Kurt Gödel, Fallacy Of False Dichotomy & Trivalent Logic
I'm not a mathematician and nor am I logician and so cum grano salis.
From what I know (not that much I'm afraid), Gödel's Incompleteness Theorems, one of them I suppose, is predicated on a statement called the Gödel sentence.
Gödel sentence (G): This statement (G) is unprovable.
The argument, as I understand it, proceeds thus:
1. Either G is provable or G is unprovable.
2. If G is provable then G is true i.e. G is unprovable.
3. If G is unprovable then G is true i.e. G is unprovable.
4. G is unprovable or G is unprovable (1, 2, 3 CD)
5. G is unprovable (4 Taut)
QED
However,
The Gödel sentence is a spin-off of the liar sentence (This sentence is false). The assumption that we make with the liar sentence is that it's a proposition and therefore that it has a truth value. Reject that assumption and no contradiction results as there are no truth values that come into opposition.
Since, the Gödel sentence is the liar sentence in some sense can't we do the same thing we did to the liar sentence: take away its status as a proposition?
Note here that, drawing from logical nihilism (@Banno), we can reinterpret what we've done here as merely expanding the possibile "states" a proposition can assume. The new "state" we've introduced in dealing with the liar sentence vide infra is a third truth value which is neither true nor false.
This third "state" for G (the Gödel sentence) would be neither provable nor unprovable i.e. we've slipped between the horns of the dilemma by coming up with a third alternative (trivalent logic)
I guess I'm accusing Gödel of committing the fallacy of false dichotomy.
A penny for your thoughts...
From what I know (not that much I'm afraid), Gödel's Incompleteness Theorems, one of them I suppose, is predicated on a statement called the Gödel sentence.
Gödel sentence (G): This statement (G) is unprovable.
The argument, as I understand it, proceeds thus:
1. Either G is provable or G is unprovable.
2. If G is provable then G is true i.e. G is unprovable.
3. If G is unprovable then G is true i.e. G is unprovable.
4. G is unprovable or G is unprovable (1, 2, 3 CD)
5. G is unprovable (4 Taut)
QED
However,
The Gödel sentence is a spin-off of the liar sentence (This sentence is false). The assumption that we make with the liar sentence is that it's a proposition and therefore that it has a truth value. Reject that assumption and no contradiction results as there are no truth values that come into opposition.
Since, the Gödel sentence is the liar sentence in some sense can't we do the same thing we did to the liar sentence: take away its status as a proposition?
Note here that, drawing from logical nihilism (@Banno), we can reinterpret what we've done here as merely expanding the possibile "states" a proposition can assume. The new "state" we've introduced in dealing with the liar sentence vide infra is a third truth value which is neither true nor false.
This third "state" for G (the Gödel sentence) would be neither provable nor unprovable i.e. we've slipped between the horns of the dilemma by coming up with a third alternative (trivalent logic)
I guess I'm accusing Gödel of committing the fallacy of false dichotomy.
A penny for your thoughts...
Comments (38)
edit: Alternatively, you could consider the sentence itself to be an axiom. E.g. "This is true" has to be taken axiomatically. Likewise, "This is false". If "This is false" is axiomatic, then, axiomatically, it must be pointing to something other than itself, otherwise it is nonsense (unsinn). To claim a different version of the statement "This axiom is false" is true is to deny its status as an axiom. Likewise for provable/unprovable.
I've focused on those aspects of Gödel's argument that are general:
1. A proposition is a sentence that has a truth value.
2. The principle of bivalence (statements can have only two truth values viz. true and false)
What means this?
I mean isn't it really more so the ego's attachment to it's own thoughts. Thinking has always just been useful, but it's not a miracle worker. We work at this problem like we're going to find some underlying wisdom, but what if there is none? What if it's just the illusion of sense created by consonants and vowels? What's the difference between a paradox and gobbledygook?
