On the different forms of ’we’ in Eldarin

Roman Rausch

Oct. 15th 2008

’Be careful, friends!’ cried Gildor laughing. ’Speak no secrets! Here is a scholar in the Ancient Tongue.’
LotR I, ch. 3




A well-known feature of Quenya is the distinguishing of an inclusive and an exclusive form of ’we’. A major revision of the associated verbal endings was carried out in 1965, so that omentielmo ’of our meeting’ in Frodo’s greeting in The Lord of the Rings was replaced by omentielvo in the second edition.
Although lots of discussions concerning this change took place among scholars in the past, the whole picture only became known with the publication of verb paradigms involving dual forms (PE16 for Early Qenya in 2006 and VT49, PE17 for later Quenya in 2007). Without knowing this whole picture the 1965 change seemed to lack a motivation, looked almost like a random permutation of suffixes. This article therefore attempts to give a chronological overview and analysis of the different 1st person plural and dual forms (mainly for Quenya, with Sindarin and Telerin cognates discussed along the way) with a special focus on concepts, mechanisms and attempts to look into Tolkien’s motivations behind the encountered changes.

1  Early Qenya

The distinction between an inclusive and an exclusive ’we’ appears already in the Early Qenya Grammar. Verb declinations at that time are marked for person by prefixes and one finds:

EQG (around 1923) (PE14:52)

Some years later Tolkien wrote down a detailed verb chart where a new dual paradigm seem to appear for the first time. There are three versions of it and the first one contains the following 1st person pl. and dual forms:

late 1920s – early 1930s (PE16:124-127)

tul- ’come’

These forms can be analyzed by supposing that there are two roots for the 1st person endings: exclusive *me- and inclusive *we-. A further marking for tense involves an additional ending.
In the aorist pl. this suffix is -n, hence -men and -ven (where intervocalic *-w- > -v- for phonological reasons). The pres. pl. in the chart is generally formed by infixing of l which seems to be a plural marker, cf. 3rd sg. neutr. pres. tulī-ta *’it comes’, 3rd pl. neutr. pres. tuli-lta *’they (neutr.) come’. The 1st pl. form is an exception, however, and is apparently formed by nasal fortification *we > -ngwe (in full symmetry one would have expected *lwe or *lve which will be found later).

The associated dual forms show a vocalic dual marker *u, so the dual roots are *mu- and *wu-|*vu-. In final position *-u becomes -o. The aorist tense goes without a further suffix, but can be optionally formed by the marker -r. On the other hand, the present tense is formed by (nasal) fortification: excl. *mu > *mbu > -mbo and incl. *wu > *ngwu > -ngo. Because of the dual marker u|o the inclusive form can be distinguished from pl. incl. -ngwe. For the development m > mb compare Tolkien’s later remarks that mb, nd, ng appear as modifications of both m, n, ñ and b, d, g (VT39:9) by a process called antoryame (’strengthening’).
There are two additional tenses in the chart, past and perfect, but they do not involve any new endings. The past is formed by appending the aorist suffixes -mur, -men, -vur, -ven to the root #tulie-. The perfect tense is similarly arrived at by appending the present endings to #tulie-.
Version 2 of the chart brings slight variants into the described system, like allowed tulime, tulive beside tulimen, tuliven. Version 3 has tulimur, tulivur instead of [tulu]mur, [tulu]vur, but the pattern is otherwise the same.

2  The Lord of the Rings

A grammatical description from the 1940s (i.e. during the writing of LotR) involves a historical discussion of the 1st pl. roots. It is unpublished in full length, but a summary is given in PE17:14.

1940s (PE17:14)

The replaced inclusive form khe might have been intended to yield a form like earlier qe /kwe/ (possibly by a blending with we?).

Although 1st pl. and dual forms appeared in grammatical charts, they was no occasion to use them in the various texts that Tolkien wrote. But this changes by the time of The Lord of the Rings as the famous Quenya greeting appears. At first it takes the form:

The usage of an exclusive form here seems to be a slip (unless there was a temporary conception of me being inclusive), so that Tolkien wrote: but this should be inclusive form, omentienwa ? -lwa. The introduced inclusive possessive #-nwa with nasal strengthening can be related to earlier tuli-ngwe and -lwa is what one would expect with the plural marker l. This plural marker also appears in the verbal ending -lto ’they’ around the same time (cf. the separate pronoun toi) in Fíriel’s Song (LR:72, c.1937). The corrected greeting now reads:

Another note dated 1955 (thus shortly after the publication of LotR) reads:

