ISSN 0439-755X
CN 11-1911/B

Acta Psychologica Sinica ›› 2015, Vol. 47 ›› Issue (12): 1445-1453.doi: 10.3724/SP.J.1041.2015.01445

Previous Articles     Next Articles

Information Amount and Obviousness Influence Hypothesis Generation

LIU Zhiya; ZHENG Chen   

  1. (Center for Studies of Psychological Application/School of Psychology, South China Normal University, Guangzhou 510631, China)
  • Received:2014-10-16 Published:2015-12-25 Online:2015-12-25
  • Contact: LIU Zhiya, E-mail:


 This study focuses on the availability of rule learning. Cherubini, Castelvecchio & Cherubini (2005); Cherubini, Rusconi, Russo, Di Bari, & Sacchi (2010) confirmed that the availability of rule learning was influenced by the information amount of the rule. Information amount was explained by how many examples could be covered by a rule. For a rule, the more number of examples could be converted, the less information amount would have. For example, in 2-4-6 task, the information amount in the rule of “even number increase” is 1/n and in the rule of “the third number is the sum of other two” is 1/n2. The information amount theory suggests that a rule with higher information amount is generated more easily than a lower one. However, Some researches (Barsalou,1982; Rips,1989; Medin, Lynch, Coley, & Atran,1997; Shafto, Coley, & Baldwin,2007; Guhe, Pease, & Smail,2011) showed that rule learning would be impacted by the information background of participants.

In this paper, information background was defined as the obviousness of the rule. Inspired by dual process model of deductive reasoning (Evans, 2003, 2010; Sloman, 1996; Barrouillet, 2011), This study assumed that the cognitive process of rule learning might be impacted by the information amount and obviousness both. Dual process model suggested that there were two independent cognitive systems, system 1 was usually described as unconscious and automatic; the system 2 was inherently conscious and controlled. This paper assumed that there might be two independent cognitive systems that manipulating rule learning process. This hypothesis was tested by experiment 1. Additionally, Ashby (1998) also suggested that there were two kinds of category learning. One was the rule-base category learning, the other was information integration. In the case of rule-based learning, participants could abstract a linguistic and explicit rule from materials, while they cannot discover an explicit rule but still can classify materials when doing information integration tasks, which seems to be implicit. This article assume that rule learning process may also conducted by both explicit and implicit systems and which system would be adopted may related to the information amount and obviousness of rules. Experiment 2 was designed to test this hypothesis. With 70 college students' participated, a revised 2-4-6 task was used to examine our hypothesis. Both experiments were presented by Psychtoolbox 3.0 on MATLAB.

Experiment 1 found that there were two independent factors, the information amount and the obviousness of the rule, significantly influence availability of rule learning. Experiment 2 is the same as experiment 1 except a rule description between every block of learning. The result of experiment 2 indicated that rules with high information amount and obviousness are more easier to be learned and expressed, while rules of low information amount combine with less obvious could be learned either but hardly be expressed clearly. These results consist with the dual process model in deductive reasoning and reveal that the rules with high amount information and obviousness are processing by an explicit rule system, and with lower amount information and less obviousness are processing by an implicit rule system.

Key words: learning, hypothesis generation, reasoning, implicit learning