Zittrain's Future of the Internet is a book about Internet governance, despite the fact that the author refuses to use that term and, near the end of the book, rather testily distances himself from the Internet Governance Forum and related attempts to develop and reform global Internet governance. Despite that, the book is a well-researched and well-constructed analysis of the architectural issues underlying some of the policy problems (and opportunities) facing the Internet. It is worth reading – and worth critiquing – for the debate will shed light on some of the key choices we have to make regarding the governance of the Internet.

Zittrain’s basic argument is set out in Part 1. Here he establishes a compelling and important contrast between two models for our environment of digital devices and networks. One is a world of sterile information appliances tethered to a network of control. The devices are locked down – their owners cannot reprogram them or install new applications from third parties (an example is the iPhone). The networks are proprietary walled gardens, where the owners act as gatekeepers to the products and services that are allowed on them. The other model, the model which made the Internet into an explosive success, combines open network protocols and programmable devices like your typical personal computer (PC). He calls this “generativity” (although the more one reads the less impressed one becomes with what this term adds to our conceptual toolkit). In the open, participatory Internet environment, the devices were not designed to optimize a single function, but relied on a general-purpose operating system that could run code from any compatible source and perform any number of functions, including ones never anticipated by the device maker. And the same was true of the network, the design of which was nonproprietary and incomplete, and based on the “procrastination principle” which allows problems to be solved later by others.

We all know the story of the Internet’s success. Instead of uncritically singing its praises, however, Zittrain forthrightly recognizes and documents the problems that were latent in it. The very properties that made the Internet innovative and revolutionary are also the ones that now threaten to bring it down. Open programmability and open networks foster not just innovation and freedom but also spam, viruses and organized cybercrime. The profusion of bad code is not a Microsoft problem, as some wish to pretend, it is inherent in the regime’s open architecture. The “future of the Internet” implied in the title is that its very success might push us all back to the model of locked down devices and tethered, centrally controlled information appliances. And as explained in Chapter 5, an especially good critique of regulation through tethered appliances and “pre-emptive design,” centralizing and extending regulation through tethered appliances does not look like a very desirable future.

This is the problematic of the first third of the book, and there is no better place to get a detailed and insightful description and analysis of this dilemma. The natural next step in such an argument is to propose some criteria, policy or principle for retaining the benefits of openness while dealing with its problems. Here the book stumbles a bit. There are many creative ideas and worthwhile discussions, but there are three problems with Zittrain’s discussion of solutions.

  • The first is that the concept of generativity turns out to be too ill-defined to bear the weight of becoming a guiding principle for policy.
  • The second is that Zittrain’s ideas about harnessing new forms of Internet-based peer production to solve the problems of the Internet, while intriguing and worth pursuing, turn out to be hedged with so many qualifications and “on the other hands” that one is not sure where they will actually lead us, if anywhere.
  • The third is Zittrain’s disengagement from the political and his tendency to look exclusively to “new technical tools and protocols” for solutions to what are often political problems.


Generativity is core to the book’s analysis but it is not well defined. At times he speaks of it as if it were an intrinsic property of tools (indeed there is an entire table to tell us whether hammers and duct tape are more “generative” than anchor bolts or Lego blocks). At other times it is defined as if it were part of a system of relationships among tools and people, and the operant property is the level of end user control in that system (a completely different concept from the first one). He admits that generative systems can be built using non-generative tools or platforms. (As he puts it, “no technical reason prevented Compuserve from developing Wiki-like features and inviting its subscribers to contribute to something resembling Wikipedia.”) With respect to information technology, what he seems to mean by generativity is basically nothing more than that users or third parties authorized by users can control the applications running on their PCs or engage in unmediated participation in the shaping of content on web sites. We don’t need the conceptual overlay of “generativity” for that.

It is really the human mind that is generative, not tools. Tool systems can be structured to be open to human creativity and autonomy or they can be structured to stifle or contain it. Politics, economics and historical path-dependency play the biggest roles in determining where we fall on that spectrum, not ex ante design choices. Generativity as Zittrain describes it seems to be an outcome rather than a design property of a specific technical system; it is a dependent variable not an independent one. Since the definition of generativity is so fuzzy, it is not clear that policy should, or even could, optimize for it, nor does it provide clear guidelines for how to balance it against other values.

Using the Internet to solve the problems of the Internet

Think of the concept of generativity as a detour in an otherwise useful narrative. You can probably skip Chapter 4, which describes and analyzes the concept, without losing any of the value of this book. What Zittrain has to say about creativity on the net is ultimately not all that different from what Yochai Benkler, Clay Shirky, Eric Raymond and other apostles of peer production and new networked forms of organization are saying. Bearing that in mind, Zittrain proposes to harness these collaborative capabilities to fight the propagation of bad code by allowing people to enter into a social configuration in order to attack it. We need, he says, tools to counter the “autistic nature of the Net experience,” which makes PC users unaware of their digital environment and therefore unable to act on social cues signaling approval or danger. This very interesting argument, which draws on the work of his colleague and mentor David Clark of MIT, points in the direction of software tools that enhance social awareness about things like reputation, or the effects of participating in various web sites or installing certain kinds of software. It is an attempt to bring the creativity of the Internet back to bear on the resolution of its own problems.

