Early Noldorin Plural Patterns


It is well known that the celtic-themed of Tolkien's Elvish languages (called Goldogrin at first, later becoming Noldorin and Sindarin) underwent quite a few changes in vocabulary and grammar from early to late times. The 'Early Noldorin Fragments' edited by Christopher Gilson, Bill Welden, Carl F. Hostetter and Patrick Wynne (published in Parma Eldalamberon (PE) 13) shed some light on Tolkien's ideas when the language ceased to be known as Goldogrin. In all, we have a small grammar (ENG), a wordlist arranged in alphabetical order (ENW) and a lexicon (ENL) where the entries are rearranged according to etymology and some forms have been emended.

Arguably the most interesting feature of the grammar is the plural formation - Early Noldorin was written down by Tolkien during the conceptional transition from marking plurals by endings as in Goldogrin [1] to the marking of plurals by vowel mutation familiar from the Noldorin of the Etymologies [2] or later Sindarin [3] as the dominant mechanism. Consequently, Early Noldorin is characterized by a variety of different plural patterns among which we see some of the later patterns emerging.

Clearly working out the proper plural of a given word was important for Tolkien - unlike in later sources, about 2/3 of nouns and adjectives have their plurals listed. Thus, all in all we have more than 250 attested examples of Early Noldorin plural formation (which is a fairly large number by the standards of Tolkien-linguistics).

General observations

It is not uncommon for a noun to have more than one possible pluralization, cf. crunc pl. crync, cryngir 'crow' (PE13:141). In some cases we can demonstrate conceptual development from the earlier ENW to the later ENL as for example in the case of aiw pl. aiwyn 'bird' (PE13:136) in ENW but aiw pl. aiwiath, aiwin (PE13:158) in ENL in which the phonology is changed and a new alternative plural with an ending is introduced. It is plausible that in the case where more than one plural form is acceptable we are dealing with historical development and analogical replacement. This can be studied in some cases:

The entry hont pl. hynt recent hontath 'trumpet' (PE13:163) shows the pluralization by means of the ending -ath is denoted as the recent development. Likewise interesting is arog pl. erig, older arogin 'swift' (PE13:137) from the ENW and arog pl. arogin later eryg 'swift' (PE13:160) from ENL. Here we see an older plural pattern using the ending -in and a more recent form eryg/erig found by vowel mutation. On the other hand peth pl. petha, †pith 'word' (PE13:164) suggests that the older pattern †pith was not favoured later - instead petha involving an ending seems to be the regular form. Thus, one cannot make a general statement that endings or vowel affection would be the earlier pattern and the other mechanism would be subsequent analogical replacement - rather both mechanisms must have been present over a longer timespan in the fictional development of the language (although Tolkien ultimately seems to derive vowel mutation from the presence of an ending, cf. the discussion in PE13:121).

In some cases we find collective plurals in addition to the regular plurals, e.g. gwenn pl. gwinn coll. gwenlhai 'light-elf' (PE13:146) or golodh pl. gelydh coll. golodhrim 'gnome' (PE13:145). For one word, a separate dual form is given: hen(n) pl. hinn dual him 'eye' (PE13:122).

Primary evidence

In the following we will present the attested examples of plural formation in a way that is most suitable to discuss vowel affection, i.e. we will sort into monosyllables and polysyllables and present the plural patterns for a given vowel or vowel combination.

Monosyllables A

The majority of monosyllabic words with a forms plural by vowel affection a, â, á > ai.

When the a is followed by a 2-consonant cluster involving r the pattern of vowel affection changes to a > e, ei (but note the optional arn > airn above) - in may cases the detailed choice is indicated as optional:

A fair share of plural patterns involves the endings -ir, -ath, -ain, -in, -lith, -rim and -a:

However, in the majority of cases the endings -ir, -in, -yn (all involving i) cause internal i-affection:

Monosyllables E

The most common pattern in monosyllabic words with e is the change e > i in plural:

The only example with a long ê exhibits ê > ai in plural:

An ending -as (in the later ENL -a) can be found in

The words for 'man', 'people' do not fit into any typical pattern, presumably they are irregular:

Monosyllables I

All monosyllabic words with i, í form their plural by means of an ending. There is a wide variety of possible endings, we see -(i)ath, -in, -i, -ir and -iant:

Monosyllables O

The most common pattern for monosyllabic words with o is vowel mutation o > y:

Occasionally, vowel mutation o, ó > ui, uai is observed:

Endings are also frequently seen, in the case of the endings -in, -ir i-affection o >e and ó > ui is observed:

Monosyllables U

Monosyllabic words with u usually form plurals by vowel mutation u > y:

In one case, u > ui is seen instead. In addition, crunc forms an alternative plural by means of an ending: Long û,ú never leads to vowel mutation - in these cases the endings -ir, -ath and -ain are seen:

Monosyllables AI

Usually words with ai pluralize with various endings. In rare cases, endings involving i lead to i-affection ai > ei.

