Since at least the publication of The Lost Road in 1987, it has been widely recognized that the Eldarin tongues, at least as exhibited from the latter half of the 1930s through the publication of The Lord of the Rings, have two primary classes of past-tense verb formation, which we can call (using Tolkien’s own terminology in Quendi and Eldar, itself borrowed from Indo-European linguistics): strong (indicating internal modification) and weak (indicating modification by affixes). The strong past tense, which class includes chiefly verbs arising directly from basic stems (that is, verb-stems arising directly from simple bases with verbal meaning, such as KAT- shape, BAT- tread, SUK- drink, etc.), is characterized by lengthening or otherwise strengthening the root syllable and suffixion of original final *‑ē. Within the strong past tense there are, again, two chief means of root strengthening: a) lengthening or fortification of the root vowel, and b) nasal infixion. The weak past tense, which class includes chiefly verbs arising from derived verb-stems (that is, verb-stems formed from a root by suffixion of some derivational element, such as *‑ā, *‑tā, *‑yā, etc.), is characterized by the lack of root syllable strengthening and the suffixion of some past-tense marker, very often *‑nē. These two main classes are neatly exemplified by the two Qenya past-tense verbs arising from the base ONO- beget in Etymologies, óne (strong) and ontane (weak), the former arising directly from the base and the latter from the derived stem onta-. For Quenya of The Lord of the Rings, the same mechanisms are exemplified by the past-tense verbs found in Galadriel’s Lament, which includes, on the one hand, unduláve down-licked, where ‑láve ‑licked is a strong past-tense verb formed on LAB- lick; and, on the other hand, ortane uplifted, a weak past-tense verb (in ‑ne) based on the derived stem orta- raise (cf. ORO-).1
Starting from this most basic formal distinction, we can classify the past-tense verbs in the Noldorin of the Etymologies by their manner of formation thusly (in the following, all words are Noldorin and all words and bases are cited from the Etymologies and/or from the Addenda and Corrigenda to the Etymologies, unless otherwise indicated):
Strong past I: Root vowel lengthening or A-infixion of basic stems
In the Etymologies all past-tense verbs in this formation class arose directly from Eldarin bases ending in a liquid (i.e., ‑L or ‑R) or in a voiced stop (here, only ‑D occurs).
- †daul *hid, concealed (s.v. DUL- hide, conceal). This is listed as a poetic past tense of two derived verbs: dœlio / delio (apparently < *dolyā- < *dul-yā-) and doltha (apparently < *doltā < *dul-tā). The past-tense form itself, however, shows derivation directly from the base, presumably by A-infixion; thus < *d-a-ul-ē < DUL-.
- haul *lifted (pa.t. of heli lift < KHAL2- uplift). The lengthening of the root vowel can here be due either to A-infixion or simply to lengthening, as they would both produce the same result, *ā, regularly diphthongizing to *au (and remaining a diphthong in stressed monosyllables; cf. N bauð judgement < bād- < BAD-); thus < *khāl-ē < KHAL2-.
- †trenor *recounted, told to end (pa.t. of treneri *he recounts, tells to end < NAR2- tell, relate). The prefix tre- through (< TER-, q.v.) is unrelated to the past-tense formation, which concerns only the verbal element ‑nor. In unstressed word-final syllables, N o can result, inter alia, from original long *ā, which given the base form is most likely the case here. The lengthening of the root vowel can again here be due either to A-infixion or simply to lengthening; thus < *nār-ē < NAR2-.
- nîð hurt, bruised (s.v. NID[2]- hurt, injure), stated to be < ON nīde, itself apparently < *nīd-ē < NID[2]-.2
Strong past II: Nasal infixion
This formation was predominantly applied to basic stems, but there is also a small class of derived stems that appear to exhibit nasal infixion.
In the Etymologies all past-tense verbs in this formation class arose directly from Eldarin bases ending in a stop (either voiceless — here, ‑P , ‑T , and ‑K occur — or voiced — here, only ‑D) or in a nasal (here, only ‑M). Not surprisingly, each of these two subclasses of Eldarin bases gave rise to a slightly different subclass of development from the nasal-infixed form.
i. Basic stems ending in a stop consonant
- trevant *traversed (pa.t. of trevedi *traverse < BAT- tread). Again, the prefix tre- through (< TER-, q.v.) is unrelated to the past-tense formation, leaving us with the verbal element ‑vant. The development appears to be by lenition from *bant < *ba-n-t-ē by nasal infixion < BAT-.
- echant *fashioned, shaped (pa.t. of echedi fashion, shape and stated to < *et-kat- < KAT- shape, with the contact *‑t-k- yielding N ‑ch-). Tolkien’s own etymological note shows that formation contains an original adverbial prefix *et- (< ET- out, forth, q.v.), leaving us with the verbal element *cant, apparently < *ka-n-t-ē by nasal infixion < KAT-.
- hant *sat (pa.t. of hað- *sit < KHAD- sit), apparently < *ha-n-d-ē by nasal infixion < KHAD-.
- hant *hurled (pa.t. of hedi hurl < KHAT- hurl), apparently < *ha-n-t-ē by nasal infixion < KHAT-.
- [†eglent] *went into exile (pa.t. of egledhi or egle[d]hio go into exile < LED- go, fare, travel), again showing the adverbial prefix *et- (< ET- out, forth) with the regular development of the contact *-t-l- > -gl-, leaving us with the verbal element *lent, apparently < *le-n-d-ē by nasal infixion < LED-.3
- adlenc *released, freed (pa.t. of †adlegi < LEK- loose, let loose, release). The Noldorin prefix ad- back, again, re- is given under the base AT(AT)- again, back; so the verbal element ‑lenc can be isolated, and explained as < *le-n-k-ē by nasal infixion < LEK-.
- danc *slew (pa.t. of degi to slay < NDAK- slay), apparently < *nda-n-k-ē by nasal infixion < NDAK-.
- rhimp *rushed, flew, flung (apparently the pa.t of rhib- < RIP- rush, fly, fling), apparently < * rhi-m-p-ē < by nasal infixion < RIP-.
- nestanc *inserted, thrust in (given under both NĒ˘- *in, inside and STAK- split, insert, pa.t. of nestegi insert, thrust in, which is stated under the former base to be derived from *ne-stak- insert, thrust in, sting). The prefix ne- *in is not part of the past-tense formation, so the verbal element ‑stanc can be isolated, and explained as < *sta-n-k-ē by nasal infixion < STAK-.
- sunc *drank (pa.t. of sogo drink < SUK- drink), apparently < *su-n-k-ē by nasal infixion < SUK-.
- gwend (also gwenn) *bound (past-tense of gwedi < WED- bind), apparently < *we-n-d-ē by nasal infixion < WED-.
As can be seen, all of the past-tense verbs in this class end in a consonant combination having the pattern homorganic nasal + stop, and further that all but one have, more specifically, homorganic nasal + voiceless stop (here, either ‑nt or ‑nc). This latter pattern appears to have been particularly prevalent and may explain why we have gwend < WED-, but hant < KHAD- and eglent < LED-: whereas by regular phonological development we would expect ‑nd in all cases, in the cases of hant and eglent it appears that the final stop was devoiced due to the analogical influence of the many other past-tense verbs of this pattern.
ii. Basic stems ending in a nasal consonant
- †dramp *hewed (pa.t. of dravo hew < DARÁM- beat, hew).
- hamp *sat (pa.t. of haf- *sit < KHAM- sit).
Both verbs attested for this class arose directly from bases ending in ‑M, and have the termination ‑mp. This final ‑p is interesting, as it cannot be explained as arising by regular phonological development from any nasal infix combination with the final basic consonant ‑M, which in fact ought to have yielded ‑mm (as, for example, in the accompanying, apparently 1st sg. pa.t., forms Tolkien gives, drammen and hemmin, respectively). Unless these verbs reflect an otherwise unattested weak past-tense formation in suffixed *‑pV (where V stands form some vowel now lost), which seems unlikely, it would appear that the ‑mp arose from other than regular phonological development.
Noting again the termination pattern homorganic nasal + voiceless stop so prominent in the previous class, it is significant that the termination ‑mp seen in this class continues this pattern. Indeed, it would appear that ‑mp here is an analogical reformation of the expected historical termination *‑mm due to the influence of the prominent homorganic nasal + voiceless stop pattern, with the final *‑m (voiced bilabial nasal), devoiced and stopped to ‑p (voiceless bilabial stop). It is noteworthy, too, that the cognate noun corresponding to dravo, EN dramb, dram(m) a heavy stroke, a blow, shows derivation from the base by an apparent suffix *‑bā, with an initial stop; perhaps the stop of this form influenced this formation as well.
Finally, there is a small class of strong nasal-infixion verbs, which however are formed from derived stems in *‑tā, including:
- lhimmint *moistened (pa.t. of lhimmid moisten < LINKWI- *wet), apparently < *linkwi-n-t-ē < *linkwi-tā < LINKWI-.
- nimmint *whitened (pa.t. of nimmid to whiten < NIK-W-), apparently < *ninkwi-n-t-ē < *ninkwi-tā < NIK-W-.
- †sint *had knowledge (pa.t. of isto to have knowledge (intr.) < IS- *know), apparently < *sintē < *isintē, apparently arising by resolution of syllabic nasal < *is-n-t-ē < *is-tā).4
Noting that, without exception, every other class of strong past-tense formation identified above includes only basic stems, it is interesting to find nasal infixion applied to these derived stems. The question arises, why these derived stems in *‑tā do not instead form their past tense like all other attested derived stems, sc. as weak verbs taking various past-tense suffixes (see below)?
A possible answer to this question arises from consideration of the semantic nature of the verbs in this class, in comparison with the class of weak intransitive pa.t. verbs in ‑as. First, notice that isto is specifically stated to be an intransitive verb. Further noting the inherent semantic relationship of intransitive verbs with inchoative verbs — both denote a verbal concept that describes or affects the subject of the verb — it is in this light unsurprising that the class otherwise contains derived verbs that are potentially at least partly inchoative and hence at least partly intransitive: thus nimmid to whiten, perhaps indicating both the causative and transitive meaning *to whiten (a thing), to cause a thing to become white(r) and the inchoative, intransitive meaning *to become white(r); likewise lhimmid moisten, perhaps both causative and transitive *to moisten (a thing), to cause a thing to become moist(er) and inchoative, intransitive *to become moist(er). Finally, note that †sint, a strong intransitive pa.t. verb in ‑nt, also has a weak intransitive pa.t. counterpart in ‑as, sc. istas (see Weak II below). From this it seems evident that this class represents the strong past-tense formation of intransitive and inchoative verbs, and thus represents the strong counterpart to the weak formation in ‑as of intransitive verbs.
Interlude: Two Poetic/Archaic Past-Tense Verbs in ‑nte
Before proceeding, some account must be given of two poetic, and phonologically seemingly quite archaic, past-tense verbs:
- †narante *told (pa.t. of †naro tell < NAR2- tell, relate).
- †oronte arose (apparently the original pa.t. of erio rise < ORO-).
From their manner of presentation it seems that both of these verbs are in fact intended to stand as poetic/archaic, but nonetheless specifically Noldorin, not Old Noldorin, forms. The first noteworthy feature of these forms is the presence of a final vowel in these words, which ought to have been lost just as were all final vowels in the transition between Old Noldorin and Noldorin proper. It would seem that these forms are essentially linguistic fossils, preserved in form from (possibly very late) Old Noldorin, and retained as historical relics. But for the present work the most interesting feature of these verbs is the fact that they exhibit an apparently fossilized pa.t. ending ‑nte that could be attached both to basic stems ending in a vowel, as apparently in †oronte < ORO-; and to derived stems, as apparently in †narante < *nara- < NAR2-.5 Noting that an archaic/Old Noldorin pa.t. ending ‑nte would by regular phonological development yield ‑nt in Noldorin proper, this old ending seems likely to underlie the pa.t. suffix ‑(a)nt seen throughout the next major formation class.
Weak past I: Suffixion of ‑(a)nt
In this class we find past-tense verbs built directly from the unmodified base, and those built from a basic stem exhibiting internal i-affection.
- hadhant *sat (pa.t. of hað- *sit < KHAD- sit).
- hafant *sat (pa.t. of haf- *sit < KHAM- sit).
- trenerant *recounted, told to end (pa.t. of treneri *he recounts, tells to end < NAR2- tell, relate).6
- degant *slew (pa.t. of degi to slay < NDAK- slay).
- gwedhant *bound (pa.t. of gwedi *bind < WED- bind).7
As can readily be seen, each of these past-tense verbs exhibits the same vowel as the internal i-affected form found in the corresponding infinitive (the root vowel of gwedhant, e, being insusceptible to internal i-affection, remains unchanged). This strongly suggests that this formation arose only after the phenomenon of internal i-affection expressed itself in the infinitive, and thus relatively late in the historical development of Noldorin. Noting further that each of these three verbs also has a corresponding strong basic past-tense form (†trenor, danc, and gwend, respectively), it seems further likely that this formation arose as a late, analogical weak past-tense of basic verbs.
