【作者】 王剑锋；

【导师】 徐飞；

【作者基本信息】 中国科学技术大学 ， 科技哲学， 2014， 博士

【摘要】 认识的不确定性是科学活动的基本特征之一。随着科学在现实社会应用的日益普遍,科学认识的不确定性对社会的影响愈发深远。因此,研究科学认识的不确定性、以及由此而产生的风险,对于全面准确理解科学的本质,促进科学的合理应用,无疑具有重要的理论价值和实践意义。本文引入风险的概念来理解科学认识的不确定性。科学认识的风险包含科学认识的不确定性以及科学家对这些不确定性所导致的(认识的和非认识的)后果的价值评价。本文从实在论、认识论、理性与合理性评价等三方面对科学认识的风险进行了系统的阐述和分析。本文采用假设实在论和自然主义实在论的本体论预设,以胡克(CliffordHooker)所提出的调节系统的形而上学理论框架作为对这一实在论预设的具体阐述。在调节系统的理论框架中,客观世界和认识主体都被阐述为调节系统：自然被阐述为一个处于开放上升的进化过程的调节系统,包括哲学和科学在内的人类认识被阐述为这一系统的延伸。基于以上预设和阐述,本文提出对科学认识的风险的三种阐述：第一,科学认识的风险与生物生存和进化的风险存在形式上的类似,我们可以在比喻的意义上用科学认识的风险指涉生物过程的风险在科学认识中的对应物；第二,科学认识的风险是生物过程的风险的延伸,认识过程与生物过程的相互联系是我们将风险的含义从生物过程延伸至认识过程的本体论依据；第三,科学认识的风险与生物过程的风险统一在更高阶层的形而上学中,这种风险是调节系统的动态过程的风险。对科学认识的风险的三种阐述逐渐抽象,最终统一在一种整体的形而上学中,调节系统是对这一整体的形而上学的一种阐述。科学认识是科学家的冒险决定。科学家在做决定时需要考虑决定的可能后果。进一步地,科学家能对决定的风险进行预见,这种对不利后果的事先评估的能力是科学认识以及智慧行为的重要特征之一。这种预见能力一方面来自科学认识的成功所带来的对世界的理解的深入或普遍,另一方面也来自从认识的错误中的学习——确认和定位认识的错误,进而获得避免错误的认识方法；两方面的过程相互促进,共同构成了对科学认识发展的双重驱动。这样一种双重驱动的过程被阐述为一种“自我指导的预见学习”(self-directed anticipative learning)的认识机制。科学家通过自我指导的预见学习,在不断取得认识的成功和提高认识能力的过程中获得了对潜在的错误的越来越好的预见,他们预见潜在错误的能力也逐渐提高。认识的错误与成功共同构成了对科学认识的双重驱动,使科学家在不断改善认识表现的同时也提高了自身的认识能力。本文将冒险的启发式、非形式化的判断理解为理性的基本要素,这些要素进一步揭示了理性对情境和心理结构的依赖,并且构成了科学家在应对认识的不确定性时的冒险决定、以及自我指导的预见学习的基础。根据这种对理性的理解,科学认识的不确定性和科学家的价值判断不再是对理性的否定,相反,冒险是理性(尤其是科学认识)的普遍特征之一。科学家冒险的、非形式化的决定、科学家对认识的不确定性及其后果的价值判断、科学家自我指导的学习等活动都包含科学家对认识的工具和目的的评估、权衡和选择,体现了认识的合规律性与合目的性的统一,因此可以被评价为合理的。价值在科学认识中具有合理的间接作用,对科学认识风险的可接受水平的判定就是价值的一种合理的间接作用。更多还原

【Abstract】 Epistemic uncertainty is among the basic characters of science, the applications of which prevailing the society along with their far-reaching influences. Thus signify the studies of the epistemic uncertainty of science as well as the resulted risk in intellectual and practical searching for comprehensive understanding and legitimate use of science.The concept of risk has been resorted to characterize the epistemic uncertainty of science as the epistemic risk of science, the latter constituted by the epistemic uncertainty and scientists’value judgments of its epistemic and non-epistemic outcomes. Theorizing the epistemic risk of science in the aspects of realism, epistemology, reason and rationality promises a systematic characterization of it.Hypothetical realism and naturalistic realism constitute our presupposition for the epistemic risk of science, a specification of which is drawn from Clifford Alan Hooker’s theoretical framework of regulatory systems. The world and subjects are both descripted as regulatory systems, the former as one such system in an open-ended evolutionary process with science and philosophy as its extensions. The ongoing presupposition and specification underline the following characterizations of the epistemic risk of science:firstly, there are formal analogies between the risks in biological dynamics and the epistemic risk of science; secondly, the epistemic risk of science can be modeled as the extension of the biological risks; thirdly, those two kinds of risks are unified in one metaphysics with regulatory systems framework among its characterizations.A risky decision account of science is proposed to point out that scientists must consider the outcomes of their decisions in an anticipative manner. The ability to evaluate the harmful outcomes anticipatively is seen as a key feature of science and of intelligent behaviors. With deeper and more extensive conceptions of the nature of the world comes the anticipative ability, which has another source in learning from errors. The error-avoiding methodologies can be learned from successful confirmation and location of errors. An account of error and success driven self-directed anticipative learning process is developed to model the learning process in which scientists get better anticipations of potential errors and more anticipative power in their successes as well as their growing power for success. Science is dual-driven by success and error with scientists getting more power from their improved performances.Risk-taking heuristics, non-formal judgments, context-dependence and psychological structure dependence are essential features of reason. Those features support scientists’ risky decisions and self-directed learning. This conception of reason rationalizes the epistemic uncertainty of science and scientists’ value judgments. What is risky is reason and science generally.The evaluation, weighing and choice of means and ends prevails in scientists’ risky non-formal decisions, value judgments on the epistemic uncertainty of science and on its outcomes, as well as their self-directed anticipative learning. Those activities can be evaluated rational on behalf of their conformity with law and purpose. Value can play rational roles in science, one of which is the decision of the acceptable level of epistemic risk of science.更多还原