www.richmond-philosophy.net - © COPYRIGHT 2002-10 RUTC - Designed and maintained by Mat Carmody.

Putnam and Scepticism

Mat Carmody

1. Introduction

How do I know these experiences of mine are of a real world, a world that exists independently of my mind and is as I perceive it to be? How can I be sure that the common-sense view that I am really in the world that I experience is not actually utterly mistaken? I am able to experience the world thanks to my sensory apparatuses and the brain that they send their data to. Now, we are all familiar with the fact that we can be easily fooled. Mirages, Escher drawings and hallucinations make us 'see' what is not really there. Amputees who suffer from phantom limb sensations 'feel' itches and pains in limbs that they no longer have. By artificially stimulating our nerves our by stimulating the brain itself directly, we can induce experiences.

This opens up a devilish possibility. You think you are sitting reading this article. But you are not. At least, not in the way you think you are. Last night, whilst asleep, you were spirited out of your bedroom and taken to the secret laboratory of the crazed scientist, Dr. Borsuk, deep underground in the woods of southern Poland. Your brain was carefully extracted from your skull and placed in a vat of nutrients. The severed nerve endings were grafted onto a battery of wires that are connected to a vast supercomputer. The computer is able to send the same sorts of signals to your brain as your eyes, ears, tongue, nose and skin do. So, it is able to create a virtual world that is indistinguishable in so far as how it feels to you, the subject, from the 'real world' experiences you had had up to now. You are not really reading this article. It is the creation of the supercomputer.

This possibility is of course the 'evil demon' possibility of Descartes or the Matrix of the film of the same name. The sceptical problem it raises is obvious. Imagine saying the following:

(1) I know I have hands.

This would seem a blindingly obvious thing to say: so obvious, indeed, that it would be hard to imagine ever needing to say it to convince anyone. Now, (1) appears incompatible with (2)

(2) I don't know I am not a brain in a vat.

If (1) is true, I do have hands. So, (2) is false: brains in vats don't have hands. In the other direction, if (2) is true, then (1) is false. For if the possibility is open that I am a brain in a vat, then the possibility is open that I don't have hands. If I don't have hands, then of course I can't know that I do.

The sceptic tells me that I can't know I am a brain in a vat. For if I am, Dr. Borsuk's computer will feed me experiences indistinguishable from those I'd get were in not in the vat. Judging 'from the inside' on the basis of my conscious experiences, I could not tell the difference. So, since I can't know I am not a brain in a vat and this is incompatible with (1), I can't know that I have hands. Or, indeed, anything else I'd ordinarily claim to know.[1]

The challenge is to show that we can know that we are not envatted brains and hence show that we can know many, if not all, of the things we ordinarily claim to know. The sceptic says that it can't be done. In a previous article, we looked at how Moore and Wittgenstein responded to scepticism. In this article, we'll look at the strategy of the philosopher Hilary Putnam. In fact, Putnam never set out to dissolve the sceptical threat. His analysis of scepticism emerged ultimately from his revolutionary work in the philosophy of language and we must examine it in some detail. The paper will therefore have two halves. In sections 2-6, we will examine the essentials of philosophy of language. I will follow the history of its development; if the reader wishes to jump the material on natural-kind and social externalism for Putnam's general point about representation (section 6), then click here. In sections 7-10, we shall see how it supposedly helps with scepticism and, of course, some criticisms.

In brief, what Putnam will be arguing is that a meaningful world such as 'brain' or 'vat' can only have a meaning because its users have interacted with the things it has referred to. The word 'brain' has its meaning because there are brains in my environment. Now, if I were in the vat, I would have had no interaction with my brain or the vat in which I am placed. So, I could not talk about or think about my situation. But I can. So, this can't be the case. I can't be a brain in a vat.

2. Internalism and Externalism

What does the English word 'gold' mean? Let us suppose that Alfred has never encountered gold before. He has come across the word in a book. What would you say to him by way of explanation?

Here's part of the definition from a free on-line dictionary:[2]

A soft, yellow, corrosion-resistant element, the most malleable and ductile metal, occurring in veins and alluvial deposits and recovered by mining or by panning or sluicing. A good thermal and electrical conductor, gold is generally alloyed to increase its strength, and it is used as an international monetary standard, in jewellery, for decoration, and as a plated coating on a wide variety of electrical and mechanical components. Atomic number 79; atomic weight 196.967; melting point 1,063.0 oC; boiling point 2,966.0 oC.

The definition makes it clear that we know gold to be an element with certain physical properties. This knowledge is relatively recent and relatively specialised. Ask most people in the street, and they would tell you that gold is a shiny, expensive yellow metal used in jewellery. If we were to go back in time a thousand years, no-one would have been able to give us the scientific definition. Perhaps gold would have been defined as an hard, precious, yellow stuff that comes out of the ground.

