Like English, the direct object of a phrase generally follows the verb in Sindarin. Unlike English, the direct object undergoes soft mutation to mark it as an object, a process I refer to as (grammatical) lenition. Examples include:
- agarfant beth “he spoke words (peth)” (PE17/126).
- ar e aníra ennas suilannad mhellyn în phain “and he desires to greet there all his friends (mellon-plural)” (SD/129).
- fennas nogothrim, lasto beth lammen “doorway of the Dwarf-folk listen to the word (peth) of my tongue” (LotR/307; PE17/45).
- penim vast “we have no bread (bast)” (PE17/144).
It isn’t the mere act of following the verb that causes mutation. In Sindarin the accusative (object of the verb) is lenited, but the vocative (the person or thing being addressed) is not:
- Lacho calad! Drego morn! “Flame light (calad)! Flee night (morn)!” (UT/65).
- pedo mellon a minno “speak, friend (mellon), and enter” (LotR/305).
In these phrases calad, morn and mellon are not lenited because they are the things the commands are being addressed to rather than the object of the verb. A direct object would be lenited, as in pedo beth mellon “say the word (peth) mellon” (VT44/26).
There is some indication that the direct object would still undergo lenition even if it is displaced from its usual position after the verb. Consider the following from the Praises of Cormallen:
- Daur a Berhael, Conin en Annûn! Eglerio! “Frodo and Sam, princes of the west, glorify (them)” (LotR/953; Let/308).
The Sindarin name of Samwise, Perhael, is lenited in this phrase, and Tolkien considered several reasons why this might be the case. In notes on the word a “and” he explored the possibility that this conjunction was the source of the mutation (PE17/41), but neither of the two likeliest mutations (stop mutation or sibilant mutation) would cause p → b, something that only occurs in soft mutation. Examples elsewhere indicate that the conjunction a “and” does not generally cause soft mutation:
- pedo mellon a minno “speak, friend, and enter (minno not **vinno)” (LotR/305).
- si loth a galadh lasto dîn “*here flower and tree (galadh not **’aladh) listen [in] silence” (LB/354).
In another place Tolkien considered the possibility that Daur “Frodo” was also lenited (from unattested *Taur):
Daur. or lenition of base t. dāra, wise. Q tāra (PE17/102).
If both Daur and Berhael are lenited, the mutation is probably grammatical rather than phonological, and I think the likeliest reason is that it is because both are direct objects of the verb eglerio “praise”. David Sado suggested the same in his book, Gateway to Sindarin (GS/100). There are other examples of lenited pronouns displaced before the verb:
- im Narvi hain echant “I, Narvi, made them (san-plural)” (LotR/305; PE17/42).
- Mae g(ī)’ovannen “Well [art] thou (ci) met” (PE17/17).
In the second example, the direct object pronoun ci might instead be mutated by the preceding adverb mae, but the first is a more unambiguous lenition of the pronominal object. There are other examples, though, where direct objects displaced in a phrase are not lenited:
- ar·phent Rían Tuor·na: man agorech? “*and said Rían to Tuor: what (man) have we done?” (VT50/5).
- edregol e aníra tírad i Cherdir Perhael ... ar Meril bess dîn “in especial he desires to see Master Samwise ... and Rose (Meril) his wife” (SD/129).
In the first example, perhaps the object man avoids mutation because it is an interrogative pronoun. It’s unclear why the direct object Meril is not lenited in the second example, however; David Salo suggested it was because it was too far removed from the initial direct object i Cherdir Perhael (GS/100). Many of the attested examples of direct objects have a definite article, so that they are already mutated by the article.
It is definitely the case that a direct object need not immediately follow the verb. In addition to the examples above, there are phrases where an adverb or indirect object intervene between the verb and the direct object:
- anno ammen sír i mbas ilaurui vín “give us this day our daily bread” (VT44/21).
- ar díheno ammen i úgerth vin “and forgive us our trespasses” (VT44/21).
- i glinn hen agorer Edain mi Velerian “*this song Men made in Beleriand” (VT50/5).
This last example is unglossed and seems to be something like the passive voice, but the initial noun in the phrase (i glinn hen “this song”) is definitely not the subject, because the verb (agorer “made”) is in the plural.
Conceptual Development: There is no mention of any special behavior for the accusative in the Gnomish Grammar of the 1910s, and the few examples we have from this period show no clear signs of lenition for direct objects:
- G. on iltathi nin pieg “*he stuck in to me a pin (pieg)” (GL/51).
- G. ôni cailthi mabir gleni nan·hirilion “he kissed the slender hands of the ladies; (lit.) gave a kiss (cailthi) to the hands (mab-dative-plural)” (GG/11).
Lack of examples over the years makes it difficult to track the conceptual development of this feature.
Neo-Sindarin: For purposes of Neo-Sindarin, I would assume that an indefinite direct object is marked by lenition regardless of where it appears in the phrase, whereas a definite direct object follows the usual mutation of its article (soft mutation if singular, nasal mutation if plural). It is conceivable the actual rules are more complex, though. In Welsh, for example, a direct object is not mutated after a non-finite verb, though this is not the case in Sindarin: ar e aníra ennas suilannad mhellyn în phain “and he desires to greet there all his friends (mellon-plural)” (SD/129, given above).