I may have misunderstood what you were saying but it has been a great help to me regardless.
When a concept is defined as being shaped by a value, it is a definition of a dependent concept. So then simply defining a concept as being dependent is not enough from my point of view. A concept has to be defined within a system of interacting (interdependent) concepts. That is what I realized last night, and that is how my fundamental theories are different than everyone who is using a value as a method for shaping a concept (a concept-like object) in their fundamental definitions. Part of a fundamental definition of a concept within the Conceptual Relativism theory is that one has to be able to give a fair accounting of how the concept might be developed and subsequently refined (or rejected) within a system of interrelated concepts and that means that a fundamental definition of such systems have to be made. The question of, what does the concept do, is a question of how would the concept work with other concepts. Different concepts have different structures just as the structure of a computational syntax is based on different data objects such as types, operands and operations. Only a conceptual type can be much more varied than that. Concepts may be defined as independent but typically a concept is not going to be independent. So a concept that can shape another concept may in turn be shaped by that other concept. Isn't this what happens when we analyze a situation using other knowledge that we possess or acquire? Don't we learn that the methods used in the analysis of a situation can be refined or shaped to be more effective when examining the situation? The idea that the concepts used in understanding could never be molded in response to the attempt to use them while shaping another concept is nonsense. A system of concepts used together that may mutually shape the other concepts as they are being used is a good definition of conceptual relativity. So the (conceptual) tools that are used to examine some concept structure may themselves be shaped by the subject concepts of the examination. That is an example of conceptual relativity. I believe that a fundamental definition of a concept which defers the definition of the sources of the values that might shape the concept is characteristic of the old failed methods of narrow AI. It is obvious that there has been something fundamental missing in these kinds of definitions and it may be the discovery and definition of the mechanisms of the interrelations of interdependent concepts that is needed to make a strong AGI definition. No I am not saying that there is anything wrong with traditional approaches to discussing and defining the objects of AI programs. What I am saying is that we have to have some mechanisms to keep these referential complications from disabling the actions of the objects derived from the fundamental definitions. The first step is by understanding that there are interrelations like those that I call Conceptual Relativity at work and then try to find the mechanisms that can discover these interrelations in the referential world and manage them so that they do not create a devastating loss of traction when they are applied to some problem. These definitions do not have to be extensive and they can be made with abstractions and generalizations. So it can be used in simple feasibility testing. A well known problem with (what I call) Conceptual Relativity in human thought is that of changing the mechanisms a test to measure or analyze a concept so that the results will fit the desired results. So then this is one kind of problem that my theory of Conceptual Relativism might be expected to produce. How can I develop my definitions of conceptual relativity to avoid that particular situation? Well I can't be sure that it will never occur but I can give the system some tendency to avoid it. However, it is a tricky situation because there are times when you need to change your methodology of evaluation in order to develop greater insight into a problem. But by thinking about the issue this way I might be able to emphasize how certain aspects of the concepts play different roles in the situation that I am concerned about and work it out so that the different 'characters' in my conceptual theater will act to balance the trickier changes introduced by another character in the system. If you move the goal posts then you have to change the game. You can't redefine the goals and say it is the exact same game as it was before. That is a simple mechanism which might be defined and managed by using the assignment of roles in the conceptual system. Jim Bromer On Fri, Oct 10, 2014 at 3:19 PM, Mike Archbold via AGI <[email protected]> wrote: > On 10/9/14, Jim Bromer <[email protected]> wrote: >> Mike Archbold said: >> Jim, I think about the issue you emphasize of no 'independent >> concepts' frequently. It plays a role in my latest approximate >> design. Mike A >> >> >> >> The idea of using systems of interdependent concepts is something that >> can be simulated easily since the interdependence is something that >> can be abstracted in computational terms. So, if someone wanted just >> to try it with meaningless abstract tokens or objects he could do >> that. He could use values or discreet interrelations. New relations >> could be introduced and studied. The concepts could be used with >> different syntactic structures (or different functional relations >> based on different characteristics) and so on. The same kind of thing >> can be used with concepts that have very simplistic meanings. >> >> Because this idea that you described as "no independent concepts" is >> (itself) computationally feasible that can lead to all sorts of simple >> programmable possibilities. I am definitely going to try this >> >> So a simulation of highly interdependent concepts and interdependent >> conceptual 'roles' is something that can be very simple. Even though >> conceptual relativism is something that goes beyond the >> interdependence of the concepts, your abstraction of this essential >> feature of the concept may be a help to my finding a good starting >> point for my next attempt to write a simple AGI program >> Jim Bromer >> > > There has been some talk here lately of Tononi's integrated > information theory. I only know the summary generalities, but it is a > (I think) holistic approach with mathematics holding it up. In the > wiki, at least, they are not talking about having a bunch of stand > alone "independent concepts." > > > ------------------------------------------- > AGI > Archives: https://www.listbox.com/member/archive/303/=now > RSS Feed: https://www.listbox.com/member/archive/rss/303/24379807-653794b5 > Modify Your Subscription: https://www.listbox.com/member/?&; > Powered by Listbox: http://www.listbox.com ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657 Powered by Listbox: http://www.listbox.com