Alfred's ancestor Aelfric was alive back in 1008 and this is what he would have said. He was familiar with gold, as he saw examples of it worn by the finest people in the land. Would the word 'gold' in his mouth meant what it means in the scientists' mouths now? It might seem that it could not have meant the same thing. How could Aelfric be said to share our understanding of the word 'gold' when concepts such as atomic number and electrical conductivity were utterly alien to him? In the same way, therefore, when Alf learns what gold is, can he be said to mean the same thing as the scientist, given that these concepts are alien to him too?

According to an internalist theory of meaning, what a word means for a person P is determined by facts about the intrinsic properties of the person P: that is, facts about him independent of his environment. Intuitively, what someone means by his words is determined by what is in his head. What Aelfric means by 'gold' is determined by what he would call 'gold' and what he would say about it. So, since Aelfric and you and I differ in our knowledge about gold, we differ in what we mean by 'gold'.

Putnam now asks us to consider the following thought-experiment. Imagine a planet called Twin-Earth, like Earth in every way, except that there is no gold. Instead, there is a substance that is like gold in all superficial ways: it is yellow, hard, metallic, malleable and so forth. Chemically, it is not gold. It is a substance not found on earth that we'll call 'Z'. As on Earth, Taelfric and his people have been in contact with Z. Some mine it, some work it, some sell it, some wear it, and so on. Imagine this world back in 1008 A.D. its time. There's a man there called Taelfric who is Aelfric's 'twin', in the sense that their histories look very similar.[3] The language Taelfric speaks is Tinglish, or Twin-English. It sounds identical and indeed all the words appear to mean the same thing as in English. There is a word 'gold' that refers to the substance Z. Now, suppose Aelfric goes on a journey and is magically spirited away in his sleep to Twin-Earth. He meets Twin-Earthers and they talk. They sound the same and seem to communicate. This remains so even when they talk about gold. The same would be true in the reverse story where Taelfric is magically brought to Earth.

Aelfric and Taelfric would say exactly the same things were they asked to say what 'gold' is. They would both identify the same range of stuffs. They could be transplanted from one world to another without noticing the difference. So, they mean the same thing by 'gold'. Not according to Putnam. Aelfric means gold - the metal with atomic number 79 - whereas Taelfric means Z.

3. Natural Kinds

Why? The word 'gold' is a natural-kind term. Many philosophers, ancient and modern, believe that the fabric of reality has a certain natural structure. The universe contains protons, neutrons and electrons that combine in a limited number of ways to form a limited number of elements. These elements are further distinguished by their different properties: some are gases, some conduct electricity, some form acids in water, and so on. The distinct natures of these things are part of reality and there for us to discover.

Contrast 'gold' with 'telescope'. Now telescopes are a kind of thing. But they are not a natural kind of thing. We invented them. Contrast also 'gold' with 'things in my front room'. In my front room are many books, a table, some magazines, an empty coffee-cup, plenty of dust and a vast number of nitrogen atoms. Is there any natural connection between these things that unites them into a kind which I name with 'things in my front room'? No. There is nothing natural here at all. They form a kind because I have chosen to group them in a certain way.

Once upon a time, human beings started naming things in the world around them. Some of these terms were natural kind terms. People invented labels for water, humans, lions, mountains, and so forth. They were discerning the natural structure of reality. Now, although they were deeply ignorant of the complex nature of reality, Putnam claims that they meant the same by their natural-kind terms as we do now. When stone-age man asked for a glass of 'water', he meant: H2O. When he was overjoyed to discover a nugget of 'gold', he meant: element with atomic number 79. The meaning of his words is determined by the natures of the things they refer to. They are determined by facts external to the mind of the speaker. The speaker, to repeat, may not have the faintest idea what gold is. No matter: thanks to having interacted with examples of gold, he means gold by 'gold'. The immediate upshot of this is that Aelfric's word 'gold' means gold, whereas Taelfric's word 'gold' means Z.

The central thought is the following:

(NK) Where T is a natural-kind term in a language L that refers to a kind K, then T can only mean K because speakers of L have interacted with samples of K.

So, I mean H2O by 'water' because that's the stuff in my environment I pick out with 'water'. I may not know the science: again, no matter. So, going back to Alfred, even if he doesn't know much about gold then, like Aelfric, his word 'gold' means gold - the stuff with atomic number 79.

4. Externalism and Natural-Kind Terms

We have answered one question and raised another. Why believe (NK)? The answer we will look at here that it is natural given reflections on scientific practice.

Sometime in the (relatively recent) past, it was discovered that gold was an element and that it had atomic number 79. Let us suppose the discoverer of this was a Professor Zloto. Zloto had in fact thought that gold was a natural compound of copper and some other metal. He discovers his error. He says,

(4) Yesterday, I thought gold was a compound. Now I know gold is an element.

Suppose now that we take the view that 'gold' means what the speaker thinks it does. Let us suppose that the yesterday in question was a Monday and the discovery day was a Tuesday. Then, on Monday, 'gold' in Zloto's mouth means (something like) shiny yellowish compound of copper and on Tuesday element with atomic number 79.

