Review of Dimensions of Possession
Date: Sat, 11 Jan 2003 17:04:26 -0500
From: Judy Dick <firstname.lastname@example.org>
Subject: Typology: Review of Baron, Herslund, and SØrensen (eds) (2001)
Irene BARON, Michael HERSLUND and Finn SØRENSEN
(Copenhagen Business School) (eds.) (2001), Dimensions of Possession
Amsterdam: Benjamin's Typological Studies in Language ISSN 0167-7373;
47 vi, 337 pp. ISBN 90 272 2951 Hardcover
US & Canada: 1 58811 062 1 / USD 86.00
Rest of world: 90 272 2951 1 / EUR 95.00
Linguist List Book Notice Tue, 05 Feb 2002
Judith P. Dick , Software Mechanics Nepean ON, Canada
The articles in this collection were selected from those presented at an
international workshop held in Copenhagen in May 1998, according to the
publisher's notice. However, I was able to find nothing in the volume to
confirm that description.
The purpose of the collection seems to have been to view linguistic typology
through the lens of possessive expressions. The essays cover the recognized
problems of interpreting possession in across many languages. Both synchronic
and diachronic approaches are represented. The most common theme is the
problem of determining which expressions have been grammaticalized.
The essays will interest those who are interested in Functional, and
Lexical-Functional grammar. Those interested in case grammar will also find
Introduction; Dimensions of possession, by Michael Herslund and Irene Baron..
The editors set the tone with a summary of thought on "possession" including multilingual references, glimpses into the papers that follow, and a lengthy list
of references. Their own approach is onomasiological. "Possession" expresses
the semantics of "Possessor" linked with "Possessee". The relation parallels
"Location-Argument", and "Experiencer-Stimulus" pairs, more readily than
the "Agent-Patient" schema. "Location" is the primary source of their concept
of "possession", as it is primitive and concrete. The division represented by
"have" and "belong" is fundamental. Both verbs express "ownership", a
subcategory of "possession".
1. The operational basis of possession; A dimensional approach revisited,
by Hansjakob Seilor.
Seilor analyses use patterns on Universal, General Comparative Grammar (GCG)
and Local levels, then presents a menu of GCG morpho-syntactic techniques
and categories denoting possession. He says that "possession" expresses the
ego's interaction with the world in gaining property as "acquiring", "owning",
and "belonging". The menu is extensible. Items range from "inherent" to
"established" and include "connective", "classifier", "case" "location" and "verb" categories. Expressions represent degrees of alienability in the range. Values are
not specified. Individual languages select from it locally. Examples from Cahuilla
demonstrate flexibility in its application at the Local level.
Examples of contradictory case marking examples are in the middle of the range. Genitives denote "appurtenance" not "possession" and imply more
intimate associations than the dative. The dative expresses "inherence".
Both may represent, according to context, "more inalienable", "more
alienable" or "neutral alienability". Lack of specificity and inversion mark
the middle area, with verbs between Possessor and Possessum.
Acquired property units relate back to the Possessor. The dynamic is reversed
in some languages. A diachronic description of possessive verbs demonstrates
that transitives express explicitness in acquisition. Less focused concepts are
stative or reflexive expressions.
2. The concept of possession in Danish grammar, by Ole Togeby.
Danish possessives are not grammaticalized, Togeby says. The possessor is a
"reference point", used to establish mental contact with the "target", making
it accessible to the mind. This is part of a larger context of "reference point",
following Langacker. His forms of possession figure in the discussion, as does
the author's case system (Togeby, 1996).
Togeby isolates six possessive forms: genitive definite (English genitive),
prepositional phrase, indirect object, ethical dative + preposition + definite
form (about body parts and clothing), and verb. He analyzes examples of each,
taken from "The Tinder box" (H.C. Andersen). A lengthy excerpt is included
in Danish and English. Each syntactic form has multiple interpretations.
Meronymy, location, subjecthood, control, effect, and experience are among them,
as well as possession. Possession is not primary.
Predicative possessives owe much to lexical interpretation, since "possession"
is part of the semantics of verbs such as "have", "get", "give", and "take". He
explains the Danish indirect object as a reduced subject-verb relation (Diderichsen,
1946). "Have" syntactically designates the indirect object. The "have" role implies
"possession", although its scope is broader.
3. Possession spaces in Danish, by Finn Sorensen.
Editor Sorensen, outlines a relational, localist theory of possession. He seeks
to unify "possession" and "location" by using "space", a construct denoting an
"empty and extended abstract object" (p. 58) that exists as part of a location
structure in which some object is claimed to be in the "space".
He defines the pattern form as <L, o, s(o')> (p. 58) in which "L" is a location
relation, "o" and "o'" are objects and "s", is a space function. The function, "s",
is constrained by a locative, binary relation. Each space has an object. The
space function assigns the role of "possessor", denoted by the operator "pos",
to its' object. "Possession" is the relation between the space and its' possessor
object. Other elements in the space contribute to the interpretation of the space.
Some relation, other than "pos", within a space locates a possessee object.
Prepositions indicate spaces within the range of their objects.
To demonstrate, he examines "give", "have", and "genitive" constructs. All are semantically flexible, but the genitive is the least constrained. His strongest
examples are genitive constructions and whole-part relations. In all
spaces, one can distinguish between those internal to the object used to create
the spaces, and those external to those objects. He hopes shortly to
model in/alienability relations. Spaces reconcile "possession" and "location",
but are said to be capable of distinguishing between them when selections are
made by some constructs.
4. The verb HAVE in Nyulnyulan languages, by William McGregor.
Most "have" verbs in the Nyulnyulan languages of Australia are classed
formally as intransitives. McGregor conjectures they are "to some extent
transitive" and express predicative possession.
In each Nyulnyulan language "have" bears its own meaning but may also
express a contextual meaning. The inherent meaning is not defined, but three
features in each "have" construction are. They are: a semantic association
between subject and object, the expression of an ongoing happening or
situation, and subject-object asymmetry.
Nyulnyulan "haves" are not copular as syntactic and semantic evidence shows.
Although they have marginal use as a type of auxiliary, they are not
like "have" in English and "avoir" in French. Nyulnyulan "have" is said to be
more dynamic than "have" and "avoir", but less transitive at the lexical level.
The Nyulnyulan "haves" are used mainly in non-material senses, so "hold",
"take", and "grasp" senses are imposed by the context.
The verbs are fully transitive in terms of the clause types they "slot into".
The "have" clause expresses "possession" as its primary predicate,
but is not a predicate possessive construction. The associative link between
subject and object is specified by the lexical verb of having, and their interpretation
is said to depend upon the semantics of that link.
5. Semantics of the verb HAVE, by Irene Baron and Michael Herslund.
Editors Baron and Herslund discuss hierarchies of sentence constituents
consistent at each of three levels: local, inclusional and grammatical. They
use "have" possessives that highlight asymmetry to show an inclusional
hierarchy is superimposed upon local relations. The examples are Danish
but the patterns are universal, they say.
"Have" establishes a state relation. The semantic link between subject and object specifies each relationship. The denotation of the subject includes the denotation
of the object. There are three inclusionary roles described for the object:
"part of a whole" for relational nouns, "part of a subject's possessions", for
animate, non-relational nouns, and "semantic feature of the subject noun".
"Have" allows adverbial expansion. A locative prepositional phrase may join the
complement. The locative object is located with respect to the subject. The
authors designate a "subplace" for it before the object. It is related remotely to
the subject, and proximately to the object by its preposition.
Sentence constituents, reflecting denotative variations, are organized hierarchically.
The hierarchy is imposed semantically, although its elements are syntactic.
The subject always includes the object or the subplace. The object never includes
either of the other constituents. The authors demonstrate both two and three
constituent inclusion. They test by transposing "have" constructions as genitive
noun phrases. The nominal construction presupposes the inclusion of object
denotation in subject denotation. The test fails where a subplace is included in
the relation since the object inclusion relation is blocked by the adverbial locative attaching to the subject.
6. Possessum-oriented and possessor-oriented constructions in Russian, by
In Russian, two case-marked possessive constructions are well established, the possessum- oriented with the dative, and the possessor- oriented with the
instrumental. The genitive does not express possession. A "synthetic passive"
construction is offered as evidence that the analysis is correct.
The Russian mood system determines the case system. The genitive is an oblique,
outer case. Although it denotes separation or remoteness, it may be interpreted as
a possessive. However, oblique inner cases, dative and instrumental, express
Russian verbs of all classes contain state descriptions and so can express
possession. Activity verbs are state verbs with a described activity like
"zavedovat" ("manage"), "rasporjazat'sja" ("control"), and others. The
activity description logically entails a state description-location, experience,
possession, or others. "The role of the state description can be extracted from
the role that the very state plays in a real activity" (p. 110). Activity verbs are
possessor-oriented and denote an inherent property described in the entailment
The possessum-oriented construction does not appear with activity verbs. Russian grammar prohibits passives with intransitive verbs. Nevertheless, a "synthetic
passive", possessum-oriented construction occurs in common parlance
though not in grammar texts. In it, the surface subject is the possessum the verb is
a possession-based activity verb, and the possessor is an oblique instrumental object. Durst-Andersen offers the construction as evidence that the analysis is correct,
and that the possessives are grammaticalized.
7. Datives and comitatives as neighbouring spouses; The case of indirect objects
and comitatives in Danish, by Lars Heltoft.
The author regards the analysis of language- specific content form as essential to typological analysis. He reduces the number of semantic roles using "receptive"
and "comitative" frames and claims to produce core possessive meaning across
cases. "Possession" is defined as "Zugehorigkeit".
Heltoft uses "content subject" to explicate the frames. It is a constituent that
complies with the semantic subject restrictions of a given verb.
In Danish, a subject is not semantically coded but restricted by associated lexical
items or syntactic constructions. Objects may not express agentivity. Frames are described as "abstract meronymy" or "semantic fields". Each brings its'
own subject content rules. Rather than semantic roles, frames constitute
"differently organized relations". In Danish, they are non-local. The
content form has two parameters, "telicity" and "presupposition vs. assertion".
The "receptive" is an indirect object (IO) frame. IOs pattern with telic verbs in
the first object position. The frame is a content subject, restricted as is the
expression subject of "fa" ("get"). With an IO, the direct object is
interpreted as the framed entity. The "comitative" is a co-subject (CS) frame.
CS is a phrase with preposition "med" ("with") a static argument, framed by
what is denoted by the expression subject of "have" ("have"). The CS
frame presupposes that the subject and CS share the same frame, and the
same activity. Variant CS semantic roles are discarded as structurally
superfluous. The frame covers any of the static co-part relationships
presupposed or entailed, for example, sociatives, or instrumentals.
8. Towards a typology of French NP "de" NP structures or how much
possession is there in complex noun phrases with "de" in French?, by
The author presents a two-part model of interpretation of "de" in order to
Identify possessives. Individual "NP de NP" structures are treated separately
from those in discourse. The division is justified by syntactic tests employing
the pronominalization of human N2s in "de lui" and the im/possible use of "etre
de N2" (Milner, 1982) extended to include attributive cases (Bartning,
1990). Bartning contends that contextual uses are grammaticalized. Examples
show prototypical uses accepting associative anaphora, but not discursive
ones. The possessive determiner works at the discourse level in pragmatic cases,
but not in contextual cases. At the microlevel, it works with all possessives.
Microlevel prototypes constitute the "kernel meaning group". They are
distinguished by the properties of N1, and occasionally N2. Bartning
classes the prototypes as "possessive" or "origin" based on their syntactic
behavior and the original meaning of "de". "Possessive" relations are paraphrased
as either "N1 avoir N2" or "N2 appartenir a N1" expressions, and "origin"
relations as "etre a" expressions.
Discourse interpretations include extra-linguistic knowledge and meaning from
Previous discourse and textual intervals. The relations are paraphrased as "N1
is associated with N2" at the textual level. The adnominal NP 2 and
possessive determiner do not express possession. "Avoir" and "appartenir a"
paraphrases are not acceptable.
9. Spanish N "de" N structures from a cognitive perspective, by Henrik
The author regards "de" as a cognitive primitive, whose function is to combine
entities in a given way. A cognitive primitive functions on a pre-possessional
level. The interpretation of the "N de N" structure depends upon the head
noun's restrictions. Restrictions may come also from contextual properties,
or world-knowledge. A more general semantic notion encompassing locative
and possessive meanings is thought possible.
"De" establishes a semantically indeterminate relation between linguistic
representations of two entities. There is no systematic correlation
between the semantic and syntactic levels. Muller says that a noun is neither
relational nor non- relational. What is significant is that the other noun or the
context can activate relations in some head nouns because some nouns are more
likely to be parts than wholes, given our mental representations.
10. The grammatical category "Possession" and the part/whole relation in
French, by Martin Riegel.
The author developed a prototype for French possessives, using "participation"
for generic "possession". French is thought to have a limited concept of
possession with a central part/whole relation. Riegel joined "participation"
and syntactic criteria with a definition adaptable to "mereological calculation".
The French examples are not translated.
To a participation category he added three syntactic patterns, "ownership",
"parental kinship" and "partitivity" (from Fonagy, 1975), integrated with the
central concept to varying degrees. They are subsumed relations. He added a
"collection" entity and a "set" relation with its elements, and formalized the whole.
The syntactic structures accommodate other relations, among them "state-entity".
Riegel produced possessive syntactic variations to demonstrate. The prototype
is said to be applicable to additional, peripheral possessives. He discussed legal ownership briefly.
11. Kinship in grammar, by Osten Dahl and Maria Koptjevskaja-Tamm.
The authors contribute to the understanding of grammaticalization by providing
a description of kinship terms, their unusual semantic and pragmatic features.
They are sympathetic to both synchronic and diachronic linguistic approaches.
They choose "degree of closeness to parental prototypes", "ascendance", and
"lineal descent" as the most meaningful kinship concepts for linguistic analysis.
Kinship nouns are animate, similar to proper names and usually definite. They
are typically syntactic subjects, egocentric, and pragmatically anchored.
Possessive nouns and verbs co-exist in a number of languages. Then the verbs
Are restricted in tense, mood and aspect. Lexical integration, accounts for many
changes in terminology and interacts with grammaticalization.
The authors compare in/alienability relations with kinship terms to those with
body parts drawing examples from many languages. Significant features discussed
include obligatory possessor marking and discourse status
12. (In)alienability and (in)determination in Portuguese, by Anne-Marie
Spanoghe tried to determine empirically if there is an in/alienable possession
parameter in Brazilian Portuguese. She strictly controlled the interaction
between the test language and the experimental model, to avoid circularity.
Results were indeterminate.
In Portuguese, the possessor is not explicitly expressed. Inalienability is
interpreted using pragmatic and discourse information. Spanoghe's
database consisted of 2700 items from contemporary writings. She posed 3
questions about the interface between the body-part continuum and the
1. What is the impact of the use of a non- prototypical (Seiler, 1983) body part
on the type of determiner used?
2. Is the definiteness of the body-part noun phrase the result of the possessor's
3. Does a non-prototypical possessor have an influence on the appearance of an
Results of 1. showed that the majority of cases with non-prototypical body-part
phrases had definite articles. In the answers to 2. The possessor fades, as the
statements move toward alienability. The results of 3. Were indeterminate.
She describes the use of the possessive determiner in discourse, its forms
and the morphological oppositions found, expressing the degree of recognition
of the possessor. She assumes European Portuguese is like Brazilian.
13. Possessives with extensive use; A source of definite articles?, by
The author examined possessive changes as a source of definite articles. In Komi
and Udmurt, (Permic languages of the Uralic family), Turkish, and
Yucatec Maya some possessives may grammaticalize into definite articles, as do demonstratives in some other languages. Where there are non-lexical
attributive possessives and expressions with possessive pronouns
grammaticalization starts with a possessive extension with associate anaphora.
Degree of referential directness is the test. Constructions with clitics or affixes
Functioning as possessive determiners, or modifiers may be
anchored in several ways. Associativity is more significant with them than
referentiality. The author made frequency comparisons with English and
Swedish, but avoided a Eurocentric approach to grammar. Both quantitative
and qualitative results were indeterminate.
In each test language, a free morpheme changed to a clitic or affix without an accompanying change in use. Possessive usage is not obligatory with
any of the extended constructions, associative or non-associative. Extensive
and non-associative possessive uses continue along with the extensive
use of demonstratives. Qualitative results varied according to language, especially
for "article- like" terms. Both extended associative anaphoric uses, and direct
anaphoric uses appeared. Both "article-like" and discourse extensions occurred.
Fraurud concluded that extended use is about referentiality and especially, focus
of attention, not about possession. Historical data indicate stability over time, with strong resistance to morphological change, but less to syntactic.
14. Possessors and experiencers in Classical Latin, by A. Machtelt Bolkestein.
Bolkestein emphasizes the difference between datives and genitives in possessive constructions. He argues for the use of "Experiencer" rather than "Dative" on some
levels of Dik's (1998) hierarchical structure.
The Experiencer possessive is a satellite on the propositional and utterance levels.
The Genitive links associated participants and predicates properties of
possessees. Bolkestein draws upon Classical Latin for examples in particular,
the dativus Sympatheticus to profile relevant constructions. The analysis is
confirmed by correlation with semantic properties of the dative constructions.
Dative-like Experiencers are "+animate". The predicated state of affairs in
the dative construction must be able to affect the Experiencer. Stative verbs work,
but dynamic verbs are more commonly used with them. Propositions are expanded
by the attachment of modal operators and/or satellites. The Latin examples clearly
show dative constructions delimiting the truth-value of the content proposition as
less than universally accepted. Bolkestein uses "Experiencer" on the
propositional level to mean "Experiencer of the truth-value of the proposition
contained in the utterance" (p. 281). Datives in the exchange of utterances
between participants, which are not part of nuclear predication or propositional
content, have no affect on truth-values. Their label is "Experiencer of the speech situation" (p. 281).
15. The difference a category makes in the expression of possession and
inalienability, by Marianne Mithun.
The author describes possessive constructions in three North American
languages: Lakhota (Siouan), Kathlamet (Chinook), and Mohawk (Iroquoian).
Although the constructions are apparently attributive, she says they express "affectedness of a participant" in verbal constructions.
Some verbal affixes that are clause markers are commonly translated as possessives.
Each pronominal prefix indicates a participant. The participant's role, if he is
directly affected, is patient or absolutive. If he is indirectly affected it is dative
or beneficiary. Affectedness suggests possession. Direct affectedness suggests
inalienability; indirect affectedness suggests alienability. The grammatical
construction is warranted by the effect of the event or action upon the participant.
It does not specify possession grammatically. If the hearer understands
"possession", he infers it from context and cultural knowledge.
The author examines the interrelationships of the test constructions with other constructions expressing common relations including inalienability, partitivity,
animacy, intimacy of involvement, and topicality. She demonstrates
that none of those constructions specify the named meaning grammatically.
Hearers infer those meanings on occasion. Context and speaker's
lexical choices differentiate among them.
16. Ways of explaining possession, by Bernd Heine.
Heine recommends a phenomenological approach to linguistic analysis in order
to understand with certainty, why possession is encoded as it is cross
linguistically. External knowledge affects diachronic derivation, he says.
Possession is an abstraction derived from simple, concrete constructs. We need
to know about those constructs to understand the abstraction. Good explanations
depend upon asking good questions, then looking hard enough for answers.
Possessive grammatical forms are more likely to result from external influences encountered in trying to communicate successfully, than from non-possessive
forms. He analyzes use patterns in preference to developing rules, and eschews monocausal explanations as simplistic.
Predicative possessives interest him here. He demonstrates with examples from
Kxoe (an SOV Khoisan click language of Africa). Using portmanteau postpositions,
he shows that language differences can complicate the correlation between source schemata and morphosyntactic patterns. With an example of structural change
from locative to possessive, he demonstrates loss of paradigmatic variability.
The new form is grammaticalized and hence conventionalized, losing its
Paradigmatic variability. Eventually the old schemata give way to new possessive relations, for which no precedents exist. With a semantic example of distil and postal postpositions, he demonstrates that grammar is derived from semantics. Case
changes relate to semantic distinctions. Semantically motivated patterns ultimately become conventionalized, using extant grammatical modes.
Possession as a focal point for typological analysis functioned as a prism,
rather than a lens. It produced a broad range of interesting results, but limited
solutions. Most papers are substantive and some are imaginative.
Contributions are arranged well in a meaningful sequence. There are no abstracts,
or biographies of authors and no index. Many languages are cited and much
analytical description given. Research was diligent, but the writing
occasionally lacks focus. The full problem definition often does not emerge until
near the end of the argument when the reader has seen the details without knowing
the author's thinking. And he has occasionally found himself trapped in thickets
of descriptive analysis.
Functional and Lexical Functionalist approaches predominated and seem
appropriate for typological analysis. They define language structure and allow
for variation in individual languages. The papers show progress in analyzing the
usual enigmas, in/alienability relations, part/whole relations, kinship and body
part expressions. The hierarchical analysis of grammatical constituents has
provided explanations for interpretive anomalies, grammatical features, and grammaticalization processes. Both synchronic and diachronic analysts
contributed. The diachronic group faces an intimidating need for information to accomplish its tasks it seems.
Grammaticalization is the most common theme. Constructions are compared
across languages and across time. At least one contributor used asymmetry (5.).
The papers contribute some pieces to the puzzle, but language types don't emerge clearly. Another writer anticipates the development of networks
of grammatical patterns eventually coming together in a general picture of
language (7). The dream is distant; we need more knowledge.
Reductionist techniques including prototypes, frames, and case consolidations
deal with complexity. Reductionism is expedient. The techniques work for
restricted tasks but reduce semantic richness. Building upon a reductionist
scheme involves risk to data integrity that may or may not be calculable.
Case remains an important vehicle for communicating syntax-semantics notions.
Not all authors see the genitive as primarily possessive. Some understand genitives
as associating entities. Others are aware of its transpositional quality. Attributive possessives were discussed often. Origin relations were distinguished from
possessive genitives. Possessive gerundives were mentioned in passing.
Some analyses dealt with possessive verbs and their complements.
Some contributors used historical datives and examples from archaic languages to
sort problems in current usage. We still have imprecise predicate datives with
object alternation between experiencer and recipient (indirect object transferee) interpretations. There was some analysis of the complex constraints on these
constructions needed to understand possessive semantics. Some writers discussed localist and possessive interactions. Little was said about psychological content,
even in relation to experiencer interpretations although process was mentioned and
word order. Instrumental, comitative and benefactive cases were included in some
analyses. Greater emphasis on the lexicon highlights the weakness of our knowledge about the syntax-semantics interface. Lexical content and grammatical expression
seem to be used as they come to hand to explain interpretations. Although some
writers agreed that grammar derives from semantics, the semantic analysis was
Many contributors agreed that "possession" is derived from "location" constructs.
There are cognitive relations between physical property and physical location, but "possession" has other dimensions. The contributors agreed on the
diversity of "possession" but their working definitions were surprisingly similar.
They reflected linguistic tradition rather than usage. "Possession" is an open-ended concept that we are incapable of defining precisely, most agreed. Physical control
of something denotes "possession" for some as do "belonging: and "appurtenance"
for others. "Ownership" is commonly equated with "belonging", although many
people who "belong" would not describe themselves as "owned". "Ownership"
was considered by some as a subcategory of "possession", although "possession"
may also be "limited ownership". One might argue for degrees of "possession"
and of "ownership" some of which overlap. Retention is significant since all
physical possession is limited by time, and control is seldom, if ever, absolute.
It would be useful to know what elements are grammatically entailed by
ownership without possession as well as by possession without ownership.
Some papers discussed aspects of the semantics of "have"; others focussed on the
"get" verbs and other possessives. There was less information about verbs of
possession than one might expect. Verbs expressing acquisition, retention transfer,
and dispossession are all of interest. Some examples of problems of current interest involve the "switch, swap and exchange" group, anything useful in negotiating
transfers. There is too little understanding of the components "possessor"
and "possessee". If one gives knowledge, a skill, or life, he also still possesses it.
If he delivers a message, the recipient is possessor and the giver is dispossessed.
But what if the message is oral;do not both giver and receiver possess the
content? Although these ideas may be relegated to the pragmatic level for
interpretation in context, they still reflect our lack of knowledge about the
semantics of "possession" and the grammatical
ABOUT THE REVIEWER:
Judith P. Dick works on full text R. & D. at Software Mechanics
Object-Oriented Consultants in Ottawa. Her doctorate from the
University of Toronto included linguistics. Her interests include ontology,
semantic analysis, legal research and object-oriented development and