Of course, the Barabási study has obvious limitations. Its recipe index is necessarily incomplete, as is the data source on flavor compound ingredients. It makes no predictions, such as of pairings that, some groups might find repugnant. Most significantly, the team's analysis ignores variables such as how much of any chemical is present in the ingredient and in the dish (that is, is it detectable to the palate), what form of the ingredient is used in any given recipe, and how the food is prepared.
But those criticisms are beside the point. Analyses like these are essays in the craft of computational analysis. Online recipes represent massive, easily accessible, and unstructured data sources. Wouldn't it be fun to download them, parse them, and see what shakes out? Indeed, at least two other research teams have published network analyses of online recipes: Lada Adamic at the University of Michigan and Ronaldo Menezes at the Florida Institute of Technology, who combed these resources looking for patterns of recipe substitutions (such as apple sauce for olive oil) and cultural influences, respectively.
“To be extremely honest with you, we did it initially just for fun,” says Menezes.
Yet these studies are intriguing, if for no other reason than they demonstrate the application of systems biology (and social networking) approaches to very non-biological datasets.
“I think it's an interesting exercise,” says Marcia Pelchat, a neuroscientist and Associate Member of the Monell Chemical Senses Center in Philadelphia, of the Barabási paper. “It shows how far we've come in our ability to analyze the flavor components of food.”
Robert Wolke, a professor emeritus of chemistry at the University of Pittsburgh, is the author of several books on the science of food, including What Einstein Kept Under His Hat: Secrets of Science in the Kitchen. A former nuclear chemist, Wolke spent 10 years writing a food-science column for the Washington Post in which he answered readers’ questions and debunked many food myths that permeate American kitchens. Over the years, Wolke has tackled such questions as why some hams can be stored at room temperature (Answer: Because they're cured), and why a woman's Dutch oven exploded (Answer: alcohol vapor, meet heating element).
While Wolke wasn't intimately familiar with network analysis in general, or Barabási's work in particular, he sees promise in such approaches. “It sounds like a very interesting start on organizing food and flavors, which is an incredibly complex, wide-ranging field.” He allowed that: “When it matures—and I don't think anybody would claim it's mature now—it can be very useful.”
In many ways, understanding the chemistry (and physics and biology) behind cuisine is at the heart of a scientific discipline known as “molecular gastronomy.” Coined by French chemist Hervé This and English physicist Nicholas Kurti, molecular gastronomy looks at the mechanisms underlying certain kitchen phenomena.
This has applied approaches including fluorescence spectroscopy, NMR, and GC-MS, to questions such as why carrot stock turns brown (it can't be the carotinoids, they are not soluble in water) and why a chemical found in tarragon is toxic on its own, but not in tarragon leaves themselves. He works with chef Pierre Gagnier to put some of his observations to the test.
But when it comes to the Barabási study, This was unimpressed. On the technical side he questioned the team's methods, asking, for instance, just what exactly constitutes a “Western European” or “North American” diet. “People in Germany don't eat [the same] as people in England. And people in England don't eat as people in Provence.”
More importantly, he says, trying to apply mathematical formulas to cuisine is pointless, as cuisine is more art than science. “A good artist can make any mix, it will always be good,” he says, just as a world-class musician can do more with seemingly dissonant notes than the average atonal neophyte. On the Food Network program, Chopped, for instance, chefs must pair bizarre, seemingly impossible ingredients, and make them work. (One recent episode required contestants to fashion appetizers from pancake mix, strawberry papaya, blue foot mushrooms…and precooked chicken feet.)
“It's not a question of technique, it's a question of art,” says This. “And [for] the question of art, there are no rules, by definition.”