|
Post by Eugene 2.0 on Aug 19, 2021 15:43:29 GMT
It's possible to use axioms to formalize something. Here's three ways of doing it:
1. Assuming every statement is an axiom 2. Assuming that there are axioms and non-axioms (i.e. free statements) 3. Assuming that there's no axioms
The third is irrelevant. The first is useless, so only the second is good. But here's another problem that is unavoidable:
A. We can add new one axiom that says what we should maintain as axioms. (Another way is to introduce a criterion for it.) B. No adding extra axioms.
According to Gödel there's no way to avoid using new axioms, at least, this extra axiom is implicit to any powerful system that bases as on arithmetics so on the formal logic with predicates.
And each such extra axiom would be necessary a free statement. Why so? – It has to be clear:
i) this extra statement is either an axiom, or not ii) if this statement is an axiom, there must be another axiom that confirms this new axiom as an extra axiom iii) if the extra isn't an axiom – is a free statement
|
|
|
Post by karl on Aug 19, 2021 22:21:10 GMT
If you have an axiomatic system, and you add an axiom, you don't need to specify that it's an axiom. It's enough that it's made clear that it's true. If this is part of a software, then the computer will deduce the same from that axiom whether it's stated to be an axiom or not.
|
|
|
Post by Eugene 2.0 on Aug 20, 2021 4:26:01 GMT
If you have an axiomatic system, and you add an axiom, you don't need to specify that it's an axiom. It's enough that it's made clear that it's true. If this is part of a software, then the computer will deduce the same from that axiom whether it's stated to be an axiom or not. Not exactly. There's a property of independency of axioms such as if {a_1, a_2, ..., a_n} are axioms, then for any two axiom is false that a_i → a_j. By adding new axiom you either change your theory, or not. To prove it, that's true, you can use your computer to check it, while it may appear to be false or non-sequitur (i.e not inferenced), and this time the computer won't accept it as a true or as another axiom till you put him a command to accept it.
|
|
|
Post by karl on Aug 20, 2021 17:05:14 GMT
If you have an axiomatic system, and you add an axiom, you don't need to specify that it's an axiom. It's enough that it's made clear that it's true. If this is part of a software, then the computer will deduce the same from that axiom whether it's stated to be an axiom or not. Not exactly. There's a property of independency of axioms such as if {a_1, a_2, ..., a_n} are axioms, then for any two axiom is false that a_i → a_j. By adding new axiom you either change your theory, or not. To prove it, that's true, you can use your computer to check it, while it may appear to be false or non-sequitur (i.e not inferenced), and this time the computer won't accept it as a true or as another axiom till you put him a command to accept it.
Consider an arithmetic system based on 10 axioms. You add a statement to the system which you tell the computer is true. That statement happens to not be deducible from the 10 existing axioms, and is hence an 11th axiom.
My point was that you don't have to tell the computer that the statement is, in fact, an axiom. Meaning, that it's not deducible from the other 10th axioms. You only need to make clear that it's true.
|
|
|
Post by Eugene 2.0 on Aug 20, 2021 17:25:11 GMT
Not exactly. There's a property of independency of axioms such as if {a_1, a_2, ..., a_n} are axioms, then for any two axiom is false that a_i → a_j. By adding new axiom you either change your theory, or not. To prove it, that's true, you can use your computer to check it, while it may appear to be false or non-sequitur (i.e not inferenced), and this time the computer won't accept it as a true or as another axiom till you put him a command to accept it.
Consider an arithmetic system based on 10 axioms. You add a statement to the system which you tell the computer is true. That statement happens to not be deducible from the 10 existing axioms, and is hence an 11th axiom.
My point was that you don't have to tell the computer that the statement is, in fact, an axiom. Meaning, that it's not deducible from the other 10th axioms. You only need to make clear that it's true.