In this class, *‑ā-, *‑bā-, *‑tā-, and *‑yā-stem verbs are represented.
- dravant *hewed (pa.t. of dravo hew < DARÁM- beat, hew).
- garant *held, possessed (pa.t. of garo *hold, possess < GAR- hold, possess).
- melant *loved (pa.t. of melo *to love < MEL- love (as friend)).
- sogant *drank (pa.t. of sogo *drink < SUK- drink).
- dammant *hammered (pa.t. of damma- to hammer < *ndam-bā- < NDAM- hammer, beat).8
- orthant *raised (pa.t. of ortho raise < ORO- up; rise; high).
- diliant *stopped (pa.t. of dilio to stop < DIL- stop up, fill up hole, etc.).
- deliant *hid, concealed (pa.t. of doelio (doltho) *hide, conceal < DUL- hide, conceal (first pencil version)).
- tiriant *watched, guarded (pa.t. of tiri or tirio < TIR- watch, guard).
Note that the derivative ending ‑io arising from *‑yā- causes internal i-affection in root-vowels susceptible to it.
There remains one past-tense verb in ‑(a)nt that does not clearly fit into any of these formation classes:
- †eglant *went into exile (pa.t. of egledhi or egle[d]hio go into exile < LED- go, fare, travel).9
This is a quite curious form. Assuming it is not a slip for *†egleðant — which would presumably belong with the verbs in Weak I-A-ii, formed by suffixion of ‑(a)nt to the internal i-affected stem (e being insusceptible to internal i-affection) — it appears to represent an analogical formation due to the shape and two-syllable, stem + ‑ant structure of nearly all of the other verbs ending in ‑ant. Such a reformation would be assisted if the prefix eg- < *et- < ET- forth, out (with the regular development of ‑gl- from the contact *‑t-l-) in this verb had ceased to be felt to be such, due to its phonological alteration, so that the combination egl- in effect came to be felt to carry a verbal force; i.e., as if egl- meant *go forth, go out. Such a reshaping would be further assisted in this particular case by the shape and meaning of Strong II-A-i trevant *traversed, and by the shape of other such adverbially-prefixed forms as echant and nestanc, all of which indeed may have attracted the strong verb from LED- to its class. But against this explanation, it can be observed that no such reshaping is evident in this pa.t. verb's cognate and counterpart, egleðas.
Weak past II: Suffixion of ‑(a)s
- istas *had knowledge (pa.t. of isto to have knowledge (intr.) < IS- *know).
- egleðas *went into exile (pa.t. of egledhi or egle[d]hio go into exile < LED- go, fare, travel).
- mudas *laboured, toiled (pa.t. of mudo labour, toil < MŌ-).
- erias *arose (pa.t. of erio rise < ORO- up; rise; high; etc.).
Notice that we find both basic (egleðas) and derived stems (*‑tā-stems istas and mudas, and *‑yā-stem erias) in this category.
The most striking common feature of the verbs in this class is that they are all intransitive, while with the exception of the intransitive/inchoative verbs under Strong II-B and of hant *sat (pa.t. of hað- < KHAD- sit) and hamp *sat (pa.t. of haf- < KHAM- sit), the verbs in all other past-tense formation classes identified here are either primarily or entirely transitive. It appears that the present class represents the chief weak pa.t. formation of intransitive verbs, with Strong II-B its strong counterpart; while the other strong and weak formation classes are applied primarily to transitive verbs.
It should be noted that the suffix ‑(a)s must have arisen from an earlier form with a double consonant, as original final *‑s or *‑s coming to stand finally after the loss of an original final vowel would regularly be lost in Noldorin. This earlier form might have been *‑ssV, where V stands for some vowel; but given the prevalence within the past-tense formation classes of combinations with a homorganic nasal, it seems more likely that the earlier form was instead *‑nsV, which would by regular development > *‑ssV > ‑s (regular, that is, if the development exhibited by S Arassuil < *aran-suil (LR:1014, XII:196, 211) can be assumed to apply to Noldorin of the Etymologies; and there appear to be no counterexamples to this development therein). If this is the correct derivation, it forms a nice parallel with the *‑nt- of the Weak I formation class.
