This work is protected under international Copyright law. This work (or portions of this work) may be downloaded, distributed, quoted, or referred to for the purpose of criticism, evaluation, or for furthering knowledge and research in this field. When referring to the work of others, please remember to cite the author's name, the title of the work, the place and date of publication (or internet address).

Abstract
1. Introduction
2. Memes in the Minds of Animals
3. Is this stretching the Meme too far?
3.1 What is a meme?
3.2 Why do British Birds not have a milk-bottle opening meme?
3.3 Does this mean that animals don't have memes?
4. What have I shown?

Milk-bottle opening behavior in a species of bird known as the British tit has been put forward as an example of a meme in a non-human animal. The existence of this type of case has lead some thinkers to believe that non-human minds can acquire memes. I believe that the British tit's behavior has been misinterpreted as memetic. In this ePaper, I argue that milk-bottle opening in the British tit can be explained by appealing to its innate behavioral repertoire. I then suggest that the question of memes in animal minds should be considered on a case by case basis.

In their paper "Do Animals Have Memes?" (1999) Simon Reader and Kevin Laland suggest that animal minds are suitable habitats for memes. Their ideas come about after a critical examination of points that Susan Blackmore raises in her book The Meme Machine (1999). Blackmore argues that humans are the only animals with minds capable of supporting memes. Reader and Laland disagree and attempt to find some examples of memetic transfer in non-human animals.

In this ePaper, I will examine Reader and Laland's claims. I will focus mainly on their suggestion that the British tit's milk-bottle opening behavior is a meme. My objective will be to show that Reader and Laland's claim is wrong, and that the British tit's behavior is not a meme. I will start by outlining Reader and Laland's argument and will show that their claims rely on an incomplete definition of what a meme is. I will then offer a firm definition for the meme, after which I will provide an alternative explanation for the milk-bottle opening behavior exhibited by the British tit. My next point will be that the question of memes in animals should be answered on a case by case basis. My conclusion will be that examples such as milk-bottle opening in the British tit do not provide evidence of memes in animals.

According to Susan Blackmore (1999), imitation is the primary mechanism by which a meme replicates and finds its way into a new mind. If she is right, it would seem that memes are exclusive to animals that have the capacity to imitate behavior. So far this ability seems to be found only in humans and some species of birds, members of which can mimic certain sounds and calls.

Reader and Laland agree that there is limited evidence of imitation in non-human animals, but argue that imitation should not be the defining feature of a meme (Reader & Laland 1999). They suggest that limiting memetic transfer to acts of imitation is restrictive and cuts out a significant range of behavioral transmission. For Reader and Laland, the psychological process underlying the transmission of information is not a determining feature of memetic replication. Instead, they argue that transmission fidelity is the key feature of memetic replication. This is to say that regardless of the means by which a behavior is replicated, a memetic transfer can be said to have taken place if the resulting behavior is a high quality copy of the original behavior. Reader and Laland are right to make this point. If acts of imitation were the only way that memes could spread, then it would seem senseless to speak of memes entering human minds from written material or from musical notation. In these cases information that produces behavior is transferred with no observation of the original behavior -- it is not mimicked. Music can, of course, be imitated, but most of the time musical memes are encoded in a musical score and are assimilated to new minds when the score is read.

To support their claims, Reader and Laland point to evidence of behavioral transmission in non-humans animals that does not come about through imitation. The examples used are cases of social learning -- learning that occurs when behavior is influenced by observation of other animals (Reader & Laland 1999). Perhaps the most significant example for Reader and Laland is milk-bottle opening in British birds. In some areas of Britain, a species of bird known as the British tit has acquired the behavior of opening milk bottles to get cream. The tit finds milk bottles at house doors and pecks away at the foil bottle top in order to get access to the cream that sits on top of the milk. When other tits observe this behavior, they try it out themselves and soon learn that if they open milk bottles, they are rewarded. It is important to note that the tits are not imitating each other. They are simply being attracted to objects that other tits are pecking at. According to Reader and Laland, the spread of this behavior throughout the tit population in Britain shows that memetic transfer has taken place. Opening milk bottles is not innate to the tit -- it is learned -- and the behavior replicates with high fidelity showing that it is a meme. Since the British tit can have memes, Reader and Laland conclude that animals have memes.

To effectively assess Reader and Laland's claims we need to know exactly what a meme is. Without a firm definition, it is possible for theorists to make any claims about memetic transmission. Reader and Laland want to relax the criteria that past thinkers have placed on the meme, but I think they are relaxing the criteria too far. If we travel down the road that Reader and Laland are opening up, we will end up describing all conditioned behavior as memetic. On the surface, this might seem acceptable, but it is only acceptable if we have a very loose definition of a meme.

Why is such a loose definition unacceptable? The reason is that it makes the term 'meme' redundant. The best way to understand this is to consider the parallel situation in genetics. A gene is a packet of instructions encoded in DNA, which directs the development of cells. Now, if someone comes along and claims that all chemical processes are genetic, then the term gene would apply to everything that goes on in the body. Not only would we describe an animal's physical and mental characteristics as genetic, we would also describe the chemical process underlying digestion as genetic. But this is not what genetics is about. Genetics attempts to explain things at a higher level than common chemical process. The same is true of memetics. Memes are used to describe high level behavior that is not an innate part of an organism's behavioral repertoire.

In this section, I will offer a firm definition of the meme and will use this definition to show why the spread of milk-bottle opening in British birds is not memetic.

3.1 What is a Meme?

Memes are best thought of as sets of instructions that can be followed to produce behavior (see Silby 2000c). Instructions can be encoded in a number of formats, including:

1) musical notation,

2) written text,

3) visible (or vocal) action,

4) connectionist networks such as the neural structure of the brain.

A set of instructions that produces the behavior of, say, whistling the first 4 notes of Beethoven's fifth symphony (a well used example) can be encoded in any of these mediums. When the instructions are followed, the same behavior will result. When a mind encounters an instruction set that produces behavior (say a musical score), it can reproduce that behavior by creating an appropriate neural "program". To understand this, consider a similar situation in the world of robotics. Imagine that a robot is developed that contains a number of built in programs. These programs provide it with behavior essential for its survival. The robot's behavior might include walking, avoiding obstacles, grasping at objects, and the production of certain vocal sounds. Suppose that engineers also give the robot a program that gives it the ability to write small behavioral routines. In effect, this program gives the robot a means by which it can alter its own behavior by writing new programs. Suppose that a feature of this innate behavioral program is that it allows the robot to observe the behavior of other robots and write programs that produces the same behavior. In other words, the robot can imitate behavior. Now, the question is: what, exactly, are the programs that the robot writes for itself? They are not a part of the robot's innate behavior and they are produced primarily through imitation. They can be translated into different languages and written down on paper. They can also be transmitted to other robots who read the instructions or who imitate the behavior and write their own programs. These behavioral programs are what we are talking about when we speak of memes.

This is precisely the sort of process that goes on in humans. At some point in our history, biological evolution provided our ancestors with a capacity to imitate behavior. This meant that when humans observed the behavior of others, their brains would create the neural wiring needed to produce the same behavior. Such neural wiring patterns are lists of instructions, which can be translated into other mediums such as written language, outward behavior, or computer code. A list of instructions that produces behavior is the thing that spreads into the minds of others. A list of instructions that produces behavior is a meme.

3.2 Why do British Birds not have a milk-bottle opening meme?

Given the above definition of the meme, it would seem that British tits do not have an 'open milk-bottle' meme. This is because British tits have not acquired any new behavior. Blackmore (1999) suggests that the British tits already knew how to peck and it was simply a matter of one tit being attracted to another -- who happened to be sitting on a milk bottle -- and then carrying out its innate pecking behavior. I think it is also possible that a tit would be attracted to a milk bottle even if no other tit happened to be around. Most birds have an innate attraction to bright shiny objects, and milk-bottle-tops are usually made out of a silvery foil. Milk-bottle-tops must stand out like bright beacons for passing birds, who feel compelled to land and carry out their innate pecking behavior. Encountering cream beneath the silver foil would reinforce the behavior and make it more likely to occur in the future. However, because no new information has entered the bird's mind and no new behavior has been produced, milk-bottle opening in birds cannot be counted as a meme.

Reader and Laland believe they have an answer to this objection. Their answer draws a parallel between the British tit's milk-bottle opening behavior and the behavior of a human tennis player. For Reader and Laland, a tennis player acquires tennis playing memes by observing other tennis players. These memes carry information about the appropriate location to play, the objects a player uses, and how the player should move in a certain way to hit the ball. They go on to suggest that the tennis player does not learn any knew motor behavior such as running, holding, waving arms, or hitting, because these are already a part of their behavioral repertoire. The tennis player is carrying out innate behavior in order to achieve a non-innate goal. Reader and Laland then state that "Exactly the same logic applies to milk-bottle-top opening birds" (Reader & Laland 1999). The tits are not learning how to peck at milk-bottle-tops, just as the tennis player is not learning how to run or swing her arms around. They are, however, learning to peck a particular object to get cream, and this is a behavior and goal that is not innately defined. Therefore, milk-bottle opening in the tit is a meme.

The problem with Reader and Laland's suggestion is that they are glossing over important differences between the two cases. It is true that tennis players are carrying out innate movements, and it is true that the British tit is carrying out innate behavior. But the difference is that the tennis player's movements are combined in a non-innate fashion, whereas the tit's behavior is exactly the same as it would be if it was landing and pecking at any object. We would not be surprised if we saw a tit land on an apple and start pecking at it. Nor would we be surprised if the tit started pecking at the ground. So what is so special about a tit pecking at a milk bottle? Pecking is simply something that the British tit does. Also, consider the different reasons behind the tennis player and tit's behavior. The tennis player behaves in a specific way in order to win the game, which is a socially defined goal. The goal of the tit is to feed, and regardless of what food it happens to find, feeding is an innate objective.

There are also important differences in how the behavior is learned. The tennis player learns through imitation and verbal instruction, which are methods of passing on information that the tennis player does not already possess. The tit's primary method of learning is through behavioral conditioning. When a desirable outcome occurs (getting the cream), the behavior of pecking milk-bottles is strengthened. But this is not a memetic transfer of information, because no new behavioral information has been assimilated by the tit. No information (or instruction set) is being moved from one mind to another so no memes have been copied. At most, the case of the British tit is an example of innate behavior being triggered through observation and reinforced through reward.

3.3 Does this mean that animals don't have memes?

The quick answer to this question is, no! All I have shown is that the spread of milk-bottle opening in British birds is not an example of memetic transfer in animals. There may be a number of animal minds that can support memes, but I think caution is needed before we make any definite claims. Because of the diverse range of cognitive ability among species on Earth, we should consider the evidence on a case by case basis. The case of milk-bottle opening in the British tit is not a meme, but there are cases of bird behavior that definitely point to the existence of memes. For example, certain species of bird can imitate sounds ranging from the calls of other birds to the sound of a telephone ring. The production of such sounds is not innate to the bird, so we must conclude that some sort of memetic transfer takes place. Something in the bird produces the neural wiring required to enable it to reproduce the sounds it hears in the environment.

One way to check if memes are involved in animal behavior is to see if the behavior is spread through imitation. For example, if a dog (Fred) watches a person (Roger) jump through a hoop and then copies the behavior and jumps through the hoop himself, we could suppose that the idea of jumping through the hoop was passed from Roger to Fred. The hoop jumping meme would have copied itself into Fred's mind through the process of observation and imitation. I think it would be unlikely for this to happen. Dogs do not seem to have the capacity to observe and then copy behavior. Getting a dog to jump through a hoop requires a long period of behavioral conditioning. Furthermore, it is possible that coordinated jumping is actually innate to dog behavior, so we might not be correct in suggesting that jumping through a hoop is a meme.

There are some examples of animal behavior that look memetic. Some dogs can be trained to balance and walk on a moving ball. I find it difficult to believe that such behavior is innate, so it might be accurate to consider 'walking on a ball' behavior to be memetic. Of course, a dog cannot learn to balance on a ball by watching another dog. The only way that a dog can learn to walk on a ball is through extensive training and reinforcement. It seems that if walking on a ball is a meme, then it is a meme that has a very limited potential for replication. Since dogs cannot copy each other, the 'walk on a ball' meme can only replicate when the behavior is viewed by a human and then implanted into another dog's mind through a similar process of training.

Other possibilities of memes in animals include the choreographed walking of dressage horses, the transmission of directional information in honey bees, and food washing in Japanese macaques (as mentioned by Reader and Laland). The important point to note is that the possibility of memes in these animals is not enough to conclude that there are memes in these animals. Each case must be assessed independently. Perhaps the best way to test whether a behavior is innate or memetic is to isolate a member of the species in question. If, for example, a Japanese macaque was isolated from birth we could observe its behavior and find out if it has the compulsion to wash its food. The result of this test would settle the question for the Japanese macaque, but would leave the question of other animals open.

My goal has been to show that milk-bottle opening in the British tit is not a meme. To accomplish this goal, I offered a firm definition of the meme and worked with this definition to show that the spread of milk-bottle opening behavior in the tit does not exhibit the features of memetic transmission. According to the definition I offered, a meme is a collection of instructions that, when followed, give rise to behavior. These instructions can be encoded in a number of mediums including the neural pathways of the brain. The most important feature of memetic transmission is that the instructions transferred are new, and did not already exist in the mind of the receiver. In this way, we can say that the acquisition of a meme gives rise to a new, non-innate behavior. Since the British tit has an innate attraction to other tits, and since pecking at objects is innate to the tit, we must conclude that milk-bottle opening is not the result of new behavior and is therefore not memetic.

I cannot show that non-human animals do not have memes. In fact, I have offered some possibilities of memes in non-human animals. I have however, shown that milk-bottle opening in the British tit is not memetic, and that the existence of this behavior is not sufficient to conclude that animals have memes. Because of the diversity of cognitive ability among animals, I do not believe the question of memes in animals can be answered definitively by pointing to a few cases. Some animals may have memes while others do not. We must proceed cautiously and take a case by case approach. By doing this we will be able to construct a catalog of animals whose minds have memetic compatibility.

Blackmore, Susan. (1999). The Meme Machine. Oxford University Press.

Dawkins, Richard. (1976). The Selfish Gene. Oxford: Oxford University Press.

Dawkins, Richard. (1986). The Blind Watchmaker. London: Longmans.

Dennett, Daniel C. (1991). Consciousness Explained. Penguin Books, 1993.

Dennett, Daniel C. (1995). Darwin's Dangerous Idea. Penguin Books, 1996.

BRENT SILBY
Department of Philosophy
University of Canterbury

Nanomachines are devices built from individual atoms. Some researchers believe that nanomachines will one day be able to enter living cells to fight disease. They also hope to one day build nanomachines that will be able to rearrange atoms in order to construct new objects. If they succeed, nanomachines could be used to literally turn dirt into food and perhaps eliminate poverty.

In this article I will outline some of the possible uses of nanomachines. I will then assess some of the problems involved in producing such machines. One of the problems I will look at is that of producing self-replicating machines. Will these machines be controllable? Or will their reproduction escalate exponentially, thus putting our whole planet in danger.

My conclusion will be that nanomachines offer humanity hope for the future, so the research should be pursued. However, I will also suggest that the dangers involved in producing self-replicating machines out weigh the potential gains and for this reason, self-replicating machines should not be built.

As the terminology implies, nanomachines are extremely small devices. Their size is measured in nanometers (a nanometer is about 1 billionth of a meter) and they are built from individual atoms. During the 1980's and 1990's, futurist and visionary K. Eric Drexler popularized the potential of nanomachines. For Drexler, the ultimate goal of nanomachine technology is the production of the 'assembler'. The assembler is a nanomachine designed to manipulate matter at the atomic level. It will be built with extremely small 'pincers' (as small as a chain of atoms) which will be used to move atoms from existing molecules into new structures. The idea is that the assembler will be able to rearrange atoms from raw material in order to produce useful items. In theory, one could shovel dirt into a vat and wait patiently for a team of nanomachine assemblers to convert the dirt into an apple, a chair, or even a computer. The machines in the vat would have a molecular schematic of the object to be built encoded in their 'memory'. They would then systematically rearrange the atoms contained in the dirt to produce the desired item.

Another goal of nanotechnology is to design nanomachines that can make copies of themselves. The thought is that if a machine can rearrange atoms in order to build new materials, it should also be able to build copies of itself. If this goal is achieved, products produced by nanomachines will be extremely inexpensive. This is because the technology (once perfected) will be self-replicating and will not require specific materials, which might be rare and therefore cost money. Arthur C. Clarke has predicted that nanotechnology will herald an end to conventional monetary systems.

If scientists manage to build nanomachines that can rearrange atoms, a world of exciting possibilities will open up. Purpose designed nanomachines could be used to provide breakthrough treatments for many diseases. Medical nanomachines programmed to recognize and disassemble cancerous cells could be injected into the bloodstream of cancer suffers, thus providing a quick and effective treatment for all types of cancer. Nanomachines could be used to repair damaged tissue and bones. They could even be used to strengthen bones and muscle tissue by building molecular support structures by reassembling nearby tissue. With the ability to manipulate human cells at the atomic level, medical science will rapidly devise treatments for most human illnesses. And since nanomachines will be designed to make copies of themselves, these treatments will be inexpensive and available to the entire population.

Food shortages and starvation will be a thing of the past if nanotechnology is perfected. Nanomachines will be able to turn any material into food, and this food could be used to feed millions of people world wide. Again, since the technology is self replicating, food produced by nanomachines will be low cost and available to all.

As well as food, nanomachines will be able to build other items to satisfy the demands of our growing population of consumers. Clothing, houses, cars, televisions, and computers will be readily available at virtually no cost. Furthermore, there will be no concern about the garbage produced by the new consumerist society because nanomachines will convert it all back into new consumable goods.

Environmental problems such as ozone depletion and global warming could be solved with nanotechnology. Swarms of nanomachines could be released into the upper atmosphere. Once there, they could systematically destroy the ozone depleting chlorofluorocarbons (CFCs) and build new ozone molecules out of water (H2O) and carbon dioxide (CO2). Ozone (O3) is built out of 3 oxygen atoms, and since water and carbondioxide both contain oxygen, the atmosphere contains a plentiful supply of oxygen atoms. While the ozone construction teams are at work in the upper atmosphere, teams of specialized nanomachines could be employed to destroy the excess CO2 in the lower atmosphere. CO2 is a heat trapping gas, which has been identified as one of the major contributors to global warming. Removing excess CO2 could help halt global warming and bring the planet's ecosystem back into balance. This will benefit all species on Earth.

The perfection of nanotechnology and the production of nanomachines could herald a new age for humanity. Starvation, illness, and environmental problems could quickly come to an end. But how realistic are the goals of nanotechnology? Will it ever be possible to produce machines the size of atoms? And if so, how feasible is it to build nanomachines that can build objects from the atom up? Is it possible for nanomachines to build copies of themselves? Before we get carried away with the promises of nanotechnology, we should take a look at some of the problems that are yet to be solved.

An important challenge to overcome is one of engineering. How can we physically build machines out of atoms? Rearranging atoms into new shapes is essentially building new molecules (nanomachines are sometimes called 'molecular machines') and this is no easy task. Using contemporary technology to rearrange atoms has been said to be analogous to assembling LEGO blocks while wearing boxing gloves. It is virtually impossible to snap individual atoms together. All we can do is crudely push large piles of them together and hope for the best. Scientists hope that once this initial challenge is overcome, nanomachines will usher in a new age of molecular engineering and previous problems will be a thing of the past. The new nanomachines will allow scientists to take off the boxing gloves and accurately snap together individual atoms to build virtually any molecule (within the laws of physics, of course).

This is nice in principle, but the question of how to build the first nanomachines remains. Nanotechnologists think that it will be impossible to build the first nanomachines by using large scale equipment (Chen C. 2000). Although progress is being made in the miniaturization of integrated circuits and in the ultra-fine finishing of high quality optical components, the large scale technology being used doesn't let us take off the boxing gloves. There is a limit to how far down these machines can go. Super smooth lens polishing is one thing, but moving individual atoms is something else all together. Nanotechnologists need to get the boxing gloves off before they can build the first nanomachines.

One way to work without boxing gloves is to patiently experiment with chemical synthesis. The idea is to build molecules of increasing complexity by allowing atoms to assemble or rearrange in natural ways. When molecules are mixed, they naturally form new molecules. Through extensive experimentation, more control can be gained over how molecules are formed. In time, it is conceivable that chemists will be able to position individual atoms by using a range of techniques developed in chemical synthesis.

One of these techniques might involve the removal and relocation of hydrogen atoms. This technique could be developed with knowledge of how hydrogen atoms interact with other atoms. For example, it is known that the propynyl radical C3H3 (its made out of 3 carbon atoms and 3 hydrogen atoms) is 'attracted' to hydrogen. It is also known that this radical has two ends. At one end there is a highly reactive radical, while at the other end there is stable carbon. This feature means that chemists may be able to synthesize a larger molecule with the propynyl radical at one end (the rest of the molecule would be built from the stable carbon end). If this larger molecule was held on a positioning device, it could be used to extract hydrogen from a range of different molecules by passing them by the reactive radical (Merkle R.C. 1993).

Chemical synthesis is promising. In computer simulations, molecularly stable gears and cogs have been formed through chemical synthesis.

If chemists and engineers succeed in building nanomachines the hope is that these machines will be able to build a whole range of new molecules from the atom up. If all goes well, scientists will never have to move atoms round while wearing boxing gloves and the lengthy experimental process of chemical synthesis will no longer be required. But will it be that easy?

In order to make new molecules, a nanomachine has to somehow 'grab' individual atoms with its pincers and move them into new positions or attach them to other molecules. This seems to be quite simple, but as George M. Whitesides (2001) points out, there are serious problems that need to be overcome. Consider, for example, the fact that a nanomachine's pincers will be made out of several atoms and will therefore be larger than the individual atoms that it needs to move around. This means that the intricacy and accuracy of the nanomachine's movement will be severely limited. It will be clumsy. Assembling atoms would be like trying to piece together a mechanical wristwatch with your fingers rather than small tweezers.

Another problem arises from the fact that individual atoms are compelled to 'attach' to other atoms. Some atomic bonds can be extremely strong (especially with carbon atoms) so pulling them apart will require large amounts of energy. Furthermore, since carbon atoms attach to just about anything it seems likely that they will bond to the nanomachine's pincers after they've been pried away from their original molecules (Whitesides 2001). The only way to remove them could be to move them to molecules that they are more strongly attracted to. But then there is the possibility that the entire nanomachine will stick to the molecule. The situation is analogous to trying to build a wristwatch with magnetized tweezers and screwdrivers. It can't be done because the individual components stick to the tools.

Drexler et al (2001) brush aside these problems. They suggest that such concerns arise from a misunderstanding of how nanomachines work. For example, the idea that nanomachines use 'pincers' to move objects around is nothing more than a poor metaphor. In reality, nanomachines might contain an active tip (like the hydrogen extractor described above), which is no larger than the atom it is designed to manipulate. So Whitesides' concerns about the size of a nanomachine's pincers are easily answered. However, his concerns about the bonding of carbon atoms to nanomachines seem more difficult to answer. Drexler attempts to bury the problem by citing theoretical work done with the hydrogen extraction tool and by referring to experimental work done with hydrogen atoms. He doesn't directly address concerns about manipulating carbon atoms. This is important, because carbon is one of the most common atoms found on Earth and will no doubt be involved if nanomachines are used to build new molecules. Progress made with hydrogen might not translate easily to future work on carbon atoms.

Drexler does, however, mention some very promising work by Wilson Ho and Hyojune Lee. In an experiment, Ho and Lee

"...used an STM tip first to locate two carbon monoxide (CO) molecules and one iron (Fe) atom adsorbed on a silver surface in vacuum at 13 K. Next, they lowered the tip over one CO molecule and increased the voltage and current flow of the instrument to pick up the molecule; then they moved the tip-bound molecule over the surface-bound Fe atom and reversed the current flow, causing the CO molecule to covalently bond to the Fe atom, forming an iron carbonyl Fe(CO) molecule on the surface. Finally, the researchers repeated the procedure, returning to the exact site of the first Fe(CO) and adding a second CO molecule to the Fe(CO), forming a molecule of Fe(CO)₂, which in subsequent images of the surface appeared as a tiny "rabbit ears" structure, covalently bound to the silver surface. Ho's group has also demonstrated single-atom hydrogen abstraction experimentally, using an STM" (Drexler et al. 2001).

This type of work will hopefully lead to more complex manipulation of atoms, and this could result in the development of tools that successfully 'pick and place' carbon atoms.

As our technological capacities develop, the promise of nanomachine technology becomes more of a reality. We may one day see the successful creation of nanomachine assemblers. These machines could end hunger and bring in a new age of advancement for humanity. Nanotechnology offers us big promises in a small package. However, the advantages it promises do not come for free. They come with some very big risks.

Cutting edge technology can take a while to catch on in the commercial world. However, there is one place in which it catches on very quickly: The Military! During humanity's history, technological research has moved fastest when there is a potential military application. The danger is that this trend will continue with nanotechnology. Imagine the possible uses of nanomachines in warfare. Self replicating nanomachines designed to target and destroy organic material could be released over enemy territory reducing the population to dust within a matter of hours. If these machines were designed to destroy each other after (say) 24 hours, the enemy's country would be left empty and safe to be invaded by military forces. Biological warfare would be a thing of the past since nanomachine warfare would be so much safer (well, for the 'good guys' anyway).

The only way to prevent this use of nanomachines would be through international agreements. Unfortunately, not all countries are willing to sign such agreements. And those who do sign might be tempted to develop the technology in secret--just incase the enemy is doing the same thing. Perhaps the most we could hope for would be a stalemate situation like the one between the United States and the U.S.S.R during the cold war. If both sides have the technology, they might be too nervous to use it, since they know that the other side will retaliate.

A more serious danger of nanomachine technology involves the ability to self replicate. Imagine that a nanomachine has the ability to make a copy of itself by rearranging the atoms contained in any nearby matter. Since it is producing an exact copy of itself, it is likely that the 'offspring' machine will be able to replicate. This is, after all, the way in which nanotechnologists intend to keep the cost of nanomachines down.

So now we have 2 nanomachines that can replicate. One more cycle will produce 2 more, which leaves a total of 4.

4 becomes 8.

8 becomes 16.

16 becomes 32, and so on.

After only 27 generations we would have over 134 million nanomachines on our hands. Since they are molecular, this doesn't seem like a big number. But the number could keep growing. After 39 generations there would be over 549 billion nanomachines on the planet. The point is obvious. Without a way of controlling the reproduction of nanomachines, the planet is in danger of being over run. Furthermore, since the nanomachines are using the planet's resources as raw material with which to replicate, the danger is that the planet could eventually be transformed into a seething mass of nanomachines.

George Whitesides (2001) responds to this problem by pointing out that Earth has already been ravaged by molecular machines--namely, biological cells. This is true. Earth was a much different place 3.5 billion years ago before the emergence of life. Self replicating cells have, over 3.5 billion years, completely transformed the planet. They have changed the planet from a world of inorganic minerals with a CO2 rich atmosphere, to a world that is perfect for biological life.

But this fact doesn't negate the danger in creating replicating nanomachines. In fact, Whitesides has reminded us that it is possible for molecular machines to replicate exponentially and transform the planet. If self replicating nanomachines get out of control, then they could alter the planet to such an extent that it is no longer suitable for biological life.

A possible solution to the problem is to limit the replicating abilities of nanomachines. For example, a mechanism could be developed by which new nanomachines are tagged with a number. This number could represent their generation. So, a nanomachine labeled 'gen 2' would produce offspring labeled 'gen 3', and their offspring would be labeled 'gen 4'. The replicating algorithm could be designed to only function if the generation number is less than 4. Also, nanomachines with a generation number higher than 1 could be encoded with a function that limits the number of reproductive cycles they can execute. By building in these safeguards, we may be able to control the population of nanomachines while at the same time allowing the existence of a number necessary to facilitate some of the advantages mentioned earlier.

However, these safeguards may not be enough. The biological world has shown us that evolution occurs and cannot be stopped. The same may be true of the nano-world. Consider the idea that each time a nanomachine makes a copy of itself, there is a possibility that an error could be made during the copying process. Such errors could be very small--perhaps no larger than a single 'bit' of information. Now imagine what would happen if an error occurred while a nanomachine was building its offspring's copying mechanism. To be more specific, imagine that a single 'bit' error occurred when encoding the function that limits the machine's replicative abilities. So, instead of checking that the machine's generation number is less than 4, it checks to see that it is less than 40. When this error is passed on to the machine's offspring, they will reproduce providing their generation number is less than 40. Since the error will be passed on to each subsequent generation, there will be a substantial explosion in nanomachine population. A single error could have the potential to send the nanomachine population out of control. And the more reproducing nanomachines there are, the greater the chance of another error occurring in at least one of them.

