In this feature, we look at the coding of possessor noun phrases, i.e. full noun phrases (not personal pronouns) that express the possessor in an ownership, body-part or kinship relationship, as in the girl’s bed, Pedro’s foot, or mother’s uncle. Personal pronoun possessors are treated in the preceding chapter. In this chapter, only coding by adposition or case, by person-indexing, or by juxtaposition is considered. Word order is the topic of Chapter 2 and is not taken into account here. The corresponding WALS chapter is Nichols & Bickel (2005).
There are just three possibilities that are distinguished here:
|Adpositional or case marking of possessor||30||26||56|
|Person-indexing on possessum||2||11||13|
In the most frequent case, the possessor is marked by an adposition or a case affix on the possessor (value 1), as in English (preposition: the house of my brother, case suffix: my brother’s house) or the Romance languages (Spanish, preposition: la casa de mi hermano).
In many languages, there is a preposition that is derived from the ‘of’ preposition of the lexifier, e.g. Afrikaans van, Cape Verdean Creole di, Guinea-Bissau Kriyol di, Zamboanga Chabacano di/de, and Mixed Ma’a/Mbugu -a. Two further examples are given in (1)-(2):
Pidgin Hindustani uses the postposition ke, from its lexifier Fiji Hindi:
Mixed Ma’a/Mbugu has the gender-agreeing preposition -a, while Fanakalo and Lingala (other Bantu-based languages that have lost gender) have the fossilized prepositions ga and ya (deriving from -a), respectively.
But there are also quite a few APiCS languages which have a newly developed possessive preposition. The Pacific English-based languages have a preposition deriving from belong (e.g. Kriol blanga, Tok Pisin bilong), while quite a few other languages have a preposition deriving from ‘for’ (e.g. Sranan fu, Saramaccan u from English for, Tayo pu from French pour).
Similarly, the Juba Arabic preposition ta comes from a noun meaning ‘possession’ (used as a possessive preposition in many Arabic vernaculars), and Guadeloupean and Haitian Creole a apparently comes from the French preposition à ‘to’. In other languages, the origin of the possessive adposition is unknown (e.g. Sango ti, Ambon Malay postposition pung, Sri Lankan Malay genitive suffix -pe).
Less expectedly, a possessive postposition may derive from a personal pronoun, in an indexing construction like the one described in §4 below. This occurs in Korlai (su), Sri Lanka Portuguese (-su), and Papiá Kristang (sa), as well as in Afrikaans (se).
This Korlai pattern clearly derives from an original construction of the type ‘father, his house’. That su is now a postposition in Korlai is clear from the fact that the possessive postpositional phrase can alternatively be postposed (kadz pay su). Another language where this occurs is Papiamentu (see 9a-b). That su is now a postposition in Papiamentu rather than a possessive pronoun is clear from the fact that it occurs not only with 3rd person possessors, but also with 2nd person possessors (see 9b).
The possessor and the possessum may simply be juxtaposed, with no segmental marking (value 2). This occurs widely in the English-based and Romance-based creoles of the Atlantic, but also elsewhere. However, most of the languages that allow juxtaposition also have an alternative construction with overt marking, sometimes with a subtle meaning difference.
It seems that these constructions derive from prepositional or case-marked constructions in the lexifier from which the overt marker (’s, de) has simply been dropped: This is apparent from the word order, which is identical to the word order of the lexifier: preposed possessor in English-based languages, postposed possessor in French-based languages. (Interestingly, English-based languages never have possessum-possessor order in constructions without marking, so simple dropping of of does not seem to have occurred, for whatever reason.)
A construction with no marking is also found in a number of pidgins:
In our third type, the marking is by a person form associated with the possessum, as in ‘the father, his house’ (value 3). This construction is sometimes called head-marking (Nichols 1986), but here it is called person-indexing (Haspelmath 2012+).
Krio and Chinuk Wawa are the only languages where this is reported to be the only existing construction. This construction is particularly common in the West African English-based languages. As in the Korlai and Papiamentu constructions discussed in §2, the indexing person forms in these constructions may be on their way toward reanalysis as postpositions, but here we have no evidence that they are not simple person-indexing constructions.