Everything is Miscellaneous | David Weinberger | Talks at Google

Thank you so much. Thank you, Kevin Marks, for arranging this. Thanks for having me in. I have a book out, so I don't know what to say. So I'm David Weinberger. I'm at the Harvard Berkman Center. I'm a writer. This is a book that came out on May 1, so that was like a week ago. This is our actual miscellaneous draw in our kitchen. I know it's fascinating. You can tell a lot about us from the that we occasionally open cans and have difficulty opening large jars. The book is called Everything is Miscellaneous the Power of the New Digital Disorder subtitle. And the basic idea of it is. So I'm going to talk for too long, first of all, a B. I will get to a point where you'll be sure that I'm done because it sounds like I say and so. But it's a false ending. And I go on from there and I want to get you over your little disappointment. The basic idea of the book is that we're really good at organizing stuff. We've had, I don't know how to count but say 10,000 years figuring out how to organize. Also, I'm going to talk way too fast, and if I'm going too fast, let me know. I may speed up as I see time dwindling. We've gotten really good at organizing physical objects, just fantastic at it. And also then organizing ideas and knowledge. And the basic idea has been that you make some type of organization that could be as neat as this one. We do like these trees very much. And then there's always one section of it where the stuff that doesn't fit, you shovel everything into that. And if that section, and this is true, whether it's an org chart or it's a. Although org charts actually don't have miscellaneous positions now they think of it, maybe they should, but they do. They're not titled miscellaneous. Physically. It would be the drawer in your kitchen, the one that I just showed you a picture of, where if you. If that becomes too big, then your organization scheme has failed traditionally, then you've just got to start over again. What the book is suggesting is that what's going on online as a strategy, and it's a good strategy, is that the miscellaneous drawer is taking over the org chart. And that's actually incredibly healthy. It's a huge step forward for us in business, but also in science and education and politics. And the stuff that's touching knowledge is actually a really good thing. PowerPoint scales, fonts, horribly. So we're all used to playlists. I mean, they're very common. They're really, really. They're tiny things, but they say something very big about what we're doing. The fact that we're able to fluidly divide up our stuff any way we want is actually quite new, especially if you compare it to what we had to do. What we have to do in the physical world, which is we have to come up with a set of categories. One person gets to come up them, or a committee, but there's only one. In the physical world, you only get to order things one way. And so what the growth of the miscellaneous is actually about, Although the book, for marketing purposes, doesn't say this correctly, for marketing purposes, the book is actually an argument against the idea. And we could put it into Aristotle's mouth if you want. If you really want to drive down sales, we'll say it's an argument against Aristotle that there's one right way of ordering the world. Well, there's had to be. We've had to come up with singular ways of ordering the world because we're ordering physical things, and physical things have to go in one spot. It's just the way it works. It's as if reality's secret plan is just to keep things apart, to make it as no matter how hard you try, you cannot have two things in the same spot. So that's the first principle that has guided our physical organization. And the second, obviously closely related to that, is that everything has to go somewhere. So in every domain, we come up with ways of dividing the stuff. And it's different in every domain, and it's appropriate to each domain, but it's in every domain everything has to go someplace, and it can only go one place. And we're amazingly good at doing this work after thousands of years of doing it. Nevertheless, those are very severe limitations. They're not natural limitations on thought. They are on physical stuff, but not on thought. And there's an immediate political consequence to having to do this, to having to put things in one and only one place, which is somebody's got to make that decision, and that decision gives that person power and authority. It's just the way it works. So when somebody decides what goes on the front page of the newspaper, and they have to, because you only have those two square feet or so to give the most important news of the day, that's a big deal. So we have editors who are good at doing that. And generally they make really good decisions, generally, but they never can make a decision that's right for everybody. You can't do it. It's not because they're not good at it, it's just that it can't be done. Same thing with encyclopedias and books in general. Somebody decides what gets published. The 32 volumes of the Encyclopedia Britannica are not decided upon because that's how much knowledge there is. That's what you can publish economically. You cannot have an infinite encyclopedia until you move online and then you can. So 32 volumes, 65,000 topics. So you got to pick the best ones you can. And we make reasonable decisions, and we make reasonable decisions about how to divide up those topics, how to organize the information, either in an encyclopedia or a textbook, or in a newspaper or in a store. Nevertheless, it's a decision being made for everyone. And this pervades our culture. It has to. Because we have lived in a world of atoms where we have to classify things in order to decide which movies you get to see. This was an X rated movie at the time, which is now laughable, involves things that aren't so important, like who gets to compete in tournaments, in the Olympics and how an online e commerce site is going to divide up its stuff so we can browse through it in ways that they hope will find useful and appealing. Still, you make that one decision. The question of categorization determines who gets to participate in presidential debates, for example, is an immediate effect on how democracy works. And then for much more important things like apartheid, where it was 3,000 pages of race rules, of race laws, if you were not white, you carried around in South Africa a 50 page book about your race. This was insanity. It was so finely divided that they had to come up with a category of honorary whites. This is because people weren't. They knew that some people were not supposed to be oppressed, but they couldn't figure out how to get the laws to work that way. So they came up with a category. There's a famous case of one guy in South Africa who by the time he, he was 50, had been reclassified racially because the laws changed. And so now he was a different race five times by the time he was 50. In one occasion having to leave his wife and family because his race changed. As he sat there, his race changed. Likewise. This is the American Psychiatric association meeting in 1972. This guy, it's not a great picture, but this guy doesn't look like this. This is a mask. This is John Fryer in a mask, in a wig, in two suits of clothing, so his body would be hidden and a voice Disguiser so that he could do this outrageous thing of be on a panel and acknowledge that he was gay. Because at