You loaded to a computer 10 axioms. And let's assume computer is not a mixer or a blender of axioms, but it checks any different statements to answer which of them are true, and along with it, which of them are axioms, which are not, etc. If you put a sequence to a computer it can reject it signalizing about it (somehow, or a master can discover the problem; if it's some upper level langs, then this task would be easier). So, the first important things that the computer check different statements. It has to to be able to prove theorems. Also, it's not like that 10th axioms are true for a computer. Computer doesn't know what is true, or what is not. It can be like 1st axiom has a meaning of T, 2nd has a meaning of T, and so on. The important thing is to make a computer able to decide which statements can have T value (or meaning). A certain statement would have the value T if it could be derived from a set of statements which ones had the value of T. And we have to input also another statement (i.e. an axiom) to explain what statements will have T value, or by which mechanism we add T value to statements.
|
|
|
Post by karl on Aug 20, 2021 18:17:55 GMT
Consider an arithmetic system based on 10 axioms. You add a statement to the system which you tell the computer is true. That statement happens to not be deducible from the 10 existing axioms, and is hence an 11th axiom.
My point was that you don't have to tell the computer that the statement is, in fact, an axiom. Meaning, that it's not deducible from the other 10th axioms. You only need to make clear that it's true.
You loaded to a computer 10 axioms. And let's assume computer is not a mixer or a blender of axioms, but it checks any different statements to answer which of them are true, and along with it, which of them are axioms, which are not, etc. If you put a sequence to a computer it can reject it signalizing about it (somehow, or a master can discover the problem; if it's some upper level langs, then this task would be easier). So, the first important things that the computer check different statements. It has to to be able to prove theorems. Also, it's not like that 10th axioms are true for a computer. Computer doesn't know what is true, or what is not. It can be like 1st axiom has a meaning of T, 2nd has a meaning of T, and so on. The important thing is to make a computer able to decide which statements can have T value (or meaning). A certain statement would have the value T if it could be derived from a set of statements which ones had the value of T. And we have to input also another statement (i.e. an axiom) to explain what statements will have T value, or by which mechanism we add T value to statements.
Well, you're into the technical aspects of how this is actually done in a computer, which I don't know enough about. So I don't know whether we actually disagree, or if we're using different words for the same thing. It's the first time I've heard the term "T value".
I just want to point out that the reason why every consistent arithmetic system (of minimum complexity) is incomplete is that there will always exist statements that are true within that system, but that the system itself could not express. Trying to find that statement within the system, would lead to a never-ending sequence of calculations, because within the system it would be of infinite complexity. But when not constrained by the existing axioms, it can be formulated, and then added to the system as a new axiom.
|
|
|
Post by Eugene 2.0 on Aug 20, 2021 19:14:23 GMT
You loaded to a computer 10 axioms. And let's assume computer is not a mixer or a blender of axioms, but it checks any different statements to answer which of them are true, and along with it, which of them are axioms, which are not, etc. If you put a sequence to a computer it can reject it signalizing about it (somehow, or a master can discover the problem; if it's some upper level langs, then this task would be easier). So, the first important things that the computer check different statements. It has to to be able to prove theorems. Also, it's not like that 10th axioms are true for a computer. Computer doesn't know what is true, or what is not. It can be like 1st axiom has a meaning of T, 2nd has a meaning of T, and so on. The important thing is to make a computer able to decide which statements can have T value (or meaning). A certain statement would have the value T if it could be derived from a set of statements which ones had the value of T. And we have to input also another statement (i.e. an axiom) to explain what statements will have T value, or by which mechanism we add T value to statements.
Well, you're into the technical aspects of how this is actually done in a computer, which I don't know enough about. So I don't know whether we actually disagree, or if we're using different words for the same thing. It's the first time I've heard the term "T value".
I just want to point out that the reason why every consistent arithmetic system (of minimum complexity) is incomplete is that there will always exist statements that are true within that system, but that the system itself could not express. Trying to find that statement within the system, would lead to a never-ending sequence of calculations, because within the system it would be of infinite complexity. But when not constrained by the existing axioms, it can be formulated, and then added to the system as a new axiom.
