The Quenya possessive case is formed with the suffix -va (or -wa) and is somewhat equivalent to English’s possessive suffix denoted by an apostrophe-s (’s). One of its main functions is to indicate possession: róma Oroméva “Orome’s horn” indicates a horn that (currently) belongs to Orome, whereas a genitive róma Oromeo “horn of Orome” would indicate a horn that originated from Orome but is not necessarily currently one of his possessions (WJ/368).

This was not the only function of the “possessive” case. It was sometimes used in adjective formation, and this was its original function. For example, Eruva can serve as a possessive “God’s” but it also means “divine” (VT44/18). Thus the “possessive” also has an attributive function: alcar Oroméva “Orome’s glory” indicates this glory that is a permanent attribute Orome, whereas alcar Oromeo “glory of Orome” would indicate glory originating from Orome in a specific moment only. Genitive aran Eldaron would have the connotation “king of the Elves, the one among the Elves who is king” whereas aran Eldava would mean “Elvish king, the king who is Elvish” (WJ/369, 407).

Tolkien discussed the use of the possessive suffix in his Quendi and Eldar essay from around 1960, where he labeled it the “possessive-adjectival” case. The relevant quote (WJ/368-369) appears in full in the entry on the genitive. Tolkien also discussed its original function in Common Eldarin: Noun Structure from the early 1950s:

In Q. it forms adjectives which function as possessives: as Ulmóva, Ulmo’s. They are also often more widely used of a genitival relation that could be expressed by “loose composition”, except that the latter cannot (in Q.) be any longer employed when the qualifying noun is itself qualified, e.g. by an adjective or genitive. Thus Tyalie mar could be used for “House of Mirth”, but also Mar Tyaliéva, and obligatorily Mar Vanwa Tyaliéva “House of Past (or Departed) Mirth”: one of the names of the House of Elrond in Imladris.

NB. as noted above, though Ulmóva can mean “Ulmo’s, of Ulmo (a person),” no personification of Mirth is implied in the name cited above. In some cases the meanings of singular -o coalesced with -va: as e.g. in kirya tyulma “a ship-mast, ship’s mast, mast of an unspecified or any ship”; tyulma kiryo, “the mast from some ship, of some ship”; tyulma i kiryo, “the mast of the ship” or i tyulma kiryava (PE21/79-80).

As indicated by this quote and also in the Quendi and Eldar essay, the line between the possessive and the genitive cases was gradually blurred to the point where they were frequently interchangeable, especially in Exilic Quenya (Tarquesta). The Namárië poem, for example, has a number of uses of the genitive and possessive that were “improper” according to their original (Classical) function (PE17/76). For purposes of Neo-Quenya, most authors simply use the possessive case where English would use apostrophe-s and the genitive case where English would use “of”, but even English is ambiguous in that respect: “the soldier’s sword” and “the sword of the soldier” mean essentially the same thing.

The possessive in phrases: Unlike the genitive, which can appear either before or after the word it modifies, the possessive usually appears after, though there are exceptions, especially in titles: Fëanoreva Tengwasse “Feanorian Alphabet” (PE22/149), Eldarinwe Leperi ar Notessi “Eldarin Fingers and Numerals” (VT47/14). Also unlike the genitive, it seems that the possessive is immune to the “last declinable word” rule, and can only be applied to a noun (or perhaps to an adjective being used as a noun). There are, however, examples of the possessive being applied to nouns that themselves have modifying adjectives: aran linta ciryalíva “king of swift ships, [lit.] the swift ship’s king”, ómo i·aire táríva “of the voice of the holy queen”.

As further evidence of this, the possessive behaves as if it were an adjective, in the sense that it is declined to agree with the plurality of the noun it modifies (the possession). See the example Eldarinwe above, which is a pluralized form of the possessive Eldarinwa “Eldarin’s” in agreement with the plural nouns Leperi ar Notessi. Another example is quenderinwe coar “Elvish bodies, bodies of Elvish kind” (PE17/175). Thus the possessive adjective has two dimensions of plurality:

• Plural possession: macili Eldave “the Elf’s swords”.
• Plural possessor: aran Eldaiva “the Elves’ king”.
• Both plural: macili Eldaive “the Elves’ swords”.

The possessive and definiteness: The possessive can have some complicated interactions with the definite article. See the entry on definiteness and the genitive/possessive for further details.

Forming the possessive: The base possessive suffix is -va after vowels and -wa after consonants. For consonantal nouns where the -wa cannot be used to form a suitable cluster, either (a) a joining vowel e is used or (b) the suffix reduces to -ua. As an example of the first, consider Fëanoreva from (Sindarin) Fëanor (PE22/149); as an example of the second see mantua < mantwa (PE19/100), though Tolkien did not mark this form as a possessive.

