The Discourse Function of the Japanese Dative Subject Construction in Written Text

ABSTRACT. This paper considers the discourse function of dative case marking on subjects in written Japanese texts (novels and newspaper articles) by examining transitivity components in discourse (Hopper and Thompson 1980). It has been shown that three parameters for the dative subject construction--participant type, the topicality nature of dative-marked phrases, and the predicate types including certain types of adjectives and nominals--do not exist arbitrarily at the sentence level, but rather co-exist within the discourse notion of backgrounding. This article shows that the Japanese dative case on subjects is best understood as a perspective marker. *

1. INTRODUCTION. It is well known that in many languages, the dative case often marks experiencers (Klaiman 1980, 1981; Verma and Mohanan 1990; Shibatani and Pardeshi 2001; Haspelmath 2001; Pardeshi 2004). Dative-marked nominals used as subjects are variably called dative subjects, experiencer subjects, oblique subjects, quirky subjects, indirect subjects, or non-canonically-marked core arguments. Due to the unique characteristics of the grammatical status of dative-marked subjects, the so-called dative subject construction (DSC) has been widely discussed under different theoretical frameworks for at least the last three decades.

Numerous articles in the syntactic literature have argued that the DSC is a transitive structure due to the syntactic properties associated with grammatical subjects (e.g. verb agreement and reflexive binding) (Gair 1990, Masica 1991, Kishimoto 2004). In the semantic/cognitive literature, on the contrary, the DSC is claimed to be motivated by semantic/cognitive factors, reflecting the nature of the agent/experiencer continuum. Cross-linguistically, a dative-marked subject has the experiencer role in a non-canonical construction, as opposed to a nominative-marked subject with an agent role in a canonical transitive construction. As such, the DSC is regarded as an intransitive structure typologically (Dixon 1994, Shibatani 1999, Shibatani 2001, Shibatani and Pardeshi 2001).

This paper proposes a new perspective for the issue of whether the DSC is transitive or intransitive by analyzing the actual usage of the DSC in terms of a continuum, allowing the possibility of an intermediate stage between transitive and intransitive. This way, languages can be located at a particular point on the transitive-intransitive scale as suggested in Haspelmath (2001) and Pardeshi (2004). (1) The distribution of the DSC in natural discourse within one language can be better accounted for using this approach. Another advantage of examining actual usage of the DSC is that we can dispense with out-of-context tests of syntactic properties, some of which give inconclusive results or are not even applicable to European languages (Haspelmath 2001:68). To elucidate this point, this paper will examine the DSC in discourse with an emphasis on the TRANSITIVITY components discussed in Hopper and Thompson (1980).

Hopper and Thompson (1980) claim that transitivity plays a central role in the understanding of clause structure in human language. According to them, the properties of a clause relevant to transitivity exist on a continuum along which various points cluster and tend to strongly co-occur. These properties include kinesis, aspect, punctuality, volitionality, affirmation, mode, agency, affectedness, and individuation. For instance, a comparison of sentences (1) and (2) with respect to the cluster components kinesis, aspect, punctuality, affectedness, and individuation, shows that (1) is higher in transitivity than (2):

(1) Jerry knocked Sam down.

(2) Jerry likes beers. Hopper and Thompson (1980: 253)

                             (1)                    (2)

Kinesis:                     action                 non-action

Aspect:                      telic                  atelic

Punctuality:                 punctual               non-punctual

Affectedness of object:      total                  not affected

Individuation of object:     high, referential,     non-individuated
                             animate, and
                             proper

Importantly, these properties are related to independent discourse notions such as FOREGROUND (main points of the discourse) and BACKGROUND (part of a discourse which does not immediately contribute to the speaker's goal but amplifies or comments on it). Drawing on examples from multiple languages and discourse analyses, Hopper and Thompson (1980) argue that high transitivity is likely to be associated with foregrounding the actual sequential events, while low transitivity is associated with backgrounding, scene-setting statements and evaluative commentary. For instance, in the foregrounded clauses in three different texts that Hopper and Thompson (1980) examined, there was frequent occurrence of two participants (high transitivity), while the majority of backgrounded clauses contained only one participant (low transitivity). Moreover, the foregrounded clauses in their data sets never include negation, while some negation occurs in the backgrounded clauses. This is understandable if one considers that negation can function to signal digression into a possible world.

Another important claim by Hopper and Thompson (1980) relevant to the present study is that sentence-level or sentence-internal accounts of morphosyntax have limitations. They argue that a complete theory of language must be based on an analysis of authentic discourse where individual sentences actually occur. Applying this line of reasoning, the present study predicts that the selected parameters for the DSC neither exist arbitrarily nor independently at the sentence level, but co-exist with discourse strategies used in real communication.

