From: herbzet on


Nam Nguyen wrote:
> Daryl McCullough wrote:
> > Jim Burns says...
> >
> >> Charlie-Boo wrote:
> >>> In a consistent system, can a true sentence imply a false one?
> >> The most straightforward answer to your question is "No".
> >>
> >> The reason is hard to express in a way that actually
> >> makes the situation clearer, in much the same way that
> >> it is hard to explain /why/ a bachelor cannot be married.
> >>
> >> If A is a false sentence, then ~A is a true one.
> >> If B is a true sentence and B implies A and A is false,
> >> then we can assert
> >> ~A
> >> B
> >> B -> A
> >>from which it follows
> >> A & ~A
> >>
> >> Another way of looking at it is that having
> >> true sentences imply only true sentences is what
> >> implications are /for/. If they didn't do that, then we
> >> would be spending our time looking at some other
> >> logical function, or something, anything else which
> >> served a similar purpose: pushing out the envelope of
> >> the known.
> >
> > Charlie said *consistent* system, not *sound* system. A consistent
> > system only guarantees that you can't derive a contradiction. There
> > is no requirement that you can't derive false conclusions.
> >
> > So, for instance, if A is a false statement, and A is an *axiom*,
> > and B is a true statement, then of course
> >
> > B -> A
> >
> > is derivable.
> >
> > What you can say is this: If a system is consistent, then a provably
> > true statement can never imply a provably false statement. Not every
> > true statement is provably true, and not every false statement is provably
> > false.
> >
> > On the other hand, a *sound* system has the property that only
> > true statements are provable. So for a sound system, a true statement
> > can never imply a false statement.
>
> It goes without saying that by "true statements" we assume to mean
> "arithmetically true statements".

Hm -- probably you're right that Daryl meant that.

[...]

--
hz
From: Daryl McCullough on
herbzet says...

>Jim Burns wrote:
>
>> If A is a false sentence, then ~A is a true one.
>> If B is a true sentence and B implies A and A is false,
>> then we can assert
>> ~A
>> B
>> B -> A
>> from which it follows
>> A & ~A
>
>Specifically, we have (classically):
>
>1) ~A Given
>2) B Given
>3) B -> A Given

The fact that A is false means that ~A is true, but it doesn't
mean that ~A is derivable in the system.

I guess it was a little ambiguous what it means to say that
B implies A. I assumed that he meant that the implication
B -> A is derivable in the system, not that it is true.

--
Daryl McCullough
Ithaca, NY

From: Daryl McCullough on
In article <4C43AA70.2E4BE5CA(a)gmail.com>, herbzet says...
>
>
>
>Aatu Koskensilta wrote:
>> Charlie-Boo writes:
>>
>> > In a consistent system, can a true sentence imply a false one?
>>
>> Sure.
>
>Nah -- implication is truth-preserving, by definition.

Let's consider the theory T with the following axiom:

0 = S(0)

That's a false sentence.

0 = 0

That's a true sentence. But

0 = 0 -> 0 = S(0)

is derivable in this theory.

--
Daryl McCullough
Ithaca, NY

From: Daryl McCullough on
herbzet says...
>
>
>
>Daryl McCullough wrote:
>> Jim Burns says...
>> >Charlie-Boo wrote:
>
>> >> In a consistent system, can a true sentence imply a false one?
>
>[...]
>
>> What you can say is this: If a system is consistent, then a provably
>> true statement can never imply a provably false statement. Not every
>> true statement is provably true, and not every false statement is provably
>> false.
>
>What is a "provably true statement"? Is this a redundant way of saying
>"provable statement"?

No, a provably true statement is one that is both provable and true.

>Is "provably false statement" a way of saying "negation of a provable
>statement"?

No, a provably false statement is one that is both disprovable (the negation
of a provable statement) and false.

--
Daryl McCullough
Ithaca, NY

From: herbzet on


Daryl McCullough wrote:
> herbzet says...
> >Aatu Koskensilta wrote:
> >> Charlie-Boo writes:
> >>
> >> > In a consistent system, can a true sentence imply a false one?
> >>
> >> Sure.
> >
> >Nah -- implication is truth-preserving, by definition.
>
> Let's consider the theory T with the following axiom:
>
> 0 = S(0)
>
> That's a false sentence.

Yes, as Nam pointed out, it appears that by "false" you mean
"arithmetically false". I didn't assume Charlie-Boo meant this,
which may be my mistake.

> 0 = 0
>
> That's a true sentence. But
>
> 0 = 0 -> 0 = S(0)
>
> is derivable in this theory.

Sure, by positive paradox A -> (B -> A).

Do you see our (yours and mine) statements as contrary assertions?

--
hz

P.S. -- The fun part here is trying to rattle Aatu's cage.