# Fuzzy Logic

1. Dec 23, 2003

### caumaan

I don't really know a lot about computers, so maybe I should just say this...

What is fuzzy logic and what are some of its applications?

2. Dec 23, 2003

### Jimmy

3. Dec 24, 2003

### Nephtys

I always thought that fuzzy logic was essentially binary code carried to the next level, with each digit having a number of possibilities between 0 and 1. Isn't that right?

4. Dec 24, 2003

### Jimmy

This is taken from the link I gave in my previous post:

It seems you are correct, Nephtys.

Last edited: Dec 24, 2003
5. Dec 24, 2003

### StarThrower

Fuzzy logic is total nonsense.

Any statement is either true or false. There is no such thing as a statement being 1/3 true and 2/3 false, and that is exactly what fuzzy logic claims is possible.

Basically, of all possible logics that a reasoning agent can use, only one will be free from contradiction, and that is binary logic.

The following statement is true:

For any statement x, either X is true or X is false, and not (X is true and X is false.

We can condense this, using the 'exclusive or' operator XOR as follows:

For any statement X, |X|=0 XOR |X|=1

If you choose any logic other than binary logic, the statement above is false.

Thus, in any logic other than binary logic, there is at least one statement X such that not (|X|=0 XOR |X|=1), from which it follows that

not (|X|=0) if and only if not (|X|=1).

Now, if you do a temporal modal analysis of the meaning of the 'if and only if' operator, you will see that in an expression "A if and only if B" which is true, implies that the truth values of A,B are true simultaneously. And it is easy to show that

A if and only if B = not A if and only if not B

thus we have:

|X|=0 if and only if |X|=1

So what this tells us is that the statement

|X|=0

and the statement

|X|=1
are true simultaneously, therefore it follows that

0=|X| and |X|=1

And now using the transitive axiom of equality we can conclude that

0=1, which is false in all systems of logic.

6. Dec 24, 2003

### Hurkyl

Staff Emeritus
If you treat all the logical operators as if they were operating on binary logical values, then of course you're gonna derive a contradiction if the values aren't restricted to be 0-1.

Fuzzy logic is one way to avoid hard decision making. For example, 5'9" is the average height for a male (or it was at one time), and 6'3" is tall. What do you call a person who is 5'11"? Average height? Tall? Does it even make sense to have a cutoff? Fuzzy logic lets you assign a number between 0 and 1 for both "average" and "tall", thus avoiding making this hard decision.

7. Dec 24, 2003

### Jeebus

Exactly.

{Part II}Or you can go with the Weisstein definition:

Such a fuzzy logic systems can be extremely useful in designing controlled variables for real-world systems such as elevators. Same type of analogy Hurkyl gave above.

8. Dec 25, 2003

### phoenixthoth

hurkyl, that's exactly the example i was going to give.

if every statement is either true or false, then tell me whether the following statements are true or false, ok?

1. i am tall (i'm 5'11'')
2. i am very tall
3. i am somewhat tall
4. i am roughly of average height

5. i am beautiful
6. i am one of the most beautiful people in the world

7. i always lie
8. this statement is false

9. the barber who shaves every man who doesn't shave himself shaves himself
10. the barber who shaves every man who doesn't shave himself does not shave himself

11. the set of all sets that are not members of itself is a member of itself.
12. the set of all sets that are not members of itself is not a member of itself.

13. this statement is true but not provable.
14. this statement is true but you can prove that it is false.

or did you just mean that all well formed formula in binary logic are either true or false?

this is the third thread with basically the same topic now.

Last edited: Dec 25, 2003
9. Dec 29, 2003

### StarThrower

There is a difference between a sentence and a statement. A statement must either be true or false, and no statement can be true and false simultaneously. Now, in order to have truth value, a statement must have meaning. Consider the sentence, "I am tall". It is neither true nor false, hence it isn't a statement. Now, you could try to define 'tall' using ruler measurement, and say that such and such is tall if and only if its length is greater than the length of this ruler, but clearly if you feel these sentences are paradoxes of some kind, then that isn't what you mean by the adjective 'tall'. Thus, all your sentences which involve the word 'tall' are not statements. The same goes four the fourth sentence. The word 'beautiful' is also meaningless.
As for the sentence "I always lie", simple statement mechanics reveal things to you. Suppose the sentence is a statement. Next suppose it is true. Now, we need to know what it means. Suppose that it is assigned a meaning such that if it is true then it is false. It will now follow that the statement which the sentence denotes is false. Before you can assess the truth value though, the meaning of the sentence must be clear. If the meaning of the sentence isnt clear, then the sentence is meaningless, and hence cannot denote a statement.

