In English, which is familiar to us Japanese, possessive pronouns cannot co-occur with definite articles (e.g.*
the my book vs. my book). This is also the case with French (e.g.*
le mon livre vs. mon livre). But the obligatory or optional co-occurrence relations are observed in the other Romance languages; in many modern Romance languages (i.e. Italian, Provencal, Rumanian, Leonese, and Aragonese) their obligatory co-occurrences are required obligatorily (and almost obligatorily in Portuguese and Catalan) with exceptions of Castilian (i. e. standard Spanish) and French:
Ital.
la sua sedia (his chair), Prov.
lo mieus cavals (my horse), Rum.
profesorul meu (my teacher), Leo.
la to casa (your house), Arag.
la tuya casa (your house), Port.(o)
meu livro (my book)
livro meu (my book), Cat.
el meu llivre (my book) (el)
nostre llibre (our book), Cast.
mi libro (my book)
el libro mio (my book)
The well-formedness of the string
Det [+
Art, +
def] +
Adj [+
genitive] + N implies that the definite article and the possessive adjective are in syntagmatic relation, i.e. belong to two different syntactical categories, while its ill-formedness implies that they are in paradigmatic relation, i.e. belong to the same syntactic category.
But from the diachronic point of view, the co-occurrence relations observed frequently in Old Spanish and French are not found completely in Modern Castilian and French. The co-occurrence of definite articles with possessive adjectives may have been considered to be semantically redundant, because a semantic feature [+ definite] became intrinsic in the possessive adjectives. This semantic redundancy was reflected in syntax, i. e. the definite articles were deleted syntactically, which is explained by Definite article deletion rule in the framework of a transformational grammar, as specified in Langacker (1968) and Ariyoshi (1975).
Next we will examine the co-occurrence between demonstrative adjectives and definite articles. The string
Det [+
Art, +
def] +
Adj [+
dem] +
N can not be observed in any Romance language as well as in the other languages like English, which implies that the two syntactic constituents are in paradigmatic relation, i. e. belong to the same syntactic category. Perhaps, this is due to the historical fact that the definite articles in Romance languages were derived from the demonstratives in Classical Latin. Therefore, we my well assume that the two syntactic constituents are semantically and syntactically redundant not only from the synchronical but also from the diachronical point of view. It is interesting to note that here remains still a Latin influence.
On the other hand in Rumanian, Castilian, and Catalan the string
Det [+
Art, +
def] +
N+
Adj [+
dem] is grammatical: Rum.
omul acesta (that man), Cast.
el pariente aquel (that relative), Cat.
el pariente aquelle (that relative). This phenomenon is explained by Demonstrative adjective movement rule which leaves behind a definite article, as specified in Contreras (1968) and Ariyoshi (1975). From this it follows that such an absolute paradigmatic relation is actualized as a syntagmatic relation in the specific construction.
抄録全体を表示