The only way to avoid the problem of uncontrollable replication is to avoid building self-replicating nanomachines. It may be true that self-replicating machines are the only way to ensure a cheap supply of nanomachines, but the potential risks outweigh the benefits. If nanomachines are built individually in labs, they will still be useful to cure disease and they will still be potentially useful for rearranging molecules to build new objects such as food. The only drawback is that none of it will come for free. Someone will still have to pay for the construction of the machines, and this means that their products will have to be paid for by consumers. So, poverty will not be eradicated. However, it could be that producing food with nanomachines is faster and cheaper than conventional means, which will mean that poverty may be eased a bit. Furthermore, if governments are willing to invest in the technology, nanomachines may be able to be used to fix some of our environmental problems by repairing the damage we've done to the atmosphere. So the research is worth continuing.

Nanomachines offer humanity hope for the future. The idea that we could one day cure diseases, fix the atmosphere, and reduce poverty in the world is an exciting one. If scientists can overcome the technical difficulties involved in producing nanomachines capable of these goals, then the fruits of their efforts will benefit us all. However, we must be cautious. The temptation to build self-replicating machines is strong, since it will give us an endless supply of new nanomachines at virutally no cost but self-replicating machines have the potential to get out of control. The best efforts to limit their replicative abilities may be insufficient, and our planet could be at risk of being over run by machines that can consume anything to produce more machines at an astounding rate.

The benefits of building nanomachines that can manipulate matter are real and cannot be ignored, so the technology should be pursued with vigor. However, the risks in producing self-replicating machines outweigh the benefits, so I conclude that self-replicating nanomachine technology should not be pursued. We should focus our efforts on perfecting machines that can produce the benefits outlined in this article while never building machines that can make copies of themselves.

© Copyright BRENT SILBY 2002
=ALL RIGHTS RESERVED=

Drexler, K. Eric (1992). Nanosystems: Molecular Machinery, Manufacturing, and Computation (New York: Wiley, 1992).

Drexler, K. Eric (1991). Unbounding the Future : Nanotechnology Revolution (New York: William Morrow, 1991)

This work is protected under international Copyright law. This work (or portions of this work) may be downloaded, distributed, quoted, or referred to for the purpose of criticism, evaluation, or for furthering knowledge and research in this field. When referring to the work of others, please remember to cite the author's name, the title of the work, the place and date of publication (or internet address).

Astronomers are showing us that planets are a relatively common phenomenon in our galaxy. Until the mid 1990's, there was no evidence of 'exoplanets' (that's what scientists call planets that orbit stars other than our sun), and any theories regarding the existence of planets outside our solar system were mere speculation. As far as we knew, our solar system was the only one that contained planets. If this had turned out to be true, our hopes of finding extra-terrestrial life would have been virtually non-existent. The search would have ended there.

But now our technology has improved and scientists have discovered over 50 exoplanets. Interestingly, astronomers cannot actually see these planets. This is because planets are not self luminous. Usually the only way to see a planet is from reflected light, which usually originates in a nearby star. But exoplanets are so far away from Earth that their reflected light is obscured by the brightness of their parent star, which means that astronomers have to find them in other ways. One such way is to infer their existence by looking at the gravitational effect they have on their parent star. Astronomers observe stars to see if they are being pulled by something close by -- something that can't be seen. Small 'wobbles' in a star's movement betray the presence of an orbiting planet. Observations of stars in this galaxy show that such wobbles are common place. If scientists are correct in supposing that these wobbles are caused by planets, then it seems that planets are relatively common. Enough evidence exists to suggest that the galaxy is full of planets.

By analyzing the magnitude of star wobbles, astronomers have determined that most of the exoplanets discovered so far are gas giants like Jupiter. Such planets are inhospitable and are not suitable for life as we know it. But we are just at the beginning of the search. As technology improves, smaller planets may be found. One day we may be able to use spectral analysis to determine what sorts of atmosphere they have. Perhaps we will find some that have Earth-like atmospheres. This possibility staggers the imagination. As far as we know, Earth is the only planet that supports life, but if we discover other Earth-like planets, we will face the possibility that some of them might support life -- some of which might be intelligent.

The discovery of extra-terrestrial life will be the most profound discovery in human history. It will change the way we view ourselves and the way we think about our place in the universe. But the question is still open. The discovery is still waiting to be made. Will we ever discover life on other worlds? Since other star systems are so far away, we cannot hope to explore them in the foreseeable future. With our current technology, a trip to the nearest star system would take over 40,000 years. And even if we developed propulsion systems that would move us at the speed of light, we would have Einsteinian relativity problems to overcome. One of these problems is the so-called 'time dilation' effect, which suggests that time moves slower for a person who is travelling at great speed. In effect, Einstein's theory states that if I travel to a nearby star system at the speed of light, from my point of view the trip would be instantaneous, but from the point of view of my friends on Earth, I will have been gone for several years. Given such transportation difficulties, it appears that the best way to discover life in other star systems is to observe it from here.

Finding extra terrestrial life is no easy task. Since we are stuck here, on Earth, the only type of alien life we can search for is intelligent life capable of radio communication. This limits the scope of our search, but for now there is no choice. SETI (the Search for Extra Terrestrial Intelligence) is an organization that is committed to discovering intelligent life in the galaxy. They search for life by systematically scanning the sky for radio signals that do not originate on Earth. The hope is that we might be able to eavesdrop on the internal communications of another civilization. We know that the chance of picking up a signal that was deliberately sent to us is extremely low, but we also know that intelligent civilizations probably use radio for their own communications. This is what SETI is trying to find. SETI have organized a system by which millions of personal computers around the world work on small chunks of radio data that has been picked up by radio telescopes. These computers sort through random background noise, and search for patterns that have no astronomical origin -- patterns that might have an intelligent origin. With any luck, SETI will find a star system that contains a planet, which is inhabited by intelligent life that has attained the capability to communicate through radio. If this happens we can decide whether we should attempt to communicate directly with the civilization -- a pretty big decision -- or leave it alone.

The continuing discovery of exoplanets gives us hope. We now know that the nine planets orbiting our sun are not alone in the universe. Knowing that there are planets elsewhere in our galaxy gives us good reason to suppose that extra terrestrial life exists somewhere in the cosmos. After all, if life can arise on Earth, then life can arise on other suitable worlds. This thought reaffirms the wisdom of searching for alien civilizations and gives the human species a goal to work towards.

Department Of Philosophy, University of Canterbury, New Zealand

Copyright (c) Brent Silby 1998

This work is protected under international Copyright law. This work (or portions of this work) may be downloaded, distributed, quoted, or referred to for the purpose of criticism, evaluation, or for furthering knowledge and research in this field. When referring to the work of others, please remember to cite the author's name, the title of the work, the place and date of publication (or internet address).

Consciousness lies at the very heart of the mind/body problem. So far it has been resistant to scientific analysis and explanation. Essentially, the mind/body problem consists in the realisation that while it may be possible to discover physical principles by which the mind works, the phenomenon of consciousness may escape the physical story. It seems that we have what the literature refers to as an explanatory gap, and that consciousness, by its very nature, cannot be accounted for by physical theories of the mind/brain. This type of thinking suggests a metaphysical dualism between the body and the mind. Anti-physicalist arguments have taken a similar form since Descartes, and often involve conceivability arguments. In other words, the anti-physicalist puts forward thought experiments that involve the conceptual possibility of certain scenarios (Levine 1993: p543). These thought experiments involve imagining situations in which a complete physical knowledge of the world (including the brain) fails to provide an account of conscious experience. For example, it is claimed that we can imagine a situation in which there exist creatures who are physically and behaviourally identical to us, yet who have no conscious experience. The conceivability of such creatures is supposed to show that consciousness is not logically supervenient on the physical.

It is something of an irony that while our minds are the one thing of which we are directly aware, the existence and nature of our minds is one of the most puzzling mysteries that we have ever grappled with. Being a part of the physical world, we would like to be able to explain the workings of our minds in scientific terms and there has been a great deal of success in this area. Neuroscientists are working hard to provide us with neural models of the brain and progress is being made in understanding the way the brain works. However, despite the progress being made, philosophers such as David Chalmers believe that consciousness escapes reductive explanations.

In his book `The Conscious Mind', Chalmers makes significant use of thought experiments involving zombies. These thought experiments are supposed to show that consciousness is not logically supervenient on the physical and that conscious states, or the properties of those states, called qualia, are resistant to physical and functional analysis.

In this paper I will first explain the concept of supervenience, then describe the problem of qualia and explain why qualia are a problem for physical theories of mind. I will then describe the zombie argument and will explain Chalmers' attempts to use it as an argument for the conclusion that qualia do not logically supervene on the physical. My goal will be to show that the supposed conceivability of zombies does not offer any serious refutation of physical or functional theories of the mind.

The notion of supervenience is an important part of Chalmers' argument. The concept of supervenience allows us to formally locate relations of dependence between sets of properties. If A-properties supervene on B-properties, then whenever you find B-properties, you will find A-properties. Now, there is an important distinction to be made between two types of supervenience - logical supervenience and natural supervenience. When we say that B-facts are naturally supervenient on A-facts, we are making a claim about the empirical world. For example, in our world the facts and laws of gravity supervene on the facts and laws relating to the mass of objects. There is a law-like correlation between how massive objects are and the gravitational forces that they give rise to. So, in this world (or universe), whatever object we choose to examine, that object will be surrounded by a gravitational field that is proportional to its total mass. There are no exceptions to this law. Gravity naturally supervenes on mass. However, we can imagine a possible universe in which the laws of physics are vastly different to the laws in our universe. Perhaps in our imagined universe there is no gravity. In our imagined universe gravity would not supervene on mass because in that universe gravity does not exist. Our ability to imagine such a universe shows us that gravity is not logically supervenient on mass. For B-properties to be logically supervenient on A-properties there must be no logically possible situations that share identical A-properties and yet remain distinct with regard their B-properties (Chalmers 1996: p35). Logical supervenience is defined in terms of logically possible worlds. For example, we could say that biological properties supervene logically on physical properties, because it would be logically impossible for a world to exist that was physically identical to ours and yet was biologically distinct (Chalmers 1996: p35). When all the physical properties of our world are in place, the biological properties come along at no extra charge. Even God could not create a universe that was physically identical to ours (including the position of every particle), while keeping it biologically distinct. If penguins exist in our world, then any world that is physically identical to ours also contains penguins. Biology is logically supervenient on the physical.

Now, Chalmers has made the claim that consciousness is not logically supervenient on the physical. So, for Chalmers, there is a possible world that is identical to our world in all physical respects, but is distinct in the fact that consciousness (in humans) does not exist. But what is Chalmers referring to when he speaks of consciousness? The word `consciousness' can be used to refer to a number of different aspects of human life, but Chalmers has a specific aspect of consciousness in mind. He is concerned with the existence of qualia.

One of the central problems in the contemporary philosophy of mind is the problem of qualia. The word `qualia' refers to the intrinsic properties of our experience, including phenomenal properties such as colours, sounds and sensations of pains. These `raw feels' seem to be intrinsic properties of our experience and yet cannot be found in the non-mental world. According to tradition, qualia are defined as ineffable, intrinsic, private and directly apprehensible in consciousness (Dennett 1988: p622). Sensations of colours, pains and sounds seem very real to us and yet when we turn to science for an explanation of the colour blue, all we learn is that there are certain wavelengths of electromagnetic radiation that are being reflected from the surfaces of various objects. There is no blue, just radiation. The question we are left with is: where is the blueness if it is not in the outside world? An intuitive answer to this question could go like this:

This story answers the question badly. It may feel as if the blue colour is there, in experience, but seeming like something is the case does not show that it is the case. Daniel Dennett suggests that it would be a waste of resources for the brain to work in this fashion. Once a sensory discrimination has been made, there is no need for it to be made again by other parts of the brain. In other words, the process from the retina to the visual cortex is the process by which the colour is discriminated so there is no need for a `colour box'. The process would have been wasted if it simply supplied its data to a box where the colour to be discriminated was reproduced. But this brings us back to the question: Where is the blue that we experience? There is nothing but wavelengths of electromagnetic radiation in the non-mental world, and there is no box filled with blue colour inside our heads so why do we have the blue sensation? Will it ever be possible to use physicalist functionalism to account for qualia? Leibniz said:

Supposing that there were a machine whose structure produced thought, sensation, and perception; we could conceive it as increased in size with the same proportions until one was able to enter into its interior as he would into a mill. Now, on going into it he would find only pieces working on one another, but never would he find anything to explain perception (Leibniz, quoted in Rey 1997: p313).

In this passage, Leibniz is pointing to the mysteriousness of consciousness and the idea that the mere physical and functional organisation of a brain cannot explain the qualities of experience. This leads us to wonder if qualia supervene on the physical and functional organisation of the brain. If qualia do not supervene on the physical and functional states of the brain, this would mean that even if we had a complete understanding of the brain's physical composition and functional organisation, we would not be able to account for qualia because they cannot be captured by a physicalist account of the mind. So we could imagine a situation in which the brain's physical and functional structures remain the same and yet qualia do not exist. There are many thought experiments to show the conceivability of an entity functionally the same as us yet being devoid of qualia. Such thought experiments imply that consciousness is a property that cannot be accounted for by physicalism. David Chalmers is a supporter of this view and believes that consciousness cannot be reductively explained. In `The Conscious Mind', Chalmers attempts to show that no physical explanation of the brain can account for consciousness and for this reason, he holds that consciousness escapes any attempt at a reductive explanation (Chalmers 1996: p93). Chalmers claims that the logical possibility of entities that are functionally the same as us but devoid of consciousness, is one way of showing that consciousness is not logically supervenient on the physical composition or functional structure of the brain. The entities that Chalmers refers to are known as zombies.

A philosopher's zombie is a hypothetical entity that is both physically and functionally identical to a normal human being. Zombies are molecule for molecule copies of humans and yet they are totally devoid of conscious experiences. My zombie twin is physically identical to me and is therefore functionally identical to me. We can imagine that he lives on a duplicate planet Earth and has led a life that is exactly the same as the life that I have lived. The important point to note is that for Chalmers, my zombie twin is also identical to me with regards his behaviour and psychology. Psychology, for Chalmers, refers to the functional organisation of the mind. This includes behavioural links and issuing verbal reports. A psychological state is defined by the role it plays in the collective cognitive system and the behaviour that it gives rise to. For example, if my zombie twin looks at the blue sky, he will make the same comments about his experience as I would. He may state that he enjoys eating chocolate but is hesitant to do so because of the pain he experiences from a hole in one of his teeth. The fact that my zombie twin is functionally and psychologically identical to me follows from the fact that he is physically identical to me. Our brain states play exactly the same functional roles and give rise to an identical psychology. If a certain brain state has the functional role of making me wince in pain, then my zombie twin will behave in exactly the same way when his brain is in that state.

But what about conscious experiences? This is where my zombie twin and I differ. My zombie twin is similar to me in all respects except that his life contains no phenomenal feel. There is nothing it is like to be a zombie. As David Chalmers puts it `All is dark inside' (Chalmers 1996: p96). Accepting the possibility such zombies leads to the view that consciousness is not logically supervenient on the physical. This view is closely related to the doctrine of epiphenomenalism, which claims that consciousness has no causal powers and has no effect on the world (Guzeldere 1997: p41). If we were to accept the epiphenomenalist view of qualia, we would have to accept that qualia cannot be defined in terms of a role that they play in the formation of beliefs and the production of behaviour. Qualia would seem to be nothing more than a needless feature of human experience, which nature has provided us with for no reason whatsoever. But if qualia serve no purpose, why would nature provide us with them? Before accepting epiphenomenalism, we should explore Chalmers' claims and decide whether or not zombies are possible.

For Chalmers' argument, the crucial point is that zombies are logically possible and there is no contradiction or conceptual problem with imagining their existence. Chalmers admits that in the natural world it is likely that anything physically identical to us would have conscious experience just as we do and if a creature was devoid of qualia, it would probably differ from us in some important physical respects. However, even if zombies are naturally impossible, their logical possibility can still be used to argue for the logical non-supervenience of qualia on the physical. Examining the logical possibility of zombies can also be used to establish physicalist conclusions about qualia. If it could be shown that zombies are not logically possible, we would then be able to pursue reductive functional explanations of consciousness.

It seems that establishing logical possibility is a simple task. For example, we can easily conceive of starships that can traverse the galaxy in just a few hours by imagining a possible world in which humans have developed the technology to make this possible. In the natural world travelling across the galaxy in such a short space of time would not be possible because to do so we would need to violate Einstein's mass/energy equations. In order to travel across the galaxy in just a few hours, a starship would need to travel faster than light and as such would require an infinite amount of energy. Nevertheless, it is possible to imagine a world in which the laws of physics are different in such a way as to allow such starships to exist. We can use the same technique to establish the logical possibility of many things that do not exist in the natural world. It certainly seems conceivable that there is a possible world in which zombies exist just as it is conceivable that a world exists in which starships can travel faster than light. Chalmers believes that there is nothing incoherent in imagining such a world and he maintains that the burden of proof lies on those who claim that zombies are not logically conceivable (Chalmers 1996: p96). Someone who states that zombies are not logically possible must show that the concept of a creature being both physically and psychologically identical to us but lacking in conscious experience is contradictory. If no such contradiction can be found, we must conclude that consciousness does not logically supervene on the physical.

Chalmers often characterises the zombies' lack of conscious experience by using the phrase `all is dark inside'. Humans in this world, on the otherhand, have a full and rich conscious life and experience sounds, smells, touches and colours. Given the distinction between zombies and humans, would it be correct to say that for humans `all is colourful inside'? Probably not. As I stated earlier, according to Dennett, there does not appear to be any place in the brain that contains colour or any other quale. The human brain, like a zombies brain, is dark inside. This, of course, expresses scepticism about the very existence of qualia. Chalmers believes that for humans, there is a `mental' box that contains instances of qualia but in zombies, this box does not exist. When Chalmers speaks of darkness inside a zombie, he is using the term darkness to represents the zombies' total lack of qualia. But this is misleading in some ways and betrays some fundamental problems in imagining the possibility of zombies. Can we really imagine zombies as being all dark inside? Surely the experience of darkness or lack of qualia is a quale in itself (Cottrell 1996). Attempting to imagine a zombie involves imagining `what it would be like' to be a zombie and have no conscious experiences. But thinking of a zombie in these terms cannot work because according to the zombie hypothesis `there is nothing it is like to be a zombie'. How can we imagine what it is like to be something when there is nothing it is like to be it? There is no conception of darkness or of a lack of experience for a zombie - there is nothing at all going on in there. Imagining what it would be like to be a zombie is not at all like imagining being blind or deaf because presumably there is `something it is like' to be blind or deaf. Trying to imagine what it is like to be a zombie is like trying to imagine what it is like to not exist, and I maintain that this is an impossible task. The obvious reply to this would be to ask why it is necessary to imagine what it is like to be a zombie. People can imagine unicorns as being logically possible without having to imagine what it is like to be a unicorn, so why can't the same be true of zombies? I would respond to this by suggesting that there is a difference between imagining a unicorn and imagining a zombie. To imagine a unicorn, one simply has to take two naturally existing things in the world and join them together - a horse and a horn. We can imagine a world in which evolution gave horses horns because we see creatures with horns in our world all the time. But imagining our zombie twins is more difficult. We can imagine our twins as being identical to us easily enough, but we also have to imagine that they have no conscious experiences. To accomplish this, we have to imagine ourselves as having no conscious experiences and this is where the problem lies. We cannot imagine what it is like to have absolutely no conscious experiences. This is why philosophers such as Daniel Dennett believe that people inevitably fail at the task of imagining zombies. Often the assumptions people make about the zombies' lack of an internal life is incompatible with the statements they make about the zombies' external behaviour. Their description of zombies often contradicts the idea that zombies are behaviourally and functionally identical to us (Dennett 1995a: p172). Indeed, as I have already explained, it seems to be impossible to imagine completely normal external behaviour at the same time as imagining a total internal absence of experience. So it would seem that one of the major problems with imagining zombies is that people underestimate exactly what is involved and they often include facts that are not really imaginable. Many of the properties ascribed to zombies are stipulated rather than imagined. The problem here is that it is possible to make any stipulation or assertion that we feel is necessary to make a thought experiment work even though these stipulations may be totally unimaginable (Horgan 1987: pp496-500). For example: Imagine a creature that is totally identical to yourself in every physical and psychological respect. Imagine also that this creature has a full conscious experience that is identical to your own conscious experience. Now imagine that this creature is not alive (Dennett 1995a: p176). Given that this creature is physically, psychologically and behaviourally identical to us, we may think that we have discovered a mysterious property that cannot be explained. Life! We are alive but our imagined twins are not alive so what is this thing called life? Dennett suggests that this mystery would only arise if we thought that life was some extra property that could be removed or added to the totality of a normally functional human (Dennett 1995a: p176). But regardless of what we think, life is not this sort of property. Perhaps we should not think of consciousness as some property that could be removed from the totality of human function. Some people may claim to able to imagine such a creature as described above, but I suspect that their imaginations would fail to fully conceive of such a scenario. I believe that if people claim to have imagined duplicates of themselves that are not alive, they have not actually imagined the situation, they have only made stipulations about what the creature should be like. Perhaps the same is true when we attempt to imagine zombies.

Shoemaker has pointed to another problem with the conceivability of zombies and has used the zombie type argument to argue for reductive explanations of consciousness. As humans, we make judgements and form beliefs about our experiences and qualia. We can do this through introspection. If this is true, it would seem that our experiences cause our introspective beliefs. Now, remember that zombies are physically, functionally, psychologically, and behaviourally identical to humans. This means that zombies and humans share identical introspective mechanisms. Of course, since zombies exhibit the same behaviour as humans, it follows that they make the same claims about their experiences as we do. They talk about enjoying a rich and full conscious life in exactly the same way as humans do. Zombies form beliefs about their experiences through their introspective mechanisms, and since they have no qualia, we must accept that qualia do not play a role in the formation of their beliefs. But, if we have the same introspective mechanisms, then it would seem that qualia do not play any causal role in our formation of beliefs. Zombies insist that they are not zombies, just as we do. But if we are forming those beliefs by using exactly the same mechanisms as zombies, then we should not be able to tell whether or not we are zombies (Shoemaker 1975: pp291-315). Qualia, for humans, seems to be accessible to introspection and therefore seems to play a role in our formation of beliefs and subsequent behaviour. If zombies have the same mechanisms as we do, then given that they have no qualia, they should behave in a different way. But according to the argument, zombies behave in exactly the same way as humans. This is a contradiction, and as such, highlights a very important problem with the logical possibility of zombies. Chalmers suggests that a response to this problem would be to argue that causal theories of knowledge and belief acquisition are not appropriate for accounting for our knowledge of consciousness (Chalmers 1996: p193). By taking a dualist stance, we could claim that knowledge of conscious experience is fundamentally different to other forms of knowledge, and that our knowledge of conscious experience is not brought about by any sort of causal relationship between knowledge and experience. Our knowledge of conscious experience consists in another sort of relationship. But Chalmers does not offer this alternative.

"... there is good reason to believe that the epistemology and semantics of experience cannot be essentially causal, and should instead be understood in other terms. ... A full understanding of these issues would require a lengthy separate investigation ..." (Chalmers 1996: p209)

So the mystery remains. Chalmers has replied to Shoemaker's objection by claiming that causal theories cannot account for our knowledge of experience, but he offers us no alternative theory. He merely suggests that we might go about accounting for knowledge by appealing to a dualist point of view. At this stage I see no reason to reject the causal accounts of knowledge acquisition. Shoemaker's objection should therefore be taken very seriously.

Despite the problems associated with the conceivability of zombies, philosophers such as Chalmers maintain that the concept of zombies is perfectly coherent and leads to no logical contradiction. But why do people think it is so easy to imagine zombies? Allin Cottrell suggests that the attraction in thinking of zombies as conceivable lies in our view of human cognition as an information processing system (Cottrell 1996). Many contemporary theories of mind make the claim that the brain is an information processing system and is in essence some sort of biological computer. These theories assume that our mental states should be understood as functional states, which are to be understood as connectionist or symbolic computational states. If the brain is to be thought of as a computer, it is very easy to imagine the brain carrying out its computational tasks without any qualia or subjective experience, just as it is natural for us to assume that computers have no internal experiences. In fact, it would be difficult to conceive of a present day computer as having subjective experiences and enjoying the raw feels we experience.

Computational theories of mind may offer some insight into why some believe it is easy to imagine zombies, but I think that imagining zombies in this fashion falls short of capturing the complexity of the human brain. It is probably true that the brain is nothing more than a sophisticated information processing device, but appealing to the computational nature of human cognition should not make it any easier to imagine zombies. We assume that computers have no qualia, but zombies are more sophisticated than electronic computers. Zombie brains are identical to humans brains in every respect. Human brains do not work in the same way as the other computers we come into contact with. Imagining desktop PC's as being devoid of qualia does not help us imagine creatures with human brains who lack qualia because they are totally different types of systems. On the otherhand, imagine that a computer is constructed that works in exactly the same fashion as a human brain. If this system was designed as a copy of an existing biological brain and was installed in an android body, it would be hard to imagine that it did not have qualia because like a zombie, it would have the same psychology and behaviour as us. If it could be shown that such a device was devoid of qualia, we would then have to discover how to add qualia to the system that was already complete. This would lead us to an epiphenomenal position, which implies that it would be possible to remove the qualia from a human subject while keeping all of her physical and psychological functions in place.

Given the difficulties in adequately imagining the possibility of zombies it would seem that zombie supporters cannot be justified in stating that zombies are logically possible. If, however, a zombie supporter could find an existing case of zombiehood, the case for zombies would become convincing. As we have seen there would be difficulties in classifying an entity as a zombie because we only have access to its verbal claims about consciousness and these claims would lead us to believe that it was conscious in the full human sense. But all is not lost for the zombie supporter yet. Examples have been found of people who could be considered to be partial zombies. These people are missing the subjective conscious experience associated with a particular aspect of cognition and can verbally report upon its absence. In the literature, the most cited example of this type of deficit involves the curious phenomenon known as `blindsight'.

Lesions to the visual cortex can cause a neurological condition known as a scotoma (a loss of vision in the corresponding area of the visual field). People who have this condition retain most of their normal vision, but are aware of a blind spot in their visual field. The interesting feature of this disorder is that in some cases, there is evidence that some visual processing is going on within the blind area of the visual field. Experiments have been carried out in which patients have had a stimulus (a flash of light) presented to their blind region and were asked if they had seen anything. Obviously, because they are blind in that area, the patients reported that they were unaware of the flash of light. However, when forced to take a guess and point to where the flash occurred, many subjects exhibit surprising accuracy. Somehow these people are detecting stimuli that are presented to their visual field. These experiments have been carried out many times and there is now little doubt that these subjects are performing much better than chance. It seems that in these cases, the visual system is carrying out some of its normal information processing tasks without the patient's conscious awareness. There is vision without qualia. This phenomenon is known as blindsight and has been used, by many, as a natural example of partial zombiehood (Block 1995: p385). The fact that blindsight can be used to show that it is possible for people to see things without the phenomenal feel associated with sight makes the phenomenon very attractive to many zombie supporters. Blindsight cases certainly seem to show that (in the case of vision at least) consciousness is not an essential part of the cognitive process that goes on within the brain. However, despite its intuitive appeal, the phenomenon of blindsight has not convinced everyone that partial zombiehood is possible. The first problem with using blindsight as an example of partial zombiehood should be obvious. One of the zombie argument's main premises states that zombies should be physically and functionally identical to human beings. But blindsighted patients have suffered physical damage to part of their cortex and their brains are no longer physically identical to their `predamaged' brains. Using the impaired visual processing systems of a blindsight patient to count as an example of partial zombiehood constitutes a failure in recognising the implications of such an exercise. One may think that one has found an example of partial zombiehood, but this would be failing to take into account the difference between a blindsighted person and a fully functional human. This in itself does not refute Chalmers' claim that experience is not logically supervenient on the physical. It does, however, show that experience is naturally supervenient on cortical activity.

Gerald Vision (1998: p137) claims that arguments that use blindsight as evidence of vision without phenomenology usually rely on misconceptions about blindsight. Daniel Dennett (1993: p331) makes similar claims. He suggests that philosophers who are reliant on blindsight as an example of visual perception without consciousness are overlooking certain basic facts about the phenomena. In some cases, it could be possible that the blindsight subjects are receiving unintended hints from the experimenter and this could explain the fact that sometimes subjects do better in the experiments than other times. Of course, in an experimental setting, these things are carefully controlled, but we can never fully dismiss the possibility of unknown cues from the experimenter.

Dennett has suggested that the blindsight phenomenon does not show that blindsighted people lack visual consciousness. For Dennett, blindsighted people have an impoverished visual content within their blind region, and as such, information contained in that region has a very low degree of influence on behaviour (Dennett 1995b: p417). This would explain why blindsighted patients do not spontaneously offer reports about the content of their blind spot. The visual content is not strong enough to have that type of influence on their behaviour. They have to be cued from an external source. Dennett claims, however, that it would be possible for a blindsighted patient to be trained to guess correctly when to make a guess about the contents of her blindspot. If she still performed above chance in the experiments, she would qualify as being visually conscious. Such a person could treat her visual stimuli in the same way as we treat conscious experience. She could think about and report on the stimuli within her visual field and for all intents and purposes be conscious in the sense that consciousness consists in reportability or an availability for deliberation. Presumably in this scenario, the content of the blind region would have a strong enough influence on behaviour to allow the blindsighted patient to respond to cues from other centres in the brain. But would this count as visual consciousness? What about the phenomenal feel of visual experience? Gerald Vision has suggested that blindsighted patients would not be able to become visually conscious as Dennett suggests. Blindsight patients have lesions that block crucial inputs to the ventral system, which is involved in object recognition. It is unlikely that the neural pathways supporting visual recognition could be regrown by simply training a patient to recognise when to guess (Vision 1998: p153). This point is obvious. Training a patient to guess when to guess about the content of their blind region would not be sufficient to repair damage to neural pathways. If a patient was successfully trained to recognise when to guess, visual consciousness would still be missing. Such training would probably involve recognising non-visual cues such as muscle readiness.

Vision claims that arguments from blindsight usually involve misconceptions about the phenomenon. It is often assumed that blindsight patients have some sort of visual beliefs about the features of the objects within their blindspot and that these beliefs can be discovered through verbal reports of blindsight patients. This assumption is misguided, however, because evidence has shown that the success rate of verbal responses to stimuli presented to the blind region is only slightly above chance. In experiments carried out by Zihl and Von Cramon, blindsight subjects had a light flashed in their blind regions and were asked to respond in three different ways: by pressing a button, blinking or saying "yes". While the success rate for pressing the button or blinking was well above chance, the subjects performed significantly worse with the verbal response (Vision 1998: p146). Blindsight patients can reach and grasp at objects and can even rotate their hand in relation to the orientation of an object, however they usually cannot issue any verbal report about the object. In a simplified account of the visual process, Gerald Vision draws a distinction between two functionally discrete streams of visual processing - the ventral stream and the dorsal stream. The ventral stream is responsible for computing object features such as colour, shape and motion, while the dorsal stream processes information with regards to location. The dorsal stream is used for immediate responses to stimuli such as grasping for food or avoiding obstacles while the ventral stream is responsible for object recognition and classification. This distinction may provide an explanation of why blindsight patients retain visuomotor skills, such as reaching, while higher level aspects of vision that may be responsible for verbal reportability are lacking. Vision suggests that in cases where blindsight patients manage to achieve higher than chance success in verbal response, patients may be relying on clues from premotor readiness in their muscles (Vision 1998: p151). This would imply that the verbal reports of blindsight patients relate only to muscle readiness rather than what is being visually perceived. For Vision, blindsight is not the result of dissociations within a single unified visual system, its occurrence can be explained more adequately by considering the interactions between distinct self contained systems.

An understanding of blindsight shows us that the phenomenon cannot easily be used as an example of partial zombiehood. Zombie arguments are supposed to convince us that normal behaviour is possible without qualia and that consciousness is not logically supervenient on the physical. Blindsight, on the otherhand, shows us that physical changes to the brain can affect certain aspects of vision and that in humans, the visual process probably consists of interactions between several separate systems, some of which continue to function normally after damage has occurred to the brain. The point is that blindsight, while an interesting phenomenon, does not provide an example of physical and behavioural sameness without qualia. It is true that in blindsight, the visual qualia are missing, but this is the result of physical and psychological changes within the brain and so the blindsight phenomenon cannot be used to argue for the logical non-supervenience of qualia on the physical. If anything, blindsight shows that the physical state of the brain is an essential condition for visual qualia. This, however, is not sufficient to show that the qualia themselves are physical, it merely shows us that there is a causal link between physical processes and mental states.

There is an example, which we may all be familiar with, that can serve as an instance in which consciousness seems to be missing for short periods of normal waking time. Imagine that you are a long distance truck driver. After driving for long periods of time, it is quite common to suddenly become aware of the fact that for some time you have been driving without being aware of what you have been doing (Armstrong 1981: p723). You may have been on a long trip and suddenly realise that you have no recollection of events that have taken place during the last part of the trip. It is a sobering experience that most drivers have had. It even happens on short trips. You find yourself arriving home without any memory of your actions during the trip. But during the trip, you were engaged in high level control operations including motor control and decision making. You were negotiating traffic signals, avoiding cars, and changing gears, but lacked consciousness of these actions and events. Would it be correct to say that this is an example of part-time zombiehood? David Armstrong believes that the driver in this scenario had minimal consciousness. This is to say that the truck driver was perceptually conscious of the road, and that there was mental activity involved in guiding the driver's actions. For Armstrong, what the driver lacked was introspective awareness - an awareness of the activities going on in her mind during the trip. In this sense, the driver's actions were driven by a primitive level of mental functioning. This mental functioning could be considered to be `unconscious' (Armstrong 1981: p724). Fred Dretske disagrees with this claim and believes that the only sense in which the driver's mental states were unconscious, is that the driver was not conscious of having those states. This does not mean that the states themselves were unconscious (Dretske 1993: p778). Dretske makes a distinction between consciousness of objects and consciousness of facts. For Dretske, the driver could have had conscious experiences of objects - for example green traffic lights - without having any consciousness of the facts about those objects. We can have a conscious experience of a green light without having a conscious belief that it is a green light. I think Dretske is correct and this shows that the case of the absent minded truck driver cannot be used as an example of zombiehood.

Another way of explaining the case of the absent minded truck driver involves memory. It could be the case that the experiences of the truck driver were not encoded in long term memory, and that is why she cannot introspect and make claims about the long journey. It may seem to the truck driver that `there was nothing it was like' to be on the long drive, but that thought occurs after the event. If we were to probe the driver at any point during the trip, we would probably get reports of a rich and full conscious experience. The point is that if the truck driver cannot remember conscious experiences, it does not follow that they did not exist. I cannot remember any of my conscious experiences from Wednesday the 1st of October 1997, but that does not mean that I was a zombie on that day. Zombies are defined as being identical to humans in every respect, except that they lack qualia. In the case of the absent minded truck driver it is far from clear that the driver lacked qualia. I think the alternative ways of describing what happens in this case are more convincing than the claims that the driver had no conscious experience of her trip.

The possibility of zombies has been asserted by philosophers in an attempt to show that consciousness is not logically supervenient on the physical structure of the brain. Zombies are physically, functionally, behaviourally and psychologically identical to humans and yet they have no qualia. In order to construct arguments that appeal to the possibility of zombies, the zombie supporter must first show that zombies are logically possible. For David Chalmers, that simply involves imagining a possible world in which zombies exist. He claims that the burden of proof lies with those who wish to deny the logical possibility of zombies. But, we have seen that imagining zombies is not as easy as Chalmers has lead us to believe. Our imaginations inevitably fail at the task. We cannot imagine a zombie. We can only make stipulations about what features a zombie might have (or lack). Furthermore, Shoemaker's argument has shown us that the definition of a zombie leads us into a contradiction. Qualia, for humans, are accessible to introspection and therefore seem to play a role in our behaviour. If zombies have the same introspective mechanisms as we do, then given that they have no qualia, they should behave in a different way. But according to the argument, zombies behave in exactly the same way as humans.

An option for the zombie supporter might have been to find a case of zombiehood in the real world. If this could be done, the zombie arguments would still work. The problem is that there does not seem to be a clear cut case of zombiehood in our world. For both the case of blindsight and the absent minded truck driver, we could easily find alternative explanations for the phenomena without resorting to the conclusion that they provide us with true examples of zombiehood.

With no existing examples of zombiehood, the zombie supporter must appeal solely to conceivability arguments. But after highlighting the problems with imagining zombies, it would seem that the zombie supporter has a lot of work to do.

I cannot show that consciousness is logically supervenient on the physical, though I think it is. I have, however, provided reasons for which arguments involving zombies cannot show that consciousness is not logically supervenient on the physical. The zombie arguments remain unconvincing and do not offer any serious challenge to physicalist descriptions of the mind.

Armstrong, D. (1981), "What is Consciousness" in The Nature of Consciousness, Edited by Block. N., Flanagan. O., and Guzeldere. G., MIT Press, A Bradford Book, 1997.

Block, N. (1995), "On a Confusion about a Function of Consciousness" in The Nature of Consciousness, Edited by Block. N., Flanagan. O., and Guzeldere. G., MIT Press, A Bradford Book, 1997.

Chalmers, D. (1996), The Conscious Mind, Oxford University Press, 1996.

Cottrell, A. (1996), "On the conceivability of zombies: Chalmers v. Dennett", http://ling.ucsc.edu/~chalmers/zombies.html, 1996, (Downloaded 1/7/98).

Dennett, D. (1988), "Quining Qualia" in The Nature of Consciousness, Edited by Block. N., Flanagan. O., and Guzeldere. G., MIT Press, A Bradford Book, 1997.

Dennett, D. (1993), Consciousness Explained, Penguin Books, 1993.

Dennett, D. (1995a), "The Unimagined Preposterousness of Zombies" in Brainchildren, Penguin Books, 1998.

Dennett, D. (1995b), "The Path Not Taken" in The Nature of Consciousness, Edited by Block. N., Flanagan. O., and Guzeldere. G., MIT Press, A Bradford Book, 1997.

Dretske, F. (1993), "Conscious Experience" in The Nature of Consciousness, Edited by Block. N., Flanagan. O., and Guzeldere. G., MIT Press, A Bradford Book, 1997.

Guzeldere, G. (1997), "Introduction: The Many Faces of Consciousness: A Field Guide" in The Nature of Consciousness, Edited by Block. N., Flanagan. O., and Guzeldere. G., MIT Press, A Bradford Book, 1997.

Horgan, T. (1987), "Supervenient Qualia" in Philosophical Review Vol xcvi, no.4, October 1987, Edited by Wood, A., Irwin, T., Appiah, A., Cornell University.

Levine, J. (1993), "On Leaving Out What It's Like" in The Nature of Consciousness, Edited by Block. N., Flanagan. O., and Guzeldere. G., MIT Press, A Bradford Book, 1997.

Rey, G. (1997), Contemporary Philosophy of Mind, Blackwell Publishers, 1997.

Shoemaker, S. (1975), "Functionalism and Qualia" in Philosophical Studies 27, 1975.

Vision, G. (1998), "Blindsight and Philosophy" in Philosophical Psychology Vol 2, Number 2, June 1998.

This work is protected under international Copyright law. This work (or portions of this work) may be downloaded, distributed, quoted, or referred to for the purpose of criticism, evaluation, or for furthering knowledge and research in this field. When referring to the work of others, please remember to cite the author's name, the title of the work, the place and date of publication (or internet address).

For Block, the word "consciousness" refers to many different concepts and phenomena that have been bundled together under the one concept. Block suggests that we run into problems when we analyse certain aspects of consciousness using premises that cannot be applied to other aspects of consciousness. In an effort to clear up the confusion associated with reasoning about consciousness, Block breaks consciousness down into several different concepts. In this paper I will be concerned only with what Block calls access consciousness and phenomenal consciousness. These two concepts appear to constitute his primary distinction and deserve attention. I will consider David Chalmers' contribution to the issue and will then outline an alternative view offered by Daniel Dennett.

Ned Block draws a distinction between two different types of consciousness - phenomenal and access. This distinction arises from the thought that the phenomenal properties of consciousness are of a different character to the cognitive, intentional or functional properties of consciousness. For Block, the phenomenal properties of consciousness are experiential properties. These properties are categorized as being properties of phenomenal consciousness (P-conscious properties). P-conscious states include the experiential states we have when we see, hear and have pains.

Block believes that A-Consciousness and P-Consciousness usually occur together but in some cases they may not.

In order to help us acquire a full understanding of the difference between P-consciousness and A-consciousness, Block provides some examples of A-consciousness without P-consciousness and of P-consciousness without A-consciousness. These examples are intended to clear up any confusion we may have between these two distinct categories of consciousness.

The ventral system is responsible for object recognition and classification, while the dorsal system is involved in computing spatial features such as location and motion. Block believes that because the visual system is comprised of these two visual subsystems, it would also be conceptually possible to find cases of P-Consciousness without A-Consciousness. This might occur if someone incurred damage to their dorsal system, while their ventral system remained intact. Of course, if Block's distinction is accurate, we would probably not know if someone was P-Conscious of events in their visual field without being A-Conscious of those events because a lack of A-Consciousness implies that content is not poised for the control of behavior. This includes behavior such as making the statement: "I see a red object."

There are other possibilities of P-Consciousness without A-Consciousness because, obviously, the P-Conscious / A-Conscious distinction does not only apply to visual events. The distinction applies to all events involved in behavior and awareness. As an example of P-Consciousness without A-Consciousness, Block asks us to imagine a situation that involves the auditory system. Suppose that you are involved in a conversation with someone when suddenly you notice the existence of a constant noise that has been occurring throughout the entire conversation. Perhaps you suddenly notice the steady ticking of a clock. The sound has been there all along and you were aware of it all along but you were not consciously aware of it. According to Block, you were P-Conscious of the noise, but you were not A-Conscious of it. In other words, even though you were P-Conscious of the sound of the clock, that information was not poised for the direct rational control of action until you noticed it. It was at that point that the noise of the clock had an influence on your behavior and thoughts. I think that if, at this point in time, you thought back over the past few minutes, you might realize that you were aware of the ticking sound all along. The important point is that it took A-Conscious awareness of the sound to shift your attention to the sound of the clock and to enable you to even consider that it had been there all along.

On the surface, Block's distinction between access and phenomenal consciousness looks like a useful way of explaining the problem of consciousness. It allows us to seek cognitivist explanations for behavior without having to find a way of including the phenomenal properties of experience in those explanations. Perhaps once we have gained a complete understanding of how access consciousness works, we could turn our attention to phenomenal consciousness. I think, however, that in making the distinction between A-Consciousness and P-Consciousness, Block has left us with a more difficult problem. What purpose does P-Consciousness actually serve? In providing us with the conceptual possibility of A-Consciousness without P-Consciousness (the super-blindsight patient), Block has eliminated the need for P-Consciousness. The reason for this will become clear in a moment when I look at David Chalmers view of the distinction.

Chalmers claims that a clear conceptual distinction can be made between access and phenomenal consciousness when one considers the fact that we can imagine P-Consciousness without A-Consciousness and A-Consciousness without P-Consciousness, and the fact that A-Consciousness can be accounted for by cognitivist explanations while P-Consciousness is resistant to such explanations. Unlike Block, however, Chalmers believes that A-Consciousness and P-Consciousness always occur together. Chalmers also offers an alternative way of describing A-Consciousness by playing down the role of rationality. Block had defined content as being A-Conscious if it was poised for the direct rational control of action. For Chalmers, it is enough to say that content is A-Consciousness if it is directly available for use in directing behaviors. On Chalmers account, the case of the background sound of a ticking clock can be described in a slightly different way. We could say that the information was available all along but it was not accessed. If we accept this view, it would seem that there was P-Consciousness and A-Consciousness of the clock noise throughout the entire event. The phenomenal aspect of the noise was always present, and at the same time the information was available for the direction of behavior; it was just not accessed.

Chalmers points out that the problem with making the distinction between A-Consciousness and P-Consciousness (his modified distinction) is that we are left with the question as to why the two always seem occur together. It would seem that there is no role for P-Consciousness to play in the collective cognitive economy. If Chalmers is right, A-Consciousness is all that is required for the control of behavior in an organism. This leaves open the conceptual possibility of zombies and other functional isomorphs who are identical to us in all respects except that they lack P-Consciousness. Chalmers concludes that P-Consciousness has no role in cognitive functioning and that A-Consciousness does all the work. It is interesting to note that while Chalmers has pointed to this problem, it does not seem to bother him much. In fact, he seems to embrace it.

The problem with this statement is that it implies that the phenomenal aspects of our experience cannot be accounted for by physicalist explanations of the mind. Pursuing this line of thought would lead us to an epiphenomenalist position and would leave the question of phenomenal consciousness largely unanswered. Phenomenal consciousness would be described as having no causal function on our mental lives and would seem to be nothing more than a `bonus' feature of experience. Before we accept this view of phenomenal consciousness, I think we should explore other options.

In response to Block's paper, Dennett offers a different solution to the confusion about the role of consciousness. In his paper "The Path Not Taken", Dennett agrees that Block is right to locate the source of the confusion in the apparent difference between phenomenal and access consciousness. However, Dennett believes that Block runs into trouble when he attempts to defend his views.

As I described above, there is evidence to show that the dorsal subsystem is responsible for computing basic features of objects, while the ventral system is involved in higher level activities such as object recognition. Block believes that the ventral system is closely connected to P-Consciousness, while the dorsal system is related to A-Consciousness. In blindsight there has been damage to the ventral system. Thus there is A-Consciousness without P-Consciousness. This is a worthwhile observation, and the accumulating evidence of these two subsystems must be taken into account. However, the existence of two distinct visual subsystems does not necessarily refute Dennett's position. Dennett might argue that these two subsystems offer nothing more than evidence of modularity in a unified visual system. Damage to the ventral system merely reduces content and thus reduces the degree of influence on behavior. If the patient had instead only suffered damage to the dorsal system, other behavioral effects might occur. Content would be impoverished and so the degree of influence on behavior would also be reduced. Perhaps the patient may be able to describe the shape of an object but not be able to offer any report on the object's motion.

If this was the case then we would have to say that the sound of the clock was guiding behavior.

The main reason for avoiding the distinction that Block makes between access consciousness and phenomenal consciousness is that it only gives us the scope to explain how access consciousness works. This is because access consciousness can be isolated as being a cognitive, or computational type of system. Phenomenal consciousness, on the other hand, seems to be resistant to our explanatory techniques. If Chalmers is correct, then phenomenal consciousness may have no causal role to play in the cognitivist's story. I think that before we simply accept that conclusion, we must consider other possibilities. Dennett's alternate view seems to me to be a good candidate. As we have seen, we can explain blindsight and other phenomena in a direct way without appealing to different types of consciousness. If we can avoid making the distinction between access and phenomenal consciousness, we may be able to eventually come up with a complete account of consciousness that does not leave the apparent existence of phenomenal experience unexplained.

1 Block. N., "On a Confusion about a Function of Consciousness" in The Nature of Consciousness, Edited by Block. N., Flanagan. O., and Guzeldere. G., MIT Press, 1997, Page 376.

2 Block. N., "On a Confusion about a Function of Consciousness" in The Nature of Consciousness, Edited by Block. N., Flanagan. O., and Guzeldere. G., MIT Press, 1997, Page 384.

3 Block. N., "On a Confusion about a Function of Consciousness" in The Nature of Consciousness, Edited by Block. N., Flanagan. O., and Guzeldere. G., MIT Press, 1997, Page 386.

4 Chalmers. D., "Availability: The Cognitive Basis of Experience" in The Nature of Consciousness, Edited by Block. N., Flanagan. O., and Guzeldere. G., MIT Press, 1997, Page 421.

5 Chalmers. D., "Availability: The Cognitive Basis of Experience" in The Nature of Consciousness, Edited by Block. N., Flanagan. O., and Guzeldere. G., MIT Press, 1997, Page 423.

6 Dennett. D., "The Path Not Taken" in The Nature of Consciousness, Edited by Block. N., Flanagan. O., and Guzeldere. G., MIT Press, 1997, Page 417.

7 Block. N., "On a Confusion about a Function of Consciousness" in The Nature of Consciousness, Edited by Block. N., Flanagan. O., and Guzeldere. G., MIT Press, 1997, Page 386.