Categories - Intertwingled: Information Changes Everything (2014)

Intertwingled: Information Changes Everything (2014)

Chapter 2, Categories

Do I contradict myself?

Very well then I contradict myself,

(I am large, I contain multitudes.)

– Walt Whitman

Sitting on the floor, legs crossed, eyes closed, I watch my breath. As the air flows in and out, the unity of mind, body, and environment is clear. I dwell in the moment, calmed by concentration, inviting mindfulness. I’m aware all formations are impermanent, more sandbar than substance, and that the essence of life is dukkha, suffering, unsatisfactoriness. Inhale, exhale, the “I” fades into anattā, there is no self. The mind is at peace and open to insight, wisdom, nibbanā. In, out, a glimpse into the process of perception, a sense of interbeing, breathing towards an understanding deeper than words. And then my alarm interrupts the practice. My fifteen minutes are up. It’s time to get back to work. This book won’t write itself.

So, what’s Buddhism doing in a chapter about classification and its consequences? Well, for starters, Siddhārtha Gautama, the person who became known as the Buddha, was an information architect. Living in India, two and a half thousand years ago, he rejected the rigid hierarchy of the caste system – the fourfold division of persons into brahmins, rulers and warriors, farmers and traders, and servants – and embraced universalism, believing enlightenment is open to all.xx Then, he shaped several new taxonomies, including the three marks of existence, the four noble truths, the five hindrances, and the noble eightfold path. Of course, the deepest, most difficult ontology Buddha taught is anattā, non-self.

This notion there’s no self – that while we’re more stable than a tornado or a sandbar, we belong in the same category – is counterintuitive and disturbing, particularly to those of us in individualistic Western cultures. Self as process not substance is outside our model of the system. So while we may try samatha, calming through mindful breathing, we’re less at ease with vipassanā, which aims for “the permanent and radical transformation of your entire sensory and cognitive experience.”xxi Consider this explanation of its benefits.

You search for that thing you call “me” but what you find is a physical body and how you have identified yourself with that bag of skin and bones. You search further, and you find all manner of mental phenomena, such as emotions, thought patterns, and opinions, and see how you identify the sense of yourself with each of them. You watch yourself becoming possessive, protective, and defensive over these pitiful things, and you see how crazy that is. You rummage furiously among these various items, constantly searching for yourself – physical matter, bodily sensations, and emotions – it all keeps whirling round and round as you root through it, peering into every nook and cranny, endlessly hunting for “me.” You find nothing. In all that collection of mental hardware in this endless stream of ever-shifting experience, all you can find is innumerable impersonal processes that have been caused and conditioned by previous processes. There is no static self to be found; it is all process. You find thoughts but no thinker, you find emotions and desires, but nobody doing them. The house itself is empty. There is nobody home.xxii

If you found this passage odd, foreign, threatening, then you’re ready for the upwind message we’ve been tacking for. Classification is as deep as it gets. It’s what binds us and separates them. Understanding and behavior are rooted in taxonomy, as are religion, philosophy, reason, and ethics. Ontology is behind our senses of fairness, risk and reward, even visual perception. Categories are the cornerstones of cognition and culture. That’s why it’s so hard for us to grok Buddhism. We’re weird. And by that I mean western, educated, industrialized, rich, and democratic (“WEIRD”).

It is not just our Western habits and cultural preferences that are different from the rest of the world, it appears. The very way we think about ourselves and others – and even the way we perceive reality – makes us distinct from other humans on the planet, not to mention from the vast majority of our ancestors.xxiii

Of course, when we’re not consumed by the pursuit of happiness, we too struggle with the ontology of existence. While we may have inherited the mind/body dualism of our favorite reductionist, René Descartes, who concluded the mind (or soul) can exist without the body, we need not stay bound by what Gilbert Ryle called “the ghost in the machine.”xxiv

In recent decades, the countervailing framework of embodied cognition has built momentum with respect to empirical research. This thesis holds that the nature of the mind is largely determined by the form of the body. Unlike computationalism, which views the brain as a central processing unit with inputs (sensory) and outputs (control), this theory of mind recognizes that how and what we think is shaped by the body’s systems of perception, action, and emotion. Our bodies constrain the nature and content of our thoughts, and cognitive processing is distributed beyond our brains. In short, cognition isn’t just in the head.


Figure 2-1. Embodied cognition.

Furthermore, according to the related theory of extended mind, thinking isn’t limited to skin and skull. Cognition is shaped by and extends into the surrounding environment. When we use a pencil to sketch ideas, the pencil becomes an extension of our bodymind, and the marks we make change the course of thought. We literally think on paper. In the words of cognitive scientist Andy Clark, human cognition includes “inextricable tangles of feedback, feed-forward, and feed-around loops: loops that promiscuously criss-cross the boundaries of brain, body, and world.”xxv

Our tools, like our bodies, become “transparent equipment.” We see through them to the task at hand. Brain imaging studies have shown that as we build fluency, we incorporate tools – pencils, hammers, bicycles, words, numbers, computers – into our bodymind schema. Then, in accordance with the principle of least effort, we strategically distribute work through the whole system of mind, body, environment. We use calculators for math. We offload memory to contacts and calendars. We rely on Google for retrieval, so there’s less need for recall. And when we play Scrabble or Tetris, before we see a solution, we move the tiles with our fingers, because it’s faster than modeling those shifts in our minds.xxvi


Figure 2-2. Extended cognition.

Embodiment is heady stuff, so let’s explore a basic example. How about the colors of a rainbow? In school we learn the spectral colors of red, orange, yellow, green, blue, indigo, and violet are produced by the light of a single wavelength, and all are visible to the human eye, except for indigo, which Isaac Newton added so the number of colors would match the number of planets, notes in a major scale, and days in a week.

This all makes sense, sort of, until you learn that in Japan, people say traffic lights are red, yellow, and blue, even though ‘Go’ is green.xxvii The distinction gets lost in translation since until the twentieth century, Japanese had only one word, ao, for both blue and green. It wasn’t until 1917, when crayons were imported into Japan, that midori, which began as a shade of ao, was redefined as a new category, green. This split left scars, which is why apples, novices, and traffic lights are blue.

Interestingly, cross-cultural studies reveal structural similarities behind these colorful distinctions. In the late 1960s, researchers discovered that while the number of color categories varies from two to eleven, there’s a common path that languages follow towards increasing specificity.xxviii


Figure 2-3. The evolution of color.

In languages with two, it’s black and white. Add one, it’s red. Next up, green or yellow. For six, green splits in two, creating blue. You get the picture. What’s odd about this conformity across cultures is that the spectrum is continuous. There are no seams in a rainbow, and yet we see them anyway. This illusion is based in biology and refined by language and culture. The red, green, and blue cones of the human retina place a hard limit on the visible spectrum. Within that continuum, the seams we see were teased out by evolution, enabling us to distinguish food, water, and predators from their surroundings. And language is layered on top. Once we have words for colors, it’s near impossible to un-see the seams.

When categories are absorbed in a culture, they become nearly as irrevocable as they are invisible. They retain power even when the ideas behind their conception are obsolete. For instance, while Cartesian dualism has been dismissed by modern philosophers, neuroscientists, and physicists, the system of Western medicine remains organized around mind/body reductionism. Doctors specialize in physical problems. Psychiatrists focus on mental disorders. In orthodox medical practice, the mind-body connection is a missing link.

I learned this myself in 2005 when I suffered terrible, chronic back pain while juggling a heavy consulting load and writing Ambient Findability. I first blamed poor posture and bought an ergonomic chair. It didn’t help. After weeks of agony, I visited my doctor. Ignoring my hint that stress could be a factor, she prescribed physical therapy and three Advil, three times a day. I followed doctor’s orders, but the pain got worse. In desperation, I went to Google with “back pain stress.” I found and read a book, Healing Back Pain: The Mind-Body Connection by Dr. John Sarno. He argues that many musculoskeletal pain disorders are rooted in repressed emotion. To distract us from anxiety, our autonomic nervous system reduces blood circulation to specific muscles, tendons, or ligaments, thereby causing oxygen deprivation and severe chronic pain. For treatment, he suggests that patients acknowledge the psychosomatic basis and repudiate any structural diagnosis. This means no pills, no physical therapy, and resumption of all normal activity. It sounds a bit odd. But you know what? It worked. Completely. My back was healed by a book.

This experience made me see the missing link, the mind-body connection, that I’d never known I was missing. It challenged my theory of existence and raised prickly questions about our culture. Suddenly, I was unable to un-see the weirdness of Western medicine. Science and technology bring us medical miracles, but success blinds us to the dark side. The over-prescription of drugs and surgery is an epidemic. We’re fixing things that aren’t broken, and the cost is astronomical. For instance, seventy percent of us suffer severe back pain, and in the U.S. alone this results in tens of thousands of surgeries a year, but the herniated, ruptured, and bulging discs commonly attributed to back pain are seen almost as often in the MRIs of healthy people.xxix All too often, diagnostic classifications are made based on visible but harmless imperfections by doctors who are blind to the invisible but powerful connections between mind and body.

Of course, we’re not all in the dark. Many doctors regularly prescribe placebos. They trust in the efficacy of mind over matter. And the market for complementary and alternative medicine – over $35 billion spent out of pocket each year – shows some patients grow impatient with medical orthodoxy. But mostly the $3 trillion healthcare industry rolls on.


Figure 2-4. Complementary and alternative medicine.

There are many reasons why medicine is a mess. Patients want a quick fix. Doctors hate saying “I don’t know.” And the truth is obscured by industry-funded research, advertising, and a government beholden to special interests. Marcia Angell, former editor-in-chief of the New England Journal of Medicine states “It is simply no longer possible to believe much of the clinical research that is published, or to rely on the judgment of trusted physicians or authoritative medical guidelines.”xxx

Considering the source, that’s a scary conclusion, but ending fraud won’t solve the problem. Ontology is a root that runs deeper than corruption. The separation of mind-body is a category error that’s hard to undo. In the 1600s, Descartes set out to validate his mechanical philosophy – the body is a machine made of parts – and to prove the existence of an immortal soul, so he wouldn’t risk heresy charges from the Catholic Church and the Inquisition. Centuries later, our culture and language are subject to reductionism and dualism. The consequences of classification expand and endure.

That’s why the origin of our work is ontology. Organizing for users isn’t just about findability. In designing taxonomies and vocabularies, we serve as architects of understanding. We shape how users view the business, the topic, the task. For better or worse, our groups and labels endure across channels and platforms. Mega-menus designed for desktop get stuffed inside mobile hamburgers, and users are stumped. Departments conceived for retail store layouts are mapped to navigation menus on e-commerce sites, and luggage gets lost.

Our work is hard to undo, so we must resist the urge to rush. While our colleagues may run screaming from the abstraction and ambiguity of this ontological inquiry, we must have the courage to dwell in discomfort. Time spent wisely at the start of a project or journey may return dividends for years.

Of course, the start is often too late, which is why we must engage with strategy and governance. Ontology begins with the org chart. In making frameworks for collaboration, we must think about goals, metrics, roles, and relationships, because how we organize ourselves changes everything. The categories we choose and the words we use to describe the project, program, process, product, service, or ecosystem will alter the path and destination invisibly and irrevocably. A digital strategy team is blind to physical touchpoints. A user experience designer ignores content creators. A search engine optimization project ruins the information architecture. Words are the interface, not just on the Web, but in our minds. As a wise woman wrote “Language as an articulation of reality is more primordial than strategy, structure, or culture.”xxxi

To avoid blind spots, we must see (and speak) differently, using averted vision to shift focus from center to beyond the periphery. Imaginative re-classification reveals invisible structures, unspoken assumptions, hidden values, and novel possibilities. But it’s not easy to invert the canon. Our biology, culture, education, and language all conspire to convince us there is a single, right way to organize things. Blue and green are distinct colors. History and science are separate subjects. Europe is above Africa. Books belong in fiction or nonfiction. The tomato is a fruit. Now turn the last five periods into question marks, then consider the contrary. Go ahead, give it a try. Like meditation, this intellectual yoga takes practice.