If this is so, then (4) can't be true. Remember that (4) is uttered on a Tuesday. On Tuesday, 'gold' means element with atomic number 79. So, what (4) means is (5):

(5) Yesterday, I thought the element with atomic number 79 was a compound. Now I know the element with atomic number 79 is an element.

Now (5) misrepresents Zloto. For, were he asked about his views about gold on the Monday, he would not have said:

(6) The element with atomic number 79 is a shiny yellowish compound of copper.

If the meaning of 'gold' changed from Monday to Tuesday, then Zloto can't talk about his past beliefs about gold because he had no such things. For he didn't believe anything about the element with atomic number 79 in the past.

What Zloto could do is say:

(7) Yesterday, I used the word 'gold' to mean a shiny yellowish compound of copper. Today, I use the word 'gold' to mean atomic number with element 79.

This is not an incoherent thing to say. Consider someone who completely gets the meaning of a word wrong. Young Alfred is learning English and mishears a conversation. He thinks that 'brass' refers to the green stuff that grows on the lawn. Shortly afterwards, he is corrected. It would be odd if Alfred were to report this with (8):

(8) Yesterday, I thought that brass was a plant whereas now I know it is an alloy of copper and zinc.

Alfred didn't think that the stuff brass was a plant! He made a simple linguistic error, not a bizarre scientific one. He would report his mistake with (9):

(9) Yesterday, I thought that 'brass' referred to that stuff that grows on the garden lawn, whereas now I know that that stuff is called 'grass': brass is actually an alloy of copper and zinc.

Now, it's not incoherent for young Alfred to say (9) because his mistake is linguistic. But Zloto's mistake was not linguistic. It was scientific. He was wrong about what gold is, not what 'gold' means. Zloto has been talking about gold - the stuff with atomic number 79 - all his life. What changed from Monday to Tuesday were his beliefs about what gold was. This is just what is captured by the perfectly natural (4).

So, Putnam's point is that when we do science, we learn more about what we've always been talking about. This is why it makes sense to say things like:

(10) Since the earliest time, man has wondered what the stars are. Some perhaps thought the stars were beings in the sky. The Ancient Greeks thought they were holes in the fabric of the night. Today, we know that stars are giant balls of gas.

Were Neolithic man, Aristotle, Galileo and Stephen Hawking able to talk together about astronomy, they would all be talking about the same things: stars. They would not mean different things by their words 'star' or else they could not communicate. They differ not in what they mean but in what they know about stars.

In conclusion, Putnam says that our natural ways of talking about how we learn more show that we are committed to externalism about the meaning of natural-kind terms. We do think that they are words that gain their meaning from the environment in which they are used. So, despite, appearances, we must conclude that Aelfric and Taelfric mean different things by 'water' in their respective mouths.

5. Artifactual Kinds

Not all terms are natural-kind terms. Terms like 'sofa', 'shoe' and 'thermometer' are not. It did not take long for philosophers to question whether externalism was true of these and other sorts of terms too. The philosopher Tyler Burge argued that externalism is true of terms like 'sofa' and 'arthritis'. To find out how, read on. Alternatively, click here to move to the next section in which we shall consider a more general line of Putnam's thought.

Alfred is a member of a community like ours, in which sofas are items of furniture designed to be sat on. Deep in the Papua New Guinean jungle, a sofa is lodged in the branches of a tree. It has fallen out of a plane. The locals happen upon it one day and think it is a sacred object sent by their God, Kanapa. They venerate the object and make copies that are equally venerated. To sit on them would be sacrilegious. They build up a complex set of beliefs about their origin and their powers. They believe, for example, that one can imbue a sword with power by putting it on the sofa overnight when there's a full moon.

Now, by a lucky coincidence, the word 'sofa' is the name in the language of this people for this type of sacred object. Derfla is a member of this people at a time when the original discoverers of the 'sofa' have long since died. Everyone in her society just accepts them as religious objects with a certain (mythological) history and supposed powers. Not Derfla. She has got it into her head that they are just items of furniture. Both she and Alfred would identify the same objects with 'sofa'. Both she and Alfred would voice the same views about the things they identify. So, do they mean the same thing? Burge would say that they don't. This time, it is not the natural environment that matters but the dominant beliefs of the community. Derfla belongs to a body of people in which 'sofa' means sacred object with such-and-such a shape and structure that can be used to heal the sick etc. etc. Her odd beliefs don't 'penetrate' to the meanings of her words.

Think now about your linguistic knowledge. You know the meanings of a lot of words. You know the meanings of about 20,000 words. Now, this does not mean that you can define them, dictionary-wise. It means you can recognise things that fall under them (in certain conditions) and perhaps you can say a little bit about them or about what you understand by them. Consider words like 'telescope', 'portrait', 'cat', 'curry', 'game', 'distress'. They are all complex terms and I would expect no-one to be able to define any of them exhaustively. But nor is it easy to define words like 'shoe' or 'pencil' or 'vat'. Very few, if any words, are easy to define.

