Full Text (truncated)
… [… snip …] …
- This book is a prolegomenon to the enterprise of building a person. It is a defense of three theses: token physicalism1, agent materialism2, and strong AI.
- Token physicalism3 is the thesis that mental events are physical events. In human beings, they are presumably neurological events.
- Agent materialism4 is the thesis that persons are physical objects having a suitable structure.
- Strong AI is the thesis that one can construct a person (a thing that literally thinks, feels, and is conscious) by building a physical system endowed with appropriate "artificial intelligence5".
- It is argued that what is required for the construction of a person is a physical system that mimics human rationality in a sense made precise in the book. The project of building such a system is actually underway in the OSCAR project, and this book discusses the theoretical underpinnings of the OSCAR project at some length. The precise objective of the OSCAR project is a formulation of a general theory of rationality and its implementation in a computer program. It follows from the theses of the book that a system running this program and appropriately connected to the world will literally be a person.
- OSCAR represents the dream of AI since its infancy, but it is a dream that has faded in much of the AI community. This is because researchers in AI have made less progress than anticipated in achieving the dream.
- My claim in this book is that the failure is not intrinsic to the task, but stems from the fact that many of the problems involved are essentially philosophical, while researchers in AI have not usually been trained in philosophy. Training in philosophy is not by itself sufficient to solve the problems, because they are hard problems and have difficult non-philosophical ingredients as well, but input from philosophers is probably a necessary condition for their solution.
- The intent of this book is to begin the process of providing a recipe for building a person, and the intent of the OSCAR project is to implement that recipe.
- The central claim of the book is that building a person reduces6 to the task of constructing a system that adequately models human rationality. Some time is spent making this claim precise, and it is urged that a system modeling human rationality in the appropriate sense will experience qualia, will be self-conscious, will have desires, fears, intentions, and a full range of mental states.
- I am not content to give a general theoretical argument to the effect that a system modeling human rationality will be a person. I want to spell out precisely what that involves and then actually build such a system. In order to accomplish this, we must have a theory of human rationality of sufficient precision to make computer modeling possible. This will include an account of theoretical reasoning (epistemology) and an account of practical reasoning.
- This book is not an appropriate vehicle for laying out an entire theory of rationality, but the last chapter sketches how the general theory will go, making reference where possible to the more precise accounts I have given elsewhere. In effect, OSCAR will be an implementation of my own theories of rationality.
- I find that most of the work I have done in philosophy over the last twenty-five years is directly relevant to this problem. In an important sense, I have been doing AI all along without realizing it. An increasing number of technically minded philosophers are coming to the same conclusion. On the other hand, no one can claim to have solved all the problems, and anyone working in this field quickly discovers that implementation and theory construction go hand in hand.
- The attempt to implement a philosophical theory of reasoning is usually quick to reveal inadequacies in the theory. In effect, the computer becomes a mechanical aid in the discovery of counterexamples. So by immersing themselves in AI, philosophers are not giving up their more traditional interests. Instead, they are adopting powerful new tools that will be of tremendous help in the solution of old problems.
- This book grew out of a series of articles in the philosophy of mind, together with the felt need for writing something that ties all of my work together and shows how it is all relevant to the OSCAR project and how the OSCAR project is relevant to traditional philosophical concerns. The book consists largely of descendants of bits and pieces of the following articles:
Text Colour Conventions (see disclaimer)
- Blue: Text by me; © Theo Todman, 2021
- Mauve: Text by correspondent(s) or other author(s); © the author(s)