that point homosexuality was classified as a disease. And you could not have that disease and be a psychiatrist. You would be drummed out of the apa. The next meeting, a guy gave a talk, also gay, called Stop it, you're making me sick. And it was literally, that was literally the case that the rules were classifying him as diseased. And the instant the rules were changed, he was healthy. It was just that easy. So there are obviously really serious consequences to how we categorize and classify. Plato had really put this quite beautifully when he said that those who know how to talk and head towards the truth. Key principle is to carve nature at its joints. And he had in mind literally butchering an animal. It's a wonderful metaphor. I mean, you know that there are places where you just have to hack in the bad sense. You have to hack at the bone. That's not a good butcher. Good butcher knows where the natural places to divide the world up are. And that's what we should be doing. So once again, this is the metaphor of the physical guiding our idea about how ideas work. So this is what he. I just like this traitor to his species here. But this is how the world looked to Plato. It's also how it has looked to us. Even though we tend to say no, no, we're not that naive, but in some ways, frankly we are. This is a very hard belief to give up. So here is the solar system, except as of the summer, here is the solar system. There goes Pluto. What happened was, as you know, the International Astronomical Union had committees that met for years and got embroiled in politics. I mean, they had to reformulate the committees. And when it finally came to a vote, the work of the committees was largely ignored. They put it to a majority vote. What is a planet? A majority vote among scientists. So they came up with a definition. And the definition has basically two parts definition of a planet which has never had a definition before. I mean, these are important objects of scientific study. They were undefined from the ancient Greeks. The word planet comes from wanderer. It's the lights in the heavens that don't follow the circular motion. But once we discovered ones that were not visible with the naked eye, we don't have a. There's no definition thousands of years. So two part definition. An object that is large enough to round itself because it has enough mass that it will round itself. The natural action of gravity. And the second is that it will clear space. Well, so why that first definition? Because. It's because they wanted to say that a planet is big, but big is not scientific. And one of the proposals was, well, we'll say Mercury or half the size of Mercury, but that's not scientific either. So let's come up with scientific. And what scientific is, oh, well, gravity has this effect. Okay, that's true. But why care about that? Why is that a criterion? Who cares? Likewise, clear at space. The only reason that's in there is to make sure that we get, well, eight out of the nine planets. The aim of this was to preserve the planets. It's that simple. And we would come up with whatever definition we had to in order to do that. So we ended up pointing at attributes that are trivial, that are as trivial as Paris Hilton boobs. There's no scientific interest in objects that are large enough to round themselves. I'm sorry, that came out entirely wrong. I've talked about this before. I've never made the comparison, and I didn't know it was going to end up that way. Oh, my God. I have no idea where I was. Paris Hilton boobs. So there's no. The largeness of the mass of planets is completely uninteresting. What's interesting about planets, maybe, or about bodies circling the sun, is that some of them have atmospheres. That's pretty interesting because other things follow from it. Or they have water. That's pretty interesting because maybe there's life on these planets, but big and round, who cares? It's only because we wanted to keep those nine. So there are four objects, apparently, that have atmospheres in the solar system. And I'll probably get this wrong, but you'll correct me, there are four, and one of them is Titan. It's a moon. That's an interesting way of slicing things up. Doing it this way is really, really scientifically uninteresting. What's far better would be to take a miscellaneous view. That is to say, there's tons and tons of attributes depending on what we're interested in. We'll slice and dice different ways. If we're interested in extraterrestrial life, then water is very interesting. If we're interested in mining, then presence of copper becomes interesting. If we're interested in things that may hit us, then other objects become interesting. So the notion that there are nine things we need to preserve right back to Plato. It's far better to be able to slice and dice along many, many different attributes. In fact, it may be. I'll go so far as to say it is the case that the universe to Plato looked like this, but in fact is much more like this, where there are just so many attributes of things, and thus so many ways we can cluster and slice and dice, that to limit yourself to just some set of them, to say, well, there are natural joints, well, every attribute is a joint. Every attribute of every object is a way that you might want to cluster and join. That's not to say every way of slicing and dicing is right, that the world is in flux and anything you want goes. That's not the case. As Umberto Eco, the novelist and academic philosopher, by the way, has pointed out, there are many cuts of animal, but there's none that connects the snout and the tail. Not everything is possible, but there are so many. There's an indefinite number of ways we can slice and dice the universe. So we do. So we make categories. And by making categories, all that we're doing is saying the two things are alike in some way. But we get to choose what the principle of likeness is, depending upon what our interests are, what we're trying to do, what our project is. We can be wrong, right? I mean, we can be wrong, or we can be right. We can misclassify. For sure there are limits. But it's also the case that any two things we bring together, we probably can find a way in which they can be brought together. There's some principle undergirding them if we want them to be so. We have enormous latitude in how we do this, but we don't do it randomly. We do it because we're trying to do something. That's the problem with thinking there's only one order of the universe. The order of the universe, that one is not based upon our projects, and thus is the one order that nobody cares about. The rest of the orders that we see, the real ones, always emerge because they're trying to do something that matters to us. We're able to do this because not only do things have attributes, the attributes tend to cluster. So if you come across something that looks like this, smells like this, and feels like this, very likely it also tastes like it. And occasionally you're wrong and you end up picking waxed fruit out of your mouth. But generally you're not. The world just seems to be clustered, to have these clusters of attributes. It seems to be a condition for, you know, at the end of the universe, when everything's homogenized, that'll be different. But right now, things are clustered very nicely for us. There's a lot to explore and to know. And we get to do it the way that makes sense to us. So we make these categories and we make further categories. And because we're a clever