--- A long standing puzzle in decision theory is how decision theory could be applied to itself - that is, how could the cognitive provesses of decision making be applied to reason about these very proccesses (reflection). The reason there is not yet any 'Reflective Decision Theory' is likely to be that the very concept itself is ill- conceived. That is, there is no 'Reflective Decision Theory'. Concepts of 'Utility', goals and decisions about how to most effectively achieve these goals are the domain of a cognitive 'decision making' system. And 'Decision Theory' is the science of such systems. But the concepts can only be applied to 'external goals' (i.e goals in the external world). To attempt to solve 'Reflectivity' by trying to deploy the same concepts of decision theory to the internal workings of the cognitive system is simply an invalid use of these concepts. The first step to solving reflectivity then, involves attempting to ascertain the true nature of 'Reflection'. For 'Reflectivity' is *not* in fact, in the decision making business. The true role of 'Reflection' it seems is *Communciation*. That is to say, it appears that 'Reflectivity' should be thought of, not as part of *Decision Theory*, but instead as part of *Communication Theory*. This is because any cognitive system of sufficient complexity to achieve genuine intelligence appears to require the division of the system into seperate modular 'sub-agents' which interact with each other to achieve desirable results. Marvin Minsky wrote a famous book 'The Society of Mind' emphasizing this soup of many interacting agents. It's not enough for a cognitive so composed to merely have an effective system of decision making. There must also be an effective *Communication System* to integrate and co-ordinate the behaviour of the all the sub-agents into an effective whole. And this is the aspect of AI research that has been neglected. Further, the connection between 'Communication' and "Reflectivity' has appeared to elude the minds of the best and brightest. But it is here being established that an effective communication system *is precisely the solution to reflectivity*. The two problems are one and the same. If the hypothesis is correct, new strategies for atatcking the reflectivity puzzle can be formulated. For one thing, there is a wide body of pre-existing knowledge on Communications Theory which can start to brought to bear on the reflectivity puzzles. For another, analogies from the field of computer networking can be ported over to the reflectivity problem. For instance: Consider sub-agents as nodes, the combined actions of the sub-agents as networks and the interactions of the sub-agents as data transfers. But what justification is there for thinking that this hypothesis is applicable to reflectivity? To see the reasons, let us consider that other great puzzle, subjective consciousness, or subjective experience. What is consciousness. For all the huge volume of past words expended in this debate, there emerge three key points: The first point is that consciousness is not a *thing*. It is a *process*. The second point is that consciousness is not something concrete. It is not for instance, a process similar to digestion. The process instead appears to involve *asbtract patterns* (functionalism). Patterns are abstractions which are the essence of mathematics. Thus we can say that consciousness is a *mathematical process*. And the third point is that consciousness appears to involve a cognitive system examining aspects of its own internal operation. All three points should immediately lead us to suspect that consciousness is connected to 'Reflectivity'. On the first point, conscious as a proccess ; a working reflection system is also a process. On the second point, consciousness as patterns (mathematical abstractions) ; a working reflection system involves reasoning about reasoning ; reasoning uses predicate logic and probabilities - fields of mathematics. Further, patterns are both the essence of mathematics and representations of knowledge itself. So a series of reasoning steps (an algorithm) is really a mathematical construction. Finally on the thrid point, a cognitive system examining its own internal working smacks of reflection immediately. Thus a reasonable a priori suspicion is that consciousness and 'Reflectivity' are in fact one and the same. Let us now apply a unique new perspective on mathematics - we shall now attempt to view mathematics through the lens of the object oriented framework. That is to say, consider mathematics as we would try to model it using object oriented programming - what the classes, methods and objects of math? This is a rather un-usual way of looking at math. Mathematical entities, if they are considered in this way at all, are not regarded as 'Objects' (things with state, identity and behaviours) but merely as static class properties. For instance the math classes in the Java libraries consist of static (class) variables and class methods. But consider instead that there could be mathematical 'objects' (in the sense of entites with states, identities and behaviours). What could these mathematical 'objects' look like? if there are mathematical objects they have to be dynamic. This conflicts with standard platonic pictures of math as entities which are eternal and static. What could these 'dynamical mathematical objects' be? The obvious answer, based on the previous points made, is that *algorithms* (reasoning procedures) are identical to *mathematical objects*. The implementation of the algorithm (ie running the program) would be equivalent to the state changes in the mathematical object. The advantage of this perspective on math is that it points to a way to unify the mathematical sciences with the computational sciences. Once we are clear about the equivalence of computation with dynamical mathematical objects, we can begin to see why reflectivity might be identical to consciousness. What is the fucntion of consciousness? The perspective discussed suggests an answer. It's got to have something to do with *knowledge* (because a mathematical process - a 'dynamical mathematical object' is equiavlent an algorithm, and astract mathematics - the language of math -are also the language of logic) . Now return to the discussion about the sub-agents in an intelligent cognitive system and the need for all these sub-agents to communciate effectively. Rather than a reasoning system consisting of a single big algorithm then, we should think of the system as composed of many small algorithms which interact together to produce intelligent behaviour. From what we said above, all of these sub-algorithms should be considered to be 'dynamical mathematical objects'. Thus, the picture we obtain is analogous to a network - the mathematical objects are the nodes, the system is the network, the co-ordination and interaction of the sub-agents is the data communication. Consider that each mathematical node represent a 'node of knowledge'. Then we see that the function of consciousness must invovle 'knoweldge integration' (nodes working together to achieve a co-ordinated whole). And we can push this hypothesis further and conjecture that: *THE PURPOSE OF CONSCIOUSNESS IS AS AN INTERNAL COMMUNICATION SYSTEM OF THE MIND. CONSCIOUSNESS IS WHAT ENABLES ALL THE SUB-AGENTS IN A COGNITIVE REASONING SYSTEM TO TALK TO EACH OTHER AND THUS INTEGRATE THEIR BEHAVIOUR INTO AN EFFECTIVE WHOLE. This, in bold-type, is the first 'Key Hypothesis' of this peice. If in fact we have nailed the purpose of consciousness and the perspective offered is roughly correct, a second key hypothesis presents itself by pushing the first hypothesis furtehr and considering again the perspective offered. If the function of consciousness is as an internal communication system of the mind, why not say that consciousness itself is a 'DP Modelling Language' (Data Process Modelling language)? This is strongly suggested by the picture of the equiavlence between 'mathematical nodes of knowledge' communicating and consciousness enacted as an algorithm. Knowledge is communciated by a data process modelling language. This second 'big idea' can be summarized thusly: *CONSCIOUSNESS IS PRECISELY THE LANGUAGE OF REFLECTIVITY. CONSCIOUSNESS IS THE DP MODELLING LANGUAGE OF THE MIND WHICH THE MIND USES TO REFLECT (End of part 1) --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Everything List" group. To post to this group, send email to everything-list.domain.name.hidden To unsubscribe from this group, send email to everything-list-unsubscribe.domain.name.hidden For more options, visit this group at http://groups.google.com/group/everything-list?hl=en -~----------~----~----~----~------~----~------~--~---Received on Wed Aug 22 2007 - 00:18:50 PDT
This archive was generated by hypermail 2.3.0 : Fri Feb 16 2018 - 13:20:14 PST