S. short [e], [o] became [i], [u] in syllable before final [i]; [-{ĕŏ}{C|CC}i] > [-{iu}{C|CC}i]

S. short [e], [o] became [i], [u] in syllable before final [i]; [-{ĕŏ}{C|CC}i] > [-{iu}{C|CC}i]

In Sindarin and Noldorin the short vowels [e] and [o] were raised to [i] and [u] in the syllables before a final [i]. This process is the first stage of i-affection, called i-raising in this lexicon following the suggestion of Elaran from a private Discord chat on 2018-08-25. A similar set of sound changes occurred in Welsh (ultimate i-affection), but the details were different (WGHC/§69). This sound change was described as “final i-affection” in Jim Allen’s Introduction to Elvish published in 1978 (pp. 114, 116, in an article written by Chris Gilson and Bill Welden). Later analyses were published in the 2000s by David Salo in Gateway to Sindarin (GS/§4.94, 2004) and by Bertrand Bellet in his essay on Vowel Affection in Sindarin and Noldorin (VASN, 2005). In this entry I follow the system outlined by Bertrand Bellet, which I think is the one that is the most consistent with the available evidence.

Tolkien explicitly mentioned this sound change (for e at least) in conjunction with notes from the late 1960s on the phonetic development of S. certh:

e > i only before vanishing ī/j [that is, where ī/j was final and then later vanished] (PE22/150).

This sound change was an important factor in the development in the Sindarin plural, and explains why the mutated plural forms of e and o are different in final syllables (where they originally appeared in the syllable immediately before the lost plural suffix ) and non-final syllables (further removed from the lost plural suffix). Thus we see that the plural of Edhel “Elf” is Edhil (from *edelī) and the plural of orod “mountain” is eryd (from *oruti < *orotī).

Most of the evidence for this sound change comes from plural forms in Sindarin and Noldorin, but there is at least one example that is not a plural:

Here the phonetic development was [o] > [u] > [y], with the first change due to i-raising and the second change due to the second stage of i-affection: i-fronting. In addition to the example above, there are several explicit examples of the phonetic development of Sindarin plurals which also illustrate i-raising:

There are quite a few Sindarin plurals whose patterns are consistent with this sound change:

The last example includes an archaic form indicating the non-final development of o in plurals: o > œ > e. This later sound change is discussed in more detail in the entry on the next phase of i-affection: i-fronting, whereby short [a], [o], [u] became [e], [œ], [y] preceding [i].

There a couple Sindarin plurals (and one late Noldorin plural from the 1940s) that show ui instead of the expected y in plurals of monosyllables with long ô which may represent a variation on i-raising:

I think these are actually variations on the behavior of i-intrusion (the last phase of i-affection), and discuss these examples in that entry on that phonetic development instead.

Conceptual Development: The process of i-affection emerged early in the conceptual history of Sindarin and its precursor languages, but the entire process is very complex. Determining when exactly i-fronting was or was not part of these phonetic developments isn’t straightforward.

Gnomish Developments (1910s): There is no sign of i-affection in the first iteration of Gnomish in the 1910s, as represented in the Gnomish Lexicon. However, as discussed by Roman Rausch in his Historical Phonologies of Ilkorin, Telerin and Noldorin around 1923 (HPITN/§4.2.1) there are signs of i-affection appearing in the Gnomish Lexicon Slips along with new i-mutated plural forms, some of which may demonstrate i-fronting:

Lack of examples makes determining the exact process very difficult. Based on the first example, it seems Sindarin-style i-fronting was present for o > u > y, but the development for e is less clear.

Early Noldorin Developments (1920s): More information on the i-mutational plural system appears in the Early Noldorin Grammar of the 1920s, but the plural patterns are not quite the same as established later in Sindarin:

Examples of o > u before i, ı̯ are probably only apparent; as tonn (tond) pl. tunnin, where Q. shows tunda “hole” (PE13/122). Note where e is from i by a-mutation, before original plural ending the i reappeared, so we get apparent mutations e < i (PE13/122).

Based on this note, it seems that any apparent i-fronting in the Early Noldorin Grammar was actually the result of inhibited a-affection. Examples on the same page seem to indicate that the normal plural form of o and e were ui and ai (probably via oi and ei):

These rules are consistent with most of the relevant plurals appearing elsewhere in the Early Noldorin Grammar, with occasional variants iai and uai instead of ai/ui:

These Early Noldorin Grammar plurals might be explained by a phonological history in which there was i-intrusion but no i-fronting.

However, there seems to be a transition in e/o plural forms in the early Noldorin Word-lists and Noldorin Dictionary, as discussed by Thorsten Renk in his article on Early Noldorin Plural Patterns (ENPP). There are still occasional plurals where e and o in final syllables becomes ai and ui, but they become increasingly rare:

Much more frequently, e and o in final syllables became i and y, as they did in Sindarin:

For the last example, compare the plural dylch (PE13/142) to the corresponding plural from the Early Noldorin Grammar: duailch (PE13/124-5). The presence of Sindarin-style plural pattern indicates that something like i-raising was probably in effect at this conceptual stage. However, there are enough variations in the Early Noldorin plural patterns that figuring out the exact phonetic rule is difficult. Roman Rausch discusses some possible interpretations in his Historical Phonologies of Ilkorin, Telerin and Noldorin around 1923 (HPITN/§4.2.1).

Noldorin Developments (1930s): The mechanisms for Noldorin plurals in The Etymologies of the 1930s are broadly the same as those appearing in the early Noldorin Word-lists and Noldorin Dictionary, but a few new plural patterns emerge. As discussed by Bertrand Bellet in his essay on Vowel Affection in Sindarin and Noldorin (VASN), the plurals of words containing e match the Sindarin pattern, becoming i in final syllables but with e remaining in non-final syllables:

Not all of these examples would be the result of i-fronting: the last example malen “yellow” developed from a primitive ᴹ✶smalinā, so its plural was likely derived from (s)malinī. If so the presence of i in the final syllable of this plural was actually the result of inhibited a-affection rather than i-raising. This was true of many Noldorin (and Sindarin) adjectives, where the suffix -en (from primitive ✶-inā) was common. The remaining examples, however, are most easily explained by the fronting of [e] > [i] in the syllable immediately before the (lost) final plural suffix .

The Noldorin plural patterns for o > y also match Sindarin for monosyllables, and could therefore represent similar i-fronting of [o]:

In the last two examples, the sound change is somewhat obscured by the fact that short vowels generally lengthened in monosyllables. While the monosyllables match the Sindarin development, polysyllable plural patterns for o are more varied. Sometimes the result is y, but other times it is ei:

The reason for these varying developments isn’t entirely clear. However, it might be significant that for two of the examples with y, the primitive plural suffix was separated from the o by two consonants: *stalgondī > thelyn; *annondī > ennyn. Conversely, the examples showing o > ei were probably separated from by only a single consonant: *(ñ)golodī > geleidh; *watorī > gwedeir; *thoronī > therein; *toronī > terein. These last four examples do not show i-raising, but do seem to show i-fronting (o > œ > e) and i-intrusion (eCi > eiC), so perhaps under these conditions i-raising was somehow inhibited and only the other two sound changes applied.

In some cases the diphthong ei was further reduced to e in final syllables:

As these examples make clear, this [ei] > [e] was a later phonetic development, and otherwise these examples followed the plural patterns whereby o becomes ei in final syllables. The last plural ered “mountains” is noteworthy, because it is a Noldorin-style plural that made into the text of The Lord of the Rings, along with Ened as a plural of S. Onod “Ent” (Let/224). The plural of Onod was updated to Enyd in the later editions of LotR, but ered was retained. For Sindarin, Tolkien speculated that this might have been a late sound change of [y] > [e] (PE17/33), possibly the normal Gondorian pronunciation (Let/224). In any case, it is clear that ered was not intended to be the normal Sindarin plural, which was clearly given as S. eryd elsewhere (WJ/192; PE17/33, 89).

The plurals forms N. emuin and emyn of N. amon (Ety/AM²) hint at an alternate development of [o] after raising to [u] (-oni > -uni > -uin > -yn), but that topic is more properly covered in the entry on i-intrusion, a later phase of i-affection.

