Understanding Written and Pictured Information

这项研究是波特目前由国家科学基金会资助的研究项目的更新，将研究阅读、理解和记忆单词、句子和图片所涉及的非常快速的过程。这项工作的主要重点是处理的级别或阶段：处理的级别是什么，它们的时间进程是什么，它们是如何相互作用的？在阅读中，单词识别的第一阶段是(A)像模块理论所说的那样只对视觉单词本身作出反应，还是(B)像“交互”理论所说的那样受到先前句子语境的影响？非词错觉和字母迁移是反映词识别早期阶段的两种现象，如果它们表现出句子语境的影响，则支持(B)。“非客体”知觉的延伸正在计划之中。当一个句子被阅读时，可能会计算出几个层次的表征：句法结构，单词组合的字面意义，以及对文本上下文中意义的更广泛的解释。这些句子处理水平将随着它们随着时间的推移而变化，通过改变阅读速度和通过在呈现之后以几个间隔探测句子中的单词的识别来跟踪。在读完一句话后，立即“逐字”回忆是可能的，即使是对超过正常记忆范围的七个左右不相关的单词。这表明一个人对句子的记忆是有效率的。这项研究将检验这一假设，即该结构主要是基于意义的；回忆，就像自发产生的句子一样，从说话者想要表达的一系列想法开始，它依赖于动词来产生适当的句子结构。用来表达原句意思的词是从最近激活的词中挑选出来的，因此在正常情况下回忆的准确率很高。总而言之，这些研究将描绘出句子阅读、理解和回忆过程中的几个层次的加工过程，并将在某些情况下与图片加工相类似。这项研究将增加对阅读过程的理解，并可能有助于设计改进的阅读教学方法。

This research, a renewal of Potter's current NSF-supported research project, will study the very rapid processes involved in reading, comprehending, and remembering words, sentences, and pictures. The main focus of the work is on levels or stages of processing: What are the levels of processing, what is their time course, and how do they interact? I. In reading, is the first stage of word recognition (a) responsive only to the visual word itself, as modular theories hold, or (b) is it influenced by the previous sentence context, as "interactive" theories propose? The nonword illusion and letter migration are two phenomena that reflect early stages of word recognition; if they show an effect of sentence context, that will support (b). An extension to "nonobject" perception is planned. II. As a sentence is read, several levels of representation are presumably computed: a syntactic structure, the literal meaning of the word combination, and a broader interpretation of the meaning in the context of the text. These levels of sentence processing will be tracked as they evolve over time, by varying the rate of reading and by probing recognition of a word in the sentence at several intervals after presentation. III. Immediately after reading a sentence, "verbatim" recall is possible, even for long sentences that exceed the normal memory span of seven or so unrelated words. This shows that one's memory for a sentence is efficiently structured. The research will test the hypothesis that the structure is primarily meaning- based; recall, like spontaneous sentence production, starts from a set of ideas the speaker intends to express, and it relies on the verb to produce the appropriate sentence structure. The words used to express the ideas of the original sentence are selected from recently-activated words, resulting in high accuracy of recall under normal circumstances. Together, these studies will give a picture of several levels of processing during the reading, understanding, and recall of sentences, and will draw parallels to picture processing in certain cases. The research will increase understanding of the reading process and may help in the design of improved methods of reading instruction.