Grammar

Nouns

Plurals

The ordinary productive plural was originally formed by a suffix *-i, which has long since been eroded away. Nowadays, the plural is marked by umlaut of the final vowel (if it was originally a back vowel), plus lengthening of a final vowel which is followed by a single root consonant.

There are probably a few irregular forms formed by analogy and other processes as well; these are usually marked in the lexicon.

A few words also have dual forms; these are usually nouns that would naturally occur in pairs, such as eyes, ears, etc. The modern dual is marked by -Vt; the historical form was likely a gemination of the final consonant (or something that evolved into a gemination of the final consonant) plus *-ət.

Possession

These will trigger length changes and, as necessary, umlaut.

There is probably an idafa-like construction that is used to express inalienable possession, whereas a preposition is used for alienable possession by a noun.

Miscellaneous

There is no evidence of case as a grammatical category in this language, at any stage of its development that can be reconstructed.

With the possible exception of the ergative personal pronouns, which can be analyzed as possessive forms of a defective root, a personal pronoun cannot be the object of a preposition. Traditionally, a possessed noun is used to paraphrase these meanings — commonly a body part (including jelot 'soul, self').

Numbers

The numbers from 1 to 10 are:

Tens formed regularly as multiples of shekr, and hundreds as multiples of thism '100', with the multiplier preceding the power of ten. If multiple places are specified, the order is (as in English) big-endian, with the largest place coming first. Places can be optionally separated by ash 'and'; shekr (but not thism) can also be dropped before ash: paksh ash taal '23'.

Numbers from 1,000 to 9,999 are expressed as a two-digit multiple of thism followed by another two-digit number (optionally separated by ash, as always): sitaath shekr taal thism ash paksh shekr jöst '4,321', laghm ash shukh thism qum ash jilaak '9,765'. Even higher numbers can be expressed using multiples (with a multiplier of up to 9,999) of äthsiim 'myriad'.

For numbers from 100 million up, there are two options. One is to use the preposition le to indicate a power (usually of myriad): äthsiim le paksh 'myriad to the power of two'. However, a slightly more modern/technical strategy is to apply the pattern läCCiiC to a numerical root, creating a word meaning 'myriad to the power <number>'. This new pattern was consciously created by an early mathematician, probably by combining the preposition le with the augmentative pattern äCCiiC found in äthsiim: läpkiish '100,000,000', läthiil '1,000,000,000,000'.

Note that the strategy with le is still the default way to express '<number> to the power <other number>' explicitly. Numbers formed with läCCiiC are best compared to the (long or short scale) terms billion, trillion, quadrillion, etc. It is still unclear what le meant, but probably something like 'by' or 'through'.

Numbers precede nouns, and do not require the plural to be explicitly marked.

Verbs

Subject Markers

There is also an archaic third person dual form jet-, which only occurs in very affected language nowadays.

TODO: sound changes when plural prefixes are added to a stem that already begins with two consonants, especially if the first is nasal.

For many verbs, the Accusative declension (with these markers) now has a future meaning. These are probably non-static verbs, although I'm still not sure.

There is also a past/perfective tense (possibly Ergative), plus a progressive for non-static verbs (gerund, which agrees with object using possessive markers; subject is marked on auxiliary or just with a pronoun).

Possibly the auxiliary for present progressive is W-K-J with CCaC? Then singular forms would be nüüke, tooke, and jööke. But I'm not sure about plural. Those might be slightly irregular after all these years, or even suppletive.

Passive continuous: possessive marking on gerund, preceded by the auxiliary week or in some dialects we(e), shortened from jeweeki, the dynamic passive of the above verb. In formal speech, you can still hear the full form.

Past continuous (ergative) probably uses the auxiliary juuki (W-K-J with CuCC pattern, plus sound changes). This is probably shortened to juuk or ju(u) by most speakers in ordinary speech.

You can derive an adjectival participle with the prefix u-. This is a common way to form e.g. professional names.

The prefix jo- added to a verb indicates the place where something happens.

Ergative Personal Pronouns

Possibly also jetli attested in some ancient text somewhere?