by Fredrik Ekman ( 2003

One of the most important sources of inspiration when Mike Singleton created his epic Lords of Midnight was J. R. R. Tolkien's The Lord of the Rings. This is particularly obvious in the first game in the series, but can be clearly seen throughout the three games.

In Tolkien's case, the driving motivation was the love of language. He just wanted to find out who were speaking the languages he invented, and thus he came to develop one of the most influential literary worlds ever.

For Singleton it was different. He did not start with the languages. In fact, most names of places and persons from the first two games are all too typical of the sort of thing you will find in most hack fantasy novels. There appears to be little or no conscious thought of phonology, morphology or etymology, or any of the other ologies which form the linguistic science.

Therefore it is all the more surprising and fascinating to find, in the novelette introduction to The Citadel, a few phrases of what appears to be a constructed language, the ancient tongue of the Wise. Not only is there a constructed language, but it seems to be elaborated at least to the level of being a fairly complex sketch, complete with some simple grammar rules and a vocabulary of some twenty words. Either that, or Singleton did a brilliant job of fooling me.

This article is an attempt to analyse the known fragments of the ancient tongue of the Wise (henceforth called "the language"). It is my hope that some day in the future we shall either find more phrases within the game itself, or be able to ask Singleton to reveal the details of the language.

Description and Fragments

In the third chapter of the introduction to The Citadel, we read the following: "He studied until his head ached and his eyes throbbed but, try as he might, he could not understand. Each new word seemed to have a dozen meanings, each part of speech a thousand rules." (Lords of Midnight Chronicles, p. 70). This is all that Singleton offers in the way of descriptions, but it is enough to convey an image of a very complex language.

The following is a summary of the know fragments.


Fragment 1: (p. 70)
"Ara darith uranar garak tha-ithil." is said to mean "I wish I could understand."


Fragment 2: (p. 70)
"Du-aran ara!" is translated as "I can speak it!"


Fragment 3: (p. 71)
"Idronel, ara b'ka e irin ur-anar!" means "Brother, I shall have sweethearts too!"


Fragment 4: (p. 72) & (p. 73)
"Garog ithar-harak!"  is a magical spell which is supposed to turn someone into stone, although the caster "missed a word out"


Fragment 5: (p. 86)
"Marish" is a geographical name which means "desolate land"


In addition to these confirmed fragments there are also a handful of geographical and personal names from the first two games that could be part of the language. It can be assumed that the language had not yet started to be developed at the time these were written, but it is possible that Singleton retrospectively worked them into the language. Names that are directly connected with the Wise include Rorthron, Gryfallon and Lorgrim (Lords of Midnight), plus Shareth and Kahangrorn (Doomdark's Revenge). These names have not been considered in the analysis. Indeed, they do have a different feel to them.


Phonology is the scientific study of sounds and sound combinations in language. In this case, it is actually quite impossible to make an accurate phonological analysis of the language, since we do not know how it is supposed to be pronounced. However, by assuming that there is a one-to-one correspondence between the written letters and the sounds they represent, some conclusions are possible. (In reality, this becomes a graphological analysis, or in other words an analysis of aspects of the language as it is written in the fragments.)

Five letters probably represent vowels, namely a (22 occurrences), i (8), u (3), e (2) and o (2). Other vowels may well exist and the five known ones may be pronounced differently in different contexts. There appear to exist no diphtongs.

There are a total of eleven letters and digraphs that appear to represent consonants, namely r (16), n (5), th (4), g (3), k (3), d (3), l (2), m (1), sh (1), h (1) and b (1). Again, these may not be the only ones in the language.

Thus we see that there is a marked predominance of a and r. These two taken together represent a staggering 49% of all letters in the fragments. Even though the repetition of the word ara (see below) raises the percentage somewhat, it can be assumed that this predominance leads to many similar-sounding words, which adds to the abovementioned complexity. Indeed, we can see from the fragments that there are many words that appear very alike.

In addition to the letters representing vowels and consonants, there is also an apostrophe in fragment 3. The significance of this is obscure. It may represent a sound of some sort, or it may be an indication that the word is a contraction, for example. It is interesting to note that it appears between two plosives (consonants produced with a sudden release of air), a combination which does not occur anywhere else.

Syllables seem to consist of a single vowel that can be preceded and/or followed by a consonant. This results in the following possible syllable structures (where C is a consonant and V is a vowel): V, CV, VC, CVC. The only seeming exception to this rule is the word b'ka (fragment 3), but then it should be remembered that we do not know what the apostrophe represents. Due to the limited data we have, nothing can be concluded regarding how syllables can be combined. It is, however, notable that there are no examples of a vowel following a vowel, except in a compound word such as tha-ithil.


Syntax is the scientific study of how words can be combined into phrases and sentences. The key to begin deciphering the syntax of the language is the word ara, which happens to appear in three of the five fragments. It must mean "I," which appears in all the corresponding translations.

Ara, which is in all three cases the subject of the phrase, is positioned first in the phrase in two instances and last in one. This is a deviation from most natural languages, where the subject has a fixed position in the phrase; first, last or in the middle. There could be several reasons for this. One is that the language has a free syntactic structure, where word form rather than word order decides the meaning of the phrase. Or it could be that fragment 2 is an exception to a rule that says that the subject is usually first. Finally, it is possible that word order varies depending on, for instance, temporal or modal aspects of the verb.

Each of the four complete sentences has a single compound word. One possible hypothesis is that this somehow represents the main verb. One of the parts would then be the verb itself and the other (probably the first, which is usually shorter) is some sort of grammatical marker, possibly denoting tense. Since all four sentences are written in varying tense, there is no way to confirm this.

An argument against this hypothesis is that fragment 1 in fact consists of two verb phrases, but there is only one compound word. Another option, therefore, is that the compounds represent a combination of auxiliary verb and main verb. Fragments 1 through 3 contain auxiliary verbs in the English translation and fragment 4 is not properly translated, so this remains a possibility. Fragment 2 further supports this hypothesis, since it only contains two words whereas the translation has four. The object ("it") can possibly be understood from the context, but the auxiliary verb (or a comparable grammatical construct) is needed at some position.

It is also possible that the word uranar in fragment 1 is in fact identical with ur-anar (fragment 3), in which case the hyphen has accidentally been left out in the first instance. That could mean that uranar represents the verb in the main phrase ("I have a wish"), while tha-ithil would be the verb in the subordinate phrase ("that I could understand"). In that case, garak may be another form of ara with the more specific meaning "that I."

Except for fragment 2, the above hypotheses suggest that the basic rule is an SOV sentence structure, meaning that the subject comes first, followed by the object and finally the verb. (English, along with most other European languages, has an SVO structure.) But fragment 2 also reminds us that, even if this is correct, there are exceptions.

Fragment 3 contains a word in plural form ("sweethearts" in the translation). Unless the plural is optional or non-existent in the language, there is hidden somewhere in that sentence the grammatical marker for plural. I can see three main suspects: First, the b in the word b'ka. Second, the word e, in which case either b'ka or irin is the singular for "sweetheart". Third, the word irin could be the plural for "sweetheart," in which case some morpheme (the smallest meaningful part of a word) denote the plural, for example the final -n. (Note for terminology freaks: This paragraph borders on the domain of morphology; the structure within words.)


The semantics of a language is the meanings of its words and phrases. In this case, we know more or less what all the fragments mean, but since our knowledge of the syntax is so scarce, we must guess how most words in the fragments and the translations correspond syntactically. In this section, I will concentrate on those words where we can be relatively certain about the meaning.

I have already mentioned ara above, which must almost certainly mean "I." The final syntactic analysis of fragment 1 above further gives darith as the noun "wish," uranar (or ur-anar) as a form of "to have" and tha-ithil as a form of "to understand." If that is so, then anar and ithil could be the common cases of those respective verbs.

Assuming that ara is "I" leaves the single compound word du-aran in fragment 2, which therefore means "can speak it." I have already speculated that the object ("it") is perhaps understood, in which case du-aran actually means "can speak." Aran could then be the common case of "to speak."

In the translation of fragment 3 we can see that the word "brother" is not actually part of the phrase that follows it. The corresponding word in the fragment should therefore be either first or last. Since the first word is followed by a comma (just as "brother" is), I conclude that idronel means "brother."

If fragment 4 means something like "turn to stone" or "become stone" then one of the words must be the word for stone. If the hypothesis about SOV word order holds, then that must be the first word in the phrase. It follows that garog is possibly the word for "stone." It should be remembered, however, that one word was accidentally left out of the spell, but it was partially effective (the target froze for a second), so that one word was probably not the word for "stone."

The meaning of Marish is known, but since it is probably a compound it would also be interesting to know the meaning of the components. There are only two meaningful ways to divide Marish, either Ma-rish or Mar-ish. But then we do not know the normal word order for adjective-noun compounds in the language. The result is a total of four alternatives (two divisions and two word orders) for the words for "desolate" and "land." And this assumes that none of the original words have been mutated in the compounding process.

It is easy to see that we know quite as little about semantics as about any other aspect of the language. The meanings of only two words, ara and idronel, can be assumed with any degree of certainty. In order to progress further, we must either find many more fragments or achieve a better understanding of the syntax and morphology.


The following sums up what we currently know (or think we know) about The Ancient Tongue of the Wise:

  • It is a constructed language that was developed by Mike Singleton for the fantasy world in the Midnight series of computer games.
  • It is a highly complex language.
  • Five fragments of the language are currently known.
  • The phonology (or at least graphology) has a considerable predominance of a and r.
  • Syllables appear to be usually of the forms V, CV, VC or CVC. Two vowels probably cannot follow each other, except in compound words.
  • Verbs often appear to consist of two components, where one component is the verb itself and the other may be some kind of grammatical marker or auxiliary verb.
  • Normal word order may be SOV, although there are exceptions.
  • The meanings of ara ("I"), idronel ("brother") and Marish ("desolate land") are as good as known, but other words can, at best, be guessed. However, the meanings of all the fragments are known on a phrase level.

If there is anything you want to ask about or add to this article, please contact me! My e-mail address is


The author wishes to thank Garth Wallace and, in particular, Roger Mills for reading and commenting upon this article. Without their help and advice, the syntax section in particular would have been much less interesting.


There are many good sites about Tolkien's constructed languages. My favourite is Ardalambion.

A good site about language construction in general is Richard Kennaway's Constructed Languages List.


Tower of the Moon