Nevertheless, says, Burge, you mean telescope by 'telescope'. There are experts out there who do know plenty about telescopes. You defer to them. When talking about telescopes, you are in effect relying on them to pad out your scanty knowledge of them. When it comes to vats, there may be no vat experts. Here, democracy rules. There are majority beliefs about vats. Few, if any, think vats are alive or must be yellow or can't contain liquids on a Tuesday. Most think they are containers.

So, if Putnam and Burge are right, then what we mean by our words is determined by facts outside our heads: by our environments and by our fellow speakers.

6. What's Essential For Reference

gbdreb bładźę. You are not looking at a misprint. You are looking at a string of symbols that are meaningless. They do not represent anything. Re-arrange them to form 'badger' and 'łabędź' and they do. 'Badger' is the word used to refer to badgers. 'łabędź' has no meaning in English - indeed, some of the letters aren't English letters - but it means something to a Pole. It means swan.

So, nothing represents or fails to represent intrinsically. No marks, sounds or pictures have or lack meaning all by themselves. Something has to make them stand for what they stand for. How? There is no easy or agreed answer to this question. For Putnam, a key element is causation. In order for 'badger' to mean badger, there must have been some interaction between badgers and people using that word. We have seen that 'badger' is a natural kind term and that there is reason enough for this (for Putnam, at least) from scientific practice. The point being made here is more general. Whether a natural kind term or not, there must be causal interaction between words and referents for them to take on a meaning.

This needs a bit of qualification. It is not true of all words, only those that refer. The word 'if' does not refer to anything. Nor does 'however' or 'nevertheless'. Nevertheless, Putnam's general point still stands. There is nothing magical about the symbols in 'nevertheless' that makes them have meaning. Something has to hook them up to the world - if not to a type of thing, perhaps to a type of implication or a type of practice.

Second, we can think of types of things we have not encountered by building new concepts out of old ones. Let 'cugoron' be the name for a mix of copper, gold and iron. Perhaps no-one has ever melted and combined these metals together. It has meaning through its parts having meaning. In the same way, I could think about H2O without ever having encountered it, so long as the concepts represented by H and 2 and O have acquired their meaning in the right way.

What about words such as 'dragon', 'ghost', 'unicorn', 'satyr', and so on? Such words do not refer to anything: there are no dragons or ghosts or unicorns or satyrs. Externalists say that such words are meaningful in one sense but another. They are not meaningful in the way that 'water' and 'gold' are because they don't pick out a kind of thing and hence there can be no clear idea of what such things really are. They are meaningful in the sense that people can wrongly think they mean something in the first sense just mentioned. How? They inherit their meaning from the stories we tell about them. It is the stories about ghosts that define (vaguely) what they are. Note therefore that to have such concepts, we need to have a language up and running whose terms have meaning in the more basic sense of being connected to the world. Our foundational layer of language connects us to the world. With that in place, we can be creative.

We can now return to scepticism. What we need to bear in mind is the thought that for words to have meaning, there must be some type of connection between them and the world. They can't refer 'magically'.

7. I Am Not A Brain In A Vat

Henceforth we shall abbreviate 'brain in a vat' to biv. Either I am a biv or I am not. You can say the same thing. There are two possibilities. If we can only know that the disjunction is true without knowing which disjunct makes it true, the sceptic has won.[4]

If I am not a biv, then I am a real flesh-and-blood person in a real flesh-and-blood world, just as common sense would have it. My word 'brain' refers to brains, those flesh-and-blood things in the heads of animals. My word 'vat' refers to a certain type of container. This is because I am part of a community of language-users some of who have interacted in the right way with brains and vats. So, if I say, 'I am not a biv', I say something true. So far, so good and so obvious.

Let's now consider the more interesting case. Suppose I am a biv and have always been a biv. I think to myself, 'I am not a biv'. Now, what do we mean by 'brain' and 'vat' here? 'Brain' cannot refer to the flesh-and-blood things that are inhabitants of the real world as I have never interacted with such things. They are not part of my environment. Nor are they part of the environment of any of my fellow language-users in the computer. (We may either imagine them to be other brains in the same vat or as programs themselves.) Let us be clear: I am a brain in that sense but I have never seen or felt or touched a real brain as all of my experiences are created for me by a computer. In my computerised-world, I experience brain-like things. The computer has generated images of such things for me. Let us call such a thing a brain*. A brain* is then a bit of computer software responsible for a certain type of experience.[5]

What goes for the brain goes for the vat. Although envatted in a real vat, I have never interacted with real vats. I have only experienced vats*. A vat* is a piece of code. When, in the computer world, I walk into the World of Vats department store, the computer calls up the vat*-code to generate the sequence of electrical impulses fed into my brain - a real brain, remember - that generates experiences.

