Tuesday, October 21, 2008

Richard Brown, Chocobos, Word Definition, Kripke

Philosophy, using Final Fantasy as fodder. Because, you know, it's a good series.

The later parts are essential to fully explore this following wall of text. If I'm missing something, do look for it further down.

A chocobo is a set of interactions with us. We can ride them, they appear yellow, they emit noises that we interpret as 'Wark!' and, among other traits, they develop a bizarre deformity when severely inbred; they turn blue and, more bizarrely, get faster. This deformity also disables their hydrophobia, allowing them to traverse lakes and streams. Were we to explore a new place and find an entity with this exact set of interactions, it would not only be possible, but practically necessary to call these entities 'chocobo.' Further, even a subset of these interactions would make it desirable - for instance, the blue deformity would require a concurrent discovery of magic - so perhaps this chocobo would have a different wing-to-leg proportion, or have the wrong number of toes, but in general be very like the pre-existing definition of chocobo. This is supported by the inconsistent representation of the chocobo's property/interactions in the source material. The most important trait is simply the subjection impression of chocobo-ness.

Still, at some point, we have something that merely reminds us of a chocobo. Perhaps it is far too small to ride, has somewhat the wrong proportions, the beak is a muddy ivory instead of brilliant yellow, and behaves aggressively toward humanoids. But this is a continuous process; as elements of the originial chocobo are taken away, it simply reminds us less and less of a chocobo. There is no non-arbitrary cutoff where we can say that anything that reminds us of a chocobo less than this is not a true chocobo, and anything more is, as exemplified in the recognition of the no true scotsman fallacy. This is essentially why the impression of chocobo-ness is most important; seeing as there is no ought from is, we cannot say we ought to give the mantle of chocobo to anything in particular from any fact but the fact that we want to. And I can easily show the problems with trying to do it objectively. If we declare any reminder of a chocobo to be a chocobo, we catch anything with beaks, but even within Final Fantasy a strict definition would work for only one game, and we would have a plethora of different things all called 'chocobo.' The point of logical definition is to detail an intuition, not to break it, and both cases violate this purpose.

At best, I can propose a rule; anything in reality that reminds us of a chocobo more than any other candidate should be called 'chocobo.' I could show the unfortunate consequences of not following this rule in the appropriation of other mythological names, but unfortunately I can't recall any specific details right now. So, I'll make one up. 'Soul' has been discredited as a non-fictional construct,* and as such is applied as a word for psychological make-up. But it could also be applied to a putative energy life-form, which seems even more soul-like to us. However, as 'soul' would have already been co-opted for personality, it could not be now used for these energy beings; calling them 'souls' would be equivalent to calling them 'people;' it is now a part of them, rather than the whole.

*(Note that this is not equivalent to not true, but is merely incredible.)

From something else entirely,
"Similarly, in extending the concept ‘wood’ to an alien form of life, we might find that it resembles what we have already called ‘wood’ in certain ways but not others and a decision will have to be made."
Basically, if it has all the familiar properties of wood without necessarily replicating the processes by which wood obtains those properties, we can simply term it as a special alien kind of wood. Otherwise, again, it simply reminds us strongly of wood, and should perhaps just be known as 'like wood, but from Alpha Centauri,' a bit similar to calling a twinkie 'food.' It's certainly not food in the sense that our ancestors used 'food' - it will not sustain you, basically. In fact, a twinkie is rather closer to a poison in its interactions - trying to live off bread alone is possible for quite a long time, while trying to live off twinkies will make you ill. Alpha Centauri 'wood' may also not be useful as wood for certain contexts, in which case using the moniker 'wood' is only misleading, as is calling a twinkie 'food.' Perhaps 'wood' works fine for building, and appears rather wood-like, but cannot be used to make paper.

"He is widely hailed as perhaps the greatest living philosopher and yet the highest degree he holds is a B.A. in philosophy from Harvard. While he was getting his B. A. from Harvard he was asked to teach a graduate seminar at M.I.T. on the logic of necessity and possibility (called ‘modal logic’) which made finishing his B.A. a rather moot point! He was writing papers solving problems in modal logic at a very tender age.
At any rate, the attack on description theories changed the face of analytic philosophy dramatically."

