In addition to the “general” plural, Quenya has a special “partitive” plural used when describing a portion of a group. The partitive plural is formed using the suffix -li, variously translated as “some, many, a lot of” (PE17/62, 127, 135; VT47/12). It is derived from the primitive root √LI “many” (Ety/LI, VT48/25). Tolkien described the use of this suffix on several occasions:
The Eldarin languages distinguish in forms and use between a “partitive” or “particular” plural, and the general or total plural (Let/178, in a 1954 letter to Naomi Mitchison).
Both Quenya & Sindarin have for most nouns two plural formations: the general or group, and the partitive or special. The plural element in nouns is [i] as a suffix [ī]. Plurals formed with this are in Quenya “general”, and the partitive is formed with -li (PE17/62, draft of a 1955 letter to David Masson).
Both languages had a 2nd plural formed with a suffix. Q -li, i-falma-li-nna-r, “the-foam wave-many-towards-pl. ending”, an i falmalī ... Q elen, eleni, elelli (PE17/127, undated linguistic notes from the late 1950s or early 1960s).
... the two plural forms, the one casual and indefinite, the other particular and complete; so Eldar “Elves as a kind or people, or all the Elves concerned”; Eldali “some Elves, a lot of Elves” (PE17/135, in a 1964 letter to W. R. Matthews).
Tolkien called this form the partial, particular or partivive plural; of these, the last is most commonly used in the literature about Quenya. One of the longest descriptions of this form appears in notes on Common Eldarin Noun Structure from the early 1950s:
A curious feature of Eldarin, in some degree common to all branches, though the means of denotation differ, is a tendency to distinguish between two kinds of plural: the group-plural (or definite); and the partitive or indefinite. This agrees with the distinction in the dual (see below) between the dual of natural pairs and the numeral dual. Usually it was the old inflexional elements described above that produced the definite plural. The other was expressed by affixes of later origin. Definite were plurals referring to whole classes, to things naturally or habitually considered in plurality (as English heavens = “the sky”, the sands = “all the sand in a given locality”, etc.), and in the syntax of many languages a plural with a definite article, meaning all the members of a group previously mentioned, or in mind. Thus in Q. Eldar (not with article!) = “Elves, The Elves, All Elves”; i Eldar = “(all) the Elves previously named” (and in some cases distinguished from other creatures); but Eldali “Elves, some Elves”. With Eldali the definite article is seldom used (PE21/73).
Following this usage, perhaps the simplest equivalent to the suffix -li is the English word “some”, in the sense of “some but not all”. If you were to say Eldar mate massa “Elves eat bread”, you would be implying that all Elves eat bread, but if you were to say Eldali mate massa “some Elves eat bread” you would be implying that a portion of Elves eat bread, but another portion of Elves do not. In this context:
- Eldar “[all] Elves” (general plural) > Eldali “some Elves” (partitive plural)
As noted above, the definite article is not typically used with the partitive plural. In some respects it functions like a plural indefinite article, much how “some” can be used in English: “a man, some men”. But there is at least one example of the partitive plural being used with a definite article, and this seems to have an idiomatic influence on its meaning:
- i falmalinnar imbë met “upon the (many) foaming waves between us (two)” (LotR/377, RGEO/59).
It seems that, in combination with the definite article, the “many” aspect of the suffix -li is emphasized. Perhaps i falmannar imbe met = “the waves between us” but i falmalinnar imbe met = “the many waves between us”, as a definite group less than totality of all waves, but nevertheless quite numerous. I’m not sure who first suggested this possible idiomatic usage of -li, but I got it from Thorsten Renk’s Quetin i Lambë Eldaiva Quenya course (QLE/25). If this interpretation is accurate, then in this context:
- i falmar “the waves” (general plural) < i falmali “the many waves” (partitive plural)
It is not clear how Quenya would transition within a narrative between an indefinite partitive plural -li and a definite general plural. In English, you might say “Some Elves came to the house. The Elves sat and ate bread”. Perhaps in Quenya you would do the same: Eldali túler i coanna. I Eldar harner ar manter massa. It is conceivable, though, that Quenya would continue to use the partitive plural: Eldali túler i coanna. Eldali harner ar manter massa. It is also possible that Quenya would require a stronger demonstrative to transition between the indefinite partitive and the definite general: Eldali túle i coanna. Tai Eldar harner. I Eldar manter massa. “Some Elves came to the house. Those Elves sat. The Elves ate bread”. Absent further evidence, I would assume the English pattern: partitive plural at first, switching to the definite general plural once the group was specified.
