The Power and Pitfalls of Analogical Thinking in Science
Paragraph 1
Scientists have long acknowledged the power of metaphor and analogy. Properly understood, analogical and metaphorical thinking are not merely ornamental aspects of language, but serve as a bridge from the known to the unknown. Perhaps the most important example of this process was the one that epitomizes the scientific revolution: Isaac Newton's realization that both heavenly and terrestrial bodies were governed by the same physical laws. A precise mathematical analogy exists between the falling of an apple and the orbit of the moon around the earth. The moon can be thought of as a really big and far-away apple that is "perpetually falling". Newton's analogy rests upon a broadening of the concept of free-fall — in other words, it involves a more abstract concept of motion. A couple of centuries later, James Clerk Maxwell recognized the process of generalization and abstraction as central to the scientific enterprise. The new sense of an idea, "though perfectly analogous to the elementary sense, is wider and more general. These generalized forms of elementary ideas may be called metaphorical terms in the sense in which every abstract term is metaphorical." We might go so far as to call metaphor the alchemy of thought — the essence of creativity.
Paragraph 2
Words like "abstraction" and "generalization" can often appear neutral, or even positive, depending on your intellectual tastes. Nevertheless, there are drawbacks to these unavoidable consequences of analogical thinking. The one that most often receives comment from scientists and philosophers is the fact that analogies are only ever partial: there are always differences between things and processes — that is how we know that they are not identical in the first place. In other words, every abstraction involves a loss of specificity.
Paragraph 3
If scientists discover processes that are "perfectly analogous" to each other, as in the Maxwell quote above, then this loss is so minuscule that it does not cause any real problems. However, in areas of active research, much more circumspection is required. When we propose that one system serve as a model for another that we do not understand, we must be careful not to lose sight of the inevitable differences between the model and reality. Otherwise, we may confuse the map with the territory, forgetting that a map can only serve as a map by being less detailed than what it represents. As a model becomes more detailed, it eventually becomes just as complex as the real thing, and therefore useless as a tool for understanding. As the cybernetics pioneers Arturo Rosenblueth and Norbert Wiener joked, "the best material model for a cat is another, or preferably the same cat."
So the loss inherent in the process of analogy cannot be avoided through additional detail or specificity. In any case, any fastidious adherence to strictly literal language severely retards our ability to create new knowledge. If we seek any kind of usable understanding, we have to use analogy, taking care to watch out for the inevitable places where our analogies will inevitably break down.