Table 1 below provides a tabulation of all of the attested formations, grouped by base and then by the four major formation categories identified above.
| Base | Strong I | Strong II | Weak I | Weak II |
| BAT- | — | trevant | — | — |
| DARÁM- | — | †dramp | dravant | — |
| DIL- | — | — | diliant | — |
| DUL- | †daul | — | deliant | — |
| GAR- | — | — | garant | — |
| IS- | — | †sint | — | istas |
| KAT- | — | echant | — | — |
| KHAD- | — | hant | hadhant | — |
| KHAL2- | haul | — | — | — |
| KHAM- | — | hamp | hafant | — |
| KHAT- | — | hant | — | — |
| LED- | — | [?†eglant/eglent] | — | egleðas |
| LEK- | — | adlenc | — | — |
| LINKWI- | — | lhimmint | — | — |
| MEL- | — | — | melant | — |
| MŌ- | — | — | — | mudas |
| NAR2- | †trenor | — | trenerant | — |
| NDAK- | — | danc | degant | — |
| NDAM- | — | — | dammant | — |
| NID[2]- | [?nîð] | — | — | — |
| NIK-W- | — | nimmint | — | — |
| ORO- | — | — | orthant | erias |
| RIP- | — | rhimp | — | — |
| STAK- | — | nestanc | — | — |
| SUK- | — | sunc | sogant | — |
| TIR- | — | — | tiriant | — |
| WED- | — | gwend | gwedhant | — |
It can readily be seen from Table 1 that the past-tense verbs derived from a base very often have more than one form, frequently both a strong and a weak form. In only about half of these cases is the strong form labelled as poetic, suggesting that both forms were equally current for many such verbs. Thus it appears that there can be more than one past-tense form for any given Noldorin verb.
It can also be seen that, with the sole (and somewhat dubious) exception of nîð, there is a strict division in the Strong I and Strong II formations according to the phonetic class of the final basic consonant: all verbs arising from bases ending in a liquid (here, ‑L and ‑R occur) form their strong pa.t. by root vowel strengthening or A-infixion (Strong I), while with one possible exception all verbs arising from bases ending in a stop or a nasal (here, ‑P, ‑T, ‑K, ‑D, and ‑M occur) form their strong pa.t. by nasal infixion (Strong II). This suggests, inter alia, that the strong pa.t. of verbs like garo (< GAR-) and melo (< MEL-), if in fact they ever had strong forms, would be Strong I: i.e., *gaur < *gār-ē and *mêl < *mēl-ē. Note further in this connection that nowhere do verbs arising from bases ending in a liquid form a pa.t. by simple suffixion of ‑n (i.e., as though employing the weak pa.t. ending *‑nē evidenced in such weak pa.t. Quenya verbs as ortane 'uplifted', LR:368). Hence we see no evidence whatsoever for pa.t. formations like *garn, *tirn, or *mell.
Finally, it is interesting to note that the small corpus of Sindarin verbs found in The Lord of the Rings exhibits three of the four main mechanisms of past-tense formation that we have identified for Noldorin; thus:
Strong past I: Root vowel lengthening or A-infixion of basic stem
- ónen I gave (LR:1036); removing the apparent 1st sg. ending ‑n I isolates the pa.t. stem óne-, probably < *ān-ē- < ANA1- *give.10
Strong past II-A-i: Nasal infixion of basic stem ending in stop consonant
- echant made (in the phrase Im, Narvi, … echant, I, Narvi, … made, LR:298), presumably like the Noldorin verb < *et-kant- < *et-ka-n-t- by nasal infixion < KAT- shape.
Weak past I-B-iii: Suffixion of ‑(a)nt to derived *‑tā-stem
- teithant drew (in the phrase Celebrimbor … teithant Celebrimbor … drew, LR:298) apparently < *tek-ta-ntē < TEK- make a mark, write or draw (signs or letters).
Missing only is the Weak past II (suffixion of ‑(a)s). Of course, none of the three past-tense verbs in The Lord of the Rings is intransitive, so this lack is not surprising, and neither suggests nor disproves that the class existed for Sindarin of that conceptual era.
Sindarin of a somewhat later conceptual era, though, has a class of past-tense formation not exhibited anywhere in the Etymologies, as seen in and described for a single example in the c. 1959 essay Quendi and Eldar (XI:415):
- agor *made, did < *akāra < KAR-.
