Sunday, October 01, 2000

Virtual Unreality: Will the Web Become Our Collective Mind?

I of the Vortex

By: Rodolfo Llinás M.D.

In I of the Vortex, to be published by MIT Press next year, a pioneering neuroscientist sets forth his conception of the mind “from neurons to self.” In this excerpt from the final chapter, Llinás looks at the future of the brain and the role of the World Wide Web. Someday there may be thinking and feeling machines, he speculates, but they can never duplicate the extraordinary self-aware human brain.


In the concluding pages of I of the Vortex, Rodolfo Llinás warns: “The event we should fear most is the possibility that as we develop better forms of communication with one another, we may cease to desire interaction with the external world…Keep in mind that the only reality that exists for us is already a virtual one—we are dreaming machines by nature!”

The prophecy of a science fiction writer prone to drama?

Far from it. Rodolfo Llinás is one of our era’s foremost neuroscientists, a pioneer in areas from how we initiate and control movement to how electrical oscillation in the brain may mediate our sensorimotor system and ultimately thought itself. A report by Llinás and his colleagues at the annual meeting of the Society for Neuroscience in 1999 made international headlines with a report on how timing in the thalamus may underlie disorders such as Parkinson’s disease.

In I of the Vortex, scheduled for publication by MIT Press next year, this distinguished scientist for the first time sets forth his grand conception of the mind “from neurons to self.” This issue of Cerebrum presents an excerpt from the final chapter, “The Collective Mind?” in which Llinás looks at the future of the brain and the role of the World Wide Web. “The Web’s presence,” he writes, “has already profoundly reshaped the most developed societies, and will continue to do so in ways that are difficult to imagine now.”

It is difficult to summarize in a few paragraphs the view of the brain and the rigorous, painstaking argumentation that set the stage for Llinás’s final chapter. For Llinás, brain and mind are inseparable events, although mind is but one of several states generated by the brain. Mind comprises all brain states in which sensorimotor images, including self-awareness, are generated. Reduced to essentials, mind evolved because living organisms act, and to act successfully must predict their environment. Indeed, we must predict just to walk across the room without bumping into furniture.

Much of the I of the Vortex is devoted to a systematic, detailed explication of how neurons in the central nervous system organize and drive bodily movement, create sensorimotor images used in prediction, and generate thoughts. Llinás’s research on the nature of movement, including pioneering studies of the intrinsic oscillatory electrical properties of the brain, are at the heart of his argument. Llinás explains the brain’s essential operations in terms of the simultaneity of this electrical activity in our neurons and networks of neurons— “the electrical glue that allows the brain to organize itself.”

What, then, is our self? “Self is the centralization of prediction.”

The great problem for the brain in controlling motion is the potentially infinite number of possible movement combinations. In a cogent, wry analysis, Llinás argues that if the brain, even with its staggering data capacity, had to consider every possible variation on reaching for milk in the refrigerator we would have too few neurons and too little time. Evolution’s answer is the “fixed action pattern,” or FAP, a acronym readers will encounter in the except below. FAPs are “sets of well-defined motor patterns, ready made “motor tapes” as it were, that when switched on produce well-defined and coordinated movements [bird songs are one fascinating example, but so is walking or swallowing]...”

Using the concept of FAPs, Llinás explains emotions and language, which are FAPs where “the actions are not motor but premotor.” As useful as fixed patterns are in saving time and energy, the nervous system also requires flexibility. An organism adapting to a changing world must be able to modify the range of an FAP. Enter learning and memory.

The sweep of integration in I of the Vortex is one readers must experience for themselves. “Language, and in particular human language, arose as an extension of premotor conditions...” write Llinás, making possible “the sophistication of purposeful movement through modifying and overriding the existing FAPs.”

To the living creature’s problem of acting in a changing world, some 700 million years of evolution—trial and error endlessly at work—have yielded a solution of incalculable complexity and beauty: the mind. But that story, suggests Llinás, “like all things biological, is still being written.” What is next?


From I of the Vortex by Rodolfo Llinás. Reprinted by permission of The MIT Press.

 [A] squirrel may show another squirrel or animal that violent jerking and wriggling may help it escape a foe, [but] it cannot tell another that this is so. With all due respect to squirrels, the message could be conveyed more quickly with the use of spoken language, which can facilitate the communication of internal abstraction(s), both in detail and accuracy.

Spoken language, as opposed to prosodic body or facial gesturing, in many ways extends the range of communications as well because it extends the range of the senses. How? Almost any example will do. A friend stands on my shoulders to look over a very high wall.

“What do you see?” I ask. And he tells me.

Spoken language here clearly allows me to “see” where I cannot.

Or my friend extends his arm over and touches an object in the other side. “What’s in there, what do you feel?” Now I can touch what I could not touch. He sticks his head over the wall; now it is my sense of smell that has been extended in range.

This raises two points. First, while it is possible that what my friend saw, felt, and smelled could be communicated to me by means of prosody—body and facial gesturing— this would most likely suffer in detail and clarity and thus speed, if the desired end result is to transmit the information accurately. This is usually the case with communications by any means, including deception through words or actions (the opposum who plays dead). For deception of any kind to achieve its desired result demands the clear outward expression of the internal abstraction; how truly this abstraction represents external reality is irrelevant. Only the clarity and accuracy with which the intent is conveyed is important. 

We can only yell so loudly and we can only hear from so far away. Thus, there are clear boundaries to this form of communication— and they seem rather constraining. But are they?

A second point regarding the extension of the senses by means of spoken language concerns boundaries. The range of this extension is bounded by the range of the combination of vocal and auditory elements of (this) communication. We can only yell so loudly and we can only hear from so far away. Thus, there are clear boundaries to this form of communication— and they seem rather constraining. But are they?

Let’s say my friend is on a ladder now, looking over the same wall. I am just barely within earshot, and he yells to me what it is he sees. I turn and yell to another person just within earshot, and so on and so forth. This chain of communication can allow someone far away to “see” with my friend’s eyes as he peers over the wall.


There is no doubt this extension in the range of communication extends the range of the senses. There is also no doubt that early man learned this and used it to advantage, by sending messengers on foot and on horseback and by means of semaphores such as flags, smoke signals, and reflective surfaces added between nodes in this chain where word-to-word communication kept it moving. Here was a technique that conveyed information across great distances in segmented or nodal fashion, not at all unlike the conduction of the action potential signal.

Although language increases the range of communication (and thus the theoretical range of the senses), it does so at the loss of both speed and accuracy. If you have noticed the distortion of information through gossip you will know what I mean.

But unlike the unfailing, unchanging action potential, the form of communication described above is limited. Although language increases the range of communication (and thus the theoretical range of the senses), it does so at the loss of both speed and accuracy. If you have noticed the distortion of information through gossip you will know what I mean. By the time the story cycles back it has undergone a noticeable transformation and is distorted to the point of barely resembling the original at all. Although such distortion is often hilarious, in conditions where “pass the word” is important, it is not quite so funny when “one if by land, two if by sea” gets mixed up. The old adage that too many links makes the chain weak is very true in regards to spoken language, with “weak” here meaning deficient in detail, accuracy, and speed. Even if somehow the transmission of information is reliable and unchanging at each node from source to end in all chains of flow (pathways of communication), distortion of the overall signal will occur because of differences in the timing of reception on the part of the receiving elements. For appropriate impact, a given signal (message) almost always needs to reach many destinations rather than just one.

Let us look at information flow from a broader evolutionary context. Just as it took a long time for single cells to become animals, it has taken a long time for humans to evolve into a closely knit society, and the reason or reasons for this are basically the same. In the case of single cells...cell grouping into multicellular animals required communication— meaning—between cells. This took a tremendous amount of time to develop. In even the most primitive cell colonies, the importance of simultaneity in signal reception is clear when we think of the combination of motor elements that must act in synchrony to perform successfully even the simplest of movements, such as the organized FAP of swimming in the lamprey. As the nervous system developed over evolution into its own society of cells, simultaneity of activation as a modular form underlying function was not only conserved, but increased in capability. As the need for more complicated movement increased, synchronous activation of widely differing muscle synergies became necessary. This was accomplished through both a coordinating timing signal from such sites as the inferior olivary nucleus, and the varied speed of conduction of fibers of different lengths assuring simultaneous arrival of the signal at target destinations across a wide range of distances.

