Thought Experiments

History

Actual thought experiments predate the term thought experiment, which was first used by Georg Christoph Lichtenberg in 1793 and only acquired wide currency after Mach's 1897 essay “Gedanken experiment.” Two thousand years before the term was invented, poets imported thought experiments into metaphysical verses about the nature of the universe. In his epic De Rerum Natura, the Roman poet Titus Lucretius supports the infinity of space by having the reader imagine that space has a boundary. If a man threw a spear at the edge of space, either the spear passes through, and so there is no edge of space, or it bounces back, in which case there is a barrier that lies beyond space and yet is in space.

Rationalists welcome thought experiments. This a priori method shows that we can learn about the world without experience! Explaining how this happens is a classic problem—one that Plato hoped to solve with his Theory of Forms. Contemporary philosophers, such as James Robert Brown, continue to work on an explanation within the Platonic framework.

Empiricist Criticism

Empiricists have trouble understanding how thought experiments could be any more informative than poetry. If everything is learned from experience, then no information about the world can be extracted from a hypothetical experiment.

Yet many attacks on Aristotle's theory of motion are launched from Galileo's armchair. Consider his critique of Aristotle's principle that heavier objects fall faster than lighter objects. Galileo supposes that a smaller object becomes attached to a heavier object during its descent. On [Page 1507]the one hand, the composite object must slow down because the heavy component is retarded by the slower component. On the other hand, the composite object must accelerate because it is heavier than the heavy component.

Empiricist reactions to thought experiments fall along a spectrum. Pierre Duhem condemns thought experiments as illusory sources of support. They have no more place in science than Plato's numerological arguments for the existence of seven planets.

John Norton thinks the success of Albert Einstein's thought experiments shows that Duhem was too dismissive. According to Norton, thought experiments should be heeded exactly to the degree that they can be reduced to cogent arguments. After all, empiricists endorse changes of opinion that are based on surprising deductions and calculations (even if the premises and data were already known).

Mach's Empiricist Account of Thought Experiment

Einstein's early philosophical mentor, Mach, was more concessive than Norton. Mach believed that thought experiments tap into stored “instinctive knowledge.”

The pictorial form of this knowledge is evident from how questions such as “How many corners are in the capital letter preceding G?” are answered. A mental image of the letter is formed and the corners counted. In a sense, the number of corners was not known (they had to be counted), and in a sense, it was known (all the information was available to quickly answer the question). Information is stored in mental images in the way information is stored in crime scene photographs (which, if need be, can be consulted later for missed clues).

...