The Computer Revolution in Philosophy: Philosophy Science and Models of MindSloman, Aaron (1978) The Computer Revolution in Philosophy: Philosophy Science and Models of Mind. Harvester Studies in Cognitive Science. The Harvester Press, Hassocks, UK. ISBN 0-85527-389-5 This is the latest version of this eprint. Full text available as:
Alternative URL: http://www.cs.bham.ac.uk/research/projects/cogaff/crp/ AbstractThe book is about the nature of science, the nature of computation, the nature of mind, and how computational models, and especially the idea of a virtual machine with multiple interacting components, can transform the study of mind including consciousness. Part 1 is about methodology, including providing an analysis of science as primarily concerned with expanding our understanding of what is possible (as opposed to simply discovering laws and correlations) and a tutorial on conceptual analysis. Part 2 aimed to illustrate the new way of studying mind: by analysing requirements for the whole architecture of a human mind consisting of many concurrently active, interacting components, in a way that allowed for the possibility of implementation in artificial minds; by discussing requirements for diverse forms of representation (as opposed to doing everything in logic, or in neural nets, for example); by exploring the growth in self-understanding involved in a child learning about numbers; by illustrating the architectural requirements for a visual system using a multi-level ontology with concurrent top-down and bottom-up processing (implemented by the author and colleagues in the POPEYE program); and by discussing how some aspects of self consciousness and our notion of free will relate to architectural control requirements, including meta-level control. An epilogue briefly discusses varieties of opposition to these ideas and predicts a need for a society for the prevention of cruelty to robots. Unlike most publications promoting AI, this book consistently emphasised the difficulties of the problems of replicating human-like minds and predicted that it would take a very long time. It also emphasised the need to blur the agent/environment distinction. THE ONLINE VERSION: The book went out of print about 20 years ago, and all rights reverted to the author. One of his students made a photocopy of the book in 2000, and some months later it was scanned in using an OCR package. The resulting HTML files, with newly created diagrams were placed online in 2001, with a number of additional notes and retrospective comments, which have subsequently been expanded from time to time, including references to later relevant work. In 2007 the whole thing was converted to PDF using OpenOffice, in a total of 215 pages, with a short additional section on the history of that version. That is the version that is being submitted to ASSC Eprints. Comments/DiscussionThe submission menu forced me to choose only one discipline and one topic. I would have preferred to specify that this is a multi-disciplinary book straddling at least philosophy, artificial intelligence and theoretical psychology. and multiple topics, not just 'cognition'. See the list of keywords. ORIGIN OF THE BOOK: I started learning about Artificial Intelligence while I was a lecturer in philosophy at Sussex university in 1969. I gradually became convinced that the things we were learning to do with computers provided radically new ways about thinking about all aspects of mind, including relations between mind and body. This book attempted to show how philosophical conceptual analysis combined with design of working systems, guided by empirical facts about human competences could transform our ways of thinking about the nature of mind. It merely scratched the surface, but it helped convince a number of quite important philosophers that AI was worth taking seriously as an approach to mind. It also influenced some other people, e.g. some researchers on executive functions. The book was written before the rise of connectionism, but much of the discussion is concerned with design requirements that need to be met whether by neural or other mechanisms. Some of the notes added to the online version mention relevant later developments. An unusual feature of the book is that a lot of it was driven by the author's belief (as a mathematics graduate) that current views on the nature of mathematics and how mathematics is learnt were inadequate, and needed to be based on a new theory of mind as having an architecture that allowed reflection on its own processes, including counting processes, for example. Some of the ideas presented in the chapter on learning about numbers are closely related to Annette Karmiloff-Smith's notion of 'representational redescription', presented in her book 'Beyond Modularity'. Although most of the book was written over 30 years ago, the ideas are still being developed and the author welcomes comments and criticisms. Note: 16 June 2007 A reader reported typos which I have corrected in the updated version.
Available Versions of this Item
Repository Staff Only: edit this item |