For possessive plurals the suffix is added to the plural form, except with vocalic nouns where the form -iva is added to the stem; the possessive preserves the more archaic i-plural for vocalic nouns. For the dual, presumably -va is added to u-duals and -wa to t-duals, but neither are attested in later writings. The only declension chart with a distinct dual possessive is the Entu, Ensi, Enta Declension (EEED) from the 1930s where it is simply the dual form with the suffix -va added: assu vs assuva, endu vs endúva (VT36/8).

We don’t know how the possessive plural was formed for e-nouns and consonantal nouns in the 1950s and 60s. It is very likely the -va suffix is added to the i-plural form, but it may also be that the í lengthened as well, as first suggested by Helge Fauskanger (HFQC/Lesson 12). These forms appeared for e-noun and consonantal plural possessives in declension charts from the 1930s (PE21/50-52), which lends weight to this theory.

Likewise, adding -va to a u-dual might result in a long ú for similar reasons, as indicated by the EEED declension noted above: endu vs endúva (VT36/8). The Plotz declensions also indicates that lengthening happened when the suffix -va was added to the partitive plural, but this is the expected result from prosodic lengthening.

To summarize:

• The possessive is formed by adding -va to forms (singular or plural) ending in a vowel.
• The possessive is formed by adding -wa to forms (singular or plural) ending in a consonant.
• Where the addition of -wa produces an unsuitable cluster, the suffix becomes either -eva or -ua, the latter probably more archaic.
• For the plural possessive of vocalic nouns, the suffix -iva is added to the stem.
• The final i, u is lengthened when -va is added to other plural, partitive plural or dual forms.

Possessives	Sg.	Du.	Part. Pl.	Pl.
vocalic: cirya	ciryava	*ciryatwa	ciryalíva	ciryaiva
e-noun: lasse	lasseva	*lassetwa	lasselíva	*lassíva
consonantal: atan	atanwa	*atanúva	*atallíva	*ataníva

Forms marked with a * are unattested in the 1950s and 60s, though as noted above, forms like lassíva and atanúva appeared in the 1920s and 30s.

As the chart above indicates, the possessive suffix would often trigger prosodic lengthening, and did so universally when added to trisyllabic or longer nouns ending in two light syllable such as -ie: Eldaliéva (S/59), tyaliéva (PE21/80), Oroméva (WJ/368). The possessive suffix was abnormal in that it sometimes triggered prosodic lengthening even for nouns with only two syllables: táríva (PE17/76; PE21/14), Ulmóva (PE19/60; PE21/80), Huinéva (Ety/PHUY). There are more examples of disyllabic possessives where lengthening did not happen (ciryava). One possible explanation appears in notes from the 1930s:

The treatment of medial long vowels was however complicated by many reformations. Phonetically oxytone nouns ending in a long vowel would retract forms uninflected or followed by an asyllabic inflexion (so Ulmṓ, Ulmṓn > Úlmo, Úlmōn) but the earlier accent would remain before a syllabic inflexion, gen. Ulmṓvā. But in the declensions the oxytone type became a special form limited to animates [emphasis mine]. Hence kiryā́, kiryā́va > kírya, kíryăvă after type kándā, kándāvā̀ > kandava, aided also by fact that the two coalesced in longer inflected forms as kiryālī́nen, kandālī́nen > kiryalīnen, kandalīnen (OP1 from the 1930s; PE19/60).

According to this note, the abnormal stress patterns in Ulmóva and táríva were limited to two-syllable nouns describing animate beings, while other nouns reformed to more common stress patterns. However this doesn’t explain Huinéva, and there are other examples of possessive animates without any unusual stress, such as Eldava (WJ/407) and Eruva (VT44/18).

Finally, as noted above, the possessive was treated like an adjective in that its final -a became -e when it modified a plural noun. This gives the possessive case “two dimensions of plurality” as discussed above.

Origins of the possessive: Tolkien discussed the origin of the possessive suffix in Common Eldarin: Noun Structure from the early 1950s:

A suffix -bā, wā. This as its form shows (with labial consonant) is not a primary suffix or true inflexion, and not old. It is only found in Quenya and Telerin, and only in Q. became part of the regular declensional equipment of nouns. In Q. it forms adjectives which function as possessives: as Ulmóva, Ulmo’s (PE21/79-80).

According to note #40 on PE21/79, Tolkien wrote -bā first, then wā above it, with a marginal note stating “take it from u̯ā which > vā in Q. only in long words”. This same etymology appears in a “Words Published” list (WP1) compiled in the late 1950s: “Ending -va should be from -u̯ā” (PE17/59). This derivation from -u̯ā (-wā) is consistent with its form -va after vowels and -wa after consonants, since [w] became [v] between vowels in Quenya’s phonological history.

