Pp. ix, 222 , Edinburgh , Edinburgh University Press , 2007 , £17.99.
Science fiction writers have long used the concept of artificial intelligence to reflect upon issues of identity and consciousness, holding it up as a mirror with which to examine humankind. In his classic Do Androids Dream of Electric Sheep? (1968), Philip K. Dick makes use of his android characters to comment on the artificiality of much of human life, questioning what criteria privileges human life over that of ‘andys’ when humans jigger their mental states with artificial mood stimulators and surround themselves with robotic animals. Artificial intelligence has long been a literary tool for the collective introspection into the human condition.
Like good science fiction, Matt Carter's Minds and Computers essentially constitutes an exploration into what makes human beings what they are. As with all storybook madmen, it is ourselves that, through the pursuit of artificial intelligence, we are seeking to create or recreate, and so any investigation into the subject must start with what constitutes the human mind.
Carter begins as do most other theorists of consciousness by analyzing critically the Cartesian dualism that is still the prevalent image people possess of consciousness, as well as other dualistic systems that try to work out the body/mind divide, showing them to be limited in their usefulness as models, if not outright false. The author then presents the merits and demerits of behaviourism and Australian materialism, with a brief foray into the basics of neuroanatomy, before moving on to fuctionalism, a theory of mind which provides a ‘substrate independent and, hence, ontologically neutral’ model of mental states and is thus the only real ‘framework [which] allows for the possibility of artificial intelligence’ (p. 47). After a brief summary of the nature of formal systems and computation, outlining the Church/Turing thesis and the concept of Gödel coding, the book then moves on to a thorough presentation of computationalism, a species of functionalism which ‘fleshes out the fuctionalist framework’ by ‘confer[ring] a clear methodology for investigating mentality’ and ‘specify[ing] the relationship between the mind and the brain’ (p. 101).
Of course, the only way that one might recognize an artificial intelligence as exactly that is through communicating with it; such is the basis of the Turing test, and Carter demonstrates his interdisciplinary approach by spending a few chapters to delve into the nature of language, expounding upon Chomsky's theories of syntax as well as the necessary relationship between syntax and semantics in human mental life, as demonstrated by the Chinese room thought experiment. The lesson which Carter takes from the Chinese room ‘is that embodied experience is necessary for the development of semantics. In order for our mental states to have meaning, we must have antecedent experience in the world, mediated by our sensory apparatus’ (p. 179). Mental states that possess any meaning are representational states.
In the penultimate chapter, Carter explores how the connectionist paradigm of human cognitive architecture allows for the creation of artificial neural networks, even presenting some diagrams for very simple such networks and demonstrating how they can be used to synthesize human speech. The author admits that ‘artificial neural network models remain gross simplifications of the biological neural activity which they seek to model’ but holds out hope that ongoing investigations into the brain may allow us ‘to develop yet more sophisticated models which implement the neurobiological principles we uncover’ (pp. 200–01).
As an introduction to the philosophy underlying artificial intelligence, this is a perfect book for students, complete with a glossary of terms, a not too exhaustive list of suggested readings, and exercises throughout. The exercises, most of which deal with core computational concepts, are particularly helpful in that they make concrete certain ideas which would otherwise remain simply abstract. In fact, the author frequently urges his readers to carry out these exercises using coins or other tangible objects, thus making all this philosophy that much more real for the novice. Carter consistently exhibits the sort of clear understanding of what guidance the average student needs; he admits to glossing over some of the more technical aspects of the theories presented but always points his reader to where further information may be obtained. In short, Minds and Computers is a teaching tool par excellence and should find its way into every classroom where the philosophy of mind is being studied.