ORB Visualization
(soon)
National Knowledge Project
3/10/2004 8:56 AM
Paul (Prueitt)
The intuition you are looking for is in the physical notion of "degrees of freedom". One seeks to enumerate all the possible situations or geometries or conditional interactions that might lead to a physically distinct reality. You can picture your n-dimensional space as being defined by n independent variables, and for each variable a set of instances over which each ranges. A single n-ary relationship connects together one instance each from all n variables. So each n-ary represents one possible reality, and the bundle of "all" relationships, describes every reality thought physically meaningful. [comment from Paul Werbos]
Then you can expand you space still further, if you want to add dimensions of choice, intent, and all the constraints of influential theories. That expands the problem beyond information (physical space) into a wider world of theory (metaphysical influence). Now information states are a' posteriori, known only after the fact. So they remain variable up to the moment. But to have influence theories must be a' priori, learned and long practiced constraints in reasoning.
The semantic web is a grammar by which we capture and express the mutual constraints within our theories. So our own ontological and rational commitments create a web of reasoning paths and decision branches.
To the extent that these are complete in all situations, they constrain us to move around in ways that we think make sense. Picture perhaps a monkey on a jungle gym. This jungle gym world and our movement through it is still conditional and influenced by the situational (information) n-ary states, but the jungle gym world is pure declarative. Let go of looking only at a process centric causal flow. We can reason forward with "how" or backward with "why" or switch around front or back or through the center with "what if". Only processes flow just one way.
Information flows into the intertwined web through selection of "particular information n-ary vector states" and these touch selected choice options on the rational declarative web. So the trigger may be a sudden unbid reality, the testing of a plan, or a question that focuses on particular situations or a narrowly constrained bundle of possibilities.
Within our decision support applications the resulting reality flow is apparent when tracing (browsing) the different rational paths and investigating the consequences of taking this or that course of options. Humans do this serially, but with "trace op" agents a great many paths can be examined in parallel and then marked for closer analysis.
In any case, what we know is that the mediating structure that satisfies our need will be a chain. [#]
The links within this chain will may embrace many different theories, relationships, and influences. When first used, each of these deserves attention and questioning, but with experience the long chains are reduced to minimal n-ary representations that include only the most critical decisions or history changing events.
Engineers and managers call these the "problem drivers". Factor and sensitivity analyses pick these out easily. To manage a difficult trade-off, you grab the drivers that have greatest impact on putting you close to the goal you seek, then you fine tune to pick up all the other advantages and "low hanging fruit". The constraint browsers are the power tools for this lightning fast work.
Dick