Khuzdul – the language of the Dwarves.

 

 

Preface

This is by no means a complete grammar. Actually, it is little more than a compilation of my theories on Khuzdul and my attempts to make up new aspects of this language. My primary sources are of course the works of JRR Tolkien, though I rely heavily on the excellent compilation of the current knowledge on Khuzdul by Helge Kĺre Fauskanger. I am no professional linguist; this is just a hobby of mine and I am but trying to make as much of my limited knowledge as possible.

Some of the elements in here are directly attested, some I conjectured, and some I simply made up.

If you have any comments or additional ideas on this (and I’m rather sure you do), please do not hesitate at telling me.

 

PS: Since English is not my native tongue, there may be some grave mistakes in here – I apologise for that in advance.

 

1.       The Noun

 

1.1    Substantive

 

Numbers: There are four complete numbers in Khuzdul: singular, septal, indefinite plural and definite plural. All plural forms employ a complex system of broken derivation (detailed below); yet there is a set of plural suffixes, which is used only with derived nouns today (that is, agentive or abstract forms of adjectival or verbal roots). It is said those suffixes were more common in Ancient Dwarvish, and that they could be applied to all nouns, back in the days when Khuzdul was more akin to Valarin and thus used sound derivation more often than today.

The definite plural is also used as a collective form.

In addition, we have a fragmentary dual that is only employed when referring to body-parts like eyes, ears, feet etc. The case of hands is an especially interesting one: the root [M-K] seems to appear in dual only, as do all related roots, like the ancient emphatic form [M-H-L]; yet we have a strange singular construction for the right hand only (for most Dwarves are right-handed), with prefixed the- (in effect themęk, the right hand), which seems to be related to the definite prefix thu-.

 

The Dwarves have always associated ‘creation’ with ‘hands’ – quite understandably, given the kind of work so many of them excel in – and thus the Seven Fathers invoked Aulë by the name męk, ‘hands’, in the language he had given them. But the desire arose in them to find a more glorious name for their greatest deity than a term for two appendages every dwarf was equipped with; and so it came that when they learned Aulës true name, a3űlęz, they merged this strange but noble word with elements from their own tongue, thus forming ma3âl, which not long after became mahal (for the sound of 3 was alien to their tongue). And to this day, the Dwarvish name for Aulë is Mahal.

 

1.2    Genitive

 

We have an attested genitive structure, a plain broken declination with some pretty straightforward rules. This is called the status constructus or Construct State (CS) in hamito-semitic linguistics, as opposed to the status absolutus. The CS word order is X(CS) Y, roughly meaning X of Y, with X being put into the CS. (The word order X Y in the status absolutus is discussed below).

If X is a compound, the change to CS afflicts only the second element of X, if both elements are nouns; if not, only the noun is put into the CS. Exception: If Y is a compound, only the second (or noun) element of Y is put into the CS (e.g. uzbad khazad-dűmu instead of *izbid khazad-dűm).

In contrast to the substantive, the genitive of derived nouns is formed by a special broken structure.

The genitive also functions as the basic possessive structure, following the same rules: The possessive object in Construct State is placed before the subject. If, however, it is necessary to distinguish genitive from possessive forms in a sentence, the genitive is constructed as detailed, while bound pronouns (see below) are employed to denote the possessive, followed by a noun in the Construct State, in which case the CS takes on the function of a reflexive structure.

 

1.3    Nominative

 

The function of the nominative is to be the subject of the verb. Basic nominative is typically identical with the substantive; but there is one case in which the subject appears in the Construct State: if bound pronouns are employed as possessive particles, as detailed below.

The nominative is defined by its position in a sentence. The first substantive in a sentence is always the subject, except if there is a free pronoun in the sentence, or bound pronouns are employed as possessive particles. In the first case, the first substantive is the object, while the free pronoun acts as the subject. In the second case, the first substantive is the object as well, but the noun in the Construct State following the possessive particle is the subject.

                Example sentence:

Khazâd ai-męnu.

The Dwarves are upon you.

                Example sentence:

Rukhs âl zughul.

Orc he battle.

                Example sentence:

Baraz [RM2]th’[RM1]  ussu [RM3] uzbad-lę bundu [RM4] gabil[RM5]. 

Red the horse-of the king-his head-of is great. [The head of the king’s red horse is great.]

