By Craig Daniel

Edited by H.K. Fauskanger

[In the middle of December '01, Craig sent me this analysis of the Black Speech. While I don't necessarily agree with all the views here presented, I thought the analysis was certainly good enough to be published, and I asked Craig if I could place it on Ardalambion as a 'second opinion' on the Black Speech. He kindly granted me permission to use it.]

This is my own reconstruction of the Black Speech, from the same corpus as that of the Ardalambion article (http://www.uib.no/People/hnohf/orkish.htm) but with a different conclusion. While I am aware that Tolkien hated the Black Speech, I am also aware that it is his way of portraying Sauron as a more complex character - Tolkien himself was an avid conlanger, and yet the Black Speech, created by Sauron as an auxiliary language for his officers, is the only known conlang in all of Middle Earth.


Here is the corpus of all Black Speech quotations:

Ash nazg durbatulûk, ash nazg gimbatul,
ash nazg thrakatulûk agh burzum-ishi krimpatul One ring to rule them all, one ring to find them,
one ring to bring them all, and in the darkness bind them

Uglúk u bagronk sha pushdug Saruman-glob búbhosh skai!

This quote has two different translations. For reasons which will be discussed below, I will use this one: Uglúk to the dung-pit with stinking Saruman-filth, pig-guts, gah!


From this, I provide the following lexicon of Black Speech words, attested or implied by attested forms. The unattested forms are marked by asterisks. Ardalambion specifically marks the words of the 'debased Black Speech' spoken by the Orcs at Lugbûrz. I do not, on the grounds that I think they are more likely to speak with poor grammar than not be at least reasonably close to Sauron's standard. All those words in Ardalambion's wordlist which do not appear in either of the quotations are included here with their meaning; it has been to long since I have read the books for me to make any claim that the list is complete.

agh - and
*at - ending indicating 'intentive' form of verb. Here I differ from Ardalambion, as I do not think it reasonable to say that a language which would have to be simple (for the non-philologist Orcs) would use an infinitive suffix. It is much more likely that an auxlang would use the naked stem as the infinitive.
ash - one
*bag - dung
bagronk - dung-pit (probably bag-ronk)
*bûb - pig
bûbhosh - pig-guts (probably bûb-hosh)
burz - dark (In the place name Lugbûrz this is given a macron, but it is not in the Ring inscription. I will treat the inscription version as correct, as this will yield the correct tengwar for 'burzum-ishi' in TengScribe.)
burzum-ishi - in the darkness (probably burz-um ishi)
*durb - rule
*dug - ending indicating participle form of verb
*gimb - find
glob - filth
gûl - evil spirit serving Sauron
*hai - race
*hosh - guts
*ishi - in (postposition)
*krimp - bind
* - man
lug - tower
*olog - troll
olog-hai - Trolls
nazg - ring
nazgûl - Ringwraith
*push - stink
pushdug - stinking (probably push-dug)
*ronk - pit
sha - and (binding nouns rather than sentences, for which agh is used); with
*shar - old
sharkû - old man (probably shar-)
skai - [an interjection] (The 'cesspool' translation keeps 'skai' as the English form; the 'dung-pit' translation uses 'gah!'. The difference is probably in that 'gah!' will be understood by English speakers whereas 'skai' will not.) Ardalambion chooses to call it an expression of contempt, but I say that the opinions he is expressing are such that it is closer to disgust.
snaga - slave
*thrak - bring
u - to (preposition, but probably used erroneously as such and much more likely a postposition in the standard form of the tongue.)
*ûk - all
*ul - them (Though the rest of the Black Speech that is known does not mark for number, so it is probably a generic third-person pronoun.)
*um - a suffix similar to English -ness.
*uruk - Orc
uruk-hai - Orcs


Notes on the probable grammar of the Black Speech, in roughly the order I worked on the various aspects:

In the name Lugbûrz 'dark tower' the adjective follows the noun. However, in the Orc curse, the adjective precedes the noun - possibly an example of the 'debased' form of the Black Speech. However, there is a reasonable chance that a place name would be in a somewhat different, more poetic style. The 'adjective then noun' form is supported by such compounds as nazgûl (nazg-gûl, ring wraith). Thus we conclude that adjective before noun is the norm. I differ in this conclusion from Ardalambion, which ignores the case of nazgûl.

Postpositions are used in the Ring inscription, as in burzum-ishi 'in (the) darkness'. However, the Orc uses u as a preposition. This is probably the debased form rearing its ugly head, as *ishi-burzum would scan equally well in the poetry. This brings us to the conclusion that, had he been speaking standard Black Speech, the Orc would have begun with *'Uglûk bagronk u' instead of 'Uglûk u bagronk'.

No articles appear anywhere. This means that the Black Speech, like Russian, has no need to mark nouns for definiteness, nor does it need to mark them for number in many cases (which is not at all unreasonable; lojban for one doesn't do either).

The word 'nazgûl' seems to be used as both singular and plural; thus we see that in the only case in which we see both there is no plural marking. I will assume this is the norm in the Black Speech, as there are also no articles to mark number. Thus ul does not actually mean them but rather is a generic third-person pronoun. However, it is translated as them in the ring inscription, where its first appearance is marked with ûk, which specifically quantifies it as plural.


and 'ishi' present problems. All the other words except for names of races are monosyllabic unless they are compounds. They all break down neatly into CVC form. 'Sharkû' breaks into either *shar- or *shark-. I therefore conclude that means man, and shar means old. However, we are left with exactly one word which is bisyllabic and is not a compound. 'ishi' should break into *ish-i, but unfortunately the concept of one thing being in another does not break down. Thus, we are forced to wonder whether our analysis of 'ishi' is correct (but no other presents itself, so we should assume it is for lack of an alternative).


There are two translations of the Orc's curse. I will refer to them by the first content word, which is the translation of the compound 'bagronk'. The Ardalambion article on the Black Speech uses the 'cesspool' version; I will use 'dungpit' instead.

'Cesspool' offers an older translation, but there is really no good reason for choosing either. Since I enjoy the linguistic detective work of decrypting the Black Speech, 'dungpit' offers the following advantages:

Nobody has deciphered the Black Speech with the use of the 'dungpit' version before.

'Dungpit' offers a regular pattern of adjectives preceding nouns, 'cesspool' gives it the other way around - however, there is independent confirmation in the uses of ash nazg 'one ring' instead of *nazg ash, and nazgûl over *gûlnazg, both of which are known to be standard Sauron-style Black Speech.

Either 'pushdug' or 'bûbhosh' will be a compound word standing by itself. Therefore, the different meanings of the other will become useful. In 'cesspool', bûbhosh probably means 'great' while in 'dungpit' the word pushdug means 'stinking'. 'Dungpit' therefore offers a more complete grammar (telling us how to form participles) where 'cesspool' offers us an additional bisyllabic word which cannot be a compound.

For these reasons, I have chosen to differ from the Ardalambion reconstruction by using the 'dungpit' translation.


To conclude the analysis above, I offer the following rudimentary and incomplete grammar:

Naked verb stems are probably used to form the infinitive. There are two known verb suffixes. Attaching -at to a verb gives the intent (ash nazg durbatulûk thus means one ring whose purpose is to rule them all). This could possibly be simply a subjunctive form, with other uses as well. To form present participles, -dug is added. I will use this also as a progressive tense, as there is nothing to suggest that there is any form of 'to be' in the Black Speech (why is it 'Saruman-filth, pig-guts' rather that something which would mean 'Saruman filth, he is pig-guts'? The cesspool version has the same problem, but with push-dug instead.)

The pronoun ul is a generic third person pronoun. Nouns and pronouns have no number, so to indicate 'all' the suffix -ûk is used. Other suffixes with similar purpose are not given.

The Black Speech uses subject-verb-object word order, like English. This interpretation, with durbatulûk being durbat ul-ûk (where ul-ûk is the object of the durb) is much more likely than the idea that the object would be a part of the verb, even though it is written as if this were the case.

Postpositional phrases come before the verb ('burzum ishi thrakat ul', not 'thrakat ul burzum ishi') in the one attested case of them occurring together. This may be a feature of the Black Speech's grammar, or it may have been done to make the inscription rhyme. This may also be left up to personal style. I choose to treat it as a feature of the grammar, as there is no evidence to the contrary and no reason to assume that it would have to follow the English norms. When deriving grammars of unknown languages (and some would say in life), of course, one must always be wary of poetry.

Postpositional phrases are formed by putting the postposition after its object (as opposed to the prepositions of English). However, in colloquial Orkish speech, the postposition moves to before its object.

There is no article. Nouns are thus not inherently marked for number.


Let us now move to the sound system of the language.

The black speech has a CVC syllable structure.

We will assume for the moment that gh is one sound, rather than a g sound followed by an h sound (which would be unpronounceable finally).

