Nominal and abstract endings
Quenya makes extensive use of affixes, prefixes and suffixes, to form words. Relatively few words consist of a naked root. (However, some of the formations are very old; not all the endings listed below were actually productive in late Valinórean or Exilic Quenya. Some methods of derivation that belong to Primitive Quendian rather than to Quenya are passed over, though Quenya vocabulary may include descendants of words so derived.) If the affixes listed below are used to derive new words, care must be taken to avoid combinations (especially of consonants) that are impossible in Quenya.
This is a list, not intended to be exhaustive, of endings occurring on Quenya nouns. (The stems referred to, KOR, GALA, PAR etc., are found in the Etymologies in LR:347-400 unless any other reference is given.) Besides the endings listed here, common nouns can be derived from the naked stem by adding any of the vowels -a, -ë, -o or (very rarely) -u; this is sometimes combined with lengthening of the stem vowel, sometimes not: porë "flour" from POR, mírë "jewel" from MIR, róma "loud sound" from ROM, malo "pollen" from SMAL. (The few nouns in -i seem to be feminine; see Feminine endings below.) The final consonant of the stem may be doubled or undergo nasal infixion before the final vowel is added (e.g. quetta "word" from KWET "say", quinga "bow" from KWIG; primitive forms *kwettâ, *kwingâ).
-at: in hyapat "shore", lanat "weft", sarat "Rúmilian letter" (SKYAP, LAN, WJ:396). Basic meaning unknown; may represent simply an extended form of the stem. It some cases it seems to denote something produced by the corresponding verbal action, like lanat "weft" from LAN "weave". Very likely, the words in -at are examples of the so-called kalat-stems, formed by suffixing of the stem vowel (so-called ómataina, WJ:417) and adding -t (see WJ:392). If so, the ending is not actually -at, but only -t (cf. rukut derived from RUKU [WJ:389]; this word does not seem to have any descendant in Quenya).
-ba: perhaps an allomorph of -wa (see below) occurring after m: romba "trumpet" from ROM "loud noise, horn-blast". Alternatively, the B of -ba is simply part of a "medial fortification" M > MB.
-ë combined with lengthening of the stem-vowel is used to derive what is properly verbal nouns. Sometimes the sense of the derived words drifts from pure abstract to the more concrete, denoting an object or phenomenon that is produced by the corresponding verb: nut- "tie", nútë "knot" (etymologically *"tying?"), lir- "sing", *lírë "song" (etymologically *"singing"; the word is asterisked because it is only attested in the instrumental case: lírinen); cf. also sírë "river" (etymologically "flowing") from sir- "flow". This method of derivation seems to be limited to basic verbal stems of the pattern (consonant-)vowel-consonant. But the ending -ë may also be used to derive abstract nouns from adjectives in -a: aira "holy", airë "sanctity" (PM:363).
-ië: abstract nouns. In WJ:394 tengwestië "Language [as abstract or phenomenon]" is called an "abstract formation" based on tengwesta "system or code of signs", *"[any individual] language". (Tengwesta is also glossed "grammar" [TEK], but only referring to the grammar or system of a specific language, not "grammar" as an abstract). Examples of -ië from the Etymologies include verië "boldness" from the adjective verya "bold" (or the verb verya- "to dare", BAR) and voronwië "endurance, lasting quality" from the adjective voronwa "enduring, long-lasting" (BORÓN). Note that this ending displaces final -a and the whole ending -ya. Sometimes it may denote a collection of something: sarna "of stone" (SAR), sarnië "shingle, pebble-bank" (UT:463). Cf. also olassië "collection of leaves, foliage" (< lassë "leaf"); the prefix o- means "together" (Letters:282). The word enquië "[six-day] week" from enquë "six" refers to a unit or collection of six (days, in this case).
