Sindarin plural forms are a distinctive feature of the language. Sindarin plurals are (for the most part) formed using vowel mutations rather than a suffix like English’s plural suffix “-(e)s”. This makes Sindarin plurals more like German which often changes vowel sounds in the plural, or like “irregular” English plurals like “man” vs. “men” or “mouse” vs. “mice”. This mutation in Sindarin plurals is called i-affection, and is the result of an ancient plural suffix ī which has been lost in modern Sindarin plurals. Tolkien discussed Sindarin plural formation on numerous occasions:

The plural element in nouns is [i] as a suffix [ī] ... In Sindarin the old ī plurals causing “affection” (PE17/62).

Sindarin ... had no declensions, but formed its plurals by the addition of -ī which had vanished b[ut] affected the preceding vowels (as in Welsh and English, “angel”, pl. “engyl”, “man”, pl. “men”, S adan, pl. edain, orch, pl. yrch) (PE17/127).

In Sindarin plurals were mostly made with vowel-changes: Adan, Edain; orch, yrch; etc. (RGEO/66).

Sindarin plurals were made usually by vowel change due to [?influence] of the primitive plural ending -ī. silivren, pl. silivrin. el(en), [pl.] elin. craban, [pl.] crebain. ae no change. ū [>] ui. au, o [>] oe. au ... (draft of the above, PE17/25).

Vowel Mutation in Plurals: The historical origins of i-affection are discussed in the entry on vowel mutations, but the results are summarized here for purposes of discussion. I divide i-affection up into three “flavors”: internal i-affection, final i-affection, and final i-intrusion. The common plural mutations, along with examples, are as follows:

• Internal i-affection occurs only in non-final syllables:
  • Non-final a → e: adan “man” → edain “men” (LotR/1128; RGEO/66).
  • Non-final o → e: onod “Ent” → enyd “Ents” (LotR/1130; Let/224).
  • Non-final u → y: [N.] tulus “poplar” → tylys “poplars” (Ety/TYUL).
  • Other vowels in non-final syllables (e, i, y) do not mutate.
  • Long vowels in non-final syllables (í, ó, ú, ý) do not mutate.
  • Archaically o → œ: golodh “Gnome” → gœlydh “Gnomes” (WJ/364).
• Final i-affection occurs only in final syllables:
  • Final a → e: narn “tale” → nern “tales” (MR/373).
  • Final e → i: edhel “elf” → edhil “elves” (WJ/364).
  • Final ê → î: hên “child” → hîn “children” (WJ/403).
  • Final o → y: orch “orch” → yrch “orcs” (LotR/345, 1131).
  • Final u → y: [N.] tulus “poplar” → tylys “poplars” (Ety/TYUL).
  • Other short vowels in final syllables (i, y) do not mutate.
• Final i-intrusion happens when the last syllable ends in a single consonant:
  • Intruded a → ai: adan “man” → edain “men” (LotR/1128; RGEO/66).
  • Intruded â → ai: bâr “home” → bair “homes” (PE17/97; SD/129).
  • Intruded ô → ui: thôn “pine” → thuin “pines” (PE17/81); monosyllables only.
  • Intruded û → ui: dûr “dark” → duir “dark (plural)” (UT/280); monosyllables only.
  • Intruded au → oe: naug “dwarf” → noeg “dwarves” (UT/100).
  • The long vowels î, ŷ in final syllables do not mutate. The vowels e/ê, o, u undergo final i-affection instead (to i/î, y, y).
• With the exception of au, diphthongs (ae, oe, ei, ai, ui) do not mutate.

Sindarin words with short u are rare; there are no attested examples of Sindarin plurals with u → y, so the best example is Noldorin tulus → tylys, but this word is still consistent with Sindarin phonology. Likewise, there are no attested Sindarin nouns showing i-intrusion for long û in monosyllables so the example given is an adjective; Tolkien also specifically mentioned this plural mutation in the quote from PE17/25 above, with “ū [>] ui”. Finally, the lack of mutation of long ú in non-final syllables is most clearly demonstrated by the (Noldorin) adjective: N. dúven “*western” → dúvin (Ety/NDŪ; EtyAC/NDŪ).

In a few examples, the mutation of short o in non-final syllables was to œ rather than e, which in some documents is ambiguously written as oe (and thus easily confused with diphthongal oe). This œ is archaic, a rounded e in much the same way that y is a rounded i. In modern Sindarin, [œ] became [e]. For a discussion of the lack of mutation for long ó in non-final syllables, see the entry on unusual plurals.

Words with short a in final syllables are a special case, since their plural form depends on whether or not they end in a single consonant (adan → edain) vs. a consonant cluster (narn → nern). Consonant clusters inhibit i-intrusion, so ordinary final i-affection (a → e) is the result. The inhibition of i-intrusion applies to some words that used to end in consonants clusters, but don’t any longer:

• Final m: cam “hand” → cem “hands” (VT50/22).
• Final ng [ŋ]: Anfang “Longbeard” → Enfeng “Longbeards” (PM/321; WJ/10).

The second example superficially resembles a consonant cluster, but its final consonant is actually pronounced like single velar spirant [ŋ] (LotR/1114). Both final m and final ng [ŋ] can only arise from more ancient clusters, because a single ancient m became v and a single [ŋ] vanished. Thus there were no forms that might draw such plurals back to i-intrusion by analogy, and they retained their simpler (non-intrusive) plural forms even after the final clusters reduced to a single consonant.

Conversely, some “clusters” behave like single consonants in that they fail to inhibit i-intrusion. The most notable example is -ss as in lass “leaf” → lais “leaves” (PE17/62, 97). The same seems to be true for other ancient clusters as well, such as with the plural Periain of Perian “Halfling”, whose class-plural Periannath indicates it once ended with long nn. I think the three double-consonant clusters ss, ll, nn were probably particularly “weak” and unable to inhibit i-intrusion, since all of them were weak enough to shorten at the end of polysyllables. I think the same applies to monosyllables ending in nn, but not those ending in nd, hence *ain “gifts” plural of ann (< ✶annā) but *bend “prisons” plural of band (< ✶mbandō).

Also remember that the digraphs ph, th, ch, dh represent single consonants at the end of words and would likewise have i-intrusion plurals, as in rath “street” → raith (RC/526) or galadh “tree” → gelaidh (PE17/60). Finally, there are a few Sindarin examples of i-intrusion plurals for words ending in clusters, notably morchant “dark shape” → morchaint (VT42/9) and alph “swan” → eilph (UT/265); compare the second of these to [N.] alf “swan” → elf (Ety/ÁLAK; KHOP). It is unclear whether these represent conceptual vacillations or are additional exceptions to clusters resisting i-intrusion.

Nasal Mutation in Definite Plurals: The definite article i has a plural form in. This form is retained before plurals beginning with vowels, as in in Edhil “the Elves” (RGEO/62), but before consonants the n generally vanishes and causes nasal mutation. Examples include:

• hên “child” → i chîn “the children” (S/198).
• têw “sign” → i thiw “the signs” (LotR/305).
• belegorn “great tree” → i melegyrn “the great trees” (WJ/185).
• bâr [mb-] “home” → i mbair “the homes” (SD/129).

See the entry on nasal mutation for a more lengthy discussion of the rules for this consonant mutation. One of the biggest challenges with Sindarin is recognizing that all the various mutations represent different forms of the same word: i dâl “the foot”, tail “feet”, and i thail “the feet” are all different forms of tâl “foot”.

Plurals of Compounds: Normally the plural mutation applies to the entire word, but in compounds it frequently applies only to the final element of the word. As Tolkien described it:

As a pronoun, usually enclitic, the form pen, mutated ben, survived. A few compounds survived, such as rochben “rider” (m. or f.), orodben “a mountaineer” or “one living in the mountains”, arphen “a noble”. Their plurals were made by i-affection, originally carried through the word: as roechbin, oerydbin, erphin, but the normal form of the first element was often restored when the nature of the composition remained evident: as rochbin, but always erphin (WJ/376).

As indicated by this quote, the compound must still be recognized for the plural to be limited to the final element. The compound rochben is formed with the still active soft mutation making it easy to recognize. However, arphen is formed with older and (possibly) no longer active sound change rp > rph, making it seem more like a single word. Furthermore, Tolkien was rather inconsistent on whether he applied the rule that only the final element of a compound is pluralized:

• celerdain plural of calardan “lampwright” = calar + tân (PE17/96; RC/523).
• Celbin plural of Calben “Light Elf” = cal + pen (WJ/362, 376).
• Enfeng plural of Anfang “Long Beard” = and + fang (PM/321; WJ/10).

Compare the above to:

• Eruchîn (not **Erychîn) plural of Eruchên “Child of God” (Let/345).
• galadhremmin (not **geledhremmin) plural of galadhremmen “tree-tangled” = galadh + remmen (LotR/238).
• Morbin (not **Merbin) plural of Morben “Dark Elf” (WJ/362); this word had an archaic plural †Mœrbin.
• morchaint (not **merchaint) plural of morchant “dark shapes” (VT42/9).
• tad-dail (not **ted-dail) plural of tad-dal “two foot” (VT42/9).

Whether these variations represent some specialized rule or are simply vacillations on Tolkien’s part is unclear, but it might be notable that all of the exceptions involve the mutation of a → e, which may have been more prone to mutation than other vowels in non-final syllables. Perhaps the vowel a still mutates in compounds unless (a) it is a loose compound like tad-dal or (b) a non-mutating vowel intervenes as in galadhremmin.

Conceptual Development: The Gnomish Grammar and Gnomish Lexicon of the 1910s did not use i-affection for plurals, and instead had various plural suffixes like -in and -th (GG/12-13). Plurals with i-affection began to appear in the Gnomish Lexicon Slips, addendums to the Gnomish Lexicon, such as: ailf plural of G. alf “swan” (PE13/109), emyn plural of G. amon “hill” (PE13/110). In the slips these i-affection plurals supplemented but did not entirely replace plural suffixes, though the suffix -in itself sometimes caused i-affection: elmin plural of G. alm “back” (PE13/109).

This remained the case in the Early Noldorin Grammar, with pure i-affection plurals being mixed with suffixal plurals, where some suffixes also caused i-affection (PE13/121-123). The use of i-affection alone became the norm in Noldorin plurals appearing in The Etymologies of the 1930s, though suffixal plurals never completely vanished, for example: conin plural of S. caun “prince” (LotR/953; PE17/102); see the entry on unusual plurals for further discussion.

A complete discussion of the earlier forms of i-affection is beyond the scope of this entry, but it is worth highlighting some of the major differences between Noldorin plurals of the 1930s and Sindarin plurals of the 1950s and 60s.

• The plural mutation of monosyllables with au was ui instead of oe as in: N. plural nuig vs. S. plural noeg for naug “dwarf” (EtyAC/NAUK; UT/100).
• The plural mutation of monosyllables with ô was ŷ instead of ui as in: N. plural pŷd of pôd “animal’s foot” vs. S. plural thuin of thôn “pine” (Ety/POTŌ; PE17/81).
• The Noldorin diphthong ei did not develop into ai, so the i-intrusion plural of a was ei instead of ai: N. plural beir vs. S. plural bair for bâr “home” (PE22/36; PE17/164).
• In Noldorin the vowel o was sometimes not subject to i-raising, and instead went through i-fronting so that o > œ > e in both final and non-final syllables; the net result is that the plural of o could be either ei or e in final syllables rather than y: N. ered, geleidh vs. S. eryd, gelydh plurals of orod “mountain” and Golodh “Gnome” (Ety/ÑGOLOD, ÓROT; S/238, WJ/192).

The variation between e and ei in the final syllables of Noldorin plurals is a bit tricky; it seems that in Noldorin [ei] sometimes became [e] in unstressed final syllables (PE22/39), as with: ᴹ✶talrunya > tellein > tellen (Ety/TAL), so the appearance of e vs ei in final syllables was sporadic. In any case, the Noldorin plural mutations o → ei/e seems to have been abandoned by the 1950s.

For reasons unclear, Tolkien chose to retain the Noldorin-style plural ered for orod in The Lord of the Rings itself; this is the only remnant of Noldorin plural mutation o → e in final syllables appearing in Tolkien’s later writing. This forced Tolkien to contrive a new explanation for this plural, and he decided y > e was a late (possibly Gondorian only) sound change in limited circumstances (Let/224; PE17/33). Indeed, Tolkien fully intended to use eryd when he got around to publishing the Silmarillion (PE17/33). However, Christopher Tolkien chose to leave this plural as ered when he posthumously edited his father’s work, most likely because Christopher did not want to confront readers with a new Sindarin plural form that they would (incorrectly) assume was mistaken.

Neo-Sindarin: The plural system proposed above has mostly been long-established in Neo-Sindarin, appearing in courses and books from the early 2000s. However, many Neo-Sindarin writers suggest that the plural of monosyllabic words with ô should be ŷ based on attested Noldorin plurals and what is known about Sindarin phonology. I instead propose ô → ui in monosyllables.

In researching Sindarin phonology I (re)discovered that there were a number of Sindarin words with o/ô → ui in monosyllabic plurals in the 1940s and 50s:

• N. dôl “hill” → duil (deleted word in Lord of the Rings drafts from the 1940s, TI/268).
• thôn “pine” → thuin (notes on The Lord of the Rings from the late 1950s or early 1960s, PE17/81).
• thôl “helmet” → thely or thuil (late etymological notes associated with The Lord of the Rings of uncertain date, PE17/188).

The last example is a bit peculiar, in that it gives two plurals. The first plural thely is probably the “historical” plural derived from eroded ancient forms, probably *tholosī (compare Q. solos < *tholos). The second is probably a “reformed” plural converted to fit more modern plural patterns. This only makes sense if ô → ui is the normal plural mutation in monosyllables.

Conversely, the last attested Noldorin plurals showing ô → ŷ date back to the 1930s or (at the latest) early 1940s:

• N. pôd “animal’s foot” → pŷd (The Etymologies of the 1930s, Ety/POTŌ).
• N. dór “land” → i·nnýr (The Feanorian Alphabet, late 1930s, PE22/33).

Therefore, there is very strong evidence that, starting with the 1940s, monosyllables with ô have plurals with ui. The main problem with this plural pattern is that it is hard to explain phonologically. We know that long ô in monosyllables only arises from ancient short ŏ (ancient long ō became ū), which should have undergone i-raising to y. Such a y resisted i-mutation in other circumstances, such as polysyllabic amon → emyn, or mŷl plural of mŷl “gull” (probably < *miulē). My best explanation is that the monosyllabic y from raised o was somehow of slightly different character and broken into the diphthong ui under i-intrusion.

See the phonetic entry on how the final [i] intruded into preceding syllable for further discussion. In any case, I recommend the plural mutation of ô → ui in monosyllables instead of the more commonly recommended ô → ŷ: *duir plural of dôr “land”, *bruig plural of brôg “bear”. For polysyllables, however, o → y is the correct plural mutation for Sindarin, and likewise for monosyllables with short o: *rych plural of roch “horse”.

@@@ extract phonetic info