I very much appreciated the post over at Incredible Hallq. Perhaps he or she would enjoy this link:
http://www.lrb.co.uk/v26/n20/pri...nt/
fodo01_.html
It's called "Water's Water Everywhere," and Jerry Fodor has a field day with what's wrong with some approaches to analytic philosophy.
I think someone at Blogspot is trying to undermine your efforts at saving the world. When I opened the Comments window, I saw the following advert:
Ann Coulter Weekly - Free
Be among the first to read Ann weekly column by email
- Free
I think someone's trying to woo your audience over to the Dark Side.
Interesting post (particularly the science bent from "Brains and Behavior" which I do not know well), and I'll add to it reasons why the related argument he offers in "The Meaning of 'Meaning'" is so significant. Partly just to get out on the table the other implications of the Twin Earth thought experiment, and mostly just because it's so much fun.
Putnam is pointing out that two seemingly central theses to the theory of Meaning (the nature of meaning) are really inconsistent. These theses are:
1. Meaning determines extension, that is, meaning determines what in the world a word refers to.
2. Meanings are psychological or mental phenomena.
When you take the Twin Earth thought experiment back in time to Archimedes, the inconsistency becomes apparent. If (2) is true, then meaning cannot determine extension, since the phenomenological properties that Archimedes is aware of are the same as the ones that Twin Archimedes is aware of. Conversely, if (1) is true, then the relevant properties needed to distinguish H20 from Water are unavailable to Archimedes, and not in his head at all.
These are awfully important implications! It suggests that the meaning of a term is in part external to speakers! One can therefore use a term, and use it reliably and correctly, without knowing its meaning. If the question, "what is the nature of meaning? how do our words have meaning at all?" is interesting, then surely Putnam's argument is substantive.
So, Jeff, did you talk about this at all when you and Hilary were lunching together at the Harvard Club?
his definition, it turns out, is wrong.
Not really. He means "this substance I observe", but as indicated, hasn't the foggiest as to what H20 means or entails in the slightest. As such, for Archimedes, Earth Water, and Twin Earth water are precisely the same thing. They are identical to his observations. He's right. His word water just doesn't mean what your word water means.
When I say "water", I mean what our friend the physicist entails as water. When I say "live oak", I mean what a biologist/botanist would intend as "live oak", and if my usage does not mesh with said usage, then I concede that I was wrong.
Archimedes was a rather intelligent person. His heirs, physicists, would tell him that Earth Prime Water, and Earth Two water do not have the same structure, and with more careful observations the differences are apparent. He would then concede that they are right. Of course, this means that his definition has changed to accommodate additional information, to which he was not previously privy. Language is not strictly personal. Language is shared. That word may well not mean what you think it meas.
Of course, the existence of inside jokes indicates that language is personal, though it is also shared. That's a marvelous thing if you think about it. It also opens the door for all sorts of fun, and all sorts of problems. Lucky us, eh?
His heirs, physicists, would tell him that Earth Prime Water, and Earth Two water do not have the same structure, and with more careful observations the differences are apparent.
I see what you did there.
i am stunned that you chose today to write this article... mainly because i read this comic (in the URL) today. Coincidence?
I'll show my ignorance by giving my thoughts on this without the benefit of training as a philosopher!!! Stand back people, I'm working without a net.
We don't always mean the same thing when we say "water". "Water", in fact, does NOT "mean" H2O. The definition of "water" is not fixed, but clearly tied to context.
If a scientist is talking with her colleague about an experiment in which 30ml of water were titrated into a beaker during a reaction, she "means" H2O. If that same scientist asks her colleague if he'd like a glass of cool water, she "means" an "aqueous solution composed of H2O and some mix of solubles". It's not even a matter of who is speaking; it's a matter of context. That same scientist on Twin Earth would agree that "water" falls from the sky, fills the oceans, and tastes great after a long hike. She would not agree that "water" has the same properties in a lab.
Note that I also used sc(air) quotes around "mean". That's because the concepts of meaning and definition are flexible. There are multiple definitions for words because they are used in different ways. Water is an ambiguous term that has multiple meanings. H2O is far less ambiguous a term. Scientists, like other specialists, use a particular argot in order to be clear in meaning. It is when they use terms that also exist in other argots, or in the general language that ambiguity can arise.