species, we go up a level and we start making more categories of categories and categories of categories until we get entire trees. These are incredibly useful objects. They're a really useful way of understanding the universe because they let us know these things inherit from the top of it. So if you know that apple is a fruit, you also know that it's part of a plant, it's part of vegetative life, and so forth. But you don't have to know that every time you bite into an apple, it's there for you to know. This is a very compact way of representing the universe and very powerful. With Aristotle, you get everything put into these trees, animals, the best example, maybe, where to be A bird means to be like these other things, but also to be a particular bird, which means you're not like other members of the category. A very sophisticated sense of similarity and difference that enables you to build a structure that represents the entire, that captures the perfect order of the entire world, in which every thing has a single place. And furthermore, that place is determined by a strict definition that tells you what goes in this category and what doesn't. And to know, for Aristotle was to know what these definitions are. Without these definitions, there's no possibility of knowledge. This is what knowledge is about. So it depends upon having a view that everything is perfectly ordered and ordered by perfect definitions. And that's the view that we are rapidly moving away from, especially in the digital world. We love these trees. These trees are just so awesome and so useful that we use them all over the place. But the, I guess, dirty little secret about these trees is that let's say you're putting away your laundry, okay? This is the clean laundry. So what you do is first you divide it up, you lump and you split. So you have a big lump and you divide it, you split it up into smaller lumps by person, typically, and then typically by body part and so forth until you're ready to put things away. And one of the. One of the consequences of this, of the process, is that when you are filing, you've got the new pair of socks. You have to decide which of two piles it's going to go into. And you can use whatever system you want. You know, divide sports socks from dress socks or winter socks. That doesn't matter. You have to be binary about it because that's how atoms work. They have to go in one place so you make your decision. So after you've finished this process, if you were to map the decisions you went through, you would have a tree. You have to, you have to make these binary splits. First you divide it by person, then body part and so forth. These trees represent. The trees we've been using for organizing our ideas are in fact what you get following the limitations of organizing real things. Everything has to go into one or the other pile. Limiting thought, the organization of thought by the limits of how you put away your laundry is a terrible limitation that we've suffered under because we've represented knowledge, ideas we've had to in physical objects. We've had to file the physical stuff. And so we've absorbed into our idea system the very limitation that's on the physical. This is a terrible, terrible limitation. Libraries, God bless them, but these are just as reality is designed to keep things apart, libraries are inevitably not designed to, but have to keep ideas apart. So if you come across and books have links in them, footnotes. If you come across a footnote here, if you're in the Harvard Library, which this is not, this is a much flatter library than the Harvard Library, the Widener, the great Widener Library. If you come across a footnote and you want to trace it, you know you're going to be traveling down a hall, going up an elevator across, because there's a bridge down the bridge, through the tunnel, through the literal damn tunnel to get to the next section of it, where you'll find the book. Except it won't be there. It's either out or it's in a different library, one of the 90 libraries. It's as if libraries were designed to keep ideas apart, whereas online you'd follow it like that. So we've managed to internalize the representation of the organization of ideas using the limitations of the physical. And the easiest guy to kick on this is Melville Dewey. It's just. I didn't even want to. But it's irresistible. The Dewey decimal system. Who 90% of US public libraries and public school libraries use it. 1876, he was a 21 year old punk. He was 21 years old. He hadn't even totally graduated from his orthodox Christian isolated college in the middle of Massachusetts at tiny provincial college. And he decides he's going to organize all knowledge. And so he did. He was an extreme rationalist. He was a spelling simplificationist, he was a metric person. All of which are. This is if you think rationalism is going to solve all Problems. That's what you do. So he spelled his name like this for a while, and then when he became the first librarian at Columbia, he changed it back to Dewey. He was slightly crazy. So he liked decimals. He just loved tens. He acknowledged this. He just loved tens. He loved tens so much. And I'm not making this up, that he would try to arrange travel so that he would arrive on a date that ended with a zero just made him feel better. He's crazy. This is where rationalism steps over into superstition and craziness. So he said 10. 10 is such a good number. Let's divide everything up into 10. We'll have 10 basic root categories, and each will have 10 subcategories. And each of those will have 10 subcategories. And so we have 1,000 integers to play with. Well, okay, you know, that's fine. And then you get an indefinite number in the fraction, so you can accommodate everything. So it wasn't completely crazy. Still allowed for that. Well, okay, so let's say you're going to build a new kitchen. You're going to design it, and you were doing, you know what, 10. I'm going to have 10 cabinets, and each cabinet is going to have 10 shelves, and each shelf is going to have 10 slots. Oh, that will be beautiful. So now let's put in my spices. I need to have 10 spices. I only have eight. Gee, I got the eight. The first eight were really easy. I need another two. What can I count as a spice? Oh, chocolate sprinkles their spices. Popcorn, salt. That's a spice. Well, okay, maybe it is, maybe it isn't. Chocolate sauce. Probably not, but you need to have 10. But that's exactly what he did for knowledge and for books. Have to have 10 of each. The requirement that this is a reflection of personal is apparent when you think about how your children would organize. If you told them, organize the kitchen at tens, they would have a different set of basic categories. So you see that immediately in the Dewey decimal system. So, for example, in the 1000s, philosophy and psychology and the 1000s, by the way, are the most important because this is hierarchical in three dimensions. It's very hierarchical system. So philosophy, the queen of the sciences. Here are the hundreds here. So the. The integers. Oh, my God. Paranormal, occult, parapsychology, dreams and mysteries, Divinatory graphology, predicting the future by your handwriting and phrenology, predicting character by the bumps on your head. This is still there. These are the top level categories. And just one other example, again, because it's Just too easy is religion, which is the second most important. It's the two hundreds where choose. We get our own integer.

Yes.

Okay, Islam shares a little bit. They updated this a few years ago. 297 is now Islam, Babism, and Baha'.

I.

Which is fine, except that there are many Muslims who do not consider those other two religions to be as serious as Islam. They're 19th century religions, so they don't quite have the pedigree, but nevertheless. And then Zoroastrianism. Yes, its own integer. And the Buddhists to the right of the decimal point, they didn't make it. It's just too bad, you know, they should have tried harder or gotten a better PR agent or something, because it's frank. It's a little embarrassing to be only part of an integer. So why don't they fix it? Well, they do continually update it and polish and get it better, but it still remains deeply broken. Deeply broken. Because there's really not that much point in trying to fix it. So let's say they spend. I don't know how long to redo everything and start from the beginning. And they redo it and they announce it. And all the librarians go out with the razor blades and they scrape off the white letters from the. From the spines of all the books and they put on the new ones. And within days, less than that, the disputes will begin because the bosn. Oh my God, the Shiites and the Sunnis either are or are not at the same level. The Jews for Jesus category, it's under the Jews. The Jews don't want that. It goes under the Christians. The Christians aren't crazy about that. All of the feminist and gender issues that arise. There is no possibility of agreeing on this. The librarians will be holding. They'll have to weld the razor blades onto their fingernails because it's never going to be settled. And it's not because the people who own the Dewey decimal system aren't smart and don't care about issues of equity. And quite the contrary. Quite the contrary. It's that we don't agree is the problem, and we never will agree. How we organize the world is a deep issue and we are just not going to agree on it. There is no single way of organizing the world. There wasn't for Plato. There can't be for Dewey. When you're organizing physical books, you got to come up with something, do the best job you can, but people are still going to be pissed at you all the time. All the time. Can't help it digitally, it's different because we're digitizing everything and that changes the basic set of rules. So I want to quickly go through four principles that change and then talk about some of the, I think, sort of larger implications of this beyond the Dewey Decimal System. So I think it's useful to think about there being three orders of order in the first order. This is the Bettman Archive, largest, most important collection of historic photos in the U.S. 11 million photos that were buried. It's 220ft below ground in the middle of Pennsylvania. To keep it safe, in the first order, you organize the things themselves. You put the pictures in the filing cabinets or put the books on the shelves. Second order. So you do the best you can there and you preserve the materials. Second order, you go through the airlock at the Bettman. I mean, literally there's an airlock because they're dropping the temperature in the first order collection. And you have the second order where you separate the metadata about the objects and very familiar way of doing it. There are lots of ways, but this is the most familiar where you take this rich information. So you're in a library, you got a whole book, you reduce it to a three by five card and you make good decisions because we've gotten very good at doing this. Library science, as you know, is pretty advanced. So you make good decisions about what information is going to fit on that three by five card. But nevertheless, you are excluding huge amounts of useful information because you have to, because you don't want. Just for physical reasons. Card catalog can't get too big. So and you get maybe two or three sort orders now instead of just the one author, subject, title, maybe four, but not a lot more than that because it gets unwieldy, because atoms just suck. In the third order, everything is digital. The contents and the information about it, the data and the metadata are all digital. And that changes the basic principles. Four basic principles of change. The first is that we are very used to the idea. If this is a physical store and you get a new camera, you're going to put it onto some shelf, probably on the camera shelf, maybe in the sporting goods shelf, maybe in the new arrival. But you're going to pick one. If it's a digital store, you can put it into every shelf you possibly can. Amazon's brilliant at this, by the way. Why? Because. Why not? It's free. More people will find it. They'll be happy, you'll be happy. So now a leaf can go on many branches, but this totally would shock Aristotle. Aristotle would say, you've just betrayed the principle of being. To be a robin is to be in this one branch, and the being on multiple branches, that is. That's chaos. Knowledge cannot exist in such a system. But nevertheless, we want to sell cameras, so we'll put them into many categories as we can. The second is that in the real world, messiness is a big problem because you can't find stuff online. Messiness in a couple of senses is really good. And in one sense, if you post something and there's so many links to it, you can't even follow them anymore. Your post was a huge success. That's fantastic. The messiness of these relationships accrete meaning. They are the accretion of meaning. And I'll come back to that. But that enriches in important ways the knowledge and information and ideas we're putting up. It's not so in the third order. Messiness is okay because you're not rearranging the actual stuff when you're making a playlist. You're not actually moving items around. You're just doing new metadata. And so you can do as much as you want. You never have to actually disturb the stuff. We can do layers and layers and layers on top the of of it. The third thing that changes is that we used to be pretty sure there was a difference between data and metadata, and now not so much. So it used to be that metadata might be. If you're looking for the. You know the name of the author, you can't remember what the book was he wrote. So you type in Herman Melville and you press the search button on some search engine, you know, whichever one you prefer. And. And you'll get back not just the name of the book now, because everything's digital, you'll get back the actual contents. Thank you for Google Print. You'll get back the actual contents. Okay, so now you have a different problem. You remember some content, you say, what was that book? It has the phrase, maybe the first line in it. Call me Ishmael. Wonder who wrote that? And so now you search on that and you get back the authority. And what this means is there is no longer any distinction between metadata and data, except sort of this functional one, that metadata is what you know and data is what you're looking for. And you use the metadata to pry up the data. Everything is metadata. And this matters