 

Note the final position of the verb. I choose to believe that all Khuzdul verbs maintain this position, either at the end of the complete sentence, or at the end of a sequence.

 

1.4    Accusative

 

The function of the accusative is to be the object of the verb. [My ideas on the exact form of the accusative aren’t quite clear yet]

 

1.5    Dative

 

The dative is defined by a simple nunation (final -n or -in/-’n after a vowel) added to the object. Its meaning is roughly ‘to’/‘for’.

 

1.6    Noun chart

 

                Key: Small letters, e.g. x, a, u, denote vowels; x denoting any vowel (in irregular forms). Digits, e.g. 1, 2, 3, denote the root consonants or radicals.

 

1.6.1          Substantive

 

For genuine nouns

Singular

Dual

Septal

Indef. Plural

Def. Plural

One radical

1x

Unknown

ę1

u1

ű1

Two radicals

1â2a

1ę2

ę1â2

1u2

1ű2

Three radicals

1u23 or 1i2a3 or 1a2o3

1a2â3

ę1e23 or ę1î2o3

1a2a3 or 1a2ű3 (archaic)

1a2â3

 

For derived nouns

Dual

Septal

Indef. Plural

Def. Plural

After vowel

–dab/–ab

–ten

–tash/–ash

–tab/–ab

After consonant

–ab

–en

–ash

–ab

 

Since the Khazâd seem so obsessed with the number seven, I have found it appropriate to include a septal form. The construction with initial ę (which is of course completely hypothetical) probably occurs only in septal. Thus, if we look at the radicals [Kh-Z-D], assuming that – as Mr Fauskanger suggested – they have something to do with the number seven (which becomes even more plausible when compared with Adűnaic hazid), we could derive a hypothetical Khuzdul word for ‘seven’ in a form like *ękhezd(?).

The plural noun dűm is well attested and is the source for my plural structure for two-consonantal roots 1ű2. The singular structure 1â2a I derived from the element –nâla in Kibil-nâla, Khuzdul for the river Silverlode (Celebrant). Indefinite Plural is again formed by shortening the vowel.

Words with three radicals seem to be the most common form in Khuzdul. They appear in a variety of singular forms, like 1u23, attested in bund; 1i2a3 in zirak and bizar (and presumably in *mibar, see below); and 1a2o3, attested in gathol only. The definite plural construction 1a2â3 we know from khazâd, tarâg etc.; according to my theory, indefinite plural 1a2a3 is attested in mazar (as in Mazarbul; see below). The indefinite plural form 1a2ű3 is to be found in shathűr.

For monoconsonantal roots we have only the plural űl, from which comes the definite plural form ű1, and the indefinite plural ul in Azanulbizar.

 

1.6.2          Genitive

 

For genuine nouns

Singular

Dual

Septal

Indef. Plural

Def. Plural

One radical

1^x

Unknown

ę1u

u1ű

u1u

Two radicals

u12u

1ę2

ę1u2

1u2u

1ű2u

Three radicals

1u23u

1a2â3

ę1î2u3

1a2u3

1â2ű3

 

For derived nouns

Singular

Dual

Septal

Indef. Plural

Def. Plural

One radical

i1

 

 

 

 

Two radicals

i12i

1i2

ę12i

1i2i

1i2i

Three radicals

i12i3

1i2ę3

ę12i3

1i2i3

1i2i3

 

There is no example of a one-consonantal root in the construct state in the Khuzdul corpus. We know one two-consonantal root in CS indefinite plural, dűmu, which provided the basis for the structure 1ű2u. Indefinite plural CS for three-consonantal roots 1a2u3 is attested in baruk khazâd, ‘axes of the Dwarves’ (note indefinite baruk for ‘axes’ and definite khazâd for ‘the Dwarves’).

Construct State pattern pl. for verb agentive and verb abstract forms: hence iglish, of hands, CS of aglash. There is no structure for one-radical derived nouns in all the plural forms, since the plural suffixes for those nouns are inflected as well; so there simply are no one-radical derived nouns in dual, septal, indefinite and definite plural.

 

1.7    Gender suffixes

 

Khuzdul employs no basic gender structure. All nouns are basically one category, although there is a derivative suffix used to denote the gender of living beings only. For example, if a root [T-D] denotes the basic idea of a ‘(domesticated) pig’, then we could assume a female form form tâda-a (>tadâ) to mean ‘sow’.

