Development of conditional reasoning: Mental models and information generation

条件推理涉及基于如果-那么前提进行演绎推理。它是高级推理的重要组成部分，理解它的本质和发展过程是认知心理学的一个关键问题。该研究计划是一个正在进行的计划的扩展，该计划将心理模型理论(Johnson-Laird&Byrne，1999；2002)与模型生成的显式信息处理描述(Markovits&Barrouillet，2002)相结合。这个模型的一个关键假设是，儿童和毫无准备的成年人的条件推理严重依赖于从长期记忆中提取信息的过程(Janveau-Brennan&Markovits，1999；Markovits&Quinn，2001)，这可以与更基本的概率推理(Markovits，Lortie Forgues，&Brunet，2010；Markovits&Thompson，2008)区分开来。重要的是，通过将信息检索的概念推广到信息生成的概念，这个模型最近已经扩展到抽象推理(Markovits，2007；Markovits&Lortie Forgues，出版社)。具体地说，我们提供了证据，证明抽象推理是通过在与事实相反的背景下生成关于潜在备选方案的信息的能力来促进的。这种备选方案生成的概念提供了一个更连贯的上下文，在其中可以理解不同抽象级别的推理能力的发展。拟议的计划将以下列方式扩展这些结果。主要的重点将集中在理解备选方案生成和推理之间的关系中的发展进程。6至15岁的儿童将被给予适合年龄的推理任务，以及替代生成任务，使用实验和相关设计。另一组研究将通过使用启动和错误识别任务来检验或多或少有能力的推理者使用信息检索的方式。

Conditional reasoning involves making deductive inferences based on if-then premises. It is a key component of advanced reasoning, and understanding its nature and developmental course is a key problem in cognitive psychology. The proposed program of research is an extension of an on-going program that has combined mental model theory (Johnson-Laird & Byrne, 1999; 2002) with an explicit information processing account of model generation (Markovits & Barrouillet, 2002). A key postulate of this model is that conditional reasoning in children, and in unprepared adults, is heavily dependent on processes involved in information retrieval from long term memory (Janveau-Brennan & Markovits, 1999; Markovits & Quinn, 2001), which can be distinguished from more basic probabilistic reasoning (Markovits, Lortie Forgues, & Brunet, 2010; Markovits & Thompson, 2008). Importantly, this model has recently been extended to abstract reasoning (Markovits, 2007; Markovits & Lortie Forgues, in press) by generalizing the notion of information retrieval to that of information generation. Specifically, we have provided evidence that abstract reasoning is facilitated by the ability to generate information about potential alternatives within a contrary-to-fact context. This notion of alternatives generation provides a more coherent context within which the development of reasoning competencies of varying levels of abstraction can be understood. The proposed program will extend these results in the following ways. The major thrust will focus on understanding the developmental progression in the relation between alternatives generation and reasoning. Children at ages between 6 and 15 will be given age-appropriate reasoning tasks, and alternatives generation tasks, using both an experimental and correlational design. Another set of studies will examine the way that information retrieval is used by more or less competent reasoners by using priming and false recognition tasks.