next up previous
Next: References Up: SITUATION CALCULUS WITH CONCURRENT Previous: Acknowledgments


This section will be removed from the final paper. It is included now only so that its remarks will appear when the document is latexed.

The word ``but'' can play a role in narrative. Suppose an event leads to a situation, but the properties being asserted about the situation are not what would normally follow from the occurrence of the event. Perhaps a good narrative should label the anomalous fluents of the new situation with ``but''.


We have not yet treated being able to prove that a person cannot accomplish something or that something cannot happen. The easiest way to think about this may be to have Junior try to prevent Daddy from doing something.

quotes from Russell and Bell

common sense informatic situation

facts vs. what is known

It seems that < should not be transitive. Narratives should be objects.

Fluents should be inferred to persist as long as there is no event in the narrative or directly following from the narrative that would change this.

It may be advantageous to treat processes by introducing a fluents that persist and whose persistence determines that some secondary fluents change in a specified way, e.g.


We need to be able to declare some fluents as dependent on others so that their change or persistence is not inferred separately, e.g.


Actually it might suffice to write


Probably Vladimir has thought about this possibility and should be asked. Whether a fluent is dependent may depend on context. Maybe we should distinguish between asserting the dependence of fluents and that of functions whose value is a fluent (as in the present example).

It looks like we may need priorities to handle the rules about what persists because of the narrative or what follows from the narrative.

Somehow he got to Moscow after losing his ticket. What is the semantics of ``somehow''?

losing ticket
buying ticket
moving block
selling block
sending telegram with messaage
sending money

Can these be handled in a uniform way? An event is realized by a sequence of subevents, actually by a subnarrative. Indeed suppose that how a person performs an action involves delegating some of the work to another person. It is usually unnecessary to completely specify the sequential or temporal relations of the work performed by the different people.

The biggest strain on the single history interpretation will come with counterfactuals or trying to compare the outcomes of different strategies. That's where the free will approximation comes in. Thus free will is an approximate theory in a determinist world.

next up previous
Next: References Up: SITUATION CALCULUS WITH CONCURRENT Previous: Acknowledgments

John McCarthy
Thu Apr 27 22:02:22 PDT 2000