In the minority of situations, vowel mutation ai > i, í indicates the plural:

Monosyllables AU

For monosyllabic words with au there are three possibilities: The word can receive an (optional) ending or it can undergo (optional) vowel mutation au > oi

Monosyllables OI

Monosyllabic words incorporating oi can be pluralized via vowel mutation oi >ui: Alternatively they can be pluralized using endings:

Monosyllables UI

Monosyllabic words with ui are always attested forming plural with various endings:

Polysyllables A-A

Polysyllabic words with a-a in the last two syllables show a rather consistent pattern. The first a is always subject to vowel mutation a > e. The second a usually undergoes vowel mutation a > ai but when the ending is such that for a monosyllabic word a > e(i) is required, this occurs in the last syllable: A single final -a has a different origin (as we will explore below in some detail) and undergoes a > y: In a single instance, an optional ending is seen, causing i-affection in the first syllable:

Polysyllables A-E

For polysyllabic words with a-e in the final syllables, the vowel mutation a-e > e-i is by far the dominant plural pattern: In two cases, i-affection isn't carried through the first syllable. A possible explanation in one case may be that the first syllable is actually a prefix, although i-affection seems to be carried out through prefixes in Noldorin as a rule. In a single case, the rather irregular pattern a-e > e-ai is seen: Likewise, in a single case plural by endings is observed:

Polysyllables A-I

There are not many polysyllabic words with the combination a-i - presumably beacuse usually internal i-affection would shift the pattern to e-i, so there are only words left where this doesn't occur (presumably because a prefix is recognized). Consequently, the words are pluralized based on the last syllable, i.e. using an ending and the first syllable remains unchanged.

Polysyllables A-O

The usual plural pattern for polysyllabic words with a-o is vowel mutation a-o > e-y, rarely a-o > e-i. Note that in our only trisyllabic example the somewhat surprising pattern o-a-o > oe-e-y is seen. In one case, onle pluralization by means of an ending is seen - this is presumably becaus the word is a compound word and its second element undergoes pluralization by an ending graug > graugir. In other cases, pluralization by an ending is optional:

Polysyllables E-E

Usually polysyllabic words with e-e seem to follow the pattern e-i(y) in plural: In one case however the pattern is different, reflecting the plural pattern of the second element of the compound gewg > gwaith:

Polysyllables E-I

If the last syllable in a polysyllabic word contains i, the plural pattern proceeds as in monosyllabic words with i by means of an ending:

Polysyllables I-A

We have one attested example showing ia > iai in plural:

Polysyllables I-E

Words with the combination i-e consistently form plural by vowel mutation i-e > i-i:

Polysyllables I-O

Words with i-o (usually the ending -ion often pluralize with en ending, internal i-affection is not seen in theses cases: However, there is also the possibility of vowel mutation in which case io >y occurs. The i-affection of the remaining syllables is consistent with the overall pattern:

Polysyllables O-A

The plural of nouns with the vowel combination o-a is consistently seen to be formed with vowel mutation o-a > (o)e > ai (note that the prefix or- in the first example is subject to i-affection here):

Polysyllables O-E

In words with the combination o-e the second syllable always undergoes vowel mutation e >i in plural. However, while the first syllable at times undergoes o > (o)e, it is also frequently seen unmodified, in one case the decision is optional:

Polysyllables O-I

Since this combination can only occur when internal i-affection doesn't modify the first syllable in the first place, vowel mutation would not lead to a distinct plural pattern and the plural is formed by an ending as expected:

Polysyllables O-O

The most common pattern seen for polysyllabic words in o-o in plural is vowle mutation o-o > e-y or alternatively o-o > y-y (in one case also o-o > y-i): Alternatively, in some cases endings are seen, in no case causing internal i-affection:

Polysyllables U-E

The few attested examples of polysyllabic words in u-e are consistently seen to undergo vowel mutation u-u > y-i:

Polysyllables U-U

The only word in this group shows the vowel mutation pattern u-u > y-y in plural:

Various other polysyllabic words

There are various other combinations of polysyllabic words involving a diphthong. However, since we have only one example for each of these groups, we can't really tell what is regular and what is irregular. We will therefore group the attested examples only in those forming plural by means of an ending and those forming plural by vowel mutation. The first group includes The second group includes:


The first and possibly most fundamental question is: Is there any pattern which would enable us to reliably tell if a given word would form plural by means of an ending or by vowel mutation? The frequent occurrance of optional ending/vowel mutation suggests that such a pattern does not exist in general. However, one pattern is observed rather consistently: Words with either i or ui in the final syllable (which would not have a distinct vowel mutation) always form plural by means of an ending. Apart from that, it is difficult to understand why we would see e.g. arch > eirch but barch > barchath.