2. STUDIES OF JAPANESE DATIVE SUBJECT CONSTRUCTION. Like many Asian languages, some Japanese predicates (e.g. verbs of possession, verbs of capability, and adjectives of sensation, desire, and evaluation) exhibit a dative-nominative case-marking pattern, as shown in sentences (3) and (4) below. (2)

(3) Ken-ni eigo-ga hanas-e-ru (koto)
    Ken-DAT English-NOM speak-POT-PRES that
    '(that) Ken can speak English.' (Shibatani 1999: 49)

(4) Boku-ni-(wa) Ken-ga totemo niku-i
    I-DAT-(TOP) Ken-YON very hateful-PRES
    'To me, Ken is very hateful." (Shibatani 1999:48)

The grammatical status of the dative/ni-marked noun phrase and the nominative/ga-marked noun phrase has been debated. It has been argued that ni in examples (3) and (4) is a dative case marker, not a postposition, and its use is syntactically motivated (Kuno 1972; Shibatani 1977, 1978; Saito 1982; Takezawa 1987; Miyagawa 1989; Dubinsky 1992; Sadakane and Koizumi 1995; Kishimoto 2004). Kuno (1972), for instance, suggests that there are variations in the casemarking for the noun Ken as shown in (5): ga-marking or ni-marking. He argues that the dative-marked noun phrase is the result of an avoidance of ga-ga repetition. Shibatani (1977, 1978) argues syntax requires that at least one ga-marker occur in a non-embedded sentence.

(5) Ken-ga/ni eigo-ga hanas-e-ru (koto)
    Ken-NOM/DAT English-NON speak-POT-PRES that
    '(that) Ken can speak English.'

In contrast, semantically-based studies suggest that the DSC is motivated by semantic and cognitive factors (McGloin 1980; Watanabe 1984; Sugimoto 1986; Kabata and Rice 1997; Kabata 1998; Masuda 1999; Shibatani 1999, 2001: Kumashiro 2000; Shibatani and Pardeshi 2001 ; Pardeshi 2004; Arnett and Masuda 2004). For example, Kabata (1998) has suggested that the ga/ni alternation in (5) is better accounted for as a cognitive phenomenon, which reflects the nature of the agentive/cognitive continuum. By employing native speakers' grammatical judgments, it has been shown that lower transitivity elements such as negation or inchoative acts tend to go with ni-marking, white higher transitivity elements such as volitionality tend toward ga-marking (Kabata 1998, Masuda 1999).3 It is noteworthy that Shibatani (1999, 2001 ) and Shibatani and Pardeshi (2001) argue against syntactic approaches, instead proposing a semantically and cross-linguistically unified account. They claim that there are different degrees of Japanese subjecthood, and that the DSC is cross-linguistically closer to the double subject construction motivated by external possessors or possessor ascension constructions. (4) They argue that the ni-marked phrase of a dative subject predicate, functions as a cognizer on which a clause is less dependent and that the DSC is intransitive, representing a state of affairs that is not controllable.

To date, however, few linguists have attempted to investigate the DSC with naturally occurring discourse data. One of the few exceptions is Sadler (2007) who provides a diachronic and synchronic analysis based upon naturally occurring data. After examining texts from classical Japanese as well as spoken and written modern Japanese, Sadler discovered that the DSC occurs differently from text to text. It is hardly used in spoken language, but when it is used it is more likely to be used with third person referents. In written language, the DSC is used with either first-person referents or the main characters of a story who most frequently demonstrate subjective viewpoints. As shown in Ono and Thompson (2003), explicit mention of the first-person referent is not necessary in Japanese spoken discourse, unlike English, where an obligatory argument is needed. Thus, the rarity of the DSC in spoken discourse is not peculiar. Sadler (2007) showed that the rarity of the DSC in spoken discourse is due to a lack of overtly mentioned noun phrases with a particle in Japanese spoken discourse. In order to investigate the nature of the DSC in naturally occurring data in depth, the present study uses written discourse, as it is likely to provide more examples of the DSC.

Sadler's insightful findings are, however, limited by her choice of data--a single genre of written discourse, the novel. There is a possibility that other genres will exhibit other stylistic characteristics (Biber 2000). Thus it is crucial to examine whether or not the phenomenon explained in one type of written discourse is also present in other types of written discourse. Secondly, in Sadler's analysis, the narrative portion and the conversational portion were carefully compared. Four types of novels (first-person novels by male or female writers, and third-person novels by male or female writers) were included in an equal amount (1230 clauses for each type of novel). Kuroda (1973: 382) points out that the first-person and the third-person stories may follow a different paradigmatic schema of linguistic performance. Thus, it would be worthwhile to compare first-person stories with third-person stories and perform a quantitative analysis to see if the DSC is differently distributed.

The goals of the present study are two-fold. First is to examine whether the prototypical usage of the DSC is consistent or varies across different types of written discourse texts such as first-person novels, third-person novels and newspaper articles. The last genre has been added because, as Maynard (1996) argues, the purpose of writing newspaper articles is to convey one's view of current events. Newspaper data contain different types of ni-marked subjects since they include news stories in which editors and reporters express their points of view. This is especially apparent in voice, life advice or the editorial sections, where editors or reporters are expected to express views or opinions on current events and issues. Second, this present study demonstrates how prototypical usage of the DSC in each genre can be accounted for under a unified discourse account. As Hopper and Thompson (1980) argue, a complete theory of language should include attention to discourse factors. This paper aims to demonstrate the relevance of this point by examining selected parameters of the DSC in discourse.

3. DATA. There are two different genres of written data sources examined in this paper. …

To read the full text of this article and others like it, subscribe today!

Searching…