X = this statement is false = X is a statement AND |X|=0

Suppose X is a statement.

Suppose |X|=1. Therefore, |X|=0 and |X|=1, therefore 0=1, which is false. Hence not (|X|=1), hence |X|=0. Thus, not (X is a statement) or not (|X|=0). Hence, X isn't a statement or |X|=1. And we know that not (|X|=1), hence X isn't a statment.

All your other sentences are completely analyzable using binary logic. What must happen is this. The sentences which are statements are true or false but not both simultaneously, and the rest of the sentences arent statements.

Consider the sentence:

the barber who shaves every man who doesn't shave himself shaves himself

Suppose the barber exists.

Now, it is asserted that for any man in town, if that man does not shave himself then the barber shaves that man AND the barber shaves himself.

Let B denote the barber, let S denote the binary relation 'shaves' and let X denote an arbitrary man from the town. The sentence is asserting that:

For any X, if not (X-S-X) then (B-S-x)

The previous sentence is a first order logical statement, and the domain of discourse is the set of men in some town. Supppose that B is an element of the set of men in this town, then B is substitutable for X, which leads to the following statement:

if not (B-S-B) then (B-S-B), which can be shown to be tautologically equivalent to
B-S-B

Now, we can eliminate the assumption that the barber is a man in this town, and right the following absolutely true statement:

Either the barber is not a man in this town, or the baber shaves himself. It was also asserted that the barber shaves himself, and this does not contradict any of the reasoning.

In 10 however, we would reach a contradiction by getting the compounnd statement B-S-B and not(B-S-B) from which it follows that the barber is not a man in the town, still no paradox at all.

11 and 12 mimic 9,10

13,14 if statements, are compound statements which assert they are statements, and in that sense they are similar to sentence 8. The error in the Godel theorem is that the sentence isn't a statement, but is processed as a statement. The break in reasoning should occur by concluding that the sentence isn't a statement.

10. Dec 29, 2003

### Canute

It doesn't matter whether Goedel's sentence was a sentence or a statement in this context. He did not suggest that some statements are neither true nor false. His proof was of the fact that any system of truths and falsities (formal axiomatic systems of a certain level of complexity etc.) must contain statments which are undecidable within the system.

As this is true for all such systems then there will always be statements which are undecidable. However this does not mean that they are neither true nor false in reality, just that they cannot be decided without extending the system.

But some people take this to mean that all such systems cannot produce truth, and infer from this that there are statements which are neither true not false in reality , just as non-dual philosophers assert.

As Lao Tsu said " Words that are strictly true seem to be paradoxical" (Tao Teh Ching). This relates directly to Goedel, in that the more true (provable) words become, the more inconsistent the system used to prove them appears. This relates also to something called the 'Quine-Duhem thesis', which states (by implication) that all axiomatically derived truths are only relatively true, and cannot ever be shown to be completely true.

Drawing conclusions from all this is tricky however, and academics argue incessantly about the real meaning of these proofs.

Last edited: Dec 29, 2003
11. Dec 29, 2003

### StarThrower

I've already completely analyzed the Godel sentence, and isolated an error in its analysis, as I now see it, there is nothing to argue about. It certainly does matter whether or not the Godel sentence is a statement, because there is no reason to use binary logic to try and process sentences which don't denote statements, no reason at all.

I can post my old work which isolates Godel's reasoning error if you wish.

12. Dec 29, 2003

### Canute

I think you're missing the significance of Goedel's theorems. Certainly not many people share your view that he was mistaken. Most people, even today, think he should have got a Nobel prize. Perhaps you know better.

The question is only whether Goedel's sentence was a true theorem within the mathematical system. Statements and sentences don't come into it. If you're saying that undecidable theorems are invalid as theorems because they are self-referential then you need to go back to the mathematics. Goedel wasn't a fool, and he dealt with this problem.