Since Khuzdul was invented by a male being (the god Aulë, who, in the first place, created only male Dwarves, by the way), it is obvious that the male gender should be the one employing the basic form of the noun; accordingly, the female suffix incorporates the indicative phoneme â, pointing away from the speaker (as opposed to î, for example, which would indicate something near to the speaker, if not the speaker himself; cf. French voicivoila, English this – that, German hier – da etc.). The suffix –a would probably come from Ancient Dwarvish –đâ (>–hâ >–a); hence use of –ha if employed after long vowels, after o, or after e (for –u-a becomes –ô, –i-a becomes –ę, and –a-a becomes â).

 

I have invented the consonant đ for Ancient Dwarvish as a predecessor for modern h; it would, in effect, be the Aulian equivalent of Valarin 3, yet Aulë knew the sound 3 as well and used it in his own name A3űlęz, from which it was adopted by the Dwarves into their name for their creator.

 

1.8    Definite articles

 

Nouns in singular, dual and septal are normally indefinite; there is, however, a definite prefix thu- (or th’- before vowels), although this is only rarely used, since definite nouns are mostly obvious from the context.

 

1.9    Pronouns

 

1.9.1          Free Pronouns

 

Most free personal pronouns are only employed as subject. They are placed directly before the verb.

The Khuzdul free personal pronouns are: 1st person singular khîl; 2nd person singular tęl; 3rd person singular âl; 1st person plural nôl; 2nd person plural męn; 3rd person definite plural khâd (‘those’, also in an indicative sense); 3rd person indefinite plural khad (‘they’); and 3rd person septal khęd (‘the seven’). The broken forms khad and khęd are sometimes replaced by the sound variation dukhad (= indefinite plural khad + definite article thu-) and thękhed (a combination of broken and sound derivation + the definite article th-, meaning, in contrast to khęd, ‘those seven’ with a certain indicative implication); but those constructs are only rarely used and considered obsolete or archaic by some.

 

1.9.2          Bound Pronouns

 

Bound pronouns have a special status in Khuzdul; their meaning depends entirely on their position. They are employed as suffixes only.

The Khuzdul bound personal pronouns are: 1st person singular –helu; 2nd person singular –tî; 3rd person singular –lę; 1st person plural –nâlu; 2nd person plural –męnu; 3rd person plural –hadu; and 3rd person septal –hędu.

Those suffixes express possessive meaning when attached to a substantive (e.g. męk-lę, ‘his hands’). They are also used as basic possessive particles if it is necessary to distinguish genitive from possessive forms in a sentence; in this case, they are attached to the genitive object as described above, but are followed by the possessive subject in the Construct State

When bound pronouns are attached to a verb or to a preposition, however, they act as objective (e.g. khazâd ai-męnu, ‘the Dwarves are upon you’).

 

I have modelled the pronoun structure very closely after the system of Old Hebrew, because it fits perfectly into the known corpus and the overall structure of Khuzdul as presented herein.

 

1.10Verbs as Nouns

 

Nouns can be derived from verbs in several forms. We know of one verb agentive pattern: a1 for one-radical roots; a12(a) for two-radical roots, and a1a2â3 or u12ad for three-radical roots.

For action nouns, the durative present is used to form a gerund which may act both in an adjectival and in a nominal sense; that is, u1a for one-radical roots, u1a2 for two-radical roots and 1u23(u) for three-radical roots; this is also employed as a verb abstract structure, whence gund[RM6] (u), ‘the process of delving’, but also ‘underground hall’, from the verbal root gunud [G-N-D], ‘to delve underground, excavate, tunnel’.

In addition, there is a verb instrumental structure used to denote the tool or object with which the action defined by the verbal root is carried out. This is 1e for one-radical roots, â1e2 for two-radical roots, and 1e2a3 for three-radical roots, whence felak, ‘a broad-bladed chisel’ from felek, ‘hew rock’ (which is a curious case, by the way, for felak has been re-adopted to verbs, as such meaning ‘to use such a tool’).

 

The verb agentive patterns a1a2â3 and u12a3 are a conjecture from azaghâl, warrior, and uzbad, ruler. I haven’t yet found any better explanation for uzbad; so I decided to stick with alternative verb agentives. Assuming that Aulë wasn’t that great a linguist.

