Sindarin has a number vowel mutations that serve various grammatical functions. The best known is i-affection, which plays a major role in Sindarin plural nouns and adjectives, but is a factor in the conjugation of the Sindarin present tense as well. I divide i-affection into 3 different “flavors” of changes: internal i-affection, final i-affection and final i-intrusion. There is also a vowel lengthening mutation that plays significant role in the Sindarin past tense. This entry is principally concerned with the historical origins of vowel mutations, to centralize that information in one place. Their grammatical uses are discussed elsewhere.
Internal i-affection: Internal i-affection happens to vowels in non-final syllables of words, where the syllable originally precedes one containing an i (which may or may not have vanished). This mutation causes the vowels a, o, u in non-final syllables to become e, e, y. This happens mainly (a) in the non-final syllables of noun and adjective plurals and (b) in the present tense of basic verbs that have a pronominal suffix. Some examples:
- Non-final syllable a → e: adan “man” → edain “men” (LotR/1128; RGEO/66).
- Non-final syllable o → e: onod “Ent” → enyd “Ents” (LotR/1130; Let/224).
- Non-final syllable u → y: [N.] tulus “poplar” → tylys “poplars” (Ety/TYUL).
- Present tense a → e: car- “to make” → cerin “I make” (PE17/132).
- Present tense o → e: nor- “to run” → nerin “I run”.
- Present tense u → y: *run- “to rub” → rynin “I rub”.
Words with short u are rare in Sindarin and none have attested plural forms with u → y, so the example tulus is a Noldorin word from the 1930s which it still fit Sindarin’s phonology. Sindarin basic verbs with u are incredibly rare (none are attested), and the example verb *run- “to rub” is invented for demonstration purposes.
Internal i-affection originates from two phonetic developments in Sindarin’s history:
- First a short [a], [o], [u] became [e], [œ], [y] preceding [i]; this is also known as i-fronting.
- Later the [œ] became [e]
The first sound change is the result of an a, o, u moving forward in the mouth to e, œ, y (fronting) to assimilate to the following i-sound and the second is the merging of the rounded vowel œ with e (IPA [ɛ]). In the case of the present tense, this i-sound still appears in the word, but in the case of plurals the i was originally at the end of the word and was thus lost as was the case with (almost) all final vowels in Sindarin.
Internal i-affection can also be seen in some older compounds such Beleriand “Country of Balar” = Balar + -iand, but this is a historical development and not an active feature of modern Sindarin compounds, as with: randir “wanderer” = ran- + dîr.
The sound changes of internal i-affection date all the way back in Early Noldorin plurals and present tense forms of the 1920s.
- Non-final syllable a → e: ᴱN. aglar “glory” → eglair (PE13/136, 158).
- Non-final syllable o → e: ᴱN. Golodh “Gnome” → gelydh, †goelydh (PE13/162).
- Non-final syllable u → y: ᴱN. tunneb “full of holes” → tynnib (PE13/165).
- Present tense a → e: ᴱN. dag- “to slay” → degib “you slay” (PE13/130).
- Present tense o → e: ᴱN. gonod- “to count” → genedig “they (masc.) count” (PE13/162).
Similar mutations are found in the non-final syllables of plural forms in the Noldorin of the 1930s. They are also seen in 1930s Noldorin infinitives, such as: teli “to come”, the infinitive of N. tol- (Ety/TOL).
Final i-affection: Final i-affection happens to vowels in the final syllable of plural nouns and adjectives (including monosyllables), where those syllables originally preceded an ancient plural -ī suffix that was ultimately lost in Sindarin. This mutation causes the vowels a, e, o, u in final syllables to become e, i, y, y. Some examples:
- Final syllable a → e: narn “tale” → nern “tales” (MR/373).
- Final syllable e → i: edhel “elf” → edhil “elves” (WJ/364).
- Final syllable o → y: orch “orch” → yrch “orcs” (LotR/345, 1131).
- Final syllable u → y: [N.] tulus “poplar” → tylys “poplars” (Ety/TYUL).
As noted above, words with short u are rare in Sindarin and there are no attested Sindarin plurals with u → y, but this Noldorin example still fits Sindarin’s phonology.
Final i-affection differs from internal i-affection due to an earlier Sindarin sound change whereby e, o were raised to i, u, but only in the syllable immediately before an i at the end of the word (this final i eventually vanished). This historic sound change is called i-raising, and final i-affection results from a combination of i-raising followed by i-fronting:
- a, e, o, u > (raising) a, i, u, u > (fronting) e, i, y, y.
- *narnī > (fronting) nernĭ > nern
- *edelī > (raising) edhilĭ > edhil
- *orkī > (raising) urchĭ > (fronting) yrchĭ > yrch
- *tulussī > (fronting) tylyssĭ > tylys(s) > tylys
In the case of short a, simple final i-affection of a → e only applies to words that end in a consonant cluster, or single consonants that were originally clusters such as m or ng [ŋ], for example: cem plural of cam “hand” (VT50/22). When a word ends in a single consonant, i-intrusion comes into play (adan → pl. edain), as described below. Clusters resist i-intrusion in Sindarin, which is why ordinary final i-affection applies to short a before clusters.