-lë is typically used to derive verbal nouns: horta- "speed, urge", hortalë "speeding, urging" (KHOR), intya- "guess, suppose", intyalë "imagination" (lit. *"guessing, supposing", INK), vesta- "wed", vestalë "wedding" (BES). These verbal nouns may be formed directly from the stem when it ends in a vowel: tailë "lengthening" (TAY [or *TAI] "extend, make long(er)"), cuilë "life, being alive" (KUY "come to life"). In the case of basic stems ending in a consonant, the ending -lë may be added to a nasal-infixed form of them: mancalë "commerce" from manca- "trade", that is in turn derived from MBAKH "exchange", or quentalë "account, history" from KWET- "speak". The ending -lë is also used to derive concrete nouns from an adjective: oia "everlasting", oialë "everlasting [?age]" (Tolkien's handwriting was illegible; OY), aica "sharp", aicalë "a peak" (AYAK), merya "festive", meryalë "holiday" (MBER).
-ma: ending denoting a thing having something to do with the root meaning, either having its properties or being produced by the verbal action in question, or even being a tool used to accomplish it: corma "ring" from KOR "round" (corma is not found in the Etymologies, but cf. cormacolindor "Ringbearers" in LotR3/VI ch. 4/Letters:308), parma "book" from PAR "compose, put together", neuma "snare" from SNEW "entangle". The "thing" may be abstract or concrete: alma = abstract "good fortune, weal" or more concretely "wealth" (stem GALA "thrive", cf. Quenya alya "prosperous, rich"). It is possible that -ba and -wa are allomorphs of this ending, being used after m and n, respectively.
-më: usually denotes abstract, or at least rather intangible, things: melmë "love" (mel- vb. "love"; MEL), qualmë "agony, death" (KWAL "die in pain"), hormë "urgency" (KHOR "urge on"), milmë "greed" (MIL-IK), nilmë "friendship" (NIL "friend"). Less abstract, but still intangible is lúmë "time, hour" and lómë "night" (LU and DO3, DÔ, root meanings not given). Sometimes the basic abstract sense is expanded to include something more concrete: one example is holmë "odour", whereas PQ *ñolmê (my reconstruction) was a verbal noun "smelling" from derived from ÑOL "smell (intr.)", sc. give out a smell (cf. also laimë "shadow" from DAY "shadow" [as verb?]). Likewise the word telmë "covering" may also be used for a concrete object: "hood" (TEL). Cf. also silmë "starlight" (Appendix E) or "light of Silpion" (Telperion) from the stem SIL "shine silver". In a few cases, -më simply functions as a nominal ending: palmë "surface" from PAL "wide (open)". It this and some other cases it may be said to have a local meaning: undumë "abyss" from undu "under", erumë "desert" from ERE "be alone, deprived", celumë "stream, flow" from KEL "go, run (especially of water)". (It is not clear where the u of erumë and celumë comes from; we must probably assume that the stems also occur in the forms *ERU, KELU; a root kelu- "flow out swiftly" is actually mentioned in UT:426.)
-në: evidently a nominal counterpart of the adjectival ending -na; compare the adjective corna "round" (KOR) with the noun cornë "[round] loaf" (LT1:257), compare also sarna "of stone" and sarnë "strong place" (lit. *"something strong as stone"? SAR), cf. also lannë "cloth" from LAN- "weave" (lannë being the nominal counterpart of adjectival *lanna "woven", hence lannë = "something made by weaving").
-on (-ond-) in andon "great gate", aldëon "avenue" - see -on under Masculine endings below.
-rë: it seems that X-rë means "state of having/being X" (almarë "blessedness" from alma "good fortune, weal, wealth"). Not to be confused with the feminine ending -rë.
-së: ending seen on some nouns, like lapsë "babe", litsë"sand" (stems LAP, LIT, meaning not given), also taxë (tacse) "nail" from TAK "fix, make fast", primitive form given as *taksê. In the case of nixë (nicsë) "frost" from the stem nicu- "be chill, cold" it should be noted that the ending -së displaces the final vowel of the stem (WJ:417). Does this ending also occur on essë "name" from ES "indicate, [?to] name"? Or is it just the final consonant reduplicated?
-sta: seen in tengwesta "grammar" (TEK) or "system or code of signs" (WJ:394). The Etymologies and WJ:394 do not agree on the origin of the word tengwesta, but if we accept WJ:394 that is the later source, this word is derived from tengwë "indication, sign, token", indicating that X-sta means "set of Xs, system of X's". However, under KHAW primitive *khau-stâ is defined as "rest-ing", indicating that -stâ (> Quenya -sta) is simply a verbal noun ending. Should we ignore this older source or conclude that -sta has several shades of meaning? Whatever the case, this ending does not seem to be productive in Quenya.
-t: in nat "thing" from NÂ2 "be": literally *"something that is". This is almost certainly the same ending -t that is suffixed to the kalat-stems; see -at above.
-wa: in lanwa "loom" from LAN "weave"; possibly an allomorph of -ma used following n. Not to be confused with the adjectival ending -wa.
-wë: basically abstracts, like voronwë "faithfulness" (UT:305, 317) evidently from the stem BORÓN. After n, as in this case, -wë may be seen as an alternative form of -më. Words in -wë may also denote something produced by the action described by the root: Thus SKAR- "tear, rend" produces harwë "wound" (primitive *skarwê; there was probably a semantic glide from full abstract "tearing, rending" to a concrete rent or wound).
Many or most of these endings are sometimes agental, denoting one who does what the meaning of the stem expresses, like English -er in thinker derived from think, but sometimes they simply denote masculine gender.
-do: evidently an allomorph of -no (see below) used primarily after l and n. Agental in lindo "singer" from LIN2 "sing". Also in noldo, representing primitive -dô (*ñgolodô, WJ:383), so this ending must have been distinct from -nô (in form if not in meaning) already in the primitive language. (Following l, -no could have become -do by common dissimilation, but not following n. Cf. Nando, said to descend from *ndandô, WJ:412.)
-indo: masculine agental suffix, attested in melindo "lover" (m.) and colindo "bearer" (Cormacolindor "Ring-bearers", LotR3/VI ch. 4). The corresponding feminine ending is -indë.
-ion: in morion "dark one", referring to Morgoth (LR:72). Perhaps actually -on (see below) suffixed to the old word *mori "black" (> Quenya morë as an independent word, MOR). Otherwise, -ion is a patronymic ending "-son" (YON).
-mo: Tolkien notes that "the ending -mo often appeared in names or titles, sometimes with an agental significance: Ulmo was interpreted as 'the Pourer' < *UL 'pour out'." (WJ:400. This interpretation of Ulmo was actually another Elvish folk etymology, for the name of this Vala was adopted from Valarin Ulubôz, Ullubôz.) But in ciryamo "mariner" the ending -mo has no agental significance; it is simply added to cirya ship, so the meaning is literally *"ship-person" or something similar. Likewise Súlimo, Manwë's title, seems to mean *"wind-person" (súlë, súli- + mo). Other examples are sermo "friend" from SER "love, be fond of (of liking, friendship)" and ingolmo "loremaster" (WJ:383); cf. n(g)ólë "lore"; cf. also the name of the Vala Irmo, "Desirer" (WJ:403). The feminine counterpart of -mo is -më, but this ending is rare.
-no: yet another masculine ending that is sometimes agental, sometimes not: simply masculine in otorno "(sworn) brother" (< TOR "brother"), agental in tirno "watcher" from TIR "watch, guard" (cf. SKAL2), may be either in samno "carpenter, wright, builder" (meaning of stem STAB not given).
-o: masculine ending, sometimes with agental significance: tyaro "doer, actor, agent" from the verbal stem tyar- "cause", Pityo nickname *"little one" from pitya "little" (PM:353). In PM:340, this ending (there with an undefined diacritical mark) is called a "pronominal suffix" and defined as "a person, somebody". It appears that this ending is properly masculine.