 

1.11Adjectives as Nouns

 

1a23űn denotes, as Mr Fauskanger pointed out, a ‘person, thing or place characterized by the root meaning’. There is no hint in the corpus at what the corresponding structure for one- and two-consonantal roots might be, so I made up some patterns of my own: a1űn for one-consonantal roots (e.g. alűn, ‘river-man’, ‘Ulmo’), and 1î2űn for two-consonantal roots (e.g. bîlűn, ‘treasure-hold’ or ‘dear person’).

1a23űn can also be applied as a noun agentive pattern, as detailed in the Derivation section below.

 

 

Stem Modifications

We know three types of roots: one-, two- and three-radical roots. The declinations are different for each of those types (see below). – I have invented a new type of root relation for Khuzdul, working on the basis of a structure found in some Kuschitic dialects: in some cases of stems with one or two radicals, an intensive form can be derived by prefixing the final radical in a softened form. For example, if we assume a hypothetic root [K-T] meaning ‘fire’, we could derive an intensive form [D-K-T] (final T>D and added as intensive prefix) with the possible meaning of ‘sun’. Similarly, we could derive a root [B-Zh-P], ‘gap’ or ‘cleft’ from [Zh-P], ‘opening’, ‘slit’.

 

Some Analysis

Barazinbar: a compound of baraz [B-R-Z], red, and inbar [M-B-R], horn or mountain(!). The original singular form of [M-B-R] would probably have been mibar according to the pattern 1i2a3; but this was changed to imbar in ancient times due to the influence of Elvish imbar, ‘The Habitation’, *‘home’ (WJ:419), which the Dwarves of course equalled with ‘mountain’. The dissimilation of mb>nb, thus forming modern inbar, may be due to the ‘semi-adopted’ nature of the word (for the Dwarves were quite hesitant at changing their own language, as is well known).

Mazarbul: known from ‘the chamber of Mazarbul’, translated by Tolkien as ‘the chamber of Records’; however, I would not accept this as a literal translation, thus equalling mazarbul with ‘records’, due to the strange structure of the word. If, as Mr Fauskanger suggested, the last element be the suffix –ul (see below), we would be faced with a four-consonantal plural word in the unlikely form 1a2a34 or, as Mr Fauskanger put it, a past participle ma1a23 of [Z-R-B]. I, for my part, choose to render mazarbul as a compound of mazar (indefinite plural of [M-Z-R] meaning something like ‘tales’, ‘stories’) and bul (another indefinite plural, of [B-L], denoting a ‘treasure’ or ‘valued item’).

Nargűn: This is the adjectival root [N-R-G], ‘black’, in the agentive pattern 1a23űn (see above), meaning ‘Black Land’ (Mordor). Another example of the same pattern is tharkűn, ‘staff-man’ (Gandalf) from the radicals [Th-R-K], staff).

Tharkűn: cf. Nargűn

–ul: This is an adjectival ending applied to nouns that end in a consonant; we do not know what form would be applied if the suffix was used after a vowel, but I suggest *–dul (since –ul probably evolved from –űl, which in turn could be a worn-down form of *–đul; hence Tolkien’s spelling Khuzdűl in some places). In addition, -ul is used to form patronymics, as attested in Fundinul, ‘son of Fundin’. Obviously, the suffix is applied to the genetive object here (e.g. in patronymics to the father rather than to the son).

Gundabad: Gund(u), underground hall, + abad, sacred or holy. In this case, the adjective follows the noun it refers to; this is not so in other compounds like sharbhund, kibilnâla etc. The ‘correct’ form would be abad-gundu, ‘holy mountain’; but Gundabad is an archaic word, and the structures may have changed since that time. [Or else gund could in fact be the gerund structure. I haven’t quite sorted this out yet.]

Aglâb and Iglishmek: Verb agentive of [G-L], to speak: agl(a) (pattern a12(a), mentioned above). –ab, definite plural suffix for derived nouns; thus forming agla-ab>aglâb, the words, ‘spoken language’. –ish, indefinite plural suffix for derived nouns; whence aglash, words; Construct State thereof iglish, + męk, dual of [M-K], hands = iglishmęk, words of hands, ‘gesture-language’.

Sharbhund: Adjective shar, bald, + singular of archaic root [Bh-N-D], which eventually evolved to [B-N-D]. [The most obvious solution. I just couldn’t think of anything better.]

 

Derivation

1a23űn, noun agentive pattern; congruent with adjective agentive pattern (for one- and two-consonantal roots, see above). This is only valid for genuine nouns and adjectives. For derived nouns and adjectives, 1e2a3 is employed.

