Supplement 1: Historical Phonology — How Sindarin Evolved

Why Historical Phonology Matters

Most learners encounter Sindarin mutations as a wall of tables: soft mutation does this, nasal mutation does that, mixed mutation is something else entirely. Without context, these tables are arbitrary rules to memorize — and they will be forgotten as quickly as they are learned.

Historical phonology changes everything. Once you understand why mutations exist — that they are the fossilized remains of ancient sound changes that swept through the Elvish language over thousands of years — the rules become predictable. You no longer memorize that p becomes b after the article i; instead, you understand that i historically ended in a vowel, that voiceless stops weakened to voiced stops between vowels in Old Sindarin, and that when the final vowel of i was lost, the mutated form of the following consonant was preserved. The "mutation" is just a vowel that no longer exists.

This supplement is for learners who have completed lessons 1–41 and want to understand the system at a deeper level. It is also essential preparation for:

Reading Tolkien's linguistic essays (Parma Eldalamberon, Vinyar Tengwar, The War of the Jewels)
Doing your own Neo-Sindarin reconstruction work
Understanding why different scholars reconstruct disputed forms differently

A note on notation: PE = Primitive Elvish (also called Primitive Quendian); CE = Common Eldarin; CT = Common Telerin; OS = Old Sindarin; S. = Classical Sindarin (the language of the Lord of the Rings period). Reconstructed forms are marked with an asterisk (*).

The Elvish Language Family Tree

Tolkien was a professional philologist, and he constructed his languages with the same rigorous family-tree logic that historical linguists apply to real language families like Indo-European. The tree below shows the branching that produced Sindarin:

Primitive Quendian (language of the first Elves at Cuiviénen)
    │
    ├── Avarin languages (Elves who refused the Great Journey westward)
    │       Six known Avarin languages; mutually unintelligible by the Third Age
    │
    └── Common Eldarin (language of the Eldar who began the westward march)
            │
            ├── Quenya branch
            │       ├── Quenya proper (Noldor + Vanyar in Valinor)
            │       └── Vanyarin Quenya (slightly divergent dialect of the Vanyar)
            │
            └── Common Telerin
                    │
                    ├── Telerin proper (Falmari, coastal Elves of Aman)
                    │
                    └── Sindarin / Noldorin branch
                                │
                                ├── Classical Sindarin (LotR period; Gondor and Rivendell)
                                ├── Doriathrin (archaic dialect of Thingol's court in Doriath)
                                └── Falathrin (coastal dialect of the Falas; Círdan's people)

The key branching point for our purposes is the split between Quenya and Common Telerin. After this split, the two branches evolved independently for thousands of years — long enough to produce dramatic differences, yet sharing the same root vocabulary. This is why Sindarin and Quenya look so different on the surface but have systematic correspondences underneath: elen (Q.) and êl (S.) are the same word, both descended from PE *elen.

The internal diversity within Sindarin is also important: Doriathrin is older and more conservative (it preserved some features that Classical Sindarin lost), while Falathrin shows some unique innovations. Most Neo-Sindarin composition uses Classical Sindarin as its target.

Stage 1: Common Eldarin to Common Telerin

The first stage of divergence from Common Eldarin produced Common Telerin, the ancestor of both Telerin and Sindarin. The most important changes at this stage were:

Labiovelar Consonants

Primitive Elvish had the labiovelar series: *kw, *gw, *ŋgw (as in English "queen," "Gwen," and a nasalized "Gwen"). In Common Eldarin, these were maintained. The Quenya branch eventually simplified some of these (merging *kw with *p in some environments), but Common Telerin preserved the full system longer.

Sindarin ultimately simplified the labiovelars differently:

*kw → c (before non-u vowels) or cw (still written in some early Sindarin texts)
*gw → gw (maintained; cf. gwaith "people")
*ŋgw → ngw (maintained in initial position)

Intervocalic *b

Ancient *b between vowels weakened to v in Common Telerin. This spirantization (stop → fricative) between vowels is a typologically common process, found in Spanish (where Latin b between vowels became v/[β]). This is the same process that later, in Old Sindarin, drove the entire soft mutation system.

Stage 2: Common Telerin to Old Sindarin — The Great Vowel Loss

This is the single most important stage in Sindarin's development. Without understanding it, nothing else makes sense.

The Loss of Final Short Vowels

In Old Sindarin, all word-final short vowels were lost. Every word that ended in a short -a, -e, -i, -o, or -u dropped that final vowel entirely.

The consequences were massive:

Two-syllable words became monosyllables:

CE *atar (father, 2 syllables) → OS → S. adar (the vowel changes are separate; this shows the structural point)
CE *elen (star) → OS *eln → (vowel adjustment) → S. êl
CE *ondo (rock, stone) → OS *ond → S. ond

Three-syllable words became two syllables:

CE *eleni (stars, plural) → stage-by-stage loss of -i → changes detailed in Stage 3

Long vowels were generally preserved, and final long vowels (like -â, -ê, -ô) survived longer before eventually shortening in later stages.

This loss of final short vowels is why Sindarin looks so different from Quenya. Quenya was developed in Valinor, where the language evolved in a relatively conservative direction — Quenya words still typically end in vowels (elen, ando, findo). Sindarin, developed in Middle-earth over thousands of years of the Great Journey and subsequent history, lost all those vowels and became a language of crisp, often monosyllabic words (êl, annon, find).

Consonant Clusters at Word-End

When final vowels were lost, consonant clusters that had previously been word-medial (protected by the following vowel) became word-final. Many of these simplified:

-nd (final) → -nn in most environments
-mb (final) → -mm
-ng (final) stayed -ng (this nasal cluster was stable)

These simplifications explain many apparent irregularities. When a word unexpectedly gains a cluster in a derived form, it is often because the single-consonant final was itself a simplification of an earlier cluster that the derived form restores.

Stage 3: The I-Affection (Prestanneth)

I-affection (Sindarin prestanneth "disturbance/influence") is the single most important process for understanding Sindarin plurals. It is a form of umlaut — the same process that produces English "foot/feet," "mouse/mice," and German Mutter/Mütter. In all these cases, a vowel in a later syllable causes a vowel in an earlier syllable to shift, and then the triggering vowel may be lost.

The Mechanism

In Common Telerin, the plural suffix was -î (a high front vowel). Words had forms like:

*elessî (stars, plural) with the -î marking the plural

This -î was a high front vowel: it was pronounced in the front of the mouth with the tongue raised high. When the preceding syllables contained vowels that were pronounced further back or lower (a, o, u, e), those vowels shifted toward the position of the -î under the influence of anticipatory articulation.

Then, in the Old Sindarin period, when final short vowels were lost, the -î itself disappeared — leaving only the changed vowels in the preceding syllables as the sole marker of plurality.

The plural markers are, in other words, the ghost of a suffix that no longer exists.

Specific Vowel Changes from I-Affection

Vowel before lost -î	Result	Example (sg → pl)
a (non-final syllable)	→ e	aran (king) → erain (kings)
a (final open syllable)	→ ai	tal (foot) → tail (feet)
â (long a)	→ ai	dân → dain
e (non-final syllable)	→ i	edhel (elf) → edhil (elves)
ê (long e)	→ î → i written	têw (letter) → tîw (letters)
o (any position)	→ y (front rounded)	orch (orc) → yrch (orcs)
ô (long o)	→ ŷ	dôr (land) → dŷr (lands)
u	→ y	tulus → tylys
û (long u)	→ ui	dûr (dark) → duir
au (diphthong)	→ oe	naug (dwarf) → noeg (dwarves)

The y vowel (see Stage 7 below) is the most distinctive product of this process.