because we use metadata. We always have to locate the stuff we're looking for, to use what we know, to find what we don't know. And if everything is metadata. Now, if everything is a lever, we just got way smarter than we were 10 years ago. The ability to find. Okay, so I'm tempted to repeat it again because I think to repeat it because I think it's an important point, but I don't want to repeat it. So that would be bad for all of us. And the fourth thing that changes is that we're very used to the idea that, well, so we go into a store, clothing store, and the rational thing to do is to go through and to pull off all the clothing that's in your size, because everything else is just noise. So you make, go through with your shopping cart and you start making this pile, you'll be thrown out in 90 seconds. Just get out. But if you go online and they do the equivalent of what happens to you in a real world store where they make you walk past aisles of stuff you don't care about, it's the wrong gender, it's the wrong size, it's the wrong season, you would leave there in 30 seconds. And so what's changed is it used to be that the people who owned the stuff also owned the organization of it. Now we do, we own the organization of it. And so obviously one of the most fertile fields around is developing the tools by which we, the users, get to organize other people's stuff. One of them is faceted classification. Are people familiar with that? Very briefly then, just because not everybody nodded. Faceted classification, this is North Carolina State University's library, allows you to, as a user to browse in a tree like fashion, because that's a very convenient way of browsing, except you get to pick what's the root and what's the next branch. So if you want to browse first by century and then by the author's place, country of birth, and then by male or female, you can. And if in the next instance you say, you know what matters to me, I want to separate, I want to browse only male authors. You browse first by gender and then you browse by year or by language. You get to construct the tree. So it has the virtue of trees without the inherent limitation of trees, namely that you have to browse the tree the way somebody else thought you would want to. So this is being used all over. This is Newegg, which has a very nice guided navigation system, as it's called, but it's a faceted system. IBM uses it to put together teams of consultants very quickly. This is a good business example because they save millions of dollars doing this instead of Just using a flat of using a database to do it. And there's a company that makes engineering parts. Physical engineering parts have 25 million parts, and they have a faceted system with thousands of facets of faceted classification. Yeah. So this stuff actually works. And it's showing up very rapidly. It's showing up in lots of. Of places, and it gives the control back to the users. Perhaps you've heard of this thing called tagging, where obviously you get to tag, in this case, web pages, but all sorts of things are being tagged now with the tags that matter to you. And obviously the important part of it is not only does it help you locate the stuff that you've tagged, which is a handy tool, but in so doing, you're doing it publicly, it turns out. And so everybody else can also browse by the tag. So in other words, you can have the rest of the world do research for you. You can see what everybody else. You're interested in design or in robotics or whatever. Obviously you can see what pages they've turned up. Experts emerge. And because we're an insanely social species, we start noticing who's been contributing tags and then we talk with them and we fall in love and we get married. We're just, you know, that's pretty much what we do. You can tell a lot about a person buyer tag. Cloud tagging is. I'm very enthusiastic. I think it's very important. There are big practical advantages to it. I'm skipping over all of that because I assume it's all familiar to you. Okay. Nevertheless, there's also a side of it that I think should be admitted to in public, which is one of the reasons, I think, that folksonomies and tagging has caught on with such fervor among many of us, is that it is a little bit way of sticking it to the man. We get to create our own categories at long last. We don't have to. You know, the postmodernists have been reviled for saying things like, authors are not the best judges of the contents of their work. The author's intention doesn't count for that much. Well, you know what? They're right. And tagging shows that it's the reader's decision about what the book is about that really, really counts. And the author gets to contribute one tag, but the rest of us get to contribute our own tags. And there is something satisfying about that. So it seems to me, not at the end. Soon, though, soon, it seems to me that we're going from thinking that the universe is organized in one way and preferably into this beautiful tree of perfect categories to pulling down all the leaves, making this huge pile of miscellaneous stuff, enriching it with metadata as much as as we possibly can. Some generated by the owners, most of it generated by the users and changing the basic idea, which used to be you want to exclude all of the crap, because who has time? So we have experts who filter and show us what we need to see and they organize it into categories for us, which takes experts to do. That's how we used to do things. Now I think the idea is the best strategy in most information instances. Include everything. It costs more to delete stuff than it does to save it, which you know from looking in your digital camera folders, the endlessly DSC001175 that JPEG. You don't even know what's in there. In order to go through and delete that, you got to look at the pictures and make decisions. It's just easier to preserve than to delete. So capture everything. Capture the I can't believe I'm making another reference to Paris Hilton Hilton's boob job. But capture that too, because even though it is by any measure trivial, you're certain we can be certain that within five years, 50 years, there's going to be a graduate student doing a dissertation on the media's treatment of Paris Hilton, of trivia. And if that stuff is not included, we've lost it. So include it. Why not? So include everything and postpone the moment when we taxonomize, when we organize it, until the user does it herself, because the user has some particular interest, which you simply nobody can anticipate. So we are rapidly heading towards mash ups, right? I mean, maybe you've heard of them where we do things at a real estate site like capture everything, including of course, what's available, the school systems, the crime reports, the Republican versus Democratic areas. This is Massachusetts, where the Starbucks are, because that may matter to somebody, or where Starbuck free zones are. So even if you go four miles offshore, somebody's going to want to know where are the graveyards because they got some weird fetish thing going on or because they do the etchings, what do you call them? The rubbings. Which also sounds fetishistic when you think about it, but that's their hobby. And somebody else is going to want to know where the good dog parks are and where dog owners are, and somebody's going to want to know the intersection for God knows what reason, but somebody's going to Want to know that Flight paths are obvious, but most people are going to not want to live under them. My old thesis advisor, Thomas Langan, the late Thomas Langan, he liked to sit out in his backyard in Toronto. He's actually pretty obnoxious, to tell you the truth. And he'd sit there with his binoculars and you're trying to talk about your dissertation. And he would. Every time a plane went over, he wouldn't even use the binoculars. He would say, oh, that's the left. That's flight 1127. It's on its way to Paris. It's a little late. It's a 7 27. They must have swapped it out for the buy engine. So he wanted to live under the flight path. That's my point. I'm sorry, I'm getting lost. You can't know what people are going, what their interests are. So include everything and let us sort it the way we want to. So, three areas. I think this matters. I am going to go like the wind. First is simplicity. We've been living in an age for a number of reasons, including broadcast, that has intensified the emphasis on keeping things simple. So the President makes a speech, and you know that every presidential speechwriter says the same thing. Keep it simple. A year ago, President Bush gave a speech on immigration. Terribly complex problem. Keep it simple, keep it simple. That's what the speech writers did. 2,400 words later, he was done. Not a bad speech. Well, not a bad speech. Typical speech. Don't go there, don't go there. So within an hour, a couple hours after, afterwards, I checked at Technorati and there were 20. Oh, sorry, faux pas. No, no, no, no. Okay. I checked at Technorati and There were over 2,400 blog posts about this speech. So it's like there were one per word. And without looking at all of them, you know what? Pretty much what each one did, which is, you know, they look at the speech and they say, oh, interesting. Did you notice that he changed his position from when he was governor? Huh, didn't know that. Do you notice that this contradicts what he said? This Arnold Schwarzenegger is saying something different. Oh, that's interesting. Bloggers do this thing where they take a simple object and they find what's complex about it. And that's not so unusual because that's exactly what we do in conversation. In every conversation. That's what we do. We don't just sit there and say, yes, I agree. Now I want to agree further with you. That's interesting. Let's agree. We find what's different and interesting. We make things complex. And now we can do this on an order of magnitude we've never been able to do before. And we're thrilled. That's one of the reasons why people, I believe, are so thrilled about blogging after being treated like freaking morons by the media for 100 years, in order to get the economics of reaching the maximum number of people, the mass audience. At long last, we get this moment of revenge where we get to be interesting and complex again. So I wanted to come back briefly to Aristotle. Aristotle, who was an enormously complex thinker, by the way, wouldn't want to slight him for a second. Incredible genius. Just amazing. To know is to know what the simple definitions, definitions that cut as cleanly as a knife. Well, Eleanor Rosch, who is cognitive psychology at UC Berkeley, has a very different idea, which is prototype theory. So this is going to be fast, really quickly. Her idea is, what does it mean to be a robin? Well, it doesn't mean to be perfectly defined what a robin is. In fact, if I. You can all recognize robins, but if I asked you to define them, you probably couldn't give me a great definition. Nothing personal. You probably can't. I couldn't. You know what skipping is? I've never even heard anybody give a good definition of skipping. Just save that for the Q and A. So what happens? She says, well, you know, you know, robin is a bird because your parents, when you were young, pointed at and said, oh, birdie, birdie. And that became your prototype. And they pointed a sparrow and said birdie and you said birdie, and those became your prototype birds without a definition. And so over the course of your life, you're surprised as a child to find out eventually that a penguin is a bird, but you accept that it's not a very good example of a bird. Somebody wanted to know what a bird is. They genuinely don't. You wouldn't point to a penguin, you'd point to a robin. And so she says, this is how we learn. And this is. This is how we organize our world through prototypes. And some are good examples, some are bad examples. Is this a blog? I'm reading this thing. I'm not sure it's a blog. What is the definition of a blog? I don't know, but I know who some really exemplar bloggers are. You want to know what a blog is? Look at this one, look at that one. And this one doesn't have comments, so maybe it's not a perfect example of a blog, but it's still a blog. It's just not a great example of one. This, by the way, echoes stuff that Wittgenstein did. So I'll just tip of the hat to Wittgenstein. So the reason this is interesting within the digital world, I believe within tagging, is that Joshua Schachter of Delicious says something can be 73% of a category. Not for Aristotle, you can't be. But you can if you tag stuff. That's what folksonomies do. They're not simply replacements of taxonomies. They're representation of a wave front in which things can be 73.3percent of the category. That's important. That's not how we thought our world was defined. Okay, second point out of three, homestretch here. If you go to Wikipedia and look at tomato, there's a section of the tomato article on tomato versus tomato. You know, it's something we're interested in. So I want to imagine. And of course, there's a discussion page about the tomato versus tomato. So I want to imagine that three people contributed to this, and I'm making up the example, so you have to go along with me. The first person, by the way, it's Aristotle, is the world's greatest expert on tomato versus tomato. And he writes up the piece. He puts some time into it, gets it right. Second person comes along, changes it. First person says, greatest expert. I'm reverting it back to where I was. I mean, I know what I'm talking about. I don't even know who this person is. Third person changes it again. First person says, I'm not playing this game. You keep changing it. World's greatest expert genuinely is my example. So I'm leaving. Which is absolutely his or her right, of course. Nevertheless, that means this person has now become functionally irrelevant. World's greatest expert, irrelevant. Meanwhile, it seems to me this person's ideas would be improved by entering into conversation, even if he has to correct some moron. And very likely, these people