Division of English Translation and Interpreting, Department of Western Languages and Literature, Faculty of Arts and Sciences, Kırıkkale University, 71450, Kirikkale, Turkey.
Institute of Cognitive Science, Osnabrück University, Osnabrück, Germany.
J Psycholinguist Res. 2020 Aug;49(4):511-539. doi: 10.1007/s10936-020-09698-4.
The goal of this study is to investigate whether or not Turkish has a particular preference in ambiguity resolution of relative clause attachment when two possible NPs are available as in "Someone shot [NP1 the servant] of [NP2 the actress] [RC who was on the balcony]ˮ. The relevant literature has showed that RC attachment preferences-whether NP1 or NP2-vary across languages, which results in contradictory evidence if some universal processing principle is assumed. Turkish differs typologically from English and other European languages in the construction of RC using complex "genitive-possessor" construction. In order to make a valid cross-linguistic comparison it is therefore of particular importance to carefully control potential extraneous factors which might obstruct true attachment preferences-if they exist. The present study, which controls various confounding factors, reveals that both syntactic and non-syntactic factors should be taken into consideration when constructing the stimulus sentences for testing attachment ambiguity resolution. Specifically, we propose that the semantic relations (e.g. part-whole relations) between the noun phrases of the genitive-possessive construction and the semantic associations with the proximal as well as with the distal predicate play a key role in the attachment preferences of monolingual Turkish native speakers in this type of ambiguous sentences. When these extraneous factors were controlled, no preference was observed.
本研究旨在探讨在“某人枪击了[NP1 女演员的仆人] [RC 他在阳台上]”这样的存在两个可能的 NP 时，土耳其语是否对关系从句附着的歧义消解有特殊偏好。相关文献表明，RC 附着偏好（无论是 NP1 还是 NP2）在不同语言之间存在差异，如果假设存在普遍的处理原则，则会得出相互矛盾的证据。土耳其语在使用复杂的“属格-所有格”结构构造 RC 方面与英语和其他欧洲语言在类型上存在差异。因此，为了进行有效的跨语言比较，仔细控制可能存在的阻碍真实附着偏好的额外因素非常重要。本研究控制了各种混杂因素，揭示了在构建用于测试附着歧义消解的刺激句子时，应考虑句法和非句法因素。具体来说，我们提出属格所有格结构中的名词短语之间的语义关系（例如部分-整体关系）以及与近端和远端谓词的语义关联在这种类型的歧义句中对单语土耳其语母语者的附着偏好起着关键作用。当控制了这些额外因素时，没有观察到偏好。
