The Giving Grammar, Part II

I have examined the languages of the Pacific a little more, since Judd's grammar preceded the concept of the ergative. As it turns out, western Pacific languages, such as Hawaiian, tend to be nominative-accusative system rather than eastern Pacific languages, which tend to be ergative-absolutive. This means that the nominative system of Hawaiian must have developed out of a previously ergative-leaning system. Even in Hawaiian, the "passive" system seems to be favored, such that Ua ai ia puaa e au, "The pig was eaten by me" is the almost the same as Ua ai au i puaa "I ate the pig." In an accusative language, the first sentence would be verb-nominative-ablative, and the second would be verb-nominative-accusative; in an ergative-absolutive language, the first sentence would be absolutive-ergative, and the second would be absolutive-preposition phrase.

Returning to the previous examples, the nouns in the phrase Ke haawi aku nei au i keia ia oe
would be nominative-accusative-accusative. It is not uncommon for languages to mark the direct an indirect object in the same manner, and the syntax here provides the information needed to distinguish the two. The second example, Owau ke haawi aku nei i keia ia oe merely fronts the nominative and add the topic marker o. In the third example, O keia ka'u e haawi aku nei ia oe, the accusative direct object is fronted and the nominative au is transformed into the a-class possessive form ka'u.  Judd translates this as "This is mine to give to you"; although the form ka'u possesses an appropriate case
and degree of control (a-class) suitable for an ergative, the shift from ke to e would not be necessary in an ergative language. If one wished to place emphasis on the indirect object, the indirect object must be fronted. The subject, however, becomes embedded in a structure ka/na mea a'u. The form a'u is the a-class genitive form of au.; thus it contains the degree of control necessary for a subject, and a case (genitive) often associated with the ergative. The word ka mea can mean "person" or "thing," but can also mean "cause," an appropriate form for the agent of the sentence. Given the habitual dropping of mea from compound phrases, I would not be surprised if the forms ka mea a'u and ka'u had the same origin. The last example, Na'u keia e haawi aku nei ia oe, places au in a dative form in front of keia. Although Judd defines na'u as a dative,  it does not serve as a marker of the indirect object, but rather means "for me, concerning me, on account of me." The sense "on account of me" approaches, but perhaps does not reach, the degree of control that an ergative language might use.

Hawaiian is a nominative-accusative language, but with structures that betray its ancestral ergativity.