Wednesday, September 9, 2009

David McFarland: Guilty Robots, Happy Dogs - The Question of Alien Minds

(My review will appear in Philosophy in Review)

David McFarland Guilty Robots, Happy Dogs: The Question of Alien Minds. Oxford: Oxford University Press 2008. Pp. 256. US$34.95 (cloth ISBN-13: 978-0-19-921929-2); US$15.95 (paper ISBN-13: 978-0-19-921930-8).

David McFarland is a retired professor of animal behavior and robotics who pursues an active interest in philosophy. In this book he treats the question of whether non-human animals and future robots could be said to have minds. While much of the book brings research in animal behavior and robotics to bear on its theme, McFarland is also well acquainted with the philosophy of mind and he readily appropriates philosophical concepts and terminology -- ‘philosopher-speak’ (50) -- when needed. Although several chapters are almost completely devoted to philosophical ideas, I think that philosophers will actually find the non-philosophical sections more stimulating.

Chapter 1, ‘Mindless Machines’, sets up a taxonomy of increasingly sophisticated robot behaviors and abilities. Robots can be simply reactive to certain elements of their environment; they can demonstrate ‘stigmercy’, or ‘[t]he production of behavior that is a direct consequence of the effects produced in the local environment by previous behavior’ (219); a robotic ‘goal-achieving system’ can change its behavior (by stopping, for instance) when a certain goal has been achieved; a ‘goal-seeking system’ is designed to work towards the accomplishment of a certain goal ‘without the goal being represented within the system’ (11), while the behavior of ‘goal-directed’ systems is informed by such representations. Next comes a discussion of different types of autonomy. Robots can be autonomous as regards their procurement of energy (as exemplified by the fascinating ‘slugbot’ slug-hunting robot). If a robot is capable of determining which of its goals it will pursue in different circumstances -- choosing, for example, either to continue working or to recharge its batteries -- it is ‘motivationally autonomous’.

Chapter 2, ‘Design of Animals and Robots’, makes an interesting attempt to explain both the origins of species in nature and the development of robots in a market economy in terms of a single broad evolutionary framework. It also outlines a model of resource management as practiced by both animals and robots. Robots seem very different from biological organisms insofar as they depend upon humans to 'reproduce' them; however, many domesticated animals also reproduce with human assistance. Parasites in nature are similarly highly dependent upon their host organisms. More generally speaking, robots and animals alike are designed to fit particular niches and serve particular functions: ‘There is no such thing as a generalized animal; there will never be successful generalized robots’ (30). Here McFarland is already gesturing towards one of the book’s main points: that human beings are designed to perform human activities in human environments and thus are endowed with human intelligence and human minds, while other animals are designed to engage in other kinds of activities in other kinds of environments and thus are likely to possess other kinds of intelligence and other kinds of minds.

Chapter 3, ‘Interpreting Behaviour’, is more philosophical than the previous chapters and perhaps less interesting to philosophers. It offers an overview of the debate over folk psychology and eliminative materialism, continues with a more detailed account of Daniel Dennet’s notion of the ‘Intentional Stance’ and rationality, and concludes with the ‘rule of thumb’ that ‘for an animal or robot to have a mind it must have intentionality (including rationality) and subjectivity’ (95), where subjectivity requires not only that the animal/robot has experiences, but also that it knows that it has experiences.

The next three chapters introduce a number of phenomena associated with human mental activity (having a ‘theory of mind’, tool use, qualia, and self-awareness) and describe attempts to uncover their presence in animals through empirical research. I was particularly intrigued by the idea that the presence of subjective experience might be detected by experiments in which animals seem to choose the optimal available combinations of various pains and pleasures. Chapter 7, ‘The Material Mind’, returns to an overview of standard themes in the philosophy of mind, including mental causation, Searle’s Chinese Room, and various brands of functionalism. This is not a particularly strong chapter, and the material is better covered by many introductory texts. Chapter 8, ‘Mental Autonomy’, includes further discussions of subjective experience and self-awareness and develops a ‘wish list’ of powers and characteristics that an animal or robot should possess in order for it to be thought of as having a mind. This leads to questions of moral accountability; when something goes wrong, ‘Who is to blame, the robot or its designer?’ (188). Going beyond the ‘motivational autonomy’ described in Chapter 1, McFarland suggests that a robot might be designed to step outside pre-programmed decision-making algorithms, to ‘take the initiative’ and ‘break the rules’ when necessary (but how would it know when this is necessary?). Perhaps ironically, he mentions in an aside that, ‘[w]e hope its values would be such that it would not do anything stupid’ (198).

The book’s surprise conclusion is reserved for its epilogue, ‘The Alien Mind’. McFarland holds out little hope for the possibility that science will solve the problems of philosophy. He thinks that the empirical data can accommodate just about any of the various and conflicting theories at play in the contemporary philosophy of mind. The final closing paragraphs suggest that the attribution of mindedness to animals or (future) robots will, at the end of the day, be a matter of social convention.

Despite its anticlimactic conclusion, the book is certainly worth reading. Even in this age of fashionable naturalism, philosophers will still have something to learn from McFarland’s style of thinking, with its thorough grounding in robotic engineering and empirical studies of animal behavior. They will certainly come away with a new stock of practical examples and ideas to philosophize about. One odd deficiency of the book is that it seems a bit dated. Its references peter out towards the end of the 1990s and it has nothing to say about the great recent advances in brain research resulting from new imaging technologies. The book is generally well written and is equipped with a useful glossary, which should make it accessible and even interesting for lay readers and undergraduates. There are, however, better first introductions to its more strictly philosophical content.

Berel Dov Lerner
Western Galilee College, Israel