Why This Matters for Learning

When you see a Sindarin plural that looks nothing like the singular, i-affection is almost always the cause. Once you know the table above, you can predict plurals from singulars and singulars from plurals, rather than learning each as a separate vocabulary item:

See yrch, know immediately the singular is orch
See edhil, know immediately the singular is edhel
See erain, know immediately the singular is aran

Stage 4: The Origin of Soft Mutation (Lenition)

Soft mutation is the most frequently encountered mutation in Sindarin texts. It appears:

After the definite article i
On direct objects of verbs
After many common prepositions (o, na, nu, an in certain uses)
On adjectives following their nouns

None of this is arbitrary. All of it traces back to a single phonological process in Old Sindarin.

The Phonological Environment

In Old Sindarin, a vowel-final word immediately preceding a consonant-initial word created an environment in which the initial consonant of the second word was flanked by vowels on both sides: [vowel] + [consonant] + [vowel]. This is the classic intervocalic position — the position in which consonants weaken most readily across all human languages.

The weakening followed a predictable path:

Step 1: Voiceless stops weaken to voiced stops: *p → *b, *t → *d, *c (k) → *g

This is natural lenition: the voicing feature spreads from the surrounding vowels into the consonant. Every stop became voiced.

Step 2: Voiced stops weaken further to voiced fricatives (spirants): *b → *v (bilabial spirant), *d → *dh (dental spirant), *g → *ɣ (velar spirant, eventually deleted)

A spirant is a "leaky" sound — instead of fully stopping the airflow (like a stop) and then releasing it, the articulators approximate without full closure, allowing continuous airflow with turbulence. Voiced spirants are weaker than voiced stops.

Step 3: Final vowels of trigger words are lost:

This is the crucial moment. The vowel that caused the weakening — the vowel that ended the article i, the preposition o, the pronoun le — was itself deleted by the same final-vowel-loss process that we described in Stage 2. But the mutated consonant remained.

Result: The definite article i historically ended in a vowel (*ī or *ia or similar). That final vowel triggered lenition of the following consonant, then was lost. Now i ends in nothing — but the following consonant is still in its lenited form, as if the vowel were still there.

This is what a mutation is: a fossilized sound change that operated in a phonological environment that no longer exists.

The Complete Soft Mutation Cascade

Original consonant	After voicing (Step 1)	After spirantization (Step 2)	Sindarin result
p	b	(stops here for p)	b
t	d	(stops here for t)	d
c (k)	g	(stops here for c)	g
b	(already voiced)	v	v
d	(already voiced)	dh	dh
g	(already voiced)	ɣ → deleted	(disappears)
m	(nasal, different path)	v	v
lh	l (voiced)	stays l	l
rh	r (voiced)	stays r	r
hw	w (voiced)	stays w	w

The disappearance of mutated g is explained by this cascade: g was the voiced velar stop; its spirantized form ɣ (like the sound in Spanish lago or Greek γ) is very weak and easily deleted entirely. So i + galadh → i 'aladh (the tree; the apostrophe marks the lost g).

Stage 5: The Origin of Nasal Mutation

The definite plural article in Old Sindarin was *in — a form with a preserved final nasal. When this nasal-final article preceded a consonant-initial word, the nasal assimilated to the following consonant rather than the consonant weakening (as in soft mutation).

Nasal Assimilation

A nasal consonant immediately before a stop tends to pull that stop toward itself, creating prenasalized stops or causing full nasalization:

Voiceless stops:

in + p → *imp → *imb → mb → the voiceless stop gains a nasal onset
in + t → *int → *ind → nd
in + c → *iŋk → *iŋg → ng (ŋg)

Voiced stops:

in + b → *imb → mb → eventually the b fully assimilates to m
in + d → *ind → nd → n
in + g → *iŋg → ng → ng (stays as the nasal velar cluster)

The Scholarly Debate

The nasal mutation of voiceless stops is one of the most debated points in Neo-Sindarin reconstruction. Tolkien's own notes show some inconsistency, and the attested data is sparse. Two main positions exist:

Position A (aspiration): p → ph, t → th, c → ch after in. This treats the nasal as having aspirated the voiceless stop rather than voiced it.