It is also not clear how Quenya would transition between a definite group and a subset of that group. In English: “The Elves came to the house. Some of those Elves sat and ate bread.” In Quenya, perhaps: I Eldar túler coanna. Eldali harner ar manter massa. Here exact English pattern cannot apply, because Quenya does not have an independent word for “some”. It seems likely that in Quenya, a partitive plural could refer to a portion of a previously mentioned (definite) group, though it is not clear whether the subgroup would become indefinite (Eldali) or would remain definite (i Eldali). Given that “with Eldali the definite article is seldom used (PE21/73)”, I suspect the article would be omitted.
There is one further idiomatic expression of note associated with the partitive plural. When you want to indicate a small group, you can use the partitive plural in combination with the adjective nótima “countable, limited in number”, which must be inflected in the plural to match the noun. Thus cennen nótime eldali “I saw a few Elves” (PE22/155), where “few” = “some but countable”. This further supports the idea that an otherwise unmodified partitive plural implies a large group (“many”).
Based on the Plotz letter, the partitive plural suffix -li is added to a vocalic noun first followed by the (plural) noun case inflection, using the forms for inflecting a regular i-plural noun:
- ciryalin [partitive plural dative] = cirya-li-n.
- ciryalion [partitive plural genitive] = cirya-li-on.
- ciryalinna(r) [partitive plural allative] = cirya-li-nna(r).
- ciryalillo(n) [partitive plural ablative] = cirya-li-llo(n).
- ciryalisse(n) [partitive plural locative] = cirya-t-sse(n).
- ciryalínen [partitive plural instrumental] = cirya-li-nen.
- ciryalíva [partitive plural possessive-adjectival] = cirya-li-va.
Because the partitive plural suffix already indicates the plurality, the extra plural marker (r or n) is optional in the allative, ablative and locative cases, as indicated by the parenthesis: -linna(r), -lillo(n), -lisse(n).
For the instrumental and possessive-adjectival cases, the added case suffix triggers prosodic lengthening because otherwise the result would be a suffixed form ending in three light syllables. Likewise, the partitive plural suffix by itself (and in its dative form) would trigger prosodic lengthening when added to a vocalic noun ending in two light syllables: tyaliéli “some plays”, teluméli “some domes”, tyaliélin “for some plays”, telumélin “for some domes”. The genitive partitive plural suffix -lion would trigger prosodic lengthening when added to any trisyllabic or longer vocalic noun: malinornélion “of [many] yellow trees” (PE17/80). It thus wins the prize for “suffix most likely to trigger prosodic lengthening”.
We have very few examples of the suffix -li being added to consonantal nouns. The few examples we have seem to indicate the suffix is added directly to the stem, with various assimilations:
- Kasalli partitive plural of Kasar “Dwarf” (WJ/388) with rl > ll.
- elelli partitive plural of elen “star” (PE17/127) with nl > ll.
If this is the general pattern, then nouns ending in l, s, t would probably be inflected as follows:
- *aranelli partitive plural of aranel “princess”.
- *sololli partitive plural of solos “helmet” with sl > ll.
- *saralti partitive plural of sarat “letter” with tl > lt.
As support for -sli > -lli, look at [ᴹQ.] Telelli vs. Teler, derived from the root ᴹ√TELES (Ety/TELES). Assuming the above holds true, the partitive plural would erase (almost) all distinctions between noun stems ending in a single consonant. However, some consonantal nouns have stems that end in more than one consonant, especially reduced compounds like Moriquen (Moriquend-) “Dark Elf”. It is very unlikely the partitive plural could assimilate in those cases.
For nouns whose stems end in consonant clusters, probably a vowel would be inserted between the stem and the partitive plural suffix: Moriquendeli. This vowel would likely be either (a) a restoration of any primitive vowel originally lost or (b) some common joining vowel like -e-. I suspect that analogical levelling would most likely to produce (b) since e would be one of the more common “restored vowels”; most Neo-Quenya authors suggest using -e- as a joining vowel with consonantal nouns for the suffix -li. It is even possible that in Post-Exhilic Quenya (Tarquesta) this was generalized to some noun stems ending in a single consonant, so saratéli (with prosodic lengthening) rather than saralti.
Regardless, any further noun case inflections would be added to the partitive plural form, as with the vocalic noun inflections given above. Furthermore, the partitive plural suffix is probably added after any possessive pronoun suffixes, which would therefore be declined like vocalic nouns: ciryanyali “some of my ships”.
If a partitive plural noun is the subject of a sentence or is modified by an adjective, the verb and/or adjective must be inflected to match the number of the noun. Since there is no special “partitive plural” inflections for adjectives and verbs, the regular plural inflection is used: saile eldali quenter ta “some wise elves said that”.
Conceptual Development: Unlike the dual, there is no sign of the partitive plural in Tolkien’s earliest conception of Elvish. There is no mention of it in the Early Qenya Grammar of the 1920s. There is, however, a mention of a partitive noun case: “-ĭnen partitive ‘of, out of’ (PE14/46)”. This partitive inflection seems to be mainly a partitive genitive, used to describe when one thing is composed of another: i tolma kuluinen “the helm of gold” (PE14/46). It was used in comparison: i mitta ’n·felda aksínen “the worst piece of steel” (PE14/48), aksínen = ᴱQ. akse “steel” + -inen. It was also used as a “partitive of collective”:
but all numerals even er, except kea, -kea, can be used with partitive of collective in sense of one (two, three, etc.) pieces of, as yúyo kuluinen “two pieces of gold” (PE14/50).
In these early appearances, the partitive does not seem to have a specific plural function. Tolkien continued to refine this partitive inflection in his tables of Qenya Declensions composed in the 1920s (PE16/111-115) where the suffix became -ika (PE16/112). In Tolkien’s larger document describing the Declension of Nouns from the early 1930s, this -ika inflection (kiryaika) became “(11) in plural only a partitive genitive, which is now obsolescent and replaced largely by plural of (4) [the regular genitive]” (PE21/3-4); this “plural only” partitive inflection may have been the been the earlier appearance appearance of something like the partitive plural. The partitive inflection eventually disappeared in later declension tables (PE21/53).
Even though the partitive plural does not appear in Tolkien’s earliest writings, the plural suffix ᴱQ. -li does appear as the general plural suffix of consonantal nouns rather than as a special plural (QL/53; PE14/43-44). In Early Qenya, the consonantal and vocalic plural suffixes are -i and -li rather than -i and -r as they were later on (starting in the mid-1930s). Thus ᴱQ. Noldoli but ᴹQ., Q. Noldor. This use of -li as an ordinary plural suffix in Early Qenya makes it hard to pin-point exactly when Tolkien introduced the partitive plural.
The general plural Noldoli continued to appear in Silmarillion drafts up to the 1930s (LR/119) but was abandoned sometime in the mid-thirties in favor of Noldor, which had appeared sporadically prior to this point (for example on LT1/162). The r-plural was the norm in The Etymologies, but the suffix ᴹQ. -li did appear as a “pl. suffix” (Ety/LĪ). In that document it was used only once in an actual word, Telelli the plural form of Teler (Ety/TELES), but this use does not seems to be partitive in nature. The first explicit mention of a partitive plural is in the Primitive Quendian Final Consonants document which was dated 1936 (PE21/xiii). The relevant text:
In keeping with this dual distinction between the partitive “2” and the whole or dual-group, the Eldarin languages show a tendency also in the plural to make a distinction between the “partitive plural” — men, some men; and the group plural — the men, all the men of a group or kind, the whole body. Thus in Q. Eldar “elves”, Eldali “the Elves, the whole Elda-race”. The latter form was (as the duals) originally syntactically singular, and was hardly distinguished in use or sense from the cpd. noun Eldalie “Elf-folk” (PE21/57).
In this text, the -li suffix has the opposite of its later meaning: it functions as a collective plural not unlike Noldorin/Sindarin -ath. This is in keeping with the apparent use of Telelli in The Etymologies (above). The first clear use of the -li suffix as a proper partitive plural is the early 1950s quote from Common Eldarin Noun Structure, given above (PE21/73).
Neo-Quenya: As mentioned above, most Neo-Quenya authors recommend using the joining vowel -e- with the suffix -li for consonantal nouns, using the assimilated forms only in a few cases. Thorsten Renk suggested using an assimilated form only for nouns ending in -l or -r, where Tamas Ferencz suggested doing so for -n as well. I think using either an assimilated form or a joining vowel would be fine in most case. I personally would do so only for nouns ending in l and maybe r, based mainly on how the word sounds. I think atanéli “some men” sounds better than atalli, but either could be correct; likewise ataréli “some fathers” may be preferable to atalli. As indicated by these example, the added -e- would often be subject to prosodic lengthening.
A strong argument can be made that the joining vowel should be -i- rather than -e-, based on general-plural examples like elenillor (MC/222). This is an interesting idea (suggested to me by Shihali), but until we get clearer evidence, I personally will stick with the established Neo-Quenya convention of -e- for now.