It is possible, I suppose, that Twin Earth could have even greater differences than the makeup of water. If that were the case, perhaps H2O would not be specific enough. Perhaps Twin Earth's water is dihydrogen oxide. Maybe there's a difference in electron orbits in the universe where TE resides leading to a different angle for the "water" molecules. If that were the case, a whole new set of terms of precision would be needed. That doesn't make the old terms bad, it just means they are not precise enough for some cases.
The fundamental problem as I see it, is in believing that any word or phrase can be guaranteed, with absolute certainty, to be precise enough for all cases and all eternity. I just don't see how that's possible. If you have any doubts, look at legal documents. Knowing that language is imprecise, lawyers are more verbose and particular with language than most anyone else. Words with extremely specific meanings within the law are still often defined in place with yet more specific language. And still? Did the founders mean a militia, or an AK-47 in your bedroom?
begin geek interlude
The differences in chemical composition of water between Earth Prime, Twin Earth, Earth-2, etc. is what makes the multiverse possible. It was in trying to end the semantic dissonance that was giving him a migraine that caused Superboy Prime to go all evil. Sorry. I really tried to resist.
end geek interlude
+50 points are awarded to R. Porter
Steve, you do realize I've wasted an entire day of work solely as a result of reading this post, don't you? The worst part is, I still have nothing intelligent to say about it. But I do agree w/ C Ewing, 50 pts to R. Porter.
There are many things in R. Porter's comment that deserve discussion all their own, but for now I only want to add one brief thing...
For the most part, philosophers like Putnam (and certainly Putnam himself) are quite open to contextualism in language. Philosophers do disagree about the scope and depth of contextual determination of meaning, but it is an issue at hand.
The question, however, is differentiating between semantic content and pragmatic content. In some cases where we might colloquially say that two words mean different things in different contexts, it really is that the two words are being used to do two different things. This is a question of what one does with a language.
There is a second question, and that is the semantic question - what do the words/sentences mean? Here, contextualism takes a different line, because then the differences are actually differences in meaning. When you have differences in meaning, then you really have two different words with the same phonetic or symbolic representation.
So talk about the meaning of sentences or words does not preclude a deep awareness of the prevalence of contextualism in language.
Turning back to Putnam in particular, I don't think R. Porter's comments are stand as an objection to Putnam's claim that semantic meaning is partly external. Let me first make a clarification, and then show why Putnam survives the challenge. Putnam does not think that water means "H2O." Part of the meaning of water is that it refers to that stuff, identifiable by its chemical composition "H2O." It's the stuff and not the composition that Putnam is talking about.
He could easily appeal to the semantics / pragmatics distinction here. In both of the cases, where the scientist is doing an experiment and when she is just asking for some refreshment, she uses the word water. In both cases, that word has a meanining comprised of a "stereotype" (Putnam's term) which is a number of non-necessary phenomenological properties and water, the stuff the term refers to. In both cases, the scientist is referring to the same thing, water. What difference is happening here? Well, in one case she has different intentions than the other, and a different battery of beliefs and desires. She is also trying to do two different things. These are all pragmatic questions. This leaves the semantics of the word "water" still standing, meaning that the different uses of terms in different contexts does not rule out the notion that words have a stable and statable meaning.
The XYZ example has always bugged me. To hypothesize that "It turns out that XYZ has many superficial observable properties in common with H2O, it just has a radically different chemical make up" is contradictory. If the observable properties it shares with H2O include chemical reactions, then saying twin-Earth water is made of XYZ is a distinction without a difference. (Unless we're talking about different isotopes--D2O or something. But hey, we call that water.) If XYZ does not chemically react in all the same ways as H2O, then it could hardly be a functional substitute for something as ubiquitous (and important for life as we know it) as H20. If we start a thought experiment with a contradiction, why should we think our intuitions about the results are informative?
And...why not say Aristotle's definition is correct and we have now discovered two kind of water that now each need a more specific name? Jade (the rock) provides a real world example of this. Both jadeite and nephrite are called jade, and the fact that they were distinct minerals was not discovered until the 19th century. On that model, XYZ would be water.
"If we start a thought experiment with a contradiction, why should we think our intuitions about the results are informative?"
Putnam writes:
"I shall suppose that XYZ is indistinguishable from water at normal temperatures and pressures." (The Meaning of 'Meaning')
I'm not sure I see the contradiction. He sets it up so that they are indistinguishable in a certain context, one in which they do not have access to the experiments which yield different results when applied to XYZ. He is not assuming that no such tests exist. That's in fact the reason why he considers the case in a 1750, much like Steve used the example of Archimedes.
Ironically, Putnam also uses the jade example, to reach a somewhat similar conclusion. He writes "if H2O and XYZ had both been plentiful on Earth, then we would have had a case similar to the jadeite/nephrite case: it would have been correct to say there were two kinds of 'water'" (241). The point, however, is that we still rely on the extension to determine the meaning of the term, it's just with the term "jade" we place particular emphasis on the superficial characteristics. This supports his claim that what determines meaning is in part external.
And still? Did the founders mean a militia, or an AK-47 in your bedroom?
This is a bad analogy on multiple accounts. First and foremost, H20 or any "substance" (if I'm allowed to use such an archaic term) is far more static than a legal construction. Also, language evolves. That's a difficulty in any language question over time. Also, the U.S. Constitution is intended to be both flexible, and able to be adjusted/altered. This includes reinterpretation. Nobody knew there would be the internet back in the days of the Whigs and Tories, but we have to make adjustments. That's life, and the U.S. Constitution was created in the hope that we would be able to make it work. That's not even including the evolution and changes of our culture/society.
This was a bit of an aside, but I thought it was worth mentioning. That, and I can't sleep after Dunkel. I choose to blame Hanno.
But I do agree w/ C Ewing, 50 pts to R. Porter.
I'm always unerringly right about the most inane and inconsequential of things. It's a gift and a curse.
Can I use my 50 points to pay for my subscription to the Playground for the next year?
Jeff, I had forgotten exactly how Putnam put it. You're right--it is not technically a contradiction. But given what we know about chemistry and physics (leaving aside inductive skepticism), it is extremely implausible that a substance would be indistinguishable from water at normal temperatures and pressures but divergent at the extremes. It's kind of like saying "suppose there's a particle with the same mass, charge and spin as an electron except when they are in the interior or a star."
At any rate, such a discovery would entail so many major revisions to accepted scientific theories that it seems naive to only ask about whether XYZ is water or not.
I'm sympathetic to the view that what determines meaning is in part external; I just don't think the twin earth example is particularly enlightening. If we think it's non-trivial to ask "Is Spaceman Spiff wrong when he refers to XYZ as 'water'?", then we are already assuming at least some level of conventionalism. But the assumption that communication is _possible_ gets us that far.
Don't have time for a full reply at the moment, but in the interim, are you sure we're reading the same Putnam articles? Most of "Brains and Behavior" is on an entirely different subject, with the main thought experiment being Super-Spartans, while the part I was referring to was about the discovery of the cause of polio. I'm not sure water even comes up in "Brains," sorry if that through you.
it is extremely implausible that a substance would be indistinguishable from water at normal temperatures and pressures but divergent at the extremes.
Its not about plausibility. Its about the meaning of words. The mental state in the one case is the same as the mental state in the other. If meaning is tied to mental state, and not to the natural kind, then the words would have the same meaning. But it turns out, the fact that the stuff is not H2O on twin earth means its not water, even though the mental state is the same. And that means that meaning is not limited, in natural kind terms, to mental states.
Personally, having read Kripke first, all this did not strike me as surprising or controversial.
Hanno, I understand that plausibility is not supposed to be an issue. But the implausibility leads me to question the intuition that "the fact that the stuff is not H2O on twin earth means its not water". It seems equally likely that we would say "we've found a new kind of water, with such and such behavior at extreme conditions." It seems a lot like the discovery that there can be different isotopes of elements. Normal hydrogen contains a proton and an electron. Deuterium contains a proton, a neutron, and an electron, and it is still classified as (an isotope of) hydrogen. And water composed of 2 molecules of deuterium and one of oxygen is still water, albeit sometimes called heavy water. Following that example (or the jade example), I think we'd be equally likely to revise the extension of the word 'water'.
Of course, extension matters either way your intuition goes. But it's the fact that we think it's non-trivial to ask, "Is XYZ water?" that tells us that extension matters. Assuming that the question can be debated also assumes that there is at least an inter-subjective component to meaning, and thus meaning can't all be in the head. (I'd have to go back and double check, but I think this means I agree with Putnam's conclusion here, but I question this particular argument for it.)
Your "Twin Earth" is absurd. Nothing is like water that isn't H20. That's why D2O is H2O. Nor could "Twin Earth" look at all like Earth, if it flowed with XYZ instead of H2O. Archimedes would have to be in a space suit, and he wouldn't dream of drinking XYZ, no matter how like water it appeared to flow and float a boat.
Also "definition" is funny word. I think definitions are theoretical and after-the-fact of word use, and more or less peripheral to it. I would say I have an idea or ideas of water, and these unconsciously guide my use of "water." Definitions I think are for defining, and "defining" typically is more description or constraint than it is prescription or formulation.
It seems equally likely that we would say "we've found a new kind of water, with such and such behavior at extreme conditions."
We could, but we don't.
If we take another kind term, like 'jade,' we discover that there are actually two types that outwardly look the same and act the same, jade A and Jade B. But when you ask people in another culture which is the 'real jade' so that we can know which molecular structure counts, you get a puzzled look. And when they here that there are two types, you get a shoulder shrug. So 1) it is perfectly possible to have a term not apply to molecular structure, but to the collection of observables but also 2) that is not English and water. For us, it is water iff it is H20 at the appropriate temps and pressure.
We could go 'we found a new water,' but we don't. And we don't because we think there is an essential structure to water and that we found it. Both that there are essential structures and that we discovered them are philosophically significant.
I've never understood how philosophers get away with getting credit for interesting observations about implausible hypotheticals. Seems like cheating to me.
I've never understood how philosophers get away with getting credit for interesting observations about implausible hypotheticals. Seems like cheating to me.
Shhhhhhh. C'mon, man, what are you trying to do, make me go out and get a real job?
I've never understood how philosophers get away with getting credit for interesting observations about implausible hypotheticals. Seems like cheating to me.
As stated, the plausibility is not the point nor is the hypothetical nature of the question.
There may be no point, if the hypothetical is absurd, and if there is one, it may becomes up for grabs. Of course, bets are off when it comes to grading. Beware the implicitly postulated logical impossibility of outsmarting the exam writer.
MT,
Not sure if that's the case at all. In this situation the idea was to formulate a situation in which we could see two people with the same brain state, but meaning different things. It is an important and longstanding position in philosophy of language that meaning is in the head. But what Putnam does here is set up a situation we can all understand in which Archimedes and Twin Archimedes have identical brain states (indeed they have identical brains) and yet they could reasonably be said to mean different things by the same simple statement "This is water."
I guess I'm just not sure what the problem is that you, SteveD, and Michael Schmidt are worried about here. O.k., there is no XYZ and there likely can't be one, but that seems to miss the point here because absolutely nothing hinges on the chemical aspects of example. We just need two different situations in which two individuals approach with the same brain state. If that is possible, then Putnam can argue that meaning ain't in the head.
The hypothetical may not be the case, and there may even be good reason to say it can't be the case, but it is merely an illustration of a possible case of same brain state, different worlds, therefore different meanings, and thus meaning must be part of the world. Not sure what the problem with the example is that would undercut the argument.
There may be other grounds on which to object to Putnam, but I must admit I don't really get the problem you guys are pointing out.
Then I would argue that Putnam did a lousy job of communicating the essentials of the argument by choosing a needlessly fanciful illustration. Illustrations should facilitate understanding, not distract from it.
Why not use the example of jade, cited above? I seem to remember some other scholar back in the nineties making similar arguments about "lilies," in which florists may consider things lilies in the florist context which botanists cannot admit to their lily category in the botanical context. Is this the same sort of thing that Putnam is talking about? Or am I missing some subtlety here?
SteveG:
As I pointed out, using Archimedes does seem to be a little different. With no knowledge of atomic composition, Archimedes cannot intend H20, but only that observable substance with particular qualities. Thus, Archimedes does not, and cannot intend H20 when he says water initially. However, when he realizes that they are initially identical in behavior, but have different compositions, he may well say, "My bad, dude", or whatever his equivalent would be. And indeed, he should say something along those lines.
