Bardh
Sindarin Crash Course by Rínor
Contents
Creating names through compounds involves more than joining two dictionary headwords. Earlier word forms, sound changes, the meaning of each element, and the time when a name was formed can all affect the result. Fiona Jallings’ consonant chart is a useful starting point. You can also explore Telpefindelë | Celephinnel’s Quettamorphosis.
The names below are my own creations unless identified as an attested comparison. Their quoted glosses record my intended meanings. The source notes distinguish Tolkien’s Sindarin vocabulary, older Gnomish or Noldorin material adopted for Neo-Sindarin, and uncertain reconstructions. A documented element does not by itself establish the whole name or every step of its proposed history.
Open My original proposed process to read my working chains. The six-pointed star (✶) marks an early form; it does not by itself tell you whether Tolkien supplied that form or a later writer reconstructed it. Eldamo uses a separate asterisk (*) for a reconstruction. See its notation guide and the source notes below.
For example, I created the sword name “Iron Light” using ang for “iron” and calad for “light”.
Source status: My own compound using Sindarin elements. ang and calad are documented, but the combined chain below is my proposal. The exact vowel loss and cluster reduction need their own support; the general cluster entry does not uniquely determine this result.
Let's run it through the process: ✶Aŋga·kalata > Angalad, AN-ga-lad, | [aŋga·kalata] > [aŋga·kalat] > [aŋga·galad] > [aŋggalad] > [aŋgalad]
Here is an attested example: ✶Gondo·ndor > Gondor, GON-dor, | [gondo·ndor] > [gondndor] > [gondor]
Attested comparison: Eldamo records this development of Gondor, citing Tolkien’s explanation in PE17/28. An attested reduction in one name is useful evidence, but its conditions still matter when considering another name.
A later compound can join already developed Sindarin elements. A reconstructed ancient compound starts with earlier forms and must account for the environments that existed at each stage. These approaches can produce different histories. You do not need to force every modern coinage through the entire historical rule list.
An example I received from Elaran on The Tolkien Forum contrasts Ceiradan and the more archaic proposal Ciriadan, with the intended meaning “Ship-Man”, using cair “ship”. These are forum proposals here, not examples of Tolkien attesting those complete name histories. The supplied development of the independent word is:
KIRjā > kirja > kerja > keria > keri > keir > kair
In the usual later pattern, ai occurs in a final syllable and ei appears when that syllable becomes non-final. The original examples erain and cair illustrate final ai. There are exceptions, so “ai can only appear in a final syllable” is too absolute. Compare rule 03900 and the historical account of cair. An element compounded before those changes occurred need not follow the history of the independent word.
Names do not always require two elements. Tolkien’s remarks below concern differentiation when using the names of things. They should not be expanded into a universal ban on a one-word name or epithet. Thurin is an attested name Finduilas gave Túrin, providing a direct counterexample to that broader claim.
The mere names of things, such as “hill, river, tree”, and especially of unique things, as “sun”, were not used, at any rate without differentiation: [...] for that would imply some kind of total equation or identity. An Elf (or Man) would not be called Anar “Sun” even to depict great glory or radiant vigour. [...] But an Elf or Man could be called Anárion, Anardil [etc.] [PE21/86]
Civaglar, ci-MAG-lar, “Blade of Glory”, cim “blade”, + aglar “glory, brilliance, splendour, radiance”
Source status: Neo-Sindarin adoption of older vocabulary. The name is presented here as my own creation.
cim comes from archaic Gnomish. Paul Strack adopts it for Neo-Sindarin and suggests a primitive form with mb. That reconstruction does not establish the simple m to v development in my chain, since historical mb follows a different path. aglar is Sindarin. The spelling of the name and its printed pronunciation also disagree at v/m; the original reading is retained below for review.
Process: ✶kim·aklara > kimaklar > kimaglar > kivaglar.
Cumeleg, CU-me-leg, “Mighty Bow”, cû “bow”, + beleg “great, mighty; large, big”
Source status: Sindarin elements; proposed historical compound. The name is presented here as my own creation.
cû has a documented earlier form containing m, so bringing that consonant into a historical analysis has a basis. beleg is also Sindarin, although its root history varies in Tolkien’s notes. The complete chain is my reconstruction. The restoration of a stop after a nasal and the later treatment of mb need to be evaluated in that historical order, rather than applied to any modern pair of words.
Process: ✶kūma·beleke > kūma·veleke > kūma·velek > kūma·veleg > kūmveleg > kūmbeleg > kumbeleg > kummeleg > kumeleg
Athaedor, a-THAE-dor “Healing Wood”, athae “*healing herb, healing”, + taur “forest, wood”
Source status: Sindarin elements; meaning and formation need qualification. The name is presented here as my own creation.
athae is a Sindarin element, but its exact sense and whether it is a noun, an adjective, or both are inferred. Tolkien’s later account treats it as a Quenya borrowing. taur refers to a forest or woodland, so my intended “Wood” should not be read as a claim that this element means wood as a material. Diphthong reduction is variable: au to o is common, while ae to e is rare. The full inherited history proposed below is therefore not established simply by joining these elements.
Process: ✶atʰajā·taurē > atʰaja·taure > aθaja·taure > aθaia·taure > aθaia·taur > aθaia·daur > aθaidaur > aθaedaur > aθaedor
Haldagnir, hal-DAG-nir, “Shadow Slayer”, N. hall¹ “veiled, hidden, shadowed, shady”, + dagnir “slayer”
Source status: Noldorin vocabulary combined with Sindarin. The name is presented here as my own creation.
hall¹ is a Noldorin adjective. Its adjectival meaning does not by itself establish the noun reading in my intended “Shadow Slayer”. dagnir is Sindarin and has a historical initial nasalized stop. The source history of the first element also needs to be distinguished from the starting form in my chain. Haldir offers a comparison for the first element, but does not prove every reduction in this longer cluster. Cluster simplification is a group of context-dependent changes, not permission to remove any middle consonant.
Process: ✶skalla·ndak·ndēr > sxalla·ndak·ndēr > xalla·ndak·ndēr > xalla·ndak·ndīr > halla·ndak·ndīr > halla·ndag·ndīr > hallndagndīr > hallndagndīr > haldagnīr > haldagnir.
Hennurin, HEN-nu-rin, “Secret Child”, hên “child”, + thurin “secret, hidden”
Source status: Sindarin elements; the proposed sound history needs review. The name is presented here as my own creation.
thurin is Sindarin, attested in Tolkien’s later writings. Its Ilkorin predecessor is thúren. The primitive reconstruction used in my chain belongs to the discussion of that Ilkorin form; it does not automatically explain Sindarin thurin. Eldamo explicitly notes uncertainty about the vowel history. hên is also Sindarin, but the vowel developments of the full compound need to be assessed in their own environment. Finally, internal nθ and n plus th at a recognized compound boundary need not behave alike. Compare boundary nasal loss and medial assimilation; the final step shown below is a proposal, not an automatic rule for every such compound.
Process: ✶kʰinā·tʰūrinā > kʰina·tʰūrina > xina·θūrina > xena·θūrina > hena·θūrina > hena·θūrin > henθurin > hennurin
Mirwen, MIR-wen, “Jewel Maiden”, mîr “jewel, precious thing, treasure”, + gwend “maiden”
Source status: Sindarin elements; an established name ending. The name is presented here as my own creation.
mîr and gwend are Sindarin, and -wen is an established feminine name ending. However, Tolkien’s discussion suggests that names used a form without the final dental consonant even in older spelling. The complete sequence through that consonant in my chain should not be presented as the demonstrated history of the ending. Shortening a long vowel before a cluster has a phonetic basis; the particular name history remains my analysis.
Process: ✶mīri·wende > mīre·wende > mīre·wend > mīrwend > mirwend > mirwenn > mirwen
Alweth, AL-weth, “Wholesome One”, *alu “wholesome”, + -eth “feminine ending”
Source status: Sindarin vocabulary with a Neo-Sindarin normalization. The name is presented here as my own creation.
Tolkien wrote alw; *alu is the usual Neo-Sindarin rendering described by Eldamo. -eth is an attested feminine ending with a documented primitive ancestor. It is not a generic gender-neutral ending meaning “one”. The gloss above records my intended name meaning; the proposed full development is not an attested Tolkien derivation.
Process: ✶alwa·itta > alwa·ittʰa > alwa·iθθa > alwa·eθθa > alwa·eθθ > alweθθ > alweθ
Tathreniel, tath-RE-ni-el, “Daughter of Willows”, tathren “of willow(s)”, + -iel¹ “daughter; feminine suffix”
Source status: Sindarin adjective and name ending; a later formation. The name is presented here as my own creation.
tathren is an adjective. -iel¹ can function as a daughter element or more generally as a feminine ending, so the kinship reading in my intended gloss is a choice, not something the ending guarantees. This short process joins already developed components. It illustrates a later formation, not a complete reconstruction from Primitive Elvish.
Process: ✶tatʰren·iel > tathreniel
Square brackets describe sounds. Historical notation can include sounds no longer present in later Sindarin. These tables retain the consonant-series and vowel-group layout used in this course. See Eldamo’s phonetic descriptions for its notation.
| Classifications | p-series | t-series | k-series |
|---|---|---|---|
| voiceless-stops | p | t | k (c) |
| voiced-stops | b | d | g |
| voiceless-spirants | f (f, ph) | s, θ (th) | h, x (ch) |
| voiced-spirants | v (v, f) | ð (dh) | †ɣ (gh) |
| nasals | m | n | ŋ (ñ) |
| voiceless-continuants | w̥ (wh) | l̥ (lh), r̥ (rh) | |
| voiced-continuants | w | l, r | j (i) |
| Classifications | Vowels |
|---|---|
| long-vowels | ī, (î, í), ē, (ê, é), ā, (â, á), ō, (ô, ó), ǭ, ę̄, ȳ, (ŷ, ý), ū, (û, ú) |
| short-vowels | ĭ (i), y, ĕ (e), †œ, ă (a), ŏ (o), ŭ (u) |
| diphthongs | ae, ai, ei, oe, ui, au, iw, eu |
| others | †ou; ø (no sound) |
The table follows the 78 entries in Paul Strack’s Sindarin phonetic index, checked on 14 September 2026. The columns show the rule, its example, and its number. The numbers follow Eldamo’s sequence; they are not instructions to apply every rule to every new compound.
The examples show selected stages, with the relevant change in bold. Click a rule title for the source discussion, or a [note] link for qualifications below the table. Noldorin comparisons are identified where needed. Quoted examples from Eldamo are attributed to it, since its phonetic analyses can include reconstructed intermediate stages.
Noldorin is a conceptual predecessor in Tolkien’s development of the language. Old Sindarin concerns an earlier stage within the fictional history. Historical sound change, grammatical mutation, and changes at compound boundaries must each be considered in their own context.
Rule titles and cited source examples follow Eldamo, credited to Paul Strack under CC BY 4.0. The notes distinguish the course’s working material from the source’s own analyses and variant forms.