There are a couple of additional aberrant Noldorin plurals that further break the patterns described above: N. ceif as the plural of N. cef (Ety/KEM), and N. eirch or erch as the plural of N. orch (LR/406). The first example is hard to explain, but may represent an instance where e also developed without i-raising. The second example is probably related to a couple of deleted plurals in The Etymologies: erch a rejected plural of N. orch (EtyAC/ÓROK.017) and ern a rejected plural of N. orn (EtyAC/ORO), both replaced by yrch and yrn respectively. These two examples were pointed out Bertrand Bellet (VASN, note #19), and probably represent a rejected plural pattern.

Summary of Conceptual Developments: There are hints of i-raising as early as the Gnomish Lexicon Slips from the late 1910s, but the Early Noldorin Grammar clearly indicates that any “apparent” i-raising was actually the result of inhibited a-affection. Despite this, there are numerous examples in the early Noldorin Word-lists and Noldorin Dictionary from the 1920s consistent with Sindarin plural patterns, which strongly indicate i-raising existed in some form at this early stage. The Noldorin plurals of the 1930s and 1940s follow the Sindarin pattern for the plurals of words containing e, as well as for monosyllables with o. However, the Noldorin developments of o in polysyllabic plurals show variations (y vs. ei) for reasons that are not entirely clear, but may have something to do with the number of consonants between the o and the lost plural suffix i.

Reference ✧ PE22/150

Order (01600)

Before 01700 short [a], [o], [u] became [e], [œ], [y] preceding [i] Ossai > ossī > ussi > S. yssı̯ WJ/400

Phonetic Rule Elements

[-ĕCi] > [-iCi]
[-ĕCCi] > [-iCCi]
[-ŏCi] > [-uCi]
[-ŏCCi] > [-uCCi]

Phonetic Rule Examples

kerθi > kirθi -ĕCCi > -iCCi KER > S. cirth ✧ PE22/150
teŋmi > tiŋmi -ĕCCi > -iCCi tekmāi > tekmī > tiŋw > S. tīw ✧ PE17/43
teŋmi > tiŋmi -ĕCCi > -iCCi teñmi > tiñm(i) > S. tîw ✧ PE17/44
eðeli > eðili -ĕCi > -iCi edelō > S. Edhil ✧ WJ/364
eleði > eliði -ĕCi > -iCi eledā > S. elidh ✧ Let/281
eleni > elini -ĕCi > -iCi elenā > S. Elin ✧ WJ/363
eleni > elini -ĕCi > -iCi elenī > S. elin ✧ PE22/150
leperi > lepiri -ĕCi > -iCi leper- > S. lebir ✧ VT47/10
leperi > lepiri -ĕCi > -iCi leperī > S. lebir ✧ VT47/24
leperi > lepiri -ĕCi > -iCi leperī > S. lebir ✧ VT48/5
andondi > andundi -ŏCCi > -uCCi ANA > S. Ennyn ✧ PE17/40
dombi > dumbi -ŏCCi > -uCCi DOM > S. dym ✧ PE22/153
dorni > durni -ŏCCi > -uCCi DOR > S. Dyrn ✧ PE17/181
tainakolli > tainakulli -ŏCCi > -uCCi tana > S. Taengyl ✧ MR/385
tannakolli > tannakulli -ŏCCi > -uCCi tana > S. Tengyl ✧ MR/385
orxi > urxi -ŏCCi > -uCCi urkō > S. Yrch ✧ WJ/390
ossi > ussi -ŏCCi > -uCCi Ossai > ossī > ussi > S. yssı̯ ✧ WJ/400
goloði > goluði -ŏCi > -uCi ñgolodō > S. Goelydh ✧ WJ/364
goloði > goluði -ŏCi > -uCi ñgolodō > S. gœlyð ✧ PE17/139
oroni > oruni -ŏCi > -uCi oronī > S. Eryn ✧ PE17/33
oroti > oruti -ŏCi > -uCi OR/ORO > S. eryd/ered ✧ PE17/64
θolohi > θoluhi -ŏCi > -uCi ÞOL > S. thely ✧ PE17/188
θoli > θuli -ŏCi > -uCi ÞOL > S. thuil ✧ PE17/188
θoni > θuni -ŏCi > -uCi thŏno > S. thuin ✧ PE17/81

N. short [e], [o] became [i], [u] in syllable before final [i]; [-{ĕŏ}{C|CC}i] > [-{iu}{C|CC}i]

GS/§4.94 (raising)

Phonetic Rule Elements

[-ĕCi] > [-iCi]
[-ĕCCi] > [-iCCi]
[-ŏCi] > [-uCi]
[-ŏCCi] > [-uCCi]

Phonetic Rule Examples

felɣi > filɣi -ĕCCi > -iCCi ON. phelga > N. fili ✧ Ety/PHÉLEG
telxi > tilxi -ĕCCi > -iCCi ᴹ√TÉLEK > N. tilch ✧ Ety/TÉLEK
xendi > xindi -ĕCCi > -iCCi ᴹ√KHEN-D-E > N. hint/hinn ✧ Ety/KHEN-D-E
breθeli > breθili -ĕCi > -iCi ᴹ√BERÉTH > N. brethil ✧ Ety/NEL
dūmeni > dūmini -ĕCi > -iCi ᴹ√NDŪ > N. dúvin ✧ EtyAC/NDŪ
eleði > eliði -ĕCi > -iCi ᴹ√ÉLED > N. Elidh ✧ Ety/ELED
ereki > eriki -ĕCi > -iCi ᴹ√ERÉK > N. erig ✧ Ety/ERÉK
fereni > ferini -ĕCi > -iCi ON. pheren > N. †ferin ✧ Ety/BERÉTH
gwaθeli > gwaθili -ĕCi > -iCi ON. wathel > N. gwethil ✧ Ety/THEL
gwenī > gwini -ĕCi > -iCi ᴹ✶winyā > N. gwîn ✧ Ety/WIN
keperi > kepiri -ĕCi > -iCi ᴹ√KEPER > N. cebir ✧ Ety/KEPER
lalmeni > lalmini -ĕCi > -iCi ᴹ√ÁLAM > N. lelvin ✧ Ety/ÁLAM
lalmeni > lalmini -ĕCi > -iCi ᴹ√LÁLAM > N. lelwin ✧ Ety/LÁLAM
neni > nini -ĕCi > -iCi ᴹ√NEN > N. nîn ✧ Ety/NEN
pelehi > pelihi -ĕCi > -iCi ON. pelehi > N. peli ✧ Ety/PEL(ES)
peneði > peniði -ĕCi > -iCi ᴹ✶kwenedē > N. penidh ✧ Ety/KWEN(ED)
xarmeni > xarmini -ĕCi > -iCi ᴹ√KHYAR > N. herfin ✧ EtyAC/KHYAR
xeni > xini -ĕCi > -iCi ᴹ√KHEN-D-E > N. hîn ✧ Ety/KHEN-D-E
andondi > andundi -ŏCCi > -uCCi ᴹ√AD > N. ennyn ✧ Ety/AD
dolθi > dulθi -ŏCCi > -uCCi ᴹ√NDOL > N. dylt ✧ Ety/NDOL
tolli > tulli -ŏCCi > -uCCi ᴹ✶tollo > N. tyll ✧ Ety/TOL²
θalɣondi > θalɣundi -ŏCCi > -uCCi ᴹ✶stalgondō > N. thelyn ✧ Ety/STÁLAG
orni > urni -ŏCCi > -uCCi ᴹ✶ÓR-NI > N. yrn ✧ Ety/ÓR-NI
orxui > urxui -ŏCCi > -uCCi ON. orkui > N. †yrchy ✧ EtyAC/ÓROK
amboni > ambuni -ŏCi > -uCi ᴹ√AM² > emuin > N. emyn ✧ Ety/AM²
dori > duri -ŏCi > -uCi N. †dy̆r > N. dȳr ✧ PE22/38
poti > puti -ŏCi > -uCi ON. poti > N. pŷd ✧ Ety/POTŌ