Both English and Quenya make use of “impersonal” verbs, which are verbs with no determinate subject. An obvious example in English is “it rains”. The verb “rains” describes the entire action, and there is not really any entity performing the act of “raining”. English addresses such situations by using “it” as an empty subject, because English cannot have a sentence without a subject. Quenya handles impersonal verbs somewhat differently, by omitting the subject entirely. Thus “it rains” in Quenya is simple the impersonal verb ule. This impersonal verb construction was described in Common Eldarin: Verb Structure (EVS2) written in the early 1950s:

Where no subject was expressed or where the action was strictly impersonal, as in: “(it) rains, (it) seems”, the bare stem of the verb or tense was used without inflexion (PE22/128).

Quenya uses such impersonal verb constructions much more frequently than in English, including many situation where English would have a determinate subject. For example, Quenya has an impersonal verb ek- which means “there is a chance, opportunity or permission”, and it is used in situations where English might use the modal verbs “can” or “may”. Where in English you would say “I can do it”, in Quenya you would say eke nin kare sa, more literally “there is an opportunity for me to do that” (VT49/20). Alternately, in Quenya “I may do that tomorrow” is ekuva nin kare sa noa, literally “there will be a chance for me to do that tomorrow”. Other impersonal constructions include:

• mára ná “[it] is good” (PE17/93).
• mára tyen [ná] “[it] is good to you = you like it” (PE22/166).
• ore nin karitas “[it] urges me to do it = I feel an urge to do it” (VT41/13).
• *óla nin “[it] dreams to me = I dream”.

In the last example, we know óla- “dream” is an impersonal verb (UT/396), but we don’t know exactly how it is used. However, in all of Tolkien’s later examples of impersonal verb construction, the putative “subject” of the action is put into the dative, as with nin “to me” and tyen “to you” in the examples above. It is reasonable to assume this is the common practice. It also says something interesting about Elvish psychology that urges and dreams are viewed as events acting on the subject rather than actions of the subject.

One challenge in identifying impersonal constructions is that Tolkien also used the term “impersonal” for the verbal suffixes -r or dual -t used for subject-verb agreement, as in elen atta siluvat “two stars will shine” (VT49/44) or lassi lantar “leaves fall” (RGEO/58). Many of Tolkien’s charts of pronominal subject suffixes included a row for these two “impersonal” verb inflections, and it seems Tolkien just used the same term for different things: impersonal inflections (used for subject-verb agreement) vs. impersonal verbs (for subjectless verbs). However, there is at least one example of a “plural impersonal” phrase without an explicit subject: quetir en “they still say” (PE17/167), so it seems the two concepts were not entirely disconnected.

Certain verbs require impersonal (subjectless) constructions, like uluva “it will rain” (PE22/167) or hrinse “it snowed” (PE17/168). However, it seems that Quenya can use impersonal constructions more generally with any verb, as a way of expressing passive voice. This passive voice construction was discussed in the Quendian & Common Eldarin Verbal Structure (EVS1) written in the late 1940s:

This has the result that the normal Eldarin verb expresses both the passive and the active: e.g. √MAT “eat”: aorist stem mati; 1 pers. matini (agent unspecified) “one eats me; I am eaten” (PE22/93).

In the late 1940s, object suffixes were always appended to the verb, so in Tolkien’s later syntax this might become mate ni “[it] eats me = I am eaten”. According to this paradigm, a sentence that would be in passive voice in English, such as “the bread was eaten”, would in Quenya use an impersonal verb construction: mante i massa “was eaten the bread” or more literally “[it] ate the bread”.

In cases where a subjectless verb was ambiguous or confusing, Quenya also had an indefinite pronoun that could function like “it” or “one” in English. This pronoun was first described in Quendian & Common Eldarin Verbal Structure (EVS1) from the late 1940s, where the form was a:

In consequence, if there is only a dative pronoun and the subject is not expressed, the indefinite subject [CE ga, Q a] was used. ni·antā = “I give”; e·ni·anta = “he gives (to) me”; but ga·ni·anta·t(e) = “one gives me it, I am given it” (PE22/94).