So, when I say 'I am not a biv' when I am a biv, my words have the following meaning: I am not a brain* in a vat* = I am not a biv*. Now this is true. For I am not a piece of computer code (brain*) somehow 'in' another bit of computer code (vat*). I am a real brain in a real vat.

We can now conclude the argument. If I am not a biv, then, of course, my utterance of 'I am not a biv' is true. But if I am a biv, then my utterance of 'I am not a biv' is true. But if it is not true, then, I am not a biv. The very idea that I could be a biv has been shown to be incoherent.

8. Two Vat Stories

This has the feel of an argument that is too good to be true. I merely need to be able to think that I am not a biv to have a true thought and prove that I am not a biv! As with all such arguments, however, it is never quite so easy.

Let us start with a simple objection. It is that we can imagine a situation in which when I say, 'I am not a brain in a vat', I say something false. In Vat Story #1, I have always been in the vat. We don't even need to suppose that anyone built the vat or the computer: they could have come into existence as the result of a fantastically improbable collision of matter in a previously life-less universe. In Vat Story #2, I was plugged into the computer only last night, previous to which I had lived a happy life amongst real brains and real vats.

Now, in Vat Story #1, Putnam is right, says the objector. I do say something true when I say 'I am not a brain in a vat', regardless of whether I am in the vat or not. So, I prove I am not a brain in a vat. But in Vat Story #2, I do not say something true. For my words 'brain' and 'vat' still refer to brains and vats, not brain*s and vat*s. They still bear the imprint of my previous existence. Suppose all the gold in the universe is destroyed unbeknownst to me. Still, it is gold that I am talking when I use the word 'gold' about because that gold was the stuff around when I learned the word. The same goes for 'brain' and 'vat'.

Now, if I could distinguish whether, if envatted, my story is Vat Story #1 or Vat Story #2, then all would be well. But I can't. My conscious perspective on things, memories, and so forth cannot help me determine whether I am in the vat or not and hence cannot help me determine whether I am and how I got here. So, Putnam's proof can't help us determine where we are.

9. Truth and Meaning

In a sense, the sceptic has done enough to win. Let us therefore alter our focus slightly. Could it be possible that we have always been envatted in the manner of Vat Story #1. Of course, that story seems wildly improbable, so let us strip away the details of the story to bring out the philosophical core. Could the reality I am so familiar with be a complex illusion? Could I be radically wrong about the universe around me because I am, perhaps by dint of my construction, in some sense inescapably envatted?

Let's recall the core of the argument:

(1) If I am not a biv, then 'I am not a biv' means I am not a biv and is true (as I'm a flesh-and-blood person).
(2) If I am a brain in vat, then 'I am not a biv' means I am not a brain* in a vat* and is true (as I am biv not a computer program 'in' another computer program).

You might think that there is an obvious flaw in the argument. Suppose I say, 'I am not a brain in a vat'. I say something true. What do I say? It seems tempting to say that I say that I am not a brain in a vat. Consider (3) and (4):

(3) 'X' is true.
(4) X.

These seem closely connected. You ask me a question and I say, ' 'X' is true'. I could have said the same thing by saying 'X'. Example: you say, 'where are my keys?' I could equally say, ' 'They are on the kitchen table' is true' or 'They are on the kitchen table.'

It is tempting to say that we can connect the two into (5)

(5) 'X' is true if and only if X

Under what conditions is a mentioned sentence 'X' true? When X is the case. Philosophers call this principle the disquotation principle for truth. It is part of the concept of truth that the conditions under which a sentence 'S' are true are give by S. Furthermore, it seems such an obvious principle that an ordinary speaker such as you or I can know a priori that (5) is true.

So, it is true that I am not a brain in a vat. The flaw, says the sceptic, is that it is not clear what this means. I have said something true but I don't know what I've said.

To see this, consider Bernard. Bernard grasps the connection between truth and quotation in (5). He knows that you can move from a (3)-type sentence to a (4)-type sentence. So, he knows and can confidently assert (6)

(6) 'The glyphs on the stela are arranged boustrophedonically' is true iff the glyphs are arranged boustrophedonically.

Bernard has no idea what 'glyph' or 'stela' or 'boustrophedonically' mean. So, even if he knows 'The glyphs on the stela are arranged boustrophedonically' is true, because his professor-friend has told him so and hence even if he knows that the glyphs on the stela are arranged boustrophedonically, he doesn't really know what this amounts to.[6]

So, when I say, 'I am not a brain in a vat' and when I know this means that I am not a brain in a vat, the problem is that I am not sure what this amounts to.

10. A Sceptical Attack...

To strengthen the complaint, we can use Putnam's own line of thought. For many years, people knew that water was the stuff in the lake but they did not know what water was. Indeed, many people had radically false theories of water. The same goes for ever so many natural-kind terms. So, I could be so wrong about what brains are - they could be computer-generated entities.

