The verbal overshadowing effect in face memory : The conditions where it occurs, and its mechanisms

从记忆中口头描述目标面孔对随后的识别任务有什么影响，例如，在目击者的列队决定中？一些实验研究发现了干扰效应（言语遮蔽效应），而另一些则发现了促进效应或没有效应。本研究的目的是探讨影响言语描述对人脸识别效果的变量。基于BEAS模型（Itoh, 2005）进行的三项研究假设，当记忆的整体、非分析性方面占主导地位时，语言化会干扰人脸识别，而当记忆的部分和分析性方面占主导地位时，语言化会促进人脸记忆，而当记忆较差（例如，记忆保持时间较长）时，部分和分析性方面占主导地位。第一项研究通过对复合人脸的识别实验发现，长时间的记忆间隔使人脸记忆的部分和分析方面占据主导地位。研究还表明，这种倾向是由任务引发的一种心理环境造成的。前者的发现支持BEAS模型，而后者则不支持。进一步的讨论可能需要更多的数据。第二项研究表明，即使口头描述不是明确要求的，自愿的口头描述也会干扰面部记忆。我们现在正计划进行一项研究，以检验语言活动是否真的与这种现象有关。第三项研究检查了BEAS模型的假设，敦促整体，非分析或部分，分析处理的任务使用纳冯图。在这项研究中，使用了人脸以外的材料（云的图片和指纹）。虽然已经获得了一些可能支持模型假设的数据，但我们需要更多的数据，并且正在继续研究以使结论更加坚定。

What is the effect of verbally describing a target face from memory on succeeding recognition task, for example, in an eyewitness' line-up decision? Some experimental research found interfering effects (verbal overshadowing effects) whereas others found facilitating effects or no effects. The purpose of this research is to examine the variables that influence the effects of verbal description on face recognition. Three researches were conducted based on the BEAS model (Itoh, 2005) that hypothesized verbalization interfered with face recognition when the holistic, non-analytical aspects of memory were dominant whereas it facilitated face memory when the partial and analytical aspects were dominant, and the partial and analytic aspects were dominant when memory was poor, for example, because of a long retention interval.The first research revealed that a long retention interval made the partial and analytic aspects of face memory dominant through recognition experiments using compound faces. It also suggested that a kind of mental settings elicited by the task caused this tendency. The former finding supports the BEAS model whereas the latter does not. More data may be needed for further discussion. The second research suggested that voluntary verbal description interfered with face memory even though verbal description was not required explicitly. We are now planning a research to examine if verbal activity actually is involved in this phenomenon. The third research examined the hypotheses of the BEAS model urging holistic, non-analytic or partial, analytic processing by a task using Navon figures. In this research, materials other than faces (pictures of cloud and finger prints)were used. Although some data that may support the hypotheses of the model were obtained, we need more data and are continuing the research to make the conclusions firmer.