| Rule | Example | No. |
|---|---|---|
| initial [w] became [gw] | ✶wath > gwath | [watʰa] > [waθa] > [gwaθa] > [gwaθ] | 01 |
| initial nasals vanished before stops | ✶mbara > bâr | [mbara] > [mbar] > [bar] > [bār] | 02 |
| final nasals vanished after vowels | ✶elen > êl | [elen] > [ele] > [el] > [ēl] | 03 |
| initial [s] vanished before spirants | ✶spindē > find | [spindē] > [spinde] > [sɸinde] > [ɸinde] > [finde] > [find] > [finn] | 04 |
| initial voiceless [j̊] became [x] [note] | [j̊erwa] > [xerwa] (Eldamo example, PE23/139) | 05 |
| voiced stops became spirants after liquids | ✶gardā > gardh | [gardā] > [garda] > [garða] > [garð] | 06 |
| [zb], [zg] became [ðβ], [ðɣ] | ✶buzbō > buðu | [buzbō] > [buzbo] > [buðβo] > [buðvo] > [buðv] > [buðw] > [buðu] | 07 |
| short [i], [u] became [e], [o] preceding final [a] [note] | ✶khīnā/khinā > hên | [kʰinā] > [kʰina] > [xina] > [xena] > [xen] > [hen] > [hēn] | 08 |
| voiced stops became spirants after vowels | ✶galadā > galadh | [galadā] > [galada] > [galaða] > [galað] | 09 |
| [ɸ], [β] became [f], [v] | ✶ap·pata > aphad- | [appat-] > [appʰat-] > [aɸɸat-] > [affat-] > [affad-] | 10 |
| medial [j] became [i] | ✶gilji > gîl | [gilji] > [gilī] > [gil] > [gīl] | 11 |
| short [e], [o] became [i], [u] in syllable before final [i] | ✶edelī > Edhil | [edelī] > [edeli] > [eðeli] > [eðili] > [eðil] | 12 |
| short [a], [o], [u] became [e], [œ], [y] preceding [i] | ✶mbarī > bair | [mbarī] > [mbari] > [bari] > [beri] > [beir] > [bair] | 13 |
| [ē], [ō] became [ī], [ū] | ✶kēwā > cîw | [kēwā] > [kēwa] > [kīwa] > [kīw] | 14 |
| [ɣ] vocalized before [l], [r], [m], [n] [note] | ✶tagra > taer | [taɣra] > [tagra] > [taɣra] > [taira] > [tair] > [taer] | 15 |
| [x], [ɸ] vocalized between a vowel and [θ] | ✶ektā > eitha- | [ektā-] > [ekta-] > [ektʰa-] > [exθa-] > [eiθa-] | 16 |
| non-initial [xʲ] vocalized to [ix] [note] | See Eldamo’s discussion and examples. | 17 |
| [iu] and [ju] became [ȳ] | ✶keurā > cŷr | [keurā] > [keura] > [kiura] > [kȳra] > [kȳr] | 18 |
| short [u] often became [o] [note] | ✶ñgurotho > gorth | [ŋgurotʰo] > [ŋguroθo] > [guroθo] > [goroθo] > [goroθ] | 19 |
| [nm], [ŋm] became [nw], [ŋw] | ✶tekmā > tēw | [tekmā] > [tegmā] > [tegma] > [teŋma] > [teŋwa] > [tēwa] > [tēw] | 20 |
| [ŋ] vanished with compensatory lengthening | ✶tekmā > tēw | [tekmā] > [tegmā] > [tegma] > [teŋma] > [teŋwa] > [tēwa] > [tēw] | 21 |
| [ǭ] became [au] | ✶astāra > astor | [astāra] > [astār] > [astǭr] > [astaur] > [astor] | 22 |
| [ę̄] became [ai] [note] | N. coe | [kǣm] > [kaim] > [kai] > [koe] (Noldorin comparison, Ety/KEM) | 23 |
| short final vowels vanished | ✶mapo > mâb | [mapo] > [map] > [mab] > [māb] | 24 |
| final [i] intruded into preceding syllable [note] | ✶edelī > Edhil | [edelī] > [edeli] > [eðeli] > [eðili] > [eðil] [beri] > [beir] (Eldamo example, PE17/164) |
25 |
| final [w] sometimes intruded into preceding syllables [note] | See Eldamo’s discussion and examples. | 26 |
| initial [x-] became [h-] | ✶khīnā/khinā > hên | [kʰinā] > [kʰina] > [xina] > [xena] > [xen] > [hen] > [hēn] | 27 |
