The Goldogrin Grammar - An Introduction

Abbreviations used:

G: Gnomish
GL: Gnomish Lexicon
GG: Gnomish Grammar

PE: Parma Eldalamberon


Oddly enough, Goldogrin (Gnomish) is at the same time a very well-known language and a completely unknown language. This very early ancestor of later Sindarin, Tolkien's concept for the language of the Noldoli at the time of The Lost Tales, can be considered well-known because there is a 58 page lexicon available, there are grammatical notes on the Goldogrin noun (partially covering adjectives as well) and there are easily ten times as many examples for past tense formation than for Sindarin. The corpus of sentences in the language is comparable with the Sindarin corpus - so in fact the language ought to be more readily accessible than the rather elusive Sindarin.

Yet, at the same time, few people knowing Sindarin fully well are even aware of the existence of Goldogrin and while there are virtually hundreds of pages claiming to teach Sindarin in the web virtually nothing has been published about Goldogrin grammar (with the notable exception of two articles by Patrick H. Wynne which I found very helpful for this project).

Maybe it is because Goldogrin lacks the charme of the later Sindarin, both in phonology and in the connection to the LOTR elves, maybe the primary source (Parma Eldalamberon 11, containing the Gnomish Lexicon) is not readily available to readers. Be it as it may - it is the aim of this article to wake the language from its decade-long sleep and to demonstrate that we can in fact hope to write small bits and pieces in this early language by Tolkien just as we can use Sindarin and that Goldogrin is not an abandoned bit of unusable work only interesting for the academic-minded.

In the following essay, frequently Tolkien-made examples occur alongside deduced forms. In order to make a clear distinction, every deduced form, however plausible, is marked with a *. As a rule, no references to single words are given, I trust the reader knows how to look up word in a dictionary.

The Goldogrin corpus

Quite a few sentences are scattered throughout the grammatical explanations (the GG) and the wordlist (the GL proper). They allow some conclusions regarding the sentence structure:

i·vrog na cuid arog 'the horse is a swift animal' (PE11:9)
i·weg na an fofrin 'man is a foolish creature' (PE11:9)
i·weg fof '*man is a fool' (PE11:9)
i·winin na gwandron 'women are beautiful' (PE11:9)
ôni cailthi a·mabwid glen irtha 'He pressed a kiss upon her slender hands.' (PE11:11)
ôni cailthi mabir gleni nan·hirilion 'He kissed the slender hands of the ladies.' (PE11:11)
nôbi·i·mab 'len suila ontha. '*He took the slender hand of his daughter' (PE11:11)
nûmi Galmir i·saroth (nuir Saroth) 'Galmir sank into the sea' (PE11:12)
talwi i'loss ar gwandra nan·Idril 'the beautiful white feet of Idril' (PE11:15)
talin i'lossi ar gwandron nan·Idril 'the beautiful white feet of Idril' (PE11:15)
cuithos hû le mui 'a cat and dog life' (PE11:27)
i·fesc ar i·dusc 'a red rag to a bull' (lit. 'the irritating and the irritable') (PE11:34)
u 'wirn u 'wethrin 'not unwelcome or welcome' (PE11:47)
on iltathi nin pieg '*he stuck in a pin to me' (PE11:51)
inig bast no odog saith '*A small bread means great hunger later.' (PE11:51)
en nin·ista mai 'I am well aware of that.' (PE11:52)
nin·ista feg 'I feel ill.' (PE11:52)
a·laithra nin 'I forget it.' (lit. '*it is lost for me') (PE11:52)
u lâ fin sî 'No room for you here.' (PE11:52)
u laudin laithin hastath unweg '*Floods and times do not wait for anyone.' (PE11:53)
u laud u laith hasta unweg '*Neither flood nor time wait for anyone.' (PE11:53)
o·gwath lemp nin 'He beckons.' (lit. '*he shakes a crooked finger for me') (PE11:53)
im len 'I have come.' (PE11:53)
um lenin '*We have come.' (PE11:53)
luista nin 'I am thirsty.' (lit. '*it thirsts for me') (PE11:55)
i·bridwen a·vridwen 'judgement of fate' (PE11:64)

The noun

Most of the Gnomish grammar available in PE11 deals with the noun and the case system, hence we have a rather good knowledge of this part of the Goldogrin grammar. The following is a slightly shortened (leaving out the historical developments) and re-ordered summary of Tolkien's description.

Nouns and the definite article

Goldogrin nouns (and unlike in English, also names) can be determined by a definite article. In nominative, this article is in singular and plural if the following noun starts with a consonant and in· if it starts with a vowel. Hence

acha 'waterfall' -> *in·acha 'the waterfall'

Frequently, this in· is shortened to or 'n·:

aithi 'sword' -> *n·aithi 'the sword'

The definite article causes consonant mutations for the following word according to the pattern

basic lenited basic lenited
c ·g dr ·dhr
cr ·gr g ·'
cl ·gl gw ·'w
cw ·gw gl ·'l
t ·d gr ·'r
tr ·dr b ·v
p ·b br ·vr
pr ·br bl ·vl
pl ·bl h ·ch
d ·dh    
Apparently, any other consonants are not subject to consonant mutation. Hence e.g.

cuid 'animal' -> *i·guid 'the animal'
dorm 'summit' -> *i·dhorm 'the summit'
gwinc 'spark' -> *i·'winc 'the spark'
haig 'road' -> *i·chaig 'the road'
pui 'child' -> *i·bui 'the child'
falm 'wave' -> *i·falm 'the wave'
lôm 'shade' -> *i·lôm 'the shade'
madri 'food' -> *i·madri 'the food

