It is often assumed that all Sindarin verbs ending in -a are derived verbs. However, this doesn't have to be so, if the root itself contains an -a the verb may end in -a but still be a primary verb. An example for this may be gala- 'to grow' (PE17:131,173) from the root GAL. Here, the final -a may just be an extension of the root through another insertion of the sundóma (the stem vowel). In LR:357 the relevant root is given as GALA- and in PE17:131 the 3rd person sg. shows variants gala, gâl which would support this idea.
Thus, there was a past tense formation in primitive Elvish which was characterized by 1) a prefix consisting of the stem vowel and 2) a lengthening of the stem vowel. Thus, the verb *kwe- 'to speak' was changed to ekwē 'spoke' and the verb *kar- 'to do' into akāra 'did' (the ending -a probably represents an extension of the stem - the Quenya kar- > #káre (SD:246) indicates that it was dropped subsequently).
However, this was not the only mechanism of forming the past tense for primary verbs. The form echant 'made' (LOTR) is derived in PE17:42 as echant 'shaped out' < et-kantē for edagant. The basic verb is identified as et-kat-. The form can be analyzed as the result of an ending -nē used to form past tenses in primitive Elvish with subsequent change of the position of the nasal, resulting in nasal infixion, thus *et-kat-nē > *et-ka-n-tē > echant. Such nasal infixion is well documented for Noldorin [1] and Quenya [2], likewise the ending -ne is very common in Quenya past tenses. Even more revealing is the form edagant which would appear to show the same prefix applied to a form
In order to appreciate Tolkien's choice of echant over edagant, one has to keep in mind that 'Words, Phrases and Passages' is a commentary on the published Text of LOTR. While LOTR was written, there is no sign that Tolkien ever envisioned the use of an augment in the Sindarin or Noldorin past tense. However, it is clear from the above that the augment was a crucial part of the Sindarin past tense in the later conception. Nevertheless, Tolkien could hardly alter the published text, thus he needed to find an explanation for the lack of the augment for each published form.
Further light on the augment is shed by PE17:93 where we find with regard to past tenses for the verb derived from the root MEN as MEN have as object, (in)tend, proceed, make for, go towards: [Q:] menta-, but past tense mennē-, perfect emēnie: in Sindarin these were blended into a past tense form emēnē-.
The implication seems to be that the augment survived in the Quenya perfect tense but not usually in the Quenya past tense (this is again confirmed in [2]) but that these forms blended into the Sindarin past tense (no distinct perfect tense for the Sindarin verb is known), leaving it with an augment and vowel lengthening (which is also characteristic of the Quenya perfect tense). Note that the given form can hardly be Sindarin but is probably rather Old Sindarin - the standard sound shift would create *evín 'went towards'.
With regard to the past tense of derived verbs, PE17:93 reveals that Sindarin verbal history is complicated and the use of -ant as 3rd person past tense ending is rather like that of Medieval Welsh -as or modern Welsh -odd. In other words, the Sindarin past tense of derived verbs cannot readily be understood from primitive Elvish developments but is rather an analogical later development.
There are also a few instances of past tenses by nasal infixion:
An interesting example is provided by
From these two examples, a suffix -nt (consistent with a weak past tense suffix found for derived verbs in Noldorin [1]) can be identified. This is the same suffix which by analogy leads to weak past tenses carfant, agarfant (see above).
Tolkien describes this in PE17:126 in more detail: -nt Sindarin past tense of transitive verb, -ir of intransitive agarfast 'he talked' agarfant beth [perhaps read agarfant bith] 'he spoke words'. The statement about -ir is somewhat puzzling as the example immediately below shows an ending -st in intransitive use. The latter ending reminds of the ending -s seen in weak intransitive past tenses in Noldorin [1].
Another example may be seen in unidentified forms in PE17:167 where to the verb raitha- '*to try' forms rithant, rithas are given which could represent transitive 'to try something' and intransitive 'to strive'. A following rithessin could then be taken as the first person 'I strived'. As the root is given as RIK, the pattern certainly agrees with what is expected for a derived verb. Hence perhaps
The last attested example shows a strong past tense for a primary verb: anta- has in Quenya the normal regular 'weak' past tense of such verbs (formed with transitive (accusative) -tā), antane. But in Sindarin is found a 'strong' past tense formed on the analogy with verbs using intransitive -tă only in the present/aorist (..) The S[indarin] antha- > anha 'give' pa.t. ōn- older ānē- [the sentence ends without a verb] (PE17:93) (the verb is also given as S: anta- in PE17:147 which seems an odd form given Sindarin phonology). Thus
However, there is a class of verbs which are derived with a suffix -tă which appears only in present tense and aorist. These verbs form a strong past tense in which the suffix is lost. Thus, had the verb been **ana-tă, its primitive past tense would have been **ānē and this form by analogie became the actual past tense. This would reveal much about the past tense of other derived verbs, if we could be sure about the precise derivation.
