Mellonath Daeron - The Language Guild of the Forodrim

Attested Sindarin Plurals

with thanks to the guildmembers, especially Findegil, Erestor and Vidumavi,
and other scholars of Elvish,
whose observations and ideas the compiler has made use of.

Contents

Introduction

• Scope and Aim
• References
• Diacritic Characters

Attested Plural Forms

1. Plurals Formed by Vowel-affection
   
   1. Plurals with Known or Constructible Singulars
      
      1. Plurals of Monosyllables, sorted by vowel in singular:
         

         	E	A	O	U
         Í	É			Ú
         	EI	AI		UI
         		AU

      2. Plurals of Bisyllables, sorted by vowels in singular:
         

         	-I	-E	-A	-O	-UI
         I-		I-E	I-A	I-O	I-UI
         Í-		Í-E	Í-A
         EI-		EI-E
         E-		E-E	E-A	E-O
         AE-		AE-E
         A-		A-E	A-A	A-O	A-UI
         O-		O-E	O-A	O-O
         Ó-		Ó-E

      3. Plurals of Polysyllables
   2. Plurals with Uncertain Singulars
2. Plurals with -in

Appendix: Notes about Problematic Forms

• Note on -(i)on
• Note on Belain
• Note on Sarn Gebir

Introduction

Scope and Aim

The following table in intention lists every attested Sindarin nominal plural form. This means:

Attested form:

For plurals that are only attested as bases in compounds, or in mutated forms, unmutated simple forms are construed. Such a construction is marked by a preceding asterisk. Plurals attested without mutation in hyphenated forms (like my^l in Bar-in-My^l) are quoted without asterisks.

When a plural is attested as the base of a compound, like *tain attested in celerdain, the compound itself is usually also a plural form and included in the table (but there are cases like Lebennin, 'Five-waters', where the compound is a singular name, despite the plural base). In a couple of cases where the simple form of the base in a plural compound presents a problem, only the compound is listed: en-feng (is the simple plural *feng or *faing?), lech-hind ('flame-eyed': would an adjective 'eyed' exist in standard Sindarin?). Since the simple word pen survived only as a usually enclitic pronoun (QE B Sindarin 1:2), celbin &c. are not taken as attestations of a plural *pin (singular pronouns do not in general have morphological plurals: e.g. no pronoun **somebodies in standard English).

A plural like emyn that occurs in many combinations (Emyn Beraid, Emyn Muil, Emyn Arnen) is still only one attested form.

Sindarin:

In a preceding stage of J.R.R. Tolkien's development of this language it was named "Noldorin" and conceived as the hereditary language of the Noldor, different from the Ilkorin language spoken by the Elves of Doriath. The name "Sindarin" marks a new conception, in which this language is the hereditary language of the Elves of Doriath and their kindred, adopted by the Noldor in Beleriand.

When the main text of The Lord of the Rings was written the language was Noldorin. The new conception entered in the "Grey Annals" (see WJ 1 GA [text] "Excursus on the languages of Beleriand"), belonging to the 1950-51 period of work, and was then introduced into revisions of the Appendices to The Lord of the Rings before the publication of that work.

Although the changes in the language from the inception of work on The Lord of the Rings to its publication were not drastic they are discernible, and they appear to comprise differences in plural formation. The Noldorin described in the "Etymologies" is therefore here technically treated as a different language than Sindarin, but Noldorin forms are sometimes quoted in support of constructions. In the absence of a detailed analysis of when in the process the later patterns were established, the nominal change to "Sindarin" from "Noldorin" is here used to delimit the later stage of the language from the earlier, potentially discordant, material.

Grey-elven plurals in The Lord of the Rings, The Road Goes Ever On and "Guide to the Names in The Lord of the Rings" are therefore accepted by definition. The criterion 'attested in the "Grey Annals" or later' means that the other principal sources are found in Unfinished Tales, Morgoth's Ring, The War of the Jewels and The Peoples of Middle-earth. Christopher Tolkien in Sauron Defeated (1 XI Second:75-78) points to evidence that the second version of the Epilogue to The Lord of the Rings is later than the Waldman letter (L 131), ascribed to late 1951: the traditional inclusion of "The King's Letter" in the Sindarin corpus can thus be maintained on this criterion as well.