That’s what we’re doing in this book. By framing and re-framing, we build the mental muscles of curiosity and imagination, and we nurture our capacity to be cheeky, sassy, wise. Buddha opposed the caste system and dogma in general. He said “place no head above your own.” Of course, to question the categories of custom, convention, rule, and order is to risk your neck. Galileo was found “gravely suspect of heresy” for confirming the Copernican re-classification of the universe, Joan of Arc was burned to death for “dressing as a man” and Nelson Mandela was categorized as a domestic terrorist by South Africa and the United States for defying the taxonomy – black, white, coloured, Indian – of apartheid.

Mostly what we do isn’t quite so heavy. But it’s unwise to ask certain questions before understanding politics and culture. In all organizations, from libraries, nonprofits, and government agencies to Fortune 500s and Silicon Valley startups, visible categories are built on invisible fault lines. So speak softly and carry some Silly String, because the dark paths that wander betwixt taxonomies and org charts are riddled with tripwire. xxxii

Organizing for Users

Of course, since users are the center of our universe, it’s our duty to take risks on their behalf. And while we tend to talk about the visible leaves and branches of information architecture – menus, buttons, links, labels, tags, facets, search, navigation, personalization – categories are the root of all this work. They cleave concepts and channels together and apart.

In retail, the interfaces are different – store, catalog, website, app – but the categories are consistent across channels. This accord makes it easy for users to switch systems or devices while letting managers stick to the same old org chart. But, when implemented inflexibly, users get lost. For instance, where would you look for luggage in this menu?


Figure 2-5. Users (and luggage) get lost.

It’s a big question because suitcases and travel bags are high-margin products. In retail, luggage is among the most profitable categories. But our e-commerce client had it hidden under “for the home” because that’s where it lives in the store. Users were lost – analytics showed “luggage” to be the most common search term – and undoubtedly so were sales.

The problem of lost luggage was even worse on mobile. Unable to rely upon mega-menus to reveal category maps on rollover, our client served up the hamburger icon. Beneath this visible tip of the iceberg were their invisible products, hidden under multiple choices, clicks and categories.


Figure 2-6. Most of the iceberg is invisible.

Once in the store, shoppers persist. They ask, browse, and will even use a map. Online, it’s easy to shop elsewhere, so taxonomies must be tuned for findability. And it’s no good responding to desktop without adapting for mobile and tablet. Stuffing categories into a big hamburger will bring bellyaches to shoppers and sellers. Users can’t buy what they can’t find.

While findability comes first, we must also remember that categories are about more than retrieval. Classification helps our users to understand. Through splitting, lumping, and labeling, we reveal choices and invite questions. Hardsides protect, spinners roll, carry-ons fit, and backpacks are hands-free. Which features matter most? Which bag is best for you?


Figure 2-7. Categories reveal choices.

Of course, all taxonomies are imperfect, as is the language they’re built upon. Let’s say you want a messenger bag. Is that under backpacks or duffels? Or how about a lightweight, hardside carry-on with two wheels? Does that even exist? Like maps and myths, taxonomies hide more than they reveal. They bury complexity to tell a story, and they always miss someone out. Some things, like luggage, get lost by accident, while others – people, places, and ideas – are buried by design.

Either way, each glitch in the matrix subtly changes understanding and behavior, which is why this work often has moral weight. Classification has consequences, as Geoffrey Bowker and Susan Leigh Star argue in Sorting Things Out.

Each category valorizes some point of view and silences another. This is not inherently a bad thing – indeed it is inescapable. But it is an ethical choice, and as such it is dangerous – not bad but dangerous.xxxiii

Taxonomies are treacherous because the easier they are to use, the harder they are to see. We grab handles without scanning contents. We trust labels without knowing origin. In the Dewey Decimal Classification, the system used in public libraries and taught in public schools, of the 100 numbers for religion, 88 are reserved for Christianity. Islam and Judaism each get one, while Buddhism is lucky to get its own decimal point. xxxiv


Figure 2-8. Religion in Dewey Decimal.

What values are implicit in this scheme? What is the intent and impact? Who does it help? Who gets hurt? What are its alternatives? And why is this the one we use? Why does it endure? We must subject all taxonomies to such questions because their imprint belies their impact. Consider, for instance, the Facebook Like. It’s only a word, but it implies an ontology and shapes understanding and behavior profoundly.


Figure 2-9. Facebook hides more than it reveals.

Unlike Share or Retweet, Like nudges us into “friendly world syndrome.” We have a hard time “liking” bad news, so most of the sad stories simply fade away, leaving us in a safe, happy place that’s good for business. When we think about taxonomies, we tend to focus on whole systems like Dewey Decimal, but like Like, a single word can embody a worldview.

Words are interface and infrastructure. They are the handles that help us complete tasks and find content, but they are also symbols that represent concepts and categories. Scholars in semiotics and semantics have delved deep into the complex relationships between sign (signifier) and meaning (signified), and the extent to which that meaning is defined by intent or interpretation. One insight, extracted from the cold, dark depths of that rabbit hole, is that all words have baggage.


Figure 2-10. Words are handles.

Words are the interface. Hide the words, and most software, websites, and taxonomies are rendered unusable. But words are also infrastructure. They are the parts through which we understand the whole. In business, salesmen in suits are replaced by websites with words. The depth and breadth of products and services is defined by categories. In all types of organization, trust is built (in part) on taxonomy. Words that fit together are indicative of a strong mission, vision, and brand. While words play on the surface, they are accepted as tokens of substance. Experts use big words to garner authority. Politicians use small words to gain power. In taxonomies and poems, semantics is more than it seems.

Of course, half the time our words are wrong, and the consequences of classification are unintended. When we choose a word, it’s packed with meaning, but lots gets lost in transit. Consider the placement of Buddhism in Religion in Dewey. The interesting question is not whether the reality of half a billion Buddhists merits more than a decimal point, but whether Buddhism belongs in Religion (or Philosophy) at all.

