is given), present tense laia, past tense lāne, perfect alaie,future lauva and imperative ala, alā. However, the idea that should form a verb which is inflected for tense was subsequently abandoned.

One paradigm of the verb to be is found in VT49:27. Here we find Aorist na- with endings -in, hence nain 'I am', present tense with nánye 'I am being', past tense nāne, perfect tense anaie and future tense nauva. However, a past tense is given in VT49:28 which in the presence of endings becomes ane- such as anen 'I was'. A second paradigm of is found in VT49:30 which has as both Aorist and present tense with a shortening when endings are appended such as nanye 'I am', a past tense and an irregular future uva. No other tenses are given.

VT49:29 has a paradigm of the verb ea 'to exist', where we find ea as both Aorist and present tense, a past tense enge, perfect tense éye or the analogical form engie and future tense euva.

Most of these ideas are exemplified in variants of the Ambidexter sentence. One finds:

i Eldar nāner ataformor 'Elves were ambidexters' (VT49:6) and the later variant
Eldar ataformaiti 'Elves are/were ambidexters' (VT49:7)

from which it becomes clear that the copula can be left out in past tense as well. There are also several variants using the negative verb:

hyarmen láne 'sinister' '*left was not sinister' (VT49:6)
hyarmen aune 'sinister' '*left was not sinister' (VT49:6)
hyarmen láne ulca '*left was not evil' (VT49:6)

Let us review all the evidence. First of all, we observe that there are two distinct verbs in use to express 'to be', one based on na/ná/nâ and the other ëa.

Let us first focus on na. Almost consistently we observe the use of the short na in the various versions of the Ataremma whereas the long is found elsewhere - this seems to indicate Tolkien's uncertainty if the form should be written with a long vowel rather than a grammatical significance. The difference between na and for sure is not used to mark present tense and Aorist (cf. the lengthening of the vowel when comparing Aorist quetë 'speaks' (VT41:11) and present tense quéta 'is speaking' (VT41:13)) - the sentence alasaila ná lá kare tai mo nave mára is explicitly called Aorist in VT43:34.

The form is only found in the earliest Ataremma and never shows up later, thus it can be probably dismissed as being a (later rejected) try.

If na/ná is used at all seems to be optional - we find e.g. lá karita i hamil mára alasaila (ná) or the contrast of ataremma menelzea na and A Ataremma i menelzea - we find no form of 'to be' in all three slightly different versions of ar aistana i yáve mónalyo Yésus. In VT47:30, Tolkien's late view on the usage of the copula can be found: As a copula 'be, is' is not usually expressed in Quenya where the meaning is clear: sc. in such expressions as 'A is good' where the adjective (contrary to the usual order in Quenya for a qualifying adjective) follows: the normal Quenya for this is A mára. But when the subject is not expressed, as usual in the impersonal 'neuter', e.g. 'it is good' = that is good, very well, is used: so mára ná, also when it is postponed as in Galadriel's Lament, Sí vanwa ná, Romello vanwa, Valimar.

As for the word order, Tolkien's note regarding na aire 'be holy' aire na 'is holy' seem to indicate a turning point - before the note, we find na with the meaning 'is' at the beginning of a sentence, cf. na héru olesse in Aia María I, but after the note, na in front of an adjective (or in fact a verb) consistently denotes imperative (or optative), cf. na care indómelya lit. '*be that [someone] does thy will' or aranielya na tuluva lit. '*be [that] thy kingdom will come'. This seems to agree nicely with the use of nai in e.g. Nai hiruvalyë Valimar! lit. '*be that you shall find Valinor', and in fact nai is decomposed into na + i (be [it] that) in RGEO:68. The only possible flaw in this theory would be the A ná kalima lá B formula quoted in VT42:32 where the form of 'to be' occurs in fact before the adjective - it is however impossible to say from the context if the document referred here might not in fact be earlier than the first Ataremma.

In fact, even in imperative use na apparently can be left out as alcar i ataren... 'glory [be] to the father...' indicates.