-on: "ending (of male names)" (WJ:400). This is from a context dealing with Sindarin, but this ending is valid in Quenya as well: compare the names Sauron and Ancalimon with the adjectives saura "foul" and ancalima "most bright". According to Letters:380, Sauron was originally Thaurond (th there being spelt with a Greek letter), and the final d may be preserved before an ending (e.g. genitive *Saurondo). Compare Sindarin lhathron "listener" from primitive *la(ns)ro-ndo (LAS2) and Quenya fion "?hawk" (Tolkien's handwriting was illegible) from the stem PHI; the plural is given as either fioni or fiondi, so the primitive form may have been *phiondo (my reconstruction). We also find andon "great gate" (andond-) from ando "gate" (AD). These words indicate that the ending -on is not exclusively used on names. Cf. also aldëon "avenue" < adj. aldëa "tree-shadowed" (LT1:249), though this is very early "Qenya" and may not may full authority. These words are obviously not masculine; they do not even denote animates.
-r or -ro: agentive endings (WJ:371), like English -er: ista- "know" > istar "wizard, *knower" (in Letters:202, Tolkien translates Istari as "those who know"); *envinyata- "renew" > Envinyatar "renewer", The endings -r and -ro may also be added to nouns: X-r(o) then meaning "person having X, having to do with X", as istya "knowledge" > istyar "scholar, learned man". It is possible that the ending -r does not mark sex, while -ro is explicitly masculine (like -rë is explicitly feminine). Cf. ontaro, ontarë "parent", m. and f., respectively (ONO). It seems that the ending -ro forms its plurals in -ri; since this would also be the plural form of -rë, the distinction of sex is lost in the plural: ontari "parents".
-u: masculine ending, sometimes with agental significance: ERE- "be alone" > Eru "The One, God", KHER- "rule, govern" > heru "lord" The word ainu is a special case. This word, denoting one of the angelic spirits originally brought into being by the One Creator, was actually a loan from Valarin ayanûz. But the Elves thought ainu looked like a personal, nominalized form of a (till then) non-existing adjective *aina, and so they actually started to use this adjective, giving it the meaning "holy", holiness being the prime characteristic of the Ainur (WJ:399). This folk etymology indicates that the ending -u (beside -o) was very frequently used to derive personal, nominalized forms from adjectives. The feminine equivalent of -u seems to be -i; see below. (But plural forms like Ainur evidently refer to the entire race, with no distinction of sex. This is probably true of several of the masculine endings given here.)
-wë: according to LR:398 an "abstract suffix" occurring in names like Manwë, Elwë, Ingwë, Finwë. However, Tolkien later decided that it was simply an element meaning "person", "generally but not exclusively masculine" (PM:340 - the sole attested case of a woman having a name in -wë is Elenwë). In Letters:282, Manwë is translated "blessed being". (It has also been explained as a borrowing from Valarin Mânawenûz; see WJ:399.)
These are often direct counterparts of the masculine endings.
-ë: feminine ending, evidently the counterpart of masculine -o: antë "[female] giver" from anta- "give" (the Etymologies, entry ANA1, gives anto "[male] giver", though in LotR anto is said to mean "mouth"). Not to be confused with the abstract or adjectival ending -ë.
-i: feminine ending, evidently the counterpart of masculine -u. Compare heru lord with heri lady, cf. also tári "queen", aini "female ainu".
-ië: feminine ending. Valië "female Vala"; cf. also feminine names like Amárië. As is evident from the example Vala/Valië, this ending can displace a final vowel. Not to be confused with the abstract ending -ië.
-iel: "daughter", as in Uinéniel "Daughter of Uinen" (UT:182).
-issë: feminine agental suffix, attested in melissë "lover" (f.). Cf. also PM:345.
-indë: feminine agental suffix, apparently the feminine equivalent of -indo, attested in Serindë "Broideress" (though translated "Needlewoman" in PM:333).
-llë: feminine agental suffix, only attested in Tintallë "Kindler" < tinta- "kindle, make to sparkle". Note: -llë is also used as a diminutive ending, see below.
-më: the feminine equivalent of the masculine ending -mo: sermë "(female) friend", sermo "(male) friend" - both from SER "love, be fond of (of liking, friendship)". This ending appears to be rare, perhaps because it is easily confused with the nominal ending -më.