aren't entire morons. Bugs get driven out of ideas through discussion, through the public negotiation of conversation. That's where knowledge is. So a really clear example of this is you're on mailing lists. I've been on some mailing lists for 15 years, maybe more. Some of them are moderately specialized things I care about. And on those lists, I find out what the news is in the area. And I see people who know way more than I do talking about this new stuff, arguing amongst themselves. These are experts in conversation. It's very clear that the mailing list itself knows more than any of the experts on it do. And this is a really, really good thing, this public negotiation of knowledge through conversation. One of the consequences of this is that we end up having to let infallibility into our notion of knowledge, which has been driven out until now. Well, pretty much. So Wikipedia is very happy to have these notices stuck in. They have over 100 of them. You can make up your own if you don't like them. These are all the ways encyclopedia articles can go wrong. It's not neutral. Contradicts an article, contradicts itself, isn't worthy of an encyclopedia, doesn't cite sources, reads like an advertisement, reads like a sermon, contains weasel words. They're very happy to put these. They encourage you to put up these notices saying, this article isn't good enough. Don't fully believe this article. And because of that, we know that Wikipedia is on our side. It's very. This becomes more credible because they're willing to admit their lack of credibility, their lack of authority. It's not trying to convince us that it's the world's greatest authority. It's trying to help us know. And the metadata here, the fallibility, metadata, is crucial to that. The question is why you will never see this here. And it's because these sources have a vested interest in appearing authoritative. Third and final. This is the philosopher Martin Heidegger, 1927, in the early Nazi years. We can talk about that later if you want. If it's not clear, I'm a Jew. I'm not crazy about Heidegger as a Nazi. I did my dissertation on him because I especially. Because I think, especially in what I'm about to talk about, I think he's right, and it's a fairly simple idea. He asked, what does it mean to be a hammer, to be a thing. Take a hammer as an example. And he doesn't mean if you're a marsh. He means on this planet. What does it mean? He says, well, to know what a hammer is, as a hammer, you got to know about nails. Just be weird. You'd say it's for hammering nails. But to know what nails are, you really have to understand wood. And how would except. And to know about wood means you have to know about lumber. Lumber, trees, trees, forests. You need an economic system, you understand, for connecting these things. Is that how it works? You don't really understand forests unless you understand they grow on an earth, and they do so in a sky with a sun. And all of that is what it means to be a hammer. A hammer is its place in this referential context of significance, of meaning. A Semantic Web, a real Semantic Web, the one that we live in, that's called the world. That's what it means to be a hammer. So for the moment, pretend that's true. I actually think it is true. Many have pointed out that we as a species externalize functions of consciousness. This is how we advance. So we externalize memory in writing and knowledge and books and arithmetic and calculators, and that's a really good thing. So what I want to speculatively advance is the idea that we're now externalizing meaning, this connection, this rich implicit connection of significance, of semantics. And we're doing this every way that we can. We do it every time we tag publicly. We're adding to this new infrastructure of meaning that we've never had before, because these tags connect things and are publicly available. Taxonomies. We still need experts, we still need taxonomies. They're incredibly valuable. Every time you do one, you are adding to this miscellaneous pile of stuff through the Semantic Web, writ small, writ large, and it's pointing out, building these connections through blogging and everything that we post, we're making connections between the pieces and we're contextualizing them. Every playlist, whether it's for music or at Harvard, there's a playlist for syllabi. Playlists are also powerful ways of making connections, of drawing the pieces together through sites like Digg and Reddit, which are doing it for news, for every. Fundamentally, every hyperlink makes a connection of meaning between pieces, and it's cumulative. We'll be doing this for generations, building this rich layer of meaning that we can then draw upon. And the most important thing about this is that this is not being done for us. It's not being done by experts. It's not well vetted, it's good information, it's bad information, but it's ours. It's what we're doing for ourselves. If you wanted to know what we're interested in, you shouldn't look at the Britannica. You should look at Wikipedia, all 1.7 million articles in English, and for the deep fried Mars Bars article and the history of umlauts and heavy metal bands for all that stuff. It's a better reflection of what we're interested in than the New York Times, than the Britannica, than any controlled source can be. This is the natural reflection, and it's ours. We are building it for ourselves. It's up to us to decide what to do with it. But the fact that we are building this infrastructure of meeting seems to me to be a generational task and of tremendous importance. Thanks. God, I hate myself. Yes, you mentioned that it's to the experts benefit on which Wikipedia to talk to the less expert people who are editing. But there's a severe one to many problem in which the expert winds up spending all his time teaching, let us say, Sociology 101. Yes, it's tough to be an expert approaching Wikipedia because as you say, you may be tutoring constantly and then permanently guarding, checking back, getting notifications of when edits are being made. So you may not. That may not be your role as an expert. That may take another expert or somebody who cares deeply about it to patrol it and to keep it at the level at which you've left it. So far it seems to be working pretty well at Wikipedia. Wikipedia seems to be a pretty good resource. It may. Who knows what will happen over time? And it may be there'll be a type of expert decay as experts lose interest in it. There's also been talk about. So I'm still waiting for somebody to do another layer on top of Wikipedia, not on the site, but that's for example, the Lord's Wikipedia, that's just metadata that points to the articles that it finds acceptable, or the American Psychiatric Association Wikipedia that points at the articles at the state, the particular version of an article that they think is okay, and pull all that together. So maybe that's what will happen. Also, we'll get frozen pages frozen by meditating metadata and authenticated by people who we trust. Don't know, but maybe. Oh, okay.

