Ükäntel

Consonants

 * Dental consonants are laminal, with /θ/ being interdental in most environments.
 * Alveolar consonants are apical.

Allophony and Assimilation

 * Obstruents become voiced under a few conditions:
 * following a nasal consonant;
 * when intervocalic and not juxtaposed with a stressed syllable.
 * The nasal consonants /n ŋ/ are indistinct in a consonant cluster, assimilating to:
 * [n̪] before or after a dental consonant, while the sequence /nl̪/ is [l̪:];
 * [n] before or after an alveolar, the sequence /nɾ/ is [r], the sequence /ɾn/ is [n:];
 * [ɲ] before or after /ʃ/, however, the sequence /nj/ remains so;
 * [ŋ] before or after a velar consonant or /h/.
 * Short /r/ is tapped while geminate /r/ is trilled.
 * When /t͡s/ appears geminate, the stop component is lengthened.
 * The distinction between /t͡s/ and /s/ is neutralized in certain environments, merging to:
 * [s] before a plosive or /n/, as well as before /ɾ/
 * [t͡s] following /n/.
 * Dental consonants and alveolar consonants most often assimilate in place to each other regressively.
 * Regressive POA assimilation occurs when both consonants of a cluster come from the set of consonants /n t̪ s l̪ ɾ/, or in the case of /nθ/, and except /l̪ɾ ɾl̪ nɾ ɾn/.
 * The clusters /l̪ɾ ɾl̪/ are realized as [r l̪:].
 * The clusters /nɾ ɾn/ are realized as [r n:]
 * The cluster /θn/ is realized as [θn̪].
 * /t͡s/ forces a preceding or following dental consonant to become alveolar, notably with /θ/ becoming [s].
 * /θ/ does not assimilate except to /t͡s/ and forces a following alveolar consonant to be laminal in articulation with /ɾ/ remaining apical. It may be important to note that /sθ θs/ remain the same.
 * /h/ is subject to significant allophony and assimilation:
 * /h/ is realized as:
 * [h] at the beginning of a word, following a back vowel and before a voiceless consonant other than /k/, and following another consonant;
 * [ɦ] intervocalically and in between a vowel and a sonorant consonant;
 * [ç] word-finally after a front vowel and following a front vowel and before a voiceless consonant other than /k/;
 * [x] word-finally after a back vowel and always before /k/;
 * [x:] when geminate regardless of vocalic environment.
 * /h/ will assimilate to a fricative that precedes it, resulting in a geminate fricative, i.e. /sh/ is realized as [s:]. This does not happen following an affricate.

Allophony

 * /i y u/ are realized as [ɪ ʏ ʊ] before sonorant consonants.
 * /ɤ ɤ:/ are frequently centralized to [ɘ ɘ:] or even [ə ə:] especially when unstressed or when comprising the closing element of a diphthong.

Assimilation
Due to the high rate of inflection, vowels are frequently juxtaposed next to one another, and many combinations of vowels assimilate to another diphthong or to a long vowel because of historical sound changes. These combinations are detailed in the table below according to Ukantel orthography. Blank squares indicate the combination does not exist.

Phonotactics
Ukantel syllables minimally consist of only a vowel and maximally are CVC, with V being any short vowel, long vowel, or diphthong. The syllable structure can thus be transcribed as (C)V(C). Outside of nasal neutralization, any consonant can appear word-initially or word-finally, or as the coda or onset of syllables not at word boundaries. Consonant clusters thus are not allowed to be longer than two segments; anaptyxis is employed to avoid this. Certain clusters are subject to assimilation, which is detailed in the above consonant allophony section. More than two vowels in a row are separated by hiatus and do not assimilate furthermore, with every grouping of two vowels being subject to assimilation.

The syllable structure is obeyed only at the word level, meaning that roots and affixes can both take apparently non-legal forms, including roots ending with a consonant cluster or geminate consonant or affixes taking the form of a sole consonant cluster or consonant.

Phonemic Length and Diphthongs
Length is distinctive on all phonemes. Generally, geminate consonants and long vowels are analyzed as sequences of the same phonemes. Double consonants always fall across a syllable boundary; double vowels always belong to the same syllable coda. Despite this, long vowels do not occur in roots or in affixes and always result from juxtaposition of vowels due to inflection; geminate consonants do occur in roots, as well as in affixes. Due to word-level anaptyxis, however, geminate consonants never form part a cluster. Diphthongs pattern and function the same way as long vowels sylabically and phonetically; in other words, they are inherently long and posess only one length distinction. However, diphthongs do occur independently in roots in addition to appearing from vowel juxtaposition.

