Implicit Analogical Inferences In Text Comprehension

Inference, an advanced cognitive processing, is often used to explore the unknown world. Analogical inference plays an important role in understanding unknown fields. Cognitive psychologists usually make use of these materials such as stories, short passages, etc., to study analogy in a variety of ways. A large number of studies on text comprehension suggest people may make predication and inference of unknown information during reading. Psycholinguists and cognitive psychologists have done many studies of predicative inference, bridging inference and inductive inference, but they didn't pay more attention to analogical inference in text reading. What's more, during experimentations on analogy, participants were often aware how they implemented with the analogical inference, therefore unconscious analogical inference (implicit analogical inference) never attracted most psychologists'attentions. So, we used the following methods, such as fixed windows technique, self-pacing reading and so forth, to explore the implicit analogical inference in text comprehension. At last we discussed the results of the experiments in terms of models of analogical reasoning and text comprehension.The research includes three experiments, all of which are programmed with Inquisit 2.0. Senior middle school students in grade 2 conduct the experiments with computers. The aim of this study is to examine whether participants'understanding and interpreting of the passages being read is affected by the passages that both were previously read and have similar relational structures with the ones being read.Participants are requested to read some short passages in every experiment. The target passages are similar to the base ones in relational structure in some points. In Experiment 1, participants read some short passages, some of which have similar relational structures. The goal is to check whether the participants map unconsciously the structures of base passages into the ones of target passages during reading and make predication of relational facts of the target passages. In Experiment 2 and Experiment 3, we try to rule out some other possible interpretations. In Experiment 2 we revise the base passages in Experiment 1 and explore the role of the structures of the base passages and the semantic activation about the base passages in long-term memory in understanding and making inferences of the target passages under the conditions of the similar structures between the base passages and the target ones. In Experiment 1 and Experiment 2, participants judge the items about the passages after reading, therefore when they understand the target passages and draw inferences, it is more likely to be judging the items during reading rather than after reading. The fixed windows technique and self-pacing reading time method are used in Experiment 3 so that the computer can record the reading times of every sentence and monitor the online processes of the target passages. Thus, the moment when participants make inferences to the target passages can be affirmed.The results are as follows:1. The passages read by the readers early can affect the understanding of the passages being read subsequently.2. The readers construct two representations in their minds, and map common relational structures of two representations into the ambiguous places in the target passages to draw analogical inference. However, they aren't conscious of this process.3. Under the conditions of the similar structures between the base passages and the target ones, the semantic information activated by the target passages doesn't have a strong impact on the understanding of the target passages.4. Even the readers don't know the purpose of reading the passages previously, they can also understand the passages and make online interpretation. That's to say, when the readers are reading the passage unintentionally, they can still construct the meaning of the passages and make the appropriate inferences according to their own knowledge and to information provided by the previous passages.The results of this study suggest that the understanding of the passages being read can be affected by the ones read previously and this process doesn't need for elaborate processing or controlled processing, so this kind of understanding is an unconscious processing.In summary, in three experiments of this study we provide evidence for implicit analogical inference during text reading. The Resonance Model in the field of text comprehension and MAC/FAC in the study of analogy can give satisfactory explanations to our findings.