Sep. 18, 2013 at 5:32 PM ET
For years, humans have had categories for colors, flavors and sounds, but when it comes to the sense of smell, things are a mess. Now, just as there are three primary colors and five basic tastes, researchers propose that odors can fall into 10 basic groups.
In the new scheme, woody smells like pine or fresh cut grass fall into one group. There are the sweet scents, like caramel, chocolate, vanilla. Florals and perfumes are one category, citrus fruits get another, and the rest of the fruit world is lumped into a third.
Less pleasant odors like sour milk, gasoline and rotten meat have their own fetid designations. Rounding out the list are minty smells — which include eucalyptus and camphor — and the smells of toasted, nutty snacks like popcorn, peanut butter and almonds.
More complex aromas like baked bread or fresh-brewed coffee might be best described as a combination of two or more of these 10 elements, just "as a combination of pitches make up a chord," Jason Castro, associate professor of neuroscience and one of the authors of the new study published in the Sept. 18 issue of PLOS ONE, told NBC News.
Humans have arranged colors in grids and pitches on a scale. We know that the wavelength of light determines the colors we see, and that we understand the frequency of a tone as a pitch. But an organizing principle for colors? "It doesn't exist in any culture," Castro said, and part of that is because the science of smell is far more complex.
Some of that has to do with biology. "In the case of color, it's simple," Hiroaki Matsunami, a molecular biologist who studies olfaction at Duke University, told NBC News. Three primary colors — red, blue, green — correspond to three primary color receptors, and a combination of those yield any color the human eye can see.
But when it comes to our sense of smell, "we have 100s of receptors so therefore by analogy we should have 100s of primary odors," he said.
So Castro and his colleagues Arvind Ramanathan and Chakra Chennubhotla didn't start out with 10 categories firmly in mind. Instead, they started with a historic list of smells, an olfactory "atlas" created in 1985 at the Institute of Olfactory Sciences in Park Forest, Ill. From this group, they extracted 144 words like "fruity" and "pungent" and "smoky," then sought out statistical relationships between them. Did people tend to use certain words in key combinations?
From the statistical mill, links emerged and categories materialized. "If you take any given smell ... it would be best characterized by one of these perceptual 'buckets,'" Castro said.
Besides describing fine wine, why would humans even need so many categories for smells? In ancient times, distinct smells of meat, or fruits, or citrus, could ensure humans ate those food groups, or were perhaps more sensitive to those smells because humans needed those citrus and meat to survive, Castro supposes.
"One interpretation we like is that these may reflect the essential portions of diet ... things you will actively seek in the environment," Castro said.
And equally important, smells tip us off to what not to eat: Words that people picked to describe nastier smells like rotting garbage, or decaying meat, gas, sour milk or poo, clustered into three clear groups (with a few palatable anomalies thrown in, like garlic and vinegar).
The new study doesn't explain the chemical experience of smelling, it explores how humans understand and then describe the experience. "I can see that it's not perfect, but if I believe what they say, this would be a good approximation of how we process odor," Matsunami said of the new study.
More about the science of smell:
The authors of "Categorical Dimensions of Human Odor Descriptor Space Revealed by Non-Negative Matrix Factorization," are Jason Castro, Arvind Ramanathan and Chakra Chennubhotla.