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 voici – voila,
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!