Position B (voicing): p → b, t → d, c → g after in. This treats nasal mutation of voiceless stops as equivalent to soft mutation (voicing only).

The intermediate forms (mb, nd, ŋg) are where these positions disagree about what happened next. Did mb → m + aspirated p (giving ph), or did mb → b (losing the nasal and keeping only the voiced result)?

Working recommendation for Neo-Sindarin composers: Use p → b (or ph), t → d, c → g for the voiceless stops under nasal mutation. Regardless of position, the voiced stops are uncontroversial: b → m, d → n, g → ng.

Stage 6: The Origin of Mixed Mutation

Mixed mutation occurs after prepositions that incorporate the definite article: erin (er + in), uin (o + in), nan (na + in), and similar compounds.

The "mixed" character comes from two simultaneous pressures:

The preposition itself was historically vowel-final (or had a form that triggered soft mutation)
The incorporated article in was nasal-final (triggering nasal mutation)

When both pressures applied simultaneously to the following noun:

Voiceless stops felt the soft-mutation pressure (voicing): p → b, t → d, c → g
Voiced stops felt the nasal-mutation pressure (nasalization): b → m, d → n, g → ng

This "split" behavior — leniting the voiceless but nasalizing the voiced — is exactly what distinguishes mixed mutation from the other two. It preserves a historical two-step process that operated differently on consonants of different voicing.

Stage 7: The y Vowel — Sindarin's Most Distinctive Feature

The vowel spelled y in Sindarin (and pronounced as German ü, French u, or IPA [y]) has no precedent in Primitive Elvish. It arose entirely through i-affection operating on back rounded vowels.

The Derivation of y

When o or u underwent i-affection (Stage 3), they did not simply shift forward to e or i like the unrounded vowels did. Instead, they shifted forward while keeping their lip-rounding, producing a front rounded vowel [y]. This is a typologically normal result: French developed [y] the same way from Latin u under the influence of -i suffixes.

o + -î → [y] (written y)
u + -î → [y] (also written y)
ô (long) + -î → [ŷ] (long front rounded)
û (long) + -î → ui (the long vowel diphthongized rather than remaining monophthong)

Practical examples:

orch "orc" (with o) → yrch "orcs" (with y)
tulus "poplar tree" → tylys "poplar trees"
mund → mynd (in some derivations)

This vowel is also found in non-plural contexts wherever original u or o was in an environment that triggered fronting, and it explains why some Sindarin words have y where their Quenya cognates have o or u.

The Welsh language uses a very similar vowel — Welsh u is pronounced [ɨ] in north Wales (close-central) and often approximates [y] in south Wales. Tolkien, a lover of Welsh, designed Sindarin's phonology to include this sound specifically because it creates the distinctive "Celtic" flavor he was aiming for.

Stage 8: Other Key Late Changes

The Development of ae and oe

The diphthongs ae and oe are characteristically Sindarin (Quenya does not have them). They developed from:

*ai in certain positions → ae: CE *aira → S. aer (holy; the diphthong shifted)
*au with i-affection → oe: CE *nauki → naug → noeg (dwarves) (the au was pulled toward oe by i-affection)

The Aspirated Liquids lh and rh

Primitive Elvish had voiced l and r which were normal liquids. In initial position in Sindarin, these developed a voiceless/aspirated variant:

Certain initial l → lh [ɬ] (the Welsh ll sound, a voiceless lateral fricative)
Certain initial r → rh [r̥] (a voiceless r, like Welsh rh)

These arose in positions where the consonant was historically preceded by a voiceless consonant that was later lost, leaving the voicelessness as an aspiration on the surviving liquid. In soft mutation, lh → l and rh → r (the aspiration is removed when the leniting environment adds voicing).

The Sound Change Summary Table

The following table shows 15 key words from Primitive Elvish through to Classical Sindarin, with intermediate stages where relevant. Intermediate forms marked with * are reconstructed.

PE form	CE / CT form	OS form	Classical Sindarin	English
*kwendî	*kwendi	*chwend	cwend	Elf (one of the speaking people)
*ñolodô	*ñoldo	*ngolod	golodh	Noldo / deep one
*ataré	*atar	*adar	adar	father
*eleni	*eleni	*elin	elin	stars (plural)
*elen	*elen	*el	êl	star (singular)
*erektê	*ereg	*ereg	ereg	thorn
*ñdûrô	*ndūro	*ndūr	dûr	dark
*NAUKÔ	*nauko	*naug	naug (sg.) / noeg (pl.)	dwarf
*orko	*orko	*ork	orch (sg.) / yrch (pl.)	orc
*talat	*talat	*talt	talt	slipping, giving way
*kwettā	*kwetta	*peth	peth	word
*anaro	*anar	*anor	Anor	sun
*ithilê	*ithil	*ithil	Ithil	moon
*morô	*moro	*mor	mor	dark (in compounds)
*galadā	*galada	*galad	galadh	tree

Note on *ataré → adar: the e → a change is an example of a back-shift in unstressed syllables that operated in Old Sindarin.

Note on galadh: the final -dh comes from older *-d spirantizing to [ð] (the voiced dental fricative) in word-final position — another lenition process, this time word-internally rather than in the sandhi environment.

Sindarin and Welsh: The Phonological Parallels

Tolkien was open about his deep love for Welsh and his intention to give Sindarin "the Welsh feeling." The following correspondences are not coincidences — they are deliberate design choices:

Welsh Feature	Welsh Example	Sindarin Equivalent	Sindarin Example
Soft mutation (treiglad meddal)	pen → ben	Sindarin soft mutation	peth → beth
Nasal mutation (treiglad trwynol)	pen → mhen/mhen	Sindarin nasal mutation	pen → ben (or phen)
Aspirate mutation (treiglad llaes)	pen → phen	Sindarin mixed mutation	pen → ben (in mixed)
Vowel u pronounced [ɨ] or [y]	tŷ (house)	Sindarin y [y]	yrch (orcs)
Voiceless lateral ll [ɬ]	llan (parish)	Sindarin lh [ɬ]	lhaw (ears)
Voiceless rh [r̥]	rhedeg (to run)	Sindarin rh [r̥]	rhass (precipice)
Double ff [f]	ffon (staff)	Sindarin ph [f]	Pheriain
Umlaut plurals	brawd → brodyr (brother/brothers)	Sindarin i-affection plurals	orch → yrch
Prefixed definite article	y/yr/y'r	Sindarin i/in	i mbas (the bread)
VSO word order (tendency)	Mae John yn canu (Is John singing)	Sindarin VSO	pêd Aragorn (speaks Aragorn)

The most striking parallel is the umlaut plural system. Welsh, Scottish Gaelic, and other Celtic languages have systems in which vowels in noun stems change to mark plurality, rather than (or in addition to) adding suffixes. Tolkien imported this entire grammatical mechanism into Sindarin, grounding it in a historically plausible backstory (the -î suffix).

Sound Correspondences: Sindarin vs. Quenya

Because both languages descend from Common Eldarin, they have systematic correspondences — the same root surfaces differently in each language by regular rule. These correspondences are important for:

Identifying cognates (same root, different language)
Reconstructing Neo-Sindarin words from attested Quenya roots