However, modern Goldogrin makes an (optional?) distinction between ·'r-, ·'l- (as the result of a lost gr-, gl-) and true ·r-, ·l- (which remain unmutated). The latter two consonants are treated as vowels as far as the definite article is concerned, hence

glâm 'hatred' -> i·'lâm 'the hatred'
lam 'tongue' -> in·lam 'the tongue'

Case endings

The Goldogrin noun has three distinct cases, nominative, genitive and dative and three numbers, singular, plural and dual. The dual, however, is lost for most nouns and only preserved in a few words. According to how the noun is inflected for case and number one can roughly identify four different classes (the classification here different from Tolkien's own): In order to provide examples for the function of the various cases, we introduce the first declination (monosyllabic words ending in a consonant) here and discuss the others later on. For the first declination, we have the following standard pattern of case markers:
1st Declination Nominative Genitive Dative
Singular   -a -i
Plural -in -ion -ir
Dual -wi -wint -wid

Usually long syllables in such words become shortened as soon as an ending is present. For example the word glôr 'gold' follows the pattern

  Singular Plural
Nominative glôr glorin
Genitive glora glorion
Dative glori glorir
The nominative is used for the subject of a sentence, for the direct object and with all locative prepositions, occasionally it can be used as locative itself. Moreover, it can also be used as a reference. Hence

gweg 'man' -> *gweg amra 'a man wanders'
sîr 'river' -> *art sîr 'beside a river'
bar 'home' -> bar 'at home'
nôs 'birth' -> nôs mora 'good by nature'

The genitive is used by itself in possessive or partitive sense. It usually follows prepositions of ablative or derivative sense and occasionally can occur by itself as an ablative. Hence

gweg 'man' -> *ais gwega 'a man's knowledge'
geth 'clan' -> *gweg getha 'a man of a clan'
bar 'home' -> bara 'from home, away'

Finally, the dative is used for the indirect object of a sentence (the one benefiting from the action done on the direct object). Further uses include the allative, both in thought and in real motion, and the dative follows all prepositions with allative sense. Hence

gweg 'man' -> *golda antha aithi gwegi 'a gnome gives a sword to a man'
banc 'trade' -> *gweg tul banci 'a man comes for trade'
arn 'son' -> *ador tir ar arni 'a father looks at a son'
bar 'home' -> bari 'homewards'

Unlike in English (or the later Sindarin or Quenya for that matter), the definite article changes accordingly when a noun is inflected for case. We find

  before consonants before vowels
Nominative in·
Genitive na· nan·
Dative i· or ir· ir·
As far as we can tell, the inflected versions of the article cause mutations just like the uninflected forms if the noun starts with a consonant, with the exception of preconsonantal ir· which does not seem to cause mutation. Showing the declination of the definite article again using the example glôr 'gold':
  Singular Plural
Nominative i·'lôr i·'lorin
Genitive na·'lora na·'lorion
Dative i·'lori i·'lorir
Hence e.g.

tûr 'king' -> *i·vess na·dura 'the wife of the king'
gweg 'man' -> *golda antha in·aithi ir·gwegi 'the gnome gives the sword to the man'
roth 'cave' -> *'n·anna nan·roth 'the entrance of the cave'

Polysyllabic words ending in a constant (the second declination) differ from the first declination by the fact that a final vowel in such words is usually lost, hence there are no singular case endings (there's however an auxiliary mechanism to mark case if the definite article is not used for this purpose, see below). Therefore the pattern of case endings is

2nd Declination Nominative Genitive Dative
Plural -in -ion -ir
Dual -wi -wint -wid

Without the definite article, the singular cases are hardly distinct. The definite article can (partially) resolve this ambiguity. We show the example celeb 'silver':

  Singular Plural
Nominative i·geleb i·gelebin
Genitive na·geleb na·gelebion
Dative i·geleb i·gelebir

The third declination (monosyllabic words ending with a vowel) has a different set of (mostly consonant) case endings. Since those do not alter the number of syllables in singular, typically a long vowel stays long in those words. The relevant system of case endings is

3rd Declination Nominative Genitive Dative
Singular   -n -r
Plural -th -thon -thir
Dual -wi -wint -wid

We give the example of gwâ 'wind':

  Singular Plural
Nominative i·'wâ i·'wath
Genitive na·'wân na·'wathon
Dative i·'wâr i·'wathir

The pattern relevant for the fourth declination is in fact quite the same, however a final vowel -a, -u in a polysyllabic word is usually changed into o in genitive and dative singular. Hence we find (using the example of coma 'disease'):

  Singular Plural
Nominative i·goma i·gomath
Genitive na·gomon na·gomathon
Dative i·gomor i·gomathir

Nouns ending in -u may form the plural with -in, -ion, -ir as an alternative to the forms above, this often yields -wi- as in the example of culu, a poetic description of 'gold':

  Singular Plural
Nominative i·gulu i·gulwin
Genitive na·gulon na·gulwion
Dative i·gulor i·gulwir

Note however that culuth, culuthon, culuthir are valid forms as well.

Case prefixes

As we have seen, the 2nd declination does not have a case distinction in singular if the definite article is not present. Especially in this case, case-marking prefixes are useful. These prefixes are hence most important for the second declination, are the most common for the 3rd declination but can in principle be used for all nouns.

The prefixes are for genitive, go· for ablative and no· for dative. is here the exception, it is prefixed to the nominative whereas go· and no· are prefixed to the word having the appropriate case endings. Before vowels, these prefixes become an·, *gon· and *non·. The following example illustrates how the case distinction in the 2nd declination emerges again:

  Singular Plural
Nominative celeb celebin
Genitive a·geleb a·gelebin
Ablative go·geleb go·gelebion
Dative no·geleb no·gelebir

The distribution of endings across the various cases is best apparent for a 4th declination example:

  Singular Plural
Nominative coma comar
Genitive a·goma a·gomath
Ablative go·gomon go·gomathon
Dative no·gomar no·gomathir

Various auxiliary constructions

There is a suffix -lim that denotes a collective plural. However, most notably for nouns of the 3rd declination, this replaces the ordinary plural (case is then marked by prefix), hence both gwath and gwâ·lim are acceptable plural forms of gwâ 'wind', 'of winds' would then be either gwathon or, more common a·walim.

The most common use of this suffix is to denote races and people, often also -thlim or -thrim (race) appear. The choice which one is taken in a particular instance is usually determined by euphony. Hence e.g. gondothlim 'the people of Gondolin'.

The genitive can also be without any case marking. In this case, the word in genitive follows and is after a singular noun subject to mutation. Hence e.g. *faigli geleb 'long hair of silver' as an alternative to *faigli a·geleb. After a noun in plural, the genitive remains presumably unmutated.

Re-emerging stems

Sometimes, the nominative forms do not represent the proper stem to which the endings have to be appended. Often, these stems are archaic and not used in modern Goldogrin, e.g. celb- as the stem for 'silver' with the archaic genitive celba and dative celbi, but occasionally the distinction survives into modern Goldogrin. In this case, the stem or the plural is usually given in the lexicon, e.g. bactha 'a leg' with plural bacthin, hence this is bactha (bacth-).

The known duals

The few Goldogrin words with a known dual are

'hand' -> mabwi 'a pair of hands' (the GL gives mab as dual, though)
tâl (tald-) 'foot' -> talwi 'a pair of feet'
hen 'eye' -> henwi 'a pair of eyes'
heth 'brother or sister' hethwi 'a brother and sister'
hunt 'nose' (*'a pair of nostrils')

Some of these forms take the prefix go- as a kind of augmentation, hence gomabwi, godalwi and gochenwi.

The verb

There are no notes by Tolkien about the Goldogrin verb, hence we have to rely on the interpretation of verbs in the corpus and the wordlists to deduce the use of Goldogrin verbs.

Verb types

Like the later Noldorin and Sindarin, Goldogrin verbs fall into two main classes - primary (basic, stem) verbs and derived verbs. They are easily distinguished - stem verbs end with a consonant (e.g. mel- 'to love') whereas derived verbs typically show a derivational ending -tha, -na, -a and hence end always with the vowel -a (e.g. antha- 'to give').

In addition, there are a few verbs with a different pattern, most important among them the group of derived verbs with endings in -u, e.g. felu- 'to seem'. A few verbs are denoted as 'irregular' or 'special' and we will list them later.

Verb inflection - number

The verb seems to be inflected for number, compare the use of hasta- in u laud u laith hasta unweg '*neither flood nor time wait for anyone' where the verb is in singular to agree with singular laud and laith with u laudin laithin hastath unweg '*Floods and times do not wait for anyone.' (PE11:53) where the verb acquires a plural ending -th to agree with the pluralized laudin and laithin.

It would be tempting to argue that the verb ought to be pluralized in the first example since it refers to two things, and indeed it is pluralized in English. But in German, for example, the verb ends up in singular in a similar construction, hence one has to accept that Gnomish does not pluralize the verb in 'neither...nor...' constructions.

The verb na- 'to be' is said to have plural nain, the negative verb û has a plural uin, hence it seems that the verb essentially shares the noun plural markers -th and -in.

In the sentence i·winin na gwandron (PE11:9) the verb does not agree with the pluralized noun. This apparent discrepancy may be due to the fact that the GG predates the GL and that Tolkien decided on the plural nain later. After all, the example is quite generic for the use of the copula 'to be' and does not seem to involve any irregularity.

We may tentatively assume that the underlying principles for the choice of the plural pattern are the same as for nouns, i.e. derived verbs ending in -a would primarily form their plural in -th whereas stem verbs would prefer the plural -in; however verbs with the ending -tha may prefer alternative plurals -thain over -thath to avoid the repetition. If this is true, the plural of na- and û- would be specifically given since these verbs don't follow this pattern.

We do not know what number a verb would have to be in if the subject of the sentence is in dual - the most likely guess (based on number agreement of adjectives) is singular.

Verb inflection - person

Goldogrin has emphatic and non-emphatic pronouns (see Goldogrin Pronouns for a reference). Only the non-emphatic pronouns are joined to the verb by using them as prefixes.

The complete set of non-emphatic prefixes is as follows:

  Singular Plural
1st Person ni· me·
2nd Person fi· gwe·
3rd Person male *a·
3rd Person female *i· *a·
3rd Person neuter *a· *a·

In this table, it has been assumed that can be used as a general plural form for both male and female - after all, what of mixed groups?

