A study on the subjective property of kara : Corpus search and sentence processing experiment

DOI IR HANDLE Web Site Research Data 1 References Open Access

Bibliographic Information

Other Title
  • カラの主語性に関する研究 : コーパス検索および文処理実験
  • カラ ノ シュゴセイ ニ カンスル ケンキュウ : コーパス ケンサク オヨビ ブン ショリ ジッケン

Search this article

Description

The present study investigated the subjective property of kara using corpus search and a sentence processing experiment. Based on three different genres of corpora, newspapers, public announcements, and magazines taken from the Balanced Corpus of Contemporary Written Japanese (BCCWJ), the “noun + kara + verb” pattern was searched and classified according to the 11 different usages of kara. Kara used as subjects were counted as once out of 696 in newspapers, once out of 1,138 in public announcements, and 13 times out of 3,053 in magazines. Based on the frequency patterns of the 11 kara usages, entropy and redundancy were calculated in each corpus. These indexes showed similar values across the three corpora, suggesting that the kara use for subjects is rare. In addition to the corpus search, the present study also employed an experiment on the processing of sentences containing kara-marked subjects. More specifically, it was investigated whether kara-marked subjects share the same features as ga-marked subjects. In modern Japanese ga-marked subjects are generated inside the VP domain and then moved to the specifier position of the IP as [IP NP-ga [VP t NP-o V]] (c.f., Kitagawa, 1986; Kuroda, 1988; Hasegawa, 1999). In contrast, Ueda (2003) and Inoue (2002) claim kara-marked subjects must stay in the inside the domain of νP as [IP [νP NP-kara [VP NP-o + V]]]. By inserting a time adverb (e.g., yesterday, today, tomorrow) before and after the subject NP, sentence stimuli were created in two canonical orders of [IP Adv NP-ga [VP NP-o V]] and [IP NP-ga Adv [VP NP-o V]] for the subject NP-ga, and in canonical order [IP Adv [νP NP-kara [VP NP-o + V]]] and scrambled order [IP NP-kara Adv [νP t [VP NP-o + V ]]] for the subject NP-kara. Accordingly, the processing of each sentence was compared to determine relating features of ga- and kara-marked subjects. A 2 (NP-ga and NP-kara) × 2 (word order) ANOVA for reaction times and accuracies indicated the significant main effect for NP type, but not for word order; additionally, no interactions were found. As with ga-marked subjects, the findings reveal that kara-marked subjects are moved to the specifier position of the IP. Consequently, kara-marked subjects share the same subjective property with ga-marked subjects.

Journal

  • ことばの科学

    ことばの科学 28 71-90, 2014-12-05

    名古屋大学言語文化研究会

References(1)*help

See more

Keywords

Details 詳細情報について

Report a problem

Back to top