If cognitive science is to become a discipline in its own right, and not just a congeries of parts of other disciplines, perhaps its best hope lies not only in such multi-pronged attacks on common problems, as just discussed, but in single research groups whose members come from different disciplines yet who work together on common problems of cognition. The range of disciplines, and the levels of analysis, are by no means settled and, indeed, are widening in scope. As Donald Norman (1981b: 3) observes, ``To some, the very essence of a cognitive system is that of a symbol processing system'', while to others the essence is that of connectionist neural networks. To yet others, the essence of cognitive science is more holistic, viewing the mind as an integral component of the larger world--of society, of culture--not (solely) understandable in terms of symbol manipulation (syntax) but in need of a semantics--an understanding of the relations of the mental symbols to the external world. Thus, two major open issues for a complete understanding of cognition are--looking inwards--how the mind is implemented and how the very fact of its implementation in particular kinds of physical or biological mechanisms influences the nature of cognition, and--looking outwards--how and to what extent the nature of cognition is shaped by the socio-cultural world that minds find themselves in.