THE INTERNET started out as the Information Highway, the Great Emancipator of knowledge, and as an assured tool for generating a well-informed citizenry. But, over the past 15 years, that optimism has given way to cynicism and fear — we have taught our children that the net is a swamp of lies spun by idiots and true believers, and, worse still, polluted by commercial entities whose sole aim is to have us click to the next ad-riddled page.

Perhaps our attitude to the net has changed because we now see how bad it is for knowledge. Or perhaps the net has so utterly transformed knowledge that we don’t recognize knowledge when we see it.

For philosopher Michael P. Lynch, our fears are warranted — the internet is a wrong turn in the history of knowledge. “Information technology,” Professor Lynch argues in his new book, The Internet of Us, “while expanding our ability to know in one way, is actually impeding our ability to know in other, more complex ways.” He pursues his argument with commendable seriousness, clarity, and attunement to historical context — and yet he misses where knowledge actually lives on the net, focusing instead on just one aspect of the phenomenon of knowledge. He is far from alone in this.

The net in fact exposes problems that have long lurked in our epistemology, problems that come into stark relief when knowledge is freed of paper, and we freely connect with it and through it across all boundaries of time and place. There’s something about how we’ve been thinking about knowledge — something inherent in traditional epistemology — that blinds Lynch and many others to the knowledge-enhancing aspects of what’s happening on the screens in front of us.

Knowing Beyond Google

The rhetoric of internet criticism often follows a typical pattern, such as, “Yes, but,” as in: “The internet does many wonderful things, but …,” followed by a long list of everything that’s wrong with the net. Lynch is far more thoughtful than most, but his arguments nonetheless suffer from the flaws that typically bedevil this type of criticism. His idea of what the net gets right — the “Yes” before the “but” — is itself qualified, and a good place to start. He characterizes knowing on the net as “Google-knowing,” or what we might otherwise call “looking something up.” He acknowledges that having instantaneous access to facts via Google-knowing ranges from handy to lifesaving, but his conception misses the full picture in two important ways.

First, we should include in Google-knowing more than just factual look-ups. If, in the 19th century, almanacs commoditized facts, then the net has taken the process one big step further, commoditizing the encyclopedia article. Not only can we look up when Thomas Jefferson was the United States minister to France (May 1785-September 1789), we can also get some quick context about what that meant to the Colonies, to France, and even to him. (To his credit, Lynch does not engage in the usual Wikipedia bashing.)

More important, to focus a discussion of internet knowledge on people looking things up on Google, or even in Wikipedia, is like describing libraries as places where people use the encyclopedia. Far more typically, when we’re on the net, someone or some service links us to a bit of news that’s interesting to us. It might be politics, quantum physics, or the Kardashians. Whatever it is, let’s say we want to know more. Before the era of the net, the reader’s curiosity was bounded by the physical rectangle in the newspaper within which the article sat. On the net, if the source of the tidbit — a tweet, a Facebook post, an email — doesn’t itself contain links to additional information, then we can pop some terms into a search engine and find more avenues to explore; push a comment or question back into the social medium through which we first learned of the topic; hop it over to a different social network; or, even reach out to the thinker or writer who stimulated the discussion in the first place.

This is the phenomenon of knowing these days. Surprisingly, it does not figure much in Lynch’s book except in examples of how knowledge goes wrong. “Sometimes we need to know more than the facts; sometimes we want to understand,” he asserts, before concluding that “it is our capacity to understand that our digital form of life often undersells, and which more data alone can’t give us.” That’s a trenchant criticism of the net, but only if we assume that on the net, all knowing is Google-knowing.

In fact, knowledge is now networked: made up of loose-edged groups of people who discuss and spread ideas, creating a web of links among different viewpoints. That’s how scholars in virtually every discipline do their work — from their initial research, to the conversations that forge research into ideas, to carrying ideas into public discourse. Scholar or not, whatever topic initially piques our interest, the net encourages us to learn more. Perhaps we follow links, or are involved in multiyear conversations on stable mailing lists, or throw ideas out onto Twitter, or post first drafts at arXiv.org, or set up Facebook pages, or pose and answer questions at Quora or Stack Overflow, or do “post-publication peer review” at PubPeer.com. There has never been a better time to be curious, and that’s not only because there are so many facts available — it’s because there are so many people with whom we can interact.

While The Internet of Us mentions many of these new formations, it is usually to dismiss them: “Social media and the blogosphere are filled with ‘reasoning’” that’s either “blatant marketing or aimed only at supporting what people already believe,” writes Lynch. He deals with MOOCs, or massive open online courses, as if they’ve already found their final form and as if MOOCs are all the net has done for education. There is only a single mention of the Open Access movement late in the book. Although Lynch concedes that data analysis in the digital humanities can sometimes yield interesting results, that admission comes only after he has spent several pages demonstrating in detail the flaws in the sole digital humanities article he uses as his example.

Why does Lynch miss so much? He has written an intelligent book that struggles honestly with important questions: Is the net turning us into passive knowers? Is it degrading our ability to reason? What can we do about this?

As our friends on Reddit might put it: He misses it because classic epistemology.

Setting the Rules of the Knowledge Game

Lynch acknowledges from the start that he’s writing as a philosophy professor embedded in the classical tradition: “My aim is to examine the philosophical foundations of what I’ll call our digital form of life,” he declares, using an apt Wittgensteinian phrase.

He thus does not claim to be breaking new ground in epistemology so much as applying the wisdom of “The Tradition” to our digital world. He writes for an undergrad rather than specialized level, and introduces concepts as they become relevant. For instance, he nicely explains the Socratic distinction between right opinion and knowledge: knowledge entails justifying one’s opinions. This throws him straight into the question that has long occupied traditional epistemology: what constitutes justification?

His answer is also traditional. He starts with the lowest level of knowledge: the type of knowing that we share with other animals. This “receptive knowledge” springs straight from sensation: “if our visual system, for example, is working as it should, and our representation of the world is accurate — if we see things as they are — then we come to know” in this basic way. True, like so many others, he modernizes the vocabulary, using phrases like “downloading” information, but his description nonetheless constitutes straightforward empiricism.

He then climbs the empiricist’s ladder. How about knowledge that is more abstract or general — the knowledge that can, for instance, be found on the net? He dismisses the idea of the heroic single knower, acknowledging that “reflective” knowers — a term he pairs with “responsible” — need to rely on others. This reliance requires learning and applying the rules of reason in public spaces, where, heeding Kant, we must think for ourselves and treat others as equally autonomous agents. But, suppose at bottom, all chains of reasons end with mere belief?

In a rigorous and frank chapter, Lynch faces this question, which has dogged epistemology from its origin, and which he believes the internet is making “even more pressing.” Invoking Wittgenstein again, he writes that, in order to be reasonable, we have “to be willing to play the game of giving and asking for reasons …” This edges him dangerously close to the postmodern world in which such games cannot confidently assert their privilege over other games. Or, if you call this a “paradigm” instead of a game, then you are taken to the territory of Thomas S. Kuhn’s 1962 The Structure of Scientific Revolutions. Or to a Heideggerian critique of rational scientific knowing as one way of disclosing the world, which needs to be grounded in a more basic analysis of disclosure itself. Or, perhaps, to a Foucauldian examination of knowledge as a historical expression of power. Or to Bruno Latour’s explication of the dependence of knowledge on its instruments and the networks in which they reside. Or to the “embodied mind” theory put forward by Francisco J. Varela, Eleanor Rosch, and others that takes the mind to include the body acting in the world.

None of these are necessary paths to take. Some may be silly. But Lynch takes none of them, declaring that “the game of giving and asking for reasons” is to be played “by rules most of us could accept were we to stop to think about it.” He stays on the Enlightenment side of the divide, well clear of the postmodern epistemic world.

Indeed, that thoughtful people would come up with the same rules remains the great hope of the democracies spawned by the Enlightenment. It buttressed Barack Obama’s patient insistence that opposing sides can find common ground. It is at the heart of Al Gore’s The Assault on Reason, an assault he believes the internet can help counter, although it was published in 2007, before the current internet backlash had taken over the discourse.

But if the internet has taught nothing else, it has taught us that we will never agree about anything. For every fact on the internet, there is an equal and opposite fact. It doesn’t help that not every purported fact is in fact a fact, or that facts tend to have multiple layers of truth. Facts simply are not going to play the role in building consensus that we had hoped.

This would be a source of pure despair if the internet were not also enabling us to see that before it existed we never agreed about anything either. Before the net, what we read and saw was so tightly controlled by cartels of well-intentioned professionals that dissenting voices were barely heard. True, many of those dissenting voices were wrong and sometimes they were spouting lunacy, but we marginalized all but the one percent of the epistemically privileged. We achieved great science but at a high price throughout the rest of the cultural landscape, and sometimes within science, too.

This fragmentation of knowledge is a fact that knowledge cannot overcome. How, then, do we best live with it? How do we flourish now that we can’t reason ourselves back together?

The Sound of One Echo Chamber Clapping

First, there is little cause for panic or despair. Knowledge has always been a local product. Now, with our localities connected globally, we are more clearly seeing what has always been the truth of our condition.

And it’s not as if we are oblivious to this issue. Handling the challenges posed by networked knowing is at the core of almost every website and internet service. Each site or service takes steps to establish the rules of engagement so that users can come to appropriate levels of trust about the claims made by the site and by other users. For Uber, it’s a ratings and comments system for drivers and passengers. For Public Library of Science (PLOS), it’s a peer review process and post-publication comments. For Reddit, it’s the up- and down-voting of vigorous and extended discussion threads. None is perfect, but then knowledge has never been perfect. Each is, however, appropriate for its domain.

But, instead of flawed, domain-specific ways of knowing on the net, Lynch sees in-grown, hermetic clusters of people who simply confirm each other’s baseless beliefs and drive themselves to extremes. In short: He sees echo chambers, which Cass Sunstein notably discussed in his 2001 book, Republic.com.

Echo chambers are indeed a problem. Large commercial sites abet them, making money by getting us to click on links to yet more ad-laced pages. And, of course, these sites exploit the fact that we’re more likely to click on a link to an idea with which we agree than one we hate. Even respectable news organizations have learned that clickbait headlines work best if they play into our existing beliefs. “Filter bubbles,” as Eli Pariser calls them, pay.

But thinking of the net as a set of echo chambers is just bad phenomenology. If you’re using email, Twitter, or Facebook, or if you do your Google-knowing via a search engine, you are inevitably face to face with at least some ideas and links that don’t come from your particular echo chamber. Plato’s Cave is now shot through with holes that transmit light, sounds, and smells that you can follow just by clicking your finger.

Do we have to follow them? Of course not. We tend to follow those that make sense to us. Our interests turn out to be distressingly parochial and local in most cases, as Ethan Zuckerman demonstrates in Digital Cosmopolitans. But, since epistemologists like Lynch insist echo chambers have degraded knowledge, we have to ask what our media diet consisted of before the net.

The answer isn’t pretty: Three channels of nightly news; a small handful of newspapers in major cities; a nice spread of national magazines, each one its own echo chamber; a Great Books series launched in 1952 that consisted of works by 130 authors, not one of whom was a woman or black, and almost all of whom were within the European tradition.

Of course, the decades after the 1950s offered a greater diversity of voices, but the basic mechanisms of knowledge continued to do the work of echo chambers: identify and publish the work of experts, vet knowledge to exclude what does not fit, and create a coherent body of knowledge that becomes harder and harder to challenge precisely because of its curated cohesion. Certainly the content was incomparably better than what comes out of toxic internet echo chambers. But the problem with echo chambers has to do with their structure, and traditional knowledge comes from that same structure.

The net is making clear how important “echo chambers” are to knowledge and even more so to understanding. If you care about molecular gastronomy and hear about a new technique, you’ll go to your favorite molecular gastronomy sites to learn more. If you’re a supporter of Net Neutrality and there’s a court ruling you don’t understand, you’ll go to a site that shares your values to get the explanation. If you are a feminist and a new pay equity law passes, you’re not going to go to a male supremacy site to find out what it means for you. Knowledge and culture depend on like-minded individuals joining together and iterating over tiny differences. This is how the net works. This is also how traditional knowing works. We did not like to acknowledge that. Now we can’t avoid it.

The Medium is the Knowledge

Perhaps our chief epistemic avoidance mechanism was turning knowing into the production of a type of content — knowledge — that we convinced ourselves had to be independent of the knower in two senses.

First, we devised methodologies that try to keep the vagaries of the individual out of the process of creating knowledge. The scientific method works. Journalistic objectivity continues to be reevaluated. (I should note that on the question of objectivity Professor Lynch contrasts his views explicitly with mine, as he does when he considers networked knowledge. He does so without animus, and I certainly respect and appreciate his criticism.)

Second, we physically separated knowledge from individuals by externalizing it (e.g., books). What started in Greece as a particular class of belief became a body of printed statements that could be called knowledge even if there was no one left to believe them. Obviously, this has been wildly successful for our species, but it also meant that the medium of externalization — paper — has shaped knowledge to fit its peculiarities. Knowledge is settled belief because ink settles into paper. Knowledge divides into topics because books can only be so big. Our most revered form of knowledge consists of long chains of reasoning because books are sequential and not suitable for optional digressions. Knowledge is something most of us read passively because books are inert — just as Socrates had opined in the Phaedrus. Knowledge is a type of content independent of believers because books are physical objects that outlast their authors and readers.

There is nothing natural about knowledge, so it is not surprising that it shaped itself to its medium of storage and communication. But now we have a new medium, and knowledge is taking on its properties. That is why looking for traditional knowledge on the net leads us to miss the phenomenon of knowledge there. It’s also why the idea of knowledge is not much discussed on the net. It’s possible that, in the next generation, the noun “knowledge” will become old-fashioned and, in the generation after that, archaic.

That wouldn’t mean that we have given up on the project of knowing our world. Instead, it would mean that that project has changed shape — from content to networks that mix ideas and sociality. That’s already happening. Knowledge is becoming what happens when links connect differences and people.

This is new but it should also sound quite familiar, for it is very close to how scholarship works in the traditional world: people who care about a topic argue about it among themselves. Networked knowledge opens up the discussion and the participants remain linked and engaged, not always expecting final resolution. Knowledge exists in the linkages of difference and disagreement.

The net is demonstrating the weakness of knowledge as finished, settled, and static content. It’s doing so by plunging us deeper into knowing.

The Networking of Knowledge

Lynch disagrees. He sees more misinformation than knowledge on the net, and he believes most of us are abdicating our responsibility for sorting the one from the other. Google-knowing is, for him, akin to consulting an unreliable almanac. That may be an inevitable consequence of scaling up almanacs to include answers to just about every question imaginable, although I suspect that Google does quite well with questions concerning information literally in almanacs. Curated sites such as WolframAlpha.com undoubtedly do even better.

But let’s grant Lynch his concerns. They are real even if they are not the whole story. So, what do we do about them?

Once again Lynch’s answer relies on a traditional assumption that our experience of the net calls into question: agreeing on our “epistemic principles” because if we’re not playing by the same rules, then it’s “game over,” as he puts it. But which rules? Lynch tells us that no matter what we believe in our private lives — “the Bible, or the Koran, or Dianetics” — we should “support institutions like the National Science Foundation or the National Endowment for the Humanities …” for they are “doing the work of democracy.” (Emphasis in the original.)

Like everyone else I know who packs their Whole Foods groceries into NPR tote bags, I, too, like those institutions. But if I am a committed Christian, Muslim, or Scientologist, those religious beliefs are at the heart of what I know. Telling me to put those beliefs aside so I can engage in a reasonable discussion is to ask me to play by someone else’s rules. And the problem is repeated if we accept that it’s not just religious people who disagree about the rules of engagement for developing knowledge. Arguments over what constitutes acceptable evidence are common in just about every field. (Lynch’s condescension about religion might be overcome if he were able to engage with the faithful on their own terms.)

I do agree that when it comes to making public policy, we want evidence-based discussions whose rules do not require private access to sources of truth. Likewise, I want the scientific method to settle questions of science and the jurisprudence system to settle questions of criminal culpability. But the net has proven what philosophers, historians, neurologists, and sociologists have already told us: it’s not that simple. Science is deeply embedded in systems of instruments, power, economics, gender, psychology — in short, in all the different facets of being human.

Although we agree that it’s good to “promote institutions that encourage cooperation, and even face-to-face contact with people who have very different views,” perhaps the way forward seems clearer to Lynch than it does to me for the following reason: beneath his traditional epistemology is a traditional ontology that takes the world as that real stuff that is outside our heads. Citing John Locke and the internet philosopher, Luciano Floridi, he argues that digital objects are more abstract than real things because they are purely formal — they are copies without originals — and thus, he implies, they are less real. He warns us that we shouldn’t let the digital world “blind us to the world that is not constructed, to the world that predates our digital selves,” the world of things that are found and not made, as he puts it. But where is the found thing that we don’t make by taking it in one way or another? In taking the clock as a time-piece and the bullet as a way to create unasked-for holes at a distance, we are relying on a contextualizing world that has been constructed even before we got here.

Lynch is able to treat the net as a simulacrum because he thinks knowing is something we do in our heads. We build up to meaning by starting with sensation. But the net, in his view, is sensation without a real referent. It is a representation of a representation. It is therefore too bad that he dismisses Andy Clark and David Chalmer’s “extended mind” idea by saying, “it might be right but we don’t have to go that far” because the mind is already extended, by which Lynch means that we rely on the testimony of others to justify our beliefs. But the extended mind concept says something more: we think with tools. The physicist cannot think about a problem without using a white board. An accountant needs a calculator. The philosopher needs books and writing materials and perhaps a fire and a glass of sherry. We think with tools. We think out in the world, not in inner representations of the world. And now we have new tools for thought. These tools include not just search engines, but everything from web pages to complex multi-modal networks of experts and amateurs. That is where thinking and knowing is now happening.

More to Lynch’s point, it is where understanding is happening. Lynch acknowledges that we are “cognitively interconnected” by the internet, but, he writes, it’s our Google-knowing that’s becoming networked, and, as he notes, understanding needs more than that. When he explores the idea of networked knowledge, he uses as his examples the wisdom of crowds as expounded by James Surowiecki, as well as prediction markets. These are important and interesting network phenomena, but they are online crowds in which knowledge is generated by averaging the contributions of individuals. This leads Lynch to his summary dismissal of networked knowledge: “You can’t take the individual out of the equation.”

No, you can’t, and the most important new form of networked knowledge does not. A network in our new age is a set of individuals who maintain their individuality while in relation to one another — that is, while maintaining their differences. The knowledge exists in their individual expressions and, most importantly, in their links to one another. This is knowledge that is not an averaging of inputs, but a maintenance of differences-in-connection.

I suspect that Lynch misses all those places where understanding is happening on the net because he is looking for traditional knowledge that arrives at a definite conclusion, is written down, and serves as a reliable authority. But networks are conducive to a type of knowing that paper makes difficult. Networked, collaborative understanding happens in posts, blogs, messages, tweets, and discussion threads. It even happens in exchanges that occur over the course of 20 years in a closed forum like a mailing list — an echo chamber if you want, or a tiny Republic of Letters if you prefer. It just doesn’t have the earmarks of traditional paper-based knowledge.

Lynch sees these new locales of knowledge, but thinks networked knowledge is never resolved because it can’t agree on what grounds a chain of reasoning. Lynch counters with the primacy of “the objective world itself,” which we can know reliably through the senses. So we are back to empiricism. This leads Lynch to advise us to literally get “up off the couch” and plunge “into the whirlpool of actual experience.” “To escape your circle of justification, do what you do with any circle: step outside its borders and breathe in the environment on the outside.”

But when you do so, your every perception will further confirm your existing assumptions about how the world works — because understanding is, fundamentally, the act of assimilating the new to an existing context. The inside of your skull is an echo chamber.

There’s tremendous value in consulting existing bodies of well-vetted beliefs, and, to their credit, teachers like Professor Lynch expose us to that value. But there is also value in the networking of knowledge in which ideas are linked in their differences. We can go wrong in those networks, but we can also go very right, achieving a new sense of how knowledge goes together even if it never fully coheres.

The networking of knowledge does not achieve the aims traditional knowing has set for itself. It is settled only within a community of believers — and not all communities of believers are right. It is inextricable from its social context. It inevitably contains differences, but those differences are now linked. It is as discursive as the net itself. It often comes in small bites, but those bites are linked out to a world larger than all the libraries that ever existed. Everyone gets to speak, even stupid, evil people. Authority generally has to be earned, not declared. The rules of reasoning vary not only within domains but within each conversational forum. Knowledge is being replaced by knowing, and knowing is now a free-for-all. At its best, this knowing does what Lynch recommends: it thinks explicitly about its rules of justification. At its worst, it’s a howling mob.

There is endless evidence to support pessimistic or optimistic views, for both are true. This is the greatest time to be a curious person who wants to learn, and it is the greatest time to be a complete idiot. The net is revealing both the power of our traditional ways of knowing and the fact that traditional knowing has always been a product of flawed humans going wrong and going right together. Knowledge cannot liberate itself from this. Ultimately, knowledge’s only hope is for more and better humanity.

¤

David Weinberger is a senior researcher at Harvard’s Berkman Center for Internet & Society who writes about how the internet is changing our ideas.