Since the text of The Silmarillion is chronologically heterogenous caution dictates that attestation always be sought elsewhere. Two plurals only known from Christopher Tolkien's philological appendix to The Silmarillion (cebir and morchaint) are included on the strength of being clearly connected to Sindarin words in The Lord of the Rings.

Even so, all forms appearing in this table do not belong to one finished and fully consistent system, as is evident for instance from the comment on echil (under "Plurals with Uncertain Singulars").

Nominal:

Adjectives and nouns appear to form plurals by the same patterns, and are not separated in the table.

Plural pronouns are not in general morphological plurals: English we is not a plural form of I, and there is no singular counterpart to few. The Sindarin plural personal pronouns ammen and hain are not included in the table, but hin / hîn and *pain that can be classed as pronouns found in adjectival use are listed among "Plurals with Uncertain Singulars", as they agree in form with nominal plural patterns.

Plural:

The term is to be understood in a narrow sense that does not include collectives ('group' or 'general' plural formations) like Crissaegrim, Glamhoth and Argonath, even though subordinate adjectives take, or may take, a plural form (as in Pinnath Gelin).

It is unclear which kind of plural meaning is expressed by the ending -(i)on that occurs in a handful of forms like Nanduhirion; this problem is discussed in a note at the end.

Mellonath Daeron will be grateful for being alerted to any form falling within these limits that may have been overlooked by the compiler.

The aim of the table and accompanying notes is to present the material and the problems of identification and construction. A discussion of what rules for plural formation can be inferred from the material is planned to follow in a separate document.

Since plural forms are the primary concern, the plural is given first when a lexeme is cited in both numbers (like "edain : adan"). The two main tables, however, are organized according to the vowels in the singular forms, and for each lexeme show the plural(s) in bold letters below the singular.

References and annotations here assume that the reader is well acquainted with Elvish philology and only needs short reminders. (For instance, the plural *sîr is said to be contained in Ossir: the reader is expected to remember why this may be so (or look up the reference and catch the hint), see the relevance of the further annotation "cf. Lebennin", and assess the credibility of the construction.)

References

One reference is given for each attested plural or singular form. In many cases the reference only specifies one of Tolkien's published titles. This means that the listed word, or a combination starting with the listed word, can be looked up in the index to that work (or in one of the indices, or in an appendix or supplement to the index, as the case may be).

The following work abbreviations are used. (The work titles are linked to descriptions in Åke Bertenstam's A Chronological Bibliography of the Writings of J.R.R. Tolkien.)

If the form is not indexed, the work abbreviation is supplemented by one or more of the following:

1. an indexed form containing the listed form, like "Emyn Duir" in the reference for duir;
2. a phrase like "In Index s.v. Naugrim" in the reference for naug, when a word is indexed under another word;
3. a phrase like "In Ivonwin" in the reference for *gwin, when a word is only attested in a compound or mutated form;
4. a structural reference to a locus in the text, like "App.E II Cirth:1" in the reference for cirth, when the index does not serve. (A separate page on structural references describes this reference format.)

Where unattested singulars are construed, references to forms supporting the constructions are given in footnotes.

A large number of plurals are found in the text "Quendi and Eldar" ("Essekenta Eldarinwa") in The War of the Jewels; these are marked 'QE' rather than 'WJ'. Where plural and singular are both referenced with only 'QE', it means that one or the other of the forms is found in the index to The War of the Jewels. If neither of them occurs in the index, a more specific reference is given only for the plural, since the two forms are in all such cases found close together.

When a form in the "Etymologies" in The Lost Road and Other Writings is quoted for comparison, the reference is written "Etym." plus the relevant root.

An Introduction to Elvish, mentioned at a couple of places, means: Jim Allan, ed., An Introduction to Elvish .. . . (Hayes, Middlesex: Bran's Head, 1978).