And we're showing red to the bull here. This is the set of interactions that I happen to label a 'challenge.'

The idea is somewhat fragmented in the original text, so I will paraphrase rather than quote.* Essentially, what you describe may not be what you are referring to, and yet this doesn't prevent you from referring to it. Further, since descriptions

*(Still think the original is worth reading.)

It can work if you really want it to, but it's an awfully torturous way to go about things. A better basic conclusion is that either the label or the description may be wrong.

The first thing to realize is that most people don't have strict logical reasons for labelling things 'chocobo' or 'tiger.' It's just, as I said above, the pure sensation of tigerness that makes people think of tigers, regardless of whether their logical description of tigers is accurate or not.

That is, this isnt' a logical process. It's a physical or phenomenal process, and strictly arational. It happens, it is not thought out.

These filters are Bayesian. A person builds up an abstract of 'tiger' from all instances of 'tiger' they've ever seen. (Also similar to Restricted Boltzmann Network training.) Then their brain simply associates the label 'tiger' with this abstract, and so when something matches the abstract to a high probability, they exclaim 'tiger!' as contrasted to 'moose' or any other possible label. You can see that you have access to the calculated probabiliy as well, as there is a 'perfect or stereotypical example of a tiger,' and less perfect examples, and finally things like housecats which are only vaguely tiger-like, and the sheer sensation of tigerness decreases continuously as you go farther from that perfect specimen. The cutoff between 'tiger' and 'not-tiger' is arbitrary, not strict.

This interpretation can, as far as I know, fully contain the one described in the paper. They explain the same data.

Similarly, there seems to be a tortuous interpretation of language. The point of language is shared symbology. So, I have a thought, and I convert it into as symbol which I can send to you. If we use the same language, this symbol will have the same meaning to you as it does to me, and I'll succeed in communicating or sharing my thought.
"So, take words like ‘tiger’. The meaning of this word cannot simply be some description like ‘yellow stripped carnivorous mammal’ for the same reasons as given against such a theory for the meaning of names."
Those reasons being that your description may be mistaken. However, either the label or the description can be mistaken. The label is decided on the basis of language - that is, there's no point in defining it in such a way that you no longer think of the same thing as I do for the same symbol, as that contravenes the communication purpose of language. That is, these things are best defined socially, as this maximized the number of people who you can effectively communicate with.

This means I've logically reconstructed what people normally do anyway. If you have a mismatch between description and label, then someone will correct you by changing the label of the description to the correct symbol, and letting you know which description is supposed to go with your previous label.

So, when you say Einstein invented the atomic bomb, are you really referring to Einstein? It depends. Did you want the correct description for the label 'Einstein' or did you want to refer to the correct label for the decription 'inventor of atomic bomb?' Which did you intend to communicate? Once again, it all comes down to what the subject wanted to do, because there is no ought from is, except the 'is' that you ought to do what you want to do.

Because I can see how this would be unclear, I'll repeat, in the hopes that I stumble on the right words. So, you can be mistaken that Einstein invented the atomic bomb, or you can be mistaken that the inventor of the atomic bomb was Einstein. If you try to communicate with the symbol 'Einstein,' somebody who knows better will notice the contradiction, and attempt to come to agreement with you on what descriptions go with what symbols.

It's worth noting, I think, that the symbols are arbitrary, they just have to be consistent across subjects.

So, Twin Earth's water-like substance, chemical composition XYZ, which they call water? There are two possibilities.

The simplest is that Twin Earthlings speak a language that sounds like English but 'water' doesn't refer to H20, precisely. It's a slightly different language. Twin Earthlings would call our water 'Other-Earth water' and we would call theirs 'Twin-Earth water.'

There is a problem, however, in that XYZ, while having a slightly different X-ray diffraction pattern than H20, is identical in the description. Twin Earthlings can slake their thirst on our water and vice-versa. Theirs expands when it freezes, the heavy version can be used in nuclear reactors, and so on. In other words, except for any test that tests directly for chemical formula, Twin Earth water is physically identical to our water. (Notice I'm having no trouble distinguishing their water from ours. This referral problem is a non-problem.)

It's not really a different substance.

(Well I suppose it uses magic, since chemical structures determines chemical properties, and so it is not otherwise logically possible for a different chemical to act like water without just being water.)

