Metaphor is not just a variety of figure of speech. Metaphor is a scientific tool, an essential part of the scientific method. Scientists, and the rest of us, understand natural and artificial phenomena by comparing them to previously analyzed phenomena, or to the operation of human artifacts. Newton likened the movement of the planets to a clock, and psychologists compare the operation of a brain with the functioning of a computer.
None of this is unusual, since metaphor is also an essential part of understanding and language for all of us. Our daily language and our thoughts are littered with metaphors, used both consciously and unconsciously. Some have been with us so long that their nature is forgotten: chairs have "arms," clocks have "faces," and language can be "littered" with metaphors. We can understand the meanings of abstract concepts as the result of layers of metaphor, such as when we write that the Federal Reserve is "pushing" interest rates up or down.
Attempts to uncover the linguistic structure of metaphor have led to the unavoidable conclusion that it is not merely an expressive tool. Rather, it is clear that metaphors play an important role in the way we understand the world: how we construct our expectations and analyses of the events around us. When metaphors are used to explain, we often use them as a tool to exclude context from consideration. This is doubly true for scientific use. A comparison is made to highlight the relevant parts of the system under discussion, and to brush aside what are considered to be the unimportant parts of some system. This is an important role; it's a messy world out there. But it's equally important to remember what is left out of a comparison, and to remain aware that what is left out is done so by the choice of the metaphor, not necessarily by the data. For example, to compare a neuron carrying an electro-chemical charge to a copper wire carrying an electrical current is to imply that the electrical features of a neuron are the important ones to consider. For the moment, while considering the metaphor, we are allowed to ignore the many ways in which a neuron is unlike a wire, such as its susceptibility to hormones, or the chemical nature of the signal it carries.
There are several dangers here. For one, it often turns out that the real-life context excluded from the metaphor is far more important than was previously thought, perhaps even more important than the phenomenon described. You can see this in several cases:
Sometimes the reverse happens, when a comparison is made to a simpler system that turns out to be not really that simple.1 Economists often compare the effect of market forces on commodity prices to the action of a spring. In the sense that both exert forces toward an equilibrium position, this could be considered true. But even if we grant that market forces push toward an equilibrium, it doesn't mean the resulting system is simple. Springs may also push toward an equilibrium position, but that doesn't mean that the load they're pushing will ever get there, or that it will stop when it does. Nature is filled with complex phenomena that result from forces as simple or simpler than a spring; acoustics, planetary motion, and radio communication are all based on quite "simple" forces. But the resulting dynamic systems are far from simple.
Another danger lies in making the slip from comparing a natural phenomenon to a simpler system to thinking that the natural phenomenon is the simpler system. Brains are like computers in some general sense of information processing, and even have some superficial similarities in terms of structure. But to identify them as computers, given the kinds of computers with which we are familiar, is simply mistaken. Whatever their general similarities, they are significantly unlike any electronic computer yet designed, built, or imagined.
Still another source of peril is when the metaphor proves more durable than the science that produced it. In several cases, metaphors have been popularly adopted and retain wide support, even though the science that produced them has long since moved on. The simple picture of electrons orbiting atomic nuclei, like planets around the sun, was discarded by physicists 75 years ago, but continues to be part of the popular image of what an atom looks like.
Evolutionary biology provides a comparable example. There is a great deal of friction among evolutionary biologists about how to understand the process of natural selection. What these various factions can agree on, however, is that the simple equation of evolution with progress toward greater complexity and sophistication of purpose is a simplistic and misleading conception. But that's the way "evolution" is broadly defined, that's the way it gets used in daily discourse, when you hear about the "evolution" of a car, an ad campaign, or a movie script, and most importantly, that's the way people understand it when they re-apply it to the science that produced the term in the first place.
One can't argue that metaphor should not be used in scientific analysis. Metaphor is one of the fundamental ways we understand the world; we could no more give it up than we could volunteer never to use our eyes. The comparison between DNA and language, for example, provided a crucial insight: that DNA ought to have punctuation. The analogy provided the prediction, and subsequent analysis proved it quite true. But other inferences from the same metaphor have not held up; DNA has no "sentences", and even "words" are a troublesome concept in that context. What's more, DNA actually represents only a fraction of the information needed to make a growing child.
There is an important contrast to make between a scientific paradigm--the framework within which explanations are offered--and the use of scientific metaphors, a more literary exercise. The progress of physics during the 1920's advanced largely in spite of the lack of available metaphors for the disturbing findings of the quantum mechanics revolution. Contrast this with the comparable revolution in genetics during the 1950's, when the field became so taken with the literary value of the metaphor that contributions from people like Barbara McClintock were ignored at least in part because they couldn't be accommodated by the language.
The thesis presented here does not argue, as several post-modernist theories of science and language do, that reality is a construct of the ideologies and predispositions of the observer. The Computer Between My Ears is about how accurately certain metaphors reflect the underlying reality. That is, some constructs are demonstrably wrong, a point that eludes many theorists. The reality is there, the issue is how we talk about it, and the inferences we make from the language we use.
Metaphors such as the ones mentioned here are valuable analytical tools, but they must be used with care, with due consideration of the aspects of a system left out of the story by the comparison used to explain it.
This proposal is for a book, The Computer Between My Ears, containing a treatment of the general subject of the misuse of metaphor in science, as well as a closer look at several fields where the dominant metaphors have demonstrated--and in a few cases caused--serious problems for the progress of that field.