Nov. 20, 1955

The inclusive root is changed here once again, one could guess in order to have a symmetrical pattern where both forms of ’we’ have different nasals (note that the 1st sg. pronoun is ni with a dental nasal). But also important is the Quenya invention lme where l is not a plural marker anymore, but rather the characteristic consonant of the 2nd person. This introduces a new mechanism of derivation – the original exclusive me ’we’ is combined with this characteristic consonant of ’thou/ye’ to create a new inclusive form which probably replaces the older ñe. The reason behind introducing such a new suffix might be sought in the phonological difficulties arising in association with the velar nasal ñ (e.g. the development of *lñe with the plural marker l).
So the new dative possessive form -elman appears now below elen síla omentielwan and the greeting finally takes the form:

This is how the phrase was published in LotR and Tolkien analyzes it in the following way: -lma possessive suffix of inclusive plural ’we’ = you and I/me, in genitive form -lmo (PE17:13). Another occurrence of the same suffix in LotR is andave laituvalmet! in the Cormallen praise, translated ’long we will praise them’ (Let:230, 8 June 1961). This seems to be a somewhat rhetorical exclamation, the crowd is praising Sam and Frodo and is addressing itself.
Tolkien was content with the outlined basic scheme for a long time, until 1965. On 21 Feb. 1958 he wrote: An enquirer (among many) asked what the L[ord of the] R[ings] was all about, and whether it was an ’allegory’. And I said it was an effort to create a situation in which a common greeting would be elen síla lúmenn’ omentielmo, and that the phase long antedated the book. (Let:205)

3  Quendi and Eldar

In 1959-60 the essay Quendi and Eldar was written and the Quenya greeting there takes at first the form elen síla lúmenna omentielmo (WJ:367,417), changed to omentielvo; Telerin ēl sīla lūmena vomentienguo. The exclusive suffix -mme appears in the essay in the examples avamme, vamme ’we won’t’ (WJ:371). The question arises: What is the full picture of the pronominal conception of that time?

At first it is important to note that the essay distinguishes o-men-t-ie ’meeting or junction of the directions of two people’ (WJ:367) from yo-men-ie ’meeting, gathering (of three or more coming from different directions)’ (WJ:407). The former is clearly dual as compared to the latter and seems to contain the dual marker t, yet the suffix -lma is plural, not dual. Perhaps this was the reason for the change omentielmo >> omentielvo. Note that late explanations of omentie involve either a compound of men- ’go, proceed’ and tie ’path, road’ or the gerund of a verb menta- (PE17:13) without dual connotations, thereby justifying plural -lma. At the time of Quendi and Eldar we thus might be dealing with the following scheme:

1959-60 (reconstructed) (a)

According to this explanation l cannot be a plural marker, but has to be the characteristic consonant of the 2nd person because it appears as part of a dual suffix. The Telerin form was left unchanged which could have various explanations. Perhaps Telerin does not have separate dual forms at all (cf. the statement that the dual inflections are peculiar to Quenya (VT49:51)) or perhaps it forms the dual differently: we/ñe/ñwe > dual incl. *-ngue, dual poss. #-ngua (cf. Early Qenya).

On the other hand there may be another explanation of the apparent plural possessive suffix in omentielmo. Note that in the PE17:132 chart (discussed below) dual forms are formed by -o < *-u, but the possessive forms coincide, e.g. 1st pl. incl. -lme, 1st dual incl. -lmo, but poss. -lma in both cases. In such a case one would be dealing with the following scheme by the time of Quendi and Eldar:

1959-60 (reconstructed) (b)

4  Writings in 1960-1965

Several pronominal charts follow Quendi and Eldar in the 1960s which often employ kar- ’do’ as example.

13 December 1962 (PE17:132)

kar- ’do’ and ala- ’grow’

Here the dual is again formed by -o < *-u. Furthermore the strengthening m > mb appears again, but this time in the plural. It seems that the derivatives of we/ñe/ñwe completely disappear here in favour of l-me.

June 1964 (PE17:75, VT49:48)

This chart is almost identical to the previous one except that there is no m > mb and that pl. -ngwe, dual -ngo < *-ngwu < we/ñe/ñwe appear as variants beside l-me.

There is another verb paradigm in PE17:57 which is not explicitly dated, but is written on the reverse side of a sheet with an essay on comparison. Among other things it deals with a revision Vanimalda >> Vanimelda. The former appears both in LotR and Quendi and Eldar but was changed to Vanimelda in the second edition of LotR (1965), so the chart could have been written in between.