This point seems significant, and does not seem to counter, but rather seems to support Putnam. I will correct my language usage when I realize I am using a word inappropriately. After all, I can't communicate otherwise. And what the hell point is there in language usage if not to communicate? Thus, it ain't all in Archimedes' head, but nor can he mean (read: intend) H20 without any comprehension or even knowing of the existence of said components. When Archimedes is brought into question, it seems to complicate things a bit.
In much the same way, Shorty cannot mean F.U., when she has yet to come to knowledge of such a phrasing (or so I shall assume). And thus does not, and cannot intend such with any given gesture.
C. Ewing,
I think you are exactly right about the importance of using Archimedes. In "The Meaning of 'Meaning'" Putnam first asks us to consider the thought experiment in a modern day Twin Earth, but then asks us to think about it if both Earth and Twin Earth were in the year 1750. Since, like the Archimedes case, there is no way for the agent in question to know about the chemical composition of water, that which is in his head cannot determine the difference between H2O and XYZ. It is vital to his argument that he can do this, precisely because it allows him to fix the same mental content in both Archimedes on Earth and Twin Archimedes.
Archimedes can intend an unidentified atomic structure, even if he does not know modern atomic theory. He does indeed have the idea that there is an essence to gold, to water, and he thinks that there are particles too small to see in virtue of which gold is gold and water is water.
Just as I can name Pi without knowing what Pi is, why cant I name H2) without knowing what H20 is? Hell, I can even eat the Pi when I am done naming it.
Michael Schmidt wrote...
"Why not use the example of jade, cited above? I seem to remember some other scholar back in the nineties making similar arguments about "lilies," in which florists may consider things lilies in the florist context which botanists cannot admit to their lily category in the botanical context. Is this the same sort of thing that Putnam is talking about? Or am I missing some subtlety here?"
The reason the Twin Earth thought experiment is more efficacious than the jade case is that the linguistic intuitions (data for linguists and philosophers of language) cannot easily be explained by appeal to other facts. In the Twin Earth case, Putnam is able to fix sameness of mental content, allowing him to test whether or not the meaning-determines-extension claim is consistent. In a way, Putnam is setting up the problem so that we only have to solve for one variable.
As to your lilly example, it is hard to tell if it is the same phenomena without knowing what exactly is going in those classifications. I suspect, however, that what is happening is that the two different scientific fields have two different classification schemes based on apparent properties. If the distinction is made based on these properties, then the point is a different one than Putnam's because that is a case where people have different sets of associated properties assigned to the same reference.
Putnam is trying to show that two theses are incompatible:
1. Meanings are psychological.
2. Meanings determine extension, or meaning determines reference.
If two people share (1) but are talking about different things, then (1) and (2) are incompatible. So, since Archimedes and Twin Archimedes have the same mental state about what they call "water" and they refer to different things, it cannot be the case that their mental state determines the reference of the term - otherwise the terms would have to co-refer.
This shows that meaning is in some part external, "water" means water in virtue of the fact that it refers to that stuff, the stuff with the chemical composition H2O. So, in the lily case, florists and botanists might be in different psychological states when thinking about lilies, so, if I am interpreting the case rightly, it is a slightly different problem than the one Twin Earth is designed to address.
He does indeed have the idea that there is an essence to gold, to water, and he thinks that there are particles too small to see in virtue of which gold is gold and water is water.
True. Atom is a Greek idea, after all. But he does not mean H20, because he has no concept of hydrogen or oxygen. Can you show me where Archimedes spoke of nuclear forces?
But he can mean the essence of the thing. The "essential nature of water", for instance, but one cannot intend for another to understand a concept they themselves do not comprehend and cannot fathom. Ants do not and cannot write the name, "Johannes Bulhof", regardless of whether their trail in the sand spells the name.
He may well mean something like H20, since H20 is as close as we get to an "essence" of water (provided you do not mean it in some silly, esoteric sense), but he does not and cannot intend H20, because he doesn't yet know what hydrogen and oxygen are.
If he means the essence of the thing, the case is closed. He does not have to know the essence to refer to it or name it.
Hell, I can even eat the Pi when I am done naming it.
Hanno lists one of the many reasons Pi is superior to cake. You can't name your cake and eat it too.
