ABSTRACT Background Cross-linguistic studies on time reference in highly inflected languages have shown that tense inflection is particularly vulnerable in agrammatic speakers. According to the PAstDIscourseLInking Hypothesis (PADILIH), an asymmetry is predicted between past and non-past forms, due to the extra discourse linkage the former type imposes. Aims The present paper investigates whether Greek agrammatic speakers are able to correctly use tense markers with respect to the relevant reference point, analyzing data from three different production tasks to understand how performance is modulated by different methodologies. Methods & Procedures seven agrammatic speakers and a control group participated in three experimental tasks (a) an elicited picture description, (b) a semi-standardized interview and (c) a sentence production priming task. Outcomes & Results Different outcomes were elicited across different tasks. Agrammatic speakers tended to accurately use past tense forms when they could freely select the content of their narration, as in the case of the two narrative tasks (the elicited picture description, and the semi-standardized interview). However, the same participants experienced significant difficulties referring to past and future events when using a sentence production priming task. Consequently, the predictions of PADILIH are not fully supported by the Greek data, given that, in addition to past tense deficits, the future tense was also severely compromised. Conclusions Our data clearly suggest that language performance is affected by the processing demands placed on the patient’s linguistic system by the experimental task used. Moreover, future tense deficits in Greek are interpreted as difficulties in the processing of conversational implicatures.

中文翻译：

---

希腊文法中的时间参考和时态标记：来自叙事和句子生产启动任务的证据

摘要 背景 对高度屈折语言中时间参考的跨语言研究表明，时态屈折在不合语法的人中尤其脆弱。根据 PAstDIscourseLInking Hypothesis (PADILIH)，过去和非过去形式之间会出现不对称，这是由于前一种类型强加了额外的话语联系。目的 本文研究希腊语语法使用者是否能够正确使用时态标记与相关参考点相关，分析来自三个不同生产任务的数据以了解不同方法如何调节性能。方法和程序 七名不讲语法的人和一个对照组参与了三项实验任务 (a) 引出的图片描述，(b) 半标准化面试和 (c) 句子生成启动任务。结果和结果 不同的任务产生了不同的结果。当他们可以自由选择他们的叙述内容时，语法说话者倾向于准确地使用过去时形式，如在两个叙述任务（引出图片描述和半标准化采访）的情况下。然而，在使用句子生成启动任务时，相同的参与者在提及过去和未来事件时遇到了重大困难。因此，希腊数据并不完全支持 PADILIH 的预测，因为除了过去时的赤字外，未来时也严重受损。结论 我们的数据清楚地表明，语言表现受到所使用的实验任务对患者语言系统的处理需求的影响。此外，希腊语的未来时态缺陷被解释为处理会话含义的困难。