Diacritic Characters

Most accents and other diacritics can be encoded in HTML. Some, however, cannot. Among these are the macron ( ¯ ), the circumflex ( ^ ) when placed over a Y or y, and also the tilde ( ~ ) when placed over any character other than A, a, N, n, O or o. In accordance with Mellonath Daeron's standard (described on the page On diacritic characters on the guild's home page) the macron is here written as a vertical bar after the vowel it belongs to, and a circumflex or tilde belonging to an 'impossible' character is placed after this character. Examples: THO|N-, my^l.

Attested Plural Forms

I. PLURALS FORMED BY VOWEL-AFFECTION

A. PLURALS WITH KNOWN OR CONSTRUCTIBLE SINGULARS

1. For the connection between senses 'land' and 'house' (as in e.g. Bar-en-Danwedh) see Etym. MBAR- ('dwell, inhabit'), and cf. Q ambar, and also -mar in for instance Eldamar 'Elvenhome'.

2. Concerning gw- see next note. With S *gwen cf. N gwend, gwenn, Ilk. gwen (Etym. WEN-, WENED-).

3. Initial W is regularly turned into GW in simple and unmutated Sindarin words, and initial GW is turned into W in composition and lenition (for instance, when the second element in Elwing was taken as Sindarin its simple form was given as gwing; PM XII, note 24).

4. Cf. echant (LR 2 IV, ill. (to §§93-98)); for the connection see Etym. KAT-.

5. It is stated that el, plural il, only survived as a final element, wherefore the forms are here cited as -el and -il.

6. Nîn-in-Eilph is translated "Waterlands of the Swans", and nîn could conceivably be a plural of the adjective nîn 'wet' (GN Places, Wetwang) used as a noun. But cf. Lebennin.

7. The word is cited as tal in S App., but possibly the correct form is *tâl (as it was in Noldorin, see Etym. TAL-).

8. In levain tad-dail the absence of mutation shows that it is a (qualitative) genitive construction, 'animals of two feet'. Its singular would be *lavan tad-dail. Hence, though the base of tad-dail is a plural noun it is not itself a plural of (an adjective) **tad-dal 'two-foot (e.g. animal)', but an indeclinable compound form.

9. Cf. Etym. TAN-. The vowel in the simple form may possibly be â (sc. *tân).

10. Concerning *t- see the preceding note.

11. Cf. Lebennin.

12. This in itself probable form is attested in a somewhat roundabout way. Pelargir is glossed 'Garth of Royal Ships' in the Index to The Silmarillion, attesting a plural *argir 'royal ships'. The corresponding singular would be a compound ar- (cf. Argonath) + cair; hence *-cir in Pelargir is an occurrence of cair in a plural form. The quantity of the vowel in *cîr is an assumption.
It is, however, possible that the singular of *argir is *argir, with contracted diphthong: cf. the contraction of cair in Círdan. If so, *-cir in Pelargir need not tell anything about the plural of cair. (Círdan can not, pace Welden in An Introduction to Elvish, be presumed to contain a plural of cair; see the discussion of compounds in "Note on -(i)on" below. Moreover, the Noldorin name Cirdan appears in "Etymologies" under TAN- as "C(e)irdan shipbuilder", clearly showing a contraction of N ceir 'ship', singular (Etym. KIR-).)

13. Cf. Dor. muil (Etym. MUY-).

14. The identification of this muil as a plural adjective rather than a noun (that could be singular) relies on Celeborn's presentation of the locale as "the bleak hills of Emyn Muil" (LR 2 VIII:49) and the English equivalents Graydon Hills, Grailaws and Hazowland (TI XXIII:17): thus muil = 'grey, bleak'. But note that there is no authorial translation. The Doriathrin muil (see previous note) is a noun, 'twilight, shadow, vagueness'. Further, if muil is an adjective, but not if it is a noun, lenition would be regular, sc. **Emyn Vuil (cf. Eryn Vorn (UT)). However, Tolkien once stated: ". . . ; but grammatically before actual stems of verbs, the soft mutation only was normally used in later S., to avoid confusion with other verb stems; and the soft mut. of m > v~ > v was also often not used for the same reason" (L 347:8). NOTA BENE: This is the reading of the manuscript letter; in the printed text a not is lost (as attentive readers can guess), and the second semicolon is turned into a comma. The clause about M > V is thus separate from that about verb forms. Cormallen and Ered Mithrin provide examples of unmutated adjective attributes.

