Non-monotonic Logic includes Abductive reasoning, Reasoning about knowledge, first-order circumscription, closed-world assumption, and autoepistemic logic.

===

non-monotonic reasoning is defeasible. A good amount of how humans reason is via non-monotonic reasoning.

=====

Another example: You turn on a device, expecting it to work. But then, you hear a “pop” and the circuit breaker blows.

Defeasible reasoning. Your reasoning that it would work is based upon an internal model of “How things should be”.

It is only through our ability to use non-monotonic reasoning that we are able to troubleshoot problems. We can do so because we have a model (or database) in our minds that reality is not conforming to.

Working through “what the problem _could_ be” is defeasible at every turn – non-monotonic – until you solve it, and model and reality match again.

Closed-world assumption.

===

t goes to 0. You expect it to work. It does not. Probability that it is working is 0.

====

In the situation you are evaluating, turning on =/= on.

You can repeatedly test and retest by turning it off and on and off and on and off and on.Each time it does not turn on.Now you are faced with belief revision.Turning it on does not always turn it on.What next?

You can repeatedly test and retest by turning it off and on and off and on and off and on.Each time it does not turn on.Now you are faced with belief revision.Turning it on does not always turn it on.What next?

=====

It looks as if the heyday of non-monotonic logic was from the early 1980s to the early 1990s, during the era of Expert systems AI, and died off during the AI Winter of the 1990s.

I’m currently converting: “Handbook of Logic in Artificial Intelligence and Logic Programming: Volume 3: Nonmonotonic Reasoning and Uncertain Reasoning”, 1994 to txt, after converting it from a postscript. PS to PDF.

552 pages – and it’s easy to read so far.

======

t’s the subspace and reasoning one’s way out of it back to normal that’s of interest to me

====

But you’d need an internal troubleshooting database. If you lack that, what do you do?

======

had to write a darn batch file. program only split PDF to jpgs. i feel so primitive.

——

Right, but you’d need a set to iterate through.

—-

Right. But if your internal model of “device” is: “Black Box”, there’s nothing in a troubleshooting set.

=====

“if it’s a good model”.Compare:

Computers. I can troubleshoot a computer by the noises it makes.

Cars: “it’s making a weird noise”. END PROGRAM.

Computers. I can troubleshoot a computer by the noises it makes.

Cars: “it’s making a weird noise”. END PROGRAM.

=====

What computer? I’m talking about human reasoning.

======

What probability is there in you that your model is incorrect or incomplete?

=====

I know Bayes, frequentist and several others, at least by familiarity. I’m asking a direct question.If human reasoning is Bayesian and you are human, your belief in Bayesian reasoning as human reasoning will come with a percentage of probability or at least of belief %.What’s yours?

======

So, you are uncertain and undecided whether or not Bayesian is a pretty good, close approximation, to natural human reasoning.But you made that statement. As far as I know, retracting statements is non-monotonic but not allowed in Bayesian.

======

My belief is that it is partially Bayesian, partially non-monotonic, partially frequentist. Human brains are flexible and, after all, it is human brains that came up with ALL of these logics.

=======

Given a set of all logics over a set of all human brains, which logics do not map to human brains?

=====

Using specific case to prove a general case. A fictional case. Can I form a probability of a fiction?

======

So, I have to update my internal model of reality to accept a hypothetical for the sake of an argument in order to arrive at your conclusions.

======

No. I’m saying that different reasoning systems can operate at different times for different needs.

=====

That is your assertion that there is only two possible cases.

====

Closed-world assumption.

That is a trait shared between first-order logic and nonmonotonic logic.

That is a trait shared between first-order logic and nonmonotonic logic.

======

Yes, I know how first-order logic works.

=====

In “Closed-world assumption” logics, anything that is outside the rules of that logic is rejected. There are different criteria for negation in each logic.

======

In non-monotonic logic, which I assert is more akin to natural human reasoning, defeasibility is negation.

======

Oh, the axioms of non-monotonic logic are really quite simple. Far simpler than first-order logic.

But its drawback is that it’s far too human as it were.

======