Inside Information

<v Narrator>There's too much information making your head spin. <v Narrator>Do you feel at the mercy of modern technology? <v Narrator>Does reality sometimes seem like an illusion? <v Narrator>Well, don't worry. It's all in your head. <v Narrator>Your personal computer has the inside information. <v Narrator>Next. <v Announcer>Major funding for this program is provided by the L.K. <v Announcer>Whittier Foundation. <v Narrator>A movie set, a typical scene. <v Narrator>It'll last maybe 20 seconds in the theater. <v Narrator>But right now, it's taking the energies of dozens of people working together. <v Speaker>[Shots fire] <v Narrator>A movie, Alfred Hitchcock once said, is life with the dull

<v Narrator>parts cut out. And there's an element of truth <v Narrator>in that. <v Narrator>We go to a movie to be, well, moved. <v Narrator>It isn't exactly brain surgery, but it can quite literally <v Narrator>change your mind. When you look at a movie you're really watching the end <v Narrator>product of a great deal of behind the scenes activity, not just <v Narrator>on the set. I'm talking about backstage in your brain. <v Narrator>Because in your mind's eye, although you're no more aware <v Narrator>of it than you are of the magic behind the movies, there's a great deal of <v Narrator>activity going on. It's as if you have a kind of personal production company up <v Narrator>here putting together your picture of reality. <v Narrator>What's the picture made of? <v Narrator>Well, nowadays we'd say information. <v Narrator>Our brains are the stars of the information processing business. <v Narrator>How do they do that? How well do they handle information?

<v Narrator>How is the very idea of information changing the way we think about ourselves? <v Narrator>Well, that's what this program is about. <v Narrator>Behind the scenes, inside information. <v Narrator>This was a watershed in the story of information <v Narrator>before its invention. Information could travel only as fast as a horse could <v Narrator>gallop or a ship could sail. <v Narrator>The telegraph, slicing electricity into a code of dots and dashes <v Narrator>allowed a message to be transmitted at nearly the speed of light.

<v Narrator>At its destination, the information could be cast in tangible form, hard <v Narrator>copy. Then circulated to a wider audience. <v Narrator>This linotype machine invented a century ago was a giant step in <v Narrator>the automation of printing. <v Narrator>And with increases in speed and efficiency, printed knowledge began to grow. <v Narrator>In the 1660s, only a handful of journals recorded the discoveries <v Narrator>of science. Today, there are about 40,000. <v Narrator>The average American newspaper has twice as many pages as it did 20 years

<v Narrator>ago. <v Narrator>And it's estimated that total printed knowledge doubles every eight years. <v Narrator>Add to that radios, cellular phones, laptop computers, fax machines, <v Narrator>VCRs wall to wall communication systems and earth <v Narrator>wired for sound and vision. <v Narrator>And it's not surprising that we think of this as the age of information. <v Narrator>But what does that really mean? <v Narrator>Out here there's no technology, but I'm still processing information <v Narrator>at a rate that would make a supercomputer blow its circuits. <v Narrator>Sights, smells, sounds. <v Narrator>That's information. <v Narrator>The extra amount of information we have to process because we live in the age <v Narrator>of information is actually insignificant. <v Narrator>Here's an example of information transmission.

<v Narrator>The original message was information is a difference that makes a difference. <v Narrator>And here's one of the difficulties of communication. <v Narrator>A message can be distorted. <v Narrator>Information reduces uncertainty. <v Narrator>Right now, we don't know if the message will be heads or tails. <v Narrator>And now we know. That's information. <v Narrator>Information <v Narrator>must be expressed in some kind of code. In Morse code letters and numbers <v Narrator>are expressed in dots and dashes. In this Japanese dance form called kabuki, cultural <v Narrator>messages are carried by color, gesture and facial expression. <v Narrator>The idea that information might be a kind of universal vocabulary is recent <v Narrator>and revolutionary. Only since the 1940s have we begun to appreciate <v Narrator>the power and richness of information as a language for describing reality.

<v Narrator>Life itself can be thought of as information. <v Narrator>Genetic messages written in the code of DNA <v Narrator>and our invention of machines that process and store information has offered a new <v Narrator>way of thinking about the most powerful computer known. <v Narrator>It uses the equivalent of 20 watts to do the work of millions of supercomputers. <v Narrator>It's the ultimate personal computer. <v Narrator>Your brain. The brain's computing power is distributed over 10 <v Narrator>billion neurons. Caltech physicist Carver Mead specializes <v Narrator>in microelectronics. <v Carver Mead>The more you distribute information processing, the more effective it is. <v Carver Mead>So anyone that wants to hoard it simply loses <v Carver Mead>effectiveness. <v Carver Mead>It's the nature of information processing, and that is a deep thing. <v Carver Mead>That's not something we knew a long time ago. <v Carver Mead>We thought maybe it would be like steam engines.

<v Carver Mead>You built bigger ones, they'd be more efficient. <v Carver Mead>They're not. The bigger they are, the less efficient they are. <v Carver Mead>That's why the personal computer has taken over most of the computing in the world. <v Carver Mead>It's far more effective than the big computers are. <v Carver Mead>But it goes much further than that. If you divide computing even more finely, <v Carver Mead>it gets even more effective. <v Carver Mead>And the ultimate in that endeavor is the brain. <v Narrator>What does a computer do? <v Narrator>It performs a sequence of steps called a program or algorithm. <v Narrator>Each step must be so clearly described that a machine could carry it out. <v Narrator>An algorithm leaves nothing to the imagination. <v Narrator>A knitting algorithm might go knit one, purl two. <v Narrator>Knit one. Purl two. <v Narrator>An algorithm is a precise series of rules for doing a particular job. <v Narrator>In this case, making a sweater. <v Narrator>Algorithms are humble things, but according to some researchers, they're what minds

<v Narrator>are made of. Oxford University biologist Richard Dawkins. <v Richard Dawkins>I think it may be disturbing for people to realize, <v Richard Dawkins>as I believe they one day will have to realize, that everything that goes <v Richard Dawkins>on inside their minds, their brains and their minds, their soul <v Richard Dawkins>could be replicated on a computer like this. <v Richard Dawkins>That, I think, is a very disturbing thought. <v Richard Dawkins>It disturbs me, but I think it's true. <v Narrator>How could something so apparently spontaneous and intangible as human <v Narrator>creativity be reduced to a set of instructions? <v Narrator>Surely it's a thing of emotions, not equations. <v Narrator>And yet inspiration doesn't come from nowhere. <v Narrator>It's a product of the workings of our brains. <v Narrator>At UC San Diego, Harold Cohen is trying to understand the unconscious <v Narrator>mental processes of the artist by codifying some of that knowledge

<v Narrator>in a computer program he calls Aaron. <v Narrator>Aaron creates all the drawings in this studio. <v Narrator>But Cohen does more than add the finishing touches. <v Narrator>He taught Aaron everything it knows. <v Harold Cohen>I'm responsible for giving the program its own internal model of the outside world. <v Harold Cohen>The program knows enough about not only about how a human being is built. <v Harold Cohen>That is head, body, arms, legs and so forth, <v Harold Cohen>it knows where the articulations are. It knows what the legal range of movement is at <v Harold Cohen>each articulation. <v Harold Cohen>And most important of all, he knows how the body behaves. <v Harold Cohen>A coherently moving body is not a rag doll. <v Harold Cohen>There are laws governing its behavior, how it balances, how it gestures. <v Harold Cohen>All those things are things that the program has knowledge about. <v Narrator>Now, Aaron can tap that knowledge in different ways so that each drawing <v Narrator>is unique, unpredictable.

<v Harold Cohen>We are faced with a completely new set of circumstances. <v Harold Cohen>This has never happened before in human history. <v Harold Cohen>The ability to write a program that in some significant sense can <v Harold Cohen>behave as a sort of surrogate human being. <v Narrator>While Cohen is analyzing the eye of the artist, Carver Mead is studying the eye <v Narrator>of the beholder. The human visual system. <v Narrator>He's trying to create a silicone retina. <v Narrator>A computer chip model of the early stages of visual processing. <v Narrator>His goal is to understand better how our sensors work. <v Carver Mead>Your nervous system takes in almost unlimited amount of stuff <v Carver Mead>and sifts it and sifts it, and sifts it down until only <v Carver Mead>the relevant salient features are the ones that finally come through. <v Carver Mead>So it's wonderful at discarding garbage and <v Carver Mead>winnowing out the gems of information that are in this <v Carver Mead>babbling confusion of stuff that comes in through your senses.

<v Narrator>And what does the brain do with the information? <v Narrator>It makes patterns. As the Nobel Prize winning physiologist Sir Charles <v Narrator>Sherrington wrote, the brain is an enchanted loom where millions of <v Narrator>flashing shuttles weave a dissolving pattern, always a meaningful pattern, <v Narrator>though never an abiding one. <v Narrator>Poetic, perhaps, but it highlights an important difference between our brains <v Narrator>and the kinds of computers most of us are familiar with. <v Narrator>Caltech neuroscientist John Hopfield. <v John Hopfield>A given computer looks very good at some problems and very bad at other problems, <v John Hopfield>and it depends on what the designer designed it for. <v John Hopfield>In the case of digital machines that were primarily designed to do arithmetic. <v John Hopfield>And they're just brilliant there. And when you get down to it, biology <v John Hopfield>has no real designer. But evolution emphasized pattern <v John Hopfield>recognition was a way to survival and the evolution designed the brain, which is

<v John Hopfield>tremendous at pattern recognition and very poor at logic. <v John Hopfield>Logic doesn't help us survive very much. <v Narrator>When we make machines, we aim for precision. <v Narrator>We try to avoid the spontaneous, the haphazard evolution is a different <v Narrator>kind of watchmaker, as UC San Diego vision scientist V.S. <v Narrator>Ramachandran explains. <v V.S. Ramachandran>It's common to assume that the brain is in some <v V.S. Ramachandran>sense perfect, that it looks at the world and records <v V.S. Ramachandran>what's really out there. But we know that the brain is full of imperfections. <v V.S. Ramachandran>Like any other biological system, like the liver or heart or pancreas, <v V.S. Ramachandran>it's beautifully engineered. <v V.S. Ramachandran>But it's certainly not optimal or perfect. <v V.S. Ramachandran>In fact, I often point this out to my colleagues that the brain is <v V.S. Ramachandran>really a bag of tricks. <v V.S. Ramachandran>And it's interesting to compare the functions of the brain, especially the human brain, <v V.S. Ramachandran>or indeed of any biological system with a computer.

<v V.S. Ramachandran>And the key difference, I think, is that a computer <v V.S. Ramachandran>can be designed from scratch. <v V.S. Ramachandran>There's a master plan or an engineer who says, I want to get this job done. <v V.S. Ramachandran>What's the best way to do it? The brain, on the other hand, is really a patchwork <v V.S. Ramachandran>job for millions of years of evolution. <v V.S. Ramachandran>What seems to have happened is that the brain seems to have tried <v V.S. Ramachandran>different tricks, if you like, or shortcuts and picked the ones that work best. <v V.S. Ramachandran>So what you end up with is an enormous patchwork job rather than a perfect machine or <v V.S. Ramachandran>a perfect computer. <v Narrator>Here's an example of how our patchwork brains can make mistakes. <v Narrator>Most fans and coaches, too, will swear that players have hot shooting <v Narrator>streaks. But after sinking a few shots in a row, the chance of <v Narrator>hitting the next one is particularly good. <v Narrator>Well, it's not.

<v Narrator>The chance is just the same as on any other shot. <v Narrator>The hot hand is an illusion. <v Narrator>In the random sequence of hits and misses, we see patterns that aren't really there. <v Speaker>We have main engine start. <v Narrator>Here's another example of a thought illusion. <v Narrator>At one point, NASA engineers estimated the chance of space shuttle failure at one <v Narrator>in 100000. <v Narrator>Even the booster rockets had failed once every 57 firings. <v Speaker>Challenger go at throttle up. <v Narrator>The engineers were overconfident about their knowledge of the system. <v Speaker>Obviously a major malfunction. <v Narrator>NASA experts knew a great deal about the shuttle, but they had illusions about how <v Narrator>complete that knowledge was. <v Narrator>Political pressures might explain NASA's overconfidence in this particular case. <v Narrator>But psychological studies show again and again that most people, <v Narrator>technical experts included, overrate the reliability of their own knowledge. <v Narrator>The fact is, we don't know much about how much we know.

<v Narrator>How well do we perform when our powers of judgment are put on trial? <v Narrator>Members of the jury, your duty today is to determine the guilt or <v Narrator>innocence of the accused. <v Narrator>Now, you have heard instructions from the judge. <v Narrator>You've listened to testimony from witnesses. <v Narrator>And you've seen exhibits presented by the prosecution and the defense. <v Narrator>Your task now is to weigh the evidence and deliver your verdict. <v Speaker>If you are impressed with the gravity of this situation, the fact that someone's <v Speaker>fate may be determined by your power of judgment. <v Speaker>Well, I put it to members of the jury. <v Speaker>That's the case every day of your lives. <v Narrator>Every day you make decisions by combining the best information <v Narrator>you have available, just as in a courtroom and in a complex society like

<v Narrator>ours, the judgment of other people can affect your life. <v Narrator>So how good are we at making judgments? <v Narrator>Take the case of experts whose job it is to answer questions like <v Narrator>which prisoners should be paroled. <v Narrator>Who should be admitted to graduate school. <v Narrator>Which heart patients are likely to survive the next three years? <v Narrator>Well, in over 100 studies of cases like these there wasn't one <v Narrator>in which the judgment of an expert produced significantly better results than a computer <v Narrator>combining the same information according to a simple equation. <v Narrator>You see, we humans are good at deciding which information is relevant <v Narrator>to a problem, but we're not so good at combining that information to make <v Narrator>accurate judgments. <v Narrator>How do we see ourselves? <v Narrator>How accurate is the information we collect to form our self images? <v Narrator>UCLA psychologist Shelly Taylor.

<v Shelley Taylor>People see themselves as the central character and indeed <v Shelley Taylor>the hero of their own lives. <v Shelley Taylor>We tend to interpret our behavior in very benign terms. <v Shelley Taylor>We see a behavior that other people might interpret as self centered <v Shelley Taylor>or tedious, as witty and engaging. <v Shelley Taylor>We we make very self aggrandizing interpretations of what we <v Shelley Taylor>do, and then we encode the information that way in memory. <v Shelley Taylor>So when we come to retrieve instances of our behavior, we have all <v Shelley Taylor>this wonderful data stored away off of our talents <v Shelley Taylor>and relatively little data stored away regarding our faults. <v Narrator>We edit our experiences, especially those that concern ourselves. <v Narrator>We throw out the bad takes, keep the good ones.

<v Shelley Taylor>This is not altogether a self-deception. <v Shelley Taylor>It's a self-deception with one eye open. <v Shelley Taylor>I'll give you an example from one of our cancer patients that we talked with. <v Shelley Taylor>This was a woman who had battled very hard against an initial and very severe bout <v Shelley Taylor>of cancer. And she said to me during the course of the interview, I will never, <v Shelley Taylor>never get cancer again. <v Shelley Taylor>I can keep it from coming back. <v Shelley Taylor>And then later in the interview, she said, and if it does come back, I'll just fight it <v Shelley Taylor>as hard as I did the first time. So she knew in the back of her mind that <v Shelley Taylor>there was a very good chance that she would have a recurrence. <v Narrator>According to Taylor, getting through tough times can be easier when we <v Narrator>put the best face on a situation. <v Narrator>Coping with the drama of life requires the right mental moves. <v Narrator>And we've been training for that for our entire evolutionary history.

<v Narrator>Chess is a kind of mental aerobics. <v Narrator>This game can be a real workout, and if current trends in computer <v Narrator>chess continue, there'll be more to sweat about. <v Narrator>The world's computer chess champion, which is a complex program, <v Narrator>has already started to beat human grandmasters at their own game. <v Narrator>And yet hardest chance is there are much harder things that we humans <v Narrator>do without a thought. <v Narrator>Just seeing these pieces is, in a sense, a much more difficult <v Narrator>problem than actually playing the game. <v Narrator>Computers can beat most of us at chess, but we can still beat every computer at <v Narrator>seeing, at processing sensory information. <v Carver Mead>I think probably the biggest single contribution to our understanding <v Carver Mead>of the brain that's been made by computers is our inability

<v Carver Mead>to do sensory processing with them. <v Carver Mead>We've just totally failed to be able to do real-time sensory processing <v Carver Mead>with the highest horsepower digital technology we have. <v Carver Mead>Remember, this is a technology that has evolved over a factor of about one hundred <v Carver Mead>million since we built the first computer. <v Carver Mead>And we're still not close to being able to do the simplest things that <v Carver Mead>the brain does. <v Narrator>The first computers were just turbo charged adding machines. <v Narrator>We built them to do things that were difficult for us to do by hand or by <v Narrator>brain, things like precise, elaborate calculations. <v Narrator>They were flashy and fast, but none too bright. <v Narrator>The question was, could they solve more interesting problems? <v Narrator>Do things that people do well? <v Narrator>Could an artificial intelligence be kindled within a computer? <v Narrator>In January 1956, <v Narrator>Herbert Simon, who would later win the Nobel Prize in economics, walked

<v Narrator>into his classroom and announced to his students that during the Christmas vacation, <v Narrator>he and a colleague had invented a thinking machine. <v Narrator>They'd actually written a computer program that went on to prove mathematical <v Narrator>theorems like this one, not just arithmetic, not just routine number <v Narrator>crunching. Proving a theorem like this is one of the hardest things that we do. <v Narrator>And so some scientists thought if a computer can do that, maybe <v Narrator>simulating some of the other things that people do will be just a mopping up operation. <v Narrator>Well, they're still mopping. <v Narrator>Thinking of the mind as a computer has taught us some important lessons. <v Narrator>The things we expected to be the most difficult to program, like playing <v Narrator>chess or proving theorems, turned out to be much easier <v Narrator>than the everyday skills we take for granted. <v Narrator>Vision is a good example.

<v Narrator>We make sense of the patterns of light that come into our eyes. <v Narrator>We see colors and shapes, recognize objects, judge their distance and direction. <v Narrator>All this is fast, subconscious and accurate. <v Narrator>Our perceptions are usually on the mark. <v Narrator>Let me tell you a story about survival, about <v Narrator>Mother Nature and Mrs. Winchester. <v Narrator>This is the house in San Jose, California, that Mrs. Winchester built. <v Narrator>Her name was Sarah, and she was heiress to the Winchester family <v Narrator>fortune. When she moved here in 1884, there were just eight <v Narrator>rooms when she'd finished. <v Narrator>There were over 160.

<v Narrator>Why did she do it? Well, it seems that a spirit medium had told her that she could <v Narrator>avoid death and perhaps escape the curse placed on her <v Narrator>by the spirits of all those people killed by Winchester rifles if she moved <v Narrator>west and kept building. <v Narrator>So she hired a crew of carpenters and kept them busy around the clock <v Narrator>for 38 years building, well, strange things like <v Narrator>a window that opens onto an elevator shaft. <v Narrator>A door that opens onto an eight foot drop into the kitchen. <v Narrator>There are cupboards that are a bit on the <v Narrator>small side. <v Narrator>And some cupboards that aren't cupboards at all. <v Narrator>Nobody building this house was working from a master blueprint because there simply <v Narrator>wasn't one. What you see here is the result of a series of spontaneous

<v Narrator>additions. In a sense, you could say the same thing <v Narrator>about the human brain. <v Narrator>There was no master blueprint, no designer. <v Narrator>It evolved. Mother Nature's handiwork, not Mrs. Winchester's. <v Narrator>Over millions of years, bits and pieces were added to our patchwork brains. <v Narrator>The strange architecture of this house is a record of Mrs. Winchester's <v Narrator>struggle to survive. <v Narrator>And the strange architecture of your brain is a record of the human struggle <v Narrator>to survive. <v Narrator>Here's the kind of curious design that evolution produces. <v Narrator>It's a map of the areas in a monkey's brain that process visual information, <v Narrator>32 distinct brain regions are involved, communicating over three hundred and <v Narrator>five pathways. <v Narrator>It's a patchwork, but it gets the job done beautifully. <v Narrator>In fact, visual systems are so good that you have to go to great lengths to fool

<v Narrator>them. <v Narrator>Take a look at this. At first glance, it seems to be a perfectly ordinary <v Narrator>triangle, but if you study the joints, you'll see that you could not build <v Narrator>a triangle like this. <v Narrator>Watch. <v Narrator>There's something very strange going on here. <v Narrator>It's your brain trying to tell the best story it can with the available <v Narrator>information. And this time it was fooled. <v Narrator>Keep your eyes on the red mark at the bottom of the screen, but pay attention to the <v Narrator>moving shapes. You should see the yellow patch on the left moving up <v Narrator>and down.

<v Narrator>Now, look at it. It wasn't moving at all. <v Narrator>Our brains used a gimmick that in this case led to the wrong answer about what was <v Narrator>moving. <v Narrator>Seems like a perfectly ordinary room, but there's <v Narrator>more to it than meets the eye. <v Narrator>Watch. <v Narrator>You see this room is in the Exploratorium in San Francisco. <v Narrator>And no, this is not earthquake damage. <v Narrator>It was designed this way. <v Narrator>It's a visual illusion and it's based on the fact that our brains <v Narrator>make assumptions. They make assumptions about the angles in this <v Narrator>room, that they're 90 degrees, right angles. <v Narrator>Except that in this case, those are the wrong angles. <v Narrator>If they were right angles, I'd be the same distance away in each corner. <v Narrator>But the walls and ceiling are actually askew.

<v Narrator>I'm really closer and so appear bigger on the right. <v Narrator>We don't get much insight into vision from our intuitions. <v V.S. Ramachandran>Typically, we ask a man in the street what goes on in your head when you perceive the <v V.S. Ramachandran>world, He will tell you there's an image in the eyeball, the retina, <v V.S. Ramachandran>and there's images transmitted faithfully through the optic nerve. <v V.S. Ramachandran>And you have a screen in the brain, a screen called the visual cortex, <v V.S. Ramachandran>where images reproduced. <v V.S. Ramachandran>That's how you perceive the world. Well, a moment's reflection will reveal that this is <v V.S. Ramachandran>complete nonsense. <v V.S. Ramachandran>There's a basic logical fallacy here, because if you create an image inside the <v V.S. Ramachandran>brain, inside the head, then you need another person in the head <v V.S. Ramachandran>looking at that image. <v Narrator>But how does this little person see? <v Narrator>Well, that's the problem we started with. <v Narrator>He must have an even smaller person in his head. <v Narrator>And it goes on and on. <v Narrator>This way of thinking about vision doesn't explain a thing. <v V.S. Ramachandran>There is no replica of the world inside the head.

<v V.S. Ramachandran>What you have instead is a very abstract, symbolic description of the world. <v V.S. Ramachandran>What we're all trying to do as scientists studying the brain is we're trying to <v V.S. Ramachandran>understand what the nature of that symbolic description is, what you're trying to - <v V.S. Ramachandran>you're cryptographers, if you like, trying to crack the code that the brain uses <v V.S. Ramachandran>when perceiving the external world. <v Narrator>What is a code? It's a system for expressing information in the symbols. <v Narrator>Here's a kind of code. This is the talking drum once used in Africa <v Narrator>to send messages from town to town. Musician Francis <v Narrator>Harway can translate my speech into a pattern of drum tones. <v Narrator>This is a telecommunication system, <v Narrator>a long distance carrier. <v Woman>You called me the other night. <v Woman>Oh, a few weeks ago on the telephone.

<v Woman>Yes. And I was so surprised. <v Narrator>Here's another way to reach out and touch someone. <v Narrator>We're most familiar with the sight and sound of language. <v Narrator>But for Rick Joy of Santa Rosa, California. <v Narrator>Blind and deaf from an early age. <v Narrator>Language has another dimension. <v Narrator>Using a system called Tadoma, he catches up on old times with a former <v Narrator>teacher. For Rick, the flow of air and vibrations of her jaw <v Narrator>and lips are a tactile code that allow him to understand speech. <v Woman>When, what year? <v Narrator>Inn the brain, <v Narrator>the activity of neurons is a code for representing information. <v Narrator>A neuron receives messages from other neurons, processes them and sometimes <v Narrator>fires, transmitting an electrical pulse. <v Narrator>Our picture of reality is sketched in neural codes. <v Narrator>Here's a way to see that we construct reality rather than somehow grasp it directly.

<v Narrator>We see this as a walking person because our brains connect the dots in a certain way <v Narrator>like this. <v Narrator>But we could have connected the dots in other ways like this. <v Narrator>Or like this. <v Narrator>There are many ways to interpret these moving dots. <v Narrator>Our brains pick this one. <v Narrator>To perceive. Is to make choices among interpretations. <v Narrator>And we make those choices with computation. <v Narrator>Here's another example of computing reality color. <v Narrator>The colors we see are the result of a mental algorithm. <v John Hopfield>When you first think about color and you learn as a student about light, you say <v John Hopfield>something is green because green light is coming to be it for that object. <v John Hopfield>And in a certain sense, that's true. <v John Hopfield>On the other hand, you know that by candle light and by sunlight,

<v John Hopfield>you know, somebody has red and yellow shirt, looks red and yellow, and the light <v John Hopfield>coming from that object is just radically different of the two circumstances. <v John Hopfield>And you do have to ask, why don't you see the change? <v John Hopfield>And you don't understand why that should be the case until you go <v John Hopfield>back and say, what's the purpose of having color vision? <v John Hopfield>We ought to be able to recognize objects. <v John Hopfield>Now, what's the same about that object between candlelight and sunlight? <v John Hopfield>Is the fact that the dyes in it haven't changed. <v John Hopfield>And so what you really want to recognize is what dyes are present at, not what light is <v John Hopfield>coming from it. <v Narrator>The mental algorithm for computing color isn't yet known. <v Narrator>But researchers think that the color of an object is based on comparing the light <v Narrator>from that object with the light from neighboring objects. <v V.S. Ramachandran>Looking at the brain as an information processing device says that you've not <v V.S. Ramachandran>completely understood everything yet. <v V.S. Ramachandran>I mean that what we think of as the world, what do you think of

<v V.S. Ramachandran>as the self is really convenient fiction, which happens <v V.S. Ramachandran>to work for the time being. <v V.S. Ramachandran>But it gives you highly - it gives you a clear idea that <v V.S. Ramachandran>our conception of reality and our conception of ourselves is highly tentative <v V.S. Ramachandran>in nature, that it can change any any time when there's new information available. <v V.S. Ramachandran>And I think this gives you ultimately gives you a sense of humility. <v Narrator>Let's go back to the drawing board. <v Narrator>Our brains are continuously computing what we call reality, making sketches <v Narrator>of the world. <v Narrator>We're not aware of most of that mental activity, that computation. <v Narrator>It seems to us as if we're aware just of the end results. <v Narrator>And as we've seen, although the models we make of the world are far from perfect.

<v Narrator>Most of the time we do just fine. <v Narrator>It's some measure of how little we know about the brain that we're always comparing it to <v Narrator>something. You know, it's like an enchanted loom or a telephone switchboard. <v Narrator>How about it's like a mail sorting office. <v Narrator>Messages are constantly arriving from the far corners of our sensory world. <v Narrator>They get sorted according to destination. <v Narrator>It's a fast, efficient system. <v Narrator>Mostly. <v Narrator>It may not be perfect, but it gets the job done.

<v Narrator>And here it is. <v Narrator>Message received. <v Narrator>But of course, the world isn't delivered right to us. <v Narrator>We have the impression that we take in our surroundings at a glance, <v Narrator>but that's an illusion. <v Narrator>We construct our picture of the world by moving our eyes several times a second. <v Narrator>Each glance takes in a part of the scene. <v Carver Mead>There's a place in your brain. There must be a place, nobody knows where it is, where <v Carver Mead>this collage is put together into a single unified percept of <v Carver Mead>the outside world. But that's a computed construct. <v Carver Mead>It's not something you see. <v Carver Mead>You piece it together from these individual little vignettes. <v Narrator>Sampling a glass of wine is a practical demonstration of how <v Narrator>the brain works as an information processor.

<v Narrator>Wine is literally a message in a bottle. <v Narrator>And to read it, without cheating by looking at the label, <v Narrator>we use our senses. <v Narrator>The first clue is visual. <v Narrator>The color, the appearance, but clarity. <v Narrator>And in fact, there are some experts who can tell you the year the grapes were harvested. <v Narrator>The vintage of the wine just by looking at its color. <v Narrator>Next, the smell, what professionals call the bouquet. <v Narrator>And again, there are experts who can pinpoint the vineyard just by sniffing <v Narrator>the wine. Finally. <v Narrator>Taste. Now, those three different sensory <v Narrator>mechanisms developed during our evolutionary history to solve <v Narrator>different problems that were essential to our survival. <v Narrator>We have different mechanisms for collecting the sensory information and different <v Narrator>specialized parts of our brains for processing it.

<v Narrator>So in a sense, a glass of wine reveals the compartments, <v Narrator>the modules of the mind. <v Narrator>A film editor works with two different kinds of information, picture <v Narrator>and sound. The sound is stored in a magnetic stripe. <v Narrator>One part of this machine is specialized for decoding it. <v Narrator>There's a division of labor in the brain as well. <v V.S. Ramachandran>Different parts of the brain are specialized for different jobs, and they carry out these <v V.S. Ramachandran>functions relatively independent of other parts of the brain. <v V.S. Ramachandran>Of course, it doesn't mean that these different parts of the brain don't communicate with <v V.S. Ramachandran>each other. They do. And that's why you are one person. <v V.S. Ramachandran>But there is a surprising degree of autonomy and function. <v V.S. Ramachandran>And this is the I think, the new insight into how the brain works. <v Narrator>This clinical recording by neurologist Barbara File

<v Narrator>shows evidence for this kind of design. In this patient, a stroke <v Narrator>has cut off vision from other compartments of the mind. <v Narrator>She sees well enough to read, but she's unable to name objects by looking <v Narrator>at them. Her other senses, however, work fine. <v Narrator>The syndrome is called visual agnosia. <v Barbara File>What is this? <v Patient>That looks like, uh... <v Narrator>Looking at a clothespin, she's baffled. <v Narrator>But she recognizes it quickly with another sense. <v Barbara File>You can touch it then tell me what it is. <v Patient>Oh I can touch it? Now I can see it already. It's a clothespin. <v Barbara File>Okay. What's this here? <v Narrator>Here, she can't recognize a candle by looking at it. <v Narrator>Her sense of touch lets her make a good guess. <v Narrator>But this time, it's her sense of smell that provides the answer. <v Patient>Oh, <v Patient>it's a candle.

<v Narrator>Researchers are now finding that the brain's compartments are even further subdivided. <v Narrator>Take vision, for example, different kinds of visual information <v Narrator>seemed to be processed separately. <v Narrator>Some of the evidence for that comes from studying patients with brain damage. <v Narrator>And there's a classic case of a woman with a disorder called motion blindness. <v Narrator>For her, even pouring a drink became a guessing game. <v Narrator>The sub compartment of her brain that handles visual motion was damaged. <v Narrator>She had no trouble seeing cars, but she couldn't judge their speed. <v Narrator>So crossing the street became an ordeal. <v Narrator>She avoided parties because of the way people suddenly seemed <v Narrator>to appear in front of her. <v Narrator>And how do you follow a conversation when you can't see the motion of someone's <v Narrator>lips? <v Narrator>Motion is only one of several visual sub compartments in the brain. <v Narrator>You know, those days when life seems colorless?

<v Narrator>Well, for some people suffering from a disorder called achromatopsia it really <v Narrator>is. Their retinas work fine but the part of the brain <v Narrator>that processes color is damaged. <v Narrator>So their world really is black and white. <v Narrator>If the mind is made of modules, how do they act together? <v Narrator>In fact, how do billions of neurons act together to produce art, <v Narrator>poetry, flights of the imagination? <v Narrator>Here's a simple analogy. <v Narrator>A single bird can do many things. <v Narrator>It can soar. Build a nest and perch. <v Narrator>But one thing it can't do is flock. <v Narrator>How do individual birds fly together as if they were of one mind? <v Narrator>An answer comes from computer simulations of flocks.

<v Narrator>In this animation, every bird independently follows a few simple rules. <v Narrator>So this isn't a cartoon where every motion is specified by an animator. <v Narrator>It's more of an experiment within a computer. <v Narrator>Intuition might lead you to expect an air traffic controllers nightmare. <v Narrator>And yet, out of these rules spread throughout the group, a flock emerges. <v Narrator>This is a computer reconstruction of a neuron from a human brain. <v Narrator>A cell like this can receive process and transmit information, <v Narrator>but it can't see or think. A collection of billions of <v Narrator>neurons - a brain c- an see and think.Tthe neurons <v Narrator>somehow flock together to form a mind. <v Narrator>If it looks like a duck, walks like a duck and quacks like a duck, <v Narrator>it's probably a duckreBut these different kinds of information, sound

<v Narrator>motion appearance, are processed in different modules of the mind. <v Narrator>What ties them together to give us a unified experience of a duck? <v Narrator>Researchers are beginning to speculate about a possible answer. <v Narrator>It seems that when neurons are carrying information about different objects, they <v Narrator>fire independently of one another. <v Narrator>But when they carry information about a single object, they seem to fire together about <v Narrator>40 times a second. <v Narrator>Being in sync may turn out to be the glue that binds the pieces of our <v Narrator>perceptual world together. <v Narrator>This isn't something that could have been discovered by studying individual neurons. <v Carver Mead>It isn't the pieces that carry the secret of thought. <v Carver Mead>It's the way the whole system works together. <v Carver Mead>It's a much more holistic view of what the question is. <v Carver Mead>How does the system work? <v Carver Mead>And that's been hard because science did so well for so long, being reductionist, <v Carver Mead>going down and getting the one problem at the bottom and solving

<v Carver Mead>that. And then everything flowed from there. <v Carver Mead>It's hard for us to get our heads around the fact that that <v Carver Mead>isn't the question in neurobiology. <v Carver Mead>The question is, how is the system organized? <v Carver Mead>How can it carry out these fantastic computations? <v Carver Mead>How would you even imagine organizing a system that could do that? <v Narrator>For years, researchers have studied the minds' computations with serial <v Narrator>computers, machines that carry out instructions one step at a time. <v Narrator>Some scientists are saying this approach may be a dead end. <v Narrator>The possibilities of parallel computing are now being explored using

<v Narrator>neural networks inspired by the architecture of the brain. <v Narrator>Each computing unit of the network sends messages to other units which process <v Narrator>the information and relay new messages. <v Narrator>The strength of their connections can be changed to fine tune the performance of the <v Narrator>network. The idea is simple. <v Narrator>Instead of one powerful central computer, there are many more modest processors <v Narrator>linked together. <v Narrator>Of course, real neurons are much more complicated than the units in these networks. <v Narrator>And yet researchers hope that these connectionist models may offer clues <v Narrator>to the way minds work. <v Narrator>For example, different memories in a single network may have processing <v Narrator>units in common. <v Narrator>Jeff Hinton is a psychologist and computer scientist with the Canadian Institute <v Narrator>for Advanced Research. <v Geoff Hinton>If I, for example, tell you a fact that you probably didn't know - that chimpanzees just <v Geoff Hinton>love eating onions - and now I ask you, do you think gorillas like like eating

<v Geoff Hinton>onions? Well, you don't know, but you think it's a bit more likely than you thought it <v Geoff Hinton>was before. That will happen automatically within your own network because internally <v Geoff Hinton>it's representation for a chimpanzee. <v Geoff Hinton>We'll share a lot of neurons in common with his representation for a gorilla. <v Geoff Hinton>And when you learn that chimpanzees like onions, you'll change some of the connection <v Geoff Hinton>strengths coming from the neurons that are active in the representation of chimpanzee. <v Geoff Hinton>And that will automatically change some of the connections strengths coming from the <v Geoff Hinton>neurons that are active in the representation of gorilla. <v Geoff Hinton>So you'll get this automatic generalization to similar things. <v Narrator>Parallel computation seems to be the right approach for many of the problems that animals <v Narrator>need to solve. <v Narrator>A beehive can be considered a parallel information processor. <v Narrator>Individual bees are the processers, collecting information and distributing it. <v Narrator>The result? An efficient society. <v Narrator>A free market is another case of information processing spread over many individuals. <v Narrator>Every sale can be seen as an exchange of information, and the result

<v Narrator>of these parallel computations, according to some economists, is an efficient <v Narrator>distribution of goods. <v Narrator>Carver Mead is an advocate of parallel computation, and he is putting his ideas <v Narrator>to the test. Trying to copy some of the brains decentralized computing <v Narrator>power in a silicon chip. <v Carver Mead>What you're looking at is not a television picture. <v Carver Mead>It's the result of an image processed by a silicon retina. <v Carver Mead>The first stage in the processing of your visual system <v Carver Mead>is done by a set of cells just behind your eyeball that <v Carver Mead>extract the first kinds of salient information from <v Carver Mead>the image coming in through your lens. <v Carver Mead>The silicon retina does a similar kind of computation. <v Carver Mead>It's extracting the salient information and discarding <v Carver Mead>less interesting information.

<v Narrator>The word retina comes from the Latin word for net, and that's a good way to think <v Narrator>about Mead's silicon version. <v Narrator>It's a net for catching information. <v Narrator>Like the human eye, the chip is a surprise detector. <v Narrator>It finds places in an image that are darker, brighter or moving <v Narrator>compared to their neighbors. <v Narrator>This image, remember, is not produced by a camera. <v Narrator>A camera simply records. <v Narrator>The silicon retina, like the human eye, interprets, computes. <v Narrator>The silicon retina uses parallel computation. <v Narrator>Knitting a sweater is a stepwise or serial algorithm. <v Narrator>Different styles of computation fit different problems. <v Narrator>This is the end product of a knitting algorithm. <v Narrator>It's one of the things you can get from following a set of instructions. <v Narrator>I'm another.

<v Narrator>My DNA is an algorithm. <v Narrator>It contains the instructions for building me. <v Narrator>Incidentally, the lady who carried out the sweater algorithm also <v Narrator>helped create my algorithm. <v Narrator>She's my mother. <v Narrator>Part of her DNA became my DNA. <v Narrator>This molecule contains an algorithm for building a living thing. <v Narrator>I'm the Three-Dimensional output of my DNA program, constructing <v Narrator>an organism from its genes m aybe the first information processing problem <v Narrator>solved on Earth. <v Richard Dawkins>Nowadays, genetics is an intensely digital process. <v Richard Dawkins>I would almost go as far as to say genetics is just a branch of computer science or vice <v Richard Dawkins>versa. There's there's very little difference between the way you can think about genetic <v Richard Dawkins>information and the way you think about information in a computer. <v Narrator>DNA - primordial information - has spawned an information <v Narrator>processor, the human brain.

<v Narrator>And this computer has culture. <v Narrator>Culture is information that is learned. <v Narrator>Information like beliefs, ceremonies and traditions. <v Narrator>Culture is a second evolutionary path and it's faster than biological <v Narrator>evolution, where information must cross a generation gap as it travels from gene <v Narrator>to gene. Cultural information travels from brain to brain. <v Narrator>Christianity is cultural information. <v Narrator>So is architecture and music and the stories we tell about where we <v Narrator>came from and where we're going. <v Narrator>And so is science. <v Narrator>And what is the culture of science doing? <v Narrator>It's creating more and more sophisticated computers and programs, programs <v Narrator>that are now beginning to simulate minds.

<v Narrator>But they may go farther than that. <v Narrator>Some researchers imagine a day when computer programs will create their own successes. <v Narrator>The computer, our brainchild, may then head off on yet another evolutionary <v Narrator>path. The journey that began with DNA is leading to unexpected <v Narrator>destinations. <v Richard Dawkins>Genetic evolution got brains up to the point where they were complicated <v Richard Dawkins>enough to start developing language and start developing culture. <v Richard Dawkins>And that enabled a new kind of evolution to start. <v Richard Dawkins>So there's been a kind of takeoff and it runs a lot faster than genetic evolution. <v Richard Dawkins>After a few centuries of this cultural evolution - or a few millennia of this cultural <v Richard Dawkins>evolution - brains started to develop artifacts so <v Richard Dawkins>complicated - computers I'm talking about - that they themselves were potentially <v Richard Dawkins>capable of yet another takeoff point. <v Richard Dawkins>And that may be what we're now about to see. <v Richard Dawkins>There may be yet another kind of cultural evolution which leaves humans

<v Richard Dawkins>out, which ultimately leaves humans out. <v Richard Dawkins>I suppose this is yet another blow to our vanity, if finally, <v Richard Dawkins>we become superfluous to civilization. <v Narrator>Might these be the first steps towards a silicon civilization? <v Narrator>This is a computer animation by Michael McKenna of M.I.T.'s Media <v Narrator>Lab. Unlike traditional animation, the exact motions <v Narrator>in this film aren't scripted in advance. <v Narrator>There's a program in this bug. <v Narrator>The animator has set out the path and speed of a cockroach. <v Narrator>But an algorithm automatically computes the motion of the legs. <v Narrator>This could be a path that leads towards artificial life, a pattern of activity <v Narrator>within a computer so complex and adaptive we might be tempted to think

<v Narrator>of it as a alive. <v Narrator>We are sometimes made uneasy by the achievements of science. <v Narrator>We feel perhaps that with each discovery that opens us to closer investigation <v Narrator>and strips away some of the mystery of how we think or the mechanisms of creativity, <v Narrator>that we're somehow diminished. <v Narrator>But something is lost. <v Narrator>And that prospect is fundamentally disturbing. <v Narrator>Comparing the brain to an enchanted loom was really an admission of our ignorance. <v Narrator>It was just a metaphor, not a working model. <v Narrator>No scientists installed looms in their laboratories to study the warp and weft <v Narrator>of the fabric of mind. <v Narrator>But the computer is different. <v Narrator>We now routinely use the invention to study the inventor. <v Narrator>The notion that we are information processes that we perform computations

<v Narrator>to assemble, pictures of reality has yielded new insights. <v Narrator>How has this model changed our image of ourselves? <v Narrator>Well, for one thing, we've learned that our brains are far more sophisticated than we <v Narrator>could have imagined before we tried to reproduce our performance in silicon. <v Narrator>But we've also discovered that our brains are patchwork, not perfect. <v Narrator>It's perhaps ironic that the computer, for some a symbol of the dehumanizing <v Narrator>power of technology, is becoming our collaborator in our search for self-knowledge. <v Narrator>The information processing model of mind can even be thought of as liberating. <v Narrator>Focusing on what human beings have in common as thinking machines <v Narrator>is helping us to appreciate more clearly what it is to be human. <v Narrator>You do find it an exciting time, don't you? <v Carver Mead>It's fantastic. There's never been a time like this in human history. <v Carver Mead>This is the greatest intellectual quest of all time.

<v Carver Mead>Hands down. <v Announcer>Major funding for this program is provided by the L.K.

<v Announcer>Whittier Foundation.