15. Although Nibin-noeg was replaced by Noegyth nibin, noeg supposedly remains the correct plural of naug.

1. The word is presumably cited in singular form, as it is being compared to the singular fanui.

2. In Mithrellas (UT)? Cf. mith (GN Places, Hoarwell) and angren (UT), silivren (LR 2 I:169).

3. Cf. mîr (LR App.E I Consonants:2, footnote) and *tain : *tan.

4. See also "Note on Sarn Gebir" below.

5. Cf. rem (ibidem) and abonnen (QE).

6. Cf. Annúminas (LR) and lithui (see note 1).

7. Cf. N dangen (Etym. NDAK-).

8. Cf. gwath (UT 2 IV App.D:3) and angren (UT), silivren (LR 2 I:169).

9. Cf. *caint : *cant.

10. Cf. Rhovanion (LR Map of Middle-earth).

11. Cf. ending -on in person nouns like mellon (LR).

12. See also "Note on Belain" below.

13. From Etym. LAM-, the 'Nolodorinized Doriathrin' names Dor-lómen and Ered Lómin are seen to contain an adjective lómen 'echoing'. In earlier texts, the land was named Dor-lómin (then meaning 'Land of Shadows'), and later the form Dor-lómen in general again gave way to Dor-lómin. Thus, although the form Dor-Lómen in WJ is used on the map of Beleriand and occurs once in a text, this attestation of lómen in post-LR sources is dubious.
In Eryd Lómin, lómin can be a plural adjective or a noun. In Dor-lómin, lómin can be a noun or a singular adjective. The shifting of forms makes the interpretation uncertain. But it seems easiest to assume that Eryd Lómin, always called the 'Echoing Mountains', contains an adjective with singular lómen. Presumably, a singular lómin would also have the plural lómin, but -en is a common adjective ending, and lómen occurs in Doriathrin, as noted above.
For Dor-lómin, see under "Plurals with -in" below.

B. PLURALS WITH UNCERTAIN SINGULARS

1. See the footnote on *cîr above.

2. As noted in WJ Index, the existence of echil seems denied by the statement in QE App.A:5 that "Hildor, since the stem *KHILI 'follow' was not current in Sindarin, was rendered by Aphadon, pl. Ephedyn". But this does not touch the standing of echil as a grammatical plural formation.

3. 2 IV, ill. (to §§93-98). In the transcription at the bottom of the illustration the last word is written hin, but the writing in tengwar seems to say hîn.

4. UT Index cites the translation 'Pools of the golden water-flowers' for Loeg Ningloron. Since UT 2 IV App.D:7 gives lô < loga as a word for a fenland, with an editorial insertion saying that the stem is log-, it seems likely that loeg is the plural of a word from a variant of the same stem (*logok-?), or an archaic plural of lô.

5. The name Ras Mewrim (WJ), an alternative to Bar-in-My^l, suggests that the etymological root vowel was EU or IU (the Noldorin precursor to *mew is maew, found under Etym. MIW-). Sindarin Y is said to be often derived from EU and IU (LR App.E I Vowels:3): the example that is cited, the Sindarin equivalent to Q leuca, is given as ly^g in the first edition, and the later loss of the circumflex may be inadvertent.

II. PLURALS WITH -in

In An Introduction to Elvish, Welden in his benchmark discussion of Sindarin plurals cites Hithaiglin as a plural with -in. But in the Index to The Silmarillion Christopher Tolkien admitted that this form (found on the version of "Map of Middle-earth" that he prepared for the original edition of The Lord of the Rings) was an error for Hithaeglir. (Probably *aeglir is a Sindarin version of Noldorin oeglir 'range of mountain peaks' (Etym. AYAK-), thus a singular.)

There are now four words in this group:

1. Dor-lómin; Christopher Tolkien translates this 'Land of Echoes' (LRW App. II, Eredlúmin). In terms from "Etymologies", lómin can be the plural of Doriathrin lóm 'echo'. The easiest assumption seems to be that this analysis is valid also for the post-LR use of the name (perhaps with 'Doriathrin' exchanged for 'North Sindarin').