The dual forms in this paradigm are very different. The exclusive dual and plural forms are now identical, but the mechanisms of creation are probably different. Plural forms in this chart are usually formed by nasal strengthening: 2nd pl. *ce > cari-nce, *le > *carin-le > cari-lle (beside carilde < *caril-ne), 3rd pl. te, ta > cari-nte, cari-nta. On the other hand there are two sets of dual forms. One is apparently formed by reduplication of the consonant: 2nd dual *ce > cari-cce, *le > cari-lle, 3rd dual *te > cari-tte. Both mechanisms of course yield same ending in the case of *me (and perhaps *le). The inclusive plural -ñgwe thus might be a result of reduplication *-ññ(w)e (see below).
Incl. dual carinque can be either explained by nasal strengthening of *que (cf. Early Qenya qe-) or by a combination *ni-ki of the first and second characteristic consonants leading to inke > inque with later epenthesis of w, cf. VT49:51.
However, somewhat surprising is the exclusive dual -nwe which has always been inclusive before. Perhaps we is a dual stem in this conception. In the Qenya Lexicon, ATA and WI̯- (U) are mentioned as dual stems (QL:33), cf. e.g. dual henwi *’pair of eyes’ (QL:82). Obviously there is a connection with the dual marker u. Duality of we is also briefly mentioned in VT49:10.

There are two other lists of personal endings where no dual forms appear. The plural forms consistently show excl. -mme, incl. -lme or similar forms:

PE17:75 (bottom of the page)

Here, the first variants are written below a set of long forms like -nye, -tye, -re (1st, 2nd and 3rd sg. respectively), the others below short forms like -n, -t, -s. The affixed -i would suggest -mmeni, -lmeni as alternative long forms. However, amātielmi *’we have eaten’ appears as an example, showing -lmi. In any case it seems certain that -i is a plural marker parallel to dual -u.


5  The 1965 change and subsequent writings

A major conceptional change of the discussed endings occurs around 1965 (see VT49:49). Tolkien now states about omentielmo: Wrong. -elmo, -lme was exclusive form. The inclusive was elwe > elve. Frodo made mistake which was probably glossed over (PE17:13). In the second edition of LotR omentielmo is therefore changed to omentielvo (but laituvalmet is allowed to stand). The ending -mme now becomes dual exclusive. This change does not affect the basic roots, just the derivation of the respective forms. The new conception is used consistently afterwards:

c. 1968 (VT49:50-51)

Associated with this paradigm is an explanation that dual declensions are formed by reduplicating the consonants: -mme, ? wwe, d-de, t-te, but since -wwe did not yield a satisfactory form it was replaced by nki 1a [1st sg. ni] + 2a [2nd sg. familiar ki] since dual was mainly familiar and maintained as such. The same explanation likely applies to carinque (PE17:57) above.

c. 1968 (VT49:16)

The only difference to the preceding chart is inclusive dual -ngwe instead of -inque. Another discussion of the 1st person pronouns involves the inclusive root we, ñwe which was derived from me and not the reverse: it had as it were a nasal stem similar to but not identical with m. The Q[uenya] independent stem we- shows loss of ñ (it would by unhindered change have remained ñwe, since initial ñ was not lost in Q[uenya]), partly perhaps to increase the dissimilarity of me and ñwe, and also because in verbal inflexion ñ was lost after the plural l-infix that in Q[uenya] preceded the pronominal subject elements (VT49:10). Hence caringwe could be explained as being derived directly from ñwe, perhaps by reduplication *ññwe.

One is left to think about Tolkien’s motivation behind such a revision of pronominal endings. The dual forms are perhaps the reason for it, consider a following scenario: Although consonant reduplication was often involved in their derivation, it was always combined with the dual marker -o < *-u, so that dual forms could always be distinguished from plural which had the final vowel -e. But apparently Tolkien became dissatisfied with final -o and so the late dual endings are formed by reduplication of the consonants only. This necessarily makes me > -mme dual which is identical to the plural form in the PE17:57 paradigm. In order to keep plural and dual distinguished, he decided to use l as a plural marker leading to excl. pl. l-me and incl. pl. l-we.

Tolkien could not make the change just like that, apparently he felt that he needed to explain why the first edition of LotR had omentielmo. Several variants of the explanation can be found in PE17:129-131. In short, the combination lw is pronounced lb in later Quenya, so that -lme and -lwe > -lbe become very close in pronunciation and are often confused by non-native Quenya speakers, especially men, but also Noldor born in Middle-earth (according to one of the variants at least). Therefore Frodo made a mistake to which Gildor politely made no remark. Since the confusion was phonological, due to a similar pronunciation, it did not appear in the dual where the forms were sufficiently dissimilar.

