Can humans really tell apart a trillion smells?

Nobody knows. A famous 2014 study said that humans can tell apart at least a trillion smells, but other scientists responded that the calculation behind that number does not hold up, so the honest answer is that the true number is still unknown.

For decades the standard figure was 10,000. The lay and scientific literature typically claimed that humans can discriminate 10,000 odours, but that number had never been empirically validated. It came from impressionistic findings in the 1920s. Then the 2014 researchers tested people on their ability to discriminate odour mixtures with varying numbers of shared components, and from the results they calculated that humans can discriminate at least 1 trillion olfactory stimuli. They said even their worst performer could probably distinguish 80 million scents.

Critics struck back. The Caltech scientist Markus Meister wrote that the study's claims were based on errors of mathematical logic, arguing that the failure lies in the method used to infer the size of odour space from a limited experimental sample. A second group argued that the estimate came from a fragile framework that could produce nearly any result from the reported data, and that the formula gives an upper bound, so the calculation supports at most a trillion, not at least a trillion. They concluded that there is no evidence for the original claim. The original authors answered Meister's logic with criticism of their own.

There is also a subtlety hidden in the headline. Telling smells apart is not the same as naming them: the study's own authors noted that people could not typically identify the individual components within the mixtures. So the safest reading is this: a trillion is a claim rather than a measurement, and nobody has yet pinned down the real count.