Scientific Rationality In Computer Simulations

The trend toward technicalization and computerization is increasingly evident in the advance of contemporary science,and one of the most typical examples of this is computer simulations.From the early use of nuclear physics and climate science in the1950 s to the spread of simulation methods in natural and engineering science,and the maturing of computational social science in the last two decades,almost all branches of scientific research have now benefited from the introduction of computer simulation.While this method has achieved great success,due both to the technical complexity of the method itself and to the flexibility of the simulation practice,it also brings us a series of issues in philosophy of science,including the possible functions of computer simulation,how should its role be located in the “methodological map”,how the knowledge it produces can be justified,etc.Such issues can be collectively referred to as problems of the methodological foundation in computer simulations,and they challenge our conventional understanding of scientific rationality to vary degrees.With this in mind,this dissertation,based on current states of science in practice,aims to analyze and interpret the methodological foundational issues of computer simulation in different dimensions,and ultimately to elucidate some distinctive features of scientific rationality in the age of computer simulation.The main content of the dissertation could be divided into three parts.The first part consists of Chapter 1,the second part covers Chapters 2 to 4,and the third part consists of Chapter 6.The first part focus on some conceptual problems concerning computer simulation methods.Due to the various applications of computer simulation in current scientific research,it is necessary to make some general remarks about the basic concepts involved.It is proposed to understand the concept of computer simulation from two points: replaceability and immateriality.Then I give an overview of the technicalhistorical background of computer simulation,its relationship with some neighbor concepts,and its types and functions in practice.In addition,this part also emphasizes that compared with context of traditional scientific methods,there is a distinctive tension between computer simulations and rationality in traditional scientific practices,which implies the necessity of methodological interpretation.In the second part,the methodological foundation of computer simulation is discussed in four different aspects,and respective interpretations and analyses are given.Chapter 2 is concerned with how computer simulations represent the target system.Through a case study on protein structure simulation,the various constituent elements involved in computer simulation are described,and it is pointed out that the simulation model is the key unit that bears the representation ability.The representational relationship in the simulation model cannot be explained in traditional representation theories,and a functional-realization-based approach is proposed.Chapter 3 turns to a special type of simulation,namely multiscale simulation,and gives an interpretation on its credibility.This chapter argues that multiscale simulations challenge many of the methodological values that are the default in ordinary modeling practice.and proposing a nomological concept of scale to reconcile this challenge.Chapter 4 is concerned with the problem of opacity(or epistemological opacity)in computer simulations.It begins with a review of the concept,sources,and types of opacity,then introduces interpretive techniques used in current scientific practices which aims to address the challenge of opacity.The effectiveness and limitations of such a strategy are discussed.Furthermore,this chapter also discuss how to defend the methodological rationality to use simulation methods when some kind of opacity is unavoidable.Chapter 5 discusses the unique significance on scientific realism of computer simulations,which shed new light on this long-time debate.Through the case of climate simulations and artificial organisms,this chapter suggests the scientific success based on the simulation method,prompting us to consider the realist significance of virtual entities in a weak sense.Subsequently,a modest realist account is put forward to illustrate the plausibility of this view.The third part returns to the initial concern about rational tension in computer simulations.For contemporary science,which is shaped by computer simulation methods,its scientific rationality is distinctive in three respects: the expansivity of the scientific horizon,the uncertainty of scientific knowledge,and the essentiality of practical constraints.Two other theoretical positions in recent literature are also discussed,that is,instrumentalism and non-anthropocentrism,and it is demonstrated that neither of these choices is the best way to grasp scientific rationality in computer simulations.Finally,the concluding remarks are both a look back and an outlook based on the main content of this dissertation.It suggests that computer simulations should be treated as part of the broader wave of scientific computerization,which is and will continue to change the image of science and is driving profound shifts in the meaning of various core concepts in philosophy of science.