The principal parts of a Kymna verb are the conclusive form, always ending in -um, and the past tense conclusive, ending in -ende. The conclusive form is used as a predicate finishing a sentence, and is opposed to the inconclusive -e when more information is to follow.
In Proto-Tallic, the only verb having a past tense was the copula ’to be’ (its exact form is unknown), the past tense of other verbs was expressed by a participial form + ’was’. In addition, there was evidentiality marking by i-infix (for ascertained events) and u-infix (for doubtful events). Since events of the first kind more often took place in the past and those of the second kind in the future, Talmit began to use i- and u-infixion to mark tense. Kymna on the other hand dropped the copula out of the past tense construction, just leaving the participle. Thus the participial ending became the past tense marker (just like Slavic -лъ, -ла, -ло).
Hence Proto-Tallic *taplende + be-pst ’was saying’, *taplendei + be-pst ’was evidently saying’ from √tapl ’to say’ (compound form *taplen- + *de ’action, event’) result in taplendίi(-nójo) ’was saying’ in Talmit; and in Kymna simply tālende ’said’.
A fossilized u-infix is found in Kymna in the word ganimo, ganymou- ’future’, coming from *ganen-pau ’something which might come’.
Since the conclusive form -um triggers u-mutation, while the inconclusive -e and past tense -ende do not, some verb classes arose:
There are some irregular formations where past tense forms come from different roots:
In a synchronic analysis of Kymna one would say that there are two verb stems (e.g. toul- and tāl-) and would probably regard the u-affected as basic, since it appears in the common conclusive form. Verb forms using the u-affected root are:
Verb forms using the unaffected root are: