Locus is a convenient one-word term for what is also known as head/dependent marking. In any kind of phrase, overt morphosyntactic marking reflecting the syntactic relations within the phrase may be located on the head of the phrase, on a non-head (i.e. on a dependent), on both, or on neither. Examples illustrating marking of various kinds in possessive noun phrases are (1)-(4) below. In possessive phrases, the possessed noun is head and the possessor is dependent.
The following locus types are distinguished on the map:
|Go to map|
|Possessor is head-marked||78|
|Possessor is dependent-marked||98|
|Possessor is double-marked||22|
|Possessor has no marking||32|
These five types are illustrated in the following subsections.
‘my brother's house’ (lit. ‘my-brother his-house’)
In these examples the possessed noun (the head) agrees in person and number with the possessor noun, the most common pattern for head-marked noun phrases. Agreement in gender is also fairly common. A few languages have a non-agreeing marker on the head noun. In Fijian, a possessive affix -i marks possessed nouns; that it does not vary for person or number of the possessor is shown in (2b):
In Yoruba, the possessed noun has its final vowel lengthened on a mid tone before a consonant-initial possessor noun:
‘Dada’s money’ (cf. owó ‘money’)
(4) Chechen (Nichols, own data)
‘lion cub’, ‘lion's cub’
(lit. ‘of-lion cub’)
‘the price of a car’
(lit. ‘of-car price’)
‘dog's tail’ (lit. ‘of-dog its-tail’)
‘(a/the) crocodile's tail’
(lit. ‘crocodile tail’)
‘Warse's canoe’ (lit. ‘Warse canoe’)
‘my house’ (lit. ‘I house' or ‘me house’)
In the Acoma examples in (1), the possessed noun (the head) is inflected, agreeing in person and number with the possessor. In the Chechen examples, in contrast, the possessor noun (the non-head) is inflected, bearing the genitive case. In Southern Sierra Miwok both words are inflected. In the Tiwi and Asmat examples there is no inflection of either word; possessive phrases are formed by juxtaposition of the uninflected possessor and possessed nominals. (This is not the only adnominal construction of Tiwi; see the next paragraph.)
In these examples the inflection is affixal. However, the same work can also be done by separate words. Head marking using a separate word is illustrated by (8) from Tiwi, where the marker of possession is the uninflected pronoun ŋara ‘he’. That it is syntactically attached to the head noun ‘tail’ and not the possessor noun ‘crocodile’ is shown when the order of possessor and possessed nouns is inverted: in (6a-b) ‘he’ immediately precedes ‘tail’ regardless of the latter's position relative to ‘crocodile’.
‘the crocodile's tail’
Dependent marking using a separate word is illustrated by (9), where na is a postposition, and by English phrases with of, e. g. the price of oil.
The division into head, dependent, double, and zero marking does not exhaust the possible types. There are several low-frequency but systematic further patterns, all of them grouped together as “Other” on the map. One of them is free (or floating) marking, where the marker is positioned with respect not to the head or the dependent of the phrase but relative to the phrase boundaries. The commonest such position is probably second, or Wackernagel, position, with the marker following the first word or similar unit of the phrase, as in the following NP examples from Chamorro, where =n is a second-position clitic following the first tonic word in the phrase (“=” is a clitic boundary).
‘the student's book’
That the =n is in second position is shown by the fact that in phrases with attributive modifiers, where word-order inversion is possible, it follows whichever word comes first:
‘the big man’
‘Brazil’ (lit. ‘Brazilians' land’)
This is different from (8) in that the pronominal piece in (12) is case-inflected and is therefore a syntactic word, while that in (8) has no case and can therefore be regarded as a phonologically word-like grammatical formative.
Though a number of languages have more than one adnominal construction, for this map we sought just one type per language. We have mapped only possessive constructions with overt heads and noun possessors. For all languages in which there was more than one marking pattern for possessive NP's, we chose whichever is default or has the fewest restrictions. For instance, if a closed or restricted lexical group of nouns (such as kin terms and/or body parts or “inalienables”: see Chapter 59 on possessive classification) has a distinct form of possessive marking, we have not included this but have mapped only the treatment of the open, or default (or “alienable”), set of nouns. If nouns and pronouns have different marking we have mapped only the noun marking; if only first and second persons have a distinctive form of marking we have not included it; and likewise for any enumerable or delimitable special forms of marking. Though kin terms and body parts are prototypically possessed and might be thought the best possible examples of possessive phrases, in the languages that formally differentiate “alienable” from “inalienable” possession it is always the “inalienables” that are the closed or defined class. Hence they are not the open, default construction we seek in this survey. Furthermore, the open class of possession reflects the language's typological preferences for marking locus, whereas “inalienable” possession shows a clear worldwide preference for head marking. More generally, we found that closed-class and minority patterns are so driven by universal preferences in their marking as to yield little interesting typological or geographical variation: pronouns often inflect for case or are otherwise dependent-marked in languages that have no cases on nouns (see Chapter 50), and they are also prone to cliticize to heads when nouns are not.
The result of surveying only open-class marking is fewer data points than have been used to define locus types in earlier literature (e.g. Nichols 1986, 1992, Cysouw 2002). This is an exemplar-based survey which takes the most typologically revealing and least universal-driven pattern to represent the language's type. In addition to greater typological clarity it also produces a clearer picture of continental and larger geographical distributions, albeit at the price of a more schematic description of each individual language's morphosyntax. (For further discussion of the examplar-based method, see the section “Defining locus types” in Chapter 23, and Bickel and Nichols 2002; for other applications, see Chapters 20, 21, 22, and 25.)
Head-marked possessive NPs are common in the Americas and the Pacific (chiefly Melanesia) and infrequent elsewhere. Dependent-marked NPs have a roughly complementary distribution to this: they are frequent in all parts of Africa, Eurasia, and Australia-New Guinea. The only area where the two types overlap appreciably is in New Guinea. Double-marked possession is rare; it is found around the Eurasian periphery, in the Himalayas, and along the Pacific coast of North America (see Chapter 22 on synthesis for a similar distribution). Zero-marked possession is also uncommon; instances of it on the map are mostly found near the equator, but apart from this curious distribution it does not form any true clusters.
This research was supported by U.S. National Science Foundation Grant No. 96-16448 (Nichols, P.I.), Swiss National Science Foundation Grants No. 08210-053455 and 610-062717 (Bickel, P.I.), and the Institute for Slavic, Eurasian, and East European Studies, UC Berkeley. We thank Aimee Lahaussois Bartosik, David A. Peterson, Rebecca Voll, and Suzanne Wilhite for help with data collection.