Final i-affection was somewhat different in the Early Noldorin of the 1920s, because clusters did not resist i-intrusion and i-diphthongs were the norm for final syllables of plurals with short a. Note also the aberrant Sindarin examples morchaint plural of morchant “dark shape” (VT42/9) and eilph plural of alph “swan” (UT/265), which may be remnants of those earlier ideas.
By the 1930s Noldorin mostly followed the Sindarin pattern of final i-affection. The main exception was o which was not subject to i-raising. Short o mostly had the same mutation in initial and final syllables in Noldorin, though the details of its final vs. non-final historical development were not the same. For orod “mountain”, compare the Noldorin and Sindarin plurals: ered (Ety/ÓROT) vs. eryd (WJ/192). Since the 1930s-style Noldorin plural ered permeated The Lord of the Rings, Tolkien had to concoct a new explanation for it as a very late (probably Gondorian-only) sound change in limited circumstances (PE17/33; Let/224).
Final i-intrusion: Final i-intrusion happens to vowels in the final syllable of plurals that end in a single consonant. Due to historical accidents it mostly applies monosyllables with long vowels. In these circumstances, not only did the final i mutate the preceding vowel, it also intruded into the preceding syllable to produce an i-diphthong. These mutations arose from the sound change whereby final [i] intruded into preceding syllable. The long vowels î, ŷ were immune to i-affection, whereas long ê (originally a short e lengthened in monosyllables) followed the simpler final i-affection to î, as in: hên “child” → hîn “children” (WJ/403).
Some examples of final i-intrusion:
- Final intrusion a → ai: adan “man” → edain “men” (LotR/1128; RGEO/66).
- Final intrusion â → ai: bâr “home” → bair “homes” (PE17/97; SD/129).
- Final intrusion ô → ui: thôn “pine” → thuin “pines” (PE17/81).
- Final intrusion û → ui: lhûg “snake” → *lhuig “snakes”.
- Final intrusion au → oe: naug “dwarf” → noeg “dwarves” (UT/100).
* There are no attested Sindarin monosyllabic plurals with û → ui, with the possible except of fui “paths” which might be a plural of unattested fû (and in any case it is not really an example of intrusion). However, there is solid evidence for this phonetic evolution in other words (e.g. S. ruin) and it is widely accepted as the likeliest plural of û monosyllables in Neo-Sindarin.
The example noeg does not appear at first to be an i-diphthong, but that is because it is the result of a later sound change whereby [oi] became [oe]. This means all of the intrusions resulted in an i-diphthong, and the case of au → oe is obscured by some additional sound changes afterwards. The probable phonetic development of the plural of naug < *naukā is a bit convoluted:
- *naukī > nǭgĭ > nǫig > noeg; in Noldorin of the 1930s it became ui instead (EtyAC/NAUK).
The i-diphthong for a → ai likewise originates from a convoluted set of sound changes. In particular, it passed through the diphthong ei first, and only later did the ei become ai as it always did in final syllables (including monosyllables):
- *(m)barī > (raising) berĭ > beir > bair; this last transition didn’t happen in Noldorin of the 1930s (PE22/36).
Finally, there are several attested Sindarin and 1940s Noldorin plurals and where monosyllabic ô → ui, but the exact phonetic evolution is unclear. In fact, many Neo-Sindarin writers assume the plural of these monosyllables have ô → ŷ, as was the case in Noldorin of the 1930s (e.g. pŷd plural of N. pôd, Ety/POTŌ), because it fits the known phonetic developments better. It is definitely the case that o → y in the final syllable of polysyllables, as in eryd plural of orod and emyn plural of amon. See the phonetic entry on how final [i] intruded into preceding syllable for my arguments for the validity of ô → ui plurals.
Final i-intrusion was a feature of Sindarin’s conceptual precursors starting with Early Noldorin plurals in the 1920s, but there were so many shifts in diphthongal developments that the actual plural forms varied considerably.
Vowel Lengthening Mutation: There is a mutation appearing in the Sindarin past tense of many basic verbs, where the base vowel was lengthened in ancient times. This mutation originates from differences in the phonetic development of short and long vowels. In particular lengthened a, e, o became ó, í, ú, vs. lengthened i, u which simply become í, ú. In past tenses with further suffixes, the vowel remains long, but where the past tense has no suffix, the vowel (usually) shortens again but retains its quality. Examples include:
- Lengthened a → ó: car- “to make” → agor “made” (WJ/415).
- Lengthened e → í: dew- “to fail” → edíw “failed” (PE17/151).
- Lengthened o → ú: nor- “to run” → onur “ran” (PE17/168).
These mutations are the result of two distinct historical sound changes:
The long vowel then shortened at the end of the polysyllable when the verb was unsuffixed, but was retained when the verb has a suffix. This mutation is somewhat speculative, given the small number of attested Sindarin past tenses and Tolkien’s inconsistency in marking their vowels as short or long.
Since these vowel-lengthening Sindarin past tenses were not introduced until the early 1950s, this mutation was not a part of the morphology of Early Noldorin of the 1920s and Noldorin of the 1930s, though it was a feature of Gnomish in the 1910s were vowel lengthening was used in past tense formations.