I accept that extension (reference) matters and that meanings aren't just in the head. (In fact, I think the existence of linguistic communities at all suffices to show it.)
I don't accept that the claim that XYZ can't be (English, non-twin) water because being H2O is an (or the) essential property of water. If we were confronted with XYZ, I contend that we might revise our notion of the extension of 'water' to include XYZ. But either way, the decision would be based on observations (both casual and scientific) of XYZ and not on anyone's linguistic intuitions.
So either way, extension/reference matters. My contention is that by accepting the question "Is XYZ water?" as coherent or meaningful, we also accept that extension must be part of the story. Different answers may highlight some further intuitions about reference, but allowing the question at all is enough to get us out of our heads when it comes to meaning.
If he means the essence of the thing, the case is closed. He does not have to know the essence to refer to it or name it.
Then his statement can be replaced with, "that thing I know not what", which is not interchangeable with H2O, and now the case it closed.
I guess the problem the hypothetical poses is for my rebuttal, because I feel forced to address a seemingly unreal world or to work within a framing of the issue that feels crooked. That feels unfair. We should play a game we all like.
Anyway, water's "definition" is neither outside nor inside the mind. "Water" is a mental pointer to a real stuff. A laundry list of ideas about water can accumulate with the advance of science, but "water" never changes, because it's still pointing to the same place. That's the real "definition" of "water."
"So, in the lily case, florists and botanists might be in different psychological states when thinking about lilies, so, if I am interpreting the case rightly, it is a slightly different problem than the one Twin Earth is designed to address."
OK, I understand better now. Lilies is not a good example.
"Water" is a mental pointer to a real stuff.
Define real. Define stuff. You've just opened up an entirely new discussion, one which we haven't the space nor time nor inclination to tackle on a blog.
At least, I'm assuming such. Maybe someone else has the time or inclination, but I'm not sure if SteveG would like to be host to such.
"Define real. Define stuff."
Oh, come on--I'm talking about water, not something even slightly controversial like Big Foot or the Loch Ness monster.
'But what Putnam does here is set up a situation we can all understand in which Archimedes and Twin Archimedes have identical brain states (indeed they have identical brains) and yet they could reasonably be said to mean different things by the same simple statement "This is water."'
If I look at you and point and say "Look, that's Steve Gimbel!" you are changing from the beginning to the end of my utterance--loosing atoms of carbon, gaining atoms of O2, entertaining new thoughts, losing synaptic strength here, gaining it there, cells circulating, etc. So I could reasonably be said to mean different things by the simple statement "Look, that's Steve Gimbel" from start to finish. My "definition" of Steve Gimbel in this hypothetical, one might reasonably say, would include, among other ideas, "that's Steve Gimbel" (i.e. the body I'm pointing at), although this is not part of my definition before I see Steve or after I have forgotten seeing Steve (not to suggest that meetings with Steve are ever less than memorable in real life--this is a hypothetical). Names for things refer to things, but things are ever changing physically, and so are our brains and minds. It just seems obvious to me that our "definitions" for names and other "designatory" words like "water" are in effect operational and something like pointing.
Oh, come on--I'm talking about water, not something even slightly controversial like Big Foot or the Loch Ness monster.
The question is whether or not we have knowledge of content, which gets into Hanno and the up-and-coming TR's field(s), but I have enough respect for it to leave it the hell alone.
It just seems obvious to me that our "definitions" for names and other "designatory" words like "water" are in effect operational and something like pointing.
It's certainly operational. It's for the purpose of understanding and communicating. However, pointing is indicative while definition is, well, definitive. Pointing at a car does not define "car". It merely gives an example of a given car. That distinction seems fairly important. Pointing at SteveG is not a good definition of SteveG, because SteveG is (it would seem) a more complex being than can be summed up in totality with a cursory view. Indeed, you can know someone quite well and never fully come to a conclusive, and all-inclusive definition of said person.
The pointing is an (essential) extra dimension to the definition, as I'm talking about it, that defines the operation to which the laundry list of qualities we associate with water or SteveG accrue. I never meant to suggest pointing was sufficient--except in so far as "definitions" of the dictionary kind seem deficient with regard to how we actually communicate.
Commenting by HaloScan