What if the anti-sceptic says that they know that brains are not computer-generated entities: brain*s. For they know that a brain is not a brain*. I can know this sort of thing even if I can't articulate the difference very well. For example, as Putnam himself pointed out, I can know that there are trees that are elms and trees that are beeches even though I can't tell the difference between the two. I know, in other words that 'elm' and 'beech' mean different things even if I don't know what the differences are. I know that there are differences thanks to the tree-experts that tell me so.

The sceptic asks us to remember Lois Lane. Lois worked in the same office as Clark Kent. She had, by anyone's standards, interacted with him. Lois had also been flying with Superman, which is again an intimate, if unusual form of interaction. Asked if Clark was Superman, Lois would have said no. She had concepts of Clark and Superman that represented them as different even though they pick out the same person. For a real-life example, consider the ancient astronomers who thought the Morning Star and the Evening Star were distinct.[7] So, although I think that brains and brain*s are different, couldn't I be wrong?

No. Lois and the ancient astronomers did not know enough to eliminate their error. But we do. We know what brains are. They are things inside our head. We know what brain*s are. They are computer-generated entities. At the moment, there are no such things. Nevertheless, we have a good idea of what they are like. We are familiar with virtual worlds such as that of Second Life or the 'worlds' of first-person-perspective computer games like Halo 3. We know the difference between a human being and a human-being*, the latter being a bit of data in a computer that is visually rendered as a structure of polygons. So, even though we don't know all there is to know about brains and brain*s, we know enough to establish that they are distinct.

But, replies the sceptic, even if you know that brains aren't brains*s, you still don't know what brains are! Just as people interacted with water without knowing what it was, you have interacted with brains without knowing what they are! This section has raised interesting questions, but has really been a detour. You could still be in the vat!

11 ...And An Anti-Sceptical Reply

Not so fast, says the anti-sceptic. We agree I understand the difference between brains and brain*s and vats and vat*s. I'm not like Bernard from section 9. Indeed, I have to, in order to understand the sceptical story. Now, since this is so, I can say, truly that 'I am not a brain in a vat' is true I am not a brain in a vat. I use 'brain' to talk about brains, not brain*s. It's obvious! That's why we have two different words, 'brain' and 'brain*'. So, I can say:

(i) 'I am not a brain in a vat' says something true.
(ii) 'I am not a brain in a vat' is true just in case I am not a brain in a vat.

And deduce (iii) after all:

(iii) I am not a brain in vat.

The sceptic will not be impressed. Put yourself in the vat. Now, since being envatted is supposed to be indistinguishable from not being envatted, a biv must be able to distinguish a brain from a brain* as well. When a biv talks uses the word 'brain', it refers to what we outside the vat call a biv*. When it talks about a biv*, it refers to what we outside may call a biv**. This is the imaginary entity it creates on the basis of its understanding of its word 'brain', which means biv in our language. So that we don't get lost, imagine a man Alf making a film in which a man Bernard is making a film about a man Charlie. Bernard is like the biv. It thinks it is in the real world despite being in a film. So, from our view, when it imagines a biv, it is like a film character talking about a character in the film.

The sceptic's point is then that since the biv can utter just the same words as us - we who assume (for the sake of argument) are 'outside' the vat - we cannot after all know whether we really are inside our outside of the vat. Both perspectives are indistinguishable from one another.

(Note how careful we have to be here. We cannot say that the biv can say just what we say, if that implies that we utter the same words with the same meaning. For my word 'brain' refers to a real brain whereas its word 'brain' refers to a brain*. We must say that we utter the same words - make exactly the same noises - as one another. Our question is whether we can know what our words mean and hence whether we are inside or outside of the vat.)

The anti-sceptic might then try the following line of thought. We've just painted a three layer picture: a real world, an envatted brain and an envatted brain as imagined by an envatted brain. We have distinguished these things with 'brain', 'brain*' and 'brain**'. And now I can run my argument once again. Since I understand the difference between 'brain', 'brain*' and 'brain**', I know that 'I am not a brain in a vat' is true just in case I am not a brain in a vat. I can rule out the truth-conditions I am not a brain* in a vat* and I am not a brain** in a vat** because, once again, I know the difference between the meanings of 'brain', 'brain*' and 'brain**' and so I am not confused about the conditions under which my words are true.

Perhaps you can see where this goes next. The sceptic says: fine! The biv can have that line of thought too. But then from the outside, we now must distinguish four terms: 'brain', 'brain*', 'brain**' and 'brain***'. The anti-sceptic points out once again that since he understands the differences, he is not confused. Furthermore, no matter how many levels the sceptic introduces, the anti-sceptic says he understands them and hence can always know that he is at the top level.[8] So, the anti-sceptic wins!

12. Climbing Up A Tower of Contrasts

What the sceptic would like to say is that we have no idea whether the top level is reality as we know it. Isn't it conceivable that all of this is just an illusion? The problem is that as soon as I make sense of the idea of being contained, I have painted a top-level picture. I can then use Putnam's proof to say that I am at the top level. So, the sceptic has to say that I can't even imagine the situation I might be in. But if I can't do that, I have no reason to listen to him. Something that can't be thought about at all is not distinct from nothing at all.