Appendix: Notes about Problematic Forms

NOTE ON -(i)on

In half a dozen names the ending -(i)on has or may have a plural meaning:

Caras Galadhon

'City of the Trees' (LR)

Dorthonion (LR)

'Land of Pines' (glossed so in the Index to Unfinished Tales: if the gloss was taken from the pre-Sindarin "Etymologies" (entry THO|N-), Fangorn's significant mention of "the pine-trees upon the highlands of Dorthonion" supports it)

Dorwinion

(which falls within the domain of Sindarin by being among the extra names entered, supposedly on Tolkien's direction, on Pauline Baynes' painted Map of Middle-earth; cf. UT 2 IV App.D:3, footnote)

Loeg Ningloron

'Pools of the golden water-flowers' (UT)

Nanduhirion (LR)

= Nan Duhirion "the valley of the dim (overshadowed) rills" (GN Places, Dimrill Dale)

Orgilion

'day of the Stars' (LR App.D Calendars:17)

It is useful to observe here the difference between a compound like Edhellond and one like Arthedain. Like in English, a genitive attribute prefixed to the main noun is indeclinable: Edhellond, like Elfhaven, could mean 'haven of the elf' as well as 'haven of the elves'. A form like Arthedain, on the other hand, is in effect an ordinary genitive phrase run together (in several cases, like Nanduhirion, a non-compound form is attested beside the compound one), with the attribute declined by number. Hence, using Orgilion as an example, if gilion is not a plural form it must be a singular with a meaning (like 'star-multitude' or 'star-canopy') that can be paraphrased by "the Stars".

In An Introduction to Elvish -ion is taken as an archaic genitive plural ending, the same as Quenya -(r/i)on. This would be semantically possible in all the cases (including those with -on), but the interpretation seems refuted by the statements that in Telerin the genitive -o (< o| ) "did not receive -n addition in the plural" (QE A *HO Telerin), and that Sindarin had perhaps never even had -o| (QE A *HO Sindarin).

Possibly plural -(i)on is a quenyaism: this is especially likely in the case of Orgilion, considering that Orgaladhad and Orbelain are known to be quenyaisms (cf. "Note on Belain" below). On the other hand, Caras Galadhon is said to be "probably of Silvan origin, adapted to Sindarin" (LR App.F I Of the Elves:1, footnote), which hardly allows a quenyaism, whether -on is ascribed to the origin or to the adaption. (Dorthonion may also show a Silvan influence, considering that in "Etymologies" it was an Ilkorin, not a Noldorin, name.)

It is also possible that -(i)on is yet another collective ending, derivable from the same origin as the augmentative -(i)on in aearon and Tirion (see R O! Elbereth:9-10). If so, gilion is a synonym to giliath.

NOTE ON Belain

In his glossary of Elvish name elements that forms an appendix to The Silmarillion, Christopher Tolkien cites "Sindarin Balan, plural Belain" (s.v. val-). Since the plural form in Noldorin was Belein, Belen (Etym. BAL-), Belain is presumably taken from some later source (unless Christopher Tolkien has supplied a deduction from analogy, which would be most surprising). Although the form does not occur in any of the published texts it is phonologically sound. The reason to nonetheless hesitate, until stronger attestation surfaces, to accept Balan and Belain as Sindarin is a combination of facts:

1. Inias Valannor, Noldorin for 'The Annals of Valinor', occurred on all the three variant titlepages of the QS typescript (1937/38); on one of them it was later changed to Inias Balannor. Since the absence of mutation in genitive constructions seems to be a rule established during the writing of The Lord of the Rings, the correction probably belongs with the revisions made to this typescript after the completion of The Lord of the Rings, sc. in the Sindarin stage of the language. The amanuensis typescript LQ 1 made from the revised QS manuscript and typescript in 1951/52 has now no titlepage, but Christopher Tolkien surmises that one existed and that the titlepage of the later (1958) typescript LQ 2 was copied from it: it has Inias Valannor, but now this Noldorin phrase was emended not to Inias Balannor but to Ínias Dor-Rodyn. There is no other occurrence of Balan in the sources for Sindarin.