The Buddha himself did not employ the concepts of religion and philosophy. These are our concepts, not his. Hence, to interpret his teaching as one or the other of these is to put it in a framework he would not have recognized. This need not mean the interpretation is simply wrong, but it may well be misleading.xxxv

While Buddhism has much in common with religion and philosophy, there are meaningful distinctions. Like religion, Buddhism has rituals, beliefs, and ethics, but it is not a system of faith and worship centered around a divine being. There is no God. Buddha was a man. And like philosophy, Buddhism appeals to reason and aims for insight, but its practice of meditation seeks an understanding deeper than words.


Figure 2-11. Buddhism isn’t religion or philosophy.

When we use the handle of religion or philosophy, we introduce baggage. And we can’t be sure of its contents, because the meaning changes from intent to interpretation. This is not a problem we can avoid. It’s the nature of language and categorization. But such ambiguity begs for awareness.

On one level, when we use words to make places, the map is the territory, the word is the thing, and language is the environment where experience and exploration occur. But on another level, that’s simply not true. Since understanding arises through the unity of mind, body, and environment, language can’t contain meaning, or to put it poetically, in the words of Hui-neng, the patriarch of Chinese Buddhism:

Truth has nothing to do with words. Truth can be likened to the bright moon in the sky. Words, in this case, can be likened to a finger. The finger can point to the moon’s location. However, the finger is not the moon. To look at the moon, it is necessary to gaze beyond the finger.

In the systems we design, our words do lots of pointing. They serve as symbols for categories, concepts, tasks, and content. Words are necessary. They help users find what they need and understand what they find. But words are never sufficient. Meaning is irretrievably lost in translation.


Figure 2-12. Words are only fingers pointing to the moon.

As information architects, the first step is awareness of the weakness of words. Once we accept the limits of language, we go beyond them. The same is true of taxonomy. Every single one is flawed. Once we admit the problems of ontology, we solve them. Classifications, like colors, exist on a spectrum. The objective ones – alphabet, numbers, geography – are easy to use but not always useful. The subjective ones – topic, task, audience – are useful but not so easy. In the white pages, we can quickly find a friend, but it’s impractical to discover a plumber. In the yellow pages, we can easily locate a theater, provided it’s not called a cinema or drive-in.

The first step in taxonomy construction is defining its purpose. What’s the goal? Who are the users? How will we measure success? But this isn’t a linear process. A taxonomy merits Agile, not Waterfall. To put objectives before ontology is good, but we must also pair classification with context. Where will the taxonomy exist? What parts will users encounter and when? Will they touch it on mobile? Will they see it on TV? A taxonomy is part of a cross-channel information architecture. To make the whole work, we need a build-measure-learn loop that gets how the parts fit together.

Simple metrics are seductive. We move t-shirts up a level and look for a sales boost. We test and refine within-category similarity and cross-category difference, and hope for customer satisfaction. But each taxonomy has many touchpoints. Categories appear in search results, filters, facets, menus, page titles and copy, product metadata, messaging, and advertising across channels. Plus, ontologies are embodied in org charts, and vice versa. We shape our taxonomies; thereafter they shape us. For all these reasons, the detailed work of taxonomy design must be informed by the big picture. In semantics and in health, to classify without context is gross malpractice.

Years ago, I consulted with one of Canada’s regional health authorities. On their site, the only way to browse was by bodily system. They’d used a taxonomy created for doctors and nurses to serve the public. And it didn’t work. In our research, most people couldn’t find diabetes. The taxonomy failed to meet the needs of its new audience.


Figure 2-13. A taxonomy designed for health professionals.

So we added new navigation, including common conditions, an A-Z list of disorders, a symptom checker, and sections especially for men, women, teens, and kids. We used multiple ontologies in service of the objectives.

When I worked with the National Cancer Institute in the United States in 2003, wayfinding within was my client’s objective. But I asked hard questions about findability via Google and added SEO into the mix. To serve people and search engines, we listed major types of cancer on the home page. And we enabled our human users to browse by bodily system, narrow by audience, search the full text, or consult an A-Z list. A decade later, the design has changed, the content is updated, but the information architecture remains untouched. Well-built structures stand the test of time. That’s why it’s so important we get ontologies and taxonomies right.


Figure 2-14. Navigating the National Cancer Institute.

Recently, when I advised Polar Bears International, this pattern unfolded once more. Their site relied solely on main navigation. Their top level categories – programs, research, education – weren’t bad. But there simply wasn’t enough infoscent for people or search engines to follow.

So after rejiggering the main nav, we added topic and format. Instead of a “browse by topic” link, we exposed its content, and the same for format. Very few users would click “browse by format,” but since everyone loves polar bear pictures, the image and video links became really popular. These changes led to a 39% increase in visits to the site the very next year.


Figure 2-15. Polar bears by topic and format.

The moral of this story is important. Due to the weakness of words, it’s hard for any label at the top of any taxonomy to stand on its own. We need the root categories of main navigation so users understand the full scope, and so there’s a place for all content, now and in the future. Breadth lets the system scale over time. But the top taxon is too abstract for users and the bottom too specific. The action is in the middle. So we must surface sample subcategories. Instead of burying them under Topic, we should bring out the Polar Bear Cubs.

We must reveal what cognitive scientists call “basic level categories.”xxxvi In a taxonomy, the basic level is the largest class of which we can easily form a concrete image. It’s hard to imagine furniture, but we can all picture a chair. Few people use “pinniped” or can distinguish harp from harbor, but like polar bears, we know a seal when we see one. At the basic level, we use the simple names of folk taxonomy rather than the terminology of scientific classification. They are the first categories that kids understand and have the most cultural significance for adults. Due to the idiosyncrasies of human psychology and perception, they are optimal for learning, recognition, memory, and knowledge organization. They are artifacts of embodied cognition and vital tools for design.

So far, in talking taxonomy, we’ve been dancing around a central point. Since every classification is flawed, we should usually use more than one. It’s a simple idea that’s hard to accept. We’re hardwired to believe there’s one right way to organize things. Dewey’s classification is a monument to that spirit. But we are getting better at providing multiple maps and paths, and it’s helping our users enormously. In universities, for instance, we’ve learned to complement the main navigation menu with alternate pathways such as audience, school, task, an A-Z index, and search.


Figure 2-16. Universities offer multiple pathways.

And in e-commerce, faceted navigation is nearly ubiquitous. In the 1990s, when we began talking about facets and Shiyali Ramamrita Ranganathan, each online store had a single taxonomy, and nobody knew who he was. They still don’t, but we’re all indebted to the mathematician and librarian from India who realized that a single taxonomy isn’t nearly enough.


Figure 2-17. Facets offer users a map of search results.

Faceted navigation serves up a custom map to search results that helps people understand what they’ve found. Users can then select filters to clarify and refine their query. Is this for a man or woman? Do you prefer black or unorthodox orange? A vegan wallet for $50? Yes, we have a few of those. Search is transformed into an iterative, interactive conversation in which users build complex queries, one simple step at a time.

And, in social, categorization is as easy as a #hashtag. Free tagging is descriptive classification without authority control. There is no hierarchy. Each tag is a category. Each object may have many tags, and vice versa. It’s messy but it works. Heck, #occupywallstreet launched a movement. And event tags like #barcamp are the timely ties that bind us together.


Figure 2-18. A tag cloud at LibraryThing.

Folksonomy has a light footprint, as it’s hard to see the whole. The glimpse we get through clouds isn’t nearly as satisfying as the view from the top-level of a taxonomy. But to fix on that contrast is to miss the point. Tagging flips the model. Rather than place each object in a hierarchy, taggers describe objects any way they want. Tag creation is idiosyncratic, bottom-up, and object-centered, and so are its use cases. The value of tags is realized in the strange connections (and descriptions) that appear once a user finds an object.


Figure 2-19. Three types of classification.

Each way of organizing has strengths and weaknesses. Taxonomy affords a view from the top, facets help us muddle through the middle, and tags build bridges at the bottom. As information architects we must define the right mix for each system by balancing value and cost. And we should be open to ideas outside this ontology of taxonomies, facets, and tags.

Netflix, for example, illustrates thinking outside the box. They had a genre taxonomy and myriad forms of navigation and personalization, but they decided to make the system better.


Figure 2-20. Netflix navigation and taxonomy.

Netflix invented a unique classification scheme that blends taxonomy, facets, and tags. They built an ontological model with over 1,000 microtags, and hired a team of taggers to describe 14,000 movies and TV shows. Then they designed algorithms and a grammar for stitching facets and tags together into a colorful array of 76,897 microgenres.xxxvii


Figure 2-21. A sampling of Netflix microgenres.

Users experience these microgenres as categories. They are easy to understand and use. But there’s no taxonomy. There are no subsets or supersets of Cult Evil Kid Horror Movies. Based on viewing history and interests, users are shown a few microgenres, each with several matching titles. It’s a great way to find movies that also invites introspection and understanding, since most of us don’t realize that we like Emotional Independent Movies Based on Books until we’re told so.

Netflix has earned a sustainable competitive advantage by creating an exceptional information architecture. They got the basics right, invented new forms of categorization and personalization, and assembled these elements into a coherent whole. This is our challenge. Organizing for users is harder than we may think, and more important. Categories are the root of this work, but we should not build them before realizing their connectedness to the whole of the system. Since there are infinite ways to organize, objectives before ontology is vital, and context is the key to classification.

Making Frameworks

So, on one level, our organizational agility has improved. We have many ways to lump and split for users. But, at a higher level, we haven’t absorbed this lesson by reinventing how we organize ourselves. We use tags and facets for objects, but fall back on simple taxonomies for people. John’s a developer, Jane’s a designer, Sara works in Marketing, and Dave is in Support. Once we split into silos, it’s hard to work together.

That’s why the biggest barriers in user experience aren’t design and technology but culture and governance. We can’t create good services without well-defined goals, roles, processes, relationships, and metrics, but all too often we oversimplify. Plan and build get split, and we fail to learn. Us and them are divided, and we fall apart. Inevitably, categorization shapes collaboration in tricky, invisible ways.

To improve these frameworks for making we must classify more carefully by starting to ask where our categories come from. For instance, without thinking, we build organizations on bodily metaphors. We employ department heads and governing bodies to make folks toe the line. And we routinely use a handful of “kinesthetic image schemas” as short-cuts.xxxviii


Figure 2-22. The experiential basis of metaphors.

There’s nothing wrong with using metaphors, provided we’re aware of their source, and realize they contain baggage that shifts from intent to interpretation. Using “department head” may induce cognitive dissonance in an organization that’s flipped the org chart by practicing servant leadership. Isn’t the head on top, like the upper class? Our corporeal experience is embodied in language and subtly changes how we think. This occurs all the time in our use of binary oppositions.

In-Out, Up-Down, Front-Back, Self-Other, Us-Them, More-Less, Male-Female, True-False, Fact-Fiction, Public-Private, Open-Closed, Yes-No, Hot-Cold, Reason-Emotion, Mind-Body, Man-Nature, Love-Hate, Win-Lose, Good-Evil

While there are no opposites in nature, we use dualism to create order and make sense of experience. These opposites generate meaning. We understand hot in relation to cold, light in relation to dark. This dualism runs deep. Studies show “the binary opposition is a child’s first logical operation.”xxxix We start with self-other, edible-inedible, and work our way up to good-evil, digital-physical, map-territory. The pairings are usually hierarchical, and the first tends to be primary. It’s better to be in than out, up than down, true not false, us not them. xl

Now we sense the dangers of embodied cognition. While some oppositions appear to be self-evident, others are clearly value-laden and ethnocentric. Dualism works because it’s simple, but that’s also why it fails. Politicians win by painting in black and white. They say folks are either with us or against us. But this path leads to tribalism and genocide. Most horrors of human history begin with the categories of us and them.

Even when it’s office politics, dualism is serious business. It divides people and obscures the truth. Is digital the opposite of physical? Is that a sensible way to split the staff? Like the Wikipedia, binary opposition can be a good place to start but a terrible place to end. Benchley’s Law – there are two kinds of people in the world, those who believe there are two kinds of people in the world, and those who don’t – points us in the right direction. To collaborate, we must admit ambiguity and complexity, and avoid premature classification.

For instance, teamwork is possible through greater awareness of how we (can) organize ourselves. The classic default is the bounded set of the childhood sandbox. There’s a clear boundary and things are in or out. We use it because it’s easy, but that doesn’t make it right. Spatial ordering of physical objects isn’t how ontology works. Wittgenstein famously debunked this classical theory by questioning the category of “games.” It has no clear boundary as no common properties are shared by all games. Some involve skill, others luck, some you can win, others you cannot. Instead the category is united by overlapping similarities or family resemblances. It’s hard to define a game, but we know one when we see it.

Fuzzy sets have a center and periphery. Some members are better than others. A robin is a better bird than an ostrich. An orange is a better fruit than a tomato. Madonna is a better singer than Bill Clinton. Terror is a better feeling than detachment. Most sets are bounded on the surface but fuzzy beneath. We think we can define them until we can’t. In this failure lies freedom. When we admit they’re not sets in stone but embodied in cognition, we’re able to classify creatively.


Figure 2-23. Multiple theories of categorization.

Paul Hiebert, the world’s leading missiological anthropologist, did just that when he invented the concept of centered sets. His work as a missionary in India led him to ask the question “Can an illiterate peasant become a Christian after hearing the Gospel only once?”xli By tradition, the church was organized as a bounded set with clear definitions of membership and carefully circumscribed beliefs and values. Hiebert proposed a more inclusive, dynamic way to form categories by defining a center, and by paying more attention to direction than location. In his model, a Christian is anyone who moves towards Christ. Some are closer to the center in knowledge and maturity, but all are equal members of the set. It’s an ontology that values openness, change, diversity. It amps up permeability and softens the boundary between us and them.

In 2012, Dan Klyn borrowed this theory to re-frame the relationship between user experience and information architecture. In his account, using centered sets is like herding cats. The center is a pail of milk that draws cats. For user experience designers “the pail is design, and it’s situated in a place where users and their experiences are the center of gravity.” xlii And what about information architects? What’s their center? Well, those crazy cats are centered on meaning. Or is it placemaking or planning or cognition?

To be sure, we can (and should) argue about the centers, but that’s not the most pivotal point. If we connect the dots from facets and tags to fuzzy, centered sets, we begin to see the silliness of playing zero-sum games. As Schrödinger tried to tell us, a cat can exist in multiple categories at once.

There’s a wonderful scene in Life of Pi in which young Pi and his mother and atheist father are walking down the street and bump into Pi’s pandit, priest, and imam all together. After an angry debate, Pi is told “he can’t be a Hindu, a Christian, and a Muslim. It’s impossible. He must choose.” In response, Pi blurts out “Bapu Gandhi said, ‘All religions are true.’ I just want to love God.”xliii Sometimes, we must choose which story we prefer, but not always. All too often we use radio buttons when checkboxes or sliders would reveal the truth. We do it to users and we do it to ourselves.


Figure 2-24. Radio buttons, check boxes, and sliders.

We can and will do better. It starts with awareness. There’s more than one way to classify a cat. Once that door of perception is open, we can nudge ourselves and our colleagues towards celebrating both difference and similarity.

The org chart is a place to start. Is the hierarchy reinforcing the unhealthy division of disciplines? Might a “holacracy” of self-organizing, multi-disciplinary, cross-functional teams work better?xliv In holacracy, authority and decision-making are distributed, and members can be in more than one circle. Zappos and Medium are giving it a try. Maybe we should too.

Once the org chart’s okay, layout is a lever worth a pull. Where we sit relative to our colleagues can unlock creativity and drive collaboration. While hot-desking is going too far, musical chairs reminds us we’re not stuck in our seats. We shape our buildings; we can re-shape them too.

Finally, how we frame our work changes its outcome. Jane touches on this in her critique of public housing projects.

One of the unsuitable ideas behind projects is the very notion that they are projects, abstracted out of the ordinary city and set apart. To think of salvaging or improving projects, as projects, is to repeat this root mistake. The aim should be to get that project, that patch upon the city, rewoven back into the fabric – and in the process of doing so, strengthen the surrounding fabric too.xlv

The same is true of our projects. Often they are better understood as programs or parts of systems. Historically, many of us in user experience have ignored content strategy. We neglect the people, process, and tools of the content lifecycle and everyone suffers including our end-users. When we ignore the ecosystem, our structures are certain to collapse.

Recently, I participated in an event that brought library directors together to talk about digital strategy. Lee Rainie delivered a brilliant keynote in which he presented the results of a Pew Research Center study aimed at learning how and why Americans value public libraries. He concluded by noting the data indicates that “libraries have a mandate to intervene in community life.” xlvi Later we were discussing “the vision for the library” and one participant advised the public librarians to aspire towards “a vision for the community” instead. This re-framing opened the door to an invigorating conversation about interventions and partnerships to address literacy, poverty, crisis informatics, and more. To shift mindsets from insular to open is to change the world for the better.

One shift that can help us all is to change our minds about planning. Like search, planning is a literacy that’s not taught in school, and yet it’s a key to success in life and work. We plan events, trips, families, sites, systems, companies, and cities. We do it all the time but make the same mistakes. First, we procrastinate. We fear complexity, so we start too late. Then, in a hurry, we split ideas and execution into phases or roles. We draw lines in our minds that segregate. The binary oppositions of think-do and plan-build are myths. Like yin and yang, these seemingly separate forces are interrelated and entangled. You can’t do one (well) without the other.


Figure 2-25. The yin-yang of idea and execution.

In planning my trip to Isle Royale, I read and made lists. But I also tried things out. In our backyard, I burned myself on the pocket stove, then learned how to foil the wind. In the living room, I modeled an “emergency poncho” for my wife. She laughed until she cried. Thinner than a dry-cleaning bag, it would have been ripped to shreds by foliage. Then she found me a real poncho. And in the bathtub, I tested the water filter, because as noted earlier, learning by failure feels like a game until someone gets larval cysts in the brain. We learn too late when we put too little do in our plan, and vice versa.

We should take these lessons online to plan-build sites and systems, because the binary opposition of agile-waterfall is just as much a myth. The Agile Manifesto backs “responding to change over following a plan” but makes a point of saying that both have value. Yet Agile is used often as a platform for proclaiming the wireframe is dead. Meaning shifts from intent to interpretation, and plans go out the window. We all know death by documentation sucks, but to pivot and sprint into an Agile death spiral isn’t a whole lot of fun either. Fortunately, the plan-build pendulum is now swinging back to the middle thanks to a succession of expensive unplanned disasters.

The timing is good. As the complexity of our ecosystems grows, we will need plans and prototypes more than ever before. To wrangle strategy, structure, and schedule in our heads is absurd. We must put ideas into the world so we can see them. Architects have been doing this forever.

It’s part of the timeless way of building that Christopher Alexander draws upon to generate the quality without a name.xlvii In planning Eishin Gakuen, a combined college and high school he built outside Tokyo in the 1980s, Alexander used many tools to extend cognition. First, instead of simply interviewing students and teachers, he invited them to co-create a pattern language – a word-picture that describes the wholeness of a place – since “it is immensely hard to help people tell you what they want.”xlviii Together they sketched out 110 essential patterns for the campus, including:

2.2 The Small Gate marks the outer end of the Entrance Street. It is a small, imposing building, which has height and volume. Hosoi, Nodera, Suzuki

6.6 The Library, also a two story building, has a large quiet reading room on the second floor, with shelves, and tables, and carrels, and beautiful windows. Kajiyama, Oginawa, Tomizu, Sato

7.7 There is also one garden, so secret, that it does not appear on any map. The importance of this pattern is that it never must be publicly announced, must not be in the site plan; except for a few, nobody should be able to find it. Hosoi

In parallel, he and his team mapped the topography – land forms, slopes, trees, ridges, roads – of the physical site. They then began the hard work of bringing the two systems of centers, patterns and places, together into a simple, beautiful site plan. They planted hundreds of six foot tall bamboo sticks topped with colorful ribbons to identify places, spaces, and relationships. By seeing-moving these flags for months, they were able to discover the plan. They augmented this visualization with topographic models of the site, using pieces of balsa wood for buildings. After trial and error, they fit all these patterns and places into a wonderful, generative whole.

In this story we see the synthesis of embodied and extended cognition. There are more dimensions to architecture than Tetris, so it’s even more vital we use models in the world to shift minds. Planning is making. Maps, sketches, words, and wireframes are still essential, but it’s also vital that we design in the medium of construction. How else will we imagine cross-channel experiences and the Internet of Things into life?

Last year, I worked on a responsive redesign for a database publisher. Our team built wireframes and design comps to conduct quick, cheap experiments, and then an HTML prototype to enable new loops of build-measure-learn. Each of these cognition amplifiers is unique. Together they teach us that one way is the wrong way. As architects, designers, and developers, we each bring discrete value to think-do and plan-build. All too often, classification obstructs collaboration. It splits us and them, and our products show the seams, and our users bear the scars. The things we make are reflections of how we see and sort ourselves, so let’s classify-plan accordingly, and be mindful that making frames is work.


Recently, I enjoyed a tour of the Inspired Teaching School. It’s a public charter in Washington, D.C. that cultivates inquiry-based learning by transforming the role of the teacher from information provider to “instigator of thought.” Instead of showing students how to do their work, teachers challenge the kids to do it themselves. It’s a tiny habit called “don’t touch my pencil” that makes a big impact. The other part I recall is the artwork. I remember a colorful drawing of animals in three categories – real, imaginary, impossible – and being inspired by the freedom with which kids invent impossible creatures.

Over time, lest we’re careful, our bodyminds grow inflexible. We imagine the impossible less and less until we can’t. Paul Graham says entrepreneurs must be cheeky, always believing there’s a better way. Similarly, information architects must be contrarians, always re-framing ideas and beliefs in a different way. Of course, Richard Saul Wurman, the infamous architect of “information architect” would beg to disagree.

I rather worship the space between things, the silence between good friends, the time between the notes of music, the break time during a conference, the space between buildings, negative space…It’s the way I approach everything. I look for a solution which has a valid oppositeness. Not a ‘different way’ of looking at things, but an opposite way.xlix

In fact, Mr. Wurman is so oppositional, he’s easy to ignore. His own wife lovingly described his self-image as “a little piece of shit at the center of the universe.”l But Dan Klyn is right. We must listen to our Dutch It’s the folks who challenge social norms that have the most interesting things to say. After all, only a contrarian would single-handedly transform the American Institute of Architects 1976 national convention into a conversation about “the architecture of information.”

That’s why I’ve chosen to call myself an Information Architect. I don’t mean a bricks and mortar architect. I mean architect as used in the words architect of foreign policy. I mean architect as in the creating of systemic, structural, and orderly principles to make something work – the thoughtful making of either artifact, or idea, or policy that informs because it is clear.lii

Since that event, the borders of our practice have shifted like sandbars. We’re drawn to architecture, information, planning, meaning – it seems the centre cannot hold. But that’s the strength of our discipline, not a fault. We define and reframe. We destroy and rebuild. The center of information architecture is cognition. In re-re-framing, we understand.

One of Wurman’s most repeated wisdoms is: “The ways of organizing information are finite. It can only be organized by location, alphabet, time, category, or hierarchy.”liii At first the last was continuum, but he changed it to make LATCH. And, in that swap, we see his scheme is arbitrary. The acronym is catchy, but it’s the opposite of right. The ways of organizing information are infinite. As uncle Buddha once said, put no head above your own, because even Dutch uncles are wrong.

To build strength and flexibility, we should open our minds to people and ideas we don’t like, and pick fights with those we do. For instance, Stewart Brand’s concept of pace layering gets a lot of love. He argues that in complex systems, it’s vital that distinct layers can change at differing rates. The combination of fast and slow creates resilience. Fast learns, slow remembers. Fast gets our attention. Slow has all the power.liv


Figure 2-26. The pace layers of civilization.

It’s a model he used brilliantly to explain buildings – site, structure, skin, services, space plan, stuff – and how, in time, they learn. lv It’s since been adapted widely in many fields. The order of layers affords comfort. It belies a measure of control. But all maps are traps. This is too. So, what’s the opposite of pace layers? Is it everything’s intertwingled?


Figure 2-27. Everything is deeply intertwingled.

The layers exist. There are no layers. Both statements are true and useful. Everything depends on context. In the 1990s, the design of hardware and software as separate layers was clearly the right strategy, until Steve Jobs returned to Apple and proved the power of synthesis and integration.

We find similar opposition in our work on the Web. To define projects, managers limit by layer. We aim to refresh the interface without touching architecture. We optimize search with no content strategy. We stretch across silos, then trap ourselves in layers. To escape, we must help folks see how a simple change to a single page can send ripples from code to culture. When we limit by layer, it’s vital we look for levers.

We should also look to nature for insight. For instance, coral reefs are made of layers, but that’s only one way to see them. Jamaica’s reefs in the 1950s served as a lovely archetype of a thriving ecosystem. Hundreds of species – sharks, snappers, parrotfish, jacks – swam among the colorful sponges and feathery octocorals sprouting from the hard coral base. In the ensuing decades, the reefs were subjected to stress and shock. Fishing and tourism grew rapidly. A fierce 1980 hurricane caused major damage. But the system appeared to bounce back. Biologists were impressed by the reef’s resilience. Then, in 1983, an unidentified pathogen decimated the long-spined sea urchin population. Left unchecked, algae quickly covered and killed all the coral, and the whole system crashed.

On a healthy reef, a new pathogen decimating a single species (like the urchin) might not have had catastrophic consequences, because an essential reef function – like keeping algae in check – could be performed by more than one species. On the highly compromised Jamaican reef, however, the continued flourishing of the ecosystem as a whole became entirely dependent on a single species continuing to do that job. The loss of the urchins, an otherwise modest trigger, caused the reef to collapse virtually overnight.lvi

It’s vital to note that nobody predicted this chain of events. Cross-layer relationships that are easy to see after are often invisible before the event. Resilience is the capacity of an ecosystem to respond to disturbance through resistance and recovery. In our systems, we can respond better if we frameshift from layers to the levers that bind together fast and slow.

Tranquility or Insight

This chapter began in meditation and the search for anattā, since no ontology is more perilous than how we see ourselves. Historically, we defined man in opposition to nature, yet the boundary doesn’t exist. I love our national parks, but we can’t stop the clock by encircling wilderness. As Alex Steffen tells us, the future of environmentalism is bright green.

Dream of living your one-planet life in a bright green city on a sustainable and thriving planet…We need, through brilliant innovations, bold enterprise and political willpower, to make sustainability an obligatory, universal characteristic of our society, not an ethical choice. We must remake the systems in which we live. We need to redesign civilization. Anything less is failure. lvii

Our old categories create externalities. We enable ourselves to cause pollution, suffering, and collapse outside our model of the system. But every action has an unequal and non-opposite reaction. We’re dealing with karma, not physics. Man is not apart from but a part of nature, which makes the equations far more complex. It’s unclear how we’ll change the course we’re on, but it starts with redefining ourselves.

In 2005, we added a Sheltie to our family. At first, I forbade her from the couch. But our six year old daughter Claire broke me down by tearfully proclaiming “Knowsy is a person too.” In 2013, the Indian government followed suit by banning all cetacean captivity and declaring that “dolphins should be seen as ‘non-human persons’ and as such should have their own specific rights.”lviii Now, in the United States, the Nonhuman Rights Project aims to reclassify animals as persons, not things.

Our goal is, very simply, to breach the legal wall that separates all humans from all nonhuman animals. Once this wall is breached, the first nonhuman animals on earth will gain legal ‘personhood’ and finally get their day in court – a day they so clearly deserve. lix

Civilization is arguably a story of an expanding moral circle. Over time, we’ve extended kindness from kin to tribe to nation and beyond. In 1776 when Thomas Jefferson declared “all men are created equal” he implicitly excluded women, African Americans, Native Americans, Jews, Quakers, Catholics, men without property, and anyone under 21. The Rights of Life, Liberty, and the pursuit of Happiness applied to less than ten percent of the human population. Whether we’ll extend ourselves further, or not, is unclear, but shouldn’t we take the time to consider our principles of classification? Is the boundary of our moral circle fuzzy or fixed? Is the center sentience or suffering? How do we circumscribe empathy? Is it emotion, fairness, or simply might makes right?

In meditation, Buddha learned everything is process, there is no self. Two and a half thousand years later, modern science is proving him right. The average age of cells in the body is 7 to 10 years, and our whole skeleton is replaced every decade.lx A person is a pattern that doesn’t exist. And it’s not just impermanence that blurs what’s mine. We have fuzzy borders too. The body is an ecosystem of ten trillion cells that also contains one hundred trillion bacteria that together affect digestion, weight, health, and even our mood. Each of us includes 2 to 5 pounds of them. This gives new meaning to Walt Whitman’s “I am large, I contain multitudes.”

We don’t know our limits. Francis Crick speculated that the claustrum, a thin layer of tissue beneath the insular neocortex that has two-way links to nearly all regions of the brain, may be responsible for integrating myriad sensations – sight, sound, touch, taste, smell – into the single, unifying experience of consciousness.lxi Of course, whenever we unify, we also divide. We invent self-other as one in what Albert Einstein famously called the “optical delusion of consciousness.” To make sense of an infinite universe, we create categories to reduce complexity. And we use tools and language to spread the load across mind-body-environment.

Despite these devices, our search for the truth is limited by a very small flashlight. So we must spin our categories like tetrominos. We must turn our ontologies downside-in and upside-out. We must seek monsters and cyborgs in the borderlands, and be mindful to watch for “black swans.”lxiiWe can’t make change unless we’re playful, since learning means letting go. E.M. Forster wrote “the song of the future must transcend creed” and asked “how can I know what I think till I see what I say?” There’s wisdom in those words, but to stare at the finger is to miss the moon.

In the beginning was yathā bhūta, reality as-it-is, unmediated by concepts or classification or culture. Now, trapped in our own maps, we meditate, in search of the untranslatable, an understanding deeper than words. We watch our breath to free our selves, to become aware of tranquility and insight, and categories and connections, as all a part of a single path.