Sorry, but how does [math]\uparrow[/math] prove that, and I quote, "It just kind of begs the question of thinking itself"?
Quoting TheMadFool This is always true
Quoting TheMadFool
This is never true. We defined G as unprovable. There may be a statement that looks exactly the same like G; but it's not G because G is per definition unprovable.
Quoting TheMadFool
This is always true because G is defined as being unprovable.
Quoting TheMadFool
Yes, we said so in the beginning!
The fallacy here is that
Quoting TheMadFool
this assumes something that is impossible. It's an invalid argument. An error in definition.
Likewise,
the liar paradox boils down to exactly this.
"I am lying" = "G is provable"
I know that you like yourself some paradoxes - but are paradoxes not just pure fiction? It seems like they exclusively stem either from a lack of knowledge or from assuming statements that are impossible to begin with.
It is, I fear, not what you think it is.
Quoting Hermeticus
Nothing that wasn't obvious.
Quoting Hermeticus
Okey dokey.
Quoting Hermeticus
Yep.
Quoting Hermeticus
It's not a statement that we made so much as it is a statement we discovered.
Quoting Hermeticus
What exactly have we assumed that's impossible?
The Liar Paradox
The liar sentence: This1 sentence is false.
The negation of the liar sentence: This2 sentence is true.
For there to be a paradox, This1 = This2 but they're, I discovered, not.
This1 refers to :point: This1 sentence is false ( L ).
This2 refers to :point: This2 sentence is true (~L).
(Notice the difference? False (This1) and True (This2)
Imagine the following:
One person saying "I am black and I am not black" (contradiction).
and
Two people saying "I am black" and "I am not black" (no contradiction).
---
Gödel's proof
Gödel sentence = G = This3 sentence (G) is unprovable = G is unprovable.
The negation of the Gödel sentence = ~G = This4 sentence is unprovable = G is provable.
Gödel's argument requires that This3 = This4, he needs a contradiction.
However,
This3 refers to :point: This3 sentence is unprovable (G)
This4 refers to :point: This4 sentence is provable (~G)
Note here too a difference: Unprovable (This3) and Provable (This4).
The liar sentence's problem is precisely because of self-reference (namely itself) which the Gödel sentence also repeats/reenacts.
Moreover, Godel's proof is purely about syntax and the proof itself does not rely on a notion of truth. That is the important way the Godel sentence is different from the liar statement. Whatever is problematic about the liar statement does not apply to the Godel sentence since the Godel sentence does not mention truth conditions nor is the Godel sentence self referential in the particulary problematic way of the liar statement. (It is only in an addition to the basic proof about syntax that we go on to note the the Godel sentence does happen to be true. And the conclusion from that is not "The Godel sentence is not provable in the system". Rather the conclusion is "If the system is consistent then the Godel sentence is not provable in the system".)
And Godel's proof can be formulated in primitive recursive arithmetic (purely finitistic combinatoric arithmetic) and with only intuitionistic logic. That would seem to be the minimum platform for mathematical reasoning. If there is any objection to Godel's proof then it would entail objection to even bare finitistic reasoning about natural numbers.
And fundamentally wrongheaded is the argument that Godel uses "false dichotomy" in the manner described in the first post. The proof is in intuitionistic logic, which does not provide "P or not-P" as a basis.
It is correct though that if one rejects the law of non-contradictions, such as certain systems of paraconsistent logic, then Godel's argument does not hold. Yes, one way out of Godel's theorem is to instead propose that mathematics be formulated paraconsistently. But that still does not impugn Godel's proof as much as it merely says there are alternative contexts in which the argument does not hold.
And beyond those points, the preponderance of posts above in this thread are an abysmally mangled misrepresentation of any aspect of Godel's proof. This is yet another example of abuse of the Internet to post serial misinformation and confusion. Posters who wish to present objections to Godel's argument should at least study the subject well enough to have informed and coherent notions about it.
The best book for the layman about Godel's incompleteness proofs is:
Godel's Theorem - Torkel Franzen
There are two separate and distinct definitions/usages of the word truth and its derivations.
The Correspondence Theory of Truth
This is the plain language and accepted standard usage of the word truth. When you swear to tell the truth, the whole truth, and nothing but the truth? You are asserting that your statements - the sentences that you say, write, sign, etc - will accurately describe events in the real world (AKA existence, AKA the universe, AKA "everything that is the case", etc, etc)
Usage in Logic & Mathematics
Propositions in logic/math are true if they are derivable from the basic axioms & rules of the particular system you happen to be working in. I am not an expert in this area, but here is a good starting point if you want to learn: https://en.wikipedia.org/wiki/Peano_axioms
The so called Liar Sentence "This1 sentence is false" is clearly not formulated in any sort of mathematical/logic framework, so it must be treated as an plain language statement. As such, there is no paradox at all, since it does not make any assertion about any event in the real world. As such it does not take a truth value, it is neither true or false. It is a collection of words that is constructed to make a grammatically correct sentence - you can think of it as a sort of poetry.
William Hughes Mearns
[i]Yesterday, upon the stair,
I met a man who wasn't there
He wasn't there again today
I wish, I wish he'd go away...[/i]
As far as your comments about Godel's Theorem go, I suggest you take @TonesInDeepFreeze's comment above to heart. You might also want to read Gödel, Escher, Bach by Douglas Hofstadter. Besides giving an excellent overview of Gödel, it's a great read, very entertaining.
1. The Liar Paradox
The Liar sentence: L = This sentence[1] is false.
The liar argument:
1. L is true (assume for reductio ad absurdum)
2. If L is true then L is false (premise)
3. L is false (1, 2 MP)
4. L is true & L is false (1, 2 Conj)
5. L is false (1 - 4 reductio ad absurdum)
6. If L is false then L is true (premise)
7. L is true (5, 6 MP)
8. L is true & L is false (5, 7 Conj)
9. L is true ( 5 - 8 reductio ad absurdum)
10. L is true & L is false (5, 9 Conj) [The liar paradox]
So far so good.
The Liar sentence: This sentence[1] is false.
What does L is false mean? It means ~L is true.
~L = This sentence[2] is false.
The paradox (line 10) = L is true & L is false = L & ~L
L & ~L = This sentence[1] is false & This sentence[2] is true.
However, L & ~L, despite how it looks, is not a contradiction. For it to be a contradiction, This sentence[1] = This sentence[2] but they're not:
This sentence[1] = This sentence[1] is false.
This sentence[2] = This sentence[2] is true.
This sentence[1] and This sentence[2] refer to two different sentences as shown above.
2. Gödel's argument.
Gödel sentence: G = This sentence[3] is unprovable.
What is the negation of G? This sentence[4] is provable.
If, in Gödel's proof, it's necessary that This sentence[3] = This sentence[4] (perhaps to arrive at a contradiction) then Gödel's proof fails because:
This sentence[3] = This sentence[3] is unprovable
This sentencea[4] = This sentence[4] is provable
This sentence[3] and This sentence[4] are not the same sentence.
As I (along with several other people) explained, this is wrong. L cannot take a truth value since it has no semantic content that can possibly be verified or denied.
Quoting TheMadFool
Yes. This stuff is difficult. I am not an expert in this, but you can become better educated. Both @TonesInDeepFreeze & I gave you several excellent books that can point you in the right direction.
http://arxiv.org/abs/1510.04393
Some make similar arguments like you. Haven't been successful.
You see, even if it comes close to be like the liar paradox, it isn't a paradox. It just clearly shows the limits of provability. And that's it.
That's the whole point of Gödel's incompleteness theorems.
I think an issue is with treating Truth/Falsity like an objest as Frege did. It creates contradictions by asserting the existence of false things (and if false things exist then what is truth aptness really doing).
That is my thoughts exactly, yet so far we have not even done this
to the actual Liar Paradox itself: "This sentence is not true."
When we formalize the exact same structure as the Liar Paradox
in Prolog we see that Prolog rejects it as unsound because it has
a cycle in the directed graph of its evalulation sequence.
?- LP = not(true(LP)).
LP = not(true(LP)).
?- unify_with_occurs_check(LP, not(true(LP))).
false.
We get the exact same thing with your simplified Gödel sentence.
He sums up his own work this same way:
...We are therefore confronted with a proposition which asserts its own unprovability. 15 ...
(Gödel 1931:43-44)
Gödel, Kurt 1931.
On Formally Undecidable Propositions of Principia Mathematica And Related Systems
This :up: . It's a statement about provability for statements in a certain class of consistent systems (those than can encompass arithmetic), using "effective procedures," roughly those that can be described with a set of instructions where the results should be the same each time and are completed in a finite number of steps and which don't require any creativity (i.e. introducing things outside the instructions).
x := y means x is defined to be another name for y
https://en.wikipedia.org/wiki/List_of_logic_symbols
This seem to be the only way that we get actual self-reference
all of the textbooks merely approximate self-reference with ?
G := ¬(F ? G) // G asserts its own unprovability in F
Proving G in F requires a sequence of inferences steps in F that proves there is no such sequence of inference steps in F.
Copyright 2023 PL Olcott
...We are therefore confronted with a proposition which asserts its own unprovability. 15 ...
(Gödel 1931:43-44)
Gödel, Kurt 1931.
On Formally Undecidable Propositions of Principia Mathematica And Related Systems
To the original question, while I'm not super familiar with paraconsistent logic, I do think this changes things. Paraconsistent logics can retrieve naive set theory. They aren't susceptible to the principle of explosion, that anything can be proved from a contradiction.
The axiom of choice also becomes trivial in this case. Disjunctive syllogism also goes out the door though.
Such systems aren't necessarily complete, but I do believe they can be complete in cases where a classical version would not be. Here is a dialetheist (allowing for "true" contradictions) arithmetic that and can prove itself, but which is inconsistent.
Consistency is less important when explosion isn't a factor (to some degree) and the undefinability of truth is also less of an issue.
https://academic.oup.com/mind/article-abstract/111/444/817/960902
Reading some out there stuff on category theoretic attempts to formalize Hegel that I didn't understand, apparently that sort of system gets at these "perks" of not having contradiction result in explosion.
Stop right there. It's about limitations in mathematics.
To talk about "certain classes of consistent system" can mislead someone to thinking Gödel is talking about something obscure. Yet it is the limited obscure fields in Mathematics which don't encompass arithmetic, which are the fields that need long descriptions to formalize them. And just what you can do with them (as they are likely to be extremely simplistic) more than give a theoretical description about them is usually even more difficult.
"a proposition which asserts its own unprovability. (Gödel 1931:43-44)"
Quoting PL Olcott
That this proposition is self-contradictory:
Proving G in F requires a sequence of inferences steps in F that proves there is no such sequence of inference steps in F.
thus the assessment that its formal system is incomplete is incorrect.
If the mathematical notion of incomplete is incorrect for one input then it
cannot be trusted for other inputs.
Copyright 2023 PL Olcott
I don't think anything I said gives the impression that the above is not the case. I was just thinking in terms of the ways that philosophers have attempted to generalize Godel (and Tarski's) findings beyond the scope of mathematics. Plus, more importantly, given the context of this thread's topic, that this doesn't hold the same way for paraconsistent systems (granted such systems won't be able to handle arithmetic and be consistent.)
See: my last post, TonesInDeepFreeze's post, etc.
Fair enough. But usually there isn't much discussion of just what is the impact of this (or similar) findings.
How can 'This statement' be true or false?
Very naive thoughts I know, but I seem to have a mental block on this topic.
It's certainly a valid take on the subject, don't consider your intuition here naive. You're very possibly right to find it incomprehensible.