Karl, every your notification - is indeed important to me. I mean it's just a bad style of mine to speak in such a manner (I hope I don't hurt anyone by it). That's explanation wasn't strictly from Godel that it must've been seen. But one idea had been floating in my mind for years, and I couldn't find it to be very similar to the Godel's project. Besides, it was about the same: the completeness of a system and the value. - And again, I might be wrong, so all in the power of that idea. I'll try to explain it. When we say that "water boils at 100 degrees celcius" is almost the same as to say "the 'water' and 'the process of boiling' = '100' and 'Celcius'". (Or in other words 'a property p of A' = 'a property q of B'.) The same would be in many other cases. (Surely, we can use some other operators, than the equality. The equality is a useful operator, and it has an important logical ability - it saves the value over some changes.) So, ok what does it mean that "water boils at 100 degrees of celcius"? It mean we have to look at some meters or some measuring devices. Let's say it would be a plain thermometer. But what does it mean '100 degrees of Celcius'? It means when the water reaches the point the mercury liquid in the thermometer reaches some limit. But any of such things wouldn't save us from illusion or some mistake. We won't never find anything in the world that would be safe. All the ethalons or measuring ideals can be broken, dried out, or disformed. We have to determine everything: starting from 'water', and ending with 'numbers'. As we know it's impossible to finish: no a posteriori truth is a necessary truth. So, that's why a mistake might occur any time. But the same is about any theory since we have to express our ideas. I mean since we use language (i.e. names, sentences), we can confuse the names or sentences by the same reasons we confuse measuring ideals. (And we also can encounter some paradoxes with names like: "9 in necessary greater, than 7", "9 - is a number of planets", "The number of planet is necessary greater, than 7".) So, we don't have any insurances to prevent such weird cases. As soon as we have the predicate logic we would achieve a powerful tool that allows us to imply the formulas that aren't impossible to get in the propositional calculus: in the predicate logic we can deduce c from (a&b): (a&b)->c. But to do this we have to use quantifications. In other words, we achieve almost the same procedures we use in the situation with physics when we equal the boiling of water to the lenght of the thermometer; we can equal or disequal some values. If we're enough tricky we can use that form to imply some unusual and weird results. It is important - that in any step we can deduce the rules of the systems. It should be clear: if the rules of the systems are deducible, than they are true. So, even if we won't add any new extra axiom, it still means that we are able to imply the rules of the system. And as soon as it's possible we can deduce something about the rules themselves. And achieving reflexive forms of statements we can interpret them as self-contradiction statements: they would say that those statements had no T value. In other words, by adding new extra axiom we just say something about our own system. I think that the most confusing part of the topic of mine was the word "axiom". I take a notion 'axiom' in a quite broad sense: it's the same as a statement that has a number < than a number n. (Considering that n - is a number of axioms. So, in this case if a computer would check a statement it would check a statement's number, and if it would satisfy x < n, then it would take the statement as something that has the value of T.)
|
|
|
Post by jonbain on Aug 21, 2021 9:32:38 GMT
Simply explained thus:
You cannot use a finite calculation to account for infinite possibilities.
|
|
|
Post by Eugene 2.0 on Aug 21, 2021 10:29:57 GMT
Simply explained thus: You cannot use a finite calculation to account for infinite possibilities. It's possible. I'd also try to make my post simply to this: It's impossible to define things using finite undefined definitions.
|
|
|
Post by jonbain on Aug 21, 2021 12:53:08 GMT
Simply explained thus: You cannot use a finite calculation to account for infinite possibilities. It's possible. I'd also try to make my post simply to this: It's impossible to define things using finite undefined definitions.
You cannot write a computer program that can resolve all errors in all other computer algorithms.
The set of infinite objects cannot be contained in a finite set of finite objects.
Its not possible to write an algorithm for all events in the universe, if that algorithm is itself contained in the universe.
|
|
|
Post by Eugene 2.0 on Aug 21, 2021 14:38:32 GMT
It's possible. I'd also try to make my post simply to this: It's impossible to define things using finite undefined definitions.
You cannot write a computer program that can resolve all errors in all other computer algorithms.
The set of infinite objects cannot be contained in a finite set of finite objects.
