索引词及其语义——兼谈汉语中的“我”和“自己”

Semantics of Indexicals - with a Special Reference on Chinese‘wo' and‘ziji'

索引词不同于专名和限定描述语,因此需要一个专门的语义理论对其进行解释。本文首先介绍了索引词的经典解释理论——Kaplan([15])的二维语义论,然后考察了反驳Kaplan理论的四类反例:(a)"only"辖域中受宽松约束的"Ⅰ";(b)第一人称涉己信念报告;(c)转换索引词;(d)延指现象。综合这四类反例的最新研究成果,我们得出的结论是:这些反例不足以否定Kaplan关于索引词直接指称的论断,但Kaplan所否认的"怪兽"确实是存在的。基于此,我们指出需要进一步研究的问题,并在文末从索引词研究的角度探讨了汉语中的"我"和"自己"。

In his classic paper Demonstratives, Kaplan claims that （i） indexicals are directly referential; （ii） no monster exists. Although his theory is quite influential, it has problems dealing with these four kinds of cases： （a） sloppily bound I under only; （b） first-person de se belief reports; （c） shifted indexicals; （d） the phenomenon of deferred reference. In this paper, through our examination of the recent literature on these aspects, we find that none of these problematic cases is sufficient to refute Kaplan＇s claim on the rigidity of indexicals, though his report semantics does need refining. More specifically, Maier （[20]） have shown that High-Order Unification and relational analysis can save Kaplan＇s rigid I from case （a） and case （b）. Besides, given the fact discovered by Anand ＆ Nevins （[1]） that the properties of Shifted Together and No Intervening Binder are satisfied by shifted indexicals in Zazaki, one has good reasons to believe that the problem raised by shifted indexicals is due to Kaplan＇s report semantics, rather than the rigidity of indexicals. Furthermore, it is more convincing of Hunter＇s （[14]） explanation to the data involved in case （d） through the interaction between indexicals and certain operators than Nunberg＇s （[21]） descriptive interpretation of indexicals. In the end, we explore the differences between Chinese wo and ziji from the semantic perspective of indexicals. Our findings are as follows： Chinese wo, just like Enlgish I is an indexical with direct reference; Chinese ziji has distinct uses, though it may denote the speaker when used as sentential free; it is inappropriate to treat long distance zijis as shifted indexicals as proposed by Anand （[3]）.