Abstract

Most of the recent work in chaos theory has been the development of data analysis tools for analyzing chaotic data. It is based upon the results of the application of these tools that many researchers have made claims that such phenomena as heartbeats, planetary orbits, and chemical reactions are chaotic. ;The first part of my dissertation is concerned with investigating the standard methods that are used to determine whether a system is chaotic, and the requirements of these methods. I begin by explaining chaos theory, and providing a generally accepted definition of chaos. I then explain the current methods that have been developed for experimentally studying chaotic systems. With this in mind, I turn to a more theoretical investigation of the reliability of these methods. Within the framework of formal learning theory, I discuss the conditions under which a scientist can reliably determine whether a system is chaotic. ;Generalizing this problem leads to the philosophical center of my dissertation: an investigation that relates how the conceptions of reliability of the scientists who work with chaos theory fit into a larger philosophical picture of reliable knowledge, and whether and in what sense or senses scientists are justified in their beliefs that particular systems are chaotic. ;I argue that Alvin Goldman is correct in a general sense; cashing out "justified" as "produced by a process that reliably produces true beliefs" captures a very strong intuition about knowledge and justification. I also argue that defining reliability, as Goldman does, in terms of "truth-ratios" does not capture our intuitions about reliability, and is often ill-defined. Instead, we need to define a reliable process as one which does one of two things: either the process always produces true beliefs, or the process gives specific conditions under which it will fail to produce true beliefs. ;Finally, I return to the discussion of a particular instance of scientific knowledge, the knowledge that a system is chaotic. I argue that, under my conception of reliabilism, and under certain conditions, scientists are in fact justified in their beliefs that particular dynamical systems in nature are chaotic