While all verbs with strong past tenses show an augment unless they start already with a vowel (with the possible exception of carfant which may have lost the augment by analogy along with the original past tense), none of the derived verbs with weak past tense shows an augment. The (Old Sindarin?) form emēnē as the past tense to *men-tă indicates that derived verbs can have an augment if they show a strong past tense.
The chief past tense formation mechanism for derived verbs seem to be endings -nt (which appears as such in Noldorin for transitive verbs and -st/-s (which has a Noldorin counterpart in -s) for intransitive verbs. This is indicated as an analogical development from frequent strong past tenses by nasal infixion.
The fact that as many as two analogical past tenses appear in a sample as small as the one presented here indicates that analogical past tenses play a considerable role, but given the relatively small sample of verbs, the precise impact is difficult to gauge.
[2] The Q(u)enya Past Tense by Thorsten Renk
[3] 'Words, Phrases & Passages in The Lord of the Rings' by J.R.R. Tolkien, Parma Eldalamberon 17, edited by Christopher Gilson
Attested past tenses of primary verbs
Attested past tenses of primary verbs are mostly formed with vowel lengthening. We find:
These forms can be explained assuming primitive a-kāra, *o-nōre, *e-dēwe and *a-sāve > *asauve > ahawv.
A few forms seem to have the augment, but no apparent vowel lengthening (note though that the forms are from extremely rough notes written on top of each other, thus the discrepancy should not be overstressed:
Finally, gala- shows a past tense with vowel lengthening, giving additional support to the idea that it is in origin a primary verb:
In PE17:131 also the analogical past tense angol is found found as alternative to aul. The analogy seems to be with verbs derived from stems in NG, thus the form can be explained as *a-ngāle > *angaule > angol (which may be a plausible past tense of gal- 'to shine' (PE17:163) which is from such a root).
The last example requires some explanation. It is derived from the root AWA, WĀ and the present tense gwaen presumably comes from *wā-i-njē. The past tense however is derived from the inverted root AW(A), thus *aw-nē > awn or with nasal infixion *aw-nē > *anwē > *anw > anu and 1st person *aw-nē-njē > anwen.
This would seem to show an augment with the past tense suffix -nt of a derived verb (see below), but we will argue that the verb is in fact a primary verb in origin. The underlying root is given as KARAP and crossed out forms show a past tense carfant (agramp) . The last form is interesting in the light of the Quenya past tense karampe. Plausibly both the Sindarin and the Quenya form can be explained by a primitive *a-karap-nē. In Quenya, the augment as lost in past tense (although not in perfect tense, see above), but in Sindarin the subsequent development would have been *a-karap-nē > *a-k'rap-nē > *akrampē > agramp whereas in Quenya *a-karap-nē > *karap-nē > karampe. Thus, the present tense ending -a would reflect an extension of the sundóma rather than a derivational suffix. The past tenses carfant, agarfant would then be later analogical developments coined after the fashion of true derived verbs.
Attested past tenses of derived verbs
Unfortunately, the attested number of derived Sindarin verbs is rather small. We have with a suffix -nt
The last example is given in two variants, cova- and covad-, the last of which rather reminds of a primary verb, but the root is clearly given as KOB/KOM, thus some derivational suffix is required in any case. Likewise, cova- cannot plausibly be explained by an extension of the root by a sundóma as this would result in **covo-.
The origin of a suffix -nt seems to be by analogy with frequent strong past tenses for primary verbs with nasal infixion, cf. the -nt created in echant. Tolkien's remarks to -ant having parallels in Welsh past tense endings on PE17:43 support this idea.
From the above paragraph, the following information can be deduced: anha- is derived from a root ANA, NĀ using a suffix -tā presumably as *ana-tā > *an'tā > antha > anha. As derived from this particular suffix, it would have had a regular past tense *ana-tā-nē leading to Quenya antane and presumably Sindarin *anhant. Discussion
The past tense formation of Sindarin largely seems to be a continuous development from the Noldorin past tense, the chief new element being the augment seen for strong past tenses. Primary verbs in Sindarin (as in Noldorin) show two main ways of past tense formation - lengthening of the stem vowel (for final consonants -l-,-r-, -v-, -w- and in one case -k-) and nasal infixion (for final consonants -k-, -p-, -t- and -w-. As seen from carfa- with past tenses carfant and agramp or from #dag- with past tenses aðag, aðanc a verb may have more than one possible past tense.References
[1] The Past-Tense Verb in the Noldorin of the Etymologies by Carl F. Hostetter.
Thorsten Renk
thorsten@sindarin.de