Since na (and possibly ) in front of an adjective or verb signify imperative/optative, with the meaning 'is' consequently has a tendency to move towards the end of a phrase, quite contrary to the standard English word order. This is apparent from a number of different texts, cf. Sí vanwa ná 'now lost is', san cemende ya menelde na 'on earth as it is in heaven', lá karitas (...) alasaila ná 'not to do it (...) is unwise'. The only real exception is alasaila ná lá kare... 'unwise is not to do...' where is only moved behind the adjective - however, that is the end of a subphrase, and moving all the way to the end of the rather lengthy following phrase presumably destroys the coherence of the whole sentence.

The form seen in Mana i·coimas Eldaron (PM:396) could be a second example in which a form of na is not moved towards the rear of the sentence - it reminds of the earlier Man-ie? 'What is it?' (LR:59) and could be understood as a compound man + ná/na > mana.

We see the verb in use with the predicative adjective in e.g. alasaila ná 'unwise is' and connecting with a prepositional phrase in e.g. ataremma menelzea na. We don't see it denoting 'one of a class' or in any auxiliary tenses. Note that the explicit distinction between Aorist and present tense in Quenya creates no real need to create an auxiliary present progressive tense using a present active participle and 'to be' analoguous to the English 'is doing'.

The form ëa occurs more rarely. It two times found occuring in the connection with Eru (God) which may carry some significance. In VT39:7 ëa is translated as 'exists' rather than 'is' which points to the relevant difference between the two words.Presumably it can only be used in sentences in which 'is' can be replaced by 'exists'. Furthermore, it may be guessed from the example tëa 'indicates' past tense tenge (VT39:7) that the past tense of ëa likewise would be enge.This agrees well with the form found in tambe enge 'as it was' (VT43:36) and the paradigm in VT49:29.


There are very few clear examples of the use of a verb 'to be' in Sindarin. The best examples are:

Ae Adar nín i vi Menel '*Oh my father who is in heaven' (VT44:21-22)
no aer i eneth lín 'holy be thy name' (VT44:21-22)

Some writers have included the line

bo Ceven sui vi Menel 'on earth as (it is?) in heaven' (VT44:21-22)

into the analysis, assuming that the 'it is' in the bracket needs to be added, however there is no real need to do so, a translation along 'on earth as in heaven' seems certainly fine. Furthermore one may be tempted to include

Naur an edraith ammen!'Fire be for saving of us' (LOTR)

where the imperative 'be!' seems to be unwritten in Sindarin, but Tolkien may not have provided a literal translation, cf. the structurally similar sentence

Gurth an Glamhoth! 'death to [the] din-horde' (UT:39)

where there's no need to assume a missing imperative 'be!' from the given translation.

Thus, for the most part we find that the verb is left out, its only trace is a form no which seems to be used in a similar construction as Quenya na, cf. the similarity S: no aer Q: na aire 'be holy'. Thus, it may be that the form no is a fossilized optative particle rather than the imperative of an inflected verb 'to be' - it is impossible to know from this one example.

3. Remarks on the development

Considering the whole evidence from all languages, we can identify some trends. First of all, the dominant root for the verb to be, NA, goes all the way back to Goldogrin, but it was not considered relevant for Quenya till the writing of the Etymologies when it replaced an earlier form e/ie/ye. This form is very likely obsolete in later Quenya - not only is it not found in later sources, but the form as an ending -ie also clashes with the infinitive-gerund ending -ie seen in enyalien 'for recalling' (UT:305) and with the perfect which likewise is characterized by the same ending, vowel lengthening and an augment, cf utúlien 'I am come' (LOTR).

The form utúlien, in contrast to the Goldogrin im len with the same meaning may also provide some indication regarding the fate of the auxiliary tenses. Quite likely they have vanished in the later development of the languages, being replaced by genuine forms for present tense and perfect (the situation in Sindarin is less clear, however).

This agrees well with the observation that the use of 'to be' seems to become progressively different from the use in English as the languages develop: While we see a very literal translation of the English wording preserving the word order in the Goldogrin i·winin na gwandron, the same word order in Quenya could not be maintained - the similarly ordered esselya na aire doesn't mean 'thy name is holy' but takes imperative meaning due to the word order 'holy be thy name'.

One main idea seems to be preserved through the whole language evolution, though. The verb 'to be' can always be left out, and we have examples showing this phenomenon from Goldogrin to Noldorin, Qenya, Sindarin and Quenya. We see this omission only in present tense, though, not much is known if it is permissible in past tense or future (then again, we rarely know all three forms in a given language).

