Passive participles in Sindarin (and Noldorin) are reasonably well attested. For half-strong and derived verbs, its formation is straighforward: add -en to the past tense, with the final -nt becoming -nn- medially as usual. The clearest Sindarin example of this is the half-strong verb covad(a)- “(make) meet” with past tense covant, passive participle covannen “met” (PE17/16, 158). Tolkien also stated that covannen is a late Sindarin formation for archaic †covan(n). Other similar passive participles for derived verbs include:

• N. ortha- “raise” → orthannen “raised” vs. past tense orthant (Ety/ORO; PE22/31).
• N. presta- “affect” → prestannen “affected” vs. past tense *prestant (Ety/PERES).

For basic verbs, the picture is more complex. For basic verbs with nasal-infixed past tenses (that is, those whose stems end in the stops b, d, g) the pattern is similar, adding -en to the (unaugmented) past tense:

• dag- “slay” → dangen “slain” vs. past tense adhanc [or *annanc] (PE17/131, 133).
• mad- “eat” → mannen “*eaten” vs. past tense *avant (PE17/131).

The second example is interesting in that Tolkien describes the origin of this passive participle:

gp. -n-ĭna. mantinā > manthen, mannen. old aorist without n. matina > maden (PE17/131).

It seems that the Sindarin passive participle originated from an ancient past passives participle (✶mantinā) vs. the ancient aorist passive participle (✶matinā). These were (respectively) the combination of the primitive passive participle suffix ✶-inā with either the ancient past (✶mantē) or aorist (✶matĭ) stems. In Sindarin, it seems the past passive participle (mannen) took over both these functions, and the aorist passive participle (†maden) fell out of use.

However, the Sindarin passive participle is not directly tied to the past tense in “modern” Sindarin, since many basic verbs form their past tense using a vocalic augment and vowel lengthening (based more on the Common Eldarin perfect tense) whereas the passive participle is in all cases based on the ancient nasal-infixed past tense. Consider the following examples:

• tir- “guard” → tirnen “guarded” (S/168) vs. past tense *edir [or perhaps *idir] from the root √TIR.
• *hol- “close” → hollen “closed” (PE17/98) vs. past tense *ochul from the root √SKOL or √KHOL.

The passive participle tirnen is attested only in its lenited form dirnen, and the (possible) passive participle hollen “closed, shut” is attested but its independent verb form is not (though the root is attested). Nevertheless these two examples indicate that nasal-infixion or suffixion is used with passive participles even in those cases where the past tense would be based on vocalic augment and vowel lengthening. From the root √KHOL “close” the probable development is: khonlinā > hollen(a) > hollen. In her book, A Fan’s Guide to Neo-Sindarin, Fiona Jallings lists the possible participle forms for basic verbs whose stems end in various consonants (FGNS/251). Following her system:

• b → mm: heb- “keep” → hemmen “kept”.
• d → nn: mad- “eat” → mannen “eaten”.
• dh → nn: badh- “judge” → bannen “judged”.
• f (ph) → mm: *raph- “seize” → rammen “seized” [a neologism from the root √RAPH].
• g → ng: dag- “slay” → dangen “slain”.
• l → ll: hol- “close” → hollen “closed”.
• n → nn: cen- “see” → cennen “seen”.
• r → rn: tir- “guard” → tirnen “guarded”.
• th → nn: path- “smooth” → pannen “smoothed” [a neologism from the root √PATH].
• v → mm: lav- “lick” → lammen “licked”.

Given the frequent collisions in these forms, it seems possible that dissimilation would have separated the passive participles of various verb classes (e.g. perhaps for some the aorist passive participle form was preserved instead), but we have no idea whether or how this would happen.

Verbs ending in w are also a bit of a special case, in that the results would vary by whether the nasal was an infix or a suffix: nw vs. wn. The one example we have seems to indicate an infix would be used: gwanwen “departed” (WJ/378). This seems to be derived from an ancient verb: wa(w)i → wanwinā > gwanwen(a) > gwanwen; it may be the passive participle of the (very irregular) verb gwae- “go (away)” (PE17/148). But in another place Tolkien gave what seem to be “short participles” of the form gwanu or gwawn, which may either be a distinct participle formation or the equivalent of the Quenya perfect participle vanwa (PE17/148).

Another special case are derived verbs whose stems end in -nna. The passive participle of the verb onna- (appearing as an element in [N.] edonna- “beget”) seems to be onnen in the name Abonnen “after born” (WJ/387) and the passive particle of danna- “fall” seems to be [N.] dannen “fallen” (Ety/DAT); in The Etymologies the verb form is (archaic?) dant-. The expected passive participles would be onnannen and dannannen, but it seems these underwent some haplological reduction to onnen and dannen. It is unclear whether this is a general phenomenon for all derived verbs with nn in the stem.

Indeed, it not even certain whether the examples above are passive participles at all, since -en is also a common adjectival suffix in Sindarin. The adjective [N.] thoren “fenced” is sometimes assumed to be the passive participle of the verb [N.] thora- “to fence” < ᴹ√THUR (Ety/THUR), but the primitive form of the adjective is given as ᴹ✶thaurēnā, meaning it is not directly derived from the verb (which is probably from primitive *thurā-). Some of the “passive participles” given above may just be adjectives derived from the same root as the verb, making analysis of these forms more difficult. The archaic passive participle †covan(n) of covad(a)- and “short participles” gwanu, gwawn of gwae- also hint there may be complexities in the Sindarin participle system that we just can’t analyze at this time.

