• Security incident: ISF was recently accessed by intruders. Please change your password, and change it anywhere else you used it. Read more

Cont: Deeper than primes - Continuation 2

Status
Not open for further replies.
That's not how definitions are established, nor do you get to substitute your confusion for well-founded Mathematics.
You are simply unaware of your relative-only reasoning, that (in the case of limits) forces the finite on relative infinity.

Moreover, your reasoning does not distinguish between relative infinity (the level of EVERYx) and actual infinity (the level of YESx).

As a result your criticism has nothing to do with http://www.internationalskeptics.com/forums/showpost.php?p=10834527&postcount=698 and http://www.internationalskeptics.com/forums/showpost.php?p=10834793&postcount=699.

No wonder that you are totally missing http://www.internationalskeptics.com/forums/showpost.php?p=10818588&postcount=673.
 
Last edited:
You are simply unaware of your relative-only reasoning, that (in the case of limits) forces the finite on relative infinity.

I am aware of something very significant and important that you are not: Mathematics. It works. Your bizarre ramblings, not so much.
 
Mathematics continues to work, despite your disapproval.
Mathematics continues to be developed beyond your artificial limitations that actually ignore already agreed frameworks like transfinite cardinals, in case of the mathematical notion of Limit.

The weakness of the traditional mathematical notions is very clearly exposed in:

http://www.internationalskeptics.com/forums/showpost.php?p=10670991&postcount=653

http://www.internationalskeptics.com/forums/showpost.php?p=10677971&postcount=662

After all, what can one say to a person that definitions are its religious dogma (as clearly seen in http://www.internationalskeptics.com/forums/showpost.php?p=10669094&postcount=620)?

Such a person simply unable to realize that definitions are not immune from re-examination, re-search, etc. exactly because they have also relative aspect (SOMEx and EVERYx) that deals also with collections.

So, there is no wonder that his reply to http://www.internationalskeptics.com/forums/showpost.php?p=10834793&postcount=699 is http://www.internationalskeptics.com/forums/showpost.php?p=10834806&postcount=700.

jsfisher, mathematics does not care about your dogmatic invariant-only approach of definitions, which are also variant.

As I have said in http://www.internationalskeptics.com/forums/showpost.php?p=10656499&postcount=567, also my definitions have variant aspects so they are not considered as religious dogma, as it is done by your traditional notion of definitions.

Mathematics continues to work
Also Newtonian physics continues to work. Is the fact that it works means that GRT must not be developed?
 
Last edited:
In http://www.internationalskeptics.com/forums/showpost.php?p=10834793&postcount=699 I wrote:
By this definition "that is summed" is at the relative level (the level of collections), where L>0 is at the level of actual infinity (the level of YESx, which is beyond collections)).

It has to be:

By this definition "that is summed" (in the case of the following diagram (the diagram in http://www.internationalskeptics.com/forums/showpost.php?p=10834793&postcount=699)) is at the relative level (the level of collections) because more than one value (where each value>0) is involved in order to get value L>0, where L>0 is at the level of actual infinity (the level of YESx, which is beyond collections)) because L is not one of the summed values (or more generally, it is not one of the collected values)).
 
Last edited:
By using symbolic representation, YESx can be represented, for example, by the outer "(" and ")", which is equivalent to the straight orange line in http://www.internationalskeptics.com/forums/showpost.php?p=10834793&postcount=699 without any sub-lines (notated, for example, as a,b,c,d, ... etc.) along it.

EVERYx (according to the diagram in http://www.internationalskeptics.com/forums/showpost.php?p=10834793&postcount=699) is equivalent to relative infinity of aleph0 sub-lines (notated, for example, as a,b,c,d, ... etc.) along YESx, that logically are not actual infinity YESx (they are between the outer "(" and ")", for example (a,b,c,d,...) ).

In this post the example of relative infinity (EVERYx) is given by a collection of aleph0 elements, but actually any given transfinite cardinal number is no more than relative infinity (it is between the outer "(" and ")", which represents actual infinity or YESx).

More details are given in http://www.internationalskeptics.com/forums/showpost.php?p=10656499&postcount=567 and http://www.internationalskeptics.com/forums/showpost.php?p=10677971&postcount=662.
 
Last edited:
Let's carefully example the notion of Length.

It is based on two building-blocks, which are:

Length=0

Length>0

There are no intermediate states between Length=0 and Length>0, such that Length=0 and length>0 are used as building-blocks (non-composed objects) in order to determine the values of each other, for example:

Code:
2      5    7
.______.____.
    3     2

In this example there are 3 length=0 building-blocks and 2 length>0 building-blocks that determine the values of each other.

It means that Length=0 and length>0 can't be related to values <> 0 without associated with each other (in the example above the 3 length=0 building-blocks are related to values 2, 5 and 7, where the 2 length>0 building-blocks are related to values 3 and 2).

Actually constants like Planck length (https://en.wikipedia.org/wiki/Planck_length) can't be defined without Length=0 and length>0 association, so it is not true that Length=0 is not involved in the determination of Planck length value (which is > 0).

Actually the essence of QM is not some measured value > 0, but it is the lack of intermediate states between Length=0 and Length>0, exactly because both of them are building-blocks (they are naturally non-composed).

Length>0 building-block is at the level of actual infinity, where any collection of Length=0 building-blocks along it, is at most at the level of relative infinity.
 
Last edited:
Actually the essence of QM is not some measured value > 0, but it is the lack of intermediate states between Length=0 and Length>0, exactly because both of them are building-blocks (they are naturally non-composed).

The term you have been so desperately searching for is "discrete".

Length>0 building-block is at the level of actual infinity, where any collection of Length=0 building-blocks along it, is at most at the level of relative infinity.

Word salad.

Everything else you posted is just an extremely clunky, poorly-worded way of pointing out the blindingly obvious. This, however, is nonsense.
 
There are persons that do not understand that a non-composed length>0 (straight of curved) is naturally the continuum, and no collection of length=0 objects along it is the continuum.

For example, the collection of all |R| length=0 objects along a given non-composed length>0 (straight of curved) is not the continuum, and so is the case of all P(|R|), P(P(|R|)), P(P(P(|R|))), ... etc. length=0 objects along a given non-composed length>0 (straight of curved).

Because of the misunderstanding of YESx (the actual continuum), these persons understand the continuum in terms of EVERYx (the level of collections), which is not the actual continuum.

No wonder that these persons misunderstand http://www.internationalskeptics.com/forums/showpost.php?p=10652239&postcount=539 or http://www.internationalskeptics.com/forums/showpost.php?p=10829458&postcount=696.
 
Last edited:
Let's extend http://www.internationalskeptics.com/forums/showpost.php?p=10841149&postcount=708 as follows:

In my opinion the understanding of Entropy (as observed and defined in, so called, closed systems) may have a profound impact on our survival as complex phenomena, and I think that intuition and rigorous treatment have to be combined in order to understand better Entropy, where infinity is a main factor of this research.

Let's carefully examine the notion of Length.

It is based on two building-blocks, which are:

Length=0

Length>0

There are no intermediate states between length=0 and length>0, such that Length=0 and length>0 are used as building-blocks (non-composed objects) in order to determine the values of each other, for example:
Code:
2      5    7
.______.____.
    3     2

In this example there are 3 length=0 building-blocks and 2 length>0 building-blocks that determine the values of each other.

It means that length=0 and length>0 can't be related to values ≠ 0 without associated with each other (in the example above the 3 length=0 building-blocks are related to values 2, 5 and 7, where the 2 length>0 building-blocks are related to values 3 and 2).

Actually constants like Planck length (https://en.wikipedia.org/wiki/Planck_length) can't be defined without length=0 and length>0 association, so it is not true that Length=0 is not involved in the determination of Planck length value (which is > 0).

Actually the essence of QM is not some measured value > 0, but it is the lack of intermediate states between length=0 and length>0, exactly because both of them are building-blocks (they are non-composed).

According to this observation, no amount of collection of length=0 building-blocks along a given length>0 building-block, is that building-block or in other words, the continuum is actually not less than a length>0 building-block.

Moreover, being a length>0 building-block the continuum, is exactly the reason of why between any arbitrary closer given pair of length=0 objects along a given length>0 building-block, there is at least one intermediate length=0 object.

The logic of this observation is as follows:

Given length=0 building-block along a given length>0 building-block, it is at the domain of the length>0 building-block, where the given length>0 building-block is at AND not at the domain of the given length=0 building-block along it.

Being at AND not at a given domain is logical contradiction, if the observation is done from the length=0 building-block with respect to the length>0 building-block.

But being at AND not at a given domain is not logical contradiction, if the observation is done from the length>0 building-block with respect to the length=0 building-block.

No matter what amount of length=0 building-blocks are defined along a given length>0 building-block, each length=0 building-block is at the domain of the given length>0 building-block, where the given length>0 building-block is at AND not at the domain of each given length=0 building-block along it.

One claims that two length=0 building-blocks if used as endpoints of a given length>0 building-block, then it is at the domain of these two length=0 building-blocks.

This claim is true not because of these two length=0 building-blocks, but because any given building-block is its own domain, or in other words, length>0 building-block is its own domain, whether the two length=0 building-blocks are defined as its endpoints, or not.

By using the notion of length=0 and length>0 building-blocks, one can observe Zeno's Achilles\Tortoise Race, as follows:

If only length=0 building-block is used than there is no room for race.

If only length>0 building-block is used than there is no race because there are no racers along it.

So the race is possible only if the two building-blocks are associated with each other, as follows:


Case A: Achilles wins against the Tortoise, and the Race stops.

Case B: Achilles does not win against the Tortoise, and the Race doesn’t stop.



Case A:

Distance = Speed * Time

The next position of Achilles and the Tortoise along the Race = previous position + Distance.

Case A exists only if neither Speed nor Time are changed during the Race. Let us show it by using an algorithm (no particular programming language is used here):

Position X1 = 0

Position X2 = 10

Achilles Speed = Aspeed = 10

Tortoise Speed = Tspeed = 1

Time = 1



Do Loop K from 1 to ∞

Achilles position = position X1 + distance ( = Aspeed * Time)

Tortoise position = position X2 + distance (= Tspeed * Time)

Position X1 = Achilles position

Position X2 = Tortoise position

If X1 > X2 then STOP

Next Loop K



Loop K 1:

Achilles position = 0 + (10 * 1) = 10

Tortoise position = 10 + (1 * 1) = 11

Position X1 = Achilles position = 10

Position X2 = Tortoise position = 11

(The Race continues after Loop K 1, since Tortoise position > Achilles position)



Loop K 2:

Achilles position = 10 + (10 * 1) = 20

Tortoise position = 11 + (1 * 1) = 12

Position X1 = Achilles position = 20

Position X2 = Tortoise position = 12

(The Race stops after Loop K 2, since Tortoise position < Achilles position)




Case B:

Distance = Speed * Time

In case B, Time is changed during the Race. Let us show it by using an algorithm (no particular programming language is used here):

Position X1 = 0

Position X2 = 10

Achilles Speed = Aspeed = 10

Tortoise Speed = Tspeed = 1

Time = 1



Do Loop K from 1 to ∞

Achilles position = position X1 + distance ( = Aspeed * Time)

Tortoise position = position X2 + distance (= Tspeed * Time)

Position X1 = Achilles position

Position X2 = Tortoise position

If X1 ≥ X2 then STOP

Time = Time / Aspeed (Achilles Speed = Aspeed = 10)

Next Loop K



Loop K 1:

Achilles position = 0 + (10 * 1) = 10

Tortoise position = 10 + (1 * 1) = 11

Position X1 = Achilles position = 10

Position X2 = Tortoise position = 11

Time = Time / Aspeed (Achilles Speed = Aspeed = 10) = 0.1

(The Race continues after Loop K 1)



Loop K 2:

Achilles position = 10 + (10 * 0.1) = 11

Tortoise position = 11 + (1 * 0.1) = 11.1

Position X1 = Achilles position = 11

Position X2 = Tortoise position = 11.1

Time = Time / Aspeed (Achilles Speed = Aspeed = 10) = 0.01

(The Race continues after Loop K 2)



Loop K 3:

Achilles position = 11 + (10 * 0.01) = 11.1

Tortoise position = 11.1 + (1 * 0.01) = 11.11

Position X1 = Achilles position = 11.1

Position X2 = Tortoise position = 11.11

Time = Time / Aspeed (Achilles Speed = Aspeed = 10) = 0.001

(The Race continues after Loop K 3)

...



Loop K ∞

The Race doesn’t stop because any given position of Achilles and the Tortoise that comes next is the result of previous positions + Distances values, where each Distance value > 0, no matter how many loops (finite or infinite) are used (Achilles position < Tortoise position is an invariant state exactly because no collection of length=0 building-blocks (where each given position is some distinct length=0 building-block) is the continuum, which is not less than a given length>0 building-block).

Since no collection is the continuum, then we define an entropy-free realm, which can’t be defined by Standard Analysis (its fruitfulness can't deal with entropy-free realm).

The same limitation holds among the notion of mapping between ur-elements like length=0 and length>0 building-blocks (where building-blocks are non-composed objects), such mapping actually ignores the difference between being length=0 building-block and being length>0 building-block, and all it cares is to define the 1-to-1 correspondence between them.
 
Last edited:
I have looked at it again and have found that case b holds only if K < ∞.

If K = ∞, then Achilles wins against the Tortoise, and the Race stops as follows:

Loop K ∞:

Achilles position = 11.111... + (10 * 0.000...1) = 11.111...10

Tortoise position = 11.111... + (1 * 0.000...1) = 11.111...1

Position X1 = Achilles position = 11.111...10

Position X2 = Tortoise position = 11.111...1

If X1 ≥ X2 then STOP

...

Please pay attention that if K=∞ then X1 = X2 does not hold.
 
I worked the math out and got "potato".
"you say potato, I say poteto,

poteto...potato...potato...poteto..."

By "potato" ___ is defined by a collection of "."

By "poteto" ___ is not defined by a collection of "."
 
I have found the notion of downward causation as understood by Prof. Douglas Hofstadter is his book "I am a strange loop".

In this book Hofstadter uses the notion of downward causation which according to it, there are also conditions where higher level of truth is not provable by lower level of truth.

For example, according to Godel's incompleteness theorems, a given mathematical framework of axioms (the lower level) that is strong enough to deal with arithmetic, can't prove or dis-prove the truth of some of its own sentences (the higher level).

In that case the causation of such sentences (the higher level) is not determined by the level of the axioms (the lower level).

Moreover, we get downward causation from the higher level to the lower level, because the non-determination of the truth of the higher level is not changed, no matter how many axioms (the lower level) are added to such mathematical framework.

This is exactly the case among length>0 (the higher level) and a collection of lengths 0 along it (the lower level), they do not determine the truth of length>0, no matter how many lengths 0 (the lower level) are defined along it.

So by using this notion, Mathematics is not less than a framework of downward causation or upward causation, and we define Mathematics itself in terms of a strange loop.
 
Last edited:
It is possible to prove that two infinite sets have the same amount of members by using a given function only once (for example, a bijective function (from set A to set B) AND (from set B to set A) (where, for example, A is the set of all natural numbers and B is the set of all even number (which is also also a proper subset of set A)).

Is it also possible to prove that two infinite sets do not have the same amount of members by using a given function only once (for example, by using a given function only once between S and the power set of S, where S is an infinite set)?
 
It is possible to prove that two infinite sets have the same amount of members by using a given function only once (for example, a bijective function (from set A to set B) AND (from set B to set A) (where, for example, A is the set of all natural numbers and B is the set of all even number (which is also also a proper subset of set A)).

I am uncomfortable with your use of the word 'using'. Two sets (be they finite or infinite) can be shown to be equi-numerous by proving a bijection exists from one set to the other. It is an existence thing, and I don't think that is exactly what you had in mind for 'using'.

Is it also possible to prove that two infinite sets do not have the same amount of members by using a given function only once (for example, by using a given function only once between S and the power set of S, where S is an infinite set)?

No.
 
In order to be clearer, I'll give an example of using only one time a given function between S and it's power set:

a --> {a}
b --> {b}
c --> {c}
...

By using this one particular case in Cantor's theorem, {} is the member of the power set of S that is not mapped with any member of S.

Can't we take this particular case without loss of generality, in order to conclude that |S| < 2|S| ?

If the answer is no, then please explain why, and how many times a given function between S and the power set of S has to be used in order to prove that |S| < 2|S| (where S is an infinite set) ?

-----------

EDIT:

If S is a finite set, for example, S={a,b,c}, it is clear that by using only one time a given function between S and it's power set {{},{a},(b},{c},{a,b},{a,c},{b,c},{a,b,c}}, by the particular case

a --> {a}
b --> {b}
c --> {c}
...

we get (by using Cantor's theorem) {} that is not mapped with any member of S, and by this particular case we can use |{a,b,c}| < |{{},{a},(b},{c}}| in order to conclude that |{a,b,c}| < |{{},{a},(b},{c},{a,b},{a,c},{b,c},{a,b,c}}|.

Is this example among finite sets does not hold among infinite sets because, for example, |S| = |S|+1 ?
 
Last edited:
In order to be clearer, I'll give an example of using only one time a given function between S and it's power set:

a --> {a}
b --> {b}
c --> {c}
...

By using this one particular case in Cantor's theorem

This has nothing to do with Cantor's Theorem.

...{} is the member of the power set of S that is not mapped with any member of S.

Can't we take this particular case without loss of generality, in order to conclude that |S| < 2|S| ?

If the answer is no, then please explain why

The answer is no. It is necessary to prove that no function exists with the required property. That is different from showing one function exists that doesn't have the property.

...and how many times a given function between S and the power set of S has to be used in order to prove that |S| < 2|S| (where S is an infinite set) ?

As I indicated in my previous post, you are introducing some strange concept of "using a function" that has nothing to do with proofs related to cardinality.

-----------

EDIT:

If S is a finite set, for example, S={a,b,c}, it is clear that by using only one time a given function between S and it's power set {{},{a},(b},{c},{a,b},{a,c},{b,c},{a,b,c}}, by the particular case

a --> {a}
b --> {b}
c --> {c}
...

we get (by using Cantor's theorem)

"By using Cantor's theorem" we don't need to consider any mappings between S and P(S). Cantor's theorem tells us |S| < |P(S)| directly.

...{} that is not mapped with any member of S

Cantor's theorem tells us nothing of the sort.

...and by this particular case we can use |{a,b,c}| < |{{},{a},(b},{c}}| in order to conclude that |{a,b,c}| < |{{},{a},(b},{c},{a,b},{a,c},{b,c},{a,b,c}}|.

For finite sets, the existence of an injective mapping that isn't bijective is sufficient to show that no bijective mapping exists.

Is this example among finite sets does not hold among infinite sets because, for example, |S| = |S|+1 ?

No. It is because the existence of an injective mapping that is not bijective is not sufficient to show no bijective mapping exists for infinite sets.
 
Status
Not open for further replies.

ISF - Join now!

Every member here is approved by hand. No bots, no spam, just people who care about evidence and honest debate.

Membership is free!

Create your free account

Back
Top Bottom