Tolkien describes this formation as characterized by the augment, or reduplicated base-vowel, and the long stem-vowel. As such, it combines the characteristics of the Noldorin Strong I formation above, with a weak formation, sc. prefixion of the sundóma; and further, it shares two of the three formation characteristics of the Eldarin perfect (see footnote 1 below) — probably not coincidentally. Tolkien further describes this formation as usual in Sindarin strong or primary verbs, by which he seems to mean what we have described here as basic verbs. But as this formation is nowhere in evidence among even the many basic verbs in the Noldorin of the Etymologies, it seems likely that this past-tense formation was one Tolkien added into his later conception of Sindarin — or rather, restored to it, as this same formation is found in Leeds-era Noldorin, in e.g. the past-tense form agor (PE13:161 s.v. cara).
The formal classification presented here is a revised and greatly expanded version of one included in a post to the Lambengolmor mailing list on July 11, 2003 (message 464).
1. By way of illustrating the role played in Eldarin tense formation by contrastive variation of root-vowel strengthening (or absence thereof) in conjunction with suffixion of a characteristic vowel, it is interesting to note that the Eldarin present tense is characterized by lengthening of the root vowel, but with suffixion of *‑ā: thus Q síla shines, is shining (LR:79); while the aorist is characterized by the absence of strengthening and the suffixion of *‑i: thus Q sile shines (VI:280, 325); and the perfect is characterized by lengthening of the root vowel together with prefixion of the root-vowel and suffixion of a termination yielding ‑ie in Quenya: thus Q u-túv-ie-nye-s I have found it (LR:950).
2. This analysis assumes that N nîð hurt, bruised < ON nīde is a past-tense verb, and not an adjective. Tolkien does not indicate the part of speech explicitly, and the gloss he provides, hurt, bruised, is ambiguous. However, the fact that Tolkien gives a different, more clearly adjectival, Noldorin form, neðw scarred, bruised, derived from ON nidwe, nidwa and cognate with Q nirwa scarred, tends to suggest that N nîð is not an additional adjectival form, but instead a past-tense verb. In any event, it seems best to err on the side of inclusion here.
3. This analysis assumes that correct reading of this past-tense verb is indeed eglent, as suggested by Helge Faukskanger (see Lambengolmor list message 728, Aug. 30, 2004). With the proposed primitive form *lendē cf. Q lende went, departed (V:47, 56, 368 s.v. LED-). For analysis of the alternative reading, eglant, see under Weak I-B-.
4. The existence of the hypothetical intermediary form *sintē is also implied by the cognate Qenya pa.t. form sinte; also given s.v. IS-.
5. Note that this may not be the historical means of derivation. For example, as with †sint < *is‑n‑t‑ē (see above), it seems possible that †narante arose by resolution of syllabic nasal from an infixed form like *nar-n-t-ē; but however it arose, it seems clear from these two examples that at least by the end of the Old Noldorin period a pa.t. suffix -nte had arisen, which could be attached directly to both basic and derived stems.
6. The form trenerant is derived from my interpretation of the fact that Tolkien overwrote original nennar-, nennarant, with trenor, trener. In the manuscript, the final letters ant of nennarant were not overwritten, and immediately follow the replacement trener. I take this to mean that Tolkien intended the new pa.t. forms to be (strong) trenor, (weak) trenerant.
7. The infinitive gwedi corresponding to pa.t. gwedhant is noteworthy as not exhibiting the development of original intervocalic *d > ð/dh seen elsewhere, and as would seem to be expected given gwedhant. But both the infinitive and pa.t. forms are very clearly written as such in the manuscript.
8. This classification takes the Qenya cognate verb-stem namba- as revealing the derivation of the Noldorin verb. It further depends on dammant being the correct reading, rather than dammint, as suggested in the Addenda and Corrigenda to the Etymologies. But it should be noted that if the correct reading is in fact dammint, it could be explained as an analogical formation due to the influence of other derived past-tense verbs in ‑mmint, exemplified in the Etymologies by lhimmint and nimmint.
9. For analysis of the alternative reading eglent see under Strong II-A-i.
10. The long initial ó of the verb is in accordance with the earliest editions and printings of The Lord of the Rings; in later printings, the accent was dropped — according to Douglas Anderson (private communication), most likely inadvertently.
Hostetter, Carl F. and Patrick H. Wynne. Addenda and Corrigenda to the Etymologies — Part One. Published in Vinyar Tengwar 45, Nov. 2003.
———. Addenda and Corrigenda to the Etymologies — Part Two. Published in Vinyar Tengwar 46, July, 2004.
See also the general Tengwestië Bibliography.
- Lambengolmor list message 620 (Jan. 22, 2004)
- Lambengolmor list message 728 (Aug. 30, 2004)
