#171 from R&D Innovator Volume 4, Number 8          August 1995

Analogy: The Creative Spark for Invention
by Keith J. Holyoak, Ph.D., and Paul Thagard, Ph.D.

Dr. Holyoak is professor of psychology at the University of California, Los Angeles, and Paul Thagard is professor of philosophy at the University of Waterloo, Canada.  They have collaborated for many years in studying the use of analogy--reasoning from past cases to new problems--in applications ranging from everyday problem solving to scientific discovery.  This article is based on their book Mental Leaps: Analogy in Creative Thought (MIT Press, Cambridge, 1995). www.psych.ucla.edu/faculty/holyoak/default.htm

The launch of the Hubble Space Telescope in 1990, heralded as a triumph for the beleaguered NASA, quickly became a high-flying embarrassment when the telescope's would-be users realized that its mirror had been ground improperly.  The most powerful eye on the universe turned out to have blurry vision, and a fix was desperately needed.

Where did the idea for a solution originate?  In the shower of a German hotel room, when a soggy NASA engineer James Crocker was looking at a shower head that could be extended to adjust to the user's height.  Crocker made an analogy between this plumbing fixture and the myopic telescope, and invented the idea of placing corrective mirrors on automated arms that would reach inside the telescope and adjust to the correct position. 

The resulting device, COSTAR (corrective optics space telescope axial replacement), was considerably more complex than a shower head.  It consisted of eight motors attached to five metal arms holding ten coin-sized mirrors.  In 1994, NASA used COSTAR to turn Hubble from a debacle into a triumph, after a space shuttle mission installed the corrective lenses with millionths-of-an-inch precision into a seemingly inaccessible part of the space observatory, while Hubble orbited 300 miles above earth.

Other Examples

The leap from an adjustable shower head to a corrective lens is only a recent example of a creative process that has often sparked major advances in science and technology.  Historians, philosophers, and psychologists of science have documented many instances of analogical thinking. 

The first recorded use of analogy in a landmark scientific theory dates back at least 2,000 years, to the recognition that sound is propagated as waves.  A Roman architect and engineer, Vitruvius, noticed several important similarities in the behavior of sound and water waves, such as the parallels between echoes and the rebound of water waves after striking an obstacle.  In modern jargon, the water waves provided the familiar source analog, which Vitruvius used to understand the unfamiliar target analog, the nature of sound.  This mental leap was eventually followed in the seventeenth century by another giant advance--a new analogy drawn between sound and light, which led to the wave theory of light.

Since science is the foundation for technology and engineering, it’s not surprising that analogies have contributed to new inventions, of which COSTAR is neither the first nor the most significant.  For example, when Alexander Graham Bell was inventing the telephone, he explicitly used an analogy to nature:  "Make transmitting instrument after the model of the human ear.  Make armature the shape of the ossicles.  Follow out the analogy of nature."  He drew diagrams that made explicit the structure of the ear and then drew analogous diagrams of physical devices that could transmit vibrations in a similar manner.  His analogy thus employed visual as well as verbal representations.

Another technological breakthrough based on a visual analogy to nature is the Velcro fastener.  In 1948, George deMestral noticed that burdock burrs stuck to his dog's fur with tiny hooks.  He figured out how to produce the same effect artificially, and now shoes and countless other objects are fastened with burr-like hooks and cloth-like loops. 

Like the telephone, Velcro was a technology that drew on a natural analog, but in this case the technology was later used as a source analog.  Thus the hook-loop concept led to discoveries such as abdominal closure in surgery, epidermal structure, molecular bonding, antigen recognition, and hydrogen bonding. 

Of course, biological analogs may fail, especially when they are shallow (remember Icarus' vain attempt to glue feathers to his arms to emulate a bird's wings?).  Or sometimes they may be partly successful:  obviously, the Wright brothers used aerodynamic principles of lift and drag, rather than use the flapping of wings to produce lift in the first working airplane.  Nevertheless, they did use an analogy with birds: when Wilbur Wright noticed that soaring birds twist their wings to restore balance, he warped the wings of the aircraft to achieve the same effect. 

Making the Mental Leap

To propose an analogy, or simply to understand one, we must take a kind of mental leap.  Like a spark that jumps across a gap, an idea from the source analog is carried over to the target.  The two analogs may initially seem unrelated, but the act of making an analogy creates connections between them.  Nothing guarantees that the target will behave the way the source suggests it might.  Some analogical mental leaps have ended in creative triumph, such as the wave theories of sound and light, and the invention of the telephone; others have ended in dismal failures.

Although making an analogy requires a leap, that leap need not be blind or random.  Recent work in cognitive science indicates that reasoning by analogy is guided by a few cognitive constraints, including these three basic types:

•           finding correspondences based on shared structure.  Sound waves and water waves, for example, share such structural properties as rebounding after striking a barrier, and diminishing in intensity with distance from their origin.

•            mapping elements that are directly similar.  This is why it seemed so natural to model airplane wings on bird wings--both are immediately recognized as belonging to the same category.

•            identifying correspondences that will achieve their purpose in using the analogy.  The engineer who jumped from the adjustable shower to the near-sighted telescope focused on how the shower's height could be continuously adjusted (rather than, say, the temperature of the water it emitted) because height adjustability was relevant to his problem. 

The use of good analogies can be advanced by social means, and research groups that foster analogies can gain a creative edge.  Kevin Dunbar has described (in Mechanisms of Insight, MIT Press, 1994, Cambridge, MA; R.J. Sternberg and J. Davidson, eds.) how the social structure of laboratories can greatly affect the use of scientific analogies.  In labs where researchers have diverse training, productive analogies frequently emerge in meetings.  A target problem that has stumped one researcher can be approached with a source analogy from another researcher with a different background.  Like broader human culture, science and technology are products of social interaction and individual cognition.

Finally, recall that analogy provides plausible conjectures, not infallible deductions.  Inferences generated by analogy must always be tested to see if they’re actually helpful.  But when done carefully, thinking by analogy tends to produce ideas that are worth considering, even if they ultimately prove flawed.