The following consonants are attested: sh, d, r, b, th, k, m, p, t, l, k, gh, z, g, n, h, s.

We know that the l and r are pronounced at the back of the mouth (similar to English but without exception) - this is a sound which the Elves find extremely unpleasant.

Some consonant clusters occur; these are thr, kr, gl, sk initially, and zg, mb, mp, rz, nk finally. Medial consonant clusters usually result from compounding or affixing. When this would result in a double consonant (note that this includes such things as *ghgh) the sounds merge.

It is probably reasonable to assume that there exist voiced *dh and *zh sounds and an unvoiced *kh, as this is an auxlang and therefore likely possesses a completely regular phonology. *Dhl (a difficult sound to pronounce using a frontal or 'clear' l sound, but not with the more rearward 'dark' l - which may be the origin of this pronunciation of l and r in the black speech) and zg are probably therefore permitted initially, while *ls (and possibly *rs and *lz) and *ng are permitted finally *sk. If so, note that *ng would be like that of the English word 'finger' in that it would be /Ng/ rather than /N/.

There are five vowels, a, i, o, u, û. The vowel o is rare but not unheard of; e is absent. It is reasonable to assume that the Orcs, some of them having been Elves once upon a time, would find the Black Speech easier to learn if the vowels a, i, o were as in Quenya. However, to better distinguish the û and u letters, I would be willing to guess that the u is lax (think u in English put) at least a reasonable fraction of the time, while û is always long.

The other reasonable assumption is that u represents the standard [u] vowel, and û is a rounded front vowel (like German ü), which would be more likely to cause sharkû to be corrupted into 'Sharkey' as the nearest vowel in any other language in Middle Earth is i. [Editor's note: The Middle-earth language Sindarin actually does have a vowel similar to ü, normally spelt y, e.g. in yrch 'Orcs'. But 'Sharkey' must be a Westron form presented in Anglicized spelling, and Westron apparently did not have this vowel.]

It is not unreasonable to assume that, in rapid speech, vowels would merge as the consonants do. Thus, I would use *shar-kûk for 'all the old men' instead of shar--ûk.

Rules for stress are not given, as there is not enough Black Speech quoted in the books to warrant doing so. Therefore, this aspect is a mystery to us, but since this is an auxlang we can assume that whatever the rules are they are quite regular.

Let us double-check that this is a reasonable phonology by computing the number of permitted monosyllables (to make sure there are no permissible clusters which we are unaware of, and maybe solve the mystery of 'ishi')

There are 20 consonants which may be initial or final. In addition, there are 6 initial clusters and 10 final clusters. This yields 27 possible initials (words can begin with vowels) and 31 possible finals (They may also end with vowels). There are five vowels, plus two diphthongs (ai and au), for a total of seven. Since each root has one of each, this yields 27 x 6 x 31 possible words, which comes to 4522 possible monosyllabic root words, meaning that the syllables above account for only about one half of one percent of the possibilities. This is significantly more than enough, making ishi more of a mystery than ever.


While the words are somewhat run together in Tolkien's rendering of the Tengwar of the ring's inscription, I propose the following, more divided form instead:

Ash nazg durb-at ul-ûk, ash nazg gimb-at ul, ash nazg thrak-at ul-ûk, agh burz-um ishi krimp-at ul.

The hyphens have been used here to separate morpheme boundaries within words; verbs have been separated from their objects (i.e. durbatulûk -> durb-at ul-ûk). My examples below will use this hyphenation scheme, to convert to standard writing attach any pronouns in the object to the verb, remove all hyphens, and use a hyphen instead of a space to separate postpositions from the words they immediately follow. However, when using the Orc's colloquialism of using them as prepositions, I will separate them from the next word with a space.

Here is the Orkish curse, similarly divided into its constituent roots and with its grammar corrected back to the standard of Sauron and repunctuated to follow English conventions:

*Uglûk bag-ronk u sha push-dug Saruman glob, bub-hosh. Skai!


To test out whether this grammar is reasonable, I have put together a few sample sentences using only the vocabulary above. Sadly, while it is easy to vent one's feelings with this vocabulary, it is hard not to talk about vile things. Clearly there are many more words which do not appear in Lord of the Rings.

Uruk glob ishi krimp shar-kûk.
The Orc binds all the old men in the filth. Nazgûl ronk ishi gimb olog-hosh sha uruk-bag.
The Ringwraith finds troll guts and Orc dung in the pit.

Shar-gûl thrak-dug nazg.

The old wraith is bringing the ring.

Ardalambion Index