| voiceless stops voiced after vowels | ✶akāra > agor | [akāra] > [akǭra] > [akaura] > [akaur] > [agaur] > [agor] | 28 |
| short vowels generally lengthened in monosyllables | ✶mbara > bâr | [mbara] > [mbar] > [bar] > [bār] | 29 |
| final [ɣ] became [a] after a consonant | ✶phelga > fela | [pʰelga] > [ɸelga] > [ɸelɣa] > [felɣa] > [felɣ] > [fela] | 30 |
| [ɣ] became [i] between sonants and vowels | ✶targasse > tarias | [targasse] > [tarɣasse] > [tarɣass] > [tariass] > [tarias] | 31 |
| [ɣ] otherwise vanished | ✶magē > mae | [magē] > [mage] > [maɣe] > [mae] | 32 |
| final [-wi] became [-y] [note] | See Eldamo’s discussion and examples. | 33 |
| [ls], [rs] became [lθ], [ss] | ᴹ✶olsa- > oltha- | [olsa-] > [olθa-] | 34 |
| [h] vanished after vowels | ✶asa > a | [asa] > [aha] > [ah] > [a] | 35 |
| final [i], [u] generally vanished | ✶certi > cirth | [kertī] > [kerti] > [kertʰi] > [kerθi] > [kirθi] > [kirθ] | 36 |
| short vowels vanished before morpheme boundaries [note] | ✶imbi·lata > imlad | [imbi·lata] > [imbi·lat] > [imbi·lad] > [imblad] > [imlad] | 37 |
| [ai], [oi] became [ae], [oe] | ✶magiti- > maed | [magiti] > [maɣiti] > [maɣit] > [maɣid] > [maid] > [maed] | 38 |
| later [ei] became [ai] in final syllables [note] | ✶mbarī > bair | [mbarī] > [mbari] > [bari] > [beri] > [beir] > [bair] | 39 |
| diphthongs [yi], [yu] became [ui] [note] | See Eldamo’s discussion and examples. | 40 |
| [œi] became [ui] or [y] | ✶oronyē > eryn | [oronjē] > [oronje] > [oronie] > [œrœnie] > [œrœni] > [œrœin] > [œryn] > [eryn] | 41 |
| [nr] became [ðr] | ✶caran·rass > Caradhras | [karanrass] > [karaðrass] > [karaðras] | 42 |
| dissimilation of dental spirants [note] | ✶thakta- > úthaes | [tʰakta] > [tʰaktʰa] > [θaktʰa] > [θaxθa] > [θaxθa] > [θaiθ] > [θaeθ] > [θaes] | 43 |
| final [mf], [nθ], [ŋx], [lθ] became [mp], [nt], [ŋk], [lt] | ✶gampa > gamp | [gampa] > [gampʰa] > [gamɸa] > [gamfa] > [gamf] > [gamp] > [gamp] | 44 |
| nasals vanished before spirantal clusters [note] | ✶camprū > cāfru | [kamprū] > [kampru] > [kampʰru] > [kamɸru] > [kamfru] > [kafru] | 45 |
| nasals vanished before morpheme boundaries [note] | Examine rule for more details: in·tîw > in·thîw > i·thîw | 46 |
| [ð] vanished before nasals at morpheme boundaries [note] | ✶Galada·ndil > Gelennil | [galada·ndil] > [galada·ndil] > [geleðe·ndil] > [geleðndil] > [gelendil] > [gelennil] | 47 |
| voiced spirants restopped after nasals [note] | No examples at this time: [-{mnŋ}{vðɣ}-] > [-{mnŋ}{bdg}-] | 48 |
| medial [mf], [nθ], [ŋx], [lθ] became [mm], [nn], [ŋg], [ll] [note] | ✶danta- > danna- | [danta-] > [dantʰa-] > [danθa-] > [danna-] | 49 |
| voiceless nasals were voiced | ✶snagdē > naeð | [snagdē] > [snaidē] > [snaide] > [n̥aide] > [n̥aiðe] > [n̥aið] > [n̥aeð] > [naeð] | 50 |
| long vowels shortened before clusters | ✶nēna·talma > nindalf | [nēna·talma] > [nīna·talma] > [nīna·talm] > [nīna·dalm] > [nīndalm] > [nindalm] > [nindalv] | 51 |
| [ī], [ū] often shortened in polysyllables [note] | ✶al·phirīne > alfirin | [alpʰirīne] > [alɸirīne] > [alfirīne] > [alfirīn] > [alfirin] | 52 |
