The role of lexical factors in short-term ordered recall

记忆最基本的功能之一就是让过去支持和指导我们现在与世界的互动。从历史上看，当前的认知操作是通过调用一个专门的记忆系统来建模的，比如工作记忆系统（Baddeley, 1986），而理论化在很大程度上与记忆的其他领域所做的工作脱节。然而，在过去的二十年里，词汇-语义属性对即时记忆的贡献已经得到了很好的证明（例如，Thorn & Page, 2009）。最初，关于词汇语义因素影响的假设表明，它们的影响仅限于回忆过程和项目信息。然而最近，另一类模特正变得越来越有影响力。这些模型表明，语义和词汇表征与短期回忆的关系比之前提出的更密切，并且可以影响项目和订单信息。在这些模型的背景下，我们最近提出了一个名为ANet的理论框架，用于激活网络视图（Poirier, Saint-Aubin, maair, Tehan, & Tolan，出版）。目前的研究计划旨在对ANet框架进行系统的调查。更具体地说，与我的学生一起，通过四条研究线，我们将测试ANet在订单信息回忆方面的关键预测。首先，词汇语义表征的激活水平将被操纵以诱导ANet预测的特定顺序错误。对于第二条研究路线，我们将通过Acheson等人（2011）最近提出的一项新任务，通过干扰词汇语义表征来测试词汇语义表征对顺序信息回忆的影响。在研究的第三条线中，我们将使用向后回忆过程来进一步测试ANet，并生成用于评估该范式下的性能的记忆模型的关键数据。最后，我们将通过使用个体差异方法寻求对ANet的收敛支持。该研究项目具有很强的预期理论意义，因为它将为评估ANet和该领域的替代账户提供关键数据，并为开发将即时记忆与长期表征联系起来的完整模型提供基础。拟议的研究计划也是为培养高素质人才而量身定制的，允许各级培训的学生参与其中。

One of the most fundamental function that memory performs is to allow the past to support and guide our present interactions with the world. Historically, current cognitive operations were modeled by calling upon a dedicated memory system, like the working memory system (Baddeley, 1986), and theorizing was largely divorced from work done in other areas of memory. However, over the last two decades, the contribution of lexical-semantic properties of words to immediate memory has been well documented (e.g., Thorn & Page, 2009). Originally, the hypotheses called upon to account for the influence of lexical-semantic factors suggested that their influence was limited to recall process and to item information. Recently, however, another class of models is becoming increasingly influential. Those models suggest that semantic and lexical representations are more closely related to short-term recall than suggested earlier and can affect both item and order information. In the context of those models, we recently proposed a theoretical framework called ANet for Activated Network view (Poirier, Saint-Aubin, Mair, Tehan, & Tolan, in press). The current research program aims to conduct a systematic investigation of the ANet framework. More specifically, with my students, through four lines of research, we will test key predictions of ANet with regards to order information recall. First, the level of activation of lexical-semantic representations will be manipulated to induce specific order errors as predicted by ANet. For the second line of research, we will test the impact of lexical-semantic representations on order information recall by interfering with them through a novel task recently proposed by Acheson et al. (2011). In the third line of research, we will use a backward recall procedure to further test ANet and to generate critical data for evaluating memory models accounting for performance in this paradigm. Finally, we will seek converging support for ANet by using an individual differences approach. This research program has strong anticipated theoretical significance, as it will produce critical data for evaluating ANet and alternative accounts in the field as well as providing the foundation for developing a full model linking immediate memory to long-term representations. The proposed research program is also custom tailored for training highly qualified personnel by allowing students at all levels of training to take part in it.