In 2024, I published in this blog an essay called "The Meaning of Meaning". When rereading it the other day I found to my chagrin that it was far less clear than I had thought it was. I was unaware when I wrote it then that there is famous essay by Hillary Putnam also called "The Meaning of 'Meaning'". In this essay, the third chapter of a book by Putnam, he proposed a very influential thought experiment. It is the thought experiment that spawned a million doctoral dissertations and is known as the Twin-Earth Thought Experiment. The purpose of the essay you're perhaps only skimming through now is to present this thought experiment and then to lay out and justify my objections to Putnam's argument. Putnam's thought experiment itself is fascinating and so even if you want to side with Putnam and object to my objections, criticise my criticisms, I think you might find something interesting and engaging in this blogpost.
Imagine another Earth superficially exactly like this one. In particular, this Twin Earth has lakes and rivers, rainstorms, glaciers, clouds of all types, steam, cups of tea, etc. In this world there is a species in all ways exactly like the human race who inhabit our Earth and a language called English which contains the word 'water'. This word 'water' is employed by the English speaking inhabitants of Twin Earth to describe all manner of phenomena which resemble at a macroscopic level what we call water in this world, water in its liquid, solid, and gaseous states. However, in Twin Earth, 'water' is not composed of H2O but rather of XYZ, an entirely different chemical recipe. Putnam defends the view that the 'water' in Twin Earth is not really water at all but something else. The Stanford Encyclopaedia suggests we use the word 'twater' to describe the 'water' on Twin Earth and so I shall adopt this convention for the rest of the essay. Putnam argues that water and twarter are two different natural kinds and so the inhabitants of Twin Earth who have got into the habit of calling twarter 'water' are misusing the word. The word 'water', Putnam insists, regardless of the macroscopic similarities between water and twater, only refers to H2O and never to XYZ.
The thought experiment can be made more interesting. Let us imagine that human civilisation on Earth and Twin Earth have developed absolutely in parallel up until the late eighteenth century at which point the two histories deviate from each other. On Earth, water is discovered to be H2O and on Twin Earth water it is discovered to be XYZ. If we can rewind the clock and consider the thoughts and psychological states of humans in our world and t-humans on Twin Earth in 1750, we would find that the conception of 'water' or 'twater' among both sets of people are identical. Nevertheless, Putnam contends, we would still say that the word 'water' only refers to the water we find in our own world and never to 'twater' and this was true even back in 1750. For Putnam, the proposition "There is a possible world in which water is not H2O" is necessarily false. The notion from Kripke's theory of 'rigid designators' is relevant here and you might find it worthwhile to look up Kripke's theory after you've finished this essay, assuming you haven't heard of it already. Water, according to Putnam, just is H2O; its chemical composition is its identity. Water may not have been known to be H2O back in 1750 but nevertheless it was. Of course, presenting the Twin Earth argument this way immediately makes one think of an objection. Couldn't the English speaking inhabitants of Twin Earth just say, "There is no possible world in which water is not XYZ"? There is a kind of parochialism associated with Putnam's conclusion that when water is discovered to be H2O in this world, we have learned something necessary about water that is true in all possible worlds. I may come back to this objection later in the essay.
What was Putnam's aim when he dropped the Twin Earth thought experiment into the muddy waters of twentieth century philosophy? He was trying to say something about the nature of meaning. For Putnam, a natural kind word like 'water' or 'tiger' or 'beech tree' is associated with both an intension and an extension. The intension is like a concept or notion in the mind. It is something like a psychological state. In his essay, Putnam elaborates on the idea that the intension is a stereotype. For instance, the intension or stereotype associated with the word 'tiger' is 'large striped cat like animal'. The intension of a word, as I understand Putnam's argument, is shared among the linguistic community – important aspects of an intension may sometimes only be known by experts, the people who can, for instance, distinguish between elm trees and beech trees. However a natural kind word also has an extension. The extension is the set of all things in the world that the word is true of. There is thus a profound puzzle involved. How do we explain the relationship between the intension and extension of a word? How make the move from mind to world? Putnam does not clearly answer this riddle but he is adamant that because meaning involves an extension as well as intension, meaning cannot be wholly in the head. Is it in the head at all? Although Putnam elaborates on his notion of stereotypes in his essay, one can sense reading it that he really wants to say that meaning must primarily be found in the world. This is why his theory of meaning has been labelled 'semantic externalism'.
In order to push his idea that meaning is found outside the head, Putnam sought to show that a word can have the same intension for two different people but different extensions for each. This is the point of the Twin Earth thought experiment. Imagine a person called Oliver living in our world before 1750 and a doppelgänger of Oliver living on Twin Earth at the same time. Both Oliver and his counterpart possess the same psychological state when considering the meaning of the word 'water'. The word 'water' appears to have the same intension for both. Oliver can point to a glass of water and say that it bears a kind of sameness relationship to other phenomena that he and I call water – he can also state that, say, when ice melts it becomes liquid water. Oliver's doppelgänger can say something similar about twater. But water in Oliver's word is metaphysically different to the twater found on Twin Earth. Putnam argues that because the extension of the word 'water' in Oliver's world is different to the extension of the word 'water' in the doppelgänger's world, that when using the same word, with the same intension, they are meaning different things by it. In fact Oliver's counterpart is always misusing the word 'water' when he describes XYZ using that word. Meaning must therefore depend on things in our actual world referred to by that word.
The Twin Earth thought experiment is a thought experiment and so I might say something briefly about thought experiments in general. They are a big part of modern philosophy – Derek Parfit used them routinely in his own philosophical corpus. Thought experiments are intended as intuition pumps. They are supposed to help us hone our concepts. But I think we can best see the flaws in Putnam's argument if we take the thought experiment seriously. Let us start with the idea that twater is indistinguishable from water at a macroscopic level but has an entirely different chemical composition. When Putnam decided to say in his essay that on Twin Earth twater is chemically XYZ, he was simply picking three letters at random. Do these three letters represent elements found in our own periodic table or does Twin Earth have an entirely different periodic table? Either way, it seems that the laws of chemistry and physics must be totally different on Twin Earth than our own Earth – this is the only way we can explain how an entirely different chemical structure can exhibit the same human level characteristics that water has in our own world, such as freezing at 0 degrees centigrade and boiling at 100 degrees centigrade (at normal atmospheric pressure). It follows that, if the laws of chemistry and physics are different on Twin Earth than they are with respect to our own world, H2O must exhibit behaviours on Twin Earth that are different to the behaviours it exhibits here in our terrestrial environs. Putnam is saying, in effect, that the sentence "There is a possible world in which water is not H2O" is necessarily false but that the sentence "There is a possible world in which the triple point of H2O is not 0 degrees centigrade" is true. One might ask, then, why the truths we find through chemistry are supposed to be necessary but the truths we find through the study of physics, particularly quantum physics, contingent? Why should chemical truths be true in all possible worlds but physics truths be allowed to vary among possible worlds?
There is a philosophical position known as nomological determinism. Adherents to this position think that the laws of physics are necessary and that everything that follows from them is necessary. Nomological determinists think that the sentence "The triple point of H2O is 0 degrees centigrade" is not just true but necessarily true. If everything occurs as the result of physical laws acting on prior circumstances, an argument can be made that not only are the chemical compositions of physical substances necessary but that everything that occurs is necessary. Not only is water necessarily H2O but the triple point of water is necessarily 0 degrees centigrade. It seems that given some initial set of conditions, which we can take to be those obtaining at the Big Bang, everything that happened subsequently is the necessary entailment of these initial conditions and the laws of physics. In proposing that everything is necessary, I am setting aside quantum indeterminacy and free will. It is a thesis that I have discussed in previous blogposts and so I will not attempt to provide a detailed argument in favour of this position here. All that needs to be said is that if we suppose that water is necessarily H2O, there is a slippery slope which, should we slide down it, will lead us to suspect that everything in the actual world is necessary. It follows from the premise that everything in the actual world is necessary that when we engage in counterfactual reasoning and thinking, we are not imagining possible worlds bur rather impossible worlds. This enables us to say that the sentence, "There is an (im)possible world in which water is not H2O" is true. If all the 'possible' worlds that differ from the actual world are equally impossible, there is no limit on what we can pretend to be true of them. There is nothing to stop us from saying that on Twin Earth water is not H2O because the world described by the Twin Earth thought experiment is not really a truly possible world anyway. There is nothing we cannot say truly about such putatively 'possible' worlds.
If I can use a strange metaphor, I would like to say that the slippery slope is inclined in both directions. In the same way that if we concede that some properties of natural kinds are necessary it follows that everything may be necessary, if we concede that some properties of natural kinds are contingent this could lead us to suppose that everything is contingent. If we slide down the slope in this opposite direction, it follows that just as we can say that the sentence "The triple point of H2O is 0 degrees centigrade" is true only in some possible worlds, we can also say "Water is H2O" is true only in some possible worlds. The issue hiding behind Putnam's thought experiment is that some properties of natural kinds are supposed to be essential and other properties are supposed to be accidental. This distinction goes back to Aristotle who used the terms per se to mean essential and per accidens to mean accidental. Putnam is saying that the sentence "Water is H2O" is describing an essential fact about water whereas the sentence "The triple point of H2O is 0 degrees centigrade" is describing an accidental fact about H2O. It seems that Putnam thinks that the truths revealed by chemistry are necessary and the truths revealed by physics are contingent. But why should this be the case? In the same way that Putnam seems to be displaying a parochialism with respect to who can use the English word 'water' accurately, he is displaying a prejudice in favour of chemistry. In fact, at one point in his discussion, Putnam goes so far as to imply that the only essential thing about a peach is its DNA sequence.
