The Early Noldorin past tense
Jan. 5th 2008
After the composition of the Gnomish Lexicon and the Qenya Lexicon the next major stage of Tolkien’s linguistic creation was the writing of the Early Qenya Grammar (EQG) and the Early Noldorin Grammar (ENG) some time around 1923 while he was at Leeds. The ENG is also accompanied by two compilations of vocabulary, the Noldorin Word-lists (NW) and the Noldorin Dictionary (ND).
There are several verb paradigms given in association with the grammar, and apart from that both NW and ND list about 20 verbs, each with their respective past tenses.
The aim of this article is thus to classify the Noldorin past tense patterns of that time. It will be seen that the conceptions in NW and ND are not quite the same and therefore the two will be treated separately before a final comparison with the verb paradigms will be given. Citations are given by the page numbers of PE13 in brackets, unless noted otherwise.
In Qenya of the EQG the past tense is obtained by the suffixes -ye, -ie, -ne. The commonest is -ie and is normally accompanied by stem strengthening consisting out of (EQG:56):
The situation is not very much different in Noldorin. The strong past tense is obtained by nasal infixion and i-mutation throughout the word apparently caused by the suffix -ye. So mad- > maint (< *mantye) ’ate’ where Qenya has *mantye > mansie (EQG:57). Note that some forms hint at a past tense suffix -ya rather than -ye, e.g. 3rd masc. medenniog < *mat-ant-yā-k-, fem. [med]ennias < *mat-ant-ya-s- (163), but this weak past tense formation is already analogical.
For a discussion of Noldorin vowel mutation at that time see .
All verbs can be grouped into ’basic verbs’ which contain no derivative suffixes (called strong -(i) verbs by Tolkien (165)) and ’derived verbs’ which are formed with -a, -ta, -nta, -ya, -tya (EQG:56) in Qenya. In Noldorin these suffixes usually appear as -a, -d, -ia.
Tolkien lists the verbs either by their roots, the infinitives, or by the 3rd singular impersonal form. The infinitive is usually formed with the suffix -i attached to consonantal stems and -d (< *-t) to vocalic stems (and therefore mostly to derived verbs). In the former case i-mutation is caused. But *-t is sometimes also appended to consonantal roots where it causes further changes, as alaith ’inf. shielding, warding off, protect(ion)’ < *alakt-, stem *alag- or gonoth ’inf. to count’ < *go-nod-t, stem gonod. Finally, derived verbs ending in -ata > -ad usually lengthen the last vowel in the infinitive: *-āta > -od, so tangod ’to fix, fixing’, stem *tangad-.
The 3rd sg. impersonal form is usually endingless, but the same basic verbs that form the infinitive with *-t obtain the 3rd sg. by i-mutation (in Welsh the 3rd sg. present is also either the bare stem or the i-mutated bare stem). So elaig *’shields, wards off’ and genyd, goenyd *’counts’. The reason for that might be that the derivational suffix -ya only occurs in the present. Compare Tolkien’s description of -(n)ta as being mere present-formative in Qenya – it means that this suffix appears in the present (kapta ’leap’) and the future (kaptuva), but is omitted in the past (kampie). In this light the present forms elaig < alak-ya, masc. elegiog < *alak-ya-k-, fem. elegiais < *alak-ya-is- become clear. On the other hand *nuv- ’sinking, going down, to sink, set’ forms the 3rd sg. impersonal ný or ní by i-mutation, but otherwise behaves like a basic verb, so masc. nuveg rather than *nyviog. Hence I have treated this last pattern as a basic verb. Perhaps i-mutation becomes a morphological device by analogy.
To keep the etymologies clear, verbs will be cited by their roots (reconstructed if necessary).
The strong past tense of basic verbs is formed by a combination of nasal infixion and i-mutation. The plural is formed by the suffix -i with medial change -mp-, -nt-, -nc- > -mm-, -nn-, -ng-. The forms crimp and later (2.1) pint are actually glossed ’Old Noldorin’ (therefore medial -nth- in pl. pinthi; late Noldorin would be *pinni), in other cases the weak past tense (1.2) is just mentioned as ’more recent’. So it appears that this pattern (1.1) becomes archaic in the course of time, displaced by an analogical formation.
Several examples indicate that the past tense of verbs with a syllabic consonant was originally obtained by lengthening. In Noldorin this results in an apparent a-infixion. So hist ’spits’ < hṣt and lhif < sḷq, sḷp are due to the resolution of syllabic ṣ > is and ḷ > il . The past tense of lhif is heilf (later hailf) from sl:q- [Tolkien’s spelling] > *salf-, subsequent i-mutation and the common change s- > h- initially. It seems that long syllabic s: becomes as, hence *hs:t- > *hast- and finally haist.
