The Problem of the Language of Smell. A single paragraph that describes our utter lack of a common language to talk about smell.

In fact, the perfumer John Stephen readily agreed with Turin that the Carbon and the Sila smelled very different, the sila being vastly nastier and greener. But simply the ways Turin and Stephen had described the molecule's odors plunged him (when he put them side by side) once again frustratedly into this problem of how to talk about smell. "We run into this problem of the non-standardization of smells," he sighed. "Look at the two of us. I describe the Carbon as 'eucalyptus, tiger balsam, camphoraceous, rich and pleasant.' But John describes the same molecule as 'solvent-like, ethereal, with a sweet fruity powdery backnote, almost marshmallow.' And our descriptions are manifestly different but they are not materially different, and when I smell what he smells, I know we're smelling the same thing because I'm saying Paris is 300 miles south-east of London, and he's saying Paris is 200 miles north-west of Lyon. Still." He ended, tightly, "It's going to be difficult to do data mining, molecule-odorant matching, simply because the language will be so different."