Join the Blue Ribbon Anti-Censorship Campaign!

This document requires a Unicode-compatible browser. If you don't see the IPA characters, download the Thryomanes font and set it as the default font for Unicode documents, or as a user-defined font. Some versions of Microsoft Internet Explorer may assume that you have the Lucida Sans Unicode font on your system and display the IPA characters improperly if you don't have that font.

Czirehlat is a new "branch" of the Tirèlhat language, which separated in December 2000. Originally begun in the spring of 1999, Tirèlhat is one in a series of personal languages that includes Gjarrda and a number of less successful language experiments. In the course of developing Tirèlhat, I made frequent and sometimes radical revisions in the phonology and morphology. The lack of stability made it difficult to use and almost impossible to learn. The idea for Czirehlat was to create a stable "dialect" of Tirèlhat, more closely resembling an earlier, more pristine state of the language. Since "Tirèlhat" originally meant "wren language" (before the Gjarrda word "trril" was borrowed to mean "wren"), the new branch was named Czirehlat, "the goldfinch language". Although Czirehlat may undergo occasional changes, they will be more rare and less radical than they were in the development of Tirèlhat.

Spelling and phonology

The phonology and spelling of Czirehlat is now fixed. There will be no further changes to the segmental phonemes or their spelling, although minor changes in the pronunciation may still be made from time to time, and new diacritic marks may be added for various purposes.

Czirehlat uses a variation of the Latin alphabet, which lacks the letter "w". The digraphs "cg", "cz", "dz", "hl", "hr", "sz", and "tz" are treated as separate letters of the alphabet. The letters "c" and "h" don't appear as independent letters, but only as the initial component of digraphs. Czirehlat text is sorted according to normal Latin alphabetical order; the spelling has been carefully selected to allow this. With the exception of /ŋ/, each phoneme of the language has a unique spelling. Pronunciation is as follows:

spelling phonemic phonetic
a /a/ [a] when stressed; [ɐ] when unstressed.
b /b/ [b]
cg /ǰ/ [ʤ]
cz /č/ [ʧ]
d /d/ [d]
dz /ʣ/ [ʣ]
e /e/ [ɛ] when stressed; [ə] when unstressed.
f /f/ [f]
g /ɡ/ [ɡ]
hl /L/ [ɬ]
hr /R/ [r̥]
i /i/ [i] when stressed; [j] when unstressed adjacent to another vowel; [ɪ] when unstressed in other environments.
j /ʒ/ [ʒ]
k /k/ [k]
l /l/ [ɮ]
m /m/ [m]
n /n/ [n] (except before "g" or "k", where it sounds like "q")
o /o/ [ɔ] when stressed; [o] when unstressed.
p /p/ [p]
q /ŋ/ [ŋ] (written as "n" before "g" or "k")
r /r/ [ɾ] between vowels; [r] elsewhere.
s /s/ [s]
sz /ʃ/ [ʃ]
t /t/ [t]
tz /ʦ/ [ʦ]
u /u/ [u] when stressed; [w] when unstressed adjacent to another vowel; [ɯ] when unstressed in other environments.
v /v/ [v]
x /x/ [x]
y /ɣ/ [ɣ]
z /z/ [z]

Stress in Czirehlat, which is marked by an acute accent on the vowel, may be placed on any syllable of a word. Compound words may have more than one stressed syllable. In actual Czirehlat text, the stress is rarely marked, most frequently in names, in words that would otherwise be confused, and in a few special cases where stress marking is mandatory. Stress is always marked if it falls on the final letter of a word (as in iurá), or on a stressed suffix or infix (as in nejáka). Stress on the first syllable of a word of two or more syllables (as in jíra) is always unmarked in text. In other cases, stress is ordinarily unmarked, but may be marked for emphasis or resolving ambiguity.

The unstressed phonemes /i/ and /u/ may be non-syllabic when adjacent to another vowel in rapid speech, combining with the other vowel to form a diphthong. When unstressed /i/ and /u/ come together, the first of the two vowels is non-syllabic: /iu/ is pronounced as [jɯ], and /ui/ as [wɪ]. Other combinations of vowels, or sequences of more than two vowels, are considered to form separate syllables. When two identical vowels come together, either between words or in borrowed names such as Xaváii [xɐˈvajʔɪ] "Hawaii", a glottal stop is inserted between them. This glottal stop is not a separate phoneme in Czirehlat, since it is automatic.

Pronunciation examples: pa-lán [pɐˈɮan] "cloud", kuér-tiu [ˈkwɛrtjɯ] "typewriter keyboard", kiú-zi-ka [ˈkjuzɪkɐ] "dandelion", czé-lit [ˈtʃɛɮɪt] "sparrow", Ték-sas [ˈtɛksɐs] "Texas", Szi-ká-go [ʃɪˈkaɡo] "Chicago", O-xá-io [oˈxajo] "Ohio".


