Kymna has five cases, four of which appear in two variants: neutral and affected. Marking for affectedness is not determined by the type or meaning of the verb or noun involved1, but is rather akin to topic marking or the definite/indefinite distinction. Instead of using the arguments of a verb (subject and object) as formal slots, the speaker puts emphasis on what actually matters in the context as the affected quality.
Taking a simple example:
would imply that the boy needed sustenance or was maybe poisoned, and thus was somehow affected by the eating. The fate of the apple itself is irrelevant, probably it was not mentioned before and hence the translation is ’an apple’.
would imply that the apple is now gone and thus could be an answer to ’Where did the apple go?’. It is not important in the context who ate it, so the translation has an indefinite article: ’a boy’.
On the other hand, if the eating of an apple had a lasting effect on the boy, then the resulting state is exactly what is normally described by the perfect. So in a different narrative context, one could translate the above sentences in the following way:
Affectedness also works with ditransitive verbs:
could be claimed as merit in a discussion of one’s vices and virtues2.
could be a measure taken to utilize surplus bread and clothes.
could be answered to a question about what has been done to help the hungry and naked.
Intransitive verbs can have an accusative affected and a dative affected argument beside the subject. It marks a further nuance of a positive or negative affection:
Of course, it is also possible to mark several objects as affected, so that a simple sentence like ’I love you’ has four possible formulations. One would mark both the subject and the object as affected for mutual love, and only the subject as affected for unrequited love.
A noun in the genitive can be a verbal argument as well, the subject of a stative verb, or a verb where it has rather the role of a patient (e.g. ilollonum ’to be asleep’, gȳrum ’to see’). The reason for this is mainly phonetic – the confusion of the conclusive verb ending *-un and the verbal noun *-on, both resulting in -um in Kymna. But it is very likely that semantic considerations of distinguishing a patient came into play, instead of ’I see’ constructions like *’there is my seeing’ might have been more common.
Hence for example *pan-mo idlondlonon ’the child’s sleeping’ became ponuma ilollonum ’the child is asleep’, *i-mo gwimron ’my seeing’ became ima gȳrum ’I see’.
This can also be put into the genitive affected. From here, the genitive affected was actually resubstituted back to nouns in order to denote inalienable possession: ima savi ’my clothes’, but namon āra ’my arm’.
The fifth case is the resultative and is used whenever something is created. For example, the direct object of a verb like ’write’ may be ’ink’, of a verb like ’build’ may be ’bricks’, but ’letter’ or ’house’ will stand in the resultative case, thus literally ’write ink into a letter’, ’build bricks into a house’. For historical reasons, resultative -man is always in the affected form. It traces back to a modification of dative *ma. The Talmit cognate mére denotes changing into a state.
The case endings in Kymna derive from agglutinated Proto-Tallic postpositions. The following table gives a general overview of the endings, but variations appear depending on the final consonant of a noun:
Note that gen. *mo > ma in Kymna fell together with dat. *ma. Therefore, instrumental *zo became the new dative.
The dat. and acc. sg. neutral -as and -on after consonants are substituted by analogy from the very common nouns ending in -a.
The plural affected suffixes are straightforward pl. *-mi or *-te + affected singular. The m of *-mi fell away with compensatory lengthening by dissimilation with a following m, hence mīon < *-mimon, -mīa < *-mima. The long vowel was by analogy extended to -mīva for *-miva < *-miwa, -mīdzon for *-midzon < *-mizon and -mīnas for *-minas < *-minajə.
The plural ending -te is preferred after -s, -n to prevent adding a syllable by an anaptyctic vowel. In colloquial speech, contractions -teva > -tva, -temon > -tmon are common. Since this is a very late development, there is no further change of consonants.
A special class of nouns are those with final -n hailing from syllabic *-ṇ, like sumun ’sound’, vomun ’sky’, vuchun ’spine’, hymun ’light’ (cf. T. hómne, bámne, búkne, ϕjómne). They take plural endings with -t- (sumunte, sumuntva, sumuntmon etc.), but behave like vocalic stems sumnu-, vomnu-, vuchnu-, hymnu- in the singular (sumnuma, sumnus etc.; exception: acc. sg. aff. sumunnas). Compare the ager-type second declension in Latin.
Nominalized adjectives also use the t-scheme by analogy to their nominalization suffix -t(a). Thus grodzuma ’strong’, grodzumat(a) ’strong one’, pl. grodzumate ’strong ones’ and so on. A common contraction is *-enna-t- > -ent- (havenna ’hungry’, haventa ’hungry one’, pl. havente).
Nouns in -e form the accusative by -in rather than *-ën, with loss of rounding in unstressed position, e.g. roume ’language’ > acc. sg. neutr. roumin.
Nouns with m in the last syllable receive dissimilated suffixes with n instead of m, and those in -me/-mi usually lose the last vowel, e.g. gomi ’nose’ is declined in the plural neutral as follows: gomni, gomnīa, gomnis, gomnin and in the plural affected: gomnīva, gomnīon, gomnīdzon, gomnīnas.
In words which end in -e, where final -e is a shortening of -ei in unstressed position, -ei- reappears when a whole syllable is attatched as an ending, e.g. ame ’city’ > acc. sg. neutr. amin, but acc. sg. aff. ameinas. This includes the large group of abstract nouns ending in -me < *-mai, but not those ending in -me < *-mne (as salme ’water’), although the two groups are often confused towards one or the other by analogy.