Abstract: Whether inductive inference is domain general or content-dependent remains elusive in the literature. Most previous studies are conducted using blank-content reasoning method, which sets the precondition that inductive inference must be independent from special contents and thus fails to distinguish aforementioned two approaches.
Recent studies reveal that none of domain general models can explain all eleven forms of inductive inference and deductive inference is content-dependent, indicating that content might play an important role in regulating inductive inference. This view is also supported from an evolutionary psychology perspective. Inductive inferences could be a group of adaptations that are gained during the evolution process to address certain events critical to survival.
In the current study, we aim to address this issue by using a novel paradigm to explore the role of content on inductive inference. Two tasks including embrace-advantage and resist-disadvantage are utilized because they are closely related to human survival and reproduction. Each task contains two forms of reasoning. Subjects are asked to either estimate the general possibility of embrace-advantage or resist-disadvantage, or express their own will of action in case of embrace-advantage or resist-disadvantage. Thus, a total of four kinds of reasoning tasks were conducted including reasoning in embrace-advantage, action-will in embrace-advantage, reasoning in resist-disadvantage and action-will in resist-disadvantage. A blank-content test serves as control for these reasoning tasks.
The results demonstrate that embrace-advantage inference is markedly different from resist-disadvantage inference, indicating a potent content-dependent effect. Embrace-advantage shows under generalization whereas resist-disadvantage shows the trend of over generalization, and the embrace-advantage inference is under generalization. We reason that these two tasks are driven by two different psychological mechanisms. Embrace-advantage follows the strategy of sufficiency-necessity, as compared to the sole sufficiency strategy utilized by resist-disadvantage tasks. Such differences have fulfilled the constitutive standards of domain specificity and we conclude that inductive inference is based on domains.
One innovating feature of the current study is the use of artificial concepts to build up the reasoning propositions. Because content is an important independent variable, artificial concepts can avoid the confounding effect of previous knowledge on subjects’ responses. The results indicate that it is an excellent method to control extraneous variable in reasoning research.

Key words: inductive inference, content effect, domain specificity, evolutionary psychology