In the Quendi and Eldar essay written around 1960, Tolkien described the origin of the two plural forms of -va:

Similarly with -va; but this was and remained an adjective, and had the plural form -ve in plural attribution (archaic Q -vai); it could not, however, indicate plurality of source, originally, and the Q distinction Eldava “Elf’s” and Eldaiva “Elves’” was a Q innovation (WJ/407 note #2).

Thus the possessive adjective agreement inflection (-ve) had the same origin as adjective agreement in general (ancient [-ai] became [-e]), whereas the pluralization of the possessor (-iva) was a later innovation, probably borrowed from other Quenya inflections like the dative plural.

Conceptual Development: The -va suffix appears quite frequently in adjectives from the Qenya Lexicon of the 1910s, and in the Early Qenya Grammar of the 1920s Tolkien described it as an “adjective suffix”.

There [is] also a general adjectival suffix -va (consonantal -uva), which is freely employed, often practically equivalent to genitive; thus tyalie “play”, tyaliéva “in play, playful, etc.” Note Mar vanwa·tyaliéva used (? for rhythmic reasons) usually in preference to mar vanwa·tyalien (PE14/47).

Thus it seems that Tolkien at first considered -va to be mainly a suffix for turning nouns into adjectives, but thought of it as a pseudo-inflection as early as the 1920s. Its regular confusion with the ordinary genitive case also dates back to this conceptual period. The suffix -va appeared as an “adjective” form in some later declension charts, but in singular forms only and thus probably not yet a true noun case (PE16/113). In the Declension of Nouns from the early 1930s Tolkien said -va also had the function of marking possession:

(9) Adjectives formed with suffixes ƀā, u̯ā, yā function in Q. as genitives possessive and relational only (PE21/3).

The extra yā primitive form is of interest, because in this conceptual period the adjective/possessive form of consonantal nouns sometimes showed -ya (PE16/113, PE21/24). In this document Tolkien introduced a distinct plural form -(l)inwa, but said the dual form was the same as the regular genitive dual suffix -hta (PE21/4, 20); this is properly a genitive suffix rather than a possessive suffix.

In declension charts after this, the possessive got a new “short plural” form -iva for vocalic nouns (PE21/43). Its regular plural form became -(l)íva (PE21/42), with -íva used for e-nouns and consonantal nouns (PE21/50, 52). At this point the declensions of -va (still called the “genitive adjective”) was more or less aligned with the forms used in Plotz, except that the consonantal forms still sometimes showed -uva instead of -wa (PE21/52). The only declension chart with a distinct dual possessive is the Entu, Ensi, Enta Declension (EEED) where it is simply the dual form with the suffix -va added: assu vs assuva, endu vs endúva (VT36/8).

The full set of conceptual developments is given in the table below, using the version numbers for the declension charts from PE16 and PE21, with EQG for the Early Qenya Grammar and LQ for Late Quenya forms (in Plotz and elsewhere). The Entu, Ensi, Enta Declension (EEED) is inserted between versions 2 and 3, and the Bodleian Declension (BD) between versions 5c and 6, as discussed in the entry on noun cases. Changes to existing forms are indicated by bold. The abbreviation Sh. Pl. = “Short Plural” for those declension charts that had archaic short plurals for vocalic nouns.

V	Possessives	Sg.	Pl.	Sh. Pl.
EQG	vocalic	-va
EQG	consonantal	-uva
EEED	vocalic	-va
EEED	consonantal	-wa
3	vocalic	-va
3	consonantal	-wa/-ya
4	vocalic	-va	-linwa
4	consonantal	-wa/-ya	-inwa
5a	vocalic	-va	-linwa¹	-iva
5b	vocalic	-va	-líva¹	-iva
5c	vocalic	-va	-líva¹	-iva
5c	consonantal	-uva/-wa	-íva
BD	vocalic	-va
LQ	vocalic	-va	-iva
LQ	consonantal	-wa	?

¹ e-nouns use -íva for plural forms.

Neo-Quenya: Given that none of the earlier declensions charts had a distinct dual form for the “adjective genitive”, this may also be true in the 1950s and 60s, so that the possessive has only singular, plural and partitive plural forms. Most Neo-Quenya authors assume the existence of a dual possessive, however, with -twa and -uva/-úva being the most common suggestions for t-duals and u-duals respectively (I personally prefer -úva for u-duals).

Note that Tolkien himself never used the label “possessive” for this noun case. Even in the 1960s he described it as the “possessive-adjectival genitive” (WJ/369). However, the label “possessive” is much more convenient, even if it is less accurate.