There is a philosophical position, a position strongly at variance with Putnam (and Kripke), holding that when we talk about the essential properties of natural kinds, we are really talking about properties we are very reluctant to give up. Let us now set aside both the argument that everything in the actual world is necessary and the argument that everything in the actual world is contingent and imagine that water and twater have both essential properties and accidental properties, that Earth and Twin Earth have both necessary and contingent facts associated with each. How do we decide which facts are necessary and which are contingent? Because the laws of physics are different in Twin Earth than in our world, it may be impossible for a person to simultaneously hold the view that "Water is H2O" and the view that "The triple point of water is 0 degrees centigrade" when considering water and twater in Twin Earth. It is impossible because on Twin Earth it is not H2O that freezes at 0 degrees centigrade but rather twater, XYZ. Putnam says that, if we have to choose between them, we should accept the first sentence as a necessary truth and discard the second, suppose the second is contingent, but there seems to me no reason why we should not discard the first sentence and keep the second as being necessary. We could say that it is essential to water that it form lakes and rivers, rainstorms, glaciers, clouds of all types, steam, cups of tea, etc. We could also say that it is essential that, at normal atmospheric pressure, water freezes at 0 degrees centigrade and evaporates at 100 degrees centigrade. We could say that it is the macroscopic properties of the substances that both inhabitants of our planet and the inhabitants of Twin Earth call 'water' that are necessary properties; and we could say that the exact chemical compositions of both water and twater are accidental rather than necessary, essential. We could say that both water and twater share the same essential properties and that the differences between them in terms of chemical constitution are contingent, accidental. It is because we are prejudiced in favour of the identity statements made by chemistry that we are reluctant to give up the idea that water is H2O. It is because we are prejudiced in favour of chemistry that we are prepared to say that a substance that behaves exactly like water at a macroscopic level isn't water at all. But there is no reason why we cannot say that there are two different kinds of water, both behaving exactly the same at a macroscopic level but having accidental differences at the level of chemical composition.
I shall turn now to a different objection to Putnam, an objection based on the notion of 'psychological state'.
To restate, Putnam's aim is to show that two people can have in mind the same intension for a word like 'water' but that the word's extension can differ between the two. Both Oliver in our world circa 1750 and his doppelgänger in Twin Earth circa 1750 are in the same psychological state when considering the word 'water', appear to be referring to the same stuff – the stuff that one drinks, swims in, which forms lakes, rivers, and glaciers, falls from the clouds as rain, is transparent (usually), freezes at 0 degrees centigrade and evaporates at 100 degrees centigrade, etc. However, again to restate, Oliver is supposedly referring to H2O and his doppelgänger to XYZ. The objection I want to raise is this. Oliver is made up approximately of 60 percent water, H2O, and his doppelgänger must be made up approximately of 60 percent twater, XYZ. Furthermore the laws of physics and chemistry must be different on Twin Earth than in our planet. If a psychological state supervenes on a physical state in the brain, then because the physical state of Oliver's brain must be different to the physical state of his doppelganger's brain, they must have different psychological states and so, therefore, they must have different intensions in mind when considering the word 'water'. Therefore Putnam's aim, to show that even if two people possess the same intension for a word like 'water' they can still be referring to different things, is not realised by his thought experiment because Oliver's psychological state when using the word 'water' must be different to the psychological state of his doppelgänger.
Putnam has an answer to this objection. Although he does not discuss this in "The Meaning of 'Meaning'", Putnam proposed a theory of mind known as functionalism. He argued that the brain is like hardware and the mind is like software. He argues that two different brains can run identical software programs. I feel that functionalism is not enough to save Putnam's conclusion because it is impossible to access the mind of Oliver's doppelgänger on Twin Earth to see if he is in the same psychological state as Oliver on our earth when using the word 'water', and so it is impossible to know if Oliver and his twin have the same intensions in mind when using the word 'water'. This is not an absolutely water-tight response to Putnam's answer. I may have to think about this some more and my readers may want to consider this line of thought themselves and see if they can find the next step in reasoning about this issue, a step that may either endorse or undermine Putnam's own thinking about this puzzle.
To conclude, we might say at a first pass that either everything in our own world is necessary or that everything is contingent. If we opt for the first prong in the fork, the Twin Earth scenario is impossible and so we can say anything we want about water or twater in that world. If everything is contingent, similarly, the statement "Water is necessarily H2O" is false because there are no necessary truths at all. If water has both necessary and contingent properties, there is no reason why we should say that "Water is H2O" is necessarily true and "Water boils at 100 degrees" is only contingently true. We could equally well say the reverse. Finally, we can question the idea that a person circa 1750 in our world has exactly the same intension for the word 'water' in mind when talking about it as a person circa 1750 in Twin Earth. This essay has not categorically debunked Putnam's argument but I hope that it has provided fodder for the philosophically curious among my readers. If you take anything away from this discussion, it is that for there to be a Twin Earth in which twater, XYZ, has all the same macroscopic properties as water does in this world, all the laws of physics and chemistry obtaining in Twin Earth must be different to those obtaining in our world. This is the aspect of Putnam's thought experiment that he does not say anything about himself. And it may be the fatal hole in his argument.