2. In The Lord of the Rings the Sindarin name of the weekday dedicated to the Valar is given as "Orbelain (or Rodyn)" (App.D Calendars:17). Tolkien commented on this name in a letter: "Orbelain is certainly a case of 'phonological' translation (of which the Noldor were quite capable), since Valanya (adj.) must be from older *Balania| which would > S. *Belain, but no such form existed in S." (L 347:8).

3. Is this a denial of the occurrence of any Sindarin form Belain, or merely of a Sindarin adjective Belain corresponding to Valanya?

   Why was the comment on Orbelain made at all? It is the end of a paragraph explaining that ath is not a dual element. Tolkien says that Sindarin duals "early became obsolete . . . [a] case occurs in Orgaladhad [dh written with an edh] 'Day of the Two Trees', but since these S. nouns were all derived from Quenya names of the 6-day week . . . it may be due to an attempt to imitate Q. duals, such as ciriat 2 ships". Although there is one sentence intervening between that and the one about Orbelain, the latter is clearly added as a support for the former. The thought is this: '-galadhad may be a construction imitating Quenya, rather than a current Sindarin form; that such constructions were made is shown by -belain since no such form existed in Sindarin'. If Orbelain could be analyzed as a compound of aur with an existing plural noun belain, the point would be lost.

   But the sentence about Orbelain is mystifying in another respect. It seems to say that since Belain is indeed the phonological counterpart of Valanya, Orbelain must be a case of phonological translation. But the forms attached to Or- in the other weekday names are apparently not actual or construed counterparts of the nominalized adjectives Elenya, Anarya, &c. (we do not have **Orelenui, **Orenair or suchlike) but counterparts of the corresponding nouns (eleni, anar, &c). Hence, if Belain is a construction, one would expect it to be a construed counterpart of Valar, not of Valanya. From the formal equivalence of Valanya and Belain it follows that Orbelain is a possible case of phonological translation, but if so, it is not formed analogically with the other weekday names. Unless Tolkien was confused, the full thought behind the sentence must be like this: 'since the second element in Orbelain is not a genuinely existing form in Sindarin, but would be the correct phonological translation of Valanya, and no less exceptional derivation can be found, a phonological translation is certainly what it is'. This implies that Belain could not possibly be an etymologically construed counterpart of Valar, perhaps because there could be no reason to make an N-extension of the root in the counterpart.

   If Belain is a counterpart of the adjective Valanya, it would regularly mutate in attributive position (sc. **Orvelain). However, as noted above (in the footnote on muil in the table) the mutation M > V is "often not used"; the name Dol Baran (LR) shows that sometimes B > V is not used either.

All this considered, the status of Balan and Belain as actual Sindarin forms seems dubious; further textual evidence is desirable.

As a related observation, it is intriguing that the Sindarin name for the sixth Elvish weekday is given as "Orbelain (or Rodyn)". The second form might indicate a purer Sindarin alternative to the quenyaism Orbelain, and that Balannor was exchanged for Dor-Rodyn perhaps supports this surmise.

On the other hand, that there are two Sindarin names for 'Friday' may simply reflect there being two Quenya names for this day, "Valanya (or Tárion)" (App.D Calendars:17).

In any case, while the Quenya names are nominalized attributes (adjective and genitive) of an implied Ré 'Day', Sindarin has weekday names with an explicit Or-. Rodyn is apparently just a plural noun 'Powers' (and has been treated so above). An earlier manuscript for this passage gives the Noldorin names for the last weekday as "Arvelain (or Ardórin)" (PM IV The Calendar:17); it is not unlikely that what Tolkien intended when he listed the Sindarin names was "Orbelain (or -Rodyn)"; that the alternative to Orbelain is Aur-Rodyn (or possibly Orrodyn).

NOTE ON Sarn Gebir

Tolkien wrote of the name Sarn Ford: "The name is a half-translation (of Sarn-athrad 'stony ford'), . . . The Elvish Sarn is also seen in Sarn Gebir" (GN Places, Sarn Ford). Unfortunately, he did not give a fuller analysis of the latter name.

There was a Sarn Athrad in Beleriand too, and in his appendix to The Silmarillion Christopher Tolkien glossed the first member of the name thus: "sarn '(small) stone' in Sarn Athrad . . . ; also in Sarn Gebir ('stone-spikes': ceber, plural cebir 'stakes'), rapids in the river Anduin." Since the pair cebir : ceber is included in the table on the strength of it, this gloss and the somewhat puzzling name Sarn Gebir require a closer examination.

The Noldorin word sarn is explained as "stone as a material, or as adj.; cf. Sarnathrad" (Etym. SAR-). In Sindarin it is, beside Sarn Athrad and Sarn Gebir, found in Edhelharn (SD 1 XI Second:67), the equivalent of Q Elessar. While in the latter word the gloss '(small) stone' is applicable, in Sarn Athrad it makes, pace Christopher Tolkien, bad sense. The authorial gloss "Sarn-athrad 'stony ford'" apparently instead depends on a sense of sarn like (either one of) those in Noldorin.

In Sindarin the "adjectival position [is] . . . after the primary noun" (R O! Elbereth:18). Of two nouns in sequence, the second is the modifier: a genitive attribute, or more rarely an epithet. If sarn and a word X are nouns, sarn X is expected to mean 'the stone of X' (or possibly 'the stone (titled) X'). A number of discarded names for the eventual Argonath, like Sarn Aran, Sarn Torath, Sern Aranath, Sairn Ubed (WR 2 I, note 5; IV:7), are indeed interpretable on the genitival pattern: 'the stone of the king', 'the stone of the kings', 'the stones of the kings', 'the stones of denial'. (If one of the latter two had been kept, a plural sern or sairn would have been attested; cf. nern : narn above). On the other hand, in Sarn Athrad, judging from Tolkien's gloss, athrad 'ford' is the primary noun and sarn the modifier. Is Sarn Gebir analogous to Sarn Torath or to Sarn Athrad?

It emerges from The Treason of Isengard and The War of the Ring that the application of the name Sarn Gebir shifted during the writing of The Lord of the Rings; its proper analysis may have shifted as well. It was first introduced as the name (in variants Sarn Gebir, Sarn-gebir, Sern Gebir and Sern-gebir) of a range of hills, perhaps as a translation of Stone Hills, that it apparently substituted, while the eventual Sarn Gebir was named Pensarn. In one note Sarn-gebir is equated with "Grailaw or Graidon Hills"; the hills were later given the final name Emyn Muil. Pensarn was succeeded by Sarn Ruin, until Sarn Gebir reappeared in a new role as the name for the rapids.

A group of "disconnected jottings, all struck out" in the K section of "Etymologies" is concerned with "N ceber pl. cebir and Sern Gebir". The root is given as KEPER- 'knob, head, top', which would accord with 'Stone Hills' as the sense of Sarn Gebir. At the same time, Christopher Tolkien explains that the meaning of Sern Gebir in these notes "seems to be 'lone stones'", thus with cebir as an adjective 'lone' modifying the plural noun sern. In any case, the gloss 'stakes' is then presumably taken from a later source, making it correspondingly likely to be the interpretation intended for the name in its final application.

If the second member of Sarn Gebir, then, is cebir, it suffers lenition in that name, which would be ungrammatical in a genitival phrase like Sarn Torath (see L 347:8). This supports Christopher Tolkien's gloss in The Silmarillion in its making Sarn Gebir analogous to Sarn Athrad in the relation between its members.

The remaining question is then why the modifier sarn in these two names is not placed in "the adjectival position". Such an inversion is quite exceptional (galadhremmin ennorath with an adjective preceding its governing noun is poetical). In Sarn Gebir, sarn is singular despite the plural cebir, and thus apparently a noun 'stone (as a material)' rather than 'stone' as an adjective. The name signifies then 'stakes consisting of stone', and Sarn Athrad is 'a ford consisting of stone'. The relationship between the members is a kind of qualitative genitive denoting the material something consists of, and it may be that in such genitives the mass noun is (or may be) placed before the primary noun. In other words, it may be that the semantic difference between a genitive like value of stone and one like statue of stone is grammaticalized in Sindarin.

This document has been online since March 13 1998. Last updated April 8 2000 by Elros, Måns Björkman.