Putting aside the magic chemical structure, if it quacks like a duck, smells like a duck...it is a duck. Twin Earth water is identical to our water. It is our water, regardless of assumptions otherwise. Its interactions are all identical.

Unless, of course, it isn't identical, but if we have one word and two different objects, then the problem is entirely different. The description of it being identical can be thrown out entirely.

Imagine instead of the property that's different being a magical chemical structure, the different property is that it's magma, they just drink it and call it water anyway. (As long as there is some difference, it doesn't matter how small or big it is.) Suddenly the mystery disappears. Of course it's not water, regardless of what Twin Earthlings call it! Again, they speak something that sounds like English, but their symbol 'water' corresponds to the same concept as our symbol 'magma,' and it is not the same language by the fact that it fails to communicate.

This technique comes from physics. The easiest way to understand what's going on is simply to describe the chain of events, the things that are happening. Ask, what are the objects? (Words, which are symbols) What are their properties? (They symbolize a concept.) What are their interactions? (They are used for communication.) Often, you're done at this point. (Ah. When you have one word symbolizing two concepts, it won't work for communication.)

So either it's identical, and there's no problem, or it's not and there's no mystery. By the law of excluded middle, there is no third option. Kripke and Putnam have tied themselves in knots to no purpose, because they refused to look at the tree of possibilities, because they haven't taken enough physics.
"That is, we can imagine that there are some animals that look very different from the way that tigers around here look. But we could, after looking at their DNA, discover that they really were tigers."
So let's apply my little theory. There's two things going on here; 'tiger' as the label that matches the description 'yellow stripped carnivorous mammal' and 'tiger' that matches the decription 'An organism organized by a particular chain of nucleotides.' Either can be valid, but not both at once. Essentially this is equivocation. To see this, let the first 'tiger' be simply "Ex" and the second 'tiger' "Wye." (Taking math is also helpful for philosophy.)

"That is, we can imagine that there are some animals that look very different from the way that Exs around here look. But we could, after looking at their DNA, discover they really were Wye."

In short, if this statement were made in Lojban, it would have been obviously incoherent. If they are truly identical, there is no problem. If they are not identical, there is no mystery.

That was fun. Let's do it again.

"In support of Kripke’s claim about the meaning of natural kind terms like ‘tiger’ not simply being a description, Putnam asks us to imagine that we found out that the things that we call ‘cats’ were really Martian robots sent to spy on us before the big invasion. If this were true, Putnam argued, ‘cat’ would not pick out a class of mammals but would instead pick out a class of Martian robots but it would be the same class of things that it picked out before we found out that they were Martian robots."

Or the label 'cat' has been mistakenly applied to Martian robots, when we intended for it to describe small furry mammals.

Even in this example, it is clear that after we discover the error, unless our name is Putnam, we're going to say 'Those aren't cats! They're Martian robots!" We would not say, "Martian robots are really cats!" as would be completely valid under this interpretation. We would suddenly learn that what we thought of as 'cats' were really mythological instead of common. Simply put, Putnam is mistaken about how people use language. Of course he's allowed to redefine 'refer' to this new meaning, but that will simply mean his symbol 'refer' will not correspond to most people's symbol, and he will be unable to effectively communicate.

"So ‘being a mammal’, or ‘being carnivorous’, cannot be part of the definition of ‘cat’. The same argument can be given for any proposed description. Given these arguments there seems to be a close affinity between natural kind terms and names with respect to rigid designation."

Oh dear. I didn't realize this, but Kripke's idea does lead to it, doesn't it?

This is preposterous. Kripke would fail immediately if he tried to do math. When you find that your definition doesn't match the thing you were applying your label to - if you find that the equation you're looking for isn't the equation of a line after all - you don't get to change the definition. You have to find a new definition that does match what you're looking at. Just as most people would say, "Cats aren't really cats! They're Martian robots!" not, "Cats aren't actually cute and fuzzy. They're Martian robots!" Actually that doesn't work either, but it's the closest I can come. What would people say if they found out that 'cat' actually refers to Martian robots? "Cat doesn't mean fuzzy thing! It means Martian robot!" Well, that's factually untrue. If I share the symbol 'cat' with you, you think of a fuzzy thing, not a sinister robot planning domination.