Hey, one of our products is the Google Toolbar, and we have constant debates about what order to put the buttons in and how to group them. You can group them by what they do or how often they're clicked on, or what the user wants them to do or whatnot. And this comes up and we always end up making the generalization that users don't spend the time to categorize things. They're better off having them pre categorized because we can measure how small of a percentage of people actually bother to make categorizations, even though we're kind of nerds and we like to tweak things. So we all think that everybody's going to want to rearrange the buttons. And one of our toolbars has that capability. But it was just interesting listening to your talk about whether things should be categorized based on function or how often they're used, or what what do you think is the right answer in terms of like making people categorize things or having them pre categorized by an authority to make things simple?

I guess this is an intensely practical question which should be addressed to somebody who's intensely practical. It does seem to me that there's a sort of 1% rule at work in a lot of categorization, including perhaps tagging. Although there was a Pew study that gave incredibly high figures for the percentage of people who were tagging. It seems quite plausible to me that 1% of a population that does the tagging and categorizing is sufficient to add enough metadata that everybody benefits wildly from it. So it may be that 1% or fewer actually bothered to rearrange the buttons on the toolbar. But if those buttons, if those. The toolbars they created were publicly available, as a second. You have to do something for default, of course. But as a second possibility, a third possibility, some people might, a small percentage of people might find that really useful. A larger percentage than the number who actually redesign the toolbar.

Hi, thank you for coming here. Thanks for having me. I noticed in the book you mentioned two writers, one Borges, who I also.

I love, you can't write a book about taxonomy, although we don't say it's about taxonomy, but you can't write a book about taxonomy without using the Borges looks like flies from a distance, which. A line that just floors me every time I read it.

And. And the other one you mentioned is, I think, is it Michael Gorman?

Oh, God, yeah. Oh, Jesus. Yeah. President of the American Library association for a brief time. Oh, yeah. Okay, good.

So anyways, I mean, basically sort of my question there is. Borges wrote about all these issues over half century ago, and yet we're still having these wars. How. What sort of reconciliation or where do things go where the people who like little buckets can get along with the people who don't?

Well, so when Borges was writing, it was science fiction. Imagine, if you will, an infinite library. Imagine, if you will, that you. Now, we have very intensely practical reasons why we want to be able to get the information that we want. We have social reasons. And this is the imponderable. I think this is not simply a task for librarians who want to divide up knowledge, because the Internet is not even primarily, from my point of view, about information. It's about social connection. And so we have people who, for deeply social reasons, want to divide up the world in a way that makes sense to them and connect with people who do the Same. So the percentage of people who are actually, actually going to think about buckets and how can I label my buckets and how can I arrange. It's very, very small. The percentage of people who want to be able to find all the email at Gmail that talks about that's very large. That's all of us. The percentage of people who want to be able to find the photo of Aunt Sarah when she was on the beach, and they can't because they have 10,000 photos on a desktop that will be all of us within just a couple of years. Years. They're not thinking about this as taxonomy. Why should they? They're thinking about how they're. How things cluster in a meaningful way in their lives. And the tools that we give them, that enable them to find the meaningful clusters are the tools that have and will succeed, the ones that let them do the task of metadata. Oh, my God, no. We have lots of history that says that lets us know that people do not like dealing with, with explicit metadata. We need no further evidence of that. But I'll give you some Microsoft word, at least 10 years ago, when I checked, they have a metadata sheet, right? I mean you can, for every document you can fill in the form and it's very, very useful for finding the documents. Again, it's a low single digit percentage of people who do that. It just feels like red tape. Lots more people are willing to tag stuff. So percentage of people working on buckets. Tiny percentage of people who want to organize into clusters. All of us, I think. Okay, I'm not sure you're satisfied, but. And I think that wraps it up. So thank you very much for coming. Oh, thank you so much for having me.