Zittrain understands, however, the double-edged nature of networks of self-governance. And so for every clarion call to employ the Net to solve the problems of the Net we have second thoughts and worries. He takes a very equivocal position on Net neutrality, arguing (not unreasonably, but no time to go into that here) that we might want to increase third-party liability for ISPs, and does not dwell on the dangers of making ISPs into a part of the “tether.” Invoking tools such as Blossom, which allow Internet users in censored nations to view the internet from the perspective of a less-censored country, he seems to defend “the unfettered development of generative tools that can defy central authority in proportion to the number and passion of those willing to use them.” (p. 196-7) At the same time he recognizes that underground communities which illegally share copyrighted files, sexual images of children or stolen credit card numbers use exactly the same techniques. He recognizes that better, more global identification of users might facilitate more responsible and open sharing of resources, but he also worries that authoritarian regimes could use such identification to track down subversives in their jurisdiction. Perhaps as a consequence of his focus on technical rather than political dimensions, Zittrain hasn’t fully come to grips with the problem of anarchy. (At least Vaidhyanathan’s Anarchist in the Library (2004) explicitly recognizes the issue, but its treatment, too, fails to get us very far beyond the “on the one hand,” “on the other hand” listing of its bad and good features.)

Code vs. politics

Although Zittrain inevitably couches his argument in terms of technical architecture, “tools” and “code,” the problems the book raises are often more political than technical. His analysis suffers a bit from the lack of a more sustained and self-conscious engagement with the political, international and institutional aspects of Internet governance.

A good example is Zittrain’s mention of North Korea, which builds radios that can only be tuned to official government broadcasts. (p. 106) To him it is the ultimate example of the tethered appliance. To me, it shows that we are dealing with a political and institutional problem more than one of technical architecture. In North Korea the relationship between the radio receiving device, spectrum allocation, transmission and content production has been architected to maintain centralized control by the state. This occurs, and probably could only occur, in the context of a one-party dictatorship with full ownership of the means of production. The ability of North Korean people to escape such a system has little to do with technological design. It has to do with finding the political leverage to overturn the dictatorship or establish more pluralism and polycentrism in the society. The problem here is not the presence or absence of some property of “generativity” in the tools; it is that political freedom is needed to support and permit “generativity.”

Zittrain clearly has a bias toward non-institutional, software-based solutions that bypass politics and law. He implies that the solution to the Net's problems is to be found not in politics and new governance structures but in the “computer scientists and geeks who would rather be coding than attending receptions in Geneva or Tunis.” (p. 243) In about one page, he dismisses the WSIS and IGF processes, revealing both a surprising degree of negativity for someone who has only a tangential connection to them, and little awareness of their underlying dynamics. “Their solution to the difficulties of individual state enforcement on the net,” he writes, “is a kind of negotiated intellectual harmony among participants at a self-conscious summit – complex regimes to be mapped out in a dialogue taking place at an endlessly long table, with a role for all to play. Such dialogues end either in bland consensus pronouncements or in final documents that are agreed upon only because the range of participants has been narrowed.”

Of course, IGF doesn’t negotiate or produce final documents but is meant to be a kind of interchange point for policy makers and policy takers. It is ironic that when Wikipedians engage in the same process of give and take to produce encyclopedia entries on asphalt and equipment vendor products it is valorized by some Internet experts as an amazing social innovation, but when open, multi-stakeholder dialogue under the auspices of the UN produce an ongoing international exchange around Internet governance it is dismissed as an ineffectual and stuffy “parade of meetings and consultations.” To be sure, the IGF has its relevance issues. Also, its use of online collaborative tools is inadequate, even primitive. But that can be fixed. The full incorporation of those tools into a context of transnational policy making would constitute a revolutionary organizational transformation for an intergovernmental institution, as it would involve a major shift of agenda-setting power away from states and towards global civil society. If it happens, it would be at least as significant as the production of personal home pages on a social networking site, perhaps a wee bit more so. What Zittrain does not seem to understand is that it will be political movements and public policy dialogues – not software tools per se – that ultimately determine whether efforts to keep the Internet open and free succeed or not. Of course, the IGF is only one of many venues where these politics will play out; there are also other international institutions, communities of geeks, powerful national governments like the U.S., and supranational states like the EU, BRICs, etc.

If Zittrain is worried about the Internet being re-engineered into a regulable, tethered set of appliances, and we agree that he should be, he may need to develop a broader view of what can drive change in this area and what shape the changes should take. While his emphasis on self-defined networks of self-governance is welcome, not all of the distributional conflicts that shape the politics of Internet governance can be or will be finessed through clever code. Governments determined to assert control over things they perceive to be vital to their own interests, such as sovereignty or national security, or states responding to powerful economic interests that have their ear, such as copyright holders, can pull the ground out from under end-user autonomy. Inter-state rivalries and identity politics will occur over Internet governance, just as they do elsewhere. Other kinds of fights over the levers of control are bound to happen.

About ¾ of the way through the book, Zittrain glances across the issue of new institutions while musing about the accountability problems of self-governing networks, saying “The mechanisms of due process and separation of powers adapted by [James] Madison [a leading drafter of the U.S. Constitution] to help substitute the rule of law for plain virtue will have to be translated into those online communities empowered with generative tools to govern themselves and to affect the larger offline world.” Bingo. Zittrain sort of ends where the rest of us are beginning.