Quenya feature	Sindarin corresponding feature	Example (Q. / S.)
Word-final vowels preserved	Word-final vowels lost	elen / êl (star)
qu [kw]	p (before front vowels)	quet- / ped- (speak)
ñ (initial) → n	ñ → g (initial velar nasal → stop)	Noldo / golodh
nd- (initial) → n	nd- → d	norna / dorn (hard)
mb- (initial) → m	mb- → b	mbār → már / bâr (home)
ld medial → ll	ld → lth or ll	alda / galadh (tree)
r (all positions)	r or rh (position-sensitive)	aran / aran (king; same)
-e final (noun)	lost	quenye / peth (kwet + vowel change)
-o final (noun)	lost	Noldo / golodh
nw- (initial)	dh- or n-	nwalme / gwalch
Diphthong ai	ae	naira / naer (sad)
au	o or au depending	aure / or (day; in compounds)

The most practically useful correspondence is Q. qu- = S. p- (before non-u vowels). Whenever you see a Quenya word beginning with qu, the Sindarin cognate (if it exists) will likely begin with p: quenta (tale) → pent (tale; cf. belegpent), quel- (to fade) → pel- (to wane; cf. Pelennor).

Implications for Learners: Connecting History to Practice

The following list translates historical facts into practical answers to common learner questions:

"Why does g disappear in soft mutation?" Because g is a voiced velar stop. In the soft mutation environment (intervocalic), voiced stops spirantized: g → ɣ. The velar fricative ɣ is extremely weak and was deleted entirely. The "disappearance" is deletion, not magic.

"Why does m become v in soft mutation but stay m in nasal mutation?" In soft mutation (intervocalic): m is a nasal consonant that, in the leniting environment, spirantized to v (the bilabial continuant). In nasal mutation: the nasal article -n reinforced the nasality of m, keeping it stable as m.

"Why are Sindarin plurals formed by vowel change rather than a suffix?" Because the suffix (-î) caused umlaut in the preceding syllables and then was itself deleted by final-vowel-loss. The vowel change is the residue of the suffix. There was originally a suffix; it simply no longer exists.

"Why do some words have lh and rh initially?" These are voiceless liquids arising from historical consonant clusters where a voiceless consonant preceded l or r. The preceding consonant was deleted; its voicelessness was preserved as aspiration on the surviving liquid.

"Why does the article trigger soft mutation?" Because the article historically ended in a vowel, creating the intervocalic leniting environment. The vowel was lost; the mutation was not.

"Why are Sindarin and Welsh so similar?" Because Tolkien designed Sindarin to capture the phonological feel of Welsh — especially its mutation system, its front-rounded vowel, its voiceless liquids, and its umlaut plurals. These are design choices, not coincidences, and understanding Welsh deeply helps with Sindarin.

Practice: Sound Change Exercises

For each of the following Primitive Elvish roots, trace the expected development to Classical Sindarin. Use the stages outlined above. Check your answers against Eldamo (eldamo.org) if available.

*TAWAR (wood, forest) — what Sindarin word do we expect? What is the cognate in Quenya?
*PHAN (full, complete) — the Quenya word is quanta. What should the Sindarin be, and why does the qu → p change apply?
*KHIM (adhere, stick to) — expect the initial kh to change how? What Sindarin prefix does this give?
*NDOR (land, dwelling) — initial nd- → what in Sindarin? How does this produce dôr (land)?
*GAL (shine, be bright) — this root should give a verb. What are the singular and plural noun forms you would expect, and what vowel change would i-affection produce?

Reflection questions:

After studying this supplement, go back to the soft mutation table in Lesson 11. Can you now derive every row of that table from the phonological history?
Look at the plural forms you learned in Lessons 6–7. Can you identify which vowel changes are i-affection and which are not?
Compare the word for "star" in both Quenya (elen) and Sindarin (êl). Can you trace each step from PE *elen to both reflexes?

This supplement is based on Tolkien's own linguistic essays published in Parma Eldalamberon (PE) and Vinyar Tengwar (VT), particularly PE17, PE18, and PE22, as well as the scholarly reconstructions in J.R.R. Tolkien's The War of the Jewels and Morgoth's Ring. The reconstruction framework broadly follows the work of Helge Fauskanger, Thorsten Strack, and the Lambengolmor mailing list community.