Japanese employs case markers to indicate the grammatical relationship between the NP and the predicate.  The particle 'o' marks an NP as the direct object as seen in example 1) taken from a Modern Japanese translation of Genji monogtari [The tale of Genji] written in the early 11th century:

 1) sono     senmyoo      o   yomiageru  no  ga
    that  official order  DO    read     NOM SUB
    '(a messenger) read the official order'  (20th century)

In Classical Japanese, however, the variable usage of 'o' (either o-marking or zero-marking) is observed. Examine the following example from the original text of Genji monogatari.

 2) sono     senmyoo     ___  yomu    nan
    that  official order      read  adverbial
    '( a messenger) read the official order' (11th century)

Example 2) expresses exactly the same story line, but the direct object "sono senmyoo" is not marked by 'o.'

Quantitative results from some previous studies seem to indicate that o-marking may have increased between the 8th century and the 10th centuries.  However, these results are not directly comparable since they are based on diverse genres of literary works (Matsuo 1944 and others). Furthermore, there are very few diachronic studies documenting the variability and the change of this particle after that period on until Modern Japanese.

There have also been suggestions that o-marking was correlated with discourse-level factors such as animacy, referentiality and importance of referents: human NPs, referential NPs, and major and important participants in discourse are more likely to be expressed with the particle 'o' (Matsuo 1944 and others).  These are, in fact, the factors which characterize Hopper and Thompson's "discourse-manipulability": discourse manipulable referents refer to those which have continual identity and importance in discourse (1984: 711).

Prior studies also mention that the use of 'o' correlates with the size of the NP and the presence of other elements between the NP and the predicate: 'o' is more likely to be used when the NP is complex, very short, or there are other elements intervening between the NP and the predicate (Miyagawa 1989 and others).  All of these factors seem to suggest that 'o' is used when clear segmentation of linguistic material is relevant (Ono, Thompson and Suzuki, soon to be published). This will be referred to as discourse-segmentation in this study.

The present study documents the variability and distributional  properties of the direct object marker 'o' in written Japanese in the past 900 years by quantitatively examining o-marking in Genji monogatari and its two available translations.

The results reveal that o-marking was strongly associated with the discourse-manipulability and discourse-segmentation in the 11th century. However, as o-marking gradually increased over the years (73% o-marked NPs in the 11th century text; 81% in the 18th century; 96% in the 20th century) to be fully grammaticized as the direct object marker by the 20th century, its correlation with these discourse factors weakened.

References:

Hopper, Paul J. and Sandra A. Thompson.  1984.  The discourse basis for lexical categories in universal grammar.  Language 60: 703-752.

Matsuo, Osamu.  1944.  Kyakugo hyooji no joshi o ni tsuite [The particle 'o' as the direct object marker].  Hashimoto-hakushi kanreki-kinen kokugogaku ronshuu.  Tokyo: Iwanami, 617-644.

Miyagawa, Shigeru.  1989.  Structure and case marking in Japanese.  San Diego, CA: Academic Press.

Ono, Tsuyoshi, Sandra A. Thompson, and Ryoko Suzuki.  To appear.  The pragmatic nature of the so-called subject marker 'ga' in Japanese: Evidence from conversation.  Discourse Studies.

Back to the Conference Schedule