This may be too harsh. We can make sense of an envatted brain and hence of the contrast between being outside a vat and inside a vat. Indeed, we can make sense of a many-layered system, just as we started to above. We are outside with real brains in real heads. Dr. Borsuk builds a computer that feeds a brain a virtual world: in this world, the when the brain thinks about brains, it is thinking about brain*s. In the virtual world, the virtual people build a virtual computer. This is a computer* in our language. This computer* is connected to a brain* and a second-level virtual world. The brain* attached to this world thinks about brains and is thinking about what we would call brains**. (Think about Alf, Bernard and Charlie in the layered film again). And so on.

Now, since we can make sense of a series of contrasts, can't we make sufficient sense of the idea that we are in the same position as an envatted brain? Of course, we can't say how we are in the same position. If we conceive of the reality 'outside' our situation in any detail, we'd be outside thanks to Putnam's externalism. There would have to be causal connections between my words and the elements of that outside of reality so that those words had meaning. All we can do is point to the sheer possibility that this could be our position - that we could, in fact, be completely wrong about what reality really is like.[9]

13. Conclusion

We have not resolved the question that opened this article, merely explored it. The view that reality is beyond us in such a way that we could be radically in error as to its nature is known as realism. Anti-realists argue that such a position is incoherent. Reality must be reflected in thought and language. Putnam argues for just such a conclusion: meaningfulness requires some kind of interaction with the environment. Putnam's chief target was not scepticism but realism.

Putnam did not, however, style himself as an anti-realist but as someone offering a different form of realism that he called 'internal realism'. The philosopher Donald Davidson likewise argued that 'proper' realism should not be of the 'metaphysical' kind that allows for massive error. We shall look at his arguments in another article. We shall end our exploration here for the present but see the readings below to help yourself explore further.

Mat Carmody
Richmond-upon-Thames College

Further reading

Putnam's thoughts on bivs are laid out in the first chapter of his Reason, Truth and History (1981. Cambridge: Cambridge University Press). Chapters 2-4 develop his attack on metaphysical realism and his internal realist position. The chapter is also reprinted as chapter 2 of the DeRose and Warfield collection mentioned below.

For introductory articles on Descartes and Scepticism, see James Hill's 'Descartes' Dreaming Argument and why we might be sceptical of it' [PDF], Keith Crome's 'Descartes' Evil Demon' [PDF] and Paul Sperring's 'Descartes' proof of the external world', all in the RJP.

For an introduction to Moore and Wittgenstein's responses to scepticism, see Mat Carmody and Paul Sheehy's 'Scepticism in the 20th century: Moore and Wittgenstein', also in the RJP.

There are many papers on Putnam's proof, few of which are easily accessible to the beginner. A good introduction can be found in the volume Scepticism: A Contemporary Reader by Keith DeRose and Ted Warfield (1998. Oxford: Oxford University Press.) Some important papers are:

  • Brueckner (1986) 'Brains in a Vat'. Journal of Philosophy 83: 148-67.

  • Brueckner (1992) 'Semantic Answers to Scepticism'. Pacific Philosophical Quarterly 73: 200-19. (also in DeRose and Warfield., pp. 43-60)

  • Forbes, G. (1995) 'Realism and Scepticism: Brains in a Vat Revisited.' Journal of Philosophy 92: 205-22. (also in DeRose and Warfield, pp. 61-75)

  • Wright (1992) 'On Putnam's Proof that We are not Brains-in-a-Vat'. Proceedings of the Aristotelian Society 92: 67-94.

For externalism there is a vast literature. The two classic papers are:

  • Putnam, H. (1975) 'The Meaning of 'Meaning'. In his Philosophical Papers, Vol. II : Mind, Language, and Reality. Cambridge: Cambridge University Press.

  • Burge, T. (1979) 'Individualism and the Mental'. In French, Uehling, and Wettstein (eds.) Midwest Studies in Philosophy IV, Minneapolis: University of Minnesota Press, pp.73-121.


[1] Some philosophers argue that we can know that we have hands even if we can't know that we are brains in vats.

One such approach is Contextualism. According to Contextualists, 'know' is a relative term like 'flat'. Relative to the rest of the golf course, the greens are flat but relative to a snooker table, the greens are far from flat. Contextualists argue that our standards for knowing vary in a similar way. Consider our friend Bernard. I claim that Bernard knows he has hands. If our standards for knowing are ordinary and low, then what I say in that context is true: Bernard does know. (Compare: with low standards for being flat, 'the golfing greens are flat' is true.) But if we start thinking about scepticism, we become aware that the standards for knowing are much higher. Indeed, they may be beyond reach. My claim is then not true in that context. In that same context, my claim that Bernard does not know he is a brain in a vat is true. (Compare: with high standards, 'the golfing greens are flat' is false.) So, relative to different standards - different contexts - someone can know that they have hands and not know that they are a brain in a vat.

