[Thanks to Tamina Stephenson for the tip]
Storto, Gianluca (2003) Possessives in Context: Issues in the Semantics of Possessive Constructions. PhD Dissertation. University of California, Los Angeles. (Professor Daniel Büring, Chair)
bq. Abstract
Possessive constructions seem amenable to conveying that a very heterogeneous range of relations hold between two entities (possessor and possessum). This interpretive flexibility has been accounted for by assuming that — excluding cases where the meaning of the possessive relation is determined by the semantics of the possessum (inherent interpretations) — ”the meaning of the possessive relation is entirely determined by contextual information (extrinsic interpretations).
Contra this assumption, it is shown that not all types of possessives license the unrestricted interpretive flexibility predicted by this model: only a proper subset of extrinsic interpretations are licensed by all types of possessives. In particular, it is shown that only definite and partitive possessives license an essentially unrestricted interpretive flexibility. And it is argued that the restricted interpretive flexibility that characterizes other types of possessives indicates that the meaning of the possessive relation is specified in the semantic composition of the possessive construction.
Two types of extrinsic interpretations are thus distinguished. It is proposed that this distinction is determined by a basic ambiguity of the syntactic construction that encodes the possessive relation between possessor and possessum: the meaning of the possessive relation can be specified entirely within this structure (control interpretations), or left unspecified (free interpretations). In control interpretations the meaning of the possessive is determined independently of its context of use: this meaning constrains the (pragmatic) uses that the possessive can be put to when uttered in context. In free interpretations the meaning of the possessive is determined by contextual information: the possessive relation is encoded by a free relational variable, whose value is contextually determined (unrestricted interpretive flexibility follows).
Finally, the restricted distribution of free interpretations is argued to show that referential pronouns do not constitute the paradigm for the interpretation of free variables in discourse. It is suggested that the distribution of free interpretations follows from the interaction between a general restriction on the assignment of contextually determined values to free variables (modeled on Heim’s (1982; 1983a) Novelty Condition) and the presuppositional requirements imposed by the (Fregean) semantics of the definite determiner on the predicate that embeds the variable encoding the possessive relation.
Let me make three brief comments:
{1} Yuri Lander’s Bibliography on possessives and related problems might be of some interest to those who read Storto’s dissertation.
{2} One interesting phenomenon is that in some languages it is the possessed that is marked morphologically and not the possessor (as in the genitive case), and body parts always exhibit a possessed form. This is an issue that leads to the analyses of a very ample field of data, with very rich material for any book.
{3} The abstract above is well writen, but I think it would be desireable if it had some examples to make the claims clearer to larger audiences.
June 4th, 2004, at 7:10 pm #