Grammatical suffixes and classifiers are added to a stem. In most cases the stem is the same as the dictionary form of a word. If the root ends in a consonant, -e may be added to form the stem, but this unstressed /e/ is often omitted if the resulting cluster is permissible.

líja [ˈɮiʒa] "to see"; líja-và-i [ˈɮiʒɐˌvaj] "want to see"
márga [ˈmarɡɐ] "starling"; márga-dài [ˈmarɡɐˌdaj] "flock of starlings"
fáz [ˈfaz] "snow"; fáz(-e)-rètzi [ˈfaz(ə)ˌrɛtsɪ] "so-called 'snow'"

Compounds are formed by adding the stem of the modifying word before the main word. The main word gets the primary stress, and the stressed syllable of the modifying word is converted to a secondary stress (written with a grave accent). In ordinary text, the secondary stress is unwritten. (Secondary stresses may be used in borrowed names to mark unreduced vowels.)

dév [ˈdɛv] "hare"; dèv-e-piák [ˌdɛvəˈpjak] "Leporidae"
vézi [ˈvɛzɪ] "orange (color)" + rín [ˈrin] "fish" = vèzi-rín [ˌvɛzɪˈɾin] "goldfish"
dzómi [ˈdzɔmɪ] "to spin" + xáhrsz [ˈxar̥ʃ] "storm" = dzòmi-xáhrsz [ˌdzɔmɪˈxar̥ʃ] "tornado"
szúj [ˈʃuʒ] "rain" + palán [pɐˈɮan] "cloud" = szùj-e-palán [ˌʃuʒəpɐˈɮan] "rain cloud"

Infixes are used to create new roots based on a modification of the basic concept expressed by the original word. The infix is added immediately after the stressed vowel or diphthong. If the infix itself is stressed, the preceding syllable loses its stress.

núri [ˈnuɾɪ] "blue"; nú-za-ri [ˈnuzaɾɪ] "azure".
tász [ˈtaʃ] "fly"; tá-vi-sz [ˈtavɪʃ] "gnat, midge".
néka [ˈnɛka] "polecat"; ne-já-ka [nəˈʒaka] "domestic ferret".
líja [ˈɮiʒa] "to see"; li-nái-ja [ɮɪˈnajʒa] "to watch".

Possessives are indicated by adding the unstressed particle le- (my), re- (your) or je- (his, her, its) before the possessed noun. Certain nouns (such as body parts and kinship words) can only be used with a possessive particle. For instance, my nose is lehríz and your nose is rehríz, but hríz by itself doesn't mean "nose". It can be used in compounds such as hrízdáxu "nasal consonant" or hrízbiúri szóbiyazár "red-nosed reindeer". To talk about noses in general, it is necessary to say "someone's nose", using the generic possessive particle xe-.

Representing names

Personal names and place names are represented as nearly as possible in Czirehlat pronunciation. As far as possible, personal names should be pronounced according to the way the name's owner pronounces it, and place names as people who live there most commonly pronounce it. Since in most cases this pronunciation is difficult or impossible to determine without a considerable amount of effort (e.g., touring the world to determine the popular local pronunciations of names), any reasonable approximation of the pronunciation will suffice as a temporary measure. This is true even in English for unfamiliar names; before the correct pronunciation of "Lech Wałęsa" was widely known in the English-speaking parts of the world, you might hear Anglicized versions like [ˈlɛk wəˈlɛnsə], but gradually a form more like [ˈlɛk vəˈwɛnsə] became more prevalent (still not perfect, but about as close as American English can get). At the very least, I hope not to make mistakes as glaring as [sæɹəˈheivou] for Sarajevo! But there will inevitably be errors, so the Czirehlat names of places and people may end up being revised as I find better pronunciations.

Some names familiar in English will be given a modern American English pronunciation (or as close as Czirehlat can get). William Shakespeare, for instance, comes out as Uíliem Széikspìr, regardless of how Shakespeare himself may have pronounced the name. Only an expert on 16th century English pronunciation would care much about the correct pronunciation in this case; it's more useful to have a pronunciation that's recognizable to modern speakers.

Scientific (or "Latin") names of animals and plants, and names of trademarked products, are left in their original spelling in the Roman alphabet. The usual conventions for scientific names are followed. If the Roman alphabet spelling of a trademarked name ever differs from one language to another, the English spelling is used, although accent marks are not omitted unnecessarily. Trademarks in other writing systems may be left in the original form, but a Romanized version should also be provided; not everyone will recognize 「豐田」 as "Toyota" or 「任天堂」 as "Nintendo", for instance.


The basic unmarked word order is subject-verb-object, but the object may be moved before the subject for emphasis.
Lé kajasar me hlav.
lé    kája-sàr me  hláv
I-SUB eat -can OBJ glass
"I can eat glass."

Me hlav lé kajasar.
me  hláv  lé    kája-sàr
OBJ glass I-SUB eat -can
"Glass I can eat."

Se jil hrastaxiz me lak.
se  jíl     hrásta-xì -z   me  lák
SUB red:fox jump  -HSY-IPF OBJ area:above
"The red fox jumps over something."
Adjectives and adverbs typically come before the words they modify, and prepositional phrases follow.
Se tiski marvi jil hrastaxiz me lak u hlivi jeg.
se  tíski márvi jíl hrásta-xì -z   me  lák  u   hlívi jég
SUB quick brown fox jump  -HSY-IPF OBJ over GEN lazy  dog
"The quick brown fox jumps over the lazy dog."


See the Czirehlat-English Dictionary.