The New Metaphysics
of the Web
Notes and Replies

David Weinberger
Discuss the essay
Go to the essay
Home page

Here are some comments worth discussing but which, for one reason or another, I couldn't incorporate into the body of the essay.

Note: Please post your messages to the discussion board set up for that purpose.

Overstated impact. Kevin Wehrbach points out that "we're still a long way from the Web being pervasive enough to truly change conceptions of the world."

Yes, the article is directed at our culture, as it says in the second paragraph. But I think the Web is having an effect beyond its circle of users. It is an airborne virus.

Conflation. Kevin Wehrbach makes a subtle point, saying I conflate "the notions of context, linkages and pointing." He says that a web site differs from a hammer in that the hammer has a "particular set of conotexts in which they are useful....But, since chains of hyperlinked Web pages are organized by user, they don't necessarily have a defined context. I may see a page as part of a different narrative than you do. In other words, you can have non container-based metaphysics without sympathy."

Hmm. Yes, a Web site constructs a context, whereas the cultural and linguistic context for a hammer is something we're born into. But that's why sympathy underlies the Web and not the instrumental context in which hammers make sense. (More accurately, I'd say that sympathy lies closer to the surface of the Web than in an instrumental context.) And, yes, I suppose a non container-based metaphysics without sympathy is possible, sympathy seems to me to characterize the metaphsyics of the Web.

It's not just the Web. Kevin Wehrbach says that it's not obvious that the Web will be the "dominant expression of the Net." For example, peer-to-peer apps are going to be important. But, Kevin says that peer-to-peer apps are "in some ways the epitome of your point about sympathy..." with which I agree. The fact is that I use the term "Web" loosely, as I think it is and will be used by non-techie-geeks. I'll go back through the article and will try to eliminate the ambiguities when possible. Thanks, Kevin!

It's not inherent in the technology. Kevin Wehrbach saysthat people may "filter and contain their own world ... It comes backto whether or not you're an optimist, but ultimately it's about human nature rather than anything inherent in the technology."

Yes, people will develop their own sub-worlds, or communities. Unless the Web dissolves into separated fiefdoms, I don't see this as an objection at all. But I don't think one's attitude about this is purely a matter of optimism or pessimism. In a sense, I think it is inherent in the technology, or at least in the technology as it is deployed and understood and "contextualized" in our culture(s). It is inherent in telephones that they increase the frequency of communication with distant people, and it is inherent in the Web that it rests upon and develops a common sympathy. That, anyway, is the point of the article.

Read Richard Mitchell. Mike O'Dell writes:

I saw you mentioned the hideous 'students as container' model of 'outcomes-based edutainment.' If you aren't a reader of Richard Mitchell, "circulation manager" of 'The Underground Grammarian,' you should be. He is the most lucid essayist living today. His book are out of print these days but they are available on the web, with his blessings. And funny! viciously funny.

Thanks, Mike. There's an archive of Mitchell's work at

Links are harder than I think. Mike O'Dell writes:

as for the "technical magic" in hyperlinks, i beg to differ. the "indirect reference" is one of the most fundamental and powerful objects in all of computing, and the operation of "following the pointer (arrow)", or "evaluating the reference" is one of the other most important, powerful notions. in all of computing, you have naming and binding and recursion, and from that you can get just about all of it. so there *is* indeed deep magic behind hyperlinks - the fact it *seems* simple doesn't mean it isn't deep - just ask anyone who is taking their first data structures course and is wrestling with the sublime differences between an object, the name of the object, and the address of the object.

As someone who continually makes every mistake in the book when using pointers in C, I understand Mike's point all too well. Nevertheless, Tim Berner-Lee's invention of HTML, distinguished by its use of hyperlinks, wasn't a technical breakthrough. It was an imaginative breakthrough. I don't mean to underplay the wizardry required to build a global, scalable, efficient, robust system that enables me to type in "" and get to Mike's company's site; there's a lot that goes on between my hitting the return key and the UUnet page coming up. But the genius of hyperlinks belongs to the creator of the concept (whoever that may be) more than to the implementors of it.

Containers are invented. Mike O'Dell writes:

in computing, containers are an invented abstraction. we have lots of containers: files, pages, segments of memory, processes, but they are all invented abstractions which require various amounts of machinery to implement. the surface of the disk doesn't know diddly about "your file", much less whether it contains HTML or Old Church Slavonic. the interpretation of a particular substring of bits stored on the surface of the disk is defined elsewhere - in the machinery implementing the varous nested containers involved in "defining" your file: sectors, tracks, cylinders, extents, clusters, inodes, directories, filesystems, mount points, namespaces, processes, programs, and applications which decide to interpret the bit patterns in a particular way. yes, you can think of these things as "containers", but that metaphor is just as invented as the rest of them.

In fact, as Mike knows, even at a higher level of abstraction, containers are inventions on computers. A folder in the file system, for example, is simply an entry in a directory with pointers to its members. Nevertheless, my essay is talking at a higher level of abstraction; I'm not concerned with what's "under the hood."

More than one objective position. Mike O'Dell writes in response to my saying "every place has one and only one objective location":

bzzzzt! every place has more than infinitely many "objective locations", in fact, the order is Aleph(1), which is the order of infinity describing *all* curves (including surfaces) (where Aleph(0) is the infinity of the number of points in N-space for infinite N)

this is true because there are Aleph(1) coordinate systems available to describe the location of a "place", even without General Relativity, but it's the same with it, so that's a nit. but the pictures of Reimann manifolds are sooo wonderfully non-intuitive.

also, the Web has aliases, of course, so there are many names for the "same page" - this may be a bug or a feature, and there are pages which have no name. a more interesting question to ask is whether the space is "metric" - is there a "distance function" which will tell you how far apart two points are in the the space. the best the web can do is "piece-wise metric and not complete" - which means there may be a notion of "closeness" between some places and not others.

I have no doubt that Mike is correct. But I was talking in a less technical sense. Every place on our planet was one and only one position on the map. In N-space this may not be true, but in lived space it is.

Strangers. Mike O'Dell writes:

the entire notion of "stranger" deeply implies a "preferred frame", the "known" as distinct from the "strange". a wonderful false dichotomy, like so many. as usual, the real metric is a floating point number between 0 and 1, with 0 being "self" and "unmet" being 1.

Or, as the Simpsons say in their retelling of "A Street Car Named Desire," "A stranger's just a friend you haven't met." :-)