In some languages objects [sic] — whether objects of transitive or subjects of intransitive sentences — may be alternatively marked by an accusative-nominative marker or by a partitive marker, and this marking difference does not correlate with any difference in semantic case function, but it correlates at least in some cases with one or more of the following semantic properties of the verbs or nouns involved: a. the definiteness-indefiniteness of the noun phrase; b. the extent to which the object is involved in the event; c. the completedness vs. non-completedness of the event; d. whether the sentence is affirmative or negative.
Standardized
Objects of transitive or subjects of intransitive sentences (absolutives) may be alternatively marked by an accusative-nominative or by a partitive, and this marking difference does not correlate with any difference in semantic case function, but with one or more of the following semantic properties of the verbs or nouns involved: a. the definiteness-indefiniteness of the noun phrase; b. the extent to which the object is involved in the event; c. the completedness vs. non-completedness of the event; d. whether the sentence is affirmative or negative.
As it is stated, this is clearly not a universal. But can this possibly be strengthened to a universal claim, as follows?Alternative case marking of an absolutive NP (i.e., direct objects and intransitive subjects), in particular (a) absolutive or accusative/nominative vs. (b) partitive or genitive, never reflects a relational-semantic difference but always one of definiteness, involvement (total/partial), completion of event, and/or polarity (positive/negative). There is a problem, though, in identifying the case alternations involved. For instance, in Turkish (Turkic, Altaic) the DO marking alternation is between an unmarked form (called absolute or nominative) and an accusative. In Finnish (Finnic, Uralic) the unmarked DO case is genitive/nominative, and its alternative is the partitive.Possible counterexamples: Languages like Old English , where objects show case marking alternations that are of relational-semantic relevance, like hieran with ACC ‘hear’, with DAT ‘belong to, obey’.
As it is stated, this is clearly not a universal. But can this possibly be strengthened to a universal claim, as follows?Alternative case marking of an absolutive NP (i.e., direct objects and intransitive subjects), in particular (a) absolutive or accusative/nominative vs. (b) partitive or genitive, never reflects a relational-semantic difference but always one of definiteness, involvement (total/partial), completion of event, and/or polarity (positive/negative). There is a problem, though, in identifying the case alternations involved. For instance, in Turkish (Turkic, Altaic) the DO marking alternation is between an unmarked form (called absolute or nominative) and an accusative. In Finnish (Finnic, Uralic) the unmarked DO case is genitive/nominative, and its alternative is the partitive.Possible counterexamples: Languages like Old English , where objects show case marking alternations that are of relational-semantic relevance, like hieran with ACC ‘hear’, with DAT ‘belong to, obey’.