A second denies epistemic closure. Suppose I know that the lights are on and I know that if the lights are on, Bernard is at home. It seems I can deduce - come to know - that Bernard is at home. Formally, if I know that p and I know that p entails q, then I know that q. This principle seems undeniable as it seems that this sort of reasoning is just a natural way of extending our knowledge. It is, however, this principle that the sceptic employs to deny me knowledge.

Let us flip the formal reasoning around. If I don't know that q and I know that p entails q, then I don't know that p. (If I did know that p, then I could come to know q, contrary to what we're supposing.) The sceptic says that I don't know that I am not a brain in a vat (=I don't know that q). The sceptic then claims that I must know that if I have hands, then I am not a brain in a vat (=I know that if p, then q). So, by the reasoning a moment ago, I don't know that I have hands (=I don't know that p)..

Now, it would clearly be unacceptable just to deny epistemic closure just to stop the sceptic, given how natural it seems. Those who deny it argue for its unacceptability on other grounds. Very roughly, they argue that we naturally don't accept that if we know something, we know all that that implies. Here's a famous example. On a trip to the zoo, you pass the zebra enclosure and see a zebra. You know it is a zebra. Now, you are not a zebra-expert. Were it a cleverly-painted mule, then you would still believe that it is a zebra. Nevertheless, you do know it is a zebra. It is not necessary to know something that one rules out all sorts of odd alternatives that would destroy your claim to know. It is not necessary, that is, in knowing p to know p's consequences q, r, s, t...that state that these alternatives are not actual.

Let's work through the case. You don't know that it is not a cleverly-painted mule (=I don't know that q). Nevertheless, you still know it is a zebra (=I do know that p). Now, if epistemic closure held, you would know that if it is a zebra, then it is not a cleverly painted mule (=I know that if p then q) and we'd have a problem: knowing that p would allow you to know that q. So, epistemic closure must go.

(Note: the famous zebra/mule example is from Fred Dretske's article 'Epistemic Operators', Journal of Philosophy 67, 1007-1023.)  [back] 

[2]  www.thefreedictionary.com/gold (accessed 10/03/08)  [back] 

[3] Imagine reading the biographies of Aelfric and Taelfric and finding them to be the same, word-for-word. In fact, we don't need them to be this similar at all, but it blocks out the following irrelevant consideration: were they to have led different lives, then this would perhaps account for why they mean different things by 'gold'. Indeed, we can imagine 'stronger' variations on the story in which Aelfric and Taelfric are molecular duplicates - physically identical beings and yet still mean different things when they say the word 'gold'. [back] 

[4] A disjunction is an 'or' statement such as 'Jim is in Paris or Jim is in London'. The two sentences that flank the 'or' are called the disjuncts[back] 

[5] We don't in fact need to be precise on what a brain* is. The important point is that it is not a brain. Putnam considers three possibilities (without exploring them): (i) the brain-image or experience generated by the computer; (ii) the 'kind of electronic impulse [from the computer] that normally produces this experience'; or (iii) 'the feature of the [computer] that is supposed to produce the...experience' - i.e. the relevant bit of software. I have assumed (iii) in the text. [back] 

[6] A glyph is a written character or symbol. A stela is 'standing stone slab used in the ancient world primarily as a grave marker but also for dedication, commemoration, and demarcation.' (Encyclopaedia Britannica on-line accessed 12/03/08: www.britannica.com/eb/article-9069561/stela.) Boustrophedonic writing is writing that starts off from left to right then doubles back on the next line to go right to left, and so on. (Or vice versa: from right to left to left to right.) See http://en.wikipedia.org/wiki/Boustrophedon [back] 

[7] In the morning, Venus appears a bright object on the horizon and appears thus again in the evening. Not surprisingly, early astronomers didn't realise they were the same object. Hence, they were given distinct names. The Greeks called the morning star 'Phosphorus' and the evening star 'Hesperus'. Pythagoras is credited with the realisation that they are the same amongst the Greeks. The Babylonians certainly knew that they were one and the same before Pythagoras and may well have been the first to discover this. They called it 'Ishtar' and compiled a list of times of its appearance in the morning and the evening. The text in question is written on what is known as the Venus tablet of Ammisaduqa that dates from the 7th century B.C. (though it is claimed that the observations therein recorded may date back to 1600 B.C. - see Sachs, A. (1974), 'Babylonian Observational Astronomy', Phil. Trans. Royal Soc. Lond. 276 (No. 1257): pp.43-50.). [back] 

[8] The anti-sceptic can say this because he is being asked to build on his basic knowledge. Knowing the difference of meaning between 'brain' and 'brain*', he can understand 'brain**' - it is like a film in a film. So, 'brain***' is like a film in a film in a film, and so on. [back] 

[9] Difficult but rewarding discussions of the material in this section can be found in Wright (1992) and Forbes (1995). [back]