It appears that in later Noldorin a-infixion is recognized as a new pattern, so that a new analogical present is formed by hilf, which now goes with the past tense hailf. It regularizes the disagreement of the initial consonants (h- and lh-). Also, the verb gist ’knows’ must be derived from ʒist- (146), there is no syllabic consonant here. The past tense gaist is then either formed by analogy to the verbs like hist; or maybe a-infixion is quite regular in such a case because n-infixion would probably have no effect upon the from (*ʒinst- > ʒist-).
The weak past tense of basic verbs is formed with the suffix -aint, pl. -enni (with i-mutation) for any root vowel. It seems that this suffix was later generally introduced for both basic and derived verbs. Later Noldorin sources indicate that it is an analogical formation from strong past tenses of verbs ending in -d (< *-t), like maint and tengaint (PE17:44).
As already mentioned the derivational suffix -ta is often present-formative in Qenya and is omitted in the past tense (kapta – kampie). However, in causative or transitive formations the past suffixes -ne, -ine are used (lapta *’wrap’ (cf. lapa- ’wrap, swathe, wind’ (QL:51)) – laptane). Another pattern is nasal infixion of the derivative ending (lokta – lokante) (EQG:58).
It appears that the present-formative suffix in Noldorin is rather -ya than -ta, cf. rotya- > rhoid in ND. Otherwise we find nasal infixion of the last consonant after the fashion of lokta – lokante and usual addition of *-ye causing i-mutation.
Note that curenni is probably formed with causative -(n)ta. This suffix loses its final -a in Noldorin practically yielding a consonantal stem (cf. tangod, meriad), so the infinitive is formed with -i. The intransitive past tense agyraint appears to be formed by i-affection only, without the extension *curant- > *curana-n-t. The augment seems to be quite unique – it otherwise only appears in agor, which is transitive, however.
The verb peda seems to be derived with -a, as also seen in the 3rd masc. pres. pedog < *kwetā-k-. Since there is no consonant in the derivational ending for n-infixion, n-infixion appears to act upon the root consonant, as in the case of basic verbs.
In Qenya this past tense is obtained by -ne, but Noldorin employs the suffix -(a)int, -(a)in (< *-(a)ntye, *-(a)nye), pl. -(e)nni, -(e)ni (< *-a)nti-, *-(a)ni-) instead; which is, as already mentioned, most probably an analogical derivation from regular strong past tenses like maint and tengaint (PE17:44). I-mutation is again carried out throughout the word.
There is only one example:
This is a compound of crimp *’bent’ and tha- ’to make cause to be’. The past tense is obtained by simply putting tha into the past tense aist, see (3).
The verb tha- ’to make cause to be’ is somewhat irregular, the basic form is from s’ta-, the past tense stem is asti- with a prefixed root vowel. This yields aist without further suffixes for the 3rd impers., and 3rd masc. astig.
Another irregular example is gwaist ’is aware, recognises’ (wa-ʒist, 146), past tense guist. The present form is apparently derived from ʒist ’knows’ with the prefix gwa- ’together, co(n)-’ (< *wa-), but the past tense seems to incorporate the unstressed form go- (< *wo-) of this prefix with subsequent oi > ui.
There is a main difference in the past tense formation introduced in the ND: There are now two past tenses, the imperfect and the aorist. Although Tolkien later calls a timeless or habitual verb form ’aorist’, it is traditionally a term referring to a simple action in the past, without any other aspectual restrictions. It it is therefore close to the perfect tense, although the perfect tense properly conveys the idea that an action has already occurred at the time of speaking, or the state resulting from the action. So English ’was doing’, ’did’ and ’has done’ would roughly correspond to the imperfect, aorist and perfect, respectively.
Note that there is no perfect tense in any Early Noldorin verb paradigm. In Welsh, there is an imperfect tense and a past tense; the past tense is either perfect or aorist in sense, but mostly the latter (J. Morris-Jones, An Elementary Welsh Grammar §267,270). One observes that Early Noldorin is strongly based on Welsh in phonology  and it seems that Tolkien had a very similar verbal system in mind. Of course, Ancient Greek is also well-known for having a past, perfect and aorist.
