ORB Visualization
(soon)
National Knowledge Project
John:
I can argue that all sound is a superset of every language invented. Similarly every "unnatural" mark is the superset of every symbol or system of writing. So what exactly is the virtue of such things except that they rise above the "noise" as a practical means of distinguishable communications?
As an experimental physicist or a biologist trying to understand extra-species signaling we would look at the signal characteristics, the circumstances under which they appeared, and the subsequent actions of entities capable of sensing that signal. From this study, we correlate phenomena, and then try artificially to stimulate and reproduce the signals and their effects. Semantics in terms of entities, actions, relationships, are precisely what we start with in any case and we judge the importance of our findings by the predictability of outcomes produced.
Shannon's information theory and my knowledge theory deal respectively with the impacts of sensory signals (information) and with learned conditioning (accepted or influential theory and experience). Both assert quantitative (bit) value measures for the stimulus (information) and the learned response (theory). The least probable signal has greatest information value, and the most constrained decision or action options demonstrate greatest theory value.
We seek preciseness in semantics so as to be capable of recognizing those rare signals unambiguously. The notion of "AIDS" is a good example and our theory of AIDS is now highly evolved and enormously influential and behavior influencing. Thanks to it we know that disease can kill us, but that with appropriate care those with the disease can live among us as close friends.
Metaphysical ontologies have sufficient abstractive range to allow us to speak or think generally where we have no finer-grained needs. Arguing the value of ambiguity as a substitute for abstractness is no defense. Going the other way, trying to find precision in noun phrase constructions quickly reaches a distinct descriptive lower bound on specificity. Type anything you would like into your search engine and look at the odds of your conceptual description matching any one else's.
Medical code link's are precisely the world were precision is required and extremely impactful. With the wrong codes a procedure can kill you, with the right ones the insurance company will pay your bills. That precision does not end discovery or keep us from digging deeper and deeper than our language might ever follow.
John, I know that you sense the nearness of our current language achievements to the goals set long ago by philosophers and logicians. That came across strongly in our conversations at the Knowledge Technology Conference in Seattle 2002. I can distinguish my own post linguistic goals in semantics, from those early ambitions and your notable successes. I do not know yet my bottom line, but it extends to the boundaries of machine learning and self-awareness -- that much I know.
But taking your language and logic goals for what they are, what more are you hoping to achieve and what important dimensions would you prefer to see younger scholars exploring. I will talk at length in the future about the "linguistic/semantic gap", but I think that your work, XML, and Sandy's work are signal summative achievements on this side of the gap. How do you sum them up -- there is great pride and passion still in your words and achievements. What banner do we plant? What Age is nearing completion or just beginning?
Dick