'Refer' simply means 'represent.' What does 'cat' represent? Well, it's arbitrary, but since we're trying to use it to communicate, we have to agree on something. Since it's been agreed, the matter is settled.

("All the things we thought were cats are actually Martian robots!" Not...um..."When you say cat, you mean Martian robot!" Yeah, I don't know. Help me out here.)

Isn't this supposed to be analytic philosophy? The one that's analogous to the term in math that's used like, "This equation can be solved analytically."
"There are no actual chocobos. This fictional animal was invented by a designer (Koichi Ishii) and so the word traces back to a creative act of an individual. This means that chocobos are essentially fictional."
That chocobos do not exist as physical cellular organisms - that they're fictional - has no bearing on what we would mean if we saw a physical cellular organism and exclaimed, "That's a chocobo!" Is it really a chocobo? The set of interactions that describe a chocobo may come from fiction, but that doesn't mean that this set of interactions doesn't describe a chocobo.

What people would say if they were precise is, "The properties that describe a chocobo are, right now, only found in fiction, but if we were to breed one, it would be found in reality as well."
"Suppose that we discovered an animal that matched the description of chocobos perfectly. We can even imagine that they come in different colors and that they behave in exactly the way that they are described in Final Fantasy."
So, as I described at the start.
"Kripke argues that this is not evidence that chocobos are real. Fool’s Gold perfectly resembles real gold but isn’t."
It's so tempting to use 'wow' at this point. I will resist, however. First of all this is factually untrue. I have a chunk of Fool's gold on a stand behind me, and it most certainly doesn't resemble gold perfectly. For one, it's crystalline, and real gold is found as amorphous nuggets.

Again, if it perfectly resembled real gold, it would just be real gold and we would not have a separate word for it.

If the cellular chocobo perfectly resembles Final Fantasy chocobos, it just is a chocobo, as defined by a set of interactions with us.

Again, look to what people would say. "So, are you talking about a chocobo in Final Fantasy or one of them real chocobos?" Just as they might refer to, "The horse in Pocahontas" and distinguish it from real horses. Perhaps, "Suprisingly, the horse in Pocahontas behaves exactly like a real horse." So, even though it has a full sense of horse-ness, people still distinguish graphical horses from cellular horses.

Consider the opposite situation, where we have something that extremely closely resembles a chocobo, but because of Kripke and Brown, isn't called a chocobo. Instead, I'll call it "Zed."

Whenever someone familiar with Final Fantasy saw Zed, they would be struck by how chocobo-like it is. Every time. If Kripke came and carefully explained that Zed isn't a chocobo because you can't properly reference a chocobo, they would still go, "But dang, it shore look like a chocobo." They would most likely, if they forgot the name, call it "Kripke's chocobo-thing that's not 'actually' a chocobo."

So let me repair the argument again.

'Imagine someone built a chocobo, so that it had a superficial resemblance to FF chocobos, like Fool's Gold shares a superficial yellowness with real gold. This would not be a chocobo."

Let's ask the crowd;

"This is a chocobo robot, not a real chocobo."

Come to think, given a real chocobo, lets ask them about the FF chocobo.

"This is a work of fiction that depicts a chocobo, not a real chocobo."

Right then.

Brown does have something else going on, though. He asserts that we don't know enough about chocobos to determine if cellular chocobos share the proper essence of chocoboness.
"Fool’s Gold perfectly resembles real gold but isn’t. We know this because we know what is essential for something to be gold (having a certain atomic number)."
"There is no way for us to tell what the internal makeup of chocobos is or whether moogles are reptiles or mammals unless it is explicitly built into the fictional story. So we may be imagining “Fool’s moogles”; things which look like moogles are described but that do not have the same essential trait."
Again, this runs afoul of the fact that labels can also be mistaken, not just descriptions. By 'gold' we happen to mean a particular element, but we could also have meant a metallic yellow thing, in which case there would be two kinds of gold; the valuable kind with chemical symbol Au, and the less valuable kind with is an iron sulfide. The 'essence' is arbitrary.

In this case, we don't mean an organism organized by a particular string of nucleotides. We're talking Exs here, not Wyes. I suspect the essence of Wye is the only essence Brown would accept, which is why he came to the conclusion that chocobos lack an essence.

"The same is true for fictional names. Since ‘Sephiroth’ is grounded in a creative act of Tetsuya Nomura Sephiroth is essentially fictional. But if it turns out that Sephiroth was not based on a purely creative act of Tetsuya Nomura but was, let’s say, the real name of a real mythic person (from before Time began) and that Tetsuya had discovered some ancient documents detailing the adventures of Sephiroth and Cloud Strife then those names would refer to those actual people. So whether or not chocobos are real or Sephiroth exists depend on where the causal chain ends. "

The second part is right but the first part is wrong. Yes, it matters where the causal chain ends.

This is another case where we have two concepts - graphical Sephiroth, cellular Sephiroth - but one symbol; Sephiroth. So, equivocation again.

Sephiroth is still a creative act. Tetsuya intended for it to refer to his fictional, graphical Sephiroth.

Even if he found out that he had coincidentally described, exactly, a cellular Sephiroth, this doesn't turn his act into a non-creative one. It was still created, not copied. It doesn't mean he actually intended to refer to cellular Sephiroth when he made graphical Sephiroth. Technically speaking you can't even say graphical Sephiroth represents real Sephiroth, because this could only have been true after all the work had been done, and Tetsuya finds out about real Sephiroth. I can show this by contradiction, and that's fun, so I will. I'm going to shorten it to graphic S and real S.

  • Assume Graphic S represents real S.
  • Graphic S was created without any reference to real S. That is, the concepts in Tetsuya's mind which he symbolized by Graphic S did not refer in any way to real S. (Go ahead, try to refer to something you haven't conceived of.)
  • Graphic S did not refer to real S at the time of creation.
  • Graphic S represented real S but did not refer to real S.
  • Refer and represent are synonyms.
  • Thus my assumption must be false.

You can only say he correlates to real Sephiroth. Othewise we have some Platonic form of truth where you can refer to things you don't know you're referring to. Like bloargihorble. In some alternate universe which we can never touch, perhaps bloargihorble is some very profound truth which I could use. According to the theory I'm analysing, I'm referring to it, despite the fact that to me, you, and anyone who could ever possibly read this, bloargihorble is meaningless nonsense.

(Later I read that this isn't true, that the way Kripke uses the chain of events prevents bloargihorble. However, the way he thinks of the chain of events simply contradicts his theory, isntead of repairing this issue.)

"Kripke notes that when we use a name we have to intend to use it in the same way as the link in the chain from which we get the name used it. So, take the example of Santa Claus. Though there may be an actual historical person that ‘Santa Claus’ grounds out in it is unlikely that we are referring to that person when we use the name since we no longer intend to be referring to that person. "

In other words Kripke and Brown accept the ideas about intent that I'm using, but apparently haven't realized that this contradicts everything else they're saying.
"Kripke on Fiction"

"Kripke makes several claims about fiction that are interesting but hard to interpret."
I'll say. Again, it seems he could benefit greatly from a physical description of fictional objects.

Sephiroth is a flat pattern of lights that appear on monitors, which cause particular chains of events in the eyes, which cause the owner of those eyes to interpret it as the representation of a person.

So talking about 'Sephiroth' per se is ambigous. We can talk about the representation of Sephiroth - the pattern of lights - or we can talk about what Sephiroth represents - a magical swordsman with very destructive tendencies.

We can even say the thing which is represented exists, because the pattern of concepts which make up Sephiroth exists. However, what we normally mean by 'exists' is more vague, usually referring to physical realization with cells and all that.

And, so now we have a physical description of events. Now, any question about existence, interaction, and referral is pretty easy to answer. When we ask questions about Sephiroth, we can specify whether we mean his graphics or the idea of him, as the two objects have different properties. However, they have very definite properties, and so the truth value of any statement about him is straightforward to determine. Now we can be precise.

The representation of Sephiroth exists as long as something is representing him, and the thing represented exists as long as someone has conceptualized him.

I think it's worth noting that what I'm saying here about language is empirically falsifiable. Language is symbols used for communication, and as such 'what we normally mean' is a testable hypothesis. We can ask people about the properties of the concepts that are symbolized by 'Sephiroth.' They should accord with what I'm saying.
"But not every question must have an answer according to Kripke. Some questions are indeterminate and Kripke thinks this means that some abstract objects are indeterminate."
So lets take this description on a test run.

"If some fictional account leaves some fact indeterminate, then it really is indeterminate. So, for example, if it is indeterminate in the story of Final Fantasy VII whether or not Cloud Strife really is Sephiroth then Cloud Strife does not have clearly defined identity conditions and so is indeterminately identical or distinct from Sephiroth."

"Chocobos have indeterminate identity conditions since all we know about them are accidental features (how they look to us)."
No. Chocobos have vague or non-unique identity, as the representation leaves several things up to our imagination. However, the concept of chocobos clearly have properties which aren't up to our imagination.

Again, apparently Brown means 'chocobos don't have genes, which is my standard for the essence of cellular creatures.' Brown uses Wye, but you don't have to; you can use Ex if you want to. But, like I said before, it makes communication difficult if chocobo symbolizes a string of nucleotides to Brown and he tries to talk to the rest of us, for which chocobo symbolizes the appearance of a giant yellow ostritch.

Similarly, even if we all adopt the Wye interpretation, then we have to start saying things like, "This extremely chocobo-like object's genome is highly unique due to its source in genetic engineering..." The concepts involved don't change, but the labels become more cumbersome.

"We are not always imagining what we think we are. [...] We may think that we are imagining a counter-factual situation where chocobos exist but we are really imagining a situation where there are things that look like chocobos are said to look but aren’t really chocobos (since we are imagining that they are real and not fictional)"

I'd rather say that we do exactly imagine what we think we're imagining, but it doesn't necessarily make any kind of sense.

I'm wrong, though. Imagine, for a moment, a world where A=A or A!=~A is false.

Sorry, you didn't. You can't. Whatever you imagined, it had identity, and thus A=A was true. If it had any kind of identity, that is, properties, it was distinguished from other things and non-contradiction was true. The properties of what you imagined did not correspond to the labels you applied to it.

And finally, this example from a section on an almost equally tortured theory.




1. Heads


2. Heads


3. Tails


4. Tails


Are there four possible outcomes or three? The answer depends on whether you think that state2 is a different state than state 3 and that will depend on whether you think that Kripke is right about rigid designation. On Kripke’s way of doing things we have two distinct objects, the coins Land R, and state 2 is a different state than state 3 because the two (distinct) objects are arranged differently."

Again, just make a physical description of events.

So we're flipping two coins. In the table there are four possible outcomes. The question is whether 2 and 3 are really identical, since they have this identical property; one coin is heads, one coin is tails.

So, first question. Are the coins perfectly identical? If not, then the mystery dissolves immediately, so I'll assume they are.

We're already on the home stretch. To flip a coin requires space. The two coins will be in different spatial coordinates; will have followed different trajectories. So while in the moment they are identical, and can be swapped, if you follow their past or future, they are not identical. You can distinguish them by 'the coin at (o,3) is R and the coin at (6,2) is L.' In the future, we simply refer to the position in the past tense, referring to a specific time. It will always be true that one coin was at (6,2) at time (40), and we put the label 'L' on that coin.

This is one of the issues of using physics to try to do philosophy. Physics tends to answer the question for you, but not in a way that addresses the problem you're trying to raise. The answer does not depend on whether you agree with Kripke, because the objects are not truly identical.

You could conceive of the coins being in the exact same place, but this raises a different issue. If you try to use physics to do philosophy, but don't even respect physics, you'll end up with nonsense, even if it isn't immediately obvious.

Instead, imagine two identical universes with identical coin-flipping apparati. There is no fact of the matter about the relation between these universes; you cannot say 'the one on the left.'

This scenario does depend on whether you agree with Kripke. Since there is no way to tell the universes apart, by assumption, they really are identical. The situation where one flips heads and the other tails is indistinguishable from the situation where the other flips heads and one flips tails. (If it is distinguishable, you should be able to do a substitution similar to 'Ex' and 'Wye.') The designation of each universe is arbitrary. I consider this to be a two-paragraph refutation of Kripke's theory of reference.

No comments: