Jefferson and the Internet

Posted on January 22nd, 2009 by

I just finished reading a delightful, trim, recently published book, In Search of Jefferson’s Moose: Notes on the State of Cyberspace, by David G. Post.  This book uses Thomas Jefferson as a guide for thinking about the Internet: both what it is like and how it ought to be governed.  The resulting mix of Jeffersoniana and big-picture Internet issues is intriguing, provocative, and (because so well written) a lot of fun.

The first part of the book focuses on describing the Internet, suggesting that we ought to look at it with the same scientific eye as Jefferson turned to his own New World.  Much of this part is based on Jefferson’s Notes on the State of Virginia, for example regarding exponential population growth.  (Post takes time to point out the connections between Franklin, Jefferson, Malthus, and Darwin, before drawing the further connection to the Internet.)  However, some of the most telling examples actually come from outside Virginia, including the moose from New Hampshire.

The moose stands for several propositions.  The first is that size matters; a moose is not merely larger than other animals, it is qualitatively distinct.  So too with the Internet versus other communications networks.  A second, more interesting, point concerns Jefferson’s purpose for displaying the moose in Paris.  The New World contains marvels that contradict the assumptions of the Old, and the display of those marvels may make a more convincing rebuttal to those assumptions than any theoretical argument could.  So too with the Internet, where we find, for example, that Wikipedia is working out far better than it “ought” to.

In the second part of the book, Post turns to questions of governance.  After all, Jefferson was not only the compiler of census data, the mapper of rivers, and the exhibitor of a moose.  He was also the believer in decentralized self government, the “first First Amendment absolutist,” and a cautious realist concerning what today are called intellectual property rights.

Post’s treatment of governance includes not only government action (such as criminalizing certain communications) but also the rule-setting roles played by such Internet gatekeepers as ICANN and by the designers of Internet technologies.

On this latter point, Post stakes out an admirably nuanced position.  He recognizes a degree of truth to Lessig’s point that action in cyberspace is governed not only by law, but by architectural design choices.  However, he also recognizes that those design choices are not completely unconstrained artifices — that there are natural laws, some not yet fully understood, that have some relevance to why the Internet is as it is, and which would prevent us from making it arbitrarily different.  I consider this dual recognition to be a particular strength of the book, one that I hope will be taken up by others and expanded upon.

As should by now be clear, I like this book a lot.  Therefore, I am quite willing to forgive it technical imprecision and even inaccuracy, which would have left me feeling grouchy had it appeared in a lesser book.  However, I can’t in good conscience remain totally silent.

Most of the errors have absolutely zero impact on the points Post is making.  An example would be the howler in which he divides r3 by r2 and gets r3/2, an error he compounds by declaring the result an “exponentially” growing function of r. Post’s point, that a moose is necessarily quite different from an insect, is unaffected by this blunder.

Occasionally Post does manage to undermine the direct point he is making about a specific example, but even then, his larger, more abstract point is unaffected.  For example, he states that IPv6 addresses can be statically configured into devices when they are manufactured, and concludes that end users will no longer be reliant on addresses allocated by Internet Service Providers, with potentially profound effects for their relationships.  In point of fact, the only IPv6 addresses that can be configured in this way are “link-local addresses,” i.e., addresses that can only be used within a single subnet, such as your home.  For routing on the global Internet, addresses will still need to be hierarchically allocated, as with the currently dominant IPv4.  

Post’s error regarding address assignment perhaps reflects his overly abstract view of routing, which he describes as being similar, in essence, to packets randomly bouncing around the Internet until they chance upon their destination.  (He notes that routing does not really work this way.)  This crude randomized routing serves Post’s purpose of explaining the decentralized character of routing, but fails to make clear why addresses need to be hierarchically assigned in order to be efficiently routable.

In any case, although Post is wrong about whether one particular consequence would follow from a switch to IPv6, he is right about his broader point: that technical choices such as this will influence business relationships, personal privacy, freedom of speech, etc.

And although Post’s misstep regarding IPv6 might have followed from his simplification of routing, it is worth pointing out that simplification per se is not a bad thing, if one’s goal is to show the forest rather than the trees.  Post has the wisdom to simplify even where his expertise would clearly have allowed a more detailed approach; for example, he sees no need to differentiate patents from copyrights.

My biggest complaint does not concern a technical point, but rather an economic and policy-making matter.  Post imagines a dialog between Jefferson and Hamilton in 1998, when the U.S. government was going to cease administering the Domain Name System (DNS) through its contract with Network Solutions, Inc (NSI).   Jefferson argues that the government ought to simply takes its hands off and let private parties do as they pleased. Hamilton argues that order must be maintained; he prevails, resulting in the creation of ICANN.  Although there is much to recommend Post’s analysis of this situation, it seems to me that the time was already five years too late for a truly Jeffersonian (or Hayekian) approach to the DNS; by then, the government had already badly tilted the playing field by funneling so much power to NSI.

As a professor of computer science at a liberal arts college, I’m sometimes called upon to explain what computer science has to do with the liberal arts.  Perhaps henceforth I will offer this book up as my answer. Only a dullard could be immune to its joyful story-telling blend of history and technology or fail to be impressed by its big, open-ended questions about how we ought to shape our world.

[For search engines: this is a book review.]



  1. David Post says:

    Max — Thanks for the nice, and thoughtful, review. You’re right — there’s a howler of a mistake in there (r cubed divided by r squared does not equal r raised to the 1.5th power . . . I’ll correct that in later editions, I promise you), and I appreciate the correction. And I also appreciate the comment that “only a dullard could be immune to the book’s joyful story” —
    David Post

  2. as400 guy says:

    Wikipedia is a good example, but the only reason it, and other things work, is because they are protected. They are policed in a way. If there were not, then we would have a free for all…

    Many sites start out likee wikipedia, using the masses to build them up and make them popular, but then they realise they can make more money by charging, so they shut up shop. I have seen this numerous times