Conceptual Development: There were some examples of what might be passive participles in Gnomish from the 1910s:

• G. agrectha- “despise” → agrecthion “despised” (GL/27).
• G. cartha- “finish” → carthion “complete, perfect” (GL/25).
• G. fur- “conceal” → furion “concealed” (GL/36).
• G. gwidh- “weave” → gwidhon “woven” (GL/46).
• G. heb- “bind” → hebon “bound” (GL/25).
• G. hel- “freeze” → helon “frozen” (GL/48).
• G. mel- “love” → melon “beloved” (GL/57).
• G. taitha- “teach” → taithion “educated” (GL/68).
• G. teltha- “cover in” → telthion “roofed in, sheltered” (GL/71).

If so, then it seems the suffix -(i)on was used to form the passive participle in Gnomish, but it is hard to know for sure since this was a very common adjective suffix in this period.

In the Early Noldorin Grammar of the 1920s, the passive participle seems to have been formed with -ion in the first sketches of verb conjugations: glathrion, {madion >>} maidion for the verbs ᴱN. glathra- “polish”, ᴱN. mad- “eat” (PE13/126-127), but this was changed to -og or nasal-infixion in the second sketch: glathrog, madog or nasal infixed mant, dunc, the last for the verb ᴱN. dag- “slay” (PE13/129-130). Most of these appeared with past passive variants which appended the relevant suffix to the past tense: in the first sketch glathra(i)thion, manthion, then in the second sketch glathraithiog/-athig, mainthiog/manthig.

The nasal-infixed form became the past passive participle in the final sketch of basic verbs: tangant, mant (or madant), adanc, lhovn from ᴱN. tangad-, mad-, adag-, lhuv- (PE13/131-132). The present passive participle used the suffix -ianw instead: tengedianw, medianw. The past passives had plural forms tengennin, mennin (or medennin). All these were distinct from the past tense, which used the suffix -ath in this final verb paradigms from the Early Noldorin Grammar.

Finally, there were examples of passive participles in Early Noldorin word lists from this period, mostly nasal-infixed forms or with suffixal -ant, sometimes with an augment:

• ᴱN. adag- “to build” → adanc “built” (PE13/136, 158, 165).
• ᴱN. athra- “to face” → athrant, †annarn (PE13/160).
• ᴱN. cara- “to make, do” → agarn (PE13/161).
• ᴱN. crann- “finish” → agrant (PE13/161).
• ᴱN. crib- “to bend” → cremp, †crimp (PE13/141).
• ᴱN. elaig- “to shield” → alchent/alchant, ONo. elainc, alanc (PE13/158).
• ᴱN. gonod- “to count up” → gonodant (PE13/162).
• ᴱN. hug- “to copulate” → honc or hogant (PE13/147, 163).
• ᴱN. mad- “to eat” → mant or madant (PE13/163), with another mention of present passive medianw (PE13/164).
• ᴱN. maitha- “to ravish” [< ᴱ✶mapta-] → mabant (PE13/149, 163).
• ᴱN. ped- “to say” → pent (PE13/164).
• ᴱN. tha- “to cause to be” → thant, (archaic?) ast (PE13/153).

Again, these passive participles were all distinct from past forms, which were also nasal infixed but had suffixal or intruded i in this conceptual period: past maint vs. pp. mant (PE13/163). These early nasal infixed passive participles might explain the otherwise mysterious archaic Sindarin participle †covan(n) for S. covad(a)- as mentioned above (PE17/158). Conceivably such forms could be derived from the ancient root + adjectival nā, which means there should be evidence of a-affection in these participles as with ᴱN. lhovn, honc above from lhuv-, hug- (but note archaic †crimp vs. modern cremp from crib-). In any case it seems that such participles were archaic in Sindarin, and by Noldorin of the 1930s the system of passive participles used in Sindarin was already more or less established.

Neo-Sindarin: The system described above is one widely-accepted by Neo-Sindarin writers, and is basically the one proposed in David Salo’s book, Gateway to Sindarin (2004, GS/120), with the caveat that Salo assumed (probably incorrectly) that forms like thoren “fenced” and dolen “hidden” were also passive participles; the primitive form of thoren is not a passive participle as described above (Ety/THUR), and dolen is not based on a verbal root (Ety/DUL).

Unlike the active participles, it is also generally assumed that passive participles are inflected in the plural to agree with the noun they modify, based on the forms gwenwin plural of gwanwen “departed” and †oennin plural of onnen in †Eboennin plural of Abonnen “After-born” [later plurals would be ennin, Ebennin] (WJ/378, 387). We have no examples of passive participles used in a phrase, with the exception of the (possibly fossilized) greeting mae govannen “well met” (LotR/209). Most likely the passive participles are used as an adjective when the modified noun is the object of the action, as in i adan nangen “the slain (dangen [nd-]) man” or in edain nengin “the slain men”, where the man or men are the ones who were slain.

Unlike English, however, Sindarin has a distinct past active participle which might be used instead, especially for intransitive verbs. For example, I think Sindarin for “the grown maiden” would be i ’wend ’óliel using the past active participle góliel of gala- “grow”, since it was the maiden who did the growing. Using the passive participle (i ’wend ’alannen) would imply someone else was growing the maiden as if she were a plant. Thus purely temporal distinctions can be handled with past vs. present active participles (góliel vs. galol) rather than with the passive participle (galannen), assuming the noun is the one who is (or was) performing the action.

However, it seems Sindarin lost its distinction between past and present passive participles (galannen vs. †galen), having merged both functions into the past passive participle. This is not very limiting, because if the noun is the object of the action, that action is typically something that happened in the past.