中文摘要：

采用Maddox，Glass，O’Brien，Filoteo和Ashby（2010）的标签转换范式，探究在基于规则的类别学习中，对同一类刺激分别建立对应的言语标签（即形成刺激的类别名称）与建立对应的动作标签（即形成刺激的类别反应）的情况下，两种类别标签的表征状况。实验l结果表明，在按随机顺序学习两类标签情境中，个体的建构方式遵从链条加工模型，即依从刺激．言语标签．动作标签的顺序进行建构；实验2结果进一步发现，这种言语标签优先建构的链条加工结构具有较强的稳定性，不会受到两类标签学习方式的影响。

英文摘要：

There are two major models account for the mechanism of rule-based category learning, including single association model and two associations＇ model. The former claims that a stimulus-to-label association associates the stimuli and category labels directly in the rule-based classification. While the latter holds that rule-based classification as a kind of implicit learning, is affected by both a stimulus-to-label association and a label- to-response association. The two kinds of associations were also called verbal label and response label respectively. Previous researchers found that the construction of the two kinds of labels follows a stimulus- to-label-to-response pattern, like chain processing, when people learn verbal and response label successively. To our knowledge, the question is whether this construction order of verbal label and response label was artificially arranged by these former researchers, since participants were asked to complete the response label firstly according to stimulus＇s verbal label which needs no response. In other words, it is still unknown that whether the learning order of the two labels would affect the way of the construction of two labels. The current study conducted two experiments to investigate whether there would be a stimulus-to-response pattern, like parallel processing, when participants learn the two labels randomly or reversely. Using the standard label conversion paradigm （Maddox, Glass et al. 2010）, experiment 1 explored the question whether the main way of the construction of the two labels is chain processing or parallel processing when the learning order of verbal label and response label was random. Experiment 2 further investigated whether the reverse learning order of the two labels would change the way of the construction of them. Experiment 1 found that the construction of the labels follow the chain of stimulus-verbal label-response label when the learning order of two labels was random. Experiment 2 also confirmed this chain processing pattem ev