Let us focus on plural formation by vowel mutation first. The main pluralization pattern by vowel mutation are for monosyllabic words:

Singular Plural
a ai, e(i) when -rX is following
e i
i plural by ending
o ui > y
u ui > y
ai i, í
au oi
oi ui
ui plural by ending
The patterns for polysyllabic words are:

Singular a-a a-e a-i a-o  
Plural e-ai, e-e(i) when -rX is following e-i, a-i for recognized prefix plural by ending e-y, seldom e-i  
Singular   e-e e-i    
Plural   e-i, e-y plural by ending    
Singular i-a i-e   i-o  
Plural i-ai i-i   io> y  
Singular o-a o-e o-i o-o  
Plural o(e)-ai e-i plural by ending e-y, y-y  
Singular   u-e     u-u
Plural   y-i     y-y

The seemingly irregular development of bala, gara and thala with plurals pl. bely, gery and thely has to be understood from the history of these forms - they are derived from prehistoric balgá, Old Noldorin garg and prehistoric stalga. The common element is that in all cases the final -a in Noldorin is not an original ending but a -g which became vocalized in the evolution from prehistoric Elvish to Noldorin. Thus, we may assume a development like Old Noldorin garg pl. *gergi with subsequent gi > y.

We can also understand the development ai > i, í from the history of the forms. We find prehistoric ekta > aith, kṛkta > craith, sikta > haith, nindyá > nainn and stainá > thain but camprú > caifr, daglé > dail, dagná > dain - thus, usually whenever the ai is derived from primitive a it remains and plural has to be indicated by an ending; if it has a different origin ai > i occurs in plural.

In many cases, the variations o > y vs. o > ui seems to be a conceptional change: o > ui is the pattern seen in the ENG whereas o > y is seen later. There is one exception: is derived from primitive togo and presumably involves an intermediate Old Noldorin *togi > tui.

The question of the mutation carried through prefixed elements is as puzzling as in later versions of the language - one the one hand, we see clear evidence for i-affected prefixes, cf. orfang pl. erfaing 'beardless' (PE13:156), on the other hand in other cases prefixed elements are unchanged, cf. amvenn pl. amvinn 'difficult' (PE13:159). It is difficult to determine Tolkien's ideas with confidence here.

The observed distribution of endings is difficult to understand. In general, a particular ending does not seem to be tied to a given phonological environment like a final consonant cluster, cf. ann pl. ennyn 'door' (PE13:137) but binn pl. binniath 'slope' (PE13:160) or ant pl. ennir 'face, front' (PE13:137) but hont pl. hontath, hynt 'trumpet' (PE13:163) - if there is any hidden correlation, it cannot be discovered with the given amount of examples. Likewise, usually endings don't seem to carry a particular semantic significance.

Comparison with the Noldorin of the Etymologies and Sindarin

All in all, the system of plural vowel mutations is remarkably similar to later Noldorin or Sindarin. This holds even for details such as the seemingly irregular pattern a > e observed in úgerth 'sins' (VT44:21,22) which follows precisely the pattern observed above. It is instructive to observe Tolkien's oscillation between the pluralization pattern a > ai (early Noldorin and Sindarin) and a > ei (late Noldorin). What sets Early Noldorin pluralization really apart is rather the presence of a significant amount of endings and the fact that certain conditions (i in the last syllable) demand an ending for pluralization.

Both Noldorin and Sindarin have attested plurals marked by endings, cf. drû > drúin in UT or fêr pl. ferin 'beech-tree' (LR:352, 381), however they are rare and presumably have a different explanation in the internal history of language development (cf. [2]). However, i in the final syllable doesn't seem to trigger an ending in plural in late Noldorin as we find Silevril 'Silmaril' both as singular (LR:383) and plural (LR:202) (see also the discussion in [4]).

The ending -(i)ath, denoting normal plural in Early Noldorin is used as general class plural in Noldorin and Sindarin, cf. geil 'star' class pl. giliath (LR:358, VT45:15) and Tolkien's remarks in RGEO:74.


I would like to thank Carl F. Hostetter for his help in the interpretation of some tricky passages in Tolkien's writings.


[1] The Goldogrin Grammar - an Introduction by Thorsten Renk

[2] Noldorin Plurals in the Etymologies by Bertrand Bellet

[3] Attested Sindarin Plurals by Anders Stenström

[4] Sindarin - the Noble Tongue by Helge Fauskanger

Thorsten Renk

Back to the index