In Quenya we finally see the emergence of a second verb ëa making statements about existence. It is a reasonable guess that Sindarin would have a similar verb - unfortunately we don't know it.

The attested inflected forms differ in their amount of regularity - while Quenya present na future nauva is not unexpected, Early Qenya present e, past ie and future va looks slightly unexpected. Qenya present ye future yéva looks regular in the light of antáva 'will give' (LR:72), presumably future of anta-, but usually the past tense forms show most irregularity in Tolkien's languages. Clearly Goldogrin na, past thi and participle ol is highly irregular.

4. Word order and emphasis

As we have observed above, in many situations changing 'A is B' into 'B is A' leads to a pronounced shift in emphasis in English. In order to investigate if this idea is relevant for the Elvish languages as well, we note that in all early languages (Goldogrin, Qenya and Noldorin) the word order is in all cases close to the English wording, thus presumably the same emphasis is intended. The question is more interesting in later Quenya where the word order cannot be the same as in English for the reasons mentioned above.

Thus, the phrase 'Valinor is lost' can only be rendered in Quenya as either vanwa na Valimar or Valimar vanwa na - but not as **Valimar na vanwa, because the last phrase would mean 'lost be Valinor!'. Thus, the question is if we have reason to assume that vanwa na Valimar carries more emphasis than Valimar vanwa na.

It turns out that we have ample reason to think so. Whenever the word order is 'B na A' in Quenya we have 'B is A' in English, whenever we see 'A is B' in English we find 'A B na' in Quenya, cf. the emphasis in Sí vanwa ná (...) Valimar! 'Now lost is Valinor!' (including an exclamation mark for emphasis), manna nalye mi nínaron where we have the same emphasis in English 'Blessed art thou amongst women' or Latin 'benedicta tu in mulieribus' or alasaila ná lá kare... 'it is unwise not to do...'.

Given the overall consistency of English and Elvish word order with regard to emphasis, we may reasonably assume that an example like the Noldorin Lheben teil brann i annon... '*five feet high [is] the door' does not represent a word order in the absence of an explicit verb 'to be' - it rather seems to reflect the same emphasis on the actual height of the door that is seen in the English sentence found in 'The Hobbit'.

5. Lenition as a marker in Sindarin?

In Sindarin, lenition of the direct object of a verb is often observed, cf. Lasto beth lammen 'listen to the word of my tongue' (LOTR) or ennas aníra i aran...suilannad mhellyn în 'there the king greet his friends (SD:128). It is conceiveable that a similar mechanism would lead to lenition if the verb 'to be' is involved (even if it is actually left out)- in 'A is B', the head of phrase B might be lenited. Some writers have argued that this could be the reason why we see lenition in i vi Menel 'who [is] in heaven' (VT44:21-22) or sui vi Menel 'as [it is?] in heaven'.

We observe that the lenition of direct verb objects seems to be independent of the position of the verb in the sentence, Daur a Berhael (...) eglerio! 'Praise (...) Frodo and Sam!' apparently shows lenition (The unlenited form of *Taur is not actually attested anywhere, but something must cause lenition for Perhael, and as apparent from loth a galadh (Lays of Beleriand 354) or pedo mellon a minno (LOTR) it is not the conjunction a). Thus, if that is indeed the underlying mechanism, there is no need to assume that the inverted and emphasized 'B is A' would be unlenited.

Unfortunately, we cannot check the conjecture. The closest thing we have is the Noldorin sentence Lheben teil brann i annon - and here the head of the phrase lheben teil brann would not show lenition.

6. Final remarks

It is interesting to observe how Tolkien's ideas shifted from his early concepts which are in fact quite close to English to the more sophisticated later Quenya where the word order conveys meaning and we find two different words for 'to be' with slightly different meaning. Yet at the same time some ideas, like the possibility to omit 'to be' altogether remain remarkably stable throughout the whole evolution of the languages. Given the fact that the verb doesn't appear to be too regular in any language, be it invented or not, more than sophisticated guesswork how the missing forms in Sindarin should read (if there are any) is quite imossible at this stage.

Thorsten Renk

Back to the index