Eye movements data for The effects of restrictiveness on relative clause processing in Farsi

Description We investigated whether there is any processing asymmetry between Farsi object and subject relative clauses and whether restrictiveness affects processing relative clauses by L1 Farsi speakers. To this end, with an eye-tracking method, we tested the effect of Clause Type and Restrictiveness in Farsi relative clauses. Forty-two Iranian Farsi speaking students at the University of Groningen participated in the experiment voluntarily. They were tested individually in the eye lab at the University of Groningen. Eye movements were recorded using a Tobii 1750 Eye tracker, interfaced with a laptop. The sampling rate was 50 HZ. The materials for this experiment consisted of 36 items, each consisting of four sentence versions. Initially, the items were adapted from Staub (2010) and Staub et al. (2017). For the present experiment, the items were translated into Farsi carefully and non-restrictive versions were added to the material. We divided the Farsi relative clause sentences into six regions: the antecedent, complementizer که /ke/, embedded noun phrase in the relative clause, embedded verb phrase in the relative clause, pre-matrix verb region, and matrix verb. For each region, four reading time measures of first-pass reading time (FP) or gaze duration, First-pass total reading time, Regression path duration (RPD), and Total reading time (TRT) were compared. Statistical analysis was conducted using linear mixed-effects models of the reading times. We used R (R Core Team, 2019) and the lme4 package (Bates et al., 2015) to predict the times of the eye movement measurements (first-pass, first-pass total, total reading time, regression path duration) based on Clause Type (subject versus object) and Restrictiveness (Restrictive versus Non-restrictive). As random effects, we included intercepts for subjects and items. We used treatment coding: The subject relative was the baseline for Clause Type and the restrictive condition was the baseline for Restrictiveness. First, we created a model with the main effects and then compared it with the model with the interaction. In the models, the subject relative was the baseline for Clause Type and the restrictive condition was the baseline for Restrictiveness. In the case of an interaction, an extra analysis was run to test the effect of Clause Type in the non-restrictive condition by changing the baseline for Restrictiveness into the non-restrictive condition.Description 我们探讨了以波斯语（Farsi）为母语的使用者在处理宾语关系从句与主语关系从句时是否存在加工不对称性，以及限制性是否会影响其对关系从句（relative clauses）的加工。为此，我们采用眼动追踪（eye-tracking）方法，检验了从句类型与限制性对波斯语关系从句加工的影响。 42名就读于格罗宁根大学的伊朗籍波斯语母语学生自愿参与了本实验。实验在格罗宁根大学眼动实验室单独进行。眼动数据通过与笔记本电脑连接的Tobii 1750眼动仪（Tobii 1750 Eye tracker）进行记录，采样率为50 HZ。本实验的刺激材料包含36组项目，每组项目对应4个句子版本。实验项目最初改编自Staub（2010）与Staub等人（2017）的研究，后续我们将其严谨翻译为波斯语，并补充了非限制性版本的材料。 我们将波斯语关系从句句子划分为六个区域：先行词区域、补语标记 که /ke/ 区域、关系从句内嵌名词短语区域、关系从句内嵌动词短语区域、主句前动词区域以及主句动词区域。 针对每个区域，我们对比了四种阅读时长远指标：首次注视时间（first-pass reading time, FP）即凝视时长、首次总注视时间、回归路径时长（Regression path duration, RPD）以及总阅读时间（Total reading time, TRT）。阅读时长的统计分析采用线性混合效应模型。我们使用R语言（R Core Team, 2019）及lme4扩展包（Bates et al., 2015），基于从句类型（主语关系从句与宾语关系从句）和限制性（限制性与非限制性）来预测各项眼动指标的时长，包括首次注视时间、首次总注视时间、总阅读时间与回归路径时长。 作为随机效应，我们纳入了被试与项目的截距项。我们采用处理编码：将主语关系从句设为从句类型的基线水平，将限制性条件设为限制性的基线水平。首先构建包含主效应的模型，随后将其与包含交互效应的模型进行比较。若存在交互效应，则额外开展分析，通过将限制性的基线调整为非限制性条件，检验非限制性条件下从句类型的效应。