The half-strong verbs use of the same inflections as weak verbs, but have strong past tenses formed by modifying the verb stem. Most verbs with half-strong conjugations are the formative verbs, which are the result of adding the suffixes -t(ă) or -y(ă) to a root. They are most commonly used with roots that cannot be verbs by themselves, such as: orta- “rise” from the root √OR “up(wards)” (PE22/114, 164). They are also used with verbs whose roots end in y or w that would otherwise be very awkward to conjugate: caita- “to lie (down)” (PE22/157), [ᴹQ.] lauta- “to abound” (PE22/103). The formatives are frequently intransitive verbs, which sometimes leads to variant transitive/intransitive past tenses: ortane “raised” (transitive and weak) vs. oronte “rose” (intransitive and half-strong).
One comprehensive set of conjugations for formative verbs appears in Late Notes on Verb Structure from the late 1960s (PE22/164), but many of these appear in slightly different forms in notes from the mid-1960s (PE17/77). The inflections for ta-formatives and ya-formatives are also not quite the same as each other:
- Stem: ista- “to know”.
- Aorist: ista “knows”.
- Present: istia (or istya) “is knowing” vs. istea*.
- Past: (i)sinte “knew”.
- Perfect: (is)istie or isísie “has known”.
- Future: istuva “will know”.
Compare this to:
- Stem: sirya- “to flow”.
- Aorist: sirya “flows”.
- Present: síria “is flowing” vs. siryea (or sirea)*.
- Past: sirinye “flowed”.
- Perfect: isírie “has flowed”.
- Future: siryuva “will flow”.
* See below for the discussion of the -ia vs. -ea present tenses.
Past Tense: The past tenses of many half-strong verbs were the result of inserting the nasal n before the formative suffix, a process described in the 1940s Quenya Verbal System:
This was originally formed “strong” with n-intrusion before the last consonant: as sirya: sirinye “flowed”; talta: talante. Where the stem was of √AT type this past could be made from ’ta-form: as orta: *rontē “rose”; ista: sinte “knew”. But few examples — only sinte “knew” is common — survive; the usual form is oronte “rose” (PE22/115).
Thus, typical half-strong past tenses were formed as sirya (< ✶sir-yă) → sirinye (< ✶sir-(i)-n-y-e), where the past tense inflection n-e was split before and after the formative suffix and an extra base vowel was inserted for pronounceability. For verbs made from invertible roots like ortă and istă, the past tense might instead be made from the inverted root, as noted above: ronte or sinte. Of these alternate forms, only sinte “knew” was widely used, so these variants can be treated as irregular.
Some half-strong verbs formed their past tenses by dropping the formative suffix entirely and inflecting the root as if it were a basic verb. This is most common with formatives produced from roots ending in Y or W: caita (< ✶kay-tă) → past caine (< ✶kay-ne); [ᴹQ.] lauta → past laune (PE17/72; PE22/103). This pattern is occasionally seen in other verbs forms like menta → past mennē (PE17/93). These root-based past tenses were probably favored, especially for Y and W roots, because the normal inflections could produce past tenses heavily obscured by various phonetic changes: ✶kay-tă → past ✶kayante > ceante (PE22/164), past ✶áwa-n-tē > oante (WJ/366).
Perfect Tense: The perfects of formatives tend to replace the final a with the perfect suffix -ie but otherwise follow the pattern of basic verbs: adding a base-vowel augment and (if possible) lengthening the stem vowel: nahta → anahtie, sirya → isírie. Such base-vowel lengthening is not possible for ta-formatives, but in ya-formatives the suffixal y merged with the perfect suffix -ie. In effect, the perfect was made from the root (√SIR) and vowel lengthening was possible. In some cases ta-formatives also have perfects derived from the root: compare (is)istie vs. isísie, two variant perfects of ista- (PE22/164).
There are also examples of half-strong verbs whose perfects are formed directly from the past tense, similar to weak verbs: caine → acainie perfect of caita- (PE22/157) or sinte → isintie, which is yet another variant perfect for ista- (PE17/77). Both these examples have variants fitting more normal patterns: isistie and acaitie, so they may represent irregular forms or conceptual vacillations.
Present Tense: In Late Notes on Verb Structure from the late 1960s, Tolkien gave some rather surprising present tenses for formative verbs with suffixal -ia: ortia, nahtia, síria (PE22/164). Present tense forms like these are seen nowhere else in the published corpus. Tolkien gave the origin of these forms as [or]tiyan > ortian, apparently from a primitive present suffix -iya. Elsewhere Tolkien said the primitive present tense suffix for derived verbs was ✶-ayā (PE17/77, 186) which became -ea through normal phonetic developments. Indeed, ā́ya appeared in the very same page the ia-present forms (PE22/164 note #103). This page also included a hard-to-read statement saying something like “make Q. ea as present tense invade other forms”.
In earlier notes from the mid-1960s, the present tense for formatives used the suffix -ea (PE17/77), so it seems this ia-presents were a new concept in the late 1960s. It is not clear whether Tolkien intended these ia-forms to become the new present tense for formatives, whether he imagined they were archaic and replaced by ea-forms, or whether this was all a transient idea.
Conceptual Development: There are past tense forms in both the Qenya Lexicon (QL) of the 1910s and the Early Qenya Grammar (EQG) of the 1920s that resemble the strong past tenses of ta-formatives:
- ᴱQ. olto → past olonte (QL/69).
- ᴱQ. silte → past silinte (QL/83).
- ᴱQ. lokta → past lokante (PE14/58).
- ᴱQ. tantilta → past tantilante (PE14/58).
The QL past tenses insert a base vowel (olonte, silinte) but the EQG past tenses seem to always insert a (lokante, tantilante). Not every Early Qenya ta-verb falls into this pattern; some have weak pasts or other irregular past tenses instead, such as: fanta → past fantane (QL/37); mukta → past mūke (QL/63).
As noted by Thorsten Renk in his article on the Quenya Past Tense (QPT), verbs with -ya in QL frequently have past tenses ending with -sine:
- ᴱQ. angaitya → past angaisine (QL/34).
- ᴱQ. kitya → past ekitsine (QL/47).
- ᴱQ. mauya → past mausine (QL/60).
- ᴱQ. naitya → past naïksine (QL/65).
- ᴱQ. paitya → past paisine (QL/72).
- ᴱQ. palasya → past palastine (QL/72).
- ᴱQ. pietya → past pieksine (QL/73).
- ᴱQ. sutya → past suksine (QL/87).
Most of these verb stems have ty, so it seems these developed as –ty+ne > –tine > –tsine where [ti] became [tsi] following the normal phonetic developments of the 1910s. Past tenses of other Early Qenya ya-verbs don’t fit this pattern, however:
- ᴱQ. linya → past lintine (QL/54).
- ᴱQ. listya → past listine (QL/55).
- ᴱQ. mirtya → past mirtine (QL/61).
- ᴱQ. orya → past orīne (QL/70).
- ᴱQ. panya → past pantine (QL/72).
- ᴱQ. purya → past purīne or pustine (QL/75).
- ᴱQ. qilya → past qīle or qīlline (QL/77).
- ᴱQ. tenya → past tentine (QL/91).
- ᴱQ. tintya → past tintine (QL/92).
- ᴱQ. turya → past tustine (QL/96).
- ᴱQ. winya → past wintine (QL/104).
Since we don’t know much about Early Qenya verb classes, it is hard to tell whether these variations represent distinct verb classes, conceptual vacillations or other irregularities. Tolkien began distinguishing the formatives (-tă, -yă) from the causatives (-tā, -yā) in the Quendian & Common Eldarin Verbal Structure written in the 1930s, though the terminology was not yet clearly established (PE22/98). A distinct class of half-strong formative verbs also appeared in the Quenya Verbal System (QVS) of the 1940s (PE22/113-115), already with past tenses like sirinye as described above.
In QVS, Tolkien lumped these formative “half-strong” verbs together with talat-stem verbs since their conjugations were nearly identical. In later documents their conjugations were more distinct, especially in the Late Notes on Verb Structure from the 1960s. See the discussion in the entry on talat-stem verbs for more details on the conjugations of that verb class.
Neo-Quenya: For purposes of Neo-Quenya I would ignore the ia-present forms from PE22/164 and assume they were either archaic or transient ideas. I would use the ea-present forms from PE17/77 instead, consistent with other derived verbs. I would, however, retain the half-strong past and perfect forms, as these are well established. Of course, one caveat is that it is not always possible to determine which verbs are half-strong and which are weak. One possible rule of thumb is that any ta-verb or ya-verb that is not a causative or based directly on a noun or adjective is probably a half-strong verb.