How Conceptual Metaphors are Stunting Web Innovation

Venkatesh Rao is a researcher in the Xerox Innovation Group, and the project manager for Trailmeme, a research beta technology that allows users to blaze and follow trails through web content and the Trailmeme for WordPress plugin. He blogs at

As much as we focus on developing new technologies, it is also essential that we break free of certain metaphors that bind and restrict our thinking about what these technologies can ultimately achieve. The familiar “document” metaphor, among others, has cast a long shadow on how we think about the web, and is standing in the way of some innovation.

The Conceptual Metaphor

In his classic study of media theory, Understanding Media, Marshall McLuhan wrote, “We look at the present through a rear-view mirror. We march backwards into the future.”

Consider these terms: page, scroll, file, folder, trash can, bookmark, inbox, email, desktop, library, archive and index. They are all part of the document metaphor, a superset of the “desktop” metaphor. Some elements, such as scroll, desktop and library pre-date the printing press, but all are based on some sort of “marks on paper-like material” reference.

It is important to understand that the document metaphor is more than a UI metaphor. It is in fact a fundamental way of understanding one domain in terms of another. For better or worse, we continue to understand the web in relation to how we understand documents. Unlike figurative metaphors, such as “he was a lion in battle,” which are simple rhetorical statements, conceptual metaphors (a notion introduced in the classic “Metaphors We Live By” by Lakoff and Johnson) like document-ness are pre-linguistic, and quietly ubiquitous. They infiltrate how we think about things on a much more basic level.

Did it ever occur to you that the phrase “the stock market is up” is actually a particular spatial metaphor for what is really just a number? As a result, we think of the stock market as a geography, which has non-trivial ramifications for how we make decisions about it.

This is often a good thing — conceptual metaphors can be helpful. In dealing with novel phenomena, we often have no choice but to understand the new in terms of the old, the complex in terms of the primal, the abstract in terms of the tangible (companies often pitch themselves according to this logic, i.e. “we’re like FriendFeed for dating”). Accordingly we often conceive of new features, new business lines, and new market opportunities in the same way.

The Tyranny of the Document Metaphor

Conceptual metaphors aren’t always a good thing, though, helpful as they may be. A conceptual metaphor enriches your thinking in some directions and impoverishes it in others. It can become a crutch, and a burden.

Consider the terms open and close for digital documents. Serviceable though they were in the early eighties, they make little sense for the live, constantly evolving web “page.” For a rapidly changing page, the pause, play and rewind metaphor borrowed from music player UIs is more appropriate, something the Google Wave team has recognized, for example.

As a technology evolves, the metaphor struggles to keep up. It becomes increasingly strained. McLuhan’s “medium is the message” phenomenon starts to really kick in, as users encounter the limits and biases of the medium.

In the early days of computing, we needed only a few terms, such as click and double-click, to mitigate the deficiencies of the document metaphor. Today, we make new demands of the metaphor every day, and it fails us regularly. Consider the irony of your Twitter home “page” that can “scroll” much faster than you can “read.”

The solution? Look for new conceptual metaphors.

Liberating the web, Conceptually

Let’s continue with the Twitter example.

The conceptual metaphor of a party, with many overlapping public conversations, works much better than the document metaphor. Sophisticated users keep Twitter in their peripheral vision, where it behaves more like an oral medium that you “listen” to in the background, rather than “seeing” it in the foreground, which the “document” metaphor encourages. Note the deficiencies of the conversation metaphor though: it does not cover the Twitter link economy, or asymmetric following. These are better understood through a “marketplace” metaphor, which, however, downplays the conversational aspects.

Likewise, the metaphor that we currently seem to be embracing for the web is “the stream.”

The emergence of the real-time web has finally precipitated the need for a more dynamic framing, and while the stream is accessible and understandable, it is not without its limitations. The flow of information and our “jumping in and out of the stream” may actually point us in a dangerously passive direction. We may dam a stream, redirect it or harness its power for other uses, but the stream remains a metaphor that emphasizes precisely our inability to control or effectively influence or filter it.

Such are the trade-offs in engineering new metaphors. Google Wave is based on a flux metaphor. YouTube borrows a “channel” metaphor from television. The research project I manage, Xerox Trails, is based on the tricky “trail” (as in hiking) metaphor first proposed by Vannevar Bush in 1945.


Central to all these programs of metaphor re-engineering is a recognition that the hyperlink is the basic building block of the web. Our conceptualization of the web still does not truly reflect its non-sequential, branching texture, created by hyperlinks.

We still haven’t truly understood that click and link are as fundamental today as read and write.

All in all, throwing off that burden is an immensely difficult task. It is much easier to create technology that conforms to dominant metaphors. What we need to do as we enter the third decade of the web, however, is consider what we want the web to be rather than awkwardly fitting that vision into older descriptive paradigms.

We need to finally begin articulating the metaphors that will move us beyond the book, and the document. Understanding the rhetoric of the hyperlink may be the most essential challenge we must meet before we are able to move our thinking forward and accommodate our digital ambitions.

Images courtesy of IconArchive and iStockphoto, scherbet, kickers and shulz

There’s some important concepts for consideration in this article.
As I’m writing up my PhD thesis, I recall vividly my attempts to explain how i was using weblogs in the learning context – this was not only for Ethics approval and satisfying lawyers of risk issues, but also other academics and the students themselves.
Upon reflection, I’m wondering if my descriptions did a disservice by limiting conceptual frameworks – or if the use of metaphor enabled capabilities that allowed participation and collaboration based on shared-meanings?
BTW – Lakoff & Johnson is a brilliant reference for those working with metaphors 😉