do, make, work; to cause

verb physical_action


Proto-Siouan *hi_E

Proto-Crow-Hidatsa *hre, *-hi_E

Crow -hili ‘do, work’ [not found as an independent stem] RG

Hidatsa hrÉ ‘do, make, work’ J , hirí ‘causative suffixes’


Mandan -here- ‘causative’ RTC

Proto-Mississipi-Valley *-hi_(r)E

Proto-Dakota *-_-yA

Lakota -_-yA

Dakota -_-yA


Chiwere -hi_ ~ -re

Proto-Dhegiha *_-(r)e

Omaha-Ponca _-(ð)e

Kanza/Kaw _-(y)e

Osage _-(ð)e

Quapaw _-(d)e



Biloxi -ha-_-yE PFE

Ofo -ti


Catawba če ‘cause’ KS

General comment

The causative in Mandan is a transparent serial verb construction, both an independent verb and a cliticized auxiliary verb in Hidatsa, and strictly a cliticized auxiliary in Crow and the rest of the Siouan language family.

Historically it seems to have been a verb with medial inflection for person *hi-_-E. Mandan and Hidatsa have both normalized inflection in the independent verb by moving person markers to the left of the initial element. When the Hidatsa auxiliary is suffixed to verbs in final -E, however, the original order of elements is preserved and indicated by the -hi- which appears between the final consonant of the root and the suffixed causative person markers. The historical order is also reflected in Dakotan and DH, where aspiration appears with the possessive reflexive ahead of the personal pronouns, and in the hi element which appears ahead of the pronouns in Biloxi.

The glide which appears in many of the languages between the person markers and -E is epenthetic everywhere but not reconstructible to a single Proto-Siouan glide. Here, we represent it as *(r) because that fits every language but Biloxi. This is a notational convention only, however. Except in the reflexive possessive, Dakotan and DH have lost initial *hi-. In contrast, Chiwere/Hoocąk seem generally to have lost *-(r)E, retaining it only in fossilized (unconjugated) form with some verbs (cf. ‘hard > cane, walking stick’, ‘fall > lie down’). In modern Chiwere/Hoocąk the pronouns follow *hi- as if the remainder of the causative were still present. In all probability, *-(r)E has simply been absorbed by the pronouns. This is clearly the case in Hidatsa, where the short vowels of the pronominal prefixes are long in the causative. Reconstruction is facilitated by considering the ordinary and the reflexive causatives together.

Catawba če may be related to Proto-Siouan *-(r)E, but the sound correspondences are not well understood. It would be much easier to relate the Siouan form if it were historically *ye.

Details Language Word Source