13. Dec 29, 2003

### phoenixthoth

the correct term i think you mean is "well formed formula," not statement. x is tall is a statement and a sentence. is T(x) not a well formed formula? ;)

of course if you assume all well formed formulas are going to have precisely one truth value among {T,F}, then it follows that all wffs have a truth value among {T,F}. fuzzy logic could be interpreted as meaning that statements are 1/3 true and 2/3 false, but that interpretation would be coming from the human using it and not the system itself just like how something being true in binary logic may or may not be actually true. all it means is that you have a function with domain of all wffs to {T,F} (binary logic) or {T,M,F} (ternary logic) or [0,1] (fuzzy logic) that satisfies basic properties that make it a generalization of binary logic. that it is a function guarantees an unambiguity in the "truth value" so that no wff can have more than one truth value and it must have at least one. so an extended excluded middle is there.

other logics are free of contradiction though they contradict the assumption that all wffs must have truth value of T or F. however, they don't contradict their own set of assumptions that each wff must have one truth value which is a generalization.

for more info and to look and tell us where the contradictions are, check out the following articles:
http://plato.stanford.edu/entries/logic-manyvalued/
http://plato.stanford.edu/entries/logic-fuzzy/

and if you can publish a paper on where godel was wrong, maybe you can get a fields metal! good luck; it would be wonderful if there were no undecidable statements.

14. Dec 29, 2003

### Bob3141592

I believe that's quite correct.

If I understand Goedel (now that's a fuzzy statement!) he wasn't making a statement about the nature of absolute truth. He was making a statement about the nature of formal systems. Lots of people seem to miss this point. A formal system is one in which there a finite number of explicitly defined symbols, and a finite number of explicitely stated axioms, and where there are a finite number of explicitely stated operations on those symbols. The symbols themselves do not contain "meaning," and it is not necessary nor even desirable to have any interpretation of the symbols. A formal system can be reduced to a set of algorithmic actions on the symbols or the derived statements of the system.

Very little we deal with is really a formal system. I'm not sure that mathematics itself is a formal sstem, since I don't know for a fact that set theory can be completely axiomized.

And reality itself may not be a formal system. We can't make it a formal system unless we can both axiomize all of the principles by which reality operates, which is a dubious proposition, and define without any ambiguity all of the elements on which reality is built.

So while Goedel's Proof is very important logically, it's commonly overextended and used in domains where it realy doesn't apply. Philosophically, it's trecherous to use without abusing it.

At least that's my take on it.

15. Dec 29, 2003

### phoenixthoth

what it sounds like you're saying is that it is not logically impeccable to apply mathematical theorems to philosophy, and with that i would whole-heartedly agree.

16. Dec 29, 2003

### Canute

Very well put imo.

This is where I begin to disagree. I think I see what you're saying. In a sense you're right, there is no such thing as a formal system, since all such systems must be incomplete or inconsistent and therefore not strictly formal. But we treat systems as formal, in other words we act as if we deal with formal systems. We assume, for instance, that proofs can be derived from formal systems despite their lack of complete formality. This is ok on a superficial level, (2 + 2 does truly equal 4), but it goes horribly wrong when we use axiomatic systems to attempt to prove truths about reality, for it can't be done.

As nearly all our reasoning is formally axiomatic, and certainly all our systems of proof, I would say we do deal all the time with what we treat as being formal axiomatic systems, but that, as you say, we are kidding ourselves for no such thing exists, and these systems contain no non-contingent proofs or truths. (I'd be interested to know whether you agree with this point or not)

I'd say that this is the heart of the issue. The non-dual view of reality is that it is built on something (/nothing) of which nothing completely true or false can be asserted. In this sense the totality of reality is seen as being beyond description by any formal system of truths and falsities. By this view Goedel's theorems are true as a natural consequence of the structure of reality, for all formally axiomatic systems used to describe it must leave out the most important part of it, and must therefore not be incomplete or inconsistent, for if fully developed they will be found to be wrong, only partially true.

I agree with that, but feel that the theorems are vitally significant to philosophy, so the dangers have to be faced. I like Roger Penrose's approach, but I know many people think he abused Goedel.

17. Dec 29, 2003

### master_coda

I just feel that I should point out that for something to be a formal system, it does not have been be consistent and complete.