Stress and Prosody
Stress operates at the word level and is weakly expressed in Ukantel. Stress always falls on the first syllable of every word. Clitics, which include pronouns, demonstratives, and particles, are not stressed. The realization of stress and vowel length do interact, although stressed long vowels are rare and restricted to monosyllabic roots, while stressed diphthongs are considerably more common. Vowels are slightly longer when stressed, approximately 50% longer than a short vowel. It can be said, then, that Ukantel has four vowel lengths: (ranging from shortest to longest) short unstressed, short stressed, long unstressed, and long stressed, which can be assigned comparative length values respectively of 1, 1.5, 2, and 2.5. In addition to stress being indicated by length, it is indicated by a slight upstep in pitch.

At the phrase level, Ukantel sentences have a falling intonation.

Orthography
Ukantel is written using a modified variant of the Latin alphabet. There are no digraphs in the language, instead a diacritical system exists, making use of the cedilla to mark (historically) palatalized alveolar consonants and the diaeresis to mark additional vowel sounds.

Clitics, which include pronouns, demonstratives, and particles are written as separate words and therefore are not subject to orthographic representation despite their phonological dependence on a host.

Alphabet
Making note of the placements of y, g, and d, Ukantel alphabetical ordering is observably different than the average Latin alphabet. These deviations exist because it was decided that letters with similar pronunciations should be near each other in the alphabet, based off the idea that accented letters were to follow their unaccented base forms.

Letter names
The vowel letters are named after the vowels themselves. Consonant letters are named are named after the consonant plus e, with an exception in h (named ha).
 * a, ä, ce, e, ë, ha, i, ye, ke, le, ne, ge, o, ö, re, se, şe, te, de, u, ü, we

Orthographic notes

 * Long vowels and geminate consonants are written with doubled graphemes.
 *  frequently alternates with   due to nasal assimilation in clusters. Many verbal infinitives with stems that appear to end in phonetic [n] actually end with phonemic /ŋ/ and are spelled accordingly with  . When these stems are inflected so that the nasal would appear intervocalically, the spelling is altered to reflect the pronunciation.


 * Vowel assimilation, detailed in an above table under vowel allophony, is indicated in the orthography.

Vowel Harmony
With a few exceptions, native Ukantel roots contain either only front vowels /y e ø æ i/ or only back vowels /u ɤ o ɑ/. /i/ does not have a back variant and patterns largely as a front vowel. In back vowel words /i/ is transparent, i.e. it does not change the harmony to front. Roots that contain only /i/ take front vowel suffixes. Generally, archiphonemes are used in grammatical writings, being transcribed /U E O A/ and representing pairs /y~u e~ɤ ø~o æ~ɑ/. There is no harmonization according to roundedness and thus all Ukantel grammatical suffixes have two variants, although suffixes containing only /i/ have only a single variant. It is also worth noting here that some suffixes consist only of one or two consonants, with vowels being inserted epenthetically if necessary according to harmony.

In native Ukantel roots, vowel harmony is rarely violated outside of a few roots, such as ayäceć, the Ukantel word for alphabet. Vowel harmony is not applied when compounding if roots contain differing sets of vowels. Loanwords, especially newer loans, do not necessarily adhere to vowel harmony.

Anaptyxis
Anaptyxis, or epenthesis of a vowel, is necessary with certain grammatical suffixes as the stringing together of certain morphemes would violate the Ukantel syllable structure. A (C)V(C) syllable structure is obeyed at the word level, meaning consonant clusters can only be two segments long and must appear intervocalically, and that all syllables must have a vocalic nucleus. A few suffixes have already two forms outside of epenthesis in order to avoid consonant strings, such as the dative case suffix, which appears as -t following a vowel and as -At following a consonant. Where consonant clusters of than would may appear due to inflextion, an epenthetic vowel -E is inserted to avoid that. The epenthetic vowel follows a string of two consonants that would otherwise appear before another consonant, and may even be inserted within a suffix.

Nouns
Nouns are manditorily inflected for number and case and optionally for possession. The below table outlines the detail of the nominal structure.

Stem
The stem includes nominal roots and any possible derivational suffixes attached thereto and is essentially the barest form of a noun found in regular speech, as well as being the dictionary citation form. The bare stem is identical to that of the nominative singular, and not always to the nominal root.

Number
Ukantel distinguishes between singular and plural. The singular is unmarked and the plural is formed with the suffix -k.

There are homophones where a singular noun may appear to be plural in form (or in fact the plural form of a separate noun) due to ending in -k.

Possession
Possession is indicated by a suffix in Ukantel. The head in a genitive phrase is marked with a possessive suffix, agreeing with the possessor in person and number, which is sometimes called a construct state due to the influence of Semitic linguistics. Concerning the two-variant singular possessive suffixes, the first appears following vowels and the second following consonants.

Case
Ukantel nouns decline for 8 cases, the suffixes of which are listed in the table below; the nominative is unmarked. The nominative case is used to mark the subject of a sentence, more specifically, the agent of a transitive verb and the patient (also called the experiencer) of an intransitive verb.

The accusative case is used to mark the patient (also called the object) of a transitive verb. When making a transitive verb causative, both the causee (the subject of the non-causal clause) and the original object appear in the accusative case, with the causee appearing second.

The dative case is used to mark the recipient of an object, the object of verbs implying transfer or perception, or the beneficiary of an action. -At is used following stems or a preceding morpheme that ends in a consonant and -t following those ending in a vowel.

The genitive case is used to mark the possessor of another noun, a noun which modifies another noun, description, geographic origin, relation, and composition.

The instrumental case is used to mark the means by which an action is done or the instrument used to accomplish the action, whether physical or abstract. It also marks the logical agent of a passive clause.

The comitative case is used to mark with whose company an action is done (i.e. accompaniment with the agent), adornment or posession with the copular verb tehe.

The temporal case is used to mark when an action is done and can be applied to numbers to mean at ___ o'clock, or to noun phrases to mean on _____.

The durative case is used to specify the length of an action or during what the action took place, meaning effectively for x ____ (where x is a numeral) or during _____ or throughout ______.

Personal pronouns
As in most languages, personal pronouns in Ukantel are somewhat irregular. Ukantel is a pro-drop language, meaning subject pronouns are omitted unless for emphasis. Similarly, genitive pronouns are usually omitted in a possessor-possessed structure unless for emphasis. The third-person singular pronouns are only used for human referents, whereas the demonstrative pronouns are used for any other third-person referents. This distinction is not realized with other third-person indicators (i.e. the possessive suffix or verbal agreement), as they are used to refer to all third-person referents. Compare the following two sentences:


 * Ege yën!
 * DEM.DIST-ACC eat-2S
 * Eat it/that!


 * Awa yën!
 * 3S-ACC eat-2S
 * Eat him/her!

The accusative case pronouns are used as reflexive pronouns. The dative pronouns are used similarly for verbs that require a dative object.


 * Şeh aşuhtoldëş.
 * 1S.ACC stand-CAUS-NEG-PFV-1S
 * I didn't stand up.

The genitive pronouns can be used either attributively, although they are usually omitted in this usage unless for emphasis, or predicatively.

Demonstrative pronouns
Ukantel makes a two-way proximity distinction in its demonstratives. Attributively or predicatively they do not decline and simply precede the noun as adjectives, while as pronouns, they decline as nouns. The proximal demonstrative is regular while the distal demonstrative bears irregular forms with its stem eg- which appears as en- in isolation and when followed by a consonant.

Interrogative pronouns
Interrogative words and phrases are formed regularly from a few interrogative roots either on their own, declined across cases, or with modifiers or particles. Ka 'what' is the most common interrogative root used in forming interrogative compounds.

Modifiers
Modifiers in Ukantel can be separated into three categories:
 * 1) Numerals
 * 2) Determiners
 * 3) Adwords
 * 4) Adjectives
 * 5) Adverbs

Adwords are further divided into 'adjectives' and 'adverbs' on the basis that there exist certain adwords that cannot be used in conjunction with nouns, and certain adwords that logically cannot be used to modify a verb, although the two syntactic functions are largely treated as one in the same. In most instances, adwords can modify either nouns or verbs with no restriction or change in form.

Numerals and determiners are distinct from adwords because they cannot modify verbs alone, with numerals furthermore being distinct from quantifiers because they take a larger amount of derivational affixes, detailed under the 'numerals' section. Determiners are a type of pronoun that can be used to modify.

Any modifier precedes that which it modifies. Modifiers do not agree with nouns they modify on basis of number or case and simply precede their head. Most modifiers can be used nominally, in which case they are declined as and behave syntactically as a noun, usually with the semantic meaning of someone or something possessing whichever quality (sat öylete 'with the tall man' sattë 'with the tall one, person').

Adwords
Adwords have synthetic comparative and superlative forms that can be used attributively and predicatively, and that are obligatory in structures indicating a basis of comparison. Duplicating an adjective indicates excess. The bare form of an adjective, and the citation form, is its positive form. A few examples of these forms are given below the table.


 * En şëlu acu ewäs uyuc teiniş.
 * DIST.DEM woman as polite sure be-PST-1S
 * I was surely as polite as that woman.


 * Hi ënal haşëk şi acu untë te.
 * so many friend-PL 1S.NOM as 2S.COM be
 * You have as many friends as me.


 * Sat öyle satëcë astuin.
 * tall man tall-COMP-ACC see-PST
 * The tall man saw a taller man.


 * Senire şukirë acu aslac te.
 * talking fighting than easy-COMP be
 * Talking is easier than fighting.


 * Ënalëc süälek şi acu ëriştë te.
 * much-COMP horse-PL 1SG.NOM than father-1SG-COM be
 * My father has more horses than me.


 * Hayrah harë kicäwöt üle etülin.
 * young-SUP girl face-3S-DAT on fall-PST
 * The youngest girl fell on her face.


 * Cöliş salën hayrah te.
 * sister-1S three-GEN young-SUP be
 * My sister is the youngest of three.


 * En häce hül hül tein.
 * DEM.DIST dog skinny skinny be-PST
 * That dog was too skinny.

Verbal morphology
Verbs in Ukantel are morphologically regular, bearing no suppletive forms. They take a multitude of suffixes to indicate finite grammatical information and suffixes that construct non-finite forms.

Finite Verbal Morphology
Finite verbs serve as the main verb of a clause. Suffixes attach to finite verbs in a fixed order, illustrated in the table below.

Stem
The stem of a verb in Ukantel is the bare form of the verb and is a bound morpheme.

Voice
The active voice in Ukantel is unmarked. More than one voice suffix may be applied to the verb, however, doubly causative verbs are not possible morphologically and instead use a periphrastic construction. The semantics of the voice suffixes are not always in line with their defined meanings and have unpredictable meanings with certain verb roots.

Polarity
Negation is indicated with the suffix -l.

Tense
Morphologically, Ukantel distinguishes three tenses, of which the present tense is unmarked. Compound tenses are expressed periphrastically.

Aspect
Ukantel distinguishes between two aspects morphologically. Other aspectual meanings are conveyed using auxiliary verbs.

Mood
The admonitive mode is generally used to indicate warnings and translates as "should", "must", "have to".


 * Hal öylet yukuratiş.
 * DEM.PROX man-DAT listen-ADM-1S
 * I should listen to this man.

The permissive mode indicates that the action is allowed.


 * Sänke etültölegiş.
 * tree-PL-ACC fall-CAUS-NEG-PRM-1S
 * I am not allowed to fell trees.

The abilitative mode indicates that the action is able to be performed.


 * Ükäntele senhöş.
 * Ukantel-ACC speak-ABIL-1S
 * I can speak Ukantel.

The potential mode indicates that the action is likely.


 * En häce unya sokdëwër.
 * that dog 2S.ACC bite-PRF-POT
 * That dog could have bitten you.

The desiderative mode indicates that the action is wanted.


 * Körhü unkat astuşëş.
 * night-TMP sky-DAT see-DES-1S
 * I want to see the sky at night.

The imperative-hortative mood (usually referred to solely as the imperative and likewise glossed as IMP) expresses command, request, suggestion, exhortation, and urging. The imperative cannot take tense or aspect suffixes. It can only be preceded by voice suffixes or the negative suffix, and followed by person suffixes, including -w, -Aw to indicate a 3rd person imperative. First and third person imperatives are cross-linguistically referred to as the hortative or jussive moods respectively, and those terms are akin to their semantic meanings in Ukantel.

First person imperatives have a hortative sense, whether singular or plural, suggesting that either the speaker or the speaker and addressee(s) perform the action. Second person imperatives have a general sense of command or request. Third person imperatives have a jussive sense, suggesting that a third party perform the action or be permitted to perform the action. Translations of first-person singular imperatives are often awkward when translated to English. Examples are given below.

Because it is not overtly marked, all imperatives sans 3rd person forms are phonologically identical to present imperfect forms of the verb.


 * Halë şöt sanun.
 * DEM.PROX-ACC 1S.DAT give-2S
 * Give this to me.


 * Şöt yukurun.
 * 1S.DAT listen-2P
 * Listen to me (spoken to multiple people)

Non-finite Verbal Morphology
Non-finite verb forms in Ukantel are numerous and common in the language. This includes the infinitive, gerunds, and participles. and converbs.

Infinitive
The infinitive form of a verb ends in -hE. The infinitive is the citation form of a verb. Some verb stems may appear to end in a vowel from their citation form due to an anaptyctic -E because the stem ends in a consonant cluster or geminate consonant; no distinction is made between these and actual -E-final stems. The infinitive in Ukantel is found in constructions where one finite verb is syntactically dependent on another and not a constituent thereof. Infinitives cannot take their own arguments, nor can they take personal suffixes. The infinitive suffix essentially terminates the verb-form.

Gerund
The gerund in Ukantel is the nominal form of a verb that is not entirely deverbal, meaning it can take the same morphological suffixes and arguments that a finite verb would. Gerunds can take personal suffixes and case suffixes as well. Entirely deverbal nouns are listed under derivational morphology. The gerund is marked with -irE.


 * Urunkawot yukurirë unut inyür.
 * parent-PL-2S-DAT listen-GER 2S.DAT help-FUT
 * Listening to your parents will help you.

Participles
Participles in Ukantel function as verbal adjectives, meaning they serve to modify a noun. They use the same endings as normal finite verbs, with the addition of an ending for a present tense participle, but nothing marking imperfective aspect, both of which are unmarked in finite verb forms, and do not take personal affixes. Third-person singular verb forms are often indistinct from a similarly-constructed participle because of this. To show nominative arguments in a participial phrase, nominative-case pronouns are used. In this manner, as they are not restricted in what tenses, aspects, modalities, etc, they can appear in, participles often take the place of relative clauses. The present tense ending, which is only used in participial constructions, is -sE.

As participles, verbs govern the same cases as they would if used finitely, meaning some participles make take dative objects.


 * Öttese öyle ayëşda senä tin.
 * sit-PRS.PTCP man mom-1S-COM talk-CONV be-PST
 * The sitting man was talking with my mom.


 * Oştasë röne etülined.
 * burn-PRS.PTCP building fall-PST-PFV
 * The burning building fell.


 * Nohawot är sanëtë ertäyse şi astuinëd häce sakaşot är heşä te.
 * mouth-3S-DAT in meat-ACC hold-PRS 1S.NOM see-PST.PFV dog back-1S-DAT in run-CONV be
 * The dog that I saw holding meat in its mouth is running after me.

A construction meaning "when" is constructed using a participle before the temporal case of örüs "time".


 * Şi latinë esälätöin örüshü kininätün.
 * 1S.NOM bowl-2S-ACC be.full-CAUS-PST time-TEMP ask-PST-ADM-2S
 * You should have asked when I filled your bowl.

Converbs
Converbs are verb forms that are dependent on another verb through some form of conjunction. Converbial suffixes terminate the verb form in a similar fashion to the infinitive suffix; converbs therefore cannot take personal endings nor can they take case endings. Because of this, converbial clauses are usually constructed in Ukantel as having the same syntactic subject as the main clause, which is indicated by placement of the nominative argument in the dependent clause, by personal agreement on the verb, or both especially in cases of emphasis. However, converbs can have their own syntactic constituents separate from that of the main clause; subordinating converbs, as well, can have a differing subject from that of the main clause (the first three suffixes listed below [-A, -OcE, -nU] do not belong in this category and must have the same subject as the main clause).

-A describes a repeated action contemporaneous with that of the main verb. This is also used to construct the progressive aspect in conjunction with the verb tehe.


 * Etülä en şëlu önin.
 * fall-CONV DEM.DIST woman walk-PST
 * That woman stumbled (i.e. walked falling).


 * Yenehü ölöcä teyüriş.
 * ten-TEMP leave-CONV be-FUT-1S
 * I will be leaving at 10.

-OcE describes a single instance of or a sustained action slightly prior to or contemporaneous with that of the main verb. This often corresponds to the English "by doing" or "with doing."


 * Eş köylük häcewe eytäyöce awot çinindehöşäl.
 * one night-DUR dog-3S-ACC take-CONV 3S.DAT help-PST-PFV-ABIL-1P
 * We were able to help him by taking his dog for a night.

-nU describes an action that happens entirely before that of the main verb and often has the connotation that the main verb follows this action. It translates into English using "and then" or "after." Duplicating the converb itself indicates continuity and dissatisfaction along with the main verb negated.


 * Awa aşuhtonu ölöcin.
 * 3S.ACC stand-CAUS-CONV leave-PST
 * He stood up and left.


 * Yënu wisëwë yëtoyur.
 * eat-CONV child-3S-ACC eat-CAUS-FUT
 * After eating he will feed his child.


 * Yukurnu yukurnu senirewäle nesliniş
 * listen-CONV listen-CONV talk-GER-3P-ACC hear-NEG-PST-1S
 * I kept listening but I did not hear their conversation.

-tE marks a clause as the condition under which the main action happens. It translates into English using "when", "if", or "in case."


 * Häcen yëhutoltë unya sokyurwër.
 * dog-2S eat-PASS-CAUS-NEG-CONV 2S.ACC bite-FUT-POT
 * If your dog is not fed, it may bite you.


 * Hököt ugutë sukunë ayuntoyurun.
 * sun-DAT look-CONV eye-PL-2S-ACC hurt-CAUS-FUT-2S
 * If you look at the sun, you will hurt your eyes.

-silA marks a clause as concessive, meaning that the action of the main verb happens in spite of it. It is usually translated into English as "although."


 * Ugusila sänköt sed ayëşë astulinhoş.
 * look-CONV tree-PL-DAT among mother-1S-ACC see-NEG-PST-ABIL-1S
 * Although I looked, I could not see my mother among the trees.

Word Formation
As a heavily inflected language, Ukantel regularly utilizes compounding and derivation to express concepts where a lexical gap exists.

Derivation
Derivation in Ukantel, wholly considered, is productive and utilized often in the language. There are a large array of suffixes that either redefine stems in terms of another word class, or further specify stems in their own. Derivation is largely regular, although a few irregularities exist, especially among verbalized forms.

Verbalization
Verbalization is no longer productive in modern Ukantel, as verbs are a closed class. However, many methods of verbalization are fossilized, and as a result derived forms are rarely opaque. While forms are often similar in meaning, the meaning of a derived form may not be obvious. In this section, verbs are listed with the infinitive suffix.

A small amount of verbs exist through zero-derivation, all of which are intransitive, and many of which have a translative sense:


 * sathë to grow, from sat 'tall'
 * tënhë to age, from tën 'old'
 * ökkehe to dry (intr.), from ök dry
 * hëdhë to bleed, from hëd 'blood'
 * oştahë to burn, from oşta 'fire'

Particles
Particles in Ukantel are function words that must be associated with another word or phrase to impart their meaning.

Postpositions
Ukantel has postpositions, which all force their complements into a specific case. They are a closed class and are listed below in a table by the case they govern.

Cardinal numerals
Ukantel has a decimal numeral system. There are distinct numeral roots for each unit digit (1-10), and for powers of ten, namely kosa "hundred (100)", hüttö "thousand (1,000)", and yalit "ten thousand (10,000)". In forming powers of ten higher than ten thousand, compounds of these roots and yalit are used, namely kine yalit "hundred thousand (100,000)", kosa yalit "million (1,000,000)", hüttö yalit "ten million (10,000,000)", and yalit yalit "hundred million (100,000,000)". These bases denoting powers of ten and compounds thereof are not numerals in themselves, being true nouns instead, and thus may not stand on their own when acting numerically. The word "nothing" is used in place of zero. Other numerals are formed using decimal structural principles. Teens, twenties, thirties, etc are formed using the tens word (10, 20, 30, etc) followed by a unit digit, written as a single word. Tens, except for 10 itself, are formed semi-regularly from unit digits using the suffix ''-tUk. ''When forming tens with this suffix, some of the unit digits may drop or elide their final vowel.

The formation of larger numerals is somewhat irregular from a semantic point of view. Up to 10,000, numerals are formed similarly to English: cf. 2,304 two thousand three hundred four and öt hüttö salë kosa şac or 5,649 five thousand six hundred fourty nine and saç hüttö ege kosa aynutukäd .However, ten-thousands (i.e. 10,000, 20,000, 30,000, etc) themselves are formed using 10,000 eylit. For example, 10,000 is eş yalit (lit. one ten-thousand) and 20,000 is öt yalit instead of kine kosa and öttük kosa respectively. However, thousands in between are formed as in English, meaning 21,000 is öttükeş hüttö. This is true for other bases denoting powers of ten as well.

A succinct amount of numeral forms are found in the table below.