The ordinary use of the pronomial prefixes can be inferred from a·laithra nin '*tt is lost for me.' (PE11:52) where the verb laithra- is prefixed by 'it'. As the example o·gwath lemp nin '*he shakes a crooked fonger to me' (PE11:52) indicates where the verb cwath 'to shake' occurs, the pronomial prefixes cause mutation for the following word. Finally, from nin·ista feg 'I feel ill.' (PE11:52) we may infer that the pronomial prefixes add an -n if the following verb starts with a vowel. Hence

*ni·chasta 'I wait'
*fi·chasta 'you wait'
*o·chasta 'he waits'

*me·chastath 'we wait'
*gwe·chastath 'you wait'
*a·chastath 'they wait'


*nin·antha 'I give'
*fin·antha 'you give'
*on·antha 'he gives'

*men·anthain 'we give'
*gwen·anthain 'you give'
*an·anthain 'they give'


*ni·mel 'I love'
*fi·mel 'you love'
*o·mel 'he loves'

*me·melin 'we love'
*gwe·melin 'you love'
*a·melin 'they love'

However, as nôbi·i·mab 'he took the hand' (PE11:11) indicates, the 3rd person singular may be expressed without writing the pronoun explicitly (this appears to be just as in later Sindarin).

Past tense

The past tense of Goldogrin verbs is quite complex and what is known about it is described in Patrick H. Wynne's article The Goldogrin past tense. Several mechanisms are known how the past tense is formed, and it is not uncommon for a Goldogrin verb to have more than one valid past tense. At this point, we simply provide some general rules, to find the past tense of an individual verb it is best to consult the lexicon and only to apply these rules if no explicit past tense can be found there.

Primary verbs with a final consonant -r, -l, -s, -v, -w, -m, -n or -ng primarily form their past tense by strengthening of the stem vowel. This usually implies the shifts a > ô, e > î, i > ai, o > û and u > au/û, for the latter choice, there seems to be a slight tendency of verbs ending with -m to do the shift into û. In addition to the strengthening, an -i is appended to all forms. Therefore

bas- 'to bake' -> bôsi 'baked'
mel- 'to love' -> mîli 'loved'
gil- 'to gleam' -> gaili 'gleamed'
nor- 'to run' -> nûri 'ran'
nus- 'to take note' -> nausi 'took note'
cum- 'to lie down' -> cûmi 'lay down'

Primary verbs with the final consonants -b, -d and -g have an about equal probability to form past tense by strengthening of the stem vowel (as outlined above, see e.g. sog- 'to drink' > sûgi 'drank') or by nasal infixion. Nasal infixion implies that final -m > -mf, final -n > -nth and final -g > -nc to which then an -i is appended (e.g. cap- 'to jump' > camfi 'jumped'). A few verbs even show both past tenses (e.g. nag- 'to chew' > nôgi, nanci 'chewed'). We find for example

heb- 'to bind' -> hemfi 'bound'
dod- 'to fall down' -> donthi 'fell down'
fag- 'to cut' -> fanci 'cut'

There are some primary verbs ending in a vowel. Their chief way of forming the past tense is the suffix -thi. Several other basic verbs make use of this mechanism as well, but mainly as an alternative to other past tense formation mechanisms.

'to endure' -> rôthi 'endured'
gai 'to possess' -> gaithi 'possessed'

The largest class of derived verbs with endings in -tha changes this ending in the past tense to -thi (there is, however, quite a number of derived verbs ending in -tha which form their past tense by stem vowel strengthening). Therefore

bactha- 'to walk' -> bacthi 'walked'
gultha- 'to endure' -> gulthi 'endured'
but e.g.
antha- 'to give' ôni 'gave'

Verbs ending in -u usually form a valid past tense by appending -i, the ending is then changed to -wi:

felu- 'to seem' -> felwi 'seemed'
telu- 'to end' -> telwi 'ended'

Four derived verbs form their past tense (optionally) by using the suffix -ni:

lintha- 'to ring a bell' -> linthani or linthi 'rung a bell'
rûtha- 'to dwell' -> rûthani 'dwelled'
santha- 'to show' -> santhani 'showed'
saptha- 'to dig' -> sôbi or sapthani 'dug'

It stands to reason that the inflection of these past tense forms works just like in present tense, hence:

*nin·ôni 'I gave'
*fi·chemfi 'you bound'
*gwe·mîli 'you love'
*o·fanci 'he cut'

The participles

There's no mention of the participle in the GG. Likewise, we do not have many clear examples of participles in the courpus of sentences, hence in finding out how these forms are supposedly formed we have to rely on information in the wordlist. If the dictionary would be organized in a standard way, this task would be hopeless (and English dictionary would not list regular verb forms, only irregularities), but luckily this is not how Tolkien organzied his writings. Even so, in collecting all examples of possible participles, we have to be aware of the possibility that we may well confuse a participle derived from the verb with and adjective derived from the same stem the verb originated from. Keeping this uncertainty in mind, we can nevertheless offer some educated guesses of how these forms may look like.

A present active participle is explicitly given for the verb na- 'to be', it is ol·, the gloss indicating that it ought to be prefixed to something. It is up to anyone's guess what is supposed to follow after this form - should we assume ?ol·bar 'being at home'? Or rather ?ol·gwadra 'being beautiful?' Or should this indicate that this is the way verbal participles are formed, hence ?ol·lintha 'ringing'?

The last alternative, though, does not seem to be very likely. The list of candidate forms contains several examples for which -ol seems to be used as an ending, among them

briga- 'to fear' -> brigol 'afraid' (lit. '*fearing')
gwirtha- 'to not wish' -> gwirthol 'reluctant' (lit. '*not wishing')
hadha- 'to cling to' -> hadhol 'abiding' (lit. '*clinging to')
hirtha- 'to care for' -> hirthol 'attentive, careful' (lit. '*caring')
lentha- 'to approach' -> lenthol 'approaching'

All these examples involve derived verbs, and the rule seems to be that the final -a is replaced by -ol. There is, however, also an example in which the complete derivational ending is lost before -ol is appended:

ictha- 'to excite' -> igol 'exciting'

Occasionally, we see the same thing with a consonant shift, possibly restoring an earlier consonant:

brigla- (from *bridla-) 'to change' -> bridol 'changing'

A few primary verbs are also attested with present participles:

mug- 'to keep silent' -> mugol 'taciturn' (lit. '*keeping silent')
naf- 'to suspect' -> nafol 'suspicious' (lit. '*suspecting')
uir- 'not to will' -> uirol 'unwilling'

Hence the rule appears to be the same, the ending -ol is appended to the stem.

A perfect active participle is given by the form len, pretty clearly identified as such by the example im len 'I have come'. This form seems to belong to the verb lentha- 'to come, to approach' (of which we also know lenthol 'approaching'). If so, the bare stem, after the loss of the derivational ending would act as perfect active participle. There is one more example which might support this theory, the pair camma- 'to bow' and cam 'bent, bowed' (being an intransitive verb, this apparently is 'having bowed').

This formation pattern would lead to a problem for stem verbs, since here the stem is identical with the present tense form, hence ?im mil could both be 'I love' and 'I have loved'. Presumably then, the stem verbs have a different pattern (or do not form this participle at all).

There are several examples of the formation of a form which looks like the perfect passive participle in the wordlists, among them derived verbs

beltha- 'to unroll' -> belin 'expanded, unrolled'
bentha- 'to shape' -> benin 'shapely' (lit. '*shaped')
fadra- 'to sate' -> fadin 'sated'
gultha- 'to carry' -> gulin 'burdened'
guira- 'to possess' -> guin 'owned'
ictha- 'to excite' -> igin 'excited'
lûtha- 'to pass (time) -> luin 'past'

and basic verbs

drib- 'to wear out' -> dribin 'worn out'
fag- 'to cut' -> fagin 'cut'
sam- 'to arrange' -> samin 'arranged'

Thus, the standard pattern seems to be that basic verbs append an ending -in to the word stem whereas derived verbs drop the derivational ending before -in is appended. Verbs with the stem ending in -r, -s and -l (regardless if primary or derived verb) seem to follow a slightly different pattern, though:

cartha- 'to make' -> carn 'made'
fur- 'to conceal' -> furn 'concealed'
gwirtha- 'to not wish' -> gwirn 'unwelcome' (lit. '*unwished')
îr- 'to intend' -> irn 'wished for'

We may conclude that for a stem ending in -r the ending used to form the perfect passive participle is -n rather than -in.

heltha- 'to freeze' -> helon 'frozen'
faltha- 'to strip' -> falon or falin 'stripped'

Thus, it seems like verbs with stems in -l occasionally prefer an ending -on. There is one example where a different verb does so as well, heb- 'to bind' and hebon 'bound'.

suthra- 'to hush' -> susc 'hushed, quiet'
thas- 'to shave' -> thasc 'shaven'
thista- 'to dry up' -> thisc 'dry' or thisin 'parched'

It appears as if the participle for the verbs with a stem in -s would be formed by removing the ending and changing the final -s into -sc.

The imperative

There are a few pairs of verbs and exclamations in the wordlist which may give a hint about how the imperative is formed - keeping in mind that the list might contain the exceptions rather than the rules we can analyze them as follows:

elma- 'to marvel at' -> elm! or elum! 'marvelous! think of that!'
haitha- 'to go, to fare, to walk' -> hai! 'go!'
hitha- 'to listen to' -> hith! 'hearken!'
suthra- 'to hush' -> suss! 'quiet!'
and finally we find
bâ! 'be gone!'
haiva! 'be gone!, be off!'

Unfortunately, there are no stem verbs in this list. apparently is not associated with a particular verb, and haiva evidently is a compound of hai + bâ (this is in fact supported by the way Tolkien references this form). We are left with two imperatives derived by removing the ending -a from a longer derivational ending -tha, -ma and one by removing the complete ending -tha. Giving this scarce evidence, it is difficult to judge what the normal pattern would be - probably euphony might make the decision.

The infinitive

Verbs in the GL are usually quoted with their stem (indicated by a hyphen) and the translation given sometimes including 'to' and sometimes not, cf.

ertha- (...) 'to isolate' (...)
fag- (fanci) 'cut'

However, occasionally we find verb forms with a translation in infinitive without the hyphen, e.g.

hada (hanthi) 'throw at'
lob (lompi) 'run, gallop (of animals)
loda (lonthi 'swallow'
the latter form changed from a previous entry lod- (lonti).

It is difficult to say if the intended meaning of these entries is indeed an infinitive or if the missing hyphen simply repesents an omission like the omission of 'to' in the translation. The form loda changed from lod- may indicate that stem verbs form an infinitive with the ending -a, but on the other hand, the form lob shows no such ending and yet has an equivalent translation. Given that, the best guess about the infinitive forms is that for derived verbs the bare verb can be used, for stem verbs likewise. There is no clear sign that the infinitive would be marked by a particular ending like in Noldorin.

The verb 'to be'

The verb 'to be' is glossed as na in the GL, its plural is nain, the past tense is given as thi (presumably the plural is *thin rather than ?thith) and the participle is ol·. A remark tells us that the verb is 'quite irregular', so it may not be permissible to generalize to other verb forms.

The example i·winin na gwandron 'women are beautiful' (PE11:9) is somewhat puzzling because it shows the singular verb despite the subject in plural. We may interpret this either as a slip by Tolkien or as a signal that the pluralization is optional or simply as something which was revised from the early GG to the later GL.

As the example i·weg fof '*man is a fool' (PE11:9) indicates, it is entirely permissible to leave out na alltogether. We do not know if this is true for the past tense as well.

We may assume that the following examples may be acceptable:

*in·aith [na] inc. 'The sword is small.'
*in·aithi nain inci. 'The swords are small.'
*i·dûri thin goldath. 'The kings were Gnomes.'
*ol·golda mora. 'Being a Gnome is good.'

The adjective

The GG contains about one page of writings on the adjective, many things must therefore be deduced from the corpus or the wordlists.

Number, case and mutation

Goldogrin adjectives usually follow the noun they describe and agree in number, but not in case. Adjectives ending in -a form a plural in -on, those ending in -i become -in and those in -u change into -on. Monosyllabic adjectives ending in a consonant add the plural ending -i but polysyllabic adjectives remain unchanged if they end with a consonant. Thus

bess mora 'a good wife' -> bessin moron 'good wifes'
*bess faiglion 'a wife with long hair' -> *bessin faiglion 'wifes with long hair'

As evident from the pluralized form of gwandra 'beautiful' in i·winin na gwandron 'women are beautiful' (PE11:9), not only the attributive adjective but also the predicative adjective agrees in number.

Attributive adjective following a singular noun are subject to soft mutation (given by the table above). This is not the case if the noun is in plural. Hence

mab 'loss 'a white hand'
mabin glossi 'white hands'

If the noun is in dual, the adjective remains in singular but is not mutated. Hence

mabwid glen 'slender hands' (of one person)
mabin gleni 'slender hands' (of several persons)

Based on the actual evidence, it is unclear if a predicative adjective in singular would be mutated. Presumably this is not the case, because a close association of the adjective and the noun is required for mutation. This is evident from the following rule: If a noun is qualified by a definite article, two or more adjectives and even a genitive, then the word order is noun - article - adjectives - genitive. Hence

talwi i·'loss ar gwandra nan·Idril 'the beautiful white feet of Idril'

The only mutation in this example is caused by the definite article acting on the first adjective. Occasionally adjectives may precede a noun as well.

Comparative, superlative

Goldogrin has two sets of comparative and superlative endings - the endings -ro(n) and -nta augment an adjective whereas the endings -thir and -nci diminish it. However, the superlative form does not really correspond to the English superlative but rather to an expression involving 'very, exceedingly'. For two adjectives which have a negative meaning to start with (feg 'bad' and inc 'little') the role of the forms is confused - both the diminishing form of these adjectives and the (rarer) augmenting form actually augments their meaning.

For adjectives in -a (and presumably all adjectives ending in vowels) the comparative endings take the form -dro(n) and -nthir. For the augmentative form, the final -a is shifted to -o. This shift does not take place for the diminishing form. Again, presumably adjectives in -i leave this vowel whereas -u would also change into -o, however this is not attested. One finds

gwandra 'beautiful' gwandrodron 'more beautiful' gwandronta 'exceedingly beautiful'
gwandra 'beautiful' *gwandranthir 'less beautiful' gwandranci 'really not beautiful'

Adjectives ending in a consonant take the endings -ron, -onta, -thir and -inci, various not well documented consonant shifts take place, we learn that

feg-thir > faithir 'worse', also *feg-ron > fedron 'worse'
*inc-thir > inthir 'smaller', also incron 'smaller'
moranthir > monthir '*less good' *mora-dron > modron 'better'

These forms are indeclinable, presumably implying they are not inflected for number.

Derived adjectives

Goldogrin has a rich set of endings to derive adjectives from nouns. The most frequent ones are -(i)ol, -(r)in, -og and -wed. Out of these, the first few usually denote 'of the quality of the thing in question', 'like the thing in question'. It appears that -in has the strongest tendency to denote 'of the quality' and materials, followed by -rin and usually -iol denotes only 'like the thing'. Hence, if there is an adjective in -in denoting the material, -rin denotes 'like the material':

tam 'copper' -> tambin 'of copper' -> tambrin 'like copper'
gais 'steel' -> gaisin 'if steel' -> gaithrin 'like steel'

However, if the adjective denoting substance is derived with -rin, -iol does the comparison:

celeb 'silver' -> celebrin 'of silver' -> celebriol 'like silver'
glôr 'gold' -> glôrin 'of gold' -> glôriol 'like gold'

Unfortunately, the rules are hard to predict. -og and -wed usually denote 'having the thing in question', here -og can also denote 'experiencing the thing in question'. This is most evident from

dair 'merriment' -> dairog 'merry (of persons)' -> dairol 'merry (of things)' -> dairwed 'merry' (general)

A lot of examples can be found:

gothri 'warfare' -> gothriol 'warlike'
gwilbrin 'butterfly' -> gwilbriniol 'like a butterfly'
nandri 'country' -> nandriol 'rural' (lit. '*like countryside')
sîr 'river' -> siriol 'flowing' (lit. '*like a river')

flass 'foam' -> falthrin 'foamy' (lit. '*of foam')
ger 'metal' -> gerin 'metallic'
glast 'marble' -> glastrin '(of) marble'
gling 'music' -> glingrin 'musical'

aith 'thorn' -> aithog 'thorny' (lit. '*having thorns')
gruith 'anger' -> gruithog 'ferocious'
maith 'rule' -> maithog 'having control'
polm 'strength' -> polmog 'strong'

clog 'stone' -> clogwed 'stone-covered'
fem 'venom' -> femwed 'venomous'
gonn 'rock' -> gonwed 'rocky'
thairin 'magic' -> thairinwed 'magical'

There are, however, some derived adjectives which deviate from these general trends.


Information about pronouns can be deduced from the courpus or found scattered throughout the wordlists. For the detailed analysis see Patrick H. Wynne's article Goldogrin Pronouns.

Non-emphatic pronouns - nominative

As mentioned above, non-emphatic pronouns in nominative are used as prefixes for the verb. They are

  Singular Plural
1st Person ni· me·
2nd Person fi· gwe·
3rd Person male *a·
3rd Person female *i· *a·
3rd Person neuter *a· *a·

Emphatic pronouns - nominative

In addition, Goldogrin has a set of emphatic pronouns which can act as an independent subject of the sentence. They are

  Singular Plural
1st Person im um, umin
2nd Person ?tha oth, othin
3rd Person male on onin
3rd Person female ir *irin
3rd Person neuter an atha

If the independent emphatic pronoun is used, the verb does not get a second pronomial prefix as apparent from the example on iltathi nin pieg '*he stuck in a pin to me' (PE11:51). We observe that independent pronouns can be combined with participles without the verb 'to be', cf. im len 'I have come.' (PE11:53), therfore presumably the same use is legitimate for adjectives:

*im gwandra 'I am beautiful.'
*onin fegi 'They are bad (boys).

Pronouns in dative

Goldogrin pronouns can be inflected for case - for the dative, the relevant endings are -n in singular and -r in plural, they are appended to the non-emphatic prefix forms but form independent pronouns then. In plural, often vowel shifts occur. We find

  Singular Plural
1st Person nin mir
2nd Person fin gwir
3rd Person male ?on ?ar
3rd Person female ?in ?ir
3rd Person neuter ?an ?ar

The 3rd person pronouns have been derived on the assumption that the inflection stays the same - producing forms which overlap with other pronomial forms - it is unclear if Tolkien would have accepted that clash or not.

To give some examples:

*on ôni nin celeb 'he gave me silver'
*a·santha mir in·acha 'they showed to us the waterfall'

Pronouns in genitive - Possessives

The possessive pronouns are apparently formed by appending the ending -ntha in singular and -thra in plural. We find

  Singular Plural
1st Person *nintha *methra
2nd Person fintha gwethra
3rd Person male ontha *athra
3rd Person female irtha *athra
3rd Person neuter *antha *athra

Like genitives/adjectives, they follow the noun whose possession they indicate, cf. mabwid glen irtha 'her slender hands' (PE11:11) or suila ontha 'of his daughter' (PE11:11). Hence maybe

*on·ôni nin in·aithi fintha. 'He gave your sword to me.'
*on·ôni nin aithi i·wandra fintha. 'He gave your beautiful sword to me.


Apparently the Goldogrin system of mutations is more complex than the simple table given in the GG, several remarks in the GL hint at this. However, we have little actual knowledge of other mutations and have to deduce them from words in the list, assuming that would be possible.

Exceptions to the lenition rules

There are several cases given in the GG where the normal rules of lenition as outlined in the table above are altered.

The main group consists (in fact just like in Noldorin or Sindarin) of words in which an original nasalized stop is restored when the word is subject to lenition. This usually yields d- -> i·nd-, b- -> i·mb- and g- -> i·ng-. This type of lenition is known to take place for the following words:

dor 'land' -> i·ndor 'the land'
dolm 'pit' -> i·ndolm 'the pit'
deldron 'beech' -> i·ndeldron 'the beech'
doldrin 'mole' -> i·ndoldrin 'the mole'
drith 'savour' -> i·ndrith 'the savour'

Belca 'Melko' -> i·Mbelca 'Melko'
bast 'bread' -> i·mbast 'the bread'
basgorn 'loaf' -> i·mbasgorn 'the loaf'
bar 'home' -> i·mbar 'the home'
barwen 'homestead' -> i·mbarwen 'the homestead'
Bandoth 'Mandos' -> i·Mbandoth 'Mandos'
Balrog 'demon' -> i·Mbalrog 'the demon'
bal 'anguish' -> i·mbal 'the anguish'
Bridwen 'fate' -> i·Mbridwen 'the fate'
bless 'grace' -> i·mbless 'the grace'
bothli 'oven' -> i·mbothli 'the oven'
golda 'gnome' -> i·ngolda 'the gnome'
Goldriel 'Goldriel' -> i·Ngoldriel 'Goldriel'
golma 'lore' -> i·ngolma 'the lore'
goldobar 'Gnomeland' -> i·ngoldobar 'the Gnomeland'
Gainu 'Angaino' -> i·Ngainu 'Angaino'

Furthermore, all words with the prefixes go-, gwa- 'together' belong to this group, among them

gwanos 'family' -> i·ngwanos 'the family'
gôloth 'forest' -> i·ngôloth 'the forest'
glest 'moot' -> i·nglest 'the moot'
godaithri 'education' -> i·ngodaithri 'the education'
gogel 'mouth' -> i·ngogel 'the mouth'

By analogy, all words in gi- are mutated as if they would be derived from original nasalized stops, hence

gilim 'winter' -> i·ngilim 'the winter'

After na· or , all words in ga- are also treated as if derived from nasalized stops, hence

gadron 'fellow' -> *i·'adron 'the fellow' -> *a·ngadron 'of a fellow'

Finally 'many words in go-, ga- unaccented' (we do not learn of the exceptions or precisely how many:

gothweg 'warrior' -> *i·ngothweg 'the warrior'

Liquid mutation

We can deduce from a number of prefixes that Goldogrin (unlike Noldorin and Sindarin) does not seem to have any consonant changes following r,l in a prefix. For example

tanc 'steady' -> iltanc 'unsteady'
giol 'fecund' -> ilgiol 'barren'
gultha- 'weigh' -> argulthion 'equal'

Stop mutation

The prefix a(d) 'into' is said to cause stop mutation, but no rules as to how this mutation might work are given.


Goldogrin exhibits a rich set of negation elements.

Negative prefixes

There are three chief differences in meaning in negation expressed in Goldogrin:

Hence e.g.

giol 'fecund' -> ilgiol 'barren' -> *ungiol 'not fecund (but not barren either)'
carm 'deed' -> ulcarm 'sin' -> *ugarm 'not a deed'

The negative verb

The GL includes a negative verb û 'not to be, not to do' with plural uin and past tense ûthi, ûthin. The translation suggests that the verb can be used in combination with adjectives and nouns, cf.

*nin·û golda. 'I am not a gnome.'
*nin·û feg. 'I'm not bad.'

as well as with verbs in infinitive, cf.

*men·uin thê 'We do not see.'

Negative particles

û- is glossed in the GL as 'negative prefix with any part of speech' - comparing with the list of negative prefixes above, we may assume that it is entirely legitimate to use the short u- as well. We may infer from the description that it can be used to negate nouns, adjectives and verbs, hence

*nin·u-thê 'I do not see.'
*Im u-ngolda 'I'm not a gnome.'
*Im u-feg. 'I'm not bad.'

But as u laudin laithin hastath unweg (PE11:53) seems to indicate, it maybe possible to write all these as *nin·u thê, *Im u (n)golda and *Im u feg. instead. This is nicely confirmed by the example u lâ fin sî 'No room for you here.' (PE11:52)

Indeed, we find an entry in the GL describing u -u 'not... nor', and its use is illustrated by the example u laud u laith hasta unweg '*Neither flood nor time wait for anyone.' (PE11:53). So, a single u can be used as 'not, no' whereas a repetition means 'neither... nor'.

Double negation

In English, a double negation has a tendency to cancel - 'No-one has never seen the Yeti.' is not the same as 'No-one has ever seen the Yeti.' - the latter meaning that the Yeti has not been seen whereas the former sentence implies that there is no one who never saw the Yeti, i.e. everyone has seen it.

Goldogrin is (probably) quite different here - in order to get the correct meaning of u laud u laith hasta unweg which literally translates 'not flood not time wait for no one.' one has to assume that the double negation doesn't cancel - otherwise the implied meaning would be that flood and time wait for everyone, which is presumably not the case. Hence it once a sentence is negated, it is permissible to add further negations to its elements without changing the meaning.


The corpus of Goldogrin sentences doesn't permit too many observations, but there are some facts which can be deduced from it and some scattered remarks throughout the wordlists.

Word order

It is hard to judge what the normal word order in Goldogrin ought to be: We find

It seems that the word order is rather free and emphasis determines what word is found at the beginning of the sentence. This would well agree with the observation that the set of independent (emphatic) pronouns is usually found at the beginning of a sentence, cf. on iltathi nin pieg (PE11:51) and im len, um lenin (PE11:53). It is rather less surprising to find the non-emphatic pronouns in front of the verb since they appear in the form of prefixes.

Likewise, there appears no particular preference in the order of objects in different cases. In ôni cailthi a·mabwid glen irtha (PE11:11) we see the accusative object preceding the dative, in on iltathi nin pieg (PE11:51) we see just the opposite word order. In Goldogrin (which is, unlike later Sindarin, a language with case endings) this is no problem - the case marking allows to identify the function of a particular word in any position.

According to PE11:15, adjectives may and do precede the noun frequently. None the less the normal rule is that adjectives and adjectival genitives and equivalent forms succeed the qualified noun as nearly as possible. It appears that the possessive pronouns are counted similarly among the 'equivalent forms' - we find mabwid glen irtha 'her slender hands' (PE11:11) and suila ontha 'of his daughter' (PE11:11), in both cases the possesive pronoun follows the possessed object.

Impersonal verbs

A few verbs in the wordlists are explicitly marked as impersonal or can be deduced to be such. The verb oltha- 'to appear as an apparition' is given with an explanation 'impers. c. dat. I dream'. The sentence a·laithra nin 'I forget it.', lit. '*It is lost for me.' (PE11:52) gives an explicit example - the subject of the English sentence appears in dative, in the Goldogrin sentence there is no subject.

Note that there is no translation of the English 'it' of 'it is lost for me' in the Goldogrin sentence - the English language has a subject requirement, i.e. even if there is no thing logically doing the action which could replace the pronoun (consider 'it is raining' - neither 'the weather is raining' nor 'the clouds are raining' would be used much in English) the pronoun 'it' is filled in. But other languages may not necessarily have such a requirement. In Goldogrin where it is permissible to leave out the pronoun 'it' in any case this may not be much of a deal, but it implies that **an oltha nin would not be correct Goldogrin whereas *oltha nin would be a valid expression for 'I dream'.

There are several other verbs which seem to fall into this class:

luista- is marked as impersonal and given with the example luista nin 'I am thirsty'
nictha- is glossed 'it is raining, hailing, is snowing'
fôtha- is glossed 'it snows'


This article owes much to Patrick H. Wynne's two essays Goldogrin Pronouns and The Goldogrin past tense which have been of enormous help in compiling this grammar.

Thorsten Renk

Back to the index