You are here: Resources --> 2006 --> Editorial: Overextending metaphors
Vous êtes ici : Ressources --> 2006 --> Editorial: Overextending metaphors

Editorial: Overextending metaphors

by Geoff Hart

Previously published as: Hart, G. 2006. Editorial: Overextending metaphors. the Exchange 13(3):2, 6–8.

The world is a complex place, and understanding that complexity can be a thorny challenge. Different communicators solve that problem in a range of interestingly different ways. Mathematicians, for instance, create abstractions that describe the world by combining the symbols of mathematics into equations. To those of us who work with words, this is clearly analogous to creating words from combinations of letters, and sentences from combinations of words—though the words and sentences of mathematics are entirely incomprehensible if you don't speak the language. For many aspects of reality, particularly in sciences such as physics, an equation provides a sufficiently complete and accurate description of reality that those who speak the language understand the message more clearly than would be possible with any other form of communication. For those of us who aren't mathematicians, we need something a bit more concrete to help us wrap our minds around a complex concept.

Abstraction is the process of simplifying something complex by eliminating details until only the essentials remain. Just as mathematicians abstract their world using equations, nonmathematicians also use abstractions—but more familiar ones. One effective approach to abstraction involves the creation of a metaphor that compares something we don't yet understand with something that we do understand. By demonstrating the similarities between the known and the unknown, the metaphor builds a bridge between the two. Crossing that bridge provides the necessary understanding. (Indeed, my example of bridging a gap is one such metaphor.) Seen from the perspective of how human memory works, metaphors let us connect new information with information that already exists in the brain, and such connections are a powerful tool for creating meaning.

One problem with metaphors is that they can be carried too far: because a metaphor is only a simulation of reality, it does not precisely or fully match that reality, and each mismatch can potentially lead to misunderstanding. Consider, for example, the trash can used to delete files in most graphical user interfaces. The Macintosh interface designers who chose this metaphor to describe how users discard files chose an obvious and effective metaphor because just about everyone understands how a trash can works. But unfortunately, a great many users took that metaphor places its designers never intended. When this interface choice was first made, many Macintosh owners used their computer at home or in a small graphics studio rather than in a large corporate workplace, and thus used their experience with trash cans to make the following assumption: "When I throw something in the trash, it's going to stay there forever, or at least until I can persuade someone to take out the trash." Unfortunately, the first implementation of the Macintosh trash can automatically emptied the trash when you shut down your computer. That was clearly a problem for anyone who expected the discarded files to still be there waiting for them the next day when they turned on their computer.

So many people complained about losing precious files (never mind that these files should never have been in the trash in the first place) that Apple changed the interface. Version two of the trash can accounted for this problem by leaving deleted files in the trash until someone specifically told the computer to empty the trash. That's a great idea, except by then, the world had moved on and more Macintosh users were using their computer in the workplace, where a janitor could be relied on to empty the trash each night after the workers went home. Since that wasn’t the way the software actually worked, the inevitable consequence was that files accumulated in the trash until they took over the entire computer; in other cases, people deleted files that were potentially embarrassing, not realizing the files were still there to be discovered by anyone who went poking around in the trash. Clearly, another small interface failure; unlike a spouse or roommate, the Macintosh operating system doesn't remind you to empty your trash periodically.

A future iteration of the interface will presumably strike the right balance between versions one and two by retaining information in the trash until you specifically delete it, but also by periodically providing a gentle reminder to empty the trash. This example illustrates an important rule for successful use of metaphors: you must strive to understand the consequences of the metaphor by asking yourself what users will think when they encounter it, and thus, how they can be expected to behave. Where some behaviors will prove damaging, we need to clearly communicate the problem and its solution in our documentation. Better still, we need to report the problem to the designers of a product so they can take appropriate measures to protect users from their own instincts.

Another significant problem with metaphors is that they rely on certain assumptions, and those assumptions bias how we think about reality. One of the most famous (some might say infamous) relates to a favorite device of science fiction writers: time travel. Science fiction writer René Barjavel, in pondering the implications of time travel, wondered about what quickly became known as the grandfather paradox: What would happen if you traveled back in time to a date before your parents were born, and killed one of your grandparents? Clearly, this means that one of your parents would never have lived, and thus could not have conceived you; the result, a few years into the future, is that you would never exist to return and kill that grandparent. But because you did not kill the grandparent, your parent would be born, leading to your birth and your subsequent desire to travel back in time and become a murderer. Round and round we go until we give up in frustration and choose a convenient way to avoid the problem—declare that time travel is impossible.

Whether or not time travel really is impossible, that would be an unfortunate choice, because paradoxes are crucially important in science: they reveal when we don't understand a process nearly as clearly as we thought we did. If we did understand fully, there would be no paradox. The grandfather paradox presupposes that we understand how the physics of time travel would really work, namely that there is an indestructible connection between the past and the future and that changing the past would inevitably change the future. Should we stop there, no one would ever examine time travel in more detail to see whether other possibilities exist, and that would rob us of a much richer understanding of our world. One consequence might be the elimination of the branch of mathematics that examines the "many worlds" hypothesis, in which a whole new universe is hypothesized to spring into existence as soon as we change the past. In the case of the grandfather paradox, this means that two universes (one in which you are born and one in which you are not) would move forward through time from that point onwards. In writing a story, I once proposed a different metaphor: that time is more like a VHS tape, and that if you go back and change something, this is no different from recording over an old program you've already watched. The future (the part of the tape after the new recording) isn't changed because you haven't overwritten it yet.

Both metaphors may be entirely incorrect (as seems likely based on our modern understanding of physics), but their correctness is not the important issue here: what's important is how each metaphor biases the way we think and predetermines the kind of analysis we're prepared to consider. Thus, a second rule of successful use of metaphors is that we must take great pains to understand the constraints they place on our thoughts. If we're aware of those constraints, we can attempt to work around them; if not, we won't make that effort, and that may prevent us from making crucial new discoveries.

A third problem arises if we oversimplify our description of reality and thus neglect key issues. Consider, for example, the issue of fighting forest fires. Because mature forests develop over time spans longer than the typical human life, it's natural for us to think of them as eternal. Because we now understand the value of "untouched" nature, the inevitable consequence is that we want to preserve old forests and protect them against fires. This belief is epitomized in the public consciousness by Smokey the Bear and the "only you can prevent forest fires" slogan. Although it's true that human-originated fires are a serious problem, and should often be fought, the often part is neglected. In particular, the limited worldview offered by Smokey the Bear ignores the fact that fires are a crucial part of natural ecosystems and that some forest ecosystems only develop after fires, and will eventually disappear from the landscape if natural fires are not allowed to burn.

The more general point is captured by the cliché that "the only constant is change". Ecosystems, including forests, aren't truly stable; instead, they exhibit what is known as metastability, in which what seems stable from the outside is actually changing continuously. In a forest, old trees die, unlucky trees are felled by lightning or windstorms, and new trees sprout to take their place. Rather than perfect stability, a mature forest is in equilibrium: individual components change, but the overall ecosystem stays close to its current state. Yet these equilibrium states also change; if the environment changes, or if disturbances such as fire are prevented, natural processes will lead the ecosystem to change into something new, and a new equilibrium will develop. For example, in the absence of fire, boreal jack pine forests will be replaced by shade-tolerant decidous trees that grow in the limited light beneath the forest canopy. As the older trees die, they are replaced by younger decidous trees, which produce so much shade when mature that the pines can no longer survive.

The problem with describing ecosystems as stable is that it conceals the important concept of dynamic equilibrium, and the consequence that any equilibrium will eventually shift to a new type of equilibrium. This means we can never preserve a specific ecosystem in its current state forever, and that we probably should not try. Instead, it is more important to preserve the conditions that allow a given site to evolve naturally from one equilibrium state to another ("succession"), while altering conditions elsewhere to permit the development of the desired ecosystem. Communicating more of the complexity provides the necessary bounds on the metaphor, permits a more complete understanding, and lets us choose wiser management strategies.

A third rule for successful use of metaphors is thus that we must identify critical points of failure—places where the metaphor is insufficiently complete that it leads our audience astray—and must provide the missing complexity that will prevent this misunderstanding. We must recognize that the purpose of a metaphor is to facilitate understanding, but once that understanding exists, we must build on it to provide any missing details that explain the true complexity.

As scientific communicators, we often resort to metaphors because of their power to facilitate understanding. But to use metaphors successfully, we must be conscious of the problems I've identified in this essay: we must identify mismatches with reality, implicit and explicit assumptions, and places where the metaphor is too simplistic. Understanding these three problems lets us help our audience to understand the mismatches between the metaphor and reality, remind them of the assumptions behind the metaphor so that they can challenge those assumptions and make conceptual breakthroughs, and recognize where we have oversimplified a complex reality. That oversimplification is only acceptable if it provides an initial understanding that we can subsequently build upon to create a deeper, richer understanding.

©2004–2024 Geoffrey Hart. All rights reserved.