Its not possible to write an algorithm for all events in the universe, if that algorithm is itself contained in the universe.
This is exactly true - it cannot resolve the other algorithms. Let's say all what he can is to describe the work of the other computers (along with the work of something else that it's able to check) in the language it's been programmed to check. In other words, if a computer could re-axiomatized itself, he could resolve the other programms. So, that was exactly I was going to say. Thank you for introducing the computer programms example.
|
|
|
Post by karl on Aug 21, 2021 16:04:54 GMT
Well, you're into the technical aspects of how this is actually done in a computer, which I don't know enough about. So I don't know whether we actually disagree, or if we're using different words for the same thing. It's the first time I've heard the term "T value".
I just want to point out that the reason why every consistent arithmetic system (of minimum complexity) is incomplete is that there will always exist statements that are true within that system, but that the system itself could not express. Trying to find that statement within the system, would lead to a never-ending sequence of calculations, because within the system it would be of infinite complexity. But when not constrained by the existing axioms, it can be formulated, and then added to the system as a new axiom.
Karl, every your notification - is indeed important to me. I mean it's just a bad style of mine to speak in such a manner (I hope I don't hurt anyone by it). That's explanation wasn't strictly from Godel that it must've been seen. But one idea had been floating in my mind for years, and I couldn't find it to be very similar to the Godel's project. Besides, it was about the same: the completeness of a system and the value. - And again, I might be wrong, so all in the power of that idea. I'll try to explain it. When we say that "water boils at 100 degrees celcius" is almost the same as to say "the 'water' and 'the process of boiling' = '100' and 'Celcius'". (Or in other words 'a property p of A' = 'a property q of B'.) The same would be in many other cases. (Surely, we can use some other operators, than the equality. The equality is a useful operator, and it has an important logical ability - it saves the value over some changes.) So, ok what does it mean that "water boils at 100 degrees of celcius"? It mean we have to look at some meters or some measuring devices. Let's say it would be a plain thermometer. But what does it mean '100 degrees of Celcius'? It means when the water reaches the point the mercury liquid in the thermometer reaches some limit. But any of such things wouldn't save us from illusion or some mistake. We won't never find anything in the world that would be safe. All the ethalons or measuring ideals can be broken, dried out, or disformed. We have to determine everything: starting from 'water', and ending with 'numbers'. As we know it's impossible to finish: no a posteriori truth is a necessary truth. So, that's why a mistake might occur any time. But the same is about any theory since we have to express our ideas. I mean since we use language (i.e. names, sentences), we can confuse the names or sentences by the same reasons we confuse measuring ideals. (And we also can encounter some paradoxes with names like: "9 in necessary greater, than 7", "9 - is a number of planets", "The number of planet is necessary greater, than 7".) So, we don't have any insurances to prevent such weird cases. As soon as we have the predicate logic we would achieve a powerful tool that allows us to imply the formulas that aren't impossible to get in the propositional calculus: in the predicate logic we can deduce c from (a&b): (a&b)->c. But to do this we have to use quantifications. In other words, we achieve almost the same procedures we use in the situation with physics when we equal the boiling of water to the lenght of the thermometer; we can equal or disequal some values. If we're enough tricky we can use that form to imply some unusual and weird results. It is important - that in any step we can deduce the rules of the systems. It should be clear: if the rules of the systems are deducible, than they are true. So, even if we won't add any new extra axiom, it still means that we are able to imply the rules of the system. And as soon as it's possible we can deduce something about the rules themselves. And achieving reflexive forms of statements we can interpret them as self-contradiction statements: they would say that those statements had no T value. In other words, by adding new extra axiom we just say something about our own system. I think that the most confusing part of the topic of mine was the word "axiom". I take a notion 'axiom' in a quite broad sense: it's the same as a statement that has a number < than a number n. (Considering that n - is a number of axioms. So, in this case if a computer would check a statement it would check a statement's number, and if it would satisfy x < n, then it would take the statement as something that has the value of T.)
If a statement can be deduced from the axioms in a consistent system, that statement is true within that system. To add a statement that is not deducible from the existing axioms as a new axiom in the system, is the same as to declare it as true within that system.
|
|
|
Post by Eugene 2.0 on Aug 21, 2021 16:43:21 GMT
Karl, every your notification - is indeed important to me. I mean it's just a bad style of mine to speak in such a manner (I hope I don't hurt anyone by it). That's explanation wasn't strictly from Godel that it must've been seen. But one idea had been floating in my mind for years, and I couldn't find it to be very similar to the Godel's project. Besides, it was about the same: the completeness of a system and the value. - And again, I might be wrong, so all in the power of that idea. I'll try to explain it. When we say that "water boils at 100 degrees celcius" is almost the same as to say "the 'water' and 'the process of boiling' = '100' and 'Celcius'". (Or in other words 'a property p of A' = 'a property q of B'.) The same would be in many other cases. (Surely, we can use some other operators, than the equality. The equality is a useful operator, and it has an important logical ability - it saves the value over some changes.) So, ok what does it mean that "water boils at 100 degrees of celcius"? It mean we have to look at some meters or some measuring devices. Let's say it would be a plain thermometer. But what does it mean '100 degrees of Celcius'? It means when the water reaches the point the mercury liquid in the thermometer reaches some limit. But any of such things wouldn't save us from illusion or some mistake. We won't never find anything in the world that would be safe. All the ethalons or measuring ideals can be broken, dried out, or disformed. We have to determine everything: starting from 'water', and ending with 'numbers'. As we know it's impossible to finish: no a posteriori truth is a necessary truth. So, that's why a mistake might occur any time. But the same is about any theory since we have to express our ideas. I mean since we use language (i.e. names, sentences), we can confuse the names or sentences by the same reasons we confuse measuring ideals. (And we also can encounter some paradoxes with names like: "9 in necessary greater, than 7", "9 - is a number of planets", "The number of planet is necessary greater, than 7".) So, we don't have any insurances to prevent such weird cases. As soon as we have the predicate logic we would achieve a powerful tool that allows us to imply the formulas that aren't impossible to get in the propositional calculus: in the predicate logic we can deduce c from (a&b): (a&b)->c. But to do this we have to use quantifications. In other words, we achieve almost the same procedures we use in the situation with physics when we equal the boiling of water to the lenght of the thermometer; we can equal or disequal some values. If we're enough tricky we can use that form to imply some unusual and weird results. It is important - that in any step we can deduce the rules of the systems. It should be clear: if the rules of the systems are deducible, than they are true. So, even if we won't add any new extra axiom, it still means that we are able to imply the rules of the system. And as soon as it's possible we can deduce something about the rules themselves. And achieving reflexive forms of statements we can interpret them as self-contradiction statements: they would say that those statements had no T value. In other words, by adding new extra axiom we just say something about our own system. I think that the most confusing part of the topic of mine was the word "axiom". I take a notion 'axiom' in a quite broad sense: it's the same as a statement that has a number < than a number n. (Considering that n - is a number of axioms. So, in this case if a computer would check a statement it would check a statement's number, and if it would satisfy x < n, then it would take the statement as something that has the value of T.)
If a statement can be deduced from the axioms in a consistent system, that statement is true within that system. To add a statement that is not deducible from the existing axioms as a new axiom in the system, is the same as to declare it as true within that system.
Agree. There are some misconceptions or some misusing of terms. But a system with n axioms and system with n+1 axioms, as well as if a system #1 has a set of {a, b, ..., z} axioms, and a set #2 has {ф, х, ъ, л, щ...} axioms they might be different. And also, if the property of implication is the same as to be true for a statement, then if a system X has some axioms, and these axioms can be deduced in this system (e.g. using an axiom x → x), so adding new extra axiom means: a) either this new axiom A is a part of the others axioms {a, b, c...}, so then {a, b, c...} → A, and it's true, because of it's a part of the rest axioms; b) or it is an independent axiom, and this time {a, b, c, A} → A works, because A is already withing the circle of the accepted axioms.
|
|