The fact that the Noldorin aorist is indeed a past tense can be for example seen from the translation of aorist maint ’ate’; compare past imperfect madath ’was eating’. Even more so, all the usual past tense patterns of NW are now glossed ’aorist’. So Q. alatya *’shield, ward off, protect’ (158) forms the aorist altíne – this is a pattern in good agreement with the EQG pattern tulya ’send, bring’ > pa.t. tulı̆ne or tantya ’set bouncing’ > tansı̆ne, rarely -īne; but there is no distinguishing of the imperfect and aorist in the EQG.
The variants pint and peint show different results of i-mutation: In the former case we see a raising *pentye > *pint(i), in the latter an infixion *pentye > *peint(i).
Note that †dadní is formed by i-mutation only, probably because of the early loss of the spirant v after u – this way the verb ends in a vowel and nasal infixion cannot be applied.
This weak past tense suffix is probably just the same suffix *-ntye as for derived verbs, but it is attached to the stem via the connecting vowel -i-. The plural is now formed with -ir rather than -i in NW, but is attested only for alaith.
Note that in dadnú the spirant v (bh) is lost after u in final position.
Initial p- becomes h- in Early Noldorin and syllabic -ṣ- is resolved in -is-. In the past tense we are probably again dealing with long syllabic *s: > as. Also note sk > ch.
Note that after r we are rather dealing with a nasal suffix (> rn) than infix. The verb anthra- can be very well seen falling into the same pattern as maitha < mapta or Q. lokta – for the nasal fortification it is extended, so *antra- > *antar- > pa.t. *antarnye > enneirn. Compare mapta > *mapat- > pa.t. *mapantye > mebaint and Q. lokta > lokante.
Some more explanation is needed in the case of rhynt < rhaid. The verb base is rotya- and there is no change of the root vowel except for an epenthesis o > oi, perhaps because a- and i-mutation of the o balance each other out. So if the first i-mutation raises rotyā > *ru(i)tya, the vowel u is again lowered by subsequent a-mutation > rhoid. Finally, ONo. oi is regularly changed to ai. This explanation is not compelling, however, it may be that the variants rotya- > *rh(o)eid-, *rhuid just have not been realized (the former has been realized in the 3rd masc. form rhe(i)diog beside rhoidiog).
In any case there is no a-mutation in the past tense: *rontye- > *runtye- > rhynt.
Note that the past tense suffix substitutes the whole derivational ending -ya in rot-ya- > *rot-antye > rhodaint. This is not the case in NW, where the corresponding form is rheidiain(t) < *rotya-ntye.
This past tense seems to be formed by vowel lengthening and augment *akār- > agor, but a nasal can be seen in pl. *akarni > egerni. Although this pattern is usual in later Sindarin, agor is exceptional in Early Noldorin.
The imperfect tense formation actually displays a very simple pattern – it is obtained by the suffix -ath for all verb classes alike. The suffix may cause a-mutation of the preceding vowel.
NW and ND are accompanied by a grammar entitled Lam na·NGoluith and several paradigms of verb declensions which may be compared with what has been discussed so far:
We see here the suffix -ath for derived verbs and nasal infixion for basic verbs – a distinction never met in the wordlists. Furthermore the retained sketch has columns which are probably representing the aorist. It is apparently formed by prefixes, glathra *’polishes’ > aglathra, dengin *’I slay’ > gindengin, manthin *’I eat’ > (gi)mennin. Such a pattern never appears in the wordlists either.
Furthermore there is a very detailed paradigm (131-132) with a complete declension of 5 different verbs:
This paradigm is in accordance with ND, where madath, maint, edainc also appear (tengaint appears in NW).
Note that the gloss lub- seems to reflect a primitive form, becoming *lhuv- in Noldorin. Hence the possibilities *lhuvath > lhovath with a-mutation (asterisked by Tolkien) or *lhuvath > lhuath with loss of v after u (cf. dadnú, ný above). The aorist forms *lhîf, *lhîw are also asterisked by Tolkien – they seem to be formed by i-mutation only (cf. dadnú > aor. †dadní above) and final -v is not lost because the preceding vowel is not u anymore. Instead, it becomes devoiced -f or the approximant -w (the latter might also be a vowel [u]).
The past form gwardhaf is once again dissimilated *gwardhath (cf. maithaf). Aorist gwridh seems strange at the first glance but may be explained by the fact that the primitive form has a syllabic ṛ: ngwṛþ- or gu̯rþū́ > Goldogrin gurthu ’death’ (PE11:43). Hence perhaps *(n)gwṛđ-ye > gwridh (with a labialized g); and there is no room for n-infixion.
update: Apr 4th 2008 — reference  added, some corrections
update: Jun 15th 2009 — heilf, hailf, lhimp, lhifaint added
This document was translated from LATEX by HEVEA.