<< >> TOC
So far I have
concentrated on the similarities between proper names and common nouns. The free
conversion possibilities between the two types of nominals makes a categorial
distinction seem unlikely and furthermore lead to the assumption of <<e,
t>, t> as base type for all nominals. Nevertheless, there are some
characteristics that do seem to set apart proper names. Consider the paradigms
in (20) and (21):
(20) a. There is/are a man / men / few men / two men in the garden.
b. *There is/are John / the man / everybody / most of the men in the garden.
(21) a. We were sitting at the table when – suddenly - a ghost / two ghosts / John appeared.
b. *We were sitting at the table when - suddenly – the ghost / everybody appeared.
Paradigm (20) exhibits what used to be called the
definiteness effect, since it was taken to distinguish between definite and
indefinite nominals. Enç (1991) showed this to be a misnomer: the distribution
does not depend on familiarity/novelty (i.e. definiteness in Heim’s
framework), but rather on specificity. Thus, the syntactic environment in (20)
allows the occurrence of nonspecifics but not of specifics. Proper names pattern
here with specifics, i.e. nominals that require a previously established
discourse referent for either identification (the man) or comparison (most of
the men). This is expected, since proper names are usually taken to be definite
and thus by extension specific (cf. Enç 1991: 9).
(21), on the other hand, is supposed to provide a discourse
environment in which discourse referents are newly introduced. Proper names
pattern here with indefinites, i.e. nominals that require that the discourse
referent they introduce be novel.
Proper names thus exhibit the puzzling characteristic
of being both definite and indefinite, depending on the context. I believe that
the puzzle’s solution lies in the characterization of the syntactic
requirements imposed by each of the constructions: assume that each of the
constructions is sensitive to overt markings. One could argue then that (20)
requires nominals which are overtly marked as nonspecific, wheras (21) rules out
the occurrence of nominals overtly marked for definiteness. The shifty behavior
of proper names could then be attributed to underspecification. If proper names
carry neither definiteness nor specificity features (or the corresponding
indices in Enç’s definition), syntactic constraints that are specified for
overt features will simply not apply to them. Thus, the semantics of proper
names is interdependent with one’s view on the syntactic encoding of
definiteness and specificity.
<< >> TOC
© Philipp Strazny 1998