-rë: feminine ending, with agental meaning in Vairë (older *Weirê "Weaver", stem WEY "weave"), but not in Ilmarë, the name of a Maia (from Ilma "starlight"). Not to be confused with the abstract ending -rë or the ending -rë denoting a set of something.
These are quite numerous. Note, however, that adjectives never end in -o or -u in mature Quenya.
-a: general adjectival ending: olórë "dream", olórëa "dreamy" (LT1:259).
-arwa: "having", e.g. aldarwa "having trees, tree-grown" from alda "tree" (3AR, in LR:360).
-ba: perhaps the form -wa (see below) takes after m: himba "adhering, sticking" from KHIM- "stick, cleave, adhere". In this case the ending takes on an almost participal meaning.
-ca: adjectival ending used on stems ending in a vowel: PHAU "gape" > fauca "open-mouthed, thirsty, parched, thirsty", POY (meaning not given) > poica "clean, pure". Cf. also GAYA- yielding *gayakâ (called an "adjectival form" in PM:363) > Quenya aica "fell, terrible, dire" after syncope. This ending is very old (Primitive Quendian *-kâ) and may not be productive in later Quenya. (Notice that in the Etymologies, Tolkien derived aica from a stem AYAK, not as later from GAYA- with this ending. The ending as such is nonetheless found in the Etym material as well.)
-da: see -na below.
-ë: rare adjectival ending, among our few examples is lissë "sweet", evidently derived from the stem LIS "honey" (this adjective is not found in the Etymologies, but occurs in Namárië). Some adjectives may seem to display a longer ending -në, as in carnë "red", varnë "swart". However, these words also exemplify the adjectival ending -ë, for the -n- is part of the root (KARÁN, BARÁN). This -ë descends from Primitive Elvish *-i, a common ending in colour-adjectives. - Note that -ë is also an abstract and feminine ending.
-ëa: represents either -ë + a, as in olórë "dream" > olórëa "dreamy", or earlier *-aya and *-oya, sc. the ending -ya (see below) added to a stem ending in some vowel: alda "tree", adjective *aldaya/*aldaia (my reconstruction) > aldëa "tre-shadowed" (LT1:249).
-ima: "X-ima" often means "X-able", "apt to X" or "worthy of X-ing": cf. a couple of such adjectives with the privative prefix ú- "un-": From the verbal stem not- "count" is derived únótima "uncountable", and from quet- "speak" comes úquétima "unspeakable". Note that the ending -ima causes the stem-vowel to become long if it is not followed by a consonant cluster (tyelima "final" [KYEL] and mirima "free" [MIS] do not fit this pattern; here and in some other cases -ima seems to function simply as an adjectival ending). Cf. also Fírimar, translated "those apt to die" in WJ:387 (cf. fir- "fade, die"). Here the adjective is used as a noun and takes the nominal plural ending.
-in: in qualin, firin, both meaning "dead" (KWAL, PHIR), cf. also quorin "drowned" (LT1:264).
-ina is evidently a longer form of -in: malina "yellow" (SMAL), telpina "of silver" (KYELEK). That -ina should be understood as a longer variant of the ending -in mentioned above is confirmed by the fact that an adjective meaning "open, free, cleared (of land)" is given as latin(a) under LAT.
-inqua: ending with the basic meaning "full": alcarinqua "glorious" basically means *"glory-full" (alcar "glory" + -inqua). WJ:415 also mentions an alternative ending *-unqua (only the archaic form -uñkwâ is actually given) that was used to derive adjectives "applied to things heavy, clumsy, ugly or bad". No such adjectives are attested, though.
-itë or -ítë, rare adjectival ending: hanuvoitë "male", inimeitë "female" (INI). Cf. also maitë "handy" from má "hand" (MA3) and hloníti "phonetic" (pl.; sg. *hlonítë; WJ:395), clearly derived from *hlon "sound" (only pl. hloni is attested", WJ:394).
-na: basically the ending of the past (or passive) participle, still used in Quenya, but it is sometimes difficult to tell these participles apart from adjectives, or impractical to introduce this distinction at all. Thus harna "wounded" from SKAR- "tear, rend" (primitive *skarnâ). In cuina "alive" from the stem KUY- "come to life, awake", the adjective describes the condition one is in having completed the action denoted by the verbal stem (cf. the semantic relationship between the English verb go vs. the corresponding past participle gone). The ending -na may dissimilate to -da following L, as in helda "naked" from primitive *skelnâ (stem SKEL).
-rin: an ending frequently found in the names of languages, Sindarin, Vanyarin, Valarin etc. But such words could also be used as general adjectives: "When the historians needed a general adjective 'Quendian, belonging to the Elves as a whole', they made the new adjective Quenderin (on the model of Eldarin, Ñoldorin, etc" (WJ:407). These words could be called ethnic adjectives. Sometimes expanded to -rinwa: Noldorinwa, Sindarinwa.
-sa: in telepsa "of silver" (KYELEP). Probably not productive in Quenya.
-wa: adjectival ending that sometimes seems related to the possessive ending -va, sometimes not: anwa "real, actual, true" (ANA2), Noldorinwa "Noldorin" (see -rin).
-vëa: adjectival ending with the specific meaning "-like": él "star", elvëa "starlike", pl. elvië. (The long é in él becomes short before the cluster lv.)
-viltë, -valta: "-less" (see Parma Eldalamberon #11 p. 23), evidently used to derive adjectives like "worthless" etc., but no such adjective is attested. This ending belongs to very early "Qenya", but no corresponding ending is known from mature Quenya.
-ya: general adjectival ending: númen "west", númenya "western". (Note: -ya is also a frequent verbal ending, apparently unrelated.) See also -ëa above. Adjectives in -ya (as well as other endings) may also be used and inflected as nouns. Attalya "Bipeds" (WJ:389) is clearly an adjective *attalya "two-footed, two-legged" (atta "two" + tal- "foot" + ya) with the nominal plural ending -r.
There are only a few verbal endings.
-ya: general verbal ending: sirya- "flow" from the stem SIR of similar sense. It does not seem that this ending modifies the meaning of the stem in any way. It must not be confused with the frequent adjectival ending -ya, which is apparently unrelated.
-sa: evidently a "frequentative" ending, attested in lapsa- "lick (frequentative)" (LAB). The ordinary verb lav- evidently means to lick on something once. Not to be confused with the adjectival ending -sa (that seems to be equally rare).
-ta: another general verbal ending, sometimes just as general as -ya, sometimes with a causative meaning: tul- "come", tulta- "summon" (= make come) (TUL), airë adjective "holy", airita- "hallow" (= make holy) (according to Vinyar Tengwar # 32 p. 7, this word occurs in the unpublished material). But in some cases, this ending seems to be chosen on the basis of euphony alone, sc. it is often used on stems ending in a vowel or semivowel: roita "pursue" from ROY "chase", caita "lie" from KAY "lie down" (the verb caita is not given in the Etymologies, but is attested in Namárië).
There are also examples of verbs being derived from adjectives, like cúna "bent" > cúna- "bend" (MC:223), or harna "wounded" > harna- "wound" (SKAR).
Some endings of various meaning:
-il: In siril "rivulet" from sir- "flow", the ending seems to denote an impersonal agent (but it may be just a variant form of the diminutive ending -llë, see below). Cf. also sicil "dagger, knife" from SIK (no root meaning given) and tecil "pen" from TEK- "write"; the primitive form is given as *tekla; the i evidently intruded after the loss of final short *-a to break up the final cluster *-kl. In at least one word, -il seems to function as a normal agental ending: *nacil "victor", only attested (in the form -dacil) in compounds like Hyarmendacil "South-victor", the name assumed by a Gondorian king. Surely this element is to be derived from *ndakla, the stem NDAK meaning "slay" (LR:375).
-incë: diminutive ending: atar "father", Atarincë "little father" (PM:353) In UT:195, Zamîn adresses young Ancalimë as hérincë, evidently meaning *"little lady" (heri "lady", see KHER; but the long é in hérincë may suggest that this word is derived from hér-, the form of heru "lord" that used before an ending [PM:210], indicating that the ending -incë does not show sex).
-llë: diminutive ending. Nandë "harp", nandellë "little harp" (ÑGAN. Also in nellë "brook"? [NEN] Cf. nén "water" - hence *nen-lë > nellë, lit. *"small [stream of] water"?) Not to be confused with the feminine ending in Tintallë.
-në: a set of something: carca "tooth", carcanë "row of teeth" (KARAK).
-rë: ending denoting a collection of the things in question: fanya "cloud", fanyarë "the skies...the upper airs and clouds" (MC:223). Could the ending -në, that seems to be of similar sense, simply be a misreading for -rë? Should carcanë read *carcarë?
-ssë: suffix denoting abstract or locality, not to be confused with the locative ending (though that may be related). Examples of such derivation include Vala "angelic power, god" > valassë "divinity" (BAL), laiqua "green" > laiquassë "greenness" (LT1:267), handa "intelligent" > handassë "intelligence" (KHAN), hópa "harbour" > hopassë "harbourage" (KHOP; the long ó of hópa is shortened), findë "hair" > findessë "a head of hair, a persons hair as a whole" (PM:345). Cf. also celussë "freshlet, water falling out swiftly from a rocky spring" from a root kelu- "flow out swiftly" (UT:426).
-ya: "suffix of endearment" mentioned in UT:418, seen in Anardilya *"dear Anardil" (UT:174). Not to be confused with the verbal and adjectival endings -ya.
Quenya has a number of prefixes that may be added to nouns and verbs.
ala- "not-, un-": Alahasta "Unmarred" (MR:254). This prefix seems to have the power to turn a following verbal stem into a past participle even if no explicit participial ending is present. Unlike ú- (see below), this prefix does not seem to have negative connotations.
am- "prefix am- up" (AM2), seen in amortala "heaving", literally *"up-rising", undoubtedly am + ortala (MC:222; orta- = "rise"). Evidently becomes ama- before a consonant; cf. amatixë, dot (tixë) placed above the line of writing, literally *"up-dot" or *"over-dot". Also amba- *"upwards" in Ambalotsë "Uprising-flower" (WJ:318; cf. amba "up, upwards", AM2).
an- "superlative or intensive prefix" (Letters:279), hence ancalima "most bright" from calima "bright".
apa- "after", in Apanónar "the Afterborn" (an Elvish name of Men, WJ:387/Silm ch. 12), also in *apacenya "of foresight" (pl. apacenyë attested in MR:216; this literally refers to aftersight - what will come after the present). Variant ep- in epessë "nickname" (lit. "aftername", sc. a name given after the regular name, UT:266). It would seem that ep- is used instead of apa- when the word to which it is prefixed begins in a vowel.
ata-, at- "back-, again-, re-" (AT[AT]). Of mere repetition, en- may be more usual, but ata- may apparently also imply reversal of some kind (cf. Tolkien's gloss "back").
au- a prefix that is best explained contrasted with hó-; see below.
ava- a prefix occurring on certain adjectives, indicating something forbidden or dangerous: Tolkien contrasts avaquétima "not to be said, that must not be said" and avanyárima "not to be told" with úquétima "unspeakable, impossible to say" and únyárima "impossible to recount" (e.g. because the facts are not known, not because anyone has forbidden to tell the tale). (WJ:370)
can- "four-" (KÁNAT), not attested in any actual compound; an example could be *cantil "square" (cf. neltil "triangle", see nel-).
en- "re": enquat- "refill" (future tense enquantuva in Namárië), entulessë "return" (UT:171). One early "Qenya" variant had an- instead; see LT1:114, 184.
ep- "after", see apa- above.
et- "forth, out". Used on a verb in ettul- probably *"come out, come forth" (SD:290, cf. ET, TUL)
hó- "away, from, from among", prefix used on verbs. According to WJ:368, the "point of view was outside the thing, place, or group in thought". The verb hótuli- *"from-come" thus means come away, "so as to leave a place or group and join another in the thought or place of the speaker", and similarly hóciri- *"from-cut" thus means cut off "so as to have or use a required portion". Contrast the prefix au-, that has a similar meaning *"from, away", but here the point of view stays with the thing, place or group in question. Auciri- also means "cut off", but now to get rid of a portion.
il- negative prefix *"un-"; it "denotes the opposite, the reversal, i.e. more than the mere negation" (LT1:255). Under the stem PHIR we find firin "dead" and ilfirin "immortal"; it can be seen that the negated form does not simply mean "not dead".
lin- "many" (LI), prefixed to adjectives like lintyulussëa "having many poplars" (sc. lin- "many" + tyulussë "poplar" + the adjectival ending -a). Assimilated lil- in lillassëa "many-leaved" (pl. lillassië in the Markirya poem), this is lin- "many" + lassë "leaf" + the adjectival ending -a.
nel- "tri-" (NEL), neltil "triangle" (TIL).
nu- *"under" in nuhuinenna "under the shadow" (SD:246), probably also in nucumna "humbled" (SD:246) - literally *"under-bent".
nun- *"under, below", attested in nuntixë *"underdot", a mark below the line of writing (TIK).
o- (long vowel when stressed: ó-) "a frequent prefix...used in words describing the meeting, junction, or union of two things or persons, or of two groups thought of as units. Thus: o-mentië (meeting or junction of the directions of two people) as in the familiar greeting between two people, or two companies each going on a path that crosses that of the other: Elen síla lúmenna omentielvo! 'A star shines upon the hour of the meeting of out ways.' ... This prefix was normally unstressed in verbs or derivatives of verbs; or generally when the next following syllable was long. When stressed it had the form ó-, as in ónoni 'twins', beside the adj. onóna 'twin-born', also used as a noun 'one of a pair of twins'." (WJ:367). Cf. also otorno *"with-brother", sc. a sworn brother as opposed to or in addition to a natural one (toron, torn- "brother"). Note that this prefix, unlike yo- (see below), primarily refers to two people, things or groups. However, this does not seem to be the case in olassië "collection of leaves, foliage" (< lassë "leaf"), that apparently refers to any number of leaves brought together (Letters:282).
oa-, oar- *"away", "occasionally used as a prefix in compounds of later formation" (WJ:366). Oareldi *"Away-Eldar", elves who departed from Beleriand to Valinor, as opposed to those who stayed there (the Sindar). (WJ:363 cf. 360)
ter- "through". Used on verbs this prefix may indicate continuance in time, so termar- "through-abide" (UT:305, 317) means "stand" in the sense of "endure". Also in the noun tercen "insight", literally *"throughsight" (MR:230).
ú- "no-, not-, un-" (GÛ) often though not always with negative connotations: úquétima "unspeakable", (WJ:370), únótimë "uncountable" (pl., from Namárië). Also used on nouns: Vanimo "beautiful one", úvanimo "monster", sc. exactly the opposite (BAN). Used on nouns ú may also imply absence of the thing in question: Úner "noman" (UT:211).
un- "down". In untúpa "down-roofs" (= covers) (Namárië cf. RGEO:67). This prefix may well be productive, so we may coin words like untul- "downcome" = "descend".
undu- "down" in undulávë "down-licked" (Namárië cf. RGEO:67). This is apparently a longer form of un- used when the latter would produce a consonant cluster not allowed in Quenya like **nl in this case. In LR:47 we also find a prefix unu-, that may be obsoleted by undu- from Namárië.
yo- is basically the preposition "with", together with (SD:56: yo hildinyar *"with my heirs"); it occurs as a prefix in yomenië "meeting, gathering" (of three or more coming from different directions). (WJ:407) Contrast o- above.