ABSTRACT
INTRODUCTION Our general aims are to contribute to a characterisation of intelligent systems from an ontological perspective afforded by systems theory and by our cognitive research programme at Edinburgh which has featured human and non-human primate investigations into what many have regarded as core cognitive competences viz. relational learning (Bryant 1974; Lawrenson and Bryant 1972; McGonigle and Jones 1978; Reese 1968), transitivity (Bryant and Trabasso 1971; Halford 1993; Inhelder and Piaget 1964; McGonigle and Chalmers 1977; Trabasso 1977), and ordering skills (Lashley 1960; Terrace and McGonigle 1994). As a logical extension, the programme has more recently extended into cognitive modelling (e.g. Harris and McGonigle 1994) and robotics (McGonigle 1990). In this we have taken the view that good characterisations of complex intelligent systems in particular should lead to testable designs for artificial agents (Donnett and McGonigle 1991; Nehmzow and McGonigle 1993). However, the flow is not one way: instead it is becoming more and more apparent that a symbiotic relationship between research on real and artificially intelligent systems is the most fruitful way forward-a domain which we would term synthetic intelligence.
