Imparisyllabic nouns in Sindarin

Four Rings
Bertrand Bellet — November 2004
Theoretical ArticleTheoretical Articles: A comprehensive knowledge of the works of J.R.R. Tolkien is needed to fully understand articles in this category, the subjects treated being studied in minute detail by their authors.
Abbreviations and symbols:

  • CE Common Eldarin
  • N Noldorin
  • OF Old French
  • ON Old Noldorin
  • PQ Primitive Quendian
  • Q Quenya (or Qenya)
  • S Sindarin
  • App. Appendix
  • ch. chapter
  • sg. singular
  • pl. plural
  • * reconstructed word
  • † archaic or poetic word
  • # deduced form of a word only attested inflected or compounded

While most of the Sindarin (and earlier Noldorin) nouns and adjectives make their plural only with vowel changes, there is a small class of nouns that also add a syllable. Moreover, a handful of them have two forms in the singular: a shorter and a longer on which the plural is “regularly” built by vowel changes. They may be labelled imparisyllabic nouns. Here is a list of affected words.

  • N/S âr “king” beside aran, pl. erain1) ; the reduced form appears in S in many names, for instance the kings of Arthedain Arveleg, Argeleb,Arvegil, Arvedui
  • N bór “steadfast, trusty man, faithful vassal”, pl. býr for older berein, beren2)
  • S †êl “star”, pl. elin, class plural elenath “the starry host”3)
  • N †fêr “beech-tree”, pl. ferin4)
  • S #hêl “ice” in Forochel beside N heleg5)
  • N nêl “tooth” beside neleg, pl. nelig6)
  • N oel “lake, pool, mere”, pl. oelin7) ; the word reappears in S as aelin apparently both for sg. and pl., cf. the place-name Aelin Uial “Meres of Twilight”8)
  • N ôl “dream”, pl. elei9)
  • N/S #ôr “mountain” in Erebor “Lonely Mountain” or Orthanc “Mount Fang” beside orod, pl. N ereid, ered, S eryd or ered10)
  • N pel “fenced field”, pl. peli11)
  • N †thêl “sister”, pl. thelei12)
  • N thôr “eagle” beside thoron, pl. therein13); in S thoron is found, with its class plural thoronath, and #thôr appears to be included in Belecthor “Great Eagle”, the name of a steward of Gondor14)
  • N †tôr “brother”, pl. terein15)

How the difference between the singular and the plural arose is quite clear. Most of the Q cognates of these words end in a final consonant in the singular: thus âr, êl, fêr, nêl, oel,ôl, pel, thêl, thôr, tôr are related to the Q words aran, elen, feren, nelet, ailin, olor,peler, seler, soron, toron respectively. The plurals are in -i, going back to CE . Now Old Noldorin in the Etymologies shows a clear tendency towards the loss of final consonants, and it must have been so in archaic Sindarin too. But these consonants were not lost in the plural since they were shielded by the plural ending. The difference was preserved in the later language after i-metaphony in the plural and loss of final vowels (and loss of h in the case of peli and thelei). Here are a few examples from the Old Noldorin of the Etymologies.

  • ON oro pl. oroti “mountain” (N ôr pl. ereid, ered16))
  • ON skhapa pl. skhapati “shore” (N habad pl. hebaid17))
  • ON nele pl. neleki “tooth” (N nêl, neleg pl. nelig18))
  • ON pele pl. pelesi, later pelehi “fenced field” (N pel pl. peli19), where peleki appears for pelehi, evidently an error)
  • ON thele pl. thelesi, later thelehi “sister” (N thêl pl. thelei20))

This does not account however for the existence of the longer singular forms such as neleg and habad (the latter seems to have entirely ousted the shorter counterpart which is not recorded; it would presumably have been *hâb).

Some assume that these forms were formed back from the plural. However, they can also be the product of regular sound change. Old Noldorin must have been a case language like Quenya: we have for instance an example of a genitive in -en, identical to the Qenya genitive of that time, in the ON thoronen, Q sornen21). Archaic Sindarin was certainly similar: in XI:370, Tolkien argues that Sindarin had probably developed an inflexional genitive in (related of course to the Quenya genitive in -o) in “the primitive period”. Following Tolkien’s later conceptions, we can reconstruct the CE situation like this:

Meaning Nominative singular Genitive singular Nominative plural
king *aran *aranô *aranî
shore *skjapat *skjapatô *skjapatî
tooth *nelek *nelekô *nelekî
eagle *thoron *thoronô *thoronî

Later in an archaic stage of Sindarin, before the loss of final consonants, this may have resulted in the following:

Meaning Nominative singular Genitive singular Nominative plural
king *ara *arano *arani
shore *skapa *skapato *skapati
tooth *nele *neleko *neleki
eagle *thoro *thorono *thoroni

And in S this would give (with unrecorded forms not in italics):

Meaning Old nominative singular Old genitive singular Old nominative plural
king âr aran erain
shore hâb habad hebaid
tooth nêl neleg nelig
eagle thôr thoron theryn

But the very notion of declension appears to be lost at this period, as can be seen from S texts; for all nouns ending in a vowel (the overwhelming majority) case forms must have fallen together with the loss of final syllables. Thus âr and aran, nêl and neleg, thôr and thoron etc. were no longer considered as different case forms but simply as variants. By this theory all CE nouns ending in a consonant would have developed shorter and longer singulars. As can be observed from êl, fêr, ôl etc. on the one hand, from habad on the other hand, it seems that one sometimes ousted the other. The reason to choose the one or the other is not evident, usage shows its arbitrariness in that matter. Perhaps it was determined to some extent by the liability to homonymic clashes, but our Sindarin vocabulary is not complete enough to let us know. There is a hint that Tolkien imagined things so, at least in Noldorin. The entry THOR, THORON in the Etymologies22) has:

…. Q soron (and sorne), pl. sorni eagle; N thôr and thoron, pl. therein - thoron is properly old gen. sg. = ON thoronen, Q sornen, appearing in names as Cil-thoron, orCil-thorondor ….

Finally, some double singulars have alternative explanations. In V:379 Tolkien both gives the ON forms oro and oroto, which would regularly yield ôr and orod respectively. Perhaps this goes back to a hesitation in the length of the final consonant in Primitive Quendian: *oroto and *orotô would regularly give the form attested later. Q only retains orto “mountain-peak” < *orotô. Since Q also has helke for “ice”, it may be that the variants #hêl and heleg come back to a primitive variation between PQ *kheleke and *khelekê.

There is an interesting parallel in the evolution of French. Old French (9th to 13thcentury) had a small declension inherited from Vulgar Latin, with two cases: subjective and objective. Subjective was derived from the Latin nominative and took over its functions together with the one of the vocative: it was thus used for the subject and its attributes, and for apostrophe. Objective came from the Latin accusative (blended in some instances with ablative and dative) and was used for all other functions. Different forms of the articles and the presence or absence of a final -s or -z chiefly marked the cases. In addition, for some nouns inherited from the Latin imparisyllabic third declension, or later modelled on it, the stem was modified too, reflecting the Latin opposition between a direct stem in the nominative singular (also the accusative singular for neuters) and an oblique stem for other cases, often with stress shift. The difference is illustrated by the declension of “the wall” and “the thief”, derived respectively from the Latin mûrus, i, m and latrô, latrônis, m.

Meaning “the wall” “the thief”
Subjective singular li murs li lerre
Objective singular le mur le larron
Subjective plural li mur li larron
Objective plural les murs les larrons

The declension disappeared in Middle French (14th to 16th century), partly because of phonetic change: most final consonants began to be dropped, notably the morphologically important s. Normally, the modern French forms come from the objective. But several words with two bases had both preserved as doublets till now, most often with a semantic difference. Examples:

  • gars “lad” / garçon “boy” from OF garz / garçon
  • copain “pal” / compagnon “companion” from OF compain / compaignon
  • pâtre “shepherd” / pasteur “shepherd, pastor, clergyman” from OF pastre / pastor
  • sire / seigneur “lord” from OF sire / seignor, the first chiefly used in apostrophe
  • pute / putain “whore”, same in OF

In a small number of cases it was the subjective only that survived; it concerns names of persons, and it is often explained that the subjective was preserved because they are often used in apostrophes. It includes many Christian names (Charles, Georges, Jules, Guy…) and a few words like fils “son” (OF fiz / fil), sœur “sister” (OF suer / seror), ancêtre “ancestor” (OF ancestre / ancessor), prêtre “priest” (OF prestre / provoire), traître “traitor” (OF traïtre / traïtor). For the latter the subjective was probably preferred because the objective - borrowed and preserved in English - would have become *traiteur and clashed with a homonym meaning “caterer” today.

