In Praise of Erroneous Models

Most historians of science have a favorite example of a now-discarded model whose errors themselves offered a fundamental insight. These are not mere cases of imperfect models, in the way that Copernicus’ model held almost as many inaccuracies as insights, which were refined by later thinkers who nonetheless gave Copernicus credit for getting the fundamentals correct. Instead these are models that have since been resoundingly overturned, but not before something really fundamentally useful had been wrung out of the mistake. This is especially true of the many and varied “subtile fluid” theories of the 18th and 19th centuries – the electrical fluid, the aether theory of light, the caloric theory of heat. Kuhn described the Leyden jar in this way – a groundbreaking instrument first developed because it was believed that electricity was a fluid – why else would you build a jar to put it in? Or consider Carnot, who seems to have been somewhat agnostic about caloric, but made extensive use of that “fluid of heat” theory, vividly describing the “flow” around the engine. Caloric was already a theory on the decline when Carnot published his “Reflections on the Motive Nature of Heat” in 1824. But the metaphor of a fluid only “falling” naturally from a higher place to a lower one, creating work not by being consumed but by it’s transfer of location was a such a significant leap forward that Carnot is usually credited with the formulating the second law of thermodynamics. A more current example is that of the all-pervading emphasis on homeostasis in ecosystem ecology in the 1960s and 70s. This tendency for biological systems to resist change and to remain in a state of equilibrium was stressed in all the best-selling textbooks of the period. Stability was regarded as a “central dogma” of the field. The presumption that a system in equilibrium will remain in equilibrium unless acted upon by an outside force led directly to the development of quantitative methods of calculating material and energy cycling through the trophic levels of a system. While faith in ecosystem stability has been steadily undermined in the last 50 years, calculating movement of energy and material through trophic levels remains fundamental to the discipline. We could consider these examples merely lucky breaks, isolated incidents that in retrospect happened to yield a useful insight that could be reimagined and redesigned to fit into a “new and improved” theoretical conception. It is possible, too, that they are merely a remnant of Whiggish tendencies in the history of science, of highlighting the “correct bits” of quaint old theories in light of modern answers. But I want to argue that the existence, and prevalence, of these examples suggests that erroneous models offer something more valuable – alternative lenses with which to focus on problems, or alternative metaphors to spark our imagination. By considering these examples in more detail I will suggest that we ought to give our erroneous models more epistemic significance than they are usually accorded.