At the evolutionary present, the most profound example of natural selection’s conservation, embellishment, and incorporation of simultaneity of activation is the brain’s solution to the problem of perceptual binding and its byproduct, cognition. Chapter 6, devoted solely to this issue, described the thalamocortical system as a close to isochronic (synchronized) sphere of function, binding in time the fractured elements of internal and external reality as represented by the neural activities of spatially disparate regions of the brain. According to this view, simultaneity of activation within this system results in perceptual unity: this book you feel in your hands, this voice that seems to be reading to you, the sense of the chair around you, all seem as one event, occurring now.

Imagine the problems with perceptual truth if simultaneous activation did not occur. Even within one sensory modality there would be trouble. If we could not bind in perceptual time what the tongue feels, the changing pressures the teeth feel, the sense of the roof of the mouth and inner cheeks, we would quickly destroy this multicompetent apparatus so important to us for the ingestion of food, among other things. If the timing of perception of these different tactile sensations were off even a little, the simple act of chewing one’s dinner would result in a bitten tongue and lacerated cheeks.

Without simultaneity of activation the problems are compounded if there is an attempt to orchestrate more than one sense modality. We could never play a musical instrument, because what we hear and what we feel in our fingers would never match. We would be unable to enunciate words, or ride a bicycle. In short, without coordinated simultaneity of activation, the binding activity of the various sensory systems into perceptual unity would be impossible, and without that the self would be left fragmented. Had evolution not solved the binding problem, we would not be discussing it now. Out of time, out of mind—literally.

Similarly, we can see that early in the formation of human society there was a need to solve the binding problem for information transfer. Messages were distorted by the fact that they were distributed at different speeds among different elements in the society and thus were not received simultaneously by everyone. Things change; what is important one day may not be the next, and so conflicting messages occur. The result is that consensus truth about the global—even local—state of affairs is neither complete nor stable.

Just as the evolution of the brain has solved the perceptual binding problem by its incorporation and use of simultaneity of activation, it is abstraction, a product of intrinsic brain activity, that has tightened the communicative fabric binding society together, in the sense of consensual truth of information. Beginning with communication through pictures and then the written word, abstract thinking has lead to a series of technological advances resulting in successively more accurate, detailed, and today virtually simultaneous communication between individuals separated by great distance. “One small step for a man, one giant leap for mankind” may sound banal, but as a riveting, historic moment it could be experienced by all of us on Earth at the same time, if not at the exact moment of its utterance.

In the decade of the 1990s we have experienced the latest event in this series of communicative advancement: the World Wide Web. In all seriousness, it is fair to say that the Web represents a breakthrough in communication, perhaps second only in importance to the invention of written language itself. The first great advance altered the course of human civilization, and this second one may too. Just in its infancy, the Web’s presence has already profoundly reshaped the most developed societies, and will continue to do so in ways that are difficult to imagine now.


Other than the Web, let us consider what communication systems we have. A television signal can reach millions, as can newspapers, albeit a good deal slower. Neither is interactive. A given message or opinion is stated, we receive it, make our judgments, and there it stops. We may discuss it with friends, but we are not contributing in any real sense to that unidirectional flow of information. We may write to an editor, but this is in all likelihood just pebbles at an elephant, and worse still is that this interaction is painfully slow. Compared to simple conversation at lunch, this is hardly interactive at all.

Telephone and certain forms of radio have the range and speed to allow virtually instantaneous transmission, but bidirectional communication flow quickly turns unidirectional when the number of users increases, even by a few. Any taxi driver will tell you that although one can hear the activity on a particular channel, as the number of users on that channel increases one can hardly get a word in edgewise. The frequency bandwidth is full and that happens, unfortunately, rather easily.

Telephones can connect you with just about anyone, anywhere, with negligible delay. But how many people can you interact with in this fashion? One could, in theory, fill an auditorium and address this crowd through an amplified speaker or conference phone, but if more than even two people respond, the result would be unintelligible noise. And so, again, bidirectional communication flow quickly becomes unidirectional as the number of users increases and interaction is reduced to listening or talking, but not both.

When we ask that communication flow as it does in the brain, the limitations of global information transfer are exposed. But these limitations are disappearing.

Much has changed since the time of Paul Revere, in that issues of detail, accuracy, range (serial range, anyway), and speed of communication no longer pose serious limitations to individual communication needs. The outer boundary of our capabilities becomes exposed only when we insist on bidirectional or interactive information flow that is virtually simultaneous over vast ranges and over vast numbers of sending and receiving parties. When we ask that communication flow as it does in the brain, the limitations of global information transfer are exposed.

But these limitations are disappearing. At least in theory, the Web is a nervous system-like structure in that its functioning seems to be solving, to a certain extent, society’s binding problem.

Already the Web provides communicative simultaneity of activation unlike anything the world has ever seen, allowing for a single person to post a message to thousands, hundreds of thousands, even millions of other people almost simultaneously. Moreover, interaction remains bidirectional at these numbers: any or all of these recipients may send their reply right back, the only delay being the amount of time it takes to frame one’s thoughts and compose the reply. In other words, the delay is now for the most part no longer a technical one.

We see that the flow of information through the Web is similar to, and perhaps best analogized by, the flow of information between and among neurons, but will the Web demonstrate some sort of intrinsic embedding as well? If so, what will be embedded? Repetitious patterns of neural activity become recognized and incorporated into the nervous system’s overall mode of operation (memories, FAPs, and the like), which at all times attempts to increase its computational efficiency while lowering its computational overhead. Intuitively, the increased speed and volume of information flow we see on the Web should feed well into this concept of embedding...but is this analogy real? And if so, will the results be beneficial?

If neurons beget mind, can the people— the minds—that represent each nodal point in the Web generate or become a collective mind? Can the Web support a consciousness of man, and if so, what on earth would this be like? On the surface, the Web and the brain are very different. The brain is alive and the Web is not. Can something nonbiological have a mind?

This last question is neither rhetorical nor limited to discussions concerning the Web, but is of potentially great importance to human society, demanding thorough and concerted effort across many disciplines.

At first glance the workings of the Web do appear to have some common features with the workings of the brain, but even this pseudoanalogy falls apart rather quickly on closer scrutiny. At all times throughout this book, I have stressed a perspective of functional architecture, and viewed from this perspective the Web is awkward at best. In practice, the Web as it is would not be able to support a consciousness of many. For one, it is a very noisy system. Also, although very fast in tasks such as getting a message from here to there, it is not fast enough in its integrative parameters to support consciousness the way the nervous system does (which is still our best if not only reference or standard). The nervous system, you should recall, increases its own efficiency through modularization of function. The Web as it presently exists is not modular. Among all nervous systems it might be most closely analogized to that of a coelenterate—a hydra or jellyfish. And if there is indeed consciousness present in a jellyfish, it is not of a quality that would be capable of supporting, en masse, a collective mind. Ultimately what one needs is a subsystem to collect and another subsystem to distribute, with the simplest of interactions at the node where these two would come together.

The concept of a collective consciousness is not a new one. The outcome of an election is taken as a mandate of the people, representing a collective decision by the populace. For embedding, the benefits of interacting with a greater number of other minds and the experiences that each represents should be obvious, as the nervous system pays particular attention to new stimuli, and it embeds, for the most part, based on repetition. If one person in your life cautions, “Do not play with any black spider that has an hourglass on its belly,” but then goes on to say, “You won’t believe this, but I once saw a flying whale,” you might remember the spider warning at about the time you are being bitten. On the other hand, if you were to hear this spider advice numerous times from friends, parents, teachers, and doctors, all warning you about what could happen to you if you are bitten, you would probably steer clear of these spiders the first time you saw one. What sticks in the mind is the repetition, and the sense that this knowledge evolved from the repetitive swirling of the information between and across other mind before you.

But wait: collective knowledge and collective mind are not one and the same. Although there may be many definitions of collective mind, one we all may agree upon is that the elements comprising the whole combine in such a way that when confronted as a whole, a singular decision about what to do is formed and implemented. The decision that is formed may not and most likely will not be representative of each element’s perspective or opinion, but rather is a consensus that serves to benefit the group overall. This is the same as the sacrifices made and benefits gained when single cells opt to socialize, leading eventually to multicellular organisms. This process culminates in the formation of a collective structure that assumes the role of the decision makers for the animal, namely the nervous system.

If we think seriously about what constitutes a collective mind, the Web is a promising candidate in terms of what it potentially might take to support a consciousness of many. It is certainly arguable that the Web has been created out of man’s desire to create a collective mind.

Is the Web a nervous system composed of nervous systems, a mind composed of minds? Not yet, as I’ve said, not in the classical sense of collective mind. It is communicating, true, but it isn’t thinking. A very similar form of global decision-making process is taking shape nonetheless, which is beginning—and will continue—to affect everyone, for better or worse.


Should we believe that we should eat garbage? Is this logical? Is it true? This is the problem with numbers. The tyranny of the majority has always been an issue, and as most of us know, the document known as the U.S. Constitution was designed in part to protect the citizens from this threat. Nevertheless, if enough people want to have creationism taught alongside evolution in their schools as equally viable and unproven (and implicitly unprovable) theories, then so shall it be. The ability of the mass media to affect public opinion has been demonstrated many times, although, as we have said, mass media is not interactive like the Web or as responsive to one’s ideas as the Web can be. So far, the public has been largely a passive receiver of information. However, for the first time, because of the incredible speed, range, and volume of communication the Web brings, public opinion could truly become public—with its inherent advantages and disadvantages.

Here is where the problem with numbers arises. The Web, precisely because of its speed and volume of information flow, may perpetuate the notion of weighing the value of ideas or beliefs based simply on the number of people reporting that they adhere to them. Not just tyranny of the majority, but of a biased, self-selected one at that! Two hundred thousand people contributing their opinions of an issue on the Web think this or that, ergo it must be true. Ultimately, this inertia of numbers develops a life of its own and determines whether we like or believe something or not, thus giving rise to the self-fulfilling prophecy. And this phenomenon will undoubtedly be accelerated by the machinations of the Web.

As individuals, we know that generalizations based on popular opinion are unreliable. But at the individual level, if you disagree with the inertia you will become an outsider and will therefore suffer the consequences of not being part of the group. Indeed, if anything one says can immediately be criticized by millions of people, it will quickly become very difficult to separate one’s self from the beliefs and feelings of others. Under such pressures, homogenization of thinking will, of necessity, take place. As the Web becomes more intelligent, these machinations will have a strong influence on self-perception, and the very concept of self will become redefined. The concept of ideas as belonging to oneself will be diluted by the fact that any idea given to the Web will either gain acceptance, immediately being commonplace, or will be immediately rejected. This will whittle away our ability to discern individual identity, our possession of ideas—in essence, what forms the backbone of our beliefs in and of self. A homogenization of thought cannot help but occur, and this will, by further feeding the inertia of numbers, cycle against itself in a very implosive fashion.

Homogenization of thought will lead to the homogenization of society, a sobering prospect for the future. When traveling as a youngster, I always loved to see the richness of differences in cultures, beliefs, and viewpoints. Not so much today; for example, you find that children in Asia or Europe all want the same consumer products, in part because they are bombarded by similar images from whatever media reaches them. The trend toward sameness is everywhere apparent, as everything, the good and the trite, is being copied—and in general, the trite is easier to copy than something that takes some thought. We are fast approaching a world culture of sameness not only in the external trappings, but also the character and values of societies. The strength of public media and its influence has made it almost impossible to buck this trend, and there is no reason to believe that the Web will not accelerate this process.

The true downside to homogenization is a decrease in variation, and variation is the key to survival. So the system becomes more brittle simply by the fact that options are reduced if everyone feels exactly the same about any event or given set of values. Against a background of sameness, vulnerabilities are more easily exposed—and more likely. We will look into this a bit later on.

A final point on the generation of a collective mind is that, as in evolution, trial and error must of necessity come into play. It might take as much time as it took nerve cells to make brains for us to learn how to implement such an extended awareness. If properly used this could be an extraordinary development. But as it is right now, the Web needs a serious overhaul in functional architecture to even hope to approach a collective event of the nature we have discussed. 

Is it reasonable to consider the world as being at all like the brain? Yes. What we observe is a similarity of order expressed at different levels, at all levels from cells to animals and from animals to societies.

Is it reasonable to consider the world as being at all like the brain? Yes. What we observe is a similarity of order expressed at different levels, at all levels from cells to animals and from animals to societies. One wonders if this is perhaps a universal law. The way the system organizes itself may reflect, for example, its solution to the tyranny of the second law of thermodynamics, “order will decrease with time.” There may be a deeper message here. One of the few ways in which local order can increase is through the generation of such things as a nervous system that employs modularization of function. If modularization is indeed a universal to combat disorder, such a geometric and architectural solution may have happened at other levels as well. Chances are high that the weak anthropomorphic principle, namely that we are here because the universal laws make it likely to the point of inevitability, is the underlying universal tendency, rather than the other way around (the strong anthropomorphic principle)—that a predetermined event in the distant past formed the universe in the way it did, so that we could “become.”


The spawn of the technology behind the Web presents an ominous event if not properly modulated. If allowed to expand out of all control, it could become a danger, perhaps the most serious threat that society has ever encountered, eclipsing that of war, disease, famine, or drug problems. The event we should fear most is the possibility that as we develop better forms of communication with one another, we may cease to desire interaction with the external world. If one considers the problems for society of mind-altering drugs, then imagine if people could realize their dreams, any dreams, by means of virtual communication with other real or imaginary human beings. And not just via the visual system, but through all sensory systems. Keep in mind that the only reality that exists for us is already a virtual one—we are dreaming machines by nature! And so virtual reality can only feed on itself, with the risk that we can very easily bring about our own destruction.

If you consider how many hours a day people now watch TV, the amount of time that will be spent in virtual worlds can only be more because it is not just watching but interacting. You can play the music you are hearing. You can fly a plane, hunt an elephant, experience intimate sexual contact, virtually. Whatever you wish. The possibility to disrupt society is virtually boundless. It could be the ultimate intellectual dependency, because the true boundaries that reality defines would disappear. The hard facts of life could be seriously questioned. Here is the possibility of creating a totally hedonistic state, a decadent sybaritic society rushing headlong into self-destruction and oblivion. We all know that pleasure must be titrated; it must not be inhaled too deeply. Pleasure is not an end, rather the means to an end. If we are approaching some form of collective consciousness, it may be a dangerously narcissistic one, one that could precipitate the unraveling of a society already undermined by the ominously anti-intellectual climate in which we live.

Brain research has known this for years. Place a stimulating electrode within a rat’s medial forebrain bundle, the pleasure center of the brain. Now allow the rat free control to activate this area by pressing an electrically connected lever, and the rat will forego all food, sleep, and water to stay in a state of constant bliss. And it will stay there until it is dead. Humans will lethally titrate cocaine intake in the same way. Virtual reality will be a lever in this regard, possibly greater and more powerfully addicting than any we have seen yet. Life is not a dream; it is about physical survival and continuation. Virtually reality will not fill the need.

Hopefully, the wisdom of human nature will ultimately recognize this virtual realm as nonreal: that somehow the evolutionary event would by some quirk have known that something like this was possible. That somehow our brains evolved not to allow us to act out our REM-driven dreams lest we hurt ourselves. More realistically, one can hope that evolution might resolve the problem much as it resolves great natural catastrophes, through variation and selection. A small subset of people may be found who say, “Don’t give me any of that two-dimensional sex, I want the real thing.” The culling of society through natural selection might produce a different, more thoughtful human being. It may be all that we can hope for.


Whether the Web is alive in a biological sense or not is probably irrelevant. If we consider each opinion, belief, or message from an individual as a stimulus, then the Web acts much as consciousness does: making quick, yea or nay consensus decisions about incoming stimuli and generating a solution; there simply is no time for anything else.

Discussions of this nature suggest an obvious, ultimate question: is mindness a property that can reside only within the realm of the biological, of living flesh and blood? 

Discussions of this nature suggest an obvious, ultimate question: is mindness a property that can reside only within the realm of the biological, of living flesh and blood?

Let us think for a moment about the case of flight. If it were the thirteenth or fourteenth century, we might conclude that flight is a property of biology, perhaps exclusively so from the fact that the only objects that are heavier than air and that can fly are living creatures. By contrast, every person living at the end of the twentieth century now knows that flying is not a property exclusive to biology. Similarly, one may wonder if mindness is exclusively a biological property. Computers as we know them today do not seem ready to have a mind, but that may be due more to limitations in our choice of design architecture than to any theoretical constraint on artificially created mindness.

In the case of flight, specialized skin, cuticle tissue, and feathers have all proven their worth as materials in the composite that conquers gravity—as have plastic, dead wood, and various metals. It is not just the materials, but the design that defines feasibility here.

So is “mind” a property of biology alone, or is it actually a physical property that may in theory be supported by some nonbiological architecture? Put another way, is there any serious reason to believe that biology is separate from physics? The scientific knowledge gathered over the last hundred years or so suggests that biology, in all its amazing complexity, is no different from anything else that obeys the laws of physics. Thus it should be possible for consciousness to be implemented by a physical organism, which in our case happens to be what we call a biological system.

The question that people generally ask is somewhat different: whether devices of an other than biological nature are capable of supporting consciousness, qualia, memory, and awareness, that which we consider to be the serious properties of nervous system function. That is, would a computer ever really be able to think?

The easy answer is yes, we think they can and will, but the more relevant question is, what would the physical system have to be like or look like before it can do the same as the brain? Or, perhaps, as some still feel, is there something spooky or otherwise indefinable, not knowable in brain, what in philosophy has been called the “hard problem”? It seems to me that the issue is most likely one of physical degrees of freedom of functional architecture, rather than the aliveness of biology versus the deadness of physics.

Having been a vertebrate physiologist all my life, with some forays into the invertebrate world, I have presented in this book an image of consciousness that is embodied by a particular type of neural network or circuit. But I must tell you one of the most alarming experiences I’ve had in pondering brain function. This was the realization, from discussions with Roger Hanlin at the Marine Biological Laboratory at Woods Hole, that the octopus is capable of truly extraordinary feats of intelligence. I have read of experiments in octopus by J. Z. Young, where these invertebrates have solved problems as complicated as opening a jar to remove a crab kept inside. Operating with nothing but the visual image of the crab inside and the tactile manipulation of the jar, the creature finally found that the top could be opened by applying force. And after having done so, when presented again with the same problem, the animal was immediately capable of opening the top and fetching the crab out. Astoundingly, this event could be learned with a single trial. More to the point, however, and most remarkable is the report that octopuses may learn from observing other octopuses at work. The alarming fact here is that the organization of the nervous system of this animal is totally different from the organization we have learned is capable of supporting this type of activity in the vertebrate brain. If we are faced with the sobering fact that there are two possible solutions to the “intelligence” problem, then there may well be a large number of possible architectures that could provide the basis of what we consider necessary for cognition and qualia. On the other hand, it may be that although we have observed great intelligence in animals such as the octopus or Sepia, these creatures might not in fact have anything like qualia. My position, though, is that it is the simplest assumption from what we see, and given Occam’s razor, the onus of proof lies with those who believe that these animals are devoid of qualia.

What do we make of the intelligence of an ant that as a robot demonstrates incredible computational agility with mere milligrams of neuronal mass, a brain with less mass than a single microchip?

But is there something in principle quite different from the types of embodiments that we have in modern day computers and in the nervous system itself? That is a very serious and important question to ask. One may consider, as did Alan Turing, whether it is in principle possible to make a universal machine out of a digital type of device if the appropriate algorithms are implemented. Can algorithmic computation ever be sufficiently extensive, fast, and concise enough to implement the totality of properties that a 14-watt entity such as our brain can implement with 1.5 kilograms of mass? And what do we make of the intelligence of an ant that as a robot demonstrates incredible computational agility with mere milligrams of neuronal mass, a brain with less mass than a single microchip? The fundamental issue is that brains are nothing like digital computers; they operate as analog devices and thus utilize physics directly in their measurements, as opposed to the abstracted measures of zeros and ones that are cleansed of the elements that generated them. Is the computation of digital physical computers truly comparable to that performed by analog devices? It has been stated that for a digital computer to be able to support the equivalent computational properties (capabilities) of the brain, the mass required might be many orders of magnitude larger and the power supply equally large.


There is another argument to consider in terms of the differences between brains and computers. Warren McCullough wondered long ago how it was that reliability could arise from nonreliable systems. The reader should know by now how unreliable nerve cells are as computational entities. First of all, they have intrinsic activity, and thus as conveyors and relayers of information may be extremely noisy. McCullough’s answer was rather intriguing: he felt that reliability could be attained if neurons were organized in parallel so that the ultimate message was the sum of activity of the neurons acting simultaneously. He further explained that a system where the elements were unreliable to the point that their unreliabilities were sufficiently different from one another would in principle be far more reliable than a system made out of totally reliable parts. Here, a reliable system is one with unreliability in each element as low as possible but still present.

For an instrument to be totally reliable it must ultimately be made up of unreliable—varied—parts!

This may sound almost paradoxical, but in what is considered a reliable system, the elements are reliable to about the same extent. And even if this reliability is 99.99 percent, the problem is that the elements are also all the same in their unreliability, meaning that what is unreliable is common to all the elements. It therefore becomes an issue of probabilities. In such reliable or redundant systems then, whatever tiny problem or unreliability they do have will add up. In nonreliable systems, however, the elements are not redundant and are therefore far more reliable than reliable systems. The flip side of this is that in a system with elements of differing unreliabilities, what they have in common are the reliable aspects! This is fundamental. It means that for an instrument to be totally reliable it must ultimately be made up of unreliable—varied—parts!

Here is precisely the fragility of society we may experience from the homogenizing effect of the Web on our thoughts, ideas, beliefs, and the like. As variation decreases, things become increasingly redundant and unreliability becomes the prevailing commonality across the elements—us.

Returning to the implementation of consciousness, it is possible that until we can understand the issue of unreliability and the probabilistic nature of computation in analog systems, we will not be able to generate the required architecture. With the proper functional architecture we could probably generate consciousness in a very large set of nonbiological entities.

The second issue is one of knowledge of self. Suppose a potential embodiment of consciousness is allowed the necessary freedom to explore and internalize the external world such that an image of self, however primitive it may be, is implemented. While this embodiment may measure external reality, the possibility of having an entity that is aware in the sense we mean most probably will not ever arise. We know that this is fundamental in the functioning of the nervous system. It can be seen in individuals who are given inverted prisms that make the world appear visually upside down. These individuals will learn to turn the visual image right side-up only if allowed to interact in a motor sense with that image. They must move within it to adjust. Ultimately we see that the architecture capable of generating cognition must relate to the motricity upon which such cognition was developed. When such architectures are finally realized, we may have thinking/feeling machines. However, our ability to design and build them may not ultimately be that useful in understanding brain function, in the same way that understanding airplanes may not tell us how the physiology of bats or birds enables them to fly.  

About Cerebrum

Bill Glovin, editor
Carolyn Asbury, Ph.D., consultant

Scientific Advisory Board
Joseph T. Coyle, M.D., Harvard Medical School
Kay Redfield Jamison, Ph.D., The Johns Hopkins University School of Medicine
Pierre J. Magistretti, M.D., Ph.D., University of Lausanne Medical School and Hospital
Robert Malenka, M.D., Ph.D., Stanford University School of Medicine
Bruce S. McEwen, Ph.D., The Rockefeller University
Donald Price, M.D., The Johns Hopkins University School of Medicine

Do you have a comment or question about something you've read in CerebrumContact Cerebrum Now.