In Sindarin and Noldorin the short vowels [a], [o], [u] were all moved forward in the mouth when they preceded a syllable containing the vowel [i], developing into [e], [œ], [y] respectively. This sort of sound change is a fairly common in many languages, where the front vowel [i] attracts other vowels further back in the mouth, causing them to move forward. This shift occurred even when there were other intervening syllables between the a/o/u and the syllable containing i. This sound change is the second stage of i-affection, called i-fronting in this lexicon following the suggestion of Elaran from a private Discord chat on 2018-08-25.

This sound change was described as “secondary affection” or “interior i-affection” in Jim Allen’s Introduction to Elvish published in 1978 (pp. 112-116, 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/§4.104, 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 aligns mostly closely with the available evidence. A similar set of sound changes occurred in Welsh (non-ultimate i-affection) (WGHC/§70). However, Bertrand Bellet suggested the fronting of vowels in Sindarin is actually closer to Germanic phonetic developments (VASN); these three vowel shifts are nearly identical to the German umlaut.

The clearest decription of this sound change appears in a “Table of Vowel Notations in Exilic Noldorin” from a discussion of the Noldorin use of the Feanorian Alphabet from the 1930s (PE22/38), with excerpts of relevant rows only:

Phon[etic] Value	Origin	Archaic	Later Gondolic	Late Exilic
[æ]	affection of ă	]G or ]Ö æ	> [e] q.v.
[œ̆]	affection of ŏ	hG hÖ h|	h| > [e], q.v.
[y̆]	affection of u, o	.Ö, .G ü or ® y	® y	® y

These three rows are the only “affections” mentioned in this table. As described here, the basic affections were:

• a > æ > e.
• o > œ > e.
• u > ü or y.

There is a note on the preceding page that makes it clear that [œ] and [y] are also the result of i-affection:

The new sounds [œ̆] and [y] produced by i-affection were denoted in various ways: thus [œ] by hÖ and by h|, [y] by .Ö and ® (PE22/37).

Somewhat further down there is a note indicating that the intermediate form [æ] was altered earlier than the rest:

The {mutation >>} i-affection of ă was distinguished from ĕ only in early Gondolic documents, where the spelling ]G or ]Ö transcribed [æ] appears (PE22/37 and note #122).

Because of the comparatively early nature of shifting [æ], this lexicon does not treat [a] > [æ] and [æ] > [e] as a distinct phonetic developments, combining them only as [a] > [e]. There is less evidence for the intermediate form [æ] than there is for [œ], which lingered long enough to appear in some published words, such as archaic S./N. †arnoediad [arnœdiad] “unnumbered” in S./N. Nirnaeth Arnoediad versus the more ordinary form S./N. arnediad (S/192; WJ/28; Ety/NOT; EtyAC/NOT). Because of words like these, the later shift whereby [œ] became [e] is listed as a separate phonetic development.

The i-fronting sound changes went on to influence various grammatical systems in both Sindarin and Noldorin.

I-fronting in Sindarin Plurals: The result of these sound changes can be seen most clearly in Sindarin plurals, where the ancient plural suffix -ī caused various sound shifts before it vanished. Consider the phonetic development in the initial syllables of the following three plurals:

• S. erain plural of S. aran “king” (PE17/40, 111).
• S. Gelydh plural of S. Golodh “Noldo” (S/238), with earlier form Gœlydh (PE17/139, WJ/364).
• N. tylys plural of N. tulus “poplar-tree” < ᴹ✶tyulussē [tjulussē] (Ety/TYUL).

The likely phonetic developments of these plurals are:

• *aranī > (fronting) *ereni > (intrusion) *erein > erain
• *golodī > (raising) *goludhi > (fronting) *gœlydhi > gœlydh > gelydh
• *t(j)ulussī > (fronting) *tylyssi > tylys

In the initial syllables, the development of a (> e) and u (> y) clearly demonstrates i-fronting, but the fronting of o in the initial syllable is partly obscured because later œ became e. The fronting of u > y is also clear in the final syllable of tylys. However, the fronting of o does not occur in the final syllable of plurals, because of the earlier sound change (i-raising) whereby short o became u in a syllable before final i and thereafter developed into the vowel y. Tolkien actually mentions this phonetic development of o in final syllables in his notes on the Feanorian Alphabet from the 1930s:

[œ̆]* This was never long. In final stressed syllables - where short vowels were lengthened - it did not occur, since here affection of o was > y (PE22/38).

Similarly, the i-fronting of a in final syllables of plurals is obscured by the later sound change (i-intrusion) whereby the plural suffix i intruded into the preceding syllable to produce the diphthong ei, which then developed into ai (in Sindarin, not always in Noldorin).

I-fronting in the Sindarin Present Tense: This sound change was also a factor in the Sindarin and Noldorin present tense. The bulk of the evidence is that Sindarin/Noldorin present was derived from the more ancient aorist tense, which used the primitive suffix -ĭ as the basis for its inflection. In Primitive Elvish, this short final i became e when no further suffixes were added, but was retained when there was a pronominal suffix. This can be seen most clearly in the Quenya aorist, where the uninflected form is Q. care “make, do” (PE17/144; PE22/154) but the inflected form is Q. carin “I make, I do” (PE22/152; Ety/KAR).

In Sindarin, the short final e vanished in the uninflected form (care > car), and later still the short vowel lengthened as it usually did in monosyllables to produce the uninflected present tense: S. câr “make, do” (PE17/132, 145). In inflected forms, the i was retained, so that any appropriate vowels in the preceding syllable were fronted (a > e) and the inflected present tense becomes: S. cerin “I make, I do” (PE17/132; VT50/22). Unfortunately, we don’t have very many examples of the Sindarin/Noldorin present tense, and most of the published forms demonstrate only the mutation of a > e:

• S. cewin “I taste” from S. caw- (PE22/152).
• S. sevin “*I have” from S. sav- (PE17/173).
• N. gerin “I hold, have” from N. gar- (Ety/ƷAR|GAR).

Presumably basic verbs containing o and (rare) u would show similar mutations, so we would expect the inflected present tense of S. tôl “come” (WJ/254) to be *telin “I come” (< *tœlin), and the present tense of N. tôg (Ety/TUK) “bring” to be tegin “I bring”. The form tegin actually appears (unglossed) in The Etymologies, but it was deleted (EtyAC/TUK). To muddy the waters, we do have an attested present tense form tolen “I am coming” that seems to contradict this pattern, but that appears to be a variant behavior of the Sindarin present tense rather than a change in this phonetic rule. Sindarin basic verbs containing u would be extremely rare since short u usually became o. However, a hypothetical verb like *run- “to rub, grind” (from the root √RUN² of the same meaning) would be expected to have an inflected present tense *rynin “I rub” (having a following nasal like n is one of the few cases where short u does not developed into o).

Other Examples of I-fronting: Many of the clearest examples of i-fronting come for Sindarin/Noldorin plurals and present tense forms, but there are other examples as well. Ancient compounds whose second element contains an i demonstrate i-fronting:

• S./N. Belthil “Divine Radiance” [from BAL + THIL] (SA/sil; Ety/BAL, THIL).
• S./N. Glewellin “Song of Gold” [from glawar + lin(d)] (PE17/61; Ety/LÁWAR).
• N. feredir “hunter” [from fara- + (n)dīr] (Ety/SPAR).

This was not true of late or reformed compounds, such as S./N. r(h)andir “wanderer” (PE17/60; Ety/RAN) or S./N. curunír “wizard” (PE22/151; Ety/KUR). The same is true for the plurals of recognized compounds, where the first half was generally reformed to be consistent with its singular form: S. rochben “rider” (roch + pen) has a plural form rochbin rather than expected **rechbin, because it was reformed from its archaic plural †rœchbin to better match its singular (WJ/376).

I-fronting was also a factor in the development of primitive -ya verbs, where the medial consonantal y became i:

• √AL(A) [> alya-] > S. elia- “to cause to prosper” (PE17/146).
• √THAR [> tharya-] > S. theria- “to be vigorous, flourish” (PE17/187).
• ᴹ√KHOR [> khorya-] > hoerio [hœrio] > N. herio “to begin suddenly” vs. ᴹQ. horya- (Ety/KHOR).
• √NOT [> notya-] > nœdia- > S./N. nedia- “to count” (PE17/145; WJ/28; Ety/NOT).

There is an ancient feminine suffix -ī that appears in some words, and Sindarin derivatives of these words also demonstrate i-fronting:

• ✶barathī(e) > S. bereth “queen” (PE17/23; MR/387).
• ✶herūnī [> ?heruni] > S. heryn “lady” (PE17/97-98).

These last two examples show developmental irregularities of their own, so as examples they should be taken with a grain of salt.

Possible Mechanics: The data model in this lexicon represents this sound change as occurring before an [i] in a following syllable, but it could also have occurred before consonantal y [j] or asyllabic [ı̯]. The system is complicated by the fact that medial [j] became [i], and it isn’t clear which change took place first. This lexicon assumes [j] became [i] first, because it simplifies the data model, but that need not be the case.

In addition, there are Noldorin examples that seem to show i-fronting before either the diphthong ui or the vowel y, as pointed out by Bertrand Bellet in his essay on Vowel Affection in Sindarin and Noldorin (VASN):

• ON. malui > †mely archaic plural of N. (h)mâl “pollen” (Ety/SMAL).
• ON. rankui > †rhengy archaic plural of N. rhanc “arm” (Ety/RAK).
• ON. orkui > †yrchy archaic plural of N. orch “Orc” (Ety/ÓROK).

These irregular plurals developed because in (Old) Noldorin the diphthong ui did not follow the rule whereby final i-diphthongs became long [ī] in polysyllables; instead it developed into vocalic [y] (probably from [ȳ]). As indicated by ON. malui > N. †mely and ON. rankui > N. †rhengy, the resulting final y seems to cause i-fronting: a > e. Bertrand Bellet suggested (VASN) that it is possible that the vowel y also caused i-fronting when it appeared medially, for example where it arose when primitive diphthongs [eu] and [iu] became [ȳ]. However, we have no attested examples of this, and such phonetic developments would probably have been rare.

The third example ON. orkui > N. †yrchy superficially appears not to follow the same pattern, but likely the [o] was first raised to [u] before being fronted to [y], as suggested by Bertrand Bellet (VASN). This suggests that i-fronting could also happen before final vocalic y [y], but such final y seems only to occur in Noldorin. In any case, it seems these developments were sufficiently unusual that all three archaic plurals were regularized into the normal Noldorin plural patterns: N. meil, N. rhenc, N. yrch (Ety/SMAL, RAK, ÓROK).

Conceptual Development: The initial element of polysyllabic Early Noldorin plurals from the 1920s frequently show e/e/y for a/o/y, so it seems likely that i-fronting dates all the way back to the introduction of the i-mutational plurals, or at least emerged soon after. For a complete list of examples, see Thorsten Renk’s Early Noldorin Plural Patterns (ENPP). Here is a selection of examples:

• ᴱN. eglerib plural of ᴱN. aglareb “glorious” (PE13/136, 158).
• ᴱN. emyth plural of ᴱN. amoth “shoulder” (PE13/159).
• ᴱN. beryn plural of ᴱN. boron “steadfast” (PE13/139).
• ᴱN. gelydh plural of ᴱN. Golodh “Gnome” (PE13/145, 162), along with deleted golydh (PE13/145) and [archaic?] variant goelydh (PE13/162).
• ᴱN. tynnib plural of ᴱN. tunneb “full of holes” (PE13/165).
• ᴱN. yrdyf plural of ᴱN. urduf “humorless” (PE13/156).

There are quite a few variations on plural forms, especially involving o: ᴱN. lylyth plural of ᴱN. loloth “poplar-tree” (PE13/149); ᴱN. gonnin plural of ᴱN. gonnen “of stone”, later replaced by the more regular plural ᴱN. gennin (PE13/123, 145). Thus, it seems that in the Early Noldorin of the 1920s, this sound change was still settling into its ultimate form. Roman Rausch looks at the possible phonetic developments in more detail in his Historical Phonologies of Ilkorin, Telerin and Noldorin around 1923 (HPITN/§4.2.1).

@@@ GS/§4.104 WGHC/§70 @@@