Examples of this indefinite a appear in passive voice constructions in the contemporaneous Quenya Verbal System (QVS), and in fact these make up the majority of “passive voice” formations in QVS:

• ᴹQ. (a)·matis “it is eaten (one eats it)” (PE22/107).
• ᴹQ. i·nér né raiqa ar sí aphastat “the man was angry but now is in good humor (lit. it pleases him)” (PE22/124).
• ᴹQ. vahaia nóre ëa i a-esta Valinor “far away (there) is a land called Valinor” (PE22/124).
• ᴹQ. masse akime aldar? “where are the trees (to be found)” (PE22/125).

This indefinite pronoun is distinct from the contemporaneous 3rd singular pronoun ᴹQ. te as in ᴹQ. te ye matina “it is eaten”. In Late Notes on Verb Structure (LVS) written in 1969 Tolkien updated this indefinite noun to e:

e, neuter, is an indefinite pronominal element corresponding to E. “one” (neut[er] “it, a thing”); a later development, to make the purely impersonal verb form more precise and dist[inguish] it from the bare infinitive, as is shown by its preceding the verb (PE22/154 note #57).

But this paragraph was rejected and the references to e were changed to mo so that the relevant section became:

To express the advice in general “aorist” terms one must use the separate negative: alasaila ná lá kare tai mo nave (or navilwe) mára “it is unwise not to do what one judges (or we judge) good”.

mo, indefinite personal pronoun “somebody, one”.
ma, neuter personal pronoun “something, a thing”.

For the indefinite mo the inclusive 1.pl. = “we, you and I (and others associated)” was often used, espec. colloquially, like E. “you” (PE22/154).

Thus it seems the indefinite pronoun a >> e >> mo. As in the 1940s, sentences with mo in the 1950s and 60s are more common than purely impersonal phrases. Other examples include:

• an ké mo querne kendele númenna “for if one turned the face westward” (VT49/8).
• ké mo qete ulca “*if one speaks evil” (VT49/16).

As for impersonal expressions with verbs that are not inherently impersonal, in later writing we have only the “plural impersonal” phrase quetir en “they still say” (PE17/167), as mentioned above.

Passive voice with the passive participle: Note that in English (and Quenya) avoiding an explicit subject is not the only use of passive voice. Sometimes passive voice is used to reverse the agent and the undergoer of an action, making the undergoer the grammatical subject and expressing the agent in some other way. In English you can say “the bread is eaten by the man”, which is the passive voice equivalent of “the man eats the bread”. English forms its passive voice by using its past (passive) participle, in this case “eaten”. There are a couple of examples of similar constructions in Quenya:

• a Túrin Turambar turun’ ambartanen “[oh] master of doom by doom mastered” (S/223; UT/138).
• nahtana ló Turin “*slain by Túrin” (VT49/24).

Neither of these are complete phrases, but in both cases we have a passive participle (turun[a] “mastered” and nahtana “slain”) followed by the putative subject of the verbal action (ambartanen “by doom” and ló Turin “by Turin”), expressed by either the instrumental or the preposition ló “by agency of”. This looks very similar to English passive voice constructions such as “Turin was mastered by fate” or “Glaurung was slain by Turin”. It is possible such expressions could be used in Quenya with or without the copula ná-: Turin turún’ ambaranten (ná) or Glaurung nahtana ló Turin (ná).

A similar “passive voice” expression with a passive participle appeared in QVS from the 1940s as well: te ye matina “it is eaten” (In QVS, ye- was the verb “to be”), which is contrasted with impersonal passive voice: (a)·matis “it is eaten (one eats it)” (PE22/107). Of this Tolkien said “the second expression is the genuinely verbal one pointing to the action, the other is adjectival”. This lends further support to the idea that English-style passive voice expressions using passive participles might be possible in Quenya.

