Multiple discovery

Ogburn and Thomas listed 148 inventions and discoveries that had been made independently by two or more people: the planet Neptune by Adams and Leverrier in 1845, the telephone by Bell and Gray in 1876, natural selection by Darwin and Wallace in 1858, the laws of heredity by Mendel in 1865 and then by de Vries and Correns in 1900. They noticed that the rate of such duplicates rose over the centuries, and reasoned that inherited mental ability cannot have changed in so short a time, so the accumulated stock of knowledge, not individual genius, decides when an invention appears. Merton went further in 1961: multiples, not singletons, are the normal pattern in science, and many apparent singletons are multiples that one publication forestalled. With Elinor Barber he compiled 264 of them. The prediction is uncomfortable: if you have an idea, assume someone else has it too.

It holds because discovery is a step from the current state of a field, and once the parts are in place many people can see the next step; the field's open problems point everyone's attention the same way. In practice this is why priority disputes exist at all: Merton found that before 1700, 92 percent of multiples were contested. It is also why the Darwin and Wallace case ran as it did. Darwin had drafted his theory in an 1844 essay and summarised it in an 1857 letter to Asa Gray, then on 18 June 1858 received Wallace's essay from the Moluccas setting out the same mechanism. Lyell and Hooker's answer was to have both read together at the Linnean Society on 1 July 1858, which settled priority without a fight. Speed and publication, not secrecy, are the rational response to a crowded frontier.

The model does not make discovery inevitable, and Merton said so: the ingredients of some discoveries lay in the culture for years before anyone put them together, and Mendel's 1865 paper waited 35 years to be noticed. Dean Keith Simonton points out that such premature discoveries are an embarrassment to the strong "spirit of the age" reading, and that a pure chance model fits the frequency of multiples only partly. There is also a counting problem: how alike two ideas must be before they count as the same discovery is a judgement, and the lists were compiled by people looking for multiples. The sober version is that readiness makes discovery likely and sets its date within a few years, while who takes the step, and whether it is noticed, still depends on individuals.