6  Examples of usage

before 1965

The exclusive possessive -mma is used in the following examples:

Ataremma/Átaremma *’our Father’, mastamma/massamma *’our bread’, lucassemmar / luciemmar / úcaremmar / rohtammar *’our trespasses’, rocindillomman *’from those who trespass’, fíriemmo / effíriemmo *’of our death’ (VT43:10-12,26-28), late Mélamarimma ’Our Home’ (PE17:109)
The verbal ending -mme appears in firuvamme *’we shall die’. The pronoun me appears as acc. me, abl. mello, dat. men; emphatic emme (also in compound etemme *’out of us’, mimme *’in us’ (VT43:36)) and emphatic dat. emmen; separate poss. menya *’our’ (VT43:10-12,26-28).
Most of the examples appear in translations of Catholic prayers, hence the excluded persons are those to whom they are directed. Mélamarimma ’Our Home’, literally ’our beloved dwelling’ is a descriptive term for Aman used by the Noldor in exile, hence the excluded persons are those that have stayed in Middle-earth.

The inclusive emphatic pronoun #engwe < we/ñe/ñwe (not replaced by -lme) appears in etengwe *’out of us’, mingwe *’in us’ (VT43:36), the shorter #ngwe in dat. ngwen *’for us’ (VT21:6, VT49:55 note 18).
The inclusive form -lme (and variant -lmi) appears in mantelme *’we ate’, amātielmi *’we have eaten’ (PE17:76).

after 1965

Apart from the revised Elen síla lúmenn’ omentielvo (LotR 2nd edition) with incl. -lwa, -lva the new conception finds its application in at least one phrase: navilwe ’we judge’ (incl.) (VT42:34).
It is notable, however, that andave laituvalmet ’long we will praise them’ or ’long will we praise the two’ (PE17:102) was allowed to stand and thus switched from inclusive to exclusive. If it was not an oversight, one should perhaps now interpret -t as an affixed 2nd dual pronoun (< tye, cf. hekat, hekal with reduced pronominal affixes of the 2nd person (WJ:364)), i.e. *’long will we praise you two’; but other explanations can probably be found.
Several separate pronouns appear: 1st pl. excl. , 1st pl. incl. >, 1st dual excl. met, 1st dual incl. wet (VT49:51), dat. vēna ’to us’ (VT49:14). Note that met is already found in Namárie (LotR II, ch.7).

7  Note on Sindarin cognates

A Noldorin/Sindarin chart dated c. 1949 contains the following forms of the verb car-:

c. 1949 (VT50:22)

A slightly later one, from the early 50s has almost the same:

early 50s (VT50:22)

A much later chart has a declension of S. car-, gala- parallel to Q. kar-, ala-:

13 December 1962 (PE17:132)

car- ’do’ and gala- ’grow’

The suffix -m is quite clearly the cognate of Q. -mme. The later inclusive S. -nc is probably the cognate of Q. -inke < ni + ki, while the earlier inclusive can be attributed to khe or strengthened kkhā (PE17:14, see above) which would both regularly yield -ch in Sindarin.
What is common to all these Sindarin forms is that the dual and plural forms tend tend to coincide. In one case, the plural can be disambiguated by adding the plural ending -ir, in the other it is the dual which can be specially marked by appending -id (which doubtlessly contains the lenited dual marker t). In the 1949 the two are merged completely.

In another (later) conception the difference between inclusive and exclusive disappears as well: Sindarin had lost the Common Eldarin (CE) distinction between ’we’ Pl. 1a exclusive of the person(s) addressed, and 1b inclusive. But the causes of the confusion were mainly phonetic (PE17:129). This probably refers to the fact that *-ñgwe would regularly become *-mbe > *-m in Sindarin (because of kw, gw > p, b) which would leave -m to be a general 1st pl. suffix in Sindarin.
This is perhaps also the reason why -nc is used in the 1962 table as an incl. pl. form, although it seemingly has to be dual according to the derivation (’I and you’) — it probably replaces *-m < *-ñgwe.

Inclusive -ch is most likely used in the untranslated phrase man agorech? *’what have we done?’, probably referring to the betrayal of Ulfang in the Battle of Unnumbered Tears. The phrase goes on to to describe the consequences for men, and even though it is rhetorically addressed by Rían to her son Tuor, the inclusive should probably be understood to refer to all men.

8  Conclusion

One can summarize the following conceptions and mechanisms that are at work to produce the 1st pl. forms in Elvish:


This document was translated from LATEX by HEVEA.