Conceptual Development: Impersonal verbs were mentioned all the way back in the Qenya Lexicon of the 1910s, including some words that do not seem to be inherently impersonal:

• ᴱQ. itisya- “to itch, irritate (impers.)” (QL/43).
• ᴱQ. loyo- “impers. be thirsty” (QL/56).
• ᴱQ. minty- “to remind; impers. it reminds me = I remember” (QL/62).
• ᴱQ. mart- “impers. it happens” (QL/63).
• ᴱQ. nasa- “taste nasty; impers. = I dislike” (QL/64).
• ᴱQ. tyava- “impers. it tastes of, reminds one of” (QL/49).
• ᴱQ. varka- “it bodes, impersonal. Also personal, I dread, I fear (of future things)” (QL/102).

In The Qenya Verb Forms of the 1910s, Tolkien gave distinct active voice, passive voice and medial voice (reflexive) forms of the verb (PE14/28-30). In the Early Qenya Grammar (EQG) of the 1920s, Tolkien distinguished impersonal verbs from passive voice:

-r for the impersonal (distinct from the endingless form, e.g. uqe “it rains”, but tulir “one goes, somebody goes”): this becomes a passive if pronominal elements are added, for these are in the accusative (rarely dative). In the first case (accusative) these still may retain (and always do so in archaic language) their accusative-position after the verb [presumably matsir ha “is eaten it”], but as the passive feeling has increased such expressions as ha·matsir “it is eaten” are not unusual.

In EQG is seems that -r was the marker of passive voice for verbs; in this document -l was used for plural verbs in agreement with a plural subject. In EQG passive voice formations, the object of the action could become the grammatical subject, as in ha·matsir “it is eaten”; ᴱQ. ha is the ordinary 3rd singular neuter pronoun “it” in EQG.

The notion that the impersonal verb form could function as passive voice was introduced in Quendian & Common Eldarin Verbal Structure (EVS1) written in the late 1940s, as discussed above, along with the optional use of an indefinite pronoun a. Examples of this formation also appeared in the contemporaneous QVS. In LVS written in 1969, the indefinite pronoun was changed to e >> mo (PE22/154).

Neo-Quenya: In his Quenya course Quetin i Lambe Eldaiva (QLE), Thorsten Renk proposed two possible ways of forming passive voice in Quenya (QLE/129). The first way was to use a passive participle much like English (and many other European languages): i massa matina (ná) “the bread is eaten”. The second way was to use an impersonal verb formation like mate i massa “eats (is eaten) the bread”. Both formations can be put into the past and future by inflecting the verb appropriately: i massa nauva matina “the bread will be eaten”, mante i massa “ate (was eaten) the bread”. Of the two, I think the second formation (impersonal verb) is preferable when the agent of the verb is completely unexpressed.

If the agent needs to be expressed, it is put in the instrumental in conjunction with passive participles: i massa matina i Eldanen (ná) “the bread is eaten by the Elf”. For impersonal verbs, the dative is used for the putative subject, but that seems to only be the case for intrinsically impersonal verbs which have no determinant subject: eke nin kare sa “there is an opportunity for me to do that” (VT49/20). For “generic impersonal passive voice”, it is likely no putative subject is possible (or rather, if you want an explicit subject, you can only use active voice: Elda mate i massa “the Elf eats the bread”).

Also of note is the fact that the indefinite pronoun mo can be used to avoid specifying a determinant subject. This is probably most common for general or proverbial statements: mo mate massa qui mo pene apsa “one eats bread if one lacks meat”. Based on the examples in QVS and LVS, it seems mo is often added as the phrase increases in complexity to avoid confusion: alasaila ná lá kare tai mo nave mára “it is unwise not to do what one judges good” (PE22/154). Colloquially, the inclusive “we” can be used the same way (PE22/154), much as English sometimes uses “you”: “you eat bread if you don’t have meat”; in Q. such expressions would use -lve “we (inclusive)”: matilve massa qui penilve apsa.

It also seems that “plural impersonal” expressions are also possible (PE17/167), as in quetir Eldar vanime (nar) “they say Elves are beautiful”, but an indeterminate rather than specific “they”.