Moderation Questions

Formal logic could take some lessons from programming then. Dereferencing a null pointer throws an exception, which halts execution if unhandled. In languages without exception handling it just causes a segfault and your application crashes. Suffice to say, you don't get a truth value either way.

Yes, agreed.

Will you tell us about dereferencing nulls now, professor?

I think LLMs are getting there. They’re starting to blend formal-style inference with context sensitive tracking, so they can handle things like failed references or presupposition without breaking the reasoning loop.

That's a simple problem. Then you have to address the problem of null data (or whatever nomenclature you fancy)

Lol. No. Guess you have no idea what a null pointer is, then. They must have missed that in your professor training.

It's really just about not anchoring too much. 'No I haven't stopped beating my wife' has multiple meanings. The problem is a tendency of some to simplistically anchor to. "You have a wif you have been beating her and you still beat her. That anchoring is a mistake but you wouldn't want a jury to have the opportunity do it. It's only because people anchor badly that you might not want to say No to the question. Others conclude the answer makes no sense but that is also wrong as it contains information

LLMs tend to appear to anchor too easily and then easily switch to a different anchor. It's being addressed and it's not too unlike humans. We struggle with anchors more and some are better at avoiding anchoring. I write posts on the importance of avoiding anchoring but they're not popular.

Edit: Amusing As I wrote that. D2 appears to have been busy anchoring. I allow he is just having fun but we also see the tendency of many to anchor to their preferred meaning. The problem of bias is pernicious.

Jalfrezi will probably believe your guess which is nice.

I mostly agree but for me the interesting biases are more metaphysical than cognitive. I run into naive or scientific realism as the default ‘anchor’ in debates. That’s why FH Bradley famously avoided debating Russell: under Russell’s analytic framework, idealism has no path to victory.

Actually, it can ONLY be a single proposition in Propositional Logic. There are no logical connectives in the statement 'The King is France is bald', so the best we can do is call it P.

n.b. "Is" is a logical operator in Predicate Logic.

But a) and b) can ONLY be expressed in Predicate Logic, given that we wish to cash out the logical implications of the sentence.

There is also c) There is exactly one king of France.

Yes I was going to point out that the problem is in sentential logic and sort of falls out once you switch that to predicate logic, because then there are two different predicates in the sentence, the latter of which is not the reason the truth conditions for “the present king of france is bald” evaluate to false. Then we get no problem because the negation would be “it is not the case that there exists one king of france and he is bald” is true.

Another way to “fix” the problem without resorting to quantifier would be to say that the statement is non-propositional because one of the terms in the statement doesn’t refer to anything. But then this is like saying you can have a statement where meaning is conveyed that seems to have a truth value, but it actually isn’t a proposition. Seems very odd.

The other two ways you can “fix” it is to just dispense with classical propositional logic and allow for either truth-value gaps where some propositions don’t have a truth-value, or by saying that a proposition’s negation isn’t always true when the proposition is false. Both get rid of the law of excluded middle, so they are not compatible with classical logic.

That's the point though isn't it. It appears a single proposition as stated because it has no connectives but there's a lot of meaning hidden away. We can try to fix this with multiple propositions. This may or may not be possible. It may require a higher order logic or as checkraise draw alludes to, a non-monotonic logic which doesn't have not not p = p

This is one of the general problem when translating language into logic. We can mean a lot when we say very little. A lot is assumed. Sometimes anyway.

In classical syllogistic logic, even a universal proposition is only meaningful if the predicate refers to something that actually exists. That is to say, that even universal propositions have existential import. Not so, of course, in 'modern' Predicate Logic.

I don't see how kneecapping the syllogism is a "fix."

Well I don’t want to be committed to saying that “all unicorns have horns” is meaningless, for instance. It seems like the tool is just too strong for the problem it’s trying to fix.

If your posts are anything to go by it’s easy to believe your apps kept crashing because you didn’t know how to use null pointers, so you came up with the ingenious idea of asking your interview candidates to fix it for you.

atta-boy

Chez, it's abundantly clear you have no idea what a null pointer is. Sounds like you think it's what you get when you load data into your application from a database field that contains nulls.

When you can get interviewees to fix your horrible code for you, why bother learning anything?

I'd summarily fire any manager who let chez interview someone for a janitorial position, let alone anything else.

Logic and reality don't much care what it sounds like to you.

Although maybe slightly more than I do.

That would be very wise of you. Not sure you'd have had the authority but still a wise wish for you to have.

Is that what tells the dealer to stop dealing to an empty seat, flipping the hand and paying it out to the vacated chair? Makes you wonder what rules would even require that semantic hack.

i post scripture about jesus having a fit that there's no figs that he curses a tree out of existence and this is what you focus upon?

When did you stop asking people “when did you stop beating your wife?”?