In Welsh, which is well known to be Tolkien’s chief inspiration for Sindarin, there are some traces of such processes. A few doublets can be noted as in French; Morris-Jones gives the following:

  • tymp “time” vs. tymor “season” reflecting the direct and oblique stems of the Latin tempus, temporis, n “time” (p.86)
  • ciwed “rabble” vs. ciwdod “people” reflecting the direct and oblique stems of the Latin cîvitâs, cîvitâtis, f “city, citizenship” (p.194)

But most importantly, the Welsh plural suffixes -(i)au, -(i)on, -ydd, -edd, -oedd, -i etc. come from reanalysis of the relationship between the products of direct and oblique stems. To quote Morris-Jones once again (§119):

…. The W pl. terminations are the Brit. stem-endings of imparisyllabic nouns, which were lost in the sg. representing the old nom. sg., but survived in the pl. after the loss of the pl. endings *-es, neut. *-a, §113.i. Thus latrô and its Brit. pl. *latrones gave W lleidr, pl. lladron by regular sound change; then the -on of the latter and similar nouns naturally came to be regarded as a pl. ending, and was added to nouns of other declensions where a pl. sign was needed, as to meddyg, see above. Such additions were made on some analogy, mostly of meaning, sometimes of form. ….

It is possible the same phenomenon occurred at a small extent in Sindarin: this may explain the dimly attested -in plurals, though the evidence can be analysed in more than one way; see Helge Fauskanger’s article Sindarin - The Noble Tongue, section “Plurals in -in”. But that it is evidently quite marginal: in contrast with Welsh, Sindarin favours vowel changes instead of endings to mark the plural.

In that respect, it is interesting that a rather high proportion of imparisyllabic nouns are marked as archaic or poetic, or often replaced by more common terms. Even if our corpus is far too limited to draw firm conclusions, the longer forms aran, orod, thoron appear more frequently than âr, #ôr, thôr. We are informed that êl, fêr, thêl and tôr were usually replaced by gîl / gil, brethil, muinthel, muindor. This might show that imparisyllabic nouns tended to be eliminated from the language, being more and more considered as irregular in contrast to the overwhelmingly more numerous parisyllabic nouns. On the other hand, shorter forms appear less seldom in names, perhaps because they allowed to avoid too long compounds.


Tolkien’s works are indexed following the system used in Vinyar Tengwar or Tengwestië: see the conventions. I have made an exception for The Lord of the Rings. Since its editions are numerous, I have preferred to give the book and the chapter, or the reference of the appendix.

Other works

  • Joly Geneviève, Précis d’ancien français, Armand Colin, 2002
  • Morris-Jones John, A Welsh Grammar, Historical and Comparative, fourth edition, Oxford University Press, 1953

Web resources

See also

On Tolkiendil

On the net

1) LR/II ch.4; LR/IV ch.7; LR App.A/I/ii; S:356; V:360; IX:129-31; L:426 nº 347
2) V:353
3) LR/II ch.1; X:373; XI:363; R:73-75; L:281 nº 211
4) V:352,381
5) LR App. A/I/iii and Map; V:364
6) V:376; VIII:113
7) V:349
8) S:355; V:349
9) V:370,379
10) LR/III ch.8; S:362; UT:435; V:379
11) , 19) V:380
12) , 13) , 20) , 21) , 22) V:392
14) S:322,365; LR App. A/I/ii
15) V:394
16) V:379
17) V:386
18) V:376
langues/english/i-lam_arth/imparisyllabic_nouns_sindarin.txt · Dernière modification: 06/04/2020 18:47 (modification externe)
Nous rejoindre sur
Tolkiendil - - Tous droits réservés © 1996-2024