| [awa] sometimes became [au] | ✶glawar > glaur | [glawar] > [glaur] | 53 |
| [au], [ae] became [o], [e] in polysyllables [note] | ✶aika·nār > Aegnor | [aika·nāro] > [aika·nǭro] > [aika·nauro] > [aika·naur] > [aiga·naur] > [aignaur] > [aegnaur] > [aegnor] | 54 |
| [lð] became [ll] | ✶kuldā > coll | [kuldā] > [kulda] > [kulða] > [kolða] > [kolð] > [koll] | 55 |
| [nl] became [ll] [note] | No examples at this time: See the rule for more info. | 56 |
| [mb], [nd] became [mm], [nn] | ✶ambuna > amon | [ambuna] > [ambona] > [ambon] > [ammon] > [amon] | 57 |
| middle consonants frequently vanished in clusters [note] | ✶Gondo·ndor > Gondor | [gondo·ndor] > [gondndor] > [gondor] | 58 |
| medial [s] became [θ] before [l], [r] | ✶cas·raya > cathrae | [kasraja] > [kasraia] > [kasrai] > [kasrae] > [kaθrae] | 59 |
| [wo] became [o] | ✶wo > gwo > go- | [wo-] > [gwo-] > [go-] | 60 |
| [n] assimilated to following labial [note] | ✶ndan·men > damen | [ndanmen] > [danmen] > [dammen] > [damen] | 61 |
| [œ] became [e] [note] | ✶okoli > ogyl | [okoli] > [okuli] > [œkyli] > [œgyli] > [œgyl] > [egyl] | 62 |
| final [ll], [nn], [ss] shortened in polysyllables | ✶caran·rass > Caradhras | [karanrass] > [karaðrass] > [karaðras] | 63 |
| final and initial [ŋg] became [ŋ] | ✶angā > ang | [aŋgā] > [aŋga] > [aŋg] > [aŋ] | 64 |
| non-initial [m] usually became [v] [note] | ✶amān > avon | [amān] > [amǭn] > [amaun] > [amon] > [avon] | 65 |
| [ðv] became [ðw] | ✶buzbō > buðu | [buzbō] > [buzbo] > [buðβo] > [buðvo] > [buðv] > [buðw] > [buðu] | 66 |
| [mm] shortened | ✶ndan·men > damen | [ndanmen] > [danmen] > [dammen] > [damen] | 67 |
| final [v] became [w] after [ae], [oe], and sometimes [i] [note] | ✶okma > oew | [okma] > [ogma] > [oɣma] > [oima] > [oim] > [oem] > [oev] > [oew] | 68 |
| final [w], [v] vanished after [u] | ✶dōmē > dû | [dōmē] > [dūme] > [dūme] > [dūm] > [dūv] > [dū] | 69 |
| [ou] became [au] | ✶lowā > law | [lowā] > [lowa] > [lou] > [law] | 70 |
| long voiceless spirants shortened | ✶rokkō > roch | [rokkō] > [rokko] > [rokkʰo] > [roxxo] > [roxx] > [rox] | 71 |
| final [l], [r] usually became syllabic [note] | ✶nathlo > nathal | [natʰlo] > [naθlo] > [naθl] > [naθol] | 72 |
| final [vn] sometimes became [von] | ✶kelepna > celevon | [kelepna] > [kelebna] > [kelemna] > [kelemn] > [kelevn] > [kelevon] | 73 |
| final [w] usually became [u] | ✶buzbō > buðu | [buzbō] > [buzbo] > [buðβo] > [buðvo] > [buðv] > [buðw] > [buðu] | 74 |
| final [rr] became [r] | ✶gę̄rra > gaer |[gę̄rra] > [gę̄rr] > [gaerr] > [gaer] | 75 |
| [sk], [sp] usually became [sg], [sb] [note] | ✶skalā > esgal [ṣkalā] > [ṣkala] > [eskala] > [eskal] > [esgal] | 76 |
| medial [x] became [h] in Gondorian pronunciation [note] | See the rule for more info. | 77 |
| voiced spirants unvoiced before voiceless spirants | See the rule for more info. | 78 |
These notes explain the scope and source of particular examples. A difference between a recorded word and a reconstructed stage may already be present in Eldamo’s source presentation. Such cases are identified here without changing the recorded word.