And even if consistency and completeness were a requirement for a system to be formal, that does not imply that there are no formal systems. There certainly are systems that are both consistent and complete.

18. Dec 29, 2003

### Bob3141592

That wasn't quite what I said. No "nontrivial" formal system can be both complete and consistent (nontrivial in this case means powerful enough to do multiplication, but I'll use it in a differenct sense later). That's what Godel proved. The systems are formal. And if they're formal, they're limited by the constraint of Godel's proof. If, however, they're not formal, meaning they don't meet at least one of the essential criteria of a formal system, maybe they can be both complete and consistent. Is reality a formal system? If it isn't, then it can be both complete and consistent. It may not be completely comprehensible to us in that case, but in and of itself it can be complete and consistent.

Note that I'm not saying that it is here. That would be a statement of faith that I'm not prepared to make at this point (but if you ask me later, I'll say that it is, I hope).

I can't agree with that, because I don't think all of our reasoning is formally axiomatic at all. For example, I play chess, and I play moderately well. Most people think of chess as a formal system, that it involves nothing but it's own special form of reasoning. The rules about how the pieces move are finite, and the board and pieces don't contain any hidden meanings or interpretations that are relavant to how the game is played. those are part of what a formal system is. But I don't fully understand the axioms about what's good in chess, and how the rules operate on the board and pieces. I don't know that anyone fully knows or understands them. Despite the progress computers have made, I do not believe chess can be solved with a discrete algorithm. Generally, I play by an instinct I don't fully comprehend. Still, it manages to work enough that I'd play for money if you want.

I might be wrong about a system as simple as chess. But certainly our reasoning as applied to psychology, politics and religion is not formally axiomatic. Not by a long shot. Yet we do reason about them. I don't event think the basic concepts people use to think about these things could be called axiomatic. We think based on concepts we've formed by extrapolations from our own idiosyncratic observations, and we change those concepts throughout our development. It's a rather recursive, even unstable operation. It's the very antithesis of a formal axiomatic system. In my opinion, philosophically, that may be our great strength, not a deficiency.

Now that sounds almost like a religious statement (is it acceptable to include religious notions in the discussions in these folders?). And if anything, it is religious beliefs that some, even many people believe to be a higher reality than the physical world. And about their religious beliefs, people often think of them as the absolute truth. It's everything else they're less confident about.

Myself, I'm rather areligious, since I think no concept of a non-trivial God can be consistent (in other words, makes any sense at all: and in this case, non-trivial means any active God, as opposed to the Deistic notion of a Creator who wills the universe into existence and then separates himself from it--that's what I call a trivial God, like the trivial solution to the harmonic equation). But we don't have to go there.

I take a more poetic view (poetry is another way around the restraints of a formal system, since the symbols it invokes have meaning but aren't unambiguously defined). I say the universe is nothing, just nothing that's unevenly distributed. Some places in the universe contains less nothing than other places, and some contain more nothing, so overall it balances out. And that's the truth, at least in a sense. Right?

I agree, the existence of Godel's theorem is significant even outside of the domain where it applies. It's useful in aguments by simile and by analogy, which are dangerous ways to philosophize themselves. But we have to use something to get a handle on things, and since we can't really define the symbols of our notions of the world to another, we have to do something. So trecherous or not, we forge ahead.

19. Dec 30, 2003

Staff Emeritus
Not quite. What he proved is more like any system that contains second order propositional calculus is either incomplete or inconsistent.

As I keep posting, maybe it will take one of these days, Tarski proved that geometry is complete, and Hilbert and others proved it is consistent. And you can do multiplication in geometry (application of areas, or Eudoxian proportions). But the proofs don't use set theory, so they escape the flaws that lead to incompleteness.

20. Dec 30, 2003

### Canute

Yes, I was trying out an idea about systems to see if I got away with it or not. I didn't.

I was wondering whether a system could be accurately called 'formal' if it was inconsistent or incomplete. It's a odd way of looking at it I suppose. Such as system is formal in intent, but it cannot be proved to be entirely formal in fact. There will always be some doubt, with any system subject to Goedel's limits, as to whether the system is actually formal or not. After all if we know that there has to be a contradiction in it somewhere then we can define it as not being strictly formal.

Does that make any sense?