 

1e2a3 attested in felak (felak-gundu, ‘cave-hewer’).

 

Wordlist (by no means comprehensive)

Abad [B-D], adj sacred, holy

*Alűn [L], river-man, Ulmo; = [L] noun agentive

Aya/ai, prep upon, over

Azaghâl [Z-Gh-L], warrior; = [Z-Gh-L] verb agentive

 

Baraz [B-R-Z], adj red

*Bîlűn [B-L], treasure-hold, dear person; = [B-L] noun agentive

Bizar [B-Z-R], valley, vale

*Bizhap [B-Zh-P], rift, ravine; cf. zhup

Bâla [B-L], story

Bund [B-N-D], head

Burk [B-R-K], axe

Buzun [B-Z-N], adj cut in deep, located in a hollow or depression

 

*Dukt [D-K-T], sun; cf. kut

Dum [D-M], delving, (artificial) cave, mine

Dush [D-Sh], gloom, darkness

 

Ękhezd [Kh-Z-D], seven; cf. khuzd

 

Felak [F-L-K], 1) verb to use a tool like a broad-bladed chisel 2) a small axe; cf. felek

Felek [F-L-K], verb hew rock; cf. felak

 

Gabil [G-B-L], adj great, big

Gamil [G-M-L], old/wizened Dwarf

Gathol [G-Th-L], fortress

Gimil [G-M-L], adj old, experienced

Gund(u) [G-N-D], underground hall

 

Inbar [M-B-R], antiquated poetic form of mibar

 

Kheled [Kh-L-D], adj glassy, mirror-like

Khuld [Kh-L-D], glass

Khuzd [Kh-Z-D], Dwarf; cf. ękhezd

Khuzdul [Kh-Z-D], 1) adj dwarvish 2) Dwarvish (of language); = [Kh-Z-D] sg + adj ending –ul

Kibil, adj silver (of colour; adapted from Sindarin celeb); = zigil

*Kut [K-T], fire (symbolic); cf. dukt

 

Li [L], river, stream

 

Mahal [M-H-L], dual only creating/protecting Hands, later also Creator (Aulë)

Mazar, adj treasured, highly valued

Mazarbul [M-Z-R] [B-L], record (literally: treasured stories); = [M-Z-R] + [B-L] ind pl

Męnu, pron you acc pl

Mibar [M-B-R], horn, tip; = inbar

 

Narag [N-R-G], adj black

Nargűn [N-R-G], Black Land; = [N-R-G] adj agentive

 

Rukhs [R-Kh-S], ork

 

Sâsa [S-S], horse

Shar [Sh-R], adj bald

Sigin [S-G-N], adj long

 

Tharkűn [Th-R-K], staff-man; = [Th-R-K] noun agentive

Thirak [Th-R-K], magical or wandering staff

Turg [T-R-G], beard

 

Uzbad [Z] [B-D], ruler, lord; [Z-B-D] verb agentive

 

Zughul [Z-Gh-L], verb battle, wage war

*Zhup [Zh-P], cut, cleft, ravine; cf. Bizhap

Zigil [Z-G-L], adj silver (of colour); = kibil

Zirak [Z-R-K], tine, tip (of a mountain)


 [RM1]Th’-, bestimmter Artikel; also th’ussu uzbad = das Pferd des Königs (nicht irgendein Pferd des Königs)

 [RM2]Baraz, rot, Adjektiv, dem Nomen [S-S] im CS vorgestellt; also baraz ussu uzbad = das rote Pferd des Königs

 [RM3]CS von [S-S], Pferd; bezogen auf uzbad, also ussu uzbad = das Pferd des Königs

 [RM4]-lę = Possessivpartikel, gefolgt von [B-N-D], Kopf, im CS; also ussu uzbad-lę bundu (oder auch ussu-uzbad-lę bundu) = der Kopf des königlichen Pferdes.

 [RM5]Gabil, groß = Adjektiv als Nomen verwendet; und X(Subjekt) Y = X ist Y; also ussu uzbad-lę bundu gabil = der Kopf des königlichen Pferdes ist groß.

 [RM6]So take careful note of how you inflect! If the form 1u23 is a genuine noun, you inflect it by the standard tables listed above; if it is a gerund construction, however, use the ‚derived nouns‘ table. For example, CS of gund is not gundu but ignid!