One of the claims integral to John Searle's critique of computational cognitive science and 'Strong AI' was that computation is 'observer-relative' or 'observer-dependent' (Searle, The Rediscovery of the Mind, 1992). This claim has already proven to be very controversial in cognitive science and AI (Endicott 1996; Coulter & Sharrock, Rey, and Haugeland in Preston & Bishop (eds.), Views into the Chinese Room, 2002). Those who come to the subject of computation via physics, for example, often argue that computational properties are physical properties, that is, that computation is 'intrinsic to physics'. On such views, computation is comparable to the flow of information, where information is conceived of in statistical terms, and thus computation is both observer-independent and (perhaps) ubiquitous. Connected with this are related issues about causality and identity (including continuity of), as well as the question of alternative formulations of information. This symposium seeks to evaluate arguments, such as (but not limited to) Searle's, which bear directly on the question of what kind of processes and properties computational processes and properties are. It thus seeks to address the general question 'What is computation?' in a somewhat indirect way. Questions that might be tackled include: Are computational properties syntactic properties? Are syntactic properties discovered, or assigned? If they must be assigned, as Searle argues, does this mean they are or can be assigned arbitrarily? Might computational properties be universally realized? Would such universal realizability be objectionable, or trivialise computationalism? Is syntax observer-relative? What kinds of properties (if any) are observer-relative or observer-dependent? Is observer-relativity a matter of degree? Might the question of whether computation is observer-relative have different answers depending on what is carrying out the computation in question? Might the answer to this question be affected by the advent of new computing technologies, such as biologically- and physically-inspired models of computation? Is it time to start distinguishing between different meanings of 'computation', or is there still mileage in the idea that some single notion of computation is both thin enough to cover all the kinds of activities we call computational, and yet still informative (non-trivial)? Does Searle's idea that syntax is observer-relative serve to support, or instead to undermine, his famous 'Chinese Room argument'?
Organising Committee http://extranet.smuc.ac.uk/events-conferences/aisb-symposium-2014/Pages/default.aspx Description: