LINGUIST List 14.2077

Tue Aug 5 2003

Sum: Adnominal possessives and animacy

Editor for this issue: Karen Milligan <>


  1. Anette Rosenbach, Adnominal possessives and animacy

Message 1: Adnominal possessives and animacy

Date: Tue, 05 Aug 2003 06:13:30 +0000
From: Anette Rosenbach <>
Subject: Adnominal possessives and animacy

A month ago I posted the following query (Linguist 14.1869): 

''Is anyone aware of a OV-language which has 2 adnominal possessive
constructions which differ in the position of the possessor
(i.e. which have both a preominal and a postnominal genitive) and
where there is an animacy-induced preference for either position?
Similar to English (though a VO-language) where human possessors are
preferably realized in prenominal position (1),while inanimate
possessors usually occur postnominally (2).

(1) Johns book
(2) the roof of the house

Is there anything comparable for OV-languages?

I wish to thank Gerlof Bouma, Mark Donohue, Dafna Graf, Detmar
Meurers, Hirotaka Mitomo, Pius ten Hacken, Claus Pusch, Horst Simon,
and Helmut Wei� for their responses. Heres a
summary of what they wrote:

Hirotaka Mitomo wrote that: 
Japanese is a typical OV-language, and only has a prenomimnal position
for dependents on a noun. Your (1) and (2) are translated into (1a
and (2a) respectively.

(1) a. John-no hon
 -Gen book

 b. *hon John(-no)

(2) a. ie-no yane
 house-Gen roof

 b. *yane ie(-no)

Mark Donohue pointed out Saweru, a SOV language of central Yapen
island, New Guinea, which has various possessive strategies:
Possessive prefixing (with 2 sets, alienable and inalienable) (cf. 3),
dative marking (4 and 5) and the general modification strategy
(6). While human possessors show a preference to occur prenominally
(3-5), inanimate or at best non-human possessors tend to occur
postnominally (6). (5) is restricted to human possessors and the only
possibility for pronominal possession.

(3)	ruama (afi) a=watu(n)
 woman she 3SG.GEN.ALIEN=house
 'the woman's house'

(4)	ruama (afi)=ai watu(n)
 woman she=DAT house

(5)	ruama watun=rai
 woman house=3Sg.F.DAT
 the womans house

(6)	watun=o mae
 house=LNKR dog
 'dog's house'

Horst Simon, Detmar Meurers, and Helmut Wei� pointed out German
as an underlyingly OV-language, where prenominal possessors are
restricted to human possessors (7, with 7b being non-standard), and
inanimate possessors have to occur postnominally (8)

(7)	a.	 Friedas Vater
 Frieda-GEN father
 Friedas father
	b.	der	 Frieda ihre Mutter
		the-DAT Frieda her mother
		Friedas mother
(8)	der Sattel von meinem Fahrrad
	the saddle of my bike

Claus Pusch provided data from his German dialect,

(9) John's book = em Johann si buech (dem Johann sein Buch) 
(10) the roof of the house = s'dach vom huus (''das Dach vom Haus'')

He also pointed out that he feels that the prenominal position is
particularly preferred if *both* the possessor and the possessum are
animate, as in (11) and (12)

(11) em Johann si brueder (''dem Johann sein Brueder'')
(12) em Fritz si chatz (''dem Fritz seine Katze'')

A further expanded possessum does however decrease the acceptability
of the prenominal construction (13 and 14):

(13) ? em Fritz si(s) chliis ch�tzli (''dem Fritz seine kleine Katze'') Fritz little cat
(14) s'chlii ch�tzli vom Fritz the little cat of Fritz)

Gerlof Bouma and Pius ten Hacken pointed out Dutch, which behaves
quite similarly to English and German. The following detailed data
were provided by Gerlof Bouma:

(15)	 Jans huis
 john+gen house
(16)	 het huis van Jan
 the(neuter) house of john
(17)	 het dak van het huis
 the(neuter) roof of the(neuter) house
(18)	*het huis' dak
 the house+gen roof

(18) is out, also because the gen. of `huis' simply does not
exist. Only in idiomatic constructions like:

(19) 	 's lands beste koffie
 the(neuter+gen) land(gen) best coffee

which can easily be paraphrased as:
(20) 	 de beste koffie van het land
 the(comm) best coffee of the(neut) land

Gen. forms (created by suffix -s) only exist in Dutch for
proper-names. As a native speaker I would say names for persons are
strongly preferred, but others are not ruled out. A construction with
an adjective like the following, I would allow:

(21)	 Europa's belangrijkste bankier
 Europe's most-important banker

Finally, names denoting things other than persons do not have a
genitive form if they include a (visible) determiner. The following
line is from the website of DSM, formerly `De Nederlandse

(22)	 Toen de mijnwinning groeide en DSM's verwerkingsactiviteiten
 when the mining grew and DSM(gen) processing activities
 uitbreidden ...

Replacing DSM with the old name does not work...

(23)	 * en De Nederlandse Staatsmijnens verwerkingsactiviteiten
 and the-dutch-state-mines(gen) processing activities

Finally, it _does_ work with a proper-name denoting a person, with a

(24)	 De Gooiers acteertalent
 	 de gooier(gen) talent-for-acting

 (Rijk de Gooier is a Dutch actor)

Finally, Dafna Graf pointed out Maria Koptjevskaja-Tamms
(2001) article on Adnominal possessives in:
Martin Haspelmath et al. (eds.): Language Typology and Language
Universals, Volume 2, 960-970. (Handbooks of Linguistics and
Communication Science 20. 1,2): Berlin/New York: Mouton de Gruyter.

Thanks again to everybody!

Anette Rosenbach
(Heinrich-Heine University Duesseldorf) 
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue