Scott Mayers wrote: ↑Mon Feb 25, 2019 10:36 pm
You misinterpret the meaning of 'undecidability' somehow if you think it implies that lack of being able to determine X assures that there IS no X.
Huh? No. That's a strawman! Go the other way.
What I am claiming is that given the possibility of something being an X, Y, Z, P, Q, R, S and T you need SOME machine which can tell the difference between all of those things! Some machine that can sort things into categories. You are taking all of (re)cognition and binary classification for granted! To tell 2 things apart from each other you need 1 bit of information. To tell N things apart from each other you need log(N) bits of information.
The simplest experiment I can do to demonstrate this is to ask this question
Is A = А ? Yes or no.
And you are probably going to fall for the trap everybody falls for because your optical system can't tell the difference between the Latin letter A and the Cyrillic letter А. You are incapable of retrieving the information from reality to even answer the question so your logic is a non-starter because - Garbage in, Garbage out.
This is the heart of empiricism. If you don't have SOME procedure to distinguish one thing form another, they are as good as identical.
Do you believe me when I tell you that they are not the same letter?
Here is a science experiment to prove it:
https://repl.it/repls/ShortLightgrayPiracy
The problem I am trying to point out (with ALL logic) is the GIGO problem.
Scott Mayers wrote: ↑Mon Feb 25, 2019 10:36 pm
The undecidable theorems are about whether we can 'decide' ahead of time what is true about ALL forms of reasoning that are BASED upon a foundation of the most simplest complete logics.
No. That is not what decidability is about. Decidability is STRICTLY about the capability of the machine (mind!) interpreting the logic to do data storage/retrieval and branching. Decidability is about process/procedure that can produce the correct answer.
Decidability is about the fact that where you ASSUME that something is an X, a decidable logic can determine THAT something is an X given a set of instructions.
If you don't have decidability you cannot construct a Universal Turing machine!
If you cannot construct a UTM then your mind is paralyzed to the structure of the logic which you adopted! In the most literal and absolute sense possible without decidability logic is 100% mechanical!
If that is what the "laws" of thought are about, that is an INCREDIBLY inflexible way to reason.
You are after logical completeness. I am after Turing-completeness.
Scott Mayers wrote: ↑Mon Feb 25, 2019 10:36 pm
They don't question that the original first -ordered logic is faulty, but that the upper-floors based upon those foundations are unpredictable. That is, you can't determine FROM the ground floor whether ALL upper floors are
impossible to construct logically because in any logically sound building as a whole, there is no limits to how many floors can be built conceptually.
You don't have that problem with Lambda calculus. Because recursion a Regular language is turing complete. If it's Turing complete it can implement other Regular languages which are also Turing complete.
What Lambda calculus gives you is a language that can INTERPRET ITS OWN SYNTAX AND GRAMMAR.
That's the fundamental feature that classical logics do NOT have.
Because a language can interpret itself CONSISTENTLY you are guaranteed that the language IS consistent.
There can be no contradictions in such a system because the root-causes of contradictions (INCONSISTENCY!) has been removed BY DESIGN.
Scott Mayers wrote: ↑Mon Feb 25, 2019 10:36 pm
The 'limit' is to whether there is an ideal system of reasoning that can prove which kinds of logic systems (outside of the logic you are using to determine this) are
not able to be complete without trying to reconstruct all logical systems and testing them first. That very 'theorem' itself has to be complete or its own conclusions lack validity.
OK. We have a fundamental misunderstanding here. Because you are talking about "outside of the logic you are using to determine this" you are necessarily appealing to Tarski's undefinability theorems and Godel's incompleteness theorems.
Wow! We exist in different paradigms!
Lambda calculus does not have 'foundations' or 'axioms'. It is 100% constructed and 100% conceptual first and foremost. The CONCEPT of computation rests simply on the abstract notion of a Universal Turing Machine.
Lambda calculus (also written as λ-calculus) is a formal system in mathematical logic for expressing computation based on function abstraction and application using variable binding and substitution.
You take the Boolean operators as foundational. I don't.
There are no 'axioms' in lambda calculus. It's a universal language for describing ANY behaviour that you can conceptualize and express.
Do you want Trinary operators? Define them.
Do you want Hextenary operators? Define them.
Do you want Boolean operators? Define them.
At this point I am still speaking strictly about the conceptual/theoretical realm. No rules, no logics, no axioms.
This is what you get with Turing Completeness. Just a canvas and a Universal language to describe how each and every one of the components in your imaginary universe SHOULD behave. There are no laws of physics here - only imagination and limits of computation.
The fact that current-day computers are implemented as binary machines is just implementation detail and it is completely irrelevant.
Because Turing-Completeness is Universal.
All Turing-complete logics are Turing-equivalent!
Turing equivalence – two computers P and Q are called equivalent if P can simulate Q and Q can simulate P
A Quantum computer is Turing-Equivalent to a Classical computer. Just faster!
(I am mis-representing the truth here a little, we are still trying to find whether there is any class of problems that can only be solved by quantum computers). Work is currently underway to prove/disprove whether classical computers can solve BQP-class problems (
https://en.wikipedia.org/wiki/BQP ).
Scott Mayers wrote: ↑Mon Feb 25, 2019 10:36 pm
If you have a 'theorem' that there is no floor above you that is MORE decidably complete above you, you still have to trust the foundation of reasoning of the system of logic you are using to propose such a theorem is correct. If your interpretation of this is ABOUT the foundations, then not even the system of logic the theorem is based on could be trusted, for it would be true about its own system of thinking.
You misunderstand what those theorems are asserting.
No. You misunderstand. Theorems are the consequences of axioms and deductive rules.
I have neither axioms nor rules beyond the ones I put in my mind.
I define the axioms. I construct the models!
And that is how I see all logic - a modeling tool.